Structure, dynamics and lipid interactions of serotonin receptors: excitements and challenges

Abstract

Serotonin (5-hydroxytryptamine, 5-HT) is an intrinsically fluorescent neurotransmitter found in organisms spanning a wide evolutionary range. Serotonin exerts its diverse actions by binding to distinct cell membrane receptors which are classified into many groups. Serotonin receptors are involved in regulating a diverse array of physiological signaling pathways and belong to the family of either G protein-coupled receptors (GPCRs) or ligand-gated ion channels. Serotonergic signaling appears to play a key role in the generation and modulation of various cognitive and behavioral functions such as sleep, mood, pain, anxiety, depression, aggression, and learning. Serotonin receptors act as drug targets for a number of diseases, particularly neuropsychiatric disorders. The signaling mechanism and efficiency of serotonin receptors depend on their amazing ability to rapidly access multiple conformational states. This conformational plasticity, necessary for the wide variety of functions displayed by serotonin receptors, is regulated by binding to various ligands. In this review, we provide a succinct overview of recent developments in generating and analyzing high-resolution structures of serotonin receptors obtained using crystallography and cryo-electron microscopy. Capturing structures of distinct conformational states is crucial for understanding the mechanism of action of these receptors, which could provide important insight for rational drug design targeting serotonin receptors. We further provide emerging information and insight from studies on interactions of membrane lipids (such as cholesterol) with serotonin receptors. We envision that a judicious combination of analysis of high-resolution structures and receptor-lipid interaction would allow a comprehensive understanding of GPCR structure, function and dynamics, thereby leading to efficient drug discovery.

Serotonin: an ancient neurotransmitter with wide-ranging physiology

Serotonin (5-hydroxytryptamine, 5-HT), an intrinsically fluorescent biogenic amine that acts as a neurotransmitter (see Fig. 1a), was discovered as a component of bovine serum ~ 70 years ago in a rather serendipitous manner by Irvine Page and co-workers at Cleveland Clinic (Rapport et al. 1948a, see Whitaker-Azmitia 1999 for a lucid account of the discovery of serotonin). However, the neurobiological relevance of serotonin was established a few years later when it was shown by Betty Twarog that serotonin is a crucial component of the central nervous system (Twarog and Page 1953; apparently, Irvine Page had initial reservations about finding serotonin in mammalian brain (Whitaker-Azmitia 1999)). Serotonin was crystallized (Rapport et al. 1948b) and the chemical structure was determined as 5-hydroxytryptamine (Rapport 1949) by Maurice Rapport.

Fig. 1.

Classification and molecular architecture of human serotonin (5-HT) receptors. a Chemical structure of serotonin (5-hydroxytryptamine). b Over the course of their long evolutionary history, serotonin receptors have differentiated into several subtypes. With the exception of serotonin₃ (5-HT₃) receptors (which are ligand-gated ion channels), all subtypes of serotonin receptors belong to the GPCR superfamily. The protein sequences of human serotonin receptors (serotonin_1A (P08908), serotonin_1B (P28222), serotonin_1D (P28221), serotonin_1E (P28566), serotonin_1F (P30939), serotonin_2A (P28223), serotonin_2B (P41595), serotonin_2C (P28335), serotonin_3A (P46098), serotonin_3B (O95264), serotonin_3C (Q8WXA8), serotonin_3D (Q70Z44), serotonin_3E (A5X5Y0), serotonin₄ (Q13639), serotonin_5A (P47898), serotonin₆ (P50406), and serotonin₇ (P34969)) were retrieved from the UniProt database (The UniProt Consortium 2019) and the phylogenetic tree was constructed using the MEGA software (Kumar et al. 2018). Distinct families of serotonin (5-HT) receptors are color coded. General topology of serotonin receptors belonging to the (c) GPCR family, and (d) serotonin₃ (5-HT₃) receptors which are ion channels. Serotonin receptors that belong to the GPCR family contain seven transmembrane α-helices, whereas serotonin₃ (5-HT₃) receptors have three distinct domains: a large N-terminal extracellular domain (ECD), a transmembrane domain (TMD) comprising of four α-helices and a large intracellular domain (ICD) between transmembrane helices 3 and 4

Serotonin is found in a wide variety of life forms ranging from primitive single-celled eukaryotes paramecium to humans, and is distributed across a range of tissues, both neural and non-neural (Hen 1992; Nichols and Nichols 2008; Berger et al. 2009). Serotonergic signaling plays a crucial role in the generation and modulation of various cognitive and behavioral functions such as learning, mood, stress, pain, sleep, aggression, depression, anxiety, cognition, and sexual behavior (Sodhi and Sanders-Bush 2004; Olivier 2015). Essentially, serotonin appears to modulate virtually the entire spectrum of human behavioral processes (Berger et al. 2009). This is rather surprising, in view of the fact that majority (~ 95%) of total body serotonin is localized outside the central nervous system (Gershon and Tack 2007) and about one in a million neurons in the central nervous system synthesize serotonin (Berger et al. 2009). Yet, all regions of the brain express multiple serotonin receptors in a tissue-dependent fashion (Nichols and Nichols 2008). As a result, drugs that target serotonin receptors are useful in the treatment of a variety of disorders (Fiorino et al. 2014). This makes drug development involving serotonin receptors an exciting area of research with tremendous opportunity.

Serotonin is intrinsically fluorescent and is a derivative of the naturally occurring fluorescent amino acid tryptophan. Although the intrinsic fluorescence of serotonin was observed a long time back (Udenfriend et al. 1955), a comprehensive understanding of serotonin fluorescence in terms of its excited state molecular properties was reported much later (Chattopadhyay et al. 1996; Maiti et al. 1997; Sarkar et al. 2018).

Serotonin receptors

In evolutionary terms, the serotonin receptors go back to 700–800 million years (Peroutka and Howell 1994). The diverse cellular activity of serotonin in humans is mediated by a total of seventeen receptors that belong to seven distinct subfamilies (5-HT_1–7). Out of these, members of six subfamilies (5-HT_1–2 and 5-HT_4–7) belong to the G protein-coupled receptor (GPCR) family, whereas members of the serotonin₃ (5-HT₃) receptor subfamily are ligand-gated ion channels (see Fig. 1). Serotonin receptors constitute the largest class of neurotransmitter GPCRs and non-odorant GPCRs (Chattopadhyay 2007; Nichols and Nichols 2008).

Serotonin receptors are instrumental in the generation and modulation of cognitive and behavioral functions such as sleep, mood, pain, addiction, sexual activity, depression, anxiety, alcohol abuse, aggression, learning, and memory (Lin et al. 2014; Cowen and Browning 2015; Stiedl et al. 2015; Carhart-Harris and Nutt 2017; Yohn et al. 2017; Underwood et al. 2018). Malfunctioning of serotonin receptors has been implicated in the etiology of mental disorders such as schizophrenia, depression, suicidal behavior, infantile autism, and obsessive compulsive disorder. For this reason, members of the serotonin receptor family represent a major drug target in all clinical areas (Chattopadhyay 2014; Fiorino et al. 2014; McCorvy and Roth 2015). More importantly, the design of novel drugs that would target specific serotonin receptors with improved selectivity and lower side effects represents an exciting and challenging area of drug development. This would require the availability of high-resolution structures of target serotonin receptors, both in apo and ligand-bound states and knowledge of detailed conformational change(s) upon ligand binding (McCorvy and Roth 2015).

Structural biology of serotonin receptors

Structural insights contribute to our understanding of ligand binding, receptor activation and inactivation mechanisms, conformational flexibility, and other crucial aspects of receptor function. Structure determination of a functionally active, full-length GPCR in native form is challenging due to its inherent conformational dynamics (Manglik et al. 2015; Latorraca et al. 2017; Wingler and Lefkowitz 2020) that poses a problem for X-ray crystallography. As a consequence, available crystal structures of GPCRs are engineered with truncations of N- and/or C-termini and flexible intracellular loops, since these regions interfere in the crystallization process (Wang et al. 2013; Yin et al. 2018; Pal and Chattopadhyay 2019). In addition, incorporation of thermostable mutation(s) in the native sequence of GPCRs and treatment with ligands that significantly help to stabilize the receptor (in most cases inverse agonists) were strategically utilized in almost all the available structures (Lebon et al. 2012; Zhang et al. 2015a). Interestingly, in addition to ligands, GPCR structures are often stabilized in active conformations by replacing ICL3 with T4 lysozyme, apocytochrome b₅₆₂ (BRIL) or by addition of G-proteins, β-arrestins, and fragment antibodies/nanobodies with affinities for the active conformation (Ghosh et al. 2015).

Most of the structures of serotonin receptors, including serotonin_1B (Wang et al. 2013; Yin et al. 2018; Miyagi et al. 2020), serotonin_2A (Kimura et al. 2019; Kim et al. 2020), serotonin_2B (Liu et al. 2013; Wacker et al. 2013, 2017a; McCorvy et al. 2018), serotonin_2C (Peng et al. 2018), and serotonin_3A (Hassaine et al. 2014) receptors, have been solved by X-ray crystallography. With the recent resolution revolution using cryo-electron microscopy (cryo-EM) (Safdari et al. 2018), the high-resolution structures of serotonin_1B (García-Nafría et al. 2018), serotonin_2A (Kim et al. 2020), and serotonin_3A (Basak et al. 2018a, b, 2019; Polovinkin et al. 2018; Zarkadas et al. 2020) receptors have been solved. High-field nuclear magnetic resonance (NMR) spectroscopy has emerged as a powerful tool for determination of structure of membrane proteins without any perturbation to the whole protein (Arora and Tamm 2001; Opella 2013). An additional advantage of NMR spectroscopy is that it provides insight into protein dynamics, in addition to structural details. However, no NMR structure of any serotonin receptor is available yet although NMR structure of a chemokine GPCR (CXCR-1) is available without sequence alteration (Park et al. 2012). Compared to X-ray crystallography and NMR spectroscopy, cryo-EM appears to be more advantageous for structural characterization of full-length membrane proteins and protein complexes (such as serotonin receptors coupled with heterotrimeric G-proteins). A unique advantage of cryo-EM is its ability to capture structural details of molecular assemblies (complexes) in physiologically relevant environment (García-Nafría et al. 2018; Basak et al. 2018a, b; Kim et al. 2020).

We discuss below important characteristics of serotonin receptors provided by recently published high-resolution structures, thereby highlighting structural conservation as well as diversity across the receptor subtypes. In addition, we will outline areas that need to be explored further for an in-depth understanding of the molecular mechanisms of activation and desensitization of serotonin receptors and resulting consequence in signaling cascades.

Recent advances in structural biology of serotonin receptors

High-resolution structure determination of membrane proteins involves several challenges that include poor expression of recombinant protein, difficulties in efficient solubilization and low purification yields (Arora and Tamm 2001; Kalipatnapu and Chattopadhyay 2005; Carpenter et al. 2008; Skretas and Georgiou 2008; Chattopadhyay et al. 2015; Wiseman et al. 2020). However, recent improvements in the areas of membrane protein molecular biology and biochemistry along with technical advancements in crystallography, NMR spectroscopy, and cryo-EM have tremendously helped to overcome these challenges toward determination of high-resolution structures of membrane proteins, including serotonin receptors (Lacapère et al. 2007; Bill et al. 2011; Opella 2013; Ishchenko et al. 2017a; Sjöstrand et al. 2017; Smith 2017; Huang et al. 2018; Mandala et al. 2018).

No high-resolution structure of any member of the serotonin receptor family was available until 2013, although by that time, structures of many GPCRs were available. By early 2013, the first structures of serotonin_1B receptor in complex with ergotamine (PDB ID: 4IAR, Fig. 2a) and dihydroergotamine (PDB ID: 4IAQ) (Wang et al. 2013) and serotonin_2B receptor in complex with ergotamine (PDB ID: 4IB4, Fig. 2c) (Wacker et al. 2013) were solved by X-ray crystallography. From 2013 to 2020, 30 structures from five serotonin receptor subtypes were experimentally determined (see Table 1). More than 60% of the structures of serotonin receptors were GPCRs, including serotonin_1B, serotonin_2A, serotonin_2B, and serotonin_2C in complex with a diverse array of ligands. On the other hand, ten structures of the ion channel serotonin_3A receptor are available till date and seven of them were reported in 2018 alone. Most of the available structures of serotonin receptors are truncated variants of the full receptor and therefore lack structural information from the N- and C-termini and loop regions (Fig. 2, marked as dashed lines), often necessary for proper functioning of the receptor (Pal and Chattopadhyay 2019). These truncations in the native serotonin receptor sequence were necessary to overcome the inherent receptor dynamics, thereby enabling crystallization (Chattopadhyay 2014). In addition, most of the crystal structures of serotonin receptors belonging to the GPCR family have one or more thermostabilizing mutations in their native sequence (Wacker et al. 2013, 2017a; Liu et al. 2013; Wang et al. 2013; Ishchenko et al. 2017b; McCorvy et al. 2018; Peng et al. 2018; Yin et al. 2018; Kimura et al. 2019). These modifications reduce dynamics of the receptor and help the crystallization process. However, such mutations could perturb the native conformation of the receptor which is likely to alter physiological function.

Fig. 2.

Structural organization of serotonin receptors. Crystal structures of (a) serotonin_1B (PDB ID: 4IAR) (Wang et al. 2013), (b) serotonin_2A (PDB ID: 6A93) (Kimura et al. 2019), (c) serotonin_2B (PDB ID: 4IB4) (Wacker et al. 2013), (d) serotonin_2C (PDB ID: 6BQH) (Peng et al. 2018), and (e) serotonin_3A (PDB ID: 4PIR) (Hassaine et al. 2014) receptors. The domain organization of serotonin receptors belonging to (f) GPCR family and (g) ligand-gated ion channels. Serotonin receptors belonging to the GPCR family contain a well conserved class A GPCR architecture with an extracellular N-terminus, followed by 7 transmembrane (TM) α-helices (TM1–7) connected by 3 intracellular (ICL1–3) and 3 extracellular loops (ECL1–3), followed by an amphipathic helix 8 (H8) and an intracellular C-terminus. In contrast to other serotonin receptors, serotonin₃ receptors are structurally characterized by a twisted β-sandwich N-terminal extracellular domain (ECD), 4 helical transmembrane domains (M1–4), and a large intracellular domain (ICD). The horizontal cylinders represent α-helices, solid lines denote extramembranous linker regions, and dashed lines depict regions where structural information is missing. Structures of GPCRs were generated from their respective PDB IDs using PyMOL Molecular Graphics System (Schrödinger, LLC). See text for more details

Table 1.

Available structures of various serotonin receptors till date

Receptor	#	Experimental method	PDB ID	Resolution (Å)	Bound ligand/effector	Missing structural region(s)	Reference
Serotonin1B	5	X-ray	4IAQ	2.8	Dihydroergotamine	N, ICL3	Wang et al. 2013
		X-ray	4IAR	2.7	Ergotamine	N, ICL3	Wang et al. 2013
		X-ray	5V54	3.9	Methiothepin	N, ICL3	Yin et al. 2018
		Cryo-EM	6G79	3.78	EP5/Go1β1γ2	N	García-Nafría et al. 2018
		X-ray	7C61	3.0	Ergotamine	N, ICL3	Miyagi et al. 2020
Serotonin2A	5	X-ray	6A93	3.0	Risperidone	N, C, ICL3	Kimura et al. 2019
		X-ray	6A94	2.9	Zotepine	N, C, ICL3	Kimura et al. 2019
		X-ray	6WGT	3.4	LSD	N, C, ICL3	Kim et al. 2020
		X-ray	6WH4	3.4	Methiothepin	N, C, ICL3	Kim et al. 2020
		Cryo-EM	6WHA	3.36	25-CN-NBOH/Gq/Gβ/Gγ	N, C	Kim et al. 2020
Serotonin2B	8	X-ray	4IB4	2.7	Ergotamine	N, C, ICL3	Wacker et al. 2013
		X-ray	4NC3	2.8	Ergotamine	N, C, ICL3	Liu et al. 2013
		X-ray	5TUD	3.0	Ergotamine	N, C, ICL3	Ishchenko et al. 2017b
		X-ray	5TVN	2.9	LSD	N, C, ICL3	Wacker et al. 2017a
		X-ray	6DRX	3.1	H8G	N, C, ICL3	McCorvy et al. 2018
		X-ray	6DRY	2.92	H8D	N, C, ICL3	McCorvy et al. 2018
		X-ray	6DRZ	3.1	H8J	N, C, ICL3	McCorvy et al. 2018
		X-ray	6DS0	3.19	H8M	N, C, ICL3	McCorvy et al. 2018
Serotonin2C	2	X-ray	6BQG	3.0	Ergotamine	N, C, ICL3	Peng et al. 2018
		X-ray	6BQH	2.7	Ritanserin	N, C, ICL3	Peng et al. 2018
Serotonin3A	10	X-ray	4PIR	3.5	–	–	Hassaine et al. 2014
		Cryo-EM	6BE1	4.31	–	–	Basak et al. 2018b
		Cryo-EM	6DG7	3.32	Serotonin	–	Basak et al. 2018a
		Cryo-EM	6DG8	3.89	Serotonin	–	Basak et al. 2018a
		Cryo-EM	6HIN	4.1	Serotonin	–	Polovinkin et al. 2018
		Cryo-EM	6HIO	4.2	Serotonin	–	Polovinkin et al. 2018
		Cryo-EM	6HIQ	3.2	Serotonin	–	Polovinkin et al. 2018
		Cryo-EM	6HIS	4.5	Tropisetron	–	Polovinkin et al. 2018
		Cryo-EM	6NP0	2.92	Granisetron	–	Basak et al. 2019
		Cryo-EM	6Y1Z	2.82	Palonosetron	–	Zarkadas et al. 2020

#, number of available structures; N, N-terminus; C, C-terminus; ICL3, intracellular loop 3; cryo-EM, cryo-electron microscopy; X-ray, X-ray diffraction; EP5, 2-[5-[2-[4-(4-cyanophenyl)piperazin-1-yl]-2-oxidanylidene-ethoxy]-1~{H}-indol-3-yl]ethylazanium; LSD, lysergic acid diethylamide; 25-CN-NBOH, (2S,6S)-2-(2,5-dimethoxy-4-bromobenzyl)-6-(2-methoxyphenyl)piperidine; H8G, N,N-diethyl-N′-[(8α)-6-methyl-9,10-didehydroergolin-8-yl]urea; H8D, (8β)-N-[(2S)-1-hydroxybutan-2-yl]-6-methyl-9,10-didehydroergoline-8-carboxamide; H8J, (8α)-N-[(2S)-1-hydroxybutan-2-yl]-1,6-dimethyl-9,10-didehydroergoline-8-carboxamide; H8M, (1S)-1-[(2-chloro-3,4-dimethoxyphenyl)methyl]-6-methyl-2,3,4,9-tetrahydro-1H-β-carboline

In contrast to the available crystal structures of serotonin receptors, the structure of the intracellular loop 3 (ICL3) of serotonin_1A receptor was independently solved by NMR spectroscopy (Chen et al. 2011), in which part of the ICL3 showed α-helical conformation, similar to the ICL3 structure of another GPCR, the vasopressin V2 receptor (Bellot et al. 2009). Importantly, the third intracellular loop of the serotonin_1A receptor has been shown to be crucial in coupling to and activation of G-proteins (Varrault et al. 1994; Raymond et al. 1999). In addition, it has been previously reported using site-directed mutagenesis that mutations in ICL3 of the serotonin_1A receptor alters G-protein coupling from G_i to G_s in a ligand-dependent manner (Malmberg and Strange 2000). Interestingly, in a study using the intrinsic fluorescence of the sole tryptophan residue of ICL3, it was shown that the tryptophan is localized in a restricted environment, probably due to the constraints induced by the peptide secondary structure (Pal et al. 2018). Despite the presence of a binding partner (G-protein), information on ICL3 electron density was found to be absent in the cryo-EM structure of serotonin_1B coupled to heterotrimeric G_o protein, thereby indicating that such secondary structural elements in ICL3 could be rather transient and/or flexible (García-Nafría et al. 2018). The structural characterization of the ICL3 region of serotonin receptors in the GPCR family is an emerging area and determining its structure could help our understanding of signaling mechanism of serotonin receptors as well as other GPCRs.

General structural features of serotonin receptors

As mentioned above, serotonin receptors are classified into seven subfamilies of receptors (serotonin_1–7, see Fig. 1b). Except for the serotonin₃ receptor which is a ligand-gated ion channel, the remaining six members are GPCRs which mediate canonical signaling through coupling with one of the three G-proteins (Gα_i/o, Gα_q/11 and Gα_s) (Giulietti et al. 2014). Serotonin receptors belonging to the GPCR family contain a well conserved class A GPCR structural design with an extracellular N-terminus, followed by 7 transmembrane (TM) α-helices (TM1–7) connected by 3 intracellular (ICL1–3) and 3 extracellular (ECL1–3) loops, and an amphipathic helix 8 (H8) and an intracellular C-terminus (Fig. 1c and Fig. 2f). The receptors adopt a tertiary structure resembling a barrel, with the ligand binding site within the membrane. All serotonin receptors in the GPCR family share relatively similar TM domains, except for slight differences in helix tilt and twist, which is probably due to the sequence conservation in the TM regions across serotonin receptors belonging to the GPCR family (McCorvy and Roth 2015; Zhang et al. 2015b; Cvicek et al. 2016). The conserved residues are assigned according to the well-established Ballesteros-Weinstein numbering system (Ballesteros and Weinstein 1995; Isberg et al. 2015), which is a sequence-based generic GPCR residue numbering scheme. It involves a numbering scheme based on the most conserved amino acid residue within each helix of the TM region being assigned as X.50, where X is the helix number denoted serially by 1–7 from N-terminus to C-terminus. The residues toward the C and N-termini are numbered in increasing and decreasing order, respectively. Not surprisingly, these conserved residues designated as X.50 have notable functional importance (Fig. 3). For example, N^1.50 in TM1 and D^2.50 in TM2 form the sodium ion binding site (Fenalti et al. 2014; McCorvy and Roth 2015), R^3.50 with D/E^3.49 and Y^3.51 form the conserved D(E)RY motif responsible for receptor activation and inactivation (Wacker et al. 2013), W^4.50 stabilizes the helical packing of TM2, 3 and 4 (McCorvy and Roth 2015), P^5.50 forms the conserved P-I-F motif together with I^3.40 and F^6.44 and regulates receptor activation (Rasmussen et al. 2011; Wacker et al. 2013), P^6.50 probably allows the movement of helices toward the intracellular side (McCorvy and Roth 2015), and P^7.50 together with N^7.49, Y^7.53 and two random amino acids at 7.51 and 7.52 positions form the NPxxY conserved motif, which controls the dynamics of TM7 toward GPCR signaling (Wacker et al. 2013). In addition to the TM domains, serotonin receptors in the GPCR family exhibit conserved structural features in the loop regions. Importantly, a conserved disulfide bridge is formed between C^3.25 and C^ECL2 that connects ECL2 with TM3 and regulates the dynamics of the loop in majority of GPCRs (Unal et al. 2010; Wheatley et al. 2012).

Fig. 3.

A schematic representation of serotonin receptors belonging to the GPCR family showing functionally important residues. The most conserved residues are colored red, whereas, residues with functional significance (specific functionality identified by various studies) are colored black. N^1.50 in TM1 and D^2.50 in TM2 form the sodium ion binding site (Fenalti et al. 2014; McCorvy and Roth 2015), R^3.50 with D/E^3.49 and Y^3.51 form the conserved D(E)RY motif responsible for receptor activation and inactivation (Wacker et al. 2013). The residue W^4.50 stabilizes the helical packing (McCorvy and Roth 2015), P^5.50 forms the conserved P-I-F motif together with I^3.40 and F^6.44 and regulates receptor activation (Rasmussen et al. 2011; Wacker et al. 2013) and P^6.50 likely allows the movement of helices (McCorvy and Roth 2015), and P^7.50 together with N^7.49, Y^7.53 and two random amino acids at 7.51 and 7.52 positions forms the NPxxY conserved motif which controls the dynamics of TM7 toward GPCR signaling (Wacker et al. 2013). In addition to the TM domains, serotonin receptors exhibit conserved structural features in the loop regions. Importantly, a conserved disulfide bridge is formed between C^3.25 and C^ECL2 that connects ECL2 with TM3 and regulates the dynamics of the loop in the vast majority of GPCRs

In contrast to all other serotonin receptors, serotonin₃ receptors are structurally characterized by a twisted β-sandwich N-terminal extracellular domain (ECD), four helical transmembrane domains (TMD), and a large intracellular domain (ICD, Figs. 1d and 2g). The ligand-gated ion channel is either homopentameric composed of five identical serotonin_3A subunits or heteropentameric consisting of a mixture of serotonin_3A and one member from any of the other four (serotonin_3B, serotonin_3C, serotonin_3D, and serotonin_3E) subunits (Niesler et al. 2003; O’Leary and Cryan 2010; Hensler 2012).

Structural diversity across serotonin receptor subtypes

Structural variations are exhibited not only between serotonin receptor subfamilies and subtypes, but also within a subtype member upon binding to specific type of ligands. Ligands are classified in terms of the downstream signaling responses generated upon binding to the receptor. Ligands that activate their cognate GPCRs are termed agonists, whereas antagonists are ligands that reduce agonist-mediated receptor activity to basal level. In contrast, inverse agonists inhibit even the basal activity of a receptor (Tate 2012; Wacker et al. 2017b). In addition, ligands can modulate signaling upon binding to sites other than the orthosteric binding site (allosteric sites), or can interact with both orthosteric and allosteric sites (bitopic ligands) (Wacker et al. 2017b). Crystal structures of serotonin_1B bound to agonist ergotamine (Wang et al. 2013; Fig. 4a–c) and inverse agonist methiothepin (Yin et al. 2018; Fig. 4d–f) display significant conformational perturbations. Compared to the methiothepin-bound inactive structure, the most striking difference in the agonist-bound (active) structure is the outward movement of the intracellular end of TM6, a hallmark of GPCR activation, that expand the cytoplasmic pocket of the TM bundle for coupling downstream signaling effectors (Yin et al. 2018). In addition, the side-chains of the residues in the conserved motifs (sometimes denoted as microswitches) such as D(E)RY, P-I-F, and NPxxY motifs exhibit significant positional and/or rotational displacements between ergotamine and methiothepin-bound structures (Fig. 4a–f). Similar conformational excursions have been found in agonist ergotamine and inverse agonist ritanserin-bound structures of serotonin_2C receptor, which represent the active and inactive states of the receptor, respectively (Peng et al. 2018). Molecular modeling studies revealed that while specific ligands could capture the receptor in a specific conformational state, ligand-free model of the serotonin_2A receptor shows clear distinction in the extracellular and intracellular end of transmembrane helices, and conserved microswitches from the ligand-bound structures (Mozumder et al. 2019). On the other hand, serotonin_3A receptor shows various conformational states not only upon binding to different ligands such as tropisetron and serotonin (Polovinkin et al. 2018), but also upon binding to the same ligand serotonin (Basak et al. 2018a). Structural differences upon ligand binding are largely restricted to the TMD of the serotonin_3A receptor, leading to conformational modulation of the pore.

Fig. 4.

Conformational diversity among serotonin receptors belonging to the GPCR family. Conformational states of (a, d, and g) ionic lock, (b, e, and h) P-I-F motif, and (c, f, and i) Y^7.53 of NPxxY motif in ergotamine-bound serotonin_1B (PDB ID: 4IAR (Wang et al. 2013), cyan), methiothepin-bound serotonin_1B (PDB ID: 5V54 (Yin et al. 2018), orange), and ergotamine-bound serotonin_2B (PDB ID: 4IB4 (Wacker et al. 2013), magenta) receptors. Whereas the ergotamine-bound serotonin_1B adopts a classical agonist-induced active-like state that modulates the canonical signaling pathway, ergotamine-bound serotonin_2B receptor structure exhibits conformational characteristics of both the active and inactive states which display strong functional selectivity for β-arrestin signaling. Structures of GPCRs were generated from their respective PDB IDs using PyMOL Molecular Graphics System (Schrödinger, LLC). See text for more details

In contrast to the same serotonin receptor subtypes, receptors from different subfamilies adopt different conformational states upon binding to the same ligand. Crystal structures of ergotamine-bound serotonin_1B (Wang et al. 2013) and serotonin_2B (Wacker et al. 2013) receptors show structural distinction over intracellular segment of TM5 as well as D(E)RY, P-I-F, and NPxxY motifs (Fig. 4a–c and Fig. 4g–i). Whereas ergotamine-bound serotonin_1B receptor adopts a classical agonist-induced active-like state that modulates the canonical signaling pathway, ergotamine-bound serotonin_2B receptor structure exhibits conformational characteristics of both the active and inactive states which display strong functional selectivity for β-arrestin signaling (Wacker et al. 2013). Beyond the ligand-bound structures, the structure of serotonin_1B receptor bound to the agonist donitriptan and coupled to an engineered Gα_o heterotrimer (PDB ID: 6G79) has been solved by cryo-EM (see Fig. 5, García-Nafría et al. 2018), indicating an active state conformation of the intracellular domain.

Fig. 5.

High-resolution cryo-EM structure of serotonin_1B receptor coupled to heterotrimeric G-protein. a Surface representation of the receptor (EMDB ID: 4358 (García-Nafría et al. 2018)). b Cartoon representations of serotonin_1B receptor (in green), Gα_o (in red), Gβ (in cyan), and Gγ (in yellow) molecular models (PDB ID: 6G79 (García-Nafría et al. 2018)) fitted into the cryo-EM density map (semi-transparent blue surface). c The orthosteric binding site in serotonin_1B receptor (blue mesh with the molecular model (green) in cartoon) bound to donitriptan (shown as red stick). The blue mesh is derived from high-resolution cryo-EM structure of serotonin_1B receptor coupled to heterotrimeric G-protein generated from EMDB ID: 4358 (García-Nafría et al. 2018). Structures of GPCRs were generated from their respective EMDB and PDB IDs using UCSF Chimera (https://www.cgl.ucsf.edu/chimera) and PyMOL Molecular Graphics System (Schrödinger, LLC)

Ligand binding sites in serotonin receptors

Ligand binding sites in GPCRs belonging to the serotonin receptor family are mainly formed by TM3, TM5, TM6, TM7, and ECL2. They are strikingly similar in their overall architecture with subtle divergence exhibited in the binding residues (Fig. 6). It is found that the orthosteric site is embedded deep in the transmembrane core, encompassing helices TM3, TM5, TM6, and TM7, whereas the extended binding site occupies the extracellular portions of TM3, TM5, TM6, and TM7 as well as ECL2. While the former binding site contributes to the ligand binding affinity, the latter site determines ligand selectivity (Wacker et al. 2013; Wang et al. 2013).

Fig. 6.

Comparison of ligand-receptor interactions in serotonin receptors belonging to the GPCR family. Crystal structures showing binding of (a) agonist ergotamine with serotonin_1B (PDB ID: 4IAR) (Wang et al. 2013) and (c) inverse agonist risperidone with serotonin_2A (PDB ID: 6A93) (Kimura et al. 2019) receptor. Close-up views of the ligand binding site of (b) serotonin_1B and (d) serotonin_2A receptors in two different orientations are shown. Residues involved in ligand binding are labeled and the most conserved residues at the binding site are highlighted with boxes. Structures of GPCRs were generated from their respective PDB IDs using PyMOL Molecular Graphics System (Schrödinger, LLC). See text for more details

From recent crystal structures of serotonin_1B (Wang et al. 2013), serotonin_2A (Kimura et al. 2019), serotonin_2B (Wacker et al. 2013), and serotonin_2C (Peng et al. 2018) receptors bound to agonists or inverse agonists, it is evident that amine-containing ligands mainly interact with the ligand binding pocket through salt bridge between positively charged nitrogen and the carboxyl group of the conserved D^3.32 residue (Fig. 6). Mutation of D^3.32 significantly disrupts the binding of amine-containing ligand to the serotonin_1B receptor (Yin et al. 2018) and the serotonin₄ receptor (Claeysen et al. 2003). The residue 3.36 (serine in serotonin₂ receptor, threonine in serotonin₄ receptor and cysteine in all other serotonin receptors) is localized just a turn below the residue D^3.32 and faces the ligand binding pocket. In the serotonin_2A receptor, mutation of S^3.36 to alanine significantly affects the affinity of primary amine ligands (such as serotonin and tryptamine), but not the affinity of secondary or tertiary amine ligands (such as N,N-dimethyl tryptamine (DMT) and 5-methoxy-DMT) (Almaula et al. 1996). In addition, W^6.48, F^6.51, F^6.52, and Y^7.43 are conserved aromatic residues in the binding pocket of all GPCRs belonging to the serotonin receptor family and could form a narrow hydrophobic cleft for important π-π stacking interactions with the ligand molecules containing aromatic ring systems (Fig. 6b, d). Mutations in F^6.51, F^6.52, and Y^7.43 reduce the ligand binding affinity in serotonin_2A receptor (Kimura et al. 2019), indicating significant loss of π-π aromatic interactions between receptor and ligand molecules. Importantly, various point mutations in the transmembrane domains of the serotonin_1A receptor have helped to identify key residues required for binding of the natural ligand serotonin. Mutations in Asp82 to Asn, Asp116 to Asn, and Ser199 to Ala resulted in a marked reduction in affinity for serotonin, whereas the binding characteristics of the antagonist pindolol remained unaffected (Ho et al. 1992). Interestingly, these mutations did not alter the ability of the receptor to induce GTP hydrolysis upon addition of excess serotonin indicating that the mutants were competent in terms of their interaction with G-proteins.

Compared to the orthosteric site, ligand binding residues in the extended binding sites are more diverse among serotonin receptors in the GPCR family. In serotonin_1B receptor, the extended binding pocket is much broader than that of serotonin_2B receptor due to the 3.0 Å outward shift of the top of TM5 (Fig. 6a) and the presence of T209^5.39 in the corresponding position of the bulky M218^5.39 in serotonin_2B receptor (Wang et al. 2013). It is important to note that the overall size of the extended binding pocket can significantly change depending upon ligand types. In the serotonin_2B receptor, the extended binding site shows 28.6% reduction in overall volume upon binding to lysergic acid diethylamide (LSD) relative to ergotamine (Wacker et al. 2017a). Many single point mutations, including the conserved D^3.32, within the orthosteric binding pocket reduce or abolish the binding of LSD at both serotonin_1B serotonin_2B receptors (Wang et al. 2013).

In contrast to the serotonin receptors belonging to the GPCR family, the ligand binding site of the ion channel serotonin₃ receptor is located at the interface of two adjacent subunits at the N-terminal ECD and is formed by three loops from one (the principal) subunit and portions of β-strands and four loops from the adjacent or complementary subunit (Fig. 7). Recent cryo-EM structures of mouse serotonin_3A bound to agonist serotonin (Basak et al. 2018a) and antagonist granisetron (Basak et al. 2019) show a well conserved binding pocket nestled by W156 and Y207 of principal subunit and W63, R65, and Y126 of complementary subunit (Fig. 7c). However, W168 experiences a large orientational difference upon binding to agonist or antagonist.

Fig. 7.

Cryo-EM structure of serotonin-bound serotonin_3A receptor. a Surface representation shows electron density map of the receptor (EMDB ID: 7882 (Basak et al. 2018a)); b superimposition of the cartoon structure (light blue) of the receptor (PDB ID: 6DG7 (Basak et al. 2018a)) into the cryo-EM density map (semi-transparent purple surface), with the serotonin molecule represented as a magenta stick; c close-up view of the serotonin-binding site. The residue labels on the principal and complementary subunits are marked in black and red, respectively. Structures of GPCRs were generated from their respective EMDB and PDB IDs using UCSF Chimera (https://www.cgl.ucsf.edu/chimera) and PyMOL Molecular Graphics System (Schrödinger, LLC). See text for more details

Ligand-induced activation-inactivation switch in serotonin receptors

The activation or inactivation of all serotonin receptors belonging to the GPCR family is similar to class A GPCRs and is characterized by the rearrangement of the amino acid side-chains in the conserved microswitches (Katritch et al. 2013; McCorvy and Roth 2015; Filipek 2019). The inactive and active states of serotonin receptors are originally thought to be classified based on the presence and absence, respectively, of a salt bridge between D/E^3.49 and R^3.50 in the D(E)RY motif. However, recent crystal structures of agonist (ergotamine)-bound serotonin_1B (Wang et al. 2013) and serotonin_2C receptors (Peng et al. 2018), as well as inverse agonist (risperidone)-bound serotonin_2A (Kimura et al. 2019) and antagonist (ritanserin)-bound serotonin_2C (Peng et al. 2018) receptors, show broken ionic lock in their structures, whereas the ionic lock is preserved in ergotamine-bound serotonin_2B receptor structures (Wacker et al. 2013; McCorvy et al. 2018; see Fig. 4a and g). Therefore, the salt bridge interaction in the ionic lock may not be a true indicator of the active and inactive states of the receptor. It is important to note that the intact salt bridge in serotonin_2B receptor displays strong functional selectivity of the receptor for β-arrestin signaling (Wacker et al. 2013). This implies that the salt bridge interaction could be important to determine both types of downstream signaling pathways, either canonical signaling via G-proteins or non-canonical signaling via β-arrestins.

In contrast to the D/E^3.49 and R^3.50 interaction, a conserved salt bridge between R^3.50 and E^6.30 is crucial to distinguish between active and inactive states. In inactive conformation the interaction between R^3.50 and E^6.30 remains intact, whereas upon receptor activation, the salt bridge is broken (Wacker et al. 2013; Wang et al. 2013; McCorvy et al. 2018; Peng et al. 2018). Importantly, the inverse agonist (ritanserin)-bound and agonist (ergotamine)-bound serotonin_2C receptor structures exhibit presence and absence of the salt bridge, respectively (Peng et al. 2018). Taken together, the salt bridges in the D(E)RY motif play an important role in signaling of serotonin receptors by regulating the movement of the intracellular segment of TM3 and TM6. Importantly, mutagenesis studies further supported the role of the D(E)RY motif to stabilize the ground (inactive) state of the receptor because mutations of this motif in the serotonin_2A receptor usually result in constitutive activity (Shapiro et al. 2002).

The conserved P-I-F motif appears to be a critical trigger for receptor activation and inactivation, especially in the light of crystal structures of agonist-bound serotonin_1B (Wang et al. 2013), serotonin_2B (Wacker et al. 2013; McCorvy et al. 2018) and serotonin_2C (Peng et al. 2018) receptors, and inverse agonist-bound serotonin_2A (Kimura et al. 2019) and serotonin_2C (Peng et al. 2018) receptors. Comparison of active vs. inactive structures reveals inward shifts of P^5.50 and F^6.44 as well as rotamer switch of I^3.40 upon receptor activation, resulting in a rotation and overall movement of TM6 (see Fig. 4b, e). The conserved Y^7.53 in the NPxxY motif functions as a “gatekeeper” to allow an opening for a water channel to flow through upon receptor activation (Yuan et al. 2014). However, recent crystal structures show that the NPxxY motif is crucial for receptor signaling either via G-protein or β-arrestin (McCorvy and Roth 2015). While Y^7.53 moves and rotates toward the helical bundle upon receptor activation in the canonical pathway, Y^7.53 shows more pronounced inward rotation in a β-arrestin biased state (see Fig. 4c–i). Interestingly, in serotonin_2C receptor, substituting Ala or Cys for Y^7.53 resulted in a marked increase in the basal level of cellular signaling and makes these mutants constitutively active (Rosendorff et al. 2000). The P-I-F and NPxxY motifs in ligand-free ground state structure are believed to lie close to the inactive state conformation.

On the other hand, serotonin₃ receptor exhibits significant structural difference in apo, agonist (serotonin)-bound, and antagonist (granisetron)-bound states. Conserved residues surrounding the binding site including W63, R65, Y126, W156, F199, and Y207 undergo rotameric reorientations over apo, serotonin-bound, and granisetron-bound structures. In addition, the loop C and TMD conformations of antagonist-bound serotonin₃ receptor are distinct and lie in between apo and agonist-bound conformations (Basak et al. 2019).

Significance of the conformational motions in receptor function

Serotonin receptors exhibit considerable flexibility around the loops as well as the transmembrane helices which are linked to their different signaling functions. While majority of the crystal structures of serotonin receptors provide a high-resolution yet static picture of the lowest energy state, multiple conformational states in ligand-free and -bound conditions would throw light onto the conformational plasticity of the receptor. The ligand binding pocket of serotonin receptors differs from one conformation to another depending upon the nature of the ligand. Although the binding of agonist and antagonist shares some common conserved residues in the binding pocket, the conformational arrangement of these residues significantly differs among agonist- and antagonist-bound states (Fig. 6b, d). In addition, the specificity of orthosteric ligand binding could be provided by allosteric modulators, thereby providing ligand specificity to the receptors (Wootten et al. 2013; Gentry et al. 2015). In some cases, allosteric modulators could also influence receptor responses without the binding of orthosteric ligands (van der Westhuizen et al. 2015). Importantly, information on binding pocket dynamics is important to design selective ligand molecules.

In addition to the ligand binding pocket, the transmembrane helices and the intracellular and extracellular loops exhibit substantial flexibility which is significantly modulated upon ligand binding. Extensive flexibility of the intracellular loops is crucial to accommodate a diverse range of effector molecules (such as G-protein and β-arrestin) (Latorraca et al. 2017). For serotonin receptors, ECL2 is usually the longest extracellular loop and its flexibility is essential to allow a variety of ligand molecules into the binding pocket (Wheatley et al. 2012; Kimura et al. 2019). On the other hand, intracellular loops, mainly ICL2 and ICL3, form the binding cavity for G-protein and therefore conformational dynamics of ICL2 and ICL3 are crucial for signaling by serotonin receptors (Prossnitz et al. 1993; Arora et al. 1995; Gáborik et al. 2003; Chakir et al. 2003; Pydi et al. 2014). Although the dynamic information of ICL3 is missing in available structures due to truncation of the loop region, estimates of motional constraint are available for serotonin_1A (Pal et al. 2018) and serotonin_2A (Mozumder et al. 2019) receptors. The extracellular and intracellular segments of transmembrane helices, in particular TM3, TM5, and TM6, undergo conformational transition due to different types of ligand binding. A specific conformational state of transmembrane helices is responsible for a specific receptor response over another.

Interestingly, superimposition of X-ray and cryo-EM structures of several serotonin receptors did not exhibit a significant difference in the overall secondary structural fold (see Fig. 8). However, cryo-EM structures of serotonin receptors belonging to the GPCR family showed conformational perturbations in the intracellular region of the receptor relative to the corresponding X-ray structures. This perturbation could be due to the binding of G-protein in the intracellular part of the receptor in the cryo-EM structures. This is supported by the absence of structural difference between X-ray and cryo-EM structures of serotonin_3A receptor, which is an ion channel and not coupled to G-proteins. Most of the perturbations in the serotonin_3A receptor was largely restricted to the loop regions of the receptor.

Fig. 8.

Overview of X-ray (in red) and cryo-EM (in blue) structures of serotonin receptors. Cartoon representation of (a) serotonin_1B (PDB IDs: 4IAQ (red) and 6G79 (blue)), (b) serotonin_2A (PDB IDs: 6A93 (red) and 6WHA (blue)) and (c) serotonin_3A (PDB IDs: 4PIR (red) and 6HIQ (blue)) receptors. Structural alignment of X-ray and cryo-EM structures exhibits root-mean-square deviation of 0.59 Å, 1.68 Å, and 1.15 Å for serotonin_1B, serotonin_2A and serotonin_3A receptors, respectively. Structures of serotonin receptors were generated from their respective PDB IDs using PyMOL Molecular Graphics System (Schrödinger, LLC). See text for more details

Lipid interactions of serotonin receptors: implications in function, dynamics and oligomerization

A hallmark of membrane proteins and receptors is their interaction with membrane lipids. This is due to the fact that a considerable portion of integral membrane proteins (receptors) lie buried in the membrane bilayer, thereby facilitating their interaction with proximal membrane lipids. For example, molecular dynamics simulations show that the lipid-receptor interface for rhodopsin, a representative GPCR, corresponds to ~ 38% of the total surface area of the receptor (Huber et al. 2004). This raises the obvious possibility that the membrane lipids could be an important modulator of receptor structure and function (Lee 2004). The importance of membrane lipids in the function of GPCRs was apparent even in early days of GPCR research from the adverse effects of delipidation on receptor function (Cerione et al. 1983; Kirilovsky and Schramm 1983). Although the interaction with lipids is important for the function of membrane proteins, it offers a unique challenge to structural biology of membrane proteins and receptors.

Fortunately, the most extensive literature on lipid-GPCR interaction is available for a member of the serotonin receptor family, namely, the serotonin_1A receptor (reviewed in Pucadyil and Chattopadhyay 2006; Paila and Chattopadhyay 2010; Jafurulla and Chattopadhyay 2013; Sengupta and Chattopadhyay 2015; Sengupta et al. 2017, 2018). Using a combination of biophysical, biochemical, cell biological, and molecular dynamics simulation approaches, it has been established that membrane lipids, such as cholesterol and sphingolipids, play a crucial role in ligand binding, G-protein coupling, signaling, and oligomerization of the serotonin_1A receptor. In addition, in a recent work, it has been shown that the endocytic pathway and intracellular trafficking of the serotonin_1A receptor is modulated by cholesterol (Kumar et al. 2019; Kumar and Chattopadhyay 2020). There are a few reports on the sensitivity of the serotonin₇ receptor to cholesterol and sphingolipids (Sjögren and Svenningsson 2007a, b; Sjögren et al. 2006). In addition, membrane cholesterol has been reported to play an important role in the oligomerization of the serotonin_2a and serotonin_2C receptors (Shan et al. 2012; Massaccesi et al. 2020). In contrast to the rich literature on cholesterol sensitivity of serotonin receptors belonging to the GPCR family, there is scant literature on cholesterol sensitivity of serotonin receptors which act as ion channels (Nothdurfter et al. 2010). Over a period of time, lipid (cholesterol) sensitivity of other GPCRs such as the β₂-adrenergic, cannabinoid (CB1), bitter taste (T2R4), cholecystokinin (CCK), opioid, oxytocin, galanin, and chemokine receptors has been reported (this list is not exhaustive, only representative; reviewed in Gimpl 2016; Oates and Watts 2011; Jafurulla et al. 2019). In spite of this existing literature, the exact mechanism behind cholesterol sensitivity of GPCR function remains elusive. A school of thought has developed which proposes mechanisms based on direct interaction of cholesterol with GPCRs. Yet, mechanisms based on the ability of cholesterol to modulate GPCR microenvironment in the membrane (fluidity, hydrophobic thickness and dipole potential) have also been postulated. Of course, a combination of these two mechanisms represents another possibility (for a comprehensive discussion, see Jafurulla et al. 2019).

Interestingly, closely bound cholesterol molecules have been reported in crystal structures of the serotonin_2A and serotonin_2B receptors (Liu et al. 2013; Wacker et al. 2013; Wacker et al. 2017a; McCorvy et al. 2018; Kimura et al. 2019; Kim et al. 2020; see Fig. 9). Closely bound cholesterol molecules were first observed in the crystal structure of the β₂-adrenergic receptor (Cherezov et al. 2007) and have become a hallmark for a number of crystal structures of many GPCRs (see Jafurulla et al. 2019 for a comprehensive list of GPCRs displaying bound cholesterol in their structures). In a cautionary fashion, we would like to point out that no clear link has been established yet between GPCRs displaying bound cholesterol in their structures and cholesterol sensitivity in GPCR function. The reason for this lies in the lack of experimental data demonstrating functional cholesterol sensitivity for these receptors. We believe that this link (or lack of it) would become more clear as more functional data become available for GPCRs exhibiting closely bound cholesterol in their structures.

Fig. 9.

Crystal structures of serotonin receptors with bound cholesterol molecules. Bound cholesterol molecules in crystal structures of serotonin receptors belonging to the GPCR family (the corresponding PDB IDs are indicated in parentheses): a serotonin_2A receptor (6A93), b serotonin_2A receptor (6WGT), c serotonin_2B receptor (4IB4), d serotonin_2B receptor (4NC3), e serotonin_2B receptor (5TVN), and f serotonin_2B receptor (6DRX). Snapshots of cholesterol-bound (cholesterol shown in green) structures of serotonin receptors (receptors shown in gray) were generated from their respective PDB structures using UCSF Chimera (https://www.cgl.ucsf.edu/chimera). See text for more details

Specific interaction of GPCRs with membrane cholesterol is often attributed to structural features of these receptors that allow their preferential association with cholesterol (Sarkar and Chattopadhyay 2020). In this context, cholesterol interaction motifs represent putative interaction sites in GPCRs that could induce cholesterol-sensitive function. A popular motif is the cholesterol recognition/interaction amino acid consensus (CRAC) motif characterized by the sequence -L/V-(X)_1–5-Y-(X)_1–5-R/K- (from N-terminus to C-terminus of the protein), where (X)_1–5 represents between one and five residues of any amino acid (Epand 2006; Fantini and Barrantes 2013). The serotonin_1A receptor, a prototypical GPCR displaying cholesterol sensitivity, has three CRAC motifs in TM2, 5, and 7 (Jafurulla et al. 2011; see Fig. 10). It should be noted here that mere presence of cholesterol interaction motifs does not necessarily indicate cholesterol-sensitive function of GPCRs. For example, the neurotensin type-1 and secretin receptors have cholesterol consensus motif (CCM) in their TM4 (Hanson et al. 2008). However, both these GPCRs fail to exhibit any appreciable change in downstream signaling response upon cholesterol depletion relative to untreated cells (Harikumar et al. 2005; Oates et al. 2012).

Fig. 10.

A schematic representation depicting the topological features and amino acid sequence of the human serotonin_1A receptor embedded in a membrane bilayer. The putative positions of the transmembrane helices of the human serotonin_1A receptor were predicted using the crystal structure of the human serotonin_1B receptor (PDB ID: 6G79 (García-Nafría et al. 2018)) as a template and the amino acids in the receptor sequence are shown as circles. The receptor has seven transmembrane stretches depicted as putative α-helices. The exact boundary between the membrane and the aqueous phase is not known and therefore location of the amino acid residues relative to the membrane bilayer is putative. The serotonin_1A receptor consists of three CRAC motifs in TM2 (CRAC I, shown in yellow), TM5 (CRAC II, shown in green), and TM7 (CRAC III, shown in blue). The palmitoylation at the C-terminus is shown as blue zigzag lines. Adapted and modified with permission from Sarkar and Chattopadhyay 2020, Copyright 2020 Wiley Periodicals, Inc.

Coarse-grain molecular dynamics simulations using the MARTINI force-field helped identify high cholesterol occupancy at the CRAC motif in TM5 of the serotonin_1A receptor (Sengupta and Chattopadhyay 2012; see Fig. 11a). An additional advantage offered by simulations is that it provides information on receptor dynamics, generally not available from crystal structures (Torrens-Fontanals et al. 2020). Simulations further showed that these sites are characterized by dynamics associated with cholesterol, typically in the ns-μs timescale. The corresponding energy landscape for association of cholesterol with GPCRs can be viewed as a series of shallow minima, interconnected by low energy barriers (Fig. 11b) (Sengupta and Chattopadhyay 2015). Figure 12 depicts a schematic representation of the dynamics of the fast and slow exchange sites in the serotonin_1A receptor. Taken together, the timescales of binding/unbinding of cholesterol at the CRAC sites are in the μs range, while the association/dissociation of cholesterol takes place at the ns timescale at most other sites.

Fig. 11.

Cholesterol interaction hot spots in the human serotonin_1A receptor. a Residue-wise maximum occupancy of cholesterol bound to the serotonin_1A receptor, obtained by coarse-grain molecular dynamics simulations. Maximum occupancy time (defined as the maximum time a given cholesterol molecule is found at a particular residue during the entire length of the simulation) of cholesterol at each amino acid of the serotonin_1A receptor was averaged and normalized over simulations carried out at varying concentrations of cholesterol. The transmembrane helices are represented as gray bands, and CRAC motifs are highlighted with the same color coding as in Fig. 10. Adapted and modified with permission from Sengupta and Chattopadhyay (2012) Identification of cholesterol binding sites in the serotonin_1A receptor. J Phys Chem B 116:12991–12,996 DOI: 10.1021/jp309888u), Copyright 2012 American Chemical Society. b A schematic energy landscape corresponding to cholesterol interaction sites in GPCRs. The energy landscape of cholesterol interaction is represented as a series of shallow minima interconnected by low energy barriers. The abscissa can be thought to correspond to individual occupancy sites represented by single residues or by a sub-space at the receptor surface (such as cholesterol consensus motif (CCM) or CRAC sites). The occupancy sites are most likely to be accessed via an exchange with the annular lipids and less often by direct site hopping of cholesterol. Note that the energy barriers and the minima could be modulated by other membrane lipids such as sphingolipids. Reprinted with permission from Sengupta and Chattopadhyay 2015, Copyright 2015 Elsevier B.V

Fig. 12.

Dynamics and diversity of cholesterol occupancy sites in GPCRs. The schematic shows weak, dynamic yet crucial cholesterol interaction sites on GPCRs, governed by cholesterol and receptor dynamics. A few sites show fast exchange with bulk membrane lipids and low occupancy at the receptor surface. Other sites characteristically exhibit slow exchange with bulk lipids and high occupancy at the GPCR surface. Fast ns timescale dynamics is observed in several sites. Site hopping in μs timescale between proximal sites is also observed. Reprinted with permission from Sengupta and Chattopadhyay 2015, Copyright 2015 Elsevier B.V and Sengupta et al. 2018 (https://pubs.acs.org/doi/10.1021/acs.jpcb.8b01657), Copyright 2018 American Chemical Society

Concluding remarks and the road ahead

Available structures of serotonin receptors in complex with a variety of ligands (e.g., agonists, inverse agonists, and antagonists) and effectors (e.g., G-proteins) outline the molecular mechanism of receptor activation and inactivation upon ligand binding in atomic details. Activation and inactivation mechanisms of serotonin receptors share several common yet distinct key features, including movement of helices and amino acid side-chains in the conserved microswitches. Interestingly, significant structural differences have been identified not only between active and inactive states of the receptors, but also among various active conformations, illustrating the process by which receptors activate a specific pathway over another. However, truncations and mutations in the available structures have limited our understanding of conformational dynamics of flexible loop regions that play a pivotal role in receptor activation and inactivation (Haldar et al. 2010; Huang et al., 2016; Pal et al. 2018; Pal and Chattopadhyay 2019; Dijkman et al. 2020). Determination of full-length receptor structures is clearly necessary in order to unravel the in-depth molecular mechanism of regulation of intracellular signaling pathways by serotonin receptors. This is becoming increasingly apparent as more reports appear elucidating the crucial role(s) of the extra- and intracellular loops in activation and signaling of GPCRs (Pal and Chattopadhyay 2019; Ma et al. 2020).

Serotonin receptors act as potential drug targets for a multitude of human diseases. However, drug molecules often interact with more than a single target through a process known as “polypharmacology” and the unintended interactions can lead to side effects. Interaction of a single drug with multiple serotonin receptors is increasingly evident for the treatment of obesity, drug abuse, and schizophrenia (Roth et al. 2004). Although development of ligands targeting selective serotonin receptors (prospective drug molecules) with a defined polypharmacological profile (the so called magic bullet approach) has proved to be extremely difficult, structure-based rational drug design approach has the potential to develop target-selective drugs with improved therapeutic profiles. However, it requires multitude of high-resolution receptor structures with diverse ligands to elucidate the mechanism of polypharmacology at multiple defined drug targets. The other approach to address this issue is to design non-selective drugs (sometimes termed as “magic shotguns”) that interact with several receptors as targets and could result in effective therapeutic intervention (Roth et al. 2004). Having said that, it may be noted that recent crystal and cryo-EM structures of serotonin receptors have helped to design novel drugs selective to specific serotonin receptors, including serotonin_1A (Becker et al. 2016), serotonin_1B (Tatarczyńska et al. 2004), serotonin_2B (Kovács et al. 2003), serotonin_2C (Kovács et al. 2003), and serotonin₃ (Gan 2005; Theriot et al. 2020) receptors.

Notably, several mutations of serotonin receptors are reported to be associated with different psychiatric disorders (Drago et al. 2008; Banlaki et al. 2015; Xia et al. 2018). Mapping the structural tweaking caused by the important pathological mutants of serotonin to understand the functional traits of the mutant receptors relative to the wild-type receptor would be of immense relevance. We believe that recent advances in high-resolution structural biology of GPCRs could provide new insights in this area.

As mentioned above, serotonin receptors are ancient receptors from an evolutionary perspective and represent hubs for an array of signaling pathways that are associated with human emotion, happiness, anxiety, depression, and overall health. Their usefulness as drug targets go beyond neuropsychiatric disorders and extends up to cancer (Fiorino et al. 2014). The important role played by serotonin receptors in human health and happiness has contributed to their immense popularity. Nonetheless, serotonin receptors currently appear to be at a crossroad, since some of the key serotonin receptors (such as the serotonin_1A receptor) has not lend itself to purification and structural analysis yet. Having said that, serotonin receptors (especially, the serotonin_1A receptor) enjoy a special position among GPCRs in terms of their sensitivity to lipids in general and cholesterol in particular due to the breadth and depth of available information in this area. This would help in translating emerging structural details in terms of a functional membrane interaction paradigm. With tremendous advancements in GPCR structural biology in the recent past, it is envisioned that these challenges would be overcome soon, enabling us to carry out in-depth analysis of serotonin receptors at all levels allowing fine-tuned effective drug development against the backdrop of receptor-lipidome interaction.

Funding

This work was supported by Science and Engineering Research Board (SERB) Distinguished Fellowship grant to A.C., Council of Scientific and Industrial Research (CSIR) Network project UNSEEN (BSC0113) and Department of Science and Technology (DST) Ramanujan fellowship to Sujoy Mukherjee.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.