A Rhodamine-Labeled Citalopram Analogue as a High-Affinity Fluorescent Probe for the Serotonin Transporter

Abstract

A novel fluorescent ligand was synthesized as a high-affinity, high specificity probe for visualizing the serotonin transporter (SERT). The rhodamine fluorophore was extended from an aniline substitution on the 5- position of the dihydroisobenzofuran ring of citalopram (2, 1-(3-(dimethylamino)propyl)-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-5-carbonitrile), using an ethylamino linker. The resulting rhodamine-labeled ligand 8 inhibited [³H]5-HT uptake in COS-7 cells (K_i = 225 nM) with similar potency to the tropane-based JHC 1-064 (1), but with higher specificity towards the SERT relative to the transporters for dopamine and norepinephrine. Visualization of the SERT with compound 8 was demonstrated by confocal microscopy in HEK293 cells stably expressing EGFP-SERT.

The serotonin transporter (SERT) is located in the membranes of presynaptic serotonergic neurons, and functions to transport excess serotonin into the cell from the synaptic cleft to terminate serotonergic signaling.¹ Although selective serotonin reuptake inhibitors (SSRIs) are used clinically for the treatment of anxiety and depression, drug-protein interactions and long-term effects of these drugs on SERT trafficking and function that affect clinical efficacy remain largely unknown. Fluorescent probes can be powerful tools for the detection of and structure-function studies on membrane proteins.^2–15

A fluorescently labeled ligand that selectively binds to a transporter or receptor protein can be used to localize protein distribution in tissue and monitor its trafficking, including internalization in live cells or tissue.^{3, 4} In addition, these tools can be used to study protein function, such as upstream and downstream molecular regulation,^{2, 10, 11} protein oligomerization,⁵ or potentially, clinical efficacy for certain pathological conditions.¹⁵ Combined with other techniques, these structure-function studies can aid in understanding ligand-protein interactions, such as protein conformational changes caused by ligand binding.^{7, 9} and also facilitate protein structure analysis. ^13–14

Previously, fluorescently tagged 2β–carbomethoxy-3β-(3,4-dichlorophenyl)tropane analogues (e.g. JHC 1-064, 1, Fig. 1) were synthesized and used to visualize dopamine transporter (DAT) trafficking in cultured live midbrain dopaminergic neurons.^{3, 4, 16} DAT, along with the norepinephrine transporter (NET) and SERT are three highly homologous transporters belonging to the neurotransmitter: sodium symporter (NSS) family, which share a common protein tertiary structure with twelve transmembrane (TM) domains.¹ Since the tropane-based analogues reported previously bind to all three transporters with similar affinities,¹⁶ a fluorescent molecule, based on a SERT-selective ligand, could be an important addition to the armamentarium of tools to study the structure and function of the SERT.

Figure 1.

JHC 1-064 (1) Citalopram (2) and its 5-anilino analogue (3).

The antidepressants citalopram (2, Fig. 1) and its S-enantiomer are selective serotonin reuptake inhibitors (SSRI) that bind with high affinity to SERT (K_i = 1.94 nM, 0.89 nM, respectively) and selectively over NET (ratio = 3070, 11800 respectively) and DAT (ratio = 4780, 9160 respectively).¹⁷ Structure-activity relationship (SAR) studies, based on a series of citalopram (2) analogues, have been described wherein modifications at the N-, 4, and 5-positions have identified points of attachment that maintain both high affinity and SERT selectivity.^17–20 Specifically, most modifications at the 5-position of the dihydroisobenzofuran ring were well tolerated at the SERT (K_i = 1–40 nM), and none of these analogues demonstrated high binding affinity at NET or DAT, which revealed that steric bulk is tolerated at this position.¹⁷ This is important as the fluorophore is a molecule of equal or greater molecular weight compared to the pharmacophore and thus must be attached to a position where steric bulk does not significantly diminish binding affinity. Among these compounds, the 5-(3′-NH₂-Ph) analogue 3 (K_i = 10.4 nM)¹⁷ was chosen as the template for synthesis of the first citalopram-based fluorescent ligand, as it contained a potentially modifiable NH₂ group.

In this study, we chose rhodamine as the fluorophore, as previous studies with Rhodamine RedX-NHS ester established that it was ideally suited for straightforward synthesis and visualization studies.¹⁶ In addition, rhodamine is widely used for fluorescence studies in biological systems due to its photostability and high quantum yield. It also has a suitable excitation spectrum for the use of relatively inexpensive lasers and is outside the intrinsic fluorescence from protein tryptophans.

The first synthetic strategy developed was to link compound 3 directly to Rhodamine RedX-NHS ester. Unfortunately, the aniline group was not sufficiently reactive, under basic reaction conditions, either at room temperature or at 45°C to couple to the rhodamine reagent, hence, an ethylamino linker was introduced.¹⁶ However, no reaction was detected under basic conditions using 2-bromoethylphthalimide as a reagent, even when the temperature reached >100°C. Another published method also failed to condense the aniline group of compound 3 with N-protected alanine under the condition of N-methylmorpholine and isobutylchloroformate.²¹ In addition, reaction between compound 3 and N-protected alanine anhydride²² failed and resulted in a salt form of 3.

The strategy shown in Scheme 1 wherein compound 3 was reacted with freshly prepared acyl chloride 5, made from the N-protected alanine 4, following a modified procedure²³ was successful. The resulting compound 6 was deprotected with hydrazine to give the free amine 7, which was reacted with rhodamine red-X NHS ester to give the desired fluorescent compound 8.

Scheme 1.

Synthesis of SERT fluorescent compound 8. Reagents and conditions: (a) 150°C, 2 h; (b) thionyl chloride, CH₂Cl₂, reflux, 5 h; (c) NEt₃, CH₂Cl₂, rt, overnight; (d) NH₂NH₂, rt, 5 h; (e) (i-pr)₂ethylamine, DMF, rt, overnight.

The binding affinity and SERT selectivity of compound 8 was determined measuring inhibition of [³H]5-HT uptake into COS-7 cells transiently expressing the hSERT and [³H]dopamine ([³H]DA) uptake into COS-7 cells transiently expressing the hDAT and hNET. These data were compared to the inhibition potency of 1 as determined previously¹⁶ (Table 1).

Table 1.

K_I values of novel citalopram analogues to hSERT, hDAT and hNET determined by their inhibition potency of radiolabeled substrates^a

compound	hSERT		hDAT		hNET
	[3H]5-HT inhib. Ki (nM)	n	[3H] DA inhib. Ki (nM)	n	[3H] DA inhib. Ki (nM)	n
5-HT (KM)	491[416;579]	8	nd		nd
1, JHC 1-064	392 [286;537]	3	62 [37;105]	3	194 [153;244]	3
2	16[12;22]	6	>1,000,000	2	>1,000,000	4
3	390[313;485]	6	nd		nd
8	225[208;241]	6	10900[9300;12900]	3	11800[10800;13000]	3
DA (KM)	-		1010[909;1120]	4	396[369;425]	5

^aUptake inhibition of [³H]5-HT at hSERT and [³H]dopamine ([³H]DA) at hDAT and hNET (Ki values) (mean [SE interval]). Ki values were determined from uptake inhibition experiments of indicated radiolabeled in COS7 cells stably expressing hSERT, hDAT or hNET. The K_i values were calculated from IC₅₀ values determined by nonlinear regression analysis of uptake data using the equation K_i = IC₅₀/(1+(L+ K_m)) where L is the concentration of [³H]5-HT or [³H]DA. The K_m value for 5-HT and DA were determined in parallel. nd, not determined.

The fluorescent compound 8 showed a similar potency to 1 for inhibition of [³H]5-HT uptake and was selective for SERT (~50-fold) over DAT and NET, although less selective than the parent molecule 2.

The fluorescent properties of compound 8 were determined as shown in Figure 2. The maximal excitation and emission wavelengths are at 570 and 592 nm, respectively, (I_F, fluorescence intensity. a.u., arbitrary units.)

Figure 2.

Excitation (black) and emission (red) wavelength spectra of compound 8.

To demonstrate specific labeling of SERT with compound 8, we used a HEK293 cell line stably expressing hSERT with EGFP fused to the N-terminus. Cells were incubated with compound 8 for 60 min at 37°C with or without the specific SERT inhibitor paroxetine. As shown in Figure 3, live cell imaging using confocal microscopy revealed a uniform distribution of compound 8 (red) at the cell surface overlapping with the distribution of EGFP-SERT (green). In the presence of paroxetine, no fluorescence was observed at the cell surface in the red channel. The same result was observed when incubation of compound 8 was performed in the presence of 1 μM S-citalopram or 100 μM 5-HT (data not shown). These data demonstrate that compound 8 labels SERT with high specificity and enables SERT visualization in live cells.

Figure 3.

Visualizing 8 binding to hSERT in transfected HEK293 cells. HEK293 cells stably expressing EGFP-hSERT were preincubated in buffer with or without 1 μM paroxetine for 30 min before one hour incubation with 20 nM 8 at 37°C. 8 labeling was visualized by confocal microscopy. Pictures are representatives of more than 3 individual experiments. Scale bar, 10 μm.

In summary, a novel fluorescent ligand selective for hSERT was designed and synthesized using citalopram (2) as the SERT-selective pharmacophore. The resulting rhodamine-labeled ligand 8 demonstrated suitable serotonin uptake kinetics (K_i = 225 nM) and selectivity over DAT and NET. Specific labeling of the SERT with compound 8 was visualized by confocal microscopy in HEK293 cells stably expressing EGFP-SERT. Further studies comparing compound 8 and 1 in labeling SERT in living neuronal cells are underway.

Abbreviations

SERT

serotonin transporter

NET

norepinephrine transporter

DAT

dopamine transporter

NSS

neurotransmitter:sodium (Na⁺/Cl⁻) symporter

SSRI

selective serotonin reuptake inhibitor

NBD

nitrobenzoxadiazolealanine

NMM

N-methylmorpholine

NHS

N-hydroxysuccinimide

EGFP

enhanced green fluorescent protein