Abscisic Acid Biosynthesis Gene Underscores the Complexity of Sugar, Stress, and Hormone Interactions

The plant hormone abscisic acid (ABA) plays a central role in many aspects of plant growth, including seed development, dormancy, and germination, and plant stress responses, such as drought and osmotic stress. ABA typically is described as a plant growth inhibitor, because of the inhibitory effect of exogenously applied ABA on seed germination and seedling growth, which often are used as bioassays for this hormone. However, the stunted phenotypes of ABA-deficient mutants suggest that endogenous ABA may play a role in growth promotion (Sharp et al., 2000; Finkelstein et al., 2002). Another well-known effect of ABA is the promotion of stomatal closure; ABA-deficient plants typically exhibit a wilty phenotype as a result of their inability to close stomata. ABA-deficient and ABA-insensitive mutants have helped to define the biochemical pathway of ABA biosynthesis and to elucidate the various roles of ABA in plant growth and development. ABA-deficient mutants have been identified in a variety of plant species, including Arabidopsis, maize, tobacco, bean, potato, and tomato, in genetic screens for precocious germination and/or a wilty phenotype (Seo and Koshiba, 2002). ABA controls seed dormancy by antagonizing the promotion of seed germination by another hormone, gibberellic acid. Thus, ABA-deficient mutants of Arabidopsis also have been isolated in screens for germination in the pres-ence of the gibberellic acid biosynthesis inhibitor paclobutrazol (Léon-Kloosterziel et al., 1996) and in screens for revertants in nongerminating gibberellin-sensitive mutants (Koornneef et al., 1982).

Exogenously applied Glc also exerts an inhibitory effect on seedling growth, and Glc-insensitive (gin) and Glc-oversensitive (glo) mutants of Arabidopsis have been isolated to investigate this phenomenon (reviewed by Rolland et al., 2002). Some of the gin mutants turned out to have ABA deficiency or ABA insensitivity. For example, gin1, gin5, and gin6 are allelic to aba2/sis4/isi4, aba3/los5, and abi4/sun6/sis5/isi3, respectively. In this issue of The Plant Cell, Cheng et al. (pages 2723–2743) report on the molecular cloning of Arabidopsis ABA2/GIN1 and present a detailed genetic, biochemical, and phenotypic analysis of gin1 mutants (Figure 1). ABA2/GIN1 encodes a short-chain dehydrogenase/reductase called SDR1, which catalyzes the conversion of xanthoxin to ABA aldehyde (ABAld). The probable allelic nature of gin1/aba2/sis4/isi4 and the identification of ABA2 as a putative short-chain dehydrogenase/reductase was discussed by Rook et al. (2001), who isolated isi4 (for impaired Suc induction) based on its altered expression of a negative selection marker under the control of the Suc-inducible promoter of the ApL3 gene for a subunit of ADP-Glc pyrophosphorylase. Schwartz et al. (1997) also suggested that the aba2 mutant is deficient in a short-chain dehydrogenase/reductase that functions in ABA biosynthesis, based on a biochemical characterization of the mutant phenotype. Interestingly, the salt-resistant mutants salt resistant1 and salobreño3, which are able to germinate under high-salt conditions, also were found to be allelic to aba2, confirming the importance of ABA in the plant response to salt or osmotic stress (Quesada et al., 2000; González-Guzmán et al., 2002). González-Guzmán et al. (2002) recently reported on the map-based cloning of Arabidopsis ABA2 and confirmed the identification of the ABA2 protein as a short-chain dehydrogenase/reductase that catalyzes the NAD-dependent conversion of xanthoxin to ABAld. Cheng et al. (2002) provide further confirmation for the identification of ABA2 via genetic crosses, map-based cloning, com-plementation, and biochemical analysis of the recombinant SDR1 protein. In addition, this work provides exciting new information and ideas on several fronts, including ABA biosynthesis and transport and interactions between ABA, Glc, and other hormones in regulating plant growth and plant responses to stress.

Figure 1.

Wild-Type and aba2/gin1 Mutant Plants.

The insets show strong GIN1:GUS expression in Glc-treated seedlings (bottom center), and in vascular tissue of hypocotyl and leaf veins (middle center), mature pollen (top left), the base of young siliques (top center), and funiculi of mature siliques (top right).

ABA BIOSYNTHESIS AND TRANSPORT

The ABA biosynthesis pathway proceeds from zeaxanthin to xanthoxin via several steps catalyzed by zeaxanthin epoxidase (which is impaired in the Arabidopsis aba1 mutant), an unidentified violaxanthin isom-erase, and 9′-cis-epoxycarotenoid dioxygenase. The maize viviparous14 (Tan et al., 1997) and tomato notabilis (Burbidge et al., 1999) mutants are disrupted in 9′-cis-epoxycarotenoid dioxygenase function, and this reaction was shown to be a key regulatory step in ABA biosynthesis (Qin and Zeevaart, 1999). Xanthoxin could be converted to ABA in a multiple-step process via ABAld or, alternatively, xanthoxic acid (Milborrow, 2001; Seo and Koshiba, (2002). The work of Cheng et al. (2002), together with that of González-Guzmán et al. (2002) (2002), demonstrates conclusively that the single SDR1 enzyme, encoded by ABA2, catalyzes all three steps in the NAD-dependent conversion of xanthoxin to ABAld (i.e., oxidation of the 4′-hydroxyl to a ketone, desaturation of a 2′-3′ bond, and opening of the epoxide ring) with a K_m of 19 to 20 μM. Cheng et al. (2002) further show that xanthoxic acid is not a substrate for SDR1 in vitro. Because aba2 null mutants are severely deficient in ABA, the work of both groups suggests that the conversion of xanthoxin to ABA via ABAld is the major pathway of ABA biosynthesis in Arabidopsis. It is interesting that SDR1 is a single-copy gene in Arabidopsis that does not appear to be “backed up” by other functionally overlapping or redundant genes.

Although it is known that carotenoid precursors of ABA are produced in plastids and the final steps of ABA biosynthesis are believed to be cytosolic, the precise locations of the biosynthetic enzymes and translocation pathways of the various intermediates have not been determined definitively (Seo and Koshiba, 2002). Cheng et al. (2002) show that a GIN1–green fluorescent protein fusion is localized exclusively in the cytosol of transiently transformed protoplasts. Spatial and temporal expression patterns of ABA2/GIN1 were analyzed based on the expression of the β-glucuronidase (GUS) reporter gene under the control of the ABA2/GIN1 promoter in transgenic Arabidopsis throughout the life cycle. In transgenic seedlings, GUS staining appeared in the vascular bundles of the hypocotyl and branch points of lateral and primary roots in young seedlings and then later became prominent in petioles of cotyledons and leaves (Figure 1). Adult plants showed strong staining in leaf petioles, mature pollen, and at the junction of pedicels and young siliques. Staining at the base of young siliques appeared only after the abscission of petals, sepals, and anthers, prompting the authors to suggest that ABA may play a role in early embryogenesis. Strong expression in vascular tissue, coupled with a distinct lack of GUS staining in known ABA “target sites,” such as mesophyll cells, guard cells, and seeds, suggests that ABAld, or ABA itself, may be transported to specific target sites after synthesis in another location. These interesting observations require further detailed investigation.

GLC SIGNALING AND ABA

Glc and other sugars play a role as signaling molecules in plants and other eukaryotes, providing information about the nutrient status of cells and organs and affecting transcriptional and post-transcriptional processes that allow growth and development to proceed in accord with certain environmental conditions. For example, in yeast, Glc in the medium (a preferred carbon source) rapidly induces the repression of genes involved in respiration and the metabolism of alternative carbon sources and the derepression of genes associated with fermentation and Glc metabolism (Rolland et al., 2001). In plants, sugar-induced feedback control of photosynthesis is well documented, and sugars are known to regulate gene expression associated with other developmental and metabolic processes, including germination, growth, and flowering (reviewed by Rolland et al., 2002).

The mechanism(s) of sugar sensing in plant cells is an area of intense research and considerable debate. There is evidence that hexokinase functions as a sugar sensor in plants as well as yeast (Jang and Sheen, 1997; Rolland et al., 2002). This model implies separate catalytic and regulatory functions of hexokinase, which have not been demonstrated unequivocally (Halford et al., 1999a, 1999b; Moore and Sheen, 1999). Xiao et al. (2000) and Rolland et al. (2001) suggested that there are at least three Glc-sensing mechanisms in plants: hexokinase-dependent and hexokinase-independent systems and a third system dependent on active glycolysis downstream of hexokinase. It is likely that multiple sugar-sensing mechanisms exist in plants, and the precise biochemical and molecular nature of these sugar sensors awaits further research.

It is clear that sugar signal transduction in plants is tightly linked with ABA and with other processes affected by ABA, such as osmotic stress and ethylene signaling. Analysis of the gin1/aba2 mutant by Cheng et al. (2002) provides clues to the nature of these interactions. First, the authors show that ABA deficiency is the cause of the gin1 phenotype, because exogenous ABA application restored Glc sensitivity in the mutant seedlings. Second, in wild-type plants, Glc application induces ABA2/GIN1 expression and the expression of a number of other ABA biosynthesis and ABA signaling genes. Osmotic stress also is known to induce ABA biosynthesis and signaling, and it is possible that this effect is attributable to an increased osmotic potential by the addition of Glc to the medium, rather than a specific effect of Glc. Indeed, application of mannitol as an osmotic control also caused the induction of a number of ABA biosynthesis and signaling genes. However, the responses to Glc and to mannitol appear to be distinct. For example, the expression of the ABA biosynthesis genes ABA1, ABA2, and AAO3 remained high in 6% Glc, but it was diminished in 6% mannitol. A striking difference also was seen in the response of the ABA signaling gene ABI4, which was strongly induced by 6% Glc but not by 6% mannitol. Cheng et al. (2002) also observed different responses to Glc versus mannitol for the ABA response genes ABI3 and ABI5, further supporting the notion that the effects of Glc and osmotic stress, although overlapping, are distinct.

ABI4 functions downstream of ABA biosynthesis and encodes an AP2-domain protein characteristic of transcription factors known to play roles in the regulation of plant development (Finkelstein et al., 1998). The abi4 mutant is allelic to the gin6 Glc-insensitive mutant, and the addition of ABA to the growth medium does not restore Glc sensitivity to abi4/gin6 (which is consistent with the nature of the abi4 mutant, which is not ABA deficient but is ABA insensitive). ABI4 may represent a key step in the regulation of downstream responses of ABA and sugar signaling (Arenas-Huertero et al., 2000; Huijser et al., 2000; Laby et al., 2000; Rook et al., 2001). The abi4 mutant analyzed by Finkelstein et al. (1998) exhibited a seed-specific phenotype, and expression analysis showed abundant tran-scripts in developing siliques. However, low-level expression also was observed in vegetative tissue, suggesting that the protein could play a role in vegetative growth as well. Analysis of a gin6/abi4 mutant allele confirmed a role for ABI4 in Glc-mediated developmental arrest during vegetative morphogenesis (Arenas-Huertero et al., 2000). Functional analysis of GIN6/ABI4 may help to define Glc signaling responses.

GLC SIGNALING AND ETHYLENE

Glc signaling also is interconnected with ethylene responses, and the Glc and ethylene signaling pathways have been found to antagonize each other (Zhou et al., 1998). For example, the constitutive ethylene mutant constitutive triple response1 (ctr1) exhibits a Glc-insensitive phenotype, and the mutant ethylene insensitive 2 (ein2) is oversensitive to Glc (i.e., Glc application causes repression of growth that is more severe than in the wild type). Cheng et al. (2002) show through examination of the phenotype and genetic crossing that ctr1 is allelic to gin4. They also found that Glc oversensitivity in the ein2 mutant is epistatic to gin1/aba2 in ein2 gin1 double mutant plants and that the expression of the ethylene response gene PDF1.2 is repressed by Glc in the wild type but not in the gin1 mutant. These results confirm the notion that Glc responses require ABA and suggest that the antagonistic effects of ethylene and Glc are mediated in part by ABA.

The ethylene signaling pathway is reasonably well understood in Arabidopsis (reviewed by Wang et al., 2002). In the absence of an ethylene signal, ethylene receptors (there are five in all: ETR1, ETR2, EIN4, ERS1, and ERS2) activate a Raf-like kinase, CTR1, and CTR1 in turn negatively regulates the downstream ethylene response pathway controlled by EIN2. The binding of ethylene inactivates the receptors, resulting in the deactivation of CTR1, which allows EIN2 to function as a positive regulator of downstream responses. Experiments conducted by Ghassemian et al. (2000) and Beaudoin et al. (2000) suggest that ABA inhibits root growth by signaling through the ethylene signaling pathway. However, the results of these experiments are confusing, and many of the details remain unresolved. The work of Cheng et al. (2002) suggests that Glc signaling may be an important component of the interactions between ABA and ethylene. Further analysis of gin4/ctr1 will be of particular interest in this regard. Perhaps it is more appropriate to think of the EIN2 pathway as a general signal response pathway that is regulated by various signals, including (and not limited to) Glc, ABA, osmotic stress, jasmonic acid, and ethylene.

GROWTH-PROMOTING FUNCTIONS OF ABA

Cheng et al. (2002) observed that gin1/aba2 mutant plants showed severe growth retardation compared with wild-type plants under optimal growing conditions and suggest that endogenous ABA could be critical for promoting plant growth, in apparent contrast to the well-established role of ABA as a growth inhibitor and stress hormone. It has been reported that ABA plays a role in promoting root and shoot growth independent of its effects on plant water balance (Sharp et al., 2000; Spollen et al., 2000).

Cheng et al. (2002) propose the interesting idea that ABA may possess dual functions: as a growth inhibitor in the presence of stresses (associated with high induced ABA levels) and a growth promoter in the absence of stresses (associated with low endogenous ABA levels). Alternatively, ABA could function as a growth inhibitor in certain cells or tissues (e.g., leaves) and as a growth promoter in other tissues (e.g., roots). The ability of ABA to act as either a promoter or an inhibitor of downstream processes likely depends on the downstream signaling components, such as transcription factors, that may operate in tissue-, development-, or environment-specific contexts and/or on the interactions of ABA with other hormones, such as ethylene. For example, studies in maize (Spollen et al., 2000) and tomato (Sharp et al., 2000) suggest that the stunted growth of ABA-deficient plants is caused by the overproduction of ethylene and that one function of ABA may be to prevent the overproduction of ethylene. Further studies of Glc, ABA, and ethylene mutants in the context of all three signaling pathways may help to elucidate these complex interactions.