Functional characterization of N-octyl 4-methylamphetamine variants and related bivalent compounds at the dopamine and serotonin transporters using Ca²⁺ channels as sensors

Abstract

The early characterization of ligands at the dopamine and serotonin transporters, DAT and SERT, respectively, is important for drug discovery, forensic sciences, and drug abuse research. 4-Methyl amphetamine (4-MA) is a good example of an abused drug whose overdose can be fatal. It is a potent substrate at DAT and SERT where its simplest secondary amine (N-methyl 4-MA) retains substrate activity at them. In contrast, N-n-butyl 4-MA is very weak, therefore it was categorized as inactive at these transporters. Here, N-octyl 4-MA and other related compounds were synthesized, and their activities were evaluated at DAT and SERT. To expedite this endeavor, cells expressing DAT or SERT were co-transfected with a voltage-gated Ca²⁺ channel and, the genetically-encoded Ca²⁺ sensor, GCaMP6s. Control compounds and the newly synthesized molecules were tested on these cells using an automated multi-well fluorescence plate reader; substrates and inhibitors were identified successfully at DAT and SERT. N-Octyl 4-MA and three bivalent compounds were inhibitors at these transporters. These findings were validated by measuring Ca²⁺-mobilization using quantitative fluorescence microscopy. The bivalent molecules were the most potent of the series and were further characterized in an uptake-inhibition assay. Compared to cocaine, they showed comparable potency inhibiting uptake at DAT and higher potency at SERT. These observations support a previous hypothesis that amphetamine-related (and, here, N-extended alkyl and) bivalent arylalkylamine molecules are active at monoamine transporters, showing potent activity as reuptake inhibitors, and implicate the involvement of a distant auxiliary binding feature to account for their actions at DAT and SERT.

Graphical Abstract

1. Introduction

Membrane proteins working as transporters of dopamine (DA) and serotonin (5-HT) have a prominent role controlling the homeostasis of these neurotransmitters in the brain (Volkow et al., 1997; Fleckenstein et al., 2007; Kristensen et al., 2011). Ligands of the DA transporter (DAT) and 5-HT transporter (SERT) are used clinically to potentiate monoaminergic neurotransmission toward the alleviation of psychiatric disorders. Typically, drugs selective toward SERT (e.g., selective serotonin reuptake inhibitors) are commonly prescribed as antidepressants, whereas drugs potent at DAT are psychostimulants used in the treatment of attention deficit hyperactivity disorder (e.g., amphetamine and methylphenidate) (Feighner, 1999; Sharma and Couture, 2014; Marazziti et al., 2019). Moreover, some psychostimulants are common illegal street drugs (e.g., methamphetamine and cocaine) (UNODC, 2019).

4-Methylamphetamine (4-MA) is an example of a psychostimulant sold mixed with caffeine and amphetamine in European clandestine markets. The consumption of this mixture was related to severe intoxications and fatalities (Blanckaert et al., 2013). 4-MA in experimental animals produces a dopamine-driven psychostimulant effect together with strong serotonergic action that, combined with its monoamine oxidase inhibitory activity, explain its toxicity (Wee et al., 2005; Blanckaert et al., 2013). In pharmacological studies, 4-MA is a potent and non-selective substrate at all three monoamine transporters (DAT, SERT, and, the norepinephrine transporter, NET) (Wee et al., 2005). Substrates at these transporters are taken up by neurons, interfering with the normal reuptake of neurotransmitters. In general, substrates of DAT or SERT (such as amphetamine and certain cathinones), once internalized, disrupt neurotransmitter storage, reverse the transport mechanism, and promote a massive release of these neurotransmitters (Rothman and Baumann, 2006; Fleckenstein et al., 2007; Kristensen et al., 2011). From an electrophysiological perspective, monoamine transporters are Na⁺/Cl^--dependent symporters that operate concomitant with an inward electrical current (Mager et al., 1994; Galli et al., 1995; Sonders et al., 1997). Thus, substrate-transporter interaction can further regulate electrical and ionic homeostasis, having functional consequences in neurons (Ingram et al., 2002; De Felice, 2017; Condon et al., 2019).

Previous studies characterized the action of 4-MA analogs at monoamine transporters using Ca²⁺ signals as a proxy of transporter activity in cells expressing voltage-gated Ca²⁺ channels (Ca_V) (Ruchala et al., 2014; Cameron et al., 2015; Solis et al., 2017; Battisti et al., 2018). The stepwise elongation of the N-alkyl chain of 4-MA from N-methyl to N-n-butyl resulted in a gradual loss of potency and a transition from substrate to reuptake inhibitor activities at the monoamine transporters, where N-n-butyl 4-MA was virtually inactive at these transporters (Solis et al., 2017). These results were validated using a classical method (i.e., rat brain synaptosomes) to evaluate the action of these N-alkyl analogs of 4-MA at DAT, SERT and NET (Solis et al., 2017).

Although the N-n-butyl analog of 4-MA was essentially inactive, prior studies have shown that related agents with yet longer N-alkyl chains are not inactive. The conversion of DA or 5-HT to a secondary amine by inclusion of a long, simple N-alkyl chain (up to C₁₄) resulted in an increased binding affinity to DAT and SERT (Andersen et al., 2016). Then, the production of bivalent compounds composed of two DA and/or 5-HT moieties connected through their amine groups by a aliphatic chain showed better binding affinity than the single DA or 5-HT molecules to DAT and SERT (Andersen et al., 2016). For example, an eight-carbon linker between two DA molecules increased binding affinity ~80-fold compared to the affinity of a single DA molecule at DAT (Schmitt et al., 2010). Similarly, bivalent compounds based on amphetamine and phenylethylamine connected by a simple aliphatic chain increased the binding affinity to DAT (Schmitt et al., 2010). Inspired by these results, implicating a distant auxiliary binding site, several extended agents (i.e., N-alkyl analogs of 4-MA) (Solis et al., 2017; Battisti et al., 2018) and seven not hitherto investigated analogs were synthesized and examined in functional assays. The N-n-octyl chain substitution on the 4-MA scaffold recovered potency hindering DA- and 5-HT-induced signals at DAT and SERT, respectively. Furthermore, the expansion to several related “bivalent” ligands connected by an eight-carbon aliphatic chain further improved potency at DAT and SERT inhibiting Ca²⁺ signals induced by endogenous substrates. The inhibitory effect of these bivalent compounds at DAT and SERT was validated using an uptake-inhibition assay and their potency was compared to that of cocaine. The results presented here support the hypothesis that further extension of the alkyl group of N-n-butyl 4-MA reverses the downward trend in potency as the N-alkyl chain of 4-MA is elongated, and that 4-MA-related bivalent compounds have increased potency at monoamine transporters to inhibit reuptake.

2. Results and discussion

2.1. Synthesis.

Eleven compounds were evaluated at DAT and SERT. Four of these molecules (N-methyl 4-MA, N-ethyl 4-MA, N-propyl 4-MA and N-butyl 4-MA; 1-4, respectively) were previously reported by us (Solis et al., 2017) and seven (5-11) are novel (Fig. 1).

Fig. 1. Structures of simple N-alkyl 4-MA analogs and “bivalent” 4-MA ligands examined in the present study.

*Compounds 1-4 were previously synthetized (Solis et al., 2017) and compounds 5-11 were synthetized as described in methods and results.

The synthesis of compound 1 was previously reported as its HBr salt, and compounds 2-4 as their HCl salts; (Solis et al., 2017) they were used as such for this study. Compound 5 was obtained by reductive amination of 1-(4-methylphenyl)-2-propanone (12)(Jacob et al., 1995) and 1-amino-n-octane; compound 6 was prepared in like manner from its aldehydic counterpart (4-methyl)phenylacetaldehyde (13)(Chikashita et al., 1987) (Scheme I). Reductive amination of 12 and 13 with N¹-Boc-protected 1,8-diaminooctane 14, followed by deprotection, afforded 7 and 8, respectively (Scheme I).

Scheme I.

Reagents and conditions: (a) i) H₂N(CH₂)₇CH₃, ii) HCl, Et₂O; (b) i) 14, PtO₂, H₂, EtOH, ii) HCl, Et₂O; (c) i) TFA, CH₂Cl₂, ii) HCl, Et₂O.

In the absence of a Boc protecting group, reductive amination of 13 with 1,8-diaminooctane yielded compound 9 (Scheme II). Further reductive alkylation of 9 using formaldehyde afforded the N¹, N⁸-dimethylated compound 10. Alkylation of 16 using formalin and sodium borohydride in acetic acid gave intermediate 17 that was deprotected and reductively alkylated with 13 to provide 11 (Scheme II).

Scheme II.

Reagents and conditions: (a) i) H₂N(CH₂)₈NH₂, PtO₂, H₂, EtOH ii) HCl, Et₂O; (b) i) HCOH, H₂, Pd/C, MeOH; ii) (COOH)₂, Et₂O; (c) 14, PtO₂, H₂, EtOH; (d) HCOH, NaBH₄, HOAc; (e) i) TFA, CH₂Cl₂, ii) 13, PtO₂, H₂, EtOH, iii) (COOH)₂, Et₂O.

2.2. N-Octyl modified 4-MAs and related bivalent compounds are active at DAT and SERT.

As mentioned above, monoamine transporters are electrically active; substrate-transporter interaction promotes an inward electrical current concomitant with the transport mechanism (Mager et al., 1994; Galli et al., 1995; Sonders et al., 1997). Such inward current depolarizes the plasma membrane reaching the threshold for Ca_V channel activation (Ruchala et al., 2014; Cameron et al., 2015). Thus, cells co-expressing an L-type Ca²⁺ channel and a monoamine transporter constitute a robust biosensor to study the pharmacological properties of monoamine transporter ligands using intracellular Ca²⁺ signals as a readout. This technique was used extensively to study the activity of amphetamine analogs and cathinone analogs at monoamine transporters (Solis et al., 2017; Battisti et al., 2018; Steele and Eltit, 2019; Davies et al., 2020). In this experimental setting, monoamine transporters’ substrates promote intracellular Ca²⁺ signals measured using permeable Ca²⁺-sensitive dyes, and inhibitors of these transporters hinder such Ca²⁺ signals (Solis et al., 2017; Battisti et al., 2018; Moerke et al., 2018; Yadav-Samudrala et al., 2019).

An initial screening to test the substrate or inhibitor activities of the eleven compounds described in Fig. 1 was performed at a single concentration. In these experiments, SERT- and DAT-expressing cells were co-transfected with the Ca²⁺ channel and GCaMP6s; the latter is a commonly used and very bright genetically-encoded Ca²⁺ sensor (Chen et al., 2013). Ca²⁺ signals were measured using an automated plate reader (FlexStation 3) in a 96-well plate format. A few wells were used to test the effect of DA (20 μM) at DAT or N-methyl 4-MA (20 μM) at SERT; Ca²⁺ signals induced by these substrates were considered positive controls that define the 100% response for each run (see Methods). N-Methyl 4-MA (1) was selected as the positive control at SERT because, in contrast to 5-HT, it produced a more robust response at 20 μM when tested in the plate reader assay. 5-HT and N-methyl-4MA showed the same efficacy and comparable potency (EC₅₀: 0.9 μM vs. 0.5 μM, respectively), (Ruchala et al., 2014; Solis et al., 2017) when evaluated in the fluorescent microscope using fast constant perfusion. The decrease in apparent potency of 5-HT in the FlexStation can be explained by the slow diffusion of the activating agent when injected into the well. For the initial screening using the FlexStation test compounds at 10 μM concentration were applied alone to assess substrate activity (open bars in Fig. 2) or in combination with the aforementioned DAT or SERT substrates (20 μM) to test inhibitory activity (black bars in Fig. 2).

Fig. 2. Evaluation of eleven compounds at SERT and DAT using a voltage-gated Ca²⁺ channel and GCaMP6s in an automated fluorescence plate reader.

Cells expressing DAT were co-transfected with Ca_V1.2 subunit of the Ca²⁺ channel and GCaMP6s coding plasmids. SERT-expressing cells were co-transfected with Ca_V1.2 and β₃ subunits of the Ca²⁺ channel and GCaMP6s coding plasmids. Ca²⁺ signals were measured using a FlexStation 3 fluorimeter, and the data is expressed relative to the Ca²⁺ signal amplitude induced by the application of positive control wells within the same run (N-methyl 4-MA for SERT and DA for DAT); the amplitude of the positive control wells are depicted by the 100% marks. The relative responses of tested agents applied alone or co-applied with the positive control are shown in white and dark bars, respectively (A and B). The final concentration of the tested compounds and positive control used in this assay were 10 μM and 20 μM, respectively. In the left panels the tested compounds were not preincubated. The apparently inactive compounds were tested again but including a preincubation period before the application of the positive control (dark bars in C and D). Mean ± SEM of n ≥ 4 runs are shown.

Compounds 1, 2 and 3, when applied alone, produced Ca²⁺ signals at SERT (Fig. 2A) whereas from this group only compound 1 showed Ca²⁺ signals at DAT (Fig. 2B). Compound 1 is a substrate at SERT and DAT, whereas 2 and 3 are substrates only at SERT (Fig. 2A). This pattern of effects was previously reported for 1, 2, and 3 (Solis et al., 2017). All other compound/transporter combinations tested showed no Ca²⁺ signal when applied alone, suggesting that they are not substrates at DAT or SERT (Fig. 2).

Inhibitory activity of these compounds can be determined when they are co-applied with substrates (black bars in Fig. 2A and 2B). At SERT, compounds 4 and 7 co-applied with the positive control (N-methyl 4-MA) were inactive at the tested concentration (Fig. 2A). In the same transporter, compound 8 co-applied with the positive control showed a small inhibitory effect (Fig. 2A). Compounds 5, 6, 9, 10 and 11 inhibited the N-methyl 4-MA signal (Fig. 2A). At DAT, compounds 2, 3, 4, 7 and 8 co-applied with the positive control (DA) showed Ca²⁺ signals of similar amplitude as DA alone, suggesting that these compounds are inactive as inhibitors at the tested concentration (Fig. 2B). In contrast, when 5 and 6 were co-applied with DA the signal was blocked ~50%, suggesting that they are weak inhibitors at DAT (Fig. 2B). Compounds 9, 10 and 11 clearly blocked most of the DA signal and the remaining signals were not different from compounds applied alone (Fig. 2B).

A previous report characterized 2 and 3 as inhibitors at DAT (Solis et al., 2017), but they were inactive in the initial screening presented above. To address this inconsistency, 2 and 3 and all inactive compounds in the initial screening were assayed again but including a preincubation period before the application of the substrate in the plate reader assay (black bars in Fig. 2C and 2D). In this set of experiments compounds 2 and 3 showed blockade at DAT (~75% and ~50% inhibition, respectively, Fig. 2D). On the other hand, compound 4 remained inactive and compounds 7 and 8 displayed a modest blocking effect (~30% inhibition), suggesting that they are weak inhibitors at DAT (Fig. 2D). The preincubation of compound 4 did not affect the N-methyl 4-MA-induced Ca²⁺ signal, suggesting that, in agreement with our previous report (Solis et al., 2017), it is inactive at SERT (Fig. 2C). In contrast, preincubation of compounds 7 and 8 fully blocked the N-methyl 4-MA-induced Ca²⁺ signal, showing that these compounds are inhibitors at SERT (Fig. 2C). Thus, this is the first time that a Ca²⁺-based method has been implemented to screen and discriminate between substrates and blockers, simultaneously, in an automated fluorescent plate reader. This small-scale screening clearly depicts that the new compounds synthesized here (5-11) are active as blockers at DAT and SERT and 9-11 are probably the most potent. Since this technique involves a transfection step, it is not ideal for high-throughput screenings, but it can be practical as a secondary screening tool, or in low- to mid- size screenings of new compounds at these transporters. The creation of a cell line permanently expressing the Ca²⁺ channel, the genetically-encoded Ca²⁺ sensor, and the transporter would be the next optimization step that would provide higher throughput screening capabilities.

To validate the findings observed using the plate reader approach, full quantitative dose-response experiments were performed in a fluorescence microscope equipped with constant perfusion, electronic solution switching, and temperature control. All these fluorescence microscopy experiments were performed using Ca_V1.2, β₃ and α₂δ subunits and Fura-2 as Ca²⁺ sensitive dye as described in (Solis et al., 2017). Compounds 5–11 were tested here, and compounds 1–4 were previously reported using the same technique (Solis et al., 2017); the latter results are included in Table 1 for comparison.

Table 1.

Potencies of compounds 1-11, with cocaine as control, to produce Ca²⁺ signals (EC₅₀ bold values, in μM ± SEM) or to block Ca²⁺ signals induced by DA or 5-HT (IC₅₀, in μM±SEM) using fluorescence microscopy. 95% Confidence Intervals (95% CI) are in μM

Compound	DAT	95%CI	SERT	95%CI
1*	0.2±0.02	0.17 to 0.25	0.5±0.1	0.31 to 0.73
2*	4.4±0.6	3.15 to 5.55	0.7±0.9	0.53 to 0.88
3*	18.3±3.6	11.20 to 25.32	2.4±0.4	1.79 to 3.34
4*	61.0±12.6	36.19 to 85.72	Inactive	-
5	3.9±0.2	3.53 to 4.26	3.2±0.8	1.66 to 4.71
6	4.7±0.2	4.34 to 5.00	2.8±0.2	2.37 to 3.17
7	8.3±0.5	7.26 to 9.30	1.7±0.4	0.87 to 2.48
8	12.0±1.7	8.53 to 15.29	4.6±1.4	1.95 to 7.24
9	1.5±0.1	1.34 to 1.64	0.46±0.02	0.42 to 0.49
l0	1.6±0.1	1.48 to 1.72	0.42±0.03	0.36 to 0.48
ll	1.4±0.1	1.27 to 1.53	0.21±0.02	0.18 to 0.24
Cocaine	2.2±0.2	1.86 to 2.56	2.5±0.3	1.87 to 3.06

IC₅₀ values were calculated by competing positive control at 5 μM concentration.

^*Data previously reported (Solis et al., 2017) included for the purpose of comparison.

As mentioned, 1 is a substrate at all three monoamine transporters (Solis et al., 2017). Elongation of the N-alkyl chain in compounds 2, 3 and 4 produced inhibitors with a gradual decrease in potency at DAT, in which N-butyl 4-MA (4) was the weakest (Solis et al., 2017). Further elongation of the N-alkyl chain to eight carbon atoms (i.e., 5) substantially rescued the inhibitor effect of the molecule up to the level of compound 2 at DAT (Table 1). At SERT, 1, 2 and 3 are substrates (Solis et al., 2017), and elongation of the N-alkyl chain to 4 decreased potency to produce Ca²⁺ signals (Solis et al., 2017). Compound 4 was very weak at SERT, and consequently was classified as inactive (Solis et al., 2017). Here, further elongation of the N-alkyl chain of the 4-MA scaffold to eight carbon atoms (i.e., 5) rescued activity at SERT, but emerged as an inhibitor at this transporter (Table 1). Previously it was reported that long N-alkyl substitutions on DA or 5-HT scaffolds improve binding affinity at DAT and SERT, presumably by increasing hydrophobic interactions of the N-alkyl chain extending toward the extracellular vestibule from the central S1 substrate binding site of the transporter (Andersen et al., 2016). Here similar results were observed indicating that N-alkyl extensions (C₈) in the 4-MA scaffold increase potency as compared to the C₄ substitution, but with the additional information that it works as a reuptake inhibitor at both DAT and SERT.

The α-desmethylation of compound 5 (to produce compound 6) retained activity as an inhibitor and did not affect its potency to inhibit Ca²⁺ signals in DAT- or SERT- expressing cells (Table 1). Compound 7 compared to 5 revealed slightly decreased inhibitor potency at DAT with no clear change in potency at SERT (Table 1). The α-desmethylation of compound 7 (that is, compound 8) worsened the activity at both transporters (Table 1). The addition of a second 4-methyl α-desmethyl amphetamine (or 4-methyl β-phenylethylamine) moiety to the molecule, generating the bivalent compound 9, increased potency as an inhibitor at DAT and SERT by ~1 log unit (Table 1). The α-methyl substitution of each 4-MA moiety in the bivalent compounds was not included to avoid the generation of two chiral centers that would complicate data interpretation. N-Methylation of both amines of 9 (to generate tertiary amines, compound 10) had no further effect as inhibitors of DAT and SERT (Table 1, Fig. 3A and 3B). When only one amine was methylated (compound 11), the potency to inhibit Ca²⁺ signals was slightly increased at SERT and the effect at DAT was unchanged (Table 1, Fig. 3A, 3B, 3D and 3E). The potency of a relevant non-selective inhibitor at monoamine transporters, cocaine, was tested under the same experimental conditions as a reference. Cocaine showed a potency of 2.2±0.2 (95%CI: 1.9 to 2.6) μM inhibiting DA-induced signals at DAT and 2.5±0.3 (95%CI: 1.9 to 3.1) μM to inhibit 5-HT-induced signals at SERT (Table 1, Fig. 3A, 3B, 3G and 3H). Compounds 9, 10 and 11 showed similar potency than cocaine at DAT, but were more potent than this psychostimulant at SERT, (Table 1, 3A and 3B).

Fig. 3. Potency of compound 9, 10 and 11 blocking Ca²⁺ signals at SERT- and DAT-expressing cells and comparison to cocaine.

Cells expressing SERT, DAT or parental cells expressing no transporter (Flp-In) were transfected with Ca_V1.2 plus β₃ and α₂δ subunits, and Ca²⁺ signals produced by 5-HT, DA, or high-K⁺ solution were determined using Fura-2AM and fluorescence microscopy. (A) Dose-response curves of compounds 9, 10, 11 and cocaine on Ca²⁺ signals induced by 5-HT in cells expressing SERT, (B) induced by DA in cells expressing DAT, or (C) induced by high-K+ solution in Flp-In cells. (D, E, and F) show the effect of compound 11 on Ca²⁺ signals induced by 5-HT, DA, or high K+ solution in SERT cells, DAT cells, and Flp-In cells, respectively. (G, H and I) show the inhibitory effect of cocaine on Ca²⁺ signals induced by 5-HT, DA, and high-K⁺ solution in SERT cells, DAT cells, and Flp-In cells, respectively. Data correspond to mean ± SEM of n ≥ 68 cells per concentration. List of potencies for the full series of compound tested is shown in Table 1.

One disadvantage of any indirect technique to assess potential new drugs is the possibility of having false-positive results. For instance, in this assay, molecules that directly modulate L-type Ca⁺² channels could show results that can be misinterpreted as active at transporters. In the literature very few activators of L-type Ca²⁺ channels have been reported (e.g., (S)-(−)-Bay K 8644 and FPL 64176) (Schramm et al., 1983; Franckowiak et al., 1985; Zheng et al., 1991), and whereas they can produce an increase in total current and a left-shift in their current-voltage relationship (Brown et al., 1984; Zheng et al., 1991), it is not evident whether they would depolarize the cell membrane at rest to activate the channels without the need of another source of depolarization. Thus, although possible, it is unlikely that L-type Ca²⁺ channel activators would be a significant problem for drug screening using this assay. On the other hand, inhibitors of L-type Ca²⁺ channels have been described and are commonly used antihypertensive drugs (Moser, 1990). Compounds with similar structure to known Ca²⁺ channel inhibitors must be additionally evaluated in a transporter uptake assay to confirm their direct effect at transporters. Regardless, a simple control experiment to unveil direct effects at the Ca²⁺ channel is to perform Ca²⁺ signal determinations on cells expressing only Ca²⁺ channels (no transporters). In this experimental setting, if compounds are applied alone, they should not produce a Ca²⁺ signal, and if applied in combination with high-K⁺ solution (that directly opens the Ca²⁺ channel by membrane depolarization), they should not inhibit the Ca²⁺ signal.

The most potent compounds of this series (9, 10, and 11) were assayed in parental cells (Flp-In cells) expressing the Ca²⁺ channel (but not monoamine transporters), and the Ca²⁺ channels were opened using a high-K⁺ external solution (Fig. 3C, 3F, and 3I). These compounds did not produce Ca²⁺ signals in cells, except at 30 μM where compounds 9 and 11 produced Ca²⁺ signals showing a slow onset (shown for compound 11 in Fig. 3F). Compound 10 showed a very small Ca²⁺ response on its own at 30 μM concentration (not shown). This unspecific effect at 30 μM was also seen in cells not expressing Ca²⁺ channels, suggesting that Ca²⁺ channels are not mediating such effect. Thus, these compounds must be affecting other proteins involved in intracellular Ca²⁺ homeostasis when used at high concentrations. Dose-response experiments showed that cocaine does not affect the operation of the Ca²⁺ channel up to 10 μM (Fig. 3C and 3I). The three bivalent compounds (9, 10 and 11) inhibited the K⁺-induced Ca²⁺ signals with IC₅₀ of: 12.6 ± 0.4 (95%CI: 11.6 to 13.7) μM, 10.6 ± 0.4 (95%CI: 10.0 to 11.5) μM, and 11.3 ± 0.5 (95%CI: 10.2 to 12.3) μM, respectively (Fig. 3C).

Cocaine showed no direct effect on the Ca²⁺ channel, indicating that the IC₅₀s obtained for its effect at DAT and SERT represent the direct inhibitory effect of cocaine at the transporters when evaluated using the Ca²⁺ assay. Similarly, the bivalent compounds were more than 20 times weaker at inhibiting the Ca²⁺ channels, indicating that the IC₅₀s obtained in the Ca²⁺ assay represent the direct inhibition at SERT. At DAT, the potency obtained in the Ca²⁺ assay for the bivalent compounds was only ~6-fold (less than a log unit) more potent than their direct effect at the Ca²⁺ channel (Fig. 3), indicating that the potency obtained is mainly driven by the effect at DAT plus a small overlapping contribution of the direct effect at the Ca²⁺ channel. This overlapping effect shifts the apparent inhibitory effect at DAT, appearing more potent. Thus, the IC₅₀ at DAT should be slightly weaker than what was obtained experimentally using the Ca²⁺ assay.

APP⁺ is a fluorescent substrate of DAT and SERT used to monitor the activity of these monoamine transporters in live cell imaging (Solis et al., 2012; Karpowicz et al., 2013). The inhibition of APP⁺ uptake was used to study the direct effect of 9, 10 and 11 at DAT and SERT. These compounds inhibited APP⁺ transport up to the level of non-specific entry at high concentrations. At SERT, compounds 9, 10 and 11 showed IC₅₀: 0.17 ± 0.03 (95%CI: 0.11 to 0.22) μM, 0.27 ± 0.05 (95%CI: 0.17 to 0.38) μM, 0.17 ± 0.03 (95%CI: 0.10 to 0.23) μM, respectively, and were more potent than cocaine inhibiting APP⁺ uptake that showed an IC₅₀: 1.50 ± 0.33 (95%CI: 0.82 to 2.17) μM (Fig. 4A, 4C, and 4E). Compounds 9, 10 and 11 were equipotent inhibiting intracellular APP⁺ accumulation, with IC₅₀ of 0.91 ± 0.11 (95%CI: 0.81 to 1.14) μM, 0.87 ± 0.09 (95%CI: 0.70 to 1.05) μM, 0.85 ± 0.08 (95%CI: 0.68 to 1.02) μM, respectively, in cells expressing DAT (Fig. 4B and 4F). In addition, cocaine with an IC₅₀ of 0.34 ± 0.05 (95%CI: 0.23 to 0.42) μM, was slightly more potent blocking APP⁺ uptake through DAT than these compounds (Fig. 4B and 4D). Illustrative experiments depicting the inhibition of APP⁺ uptake by cocaine and compound 11 at both transporters are shown in Fig. 4. In the APP⁺ assay the bivalent compounds were half a log unit less potent than cocaine at DAT; whereas in the Ca²⁺ assay these compounds were equipotent with cocaine. As discussed above, this discrepancy can be explained by the contribution of the direct inhibitory effect of the bivalent compounds at the Ca²⁺ channel. Thus, the results obtained by APP⁺ are more accurate than the Ca²⁺ assay for DAT, indicating that these compounds are slightly less potent than cocaine at this transporter.

Fig. 4. Activity of compound 9, 10 and 11 at DAT and SERT using an uptake inhibition assay.

The uptake of APP⁺, a fluorescent substrate of monoamine transporters, was used to trace the activity of DAT and SERT in live cells using fluorescence microscopy. The effect of several concentrations of compounds 9, 10, 11 and cocaine inhibiting SERT (A) or DAT (B). Experimental trances showing the uptake of APP⁺ in cells expressing SERT and DAT and the inhibitory action of cocaine (C and D) and the inhibitory action of compound 11 (E and F). The vertical graduation mark represents 20% of the maximal APP⁺ uptake in absence of inhibitor. The data shown are mean (black trace) + SEM (grey marks) of n ≥ 6 wells per concentration.

Altogether, these results are consistent with previous findings testing bivalent compounds derived from common substrates, showing high binding affinity at DAT and SERT (Schmitt et al., 2010; Andersen et al., 2016). The alkyl-linker length used in compound 9, 10 and 11 is not long enough to support interactions of the bivalent compounds between adjacent transporter macromolecules, suggesting that each bivalent molecule tested in this study must interact within a single transporter protomer. Previously, docking studies followed by serial mutagenesis showed that residues Ala173, Asn177 and Thr439 of the S1 site and residues Ile179 and Lys490 present in the S2 sites are involved in the binding of each 5-HT moiety of the bivalent compound 5-HT-PEG₄-5-HT at SERT (Andersen et al., 2016). Similarly, flexible docking studies of DA-(CH₂)₈-DA at DAT showed that one DA moiety occupies the S1 binding site involving residues D79, Tyr156 and Ser149, and the second DA moiety binds to the S2 site including residues Lys92 and Asp313 (Schmitt et al., 2010). These studies strongly suggest that one side of these high affinity bivalent compounds interacts with the canonical S1 site in the central cavity of the transporter, then the linker projects toward the extracellular vestibule in which the other moiety interacts with the S2 site. These extensive interactions between the bivalent compounds and the monoamine transporters increase affinity, keeping the transporter in an inhibited conformation and preventing transport to occur.

Structural work using X-ray crystallography has shown that traditional DAT ligands such as dopamine, amphetamine and cocaine (co-crystallized with the drosophila DAT) bind to the S1 site located in the center of the transporter (Wang et al., 2015). Biochemically, cocaine is an inhibitor that stabilizes the outward-facing conformation of the transporter (Ferrer and Javitch, 1998); on the other hand, substrates transit through transporters as dynamic conformational transitions occur along the transport cycle (Coleman et al., 2019). The N-alkyl extended 4MA and bivalent molecules described here should work in a similar manner to cocaine, stabilizing an outward-facing conformation of the transporter. Interestingly, an allosteric modulator of DAT decreases cocaine affinity but keeps its activity transporting DA (Aggarwal et al., 2019). Thus, potentially, certain compounds binding at allosteric sites (other than the orthosteric S1 site) would antagonize the action of cocaine, working as a cocaine “antidote”. The bivalent compounds described previously (Schmitt et al., 2010; Andersen et al., 2016) and here can be used as a starting point to explore new compounds favoring binding to the S2 site over the S1 site, that could have a therapeutic use against misuse of certain psychostimulants.

Inhibitors of monoamine transporters have a potential of abuse mainly by disrupting DA reuptake in the central nervous system (Volkow et al., 1999; Negus and Miller, 2014). A measure used to predict abuse liability from in vitro data is the DAT/SERT potency ratio. Assuming adequate brain penetration, both non-selective inhibitors at these transporters (e.g., cocaine) or selective at DAT (e.g., MDPV) are self-administered, or potentiate intracranial self-stimulation (ICSS); these effects indicate abuse liability in experimental animals, assessments that can translate to humans (Bonano et al., 2014; Negus and Miller, 2014; Watterson et al., 2014; Gannon et al., 2018; Collins et al., 2019; Baird et al., 2021). Conversely, SERT-selective inhibitors are not self-administered, depress ICSS, and are not abused (Howell and Negus, 2014; Negus and Miller, 2014). The N-alkyl 4-MA and bivalent molecules tested here, although showing some preference for SERT, are all rather non-selective inhibitors, suggesting that these compounds might potentially have abuse liability by potentiating DA neurotransmission; however preclinical evaluation of these compounds must be performed to assert this interpretation empirically.

In conclusion, a Ca²⁺ channel in conjunction with a genetically encoded- Ca²⁺ sensor (GCaMP6s) were used to identify ligands of monoamine transporters exploiting an automated fluorescence plate reader using a 96-well format. This approach was fast and effective at discerning between substrates and inhibitors of DAT and SERT at a single concentration. Then, a quantitative analysis of the newly synthesized compounds showed that the extension of a N-alkyl chain from C₄ to C₈ at the 4-MA scaffold recovers potency as reuptake inhibitor at DAT and SERT. Three related bivalent compounds were inhibitors at these transporters showing potencies comparable to cocaine at DAT and SERT. These observations support a previous hypothesis that the simultaneous binding of molecules to the S1 and S2 sites increases potency as reuptake inhibitor at monoamine transporters.

3. Methods

3.1. Chemistry.

All commercially available reagents and solvents were purchased from Sigma-Aldrich (St. Louis, MO) and used as delivered. For new compounds, melting points were measured in glass capillary tubes (Thomas-Hoover melting point apparatus) and are uncorrected. ¹H NMR spectra were recorded with a Bruker 400 MHz spectrometer; chemical shifts (δ) are reported in parts per million (ppm) relative to tetramethylsilane as internal standard. Reactions and product mixtures were routinely monitored by thin-layer chromatography (TLC) on silica gel pre-coated F254 Merck plates. The purity of the tested compounds was established by elemental analysis (Atlantic Microlabs; Norcross; GA); determined values were within 0.4% of theory (for atoms indicated).

N-(1-(4-Methylphenyl)prop-2-yl)octan-1-amine hydrochloride (5).

A mixture of 1-(4-methylphenyl)-2-propanone (12)(Jacob et al., 1995) (0.5 g, 3.37 mmol), 1-amino-n-octane (0.52 g, 4.04 mmol), PtO₂ (0.005 g, 0.02 mmol) and EtOH (20 mL) was shaken in a Parr hydrogenator under H₂ (ca. 45 psi) at room temperature for 16 h. Catalyst was removed by filtration through a Celite pad and solvent was removed under reduced pressure. The oily residue was converted to its HCl salt using a saturated solution of HCl(g)/Et₂O which, upon recrystallization from EtOH/Et₂O, yielded 0.65 g (60%) of 5 as a white solid: mp: 223–224 °C; ¹H NMR (DMSO-d₆) δ 0.88 (t, J = 6.9 Hz, 3H, CH₃), 1.19 (d, J = 6.5 Hz, 3H, CH₃), 1.18–1.39 (m, 10H, 5 x CH₂), 1.56–1.75 (m, 2H, CH₂), 2.29 (s, 3H, CH₃), 2.60 (dd, J = 10.6, 13.0 Hz, 1H, CH), 2.82–2.94 (m, 2H, CH₂), 3.23 (dd, J = 3.4, 13.0 Hz, 1H, CH), 3.31–3.41 (m, 1H, CH), 7.10–7.19 (m, 4H, ArH), 9.03 (br s, 2H, NH₂, D₂O ex). Anal. Calcd for (C₁₈H₃₁N•HCl) C, 72.57; H, 10.83, N, 4.70. Found: C, 72.59; H, 10.86; N, 4.69.

N-(4-Methylphenylethyl)octan-1-amine hydrochloride (6).

A mixture of (4-methyl)phenylacetaldehyde (13)(Chikashita et al., 1987) (0.13 g, 0.97 mmol), 1-amino-n-octane (0.15 g, 1.16 mmol), PtO₂ (0.001 g) and EtOH (10 mL) was shaken in a Parr hydrogenator under an H₂ atmosphere (ca. 45 psi) at room temperature for 16 h. The catalyst was removed by filtration through a Celite pad and the solvent was removed under reduced pressure. The oily residue was converted to its HCl salt using a saturated solution of HCl (g)/Et₂O which upon recrystallization from EtOH/Et₂O yielded 0.07 g (26%) of 6 as a white solid: mp: 268–269 °C; ¹H NMR (DMSO-d₆) δ 0.88 (t, J = 7.0 Hz, 3H, CH₃), 1.19–1.38 (m, 10H, 5 x CH₂), 1.56–1.69 (m, 2H, CH₂), 2.29 (s, 3H, CH₃), 2.83–2.96 (m, 4H, 2 x CH₂), 3.03–3.12 (m, 2H, CH₂), 7.10–7.19 (m, 4H, ArH), 8.91 (br s, 2H, NH₂, D₂O ex). Anal. Calcd for (C₁₇H₂₉N•HCl•0.2 H₂O) C, 71.02; H, 10.66; N, 4.87. Found: C, 70.94; H, 10.57; N, 4.80.

N¹-(1-(4-Methylphenyl)prop-2-yl)-1,8-diaminooctane dihydrochloride (7).

A mixture of 13 (Chikashita et al., 1987) (0.3 g, 2.02 mmol), N¹-Boc −1,8-diaminooctane (14)(Huang et al., 2014) (0.49 g, 2.02 mmol), PtO₂ (0.003 g) and EtOH (20 mL) was shaken in a Parr hydrogenator under H₂ (ca. 45 psi) at room temperature for 16 h. The catalyst was removed by filtration through a Celite pad and solvent was removed under reduced pressure. The oily residue was converted to its HCl salt using a saturated solution of HCl(g)/Et₂O which upon recrystallization from EtOH/Et₂O yielded 0.34 g (41%) of 15 as a white solid: mp: 81–82 °C; ¹H NMR (DMSO-d₆) δ 1.09 (d, J = 6.5Hz, 3H, CH₃), 1.15–1.42 (m, 17H, 4 x CH₂, 3 x CH₃), 1.51–1.72 (m, 4H, 2 x CH₂), 2.29 (s, 3H, CH₃), 2.52–2.60 (m, 1H, CH), 2.71–2.80 (m, 2H, CH₂), 2.89–2.96 (m, 2H, CH₂), 3.02–3.12 (m, 1H, CH), 3.41–3.49 (m, 1H, CH), 6.71 (br s, 1H, NH, D₂O ex), 7.10–7.19 (m, 4H, ArH), 8.03 (br s, 2H, NH₂, D₂O ex). Used in the next step without further characterization.

Trifluoroacetic acid (TFA; 10 mL) was added in a dropwise manner to an ice-cold, stirred solution of 15 (0.34 g, 0.83 mmol) in CH₂Cl₂ (10 mL); the resulting mixture was allowed to stir at room temperature for 24 h. Solvent was removed under reduced pressure and water (ca. 20 mL) was added. The aqueous layer was basified and extracted with EtOAc (4 × 10 mL). The combined extracts were dried (Na₂SO₄), filtered, and the solvent was evaporated under reduced pressure to yield 0.15 g of a yellow oil. A saturated HCl (g) solution in anhydrous Et₂O (5 mL) was added to a solution of the free base in anhydrous Et₂O (5 mL) and the reaction mixture was allowed to warm to room temperature and stir overnight. The precipitate was collected by filtration to yield a white solid that, upon recrystallization from iPrOH/Et₂O, afforded 0.08 g (29%) of 7 as a white solid: mp: 187–189 °C; ¹H NMR (DMSO-d₆) δ 1.10 (d, J = 6.1 Hz, 3H, CH₃), 1.21–1.44 (m, 8H, 4 x CH₂), 1.46–1.77 (m, 4H, 2 x CH₂), 2.28 (s, 3H, CH₃), 2.60 (dd, J = 10.5, 12.3 Hz, 1H, CH), 2.70–2.83 (m, 2H, CH₂), 2.85–2.99 (m, 2H, CH₂), 3.26 (dd, J = 2.4, 12.3 Hz, 1H, CH), 3.41–3.49 (m, 1H, CH), 7.10–7.19 (m, 4H, ArH), 8.02 (br s, 3H, NH₃, D₂O ex), 9.02 (br s, 2H, NH₃, D₂O ex). Anal. Calcd for (C₁₈H₃₂N₂•2HCl•1.2 H₂O) C, 58.27; H, 9.89; N, 7.55. Found: C, 58.34; H, 9.54; N, 7.67.

N¹-(4-Methylphenylethyl)-1,8-diaminooctane dioxalate (8).

A mixture of (4-methyl)phenylacetaldehyde (13) (Chikashita et al., 1987) (0.3 g, 2.23 mmol), N¹-Boc-1,8-diaminooctane (14)(Huang et al., 2014) (0.54 g, 2.23 mmol), PtO₂ (0.003 g, 0.014 mmol) and EtOH (20 mL) was shaken in a Parr hydrogenator under H₂ (ca. 45 psi) at room temperature for 16 h. The catalyst was removed by filtration through a Celite pad and the solvent was removed under reduced pressure. The crude product was purified through a silica gel column with CHCl₃/MeOH (95/5) to afford 0.10 g (12%) of 16 as a yellow oil; ¹H NMR (DMSO-d₆) δ 1.18–1.30 (m, 6 H, 3 × CH₂), 1.31–1.42 (m, 15H, 3 × CH₃, 3 x CH₂), 2.29 (s, 3H, CH₃), 2.48–2.53 (m, 2H, CH₂), 2.58–2.73 (m, 4H, 2 × CH₂), 2.88 (q, J = 6.6 Hz, 2H, CH₂), 3.30 (br s, 1H, NH, D₂O ex), 6.71 (br s, 1H, NH, D₂O ex), 7.10–7.19 (m, 4H, ArH). The reaction was repeated, and the product was used in the next step without further characterization.

Trifluoroacetic acid (5 mL) was added in a dropwise manner to an ice-cold stirred solution of 16 (0.09 g, 0.25 mmol) in CH₂Cl₂ (5 mL); the resulting mixture was stirred at room temperature for 24 h. Solvent was removed under reduced pressure and water (ca. 20 mL) was added. The aqueous layer was basified and extracted with EtOAc (4 × 10 mL). The combined extracts were dried (Na₂S0₄), filtered, and the filtrate was evaporated under reduced pressure to yield 0.7 g of a yellow oil. A saturated solution of oxalic acid in anhydrous Et₂O (5 mL) was added to a solution of the oil in anhydrous Et₂O (5 mL) and the reaction mixture was allowed to warm to room temperature and to stir overnight. The precipitate was collected by filtration to yield a white solid that, upon recrystallization from MeOH/Et₂O, afforded 0.5 g (45%) of 8 as a white solid: mp: 198–200 °C; ¹H NMR (DMSO-d₆) δ 1.18–1.41 (m, 8 H, 4 × CH₂), 1.44–1.69 (m, 4H, 2 × CH₂), 2.29 (s, 3H, CH₃), 2.70–2.82 (m, 2H, CH₂), 2.82–2.97 (m, 4H, 2 × CH₂), 3.02–3.16 (m, 2H, CH₂), 7.08–7.19 (m, 4H, ArH); Anal. Calcd for (C₁₈H₃₂N₂•2 C₂H₂O₄) C, 58.27; H, 9.89; N, 7.55. Found: C, 58.34; H, 9.54; N, 7.67.

N¹-(4-Methylphenylethyl)-N⁸-(4-methylphenylethyl)-1,8-diaminooctane dihydrochloride (9).

A mixture of 13 (Chikashita et al., 1987) (0.2 g, 1.49 mmol), 1,8-diaminooctane (0.1 g, 0.74 mmol), PtO₂ (0.002 g) and EtOH (10 mL) was shaken in a Parr hydrogenator under an H₂ atmosphere (ca. 45 psi) at room temperature for 16 h. Catalyst was removed by filtration through a Celite pad and the solvent was removed under reduced pressure. The oily residue was converted to its HCl salt using a saturated solution of HCl(g)/Et₂O that, upon recrystallization from MeOH, yielded 0.13 g (38%) of 9 as a white solid: mp: 320–321 °C; ¹H NMR (DMSO-d₆) δ 1.21–1.38 (m, 8H, 4 × CH₂), 1.52–1.70 (m, 4H, 2 × CH₂), 2.29 (s, 6H, 2 × CH₃), 2.81–2.96 (m, 8H, 4 × CH₂), 3.03–3.16 (m, 4H, 2 × CH₂), 7.10–7.19 (m, 8H, ArH), 8.75 (br s, 4H, NH₂, D₂O ex). Anal. Calcd for (C₂₆H₄₀N₂•HCl) C, 68.85; H, 9.33; N, 6.18. Found: C, 68.59; H, 9.30; N, 6.14.

N¹,N⁸-Dimethyl-N¹,N⁸-bis(4-methylphenylethyl)octane-1,8-diamine dioxalate (10).

A mixture of 9 (free base, 0.1 g, 0.22 mmol), formaldehyde (37%, 0.05 ml), 10% Pd/C (0.02 g) and MeOH (20 mL) was shaken in a Parr hydrogenator under an H₂ atmosphere (ca. 20 psi) at room temperature for 16 h. The catalyst was removed by filtration through Celite and solvent was removed under reduced pressure. The oily residue was converted to its oxalate salt using a saturated solution of (COOH)₂/Et₂O that, upon recrystallization from MeOH, yielded 0.07 g (56%) of 10 as a white solid: mp: 174–175 °C; ¹H NMR (DMSO-d₆) δ 1.21–1.39 (m, 8H, 4 × CH₂), 1.51–1.69 (m, 4H, 2 × CH₂), 2.25 (s, 6H, 2 × CH₃), 2.71 (s, 6H, 2 × CH₃), 2.89–3.07 (m, 8H, 4 × CH₂), 3.07–3.33 (m, 4H, 2 × CH₂), 7.07–7.19 (m, 8H, ArH), Anal. Calcd for [C₂₈H₄₄N₂•2.3 (COOH)₂] C, 63.59; H, 7.95; N, 4.55. Found: C, 63.39; H, 8.14; N, 4.83.

N¹-Methyl-N¹,N⁸-bis(4-methylphenethyl)-1,8-diaminooctane dioxalate (11).

Formaldehyde (37%, 0.9 mL), NaBH₄ (0.14 g, 3.72 mmol), and sufficient glacial acetic acid to adjust the pH to 6.0, were added to a stirred solution of 16 (0.45 g, 1.24 mmol) in MeOH (30 mL). The mixture was allowed to stir overnight at room temperature, and then quenched with NaOH solution (30%) to render the mixture strongly alkaline (pH 12–14). The aqueous phase was extracted with EtOAc (3 × 40 mL). The combined organic portion was washed with a saturated, aqueous NaCl solution, dried (MgSO₄), filtered and the filtrate was evaporated under reduced pressure to give 0.47 g (99%) of N-methyl 16 (i.e., 17) as a yellow oil: ¹H NMR (DMSO-d₆) δ 1.19–1.29 (m, 6 H, 3 × CH₂), 1.31–1.42 (m, 15H, 3 × CH₃, 3 × CH₂), 2.17 (s, 3H, CH₃), 2.29 (s, 3H, CH₃), 2.27–2.36 (m, 2H, CH₂), 2.41–2.49 (m, 2H, CH₂) 2.59–2.72 (m, 2H, CH₂), 2.89 (q, J = 6.6 Hz, 2H, CH₂), 6.73 (br s, 1H, NH, D₂O ex), 7.10–7.19 (m, 4H, ArH). Used in the next step without further characterization.

Trifluoroacetic acid (8 mL) was added in a dropwise manner to an ice-cold, stirred solution of the oil (i.e., 17) (0.47 g, 1.24 mmol) in CH₂Cl₂ (8 mL); the resulting mixture was allowed to stir at room temperature for 24 h. The solvent was removed under reduced pressure and water (ca. 20 mL) was added. The aqueous layer was basified and then extracted with EtOAc (4 × 10 mL). The combined extracts were dried (Na₂SO₄), filtered, and the solvent was evaporated under reduced pressure to yield 0.35 g (99 %) of N-methyl 8 (i.e., 18) as a yellow oil: ¹H NMR (DMSO-d₆) δ 1.19–1.31 (m, 8 H, 4 × CH₂), 1.33–1.47 (m, 4H, 2 × CH₂), 2.18 (s, 3H, CH₃), 2.28 (s, 3H, CH₃), 2.27–2.36 (m, 2H, CH₂), 2.41–2.49 (m, 2H, CH₂) 2.59–2.72 (m, 4H, CH₂), 3.32 (br s, 2H, NH₂, D₂O ex), 7.10–7.19 (m, 4H, ArH). Used in the next step without further characterization.

A mixture of 13 (Chikashita et al., 1987) (0.17 g, 1.24 mmol), 18 (0.35 g, 1.24 mmol), PtO₂ (0.002 g) and EtOH (20 mL) was shaken in a Parr hydrogenator under an H₂ atmosphere (ca. 45 psi) at room temperature for 16 h. The catalyst was removed by filtration through a Celite pad and the solvent was removed under reduced pressure to yield 0.52 g of a yellow oil. A saturated solution of oxalic acid in anhydrous Et₂O (5 mL) was added to a solution of the free base in anhydrous Et₂O (5 mL) and the reaction mixture was allowed to warm to room temperature and stir overnight. The precipitate was collected by filtration to yield a white solid that, upon recrystallization from EtOH, afforded 0.43 g (60%) of 11 as a white solid: mp: 162–164 °C; ¹H NMR (DMSO-d₆) δ 1.19–1.38 (m, 8 H, 4 × CH₂), 1.48–1.69 (m, 4H, 2 × CH₂), 2.29 (s, 6H, 2 × CH₃), 2.71 (s, 3H, CH₃), 2.79–3.02 (m, 8H, 4x CH₂), 3.03–3.21 (m, 4H, 2 × CH₂), 7.08–7.19 (m, 8H, ArH); Anal. Calcd for [C₁₈H₃₂N₂•2.6 (COOH)₂], C, 61.51; H, 7.56; N, 4.45. Found: C, 61.49; H, 7.38; N, 4.56.

3.2. Cell lines, plasmids and transfection.

Cells expressing dopamine or serotonin transporters of human origin were previously developed using the Flp-In T-REx 293 cell system by targeted integration of the cDNAs (Cameron et al., 2015; Solis et al., 2017). In control experiments where DAT or SERT expression were omitted the parental Flp-In T-REx 293 cells were used (Flp-In in Fig. 3). Cells subjected to intracellular Ca²⁺ determinations were plated in 96 well imaging plates and transfected using FuGENE 6 as transfection reagent. 6–8 h after transfection, cells were exposed to doxycycline (1μg/mL) to induce the expression of monoamine transporters for 3 days before the experiment (Cameron et al., 2015; Solis et al., 2017; Battisti et al., 2018). Voltage-gated Ca²⁺ channels were expressed by transient transfection using Fugene 6 and plasmids coding Ca_V1.2 (main subunit), and the auxiliary subunits β₃ and α₂δ. In addition, an EGFP coding plasmid was included in the transfection mix as transfection marker. This procedure was previously described in detail (Cameron et al., 2015; Solis et al., 2017; Battisti et al., 2018). Plasmids encoding GCaMP6s were provided by Dr. Douglas Kim (Chen et al., 2013) (Addgene #40753). When GCaMP6s proteins were used as Ca²⁺ indicator, the EGFP transfection marker was omitted. During the optimization process for the multi-well plate experiments, it was determined that α₂δ was not necessary for obtaining signals, and β₃ was only needed in DAT cells.

3.3. Intracellular Ca²⁺ determination using fluorescence microscopy.

Cells were placed on the stage of an Olympus IX70 fluorescence microscope equipped with a 20X 0.75NA objective, a polychromator – based light source (polychrome V, Till photonics) and an EMCCD camera (Luca Andor). Constant perfusion was achieved using a pressurized system and solution switching was controlled digitally using electronic valves (Automate Scientific). Solutions were warmed to ~35°C using ThermoClamp heater (Automate Scientific). Light excitation wavelengths, image acquisition and solution switching were controlled by Live Acquisition software (Till photonics). All tested compounds were diluted to the indicated concentration in IS solution (130 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose, in mM, pH adjusted to 7.4). To test substrates and inhibitors of monoamine transporters, protocols identical to the ones described previously were used (Solis et al., 2017; Battisti et al., 2018). Briefly, the automated protocol for substrates consisted of a 10 sec exposure to IS solution followed by a 5 sec exposure to a positive control (10 μM DA for DAT or 10 μM 5-HT for SERT), then after a wash cells were exposed for 5 sec to a given concentration of the test compound followed by a final 30 sec wash. To study inhibitors of transporters, cells were exposed for 10s to IS to set the baseline, and then were exposed for 5 sec to the positive control; after a 30 sec wash, the cells were exposed for 30 sec to a given concentration of inhibitor and then were exposed to the inhibitor plus 5 μM of the positive control for 5 sec and finally were exposed to a 30 sec wash using IS. When K⁺-induced depolarization was required, a high-K⁺ solution was used, having the same composition as IS except for an equimolar substitution of K⁺ for Na⁺. The experiments have the same timeline described above, but the high-K⁺ solution, was used instead of the monoamine substrate (see Fig. 3).

Data analysis and curve fitting to obtain EC₅₀ and IC₅₀ values were performed as described previously (Solis et al., 2017; Battisti et al., 2018).

3.4. Intracellular Ca²⁺ determination using FlexStation 3.

The FlexStation 3 is an automated fluorescence plate reader with temperature control that can inject solution to wells while measuring fluorescence, recording one column of 8 wells per run, thus a 96-well plate is read in 12 runs of 8 wells each. After 30s of basal fluorescence determination the compounds were automatically applied by the robotic injector, and the recording continue for 2,5 min for each well. For the study of a new agent in each run, a set of conditions were tested including at least 2 wells for the positive control (20 μM DA for DAT or 20 μM N-methyl 4-MA for SERT), 2–3 wells for test drug alone, 2–3 wells for test drug in combination with the positive control and in some runs wells of negative control (MDPV for DAT and fluoxetine for SERT). All measurements were taken at 34°C. All the Ca²⁺ determinations were done using IS as the external solution (see composition above). Only runs that have responses on the positive control wells were included in the analysis. The maximal GCaMP6s signal of each well was computed and the mean value of the responses of the “positive control” wells was assigned to 100% per run; the mean of the responses of “drug alone” wells and the mean of “drug plus positive control” wells per run were reported as percentage of the “positive control” wells for each run. Mean ± SEM of n ≥ 4 runs per condition are shown in Fig. 2.

3.5. APP⁺ uptake inhibition studies.

APP⁺ is a substrate of monoamine transporters that does not fluoresce in solution, but it fluoresces when taken up by cells and interacts with intracellular components (Solis et al., 2012). FlpIn-TRex HEK293 cells expressing human DAT or human SERT plated on 96-well imaging plates were transfected with a plasmid coding DsRED (Moerke et al., 2018; Yadav-Samudrala et al., 2019). Then the culturing media was supplemented with doxycycline 1 μg/ml to induce the expression of the monoamine transporters. After three days cells were placed on the stage of the Olympus IX 70 microscope described above. The DsRED signal was used to find the focal plane of the monolayer and APP⁺ uptake was measured under constant perfusion at room temperature. The wavelengths used to detect APP⁺ in cells were 460/10 nm (excitation) and 535/50 nm (emission). Cell perfusion was computer-controlled: the cells were first exposed to IS (described above) for 10 sec, then to a single concentration of the test compound for 40 sec, and finally to a mixture of the test compound and 5 μM APP⁺ for 30 sec. Each well was exposed to a single concentration of the test compound. Control wells were tested using the same protocol but omitting the test compound. Single cell analysis was done using ImageJ software and the maximal intensity of APP⁺ accumulation (after 30 min APP⁺ exposure) was recorded. The maximal fluorescent value of each cell in a well (20–40 cells) were averaged and represent the maximal uptake of that well. The mean value of each well was divided by that of the control condition (APP⁺ alone) for each experiment to obtain a normalized value. A dose-response curve was constructed using Prism 5.0 software, using an N ≥ 6 wells per concentration.

3.6. Statistics.

Non-linear-curve fitting, IC₅₀, EC₅₀, SEM and 95% confidence intervals (95%CI) were computed using GraphPad Prism 5.0 software.

Highlights:

• Pharmacology of DAT and SERT is relevant for drug discovery and drug abuse research

• Ca²⁺ channels and GCaMP6s in cells constitute fluorescent sensors for DAT or SERT

• These cellular sensors can discriminate substrates from inhibitors at DAT and SERT

• Bivalent arylalkylamine molecules were potent inhibitors at DAT and SERT

Abbreviations:

DA

Dopamine

5-HT

serotonin

NE

norepinephrine

DAT

dopamine transporter

SERT

serotonin transporters

NET

norepinephrine transporter

Ca_V

voltage-gated Ca²⁺ channel

4-MA

4-methyl amphetamine

MDPV

3,4 methylenedioxypyrovalerone

APP⁺

4-(4-(dimethylamino)phenyl)-1-methylpyridinium