Arginine 222 in the Pre-transmembrane Domain 1 of 5-HT 3A Receptors Links Agonist Binding to Channel Gating*

Ligand-gated ion channels are integral membrane proteins that mediate fast synaptic transmission. Molecular biological techniques have been extensively used for determining the structure-function relationships of ligand-gated ion channels. However, the transduction mechanisms that link agonist binding to channel gating remain poorly understood. Arginine 222 (Arg-222), located at the distal end of the extracellular N-terminal domain immediately preceding the first transmembrane domain (TM1), is conserved in all 5-HT 3A receptors and (cid:1) 7-nicotinic acetylcholine receptors that have been cloned. To elucidate the possible role of Arg-222 in the function of 5-HT 3A receptors, we mutated the arginine residue to alanine (Ala) and expressed both the wild-type and the mutant receptor in human embryonic kidney 293 cells. Functional studies of expressed wild-type and mutant receptors revealed that the R222A mutation increased the apparent potency of the full agonist, serotonin (5-HT), and the partial agonist, 2-Me-5-HT, 5- and 12-fold, respectively. In addition, the mutation increased the efficacy of 2-Me-5-HT and converted it from a partial agonist to a full agonist. Furthermore, this mutation also converted the 5-HT 3 receptor antagonist/ very weak partial agonist, apomorphine, to a potent agonist. Kinetic analysis revealed that the R222A mutation increased the rate of receptor activation and desensitization

In the nervous system, serotonin type 3 (5-HT 3 ) 1 receptors can mediate fast excitatory synaptic transmission and modulate neurotransmitter release (1). To date two 5-HT 3 receptor subunits have been identified: 5-HT 3A and 5-HT 3B (2,3). The 5-HT 3A receptor subunits can form functional channels homomerically (2), whereas the 5-HT 3B receptor subunits are nonfunctional when expressed alone (3). However, the 5-HT 3B receptor subunits can form heteromeric channels with the 5-HT 3A receptor subunits, which results in modified biophysical characteristics compared with homomerically expressed 5-HT 3A receptor subunits (3). 5-HT 3A and 5-HT 3B receptor subunits also have different distribution patterns in the nervous system. The 5-HT 3A receptor subunits are expressed in both central and peripheral neurons, whereas the 5-HT 3B receptor subunits are restricted to peripheral neurons (4). This suggests that homomeric 5-HT 3A receptors play a dominant role in 5-HT 3 receptor-mediated responses in the central nervous system. 5-HT 3 receptors belong to a superfamily of ligand-gated ion channels, which includes nicotinic acetylcholine (nACh) receptors, glycine receptors, and ␥-aminobutyric acid type A receptors (5). The subunits in this superfamily are thought to assemble as pentamers with each subunit containing a large extracellular N-terminal domain, four transmembrane domains (TM1-TM4), a large intracellular loop between TM3 and TM4, and an extracellular C-terminal domain (Fig. 1A) (6). The agonist binding sites are thought to be located in the N-terminal domain at subunit-subunit interfaces, and the lining of the ion channel is believed to be formed by the second transmembrane (TM2) domain (6). The binding of agonist to the binding sites in the N-terminal domain presumably results in a conformational change of the channel protein, which is then transduced to the TM2 domain to result in channel opening.
A number of studies have been carried out to determine the sequence elements involved in agonist binding and channel gating of 5-HT 3A receptors (7)(8)(9)(10)(11)(12)(13). However, the mechanisms that transduce the binding of agonist to the opening of the channel are still poorly understood. Understanding the function of ligand-gated ion channels at the molecular level requires understanding how agonist binding to the receptor is converted to channel opening. Arginine (Arg) 222 is of particular interest because it is located at the distal end of the extracellular N-terminal domain, immediately adjacent to the first transmembrane (TM1) domain (Fig. 1A), and thus, it is between the presumed agonist binding sites in the N-terminal domain and TM2. Sequence alignments reveal that Arg-222 of the mouse 5-HT 3A receptor is conserved in all of 5-HT 3A receptors and ␣7-nACh receptors that have been cloned from various species (Fig. 1B). Moreover, some mutations at Arg-222 can alter the sensitivity of 5-HT 3A receptors to agonists and produce channels that open spontaneously (14). In an attempt to understand the role of Arg-222 in the function of 5-HT 3A receptors, we replaced the arginine residue with an alanine using site-directed mutagenesis. The wild-type and R222A 5-HT 3A receptors were transiently expressed in human embryonic kidney (HEK) 293 cells, and the functional properties of the receptors were studied using the whole-cell patch clamp recording technique in combination with fast solution exchange. The results suggest that Arg-222 is involved in transducing the signal that couples agonist binding to channel opening in 5-HT 3A receptors.

EXPERIMENTAL PROCEDURES
Mutagenesis-Point mutation of the mouse 5-HT 3A receptor was accomplished using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutation was verified by double strand DNA sequencing using an ABI Prism 377 automatic DNA sequencer (Applied Biosystems, Foster City, CA). The cDNAs were then subcloned into the vector pcDNA3.1 (Invitrogen) for expression in HEK 293 cells.
Cell Culture and Transient Receptor Expression-HEK 293 cells (American Type Culture Collection, Manassas, VA) were grown in minimum essential medium (Invitrogen) supplemented with 10% horse serum and maintained in a humidified incubator at 37°C in 5% CO 2 . The HEK 293 cells were transiently transfected with the wild-type or R222A 5-HT 3A receptor cDNA using LipofectAMINE 2000 reagent (Invitrogen) according to the manufacturer's instructions. Green fluorescent protein (pGreen Lantern, Invitrogen) was co-expressed with the 5-HT 3A receptor subunits to permit selection of transfected cells under fluorescence optics.
Patch Clamp Recording-HEK 293 cells were recorded 1-3 days after transfection. Cells were continuously superfused with a solution containing 140 mM NaCl, 5 mM KCl, 1.8 mM CaCl 2 , 1.2 mM MgCl 2 , 5 mM glucose, and 10 mM HEPES (pH 7.4 with NaOH; ϳ340 mosmol with sucrose). Membrane current was recorded in the whole-cell configuration (15) using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) at 20 -22°C. Pipettes were pulled from borosilicate glass (TW-150F, World Precision Instruments, Sarasota, FL) using a twostage puller (Flaming-Brown P-87; Sutter Instruments, Novato, CA) and had resistances of ϳ5 megaohms when filled with pipette solution containing 140 mM CsCl, 2 mM MgCl 2 , 10 mM EGTA, 10 mM HEPES (pH 7.2 with CsOH; ϳ315 mosmol with sucrose). Cells were held at Ϫ60 mV unless otherwise indicated. Data were acquired using pClamp8.0 software (Axon). Currents were filtered at 2 kHz and digitized at 2-10 kHz. Agonists were applied with a piezoelectric device (PZ-150M; EXFO Burleigh Products Group Inc., Victor, NY) through two-barrel glass tubing (TGC150, Warner Instruments, Hamden, CT) that had been pulled to a tip diameter of ϳ200 m. The piezoelectric device was driven by transistor-transistor logic pulses from pClamp 8.0 software. Voltage applied to the piezoelectric device produced a rapid lateral displacement of the tubing to move the interface between control and agonist solutions. Solution exchange rate was estimated using the potential change induced by switching from the control solution to a 140 mM N-methyl-D-glucamine test solution. The solution exchange time constants were ϳ0.3 ms for an open pipette tip and ϳ1.6 ms for whole-cell recording.
Data Analysis-Average data are presented as means Ϯ S.E. unless noted otherwise. Data analysis and curve fitting were performed with Voltage Independence-To address whether the R222A mutation altered ion permeation, current-voltage (I-V) relationships were obtained by measuring the amplitude of current Antagonist Profile-Because some Arg-222 mutants exhibited spontaneous channel opening when expressed in Xenopus oocytes (14), we examined whether the R222A mutation produced such activity in HEK 293 cells. Spontaneous opening of ligand-gated channels can be blocked with antagonists (14,16). As shown in Fig. 4A, left, in cells expressing the wild-type receptors the application of 300 nM MDL 72222, a competitive 5-HT 3 receptor antagonist, for 10 s in the absence of 5-HT did not alter the holding current. Similarly, in cells expressing the R222A receptors the application of 300 nM MDL 72222 for 10 s did not alter the holding current (Fig. 4A, right). We also tested the effectiveness of MDL 72222 in antagonizing 5-HT-activated currents. MDL 72222 at 300 nM completely blocked the current activated by 30 M 5-HT in both the wild-type and the R222A receptors (Fig. 4B). Blockade of 5-HT-activated responses by MDL 72222 was readily reversed by washing (data not shown). Apomorphine has been reported to be an antagonist/very weak partial agonist in the 5-HT 3 receptors (17). Fig. 4C, left, shows that in the wild-type receptors 30 M apomorphine activated a very small current when compared with the current activated by 30 M 5-HT. In addition, when co-applied with 5-HT in the wild-type receptors, 30 M apomorphine inhibited the current activated by 30 M 5-HT. On the other hand, 30 M apomorphine activated a significant inward current in the R222A receptors (Fig. 4C, right). The bar graphs in Fig. 4D show the average action of apomorphine in both the wild-type and the R222A receptors. Apomorphine inhibited 5-HT-activated current by 82.1 Ϯ 3.1% in the wild-type receptors (p Ͻ 0.01). In the wild-type receptors the average amplitude of the current activated by 30 M apomorphine was less than 1% that activated by 30 M 5-HT; however, in the R222A receptors the average amplitude of the current activated by 30 M apomorphine was ϳ 60% of the amplitude of the current activated by 30 M 5-HT (wild type, 0.88 Ϯ 0.35%; R222A, 60.8 Ϯ 4.3%, p Ͻ 0.01).
Activation Kinetics-To evaluate the kinetics of 5-HT 3A receptor activation, 5-HT or 2-Me-5-HT was applied at different concentrations for a sufficient duration to allow the receptoragonist interaction to reach equilibrium.  (Fig. 5B), respectively, from cells expressing the wild-type or the R222A receptors. The records show that high agonist concentrations elicited a faster activation than low agonist concentrations for both the wild-type and the R222A receptors. In addition, the rate of receptor activation was agonist-dependent for both the wildtype and the R222A receptors; fast for 5-HT and slower for  (Fig. 5D) concentration for the wild-type and the R222A receptors. Although the concentration at which the activation rate was half-maximal was 3.7 M 5-HT for the R222A receptor, 25 M 5-HT was required to achieve the same activation rate for the wild-type receptor. Similarly, 23 M 2-Me-5-HT produced the half-maximal activation rate in the R222A receptor, but extrapolation suggests that 1800 M 2-Me-5-HT would be needed to produce the same activation rate for the wild-type receptor.
Gating-Activation of ligand-gated ion channels involves agonist binding and conformational changes that lead to gating. Previous studies on ligand-gated ion channels suggest that at low agonist concentrations, agonist binding is the rate-limiting step; however, when the agonist concentration is high, gating becomes the rate-limiting step, and the activation rate approximates the channel opening rate (18). In view of this, we examined the activation of both the wild-type and the R222A receptors with saturating concentrations of 5-HT. The increase in activation rate with increasing agonist concentrations appeared to reach a plateau at ϳ300 M 5-HT for both the wildtype (Fig. 6A) and R222A (Fig. 6B) receptors, since the activation rate did not appear to be accelerated by increasing 5-HT concentration above 300 M (up to 3 mM). On the other hand, the activation rate was faster in R222A receptors than in wild-type receptors (Fig. 6C). In the wild-type receptors, for 5-HT concentrations Ն300 M, average activation rates (Fig.  6D) were not significantly different (analysis of variance, p Ͼ 0.5). Similarly, in the R222A receptors, average activation rates (Fig. 6D) were not significantly different for 5-HT concentration Ն300 M (analysis of variance, p Ͼ 0.3). However, the R222A mutation significantly accelerated the activation rate for 300 M (p Ͻ 0.01), 1 mM (p Ͻ 0.001), and 3 mM 5-HT (p Ͻ 0.001) compared with the wild-type receptors (Fig. 6D).
Deactivation Kinetics-Deactivation is the process of agonist unbinding and channel closing, and it is generally considered as an index of agonist affinity (19,20). To determine whether the R222A mutation alters the rate of deactivation, we examined the current deactivation kinetics after rapid removal of the agonist. The records in Fig. 7, A and B, illustrate the current deactivation after rapid removal of 1 mM 5-HT (Fig. 7A)
Desensitization Kinetics-In the continued presence of agonist, the activated current exhibited a marked decrease in amplitude after reaching a peak, indicating receptor desensitization. Examples of the desensitization resulting from the ap-  Fig. 8, A and B, respectively. For the wild-type 5-HT 3A receptors, the desensitization of the current was relatively slow and well described by a mono-exponential function. In addition, the desensitization time constant for 2-Me-5-HT in the wild-type receptors was significantly slower than for 5-HT (p Ͻ 0.04). For the R222A receptors, a bi-exponential function was required to adequately fit the desensitization of the current activated by 5-HT or 2-Me-5-HT. Both the fast and slow desensitization components of the R222A receptors were similar for 5-HT and 2-Me-5-HT (p Ͼ 0.50). The fast component contributed ϳ60% of the total current decay for both agonists (Fig. 8, C and D). The slow component of the decay in the R222A receptor for either 5-HT or 2-Me-5-HT was faster than the current decay in the wild-type receptor (p Ͻ 0.01).

DISCUSSION
In the present study, we investigated the role of amino acid residue Arg-222 in the function of 5-HT 3A receptors. We found that the R222A mutation increased the apparent potency of both the agonist, 5-HT, and the partial agonist, 2-Me-5-HT. The R222A mutation also increased the efficacy of the partial agonist, 2-Me-5-HT, and converted the antagonist/very weak partial agonist, apomorphine, to a potent agonist. In addition, the R222A mutation accelerated the rate of 5-HT 3A receptor activation and desensitization, but it did not alter the rate of 5-HT 3A receptor deactivation.
Ligand-gated ion channels exist in at least three interconvertible states (resting, open, and desensitized), and the function of these channels is thought to be determined by the transitions among these states (6). Activation of ligand-gated ion channels involves the binding of agonist to the receptor, which results in a conformational change of the protein that opens the channel. In the present study, how the R222A mutation affects 5-HT 3A receptor activation was examined.
Because deactivation is defined as the process of the unbinding of agonist from the open state of the receptor and the channel returning to the closed resting state, it is expected that receptors with a higher affinity for agonist in the open state would decrease the probability of agonist unbinding from the receptor. Thus, agonist affinity could be assessed from agonist unbinding rate (19,20). If the R222A mutation increases agonist affinity, the deactivation rate of the receptor would be expected to be slower. We found, however, that the R222A mutation did not change the rate of deactivation, because currents for both wild-type and R222A receptors decayed with similar kinetics upon removal of agonist. The observation that the mutation did not alter the deactivation rate for either 5-HT or 2-Me-5-HT is, thus, inconsistent with an increase in agonist affinity for R222A receptors. This is consistent with the observation in a single-oocyte binding assay (14). However, we cannot exclude the possibility from this study that Arg-222 is part of the agonist binding pocket.
2-Me-5-HT is a partial agonist at mouse wild-type 5-HT 3A receptors (21). Because the relative efficacy of an agonist is dependent in part upon the gating process (22), the conversion of the partial agonist, 2-Me-5-HT, into a full agonist suggests the possibility of an enhanced gating efficacy in the R222A receptors. The increase in efficacy by R222A mutation is not unique to 2-Me-5-HT, because this mutation also converted the extremely weak 5-HT 3 receptor partial agonist, apomorphine, to a potent agonist. In addition, the R222A mutation was found to enhance the rate of 5-HT 3A receptor activation at saturating concentrations of 5-HT, which is also consistent with a facilitation of the gating process. For the ligand-gated ion channels, it has been found that a mutation that facilitates gating may only cause a shift to the left of the agonist concentrationresponse curve for a full agonist (22). The observations that the R222A mutation accelerated activation of 5-HT 3A receptors by 5-HT and increased the apparent potency of 5-HT suggest that the efficacy of 5-HT to gate the channel is also enhanced by the mutation. Alterations in apparent potency and efficacy of agonists by mutations have been observed for other ligand-gated ion channels. For example, point mutations in ␣7-nACh receptors at Leu-247 (23) and ␣-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid receptors at Ala-636 (24) converted an antagonist into an agonist and increased the apparent potency of agonists. On the other hand mutations in the NR1 subunit of N-methyl-D-aspartate receptors at Asp-732 (25) and the ␣2 subunit of nACh receptors at Asp-200 (26,27) decreased apparent agonist potency and converted partial agonists into competitive antagonists.
The distance between the agonist binding sites and the channel pore of the nACh receptor is estimated to be ϳ30 Å (28). The agonist binding sites of the 5-HT 3A receptor are thought to be comparable with those of the nACh receptor (29). Because the agonist binding sites are at a distance from the channel gating mechanism, the binding signal must involve a transduction mechanism to open the channel. The binding of agonist to the receptor is thought to provide the driving force to open the channel (30,31). It is proposed for nACh receptors that the binding of agonist triggers a localized disturbance at the binding sites that transmits to transmembrane domains of the receptor through a small rotation of the N-terminal domain (32). Our observations on the role of Arg-222 in the gating of 5-HT 3A receptors suggest that this amino acid residue may be involved in the transduction of the signal from agonist binding to channel gating. The pre-TM1 region and TM2-TM3 loop of several ligandgated ion channels have been found to be critical in the coupling of agonist binding to channel gating. Certain residues of the TM2-TM3 loop in nACh receptors (33,34), glycine receptors (35), and ␥-aminobutyric acid type A receptors (36) have been found to participate in the transduction of agonist binding to channel gating. For N-methyl-D-aspartate receptors, the pre-TM1 segment was found to link agonist binding to channel gating and to affect entry into open and desensitized states (37). In addition, a recent study suggests that an interaction of residues in TM1 of nACh receptors, which are near Arg-222 in 5-HT 3A receptors, and TM2 contributes to the gating process (38). Our data suggest that Arg-222, located in pre-TM1 of 5-HT 3A receptors, may also play a crucial role in the transduction mechanism coupling agonist binding to channel opening and entry into the desensitized state. In this context, Arg-222 may serve as a constraint to gating of the 5-HT 3A receptors by maintaining the wild-type receptors in the resting closed state. The R222A mutation may release this constraint and reduce the energy barrier for channel opening, increasing apparent agonist potency and partial agonist efficacy as a result of enhanced coupling between agonist binding and gating.
In the present study, it appears that the R222A mutation reduced the energy barrier for agonist to open the channel. In this regard, it should be noted that arginine and alanine residues differ in their size, polarity, and hydrophobicity. Facilitated coupling resulting from the R222A mutation might be due to the decreased side-chain size, the increased hydrophobicity, or the reduced charge. Any of these changes may reduce the energy barrier for channel opening and, thus, facilitate the transition from a closed to an open state. Mutations at Val-385 of the ␣ subunit of muscle type nACh receptors revealed that both volume and stereochemistry contribute to channel gating but not to binding affinity (39). The observation that some 5-HT 3A receptor mutants, such as R222F and R222I, exhibits spontaneous channel opening (14) is consistent with the notion that Arg-222 mutations can reduce the energy barrier for channel opening.
It has been proposed that the closed resting state is the most stable conformation in the absence of agonist, whereas the closed desensitized state is the most stable conformation in the presence of agonist (6). Desensitization is a widespread phenomenon among ligand-gated ion channels (40). Our observation that the R222A receptors exhibited a significantly faster desensitization than the wild-type receptors suggests that although the R222A mutation facilitated gating of the receptor, it also increased the proportion of receptors entering into the desensitized state from the open state. It has been proposed for the nACh receptor that gating and desensitization are energetically coupled events (41). Our observation that an increase in gating efficacy is associated with an enhanced desensitization of 5-HT 3A receptors is consistent with such a notion and suggests that desensitization of 5-HT 3A receptors is positively coupled to gating.
In summary, we have found that the R222A mutation can enhance the apparent potency of agonists and the efficacy of partial agonists in 5-HT 3A receptors. Kinetic analysis indicates that these alterations are associated with faster activation and desensitization of this receptor channel. The unique position of Arg-222 may allow it to couple the binding of agonist to the opening of the channel, since the R222A mutation appears to reduce the energy barrier for gating. A reduced energy barrier may speed up activation, which in turn may increase the apparent potency of agonists and the efficacy of partial agonists. Thus, our data suggest that Arg-222 has a functional role in the signal transduction mechanism of 5-HT 3A receptors.