A Screen for Genes That Function in Abscisic Acid Signaling in Arabidopsis thaliana | Genetics

A Screen for Genes That Function in Abscisic Acid Signaling in Arabidopsis thaliana

  1. Eiji Nambara*,†,
  2. Masaharu Suzuki‡,
  3. Suzanne Abrams§,
  4. Donald R. McCarty‡,
  5. Yuji Kamiya† and
  6. Peter McCourt*,1
  1. *Department of Botany, University of Toronto, Toronto, Ontario M5S 3B2, Canada
  2. †Plant Science Center, The Institute of Physical and Chemical Research (RIKEN), Wako, Japan 351-0198
  3. ‡Horticultural Sciences Department, University of Florida, Gainesville, Florida 32605
  4. §Plant Biotechnology Institute, National Research Council of Canada, Saskatoon, Saskatchewan S7N 0W9, Canada
  1. 1 Corresponding author: Department of Botany, University of Toronto, 25 Willcocks St., Toronto, ON M5S 3B2, Canada. E-mail: mccourt{at}botany.utoronto.ca

Abstract

The plant hormone abscisic acid (ABA) controls many aspects of plant growth and development under a diverse range of environmental conditions. To identify genes functioning in ABA signaling, we have carried out a screen for mutants that takes advantage of the ability of wild-type Arabidopsis seeds to respond to (−)-(R)-ABA, an enantiomer of the natural (+)-(S)-ABA. The premise of the screen was to identify mutations that preferentially alter their germination response in the presence of one stereoisomer vs. the other. Twenty-six mutants were identified and genetic analysis on 23 lines defines two new loci, designated CHOTTO1 and CHOTTO2, and a collection of new mutant alleles of the ABA-insensitive genes, ABI3, ABI4, and ABI5. The abi5 alleles are less sensitive to (+)-ABA than to (−)-ABA. In contrast, the abi3 alleles exhibit a variety of differences in response to the ABA isomers. Genetic and molecular analysis of these alleles suggests that the ABI3 transcription factor may perceive multiple ABA signals.

THE plant hormone abscisic acid (ABA) controls numerous physiological processes, ranging from inhibition of germination and the establishment of seed dormancy to adaptive responses to a variety of abiotic stresses (Zeevaart and Creelman 1988, see review; Leung and Giraudat 1998). Recent genetic studies in Arabidopsis have demonstrated that plant responsiveness to ABA is controlled by a number of molecular processes including transcription (Giraudatet al. 1992; Finkelsteinet al. 1998; Finkelstein and Lynch 2000b; Lopez-Molina and Chua 2000), RNA processing (Hugouvieuxet al. 2001), post-translational modifications (Leung et al. 1994, 1997; Meyeret al. 1994; Cutleret al. 1996), and metabolism of a second messenger (Xionget al. 2001). Elucidation of the relationships by which these and other factors act to transmit the ABA signal is essential for understanding the disparate roles of ABA on plant growth and development.

We have chosen the Arabidopsis seed as a model system for studying the role of ABA-mediated signal transduction in the control of seed dormancy and germination for a number of reasons. An allelic series of mutations that decrease ABA biosynthesis demonstrate that the level of seed dormancy in Arabidopsis is dependent on embryonic ABA concentrations (Karssenet al. 1983). Thus, seed germination is an excellent biological assay for ABA responsiveness. Moreover, because exogenously applied ABA can inhibit wild-type germination, mutations that decrease seed responsiveness to ABA can be easily identified and, to date, five loci (abi; abscisic acid-insensitive) have been characterized. Dominant mutations in two genes that encode homologous type 2C protein phosphatases, designated ABI1 and ABI2, reduce ABA sensitivity in both embryonic and adult plants (Leung et al. 1994, 1997; Meyeret al. 1994). By contrast, recessive mutations in the ABI3, ABI4, and ABI5 genes mostly affect seed responses and identify three different classes of transcription factors (Giraudatet al. 1992; Finkelsteinet al. 1998; Finkelstein and Lynch 2000b; Lopez-Molina and Chua 2000). Although the relationship between these proteins is unclear, recently it has been shown that ABI3 and ABI5 interact in a yeast two-hybrid protein assay (Nakamuraet al. 2001). Furthermore, Soderman et al. (2000) reported that ectopic expression of the ABI3 or ABI4 gene increases in the accumulation of ABI5 mRNA, and these genes could act cooperatively in vivo.

Critical to our understanding of how ABA activates seed dormancy and inhibits germination is the identification of all the genes that are involved in the transduction of the hormone signal. To identify new factors involved in ABA signaling and expand the collection of mutant alleles that alter seed ABA sensitivity we have taken advantage of the ability to separate ABA enantiomers from a chemically synthesized mixture of the naturally occurring (+)-(S)-ABA [(+)-ABA] and its mirror image (−)-(R)-ABA [(−)-ABA]. The two molecules are very similar in shape, differing only in the disposition of the methyl groups on the ring (Figure 1). The vinyl methyl and the gem dimethyl groups are reversed in the mirror image forms. Comparing the structures of the enantiomers in the conformation adopted by ABA in the crystal structure, the 7′- and 9′-methyl groups are almost identical, with the major difference being the location of the axial 8′ methyl group.

The premise of the screen is that it might be possible to identify mutants that differentiate between (+)-ABA and (−)-ABA. The choice of these compounds was based on the observation that natural genetic variation may have caused subtle differentiation between these two stereoisomers in a number of plant species. In wheat embryos, for example, (−)-ABA effectively induces gene expression of dhn and lea genes, but the effects of this enantiomer on another ABA inducible gene Em is relatively minor (Walker-Simmonset al. 1992). In Arabidopsis protoplasts, (−)-ABA fails to induce the ABA responsive gene Rab18, but can increase the conversion of natural ABA to its breakdown product phaseic acid (Windsor and Zeevaart 1997; Jeannetteet al. 1999). The screen was modeled after the successful screens for mutations that confer a seed ABA-insensitive phenotype to ABA stereoisomer mixtures (Koornneefet al. 1984; Finkelstein 1994). Genetic manipulation of the germination response to one ABA isomer vs. the other should lead to a finer scale dissection of the ABA response, which will be useful in identifying molecules that, although essential to ABA signaling, may have only a minor effect on it. For example, if activation of a redundant ABA response has a preference for one stereoisomer over another, then mutations that disrupt its function may change the ratio of response to the two isomers. Here we report isolation and characterization of (−)-ABA-insensitive mutants based on inhibition of germination. Although all these mutants also exhibit insensitivity to (+)-ABA, in contrast to wild type the degree of insensitivity to these stereoisomers is different among mutants. The mode of ABA action in hypothetical ABA signaling pathways is discussed.

MATERIALS AND METHODS

Plant materials and growth conditions: Arabidopsis thaliana M2 ecotype Columbia seeds mutagenized by ethyl methane-sulfonate (EMS), fast neutron irradiation, and gamma ray irradiation were purchased from Lehle Seeds (Round Rock, TX). Strain names containing E, F, and G designated mutant lines isolated from EMS-, fast neutron-, and gamma ray-mutagenized M2 populations, respectively. The M2 pools designated as E31 to E48 or F4 contain gl1 mutation as a genetic marker. Strain T45-3 is a mutant strain isolated from T-DNA insertion lines; however, this strain does not contain T-DNA (data not shown). The abi4-1 and abi5-1 mutants used for the allelism tests were obtained from Dr. Ruth Finkelstein (Finkelstein 1994). Sterilizing seed and growing plants under sterile conditions was done as previously described (McCourt and Keith 1998).

Mutant screen and germination test: (+)-ABA and (−)-ABA were purified by chiral HPLC (Dunstanet al. 1992). Seeds were chilled for 4 days and transferred to continuous light conditions at room temperature (24°). For germination tests, the percentage of germination was scored each day with expansion and greening of the cotyledons used as a criterion for germination. At least three independent experiments were performed and the representative results were shown. For sugar addition experiments 2% glucose was added to a standard 0.5× MS plate with or without 1 μm (+)-ABA. Seeds were chilled for 4 days and transferred to room temperature with continuous light conditions for a week and scored as described above.

Figure 1.
Figure 1.

Structure of (+)-(S)-ABA and (−)-(R)-ABA. (−)-ABA is the unnatural stereoisomer of natural (+)-ABA.

Mapping of the cho mutations: Mutant lines were crossed to Landsberg erecta. SSLP and CAPS markers were used to map the mutations (Konieczny and Ausubel 1993; Bell and Ecker 1994). Information on the markers was obtained from the TAIR web site (http://www.arabidopsis.org/aboutcaps.html).

DNA sequencing: The abi3 gene was amplified by PCR and cloned into pT7Blue T-Vector (Takara, Kyoto, Japan). Double-stranded DNA was sequenced on both strands by DNA sequencer ABI377 (Applied Biosystems, Foster City, CA). Two independent clones were sequenced on both strands to identify the abi3 mutations.

RESULTS

Isolation of mutants that are insensitive to (−)-ABA: Genes encoding factors that respond to the unnatural stereoisomer (−)-ABA are also expected to function in response to the naturally occurring (+)-ABA (Figure 1). Therefore, mutations in these genes should be expected to also have some altered responsiveness to (+)-ABA. On the basis of this premise we first compared the germination response of Columbia wild-type seed to the different stereoisomers. Wild-type seed shows 100% inhibition at concentrations ≤2.4 μm (+)-ABA while concentrations >5 μm are required to give similar results when (−)-ABA is used (Figure 2). Similar germination curves in response to (+)- or (−)-ABA were observed with Landsberg erecta wild-type seeds (data not shown). These results indicate that (−)-ABA is sensed, although not as efficiently, by Arabidopsis seeds at the level of germination. To test if the (−)-ABA response is through a known ABA response pathway, ABI1-1, ABI2-1, and era1-2 seeds were tested for their sensitivity to the ABA isomers. The ABI1-1 and ABI2-1 seeds were able to germinate on 10 μm (−)-ABA while the era1-2 mutant seed was able to germinate on 0.3 μm (−)-ABA (data not shown), demonstrating that (−)-ABA can signal through similar ABA response pathways as does (+)-ABA.

Figure 2.
Figure 2.

Inhibition of wild-type germination by (+)-ABA or (−)-ABA. Open squares, 0 μm; open diamonds, 0.3 μm; open circles, 0.6 μm; open triangles, 1.2 μm; solid squares, 2.4 μm; solid diamonds, 3 μm; solid circles, 5 μm. (A) Inhibition of wild-type germination by (+)-ABA. Wild-type seeds were sown on various concentrations of (+)-ABA, chilled for 4 days, and cultured under continuous light conditions at room temperature. Percentage of germination is plotted against days after imbibition (DAI). (B) Inhibition of wild-type germination by (−)-ABA. Wild-type seeds were chilled for 4 days and cultured under continuous light conditions at room temperature. Percentage of germination is plotted onto DAI.

On the basis of these observations we screened ~360,000 M2 seeds derived from 75,000 M1 EMS, 29,000 M1 fast neutron-, and 13,500 M1 gamma ray-irradiation mutagenized seeds for mutants that were able to germinate in the presence of 10 μm (−)-ABA. Seeds that germinated on this concentration of (−)-ABA were propagated to the M3 generation. After retesting, 26 M3 lines that represented at least 23 independent mutations were advanced for further analysis. Mutations were deemed independent if the ABA-insensitive seed was isolated from separate M2 seed pools. Using the criterion of 50% inhibition of germination, 20 of 26 mutant lines were identified as insensitive to either 3 μm (+)-ABA or 10 μm (−)-ABA (Table 1). (−)-ABA-insensitive mutants that are able to germinate in the presence of 10 μm (−)-ABA were categorized into two classes on the basis of the ability of germination on (+)-ABA. One class, designated as class I ABA-insensitive mutants, shows (+)-ABA insensitivity on 3 μm (+)-ABA, and the second class, designated as class II insensitive mutants, fails to germinate on 3 μm (+)-ABA. Aside from (−)-ABA-insensitive mutants, the third class (class III) does not germinate on 10 μm (−)-ABA but germinates faster than wild type on 1 μm (+)-ABA, a concentration that under our assay conditions delays wild-type germination.

The class I ABA-insensitive mutants were expected to be allelic to the known abi loci because this class permits germination on both (+)-ABA and (−)-ABA. Subsequent genetic analysis demonstrated that these 20 lines were recessive and fall into three complementation groups (Table 1): 4 abi3 (4 independent), 12 abi4 (10 independent), and 4 abi5 (3 independent). By contrast, 6 lines of class II ABA-insensitive mutants that fail to germinate on 3 μm (+)-ABA were able to germinate faster than wild type on 1 μm (+)-ABA (data not shown). Therefore, these lines also have reduced sensitivity to both (+)-ABA and (−)-ABA, but the degree of insensitivity to the two stereoisomers is different from that in wild type. The 6 class II lines are recessive and define two abi3 alleles and two new loci, designated as chotto1 (cho1) and chotto2 (cho2; Table 1). Three independent cho1 alleles were identified and this locus was mapped onto the bottom of chromosome 5 with tight linkage to the CAPS marker ASB2 (seven recombinants in 640 chromatids). Although the abi2 locus is also located on this region, the cho1 locus is genetically separable from the abi2 locus and subsequent sequencing of the ABI2 gene in the cho1 mutant showed no mutation in the ABI2 gene (data not shown). The single cho2 mutation was mapped onto the bottom of chromosome 4 and is linked to the CAPS marker AG (six recombinants in 24 chromatids). No known abi loci were mapped close to this region, suggesting cho2 also defines a new ABA response gene.

ABI transcription factors participate in differential responses to ABA stereoisomers: As mentioned, all 20 strong ABA-insensitive mutants turned out to have mutations in the previously characterized three ABA response genes (Table 1). These genes encode different transcription factors with ABI3, ABI4, and ABI5 belonging to the B3, AP2, and bZIP families, respectively, based on the conserved DNA-binding domain (Giraudatet al. 1992; Finkelsteinet al. 1998; Finkelstein and Lynch 2000b). Although mutations in these genes define a role for these transcription factors on ABA responsiveness in seeds, a detailed study using multiple alleles has not been reported to measure how various mutations within each gene influence ABA response. For example, although null alleles of abi3 have a multitude of late embryogenic defects (Nambara et al. 1994, 1995), the abi3-1 allele, which contains a missense mutation in the B3 domain, causes only a mild insensitivity to ABA and otherwise has relatively normal seed development (Bies-Etheveet al. 1999). All abi4 and abi5 alleles identified to date are ABA insensitive in seeds and have no other obvious phenotypes. To discern if all these new alleles are equivalent in terms of their response to ABA isomers we tested the germination rates on 3 μm (+)-ABA or 10 μm (−)-ABA. Although these were the concentrations of hormone that were used in the initial screen, because we are monitoring germination rates rather than a static time point, we could get a kinetic measurement of insensitivity. All 10 independent abi4 mutants germinated equally fast or slightly faster on 10 μm (−)-ABA than on 3 μm (+)-ABA; the 3 independent abi5 mutants germinated faster on 3 μm (+)-ABA than on 10 μm (−)-ABA (Figure 3; data not shown). It therefore appears that a defect in the ABI5 gene causes a stronger insensitivity to (+)-ABA vs. (−)-ABA compared to abi4 mutant alleles.

View this table:
TABLE 1

List of the (−)-ABA-insensitive mutants

In contrast to abi4 and abi5 mutants, which fall into discrete classes with respect to their differential response to ABA stereoisomers, the pattern of germination response to (+)- and (−)-ABA is not consistent between various abi3 alleles (Figure 3; data not shown). For example, although some abi3 alleles such as abi3-9 showed no significant difference in insensitivity to either ABA isomer, one abi3 allele, abi3-8, was able to germinate much faster on 3 μm (+)-ABA vs. 10 μm (−)-ABA whereas abi3-12 showed the opposite effect (Figure 3).

The ABI3 gene has been identified and found to be highly homologous to the maize seed-specific transcription factor VP1. Moreover, many of the phenotypes seen in abi3 null mutants are reflected in loss-of-function viviparous1 (vp1) alleles (McCartyet al. 1989; Nambaraet al. 1995). Molecular comparisons between VP1 and ABI3 have defined a number of conserved domains. Three basic regions, designated B1, B2, and B3, appear to be involved in protein-protein and DNA-protein interactions (Suzukiet al. 1997; Nakamuraet al. 2001). The B3 domain in concert with B2 binds DNA in vitro (Suzukiet al. 1997); however, the B3 domain has been shown not to be necessary for VP1-dependent ABA response (Suzukiet al. 1997). Sequence analysis of our abi3 alleles showed that abi3-13 contains a missense mutation in an invariant aspartic acid residue in the B3 domain. In contrast abi3-11 and abi3-12 contain mutations that result in a premature stop codon prior to the B3 domain (Figure 4). The abi3-9 and abi3-10 alleles contain missense mutations within the B2 domain (Figure 4). The abi3-8 mutation causes an amino acid substitution with conversion of leucine 298 to a phenylalanine within the B1 domain (Figure 4). All amino acid residues with missense mutations identified in this study are highly conserved in a variety of ABI3 orthologs. Although two abi3 alleles were isolated in the weak ABA-insensitive class (Table 1), no weak abi4 or abi5 alleles were identified.

Differential responses of the abi3 alleles to ABA in the presence of glucose: The varied responses of abi3 alleles to (+)- and (−)-ABA suggested that subtle phenotypes of various alleles can be uncovered and perhaps these mutations can define the roles of different protein motifs of ABI3. To further pursue this idea we tested the germination and subsequent seedling growth of these abi3 alleles on (+)-ABA in the presence and absence of glucose. Externally applied sugar can have a myriad of effects on Arabidopsis germination and growth (see Gibson 2000 for review). At low concentrations, sugar stimulates wild-type germination as measured by radicle emergence and inhibits the effects of exogenous ABA on inhibition of germination (Finkelstein and Lynch 2000a). Conversely, at higher concentrations, sugar represses cotyledon development and early seedling growth and these responses have been used to identify sugar-insensitive mutants in Arabidopsis (Pegoet al. 1999; Arenas-Huerteroet al. 2000; Huijseret al. 2000; Labyet al. 2000). Many of these sugar-insensitive mutants turn out to be new alleles of ABA auxotrophs or loss-of-function alleles of ABI4. Interestingly, ABI1-1, ABI2-1, and abi3-1 mutants do not confer a sugar-insensitive phenotype, suggesting that only certain ABA response genes are involved in sugar sensing (Arenas-Huerteroet al. 2000; Huijseret al. 2000; Labyet al. 2000). Because our collection of the abi3 alleles appear to discriminate some of ABA signaling pathways, these might give us further insight to understand the interaction between ABA and sugar signaling. We therefore tested germination response of our abi3 alleles on low concentrations of (+)-ABA in the presence of 2% glucose, a concentration that does not normally inhibit Arabidopsis germination or seedling growth. Radicle emergence of the wild-type seeds on 1 μm (+)-ABA was retarded, but this effect is rescued by coapplication of glucose (Finkelstein and Lynch 2000a; data not shown). However, the combination of glucose and ABA caused an inhibition of early wild-type seedling growth that resulted in underdeveloped plantlets that were chlorotic and anthocyanic (Figure 4). Although all abi3 alleles were able to germinate and grow on ABA in the absence of glucose, the presence of glucose had varied effects on different alleles. Both abi3-9 and abi3-12 seedlings were similar to those of the wild type, whereas the remaining abi3 alleles tested were insensitive, showing green cotyledons and root growth on glucose plus ABA (Figure 4). Unlike the abi3-8 mutation that maps to the B1 domain, the two other glucose-insensitive abi3 alleles, abi3-10 and abi3-11, are the result of a base substitution in the B2 domain and a premature stop codon just outside the B2 domain, respectively (Figure 4). In contrast to the abi3 mutants, abi4 mutants identified in this study exhibited early seedling growth in the presence of 1 μm (+)-ABA plus 2% glucose (data not shown).

Figure 3.
Figure 3.

Germination of the abi3, abi4, and abi5 mutants on (+)- or (−)-ABA. Mutant seeds were sown on 10 μm (−)-ABA (solid squares) or on 3 μm (+)-ABA (open triangles), chilled for 4 days, and cultured under continuous light conditions at room temperature. Percentage of germination is plotted onto DAI.

DISCUSSION

ABA signaling during seed germination: We isolated at least 17 strong ABA-insensitive mutants and 5 weak ABA-insensitive mutants using (−)-ABA. All the mutants exhibit, more or less, insensitivity to both (+)-ABA and (−)-ABA, but the degree of insensitivity is different among these lines. Although we used an artificial compound to screen mutants, all lines identified in this study also show altered response to (+)-ABA, suggesting that these loci are involved in (+)-ABA signaling in vivo.

Among the strongest group, 4 abi3, 10 abi4, and 3 abi5 alleles were isolated. Mutations in these loci have been identified in other ABA response screens, which is consistent since reduction in these gene functions causes decreased sensitivity to both isomers (Koornneefet al. 1984; Finkelstein 1994; Table 1). Moreover, this study demonstrates that establishment of correct ABA responsiveness in the seed requires at least two additional genes, CHO1 and CHO2. These mutants exhibit significant reduction in responsiveness to (−)-ABA and subtle reduction in responsiveness to (+)-ABA. Although the molecular identity of these genes is unknown, genetic mapping suggests that the (−)-ABA-insensitive screen does identify new genes that contribute to ABA signaling in the seed. Finally, the weakest (−)-ABA-insensitive mutants appear to uncover genes involved in other plant hormone responses such as auxin and ethylene. Consistent with this, ctr1 has been identified as an enhancer of ABI1-1 and this ethylene constitutive mutant confers an ABA insensitivity to the seed in the ABI1-1 mutant background (Beaudoinet al. 2000; Ghassemianet al. 2000). Furthermore, some already characterized auxin-resistant mutants and mutants defective in auxin transport exhibit a subtle insensitivity to (+)-ABA (data not shown).

In principle, mutants that show a differential response to the ABA stereoisomers could contain mutations in genes involved in ABA reception, ABA degradation, or ABA transport. For example, in suspension-cultured barley cells, (−)-ABA is less effective than (+)-ABA in inhibiting saturable uptake of 3H-(+)-ABA and is itself not transported through the carrier as efficiently as natural (+)-ABA (Perraset al. 1994). Mutations that increase degradation or decrease transport of (−)-ABA might show a preferential decrease in sensitivity to this unnatural ABA isomer. The fact that (−)-ABA can induce some response in Arabidopsis protoplasts while having little or no effect on others also suggests that perception and/or downstream ABA signaling may be involved in the differential response between these two isomers (Windsor and Zeevaart 1997). Possibly, an ABA receptor exists that has different affinities for the two isomers. This appears likely for receptors involved in ABA reception in stomatal closure in which there is a strict requirement for the natural isomer (Sondheimeret al. 1971). In such a model, downstream outputs are differentially activated by differences in signaling flux (Figure 5). Alternatively, multiple receptors that differentiate between the two stereoisomers may signal to different ABA response outputs.

Figure 4.
Figure 4.

Effects of various abi3 mutations on sugar responsiveness. (A) Top: 7-day-old seedlings on 1 μm (+)-ABA. Bottom: 7-day-old seedlings on 1 μm (+)-ABA plus 2% glucose. (B) A schematic diagram of the ABI3 protein with the location of the various mutations identified in this study. The mutations are as follows based on GenBank accession no. X68141: abi3-8 (C- to-T transition at position 1297), abi3-9 (C-to-T transition at position 1789), abi3-10 (G-to-A transition at position 1790), abi3-11 (C-to-T transition at position 1873), abi3-12 (C-to-T transition at position 2029), and abi3-13 (G-to-A transition at position 2242). The four conserved domains have been described by Giraudat et al. (1992).

ABI3 appears to play a complex role in ABA signaling and sugar sensing: Aside from identifying new ABA response genes, our screen has also allowed the finer dissection of known ABA response genes in terms of their roles in ABA signaling. For example, a collection of abi3 missense and nonsense alleles has been useful in further understanding the role of protein motifs in ABI3 functions. The ABI3 gene is composed of four amino acid domains that are highly conserved between ABI3 orthologs. These are the A1 domain, a region in the acidic N terminus of the protein, and three COOH terminal basic domains, designated B1, B2, and B3 (Giraudatet al. 1992). In the ABI3 ortholog of maize, VP1, the B3 domain has been shown to act cooperatively with the B2 domain to bind the Sph element, an enhancer sequence that is widely conserved in seed-specific promoters (Suzukiet al. 1997). Several of our new abi3 alleles contain nonsense mutations that should produce immature proteins lacking the B3 domain. Although these abi3 nonsense alleles show reduced ABA sensitivity, their phenotypes represent only a subset of the phenotypes seen in more severe alleles of ABI3. For example, the abi3-6 mutant, a deletion allele of abi3, cannot complete late embryogenesis and its seeds are desiccation intolerant (Nambaraet al. 1994). Together, these results suggest the B3 domain of the ABI3 is not essential for completion of late embryogenesis or the acquisition of desiccation tolerance. Similar conclusions have been drawn using a truncated mutant lacking the B3 domain of VP1 in maize (McCartyet al. 1989). Possibly other B3 domain-containing proteins in Arabidopsis might complement the truncation of ABI3 B3 domain. A likely candidate is FUS3, which has been shown to act with ABI3 in the numerous aspects of seed maturation (Parcyet al. 1997; Nambaraet al. 2000).

Figure 5.
Figure 5.

A hypothetical model of parallel ABA signaling pathways. Two ABA binding proteins (ABP1 and ABP2) bind (+)-ABA or (−)-ABA to different degrees. Dotted line represents less activation than a solid line. The ABPs then activate separate parallel signaling pathways that intersect at the ABI3 transcription factor. ABI5 interacts with ABI3 to preferentially regulate the (+)-ABA response pathway.

Another abi3 allele isolated in our screen, abi3-8, confers an increased insensitivity to (+)-ABA vs. (−)-ABA. By contrast, the other abi3 alleles showed no difference or are more insensitive to (−)-ABA than to (+)-ABA. Furthermore, abi3-8 has a similar differential response to ABA isomers that were observed in the abi5 alleles. The similarity of phenotypes of these mutants suggests ABI3 and ABI5 may interact in the same ABA-dependent pathway. Recently, the B1 domain of ABI3 has been shown to interact with the ABI5 protein in a yeast two-hybrid assay (Nakamuraet al. 2001). Consistent with this, abi3-8 contains a missense mutation at leucine 298 in the B1 domain. This leucine is invariant in ABI3 orthologs identified so far. Perhaps this mutation disrupts the ABI3-ABI5 interaction, thus resulting in a preferential insensitivity to (+)-ABA. On this point, all abi5 alleles identified showed an increased insensitivity to (+)-ABA vs. (−)-ABA, suggesting ABI5 may preferentially respond to a hypothetical (+)-ABA signal vs. (−)-ABA signal (Figure 5). This is in contrast to abi4 mutants that either do not show a differential response or in some cases are more insensitive to (−)-ABA. Recently, ABI5 has been shown to be post-translationally modified in an ABA-dependent manner (Lopez-Molinaet al. 2001). It is likely that phosphorylation of the ABI5 may be one of the components of the (+)-ABA signal. Further genetic, molecular, and biochemical analysis of these alleles should reveal how ABI5 and perhaps ABI4 in combination with ABI3 may differentiate between these two isomers and provide insights into the subtleties of ABA signaling.

Recently, a number of genes that determine the response of plants to the hormones ethylene and abscisic acid have also been shown to be involved in early seedling sugar sensing (see Gazzarrini and McCourt 2001 for review). Although results suggest that ABA signaling and carbon homeostasis are tightly coupled, these interactions are complex since only a subset of ABA response mutants alter the response of plants to high sugar. Reductions in ABA biosynthesis and loss-of-function mutations in ABI4 or ABI5 confer a sugar-insensitive phenotype, but the ABI1-1, ABI2-1, and abi3-1 mutations that reduce ABA sensitivity do not show an altered sugar sensitivity. In this study we observed that although all our abi3 alleles have reduced sensitivity to exogenous ABA, specific alleles do show an altered sensitivity to glucose in the presence of ABA. It therefore appears that ABI3 has a role in sugar-ABA interactions, but this function appears to be allele specific. The lack of sugar insensitivity in some alleles may reflect the severity of the abi3 allele. However, there is no clear correlation of ABA-insensitive phenotype and the sugar response between different alleles, suggesting that the allele specificity is complex. This is further verified at the molecular level. For example, when arginine 462 is converted to a glutamine the seedling becomes sugar insensitive in the presence of ABA, whereas if the same arginine is mutated to a tryptophan decreased sugar sensitivity is not observed (Figure 4). The lack of a clear clustering of mutations between alleles that are sugar insensitive vs. sugar sensitive in the presence of ABA suggests that ABI3 perceives the sugar signal not by a single conserved domain.

The (+)/(−)-ABA insensitivity screen: Although there are advantages to using stereoisomers to identify mutations in ABA responsiveness, there are also limitations. As noted earlier, the premise of the screen was to identify mutants that showed a differential response to the ABA isomers. However, we screened first for reduced sensitivity to (−)-ABA and then further tested the seed germination on (+)-ABA in the next generation. Therefore, identification of mutants more insensitive to (+) vs. (−) isomer would be biased against. This scenario is illustrated by the fact that only 3 independent alleles of abi5 were identified, whereas 10 alleles of abi4 were uncovered. Since the ABI4 and ABI5 genes are approximately the same size, this bias is most likely due to the fact that loss-of-function abi5 mutants are more insensitive to (+)- than to (−)-ABA. We are presently testing this hypothesis by screening first for (+)-ABA insensitivity mutants and then retesting them on (−)-ABA.

Still other genes may have been missed because we used reduced sensitivity to ABA as a screening criterion. Although the use of purified isomers improves the chances of uncovering redundant functions it is still possible that loss of one redundant component causes too mild a phenotype to score.

On this note, during this screen a number of lines (class III) that failed to germinate on 10 μm (−)-ABA, but were able to germinate much faster than wild type on 1 μm (+)-ABA, were identified. Some of these mutants exhibited ethylene constitutive triple response phenotypes in the dark similar to those observed for ctr1 and eto mutants of Arabidopsis. The phenotypes of these putative mutants is consistent with previous reports that show mutations in ethylene responses alter ABA responsiveness in the ABI1-1 mutant background during seed germination (Beaudoinet al. 2000; Ghassemianet al. 2000). The fact that putative ethylene-related mutants that are very weakly insensitive to ABA were identified in our screen is consistent with the view that weak ABA-insensitive mutants can be isolated. With this said, we did identify two new loci that cause a reduction in responsiveness to (−)-ABA with only a subtle reduction in responsiveness to (+)-ABA. This demonstrates the usefulness of our screening procedure in that a different spectrum of mutants was identified compared to previous ABA-insensitive screens. This also supports the contention that the difference of the response to the ABA isomers is likely because of the difference in biological properties of these stereoisomers. The cloning of CHOTTO1 and CHOTTO2 will further test this assumption.

Acknowledgments

We thank Dr. R. Finkelstein for providing us abi4-1 and abi5-1 seeds. Funding for much of this work was provided by a grant from the National Sciences and Engineering Research Council to P.M.

Footnotes

  • Communicating editor: C. S. Gasser

  • Received December 13, 2001.
  • Accepted April 5, 2002.

LITERATURE CITED

    1. Arenas-Huertero F.,
    2. Arroyo A.,
    3. Zhou L.,
    4. Sheen J.,
    5. Leon P.
    , 2000 Analysis of Arabidopsis glucose insensitive mutants, gin5 and gin6, reveals a central role of the plant hormone ABA in the regulation of plant vegetative development by sugar. Genes Dev. 14: 2085–2096.
    1. Beaudoin N.,
    2. Serizet C.,
    3. Gosti F.,
    4. Giraudat J.
    , 2000 Interactions between abscisic acid and ethylene signaling cascades. Plant Cell 12: 1103–1115.
    1. Bell C. J.,
    2. Ecker J. R.
    , 1994 Assignment of 30 microsatellite loci to the linkage map of Arabidopsis. Genomics 19: 137–144.
    1. Bies-Etheve N.,
    2. da Silva Conceicao A.,
    3. Giraudat J.,
    4. Koornneef M.,
    5. Leon-Kloosterziel K.,
    6. et al.
    , 1999 Importance of the B2 domain of the Arabidopsis ABI3 protein for Em and 2S albumin gene regulation. Plant Mol. Biol. 40: 1045–1054.
    1. Cutler S.,
    2. Ghassemian M.,
    3. Bonetta D.,
    4. Cooney S.,
    5. McCourt P.
    , 1996 A protein farnesyl transferase involved in abscisic acid signal transduction in Arabidopsis. Science 273: 1239–1241.
    1. Dunstan D. I.,
    2. Bock C. A.,
    3. Abrams G. D.,
    4. Abrams S. R.
    , 1992 Metabolism of (+)- and (−)-abscisic acid by somatic embryo suspension cultures of white spruce. Phytochemistry 31: 1451–1454.
    1. Finkelstein R. R.
    , 1994 Mutations at two new Arabidopsis ABA response loci are similar to the abi3 mutations. Plant J. 5: 765–771.
    1. Finkelstein R. R.,
    2. Lynch T. J.
    , 2000a Abscisic acid inhibition of radicle emergence but not seedling growth is suppressed by sugars. Plant Physiol. 122: 1179–1186.
    1. Finkelstein R. R.,
    2. Lynch T. J.
    , 2000b The Arabidopsis abscisic acid response gene ABI5 encodes a basic leucine zipper transcription factor. Plant Cell 12: 599–609.
    1. Finkelstein R. R.,
    2. Wang M. L.,
    3. Lynch T. J.,
    4. Rao S.,
    5. Goodman H. M.
    , 1998 The Arabidopsis abscisic acid response loci ABI4 encodes an APETALA2 domain protein. Plant Cell 10: 1043–1054.
    1. Gazzarrini S.,
    2. McCourt P.
    , 2001 Genetic interactions between ABA, ethylene and sugar signaling pathways. Curr. Opin. Plant Biol. 4: 387–391.
    1. Ghassemian M.,
    2. Nambara E.,
    3. Cutler S.,
    4. Kawaide H.,
    5. Kamiya Y.,
    6. et al.
    , 2000 Regulation of abscisic acid signaling by the ethylene response pathway in Arabidopsis. Plant Cell 12: 1117–1126.
    1. Gibson S. I.
    , 2000 Plant sugar-response pathways. Plant Physiol. 124: 1532–1539.
    1. Giraudat J.,
    2. Hauge B. M.,
    3. Valon C.,
    4. Smalle J.,
    5. Parcy F.,
    6. et al.
    , 1992 Isolation of the ABI3 gene by positional cloning. Plant Cell 4: 1251–1261.
    1. Hugouvieux V.,
    2. Kwak J. M.,
    3. Schroeder J. I.
    , 2001 An mRNA cap binding protein, ABH1, modulates early abscisic acid signal transduction in Arabidopsis. Cell 106: 477–487.
    1. Huijser C.,
    2. Kortstee A.,
    3. Pego J.,
    4. Weisbeek P.,
    5. Wisman E.,
    6. et al.
    , 2000 The Arabidopsis SUCROSE UNCOUPLED-6 gene is identical to ABSCISIC ACID INSENSITIVE-4: involvement of abscisic acid in sugar responses. Plant J. 23: 577–585.
    1. Jeannette E.,
    2. Rona J.-P.,
    3. Bardat F.,
    4. Cornel D.,
    5. Sotta B.,
    6. et al.
    , 1999 Induction of RAB18 gene expression and activation of K+ outward rectifying channels depend on an extracellular perception of ABA in Arabidopsis thaliana suspension cells. Plant J. 18: 13–22.
    1. Karssen C. M.,
    2. Brinkhorst-van der Swan D. L. C.,
    3. Breekland A. E.,
    4. Koornneef M.
    , 1983 Induction of dormancy during seed development by exogenous abscisic acid: studies on abscisic acid deficient genotypes of Arabidopsis thaliana (L.) Heynh. Planta 157: 158–165.
    1. Konieczny A.,
    2. Ausubel F. M.
    , 1993 A procedure for mapping Arabidopsis mutations using co-dominant ecotype-specific PCR-based markers. Plant J. 4: 403–410.
    1. Koornneef M.,
    2. Reuling G.,
    3. Karssen C.
    , 1984 The isolation and characterization of abscisic acid-insensitive mutants of Arabidopsis thaliana. Physiol. Plant 61: 377–383.
    1. Laby R. J.,
    2. Kincaid M. S.,
    3. Kim D.,
    4. Gibson S.
    , 2000 The Arabidopsis sugar-insensitive mutants sis4 and sis5 are defective in abscisic acid synthesis and response. Plant J. 23: 587–596.
    1. Leung J.,
    2. Giraudat J.
    , 1998 Abscisic acid signal transduction. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49: 199–222.
    1. Leung J.,
    2. Bouvier-Durand M.,
    3. Morris P.-C.,
    4. Guerrier D.,
    5. Chefdor F.,
    6. et al.
    , 1994 Arabidopsis ABA response gene ABI1: features of a calcium-modulated protein phosphatase. Science 264: 1448–1452.
    1. Leung J.,
    2. Merlot S.,
    3. Giraudat J.
    , 1997 The Arabidopsis ABSCISIC ACID-INSENSITIVE2 (ABI2) and ABI1 genes encode homologus protein phosphatase 2C involved in abscisic acid signal transduction. Plant Cell 9: 759–771.
    1. Lopez-Molina L. L.,
    2. Chua N.-H.
    , 2000 A null mutation in a bZIP factor confers ABA-insensitivity in Arabidopsis thaliana. Plant Cell Physiol. 41: 541–547.
    1. Lopez-Molina L. L.,
    2. Mongrand S.,
    3. Chua N-.H.
    , 2001 A post-germination developmental arrest checkpoint is mediated by abscisic acid and requires the ABI5 transcription factor in Arabidopsis. Proc. Natl. Acad. Sci. USA 98: 4782–4787.
    1. McCarty D. R.,
    2. Carson C. B.,
    3. Lazar M.,
    4. Simonds S. C.
    , 1989 Transposable element induced mutations of the viviparous-1 gene of maize. Dev. Genet. 10: 473–481.
    1. McCourt P.,
    2. Keith K.
    , 1998 Sterile techniques in Arabidopsis. Methods Mol. Biol. 82: 13–17.
    1. Meyer K.,
    2. Leube M.,
    3. Grill E.
    , 1994 A protein phosphatase 2C involved in ABA signal transduction in Arabidopsis thaliana. Science 264: 1452–1455.
    1. Nambara E.,
    2. Keith K.,
    3. McCourt P.,
    4. Naito S.
    , 1994 A deletion mutant of the Arabidopsis ABI3 gene. Plant Cell Physiol. 35: 509–513.
    1. Nambara E.,
    2. Keith K.,
    3. McCourt P.,
    4. Naito S.
    , 1995 A regulatory role for the ABI3 gene in the establishment of embryo maturation in Arabidopsis thaliana. Development 121: 629–636.
    1. Nambara E.,
    2. Hayama R.,
    3. Tsuchiya Y.,
    4. Nishimura M.,
    5. Kawaide H.,
    6. et al.
    , 2000 The role of ABI3 and FUS3 loci in Arabidopsis thaliana on phase transition from late embryo development to germination. Dev. Biol. 220: 412–423.
    1. Nakamura S.,
    2. Lynch T. J.,
    3. Finkelstein R. R.
    , 2001 Physical interactions between ABA response loci of Arabidopsis. Plant J. 26: 627–635.
    1. Parcy F.,
    2. Valon C.,
    3. Raynal M.,
    4. Gaubier-Comella P.,
    5. Deleseny M.,
    6. et al.
    , 1997 The ABSCISIC ACID-INSENSITIVE3, FUSCA3, and LEAFY COTYLEDON1 loci act in concert to control multiple aspects of Arabidopsis seed development. Plant Cell 9: 1265–1277.
    1. Pego J. V.,
    2. Weisbeek P. J.,
    3. Smeekens S. C. M.
    , 1999 Mannose inhibits Arabidopsis germination via a hexokinase-mediated step. Plant Physiol. 119: 1017–1023.
    1. Perras M.,
    2. Abrams S. R.,
    3. Balsevich J. J.
    , 1994 Characterization of an abscisic acid carrier in suspension-cultured barley cells. J. Exp. Bot. 280: 1565–1573.
    1. Soderman E. M.,
    2. Brocard I. M.,
    3. Lynch T. M.,
    4. Finkelstein R. R.
    , 2000 Regulation and function of the Arabidopsis ABA-insensitive4 gene in seed and abscisic acid response signaling network. Plant Physiol. 124: 1752–1765.
    1. Sondheimer E.,
    2. Galson E. C.,
    3. Chang Y. P.
    , 1971 Asymmetry, its importance to the action and metabolism of abscisic acid. Science 174: 829.
    1. Suzuki M.,
    2. Kao C. Y.,
    3. McCarty D. R.
    , 1997 The conserved B3 domain of VIVIPAROUS1 has a cooperative DNA binding activity. Plant Cell 9: 799–807.
    1. Walker-Simmons M. K.,
    2. Anderberg R. J.,
    3. Rose P. A.,
    4. Abrams S. R.
    , 1992 Optically pure abscisic acid analogs—tools for relating germination inhibition and gene expression in wheat embryos. Plant Physiol. 99: 501–507.
    1. Windsor M. L.,
    2. Zeevaart J. A. Z.
    , 1997 Induction of ABA 8′-hydroxylase by (+)-S-, (−)-R-and 8′,8′,8′-trifluoro-S-abscisic acid in suspension cultures of potato and Arabidopsis. Phytochemistry 45: 931–934.
    1. Xiong L.,
    2. Lee B.-H.,
    3. Ishitani M.,
    4. Lee H.,
    5. Zhang C.,
    6. et al.
    , 2001 FIERY1 encoding an inositol polyphosphate 1-phosphatase is a negative regulator of abscisic acid and stress signaling in Arabidopsis. Genes Dev. 15: 1971–1984.
    1. Zeevaart J.,
    2. Creelman R.
    , 1988 Metabolism and physiology of abscisic acid. Annu. Rev. Plant Physiol. Plant Mol. Biol. 39: 439–473.