Delineation of the functional properties and the mechanism of action of TMPPAA, an allosteric agonist and positive allosteric modulator of 5-HT3 receptors

Biochemical Pharmacology 110–111 (2016) 92–108

Contents lists available at ScienceDirect

Biochemical Pharmacology journal homepage: www.elsevier.com/locate/biochempharm

Delineation of the functional properties and the mechanism of action of TMPPAA, an allosteric agonist and positive allosteric modulator of 5-HT3 receptors Agnes Gasiorek a, Sarah M. Trattnig a, Philip K. Ahring b,1, Uffe Kristiansen a, Bente Frølund a, Kristen Frederiksen c, Anders A. Jensen a,⇑ a

Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, DK-2100 Copenhagen, Denmark Saniona A/S, Baltorpvej 154, DK-2750 Ballerup, Denmark c H. Lundbeck A/S, Otiliavej 9, DK-2500 Valby, Denmark b

a r t i c l e

i n f o

Article history: Received 28 February 2016 Accepted 6 April 2016 Available online 13 April 2016 Keywords: Cys-loop receptor 5-HT3 receptor Allosteric agonist Allosteric modulation Ago-PAM TMPPAA

a b s t r a c t We have previously identified a novel class of 5-hydroxytryptamine type 3 receptor (5-HT3R) agonists sharing little structural similarity with orthosteric 5-HT3R ligands (Jørgensen et al., 2011). In the present study we have elucidated the functional characteristics and the mechanism of action of one of these compounds, trans-3-(4-methoxyphenyl)-N-(pentan-3-yl)acrylamide (TMPPAA). In electrophysiological recordings TMPPAA was found to be a highly-efficacious partial agonist equipotent with 5-HT at the 5-HT3A receptor (5-HT3AR) expressed in COS-7 cells and somewhat less potent at the receptor expressed in Xenopus oocytes. The desensitization kinetics of TMPPAA-evoked currents were very different from those mediated by 5-HT. Moreover, repeated TMPPAA applications resulted in progressive current rundown and persistent non-responsiveness of the receptor to TMPPAA, but not to 5-HT. In addition to its direct activation, TMPPAA potentiated 5-HT-mediated 5-HT3AR signalling, and the allosteric link between the two binding sites was corroborated by the analogous ability of 5-HT to potentiate TMPPAA-evoked responses. The agonism and potentiation exerted by TMPPAA at a chimeric a7-nACh/5-HT3A receptor suggested that the ligand acts through the transmembrane domain of 5-HT3AR, a notion further substantiated by its functional properties at chimeric and mutant human/murine 5-HT3ARs. A residue in the transmembrane helix 4 of 5-HT3A was identified as an important molecular determinant for the different agonist potencies exhibited by TMPPAA at human and murine 5-HT3ARs. In conclusion, TMPPAA is a novel allosteric agonist and positive allosteric modulator of 5-HT3Rs, and its aberrant signalling characteristics compared to 5-HT at the 5-HT3AR underline the potential in Cys-loop receptor modulation and activation through allosteric sites. Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction The neurotransmitter serotonin (5-hydroxytryptamine, 5-HT) mediates its physiological effects through a plethora of G-protein coupled receptors and a family of ligand-gated cation-selective ion channels, the 5-HT type 3 receptors (5-HT3Rs) [1]. The 5-HT3Rs belong to the Cys-loop receptor superfamily, which also comprises nicotinic acetylcholine receptors (nAChRs), ⇑ Corresponding author at: Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, DK-2100 Copenhagen, Denmark. E-mail address: [email protected] (A.A. Jensen). 1 Current address: Faculty of Pharmacy, The University of Sydney, Sydney, NSW, Australia. http://dx.doi.org/10.1016/j.bcp.2016.04.004 0006-2952/Ó 2016 Elsevier Inc. All rights reserved.

c-aminobutyric acid type A receptors (GABAARs), glycine receptors (GlyRs) and a zinc-activated channel [2–7]. The receptors are membrane-bound homomeric or heteromeric complexes assembled from five subunits. In the case of human 5-HT3Rs, five subunits (5-HT3A-E) assemble into homomeric 5-HT3A and heteromeric 5-HT3AB, 5-HT3AC, 5-HT3AD and 5-HT3AE receptor complexes (5-HT3A and 5-HT3AB being the major physiological receptor subtypes), whereas other species only express 5-HT3A and 5-HT3B subunits [2,3,8,9]. The functions mediated by 5-HT3Rs in serotonergic transmission and as heteroreceptors regulating the activity in other neurotransmitter systems make the receptors interesting therapeutic targets. Competitive 5-HT3R antagonists are clinically administered drugs in the treatment of nausea and emesis arising after operations or from radiation and chemotherapy. Further,

93

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

5-HT3R ligands are investigated as putative therapeutics in gastrointestinal disorders, various psychiatric and cognitive disorders, and pain [8,10,11]. As other Cys-loop receptors, the pentameric 5-HT3R complex consists of an extracellular domain (ECD) made up by the N-termini of the subunits, a transmembrane domain (TMD) consisting of the four transmembrane a-helices TM1-4 in the subunits, and an intracellular domain (ICD) composed by the intracellular loops in the subunits [3,4,12]. Signal transduction is initiated by agonist binding to the orthosteric sites located at subunit interfaces in the ECD of the receptor, which subsequently elicits the opening of the ion channel in the TMD hereby enabling the flux of ions through the channel [3,4,13,14]. Following this activation, the receptor will either return to its resting unbound/closedchannel conformation upon release of the agonist from its binding site (deactivation) or transition into an agonist-bound/closedchannel conformation (desensitization). The desensitized receptor can subsequently convert back to the active state or return to its resting state (reactivation and recovery from desensitization, respectively) [4,14–16]. The numerous allosteric transitions underlying Cys-loop receptor function means that the signalling events elicited by the orthosteric agonist can be modulated by concomitant binding of ligands to numerous allosteric sites topographically distinct from the orthosteric site [4,14,15]. Positive allosteric modulators (PAMs) enhance the affinity/potency and/or the efficacy of the orthosteric agonist at the receptor, whereas negative allosteric modulators (NAMs) do the opposite. Allosteric agonists capable of directly activating the receptor have also been identified, and some of these exhibit PAM activity in addition to their intrinsic activity and are referred to as ‘‘ago-PAMs”. The receptor modulation exerted by any allosteric ligand arises from its effects on one or several of the multiple energy barriers associated with the allosteric transitions between different receptor conformations. As a result hereof, the modulatory characteristics displayed by a range of allosteric ligands can differ dramatically [4,14–16]. In contrast to the abundance of potent and selective allosteric modulators of nAChRs and GABAARs published over the years [15–17], allosteric modulators of 5-HT3Rs reported to date have typically emerged from discoveries of 5-HT3R activity in drugs with already identified pharmacological activities at other targets (see [18,19] and references herein). The fact that most of these modulators display significantly lower potencies at the 5-HT3Rs than at their original targets has limited their potential as pharmacological tools, and thus the development of novel allosteric modulators could potentially be valuable for future explorations of the receptors. In 2011 we reported the discovery of a series of benzamide analogues and homologues as a novel class of 5-HT3R agonists [20]. The structures of these compounds were substantially different from those of orthosteric 5-HT3R ligands, not least in their lack of a positive charge at physiological pH, an essential pharmacophore element for orthosteric ligand binding to 5-HT3Rs and other Cysloop receptors [3,13,19]. In agreement with this, the compounds did not compete with the orthosteric radioligand [3H]GR65530 for binding to the 5-HT3R. Further, analogues from the series were shown to potentiate 5-HT-mediated 5-HT3R signalling in a fluorescence-based functional assay [20]. Hence, we proposed that the compounds could act as ago-PAMs at 5-HT3Rs while stressing that additional investigations were needed in order to fully draw this conclusion [20]. In the present study we have characterized the functional properties of one of these compounds, trans-3-(4methoxyphenyl)-N-(pentan-3-yl)acrylamide (TMPPAA, analogue 3d in the original paper [20], Fig. 1), at the 5-HT3A receptor (5-HT3AR) in elaborate electrophysiology studies and elucidated the molecular basis for its activity at the receptor.

Fig. 1. Chemical structure of trans-3-(4-methoxyphenyl)-N-(pentan-3-yl)acrylamide (TMPPAA).

2. Materials and methods 2.1. Materials Culture media, serum, antibiotics, buffers for cell culture and the LipofectAMINETM Plus regent transfection kit were obtained from Invitrogen (Paisley, UK). The Polyfect transfection reagent was obtained from Qiagen (Hilden, Germany). ACh, 5-HT, metoclopramide, probenecid, poly-D-lysine and the 3,30 ,5,50 -tetramethyl benzidine liquid substrate system were purchased from Sigma (St. Louis, MO). PU02 was obtained from Chembridge Corporation (San Diego, CA) and ondansetron, (±)-zacopride, clozapine, picrotoxin and propofol from Tocris Cookson (Bristol, UK). TMPPAA was synthesized in-house essentially as described previously [20]. The FLIPRTM Membrane Potential Blue (FMP) assay and the Fluo-4/AM dyes were purchased from Molecular Devices (Crawley, UK) and Molecular Probes (Eugene, OR), respectively. Defolliculated stage V–VI Xenopus laevis oocytes were obtained from Lohmann Research Equipment (Castrop-Rauxel, Germany). The cDNAs encoding for the rat a7 nAChR subunit (ra7), human 5-HT3A (h5-HT3A) and mouse 5-HT3A (m5-HT3A) subunits were kind gifts from Drs. J.W. Patrick, J. Egebjerg and D. Julius, respectively, and the stable h5-HT3AR-HEK293 cell line was a kind gift from Dr. C. Rojas [21]. 2.2. Molecular biology The cDNAs used in the study were in pCIneo (h5-HT3A, m5-HT3A and ra7/m5-HT3A) and pGEMHE (human a7 (ha7) and h5-HT3A used for the oocyte experiments) vectors. The numbering of amino acid residues in the receptor subunits in this study [h5-HT3A (BC002354.2), m5-HT3A (NM_013561.2) and ra7 (NM_012832.2)] is based on the non-mature proteins (i.e., including the signal peptides). The chimeric and myc-tagged subunits were generated by splicing by overlap extension PCR [22]. The fusion points in the chimeric constructs were as follows (m5-HT3A sequences underlined): h5-HT3A/m5-HT3A (. . .F234T235-V236-/-I242-I243-R244. . .), T240-V241-/-V237-I238-R238. . .)

m5-HT3A/h5-HT3A and

(. . .V327-R328-L329-/-V335-H336-K337. . .

(. . .F239-

h5-HT3Am5-HT3A-ICL2 and

. . .R426-V427-

G428-/-S447-V448-L449. . .). The myc-m5-HT3A construct was made by insertion of the myc epitope sequence (n-EQKLISEEDLc) immediately after the m5-HT3A signal peptide sequence between residues G23 and S24 (predicted by SignalP 4.1 [23]). The construction of the ra7/m5-HT3A (. . .T223-V224-/-I242I243-R244. . .) and myc-h5-HT3A constructs has been described previously [24,25]. Point mutations were introduced into the cDNAs for h5-HT3A, m5-HT3A and myc-tagged versions of these using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). The absence of unwanted mutations in all cDNAs created by PCR was verified by sequencing (Eurofins MWG Operon, Ebersberg, Germany). 2.3. Cell culture and transfections All cell lines were cultured in a humidified atmosphere of 5% CO2 and 95% air at 37 °C. The stable h5-HT3AR-HEK293 cell line

94

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

[21] was maintained in RPMI 1640 containing penicillin (100 U/ml), streptomycin (100 lg/ml), 10% foetal bovine serum, and 0.5 mg/ml G-418. The tsA201 cells were grown in Dulbecco’s Modified Eagle Medium + Glutamax-I (DMEM) supplemented with penicillin (100 U/ml), streptomycin (100 lg/ml), and 10% foetal bovine serum. For the transfections of tsA201 cells, 1.2  106 cells were split into a 6 cm tissue culture plate and transfected the following day with a total of 4 lg cDNA using Polyfect as transfection reagent. The cells were assayed 36–48 h after the transfection. The COS-7 cells used for the whole-cell patch-clamp recordings were cultured in DMEM supplemented with penicillin (100 U/ml), streptomycin (100 lg/ml) and 10% foetal bovine serum. The cells were split into T25-flasks (0.8  106 cells) and transiently transfected with 1 lg h5-HT3A-pCIneo or m5-HT3A-pCIneo together with a plasmid encoding for green fluorescent protein using the LipofectAMINETM Plus regent transfection kit. The cells were used 36–72 h after transfection. 2.4. Ca2+/Fluo-4 assay The experiments with the h5-HT3AR-HEK293 cell line [21] in the Fluo-4/Ca2+ assay were performed essentially as described previously [20]. Cell lines were split into poly-D-lysine-coated black 96-well plates with clear bottom (BD Biosciences, Bedford, MA). Following a 16–24 h incubation, the culture medium was aspirated and the cells were incubated in 50 ll assay buffer [Hank’s Buffered Saline Solution containing 20 mM HEPES, 1 mM CaCl2, 0.5 mM MgCl2 and 2.5 mM probenecid, pH 7.4] supplemented with 6 mM Fluo-4/AM at 37 °C for 1 h. Then the buffer was aspirated, the cells were washed once with 100 ll assay buffer, and then 100 ll assay buffer was added to the cells (in the antagonist experiments, various concentrations of the antagonist were dissolved in this buffer). The 96-well plate was assayed in a NOVOstarTM microplate reader (BMG Labtechnologies, Offenburg, Germany) measuring emission [in fluorescence units (FU)] at 520 nm caused by excitation at 485 nm before and up to 60 s after the addition of 33 ll agonist solution in assay buffer. The experiments were performed at room temperature in duplicate at least three times for each compound. 2.5. FLIPRTM Membrane Potential Blue (FMP) assay The experiments with transiently transfected tsA201 cells in the FMP assay were performed essentially as described previously [26]. 16–24 h after the transfection, the cells were split into polyblack 96-wells plates (8  104 cells/well) with clear bottom (BD Biosciences, Bedford, MA). 16–24 h later, the medium was aspirated, and the cells were washed with 100 ll Krebs buffer [140 mM NaCl/4.7 mM KCl/2.5 mM CaCl2/1.2 mM MgCl2/11 mM HEPES/10 mM D-Glucose, pH 7.4]. 50 ll Krebs buffer was added to the wells (in the antagonist experiments, various concentrations of the antagonist were dissolved in this buffer) after which an additional 50 ll Krebs buffer supplemented with the FMP assay dye (1 mg/ml) was added to each well. Then the plate was incubated at 37 °C in a humidified 5% CO2 incubator for 30 min and assayed in a NOVOstarTM plate reader measuring emission [in fluorescence units (FU)] at 560 nm caused by excitation at 530 nm before and up to 1 min after the addition of 33 ll agonist solution. In the experiments with the chimeric ra7/m5-HT3A receptor and in the studies of the inhibition of 5-HT- and TMPPAA-mediated signalling by ondansetron and metoclopramide, the plates were assayed in a FLEXStation3 reader (Molecular Devices, Crawley, UK). Hank’s Buffered Saline Solution supplemented with 20 mM HEPES, 1 mM CaCl2, 1 mM MgCl2 (pH 7.4) was used as assay buffer for the experiments with the chimeric ra7/m5-HT3A receptor. The experiments were performed at room

D-lysine-coated

temperature in duplicate at least three times for each compound at all receptors. 2.6. Enzyme-linked immunosorbent asssay (ELISA) The total and cell surface expression levels of myc-tagged versions of selected wild type (WT), chimeric and mutant receptors were quantified in an ELISA. 16–24 h after transfection, tsA201 cells transfected with myc-h5-HT3A, myc-m5-HT3A, myc-h5HT3A/m5-HT3A, myc-m5-HT3A/h5-HT3A, myc-h5-HT3AM470T, myc-m5-HT3AT481M or untagged h5-HT3A (unspecific binding) cDNAs were plated into PDL-coated 48-well plates (2.5  105 cells/well). The following day, the ELISA was performed essentially as described previously [25]. Briefly, cells were washed and fixed in 4% paraformaldehyde, washed with phosphate-buffered saline (PBS), and incubated with blocking solution (3% dry milk in 50 mM Tris–HCl, 1 mM CaCl2, pH 7.5) for 20 min on ice, after which the rest of the assay was performed at room temperature. Following several rounds of washing with PBS, the cells were incubated with mouse anti-myc antibody (Invitrogen, Paisley, UK) diluted 1:1000 in blocking solution for 45 min. Following several rounds of washing with PBS, the cells were incubated with blocking solution (20 min) prior to incubation with a horseradish peroxidaseconjugated goat anti-mouse secondary antibody (Molecular Probes, Eugene, OR) diluted 1:400 in blocking solution (45 min). Cells were then washed three times with PBS before protein expression was determined using the 3,30 ,5,50 -tetramethylbenzi dine liquid substrate system and H2SO4. The absorbance of the supernatant was determined at 450 nm. Total expression levels of the myc-tagged proteins were determined using Triton X-100 (0.1% in blocking solution) during the first blocking step and during the incubation with the primary antibody. 2.7. Xenopus oocytes and two-electrode voltage clamp (TEVC) electrophysiology The cDNA constructs encoding for h5-HT3A, ha7 and ra7/ m5HT3A were linearized, transcribed, and capped using the mMessage mMachine T7 kit (Ambion, Austin, TX) according to the manufacturer’s protocol, after which the reaction mixtures were subject to an additional 15 min Turbo DNase reaction at 37 °C to remove the remaining cDNA template. The cRNAs were purified using the NucleoSpinÒ RNA Clean-up XS kit (Macherey–Nagel, Düren, Germany) and it was confirmed that the cRNAs had the correct sizes on a gel. The h5-HT3A, ha7 and ra7/m5-HT3A cRNAs were injected into oocytes in amounts of 0.02–0.92 ng, 0.6– 5.7 ng, and 27.6–49.3 ng, respectively (injection volumes 4.6– 41.4 nl), and incubated at 18 °C in modified Barth’s saline solution [in mM: 88 NaCl, 1 KCl, 15 HEPES, 2.4 NaHCO3, 0.41 CaCl2, 0.82 MgSO4, 0.3 Ca(NO3)2, pH = 7.5] supplemented with penicillin (100 U/ml), streptomycin (100 lg/ml). Whole-cell electrophysiological recordings were performed using the TEVC technique on oocytes 2–4 days after injection. The oocyte was placed in a recording chamber and superfused at a rate of 5 ml/min with saline solution. For the recordings on h5HT3AR- and ha7-expressing oocytes, extracellular saline solution S1 [in mM: 115 NaCl, 2.5 KCl, 10 HEPES, 1.8 CaCl2, 0.2 MgCl2, pH = 7.5] was used, whereas the recordings on ra7/m5-HT3Aexpressing oocytes were performed using extracellular saline solution S2 [in mM: 90 NaCl, 1.0 KCl, 10 HEPES, 0.5 BaCl2, 0.01 EDTA, pH = 7.5] on account of the reported Ca2+-mediated inhibition of the chimeric receptor [27]. The oocytes were clamped at 60 mV using the OC-725C TEVC Amplifier (Warner Instruments, Hamden, CT) and both voltage- and current electrodes were agar-plugged, filled with 3 M KCl, and displayed resistances of 0.5–2 MX. The compounds (in saline solutions S1 or S2) were applied until peaks

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

of the responses were reached (30–60 s) followed by a washout period of 2–3 min. In the antagonist experiments, the oocyte was preincubated for 60 s with the test compound before coapplication of the test compound and the agonist. In the PAM experiments, the oocyte was either preincubated for 30 s with the test compound before co-application of the test compound and the agonist (‘‘equilibrium application”), or preincubated for 30 s with the test compound before application of the agonist (‘‘close channel” application”).

95

TEVC data (mean ± S.E.M. values) were analysed using a paired two-sided t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Statistical analysis of data from the two fluorescence-based assays was performed using an unpaired two-sided t-test (*p < 0.05, **p < 0.01, ***p < 0.001), in some cases with the addition of a Bonferroni correction for multiple comparisons (*p < 0.025, **p < 0.005, ***p < 0.0005).

3. Results 2.8. Whole-cell patch-clamp electrophysiology

3.1. Electrophysiological characterization of TMPPAA at 5-HT3ARs

COS-7 cells transiently expressing h5-HT3AR or m5-HT3AR were subjected to recording at room temperature in the whole-cell voltage-clamp configuration essentially as previously described [26,28,29]. The cells were rinsed with PBS and detached from the culture flask by tripleX [0.1% (w/v)] digestion for 2 min at 37 °C and seeded on the day of experiment. Glass cover slips (3.5 mm) precoated with poly-D-lysine [0.005% (w/v)] were placed in Petri dishes and cells were added at a suitable density. Coverslips with cultured cells were transferred to a perfusion chamber mounted on the stage of an inverted microscope supplied with Nomarski optics and superfused with extracellular solution (in mM; 140 NaCl, 11 glucose, 10 HEPES, 4.7 KCl, 0.1 CaCl2, adjusted to pH 7.4 with NaOH). Patch pipettes were pulled from borosilicate glass using a horizontal electrode puller (Zeitz Instrumente, Augsburg, Germany). When filled with intracellular solution (in mM; 120 KCl, 1.8 MgCl2, 10 EGTA, 10 HEPES, adjusted to pH 7.4 with KOH) and immerged in extracellular solution patch pipettes had a resistance of approximately 2 MX. All data were obtained with an EPC-9 amplifier (HEKA Electronics, Lambrecht, Germany) run by a Windows XP personal computer. Experimental conditions and data acquisition were set and obtained using the PULSE-software accompanying the amplifier. Data were low pass filtered and sampled directly to the hard disc, cells were held at 60 mV during recordings. Compounds were dissolved in extracellular solution and applied to the patched cell through a double-barrelled application pipette. Application pipettes were fabricated from theta glass tubes (1.5 mm outer diameter; WPI, Sarasota, FL) and mounted on a piezoelectric device (PZS-100HS; Burleigh Instruments, Quebec, Canada) connected to a piezo-driver (PZ-150 M; Burleigh Instruments) driven by TTL pulses from the EPC-9 amplifier. Approximately one minute after the onset of the gravity flow, a PULSE protocol was initiated and the current was recorded until stably separated by 45-s waiting periods.

Initially the functional properties of 5-HT were characterized at h5-HT3ARs expressed in COS-7 cells in whole-cell patch-clamp electrophysiological recordings. In good agreement with the findings in previous electrophysiology studies of 5-HT3AR in mammalian cells [30–33], 5-HT elicited currents in a concentrationdependent manner. Fitting averaged peak current amplitudes to the Hill equation revealed an EC50 value of 8.1 lM (pEC50 ± S.E. M.: 5.10 ± 0.08) and a Hill slope of 1.1 ± 0.27 at h5-HT3AR (n = 6, Fig. 2A). These 5-HT data are identical to those in a previous publication [26], because the patch-clamp recordings for the two studies were performed in parallel at the same patch-clamp set-up.

2.9. Data analysis All analysis and curve fitting were performed using Prism (version 6.0d; Graphpad Software, San Diego, CA). As for the data from the whole-cell patch-clamp recordings, macroscopic currents were extracted from PULSE (HEKA, Lamprecht, Germany) and subsequently analysed using IgorPro software (Wavemetrics, Lake Oswego, OR). For the data from the TEVC recordings, the currents were analysed using Clampfit (version 10.4, pCLAMP Software, Molecular Devices, Sunnyvale, CA). Concentration–response data for the agonists were fitted to a sigmoidal curve with variable slope using nonlinear regression: Y = Bottom + (Top  Bottom)/[1 + (10^(log EC50  X)  nH)], where X is the logarithm of the agonist concentration, Y is the response and nH is the Hill slope. EC50, nH and Imax values were derived from these equations. Statistical analysis of data from the TEVC recordings at ha7 nAChR (mean ± S.E.M. values) was performed by one-way ANOVA analysis followed by Dunnett’s Multiple Comparisons test (p > 0.05), whereas other

3.1.1. Basic properties of TMPPAA as a 5-HT3AR agonist In agreement with its agonism at h5-HT3A and h5-HT3AB receptors in fluorescence-based functional assays [20], TMPPAA elicited currents in a concentration-dependent manner in both h5-HT3ARand m5-HT3AR-expressing COS-7 cells (Fig. 2B and C). As will be addressed in further detail below, repeated applications of high concentrations of TMPPAA (10–300 lM) resulted in progressive current amplitude run-down, something that was not observed upon repeated 5-HT applications. Since it thus was not possible to record complete concentration–response relationships for TMPPAA at individual cells, the curves in Fig. 2B and C are based on recordings from numerous cells where the current elicited by a specific TMPPAA concentration has been normalized to that evoked by 100 lM 5-HT at the same cell. The curve fitted to these data points revealed an EC50 value of 7.1 lM, a Hill slope of 2.4, and an Imax of 84% (of that mediated by 100 lM 5-HT) for TMPPAA at the h5-HT3AR (n = 3–5 for each data point, Fig. 2B). Thus, TMPPAA was a high-efficacious partial agonist (in terms of its peak current amplitudes) and equipotent with 5-HT at this receptor. Interestingly, TMPPAA displayed 3-fold lower agonist potency at the m5-HT3AR, whereas its Hill slope and maximal response at this receptor were comparable to those at the h5-HT3AR (EC50 = 20 lM, Hill slope = 2.9, Imax = 92%, determined from fitted curve, n = 3–5 for each data point, Fig. 2C). With both receptors, rebound currents were often observed following termination of the applications of high TMPPAA concentrations (100 lM and 300 lM) (Fig. 2B). Importantly, TMPPAA did not evoke significant responses in non-transfected COS-7 cells, and TMPPAA-mediated responses in h5-HT3AR-expressing cells were completely inhibited by coapplication of the 5-HT3R antagonist PU02 (100 lM) (data not shown). The agonist properties of TMPPAA were also studied at the h5-HT3AR expressed in Xenopus oocytes in TEVC recordings. In good agreement with previous studies of 5-HT3ARs in oocytes [34,35], 5-HT exhibited an EC50 value of 1.9 lM (pEC50 ± S.E.M.: 5.72 ± 0.03, n = 2) at the receptor. However, TMPPAA was a less potent h5-HT3AR agonist in this expression system compared with COS-7 cells, as it only elicited significant currents in the oocytes at concentrations of 30 lM and higher (Fig. 2D). Because of this lower potency a complete concentration–response relationship for TMPPAA could not be recorded, and thus its efficacy at the receptor

96

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Fig. 2. Electrophysiological characterization of 5-HT and TMPPAA as agonists at h5-HT3AR and m5-HT3AR expressed in COS-7 cells and at h5-HT3AR expressed in Xenopus oocytes. In all current traces in A–C, the vertical bars represent 200 pA and the horizontal bars 5 s. In D, the vertical bar represents 200 nA and the horizontal bar 30 s. (A) Representative currents recorded for various concentrations of 5-HT (left) and the averaged concentration–response curve for 5-HT at h5-HT3AR-expressing COS-7 cells (right). Each data point represents mean ± S.E.M. of n = 6 cells from at least 3 different transfections on 3 different days. Inset: Representative current trace of the current evoked by prolonged application of 100 lM 5-HT at h5-HT3AR-expressing COS-7 cells. (B) Representative currents recorded for various concentrations of TMPPAA (left) and the averaged concentration–response curve for TMPPAA at h5-HT3AR-expressing COS-7 cells (right). Each of the data points was recorded from individual cells and their peak current amplitude values were normalized to the peak current amplitude induced by 100 lM 5-HT at the same cells. Each data point represents mean ± S.E.M. of n = 3–5 cells from at least 3 different transfections on 3 different days. (C) Averaged concentration–response curve for TMPPAA at m5-HT3AR-expressing COS-7 cells. Each of the data points was recorded from individual cells and their peak current amplitude values were normalized to the peak current amplitude induced by 100 lM 5-HT at the same cells. Each data point represents mean ± S.E.M. of n = 3–5 from at least 3 different transfections on 3 different days. (D) Representative currents (left) and all current amplitudes recorded (right) for 10 lM, 30 lM and 100 lM TMPPAA at h5-HT3AR-expressing oocytes. The currents (in % of the currents evoked by 30 lM 5-HT (EC90) at the same oocyte) are given with S.E.M. values.

could not be determined. The currents evoked by 100 lM TMPPAA in h5-HT3AR-expressing oocytes were completely inhibited by

preincubation and co-application with 10 nM ondansetron (data not shown).

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

3.1.2. Kinetic characteristics of TMPPAA as a 5-HT3AR agonist As can be seen from the currents traces for 5-HT and TMPPAA at the h5-HT3AR expressed in COS-7 cells, the kinetic properties of the two agonists at the receptor differed dramatically (Fig. 2A and B). Since 5-HT and TMPPAA were equipotent receptor agonists in these recordings, these differences can be observed by direct comparisons of the characteristics of currents evoked by identical concentrations of the two agonists. The current rise kinetics of the responses evoked by TMPPAA were considerably slower than those mediated by comparable concentrations of 5-HT, as reflected in the longer application times needed to reach peak current amplitudes for TMPPAA (up to 4 s) than for 5-HT (up to 0.5 s). This difference was in particular pronounced for 5-HT and TMPPAA concentrations evoking submaximal responses (10 lM and 30 lM), whereas the rise kinetics of currents produced by saturating concentrations of the two agonists were more comparable (Fig. 2A and B). Activation (current rise) kinetics are the result of multiple components: the ligand application rate of the system, the rate of ligand diffusion to its binding site, the kinetics of ligand binding, and the kinetics of the receptor gating induced by the ligand. Since an ultra-fast application system was used for the recordings, the slow activation kinetics of the TMPPAA currents cannot be ascribed to the rate of ligand application being a limiting factor, something which is also evidenced by the fast rise kinetics observed for 5-HT currents (Fig. 2A). Moreover, the fairly similar current rise kinetics observed for saturating TMPPAA (100 lM and 300 lM) and 5-HT (100 lM) concentrations suggest the receptor gating kinetics of the two agonists do not differ substantially. Thus, the different current rise kinetics elicited by sub-maximal concentrations of TMPPAA and 5-HT seem to arise from a slower diffusion of TMPPAA to its binding site and/or slower binding kinetics to the receptor. The differences in the current decay kinetics between the two agonists were even more noticeable than those observed for the current rise. Whereas saturating 5-HT concentrations desensitized the receptor almost completely within 5 s (inserted current trace, Fig. 2A), the responses mediated by saturating concentrations of TMPPAA were characterized by very slow decay (Fig. 2B). Thus, although TMPPAA was a partial agonist in terms of its peak current amplitudes, the total charge passing the membrane during a sustained application exceeded that evoked by a comparable 5-HT concentration. In contrast to its aberrant desensitization kinetics,

97

the fact that TMPPAA-evoked currents quickly returned to basal levels when applications were terminated indicated relatively fast receptor deactivation/unbinding kinetics of the agonist (Fig. 2B). 3.1.3. Current run-down observed for TMPPAA as a 5-HT3AR agonist An interesting difference between 5-HT- and TMPPAAmediated signalling noted above was the progressive current amplitude run-down observed upon repeated TMPPAA applications, which ultimately rendered the h5-HT3AR non-responsive to the agonist. As exemplified in Fig. 3A, peak amplitudes of the currents evoked by a saturating concentration of TMPPAA (300 lM) decreased with repeated applications, and the kinetics of the evoked currents likewise changed gradually. A similar rundown was also observed upon repeated applications of a 10-fold lower concentration of TMPPAA (30 lM), although it required more applications to make the receptor completely nonresponsive to the agonist. Increasing the wash periods between TMPPAA applications from 2 min to up to 5 min did not revoke or diminish this run-down (data not shown). In contrast, a similar current run-down was not observed using 5-HT as agonist. Repeated applications of a saturating 5-HT concentration (100 lM) yielded currents characterized by stable peak amplitudes and unchanged kinetics (data not shown). Interestingly, h5-HT3AR-expressing cells rendered non-responsive to TMPPAA by repeated applications of this agonist remained sensitive to 5-HT, as 5-HT (100 lM) produced robust currents that were comparable in peak amplitudes and kinetics to those recorded in naïve cells (Fig. 3A). Moreover, a single application of 5-HT (100 lM) on h5-HT3ARs made non-responsive to TMPPAA appeared to ‘‘re-sensitize” the receptors to TMPPAA, as this agonist subsequently was able to elicit robust currents in these cells again (data not shown). The apparent ability of 5-HT to ‘‘cancel out” the run-down effect of TMPPAA was also reflected in the complete lack of run-down observed upon repeated applications of 30 lM TMPPAA together with a trace concentration of 5-HT (0.3 lM) (Fig. 3B). 3.1.4. Properties of TMPPAA as a 5-HT3AR PAM To investigate whether TMPPAA in addition to its agonist activity also possessed PAM activity at the h5-HT3AR, we assessed whether the 5-HT-mediated receptor response was potentiated by the presence of a concentration of TMPPAA below those

Fig. 3. The progressive run-down of TMPPAA-evoked currents in h5-HT3AR-expressing COS-7 cells. In all current traces in the figure, the vertical bars represent 200 pA and the horizontal bars 5 s. (A) The progressive run-down of TMPPAA-evoked currents in h5-HT3AR-expressing COS-7 cells. Upon repeated applications of 300 lM TMPPAA (one every 2 min), peak current amplitudes decreased and current kinetics changed gradually, until TMPPAA application did not induce a significant response. Immediately after these applications, three consecutive applications of 100 lM 5-HT evoked currents characterized by stable peak amplitudes and kinetics. (B) The presence of a trace concentration of 5-HT prevents run-down of TMPPAA-evoked currents in h5-HT3AR-expressing COS-7 cells. Repeated co-applications of 30 lM TMPPAA and 0.3 lM 5-HT (one every 1 min) did not result in current run-down. The traces in A and B are from representative recordings from a total of 3 to 5 recordings from cells from 2 different transfections on 3 different days.

98

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

mediating direct activation. Whereas application of 10 lM TMPPAA on its own did not elicit significant currents in h5-HT3AR-expressing oocytes (Figs. 2D and 4A), preincubation and subsequent co-application of this TMPPAA concentration with 1 lM 5-HT (EC10) yielded currents characterized by significantly higher peak amplitudes than those mediated by 5-HT alone (Fig. 4A). This augmentation of the 5-HT-evoked currents was too substantial to be ascribed to simple addition of the negligible agonist activity of 10 lM TMPPAA on top of the 5-HT response, suggesting that TMPPAA in addition to its intrinsic activity also potentiates the 5-HT3AR signalling mediated by the orthosteric agonist. In a reverse experiment, it was investigated whether the TMPPAA-induced h5-HT3AR signalling could be modulated by a trace concentration of 5-HT. Whereas 0.3 lM 5-HT in itself did not elicit a significant agonist response in h5-HT3AR-expressing COS-7 cells, co-application of 0.3 lM 5-HT with increasing TMPPAA concentrations resulted in a left-shifted concentration– response curve for the allosteric agonist, the potency of TMPPAA in the presence of 0.3 lM 5-HT being higher than that for the agonist alone [EC50 = 4.2 lM (n = 3–4 for each data point) vs. EC50 = 7.1 lM (n = 3–5 for each data point)] (Fig. 4B). This modulation was subtle, but bearing in mind the trace concentration of 5-HT used in the recordings this potentiation arises from a low degree of 5-HT occupancy of the orthosteric sites in the receptors. Thus, this suggests that 5-HT acts as a potentiator of TMPPAAmediated h5-HT3AR signalling.

3.2. Inhibition of 5-HT- and TMPPAA-mediated 5-HT3AR signalling To elucidate whether the signals elicited by 5-HT and TMPPAA at 5-HT3ARs differ in terms of their sensitivity towards inhibition, the functional properties of various reference 5-HT3R antagonists were characterized at a stable h5-HT3AR-HEK293 cell line in the Ca2+/Fluo-4 assay using EC80 concentrations of the two agonists. The seven 5-HT3R antagonists used in these studies were the orthosteric antagonists ondansetron, (±)-zacopride, clozapine and metoclopramide and the allosteric antagonists PU02, propofol and picrotoxin. Ondansetron and (±)-zacopride both displayed IC50 values around 1 nM at the 5-HT EC80-induced h5-HT3AR signalling, whereas clozapine was a 1000-fold weaker antagonist (Fig. 5A, Table 1). These values are in good agreement with previously reported binding affinities and antagonist potencies for the compounds [19,20,25,36]. Metoclopramide exhibited an IC50 of 4.0 lM, which is in the high end of the considerable range of functional IC50 values displayed by the antagonist at 5-HT3ARs in previous studies [31,32,37,38]. Finally, the 5-HT3R NAMs PU02 and propofol displayed antagonist potencies in agreement with previously reported values [26,39,40], whereas the channel blocker picrotoxin displayed a 10-fold lower IC50 value at the h5-HT3AR in this assay than those reported from previous studies in oocytes [41,42]. When TMPPAA EC80 was used as agonist, ondansetron, (±)-zacopride and the three allosteric antagonists inhibited

Fig. 4. Electrophysiological characterization of 5-HT and TMPPAA as positive modulators at the h5-HT3AR expressed in Xenopus oocytes and in COS-7 cells. (A) Positive allosteric modulation of 5-HT-evoked currents in h5-HT3AR-expressing oocytes by a trace concentration of TMPPAA. Representative currents (left) and all recorded current amplitudes (right) for the potentiation mediated by 10 lM TMPPAA at the response evoked by 1 lM 5-HT (EC10) through the h5-HT3AR. In the current trace, the vertical bar represents 200 nA and the horizontal bar 30 s. The current amplitudes are given (with S.E.M. values) in % of the current amplitude evoked by 30 lM 5-HT (EC90) and are based on n = 4 from 2 different injections on 3 different days. The control represents the response evoked by 1 lM 5-HT on its own. (B) Positive modulation of TMPPAAevoked currents in h5-HT3AR-expressing COS-7 cells by a trace concentration of 5-HT. Representative currents recorded for various concentrations of TMPPAA co-applied with 0.3 lM 5-HT (left) and averaged concentration–response curves (right) for TMPPAA alone (hatched curve, from Fig. 2B) and in the presence of 0.3 lM 5-HT. In the current trace, the vertical bar represents 200 pA and the horizontal bar 5 s. Data represent mean ± S.E.M. of n = 3–5 from 2 different transfections on 3 different days.

99

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Fig. 5. The inhibition of 5-HT- and TMPPAA-evoked signalling through the h5-HT3AR exerted by ondansetron and metoclopramide. The figures depict representative data given as mean ± S.D. values based on duplicate determination. (A) Concentration–inhibition curves for ondansetron and metoclopramide at 5-HT EC80- and TMPPAA EC80mediated responses in stable h5-HT3AR-HEK293 cells in the Ca2+/Fluo-4 assay. (B) Concentration–inhibition curves for ondansetron and metoclopramide at 5-HT EC80- and TMPPAA EC80-mediated responses in tsA201 cells transiently expressing the h5-HT3AR in the FMP assay.

Table 1 Properties of seven reference 5-HT3R antagonists at 5-HT- and TMPPAA-mediated h5-HT3AR signalling in two fluorescence-based functional assays. The Ca2+/Fluo-4 assay and the FMP assay were performed using a stable h5-HT3AR-HEK293 cell line and tsA201 cells transiently expressing h5-HT3AR, respectively. EC80 (EC70–90) concentrations of 5-HT and TMPPAA (determined on the day of experiments) were used as agonists. The IC50 values (in lM with pEC50 ± S.E.M. values in brackets) obtained for each antagonist using 5-HT and TMPPAA as agonists are given together with the ratio between the two. Data from the Ca2+/Fluo-4 assay represent mean ± S.E.M. of 3 individual experiments, whereas data from the FMP assay represent mean ± S.E.M. of 6 (ondansetron) or 4 (metoclopramide) individual experiments performed in duplicate as described in Section 2. ***P < 0.001.

a

5-HT IC50 (lM) [pIC50 ± S.E.M.]

TMPPAA IC50 (lM) [pIC50 ± S.E.M.]

Ratio ICTMPPAA /IC5-HT 50 50

Ca2+/Fluo-4 assay Ondansetron (±)-Zacopride Metoclopramide Clozapine PU02 Propofol Picrotoxin

0.0018 [8.73 ± 0.20] 0.0016 [8.80 ± 0.13] 4.0 [5.40 ± 0.10] 1.1 [5.94 ± 0.02] 1.7 [5.76 ± 0.13] 40 [4.40 ± 0.09] 2.3 [5.64 ± 0.04]

0.00091 [9.04 ± 0.12] 0.0020 [8.69 ± 0.11] 160 [3.79 ± 0.04] 6.9 [5.16 ± 0.06] 1.5 [5.84 ± 0.18] 78 [4.11 ± 0.05] 2.3 [5.64 ± 0.11]

0.51 1.3 40 6.3 0.88 2.0 1.0

FMP assay Ondansetron Metoclopramide

0.0014 [8.85 ± 0.05] 1.4 [5.83 ± 0.02]

0.0027 [8.59 ± 0.04] 500 [3.3]a

0.51 360

⁄⁄⁄ ⁄⁄⁄

The concentration–inhibition curve was not complete. The IC50 value is estimated from the fitted curve.

h5-HT3AR signalling with IC50 values that were not significantly different from their respective IC50 values in the 5-HT experiments (Table 1). In contrast, clozapine and in particular metoclopramide displayed significantly higher IC50 values at the TMPPAA-evoked signalling than that of 5-HT (Fig. 5A, Table 1). This difference was verified by additional functional characterization of ondansetron and metoclopramide at the h5-HT3AR transiently expressed in tsA201 cells in the FMP assay. Here, ondansetron displayed comparable IC50 values at the 5-HT EC80- and TMPPAA EC80-mediated

responses, whereas metoclopramide was >300-fold more potent as antagonist of the 5-HT response than of the TMPPAA response (Table 1, Fig. 5B). 3.3. Delineation of the molecular basis for TMPPAA activity at the 5-HT3AR The different signalling mediated by 5-HT and TMPPAA via distinct binding sites in the 5-HT3AR prompted us to investigate

100

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Fig. 6. Functional characterization of TMPPAA at the chimeric ra7/m5-HT3A receptor. (A) Topology of the ra7/m5-HT3A chimera. (B) Representative currents (top) and all recorded current amplitudes (bottom) of the inhibition mediated by 10 lM and 100 lM TMPPAA of the response evoked by 30 lM ACh (EC10) in oocytes expressing the ha7 nAChR. In the current trace, the vertical bar represents 200 nA and the horizontal bar 30 s. The current amplitudes are given (with S.E.M. values) in % of the current amplitude evoked 30 lM ACh (control) and are based on n = 3–6 from 2 different injections on 2 different days. (C) Representative currents for the potentiation exerted by 100 lM TMPPAA at the response evoked by 30 lM ACh (EC10) in oocytes expressing the ra7/m5-HT3A receptor using ‘‘equilibrium application” (left) and ‘‘closed-channel application” (right) modes. In the current traces, the vertical bars represent 200 nA and the horizontal bars 30 s. (D) All recorded currents of the potentiation exerted by 100 lM TMPPAA at the response evoked by 30 lM ACh (EC10) in oocytes expressing the ra7/m5-HT3A receptor using ‘‘equilibrium application” (light-grey squares) and ‘‘closed-channel application” (dark-grey circles) modes. The current amplitudes are given (with S.E.M. values) in % of the current amplitude evoked by 1 mM ACh (n = 3). The control represents the response evoked by 30 lM ACh on its own. (E) Representative concentration–response curves for TMPPAA applied alone or together with 3 lM ACh (EC10) at tsA201 cells transiently expressing the ra7/m5-HT3A receptor in the FMP assay. Data are given as mean ± S.D. values based on duplicate determinations from a representative experiment out of a total of 3 experiments.

the molecular basis for the TMPPAA activity. This was done in a three-step process, where the functional properties of TMPPAA were sequentially characterized at a chimeric ra7/m5-HT3A receptor, at chimeric h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors, and at mutant h5-HT3ARs and m5-HT3ARs.

3.3.1. Functional properties of TMPPAA at the ra7/m5-HT3A receptor Initially, the functionality of TMPPAA was characterized at the receptors formed by the classical ra7/m5-HT3A subunit chimera comprised by the ra7 ECD and the m5-HT3A TMD/ICD (Fig. 6A) [27]. However, in order to be able to interpret the functionality

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

displayed by TMPPAA at this receptor, we first characterized its pharmacological properties at the WT a7 nAChR. TMPPAA (100 lM) did not evoke significant currents through ha7 nAChRs expressed in Xenopus oocytes, nor did it potentiate the currents mediated by ACh EC10 (Fig. 6B). Instead, TMPPAA inhibited the ACh EC10-evoked responses with an IC50 of approximately 10 lM (Fig. 6B). This antagonist activity contrasts our previous findings for TMPPAA at the a7 receptor, as the compound was found to be inactive at concentrations up to 300 lM when tested for agonist, PAM and antagonist activity at a stable ha7-GH3 cell line in the Ca2+/Fluo-4 assay [20]. This discrepancy could be rooted in the different expression systems, but the use of a buffer containing genistein (a tyrosine kinase inhibitor that potentiates a7 signalling [43]) in the Ca2+/Fluo-4 assay may also have disrupted or somehow masked the inhibitory activity of the compound at the receptor. In any case, we believe that the antagonism displayed by TMPPAA at the a7 nAChR in the TEVC recordings is a true reflection of its pharmacology at the receptor, and that TMPPAA analogously to numerous other ligands thus exhibits dual a7/5-HT3R activity [26,44,45]. In concordance with previous studies of the ra7/m5-HT3A receptor expressed in Xenopus oocytes [27,46], ACh exhibited an EC50 of 65 lM (pEC50 ± S.E.M.: 4.19 ± 0.04, n = 3) at the chimeric receptor (data not shown). In contrast, TMPPAA did not evoke significant currents in ra7/m5-HT3A-expressing oocytes at concentrations up to 100 lM (Fig. 6C). However, preincubation and subsequent co-application of 100 lM TMPPAA with 30 lM ACh (EC10) at the receptor (‘‘equilibrium application”) resulted in a significantly enhanced response compared to that induced by 30 lM ACh alone (Fig. 6C and D). Preincubation with 100 lM TMPPAA followed by application of 30 lM ACh alone (‘‘closed channel application”) resulted in a considerably higher degree of potentiation of the ACh-mediated response (Fig. 6C and D). The functionality displayed by TMPPAA at these oocytes was supported by studies of the ra7/m5-HT3A receptor transiently transfected in tsA201 cells in the FMP assay. Here, TMPPAA displayed significant intrinsic agonist activity at the chimeric receptor, and coapplication of TMPPAA and 3 lM ACh (EC10) at the cells significantly augmented the TMPPAA-induced responses, thus further supporting the existence of an allosteric link between the ACh and TMPPAA binding sites in the chimeric receptor (Fig. 6E). Interpretations of the PAM and ago-PAM properties displayed by TMPPAA at the a7/m5-HT3A receptor in the oocyte recordings and in the FMP assay, respectively, are obviously complicated by the fact that TMPPAA possesses activity at both receptors making up this chimera. Whereas it seems highly unlikely that TMPPAA would exert its ago-PAM activity at 5-HT3AR and its antagonist activity at a7 through different sites in the two receptors, TMPPAA could be speculated to mediate its ago-PAM/PAM effects at the chimeric receptor as well as its a7 antagonism through a site in the a7 ECD part in both receptors. However, considering that the ago-PAM properties of TMPPAA at the 5-HT3AR are retained in a chimera comprising the TMD/ICD of this receptor, we propose that TMPPAA is much more likely to act through this part of the chimeric receptor. In this scenario, the TMPPAA-mediated inhibition of the a7 nAChR would also be mediated through the TMD/ICD of this receptor. 3.3.2. Functional properties of TMPPAA at the h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors In another line of experiments 5-HT and TMPPAA were characterized functionally at receptors formed by the h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A chimeras that analogous to ra7/m5-HT3A consist of the ECD of one subunit and the TMD/ICD of the other (Fig. 7A). In concordance with its 3-fold higher potency at h5HT3AR than at m5-HT3AR in the patch-clamp recordings (Fig. 2B and C), TMPPAA exhibited a 10-fold difference in its

101

potencies at h5-HT3AR and m5-HT3AR transiently expressed in tsA201 cells in the FMP assay (Fig. 7B, Table 2). In contrast, 5-HT displayed similar EC50 values at the two receptors in the assay, and the equipotent activity of 5-HT at the receptors was thus used as reference points in our evaluation of the functional properties displayed by TMPPAA at the chimeric receptors. The potencies displayed by TMPPAA at the chimeric h5-HT3A/ m5-HT3A and m5-HT3A/h5-HT3A receptors correlated completely with the identity of the TMDs/ICDs of the receptors. The compound was an equally weak agonist at WT m5-HT3A and h5-HT3A/m5HT3A receptors, just as its EC50 values at the WT h5-HT3A and m5-HT3A/h5-HT3A receptors did not differ significantly (Table 2, Fig. 7B). Although this pattern of TMPPAA potencies at the four receptors was very clear-cut from an isolated perspective, it should be noted that the EC50 values displayed by 5-HT at the two chimeric receptors were changed in the same directions compared to the WT receptors, albeit to smaller degrees than those observed for TMPPAA (Table 2, Fig. 7B). To investigate whether the different functional properties displayed by 5-HT and TMPPAA at the four receptors could be ascribed to different expression levels of the receptors, we quantified the total and cell surface expression levels of the receptors formed by myc-tagged versions of the four subunits in tsA201 cells by ELISA. Importantly, the functionalities displayed by 5-HT and TMPPAA at myc-h5-HT3ARs and myc-m5-HT3ARs in the tsA201 cells were indistinguishable from those at the untagged WT h5HT3AR and WT m5-HT3AR (data not shown). As can be seen in Fig. 7C, both total expression and the cell surface expression levels determined for the myc-h5-HT3A and myc-m5-HT3A receptors in the ELISA were comparable, whereas the expression levels of myc-h5-HT3A/m5-HT3A and myc-m5-HT3A/h5-HT3A receptors were somewhat lower and higher, respectively, than the levels for the two myc-tagged WT receptors. These differences in expression levels combined with the small changes in 5-HT EC50 values observed for the two chimeric receptors complicated the interpretations of the results obtained by this approach somewhat. 3.3.3. Functional properties of TMPPAA at mutant h5-HT3ARs and m5HT3ARs In the final phase of the delineation process, we searched within the 5-HT3A TMD and ICD for putative molecular determinants responsible for the differential TMPPPAA potency at the h5-HT3A and m5-HT3A receptors (Figs. 2B, C and 7B). Although the intracellular loop 2 (ICL2) is the TMD/ICD region displaying the highest diversity between the h5-HT3A and m5HT3A subunits, the intracellular location of residues in this loop makes them highly unlikely contributors to TMPPAA binding (Fig. 8A). Nevertheless, to probe for putative contributions from this loop to TMPPAA activity, we characterized 5-HT and TMPPAA at a chimeric h5-HT3Am5-HT3A-ICL2 receptor assembled from a h5HT3A subunit in which the ICL2 was replaced by the corresponding loop in m5-HT3A (Fig. 8A). Neither 5-HT nor TMPPAA displayed substantially different potencies at the h5-HT3Am5-HT3A-ICL2 receptor compared to WT h5-HT3AR (Table 3), and thus the 5-HT3A ICD does not seem to contain molecular determinants of the observed species difference in TMPPAA potency at the two WT receptors. Next we took advantage of the high degree of conservation between the h5-HT3A and m5-HT3A TMDs and substituted the 11 non-conserved residues in the h5-HT3A TMD with the corresponding m5-HT3A residues (Fig. 8A). Whereas introduction of the V246A mutation in the h5-HT3AR had detrimental effects on both 5-HT- and TMPPAA-mediated signalling and thus appeared to disrupt general receptor function, the 10 other mutant receptors were functional. Moreover, in contrast to the somewhat ambiguous data obtained for the chimeric h5-HT3A/m5-HT3A and

102

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Fig. 7. Functional characterization of TMPPAA at the h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors. (A) Topologies of the h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A chimeras. (B) Concentration–response curves for 5-HT (left) and TMPPAA (right) at the WT h5-HT3A, WT m5-HT3A, h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors transiently expressed in tsA201 cells in the FMP assay. Both graphs depict data given as mean ± S.D. values based on duplicate determinations from a representative experiment out of a total of 3 experiments. (C) Total expression (left) and cell surface expression (right) levels of myc-tagged versions of WT h5-HT3A, h5-HT3A/m5-HT3A, h5HT3AM470T, WT m5-HT3A, m5-HT3A/h5-HT3A and m5-HT3AT481M receptors transiently expressed in tsA201 cells determined by ELISA. All absorbance was background corrected using tsA201 cells transfected with untagged WT h5-HT3A. Data are given as mean ± S.E.M. values of two independent experiments performed in triplicate.

m5-HT3A/h5-HT3A receptors a clear-cut picture emerged from the functional profiling of 5-HT and TMPPAA at these mutant receptors. In concordance with its equipotent agonism at WT h5HT3AR and WT m5-HT3AR, none of the EC50 values exhibited by 5-HT at 9 of these mutants (V237I, M259V, Y265C, N269D, S447Y, K451R, H455R, Q486H and A488S) differed substantially from that at WT h5-HT3AR (Table 3). Analogously, the potencies displayed by TMPPAA at these 9 mutants did not differ substantially from that at the WT receptor (Table 3). It should be noted that statistical analysis did find a few of the 5-HT and TMPPAA

potencies at these 9 mutants to differ significantly from that at WT h5-HT3AR, but we do not consider these minute differences to be pharmacologically pertinent. In contrast to these mutants, substitution of the Met470 residue located in the extracellular part of TM4 in h5-HT3A with a Thr residue resulted in a substantially right-shifted TMPPAA concentration-relationship at h5-HT3AR, so much so that its potency at the h5-HT3AM470T receptor was comparable to that at WT m5-HT3AR (Fig. 8B, Table 3). Conversely, introduction of the reverse T481M mutation in m5-HT3AR left-shifted the concentration-relationship of TMPPAA significantly, albeit its

103

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Table 2 Agonist properties of 5-HT and TMPPAA at WT h5-HT3A, WT m5-HT3A and chimeric h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors transiently expressed in tsA201 cells in the FMP assay. EC50 values (in lM with pEC50 ± S.E.M. values in brackets), Hill slopes and the maximal responses (Rmax) (in % of the 5-HT Rmax at the relevant WT receptor measured at the same plate: h5-HT3A/m5-HT3A data vs. 5-HT Rmax at WT h5-HT3AR, and m5-HT3A/h5-HT3A data vs. 5-HT Rmax at WT m5-HT3AR) are given. Data represent mean ± S.E.M. of 3 individual experiments performed in duplicate as described in Section 2. Statistical analysis has been performed on the pEC50 values for 5-HT and TMPPAA at h5-HT3A/m5-HT3A (against WT h5-HT3AR) and at m5-HT3A/h5-HT3A data (against WT m5-HT3AR). ***P < 0.001. 5-HT

WT h5-HT3A WT m5-HT3A h5-HT3A/m5-HT3A m5-HT3A/h5-HT3A

TMPPAA

EC50 (lM) [pEC50 ± S.E.M.]

nH ± S.E.M.

Rmax ± S.E.M. (%)

EC50 (lM) [pEC50 ± S.E.M.]

nH ± S.E.M.

Rmax ± S.E.M. (%)

0.25 0.28 0.85 0.12

3.1 ± 0.1 3.4 ± 0.3 3.2 ± 0.2 3.8 ± 0.3

100 100 94 ± 6 98 ± 6

7.8 [5.11 ± 0.03] 100a 100a 18 [4.75 ± 0.11]

3.3 ± 0.2 n.d. n.d. 2.5 ± 0.4

115 ± 4 46 ± 1a 59 ± 5a 112 ± 4

[6.60 ± 0.02] [6.56 ± 0.02] [6.07 ± 0.03]⁄⁄⁄ [6.92 ± 0.02]⁄⁄⁄

n.d., not determinable. a The TMPPAA concentration–response curve at this receptor was not complete. The EC50 value is estimated from the fitted curves based on the assumption that the maximal response for TMPPAA is roughly comparable to those for 5-HT at the receptor. The values listed in the Rmax column are the response evoked by the highest TMPPAA concentration normalized to the 5-HT Rmax values at the receptor.

potency at the m5-HT3AT481M receptor still was 4-fold lower than that at the WT h5-HT3AR (Table 3, Fig. 8B). Importantly, the EC50 values displayed by 5-HT at the h5-HT3AM470T and m5-HT3AT481M receptors did not differ significantly from those at either of the two WT receptors (Table 3, Fig. 8B). Moreover, the total expression and cell surface expression levels exhibited by the myc-h5HT3AM470T and myc-m5-HT3AT481M receptors in the tsA201 cells were not substantially different from those exhibited by myc-h5HT3AR and myc-m5-HT3AR, respectively (Fig. 7C). Hence, the different functionalities displayed by TMPPAA at the four receptors cannot be ascribed to differences in expression levels. The importance of the Met470/Thr481 residue for the agonist properties of TMPPAA at the 5-HT3AR was further probed by functional characterization of 5-HT and TMPPAA at six additional h5-HT3AM470X receptor mutants. 5-HT was essentially equipotent as agonist at all six of these mutants compared to the WT receptor (Fig. 8C, Table 3). As for TMPPAA, conservative mutations of Met470 in the form of aliphatic residues (Ala, Val, Leu and Ile) resulted in modestly increased EC50 values (2-fold), and introduction of a Ser residue in this position significantly decreased its potency by 3-fold. Conversely, substitution of Met470 with the aromatic residue Phe resulted in a significant (3-fold) left-shift of the concentration–response relationship for TMPPAA at the receptor (Fig. 8C, Table 3). Thus, the original demonstration of the Met470/Thr481 residue as a key molecular determinant of the species difference in TMPPAA potency was further substantiated by the effects arising from these six additional mutations of Met470 in h5-HT3AR. 4. Discussion The complexity of Cys-loop receptor function means not only that the receptors can be modulated or activated via multiple allosteric sites but also that the modulation/activation induced through these sites can differ dramatically depending on the specific allosteric transitions affected in the receptors [4,14,16]. Several ago-PAMs have been identified for GABAARs, nAChRs and GlyRs [47–52], and two recent studies have identified the monoterpenes carvacrol and thymol as ago-PAMs of 5-HT3Rs [53,54]. Following our previous identification of a series of putative 5-HT3R ago-PAMs [20], we here present the functional properties and the molecular basis for the activity of one of these compounds, TMPPAA. 4.1. Functional characteristics of TMPPAA at the 5-HT3AR In electrophysiological recordings TMPPAA was found to be a highly-efficacious partial agonist with similar potency as 5-HT at the h5-HT3AR in COS-7 cells, whereas it exhibited a somewhat

lower potency at the receptor in oocytes. However, these studies also revealed that 5-HT and TMPPAA mediate very different forms of 5-HT3AR activation (Figs. 2 and 3). While a partial agonist in terms of its maximum peak current amplitude, TMPPAA would be expected to mediate the influx of much higher total current charge into the cell over time than 5-HT. On the other hand, the progressive run-down of current amplitudes and the resulting non-responsiveness of the receptor to TMPPAA observed upon repeated applications of the agonist seem to suggest that the longer periods of gating arising from the slow onset of desensitization could be followed by a prolonged period of inactivity. The ability of 5-HT to nullify this run-down of TMPPAA-evoked currents is interesting, as it suggests that the presence of even small concentrations of the endogenous agonist could be important for the effects of TMPPAA in native tissues. While the mechanisms underlying the distinct kinetics of TMPPAA as a 5-HT3R agonist admittedly cannot be unequivocally deduced from the recordings in this study, possible explanations for its signalling characteristics are presented in Fig. 9. As outlined in Section 3, we propose that different rise kinetics of the currents evoked by sub-maximal TMPPAA and 5-HT concentrations are based in slower diffusion and/or slower binding kinetics of TMPPAA to the receptor. It certainly seems plausible that the proposed transmembrane TMPPAA site would be less accessible than the orthosteric site, and that it thus would take a longer time for the allosteric agonist to reach its binding site than 5-HT. The slower desensitization kinetics of the TMPPAA currents indicate that the agonist stabilizes an active receptor conformation less susceptible to undergo desensitization than that stabilized by 5-HT and/or that the TMPPAA-bound desensitized receptor is much more disposed to undergo reactivation than the 5-HT-bound desensitized receptor (Fig. 9). As for the progressive current rundown observed upon repeated TMPPAA applications, we speculate that this in part could reflect that the small fraction of receptors desensitized by TMPPAA during each application are unable to recover and return to the resting receptor state, thus remaining non-responsive to TMPPAA in the subsequent applications (Fig. 9). This proposed mechanism for the current run-down would require the slow decay of TMPPAA currents to be rooted in slow desensitization and negligible reactivation of the receptors, but it does not explain the gradual change in current kinetics observed with each TMPPAA application. Finally, the conformational changes underlying the reactivation of these TMPPAAnonresponsive receptors by 5-HT is likely to lead to TMPPAA unbinding and the release of molecular constraints obstructing the spontaneous return of the receptors to resting conformations (Fig. 9). This is somewhat analogous to the resensitization mediated by some PAMs at a7 nAChRs desensitized by orthosteric agonists [15,55].

104

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Fig. 8. Functional characterization of TMPPAA at mutant h5-HT3A and m5-HT3A receptors. (A) Topology of the h5-HT3A TMD/ICD (left) and the amino acid sequences of IL2 in h5-HT3A and m5-HT3A (right). The non-conserved residues between the TMDs in h5-HT3A and m5-HT3A are indicated together with the mutations introduced into h5-HT3A. (B) Concentration–response curves for 5-HT (left) and TMPPAA (right) at the WT h5-HT3A, WT m5-HT3A, h5-HT3AM470T and m5-HT3AT481M receptors transiently expressed in tsA201 cells in the FMP assay. (C) Concentration–response curves for 5-HT (left) and TMPPAA (right) at the WT h5-HT3A and six mutant 5-HT3AM470X receptors transiently expressed in tsA201 cells in the FMP assay. Data in B and C are given as mean ± S.D. values based on duplicate determinations from a representative experiment out a total of 3 experiments.

The slower current rise and decay kinetics of TMPPAA compared to 5-HT at 5-HT3ARs are similar to previously reported differences in kinetic properties for orthosteric agonists and agoPAMs at other Cys-loop receptors [15,50,52,56]. Interestingly, carvacrol- and thymol-mediated activation of h5-HT3ARs expressed in oocytes were reported to be slower than that by 5-HT, whereas the desensitization kinetics of the three agonists were comparable [53]. Although one should be cautious when comparing data from mammalian cells and oocytes, the apparent different signalling characteristics of carvacrol/thymol and TMPPAA are likely to be founded in distinct mechanisms of action. Thus, analogous to the classification of a7 nAChR PAMs into two types based on their impact on receptor desensitization, 5-HT3R ago-PAMs also seem to come in different types [15,55].

In addition to its agonism, TMPPAA also acts as a PAM of the 5-HT-evoked 5-HT3AR signalling, and the analogous ability of 5-HT to potentiate TMPPAA-mediated signalling further demonstrates the allosteric link between the orthosteric site and the TMPPAA site in the receptor (Fig. 4). This reciprocal potentiation of each other as agonists is in good agreement with previous studies of the interplay between orthosteric and allosteric agonists at nAChRs and GABAARs [52,57], and the allosteric linkage of the TMPPAA site to the orthosteric site as well as to other allosteric sites in the h5-HT3AR complex is further evident from the abilities of various orthosteric and allosteric 5-HT3R antagonists to inhibit TMPPAA-mediated signalling (Fig. 5, Table 1). It is by no means a given that an orthosteric antagonist will inhibit the signalling evoked by an allosteric agonist, as this depends on the allosteric link between the two sites and on

105

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

Table 3 Agonist properties of 5-HT and TMPPAA at WT and mutant h5-HT3ARs and at WT and mutant m5-HT3ARs transiently expressed in tsA201 cells in the FMP assay. EC50 values (in lM with pEC50 ± S.E.M. values in brackets), Hill slopes and the maximal responses (Rmax) of TMPPAA (in % of 5-HT Rmax at the WT receptor measured at the same plate) are given. Unless otherwise indicated, data representmean ± S.E.M. of 3 individual experiments performed in duplicate as described in Section 2. Statistical analysis has only been performed on the pEC50 values for 5-HT and TMPPAA, and here the respective mutants have been compared with the WT receptor based on data from experiments where both receptors were included (n = 3). *P < 0.025, **P < 0.005. 5-HT

TMPPAA

EC50 (lM) [pEC50 ± S.E.M.]

nH ± S.E.M.

Rmax ± S.E.M. (%)

EC50 (lM) [pEC50 ± S.E.M.]

nH ± S.E.M.

Rmax ± S.E.M. (%)

h5-HT3ARs WTa h5-HT3Am5-HT3A-ICL2 V237I V246Ab M259V Y265C N269D S447Y K451R H455R Q486H A488S M470T M470A M470V M470L M470I M470S M470F

0.26 0.39 0.39 n.d. 0.38 0.25 0.26 0.12 0.26 0.32 0.27 0.30 0.41 0.35 0.33 0.27 0.26 0.35 0.23

3.5 ± 0.2 5.5 ± 0.4 4.1 ± 0.1 n.d. 3.9 ± 1.1 3.5 ± 0.2 2.9 ± 0.2 3.1 ± 0.2 3.8 ± 0.3 3.7 ± 0.5 3.9 ± 0.3 3.5 ± 0.4 4.1 ± 0.2 3.9 ± 0.5 4.0 ± 0.4 3.5 ± 0.2 3.5 ± 0.1 3.9 ± 0.3 3.2 ± 0.1

100 121 ± 21 89 ± 7 n.d. 167 ± 16 96 ± 13 139 ± 3 71 ± 10 72 ± 11 60 ± 12 112 ± 14 105 ± 11 113 ± 3 120 ± 21 138 ± 27 103 ± 16 105 ± 17 130 ± 25 130 ± 11

9.1 [5.04 ± 0.02] 17 [4.78 ± 0.03]⁄ 13 [4.89 ± 0.02] n.d. 6.3 [5.20 ± 0.02]⁄⁄ 6.8 [5.17 ± 0.01]⁄

12 [4.93 ± 0.03] 17 [4.76 ± 0.03]⁄ 30 [4.52 ± 0.06]⁄⁄ 2.8 [5.56 ± 0.04]⁄⁄

2.8 ± 0.2 3.1 ± 0.3 2.6 ± 0.1 n.d. 3.6 ± 0.2 2.8 ± 0.4 2.7 ± 0.1 1.9 ± 0.2 2.4 ± 0.1 2.8 ± 0.1 2.9 ± 0.4 2.4 ± 0.4 n.d. 3.1 ± 0.1 3.1 ± 0.2 2.8 ± 0.3 3.3 ± 0.6 3.2 ± 0.4 2.5 ± 0.1

127 ± 3 133 ± 23 94 ± 8 n.d. 165 ± 22 96 ± 12 144 ± 2 72 ± 11 67 ± 8 58 ± 9 107 ± 14 113 ± 10 92 ± 4c 131 ± 20 151 ± 28 102 ± 16 113 ± 17 146 ± 27 126 ± 15

m5-HT3ARs WT T481M

0.29 [6.53 ± 0.04] 0.26 [6.58 ± 0.06]

4.0 ± 0.1 4.0 ± 0.2

100 107 ± 10

100c 25 [4.61 ± 0.03]

n.d. 4.1 ± 0.5

40 ± 9c 117 ± 19

[6.58 ± 0.01] [6.41 ± 0.05] [6.41 ± 0.02] [6.42 ± 0.03] [6.60 ± 0.02] [6.58 ± 0.03] [6.92 ± 0.02]⁄⁄ [6.58 ± 0.05] [6.50 ± 0.05] [6.57 ± 0.05] [6.52 ± 0.04] [6.39 ± 0.01] [6.45 ± 0.02]⁄ [6.48 ± 0.05] [6.57 ± 0.02] [6.59 ± 0.09] [6.45 ± 0.01]⁄⁄ [6.63 ± 0.02]

8.1 [5.09 ± 0.03] 5.2 [5.28 ± 0.02]⁄ 7.9 [5.10 ± 0.03] 15 [4.81 ± 0.07] 12 [4.92 ± 0.06] 9.8 [5.01 ± 0.01] 50 [4.3]c 20 [4.69 ± 0.03]⁄⁄ 23 [4.63 ± 0.04]⁄⁄

n.d., not determinable. a Data for WT h5-HT3AR are from 13 independent experiments. b Both 5-HT and TMPPAA elicited negligible responses in tsA201 cells expressing this receptor. c The TMPPAA concentration–response curve at this receptor was not complete. The EC50 value is estimated from the fitted curves based on the assumption that the maximal response for TMPPAA is roughly comparable to those for 5-HT at the receptor. The values listed in the Rmax column are the response evoked by the highest TMPPAA concentration normalized to the 5-HT Rmax values at the relevant WT receptor (WT h5-HT3AR or WT m5-HT3AR).

the binding kinetics of and the receptor conformations stabilized by the respective ligands [50,52,58,59]. In this study, the markedly different antagonist potencies displayed by metoclopramide at 5-HT- and TMPPAA-mediated h5-HT3AR signalling constituted an interesting outlier compared to the properties exhibited by the other antagonists (Table 1). Although metoclopramide is believed to target the orthosteric site in the 5-HT3R, it and several other orthosteric 5-HT3R antagonists have been reported to exhibit mixed competitive/non-competitive antagonism at the receptors in vitro [60]. Interestingly, metoclopramide has in patch-clamp recordings been found to be a 300-fold less potent antagonist of 5-HT-induced h5-HT3AR responses in an open-channel application (co-application with 5-HT) than in an equilibrium application (preincubation followed by co-application with 5-HT) [32]. In view of the distinct signalling kinetics of 5-HT and TMPPAA, we thus speculate that the differential inhibition exerted by metoclopramide at the h5-HT3AR signalling mediated by the two agonists at least in part could be rooted in different binding affinities of the antagonist to the resting and the TMPPAA-bound active and desensitized conformations of the receptor. However, we are unable to explain why metoclopramide would differ from ondansetron and (±)-zacopride in this respect, and thus this observation should mainly be considered as another indication of the fundamentally different characteristics of the h5-HT3AR signalling evoked by 5-HT and TMPPAA. 4.2. Molecular basis for TMPPAA activity at the 5-HT3AR We propose that the robust ago-PAM/PAM activity displayed by TMPPAA at the ra7/m5-HT3A receptor, its functional properties at the h5-HT3A/m5-HT3A and m5-HT3A/h5-HT3A receptors, and the

almost complete conversion of its different agonist potencies at h5-HT3AR and m5-HT3AR brought on by mutations of a single non-conserved TM4 residue collectively identify the TMD as the 5-HT3AR region targeted by the compound (Figs. 6–8, Tables 2 and 3). Considering that TMPPAA possesses activity at all three WT receptors applied in these studies and the putative involvement of a residue in the extracellular part of TM4 in its binding to 5-HT3AR, TMPPAA could hypothetically act through a site formed by both ECD and TMD residues, analogously to what recently has been proposed for the GABAAR NAM dipicrylamine [61]. However, we find this scenario unlikely for the reasons outlined below. Although binding of some allosteric ligands to nAChRs and GABAARs have been proposed to involve residues located in the extracellular part of TM4 or in the C-termini of the receptors [49,61,62], the importance of the Met470/Thr481 residue for TMPPAA functionality at the 5-HT3AR is not necessarily rooted in a direct involvement in ligand binding. The residue does not align with any of these previously identified residues [49,61,62], and judging from the spatial orientation of Thr481 in a m5-HT3AR crystal structure (Protein Data Bank ID code: 4PIR, [12]) its side chain projects out from the top of TM4 into the membrane surroundings and not towards other TMs in its own or neighbouring subunits. Thus, an obvious putative TMPPAA binding site comprising this residue does not present itself from this structure. Instead, TMPPAA could be envisioned to target a site in another part of the 5-HT3AR TMD (possibly binding with similar affinities to h5-HT3AR and m5-HT3AR), with the different side chains of Met470/ Thr481 exerting their respective effects on its functionality through an allosteric mechanism. Although we will refrain from elaborate speculations about the exact location of this binding site in the

106

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

interfaces in heteromeric receptors [17,63–65], and such interfaces are obviously absent in the homomeric 5-HT3A and a7 receptors. Furthermore, as mentioned above, the slow rise of the currents evoked by sub-maximal TMPPAA concentrations through 5HT3ARs could indicate that the compound targets a not easily accessible binding site, for example one embedded in the TMD of the receptor. The slow desensitization kinetics of TMPPAA currents is also reminiscent of those exhibited by other ago-PAMs and PAMs acting through TMDs in Cys-loop receptors [15,47,50,56,66,67]. Finally, the notion that the differential modulation exhibited by TMPPAA at 5-HT3A and a7 receptors could be mediated through a common allosteric site in the two receptors is supported by the diverse functionalities displayed by close structurally related modulators at other Cys-loop receptors or by specific modulators at different receptor subtypes in previous studies [67,68]. This study has not addressed whether the agonist and PAM activities of TMPPAA at 5-HT3AR are mediated through the same site(s) or via discrete sites as it has been proposed for other Cys-loop receptor ago-PAMs [49,51]. However, the functionality of TMPPAA at the ra7/m5-HT3A receptor suggests that both activity components could originate from the 5-HT3AR TMD. In conclusion, the detailed delineation of the functionality of TMPPAA at the 5-HT3AR in the present study demonstrates just how much the characteristics of the signalling induced by allosteric ligands at Cys-loop receptors can differ from those mediated by orthosteric ligands. Moreover, the distinct ago-PAM properties exhibited by TMPPAA and carvacrol [53] suggest that the nature of the 5-HT3R signalling evoked through different allosteric sites can differ substantially. We find the aberrant signalling characteristics of TMPPAA highly interesting and propose that the compound could make for an interesting pharmacological tool in future explorations of 5-HT3Rs. Furthermore, while this has not been the focus of this study, the a7 NAM activity displayed by TMPPAA is also interesting and could be pursued in future work. Finally, the availability of an arsenal of allosteric ligands with diverse functionalities at 5-HT3Rs combined with the recently demonstrated ability to obtain high-resolution crystal structures of the full-length receptor complex [12] will hopefully lead to an atomic-level insight into the molecular bases for allosteric modulation and activation of these receptors. Fig. 9. Possible mechanistic explanations for the observed different signalling characteristics of 5-HT and TMPPAA at the 5-HT3AR. Simplified scheme outlining the equilibriums between resting [R, white], active [A, grey] and desensitized [D, black] receptor conformations, the five processes underlying the transitions between these, and the effects mediated by 5-HT (bottom of figure) and TMPPAA (top of figure) at them. The possible mechanistic explanations of the distinct signalling characteristics of TMPPAA relative to those of 5-HT are indicated in red. While it is assumed that up to five TMPPAA molecules bind to each 5-HT3AR complex, it should be stressed that the TMPPAA binding sites depicted in this figure are intended for illustrative purposes only, since the specific location of these in the receptor has not been delineated. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

TMD, it is interesting to note that TMPPAA functionality at h5-HT3AR is unaffected by the M259V mutation reported to eliminate agonist activity of carvacrol at the receptor as well as by the S248N and I290N mutations that both render the receptor insensitive to the NAM PU02 (Table 3 and data not shown) [26,53]. Carvacrol and PU02 have been proposed to target different sites in the transmembrane subunit interface in the 5-HT3R complex [26,53], and the insensitivity of TMPPAA to these mutations thus suggests that the compound either acts through another site in this interface or through a site elsewhere in the TMD. The notion of a binding site for TMPPAA in the 5-HT3AR TMD is supported by several observations in the literature. The majority of allosteric modulators demonstrated to target the ECDs of Cys-loop receptors to date appear to act through noncanonical subunit

Acknowledgements The authors thank the Novo Nordisk Foundation and the Lundbeck Foundation for financial support of this study. Drs. Dorte Strøbæk (Saniona A/S) and Rikke Søgaard (University of Copenhagen) are thanked for giving us access to electrophysiology set-ups, and Professor John A. Peters (University of Dundee, UK) is thanked for inspiring discussions about parts of this study. Drs. J. Egebjerg, J.W. Patrick, D. Julius and C. Rojas are thanked for their generous gifts of cDNAs and cell lines. References [1] M. Berger, J.A. Gray, B.L. Roth, The expanded biology of serotonin, Annu. Rev. Med. 60 (2009) 355–366. [2] A.A. Jensen, P.A. Davies, H. Bräuner-Osborne, K. Krzywkowski, 3B but which 3B? And that’s just one of the questions: the heterogeneity of human 5-HT3 receptors, Trends Pharmacol. Sci. 29 (2008) 437–444. [3] N.M. Barnes, T.G. Hales, S.C. Lummis, J.A. Peters, The 5-HT3 receptor – the relationship between structure and function, Neuropharmacology 56 (2009) 273–284. [4] A. Taly, P.J. Corringer, D. Guedin, P. Lestage, J.P. Changeux, Nicotinic receptors: allosteric transitions and therapeutic targets in the nervous system, Nat. Rev. Drug Discov. 8 (2009) 733–750. [5] R.W. Olsen, W. Sieghart, International Union of Pharmacology. LXX. Subtypes of c-aminobutyric acidA receptors: classification on the basis of subunit composition, pharmacology, and function. Update, Pharmacol. Rev. 60 (2008) 243–260.

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108 [6] J.W. Lynch, Native glycine receptor subtypes and their physiological roles, Neuropharmacology 56 (2009) 303–309. [7] P.A. Davies, W. Wang, T.G. Hales, E.F. Kirkness, A novel class of ligand-gated ion channel is activated by Zn2+, J. Biol. Chem. 278 (2003) 712–717. [8] J. Walstab, G. Rappold, B. Niesler, 5-HT3 receptors: role in disease and target of drugs, Pharmacol. Ther. 128 (2010) 146–169. [9] S.P. Alexander, A. Mathie, J.A. Peters, Guide to Receptors and Channels (GRAC), 5th edition, Br. J. Pharmacol. 164 (Suppl. 1) (2011) S1–S324. [10] M. Engel, M.P. Smidt, J.A. van Hooft, The serotonin 5-HT3 receptor: a novel neurodevelopmental target, Front. Cell. Neurosci. 7 (2013) 76. [11] K.N. Browning, Role of central vagal 5-HT3 receptors in gastrointestinal physiology and pathophysiology, Front. Neurosci. 9 (2015) 413. [12] G. Hassaine, C. Deluz, L. Grasso, R. Wyss, M.B. Tol, R. Hovius, et al., X-ray structure of the mouse serotonin 5-HT3 receptor, Nature 512 (2014) 276–281. [13] H.A. Lester, M.I. Dibas, D.S. Dahan, J.F. Leite, D.A. Dougherty, Cys-loop receptors: new twists and turns, Trends Neurosci. 27 (2004) 329–336. [14] P.S. Miller, T.G. Smart, Binding, activation and modulation of Cys-loop receptors, Trends Pharmacol. Sci. 31 (2010) 161–174. [15] D. Bertrand, M. Gopalakrishnan, Allosteric modulation of nicotinic acetylcholine receptors, Biochem. Pharmacol. 74 (2007) 1155–1163. [16] M. Grupe, M. Grunnet, J.F. Bastlund, A.A. Jensen, Targeting a4b2 nicotinic acetylcholine receptors in central nervous system disorders: perspectives on positive allosteric modulation as a therapeutic approach, Basic Clin. Pharmacol. Toxicol. 116 (2015) 187–200. [17] W. Sieghart, Allosteric modulation of GABAA receptors via multiple drugbinding sites, Adv. Pharmacol. 72 (2015) 53–96. [18] P.A. Davies, Allosteric modulation of the 5-HT3 receptor, Curr. Opin. Pharmacol. 11 (2011) 75–80. [19] A.J. Thompson, Recent developments in 5-HT3 receptor pharmacology, Trends Pharmacol. Sci. 34 (2013) 100–109. [20] C.G. Jørgensen, B. Frølund, J. Kehler, A.A. Jensen, Discovery of benzamide analogues as a novel class of 5-HT3 receptor agonists, ChemMedChem 6 (2011) 725–736. [21] C. Rojas, M. Stathis, A.G. Thomas, E.B. Massuda, J. Alt, J. Zhang, et al., Palonosetron exhibits unique molecular interactions with the 5-HT3 receptor, Anesth. Analg. 107 (2008) 469–478. [22] R.M. Horton, H.D. Hunt, S.N. Ho, J.K. Pullen, L.R. Pease, Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension, Gene 77 (1989) 61–68. [23] T.N. Petersen, S. Brunak, G. von Heijne, H. Nielsen, SignalP 4.0: discriminating signal peptides from transmembrane regions, Nat. Methods 8 (2011) 785–786. [24] A.A. Jensen, I. Mikkelsen, B. Frølund, H. Bräuner-Osborne, E. Falch, P. Krogsgaard-Larsen, Carbamoylcholine homologs: novel and potent agonists at neuronal nicotinic acetylcholine receptors, Mol. Pharmacol. 64 (2003) 865– 875. [25] K. Krzywkowski, A.A. Jensen, C.N. Connolly, H. Bräuner-Osborne, Naturally occurring mutations in the human 5-HT3A gene profoundly impact 5-HT3 receptor function and expression, Pharmacogenet. Genomics 17 (2007) 255– 266. [26] S.M. Trattnig, K. Harpsøe, S.B. Thygesen, L.M. Rahr, P.K. Ahring, T. Balle, et al., Discovery of a novel allosteric modulator of 5-HT3 receptors: inhibition and potentiation of Cys-loop receptor signaling through a conserved transmembrane intersubunit site, J. Biol. Chem. 287 (2012) 25241–25254. [27] J.-L. Eiselé, S. Bertrand, J.-L. Galzi, A. Devillers-Thiéry, J.-P. Changeux, D. Bertrand, Chimaeric nicotinic-serotonergic receptor combines distinct ligand binding and channel specificities, Nature 366 (1993) 479–483. [28] O.P. Hamill, A. Marty, B. Sakmann, F.J. Sigworth, Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, Pfluegers Arch. 391 (1981) 85–100. [29] J.K. Christensen, T. Varming, P.K. Ahring, T.D. Jorgensen, E.Ø. Nielsen, In vitro characterization of 5-carboxyl-2,4-di-benzamidobenzoic acid (NS3763), a noncompetitive antagonist of GLUK5 receptors, J. Pharmacol. Exp. Ther. 309 (2004) 1003–1010. [30] M. Barann, G. Molderings, M. Bruss, H. Bonisch, B.W. Urban, M. Gothert, Direct inhibition by cannabinoids of human 5-HT3A receptors: probable involvement of an allosteric modulatory site, Br. J. Pharmacol. 137 (2002) 589–596. [31] A.M. Brown, A.G. Hope, J.J. Lambert, J.A. Peters, Ion permeation and conduction in a human recombinant 5-HT3 receptor subunit (h5-HT3A), J. Physiol. 507 (Pt 3) (1998) 653–665. [32] J. Walkembach, M. Bruss, B.W. Urban, M. Barann, Interactions of metoclopramide and ergotamine with human 5-HT3A receptors and human 5-HT reuptake carriers, Br. J. Pharmacol. 146 (2005) 543–552. [33] D. Kurzwelly, M. Barann, A. Kostanian, S. Combrink, H. Bonisch, M. Gothert, et al., Pharmacological and electrophysiological properties of the naturally occurring Pro391Arg variant of the human 5-HT3A receptor, Pharmacogenetics 14 (2004) 165–172. [34] I.D. Mair, J.J. Lambert, J. Yang, J. Dempster, J.A. Peters, Pharmacological characterization of a rat 5-hydroxytryptamine type3 receptor subunit (r5-HT3A (b)) expressed in Xenopus laevis oocytes, Br. J. Pharmacol. 124 (1998) 1667– 1674. [35] M. Lochner, S.C.R. Lummis, Agonists and antagonists bind to an A-A interface in the heteromeric 5-HT3AB receptor, Biophys. J. 98 (2010) 1494–1502. [36] G. Rammes, B. Eisensamer, U. Ferrari, M. Shapa, G. Gimpl, K. Gilling, et al., Antipsychotic drugs antagonize human serotonin type 3 receptor currents in a noncompetitive manner, Mol. Psychiatry 9 (2004) 846–858.

107

[37] H. Bönisch, M. Barann, J. Graupner, M. Gothert, Characterization of 5-HT3 receptors of N1E-115 neuroblastoma cells by use of the influx of the organic cation [14C]-guanidinium, Br. J. Pharmacol. 108 (1993) 436–442. [38] D.L. Downie, A.G. Hope, J.J. Lambert, J.A. Peters, T.P. Blackburn, B.J. Jones, Pharmacological characterization of the apparent splice variants of the murine 5-HT3 R-A subunit expressed in Xenopus laevis oocytes, Neuropharmacology 33 (1994) 473–482. [39] M. Barann, J.P. Dilger, H. Bonisch, M. Gothert, A. Dybek, B.W. Urban, Inhibition of 5-HT3 receptors by propofol: equilibrium and kinetic measurements, Neuropharmacology 39 (2000) 1064–1074. [40] M. Barann, I. Linden, S. Witten, B.W. Urban, Molecular actions of propofol on human 5-HT3A receptors: enhancement as well as inhibition by closely related phenol derivatives, Anesth. Analg. 106 (2008) 846–857. [41] P. Das, C.L. Bell-Horner, T.K. Machu, G.H. Dillon, The GABAA receptor antagonist picrotoxin inhibits 5-hydroxytryptamine type 3A receptors, Neuropharmacology 44 (2003) 431–438. [42] I.M. Paulsen, I.L. Martin, S.M. Dunn, Isomerization of the proline in the M2–M3 linker is not required for activation of the human 5-HT3A receptor, J. Neurochem. 110 (2009) 870–878. [43] A. Wiesner, C. Fuhrer, Regulation of nicotinic acetylcholine receptors by tyrosine kinases in the peripheral and central nervous system: same players, different roles, Cell. Mol. Life Sci. 63 (2006) 2818–2828. [44] S.C. Lummis, A.J. Thompson, M. Bencherif, H.A. Lester, Varenicline is a potent agonist of the human 5-hydroxytryptamine3 receptor, J. Pharmacol. Exp. Ther. 339 (2011) 125–131. [45] C. Beinat, S.D. Banister, M. Herrera, V. Law, M. Kassiou, The therapeutic potential of a7 nicotinic acetylcholine receptor (a7 nAChR) agonists for the treatment of the cognitive deficits associated with schizophrenia, CNS Drugs 29 (2015) 529–542. [46] M. Castillo, J. Mulet, J.A. Bernal, M. Criado, F. Sala, S. Sala, Improved gating of a chimeric a7-5HT3A receptor upon mutations at the M2–M3 extracellular loop, FEBS Lett. 580 (2006) 256–260. [47] H.J. Feng, R.L. Macdonald, Multiple actions of propofol on abc and abd GABAA receptors, Mol. Pharmacol. 66 (2004) 1517–1524. [48] K.A. Wafford, M.B. van Niel, Q.P. Mab, E. Horridge, M.B. Herd, D.R. Peden, et al., Novel compounds selectively enhance d subunit containing GABAA receptors and increase tonic currents in thalamus, Neuropharmacology 56 (2009) 182– 189. [49] A.M. Hosie, M.E. Wilkins, H.M. da Silva, T.G. Smart, Endogenous neurosteroids regulate GABAA receptors through two discrete transmembrane sites, Nature 444 (2006) 486–489. [50] Q. Shan, J.L. Haddrill, J.W. Lynch, Ivermectin, an unconventional agonist of the glycine receptor chloride channel, J. Biol. Chem. 276 (2001) 12556–12564. [51] N.A. Horenstein, R.L. Papke, A.R. Kulkarni, G.U. Chaturbhuj, C. Stokes, K. Manther, et al., Critical molecular determinants of a7 nicotinic acetylcholine receptor allosteric activation: separation of direct allosteric activation and positive allosteric modulation, J. Biol. Chem. (2016). [52] J.K. Gill, M. Savolainen, G.T. Young, R. Zwart, E. Sher, N.S. Millar, Agonist activation of a7 nicotinic acetylcholine receptors via an allosteric transmembrane site, Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 5867–5872. [53] S.J. Lansdell, C. Sathyaprakash, A. Doward, N.S. Millar, Activation of human 5hydroxytryptamine type 3 receptors via an allosteric transmembrane site, Mol. Pharmacol. 87 (2015) 87–95. [54] P.M. Ziemba, B.S. Schreiner, C. Flegel, R. Herbrechter, T.D. Stark, T. Hofmann, et al., Activation and modulation of recombinantly expressed serotonin receptor type 3A by terpenes and pungent substances, Biochem. Biophys. Res. Commun. 467 (2015) 1090–1096. [55] J.H. Grønlien, M. Hakerud, H. Ween, K. Thorin-Hagene, C.A. Briggs, M. Gopalakrishnan, et al., Distinct profiles of a7 nAChR positive allosteric modulation revealed by structurally diverse chemotypes, Mol. Pharmacol. 72 (2007) 715–724. [56] M.M. Palczynska, M. Jindrichova, A.J. Gibb, N.S. Millar, Activation of a7 nicotinic receptors by orthosteric and allosteric agonists: influence on singlechannel kinetics and conductance, Mol. Pharmacol. 82 (2012) 910–917. [57] P. Li, J.R. Bracamontes, B.D. Manion, S. Mennerick, J.H. Steinbach, A.S. Evers, et al., The neurosteroid 5b-pregnan-3a-ol-20-one enhances actions of etomidate as a positive allosteric modulator of a1b2c2L GABAA receptors, Br. J. Pharmacol. 171 (2014) 5446–5457. [58] N. Karim, J. Curmi, N. Gavande, G.A. Johnston, J.R. Hanrahan, M.L. Tierney, et al., 20 -Methoxy-6-methylflavone: a novel anxiolytic and sedative with subtype selective activating and modulating actions at GABAA receptors, Br. J. Pharmacol. 165 (2012) 880–896. [59] S. Ueno, J. Bracamontes, C. Zorumski, D.S. Weiss, J.H. Steinbach, Bicuculline and gabazine are allosteric inhibitors of channel opening of the GABAA receptor, J. Neurosci. 17 (1997) 625–634. [60] B. Costall, R.J. Naylor, Neuropharmacology of 5-HT3 receptor ligands, in: H.G. Baumgarten, M. Göthert (Eds.), Serotoninergic Neurons and 5-HT Receptors in the CNS, Springer Science Publishers, Berlin, 1999, pp. 409–438. [61] A. Limon, A. Estrada-Mondragon, J.M. Reyes-Ruiz, R. Miledi, Dipicrylamine modulates GABAq1 receptors through interactions with residues in the TM4 and Cys-loop domains, Mol. Pharmacol. (2016). [62] K. Paradiso, J. Zhang, J.H. Steinbach, The C terminus of the human nicotinic a4b2 receptor forms a binding site required for potentiation by an estrogenic steroid, J. Neurosci. 21 (2001) 6561–6568.

108

A. Gasiorek et al. / Biochemical Pharmacology 110–111 (2016) 92–108

[63] S. Seo, J.T. Henry, A.H. Lewis, N. Wang, M.M. Levandoski, The positive allosteric modulator morantel binds at noncanonical subunit interfaces of neuronal nicotinic acetylcholine receptors, J. Neurosci. 29 (2009) 8734–8742. [64] M. Moroni, R. Vijayan, A. Carbone, R. Zwart, P.C. Biggin, I. Bermudez, Nonagonist-binding subunit interfaces confer distinct functional signatures to the alternate stoichiometries of the a4b2 nicotinic receptor: an a4-a4 interface is required for Zn2+ potentiation, J. Neurosci. 28 (2008) 6884–6894. [65] J.A. Olsen, J.S. Kastrup, D. Peters, M. Gajhede, T. Balle, P.K. Ahring, Two distinct allosteric binding sites at a4b2 nicotinic acetylcholine receptors revealed by NS206 and NS9283 give unique insights to binding activity-associated linkage at Cys-loop receptors, J. Biol. Chem. 288 (2013) 35997–36006.

[66] H.J. Feng, M.T. Bianchi, R.L. Macdonald, Pentobarbital differentially modulates a1b3d and a1b3c2L GABAA receptor currents, Mol. Pharmacol. 66 (2004) 988– 1003. [67] J.K. Gill-Thind, P. Dhankher, J.M. D’Oyley, T.D. Sheppard, N.S. Millar, Structurally similar allosteric modulators of a7 nicotinic acetylcholine receptors exhibit five distinct pharmacological effects, J. Biol. Chem. 290 (2015) 3552–3562. [68] H. Hammer, B.M. Bader, C. Ehnert, C. Bundgaard, L. Bunch, K. HoestgaardJensen, et al., A multifaceted GABAA receptor modulator: functional properties and molecular mechanism of action of the sedative-hypnotic and recreational drug methaqualone (Quaalude), Mol. Pharmacol. 88 (2015) 401–420.