Revealing the location and dynamics of a concealed binding site in the dopamine transporter

Abstract

The dopamine transporter (DAT) is linked to neuropsychiatric disorders including ADHD, Parkinson's disease, and substance use disorders. Accordingly, DAT is the target of illicit drugs and clinically important medicines. However, the number and function of ligand binding sites in DAT is enigmatic due to conflicting data from available structures and molecular pharmacology. Herein, we design force sensors with DAT ligands and measure their interaction forces with wild-type and mutated DATs, from which two distinct populations of unbinding strengths and off-rates are detected. The high-force population is reduced by V152I and S422A mutations, or by substituting Na⁺ with K⁺ or NMDG⁺. In contrast, several modifications including mutation G386H, acetylation of K92 and K384, mutation K92A, mutation K384A, or protonation of H477 decrease the low-force population. The present data delineate the threshold of binding strength, which may account for certain ligand binding sites to be imperceptible in crystal or cryo-EM structures. Furthermore, the force spectra provide the information on the position and kinetic rates of a herein detected ligand binding site in DAT.

Fig. 1. The cocaine analogue MFZ2-12 binds to two sites in human DAT.

a The drug molecule MFZ 2-12, was covalently conjugated to the AFM cantilever tip via its N-ethylamino derivative and an azido-PEG-NHS linker (Supplementary Fig. 1). b The force measurements (Supplementary Fig. 2) revealed that, in the presence of 10 µM Zn²⁺, the binding activity of MFZ2-12 to DAT is significantly higher than that in the absence of Zn²⁺ (one-tailed t-test with equal variance assumed, n = 5 cells. The box range indicates the standard deviation, the small square in the box indicates the mean value, the full width bar indicates the median value, the whisker indicates the range of outlier with coefficient equals to 1). c Two populations of unbinding events were observed from repeatedly recorded force curves. Ten example curves are shown for each. The noise in the force curves originates from thermal fluctuation of the cantilever with a mean value of 0.6 nm (corresponding equation and calculation in the Methods section). d Statistic probability density functions (PDF) of the unbinding force show two peaks on cells expressing DAT, whereas the PDF on CHO cells without DAT has much lower binding probability (area under the curve) and a single peak different from those on DAT. All PDF measurements were conducted on at least two biological replicas with very similar results. e A mutation within the S1 site of DAT (S422A) abolished the second peak (measured in the presence of 10 µM Zn²⁺) in the force PDF. f Substitution of Na⁺ with Li⁺ reduced the population of stronger unbinding events of the wtDAT. Measurements in panel d, e, and f were performed at pulling speed of 6 µm/s. Within each panel, the same cantilever was used to compare different condition. For different panels, the cantilevers are different. Source data are provided as a Source Data file.

Fig. 2. Visualization of the substrate binding site S1 and S2, based on the cocaine-bound structure of the human DAT.

The inhibitor cocaine is bound to the S1 and shown in space-filling mode in blue. The residues of the S1 in contact with cocaine are colored violet, the residues of the S2 are shown in yellow, the residues (V152, S422, G386, K384, and K92) tested by mutation are highlighted in red, the residues in contact with the inhibitor MRS7292 are shown in green. a Full transporter model with the scaffold domain shown in light salmon, the bundle domain in light blue. b Zoom into the substrate binding sites. The hydrophobic gate residue F320 (colored in orange) is prominently in contact with substrate in S1 and inhibitors bound to S2.

Fig. 3. Dopamine binds to two sites of human DAT.

a The dopamine molecule was covalently conjugated to the AFM cantilever tip via NHS-PEG-maleimide linker (Supplementary Fig. 8a). Two populations of unbinding events, red solid lines, measured in the absence of Zn²⁺, (at scanning speed of 1.5 µm/s) were observed from repeatedly recorded force curves. Substitution of Na⁺ in the buffer by K⁺ (b) or NMDG (c) diminished the population of the stronger binding events (around 20 pN), indicating that the stronger binding events originate from the central S1 site, which is sodium dependent. Mutation within the S1 site of DAT (V152I) also reduced the population of stronger binding events (d) while the overall binding activity is preserved. Mutation in the extracellular vestibule (G386H) results in weaker binding (shift of the first force peak to the left) but does not affect the binding force at the S1 site e. Source data are provided as a Source Data file.

Fig. 4. Identifying the location of S2 site of human DAT.

a Cells expressing wtDAT were treated with 3 mM sulfo-NHS-acetate (SNA) for 30 min to acetylate K92 and K384 of DAT. After the treatment, the population of the weaker unbinding events decreased substantially (dashed blue curve in b, measured in the absence of Zn²⁺, at pulling speed of 6 µm/s). For further identification, DAT mutants with single mutation K92A or K384A were measured, both of which reduced the population of weaker unbinding events (green and magenta thin solid curves in b). c The pH value of the buffer was reduced from 7.4 to 5.5 to examine the role of H477 in supporting binding to the S2. Histidine is the only amino acid residue with a pKa value within this pH range and there is no histidine at S1 site (Supplementary Table 1). The drop in pH reduction has a markedly more pronounced effect on unbinding from the S2 site (the first peak around 12 pN of the red solid line in c, measured in the absence of Zn²⁺, at scanning speed of 1.5 µm/s) than from the S1 site (the second peak around 20 pN). d Kinetic on-rate at the S1 site of untreated wtDAT is significantly higher than those of DATs changed at K92, K384, or H477, indicating that the S2 site helps dopamine enter the S1 site. e Kinetic off-rate at the S2 site is much higher than that at the S1 site. (One-tailed t-test with equal variance assumed, n = 3–8 cells. The box range indicates the standard deviation, the small square in the box indicates the mean value, the full width bar indicates the median value, the whisker indicates the range of outlier with coefficient equals to 1). Source data are provided as a Source Data file.