Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The 5-hydroxytryptamine (5-HT)₃ receptor is a member of the family of ligand-gated ion channels. It consists of five subunits arranged pseudosymmetrically around an integral cation channel (Boess et al., 1995; Green et al., 1995). Initially, a single subunit was cloned (5-HT_3A; Maricq et al., 1991), and two splice variants have been identified in various species, designated 5-HT_3AL, with an additional five or six amino acids between the third and fourth hydrophobic segments, and 5-HT_3AS, without the insertion (Hope et al., 1993; Isenberg et al., 1993; Uetz et al., 1994; Werner et al., 1994; Belelli et al., 1995; Miquel et al., 1995). The 5-HT_3A cDNA forms functional receptors that, when expressed in various cell lines, exhibit positive cooperativity to 5-HT and show a similar pharmacological profile as native 5-HT₃ receptors (Jackson and Yakel, 1995). Recently, a second subunit has been cloned, designated 5-HT_3B, that forms functional 5-HT₃ receptors when expressed together with 5-HT_3A but not when expressed alone. The hetero-oligomer exhibits distinctive pharmacological properties compared with those found in the homo-oligomer and a markedly increased single-channel conductance (Davies et al., 1999).

The deduced amino acid sequence of the 5-HT_3A receptor protein shows greatest similarity within the ligand-gated ion channel family with the nicotinic acetylcholine (ACh) receptor (nAChR) 7 subunit, with a sequence identity of 31% (55% including conservative substitutions). The construction of a chimera with the amino terminus of nAChR 7 subunit together with the carboxyl terminus of 5-HT_3AL subunit exhibited agonist activation by ACh but not 5-HT (Eisele et al., 1993). This provided definitive evidence that agonist recognition is encoded by the amino terminus, thought to be extracellularly located. Biochemical and site-directed mutagenesis studies of the nAChR have identified three hydrophilic loops within the -subunit amino-terminal domain that are involved in ligand recognition (Changeux et al., 1992; Karlin and Akabas, 1995). In particular, photoaffinity labeling of the tyrosine residue (Y93 in the Torpedo -subunit) in the most amino terminal of these hydrophilic loops, loop 1, with [³H]ACh mustard (Cohen et al., 1991), p-dimethylamino-benzenediazonium fluoroborate (Galzi et al., 1990), and subsequent mutagenic studies (Aylwin and White, 1994a,b; Nowak et al., 1995) have implicated this residue in the recognition of ACh (Fig. 1).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Alignment of deduced amino acid sequences of the 5-HT3A receptor subunit and the -subunits of nAChRs in recognition loop 1. Conserved residues are shown boxed, and the residues referred to in the text are in bold. The C-terminal residues are numbered.

Within the 5-HT_3A receptor sequence, the amino acid triplet NEF aligns with YNN in the Torpedo -subunit and YNS in the 7 sequence (Fig. 1). The markedly different character of this amino acid triplet suggests that this may play an important role in the differential ligand recognition of these two sequences. In an extension of our previous studies in which we carried out mutations of glutamate (E106) of the 5-HT_3AL sequence (Boess et al., 1997), we have now mutated the aromatic amino acid phenylalanine, F107, conservatively to tyrosine (F107Y) or to asparagine (F107N). We characterized the mutants with both radioligand binding and whole-cell patch-clamp studies.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. BSu36I was obtained from Promega (Madison, WI). SalI was obtained from Stratagene (La Jolla, CA). The Altered Sites Mutagenesis Kit was purchased from Promega and included the vector pAlter. The expression vector pRC/CMV was supplied by InVitrogen (San Diego, CA). HEK 293 cells were obtained from American Type Culture Collection (Rockville, MD). TSA201 cells were obtained from COR Therapeutics (San Francisco, CA). Dulbecco's modified Eagle's medium, Hanks' balanced salt solution (HBSS), and penicillin/streptomycin were purchased from BioWhittaker (Walkersville, MD). Fetal bovine serum was purchased from Life Technologies (Burlington, Ontario, Canada). 5-HT creatinine sulfate, N-methyl-5-HT oxalate, and ACh chloride were obtained from Sigma-Aldrich Canada Ltd. (Mississauga, Ontario, Canada). m-Chlorophenylbiguanide (mCPBG), 2-methyl-5-HT, and metoclopramide were obtained from Research Biochemicals International (Natick, MA). [³H]GR65630 (61.4-64.4 Ci/mmol) was obtained from DuPont-New England Nuclear (Boston, MA). [³H](S)-Zacopride (84.0 Ci/mmol) was purchased from Amersham Life Science Inc. (Arlington Heights, IL). Ondansetron was donated by Glaxo (Ware, UK). Renzapride and granisetron were donated by SmithKline Beecham Pharmaceuticals (Harlow, UK). All drugs were prepared in HEPES (10 mM, pH 7.5). Other chemicals and reagents were purchased from Sigma-Aldrich Canada Ltd., BDH (Toronto, Ontario, Canada), and FischerBiotech (Nepean, Ontario, Canada).

Site-Directed Mutagenesis of 5-HT_3AL cDNA. The 5-HT_3AL receptor cDNA from NG108-15 cells (generously provided by Dr. Eric Kawashima, Glaxo Molecular Biology Institute, Geneva, Switzerland) was subcloned into the mutagenesis vector pAlter and the eukaryotic expression vector pRC/CMV. The Altered Sites Mutagenesis Kit (Promega) was used to introduce mutations of F107 to tyrosine (F107Y) and asparagine (F107N) using the following mutagenic oligonucleotides: WT 5'-AGA-CTT-CCC-CAC-GTC-CAC-AAA-CTC-ATT-GAT-GAG-AAT-3', F107Y 5'-AGA-CTT-CCC-CAC-GTC-GAC-ATA-CTC-ATT-GAT-GAG-AAT-3', and F107N 5'-AGA-CTT-CCC-CAC-GTC-GAC-ATT-CTC-ATT-GAT-GAG-AAT-3'.

Transient Expression of cDNAs, Cell Culture, and Membrane Preparation. WT and mutant cDNAs in the eukaryotic expression vector pRC/CMV were transiently expressed in either HEK 293 or the HEK 293-derived cell line TSA201 (Heinzel et al., 1988). TSA201 cells produced increased expression levels compared with HEK 293 cells. The results obtained with WT and mutant receptors expressed in the two cell types were similar, and the data were pooled. HEK 293 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM glutamine. The same medium was used to culture the TSA201 cells with the exception of low glucose (1 g/l). Cells were incubated at 37°C in a humidified 7% CO₂ atmosphere on either 150-or 35-mm-diameter plates.

Radioligand Binding. Membranes were thawed and centrifuged (27,000g, 4°C for 20 min), and the pellets resuspended in ice-cold HEPES (10 mM, pH 7.5) for both [³H]GR65630 and [³H](S)-zacopride binding. Assay tubes contained 800 µl of the HEPES buffer, with or without the competing drug, and 100 µl of radioligand [³H]GR65630 (for competition studies, 0.6 nM; for saturation studies, 0.02-26.7 nM) or [³H](S)-zacopride (for competition studies, 0.5 nM; for saturation studies, 0.02-10 nM). Metoclopramide (300 µM) was used to define specific binding. The assay tubes were preincubated on ice for 2 min before the addition of 100 µl of the membrane suspension, equivalent to approximately 100 µg of protein. After further incubation either on ice for 2 h with [³H]GR65630 or at 37°C for 2 h with [³H](S)-zacopride, the assay was terminated by rapid filtration through Whatman GF/B filters that had been pretreated with 0.3% polyethylenimine in the above HEPES buffer. The filters were then washed with 2 × 5 ml of ice-cold HEPES buffer, and the resultant radioactivity was determined by conventional liquid scintillation counting (Ecoscint; National Diagnostics, Atlanta, GA) at an efficiency of about 60%.

Electrophysiology. Membrane currents were recorded under voltage-clamp from single cells with the whole-cell configuration of the patch-clamp technique using an Axopatch-1D amplifier (Axon Instruments, Foster City, CA). Series resistance compensation was used, and recordings were performed at a holding potential of 60 mV. Data acquisition, storage, and analysis were performed using pClamp 6 (Axon Instruments). Pipettes were pulled from borosilicate glass and had resistances of 2 M when filled with pipette solution consisting of 135 mM CsCl, 0.5 mM MgCl₂, and 10 mM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid tetrapotassium salt, pH 7.2, with CsOH. Ligands were applied to the cells using a gravity-feed rapid perfusion system (half-time of solution change was 100 ms) based on the design of Carbone and Lux (1987). The cell was continually perfused with solution containing 130 mM NaCl, 5 mM KCl, 1.8 mM CaCl₂, 1.2 mM MgCl₂, 10 mM HEPES, and 5 mM glucose, pH 7.4 with NaOH. The solutions were changed by means of a manually operated valve that housed a manifold connected to solution reservoirs. Recordings were performed at room temperature (19-21°C). Cells with phase-bright "inclusion bodies" that were visible with a phase contrast microscope were generally found to express receptors. To allow recovery from desensitization, 3 min was allowed to elapse between agonist applications. Because the current was subject to run-down during the course of recording, data for concentration-effect curves were collected using the following sequence: maximal concentration, test concentration, maximal concentration. Data were accepted for analysis only if the current for the second maximal concentrations did not decline by more than 10%. For concentration-effect curves, data were pooled from different cells.

Data Analysis. Radioligand binding data were analyzed by computer-assisted iterative curve fitting (Kaleidagraph; Synergy Software, Reading, PA) according to the equation: B = B_max[L]ⁿ/([L]ⁿ + Kⁿ), where B is the bound ligand, B_max is the maximum binding at equilibrium, K is the molar equilibrium dissociation constant (K_d) for saturation studies or molar concentration of competing compound to reduce the specific binding to 50% for competition studies (IC₅₀), L is the free molar concentration of radioligand for saturation studies or molar concentration of competing compound for competition studies, and n is the Hill coefficient. The Cheng-Prusoff equation was used to calculate the K_i values of competing drugs, with K_i = IC₅₀/(1 + ([L]/K_d).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Radioligand binding studies were performed with both [³H]GR65630 and [³H](S)-zacopride on either HEK 293 or TSA201 cell membranes expressing WT, F107Y, or F107N mutant receptors. The densities of mutant and WT receptors determined with either radioligand were similar, suggesting that the mutations had not compromised expression. Saturation studies with either [³H]GR65630 (Fig. 2) or [³H](S)-zacopride appeared to label a homogeneous population of receptors, and there were no significant differences in the K_d values for either ligand between the WT and mutant receptors (Table 1 and 2).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Equilibrium saturation binding isotherms determined with [3H]GR65630. Studies were carried out on cell membranes obtained from HEK 293 cells expressing WT (F107), F107Y, or F107N mutant receptors. Specific binding was defined with 300 µM metoclopramide. Results are shown from a single experiment that was repeated 3 to 10 times. Scatchard transformation of the data in all cases was consistent with the expression of a single homogeneous receptor population.

Mutant F107Y. Competition studies in the presence of [³H]GR65630 indicated that the affinity for the natural agonist 5-HT was decreased by a factor of 13 in F107Y compared with the WT (K_i = 203 and 15.6 nM, respectively; Fig. 3, Table 1). The direction of the change for the close structural analog 2-methyl-5-HT was the same, although the magnitude of the change was only 3.5-fold (K_i = 610 and 173 nM for the mutant and WT, respectively; Table 1). The agonist N-methyl-5-HT had a 35-fold decrease in affinity compared with WT (K_i = 6695 and 187 nM, respectively; Table 1). This mutation had no significant effects on the affinity for the agonist mCPBG or the antagonists ondansetron, granisetron, or renzapride (Table 1). Competition studies, in which [³H](S)-zacopride replaced [³H]GR65630, gave similar results for this mutant, although the magnitude of the observed changes was reduced (Table 2).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Competition of 5-HT for [3H]GR65630 binding to WT (F107), F107Y, and F107N mutant receptors on membranes obtained from expressing HEK 293 cells. Data are expressed as the mean ± S.E. (n = 4-8).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Concentration-effect curves for 5-HT from WT (F107), F107Y, and F107N mutant receptors transiently expressed in HEK 293 cells. I and Imax are the currents at a given agonist concentration and the maximal current, respectively. Data are expressed as the mean ± S.E. (5-11 points were used for each determination).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Comparison of the agonist-induced inward currents in HEK 293 cells transiently transfected with WT (F107), F107N, and F107Y receptors. Currents were recorded from cells that were voltage-clamped at a holding potential of 60 mV. A, 5-HT application at micromolar concentrations indicated. In each case, current traces are shown at approximate EC50 and 10 × EC50 agonist concentrations. In the case of the mutant F107Y, an additional trace is shown at 1000 × EC50. B, mCPBG application at a concentration of 10 µM. The horizontal bar indicates the duration of agonist application. Note that the time to maximum current amplitude is markedly longer for the current elicited by 100 µM 5-HT (10 × EC50) in mutant F107Y receptors compared with the current elicited by 10 µM 5-HT (10 × EC50) in WT receptors. Note also that the time to maximum current amplitude is reduced by 10 mM 5-HT (1000 × EC50) for mutant F107Y receptors.

Mutant F107N. Competition studies with [³H]GR65630 showed that the affinity for the agonist 5-HT was increased 10-fold for mutant receptor F107N compared with the WT (K_i = 1.62 and 15.6 nM, respectively; Fig. 3 and Table 1), whereas the increase in affinity for 2-methyl-5-HT was only 4-fold compared with WT (41.7 and 173 nM, respectively; Table 1). The affinity for N-methyl-5-HT increased 35-fold in mutant F107N compared with WT (K_i = 5.2 and 187 nM, respectively; Table 1). mCPBG, ondansetron, and renzapride exhibited no change in affinity for the mutant F107N compared with WT, although the affinity of granisetron was decreased 10-fold (Table 1). In competition studies with [³H](S)-zacopride, granisetron showed no difference in affinity to the WT receptor, whereas renzapride exhibited a significant increase in affinity compared with WT (Table 2).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Comparison of the whole-cell currents recorded in WT (F107) and mutant F107N receptors as a consequence of 5-HT or ACh application. A, agonist application is indicated by the horizontal bars. The concentration of 5-HT was 10 µM, whereas the concentration of ACh was 10 mM for F107 and 3 mM for F107N. B, renzapride application, at a concentration of 100 nM, preceded that of ACh by 30 s and was followed by coapplication with ACh.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The mutation of the aromatic residue F107 in the 5-HT_3AL receptor subunit sequence produced parallel shifts in radioligand binding affinity (K_i) and EC₅₀ value determined electrophysiologically for the natural agonist 5-HT. The apparent affinity for the mutant F107Y decreased by about 10-fold compared with the WT receptor, whereas for the mutant F107N, the apparent affinity for 5-HT increased by a similar magnitude. The mutations exhibited similar characteristics when N-methyl-5-HT replaced 5-HT in these experiments. The change in binding affinity of 5-HT for the mutant receptors, determined by radioligand binding, was mimicked qualitatively by its structural analog 2-methyl-5-HT, although the magnitude of the change for this agonist was smaller. Accurate determination of the EC₅₀ values for 2-methyl-5-HT was compromised by its low efficacy in this preparation, approximately 20% of that found for the natural agonist. The mutations did not affect the affinity of the agonist mCPBG, and the currents elicited by it were indistinguishable from those found in the WT receptor. The mutations did not significantly compromise the affinity of the majority of the antagonist ligands studied, although in competition studies with [³H]GR65630, the affinity of granisetron for mutant F107N was decreased about 10-fold compared with WT. It is interesting in this regard that Yan et al. (1999) located, by alanine-scanning mutagenesis, a region of the 5-HT₃ receptor sequence, N-terminal to that studied here, that exhibits distinct amino acids that interact with the agonist 5-HT and the antagonist granisetron. Although our observation in this regard merits further exploration, the main focus of this work concerns the agonist interaction with these mutants.

The mutation of F107 to tyrosine (F107Y) produced a decrease in radioligand binding affinity of 5-HT and a concomitant increase in EC₅₀. In addition, the rate of activation of the 5-HT current was markedly slowed compared with WT. Maximum 5-HT-induced currents were obtained in both WT and F107Y mutant receptors by agonist activation at 10 × EC₅₀. In the WT receptor, the time to half-peak current amplitude was 108 ± 25 ms (10 µM 5-HT), whereas in the mutant F107Y receptors, this was increased to 1795 ± 783 ms (100 µM 5-HT). However, on application of a supramaximal 5-HT concentration (10 mM, approximately 1000 × EC₅₀) to mutant F107Y, the activation rate of the macroscopic whole-cell current was markedly increased, showing a time to half-peak current of 494 ± 180 ms. These observations are consistent with decreased association rate of the agonist 5-HT for the F107Y mutant receptor. Currents induced by the agonist mCPBG were indistinguishable, in both magnitude and rate of activation, from those found in the WT receptor.

(1)

There is no state model currently available that reproduces all of the properties of 5-HT₃ receptor activation. However, it can be seen in this simplified agonist binding eq. 1 that the rates of both agonist association/dissociation (k₊₁/k_1, k₊₂/k₂) and channel opening/closing (gating, /) of the receptor determine its apparent affinity (EC₅₀). However, the observation that 5-HT activation in F107Y mutant receptors is markedly increased by supramaximal 5-HT concentrations suggests that it is the decreased agonist association rate that is responsible for the decrease in 5-HT apparent affinity for the F107Y receptor. Once maximum current amplitude is reached, the agonist association rate alone affects activation rate, as channel opening and closing rates are independent of concentration. The observation that both the K_i and EC₅₀ values obtained for mCPBG are unaffected by this mutation is consistent with the suggestion that this mutant does not compromise channel gating. mCPBG recognition appears to be primarily associated with an amino acid segment immediately before the first hydrophobic domain of the sequence (Lankiewicz et al., 1998). Interestingly, mutation of the neighboring tyrosine residue to phenylalanine in the nAChR 7 subunit results in decreased agonist affinity for the natural agonist (ACh; Galzi et al., 1991). Aylwin and White (1994b), using mouse muscle cDNA-derived nAChRs, suggested that this was most likely due to a decrease in agonist association rate because, using single-channel analysis, they determined that the channel opening and closing rates were similar to those of WT. The low conductance of 5-HT₃ receptor homo-oligomer used in this study does not permit single-channel analysis, and we have thus been unable to further explore this.

The most parsimonius explanation of this observation is that the phenolic hydroxyl substituent in this mutant sterically compromises the access of 5-HT to its recognition site, an effect that can be overcome by driving the interaction with increased ligand concentrations. The increased bulk introduced with an additional aliphatic methyl group in N-methyl-5-HT reduced the radioligand binding affinity by a factor of 35, although the magnitude of this decreased affinity was not entirely reflected in the EC₅₀ determinations. A detailed electrophysiological characterization was not carried out with this compound.

The mutation of F107 to asparagine (F107N) caused an increase in affinity for 5-HT. The parallel changes in K_i and EC₅₀ values are again consistent with a direct effect on binding of the agonist at the recognition site rather than with an indirect effect on channel gating. Asparagine is able to both accept and donate hydrogen bonds. However, phenylalanine acts only as a weak hydrogen bond acceptor by virtue of its delocalized -electrons (Levitt and Perutz, 1988). The relatively small effects of the mutation on affinity of 2-methyl-5-HT, in which the methyl substituent will increase the partial charge on the pyrrole nitrogen (Fig. 7), seem to preclude an interaction with this site. In comparison with 5-HT, the analog N-methyl-5-HT exhibits more pronounced changes in both K_i and EC₅₀ for this mutant. The additional methyl substituent in this compound will increase the electron density on this aliphatic amine compared with 5-HT, thus making N-methyl-5-HT a stronger hydrogen bond acceptor than 5-HT but a weaker hydrogen bond donor (Fig. 7). This suggests not only that it is the aliphatic amine of 5-HT that interacts directly with the amino acid in this position but also that in F107N, asparagine donates a hydrogen bond to the aliphatic amine of the indoleamines.

View larger version (10K): [in this window] [in a new window]   Fig. 7.   The structures of 5-HT, N-methyl-5-HT, and 2-methyl-5-HT. The numbers in parentheses adjacent to the pyrrole nitrogen and the aliphatic amine, in each structure, are pKa values calculated with the PALLAS software system.

The mutant F107N also exhibited a significant affinity for ACh (EC₅₀ = 256 µM), producing similar maximal currents to those found with 5-HT. Both the 5-HT and ACh responses were blocked, in an apparantly competitive manner, by renzapride, the specific 5-HT₃ receptor antagonist. This mutation is thus perhaps simply permissive of the ACh interaction in that the majority of nAChR -subunits have a conserved asparagine at this position (Fig. 1), although in the 7 subunit, the equivalent position is occupied by serine, which is also a hydrogen bond donor and acceptor.

Previous studies with the nAChR have clearly shown the requirement for an aromatic residue at position 93 in the nAChR -subunits to support a high affinity for the natural agonist. The studies of Nowak et al. (1995), in which a variety of unnatural amino acids were substituted at this position, suggested that this particular tyrosine was involved in hydrogen bond donation. In this 5-HT₃ receptor mutant, the equivalent alignment position is occupied by asparagine (105; Fig. 1), which will also act as a hydrogen bond donor. Our observation that the 5-HT₃ receptor mutant F107N but not the WT receptor is activated by ACh indicates that the asparagine at position 107 may allow hydrogen bond formation, further stabilizing the interaction with the acetyl moiety of ACh. This would increase the energy of the interaction between ACh and the receptor, allowing receptor activation. Although it is reasonable to suppose that this is a direct binding interaction, our attempts to identify high-affinity ACh displacement of [³H]GR65630 in mutant F107N were unsuccessful, and it is thus possible that the mutation simply provided a means by which the receptor conformation could proceed to the channel open state. The apparent affinity of this 5-HT₃ receptor mutant for ACh is reduced by only a factor of about 2 over that found for the chick 7 homo-oligomer.

The results of mutational analysis of phenylalanine 107 in the 5-HT_3AL sequence suggest that this position provides important recognition properties for the natural agonist 5-HT. The mutant F107Y results in a marked decrease in the association rate of the agonist 5-HT. Our studies with both 5-HT and N-methyl-5-HT in mutant F107N have been rationalized with the suggestion that the amino acid in this position acts as a hydrogen bond donor to the terminal aliphatic amine of 5-HT. In this mutant, ACh functioned as a full agonist, and this response was inhibited by the 5-HT₃ receptor antagonist renzapride.