CYP707A3, a major ABA 8′-hydroxylase involved in dehydration and rehydration response in Arabidopsis thaliana

Authors

  • Taishi Umezawa,

    1. Gene Discovery Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    2. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Kouya-dai, Tsukuba, Ibaraki 305-0074, Japan,
    Search for more papers by this author
  • Masanori Okamoto,

    1. Growth Regulation Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    2. Department of Biological Sciences, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo 192-0397, Japan,
    Search for more papers by this author
  • Tetsuo Kushiro,

    1. Growth Regulation Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    Search for more papers by this author
    • Present address: Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan.

  • Eiji Nambara,

    1. Growth Regulation Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    Search for more papers by this author
  • Youko Oono,

    1. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Kouya-dai, Tsukuba, Ibaraki 305-0074, Japan,
    2. Plant Functional Genomics Research Team, Functional Genomics Research Group, RIKEN Genomic Sciences Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan, and
    Search for more papers by this author
  • Motoaki Seki,

    1. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Kouya-dai, Tsukuba, Ibaraki 305-0074, Japan,
    2. Plant Functional Genomics Research Team, Functional Genomics Research Group, RIKEN Genomic Sciences Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan, and
    Search for more papers by this author
  • Masatomo Kobayashi,

    1. Experimental Plant Division, RIKEN Bioresource Center, 3-1-1 Kouya-dai, Tsukuba, Ibaraki 305-0074, Japan
    Search for more papers by this author
  • Tomokazu Koshiba,

    1. Department of Biological Sciences, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo 192-0397, Japan,
    Search for more papers by this author
  • Yuji Kamiya,

    1. Growth Regulation Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    Search for more papers by this author
  • Kazuo Shinozaki

    Corresponding author
    1. Gene Discovery Research Group, RIKEN Plant Science Center, 1-7-22 Suehiro-cho, Yokohama, Kanagawa 230-0045, Japan,
    2. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Kouya-dai, Tsukuba, Ibaraki 305-0074, Japan,
    Search for more papers by this author

*(fax +81 29 836 9060; e-mail sinozaki@rtc.riken.jp).

Summary

Abscisic acid (ABA) catabolism is one of the determinants of endogenous ABA levels affecting numerous aspects of plant growth and abiotic stress responses. The major ABA catabolic pathway is triggered by ABA 8′-hydroxylation catalysed by the cytochrome P450 CYP707A family. Among four members of Arabidopsis CYP707As, the expression of CYP707A3 was most highly induced in response to both dehydration and subsequent rehydration. A T-DNA insertional cyp707a3-1 mutant contained higher ABA levels in turgid plants, which showed a reduced transpiration rate and hypersensitivity to exogenous ABA during early seedling growth. On dehydration, the cyp707a3-1 mutant accumulated a higher amount of stress-induced ABA than the wild type, an event that occurred relatively later and was coincident with slow drought induction of CYP707A3. The cyp707a3 mutant plants exhibited both exaggerated ABA-inducible gene expression and enhanced drought tolerance. Conversely, constitutive expression of CYP707A3 relieved growth retardation by ABA, increased transpiration, and a reduction of endogenous ABA in both turgid and dehydrated plants. Taken together, our results indicate that CYP707A3 plays an important role in determining threshold levels of ABA during dehydration and after rehydration.

Introduction

Abscisic acid (ABA) is a phytohormone that regulates various processes of plant development such as seed dormancy, germination and senescence. Abscisic acid also functions in relation to adaptive responses to environmental stresses, including stomatal regulation and ABA-responsive gene expression (Nambara and Marion-Poll, 2005; Shinozaki and Yamaguchi-Shinozaki, 2000). The endogenous ABA content is thought to be the major determinant of these physiological processes. This hypothesis is supported by several lines of evidence that ABA content in plants fluctuates widely during seed development, germination or drought stress (Black, 1992; Kiyosue et al., 1994; Zeevaart, 1980). Furthermore, the lack of ABA-biosynthetic genes causes reduced seed dormancy and wilting (McCarty, 1995). In theory, endogenous ABA content should be maintained by a balance between biosynthetic and catabolic activities.

Recently, a number of enzymes for ABA biosynthesis have been identified from various genetic or biochemical approaches (reviewed by Nambara and Marion-Poll, 2005). One of these, the 9-cis-epoxycarotenoid dioxygenase (NCED), cleaves 11, 12 double bonds of C40 carotenoids and produces the C15 precursor of ABA. This step is thought to be a critical reaction for de novo ABA biosynthesis in plants (Kende and Zeevaart, 1997; Schwartz et al., 1997). There are five members of the NCED family in the Arabidopsis genome. As the NCED family members exhibit various tissue specificities and expression patterns, it is suggested that each plays a distinct role (Tan et al., 2003). Drought stress induced AtNCED3 predominantly among Arabidopsis NCED genes, therefore AtNCED3 is regarded as the most important enzyme for drought-inducible ABA biosynthesis (Iuchi et al., 2001).

Although much is known about ABA biosynthesis in plants, our knowledge about the catabolic pathway of ABA is still quite limited. Abscisic acid is catabolized into inactive forms by either oxidation or conjugation (reviewed by Nambara and Marion-Poll, 2005). The oxidative pathways play a pivotal role in various physiological processes. The major oxidative pathway is triggered by the hydroxylation of C-8′ to form 8′-hydroxy ABA (8′OH-ABA). In the subsequent step, 8′OH-ABA is spontaneously isomerized to phaseic acid (PA). This reaction is then followed by conversion of PA to dihydrophaseic acid (DPA). It has been predicted that ABA 8′-hydroxylase is a cytochrome P450 monooxygenase that is induced by ABA in maize or Arabidopsis cells (Krochko et al., 1998; Windsor and Zeevaart, 1997). Recently, two independent research groups identified CYP707A1 to A4, one of the cytochrome P450 family in Arabidopsis, as the enzyme responsible for catalysing ABA 8′-hydroxylation (Kushiro et al., 2004; Saito et al., 2004). It has been suggested that CYP707As have distinctive roles in plants because each shows its own tissue specificity or expression patterns, for example, CYP707A2 has a significant role for seed dormancy, while CYP707A3 does not affect that (Kushiro et al., 2004).

CYP707As are induced by exogenous ABA treatment, dehydration or rehydration in Arabidopsis (Kushiro et al., 2004; Saito et al., 2004). The induction of CYP707As is likely to be important for the maintenance of endogenous ABA levels, especially when plants have to inactivate ABA promptly after release from dehydration (Zeevaart, 1980). However, at the present time it is still unclear which members of the CYP707A group take charge and initiate such a response. In this study we focused on a single CYP707A member from Arabidopsis, CYP707A3, a gene that is induced most abundantly in Arabidopsis during the dehydration and rehydration processes. We characterized the regulation of CYP707A3 expression under dehydration or rehydration, and dissected the in planta function of CYP707A3 using loss-of-function mutants and gain-of-function transgenic plants. Our data demonstrate that CYP707A3 plays significant roles in seedling growth and responses to dehydration and rehydration in Arabidopsis. In addition, our results from CYP707A3-overexpressing plants have opened up new insights for the manipulation of endogenous ABA content in plants.

Results

Expression analysis of CYP707As under dehydration and rehydration conditions

In Arabidopsis there are four ABA 8′-hydroxylase genes, designated CYP707A1–CYP707A4 (Kushiro et al., 2004; Saito et al., 2004). It has been shown previously that their gene expression is upregulated by ABA, dehydration and rehydration (Kushiro et al., 2004; Saito et al., 2004). The interesting expression pattern of CYP707As provoked us to determine which CYP707A member takes a central role for ABA catabolism during the dehydration or rehydration process. We compared each expression level of CYP707A1-4 by RNA gel-blot analysis, using Arabidopsis seedlings subjected to dehydration and rehydration treatments (Figure 1). The results showed that expression of CYP707A1, CYP707A3 and CYP707A4 could be detected during dehydration or rehydration, and CYP707A2 was expressed only slightly in seedlings. During both dehydration and rehydration processes, CYP707A3 exhibited the highest expression levels among the CYP707A family. Notably, rehydration-responsive CYP707A3 expression occurred rapidly within first 0.5 h, in contrast to the slower induction of CYP707A1 and CYP707A4 which peaked at 1 or 2 h after rehydration. These data suggest that CYP707A3 plays a prominent role in ABA catabolism during the dehydration and rehydration processes.

Figure 1.

RNA gel-blot analysis of CYP707As and stress-marker genes during dehydration and rehydration.
Col(WT), aba2-2, nced3-2, Ler(WT), abi1-1 and abi2-1 were grown on agar plates for 2 weeks, then dehydrated for 1, 2 and 5 h (DH), and subsequently rehydrated for 0.5, 1, 2 and 5 h (RH). Total RNA was extracted from each sample, and 7 μg RNA was loaded into each lane. Ethidium bromide-stained rRNA is shown as loading control. Expression patterns of CYP707A1-4, ProDH, NCED3, RD29A, RD29B and RAB18 were analysed by hybridization with 32P-labelled cDNA probes, and detected by autoradiography.

It is noteworthy that our expression data are a little different from previous reports, in which all CYP707As were induced by dehydration or rehydration (Kushiro et al., 2004; Saito et al., 2004). Such confusion may be attributed to differences in experimental methods: PCR techniques were employed by previous studies, while RNA gel-blot analysis was performed in this study. In a previous report (Kushiro et al., 2004) the expression of CYP707A1-4 was presented as fold-change values, thus we could not compare each expression level specifically. Dot-blot hybridization was used to confirm that each probe of CYP707A1-4 efficiently hybridized to its own sequence (Figure S1). It is therefore concluded that our experiments (Figure 1) accurately reflect the amount of transcript in plants.

ABA signalling and expression of CYP707As

At the present time, the regulatory mechanism for expression of CYP707As is still unclear, especially for the response to dehydration or rehydration. We have indirect evidence to suggest that ABA-dependent regulation may be involved in CYP707A gene expression, because we detected the presence of ABA-responsive elements (ABRE) in their promoter regions (data not shown). In order to characterize this mechanism, we used several ABA-deficient (aba2-2 and nced3-2) or ABA-insensitive (abi1-1 and abi2-1) mutants. If CYP707As are indeed regulated by ABA signalling, even under dehydration or rehydration conditions, their gene expression should be repressed in ABA-deficient or ABA-insensitive mutants.

As a positive control, we analysed ABA-responsive genes (RD29B and RAB18) that are under direct regulation through ABRE (Mantyla et al., 1995; Uno et al., 2000; Yamaguchi-Shinozaki and Shinozaki, 1994), and our RNA blot analyses confirmed that they were significantly repressed in abi1-1, abi2-1, aba2-2 and nced3-2 plants (Figure 1). In the same ABA-related mutants, investigation of the response of CYP707A genes to rehydration showed that expression levels of CYP707A1 and CYP707A4 were slightly inhibited. The strong induction of CYP707A3 subsequent to rehydration normally occurred at 0.5 h, even in ABA-related mutant backgrounds. However, repression of CYP707A3 transcripts at 1 h after rehydration was promoted in the ABA-related mutants. Thus it can be concluded that CYP707A3 expression is largely independent, but partly dependent on ABA signalling under dehydration or rehydration conditions.

RD29A and AtNCED3 are well known dehydration-responsive genes (Iuchi et al., 2001; Yamaguchi-Shinozaki and Shinozaki, 1993, 1994), and ProDH is a typical rehydration-responsive gene (Nakashima et al., 1998; Satoh et al., 2002). Sufficient induction of these positive control genes confirmed our experimental conditions and the validity of mutant backgrounds (Figure 1). Even under non-stress conditions, ProDH showed a relatively strong expression pattern in ABA-related mutants. These data suggest that ABA has some effects on the regulation of ProDH. It is important to note that the expression patterns of CYP707A3 and ProDH were quite different from one another (Figure 1). Whereas ProDH increased gradually in response to rehydration, CYP707A3 exhibited a maximum expression level at 0.5 h subsequent to rehydration. Such an early induction of CYP707A3 suggests that this gene is affected by some positive regulations which are not only independent from ABA, but are also different from ProDH expression.

CYP707A3 participates in post-germination developmental arrest by ABA

The phytohormone ABA efficiently inhibits early seedling growth by preventing developmental phase transition from embryos (Lopez-Molina et al., 2001). This process is known as post-germination growth arrest, a phenomenon regulated primarily by the ABI5 transcription factor (Lopez-Molina et al., 2001, 2003). Concomitant with a decrease in the CYP707A2 transcript level, the expression of CYP707A3 and CYP707A1 is upregulated during and after seed germination (Kushiro et al., 2004). As a result, it is likely that they affect early seedling growth through the degradation of ABA. As a means to examine whether CYP707A3 participates in this process, we checked the ABA-responsiveness of cyp707a3-1 and cyp707a3-2 knockout mutants (Kushiro et al., 2004). These mutants exhibited no visible phenotypes when grown normally under our growth conditions. The results of these analyses showed that the greening rate of cyp707a3-1 and cyp707a3-2 was significantly restricted in the presence of 0.1 or 0.5 μm (+)-S-ABA (Figure 2). These data suggest that CYP707A3 is at least one of the ABA 8′-hydroxylases that function to degrade endogenous ABA at the post-germination stage. It is important to note that the differences between wild type and mutants were not observed when plants were exposed to (−)-R-ABA, an isoform that is not the substrate of CYP707As (Kushiro et al., 2004; Saito et al., 2004).

Figure 2.

Effects of (+)-S-ABA or (−)-R-ABA on early seedling growth of wild-type or cyp707a3 mutants.
Seeds were sown on normal GM medium and GM medium containing 0.1 or 0.5 μm (+)-S-ABA or 1 μm (−)-R-ABA. After stratification at 4°C for 4 days, plates were maintained at 22°C under continuous light. The greening rate was calculated from the number of expanded cotyledons within 50 seeds for each plant type: Col (open circles); cyp707a3-1 (solid triangles); cyp707a3-2 (solid squares). Bars, ±SE (n = 3).

CYP707A3 supports the drought tolerance of plants via maintenance of ABA levels

In general, plants’ responses to drought are largely dependent on ABA. It is therefore reasonable to consider that CYP707As play an important role in drought tolerance via the regulation of ABA degradation. In support of this supposition, it was recently shown that chemical inhibition of CYP707A activity led to higher amounts of ABA and elevated drought tolerance (Kitahata et al., 2005). CYP707A3 was shown to be the most abundant gene in vegetative tissues, such as rosette leaves or roots (Kushiro et al., 2004; Saito et al., 2004), therefore it is likely that CYP707A3 participates in the regulation of ABA content in vegetative tissues in response to abiotic stress. In order to determine whether CYP707A3 affects the drought response of Arabidopsis, we examined the drought tolerance of cyp707a3 knockout plants. Under our experimental conditions, the majority of wild-type plants could not survive after 14 days’ withholding of water. Despite this extended exposure to drought stress, both cyp707a3-1 and cyp707a3-2 mutant backgrounds maintained higher survival rates than the wild-type control plants (Figure 3a,b). The enhanced survival rate of mutant plants was supported by the observed moderate suppression of transpiration of cyp707a3 mutants (Figure 3c). These data suggest that CYP707A3 partly contributes to stomatal regulation in Arabidopsis leaves by balancing endogenous ABA levels.

Figure 3.

Drought-tolerance analysis for cyp707a3 mutants.
(a) Three-week-old plants of wild type (Col), cyp707a3-1, cyp707a3-2 and nced3-1 were subjected to drought conditions by withholding water. Photos were taken 14 days after initiation of the experiment. The experiment was replicated three times; five pots (in total 15 plants) for each plant type were used within each replicate experiment.
(b) Plants of wild type (Col), cyp707a3-1, cyp707a3-2 and nced3-1 that survived the stress were counted after 14 days’ drought treatment and their survival rate was calculated. Bars, ±SE (n = 4).
(c) Transpiration rates of wild-type (Col), cyp707a3-1, cyp707a3-2, nced3-1 and aba2-2 plants were measured with an LI-6400 portable photosynthesis system under non-stress conditions. Bars, ±SE (n = 6).

As a method to confirm that CYP707A3 plays a significant role in the maintenance of ABA levels in plants, we measured endogenous contents of ABA, PA and DPA in wild-type and cyp707a3-1 plants. In both plant backgrounds, endogenous ABA content was elevated during dehydration and decreased rapidly in response to rehydration. However, ABA content in cyp707a3-1 was higher than in wild-type plants, not only during dehydration/rehydration, but even under the initial conditions before the experimental treatments (Figure 4a). Thereafter, the difference was extended under dehydration, and the ABA level was still maintained in cyp707a3-1 during the rehydration process. On the other hand, PA production in cyp707a3-1 was less than in the wild type during both dehydration and rehydration treatments (Figure 4a).

Figure 4.

Effects of CYP707A3 disruption on endogenous abscisic acid (ABA) and gene-expression patterns under dehydration or rehydration conditions.
(a) Endogenous ABA, phaseic acid (PA) and dihydrophaseic acid (DPA) contents in Col (circles) and cyp707a3-1 plants (squares). Plants were grown on agar plates for 2 weeks, then dehydrated for 3 and 6 h (DH, solid symbols) and subsequently rehydrated for 1 and 2 h (RH, open symbols). Plants were harvested to measure endogenous ABA, PA and DPA content at the times indicated. Inset, initial levels of ABA in Col (open bars) and cyp707a3-1 plants (solid bars). Bars, ±SE (n = 3).
(b) RNA gel-blot analysis of Col(WT), cyp707a3-1, cyp707a3-2 and cyp707a2-1 plants. Two-week old plants were subjected to dehydration for 1, 2 and 5 h (DH) and then to rehydration for 0.5, 1, 2 and 5 h (RH). Total RNA was extracted from each sample and 7 μg RNA was loaded equally into each lane. Ethidium bromide-stained rRNA is shown as a loading control. Expression patterns of CYP707A1-4, RD29B, RAB18, NCED3 and ProDH were analysed by hybridization with 32P-labelled cDNA probes and detected by autoradiography.

In association with the increasing ABA content in cyp707a3-1 plants, RNA gel-blot analysis showed a strong induction of known ABA-responsive genes such as RD29B or RAB18 (Figure 4b). Expression levels of CYP707A3, NCED3 or ProDH were recorded as experimental controls. Full-length CYP707A3 transcripts disappeared completely in cyp707a3-1 and -2, but the induction of NCED3 and ProDH could be observed typically in response to dehydration and rehydration, respectively.

Overexpression of CYP707A3 affects various responses to ABA

The ABA-dependent regulation of seed dormancy or stress response highlights the importance of endogenous ABA content in plants for biotechnological applications. Genetic engineering is a promising technology that may enable scientists to manipulate endogenous ABA levels artificially. We were interested to determine whether CYP707As may serve as potential candidate genes that may function as biotechnological tools and enable the control of endogenous ABA levels. To answer this question, we tested the effects of constitutive expression of CYP707A3, which was driven by the CaMV35S promoter. Three independent lines (numbers 1, 12 and 17) of 35S::CYP707A3 plants were obtained and used for the subsequent detailed analyses.

We measured transpiration rate of 35S::CYP707A3 plants to check the effect of CYP707A3 overexpression on stomatal regulation. Under identical experimental conditions, we confirmed that the transpiration rate of 35S::CYP707A3 plants was higher than that of control plants (Figure 5c). An individual transgenic plant line (genetic line 1) showed the most strong pleiotropic phenotype as well as aba2-2 (Figure 5b), an observation that suggests the endogenous ABA content in line 1 should be reduced significantly. Although the expression levels of CYP707A3 were nearly similar among transgenic lines 1, 12 and 17 (Figure 5a), the phenotypic changes in each line were conflicting. It is reasonable to suggest that their phenotypes reflected diverse endogenous ABA levels, and that the enzymatic activities of the introduced CYP707A3 gene could be different between transgenic lines due to translational or post-translational events. It is important to note that we did not check the drought tolerance of 35S::CYP707A3 plants because of differences in plant size or morphological changes, as shown in Figure 5(b). These differences may involve primary or secondary effects on stress tolerance, and prevented us from comparing drought tolerance qualitatively and quantitatively among the transgenic plant backgrounds.

Figure 5.

Pleiotropic phenotype of 35S::CYP707A3 plants.
(a) Expression of CYP707A3 in vector control (VC) and 35S::CYP707A3 plants (lines 1, 12 and 17). Total RNA samples were extracted from three independent seedlings for each line and examined by RT-PCR.
(b) Photographs of VC, 35S::CYP707A3 (lines 1, 12 and 17) and aba2-2 plants taken 3 weeks after germination. In addition to aba2-2, 35S::CYP707A3 plants, particularly line 1 exhibited small plant size and wilting leaves.
(c) Transpiration rates of VC and 35S::CYP707A3 plants were measured with an LI-6400 portable photosynthesis system. Bars, ±SE (n = 6).

The 35S::CYP707A3 lines were sown on agar medium containing 0.1 or 0.5 μm (+)-S-ABA, or 1 μm (−)-R-ABA, and their greening rates were measured subsequently (Figure 6). In contrast to cyp707a3 mutants, the effects of (+)-S-ABA on 35S::CYP707A3 plants were less than on control plants. However, no detectable differences were observed in the presence of (−)-R-ABA, suggesting that the overexpressed CYP707A3 maintained accurate selectivity to exogenous (+)-S-ABA in transgenic plants.

Figure 6.

Effects of CYP707A3 overexpression on early seedling growth in the presence of (+)-S-ABA or (−)-R-ABA.
Seeds were sown on normal GM medium and GM medium containing 0.1 or 0.5 μm (+)-S-ABA or 1 μm (−)-R-ABA. After stratification at 4°C for 4 days, seedlings were grown at 22°C under continuous light. The greening rate was calculated from the number of expanded cotyledons within 50 seeds for each plant type: VC (open circles); and 35S::CYP707A3 of lines 1 (solid triangles), 12 (solid squares) and 17 (solid circles). Bars, ±SE (n = 4).

Overexpression of CYP707A3 enhances ABA catabolism

As described above, 35S::CYP707A3 plants showed insensitivity to exogenous ABA, and exhibited elevated transpiration rates. These data suggest that the resultant increase in ABA catabolic activity, which was caused by the constitutive expression of CYP707A3, lowered endogenous ABA contents in 35S::CYP707A3 plants. In order to confirm this supposition, we tested a particular transgenic plant background (line 1) and measured its endogenous ABA contents and its catabolites, PA or DPA (Figure 7). In comparison with control plants, under non-stressed conditions the endogenous levels of ABA were significantly reduced in 35S::CYP707A3 transgenic plants (Figure 7, inset). After drought stress, although ABA contents were elevated in both controls and 35S::CYP707A3 plants, ABA levels were maintained at lower levels in 35S::CYP707A3 plants. However, PA or DPA contents were increased dramatically in transgenic CYP707A overexpression plants. After rehydration, the degradation of ABA and accumulation of PA/DPA were enhanced in 35S::CYP707A3 plants. These results suggest that the overexpression of CYP707A3 caused the efficient reduction of endogenous ABA content by elevated ABA catabolic activity in transgenic plants.

Figure 7.

Endogenous abscisic acid (ABA), phaseic acid (PA) and dihydrophaseic acid (DPA) content in 35S::CYP707A3 plants.
Two-week-old plants of vector control (circles) and 35S::CYP707A3 line 1 (squares) were grown on agar plates for 2 weeks, then dehydrated for 3 and 6 h (DH, solid symbols) and subsequently rehydrated for 1 and 2 h (RH, open symbols). Plants were subsequently collected to measure endogenous ABA, PA and DPA contents at the time points indicated. Inset, initial levels of ABA in vector control (open bars) and 35S::CYP707A3 line 1 (solid bars). Bars, ±SE (n = 3).

Discussion

Physiological functions of CYP707A3

Regulation of ABA metabolism within plants depends on internal or external signals, such as water deficit or developmental stages. The equilibration of ABA biosynthesis and catabolism is a determinant of endogenous ABA content in plants. The ABA content fluctuates widely during the dehydration and subsequent rehydration processes (Nambara and Marion-Poll, 2005). The accumulation of endogenous ABA that occurs under drought stress is rapidly reversed through ABA catabolism subsequent to removal of the stress (Kiyosue et al., 1994; Zeevaart, 1980). It is important to note that these reactions occur concomitantly with the induction of CYP707As gene expression (Kushiro et al., 2004). As such stress-responsive expression of CYP707As appears to be important for plant responses to drought, further analysis has been required to determine whether CYP707As are functional for ABA catabolism in response to dehydration stress.

Although it was previously known that all four members of CYP707As are induced by dehydration and rehydration (Kushiro et al., 2004), it is still unknown which functions predominantly in such a situation. We conducted RNA gel-blot analysis, using Arabidopsis plants exposed to dehydration and rehydration, and our results suggested that CYP707A3 is important for ABA catabolism under stress conditions, as its expression level was clearly the most abundant of the four CYP707A members (Figure 1). In addition, our expression data support the functional analysis, which demonstrated that disruption of CYP707A3 caused an elevated accumulation of endogenous ABA, especially under dehydration (Figure 4a). The significant increase of endogenous ABA in cyp707a3-1 indicates that CYP707A3 is one of the determinants for the threshold level of endogenous ABA during dehydration stress. This proposition was also supported by phenotypic changes in cyp707a3-1 and cyp707a3-2 plants. We found that their survival rate under drought conditions was relatively elevated compared with wild-type plants (Figure 3a). Furthermore, their survival was associated with a lower transpiration rate (Figure 3b) and a higher induction of ABA-responsive genes (Figure 4b). These results suggest that the maintenance of endogenous ABA levels by CYP707A3 is important for stomatal regulation and gene expression.

Other CYP707As may also potentially participate in ABA degradation under dehydration and rehydration conditions. This conclusion is reasonable as ABA was reduced rapidly, even in cyp707a3-1 after rehydration (Figure 4a), and other CYP707A transcripts were also apparently induced by dehydration and rehydration (this study; Kushiro et al., 2004; Saito et al., 2004). It is likely that the physiological functions of each of the CYP707As are redundant. However, the observation that they have distinct expression patterns in tissues, organs and stages supports the notion that they probably have distinctive functional roles. For example, CYP707A3 was relatively abundant in vegetative tissues, and the disruption of this gene had little effect on the seed dormancy of Arabidopsis (Kushiro et al., 2004). In this study, we found that cyp707a3 plants showed a significant delay in growth after germination on (+)-S-ABA-containing medium, but that the trait was relatively limited (Figure 2). Although CYP707A3 functions mainly in post-germination growth arrest, other CYP707As may also function simultaneously for ABA catabolism.

Our experimental data suggest that CYP707A3 is a major enzyme for ABA catabolism in Arabidopsis during dehydration and rehydration, whereas AtNCED3 has a significant role in dehydration-responsive ABA biosynthesis (Iuchi et al., 2001). We propose that the pair of AtNCED3 and CYP707A3 has a pivotal role for ABA biosynthesis and catabolism during dehydration or rehydration. The balance between ABA biosynthesis and catabolism can be explained as follows. During dehydration stress, transcription of AtNCED3 is induced and ABA biosynthesis is fully activated (Iuchi et al., 2001). CYP707A3 (and other CYP707As) is also induced subsequently (Kushiro et al., 2004; Saito et al., 2004), and can provide significant ABA catabolic activity, even under dehydration. It is likely that dehydration-responsive induction of CYP707A3 helps not only to maintain endogenous ABA levels within the permissible range, but also to prepare the plant for degradation of ABA after removal of the stress. After rehydration, ABA-biosynthetic activity decreased quickly due to the disappearance of AtNCED3 transcripts. Instead, expression of CYP707A3 is induced and its accumulation allows for rapid degradation of ABA. We propose that ABA-biosynthetic and catabolic activities co-operatively determine endogenous ABA levels during dehydration and rehydration processes in Arabidopsis.

Transcriptional regulation of CYP707A3

Previous reports indicated that CYP707As were induced by dehydration, rehydration and applications of exogenous ABA (Kushiro et al., 2004; Saito et al., 2004). However, at present the regulatory mechanism of expression of CYP707As is not well characterized. Some ABRE or ABRE-like sequences can be found in the promoter region of CYP707As (data not shown), suggesting that the expression of CYP707As involves ABA-dependent regulation. As a result, we believed that utilization of several ABA-deficient or ABA-insensitive mutants would prove valuable tools for the study of CYP707As under dehydration or rehydration conditions. However, the results proved to be complicated (Figure 1) where CYP707A1 and CYP707A4 were induced by dehydration and rehydration, respectively, and they were repressed in ABA-related mutants. Interpretation of these data suggests that their expression depends considerably on ABA signalling. A previous study also showed that CYP707A1 was downregulated in abi1-1 (Hoth et al., 2002). On the other hand, CYP707A3 was induced by both dehydration and rehydration, but its expression pattern was not clearly altered in ABA-related mutants, especially for early induction after rehydration. The effect of ABA could barely be observed for CYP707A3 repression in the latter period of rehydration. It is important to note that this result does not indicate that expression of CYP707A3 occurs independently from ABA. This conclusion can be substantiated as CYP707A3 is induced simply by exogenous ABA and is restricted in abi1-1 and abi2-1 mutants (data not shown). Therefore it can be concluded that CYP707As are regulated partly by an ABA-dependent pathway, but other regulatory systems remain to be identified, especially on rehydration.

Recently, results from a comprehensive microarray analysis identified numerous rehydration-responsive genes (Oono et al., 2003). Rehydration-responsive genes were classified into two groups according to regulation of the ACTCAT cis-acting element. This particular element is associated with the ATB2 subfamily of bZIP transcription factors, and is accountable for the rehydration-responsive induction of ProDH (Satoh et al., 2002, 2004). The ACTCAT sequence could be detected only in some of the rehydration-responsive genes (Oono et al., 2003), therefore multiple regulatory systems must be involved in the perception of rehydration signals. We were unable to detect any ACTCAT sequences in the promoters of CYP707As. In addition, whereas ProDH was induced gradually by rehydration, the induction of CYP707A3 occurred rapidly and transiently. These data suggest that some novel regulatory pathways, other than ACTCAT, may be involved in the regulation of CYP707A3 transcription. For future studies it will be interesting to determine the cis- or trans-factors responsible for the rehydration-responsive expression of CYP707A3.

Effects of CYP707A3 expression as an ABA 8′-hydroxylase in plants

According to reliable biochemical characterizations, CYP707As are undoubtedly acknowledged as ABA 8′-hydroxylases, and employ strict substrate specificity to produce 8′-hydroxy ABA (Kushiro et al., 2004; Saito et al., 2004). Although ABA forms two types of optical isomer, (+)-S-ABA and (−)-R-ABA, a natural form is (+)-S-ABA, which comprised a major portion of ABA pools in plants. CYP707A proteins specifically recognize (+)-S-ABA, but not (−)-R-ABA (Kushiro et al., 2004; Saito et al., 2004). In this study we could not observe any differences between wild-type and cyp707a3 plants when (−)-R-ABA was supplied, despite the fact that (−)-R-ABA inhibited plant growth after germination (Figure 2). As well as cyp707a3 plants, 35S::CYP707A3 plants also showed specific insensitivity to (+)-S-ABA (Figure 6). These results clearly indicate that the phenotypic changes of cyp707a3 mutants or 35S::CYP707A3 plants depend solely on the enzymatic activity of CYP707A3.

Constitutive overexpression of CYP707A3 leads to smaller plant size and wilting leaves in transgenic plants (Figure 5b). Interestingly, these morphological changes were similar to those in ABA-deficient mutants such as aba2-2. In addition, the transpiration rates were significantly upregulated in 35S::CYP707A3 plants. Collectively, these data suggest that endogenous ABA content is reduced by a strong activity of ABA 8′-hydroxylation, which results from accumulation of CYP707A3 protein. In 35S::CYP707A3 plants, endogenous ABA was reduced significantly, and after drought stress ABA catabolites were produced dramatically (Figure 7). Our data confirm that constitutive expression of CYP707A3 effectively reduces endogenous ABA levels in plants. In future these data may prove useful for understanding technical aspects of ABA metabolic engineering. Both biosynthetic and catabolic enzymes of ABA are already available, therefore it is feasible that, in the future, scientists will be able to manipulate the fine regulation of endogenous ABA content in plants. These days we have many breeding targets by controlling ABA content in plants, for example, the manipulation of seed dormancy or drought tolerance. To date, several researchers have attempted to control ABA content in plants via the manipulation of ABA-biosynthetic genes, for example NCED (Iuchi et al., 2001; Qin and Zeevaart, 2002; Thompson et al., 2000). However, the constitutive expression of NCED resulted in only a moderate increase in endogenous ABA content. It is likely that co-operative handling of ABA biosynthesis and catabolism will prove a more effective strategy for genetic engineering of endogenous ABA levels in plants.

Conclusion and perspectives

In the present study we characterized the biological functions of an Arabidopsis ABA 8′-hydroxylase enzyme, CYP707A3. First, we were able to suggest that CYP707A3 plays a prominent role in ABA catabolism in the vegetative tissues of Arabidopsis, especially under dehydration or rehydration conditions. Now we know the physiological functions of two CYP707As: CYP707A2 for seed dormancy (Kushiro et al., 2004); and CYP707A3 for vegetative tissues (this study). Further analyses will be required to understand the in planta functions of other CYP707As, and these studies will help to discriminate functional overlaps and differences among the CYP707A gene family. Another significant finding of this study is that the constitutive expression of CYP707A3 is an effective method to reduce endogenous ABA content in Arabidopsis. These data have opened up the possibility that CYP707A3 may serve as a tool for controlling ABA content in plants. This possibility should be tested for biotechnological applications in stress tolerance and regulation of seed dormancy in the future.

Experimental procedures

Transgenic plants and T-DNA insertional mutants

The full-length cDNA of CYP707A3 was obtained from an RAFL09-13-J06 clone available in the RIKEN Arabidopsis full-length cDNA collection (Seki et al., 2002b, 2004). The open reading frame of CYP707A3 was inserted downstream of the CaMV35S promoter within the pBE2113 vector (Mitsuhara et al., 1996), resulting in the 35S::CYP707A3 plasmid. Arabidopsis plants (Col) were transformed using the Agrobacterium tumefaciens C58 strain, which harboured the 35S::CYP707A3 plasmid using methods described previously (Valvekens et al., 1988). The 35S::CYP707A3 plants were obtained after selection with kanamycin, and lines 1, 12 and 17 were utilized for subsequent analyses. Mutant cyp707a3-1 and cyp707a3-2 lines were previously isolated by Kushiro et al. (2004) from the SALK T-DNA-tagged lines (Alonso et al., 2003).

Analysis of cyp707a3 mutants or CYP707A3-overexpressing plants

Seeds of Columbia ecotype (Col), cyp707a3-1 and cyp707a3-2, or 35S::CYP707A3 transgenic lines 1, 12, 17 and vector control (VC) were sown on Gamborg's B5 medium agar plates with/without 0.1 or 0.5 μm (+)-S-ABA (Toray Co Ltd, Japan) or 1 μm (−)-R-ABA, as described previously (Kushiro et al., 2004). After 4 days’ stratification at 4°C, plates were placed at 22°C under continuous light, and their subsequent greening rate was calculated. For analysis of drought tolerance, 2-week-old seedlings were transferred from plate conditions and grown in soil. Plants were exposed to drought stress by withholding water as described previously (Umezawa et al., 2004), and their survival rate was counted after 14 days. Transpiration rates were measured for the 4-week-old plants of Col and cyp707a3-1 or -2, or 35S::CYP707A3 and control plants, with a LI-6400 portable photosynthesis system (Li-Cor, Lincoln, NE, USA) according to methods described previously (Umezawa et al., 2001).

Expression analysis of CYP707As and stress-marker genes

Columbia ecotype (Col), aba2-2, nced3-2, Landsberg erecta ecotype (Ler), abi1-1 and abi2-1 were grown under continuous light at 22°C for 2 weeks. Plants were exposed to dehydration and rehydration conditions as described previously (Oono et al., 2003; Seki et al., 2002a), and subsequently harvested at each time point. Total RNA was extracted according to a method described previously (Nagy et al., 1988). RNA gel-blot analysis was performed according to standard protocols with slight modifications (Sambrook et al., 1989; Umezawa et al., 2002). Total RNA (7 μg) was resolved in a 1% agarose gel containing formaldehyde and subsequently blotted onto a Hybond-N membrane (Amersham Biosciences, Amersham, UK). Nine full-length cDNA probes of CYP707A1 (GenBank AB122149), CYP707A2 (GenBank AC005315), CYP707A3 (GenBank AB122150), CYP707A4 (GenBank AP000419), ProDH (GenBank NM_113981), AtNCED3 (GenBank NM_112304), RD29A (GenBank NM_124610), RD29B (GenBank NM_124609) and RAB18 (GenBank NM_126038) were prepared by reverse-transcription PCR and their nucleotide sequences verified after cloning to pGEM-T vector (Promega, Madison, WI, USA). The cDNA fragments were labelled with [α-32P]dCTP (Amersham) using a BcaBEST DNA labelling kit (Takara Shuzo, Kyoto, Japan). The blot was hybridized with a 32P-labelled probe in 5× sodium chloride, sodium phosphate, EDTA buffer (SSPE) (750 mm NaCl, 43.25 mm NaH2PO4, 6.25 mm EDTA pH 7.4), 1% sodium dodecyl sulphate (SDS), 5× Denhardt's solution and 20 μg ml−1 salmon sperm DNA at 60°C overnight. The membrane was washed in 2× sodium chloride, sodium citrate (SSC) at room temperature for 5 min, 2× SSC at 60°C for 15 min, 1× SSC at 60°C for 15 min and 0.1× SSC at 60°C for 15 min. Radioactivity was detected by autoradiography using RX-U X-ray film (Fujifilm, Tokyo, Japan).

RNA gel-blot analysis was also performed using Col, cyp707a2-1, cyp707a3-1 and cyp707a3-2 plants. Eight cDNA probes of CYP707A1, CYP707A2, CYP707A3, CYP707A4, ProDH, AtNCED3, RD29B and RAB18 were used for hybridization and the detection was conducted as described above.

To check the expression of CYP707A3 in 35S::CYP707A3 overexpression plants, we performed reverse transcription (RT)-PCR analyses. First-strand cDNA was synthesized from total RNA of VC and 35S::CYP707A3 plants (lines 1, 12 and 17) with SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA). The cDNA generated was then subjected to PCR using the following oligonucleotides: forward primer 5′-ATGGATTTCTCCGGTTTGTTTC-3′ and reverse primer 5′-CTATGGTTTTCGTTCCAAGG-3′.

Measurement of ABA, PA and DPA levels

Arabidopsis plants were grown on GM medium for 2 weeks, and then plants were exposed to dehydration. Stress conditions were basically adjusted between each plant line as follows. After 3 h dehydration the relative water content of plants was approximately 50%, and it reduced to <40% after 6 h dehydration. Samples were harvested and extracted for ABA, PA and DPA as described previously (Kushiro et al., 2004). Purified ABA was quantified by GC-selected ion monitoring (SIM) using methods described previously (Cheng et al., 2002). The PA and DPA fractions were trimethylsilylated with N-methyl-N-trimethylsilyltrifluoroacetamide at 60°C for 10 min. For GC-SIM analysis, the 265 m/z (deuterated) and 262 m/z (endogenous) peaks were used for PA quantification, while 429 m/z (deuterated) and 426 m/z (endogenous) peaks were used for DPA quantification. Back-up ions (deuterated PA 355, 328 and 306; endogenous PA 352, 325 and 303; deuterated DPA 339 and 250; endogenous DPA 336 and 247) were also monitored for peak confirmation.

Acknowledgements

We wish to thank Dr T. Asami (RIKEN; Saitama, Japan) for kindly providing the [1,2-13C2]-(±)-ABA. We also wish to thank Dr N. Hirai (Kyoto University, Japan) for providing the deuterium-labelled PA and DPA; and the Arabidopsis Biological Resources Center at Ohio State University (Columbus, OH, USA) and the Salk Institute (Carlsbad, CA, USA) for providing the T-DNA-tagged lines used in this research. This work was supported by a grant from the Bio-Oriented Technology Research Advancement Institution of Japan to K.S., and Grant-in-Aids for Scientific Research for Young Scientists (B) 14760075 and 16770043 from the Ministry of Education, Science, Sport, and Culture of Japan to T.U. M.O. was supported by a research fellowship from the Japan Society for the Promotion of Science for Young Scientists.

Ancillary