The Cotton Wall-associated Kinase GhWAK7A Mediates Responses to Fungal Wilt Pathogens by Complexing with the Chitin Sensory Receptors

Review timeline: TPC2019-00867-RA Submission received: Nov. 5, 2019 1st Decision: Nov. 18, 2019 manuscript declined TPC2019-00950-RA Submission received: Feb. 24, 2020 1st Decision: April 3, 2020 revision requested TPC2019-00950-RAR1 1st Revision received: June 26, 2020 2nd Decision: July 27, 2020 revision requested TPC2019-00950-RAR2 2nd Revision received: Aug. 12, 2020 3rd Decision: Sept. 6, 2020 acceptance pending, sent to science editor Final acceptance: Sept. 30, 2020 Advance publication: Oct. 9, 2020


TPC2019-00950-RA 1 st Editorial decision -revision requested
April 3, 2020 We have received reviews of your manuscript entitled "Cotton wall-associated kinase GhWAK7A mediates responses to fungal wilt pathogens by modulating chitin sensory complex." Thank you for submitting your best work to The Plant Cell. The editorial board agrees that the work you describe is substantive, falls within the scope of the journal, and may become acceptable for publication, pending revision and re-review.
We ask you to pay special attention to the following points in preparing your revision: A major concern raised by 2 reviewers is that insufficient evidence is presented to support the involvement of GhWAK7 in regulating the chitin-mediated defense against verticillium and Fusarium . They provide several suggestions on the data could be strengthened so that they better substantiate the claims made. This consists largely of quantifying data presented in several of the figures (see detailed comments from Reviewer 1). A second main concern regards the specificity of WAK7 function: does it regulate the chitin-mediated defense response, rather than recognition of an unknown ligand. This should be more thoroughly discussed. Reviewer comments on previous submission and author responses: We thank the editors and reviewers for their thorough review of our manuscript and insightful comments, to which we respond point-by-point in the following. We appreciate the reviewers for their time and effort in raising these constructive comments, which help us significantly improve our manuscript. We have carefully taken these comments into account by performing additional experiments and modifying the manuscript textually.

Editors' comments:
We ask you to pay special attention to the following points in preparing your revision: 1. A major concern raised by Reviewers 1 and 2 is that insufficient evidence is presented to support the involvement of GhWAK7 in regulating the chitin-mediated defense against verticillium and Fusarium. It is essential that you address these concerns. The reviewers provide several suggestions for how the existing data could be strengthened, so that they better substantiate the claims made. In several instances quantifying the data already presented may be sufficient, however addressing other concerns may require additional experimentation (e.g. Kinase assays shown in Fig 6-see Reviewer 2 "minor comment" 13).
Our response: We thank the editor for the suggestions. Following the suggestions from Reviewers 1 and 2, we have provided quantification data of the relative intensities of immunoblot bands from three independent biological repeats for the chitin-induced MAPK activation in VIGS-GhWAK7A cotton plants (New Figure 3C), and VIGS-GhMPK6 and -GhMPK3 cotton plants (New Supplemental Figure 9B), and for the effect of GhWAK7A on the association of GhCERK1 and GhLYK5 (bottom panels in Figure 8A and B).
As suggested by Reviewer 2, we have performed the in vivo kinase assays using anti-phosphoserine/threonine antibodies and the result indicated that GhWAK7A enhanced, whereas the kinase-inactive GhWAK7A K451M reduced the phosphorylation of GhLYK5 (New Figure 8E). In addition, we have carried out an in vitro kinase assay with a reduced protein amount of GhWAK7A CD as the kinase and additional controls. The new results further supported the previous observation that GhWAK7A phosphorylated GhLYK5 in vitro (New Figure 8C, D and Supplemental Figure 17C). Moreover, we have performed the split-luciferase assay to further validate the in vivo association between GhWAK7A and GhLYK5 in cotton (New Figure 7C).

2.
A second concern that must be addressed is the specificity of WAK7. The data providing evidence that it regulates the chitin-mediated defense response, rather than recognition of an unknown ligand needs to be more carefully documented (see Reviewer 2, Major comment 2).
Our response: Our data showed that GhWAK7A interacted with the GhLYK5/GhCERK1 chitin sensing complex and regulated chitin signaling. The chitin binding assays showed that GhLYK5 and GhCERK1, but not GhWAK7A, bind to chitin (New Figure 6F). During the revision, we have addressed the specificity of GhWAK7 in chitin signaling via different ways. We first showed that GhWAK7 is not involved in cotton drought and salt responses as concerned by Reviewer 1 (New Supplemental Figure 8A and B). Secondly, considering that WAKs have been indicated in sensing OGs in Arabidopsis (Kohorn and Kohorn, 2012), one possible ligand for GhWAK7 is OGs. In the revised manuscript, we have strengthened our observation that GhWAK7 is not involved in OG signaling.
In addition, we have been collaborating with Dr. Simone Ferrari at Sapienza Università di Roma working on OGtriggered immune responses, to show the purity of OGs during preparation and characterize their responses in Arabidopsis and cotton (New Supplemental Figure 10A-F, Figure 3G and H). Finally, we tested whether GhWAK7A is involved in mediating signaling triggered by cell wall extracts from Fusarium oxysporum (FovCWE). FovCWE triggers PTI responses in cotton and its biologically active component is unlikely to be chitin (manuscript in preparation). The new result indicated that silencing GhWAK7A did not affect FovCWE-induced GhMAPK activation (New Figure 3I). Taken together, the data implied that GhWAK7A may not be involved in OG-and FovCWE-triggered responses, but is required for chitin-triggered immune responses. However, we agree with the reviewers that we cannot completely rule out the possibility that GhWAK7A might be involved in the perception or signaling of an unknown ligand, which we have added in the Discussion (Line 664-671).
3. Please also seriously consider the recommendation of Reviewer 3 to include a model figure, as we agree that this would be helpful to interpret the results.
Our response: We thank the editor and this reviewer for the suggestion. We have included a working model to summarize the potential role of GhCERK1, GhLYK5, and GhWAK7A in response to fungal infections in cotton ( Figure 8F) and have discussed this model in Discussion (Line 589-595) as suggested by Reviewer 3.

Reviewer #1:
Wang et al present an extensive analysis of the response of cotton to chitin fragments, including an analysis of the LYK and WAK cotton families and some of their activities. The work is fairly well presented but while the data do indicate that CERK and LYK mediate the chitin response (as well documented in another species), the data fall short of convincing the reader that WAKs are involved in the process.
None of the data provide evidence that the cotton WAKs are indeed wall associated. Indeed, most of the biochemical and molecular analysis is in vitro or in cultured cells. As such the authors should call these genes/proteins WAK-like or WAKLs.
Our response: We thank this reviewer for the comments and suggestions. During the revision, we have performed additional experiments, including the split-luciferase assay (New Figure 7C), in vitro kinase assay (New Figure 8C, D, and Supplemental Figure 17C), and in vivo kinase assay (New Figure 8E), and quantified our data from three independent biological repeats (New Figure 3C, Bottom panels of Figure 8A and B, Supplemental Figure 9B) to further strengthen the involvement of GhWAK7A in chitin responses (please see details below). We agree with this reviewer that our data do not provide evidence that GhWAKs are wall-associated and appreciate the suggestion to call them WAK-like proteins (WAKLs Our study here is mainly focused on the genome-wide identification of cotton GhWAK genes. We eliminated the GhWAKL genes via three criteria. Firstly, the five AtWAK proteins were used as queries to blast against three cotton genomes. Five AtWAKs are specific and different from the AtWAKLs in the extracellular domain configuration (Verica and He, 2002) and clustered separately based on the phylogenic analysis (Response Figure  1). Pairwise comparisons of the WAKLs and WAKs showed that they are only 18% to 22% identical in their extracellular regions (Verica et al., 2003). Secondly, the protein list we obtained from the blast search was further amended according to the protein annotation in the cotton genome to eliminate the non-GhWAK genes including GhWAKL genes. Finally, we manually checked the protein motifs of each GhWAK protein candidates and only the proteins that contain all the motifs present in AtWAKs including an extracellular signal peptide, WAK-GUB, EGF motif, EGF-Ca 2+ motif, and intercellular kinase domain, were assigned as cotton GhWAK proteins (Supplemental  Tables 1, 2 and 3).
In addition, we performed the phylogenetic analysis of cotton GhWAKs, AtWAKs, and AtWAKLs (except AtWAKL7, 12, 14, and 16, which are not canonical RLKs as illustrated in Response Table 1). The phylogeny revealed that cotton WAKs we identified in this study are in the same clade with AtWAKs, not AtWAKLs (New Supplemental Figure 1), demonstrating that the proteins we identified from cotton in this study are WAKs but not WAKLs. Therefore, to follow the gene nomenclature of WAKs and WAKLs, and avoid confusion, we have named these cotton candidates as GhWAKs. We have included these descriptions in the revision (Line 161-164). The effect of WAKL VIGS on pathogen infection is small but as presented significant. But since the effect is small it is possible that any stress and not just these two pathogens would cause some symptoms of VIGS WAKL. Indeed the past literature predicts this, and raises the question about the specificity of the observed response in Fig  2. The specificity is also further questioned but the comments below concerning the molecular experiments.  Figure 2H). The statistical analysis of disease index using one-way ANOVA indicated a significant difference between control and VIGS-GhWAK7A plants for both Vd and Fov infections. We also included the closeup images of plant leaves after infections (Middle panels in New Figure 2E and G).

Response
To address the concern whether any other stresses may contribute to disease symptom development of VIGS-GhWAK7A plants, we have performed drought treatment in soil-grown cotton plants (New Supplemental Figure 8A) and salt stress treatment with 200 mM NaCl in hydroponic-grown cotton plants (New Supplemental Figure 8B). VIGS-GhWAK7A plants showed similar wilting rates as the vector control plants in response to the water withdrawal treatment (New Supplemental Figure 8A). In addition, in response to the treatment of 200 mM NaCl, both control and VIGS-GhWAK7A plants showed a similar level of stunt growth with chlorosis, yellowish and wilting leaves (New Supplemental Figure 8B). The new results indicate that silencing GhWAK7A does not show altered responses to drought and salt stresses, supporting the specificity of GhWAK7A in cotton response to Vd and Fov infections. Fig 3B: The authors claim that there is a reduction of MAPK6 and MPK3 in VIGS WAKL7 plants treated with chitin. The reduction in MPK3 is minor if any at all, while indeed there is a reduction in MPK6. The authors a state the experiment was done at least 3 times (bottom of legend) but no quantitation of these triplicates is provided and should be. Also essential is a western of total MPK6 and MPK3 protein levels to be sure that these are not changing.
Our response: We thank this reviewer for the suggestion. The reduction of GhMPK6 and GhMPK3 in VIGS-GhWAK7A plants treated with chitin has been consistently observed in multiple independent repeats. During the revision, we have repeated this experiment again and observed a similar trend (New Figure 3B). We also have examined the GhMPK6 and GhMPK3 protein levels by immunoblots using anti-MPK6 and anti-MPK3 antibodies ( Figure 3B, third and fourth panels). The new data suggested that VIGS-GhWAK7A or GhWAK5A did not affect the GhMPK6 and GhMPK3 protein levels. In addition, quantification of the relative intensities of immunoblot bands of phosphorylated GhMPK3/6 relative to the protein inputs from three independent repeats was included to show the compromised GhMPK6 and GhMPK3 phosphorylation in VIGS-GhWAK7A plants treated with chitin (New Figure  3C).

Fig 3G:
The authors claim that WAKL VIGS does not affect OG induced MAPK activity. But the authors do not provide in methods how the OGs were produced, and indeed suggest that they are using a crude extract of treated leaves; these are not pure "OGs" and likely include multiple inducers of events that could be activating MAPKs independent of the OG response.
Our response: Thanks for pointing this out. We have been collaborating with Dr. Simone Ferrari at Sapienza Università di Roma, working on OG-triggered immune responses in plants, on the preparation and characterization of OGs. OGs used in this study were generated by the partial digestion of polygalacturonic acid (Na + salt) with homogeneous Aspergillus niger endopolygalacturonase. The qRT-PCR data showed that the induction of AtFRK1 by OGs was similar to that by flg22 in Arabidopsis (New Supplemental Figure 10A). Size-homogeneous OGs were analyzed by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). The data indicated OG oligomers with a distribution of the degree of polymerization (DP) between 6 and 20 and an enrichment in the DP around 12 (New Supplemental Figure 10B). This is consistent with the analysis using matrix-assisted laser desorption/ionization-time of flight-mass spectrometry (MALDI-TOF-MS) (New Supplemental Figure 10C). Cotton leaves treated with OGs triggered rapid GhMPK6 and GhMPK3 activation with a peak at 15 min (New Supplemental Figure 10D), ROS production (New Supplemental Figure 10E), and transcriptional upregulation of GhWRKY30 and GhMPK3 at 30 min in cotton (New Supplemental Figure 10F).
In addition to GhMAPK activation, we tested whether silencing GhWAK7A affected the OG-triggered ROS production and the expression of GhWRKY30 and GhMPK3. Similar to its response to OG-triggered GhMAPK activation ( Figure 3F), silencing GhWAK7A did not affect OG-triggered ROS production or gene expression of GhWRKY30 and GhMPK3 (New Figure 3G and H).

Fig 6G/H: Chitin indeed induces a LYK/CERK association in vitro
, but the WAKL effect on this seems marginal at best. And the interpretation of this gel is very hard given the inconsistent levels of the various receptors, especially of CERK. In 6H it is claimed that the WAKL kinase mutant does not increase LYK/CERK association, and this is a more robust results, yet as in G, the results need to be quantified relative to an endogenous protein, or perhaps as ratios of standardized receptor amounts. If the results were more dramatic then they might be more acceptable, but as its, since the results seems marginal, they must be accurately quantified.
Our response: We thank this reviewer for the suggestion. The original Fig. 6G and H are now Fig 8A and B in the revision. We have consistently observed an enhanced chitin-triggered association of GhLYK5 and GhCERK1 in the presence of GhWAK7A in multiple repeats ( Figure 8A). Notably, GhWAK7A K451M , the kinase-inactive mutant of GhWAK7A, appeared to have a much-pronounced effect in reducing the chitin-induced association of GhLYK5 and GhCERK1, likely due to its potential dominant-negative effect ( Figure 8B). We have followed this reviewer's suggestions and performed the quantification analysis from three independent repeats, and the new data are presented in the bottom panels of Figures 8A and B. The data represent the densitometry units of immunoprecipitated GhLYK5 normalized to the total GhLYK5 proteins before immunoprecipitation.
6I. The gel shows that LYK phosphorylates CERK (not sure why there are 2 bands) and that WAKL7 phosphorylates itself. However, the authors claim that the band in the WAKL kinase lanw(3rd from left) is LYK I assume based on the mw. How do they know that this is not instead a degradation product of the phosphorylated WAKL that happens to co-migrate (actually it is a little lower in mw) than the LYK?
Our response: To address this reviewer's concern, we have performed additional in vitro kinase assays with newly purified proteins and additional controls (New Figures 8C and D, and Supplemental Figure 17C). It is clear that GhLYK5 CD had an undetectable kinase activity (Lane 4 in Figure 8C; Lane 1 in Figure 8D), both GhWAK7A CD and GhCERK1 CD had auto-phosphorylation activity (Lane 2 in Figure 8C for GhWAK7A CD , and Lane 3 in Figure 8D, Lane 1 in Supplemental Figure 17C for GhCERK1 CD ). Figure 8C showed that GhWAK7A CD but not GhWAK7A K451M-CD phosphorylated GST-GhLYK5 CD (compare Lane 3 and Lane 1). Notably, we did not observe a degradation product of phosphorylated GhWAK7A at the position of GST-GhLYK5 CD . Figure 8D showed that the phosphorylation pattern of MBP-GhCERK1 CD did not change in the presence of GST (Lane 3) or GST-GhLYK5 CD (Lane 2), suggesting that GhCERK1 CD did not phosphorylate GhLYK5 CD . Furthermore, in the Supplemental Figure 17C, in which GST-GhCERK1 CD was used as the kinase in the in vitro kinase assay, GST-GhCERK1 CD did not increase the phosphorylation of GST-GhLYK5 CD (Lane 2) compared to GST control (Lane 1).
Given that the WAKL induced LYK/CERK binding is not well documented, that the phosphorylation of LYK by WAKL is questionable, and lastly that there is no in vivo analysis of these associations and phosphorylation, the manuscript does not adequately support the involvement of WAKL in the LYK CERK/chitin mechanism.
Our response: We appreciate this reviewer's comments, for most of which we have addressed in the previous points, such as GhLYK5 phosphorylation by GhWAK7A. To our knowledge, this is the first report about the characterization of cotton LYKs/CERK1 in chitin signaling and in response to two devastating cotton wilt diseases, and the involvement of WAK in LYK/CERK1-mediated chitin signaling in plants. We have carefully discussed this and performed extensive new experiments to strengthen our conclusions. In addition to previously performed in vivo Co-IP and FRET-FLIM assays for GhWAK7A and GhLYK5/GhCERK1 associations, we have performed the splitluciferase assays to further show the association of GhWAK7A and GhLYK5 in vivo (New Figure 7C). We have also shown that, unlike GhLYK5 and GhCERK1, GhWAK7A might not directly bind to chitin (New Figure 6F).
We also performed the in vivo phosphorylation assay with anti-phosphoserine/threonine antibody (α-pS/T, ECM Bioscience) (New Figure 8E). GhLYK5-FLAG was co-expressed with vector control, GhWAK7A-HA, or GhWAK7A K541M -HA in cotton protoplasts with or without chitin treatment, and the phosphorylated GhLYK5-FLAG proteins after immunoprecipitation using FLAG agaroses were detected by immune-blots using antiphosphoserine/threonine antibody. The result indicated that GhWAK7A enhanced, whereas GhWAK7A K451M reduced the phosphorylation of GhLYK5 (New Figure 8E). We hope that the new data about in vitro phosphorylation of GhLYK5 by GhWAK7A, and in vivo association and phosphorylation between GhWAK7A and GhLYK5 have strengthened our conclusion about the involvement of GhWAK7A in GhLYK5/GhCERK1-mediated chitin signaling in cotton.

Minor points
Line 86-EGFs do not contribute to ligand binding specificity. Line 113 reference is incorrect-this paper does not show wall binding. Reviewer #2: Wang et al. present the identification of cotton WAK and LYK receptor kinase isoforms and their potential roles in anti-fungal immunity. After identifying family members in cotton through careful bioinformatic and phylogenetic analyses, the authors use a combination of published and novel transcriptional data to identify GhWAK7A as being potentially involved in resistance to fungal wilt pathogens. The authors eventually suggest that GhWAK7A functions via regulation of the chitin-perceiving CERK/LYK4/5 complex.
While many of the findings are of interest and supported by their data, the authors do not convincingly demonstrate that WAK7A functions primarily via regulation of chitin responses, instead of recognition of an unknown ligand. There remain several issues with the manuscript that must be resolved before publication in a top-tier journal such as The Plant Cell would be suitable, which are outlined below.
Our response: We thank this reviewer for the concise summary and recognition of our comprehensive work. To address the constructive comments that this reviewer raised, we have performed additional experiments to strengthen the involvement and specificity of GhWAK7A in chitin-mediated responses (see below).
Major comments: 1. Figure 1, Supp. Figure 1, and Methods: -Which part of the protein was used for the construction of the phylogenetic tree? Full length CDS or a specific domain? Do they cluster differently if only use kinase domain?
Our response: The full-length CDS protein sequences were used for the construction of the phylogenetic trees in the original manuscript. Following this reviewer's suggestion, we have performed additional phylogenetic analysis using the kinase domains (New Supplemental Figure 4). Notably, the phylogeny with regard to the kinase domains revealed a similar pattern as that using the full-length sequences and that none of cotton WAKs from G. hirsutum, G. arboreum, or G. raimondii is interleaved to AtWAKs (New Supplemental Figure 4).
-Why are Ga03G0830.1 and Ga02G01816 in green cluster? Cluster closer to At cluster?
Our response: We apologize that we did not make this clear. From the phylogenic analysis of GaWAKs and AtWAKs, Ga03G0830.1 and Ga02G01816 belong to the same clade with other GaWAKs, in relative to the isolated clade of the five AtWAKs (Supplemental Figure 3B). These two genes are the paralogs of Ga03G0831.1, which is the ortholog of GhWAK10A from the G. hirsutum A subgenome (Supplemental Figure 5A) and Gorai.005G086700.1 from the G. raimondii genome (Supplemental Figure 5C). Since Ga03G0831.1, GhWAK10A, and Gorai.005G086700.1 were assigned as Clade V (green cluster), Ga03G0830.1, and Ga02G01816 were marked with green and assigned to Clade V of cotton WAKs (Supplemental Figure 2).
-It is not clear to me why the authors have only used the annotated AtWAKs and not the AtWAK-likes -are any of the GhWAKs actually more closely related to the AtWAKLs? Does Gh have WAKLs, and how do they cluster in the tree?
Our response: We thank this reviewer for pointing this out. As in response to the concern raised by Reviewer  Table 1).
Our study here is mainly focused on the genome-wide identification of cotton GhWAK genes. We eliminated the GhWAKL genes via three criteria. Firstly, only five AtWAK proteins were used as a query to blast against three cotton genomes. Secondly, the protein list we got from the blast search was further amended according to the protein annotation in the cotton genome to eliminate the non-GhWAK genes including GhWAKL genes. Finally, we manually checked the protein motifs of each GhWAK protein candidates and only the proteins that contain all the motifs essential for AtWAKs including an extracellular signal peptide, WAK-GUB, EGF motif, EGF-Ca 2+ motif, and intercellular kinase domain, were assigned as cotton GhWAK proteins (Supplemental Tables 1, 2 and 3). In addition, we performed the phylogenetic analysis of cotton GhWAKs, AtWAKs, and AtWAKLs (except WAKL7, WAKL12, WAKL16, and WAKL14). The phylogeny revealed that cotton GhWAKs we identified are closer to AtWAKs than AtWAKLs (New Supplemental Figure 1), demonstrating the proteins we identified from cotton in this study are WAKs but not WAKLs.
2. The authors test WAK7A VIGS for altered OG perception by MAPK assay, and from this conclude that WAK7A does not perceive OGs, which is fine. However, they cannot rule out that WAK7A perceives a different ligand to regulate immunity. This should be at least discussed. Furthermore, the OG experiment lacks a control which actually does show compromised OG perception. Also, do OGs also induce WRKY30 and/or MPK3 expression, and is this affected by WAK7A VIGS?
Our response: Thanks for the suggestions. We agree with this reviewer that we cannot completely rule out the possibility that GhWAK7A might be involved in the perception or signaling of an unknown ligand, for which we have added in the Discussion (Line 664-671). However, we have also tried our best to address that GhWAK7A is not involved in the perception or signaling of OGs or an unknown elicitor from Fusarium cell wall extracts. It has been shown that AtWAKs are associated with the perception of oligosaccharide ligands (Kohorn and Kohorn, 2012). We have been collaborating with Dr. Simone Ferrari at Sapienza Università di Roma, working on OG-triggered immune responses in plants, on the preparation and characterization of OGs. OGs used in this study were generated by the partial digestion of polygalacturonic acid (Na + salt) with homogeneous endopolygalacturonase from Aspergillus niger. The qRT-PCR data showed that the induction of AtFRK1 by OGs was similar to that by flg22 in Arabidopsis (New Supplemental Figure 10A). Figure 10C). Figure 10D), ROS production (New Supplemental Figure 10E), and transcriptional upregulation of GhWRKY30 and GhMPK3 at 30 min in cotton (New Supplemental Figure 10F). Silencing GhWAK7A by VIGS did not alter OG-induced ROS production, MAPK activation, or expression of GhWRKY30 and GhMPK3 ( Figure 3F, New Figures 3G and H). Furthermore, our lab has been characterizing the cell wall extracts from Fusarium oxysporum (FovCWE) that trigger GhMAPK activation in cotton yet their bioactive components are likely not chitin (manuscript in preparation). Our new data indicated that silencing GhWAK7A did not affect FovCWE-induced MAPK activation (New Figure 3I). Taken together, the data indicated that GhWAK7A may not be involved in OG and FovCWEtriggered responses but is required for chitin-triggered immune responses via interaction with the chitin-sensing complex GhLYK5 and GhCERK1.

How do the authors conceptually explain that GhWAK7A would regulate chitin-induced complex formation between
GhCERK1 and GhLYK5 through phosphorylation of the intracellular domain of GhLYK5 while this complex formation is actually mediated extracellularly through ligand-binding. This is at odd with current models of how extracellular ligands induce heteromerization of receptor/co-receptor complexes (Hohmann et al., Annu. Rev. Plant Biol. 2017), and certainly requires some discussion.
Our response: We thank this reviewer for bringing this point out for discussion. We concur with this reviewer that current models elucidate the ligand-induced heteromerization of receptor/co-receptor complexes mainly based on the structural studies using the extracellular domains. However, evidence also indicates the potential involvement of intracellular domains in receptor/co-receptor complex formation, especially in vivo. For example, the kinase activity of BRI1 is strictly required for the ligand-induced BRI1-BAK1 complex formation ( Our response: Thanks for pointing this out. We have included this reference and added the statement as "AtLYM2 has been shown to regulate chitin-induced stomatal closure in Arabidopsis".
3. Introduction: line 143: it is more correct to write that "GhWAK7A, GhCERK1, and GHLYK5 contribute to cotton resistance to Fov and Vd..." Our response: We have changed it as suggested.
4. Figure 2E: more wilting leaves not really obvious from those pictures. Higher resolution/closeup leaf images would be helpful.
Our response: Thanks for the suggestion. We changed the images in Figure 2E and included a close-up view of leaf images to show a clear disease phenotype. Figure 2F and lines 338-341: from the data presented it is not correct to state that GhWAK5A plays an opposite role to GhWAK7A in cotton resistance against Vd, as the disease index of the GhWAK5A VIGS plants is comparable to that of the control plants.

5.
Our response: Thanks for pointing this out. The disease index of Vd-infected GhWAK5A-silenced plants did not show a significant difference compared to that of the control plants. Thus, we have deleted this statement in the revised manuscript. Figure 2A and 4A -how do these data look if values are calculated as relative to mock? It is difficult to interpret as is, given that many of the genes show high expression on the heatmap even at 6h mock.

Gene expression heatmaps in
Our response: In the RNA-seq datasets, the FPKM values of most GhWAKs were relatively low with either 0 or less than 1. However, a few of them, including GhWAK5D, GhWAK6D, GhWAK7A, GhWAK8A, GhWAK9D, and GhWAK10A, had relatively high FPKM values (Response Table 2). Thus, we have presented the gene expression heatmaps with the original FPKM values. To clearly present the data, we changed the color schemes of the graph to the FPKM values with the colored blocks (New Figure 2A and 5A).

We have changed the description as "The activation of GhMPK3 and GhMPK6 was confirmed with VIGS-GhMPK3 or VIGS-GhMPK6 plants in which the protein expression levels of GhMPK3 or GhMPK6 were reduced".
8. Figure 4C and 4E: I agree that the discoloring of the stem is darker in GhCERK1, and for the leaves GhCERK1 also shows the strongest phenotype, but the rest also looks affected; also seen in disease index even if not significant; E: foliar no difference, stem yes for CERK1 and LYK5 Our response: We agree with this reviewer that VIGS-GhCERK1 plants in response to Vd and Fov infections showed the most pronounced susceptibility compared to VIGS-Ctrl or other LYK plants. To better monitor the disease phenotypes in cotton upon the Vd and Fov infections, we have included multiple evaluation parameters including disease index recording, phenotyping of leaf wilting and chlorosis, and stem staining, and performed multiple independent biological repeats. We have done seven repeats for most of VIGS-GhLYKs plants to Vd infections (Supplemental Table 13) and six repeats for VIGS-GhLYKs plants to Fov infections (Supplement Table  14). 9. Figure 5 B: Colors would be much easier to distinguish.

For
Our response: The graph has modified with colors (Original Figure 5B; Figure 6C in the revised manuscript) as suggested.
10. Line 608: how was chitin mixture "used as beads" to pull down the complex? No details are provided in the methods (see Figure 5D).
Our response: Thanks for pointing this out. The chitin-magnetic beads (New England Biolabs) were used for the chitin binding assay (Original Figure 5D; Figure 6F in the revised manuscript) and the detailed information and protocol have been included in Methods.
11. Line 609-611: these conclusions should be toned down as there is less GhWAK7A protein in the input.
Our response: Thanks for pointing this out. We have modified the conclusion that GhWAK7A might not bind chitin. We have repeated this assay more than three times with similar results. We showed another representative data with a comparable protein level of GhWAK7A with GhCERK1 and GhLYK5 (New Figure 6F).
12. Figure 5F: what is the authors explanation for the extremely uniform and saturated shift in their phostag gel for LYK5?
Our response: Using the phos-tag gel-based immuno-blots, we have detected the markedly up-shifted migration band of GhLYK5 upon chitin treatment (Original Figure 5F; Figure 6H in the revised manuscript), suggesting that perception of chitin triggers phosphorylation of GhLYK5. Our lab has observed the substantial difference in the level of phosphorylation-induced mobility shifts for various proteins likely due to the differential phosphorylation residues, conformation of proteins, and the composition of phos-tag in the protein gels.
13. Figure 6I: the kinase assay lacks all important controls (kinase dead versions, transphosphorylation of the GST tag along, etc). WAK7A is clearly highly active in vitro and used in huge excess, and thus the transphosphorylation of LYK5 (in comparison to that by CERK1) could easily be an artefact.
a. Related line 1093 -1 ul of 32P is not indicative of an amount, must be reported in or Bq, etc.
Our response: As this reviewer suggested, we have performed the in vitro kinase assays with a reduced amount of GhWAK7A proteins and included additional controls, including the kinase-dead mutant GhWAK7A K451M and GST proteins (New Figures 8C, D and Supplemental Figure 17C). We have observed that GhWAK7A CD possessed an auto-phosphorylation activity, which was no longer detectable for GhWAK7A K451M when a similar amount of proteins was used ( Figure 8C). GhWAK7A CD was able to phosphorylate GhLYK5 CD when a much-reduced amount of GhWAK7A CD proteins was used in the in vitro kinase assays ( Figure 8C). In addition, we purified both MBP-GhCERK1 CD and GST-GhCERK1 CD proteins and included MBP and GST proteins as controls. The in vitro kinase assays indicated that although GhCERK1 CD had an auto-phosphorylation activity, but was unable to phosphorylate GhLYK5 CD with either MBP-GhCERK1 CD or GST-GhCERK1 CD as a kinase (New Figure 8D and Supplemental Figure  17C). We have changed the unit of hot ATP to 5 µCi [ 32 P]-γ-ATP, which was included in the Methods.
14. Figure 6H: the chitin-induced phosphorylation band shift of LYK5 is not compromised by kinase-dead WAK7A, which would seem to contradict the model proposed by the authors, this should be discussed.
Our response: Thanks for pointing this out. Indeed, we did not observe the effect of GhWAK7A K451M on the chitininduced phosphorylation mobility shift of GhLYK5. This is not completely surprising since not all the effects on the protein phosphorylation will be reflected by mobility shifts. Alternatively, the endogenous GhWAK7A proteins might still phosphorylate GhLYK5-HA, which contributed to the chitin-induced GhLYK5 mobility shift as indicated for phosphorylation. We have added this in the manuscript (Line 560-562).

Figure 6E -1st panel: text below image; what is the band in line 1 and 3 of panel 2?
Our response: We thank this reviewer for pointing this out. We have changed the presentation format of this figure and make sure that the gel images were clearly labeled (Original Figure 6E; Figure 7G in the revised manuscript). Panel 2 shows α-GST immunoblot of the samples from pull-down assays using α-GST antibody. Lane1 and 3 in panel 2 (middle panel) are GST controls at ~25 kDa and the bands around 75 to 100 kDa are likely non-specific bands ( Figure 7G).
The manuscript "Cotton wall-associated kinase GhWAK7A mediates responses to fungal wilt pathogens by modulating chitin sensory complex", revealed a pivotal, positive roles of a Wall-Associated Kinase (GhWAK7A) during Vd and Fov infection in cotton.
With a comprehensive bioinformatic analysis of WAKs in Gossypium species and transcriptome analysis, GhWAK7A and GhWAK5A present notable induction after infection, then are selected to be the candidate genes. The VIGS and massive biochemical assays provide solid results to support the conclusion that GhWAK7A, GhCERK1 and GhLYK5 are essential for the cotton resistance to Vd and Fov, especially for the result that GhWAK7A interacts with GhCERK1 and GhLYK5.
This manuscript is well organized with solid experimental design and large-scale bioinformatic datasets, and presented a nice result in showing GhWAK7A induced GhCERK1-GhLYK5-involved chitin response in defensing fungal disease, Vd and Fov, showing a broad-spectrum resistance.
Our response: We thank this reviewer for the comments on our solid data collection and comprehensive bioinformatics data analysis. We have modified the manuscript accordingly and made the efforts to address all the following points.
There are several concerns could be addressed by the authors: Major: 1. In the abstract, the author states 'GhLYK5-GhCERK1' complex. In the text line 603, 'GhLYK1' used in the title, but was excluded in the following section (until line 875, author presents 'GhLYK1 and GhLYK5'). I guess 'GhLYK1' is the same gene as 'GhCERK1' (like those in Arabidopsis). Maybe unify the gene name and the complex order name are much better for readers to understand the manuscript.
Our response: Thanks for pointing this out. LYK1 was also named as CERK1 as in Arabidopsis and rice. We have unified the name in the text and used GhCERK1 as suggested.
2. I am not sure, GhWAK7A modulates chitin-induced GhCERK1 and GhLYK5 association via phospholating GhLYK5 first? Then the phosphorlated LYK5 associated with CERK1 then function? 3. You may draw a diagram in DISCUSSION to show the working model of these genes in resistance to both fungal diseases.
Our response: Thanks for the suggestion. We have included a working model to summarize the potential role of GhCERK1, GhLYK5, and GhWAK7A in response to fungal infections in cotton ( Figure 8F) and discussed this model in the Discussion (Line 589-595).
4. All the phenotype graphs ( Figure 2E, 2G...) are lacking of scales, kindly provide them. Minor: 1. The concern raised by Reviewer 5, regarding an alternative model that GhWAK7A might be a scaffolding protein, rather than a kinase. Do your data completely rule out this alternative function? If not, you might want to soften your conclusion, or include this alternative hypothesis in your discussion.
2. Please clarify or correct two confusing issues raised by Reviewer 5: a. In Figure 8E, Why does the level of Phosphorylate GhLYK5 not seem to significantly change upon chitin treatment? A better way to phrase the question might be, why is GhLYK5 already phosphorylated, in the absence of chitin treatment? b. Use of the name GhLYK1 in the model figure, instead of GhCERK1, is confusing.
3. Addressing the editorial comments of Reviewer 4.
- Reviewer comments on previous submission and author responses: We appreciate the decision from the board of reviewing editors to accept our manuscript upon revision based on the reviewers' comments. We have revised our manuscript accordingly. We hope that the revision has sufficiently addressed the reviewers' concerns and met The Plant Cell editorial expectations, and the manuscript will be accepted for publication.

Editors' comments:
We have received reviews of your manuscript entitled "Cotton wall-associated kinase GhWAK7A mediates responses to fungal wilt pathogens by modulating the chitin sensory complex." On the basis of the advice received, the board of reviewing editors would like to accept your manuscript for publication in The Plant Cell. This acceptance is contingent on revision based on the comments of our reviewers. In particular, please consider the following: 1. The concern raised by Reviewer 5, regarding an alternative model that GhWAK7A might be a scaffolding protein, rather than a kinase. Do your data completely rule out this alternative function? If not, you might want to soften your conclusion, or include this alternative hypothesis in your discussion.
Our response: We appreciate the editor and reviewer's suggestion. We agree that GhWAK7A may play a role in scaffolding chitin perception and signaling complex in addition to its function as a kinase. As detailed in response to Review 5, our data from the in vitro and in vivo kinase assays suggested that GhLYK5 phosphorylation by GhWAK7A plays an important role in the chitin perception complex assembly and the chitin-triggered immune signaling in cotton. Additionally, GhWAK7A may play a scaffolding role in chitin perception and signaling complex. We have included this in the Discussion (Line 578-579, Line 630 and Line 647) and the legend of the model figure  (Line 1303-1304).
2. Please clarify or correct two confusing issues raised by Reviewer 5: a. In Figure 8E, Why does the level of Phosphorylate GhLYK5 not seem to significantly change upon chitin treatment? A better way to phrase the question might be, why is GhLYK5 already phosphorylated, in the absence of chitin treatment?
Our response: In Figure 8E, GhLYK5 phosphorylation was detected by an immuno-blot using an α-pS/T antibody with a regular SDS-PAGE gel. This is different from GhLYK5 phosphorylation detection using a phos-tag SDS-PAGE gel in Figure 6H. The phos-tag SDS-PAGE gels separate the phosphorylated and non-phosphorylated proteins based on the phosphorylation levels and capture proteins phosphorylated on serine, threonine, and tyrosine residues, while the detection of phosphoproteins using an anti-pS/T antibody relies on the presence of phosphorylated serine/threonine sites of the proteins. Consistent results were obtained in Figure 6H and 8E  induced GhLYK5 phosphorylation was detected by an immuno-blot using an α-pS/T antibody and a regular SDS-PAGE gel. The phos-tag SDS-PAGE gels separate the proteins based on the level of phosphorylation of total phosphorylation sites, and they include both phosphorylated and non-phosphorylated proteins. Therefore, the different band mobility patterns in Figure 6H and 8E are likely due to phos-tag and regular gels used in the assays respectively. Besides, in Figure 6H, a significant portion of the GhLYK5 proteins also showed a mobility shift before chitin treatment, suggesting the phosphorylation of GhLYK5 before chitin treatment as it was reported for AtCERK1 (Liu et al., 2018;Gong et al., 2019). This is consistent with the observation in Figure 8E where the phosphorylated GhLYK5 proteins were detected by an α-pS/T antibody before chitin treatment. Different from the detection of phosphoproteins by the phos-tag SDS-PAGE gels that separate the phosphorylated and nonphosphorylated GhLYK5 proteins based on the total phosphorylation level, the α-pS/T antibody detects the phosphoserine/threonine residues of GhLYK5 proteins. Thus, the increase of phosphorylated GhLYK5 upon chitin treatment detected by an α-pS/T antibody in Figure 8E seems not as pronounced as that detected by phos-tag SDS-PAGE gels in Figure 6H. Immuno-blots using an α-pS/T antibody after immunoprecipitation confirmed GhLYK5 phosphorylation and further revealed phosphorylation at certain serine and threonine residues in vivo. This experiment was included during the revision as suggested by the reviewers. Figure 7A& B: It would have been nice to see another negative control like something GhWAKs not expected to interact with CERK1 / LYK5 -not induced by chitin?
Our response: In this work, we have included GhWAK5A as a negative control for the specificity of the association of GhWAK7A with GhLYK5 (Supplemental Figure 17B). The result showed that unlike GhWAK7A, which interacted with GhLYK5 constitutively, GhWAK5A did not interact with GhLYK5. In addition, we have included GhWAK5A as a control for the specificity of GhWAK7A in chitin-triggered defense responses, including their expression levels ( Figure 3A) and MAPK activation ( Figure 3B and C).
Given the highly-variable expression of WT vs K451M versions of GhWAK7A, the authors should reconsider their interpretation of results relating the kinase activity. Apart from Figure 8B, all the observations in the current manuscript could be explained by a purely scaffolding role for GhWAK7A. The WT vs kinase-dead versions of GhWAK7A would need to be compared side-by-side in a single experiment to properly conclude that its kinase activity is strictly required to support LYK-CERK complex formation. Furthermore, multiple kinase-dead versions of the protein (e.g. mutation of the catalytic loop D alongside the current K451M) to reach a clearer conclusion.
Our response: We agree with this reviewer on the possibility of a scaffolding role of GhWAK7A in chitin-mediated signaling. In addition to Figure 8B where the kinase-inactive mutant of GhWAK7A (GhWAK7A K451M ) reduced chitininduced GhCERK1-GhLYK5 association, the in vitro and in vivo kinase assays ( Figure 8C and F) suggest that GhLYK5 phosphorylation by GhWAK7A plays a role in chitin-triggered immune signaling in cotton. However, we concur with this reviewer that GhWAK7A may also play a scaffolding role in chitin perception and signaling. We have included this in the Discussion and Model (Line 578-579, Line 630, and Line 647).
It is unclear why the authors focus on GhLYK1 in their model, when all their experimental data has been focused on GhCERK1.
Our response: Thanks for pointing this out. We have replaced GhLYK1 in the model by GhCERK1 to avoid confusion ( Figure 8F).
The manuscript does not clearly state the origin of all antibodies used (e.g. what a-pS/T antibody is used? Is it context-specific, monoclonal, etc?). For all antibodies, supplier and catalogue # should be explicitly stated in the methods.
Our response: The suppliers and catalog numbers of all antibodies have been stated in the methods. The α-pS/T antibodies (anti-phosphoserine/threonine rabbit polyclonal antibodies) are from ECM Bioscience (Cat# PP2551). Figure 6G and 6H are missing MW indications.
Our response: Figure 6G and 6H showed the data from phos-tag SDS-PAGE gels. The regular protein molecular weight markers contain chelating agents, which distort band shapes affecting the neighboring sample lanes. We did not include the protein molecular weight markers when we performed these assays. The molecular weight for the loading control (Rubisco protein) has been labeled.
The authors seem to have addressed most of the issues that were raised in review #1. I still think it would be better to see a co-IP of the WAK and LYK/CERK from plant tissue rather than protoplasts over expressing fusion proteins.
Our response: We agree with this reviewer that it would be ideal to use cotton plant tissues to perform the co-IP assays of WAK and LYK/CERK. It is challenging to get antibodies to specifically recognize WAK and LYK/CERK in cotton particularly considering the large number of homologous genes in the cotton genome. Stable transgenic plants with an epitope tag is an alternative for co-IP using plant tissues. Cotton transformation is achievable for a few lines with reasonable regeneration efficiency through the tissue culture process, which takes eight to twelve months to obtain the primary transformants. Transforming cotton genes into Arabidopsis shortens the experimental time yet introduces complications associated with different plant species. The efficiency of the Agrobacterium-mediated transient expression system in cotton is low and needs to be further optimized. We have established protocols to use cotton protoplasts to express proteins. The transfected cotton protoplasts respond to pathogen elicitors, including chitin-triggered MAPK activation and immune receptor complex formation. Thus, the cotton protoplast system is an alternative and effective way to facilitate the functional study of cotton genes in cotton but not in other heterologous model systems.
While the data support a role for WAKs in chitin perception, I don't think the paper shows that they "modulate" the response" as stated in the title.
Our response: We have changed the title to "Cotton wall-associated kinase GhWAK7A mediates responses to fungal wilt pathogens by complexing with the chitin sensory receptors" (Line 3) and statement in ABSTRACT (Line 46).