Abstract
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.
One Sentence Summary Aux/IAA proteins promote drought tolerance by regulating glucosinolate levels.
Main Text
The Aux/IAA transcriptional repressors have a central role in auxin signaling. In the presence of auxin, the Aux/IAAs are degraded through the action of the ubiquitin E3-ligase SCFTIR1/AFB, resulting in de-repression of transcription by the AUXIN RESPONSE FACTOR (ARF) transcription factors(1, 2). Although the mechanisms of auxin perception and Aux/IAA degradation are well known, other aspects of Aux/IAA regulation remain poorly understood. In particular, the factors that regulate transcription of the Aux/IAA genes are mostly unknown. Recently, we showed that three Aux/IAA genes, IAA5, IAA6, and IAA19 are directly regulated by the DREB2A and DREB2B transcription factors and that recessive mutations in these IAA genes result in a decrease in drought tolerance (3).
To determine the molecular basis of the loss of drought tolerance in iaa5, iaa6, and iaa19 mutant plants, we used RNAseq to identify genes that were differentially regulated in Col-0 vs. the iaa5 iaa6 iaa19 (iaa5,6,19) triple mutant when exposed to desiccation stress (Table S1). A total of 651 genes were differentially expressed between the mutant and Col-0 under these conditions (FDR < 0.001), 439 down-regulated and 212 up-regulated. A gene ontology search revealed that 12 genes that function in the aliphatic glucosinolate (GLS) biosynthetic pathway are down-regulated in iaa5,6,19 (Fig. 1A, B). In contrast, expression of these genes is not significantly affected by dehydration stress in Col-0. We confirmed these results by quantitative RT-PCR (qRT-PCR) (Fig. S1A).
Because the aliphatic GLS biosynthetic enzymes are down-regulated in iaa5,6,19 mutants during drought stress, we wondered if the levels of GLSs were also affected. To test this, we measured GLS levels in stress treated Col-0 and iaa5,6,19 mutants at time intervals. Indolic GLSs were unaltered (Table S2). However, the level of 4-methylsulfinyl glucosinolate (4-MSO), the most abundant aliphatic glucosinolate in Arabidopsis (Col-0), was sharply decreased in iaa5, 6, 19 plants after 1 hr and 3 hr of desiccation (Fig. 1C). This data shows that down-regulation of aliphatic GLS biosynthetic enzymes in iaa5, 6, 19 mutants results in decreased GLS levels.
GLSs are well known for their role in plant defense and innate immunity, although recent studies also suggest that they may have a role in regulating plant growth (4-8). These secondary metabolites are found primarily in the Brassicaceae, a family that includes many economically important crops (6). GLSs are broken down by the enzyme myrosinase into isothiocyanates and related compounds, which are toxic to insect herbivores and other plant pathogens (7-10). To determine if decreased drought tolerance in the iaa5,6,19 mutant is related to reduced GLS levels, we measured the effects of mutations in GLS biosynthetic genes on response to drought. We employed two assays; growth of seedlings on agar medium containing PEG, and growth of plants in pots after withholding water. We first characterized mutants in the CYP79F1 and CYP79F2 genes, encoding enzymes that convert elongated methionine to aldoximes (Fig. 1A). Both mutants, as well as the double mutant, are less tolerant to both PEG treatment and water withholding (Fig. 2A, B; Fig. S2B, C). Similarly, loss of CYP83A1, responsible for conversion of aldoximes to aci-Nitro compounds, results in reduced tolerance to water withholding in pots (Fig. S2D). These results indicate that loss of aliphatic GLS compounds results in decreased drought tolerance, providing an explanation for the phenotype of the iaa5,6,19 mutant.
The MYB28 and MYB29 transcription factors are known to regulate genes in the aliphatic GLS biosynthetic pathway, including the CYP79F1/F2 and CYP83A1 genes (11-14). Indeed, aliphatic GLS levels in the myb28 myb29 double mutant are extremely low, whereas over-expression of either MYB28 or MYB29 results in elevated aliphatic GLS levels (12-15). Thus, it is possible that the effects of the iaa5,6,19 mutations on the pathway are mediated by changes in expression of these transcription factor genes. Examination of our RNAseq data revealed that MYB28 is down-regulated in the triple mutant in response to stress. MYB29 is expressed at a very low level in seedlings. We confirmed this by qRT-PCR (Fig. S1). When we determined the response of the myb28 mby29 double mutant to water withholding we found that it is less tolerant than the wild type, further support for the idea and GLSs are required for drought tolerance (Fig. 2C; Fig. S3A). Since overexpression of MYB28 and MYB29 increases GLS levels, we also tested the behavior of these lines in our assays. Strikingly, we found that both lines displayed strongly increased drought tolerance (Fig. 2D, Fig. S3B) further confirming that GLSs confer drought tolerance. We also tested the effect of over-expression of the AOP2 gene on drought tolerance. Over-expression of this gene in Col-0 results in an increase in aliphatic GLS levels, although less than in lines over-expressing MYB28 (16). The results in Fig 2D and Fig. S3B show that increased AOP2 levels results in a modest but statistically significant increase in drought tolerance.
If decreased drought tolerance in iaa5,6,19 plants is due to a GLS deficiency then over-expression of MYB28 or MYB29 in the triple mutant may ameliorate the effects of the mutations. Indeed, when we cross the 35S:MYB28 or 35S:MYB29 transgene into the iaa19-1 mutant the result is the restoration of wild-type levels of drought tolerance to the mutant line (Figure 2E).
Although IAA5, 6, and 19 are required for expression of MYB28, they are unlikely to directly regulate MYB28 because they are transcriptional repressors (1). To identify transcription factors that might be direct targets of the Aux/IAAs, and that might regulate MYB28 expression, we searched our RNAseq data for factors that are up-regulated in the triple mutant in response to stress. One interesting candidate was WRKY63, also known as ABA OVERLY SENSITIVE3 (abo3)(17). The abo3 mutant was originally isolated because it is hypersensitive to ABA in seedlings. Our qRT-PCR experiments confirmed that WRKY63 is up-regulated in response to dehydration in the triple mutant compared to the wild type (Fig. 3A). Examination of the promoter region of the WRKY63 gene revealed the presence of two tandemly repeated AuxRE elements (−350 to −361), known to bind ARF transcription factors (2). To determine if IAA19 binds to these sequences, we performed a ChIP-PCR analysis using a rAA19-YPet-His-FLAG line. The results, shown in Fig 3B, show that recovery of AuxRE sequences is significantly enriched in the IP from rIAA19-YPet-His-FLAG compared to the control indicating that IAA19 binds to this sequence, presumably indirectly through an interaction with an ARF transcription factor.
To determine if WRKY63 might regulate MYB28/29 expression, we measured RNA levels in two 35S:WRKY63 lines by qRT-PCR and found that expression of both genes was reduced suggesting that WRKY63 acts to repress expression of the MYB genes (Fig 3C) (18). Finally, we measured drought tolerance of the 35S:WRKY63 lines as well as the knockout mutant wrky63-1. As shown in Fig. 3D and Fig. S3B, the mutant has a normal response to water withholding. However, both over-expression lines are less drought tolerant consistent with reduced expression of MYB28/29. We note that WRKY63 is a member of small clade of 4 genes. It is possible that these genes have an overlapping function in regulation of MYB28/29. We note that our analysis of WRKY63 differs from the earlier work showing that the abo1 mutant is less drought tolerance (17). The reason for this discrepancy is unclear.
In considering how aliphatic GLSs may regulate drought response, we noted that these compounds have been reported to promote stomatal closure. In addition, the myrosinase TGG1 is one of the most abundant proteins in Arabidopsis guard cells suggesting that GLS compounds may have an important role in guard cell function (19). Further, isothiocyanate, a GLS metabolite, closes stomata in Arabidopsis (20, 21). To determine if a defect in stomatal regulation might be responsible for reduced drought tolerance in the iaa5,6,19 and myb28 myb29 lines, we first examined the stomatal response to drought in epidermal peels. As expected, the stomata on Col-0 plants close in response to drought conditions (Fig. 4A). In contrast, the stomata on both mutant lines failed to respond. We next asked whether application of 4-MSO would promote stomatal closure in wild-type and mutant plants. As a control we also applied abscisic acid (ABA) The results shown in Fig. 4B demonstrate that all three genotypes respond to both ABA and 4-MSO. We further tested the effects of 4-MSO on light-induced opening of stomata from dark-adapted plants. The results in Fig 4C show that 4-MSO inhibits this response in Col-0 and iaa5,6,19 plants. These results suggest that the primary basis for reduced drought tolerance in the iaa5,6,19 line is failure to close stomata in drought conditions. To test this possibility, we applied 4-MSO to wild-type and mutant plants subjected to water withholding in pots. The results in Fig. 4D and Fig. S2A show that application of this GLS restores drought tolerance in the mutant to wild-type levels.
To determine if all GLS compounds promote stomatal closure, we applied indol-3-ylmethylglucosinolate (I3M), an indolic GLS that is abundant in Arabidopsis, to Col-0. The results in Fig. 4C demonstrate that I3M is unable to inhibit light-induced stomatal opening in either Col-0 or the triple mutant, indicating that not all glucosinolates have the same efficacy. Similar results were obtained in an experiment where we tested the ability of I3M to promote stomatal closure in light grown plants (Fig 4E). We also tested a second aliphatic GLS, sinigrin hydrate (SH) that is abundant in many Arabidopsis ecotypes, although not Col-0 (22). Like 4-MSO, SH promoted closure in Col-0 and in the abi1-1 mutant (Fig S3C). Thus, it is possible that this activity is restricted to aliphatic GLS compounds, although additional studies are required to explore this possibility further.
ABA is known to play a key role in stomatal regulation (23). We investigated potential interactions between GLS and ABA by testing the response of iaa5,6,19 to ABA treatment. The results in Fig. 4B show that the mutant line does respond to ABA, indicating that GLS is not required for the ABA response. Further, we found that the abi1 mutant, which is deficient in ABA regulation of stomatal closure, has a normal response to GLS (Fig. 4F) (23). These results suggest that GLS and ABA can act independently to regulate stomata. This is consistent with a recent study showing that GLS and ABA have an additive effect on stomatal closure (24).
Reactive oxygen species (ROS) play an important role in stomatal regulation. One of the effects of ABA is to stimulate production of extracellular ROS through the RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD)(23). Extracellular ROS than acts through the receptor kinase GUARD CELL HYDROGEN PEROXIDE-RESISTANT 1 (GHR1) to regulate the SLOW ANION ACTIVATING CHANNEL 1(23). Since isothiocyanates (ITC), products of GLS metabolism, are known to promote stomatal closure via ROS, we wondered if GHR1 is required for response to GLS(21). The results in Fig. 4G show that the grh1 mutant is resistant to the effects of 4-MSO on stomata indicating that GLS likely acts through production of ROS.
Since loss of the auxin-sensitive repressors IAA5, IAA6, and IAA19 in the iaa5,6,19 results in decreased expression of GLS biosynthetic genes, we predict that the reduction in the levels of these proteins after auxin treatment would have the same effect. Indeed, treatment of seedlings with 10 μM IAA for 2 hours results in reduced expression of the GLS genes (Fig. S3D). Further, we predict that mutations which stabilize the Aux/IAA proteins will increase both GLS levels and drought tolerance. The TIR1/AFB family of auxin co-receptors consists of 6 members in Arabidopsis that act in an overlapping fashion to regulate auxin-dependent transcription throughout the plant. To determine the role of the TIR1/AFBs in GLS regulation we measured GLS levels in two higher order tir1/afb lines, afb1,3,4,5 and afb2,3,4,5 (25). Both lines had significantly higher GLS levels than either Col-0 or iaa5,6,19 in 7-day-old seedlings (Fig. 4H). After one hour of desiccation, GLS levels dropped in all three mutant genotypes. In the case of the iaa5,6,19 line, GLS levels were much lower than Col-0, while in afb1,3,4,5 and afb2,3,4,5, levels were similar to Col-0. We also determined the effects of water withholding on these genotypes. The results in Fig 4I and Fig. S3E show that both afb lines are significantly more drought tolerant than Col-0.
The major plant hormones are all secondary metabolites with ancient signaling functions. Here we show that GLSs also act as chemical signals that regulate stomatal aperture during drought stress. Desiccation results in a rapid decrease in auxin response in seedlings, an effect that probably results in decreased growth(3). By utilizing auxin to regulate stomatal aperture, the plant may integrate growth and stomatal regulation in drought conditions.
Funding
This work was supported by grants from the NIH (GM43644 to ME; JRE), the Howard Hughes Medical Insitute (ME and JRE), the NSF (IOS 1547796 and 1339125 to DJK) the USDA National Institute of Food and Agriculture, (Hatch project number CA-D-PLS-7033-H to DJK) and by the Danish National Research Foundation (DNRF99) grant.
Contributions
MS and ME conceived of the project. MS prepared samples. MT analyzed RNAseq data. BL and EK performed glucosinolate assays. MS generated transgenic lines, performed genetic experiments, and performed physiological experiments. DJK assisted with glucosinolate analysis. MS and ME wrote the manuscript. MS, ME, LS, JRE, EK, and DJK edited the manuscript. JRE and LS provided materials.
Competing interests
The authors declare no competing interests.
Data and materials availability
Plant lines will be available through the Arabidopsis Stock Center. RNAseq data will be deposited into NCBI as GEO#xxxxx.
Acknowledgements
We thank Jiyoung Park, Po-Kai Hsu and Julian Schroeder for abi1-1and ghr1 seeds and for help with stomatal aperture measurement, Eric Schmelz for assistance with freeze drying of samples, Jeongim Kim and Clint Chapple for cyp83a1 seeds, Venkatesan Sundaresan and Titima Tantikanjana for cyp79f single and double mutant lines, and Jim Whelan for wrky63-1 and 35S::WRKY63 seeds.