The Arabidopsis dw f 7/ste1 Mutant Is Defective in the ⁷ Sterol C-5 Desaturation Step Leading to Brassinosteroid Biosynthesis

Correspondence to: Kenneth A. Feldmann, feldmann{at}ag.arizona.edu (E-mail), 520-621-7186 (fax)

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Lesions in brassinosteroid (BR) biosynthetic genes result in characteristic dwarf phenotypes in plants. Understanding the regulation of BR biosynthesis demands continued isolation and characterization of mutants corresponding to the genes involved in BR biosynthesis. Here, we present analysis of a novel BR biosynthetic locus, dwarf7 (dwf7). Feeding studies with BR biosynthetic intermediates and analysis of endogenous levels of BR and sterol biosynthetic intermediates indicate that the defective step in dwf7-1 resides before the production of 24-methylenecholesterol in the sterol biosynthetic pathway. Furthermore, results from feeding studies with ¹³C-labeled mevalonic acid and compactin show that the defective step is specifically the ⁷ sterol C-5 desaturation, suggesting that dwf7 is an allele of the previously cloned STEROL1 (STE1) gene. Sequencing of the STE1 locus in two dwf7 mutants revealed premature stop codons in the first (dwf7-2) and the third (dwf7-1) exons. Thus, the reduction of BRs in dwf7 is due to a shortage of substrate sterols and is the direct cause of the dwarf phenotype in dwf7.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Sterols are known to play at least two critical roles in plants: as bulk components of membranes regulating stability and permeability (Bach and Benveniste 1997 ) and as precursors of growth-promoting brassinosteroids (BRs; Fujioka and Sakurai 1997 ). Sterol biosynthesis in plants has been studied extensively through enzyme purification or gene cloning (Grunwald 1975 ; Goodwin 1979 ; Benveniste 1986 ; Bach and Benveniste 1997 ). Figure 1 shows the proposed biosynthetic pathway from squalene to brassinolide (BL). A major difference between photosynthetic and nonphotosynthetic organisms is that cyclization of squalene 2,3-oxide is bifurcated to a different route for each system (Benveniste 1986 ). In animals and yeast, squalene 2,3-oxide is cyclized to lanosterol, whereas in photosynthetic organisms it is cyclized to cycloartenol (Nes and McKean 1977 ). Accordingly, photosynthetic organisms require somewhat different biosynthetic enzymes, such as cycloartenol synthase (Corey et al. 1993 ) and cycloeucalenol–obtusifoliol isomerase, which are required to open the cyclopropane ring in cycloartenol (Figure 1). However, most of the enzymatic steps are shared between the two different pathways.

View larger version (36K): [in this window] [in a new window]   Figure 1. The Proposed BL Biosynthetic Pathway from Squalene to BL. The BL biosynthetic pathway is divided into the sterol-specific pathway, squalene to campesterol, and the BR-specific pathway, campesterol to brassinolide. Common names for the compounds are labeled, and proposed enzymes involved in each reaction are boxed and labeled. Genes identified by mutants are marked. The acronyms for some compounds are in parentheses. In the inset, the carbon atoms of the sterol core rings and side chain are numbered.

In plants, sterols are subject to a series of modifications before conversion to BL. Different sterols, such as 24-methylenecholesterol (24-MC), campesterol (CR), isofucosterol, and sitosterol, are converted to the BL congeners dolicholide, BL, 28-homodolicholide, and 28-homoBL, respectively, in a species-specific manner (Fujioka et al. 1995 ; Sasse 1997 ). The BR-specific pathway diverges into the early and the late C-6 oxidation pathways. In the early C-6 oxidation pathway, introduction of a 6-oxo group occurs before the vicinal hydroxylation reactions at the side chain, whereas it occurs after these hydroxylations in the late C-6 oxidation pathway (Figure 1; Choi et al. 1997 ).

Several mutants, such as constitutive photomorphogenesis and dwarfism (cpd), deetiolated2 (det2), and dwarf4 (dwf4), have been shown to be defective in the BR-specific pathway (Li et al. 1996 , Li et al. 1997 ; Szekeres et al. 1996 ; Choe et al. 1998 ). These BR biosynthetic dwarfs share a characteristic dwarf phenotype, which includes short robust stems, reduced fertility, prolonged life cycle, and dark-green, round, and curled leaves when grown in the light. In the dark, these mutants exhibit short hypocotyls and expanded cotyledons. cpd (dwf3) mutants are only rescued by 23-hydroxylated compounds (Szekeres et al. 1996 ). The CPD gene was shown to encode a cytochrome P450 steroid hydroxylating enzyme (CYP90A1). In addition, Li et al. 1996 , Li et al. 1997 showed that det2/dwf6 is blocked in the C-5 reduction step. DET2 was found to be homologous to steroid 5-reductases. Like its animal equivalents, DET2 successfully converted progesterone (3-oxo-^4,5 steroid) to 4,5-dihydroprogesterone in a human cell line. In addition, the human 5-reductase gene effectively complemented det2 mutants (Li et al. 1997 ). Recently, we have shown that DWF4 encodes a cytochrome P450 whose amino acid sequence is 43% identical to CPD; DWF4 has been named CYP90B1 (Choe et al. 1998 ). Based on results from feeding studies using BR biosynthetic intermediates, we showed that the proposed rate-limiting step of BR biosynthesis, 22-hydroxylation, is blocked in dwf4 mutants.

In the plant sterol biosynthetic pathway, several of the genes have been cloned or identified based on heterologous expression or sequence similarity. First, Corey et al. 1993 isolated a cycloartenol synthase cDNA by heterologous complementation of yeast mutants lacking lanosterol synthase. In addition, two types of cDNAs encoding sterol methyltransferases have been isolated from soybean (Shi et al. 1996 ) and Arabidopsis (Husselstein et al. 1996 ). The Arabidopsis cDNA has been shown to mediate a second methyltransferase step leading to C₂₉ sterols (Bouvier-Nave et al. 1997 ). For the 14-demethylation reaction, Bak et al. 1997 cloned the cDNA encoding the 14-demethylase cytochrome P450 enzyme (CYP51) from Sorghum bicolor. Based on sequence similarity, Grebenok et al. 1997 identified an Arabidopsis sterol C-8 isomerase (GenBank accession number AF030357). Furthermore, an ERGOSTEROL25 (ERG25) homolog for Arabidopsis (C-4 demethylase) also has been discovered in the genome sequencing project (GenBank accession number AL021635). Finally, a sterol C-7 reductase has been cloned by heterologous expression of an Arabidopsis cDNA in yeast (Lecain et al. 1996 ).

As compared with the wealth of cloned genes in sterol biosynthesis, only one mutant has been found in these genes. Gachotte et al. 1995 screened an ethyl methanesulfonate (EMS)–induced mutant population (22,000 M₂ plants) for mutants displaying an altered sterol profile. The screen yielded one mutant, sterol1 (ste1), whose endogenous level of C-5–desaturated sterols is reduced to 30% of that of the wild type. Expression of the yeast gene ERG3 (the gene for ⁷ sterol C-5 desaturase) in the ste1-1 mutant increased the level of C-5–desaturated sterols 1.7- to 2.8-fold compared with the ste1-1 control, suggesting functional conservation of the enzymes from yeast and plants. However, visible phenotypes were not found in ste1-1 plants. Thus, the authors hypothesized that the residual 30% level of C-5–desaturated sterols was sufficient for the growth of plants.

We have been characterizing a large collection of BR dwarf mutants. Of the eight dwf loci identified to date, dwf3 (cpd; Szekeres et al. 1996 ), dwf4 (Choe et al. 1998 ), and dwf6 (det2; Li et al. 1996 ) have been shown to act in the BR biosynthetic pathway, whereas dwf2 (bri1) probably is involved in BR perception (Clouse et al. 1996 ; Li and Chory 1997 ). In this study, we provide a morphological and genetic characterization of dwf7. Based on our results from biochemical analyses, including feeding studies and quantification of endogenous levels of BRs, we hypothesize that dwf7 mutants possess lesions in the STE1 gene previously cloned by Gachotte et al. 1996 . Here, we show that the STE1 locus in dwf7 mutants contains loss-of-function mutations, and we designate dwf7-1 and dwf7-2 as ste1-2 and ste1-3, respectively.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Isolation of dwf7 Mutants
The dwf7-1 mutant originally was identified in a screen of 14,000 T-DNA–transformed lines of Arabidopsis. Genetic complementation tests with other dwf loci indicated that dwf7 belongs to a unique complementation group. dwf7-1 segregated as a monogenic recessive mutation; progeny from a heterozygote segregated 325 (wild type):98 (dwf7-1). Although dwf7-1 originated from a T-DNA mutant population, it failed to cosegregate with the kanamycin resistance marker in the T-DNA, suggesting that dwf7-1 was an untagged mutant. Furthermore, mapping the dwf7-1 mutation to the Arabidopsis genome by using simple sequence length polymorphisms (SSLPs; Bell and Ecker 1994 ) confirmed that dwf7 maps to a location different from previously isolated dwarfs. The meiotic recombination ratio between dwf7 and the SSLP marker nga172 on chromosome 3 was scored as 0 / 86, indicating tight linkage of dwf7 to nga172. According to a recent recombinant inbred map of Arabidopsis, nga172 is located 2.2 centimorgans from the top of chromosome 3.

A second allele of dwf7 was identified among 43 dwarf mutants isolated by screening >50,000 M₂ seeds of an EMS mutant population. Similar to dwf7-1, the new allele was biochemically complemented by early BR biosynthetic intermediates, including 22-hydroxycampesterol (22-OHCR) and cathasterone (data not shown), and mapped near nga172 (data not shown). Sequencing revealed a premature stop codon in exon 1 (see below).

Morphological Analysis of dwf7-1
dwf7 displays many of the characteristics of other BR dwarfs as shown in Figure 2. Compared with 1-month-old wild-type plants (Figure 2A), dwf7-1 plants grown for 5 weeks in the light possess short robust inflorescences, dark-green, round leaves, reduced fertility, and short pedicels and siliques (Figure 2C). The wild type generally terminates flowering before 7 weeks of age; however, dwf7-1 continues to produce flowers at this age, indicating a prolonged life span (Figure 2B). Additional morphological defects of 5-week-old light-grown plants are summarized in Table 1. Most noticeably, the height of dwf7-1 plants is strikingly reduced and is only 14% that of wild-type height. The leaf blade width of dwf7-1 mutants is similar to that of wild-type plants; however, the length is greatly reduced (1.8 cm) as compared with that of the wild type (3 cm), resulting in the round shape of dwf7-1 leaves. The overall morphology of dwf7-2 was similar to dwf7-1 except that it was slightly shorter and more sterile (data not shown).

View larger version (44K): [in this window] [in a new window]   Figure 2. Morphology of Wild-Type and dwf7-1 Plants. (A) Wild-type plant at 1 month of age. (B) and (C) dwf7-1 plants at 7 and 5 weeks of age, respectively. The characteristic dwarf phenotype, such as short robust stems, reduced fertility, and dark-green, round, and curled leaves, is shown in (B) and (C). At 7 weeks of age (B), wild-type plants had ceased growing, whereas dwf7-1 plants continued to grow. Bar = 2 cm.

Because null mutations in the BR pathway result in a dwarf phenotype, as well as defects in skotomorphogenesis, we compared the dwf7-1 mutant with other BR dwarfs for growth in the dark. Hypocotyl lengths from the longest to the shortest were 18 ± 1.6 (wild type; units in millimeters ±SE; n = 15), 6.3 ± 0.29 (dwf7-1), 4.1 ± 0.03 (det2/dwf6), 1.26 ± 0.09 (dwf4), 1.24 ± 0.08 (cpd/dwf3), and 1.18 ± 0.08 (bri1/dwf2). These data indicate that dwf7-1 displays a less severe phenotype (35% that of wild-type hypocotyl length) than do other BR dwarfs (e.g., 7% of wild type in dwf4; Choe et al. 1998 ). Furthermore, dwf7-1 frequently displayed closed cotyledons and hooks similar to those of the wild type, whereas severe dwarfs, including bri1/dwf2, cpd/dwf3, and dwf4, showed expanded cotyledons and open hooks (data not shown).

Unlike severe dwarfs, such as dwf4 and cpd, dwf7-1 mutants are not mechanically sterile (Figure 2B). However, the average number of seeds in a silique is reduced in dwf7-1 (n = 12) compared with that of the wild type for reasons yet to be identified (n = 49) (Table 1). Scanning electron microscopy (Figure 3A to C) demonstrates a relationship between fertility and floral structure. In the wild type (Figure 3A), the length of stamens is greater than or similar to that of the gynoecium (quantified in Figure 3d), facilitating dehiscence of pollen on the stigmatic surface. Although dwf7-1 flowers (Figure 3B and Figure 3d) possess stamens and gynoecia that are shorter than those in the wild type, the fertility of dwf7-1 flowers is possible through a concomitant reduction in the length of both organs. In contrast, only stamen elongation is affected more severely in dwf4 flowers (Figure 3C and Figure 3d). The short stamen length in dwf4 is likely to cause dehiscence of pollen on the ovary wall rather than on the stigmatic surface. In fact, when dwf4 pollen is transferred to either wild-type or dwf7-1 stigmas, viable seeds are made.

View larger version (78K): [in this window] [in a new window]   Figure 3. Analyses of Wild-Type, dwf7-1, and dwf4-3 Flowers. (A) to (C) Scanning electron microscopy of flowers from wild-type (ecotype Wassilewskija-2 [Ws-2]), dwf7-1, and dwf4-3 plants. Flowers were harvested immediately after petal opening. The wild-type flower shown in (A) has a stamen length similar to or slightly greater than that of the gynoecium, facilitating pollen dehiscence on the stigmatic surface. A fertile dwf7-1 flower is shown in (B) with a concomitant reduction in the size of the gynoecium and the stamen. A sterile dwf4-3 flower is shown in (C) with shorter filaments than the gynoecium, preventing pollen dehiscence on the stigmatic surface. Bar in (A) = 1 mm. (D) Measurements of gynoecia and stamens shown in (A) to (C). dwf7-1 displays a concomitant reduction in the length of gynoecia and stamens, whereas dwf4-3 displays a greater reduction in stamen length. Each data point represents the average length for five flowers. Standard errors are shown at each data point.

The common denominator for the various phenotypes found in dwf7-1 mutants is a reduction in longitudinal growth, which could be due to either a reduced number of cells or a failure in cell elongation. Observations made with other BR dwarf mutants suggest that the number of cells is comparable in the wild type and mutants (Kauschmann et al. 1996 ; Nomura et al. 1997 ; Azpiroz et al. 1998 ). Figure 4 displays cell size in the wild type (Figure 4A) and dwf7-1 (Figure 4B) mutants. The length of cells in the epidermis, cortex, and xylem of dwf7-1 is greatly reduced (<30% of wild type). This reduced cell size is converted to the length of the wild type in response to application of BL (Figure 4C). Thus, the reduced organ length in dwf7-1 also is due to a failure of cell elongation.

View larger version (129K): [in this window] [in a new window]   Figure 4. Light Microscopy of Wild-Type and dwf7-1 Stems Sectioned Longitudinally and Transversally. (A) Wild type. (B) dwf7-1. Cell size is reduced drastically in many different tissues, such as the epidermis, cortex, and vasculature. (C) dwf7-1 plants treated with BL. Cells were restored to their wild-type length in response to daily application of 10-7 M BL for 1 week. (D) and (E) Cross-sections of wild-type and dwf7-1 stems, respectively. Positions of the vascular bundles are marked with triangles. Eight evenly spaced bundles are found in the wild type, whereas fewer and irregularly located bundles were present in dwf7-1. Bar in (C) = 100 µm for (A) to (C) (longitudinal sections); magnification x100 for (D) and (E).

We also have examined the organization of vascular bundles in wild-type and dwf7-1 mutants. Wild-type inflorescences possess eight vascular bundles (Figure 4D). However, the number of vascular bundles was reduced to six in dwf7-1 (Figure 4E). Furthermore, the spacing between the vascular bundles in dwf7-1 is irregular. In the wild type, interfascicular parenchyma cells alternate regularly with vascular bundles; however, cross-sections of dwf7-1 show that two vascular bundles are joined without being separated by parenchyma cells. Within a single vascular bundle, the size and number of xylem cells in dwf7-1 plants generally are reduced, whereas the number of phloem cells is similar to or even greater than that in the wild type. This characteristic abnormality of vascular bundle organization has been observed consistently in other BR dwarfs (Szekeres et al. 1996 ; S. Choe, C.P. Tissier, and K.A. Feldmann, unpublished data).

Biochemical Complementation of dwf7-1 with BL
Figure 5 demonstrates that dwf7-1 seedlings grown in BL-supplemented liquid media were remarkably sensitive to BL. Growth in 1 nM BL induced significant elongation of dwf7-1 hypocotyls (160% increase), whereas the wild-type increase was marginal (5%). Treatment with 10 and 100 nM BL completely rescued dwf7-1 hypocotyls to wild-type length. The strongest response of the wild type to BL was obtained at 100 nM (Figure 5). Higher concentrations of BL (1 µM) caused a stressed morphology, including inhibition of root growth and swollen, twisted, and fragile hypocotyls in both dwf7-1 and wild-type plants (data not shown). After BL treatment of dwf7-1, cells in the treated region of the stem are similar in length to wild-type cells (Figure 4C).

View larger version (18K): [in this window] [in a new window]   Figure 5. Response of Light-Grown Wild-Type and dwf7-1 Hypocotyls to Different Concentrations of BL. dwf7-1 responds to 10-9 M BL and is completely rescued by 10-8 M BL. Error bars indicate ±SE.

Identification of the BR Biosynthetic Defect in dwf7-1
Biochemical complementation of dwf7-1 following application of BL suggested that dwf7-1 is likely to be defective in BR biosynthesis. To pinpoint the defective step in the BR biosynthetic pathway, dwf7-1 mutants were treated with BR biosynthetic intermediates. Due to undetectable bioactivity of some early intermediates (CR to 6-oxocampestanol) in bioassays (Fujioka et al. 1995 ; Choe et al. 1998 ), these were not used. Instead, we chose three biologically active compounds, 22-OHCR, 6-deoxoCT, and BL, for these feeding tests (see Figure 1). Because the 22-hydroxylation reaction is reported to be mediated by DWF4 (Choe et al. 1998 ), biochemical complementation of dwf mutants other than dwf4 by 22-OHCR places the defective step upstream of CR.

Figure 6A shows the response when inflorescences were treated daily for 1 week with these BR intermediates. Complementing compounds induced growth of internodes and strongly increased pedicel length. dwf7-1 pedicels treated with 22-OHCR and BL showed growth greater than or equal to that of the wild type. Measurements of pedicel length shown in Figure 6B demonstrated that the three compounds tested, 22-OHCR, 6-deoxoCT, and BL, all increased dwf7-1 pedicel length >200% as compared with the control, suggesting that the defective step in BR biosynthesis is located at or before the CR biosynthetic step. Similarly, 3-week-old inflorescences of dwf7-2 were tested with 22-OHCR, 6-deoxoCT, teasterone, and BL. All four compounds induced significant elongation of pedicels and internodes (data not shown), indicating that dwf7-1 and dwf7-2 share the same biosynthetic defect.

View larger version (34K): [in this window] [in a new window]   Figure 6. Wild-Type and dwf7-1 Inflorescences Treated with BR Intermediates. (A) From left to right, dwf7-1 inflorescences treated with water, 22-OHCR, BL, and a wild-type inflorescence treated with water, respectively. (B) Quantification of the inflorescence feeding test shown in (A). The lengths of pedicels treated with water, 6-deoxoCT, 22-OHCR, and BL were measured to the nearest millimeter (n > 15). The pedicels elongated greater than twofold in response to all the BRs tested, suggesting that the biosynthetic defect in dwf7-1 resides before the production of CR. Error bars indicate ±SE.

As shown in Table 2, more definitive results indicating a specific defect in BR biosynthesis have been obtained from gas chromatography–selective ion monitoring (GC-SIM) analysis of endogenous BRs and sterols in dwf7-1 plants. The endogenous levels of sterols, such as 24-MC, CR, and campestanol (CN), in wild-type plants, were 3800, 32,900, and 1140 ng/g fresh weight, respectively. However, the levels of all three sterols in dwf7-1 mutants were extremely diminished at 3.1, 1.1, and 1.4% of the wild type, respectively, suggesting that the biosynthetic block is located before 24-MC. These data are consistent with the results of intermediate feeding studies (Figure 6).

Further biochemical feeding studies with ¹³C-labeled mev-alonic acid (MVA) and compactin, a MVA biosynthetic inhibitor, were performed to identify the specific sterol biosynthetic step defective in dwf7-1 plants. In a preliminary experiment, the effects of compactin and MVA on the growth of Arabidopsis seedlings in liquid media were investigated. The growth of wild-type Arabidopsis seedlings was almost completely inhibited in the presence of 10 µM compactin. The inhibition, however, was restored to the level of controls by the simultaneous application of 4.5 mM of MVA (data not shown). Therefore, 4.5 mM ¹³C-MVA and 10 µM compactin were added to Arabidopsis seedling cultures in the metabolic feeding studies. After 11 days in culture, sterols were extracted and purified by silica and octadecylsilane (ODS) cartridge columns and ODS-HPLC. Purified samples were derivatized and analyzed by gas chromatography–mass spectrometry (GC-MS). As shown in Figure 7, ¹³C-MVA was converted to ¹³C₅-episterol and subsequent sterols, such as ¹³C₅–24-MC and ¹³C₅-CR in the wild type. However, the ¹³C₅–5-dehydroepisterol and downstream compounds were not detected in dwf7-1 mutants, whereas the precursor ¹³C₅-episterol accumulated fourfold as compared with the wild type. In addition, an uncommon sterol, ¹³C₅–7-dehydrocampestanol (24-epifungisterol), greatly accumulated (Figure 7). Two lines of evidence—a failure to convert episterol to subsequent sterols, such as 24-MC and CR, and accumulation of 7-dehydrocampestanol in dwf7-1—suggest that the defective step in dwf7-1 is the C-5 desaturation stop.

View larger version (12K): [in this window] [in a new window]   Figure 7. GC-MS Analysis of Wild-Type and dwf7-1 Seedlings Fed with 13C-MVA in the Presence of Compactin, an Inhibitor of MVA Biosynthesis. Accumulation of episterol with a simultaneous decrease of downstream intermediates, including 24-MC and CR, predicts that the C-5 desaturation step is blocked in dwf7-1 plants. The units are in micrograms per 5 g fresh weight of tissue. The designation ND (not detected) means that the quantity is lower than the detection limit. Ws-2, Wassilewskija-2 wild type.

Molecular Characterization of dwf7
An EMS-induced mutant (ste1-1) of STE1 encoding a ⁷ sterol C-5 desaturase did not possess a dwarf phenotype (Gachotte et al. 1995 ). However, because it is likely that ste1-1 is a leaky allele, we hypothesized that dwf7-1 might be a strong or null allele. We first sequenced the genomic DNA of the STE1 gene and identified two introns and three exons by comparing them with the published STE1 cDNA sequence. The organization of the STE1 gene is represented schematically in Figure 8. Sequencing the STE1 locus in the dwf7 alleles revealed mutations. The mutations found in dwf7-1 and dwf7-2 were located in the third and the first exons, respectively. Both of the dwf7 alleles contained a base change from a guanine to an adenine, converting tryptophan (TGG) to a stop codon (TAG in dwf7-1 and TGA in dwf7-2).

View larger version (8K): [in this window] [in a new window]   Figure 8. Schematic Representation of the STE1 Gene. Comparison of cDNA and genomic DNA sequences revealed three exons (thick boxes) and two introns (horizontal bars). The single open reading frame encodes a protein of 281 amino acids. The dwf7-2 (ste1-3) mutation is located in the first exon, changing a tryptophan to a stop codon. The dwf7-1 (ste1-2) mutation also changes a tryptophan to a stop codon (amino acid position 230). The three white boxes indicate the transmembrane domains, and the three histidine boxes are lightly shadowed. The figure is drawn to scale by using the GCK software (Textco, Inc., West Lebanon, NH). Bar = 120 bp.

In addition to creating a stop codon, the mutation in dwf7-1 eliminated a HaeIII restriction enzyme recognition site (GGCC to AGCC). Taking advantage of this restriction enzyme site change, we tested the linkage of this mutation to the dwf7-1 phenotype. DNAs isolated from 17 different dwarf plants from a segregating F₂ population were subjected to polymerase chain reaction (PCR) analysis by using S5D_3F and S5D_1R primers (underlines were used to distinguish forward or reverse primers from the gene acronym S5D), and the PCR products were digested with HaeIII. Agarose gel electrophoresis definitively showed that none of the PCR products from 17 mutant templates was restricted, whereas products from wild-type templates were all restricted at the HaeIII site (data not shown). These data suggest that the creation of the premature stop codon in exon 3 is the cause of the dwf7-1–conferred phenotype.

To better understand the importance of these nonsense mutations, we analyzed the sequence of STE1 in relation to other C-5 desaturase proteins isolated from fungi. The STE1 protein is composed of 281 predicted amino acids with a theoretical pI of 6.39 and molecular mass of 33 kD. Whereas yeast ERG3 (38% identical; Arthington et al. 1991 ; GenBank accession number M62623) is predicted to contain four transmembrane domains, STE1 possesses three putative transmembrane domains. The overall amino acid sequence identities of STE1 with C-5 desaturases from fission yeast (GenBank accession number AB004539) and Candida glabrata (Geber et al. 1995 ; GenBank accession number L40390) were 37 and 33%, respectively (gap creation weight of 4; gap extension weight of 1). In addition, multiple sequence alignment of STE1 with the three yeast sequences, shown in Figure 9, revealed that the transmembrane domains and histidine clusters, which were first reported by Gachotte et al. 1996 , are well conserved between the proteins. The three characteristic histidine boxes flank the last transmembrane domain. The nonsense mutations are located in the first exon (dwf7-2) and the third exon, immediately before the third histidine box (dwf7-1), indicating that at least one histidine domain is deleted in each of the dwf7 mutants as a result of the premature stop codons.

View larger version (92K): [in this window] [in a new window]   Figure 9. Multiple Sequence Alignment of DWF7/STE1 with Known Sequences for 7 Sterol C-5 Desaturases. GenBank accession numbers for the sequences are M62623 (S. cerevisiae), AB004539 (Schizosaccharomyces pombe), L40390 (C. glabrata), and AF105034 (DWF7/STE1, Arabidopsis). The conserved transmembrane domains and histidine clusters are boxed and labeled. The positions of the premature stop codons in dwf7-1 and dwf7-2 are indicated with filled circles. Histidine residues in each conserved histidine box are identified with filled triangles. A consensus sequence is shown in the bottom row of the alignment. Capital letters stand for residues conserved among all sequences, whereas lowercase letters mean 50% identical. Dashes indicate gaps introduced to maximize alignment. Multiple sequence alignment was performed using PILEUP in the Genetics Computer Group software (Madison, WI) with a gap creation penalty of 4 and a gap extension parameter of 1. The annotation of the aligned sequences was performed using the ALSCRIPT software (Barton 1993 ).

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

In this report, we present morphological, biochemical, and molecular analysis of Arabidopsis dwf7 mutants. Morphologically, dwf7-1 plants display a dramatic reduction in the length of many different organs examined, and this size reduction is attributable to a defect in cell elongation. Biochemically, dwf7-1 hypocotyls were converted completely to wild-type length with the application of BL, suggesting a deficiency in BRs. In agreement with this, BR intermediate feeding analysis, accompanied by analysis of endogenous levels of BRs and sterols by using GC-SIM, indicated that dwf7-1 is defective specifically in the ⁷ sterol C-5 desaturase step of the sterol biosynthetic pathway. Sequencing of the ⁷ sterol C-5 desaturase gene in dwf7-1 and dwf7-2 revealed premature stop codons, suggesting loss-of-function mutations. Thus, we propose that a shortage of sterols leads to a drastic reduction of BR levels in dwf7-1 and dwf7-2 and to the characteristic dwarf phenotype.

BRs Regulate Cell Elongation and Cell Differentiation
The overall morphology of plants is dependent on three factors: cell size, shape, and number (Cosgrove 1997 ). Various signals modulate these factors. Environmental signals, such as water, temperature, and light, are transduced to invoke internal hormone signals, including auxins, gibberellins, and BRs. These signals then trigger the cell elongation process, including but not limited to cell wall loosening by xyloglucan endotransglycosylases and expansins. Thus, a block in any of the signal transduction cascades from the environmental signals to the cell elongation process could result in dwarfism. Mutants resistant to or deficient in classic hormones, such as auxin (e.g., auxin resistant2 [axr2]; Timpte et al. 1992 ) and gibberellin ([ ga1 to ga5 and gai ]; Koornneef and van der Veen 1980 ; Koornneef et al. 1985 ), often result in dwarfism. Thus, we first tested whether dwf7 is either rescued by or resistant to exogenous application of these hormones. Three-week-old dwf7-1 plants sprayed with 0.1 mM GA₃ responded, as did the wild type (<10% increase of inflorescence height; data not shown); however, GA₃ did not rescue the dwf7-1 phenotype. In addition, dwf7-1 roots grown on indole acetic acid–supplemented agar media (0.1 µM) displayed stunted morphology similar to that of the wild type, suggesting that dwf7-1 is not resistant to the exogenous application of auxin (data not shown). The reduction of hypocotyl length in dwf7-1 was rescued by the application of BL (Figure 5). Both wild-type and dwf7-1 plants responded to BL, but dwf7-1 plants were hypersensitive. The length of dwf7-1 hypocotyls was increased 160% in response to 1 nM BL as compared with the untreated control, whereas the wild type responded marginally (5%). In addition, application of BRs to 3-week-old dwf7-1 plants induced the growth of many different organs, including stems, leaves, siliques, petioles, and pedicels, suggesting that the major defect in dwf7-1 is a deficiency of BL.

Apart from a reduction in cell elongation, a deficiency of endogenous BRs resulted in altered organization of vascular tissue in the inflorescence (Figure 4E). Szekeres et al. 1996 showed that the number of xylem cells in cpd was decreased as compared with the wild type, whereas the number of phloem cells was increased. The authors reasoned that this could be due to unequal division of cambial cells. Furthermore, previous studies on the effects of BRs on vascular development indicated that BRs play a role in tracheary element formation (Clouse and Zurek 1991 ; Iwasaki and Shibaoka 1991 ). Because BRs also have been found in the cambial region of pine, indicative of an important role in this tissue (Kim et al. 1990 ), we hypothesize that the deficiency of BRs in dwarf mutants caused changes in cell fate in vascular cambial cells through yet unknown mechanisms.

Auxins also are known to be a major factor affecting differentiation of the vascular system (Aloni 1987 ). Lincoln et al. 1990 showed that stem cross-sections of axr1 displayed altered development of the vascular system. The vascular bundles in axr1 mutants are located peripherally and are not as regularly spaced as compared with those in wild-type plants (Lincoln et al. 1990 ). Furthermore, as opposed to the reduced number of vascular bundles in dwf7-1 (five to seven), axr1 plants possess a greater number of bundles (eight to nine) as compared with the wild type (six to eight). Thus, it seems that auxins and BRs play opposing roles in determining the number of vascular bundles. Two other assays in which auxin and BR interactions have been demonstrated are the rice lamina bending assay and hypocotyl hook opening bioassay. Results from these assays include the fact that the degree of effect caused by the combined application of auxin and BR was greater than was the sum of the effect of each, indicative of a synergistic effect of the two hormones (Yopp et al. 1981 ; Takeno and Pharis 1982 ; reviewed in Mandava 1988 ). However, the details of the mechanisms for interactive and independent action remain to be elucidated.

It needs to be pointed out that hypocotyl growth in darkness is accomplished through both GA- and BR-dependent cell elongation processes. One piece of evidence for dependence on both GA and BR is that dwf7-1 hypocotyls elongated fivefold in response to darkness as compared with light-grown hypocotyls (data not shown), although they are still shorter than those of the wild type. Because BL levels are not detectable in dwf7-1 plants (Table 2), growth of dwf7-1 in the dark could be accomplished mostly by GA-dependent cell elongation processes. Peng and Harberd 1997 and Azpiroz et al. 1998 found that both gai and dwf4, respectively, partially suppressed the stem elongation phenotype of a light receptor mutant, hy, suggesting that hypocotyl elongation in the absence of light inhibition requires independent growth contributed by both GA and BRs.

dwf7 Plants Are Defective in Sterol Biosynthesis Leading to BRs
A defect either in a biosynthetic enzyme or a factor modulating an enzymatic activity could lead to deficiency of endogenous BRs. To place dwf7 at a specific step in the proposed BR biosynthetic pathway, we first chose to perform feeding studies with BR biosynthetic intermediates. Rescue of dwf7-1 by exogenous application of 22-OHCR suggests that the biosynthetic defect likely resides before the production of CR (Figure 6A). Consistent with the results from feeding studies, the endogenous levels of 24-MC, CR, and CN were extremely reduced in dwf7-1 (Table 2). These data indicate that the biosynthetic defect is before 24-MC; dwf7-1 contains only 3% of 24-MC as compared with the wild type. When the phenotypes of dwf7-1 are compared with the downstream biosynthetic mutant dwf4 and the BR-insensitive bri1 (dwf2) mutant (Clouse et al. 1996 ; S. Choe and K.A. Feldmann, unpublished data), it is obvious that dwf7-1 displays a weaker phenotype despite being a presumptive null mutation. This suggests that there could be an alternative sterol and BR biosynthetic pathway or that there are duplicate genes at individual steps. Providing evidence for the duplicate gene hypothesis, we recently cloned a homolog of the DWF7/STE1 gene (named HOMOLOG OF DWF7, HDF7). HDF7 is 80% identical in amino acid sequence with STE1 (A. Tanaka, S. Choe, and K.A. Feldmann, unpublished data). Similarly, Fujioka et al. 1997 reported that the endogenous level of CN in det2, which is defective in a step between CR and CN, is ~10% that of the wild-type amount. The authors hypothesized that the 10% leakage through the defective step in det2 mutants, even in a null allele, could be associated with a second copy of DET2 that lightly hybridizes in DNA gel blot analyses.

Placing dwf7 at a single sterol biosynthetic step was accomplished through feeding studies with ¹³C-MVA and compactin. A greater than fourfold accumulation of episterol accompanying the absence of downstream intermediates in dwf7-1 indicates that the ⁷ sterol C-5 desaturase step is blocked in dwf7. In addition, the feeding studies identified an accumulation of 7-dehydrocampestanol, which is an uncommon sterol in plants (Figure 7). Accumulation of this compound only in dwf7-1 suggests that sterol biosynthesis in dwf7-1 could proceed to a C-24 reduction step, skipping C-5 desaturation as well as the next immediate C-7 reduction. The C-24 reductase seems to convert episterol independently of the immediate upstream enzyme. The absence of a detectable amount of C-7–reduced compounds in dwf7-1 suggests that the enzymatic step is highly dependent on the C-5 desaturation reaction. This confirms the sequence of reactions originally proposed by Taton and Rahier 1991 , Taton and Rahier 1996 .

dwf7 Mutants Produce Nonfunctional ⁷ Sterol C-5 Desaturase Due to Premature Stop Codons
The ⁷ sterol C-5 desaturase–mediated reaction is common to both photosynthetic and nonphotosynthetic organisms. Many genes encoding a C-5 desaturase have been cloned from fungi. First, Arthington et al. 1991 cloned the ERG3 gene from Saccharomyces cerevisiae. The authors found that viable erg3 mutants, which normally accumulate ⁷ sterols, were restored to wild-type phenotype when transformed with a wild-type genomic clone of the ⁷ sterol C-5 desaturase gene. Taguchi et al. 1994 showed that the yeast mutant syr1 displays dual phenotypes, resistance to the phytotoxin syringomycin and susceptibility to higher concentrations of Ca²⁺, presumably due to altered membranes. Sequencing the ERG3 locus in the syr1 mutant revealed that syr1 is an allele of ERG3. Furthermore, Geber et al. 1995 cloned both ERG3 and ERG11 (14-sterol-demethylase) from C. glabrata. The authors found that lethal erg11 mutations can be suppressed by an additional mutation in erg3. They reasoned that formation of toxic 3ß,6-diol sterols in erg11 mutants is prevented due to the defect in C-5 desaturation in erg11 erg3 double mutants.

In plants, Gachotte et al. 1995 found that the Arabidopsis ste1-1 mutant, which is deficient in C-5 desaturated sterols, can be partially complemented by the yeast ERG3 gene. Accordingly, the authors hypothesized that ste1-1 possesses a mutation in the sterol C-5 desaturase gene. They isolated the Arabidopsis C-5 desaturase gene through heterologous complementation of a yeast erg3 null mutant with an Arabidopsis cDNA library (Gachotte et al. 1996 ). Finally, the partial human cDNA for the C-5 desaturase has been identified by Matsushima et al. 1996 . Alignment of the sequences of these enzymes revealed that C-5 desaturases from different organisms are highly conserved in overall sequence as well as in specific domains. The overall amino acid sequence identity and similarity among STE1 and ERG3 and the human ortholog is 38% (50%) and 35% (47%), respectively (similarity within parentheses). As indicated in Figure 8 and Figure 9, key domains including the transmembrane domains and the histidine clusters are well conserved between all the C-5 desaturases.

Closely spaced histidine residues, HX₃H in helices, serve as typical metal binding motifs in many proteins (Regan 1993 ). Shanklin et al. 1994 showed that three membrane-associated bacterial enzymes, fatty acid desaturase, alkane hydroxylase, and xylene monooxygenase, possess eight histidine residues that are conserved in three regions dispersed in these enzymes, HX_(3-4)H, HX_(2-3)HH, and HX_(2-3)HH (where X stands for any amino acid). DNA constructs containing site-directed mutations at any of these eight histidine residues of the rat ⁹ desaturase failed to complement the yeast mutant ole1, which is defective in the same enzymatic step, suggesting that the individual histidine residues are essential for the function of the enzyme. On the basis of these observations, Shanklin et al. 1994 hypothesized that the histidine clusters conserved in these enzymes constitute new structural domains of diiron binding centers (Shanklin et al. 1994 ). Gachotte et al. 1996 first recognized the conserved histidine clusters in STE1 and yeast proteins. We confirmed that the motifs are highly conserved in STE1 and the yeast ERG3 enzymes with the same context of HX₃H, HX₂HH, and HX₂HH (Figure 9), revealing the presence of a putative iron binding motif in ⁷ sterol C-5 desaturases.

More direct evidence of metal ion involvement in ⁷ sterol C-5 desaturase function was obtained by Taton and Rahier 1996 . These authors discovered that the enzyme prepared from maize microsomes is inhibited by cyanide, whereas it is insensitive to carbon monoxide, indicative of the involvement of a metal ion, presumably an iron, for the proper function of the enzyme. Furthermore, we noticed that the typical histidine moiety also was conserved in a different group of oxidases such as RANP-1 (Uwabe et al. 1997 ), C-4 methyl sterol oxidase (Li and Kaplan 1996 ), and aldehyde decarbonylase (Aarts et al. 1995 ). Occurrence of these histidine boxes in a wide variety of oxidases indicates that this domain plays a common and essential role in the function of membrane oxidases. Therefore, it is likely that the mutations in dwf7-1 and dwf7-2 would be deleterious to protein function. The premature stop codon in dwf7-2 would eliminate all important known domains, whereas the third histidine box and several amino acid residues that are 100% conserved in the C terminus of the protein are eliminated in dwf7-1. Intriguingly, the location of the mutations in dwf7-1 and dwf7-2 seems to be related to the phenotypic severity of the mutant alleles. dwf7-2, which contains an earlier stop codon, was shorter in height and less fertile than dwf7-1. A more precise comparison between the two alleles is not possible because the EMS allele, dwf7-2, has not been outcrossed to remove any background mutations that might have increased the severity of the phenotype of dwf7-2. Despite the differences in severity, both dwf7 alleles would be predicted to be complete loss-of-function alleles. The resulting nonfunctional enzyme would cause a block in sterol biosynthesis. This shortage of substrate sterols in dwf7-1 and dwf7-2 would lead to a deficiency of endogenous BRs and cause the characteristic dwarfism in dwf7 plants.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Plant Growth
For sterile growth of Arabidopsis thaliana plants, seeds of mutants and the wild type were sterilized (50% Clorox and 0.005% Triton X-100) for 8 min, washed three times with sterile distilled water, and dried with 95% ethanol. The seeds were sprinkled on 0.8% agar-solidified media or in liquid media containing 1 x Murashige and Skoog 1962 salts and 0.5% sucrose (pH 5.8 with KOH). For the plants grown in the dark, the seeds on the plates were illuminated for 3 hr (240 µmol m^-2 sec^-1) before being wrapped with two or three layers of aluminum foil. For the mature plants used for morphometric analysis and gas chromatography–selective ion monitoring (GC-SIM) studies, seeds were planted on soil (Metromix 350; Grace Sierra Co., Milpitas, CA) presoaked with distilled water. The flats containing the pots were covered with plastic wrap and cold-treated at 4°C for 2 days before transfer to a growth chamber (16 hr of light [240 µmol m^-2 sec^-1] and 8 hr of dark at 22 and 21°C, respectively, and 75 to 90% humidity). The plastic wrap was removed after 2 to 3 days. The pots were subirrigated in distilled water or Hoagland's nutrient solution as required.

Morphometric and Physiological Analysis
At 5 weeks of age, the various morphological traits listed in Table 1 were measured. The number of seeds per silique was determined after the plants were completely dried. Unopened siliques from each plant were selected and crushed, and the number of seeds was counted under a dissecting microscope. To measure the fresh and dry weight, we cut the aerial parts of the plants and immediately weighed them to obtain the fresh weight; the plants were then completely dried in a 60°C oven for 5 days before measuring the dry weight. Observations on the structure of flowers were made with flowers at stage 14 (Smyth et al. 1990 ), which are right beneath the cluster of developing flowers at the shoot apices. Individual organs of a flower were separated under the dissecting microscope. The length of the organs was measured to a tenth of a millimeter, and the four longest stamens for each flower were measured and the mean value calculated.

The anatomical studies using a scanning electronic microscope and a light microscope were performed as described by Azpiroz et al. 1998 .

Mapping and Sequencing of the DWARF7 Locus
The mapping of dwarf7 (dwf7) was performed using simple sequence length polymorphism (SSLP) markers (Bell and Ecker 1994 ). Briefly, dwf7-1 mutants (Wassilewskija-2 [Ws-2] background) were crossed to Columbia wild-type plants. Genomic DNA was isolated (Dellaporta et al. 1983 ) from individual F₂ dwarf plants. To locate the mutation to one of the five chromosomes, 20 individual plants were tested with at least two SSLP markers per chromosome. The polymerase chain reaction (PCR) amplified products were analyzed on 4% agarose gels in 1 x TAE buffer (40 mM Tris-acetate and 10 mM EDTA). Once the dwf7-1 mutation was shown to be linked to the nga162 marker located on chromosome 3 (recombination ratio 11.9%), we tested marker nga172, which maps at 2.2 centimorgans. No recombination was detected between the dwf7-1 mutation and nga172 when 86 chromosomes were tested, suggesting that dwf7-1 is linked closely to the nga172 marker. Linkage between the markers and the dwarf phenotype was determined according to Koornneef and Stam 1992 .

PCR products amplified using primer sets derived from the cDNA sequence of STEROL1 (STE1) were subjected to sequencing. To design sets of primers that do not fall in exon–intron junctions, we predicted possible splice sites by using the RNASPL program available at the Internet site of Baylor College of Medicine (Houston, TX; http://dot.imgen.bcm.tmc.edu:9331/seq-search/gene-search.html). Primers were designed using the Primer Selection software of DNAstar (DNASTAR Inc., Madison, WI). Oligonucleotide sequences 5' to 3' are CAGTGTGAGTAAT T TAGCAT TACTA (S5D_FF), GGAAAGATCATC-AAACAT T TACATGT (S5D_LR), GCGCAATCT TCT T TCGT T T (S5D_1F), TGGACAACAACAACACAAGA (S5D_1R), GATGCACAGAGAGCT- TCATGAC (S5D_2F), CCGGCAAATGGAGAGAGTGTAT (S5D_2R), CACCCATCATATCTACAACAA (S5D_3F), and CATCT T T TGCCG-GCGAATCTAT (S5D_4F) (underlines were added to distinguish forward or reverse primers from the gene acronym S5D). Primers were purchased from Genosys Biotechnologies, Inc. (The Woodlands, TX). For template DNA, genomic DNA was isolated from two or three leaves of dwf7-1 and wild-type plants according to the method described by Krysan et al. 1996 . Amplification of the DNA fragment spanning the whole coding region was performed with the S5D_4F and S5D_1R primer set with Taq polymerase (Boehringer Mannheim).

Standard PCR reaction mixtures, 1 x PCR buffer (10 mM Tris-HCl, 1.5 mM MgCl₂, and 50 mM KCl, pH 8.3), 0.2 µM each of forward and reverse primer, 0.2 mM each deoxynucleotide triphosphates, 1 ng of genomic DNA, and 2 units of Taq polymerase were subjected to a PCR program consisting of an initial denaturation at 95°C for 2 min and then for 35 cycles (95°C for 30 sec, 56°C for 30 sec, and 72°C for 2.5 min ), with a final elongation step of 7 min at 72°C. PCR-amplified DNA was size-separated on 0.8% agarose gels in 1 x TAE, and the resulting DNA bands were gel-purified using a DNA purification kit (Bio-Rad). The concentration of the extracted DNA was measured by comparing the band intensity with a DNA mass standard (Bethesda Research Laboratories). Sequencing of the DNA was performed at the Arizona Research Laboratory (University of Arizona, Tucson). DNA sequence analysis was conducted using software packages, including one from Genetics Computer Group (Madison, WI) and other database search tools available on the Internet.

The base change in dwf7-1 eliminated the recognition site for a restriction enzyme HaeIII by converting the sequence from GGCC to AGCC. Thus, we utilized this polymorphism to test the cosegregation of the dwarf phenotype with the mutation. The 0.8 kb of DNA spanning the mutation was amplified using S5D_3F and S5D_1R primers from 17 different dwarf plants from the mapping lines. Two microliters from each 20 µL of PCR-amplified DNA was digested with the restriction enzyme HaeIII (Boehringer Mannheim). After complete digestion, the samples were resolved on a 2% agarose gel in 1 x TAE buffer.

Genomic DNA sequence flanking the cDNA was identified by sequencing the products obtained from thermal asymmetric interlaced PCR (TAIL PCR) (Liu et al. 1995 ). Two sets of primers were used to amplify the 5' and 3' flanking DNA. Oligonucleotide sequences 5' to 3' are GTAGAAGCACCAGAGGAAACCGGAGATGAAGT (D7-5-1; melting temperature of 69°C), AAGTATAGTAGGGT TCCGGCGAGG-TA (D7-5-2; melting temperature of 64°C), ATAGAT TCGCCG-GCAAAAGATGACTC (D7-5-3; melting temperature of 63°C), TGC-AGGATACCATACGATACACCACACGACAT (D7-3-1; melting temperature of 68°C), CATACGATACACCACACGACATACAAGCAT-AACTA (D7-3-2; melting temperature of 67°C), and ATATGGATG-GAT TGGATGT T TGGCTCTC (D7-3-3; melting temperature of 63°C). The melting temperature of each primer was calculated with the formula 69.3 + 0.41 (%GC) - 650/L (Mazars et al. 1991 ), where L is length of primer. Arbitrary degenerate primers AD1, AD2, and AD3 were synthesized according to the sequence described by Liu et al. 1995 . TAIL PCR was performed according to the program originally described by Liu et al. 1995 . TAIL PCR–amplified DNA was separated on 1% agarose gels and gel extracted for sequencing.

Feeding Experiments
Biochemical complementation of dwf7-1 plants with different concentrations of brassinolide (BL) was performed in liquid media. BL-supplemented (control, 10^-9, 10^-8, and 10^-7 M) sterile liquid media (1.5 mL) was dispensed into wells of a 24-well plate (Corning Co., Corning, NY). Three seedlings, germinated on agar-solidified media, were transferred into each well. After a week of growth with continuous shaking (230 rpm), the seedlings were lightly stained with toluidine blue, and hypocotyls and roots were measured to the nearest millimeter.

Feeding experiments using biosynthetic intermediates were performed with 3-week-old mutant plants. The intermediates tested were diluted to the desired concentration with water containing 0.01% Tween 20. Two microliters of each brassinosteroid (BR) solution was applied daily to the shoot tips of plants by using a micro pipettman. After 1 week of treatment, total growth of inflorescence and pedicels was measured to the nearest millimeter (n = 15).

Analysis of Endogenous BRs
Plants were grown for 5 weeks on soil. Two hundred grams of the aerial parts of plants, including stems, flowers, leaves, and siliques, was harvested and subjected to BR extraction. The procedure for extraction and analysis of BR intermediates by using GC-SIM has been described (Fujioka et al. 1997 ).

¹³C-Labeled Mevalonic Acid Feeding Experiments
Before feeding experiments, seedlings were germinated and grown on 0.5 x Murashige and Skoog agar medium in the light at 22°C (25 mL per dish). Eight days after sowing, the seedlings were transferred to a 200-mL flask containing 30 mL of Murashige and Skoog media supplemented with 3% sucrose (Ws-2, five seedlings; dwf7-1, 40 seedlings).

Compactin (mevastatin; Sigma) was converted to its sodium salt as described previously (Kita et al. 1980 ). DL-Mevalonolactone-2–¹³C (¹³C-MVA; Isotec, Miamisburg, OH) was dissolved in methanol. Solutions of compactin and ¹³C-MVA were added aseptically to each 200-mL flask (final concentration, 10 µM compactin and 4.5 mM ¹³C-MVA) just after the seedlings were transferred, and seedlings were allowed to grow for 11 days at 22°C in the light on a shaker (110 rpm). After incubation, the seedlings (~5 g fresh weight of both Ws-2 and dwf7-1 plant materials) were extracted with methanol (250 mL), and the extract was partitioned between CHCl₃ and H₂O. The CHCl₃-soluble fraction was purified with a silica cartridge column (Sep-Pak Vac 12 cc; Waters, Milford, MA), which was eluted with 20 mL of CHCl₃. The eluate was purified with an octadecylsilane (ODS) cartridge column (Sep-Pak PLUS C18; Waters), which was eluted with 20 mL of methanol. The fraction was subjected to HPLC on an ODS column as follows: column, Senshu Pak ODS 4150-N (150 x 10 mm); solvent, methanol; flow rate, 2 mL /min; and detection, UV 205 nm. Fractions were collected every 0.5 min (between retention times of 10 to 20 min). Main fractions of each sterol were as follows: 5-dehydroepisterol (retention time of 11.5 to 12 min), episterol (retention time of 12.5 to 13 min), 24-methylenecholesterol (24-MC; retention time of 13 to 13.5 min), 7-dehydrocampestanol (retention time of 14.5 to 15 min), and campesterol (CR; retention time of 15.5 to 16 min).

Each fraction was converted to a trimethylsilyl derivative and analyzed by gas chromatography–mass spectrometry (GC-MS). GC-MS analyses were performed on a JEOL Automass JMS-AM 150 mass spectrometer (Tokyo, Japan) connected to a Hewlett-Packard 5890A-II gas chromatograph with a capillary column DB-5 (0.25 mm x 15 m; 0.25-µm film thickness). The analytical conditions were the same as previously described (Fujioka et al. 1997 ).

5-Dehydroepisterol, episterol, and 7-dehydrocampestanol were chemically synthesized (S. Takatsuto, C. Gotoh, T. Noguchi, and S. Fujioka, unpublished data).

	ACKNOWLEDGMENTS

We thank Brian P. Dilkes, Alice Traut, and Xuelu Wang for their technical assistance. We thank Pierre Benveniste and his collaborators for sharing unpublished data on ste1-1. S.C. thanks the Korean government for support as an overseas scholar (Grant No. 93-0019). A.T. thanks the JAERI and Science and Technology Agency of Japan. We also thank the staffs of the sequencing and imaging facilities in the Arizona Research Laboratory at the University of Arizona (Tucson). S.F. acknowledges support from RIKEN (President's Special Research Grant) and from a grant-in-aid for scientific research (B) from the Ministry of Education, Science, Sports, and Culture of Japan (Grant No. 10460050). K.A.F. acknowledges support from the National Science Foundation (Grant No. 9604439).

Received November 11, 1998; accepted November 18, 1998.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Aarts, M.G.M., Keijzer, C.J., Stiekema, W.J., and Pereira, A. (1995) Molecular characterization of the CER1 gene of Arabidopsis involved in epicuticular wax biosynthesis and pollen fertility. Plant Cell 7:2115-2127[Abstract].

Aloni, R. (1987) Differentiation of vascular tissue. Annu. Rev. Plant Physiol. 38:179-204[CrossRef][ISI].

Arthington, B.A., Hoskins, J., Skatrud, P.L., and Bard, M. (1991) Nucleotide sequence of the gene encoding yeast C-8 sterol isomerase. Gene 107:173-174[CrossRef][ISI][Medline].

Azpiroz, R., Wu, Y., LoCascio, J.C., and Feldmann, K.A. (1998) An Arabidopsis brassinosteroid-dependent mutant is blocked in cell elongation. Plant Cell 10:219-230[Abstract/Free Full Text].

Bach, T.J., and Benveniste, P. (1997) Cloning of cDNAs or genes encoding enzymes of sterol biosynthesis from plants and other eukaryotes: Heterologous expression and complementation analysis of mutations for functional characterization. Prog. Lipid Res. 36:197-226[CrossRef][Medline].

Bak, S., Kahn, R.A., Olsen, C.E., and Halkier, B.A. (1997) Cloning and expression in Escherichia coli of the obtusifoliol 14--demethylase of Sorghum bicolor (L.) Moench, a cytochrome P450 orthologous to the sterol 14--demethylases (CYP51) from fungi and mammals. Plant J. 11:191-201[CrossRef][ISI][Medline].

Barton, G.J. (1993) ALSCRIPT, a tool to format multiple sequence alignments. Protein Eng. 6:37-40[Free Full Text].

Bell, C.J., and Ecker, J.R. (1994) Assignment of 30 microsatellite loci to the linkage map of Arabidopsis. Genomics 19:137-144[CrossRef][ISI][Medline].

Benveniste, P. (1986) Sterol biosynthesis. Annu. Rev. Plant Physiol. 37:275-308[CrossRef].

Bouvier-Navè, P., Husselstein, T., Desprez, T., and Benveniste, P. (1997) Identification of cDNAs encoding sterol methyl-transferases involved in the second methylation step of plant sterol biosynthesis. Eur. J. Biochem. 246:518-529[Medline].

Choe, S., Dilkes, B.P., Fujioka, S., Takatsuto, S., Sakurai, A., and Feldmann, K.A. (1998) The DWF4 gene of Arabidopsis encodes a cytochrome P450 that mediates multiple 22-hydroxylation steps in brassinosteroid biosynthesis. Plant Cell 10:231-243[Abstract/Free Full Text].

Choi, Y.H., Fujioka, S., Nomura, T., Harada, A., Yokota, T., Takatsuto, S., and Sakurai, A. (1997) An alternative brassinolide biosynthetic pathway via late C-6 oxidation. Phytochemistry 44:609-613[CrossRef].

Clouse, S.D., and Zurek, D. (1991). Molecular analysis of brassinolide action in plant growth and development. In Brassinosteroids: Chemistry, Bioactivity and Applications, H.G. Cutler, T. Yokota, and G. Adam, eds (Washington DC: American Chemical Society), pp. 122–140.

Clouse, S.D., Langford, M., and McMorris, T.C. (1996) A brassinosteroid-insensitive mutant in Arabidopsis thaliana exhibits multiple defects in growth and development. Plant Physiol. 111:671-678[Abstract].

Corey, E.J., Matsuda, S.P.T., and Bartel, B. (1993) Isolation of an Arabidopsis thaliana gene encoding cycloartenol synthase by functional expression in a yeast mutant lacking lanosterol synthase by the use of a chromatographic screen. Proc. Natl. Acad. Sci. USA 90:11628-11632[Abstract/Free Full Text].

Cosgrove, D.J. (1997) Relaxation in a high-stress environment: The molecular bases of extensible cell walls and cell enlargement. Plant Cell 9:1031-1041[CrossRef][ISI][Medline].

Dellaporta, S.L., Wood, J., and Hicks, J.B. (1983) A plant DNA minipreparation: Version II. Plant Mol. Biol. Rep. 1:19-21.

Fujioka, S., and Sakurai, A. (1997) Brassinosteroids. Nat. Prod. Rep. 14:1-10[CrossRef][ISI][Medline].

Fujioka, S., Inoue, T., Takatsuto, S., Yanagisawa, T., Yokota