Super-resolution microscopy reveals functional organization of dopamine transporters into cholesterol and neuronal activity-dependent nanodomains

Abstract

Dopamine regulates reward, cognition, and locomotor functions. By mediating rapid reuptake of extracellular dopamine, the dopamine transporter is critical for spatiotemporal control of dopaminergic neurotransmission. Here, we use super-resolution imaging to show that the dopamine transporter is dynamically sequestrated into cholesterol-dependent nanodomains in the plasma membrane of presynaptic varicosities and neuronal projections of dopaminergic neurons. Stochastic optical reconstruction microscopy reveals irregular dopamine transporter nanodomains (∼70 nm mean diameter) that were highly sensitive to cholesterol depletion. Live photoactivated localization microscopy shows a similar dopamine transporter membrane organization in live heterologous cells. In neurons, dual-color dSTORM shows that tyrosine hydroxylase and vesicular monoamine transporter-2 are distinctively localized adjacent to, but not overlapping with, the dopamine transporter nanodomains. The molecular organization of the dopamine transporter in nanodomains is reversibly reduced by short-term activation of NMDA-type ionotropic glutamate receptors, implicating dopamine transporter nanodomain distribution as a potential mechanism to modulate dopaminergic neurotransmission in response to excitatory input.The dopamine transporter (DAT) has a crucial role in the regulation of neurotransmission. Here, the authors use super-resolution imaging to show that DAT clusters into cholesterol-dependent membrane regions that are reversibly regulated by ionotropic glutamate receptors activation.

Fig. 1

Dronpa-DAT distributes to cholesterol-dependent nanodomains in the plasma membrane of living CAD cells. a–d Images of untreated live CAD cells transiently expressing Dronpa-DAT. a TIRF-M image. b Reconstructed PALM image. c Enlarged PALM image corresponding to the boxed region in a, b. d DBSCAN-based cluster map of Dronpa-DAT distribution. Clustered localizations are shown by color-coding with non-clustered localizations in gray. e–h Images of live CAD cells expressing the myristoylated N-terminus of p60^SRC fused to Dronpa (Src15-Dronpa). e TIRF-M image, f reconstructed PALM image. g Enlarged PALM image corresponding to the boxed region in e, f. h DBSCAN-based cluster map of Src15-Dronpa distribution. Clustered localizations are shown by color-coding with non-clustered localizations in gray. i–l Images of Dronpa-DAT expressing CAD cells treated with 5 mM methyl-β-cyclodextrin (mβCD) to remove cholesterol. i TIRF-M image, j reconstructed PALM image. k Enlarged PALM image corresponding to the boxed region in i, j. i DBSCAN-based cluster map of Dronpa-DAT distribution. m–p Images of Dronpa-DAT expressing CAD cells treated with 5 μg ml⁻¹ cytochalasin D (CytD) to disrupt the actin skeleton. m TIRF-M image, n reconstructed PALM image. o Enlarged PALM image corresponding to the boxed region in m, n. p DBSCAN-based cluster map of Dronpa-DAT distribution. q Clustering of Dronpa-DAT measured as the fraction in % of total localizations (means ± s.e.m., *p < 0.05, one-way ANOVA with a Bonferroni’s post-test). Data are based on 12–20 cells from three independent experiments. Red arrows mark examples of DAT clustered in nanodomains. Scale bars a, b, e, f, i, j, m, n 5 μm, c, d, g, h, k, l, o, p 1 μm

Fig. 2

Visualization of DAT by STORM in CAD cells. TIRF-M (a, c, f, h), STORM (b, d, g, i, m, o), and widefield (l, n) images of CAD cells transiently expressing DAT with or without cholesterol depletion with 5 mM mβCD. DAT was visualized by immunolabeling with primary DAT antibody and Alexa405-Alexa647 conjugated-anti-rat secondary antibody. a, f TIRF-M view of DAT with (mβCD) or without cholesterol depletion (control). b, g Reconstructed STORM image of TIRF-M image in a or f. c, d, h, i Enlarged TIRF-M or STORM images corresponding to the boxed regions in a, b (control) or f, g (mβCD). e, j DBSCAN-based cluster map of the STORM images in d (control) and j (mβCD). Clustered localizations are shown by color-coding with non-clustered localizations in gray. k Fraction of particles (according to DBSCAN) clustered in nanodomains with (mβCD) or without cholesterol depletion (control). Data are % of localizations in cluster (means ± s.e.m., *p < 0.05, unpaired two-tailed t-test). l–o Visualization of DAT expressing CAD cells using inclined illumination to visualize a cross section through the cell. l Widefield image, m reconstructed STORM image of widefield image in l, n, o Enlarged TIRF-M or STORM images corresponding to the boxed regions in l, m. Red arrows mark DAT clustered in nanodomains. Data are based on 14–16 cells from three independent experiments. Scale bars a, b, f, g, l, m 5 μm, c–e, h–j 500 nm, n, o 400 nm

Fig. 3

Endogenous DAT is clustered in nanodomains in the extensions and varicosities of cultured dopaminergic neurons. a–c, f–h Cross-section views through dopaminergic cell body showing the DAT distribution in the soma in an overall view, the soma edge, and in intracellular vesicular structures. DAT was visualized by immunolabeling with primary DAT antibody and Alexa405-Alexa647-conjugated secondary antibody. a, f Widefield image and corresponding reconstructed STORM image in the cell body. b, g Widefield image and corresponding STORM image at the soma edge; c, h Enlarged widefield or STORM image corresponding to the boxed regions in a, f, visualizing DAT in a putative endocytic vesicle. Yellow arrows highlight the circular distribution of the DAT signal. d, i Widefield image and reconstructed STORM image in a representative extension with varicosities. e, j Enlarged widefield image and corresponding STORM image of extension and varicosity. k Cross sectional profile along the boxed region in j showing the signal intensity as a function of the distance (µm). l, m STORM image and corresponding DBSCAN-based cluster map showing the degree of clustering of DAT in a representative varicosity. n, o STORM image and corresponding DBSCAN-based cluster map showing the degree of clustering of DAT in a representative extension. Clustered localizations are shown by color-coding with non-clustered localizations in gray. p DBSCAN-based quantification of the clustering of endogenous DAT in dopaminergic neurons in soma, varicosities and extensions. Data are % of localizations in cluster (means ± s.e.m., ***p < 0.001, one-way ANOVA and Bonferroni’s post-test). Data are from 16 to 32 images from six individual experiments. Red arrows mark DAT clustered in nanodomains. q Varicosity from image n, where the localizations are colored by the number of nearest neighbors within a 30 nm radius. r–t Example clusters of varied densities identified in image q, where the number of nearest neighbors within a 30 nm radius are identified by color and by position on the z axis. Scale bars a, f, l 5 μm, b, d, g, i 1 μm, c, h, m 200 nm, e, j, n–q, s 500 nm

Fig. 4

Cholesterol depletion impairs DAT nanodomain clustering in extensions and varicosities of dopaminergic neurons. a, c, e, g, i, k Representative STORM images showing DAT distribution in extensions and varicosities of dopaminergic neurons under control conditions (a, c), after cholesterol removal with 5 mM mβCD (e, g), or after actin depolymerization with 5 μg ml⁻¹ CytD (i, k). DAT was visualized by immunolabeling with primary DAT antibody and Alexa405-Alexa647-conjugated secondary antibody. b, d, f, h, j, l DBSCAN-based cluster maps of the DAT signal in a, c, e, g, i, k. Clustered localizations are shown by color-coding with non-clustered localizations in gray. m, n DBSCAN quantifications DAT nanodomain localization in dopaminergic neurons after cholesterol removal or actin depolymerization, showing m clustering in varicosities and n clustering in extensions. Data are fraction of localizations in clusters in % (means ± s.e.m., ***p < 0.001, one-way ANOVA and Bonferroni’s post-test). o, p Probability distribution of the cluster sizes in varicosities (o) and extensions (p) at control conditions (black) or after treatment with mβCD (red) or CytD (blue). Data are based on from 9 to 39 images from three individual experiments. Red arrows mark DAT clustered in nanodomains. Scale bars 500 nm

Fig. 5

NMDA stimulation of dopaminergic neurons leads to decreased clustering of DAT. a, c, e, g, i, k Representative STORM images showing DAT distribution in extensions and varicosities of dopaminergic neurons under control conditions (a, c), after NMDA stimulation (5 min, 20 µM) (e, g), or after NMDA stimulation (5 min, 20 µM) in presence of the NMDA receptor antagonist AP5 (100 µM) (i, k). DAT was visualized by immunolabeling with primary DAT antibody and Alexa405-Alexa647-conjugated secondary antibody. b, d, f, h, j, l DBSCAN-based cluster maps of the DAT signal in a, c, e, g, i, k. Clustered localizations are shown by color-coding with non-clustered localizations in gray. m, n Quantifications, based on the DBSCAN, of DAT nanodomain localization in dopaminergic neurons after NMDA or after NMDA + AP5 showing m clustering in varicosities, and n clustering in extensions. Data are fraction of localizations in clusters in % (means ± s.e.m., ***p < 0.001, one-way ANOVA and Bonferroni’s post-test). o, p Probability distribution of the cluster sizes in varicosities (o) and extensions (p) at control conditions (black) or after NMDA (purple) or after NMDA + AP5 (green). Data are from 63 to 64 images from four individual experiments. Scale bar 200 nm. q Reversibility of NMDA-induced decrease in DAT nanodomain localization. Neurons were treated for 5 min with NMDA (purple) or vehicle (black) and fixed (treatment) or treated for 5 min with NMDA (purple) or vehicle (black) followed by a 1 h wash out before fixation (+1 h). Data are based on DBSCAN of the resulting STORM images and shown as fraction of localizations in clusters after NMDA as percentage of fraction of localizations in clusters at control conditions (means ± s.e.m., ***p < 0.001, two-way ANOVA and Bonferroni’s post-test). Data are from 146 images from two individual experiments. Scale bars 500 nm

Fig. 6

Cross-section through the cell body, extensions and varicosities of dopaminergic neurons showing by dual-color dSTORM imaging the distinct distribution of tyrosine hydroxylase (TH) compared to DAT. DAT was visualized by immunolabeling and Alexa647-conjugated secondary antibody. TH was visualized by immunolabeling and CF568-conjugated secondary antibody. a Somatic distribution of DAT. b Somatic distribution of TH. c Merged image of a and b. d–f Enlarged STORM images corresponding to the boxes shown in a–c. g Distribution of DAT in extension with varicosity. h Distribution of TH in extension with varicosity. i Merged image of g and h. j Distribution of DAT in extension. k Distribution of TH in extension. l Merged image of j and k. m Distribution of DAT in varicosity. n Distribution of TH in varicosity. o Merge image of m and n. Scale bars: a–c 5 μm, g–i 2 μm, d–f, j–o 500 nm

Fig. 7

Cross-section through the cell body, extensions and varicosities of dopaminergic neurons showing by dual-color dSTORM imaging the distinct distribution of the vesicular monoamine transporter-2 (VMAT2) compared to DAT. DAT was visualized by immunolabeling and Alexa647-conjugated secondary antibody. VMAT2 was visualized by immunolabeling and CF568-conjugated secondary antibody. a Somatic distribution of DAT. b Somatic distribution of VMAT2. c Merged image of a and b. d–f Enlarged STORM images corresponding to the boxes shown in a–c. g Distribution of DAT in extension with varicosities. h Distribution of VMAT2 in extension with varicosities. i Merged image of g and h. j Distribution of DAT in extension. k Distribution of VMAT2 in extension. l Merged image of j and k. m Distribution of DAT in varicosity. n Distribution of VMAT2 in varicosity. o Merge image of m and n. Scale bars a–c 5 μm, g–i 2 μm, d–f, j–o 500 nm