5-HT Receptors

5-hydroxytryptamine (5-HT, serotonin) is a monoamine that serves as a neurotransmitter, mitogen and hormone. 5-HT influences cells in the brain, nervous system and peripheral tissues by binding to, and activating, a diverse array of cell surface receptors. 5-HT receptors have been divided into seven families based on their molecular structures, pharmacology, and signal transduction linkages. One of the families of 5-HT receptors (5-HT₃) is comprised of ligand-gated ion channels (termed 5-HT_3A–5-HT_3E), whereas six of the families (5-HT₁, 5-HT₂, 5-HT₄, 5-HT₅, 5-HT₆, 5-HT₇) are comprised of heptahelical receptors that signal primarily through coupling to heterotrimeric guanine nucleotide binding and regulatory proteins (G proteins). The G protein-coupled 5-HT receptor families have been characterized by pharmacological properties, amino acid sequences, gene organization, and second messenger coupling pathways. These receptors are integral membrane proteins with seven putative hydrophobic transmembrane domains connected by three intracellular loops (termed iL1–iL3) and three extracellular loops (termed eL1–eL3). The amino terminus of each receptor is oriented toward the extracellular space, whereas the carboxyl terminus (CT) is oriented toward the cytoplasm. These receptors possess conserved or common sites for posttranslational modifications. The extracellular domains are typically glycosylated and possess cysteine residues that may participate in disulfide bonds, which provide structural constraints on the conformation of the receptors. The intracellular domains variably contain phosphorylation sites for a wide array of kinases, palmitoylation sites in the CT for lipid anchoring, and various potential sites for protein–protein interactions.

Recent work has demonstrated an unanticipated diversity of signals linked to the various G protein-coupled receptors (GPCRs), including 5-HT receptors. The work also has led to a growing awareness that GPCR signaling diversity can be modulated by diverse regulatory proteins, many of which are likely to bind directly to the receptor. Those proteins that bind directly to specific motifs on GPCRs are termed receptor-interacting proteins (RIPs). The topic of this chapter focuses on the interaction of Ca²⁺–calmodulin (CaM), which is potential 5-HT receptor RIP, with G protein-coupled 5-HT receptors.

G Protein-Coupled Receptor-Interacting Proteins

A growing body of work suggests that GPCRs can bind to a variety of proteins. This should not be entirely surprising in that, very early on, it was recognized that GPCRs could bind to G proteins. Subsequently, molecular methods allowed the identification of other proteins (arrestins, protein kinase A, protein kinase C (PKC), and G protein-coupled receptor kinases) that bind to GPCRs and modify their functions and interactions with G proteins (48,63,64). Within the last several years, novel protein–protein interactions with GPCRs have been reported, including physical interactions with transmembrane proteins such as other GPCRs (47,76), and RAMPs (receptor associated modifying proteins) (17,19,58).

The most striking and well-characterized examples of GPCR RIPs are the RAMPs. The human calcitonin receptor like receptor (hCRLR) has drastically altered the ligand-binding properties depending upon the RAMP with which it is expressed. It becomes a functional calcitonin gene-related peptide receptor when cotransported with human RAMP1 to the cell surface, whereas it becomes a functional adrenomedullin receptor when cotransported with RAMP2 (53). Another excellent example is provided by the angiotensin II AT_1A receptor, which has been shown to physically interact with ARAP1 (AT₁ receptor associated protein 1), thereby promoting recycling of the receptor to the plasma membrane, enhancing its signaling (37). The importance of this relationship was highlighted by the recent finding that transgenic mice overexpressing ARAP1 develop hypertension and renal hypertrophy (36). Although the interaction of RAMPS with the hCRLR, and ARAP1 with the AT_1A receptor, are quite remarkable, other GPCRs have more recently been demonstrated in a preliminary manner to interact with nonmembrane-spanning proteins such as 14-3-3 proteins (18,26,62,65,90), CaM (16,94), PDZ (PSD-95 discs-large ZO-1) proteins (8,75), and cytoskeletal elements (75,83,86). Most of these reports are preliminary, and the functional consequences of the interactions of GPCRs with these putative RIPs have not been completely elucidated. Indeed, unequivocal verification that these proteins actually interact with GPCRs in intact cells has generally not been completed.

For example, cytoskeletal elements (microtubules and spinophilin) have been shown to physically bind to peptides derived from the α₂-adrenergic receptor (75,82,86). The dopamine D₂ receptor i3L binds spinophilin (but not to its related protein, neurabin) in yeast two-hybrid and fusion protein interaction assays, although the functional significance of this interaction remains undefined. Spinophilin is expressed ubiquitously and contains multiple protein interaction domains (81). The α₂-adrenergic receptor i3L was shown to interact with amino acids 169–255 of spinophilin, in a region between its F-actin-binding and phosphatase 1 regulatory domains. Because the interaction between the α₂-adrenergic receptor and spinophilin was increased by agonist treatment, it would seem to serve a dynamic function (24). The same group showed that the disruption of microtubules with colchicine or nocodazole could alter the apical steady state distribution and delivery of A₁-adenosine receptors and increase the binding capacity of α₂B-adrenergic receptors expressed in MDCK-II cells (31). The authors suggested that the cytoskeleton and G proteins interact with distinct domains of the i3L of the α₂B-adrenergic receptor and that the cytoskeletal effects may be mediated, independent of G proteins. These studies show that abundantly expressed cellular proteins, such as components of the actin cytoskeleton (or CaM), might play important roles in GPCR function, possibly independent of G proteins.

At least five well-characterized examples of GPCR RIPs involve 5-HT receptors. These include the interactions of caveolin-1 and PSD-95 with the 5-HT_2A receptor, MUPP1 (multi-PD2 domain protein 1) with the 5-HT_2C receptor, and EBP50 (ezrin/radixin/moesin-binding phosphoprotein 50) and SNX27 (new sorting nexin) with the 5-HT_4a receptor. MUPP1 interacts with a PDZ domain of the 5-HT_2B receptor. PDZ recognition motifs are contained in many signaling proteins, and these domains mediate key protein–protein interactions. For GPCRs, the best characterized example of a functionally significant GPCR interaction with a PDZ protein is the β₂-adrenergic receptor, which interacts with NHERF (Na⁺/H⁺ exchanger regulatory factor)/EBP50 (32) via agonist-dependent binding of the first PDZ domain of NHERF to the CT of the β₂-adrenergic receptor, resulting in regulation of the type 3 sodium-proton exchanger, NHE-3 (33). PDZ interactions of the β₂-adrenergic receptor with NHERF have also been shown to control recycling of internalized receptors through an endocytic sorting pathway that leads to lysosomal degradation (34).

In yeast two-hybrid screens, 5-HT₂ receptors interact with a multivalent PDZ protein called MUPP1. Coimmunoprecipitations and mutagenesis studies have shown that an SXV sequence at the extreme CT of the 5-HT_2C receptor selectively interacts with the 10th PDZ domain of MUPP1 (23). This interaction is functionally important in that it induces a conformational change in MUPP1 (23), attenuates receptor phosphorylation on serine 453 (S⁴⁵³), and attenuates desensitization (35).

Bhatnagar and colleagues surprisingly demonstrated that caveolin-1 binds to 5-HT_2A receptors in rat synaptic membranes, C6 glioma cells, and transfected HEK-293 cells (14). They demonstrated that this interaction is significant in that caveolin-1 (but not caveloin-2) facilitated increases in intracellular Ca²⁺ through increased coupling of the receptor with G_qα. This specificity of the interaction between caveolin-1 and 5-HT_2A receptor was further supported by experiments showing that caveolin-1 knockdown greatly diminished both 5-HT_2A and P_2Y purinergic receptor signaling without altering PAR-1 (protease activated receptor-1 thrombin receptor) signaling. The same group also demonstrated that interaction of PSD-95 with the 5-HT_2A receptor increases coupling of the receptor to G_qα/11 and blunts receptor internalization without altering its desensitization (9,95,96).

5-HT₂ receptors are not the only 5-HT receptors that have been shown to interact functionally with RIPs. The subcellular localization of the 5-HT_4a receptor is regulated by its binding to two RIPs. EBP50 also called NHERF) and SNX27a (a new sorting nexin, also called Mrt1a) have been shown to interact directly with the 5-HT_4a receptor and to modify its subcellular localization. SNX27a redirects the 5-HT_4a receptor into early endosomes, whereas EBP50/NHERF recruits the receptor into ezrin-containing microvilli (41). In the remainder of this chapter, we will describe findings that support the idea that CaM is an important 5-HT receptor RIP.

Calmodulin

CaM is a member of the superfamily of EF-hand proteins. CaM has four EF-hand motifs, each of which is composed of two α-helices connected by a 12-amino acid loop. When intracellular Ca²⁺ levels rise to the low micromolar range, all four EF-hands bind Ca²⁺, inducing a conformational change that results in binding to various target proteins (4,5,27,97). In CaM’s non Ca²⁺-bound state, the α-helices of the EF hands are aligned in a nearly parallel fashion, which is termed the closed conformation, whereas in its Ca²⁺-bound state, the EF hands are oriented in a nearly perpendicular fashion termed the open conformation. We liken this to two outstretched arms with hands ready to grasp a target. CaM is shaped like a barbell, and when it binds to a target α-helical peptide, it folds over the peptide (27,97), assuming a globular configuration. The crystal structure of CaM and other biochemical evidence have allowed for the construction of algorithms to predict potential CaM-binding proteins (97). Those algorithms identify α-helical peptides that have groupings of hydrophobic amino acids interspersed with positively charged amino acids. In helical wheel diagrams, the positively charged amino acids are often located on the opposite “side” of the peptide from the hydrophobic amino acids. Interestingly, many GPCRs (including 5-HT receptors) possess similar motifs that could serve as CaM-binding domains.

CaM is classically activated by increases in intracellular Ca²⁺, resulting in conformational changes in CaM and activation of target proteins. However, other mechanisms of regulating CaM are possible, although poorly understood. Primary among the alternate mechanisms of activating CaM is phosphorylation of CaM on serine–threonine or tyrosine residues. CaM can be phosphorylated by receptor and nonreceptor tyrosine kinases and serine–threonine kinases (25,30,38,68,77,80). Casein kinase II phosphorylates CaM in vitro on threonine and serine residues (T⁷⁹, S⁸¹, S¹⁰¹, and T¹¹⁷) (78), whereas Ca²⁺-calmodulin-dependent protein kinase IV phosphorylates CaM primarily on T⁴⁴ (40). In contrast, the EGF (epidermal growth factor) receptor phosphorylates Y⁹⁹ of bovine brain CaM (10,28) with a stoichiometry of 1:1 (11), and the insulin receptor phosphorylates Y⁹⁹ and Y¹³⁸ of CaM in CHO-IR (Chinese hamster ovary-insulin receptor) cells (42,79). Nonreceptor tyrosine kinases Src (30) and Jak2 also phosphorylate CaM in response to a number of stimuli (31,32,49,59), but the identity of the tyrosine residues that are phosphorylated by these kinases have not been established. In any case, because nearly all of the G protein-coupled 5-HT receptors either increase intracellular Ca²⁺ or activate tyrosine kinases, any number of 5-HT receptors could reasonably be expected to activate CaM.

5-HT_1A Receptor

Because the 5-HT_1A receptor has been linked to CaM in a few preliminary reports (21,29,57,84,85,99), it is a logical candidate to study the possible interaction between the receptor and CaM. The 5-HT_1A receptor i3L contains a number of sequences that could potentially interact with RIPs. In that regard, we identified two putative CaM-binding domains in the proximal and distal juxtamembrane regions of the i3L of the 5-HT_1A receptor, using a computer search algorithm (http://calcium.uhnres.utoronto.ca/ctdb/pub_pages/general/index.htm), for which scoring is based on evaluation criteria including hydropathy, α-helical propensity, residue charge, helical class, residue weight, and hydrophobic residue content (98). As a general rule, CaM-binding regions are characterized by the presence of several hydrophobic residues interspersed with several positively charged residues, often forming amphipathic α-helices. CaM-binding regions described to date have been divided into several motifs based on the distance between key hydrophobic residues. The putative proximal i3L CaM-binding region of the 5-HT_1A receptor (Y²¹⁵GRIFRAARFRIRKTVKKVEKTG²³⁷) was identified as a 1–12 motif, with key hydrophobic residues at positions 1 and 12. The more proximal sequence is also a putative G protein contact site. The more distal sequence (A³³⁰KRKMALARERKTVKTLGIIMG³⁵²) conforms to a standard CaM-recognition motif of hydrophobic residues at positions 1, 8, and 14 interspersed with positively charged amino acids. This distal peptide is in an intriguing location, as it overlaps with a G protein-contact site and a PKC phosphorylation site in the 5-HT_1A receptor (45,46,50,51,72). The predicted locations of the two CaM-binding regions of the 5-HT_1A receptor (as well as other 5-HT receptors) are depicted in Figure 4.1.

FIGURE 4.1

Schematic illustration of putative CaM-binding domains of various G protein-coupled 5-HT receptors. The upper face of each receptor represents the extracellular domains, and the lower face represents the intracellular domains. The grey bars represent (more...)

In order to illustrate the amphipathic, α-helical nature of the two putative CaM-binding regions from the 5-HT_1A receptor, we modeled them with an α-helical wheel algorithm (http://marqusee9.berkeley.edu/kael/helical.htm), which showed that each peptide is predicted to form an α-helix with hydrophobic amino acids (λ) and charged amino acids (+) located on opposite sides of each helix (Figure 4.2), typical of the amphipathic nature of CaM-binding sites.

FIGURE 4.2

Helical wheel projections of putative CaM-binding domains of the 5-HT_1A receptor. Projections were made using a computer algorithm (http://marqusee9.berkeley.edu/kael/helical.htm). Each circle represents an amino acid residue. Grey circles indicate hydrophobic (more...)

We used a number of methods (coimmunoprecipitation, gel-shift analysis, surface plasmon resonance spectroscopy [SPRS], slot blot, dansyl chloride spectroscopy, and bioluminescence resonance energy transfer [BRET]) to demonstrate that the 5-HT_1A receptor interacts with CaM in vitro and in live cells and to calculate the affinities of the two sites for CaM (91). The apparent affinity of the proximal i3L putative CaM site was 87 ± 23 nM, and the apparent affinity of the distal i3L site was 1.70 ± 0.16 μM. Regarding the functional effects of the CaM sites, we have discovered that CaM-binding and phosphorylation of the 5-HT_1A receptor i3L peptides by (PKC) or protein kinase C are mutually antagonistic processes in vitro, suggesting a possible role for CaM in the regulation of 5-HT_1A receptor phosphorylation and desensitization (91). Furthermore, we have found that binding of CaM to the 5-HT_1A receptor decreases coupling to G_i/o G proteins as assayed by GTPgS binding to crude membrane preparations (Turner et al., unpublished data).

These data support the idea that the 5-HT_1A receptor contains high- and moderate-affinity CaM-binding regions that regulate G protein coupling, receptor phosphorylation, and (perhaps) desensitization.

The idea that CaM interferes with receptor phosphorylation is not limited to 5-HT receptors. Indeed, G protein-coupled receptor kinase (GRK)-mediated phosphorylation of rhodposin is inhibited by CaM (66). Additionally, phosphorylation of a conserved serine residue in the CT of group III metabotropic glutamate receptors by PKC inhibits CaMbinding (1). Thus, there seems to be a mutually antagonistic relationship between phosphorylation and CaM binding for GPCRs and for other signaling proteins. For example, phosphorylation of GRK2 (G protein-coupled receptor kinase 2) by PKC abolishes the ability of CaM to inhibit GRK2 (44), and phosphorylation of the MARCKS proteins (myristoylated alanine rich C-kinase substrate) inhibits CaM binding (35,52). However, this effect is not universal in that phosphorylation of a calcineurin peptide by CaMK-II does not significantly alter the binding of CaM when compared to the nonphosphorylated peptide (22).

Why would phosphorylation and CaM binding be mutually antagonistic? Although the structural basis for this antagonism is not known, it should be intuitive that phosphorylation, by adding a negative charge to a target protein, could disrupt the CaM recognition motif that relies on positive charges. Phosphorylation might also induce conformational changes in the target protein, perhaps by disrupting electrostatic tethering of the positively charged face of the peptide to acidic (negatively charged) phospholipids that are resident on the inner leaflet of the plasma membrane. The inner leaflet of mammalian plasma membranes typically contains between 15–30% of acidic lipids, primarily phosphatidylserine. There is increasing awareness that peptides can interact with plasma membrane lipids through positive charge domains and, especially, with negatively charged lipids in the inner leaflet of the plasma membrane (87). Indeed, electrostatic interactions between positive charges in the transmembrane domains of the S4 family of voltage-sensitive ion channels and negative charges on the inner plasma membrane induce a conformational change in the channels that results in pore opening and closing (56). The electrostatic interaction of positive charge domains with negatively charged lipids would be expected to occur selectively on the plasma membrane, rather than to other internal membranes, in that an inner leaflet of the plasma membrane has far more electrostatic charge than other cellular membranes (24,61).

Conversely, CaM could attenuate phosphorylation by competing with or blocking the access of various kinases to phosphorylation sites within or near to the CaM-binding domains. Alternatively, CaM could mask the positive charges in its target-binding domains, thereby disrupting the electrostatic interaction of the CaM-binding domains with the plasma membrane, resulting in conformational changes that are less favorable to phosphorylation.

Other 5-HT₁ Receptors

Virtually nothing is known regarding the interaction of CaM with other 5-HT₁ receptors, although a computer algorithm predicts that each 5-HT₁ receptor might have several CaM sites (http://calcium.uhnres.utoronto.ca/ctdb/pub_pages/general/index.htm) (98). The 5-HT_1B receptor has three potential CaM-binding domains, including a moderate- to high-likelihood site in the proximal juxtamembrane region of the i3L (L²²⁷YGRIYVEARSRILK²⁴¹), an extended high-likelihood site in the distal juxtamembrane region of the i3L (N²⁸⁸QVKVRVSDALLEKKKLMAARERKATKTLGIILG³²¹), and a moderate-likelihood site in the i1L (N⁶⁷AFVIATVYRTRK⁷⁹) (Figure 4.2).

The 5-HT_1D receptor has three potential CaM-binding domains, including a moderate-likelihood site in the proximal juxtamembrane region of the i3L (L²¹²LIILYGRIYRAARNRI²²⁸), an extended high-likelihood site in the distal juxtamembrane region of the I3L (N²⁷⁵HVKIKLADSALERKRISAARERKATKILGIIL³⁰⁷), and a moderate-likelihood site in the i2L K¹⁴⁸RRTAGHAATMIAIV¹⁶²).

The 5-HT_1E receptor has three potential CaM-binding domains, including a moderate- to high-likelihood site in the i2L (a¹²³itnaieyarkrtakraalmiltv¹⁴⁶), and two moderate-likelihood sites in the proximal and distal juxtamembrane regions of the i3L (y²⁰³riyhaakslyqkr²¹⁶ and s²⁸²strerkaarilglilg²⁹⁸).

The 5-HT_1F receptor has two potential CaM-binding domains, including a high-likelihood site in the proximal and distal juxtamembrane regions of i3L (l¹⁹⁸ilyykiyraaktlyhkrqas²¹⁸ and i²⁸³sgtrerkaattlglilg³⁰⁰). Figure 9.2 clearly shows that all of the putative 5-HT-receptor CaM-binding domains reside in juxtamembrane regions that are also thought to be critical for G protein coupling, supporting the idea that CaM-binding to 5-HT receptors either shields the receptor from G proteins and/or induces conformational changes that could alter G protein coupling.

5-HT₂ Receptors

5-HT_2A receptors in various neuronal and peripheral cells mediate numerous effects through the intermediate actions of CaM (12,23,31,33,43,69,71,88) or CaM-dependent enzymes such as CaMK-II (3,13,55) and myosin light chain kinase (70). We therefore assessed the sequence of the human 5-HT_2A receptor for possible CaM-binding sites. As a receptor that couples to G_q/11-type G proteins, the 5-HT2A receptor is capable of stimulating phosphoinositide turnover and subsequent increases in intracellular Ca²⁺ (73). Accordingly, we postulated that the 5-HT_2A receptor might exert some of its Ca²⁺-sensitive intracellular effects by directly interacting with CaM. A search of the amino acid sequence using a computer algorithm (http://calcium.uhnres.utoronto.ca/ctdb/pub_pages/general/index.htm) (98) revealed two novel potential CaM-binding motifs, located in the i2L (S¹⁸⁴RFNSRTKAFLKIIAVWTI²⁰²) and the juxtamembrane region of the CT (P³⁶⁷LVY TLFNKTYRSAFSRYIQ³⁹⁶) of the 5-HT_2A receptor. Both sequences contain consensus phosphorylation sites and are potentially important for G protein coupling, suggesting that interaction of CaM with those sites could play roles in regulating receptor function. The CaM-binding region in the i2L of the 5-HT_2A receptor conforms to a 1-8-14 motif, which is characterized by hydrophobic residues at positions 1, 8, and 14. The putative CaM-binding domain in the CT of the 5-HT_2A receptor conforms to a 1-10 motif, with critical hydrophobic residues separated by 8 amino acids.

In order to illustrate the amphipathic nature of the putative CaM-binding sequences of the 5-HT_2A receptor, we modeled them with an α-helical wheel algorithm (http://marqusee9.berkeley.edu/kael/helical.htm). Figure 4.3 shows that both putative CaM-binding domains of the 5-HT_2A receptor contain clusters of positively charged amino acids on one side of the α-helix, with mostly hydrophobic amino acids concentrated on the opposite side.

FIGURE 4.3

Helical wheel projections of putative CaM-binding domains of the 5-HT_2A receptor. Projections were made using a computer algorithm (http://marqusee9.berkeley.edu/kael/helical.htm). Each circle represents an amino acid residue. Grey circles indicate hydrophobic (more...)

We used a variety of techniques (coimmunoprecipitation, gel-shift analysis, slot blot, dansyl chloride spectroscopy, and BRET) to demonstrate that the 5-HT_2A receptor interacts with CaM in vitro and in live cells, and to calculate the affinities of each of the two sites for CaM. Peptides from each of the sites bound to CaM in a Ca²⁺-dependent fashion, with the i2L peptide binding with an apparent higher affinity than that of the CT peptide, based on mobility shifting of CaM in a nondenaturing gel shift assay. We used fluorescence emission spectral analyses of dansyl-CaM to calculate the apparent K_D values of 65 ± 30 nM for the i2L peptide and 168 ± 38 nM for the CT peptide (92).

Because the putative CT CaM-binding domain overlaps with a putative PKC site, we tested whether CaM and PKC exerted mutually antagonistic effects on this peptide sequence. We demonstrated that the CT peptide was readily phosphorylated by PKC in vitro and that CaM-binding and phosphorylation of this peptide were antagonistic. These results suggest that there is a potential role for CaM in the regulation of 5-HT_2A receptor phosphorylation and desensitization (similar to what was shown for the 5-HT_1A receptor in the section titled “5-HT_1a Receptor”). We also tested whether CaM had an effect on G protein coupling of the 5-HT_2A receptor by measuring [³⁵S]GTPγS binding in the presence and absence of 5-HT. Those experiments showed that CaM decreases 5-HT_2A receptor-mediated [³⁵S]GTPγS binding to NIH-3T3 cell membranes, supporting a possible role for CaM in modulating receptor-G protein coupling. The i2L peptide (but not the CT peptide) was able to stimulate [³⁵S]GTPγS binding, and this effect was blocked by CaM, suggesting that the i2L is critical for G protein activation. In contrast, the CT peptide (but not the i2L peptide) was a strong substrate for PKC-induced phosphorylation. These results strongly support two regulatory roles for CaM in 5-HT_2A receptor function. First, CaM appears to blunt receptor coupling to G proteins through interaction with the i2L. Second, CaM blunts phosphorylation (and possibly desensitization) of the CT peptide of the 5-HT2A receptor (92). Thus, the 5-HT_2A receptor contains two high-affinity-CaM-binding domains that appear to play important roles in both the initiation and the termination of receptor signaling.

The evidence for important roles of CaM in regulating other 5-HT₂ receptors is less clear than for the 5-HT_2A receptors. 5-HT_2B receptors potentiate NMDA-induced depolarizations in frog spinal cord motor neurons through CaM, but not through elevations of intracellular Ca²⁺ or activation of CaMK-II (39). The human 5-HT_2B receptor has one high-likelihood CaM-binding domain, (L³⁸²FNKTFRDAFGRYITCNYRATKSVKTLR⁴⁰⁹), which is located at the juxtamembrane region of the CT of the receptor.

There are two reports showing that calcineurin, which is a CaM-dependent phosphatase, inhibits the desensitization of the 5-HT_2C receptor (2,15). The 5-HT_2C receptor has a high-likelihood CaM-binding domain (V³⁶⁷YTLFNKIYRRAFSNYLRCNYK³⁸⁸) at the juxtamembrane region of the CT of the receptor, as well as two moderate-likelihood CaM-binding domains in the i2L (F¹⁶⁵NSRTKAIMKIAIVW¹⁷⁹) and distal i3L (A³⁰²INNERKASKVLGIVF³¹⁷). Thus, it is possible that all three of the major subtypes of 5-HT₂ receptors might be regulated through direct interactions with CaM, although that possibility has not been rigorously examined for the 5-HT_2B and 5-HT_2C receptors.

Other 5-HT Receptors (5-HT₄, 5-HT₅, 5-HT₆, 5-HT₇)

We were unable to find references mentioning the coupling of 5-HT₄, 5-HT₅, or 5-HT₆ receptors to CaM. However, each of those receptors has potential CaM-binding domains. The human 5-HT_4A receptor has two putative CaM-binding domains, including a high-likelihood site in iL1 (V⁴⁴CWDRQLRKIKTNYFIVSLA⁶³), and a moderate-likelihood site in the proximal juxtamembrane region of the CT (N³⁰⁸PFLYAFLNKSFRRAFLI³²⁵). The human 5-HT_5a receptor has a moderate-likelihood site in the proximal juxtamembrane region of i3L (N²⁰⁶PFLYAFLNKSFRRAFLI²²⁴) and a high-likelihood site at the distal juxtamembrane region of i3L (D²⁵²SRRLATKHSRKALKASLTLGIL²⁷³). The human 5-HT₆ receptor has two putative CaM-binding sites in the juxtamembrane regions of the i3L, including a potential moderate-likelihood site in the proximal i3L (D²⁵⁹SRRLATKHSRKALKASLTLGIL²⁷⁵) and a high-likelihood site in the distal i3L (H³¹²ERKNISIFK REQKAATTLG IIVGA³³⁵).

In contrast to the human 5-HT₄, 5-HT₅, and 5-HT₆ receptors, 5-HT_7a receptors have been functionally linked to CaM in that they have been shown to activate CaM-sensitive adenylyl cyclases, AC1 and AC8, whereas the 5-HT₆ receptor do not activate them (7). In their study, Baker and colleagues showed that the 5-HT₆ receptor behaved in a typical manner for G_s-coupled receptors in that it stimulated AC5, a G_s-sensitive adenylyl cyclase, but not AC1 or AC8. AC1 and AC8 typically are not activated by G_s-coupled receptors in vivo. In contrast, the 5-HT_7a receptor stimulated AC1 and AC8 by increasing intracellular Ca²⁺ (7).

The 5-HT_7a receptor has three potential CaM-binding domains, two of which are in the juxtamembrane regions of i3L, including a proximal moderate-likelihood site (Y²⁵⁹YQIYKAARKSAAKHKF²⁷⁵) and a distal high-likelihood site (R³¹³KNISIFK REQKAAT-TLG IIVGA³³⁵). The 5-HT_7a receptor also has an unusual CaM-binding motif in the proximal juxtamembrane region of the CT (F³⁸¹IYAFFNRDLRTTYRS³⁹⁷). This site contains a sequence that is termed an ilimaquinone or IQ motif, which is a consensus site for Ca ²⁺-independent CaM binding (74). The IQ motif is widely distributed in a variety of CaM-binding proteins, often in association with classical Ca²⁺-dependent binding motifs. IQ motif-containing proteins possess a dazzling array of biological functions including signal transduction, phosphorylation, cytoskeletal regulation, trafficking, cell cycle control, and polarized growth (74). The presence of the IQ motif in the 5-HT_7a receptor is intriguing in that IQ motifs play critical roles in neuronal growth and plasticity, and in assembling neuronal signal transduction complexes (6).

The presence of the IQ motif in the human 5-HT_7a receptor is also intriguing at the molecular level. One group has shown that the presence of the IQ motif can influence the binding of CaM to classical Ca²⁺-dependent CaM-binding motifs. They showed that the IQ motif in an abundant neuronal protein, PEP-19, accelerates the slow association and dissociation of Ca²⁺ from the C-domain of free CaM 40–50-fold, and also increases the rate of dissociation of Ca²⁺ from CaM when CaM is bound to CaMK-II. Thus, the IQ motif could regulate the dynamics of Ca²⁺-binding to both free CaM and CaM bound to target proteins. The authors of this study pointed out that this relationship is critical in that the binding of Ca²⁺ to the C-domain of CaM is a rate-limiting step for CaM-dependent enzyme activation (67). Scenarios in which the 5-HT_7a receptor IQ motif enhances the dynamic interaction of CaM with the two putative juxtamembrane CaM-binding domains in the i3L of the 5-HT_7a receptor, or with CaM target enzymes colocalized within a signaling complex with the 5-HT_7a receptor, could be easily envisioned. Indeed, we speculate that the IQ motif could facilitate CaM-binding to GPCRs that heterodimerize with the 5-HT_7a receptor. Those possibilities would be interesting to pursue through further experimentation.