The Role of Tryptophan Residues in the 5-Hydroxytryptamine 3 Receptor Ligand Binding Domain*

Aromatic amino acids are important components of the ligand binding site in the Cys loop family of ligand-gated ion channels. To examine the role of tryptophan residues in the ligand binding domain of the 5-hy-droxytryptamine 3 (5-HT 3 ) receptor, we used site-di- rected mutagenesis to change each of the eight N-termi-nal tryptophan residues in the 5-HT 3A receptor subunit to tyrosine or serine. The mutants were expressed as homomeric 5-HT 3A receptors in HEK293 cells and ana- lyzed with radioligand binding, electrophysiology, and immunocytochemistry. Mutation of Trp 90 , Trp 183 , and Trp 195 to tyrosine resulted in functional receptors, although with increased EC 50 values (2–92-fold) to 5-HT 3 receptor agonists. Changing these residues to serine either ablated function (Trp 90 and Trp 183 ) or resulted in a further increase in EC 50 (Trp 195 ). Mutation of residue Trp 60 had no effect on ligand binding or receptor function, whereas mutation of Trp 95 , Trp 102 , Trp 121 , and Trp 214 ablated ligand binding and receptor function, and all but one of the receptors containing these mutations were not expressed at the plasma membrane. We propose that Trp 90 , Trp 183 , and Trp 195 are intimately involved in ligand binding, whereas Trp 95 , Trp 102 , Trp 121 , and Trp 214 have a critical role in receptor structure or assembly. The in 35-mm plates and stained 3 days post-transfection. Radioligand Binding— Transfected HEK293 cells were washed twice with phosphate-buffered saline at room temperature: all subsequent steps were carried out at 1–4 °C. Cells were scraped into 1 ml of HEPES buffer (10 m M , pH 7.4) containing the following proteinase inhibitors: 1 m M EDTA, 50 m g/ml soybean trypsin inhibitor, 50 m g/ml bacitracin, and 0.1 m M phenylmethylsulfonyl fluoride. Harvested cells were washed in HEPES with proteinase inhibitors and frozen at 2 20 °C. After thawing, for Biotinylated anti-rabbit IgG fluorescein isothiocyanate D as were mounted in vectashield mounting (Vector), and immunofluorescence observed using a Nikon optiphot Electrophysiological

The nACh receptor is the receptor most closely related to the 5-HT 3 receptor with up to 30% amino acid sequence identity (5). Evidence from biochemical and mutagenesis studies on nACh receptors indicate that the ligand binding site is located in discontiguous regions of the extracellular N-terminal domain, and this has been further confirmed by the construction of a chimeric protein consisting of the N-terminal domain of the ␣7 neuronal nACh receptor subunit linked to the C-terminal portion of the 5-HT 3A receptor subunit, which showed nACh receptor pharmacological properties and 5-HT 3 receptor channel properties (11). Labeling and mutagenesis studies have identified a number of N-terminal amino acids in nACh subunits that are probably involved in ligand binding; these are mostly aromatic amino acids and include Trp␣ 54 , Trp␣ 86 , Tyr␣ 93 , Trp␣ 149 , Trp␣ 187 , Tyr␣ 190 , Cys␣ 192 , Cys␣ 193 , and Tyr␣ 198 (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25).
Sequence alignments between the nACh and 5-HT 3 receptors show that some of the tryptophan residues are conserved in both classes of receptor. These include nACh Trp␣ 54 , Trp␣ 86 , and Trp␣ 149 , which align to residues Trp 90 , Trp 121 , and Trp 183 in the 5-HT 3A receptor subunit. Biochemical evidence also indicates that tryptophan residues are involved in 5-HT 3 receptor ligand binding; modifying these residues with N-bromosuccinimide inhibits radiolabeled antagonist binding to the 5-HT 3 receptor in a ligand-protectable manner (26). To examine the role of tryptophan residues in the N-terminal ligand binding domain of the 5-HT 3A receptor, we have changed the eight tryptophans in this region ( Fig. 1) using site-directed mutagenesis. The resulting mutants have been expressed in HEK293 cells and characterized with radioligand binding, electrophysiology, and immunocytochemistry.

EXPERIMENTAL PROCEDURES
Cell Culture and Transient/Stable DNA Transfection-Human embryonic kidney (HEK293) cells were cultured on 90-mm tissue culture plates in Dulbecco's modified Eagles medium/F-12 medium (1:1) medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (Sigma) at 37°C and 7% CO 2 in a humidified atmosphere. HEK293 cells stably expressing 5-HT 3A receptors were developed using the eukaryotic expression vector pRc/CMV (In Vitrogen, Abingdon, UK) containing the complete coding sequence for the 5-HT 3A(b) subunit from NIE-115 cells as described previously (27). Mutagenesis reactions were performed using the Kunkel method (28) and confirmed by DNA sequencing. For transient transfections, HEK cells at 60 -70% confluency (ϳ48 h post passage) were transfected with WT or mutant plasmid DNA by the calcium phosphate precipitation method (29). For radioligand binding studies cells were grown on 90-mm plates (Falcon) and harvested 3 days after transfection. For electrophysiology experiments cells were grown on 35-mm plates, and recordings were performed 1-4 days post-transfection. In immunofluorescence experiments cells were grown on 22-mm diameter glass coverslips in 35-mm plates and stained 3 days post-transfection.
Radioligand Binding-Transfected HEK293 cells were washed twice with phosphate-buffered saline at room temperature: all subsequent steps were carried out at 1-4°C. Cells were scraped into 1 ml of HEPES buffer (10 mM, pH 7.4) containing the following proteinase inhibitors: 1 mM EDTA, 50 g/ml soybean trypsin inhibitor, 50 g/ml bacitracin, and 0.1 mM phenylmethylsulfonyl fluoride. Harvested cells were washed in HEPES with proteinase inhibitors and frozen at Ϫ20°C. After thawing, they were washed twice with HEPES buffer and resuspended, and 50 g of cell membranes were incubated in 0.5 ml of HEPES buffer containing [ 3 H]granisetron (81 Ci/mmol, DuPont) ranging in concentration from 0.05 to 40 nM, or [ 3 H]meta-chlorophenylbiguanide (mCPBG; 26 Ci/mmol, DuPont) ranging in concentration from 0.05 to 20 nM. 100 nM unlabeled granisetron or 1 M D-tubocurarine was used to determine nonspecific binding. Antagonist reactions were incubated for 1 h and agonist reactions were incubated for 2 h at 4°C and were terminated by rapid vacuum filtration using a Brandel cell harvester onto GF/B filters presoaked for 3 h in 0.3% polyethyleneimine followed by two rapid washes with 4 ml of ice-cold HEPES buffer. Radioactivity was determined by scintillation counting (Beckman LS6000sc). Protein concentration was estimated using the Bio-Rad Protein Assay with bovine serum albumin standards. Data were analyzed by iterative curve fitting (GraphPad, PRISM, San Diego, CA) according to the equation where B is bound radioligand, B max is maximum binding at equilibrium, K is the equilibrium dissociation constant, [L] is the free concentration of radioligand, and n is the Hill coefficient.
Immunofluorescent Localization of WT and Mutant 5-HT 3 Receptors-Transfected cells were washed with 3 changes of Tris-buffered saline (0.1 M Tris/HCl, pH 7.4, 0.9% NaCl) and fixed using ice-cold 4% paraformaldehyde in phosphate buffer (66 mM Na 2 HPO 4 , 38 mM NaH 2 PO 4 , pH 7.2). To label the N-terminal domain, pAb120 antiserum (30) was used at 1:300 dilution in Tris-buffered saline. Intracellular receptor expression was determined by inclusion of 0.3% Triton X-100 for membrane permeabilization. Primary antibody incubation was for 1 h at room temperature. Biotinylated anti-rabbit IgG (Vector) and fluorescein isothiocyanate avidin D (Vector) were used to detect bound antibody as per the manufacturer's instructions. Coverslips were mounted in vectashield mounting medium (Vector), and immunofluorescence was observed using a Nikon optiphot microscope.
Electrophysiological Procedures-Membrane currents from single cells were recorded by whole cell patch clamp using an EPC-9 amplifier (HEKA Elektronik, Darmstadt, Germany) controlled by Pulse software (HEKA). Patch electrodes were made from filamented glass capillary tubing (Clark Electromedical) and were back-filled with filtered (0.2 m, Millipore) electrode solution (140 mM CsCl 2 , 1 mM MgCl 2 , 10 mM EGTA, 10 mM HEPES, to pH 7.2 with CsOH). Electrodes with resistances of 2-4 M⍀ were used for recordings. Series resistance was less than 10 M⍀. Cells were voltage clamped at Ϫ60 mV in all experiments, and drugs were applied via a U-tube delivery system capable of a complete change of solution within 100 ms (31). Data acquisition, monitoring, storage, and analysis were performed using HEKA software. During recordings the dishes were continuously perfused (3-5 ml/min) at room temperature in bath solution (130 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, 30 mM glucose, pH 7.2 with NaOH). Drug applications were followed by a minimum 3-min wash period. Dose-response curves were fitted to the equation: I ϭ I max /(1 ϩ (EC 50 / [L]) n ) using Kaleidagraph (Abelbeck Software), where I is the current at a given agonist concentration, I max is the maximal current, EC 50 is the agonist concentration that elicits half-maximal current, and n is the Hill coefficient.

Wild Type and Trp 60 Mutant Receptors-Radioligand binding studies using membranes from WT-transfected HEK293 cells revealed that [ 3 H]granisetron and [ 3 H]mCPBG label a
homogeneous population of binding sites with high affinity; K d ϭ 0.17 Ϯ 0.03 nM and 1.48 Ϯ 0.09 nM respectively (n ϭ 3). Hill coefficients were not significantly different to unity. There was no difference in radioligand binding parameters for either [ 3 H]granisetron or [ 3 H]mCPBG because of modification of the 5-HT 3 receptor W60 to serine (Table I). Whole cell patch clamp revealed that 5-HT and the 5-HT 3 receptor agonist mCPBG elicited rapid transient inward currents from cells transfected with WT DNA (Fig. 2). From concentration-effect curves EC 50 s were 2.10 Ϯ 0.40 M and 0.81 Ϯ 0.09 M for 5-HT and mCPBG respectively. The EC 50 and n H values for W60S 5-HT 3 receptor constructs were not significantly different to WT (Table II).
Effect of Mutations at Positions 90, 183, and 195-Using membranes prepared from W90Y-transfected HEK293 cells, a 5.4-fold increase in K d for [ 3 H]granisetron was observed compared with WT; the Hill coefficient was not significantly changed. The change in antagonist affinity was reflected by a similar change in agonist affinity: K d for [ 3 H]mCPBG increased 5.6-fold, with no effect on the Hill coefficient. Binding of both radioligands was ablated by the mutations W90S, W183Y, and W183S. W195Y and W195S mutations did not effect the [ 3 H] mCPBG binding affinity but significantly increased the K d for [ 3 H]granisetron binding; these latter increases were not significantly different between the two mutants ( Table I).
The EC 50 values for the W90Y mutant, stably transfected into HEK293 cells, were increased 3.8-and 2.5-fold as compared with WT for 5-HT and mCPBG, respectively (Fig. 3), whereas the EC 50 for the partial agonist 2-Me-5-HT was not altered; EC 50 values were 12.7 Ϯ 0.45 M and 12.9 Ϯ 1.98 M for WT and W90Y, respectively (n ϭ 3). The Hill coefficients for all three agonists were not significantly changed by the W90Y mutation. The mutant W90S, when expressed in HEK293 cells, did not show any response to 5-HT 3 (Յ1 mM) or mCPBG (Յ500 M). The electrophysiological properties of W183Y, W183S, W195Y, and W195S mutants were also examined in transfected HEK293 cells. For the W183Y mutation, EC 50 values for 5-HT and mCPBG were increased 92-and 24-fold, respectively, compared with WT, whereas there was no response to application of either agonist for the W183S mutation. The W195Y mutation resulted in 3.9-and 2.1-fold increases in EC 50 for 5-HT and mCPBG, respectively, whereas the W195S mutation caused 9.0-and 2.7-fold increases in EC 50 , respectively. The Hill coefficients were not affected by W183Y, W195Y, or W195S mutations for either agonist (Table II). The time course of desensitization for these mutants was slower than WT (Fig. 2), suggesting that the mutations had resulted in changes to the   Table II. receptors showed no specific binding with either ligand. Furthermore, none of the mutants W95Y, W95S, W102Y, W102S, W121Y, W121S, W214Y, and W214S, when transiently transfected into HEK293 cells, showed electrophysiological responses to Յ1 mM 5-HT.

Immunolocalization of WT and Mutant 5-HT 3A Receptors-
Receptors that give rise to agonist-induced currents in transfected HEK293 cells are clearly expressed at the plasma membrane. To determine whether mutants that did not respond to agonist or demonstrate detectable radioligand binding are correctly oligomerized and expressed at the plasma membrane, the localization of nonfunctional mutant subunits was studied using a 5-HT 3A receptor N-terminal specific antibody, pAb120 (30). Cytoplasmic 5-HT 3 receptor-specific staining was observed in cells with permeabilized membranes expressing WT and all mutant constructs tested (Figs. 4 and 5). pAb120 immunolabeled cells, transfected with WT, functional mutants (data not shown) and nonfunctional mutants W90S, W183S, and W214Y with nonpermeabilized membranes, acquired a characteristic narrow ring of pAb120 labeling around the membrane. This strong fluorescent signal from the membrane was not observed in cells that did not express receptor or cells that expressed any of the other nonfunctional mutant constructs. The fluorescence in the membrane was granular in appearance, possibly corresponding to aggregations of receptors in the membrane. DISCUSSION To identify the functional roles of tryptophan residues in the N-terminal domain of the 5-HT 3A receptor, we substituted the eight tryptophans in this domain (Fig. 1) to tyrosine and/or serine. The pharmacology, electrophysiology, and localization of the altered receptors were examined using the HEK293 cell expression system. The data we have presented suggest that three of the tryptophan residues, Trp 90 , Trp 183 , and Trp 195 , play a role in ligand binding, whereas the four residues Trp 95 , Trp 102 , Trp 121 , and Trp 214 are critical for the correct receptor assembly and/or structure.
The nonconservative substitution of Trp 60 for serine results in mutant receptors that have radioligand binding and electrophysiological properties not discernible from the WT receptor, suggesting that Trp 60 does not play a role in ligand binding and appears to have little or no structural significance. Alignments of the 5-HT 3 , nACh, GABA A , and glycine receptors (Fig. 6) show that Trp 60 of the 5-HT 3A receptor subunit does not align with tryptophanyl, aromatic, or ligand binding residues from any other members of the nACh-type LGIC family. These findings suggest that Trp 60 is not an important residue for either the structure or the function of the 5-HT 3 receptor.
Mutations at Positions 95, 102, and 121-The effects of mutation at these positions to tyrosine and serine were indistinguishable from each other. All six substitutions resulted in receptors that did not bind [ 3 H]granisetron or [ 3 H]mCPBG, did not respond to 5-HT, and were not expressed at the plasma membrane, although high levels of intracellular protein expression were observed for all of the mutant receptors (Fig. 5).  These observations are consistent with subunits not assembling or assembling incorrectly or fully assembled pentamers not being sorted to the plasma membrane. The importance of hydrophobic amino acids in the N-terminal domain for mediating subunit interactions in receptor assembly has been demonstrated in the glycine (32) and nACh (33) receptors, and it is likely that in the 5-HT 3 receptor, Trp 95 , Trp 102 , and Trp 121 are similarly critical for correct assembly.
An alternative explanation is that Trp 95 , Trp 102 , and Trp 121 may be important for the determination of the tertiary structure of the 5-HT 3 receptor. Indeed both explanations may be correct because one or more of these residues may in fact modify the tertiary structure so that it prevents correct subunit-subunit interactions. Trp 95 and Trp 121 are strictly conserved in all binding and nonbinding nACh, GABA A , and glycine receptor subunits, whereas Trp 102 residue is conserved in most nACh subunits and is otherwise represented as an aromatic residue in the other subunits of this receptor family (Fig.  6). These canonical residues may therefore create an N-terminal domain backbone structure that is common to all of the LGICs in this family (17,19,34,35).
Mutations at Position 214 -The mutation W214Y resulted in a receptor which was expressed at the plasma membrane but did not bind radiolabeled granisetron or mCPBG and did not respond to 5-HT in electrophysiological assays. The effect of the W214S mutation was more severe, resulting in no binding, function, or expression at the plasma membrane. These data suggest that an aromatic residue is necessary at position 214 for correct receptor expression but is not sufficient to retain function. A tryptophan residue at an equivalent position to Trp 214 is conserved in all binding and nonbinding subunits of the nACh receptor (Fig. 6), although aromatic amino acids are not present at the homologous position in either of the anionic receptors. This tryptophan residue may therefore be an important determinant of the tertiary structure of the N-terminal domain of cationic, but not anionic, channels.
Mutations at Position 90 -The W90Y mutation resulted in a receptor with decreased radioligand binding affinities and agonist responses, whereas the replacement of tryptophan with serine caused radioligand binding and agonist responses to be abolished without affecting receptor expression at the plasma membrane. In the W90Y mutant, the similar increase in mCPBG EC 50 (agonist binding receptor in the resting, activable state) and K d (agonist binding receptor in the desensitized state) suggests that the effects of this mutation may be due to a binding site modification rather than an effect on the mechanisms of channel gating. Furthermore, Hill coefficients for radioligand binding and electrophysiological experiments were not altered by the W90Y mutation. This hypothesis is supported by the observation that the W90Y mutation did not change the EC 50 for the partial 5-HT 3 receptor agonist 2-Me-5-HT, and these data also suggest that 2-Me-5-HT binds to the receptor via interactions different from those of the full agonists, 5-HT and mCPBG. The observation that W90S mutant receptors reach the plasma membrane but do not bind ligands suggests that an aromatic group at this position is essential for ligand binding, either because of its interaction with the ligand or for the formation of the correct structure of the ligand binding pocket.
All 5-HT 3 receptor ligands possess an aromatic ring containing a polar group, and a basic nitrogen atom (36). Our data are consistent with a direct interaction of 5-HT 3 receptor ligands with Trp 90 , and we suggest this is via cation interactions between the aromatic tryptophan and the basic nitrogen present in 5-HT 3 ligands, similar to the ligand binding mechanisms postulated for the binding of the quaternary ammonium ion of ACh to aromatic residues of the nAChR (37,38). This hypothesis explains the reduction in affinity for ligands caused by the W90Y mutation, because the indole of tryptophan provides a larger and more intense region of negative electrostatic potential than a simple benzene ring, making tryptophan a more attractive cation binding site than tyrosine. The ablation of ligand binding by replacement of Trp 90 with serine is consistent with this hypothesis, because it would result in the complete removal of this putative cation binding site.
Other possible ligand stabilization mechanisms that could explain our data include dispersion, i.e. a mutual synchronization of fluctuating charge in the overlapping indole rings, or a dipole-induced dipole interaction between the polar group of the ligand and the indole of the tryptophan. Both of these interactions would be weaker than the cation bond that we propose and are therefore probably less likely, although further experimentation and/or determination of the three-dimensional structure of the binding site is required to confirm or disprove our hypothesis.
A recent study also supports a role for Trp 90 in ligand binding and further uses alanine scanning mutagenesis to provide evidence that this region may be in a ␤-strand conformation (39). Trp 90 aligns with the nAChR ␣7 Trp 54 , which is proposed to form a complimentary component of the ligand binding site, i.e. contributed by an adjacent subunit (24). ␣7 Trp 54 is highly conserved in nACh subunits that are expected to contribute a complimentary ligand binding loop, although it is also conserved in many of the neuronal ␣ subunits (Fig. 6). An aromatic residue, phenylalanine, occupies this position in GABA and glycine receptor subunits, and in the former has been shown to contribute to the binding site for GABA (40,41). Thus the data presented for the 5-HT 3A Trp 90 residue extend the identified structural and functional homologies between the receptors of this LGIC superfamily in addition to their considerable sequence homologies.
Mutation at Position 183-Mutation of Trp 183 to tyrosine caused a large increase in agonist EC 50 values and abolished radioligand binding, whereas the W183S mutant had no apparent ligand binding or function but was expressed at the plasma membrane. The cooperative nature of ligand binding was unaffected by the W183Y mutation, suggesting that its effects are unlikely to be manifest via a structural change in the receptor. Thus residue Trp 183 appears to be involved in ligand binding, probably via cation interactions as postulated above for the Trp 90 residue.
The tryptophan at position 183 of the 5-HT 3A receptor sub- Tryptophan Residues in the 5-HT 3 R Ligand Binding Domain unit aligns with trytophan residues in nACh ␣ and ␤ (except muscle type) subunits but is not conserved in the 5-HT 3B receptor subunit nor in the nonbinding nACh subunits ␥, ␦, and ⑀. Trp 183 of the 5-HT 3A receptor subunit aligns with phenylalanine residues in the glycine receptor and tyrosines in the GABA A receptor (Fig. 6). These latter aromatic residues have been implicated in ligand binding for the nACh (14,21), GABA A (42), and glycine receptors (43). Thus 5-HT 3A receptor subunit Trp 183 and its homologues are the first residues that have been demonstrated to be important for ligand receptor recognition in both the cationic and anionic receptors of this LGIC family. Because the largest change in EC 50 of the functional tryptophan mutants examined in this study was caused by the W183Y mutation, and the removal of the aromatic residue at this position ablated ligand binding, we propose that this residue is the most important tryptophan in 5-HT 3 receptor ligand binding site. Recently the equivalent tryptophan in the nACh receptor (Trp 149 ) was shown to be the primary cation binding site in the nAChR by Zhong et al. (44), using the in vivo nonsense codon suppression method of mutagenesis.
Mutations at Position 195-The mutation of Trp 195 to tyrosine and serine resulted in functionally expressed receptors with modified ligand binding and electrophysiological properties. The affinity, as measured by the EC 50 s, for 5-HT and mCPBG were reduced in the W195Y mutant compared with WT and further reduced for 5-HT in the W195S mutant, without alteration of the Hill coefficients. Interestingly the binding characteristics of [ 3 H]mCPBG binding were unchanged for either mutant, although [ 3 H]granisetron binding affinities were increased. The unaffected mCPBG binding and function and unchanged Hill coefficients suggest that this mutation does not affect channel gating. Combined with the other data, we propose that changing Trp 195 specifically effects a region of the binding pocket that has some importance for ligand binding, and, because these mutations resulted in particularly large changes in [ 3 H]granisetron binding affinity, it is likely that the affected region of the binding pocket participates more in antagonist than agonist binding.
The 5-HT 3A receptor subunit Trp 195 residue aligns with tryptophan residues in GABA A ␣ and ␤ subunits and in glycine receptors but does not have a homologue in the nAChR subunits (Fig. 6). The tryptophan at this position in anionic channels has not been reported to be important in ligand binding or channel function. Therefore our data represent the identification of a novel ligand binding region in the 5-HT 3 receptor that is not present in the other members of the Cys loop neurotransmitter family.
Conclusions-Combining the findings from mutations at Trp 90 , Trp 183 , and Trp 195 , we can postulate that the cationic ligands of the 5-HT 3 receptor interact with these aromatic residues, probably via cation interactions, with Trp 183 likely to be the principle tryptophan in the binding site. The homology between these putative ligand binding amino acids and the other members of the Cys loop family, demonstrates that tertiary folding of the N-terminal domain for these receptors is highly conserved.
In summary, we have systematically examined the roles played by all of the tryptophan residues in the N-terminal ligand binding domain of the 5-HT 3 receptor. A combination of radioligand binding, electrophysiological assays, and immunolocalization experiments allowed the identification of functionally significant residues in ligand binding, Trp 90 , Trp 183 , and Trp 195 , and those implicated in receptor structure and/or assembly, Trp 95 , Trp 102 , Trp 121 , and Trp 214 . The findings of this study further exemplify the high degree of structural and functional homology between the receptors in the Cys loop LGIC family and provide insights toward the subtle differences that may be responsible for the characteristic 5-HT 3 receptor ligand binding profile.