The ABSCISIC ACID-INSENSITIVE 3 (ABI3) gene of Arabidopsis thaliana (L.) Heynh is known to play an important role during seed maturation and dormancy. Here, we present evidence suggesting an additional role for ABI3 during vegetative quiescence processes. During growth in the dark, ABI3 is expressed in the apex of the seedlings after cell division is arrested. The 2S seed storage protein gene, a target gene of ABI3 in seeds, is also induced in the arrested apex under similar darkness conditions. In addition, β-glucuronidase expression under the control of the ABI3 promoter is abolished by treatments that provoke leaf development in the dark [sucrose and abscisic acid (ABA) biosynthesis inhibitors] and induced by treatments that prevent leaf development (darkness and ABA). Furthermore, ABI3 expression is absent in apices of dark-grown de-etiolated (det 1) and abi3 mutants, both known to develop leaves or leaf primordia in the dark. The fact that the expression of the ABI3 gene is only observed in a fraction of the analysed plants suggests that ABI3 is probably only one of the components of a molecular network underlying quiescence. In addition to the expression of ABI3 in apices of dark-grown seedlings, the ABI3 promoter confers expression in other vegetative organs as well, such as the stipules and the abscission zones of the siliques. In conclusion, apart from its role in seed development, ABI3 might have additional functions.
Plants, being sessile organisms, have developed strategies to survive changing and potentially harsh environmental conditions. Quiescence provides the persistence of the plant in an adverse environment through the partial or complete arrest of meristematic growth. This arrest is relieved immediately after the environmental limitation has been overcome. In addition to this adaptive purpose, quiescence is also an indispensable part of developmental processes, such as seed maturation. In the latter process, abscisic acid (ABA) is crucial in the prevention of precocious germination and subsequently to induce seed dormancy (Black 1991; Galau, Jakobsen & Hughes 1991). Similarities between the developmental quiescence in the seed and adaptive quiescence processes in the apex have often been discussed in physiological studies, especially with respect to tree species. Both processes are characterized by the involvement of ABA, a chilling requirement to overcome rest, the accumulation of late-embryogenesis abundant (LEA) proteins, the occurrence of lipid and protein bodies, and a low water potential (Saure 1985; Powell 1987; Borchert 1991; Sagisaka 1991; Dennis 1996; Salzman et al. 1996). Nevertheless, more evidence is needed to determine whether vegetative and seed quiescence are regulated by the same mechanisms. This led us to investigate the role of the ABA-insensitive 3 (ABI3) protein, a key regulator of seed development and dormancy, for its additional involvement in vegetative quiescence.
The Arabidopsis abi3 mutant was originally isolated for its ability to germinate in the presence of inhibiting concentrations of ABA (Koornneef, Reuling & Karssen 1984). Seeds that were homozygous for the pleiotropic recessive abi3 mutation fail to develop seed dormancy and to accumulate reserve proteins and lipids properly (Nambara, Naito & McCourt 1992). Seeds of severe abi3 alleles become desiccation intolerant and are green at maturity (Ooms et al. 1993; Nambara et al. 1994). The action of ABI3 is believed to be seed specific (Koornneef et al. 1989; Finkelstein & Somerville 1990; Giraudat et al. 1992; Nambara et al. 1994, 1995; Parcy et al. 1994). Studies with transgenic plants overexpressing ABI3, however, indicated that ABI3 can function in tissues other than seeds. Ectopically expressed ABI3 can potentiate ABA responses in vegetative tissues (Parcy & Giraudat 1997). The ectopic expression of P35S-ABI3 was necessary and sufficient to direct in response to ABA the ectopic expression of downstream target genes, such as 2S, CRC, and Em, major seed protein genes, in leaves of Arabidopsis (Parcy et al. 1994). Similarly, the expression of P35S-PvALF (the homologue of ABI3 in Phaseolus vulgaris) provoked the expression of a chimeric phaseolin-GUS gene in mature bean leaves (Bobb, Eiben & Bustos 1995). Additionally, the ABI3-homologous gene OSVP1 of Oryza sativa was found to be expressed in suspension-cultured cells (Nakagawa et al. 1996). Given its important role in seed maturation and its ability to function in vegetative tissues, ABI3 seemed a particularly interesting candidate to test the hypothesis on the similarity of at least some molecular aspects of quiescence in different organs.
Here, we show that the expression of the ABI3 gene in Arabidopsis thaliana is not limited to the seed. In agreement with the hypothesized role of ABI3 in vegetative quiescence, the expression of ABI3 was induced in the apex of dark-grown Arabidopsis seedlings after cell division had ceased. Also after transfer from light to darkness, ABI3 was expressed at the apex. In addition, ABI3 expression was abolished by treatments that provoked leaf development in the dark (sucrose, ABA biosynthesis inhibitors) and induced by treatments that prevented leaf development (darkness, ABA). Furthermore, ABI3 expression was absent in the apices of dark-grown de-etiolated (det1) and abi3 mutants, both of which are known to develop leaves or leaf primordia in the dark. Together, these data suggest a role for ABI3 in apex quiescence.
MATERIALS AND METHODS
Plant material and growth conditions
The abi3-4 and abi3-5 mutants are in the Arabidopsis thaliana (L.) Heynh. Landsberg erecta (Ler) ecotype, whereas the abi3-3, abi3-6, det1, and det2 mutants are in the Columbia (Col) ecotype. PCYC1At-GUS is in C24 (Ferreira et al. 1994) and P2S-GUS and PUSP-GUS are in Col (Bäumlein et al. 1991). The general expression pattern of P2S-GUS lines correlated with that obtained by in situ hybridization (W. Boerjan, unpublished results; Guerche et al. 1990). The transgenic PABI3-GUS and P35S-ABI3 were available in both Ler and C24 ecotypes (Parcy et al. 1994).
For germination, the seeds were surface sterilized and placed on a Murashige and Skoog medium (Murashige & Skoog 1962) supplemented with 10 g L–1 sucrose. After a cold treatment overnight for homogenous germination, the seeds were grown at 20 °C, 50 μmol m–2 s–1 light intensity, 70% relative humidity, and under a 16 h light : 8 h dark cycle. For the dark treatments, the plates were wrapped twice with aluminium foil and placed in a dark container in the same culture room. The greenhouse conditions for plant growth and crosses were as follows: 23 °C (without shielding from incident daylight), 50 μmol m–2 s–1 light intensity at plant level (MBFR/U 400 W incandescent lamps; Philips, Eindhoven, The Netherlands), 40% relative humidity, and a 16 h light : 8 h dark cycle. Harvested seeds (except for green, desiccation-intolerant, homozygous abi3) were treated for 2 weeks at 28 °C and stored at 4 °C until use.
The crosses of abi3-6, det1, and det2 (Col) with PABI3-GUS (C24) were carried out without emasculation by using PABI3-GUS as a pollinator. Homozygous abi3, det1 and det2 F2 plants were selected on medium with 50 mg L–1 kanamycin and lines homozygous for the PABI3-GUS construct were identified among the F3 plants and used for the experiments. F2 plants homozygous for the PABI3-GUS construct and heterozygous or wild type for det1 and det2 were used as controls.
All experiments were carried out with at least 50 plants, unless otherwise specified. The appearance of at least the first leaf pair was scored as positive leaf development.
Histochemical β-glucuronidase (GUS) assays
GUS was assayed as described by Jefferson (1987). After the histochemical reaction, the material was fixed with 3% glutaraldehyde in phosphate buffer for 1 h, washed twice with phosphate buffer, and passed over 30, 50, and 70 to 95% ethanol to remove the chlorophyll. The material was cleared with chlorallactophenol prior to examination under a light microscope (Beeckman & Engler 1994). No GUS activity was detected in transgenic plants that carried a promoterless GUS construct (Parcy et al. 1994), which served as a negative control.
Whole-mount in situ hybridization
Whole plants were dehydrated, digested with proteinase K, and fixed as described (de Almeida Engler, Van Montagu & Engler 1994). A 450 bp poplar ABI3 probe (GenBank Accession No. aj003166; base pairs 1538–1981 with a 65% overall identity and > 90% identity over a stretch of 250 bp to the region between 1854 and 2321 bp of the corresponding Arabidopsis sequence) was used in sense and antisense orientation for labelling with digoxigenin UTP (Boehringer, Mannheim, Germany). Southern hybridization to genomic DNA of Col and C24 revealed that the probe recognized a single gene in Arabidopsis. A final concentration of 3 μg mL–1 of probe in a 50% formamide, 5 × SSC solution (1 × SSC: 150 mM NaCl, 15 mM Na3-citrate, pH 7·0) that contained 50 μg mL–1 heparin and 100 μg mL–1 salmon sperm DNA were used to hybridize the whole plant material overnight. RNA/RNA hybrids were detected with antidigoxigenin Fab fragments conjugated to alkaline phosphatase. The alkaline phosphatase reaction was allowed to proceed in the dark until sufficient dark precipitate was observed.
For Nomarski microscopy, the plant material was fixed with 3% glutaraldehyde, passed over ethanol changes, and cleared with chlorallactophenol (Beeckman & Engler 1994). A minimum of 40 individual plants were examined for both sense and antisense probes.
PABI3-GUS, P2S-GUS, and PUSP-GUS are expressed in vegetative tissues during dark treatments
Arabidopsis plants, transformed with the chimeric PABI3-GUS construct (Parcy et al. 1994), were used to address the question whether, in analogy to quiescence in the seed, the ABI3 gene was also expressed in vegetative tissues during growth-arresting conditions. First, it was necessary to identify a developmental state at which quiescence was induced. Because it is known that etiolated Arabidopsis seedlings do not develop leaves in the dark, the apex of dark-grown seedlings was chosen to study the role of ABI3 in vegetative quiescence. The reasoning was that if ABI3 were involved in the establishment or maintenance of vegetative quiescence, the chimeric PABI3-GUS gene should be expressed in the apex at or after the arrest of cell division. Previously, it had been shown that CycB1 (new name for Cyc1At), a cell cycle gene, is strictly expressed during mitosis (Ferreira et al. 1994; Shaul, Van Montagu & Inzé 1996; Renaudin et al. 1998). To indicate the approximate day at which cell divisions in the apex were arrested, the expression of the chimeric PCYC1At-GUS gene (Ferreira et al. 1994) was followed in dark-grown Arabidopsis seedlings (C24). Whereas 16% (n = 76) of the PCYC1At-GUS seedlings still showed expression in the apex after 4 d of germination in the dark, none of the seedlings expressed the PCYC1At-GUS gene after 7 d in the dark (n = 60).
For the sake of clarity throughout the text, we will refer to three different quiescence states (Qs) that occur at the end of seed development (quiescence state 1, QS1); at the end of seedling development (quiescence state 2, QS2); and after leaf development was initiated (quiescence state 3, QS3).
Approximately 10% of the PABI3-GUS plants in C24 (n = 355) and 15% in Ler (n = 377) showed GUS activity in the apex after 10–15 d in the dark (Fig. 1a), corresponding to QS2. None of the seedlings showed GUS activity before that time point in the apex (data not shown), indicating de novo expression from the ABI3 promoter. Whole-mount in situ hybridizations confirmed that ABI3 was expressed in the apex (Figs 1b & c). No RNA signal was observed in the cotyledons.
Light-germinated plants, transferred after 5 d to the dark, showed induction of PABI3-GUS in the apices of 10% of the plantlets (n = 120 in Ler, n = 60 in C24) 12–15 d after transfer to the dark (Fig. 1d), although the growth arrest was less homogenous throughout the tested plants. GUS activity was found in the apex and/or the subapical region of the plants. Because the partial arrest was initiated after leaf development had started, we refer to this state as QS3. No GUS activity was observed in the apices of light-grown seedlings or in the axillary buds of the rosette of greenhouse-grown plants (data not shown). Most probably, the development and growth of these meristems is fast and not interrupted by an arrest under optimal growth conditions.
During seed maturation, ABI3 is necessary for the accumulation of seed storage proteins, such as the 2S napin; in vegetative tissues, overexpression of ABI3 results in the ectopic expression of 2S in response to ABA (Parcy et al. 1994). Consequently, we investigated whether P2S-GUS (in Col) was also expressed during apex quiescence in the dark. The chimeric P2S-GUS construct contains a 1890 bp sequence upstream of the Arabidopsis 2S1 gene fused to the GUS-coding sequence (W. Boerjan, unpublished results; De Clercq et al. 1990; Guerche et al. 1990). Interestingly, plants that carry the chimeric P2S-GUS construct showed an induction of GUS activity in the apex after 10–15 d of germination in the dark (Fig. 1e) and 12–15 d after transfer from light to darkness (Fig. 1f). The GUS activity in the apex or the subapical zone was always accompanied by slight vascular expression, which was most probably due to stability of the GUS protein during early seedling development. Similar results were obtained with PUSP-GUS (in Col), in which the promoter is derived from the unknown seed protein (USP) gene of Vicia faba (data not shown; Bäumlein et al. 1991). In contrast to the expression of PABI3-GUS, which was detected in only 10–15% of the apices of dark-grown plants, P2S-GUS and PUSP-GUS were expressed in all individuals tested. In conclusion, the expression of the ABI3, 2S, and USP genes, so far believed to be seed specific, was also detected in the growth-arrested apex of dark-grown Arabidopsis, suggesting parallels in the developmental programs of seed maturation and apex quiescence.
ABA and sucrose influence the establishment of growth arrest in the dark
Given the fact that PABI3-GUS was only expressed in a fraction of the individuals, PABI3-GUS expression was probably strongly dependent on the environmental conditions. Sucrose and ABA were chosen as two external factors with contrasting effects on growth in the dark and therefore possibly also on PABI3-GUS expression. To test the effects of these factors, the uppermost part of dark-grown seedlings was grown in contact with media either with or without sucrose or with sucrose supplemented with 50 μM ABA (Fig. 2). The effects on growth were investigated both in wild type and in abi3 mutants. When sucrose was present in the medium, 70–100% of the wild-type and abi3 seedlings developed leaves in the dark, 3 weeks after germination (Fig. 2). Under these conditions, no GUS activity was detected in the apex of either PABI3-GUS (n = 60) or P2S-GUS (n = 50) plants during the 4 weeks of treatment in the dark (Figs 1g & i), suggesting that when sucrose is available to the apex, leaf development is promoted and both QS2 and QS3 are absent.
The simultaneous application of sucrose and 50 μM ABA to the medium suppressed the positive effect of sucrose on leaf development in the dark (Fig. 2) and PABI3-GUS expression was induced in the apex of 43% (n = 79) of the plants (Fig. 1h). Expression of ABI3 was confirmed by whole-mount in situ hybridization (data not shown). Interestingly, when abi3 mutants were analysed, a similar sensitivity of the apex to ABA was observed, although these mutants were originally isolated as being ABA-insensitive during germination (Fig. 2). Notably, ABA application did not result in a complete repression of leaf development, but could only suppress the positive effect that was provoked by sucrose 3 weeks after germination.
The above data suggested that ABA was involved in apex quiescence. To further substantiate this observation, the effect of the ABA biosynthesis inhibitors fluridone and norflurazon on leaf development was investigated. Fluridone did not affect the frequency or kinetics of germination, but promoted leaf development in the dark at significant levels after 4 weeks (Fig. 3). Similar results were obtained with norflurazon (data not shown). This confirmed the importance of ABA to retard leaf development in the dark, as also demonstrated in the previous experiment (Fig. 2).
PABI3-GUS expression is absent in det1 mutants that develop leaves in the dark
To further substantiate our hypothesis that expression of ABI3 is associated with QSs, ABI3 expression in det mutants was analysed. The det1 mutants resemble light-grown plants when germinated in the dark. They develop chloroplast-containing leaves immediately after germination in the dark and leaf development is continuous until a rosette has formed (Chory et al. 1989). DET1 is a light-signal transduction component and acts most probably as a repressor of photomorphogenesis in the dark (Pepper et al. 1994). Conversely, det2 mutants have a short hypocotyl and expanded cotyledons after 7 d in the dark and form primary leaf buds that develop into leaves without chloroplasts (Chory et al. 1991). Leaf development in det2 mutants starts approximately 14 d after germination in the dark. Although, in the dark, det2 mutants share the phenotype of de-etiolated leaf development with det1 mutants, DET2 was found to encode a steroid 5α-reductase in the brassinolide biosynthetic pathway, which is thought to be regulated in response to light (Li et al. 1996; Chory & Li 1997). As these mutants have contrasting kinetics of leaf development in the dark, they are ideal candidates to test our hypothesis that ABI3 is expressed during quiescence and not during leaf development.
For these experiments, the PABI3-GUS construct was crossed into the det1 and det2 mutants and its expression analysed. At QS1, both det1 and det2 mutants have no defects and PABI3-GUS expression was similar to that of wild type in histochemical assays (data not shown). At the apical meristem of det1 seedlings, no PABI3-GUS expression was detected during 28 d of growth in the dark. During this period, det1 developed leaves continuously (Fig. 4b). In contrast, PABI3-GUS expression was detected in det2 mutants at QS2. At this stage, leaf development in det2 is retarded, as outgrowth of the leaves only occurs after 14 d of growth in the dark (Fig. 4c) (Chory, Nagpal & Peto 1991). After 14 d of germination in the dark, both det1 and det2 formed leaves in a qualitatively light-independent manner. No QS3 was imposed and no PABI3-GUS expression was observed (Figs 4e & f). Again, these data suggest that PABI3-GUS is only expressed at stages of vegetative quiescence, and not during active growth.
PABI3-GUS expression is absent in dark-grown abi3 mutants
abi3 mutants also develop leaves in the dark. Leaf primordia are present in 10% of the abi3 seeds (Nambara et al. 1995). To investigate whether leaf development in abi3 mutants was initiated only at the completion of seed development or also after germination in the dark, leaf development of abi3 mutants was followed for 4 weeks. The homozygous and heterozygous/wild-type progeny of single, heterozygous abi3-3 (Col), abi3-6 (Col), abi3-4 (Ler), and abi3-5 (Ler) plants were assayed for leaf development (Table 1). The discrimination of the genotypes was based on the fact that homozygous abi3 seeds of these alleles were green, whereas heterozygous and wild-type seeds were brown. Leaf development within the first 2 weeks was exclusively established in the homozygous abi3 mutants (Table 1). At later time points (3 and 4 weeks), leaf development was present in more individuals among the homozygous abi3 plants, but it was no longer exclusive (Table 1). Taken together, abi3 mutants produced shoots earlier and in more individuals than wild-type seedlings. The abi3 mutants thus behave heterochronically during seedling development in the dark, suggesting a shorter QS1 (10% of the seeds have leaf primordia; Nambara et al. 1995) as well as a faster passage through seedling development.
Table 1. . Shoot development of abi3 homozygous and heterozygous/wild-type plants
In agreement with the latter hypothesis, abi3-6 plants containing the PABI3-GUS construct showed no GUS activity in the cotyledons after 7 d of germination in the dark (data not shown), whereas in all wild-type seedlings it was detected for 14 d after germination (Fig. 1a). The reduced GUS activity in dark-grown abi3 cotyledons was not due to a decreased expression of PABI3-GUS during seed development, as ripe abi3 seeds transgenic for PABI3-GUS contained 134% of the wild-type GUS activity (Rohde et al. in preparation). This result is in accordance with the observation that the ABI3 protein is more abundant in the abi3 mutant (Parcy et al. 1997). As in det1 mutants, no PABI3-GUS activity could be detected throughout development in the dark, suggesting the absence of both QS2 and QS3 (Figs 4a & d).
In conclusion, a genetically determined ability to develop leaves in the dark was found to be associated with a lack of PABI3-GUS expression. In contrast, a genetically determined inability to develop leaves in the dark was correlated with PABI3-GUS expression.
PABI3-GUS is additionally expressed in other vegetative organs
Because we observed that ABI3 expression was not restricted to seed development, but was also found in the apices of dark-grown seedlings, we extended the histochemical analysis of PABI3-GUS plants throughout development. Interestingly, GUS activity was detected in several parts of greenhouse-grown plants (Fig. 5): at the receptacle of flowers (Fig. 5a), in the axils of pedicels and of axillary flower bracts (Fig. 5b), in the abscission zones of the siliques and the rosette leaves of older plants (Figs 5c & d), and in the stipules of light- as well as dark-grown plants (Figs 5a, d & e). The expression of the ABI3 gene in the stipules was confirmed by whole-mount in situ hybridizations (Figs 5f & g). Similar patterns of GUS activity were observed in PABI3-GUS lines of both Landsberg erecta (Ler) and C24 ecotypes. Also a PABI3-GUS (C24) construct introgressed into the abi3-6, det1 and det2 mutants revealed a similar tissue-specific expression (data not shown).
The PABI3-GUS expression in the axils of flower bracts (Fig. 5b), although less systematically observed, might indicate another case of a transient QS. The development of accessory paraclades that will originate from such an axillary bud is delayed until the main inflorescence paraclade has achieved a certain development. The ABI3 expression in stipules and abscission zones will need further investigation, as no ABA effects have been integrated into current models on the creation or on the function of these organs (Oppenheimer et al. 1991; Larkin et al. 1993; González-Carranza, Lozoya-Gloria & Roberts 1998).
As depicted in Fig. 6, several hypothetical QSs are encountered during development of wild type in the dark, during which ABI3 is expressed (Figs 1a & d, & 6). Whereas ABI3 was expressed in all seeds during QS1 (Parcy et al. 1994), it was expressed only in a fraction of the seedlings during QS2 and QS3. There was a significant delay between the cessation or retardation of cell division and the appearance of PABI3-GUS expression. This delay is most clearly illustrated at QS2 in which cell division ceased after 4–6 d of germination in the dark, and PABI3-GUS expression was only observed after 10–12 d. Therefore, we hypothesize that ABI3 is not inducing a QS, but is rather involved in maintaining it or inhibiting/retarding further development.
Irrespective of the fact that the ABI3 expression occurred in only a fraction of individuals, it was always associated with non-growth or retarded growth in the dark and never with etiolated leaf development (Figs 1 & 4). This observation holds for mutants with distinct abilities to develop leaves in the dark (abi3, det1, det2) and for environmental factors promoting or retarding leaf development (sucrose, darkness, ABA). Together, these results suggest that ABI3 expression is associated with vegetative quiescence.
Quiescence provoked solely by darkness was associated with ABI3 expression in 15% of the individuals. The growth inhibition as well as the expression of ABI3 imposed by darkness were completely overcome when sucrose was locally supplied to the apex; leaves developed in nearly 100% of the plants and none of the seedlings expressed ABI3 at the apex. When quiescence was additionally controlled by externally supplied ABA, the number of individuals with ABI3 expression raised to 43%. The ABA and sucrose thus clearly had opposing effects on leaf development and ABI3 expression in the dark (Figs 2, 3 & 7). These observations demonstrate that the metabolic state of the apical meristem is very crucial for leaf development in the dark. Therefore, the discrete metabolic state of the apical meristem is probably also responsible for the quantitative expression of ABI3 during quiescence. The fact that apices of abi3 mutants were apparently as sensitive to ABA as the apices of wild-type seedlings further suggests that ABA also influences growth inhibition independently of ABI3 (Fig. 7).
On the other hand, the expression of 2S, a target gene of ABI3 in seeds, was found in the apices of all dark-grown seedlings, as demonstrated by a chimeric P2S-GUS construct. Therefore, other factors than ABI3 must have contributed to the expression of P2S-GUS. Most probably, ABI3 and any of the additional factors control only a part of the quiescence phenotype. The observation that overexpression of ABI3 did not lead to less leaf development in the dark supports this idea (A. Rohde, unpublished results). More ABI3 could thus not compensate for the other functions. Alternatively, ABI3 expression could also be transient, so that at a given time the expression is only observed in a percentage of the individuals.
Regardless of the observation that expression of ABI3 is always associated with non-growth and its absence associated with growth, it was impossible to prove directly the role of ABI3 in vegetative quiescence. Usually, mutants are the best candidates to reveal the function of the target gene. However, because ABI3 is expressed consecutively at different developmental stages (during seed maturation and at QS1, QS2, and QS3), it is not possible to demonstrate the role of ABI3 in QS2 and QS3 only, because the defects in abi3 mutants at seed maturation and at QS1 alter the further development of the plant. Indeed, the leaf primordia of abi3 mutants were shown to contain already partially developed chloroplasts, whereas leaf primordia that develop from dark-grown wild-type seedlings did not develop chloroplasts (Rohde et al. in preparation). In addition, GUS activity (derived from expression of PABI3-GUS) was metabolized in abi3 mutants already after 7 d of growth in the dark, whereas in wild-type seedlings this took 14 d. Again, this observation indicates an altered metabolism during early seedling development in abi3 mutants. Taking these considerations into account, the increased leaf development of abi3 mutants cannot be interpreted unambiguously as a phenotype that resulted solely from the partial loss of QS2 and QS3. The proof that ABI3 indeed plays a role in vegetative quiescence might come from the analysis of a conditional dominant-negative mutant. ABI3 could then be specifically knocked-out at the shoot apex during seedling development in the dark in order to reveal its in vivo function.
In conclusion, ABI3 is expressed besides seed development, namely in vegetative quiescence processes. We suggest that wild-type ABI3 confers retardation of the default developmental program under unfavourable conditions, such as darkness (Fig. 7).
Our model integrates the earlier observations of Nambara et al. (1995) that abi3 is a heterochronic mutant at QS1 and that wild-type ABI3 inhibits precocious seedling development. We extend this hypothesis by suggesting that ABI3 acts not only at the transition from seed development to germination, but also in conditions that do not demand or are disadvantageous for the continuation of development. ABI3 would thus integrate not only developmental triggers in the signalling cascade but also be able to adjust development in case of limiting environmental factors. Finally, the ABI3 expression (as well as that of P2S-GUS and PUSP-GUS) in seed and apex quiescence suggests that parallels in the regulation of both processes exist.
The seeds of the transgenic PABI3-GUS and P35S-ABI3 plants in both Ler and C24 ecotypes were kindly provided by Drs François Parcy and Jérôme Giraudat. Seeds of abi3-3 and abi3-6 were a generous gift of Dr Eiji Nambara. For the seeds of abi3-4 and abi3-5 Drs Karen Léon-Kloosterziel and Maarten Koornneef, and for the PUSP-GUS Dr Helmut Bäumlein are gratefully acknowledged. The authors would like to thank Bart Burggraeve for skilful technical assistance, Drs D. Van der Straeten, T. Gerats, M. Van Lijsebettens and M. May for useful comments on the manuscript, M. De Cock for preparing it, and R. Verbanck for artwork. A.R. is a research assistant of the Fund for Scientific Research – Flanders (Belgium).