Arabidopsis P4 ATPase-mediated cell detoxification confers resistance to Fusarium graminearum and Verticillium dahliae

Many toxic secondary metabolites produced by phytopathogens can subvert host immunity, and some of them are recognized as pathogenicity factors. Fusarium head blight and Verticillium wilt are destructive plant diseases worldwide. Using toxins produced by the causal fungi Fusarium graminearum and Verticillium dahliae as screening agents, here we show that the Arabidopsis P4 ATPases AtALA1 and AtALA7 are responsible for cellular detoxification of mycotoxins. Through AtALA1-/AtALA7-mediated vesicle transport, toxins are sequestered in vacuoles for degradation. Overexpression of AtALA1 and AtALA7 significantly increases the resistance of transgenic plants to F. graminearum and V. dahliae, respectively. Notably, the concentration of deoxynivalenol, a mycotoxin harmful to the health of humans and animals, was decreased in transgenic Arabidopsis siliques and maize seeds. This vesicle-mediated cell detoxification process provides a strategy to increase plant resistance against different toxin-associated diseases and to reduce the mycotoxin contamination in food and feed.


Introduction
1. "The P4 subfamily of P-type ATPases (P4 ATPases) functions as phospholipid flippases that translocate specific phospholipid substrates from the exoplasmic/luminal leaflet to the cytoplasmic leaflet of biological membranes,..." All plant P4 ATPases characterized to date and most P4 ATPases from other organisms are phospholipid transporters, but several P4 ATPases from humans and yeast have now been shown to transport lipids not containing a phosphate group, so please correct his sentence to account for that fact.
2. When describing the general characteristics and physiological function of P4 ATPases in different organisms the authors cite a number of reviews published from 1996 to 2016 (refs. 52-60), several of which are too old to reflect our knowledge of the proteins. In the past 4-5 years the field has made huge advances and several updated reviews have been published, including some specifically devoted to plant 1. The authors show that ala1 and ala7 mutants have defective accumulation of fungal toxins in their vacuoles and that this phenotype can be complemented by the expression of atALA1 and AtALA7, respectively. Therefore the authors conclude: "These data reveal that, in Arabidopsis, AtALA1 is responsible for the DON vacuole accumulation and AtALA7 is for the CIA vacuole accumulation".
In order to make this conclusion, the authors are lacking at least one additional KO line for each ALA that demonstrates that the phenotype is directly related to the ALA protein. While complementation shows that the ALA proteins can complement the phenotypes observed, the authors are overexpressing the proteins all over the plants using a 35S promoter. Thus, even if the phenotypes are not directly related to the deletion of ALA1 or ALA7 but to another gene deleted in the background, the ALA proteins might have an activity that can, when overexpressed, rescue the phenotype.
2. "DON5-FAM signal appeared in RFP-RabA1d-indicated EE-like structures (Fig. 2b); the signal was finally converged in vacuoles which were surrounded by the γ-Tip (Fig. 2c)." The authors do not make a time-course of the accumulation of fluorescent DON, so they should not make any statement that refers to this. In this case, "finally" implies that they are showing the substance travelling through EE to the vacuole with time, which is not the case.
3. "In Arabidopsis root cells, EGFP-AtALA1 localized at the PM and tonoplast, which was stained by FM4-64 (Fig. 2d,e)" The localization of the protein with the tonoplast is not at all clear. The authors present a picture in which FM4-64, which clearly colocalizes with the tonoplast does not really overlap with the signal for ALA1. ALA1 seems to be present in vesicular structures surrounding the vacuole (perhaps prevacuolar compartments) and a difuse membrane surrounding the vacuole (ER?), but there is no clear tonoplast colocalization. This is supported by the low Rp value.
4. "In tobacco epidermal cells, EGFP-AtALA1 was colocalized with RFP-RabF2a (Fig. 2g)." If has been shown before that ALA1 does not leave the ER in tobacco epidermal cells in the absence of a co-expressed beta subunit (López-Marqués, R.L. et al. (2012) A putative plant aminophospholipid flippase, the arabidopsis P4 ATPase ALA1, localizes to the plasma membrane following association with a β-subunit. PLoS One 7, e33042). When co-expressed with a beta subunit, ALA1 travels to the plasma membrane and no association with endomembranes was observed in these experiments. How can the authors explain this discrepancy? 5. "In the BFA and DON5-FAM treated cells, massive aggregates of DON5-FAM and BFA induced FM4-64 aggregates were formed, but the DON5-FAM signal was not colocalized to the BFA bodies (Fig. 2h)." Can the authors rule out that what they call "massive aggregates" are not simply an accumulation of DON in the vacuoles? Trafficking from the Golgi is affected by BFA, but the authors have not shown so far that the Golgi is involved in trafficking of DON to the vacuole. In fact, the signals obtained for BFA treatment for CIA-treated plants are clearly different than those obtained for DON, which suggests the two substances do not follow the same route to the vacuole.
6. The subcellular localization of ALA7 with GFP ( Figure 3e) is not convincing. The ALA7 signal looks like the ER membrane, which is attached to the plasma membrane through contact points even after plasmolysis. To date, no ALA protein has been shown to be able to leave the ER in the absence of a beta subunit. To make these experiments credible (both for ALA1 and ALA7), the tobacco localization needs to be repeated in the presence of a beta subunit. 7. Do ala1 or ala7 plants show any cellular phenotypes in vesicular trafficking that can support their role? All the experiments are based on protein overexpression and there is no direct evidence of the involvement of any of the two ALA proteins in vesicle formation under native conditions. Discussion 1. "With amphophilic structure, in the contrary, the ability of DON to enter the cells through diffusing across ..." Amphophilic means "having an affinity for both acid and basic dyes." Do the authors mean "amphiphilic" (having both hydrophilic and hydrophobic parts)? The term is used repeatedly in the text.
2. "Possibly, AtALA1 is responsible for transport of some amphophilic molecules like DON; while AtALA7 is responsible for that of lipophilic compounds as CIA." The authors have not provided evidence that there results can be generalized like this to other molecules, and should tone down this statement.
3. The authors seem to have forgotten to write a reference to Sup. Fig. 7 in the text. In this figure, they present a model of how ALA proteins might be involved in the observed detoxification effects. As mentioned in my comments above, the BFA treatment of plants suggests slightly different transport pathways to the vacuole for both substances and a putative model should reflect these differences.  2. In the generation of transgenic plants, please include the procedure for tobacco plants.

Subcellular localization
-Please indicate the literature reference or catalog number/accession number of the different organelle markers used. If plasmids were constructed ad-hoc for this work, please describe their construction. The authors might consider including a table of plasmids used.
-"For the localization of AtALA1, roots of EGFP::AtALA1 transgenic Arabidopsis were stained by FM4-64 (4 μM, F34653, Thermo Scientific) in EGFP-AtALA1 transgenic Arabidopsis.". Please rephrase -Please indicate what type of objective was used for the microscopy experiments and add the excitation and recording wavelengths for each fluorophore, as well as the pinhole aperture for each experiment. Were all images generated under identical pinhole apertures and with the same magnification?

Microscopy observation
Please specify in the title that this is referring to a set of specific experiments Reviewer #3 (Remarks to the Author): The authors identified Arabidopsis P4-ATPase genes, AtALA1 and AtALA7, responsible for the cell detoxification of the mycotoxins produced by Fusarium graminearum and Verticillium dahlia. They showed that by AtALA1-/AtALA7mediated vesicle transport, the toxins were compartmentalized in vacuoles for degradation. They further proved that overexpression of AtALA1 and AtALA7 significantly increased the resistance of transgenic Arabidopsis and maize plants to F. graminearum and V. dahlia.
Novel data are well presented and analyzed. Except there are many language and grammatic issues which are pointed out by the reviewer in the PDF file attached. The authors are encouraged to correct and edit these problems.

Reviewer #1 (Remarks to the Author):
Wang et al. in the manuscript entitled "Arabidopsis P4 ATPase-mediated cell detoxification confers the resistance to Fusarium graminearum and Verticillium dahliae" argue that ALA1 and ALA7 detoxify fungal toxins, DON and CIA, by trafficking them into vacuoles. All data are likely to support the authors' claim, but more precise experimental results should be presented to be published in the journal. In addition, the authors are likely to pick or select only a part of data to support their argument.
1. To show cellular localization of ALA1 and ALA7, the authors generated transgenic Arabidopsis plants expressing GFP or RFP-fused proteins by 35S promoter. Using these plants or sometimes by heterologous expression in tobacco with some plant organellar markers, the authors argue that ALA1/7 are transported from the PM via endosomes to vacuole. However, as the authors may know, by overexpression proteins can be localized to additional cellular compartments to their original sites. Therefore, the authors should generate and use plants expressing fusion proteins by native promoters to precisely observe their localization in plant cells.
Thanks! It's a good idea to use native promoter to control the GFP or RFP-fused proteins to reduce the mis-localization from overexpression. As suggested, we generated transgenic Arabidopsis in which the fusion proteins were under control of native promoters (proAtALA1::EGFP-AtALA1 and proAtALA7::AtALA7-eGFP). The new data were presented in Fig 1. 2. The authors found that DON or CIA are not accumulated in FM4-64-stained vacuoles in ala1 or ala7 mutant in Fig To avoid the possibility that the failure of the accumulation of CIA and DON in vacuoles is resulted from the PCD, we began the observation of the distribution of the toxins at earlier time (for DON, at 2h; for CIA at 1h). Based on our experiments, the earliest time window for observation of the appearance of CIA in pre-vacuoles is at ~2 hours after feeding the plant with CIA FITC ; for DON, the time is ~6h. We believe that there's not enough time for CIA (~2h) and DON (~6h) to damage the transport and vacuole compartmentation of the toxins. The distribution of DON and CIA at earlier time points was added in Fig. 1d and i.
3. To test disease resistance to fungi, the authors generated transgenic Arabidopsis plants overexpressing ALA1 or ALA7 only. I wonder whether transgenic plant overexpressing fluorescently tagged proteins also show similar enhanced resistance to tested fungi? Are there any specific reasons to use ALA1 or ALA7 only to generate transgenic plants for resistance test?
Yes, the transgenic Arabidopsis overexpressing fluorescently tagged proteins also show similar enhanced resistance to tested fungi with native protein (see image below). To avoid possible negative impact from GFP fusion on the function of the proteins when expressed in other plants (e.g., tobacco and maize), we used the native AtALA1, instead of EGFP-AtALA1 in transgenic maize.
4. The authors show leaf whitening (resistance to DON and CIA), root length (ALA7 overexpression) or plant weight (ALA1 overexpression) depending on an experiment. Are there any reasons for this? The authors should show all three kinds of results regardless of different experimental approaches.
We found that 1 µg of DON (mL -1 ) can significantly inhibit the growth of wild-type roots. However, to bleach the leaves, the concentration should be 50 μg·mL -1 . Therefore, we used different concentration of DON in different experiments. Same to the CIA. 5. In many figures, arrowheads are not clearly indicated.
Thanks, we replaced the arrowheads with better one. Fig. 2b and c were obtained at the same time point after DON treatment. Why DON is distinctly localized? In addition, why localization patterns in Fig. 2e are not observed in Fig. 2d? The manuscript presents the putative role of two P4 ATPases (ALA1 and ALA7) in detoxification of fungal toxins. P4 ATPases are lipid translocators involved in vesicle formation, lipid signaling and generation of an asymmetry lipid distribution that can be used to recruit proteins to the surface of the membrane. Based on KO mutants, protein overexpression and the use of secretory pathway inhibitors, the authors demonstrate that overexpression of arabidopsis ALA proteins can be used to increase the resistance of several plant species to fungal pathogens. While their findings are interesting, the mechanism behind the observed phenotype are not clear and several points require attention.

Pictures in
General comments 1. The English language relatively good, but there are many paragraphs where the grammatical constructions get too complicated or the use of the language is not correct. E.g. "This strategy enables hosts to be prevented from the toxicity of the mycotoxins" or "Recently, it was turned out that..."-The authors should consider professional language editing.
According to the suggestion, we carefully edited the language.
2. While I am not a statistics expert, I am not sure the statistical analysis is correct for most of the figures. In most figures, the authors present a Student's t-test, which is meant to compare two means. However, most experiments include three or more means, as there is always a control and at least two mutant/complemented lines. I believe the correct way to do this analysis is using a test of the type of ANOVA, which compares several different means, even if the authors are only interested in showing the difference between each mutant and the wild type. This type of analysis has been used for Sup. Fig. 2, for instance.
We appreciate this suggestion. Accordingly, we reanalyzed our data using one-way ANOVA with Tukey multiple comparisons test.
3. In general, it is difficult to assess the quality of the bioimaging work, such as the subcellular localizations and the secretory pathway inhibitor experiments, as no mention is made of number of repetitions and no quantifications are provided in most cases. The authors should clearly state how many repetitions of each experiment were made and provide quantifications and statistical analyses.
We added the descriptions of how many repetitions of each experiment, and provided quantifications and statistical analyses in the legends. Introduction 1. "The P4 subfamily of P-type ATPases (P4 ATPases) functions as phospholipid flippases that translocate specific phospholipid substrates from the exoplasmic/luminal leaflet to the cytoplasmic leaflet of biological membranes,..." All plant P4 ATPases characterized to date and most P4 ATPases from other organisms are phospholipid transporters, but several P4 ATPases from humans and yeast have now been shown to transport lipids not containing a phosphate group, so please correct his sentence to account for that fact.
Thanks a lot for the suggestion. We have adapted the description.
2. When describing the general characteristics and physiological function of P4 ATPases in different organisms the authors cite a number of reviews published from 1996 to 2016 (refs. 52-60), several of which are too old to reflect our knowledge of the proteins. In the past 4-5 years the field has made huge advances and several updated reviews have been published, including some specifically devoted to plant P4 ATPases. While some of the references could stay, at least the oldest one should be updated. In addition, the authors should avoid using original citations for the general subjects or they should otherwise be fair and cite all the original papers on the subject. Some suggestions for the authors: We appreciate this constructive criticism. Following the suggestion, we added some new references and removed some old ones. We had adopted these references.

Results
1. The authors show that ala1 and ala7 mutants have defective accumulation of fungal toxins in their vacuoles and that this phenotype can be complemented by the expression of atALA1 and AtALA7, respectively. Therefore the authors conclude: "These data reveal that, in Arabidopsis, AtALA1 is responsible for the DON vacuole accumulation and AtALA7 is for the CIA vacuole accumulation". In order to make this conclusion, the authors are lacking at least one additional KO line for each ALA that demonstrates that the phenotype is directly related to the ALA protein. While complementation shows that the ALA proteins can complement the phenotypes observed, the authors are overexpressing the proteins all over the plants using a 35S promoter. Thus, even if the phenotypes are not directly related to the deletion of ALA1 or ALA7 but to another gene deleted in the background, the ALA proteins might have an activity that can, when overexpressed, rescue the phenotype.
We agree with the reviewer that we need more KO lines for each ALA to demonstrate the reliability of the phenotypes. Accordingly, we selected two mutants of each ALA for analysis: for ALA1, they are ala1-8 (salk_002106) and ala1 (salk_056947); for ALA7, they are ala7 (salk_125598) and ala7-24 (salk_063917). The results were added in Fig 1. 2. "DON5-FAM signal appeared in RFP-RabA1d-indicated EE-like structures (Fig.  2b); the signal was finally converged in vacuoles which were surrounded by the γ-Tip (Fig. 2c)." The authors do not make a time-course of the accumulation of fluorescent DON, so they should not make any statement that refers to this. In this case, "finally" implies that they are showing the substance travelling through EE to the vacuole with time, which is not the case.
Thanks. It was deleted.
3. "In Arabidopsis root cells, EGFP-AtALA1 localized at the PM and tonoplast, which was stained by FM4-64 (Fig. 2d,e)" The localization of the protein with the tonoplast is not at all clear. The authors present a picture in which FM4-64, which clearly colocalizes with the tonoplast does not really overlap with the signal for ALA1. ALA1 seems to be present in vesicular structures surrounding the vacuole (perhaps prevacuolar compartments) and a diffuse membrane surrounding the vacuole (ER?), but there is no clear tonoplast colocalization. This is supported by the low Rp value.
We observed the localization of AtALA1 in proAtALA1::EGFP-AtALA1 transgenic Arabidopsis, and replaced the images with new one, in which the tonoplast colocalization of AtALA1 can be clearly observed (also see response to review1, question 6).
4. "In tobacco epidermal cells, EGFP-AtALA1 was colocalized with RFP-RabF2a (Fig. 2g)." If has been shown before that ALA1 does not leave the ER in tobacco epidermal cells in the absence of a co-expressed beta subunit (López-Marqués, R.L. et al. (2012) A putative plant aminophospholipid flippase, the arabidopsis P4 ATPase ALA1, localizes to the plasma membrane following association with a β-subunit. PLoS One 7, e33042). When co-expressed with a beta subunit, ALA1 travels to the plasma membrane and no association with endomembranes was observed in these experiments. How can the authors explain this discrepancy?
To confirm the PM localization of AtALA1, we constructed proAtALA1::EGFP-AtALA1 transgenic Arabidopsis. Similar to 35S::EGFP-AtALA1 Arabidopsis, the EGFP-AtALA1 signal appeared in the PM. We think the discrepancy may be from the different transformation methods: for the stable ectopic expression of EGFP-AtALA1 (i.e., ProAtALA1::EGFP-AtALA1) in Arabidopsis, the endogenous ALISs can help ALA1 traveling from ER to the plasma membrane; for the transient expression in tobacco, the trafficking needs the help of the co-expression of its binding partner, ALIS.
5. "In the BFA and DON5-FAM treated cells, massive aggregates of DON5-FAM and BFA induced FM4-64 aggregates were formed, but the DON5-FAM signal was not colocalized to the BFA bodies (Fig. 2h)." Can the authors rule out that what they call "massive aggregates" are not simply an accumulation of DON in the vacuoles? Trafficking from the Golgi is affected by BFA, but the authors have not shown so far that the Golgi is involved in trafficking of DON to the vacuole. In fact, the signals obtained for BFA treatment for CIA-treated plants are clearly different than those obtained for DON, which suggests the two substances do not follow the same route to the vacuole.
Our data show that the entry speed of CIA into cytosol is faster than that of DON ( Supplementary Fig. 3e,f). The BFA treated time for DON 5-FAM in Fig. 2h was 12 h, while that for CIA FITC in Fig. 3d was 3 h. To address this problem that "In fact, the signals obtained for BFA treatment for CIA-treated plants are clearly different than those obtained for DON", we prolonged the observing time for CIA from three hours previously to six hours. The distribution of CIA FITC signal under BFA treatment was similar to that of DON signal (see image below). We replaced the imagining in Fig 3 d with new one.
Yes, we do not show that the Golgi is involved in trafficking of DON to the vacuole. Both DON and CIA signal are independent of BFA body, confirming that the traffic of DON and CIA to vacuoles is independent of Golgi.
6. The subcellular localization of ALA7 with GFP ( Figure 3e) is not convincing. The ALA7 signal looks like the ER membrane, which is attached to the plasma membrane through contact points even after plasmolysis. To date, no ALA protein has been shown to be able to leave the ER in the absence of a beta subunit. To make these experiments credible (both for ALA1 and ALA7), the tobacco localization needs to be repeated in the presence of a beta subunit.
We appreciate this suggestion. Accordingly, we conducted co-expression of proAtALA7::AtALA7-eGFP with a β-subunit, ALIS1. As you pointed out, AtALA7 indeed needs β-subunit to exit from ER ( Supplementary Fig.7). 7. Do ala1 or ala7 plants show any cellular phenotypes in vesicular trafficking that can support their role? All the experiments are based on protein overexpression and there is no direct evidence of the involvement of any of the two ALA proteins in vesicle formation under native conditions. Yes, we agree with that although we observe the bubble-like formation containing signals of the two proteins and two toxins emerged on the PM (Fig 1d;  Fig 1i and 3b), details about the ALA proteins in vesicle formation are required. Nevertheless, the dual labeling of ALA proteins and their transport cargos described here allows us to observe the real time process of cargo trafficking as well as dynamic alteration of vesicles, EEs, LEs, and vacuoles in vivo, thus providing a platform to study the subcellular processes of P4-ATPase mediated vesicle transport and the important aspects of the compartmental detoxification in living cells. Discussion 1. "With amphophilic structure, in the contrary, the ability of DON to enter the cells through diffusing across ..." Amphophilic means "having an affinity for both acid and basic dyes." Do the authors mean "amphiphilic" (having both hydrophilic and hydrophobic parts)?
The term is used repeatedly in the text.
Yes, we mean DON is an amphiphilic compound. Thank you for point out this mistaken.
2. "Possibly, AtALA1 is responsible for transport of some amphophilic molecules like DON; while AtALA7 is responsible for that of lipophilic compounds as CIA." The authors have not provided evidence that there results can be generalized like this to other molecules, and should tone down this statement.
In our previous study, we identified a P4-ATPse, BbCrpa, from entomopathogenic fungus, Beauveria bassiana. We show that BbCrpa is capable of transporting cyclosporine A (CsA, a lipophilic cyclic polypeptide), tacrolimus (FK506), as well as CIA. All these three compounds are lipophilic. This finding promotes us to identify the functional homologues of BbCrpa in Arabidopsis. In this study, we found ALA7 is responsible for transport of CIA, while ALA1 is for amphiphilic DON. The results from BbCrpa and ALA7/1 encourage us to propose this hypothesis. Of course, this hypothesis needs further study, which is our next research focusing on. According to the comment, in the revised manuscript, the hypothetic statement was weakened.
3. The authors seem to have forgotten to write a reference to Sup. Fig. 7 in the text. In this figure, they present a model of how ALA proteins might be involved in the observed detoxification effects. As mentioned in my comments above, the BFA treatment of plants suggests slightly different transport pathways to the vacuole for both substances and a putative model should reflect these differences.
Thank you for pointing out this omission. We added the reference to the modified text. By the way, we move the model from Supplementary into the text (Fig. 7). We observed the BFA treatment again and did not found significant difference between DON and CIA-treated cells.
4. The role of ALAs in detoxification of fungal toxins has not yet been described, but a role on detoxification of heavy metals was found for ALA4 We had replaced the image with better one. It had been added.
2. In the generation of transgenic plants, please include the procedure for tobacco plants.
It had been done.
3. Subcellular localization -Please indicate the literature reference or catalog number/accession number of the different organelle markers used. If plasmids were constructed ad-hoc for this work, please describe their construction. The authors might consider including a table of plasmids used.
A table of plasmids had been supplied in the revised manuscript (Supplementary Table 3).
-Please indicate what type of objective was used for the microscopy experiments and add the excitation and recording wavelengths for each fluorophore, as well as the pinhole aperture for each experiment. Were all images generated under identical pinhole apertures and with the same magnification?
Thanks! It had been done.

Microscopy observation
Please specify in the title that this is referring to a set of specific experiments Thanks！It had been done.
Reviewer #3 (Remarks to the Author): The authors identified Arabidopsis P4-ATPase genes, AtALA1 and AtALA7, responsible for the cell detoxification of the mycotoxins produced by Fusarium graminearum and Verticillium dahlia. They showed that by AtALA1-/AtALA7-mediated vesicle transport, the toxins were compartmentalized in vacuoles for degradation. They further proved that overexpression of AtALA1 and AtALA7 significantly increased the resistance of transgenic Arabidopsis and maize plants to F. graminearum and V. dahlia. Novel data are well presented and analyzed. Except there are many language and grammatic issues which are pointed out by the reviewer in the PDF file attached. The authors are encouraged to correct and edit these problems.
Thank you very much for your editing. According the suggestion, we carefully correct and edit these problems.