BIK1 protein homeostasis is maintained by the interplay of different ubiquitin ligases in immune signaling

Pathogen-associated molecular patterns (PAMPs) trigger plant innate immunity that acts as the first line of inducible defense against pathogen infection. A receptor-like cytoplasmic kinase BOTRYTIS-INDUCED KINASE 1 (BIK1) functions as a signaling hub immediately downstream of multiple pattern recognition receptors (PRRs). It is known that PLANT U-BOX PROTEIN 25 (PUB25) and PUB26 ubiquitinate BIK1 and mediate BIK1 degradation. However, how BIK1 homeostasis is maintained is not fully understood. Here, we show that two closely related ubiquitin ligases, RING DOMAIN LIGASE 1 (RGLG1) and RGLG2, preferentially associate with the hypo-phosphorylated BIK1 and promote the association of BIK1 with the co-receptor for several PRRs, BRI1-ASSOCIATED RECEPTOR KINASE1 (BAK1). PUB25 interacts with RGLG2 and mediates its degradation. In turn, RGLG2 represses the ubiquitin ligase activity of PUB25. RGLG1/2 suppress PUB25-mediated BIK1 degradation, promote BIK1 protein accumulation, and positively regulate immune signaling in a ubiquitin ligase activity-dependent manner. Our work reveals how BIK1 homeostasis is maintained by the interplay of different ubiquitin ligases.

The authors show the importance of the ubiquitination activity of RGLGs for BIK1 regulation by using RGLG2m which carries mutations on 4 cysteine residues. In general, mutations in the cysteine residues can change the structure of the protein. Does RGLG2m interact with BIK1 similar to the WT RGLG2 in vitro and in vivo? It is possible that the mutations may change the structure of RGLG2 and reduce binding affinity to BIK1, which causes the loss of regulation of BIK1 and the ubiquitination activity may not be important for BIK1 regulation.
Other points Line 105-106 and 109-110, please update this part with the most recent publication: https://doi.org/10.15252/embj.2020107257 For figure 2b and supp figure 3a, it is strange to not express BIK1 as a negative control. The authors should use a closely related protein (for example PBL13) as an additional negative control. For figure 3e, the BIK1-FLAG in Col-0 is barely detectable, please include an image with stronger exposure. There seems to be a weaker expression of RGLG1 and RGLG2 when co-expressed with PUB25 (Figures 4a and 4b). Are there any direct interactions between RGLG1/2 and PUB25? The label of figure 5a might be wrong. I believe it should be '--+ +' for RGLG2. If both PUB25 and RGLG1/2 can ubiquitinate BIK1, how does one stabilize the protein and the other degrades it? What is the specificity? I think the authors should at least compare the BIK1 ubiquitination pattern induced by RGLG1/2, PUB25, and RHA3B. In figure 6f, lane 5 of the western blot, why is there a band around 55KDa when PUB25 is not expressed? Is that a non-specific band? Line 342, please keep the protein names (ALT44/45 and RHA3B) consistent throughout the paper.

Reviewer #2 (Remarks to the Author):
In this study, the authors have discovered that two RING-type ubiquitin ligases, RGLG1 and RGLG2, which are closely related, compete with PUB25 for associating with BIK1. Additionally, RGLG1/2 directly ubiquitinate BIK1, and their functions in favourably controlling immunological signals depend on the activity of the ubiquitin ligase. The study is novel and interesting and reveals how the ubiquitylation of BIK1 activates the immune response. Although the findings from the study are robust, there are some issues with the data, which I have listed below: 1. In this study, the RGLG1 was identified from the transcriptomic study; however, please explain why sometimes RGLG1 was used, and sometimes RGLG2 but claiming both to have a similar effect. It is important to maintain consistency in the results by including both RGLG1 and RGLG2 in all the experiments done in the study. 2. It is important to show in the supplementary data the different isoforms of RGLG and how much conservation they possess. 3. While performing the split luciferase complementation assays, why did the relative luciferase activity fluctuate in each setup, especially in Fig S2c when the ATL83 association was measured? Also, from the figure, the treatment of flg22 did not significantly reduce the association of ATL32 and ATL6-NL with BIK1-CL. Why was BIK1 was cloned preferentially with CLuc and not with NLuc? 4. In Fig 2, it is important to include the loading control to show an equal amount of protein loaded in each well. 5. In Fig 2, the authors have claimed both RGLG1 and RGLG2 to show an association to BIK1; however, they have only shown the flg22-induced dissociation of RGLG2 and BIK1 through co-IP. Similar studies need to be undertaken with RGLG1. As mentioned in line 208, I could not find a validation wherein the BIK1 hyperphosphorylation leads to the inhibition of RGLG2 interaction. It would be important to show this in a phospho-mimic mutant of BIK1. 6. The BIK1 is a target for PBL25-mediated ubiquitination. In lines 244-246, the overexpression of BIK1 in the rglg1rglg2 background does not increase the plant immunity to DC3000 infection. So does BIK1 OE increase the levels of PBLs in the rglg double mutant background? 7. There appears to be a considerable size difference between RGLG1 and RGLG2. The BIK1-HA levels in Fig 3 g and h without RGLG1/2 proteins are dissimilar when they are just biological replicates. There should be consistency in the results. 8. In fig 4d, the input lane of the GST-BIK1 shows similar levels in lane1 and 2 when RGLG2 should show more levels of BIK1. Also similar results were seen in 4c, wherein the BIK1-FLAG levels were unaltered when present with PUB25 only (3rd lane) 9. BIK1 is plasma membrane-associated protein, RGLG1/2 is also plasma membraneassociated. BiFC or co-localization assays can be done to show the physical interaction on the cellar level. 10. A competition plot can be used to understand the competitive interaction between BIK1, PUB25 and RGLG1/2. An example of this can be found in this publication (https://www.sciencedirect.com/science/article/pii/S1046202301911788) 11. BAK1 promotes the recruitment of RGL2 with BIK1 however in Fig 5a, we observe that flg22 treatment does not promote the interaction of BAK1 with RGLG2. We find RGLG in the input lane even when RGLG is absent, which could be a non-specific band. 12. RGLG2 mediated interaction of BIK1 and BAK1 shown in Fig 5C is unclear. Pull down with FLAG followed by probing with α-GFP needs to be shown to validate the interaction of the RGLG2-BIK1-BAK1 complex. 13. In fig 5f lane 5 we observe low levels of PUB25 even when it is absent from the interaction. 14. The discussion section is repetitive in some places, for example, in lines 345-353. 15. It is strange why in Fig 4d, the size of His-FLAG-PUB25 is around 43kDa when blotted with αFLAG however, in Input, the protein size is around 55kDa. The experiment needs to be redone. 16. The working model mentioned that ubiquitination of BIK1 prevents its proteasomal degradation. It would be good to find the downstream effector of RGLG1/2 mediated monoubiquitinated BIK1. How is the monoubiquitylation of BIK1 different from that mediated by RHA3A/B? It is intriguing how RGLG1/2 plays a vital role in the mono-ubiquitination of BIK1 and maintains its levels. For the other role of RGLG1/2, they induce proteolytic turnover of PIN2, ERF53 and PP2CA. However, it is unclear how the ubiquitination of BIK1 helps to prevent its degradation and promote immune responses. Overall, the study deals with an exciting dimension of BIK1-mediated regulation of plant immunity and incorporating the changes will make the findings more robust and straightforward for a broad readership.

Reviewer #3 (Remarks to the Author):
The manuscript by Bai and Sun et al. showed that the closely related ubiquitin ligases RGLG1/2 associate with the hypo-phosphorylated BIK1 and are required for the accumulation BIK1 in Arabidopsis plants. RGLG1/2 also associate with the receptor-like kinase BAK1, which is partially required for their association with BIK1. In addition, RGLG1/2 were shown to promote ubiquitination of BIK1 in vitro and compete with PUB25 for the association with BIK1. While these findings are interesting, additional data are required to clarify the mechanism of how RGLG1/2 regulate the stability of BIK1. 1. It was proposed that RGLG1/2 increase BIK1 stability through direct ubiquitination of BIK1 (mainly based on in vitro evidence that RGLG1/2 ubiquitinate BIK1). If this is true, a considerable amount of ubiquitinated BIK1 protein should be observed on various western blots shown in the manuscript, especially when RGLG1/2 were co-expressed with BIK1-KA, as ubiquitinated and non-ubiquitinated BIK1 proteins can be easily separated by SDS-PAGE. 2. RGLG1/2 were shown to compete with PUB25 for interaction with BIK1. Are the ubiquitin ligase activities of RGLG1/2 required for their competition with PUB25 for BIK1 interaction? 3. Can RGLG1/2 ubiquitinate PUB25 to affect its ability to degrade BIK1? 4. The rglg1/2 double mutant exhibits pleiotropic phenotype as shown in Fig. 6B. Is expression of defense marker genes such as PR1 and PR2 up-regulated in the rglg1/2 double mutant like in the bik1? Does overexpression of BIK1 alleviate the morphologic phenotype of rglg1/2?

Reviewer #1:
In the manuscript entitled"BIK1 protein homeostasis is maintained by interplay of different ubiquitin ligases to regulate immune signaling" Bai et al., reported new positive regulators, RGLG1 and RGLG2, which positively regulate BIK1 protein level. Both RGLGs preferentially associate with the hypo-phosphorylated BIK1, compete with PUB25 for binding to BIK1, and suppress PUB25-mediated BIK1 degradation. They also found RGLGs directly ubiquitinate BIK1 and their ubiquitination activity is required for the suppression of PUB25-mediated BIK1 degradation. I think the paper is well written and the authors performed many different experiments to show the function of RGLGs on BIK1 homeostasis. Thus, I believe that the manuscript is potentially interesting for the readers in Nature com. Thank you so much for your valuable suggestions! 1) We have performed co-IP assays using stable transgenic lines. We confirmed that RGLG2 was associated with BIK1 in pRGLG2::RGLG2-HA/p35S::BIK1-FLAG transgenic plants (Fig. 2d), and RGLG1 was associated with BIK1 in p35S::RGLG1-HA/p35S::BIK1-FLAG transgenic plants ( Supplementary Fig. 6e).
2) We demonstrated the PM localization of RGLGs by experimental data: We showed that both RGLG1-GFP and RGLG2-GFP were localized to the plasma membrane, when they were expressed in Arabidopsis protoplasts ( Supplementary Fig.   4c, d).
2. rglg1 rglg2 mutant shows severe morphological phenotype, showing pleiotropic effects by the loss of RGLGs. In actuality, RGLG2 was shown to regulate PIN2 and Thanks a lot for these insightful suggestions! 1) The growth defect phenotype of rglg1 rglg2 was almost not restored by overexpressing BIK1 (Supplementary Fig. 12b), suggesting that the growth defect of rglg1/2 is not caused by the reduced accumulation of BIK1 proteins.
2) We provided the kinetics of ROS production in rglg1 rglg2 (New Fig. 3a).
These results showed the specific regulation of BIK1 by RGLGs in PTI pathway.
The increased association is quite weak. Moreover, the slightly increased BIK1 protein amount in the output might be due to the increased BIK1 protein level in the presence of RGLG2 and not due to the enhanced interaction.

Thanks a lot for this good question!
We performed more experiments to further support that RGLG2 enhances BAK1-BIK1 interaction.
1) We expressed BAK1-FLAG together with BIK1-HA and RGLG1/2-HA in protoplasts and performed co-IP assays. The results not only showed that both BIK1-HA and RGLG1/2-HA could be pulled down by BAK1-FLAG, but also confirmed that the association of BAK1 and BIK1 were enhanced by RGLG/2 ( Fig.   6e, Supplementary Fig. 20c).
2) Furthermore, we examined the the association of BAK1 and BIK1 in rglg1 rglg2 mutant. We found that the association of BAK1 and BIK1 was weaker in rglg1 rglg2 than in Col-0 ( Fig. 6f).

The authors show the importance of the ubiquitination activity of RGLGs for BIK1
regulation by using RGLG2m which carries mutations on 4 cysteine residues. In general, mutations in the cysteine residues can change the structure of the protein.
Does RGLG2m interact with BIK1 similar to the WT RGLG2 in vitro and in vivo? It is possible that the mutations may change the structure of RGLG2 and reduce binding affinity to BIK1, which causes the loss of regulation of BIK1 and the ubiquitination activity may not be important for BIK1 regulation.
Thank you for this good question! To exclude this possibility, we examined the association of BIK1 with RGLG1m or RGLG2m. The co-IP results showed that the association of BIK1 with RGLG1/2 was comparable to that with RGLG1m/2m in vivo; and the in vitro pull-down results showed that the direct interaction of BIK1 with RGLG1/2 was comparable to that Thank you for this suggestion! We have updated the related parts in both introduction and discussion parts.
"Interestingly, PUB4 was also shown to have a dual effect on BIK1 homeostasis: it mediates the ubiquitination and degradation of non-activated BIK1 at the resting state, but promotes the accumulation of activated BIK1 after PAMP treatment". Thank you for this suggestion! We performed new experiments using PBL13 as a negative control. We found that neither RGLG1 nor RGLG2 was able to associate with PBL13 ( Supplementary Fig. 6c, d).
For figure 3e, the BIK1-FLAG in Col-0 is barely detectable, please include an image with stronger exposure.
Thank you for this suggestion! We have added an image with stronger exposure (new There seems to be a weaker expression of RGLG1 and RGLG2 when co-expressed with PUB25 (Figures 4a and 4b). Are there any direct interactions between RGLG1/2 and PUB25?
Thank you so much for this insightful question! The review #3 also raised the similar question.
Inspired by your great question, we explored the relationship between RGLG1/2 and PUB25, and found that there is interplay between RGLG1/2 and PUB25. These results support the consumption in the title of the paper "BIK1 protein homeostasis is maintained by interplay of different ubiquitin ligases to regulate immune signaling" 1) PUB25 directly interacts with RGLG1/2 The results of co-IP assays showed that RGLG1/2 associated with PUB25 in Arabidopsis protoplasts. Furthermore, the association was reduced upon flg22 stimulation (Fig. 4a, b). We also performed BiFC assays, the results showed that PUB25-nYFP associated with RGLG1/2-cYFP at the plasma membrane, while PUB25-nYFP did not associate with RGLG3-cYFP, and RGLG1/2-cYFP did not associate with PUB39-nYFP ( Fig. 4c and Supplementary Fig. 15a). Moreover, the results of pull-down assays showed that RGLG1/2 directly interacted with PUB25 in vitro ( Fig. 4d and Supplementary Fig. 15b). Moreover, we found that PUB25 directly ubiquitinates RGLG2m in vitro ( Supplementary Fig. 15c).
2) PUB25 mediates the proteosomal degradation of RGLG1/2 Interestingly, we found that, in the presence of CHX, the proteins levels of RGLG2 were reduced when they were co-expressed with PUB25, which was blocked by MG132, an inhibitor of 26S proteosome (Fig. 4e). Furthermore, when transiently expressed, RGLG1/2 protein levels were higher in pub25 pub26 double mutant protoplasts than in Col-0 protoplasts in the presence of CHX (Fig. 4f, Supplementary   Fig. 16a, b). These results suggest that PUB25 mediates the proteosomal degradation of RGLG1/2.

3) RGLG2 suppresses the ubiquitin ligase activity of PUB25
PUB25 has ubiquitin ligase activity as demonstrated by its auto-ubiquitination (Fig.   4g). Interestingly, expression of RGLG2, but not RGLG2m led to the reduction in PUB25 auto-ubiquitination (Fig. 4g). However, the non-covalent association of PUB25 with a Ub moiety was not affected (Fig. 4g). These results suggest that RGLG2 suppresses the ubiquitin ligase activity of PUB25.
The label of figure 5a might be wrong. I believe it should be '--+ +' for RGLG2.
Thank you so much for catching this error! We have corrected the wrong labeling (New Fig. 6a) If both PUB25 and RGLG1/2 can ubiquitinate BIK1, how does one stabilize the protein and the other degrades it? What is the specificity? I think the authors should at least compare the BIK1 ubiquitination pattern induced by RGLG1/2, PUB25, and RHA3B.
Thank you for this good suggestion! We compared the BIK1 ubiquitination pattern induced by RGLG1/2 and PUB25. We found that the ubiquitination pattern mediated by RGLG1/2 was different from that by PUB25. The ubiquitination of BIK1 by PUB25 exhibited a laddering pattern, therefore, the type of BIK1 ubiquitination by PUB25/26 seems to be polyubiquitination; while that by RGLG1/2 seemed to be monoubiquitination (Fig. 5f).
For RHA3A/B-mediated BIK1 monoubiquitination, we introduced combined mutations of the C-terminal four lysines (ubiquitination sites mediated by RHA3A/B) into BIK1 and generated BIK1 C4KR . BIK1 C4KR was not able to dissociate from FLS2 upon flg22 treatment ( Supplementary Fig. 14a). However, the protein accumulation of BIK1 C4KR promoted by RGLG1/2 was comparable to that of WT BIK1; and PUB25 mediates the degradation of BIK1 C4KR as it does for WT BIK1 (Supplementary Fig.   14b, c).
In figure 6f, lane 5 of the western blot, why is there a band around 55KDa when PUB25 is not expressed? Is that a non-specific band?
Thank you for pointing this out! We have performed new experiments (New Fig. 5a).
Line 342, please keep the protein names (ALT44/45 and RHA3B) consistent throughout the paper.
Thank you for pointing this out! We have changed the proteins names and kept the name of RHA3A/B consistent throughout the paper.

Reviewer #2 (Remarks to the Author):
In this study, the authors have discovered that two RING-type ubiquitin ligases, RGLG1 and RGLG2, which are closely related, compete with PUB25 for associating with BIK1. Additionally, RGLG1/2 directly ubiquitinate BIK1, and their functions in favourably controlling immunological signals depend on the activity of the ubiquitin ligase. The study is novel and interesting and reveals how the ubiquitylation of BIK1 activates the immune response. Although the findings from the study are robust, there are some issues with the data, which I have listed below: We appreciate your great suggestions and comments! 1. In this study, the RGLG1 was identified from the transcriptomic study; however, please explain why sometimes RGLG1 was used, and sometimes RGLG2 but claiming both to have a similar effect. It is important to maintain consistency in the results by including both RGLG1 and RGLG2 in all the experiments done in the study.
Thank you for this good suggestion! Actually, RGLG1 was much harder to be handled than RGLG2 due to unknown reasons, especially for purification of RGLG1 DNA by maxi prep to do protoplast transfection, and for obtaining RGLG1 transformants in Escherichia coli to analyze ubiquitination. We tried our best and performed the following new RGLG1 related experiments during revision. Eventually, the most data actually include both RGLG1 and RGLG2. We found that RGLG1 and RGLG2 function similarly in regulating immunity. To confirm the results obtained via split luciferase complementation assays, we also performed co-immunoprecipitation (IP) assays in Arabidopsis protoplasts to examine the association of BIK1 with ATL80 or ATL6, and similar results were obtained ( Supplementary Fig. 3a, b). 2) We have modified these statements: "Furthermore, the association of BIK1 with RGLG1 or ATL83 was significantly reduced upon flg22 stimulation, and that with ATL32 or ATL6 was slightly reduced after flg22 treatment ( Supplementary Fig. 2c)".
3) We performed new split-luciferase complementation assays using BIK1-NLu, and similar results were obtained when BIK1-NLuc and RGLG1/2-CLuc were transiently expressed in N. benthamiana ( Supplementary Fig. 5a and 5b). Thank you for this great suggestion!

1) We performed new experiments with RGLG1:
A. To confirm the association of RGLG1 with BIK1, we co-expressed FLAG-epitope-tagged RGLG1 and HA-epitope-tagged BIK1 in Arabidopsis protoplasts. RGLG1-FLAG proteins were immunoprecipitated by anti-FLAG antibodies, and BIK1-HA was present in the RGLG1-FLAG immunoprecipitates, and the association between BIK1 and RGLG1 was reduced when the protoplasts were treated with flg22 (Fig. 2b); B. Similarly, RGLG1-HA were also present in the BIK1-FLAG immunoprecipitates, and the association of BIK1 and RGLG1 was reduced when the protoplasts were treated with flg22 ( Supplementary Fig. 5c); C. We also performed co-IP assays using stable transgenic lines. We found that RGLG1 was associated with BIK1 in p35S::RGLG1-HA/p35S::BIK1-FLAG plants, and the association of BIK1 and RGLG1 was reduced when the protoplasts were treated with flg22 ( Supplementary Fig. 6e); D. We found that the association of RGLG1 with the BIK1 kinase dead mutant, BIK1 K105E (BIK1Km) was stronger than with wild-type (WT) BIK1. Moreover, the flg22-induced dissociation was not observed between RGLG1/2 and BIK1Km (Fig.   2g). By contrast, the association of RGLG1 with the BIK1 phospho-mimetic mutant BIK1S236D/T237D (BIK1 2D ) was weaker than with WT BIK1 (Supplementary Fig   8b).
Additionally, we also performed performed BiFC assays in Arabidopsis protoplasts using BIK1-nYFP and RGLG1-cYFP. BIK1 associated with RGLG1 at the plasma membrane (Fig. 2e). Moreover, RGLG1 was pulled down by BIK1 fused to a GST tag, but not by GST or GST-FLS2CD (the intracellular region of FLS2) in in vitro assays ( Fig. 2f) 2). To support that BIK1 hyperphosphorylation leads to the inhibition of RGLG1/2 interaction, we performed experiments to examine the association of RGLG1/2 with the BIK1 phospho-mimetic mutant BIK1S236D/T237D (BIK1 2D ). We found that the association of RGLG1/2 with BIK1 2D was weaker than that with BIK1 ( Supplementary Fig 8b, c).  Fig. 12d); and protein accumulation of exogenous PUB25-HA was not increased by BIK1 overexpression (Supplementary Fig. 12e).
7. There appears to be a considerable size difference between RGLG1 and RGLG2.
The BIK1-HA levels in Fig 3 g and h without RGLG1/2 proteins are dissimilar when they are just biological replicates. There should be consistency in the results.
Thanks a lot for this suggestion! We have redone this experiment (New Fig. 3g). 9. BIK1 is plasma membrane-associated protein, RGLG1/2 is also plasma membrane-associated. BiFC or co-localization assays can be done to show the physical interaction on the cellar level.
Thank you for this great suggestion!
We performed BiFC bimolecular fluorescence complementation (BiFC) assays in Arabidopsis protoplasts using BIK1 tagged with an nYFP (the N-terminal half of yellow fluorescent protein) tag and RGLG1/2 tagged with a cYFP (the C-terminal half of YFP) tag. BIK1 associated with RGLG1/2 at the plasma membrane. Whereas, BIK1-nYFP did not associate with RGLG3-cYFP, and RGLG1/2-cYFP did not associate with PBL13-nYFP ( Fig. 2e and Supplementary Fig. 6f). These results verify the specificity of the association between BIK1 and RGLG1/2 at the plasma membrane.
We also performed co-localization assays. RGLG1-GFP or RGLG2-GFP was co-expressed with TaPBS1-RFP (a known PM associated protein) in Arabidopsis protoplasts. Green and red fluorescent proteins were visualized via confocal microscopy. We found that both RGLG1-GFP and RGLG2-GFP were localized to the plasma membrane (PM), when they were expressed in Arabidopsis protoplasts ( Supplementary Fig. 4c, d).
10. A competition plot can be used to understand the competitive interaction between BIK1, PUB25 and RGLG1/2. An example of this can be found in this publication (https://www.sciencedirect.com/science/article/pii/S1046202301911788) Thank you for this great suggestion! A competition plot would be very useful for understanding the competitive interaction between BIK1, PUB25 and RGLG1/2.
However, our data of protein levels are based on images in different blots probed with different antibodies, and are very hard to be used for performing the competition plot analysis, which definitely awaits future investigation.
11. BAK1 promotes the recruitment of RGL2 with BIK1 however in Fig 5a, we observe that flg22 treatment does not promote the interaction of BAK1 with RGLG2.
We find RGLG in the input lane even when RGLG is absent, which could be a non-specific band.
Thanks a lot for this question! Yes, flg22 treatment does not promote the interaction of BAK1 with RGLG2. Together with other results, we think "that RGLG2 is likely accompanying BAK1 in the same complex to regulate immunity" (Supplementary Fig.   20a, b).
Thank you so much for catching this error! We have corrected the labeling error (New Fig. 6a).
12. RGLG2 mediated interaction of BIK1 and BAK1 shown in Fig 5C is unclear. Pull down with FLAG followed by probing with α-GFP needs to be shown to validate the interaction of the RGLG2-BIK1-BAK1 complex.
Thanks a lot for this great suggestion! We followed your suggestion and performed new co-IP experiments. When BAK1-FLAG was co-expressed with BIK1-HA and RGLG1/2-HA in protoplasts, the results of co-IP not only showed that both BIK1-HA and RGLG1/2-HA could be pulled down by BAK1-FLAG, but also confirmed that the association of BAK1 and BIK1 were promoted by RGLG1/2, and also suggest that that BAK1, BIK1, and RGLG1/2 are in the same complex (Fig. 6e, Supplementary   Fig. 20c).
Furthermore, we compared the the association of BAK1 and BIK1 in Col-0 and rglg1 rglg2 plants, we found that the association of BAK1 and BIK1 was weaker in rglg1 rglg2 than in Col-0 (Fig. 6f). 13. In fig 5f lane 5 we observe low levels of PUB25 even when it is absent from the interaction.
Thank you for pointing this out! We have performed new experiments (New Fig. 5a).
14. The discussion section is repetitive in some places, for example, in lines 345-353.
Thanks for this suggestion! We have rewrote this part.
"Here, we demonstrate that another two closely related RING-type ubiquitin ligases, RGLG1 and RGLG2 promote BIK1 accumulation and positively regulate BIK1-mediated immune signaling. Importantly, our work reveal that BIK1 protein homeostasis is maintained by interplay of RGLG1/2 and PUB25 (Fig. 7)." 15. It is strange why in Fig 4d,  However, it is unclear how the ubiquitination of BIK1 helps to prevent its degradation and promote immune responses. Overall, the study deals with an exciting dimension of BIK1-mediated regulation of plant immunity and incorporating the changes will make the findings more robust and straightforward for a broad readership.
Thank you for these insightful questions and suggestions! 1) We generated combined mutations of the C-terminal four lysines (BIK1 C4KR ) that are ubiquitination sites mediated by RHA3A/B. BIK1 C4KR was unable to dissociate from FLS2 upon flg22 treatment ( Supplementary Fig. 14a). However, the protein accumulation of BIK1 C4KR promoted by RGLG1/2 was comparable to that of WT BIK1; and PUB25 mediates the degradation of BIK1 C4KR as it does for WT BIK1 ( Supplementary Fig. 14b, c). These results suggest that the role of RGLG1/2 in regulating BIK1 is distinct from that of RHA3A/B.
2) We also added new discussions in the discussion section: "The type of BIK1 ubiquitination by PUB25/26 seems to be polyubiquitination 14 (Fig.   5f), while RHA3A/B mediate the monoubiquitination of BIK1 13 . In this work, we demonstrate that RGLG1/2 directly ubiquitinate BIK1, and the mode of BIK1 ubiquitination by RGLG1/2 seems to be monoubiquitination as demonstrated by in vitro assays (Fig. 5e, f and supplementary Fig. 18). However, the context, mechanism, and outcome of BIK1 monoubiquitination mediated by RHA3A/B are distinct from those by RGLG1/2. RGLG1/2 prefer to target the hypo-phosphorylated BIK1 ( Fig. 2 and supplementary Fig. 5, 6, 8), while the flg22-induced BIK1 phosphorylation is a prerequisite for its monoubiquitination mediated by RHA3A/B 12 . RHA3A/B-mediated BIK1 monoubiquitination is required for the release of BIK1 from FLS2, and BIK1 C4KR fails to dissociate from FLS2 upon flg22 treatment, while RGLG1/2 still promote the protein accumulation of BIK1 C4KR (supplementary Fig. 14). Nevertheless, the exact outcomes of RGLG1/2-mediated BIK1 ubiquitination await future investigation".
3) Consistently, in our system, we found that the expression of RGLG1/2 reduced the protein abundance of ERF53 and PP2CA, both were shown to be the substrates of RGLG1/2 ( Supplementary Fig. 11b, c) showed that association of BIK1 and PUB25 was also reduced by RGLG1m/2m expression, although to a less extent than by RGLG1/2 ( Supplementary Fig. 19c, d), suggesting that RGLG1m/2m are also able to compete with PUB25 for BIK1, and the ubiqutin ligase activity of RGLG1/2 is partially required for their competition with PUB25. However, RGLG2m only slightly but not significantly restored the immune responses of rglg1 rglg2 with respect to resistance to Pst DC3000 hrcC - (Fig. 5c).
These results imply that the competition of RGLG1/2 with PUB25 likely make a minor contribution to the function of RGLG1/2 in regulating BIK1 homeostasis.
Thank you for this insightful question! We performed new experiments and found that RGLG2 suppresses the ubiquitin ligase activity of PUB25, which could affect its ability to degrade BIK1. "PUB25 has ubiquitin ligase activity as demonstrated by its auto-ubiquitination (Fig. 4g).
Interestingly, we found that expression of RGLG2, but not RGLG2m led to the reduction in PUB25 autoubiquitination (Fig. 4g). However, the non-covalent association of PUB25 with a Ub moiety was not affected (Fig. 4g)".
Furthermore, PUB25 directly ubiquitinates RGLG2m in vitro ( Supplementary Fig.   15c), and in the presence of CHX, the proteins levels of RGLG2 were reduced when they were co-expressed with PUB25, which was blocked by MG132, an inhibitor of 26S proteosome (Fig. 4e). Moreover, when transiently expressed, RGLG1/2 protein levels were higher in pub25 pub26 double mutant protoplasts than in Col-0 protoplasts in the presence of CHX (Fig. 4f, Supplementary Fig. 16a, b). These results suggest that PUB25 mediates the proteosomal degradation of RGLG1/2. Thanks a lot for these questions!
We measured the transcript levels and found that unlike those in bik1, the PR1 transcript levels were slightly but not significantly higher in rglg1 rglg2 than in Col-0 ( Supplementary Fig. 9d). The growth defect phenotype of rglg1 rglg2 were almost not restored by overexpressing BIK1 (Supplementary Fig. 12b), suggesting that the growth defect of rglg1/2 is not caused by the reduced accumulation of BIK1 proteins.
In the revised manuscript, the authors provided many additional data which well supports their model of BIK1 regulation by RGLGs. I think the overall data is convincing and the revised paper is well-written. However, I want to raise a few very minor points which would help to improve this paper further before publication in Nature Communication.
In Figure 3, the authors showed that RGLG1/2 promotes BIK1 protein accumulation. However, in Fig6 RGLG2 expression does not affect the BIK1 protein level. This difference would be probably because of the presence or absence of CHX. If so, the authors should explain possible mechanisms of CHX on RGLG1/2-mediated BIK1 protein accumulation. Why does RGLG1/2-mediated BIK1 protein accumulation become obvious only in the presence of CHX, while BIK1 reduction in rglg1/2 is obvious without CHX? In Figure 6, the authors showed that RGLG2 promotes the association between BIK1 and BAK1. Can the authors exclude the possibility that RGLG2 enhances the BIK1-BAK1 interaction through the accumulation of BIK1? Why BIK1 level is less in rglg1/2 mutant in Fig. 3F, but not in Fig. 6F? I think the authors should carefully explain these points to avoid confusing readers.
How conserved are RGLG1/2 in plants? Do they exist outside the Brassicaceae?
Since RGLG1/2 can also associate with BAK1, can authors check the native protein level of BAK1 in the rglg1/2 double mutant? (BAK1 expression level looks slightly weaker in rglg1/2 compared to WT in Figure 6f.) In the revised manuscript, the authors provided many additional data which well supports their model of BIK1 regulation by RGLGs. I think the overall data is convincing and the revised paper is well-written. However, I want to raise a few very minor points which would help to improve this paper further before publication in Nature Communication.
We highly appreciate your great suggestions and comments! 1. In Figure 3, the authors showed that RGLG1/2 promotes BIK1 protein accumulation. However, in Fig. 6 RGLG2 expression does not affect the BIK1 protein level. This difference would be probably because of the presence or absence of CHX.
If so, the authors should explain possible mechanisms of CHX on RGLG1/2-mediated BIK1 protein accumulation.
Thank you so much for the valuable suggestions! Yes, this difference would be probably due to the presence or absence of CHX. We have added the following new statements and new data: 2) We also added new data showing that "notably, in the absence of CHX, overexpression of RGLG1 hardly affected the BIK1 protein accumulation in protoplasts ( Supplementary Fig. 12d). This might be because transient protein overexpression in Arabidopsis protoplasts results in synthesis of large amounts of BIK1 proteins. Therefore, the contribution of BIK1 degradation to its abundance becomes unconspicuous under this condition".
2. Why does RGLG1/2-mediated BIK1 protein accumulation become obvious only in the presence of CHX, while BIK1 reduction in rglg1/2 is obvious without CHX?
Thanks for this good question!