Fluorescent non-canonical amino acid provides insight into the human serotonin transporter | Nature Communications

Nature Communications volume 15, Article number: 9267 (2024) Cite this article

Metrics details

The serotonin transporter (SERT), responsible for the reuptake of released serotonin, serves as a major target for antidepressants and psychostimulants. Nevertheless, refining the mechanistic models for SERT remains challenging. Here, we expand the molecular understanding of the binding of ions, substrates, and inhibitors to SERT by incorporating the fluorescent non-canonical amino acid Anap through genetic code expansion. We elucidate steady-state changes in conformational dynamics of purified SERT with Anap inserted at intracellular- or extracellular sites. This uncovers the competitive mechanisms underlying cation binding and assigns distinct binding- and allosteric coupling patterns for several inhibitors and substrates. Finally, we track in real-time conformational transitions in response to the interaction with Na+ or serotonin. In this work, we present a methodological platform reporting on SERT conformational dynamics, which together with other approaches will deepen our insights into the molecular mechanisms of SERT.

SERT plays an indispensable role in the synaptic homeostasis of serotonin (5-HT) by mediating the energetically unfavorable reuptake of released 5-HT1. The discovery that therapeutic manipulation of SERT function alleviated symptoms of depression, anxiety, and obsessive-compulsive disorder has made SERT the target for some of the most prescribed psychiatric medications2,3. Together with the transporters for dopamine (DAT) and norepinephrine (NET), SERT belongs to the monoamine transporter subfamily of the solute-carrier 6 (SLC6) transporters. These share the binding of numerous compounds, including therapeutic drugs for psychiatric disorders and substances of abuse, such as cocaine, amphetamine, and MDMA (ecstasy)3,4. SERT, DAT, and NET adopt a similar structure classified by 12 transmembrane domains (TMs) with TM 1-5 being inverted symmetrical to TM 6–10 (LeuT-fold)5,6,7. Substrates are thought to be transported by an alternating access mechanism where access to a central substrate binding pocket is regulated by intra- and extracellular gates1,6,8. It is believed that substrates bind to an outward-open state, stabilized by the binding of two Na+ ions adjacent to the substrate binding pocket in sites referred to as the Na1- and Na2 sites. This binding triggers the transition to an inward-open state from which substrate and Na+ are released to the intracellular environment1. While the formation of the extracellular gate involves concerted local and global structural rearrangements, that of the intracellular gate seems to rely mostly on the movement of the N-terminal half of TM1 (TM1a) and the partial unwinding of the cytosolic part of TM51,9,10,11. Despite the structural and mechanistic overlap with DAT and NET, SERT remains unusual by its functional and conformational interactions with K+12 and Li+13. Specifically, SERT counter-transports K+12,14, the binding of which promotes an inward-facing state9,10,15. Although initially thought to be unique for SERT, components of the interaction with K+ may reflect a more general mechanism also present in homologous transporters16,17,18. The binding of Li+, in contrast, is considered more SERT-specific, and its mechanistic nature is unknown13,19. Interestingly, Li+, which is prescribed as a psychiatric medication20, induces a distinct conformational- and conducting state of SERT13,21,22, which may contribute to its therapeutic effects.

Recent advancements in our understanding of SERT structure and function have mostly been driven by improvements in biophysical methodologies4,6,9,10,15,23. Nevertheless, techniques to study SERT that offer both high spatial and temporal resolution are lacking behind. Such techniques often rely on the site-specific attachment of a conformational reporter, performed mostly using chemical conjugation of thiol-reactive probes at cysteine (Cys) residues24. However, SERT contains 18 endogenous Cys residues of which some are essential for function or may not be fully accessible25. Consequently, site-specific labeling with fluorophores or spin probes via Cys-conjugation is challenging. In addition, most probes are relatively large and connected via linkers, which place the reporters relatively far from the protein backbone and may interfere with native function.

Fostered by the progress in genetic code expansion26, methods for incorporation of the fluorescent non-canonical amino acid (ncAA) 3-(6-acetylnaphthalen-2-ylamino)-2-aminopropanoic acid (Anap) into recombinant proteins have been developed for expression in yeast and mammalian cells27,28. Expressing proteins with Anap genetically encoded requires an orthogonal tRNA and aminoacyl-tRNA synthetase (aaRS) pair cognate for Anap. The anticodon of the orthogonal tRNA is complementary to the amber stop codon (UAG), implying that Anap incorporation relies on suppression of the amber stop codon, which is inserted by site-directed mutagenesis26,28,29. Conceptually, this labeling platform offers the potential to site-specifically incorporate Anap at all positions in a protein. Moreover, the size and short linker, slightly beyond those of a tryptophan, renders Anap representative of the underlying structure and less likely to compromise native protein function than classical conjugated fluorophores. Finally, Anap has a remarkable environmental sensitivity, which, based on the reasoning that local structural changes may modulate the microenvironment surrounding the incorporation site28,30, enables the use of changes in Anap fluorescence as a potential read-out for structural changes. Since its implementation, Anap has served as a reporter in studies of conformational changes for especially membrane proteins, although mostly for ligand- and voltage-gated ion channels30.

Inspired by these previous studies27,31,32, we here explore the use of Anap for the study of SERT conformational dynamics in response to the binding of substrates, ions, and SERT-modulating drugs. We find that Anap can be efficiently incorporated into SERT at the positions 86 and 556 on the intracellular- and extracellular sides, respectively, excellently positioned to monitor the gating dynamics. The fluorescence of V86Anap and F556Anap, both of which retain wild-type (WT)-like function, remains specific for purified SERT and highly sensitive to its conformation. We conduct spectroscopic experiments that allow us to report on specific changes in Anap fluorescence as a function of direct ion-, substrate-, and inhibitor binding to SERT, both under steady-state conditions and in real-time. Overall, this system enables us to bridge a methodological gap in the study of SERT, making a refining of its mechanistic models within reach.

Anap (Fig. 1a) is an environmentally sensitive fluorophore27,28, for which the emission peak shows large blue-shifts upon exchange from H2O to more hydrophobic solvents (Fig. 1a). To incorporate Anap into SERT, we co-transfected single-amber stop codon mutants together with the cognate tRNA-aaRS pair for Anap (referred to here as tRNA-aaRSAnap) in HEK293 cells supplemented with Anap in the culture medium (Fig. 1b). We screened a total of 24 sites for Anap incorporation in SERT and identified functional variants through measurements of cellular [3H]5-HT uptake (Supplementary Fig. 1a, b). Among these mutants, of which the majority exerted measurable 5-HT uptake, we selected V86Anap and F556Anap for further studies based on their 1) high uptake activity, 2) status as non-conserved positions in SLC6 transporters (Supplementary Fig. 1c), 3) complementary intra- and extracellular locations in the transporter structure, and 4) positioning in structural environments predicted to change in response to conformational transitions9,10 (Fig. 1c). Specifically, V86 resides in TM1a, which regulates intracellular gating by bending as much as 60° as the transporter interconverts between the inward- and outward-open conformations9,10. F556 is located in TM11, facing the extracellular vestibule, near helices undergoing re-arrangements during extracellular gating9,10.

a Excitation spectrum in H2O (black dashed line; peak at 352 ± 0.3 nm) and emission spectra in H2O (red; λmax at 491 ± 0.6 nm), DMSO (blue; λmax at 453 ± 0.4 nm), or ethyl acetate (EtOAc) (purple; λmax at 432 ± 0.3 nm) of free Anap (inset), normalized to the emission at 490 nm in H2O. n = 3. b Principle underlying Anap incorporation. Plasmids encoding tRNA-aaRSAnap or a TAG-mutant of SERT are, together with Anap, introduced into the cell within which transcribed tRNAAnap is charged with Anap by aaRSAnap. Charged tRNAAnap is recruited in response to the ribosome encountering the UAG (amber) codon on the SERT mRNA strand, thereby incorporating Anap into the growing protein chain (here following Met (brown) and Val (magenta)). Full-length SERT with Anap is translocated to the membrane. Created in BioRender (BioRender.com/s78o344). c Superimposed structures of outward-open (yellow; 5I6X) and inward-open (blue; 6DZZ) SERT, viewed within the plane of the membrane. The upper and lower enlarged boxes highlight position F556 and V86, respectively, as sticks in red. d [3H]5-HT uptake in HEK293 cells following lipofectamine transfection with SERT WT (gray), V86Anap (red), or F556Anap (blue), with(+) or without(-) tRNA-aaRSAnap or Anap. Data are normalized to that with WT only (open bar). n = 4. e [3H]5-HT uptake in HEK293 cells expressing SERT WT (gray circles), V86Anap (red squares), or F556Anap (blue triangles) as a function of increasing concentrations of unlabeled 5-HT. Data are normalized to a control without unlabeled 5-HT (ctrl) and modeled by a non-linear regression. This yields IC50 values for WT, V86Anap, and F556Anap of 606 [537; 684], 333 [307; 361], and 593 [552; 637] nM, respectively (mean [S.E.M. interval]), of which the IC50 for V86Anap is significantly decreased (p = 0.0026) relative to that for WT. n = 4. Data points are shown as mean ± S.E.M., with n defining the number of biological replicates. The statistical analysis is performed using a one-way analysis of variance (ANOVA) with Bonferroni multiple comparison correction. Data are provided as a Source Data file.

The factors determining the efficiency by which amber codons are being suppressed by the binding of charged orthogonal or endogenous tRNAs are not fully understood, but are believed to be dependent on the mRNA codon sequence context33,34. To estimate the extent to which the uptake activity for V86Anap and F556Anap resulted from unspecific read-through, i.e., the incorporation of endogenous amino acids in response to the inserted UAG, these mutants were transfected into cells with and without tRNA-aaRSAnap and/or Anap (Fig. 1d). Without tRNA-aaRSAnap, Anap, or both, we observed essentially no cellular 5-HT uptake by cells transfected with either of the two mutants. This indicated that unspecific read-through, by recruitment of endogenous tRNAs or erroneously charged tRNAAnap, or translation re-initiation35, was negligible. In contrast, combining tRNA-aaRSAnap and Anap in cells expressing V86Anap and F556Anap yielded 5-HT uptake corresponding to 63 ± 1 and 53 ± 1 %, respectively, of that for cells expressing SERT WT.

To assess whether the transport function of SERT was affected by the V86Anap or F556Anap mutations, we determined the Km for 5-HT by homologous inhibition of [3H]5-HT uptake (Fig. 1e). While the Km for V86Anap showed a small yet significant decrease, the Km for F556Anap was indistinguishable from that of WT. Overall, these observations suggested that Anap incorporation for V86Anap and F556Anap was highly efficient, specific, and functionally tolerated.

Previous work with Anap has relied on the characterization of membrane proteins in intact cells36 or isolated membranes37,38. While such studies show robust changes in Anap fluorescence, the signal from specifically inserted Anap will compete with those from free Anap in intracellular compartments and from unspecifically incorporated Anap in endogenous proteins30, the latter being a consequence of the fact that 23% of human genes terminate on UAG39. To reduce interference from these unspecific Anap signals, we detergent-solubilized and purified full-length V86Anap and F556Anap in conditions previously applied for the structural determinations of SERT in all major conformations6,10. To express V86Anap and F556Anap in quantities sufficient for purification, we implemented a polyethyleneimine-based transfection platform for suspension HEK293expi cells. Interestingly, it has been shown that co-expression of the eukaryotic release factor 1, carrying an E55D mutation (eRF1-E55D), can improve ncAA incorporation by reducing termination selectively at the amber codon40. Accordingly, we co-expressed V86Anap or F556Anap together with eRF1-E55D, for which we observed a >50% increase in 5-HT uptake relative to that of cells without eRF1-E55D co-expressed (Supplementary Fig. 1d, e). Moreover, the effect of eRF1-E55D in the absence of Anap remained low, indicating that the increased uptake could be attributed mostly to SERT with Anap incorporated.

To assess whether purified V86Anap and F556Anap were functional, we reconstituted these into liposomes and confirmed that both mutants retained the ability to transport 5-HT (Fig. 2a). This suggested that, when exposed to a Na+ gradient, the mutants could assume all the conformational states required for transport. Furthermore, we measured the concentration-dependent binding of the selective serotonin reuptake inhibitor (SSRI) [3H]S-citalopram using the scintillation proximity assay41 (Fig. 2b). V86Anap and F556Anap both showed specific binding of [3H]S-citalopram with affinities indistinguishable from that of WT. When repeating the binding assay with samples of purified protein expressed in the absence of Anap, the Bmax values were reduced to 16.1 ± 1.8 and 10.8 ± 1.0 % of those for V86Anap and F556Anap, respectively. This indicated that the unspecific read-through for purified SERT, probably originating from the incorporation of endogenous amino acids in response to the amber stop codon, was comparable to that determined by cellular 5-HT uptake (Fig. 2b; Supplementary Fig. 1d). Immunoblotting furthermore showed the relative binding of [3H]S-citalopram, for each condition, to match the total amounts of SERT, which, for V86Anap and F556Anap, existed mostly in the fully mature state42 (Fig. 2c). To estimate the intrinsic stability of purified V86Anap and F556Anap at room temperature (RT), we measured their ability to bind [3H]S-citalopram following increasing pre-incubation periods at RT (Supplementary Fig. 2a). For V86Anap and F556Anap, we obtained 86 ± 1 and 93 ± 4% retained activity, respectively, after 1 h, similar to that of WT. This stability was crucial to avoid the fluorescent contamination from degraded or denatured SERT.

a Time-dependent [3H]5-HT uptake into liposomes reconstituted with SERT WT (gray circles), V86Anap (red squares), or F556Anap (blue triangles). Data are fitted to a one-phase association. n = 3. b Saturation binding of [3H]S-citalopram to purified SERT WT (gray circles), V86Anap (red squares), and F556Anap (blue triangles) expressed in the presence (solid lines) or absence (dashed lines) of Anap. Data are fitted to a one-site binding model and, for V86Anap and F556Anap, normalized to the Bmax predicted for protein expressed in the presence of Anap. Affinities (Kd) for V86Anap and F556Anap are 2.3 ± 0.2 nM and 2.8 ± 0.1 nM, respectively, which are not significantly different from that of WT (2.7 ± 0.2 nM). n = 3. c Western blotting (top) and in-gel fluorescence (bottom) analysis of purified WT, V86Anap, and F556Anap expressed in the presence (+) or absence (-) of Anap, and pre-incubated with (+) or without (-) PNGase F. The smear at ~100 kDa corresponds to fully mature SERT (M), whereas that at ~60 kDa reflects un-glycosylated/immature SERT (I). The western blotting band intensities of de-glycosylated V86Anap and F556Anap expressed in the absence of Anap are 16.8 ± 3.4 % and 8.1 ± 1.4 % of those for V86Anap and F556Anap expressed in the presence of Anap. The in-gel fluorescence band intensity of de-glycosylated WT is 6.7 ± 1.4 % and 7.0 ± 2.3 % of those for V86Anap and F556Anap, respectively. Western blotting and in-gel fluorescence analyzes were repeated independently 5 and 4 times, respectively, with similar results. d Emission spectra (arbitrary units) of purified WT (gray), V86Anap (red), and F556Anap (blue), alongside that of 5 nM free Anap (black dashed line). The total fluorescence of V86Anap is 40 ± 1 % of that for F556Anap. n = 3 – 8. Data points, affinities, ratios, and λmax values represent mean ± S.E.M., with n defining the number of biological replicates. Statistical analyzes are performed using a one-way ANOVA with Bonferroni multiple comparison correction. Data are provided as a Source Data file.

To uncover the spectral characteristics in Anap fluorescence for purified V86Anap and F556Anap, we recorded their excitation- and emission spectra. The excitation peaks remained similar between free- and incorporated Anap (Supplementary Fig. 2b). However, we observed 67 ± 1 and 45 ± 1 nm blue-shifts in the emission spectra peaks (λmax) of V86Anap and F556Anap, respectively, relative to that of free Anap (Fig. 2d). These blue-shifts exceeded that observed between free Anap in water and DMSO (Fig. 1a), which could not be accounted for by detergents or other buffer components (Supplementary Fig. 2c), suggesting that the incorporated Anap resided in relatively hydrophobic environments in both mutants. To estimate the fluorescence specificity for V86Anap and F556Anap, we purified SERT WT expressed under the same conditions as those for the mutants. As such, any residual fluorescence signal would originate from contaminating Anap sources. When comparing the emission spectrum of SERT WT with those of the Anap mutants, even despite the relative protein purity (Supplementary Fig. 2d), we found the fluorescence contribution from V86Anap and F556Anap to be 87 ± 2 and 95 ± 1 %, respectively (Fig. 2d). This contribution was substantiated by in-gel SDS analyzes for which the fluorescent signal for V86Anap and F556Anap corresponded to mature SERT42 (Fig. 2c), as also observed by immunoblotting, while size-exclusion chromatographic analysis confirmed the homogeneity of the sample (Supplementary Fig. 2e). Surprisingly, the total fluorescence signal for purified samples of V86Anap was, despite the higher expression, less than half of that for F556Anap (Fig. 2d). Knowing that the fluorescence of Anap is sensitive to its environment, we denatured SERT to investigate whether the difference in Anap fluorescence between V86Anap and F556Anap originated from differences in the nearby protein environment. Indeed, following thermal denaturation (Supplementary Fig. 2f) or treatment with guanidinium chloride (Supplementary Fig. 2g–i), the total fluorescence (area under the curve) of V86Anap increased by ~50 %. In contrast, the fluorescence of F556Anap decreased by ~25 %. Taken together, the results for purified SERT suggested that the measured binding activity and Anap fluorescence were specific for SERT with Anap incorporated in either position 86 (V86Anap) or 556 (F556Anap). Furthermore, the fluorescent properties of Anap were influenced by the surrounding protein structure at both positions.

The effects of ion interactions and their mechanistic role in SERT function are poorly understood, in particular those for Li+ and K+. It has generally been challenging to measure ion binding directly through the use of ion radioisotopes (e.g. 22Na and 40K) due to their generally low affinity and the hazard associated with handling large quantities of radioactive material. Therefore, we employed V86Anap and F556Anap to investigate whether the fluorescent spectral properties of V86Anap and F556Anap were sensitive to the binding of Na+10,43, Li+13,15, and K+9,10,15,16, while keeping Cl- constant. In the presence of Na+, we observed profound differences in Anap emission spectra relative to those obtained with the inert cation N-methyl-d-glucamine (NMDG+) (Fig. 3a, b). Specifically, the intensities at the wavelength corresponding to the λmax in Na+ increased by 20.7 ± 1.7 and 69 ± 1.3 % for V86Anap and F556Anap, respectively. For F556Anap, the higher fluorescence yield was accompanied by an 18 ± 2 nm red shift in the emission peak, whereas for V86Anap, it coincided mostly with a narrowing in the spectral shape. The spectra in the presence of Li+ closely resembled those in NMDG+, while substitution with K+ caused a minor decrease in the fluorescence of V86Anap and an increase in that of F556Anap. Notably, while a change in Anap fluorescence likely reflects a local conformational shift, we cannot exclude the possibility that other factors could also influence the fluorescence spectra. Moreover, the absence of a fluorescence change does not necessarily imply that no conformational alteration has occurred. Nevertheless, the finding that the fluorescence for both V86Anap and F556Anap in their cation-free states (NMDG+) and in the presence of Li+ resembled that in K+ could suggest that the conformational equilibria of SERT in NMDG+ and Li+ mirrored that in K+ more so than that in Na+. Considering the proposed effects of Na+ and K+ binding on the equilibrium between the inward- and outward-facing SERT conformations9,10,15, the differences in Anap fluorescence for V86Anap and F556Anap in Na+ and K+ were consistent with the conformational differences between the outward- and inward-open conformations, respectively.

a, b Emission spectra of purified V86Anap (a) and F556Anap (b) following incubation in buffer containing 200 mM Na+ (red), NMDG+ (black), Li+ (purple), or K+ (blue), normalized to the fluorescence intensity at λmax in Na+. For V86Anap, the λmax values in NMDG+, Li+, and K+ are not significantly different from that in Na+, whereas for F556Anap, these are blue-shifted by 18 ± 2, 16 ± 2, and 20 ± 1 nm, respectively (p < 0.0001). The values of n are specified in Supplementary Table 1. c, d Emission spectra as a function of increasing concentrations of Na+ (NMDG+ as counter-ion) for V86Anap (c) and F556Anap (d), with the directionalities for the increasing [Na+] indicated by arrows. Data are normalized to the fluorescence intensity at λmax in Na+. e For each spectrum in (c), the spectral ratios are computed and plotted as a function of [Na+]. The control (ctrl) is without Na+. Data are fitted to a Hill model and normalized to its predicted Bmax. n = 4. f For each spectrum in (d), the spectral ratios are plotted as in (e) in a background of NMDG+ (gray squares), Li+ (blue triangles), and K+ (red circles). The Hill coefficient in Li+ is significantly less steep than those in NMDG+ (p = 0.006) and K+ (p = 0.022). n = 3. EC50 values and Hill coefficients from (e) and (f) are summarized in Table 1. Data points, λmax values, intensity changes, and Hill coefficients are mean ± S.E.M., with n defining the number of biological replicates. In (a, b, and f,) data are analyzed with a one-way ANOVA with Dunnett multiple comparison correction, comparing each mean to that of a control (NMDG+). Data are provided as a Source Data file.

To assess the conformational effects of Cl-, we exchanged Cl- with acetate, while keeping Na+ constant. Exchanging Cl- to acetate gave rise to a minor decrease in the fluorescence of V86Anap, compared to the changes observed by the substitution of cations, whereas no effect was observed for F556Anap (Supplementary Fig. 3a, b). While this suggested that the Na+-induced fluorescence changes were essentially independent of the presence of Cl−, understanding the exact role of Cl- on the conformation of SERT will require further studies.

To quantify the spectral changes promoted by the exchange of cations, we performed a spectral shift analysis44. To do so, we defined a parameter for which we averaged the fluorescence within a fixed range of wavelengths on each side of the λmax, determined for both V86Anap and F556Anap in Na+, and computed their ratio (left/right; see “methods”) (Supplementary Fig. 3c). The same parameters were then used to calculate the fluorescence ratios for all subsequent fluorescence spectra. This parameter, which we refer to as the spectral ratio, is sensitive to spectral changes that are asymmetrical with respect to λmax, i.e., blue/red shifts of the emission spectra or changes in the spectral shape (narrowing/broadening). Notably, we found significant differences in the spectral ratios by the addition of NMDG+, Li+, and K+ (Supplementary Fig. 3d, e), suggesting that the binding of ions changed the fluorescent properties of the inserted Anap probes. Encouraged by the large difference in signal with and without Na+, we proceeded by recording emission spectra following incubation in increasing concentrations of Na+, using NMDG+ as the counter-ion (Fig. 3c, d). We plotted the derived spectral ratios as a function of Na+, which yielded sigmoidal and saturable responses for both V86Anap and F556Anap. When modeled by the Hill equation, the EC50 values for Na+ were 10.4 [9.7;11.1] and 14.7 [14.3;15.1] mM for V86Anap and F556Anap, respectively (Fig. 3e, f). These were in agreement with potencies reported for e.g., Na+-dependent 5-HT uptake3,45,46. Furthermore, the model displayed Hill coefficients of 2.43 ± 0.26 for V86Anap and 1.77 ± 0.07 for F556Anap, consistent with the cooperative binding of two Na+ ions47. Besides directly estimating the steady-state affinity for Na+, these results suggested that the distinct shifts induced by Na+ binding involved environmental changes in regions encompassing both residues at position 556 on the extracellular side and 86 within the intracellular gate.

Since the fluorescence in Li+ and K+ did not differ much from that with NMDG+, we were unable to measure their binding directly from Anap spectral shifts. However, we envisioned that Li+ and K+ binding could be investigated by their ability to modulate the Na+-dependent spectral shifts. Accordingly, we repeated the Na+ titration in the presence of 150 mM NMDG+, Li+, or K+, using F556Anap (Fig. 3f). In the presence of Li+ or K+, compared to NMDG+, the EC50 for Na+ binding increased while the Bmax remained essentially unchanged. With K+, the Hill coefficient was unaffected, assigning K+ as a classical competitive inhibitor for Na+. With Li+, in contrast, the Hill coefficient was significantly decreased. Given the shared competitive inhibitory mechanism of Li+ and K+, we estimated their Ki, using the Cheng-Prusoff equation, to be 77.4 [72.7;82.5] and 91.0 [89.2;92.8] mM, respectively. Notably, this Ki for K+ is similar to those determined for DAT17 and LeuT through Na+-dependent radioligand binding studies16,18, while that for Li+ overlaps with the Li+ affinity determined for DAT by electrophysiological recordings of leak currents48.

The seemingly distinct conformational equilibria observed in Na+, Li+, and K+ might arise from the mutually exclusive binding to different conformations13,16. If so, any bias in the conformational equilibria imposed by the mutations or changes in environment might disturb the apparent ion potencies or inhibitory mechanisms. To assess whether these properties were affected by the F556Anap mutation, we measured Na+-dependent [3H]S-citalopram binding in the presence and absence of Li+ and K+ for purified SERT WT and F556Anap (Supplementary Fig. 4a). For both WT and F556Anap, we obtained EC50 values and Hill coefficients, including the decreased Hill coefficient in Li+, resembling those observed for F556Anap by fluorescence experiments (Fig. 3f; Table 1). Furthermore, when repeating the experiment with SERT WT in native membrane preparations, the results were comparable to those obtained for purified SERT (Supplementary Fig. 4b; Table 1). These results suggested that the functional impact of the F556Anap mutation and the micellar environment was minimal and that the fluorescence experiments mirrored the interactions of Na+, Li+, and K+ with SERT WT in native membranes.

SERT is the main target for most drugs used in the treatment of major depression and anxiety disorders, as well as target for widely abused psychostimulants and hallucinogens3,4. These encompass 1) classical competitive inhibitors, such as SSRIs and cocaine, that bind in the central S1 binding site and are suggested to stabilize outward-facing states4,6,49, 2) hallucinogenic drugs like ibogaine that are proposed to promote inward-facing states50, and 3) synthetic psychedelic substrates like MDMA that induce substrate efflux51. Here, we assessed whether V86Anap and F556Anap offered the possibility to distinguish between different ligand binding modes. We recorded emission spectra of V86Anap and F556Anap following incubation in Na+ with the inhibitors paroxetine, fluoxetine, S-citalopram, cocaine, imipramine, and ibogaine, as well as with the substrates MDMA and 5-HT (Supplementary Fig. 5a, b). For each spectrum, we computed the ligand-induced change in fluorescence intensity at λmax (determined in Na+) against the change in spectral ratio relative to a control (Fig. 4a, b). Of note, the compounds were applied in saturating concentrations, which did not affect the fluorescent properties of free Anap (Supplementary Fig. 5c) and were sufficient to outcompete the binding of 5 nM [3H]S-citalopram to SERT WT (Supplementary Fig. 5d).

a, b Scatter plots for V86Anap (a) and F556Anap (b) showing the shifts in fluorescence intensity (ΔFluorescence intensity) against the shifts in spectral ratios (ΔSpectral ratio) with different compounds, relative to a control (no ligand; dotted lines). The shifts in fluorescence intensities (y-axis; 425 nm for V86Anap, 448 nm for F556Anap) and shifts in spectral ratios (x-axis) are derived from the spectra in Supplementary Fig. 5a, b. Exact p-values for the statistical comparison between ligand-induced changes in ΔFluorescence intensity or ΔSpectral ratio are listed in Supplementary Table 2. c Shifts in fluorescence intensity (blue; left y-axis) and spectral ratio (red; right y-axis) for V86Anap-I172M and I172M-F556Anap incubated with S-citalopram, relative to a control (no ligand), derived from the spectra in Supplementary Fig. 5e, f. Compared to those for V86Anap and F556Anap (shown for reference as dashed bars), the spectral changes with S-citalopram are significantly reduced for V86Anap-I172M (intensity: p = 0.0004; spectral ratio: p = 0.0007) and I172M-F556Anap (intensity: p < 0.0001; spectral ratio: p < 0.0001), compared using an unpaired t test. d Emission spectra of F556Anap incubated with increasing concentrations of 5-HT and normalized to the fluorescence intensity at 448 nm for the control (no ligand). The directionality of increasing [5-HT] is indicated with an arrow. e Spectral ratios computed from the spectra in (d), plotted as a function of [5-HT], with that devoid of 5-HT labeled as control (ctrl). The data are fitted to a Hill equation, predicting an EC50 value of 1.01 [0.95; 1.08] µM (mean [S.E.M. interval]) and a Hill coefficient of 1.12 ± 0.17. n = 3. Data points and Hill coefficient are mean ± S.E.M., with n defining the number of biological replicates. The values of n for a-c are specified in Supplementary Table 1. Data are provided as a Source Data file.

Surprisingly, the emission spectra of V86Anap and F556Anap in the presence of the different inhibitors gave rise to distinct changes in both fluorescence intensities and spectral ratios (Fig. 4a, b; Supplementary Table 2). For F556Anap, only the changes with fluoxetine and paroxetine were not significantly different and resembled the control (no ligand). For V86Anap, on the other hand, the changes in intensity and spectral ratio with paroxetine were among the largest observed and were significantly different from those with fluoxetine. Consequently, when comparing the fluorescence patterns from both Anap insertions, each inhibitor left a distinct fluorescent footprint (Fig. 4a, b; Supplementary Fig. 5a, b). This suggested that among the inhibitors, even those considered to rely on the same pharmacological mechanisms, there were differences in the binding-induced structural dynamics. Such differences were not readily evident from the corresponding structures of SERT in complex with e.g., SSRIs6 or cocaine49. To substantiate that the shifts in fluorescence were indeed mediated by the binding events in the S1 site, we introduced the S1 site mutation I172M, shown previously to lower the affinity for S-citalopram by almost three orders of magnitude52,53. Notably, emission spectra of V86Anap-I172M and I172M-F556Anap in NMDG+, K+, and Na+ resembled those of V86Anap and F556Anap, indicating that the conformational equilibria and responsiveness of the I172M mutants were preserved (Supplementary Fig. 5e, f). Subsequently, we assessed whether the shifts in fluorescence promoted by the addition of S-citalopram were affected by the I172M mutation. We observed that these spectral changes for V86Anap-I172M and I172M-F556Anap were significantly reduced relative to those for V86Anap and F556Anap (Fig. 4c; Supplementary Fig. 5e, f), supporting a direct correlation between the binding of S-citalopram and the changes in Anap fluorescence.

Whereas the spectral responses to the binding of inhibitors varied, the two substrates, 5-HT and MDMA, produced similar response patterns (Fig. 4a, b). Specifically, both decreased the fluorescence intensity of F556Anap while showing no effect on that of V86Anap. Furthermore, both increased the spectral ratio for F556Anap, otherwise only seen with noribogaine, and decreased the spectral ratio for V86Anap. Although these changes differed in size, with the change in spectral ratio for F556Anap in 5-HT exceeding that in MDMA, they overall assigned substrates with a distinct fluorescent pattern unlike those of the inhibitors. To assess whether the 5-HT-induced responses were specific, we co-applied 5-HT and paroxetine. In the presence of paroxetine, we observed no effect of 5-HT on either V86Anap or F556Anap (Fig. 4a, b), suggesting that paroxetine binding was sufficient to fully inhibit the 5-HT-induced fluorescent response. Subsequently, we titrated in 5-HT to determine its EC50, for which we obtained a response in agreement with the law of mass action for a single binding site (Fig. 4d, e). This predicted an EC50 value of 1.01 [0.95; 1.08] µM, which is similar to Km values previously reported3,6,54.

The binding of Na+ and 5-HT to SERT represents the initial binding events in the transport cycle that trigger the transition to the inward-facing state1. As previously noted, we observed substantial spectral changes following the binding of Na+ for both mutants (Fig. 3a, b) and by the binding of 5-HT, although only for F556Anap (Fig. 4d). This motivated us to study the real-time conformational transitions associated with the binding of Na+ and 5-HT using stopped-flow fluorometry (Fig. 5a). To obtain real-time measurements of the Anap fluorescence during the association of Na+ to SERT, we pre-equilibrated V86Anap and F556Anap in the absence of Na+, to promote the Na+-free state, before applying 100 mM Na+ (Fig. 5b). For both V86Anap and F556Anap, the rapid application of Na+ produced relatively slow changes in fluorescence, despite the high Na+ concentration, that evolved following monoexponential time courses with similar time constants (τ) (~5 sec). Notably, Na+ has been proposed to bind on time-scales several orders of magnitude faster to both the inward- and outward-facing states of SERT55. Accordingly, the time-dependent changes in fluorescence presented here were likely to reflect a two-step process, involving 1) the binding of Na+ to the Na1 and Na2 sites, and 2) the subsequent shift in conformational population, of which the latter was responsible for the changes in Anap fluorescence. Assuming that Na+ saturated both Na+ sites on sub-second timescales55,56, the slow decrease (V86Anap) and increase (F556Anap) in fluorescence observed here could describe the transition of Na+-bound SERT to the state where it is prone to bind substrate and inhibitors. To the best of our knowledge, such real-time measurements of SERT dynamics following Na+ binding have not been reported before. We cannot, however, exclude the possibility that Na+ binding followed the spontaneous transition to the outward-open Na+-selective state through conformational selection. Next, we measured the fluorescence changes following Na+ dissociation (Fig. 5b), although mechanical limitations prevented us from diluting Na+ to an extent where its rebinding could be ignored. Instead, we diluted Na+ in a range within which a concentration jump would produce the highest signal (Fig. 3c, d). For both mutants, we observed, when fitted by a monoexponential model, a time constant of ~10 sec. Taken together, these results suggested that the site-specific insertion of Anap enabled us to probe the real-time conformational dynamics in response to association and dissociation of Na+.

a Simplified cartoon illustrating the experimental setup for stopped-flow fluorometry. The contents of two inlets (A, blue; B, red) are mixed while the sample is excited and its emission measured. The blue and red bars above the individual reactions in (b) and (d) represent the content of these inlets. b Time-dependent traces showing a baseline, the association of Na+, and the dissociation of Na+ (~8 technical replicates per n) for V86Anap (green) and F556Anap (gray). Data were fitted to one-phase exponential models, yielding the time constants (τ) specified at each trace. c Cartoon illustrating the binding events in (d). Following the equilibration in Na+ (middle), the application of 5-HT causes a conformational shift associated with a decrease in the fluorescence intensity of Anap (right). When paroxetine is present, it binds following the dissociation of 5-HT, which kinetically traps SERT and prevents the rebinding of 5-HT (left). Created in BioRender (BioRender.com/l18c793). d Traces for the baseline (~14 technical replicates per n), association (~14 technical replicates per n), and dissociation (~7 technical replicates per n) of 5-HT for F556Anap, all of which are conducted in a background of Na+ (black bar). The 5-HT association is modeled by a one-phase exponential association, whereas the dissociation is fitted to a two-phase exponential dissociation. These fits predict a pseudo-first order Kon of 2.47 × 105 ± 0.06 × 105  s−1 and a fast Koff of 0.72 ± 0.04 s−1. The rates convert to an affinity (Koff / Kon) of 2.93 ± 0.15 µM. Real-time fluorescence changes (emission measured at λ > 495 nm; arbitrary units (arb. units)) were normalized to the first (association) or last (dissociation) data point. Data points and time constants are mean ± S.E.M., with n defining the number of biological replicates. Data are provided as a Source Data file.

Similarly, we measured the real-time fluorescence changes following the binding of 5-HT for F556Anap (Fig. 5c, d). We mixed F556Anap, pre-equilibrated in 200 mM Na+, with a saturating concentration of 5-HT and observed a rapid decrease in fluorescence with a monoexponential time course, for which the time constant was more than an order of magnitude faster than that for Na+ binding. The association rate constant for 5-HT binding to SERT is, however, proposed to be in the order of 106 – 107 M−1 s−1 55, predicting 5-HT binding of to occur on time-scales even shorter than those observed here. This suggested, as for the Na+ binding, that the fluorescence decrease for F556Anap effectively captured the conformational ensemble shift following 5-HT binding. To monitor the fluorescence changes associated with 5-HT dissociation, we diluted 5-HT-bound F556Anap into a buffer supplemented with a large molar excess of paroxetine to prevent rebinding of released 5-HT (Fig. 5c, d). Accordingly, we relied on the ability of paroxetine to block the binding of 5-HT and produce a fluorescent signal similar to that with only Na+ bound (Fig. 4b). We observed a signal that returned to its initial baseline with a time-constant of ~1 sec. Interestingly, the high signal-to-noise furthermore allowed us to resolve a secondary slow component (τ = ~25 sec) of the 5-HT dissociation. This could arise from slow re-orientation from one or more states in which 5-HT was trapped, from a second bound 5-HT molecule, or from effects ascribed to paroxetine. The specific assignment will require further studies. Altogether, whereas previous studies have probed the 5-HT binding itself, our results suggested that we could follow in real-time the ensemble shifts in protein conformation that followed the 5-HT association and dissociation, of which the latter might proceed from distinct populations. While acknowledging that the observed time-constants primarily reflected conformational transitions rather than direct binding events, we estimated the pseudo-first order association rate constant for 5-HT, facilitated by the molar excess of 5-HT relative to SERT. Notably, using this rate constant and the apparent Koff (fast) for 5-HT, the resulting values translated to an affinity for 5-HT of 2.93 ± 0.15 µM, coinciding with the value determined under equilibrium (Fig. 4e).

In this study, we demonstrate the efficient suspension-scale expression and purification of SERT (~0.5 mg per l of culture) with Anap genetically encoded at an intra- (V86Anap) and an extracellular (F556Anap) site, none of which compromises function. By steady-state fluorescence spectroscopy, we show that Anap fluorescence can be exploited to study conformational changes in SERT in response to the binding of ions, inhibitors, and substrates. Each ligand can be assigned with a distinct fluorescent footprint despite experimentally resolved structures of SERT showing many of these to stabilize similar outward-open conformations. The sensitivity of the system furthermore enables us to estimate the steady-state EC50 values for Na+ and 5-HT, as well as the inhibitory potencies and mechanisms for Li+ and K+ binding. Finally, we demonstrate a stopped-flow spectrofluorometric approach to measure the real-time conformational dynamics following the association and dissociation of Na+ and 5-HT, the likes of which, we believe, have not been performed before for SERT or other related transporters.

Purified SERT in Na+- or K+-bound states is suggested to predominantly adopt outward- or inward-open10 conformations, respectively, which is likely to account for the distinct fluorescent footprints observed here. Furthermore, the observation that the fluorescence in NMDG+ resembled that in K+ (although being significantly different for both mutants), could suggest that SERT in its cation-free state is biased towards an inward-facing conformation. Applying the same rationale for the spectra of V86Anap and F556Anap in Li+ provides insight into the SERT conformations promoted by Li+ binding. For V86Anap, the fluorescence pattern in Li+ was identical to that in NMDG+, suggesting Li+ to stabilize a similar conformational equilibrium with respect to the environment around TM1a. However, for F556Anap, Li+ produced a fluorescent response different from those observed in NMDG+, Na+, and K+ (Supplementary Fig. 3c, d). The competitive nature of K+ and Li+, with respect to Na+, could suggest that K+ and Li+ compete with Na+ for overlapping binding sites. This aligns with a cryo-EM structure of the homologous NET, where a K+ ion was proposed to occupy the Na1 site57. Interestingly, while the binding of Na+ and K+ appeared to be mutually exclusive, the observation that Li+ interfered with the Na+ cooperativity could be accounted for if Li+ and Na+ bound at the same time, as previously proposed for the homologous DAT48 and GABA transporter58.

The exchange of ions seemed to cause large environmental changes around Anap, likely as a result of SERT exploiting the full conformational spectrum10. In contrast, those induced by the binding of inhibitors or substrates were less pronounced and likely to be dominated by the bound Na+ ion. Interestingly, whenever spectral changes were observed upon the binding of inhibitors, these were unlike those seen with K+. Consequently, the changes were difficult to interpret in terms of transitions between the outward- and inward-facing states. Given the distance from the S1 site to position 86 and 556, and the observation that S1-site mutations abolished the signal changes, any contributions from direct interactions between Anap and the ligands should be negligible. Accordingly, the observation that e.g., paroxetine induced changes only for V86Anap might result from local conformational changes confined to the area around TM1a. This could suggest that certain binding modes are associated with an allosteric uncoupling between the intracellular- and extracellular sites.

Despite the absence of an ion gradient, we showed that 5-HT and MDMA promoted shifts in the fluorescence differing from those of the inhibitors. As for paroxetine binding, the changes observed for V86Anap in the presence of substrate were indicative of changes around TM1a that were not evident from the corresponding crystal and cryo-EM structures6,9. For F556Anap, the decrease in fluorescence by the binding of 5-HT seemed unlikely to originate from quenching of Anap by the indole group of 5-HT59 since the conformational changes following 5-HT binding were relatively slow, and since the response with MDMA (lacking the indole group) was similar. This observation, along with the inhibition by paroxetine and estimated 5-HT EC50 value, suggested that the 5-HT response was conformationally mediated through S1 site binding.

The expression and purification of SERT with Anap site-specifically incorporated represents a system offering several advantages in the study of SERT and related transporters. First, relying on fluorescence as a read-out for protein conformation offers a label-free approach to detect the steady-state and real-time interactions with ions, ligands, and potentially interacting proteins. Second, the nature of genetic code expansion makes all positions in the transporter candidates for Anap incorporation, including those buried in the transporter structure that are inaccessible to chemical conjugation. Accordingly, appropriate positioning of Anap can add layers of conformational- and allosteric information. Furthermore, the suspension-scale expression system implemented here constitutes a platform that can be easily transferred to other targets. Since changes in the spectral properties of Anap alone do not report on the extent or nature of the underlying conformational rearrangements, Anap can be combined with FRET acceptors in the form of immobilized transition metals60 or fluorescent proteins61 to quantify distance changes. Moreover, for studies of membrane proteins conducted over expanded timescales, or under the influence of ion gradients, protein can be reconstituted into nanodiscs or liposomes, respectively. We hope that the results and methodological possibilities presented here, many of which are unparalleled, will motivate further studies of this essential class of transporters.

All reagents were purchased from Merck unless otherwise stated.

The gene for full-length hSERT was encoded with a C-terminal thrombin cleavage site followed by a Twin-Strep tag and a 12HIS tag, each separated by GGS linkers. The construct, terminated on an ochre stop codon (TAA), was synthesized (Eurofins Genomics) and subcloned into a pcDNA3.1 vector using BssHII and XbaI (New England Biolabs) at the 5’ and 3’ ends, respectively. Amber stop codons (TAG) were introduced through site-directed mutagenesis by a linear amplification whole-plasmid polymerase chain reaction (PCR). The primers (Eurofins Genomics) for V86Anap, I172M, and F556Anap had the following sequences: V86Anap (5’ – GACCTGGGGCAAGAAGTAGGATTTCCTTCTCTCAGTGATTG - 3’); I172M (5’ – CTGCATCATTGCCTTTTACATGGCTTCCTACTACAACACCATC – 3’); F556Anap (5’ – CCTGTTCATCATTTGCAGTTAGCTGATGAGCCCGCC - 3’). Gene sequences were verified by DNA sequencing (Eurofins Genomics).

Adherent HEK293 cells (ThermoFisher Scientific; tested free of mycoplasma) were cultivated at 37°C, 5% CO2, and 100% humidity in Dulbecco’s Modified Eagle Medium (in house) supplemented with HEPES, 10% FBS (Gibco), 2 mM L-glutamine, and penicillin/streptomycin (in house). ~16 h prior to transfection, cells were seeded in poly-ornithine-coated 24-well plates in 0.5 ml per well. At ~50% confluency (~1.2 × 105 cells per well), cells were switched to medium without penicillin/streptomycin. Cells were transiently co-transfected with 442 ng DNA per well, pre-mixed in Opti-MEM (Gibco) with either 1.25 µl Lipofectamine 2000 (Invitrogen) or 1.33 µl polyethyleneimine HCl (linear, avg. 20 kDa). A 2:2:1 plasmid ratio of SERT: pANAP (tRNA-aaRSAnap; Addgene #48696): eRF1-E55D (Addgene #130876), respectively, was used. For control transfections without pANAP and/or eRF1-E55D, DNA concentrations were maintained with an empty pcDNA3.1 plasmid. Following 6 h incubation, the transfection medium devoid of penicillin/streptomycin was replaced by complete medium. 0, 15, or 25 µM Anap was added to the medium during and after transfection (from 5 mM stocks prepared in 12 mM NaOH). Cells were incubated for 48 h shielded from light.

Uptake assays were conducted at RT using transiently transfected HEK293 cells. Cells were washed twice with buffer (20 mM HEPES (pH 7.4), 130 mM NaCl, 5.4 mM KCl, 1.2 mM CaCl2, 1.2 MgSO4, 1 mM ascorbic acid, 5 mM D-glucose). Uptake was initiated by the rapid application of 12.3 nM 5-[1,2-3H]hydroxytryptamine ([3H]5-HT) (40.7 Ci mmol−1; PerkinElmer). Following 2 min (mutant screen and unspecific read-through controls) or 2-4 min (5–HT uptake inhibition curves) incubation with gentle shaking, the uptake reaction was quenched by washing cells twice in ice-cold buffer. For homologous uptake inhibition curves, increasing concentrations of unlabeled 5-HT (10−9 to 10−3 M) were added in parallel with the [3H]5-HT. A control (no unlabeled 5-HT) was included. The nonspecific uptake was obtained by the addition of 10 µM paroxetine. Cells were lysed in 1% SDS, transferred to counting plates, and mixed with Opti-phase Hi Safe 3 scintillation fluid (PerkinElmer). Plates were counted for 3 min in MicroBeta2 2450 Microplate Counter (PerkinElmer). All experiments were repeated at least 3 times. Substrate depletion was avoided by adjusting the assay volume and uptake duration to ensure that the specific counts corresponded to < 10% of the total [3H]5-HT.

For suspension-scale transfection, we used Expi293F suspension cells (ThermoFisher Scientific; tested free of mycoplasma). These were cultivated in HE400AZ medium (GMEP Cell technologies) at 37 °C, 8% CO2, 80% humidity, and 110 r.p.m. At a cell density of ~4 mio. ml−1, cells were transiently co-transfected with 2 µg DNA and 6 µg PEI pr. ml culture. DNA and PEI were mixed and incubated for 10 min in Opti-MEM equivalent to 1/10 of the culture volume. A 2:2:1 plasmid ratio of SERT (WT, V86Anap, or F556Anap), pANAP, and eRF1-E55D, respectively, was used. 15 µM Anap was added directly to the medium prior to transfection. For negative controls, SERT was expressed in the absence of Anap. Following incubation for 72 hours shielded from light, cells were harvested at 6000 g and 4 °C for 20 min. Cells were washed once in ice-cold PBS and frozen at −80 °C. Thawed cells were resuspended in 30 mM HEPES (pH 8), 30 mM NaCl, 5 mM KCl, 10% (w/v) sucrose, and 7 mM MgCl2 (supplemented with 1 mM EDTA, 2 µg ml−1 DNAse, 2 µg ml−1 RNAse, 500 µM Tris(2-carboxyethyl)phosphine (TCEP), 10 µg ml−1 Leupeptin (ThermoFisher Scientific), 10 µg ml−1 benzamidine, and protease inhibitor cocktail diluted 1000-fold). Cells were disrupted by cycles of probe sonication pulses and crude membranes were isolated by successive centrifugations at 1000 g for 5 min and 125,171 g for 1 hour, respectively. Membranes were re-suspended in the same buffer (devoid of EDTA, MgCl2, and DNAse/RNAse), frozen in liquid nitrogen, and stored at −80°C. Thawed membranes were supplemented with 20 mM Tris-HCl (pH 8), 150 mM NaCl, 10 % (v/v) glycerol), 1 % (w/v) n-dodecyl-β-D-maltopyranoside (DDM) (Anatrace), and 0.2 % (w/v) cholesterol hemisuccinate (CHS) at 0.1 g ml−1 for 1.5 h at 4°C in the dark. Un-solubilized material was pelletized at 125,171 g for 1 h at 4°C and the supernatant was filtered through a 0.45 µm polyether sulfone filter (VWR) and passed over 1 ml Ni2+-NTA (HisTrapTM HP; Cytiva) in the presence of 25 mM imidazole. The resin was washed with 11 column volumes of buffer containing 25 mM imidazole. Immobilized SERT was eluted by a linear imidazole gradient in 20 mM Tris-HCl (pH 8), 300 mM NaCl, 10% (v/v) glycerol, 1 mM (0.05% (w/v)) DDM, 0.2 mM (0.01% (w/v)) CHS, 24 µM lipids (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG); molar ratio 1:1:1), 500 µM TCEP, 10 µg ml−1 Leupeptin, 10 µg ml−1 benzamidine, and protease inhibitor cocktail diluted 1000-fold. Eluted SERT was concentrated on a Vivaspin20 (50 kDa MWCO; Satorius) and further purified by size-exclusion chromatography (Superdex 200; Cytiva) in the same buffer devoid of imidazole. Top fractions were pooled, concentrated ~7-fold on a Vivaspin6 (50 kDa MWCO; Satorius), frozen in liquid nitrogen, and stored at −80 °C.

SERT WT, V86Anap, and F556Anap, expressed in the presence or absence of Anap, were incubated overnight with or without PNGase F (New England Biolabs). Samples were analyzed by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) in hand-casted 10% acrylamide gels. Equivalent amounts of SERT sample, relative to the total volume of purified sample, were loaded per well. For in-gel fluorescence imaging, the gel was imaged using a fluorescence filter with 365 nm excitation and Cy3 emission. For immunoblotting, the gel was transferred to a PVDF membrane (Bio-Rad) and blocked in skim milk powder (suspended in PBS with 0.05% (v/v) TWEEN-20) for 1 h. The membrane was mixed with hSERT antibody (ST51-2; MAb Technologis) and incubated overnight. The membrane was washed for 3 × 5 min in PBS with 0.05% (v/v) TWEEN-20 and incubated for 1 h with goat anti-Mouse secondary antibody conjugated with horseradish peroxidase (HRP) (ThermoFisher). Following another 3 × 5 min washes, the membrane was imaged upon application of HRP substrate (Amersham). An ImageQuant 800 (Cytiva) was used for imaging of the fluorescent gel and western blot. In-gel SDS-PAGE analyses and western blotting were repeated at least 3 times.

Equilibrium radioligand binding assays were performed with purified or membrane-embedded SERT. The Scintillation Proximity Assay (SPA) was used for purified SERT. Thawed SERT aliquots were centrifuged at 15,000 g and 4°C for 10 min. In a 96-well white, flat, and clear-bottom plate (Greiner Bio-One), SERT was immobilized to Yttrium Silicate Copper (YSi-Cu) His-tag SPA beads (PerkinElmer) in 20 mM Tris-Cl (pH 8), 10% (v/v) glycerol, 200 mM NaCl, 500 µM TCEP, 1 mM DDM, 0.2 mM CHS, and 24 µM lipids (POPC, POPE, POPG; molar ratio 1:1:1). For S-citalopram saturation experiments, 0.175 µg ml-1 (2.5 nM) SERT was mixed with 0.8 mg ml−1 YSi-Cu SPA beads and increasing concentrations of [N-Methyl-3H]-S-citalopram ([3H]S-citalopram) (81 Ci mmol−1; PerkinElmer). For negative controls, purified Anap variants expressed in the absence of Anap were used in amounts equivalent to those yielding 2.5 nM of SERT expressed in the presence of Anap. For time-dependent stability measurements, SERT was incubated in buffer containing 200 mM NaCl at RT. At 0, 30, and 60 min, 20 nM of SERT was mixed with 1 mg ml−1 YSi-Cu SPA beads and 50 nM [3H]S-citalopram (81 Ci mmol−1). For S-citalopram affinity and stability experiments, nonspecific binding was determined by pre-incubation with 100 µM paroxetine. For Na+-dependent [3H]S-citalopram binding on purified protein, 5 nM SERT, 10 nM [3H]S-citalopram, and 0.8 mg ml−1 YSi-Cu SPA beads were mixed in buffer with increasing concentrations of Na+ in the presence of 150 mM NMDG+, Li+, or K+. The ionic strength was maintained using NMDG+ as a counter ion and chloride was used as the corresponding anion. As control for the compound screen, 5 nM SERT WT was mixed in Na+-containing buffer, supplemented with 1 mM ascorbic acid, with the following compounds in concentrations corresponding to ~20-50xKd: 500 nM fluoxetine, 50 nM paroxetine, 100 nM S-citalopram, 1.5 µM imipramine, 20 µM cocaine, 20 µM noribogaine, 50 µM MDMA, and 12 µM 5-HT ± 1 µM paroxetine. SPA plates were sealed and incubated for ~15 min at RT and 600 r.p.m. shaking followed by overnight incubation at 4 °C. For Na+-dependent [3H]S-citalopram binding on SERT in native membranes, Expi293F cells were transfected with 2 µg SERT WT and 6 µg PEI pr. ml culture as described under ‘SERT expression and purification’. Membranes were prepared as for purification except for the implementation of an extra homogenization and ultracentrifugation step to increase the purity. In a transparent, round-bottom 96-well plate (Corning), 10 nM [3H]S-citalopram and membranes corresponding to 100 µg pr. well were mixed in 20 mM Tris-Cl (pH 8), 10% (v/v) glycerol, 500 µM TCEP, and protease inhibitor cocktail diluted 1000-fold. Ionic conditions identical to those used for the equivalent experiment with purified SERT were applied. Following 1 min incubation at intense shaking, the plate was bath sonicated for 15 sec to equilibrate potential ion gradients across vesicle membranes. The plate was incubated for 30 min at shaking after which the membranes were immobilized to a PEI-soaked filter (PerkinElmer) and washed with 0.5 L ice-cold PBS (in house) using a Tomtec cell harvester. The filter was dried and combined with Meltilex (PerkinElmer). Counts per minute (c.p.m.) for SPA and membrane binding were recorded on a MicroBeta2 2450 Microplate Counter (PerkinElmer) for 1 and 3 min, respectively. All experiments were repeated at least 3 times using protein from ≥ 2 different expressions and purifications or membranes preparations.

Liposomes with SERT were prepared essentially as described previously for Drosophila melanogaster DAT (dDAT)17. Briefly, soy polar lipid extract, cholesterol, and brain polar lipid extract (Avanti) were mixed in a 60:17:20 molar ratio, and suspended in 20 mM HEPES (pH 7.5) and 150 mM KC2H3O2 (internal buffer) to 10 mg ml−1. The suspended lipids were subjected to 5 rounds of freeze-thawing and extruded through a 0.4 µm Nucleopore Whatman Track-Etched polycarbonate filter. Liposomes were diluted to 4 mg ml−1 before the addition of SERT in a 1:400 (WT, V86Anap) to 1:600 (F556Anap) protein:lipid ratio. Empty liposomes were prepared in parallel. Following incubation for 1 h at 4 °C, three rounds of 83 mg Bio-Beads SM-2 were added with 1 h intervals before proteoliposomes were incubated overnight at 4 °C. The following day, 83 mg Bio-Beads SM-2 were added and proteoliposomes were incubated for 2 h. Bio-Beads were removed by filtration and proteoliposomes were pelletized by centrifugation at 140,000 g for 1 h at 4 °C. Proteoliposomes were suspended in internal buffer to 10 mg ml−1 and frozen in liquid nitrogen.

Proteoliposomes were thawed on ice, extruded through a 0.4 µm Nucleopore Whatman Track-Etched polycarbonate filter, and aliquoted into a round-bottom 96-well plate. The uptake assay was performed at room temperature. At different time-points, 50 nM [3H]5-HT (36 Ci mmol−1; PerkinElmer), diluted in 20 mM HEPES (pH 7.5), 150 mM NaCl, and 0.5 mM ascorbic acid, was added to the wells in a 20-fold volume excess. Nonspecific uptake was determined by performing the uptake experiments on empty liposomes in parallel. Using a Tomtec cell harvester, proteoliposomes were trapped onto a filter (PerkinElmer) soaked in PEI, and washed with 0.5 L ice-cold buffer (20 mM HEPES (pH 7.5) and 150 mM NaCl). Filters were dried and Meltilex (PerkinElmer) was melted into the filter. Counts per minute were recorded on a MicroBeta2 2450 Microplate Counter (PerkinElmer) for 1 min. The experiments were performed in three independent biological replicates.

Steady-state fluorescence intensities were recorded as the ratio between the corrected signal and corrected reference (S1c/R1c) on a FluoroMax-4 spectrofluorometer (HORIBA Scientific). 0.1 sec integration time was used. Excitation- and emission monochromators were calibrated prior to every experiment. Spectra were obtained in a 5x5 mm quartz cuvette (Hellma) with 1 nm increments. For measurements of the free Anap environmental sensitivity, Anap was diluted to 50 nM in H2O, dimethyl sulfoxide (DMSO), ethyl acetate (EtOAc), or buffer (20 mM Tris-Cl (pH 8), 10% (v/v) glycerol, 500 µM TCEP, 1 mM DDM, 0.2 mM CHS, 24 µM lipids (POPC, POPE, POPG; molar ratio 1:1:1), supplemented with 200 mM NaCl. For excitation spectra, the fluorescence intensity at 490 nm was recorded upon excitation at 270-454 nm. For emission spectra, samples were excited at 360 nm and the fluorescence intensity was measured at 375−660 nm. 4 nm excitation- and emission slit widths were used. Experiments with purified SERT were assayed using 7 nm slit widths. SERT was thawed on ice in the dark, centrifuged at 15,000 g and 4°C for 10 min. Protein was diluted to ~7.5 nM and ~12.5 nM for the F556Anap and V86Anap constructs, respectively, in 20 mM Tris-Cl (pH 8), 10% (v/v) glycerol, 500 µM TCEP, 1 mM DDM, 0.2 mM CHS, 24 µM lipids (POPC, POPE, POPG; molar ratio 1:1:1), supplemented with the ion/ligand combinations indicated. Samples were incubated for 10–30 min shielded from light before emission spectra at 375−550 nm were recorded upon excitation at 360 nm. For excitation spectra, the emission at 450 nm was recorded upon excitation at 270−435 nm. To correct for contributions from the Raman peak and fluorescent/scattering properties of the buffers and/or ligand components, emission spectra for each ion/ligand combination without protein were recorded in parallel. For denaturation experiments, SERT was pre-incubated for 30 min in buffer heated to 60 °C or in buffer supplemented with 5 M guanidinium hydrochloride. For cation screens, SERT was incubated with either 200 mM NMDG+, Li+, K+, or Na+, or with increasing Na+. The ionic strength was maintained at all times with NMDG+. Cl- was used as the corresponding anion in all ionic compositions. For the anion screen, Cl- was substituted for acetate, using Na+ as the corresponding cation. For the screen of different compounds, SERT was incubated with the compounds and concentrations specified under ‘[3H]S-citalopram binding’ in buffer containing 200 mM Na+ and 1 mM ascorbic acid. The compound quenching controls were conducted with 5 nM free Anap in the same buffer. All experiments were repeated at least 3 times with protein from different purifications.

For kinetic experiments, an SX20 stopped-flow system (AppliedPhotophysics) was used. Fluorescence traces were obtained at 500 V, 0.01 sec time increments, and 360 nm excitation. A 495 nm long-pass filter was used and reactions were measured at 20 °C. Protein was prepared as described for the steady-state experiments. For F556Anap, ~11 nM was used throughout. For V86Anap, ~30 and ~50 nM were used for the baseline/association and dissociation, respectively. For Na+ kinetics, a baseline (10 sec) was acquired by mixing SERT, diluted in the absence of Na+, with NMDG+ buffer. Next, the Na+ association (40 sec) was obtained by mixing SERT with buffer containing 100 mM Na+. The Na+ dissociation (50 sec) was acquired by diluting SERT, pre-mixed with 18.75 (V86Anap) or 37.5 (F556Anap) mM Na+, 2-fold into a Na+-free buffer. NMDG+ was used to preserve the ionic strength throughout. For 5-HT kinetics using F556Anap, all measurements were conducted in buffer containing 200 mM NaCl and 1 mM ascorbic acid. A baseline (10 sec) was obtained by mixing F556Anap with buffer. The 5-HT association (10 sec) was acquired by mixing F556Anap with 12 µM 5-HT, while the dissociation (40 sec) was acquired by diluting F556Anap (pre-mixed with 12 µM) 2-fold into buffer devoid of 5-HT and supplemented with 100 µM paroxetine. Control traces were made with F556Anap to which only paroxetine was added (without 5-HT). Experiments were repeated at least 3 times with protein from separate purifications.

All data points are mean ± S.E.M., either shown as error bars or as error envelopes (fluorescence spectra or kinetic traces). Hill coefficients, S-citalopram affinities, λmax- values, differences in fluorescence intensities, and relative uptake/binding are reported as mean ± S.E.M. Affinities, EC50, and IC50 values determined by analysis in logarithmic dimensions are reported as mean [S.E.M. interval]. Statistical analysis involving more than two means were carried out using a one-way analysis of variance (ANOVA) with either Bonferroni, Tukey, or Dunnett tests to correct for multiple comparisons. All experiments were repeated ≥ 3 times using protein or membrane preparations from at least 2 independent expressions and purifications.

[3H]5-HT uptake data for the mutant screen or unspecific read-through controls were corrected for nonspecific uptake (+paroxetine) and normalized to the c.p.m. obtained for SERT WT expressed in the absence of Anap, tRNA-aaRSAnap, and eRF1-E55D. For the expression using PEI in 15 or 25 µM Anap, data were normalized to the c.p.m. obtained for SERT WT expressed with 15 µM Anap, tRNA-aaRSAnap, and eRF1-E55D. For homologous 5-HT inhibition of [3H]5-HT uptake, c.p.m. were normalized to those obtained in the absence of unlabeled 5-HT for each individual SERT variant, while the bottom was normalized to 0. Data were modeled by a non-linear fit restraining the Hill coefficient to 1, from which IC50 values for 5-HT were predicted. For proteoliposome uptake experiments, the counts for nonspecific uptake (empty liposomes) were subtracted from those for liposomes with SERT. C.p.m. were converted to pmol substrate per mg SERT and fitted by a one-phase association model. For [3H]S-citalopram saturation binding, data were corrected for nonspecific binding (+paroxetine) and normalized to the Bmax predicted by a one-site binding model, from which Kd values were provided. The data for V86Anap or F556Anap, expressed in the absence of Anap, were normalized to the Bmax predicted for the same mutant expressed in the presence of Anap. For stability measurements, c.p.m. were corrected for nonspecific binding (+paroxetine) and normalized to the c.p.m. obtained for each mutant at time zero. For Na+-dependent [3H]S-citalopram binding, data were normalized to the Bmax predicted by the Hill equation:

Where ϴ denotes the fraction of bound ligand, Bmin and Bmax represent the lower and upper plateau, respectively, EC50 is the concentration of ligand at which the 50 % is bound, [ligand] is the ligand concentration, and n is the Hill coefficient. The resulting EC50/IC50 values were used to compute the Ki values for Li+, and K+, using the Cheng-Prusoff equation:

Where Ki is the potency of the inhibitor, IC50 represents the concentration of inhibitor at which the ligand binding is inhibited by 50 %, and Kd is the affinity of the ligand.

Steady-state fluorescence spectra with protein or free Anap were corrected for the 1st order Raman peak (410 nm) and fluorescence/scattering from the buffer/ligand components by the subtraction of spectra obtained in the absence of protein or free Anap. For emission spectra of free Anap, fluorescence intensities (S1c/R1c) from the corrected spectra were normalized to those obtained at 490 nm in H2O or, for the compound quenching controls, in Na+ buffer. Excitation spectra of free and incorporated Anap were normalized to the fluorescence intensity upon excitation at 360 nm. Emission spectra of V86Anap and F556Anap were normalized to the intensity obtained at λmax in Na+, corresponding to 425 and 448 nm for V86Anap and F556Anap, respectively. For the compound screen, ΔFluorescence intensity was computed by normalizing the fluorescence intensities to that at λmax obtained in the absence of ligands. The spectral ratios for the spectra of V86Anap were computed by taking the ratio of the average fluorescence intensity between 405–425 nm to the average fluorescence intensity between 426–476 nm. The spectral ratios for the spectra of F556Anap were calculated by taking the ratio of the average fluorescence intensity between 423–448 nm to the average fluorescence intensity between 449–494 nm. The same fluorescence intervals for V86Anap and F556Anap were used for all fluorescence data. To correct for mechanical drifts, spectral ratios were normalized to that obtained in Na+ alone. For Na+ and 5-HT titrations, the normalized spectral ratios were plotted as a function of ligand concentration and fitted to the Hill regression as described for radioligand binding. For the ion titration in the presence or absence of Li+ and K+, the apparent Ki values were estimated as described for Na+ -dependent [3H]S-citalopram binding. For stopped-flow fluorometric measurements, the traces for the baseline, association, and dissociation of 5-HT or Na+ were plotted as ΔFluorescence of the raw data, normalized to the first (association) or last (dissociation) data point. For the dissociation of 5-HT, the trace obtained in the absence of 5-HT was subtracted from that obtained in the presence of 5-HT. The data for the association of Na+ or 5-HT, and those for the dissociation of Na+, were fitted to a one-phase exponential model. The dissociation of 5-HT was modeled by a two-phase exponential function. For the 5-HT kinetics, the half-life (T½; the time (sec) at which the response is 50% of the total) was converted to a Kon:

The half-life for the dissociation of 5-HT was converted to Koff:

The 5-HT affinity was computed as:

Numerical analysis and modeling of data were conducted using GraphPad Prism 9. The multiple sequence alignment was constructed based on the Clustal Omega algorithm62 and graphically represented using ESPript 3.063. Pymol was applied for the cartoon representation and superimposition of SERT in the outward-open paroxetine-bound (PDB. 5I6X) and inward-open ibogaine-bound (PDB. 6DZZ) state. For the in-gel and immunoblot image processing, ImageJ was used. The cartoons illustrating the principles underlying Anap incorporation or the kinetic mechanisms were created with BioRender.com. The chemical structure of Anap was made with ChemDraw.

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

The data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper.

Cheng, M. H. & Bahar, I. Monoamine transporters: structure, intrinsic dynamics and allosteric regulation. Nat. Struct. Mol. Biol. 26, 545–556 (2019).

Article CAS PubMed PubMed Central Google Scholar

Bowman, M. A. & Daws, L. C. Targeting serotonin transporters in the treatment of juvenile and adolescent depression. Front Neurosci. 13, 156 (2019).

Article PubMed PubMed Central Google Scholar

Kristensen, A. S. et al. SLC6 neurotransmitter transporters: structure, function, and regulation. Pharm. Rev. 63, 585–640 (2011).

Article CAS PubMed Google Scholar

Rudnick, G. & Sandtner, W. Serotonin transport in the 21st century. J. Gen. Physiol. 151, 1248–1264 (2019).

Article CAS PubMed PubMed Central Google Scholar

Nielsen, J. C. et al. Structure of the human dopamine transporter in complex with cocaine. Nature 632, 678–685 (2024).

Article CAS PubMed Google Scholar

Coleman, J. A., Green, E. M. & Gouaux, E. X-ray structures and mechanism of the human serotonin transporter. Nature 532, 334–339 (2016).

Article ADS CAS PubMed PubMed Central Google Scholar

Hu, T. et al. Transport and inhibition mechanisms of the human noradrenaline transporter. Nature 632, 930–937 (2024).

Yamashita, A., Singh, S. K., Kawate, T., Jin, Y. & Gouaux, E. Crystal structure of a bacterial homologue of Na + /Cl–dependent neurotransmitter transporters. Nature 437, 215–223 (2005).

Article ADS CAS PubMed Google Scholar

Yang, D. & Gouaux, E. Illumination of serotonin transporter mechanism and role of the allosteric site. Sci. Adv. 7, eabl3857 (2021).

Article ADS CAS PubMed PubMed Central Google Scholar

Coleman, J. A. et al. Serotonin transporter-ibogaine complexes illuminate mechanisms of inhibition and transport. Nature 569, 141–145 (2019).

Article ADS CAS PubMed PubMed Central Google Scholar

Merkle, P. S. et al. Substrate-modulated unwinding of transmembrane helices in the NSS transporter LeuT. Sci. Adv. 4, eaar6179 (2018).

Article ADS PubMed PubMed Central Google Scholar

Nelson, P. J. & Rudnick, G. Coupling between platelet 5-hydroxytryptamine and potassium transport. J. Biol. Chem. 254, 10084–10089 (1979).

Article CAS PubMed Google Scholar

Ni, Y. G. et al. A lithium-induced conformational change in serotonin transporter alters cocaine binding, ion conductance, and reactivity of Cys-109. J. Biol. Chem. 276, 30942–30947 (2001).

Article CAS PubMed Google Scholar

Rudnick, G. & Nelson, P. J. Platelet 5-hydroxytryptamine transport, an electroneutral mechanism coupled to potassium. Biochemistry 17, 4739–4742 (1978).

Article CAS PubMed Google Scholar

Schicker, K. et al. Unifying concept of serotonin transporter-associated currents. J. Biol. Chem. 287, 438–445 (2012).

Article CAS PubMed Google Scholar

Billesbolle, C. B. et al. Transition metal ion FRET uncovers K(+) regulation of a neurotransmitter/sodium symporter. Nat. Commun. 7, 12755 (2016).

Article ADS PubMed PubMed Central Google Scholar

Schmidt, S. G. et al. The dopamine transporter antiports potassium to increase the uptake of dopamine. Nat. Commun. 13, 2446 (2022).

Article ADS CAS PubMed PubMed Central Google Scholar

Schmidt, S. G., Nygaard, A., Mindell, J. A. & Loland, C. J. Exploring the K+ binding site and its coupling to transport in the neurotransmitter: sodium symporter LeuT. (eLife Sciences Publications, Ltd, 2023).

Bhat, S. et al. Handling of intracellular K(+) determines voltage dependence of plasmalemmal monoamine transporter function. Elife 10, e67996 (2021).

Article CAS PubMed PubMed Central Google Scholar

Williams, R. S. & Harwood, A. J. Lithium therapy and signal transduction. Trends Pharm. Sci. 21, 61–64 (2000).

Article CAS PubMed Google Scholar

Mager, S. et al. Conducting states of a mammalian serotonin transporter. Neuron 12, 845–859 (1994).

Article ADS CAS PubMed Google Scholar

Lin, F., Lester, H. A. & Mager, S. Single-channel currents produced by the serotonin transporter and analysis of a mutation affecting ion permeation. Biophysical J. 71, 3126–3135 (1996).

Article ADS CAS Google Scholar

Moller, I. R. et al. Conformational dynamics of the human serotonin transporter during substrate and drug binding. Nat. Commun. 10, 1687 (2019).

Article ADS PubMed PubMed Central Google Scholar

Gunnoo, S. B. & Madder, A. Chemical protein modification through cysteine. Chembiochem 17, 529–553 (2016).

Article CAS PubMed Google Scholar

Chen, J.-G., Liu-Chen, S. & Rudnick, G. External cysteine residues in the serotonin transporter. Biochemistry 36, 1479–1486 (1997).

Article CAS PubMed Google Scholar

Chin, J. W. Expanding and reprogramming the genetic code. Nature 550, 53–60 (2017).

Article ADS CAS PubMed Google Scholar

Lee, H. S., Guo, J., Lemke, E. A., Dimla, R. D. & Schultz, P. G. Genetic incorporation of a small, environmentally sensitive, fluorescent probe into proteins in Saccharomyces cerevisiae. J. Am. Chem. Soc. 131, 12921–12923 (2009).

Article CAS PubMed PubMed Central Google Scholar

Chatterjee, A., Guo, J., Lee, H. S. & Schultz, P. G. A genetically encoded fluorescent probe in mammalian cells. J. Am. Chem. Soc. 135, 12540–12543 (2013).

Article CAS PubMed PubMed Central Google Scholar

Noren, C. J., Anthony-Cahill, S. J., Griffith, M. C. & Schultz, P. G. A general method for site-specific incorporation of unnatural amino acids into proteins. Science 244, 182–188 (1989).

Article ADS CAS PubMed Google Scholar

Puljung, M. C. ANAP: A versatile, fluorescent probe of ion channel gating and regulation. Methods Enzymol. 654, 49–84 (2021).

Article CAS PubMed Google Scholar

Huang, Y. Z. et al. TRPV1 analgesics disturb core body temperature via a biased allosteric mechanism involving conformations distinct from that for nociception. Neuron 112, 1815–1831 e4 (2024).

Article CAS PubMed Google Scholar

Soh, M. S., Estrada-Mondragon, A., Durisic, N., Keramidas, A. & Lynch, J. W. Probing the structural mechanism of partial agonism in glycine receptors using the fluorescent artificial amino acid, ANAP. ACS Chem. Biol. 12, 805–813 (2017).

Article CAS PubMed Google Scholar

Dabrowski, M., Bukowy-Bieryllo, Z. & Zietkiewicz, E. Translational readthrough potential of natural termination codons in eucaryotes–The impact of RNA sequence. RNA Biol. 12, 950–958 (2015).

Article PubMed PubMed Central Google Scholar

Bartoschek, M. D. et al. Identification of permissive amber suppression sites for efficient non-canonical amino acid incorporation in mammalian cells. Nucleic Acids Res 49, e62 (2021).

Article CAS PubMed PubMed Central Google Scholar

Kalstrup, T. & Blunck, R. Reinitiation at non-canonical start codons leads to leak expression when incorporating unnatural amino acids. Sci. Rep. 5, 11866 (2015).

Article ADS PubMed PubMed Central Google Scholar

Kalstrup, T. & Blunck, R. Dynamics of internal pore opening in K(V) channels probed by a fluorescent unnatural amino acid. Proc. Natl Acad. Sci. USA 110, 8272–8277 (2013).

Article ADS CAS PubMed PubMed Central Google Scholar

Aman, T. K., Gordon, S. E. & Zagotta, W. N. Regulation of CNGA1 Channel Gating by Interactions with the Membrane. J. Biol. Chem. 291, 9939–9947 (2016).

Article CAS PubMed PubMed Central Google Scholar

Zagotta, W. N., Gordon, M. T., Senning, E. N., Munari, M. A. & Gordon, S. E. Measuring distances between TRPV1 and the plasma membrane using a noncanonical amino acid and transition metal ion FRET. J. Gen. Physiol. 147, 201–216 (2016).

Article CAS PubMed PubMed Central Google Scholar

Liu, W., Brock, A., Chen, S., Chen, S. & Schultz, P. G. Genetic incorporation of unnatural amino acids into proteins in mammalian cells. Nat. Methods 4, 239–244 (2007).

Article CAS PubMed Google Scholar

Schmied, W. H., Elsasser, S. J., Uttamapinant, C. & Chin, J. W. Efficient multisite unnatural amino acid incorporation in mammalian cells via optimized pyrrolysyl tRNA synthetase/tRNA expression and engineered eRF1. J. Am. Chem. Soc. 136, 15577–15583 (2014).

Article CAS PubMed PubMed Central Google Scholar

Harder, D. & Fotiadis, D. Measuring substrate binding and affinity of purified membrane transport proteins using the scintillation proximity assay. Nat. Protoc. 7, 1569–1578 (2012).

Article CAS PubMed Google Scholar

El-Kasaby, A. et al. Mutations in the carboxyl-terminal SEC24 binding motif of the serotonin transporter impair folding of the transporter. J. Biol. Chem. 285, 39201–39210 (2010).

Article CAS PubMed PubMed Central Google Scholar

Tavoulari, S. et al. Two Na+ sites control conformational change in a neurotransmitter transporter homolog. J. Biol. Chem. 291, 1456–1471 (2016).

Article CAS PubMed Google Scholar

Langer, A. et al. A new spectral shift-based method to characterize molecular interactions. Assay. Drug Dev. Technol. 20, 83–94 (2022).

Article CAS PubMed PubMed Central Google Scholar

Larsen, M. B. et al. Dopamine transport by the serotonin transporter: a mechanistically distinct mode of substrate translocation. J. Neurosci. 31, 6605–6615 (2011).

Article CAS PubMed PubMed Central Google Scholar

Felts, B. et al. The two Na+ sites in the human serotonin transporter play distinct roles in the ion coupling and electrogenicity of transport. J. Biol. Chem. 289, 1825–1840 (2014).

Article CAS PubMed Google Scholar

Shi, L., Quick, M., Zhao, Y., Weinstein, H. & Javitch, J. A. The mechanism of a neurotransmitter:sodium symporter–inward release of Na+ and substrate is triggered by substrate in a second binding site. Mol. Cell 30, 667–677 (2008).

Article CAS PubMed PubMed Central Google Scholar

Borre, L., Andreassen, T. F., Shi, L., Weinstein, H. & Gether, U. The second sodium site in the dopamine transporter controls cation permeation and is regulated by chloride. J. Biol. Chem. 289, 25764–25773 (2014).

Article CAS PubMed PubMed Central Google Scholar

Yang, D., Zhao, Z., Tajkhorshid, E. & Gouaux, E. Structures and membrane interactions of native serotonin transporter in complexes with psychostimulants. Proc. Natl Acad. Sci. 120, e2304602120 (2023).

Article CAS PubMed PubMed Central Google Scholar

Jacobs, M. T., Zhang, Y. W., Campbell, S. D. & Rudnick, G. Ibogaine, a noncompetitive inhibitor of serotonin transport, acts by stabilizing the cytoplasm-facing state of the transporter. J. Biol. Chem. 282, 29441–29447 (2007).

Article CAS PubMed Google Scholar

Rudnick, G. & Wall, S. C. The molecular mechanism of “ecstasy” [3,4-methylenedioxy-methamphetamine (MDMA)]: serotonin transporters are targets for MDMA-induced serotonin release. Proc. Natl. Acad. Sci. USA 89, 1817–1821 (1992).

Henry, L. K. et al. Tyr-95 and Ile-172 in transmembrane segments 1 and 3 of human serotonin transporters interact to establish high affinity recognition of antidepressants*. J. Biol. Chem. 281, 2012–2023 (2006).

Article CAS PubMed Google Scholar

Krout, D. et al. Inhibition of the serotonin transporter is altered by metabolites of selective serotonin and norepinephrine reuptake inhibitors and represents a caution to acute or chronic treatment paradigms. ACS Chem. Neurosci. 8, 1011–1018 (2017).

Article CAS PubMed Google Scholar

Kilic, F., Murphy, D. L. & Rudnick, G. A human serotonin transporter mutation causes constitutive activation of transport activity. Mol. Pharmacol. 64, 440–446 (2003).

Article CAS PubMed Google Scholar

Burtscher, V., Schicker, K., Freissmuth, M. & Sandtner, W. Kinetic models of secondary active transporters. Int. J. Mol. Sci. 20, 5365 (2019).

Szollosi, D. & Stockner, T. Investigating the mechanism of sodium binding to SERT using direct simulations. Front Cell Neurosci. 15, 673782 (2021).

Article PubMed PubMed Central Google Scholar

Tan, J. et al. Molecular basis of human noradrenaline transporter reuptake and inhibition. Nature 632, 921–929 (2024).

Zhou, Y., Zomot, E. & Kanner, B. I. Identification of a lithium interaction site in the gamma-aminobutyric acid (GABA) transporter GAT-1. J. Biol. Chem. 281, 22092–22099 (2006).

Article CAS PubMed Google Scholar

Pospisil, P. et al. Fluorescence quenching of (dimethylamino)naphthalene dyes Badan and Prodan by tryptophan in cytochromes P450 and micelles. J. Phys. Chem. B 118, 10085–10091 (2014).

Article CAS PubMed PubMed Central Google Scholar

Dai, G., Aman, T. K., DiMaio, F. & Zagotta, W. N. The HCN channel voltage sensor undergoes a large downward motion during hyperpolarization. Nat. Struct. Mol. Biol. 26, 686–694 (2019).

Article CAS PubMed PubMed Central Google Scholar

Mitchell, A. L., Addy, P. S., Chin, M. A. & Chatterjee, A. A unique genetically encoded FRET pair in mammalian cells. Chembiochem 18, 511–514 (2017).

Article CAS PubMed Google Scholar

Madeira, F. et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic acids Res. 50, W276–W279 (2022).

Article CAS PubMed PubMed Central Google Scholar

Robert, X. & Gouet, P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).

Article CAS PubMed PubMed Central Google Scholar

Download references

We would like to thank Jonas H. Steffen for technical assistance with the stopped-flow equipment, generously provided for our use by Professor Michael J. Davies. We thank Sarah Bargmeyer for thorough revision of the manuscript. Anton Turaev is thanked for inspiring us to use what we refer to as spectral ratio. The financial support for this work was provided by the Novo Nordic Foundation (NNF22OC0079091 to C.J.L.), the Independent Research Fund Denmark (1030-00036B to C.J.L.), and the Carlsberg Foundation (CF20-0345 to C.J.L.).

Department of Neuroscience, University of Copenhagen, Copenhagen, Denmark

Andreas Nygaard & Claus J. Loland

Department of Drug Design and Pharmacology, University of Copenhagen, Copenhagen, Denmark

Linda G. Zachariassen, Kathrine S. Larsen & Anders S. Kristensen

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

The study was conceptualized and designed by A.N., C.J.L., and A.S.K. A.N. performed all experiments and analyzed all data with C.J.L. K.S.L. and L.G.Z. performed initial mutant screens. The manuscript was written by A.N. with support from C.J.L. and A.S.K. All authors commented on it.

Correspondence to Claus J. Loland.

The authors declare no competing interests.

Nature Communications thanks Baruch Kanner, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Nygaard, A., Zachariassen, L.G., Larsen, K.S. et al. Fluorescent non-canonical amino acid provides insight into the human serotonin transporter. Nat Commun 15, 9267 (2024). https://doi.org/10.1038/s41467-024-53584-9

Download citation

Received: 12 January 2024

Accepted: 17 October 2024

Published: 27 October 2024

DOI: https://doi.org/10.1038/s41467-024-53584-9

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative