Auxin-sensitive Aux/IAA proteins mediate drought tolerance in Arabidopsis by regulating glucosinolate levels

A detailed understanding of abiotic stress tolerance in plants is essential to provide food security in the face of increasingly harsh climatic conditions. Glucosinolates (GLSs) are secondary metabolites found in the Brassicaceae that protect plants from herbivory and pathogen attack. Here we report that in Arabidopsis, aliphatic GLS levels are regulated by the auxin-sensitive Aux/IAA repressors IAA5, IAA6, and IAA19. These proteins act in a transcriptional cascade that maintains expression of GLS levels when plants are exposed to drought conditions. Loss of IAA5/6/19 results in reduced GLS levels and decreased drought tolerance. Further, we show that this phenotype is associated with a defect in stomatal regulation. Application of GLS to the iaa5,6,19 mutants restores stomatal regulation and normal drought tolerance. GLS action is dependent on the receptor kinase GHR1, suggesting that GLS may signal via reactive oxygen species. These results provide a novel connection between auxin signaling, GLS levels and drought response.

In this manuscript, Salehin et al. explore the role of three Aux/IAA Arabidopsis proteins (IAA5,IAA6 and IAA19) in controlling drought resistance via glucosinolate (GLS) level regulation. In previous work by the same group it was shown that these Aux/IAA genes are involved in drought tolerance and direct upstream regulators were described. Here they characterize the downstream events by utilizing triple recessive iaa5 iaa6 iaa19 mutants. By RNA-seq they identified 12 genes that function in GLS biosynthesis that are downregulated in the triple mutant. They quantified the levels of GLS in the triple mutant and found one GLS (4-MSO) significantly downregulated. They tested mutations in GLS biosynthetic genes (CYP79F1 and F2, CYP83A1) and found that these mutants show decreased drought tolerance. These genes are regulated by two TFs, MYB28 and 29, and discovered that the double myb28 myb29 mutant is less drought tolerant. Conversely, overexpression lines show increased drought tolerance. By ChIP-PCR and RNA-seq they identified another TF WRKY63 as a direct target of IAA19 which in turn represses both MYB gene expression. Overexpression of WRKY63 results in decreased drought tolerance. Then they investigate whether this pathway affects stomata closure by testing mutants in Aux/IAA and MYB genes and treatments with 4-MSO and determined that failure to close stomata is responsible for decreased drought tolerance. They further determined that this pathway is independent from the plant hormone ABA role in stomata regulation and that it likely involves ROS and the receptor kinase GRH1 since the mutant grsh1 is resistant to 4-MSO applications. Finally, they determined that mutants in the auxin co-receptors AFB family results in increased tolerance to drought.
Understanding the molecular mechanisms regulating tolerance to drought stress is essential if we want to improve crops tolerance to extreme temperatures, so the topic of this manuscript is of broad significance. The manuscript in general is clear, concise and well-written, and the main conclusions are justified.
I have a few comments and suggestions below: 1. It is unclear whether the recombineered IAA19 line used for ChIP-PCR rescues the phenotype of the triple mutant. If this was shown in previous publication it should be mentioned, otherwise this data should be presented. 2. Fig. 1B, what is the scale in the heatmap? 3. There is quite a bit of variability in rosette dry weight quantifications in different experiments; see . Is this normal? 4. In Fig.1 A, 13 genes are highlighted as differentially regulated but 12 are mentioned in the text. 5. CYP79F2 is not down-regulated in the triple iaa5 iaa6 iaa19 mutant but cyp79f2 mutant does show less resistance to drought stress and they are both regulated by MYB28 and 29. How do you account for this difference given that in the triple mutant both MYBs are downregulated? 6. There is no mention of Fig. S4 in the text. I would also recommend the authors to include their findings on WRKY/MYB in the model. 7. Page 2: " We confirmed some of these results by…" 8. Page 3: introduce the AOP2 gene. 9. Page 4. Describe the abi1-1 mutant when first introduced in the text. 10. Please specify how many bio-replicates were used in qRT-PCR experiments. 11. Fig 4C, the scale bar is over the text; please fix. 12. Page 3: "…further support for the idea that GLSs are…" 13. Page 4: introduce myrosinase TGG1.
Reviewer #2 (Remarks to the Author): Glucosinolates (GLSs) are secondary metabolites that serve under biotic stress challenges. Here authors show that in Arabidopsis, aliphatic GLSs can additionally play a role in abiotic stress response as they are stimulated by the auxin sensitive Aux/IAA repressors IAA5, IAA6, and IAA19 under drought conditions. Loss of IAA5/6/19 results in reduced GLS levels and decreased drought tolerance, which is further associated with a defect in stomatal regulation. Remarkably, authors recorded a complementation of stomatal closure by overexpression of MYBs regulating aliphatic GSLs and by exogenous application of aliphatic GSL 4MSO in IAA5/6/19 as well as myb28/myb29 mutants. Although it was not possible to explain what makes these metabolites so special in regulating stomatal closure or to track the direct regulator of aliphatic GSL biosynthesis and the direct regulator(s) of improved drought tolerance downstream of IAA5/6/19, this work is still a highlight. It shows that secondary metabolites can have functions beyond "secondary metabolic pathways" by attributing novel features of aliphatic GSL in drought stress. Series of experiments are comprehensive and results presented solid. I assume that findings of this work will melt borders of secondary and primary metabolism and help addressing improvement of plants growth under drought stress conditions. Here are some of the questions I would like to be answered before publication: -What makes aliphatic GSLs so special in their ability to regulate e.g. stomatal closure? -Are aliphatic GSLs or ITCs the active molecules targeting essential component in plant cells? -What are potential targets of these metabolic in plant cells vs human cell? -How 4MSO is used to change stomata aperture in cells? Is active transport from vacuole or Scells is needed? -Do other metabolites will take over the function of aliphatic GSL in species containing no GSLs? -Mutual negative regulation of indolic and aliphatic GSL pathways is known to take place upon overproduction of one of these two classes of metabolites. Could it be that plants measure levels of indolic GSLs produced by monitoring levels of IAA hormone and downregulating production of aliphatic GSL as a response to high IAA levels? -A schematic drawing which shows common biosynthetic origin of IAA metabolite and indolic GSL along with mutual reciprocal negative regulation of indolic and aliphatic branches will be immensely helpful (in Supplemental Data). -This reviewer also wonders whether mechanism ascribed to aliphatic GSLs are as GSL causative as proposed. Figure 2 and Figure 4 present very impressive results suggesting this should be the case. But, is it possible that MYB28 and MYB29 have target genes beyond GSL and related to drought response? Public microarray or RNA-seq data? Furthermore, could it be that treatment of Arabidopsis plants with 4MSO induce expression of MYB28 or 29 followed by activation of non-GSL genes. Addressing/discussing a potential of having non GSL target genes in the arsenal of MYB28/29 regulators would help unveiling this doubt. Minor points: -For some but not all experiments authors used the cyp79f1/f2 mutants. Need to be mentioned why is this so and whether authors observed the phenotypic features of cyp79f1/f2 khown as "bushy". Do cyp79f1/f2 plants have different IAA levels? -Data set presented in Figure 2 (complementation of drought stress phenotype of IAA19 mutant by ovexpression of MYBs regulating GSLs) is very impressive and very nice. Is there a reason why authors used a single iaa19 mutant and not iaa5/6/19 mutant.
-Complementation of iaa19 with MYB34 or 51 would be of interest. It can eventually address the potential of indolic MYBs (and ancestor genes of MYB28 and MYB29) having ability to activate genes beyond GSLs.
Reviewer #3 (Remarks to the Author): The authors clarified that aliphatic glucosinolate (GLS) levels are regulated by the auxin-sensitive Aux/IAA repressors IAA5, IAA6, and IAA19 using Arabidopsis. These proteins act in a transcriptional cascade that maintains expression of GLS levels in plants under drought conditions. Loss of IAA5, IAA6, and IAA19 reduced GLS levels and decreased drought tolerance through stomatal regulation. In addition, application of GLS to the iaa5 iaa6 iaa19 mutants restores stomatal regulation and normal drought tolerance. The authors claimed GLS action is dependent on the receptor kinase GHR1. This topic is very interesting but there are several points to be addressed and to be improved.
The authors propose the model (Fig. S4.). In this model, myrosinase mediates the reaction from GLS to ITC. Where is the myrosinase present? Also where is the GLS? The authors tested effects of exogenous GLS on stomatal responses? In this case, can GLS be contacted with myrosinase?
In addition, the authors should provide information about solvents for 4-MSO and I3M. Islam et al. reported that ITC inhibited potassium inward-rectifying channels. The authors should cite this paper and discuss their results because they examined light-induced stomatal opening, which is strongly regulated by the potassium channels. The authors should measure ITC contents if proposing the model.
We are very grateful to the reviewers for their many insightful comments. We have tried to address each of the reviewers concerns as described below. As we considered the revised manuscript, we also realized that we had used the wrong wild-type control (Col-0 instead of Ws) for experiments involving the cyp79f1f2 double mutant. We have repeated these experiments using the Ws control line.

Mark Estelle
Reviewers' comments: Reviewer #1 (Remarks to the Author): In this manuscript, Salehin et al. explore the role of three Aux/IAA Arabidopsis proteins (IAA5,IAA6 and IAA19) in controlling drought resistance via glucosinolate (GLS) level regulation. In previous work by the same group it was shown that these Aux/IAA genes are involved in drought tolerance and direct upstream regulators were described. Here they characterize the downstream events by utilizing triple recessive iaa5 iaa6 iaa19 mutants. By RNA-seq they identified 12 genes that function in GLS biosynthesis that are downregulated in the triple mutant. They quantified the levels of GLS in the triple mutant and found one GLS (4-MSO) significantly downregulated. They tested mutations in GLS biosynthetic genes (CYP79F1 and F2, CYP83A1) and found that these mutants show decreased drought tolerance. These genes are regulated by two TFs, MYB28 and 29, and discovered that the double myb28 myb29 mutant is less drought tolerant. Conversely, overexpression lines show increased drought tolerance. By ChIP-PCR and RNA-seq they identified another TF WRKY63 as a direct target of IAA19 which in turn represses both MYB gene expression. Overexpression of WRKY63 results in decreased drought tolerance. Then they investigate whether this pathway affects stomata closure by testing mutants in Aux/IAA and MYB genes and treatments with 4-MSO and determined that failure to close stomata is responsible for decreased drought tolerance. They further determined that this pathway is independent from the plant hormone ABA role in stomata regulation and that it likely involves ROS and the receptor kinase GRH1 since the mutant grsh1 is resistant to 4-MSO applications. Finally, they determined that mutants in the auxin co-receptors AFB family results in increased tolerance to drought.
Understanding the molecular mechanisms regulating tolerance to drought stress is essential if we want to improve crops tolerance to extreme temperatures, so the topic of this manuscript is of broad significance. The manuscript in general is clear, concise and well-written, and the main conclusions are justified.
I have a few comments and suggestions below: 1. It is unclear whether the recombineered IAA19 line used for ChIP-PCR rescues the phenotype of the triple mutant. If this was shown in previous publication it should be mentioned, otherwise this data should be presented. Shani et al., 2017. 2. Fig. 1B, what is the scale in the heatmap?

Yes, the characterization of that line, including its ability to rescue the iaa19 mutant, was presented in
Our apologies. The scale disappeared in the PDF. Fig 2B-E for example (from 0.35-0.6). Is this normal?

There is quite a bit of variability in rosette dry weight quantifications in different experiments; see Col-O in
We have noted this as well. We think this is due to seasonal differences in humidity. Our chambers don't have humidity regulation. It is important to note that each experiment was repeated three times and the same trend was observed in each case. Fig.1 A, 13 genes are highlighted as differentially regulated but 12 are mentioned in the text.

Corrected.
5. CYP79F2 is not down-regulated in the triple iaa5 iaa6 iaa19 mutant but cyp79f2 mutant does show less resistance to drought stress and they are both regulated by MYB28 and 29. How do you account for this difference given that in the triple mutant both MYBs are downregulated?
CYP79F2 was downregulated in the RNAseq data but with a FDR of 0.049, it was below our FDR threshold of 0.001. Nevertheless, we assayed CYP79F2 expression by qRT-PCR. The results in Fig S1B show that this gene is downregulated in the mutant, although not as much as some of the other genes in the pathway.
6. There is no mention of Fig. S4 in the text. I would also recommend the authors to include their findings on WRKY/MYB in the model.

Done.
9. Page 4. Describe the abi1-1 mutant when first introduced in the text.

Done.
10. Please specify how many bio-replicates were used in qRT-PCR experiments.

All qRT-PCR experiments had 3 bioreplicates. We now indicate this in the Methods.
11. Fig 4C, the scale bar is over the text; please fix.
The myrosinase is already introduced on page 1.

Reviewer #2 (Remarks to the Author):
Glucosinolates (GLSs) are secondary metabolites that serve under biotic stress challenges. Here authors show that in Arabidopsis, aliphatic GLSs can additionally play a role in abiotic stress response as they are stimulated by the auxin sensitive Aux/IAA repressors IAA5, IAA6, and IAA19 under drought conditions. Loss of IAA5/6/19 results in reduced GLS levels and decreased drought tolerance, which is further associated with a defect in stomatal regulation. Remarkably, authors recorded a complementation of stomatal closure by overexpression of MYBs regulating aliphatic GSLs and by exogenous application of aliphatic GSL 4MSO in IAA5/6/19 as well as myb28/myb29 mutants. Although it was not possible to explain what makes these metabolites so special in regulating stomatal closure or to track the direct regulator of aliphatic GSL biosynthesis and the direct regulator(s) of improved drought tolerance downstream of IAA5/6/19, this work is still a highlight. It shows that secondary metabolites can have functions beyond "secondary metabolic pathways" by attributing novel features of aliphatic GSL in drought stress. Series of experiments are comprehensive and results presented solid. I assume that findings of this work will melt borders of secondary and primary metabolism and help addressing improvement of plants growth under drought stress conditions. Here are some of the questions I would like to be answered before publication: -What makes aliphatic GSLs so special in their ability to regulate e.g. stomatal closure?
At this point we don't have an answer to this important question. Because we have tested a small number of compounds, we hesitate to draw strong conclusions. We expect that specificity is related to the interaction between the ITC, and its target, perhaps a peroxidase. These intriguing questions await further experimentation.
-Are aliphatic GSLs or ITCs the active molecules targeting essential component in plant cells?
Based on Sally Assmann's work (Zhao et al., 2008), ITCs appear to be the active molecule. We now state this more directly on page 4. We have also shown the tgg1 tgg2 double mutant is sensitive to growth on medium containing PEG indicating that a myrosinase dependent breakdown product of the GLS is required for stress tolerance. We now include this data in Fig  S3C. -What are potential targets of these metabolic in plant cells vs human cell?
Our understanding is that the targets of glucosinolates in human cells are largely unknown. In any case, I don't think our study speaks to this at present. Perhaps in the future. In plants, the best-known targets of glucosinolates are insect foragers and some fungal and bacterial pathogens, although even in these cases, the molecular targets are not well understood. With respect to our work, based on earlier studies we propose that the ITCs promote ROS formation through a peroxidase. A confirmation of this model awaits further experimentation.
-How 4MSO is used to change stomata aperture in cells? Is active transport from vacuole or Scells is needed?
This is an interesting topic that we don't address in our paper. The source of the GLS that regulates the stomata is currently unknown, although we have preliminary data that addresses this question. We prefer to hold this data for another manuscript.
-Do other metabolites will take over the function of aliphatic GSL in species containing no GSLs?
This is an excellent question. Unfortunately, we don't know the answer. Previous work has shown that ITCs regulate stomata in Vicia faba and it is possible that endogenous ITC is produced through a different pathway in this and other plants. We now mention this speculation on page 6.
-Mutual negative regulation of indolic and aliphatic GSL pathways is known to take place upon overproduction of one of these two classes of metabolites. Could it be that plants measure levels of indolic GSLs produced by monitoring levels of IAA hormone and downregulating production of aliphatic GSL as a response to high IAA levels?
This is a very interesting hypothesis that we, or perhaps others, can test in the future.
-A schematic drawing which shows common biosynthetic origin of IAA metabolite and indolic GSL along with mutual reciprocal negative regulation of indolic and aliphatic branches will be immensely helpful (in Supplemental Data).
We agree that these are very interesting questions, and no doubt the regulation of glucosinolate biosynthesis and transport is complex. However, we are not convinced that such a figure would be helpful given the more focused nature of our study. We would prefer not to include this unless the reviewer insists.
-This reviewer also wonders whether mechanism ascribed to aliphatic GSLs are as GSL causative as proposed. Figure 2 and Figure 4 present very impressive results suggesting this should be the case. But, is it possible that MYB28 and MYB29 have target genes beyond GSL and related to drought response? Public microarray or RNA-seq data? Furthermore, could it be that treatment of Arabidopsis plants with 4MSO induce expression of MYB28 or 29 followed by activation of non-GSL genes. Addressing/discussing a potential of having non GSL target genes in the arsenal of MYB28/29 regulators would help unveiling this doubt.
Yes, we considered this possibility. However, because over-expression of AOP2, a downstream enzyme in GLS biosynthesis, also results in drought tolerance, we don't think other MYB28/29 targets are likely to be causal.
Minor points: -For some but not all experiments authors used the cyp79f1/f2 mutants. Need to be mentioned why is this so and whether authors observed the phenotypic features of cyp79f1/f2 khown as "bushy". Do cyp79f1/f2 plants have different IAA levels?
Both the cyp79f1/f2 and the myb28 myb29 lines have very low levels of GLS. We chose to work with the latter because it is representative.
The cyp79f1f2 double mutant does exhibit a bushy phenotype in our hands as has been described, and it has also been reported that these plants exhibit higher auxin levels (Tantikanjana et al., Genes and Dev., 2001). However, we think that it is unlikely that the change in auxin levels results in decreased tolerance since none of the other genotypes we tested (cyp83a1, myp28/29, 35S:MYB28 etc) exhibit an auxin related phenotype. Also, the fact that directly application of GLS can rescue the drought phenotype strongly argues that a change in GLS levels is responsible for this phenotype.
-Data set presented in Figure 2 (complementation of drought stress phenotype of IAA19 mutant by ovexpression of MYBs regulating GSLs) is very impressive and very nice. Is there a reason why authors used a single iaa19 mutant and not iaa5/6/19 mutant.