BAT1, a putative acyltransferase, modulates brassinosteroid levels in Arabidopsis


For correspondence (e-mail


Brassinosteroids (BRs) are essential for various aspects of plant development. Cellular BR homeostasis is critical for proper growth and development of plants; however, its regulatory mechanism remains largely unknown. BAT1 (BR-related acyltransferase 1), a gene encoding a putative acyltransferase, was found to be involved in vascular bundle development in a full-length cDNA over-expressor (FOX) screen. Over-expression of BAT1 resulted in typical BR-deficient phenotypes, which were rescued by exogenously applied castasterone and brassinolide. Analyses of BR profiles demonstrated that BAT1 alters levels of several brassinolide biosynthetic intermediates, including 6-deoxotyphasterol, typhasterol and 6-deoxocastasterone. BAT1 is mainly localized in the endoplasmic reticulum. BAT1 is highly expressed in young tissues and vascular bundles, and its expression is induced by auxin. These data suggest that BAT1 is involved in BR homeostasis, probably by conversion of brassinolide intermediates into acylated BR conjugates.


Brassinosteroids (BRs) are structurally similar to the cholesterol-derived steroid hormones of animals and the ecdysteroids of insects (Grove et al., 1979), and are essential in various plant developmental processes, including cell elongation and division, skotomorphogenesis, vascular differentiation, reproductive development and seed germination (Bishop and Koncz, 2002). BRs are synthesized from campesterol, a sterol synthesis product, that is eventually converted to biologically active castasterone (CS) and brassinolide (BL) by a series of hydroxylations, reductions, oxidations and epimerizations (Fujioka and Yokota, 2003; Bishop, 2007). Expression levels of key BR biosynthetic genes are principally regulated by BRs via a negative feedback mechanism (Mathur et al., 1998; Bancos et al., 2002; Müssig et al., 2002; Tanaka et al., 2005; Kim et al., 2006). For example, DEETIOLATED2 (DET2), DWARF4 (DWF4), CONSTITUTIVE PHOTOMORPHOGENESIS AND DWARFISM (CPD), BR-6-oxidase1 (BR6ox1) and ROTUNDFOLIA3 (ROT3) were up-regulated in BR-depleted Arabidopsis plants, and expression of DWF4, CPD, BR6ox1 and ROT3 was reduced by exogenous BL treatment. As the expression levels of these BR biosynthetic genes are correlated directly with the rate of BR biosynthesis (Tanaka et al., 2005), the feedback regulation of BR biosynthesis in response to BR is apparently essential for control of BR action.

Excessive amounts of BRs may be controlled through modification, which results in loss of activity and a decrease in bioactive BR pools (Fujioka and Yokota, 2003). Of 42 BR metabolites identified from various plant species or predicted from naturally occurring metabolic intermediates (Clouse and Sasse, 1998; Fujioka and Yokota, 2003), 19 were found to be conjugated to sugar or fatty acids in plants (Bajguz, 2007).

Three types of modification processes are known to be involved in the inactivation of BRs. BR inactivation may be mediated by sulfonation at the 22-OH group of BRs, especially 24-epicathasterone (24-epiCT), by a steroid sulfotransferase named BNST3 from Brassica napus (Rouleau et al., 1999).

Hydroxylation is another means of BR inactivation, and is catalyzed by the cytochrome P450 mono-oxygenase BAS1 (cytochrome P450 phyB activation-tagged suppressor 1-dominant)/CYP72B1/CYP734A1 (Neff et al., 1999). Over-expression of BAS1 in Arabidopsis and tobacco plants resulted in BR-deficient phenotypes, which correlated with production of inactive 26-hydroxycastasterone (26-OHCS) and 26-hydroxybrassinolide (26-OHBL) (Turk et al., 2003). A BAS1 homolog, SHK1 (CYP72C1/SOB7/CHI2, DLF), is also thought to be involved in the hydroxylation and inactivation of BRs (Kim et al., 2005; Nakamura et al., 2005; Takahashi et al., 2005; Turk et al., 2005). Whereas BAS1 preferentially binds to CS and to a lesser extent BL, as well as the inactive precursors typhasterol (TY) and teasterone (TE), SHK1 shows little or no binding affinity for active BRs, but displays significant binding to TY and other precursor molecules (Thornton et al., 2010).

Another class of inactive BR conjugates is the glucoside class. In Arabidopsis, conjugation of CS and BL with glucose is catalyzed by UGT73C5, a UDP-glycosyltransferase (UGT) (Poppenberger et al., 2005). An increase in 23-O-glucosylation activity in UGT73C5 over-expressing plants correlated with BR-deficient phenotypes, which were rescued by application of 24-epiBL. UGT73C6, the closest homolog of UGT73C5, also catalyzes 23-O-glucosylation of CS and BL in planta (Husar et al., 2011). The levels of bioactive BRs are tightly regulated by BR catabolism as well as the relative rate of BR biosynthesis; however, the process of BR conjugate formation remains largely unknown.

Here we show that BAT1, a putative acyltransferase, alters the levels of brassinolide intermediates in Arabidopsis, which may provide another regulatory system for BR homeostasis.


Identification of FOX hunting transgenic lines with altered vasculature development

To identify novel factors controlling vascular bundle development, we used the FOX (full-length cDNA overexpressor) hunting system (Ichikawa et al., 2006). We observed transverse inflorescence stem sections of 4500 5-week-old transgenic lines, and screened them for an altered pattern or number of vascular bundles. One of these FOX lines, labeled F23231, was particularly interesting because it showed a decreased number of vascular bundles in its inflorescence stem compared to the wild-type stem (Figure 1a). This line exhibited dwarf phenotypes such as small and rounded dark-green rosettes with shortened leaf petioles (Figure 1b). The phenotypes of the F23231 line were already apparent at the seedling stage, and became most pronounced after plants had formed their first true leaves. At maturity, F23231 plants resembled BR-deficient mutants, such as bri1-5, a weak BRI1 allele (Noguchi et al., 1999), in both height and rosette width, both which are approximately half those of wild-type plants (Figure 1c).

Figure 1.

BR-deficient phenotypes of F23231 FOX (full-length cDNA overexpressor) Arabidopsis lines. (a) F23231 plants have a reduced number of vascular bundles compared to Col-0. The base of the inflorescence stems was transversely sectioned when the first silique appeared. (b) F23231 plants have rounded, dark-green rosette leaves and shortened petioles compared to Col-0. (c) At maturity, the F23231 plant showed dwarfism similar to the bri1-5 mutant. (d) Transcript levels of BR biosynthetic genes CPD, DWF4 and ROT3. Total RNAs were isolated from the leaves of 4-week-old Arabidopsis lines and subjected to quantitative real-time RT-PCR. The expression levels of each BR biosynthetic gene relative to ACTIN2 in each Arabidopsis line were normalized to those in Col-0. ACTIN2 was used as an internal control. Values are means ± SE of triplicates. Asterisks indicate significant differences from wild-type (*< 0.05; **< 0.01; Student's t-test). Scale bars = 100 μm (a) and 1 cm (b,c).

As the morphological phenotype of BR deficiency in the F23231 line indicated a possible change in BR biosynthesis or levels, we examined the transcript levels of BR biosynthetic genes using quantitative real-time PCR analysis. We found that expression of CPD, DWF4 and ROT3 was highly elevated in this transgenic line compared to wild-type levels (Figure 1d).

BAT1 encodes a putative acyltransferase

To define the molecular nature of the F23231 phenotype, PCR-amplified genomic DNA was isolated from the transgenic plant using T-DNA-specific primers (Ichikawa et al., 2006). Approximately 1.5 kb of PCR fragment was sequenced, and matched a gene encoding a putative HXXXD-type acyltransferase family protein (At4g31910). We subsequently confirmed over-expression of At4g31910 in the F23231 line (Figure S1a), which may cause its BR-deficiency phenotype, and named At4g31910 as BAT1 (putative BR-related acyltransferase). The BAT1 gene consists of three exons encoding a protein of 458 amino acids that contains the putative acyltransferase domain (Figure 2a). To confirm that BAT1 is responsible for the BR-related phenotypes, we generated transgenic lines that over-express a full-length version of the BAT1 gene in Col-0 background. We found that BAT1 over-expression reverses the BR-deficient dwarf phenotype observed in the original FOX hunting line (Figure S1b). These results suggest that increased expression of BAT1 is related to the dwarf phenotypes, which may be due to changes in BR homeostasis or signaling.

Figure 2.

bat1-1, a BAT1 null mutant, exhibits enhanced growth of inflorescence stems. (a) Schematic presentation of T-DNA insertions in bat1 mutant lines. Black boxes and lines represent exons and introns, respectively. The T-DNA insertions are located inside exons (blue triangle) and introns (purple triangle). Four homozygous SALK T-DNA insertion lines were isolated and named bat1-1, bat1-2, bat1-3 and bat1-4 (see Experimental procedures). (b) bat1-1 plants have longer stems than Col-0 plants. The phenotypes were observed at 36 days after germination. (c) The bat1-1 mutant has longer inflorescence stems than Col-0 and BAT1 over-expressing lines. The lengths of primary stems were measured every day after bolting. Values are means ± SE of approximately 20 individual plants. Asterisks indicate significant differences from wild-type (*< 0.05; **< 0.01; Student's t-test). (d) The stem diameters and numbers of vascular bundles of bat1-1 are increased compared to those of Col-0. The bases of the inflorescence stems were sectioned when the first silique emerged. Scale bars = 100 μm.

To dissect the functional significance of BAT1, we isolated T-DNA insertion lines for At4g31910. Four lines (i.e. bat1-1, bat1-2, bat1-3 and bat1-4) were obtained as homozygotes. Of these lines, bat1-1 contains an insertion in the first exon of the BAT1 gene (Figure 2a), which abolished expression of BAT1 (Figure S1c). The bat1-1 plants had longer stems than wild-type or BAT1 over-expressing plants before flowering but reached a height similar to that of wild-type plants just after flowering (Figure 2b,c). In addition, the bat1-1 plants exhibited larger inflorescence stems, with an increased number of vascular bundles compared to wild-type stems (Figure 2d). However, there was no other apparent difference in morphology between bat1-1 and wild-type plants after flowering.

BAT1 localizes to the endoplasmic reticulum in young tissues and the phloem of vascular bundles

To assess the involvement of BAT1 in BR homeostasis and vascular bundle development, we first examined its subcellular localization using GFP-tagged BAT1 proteins and confocal laser scanning microscopy. BAT1-GFP co-localized with RFP-tagged BiP, an ER marker protein, and also localized in the nucleus (Figure 3a). We further analyzed subcellular fractions containing the ER using anti-GFP antibodies for BAT1-GFP and anti-calreticulin (CRT) antibodies for the ER marker (Figure S1d). Consistent with BAT1-GFP imaging, BAT1 proteins were detected in the ER fraction. We then determined the expression pattern of BAT1 using transgenic lines harboring a transcriptional reporter construct (proBAT1:GUS) in which 2.0 kb of the BAT1 promoter was fused to the β-glucuronidase (GUS) gene. Promoter activity was found to be present in most organs, especially young tissues and vascular bundles (Figure 3b–d). In young seedlings, GUS staining was mainly observed in the vascular cylinder of cotyledons, leaves and hypocotyls (Figure 3b). The activity of the BAT1 promoter was also highly detected in root tip, lateral root initiation sites and the maturation zone of the primary root (Figure 3c). In adult plants, the BAT1 promoter was active in the vasculature of the inflorescence stem, especially in phloem cells (Figure 3d). These tissue-specific expression patterns also overlap with those of key BR biosynthetic genes, including CPD and ROT3 (Mathur et al., 1998; Bancos et al., 2002; Shimada et al., 2003; He et al., 2005; Kim et al., 2005, 2006). These data, together with its putative biochemical nature as an acyltransferase, suggest a possible function of BAT1 in regulating BR homeostasis in the endoplasmic reticulum.

Figure 3.

Subcellular localization of BAT1 proteins and histochemical analyses of BAT1 expression. (a) BAT1-GFP co-localized with BiP-RFP in protoplasts. BAT1 and BiP fused to either GFP or RFP were transiently expressed in Arabidopsis mesophyll protoplasts. After 12 h incubation, fluorescent signals were visualized using confocal laser scanning microscopy. (b) BAT1 promoter activity was detected in veins of cotyledons (left panel) and leaves (right panel). (c) BAT1 promoter activity was detected in maturation zones of the primary root (top left panel), lateral root primordia (centre and bottom left panel) and the root tip (right panel). Two-week-old Arabidopsis seedlings expressing proBAT1:GUS were used to monitor BAT1 expression. (d) BAT1 expression was abundantly detected in the vasculature of inflorescence stem, especially the phloem (ph) of 5-week-old plants expressing proBAT1:GUS. Scale bars = 20 μm (a), 50 μm (c) and 100 μm (d).

Increased expression of BAT1 causes a reduction in the endogenous levels of BRs

The predicted biochemical function of BAT1, the BR-related phenotypes of over-expressing transgenic plants, and its localization and expression patterns suggested that BAT1 may have a role in the modification or homeostasis of BRs. To investigate possible biochemical consequences of BAT1 activity, endogenous levels of BRs were determined in aerial tissues of 5-week-old plants by GC-MS analysis using deuterium-labeled internal standards. Compared with the endogenous BR levels in wild-type plants, several major BR synthetic intermediates were drastically reduced in BAT1 over-expressing transgenic lines (Figure 4). In BAT1 over-expressing plants, endogenous levels of BR intermediates downstream of campestanol (CN), especially 6-deoxocathasterone (6-deoxoCT) and 6-deoxo-3-dehydroteasterone (6-deoxo3DT), were increased 1.8- and 1.9-fold, respectively, compared to wild-type levels. Interestingly, the endogenous BR intermediates of the early C6 oxidation pathway in these transgenic plants were decreased to 63% of wild-type levels for TY. Furthermore, 6-deoxoTY and 6-deoxoCS intermediates of the late C6 oxidation pathway in BAT1 over-expressing lines were decreased to 47% and 63%, respectively, compared to the wild-type. Active BL was not detectable in either wild-type or transgenic lines. In contrast, bat1-1 knockout plants did not show any statistically significant alterations in BR levels. These data clearly suggest that BAT1 alters BR levels, which may be caused by acylation of BR biosynthetic intermediates including 6-deoxoTY, 6-deoxoCS and TY, but not 6-deoxoCT, 6-deoxoTE or 6-deoxo3DT.

Figure 4.

Endogenous levels of BRs in Col-0, the BAT1 over-expressing line, and bat1-1. Aerial parts of 5-week-old plants were analyzed for free BR levels using GC-MS (see Experimental procedures). The contents (ng/g fresh weight) of BR intermediates from three independent sets of samples are displayed. n.d., not detected (below the detection limit). CR, campesterol; CN, campestanol; 6-oxoCN, 6-oxocampestanol; 6-deoxoCT, 6-deoxocathasterone; CT, cathasterone; 6-deoxoTE, 6-deoxoteasterone; TE, teasterone; 6-deoxo3DT, 6-deoxo-3-dehydroteasterone; 3DT, 3-dehydroteasterone; 6-deoxoTY, 6-deoxotyphasterol; TY, typhasterol; 6-deoxoCS, 6-deoxocastasterone; CS, castasterone; BL, brassinolide. Asterisks indicate significant differences from wild-type (*< 0.05; **< 0.01; Student's t-test).

Active BRs, CS and BL rescue the shortened hypocotyls and dwarf phenotype of BAT1 over-expressing lines

To determine whether the BR-deficient phenotype of BAT1 over-expressing plants was caused by perturbation in BR signaling or BR metabolism, we examined whether the phenotype observed could be rescued by application of exogenous BR intermediates. The transgenic lines over-expressing BAT1 had shorter hypocotyls than wild-type and bat1-1 plants under both light and dark conditions. The shortened hypocotyls of BAT1 over-expressing plants were rescued by exogenous application of active BRs, CS and BL (Figure 5a). The lengths of BAT1 over-expressing hypocotyls became similar to those of wild-type hypocotyls after treatment with 0.3 μm of CS or BL under long-day conditions. Interestingly, the shortened hypocotyls of BAT1 over-expressing lines were partially rescued by 3 μm TY, but not by TE even at high concentrations (Figure 5a). In the dark, application of 10 nm BL restored the lengths of shortened hypocotyls of the transgenic plants to those of wild-type plants, similar to the response of BR-deficient det2-1 seedlings to exogenously applied BL (Figure S2a). Only active BRs, BL and CS rescued the dwarf phenotypes of soil-grown mature plants (Figure 5b). Rescued plants exhibited much longer inflorescence stems, and slightly larger, less green and curled leaves than mock-treated plants (Figure 5b). These results suggest that the BR-deficient phenotypes of BAT1 over-expressing plants are caused by a reduction of active endogenous BR levels. To confirm this, we performed a genetic complementation analysis by crossing our BAT1 over-expression line with the DWF4 over-expressing line, which contains high levels of endogenous BRs (Choe et al., 2001; Wang et al., 2001; Nemhauser et al., 2004). The architecture of the aerial parts of parental DWF4 transgenic plants was larger than that of the wild-type, with an increased number of branches and siliques. The crossed F1 plants over-expressing both BAT1 and DWF4 showed wild-type phenotypes, with a reduced height of inflorescence stems compared to the DWF4 transgenic plants (Figure 5c), indicating that BAT1 may modulate the high endogenous BR levels in DWF4 over-expressing plants.

Figure 5.

Active BRs rescue BR-deficient phenotypes of the BAT1 over-expressing line. (a) Restored hypocotyl elongation of BAT1 over-expressing plants by BR. The hypocotyl lengths of light-grown seedlings in the media containing 0.3 or 3 μm of BL, CS, TE or TY were measured. Values are means ± SE of approximately 20 individual seedlings. Asterisks indicate significant differences from the control (*< 0.05; **< 0.01; Student's t-test). (b) Active BRs restore the BR-deficient phenotype of BAT1 over-expressing plants to wild-type phenotype. Three-week-old BAT1 over-expressing plants grown in soil were sprayed twice a week with 1 μm of CS or BL solution for another 3 weeks. (c) BAT1 over-expression suppresses the BR-related phenotype of DWF4 over-expressing lines. BAT1 over-expressing lines were genetically crossed with a DWF4 over-expressing line that over-accumulates BRs. Representative 4-week-old Arabidopsis plants are shown. Scale bars = 5 mm (b), 5 cm (c) and 2 cm (d).

BAT1 expression is up-regulated by auxin

As the expression of crucial BR biosynthesis genes is differentially regulated at different BR levels (Mathur et al., 1998; Bancos et al., 2002, 2002; Goda et al., 2002; Müssig et al., 2002; Tanaka et al., 2005; Kim et al., 2006), we examined whether the action of BAT1 is controlled at the transcriptional level by BL and other phytohormones. The expression of BAT1 was slightly, but not significantly, up-regulated by epiBL (Figure 6a and Figure S2b). However, interestingly, BAT1 expression was significantly induced by auxin. As the BAT1 promoter is activated by auxin and contains putative auxin response factor (ARF) binding sites (Figure S2c), we examined whether auxin-dependent BAT1 activation is mediated by a canonical TIR1/ARF signaling cascade (Dharmasiri et al., 2005; Turk et al., 2005). To this end, Arabidopsis protoplasts were transformed with HA-tagged ARF7, 8 or 19, and chromatin associated with ARF proteins was isolated using monoclonal anti-HA antibodies. This experiment showed that the BAT1 promoter directly bound to ARF19, but not to ARF7 or 8 (Figure 6b), indicating a key role for ARF19 in BAT1 regulation. Consistent with this observation, expression of BAT1 was substantially decreased in the loss-of-function arf19-1 mutant (Figure 6c). Furthermore, auxin-dependent activation of BAT1 was nearly abolished in tir1-1 (Figure 6d), a loss-of-function mutant of the TIR1 auxin receptor, (Kepinski and Leyser, 2005), suggesting a role for auxin in BAT1 transcription.

Figure 6.

Auxin induces BAT1 expression. (a) Transcripts of BAT1 were significantly elevated by auxin treatment. Arabidopsis seedlings were grown in half-strength MS liquid medium for 7 days under continuous light conditions, and then treated with hormones for designated time periods. IAA, indole acetic acid; epiBL, epi-brassinolide. Values are means ± SE of 20 individual seedlings. Asterisks indicate significant differences from the control (*< 0.05; Student's t-test). (b) ARF19 binds to the BAT1 promoter. ARF19-HA was transiently expressed in Arabidopsis mesophyll protoplasts for 5 h, and samples were subjected to chromatin immunoprecipitation using anti-HA antibodies. Upper panel: Quantitative real-time RT-PCR was performed using BAT1 promoter-specific primers spanning the ARF binding sites (see Experimental procedures). Lower panel: Precipitated DNA was also amplified by PCR and resolved in a 1% agarose gel. input, PCR products from total lysed chromatin DNA; output Myc, PCR products from immunoprecipitated DNAs with anti-Myc antibodies; output + HA, PCR products from immunoprecipitated DNAs with anti-HA antibodies. (c,d) Expression of BAT1 decreases in auxin signaling impaired mutants. Expression levels of BAT1 in arf19-1, a loss-of-function mutant of the ARF19 transcriptional factor that functions in auxin signaling (c), and in tir1-1, a dominant-negative allele of the TIR1 auxin receptor (d). The expression levels of BAT1 in the presence of IAA were normalized to the levels in the absence of IAA. Values are means ± SE of relative expression levels for three biological replicates. Asterisks indicate significant differences from the control (**< 0.01; Student's t test). (e) Kinetics for expression changes of a set of BR homeostasis-related genes upon auxin treatment over various time courses. Seven-day-old Col-0 seedlings were treated with 1 μm IAA for the designated time and harvested for RNA isolation. BAS1 and SHK1, hydroxylation-mediated BR inactivation genes; UGT73C5 and UGT73C6, glucosylation-mediated BR activation genes; BRX, a positive regulator of BR biosynthesis. In (a), (c) and (d), ACTIN2 was used as an internal control.

To understand the dynamics of BR homeostasis by auxin, we examined the kinetics of changes in expression of a set of BR-related genes. Upon auxin treatment, BRX, a positive regulator of BR biosynthesis (Mouchel et al., 2006), reached its highest induction 30–60 min after auxin treatment and then gradually diminished (Figure 6e). BAS1, encoding a BR inactivation enzyme involved in hydroxylation of active BRs, showed a similar activation pattern to BRX upon auxin application. However, at 3 h after auxin treatment, levels of the BAS1 transcript started to increase again. The transcript of SHK1, the closest homolog of BAS1, was highly induced at 3 h after auxin application, and this induction was sustained until 24 h after auxin treatment. However, the induction of BAT1 was slightly delayed, reaching its peak at 60–90 min after auxin treatment. Interestingly, induction of the UGT73C5 and UGT73C6 genes that encode the UDP-glycosyltransferases involved in BR inactivation by glucosylation began to slightly increase 3 h after auxin application and then decreased. These results suggest that BAT1 may be involved in BR homeostasis, together with other BR regulatory enzymes, under the control of auxin signaling, but with a distinctive temporal response from other BR-inactivating regulators.


Plants maintain extremely low levels of endogenous BRs, and BR homeostasis is tightly controlled for normal growth and development (Bancos et al., 2002; He et al., 2005; Kim et al., 2006). Unlike other plant hormones, although distributed widely throughout the plant, BRs are unlikely to be transported long distances (Symons and Reid, 2004), suggesting that endogenous BR levels are strictly regulated by local control of BR biosynthesis and metabolism in target cells or tissues. The tight control of key metabolic regulations is of interest as their effect on plant growth responses is immediate and rapid, adjusting required BR levels in response to environmental or physiological signals (Bancos et al., 2002; Shimada et al., 2003; Lisso et al., 2005; Tanaka et al., 2005). Conjugated compounds may serve as a pool of inactive phytohormones that are rapidly converted to active forms by de-conjugation reactions (Bajguz and Tretyn, 2003; Piotrowska and Bajguz, 2011). Some conjugates are thought to be temporary storage forms from which free active hormones may be released by hydrolysis after receiving the appropriate signals.

Here we identified a putative acyltransferase, BAT1, as a protein that regulates BR-specific growth responses and endogenous BR levels. BR profile analysis revealed that BAT1 over-expression substantially reduces the endogenous levels of several BR biosynthetic intermediates, including 6-deoxoTY, TY and 6-deoxoCS, which is directly related to the BR-deficient phenotype found in the over-expression lines. Taken together with the predicted function of BAT1 as an acyltransferase, we propose that BAT1 is involved in BR inactivation through acylation of BR side-chains, which may provide another regulatory mechanism for BR homeostasis. In previous studies, acyl-conjugated BRs were isolated (Asakawa et al., 1994, 1996). Recently, a member of the acyltransferase family, BIA1, was found to be involved in controlling BR levels, probably via acyl conjugation of BRs in Arabidopsis (Roh et al., 2012). This type of BR modification is often accompanied by conjugation to a fatty acid, such as lauric acid, myristic acid and palmitic acid, at a number of functional carbon positions. For example, 3,24-diepiCS and 3,24-diepiBL were found to be conjugated to 3-laurate, 3-myristate or 3-palmitate ester in serradella cells (Ornithopus sativus) (Kolbe et al., 1994, 1996). TE and its derivatives may be conjugated to 3-laurate and 3-myristate, creating several acyl metabolites in lily (Lilium longiflorum) cell cultures (Abe et al., 1994; Asakawa et al., 1994, 1996; Kolbe et al., 1997; Soeno et al., 2000a,b). These acyl BRs are easily converted to free TE, CS and BL intermediates by hydrolysis. Acyl conjugates may serve as reversible deactivated storage forms, which may be important for regulation of physiologically active BR levels. Conjugation of BRs has been shown to be important for BR deactivation and may involve various BR intermediates as substrates throughout the BR biosynthesis pathway (Bajguz and Tretyn, 2003; Piotrowska and Bajguz, 2011). It is also possible that acyl conjugates directly interfere with BR perception by competing with active BRs, or, alternatively, BR acylation may be required for its irreversible deactivation, transport, compartmentalization and stability. However, we have not yet determined the catalytic activity of BAT1 because of technical difficulties. Due mainly to the lack of labeled internal standards of acylated BRs, it is almost impossible to quantify acylated BRs in vivo, especially with respect to their distribution within different plant tissues or organs.

Co-localization of BAT1 with an ER marker clearly indicates its role in BR homeostasis in the endoplasmic reticulum. It has been shown that the rate-determining enzymes in BR biosynthesis, such as DIM/DWF1, BR6ox1 and DWF4, are also located in the endoplasmic reticulum (Schuler, 1996; Choe et al., 1998; Klahre et al., 1998; Shimada et al., 2001; Bancos et al., 2002; Kim et al., 2006), suggesting that the ER is a major organelle in BR metabolism. Previous studies have shown that root tissues contain the lowest amount of the endogenous BRs (Clouse and Sasse, 1998); however, the expression levels of several key BR biosynthetic genes, such as DWF4 and ROT3, are relatively high in roots (Shimada et al., 2003). Interestingly, BR metabolic genes, such as BAS1 (Shimada et al., 2003), SHK1 (Takahashi et al., 2005) and BAT1 (Figure 3c), are known to be highly expressed in roots, which may be responsible for maintaining a low level of active root BRs (Yuan et al., 2007). The expression of BAT1 is developmentally regulated, and is especially enriched in the vascular tissues, in which the CPD and ROT3 genes involved in BR biosynthesis are preferentially expressed (Mathur et al., 1998; Kim et al., 2005). BAT1 is also expressed to a greater extent in the root elongation zone, but not in the cell division zone, which may contribute to the regulatory mechanisms necessary for maintaining optimal BR levels of the root system, in which high levels of BRs are favorable for cell division and unfavorable to root elongation (Bancos et al., 2002; Shimada et al., 2003). Therefore, our results suggest the possibility that BR homeostasis may be adjusted by both BR biosynthesis and metabolism/conjugation in the same tissues or even in the same subcellular region of plant cells, which may be necessary for the tight regulation of local BR levels, as BRs do not undergo long-distance transport (Symons and Reid, 2004).

Combinatorial control of a small number of plant hormones may provoke a wide range of growth responses (Clouse and Sasse, 1998), and some of these hormones interact through biosynthetic regulation (Vogel et al., 1998; Collett et al., 2000; Ghassemian et al., 2000; Bouquin et al., 2001). It is known that BR biosynthesis is controlled by negative feedback regulation (Mathur et al., 1998; Bancos et al., 2002, 2002; Müssig et al., 2002; Tanaka et al., 2005; Kim et al., 2006) and stimulated by auxin signaling in roots (Mouchel et al., 2006; Chung et al., 2011). Auxin utilizes synthesized BRs for some of its growth-promoting effects in Arabidopsis. However, in this study, BAT1 expression was significantly up-regulated by auxin, but not directly by BL (Figure 6a and Figure S2b). The BAT1 promoter has a conserved auxin-responsive element (AuxRE) to which ARF proteins bind (Ulmasov et al., 1999). The ARF7 and ARF19 proteins act redundantly in various auxin-mediated plant responses (Okushima et al., 2005, 2007; Wilmoth et al., 2005; Li et al., 2006). However, unlike root tissue, the aerial portion of ARF7 over-expressing plants is indistinguishable from that of the wild-type (Hardtke et al., 2004), while the ARF19 over-expressing plants show a dwarf phenotype with narrower, elongated and mis-shapen leaves (Okushima et al., 2005, 2007), similar to the BR-deficient phenotype of BAT1 over-expressing plants. These data, together with direct binding of ARF19 to the BAT1 promoter (Figure 6b), suggest a specific role of ARF19 as a transcriptional regulator for auxin-mediated BAT1 activation. In the auxin-signaling mutants tir1-1 and arf19-1, the auxin-dependent response of BAT1 disappeared (Figure 6c,d). These data clearly demonstrate a role for auxin in BAT1 regulation, and show that auxin targets BAT1, whose product deactivates BR, as well as other BR biosynthetic genes, to adjust the BR levels required for optimal growth responses in vasculature development, leaf morphogenesis, inflorescence stem and hypocotyl elongation. In other words, auxin reduces the pool of active BRs through the activity of BAT1, creating a negative feedback mechanism. Given the biochemical nature of predicted acyltransferases for unknown substrates, it is also possible that BAT1 may modulate the activity of other proteins by acylation, which may be decisive for regulatory outputs in BR homeostasis or other cellular processes involved in regulation of BR biosynthetic or metabolic enzymes. However, we were unable to test these possibilities in this study. Nevertheless, it is clear that BAT1 provides another layer of strict regulation contributing to BR homeostasis. The modulation of BR metabolism, and eventually bioactive BRs, is known to be directly related to grain yield and biomass production in plants (Choe et al., 2001; Sakamoto et al., 2005, 2011; Morinaka et al., 2006; Wu et al., 2008). In this sense, BAT1 may be a useful target to control the level of bioactive BRs and thereby increase crop productivity. However, it is of note that additional factors, including other hormones and environmental stimuli, not only auxin and BR, may be involved in BR homeostasis, leading ultimately to proper regulation of plant growth and development processes.

Experimental procedures

Plant material and growth conditions

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was used as the wild-type in all experiments performed in this study. The FOX hunting Arabidopsis lines (Ichikawa et al., 2006) were used for screening of alterations in the vasculature. Transgene in the FOX collection was identified using a pair of oligonucleotides (GS4, 5′-GTACGTATTTTTACAACAATTACCAACAAC-3′, and GS6, 5′-GGATTCAATCTTAAGAAACTTTATTGCCAA-3′), and then subjected to sequencing. Plants were grown in soil or in half-strength Murashige & Skoog (MS) medium containing 0.8% w/v agar under 16 h light/8 h dark at 23°C. Four SALK T-DNA knockout mutant lines of At4g31910 (SALK_123922, SALK_109656, SALK_137638 and SALK_094563) were obtained from the Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH), and named as bat1-1, bat1-2, bat1-3 and bat1-4, respectively.

Histological analysis of vasculature development

Histological analysis was performed as previously described, with minor modifications (Hejátko et al., 2007). Fresh stem sections were cut by hand from the base of the inflorescence stem when the first silique appeared. The stems were fixed for 3 h in 3% glutaraldehyde in 0.1 m phosphate buffer (pH 7.2), and then rinsed twice with 0.1 m phosphate buffer (pH 7.2) and dehydrated in graded ethanol. Prepared stems were embedded in Spurr's resin (Ted Pella, for 48 h at 65°C. Sectioned specimens (0.5 or 2 μm) were cut using an MT-X ultramicrotome (RMC,, stained using 0.05% toluidine blue, and then photographed under a Zeiss Axioplan2 microscope. For fix-free staining, handcut sections were prepared using a razor blade from the base of the inflorescence stems when the first silique appeared. Then, sections were stained with toluidine blue (0.05% w/v solution in water) for 30 sec, destained in distilled water for 30 sec, mounted in 50% glycerol, and observed under a BX61 microscope (Olympus, using differential interference contrast optics.

Subcellular localization and histochemical analyses

The full-length coding region of BAT1 was amplified from Arabidopsis cDNAs, fused to GFP sequences, and inserted into a plant expression vector containing the 35SC4PPDK promoter and the NOS terminator (Hwang and Sheen, 2001). GFP-tagged BAT1 plasmid (20–40 μg) was co-transfected with BiP-RFP, an ER marker, into 4 × 104 mesophyll protoplasts from 4-week-old plants. The transfected cells were incubated under constant light at 23°C for 12 h. GFP and RFP fluorescence was observed using a confocal laser scanning microscope (LSM 510 Meta system; Zeiss, Subcellular fractions containing the ER were obtained from protoplasts by a modified fractionation protocol (Lee et al., 2011). Transformed protoplasts were resuspended in 4 ml HMS buffer (330 mm sorbitol, 50 mm HEPES/KOH pH 7.6 and 3 mm MgCl2), and filtered three times through two layers of 11 μm nylon filters (Millipore, for homogenization. To obtain fractions containing the ER, the homogenate was incubated on ice for 30 min in 3 ml HMS buffer supplemented with 20 mm EDTA and 10 mm EGTA, and centrifuged at 900 g for 6 min at 4°C to remove chloroplasts. The supernatant was then loaded onto a two-step sucrose gradient (from bottom to top: 2 ml of 60% sucrose and 4 ml of 36% sucrose in 50 mm HEPES/KOH pH 7.6, 0.1 mm MgCl2 and 3 mm EDTA), and centrifuged at 40 000 g for 90 min at 4°C. ER fractions were collected at the top of the 36% sucrose layer. The fraction was diluted in 9 ml HMS buffer, and organelles were harvested by centrifugation at 40 000 g for 40 min at 4°C.

To monitor BAT1 expression, 2.0 kb of genomic sequence upstream of the BAT1 translation start site was amplified by PCR and cloned into the binary vector pCAMBIA1303 (Hejátko et al., 2007). The proBAT1:GUS reporter construct was introduced into Col-0 Arabidopsis plants via Agrobacterium-mediated transformation (Clough and Bent, 1998). To detect GUS activities, hygromycin-resistant seedlings or tissues of several independent T2 transgenic lines at various developmental stages were incubated in staining solution [100 mm phosphate buffer, pH 7.0, 1 mm X-gluc (Duchefa,, 10 mm EDTA, 0.1% v/v Triton × -100] overnight at 37°C. After staining, the samples were destained by a series of ethanol incubations and visualized under an Axioplan microscope (Zeiss,

Quantitative real-time RT-PCR analysis

Total RNAs were isolated from 7-day-old or 4-week-old plants using TRIZOL reagent (Invitrogen, according to the manufacturer's instructions. DNA contamination of RNA samples was removed using DNAfree (Ambion,, and the first strand of cDNA was prepared using the ImProm-II reverse transcription system (Promega, The subsequent quantitative real-time PCR was performed using gene-specific primers in a Light Cycler 2.0 (Roche, with SYBR Premix Ex Taq system (Takara, ACTIN2 transcript levels were used as an internal control. All quantitative RT-PCR experiments were performed using biologically independent samples at least three times.

Quantification of endogenous BR levels

Fifteen grams (fresh weight) of aerial parts of plants were lyophilized and extracted twice with 500 ml MeOH:CHCl3 (4:1), and deuterium-labeled internal standards were added. Purification and quantification of each BR was performed using GC-MS as described previously (Fujioka et al., 2002).

BR feeding experiments

To examine the effect of BR on the length of hypocotyls, sterilized Col-0, bat1-1 and BAT1 over-expressing seeds were stratified at 4°C for 3 days in the dark. Seedlings were grown vertically for 7 days either in 16 h white light (77 μm m−2 sec−1)/8 h dark or continuous dark conditions in half-strength MS medium containing 0.3–3 μm of typhasterol (TY), teasterone (TE), castasterone (CS) or brassinolide (BL). For the mock control, an equivalent amount of ethanol was added to the medium. Seven-day-old seedlings from each plate were photographed, and the lengths of at least 20 seedling hypocotyls were measured in three independent experiments. To determine the effect of BR on adult plants, Col-0 and BAT1 over-expressing plants were grown in soil for 3 weeks. Twenty milliliters of TY (10−5 or 10−6 m), TE (10−5 or 10−6 m), CS (10−5, 10−6 or 3 × 10−6 m) or BL (10−5, 10−6 or 3 × 10−6 m) solutions containing 0.01% Tween-20 were sprayed on the shoots twice a week. For the mock control, equivalent volumes of Tween-20 (0.01%) were applied. After 3 weeks of BR treatment, inflorescence stems were photographed.

Chromatin immunoprecipitation (ChIP) assay

HA-tagged ARF7, ARF8 or ARF19 were transiently expressed in Arabidopsis protoplasts for 5 h at room temperature. Accumulation of each ARF protein in the protoplasts was examined by separation using 10% SDS-PAGE, followed by Western blot analyses using horseradish peroxidase-conjugated anti-HA antibodies (Cell Signaling, ChIP assays were performed as previously described (Lee et al., 2007) with minor modifications. After fixation of transfected protoplasts using 1% formaldehyde at room temperature for 10 min, the protoplast chromatin was isolated using lysis buffer (1% SDS, 10 mm EDTA, 50 mm Tris/HCl pH 8.1, 1 ×  protease inhibitor cocktail and sheared by sonication (Branson model). The genomic DNA fragments bound to each ARF-HA protein were precipitated using high-affinity anti-HA antibodies (Roche, The BAT1 promoter was amplified by PCR and quantitative real-time PCR using the immunoprecipitated genomic DNA as template with specific primers (Figure S2c).


This work was supported by a grant from the Advanced Biomass R&D Center (ABC) of Korea, funded by the Ministry of Education, Science and Technology (ABC-2011-0028378), and by Institute of Planning and Evaluation for Technology in food, agriculture, forestry and fisheries (309017-5), Ministry for Food, Agriculture, Forestry and Fisheries, Republic of Korea. S.C. was the recipient of a Brain Korea 21 fellowship. We are also grateful to Suguru Takatsuto (Department of Chemistry, Joetsu University of Education, Joetsu-shi, Niigata 943−8512, Japan) for supplying deuterium-labeled internal standards.