Novel Fluorescent Ligands Enable Single-Molecule Localization Microscopy of the Dopamine Transporter

Abstract

The dopamine transporter (DAT) is critical for spatiotemporal control of dopaminergic neurotransmission and the target for therapeutic agents, including ADHD medications, and abused substances, such as cocaine. Here, we develop new fluorescently labeled ligands that bind DAT with high affinity and enable single-molecule detection of the transporter. The cocaine analogue MFZ2–12 (1) was conjugated to novel rhodamine-based Janelia Fluorophores (JF₅₄₉ and JF₆₄₆). High affinity binding of the resulting ligands to DAT was demonstrated by potent inhibition of [³H]dopamine uptake in DAT transfected CAD cells and by competition radioligand binding experiments on rat striatal membranes. Visualization of binding was substantiated by confocal or TIRF microscopy revealing selective binding of the analogues to DAT transfected CAD cells. Single particle tracking experiments were performed with JF₅₄₉-conjugated DG3–80 (3) and JF₆₄₆-conjugated DG4–91 (4) on DAT transfected CAD cells enabling quantification and categorization of the dynamic behavior of DAT into four distinct motion classes (immobile, confined, Brownian, and directed). Finally, we show that the ligands can be used in direct stochastic optical reconstruction microscopy (dSTORM) experiments permitting further analyses of DAT distribution on the nanoscale. In summary, these novel fluorescent ligands are promising new tools for studying DAT localization and regulation with single-molecule resolution.

Graphical Abstract

INTRODUCTION

Dopamine (DA) is an important neurotransmitter in the central nervous system where it is involved in the regulation of locomotion, reward mechanisms, learning, and cognition¹. Alteration in DA signaling is of central importance in several neurodegenerative and neuropsychiatric diseases, such as schizophrenia, bipolar disorder, attention-deficit/hyperactivity disorder (ADHD), autism, and Parkinson’s disease (PD)^2–7. One critical component in regulating DA homeostasis is the dopamine transporter (DAT), which terminates DA signaling and ensures maintenance of a reusable pool of DA^8,9. DAT is the primary target for amphetamines, substrates that are actively transported by DAT, and cocaine and methylphenidate, blockers of DAT transport^8,9. As DAT is the primary mechanistic target for both therapeutic agents (e.g., methylphenidate and amphetamine for ADHD) and illicit/abused substances (e.g., psychostimulants such as cocaine and methamphetamine), it is of interest to understand the mechanisms underlying these actions so that safer therapeutics may be developed. We are especially interested in the development of atypical DAT inhibitors that do not have the abuse liability of cocaine or methamphetamine and may be useful as part of a therapeutic strategy to treat psychostimulant use disorders^10,11.

In order to gain further insight into the function and cellular distribution of DAT, approaches need to be developed that allow the study of the molecular and cellular processes that govern the activity and availability of this key regulatory protein in the presynaptic nerve terminals with high precision and resolution. Recent advancements in fluorescent imaging techniques have demonstrated the strength of single-molecule microscopy, including methods such as dSTORM (direct stochastic optical reconstruction microscopy) and PALM (photoactivated localization microscopy), to improve spatial resolution^12–17, and single-particle tracking to study the dynamic behavior of individual molecules in the plasma membrane of living cells^18–20. However, to employ such methods for studying endogenously expressed DAT, new fluorescent cocaine analogues need to be developed that are tagged with novel fluorescent dyes better suited for single-molecule microscopy by displaying natural switching of a fluorophore between fluorescent and dark states (blinking behavior) and/or high photostability¹⁷. Such fluorescent ligands would permit the detection and tracking of endogenously expressed single DAT proteins, which, in turn, allows the study of the plasma membrane surface dynamics and distribution of DAT in dopaminergic neurons.

Previously, we have demonstrated that the N-position of the high-affinity cocaine analogue, MFZ2–12 (1), tolerates substitution with a rhodamine-labelled linker, maintaining high-affinity (JHC1–064 (2):DAT K_i = 18 nM)²¹ and selectivity for DAT (Figure 1)²². Fluorescent cocaine analogue 2 has proven to be extremely useful for visualizing DAT in both transfected cells and neurons. Both DAT and norepinephrine transporter (NET) trafficking and post-endocytotic sorting have also been revealed using 2 and related fluorescent cocaine analogues^22,23. Furthermore, 2 has provided evidence for an outward facing conformation of DAT in neuronal filipodia²⁴ and has been used to study missense DAT mutations from a patient with PD and ADHD²⁵.

Figure 1.

Molecular structure of a high-affinity cocaine analog and its fluorescent derivatives

In this study, we have utilized the fluorescent properties of the next-generation rhodamine-based Janelia Fluor dyes JF₅₄₉ and JF₆₄₆, which have previously been shown to be well suited for single-molecule microscopy²⁶. Connecting these dyes with 1 via varying linkers led to four novel probes, two of which showed excellent utility in cell-based experiments: DG3–80 (3) and DG4–91 (4) (Figure 1). Herein, we demonstrate that these fluorescent ligands bind with high affinity to DAT in both heterologous cell systems and DA neurons while maintaining their fluorescent properties, making these tools ideal for both single-particle tracking and dSTORM experiments.

RESULTS AND DISCUSSION

Building on the successful molecular design of 2²¹, we set out to synthesize analogues where the sulforhoamine B (i.e., rhodamine red) fluorophore is replaced with JF₅₄₉ and JF₆₄₆ using the same attachment position on the tropane ring of 1. To accommodate the increased hydrophobicity of the Janelia Fluor dyes, a short polyethylene glycol (PEG) linker was first considered such that the linker length was comparable to that of 2. Analogues with other hydrophilic linkers were also synthesized in order to identify optimal linker length and chemical space.

Chemical Synthesis of Novel Fluorescent Cocaine Analogues:

High-affinity cocaine analogue 1 was synthesized from (–)-cocaine as previously described²⁷. N-Demethylation of 1 using 1-chloroethyl chloroformate (ACE-Cl) followed by methanolysis gave the nor-compound, 5²⁷. As shown in Scheme 1, alkylation of 5 with the PEG₃-linker (6²⁸) gave the PEGylated compound 8. In situ deprotection with TFA and coupling of 6-carboxy-JF₅₄₉²⁶ or 6-carboxy-JF₆₄₆²⁶ with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC·HCl) and hydroxybenzotriazole (HOBt) gave 3 and DG3–63 (10), respectively. In a similar manner described for 3 and 10, a longer PEG linked analogue was prepared, by alkylation of 5 with the PEG₆-linker (7²⁸) to give 9, which was then deprotected with TFA, followed by EDC/HOBt-coupling with 6-carboxy-JF₆₄₆ to afford 4, the PEG₆ version of fluorescent cocaine analogue 10. A piperazine-based linker was also prepared based on an earlier fluorescent analogue of 1 (unpublished data). Reductive amination of 5 with linker 11, followed by phthalimide deprotection of 12, gave the piperazine compound 13 with a terminal amine. EDC/HOBt-coupling with 6-carboxy-JF₆₄₆ gave DG2–63 (14). The JF₆₄₆-conjugated 4, 10, and 14 were isolated as colorless films, or a pale teal foam in the case of 14, indicating that the colorless lactone conformation is dominant for these JF₆₄₆ analogues in isolation. Elution on silica during purification revealed the dark blue color of the colored zwitterionic state. In contrast, the JF₅₄₉-conjugated 3 was isolated as a pinkish-red film, consistent with its propensity to remain in the colored zwitterionic state²⁶.

Scheme 1.

^aReagents and conditions: (a) 6 or 7, triethylamine, acetonitrile, reflux, overnight, 57% (8), 78% (9); (b) (i) TFA, DCM, rt, 30 min (ii) 6-carboxy-JF₅₄₉ or 6-carboxy-JF₆₄₆·TFA, EDC·HCl, HOBt, N,N-diisopropylethylamine, DCM or DMF, rt, overnight, 50% (3), 20% (10), 31% (4); (c) 11, sodium triacetoxyborohydride, DCM, rt, 79% (12); (d) hydrazine, ethanol, reflux, 20 h, 67% (13); (e) 6-carboxy-JF₆₄₆, EDC·HCl, HOBt, N,N-diisopropylethylamine, chloroform, rt, 2 d, 67% (14).

Pharmacological Profiling of Fluorescent Cocaine Analogues:

The binding affinities of 3, 4, 10 and 14 for DAT were determined with radioligand binding assays in membrane preparations from rat striatum and compared to 1 and 2²⁹. Compound 3 had the highest affinity for DAT (K_i = 6 nM) followed by 4 (K_i = 28 nM) and 14 (K_i = 38 nM) (Table 1). Compound 10 showed a decreased affinity for DAT with K_i = 112 nM (Table 1). Overall, the fluorescently labelled ligands presented a reduced affinity for DAT as compared to the parent compound 1 (K_i = 0.8 nM), but similar to that of 2 (K_i = 16 nM) (Table 1). Compound 1 was also a potent reuptake inhibitor for the closely associated NET and serotonin transporters (SERT) (Table 1)^30,31, allowing for the design of fluorescent analogues suitable for imaging experiments focused on not only DAT but also NET and SERT. In order to determine the binding affinities of 3, 4, 10 and 14 for NET and SERT, membranes from rat midbrain and prefrontal cortex, respectively, were utilized²⁹. The secondary interactions of the fluorescent portion of 2–4, 10 and 14 resulted in moderately decreased binding affinity for NET as compared to the other monoamine transporters. Meanwhile, all of these analogues had comparable relative binding affinities for DAT and SERT (Table 1).

Table 1. Binding affinities of the fluorescent analogues for DAT, NET, and SERT and their ability to inhibit [³H]DA uptake.

Data are indicated as mean ± S.E.M. from 3 independent experiments.

	pKi ± S.E.M. (nM)			pIC50 ± S.E.M. (nM)
	DAT [3H]WIN 35,428	NET [3H]Nisoxetine	SERT [3H]Citalopram	DAT [3H]DA
1	9.10 ± 0.23 (0.8)	8.18 ± 0.05 (7)	7.99 ± 0.20 (10)
2	7.79 ± 0.09 (16)	7.10 ± 0.01 (79)	7.80 ± 0.08 (16)
3	8.22 ± 0.10 (6)	6.65 ± 0.13 (225)	7.38 ± 0.08 (42)	7.55 ± 0.04 (28)
10	6.95 ± 0.08 (112)	6.33 ± 0.05 (468)	6.82 ± 0.17 (153)	5.94 ± 0.12 (1147)
4	7.55 ± 0.03 (28)	6.62 ± 0.06 (241)	7.37 ± 0.05 (43)	6.99 ± 0.15 (102)
14	7.42 ± 0.02 (38)	5.92 ± 0.03 (1205)	7.42 ± 0.08 (38)	6.80 ± 0.17 (160)

As the binding experiments were conducted on rat brain membranes, it was necessary to ensure that the ligands effectively bind to surface DAT in whole cells. In order to assess this we conducted [³H]DA uptake experiments in CAD cells transiently expressing human DAT (hDAT). All compounds tested inhibited [³H]DA uptake (Table 1).

Visualization of DAT in CAD Cells with the Fluorescent Cocaine Analogues:

To determine the specificity and labeling capability for DAT, the fluorescent cocaine analogues were visualized in CAD cells transiently expressing enhanced green fluorescent protein-tagged hDAT (EGFP-hDAT) using both confocal and TIRF (Total Internal Reflection Fluorescence) microscopy. The cells were incubated with the analogues for 10 min at room temperature (rt) after 10 min of pre-incubation with either vehicle or DAT blocker nomifensine. After washing away excess compound, the cells were fixed and immediately imaged.

Utilizing confocal microscopy, it was demonstrated that both the JF₅₄₉-conjugated 3 and JF₆₄₆-conjugated 4 selectively bind to the surface of CAD cells that transiently express EGFP-hDAT and not to untransfected cells (Figures 2 and 3). Furthermore, binding to the EGFP-hDAT expressing cells was completely blocked by pre-incubation with nomifensine. Unfortunately, any specific surface labeling of the JF₆₄₆-conjugated ligands 10 and 14 was overwhelmed by signal from internal compartments in both EGFP-hDAT-expressing and wild-type cells (Supplementary Figure 1). JF₆₄₆ exhibits a higher propensity to adopt the lipophilic lactone form and JF₆₄₆-based ligands are typically highly cell-permeable. Unlike the longer PEG linker in 4, the short PEG and piperazine linkers in 10 and 14 were apparently not sufficient to prevent cellular entry. However, it is also possible that the nonfluorescent lactone form of JF₆₄₆ could be stabilized in 10 and 14 though interaction with the linkers, MFZ2–12, DAT, or other environmental factors³².

Figure 2. Visualization of the JF₅₄₉-conjugated 3 in CAD cells transiently expressing EGFP-hDAT utilizing confocal microscopy.

Live CAD cells that transiently express EGFP-hDAT were incubated for 10 min with 10 nM 3 in the absence or presence of 10 µM nomifensine. Images were acquired upon fixation of the cells. The images shown are representatives of at least 3 independent experiments.

Figure 3. Visualization of the JF₆₄₆-conjugated 4 in CAD cells transiently expressing EGFP-hDAT utilizing confocal microscopy.

Live CAD cells that transiently express EGFP-hDAT were incubated for 10 min with 10 nM 4 in the absence or presence of 10 µM nomifensine. Images were acquired upon fixation of the cells. The images shown are representatives of at least 3 independent experiments.

TIRF microscopy was utilized to ensure that the fluorescent ligands labelled surface DAT in a homogenous fashion (Figure 4 and Supplementary Figure 2). In corroboration with the confocal microscopy, surface labeling was observed for both 3 and 4 on EGFP-hDAT-expressing CAD cells. This labeling was not observed in untransfected CAD cells. Contrasting the confocal microscopy data, specific surface labeling of EGFP-hDAT could be observed for 10, although the fluorescence was much fainter than for its counterpart 4. The low fluorescent intensity could explain why no surface labeling could be detected using confocal microscopy. As for 14, while there was surface labeling of EGFP-DAT expressing cells (which was absent on non-expressing cells), the labeling was very punctate. This could further indicate that the environment of the fluorophore is critical for its fluorescence, or that due to the lipophilic properties of the compound, it penetrated into the cell membrane. Because of the high and selective labeling of surface DAT with the JF₅₄₉-conjugated 3 and JF₆₄₆-conjugated 4, further experiments were only conducted with these ligands.

Figure 4. Visualization of the surface labeling of EGFP-hDAT-expressing CAD cells by 3 and 4 utilizing TIRF microscopy.

Untransfected CAD cells or CAD cells that transiently express EGFP-hDAT were incubated for 10 min with 10 nM 3 (A) or 4 (B). Images were acquired upon fixation of the cells. The images shown are representatives of at least 3 independent experiments.

Visualization of endogenous DAT in dopaminergic neurons with 3 and 4:

To assure that both 3 and 4 also bind to DAT in dopaminergic neurons, cultured neurons were treated with each of the fluorescent ligands, fixed, permeabilized, and immunolabeled with a rat monoclonal antibody targeting the DAT N-terminus, and then labelled with Alexa647- or Alexa568-conjugated secondary antibodies, respectively. Subsequently, images were acquired utilizing a confocal microscope, to assess whether the labeling with the fluorescent ligands overlay that of DAT immunolabeling (Figure 5). As expected from the CAD cell data, both fluorescent compounds selectively labelled DAT-expressing neurons.

Figure 5. Visualization of 3, 4, and DAT in primary dopaminergic neurons utilizing confocal microscopy.

Live dopaminergic neurons were incubated for 10 min with 10 nM 3 (A) or 4 (B). After washing and fixation, DAT was labeled with a rat monoclonal antibody targeting the DAT N-terminus, and then with Alexa647- or Alexa568-conjugated secondary antibodies, respectively. Images shown are representatives of at least 2 independent experiments.

Utility of 3 and 4 for Tracking Single DAT Molecules on the Surface of CAD Cells:

It has become evident that the surface diffusion of proteins, such as transporters, can play an important role in regulating their localization and function. For example, for the glutamate transporter GLT-1, it has been demonstrated that surface diffusion plays a critical role in its ability to effectively remove glutamate from the synapse and thus for the regulation of excitatory post synaptic currents¹⁹. Furthermore, for DAT, it has been shown that the surface diffusion is altered upon treatment with amphetamine, a common component of ADHD medications, and that it may be dysregulated by ADHD-associated DAT mutations³³.

It should be noted that past studies that have explored the role of DAT surface diffusion have been conducted with a cocaine analogue linked to a streptavidin-conjugated quantum dot^33,34 or quantum dot-labelled antibodies¹⁹. While quantum dots are brighter and more stable than the fluorophores used in this paper, the large size (15–20 nm) may impede the surface diffusion of DAT³⁵. Therefore, while the fluorescent ligands used in this study show lower brightness and photostability than quantum dot-based labels, their small size have the potential to provide more accurate measurements of DAT surface diffusion.

In order to determine the surface diffusion of DAT with 3 and 4, CAD cells transiently expressing EGFP-hDAT were sparsely labelled for 10 min with 1 nM fluorescent ligand, and then imaged with a TIRF microscope at 37 °C (Figure 6; for example videos of 3 and 4 see Supplementary Videos 1 and 2, respectively). Consecutively, the obtained videos were processed and analyzed utilizing u-track and DC-MSS software. In order to enable image-processing of the large datasets without multi-day processing times on desktop computational units, improvements were made to the memory management of the u-track software which have been incorporated into version 2.3 and later (see Methods).

Figure 6. Surface diffusion of EGFP-hDAT on CAD cells labelled with 3 or 4.

A) Representative single-molecule microscopy images showing tracking of 3 (red)- and 4 (blue)-labelled EGFP-hDAT. B-C) Bar graphs indicating the distinct DAT populations (B, %) based on surface diffusion behavior and the speed by which DAT diffuses (C, µm/s²). All data is shown as mean ± S.E.M. calculated from 14 cells from 3 independent experiments.

Utilizing the u-track and DC-MSS software, the surface diffusion of DAT can, for the first time, be quantified and categorized into immobile and mobile (confined, Brownian, and directed) motion^36,37. Measurement of DAT surface diffusion with both 3 and 4 indicated that DAT predominantly diffuses in a Brownian fashion (66%), followed by a smaller fraction that diffuses in a confined manner (18%). Only 14% of DAT was found to be immobile, and 2% appeared to be diffusing in a directed manner (Figure 6B).

Within each mobility class, the speed by which DAT diffuses (diffusion coefficient (D, µm²/s)), Figure 6C) can be determined. Interestingly, the overall diffusion coefficient observed (D = 0.24 µm²/s) was greater from that observed in previous studies that determined the diffusion coefficient of DAT^19,33,34. These studies, however, utilized quantum dots to visualize and track DAT. As it has been shown that quantum dots sterically hinder lateral mobility and can have a major impact on the frequency by which proteins can transition between different mobility classes, it is not unlikely that the bulky size of these quantum dots slowed down DAT diffusion³⁵. However, further studies are necessary to confirm if this is actually the case, and whether this can explain the differences observed between our study and previous studies. In summary, both 3 and 4 can effectively label DAT for single-particle tracking studies.

Visualization of 3- and 4-labelled EGFP-DAT by dSTORM in CAD cells:

In recent years, novel super resolution microscopy techniques, such as dSTORM, have resulted in an order of magnitude improvement in spatial resolution as compared to conventional light microscopy^12–17. The increased spatial resolution obtained with dSTORM is achieved by utilizing reducing agents to transfer the majority of fluorophores to a reversible dark state, which can then be stochastically activated¹⁷. This then allows the precise determination of the position of the activated fluorophores from their point spread function¹⁷.

Utilizing dSTORM, we have previously demonstrated that endogenously expressed membrane-bound DAT is dynamically sequestrated into cholesterol-dependent nanodomains within CAD cells as well as in neuronal projections and presynaptic varicosities of cultured dopaminergic neurons³⁸. Here, we show that both 3 and 4 can also be used to study the membrane distribution of DAT in CAD cells (Figure 7). In the presence of reducing agents and an oxygen scavenging system, we obtained stochastic blinking for both 3 and 4 (for examples see Supplementary Videos 3 and 4, respectively). The positions of the blinking fluorophores could then be determined with Thunderstorm³⁹ (Figure 7). Subsequent analysis with the density-based analysis algorithm DBSCAN⁴⁰ allowed the determination of the fraction of DAT residing in nanodomains (Figure 7C), which was similar for 3- and 4-labelled DAT (0.37 and 0.31, respectively). Of note, while similar clusters were observed previously through antibody-labeled and Dronpa-conjugated DAT³⁸, the absolute number of clustered localization varies between these methods due to differences in fluorophores, and labeling, imaging and identification methods.

Figure 7. Visualization of EGFP-DAT labeled with 3 and 4 by dSTORM in CAD cells.

A) Representative widefield images of the EGFP signal and dSTORM images of the 3 or 4 signal. B) Example of localizations (black) detected from 4. C) DBSCAN clustering analysis applied to the localizations from example B, where red indicates clustered localizations. Images shown are representatives of 1 independent experiment.

An interesting future goal is the use of ligands such as 3 and 4 to label endogenous DAT for single-molecule microscopy experiments, such as single particle tracking and dSTORM, in live cultured dopaminergic neurons. However, as our dSTORM experiments were conducted on fixed samples, it remains to be determined whether the reducing agents and high laser power used have any negative effect on the viability of live dopaminergic neurons. Alternatively, it may be possible to use other reducing agents to elicit blinking of JF₅₄₉ and JF₆₄₆^17,26, or modify the fluorophores so they spontaneously blink at physiological pH without additives⁴¹.

CONCLUSION

In conclusion, by linking 1 to the next-generation rhodamine-based fluorescent Janelia Fluor dyes JF₅₄₉ and JF₆₄₆, we have developed novel fluorescent DAT ligands with optimized linkers that are highly suitable for single-molecule localization microscopy. The ligands developed, 3 and 4, bind with high affinity to DAT expressed in catecholaminergic CAD cells as well as to endogenous DAT in primary dopaminergic neurons. Importantly, JF₅₄₉ and JF₆₄₆ retained their fluorescent properties upon conjugation to 1, making them well suited for single-molecule microscopy experiments, such as single particle-tracking and dSTORM. Utilizing these novel probes, we have now been able, for the first time, to quantify and categorize the dynamic behavior of DAT into four motion classes (immobile, confined, Brownian, and directed) providing a solid foundation from which to further explore the role of DAT surface diffusion under both physiological and pathophysiological conditions.

METHODS

Chemical Synthesis.

General.

All chemicals and solvents were purchased from chemical suppliers unless otherwise stated and used without further purification. 6-carboxy-JF₅₄₉ and 6-carboxy-JF₆₄₆·TFA were provided by Dr. Luke Lavis (Janelia Research Campus). ¹H NMR spectra were acquired using a Varian Mercury Plus 400 spectrometer at 400 MHz. Chemical shifts are reported in parts-per-million (ppm) and referenced according to deuterated solvent for ¹H NMR spectra (CDCl₃, 7.26). Reaction conditions are unoptimized. All column chromatography was performed using a Teledyne Isco CombiFlash RF flash chromatography system with silica gel disposable flash columns (20–40 or 40–60 microns). Solvent system DMA = DCM/methanol/ammonium hydroxide (e.g., 10% DMA = v/v/v 90 : 0.9 : 0.1). All final fluorescent products were partitioned into tared amber vials, concentrated in vacuo, and stored under argon (~1 mg/vial). HRMS (mass error within 5 ppm) and MS/MS fragmentation analysis were performed on a LTQ-Orbitrap Velos (Thermo-Scientific, San Jose, CA) coupled with an ESI source in positive ion mode to confirm the assigned structures and regiochemistry. Safety statement: no unexpected or unusually high safety hazards were encountered.

Methyl (1R,2S,3S,5S)-3-(3,4-dichlorophenyl)-8-(2,2-dimethyl-4-oxo-3,8,11-trioxa-5-azatridecan-13-yl)-8-azabicyclo[3.2.1]octane-2-carboxylate (8).

To a solution of 5²⁷ (0.335 g, 1.02 mmol) and 6²⁸ (0.316 g, 0.965 mmol) in anhyd acetonitrile (10 mL) was added triethylamine (0.17 mL, 1.2 mmol). The reaction was heated to reflux under argon for 24 h, concentrated in vacuo, re-dissolved in DCM and washed with 10% ammonium hydroxide. The aq phase was extracted twice more with DCM, and the combined organic phases were washed with brine, dried over magnesium sulphate, filtered, and concentrated in vacuo. The crude product was purified by flash chromatography (0 – 30% DMA) to give the title compound as a viscous yellow oil (0.318 g, 57% yield). ¹H NMR δ: 7.33–7.31 (m, 2H), 7.08 (d, 1H, J = 7.8 Hz), 5.03 (br s, 1H), 3.78 (m, 1H), 3.64 (s, 3H), 3.59–3.45 (m, 8H), 3.42 (m, 1H), 3.32 (m, 2H), 2.95–2.88 (m, 2H), 2.53–2.46 (m, 3H), 2.17–2.00 (m, 2H), 1.74–1.59 (m, 3H), 1.44 (s, 9H).

Methyl (1R,2S,3S,5S)-3-(3,4-dichlorophenyl)-8-(2,2-dimethyl-4-oxo-3,8,11,14,17,20-hexaoxa-5-azadocosan-22-yl)-8-azabicyclo[3.2.1]octane-2-carboxylate (9).

To a solution of 5²⁷ (0.503 g, 1.60 mmol) and 7²⁸ (0.740 g, 1.61 mmol) in anhyd acetonitrile (10 mL) was added triethylamine (0.27 mL, 1.9 mmol). The reaction was heated to reflux under argon for 16 h, concentrated in vacuo, re-dissolved in DCM and washed with 10% ammonium hydroxide. The aq phase was extracted twice more with DCM, and the combined organic phases were washed with brine, dried over magnesium sulphate, filtered, and concentrated in vacuo. The crude product was purified by flash chromatography (0 – 30% DMA) to give the title compound as a viscous tan oil (0.832 g, 78% yield). ¹H NMR δ: 7.32 (m, 2H), 7.07 (d, 1H, J = 8.4 Hz), 5.06 (br s, 1H), 4.37 (m, 1H), 3.77 (m, 1H), 3.65–3.42 (m, 23H), 3.30 (m, 2H), 2.94–2.86 (m, 2H), 2.49–2.44 (m, 3H), 2.20–2.00 (m, 2H), 2.75–2.55 (m, 3H), 1.44 (s, 9H).

2-(3-(Azetidin-1-ium-1-ylidene)-6-(azetidin-1-yl)-3H-xanthen-9-yl)-4-((2-(2-(2-((1R,2S,3S,5S)-3-(3,4-dichlorophenyl)-2-(methoxycarbonyl)-8-azabicyclo[3.2.1]octan-8-yl)ethoxy)ethoxy)ethyl)carbamoyl)benzoate (3).

Compound 8 (4.9 mg, 9.0 μmol) was stirred in DCM (0.75 mL) and TFA (0.25 mL) for 1 h at rt. The resulting solution was concentrated in vacuo, re-dissolved in DCM (0.5 mL) and N,N-diisopropylethylamine (0.02 mL), and used as is in the subsequent step. To a solution of 6-carboxy-JF₅₄₉²⁶ (3.6 mg, 8.0 μmol) and N,N-diisopropylethylamine (0.02 mL, 0.1 mmol) in anhyd DMF (1 mL) under argon at rt was added EDC·HCl (2.3 mg, 12 μmol) and HOBt (1.2 mg, 9 μmol). The reaction stirred for 30 min, then the previous solution was added, and the reaction continued to stir for 18 h at rt. Following concentration in vacuo at 60 °C, the crude product was re-dissolved in DCM, washed with 10% ammonium hydroxide, brine, then dried over magnesium sulphate, filtered and concentrated. Purification by flash chromatography (0 – 8% DMA) provided the title compound as a pinkish-red film (3.5 mg, 50% yield). ¹H NMR δ: 8.01 (d, 1H, J = 8.2 Hz), 7.97 (d, 1H, J = 7.4 Hz), 7.55 (s, 1H), 7.32–7.30 (m, 2H), 7.06 (d, J = 9.0 Hz, 1H), 6.76 (br s, 1H), 6.52 (d, 2H, J = 8.2 Hz), 6.20 (s, 2H), 6.07 (d, 2H, J = 8.2 Hz), 3.92–3.89 (m, 8H), 3.72 (m, 1H), 3.66–3.37 (m, 14H), 2.96–2.85 (m, 2H), 2.47–2.34 (m, 7H), 2.10–1.90 (m, 2H), 1.68–1.60 (m, 3H). HRMS: found m/z = 881.3103 (MH⁺), calcd for C₄₈H₅₀Cl₂N₄O₈ 881.3079 (MH⁺).

Methyl (1R,2S,3S,5S)-8-(2-(2-(2-(3,7-di(azetidin-1-yl)-5,5-dimethyl-3’-oxo-3’H,5H-spiro[dibenzo[b,e]siline-10,1’-isobenzofuran]-6’-carboxamido)ethoxy)ethoxy)ethyl)-3-(3,4-dichlorophenyl)-8-azabicyclo[3.2.1]octane-2-carboxylate (10).

Compound 8 (9.8 mg, 18.0 μmol) was stirred in DCM (0.50 mL) and TFA (0.50 mL) for 1 h at rt. The resulting solution was concentrated in vacuo, re-dissolved in DCM (0.5 mL) and N,N-diisopropylethylamine (0.04 mL), and used as is in the subsequent step. To a solution of (6-carboxy-JF₆₄₆·TFA)²⁶ (10.0 mg, 16.0 μmol) and N,N-diisopropylethylamine (0.04 mL, 0.2 mmol) in anhyd DCM (1 mL) under argon at rt was added EDC·HCl (4.6 mg, 24 μmol) and HOBt (2.4 mg, 18 μmol). The reaction stirred for 1 h, then the previous solution was added, and the reaction continued to stir for 24 h at rt. The pale green reaction mixture was transferred to a separatory funnel with DCM, washed with brine (1x), and the brine was extracted with fresh DCM (1x). The combined organic phases were dried over magnesium sulphate and concentrated in vacuo. Purification by flash chromatography (0 – 8% DMA) provided the title compound as a colorless film (3.2 mg, 20% yield). ¹H NMR δ: 7.97 (d, 1H, J = 7.8 Hz), 7.88 (d, 1H, J = 7.8 Hz), 7.71 (s, 1H), 7.32–7.29 (m, 2H), 7.05 (d, 1H, J = 7.8 Hz), 6.78 (br s, 1H), 6.74 (d, 2H, J = 8.6 Hz), 6.66 (s, 2H), 6.24 (d, 2H, J = 9.0 Hz), 3.91–3.87 (m, 8H), 3.70 (m, 1H), 3.64–3.58 (m, 8H), 3.48 (s, 3H), 3.47–3.40 (m, 2H), 3.35 (m, 1H), 2.92–2.89 (m, 1H), 2.84 (m, 1H), 2.46–2.33 (m, 7H), 2.10–1.90 (m, 2H), 1.70–1.55 (m, 3H), 0.63 (s, 3H), 0.57 (s, 3H). HRMS: found m/z = 923.3354 (MH⁺), calcd for C₅₀H₅₆Cl₂N₄O₇Si 923.3368 (MH⁺).

Methyl (1R,2S,3S,5S)-8-(1-(3,7-di(azetidin-1-yl)-5,5-dimethyl-3’-oxo-3’H,5H-spiro[dibenzo[b,e]siline-10,1’-isobenzofuran]-6’-yl)-1-oxo-5,8,11,14,17-pentaoxa-2-azanonadecan-19-yl)-3-(3,4-dichlorophenyl)-8-azabicyclo[3.2.1]octane-2-carboxylate (4).

Compound 9 (10.1 mg, 14.9 μmol) was stirred in DCM (0.50 mL) and TFA (0.20 mL) for 1 h at rt. The resulting solution was concentrated in vacuo, re-dissolved in DCM (1 mL) and N,N-diisopropylethylamine (0.06 mL), and used as is in the subsequent step. To a solution of (6-carboxy-JF₆₄₆·TFA)²⁶ (9.1 mg, 14.9 μmol) and N,N-diisopropylethylamine (0.04 mL, 0.2 mmol) in anhyd DCM (1 mL) under argon at rt was added EDC·HCl (4.3 mg, 22 μmol) and HOBt (2.0 mg, 14.9 μmol). The reaction stirred for 1 h, then the previous solution was added, and the reaction continued to stir for 24 h at rt. The pale green reaction mixture was transferred to a separatory funnel with DCM, washed with brine (1x), and the brine was extracted with fresh DCM (1x). The combined organic phases were dried over magnesium sulphate and concentrated in vacuo. Purification by flash chromatography (0 – 8% DMA) provided the title compound as a colorless film (4.8 mg, 31% yield). ¹H NMR δ: 7.98 (s, 2H), 7.76 (s, 1H), 7.32–7.30 (m, 2H), 7.20 (br s, 1H), 7.07 (d, 1H, J = 10.6 Hz), 6.75 (d, 2H, J = 8.6 Hz), 6.65 (d, 2H, J = 2.7 Hz, 2H), 6.25 (dd, 2H, J = 2.7, 8.8 Hz), 3.89 (m, 8H), 3.72 (m, 1H), 3.64–3.39 (m, 26H), 2.94–2.89 (m, 1H), 2.85 (m, 1H), 2.50–2.32 (m, 7H), 2.15–1.95 (m, 2H), 1.70–1.60 (m, 3H), 0.63 (s, 3H), 0.57 (s, 3H). HRMS: found m/z = 1055.4166 (MH⁺), calcd for C₅₆H₆₈Cl₂N₄O₁₀ 1055.4155 (MH⁺).

Methyl (1R,2S,3S,5S)-3-(3,4-dichlorophenyl)-8-(2-(4-(4-(1,3-dioxoisoindolin-2-yl)butyl)piperazin-1-yl)ethyl)-8-azabicyclo[3.2.1]octane-2-carboxylate (12).

To a solution of 5²⁷ (0.314 g, 1.0 mmol) and compound 11 (0.411 g, 1.2 mmol) in anhyd DCM (10 mL) was added sodium triacetoxyborohydride (0.420 g, 2.0 mmol) and cat acetic acid (5 drops), and the reaction stirred at rt for 22 h. Then, the reaction mixture was washed with saturated sodium bicarbonate, and the aq phase was extracted with DCM (3x). The combined organic phases were washed with brine, dried over magnesium sulphate, and concentrated in vacuo. Purification by flash chromatography (0 – 10% DMA) provided the title compound as a white solid (0.496 g, 79% yield). ¹H NMR δ: 7.84–7.82 (m, 2H), 7.71–7.69 (m, 2H), 7.32–7.30 (m, 2H), 7.09 (d, 1H, J = 8.6 Hz), 3.70 (m, 3H), 3.51 (s, 3H), 3.39 (m, 1H), 2.95–2.85 (m, 2H), 2.51–2.32 (m, 15H), 2.10–1.90 (m, 2H), 1.73–1.50 (m, 7H).

Methyl (1R,2S,3S,5S)-8-(2-(4-(4-aminobutyl)piperazin-1-yl)ethyl)-3-(3,4-dichlorophenyl)-8-azabicyclo[3.2.1]octane-2-carboxylate (13).

To a suspension of compound 12 (0.488 g, 0.778 mmol) in ethanol (10 mL) was added neat hydrazine (0.04 mL, 1.2 mmol) at rt, then the reaction was heated to reflux for 20 h. The resulting precipitate was vacuum filtered, and the filtrate was concentrated in vacuo. Purification by flash chromatography (0 – 30% DMA) provided the title compound as a light yellow oil (0.261 g, 67% yield). ¹H NMR δ: 7.30–7.28 (m, 2H), 7.07–7.05 (d, 1H, J = 10.6 Hz), 3.70 (m, 1H), 3.49 (s, 3H), 3.38 (m, 1H), 2.94–2.84 (m, 2H), 2.69 (t, 2H, J = 6.6 Hz), 2.50–2.19 (m, 17H), 2.10–1.90 (m, 2H), 1.69–1.57 (m, 3H), 1.52–1.44 (m, 4H).

Methyl (1R,2S,3S,5S)-8-(2-(4-(4-(3,7-di(azetidin-1-yl)-5,5-dimethyl-3’-oxo-3’H,5H-spiro[dibenzo[b,e]siline-10,1’-isobenzofuran]-6’-carboxamido)butyl)piperazin-1-yl)ethyl)-3-(3,4-dichlorophenyl)-8-azabicyclo[3.2.1]octane-2-carboxylate (14; DG2–63).

To a solution of (6-carboxy-JF₆₄₆·TFA)²⁶ (10.0 mg, 16 μmol) and N,N-diisopropylethylamine (0.04 mL, 0.2 mmol) in anhyd chloroform (1 mL) under argon at rt was added EDC·HCl (4.6 mg, 24 μmol) and HOBt (2.4 mg, 18 μmol). The reaction stirred for 1.5 h, then compound 13 (9.2 mg, 18 μmol) was added, and the reaction continued to stir for 24 h at rt. An additional dose of EDC·HCl (4.0 mg, 21 μmol) was added, and the reaction continued to stir for another 24 h at rt. Purification by flash chromatography (0 – 5% DMA), followed by preparative TLC (2000 microns, 10% DMA) provided the title compound as a teal foam solid (10.4 mg, 67% yield). ¹H NMR δ: 8.07 (br d, 1H, J = 7.8 Hz), 9.97 (d, 1H, J = 7.8 Hz), 7.74 (s, 1H), 7.32 (d, 1H, J = 8.6 Hz), 7.29 (s, 1H), 7.06 (d, 1H, J = 8.2 Hz), 6.75 (d, 2H, J = 8.6 Hz), 6.65 (s, 2H), 6.25 (d, 2H, J = 8.6 Hz), 3.88 (m, 8H), 3.73 (m, 1H), 3.49 (s, 3H), 3.49–3.43 (m, 3H), 3.10–2.80 (m, 9H), 2.80–2.40 (m, 8H), 2.39–2.28 (m, 4H), 2.20–1.97 (m, 2H), 1.75–1.60 (m, 7H), 0.63 (s, 3H), 0.56 (s, 3H). HRMS: found m/z = 975.42 (MH⁺), calcd for C₅₄H₆₄Cl₂N₆O₅Si 975.42 (MH⁺).

Radioligand Binding Studies

DAT Binding Assay.

Frozen striatum membranes, dissected from male Sprague−Dawley rat brains (supplied on ice by Bioreclamation, Hicksville, NY), were homogenized in 20 volumes (w/v) of ice cold modified sucrose phosphate buffer (0.32 M sucrose, 7.74 mM Na₂HPO₄, and 2.26 mM NaH₂PO₄, pH adjusted to 7.4) using a Brinkman Polytron (Setting 6 for 20 s) and centrifuged at 48,400 x g for 10 min at 4 °C. The resulting pellet was resuspended in buffer, recentrifuged, and suspended in ice cold buffer again to a concentration of 20 mg/mL, original wet weight (OWW). Experiments were conducted in 96-well polypropylene plates containing 50 μL of various concentrations of the inhibitor, diluted using 30% DMSO vehicle, 300 μL of sucrose phosphate buffer, 50 μL of [³H]WIN 35,428⁴² (final concentration 1.5 nM; PerkinElmer Life Sciences, Waltham, MA), and 100 μL of tissue (2.0 mg/well OWW). All compound dilutions were tested in triplicate and the competition reactions started with the addition of tissue, and the plates were incubated for 120 min at 0–4 °C. Nonspecific binding was determined using 10 μM indatraline.

SERT Binding Assay.

Frozen stem membranes dissected from male Sprague−Dawley rat brains (supplied on ice by Bioreclamation, Hicksville, NY) were homogenized in 20 volumes (w/v) of 50 mM Tris buffer (120 mM NaCl and 5 mM KCl, adjusted to pH 7.4) at 25 °C using a Brinkman Polytron (at setting 6 for 20 s) and centrifuged at 48,400 x g for 10 min at 4 °C. The resulting pellet was resuspended in buffer, recentrifuged, and suspended in buffer again to a concentration of 20 mg/mL, OWW. Experiments were conducted in 96-well polypropylene plates containing 50 μL of various concentrations of the inhibitor, diluted using 30% DMSO vehicle, 300 μL of Tris buffer, 50 μL of [³H]citalopram (final concentration 1.5 nM; PerkinElmer Life Sciences, Waltham, MA), and 100 μL of tissue (2.0 mg/well OWW). All compound dilutions were tested in triplicate and the competition reactions started with the addition of tissue, and the plates were incubated for 60 min at rt. Nonspecific binding was determined using 10 μM fluoxetine.

NET Binding Assay.

Frozen prefrontal cortex dissected from male Sprague−Dawley rat brains (supplied on ice from Bioreclamation, Hicksville, NY) were homogenized in 20 volumes (w/v) of 50 mM Tris buffer (300 mM NaCl and 5 mM KCl, adjusted to pH 7.4) at 25 °C using a Brinkman Polytron (at setting 6 for 20 s). The tissue was centrifuged at 48,400 x g for 10 min at 4 °C. The resulting pellet was suspended in fresh buffer and centrifuged again. The final pellet was resuspended in cold binding buffer to a concentration of 80 mg/mL OWW. Experiments were conducted in glass assay tubes containing 50 μL of various concentrations of the inhibitor, diluted using 30% DMSO vehicle, 300 μL of Tris buffer, 50 μL of [³H]nisoxetine (final concentration 0.5 nM; Perkin-Elmer Life Sciences), and 100 μL of tissue (8.0 mg/tube OWW). The reaction was started with the addition of the tissue, and the tubes were incubated for 180 min at 0−4 °C. Nonspecific binding was determined using 10 μM desipramine.

Similarly to what previously described²⁹, for all binding assays, incubations were terminated by rapid filtration through Perkin Elmer Uni-Filter-96 GF/B (DAT and SERT) or Whatman GF/B filters (NET), presoaked in either 0.3% (SERT and NET) or 0.05% (DAT) polyethylenimine, using a Brandel 96-Well Plates Harvester Manifold or Brandel R48 filtering manifold (Brandel Instruments, Gaithersburg, MD). The filters were washed a total of 3 times with 3 mL (3 × 1 mL/well or 3 x 1 mL/tube) of ice-cold binding buffer. For DAT and SERT binding experiment 65 μL Perkin Elmer MicroScint20 Scintillation Cocktail was added to each filter well. For NET binding experiment, the filters were transferred in 24-well scintillation plates and 600 μL of CytoScint was added to each well. All the plates/filters were counted using a Perkin Elmer MicroBeta Microplate Counter. For each experiment, aliquots of the prepared radioligand solutions were measured to calculate the exact amount of radioactivity added, taking in account the experimentally determined top-counter efficiency for each radioligand. IC₅₀ values for each compound were determined from inhibition curves and K_i values were calculated using the Cheng-Prusoff equation⁴³. When a complete inhibition could not be achieved at the highest tested concentrations, K_i values have been extrapolated by constraining the bottom of the dose-response curves (= 0% residual specific binding) in the non-linear regression analysis. These analyses were performed using GraphPad Prism version 8.00 for Macintosh (GraphPad Software, San Diego, CA). K_d values for the radioligands were determined via separate homologous competitive binding or radioligand binding saturation experiments. pK_i (K_i) values were determined from at least 3 independent experiments performed in triplicate and are reported as mean ± SEM.

Cell Culture and Transfections

Cath-a-differentiated (CAD) cells were grown in media containing HAM’s F12 and DMEM (1:1) (GIBCO, Grand Island NY) supplemented with 10% (w/v) fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA) at 37 °C in a humidified 5% CO₂ atmosphere. The cells were transiently transfected with EGFP-DAT⁴⁴ with lipofectamine 2000 (Invitrogen, 1ug:3ul DNA:lipofectamine ratio) according to the manufacturer’s protocol 48 h prior to experiments. For imaging experiments, the cells were plated into polyornithine (PORN)-coated 8-well Nunc Lab-Tek II Chambered Coverglass (Thermofisher) at a density of 40,000 cells/well 24h prior to the experiment.

Dissection and Culturing of Rat Dopaminergic Neurons

Postnatally derived rat midbrain dopaminergic neurons were isolated and grown using a modified protocol from Rayport et al. (1992)⁴⁵. In short, the ventral midbrain was isolated from P1 to P2 Wister rats, and digested in a papain solution (16 mM NaCl, 5.4 mM KCl, 26 mM NaHCO₃, 2 mM NAH₂PO₄, 1 mM MgSO₄, EDTA, 25 mM glucose, 1 mM cysteine, 0.5 mM Kyrunate, and Papain 20 U/ml) for 30 min at 37 °C while being slowly superfused with a mixture of 95% O₂ and 5% CO₂. The digested tissue was gently triturated into single cells by mechanical force. The cells were then centrifuged (300g, 10min), and seeded in warm SF1C media (50% (v/v) modified Eagle’s medium (MEM), 40% (v/v) Dulbecco’s modified eagles medium (DMEM), 10% (v/v) F-12 (all from Invitrogen) supplemented with 1% (v/v) heat inactivated calf serum (FBS, 2.5 mg/ml, Invitrogen), 0.35% (w/v) d-glucose, 0.5 mM glutamine, 5 mM Kyrunic acid, Penicillin, Streptomycin, liquid catalse (0.05%), and DiPorzio⁴⁶ (containing insulin, transferrin, superoxide dismutase, progesterone, cortisol, Na₂SeO₃, and T3)) on a monolayer of glial cells grown on coverslips. Two hours after seeding, the cells were treated with glia derived neurotrophic factor (GDNF, 10 ng/ml). The following day 5-Fluorodeoxyurdine was added to the cell media to inhibit the proliferation of glia cells. The sample was used after 8–14 days in vitro (DIV).

[³H] Dopamine Uptake

CAD cells were seeded out into PORN-coated 96-well white clear bottom tissue culture plates (20,000 cells/well, Corning, USA) and transfected the following day with hDAT (50 ng/well). Dopamine uptake experiments were conducted 48 h after transfection using [³H] dopamine (30–60 Ci/mmol, PerkinElmer Life Sciences). The cells were equilibrated in uptake buffer (25 mM HEPES, 120 mM NaCl, 5 mM KCl, 1.2 mM CaCl₂, 1.2 mM MgSO₄, 1 mM ascorbic acid, and 5 mM D-Glucose; pH 7.4) for 20 min at rt. The cells were then incubated with various concentrations of the fluorescent ligands or 10µM nomifensine to determine non-specific uptake. Dopamine uptake was initiated through the addition of 20 nM [³H]DA and further incubation of 5 min at rt. Uptake was terminated through rapid aspiration and immediate washes with ice-cold uptake buffer. Upon removal of the final wash, Opti-Phase HiSafe3 scintillation fluid (PerkinElmer Life Sciences) was added to the wells and after 1 h of incubation at rt the samples were counted in a beta counter (MicroBeta², Perkin Elmer Life Sciences). The uptake data was analysed using nonlinear regression analysis in GraphPad Prism 7.

Confocal Microscopy and Immunocytochemistry

CAD Cells.

CAD cells were labelled with 10 nM of 3, 10, 4, or 14 for 10 min at rt in artificial cerebrospinal fluid (ACSF; 120mM NaCl, 5mM KCl, 2mM CaCl2, 2mM MgCl2, 25mM HEPES, 10mM Glucose, pH 7.4) after 10 min of pre-incubation with vehicle or 10µM nomifensine, and fixed immediately with PFA (4% EM Grade, Electron Microscopy Sciences) for 15 min. Upon fixation, the cells were washed twice with phosphate buffered saline (PBS) and visualized immediately with an inverted Zeiss LSM780 confocal microscope with ZEN Black software using an oil-immersion PlanApo 63x 1.4 NA objective (Zeiss, Oberkochen, Germany). 488, 543 or 633 nm laser lines supplied by Zeiss were used for excitation in combination with a 488/543 or a 488/633 beamsplitter. The green channel emission light was collected in the 494–538 nm band, the red channel emission was collected in the 547–698 nm band, and the far-red channel emission was collected in the 636–698 nm band, all by GaAsP detectors. The transmitted 488 light was also collected using a T-PMT. Pinhole was adjusted at 1 AU for 488 nm, and sequential unidirectional line scan with 8x averaging at Nyquist sampling rate was applied.

Dopaminergic Neurons.

Dopaminergic neurons were labelled with 3 (10 nM) or 4 (10 nM) for 10 min at rt in ACSF. Upon removing access ligand followed by three washes with ACSF, the samples were immediately fixed with 4% PFA for 15 min at rt. Then, the sample was washed three times in PBS, and blocked and permeabilized in 5% (w/v) goat serum (Gibco), 0.2% (w/v) saponin (Sigma) and 1% (w/v) bovine serum albumin (BSA, Sigma) in PBS for 20 min at rt. Subsequently, the samples were incubated with rat anti-DAT primary antibody (Millipore, mab369, Massachusetts, USA; 1:1000) for 30 min at rt followed by goat anti-rat secondary antibody (1:400) conjugated to Alexa647 (Invitrogen, #A-21247) or Alexa568 (Invitrogen, #A-11077) for 30 min at rt. Immediately after labelling, images were acquired with an upright Zeiss LSM710 confocal microscope with ZEN Black software using a PlanApo 20x 0.8 NA objective (Zeiss, Oberkochen, Germany). 561 or 633 nm laser lines supplied by Zeiss were used for excitation in combination with a 561/633 beamsplitter. The red channel emission was collected in the 562–630 nm band, and the far-red channel emission was collected in the 639–735 nm band, all by GaAsP detectors. Pinhole was adjusted at 1 AU for 561 nm, and sequential unidirectional line scan with 16x averaging at 170 nm/px sampling rate was applied.

TIRF Microscopy

TIRF Microscope.

All TIRF imaging was conducted on an ECLIPSE Ti-E epifluoresence/TIRF microscope NSTORM system (NIKON, Japan) with an M-CCD camera (iXon3 897, Andor, United Kingdom) that was equipped with an incubator and temperature control unit. Maximum laser power for the 488, 561 nm and 647 nm lasers were 1.1, 1.1, and 2.3 kW cm⁻², respectively. For TIRF imaging, CAD cells expressing EGFP-DAT were labelled with 10 nM of 3, 10, 4, or 14 for 10 min at rt in ACSF. Samples were immediately fixed with PFA for 15 min (3% EM Grade, Electron Microsocpy Sciences), followed by 2 washes with NH₄Cl (50 mM) and glycine (20 mM) in PBS, and imaged. The images were captured with 500 ms exposure time, 300 EM gain multiplier, and 1x conversion gain, using 1% laser power.

Single-Particle Tracking.

For single-particle tracking studies, CAD cells were labelled with 1 nM of 3 or 4 for 10 min at rt in ACSF. Upon washing the cells twice with ACSF, the cells were placed into the microscope and warmed up to 37 °C prior to imaging. Single-particle tracking videos were captured for 2 min with a framerate of 50 frames per second, a 20 ms exposure time, 300 EM gain multiplier, and 5x conversion gain. The laser power used for 3 and 4 were 15% (561) and 10% (647), respectively. In order to determine the surface diffusion of DAT, the localizations were detected and linked utilizing a modified version of the multi-particle tracking software u-track³⁶. The in-house modifications applied to this software greatly enhanced the speed by which the videos are processed without altering the tracking. These modifications have been incorporated into version 2.3 of u-track, which can be downloaded from https://github.com/DanuserLab/u-track. In short, the sub-pixel locations of sub-resolution particles were calculated through fitting a 2D Gaussian with the standard deviation of the point spread function of the microscope around local intensity maxima. For constructing intact trajectories, the algorithm first links the localizations between consecutive frames, followed by linking of the track segments generated to close gaps. Once the tracks were derived, the tracks were segmented into different diffusion classes (immobile, confined diffusion, Brownian diffusion, super diffusion) and characterized using the transient motion analysis software DC-MSS³⁷.

dSTORM.

CAD cells expressing EGFP-DAT were labelled with 3 (10 nM) or 4 (10 nM) for 10 min at rt in ACSF. Samples were immediately fixed with PFA for 15 min (3% EM Grade, Electron Microsocpy Sciences), followed by 2 washes with NH₄Cl (50 mM) and Glycine (20 mM) in PBS. Imaging took place directly after fixation per sample, utilizing an imaging buffer containing: 10% (w/v) glucose, 1% (v/v) beta-mercaptoethanol, 50 mM Tris-HCl (pH 8), 10 mM NaCl, 34 μg ml⁻¹ catalase, 28 μg ml⁻¹ glucose oxidase, 2 mM cyclooctatetraene. dSTORM images were captured with 30000 frames, using a 16 ms exposure time 300 EM gain multiplier, and 5x conversion gain, using 100% laser power. 405 nm laser was gradually increased to <0.1 kW cm⁻² to further induce blinking. Localizations were fit to the data through ThunderStorm using the Local Maximum detector, a threshold of 2*std(Wave.F1). Drift was corrected through redundant cross correlation in matlab (https://www.osapublishing.org/oe/abstract.cfm?uri=oe-22–13-15982). Localizations within 15 nm and 3 frames were merged together. Localizations were filtered for an uncertainty less than 25 nm. Clustering was performed with DBSCAN from the sklearn python library. DBSCAN isolated localization that contained 30 other localizations within a 35 nm radius.