Brassinosteroid biosynthesis and inactivation

Authors


  • Edited by V. Hurry

*e-mail: shchoe@snu.ac.kr

Abstract

The term brassinosteroids (BRs) refers to the growth-promoting plant steroidal hormones. Various developmental programs including but not limited to cell elongation, stress tolerance, and skoto-/photo-morphogenesis are controlled by subnanomolar concentrations of BRs. Accordingly, BR mutants that are defective in BR biosynthetic or signaling pathways usually display dwarfism. Characterization of numerous BR dwarf mutants isolated from Arabidopsis, pea, tomato, and rice greatly contributed to our understanding of BR biology. Recently, an enzyme that mediates the final step in the BR biosynthetic pathways has been characterized by two different groups. The brassinolide synthases (Cytochrome P450s 85A2 and 85A3) are multifunctional enzymes that catalyze the last three consecutive steps in BR biosynthetic pathways, namely, C-6 hydroxylation, dehydrogenation, and Baeyer-Villiger type oxidation. In addition, many of the previously unknown steps have been genetically characterized. This review aims to summarize the knowledge that has been developed during the last 2–3 years in this field of BR biosynthesis and inactivation research.

Abbreviations – 
BRs

brassinosteroids

BL

brassinolide

CYP

cytochromes P450

Introduction

Life cycles of plant growth and development are regulated by plant hormones that act alone or, more usually, in concert. Varying the concentrations of bioactive hormones in a cell is an important mechanism by which hormones induce differential levels of responses from cells. The pool sizes of bioactive hormones in a cell are affected by at least two different processes: input by de novo biosynthesis and inward transport, and output through inactivation and sequestration in subcellular organelles. Accumulation or deficiency of plant hormones frequently results in detrimental effects to plant development, implying that the amounts of bioactive hormones in the cell must be kept finely controlled.

The brassinosteroids (BRs) are a group of plant-originated steroidal lactones that exert pronounced growth-promoting activities. BRs are engaged in a flurry of plant developmental aspects, including the stimulation of cell division and elongation, conferring stress tolerance, vascular system differentiation, leaf development, and photo-/skoto-morphogenesis. Plants that are defective in BR biosynthesis or signal transduction pathways display characteristic growth-deficient phenotypes, including short stature, round and curled leaves, short petioles, short pedicles and reduced fertility. In addition, when they are grown in darkness, hypocotyls are short, and the cotyledons are open without an apical hook (Choe 2004). Recently, significant progress has been made in BR biosynthetic research, especially in rice. Morphological alterations of rice BR mutants include short stature, reduction in a second internode, and erect leaves (Hong et al. 2003). Analysis of these rice BR mutants significantly enhanced our understanding of the BR biosynthetic pathways, especially the enzymatic steps that had not been discovered by Arabidopsis research (Hong et al. 2003).

BRs are biosynthesized using plant sterols including campesterol (CR) as pathway precursors (Fujioka and Yokota 2003, Fig. 1). The BR biosynthetic pathways were initially established based on metabolic conversion assay of radio-labeled compounds in brassinolide (BL)-overproducing cell lines of Madagascar periwinkle (Fujioka and Yokota 2003). Recently, numerous mutants that are defective in the biosynthetic enzymes have been recovered from different plants (Bishop 2003, Choe 2004). Biochemical characterization of these mutants facilitated further validation of the pathways, and also discovery of previously unknown steps. Most of the BR biosynthetic steps are mediated by a family of enzymes belonging to Cytochrome P450s (P450s). P450s are heme-thiolate monooxygenases that utilize a flavoprotein system to transfer electrons from NADPH and/or NADH to a substrate (Schuler 1996). P450-catalyzed biochemical reactions include hydroxylation, oxidation, isomerization, and dehydration (Werck-Reichhart and Feyereisen 2000). Plant genomes possess hundreds of P450 genes, while only three are found in Saccharomyces cerevisiae, 80 in Drosophila melanogaster, and 85 in Caenorhabditis elegans (Nelson et al. 2004a). P450s are classified according to the degree of amino acid sequence identity. P450s that share greater than 40% amino acid identity are placed under the same family number, and those more than 55% identity are placed in the same subfamily. If they possess >97% identity, they are considered allelic. Arabidopsis P450s are grouped to 45 families and 72 subfamilies, with 33 subfamilies containing a single member, while one family, CYP71, contains 54 members in just two subfamilies (http://www.p450.kvl.dk). Phylogenetic clustering of the functional P450s showed that 172 of them belong to a single clade, the A-type, while 100 belong to the non-A-type (Paquette et al. 2000). The A-type P450s are important in the biosynthesis of plant-specific metabolites, while the non-A-type carries out the reactions conserved across the kingdoms such as sterol biosynthesis. A-type P450s are likely derived from a common ancestor (Durst and Nelson 1995, Paquette et al. 2000), while the non-A-type P450s are diverse, sharing more homology to non-plant P450s than A-type P450s. Recently, many of the enzymes that are involved in BR biosynthesis in rice were found to belong to P450s, and all these newly found P450s, except pea DDWARF1 (Kang et al. 2001), are classified into non-A-type.

Figure 1.

Figure 1.

Brassinosteroid biosynthetic pathways. (A) pathways from 24-methylenecholesterol to (24R)-5α-Ergostan-3-one, and (B) pathways from campestanol to brassinolide. Depending on the order of hydroxylation reactions, the pathways are bifurcated twice and named early or late oxidation pathways. In the early C-22 oxidation pathway, 24-methylene cholesterol goes through the sterol C-22 hydroxylation reaction prior to the C-5 reduction steps (Fig. 1A). Similarly, in the early C-6 oxidation pathways, hydroxylation at C-6 position of campestanol precedes C-22 hydroxylation reaction (Fig. 1B). The enzymatic reactions that mediate each step are labeled with numbers. Dotted lines in the Fig. 1 (B) represent the reactions that are possibly mediated by CYP85 proteins, however, biochemical confirmation should be performed. The steps with double arrows in Fig. 1 (B) refer to reversible reactions, and the lengths of arrows indicate the enzymatically favourable direction. Reactions 14 and 15 inactivate BRs, its substrate include castasterone as well as brassinolide. Brassinolide is considered as the end product of the pathways, and thus the most bioactive among natural BRs. The carbon atoms for the core rings and a side chain are numbered in a separate box in Fig. 1 (A).

Figure 1.

Figure 1.

Brassinosteroid biosynthetic pathways. (A) pathways from 24-methylenecholesterol to (24R)-5α-Ergostan-3-one, and (B) pathways from campestanol to brassinolide. Depending on the order of hydroxylation reactions, the pathways are bifurcated twice and named early or late oxidation pathways. In the early C-22 oxidation pathway, 24-methylene cholesterol goes through the sterol C-22 hydroxylation reaction prior to the C-5 reduction steps (Fig. 1A). Similarly, in the early C-6 oxidation pathways, hydroxylation at C-6 position of campestanol precedes C-22 hydroxylation reaction (Fig. 1B). The enzymatic reactions that mediate each step are labeled with numbers. Dotted lines in the Fig. 1 (B) represent the reactions that are possibly mediated by CYP85 proteins, however, biochemical confirmation should be performed. The steps with double arrows in Fig. 1 (B) refer to reversible reactions, and the lengths of arrows indicate the enzymatically favourable direction. Reactions 14 and 15 inactivate BRs, its substrate include castasterone as well as brassinolide. Brassinolide is considered as the end product of the pathways, and thus the most bioactive among natural BRs. The carbon atoms for the core rings and a side chain are numbered in a separate box in Fig. 1 (A).

Previous review articles focused on BR biosynthesis, inactivation, and signal transduction pathways (Fujioka and Yokota 2003, Sasse 2003, Choe 2004, Vert et al. 2005). This article aims to review recent advances in the field of BR biosynthesis and inactivation.

BR biosynthesis

Rice BRD2: 24-methylenecampesterol to campesterol, reaction 2

DWARF1 and LKB mediate the C-24 reduction reactions in the early steps of BR biosynthesis in Arabidopsis and pea, respectively (Choe et al. 1999, Nomura et al. 1999). A double bond at C-24 is reduced to a single bond via two consecutive steps consisting of isomerization of Δ24 to Δ22 and subsequent reduction of the C-22 double bond (Choe et al. 1999, Nomura et al. 1999). Recently, Hong et al. (2005) reported that rice brassinosteroid-deficient dwarf2 (brd2) is defective in the rice gene that is homologous to Arabidopsis DWARF1. In the brd2 mutant plants, the endogenous amount of 24-methylenecholesterol, which is a precursor of the C-24 reduction step, accumulated more than 10-fold, whereas amount of the product, campesterol, was 1/1000th compared with wild-type. Interestingly, Hong et al. (2005) also noticed that the overall phenotypes of rice brd2 mutant were milder than those of rice brd1 mutant, which is defective in the BR C6-oxidation step. After examination of the endogenous BR levels, Hong et al. found that dolichosterone (DS) and other C-24 non-reduced BRs are enriched in brd2. Thus, they postulated that the C-24 reduction step could be skipped during BR biosynthesis. After careful bioassay of the rice lamina-bending system, they found that DS is also bioactive and induces typical bending at lamina joints. The accumulation of DS may explain the relatively weaker phenotypes of the brd2 mutants.

CYP90D2: 6-deoxoTE/TE to 3-dehydro-6-deoxoTE/3-dehydroTE (‘/’ separates the intermediates of the late and early C-6 oxidation pathways), reaction 10

Hong et al. (2003) reported that rice BR dwarf mutants including ebisu dwarf (d2) display characteristic phenotypes, such as inhibition of elongation in the second internode, erect leaves, and de-etiolation in the dark. On the basis of successful complementation of the mutant phenotype by exogenous application of BL, it was confirmed that d2 was defective in BL biosynthesis. A map-based cloning approach revealed that D2 encodes a novel member of a cytochrome P450 enzyme, CYP90D2 (Hong et al. 2003). Through a search of the rice genome databases, the authors found another close homolog that shares 66% identity in amino acid sequence level with D2, named CYP90D3. Currently, the CYP90D group consists in three members, two from rice and one from Arabidopsis (CYP90D1, Nelson et al. 2004b). Recently, Kim et al. (2005) proposed that Arabidopsis CYP90D1 is involved in the BR 3-dehydrogenation steps such as 6-deoxoTE/TE to 3-dehydro-6-deoxoTE/3-dehydroTE, based on the results of genetics and endogenous BR level analysis.

To elucidate a specific enzymatic step mediated by CYP90D2 and D3 of rice, Hong et al. (2003) performed intermediate feeding tests. The rice lamina-bending assay showed that only the biosynthetic intermediates 3-dehydroTE (3-dehydro-6-deoxo-TE) and its downstream compounds could induce lamina bending, whereas application of the upstream compounds such as TE and 6-deoxoTE resulted in negligible responses. Furthermore, analysis of endogenous levels of BRs showed that the level of 6-deoxoCT slightly increased, while the 3-dehydro-6-deoxo-TE level was only half that of wild-type. On the basis of the results of endogenous BR levels and feeding tests, it has been proposed that CYP90D2 mediates the 3-dehydrogenase reaction in the BR biosynthetic pathways.

CYP724B1 (DWARF11): 3-dehydro-6-deoxoTE/3-dehydroTE to 6-deoxoTY/TY, reaction 11

Similar to the reduced seed length phenotypes of Arabidopsis BR dwarfs (Choe et al. 2000), rice BR-defective mutants often display a short seed phenotype (Tanabe et al. 2005). To elucidate the underlying mechanisms of this short seed phenotype, Tanabe et al. took a map-based cloning approach to isolate the responsible gene for the dwarf11 (d11) mutant. The amino acid sequence of the cloned gene revealed that it shares 43% identity with Arabidopsis DWF4 (CYP90B1), suggesting a role in BR biosynthesis. To identify the possible enzymatic step in the pathways that this gene product participated in, these authors used the rice lamina-bending assay with various biosynthetic intermediates. In the BR biosynthetic pathways, only the intermediates from 6-deoxoTY and its downstream compounds could induce bending responses from the both d11-1 and d11-2 mutant plants (Tanabe et al. 2005). This suggests that d11 is defective in a step of BR biosynthetic pathways that converts 6-deoxo-3-dehydroTE/3-dehydroTE to 6-deoxo-TY/TY.

Arabidopsis CYP85A2 and tomato CYP85A3: 6-deoxoCS to BL via CS, reaction 13

BL is an end product in the BR biosynthetic pathways and is considered to be the most active compound among approximately 50 BRs discovered to date. Thus, identification of the enzyme that mediates the final step in the BL biosynthetic pathways has been a focal point in this research. Two different groups independently identified this enzyme, which is a Baeyer-Villiger type oxidase. The Yamaguchi group at RIKEN (Nomura et al. 2005) reported that a homolog (CYP85A3) of tomato Dwarf enzyme (CYP85A1) mediates BR-6-hydroxylation as well as the Baeyer-Villiger type oxidation step. The tomato Dwarf gene was initially discovered by Bishop et al. (1996). A transposon-tagged loss-of-function mutant, extreme dwarf (dx), displayed severe dwarfism due to a block in the BR-6-oxidation step, which is the penultimate step in the pathways (Bishop et al. 1996). Cloning of the gene responsible for the dx mutant revealed that a Cytochrome P450 (CYP85A1) is disrupted.

Unexpectedly, Nomura et al. (2005) found that dx fruit contained significant amounts of BL. Because BL is synthesized using CS as a precursor, they hypothesized that an alternative enzyme substituting for CYP85A1 may exist in the dx fruit. To clone this second copy of the CYP85A1 gene active in tomato fruit, they designed a pair of oligonucleotides based on the nucleotide sequence identity conserved in the CYP85 family of genes already known in Arabidopsis, tomato, and pea. PCR amplification of a DNA fragment using the designed oligonucleotides led to the isolation of a gene that shares 75% identity in amino acid sequence level with CYP85A1, which was named CYP85A3. As expected, the spatial expression pattern of the two CYP85 genes differed in tomato tissues; CYP85A3 was preferentially expressed in the fruits whereas CYP85A1 expressed in vegetative tissues. Thus, it was hypothesized that the elevated level of BL in the dx fruit was due to the enzymatic activity of CYP85A3. To test whether CYP85A3 indeed mediates the BR-6-oxidation reaction, CYP85A3 was heterologously expressed in yeast, and feeding experiments were performed. As was the case for CYP85A1 (Bishop et al. 1999), feeding of a deuterium- labeled substrate, [2H6]6-deoxoCS, to the CYP85A3-expressing yeast strain resulted in the production of [2H6]6-CS (Nomura et al. 2005). Surprisingly, further examination of the GC-MS profiles of the reaction products from [2H6]6-deoxoCS feeding experiments showed that in addition to the expected [2H6]6-CS peak, the authors could identify another peak that was similar to [2H6]6-BL. Comparison of the characteristic ions, m/z, from standard BL to that of the metabolic profile confirmed that the peak observed in CY85A3 was the same as BL. This means that CYP85A3 is a multifunctional enzyme that mediates three consecutive enzymatic steps in the BL biosynthetic pathways, namely C-6 hydroxylation, dehydrogenation, and Baeyer-Villiger type oxidation.

Independently, Kim et al. (2005) also found that Arabidopsis CYP85A2 mediates the final step in the BL biosynthetic pathways. Previously, they found that tomato has a significant pool of C27-BRs including 28-norCS, and these C27-BRs are bioactive in various bioassays (Kim et al. 2004). Similarly to the experiments with tomato, it was shown that the Arabidopsis enzyme preparation could convert 28-norCS to CS, suggesting that Arabidopsis has a pathway from cholesterol to BL in addition to the conventional CR to BL pathway (Kim et al. 2004). On the basis of these findings, Kim et al. (2005) hypothesized that CYP85A1 and CYP85A2 from Arabidopsis have different substrate preferences: one for deoxoCS and the other for 28-norCS. To test this hypothesis, they heterologously expressed the two genes in yeast, and performed a metabolite conversion analysis. Results proved that CYP85A2 has significantly greater overall enzymatic activity than CYP85A1, and that CYP85A2 can effectively transform more of 6-deoxo-28-norCS to 28-norCS relative to CYP85A1 (Kim et al. 2005). In addition to the differential enzyme activity, Kim et al. found that CYP85A2-expressing yeast produce a BL-like peak when they feed the cell line with either of the two precursors, 6-deoxo-28-norCS or deoxoCS. Further comparison of the BL-like peak with the profile of authentic BL revealed that the peak was indeed BL that was produced after conversion from the precursors.

There has been debate over whether the nascent BR biosynthetic intermediates are bioactive, i.e. whether intermediates need be converted to BL to get bioactivity, or whether some of BRs such as CS and TY possess nascent bioactivity without being metabolized to BL. Previously, Wang et al. (2001) showed that the BR receptor BRI1 can bind to both CS and BL, suggesting that CS might serve as a bioactive compound. In addition, thorough analysis of endogenous BR levels in different plants showed that rice and pea lack BL (Fujioka and Yokota 2003). Therefore, it has been suggested that both CS and BL are bioactive compounds. Genetic analysis of Arabidopsis mutants that are defective in either CYP85A1 or CYP85A2 gene revealed important findings concerning the nascent bioactivity of CS or BL. When the function of the CYP85A1 gene was disrupted, the mutant did not display any noticeable defects in development (Kim et al. 2005, Kwon et al. 2005, Nomura et al. 2005), possibly due to functional complementation by a redundant gene CYP85A2. However, a loss-of-function mutant for the CYP85A2 gene displayed semi-dwarf phenotype (Kim et al. 2005). Compared with a wild-type plant, rosette leaves look darker-green, rounder, and curled. In addition, filament growth was not sufficient to reach the stigmatic papillae, causing reduced fertility in the mutant (Kim et al. 2005). This suggests that BL deficiency in the cyp85a2 mutant negatively affected the reproductive organ development.

In contrast to the semi-dwarf phenotype of the cyp85a2 mutant, a double mutant defective in the two genes, cyp85a1 and cyp85a2, displayed the conventional BR dwarf phenotype (Kwon et al. 2005, Nomura et al. 2005). The double mutants possessed reduced stature, short petioles and pedicels, and dark green and rounder shape of rosette leaves (Kwon et al. 2005). In addition, the double mutant displayed a shorter hypocotyl length in the dark (Kwon et al. 2005), a characteristic of de-etiolation phenotypes of BR dwarf mutants (Choe 2004). Interestingly, the short hypocotyl in the dark was further reduced in presence of the BR biosynthetic inhibitor brassinazole. In addition, Kwon et al. (2005) also found that the dwarf phenotypes observed in the double mutants were relatively weaker than the cases of dwf7-1 or cpd-388, suggesting that the double mutant may maintain some minor pool of bioactive BRs. Examination of the endogenous BRs in the double mutants revealed that the amount of 6-deoxoCS was quadrupled, whereas the amount of CS was only 1/30th that of wild-type, possibly due to the block in the BR-6-oxidation step. However, the amounts of CN and 6-oxoCN in the double mutant were not significantly different from those of wild-type, suggesting that the BR-6-oxidation of CN (reaction 7) might be mediated by an enzyme other than the two CYP85s in Arabidopsis (Kwon et al. 2005).

The relatively mild growth defects observed in the Arabidopsis cyp85a2 mutant that is defective in the BL synthase suggests that BL might play a specific role. The developmental anomalies in a single mutant of CYP85A2 were most obvious in the reproductive organs (Kim et al. 2005). Similarly, tomato CYP85A3 expression is greatly enriched in reproductive organs and fruits (Nomura et al. 2005). In addition, Shimada et al. (2003) reported that Arabidopsis floral clusters and seeds accumulate significant amounts of BL relative to other organs, such as stems and leaves. Enrichment of BL in reproductive organs and relatively weak overall growth defects in the cyp85a2 single mutant of Arabidopsis suggest that CS is a bioactive compound responsible for overall vegetative growth, whereas BL could be more responsible for the development of reproductive organs. However, it cannot be ruled out that BL plays a role in the development of other vegetative organs. Cell-specific localization of BL and more thorough analysis of the BL synthase gene expression may shed light on the specific roles assigned to BL.

As mentioned earlier, BL has never been detected from rice plants. Thus, it is possible that rice lacks the homolog of BL synthase, CYP85A2. Currently, searching the rice pseudomolecules database (http://www.tigr.org) deduced from rice genome using Arabidopsis CYP85A2 as a probe revealed only one member of the CYP85 family. Heterologous expression of this gene and subsequent characterization should reveal whether this rice protein mediates only BR-6-oxidation.

Steps with unidentified genes

In human cells, a sterol C5α-reduction step is preceded by the isomerization of the double bond from Δ5 to Δ4 position. This isomerization reaction consists of two functionally dividable reactions: steroid 3-dehydrogenation and subsequent isomerization of the double bond from Δ5 to Δ4 position. The enzyme for this dual function is well studied in human and mouse. A searching of Arabidopsis database (TAIR, http://www.Arabidopsis.org) revealed four genes that have domains similar to the human 3-β hydroxysteroid dehydrogenase/isomerase gene. These include At1g47290, At2g26260, At2g33630, At2g43420. In addition, similar searches of the TIGR rice pseudomolecules database (http://www.tigr.org) resulted in three different genes, LOC_Os02g48460, LOC_Os03g29150, LOC_Os03g29170, and LOC_Os09g34090 that have similarity to the human gene. Currently, it is not known whether these genes are involved in plant steroid metabolism. Arabidopsis T-DNA knock-out mutants that are disrupted in one of these genes did not show any BR dwarf phenotype, possibly due to genetic redundancy of these genes (Chung, Chung, Choe unpublished data). Further studies, such as multiple mutants or heterologous expression of the genes followed by feeding experiments, may reveal a role for these genes in BR metabolism.

BR inactivation

CYP734A1: CS/BL to 26-hydroxyCS/26-hydroxyBL, reaction 14

In addition to the loss-of-function mutations in the BR biosynthetic genes, increased activity of BR-inactivating enzymes could result in dwarfism. A gene that was first shown to be involved in BR catabolism was found during suppressor mutant screening for phyB. PHYB encodes an apoprotein of the light receptor Phytochrome B. A loss-of-function mutation in this gene causes long hypocotyl phenotypes in the presence of white or red light. To isolate a suppressor mutation of phyB, Neff et al. (1999) performed activation-tagging mutagenesis against the phyb-4 and recovered a mutant in which the long hypocotyl phenotype was suppressed. This screen led to the isolation of a short hypocotyl mutant named phyb-4activation-taggedsuppressor1 (BAS1, At2g26710). Careful examination of the endogenous levels of BRs in the bas1-D mutant revealed that the amount of CS dropped to 15% that of wild-type. Thus, it was presumed that the reduced hypocotyl length in the bas1-D mutant was due to a decreased level of bioactive BRs (Neff et al. 1999). Isolation of a DNA fragment flanking the T-DNA from the activation-tagged mutants facilitated an isolation of the gene, and this gene was shown to encode a cytochrome P450 that was classified as CYP734A1 (formerly CYP72B1, Nelson et al. 2004b). To further understand the role of BAS1 in BR metabolism, Turk et al. (2003) heterologously expressed the cDNA in yeast. Feeding the BAS1-expressing yeast with CS and BL, and subsequent analysis of the metabolites revealed that BAS1 converted CS and BL to C-26-hydroxylated compounds. Separate experiments also showed that the C-26-hydroxylated BRs were not bioactive BRs. Thus, as was hypothesized, hydroxylation at C-26 position by BAS1 inactivated bioactive BRs as a part of a mechanism to remove bioactive BRs when they are not necessary in plants (Turk et al. 2003).

It is currently not known why the addition of a hydroxyl group to C-26 inactivates BRs. It might be that stearic hindrance due to the additional hydroxyl group prevents the compounds fitting into the pocket of the BR receptor. Alternatively, the C-26 hydroxyl group may be further conjugated with carbohydrates or esterified with fatty acids before losing their biological activity. Although not yet found in Arabidopsis, it has been shown that an enzyme preparation of tomato cell suspension cultures transforms 24-epiCS or 24-epiBL to 26-hydroxylated forms, and these are further glycosylated, suggesting that steroid C-26 hydroxylation in Arabidopsis may lead to further modification at this position before degradation or sequestration into a subcellular organelle like vacuoles.

CYP72C1: CHIBI2/SHRINKED/SOB7

Activation-tagging mutagenesis resulted in the identification of another locus whose gene product is involved in BR inactivation. Three groups independently isolated the gene with this approach. First, Nakamura et al. (2005) found that an activation-tagged mutant of Arabidopsis showed the conventional BR-deficient dwarf phenotype, and they named this chibi2 (chi2). Furthermore, two other groups identified activation-tagged mutants that were named shrinked1-D (shk1-D) (Takahashi et al. 2005) and a suppressorof phyb-4 7-D (sob7 ) (Turk et al. 2005). Like chi1, shk1-D and sob-D have the cauliflower mosaic virus 35S enhancer elements inserted in the promoter region, and accordingly overexpressed a cytochrome P450 gene (CYP72C1, At1g17060). At the amino acid sequence level, CYP72C1 and CYP734A1 (BAS1) share 36% identity. Judging by the dwarf phenotypes caused by overexpression of the CYP72C1 gene, the gene product was predicted to play a role as a BR-inactivating enzyme. Examination of the endogenous BR levels showed that all of the biosynthetic intermediates from 3-dehydro-6-deoxoTE and downstream of it in the pathways are reduced by more than half relative to wild-type (Nakamura et al. 2005, Takahashi et al. 2005, Turk et al. 2005). This suggests that CYP72C1 may participate in degradation of not only CS and BL but many of upstream intermediates. Because CYP734A1 was shown to hydroxylate at the C-26 position of BRs, Turk et al. (2005) tested whether the C-26 BR levels are also increased in the sob7-D mutants. Unlike bas1-D, the C-26-hydroxy BR levels were not significantly increased, suggesting that the enzymatic function of CYP72C1 may not be a C-26 hydroxylase. Further biochemical analysis should follow to determine the precise function of this BR-inactivating enzyme CYP72C1.

Arabidopsis has at least two different genes for BR inactivation, thus, it is interesting to understand the roles assigned to these two genes. The two genes may have different mechanisms of spatial and temporal regulation. To this end, Turk et al. (2005) examined different adult tissues of Arabidopsis, such as axillary leaf, cauline leaf, flower, rosette leaf, silique, and stem. The two genes were expressed in all the tissues tested. However, the degree of transcript abundance varied such that the SOB7 transcript was greatest in the silique, whereas the BAS1 transcript was significantly decreased in the rosette leaves and stem compared with other tissues (Turk et al. 2005). When the induction mechanisms were tested for light and BL, CYP734A was induced by both white light and BL treatment (Turk et al. 2005), whereas CYP72C1 was slightly induced in the dark (Nakamura et al. 2005, Turk et al. 2005). In addition, CYP734A1p::CYP734A1:GUS histochemical analysis showed that the gene is specifically expressed at the root/hypocotyl transition zone, in the elongation zone of the hypocotyl just below and often including the apical hook.

Takahashi et al. (2005) examined the spatial expression pattern of SHK1 in seedlings by using SHK1 promoter – GUS fusion gene. When the dark-grown seedlings of the SHK1::GUS transgenic line were stained for GUS activity, staining was found from wide range of the tissues including hypocotyls, roots, and the junction between hypocotyl and root, but GUS activity was not detectable from cotyledons (Takahashi et al. 2005). However, the seedlings grown in the light displayed GUS at cotyledons and roots but not in hypocotyls. Interestingly, GUS staining was also detected from flowers and siliques. However, the stain in the flowers was not found in young emerging floral buds but become clear as the flowers matured, suggesting that BRs are essential during early floral development (Takahashi et al. 2005). Considering differences in biochemical activities as well as gene expression pattern, it is likely that the two BR-inactivating enzymes, CYP734A1 and CYP72C1, play different roles to downregulate bioactive BR pools under the instruction of different upstream signals.

UGT73C5: UDP-glycosyl transferase 73C5, reaction 15

Biochemical detection of C23-glycosylated BRs from various plants (Fujioka and Yokota 2003) implied that plants have enzymes for glucosylation of BRs. Modification of steroid hormones by glucosylation has been known as a mechanism to regulate the pools of bioactive hormones in insects (Nakai et al. 2004) or human (Turgeon et al. 2003). Recently, Poppenberger et al. (2005) showed that UDP-glycosyltransferase 73C5 (UGT73C5) could participate in BL inactivation by BL-23-O-glucosylation activity. When the UGT73C5 gene was ectopically overexpressed in Arabidopsis, the transgenic plants displayed characteristic BR-deficient dwarf phenotypes, and this dwarfism was reverted to wild-type morphology by exogenous application of epi-BL (Poppenberger et al. 2005). In addition, the overexpression line possessed greatly decreased levels of BR intermediates, such as typhasterol, 6-deoxocastasterone, and CS. In contrast, when CS and BL were exogenously supplied to wild-type and UGT73C5-overexpressing line, the BL-23-O-glucoside level was increased approximately 500 times in the overexpression line compared with wild-type. In addition to UGT73C5, Arabidopsis genome possess many of closely homologous genes, future research should reveal whether these genes are implicated in BR homeostasis, which is important for understanding the precise control of plant growth and development.

Future perspectives

As progress on the elucidation of BR biosynthetic pathways is made, the pathways are becoming more networked. Currently, it has been shown that once 24-methylenecholesterol is made, it seems to serve as a precursor for BR biosynthesis (Hong et al. 2005). Even skipping the DWF1-mediated step (reaction 2 in Fig. 1), the sterols start to participate in the first reaction of BR biosynthesis that is mediated by DWF4. DWF4 seems to have a broad range of substrates; these include at least 24-methylenecholesterol, campesterol, and campestanol (Fig. 1). Once these sterols pass the committed step by DWF4, they seem to go through the downstream steps in the pathways mediated by, CPD, D2, DWF11, DDWF1, and CYP85. Although it has not been biochemically proven, the other sterols such as 22-hydroxymethylenecholesterol, (22S)-22-hydroxycampesterol, (22S, 24R)-22-hydroxy-ergost-4-en-3β-ol, (22S, 24R)-22-hydroxy-ergost-4-en-3-one, and (22S, 24R)-22-hydroxy-5α-ergostan-3-one that passed the DWF4-mediated reaction might go downstream, in that case, these networked pathways can be even pictured as a grid.

Major questions still remain in BR biology, including how are these steroids modified with fatty acids and/or carbohydrates? Many of the conjugates have been identified from various plants (Fujioka and Yokota 2003), but other molecular mechanisms of conjugation and the physiological roles of these compounds other than UGT73C5-mediated products remain unknown. In addition, modified steroids may be sequestered into subcellular organelles; however, intracellular or intercellular transporters for the conjugated molecules have not been identified. Targeted approaches using candidate proteins such as ATP-binding cassette (ABC) transporters might reveal some clues on transport mechanisms of BRs in plant cells in light of the facts that the ABC transporters are implicated in steroid transport processes in human cells (Young and Fielding 1999).

As rapid progress in BR biology is made, practical application of the BR-related knowledge to crop improvement has begun. As was outlined by the Sakamoto and Matsuoka (2004), BR-deficiency phenotypes of rice confer a variety of desirable traits. These include erect leaves that enables dense planting, short internodes that are useful for antilodging, and greener leaves which could result in increased photosynthetic efficiency. However, the severity of the known BR-deficient mutants of rice made them less attractive to plant breeders. Once the severity of the phenotype is under tight control through molecular breeding technology, it should contribute to better designing of crop plants for increased yields and improved grain quality in the future.

Acknowledgements –  This research was supported, in part, by grants (PF0330201-00) from Plant Diversity Research Center of 21st Century Frontier Research Program funded by Ministry of Science and Technology of Korean government and SRC program of MOST/KOSEF (R11-2000-081) through the Plant Metabolism Research Center, Kyung Hee University.

Ancillary