HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dubin, A. E. Articles by Erlander, M. G. Search for Related Content PubMed PubMed Citation Articles by Dubin, A. E. Articles by Erlander, M. G. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 43, 30799-30810, October 22, 1999

The Pharmacological and Functional Characteristics of the Serotonin 5-HT_3A Receptor Are Specifically Modified by a 5-HT_3B Receptor Subunit^*

Adrienne E. Dubin, Rene Huvar, Michael R. D'Andrea^§, Jayashree Pyati, Jessica Y. Zhu, K. C. Joy, Sandy J. Wilson, Jose E. Galindo, Charles A. Glass, Lin Luo, Michael R. Jackson, Timothy W. Lovenberg, and Mark G. Erlander

From the R. W. Johnson Pharmaceutical Research Institute, San Diego, California 92121 and the ^§ R. W. Johnson Pharmaceutical Research Institute, Spring House, Pennsylvania 19477-0776

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

While homomers containing 5-HT_3A subunits form functional ligand-gated serotonin (5-HT) receptors in heterologous expression systems (Jackson, M. B., and Yakel, J. L. (1995) Annu. Rev. Physiol. 57, 447-468; Lambert, J. J., Peters, J. A., and Hope, A. G. (1995) in Ligand-Voltage-Gated Ion Channels (North, R., ed) pp. 177-211, CRC Press, Inc., Boca Raton, FL), it has been proposed that native receptors may exist as heteromers (Fletcher, S., and Barnes, N. M. (1998) Trends Pharmacol. Sci. 19, 212-215). We report the cloning of a subunit 5-HT_3B with ~44% amino acid identity to 5-HT_3A that specifically modified 5-HT_3A receptor kinetics, voltage dependence, and pharmacology. Co-expression of 5-HT_3B with 5-HT_3A modified the duration of 5-HT₃ receptor agonist-induced responses, linearized the current-voltage relationship, increased agonist and antagonist affinity, and reduced cooperativity between subunits. Reverse transcriptase-polymerase chain reaction in situ hybridization revealed co-localization of both 5-HT_3B and 5-HT_3A in a population of neurons in the amygdala, telencephalon, and entorhinal cortex. Furthermore, 5-HT_3A and 5-HT_3B mRNAs were expressed in spleen and intestine. Our data suggest that 5-HT_3B might contribute to tissue-specific functional changes in 5-HT₃-mediated signaling and/or modulation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The biogenic amine serotonin (5-hydroxytryptamine, 5-HT)¹ signals through a myriad of pharmacologically defined cell surface receptors (4, 5), only one of which (the type 3 5-HT₃ receptor) is a fast activating, ligand-gated, nonselective cation channel (1, 2, 6). Activation of 5-HT₃ receptors produces a variety of effects including membrane depolarization and increase in intracellular Ca²⁺ (1, 7), modulation of neurotransmitter release (2, 8), excitation of central and peripheral neurons (1, 2, 8), release of 5-HT from enterochromaffin cells of the small intestine (9), and Ca²⁺ influx into lymphocytes (10). 5-HT₃ receptor activation mediates emetic and inflammatory responses (11) and may contribute to pain reception, anxiety, cognition, cranial motor neuron activity, modulation of affect, and the behavioral consequences of drug abuse (Refs. 12 and 13; but see Ref. 14).

The 5-HT_3A receptor subunit shares structural similarities with members of the superfamily of ligand-gated ion channels (15) and is thought to be a pentameric protein (16, 17) with multiple agonist and allosteric ligand binding sites (2, 11). Both native and recombinant 5-HT₃ receptors reveal rapid and cooperative activation by agonists and desensitization to prolonged application of 5-HT (reviewed in Refs. 1, 2, and 6). With a few exceptions, ligand-gated channels require the association of more than one kind of homologous subunit for function, and subunit composition determines the pharmacological (18) and kinetic (e.g. desensitization (19, 20)) profile of heteromeric receptors. While the 5-HT_3A subunit expressed in heterologous systems functions efficiently as homomers, different voltage-dependence, desensitization, and pharmacological properties between recombinant and native 5-HT₃ receptors suggest that native 5-HT₃ receptors may exist as heteromers (3, 21-27). Although the 5-HT_3A gene encodes splice variants in mouse (28, 29) and guinea pig (30), most of the characteristics of these variants are similar (29, 31, 32), and the subtle pharmacological differences (22, 31, 32) cannot completely account for the differences observed between recombinant homomers and native 5-HT₃ receptors. In fact, co-expression of both splice variants in oocytes could not reproduce the responses observed in the cell line from which the splice variants were cloned (22). Biochemical studies on porcine brain have revealed the existence of at least four proteins (52-71 kDa) closely associated with the 5-HT_3A subunit that may represent antigenically distinct channel subunits (33).

The cloning of a 5-HT_3B subunit that modified the pharmacological and single channel characteristics of the 5-HT_3A subunit when co-expressed in heterologous expression systems was recently reported (34). We extend this report and provide evidence for the co-existence of both 5-HT_3B and 5-HT_3A receptor subunits in native cells, a requirement for heteromeric association. We show that 5-HT_3B has no effect on the function of 22, 34, 42, and 7 nicotinic ACh receptors. Furthermore, the characterization of the pharmacology and function of heteromeric receptors in Xenopus oocyte and mammalian expression systems reported here differs from that described (34), and these differences have important consequences for the function of heteromers in vivo.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cloning of p5HT_3BR-- The amino acid sequence of the 5-HT_3A receptor sequence (accession no. P46098) was used as a query in a tblastn search of the publicly available human genomic sequence data base (35). From this search, a related sequence to the 5-HT_3A receptor (subsequently called 5-HT_3B) was found (accession no. AC002290). To obtain a full-length sequence, a human small intestine size-selected cDNA library was constructed in pSPORT as described by the manufacturer (Life Technologies, Inc.) with the addition of size selection of cDNAs to >2 kilobase pairs. To screen the subsequent library, a ~400-base pair fragment was amplified with specific PCR primers (GATCTCCCTACCTCTAAGTG; AGCACACTGGTCTTGAACAC) to exonic sequences within the genomic sequence using cDNA synthesized from poly(A) RNA extracted from human small intestine. The ~400-base pair fragment was subsequently radiolabeled and used to screen ~500,000 recombinants from the constructed cDNA library by standard molecular biology methods (36). A single positive clone was obtained and subsequently sequenced by a 377 DNA sequencer (Applied Biosystems, Foster City, CA) and determined to be full-length. The Sequencher program (Gene Codes Corp., Ann Arbor, MI) was used for removal of vector sequences and general sequence analyses. The full-length 5-HT_3B receptor was subsequently cloned into pGEM HE and pcDNA3.1zeo(+) (InVitrogen, Carlsbad, CA) for functional expression in Xenopus oocytes and HEK293 cells, respectively.

Transfection of p5-HT_3B into a Mammalian Cell Line Stably Expressing 5-HT_3A Receptor-- HEK293 cells stably expressing p5HT_3A (accession no. D49394) in pCINeo (Promega) were screened for expression using Fluo-4/AM (Molecular Probes, Inc., Eugene, OR) in conjunction with a FLIPR^TM system (Molecular Devices, FL-101, Sunnyvale, CA). The clone revealing the most robust influx of Ca²⁺ upon application of 5-HT was used for all subsequent studies (5-HT_3A/HEK). This line was stably transfected with p5HT_3B in pcDNA3.1zeo(+) (5-HT_3B/5-HT_3A/HEK), and individual clones were maintained under selection (neomycin (500 µg/ml), zeocin (200 µg/ml)).

Ca²⁺ Influx Measurements Using FLIPR^TM-- Transfected cells were plated onto poly-D-lysine-coated black-walled 96-well plates (Biocoat; Becton Dickinson catalog no. 354640) and grown for at least three days to confluency. Cells were rinsed with F-12/Dulbecco's modified Eagle's medium, incubated in Fluo-4 (2 µM; Molecular Probes) with pluronic acid (0.04%; Molecular Probes) for 1 h at room temperature, and assayed using the FLIPR. Cells were challenged with agonists (at 3-fold concentration in 40 µl added to 80 µl at a velocity of 50 µl/s), and the fluorescence intensity was captured every second for the first 2 min following agonist addition, with additional readings every 2 s for 1 min. For antagonist experiments, agonist was added in the presence of antagonist at the indicated concentrations after a 2-min preincubation. The saline used for these experiments contained 130 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 20 mM HEPES (pH 7.4).

Expression of 5-HT_3B in Xenopus Oocytes and Two-electrode Voltage Clamp Recordings-- Xenopus laevis oocytes were prepared and injected using standard methods (37). Ovarian lobes dissected from adult female X. laevis (Nasco, Fort Atkinson, WI) were shipped on ice overnight, defolliculated in 2 mg/ml collagenase (Type 1, ICN Biomedicals, Aurora, OH) in nominally Ca²⁺-free saline. Selected oocytes were maintained in ND-96 (100 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM HEPES, 2.5 mM sodium pyruvate, 50 µg/ml gentamicin, pH 7.0) for 2-24 h before injection. For in vitro transcription, pGEM HE (38) containing 5-HT_3A or 5-HT_3B cDNA was linearized with NheI and transcribed for 2 h at 37 °C with T7 RNA polymerase (T7 mMESSAGE mMACHINE; Ambion) in the presence of the cap analog m7G(5')ppp(5')G. cRNA was quantified using formaldehyde gels against RNA markers (Life Technologies).

Oocytes were injected using a 10µl microdispensor (catalog no. 53506-100; VWR Scientific) with 50 nl of 5-HT_3A receptor cRNA with or without the 5-HT_3B cRNA (0.33 and 0.0033-3.3 ng each) or 5-HT_3B (3.3 ng) together with human nicotinic ACh receptor subunits (0.2 ng of 22, 0.25 ng of 34, 0.2 ng of 42, 0.5 ng of 7). Control oocytes were injected with 50 nl of water. Injected oocytes were maintained in 48-well cell culture clusters (Costar; Cambridge, MA) at 18 °C up to 14 days in ND96. Whole cell agonist-induced currents were measured with a conventional two-electrode voltage clamp (GeneClamp500, Axon Instruments, Foster City, CA). Recording microelectrodes were filled with 3 M KCl and had resistances of 1 and 2 megaohms. Cells were maintained in a 600-µl volume bath and continuously perfused at 10 ml/min at room temperature. The membrane voltage was clamped at -70 mV unless otherwise indicated. Responses of 5-HT_3A-injected oocytes to agonists were measured in either Ca²⁺-containing saline (Ca²⁺-SOS; equivalent to ND-96 without gentamicin) or Ba²⁺-containing saline (Ba²⁺-SOS; Ba²⁺ replaced Ca²⁺, with or without 1 mM EGTA to chelate extracellular Ca²⁺).

While most of the oocytes injected with 5-HT_3A responded to 5-HT with prolonged responses in the continued presence of agonist (see "Results"), a minority (<30%) of oocyte batches responded to 5-HT with an ultrafast decay component (data not shown). The t80 of responses to 10 µM 5-HT was ~10-fold faster (p < 5e⁵) and the time to peak of the response to 5-HT (10 µM) was 3.4-fold shorter (p < 0.05) in these oocytes compared with the oocytes described in this report. The modulation of 5-HT_3A receptors expressed in oocytes by 5-HT_3B was similar with the exception of its effect on decay kinetics. The effect of 5-HT_3B was to shorten the long lasting responses of the majority of oocytes and lengthen the short responses to an identical intermediate value. The decay constant  for heteromers was significantly slower than that of homomeric responses under identical conditions (p < 5e⁶). Thus, 5-HT_3B had opposite effects on the decay of responses from the two populations of oocytes. For clarity, only data for the majority of oocytes are described. It is not immediately apparent why oocyte batches fell into two populations with regard to the kinetics of desensitization to 5-HT. However, if similar populations of responses are observed in mammalian cells, it may account for the less robust effects of 5-HT_3B co-expression on Ca²⁺ influx in populations of cells compared with single oocytes.

Whole Cell Voltage Clamp Recordings from Stable Transfectants-- The whole cell configuration of the patch clamp technique (39) was used to record ligand-induced currents from 5-HT_3A/HEK or 5-HT_3B/5-HT_3A/HEK cells lacking visible association with other cells maintained >2 days on 12-mm uncoated glass coverslips. During recordings, cells were visualized using a Nikon Diaphot 300 with DIC Nomarski optics and continuously perfused in a physiological saline (~0.5 ml/min) consisting of 130 mM NaCl, 4 mM KCl, 1 mM CaCl₂, 1.2 mM MgCl₂, and 10 mM hemi-Na-HEPES (pH 7.3, 295-300 mosM as measured using a Wescor 5500 vapor pressure osmometer (Wescor, Inc., Logan, UT)). Recording electrodes were fabricated from borosilicate capillary tubing (R6; Garner Glass, Claremont, CA). The tips were coated with dental periphery wax (Miles Laboratories, South Bend, IN) and had resistances of 0.6-1 megaohms when containing intracellular saline (100 mM potassium gluconate, 25 mM KCl, 0.483 mM CaCl₂, 3 mM MgCl₂, 10 mM hemi-Na-HEPES, and 1 mM K₄-1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid tetrapotassium salt (100 nM free Ca²⁺); pH 7.4, 290 mosM). Unless indicated, all voltages shown are corrected for a liquid junction potential of 18 mV. Series resistances were <2 megaohms and generally not compensated. Current and voltage signals were detected and filtered at 2 kHz with an Axopatch 1D patch-clamp amplifier (Axon Instruments, Foster City, CA), digitally recorded with a DigiData 1200B laboratory interface (Axon Instruments) and PC-compatible computer system, and stored on magnetic disc for off-line analysis. Data acquisition and analysis were performed with PClamp7 software. Slow changes in holding current were simultaneously detected and filtered at 2 kHz, recorded with an LPF 202A DC amplifier (Warner, Hamden, CT) and VR-10B digital data recorder (Instrutech, Great Neck, NY) onto video tape, and later analyzed at 10 Hz.

Apparent reversal potentials (V_rev) of ligand-induced conductance changes were determined using a voltage-ramp protocol (40). Voltage ramps were applied between 100 and +100 mV from a holding potential of 68 mV at a rate of 1 mV/ms every 1 s.

PCR-based Tissue Distribution of Human 5-HT_3B-- The tissue distribution of 5-HT_3B mRNAs was determined by semiquantitative PCR. A primer set specific to 5-HT_3B (TGTGTTCAAGACCAGTGTGC; TAGCTTTGGAAGAGCAGTCG) was used to complete amplification of a portion of the 5-HT_3B mRNA via PCR using cDNA templates synthesized from poly(A) RNA (CLONTECH, Palo Alto, CA) that was extracted from various human tissues (see legend to Fig. 7). Products were resolved on a 6% polyacrylamide gel and detected using a labeled oligonucleotide (TGTTGGTCAAATTCCTCCATGATGAGCAGCGTGGTGGACA).

Reverse Transcriptase-Polymerase Chain Reaction in Situ Hybridization (RT-PCR ISH)-- RT-PCR ISH was performed as described previously (41) on normal human and monkey tissues using specific primer sets to 5-HT_3B (5'-TGTGTTCAAGACCAGTGTGC; 3'-TAGCTTTGGAAGAGCAGTCG) and 5-HT_3A (5'-CAGAGGATTTCTGCTCAGGC; 3'-CCTCGTACAGAGTTATCAGG). Negative control plasmid sequencing primer sets included 5'-CGCCAGGGTTTTCCCAGTCACGAC-3', and 5'-AGCGGATAACAATTTCACACAGGA-3'. Briefly, with the exception of monkey amygdala, formalin-fixed tissues (Analytical Biological Services, Inc., Wilmington, DE) were paraffin-embedded, cut (5 µm) onto Plus Superfrost slides (Fisher). These and normal human checkerboard tissue slides (Dako, Carperturia, CA) were dried, dewaxed, and hydrated. Fresh frozen Macaque amygdala was sectioned at 10 µM on a cryostat, fixed, and frozen. Tissues were digested in 2 mg/ml pepsin (Roche Molecular Biochemicals) for 30 min at 37 °C, washed in diethyl pyrocarbonate-treated H₂O, dehydrated, and air-dried. All tissues (with the exception of positive controls) were treated with 1 unit/µl DNase I (Roche Molecular Biochemicals) overnight. Slides were placed in aluminum boats (41) on the heating block of a model 480 DNA Perkin-Elmer thermal cycler programmed for 30-min cDNA synthesis at 65 °C, denaturation for 3 min at 94 °C, and then 20 cycles of amplification at 60 °C for 90 s and 94 °C for 1 min. Ten µl of probe mixture (containing reagents of the GeneAmp Kit (Perkin-Elmer), 1 mM digoxygenin dUTP, 40 units of RNasin (Roche Molecular Biochemicals), 2% bovine serum albumin, and 20 µM primers) was added to the tissue and incubated under a fastened coverslip. The presence of reaction product was visualized with alkaline phosphatase-conjugated anti-digoxygenin antibody (1:200; Life Technologies) and nitro blue tetrazolium bromochloroindolyl phosphate (Enzo Diagnostics, Inc., Farmingdale, NY). Sections were counterstained with nuclear fast red (Polyscientific, Bay Shore, NY). Sections were visualized under a BX-50 Olympus light microscope and photographed (Fig. 8) or captured electronically using a Sony 3CCD camera using Image Pro software (Fig. 9; Phase 3 Imaging Systems, Glen Mills, PA).

Analysis of mRNA Expression in Human Tissues Using cDNA Microarrays-- Expression of mRNA from various tissues and cell types was determined using techniques previously described (42). Peripheral blood mononuclear cells were isolated using a standard Ficoll gradient from 1 unit of blood from each of five normal human donors. Monocytes (CD14+) were purified from peripheral blood mononuclear cells by positive selection and magnetic bead technology, and total RNA was isolated using the Qiagen RNeasy Mini Kit. Total RNA from normal human ascending and descending colon (CDP-061068; lot 8906066; CDP-061069; lot 8906067; BioChain Institute, Inc.) and 5 µg of each poly(A) RNA from normal human small intestine and spleen (6547-1; lot 7050565; 6547-1; lot 7050565; CLONTECH) were amplified without further treatment and labeled with cy3-dCTP. The labeled material was hybridized to DNA chips including 5-HT_3B and 5-HT_3A sequences.

Data Analysis-- Values are presented as the mean ± S.E. unless indicated. Ca²⁺ influx peak values were determined by subtracting minimum from maximum values during the agonist exposure. EC₅₀, IC₅₀, and slope values for dose-response data were determined using the Hill equation (sigmoidal dose response with variable slope) using GraphPad Prism (GraphPad, San Diego, CA).

Reagents-- 5-HT, mCPBG, 1-PBG, 2-methyl-5-HT (2-Me-5-HT), quipazine dimaleate, 1-(3-chlorophenyl)piperazine dihydrochloride (mCPP), ketanserine tartrate, LY-278,584 maleate, DA, tropisetron, histamine, tryptophan, norepinephrine, octopamine, homovanillic acid, tryptophol, -methyl serotonin, glycine, GABA, 5-hydroxyindole, -hydroxybutyrate, cis-4-aminocrotonic acid, agmatine, D-cycloserine, N-acetyl-L-cysteine, acetyl-aspartyl-L-glutamic acid, S--histamine, N--methyl histamine, melatonin, N-acetyl serotonin, (±)-epibatidine-2HCl, and tropisetron (3-tropanyl-indole-3-carboxylate HCl) were purchased from RBI. Tyramine, tryptamine, tryptophanamide, glutamate, -alanine, 5-hydroxyindolacetic acid, taurine, 6-hydroxymelatonin, 5-hydroxyindole 2-carboxylic acid, ATP, 3,4-dihydroxyphenylacetic acid, and 5-hydroxyindole 3-acetamide were obtained from Sigma. Bacterial alkaline phosphatase was purchased from Life Technologies. Y-25130 (N-(1-azabicyclo[2.2.2]oct-3-yl)-6-chloro-4-methyl-3-oxo-3,4-dihydro-2H-1,4-benzoxazine-8-carboxamide monohydrochloride) was purchased from Tocris Cookson, Inc. (Ballwin, MO).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The Amino Acid Sequence of the Human 5-HT_3B Reveals Similarity to the 5-HT_3A Receptor Subunit-- The nucleotide sequence of 5-HT_3B revealed a single large open reading frame of 1326-base pair and 5'- and 3'-untranslated extensions of 141 and about 456 nucleotides. The first in-frame methionine was designated as the initiation codon for an open reading frame that predicts a 5-HT_3B protein with 441 amino acids and an estimated molecular mass of about 50.3 kDa. The protein contained hydrophobic amino-terminal residues with sequences highly predictive of signal cleavage sites that would result in a mature protein initiating at amino acid 22 (SignalP). The predicted 5-HT_3B protein was aligned with nucleotide and protein data bases and found to be related to known 5-HT_3A receptors, revealing about 44% identical and approximately 70% highly conserved amino acids compared with the human 5-HT_3A subunit. The identity of the 5-HT_3B receptor with the 5-HT_3A receptor at the nucleotide level was only about 60%. Hydropathy analysis of the amino acid sequence predicted the presence of four putative transmembrane domains (M1-M4) with similar spanning as that predicted for the 5-HT_3A subunit (not shown). The coding sequence reported here is identical to the sequence reported by Davies et al. (34).

Interestingly, the predicted 5-HT_3B protein contains many of the features of the proposed ligand binding sites of ligand-gated receptors (2). Conserved canonical sequences in loops A, B, and C proposed to form the ligand recognition site in the first N-terminal extracellular loop are present in 5-HT_3B, including two cysteine residues found in the conserved cysteine-cysteine loop (2). However, Glu¹⁰⁶, which in the murine 5-HT_3A receptor is critical for high affinity 5-HT binding (43), is replaced by methionine in 5-HT_3B.

5-HT_3B lacks all three negatively charged residues bracketing M2 that are thought to be critical determinants for the rate of ion transport in nicotinic ACh receptors (44). The 5-HT_3A receptor is missing only the "cytoplasmic" ring (45, 46).

Five potential sites of glycosylation (i.e. NX(S/T)) are located at the extracellular amino terminus (Asn⁵², Asn⁹⁶, Asn¹³⁸, Asn¹⁶⁸, and Asn²⁰³). Four potential phosphorylation sites are located in the cytoplasmic loop between M3 and M4: one potential site for protein kinase C (i.e. (S/T)X(K/R)) at Thr³⁷⁷, three for casein kinase II (i.e. (S/T)XX(D/E)) at Thr³⁴⁶, Thr³⁷⁷, and Ser⁴¹⁰, and one for mammary gland casein kinase (i.e. SXE) at Thr³⁹⁴.

5-HT_3B Appears to Be Incapable of Functioning as an -Subunit in Homomers or in Combination with Nicotinic ACh Receptor -Subunits-- Since 5-HT_3B contains most of the consensus sequences for ligand binding sites, we tested whether 5-HT_3B could be activated by a panel of ligands. Oocytes were injected with 5-HT_3B cRNA alone (3.3 ng) or 5-HT_3B together with nicotinic ACh 1, 2, and 3 cRNA and tested 2-14 days later. 5-HT (100 µM) had no effect in oocytes injected with putative 5-HT_3B subunit alone (n = 4 (3.3 ng of cRNA/oocyte), n = 12 (0.33 ng), n = 12 (0.033 ng), and n = 3 (0.0033 ng)). Furthermore, oocytes injected with 5-HT_3B cRNA with or without nicotinic ACh 1, 2, and 3 cRNA were insensitive to 300 neuroactive compounds at 100 µM including 5-HT, ACh, histamine, tyramine, tryptamine, tryptophanamide, tryptophan, norepinephrine, octopamine, DA, 3,4-dihydroxyphenylacetic acid, homovanillic acid, tryptophol, -methyl-5-HT, glutamate, glycine, GABA, -alanine, taurine, -phenylethylamine, 5-hydroxyindolacetic acid, 5-hydroxyindole, 6-hydroxymelatonin, -hydroxybutyrate, cis-4-aminocrotonic acid, agmatine, D-cycloserine, N-acetyl-L-cysteine, acetyl-aspartyl-L-glutamic acid, S--histamine, N--methyl histamine, melatonin, 5-hydroxyindole 2-carboxylic acid, N-acetyl serotonin, and 5-hydroxyindole 3-acetamide (data not shown). To test whether phosphorylation might inhibit sensitivity to these ligands, 5-HT_3B-injected oocytes were injected with bacterial alkaline phosphatase (0.25-0.3 units) at least 30 min prior to recording, but this treatment was ineffective in conferring sensitivity to these ligands (n = 3).

Injection of the 5-HT_3B cRNA had no effect on currents through nicotinic ACh 22, 34, 42, and 7 receptors expressed in oocytes (data not shown). Furthermore, the responses to low and high concentrations of epibatidine (0.3 and 10 µM) were similar in 7 nicotinic ACh-injected oocytes in the presence or absence of 5-HT_3B, indicating that the dose-response relationship was not appreciably altered (data not shown).

5-HT_3B Modifies 5-HT_3A Receptor Function-- The high sequence similarity of 5-HT_3B to the human 5-HT_3A subunit prompted us to co-express both species of cRNA in oocytes. 5-HT₃ receptor agonists produced robust responses in oocytes injected with 5-HT_3A in the presence and absence of 5-HT_3B (Fig. 1, A and B). Interestingly, co-injection of 5-HT_3B cRNA produced striking differences in the desensitization and pharmacological profiles of 5-HT₃ receptor currents during prolonged application of low and high concentrations of agonists compared with 5-HT_3A alone (Fig. 1A, right) and altered the voltage dependence of agonist-induced currents from strongly inward rectifying to nearly linear (n = 13; data not shown). There was no detectable effect of 5-HT_3B on the reversal potential of (V_rev) of 5-HT₃ agonist-induced responses (15 ± 2 mV (n = 7) and 16 ± 2 mV (n = 13) for homomers and heteromers, respectively). Similar V_rev values were obtained when Ba²⁺ replaced Ca²⁺ in the extracellular saline. Specific antagonists could block homomeric (Fig. 1C) and heteromeric 5-HT₃ receptors.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   5-HT3B modifies the kinetics and magnitude of 5-HT3A receptor currents elicited by a subset of agonists. A, 5-HT3A-expressing oocytes responded to 5-HT, 2-Me-5-HT, mCPBG, and 1-PBG with slowly desensitizing responses (left panel). 5-HT, mCPBG, 1-PBG, and 2-Me-5-HT elicited currents with significantly faster decay kinetics when applied to oocytes expressing both subunits (right panel). Tested agonists included 5-HT (0.3 and 10 µM, superimposed responses), mCPBG (0.3 and 10 µM), 1-PBG (10 and 100 µM), DA (0.1 and 1 mM), and 2-Me-5-HT (10 µM). Agonists were applied during the time indicated by the horizontal bar above the record. The open bar extending from the solid bar indicates the longer exposure to agonist for one of the superimposed traces. Oocytes were continuously perfused with Ba2+-SOS at a rate of 10 ml/min at room temperature. Time scale bar, 40 s. B, peak currents elicited by the indicated agonists from oocytes expressing 5-HT3A (gray bars) and both 5-HT3B and 5-HT3A (solid bars). The maximal inward current to 10 µM 5-HT was similar for oocytes bathed in either Ca2+-SOS or Ba2+-SOS, and the results were combined. Peak currents were as follows for oocytes bathed in Ca2+-SOS versus Ba2+-SOS, respectively: 5-HT3A, 6.4 ± 1.0 µA (n = 15) versus 5.0 ± 0.6 µA (n = 41); 5-HT3B/5-HT3A, 8.7 ± 1.6 µA (n = 13) versus 8.1 ± 1.0 µA (n = 22). The number of oocytes tested ranged from 9 to 61; S.E. values were 10% for 5-HT (10 µM) and 21% for both mCPBG (10 µM) and 1-PBG (100 µM). p values are indicated as follows: *, p < 0.05; ** p < 0.01 (Student's t test). C, responses to 1-PBG were blocked by the 5-HT3 antagonist tropisetron. After obtaining a control response to 1-PBG (100 µM; indicated by the solid bar) (left panel), the oocyte was washed for 2 min, incubated for 30 s in tropisetron (1 µM; open bar), and then challenged with both 1-PBG (solid bar) and tropisetron. The response to 1-PBG was completely blocked (middle panel). The response to agonist recovered after a 2-min washout of antagonist (right panel). Responses to 10 µM mCPBG were similarly antagonized (n = 4).

5-HT_3B Alters the Desensitization of 5-HT_3A Receptors-- The majority of oocyte batches (>70%) expressing 5-HT_3A homomers responded to 5-HT with a slowly desensitizing current similar to that reported in other studies expressing human 5-HT_3A (45-48), however, the magnitude of the responses observed in the present study was substantially larger (>2 µA), presumably due to cRNA stabilization by including 5'- and 3'-untranslated regions of the Xenopus  globin gene (38) (Fig. 1B, gray bars). The inward current through 5-HT_3A homomeric receptors declined during the continued presence of high doses of 5-HT (Fig. 1A, left), consistent with the desensitization previously described for recombinant 5-HT_3A receptors (15, 29, 30, 45, 46, 49). The current decay could not be fit by an exponential function, so desensitization was quantified by measuring the time between the rising and falling phases of the response at 80% of the peak (t80; Table I). The duration of 5-HT-induced responses decreased as the concentration of 5-HT was increased. The decay was slightly but significantly faster in Ca²⁺-SOS (t80 = 19 ± 3 s; n = 10) compared with the decay of responses recorded in the absence of external Ca²⁺ (Ba²⁺-SOS) (t80 = 30 ± 3 s; n = 36; p < 0.05).

View this table: [in this window] [in a new window]   Table I Co-expression of 5-HT3B in Xenopus oocytes expressing 5-HT3A decreased the t80 of current responses elicited by prolonged exposure to 5-HT3 receptor agonists Values represent the mean ± S.E. (n) of the time between 80% rise and fall of the peak response (in s) from a holding potential of 70 mV. p values indicating the significance of the difference in t80 values for 5-HT3A and 5-HT3B/5-HT3A receptor responses were determined using the Student's t test. Data from oocytes bathed in Ba2+-SOS with and without EGTA were similar and combined.

In oocytes co-injected with 5-HT_3A and 5-HT_3B, nearly maximal 5-HT-induced currents decayed 5 times more quickly (Fig. 1A, left versus right panels; Fig. 2A; Table I), and the time to peak for responses to 5-HT (10 µM) was 3.3 ± 0.3 s (n = 38) (Fig. 2B, solid bars; data for oocytes expressing heteromers and bathed in Ca²⁺-SOS and Ba²⁺-SOS were similar and combined). In contrast to the complex response waveform observed in the absence of 5-HT_3B, inward currents elicited by 10 µM 5-HT were best fit to a single exponential ( = 4.9 ± 0.3 s; n = 25; Ba²⁺-SOS) in more than 80% of the oocytes expressing heteromers. 5-HT_3B co-expression decreased the t80 values of responses to 5-HT, 2-Me-5-HT, mCPBG, and 1-PBG by 5.6-, 5-, 2.0-, and 1.7-fold, respectively, suggesting agonist-dependent differences (Table I). The t80 was independent of voltage under all conditions (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   5-HT3B co-expression decreased the t80 and time to peak of the 5-HT response. A, t80 of responses evoked by 5-HT (concentrations indicated) are shown for oocytes injected with 5-HT3A (gray bars; n = 8, 28, and 9 individual oocytes, respectively) and both 5-HT3A and 5-HT3B (solid bars; n = 11, 28, and 16 individual oocytes, respectively). The differences were significant at all agonist concentrations (p < 0.05, < 5e9, and < 0.005 for 1, 10, and 100 µM 5-HT, respectively). Extracellular saline was Ba2+-SOS. B, the time to peak of the response to 10 µM 5-HT is significantly faster when 5-HT3B is co-expressed with 5-HT3A (solid bars, n = 22) compared with 5-HT3A alone (gray bars, n = 32, p < 0.0001). Extracellular saline was Ba2+-SOS.

5-HT_3B Alters Agonist Potency and Cooperativity at 5-HT_3A Receptors Expressed in Xenopus Oocytes-- The efficacies of 5-HT, mCPBG, 2-Me-5-HT, and 1-PBG were similar (see legend to Fig. 3); the peak response elicited by 1 mM DA was approximately a quarter of the maximum 5-HT response. The dose-response relationship for 5-HT is shown in Fig. 3A and reveals an apparent EC₅₀ of ~400 nM. Ratios of the responses elicited by low and saturating concentrations of 5-HT (0.3 and 10 µM), mCPBG (0.3 and 10 µM) and 1-PBG (10 and 100 µM) were determined for individual oocytes injected with 5-HT_3A alone (Fig. 3B, stippled bar) and indicate that in oocyte studies the EC₅₀ for mCPBG and 1-PBG were >400 nM and >10 µM, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Agonist dose-response relationships are altered in the presence of 5-HT3B and depend on the ratio of 5-HT3B to 5-HT3A cRNA injected into Xenopus oocytes. A, oocytes were injected with either 5-HT3A or both 5-HT3A and 5-HT3B cRNA and tested for their response to the indicated concentrations of 5-HT. The data are presented relative to the maximum response elicited by 100 µM 5-HT in the same oocytes. 5-HT3B decreased the apparent affinity of the 5-HT3A receptor for 5-HT (5-HT3A, 400 nM (95% confidence interval: 371-452 nM); the number of oocytes tested ranged from 3 (100 nM) to 19; 5-HT3B/5-HT3A: 1.75 µM (494 nM to 6.2 µM), 3-18 oocytes). Agonist was applied to the cell at a rate of 10 ml/min in bath perfusate. The data were fit with the Hill equation using GraphPad Prism. B, ratios of the response to 0.3 µM 5-HT, 0.3 µM mCPBG and 10 µM 1-PBG compared with the maximum response elicited by the same agonist are plotted for oocytes injected with 5-HT3A alone (stippled bars) or in the presence of 5-HT3B at a ratio of 1:10 (solid bars), 1:1 (hatched bars), 1:0.1 (open bars), 1:0.01 (cross-hatched bars) 5-HT3A cRNA/5-HT3B cRNA. The percentage of the maximal response obtained in individual oocytes was averaged. Significant differences are indicated by asterisks (*, p < 0.05; **, p < 0.01). These data also provided an estimate of the relative efficacy of the different agonists compared with 5-HT. In 5-HT3A homomers and 5-HT3B/5-HT3A heteromers, the ratio of the response to 10 µM mCPBG versus 10 µM 5-HT was 0.85 ± 0.04 (n = 10) and 1.1 ± 0.2 (n = 9), respectively; 100 µM 1-PBG versus 10 µM 5-HT was 0.95 ± 0.11 (n = 13) and 1.5 ± 0.2 (n = 5), respectively; 10 µM 2-Me-5-HT versus 10 µM 5-HT was 1.2 ± 0.5 (n = 3) and (0.79 and 1.07), respectively; 1 mM DA versus 10 µM 5-HT was (0.27, 0.67) and (0.12, 0.23), respectively.

Maximum responses to 5-HT, mCPBG, and 1-PBG were moderately but significantly enhanced when 5-HT_3A was co-expressed together with 5-HT_3B cRNA (Fig. 1B, solid bars). Nearly maximal responses were still elicited by 10 µM 5-HT, but the apparent affinity for 5-HT was decreased in the presence of 5-HT_3B (Fig. 3A, triangles). In contrast, oocytes expressing both subunits were more sensitive to application of low concentrations of mCPBG and 1-PBG compared with oocytes expressing 5-HT_3A alone (Figs. 1A and 3B). When ratios of peak response elicited by low and high concentrations of agonist were calculated, co-injected oocytes had a larger relative response to the biphenylguanide derivatives but smaller relative response to 5-HT, and the magnitude of these differences depended on the relative ratio of injected cRNAs (Fig. 3B). The differences in agonist potency were no longer observed when 5-HT_3B cRNA was 100-fold more dilute than 5-HT_3A cRNA (Fig. 3B, cross-hatched bars).

The agonists 1-PBG and mCPBG elicited no response in oocytes injected with 0.33 ng of 5-HT_3B subunit alone (the concentration injected to give a 1:1 ratio; n = 4), indicating that the increase in 1-PBG and mCPBG responsiveness in co-injected oocytes was not due to a direct activation of 5-HT_3B homomers by these agonists. Furthermore, the response to 1-PBG was similarly blocked by tropisetron (1 µM) and LY-278,584 maleate (1 µM) (data not shown) in a reversible manner. The selective 5-HT₂ receptor antagonist ketanserin (10 µM) had no effect on agonist responses (data not shown).

5-HT_3B Modifies 5-HT_3A Receptor Function in Mammalian Cells-- Agonist-induced increases in intracellular Ca²⁺ in mammalian cells transfected with 5-HT_3A in the presence or absence of 5-HT_3B were measured by the FLIPR system. Full agonist dose-response relationships were obtained for seven agonists. EC₅₀ for 5-HT (Fig. 4A), 1-PBG (Fig. 4B), mCPBG, 2-Me-5-HT, quipazine, DA, and mCPP are indicated in Table II. Maximal fluorescent values for these agonists were compared relative to the maximal value obtained for 5-HT in the same series of experiments (the confluency was similar for all cell lines) to obtain an estimate of their relative efficacy. The percentage of the 5-HT response was 110, 103, 91, 80, 73, and 50% for mCPBG, 1-PBG, 2-Me-5-HT, quipazine, DA, and mCPP, respectively (values are averages of four experiments; S.D. were <10% with the exception of 2-Me-5-HT (15%)).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   5-HT3B and 5-HT3A heteromers display pharmacological properties distinct from 5-HT3A homomeric receptors. A, dose response for 5-HT-activated Ca2+ influx using the FLIPR system. 5-HT3A and 5-HT3B heteromeric receptors (triangles) had significantly higher affinity for 5-HT and decreased slope compared with 5-HT3A homomeric receptors (squares). Peak responses were determined and normalized to the maximum observed response. Mean values from 4-14 separate experiments are presented. Data were fit with the Hill equation, and differences in EC50 (Table II) and slope were significant (p < 0.05 for each parameter). The Hill slopes were 3.7 and 1.7 for 5-HT3A/HEK and 5-HT3B/5-HT3A/HEK, respectively, with 95% confidence intervals of 2.5-4.9 and 1.0-2.3, respectively. B, intracellular Ca2+ levels in 5-HT3A/HEK (left panel) and 5-HT3B/5-HT3A/HEK (right panel) cells recorded using the FLIPR after challenge with the indicated final concentrations of 1-PBG. The arbitrary scale on the y axis is the same for all wells.

Robust shifts in dose-response relationships for 5-HT and the biguanide derivatives were observed in 5-HT_3B-expressing cells. In contrast to the results for 5-HT in the oocyte studies, the dose response for 5-HT in the mammalian cells was shifted to the left (Fig. 4A, triangles), a finding that is consistent with the shift observed for the biguanide derivatives applied to oocytes (Fig. 3B) and mammalian cells (Fig. 4B) expressing 5-HT_3A and 5-HT_3B. Complete dose-response relationships for 5-HT (Fig. 4A) and 1-PBG (Fig. 4B) and for mCPBG, 2-Me-5-HT, mCPP, and quipazine (data not shown) indicate a shallower slope and increased agonist affinity in 5-HT_3B-expressing cells (Table II). The potency of DA was also increased by co-expression of 5-HT_3B, but there was no detectable change in the Hill coefficient (data not shown).

Agonist-induced Ca²⁺ and ionic current responses had faster decay kinetics in 5-HT_3B/5-HT_3A/HEK compared with 5-HT_3A/HEK cells (Fig. 4B). Cells expressing both 5-HT_3B and 5-HT_3A (Fig. 4B, right) produced faster decaying Ca²⁺ signals during continued challenge with 1-PBG compared with 5-HT_3A receptor-expressing cells (Fig. 4B, left). At concentrations producing nearly maximal 1-PBG responses, the effect was particularly striking. The fast decay was also observed with mCPBG (data not shown) and 5-HT where t80 values were ~3-fold larger for 5-HT_3A/HEK compared with 5-HT_3B/5-HT_3A/HEK cells (p < 0.025).

The differences in kinetics and pharmacology observed in the Ca²⁺ influx experiments were not due to altered intracellular Ca²⁺ buffering in the Ca²⁺-influx studies. Similar results were observed for 5-HT- and 1-PBG-elicited whole cell currents (Fig. 5A). Cells were challenged with a voltage ramp protocol to determine whole cell conductance, and ramp-induced currents (spikes in Fig. 5A) are shown on a faster time scale in Fig. 5B. Fast application of 10 µM 1-PBG (solid bar) produced significantly smaller inward currents and conductance changes in 5-HT_3A/HEK cells (Fig. 5A, top left) compared with 5-HT_3B/5-HT_3A/HEK cells (Fig. 5A, top right). Subsequent application with 5-HT (open bar) revealed similar large responses in both cell types. 1-PBG (10 µM) elicited responses that were 3 ± 1% (n = 7) of the response to 10 µM 5-HT in individual 5-HT_3A/HEK and a 10-fold higher (28 and 34%) percentage of the response to 10 µM 5-HT in individual 5-HT_3B/5-HT_3A/HEK cells.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Electrophysiological recording from 5-HT3A/HEK and 5-HT3B/5-HT3A/HEK cells revealed distinct pharmacological and voltage-dependent properties. 5-HT3A/HEK cells (left) and 5-HT3B/5-HT3A/HEK cells (right) were voltage-clamped, and agonists were rapidly applied from nearby puffer pipettes. Spikes in the current record (cut off at top) represent evoked currents induced by voltage ramp protocols used to determine the change in whole cell membrane conductance every 1 s. A, top, inward currents elicited by 10 µM 1-PBG (solid bar) and 10 µM 5-HT (open bar) from a holding potential of 68 mV. Bottom, inward currents elicited by 100 µM 1-PBG (gray hatched bar). B, the voltage relationship for 5-HT-induced currents in 5-HT3A/HEK (left) and 5-HT3B/5-HT3A/HEK (right) cells obtained using a voltage ramp protocol. Voltage ramps were applied to cells before (solid circle) and during (open circle) agonist application. The reversal potential (Vrev) was similar for both responses. Asterisks above the recordings in A indicate the ramp currents shown on an expanded scale in B. The voltage axis is corrected for junction potential.

To address whether the discrepancy observed for the 5-HT dose dependence in oocyte and FLIPR studies was due to differences in the speed of agonist addition, mammalian cells were patch clamped, and the effect of rapid application of a low concentration of 5-HT (100 nM) from nearby puffer pipettes was determined. A low concentration of 5-HT had no effect on 5-HT_3A homomers (n = 4) but activated currents in 5-HT_3B/5-HT_3A/HEK cells (n = 6 of 7). Significant differences between cell lines were observed for the proportion of responsive cells (p < 0.01, ² analysis (50)). A maximal concentration of 5-HT elicited responses of similar magnitude in the two cell types (data not shown). This is consistent with the FLIPR results in which 5-HT_3B increased the affinity of 5-HT₃ receptors for 5-HT.

The voltage dependence of the 5-HT induced current was inwardly rectifying in 5-HT_3A-expressing cells and more linear in the presence of 5-HT_3B (Fig. 5B). The ratio of 5-HT-induced current measured at equivalent voltages (50 mV) positive and negative to V_rev was calculated as an indication of the degree of rectification. In 5-HT_3A/HEK cells, the inward current was 2-fold larger than the outward current (ratio of 0.52 ± 0.04; n = 11). In cells expressing both subunits, the current-voltage relationship was linear (ratio of 1.12 ± 0.10; n = 6). This difference in rectification observed between the two transfectants was significant (p < 0.0025). The V_rev of the 5-HT response was similar for both transfectants and was 7.7 ± 2.3 mV (n = 11) and 9.4 ± 1.3 mV (n = 8) for 5-HT_3A/HEK and 5-HT_3B/5-HT_3A/HEK cells, respectively.

In the presence and absence of 5-HT_3B, the Ca²⁺ influx observed during challenge with 1 µM 5-HT was completely blocked by specific 5-HT₃ receptor antagonists (Table II). Response to agonist in 5-HT_3A/HEK and 5-HT_3B/5-HT_3A/HEK cells could be completely blocked by tropisetron, LY-278,584, and Y-25130; however, Y-25130 was a more potent antagonist against 5-HT responses elicited in 5-HT_3B/5-HT_3A/HEK compared with 5-HT_3A/HEK cells (Fig. 6; Table II).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   The antagonist Y-25130 more potently blocked 1 µM 5-HT-induced Ca2+ influx into 5-HT3B/5-HT3A/HEK compared with 5-HT3A/HEK cells. Y-25130 was added after 20 s at the indicated concentrations for 2 min, and cells were subsequently challenged with 1 µM 5-HT in the presence of the indicated concentration of antagonist. The graph shows the relative response compared with the response without antagonist. The IC50 for Y-25130 was 36 nM (19-68 nM) for 5-HT3A/HEK and 4.8 nM (2.6- 8.7 nM) for 5-HT3B/5-HT3A/HEK cells.

Ketanserin, an antagonist at 5-HT₂ receptors, had no effect on 5-HT- or 1-PBG-induced Ca²⁺ responses up to 10 µM. Spiperone, an antagonist at 5-HT_2A and D₂ DA receptors, appeared to be a partial agonist at the 5-HT_3A receptor at concentrations above 1 µM (data not shown).

Distribution of 5-HT_3B and 5-HT_3A mRNA-- PCR-based tissue distribution revealed that the 5-HT_3B mRNA is expressed in human cerebral cortex including occipital, frontal, and temporal regions; amygdala; hippocampus; and testis (Fig. 7). Very low levels were observed in adrenal gland, bone marrow, lymph node, salivary gland, and thyroid gland. No detectable transcript was observed in heart, placenta, lung, liver, skeletal muscle, kidney, pancreas, spleen, thymus, prostate, ovary, small intestine, colon, leukocytes, cerebellum, medulla, spinal cord, putamen, caudate nucleus, corpus callosum, substantia nigra, and thalamus. Transcripts present at extremely low levels or having restricted cellular distribution might not be detected using this method (e.g. mRNA was not detected in small intestine, the tissue from which 5-HT_3B was cloned, or spleen (see Fig. 8)).

View larger version (182K): [in this window] [in a new window]   Fig. 7.   Tissue distribution of 5-HT3B determined by semiquantitative PCR. PCR products of the predicted size (447 base pairs) were observed in cerebral cortex (lane 2); occipital, temporal, and frontal lobes (lanes 5-7); amygdala (9); hippocampus (12); and whole brain (13, 17). Lane 1, cerebellum; lane 2, cerebral cortex; lane 3, medulla; lane 4, spinal cord; lane 5, occipital pole; lane 6, frontal lobe; lane 7, temporal lobe; lane 8, putamen; lane 9, amygdala; lane 10, Caudate nucleus; lane 11, corpus callosum; lane 12, hippocampus; lane 13, whole brain; lane 14, substantia nigra; lane 15, thalamus; lane 16, heart; lane 17, brain; lane 18, placenta; lane 19, lung; lane 20, liver; lane 21, skeletal muscle; lane 22, kidney; lane 23, pancreas; lane 24, spleen; lane 25, thymus; lane 26, prostate; lane 27, testis; lane 28, ovary; lane 29, small intestine; lane 30, colon (mucosal lining); lane 31, peripheral blood leukocyte. The gel was imaged using a PhosphorImager (model 445SI, Molecular Dynamics).

View larger version (85K): [in this window] [in a new window]   Fig. 8.   5-HT3B and 5-HT3A transcripts were expressed in human spleen. RT-PCR ISH revealed intense perinuclear labeling (purple) of 5-HT3B (a) and 5-HT3A mRNA (b) in small monocytes (arrows) in the spleen. Fast red staining of nuclei in b indicates negatively labeled cells. Scale bar, 17 µM.

We used RT-PCR ISH to determine whether both 5-HT_3A and 5-HT_3B were expressed in a variety of primate tissues. Positive mRNA labeling was visualized as purple intracellular precipitates. All cellular nuclei were intensely labeled in positive control slides in which DNase treatment was omitted (41) (data not shown). Negative control slides included the use of a nonsense primer pair as well as the lack of digoxygenin-dUTP and revealed no detectable labeling (data not shown). Primer sets specific to 5-HT_3B and 5-HT_3A (see "Experimental Procedures") were used to determine the histological expression pattern in various normal human and monkey tissues. Results indicate that both messages were localized in small monocytes of the human spleen (Fig. 8) and tonsil (data not shown). Fig. 8 (a and b, respectively) shows intense perinuclear labeling of 5-HT_3B and 5-HT_3A (arrows). 5-HT_3B and 5-HT_3A mRNA was also detected in small and large intestines, uterus, prostate, ovary, and placenta (data not shown). Consistent with this labeling pattern, 5-HT_3B and 5-HT_3A mRNA were detected in spleen, small and large intestine, and CD14+ monocytes using microarray hybridization techniques (Table III).

In an effort to determine the co-expression within a single cell, we performed RT-PCR ISH in serial sections of the monkey amygdala (Fig. 9), monkey entorhinal cortex, and human cerebral cortex (data not shown). A small population of co-labeled neurons was revealed in all three tissues. Low power micrographs of 5-HT_3A (Fig. 9a) and 5-HT_3B (Fig. 9b) show labeling patterns for both targets in neuronal profiles in serial sections of the lateral amygdala. Upon higher magnification (Fig. 9, c and d), both messages clearly co-exist in seven neurons (arrows). In the cerebral cortex, co-labeled cells tended to be pyramidal in shape (data not shown).

View larger version (138K): [in this window] [in a new window]   Fig. 9.   Co-localization of 5-HT3B and 5-HT3A transcripts in neurons in the monkey amygdala was detected using RT-PCR ISH. Seven neuronal cells (arrowheads) from the lateral amygdala express both 5-HT3A (a and c) and 5-HT3B (b and d) mRNA. Not all cells were labeled by the RT-PCR ISH technique. Fast red nuclear staining is observed only in cells negative for either 5-HT3A or 5-HT3B expression. An asterisk marks a vessel as reference in the serial sections. Contrast and brightness were increased using Image Pro version 3.0. Scale bar, 40 µM.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The present report describes the identification, cloning, localization, and pharmacological and functional characterization of a novel 5-HT₃ receptor subunit, 5-HT_3B. This subunit appears to function only in conjunction with 5-HT_3A and not as a homomeric receptor and confers unique pharmacological and functional properties upon the 5-HT_3A receptor. Importantly, the mRNA encoding 5-HT_3B is expressed in 5-HT_3A-expressing cells. During the course of these studies, the sequence and preliminary functional analysis of this novel gene was published (34). The present report corroborates and extends aspects of the study by Davies et al. (34), describes novel 5-HT_3B-dependent modification of 5-HT_3A function, and provides evidence for co-localization of 5-HT_3B with 5-HT_3A subunits in single cells.

5-HT_3B Specifically Modifies 5-HT_3A Receptor Function

Pharmacology-- The pharmacological profile of human 5-HT_3A receptors expressed in oocytes and mammalian cells was similar to that previously described (30, 45-47), with some notable exceptions. The rank order of potency was quipazine 5-HT ~ mCPBG > 2-Me-5-HT > mCPP > 1-PBG DA. Quipazine was also a potent agonist at murine recombinant 5-HT_3A receptors (51, 52). In oocyte and mammalian studies, the EC₅₀ for 5-HT was about 400 nM, ~6-9-fold lower than values observed for human recombinant 5-HT_3A previously reported (30, 34, 45, 46, 48). The differences in agonist pharmacology may be due to different levels of expression achieved for 5-HT_3A in the different heterologous expression systems (53-55), different degrees of desensitization, and/or different experimental methods. The levels of expression we observed were at least 5-fold higher than those previously reported.

We found the rank order of efficacy relative to 5-HT was mCPBG > 5-HT ~ 1-PBG ~ 2-Me-5-HT > quipazine > DA mCPP. DA and quipazine were previously shown to be partial agonists at 5-HT₃ receptors (51, 56, 57). Furthermore, mCPP, a nonselective agonist/antagonist at 5-HT receptors (58, 59) is a partial agonist at the human receptor. While 1-PBG was a full agonist in the present study, 1-PBG was shown to be a partial agonist in a previous study (46).

Co-expression of 5-HT_3B had striking effects on 5-HT_3A receptor agonist and antagonist pharmacology. First, the potency of mCPBG and 1-PBG for heteromeric receptors was significantly greater than for homomeric receptors in both electrophysiological and Ca²⁺-imaging studies. Second, the apparent affinity for 5-HT was altered; however, the direction of the shift depended on the experimental method employed. Full dose-response curves were obtained simultaneously for 5-HT_3A/HEK and 5-HT_3B/5-HT_3A/HEK on the same 96-well plate, where all cell types were treated similarly (e.g. Fig. 4B). All three agonists revealed a similar shift to higher affinity and a decrease in Hill coefficient in using the FLIPR system. This is in contrast to the rightward shift for 5-HT dose dependence measured in oocyte experiments. The discrepancy may be accounted for by differences in experimental protocols and the differential degrees of desensitization of the 5-HT₃ receptor by 5-HT and the biguanides (see below). Oocytes were challenged with a relatively shallow agonist concentration gradient prior to attaining the indicated concentration, and in the most quickly desensitizing responses (5-HT_3B/5-HT_3A expressing cells), the maximum dose may not be achieved prior to the onset of receptor desensitization. Mammalian cells challenged with 100 nM using a fast perfusion system identical to that used by Davies et al. (34) responded to 100 nM 5-HT only if they expressed both subunits (5-HT_3B/5-HT_3A/HEK cells), indicating that the affinity for 5-HT was enhanced by co-expression of 5-HT_3B in mammalian cells as determined by FLIPR and patch clamp methods. Discrepancies between our data and the previous report of a lack of effect of 5-HT_3B on mCPBG potency (34) may be due to differences in the cells used in these studies, since the modulation of 5-HT_3A receptor function and pharmacology was best observed 3 days after plating the stable cells, and Davies and colleagues used cells 2-3 days after transient transfection.

5-HT_3B contains a methionine at the equivalent position as Glu¹⁰⁶ in the mouse 5-HT_3A, which influences ligand binding (43). Substitution of an asparagine (E106N) decreased the Hill coefficient toward unity, suggesting a loss of cooperativity, and decreased the affinity for bath-applied 5-HT (43). If both subunits in the 5-HT_3B/5-HT_3A heteromer contribute to the ligand binding site, then substitution of a methionine may be responsible at least in part for altering agonist affinities.

The antagonists tropisetron, Y-25130, and LY-278,584 had low nanomolar affinity for human 5-HT_3A homomers, similar to previously reported values (30, 56, 60, 61). While no difference in affinity for LY-278,584 and tropisetron was observed in the presence of 5-HT_3B, Y-25130 had a 7.5-fold higher affinity in cells expressing heteromers. This contrasts with the decrease observed for D-tubocurarine (34).² High affinity antagonism by Y-25130 has been observed in frog DRG (IC₅₀ = 500 pM) and the indirect sympathomimetic response in the isolated rabbit heart (antagonist potency of 0.1 nM) (60, 62); interestingly, the potency of Y-25130 was 1000-fold less in blocking the contraction of the guinea pig ileum longitudinal smooth muscle (60). The differential antagonist profiles for 5-HT₃ receptor heteromer and homomers has important clinical implications, especially since there appears to be controversial evidence for the ability of 5-HT₃ receptor antagonists to improve cognition, anxiety, modulation of affect, and substance abuse (14). Both receptor populations were indistinguishable in their affinity for ondansetron (34). Future studies will examine whether the receptor populations distinguish among other antagonists such as granisetron and enantiomers of zacopride.

The observed enhancement of peak currents to 5-HT, mCPBG, and 1-PBG by 5-HT_3B co-expression in oocytes may be due in part to the increase in single channel conductance previously reported (34). 5-HT_3B was previously found to increase the efficacy of mCPBG and 1-PBG compared with 10 µM 5-HT (34).

Desensitization-- 5-HT_3B modified the duration of the response elicited by continued application of the agonists 5-HT, 2-Me-5-HT, mCPBG, and 1-PBG to 5-HT_3A receptors expressed in mammalian cell lines and the majority of oocytes. 5-HT_3B dramatically shortened responses and appeared to convert the receptors to a single population of sites whose rate of decay could be fit by a single exponential function (oocyte and mammalian cell studies). The differences in kinetics we observed in the presence of 5-HT_3B were due to alterations in 5-HT₃ receptor function and not to activation of contaminating endogenous Ca²⁺-activated Cl channels. The modulation of channel properties by 5-HT_3B was observed in salines depleted of Ca²⁺ and at membrane potentials near the chloride equilibrium potential. While 5-HT_3A receptors are permeable to Ca²⁺ (25, 34, 63, 64), the influx of Ca²⁺ appears to be insufficient to activate endogenous chloride currents (65, 66). Our quantification of kinetic parameters provides a means of comparing response durations and onset between populations of oocytes treated under similar experimental conditions. It is clearly not an accurate description of the underlying channel activity, because the method of application and oocyte geometry precludes rapid exchange of solutions (66). To further substantiate the differences in decay kinetics observed in the oocyte experiments, we recorded from mammalian cells expressing 5-HT_3A, with and without 5-HT_3B, applied agonists from nearby puffer pipettes and observed that current responses elicited by 5-HT and mCPBG decayed more rapidly in 5-HT_3B-expressing cells. Furthermore, measurement of Ca²⁺ influx using the FLIPR revealed that the decay of the response was more rapid in the presence of 5-HT_3B.

Desensitization kinetics is known to vary among cells, and even within the same cell type (1, 2). The onset of desensitization and the character of the kinetic properties vary from very slow and linear (Ref. 30; present study) to fast and mono- or biexponential (for a review, see Ref. 1). Many studies have reported a Ca²⁺ and voltage-dependence for the rate of desensitization of 5-HT_3A receptors (1). We observed a weak Ca²⁺ dependence but no voltage dependence. A lack of voltage dependence was also observed for rat nodose ganglion neurons (67) and undifferentiated NG108-15 cells (68). Other members of this ligand-gated superfamily reveal a strong dependence of desensitization kinetics on subunit composition (20, 69, 70).

Substitution at L286 in the M2 domain of the mouse 5-HT_3A receptor profoundly alters the duration of the 5-HT response through homomers (71). The equivalent amino acid in the 5-HT_3B subunit is a valine (Val²⁷⁹). While this substitution was not made in the study by Yakel and colleagues (71), a conserved change (alanine) dramatically shortened the duration of the response as well as linearized the current-voltage relation. It is possible that in the human receptor, 5-HT_3B Val²⁷⁹ lines the pore and interacts with 5-HT_3A Leu²⁸² in the heteromer such that the kinetics and voltage dependence of heteromers are faster and more linear, respectively, compared with homomers. How this mutation alters the structure of 5-HT_3A homomers is not known.

We can only speculate as to the mechanism underlying the 5-HT_3B effect on current kinetics. One possibility is t