Bikinin-like inhibitors targeting GSK3/Shaggy-like kinases: characterisation of novel compounds and elucidation of their catabolism in planta

Abstract

Background: Plant GSK-3/Shaggy-like kinases are key players in brassinosteroid (BR) signalling which impact on plant development and participate in response to wounding, pathogens and salt stress. Bikinin was previously identified in a chemical genetics screen as an inhibitor targeting these kinases. To dissect the structural elements crucial for inhibition of GSK-3/Shaggy-like kinases by bikinin and to isolate more potent compounds we synthesised a number of related substances and tested their inhibitory activity in vitro and in vivo using Arabidopsis thaliana.

Results: A pyridine ring with an amido succinic acid residue in position 2 and a halogen in position 5 were crucial for inhibitory activity. The compound with an iodine substituent in position 5, denoted iodobikinin, was most active in inhibiting BIN2 activity in vitro and efficiently induced brassinosteroid-like responses in vivo. Its methyl ester, methyliodobikinin, showed improved cell permeability, making it highly potent in vivo although it had lower activity in vitro. HPLC analysis revealed that the methyl residue was rapidly cleaved off in planta liberating active iodobikinin. In addition, we provide evidence that iodobikinin and bikinin are inactivated in planta by conjugation with glutamic acid or malic acid and that the latter process is catalysed by the malate transferase SNG1.

Conclusion: Brassinosteroids participate in regulation of many aspects of plant development and in responses to environmental cues. Thus compounds modulating their action are valuable tools to study such processes and may be an interesting opportunity to modify plant growth and performance in horticulture and agronomy. Here we report the development of bikinin derivatives with increased potency that can activate BR signalling and mimic BR action. Methyliodobikinin was 3.4 times more active in vivo than bikinin. The main reason for the superior activity of methyliodobikinin, the most potent compound, is its enhanced plant tissue permeability. Inactivation of bikinin and its derivatives in planta involves SNG1, which constitutes a novel pathway for modification of xenobiotic compounds.

Figure 1

Targets of inhibitors interfering with sterol biosynthesis, BR biosynthesis and BR signal transduction.

Figure 2

Structures of the synthesised compounds. The compounds 14 and 15 are also called bikinin (BIK) and iodobikinin, respectively. The compounds 18 and 10, also called methylbikinin and methyliodobikinin, respectively, represent their methylesters.

Figure 3

In vitro inhibitor activity. For ASK in vitro kinase assays, purified GST-ASK fusion proteins were incubated with MBP as a substrate and [γ-³²P]-ATP as a co-substrate in absence (−) or presence of 10 μM of the different compounds (the numbers correspond to Figure 2). Compounds 1 to 9 (left panel) differ in the aliphatic side chain. The influence of the position of the heterocyclic nitrogen was tested with compounds 3 and 11 (middle panel; the molecular structure shown represents compound 11). The right panel shows the effect of the halogen substituent of the pyridine ring. BIK, bikinin.

Figure 4

In vivo effects of the inhibitors. (A) 7-day-old wild type A. thaliana seedlings, the brassinosteroid synthesis mutant cpd or the signalling mutant bri1-1 were transferred to ½ MS medium containing 1 μM 24-epi-brassinolide (24-epi-BL), 30 μM compound 10 or 30 μM compound 15 and incubated for 7 days under long day conditions. All pictures were taken with the same magnification. The size bar represents 1 mm. (B) The hypocotyl length of 7-day-old A. thaliana Col-0 seedlings grown on ½ MS containing 1% sucrose and supplemented with compound 10 or 15 at the indicated concentrations were measured. The means and standard deviations were calculated from at least 25 seedlings. (C) Plants expressing BZR1-CFP were treated with the indicated compound for 1 h. Subsequently, the phosphorylation status of BZR1 was detected by western blot analysis using a GFP-antibody. A coomassie R250 stain is shown as loading control.

Figure 5

Methylated compounds are rapidly hydrolysed in planta. A. thaliana Col-0 seedlings were infiltrated with a solution containing 50 μM compound 10. (A) Control samples were taken before infiltration and analysed by HPLC. (B) Already after 15 min a second peak designated P1 appeared. (C) After 1 h two additional peaks, P2 and P3 were visible, (D) which became even more pronounced after 48 hours. (E) The same experiment was performed with sng1 mutants. The chromatogram obtained with the sample taken after 48 h is shown. (F) Overlay of the 48 h chromatograms of the Col-0 sample and the sng1 sample. The small boxes inserted into the chromatograms show the UV spectra of the peaks in the range of 220 to 400 nm. SM, sinapoylmalate; SG, sinapoylglucose; mAU, milli absorption units recorded at 250 nm.

Figure 6

Methylation increases tissue-permeability.A. thaliana seedlings were incubated in 50 μM solutions of compounds 10 and 15 in ½ MS medium. Samples were taken after the indicated time and the in situ levels of compound 15 were analysed by HPLC. The solid line represents the results for plants incubated with compound 10 and the dashed line shows the results for compound 15. The averages and standard deviations were calculated from 3 independent assays.

Figure 7

Activity of the modified compounds. The in planta modification products of compound 10 were isolated and assayed for their potency to inhibit BIN2 activity in vitro. Compound 15 was used as a control. Each substance was used at a concentration of 10 μM. The same conditions as in Figure 3 were applied except that [γ-³³P]-ATP was used as co-substrate. A stain with coomassie R250 is included as a loding control.

Figure 8

The catabolic fate of compound 10 in A. thaliana. Compound 10 is rapidly hydrolysed, likely by a lipase, and thereby activated to compound 15. Subsequently, compound 15 is inactivated by conjugation with glutamic acid or with malate. The latter reaction is catalysed by SNG1. Since SNG1 utilises glucosylated substrates compound 15 is presumably glucosylated by a UGT prior to SNG1-mediated conjugation with malate.