The synthetic NLR RGA5HMA5 requires multiple interfaces within and outside the integrated domain for effector recognition

Some plant sensor nucleotide-binding leucine-rich repeat (NLR) receptors detect pathogen effectors through their integrated domains (IDs). Rice RGA5 sensor NLR recognizes its corresponding effectors AVR-Pia and AVR1-CO39 from the blast fungus Magnaporthe oryzae through direct binding to its heavy metal-associated (HMA) ID to trigger the RGA4 helper NLR-dependent resistance in rice. Here, we report a mutant of RGA5 named RGA5HMA5 that confers complete resistance in transgenic rice plants to the M. oryzae strains expressing the noncorresponding effector AVR-PikD. RGA5HMA5 carries three engineered interfaces, two of which lie in the HMA ID and the other in the C-terminal Lys-rich stretch tailing the ID. However, RGA5 variants having one or two of the three interfaces, including replacing all the Lys residues with Glu residues in the Lys-rich stretch, failed to activate RGA4-dependent cell death of rice protoplasts. Altogether, this work demonstrates that sensor NLRs require a concerted action of multiple surfaces within and outside the IDs to both recognize effectors and activate helper NLR-mediated resistance, and has implications in structure-guided designing of sensor NLRs.

However, all the binding assays are performed without helper NLRs, only between RGA5 and effectors, which undermines the points about the HMA engineering effects on RGA4.Césari et al. report the RGA5 CC domain is required but not sufficient for repression of RGA4-mediated cell death, while the RATX1 domain (HMA domain) is dispensable (Césari et al., The EMBO Journal, 2014).In this manuscript, the authors find that RGA5-HMA5 can trigger RGA4-dependent plant cell death upon the perception of AVR-PikD.Does it mean the engineered HMA domain of RGA5 can interact and repress RGA4?Test whether RGA5-HMA/HMA3/HMA4/HMA5 can bind RGA4 with or without AVR-PikD would improve strongly on this point.The failure to trigger RGA4-dependent cell death may be caused by specific residues combination which affords the correct formation of RGA5-RGA4 heterodimer.2.Even though both Pik1-HMA/RGA5-HMA and AVR-PikD/AVR-COR39 are structurally similar, the =AF9D:579 B: 2=>%'&-43%2=>/ 5A8 30-+&-43%2=5 BD -43'%.13),5D9 8=::9D9AF "K(%K)%K* E<99F 5A8 J' <9?=H%K( =AF9D:579 D9EC97F=G9?I#$ 5A8 F<=E 6=A8=A; C5FF9DA 8=::9D9A79 @5I 69 D9?5F98 FB F<9 correct function of corresponding helper NLRs.So, engineering the residues in the RGA5/AVR-Pia or AVR1-COR39 for the RGA5/AVR-PikD interaction may bring in the complexity of results, e.g., RGA5-HMA3/HMA4 can bind to AVR-PikD but failed to trigger RGA4-dependent cell death.However, this interface difference may be used for engineering the receptor that can interact with both types of effectors.It will be exciting if the structure of RGA5-HMA with AVR-PikD or even with RGA4 can be solved.3. The three interfaces the authors identified are localized in four opposite parts of RGA-HMA5.Based on the structure of AVR-PikD, it is hard to imagine that AVR-PikD can interact with RGA-HMA all around at all sides.4. The authors should provide the peptide coverage and the independent sample number of RGA5-HMA5 in the HDX experiment.For HDX experiments, we can quickly figure out the regions that contribute to the interaction, but the authors should be cautious when claiming the role of a particular residue.The exchange rate of any overlapping adjacent peptides should be considered.5.The peptide (1030-1039 aa) in loop three identified in the HDX experiment, which corresponded to the interface of Pik1-HMA binding to AVR-PikD, is also participating in the homodimer formation of RGA5-HMA (as shown in RGA5-HMA structure and Xi et al., Mol Plant Pathol.2022).Is it possible that the different exchange rates of this peptide may come from the disruption of AVR-PikD on the formation of the homodimer of RGA5-HMA?6. Do any binding test results demonstrate that the decreasing binding affinity of Pikm1-HMA is caused by the "looping out" residue N261 in Pikp-HMA?Does the E1070 in RGA5-HMA have the same effects?Because Pikp-HMA can bind to AVR-PikD, and RGA5HMA5K/E can bind AVR-PikD but lose the capability to trigger RGA4, it looks like the key factor is the interference from the opposite charge of E1069 to K1070 in RGA5-HMA, not steric-hindrance effect.This can be further proved by the Pikp-1NK-KE, which also loses response to AVR-PikD.7. The HDX experiment results show that the two peptides in C-ter tail of HMA5 change from highly flexible to structured when the AVR-PikD is added, and this tail is crucial for derepressing RGA4 for triggering the rice cell death.Can this tail bind to AVR-PikD?Minor notes 1.In figure 1d, adding residue numbers in the alignment sequences may make the reader easier to find out corresponding residues.2. It's better to show the sequence of substituted residues of RGA5-HMA5K/E in Fig. 1 3.The contents of figure 2e are not elucidated clearly in the main text, which is mentioned in line 207, which refers to "The above results" in line 205, 4. In line 118, the PDB code of AVR-PikD and the complex of AVR1-CO39/RGA5-HMA used for structural superimposition should be provided.

Reviewer #3 (Remarks to the Author):
In this manuscript the authors successfully engineered paired NLR RGA4/RGA5a to recognize AVR-PikD which normally is not recognized by wildtype RGA4/RGA5a.The new RGA5<sup>HMA5</sup> is capable of binding AVR-PikD and then derepress RGA4 to induce defense to the blast fungus expressing AVR-PikD.To achieve this the authors had to engineer three interfaces including the one outside integrated domain.
Several studies demonstrating synthetic sensor NLRs with extended or altered effector recognition specificities have been already published including the one from the same group (https://doi.org/10.1073/pnas.2110751118) and others (https://doi.org/10.7554/eLife.47713).Also, natural resistance to AVR-PikD do exist which lessens the novelty of this manuscript.However, in this manuscript the authors have demonstrated that an interface outside of the integrated domain was crucial for the correct functioning of RGA4 which is a novel finding and may have broader impact in the future NLR engineering studies.
The authors should seriously consider generating transgenic Nipponbare expressing RGA4 with RGA5a, RGA5<sup>HMA2</sup> and RGA5<sup>HMA5</sup> altogether and test it against individual rice blast fungus strains which express AVR-Pia, AVR-Pib or AVR-PikD.Provided that such transgenic plant exhibit full resistance against all rice blast strains tested, it would mean that synthesizing and stacking novel resistance is a real possibility and thus have much broader impact.
Overall, the authors have provided enough evidence to support their claims although the details of experiments are sometimes lacking due to very limited methods section.For example, on page 15, the authors have stated that they have constructed <i>N.benthamiana</i> expression clones using pCAMBIA1305 and rice protoplast assay clones using pUC19 without identifying which promoters were used to drive the expression.I believe pCAMBIA1305 has 35S promoter but pUC19 does not have any plant promoters.Which promoter was used for rice protoplast luciferase assays?Also, qPCR assays do not describe how the comparisons were done.
Again, the results provided by the authors appears to be high quality and adequate to support their claims.However, the methods section is too brief to be reproduced stand-alone.The authors should utilize supplemental information section to describe their experimental methods in detail even if the methods used were previously published elsewhere.
Line 1: The title is too broad.The authors should change "sensor NLR receptors" to "sensor NLR RGA5<sup>HMA5</sup>", since the authors do not know if that statement is also true in Pik-1/Pik-2-based engineered NLRs.
Lines 22-23: Not all sensor NLRs have integrated domains nor are they paired NLRs.
Line 34: should remove the word "noncorresponding" Line 39: there is no such thing as "gene-for-gene diseases".Should change to "crop diseases".3. Please, include full length sequences of all synthetic constructs used in the study as supplemental information/dataset.Plasmid maps, if available, can be uploaded to Zenodo, and plasmids to Addgene where applicable.Please, include how these resources would be available to the community otherwise.
Response: We have made a supplementary dataset for full-length sequences of all the constructs, which has been uploaded to Zenodo (https://doi.org/10.5281/zenodo.8133144).

Include seed stock availability information.
Response: We generated five independent transgenic lines expressing RGA5 HMA5 with RGA4, which are stocked in our lab, but, at present, there are only a few seeds for these lines.However, we are propagating the seeds, which will be available to the scientific community upon direct request to the corresponding authors.revealed that all the RGA5 orthologs have a similar C-tail albeit evident difference between rice accessions.We also did a BLAST search of the RGA5 C-tail against the Nipponbare genome of rice and didn't find any significant hit, suggesting that the lysine-rich C tail is not present in any untranslated sequence.We generated the designer RGA5 HMA5 by using the cDNA of RGA5 from Kitaake, confirming that the lysine-rich C tail is not an untranslated sequence past the stop codon.
Nevertheless, We did pBLAST against the NCBI database and found that some non-integrated HMA proteins of rice also carry lysine-rich C-tails albeit highly diversified (Supplementary Fig S9h ), suggesting that the C-terminus of RGA5 may have originated from integration of an HMA protein.
Notably, a recent study reported that rice RGA4/RGA5 allelic NLRs have six distinct integrated domains at a similar position after the leucine-rich repeat domain, including Pias2 orthologs that have DUF761 domain for recognizing AVR-Pias, and suggested that the integrated domains are originated from integration of endogenous sequences matching the domains mediated by CID (conservation and association with IDs) motif (Shimizi et al., 2022, PNAS 119, e2116896119).Overall, this work provides a significant advance toward structure-guided designing of sensor NLRs with distinct specificity of effectors for directional crop breeding and also raises the complex effectorbinding and receptor-activation mechanisms behind NLRs.I would make the following suggestions to strengthen the current version of the manuscript.
Response: We sincerely thank the reviewer for the comments and suggestions.

Major comments
1.The authors conclude that some key residues in RGA5-HMA5 are crucial for derepressing RGA4.
However, all the binding assays are performed without helper NLRs, only between RGA5 and effectors, which undermines the points about the HMA engineering effects on RGA4.Césari et al. is crucial (required although not sufficient) for functional interaction between RGA5 and RGA4 and that the RATX1 (aka HMA) domain is not required in RGA4 repression but for AVR recognition.This conclusion doesn't exclude the regulatory role of the HMA domain of RGA5 in RGA4 derepression upon recognition of the MAX effectors.In nature, RGA4 is derepressed to trigger HR upon recognition of the MAX effector by RGA5 via its HMA.That is why, we did not investigate the impact of canonical domains within RGA5 on RGA4 repression but instead focused on the effects of resurfacing the HMA domain and the C-terminal tail following HMA to gain binding affinity with the noncorresponding effector AVR-PikD under the assumption that as long as there exists the interaction, RGA4-mediated cell death would occur.Therefore, the interaction assays between RGA5-HMA variants and the noncorresponding effector were performed without RGA4.However, we test the NLR variant RGA5 HMA5 with and without RGA4 for bioassays (Fig. 3d, e and f; Fig. 5c).Like the RATX1 domain in Cesari et al., we found that the engineered HMA domains alone do not repress RGA4.In addition, our results show that none of the RGA5-HMA, -HMA3, -HMA4 or-HMA5 domain interacted with RGA4 with or without AVR-PikD by co-IP assay in Fig S2 .Although RGA5-HMA3 and RGA5-HMA4 could bind to AVR-PikD, RGA5-HMA5, RGA5 HMA3 and RGA5 HMA4 failed to trigger RGA4-dependent cell death as RGA5 HMA5 .This discrepancy may not be caused by losing the ability to interact with RGA4.
We suppose that some amino acids in both the HMA domain and the C-terminal tail may be important to derepress RGA4 by causing allosteric changes (please refer to the description on RGA5 HMA3 , and RGA5 HMA4 , RGA5 HMA6 RGA5 HMA5K/E in the ms from lines 240-290).The structural mechanism that depresses the RGA4 activity will give the reasons for that.Our and other groups are still working to solve the complex structures of RGA4/5 with or without the effectors to understand the structural mechanism to depress the RGA4 activity.
Response: We fully agree with the comments.The structure of RGA5-HMA with AVR-PikD or even with RGA4 will give us solid data to explain the functional activity of the mutants designed by us and the Kroj's group.Unfortunately, we have been unable to obtain the crystal of the complexes, but we are still working on it.In the meantime, we here performed HDX-MS to analyze the interfaces of RGA5-HMA with AVR-PikD.The HDX-MS results suggested that multiple interfaces of RGA5-HMA5 are involved in the interaction with AVR-PikD, and the rice protoplast assays indicate that multiple interfaces but not only single interface can confer the RGA5 recognizing AVR-PikD to active the RGA4-dependent cell death.These indicate that both interfaces of Pik-1/AVR-PikD and RGA5/AVR-Pia or AVR1-COR39 in RAG5-HMA5 are indispensable for activating RGA4-dependent cell death.
We are currently optimizing the two interfaces and the C-tail to generate a RAG5 variant that can interact with both types of effectors and confer resistance to rice blast isolates that carry either of them.
3. The three interfaces the authors identified are localized in four opposite parts of RGA-HMA5.Based on the structure of AVR-PikD, it is hard to imagine that AVR-PikD can interact with RGA-HMA all around at all sides.
Response: Thank you for the comments.RGA5-HMA5 interacts with AVR-PikD by the engineered key residues on the interface 1 (RGA5-HMA/AVR-Pia) and interface 2 (Pikp-HMA/AVR-PikD), while the C-terminal tailing of RGA5-HMA on the interface 3 as well contribute to complex formation.We reckon that RGA5-HMA5 may bind AVR-PikD in multiple modes or the flexible loop in the N-terminus of AVR-PikD extending to the three interfaces in the RGA5-HMA domain may also be important for the interaction.
4. The authors should provide the peptide coverage and the independent sample number of RGA5-HMA5 in the HDX experiment.For HDX experiments, we can quickly figure out the regions that contribute to the interaction, but the authors should be cautious when claiming the role of a particular residue.The exchange rate of any overlapping adjacent peptides should be considered.
Response: Thank you for the suggestions.We performed the HDX-MS analyses in three independent experiments and obtained similar results.The peptide coverage was added in Supplementary Fig. S7.Six peptides, including overlapping adjacent peptides in RGA5-HMA5, decreased exchange upon binding of RGA5-HMA5 with C-tail by AVR-PikD.We verified two interaction surfaces by the Y2H assays and the LUC activity in rice protoplasts.Response: This is a good question.The interface (RGA5-HMA3) of the RGA5-HMA homodimer contributed to the interaction of engineered RGA5-HMA with AVR-PikD based on our Y2H results (Fig ) 9#& 7F$ .86%5AB0D9N <AIHKGJ J@= 23. @FDF<AD=H JF :AE< JF J@= AEJ=H>9;= F> J@= P) 9E< O( AE 61.,% HMA.
6. Do any binding test results demonstrate that the decreasing binding affinity of Pikm1-HMA is caused by the "looping out" residue N261 in Pikp-HMA?Does the E1070 in RGA5-HMA have the same effects?Because Pikp-HMA can bind to AVR-PikD, and RGA5HMA5K/E can bind AVR-PikD but lose the capability to trigger RGA4, it looks like the key factor is the interference from the opposite charge of E1069 to K1070 in RGA5-HMA, not steric-hindrance effect.This can be further proved by the Pikp-1NK-KE, which also loses response to AVR-PikD.
Response: Thank you for the comments.De la Concepcion et al. show that Pikm-HMA has tighter binding affinities for AVR-Pik effectors than Pikp-HMA due to the "looping out" residue N261 in Pikp-HMA (https://doi.org/10.1038/s41477-018-0194-x).Meanwhile, Pikp-1NK/KE, unlike RGA5-HMA that has a lysine-rich C-tail, can recognize AVR-PikD and mediate cell death in N. benthamiana (DOI: https://doi.org/10.7554/eLife.47713.001).RGA5-HMA could not bind to AVR-PikD, but E1070 deletion in RGA5 HMA4 , resulting in the forward shift of K1071, acquired binding affinity to AVR-PikD (Fig. 2 and 3), suggesting that the forward shift of K1071 is important for binding to AVR-PikD by RGA5-HMA4 and RGA5-HMA5.We further constructed a mutant RGA5-HMA6 carrying the E1070, which was almost the same as RGA5HMA5 except keeping the E1070 and could bind to AVR-PikD.However, RGA5 HMA6 lost the capability to activate RGA4, suggesting that the forward shift of K1071 is also required for RGA5 HMA5 to derepress RGA4 upon the AVR-PikD recognition.
7. The HDX experiment results show that the two peptides in C-ter tail of HMA5 change from highly flexible to structured when the AVR-PikD is added, and this tail is crucial for derepressing RGA4 for triggering the rice cell death.Can this tail bind to AVR-PikD?Response: Our HDX results indicate that some regions of the C-terminal tail are involved in the binding of RGA5-HMA5 to AVR-PikD.However, the Y2H assay could not detect the interaction between AVR-PikD and the C-terminal tail (Fig 5b).For this discrepancy, we speculate that the Cterminal tail may be involved in the binding of AVR-PikD after the AVR-Pia/AVR1-CO39 interface binds to AVR-PikD.
8. In figure 1d, adding residue numbers in the alignment sequences may make the reader easier to find out corresponding residues.
Response: We have added the residue numbers in Fig 1d .9. It's better to show the sequence of substituted residues of RGA5-HMA5K/E in Fig. 1 Response: Revised Fig. 1  Reviewer #3: In this manuscript the authors successfully engineered paired NLR RGA4/RGA5a to recognize AVR-PikD which normally is not recognized by wildtype RGA4/RGA5a.The new RGA5<sup>HMA5</sup> is capable of binding AVR-PikD and then derepress RGA4 to induce defense to the blast fungus expressing AVR-PikD.To achieve this the authors had to engineer three interfaces including the one outside integrated domain.
Several studies demonstrating synthetic sensor NLRs with extended or altered effector recognition specificities have been already published including the one from the same group (https://doi.org/10.1073/pnas.2110751118) and others (https://doi.org/10.7554/eLife.47713).Also, natural resistance to AVR-PikD do exist which lessens the novelty of this manuscript.However, in this manuscript the authors have demonstrated that an interface outside of the integrated domain was crucial for the correct functioning of RGA4 which is a novel finding and may have broader impact in the future NLR engineering studies.

Line 40 :
The reference #1 is in adequate.Should change to https://doi.org/10.1038/s41559-018-0793-yLine 50: "However, sensor NLRs usually carry" -this is true only in paired NLRs.ZAR1 is a sensor NLR which do not require any helper NLRs and do not contain integrated domain.Should change to "In paired NLRs, sensor NLRs usually" Line 57: references missing for Pik1 and RGA5 molecular engineering.Line 57: "in rice to confer" should change to "in rice that confer" Line 83: "designer sensor NLRs to trigger immune" should change to "designer sensor NLRs to trigger full immune" Lines 378-379: "Quik Change Site-Directed Mutagenesis Kit (Transgen)" I could not find company Transgen.Line 421: "yeast strain Y2H AH109" should change to "yeast strain AH109" AH109 should not be in italic font Line 443: GV3101 should not be in italic font Line 445: Tris-HCL should be Tris-HCl Line 447: "The homogenates were centrifuged" should describe the exact centrifugation settings Line 456: "ponceau" should be "Ponceau S" Responses to the reviewers' comments Reviewer #1: In this elegant NLR engineering study, authors rationally interchange several effector binding surfaces and uncover novel contribution of lysine rich C-terminal lysine rich 'tail' in RGA5 to the activation of the immune signaling through helper NLR, RGA4.The study combines genetics and biochemistry to dissect effector/receptor interactions and utilizes several orthogonal assays, including transient expression in Nicotiana benthamiana, rice protoplasts and transgenics to demonstrate effectiveness of synthetic NLR.Overall, I feel that the works is very polished already and deserves publication.All of the presented results are already well supported.My comments are mostly about data presentation, availability of reagents that should be required and additional evolutionary analyses that could inform the reader and put results uncovered in the specific receptor in a larger NLR biodiversity context.Response: We genuinely appreciate all of your valuable comments and suggestions.We have addressed all the concerns and suggestions in the revised version of the manuscript and made amendments as per suggestions.Please see the tracked changes in the revised version of the manuscript.Major suggestions/concerns: 1.For all figures please use box plots in place of bar graphs and include all individual data points overlayed on the graph.Response: Thank you for the suggestion!As per your suggestion, we have replaced the bar graphs with the box plots with overlayed data points obtained from three biological replications, which are color-coded.The changes were made in all the figures.2. For protoplast assays specifically, please, indicate if presented data is one experiments (if so, how many technical replicates) or an aggregate of all three experiments.Please, see Saur et al 2019 (https://doi.org/10.1186/s13007-019-0502-0)for an example of more transparent data presentation that I am looking for.Response: The data presented was obtained from three independent experiments, with three technical replicates per experiment.The three experiments are coded by distinct colors in the box plots.
5. Figure4(especially images of leaves) are very difficult to see.I would suggest to provide a larger version of the images/graphs.Response: We do have high-resolution TIFF images included in Figure4, and will furnish these TIFF images.6.Currently, the phylogenetic distribution of lysine rich tail which is one of the most exciting discoveries presented in this work is unclear.Is it presented in any other RGA5 orthologs?Is it present as an untranslated sequence past their stop codons?Is it present in any non-integrated HMA proteins?I think a simple blast search both in protein space but also in protein against genomic DNA space can reveal the origin of this functionally important region.Response:These are excellent questions!As reported in this study, the C-terminal lysine-rich tail immediately after RGA5-HMA plays a crucial role in regulating RGA4-dependent cell death.Following your suggestion, we downloaded RGA5 orthologs in seven rice accessions that have paired RGA4/RGA5 orthologs, including cultivated and wild species of rice published in a recent study(Shang et al., 2022, Cell Research 32:878-896) and made an alignment of their amino acid sequences of the RGA5 orthologs, excludingthe Pias2 orthologs that have the DUF761 domain instead of the HMA domain as in RGA5, and presented the alignment in the new supplementary Fig S9a.The comparison

7 .
A more complex analysis would be to test co-evolution of the C-terminal tail with the rest of RGA5 protein (are engineered surfaces and the lysine rich region show signature of co-evolution?) and with RGA4.I would understand if authors do not have expertise in this area and although this would greatly enhance the work, I don't see it as a requirement for publication.Response: Thank you for the suggestion!To investigate the coevolution between the C-tail and the rest of the domains of RGA5, we have performed simple domain-specific phylogenetic analyses of CC, NB-ARC, LRR, HMA and C-tail domains.The analyses, as shown in Supplementary Fig S9 c to g, suggest that the variations in the C-tail of RGA5 orthologs may be coevolved with the HMA domain but not the rest of the domains.Reviewer #2: The manuscript by Zhang et al. engineers four RGA5-HMA mutants based on the structure information and finds out one of the mutants confers complete resistance in transgenic rice plants to the M. oryzae strains expressing the noncorresponding effector AVR-PikD.The authors identify three interfaces bound by AVR-PikD in the RGA5HMA5 through HDX-MS and biochemistry assays.This work demonstrates that RGA5HMA5 requires multiple surfaces within and outside the IDs to recognize noncorresponding effectors and activate helper NLR-mediated resistance.
report the RGA5 CC domain is required but not sufficient for repression of RGA4-mediated cell death, while the RATX1 domain (HMA domain) is dispensable (Césari et al., The EMBO Journal, 2014).In this manuscript, the authors find that RGA5-HMA5 can trigger RGA4-dependent plant cell death upon the perception of AVR-PikD.Does it mean the engineered HMA domain of RGA5 can interact and repress RGA4?Test whether RGA5-HMA/HMA3/HMA4/HMA5 can bind RGA4 with or without AVR-PikD would improve strongly on this point.The failure to trigger RGA4-dependent cell death may be caused by specific residues combination which affords the correct formation of RGA5-RGA4 heterodimer.Response: This is a critical point.Césari et al. (2014), in their EMBO journal paper, performed the domain deletion analysis of RGA5 to delimit the role of individual domains thereof in effector-independent RGA4mediated cell death in Nicotiana benthamiana.The authors concluded that the CC domain of RGA5 5. The peptide (1030-1039 aa) in loop three identified in the HDX experiment, which corresponded to the interface of Pik1-HMA binding to AVR-PikD, is also participating in the homodimer formation of RGA5-HMA (as shown in RGA5-HMA structure and Xi et al., Mol Plant Pathol.2022).Is it possible that the different exchange rates of this peptide may come from the disruption of AVR-PikD on the formation of the homodimer of RGA5-HMA?
now shows the complete sequence of RGA5-HMA5.10.The contents of figure 2e are not elucidated clearly in the main text, which is mentioned in line 207, which refers to "The above results" in line 205.Response: Thank you!We have added the detail in the revised text to explain Fig 2e, which reads as follows: "To test the function of the three designer RGA5 mutants in rice, Oryza sativa cv.Nipponbare protoplasts were transfected with a combination NLRs and AvrPikD along with luciferase, which showed that the expression of RGA4 or RGA4/RGA5/AVR-Pia and RGA4/RGA5 HMA5 /AVR-PikD led to a significant reduction in luciferase reporter activity, compared with other combinations namely RGA5, RGA5 HMA5 , AVR-PikD, RGA4/RGA5, RGA4/RGA5 HMA3 , RGA4/RGA5 HMA5 , RGA4/RGA5 HMA3 /AVR-PikD and RGA4/RGA5 HMA4 /AVR-PikD (Fig. 2e, 3f).The above results indicate that only RGA5 HMA5 but not RGA5 HMA3 and RGA5 HMA4 could cause RGA4-mediated cell death in rice protoplasts upon AVR-PikD recognition ."11.In line 118, the PDB code of AVR-PikD and the complex of AVR1-CO39/RGA5-HMA used for structural superimposition should be provided.Response: Thank you!The PDB code of AVR-PikD (PDBID: 5A6W) with the complex of AVR1-CO39/RGA5-HMA (PDBID: 5ZNG) is added in line 118.