Identification and structure-guided development of triazole urea-based selective antagonists of Arabidopsis karrikin signaling

Abstract

The smoke-derived butenolides, karrikins (KARs), regulate many aspects of plant growth and development. However, KARs and a plant hormone, strigolactones (SLs), have high resemblance in signal perception and transduction, making it hard to delineate KARs response due to the shortage of chemical-genetic tools. Here, we identify a triazole urea KK181N1 as an inhibitor of the KARs receptor KAI2. KK181N1 selectively depress the KAR-induced phenotypes in Arabidopsis. We further elucidate the antagonistic, KAI2 binding mechanism of KK181N1, showing that KK181N1 binds to the catalytic pockets of KAI2 in a non-covalent binding manner. Our experiments also demonstrate the binding affinity of triazole urea compounds are regulated by the structured water molecule networks. By fine-tuning this network, we successfully develop a more potent derivative of KK181N1. We anticipate that these chemicals will be applicable to the elucidation of KARs biology, especially for discriminating the molecular and physiological aspects of KARs and SL signaling.

Fig. 1. Screening to identify a KAI2 signaling inhibitor.

a Schematic diagram of the screening using an in-house chemical library to identify KAI2 signaling inhibitor(s). Arabidopsis seeds of atd14-1, which is an SL receptor disruption mutant, were used in this screening. In the first step of the screening, the seeds were co-treated with 1 µM (±)-GR24 and 127 chemicals (5 µM each) from our in-house library. Nine chemical hits that reversed (±)-GR24-induced hypocotyl inhibition were further tested in the second step of the screening, where the seeds were co-treated with 1 µM KAR₁ and 9 candidate chemicals (5 µM each). b Chemical structure of KK181N1 with atom numbering. Two nitrogen atoms in the 1,2,3-triazole moiety are highlighted in blue. c Representative seedlings of 7-d-old wild-type Arabidopsis grown on MS medium containing 0.3 µM KAR₂ co-treated with DMSO or 10 µM KK181N1. DMSO is the mock control. Scale bar = 10 mm. d Hypocotyl growth of wild-type Arabidopsis seedlings in response to 0.3 µM KAR₂ supplemented with different concentrations of KK094 (white column) and KK181N1 (black column). Hypocotyl length was measured after cultivation under fluorescent white light (30 μmol m^–2 s^–1) at 22 °C for 2 days followed by under red LED light (35 μmol m^–2 s^–1) at 22 °C for 4 days. Values and bars represent the means ± SE (n = 35 (0 µM KAR₂, 0.3 µM KAR₂ + 3 µM KK094, 0.3 µM KAR₂ + 10 µM KK094, 0.3 µM KAR₂ + 0.3 µM KK181, 0.3 µM KAR₂ + 10 µM KK181, 0.3 µM KAR₂ + 30 µM KK181, and 0.3 µM KAR₂ + 100 µM KK181), 36 (0.3 µM KAR₂, 0.3 µM KAR₂ + 0.3 µM KK094, 0.3 µM KAR₂ + 1 µM KK094, 0.3 µM KAR₂ + 1 µM KK181), 37 (0.3 µM KAR₂ + 0.1 µM KK094, 0.3 µM KAR₂ + 0.1 µM KK181), 38 (0.3 µM KAR₂ + 100 µM KK094), and 39 (0.3 µM KAR₂ + 30 µM KK094)). Letters indicate significant differences (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. e The germination rate of wild-type Arabidopsis seeds co-treated with 1 µM KAR₂ and different concentrations of KK094 (white column) and KK181N1 (black column) was recorded. The seeds were first incubated at 30 °C under continuous fluorescent white light (20 μmol m^–2 s^–1) for 48 h and then at 22 °C for an additional 48 h. Data are the results of 6 biological replicates of around 30 seeds each. Values and bars represent the means ± SE. Letters indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. f Hypocotyl growth of Arabidopsis kai2-2, atd14-1, atd14kai2 and max2-1 seedlings in response to 1 µM (±)-GR24 supplemented with DMSO or 5 µM KK181N1. Hypocotyl length was measured after cultivation under fluorescent white light (30 μmol m^–2 s^–1) at 22 °C for 2 days followed by under red LED light (35 μmol m^–2 s^–1) at 22 °C for 4 days. Values and bars represent the means ± SE (n = 21) (Mock in kai2-2), 22 (Mock + 1 µM (±)-GR24 and 5 µM KK181 + 1 µM (±)-GR24 in kai2-2), 30 (Mock, Mock + 1 µM (±)-GR24, and 5 µM KK181 + 1 µM (±)-GR24 in d14-1, and Mock + 1 µM (±)-GR24 in max2-1), 33 (5 µM KK181 + 1 µM (±)-GR24 in max2-1), 34 (Mock in max2-1), (Mock + 1 µM (±)-GR24, and 5 µM KK181 + 1 µM (±)-GR24 in d14kai2), and 36 (Mock in d14kai2). Letters indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. n.s. indicates not significant. Source data are provided as a Source Data file.

Fig. 2. Recognition of KK181N1 by AtKAI2.

a ITC thermograms from the titration of KK181N1 into AtKAI2. Binding of KK181N1 to AtKAI2 gave a dissociation constant (K_d value) of 27.7 ± 0.4 μM, as calculated from three independent experiments. Error bar indicate standard deviation (SD) (n  =  3 technical replicates). b Crystal structure of AtKAI2-KK181N1. The structure is represented with ribbon and surface models. The cap domain is colored purple, and the core domain is green. In the close-up view of the active site of AtKAI2, KK181N1 (shown as a yellow sphere) is non-covalently embedded in the catalytic pocket. The relative spatial positions of catalytic nucleophile S95 (shown as the green stick) and four water molecules (named W1 to W4 and shown as red spheres) situated at the bottom of the pocket to KK181N1 are presented, along with the F_o-F_c omit map contoured at 2.0 σ (gray). c The hydrogen-bonding network formed at the bottom of active site. The residues involved in this network are shown as green stick models. Hydrogen bonds are shown as dashed yellow lines. d The hydrophobic interaction of KK181N1 with residues situated in the middle part of the pocket. Residues are represented as green sticks and further covered with dots to indicate the atomic surface with van der Waals radius. e Three residues (F134A, L142A, and L218A) along the boundary of the pocket are indicated. f The binding affinity of KK181N1 to AtKAI2 with mutations of pocket residues was measured by ITC. The relative affinity was calculated with K_d values for comparisons among mutants. n. d. indicates not detected. The gray, light orange, and light green shading respectively represent the relative affinity of residues with mutation at the bottom, middle, and entrance of the pocket. Data are means of 2–4 technical replicates. ITC thermograms for each mutant are shown in Supplementary Fig. 6. A summary of K_d values is given in Supplementary Table 2. Source data are provided as a Source Data file.

Fig. 3. Structural basis for KK181N1 as the KAI2-specific non-covalent antagonist.

a Structural comparison of residue F157 in apo-AtKAI2 (PDB code: 4JYP, magenta) and AtKAI2-KK181N1 (this study, green) with Y209 in apo-OsD14 (PDB code: 3VXK, gray) and OsD14-KK094CM (PDB code: 5ZHR, orange). KK181N1 is indicated by a cyan stick. Sphere models show oxygen atoms with the van der Waals radius. b ITC thermograms of the AtKAI2 mutant F157Y, producing a weak affinity with a K_d value of 445 ± 109 μM. Data are means ± SD (n  =  3 technical replicates). c The active site of AtKAI2 was separated into two pockets (1 and 2) by residue Y124. The cavities are shown as a semitransparent surface. d ITC thermograms of the AtKAI2 mutant Y124F. The K_d value is not detected (n. d.). e The germination rate of Striga hermonthica seeds treated with 10 nM (±)-GR24 and DMSO or different concentrations of KK181N1 was recorded. The conditioned seeds were applied compounds and incubated at 30 °C in the dark for 2 days. Data are the results of 6 biological replicates of around 50 seeds each. Values and bars represent the means ± SE. Letters indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. f Sequence alignment of Y124 and residues interacting with the methyl group of KK181N1. c, i, and d respectively represent the conserved, intermediate, and divergent subclades of parasitic HTLs proteins. Full sequence alignment is shown in Supplementary Fig. 14. g The germination rate of lettuce seeds treated with different concentrations of KK181N1 in the presence and absence of 1 μM KAR₁. The seeds were incubated at 22 °C in the dark for 48 h. Data are the results of 6 biological replicates of around 50 seeds each. Values and bars represent the means ± SE. Letters indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. h Structural comparison of AtKAI2-KK181N1 (green) and OsD14-KK094CM (PDB code: 5ZHR, orange). The van der Waals radius is indicated by spheres. The methyl group (red spheres) was modeled on KK094CM by PyMOL 2.5.4, producing Methyl-KK094CM. This hypothetical methyl group is thought to cause strong steric hindrance with C_β of F26 and C_δ of I193 (green spheres). i The binding affinities of AtKAI2 to 1,2,3-triazole ureas were measured by ITC assays. The relative affinity is calculated with K_d values (in the Table) for the comparisons among compounds. Data are means of 2 or 3 technical replicates (mean ± SD). ITC thermograms are shown in Supplementary Fig. 13. The chemical structures of 1,2,3-triazole ureas are shown in Supplementary Fig. 2a. Source data are provided as a Source Data file.

Fig. 4. Triazole urea compounds participate in a water-mediated hydrogen bond network in AtKAI2 and OsD14.

Comparison of the hydrogen-bonding networks in AtKAI2-KK181N1 (a, green) and OsD14-KK094CM (b, orange). KK181N1 and KK094CM are represented by yellow and wheat sticks, respectively. Four water molecules from AtKAI2-KK181N1 are labeled W1 to W4 in red shadow, and two from OsD14-KK094CM are labeled W1 and W2 in yellow shadow. Hydrogen bonds are indicated by green (a) or orange (b) dashed lines. c Chemical structures of KK122, KKT3054, and KK200. d The binding affinity of AtKAI2 to KK compounds, measured by ITC assays. The relative affinity is calculated with K_d values for comparison. Data are means of 2 or 3 technical replicates (mean ± SD). e Hypocotyl growth of wild-type Arabidopsis seedlings in response to 0.3 µM KAR₂ supplemented with different concentrations of KK181N1 (white column) and KKT3054 (black column). Hypocotyl length was recorded after cultivation under fluorescent white light (30 μmol m^–2 s^–1) at 22 °C for 48 h followed by red LED light (35 μmol m^–2 s^–1) at 22 °C for 4 days. Values and bars represent the means ± SE (n = 34 (0.3 µM KAR₂ + 100 µM KK181), 35 (0 µM KAR₂, 0.3 µM KAR₂, 0.3 µM KAR₂ + 10 µM KK181, 0.3 µM KAR₂ + 30 µM KK181, 0.3 µM KAR₂ + 0.1 µM KOKT3054, 0.3 µM KAR₂ + 10 µM KOKT3054, 0.3 µM KAR₂ + 100 µM KOKT3054), 36 (0.3 µM KAR₂ + 0.1 µM KK181, 0.3 µM KAR₂ + 1 µM KOKT3054, and 0.3 µM KAR₂ + 30 µM KOKT3054), 37 (0.3 µM KAR₂ + 0.3 µM KK181, 0.3 µM KAR₂ + 1 µM KK181, 0.3 µM KAR₂ + 3 µM KK181, and 0.3 µM KAR₂ + 0.3 µM KOKT3054), and 38 (0.3 µM KAR₂ + 3 µM KOKT3054)). *Letters indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. f The germination rate of wild-type Arabidopsis seeds treated with different concentrations of KK181N1 (white column) and KKT3054 (black column) in the presence of 1 µM KAR₂ was recorded. The seeds were first incubated at 28 °C under continuous fluorescent white light (20 μmol m^–2 s^–1) for 48 h and then at 22 °C for an additional 48 h. Data are the results of 6 biological replicates of around 30 seeds each. Values and bars represent the means ± SE. *Letter indicate significant difference (P < 0.05), as determined by one-way ANOVA followed by two-tailed Tukey’s HSD test for multiple-pair comparisons. Source data are provided as a Source Data file.