The BAHD Acyltransferase BIA1 Uses Acetyl-CoA for Catabolic Inactivation of Brassinosteroids^,

Abstract

Previous research complemented with results on BIA1 enzymatic activities shows that the enzyme regulates brassinosteroid homeostasis via mono- and diacetylation of castasterone

Dear Editor,

Brassinosteroids (BRs) are steroid hormones that are essential for normal development but also enable plants to adapt their growth to environmental conditions. The hormones are formed from the bulk sterol campesterol in a branched biosynthetic pathway that produces the biosynthetic end products castasterone (CS) and brassinolide (BL; Clouse, 2011).

BL and CS are highly bioactive, acting in picomolar to nanomolar concentrations. Thus, stringent control of their homeostasis is essential and is achieved through multiple modes, including feedback control of biosynthetic gene expression and catabolic inactivation of CS and BL via hydroxylation, glucosylation, sulfonation, and acylation (summarized in Fig. 1). The first catabolizing enzyme to be discovered was PHYB ACTIVATION TAGGED SUPPRESSOR 1 (BAS1), a cytochrome P450 monooxygenase that acts in the hydroxylation of CS and BL (Neff et al., 1999; Turk et al., 2005). It was identified in a screen by the phenotypes of bas1-d, a dominant mutant that shows characteristic symptoms of BR deficiency, including dwarf growth with dark-green, round leaves in the light and de-etiolated hypocotyls in the dark (Neff et al., 1999).

Figure 1.

BIA1 uses acetyl-CoA to acylate CS in vitro. A, Catabolic inactivation events known to date. The illustration was modified from Rozhon et al. (2019). Enzymes that participate are sulfotransferases AtST1 and AtST4a (Marsolais et al., 2007), cytochrome P450 BAS1 (Neff et al., 1999), UDP-glycosyltransferases UGT73C5 and UGT73C6 (Poppenberger et al., 2005; Husar et al., 2011), acyltransferase PIZZA/BAT1/DRL1 (Schneider et al., 2012; Choi et al., 2013; Zhu et al., 2013), and BIA1, whose enzymatic activity was previously unknown. B, Activity of recombinant GST-BIA1 against CS. Enzyme assays contained 3 μg of recombinant GST-BIA1 fusion protein and 1 mm CS, and 10 mm acetyl-CoA or 2.5 mm caffeoyl-CoA, coumaroyl-CoA, malonyl-CoA, or myristoyl-CoA were added as donors. Reaction products were separated by TLC. Asterisks show the position of the reaction product (red) and autofluorescence originating from caffeoyl-CoA (orange). C, Michaelis‐Menten kinetics of GST-BIA1. Values are means ± se of three replicates.

Other catabolizing enzymes were identified in mutant screens. One of them is the BAHD acyltransferase BRASSINOSTEROID INACTIVATOR1/ABNORMAL SHOOT1 (BIA1/ABS1; Roh et al., 2012; Wang et al., 2012). The dominant, activation-tagged bia1-1D and bia1-2D mutants showed severe BR-deficient phenotypes, which could be rescued by BR, had significantly reduced amounts of several BRs, including 6-deoxocastasterone and CS, and displayed molecular signatures indicative of BR depletion. Moreover, bia1-3, a loss-of-function line, showed phenotypes reminiscent of BR overaccumulation (Roh et al., 2012). Thus, there was clear evidence that BIA1 can inactivate BRs. However, its enzymatic activity had remained unknown, and we addressed this.

Given the phenotypes and BR profile of bia1-1D, we hypothesized that BIA1 may acylate bioactive BRs, in particular CS. To test this, we expressed BIA1 as a GST-tagged protein in Escherichia coli (Supplemental Fig. S1) and performed in vitro acylation assays with 3 μg of fusion protein (for all experimental procedures see Supplemental Table 1 and Supplemental Materials and Methods), an amount that gave satisfying results in preliminary tests (Supplemental Fig. S2). CS, as well as 24-epiCS and 24-epiBL, were selected as acceptors. Myristoyl-CoA was chosen as the most promising donor, since all BR-acyl conjugates described so far possess long-chain fatty acids (Asakawa et al., 1994; Kolbe et al., 1995; Soeno et al., 2000; Schneider et al., 2012). In addition, acetyl-CoA, caffeoyl-CoA, coumaroyl-CoA, and malonyl-CoA were used. Thin-layer chromatography (TLC) analyses of the enzymatic reactions showed, surprisingly, that CS and 24-epiCS conjugates were obtained with acetyl-CoA, while none were formed with the other CoA donors, including myristoyl-CoA (Fig. 1; Supplemental Figs. S2 and S3). Also, for 24-epiBL, a product was seen with acetyl-CoA, albeit at a lower level (Supplemental Fig. S3). A fluorescent band was similarly formed in all reactions using caffeoyl-CoA. However, it originated from caffeoyl-CoA itself (Supplemental Fig. S4) and thus is not a conjugate. No product was obtained with GST alone or heat-inactivated GST-BIA1 fusion protein, which were included as controls (Supplemental Fig. S5).

Analysis of the products by high performance liquid chromatography-quadrupole time-of-flight mass spectrometry (HPLC-QTOF) revealed that with both CS and 24-epiBL, mainly diacetylation products were formed, with only minor amounts of monoacetylation products detectable (Table 1; Supplemental Figs. S6–S11).

Table 1. Peak area% of acetylation products of CS and BL variants, measured by HPLC-QTOF.

Protonated, ammonium, and sodium adducts of the respective mono- and diacetylated products were extracted, the peak areas obtained and summed up, and the relative percentages of mono- and diacetylated products calculated.

BRs	C-2 OH	C-3 OH	C-22 OH	C-23 OH	C-24 CH3	Monoacetylation Product	Diacetylation Product
						area %	area %
CS	α	α	R	R	S	1.3	98.7
24-epiBL	α	α	R	R	R	3.9	96.1
24-epiCS	α	α	R	R	R	8.5	91.5
6-Deoxo-24-epiCS	α	α	R	R	R	3.8	96.2
tris-EpiCS	α	α	S	S	R	46.9	53.1
3,24-DiepiCS	α	β	R	R	R	100.0	0.0

To investigate structural properties required for CS acetylation in more detail, the CS variants 24-epiCS, 6-deoxo-24-epiCS, tris-epiCS, and 3,24-diepiCS were tested. BIA1 efficiently acetylated 24-epiCS, indicating that the stereochemical pattern of C-24 carrying a methyl group does not interfere with the reaction (Table 1; Supplemental Fig. S8). For 24-epiCS and 6-deoxo-24-epiCS, mainly diacetylated products were obtained, indicating that the 6-oxo group also is not required for BIA1 activity (Table 1; Supplemental Figs. S8 and S9). Tris-epiCS was efficiently utilized as well, and two bands with equal intensities were observed via TLC (Supplemental Fig. S3); in line, HPLC-QTOF showed that comparable amounts of mono- and diacetylated products were generated (Table 1; Supplemental Fig. S10). In the case of 3,24-diepiCS, just one slowly migrating band was seen (Supplemental Fig. S3), and HPLC-QTOF analysis confirmed that only a monoacetylated product was formed (Table 1; Supplemental Fig. S11). Taken together, these data imply that BIA1 mainly diacetylates CS. One acetyl residue seems to be transferred to ring A, most likely to the hydroxy group at C-3, and the second acetyl moiety to the aliphatic side chain, to the hydroxyl group at either C-22 or C-23.

Further enzymatic activity assays established that in vitro GST-BIA1 has best activity at pH 9.0 (Supplemental Fig. S12A) and a temperature of 40°C (Supplemental Fig. S12B). A linear reaction was seen for up to 8 h (Supplemental Fig. S12C). Using pH 9.0, 30°C, and 1 h incubation time, an enzyme kinetic was recorded by measuring the reaction velocity (v) at different CS concentrations by HPLC. The calculated K_M value of GST-BIA1 for diacetylation of CS was 26 μm and the maximal reaction velocity (v_max) was 1.3 μmol h⁻¹ mg⁻¹ (Fig. 1C), showing that BIA1 has a high affinity for CS.

To determine whether BIA1 is required in planta to form acylCS and diacylCS, we generated plants overexpressing yellow fluorescent protein (YFP)-tagged BIA1 under control of the constitutive Cauliflower mosaic virus 35S promoter (35S:BIA1-YFP). The strongest symptoms of BR deficiency were seen in line 4.2 (Fig. 2), which also formed the most BIA1-YFP protein (Fig. 2). In addition, the bia1-3 T-DNA insertion mutant was obtained and characterized. Sequencing revealed that the T-DNA is integrated 49 bp upstream of the stop codon and quantitative PCR analysis confirmed that BIA1 mRNA is clearly decreased (Fig. 2; Roh et al., 2012). Experiments in which bia1-1D, 35S:BIA1-YFP4.2, bia1-3, and wild-type plants were fed with CS, and analyses of the formed catabolites by HPLC-QTOF showed that wild-type plants produced a peak with a mass corresponding to CSAc₂, detectable at a retention time of 11.2 min (Supplemental Fig. S13). This metabolite was 100 times more abundant in bia1-1D and 35S:BIA1-YFP line 4.2 plants, and was absent in the bia1-3 mutant. In addition, peaks with a mass corresponding to monoacetylated CS were detected in all lines except bia1-3.

Figure 2.

BIA1 acylates CS in planta. A, Phenotypes of 3-week-old wild-type and BIA1-YFP overexpressing plants grown in long days at 22°C ± 2°C. Here, multiple images were made into a digital composite for comparison. B, Expression of the transgene in the lines shown in A was confirmed by immunoblot analysis using an anti-GFP antibody. CBB, Coomassie Brilliant Blue. C, BIA1 transcript levels in 10-d-old seedlings of the wild type and the bia1-3 T-DNA insertion line. Data were normalized to GAPC2. D, HPLC-QTOF metabolite analysis of 12-d-old seedlings incubated for 24 h in one-half strength Murashige and Skoog medium supplemented with 1 μg mL⁻¹ CS. The bars show the mean ± sd of three to four biological replicates; n.d., Not detected. Numbers are P-values from a two-sided Student's t test with heteroscedastic variance for difference compared to the wild type. Only significant (<0.05) values are shown.

In summary, our results show that BIA1 is required for the formation of monoacetylated and, in particular, diacetylated CS and uses acetyl-CoA as a donor substrate. Combined with the results from previous work, there is conclusive evidence now that BIA1 functions in BR acetylation, which is a highly effective means of bioinactivation.

Supplemental Data

The following supplemental materials are available:

• Supplemental Figure S1. Purification of BIA1 as a GST-fusion protein.

• Supplemental Figure S2. Optimization of GST-BIA1 input protein amounts.

• Supplemental Figure S3. Activity of purified BIA1 protein toward 24-epiBL and CS variants.

• Supplemental Figure S4. Caffeoyl-CoA causes a fluorescent band, indicated by pound signs in lanes 4 and 5.

• Supplemental Figure S5. Activity of recombinant, GST-tagged BIA1 protein toward 24-epiBL and CS variants with acetyl-CoA as a donor.

• Supplemental Figure S6. Analysis of CS acetylation products by HPLC-QTOF.

• Supplemental Figure S7. Analysis of 24-epiBL acetylation products by HPLC-QTOF.

• Supplemental Figure S8. Analysis of 24-epiCS acetylation products by HPLC-QTOF.

• Supplemental Figure S9. Analysis of 6-deoxo-24-epiCS acetylation products by HPLC-QTOF.

• Supplemental Figure S10. Analysis of tris-epiCS acetylation products by HPLC-QTOF.

• Supplemental Figure S11. Analysis of BIA1 acetylation products of 3,24-diepiCS by HPLC-QTOF.

• Supplemental Figure S12. Optimization of BIA1 reaction conditions.

• Supplemental Figure S13. Detection of CSAc₂ in plants fed with CS.

• Supplemental Table S1. Oligonucleotides used in this study.

• Supplemental Materials and Methods. Methodology used in this study.