A mechanism of uncompetitive inhibition of the serotonin transporter

  1. Shreyas Bhat
  2. Ali El-Kasaby
  3. Ameya Kasture
  4. Danila Boytsov
  5. Julian B Reichelt
  6. Thomas Hummel
  7. Sonja Sucic
  8. Christian Pifl
  9. Michael Freissmuth
  10. Walter Sandtner

  • Curated by eLife

    eLife logo

    eLife assessment

    This study presents the important finding of an unusual uncompetitive inhibitor (ECSI#6) of the serotonin transporter that removes the neurotransmitter serotonin from the synaptic cleft. Through careful and comprehensive analysis, the authors convincingly show that the molecule most likely binds to the inward-facing and K+-bound state and that it assists in folding and targeting the transporter. The work will be of interest to those engaged in biophysical analyses of the serotonin transporter, and colleagues developing pharmacological chaperoning strategies for transporters in general.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The serotonin transporter (SERT/SLC6A4) is arguably the most extensively studied solute carrier (SLC). During its eponymous action – that is, the retrieval of serotonin from the extracellular space – SERT undergoes a conformational cycle. Typical inhibitors (antidepressant drugs and cocaine), partial and full substrates (amphetamines and their derivatives), and atypical inhibitors (ibogaine analogues) bind preferentially to different states in this cycle. This results in competitive or non-competitive transport inhibition. Here, we explored the action of N -formyl-1,3-bis (3,4-methylenedioxyphenyl)-prop-2-yl-amine (ECSI#6) on SERT: inhibition of serotonin uptake by ECSI#6 was enhanced with increasing serotonin concentration. Conversely, the K M for serotonin was lowered by augmenting ECSI#6. ECSI#6 bound with low affinity to the outward-facing state of SERT but with increased affinity to a potassium-bound state. Electrophysiological recordings showed that ECSI#6 preferentially interacted with the inward-facing state. Kinetic modeling recapitulated the experimental data and verified that uncompetitive inhibition arose from preferential binding of ECSI#6 to the K + -bound, inward-facing conformation of SERT. This binding mode predicted a pharmacochaperoning action of ECSI#6, which was confirmed by examining its effect on the folding-deficient mutant SERT-PG 601,602 AA: preincubation of HEK293 cells with ECSI#6 restored export of SERT-PG 601,602 AA from the endoplasmic reticulum and substrate transport. Similarly, in transgenic flies, the administration of ECSI#6 promoted the delivery of SERT-PG 601,602 AA to the presynaptic specialization of serotonergic neurons. To the best of our knowledge, ECSI#6 is the first example of an uncompetitive SLC inhibitor. Pharmacochaperones endowed with the binding mode of ECSI#6 are attractive, because they can rescue misfolded transporters at concentrations, which cause modest transport inhibition.

Article activity feed

  1. Version published to 10.7554/elife.82641 on eLife
  2. eLife

    eLife assessment

    This study presents the important finding of an unusual uncompetitive inhibitor (ECSI#6) of the serotonin transporter that removes the neurotransmitter serotonin from the synaptic cleft. Through careful and comprehensive analysis, the authors convincingly show that the molecule most likely binds to the inward-facing and K+-bound state and that it assists in folding and targeting the transporter. The work will be of interest to those engaged in biophysical analyses of the serotonin transporter, and colleagues developing pharmacological chaperoning strategies for transporters in general.

  3. eLife

    Reviewer #1 (Public Review):

    This paper reports an analysis of the inhibition of the serotonin transporter, SERT, by a novel compound, ECSI#6. The authors perform a comprehensive analysis of SERT transport inhibition for the new agent and compare its properties to those of other well-characterized agents: cocaine and noribogaine, with the data pointing to an unusual noncompetitive mechanism of inhibition, a model also supported by electrophysiological recordings of transport currents. Based on the results of these experiments the authors conclude that ESCI#6 binds essentially exclusively to the inward-facing state of the transporter. The authors further present experiments suggesting that ESCI#6 can stabilize the folded form of an ER-arrested SERT mutant and recover its trafficking to the plasma membrane, with some in-vivo drosophila experiments perhaps also supporting this conclusion. Finally, kinetic simulations using a transport model with rate constants from previous experiments support the basic conclusions of the first sections of the paper.

    Strengths:
    The transport experiments and simulations here are thorough, carefully performed, and reasonably interpreted. The authors' arguments for noncompetitive inhibition seem well-thought-out and reasonable, as is the conclusion that ESCI#6 binds to the inward-facing state of the transporter. The simulations are also thorough and support the conclusions. In the discussion, the comparison of enzyme noncompetitive inhibition to the process studied here was thoughtful and interesting. Also, the care and analysis of the uptake data are a strength of the paper, with well-presented evidence of reproducibility and statistics. The electrophysiology data is more limited but does communicate the essential conclusion.

    Weaknesses:
    The most important concern about the work is the interpretation of the in-vivo drosophila data. Though the SERT fluorescence with WT protein is strong, I cannot see any fluorescence in either drug-treated image from the PG mutant. In this context, shouldn't there be additional intracellular staining for ER-resident SERT? If the cell bodies of these cells are elsewhere this should be clearly pointed out.

    Similarly, the single Western blot demonstrating enhanced glycosylation in the presence of Noribogaine or ECSI#6 could be strengthened. I can see the increased band at a high molecular weight that the authors attribute to the fully glycosylated form, but this smear, and the band below, look quite different from those in the blot shown in the El-Kasaby et al reference, raising concerns that the band could be aggregated or dimerized protein rather than a glycosylated form. This concern could easily be addressed by control experiments with appropriate glycosidases, as shown in the reference.

    The overall interest in the work is reduced given the quite low affinity of ECSI#6 for the transporter.

  4. eLife

    Reviewer #2 (Public Review):

    The authors aimed to determine the mode of inhibition of the serotonin transporter SERT as a by-product of MDMA synthesis (ECSI#6). They present a thorough kinetic analysis, using different experimental techniques (binding and transport inhibition) and kinetic modelling. They also test the predicted pharmocophore effect of the compound. In my view, the authors provide compelling evidence for an uncompetitive inhibition mechanism, in which the compound most likely binds to the inward-facing and K+-bound state. Inhibitors of this type may have the potential for therapeutic use.

  5. eLife

    Reviewer #3 (Public Review):

    This is interesting research that uncovers a novel inhibition mechanism for serotonin (SERT) transporters, which is akin to traditional un-competitive inhibitors in enzyme kinetics. These inhibitors are known to preferentially bind to the enzyme-substrate complex, thus stabilizing it, resulting in a decrease of the IC50 with increasing substrate concentrations. In contrast to this classic enzyme inhibition mechanism, the authors show for SERT, through detailed kinetic analysis as well as kinetic modeling, that the inhibitor, ECSI#6, binds preferentially to the inward-facing state of the transporter, which is stabilized by K+. Therefore, inhibition becomes "use-dependent", i.e. increasing substrate concentrations push the transporter to the inward-facing configuration, which then leads to the increased apparent affinity of ECSI#6 binding. Interestingly, this mechanism of action predicts that the inhibitor should be able to rescue SERT misfolding variants. The authors tested this possibility and found that surface expression and function of a misfolding mutant SERT is increased, an important experimental finding. Another strength of the manuscript is the quantitative analysis of the kinetic data, including kinetic modeling, the results of which can reconcile the experimental data very well. Overall, this is important and, in my view, novel work, which may lead to new future approaches in SERT pharmacology.

    With that said, some weaknesses of the manuscript should be mentioned. 1) The authors suggest that serotonin and ECSI#6 cannot bind simultaneously to the transporter, however, no direct evidence for this conclusion is provided. 2) How does ECSI#6 access the inward-facing binding site? Does it permeate the membrane and bind from the inward-facing conformation, or is it just a very slowly transported low-affinity substrate that stabilizes the inward-facing state with much higher affinity? Including ECSI#6 in the recording electrode may provide further information on this point. Additionally, it is not clear why displacement experiments were not carried out with cocaine. Since cocaine is a competitive inhibitor but does not induce transport (i.e. doesn't induce the formation of the inward-facing conformation), it should act in a competitive mechanism with ECSI#6. 3) Why are dose-response relationships not shown for electrophysiological experiments? These would be a good double-check for the radiotracer flux data.
    Despite these weaknesses, I believe that this is important work, which adds to our understanding of the pharmacology of serotonin transporters, which are of critical nature due to being a target of anti-depressant drugs. The data make a case for the proposed inhibition mechanism and the interpretation of results, as well as conclusions, are generally sound.

  6. eLife

    Author Response:

    Reviewer #1 (Public Review):

    This paper reports an analysis of the inhibition of the serotonin transporter, SERT, by a novel compound, ECSI#6. The authors perform a comprehensive analysis of SERT transport inhibition for the new agent and compare its properties to those of other well-characterized agents: cocaine and noribogaine, with the data pointing to an unusual noncompetitive mechanism of inhibition, a model also supported by electrophysiological recordings of transport currents. Based on the results of these experiments the authors conclude that ESCI#6 binds essentially exclusively to the inward-facing state of the transporter. The authors further present experiments suggesting that ESCI#6 can stabilize the folded form of an ER-arrested SERT mutant and recover its trafficking to the plasma membrane, with some in-vivo drosophila experiments perhaps also supporting this conclusion. Finally, kinetic simulations using a transport model with rate constants from previous experiments support the basic conclusions of the first sections of the paper.

    Strengths:
    The transport experiments and simulations here are thorough, carefully performed, and reasonably interpreted. The authors' arguments for noncompetitive inhibition seem well-thought-out and reasonable, as is the conclusion that ESCI#6 binds to the inward-facing state of the transporter. The simulations are also thorough and support the conclusions. In the discussion, the comparison of enzyme noncompetitive inhibition to the process studied here was thoughtful and interesting. Also, the care and analysis of the uptake data are a strength of the paper, with well-presented evidence of reproducibility and statistics. The electrophysiology data is more limited but does communicate the essential conclusion.

    Weaknesses:
    The most important concern about the work is the interpretation of the in-vivo drosophila data. Though the SERT fluorescence with WT protein is strong, I cannot see any fluorescence in either drug-treated image from the PG mutant. In this context, shouldn't there be additional intracellular staining for ER-resident SERT? If the cell bodies of these cells are elsewhere this should be clearly pointed out.

    We have modified Fig. 6 to include, in all instances, images of the posterior brain, where the neurons (FB6K) reside, from which the serotonergic projections originate. These images visualize expression of membrane-anchored GFP (mCD8GFP; in panel B), immunoreactivity of serotonin (panel B’), wild type SERT (panels C’,D’,E’) and mutant SERT-PG601,602AA (panels F’,G’,H’) in the soma. The description of these panels has been added to the pertinent sentences starting on p. 20, line 6 from bottom to the end of end of the first paragraph p. 21, which read:

    “These projections (Fig. 6A-A’’) and the FB6K-type neurons, from which they originate in the posterior brain (Fig. 6B-B’’) can be visualized by expressing membrane-anchored GFP (i.e. GFP fused to the C-terminus of murine CD8; [36]) under the control of TRH-T2A-Gal4. Similarly, when placed under the control of TRH-T2A-Gal4, YFP-tagged wild-type human SERT was expressed in the FB6K-type neurons (Fig. 6C’) and delivered to the fan-shaped body (Fig. 6C). In contrast, in flies harboring human SERT-PG601,602AA, the transporter was visualized in the soma of FB6K-type neurons (Fig. 6F’), but the fan-shaped body was devoid of any specific fluorescence (Fig. 6F). However, if three-day old male flies expressing human SERT- PG601,602AA were fed with food pellets containing 100 μM ECSI#6 or 100 μM noribogaine for 48 h, fluorescence accumulated to a level, which allowed for delineating the fan-shaped body (Fig. 6G and H, respectively). This show that ECSI#6 and noribogaine exerted a pharmacochaperoning action in vivo, which partially restored the delivery of the mutant transporter to the presynaptic territory. As expected, in flies harboring wild-type human SERT, feeding of ECSI#6 and noribogaine did not have any appreciable effect on the level of fluorescence in the fan-shaped body (Fig. 6D and E, respectively). “

    Similarly, the single Western blot demonstrating enhanced glycosylation in the presence of Noribogaine or ECSI#6 could be strengthened. I can see the increased band at a high molecular weight that the authors attribute to the fully glycosylated form, but this smear, and the band below, look quite different from those in the blot shown in the El-Kasaby et al reference, raising concerns that the band could be aggregated or dimerized protein rather than a glycosylated form. This concern could easily be addressed by control experiments with appropriate glycosidases, as shown in the reference.

    We understand that the appearance of the mature glycosylated species is being criticized, at least in part, because it differs from sharper bands, which can be found in our previously published papers. We stress that the resolution very much depends on the electrophoretic conditions. We addressed the reviewers’ criticism by carrying out the recommended deglycosylation experiments: a representative experiment is shown in (the new) panel F of Fig. 5, with lysates prepared from HEK293 cells expressing wild type SERT, from untransfected HEK293 cells and from HEK293 cells, which had been preincubated with 30 μM cocaine, 100 μM ECSI#6 and 30 μM noribogaine. The experiment confirms the band assignment with the upper band(s) M representing the mature glycostylated species (which are resistant to deglycosylation by endoglycosidase H) and the lower band C corresponding to the core- gylcoylated species (which are susceptible to cleavage that (as expected) the mature band show a representative degylcosylation by endoglycosidase H). We also think that the immunoblot in panel F ought to satisfy the aesthetic criticism: the bands are sharper/less smeared.

    The description of panel F can be found on p. 18, starting in line 7 from bottom to end of page, and reads: “We confirmed the band assignment by enzymatic deglycosylation (Fig. 5F): the upper bands (labeled M), which appeared in cells incubated in the presence of ECSI#6 and of norbogaine, were resistant to deglycosylation by endoglycosidase H (which cannot cleave mature glycans). In contrast, the core-glycosylated species (labeled C), was susceptible to cleavage by endoglycosidase H resulting in the appearance of the deglycosylated band D.”

    The overall interest in the work is reduced given the quite low affinity of ECSI#6 for the transporter.

    We agree that it would be preferable to have a compound, which works in the submicromolar/nanomolar range. However, it is worth pointing out that the EC50 is low enough for allowing in vivo rescue of the folding-deficient SERT-PG: feeding flies restores its trafficking to the cell surface and to the presynaptic specialization. Obviously, there is room for improvement, but ECSI#6 provides a starting point.

    Reviewer #3 (Public Review):

    This is interesting research that uncovers a novel inhibition mechanism for serotonin (SERT) transporters, which is akin to traditional un-competitive inhibitors in enzyme kinetics. These inhibitors are known to preferentially bind to the enzyme-substrate complex, thus stabilizing it, resulting in a decrease of the IC50 with increasing substrate concentrations. In contrast to this classic enzyme inhibition mechanism, the authors show for SERT, through detailed kinetic analysis as well as kinetic modeling, that the inhibitor, ECSI#6, binds preferentially to the inward-facing state of the transporter, which is stabilized by K+. Therefore, inhibition becomes "use-dependent", i.e. increasing substrate concentrations push the transporter to the inward-facing configuration, which then leads to the increased apparent affinity of ECSI#6 binding. Interestingly, this mechanism of action predicts that the inhibitor should be able to rescue SERT misfolding variants. The authors tested this possibility and found that surface expression and function of a misfolding mutant SERT is increased, an important experimental finding. Another strength of the manuscript is the quantitative analysis of the kinetic data, including kinetic modeling, the results of which can reconcile the experimental data very well. Overall, this is important and, in my view, novel work, which may lead to new future approaches in SERT pharmacology.

    With that said, some weaknesses of the manuscript should be mentioned. 1) The authors suggest that serotonin and ECSI#6 cannot bind simultaneously to the transporter, however, no direct evidence for this conclusion is provided.

    We assessed this point by plotting the data in Fig. 2A,B,C as Dixon plots in (the new) panels D,E,F of Fig. 2. We refer the reader to Segel’s textbook on enzyme kinetics (new ref. 18) on using Dixon plots in the presence of two inhibitors. The pertinent description is on p. 9, lines 12-22 and reads as follows: “We transformed the data summarized in Figs. 2A-C by plotting the reciprocal of bound radioligand as a function of inhibitor concentration to yield Dixon plots (Fig. 2D-F): the x-intercept corresponds to -IC50 of the inhibitor [18]. Thus, Dixon plots allow for differentiating mutually exclusive from mutually non-exclusive binding, if one inhibitor (i.e., cocaine, noribogaine or ECSI#6) is examined at a fixed concentration of the second inhibitor (i.e., serotonin) [18]: if binding of the two inhibitors is mutually non-exclusive, a family of lines of progressively increasing slope, which intersect at -IC50, is to be seen. In contrast, if the two inhibitors bind to the same site, the slope of the inhibition curves is not affected and the x- intercept (i.e, -IC50 of the variable inhibitor) is shifted to more negative values. It is evident from Fig. 2D-E that the presence of 1 and 10 μM serotonin progressively shifted the (expected) x-intercept for cocaine (Fig. 2D), noribogaine (Fig. 2E) and ECSI#6 (Fig. 2D). Thus, binding to SERT of serotonin and of these three ligands was mutually exclusive.” Based on the Dixon plots, we feel that our conclusion is justified, i.e., binding of serotonin and ECSI#6 (and of the other ligands) is mutually exclusive.

    1. How does ECSI#6 access the inward-facing binding site? Does it permeate the membrane and bind from the inward-facing conformation, or is it just a very slowly transported low-affinity substrate that stabilizes the inward-facing state with much higher affinity? Including ECSI#6 in the recording electrode may provide further information on this point.

    We did the suggested experiments: the data are summarized in panel I of Fig. 4 and described in the first paragraph on p. 15, which also explains why this experiments is possibly inconclusive due to the high diffusivity of ECSI#6:

    “Fig. 4I shows representative traces of 5-HT induced currents recorded from SERT expressing cells in the absence (in blue) and presence of 100 μM ECSI#6 (in red) in the electrode solution: when thus applied from the intracellular side, ECSI#6 did not cause an appreciable current block. The right-hand panel summarizes the current amplitude obtained from cells measured in the absence (blue open circles) and presence of intracellular ECSI#6 (open circles in red). These data seem to indicate that ECSI#6 binds to SERT from the extracellular side. Yet this conclusion can be challenged based on the following consideration: in earlier experiments, ibogaine, the parent compound of noribogaine, was found to block HERG channels when applied from the bath solution but failed to do so when added to the electrode solution [27]. However, at a lower intracellular pH (i.e., pH 5.5), ibogaine gained the ability to inhibit HERG from the intracellular side (i.e., via application through the electrode). Conversely, ibogaine was less effective when applied to an acidified bath solution. These observations led to the conclusion that ibogaine blocked HERG from the cytosolic side: because the molecule in its neutral form was so diffusive, a low intracellular pH was required to force its protonation and thus preclude diffusion from the interior of the cell. ECSI#6 is presumed to also be very diffusible given its estimated logP value and polar surface area of 2.48 and 66 Å2, respectively. However, ECSI#6 harbors an amide nitrogen (see Fig. 1A) and thus remains neutral in the experimentally accessible pH range. Hence, it is not possible to verify to which side of SERT it binds.”

    Additionally, it is not clear why displacement experiments were not carried out with cocaine. Since cocaine is a competitive inhibitor but does not induce transport (i.e. doesn't induce the formation of the inward-facing conformation), it should act in a competitive mechanism with ECSI#6.

    We did not quite understand this comment, because displacement experiments were done with cocaine (Fig. 2A, new Fig 2G/previous Fig. 2D). However, if the reviewer questions why we do not use cocaine rather than 5-HT, in the three-way competition experiment, it is precisely, because we wanted to compare the action/binding mode of ECSI#6 to that of cocaine.

    1. Why are dose-response relationships not shown for electrophysiological experiments? These would be a good double-check for the radiotracer flux data.

    These experiments were done and are shown in (the new) panels G and H of Fig. 4; the pertinent description is in the second paragraph of p. 14 and reads:

    “The protocol depicted in Fig. 4B can also be used to gauge the apparent affinity of ECSI#6 for SERT in the presence of 5-HT. Plotted in Fig. 4G is the block of the serotonin-induced current as a function of the co-applied ECSI#6 concentration. The current was evoked by a saturating concentration of 5-HT (30μM) and inhibited by 3, 10, 30 and 100 μM co-applied ECSI#6, respectively (the inset in Fig. 4G shows representative current traces). A fit of an inhibition curve to the data points yielded an IC50 value of approx. 5 μM. This value was lower but still in reasonable agreement, with the IC50 obtained in the radioligand uptake assay for the condition where the 5-HT concentration had been saturating (cf. dashed line in Fig.1C; 10 μM 5-HT). In the uptake assay the IC50 value of ECSI#6 dropped to about 0.5 mM, in the presence of a low 5-HT concentration (i.e., 0.1 μM). In contrast to uptake experiments, electrophysiological recordings also allow for assessing the apparent affinity of ECSI#6 for SERT in the absence of the substrate. This can be achieved by employing the protocol depicted in Fig. 4H (see representative current traces on the left-hand side): we first applied 30 μM 5- HT to a cell expressing SERT for 0.5 s to elicit a peak current (i.e., a control pulse). We then reapplied 30 μM 5-HT after a superfusing the cell with 100 μM ECSI#6 for 1 s (second upper trace in panel H). We chose this time period because it had been sufficient to allow for full current block in the other protocol (see Fig. 4G): the amplitude of the peak current following pre-application of 100 μM ECSI#6 was essentially identical to the prior control pulse. When we pre-applied 100 μM ECSI#6 for a longer period (i.e., 3 s) the amplitude of the two peak currents also remained the same (cf. lower traces in panel H). The right-hand panel shows the summary of several experiments. Plotted in the graph is the ratio of the second and first pulse, which was always close to one. We previously used this protocol to assess the binding kinetics of cocaine, methylphenidate and desipramine on SERT and DAT. Pre-application of these inhibitors consistently led to a concentration dependent reduction in the peak current amplitude of the second pulse in comparison to the first [23]. The lack of inhibition, thus, indicates that the affinity of ECSI#6 in the absence of 5-HT is low. To obtain estimates for the affinity of ECSI# for SERT in the absence of 5-HT we would need to apply this compound at much higher concentrations. This, however, is not possible, because ECSI#6 is poorly soluble in aqueous solutions (i.e., max. 0.03 mg/ml).”

  7. Version published to 10.1101/2022.08.11.503588v2 on bioRxiv
  8. Version published to 10.1101/2022.08.11.503588v1 on bioRxiv