The ratio of campesterol to sitosterol that modulates growth in Arabidopsis is controlled by STEROL METHYLTRANSFERASE 2;1

Authors

  • Aurélie Schaeffer,

    1. Institut de Biologie Moléculaire des Plantes du CNRS, Département Biosynthèse et Fonctions des Isoprénoïdes, Institut de Botanique, 28 rue Goethe, 67083 Strasbourg, France
    Search for more papers by this author
  • Roberte Bronner,

    1. Institut de Biologie Moléculaire des Plantes du CNRS, Département Biosynthèse et Fonctions des Isoprénoïdes, Institut de Botanique, 28 rue Goethe, 67083 Strasbourg, France
    Search for more papers by this author
  • Pierre Benveniste,

    1. Institut de Biologie Moléculaire des Plantes du CNRS, Département Biosynthèse et Fonctions des Isoprénoïdes, Institut de Botanique, 28 rue Goethe, 67083 Strasbourg, France
    Search for more papers by this author
  • Hubert Schaller

    Corresponding author
      For correspondence (fax +33 3 88 35 84 84; e-mail hubert.schaller@ibmp-ulp.u-strasbg.fr).
    Search for more papers by this author

For correspondence (fax +33 3 88 35 84 84; e-mail hubert.schaller@ibmp-ulp.u-strasbg.fr).

Summary

The Arabidopsis genome contains three distinct genes encoding sterol-C24-methyltransferases (SMTs) involved in sterol biosynthesis. The expression of one of them, STEROL METHYLTRANSFERASE 2;1, was modulated in 35S::SMT2;1 Arabidopsis in order to study its physiological function. Plants overexpressing the transgene accumulate sitosterol, a 24-ethylsterol which is thought to be the typical plant membrane reinforcer, at the expense of campesterol. These plants displayed a reduced stature and growth that could be restored by brassinosteroid treatment. Plants showing co-suppression of SMT2;1 were characterized by a predominant 24-methylsterol biosynthetic pathway leading to a high campesterol content and a depletion in sitosterol. Pleiotropic effects on development such as reduced growth, increased branching, and low fertility of high-campesterol plants were not modified by exogenous brassinosteroids, indicating specific sterol requirements to promote normal development. Thus SMT2;1 has a crucial role in balancing the ratio of campesterol to sitosterol in order to fit both growth requirements and membrane integrity.

Introduction

The plasma membrane of eukaryotic cells is made of a phospholipid bilayer which is reinforced by sterols (Bloch, 1983; Demel and De Kruyff, 1976). Cholesterol in the animal kingdom and ergosterol in the vast majority of fungi are single-pathway end products that fulfil this structural role. In comparison, the plant kingdom produces sterols which are diverse (Nes and McKean, 1977).

Higher plants contain a mixture of sterols differing essentially in the number of carbon atoms at position C24 on the side chain of the molecule. Most plant species, including Arabidopsis thaliana (L.) Heynh., present a sterol profile consisting of cholesterol, a 24-desmethyl sterol; campesterol, a 24-methyl sterol; and sitosterol and stigmasterol, two 24-ethyl sterols in which campesterol and sitosterol are predominant. Formation of 24-alkyl sterols in plants proceeds through the addition of exocyclic carbon atoms provided by S-adenosyl methionine (AdoMet) and catalysed by AdoMet-Sterol-C24-methyl transferases (SMTs) during the course of the enzymatic conversion of cycloartenol, the product of the cyclization of 2,3-oxidosqualene, into pathway end products as shown in Figure 1. Sterol-C24-methylation enzymatic reactions have been studied extensively in plants (Bladocha and Benveniste, 1985; Janssen and Nes, 1992; Malhotra and Nes, 1971; Rahier et al., 1984; Rendell et al., 1986; Wojciechowski et al., 1973) and in yeast (Gaber et al., 1989; Nes et al., 1998).

Figure 1.

Simplified biosynthetic pathway for plant sterols.

Schematic diagram of the biosynthetic pathways of cycloartenol to 24-methyl and 24-ethyl sterols. Dashed arrows represent more than one biosynthetic step not indicated here. The 24-methyl biosynthetic segment of the plant sterol pathway produces 24-methylcholesterol as a mixture of campesterol [(24-R)-epimer] and dihydrobrassicasterol [(24-S)-epimer] (Rendell et al., 1986), the proportions of which are not indicated here.

Recently, cDNAs encoding SMTs from an array of higher plants have been identified on the basis of sequence identity with the yeast gene ERG6 (Gaber et al., 1989) encoding a zymosterol-C24-methyltransferase. Gene products were then characterized by heterologous expression in the mutant erg6 or in bacteria (Bouvier-Navéet al., 1997; Bouvier-Navéet al., 1998; Grebenok et al., 1997; Husselstein et al., 1996; Shi et al., 1996). In the case of tobacco, it has been shown that expression of two distinct SMTs in the erg6 mutant results in enzyme activities corresponding to distinct steps in the plant sterol biosynthesis pathway. Indeed, an acetone powder from erg6-NtSMT1 specifically catalyses the C24-methylation of cycloartenol; and an acetone powder from erg6-NtSMT2 preferentially catalyses the C241-methylation of 24-methylene lophenol (Bouvier-Navéet al., 1998) (Figure 1). The occurrence of two SMTs involved in the first and second methylation steps was further illustrated in planta: expression in tobacco of SMT1 or SMT2 led to different sterol profiles of the transgenic plants (Schaeffer et al., 2000). Indeed, expression of a 35S::SMT1;1 tobacco transgene results in the depletion or accumulation of cycloartenol in the case of overexpression or co-suppression, respectively, of the SMT1;1 message. Likewise, overexpression of a 35S::SMT2;1 Arabidopsis transgene results in a dramatic increase of sitosterol due to elevated 24-methylene-lophenol-C241-methyltransferase activity.

The multiple sequence alignment presented in Figure 2 clearly shows the distribution of the three Arabidopsis SMTs (Bouvier-Navéet al., 1997; Diener et al., 2000) into an SMT1 family (one Arabidopsis gene) or an SMT2 family (two Arabidopsis genes) according to their identity to each other; all members of a given family share at least ≈70% identity and show not more than ≈45% identity with members from the other family (Schaeffer et al., 2000).

Figure 2.

Sequence alignment of tobacco and Arabidopsis sterol methyltransferases.

Alignment was performed using pileup in the Genetics Computer Group package run with the default parameters. Positions with a consensus residue are boxed. GenBank accession numbers are X89867(AtSMT2;1); U71400(AtSMT2;2); AF090372(AtSMT1, A. Schaeffer, P. Benveniste and H. Schaller, unpublished); U71400(NtSMT2;1); and U81312(NtSMT1;1). At, Arabidopsis thaliana; Nt, Nicotiana tabacum. Data from the Arabidopsis Genome Initiative revealed the sequences of three SMT genes: AtSMT1, BAC MSH12Genbank accession number AB 006704; AtSMT2;1, Genbank accession number ACO26234; AtSMT2;2, Genbank accession number Z83321. SMT2;1 and SMT2;2 polypeptides share 81% identity, and their respective expression in the yeast erg6 (smt null mutant) led to identical sterol profiles of the transformed yeast cells (Bouvier-Navéet al., 1997), therefore we propose the above nomenclature rather than SMT2 and SMT3 (Diener et al., 2000).

SMTs generate a complex 24-alkyl-Δ5-sterol composition which is genetically defined in higher plants. The physiological significance of these defined proportions of 24-desmethyl, 24-methyl and 24-ethyl sterols, as well as the regulation of their respective concentrations, is not yet well documented. The pivotal role of campesterol as a precursor of the plant-growth regulators called brassinosteroids has been established through biochemical approaches (for review see Yokota, 1997) and molecular characterization of a set of dwarf mutants blocked in the steroid biosynthesis pathway (for review see Schumacher and Chory, 2000). All these dwarf mutants share phenotypic traits: extremely short stature due to a lack of cell elongation; and reduced fertility. Biochemical complementation of such biosynthetic mutants was achieved by treatment of seedlings in vitro with micromolar concentrations of various brassinosteroids. This was verified in the case of the mutants impaired in the conversion of campesterol into brassinolide, namely the Arabidopsis mutants det2 (Fujioka et al., 1997; Li et al., 1996); cpd/cbb3 (Kauschmann et al., 1996; Szekeres et al., 1996); dwarf 4 (Choe et al., 1998); and the tomato mutant dwarf (Bishop et al., 1999). The biochemical complementation of dwarfism with brassinosteroids was also demonstrated in the case of mutants impaired in the conversion of cycloartenol into campesterol and sitosterol, namely the Arabidopsis mutants diminuto (Klahre et al., 1998); ste1/dwarf7 (Choe et al., 1999; Gachotte et al., 1995; Husselstein et al., 1999); dwarf5 (Choe et al., 2000); and the garden pea mutant lkb (Nomura et al., 1997). Dim and lkb accumulate Δ5,24-sterols (24-methylene cholesterol, isofucosterol) at the expense of Δ5-sterols (campesterol, sitosterol) due to a blockage of the C24-isomerization/reduction step (Figure 1). Ste1 and dwarf7 accumulate Δ7-sterols (Δ7-campesterol, Δ7-sitosterol) at the expense of Δ5-sterols due to a blockage of the C5(6)-desaturation step (Figure 1). Dwarf5 accumulates 7-dehydrocampesterol and 7-dehydrocampestanol at the expense of campesterol. Although these mutants have been described as primarily sterol mutants because of the steps that they correspond to in the sterol biosynthesis pathway, the major consequence of all these mutations is to deplete the plant cell of sufficient amounts of campesterol for brassinosteroid synthesis. In the wild type, the amount of campesterol produced is controlled by the enzyme SMT2 which controls the key branching point of the pathway leading from cycloartenol to 24-alkyl-Δ5-sterols (Figure 1). Indeed, the orientation of the sterol biosynthetic flux towards campesterol or sitosterol is achieved via SMT2 enzyme activity in plants (Schaller et al., 1998). Here we show that the modulation of SMT2;1 gene expression in transgenic Arabidopsis leads to a selective modification of the sterol profile. The ratio of campesterol to sitosterol is dramatically modified and is associated with growth and developmental modifications.

Results

Modulation of the expression of AtSMT2;1 in transgenic Arabidopsis generates novel 24-alkyl-Δ5-sterol profiles

The biosynthetic flux from cycloartenol to the 24-methyl sterol or to the 24-ethyl sterol segments of the pathway was challenged in Arabidopsis transformed with the cDNA encoding AtSMT2;1. Screening a set of 20 primary transformants led to the identification of five valuable lines based on preliminary sterol analyses (data not shown). The phenotypic characterization of these five lines was carried out with greenhouse-grown or in vitro plants from the T2 to the T4 generation. Northern blot analysis revealed that plants from this set of 35S::SMT2;1 transgenic lines may be classed in two groups. SMT2;1 transcripts accumulated in 30 plants out of the 35 analysed (seven plants per T2 family), whereas they were nearly undetectable in the remaining five plants when compared with the level of the SMT2;1 message in the wild type or in a transgenic control line harbouring only the void expression vector. A representative Northern blot analysis is shown in Figure 3. From this experiment we conclude that the collection of 35S::SMT2;1 plants contains individuals overexpressing the transgene and others co-suppressing SMT2;1. In the latter, the level of SMT2;1 transcripts is strongly attenuated owing to a gene-silencing phenomenon not investigated here and reviewed by Depicker and Van Montagu (1997).

Figure 3.

Northern blot analysis.

The expression level of SMT2;1 in wild type, transgenic control and 35S::SMT2;1 Arabidopsis grown in vitro on GM medium for 5 weeks was established using the corresponding DNA 32P-radiolabelled probes. Hybridization of the blot with an 18S RNA 32P-radiolabelled probe indicates that equivalent amounts of total RNA are loaded in each lane of the gel.

Sterol profiles of 35S::SMT2;1 lines characterized for the expression of SMT2;1 were accurately determined and are shown in Table 1. This metabolic analysis led to the distribution of transgenic plants into two groups, as did the Northern blot analysis.

Table 1.  Sterol profile a of 35S::SMT2;1 Arabidopsis
SterolWild
type
Transgenic
control
hs1hs2hc1hc2
  • a Plants were harvested after 5 weeks on GM containing 50 mm kanamycin.

  • b

    Percentage of total.

  • c

    Not detected.

  • d

    Trace amount under the area calculation limit of the GC software.

  • e

    Mean value ± SD of five analyses.

Cholesterol2b10.50.522
Cycloartenol333343
24-methylene cycloartanol424322
Obtusifoliol11163
24-methylene lophenol11ε167
Episterolc11
24-methylene cholesterolεdεεε13
Campesterol1111432944
Brassicasterol230.50.559
 
24-ethylidene lophenol11131
Δ7-sitosterol12321ε
Δ7-avenasterol11111ε
Isofucosterol343442
Sitosterol646575743421
Stigmasterol656612
 
R = campesterol / sitosterol0.18 ± 0.03e0.19 ± 0.030.03 ± 0.010.03 ± 0.011.1 ± 0.52.2 ± 0.9
 
Total (µg g−1 DW)247828963305221033093858

SMT2;1 overexpressors (lines 861 and 752) were characterized by an important drop in the amount of campesterol and brassicasterol (from 11 to 4% and from 2 to 0.5%, respectively, of the total sterols) and a concomitant increase of the amount of sitosterol (from 65 to 75% of the total). The proportion of all other sterols (biosynthetic intermediates and 24-alkyl-Δ5-sterols) remain very similar in both wild-type, transgenic control and 35S::SMT2;1 overexpressors. Such a metabolic profile indicates clearly that a re-orientation of the biosynthetic flux towards the 24-ethyl sterol segment has occurred. It is possible to calculate the ratio, R, corresponding to campesterol/sitosterol. This ratio was ≈0.2 for wild-type plants analysed as a whole at the fully expended rosette stage, whereas R ≈ 0.03 for 35S::SMT2;1 overexpressors 861 and 752 (Table 1). These are designated hs1 and hs2, respectively, for high-sitosterol plants.

Co-suppressed 35S::SMT2;1 plants (lines 751 and 1312) displayed a dramatic accumulation of campesterol which increased three to four times when compared to controls, and a spectacular depletion of sitosterol which decreased up to one-third of the amount found in controls (Table 1). In addition, other 24-methyl(ene) sterols than campesterol increased slightly, for instance, 24-methylene lophenol or obtusifoliol, and other 24-ethyl(idene) sterols than sitosterol decreased slightly (Table 1). According to this novel 24-alkyl-Δ5-sterol distribution in 35S::SMT2;1 co-suppressed Arabidopsis, we can conclude that the biosynthetic flux has been re-oriented towards the 24-methyl sterol biosynthetic segment. These transgenic plants are characterized by a ratio R ≈ 1–2 and are designated hc1 (line 751) and hc2 (line 1312) for high-campesterol plants.

The sterol measurements performed with both groups of 35S::SMT2;1 plants demonstrate the selectivity of the modifications generated by the transgene 35S::SMT2;1. Such modifications essentially involve changes in the amount of two molecular species, campesterol and sitosterol. A more detailed analysis of the sterol composition of 35S::SMT2;1 plants was carried out and compared with that of the wild type. Analyses were carried out for rosette and cauline leaves, stems, inflorescences and siliques: the R values were calculated and are reported in Figure 4. The overall R values range between 0.02 and 3.87 in this Arabidopsis‘sterol map’. In the wild type, R values differ between leaves on the one side and stems and siliques on the other (Figure 4). In the case of high-sitosterol plants, R values decreased to one-fifth of the wild-type value for leaves, and to one-tenth for stems. Interestingly, R decreased only to one-third of the wild type value in the case of siliques and inflorescences. In high-campesterol plants, R values increased up to 30 times for stems; 10–15 times for inflorescences; 8–10 times for siliques; and 7–10 times for leaves. All sterol measurements were performed on whole cellular fractions. However, in the case of rosette leaves a microsomal fraction was prepared and its sterols analysed: it was shown that the sterol profile from wild-type and 35S::SMT2;1 plants was the same when measured on whole tissue or on a microsomal fraction (data not shown).

Figure 4.

Sterol map of wild-type and 35S::SMT2;1 Arabidopsis.

Schematic diagram of Arabidopsis at maturity. The sterol profile determined for rosette and cauline leaves, stems, inflorescences and siliques led to the calculation of R = campesterol / sitosterol.

Novel Δ5-sterol profiles in Arabidopsis have developmental effects

A detailed phenotypic characterization of 35S::SMT2;1 lines hs1 and hs2 (high sitosterol) and hc1 and hc2 (high campesterol) was undertaken with plants from T3 or T4 generations. At first glance, all the transgenic lines were characterized by a size reduction at maturity ranging from two-thirds to one-third of the size of wild-type and transgenic control plants, as indicated in Table 2. Apart from this, the phenotypes of high-sitosterol or high-campesterol plants differed as illustrated in Figure 5.

Table 2.  Developmental and growth major characteristics a of 35S::SMT2;1 Arabidopsis
ParameterWild
type
Transgenic
control
hs1hs2hc1hc2
  • a

    Measurements were done with 10 plants per line. Plants were grown between November and March in a greenhouse under a photoperiod of 15 h

  • light at 23°C / 9 h dark at 18°C. Two independent experiments gave identical results.

Height at maturitya (cm)30 ± 836 ± 1216 ± 518 ± 314 ± 719 ± 6
Epidermal cell lengtha (µm)221 ± 30258 ± 24153 ± 46111 ± 32243 ± 30322 ± 35
Branching of stema17 ± 119 ± 416 ± 49 ± 227 ± 345 ± 7
Fertility (mg seeds per plant)242 ± 50191 ± 74134 ± 38121 ± 4415 ± 522 ± 4
Life span (weeks)888810–1210–12
Flowering duration (weeks)44446–86–8
Figure 5.

Phenotypes of Arabidopsis transformed with 35S::SMT2;1.

D evelopmental modifications in the case of overexpression (hs1) and co-suppression (hc2) of SMT2;1 compared with the wild type.

(a) Greenhouse-grown plants at maturity.

(b–d) Epidermis cells of the fourth internode of the stems. Bar = 20 µm.

(e–g) Representative set of green siliques. Magnification × 10.

(h–j) Representative set of flower buds and open flowers. Magnification × 16.

(k,l) Longitudinal stamen sections before anthesis.

(m) Germination on agar water of wild type, det2, fackel, hs1, hs2, hc1 and hc2 after 11 days.

(n) Etiolation response of germination in the dark on agar water of samples as in (m).

In the case of high-sitosterol plants, the growth reduction may be defined by a uniform reduction in stem and leaf size that resulted in plants exhibiting a ‘wild-type-like’ phenotype corrected by a scale factor (Figure 5a), but sizes of flowers and siliques remained unchanged (Figure 5e,f,h,i). The life span and flowering duration of those lines was identical to that of the controls; however their fertility was significantly reduced to approximately half the amount of seeds produced by control plants (Table 2), probably due to the fact that high-sitosterol plants produce fewer flowers than the wild type. Seeds from control or transgenic lines were indistinguishable in terms of morphological aspect (data not shown). The reduced growth rate and stature of high-sitosterol plants could be due to a defect in cell division and/or elongation. To address this question, we measured the size of cells taken from the fourth internode of the stems. Results as shown in Figure 5(b,c) indicate that the size of transgenic cells is significantly reduced when compared to the wild type; the length of epidermal cells was found to be reduced to 50–70% of the wild type (Table 2).

In the case of hc1 and hc2, a number of novel phenotypic traits were remarkable, in addition to the height reduction measured at plant maturity. Firstly, the life span of high-campesterol plants was prolonged (from 8 weeks for the wild type to 10–12 weeks), and this was also true of flowering duration (Table 2). Secondly, the architecture of high-campesterol plants was radically modified when compared to the wild type, or even more, to high-sitosterol plants: high-campesterol plants showed a bushy appearance due to an increased number of secondary stems originating from the axillary buds (Figure 5a). The branching appearance is characterized as an average number of stems, which is doubled when compared to controls (Table 2). Such branching in high-campesterol plants is probably typical of reduced apical dominance. Thirdly, the morphological pattern of flowers is altered: flower buds at early stages are already open and show a protruding pistil, whereas wild-type flowers from the same stage are still closed (Figure 5h,j). In addition, the shape of sepals and petals is markedly modified: the margins of these flower pieces are serrate whereas those of wild-type flowers are linear (Figure 5h,j). Fourthly, the fertility of high-campesterol plants was considerably reduced when compared to the wild type or to high-sitosterol (low-campesterol) plants: the siliques were smaller (Figure 5g) and the amount of seeds harvested per plants was decreased by a factor 10 (Table 2). Along these lines, the observation of anthers before anthesis revealed a reduction in the amount of pollen (Figure 5k,l). Moreover, fluorescein diacetate/DAPI staining of pollen samples collected at anthesis indicates that a majority of pollen grains from transgenic plants were dead, whereas those from the wild type or from high-sitosterol plants were not (data not shown).

Height reduction of high-campesterol plants was monitored. Size measurements were performed on cells taken from the fourth internode of high-campesterol stems. As shown in Table 2 and Figure 5(d), cell length in high campesterol and wild-type plants was similar, indicating that the growth alteration in the former is not due to a defect in cell elongation.

Finally, we compared germination of high-sitosterol and high-campesterol plants with germination of the brassinosteroid biosynthesis mutant det2, and of a recently characterized sterol mutant, fackel (Jang et al., 2000; Schrick et al., 2000). Brassinosteroid mutants exhibit post-embryonic growth defects (dwarfism) and constitutive photomorphogenesis in the dark, due to impaired conversion of campesterol into active brassinosteroids. Fackel mutations in a sterol-14-reductase impair the organization of cell growth and division at the embryonic stage of development, due to a depletion in Δ5-sterols (campesterol, sitosterol) and a concomitant accumulation of Δ8,14-sterol (biosynthetic intermediates). Comparative germination of wild-type, det2, fackel, high-sitosterol and high-campesterol plants in the light or in the dark is shown in Figure 5(m,n). Whereas det2 and fackel displayed reduced growth in the light (Figure 5m) and absence of normal (wild-type) etiolation response in the dark, high-sitosterol and high-campesterol seedlings grow identically to the wild type in both light (Figure 5m) and dark (Figure 5n).

Together our observations show that a variation of the ratio of campesterol to sitosterol is associated with developmental modifications in Arabidopsis. These changes are of two types which differ remarkably. In the case of a decrease of campesterol and an increase of sitosterol, growth alteration results in the overall size reduction of vegetative organs. In the case of an increase of campesterol and a decrease of sitosterol, the growth alteration is seen as an additional size reduction, modified apical dominance, novel flower morphology, and drastically reduced fertility.

Brassinosteroid treatment complements growth alteration of high-sitosterol but not of high-campesterol Arabidopsis

A probable physiological link between a modified campesterol and sitosterol composition and growth or development alterations may be brassinosteroid synthesis and/or availability in 35S::SMT2;1 lines. This possibility was challenged by germinating and growing in vitro T4 plants from lines hs1, hs2, hc1 and hc2 in the presence of 24-epibrassinolide. We germinated and grew 35S::SMT2;1 plants for 2 months on germination medium (GM) supplemented with 0.05 µm 24-epiBR. This treatment has no effect on either height or cell length of wild-type or transgenic control plants. Results given in Table 3 clearly show that treated hs1 (Table 3) and hs2 (data not shown) plants have the same height at maturity as untreated wild-type plants; the length of epidermal cells from treated transgenic plants increased close to the value of untreated control plants. Plants from the co-suppressed lines hc1 (data not shown) and hc2 (Table 3) were not affected by a treatment with 24 epi-brassinolide, indicating that growth and developmental alterations associated with a high-campesterol and low-sitosterol phenotype are probably not due to a reduced endogenous brassinosteroid synthesis or supply.

Table 3.  Morphometric analysis of 24-epibrassinolide-treated 35S::SMT2;1 Arabidopsis
Parameter24-epibrassinolide (µm)
Wild typeTransgenic controlhs1hc2
050050050050
  • a

    Plants were grown for 2 months on GM in 2 l jars (five plants per jar). Results are average ± SD taken from 10 plants. Two independent experiments produced equivalent data.

Heighta (cm)13 ± 314 ± 314 ± 313 ± 310 ± 313 ± 37 ± 58 ± 3
Epidermal cell lengtham)229 ± 48232 ± 62307 ± 68300 ± 58185 ± 46292 ± 6272 ± 64262 ± 43

We conclude from the biochemical complementation experiments carried out with 35S::SMT2;1 plants in the presence of 24-epibrassinolide that the size reduction of plants with a low campesterol content would be primarily due to a reduced endogenous brassinosteroid synthesis or availability.

Discussion

In this article we report the phenotypic characterization of A. thaliana transformed with a 35S::SMT2;1 transgene. SMT2 is the enzyme of the sterol pathway which confers to the plant cell the capability to synthesize 24-ethyl sterols (Benveniste, 1986; Bouvier-Navéet al., 1997; Fonteneau et al., 1977; Schaller et al., 1998) from a 24-methylene sterol precursor (Figure 1). As such, SMT2 controls a branching point of the biosynthesis pathway that defines two biosynthetic segments: one of those produces 24-methyl sterols such as campesterol, the other produces 24-ethyl sterols such as sitosterol, in Arabidopsis. Therefore variations in the amount of sitosterol and campesterol are based on modulations of the activity of SMT2.

Overexpression of At SMT2;1 increased the biosynthetic flux in the 24-ethyl sterol segment of the pathway at the expense of the 24-methyl sterol segment. Such plants are characterized by a high sitosterol content and a low campesterol content when compared to the wild type. On the other hand, co-suppression of AtSMT2;1 increased the biosynthetic flux in the 24-methyl sterol segment at the expense of the 24-ethyl sterol segment of the pathway. In that case, transgenic plants are characterized by a low-sitosterol and a high-campesterol content when compared to the wild type. These data show that modification of the sterol profile is highly selective because the modulation of SMT2;1 expression affects only the ratio of campesterol to sitosterol, reinforcing the demonstration of the function of SMT2;1 in Arabidopsis (Bouvier-Navéet al., 1997).

High-sitosterol Arabidopsis plants have an overall reduced stature, probably due to a reduction in cell size indicating a defect in cell elongation. A growing number of reports have recently shed light on the importance of brassinosteroids as regulators of cell elongation and enlargement. Our results indicate that growth defects of high-sitosterol Arabidopsis are complemented by an exogenous supply of 24-epibrassinolide. From this we propose that the growth defect would be due to insufficient availability or supply of campesterol. It was demonstrated in a number of plant species, including Arabidopsis, that campesterol is the precursor of brassinosteroids (Fujioka et al., 1997; for review see Yokota, 1997). Therefore a reduced amount of campesterol in the plant cell would probably lead to a reduced amount of its metabolites, active brassinosteroids, and finally to the observed morphological phenotype. The observation of such a size reduction phenotype among different 35S::SMT2;1 species, Arabidopsis (this work) and tobacco (Schaller et al., 1998), suggests that cell-size reduction is not the only cause of size reduction in low-campesterol (high-sitosterol) plants; the growth defect in high-sitosterol tobacco is due to a reduction in the number of cell divisions (Schaller et al., 1998).

Growth and developmental modifications associated with a high campesterol content in plants are diverse, and differ from the growth reduction observed in high-sitosterol plants. High-campesterol plants have pleiotropic effects on development such as increased branching, reduced growth and reduced fertility. In this latter effect, reduced seed production may result in the observed increased flowering time (Hensel et al., 1994). We have shown that 24-epibrassinolide is ineffective in complementation experiments on high-campesterol plants, suggesting that the phenotype is probably not linked to a defect in brassinosteroid synthesis or availability in the cell. Accordingly, the fact that campesterol (the committed precursor of brassinosteroids in Arabidopsis) accumulates favours non-limiting brassinosteroid production in high-campesterol plants. Along these lines, it has been shown that bri1 plants, which bear mutations at the BRI locus encoding a receptor kinase that may act as a receptor for brassinolide (Li and Chory, 1997), accumulate brassinosteroids and campesterol probably due to an impaired feedback regulation mechanism of brassinosteroid biosynthesis (Noguchi et al., 1999). Thus dwarfism of bri mutants is caused by an impaired signalling pathway or response to brassinosteroid, although these molecules and their sterol precursor are synthesized in excessive amounts in the cell. Measurements of cell length in high-campesterol plants indicates that cell elongation is not involved in reduced stature. This would suggest that cell divisions are reduced in number and, more generally, that meristematic activity in plants would be sensitive to a modified 24-alkyl sterol composition of the cell. Pleiotropic effects on development of the enrichment in campesterol and/or depletion in sitosterol may be explained alternatively by modifications of membrane properties in terms of fluidity and protein environment. It is well established that sterols in the membrane can affect or regulate the function of membrane-bound proteins. For instance, modifications of the sterol composition by inhibitors, leading to the replacement of Δ5-sterols by 9β,19-cyclopropyl sterols or 14α-methyl-Δ8-sterols, cause significant modifications of the properties of membrane H+-ATPase (Grandmougin-Ferjani et al., 1997). In the case of such a proton pump, it has been shown recently that co-suppression of the plasma membrane H+-ATPase 4 in tobacco severely affects a number of transport-dependent physiological processes, including sucrose translocation, and results in poor growth (Zhao et al., 2000). Finally, our attempts to rescue high-campesterol plants by feeding them with sitosterol were ineffective (data not shown), probably because of the hydrophobic nature of the molecules supplied.

The characterization of smt1 Arabidopsis lines isolated from a population of Ac transgenic mutants (Diener et al., 2000) shows that a defect in SMT1 gene expression selectively modifies the sterol composition of the plants that accumulate cholesterol and are depleted in sitosterol. This novel sterol profile is associated with pleiotropic developmental defects such as reduced growth and fertility, root hypersensitivity to calcium salts, and aberrant embryo morphogenesis (Diener et al., 2000). Smt1 plants are not complemented by brassinosteroid, probably due to an unchanged level of campesterol (Diener et al., 2000), which is in complete accordance with the results presented here. It appears that the developmental defects might be caused by an enrichment of plant membranes in cholesterol and/or a depletion in sitosterol, as discussed above. Alternatively, the possibility has been documented that certain sterols or biosynthetic intermediates of the pathway may play a role as signals in some aspects of development, including embryogenesis, particularly in the case of mutations in FACKEL which encodes a sterol-C14-reductase (Jang et al., 2000; Schrick et al., 2000). We have seen that, unlike smt1 or fackel, 35S::SMT2;1 lines germinate in the light and in the dark similarly to the wild type, indicating normal development at the seedling stage.

The present work shows that the ratio of campesterol to sitosterol is determined at the level of SMT2;1 in Arabidopsis; this ratio has to be properly balanced to fit the sterol needs of the cell with respect to growth requirements and membrane integrity.

Experimental procedures

Plant material

Arabidopsis thaliana (L.) Heynh ecotype C24 was obtained from the Nottingham Arabidopsis Stock Center. The mutant line det2 was provided by the Arabidopsis Biological Resources Center, Ohio. The mutant line fackel was kindly donated by Kathrin Schrick, University of Tuebingen. Plants were sown, then grown either in vitro or in the greenhouse. Surface-sterilized seeds were sown on Murashige and Skoog (MS) medium containing meso-inositol (100 mg l−1), thiamine HCl (0.5 mg l−1), pyridoxine (0.5 mg l−1), nicotinamide (0.5 mg l−1), and 0.7% (w/v) agar pH 5.7. Plants were grown in vitro at maturity on the same germination medium (GM) under a photoperiod of 16 h light at 23°C and 8 h dark at 20°C. In the greenhouse seeds were sown in soil, then germinating seedlings were transferred individually into pots of 8 cm diameter and grown at maturity. For brassinosteroid treatments, 24-epibrassinolide (CIDtech Research Inc., Mississauga, Ontario, Canada) was dissolved in ethanol as a 5 mm stock solution and added to GM medium. Comparative germination tests in the light or dark were carried out on agar (0.7%) water.

Microscopy and cell measurements

Stem epidermis was peeled off from the middle of the fourth internode of the main stem of 6-week-old plants. At least 20 cells per epidermal peel were measured under a microscope with a micrometer (Zeiss, Oberkochen, Germany). Measurements were performed on 10 plants per plant line.

Plant transformation

The 35S::SMT2;1 T-DNA construct and A. tumefaciens strain LBA4404 used in this work were as described (Schaller et al., 1998). Arabidopsis transformation was conducted according to a published procedure (Valvekens et al., 1988). Transgenic plants from succeeding generations were selected on GM medium for resistance to 50 mg l−1 kanamycin; resistant seedlings were then transferred on GM medium or in soil. Segregation analysis of the kanamycin resistance trait, as well as DNA gel-blot analysis of the transgenic lines studied in this article using 32P-labelled probes derived from AtSMT2;1, indicate the presence of at least two T-DNA loci in all lines.

Extraction of nucleic acids and gel-blot analysis

Total RNA was extracted according to the method of Goodall et al. (1990). Northern blot needed the separation of 10 µg RNA samples on formaldehyde gels as described (Sambrook et al., 1989). RNAs were blotted onto a Hybond-N+ nylon membrane (Amersham, Little Chalfont, Bucks, UK) and hybridized with 32P-labelled DNA probes obtained by random priming of a DNA template.

Extraction and dosage of sterols

Lipids were extracted from ≈5–10 mg (small-scale qualitative analysis) or 100–200 mg (quantitative analysis) of ground, dry material at 70°C in dichloromethane : methanol (2 : 1, v/v). The dried residue was saponified with 6% (w/v) KOH in methanol at 90°C for 1 h to release the sterol moiety of steryl esters. Sterols were then extracted with three volumes of n-hexane and an acetylation reaction was performed on the dried residue for 1 h at 60°C in toluene with a mixture of pyridine : acetic anhydride (1 : 1, v/v). Steryl acetates were resolved by thin-layer chromatography (Merck, Darmstadt, Germany, 60F254 precoated silica plates) with one run of dichloromethane as a single band at Rf = 0.5. Purified steryl acetates were separated and identified by gas chromatography (GC) using a Varian 8300 gas chromatograph with a flame ionization detector and a glass capillary column (wall-coated open tubular, 30 m long, 0.25 mm internal diameter, coated with DB1; J.&W. Scientific, Folsom, CA, USA) using H2 as a carrier gas (2 ml min−1). The temperature program included a fast increase from 60 to 220°C (30°C min−1) and a slow increase from 220 to 280°C (2°C min−1). Data from the detector were monitored with the varian star computer program (Varian, Walnut Creek, CA, USA). Sterol structures were confirmed by GC–MS (Fison MD800 equipped with a glass capillary column: WCOT coated with DB5, J.&W. Scientific) according to (Rahier and Benveniste, 1989). Determination of the sterol content from microsomes was as indicated above, except for the extraction procedure which started with the saponification step.

Acknowledgements

We thank K. Schrick (University of Tuebingen) and the Stock Centers (NASC, ABRC) for providing seeds. We are indebted to B. Bastian for kindly typing the manuscript.

Accession numbers At SMT1, AF 090372, At SMT2;1, X89867, At SMT2;2, U71400.

Ancillary