The recombinant 5-HT_1A receptor: G protein coupling and signalling pathways

Abstract

The 5-hydroxytryptamine 5-HT_1A receptor was one of the first G protein coupled receptors whose cDNA and gene were isolated by molecular cloning methods. Transfection of the cDNA of this receptor into cells previously bearing no 5-HT receptors has resulted in the acquisition of large amounts of information regarding potential signal transduction pathways linked to the receptor, correlations of receptor structure to its various functions, and pharmacological properties of the receptor. Transfection studies with the 5-HT_1A receptor have generated critical new information that might otherwise have been elusive. This information notably includes the discovery of unsuspected novel signalling linkages, the elucidation of the mechanisms of receptor desensitization, the refinement of models of the receptor pharmacophore, and the development of silent receptor antagonists, among others. The current review summarizes the most important studies of the recombinant 5-HT_1A receptor in the decade since the identificiation of its cDNA.

British Journal of Pharmacology (1999) 127, 1751–1764; doi:10.1038/sj.bjp.0702723

Abbreviations:

cyclic AMP

3′,5′-cyclic adenosine monophosphate

CFTR

cystic fibrosis transmembrane regulator

DAG

diacylglycerol

Erk

extracellular signal-regulated kinase

G protein

guanine nucleotide regulatory binding protein

G_i

G protein that inhibits adenylyl cyclase

GIRK

G protein-gated inwardly rectified K⁺ channel

G_o

G protein that serves functions other than to regulate adenylyl cyclase

G_q

G protein that activates phospholipase C

Grb2

protein that serves as a molecular adapter

GRK

G protein-coupled receptor kinase

GTP

guanosine triphosphate

GTPγS

nonhydrolysable GTP analogue

G_S

G protein that stimulates adenylyl cyclase

8-OH-DPAT

8-hydroxy-2-(di-n-propylamino)-tetralin

5-HT

5-hydroxytryptamine

serotonin; i2 loop

putative second intracellular loop of the 5-HT_1A receptor protein

i3 loop

putative third intracellular loop of the 5-HT_1A receptor protein

IκBα

inhibitor of NF-κB

IP₃

inositol triphosphate

Mek

mitogen and extracellular signal regulated kinase, which phosphorylates and activates Erk

NF-κB

nuclear factor-κB

PC-PLC

phosphatidylcholine-specific phospholipase C

PI-PLC

phosphatidylinositol-specific phospholipase C

PKA

cyclic AMP-dependent protein kinase

PKC

protein kinase C

PLA₂

phospholipase A₂

PLC

phospholipase C

Raf

a kinase that is activated by Ras and that phosphorylates and activates MEK

Ras

a monomeric low molecular weight G protein that activates Raf

RGS proteins

regulators of G protein signalling

ROI

reactive oxygen intermediates

R-SAT

receptor selection and amplification technology

Shc

a protein that serves as a docking platform

Src

non-receptor tyrosine kinase

WAY-100635

N-[2-[4-(2-methoxyphenyl)1-piperazinyl]ethyl)-n-(2-pyridiynl)cyclohexanecarboxamide trihydrochloride

Introduction

Serotonin (5-HT, 5-hydroxytryptamine) was discovered in 1948 by Rapport et al. as a potent vasotonic factor. For several decades, many of the effects of 5-HT were attributed to two major subtypes of 5-HT receptors (^{Gaddum & Picarelli, 1957}). With the development in the 1980s of specific ligands for the various subtypes of 5-HT receptors came the realization that there must be more than two subtypes of 5-HT receptor. Molecular cloning studies over the last decade have confirmed the existence of at least 14 subtypes of 5-HT receptors, each encoded by a distinct gene. Most of those receptors belong to a large family of receptors that transduce signals through G proteins.

One of the best-characterized 5-HT receptors is the 5-HT_1A receptor (^{Pedigo et al., 1981}). This receptor is characterized pharmacologically (like all 5-HT₁ receptors) by its high affinity for 5-HT. It also has a uniquely high affinity for 8-OH-DPAT, and for second generation, arylpiperazine anxiolytic agents such as buspirone, gepirone and ipsapirone. Physiological, clinical, and pharmacological studies have documented potential roles for the 5-HT_1A receptor in neuroendocrine function and thermoregulation (^{Balcells-Olivero et al., 1998}; ^{Seletti et al., 1995}), vasoreactive headaches (^{Leone et al., 1998}), sexual behaviour (^{Maswood et al., 1998}), food intake (^{Gilbert et al., 1988}; ^{Yamada et al., 1998}), tooth-germ morphogenesis (^{Moisewitsch et al., 1998}), memory (^{Edagawa et al., 1998}), immune function (^{Iken et al., 1995}), aggression (^{Miczek et al., 1998}), depression (^{Blier et al., 1997}; ^{Shiah et al., 1998}), and anxiety (^{Parks et al., 1998}; ^{Ramboz et al., 1998}).

The 5-HT_1A receptor was one of the first G protein-coupled receptors for which the cDNA and gene were cloned (^{Albert et al., 1990}; ^{Chanda et al., 1993}; ^{Fargin et al., 1989}; ^{Fujiwara et al., 1990}; ^{Kobilka et al., 1987}; ^{Stam et al., 1992}). This gene is intronless, and its message is expressed mainly in the brain, spleen, and neonatal kidney (^{Fargin et al., 1989}; ^{Kobilka et al., 1987}). A number of interesting observations have been derived from the primary nucleic and amino acid sequences of the 5-HT_1A receptor. The rat and human 5-HT_1A receptor nucleic acid sequences are 88% homologous with each other (^{Albert et al., 1990}; ^{Fargin et al., 1989}; ^{Fujiwara et al., 1990}; ^{Kobilka et al., 1987}; ^{Stam et al., 1992}), whereas those sequences possess significantly less homology with other members of the family of G protein-coupled 5-HT receptors such as the 5-HT_2A (19%) and 5-HT_2C (18%) receptors and the 5-HT_1D receptors (43%) (^{Fujiwara et al., 1993}). The potential importance of the 5-HT_1A receptor has been underscored by the recent cloning of putative homologues in non-mammalian species such as Xenopus laevis (^{Maracci et al., 1997}) and Fugu rubripes (^{Yamaguchi & Brenner, 1997}).

The encoded human protein, composed of 422 amino acids is characterized by a core molecular weight of ∼46,000, and an isoelectric point of 8.8. Hydropathicity analysis reveals that the 5-HT_1A receptor contains seven hydrophobic stretches that could possibly form membrane-spanning α-helices. By analogy with the β₂-adrenoceptor, and because of the presence of three consensus sequences for N-linked glycosylation on the amino terminus, the receptor is probably oriented in the plasma membrane with the amino terminus facing the extracellular domain. Hydrophilic sequences that form three intracellular and three extracellular loops connect the seven hydrophobic transmembrane regions. In the putative second extracellular domain the 5-HT_1A receptor possesses a cysteine residue (Cys¹⁸⁶) which may form a disulfide bond with Cys¹⁰⁹, which is located at the limit between the first extracellular loop and the third transmembrane domain (Figure 1). By analogy with what has been shown for the β₂-adrenoceptor (^{Dohlman et al., 1990}; ^{Fraser et al, 1989}), this disulfide bond may stabilize the receptor conformation and explain in part, why reducing agents affect 5-HT_1A receptor binding properties (^{Emerit et al., 1991}; ^{Gozlan et al., 1988}). It should be noted that the original sequence of the human receptor (^{Kobilka et al., 1987}) was modified slightly by ^{Chanda et al. (1993)}, because the initial publication of ^{Kobilka et al. (1987)} appears to have contained a sequencing error near the junction of the second intracellular loop and the fourth transmembrane domain.

Figure 1

Two-dimensional topographical representation of the human 5-HT_1A receptor. Each circle represents an amino acid. Numbering of the amino acids commences from the initiator methionine, and every 50 amino acids are marked for convenience. The amino terminus faces the extracellular space, and the carboxyl terminus faces the cytoplasm. Seven transmembrane stretches, each composed of ∼20–26 hydrophobic amino acids, are connected by three extracellular loops (termed e1, e2, and e3) and three intracellular loops (termed i1, i2, and i3). There are three potential sites of N-linked glycosylation on the amino terminus (depicted as branching trees). A disulfide bond between Cys¹⁰⁹ and Cys¹⁸⁷ putatively links the e1 and e2 loops. Transmembrane (TM) domains contain residues important for ligand binding. Asp⁸² (in TM2) and/or Asp¹¹⁶ (in TM3) are important for binding of 5-HT, perhaps by serving as a counterion for the amine group. Ser199 (in TM5) also is important for binding of 5-HT. Asn²⁰⁰ (in TM5) is important both for agonist and antagonist binding. Asp³⁸⁶ (in TM7) is critical for selective binding to β-adrenoceptor blocker like pindolol. Ser³⁹³ and Ser³⁹⁶ (also in TM7) are important for agonist binding. The intracellular domains contain sites that are important for contacting G proteins particularly the entire i2 loop and the carboxy terminal end of the i3 loop (gray circles). Sites putatively involved in phosphorylation by PKC and PKC are depicted as black circles. Specific PKC phosphorylation sites are as follows: Lys¹⁴⁷-Arg-Thr-Pro-Arg¹⁵¹ (i2); Arg²²⁷-Lys-Thr-Val-Lys²³¹ (i3); Lys²⁵²-Ser-Leu²⁵⁴ (i3); and Arg³⁴¹-Lys-Thr-Val-Lys³⁴⁵ (i3). There are two putative PKA sites: Arg²²⁷-Lys-Thr²²⁹ and Arg³⁴¹-Lys-Thr³⁴³. Underlined residues are putative phosphate acceptor sites. The carboxy terminal tail is putatively anchored to the interior face of the plasma membrane by one or two palmitoyl anchors bound to Cys⁴¹⁸ and/or Cys⁴²⁰.

The 5-HT_1A receptor is quite interesting in that it has been implicated in numerous signalling pathways in physiologically relevant model systems. In nearly every case, the signals have been shown to be exquisitely sensitive to pertussis toxin, exclusively implicating G_i/o protein in signals initiated by the 5-HT_1A receptor in physiological settings. In neurons, the major signals emanating from the 5-HT_1A receptor are inhibition of adenylyl cyclase (^{De Vivo & Maayani, 1986}; ^{Weiss et al., 1986}), and opening of K⁺ channels (^{Andrade & Nicoll, 1986}; ^{Colino & Halliwell, 1987}; ^{Zgombick et al., 1989}). The 5-HT_1A receptor has also been demonstrated to inhibit a Ca²⁺ current (^{Pennington & Kelly, 1990}), to stimulate adenylyl cyclase (^{Markstein et al., 1986}; ^{Shenker et al., 1985}), and to inhibit phospholipase C activation (^{Claustre et al., 1988}) in various neuronal preparations. In non-neuronal rat ventral prostate cells, endogenous 5-HT_1A receptors inhibit adenylyl cyclase and stimulate nitric oxide synthase (^{Carmena et al., 1998}).

The realization that the 5-HT_1A receptor couples to multiple signalling pathways in cells and tissues in which it is normally expressed reflects a relatively new understanding of the potential promiscuity of receptor-G protein signalling pathways. Nevertheless, it has long been suspected that single receptor subtypes might be linked to various second messengers in a single cell system (^{Limbird, 1988}). No matter how carefully constructed the experiments, however, there is always the consideration in complex tissues and organs that the effects may be mediated by more than one receptor subtype, which are pharmacologically similar but functionally distinct. These considerations can be eliminated by inserting a receptor through transfection methods into a cellular model which previously lacked that or related receptors. Transfected cells have the advantage of expressing single well-defined receptor subtypes. Non-transfected or dummy-transfected cells serve as excellent controls. The purpose of this review is to summarize information that has been obtained by transfection of the recombinant 5-HT_1A receptor into various cell models. We will focus first upon issues relevant to signal transduction and then upon other functions of the 5-HT_1A receptor.

Signalling linkages

Coupling of the 5-HT_1A receptor to many of the signalling pathways described in tissues has been recapitulated in various transfected cell lines, with two notable exceptions. First, positive coupling of the 5-HT_1A receptor to adenylyl cyclase has not been documented in any transfected cell line to date. Various manoeuvres in CHO cells, including prolonged treatment with pertussis toxin and/or co-stimulation with G_Sα-coupled receptors failed to reveal a positive regulation of adenylyl cyclase or cellular cyclic AMP levels. Further, no direct coupling between the 5-HT_1A receptor and G_Sα has been detected with G protein photoaffinity labelling, cholera toxin catalyzed ADP-ribosylation reactions, or co-immunoprecipitation experiments (^{Raymond et al., 1993b}; and unpublished observations). Second, coupling to nitric oxide has not been described in transfected, cultured cells, but the discovery of this pathway in rat ventral prostate was only recently described (^{Carmena et al., 1998}).

Inhibition of adenylyl cyclase

The most consistent coupling of the recombinant 5-HT_1A receptor is to the inhibition of adenylyl cyclase through pertussis toxin-sensitive G proteins. The coupling to the inhibition of adenylyl cyclase appears to be universally expressed (^{Banerjee et al., 1993}; ^{Fargin et al., 1989}; ^{Liu & Albert, 1991}; ^{Varrault et al., 1992b}), and is extremely efficient in that the efficacy of coupling is maximal at low physiologically relevant levels of receptor expression. The efficiency of coupling is probably due to abundant receptor reserve. ^{Langlois et al. (1996)} demonstrated that 5-HT_1A receptors transfected into polarized epithelial LLC-PK₁ cells were expressed on both basolateral and apical membranes. Receptors on both surfaces were able to inhibit cyclic AMP accumulation. Thus, the recombinant 5-HT_1A receptor has been shown to consistently inhibit adenylyl cyclase in multiple cells, and also on two key membrane domains in a polarized cell line.

Despite its obvious importance, the mechanism of the inhibition of adenylyl cyclase by the 5-HT_1A receptor remains largely unexplored. The family of adenylyl cyclases consists of at least nine distinct members, each possessing a diverse repertoire of regulatory inputs (^{Hurley, 1999}; ^{Mons et al., 1998}; ^{Taussig & Zimmerman, 1998}). Because the 5-HT_1A receptor can activate a multitude of signalling pathways (described below), it could regulate adenylyl cyclase in a number of ways depending both upon the specific signals generated by receptor occupancy, and upon the types and amounts of adenylyl cyclases expressed in the cells or tissues of interest. For example, G_iα subunits bind to and inhibit types 5 and 6 adenylyl cyclase. This would seem to be a rather straightforward means of regulation of adenylyl cyclase by the 5-HT_1A receptor since this receptor almost universally couples to G_iα subunits. However, G_βγ subunits can conditionally stimulate type 2 adenylyl cyclase when activated G_sα is present. Elevations of intracellular Ca²⁺ also can inhibit types 5 and 6 adenylyl cyclase. In contrast, activation of Ca²⁺/calmodulin can stimulate types 1, 3, and 8 of adenylyl cyclase. Protein kinase C can stimulate types 2 and 7 adenylyl cyclase, whereas Ca²⁺/calmodulin-dependent protein kinase II can inhibit type 3 adenylyl cyclase (^{Hurley, 1999}; ^{Mons et al., 1998}; ^{Taussig & Zimmerman, 1998}). Moreover, the 5-HT_1A receptor could regulate cellular levels of cyclic AMP downstream of adenylyl cyclase by inhibiting its destruction by phosphodiesterases or other mechanism (^{Wang et al., 1999}). Thus, what on the surface appears to be a rather straightforward regulation of adenylyl cyclase by the 5-HT_1A receptor may in fact be a complex summation of a number of distinct regulatory signals. Transfected cell systems hold promise for exploring this issue because they offer the ability to alter the amount and types of adenylyl cyclase present within the host cell.

Coupling to phospholipases

Activation of phosphatidylinositol-specific phospholipase C (PLC) results in the generation of two key second messengers. The first is inositol trisphosphate (IP₃), which regulates intracellular Ca²⁺ release (^{Berridge et al., 1998}). The second is diacylglycerol (DAG), which binds to and activates protein kinase C (PKC). PLC can be activated by receptors that couple to both pertussis toxin-sensitive and -insensitive G proteins. Activation of PLC by G_q-coupled 5-HT₂ receptors appears to be almost universal. Although 5-HT_1A receptors can clearly activate PLC, this effect is highly host cell-specific. ^{Fargin et al. (1989)} first demonstrated that the 5-HT_1A receptor could activate PI-PLC in HeLa cells. This coupling was shown to be as effective as that induced by endogenous histamine H₁-like receptors (^{Raymond et al., 1989}). However, this coupling was not as efficient as coupling to the inhibition of adenylyl cyclase in that the EC₅₀ of PLC stimulation occurred at significantly higher ligand concentrations, and because virtually no receptor reserve was apparent. This coupling may be physiologically relevant in that endogenous 5-HT_1A receptors in human Jurkat (T cell-like) cells activates PLC (^{Aune et al., 1993}). ^{Liu & Albert (1991)} were the first to demonstrate that activation of PI-PLC by the 5-HT_1A receptor is cell-specific. They showed that the 5-HT_1A receptor expressed in Ltk⁻ fibroblasts activates PI-PLC because 5-HT increased phosphoinositide hydrolysis and levels of intracellular Ca²⁺. In BALB/c-3T3 cells and in Ltk⁻ fibroblasts, 5-HT_1A receptor-mediated increases in intracellular Ca²⁺ derive from the release of intracellular Ca²⁺, rather than from influx of extracellular Ca²⁺ (^{Abdel-Baset, 1992}). In contrast, when the 5-HT_1A receptor was expressed in GH₄C₁ pituitary cells, no evidence of PI-PLC activity was detected (^{Liu & Albert, 1991}). Interestingly when the human 5-HT_1A receptor was expressed in Xenopus oocytes, it was shown to activate PLC (^{Ni et al., 1997}). In aggregate, the work of several laboratories has shown that the 5-HT_1A receptor couples to PI-PLC in HeLa cells (^{Boddeke et al., 1992}; ^{Harrington et al., 1994}; ^{Middleton et al., 1990}), in Ltk⁻ cells (^{Liu & Albert, 1991}), in Xenopus oocytes (^{Ni et al., 1997}), and in BALB/c-3T3 cells (^{Abdel-Baset et al., 1992}), but only weakly or not at all in CHO cells (^{Cowen et al., 1997}; ^{Newman-Tancredi et al., 1992}; ^{Raymond et al., 1992}), Cos-7 cells (^{Fargin et al., 1989}), NIH-3T3 cells (^{Varrault et al., 1992a}), or GH₄C₁ cells (^{Liu & Albert, 1991}).

Activation of protein kinase C (PKC) occurs in HeLa cells (^{Middleton et al., 1990}; ^{Raymond et al., 1989}), and depends upon phospholipase C activation (^{Fargin et al., 1989}). This effect is probably mediated by G protein βγ subunits, and almost certainly depends upon the expression of βγ-regulatable PLC. The coupling of the 5-HT_1A receptor to PKC is as efficacious as that induced by the endogenous histamine receptor expressed in HeLa cells. The activation of PKC by the 5-HT_1A receptor is relevant in that it increases Na⁺-dependent phosphate uptake to a level similar to that induced by the endogenous histamine receptor (^{Raymond et al., 1991}). Thus, in HeLa cells, the 5-HT_1A receptor regulates an endogenous transport process via PKC activation. However, like the activation phosphoinositide hydrolysis in HeLa cells, the coupling is much less efficient than is the coupling to the inhibition of adenylyl cyclase. Activation of PKC and Na⁺-dependent phosphate uptake is more efficacious in cells expressing ∼3 pmol of receptors mg⁻¹ protein than in cells expressing ∼500 fmol of receptors mg⁻¹ protein (^{Raymond et al., 1989}).

The recombinant 5-HT_1A receptor has also been shown to activate PLA₂ in HeLa cells (^{Harrington et al., 1994}), and to augment Ca²⁺-induced arachidonic acid metabolism in CHO cells (^{Raymond et al., 1992}). ^{Cowen et al. (1997)} showed the 5-HT_1A receptor activates phosphatidylcholine-specific phospholipase C (PC-PLC) in CHO cells. However, details of those linkages remain to be elucidated.

Regulation of channels

The recombinant 5-HT_1A receptor has been shown to regulate the function of several distinct types of channels, including inwardly rectified K⁺ channels, high conductance anion channels, CFTR Cl⁻ channels, and Ca²⁺ channels. G protein-gated inwardly rectified K⁺ (GIRK) channels mediate hyperpolarizing postsynaptic potentials in the nervous system and in the heart during activation of G_i/oα-coupled receptors, including the 5-HT_1A receptor (^{Andrade & Nicoll, 1986}; ^{Colino & Halliwell, 1987}; ^{Zgombick et al., 1989}). The regulation of GIRK channels by receptors relies upon the interaction of G protein βγ subunits (released by receptor activation) with regulatory sites on the channels (^{Doupnik et al., 1996}). ^{Karschin et al. (1991)} used a highly efficient recombinant vaccinia virus vector system to express the 5-HT_1A receptor in primary cultures of rat atrial myoctyes, and documented that the 5-HT_1A receptor could stimulate an endogenous atrial inward rectifier K⁺ current. Those studies were expanded by co-injecting rat atrial RNA with 5-HT_1A receptor RNA into Xenopus oocytes (^{Dascal et al., 1993}), which resulted in the expression of a G protein-activated K⁺ channel that could be activated by the 5-HT_1A receptor. Moreover, the 5-HT_1A receptor was also able to activate GIRK1 channels when receptor and channel RNAs were co-injected into Xenopus oocytes (^{Doupnik et al., 1997}).

One curious observation has been that the kinetical characteristics of GIRK regulation by receptors are markedly different in transfected cells vs in cells that natively express the GIRK channels and receptors. In neurons and atrial myocytes, the time courses for receptor-mediated GIRK current deactivation are 20–40 times faster than are those observed in systems in which cloned receptors and GIRK channels have been co-expressed heterologously (^{Andrade & Nicoll, 1986}; ^{Colino & Halliwell, 1987}; ^{Dascal et al., 1993}; ^{Karschin et al., 1991}; ^{Zgombick et al., 1989}). That finding suggested that additional components might be required to confer the rapid kinetical properties of the native transduction pathway. ^{Doupnik et al. (1997)} studied the effects of co-expression of the 5-HT_1A receptor, GIRK1, and various ‘regulators of G protein signalling’ (RGS proteins) in Xenopus oocytes. They found that they could restore rapid activation and deactivation to GIRK current waveforms evoked by activation of 5-HT_1A receptors by co-expression of RGS1, RGS3, or RGS4, but not by RGS2. This work provided evidence for functional regulation of 5-HT_1A receptor-mediated GIRK activation by RGS1, RGS3, and RGS4.

The 5-HT_1A receptor has been shown to regulate several other channels in transfected cells. ^{Uezono et al. (1993)} showed that the 5-HT_1A receptor expressed in Xenopus oocytes could augment the activation of CFTR Cl⁻ channels induced by β₂-adrenoceptors. This effect was indirect in that the conditional activation of CFTR by the 5-HT_1A receptor was enhanced by co-expression of adenylyl cyclase type II and G_sα, and likely proceeded via G protein βγ subunits. ^{Ni et al. (1997)} demonstrated that the 5-HT_1A receptor expressed in Xenopus oocytes could also stimulate an oscillatory Ca²⁺-activated Cl⁻ current. ^{Mangel et al. (1993)} showed that the 5-HT_1A receptor inhibited a high conductance anion channel in CHO cells through either G_iα2 or G_iα3. The potential significance of the inhibition of high-conductance anion channels by the 5-HT_1A receptor is not known, but those channels are thought to be important in cell volume regulation and the maintenance of membrane potential. ^{Liu & Albert (1991)} demonstrated that the rat 5-HT_1A receptor inhibited Bay K8644-mediated Ca²⁺ influx in GH₄C₁ cells, and that this effect required expression of G_oα (^{Liu et al., 1994}). The ability of the 5-HT_1A receptor to inhibit Ca²⁺ influx has been confirmed in a preliminary manner in several putative neuronal cell lines (NCB-20, F11, and HN2) (^{Singh et al., 1996b}). Thus, the 5-HT_1A receptor can stimulate or inhibit multiple distinct ion channels in transfected cells.

Coupling to active ion transport processes

In addition to regulating channels, the 5-HT_1A receptor has been shown to activate several active ion transport processes. In HeLa cells, the receptor stimulates Na⁺-dependent phosphate uptake through a PKC-mediated pathway (^{Raymond et al., 1989}; ¹⁹⁹⁰), and Na⁺/K⁺-ATPase through a Ca²⁺-mediated pathway (^{Middleton et al., 1990}). Both pathways probably depend upon G_βγ-mediated stimulation of PLC. In CHO cells, the 5-HT_1A receptor activates Na⁺/H⁺ exchange (^{Garnovskaya et al., 1997}; ¹⁹⁹⁸) through a pathway that requires G_iα, Src tyrosine kinase, and PI3-K, but not Mek, Ras, or Raf. The coupling to Na⁺-dependent phosphate uptake suggests a potential role for the 5-HT_1A receptor in regulating cellular energy processing, whereas the coupling to and Na⁺/K⁺-ATPase and Na⁺/H⁺ exchange suggests potential roles in cell volume regulation.

Coupling to G proteins

The ability of the 5-HT_1A receptor to couple to various G protein subunits has been carefully studied using several different expression systems. Manning's group (^{Barr et al., 1997}; ^{Butkerait et al., 1995}) used Spodoptera frugiperda (Sf9) cells for co-expression of the human 5-HT_1A receptor with mammalian G protein subunits. They assessed receptor/G protein coupling by [³⁵S]-GTPγS binding and by guanine nucleotide-sensitive agonist binding assays. Co-expression of the receptor with members of the α_i group (but not others) together with various combinations of β₁ and γ subunits increased the affinity for agonists. Using a similar system, ^{Mulheron et al. (1994)} documented that the 5-HT_1A receptor could also functionally couple to an endogenous G_o-like G protein in Sf9 cells. When the receptor was co-expressed with β₁ and γ₂, relatively equivalent coupling to α_i1, α_i2, α_i3, α_o and α_z was seen, whereas there was essentially no detectable coupling to α₁₂, α₁₃, α_s and α_q. When β₁ and α_i1 were co-expressed with various γ-subunits, the following rank order of affinity was established: γ₂∼γ₃∼γ₅>γ₁ (^{Barr et al., 1997}; ^{Butkerait et al., 1995}).

Other groups have detected measurable differences in the affinity of 5-HT_1A receptor for G_i/o/z family α-subunits in various transfected cell systems. ^{Clawges et al. (1997)} added purified G protein subunits to Sf9 cell membranes expressing 5-HT_1A receptors, and established that α_i1, α_i2, α_i3, and α_o were able to shift the receptors to a high-affinity state in the presence of either brain or retinal β₁/γ. α-Transducin (G_tα) subunits were inactive regardless of which β₁/γ preparation was used. A significantly higher affinity for agonist was observed for the 5-HT_1A receptor in the presence of α_i3 compared with either α_i2 or α_o. ^{Bertin et al. (1992)} expressed 5-HT_1A receptors in E. coli to determine a relative rank order of affinity for this receptor to reconstituted purified mammalian G protein α-subunits of G_iα3>G_iα2>G_iα1≫G_oα≫G_sα. Another group (^{Garnovskaya et al., 1997}; ^{Raymond et al., 1993b}) demonstrated agonist-promotable physical coupling of the 5-HT_1A receptor to G proteins in HeLa and CHO cells using high affinity agonist binding and co-immunoprecipitation assays. Agonist treatment induced coupling of the 5-HT_1A receptors to G proteins with an apparent rank order of G_iα3>G_iα2∼G_iα1∼G_oα>G_zα≫G_sα. Thus, despite using markedly different transfections systems (bacteria, insect and mammalian cells) there is reasonable consensus that the 5-HT_1A receptor functionally couples nearly exclusively through α-subunits of G_i/o/z proteins, with a likely rank order of G_iα3>G_iα2G_iα1G_oα>G_zα. There is currently little evidence that the 5-HT_1A receptor can couple to G_tα, G_sα, G_qα, G_12α or G_13α.

Some groups have taken these studies a step further by trying to link specific G protein subunits with signals that occur downstream of the G proteins. ^{Fargin et al. (1991)} used an immuno-neutralization method to link G_iα3 to the inhibition of adenylyl cyclase and to the activation of PI-PLC in HeLa cells. ^{Raymond et al. (1993b)} used a similar method in CHO and HeLa cells to demonstrate that 5-HT_1A receptor-mediated inhibition of adenylyl cyclase can be mediated by either G_iα3 or G_iα2, and probably G_iα1. These findings are not surprising in light of the work of ^{Wong et al. (1992)} that demonstrated that G_iα1, G_iα2, G_iα3, and G_zα, can all inhibit cyclic AMP accumulation in mammalian cells. ^{Gettys et al. (1994)} used selective photoaffinity labelling of G_iα subunits in CHO cells to demonstrate a close correlation between the ability of a panel of 5-HT_1A receptor agonists to activate G_iα2 and to inhibit adenylyl cyclase, whereas there was much less correlation between G_iα3 and the inhibition of adenylyl cyclase. Immunoneutralization studies suggested that G_iα3 and G_iα2 were about equally important for the inhibition of adenylyl cyclase in HeLa cells, whereas G_iα2 was more important in CHO cells. The differences in the relative importance of G_iα2 and G_iα3 may have been affected by the relative abundance of the G protein α-subunits in HeLa and CHO cells. In HeLa cells, G_iα3 is much more highly expressed than G_iα2 (which is almost undetectable by immunoblot), whereas in CHO cells, there is ∼9 fold more G_iα2 than G_iα3 (^{Raymond et al., 1993b}).

^{Liu et al. (1994)} used an elegant antisense approach to show that G_iα2 was primarily responsible for 5-HT_1A receptor-mediated inhibition of adenylyl cyclase in GH₄C₁ cells, whereas G_oα was responsible for the inhibition of Ca²⁺ channels. Using the same system, they have also shown that G_iα2 is responsible for 5-HT_1A receptor-mediated increases in intracellular Ca²⁺ in Ltk⁻ cells (^{Albert et al., 1996}). ^{Garnovskaya et al. (1997)} showed that either G_iα2 or G_iα3 could couple the 5-HT_1A receptor to the activation of Na⁺/H⁺ exchange in CHO cells, whilst G_iα1, G_oα, and G_zα could not. ^{Langlois et al. (1996)} presented evidence to suggest that 5-HT_1A receptors inhibit adenylyl cyclase via G_iα3 on the apical cell surface, and via G_iα2 on the basolateral surface of polarized epithelial LLC-PK₁ cells. In aggregate, these studies link the effects of the 5-HT_1A receptor to the inhibition of adenylyl cyclase, activation of Na⁺/H⁺ exchange, and activation of PLC through G_iα2 or G_iα3. The inhibition of Ca²⁺ channels appears to require G_oα.

DNA synthesis, growth, and transformation

5-HT receptors coupled to pertussis toxin-sensitive G proteins have previously been implicated as growth stimulatory (^{Ishizuka et al., 1992}; ^{Seuwen et al., 1988}). Furthermore, in glial cells, endogenous 5-HT_1A receptors stimulate secretion of the S100 protein, which has been shown to promote the growth of serotoninergic neurons (^{Lauder, 1993}; ^{Whitaker-Azmitia, 1991}). However, a direct link between the 5-HT_1A receptor and growth cascades has not been demonstrated in cells in which the receptor is expressed naturally. Therefore, three different transfected cell lines have been used to study the links of 5-HT_1A receptors to DNA synthesis, growth and cellular transformation. ^{Abdel-Baset et al. (1992)} transfected BALB/c-3T3 fibroblasts with the rat 5-HT_1A receptor, and showed that 5-HT induced enhanced incorporation of [³H]-thymidine into DNA in a clone that expressed 600 fmol mg⁻¹ of protein of [³H]-8-OH-DPAT. Long-term treatment of cell cultures resulted in phenotypical transformation and foci formation in transfected, but not in non-transfected cells. Those effects were sensitive to pertussis toxin, and were attributed to the actions of PI-PLC. Another group contemporaneously showed that the 5-HT_1A receptor expressed in NIH-3T3 cells (at 40–500 fmol mg⁻¹ of protein) induced focus formation, weakly stimulated DNA synthesis, and potentiated increases in DNA synthesis initiated by epidermal growth factor (^{Varrault et al., 1992a}). Those effects were also sensitive to pertussis toxin, and did not appear to involve PI-PLC or lowering of intracellular cyclic AMP levels. ^{Lam et al. (1996)} showed that the 5-HT_1A receptor expressed in Rat-1 fibroblasts increased proliferation as measured by the R-SAT method (^{Messier et al., 1995}). Thus, the 5-HT_1A receptor can stimulate proliferation and/or transformation in several transfected cell types as measured by different methods.

Regulation of transcriptional cascades

The 5-HT_1A receptor has been linked to two specific signalling pathways involved in the regulation of transcription, namely activation of Erk (extracellular signal-regulated kinase) family mitogen-activated protein kinases, and the transcriptional regulatory factor, NF-κB (nuclear factor-κB). ^{Cowen et al. (1997)} used transfected CHO cells to document that 5-HT_1A receptor agonists activate a signalling pathway that stimulates NF-κB, probably through accelerating the degradation of an inhibitor of NF-κB called IκBα. Phosphorylation of IκBα causes dissociation from NF-κB, resulting in nuclear translocation of NFκ-B, and accelerated degradation of IκBα via ubiquitination (^{Berg & Baldwin, 1993}; ^{Brown et al., 1993}). ^{Cowen et al. (1997)} also showed that the degradation of IκBα was sensitive to inhibition of phosphatidylcholine-specific phospholipase C (PC-PLC).

Several groups have documented that the 5-HT_1A receptor activates Erk kinases in CHO cells, and have identified a number of signalling molecules involved in that pathway. Like other G protein-coupled receptors (^{Luttrell et al., 1997}; ^{Marshall, 1995}), the 5-HT_1A receptor activates Erk through a complex pathway that involves many of the same molecules used by growth factor receptor tyrosine kinases. The activation of Erk by the 5-HT_1A receptor is initiated by βγ subunits released from pertussis toxin-sensitive G proteins. This results in tyrosine phosphorylation of Shc, a protein that serves as a docking platform. Shc phosphorylation results in the recruitment of a lipid kinase, and an adapter protein, Grb2 to the signalling complex. Grb2, in turn, binds to a Ras activator protein called Sos. Ras activation leads to sequential activation of Raf, which phosphorylates and activates MEK (mitogen and extracellular signal regulated kinase), which phosphorylates and activates Erk (^{Garnovskaya et al., 1996}). The precise role of PI3-K in propagating the Erk signal is not known, but it is clearly necessary for 5-HT_1A receptor-mediated Erk activation, and most likely operates at the level of Shc and Grb2 (^{Cowen et al., 1996}; ^{Garnovskaya et al., 1996}; ¹⁹⁹⁸). More recent studies have also implicated a Ca²⁺/calmodulin-dependent endocytosis step between Ras and Raf (^{Della Rocca et al., 1999}), and reactive oxygen intermediates upstream of Src in 5-HT_1A receptor-mediated ERK activation in CHO cells (Mukhin, Raymond and Garnovskaya, unpublished observation). An NAD(P)H oxidase enzyme probably produces the reactive oxygen species.

^{Cowen et al. (1996)} hypothesized a role for phosphatidylcholine-specific PLC (PC-PLC) in 5-HT_1A receptor-mediated Erk activation based on their observations that 8-OH-DPAT elicited release of radioactivity from cells pre-loaded with [³H]-choline. This effect and Erk activation were both attenuated by a PC-PLC inhibitor (tricyclodecan-9-yl-xanthogenate, D609). If PC-PLC is required for 5-HT_1A receptor-mediated Erk activation in CHO cells, it most likely functions at or upstream of Raf based on several studies in other systems. Introduction of bacterial PC-PLC activates Raf, and D609 blocks Raf activation induced by epidermal growth factor and serum in NIH-3T3 cells. Dominant negative Raf constructs block replication, and this block is not overcome by introduction of bacterial PC-PLC (^{Cai et al., 1992}; ¹⁹⁹³). Moreover, dominant negative Ras attenuates the activation of PC-PLC in NIH-3T3 cells (^{Cai et al., 1993}). Based on these studies and their own work, ^{Cowen et al. (1996)} hypothesized that PC-PLC augments the activation of Raf that is induced by Ras. If this were so, it would appear that the 5-HT_1A receptor in CHO cells activates Erk through the activation of separate, yet converging lipid signalling pathways. The first involves PI3-K and intersects the Erk pathway upstream of Ras, whereas the second involves PC-PLC and also intersects the Erk pathway upstream of Ras.

Desensitization and phosphorylation

Desensitization is a process through which signalling pathways become progressively less able to mount a signal after repeated exposures to a stimulus. One major mechanism through which desensitization of G protein-coupled receptors occurs is kinase-directed phosphorylation of the receptors. Thus far, three distinct protein kinases have been implicated in the desensitization and phosphorylation of the 5-HT_1A receptor, namely PKC, PKA, and GRK (G protein-coupled receptor kinase). There are four putative PKC sites and two putative PKA sites in the human 5-HT_1A receptor (see Figure 1 and its legend). Stimulation of PKC by application of phorbol esters induces a rapid phosphorylation of the receptor at a stoichiometry of two phosphates per receptor (^{Raymond, 1991}). This phosphorylation is associated with desensitization of several signals in various cell lines. In CHO cells, PKC-mediated desensitization of the inhibition of adenylyl cyclase by the 5-HT_1A receptor is manifested only by a change in potency of agonist, and is not associated with a significant change in efficacy (^{Raymond, 1991}), whereas in P11 rat pituitary cells, PKC-mediated desensitization is manifested by a change in both and efficacy of 8-OH-DPAT to cause inhibition of adenylyl cyclase (^{Hensler et al., 1996}). In contrast, phorbol ester pretreatment of Ltk⁻ fibroblasts and GH₄C₁ cells transfected with the 5-HT_1A receptor has no detectable effects on the inhibition of adenylyl cyclase (^{Lembo et al., 1995}; ¹⁹⁹⁷; ^{Liu & Albert, 1991}). In Ltk⁻ fibroblasts, activation of PKC with phorbol ester results in desensitization of 5-HT_1A receptor-mediated PLC activity as measured by increased intracellular Ca²⁺ and hydrolysis of phosphoinositides. The effects of phorbol ester on Ca²⁺ mobilization were profound, essentially eliminating subsequent responses to 5-HT. Desensitization could be reversed by mutation of three putative PKC sites in the i3 loop of the 5-HT_1A receptor (^{Lembo et al., 1995}), providing strong support for a functional link between phosphorylation and desensitization of the receptor. ^{Harrington et al. (1994)} provided further evidence for PKC-mediated desensitization of the 5-HT_1A receptor. They showed that pretreatment with phorbol ester of 5-HT_1A-receptor transfected HeLa cells rapidly uncoupled the receptor from G proteins, and that inhibitors of PKC could block this effect. Thus, the evidence is fairly clear that PKC can induce desensitization and phosphorylation of the 5-HT_1A receptor in transfected cells, although the functional consequences of the interaction varies depending on the cell type and signalling pathways being measured.

There is also strong evidence that activation of PKA can lead to phosphorylation and desensitization of the 5-HT_1A receptor in multiple cell types. In CHO cells, PKA stimulation leads to phosphorylation of the receptor with a stoichiometry of one phosphate per receptor. Pharmacological stimulation of PKA also augments PKC-induced phosphorylation and desensitization of the inhibition of adenylyl cyclase by the 5-HT_1A receptor (^{Raymond & Olsen, 1994}). Those results confirmed the observations of ^{Liu & Albert (1991)}, who first showed that activators of PKA enhanced the phorbol ester-induced desensitization of 5-HT_1A receptor coupling to PLC in Ltk⁻ fibroblasts. In HeLa cells, ^{Harrington et al. (1994)} documented rapid PKA-induced reduction in high affinity agonist binding as well as a total attenuation of the ability of the 5-HT_1A receptor to inhibit adenylyl cyclase.

The 5-HT_1A receptor can mediate its own desensitization. Prolonged treatment with 5-HT leads to downregulation and desensitization of 5-HT_1A receptors in Swiss 3T3 cells (^{van Huizen et al., 1993}). In HeLa cells, treatment with 8-OH-DPAT leads to rapid uncoupling of the receptor from G proteins and from the inhibition of adenylyl cyclase. These effects appear to depend upon PKC, PLA₂, and Ca²⁺ (^{Harrington et al., 1994}). In P11 cells, pretreatment with the agonist 5-carboxamidotryptamine results in a rightward shift of the concentration-response curve of 8-OH-DPAT-inhibited adenylyl cyclase activity (^{Hensler et al., 1996}). In Sf9 insect cells, pretreatment with 5-HT leads to rapid phosphorylation of the 5-HT_1A receptor on serine and threonine residues, and to uncoupling of the receptor from G proteins and from the inhibition of adenylyl cyclase (^{Nebigil et al., 1995}). This effect could be blocked by heparin, but not by inhibitors of PKC, and was thus attributed to the actions of a GRK. There are 17 serine and threonine residues on the intracellular loops that might serve as potential GRK phosphorylation sites.

Correlation of receptor structure and function

Despite being one of the first G protein coupled receptors to be cloned, relatively few detailed structure-function studies have been performed on the 5-HT_1A receptor. One of the major distinctive pharmacological characteristics of the 5-HT_1A receptor is a high affinity for classical β-adrenoceptor blockers such as pindolol. ^{Guan et al. (1992)} delineated a single residue critical in the binding of β-adrenoceptor blockers to the 5-HT_1A receptor. They mutated Asn³⁸⁶ in the seventh transmembrane domain of the human 5-HT_1A receptor, based on the observation that this residue is uniquely conserved in all 5-HT1A and β-adrenoceptors of different species. Mutation of this residue to valine caused a selective decrease in the affinity of pindolol and similar ligands for the mutant Asn³⁸⁶→Val receptor, whilst producing insignificant changes in the binding of other 5-HT_1A receptor ligands. Thus, Asn³⁸⁶ is critical for binding to β-adrenoceptor blockers like pindolol, but not to 5-HT. ^{Kuipers et al. (1997)} examined the effects of the same mutation on a range of aryloxypropanolamine enantiomers and further suggested that Asn³⁸⁶ functions as a chiral discriminator in that the Asn³⁸⁶→Val mutation more significantly lowered the affinities of the S-enantiomers in a ligand binding assay.

^{Ho et al. (1992b)} used a vaccinia infection-transfection method to transiently express wild-type and mutant 5-HT_1A receptors into COS-7 cells in order to study the effects of various point mutations in putative transmembrane regions on receptor ligand binding. Three substitutions, Asp⁸²→Asn, Asp¹¹⁶→Asn, and Ser¹⁹⁹→Ala, resulted in a 60–100 fold decreased affinity of 5-HT for the receptor, but had no effect on the affinity of the antagonist, pindolol. The binding of 5-HT to a fourth mutant, Thr²⁰⁰→Ala, was not measurable. Nevertheless, 5-HT induced GTPase activities for all of the mutant receptors studied. These findings indicate that Asp⁸², Asp¹¹⁶, and Ser¹⁹⁹ play important roles in the binding of 5-HT, but have little effect on pindolol binding. Thr²⁰⁰ is important in binding to both 5-HT and to pindolol. By analogy with the β-adrenoceptor, Asp⁸² and/or Asp¹¹⁶ are likely to act as a counterion for the amine group of 5-HT (^{Strader et al., 1989}). While the first acidic residue is conserved in all cloned G protein-coupled receptors reported thus far, the second one is present in only the receptors which bind the bioamines epinephrine, norepinephrine, dopamine, acetylcholine and serotonin. These findings have been nicely integrated into a model of the 5-HT_1A receptor pharmacophore by ^{Kuipers et al. (1994)}.

^{Varrault et al. (1994)} used a different approach to study potential G protein contact sites within the 5-HT_1A receptor. They constructed synthetic peptides derived from the second (i2) and third (i3) intracellular loops of the human 5-HT_1A receptor, and assessed the ability of those peptides to modulate the binding of a nonhydrolysable GTP analogue to G_i/o, and to inhibit adenylyl cyclase. A peptide consisting of the entire i2 loop (Asp¹³³-Arg¹⁵³) and a heptadecapeptide from the carboxyl terminal region of the i3 loop (Ala³³¹-Leu³⁴⁷), but not a nonapeptide from the carboxyl terminal region of the i3 loop (Ala³³⁶-Val³⁴⁴), inhibited forskolin-stimulated adenylyl cyclase activity in membranes derived from NIH-3T3 cells, S49 cells, and rat hippocampus, and increased GTPγS binding to purified bovine brain G_i/o proteins. Thus, those studies identified the entire i2 loop and a carboxyl terminal heptadecapeptide of the i3 loop of the human 5-HT_1A receptor as key G protein regulatory sites. These sequences contain key threonine residues shown to be involved in receptor desensitization and signal transduction by ^{Lembo et al. (1995}; ¹⁹⁹⁷). More recently, ^{Albert et al. (1998)} presented evidence that supported a key role for Thr¹⁴⁹ in the i2 loop of the rat receptor in coupling specifically to G_βγ-mediated signals. These studies support a model in which the i2 loop and portions of the i3 loop form amphipathic α-helices aligned such that a hydrophobic G protein interaction site is formed, as described by ^{Albert et al. (1998)}.

^{Lembo et al. (1995}; ¹⁹⁹⁷) used site-directed mutagenesis to examine the roles of intracellular threonine and serine residues in the i2 and i3 loops of the rat 5-HT_1A receptor. They found that four specific residues played roles in signal initiation, and also in PKC-mediated desensitization of the receptor. A detailed mutagenesis study of the PKC sites on the receptor was needed because PKC-mediated phosphorylation reactions could affect the function of the receptor (^{Raymond, 1991}), G_i proteins (^{Drummond, 1985}; ^{Daniel-Issakani et al., 1989}; ^{Bushfield et al., 1990}; ^{Yatomi et al., 1992}; ^{Strassheim & Malbon, 1994}) or PLC (^{Ryu et al., 1990}; ^{Ali et al., 1997}), any of which effects could lead to desensitization of this pathway. By constructing and expressing several mutant receptors with conservative point mutations, they showed that two threonines and one serine residue located within consensus PKC phosphorylation sequences in the i3 loop were needed to confer PKC-mediated desensitization of Ca²⁺ mobilization in Ltk⁻ fibroblasts. When the individual point mutations (Thr²²⁹→Ala, S²⁵³→Gly, and Thre³⁴³→A) were tested, there was no difference in desensitization when compared with non-mutated receptors. In contrast a double mutant (Thr²²⁹→Ala/S²⁵³→Gly) and a triple mutant (Thr²²⁹→Ala/S²⁵³→Gly/Thre³⁴³→Ala) became progressively more resistant to PKC-mediated desensitization. Those results suggest that there is a good correlation between the presence of those three residues and PKC-mediated desensitization of the 5-HT_1A receptor.

^{Lembo et al. (1997)} also assessed the role of Thr¹⁴⁹ in signal transduction by expressing a mutant rat 5-HT_1A receptor (Thr¹⁴⁹→Ala) in Ltk⁻ fibroblasts and in GH₄C₁ cells. They found that the mutant receptor lost its ability to elevate intracellular Ca²⁺ in Ltk⁻ cells, and also was unable to inhibit opening of BayK8644 sensitive Ca²⁺ channels in GH₄C₁ cells. In contrast, the Thr¹⁴⁹→Ala mutation only partially uncoupled the receptor from adenylyl cyclase inhibition. Interestingly, the Thr²²⁹→Ala mutation also partially uncoupled the receptor from the inhibition of adenylyl cyclase, and reduced the peak increase in intracellular Ca²⁺ in Ltk⁻ cells (^{Lembo et al., 1995}). Thus, the work of ^{Lembo et al. (1995}; ¹⁹⁹⁷) suggests those three residues in the i3 loop of the 5-HT_1A receptor (Thr²²⁹, S²⁵³, and Thre³⁴³) mediated PKC-induced desensitization of the rat receptor. Additionally, Thr²²⁹ is important for efficient coupling to both inhibition of adenylyl cyclase and elevations of intracellular Ca²⁺. Thr¹⁴⁹ in the i2 loop is highly critical in both elevations of intracellular Ca²⁺ and blockade of Ca²⁺ channels by the receptor, and plays a smaller, but significant role in the inhibition of adenylyl cyclase.

Allelic variants of the 5-HT_1A receptor

The possibility that the 5-HT_1A receptor might possess allelic variants was supported by the cloning of the rat receptor by two different groups (^{Albert et al., 1990}; ^{Fujiwara et al., 1990}) whose published nucleic acid sequences differed only by two nucleotides. One of the variants was silent at the protein level, whereas the other resulted in a predicted change in a single amino acid in the rat 5-HT_1A receptor (^{Fujiwara et al., 1993}). Several groups have subsequently described variations in the 5′-untranslated and the coding regions of the human 5-HT_1A receptor gene. Some of the nucleic acid changes in the coding block do not result in amino acid changes, whereas several others result in single amino acid changes; these include Pro¹⁶→Leu, Ile²⁸→Val, Gly²²→Ser, Arg²¹⁹→Leu; Gly²⁷²→Asp; Asn⁴¹⁷→Lys (^{Erdmann et al., 1995}; ^{Harada et al., 1996}; ^{Kawanishi et al., 1998}; ^{Lam et al., 1996}; ^{Nakhai et al., 1995}). The existence of these variant sequences raises the interesting possibility that there may be functional differences that could lead to disease manifestations. Although a detailed biochemical analysis of each of the variants has not yet been performed, ^{Brüss et al. (1995)} studied the Ile²⁸→Val variant in transfected COS-7 cells. They demonstrated that this receptor variant had ligand binding properties that were nearly identical to those of the wild-type receptor (^{Brüss et al., 1995}). In addition, to this point, no definitive links in populations between the allelic variants have been made to alcoholism, schizophrenia, bipolar affective disorder, or Tourette's syndrome.

Ligand pharmacology

Cells transfected with the 5-HT_1A receptor could provide a ready source for the study of this receptor absent of any input from other 5-HT receptors. However, the usefulness of pharmacological information obtained from transfected cell systems required some measure of validation. Several groups have directly compared the ligand binding and second messenger coupling parameters of transfected 5-HT_1A receptors with those naturally expressed in tissues. They have concluded that transfected cell systems are valid for studying the pharmacology of the 5-HT_1A receptor (^{Pauwels et al., 1993}; ^{Pou et al., 1997}). Transfected cell lines have been used to characterize radioligands (^{Sundaram et al., 1992}; ¹⁹⁹⁵) and identify and characterize partial agonists (^{Arthur et al., 1993}; ^{Assie et al., 1997}; ^{Pauwels et al., 1993}; ¹⁹⁹⁷) and inverse agonists (^{Barr & Manning, 1997}; ^{Newman-Tancredi et al., 1998}) of the 5-HT_1A receptor. They have also been used in the development of silent antagonists of the 5-HT_1A receptor, which has been a particularly nettlesome task because most putative silent antagonists actually turned out to be partial agonists (^{Routledge, 1996}). Finally, WAY-100635, (N-[2-[4-(2-methoxyphenyl)1-piperazinyl]ethyl)-n-(2-pyridiynl) cyclohexanecarboxamide trihydrochloride) was characterized as a silent antagonist in a number of in vivo and in vitro models (^{Fletcher et al., 1996}). Critically, the properties of WAY-100635 as a true silent antagonist of the 5-HT_1A receptor were confirmed in a transfected CHO cell system using GTPγS binding (^{Newman-Tancredi et al., 1996}) and measurements of the metabolic production of protons (^{Dunlop et al., 1998}).

Transfected cells have also been used to characterize the interactions of novel ligands with the 5-HT_1A receptor. For example, transfected cells were instrumental in delineating the specific interactions of modified oleamide, an endogenous sleep-associated fatty acid primary amide, with 5-HT_1A and 5-HT_2A receptors (^{Boger et al., 1998}). Transfected cells were also used to disprove an interaction between 5-HT-moduline (Leu-Ser-Ala-Leu) and the 5-HT_1A receptor and to confirm its interaction specifically with a very high apparent affinity and in a non-competitive manner with 5-HT_1B receptors (^{Rousselle et al., 1998}).

Coupling efficiency

Transfected cell systems have been used to compare the efficiency of coupling of the 5-HT_1A receptor to various second messengers and G proteins. For example, the 5-HT_1A receptor expressed in HeLa cells both inhibits adenylyl cyclase and activates PLC (^{Fargin et al., 1989}). However, the receptor has been shown to differentially modulate those activities. The 5-HT_1A receptor couples very efficiently to the inhibition of adenylyl cyclase because changing receptor expression levels from <20 fmol mg⁻¹ protein to ∼3 pmol mg−1 of protein had no effect on the potency or efficacy of agonists to inhibit cyclic AMP accumulation (^{Boddeke et al., 1992}; ^{Fenrick et al., 1996}). This tight coupling applied over a variety of full and partial agonists (Schoeffter et al., 1996). In contrast, coupling of the 5-HT_1A receptor to PLC as measured by Ca²⁺ mobilization and phosphoinositide hydrolysis was much less efficient in that the potency, and especially the efficacy, of various agonists was considerably less than for the inhibition of adenylyl cyclase. Moreover, both efficacy and potency of coupling of the 5-HT_1A receptor to PLC were improved at higher levels of receptor expression (^{Boddeke et al., 1992}; ^{Fargin et al., 1989}; ^{Fenrick et al., 1996}; ^{Raymond et al., 1989}; ^{Schoeffter et al., 1997}).

Transfected cells have been used to study other aspects of 5-HT_1A receptor coupling efficiency, including variables that affect ligand efficacy and potency. One such variable is receptor/G protein stoichiometry. ^{Newman-Tancredi et al. (1997)} studied the effects of altering receptor : G protein ratios in CHO cells by examining [³⁵S]-GTPγS binding in transfected lines bearing various numbers of 5-HT_1A receptors. Their data suggested that increasing receptor/G-protein ratio (i) augments the potency of full agonists, (ii) increases the efficacy of partial agonists and (iii) increases the negative efficacy of inverse agonists at recombinant human 5-HT_1A receptors. Those conclusions were similar to those of ^{Boddeke et al. (1992)}, who showed that increasing receptor numbers were associated with higher efficacy of partial agonists to stimulate PLC in HeLa cells. The levels of G proteins were constant among the cell lines in those two studies, whereas receptor levels varied. ^{Raymond et al. (1992)} altered the ratios of 5-HT_1A receptors to G proteins by treating CHO cells with varying amounts of pertussis toxin to neutralize fractions of endogenous G proteins. They showed that increasing amounts of pertussis toxin resulted first in a rightward shift of the IC₅₀ of 5-HT for the inhibition of adenylyl cyclase (reduction in potency) followed by a reduction in the efficacy. Those results are consistent with either the classical or operational models of receptor/effector coupling (^{Black et al., 1985}; Kenakin & Morgan, 1989), both of which models assume unconstrained interactions of the various signalling molecules. In contrast, ^{Varrault et al. (1992b)} used transfected NIH-3T3 cells to show that increasing 5-HT_1A receptor/G protein stoichiometry mainly increased the efficacies, but not the potencies of various full and partial agonists to inhibit adenylyl cyclase. In a few cases, alterations in efficacy were seen, but these were less than predicted, and varied depending upon the specific ligands being studied (^{Varrault & Bockaert, 1992}). These latter results are consistent with a constrained model of receptor/G protein interaction in which the ability to interact efficiently is damped by functional membrane compartmentalization. The findings are significant in aggregate as they suggest that cell-specific constraints may alter the effects of 5-HT_1A receptor/G protein coupling on ligand efficacy in various second messenger pathways.

Another variable that could affect ligand efficacy and potency at the 5-HT_1A receptor is the possibility that certain ligands induce conformational changes in the receptor that are more or less favourable for G protein activation. For example, partial agonists could be less efficacious than full agonists because they are unable to induce the optimal conformational change in the receptor that regulates contact with G proteins. ^{Gettys et al. (1994)} confirmed that full agonists of the 5-HT_1A receptor were more efficacious in activating G_iα subunits in CHO cells than were partial agonists. They further tested the hypothesis that partial agonists might induce relatively selective activation of specific G_iα subunits when compared with full agonists. They provided evidence in CHO cells that G_iα3 was activated with a rank order of efficacy of 8-OH-DPAT∼5-HT>ipsapirone∼rauwolscine, whereas G_iα2 was activated with a rank order of efficacy of 5-HT>8-OH-DPAT>rauwolscine>ipsapirone. Thus, both receptor/G protein stoichiometry and the availability of specific isotypes of G_iα subunits might affect the coupling efficiency of the 5-HT_1A receptor to G proteins and specific second messengers.

Transcriptional regulation of the 5-HT_1A receptor

Variables that regulate the absolute levels and the stability of receptor mRNA probably lead to critical regulation of receptor protein expression and function. For the 5-HT_1A receptor, this point was experimentally tested by ^{Konigs et al. (1995)}, who showed that levels of receptor mRNA (as measured by Northern blot and in situ hybridization) correlated very well with receptor protein expression (as measured by radioligand binding) in clones of Swiss 3T3 cells transfected with the human 5-HT_1A receptor. The levels of expression of the 5-HT_1A receptor can be altered in vivo by several distinct stimuli. For example, glucocorticoids have been shown to negatively regulate 5-HT_1A receptor mRNA expression within the hippocampus (^{Chalmers et al., 1994}). The selective 5-HT_1A receptor agonist 8-OH-DPAT differentially regulates the levels of 5-HT_1A receptor mRNA in the dorsal and median raphe nuclei (^{Razani et al., 1997}). Hyperammonaemia increased 5-HT_1A receptor mRNA and protein expression in rat hippocampus and transfected HN2 cells (^{Alexander et al., 1995}). Stress induced by serum deprivation increases 5-HT_1A receptor protein and message levels in transfected HN2 and NCB-20 cells (^{Singh et al., 1996b}). The recombinant 5-HT_1A receptor also has been shown to modulate the levels of gangliosides when expressed in HN2 hippocampal, dorsal root ganglion-derived F-11 cells and NCB-20 brain cells (^{Singh et al., 1996a}). This may represent a cellular defense response in that 5-HT_1A receptors protected primary cultures of neurons from chick embryo telencephalons from apoptotic cell death, possibly by stimulating production of NGF (^{Ahlmeyer & Kriegelstein, 1997}). Despite the potential importance of the regulation of 5-HT_1A receptor message and protein levels, almost nothing is known about the molecular mechanisms through which those various stimuli regulate the expression of the 5-HT_1A receptor. In that regard, ^{Charest et al. (1993)} have identified a septal cell line that can be induced to express 5-HT_1A receptors by retinoic acid treatment. Other non-neuronal cell lines (Jurkat cells) have also been suggested to express the 5-HT_1A receptor (^{Aune et al., 1993}; ¹⁹⁹⁴). These cell lines might prove to be useful for analysis of promoter elements of the 5-HT_1A receptor gene that regulate its expression in the central nervous system.

Conclusions

The isolation of the gene and cDNA for the 5-HT_1A receptor has resulted in an explosion of information regarding the characteristics and functions of this receptor in heterologous expression systems. This information has translated important into insights in some cases. In other cases, it has mainly given us a glimpse of what may be possible in physiological settings. The challenge of the next few years will be to separate what is only theoretically possible from what actually happens in cells in which the receptor is naturally expressed.

Acknowledgments

The authors were supported by grants from the Department of Veterans Affairs (Merit Awards to M.N. Garnovskaya and J.R. Raymond) the National Institutes of Health (DK52448 and DK53981 to J.R. Raymond and T.W. Gettys), the U.S. Department of Agriculture (NRICGP/USDA 9800699 to T.W. Gettys), a laboratory endowment jointly supported by the M.U.S.C. Division of Nephrology and Dialysis Clinics, Incorporated (J.R. Raymond), an American Heart Association Fellowship Award (Y.V. Mukhim), and a M.U.S.C. University Research Foundation award (M.N. Garnovskaya).