A key residue of the extracellular gate provides quality control contributing to ABCG substrate specificity

Abstract

For G-type ATP-binding cassette (ABC) transporters, a hydrophobic "di-leucine motif" as part of a hydrophobic extracellular gate has been described to separate a large substrate-binding cavity from a smaller upper cavity and proposed to act as a valve controlling drug extrusion. Here, we show that an L704F mutation in the hydrophobic extracellular gate of Arabidopsis ABCG36/PDR8/PEN3 uncouples the export of the auxin precursor indole-3-butyric acid (IBA) from that of the defense compound camalexin (CLX). Molecular dynamics simulations reveal increased free energy for CLX translocation in ABCG36^L704F and reduced CLX contacts within the binding pocket proximal to the extracellular gate region. Mutation L704Y enables export of structurally related non-ABCG36 substrates, IAA, and indole, indicating allosteric communication between the extracellular gate and distant transport pathway regions. An evolutionary analysis identifies L704 as a Brassicaceae family-specific key residue of the extracellular gate that controls the identity of chemically similar substrates. In summary, our work supports the conclusion that L704 is a key residue of the extracellular gate that provides a final quality control contributing to ABCG substrate specificity, allowing for balance of growth-defense trade-offs.

Fig. 1. L704F mutation in ABCG36 uncouples IBA from camalexin export.

IBA (a) CLX (b) and IAA (c) export from Arabidopsis protoplasts prepared from indicated ABCG36 loss-of-function alleles. Significant differences (p < 0.05) of means ± SE (n = 8 independent protoplast preparations) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. IBA (d) CLX (e) and IAA (f) export from N. benthamiana protoplasts after transfection with indicated mutant versions ABCG36. Significant differences (p < 0.05) of means ± SE (n = 14 independent protoplast preparations) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Confocal imaging (g) and quantification (h) of GFP-tagged versions of mutated ABCG36 after transfection of N. benthamiana leaves. Short treatment of FM4-64 was used as PM markers; bar, 50 μm. Significant differences (p < 0.05) of means ± SE (n = 25 epidermal cells) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. i ATPase activity of microsomal fractions prepared from tobacco leaves transfected with vector control or indicated mutant versions of ABCG36 measured at pH 9.0 in the presence of 50 μM indole, IAA, IBA, CLX, or ortho-vanadate. Significant differences (p < 0.05) of means ± SE (n = 3 independent transfections and microsomal preparations) were determined using Ordinary Two-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Data are presented as box-and-whisker plots, where median and 25^th and 75^th percentiles are represented by the box itself and the middle line, respectively; means are indicated by a “+”. Source data is provided as a Source Data file.

Fig. 2. Uncoupling of IBA from camalexin export is ecotype-independent.

Relative root length of two independent Arabidopsis abcg36-4 lines complemented with indicated mutant versions of ABCG36 grown for 12 days on 7.5 μM IBA (a) or 5 μg/ml CLX (b); Wt growth is set to 100%. Significant differences (p < 0.05) of means ± SE (n = 3 independent experiments with each 10 seedlings) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. c, d IBA (c) and CLX (b) export from two independent Arabidopsis abcg36-4 lines complemented with indicated mutant versions of ABCG36. Significant differences (p < 0.05) of means ± SE (n = 8 independent protoplast preparations) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Confocal imaging (e) and quantification (f) of Arabidopsis abcg36-4 lines complemented with indicated mutant versions of ABCG36; bar, 50 μm. Significant differences (p < 0.05) of means ± SE (n = 100 root cells) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Data are presented as box-and-whisker plots, where median and 25th and 75th percentiles are represented by the box itself and the middle line, respectively; means are indicated by a “+”. Source data is provided as a Source Data file.

Fig. 3. ABCG36 has a sensor-like function during infection.

5-week-old plants grown on soil were watered with buffer (untreated control) or with Fo (Fo699; 10⁷ conidia/ml). Representative plants are pictured (a) and disease symptoms were quantified (b) using a scale from 0-10 ¹⁴. Significant differences (p < 0.05) of means ± SE (n = 3 independent infection series with each 15 single plants) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. cExtracellular levels of camalexin (c) and 4MeOI3M (d) on P. infestans-inoculated Arabidopsis leaves determined by non-targeted UPLC-ESI-QTOF-MS. Significant differences (p < 0.05) of means ± SE (n = 7 with each 30 droplets) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Data are presented as box-and-whisker plots, where median and 25th and 75th percentiles are represented by the box itself and the middle line, respectively; means are indicated by a “+”.

Fig. 4. L704 is part of the extracellular gate of ABCG36 and L704F increases free energy surface for camalexin in the entrance region.

a Chemical formulas for potential ABCG36 substrates employed in this study (downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov)). b THs of ABCG36 are shown from the extracellular space. Extracellular gate residues are highlighted by sticks and balls. Docked IBA is shown using stick representation. Electronic structural deformation, the electric field lines of IBA, CLX, and IAA before (c) and after binding to the ABCG36 substrate pocket (d). The electronic binding energies and the results displayed here were obtained by ab initio calculations using DFT methods. e, f Specific ³H-IBA and ³H-CLX binding to indicated Arabidopsis microsomes calculated as the difference between total and unspecific substrate binding determined in the absence (total) and presence of a 1000-fold excess of non-radiolabeled substrate concentrations (unspecific), respectively. Significant differences (p < 0.05) of means ± SE (n = 3 independent microsomal preparations with each 3 technical replica) to the corresponding Wt were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Data are presented as box-and-whisker plots, where median and 25th and 75th percentiles are represented by the box itself and the middle line, respectively; means are indicated by a “+”. Source data are provided as a Source Data file. Free energy surfaces (FES) were calculated from metadynamics simulations with complexes of IBA (g) and CLX (h) with Wt and L704F ABCG36, respectively. Pulling of IBA (i) and CLX (j) from the central binding pocket to the extracellular space were performed in biased simulations with Wt and L704F ABCG36, respectively. Frequencies of contacts between the small molecules and the protein are plotted.

Fig. 5. L704Y mutation in ABCG36 widens the transport selectivity to indolic compounds.

IBA (a) CLX (b), IAA (c), and indole (d) export from N. benthamiana protoplasts after transfection with indicated mutant versions ABCG36. Means of mutant ABCG36 that are significantly different from Wt ABCG36 (grey fill) are indicated in red, while non-significant ones are in blue. ABCG36^A1357V (green) is included as a negative control. Significant differences (p < 0.05) of means ± SE (n = 14 independent protoplast preparations) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. Confocal imaging (e) and quantification (f) of GFP-tagged versions of mutated ABCG36 after transfection of N. benthamiana leaves. Short treatment of FM4-64 was used as PM markers; bar, 50 μm. Note that slightly enhanced PM signals of some mutant versions of ABCG36 in comparison to Wt are also found in respective FM4-64 controls and thus do not reflect higher expression levels; bar, 10 μm. Significant differences (p < 0.05) of means ± SE (n = 25 epidermal cells) were determined using Ordinary One-way ANOVA (Tukey’s multiple comparison test) and are indicated by different lowercase letters. g Free energy surfaces (FES) were calculated from metadynamics simulations with complexes of indole with Wt and L704Y ABCG36, respectively. h–j Pulling simulations of indole from the central binding pocket to the extracellular space were performed in simulations with Wt and L704Y ABCG36, respectively. Frequencies of contacts between the small molecule and the protein are plotted. j. A logistic regression model was employed to identify the key residues involved in defining transport. The model’s coefficients were color-coded from blue to red, indicating residues whose interactions negatively (blue) or positively (red) support transport, respectively. Data are presented as box-and-whisker plots, where median and 25th and 75th percentiles are represented by the box itself and the middle line, respectively; means are indicated by a “+”. Source data is provided as a Source Data file.

Fig. 6. Evolutionary analysis of key residues in the putative extracellular gate.

WebLogo presentation of the extracellular gate regions of TH5 (a) and TH11 (b) of 1853 full-size ABCGs. The x-axis displays the position of amino acids in the multiple sequence alignments according to AtABCG36. c L704 and F1375 are less conserved than F703 and F1374 in full-size ABCG sequences, the positions of ABCG substrates, IBA, and CLX are indicated. d Occurrence of amino acids corresponding to F703, L704, F1374, and F1375 in Arabidopsis ABCG36 in full-size ABCG sequences of indicated taxa.