Genetic approaches using Arabidopsis thaliana aimed at the identification of mutations affecting events involved in auxin signalling have usually led to the isolation of auxin-resistant mutants. From a selection screen specifically developed to isolate auxin-hypersensitive mutants, one mutant line was selected for its increased sensitivity to auxin (× 2 to 3) for the root elongation response. The genetic analysis of sax1 (hypersensitive to abscisic acid and auxin) indicated that the mutant phenotype segregates as a single recessive Mendelian locus, mapping to the lower arm of chromosome 1. Sax1 seedlings grown in vitro showed a short curled primary root and small, round, dark-green cotyledons. In the greenhouse, adult sax1 plants were characterized by a dwarf phenotype, delayed development and reduced fertility. Further physiological characterization of sax1 seedlings revealed that the most striking trait was a large increase (× 40) in ABA-sensitivity of root elongation and, to a lesser extent, of ABA-induced stomatal closure; in other respects, hypocotyl elongation was resistant to gibberellins and ethylene. These alterations in hormone sensitivity in sax1 plants co-segregated with the dwarf phenotype suggesting that processes involved in cell elongation are modified. Treatment of mutant seedlings with an exogenous brassinosteroid partially rescued a wild-type size, suggesting that brassinosteroid biosynthesis might be affected in sax1 plants. Wild-type sensitivities to ABA, auxin and gibberellins were also restored in sax1 plants by exogenous application of brassinosteroid, illustrating the pivotal importance of the BR-related SAX1 gene.
Cellular, molecular and genetic approaches have led to the cloning of a few genes which play a role in auxin action, but many of the events involved in auxin signalling pathways have still to be identified ( Walden & Lubenow 1996). Arabidopsis thaliana mutants affected in auxin responses have been selected on the basis of root phenotypes associated with auxin resistance. They also display alterations in plant morphology and elongation capacity (axr1, Estelle & Somerville 1987; axr2, Wilson et al. 1990 ; axr3, Leyser et al. 1996 ) and/or in gravitropism (axr2, Wilson et al. 1990 ; aux1, Timpte et al. 1995 ). In addition to auxin resistance of root elongation, all these mutations confer cross-resistance to other hormones: ethylene and cytokinins for axr1 ( Timpte et al. 1995 ), aux1 ( Pickett et al. 1990 ) and axr3 ( Leyser et al. 1996 ) or ethylene and abscisic acid for axr2 ( Wilson et al. 1990 ). These pleiotropic phenotypes of auxin-resistant mutants may illustrate the existence of cross-talk between various hormone signalling pathways. No selection procedure aimed at the screening for auxin hypersensitive mutants has been described thus far. However, the Arabidopsis mutant agr3 ( Bell & Maher 1990) that was initially selected for alterations in root gravitropism also showed an increased auxin sensitivity of root elongation ( Maher & Bell 1990). Auxin hypersensitivity has sometimes been postulated on the basis of morphological traits reminiscent of auxin action on plant development, for instance increased apical dominance in axr3 ( Leyser et al. 1996 ).
These data emphasize the necessity to search for new mutants exhibiting an increased sensitivity to auxin. Screening for such mutations, aside from broadening the spectrum of auxin response mutants, might allow the identification of negative regulators of auxin sensitivity. The present work describes the isolation of a new A. thaliana mutant, sax1 (hypersensitive to abscisic acid and auxin), originally selected for its increased sensitivity to auxins at the root level. The physiological characterization of sax1 revealed additional phenotypic traits, including an increased sensitivity to abscisic acid (ABA) of root elongation and stomatal aperture, a resistance of hypocotyl elongation to gibberellic acid (GA3) and to ethylene and a reduced size. Most of these phenotypic traits could be partially restored by application of exogenous brassinosteroids. Therefore, this mutant provides a new genetic tool for studying interactions between signalling networks involved in hormonal responses of vegetative tissues.
Mutant isolation and genetic mapping
We screened approximately 500 independent M2 families for their hypersensitivity to auxin. The screen was based on the search for plantlets exhibiting an exaggerated auxin-induced plant morphology in vitro, in the presence of an auxin concentration (0.1 μm 1-naphthaleneacetic acid, NAA), which does not significantly modify the wild-type phenotype ( Fig. 1a). Seedlings were isolated which exhibited all or some of the phenotypic traits characterizing a wild-type seedling grown on a higher NAA concentration (1 μm), i.e. a short and thick primary root, long and dense root hairs and a few secondary roots ( Fig. 1a). One mutant line was selected as a putative auxin-hypersensitive mutant on these criteria. Because subsequent physiological analysis demonstrated that this mutant is hypersensitive to abscisic acid and to auxin, the mutant was named sax1. We did not identify any other allele of sax1, either in the EMS M2 populations tested or in 4000 T2 T-DNA families screened from the T-DNA collection produced by Bechtold et al. (1993) .
As shown in Fig. 1(b), the sax1 mutant phenotype in the absence of auxin is marked enough to be used for performing segregation analyses, its main characteristics being a short primary root, the absence of secondary roots and epinastic cotyledons. The F1 plants resulting from back-crosses of sax1 plants to wild-type plants, irrespective of the female parent, all exhibited a wild-type phenotype. In the F2 progenies, the sax1 phenotype segregated in a 1:3 ratio, indicating that it is caused by a single recessive Mendelian allele. To map the mutation, homozygous sax1 plants (Columbia-0 ecotype, Col-0) were crossed to wild-type plants of the Landsberg ecotype and the F2 mutant progeny was scored for segregation of CAPS or SSLP markers between the two ecotypes. The sax1 mutation was located on the lower arm of chromosome 1, between the two markers GAPB and nga 280, at 15.5 ± 6.2 c m and 11.1 ± 4.8 c m, respectively. This region does not contain any published mutation conferring a Sax1– phenotype ( Lister & Dean 1993), and therefore no allelism tests were carried out between sax1 and other mutants.
The sax1 mutation has pleiotropic effects on plant morphology
Compared with wild-type seedlings grown for 9 days in vitro in the absence of exogenous auxin, sax1 seedlings grown in the same conditions have much shorter roots and hypocotyls and the cotyledons are dark-green, round, smaller and epinastic ( Fig. 1b). Moreover, young mutant plants exhibit a strongly reduced production of secondary roots and a root system with longer and denser root hairs. In sax1 plants, geotropism of root and hypocotyl is not affected. Under dark-growth conditions, sax1 seedlings display a phenotype closer to that of wild-type and, in particular, cotyledons are etiolated, the root has almost a normal length, and although the hypocotyl is still 50% shorter than the wild-type, it displays an approximately 10-fold increase in length in response to darkness as seen in the wild-type ( Fig. 1b).
When grown in the greenhouse, mutant plants show a strongly altered phenotype. Their rosette is small and composed of round leaves with short epinastic petioles ( Fig. 1c, Table 1). Adult sax1 plants show a very short primary stem as well as shorter secondary branches due to strongly reduced internode length (more than 10 times, see Table 1), which account for the dwarf stature of sax1 plants ( Fig. 1c). In flowering plants, sax1 shows fewer inflorescences and siliques, and fertility is reduced ( Table 1). The time-course of development up to the adult mutant plant is delayed by 4–5 days for most of the developmental events, including the occurrence of the first side branch, secondary branches and inflorescence and opening of the first flower. The largest delay (1 week) was observed for the senescence of flowers and siliques.
Table 1. Morphometric characteristics of wild-type and sax1 plants grown in the greenhouse
Seedlings were germinated and grown in the greenhouse for 6 weeks in a 16 h light/8 h dark cycle. Data represent mean ± standard error of 10 plants.
Rosette diameter (cm)
15.0 ± 0.4
4.5 ± 0.1
Height of primary stem (cm)
35.5 ± 0.7
3.6 ± 0.1
Number of side branches
3.6 ± 0.4
3.7 ± 0.3
Number of secondary branches
7.5 ± 0.3
3.4 ± 0.3
Number of inflorescences
47 ± 3
14 ± 5
Internode distance (cm)
3.1 ± 0.4
0.2 ± 0.02
(first and second ones)
Number of siliques/plant
220 ± 19
36 ± 4
Approximate number of seeds
66 ± 8
35 ± 5
per silique after 8 weeks
Root elongation in sax1 seedlings is hypersensitive to auxin
Root growth inhibition by auxins was analysed in wild-type and sax1 genotypes. The results of the different assays are expressed as a percentage of growth in the absence of hormone, allowing direct comparisons of the two genotypes despite the short root of sax1 and its slow rate of elongation. Routinely, the seeds were germinated directly on 1-naphthaleneacetic acid (NAA), 2,4-dichlorophenoxyacetic acid (2,4-D) and indole-3-acetic acid (IAA) and root length was scored after 7 days ( Fig. 2). For both genotypes, increasing auxin concentrations in the culture medium inhibited root growth up to 80–90% in the presence of 1 μm auxin. However, determination of the concentrations inducing a 50% inhibition indicate that sax1 roots are more sensitive to auxins by a factor ranging from 2 (for NAA and IAA) to 3 (for 2,4-D). In contrast, the roots of sax1 show wild-type sensitivities to two other hormones which inhibit root elongation, the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) and the cytokinins kinetin and benzyladenine (data not shown).
Root elongation and stomatal closure are hypersensitive to ABA in sax1 plants
The increased sensitivity of root elongation in sax1 is not specific to auxin. The sax1 mutation also confers an increased sensitivity to abscisic acid, as root growth was inhibited by ABA concentrations 40 times lower than those effective on wild-type roots ( Fig. 3a). This hypersensitivity to ABA co-segregated with both the hypersensitivity to auxins and the morphological phenotype in F3 mutant seedlings raised from the two successive back-crosses.
To test the ABA-induced closure of stomata, epidermal strips were incubated in the light for 2 h to fully open the stomata before adding ABA. The relative stomatal aperture is reported in Fig. 3(b) for the two genotypes as a function of ABA concentrations. Increasing concentrations of ABA decreased the aperture of wild-type and sax1 stomata. Although sax1 stomata displayed a higher variability in this response than stomata from wild-type plants, they appear more sensitive to ABA ( Fig. 3b). The stomatal density in the abaxial epidermis does not differ significantly between wild-type (297 ± 16 mm–2, n = 8 independent estimations) and sax1 leaves (326 ± 38 mm–2, n = 8 independent estimations). The mean length of guard cells is identical for wild-type and mutant stomata, 18.8 ± 0.4 μm (n = 2 independent experiments, 2 × 100 cells) and 17.6 ± 0.3 μm (n = 2 independent experiments, 2 × 100 cells), respectively. However, the stomatal pore width in the mutant is smaller than in the wild-type plants, both in the dark (0.94 ± 0.07 μm and 1.91 ± 0.16 μm, respectively, n = 5 independent experiments, 5 × 100 pores), as well as in the light (3.49 ± 0.15 μm and 5.13 ± 0.21 μm, respectively, n = 11 independent experiments, 11 × 100 pores). Thus, these properties cannot account for the ABA-hypersensitivity of the sax1 stomata.
In other respects, seed responses are not altered in sax1 plants. Germination of dry seeds from wild-type and sax1 plants exhibited the same sensitivity to exogenous ABA, with 50% inhibition occurring in the presence of 10 μm ABA ( Fig. 3c). The efficiency of germination of seeds from wild-type and sax1 plants is identical under post-imbibition conditions without cold treatment (50% germination after 60 h post-imbibition) and increases almost proportionately to the length of the chilling period ( Fig. 3d). Furthermore, the ABA-induced expression of the Rab18 and AtDi21 genes ( Gosti et al. 1995 ) is not modified in sax1 plants (data not shown).
Hypocotyl elongation of sax1 seedlings in the light is insensitive to gibberellins and ethylene
On light-grown seedlings, the elongation of wild-type hypocotyls was promoted by increasing concentrations of GA3 while the hypocotyl of sax1 plants was resistant to the hormone and almost did not elongate, with the highest GA3 concentration tested (10–4 M) stimulating the elongation by only 20% ( Fig. 4a). The dwarf phenotype of sax1 plants grown in the greenhouse could not be reversed by gibberellin treatments (1 m m GA3) although the treated plants displayed a slight increase of main stem growth and fertility, two responses characterizing GA3 action (data not shown).
When ACC (1-aminocyclopropane-1-carboxylic acid) was tested in light conditions (16 h light/8 h dark cycle), it stimulated hypocotyl elongation of wild-type plants when applied at concentrations higher than 10–7m ( Fig. 4b). This result, in agreement with data reported recently by Smalle et al. (1997) , was highly reproducible and was also obtained with ethylene (A. Barry and M. Hall, personal communication). In contrast, hypocotyl elongation of sax1 plants was affected neither by ACC, whatever the concentration tested ( Fig. 4b), nor by ethylene (data not shown), although sax1 efficiently converts ACC in ethylene (A. Barry and M. Hall, personal communication). These results indicate that hypocotyls of sax1 plants are insensitive to ethylene under light conditions although a normal triple response to ACC in dark conditions is observed (data not shown).
The dwarf phenotype of sax1 seedlings is rescued by 24-epibrassinolide
Brassinosteroids have been shown to play an essential role in plant development, especially in the control of cell elongation (reviewed in Clouse 1996). Wild-type and mutant plants were treated by the brassinosteroid 24-epibrassinolide (EBR). As shown in Fig. 5(a), the short phenotype of sax1 mutant plants grown in the light is partially rescued by addition of 10–9m EBR. The more obvious trait is the long root of sax1 seedlings after an EBR treatment, the aerial part not being significantly modified in these conditions.
A wide range of 24-epibrassinolide (EBR) concentrations were tested under light conditions on root and hypocotyl elongation of wild-type and mutant genotypes. The dose–response curves show that root elongation in wild-type plants was stimulated by at most 20% in the presence of 3 × 10–10m EBR and higher concentrations are inhibitory ( Fig. 5b). Application to sax1 seedlings of EBR concentrations lower than 10–9m induced a large increase in root length, reaching a maximal stimulation of 100% at 10–9 and 3 × 10–9m, whereas higher EBR concentrations inhibit root elongation ( Fig. 5b). EBR (3 × 10–7m) also stimulated hypocotyl elongation of wild-type and mutant seedlings up to 100% and 200%, respectively ( Fig. 5c), leading to a complete restoration of hypocotyl length in sax1 plants. In both genotypes, the hypocotyl is less sensitive to EBR than the root by two orders of magnitude. Table 2 shows morphometric characteristics of root and hypocotyl of in vitro wild-type and sax1 plants cultivated in the light, in the absence or presence of EBR concentrations that were the most efficient on roots (3 × 10–9m) and hypocotyls (3 × 10–7m) ( Fig. 5b,c). Mutant primary roots are approximately 60% shorter than those of the wild-type, whereas their epidermal cell lengths are identical ( Table 2) which suggests that mutant roots have a reduced number of cells. The addition of 3 × 10–9m EBR in the medium has no significant effect on wild-type seedlings whereas a large increase in sax1 root length is observed, with the size of epidermal cells being unchanged in the mutant root ( Table 2). In the case of hypocotyls, the epidermal cells are half the length of wild-type, which fits well with the reduced hypocotyl length. For the two genotypes, the addition of 3 × 10–7m EBR in the medium increases hypocotyl and cell lengths. Under those conditions, the values reached in sax1 are identical to those of wild-type plants ( Table 2).
Table 2. Morphometric characteristics of wild-type and sax1 plants grown in vitro in the absence (control) or presence of 24-epibrassinolide (+ EBR)
Seedlings were germinated and grown on ABIS medium for 7 days in a 16 h light/8 h dark cycle. Data represent mean (± SE) of three plants (for organ measurement, n = 3) and 30–40 cells from each of the three plants (for cell measurements, n = 90–120). A second independent experiment performed in the same conditions gave similar results.
Length (mm) control (+ 3 × 10–9m EBR)
Epidermal cell length (μm) control (+ 3 × 10–9m EBR)
21.3 ± 4.8
22.7 ± 2.3
98.7 ± 6.5
94.9 ± 5.1
8.7 ± 0.2
16.7 ± 0.4
96.7 ± 2.5
96.7 ± 8.9
Length (mm) control (+ 3 × 10–7m EBR)
Epidermal cell length (μm) control (+ 3 × 10–7m EBR)
1.9 ± 0.3
3.0 ± 0.7
82.4 ± 5.1
124.4 ± 19.1
0.9 ± 0.3
2.2 ± 0.5
63.6 ± 7.6
135.1 ± 18.1
Under dark-growth conditions, EBR concentrations ranging from 10–9 to 3 × 10–7m display an increasing inhibition of root elongation for both genotypes ( Fig. 5d). Hypocotyl elongation in sax1 is significantly stimulated by 10–8m EBR, but higher concentrations inhibit hypocotyl growth in both genotypes ( Fig. 5e).
Treatment with EBR abolishes the hypersensitivity of sax1 root elongation to ABA and auxin and restores the capacity of the sax1 hypocotyl to respond to GA but not to ACC
Apart from its dwarf stature, sax1 displays alterations in the responsiveness to hormonal regulation of cell elongation, both in the root and in the hypocotyl. The effects of EBR treatment on these phenotypic traits were evaluated. As to the root elongation, the addition of 10–9m EBR, the most efficient concentration to promote root elongation, to the media supplemented with different ABA ( Fig. 6a) or IAA ( Fig. 6b) concentrations does not affect the dose–response curves of wild-type roots but totally suppresses the shifts in sensitivity to ABA or auxin observed in sax1 root.
The sax1 hypocotyl length was unaffected by GA3 and ACC treatments ( Fig. 6c,d, respectively). When 3 × 10–7m EBR, the most efficient concentration on hypocotyl, was added to increasing GA3 concentrations in the culture medium, both genotypes displayed a high stimulation of hypocotyl elongation in wild-type plants and also clearly in sax1 plants ( Fig. 6c). This indicates that sax1 hypocotyl has recovered the capacity to respond to gibberellin in a dose-dependent manner in the presence of EBR. The addition of EBR (3 × 10–7m) to ACC-containing media inhibits ethylene stimulatory action in wild-type seedlings and therefore does not reveal a responsiveness of hypocotyl elongation in sax1 seedlings ( Fig. 6d).
We have isolated a new recessive nuclear mutation, sax1, which maps to a single locus on chromosome 1. The sax1 mutation has pleiotropic effects on the morphology of seedlings grown in vitro, as well as of plants in the greenhouse, the more striking trait being pronounced dwarfism. The sax1 mutant is also characterized by an increased sensitivity of root elongation to abscisic acid and, to a lesser extent, to auxin and by a resistance of hypocotyl growth to gibberellins and to ethylene in light conditions. All these phenotypic traits co-segregate which suggests that they originate from a single mutation.
Several dwarf mutants affected either in gibberellin biosynthesis or in response to gibberellins have been described previously ( Koornneef & van der Veen 1980; Koornneef et al. 1985 ). On the basis of the mapping, sax1 lies at a different locus. The defect in hypocotyl elongation exhibited by sax1 seedlings grown in vitro in the presence of GA3 would probably be due to a lack of responsiveness to gibberellins restricted to hypocotyl as sax1 plants in the greenhouse showed a partial response to sprayed GA3. The hypocotyl of sax1 seedlings is also affected in its sensitivity to ethylene, but only in light-growth conditions as sax1 is normally responsive to ethylene in the dark. These results suggest that the sax1 mutation does not lead to a general insensitivity to these growth regulators.
Interestingly, with regard to previously described auxin response mutants, sax1 seedlings show hypersensitivity to auxins, with root elongation being inhibited by auxin concentrations 2–3 times lower than those affecting wild-type roots. This is probably due to a shift in auxin sensitivity, as preliminary results showed no significant difference between sax1 and wild-type plants in the auxin endogenous content (R. Maldiney, personal communication) or in the accumulation and metabolism of exogenous auxin (data not shown). Up until now, only two A. thaliana mutants have been postulated as auxin hypersensitive, agr3 ( Maher & Bell 1990) and axr3 ( Leyser et al. 1996 ), but the possibility that the mutations affect a common locus can be ruled out as the corresponding genes are mapped either on chromosome 1 but far from sax1 in the case of axr3 ( Leyser et al. 1996 ) or on chromosome 5 for agr3 ( Sinclair et al. 1996 ).
The sax1 mutant displays a highly significant increase in ABA sensitivity for several vegetative responses. From preliminary assays for endogenous ABA content, no significant difference was found between sax1 and wild-type plants (R. Maldiney, personal communication), suggesting that ABA signalling pathways rather than biosynthesis are altered by the sax1 mutation. Up until now, the era mutants are the only ones selected in A. thaliana as being hypersensitive to ABA ( Cutler et al. 1996 ). The era1 mutation has not been mapped, but the corresponding gene has been cloned. ERA1 encodes the β-subunit of a protein farnesyl transferase ( Cutler et al. 1996 ). The recessive nature of the mutation suggests that ERA1 could act as a negative regulator of ABA signals. Except for an increase in seed dormancy, the physiological characterisation of the era1 mutant is rather limited. However, it seems that the morphologies of sax1 and era1 mutants are distinct in seedlings grown in vitro and in the adult plant in the greenhouse, as the authors mentioned that era1 plants exhibit relatively normal growth and development ( Cutler et al. 1996 ). In addition, sax1 in contrast to era1 does not exhibit an increased seed dormancy, making it unlikely that the SAX1 and ERA1 genes are identical. However, a genetic complementation test is needed to rule out the two mutations being allelic. The fact that the sax1 mutation affects ABA responses characterizing vegetative tissues, namely root elongation and stomatal closure, raises the possibility that sax1 could be affected in the same pathway as abi1 and/or abi2 mutants (reviewed in Merlot & Giraudat 1997). However, the ABA responses involved in germination, dormancy and in Rab18 and AtDi21 gene transcription are not affected by the sax1 mutation whereas they are strongly altered in abi1 and abi2 mutants ( Gosti et al. 1995 ; Koornneef et al. 1984 ; Leung et al. 1997 ; Roelfsema & Prins 1995). This suggests either that SAX1 lies in an ABA signalling pathway distinct from ABI1 and ABI2 cascades or, more likely, that sax1 is not primarily defective in an ABA-response component within the primary ABA transduction cascade.
The hypocotyl and root are severely shortened in sax1 plants compared to wild-type ones, thus accounting for the dwarf phenotype. This dwarf phenotype can be rescued by growing sax1 plants on EBR, suggesting that sax1 could be defective in brassinosteroid biosynthesis. The treatment of sax1 seedlings growing in light conditions by increasing exogenous EBR concentrations partially restores root elongation (100% stimulation) and totally rescues hypocotyl elongation (200% stimulation). Further characterization at the histological level revealed that the sax1 mutant has epidermal cells which are only half the length of those of wild-type due to defects in hypocotyl elongation. In the case of the retarded root elongation, reduction of cell division has to be postulated as the cause as epidermal cells have a normal length. Consequently, application of exogenous brassinosteroids to sax1 plants has two different effects – activation of cell division in the root and stimulation of cell elongation in the hypocotyl. The promotion of cell division has been observed in the upper part of brassinosteroid-treated internodes of bean plants ( Worley & Mitchell 1971) and was described recently in isolated leaf protoplasts of Petunia hybrida ( Clouse 1998), although such an effect does not seem to be general to all plants. For example, brassinosteroid failed to activate cell division in cultured carrot cells ( Sala & Sala 1985). The efficiency of EBR application on seedlings growing in dark conditions is low; however, the elongation of sax1 hypocotyl can be increased up to 50% by EBR with no further increase by the natural brassinolide (data not shown). This may result from the fact that the elongation response of sax1 plants is not strongly altered in the dark as etiolated hypocotyls elongate in the same proportion (×10) in the mutant and wild-type plants compared to light-grown seedlings. Studies on dwarf mutants of A. thaliana have shed light on the role of brassinosteroids in the control of elongation processes ( Clouse 1996; Ecker 1997). Strong similarities in terms of morphology exist between sax1 plants and a series of mutants affected in biosynthesis or signalling of brassinosteroids (reviewed in Altmann 1998). The det2 locus is located on chromosome 2 ( Li et al. 1996 ), cpd mutation on chromosome 5 ( Szekeres et al. 1996 ), the locus of the three alleles, cbb1, dim and dwf1, maps on chromosome 3 ( Kauschmann et al. 1996 ) and the mutation bri1 on chromosome 4 ( Clouse et al. 1996 ). The dwf4 mutation was mapped very recently on chromosome 3 ( Azpiroz et al. 1998 ). The location of the sax1 mutation on chromosome 1 clearly eliminates the possibility that sax1 is allelic to one of the known brassinosteroid mutants.
Other remarkable characteristics of sax1 plants are the hypersensitivity of roots to ABA and auxin and the insensitivity of hypocotyls to GA3 and ACC. These phenotypic traits co-segregate with an abnormal growth development of seedlings grown in vitro, as well as in adult plants in the greenhouse. Except for ACC insensitivity, all the hormonal phenotypes were reversed by EBR addition to the growth medium which shifted the dose–response curves to a wild-type sensitivity to IAA and ABA for root elongation and to GA3 for hypocotyl growth. Clearly, the addition of brassinosteroid to ACC-media does not restore a wild-type sensitivity to ethylene in sax1 seedlings, and abolishes the response of wild-type seedlings to ACC in the light. This effect on wild-type plants, which reveals the existence of interactions between brassinosteroid and ethylene response pathways in the light, has not been investigated further. These data demonstrate that the primary defect in sax1 concerns the synthesis of brassinosteroids, and that most of the other phenotypic traits derive from this primary alteration. Very little is known about the sensitivity of brassinosteroid mutants to other hormones. The bri1 mutant, a mutant affected in brassinosteroid perception, shows an increased response to ABA compared to wild-type plants ( Clouse et al. 1996 ). Among the brassinosteroid biosynthesis mutants, Kauschmann et al. (1996) reported that the cbb1 and cbb3 mutants react as wild-type seedlings to auxin, cytokinin, gibberellins and an ethylene releasing compound (ethrel). Few data are available concerning the interactions of brassinosteroids with other hormones, as reviewed by Mandava (1988). Basically, in many bioassays, brassinosteroids interact strongly with auxins (possibly synergistically). The responses to brassinosteroids and GA appear to be independent and additive, whereas ABA interacts strongly with brassinosteroids and reverses their growth effects. Recently, from the analysis of the dwf4 phenotype ( Azpiroz et al. 1998 ) indirect evidence has been presented that a fully active brassinosteroid pathway is necessary for a full response to GA, auxin and darkness.
We provide direct evidence here for the existence of interactions between brassinosteroids and auxin, ABA, GA and ethylene signalling pathways. The fact that brassinosteroids restore a wild-type sensitivity at least to auxin, ABA and GA3 in sax1 seedlings strongly demonstrates that three different signalling pathways controlling cell elongation are under the control of a fourth pathway involving brassinosteroids. The absence of GA response in cells defective for elongation might just mean that brassinosteroid is the main limiting factor of elongation and that GA cannot compensate for this. On the contrary, hypersensitivity to ABA and auxin is convincing evidence that brassinosteroid must at some point antagonize ABA and auxin responses. In that case, the target of the brassinosteroids might be a common signalling component to ABA and auxin pathways, or will lie downstream of the signalling pathways in a step involved in the elongation process per se. The sax1 mutation represents an original tool for identifying common targets to hormone signalling pathways involved in the control of cell elongation.
Plant material and growth conditions
Seeds from Arabidopsis thaliana (L.) Heyn, Columbia-0 ecotype (Col-0) were surface sterilized for 1 h in sodium hypochlorite (0.9° active Cl), rinsed six times with sterile water and stored overnight at 4°C. They were then plated on a medium containing 5 m m KNO3, 2.5 m m K2HPO4/KH2PO4, pH 6, 2 m m MgSO4, 1 m m Ca(NO3)2, 1 m m MES, 50 μm FeEDTA, MS microelements ( Murashige & Skoog 1962), 10 g l–1 sucrose and 7 g l–1 bacto-agar. Seedlings were grown on vertically placed Petri dishes in the growth chamber. Culture conditions were 21°C, with a 16 h day length at lighting levels of 120 μmol m–2 sec–1 photosynthetically available radiation (PAR), using neon tubes (in combination from Mazdafluor Blanc Industrie and Mazda fluor Prestiflux, Mazda Eclairage, France). For dark conditions, plates were wrapped in four layers of aluminium foil.
When plants were grown to maturity, seeds were either directly sown into soil in the greenhouse or plantlets were transferred from in vitro culture to soil. Greenhouse conditions were as follows: the photoperiod 16 h natural light with possible complementation (sodium lamp SON/T AGRO 400, Philipps, France) when light irradiance was lower than 220–260 μmol m–2 sec–1 and the thermoperiod was 23°/18°C (day/night). Irrigation was undertaken three times per week with nutrient solution (chemical fertilizer Solu-plant, 2 g l–1, Duclos International), and air humidity was regulated to approximately 55%.
Mutant selection and genetic analysis
Approximately 500 M2 lines generated following EMS mutagenesis were kindly provided by Prof. G. Belliard (Institut des Sciences Végétales, Gif sur Yvette and University of Paris XI, Orsay, France). The EMS treatment (0.3% EMS, 16 h) resulted in the segregation of chlorophyll mutations in approximately 2–5% of the M2 families. Fifty seeds from each individual M2 family were sown on the medium described above supplemented with 0.1 μm 1-naphthaleneacetic acid (NAA).
Selected sax1 plants were transferred to the greenhouse, self-fertilized and the transmission of the phenotype was confirmed in the M3 generation. Plants from the M3 generation were used as the female parent in crosses; back-crossing was repeated two successive times with wild-type plants from the Columbia ecotype. One hundred per cent of the F1 plants resulting from these crosses (359 plants analyzed) exhibited a wild-type phenotype. The F2 progenies of three populations were analyzed; the sax1 phenotype segregated in a 1 (mutant): 3 (wild-type) ratio in agreement with the Chi-square statistical test (ratios were 128:367, 95:299 and 99:293). The F2 segregating population of sax1 generated after two back-crosses was further propagated to yield F4 seeds which were used for all the experiments described.
For mapping, 33 seedlings homozygous for the sax1 mutation were selected in an F2 population issued from the cross of sax1 (Col-0 background) and wild-type plants from the ecotype Landsberg erecta (Ler). Genomic DNA for PCR was prepared from leaves using the method described by Konieczny & Ausubel (1993). Co-dominant cleaved amplified polymorphic sequence markers (CAPS) were used as described by Konieczny & Ausubel (1993) and simple sequence-length polymorphism markers (SSLP) as described by Bell & Ecker (1994). The CAPS and SSLP primer pairs were purchased from Research Genetics, Inc. (Huntsville, AL, USA). The map positions of the CAPS and SSLP markers were from the Arabidopsis map generated through the Landsberg/Columbia recombinant inbred lines by Lister & Dean (1993), released June 29 1995. The map distances relative to the CAPS and SSLP markers in centimorgans were calculated according to Kosambi (1944).
Root growth inhibition by different hormones was carried out in two ways. Seedlings were germinated and grown either on hormone-supplemented media for 7 days or on plates containing minimal medium for 4 days and then transferred for an additional 2 days onto a minimal medium supplemented with various concentrations of a given hormone (root elongation on ABA).
Routinely, lengths of primary root of 30 seedlings were measured with a graduated ruler (± 0.5 mm) under a binocular microscope. In transfer experiments, the elongated part of the root was measured by the same method 2 days after transfer. Standard errors calculated for the mean of 30 measurements per condition never exceeded 6% of the absolute value. Inhibition of elongation in the presence of an effector was calculated relative to elongation on minimal medium and expressed as a percentage for a comparison between wild-type and mutant genotypes.
To test the effect of 24-epibrassinolide (EBR), stimulation and/or inhibition of growth was assayed by measuring the length of the root and/or hypocotyl of wild-type and mutant 7-day-old seedlings germinated and grown directly in the presence of various concentrations of hormone (GA, ACC) or/and EBR. When ABA or IAA were tested together with EBR, 5-day-old seedlings grown on minimal medium were transferred onto the supplemented media for 2 days. The brassinosteroid EBR was solubilized in ethanol; the absence of ethanol effects on root and hypocotyl growth has been checked.
For whole mount light microscopy, 7-day-old seedlings grown in the light were stained with methyl-blue dye (0.001% (W/v) water solution. Determination of epidermal cell dimensions was performed directly under a microscope using an ocular micrometre (Leitz, Dialux 22). Measurements were performed on cells located in the middle part of the hypocotyl which is representative of the elongation of any epidermal cell throughout the entire growth phase of the hypocotyl ( Gendreau et al. 1997 ). In roots, epidermal cells were measured in the differentiation zone where cells are fully expanded.
Germination and dormancy assays
For germination assays, mature seeds were plated on culture medium (see Growth conditions) supplemented with various ABA concentrations. After 4 days at 4°C in darkness to break seed dormancy, plates were transferred at 21°C upon a photoperiod 16 h light/8 h dark. Germination of seeds was scored as positive when the radicle tip had fully penetrated the seed coat. The percentage of germination was determined by dividing the number of seeds that germinated at a given time by the total number of plated seeds × 100.
For dormancy studies, plants of the wild-type and mutant genotypes were grown at the same time in the greenhouse. Freshly harvested seeds were kept drying for 1 week before a chilling treatment for 0, 4, 8 and 10 days at 4°C in darkness. Germination of seeds was scored as described above.
Measurements of stomatal aperture
Plants in pots containing coarse sand were grown in a growth chamber. The temperature was 21°C during the light period (330 μmol m–2 sec–1) and 19°C during the dark period; the relative humidity was constant at 70%. To study wild-type and mutant plants at a similar stage of development, mutant seeds were sown about 10 days earlier. Measurements of stomatal aperture were carried out on epidermal strips isolated from leaves of 5- to 6-week-old plants. Methods for epidermis peeling and measurements of stomatal aperture were derived from the ones applied to Commelina communis ( Vavasseur et al. 1995 ). For each treatment, 10 epidermal strips were floated on the incubation solution (60 m m KCl, 10 m m MES, pH 6.15), first in the absence of hormone for 2 h under light (450 μmol m–2 sec–1), then in the presence of different ABA concentrations for 3 h.
The authors are grateful to Jean Guern for his encouragement and support of the work and to Jerôme Giraudat (CNRS, Gif sur Yvette) for helpful discussions and critical reading of the manuscript. We also thank Dick Kendrick and Silvère Pagant (Institute of Physical and Chemical Research, RIKEN, Japan) for a critical review of this manuscript. We thank Geneviève Belliard (CNRS, Gif sur Yvette and University of Paris 11, Orsay) and Michel Caboche and Georges Pelletier (INRA, Versailles) for providing Arabidopsis seeds from EMS mutant library and T-DNA mutant collection, respectively. We are indebted to Catherine Bellini and Herman Höfte (INRA, Versailles) for supplying 24-epibrassinolide for preliminary experiments. We wish to thank Rémy Drouen for technical assistance with plant culture. We are indebted to Régis Maldiney and Emile Miginiac (University of Paris 6, Paris) for measurements of endogenous hormone contents, and Juliette Leymarie and Alain Vavasseur (CEN Cadarache, Saint Paul lez Durance) for teaching and helping in stomata studies. C.V. was supported by a grant from the European Union Human Capital and Mobility Network (ERBCHBGCT 930488). This work was funded in part by the BIOTECH program of the European Community as part of the Project of Technical Priority 1993–1996.