STRUCTURAL BASIS FOR FUNCTIONAL MODULATION OF PENTAMERIC LIGAND-GATED ION CHANNELS
by
YVONNE W. GICHERU
Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Thesis Advisor: Sudha Chakrapani, Ph.D.
Department of Physiology and Biophysics
CASE WESTERN RESERVE UNIVERSITY
May 2019
CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis/dissertation of
YVONNE W. GICHERU
Candidate for the degree of Physiology and Biophysics*
Witold Surewicz (Committee Chair)
Matthias Buck
Stephen Jones
Vera Moiseenkova-Bell
Rajesh Ramachandran
Sudha Chakrapani
March 27, 2019
*We also certify that written approval has been obtained for any proprietary material contained therein.
Dedication
To my family, friends, mentors, and all who have supported me through this process, thank you.
Table of Contents
List of Figures ...... iv
List of Abbreviations ...... v
Abstract ...... vi
Chapter 1 ...... 1
Introduction ...... 1
1.1 Pentameric ligand-gated ion channel (pLGIC) superfamily ...... 2
1.2 pLGIC architecture ...... 4
1.3 Gating in pLGIC ...... 7
1.4. Nicotinic acetylcholine receptors (nAChRs) ...... 10
1.5 Serotonin receptors (5-HT3R) ...... 16
1.6 γ- Aminobutyric acid type A receptors (GABAAR) ...... 24
1.7 Glycine receptors (GlyR) ...... 29
1.8 Prokaryotic receptors ...... 33
1.9 Purpose of this study ...... 37
Chapter 2 ...... 43
Allosteric modulation of pLGIC function by long chain polyunsaturated fatty acids (PUFAs) ...... 43
Introduction ...... 44
Results ...... 50
DHA modulates the pH-elicited response in GLIC ...... 50
Structural changes associated with DHA binding in GLIC ...... 53
AA and EPA modulate the pH-elicited response in GLIC ...... 55
i
PUFA effect on 5-HT3AR function ...... 56
Membrane environment for structural studies ...... 57
Discussion ...... 58
Methods ...... 61
Chapter 3 ...... 78
Understanding allosteric and orthosteric inhibition mechanisms of the 5-
HT3A receptor ...... 78
Introduction ...... 79
Results ...... 85
Cannabinoid inhibition of 5-HT3AR currents ...... 85
Ligand binding sites in 5-HT3AR ...... 85
Granisetron inhibition of 5-HT3AR currents ...... 87
Granisetron binding and conformational changes ...... 88
Discussion ...... 92
Methods ...... 95
Chapter 4 ...... 114
A commentary on the relevance of a membrane environment for ion channel gating ...... 114
Foreword ...... 115
Introduction ...... 116
Impact of membrane environment on channel gating and dynamics .... 117
New insights from noncanonical domain changes? ...... 120
Chapter 5 ...... 123
Discussion and future directions ...... 123
Discussion ...... 124
ii Future directions ...... 128
Appendix ...... 133
References ...... 134
iii List of Figures
Figure 1. 1: A schematic representation of the topology and arrangement in pLGICs. ... 39 Figure 1. 2: nAChR structure from Torpedo membranes...... 40 Figure 1. 3: Representative eukaryotic pLGIC structures...... 41 Figure 1. 4: Representative prokaryotic pLGICs...... 42
Figure 2. 1: Essential long chain polyunsaturated fatty acids (PUFAs)...... 66 Figure 2. 2: DHA modulation of GLIC function ...... 67 Figure 2. 3: Effect of DHA on GLIC desensitization...... 68 Figure 2. 4: Effect of DHA pre-application on GLIC currents...... 69 Figure 2. 5: DHA has no effect on the non-desensitizing GLIC I9A/H11F mutant...... 70 Figure 2. 6: Lack of DHA effect on the non-desensitizing GLIC I9A/H11F mutant...... 71 Figure 2. 7: DHA binding site in GLIC...... 72 Figure 2. 8: R118A mutation reduces the effect of DHA on desensitization...... 73 Figure 2. 9: Conformational changes in the GLIC pore...... 74 Figure 2. 10: EPA and AA modulation of GLIC function...... 75
Figure 2. 11: DHA, EPA and AA effect on 5-HT3AR function...... 76 Figure 2. 12: Changes in M4 distance measured by DEER for GLIC in nanodiscs...... 77
Figure 3. 1: Endocannabinoids and phytocannabinoids...... 101
Figure 3. 2: Cannabinoid modulation of 5-HT3AR function...... 102
Figure 3. 3: Cryo-EM reconstruction of THC-5HT-bound full-length 5-HT3AR...... 103
Figure 3. 4: Comparison of serotonin-bound 5-HT3AR with 5-HT3AR-THC-5HT...... 104
Figure 3. 5: Granisetron inhibition of 5-HT3AR currents...... 105
Figure 3. 6: Cryo-EM structure of granisetron-bound full-length 5-HT3AR...... 106 Figure 3. 7: The granisetron-binding site...... 107 Figure 3. 8: Differences between granisetron and tropisetron binding poses...... 108 Figure 3. 9: Conformational differences between the apo and ligand-bound states. .... 109 Figure 3. 10: Effects of mutations at the ligand-binding pocket...... 111
Figure 3. 11: Functional characterization of R65A 5-HT3AR...... 113
iv List of Abbreviations α-Bgx α-Bungarotoxin
AA cis-5,8,11,14-Eicosatetraenoic acid AChBP Acetylcholine binding protein 2-AG 2-arachidonoylglycerol AEA N-arachidonoylethanolamine CBD Cannabidiol CNS Central nervous system Cryo-EM Cryo-electron microscopy DHA cis-4,7,10,13,16,19-Docosahexaenoic acid EC Endocannabinoid ECS Endocannabinoid system EPA cis-5,8,11,14,17-Eicosapentaenoic acid ELIC Erwinia chrysanthemi ligand-gated ion channel GI Gastrointestinal GLIC Gloeobacter violaceous ligand-gated ion channel
GABAAR γ-aminobutyric acid receptor type A GlyR Glycine receptor
5-HT3AR Serotonin receptor type 3A 5-HT Serotonin MDs Molecular dynamics simulations nAChR Nicotinic acetylcholine receptor PLGIC Pentameric ligand-gated ion channel PNS Peripheral nervous system PUFA Polyunsaturated fatty acid TEVC Two-electrode voltage clamp THC ∆-9-Tetrahydrocannabinol
v Structural Basis for Functional Modulation of Pentameric Ligand-gated Ion Channels
Abstract by YVONNE GICHERU
Pentameric ligand-gated ion channels (pLGICs) are transmembrane proteins which mediate fast excitatory or inhibitory neurotransmission. The composition of the plasma membrane regulates the function of these channels. Bioactive lipids and cannabinoids have been shown to allosterically modulate the function of pLGICs and the mechanisms by which they affect receptor activity continue to be actively investigated. Here, we examine how essential omega-3 and omega-6 fatty acids, and endogenous/ plant cannabinoids modulate pLGICs by functional and structural studies. This work also addresses the structural basis for inhibition in a cationic pLGIC by a clinically approved antagonist.
We show that docosahexaenoic acid (DHA) enhances desensitization in two cationic pLGICs, GLIC and 5-HT3AR, immediately upon application. DHA has no effect on a non-desensitizing GLIC double mutant confirming that DHA enhances agonist induced desensitization. We crystallized GLIC with DHA and found that
DHA binds at the channel periphery near transmembrane helix 4 (M4) and is stabilized by an Arg118 residue from the Cys-loop. The effect of DHA was reduced when we mutated the Arg to Ala thus validating the DHA binding site. Notably, the
vi upper half of the channel pore was wide open and the lower half was constricted resulting in loss of water and ions, thus stabilizing a lipid-induced potential desensitized state. Conversely, cannabinoids mediate their effect after pre- application suggesting that they likely diffuse into the membrane before affecting channel activity. A 3-D reconstruction of the 5-HT3AR in complex with ∆-9-
Tetrahydrocannabinol (THC) and serotonin (5-HT) shows the extracellular and transmembrane domains in a potentially activated conformation and a partially occluded intracellular domain at the lateral portals. This preliminary structure provides a glimpse into the potential mechanism underlying cannabinoid modulation of the 5-HT3AR.
To better understand orthosteric inhibition we show the cryo-electron microscopy structure of the 5-HT3AR receptor in complex with granisetron displaying the precise binding orientation of the ligand, the interacting residues and the conformational changes that stabilize a non-conducting conformation. We highlight the differences between the resting non-conducting and inhibited non- conducting conformations. Collectively, these results begin to shed light into the structural basis for the modulation of pLGICs by lipids, cannabinoids and antagonists.
vii
Chapter 1
Introduction
1
1.1 Pentameric ligand-gated ion channel (pLGIC) superfamily
Intercellular communication in the mammalian central and peripheral nervous systems is mediated by membrane protein complexes which convert neurotransmitters released from the presynaptic end to an ion flux in the postsynaptic end resulting in neuronal excitability, synaptic plasticity, and muscle contraction. These complexes are diverse ionotropic channels that are allosterically regulated. The superfamily of pentameric ligand-gated ion channels
(pLGICs) was first theorized by Langley who observed “receptive substances” on striatal tissue and ganglionic cells that were responsive to nicotine and curare [1].
Castillo and Katz later showed that the receptors for acetylcholine were located on the external surface of the muscle facing the nerve endings [2]. Isolation of the muscle-type nicotinic acetylcholine receptors from the electric organ of the
Torpedo ray led to the cloning of mammalian nAChRs expressed in muscles and nerve cells [3, 4].
To date, at least 40 vertebrate pLGIC subunits have been identified and classified based on the activating neurotransmitter: the excitatory cation selective nicotinic acetylcholine receptor (nAChR) and serotonin receptor (5-HT3R) and the inhibitory anion selective γ-aminobutyric acid receptor type A (GABAAR) and glycine receptor (GlyR) [5]. pLGIC homologs have also been identified in invertebrates including the glutamate-gated chloride channel (GluCl) in C. elegans, mollusks and crustaceans [5]. Prokaryotic homologs were discovered with two exemplars having significantly aided the field in understanding the structure and modulation of pLGICs [6]. GLIC (Gloeobacter violaceus Ligand-gated Ion Channel)
2 a non-selective cation channel that is gated by protons [7, 8] and ELIC (Erwinia chrysanthemi Ligand-gated Ion Channel) also a non-selective cation channel gated by primary amines including GABA [9, 10], have been presumed to maintain pH and ion concentrations in the bacterial periplasmic space [5].
Prokaryotic and eukaryotic pLGICs have ~20% sequence identity reflecting the distance in their phylogeny [11]. Nevertheless, the overall architecture and topology of the eukaryotic and homologous members of the pLGIC family is conserved. In addition, similarities in gating and modulation make the prokaryotic counterparts invaluable tools for understanding pLGIC function. Interestingly, while the finding of pLGICs began and was assumed to be present only in metazoans, phylogenetic analyses suggest a bacterial source for the animal lineage [6].
3 1.2 pLGIC architecture
Information about the shape, dimensions and arrangement of a pLGIC came from electron microscopy images of the nicotinic acetylcholine receptor from the Torpedo ray at low to medium resolutions [12, 13]. Interestingly, a soluble protein from the snail Lymnaea stagnalis aligns with the ECD of pLGICs and forms a stable pentamer without the TMD and ICD [14]. This protein, the acetylcholine binding protein (AChBP), contains conserved residues present in the nAChR and is modulated by nAChR agonists and antagonists. Thus, details about the ligand recognition and conformational changes in the ECD were initially obtained from high resolution crystal structures of the AChBP bound to a number of modulators [15]. Recent high resolution structures of the ECD of pLGICs bound to neurotransmitters show similarities to the AChBP [16-19].
pLGICs consist of five identical (homomers) or homologous subunits
(heteromers) assembled around a central ion conducting pore (Figure 1.1). Each subunit has three domains; the neurotransmitter (ligand) binding extracellular domain (ECD) which is a target for therapeutics, the ion conducting transmembrane domain (TMD) which binds alcohols, bioactive lipids, steroids and anesthetics and the intracellular domain (ICD) which is involved in regulating singe-channel conductance through lateral portals in some channels, sorting, and trafficking to the plasma membrane [13, 20]. Each receptor is about 160 Å long and 80 Å wide and is perpendicular to the membrane.
The N-terminal ECD is composed of a 10 beta-strand (β1 – β10) sandwich core with six inner sheets (β1, β2, β3, β5, β6, β8 ) and four outer sheets (β4, β7,
4 β9, β10) preceded by loops named in the same manner. Neurotransmitters bind at the interface of two adjacent subunits; the principal (+) face from one subunit and the complementary (-) face from the adjacent subunit. The binding site is referred to as having an “aromatic cage” because it is characterized loops containing predominantly aromatic residues which have been shown experimentally to interact with neurotransmitters [13, 14]. The principal subunit contributes loop A (β4-β5 loop), loop B (β7-β8 loop) and loop C (β9-β10 loop) while the complementary subunit contributes three β strands (‘loops D-F’) [21, 22].
Importantly, the signature disulfide bridge separated by 13 amino acids is present in the β6 – β7 loop (Cys-loop) of eukaryotic ECDs but not prokaryotes [11, 20, 23].
The ECD contains N-glycosylation sites which have been shown to be important for expression of functional receptors [24-28].
The ECD and TMD are connected by intersubunit loops which are critical for conveying changes from the ECD to the TMD and ICD [21, 22, 29]. The ECD-
TMD interface is lined by β1-β2, β6-β7, and β8-β9 loops in the ECD and the pre-
M1, M2-M3 linker and the C-terminal end of M4 in the TMD. Perturbations to these segments have been shown to affect EC50 in pLGICs [30]. The TMD from each subunit is composed of four membrane spanning α-helices (M1-M4) for a total of
20 helices in one receptor. M2 helices from each subunit surround the ion conducting pore which also functions as the channel gate. This gate has been investigated extensively and shown to coincide with a hydrophobic set of residues from the middle to the lower part of the channel. The hydrophobic girdle in this region creates an energetic barrier that impedes passage of a hydrated ion [31].
5 Slightly further down the M2 helices are residues that determine the ion selectivity for pLGICs and mutations to these residues have been shown to reverse charge selectivity [31-34]. M1 and M3 helices form a ring around the M2 helices and M4 helices make up the outermost ring that interacts extensively with the surrounding lipid environment. At the extracellular end, the transmembrane helices are splayed apart but gather closer together towards the intracellular side. Intra- and inter- subunit pockets in the TMD occupied by lipids can be displaced by lipophilic allosteric modulators [20].
The ICD is the extended regulatory loop/ domain (MA helical bundle) positioned between M3 and M4 helices. Prokaryotic channels intrinsically do not have an extended ICD; their M3 and M4 helices are connected by short loops [7,
35]. In eukaryotic cation channels, the end of the M3 helix is connected to a short
α-helix (MX helix) via the post-M3 loop. MX lies parallel to the plane of the membrane and forms a belt around the MA helical bundle. MX and MA helices are connected by the MA loop which in many structures is unresolved due to its flexible and unstructured nature. MA helices extend upward from the ICD and become M4 helices once they enter the membrane. Potential intra-subunit interactions holding the helical bundles in place have been suggested from recent pLGIC structures.
Notably, lateral conduits for ion passage between MA helices were shown in the first nAChR and later in the 5-HT3AR structures [13, 18, 36]. Available structures of anionic channels do not show an extended ICD like the cationic channels. It is evident that in the full-length anionic structures, the ICD is composed of short unstructured loops [19, 37] (Chakrapani Lab, unpublished data). It is unclear from
6 these structures whether there is an MX helix present and whether there are lateral portals for ions to exit the channel. Cytosolic proteins can modulate receptor targeting, assembly and trafficking by binding to the ICD such as gephyrin binding to GlyR, rapsyn to nAChR and Ric3 binding to 5-HT3AR. In addition, phosphorylation of residues in this domain is linked to changes in desensitization rates and conductance [20].
1.3 Gating in pLGIC
Ligand binding in the ECD results in conformational changes transduced through a distance of ~50 Å through the ECD-TMD interface to the TMD and ICD leading to ion conductance [38]. Gating is a reversible process in which a receptor in the closed/ resting state (R) shifts to the open/activated state (R*) allowing ions to go through [29]. The minimal gating cycle in pLGICs involves reversible transitions between at least three discrete conformations namely the resting (non- conducting, low agonist affinity), open (conducting, high agonist affinity) and desensitized (non-conducting, highest agonist affinity) states. Gating in pLGICs continues to be an active area of investigation mainly because current structures
‘trapped’ in specific conformations have not been unambiguously assigned.
Channels can open in the absence of activating ligands but at a strikingly low probability [39]. Therefore, activation by an agonist increases the probability that the channel will open. Two agonist molecules are required to fully activate pLGICs thereby increasing the gating equilibrium constant because of the greater
7 agonist affinity for the open state compared to the closed state. Hence, not surprisingly, the structure of the activated conformation is likely markedly different from the resting conformation [29]. The ligand in the binding site interacts with aromatic residues from adjacent subunits as well as hydrophobic contacts that contribute to ligand affinity [40]. In the absence of a ligand, loop C is positioned in an extended conformation and when the receptor is bound to an agonist, it moves to a “closed” conformation capping binding site. Thus, the conformation of loop C is associated with the occupancy and functional state of the channel [15]. Agonist bound structures of the AChBP show that the signal transduction pathway begins at the binding site where the inward motion of loop C brings a conserved Tyr residue into an interaction with a conserved Lys resulting in inter-residue translations that could be transmitted to the channel through the β6-β7 loop and
β10 loop (pre-M1 region) [29]. Residues in the other ligand binding loops also shift to stabilize the agonist in the binding site. Propagation of these local changes leads to a repacking of the β-sandwich core with the outer β sheets curving in towards the central pore axis [41].
The ECD-TMD interface is composed entirely of loops with β1-β2, β6-β7 and β8-β9 loops from the ECD and the pre-M1 link and M2-M3 loop from the TMD.
In the nAChR, residues from the β1-β2 loop form multiple contacts with a conserved proline in the M2-M3 linker while the β6-β7 loop inserts between the pre-M1 (β10 link to M1) and the M2-M3 linker and interacts with residues from both loops [41]. In the ligand bound state, loop C extends inward with an overall counterclockwise twist of the ECD resulting in a concerted displacement of the β1-
8 β2 and β6-β7 loops which causes a notable outward movement of the M2-M3 linker away from the central pore axis. This movement of the M2-M3 linker pulls the tops of the M2 and M3 helices in the same outward direction thereby expanding the top of the pore [21, 42]. In addition, interactions between the M2-M3 linker and the β1-β2, β6-β7 loops and pre-M1 are reduced.
Ion conduction in pLGICs principally involves the movement of the M2 helices. The extracellular end of the pore of anionic pLGICs is lined by positively charged residues (Arg19′) while in cation channels the pore is occupied by
Asp/Glu20′. The prime numbering (X′) nomenclature for the residues facing the pore in the M2 helices is from the 20′ residue at the top of M2 to the -4′ residue at the bottom of the M2 [43]. Two rings of hydrophobic residues follow in anionic channels at Ile/Gly 16′/17′ and Leu9′. Polar sidechains (Ser/Thr) line the anionic channels in between the hydrophobic residues. Similarly cation channels have hydrophobic girdles at Ile17′, Val13′, and Leu9′ and polar residues below at Thr6′,
Ser2′. The intracellular end of anionic channels generally contains a net positive charge (Arg0′) while cationic channels have a negative charge (Glu-1′, Asp-4′).
These charged residues are implicated in ion selectivity of pLGICs [34, 44]. In the resting state of cationic pLGICs, several constriction points below 3 Å radii along the pore are evident at 9′, 6′ 2′ and -1′. Likewise, anionic pLGICs show constrictions at 9′, 2′ and -2′. The activation ‘gate’ in pLGICs is considered to be at the conserved
Leu/Ile9′ while the desensitization gate is considered to be at the intracellular end of the pore [16, 17, 25, 45-47].
9 During desensitization, the agonist is still bound to the receptor but the channel is refractory to ion conduction. Questions as to whether the resting closed state and the desensitized non-conducting conformation were structurally distinct remained ambiguous [38]. The Monod-Wyman-Changeux (MWC) structural model posits that allosteric proteins exist in equilibrium in the absence of agonist between at least two global conformational states namely a resting and an active state and that binding of an agonist shifts the equilibrium of the population of conformers towards the stable highest agonist affinity state [48]. Several high resolution structures of pLGICs solved in the presence of saturating agonist concentrations are expected to be predominantly in the stable desensitized conformation and these structures have begun to provide insight into the potential differences between the resting and desensitized conformations [16-19, 25, 27, 49]. In general, gating kinetics in pLGICs are regulated by subunit composition, which underlie intrinsic mechanistic differences between anionic and cationic channels, and by membrane lipids [38, 41, 50, 51].
1.4. Nicotinic acetylcholine receptors (nAChRs)
The nAChR has a rich history from its seminal description as the ‘receptive substance’ at the nerve terminal [1], to pivotal functional studies of the neuromuscular junction [2, 52], to the identification of compounds that can bind to the nAChR [53], to cloning and isolation [4], and finally imaging of the receptor from low to relatively high resolutions [12, 13, 25, 54-56]. Thus, the nAChR is considered the symbolic head of the pLGIC superfamily [57].
10 Subunit composition and localization
Muscle nAChRs in the Torpedo electric organ and fetal vertebrates are composed of α1, β1, γ, δ subunits in a 2:1:1:1 stoichiometry. The pentamer assembles in a constrained order going in the clockwise direction: α1, γ, α1, β, δ.
The ε subunit replaces the γ subunit in the mature synapse in the same stoichiometry [58, 59]. In neuronal nAChRs, the α subunit comprises of the following subtypes: α2-α7, α9, and α10 while the β subunit has β2-β4 subtypes.
The α subunits are characterized by two adjacent cysteines which are absent in the β subunits [60]. Neuronal nAChRs can assemble as functional homomeric or heteromeric receptors and the varied subunit combinations confer specific structural and functional tunability in these receptors. The snake toxin α-
Bungarotoxin (α-Bgtx) was used to characterize the muscle nAChR which led to further classification of neuronal receptors into α-Bgtx sensitive receptors (α7, α9,
α10) and α-Bgtx insensitive receptors (α2-α6 and β2-β4) [60].
nAChRs are present in high concentrations in the postsynaptic membrane of the neuromuscular junction in the stringent (α1)2, γ, α1, β, ε stoichiometric assembly. They are also expressed in the mammalian brain and peripheral ganglia. Non-neuronal cells such as epithelial, endothelial and immune cells also show expression of neuronal nAChRs where they modulate cell proliferation and differentiation. In bronchial epithelial cells, nicotine sensitive AChRs are implicated in mediating the toxic effects of smoking [59, 61]. The predominant neuronal nAChRs in the brain are the high nicotine affinity α4β2 and the α-Bgtx sensitive α7 receptors. α7-α9 can form functional homopentamers or heteropentamers (α7/α8,
11 α9/α10). α2-α6 and β2-β4 form functional heteropentamers typically in (αx)2(βy)3 subunit combination. α5 and β3 subunits cannot form functional homo- or hetero- pentamers alone or in combination with each other. They are operational only when co-expressed with the functional receptors. Thus, although these subunits are auxiliary, they do contribute to the function of the receptor [59, 62].
Structure and function
The nAChR is composed of five identical or homologous subunits arranged around a central pore (Figure 1.2). The ~290 kDa receptors are assembled perpendicularly to the membrane. nAChRs are non-selective cation channels that mediate excitatory neurotransmission. The N-terminus ECD contains at least 2 to
5 acetylcholine (ACh) binding sites functionally connected to the channel pore. The principal (+) side (loops A-C) of the neurotransmitter binding site comes from the
α (α1, α2, α3, α4, α6, α7, or α9) subunits where a conserved cys-cys pair in loop
C is present in the α subunit of nAChRs, while the complementary (-) side (loops
D-F) is from the β (β2 or β4) or (α10, δ, γ, or ε) subunits. The auxiliary subunits (α5 and β3) do not contribute to ligand binding [57, 63]. In most cases, heteromeric receptors present two ACh binding sites while homomeric receptors present five potential binding sites [62]. Recent high resolution structures of the nAChR in the presence of nicotine showed nicotine bound at the α-β interface (high sensitivity) in the (α4)2(β2)3 and additional α-α interface (low sensitivity) in the (α4)3(β2)2 receptor combinations [25, 56].
The AChBP provided the first high resolution insight into the ligand binding domain of pLGICs. Its close homology to the α- subunits of the nAChRs and its
12 sensitivity to known agonists and antagonists made it a suitable model to investigate ligand binding in the pLGIC family. The mode of ligand binding at the interface of two subunits and ligand interaction with conserved aromatic residues was found to be similar in nAChRs. Agonists contract the ligand binding pocket with the closure of loop C while antagonists bind with loop C in an extended conformation [14, 15].
Muscle nAChRs are located at post synaptic membranes while neuronal receptors are located peri-, pre- and post- synaptically. At the neuromuscular junction, the location of the muscle nAChRs is necessary for the depolarization of the endplate resulting in muscle contraction. In the peripheral nervous system
(PNS), nAChRs are predominantly located post synaptically for mediating fast excitatory neurotransmission. In the brain, these receptors are located peri- and pre- synaptically for the purpose of modulating neurotransmitter release
(serotonin, dopamine, glutamate, norepinephrine, and acetylcholine) in the synapse. It is possible that some neuronal nAChRs mediate fast neurotransmission in the central nervous system (CNS) but their physiological importance is yet to be uncovered [59].
As alluded to in the gating section, the subunits in the pentameric assembly of pLGICs contribute to gating kinetics and selectivity and specific subunit composition regulates the pharmacological features of the ligand binding sites as well as sites of expression [63]. It is now known that the predominant neuronal
α4β2 receptor with two ACh binding sites [25, 56] is upregulated during nicotine exposure. Because heterologous expression stably presents both (α4)2(β2)3 or
13 (α4)3(β2)2 subunit compositions, it is speculated that both assemblies are present in the mammalian brain and that the (α4)2(β2)3 composition is upregulated in nicotine addiction. The two stoichiometries are distinct in their functional
2+ properties: the (α4)2(β2)3 has lower Ca permeability and single channel conductance and higher nicotine affinity compared to the (α4)3(β2)2 [64-66].
Lipids play a critical role in the function of ion channels [67] but the mechanisms by which they modulate channels are still unclear. One postulation is that membrane lipids maintain the equilibrium of the conformers with a higher probability of selecting for the resting conformation in the absence of ligands.
Attempts to isolate the nAChR from membranes led to the realization that exogenous lipids are necessary for preserving functionality [51]. Lipid compositions lacking cholesterol and anionic lipids have been shown to stabilize an uncoupled conformation of the nAChR whereby the agonist binds the receptor
(low affinity) but fails to transduce conformational changes that result in channel gating [41]. Molecular dynamics (MD) simulations of the nAChR show putative cholesterol binding sites in both non-annular and annular sites suggesting that cholesterol stabilizes the structure of the channel and supports contact between the ligand binding domain and the pore [68]. The recent α4β2 receptor structure showed cholesterol bound at intrasubunit interfaces towards the bottom of the channel [56]. Simulation studies have also shown similar cholesterol binding sites in the GABAAR [69]. Free fatty acids and anesthetics have also been shown to modulate nAChRs [70, 71].
14 Physiology and pathophysiology in nAChRs
Various physiological processes are mediated by nAChRs in the central and peripheral nervous systems and in non-neuronal cells. In the brain presynaptic nAChRs modulate the release of neurotransmitters by an increase in intracellular calcium or depolarization of the presynaptic membrane. Peripherally, they facilitate fast excitatory neurotransmission in their target cells while in non-excitable cells
(leukocytes, kidney, lungs, skin) their function in cell differentiation, proliferation and implications in side effects of compounds that act on nAChRs is still an active area of research [72]. Neuronal nAChRs also play a role in anxiety, voluntary motion, memory and attention, reward and pain [73].
Muscle-type nAChRs are clinical targets for two main classes of inherited and acquired diseases that result in defective neuromuscular transmission and muscular fatigue/weakness. Myasthenia gravis (MG) is an autoimmune disease characterized by circulating antibodies against muscle nAChRs. MG is non- inherited and can be localized to one muscle group or spread to several.
Congenital myasthenic syndromes (CMS) are heterogeneous diseases where mutations in muscle nAChRs and synaptic proteins lead to impaired neuromuscular transmission [74, 75].
Neuronal nAChRs are implicated in neuropathologies including Alzheimer’s
(loss of cholinergic transmission where nicotine therapy is considered to increase cholinergic function), Schizophrenia (where higher prevalence of mutations have been reported in α7 and α4β2 receptors in these patients), Parkinson’s disease, hereditary epilepsies (most mutations in autosomal dominant nocturnal frontal lobe
15 epilepsy are in the α4 subunit), autism and most widespread, nicotine addiction.
Smoking is a public health concern and the effects of nicotine addiction lead to different cancers as well as increasing addictive tendencies to other drugs [59].
The complexity of the nAChR system presents a challenging opportunity to find new therapeutic approaches for the treatment of the aforementioned and many other diseases and disorders relating to nAChRs.
1.5 Serotonin receptors (5-HT3R)
Serotonin (5-hydroxytryptamine, (5-HT)) was first described in 1937 and later named in 1948 after its vasoconstrictive activity [76, 77]. Studies in the 1950s led to the finding that there were two types of 5-HT receptors in the guinea-pig ileum: M receptors that could be inhibited by morphine and were probably present in the nervous system and D receptors that could be blocked by dibenzyline likely located in muscles [78]. Since then, knowledge of the serotonergic system has increased to reveal a complex system of signaling mediated by different receptor subtypes. There are seven 5-HTR subtypes namely 5-HT1 to 5-HT7. These receptors with the exception of 5-HT3R are G-protein coupled receptors. 5-HT3R are ligand gated ion channels of the pLGIC superfamily [79]. Radioligand binding studies led to the identification of 5-HT3R in the CNS and functional studies confirmed that these receptors were ligand-gated ion channels [80, 81]. The first serotonin receptor was cloned and expressed to reveal features consistent with excitatory ligand-gated ion channels [82].
16 Subunit composition and localization
Five subunits namely A-E have been identified in the 5-HT3R subfamily.
Subunit A was the first to be identified followed by subunit B a few years later [83,
84]. Genes encoding subunits C-E were identified [85] with further work describing different isoforms and splice variants of these subunits [77]. Interestingly, C-E subunits are present in most mammalian species but absent in rodents. In addition, phylogenetic analysis of 5HT3R sequences (A-E) in different mammals suggests that subunits C-E are more closely related and have evolved distinct features from the more distant A and B subunits [86]. Expression of individual genes (A, B, and
C) show localization in many tissues including the brain, stomach, intestine, colon, lung, muscle, and kidney. 5-HT3D shows expression in the colon, liver and kidney only, while 5-HT3E is present in the intestine and colon only [85].
Structure and function
5-HT3 receptors assemble as homopentamers of the A subunit alone or heteropentamers of the A subunit in combination with either the B, C, D or E subunit [87]. The subunits are symmetrically arranged around a central pore. 5HT3 receptors mediate fast excitatory neurotransmission in the central and peripheral nervous systems. They are non-selective cation channels. Each subunit has an N- terminus ECD, a membrane spanning TMD composed of 4 helices each and an
ICD between TM helices 3 and 4. The ligand binding site at the intersubunit- interface of the ECD is formed by loops from the principal (+) subunit (loops A-C) and the complementary (-) subunit (loops D-F). A conserved ‘aromatic cage’ is
17 created from residues in the binding pocket that interact with the amine and indole rings of agonists and antagonists [18, 49]. N-glycosylation sites in the ECD have been shown to be critical for receptor expression and assembly [24].
The M2 helices line the pore of the channel with the activation gate consisting of a conserved Leu9′ located in the middle of the pore. In the resting and inhibited states, the pore is constricted at this position as well as the bottom of the pore. In the activated state, the helices splay open allowing hydrated ions to pass through. The ICD is forms a helical bundle termed the MA stretch between helices M3 and M4. In 5-HT3R the ICD is a regulatory region that can be phosphorylated and thus affect desensitization [88], interact with chaperones
(RIC3), and affect single channel conductance [77, 89]. Interestingly, lateral portals in the ICD for ion conduction initially predicted in the nAChR, were evident in recent high resolution structures of the 5-HT3AR in serotonin-bound activated conformations [18, 49].
Investigation into whether subunits B-E form functional homopentamers showed that while there was some expression on membranes, western blot results for these subunits showed multiple bands at lower molecular weights than the expected ~50 kDa. In addition, functional studies of each individually expressed subunit showed currents only from the A receptor. B-E subunits did not gate even at 100 µM 5-HT indicating that they do not form functional homomeric receptors.
Heterologous expression of subunit A and B showed robustly activating 5-HT gated currents similar to the homomeric 5-HT3AR. AB receptor currents activated significantly faster than A receptor currents [86]. Further characterization of the
18 effect of the B subunit indicated that the effect of picrotoxin block was minimal in the AB receptor compared to A alone suggesting that picrotoxin preferentially binds the A subunit [90]. Single channel conductance of the A receptor is <1 picosiemens
(pS) while in the AB receptor it is ~16-30 pS owing to the 3 arginine residues present in the MA stretch of A subunit which are missing in the B subunit (Q, D,
A). The functional relevance of this difference in single channel conductance is still unclear. Replacing the arginines in A with the residues in B increased the single channel conductance of the A receptor similar to the AB receptor [91]. The AB receptor has lower Ca2+ permeability but faster activation and desensitization kinetics compared to the homomeric A receptor [84].
Co-expression of subunit A and C produced functional channels with lower currents compared to subunit A alone. Co-expression of A and D resulted in functional receptors whose activating currents and inhibition with picrotoxin were very similar to homomeric A receptors suggesting that the D subunit may not have an impact on the function of the A subunit. A and E subunits also formed functional channels that had slightly reduced peak currents compared to A alone but this was not conclusive to determine an apparent effect of subunit E. Interestingly, sequence alignments of the A subunit with C-E showed that the 3 arginine residues in the MA stretch that determine conductance in the A receptors are replaced by different residues (G, W, X in subunit C and E, Q, E in subunits D and E) in the other subunits. Because these heteromeric receptors are all functional, questions arise as to what their contributions are in the tissues that express them [86].
19 Of all the members of the pLGIC family, only the 5-HT3AR subfamily has high resolution structures in the resting/closed, activated, pre-open/desensitized and inhibited conformations. Comparisons between the resting and activated conformations reveal closure of loop C capping serotonin in the binding site at the
ECD resulting in global changes with the ECD twisting in the counterclockwise direction and the TMD rotating in the clockwise direction leading to channel opening. These motions are in line with earlier predictions and studies with prokaryotic members of the family [92]. The first structure of the 5-HT3AR was solved by X-ray crystallography in the presence of nanobodies which stabilized an inhibited conformation (Figure 1.3A). Comparisons between this inhibited state and the later solved resting state in the absence of ligands or nanobodies by cryo-
EM, showed an expanded overall conformation of the resting state with an extended (outward) conformation of loop C relative to the inhibited state. These results suggest that the inhibited state is along the activation pathway relative to the resting conformation [26, 36]. A recent cryo-EM 5-HT3AR structure bound to the antagonist tropisetron displayed similar features to the inhibited crystal structure and is also along the activation pathway but with an energetic barrier to reaching full activation. Chapter 3 of this work discusses the details of 5-HT3AR bound to the clinically approved antagonist, granisetron, providing insights into the differences between the resting state and how two structurally similar antagonists have differing binding modes.
The pre-open/desensitized conformation was solved in the presence of serotonin and showed features of the activated conformation in the ECD with the
20 counterclockwise twist but the pore of the channel remained closed [18]. It is important to note that all the 5-HT3AR structures solved thus far were all in a detergent environment. Because membrane lipids play a critical role in the function of ion channels, structures solved in a physiologically environment such as nanodiscs may shed light on and help to distinguish between states that appear similar although they are physiologically distinct.
5-HT3R are also modulated by allosteric compounds including alcohols, steroids and cannabinoids. Ethanol enhances 5-HT3AR currents at low agonist concentrations but has no effect on 5-HT3ABR currents. Suggested mechanisms by which ethanol potentiates 5-HT3AR currents include slowing desensitization, increasing the rate of activation and slowing the unbinding of 5-HT [93, 94]. Short chain alcohols (ethanol, butanol) potentiate 5-HT3AR currents while long chain alcohols (hexanol, octanol) inhibit currents. Local anesthetics and antipsychotics have been reported to allosterically and competitively modulate 5-HT3R function.
Natural and synthetic cannabinoids exhibit analgesic properties and effectively repress nausea and vomiting through different mechanisms including inhibiting 5-
HT3R [95, 96]. Cannabinoid modulation of 5-HT3AR function is discussed more in
Chapter 3 of this work. In addition, naturally occurring compounds such as ginseng and ginger modulate 5-HT3Rs [77].
Physiology and pathophysiology
5-HT3R are located pre- and post- synaptically where they mediate membrane depolarization, release of neurotransmitters, neuronal excitation and
21 release of serotonin from the enterochromaffin cells of the small intestine [84]. 5-
HT3R are widely expressed in the central nervous system with higher levels in the area postrema, nucleus tractus solitarii, nucleus accumbens, nucleus dorsalis nervi vagi, nucleus caudatus, cingulate cortex, and hippocampus. These regions control in addition to other processes, the vomiting reflex, anxiety control, the reward system (dopamine mesolimbic system), and pain processing. In the peripheral nervous system 5-HT3R expression is evident in sensory, enteric and autonomic neurons thus regulating sensory transmission and autonomic functions
[97]. In general, 5-HT3R are involved in information transfer in the brain and gastrointestinal tract thereby regulating gut motility, emesis reflex, and visceral pain perception [98, 99]. The bulk of serotonin in the human body can be released peripherally from the enterochromaffin cells of the intestine through multiple mechanisms modulating the physiological function of the GI tract [100].
Chemotherapeutic agents have been shown to induce emesis. The definitive mechanism and pathway by which the vomiting reflex is activated has been the subject of many studies. However, it became clear that 5-HT and its receptor play major roles and that inhibiting this reflex produced therapeutic effects
[101]. Setrons are 5-HT3R receptor antagonists approved for use in the early
1990s to prevent chemotherapy-induced nausea and vomiting (CINV) [102]. The release of 5-HT triggered by chemotherapy agents [100] stimulates peripheral 5-
HT3R located on the afferent vagal neurons and together with local 5-HT release in the area postrema results in nausea and vomiting. First generation setrons include ondansetron, dolasetron, granisetron, and tropisetron and have high
22 antiemetic efficacy in the acute phase. However, delayed and refractory emesis continue to present significant challenges for both patients and clinicians.
Palonosetron is the only second generation setron available and it has shown remarkable efficacy in delayed emesis and some improvement for patients with refractory emesis [103]. Palonosetron has a higher binding affinity compared to first generation setrons through additional mechanisms and a longer half-life [103,
104].
Still, not all patients get relief from CINV and the adverse reactions from some of these drugs necessitates a better understanding of the modulation of 5-
HT3R by antagonists [105]. It is important to note that setrons are not universal antiemetics as they do not prevent vomiting induced by motion sickness or morphine [100]. Dysfunctional receptors are also implicated in psychiatric and gastrointestinal conditions such as bipolar disorder, depression, anxiety, and irritable bowel syndrome [99, 106]. 5-HT3R antagonists also find use in treating chronic pain [93].
While research for more efficacious orthosteric antagonists is essential, strategies targeting allosteric inhibition mechanisms of 5-HT3R could open up an avenue for the discovery of new therapeutics that are subtype specific and do not alter the pattern of endogenous neurotransmitters. The functional relevance of the
B-E subunits of the 5-HT3R family remains unclear despite localized tissue expression for some subunits [98]. Questions remain as to the stoichiometry of these subunits with subunit A and whether heteromeric receptors of subunit and A and two of the other subunits can form functional channels. In addition, the
23 functional contribution of these subunits and their variants in different tissues and in diseases such as IBS, schizophrenia and bipolar disorder remain to be determined [107]. Information on the assembly of heteromeric receptors and their functional significance could lead to selective drugs with fewer off targets.
Heterologous expression and isolation for structural studies may shed light on receptor assembly and possibly draw functional relevance from the structures.
1.6 γ- Aminobutyric acid type A receptors (GABAAR)
GABA receptors are the major inhibitory receptors in the central nervous system. They are gated by the main inhibitory neurotransmitter γ-aminobutyric acid
(GABA). There are two types of GABA receptors; the ionotropic GABAA receptors
(GABAAR) which are chloride channels that mediate fast inhibitory synaptic transmission, and the metabotropic GABAB receptors (GABABR) [108]. GABAAR are in the superfamily of pLGICs while GABABR are G-protein coupled to activating inwardly rectifying potassium channels or inhibiting voltage gated calcium channels involved in cAMP signaling [109]. GABA receptors were pharmacologically identified by their activation by GABA and muscimol and inhibition by picrotoxin and bicuculline. The agonist baclofen was shown to be specific for GABABR and insensitive to muscimol and bicuculline [110-112].
GABAAR are also modulated by clinically relevant compounds such as benzodiazepines, alcohols, general anesthetics and neurosteroids.
24 Subunit composition and localization
To date, 19 genes have been identified that code for various GABAAR subunits: α1- α6, β1- β3, γ1- γ3, δ, ε, θ, π, ρ1-ρ3. These subunits share about 30-
40% sequence identity with the subtypes sharing about 70-80% sequence identity
[109]. GABAAR form functional heteropentamers with a general consensus for the
2α:2β:1γ stoichiometry at synaptic sites. Receptors located at extrasynaptic sites are believed to be composed of (α4)2(β3)2δ [113]. Homopentameric receptors comprising the ρ subunit and predominantly expressed in the retina were reported to be functional [108].
GABA receptor inhibition is classified into two: high amplitude phasic inhibition which occurs at postsynaptic sites and low amplitude tonic inhibition which occurs at extrasynaptic sites. Phasic inhibition is primarily mediated by α1,
α2 or α3 subunits in association with γ2 subunit and is critical for preventing neuronal over-excitation and for generating rhythmic activities (gamma and theta frequency oscillations). Tonic inhibition is mediated by α4, α5 or α6 subunits usually with the δ subunit and is important for increasing a cell’s input conductance such that an excitatory input will be reduced in size and duration.
In tonic inhibition, the leftover GABA that does not bind postsynaptically diffuses away to these extrasynaptic sites activating these receptors [114]. GABAAR are ubiquitous in the brain with the majority of receptors comprising α, β, and γ2 subunits. α2, α3, α5, subunits are expressed in the hippocampus while α4, and α6, are more abundant in the forebrain and cerebellum. The δ subunit is localized at perisynaptic sites in the forebrain partnered with the α4 or α6 subunits.
25 Defining precise subunit stoichiometries in various tissues is still a challenge but it is clear that subunit composition and assembly determines pharmacological responses, receptor localization, function and kinetics. For example, receptors containing the γ2 subunit, a β subunit and one α subunit (α1, 2, 3, or 5) will have high sensitivity to traditional benzodiazepines (BZDs) while those containing α4 or α6 subunits do not bind traditional BZDs agonists. The γ1 subunit also has lower sensitivity to BZDs compared to the γ2 subunit. Extrasynaptic receptors with a δ subunit are more resistant to desensitization compared to the γ containing subunits
[108, 114, 115].
Structure and function
GABAAR assemble as heteropentamers of homologous subunits around a central ion conducting pore. The N-terminal extracellular ligand binding domain carries the defining Cys-loop while the transmembrane domain contains the ion channel lined by M2 helices from each subunit. The intracellular loop between M3 and M4 is important for receptor modulation by binding of proteins important for trafficking, anchoring and phosphorylation (Figure 1.3B).
GABA binds at the classical interface between subunits in the ECD. Ligand binding residues are from the principal (+) β subunit and the complementary (-) α subunit. Loop C from the β subunit caps the ligand binding site in the presence of
GABA and adopts an extended conformation in the absence of ligand. Conserved
N-glycans are present at the periphery of each of the subunits [16, 19, 27].
Recently solved GABAAR structures comprising αβγ subunits present an assembly in the order β2 (or β3) –α1– β2 (or β3) –α1 –γ2 in a counter-clockwise direction.
26 Benzodiazepines (BZDs) bind at an allosteric site in the α1 (+) –γ2 (-) interface.
Bicuculline is a competitive antagonist that binds at the GABA orthosteric site.
Similar to the cationic channels, the ligand binding site of GABAAR is comprised of aromatic residues that form an ‘aromatic cage’ around the ligand [19, 27].
Other allosteric modulators such as neurosteroids, cannabinoids, picrotoxin and ethanol bind to the TMD and either potentiate or inhibit GABAAR function [116,
117].
Binding of the agonist results in capping of loop C over the binding site leading to an anticlockwise rotation of the subunit ECDs in an asymmetric manner. Bicuculline binding causes an inward flex of the tip of loop C but similar to the antagonist bound 5-HT3AR, the energy barrier presented by antagonism does not allow further conformational changes in the ECD thus the TMD remains closed. Potentially desensitized conformations of GABAAR show activated ECDs bound to agonists, open activation gate at the 9′ position and constriction at the
-2′ which is in the vicinity of the desensitization gate [19, 27]. In general, features of activation, inhibition and desensitization are conserved in pLGICs.
Physiology and pathophysiology
GABA receptors contribute to a vast number of CNS functions including sensory and motor processing, cognition, emotions, sleep and wakefulness. They are important targets for compounds (BZDs, etomidate, barbiturates) with anti- anxiety, sedative, anti-convulsant, anesthetic, muscle relaxant and analgesic
27 properties. While many of these compounds are for clinical use, they are also commonly used recreationally and in substance abuse. Dysfunction of GABAAR results in many diseases and disorders including epilepsies, sleep and anxiety disorders, neurodevelopmental and neuropsychiatric disorders, sensorimotor processing, substance abuse and problems with learning and memory [118].
Mutations in GABAAR genes have been implicated in idiopathic epilepsy syndromes. These mutations are present in all three domains of the receptor resulting in neuronal hyperexcitability and thus predisposing the individual to seizures [119]. In catamenial epilepsy, upregulation of δ-GABAARs during high levels of circulating progesterone results in increased tonic inhibition and decreased seizure susceptibility and anxiety. Progesterone and other neurosteroids including 5α-pregnan-3α, 21-diol-20-one, (3α, 5α –THDOC) have been shown to have a positive effect on GABAARs [120, 121]. Enhanced GABAAR function by neurosteroids results in naturally occurring analgesic, sedative, anticonvulsant, anesthetic and anxiolytic effects [122].
The main CNS endocannabinoid, 2-arachidonoylglycerol (2-AG), potentiates GABAAR currents at low GABA concentrations and works synergistically with the neurosteroid THDOC [123]. Long chain polyunsaturated fatty acids (PUFA) DHA and AA have been shown to potentiate and enhance desensitization of GABAAR (α1β2γ2s) currents at concentrations below 300 µM
[124]. The use of BZDs in clinical treatment and addiction has been studied extensively [125]. However the molecular basis for the action of BZDs remained unclear until the recently solved structures of the GABAAR in the presence of
28 diazepam and alprazolam and the competitive antagonist for the BZD site [19, 27].
A mechanical understanding of modulation by these widely used drugs could have impactful ramifications on the designing compounds without addictive properties.
Moreover, the complexity of the GABAAR area of research presents expansive opportunities to develop fine-tuned therapeutics that are isoform-specific with fewer side effects.
1.7 Glycine receptors (GlyR)
Glycine is the other major inhibitory neurotransmitter in the adult CNS.
Glycine binds to and activates glycine receptors thus mediating fast inhibitory neurotransmission in the brainstem and spinal cord. Radiolabeled strychnine showed the postsynaptic localization of GlyR and that glycine potently displaces radiolabeled strychnine [126]. GlyR was purified from rat spinal cord using amino- strychnine agarose chromatography isolating three polypeptides of 48 kDA, 58 kDa and 93 kDa. Later studies identified the smaller molecular weight (MW) polypeptides to be two different subunits (α and β) of the GlyR and the larger MW to be an intracellular scaffolding protein, gephyrin [127-129].
Subunit composition and localization
In vertebrates, four genes encoding the GlyRα subunit (α1- α4) and one gene encoding the GlyR β subunit have been identified. The α subunits have greater than 80% sequence identity and they have been shown to form functional homomeric receptors. Less than 50% sequence identity is shared between the α
29 and β subunits and when expressed alone, β subunits do not form functional receptors [130]. GlyR are mainly localized in the dorsal horn of the spinal cord and brainstem with some expression in the mammalian retina. Recently it was shown that α2 and α3 subunits are expressed in the forebrain where they regulate neuronal excitability [131, 132].
Structure and function
GlyR are anion selective members of the pLGIC superfamily. They assemble as homopentamers of the α subunit or heteropentamers of the α and β subunit. Each subunit is composed of an N-terminal ECD, four membrane spanning TMD α-helices (M1-M4) and an unstructured ICD between M3 and M4
(Figure 1.3C). The ECD carries the signature cys-loop and binds ligands at the interface of adjacent subunits. The neurotransmitter binding site contains the characteristic ‘aromatic cage’ from residues on the principal (+) surface and the complementary (-) surface. Similar to other members of the pLGIC family, agonist binding results in closure of loop C while antagonists present a more open or extended conformation of loop C. In the presence of the agonist the TMD undergoes a clockwise rotation that leads to an expansion of the pore lining helices allowing ion conduction. Antagonists stabilize an inhibited conformation that is defined by a constricted pore [17, 133].
The reported GlyR stoichiometric assembly of the heteropentamer is 2α1:3β with a β-α-β-α-β arrangement [134, 135] and the β subunit predominating ligand binding. Available high resolution structures of the GlyR are homopentamers
(human α3 and zebra fish α1) and present ligands bound at all five sites. One study
30 engineered residues specific to the β subunit into the homomeric GlyR guided by the anionic pLGIC from Caenorhabditis elegans, the glutamate gated chloride channel (GluCl) that has a high sequence homology to GlyR. They determined the contributions of the αβ, βα, and ββ interfaces to ligand binding. The results of the study showed that all five interfaces in the heteromeric channel respond to agonist activation with differing affinities suggesting subtype specificity in drug targeting
[136]. Structural studies of heteromeric channels coupled with this functional work would better confirm the role of the additional interfaces in ligand recognition in the different α subunits and potential relevance in tissue localization. The β subunit plays a crucial role in GlyR trafficking and postsynaptic clustering through its intracellular loop that binds to the scaffolding protein gephyrin. The ICD also interacts with other membrane transport proteins including neurobeachin and
Vps35 [130].
GlyR are modulated allosterically by compounds such as general anesthetics, alcohols, ivermectin and specific neurosteroid derivatives by potentiating GlyR Cl- currents. These hydrophobic compounds bind to the TM domain of GlyR to modulate channel function. Interestingly, neurosteroids such as progesterone inhibit glycine-activated currents while potentiating GABA-activated currents. GlyR subunit composition defines the bidirectional effects of neurosteroids and possibly the difference with the GABAAR modulation [17, 130,
137-139]. Endogenous and exogenous cannabinoids also regulate GlyR function bilaterally based on GlyR subunit composition and the charge on the cannabinoids.
Neutral endocannabinoids such as AEA, N-arachidonoyl-dopamine (NADA) and
31 noladin ether (NOLE) potentiate α1-α3 currents while acidic endocannabinoids such as N-arachidonoyl glycine (NA-Gly) and N-arachidonoyl-L-serine (NA-Ser) and the long chain PUFA arachidonic acid (AA) inhibit α2 and α3 and potentiate
α1 receptors [140]. THC has been shown to potentiate GlyR currents with the maximal effect seen with lower glycine concentrations [141, 142].
Physiology and pathophysiology
Binding of glycine allows conduction of Cl- ions into the postsynaptic cell resulting in hyperpolarization that regulates neuronal excitability. Glycine receptors are important for motor and sensory information controlling movement, peripheral and central pain sensitivity, vision and auditory functions [127, 143]. Mutations in the α1 and β genes that affect glycine mediated neurotransmission cause the hereditary disease hyperekplexia. This disease is characterized by an exaggerated startle response to touch and sound. GlyR carrying these mutations have decreased single channel conductance and agonist sensitivity. The α3 gene has been implicated in generalized epilepsy, autism and amyotrophic lateral sclerosis
[130]. In addition, inhibition of GlyR α3 by prostaglandin E2 induces phosphorylation that underlies inflammatory pain sensitization [143]. The differential modulation of GlyR by allosteric modulators makes these receptors an attractive target for therapeutics that are subunit specific.
32 1.8 Prokaryotic receptors
Prokaryotic pLGICs were first identified in bacteria by using the sequences of the eukaryotic pLGIC ligand binding domains to search bacterial genomes.
Significant hits were found in cyanobacteria (Crocosphaera watsonii, Gloeobacter violaceus), α-proteobacteria (Rhodopseudomonas palustris) γ-proteobacteria
(Erwinia chrysanthemi, Methylococcus capsulatus) and one archea
Methanosarcina [6]. The first high resolution pLGIC structures were solved by X- ray crystallography from two prokaryotes ELIC and GLIC (Figure 1.4) with a sequence identity of about ~18% [7, 8, 10]. GLIC and ELIC are both non-selective cation channels gated by hydrogen ions and primary amines respectively [8, 144].
Prokaryotic pLGICs are located on the inner membrane of gram-negative bacteria.
However, their physiological relevance is still unclear but roles in environmental adaptation for ELIC and pH regulation for GLIC have been suggested [5, 145].
More recently, a gammaproteobacterial channel known as sTeLIC was discovered and shown to activate at high pH [146]. Before high resolution structures from eukaryotic pLGICs became available, prokaryotic structures were used as surrogates to understand the fundamentals of pLGIC function because of the conserved core architecture. In fact, GLIC function is altered by pharmacologically relevant compounds in a similar manner to eukaryotic receptors which made it a reliable model for understanding modulation. Conversely, ELIC is modulated by fewer of the same compounds at lower affinities than GLIC, indicating that the ELIC is pharmacologically less similar to GLIC and the eukaryotic receptors [144]. The
33 similarity of GLIC to eukaryotic function made it an attractive model for our experimental work.
Subunit composition and structure
Two GLIC crystal structures in potentially open conformations were solved at 2.9 Å and 3.1 Å at pH 4.6 and pH 4. Initial studies with GLIC showed that at activating pH, the channels do not desensitize. This led to the conclusion that under the crystallization conditions, the channels would likely be in an open conformation [35]. However, later studies showed that GLIC currents desensitize slowly (seconds to minutes timescale) and that the crystal structures likely represent desensitized or intermediate conformations [147, 148]. Interestingly the pore in both structures is wide open at the 9′ position and slightly under 3 Å at the
-2′ position suggesting that it could conduct ions. It has been postulated that the detergent molecules in the GLIC pores stabilize the channel open in what would be a non-conducting pore. A resting conformation of GLIC was also solved at neutral pH [149]. The crystal structure of ELIC was initially solved in the absence of ligands at 3.3 Å and when activating ligands were identified, ELIC was crystallized in the presence of GABA and benzodiazepines [150].
Prokaryotic channels are composed of five identical subunits arranged around a central ion conducting pore. Each subunit has an N-terminus ECD composed of 10 β-strands, a TMD composed of four α-helices (M1-M4) and a short loop connecting M1 and M2, M3 and M4 on the intracellular side. The β-strands of
34 the ECD are arranged into two sheets forming a β-sandwich. Prokaryotes lack the functionally important cysteine disulfide bridge in the β6-β7 loop of the ECD that defines eukaryotic pLGICs. Agonists bind at the interface of two subunits and in both GLIC and ELIC the overall conformation of this interface is conserved with the capping of loop C over the binding site. However, because GLIC is gated by protons, the ligand binding site contains ionizable residues [7]. ELIC is gated by
GABA and modulated by benzodiazepines thus its orthosteric site contains aromatic residues similar to the eukaryotic binding sites [150]. The M2 helices line the ion conducting pore with hydrophobic constrictions at the 9′, 13′ and 16′ positions, polar rings at the 2′ and 6′ positions and negatively charged residues at the selectivity filter. This pattern of residues lining the pore is similar to the eukaryotic cationic channels.
In the resting conformation of GLIC, the subunits at the ECD have somewhat few interactions and appear loosely packed. Thus, the solvent accessible surfaces at the subunit interfaces is large. In the presence of activating low pH, the ECD moves radially towards the central pore and tangentially in an anti-clockwise direction resulting in increased intersubunit interactions. In addition, the transition from pH 7 to pH 4 results in an inward movement of loop C similar to the previously established AChBP with a capped loop C when bound to an agonist.
In the TMD, the middle parts of M2 are bent while the upper parts tilt outward toward M3 helices from the same subunit and M1 helices in the neighboring subunits. The Ile9′ residues which form the activation gate point away from the pore opening the gate. The overall movement of the TMD helices is a clockwise
35 radial twist in the activated conformation relative to the resting conformation. The motions elicited in the transition from the resting to the activated state of the GLIC channel were later shown to be similar in the eukaryotic channels [18, 92].
GLIC function is inhibited by anesthetics at clinical concentrations similar to the nAChR [71, 151]. Crystal structures of GLIC with desflurane and propofol show the ligands bound in the upper portions of the TM domain [152, 153]. Interestingly, ketamine has been shown to bind in the ECD about 10 Å below the orthosteric site. Alcohols modulate GLIC function bidirectionally. The short chain alcohols methanol and ethanol potentiate GLIC currents at concentrations higher than 600 mM. Alcohols with chains longer than ethanol inhibit GLIC currents with increasing potency as chain length increases. A mutation in the M2 helix at the 14′ position markedly increased GLIC sensitivity to alcohol potentiation. A crystal structure of
GLIC F14′A with ethanol showed density for ethanol at the intersubunit interface
[154]. Replacing the bulky phenyl ring with a smaller side chain results in an expanded cavity. The alcohol binding site appears to be distinct from the intrasubunit anesthetic binding site but it is also possible for alcohols to occupy intrasubunit sites but with lower occupancy [152].
While lipid components have been shown to affect the activity of eukaryotic receptors, GLIC function is not dramatically altered by the lack of certain lipids
(discussed further in Chapter 4). GLIC crystal structures show lipid molecules bound to the channel between M3 and M4 and M1 and M4 suggesting that some lipids are strongly bound to the protein and may be necessary for channel stability and function. Cholesterol and long chain polyunsaturated fatty acids have been
36 shown to enhance desensitization of GLIC currents similar to the α7 nAChR [70,
155, 156]. This similarity in modulation allowed us to probe the structural basis for the effect of fatty acids on GLIC function to gain insight on how these lipids likely regulate eukaryotic pLGICs.
1.9 Purpose of this study
The aim of this work was to understanding the structural basis for modulation of pLGICs by lipids, cannabinoids and antagonists. Components of the plasma membrane have been shown to influence channel function [41, 51, 157].
Specifically, ω-3 (DHA and EPA) and ω-6 (AA) fatty acids allosterically enhance or inhibit the function of pLGICs. Endogenous cannabinoids (AEA and 2-AG) derived from arachidonic acid (AA) have also been shown to modulate pLGICs.
Further, plant cannabinoids (THC and CBD) affect pLGIC function similar to the endogenous cannabinoids through cannabinoid-receptor independent mechanisms. However, the mechanisms underlying fatty acid and cannabinoid allosteric modulation of pLGICs remain unclear. Localization of fatty acids in intra- and inter- subunit interfaces suggests that channel function may be altered through direct interaction [158]. Electrophysiology and structural biology techniques (X-ray crystallography and Cryo-EM) allowed us to test the hypothesis that direct lipid- protein interaction is one of the primary ways that channel function is regulated.
The same methods were utilized to test whether cannabinoids can bind to pLGICs
37 to alter their function. Because 5-HT3AR are important targets for CINV, extensive predictions of how antagonists bind to the receptor have previously been made.
We showed the precise binding orientation of a clinically used 5-HT3AR antagonist and conformational changes that result in channel inhibition.
Collectively, this work provides a structural basis for the allosteric modulation of pLGICs by lipidic molecules and orthosteric regulation by antagonists. This insight has the potential to open avenues for the design of a new class of therapeutics that confer subtype selectivity and target allosteric sites. Additionally, designing more efficacious orthosteric antagonists will benefit from the high resolution information presented here.
38 Figure 1.1
Figure 1. 1: A schematic representation of the topology and arrangement in pLGICs.
Each subunit is composed of an N-terminus extracellular domain (ECD), four α-helices (M1-M4) in the transmembrane domain (TMD) and intracellular loops linking the α-helices in the intracellular domain (ICD) (left panel). Five subunits form a pentamer around a central ion conducting pore lined by M2 helices. Individual subunits shown in color and arrow indicates ion flow (right panel).
39
Figure 1.2
Figure 1. 2: nAChR structure from Torpedo membranes.
The left panel shows a side view of the nACh receptor showing the three domains. The grey bars delineate the membrane bilayer. The right panel shows a single subunit. The post-M3 loop, MX helix and MA loop are not evident in this structure. Figure made from PDB ID: 2BG9 [13].
40 Figure 1.3
Figure 1. 3: Representative eukaryotic pLGIC structures.
(A) Crystal structure of mouse 5-HT3AR in the presence of stabilizing nanobodies (VHH15). MX helix and the large MA bundle were well resolved (Figure made from PDB ID: 4PIR)
[26]. (B) Full length human α1β3γ2 GABAAR Cryo-EM structure solved in the absence of ligands. The intracellular domain is flexible and only short loops could be resolved (Figure made from PDB ID: 6I53) [37]. (C) Cryo-EM structure of the zebra fish α1 GlyR bound to the antagonist strychnine. The M3-M4 loop was replaced with the Ala-Gly-Thr tripeptide in this structure (Figure made from PDB ID: 3JAD) [17].
41 Figure 1.4
Figure 1. 4: Representative prokaryotic pLGICs.
(A) GLIC crystal structure solved at pH 4 in an open conformation (Figure made from PDB ID: 3EHZ) [7]. (B) ELIC solved in the absence of any ligands and hence believed to be in a closed conformation (Figure made from PDB ID: 2VL0) [10]. Both structures have a similar architecture to the eukaryotic pLGICs but they lack the disulfide bridge in the ECD and the large intracellular domain. Approximate membrane boundaries indicated by gray bars.
42
Chapter 2
Allosteric modulation of pLGIC function by long chain polyunsaturated fatty acids (PUFAs)
Portions of this chapter were published in: Basak S, Schmandt N, Gicheru Y, Chakrapani S. Crystal structure and dynamics of a lipid-induced potential desensitized state of a pentameric ligand-gated channel. eLife. 2017; 6: e23886. doi: 10.7554/eLife.23886 Structural studies were conducted by S.B and N.S. Functional studies were conducted by Y.G.
43
Introduction The lipid bilayer
The plasma membrane is a physical barrier around a cell that delimits the subcellular compartment from the external environment. The membrane is composed of diverse lipids in a bilayer that regulates signaling information systems. Integral membrane proteins involved in signal transduction are anchored in the lipid bilayer. Eukaryotic cell membranes are primarily composed of glycerophospholipids namely phosphatidylcholine (PtdCho), phosphatidylserine
(PtdSer), phosphatidylethanolamine (PtdEtn), phosphatidylinositol (PI) and phosphatidic acid (PA). Diacylglycerol (DAG) is the hydrophobic backbone of glycerophospholipids and when a phosphate group is attached, it becomes PA.
Esterification at the phosphate group with a choline, serine, ethanolamine, or inositol group makes up the aforementioned glycerophospholipids. DAG carries two fatty acyl chains of varying lengths and saturation (cis-unsaturated).
Sphingolipids (sphingomyelin and glycosphingolipid) are a different class of lipids with a hydrophobic ceramide backbone and saturated or trans-unsaturated fatty acyl chains. The cis fatty acyl chains in glycerophospholipids renders them fluid at room temperature while the trans acyl chains in sphingolipids enables a tighter packing of the lipids adopting a gel-like phase. Sterols like cholesterol preferentially mix with sphingolipids thereby increasing their fluidity. Membrane leaflets tend to be asymmetric with the inner leaflet composed mainly of PtdSer and PtdEtn and the outer leaflet mainly comprising PtdCho and SM [159, 160].
44 Integral membrane proteins are anchored in lipid bilayers and their function is regulated by the composition of the lipids around them. Numerous studies on the nAChR demonstrate the critical role lipids play in protein stability and function
[41, 51, 67]. Experimental data using electron-spin resonance techniques showed that lipids interacting with the surface of the protein are fairly immobile compared to the bulk lipids. In addition, the nAChR preferentially associates with cholesterol, fatty acids and PA compared to other lipids [67]. Other studies established that nAChRs reconstituted in PC membranes alone were non-functional and that cholesterol and anionic lipids were critical for channel activity [41, 161-163].
Fatty acids and their function
Fatty acids (FAs) have also been shown to modulate the function of membrane proteins through direct or indirect means. Free fatty acids (FFA) released from phospholipids are generated and degraded through regulated mechanisms. Direct modulation describes the interaction of the free fatty acid with the protein without the fatty acid undergoing any modifications. Free fatty acid metabolites characterize the indirect modulation [164]. Certain stress conditions such as hypoxia, ischemia or kainite-induced epilepsy in the brain result in decreased intracellular pH. This leads to increased calcium concentrations which result in the activation of phospholipases (A2 and C) that augment intracellular FFA such as linolenic acid which has been shown to have neuroprotective functions
[165].
Essential fatty acids are those that the body cannot make and must be ingested. They are divided into two classes: ω-6 (n-6) and ω-3 (n-3)
45 polyunsaturated fatty acids (PUFAs). The precursor to ω-6 fatty acids is linoleic acid (LA, 18:2, n-6) which undergoes desaturation and elongation to other FAs including arachidonic acid (AA, 20:4). LA is mainly found in vegetable oils such as safflower, soybeans, and corn. α-Linolenic acid (ALA, 18:3, n-3) is the precursor for ω-3 FAs and also undergoes elongation and desaturation to make eicosapentaenoic acid (EPA, 20:5) and docosahexaenoic acid (DHA, 22:6) (see
Figure 2.1). Food sources for n-3 PUFAs are animal products especially fish such as tuna, salmon and trout. The n-6 and n-3 PUFAs share the same series of conversion enzymes but are functionally distinct and not interconvertible. AA,
DHA and EPA are esterified to phospholipids (PtdEtn, PtdCho, PtdSer) and active processes maintain their concentrations in the cytosolic leaflet of membranes. DHA and AA are more abundant than EPA in phospholipid membranes [166-169].
As structural elements of the plasma membrane, PUFAs regulate the elastic properties of the lipid bilayer and thus the function of the membrane proteins anchored therein. PUFAs released from phospholipids and their derivatives can directly modulate receptors with synaptic functions as described previously. PUFAs also exert their effects on gene expression. n-3 PUFAs directly inhibit inflammatory signaling through NFκB and are potent activators of the nuclear transcription factors PPARs which are important for lipid oxidation
[166, 170]. PUFAs play a role in neurogenesis and neuroprotection through
PUFA metabolites- eicosanoids (EPA) and autacoids (DHA). The EPA derived eicosanoids: prostaglandin-3-, leukotriene-5- and thromboxane-3-series have
46 anti-inflammatory, anti-arrhythmic and vasodilatory effects. Resolvins, docosatrienes and neuroprotectins from DHA suppress NFκB, trigger synthesis of anti-apoptotic proteins to protect against neuronal death, and promote survival through BDNF (brain-derived neurotrophic factor) synthesis. Arachidonic acid on the other hand has pro-inflammatory, pro-arrhythmic and vasoconstrictive effects through its eicosanoids: thromboxane-2-, prostaglandin-2-, leukotriene-4-series
[166, 170]. Because of the opposing effects of these PUFAs, diets with a higher ratio of n-3 to n-6 PUFAs are associated with a reduced risk of developing neurodegenerative diseases such as Alzheimer’s with an inflammatory component. Clinical studies suggest that a diet with a 1:1 or 2:1 ratio of n-3 to n-6 may decrease neuroinflammation, decrease cardiovascular risk factors and improve cognition with normal aging [167, 170].
Psychrophilic and piezophilic bacteria have been shown to synthesize long chain PUFAs including DHA and EPA through the anaerobic polyketide pathway.
This pathway is catalyzed by PUFA synthase (encoded by pfa genes) which is an enzyme complex with multiple catalytic domains. Because the fatty acids are produced in marine microorganisms (deep seas, fish, high pressure environments,
Polar Regions), the classic explanation for their synthesis was in the maintenance of membrane fluidity. Other emerging functions of long chain PUFAs are in protecting bacteria against reactive oxygen species. Bacteria with long chain
PUFAs were more resistant to ROS than those lacking them. Studies conducted on marine bacteria that can synthesize EPA (Shewanella livingstonensis Ac10) and EPA-deficient mutants showed that EPA regulates bacterial cell division only
47 at low temperatures suggesting a functional role for EPA. Interestingly, EPA has also been implicated in bacterial resistance against antibiotics. E.coli DH5α heterologously expressing EPA compared to E.coli not expressing EPA were tested for resistance against different classes of antibiotics and the EPA synthesizing bacteria showed higher minimum inhibitory values than the E.coli without EPA. The health benefits of long chain PUFAs on human health have led to the commercialization of producing PUFAs mainly from eukaryotic microorganisms such as microalgae. Genetic engineering of PUFA producing marine bacteria may boost the typically low amounts of PUFAs they produce, thus providing an additional sustainable source of long chain PUFAs [171-174].
Voltage-gated sodium, calcium and potassium channels, and members of the pLGIC family are modulated by cis-unsaturated fatty acids [70, 124, 164, 175].
The sn-2 position of glycerophospholipids typically carries the cis-unsaturated fatty acid which can be released through the actions of phospholipases. The free fatty acids can either directly bind to membrane proteins or act as second messengers in downstream signaling [159]. Arachidonic acid (AA) has been shown potentiate
NMDA receptor currents. AA released from activation of NMDA receptors leads to opening of more NMDA receptors resulting in amplification of calcium concentrations thus inducing long-term potentiation at hippocampal synapses. In what is potentially a positive feedback loop, it is not clear whether AA binds to the receptors directly thus modulating their function or if the changes evident are due to membrane properties [176]. On the other hand, activation of neuronal α7 nAChRs results in the release of AA which strongly inhibits nicotine activated
48 currents in a potential negative feedback loop. Here too, the mechanism of inhibition remains to be understood [177]. In two separate studies, DHA was shown to affect channel function not by direct interaction, but rather by altering the properties of the lipid bilayer [70, 178].
The organization of the TMD in pLGICs places the pore lining M2 helices in the innermost ring. A layer around the M2 helices is made by M3 and M1 helices which securely create a barrier from the outer environment. The outermost M4 helices form the final layer and these helices interact the most with the lipid environment. In fact studies have shown that perturbations in the M4 helix affect gating kinetics, lipid sensitivity and disease-like states [179-182]. Two types of lipid sites have been described based on their location relative to integral membrane proteins: annular and non-annular lipids. Annular lipids comprise the first shell of lipids around the protein and exchange with bulk lipids relatively quickly. Annular lipids do not interact with membrane proteins in a specific manner. Non-annular lipids are located in the spaces between transmembrane helices, display specificity and are considered essential for channel function. It is presumed that the rate of exchange of non-annular lipids is significantly slow owing to the specific interaction of the lipids to the proteins. Endogenously released and exogenously added FFA were shown to localize at both annular and non-annular sites of the nAChR [165].
The importance of long chain PUFAs in maintaining brain and overall body health cannot be overstated. The modulation of synaptic and other membrane proteins by PUFAs has been shown experimentally through different techniques.
49 However, evidence for a PUFA binding site on these channels is still lacking [67,
70, 124, 164, 175-177].
In this work, we attempted to elucidate the structural basis for PUFA action on pLGIC by electrophysiological measurements and X-ray crystallography structural studies. DHA, EPA and AA enhance the desensitization rate of GLIC and
5-HT3AR currents. EPA has an additional effect of strongly reducing peak currents.
We focused first on the effects of DHA on GLIC and hypothesized that DHA (likely
EPA and AA) stabilizes an agonist-induced desensitized state. To confirm this, we generated a non-desensitizing GLIC double mutant. We reasoned that if DHA enhances agonist-induced desensitization, then the effect of DHA on this mutant should be significantly reduced. To take this work further, we crystallized GLIC in the presence of DHA and the notable features of this complex were the DHA binding site at the channel periphery near M4 and the constriction of the pore from the middle to the intracellular end.
Results
DHA modulates the pH-elicited response in GLIC
To test the effect of DHA on GLIC function, we expressed GLIC in Xenopus laevis oocytes and measured currents by two-electrode voltage-clamp (TEVC) techniques (see methods). As previously shown, GLIC is activated by extracellular protons (pH 4.5) and the currents display a slow decay as the channels desensitize
(Figure 2.2A) [7, 35]. When DHA (50 µM) was co-applied with pH 4.5, the
50 macroscopic decay from the peak was accelerated, leading to much smaller steady-state currents (Figure 2.2A, blue arrow). Upon deactivation at pH 7.0, subsequent change to pH 4.5 resulted in currents with peak amplitudes and decay phases indistinguishable from the first pulse, revealing that the effect of DHA was fully reversible. In addition to the effect on current decay, DHA decreases the amount of steady-state current (measured at 2.2 min from the start of application) as shown in the plot of the steady-state-to-peak ratio, suggesting that both the rate and the extent of desensitization are increased (Figure 2.2B).
A detailed analysis of the effect of DHA at different concentrations and at various activating pH is shown in Figure 2.3A. The effect on desensitization was observed at DHA concentrations above 5 µM, and was more pronounced at higher proton concentrations, suggesting that channel activation promotes the effect of
DHA. Additionally, in the presence of DHA, a small left-shift in pH-response is observed for GLIC; pH50 4.87 ± 0.04 in the absence of DHA and pH50 5.02 ± 0.03 in the presence of DHA (Figure 2.3B). These findings are in fact expected for a modulator that promotes desensitization (the conformational state with the highest agonist-affinity). Further, outward current decay was also accelerated in the presence of DHA (Figure 2.3C), similar to the effect on inward currents. Upon pre- application of DHA (at pH 7.4) prior to co-application at pH 4.5, additional decrease in peak amplitudes is observed, suggesting that enhanced availability of DHA could result in further effects on GLIC currents (Figure 2.4). Nevertheless, the predominant effect is still the enhanced desensitization and not stabilization of a closed state. If the closed state were to be stabilized, then the reduction in peak
51 current would have been significantly more. The effects of DHA were fully reversible in all of the measured conditions.
To further confirm that DHA indeed promotes an agonist-induced desensitized state rather than a pre-open resting state, we studied the effect of
DHA on an alanine mutation at the Ile9 position in M2. Mutation at the equivalent position in several pLGIC has been shown to increase agonist sensitivity and slow desensitization [35, 183-185]. The prediction is that perturbations that destabilize the desensitized state should lower the effect of DHA. The I9A mutant exhibits a gain-of-function phenotype [35, 147, 186] resulting in leaky oocytes, an effect that can be offset with a background mutation (H11F) [187, 188] that reduces pH- sensitivity [189].
We found that the double-mutant (I9A /H11F) shows robust non- desensitizing currents with a pH50 5.17 + 0.19 (Figure 2.5A). Quite remarkably,
DHA had no effect on this mutant over a range of pH conditions, and even up to a
100 µM concentration (Figures 2.5B, 2.5C, and 2.6). Since the mutant’s pH response is close to wt, the DHA effect cannot be explained by stabilization of the resting or pre-open states. However, technical limitations of TEVC, that include slower perfusion rates, preclude us from resolving fast kinetic components of desensitization. We therefore, at this point, cannot ascertain which of the multiple desensitized states that DHA stabilizes. Nevertheless, above findings demonstrate that transitions to the desensitized state are necessary for the DHA effect, and that
DHA may stabilize a desensitized conformation induced by the agonist during gating.
52
Structural changes associated with DHA binding in GLIC
To better understand the mechanism of DHA action, co-crystals of GLIC were grown in the presence of 50 µM of DHA at pH 4.5. The crystals diffracted up to 3.25 Å and the structure was solved using GLIC-pH4 (PDB ID: 4HFI) as the starting model [190]. Well-defined electron density was found for DHA in all the subunits in a pocket at the channel periphery; lined by M4, the M2-M3 linker, and the β6-β7 loop (Figure 2.7A, and 2.7B). The DHA molecule appears bent with a twisted/curled orientation and makes a salt-bridge interaction with an arginine sidechain (Arg118) in the β6-β7 loop. Besides this interaction, DHA does not appear to engage with the rest of the protein.
To determine if the interaction of DHA with Arg118 was necessary for the observed effect on channel gating, we probed the functional consequence of mutating Arg118 to Ala and studied the effect of DHA on R118A desensitization.
The R118A mutant showed robust pH-induced currents in oocytes, however in comparison to GLIC wt, the current decay was much less affected by 50 µM DHA
(Figure 2.8).This finding thereby validated the crystallographically-captured DHA- binding site and further implicates a novel role for the Arg118 in lipid-channel interactions.
The most prominent difference noted in the GLIC-pH4-DHA structure in comparison to other GLIC structures (at pH 7.0 and pH 4.0), was at the level of M2 lining the channel pore (Figure 2.9). The conformation of the GLIC-pH4-DHA pore
53 did not align with either the resting (GLIC-pH7) [149] or the putative open state
(GLIC-pH4) [190]. In this new M2 conformation, the pore is funnel-shaped with the hydrophobic extracellular-end (Ala13 to Thr20) wide-open to a pore radius greater than 5 Å, reminiscent of the GLIC-pH4 structure, while the polar intracellular-half
(between Ile9 Thr2) is constricted to 2.5 Å, resembling the GLIC-pH7 closed structure (Figure 2.9A). The reduction in the GLIC-pH4-DHA pore radius was accompanied by notable changes in the occupancy of detergent molecules, water, and ions within the pore. While electron density was observed for the bundle of six dodecyl-maltoside molecules in the upper M2, they were more disordered in comparison to the GLIC-pH4 structures (Figure 2.9B). The reduced detergent occupancy is expected, due to the effect of pore constriction at Ile9. Additionally, the channel pore in the GLIC-pH4 structure (PDB ID: 4HFI) reveals a cation binding site (at the level of Thr2 coordinated by water molecules beneath it) and an ordered pentagonal-ring of water molecules that are in-plane with the γ-O atoms of Ser6 [190] (Figure 2.9B). These densities were also noted in lower-resolution structures (comparable in resolution to GLIC-pH4-DHA), although they appeared more diffuse [7, 8]. In contrast, the GLIC-pH4-DHA structure showed a distinctly different hydration profile at the intracellular half of the pore, with loss of densities for water and ions. Interestingly, the orientation of Ser6 sidechains is implicated in directly influencing the organization of the water ring at this position [190]. A small change at this position in GLIC-pH4-DHA brought about by the compression at Ile9 may thus be responsible for the loss of ions and water. Collectively, these results confirmed that DHA stabilizes an agonist induced desensitized state.
54
AA and EPA modulate the pH-elicited response in GLIC
We were interested in testing the effect of the 20 carbon chain PUFAs (AA,
20:4 and EPA, 20:5) on GLIC function to see if there were differences in modulation based on carbon chain length and number of double bonds. We expressed GLIC in Xenopus laevis oocytes and measured currents by TEVC. When EPA or AA (50
µM) was co-applied with pH 4.5, the macroscopic decay from the peak was accelerated, leading to much smaller steady-state currents (Figure 2.10, blue arrow). Peak current amplitudes in the presence both PUFAs amounted to the steady state current in the previous pulse in the absence of the PUFAs. This is in contrast to peak currents that were similar in the presence or absence of DHA.
Remarkably, the effect of EPA on desensitizing GLIC currents was significantly stronger compared to both DHA and AA and the steady state current much less than the other two PUFAs. Further functional characterization of the effect of EPA and AA at different concentrations and activating pH may provide some insights into the striking differences. It may be possible that the effects of EPA are concentration dependent and that the 50 µM concentration is on the saturating end.
We attempted to co-crystallize GLIC with EPA and AA and the crystals diffracted to below 3.5 Å resolution. GLIC-EPA and GLIC-AA structures were solved in detergent and E.coli lipids at pH 4. Overall, there were no differences in the conformation of these structures to the putative GLIC open structure (PDB ID:
4HFI). Importantly, there was no new continuous density near M4 or elsewhere on
55 the channel that would have corresponded to EPA or AA. The presence of detergent and E. coli lipid densities in the pore and around the protein respectively matched those in GLIC 4HFI. The lack of relevant new densities and structural changes in protein and made it difficult to pursue these structural studies any further.
PUFA effect on 5-HT3AR function
Advances in cryo-electron microscopy (Cryo-EM) enabled us to move on to the eukaryotic members of the pLGIC family. These proteins are about 250 kDa and have a five-fold symmetry which makes them ideal for visualization under cryogenic conditions. Previous studies on the GABAAR and nAChR showed that
DHA enhanced desensitization in both receptors [70, 124]. In addition, AA potentiated currents in the NMDA receptor and strongly inhibited α7 nAChR currents [176, 177]. The varied effects of PUFAs on different systems made us seek to understand their effects on the 5-HT3AR. We wondered whether DHA, EPA and AA would have similar effects as in GLIC and in particular whether EPA would display higher potency. We tested the effect of the three PUFAs on 5-HT3AR function by injecting full length 5-HT3AR cRNA into Xenopus oocytes and measured currents by TEVC.
Co-application of DHA and AA with 1 µM serotonin (5-HT) resulted in peak current reduction compared to peak current in the absence of the PUFAs.
Interestingly there was no difference between the first and third pulses indicating
56 that the effects of DHA and AA were reversible (Figure 2.11). What was not obvious in the DHA and AA currents was the effect on desensitization. We normalized one pulse in the absence and presence of DHA and AA and found that both PUFAs enhance desensitization of 5-HT3AR currents although to a lesser extent compared to GLIC currents. Notably, when EPA was co-applied with 1 µM
5-HT, the peak current was significantly reduced relative to the reduction in peak current with DHA and AA. Previous and subsequent peak currents to the EPA peak currents were similar indicating that the effect of EPA was reversible. Here too, further detailed analysis of different PUFA concentrations and various 5-HT concentrations may shed more light into the apparent differences especially with
EPA. Finally, we also tested the effect of DHA and EPA on the α1 GlyR. The effect of DHA on desensitization was not apparent. EPA application resulted in rapidly decaying currents with significantly less steady state currents (data not shown).
Further characterization and optimization of the GlyR with PUFAs are necessary.
Membrane environment for structural studies
A mechanistic understanding of PUFA modulation of the 5-HT3AR will benefit from Cryo-EM structural studies where the structure of the 5-HT3AR is solved in the presence of each of the PUFAs. There are inherent challenges with understanding membrane protein function in detergent as discussed earlier in studies with the nAChR [41, 51]. Indeed, the TM domain of GLIC crystal structures in the resting and open conformations in detergent are almost identical with the exception of the pore lining M2 helices. This was a surprising finding because
57 ligand binding in the ECD leads to conformational changes in the ECD and the signal is transduced through the ECD-TMD interface leading to overall rearrangement of the TMD resulting in channel opening. We surmised that a detergent environment potentially masks protein dynamics making accurate structural interpretation difficult in the absence of complementary techniques. We optimized the nanodisc reconstitution protocol (see methods) for GLIC and the eukaryotic systems in the lab. The nanodisc technology creates a membrane environment around the membrane protein held in place by a membrane scaffold protein. Structure and dynamics experiments conducted on proteins reconstituted in nanodiscs have the potential to capture subtle but critical conformational changes that may be key to understanding gating and modulation of ion channels.
We successfully reconstituted GLIC into lipid nanodiscs and DEER data obtained from GLIC pH7 and GLIC pH4 nanodiscs indicated that M4 underwent outward movement from the resting to the desensitized conformation (Figure 2.12). The dynamics of M4 were unveiled in a lipid environment in line with the extensive interaction of this helix with the membrane environment.
Discussion
This study was aimed at elucidating the mechanism underlying PUFA modulation of pLGICs. In the GLIC system, the effects of DHA, EPA and AA were similar in that they enhanced desensitizing currents in GLIC and significantly reduced steady state currents. The effect of DHA was reversible and there was no difference in peak amplitude in the presence or absence of DHA. EPA and AA both
58 displayed enhanced desensitization, reduced steady state currents and reduced peak amplitude currents. In addition, GLIC currents strongly and rapidly desensitized in the presence of EPA relative to DHA and AA. Similar results were obtained with the 5-HT3AR system. It is not clear why EPA has such a significant effect yet it differs by one double bond compared to DHA and AA. It is possible that EPA has additional binding sites on pLGICs or that it partitions more effectively into the bilayer to exert additional effects though the membrane. The rapid onset of PUFA effect upon co-application suggests that the lipids are likely to interact with the protein at the protein-lipid interface. The DHA binding site in GLIC supports this idea. In addition, the full recovery of subsequent pulses after application of lipids suggests that the lipids may act near the interface where they may be easily “washed off”. However, if PUFAs were to permeate the oocyte membrane it would be expected that the membrane fluidity would change enough to demonstrably affect channel function over time. Because this is not what the experimental results show, it is possible that the surface area of the oocyte exposed to the PUFAs acts as a large reservoir that would take a long time to fill and show effects. Single channel recordings may provide insight into the effect of free PUFA permeation into a small patch of membrane.
DHA seems to enhance entry into the desensitized state and the non- desensitizing I9A /H11F double mutant confirms this hypothesis. pLGIC structures solved in apparent desensitized conformations suggest that channel pores are only constricted at the intracellular 2′ or -2 ′ desensitization gate with the activation gate at the 9′ location wide open [16, 17, 19, 25, 27]. Our structure shows
59 constrictions from the 9′ to 2′ location enough to impede a hydrated Na+ [191]. It is possible that our structure represents one of many desensitized states and that a lipid mediated desensitized conformation may be structurally different from an agonist induced desensitized state. The GLIC-DHA structure has been purported to represent a pre-active intermediate conformation along the activation pathway based on the reduction in peak current when DHA was pre-applied at neutral pH
[192]. While that may be a possibility, simple kinetic models cannot explain the complexities that may involve membrane effects as well.
The membrane environment that surrounds the protein has been suggested to modulate the equilibrium between resting and desensitized conformations. The nAChR reconstituted into PC only membranes were mainly in a desensitized-like conformation due to the tighter packing density of the bilayer. Membranes that contained cholesterol and PA were more fluid and thus stabilized a more resting like conformation that could undergo agonist induced changes [163]. An extensive study on the effect of PUFAs on the nAChR showed that exogenously applied
PUFAs shifted the receptor from the resting state but trapped it in an intermediate state [193].
The effects of PUFAs are clearly complex and may be an interplay of direct interactions with integral membrane proteins and changes in the fluidity of the lipid bilayer. We can attempt to answer some of the lingering questions by conducting structural studies to determine the binding site of these PUFAs on the 5-HT3AR by
Cryo-EM and functional studies to further elucidate whether the effects are concentration dependent.
60 Methods
Cloning and functional measurements in oocytes The gene encoding GLIC was inserted into the pTLN vector for oocyte expression and confirmed by DNA sequencing. The DNA was then linearized with the Mlu1 restriction enzyme overnight at 37°C. The cRNA was synthesized using the mMessage mMachine kit
(Ambion, Life Technologies, Carlsbad, CA), purified with RNAeasy
(Qiagen, Germantown, MD), and injected (5–15 ng) into Xenopus laevis oocytes
(stages V-VI). Control oocytes were injected with the same volume of water to verify endogenous currents were not present. Oocytes were maintained at 18°C in
OR3 media (Leibovitz media, GIBCO BRL: Life Technologies, Carlsbad, CA) containing glutamate, 500 units each of penicillin and streptomycin, pH adjusted to 7.5, osmolarity adjusted to 197 mOsm). Two electrode voltage-clamp experiments were then performed at room temperature 2–5 days after injection. A
Warner Instruments (Hamden, CT) Oocyte clamp OC-725 was used for the measurements, and the current was sampled and digitized at 500 Hz with a
Digidata 1440A (Molecular Devices, Sunnyvale, CA). Oocytes were clamped at a holding potential of −60 mV, and current traces were recorded in response to ligand application. Solutions were changed using a syringe pump perfusion system flowing at a rate of 2 ml/min. The electrophysiological solutions contain 96 mM
NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES (pH 7.4, osmolarity adjusted to 195 mOsm) or 5 mM Sodium Citrate [7, 147] (at acidic pH buffer, pH adjusted to indicated value (4.0–6); osmolarity adjusted to 195 mOsm).
All chemical reagents were purchased from Sigma-Aldrich. DHA, EPA and AA
61 stock solutions were freshly prepared in DMSO prior to each experiment and diluted to the final concentration in the electrophysiology solutions. The highest
DMSO concentration used was ~0.016% and comparison with control experiments were made using the highest concentration of DMSO in the test conditions. The traces were analyzed by Clampfit 10.2 (Molecular Devices, Sunnyvale, CA). Dose response curves were fit in Origin (OriginLab, Northampton, MA) to determine the pH50 and Hill coefficient (nH). The values for ‘n’ in the figure legends refer to the number of oocytes. Zebra fish GlyR α1 and mouse 5-HT3AR genes were also inserted into the pTLN vector and cRNA synthesized as described above. 5-HT and glycine solutions were prepared fresh before each experiment.
Electrophysiology experiments were conducted as described above.
Cloning and protein expression The GLIC gene cloned into a modified pET26b vector was expressed as a fusion construct with N-terminal maltose binding protein
(MBP) as previously described [7]. The protein was expressed and purified as previously described [7, 8, 30]. Briefly, C43 E.coli cells (Lucigen Corporation,
Middleton, WI) transformed with the construct were grown in terrific broth media containing 50 μg/ml kanamycin at 37°C to O.D600 of 1.0. Cells were induced with
0.2 mM isopropyl 1-thio-β-d-galactopyranoside (Gold Biotechnology, Olivette, MO) overnight at 18 °C. Membranes were prepared by homogenizing the cells in Buffer
A (100 mM NaCl, 20 mM Tris-HCl (pH 7.4)) with protease inhibitors and centrifuged at 100,000 x g for 45 min. Membranes were solubilized in Buffer A using 40 mM
DDM (n-dodecyl-β-d-maltopyranoside, Anatrace Inc, Maumee, OH) at 4°C. The protein was purified by binding to amylose resin and eluting with 20 mM maltose.
62 The maltose binding protein tag was cleaved with human rhinovirus 3C protease
(GE Healthcare, Wauwatosa, WI), and the pentameric protein was separated from
MBP using size exclusion chromatography on a Superdex 20/200 column (GE
Healthcare, Wauwatosa, WI) with Buffer A and 0.5 mM DDM.
Crystallization GLIC wt in 100 mM NaCl, 10 mM Tris-HCl (pH 7.4), and 0.5 mM
DDM was concentrated to between 9–10 mg/ml with an Amicon Ultra 50 KDa cutoff concentrator (EMD Millipore, Billerica, MA). Prior to crystallization setup, the protein was supplemented with 50 μM DHA (from 300 mM DHA stock in ethanol) and 0.5 mg/ml E.coli polar extract (Avanti Polar Lipids) and incubated on ice for 1 hr. The protein was crystallized at 4°C by sitting drop vapor diffusion in Cryschem plates (Hampton Research, Aliso Viejo, CA) with a 1:1 mixture (1 μl each) of protein and reservoir solution (225 mM ammonium sulfate, 50 mM sodium acetate, pH 3.9–4.2 and 9–12% PEG4000). Crystals typically formed within one week and typically took 2–3 weeks to reach full size. The crystals were cryoprotected by adding 6 μL reservoir solution supplemented with 30% ethyleneglycol to the drop, and directly frozen in liquid nitrogen using appropriately sized microloops
(MiTeGen, Ithaca, NY) or cryoloops (Hampton Research).
Structure determination X-ray diffraction data were acquired on NE/CAT beamlines 24ID-C at the Advanced Photon Source at Argonne National
Laboratory. The data was indexed using iMosflm [194] and further processed using programs within the CCP4 suite [195]. The crystals belong to space group C121 with one pentamer in the asymmetric unit. Initial phases were obtained by molecular replacement using PHASER [196] with GLIC (PDB ID: 4HFI) [190]
63 crystal structures as a search model. The initial model was refined with REFMAC5
[197] and was followed by manual model building/fitting in COOT [198]. MOLE
[199] and HOLE [200] software were used to compute pore radius profiles.
Expression of membrane scaffolding protein and GLIC reconstitution in nanodisc Membrane scaffold protein (MSP1E3D1) was expressed and purified as previously described [201] with some modifications. The MSP1E3D1 gene in pET
28a (a gift from Stephen Sligar : Addgene plasmid # 20066) [202] was transformed in E. coli BL21(DE3) cells (Agilent Technologies, Santa Clara, CA) and plated on
LB-agar plates supplemented with kanamycin (25 μg/mL). An overnight culture from a single colony was set up with LB supplemented with kanamycin (25 μg mL−1) and 1% glucose. The overnight culture was used to inoculate a 1L culture of terrific broth supplemented with kanamycin (25 μg/mL) and 0.2% glucose. The culture was grown at 37°C with shaking to an OD600 of ~1.0, and induced by with
1 mM IPTG for 4 hr at 37°C. Cells were harvested by centrifugation and the cell pellet resuspended in Buffer A containing 1 mM PMSF and Complete EDTA-free protease inhibitor cocktail tablet (Roche) and lysed by homogenization. The lysate was centrifuged at 30,000 x g for for 30 min and the supernatant was bound to Ni-
NTA equilibrated with Buffer A. The resin was washed with four bed volumes of
Buffer B (300 mM NaCl, 40 mM Tris-HCl, and pH 8.0) containing 1% Triton X-100, four bed volumes of Buffer B containing 50 mM sodium cholate, four bed volumes of Buffer B, four bed volumes of Buffer B containing 20 mM imidazole, and eluted with Buffer B containing 300 mM imidazole. The eluted MSP1E3D1 was passed through a desalting column equilibrated with Buffer C (100 mM NaCl, 50 mM Tris-
64 HCl, 0.5 mM EDTA, and pH 7.5), and the concentration was determined by absorbance at 280 nm (extinction coefficient = 29,910 M−1 cm−1). The purity was assessed by SDS–PAGE and size-exclusion chromatography.
Detergent-solubilized spin-labeled GLIC mutants passed through gel-filtration columns were incorporated into lipid nanodiscs as previously described with some modifications [201]. Briefly, asolectin dissolved in chloroform was dried using nitrogen stream and rehydrated in Buffer A supplemented with 2mM DDM. Each spin labeled mutant was mixed with rehydrated lipids and MSP1E3D1 in the
GLIC:MSP:Lipid = 1:3:360 molar ratio. The mixture was incubated at 4°C for 30 minutes with gentle rotation. Bio-beads SM-2 (Bio-Rad Laboratories) were added to initiate reconstitution overnight at 4°C with gentle rotation. Bio-beads were then removed and the reconstitution mixture assessed by size-exclusion chromatography and SDS–PAGE. Samples were equilibrated in pH 7.0 (closed state) or pH 4.0 (desensitized state). We used pH 4.0 for DEER measurements due to potential instability issues of membrane scaffolding proteins in nanodisc at more acidic pH conditions. The peak corresponding to pentameric GLIC reconstituted into nanodiscs was collected and flash-frozen in 20% glycerol for
DEER measurements.
65 Figure 2.1
Figure 2. 1: Essential long chain polyunsaturated fatty acids (PUFAs).
(Top) Linoleic acid is the ω-6 precursor to arachidonic acid (AA). AA contains 20 carbons and four double bonds. (Bottom) α-Linolenic acid is the ω-3 precursor. The two ω-3 PUFAs are eicosapentaenoic acid (EPA) with 20 carbons and 5 double bonds and docosahexaenoic acid (DHA) with 22 carbons and 6 double bonds. The omega notation refers to the carbon numbering closest to the double bond from the methyl end. (Adapted from Sokoła-Wysoczanska et al, Nutrients 2018 [203] distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0)).
66 Figure 2.2
Figure 2. 2: DHA modulation of GLIC function
(A) The trace shows a continuous recording of GLIC currents in oocytes measured by two electrode voltage-clamp (TEVC) in response to multiple pH-4.5 pulses. The pH-pulses were interspaced by perfusion with the pH 7.4 solution (for deactivation and recovery). Currents were measured in the absence (marked by red lines) or presence of 50 μM DHA (marked by blue lines). The baseline is marked as a dotted black line. DHA inhibits GLIC currents by increasing desensitization (faster current decay and lower steady-state currents; highlighted by vertical blue arrow and dotted blue/red lines). The effect of DHA on the current decay was fully reversible, as seen in the second and fourth pH-pulses. (B) A plot of the ratio of steady-state (measured at t = 2.2 min) over the peak current amplitude for the two conditions (n = 12) with s.d shown as error bars.
67 Figure 2.3
Figure 2. 3: Effect of DHA on GLIC desensitization.
(A) The effect of DHA at various concentrations (left) on currents at pH 4.5 and the effect of 50 μM DHA on currents elicited by various extracellular pH (right). The currents were recorded by TEVC at a holding potential of -60 mV. In each case, a ratio of the steady- state current (measured at t = 4.5 min) to the peak amplitude was plotted. The error bars denote s.d (n = 6). (B) Normalized peak amplitudes in the presence (blue) and absence (red) of 50 μM DHA plotted as a function of pH, and the data were fitted with the Hill equation to yield pH50 4.87 + 0.04 and nH 1.6 + 0.2 in the absence of DHA; pH50 5.02 +
0.03 and nH 1.9 + 0.3 in the presence of DHA. The error bars denote s.d (n = 3). (C) Outward currents recorded at pH 4.5 and +60 mV holding potential in the presence and absence of 50 μM DHA. The dashed lines and arrow mark the level of steady-state currents under the two conditions.
68 Figure 2.4
Figure 2. 4: Effect of DHA pre-application on GLIC currents.
Representative pH-elicited GLIC-currents recorded by TEVC at -60 mV membrane potential in response to pre-application of 50 μM DHA at pH7.4 (2.2 min duration) (middle trace) and compared with currents recorded without DHA pre-application (first and third pulses). All the three pH 4.5-pulses had 50 μM DHA. The ratio of the peak-current amplitudes measured with and without pre-application of DHA was 0.64 + 0.11 (n = 5).
69 Figure 2.5
Figure 2. 5: DHA has no effect on the non-desensitizing GLIC I9A/H11F mutant.
(A) Normalized pH-response for GLIC-wt (red) and GLIC I9A/H11F double mutant (orange) at -60 mV. The error bars denote s.d and the curve is a fit to the Hill equation.
GLIC-wt (pH50 4.87 + 0.04 and nH 1.6 + 0.2; n = 3) and GLIC I9A/H11F (pH50 5.17 + 0.19 and nH 0.88 + 0.29; n = 7). (B) Macroscopic currents measured by TEVC for GLIC I9A/H11F in response to pH jumps (from 7.4 to 4.5), at -60 mV holding potential, in the presence or absence of 50 μM DHA. (C) Currents recorded at pH 4.5 in the presence of either 50 μM DHA or 100 μM DHA.
70 Figure 2.6
Figure 2. 6: Lack of DHA effect on the non-desensitizing GLIC I9A/H11F mutant.
Macroscopic currents measured by TEVC for GLIC I9A/H11F in response to pH jumps (from 7.4 to the indicated pH value), at -60 mV holding potential, in the presence of either 50 μM DHA or 100 μM DHA.
71 Figure 2.7
Figure 2. 7: DHA binding site in GLIC.
(A) Side-view of the GLIC-pH4-DHA structure at pH 4.0 solved to 3.25 Å resolution with a bound DHA molecule shown in stick representation. Only one subunit is colored for clarity (The TM helices are colored as: M1-blue, M2-green, M3-cyan, and M4-wheat). The 2Fo- Fc electron density map for DHA, contoured at 1.0 σ-level, is shown as a blue mesh. The phospholipid molecule (PLC), shown in sticks, was also present in previously reported GLIC structures at acidic pH. (B) Zoomed-in views of the region marked by the inset in panel A (top). The GLIC-pentamer viewed from the extracellular side, overlaid with the Fo-
Fc “omit” electron density map generated by excluding DHA molecules from structure factor calculations. The green mesh is contoured at 2.0 σ and DHA molecules are drawn as ball-and-sticks. Prominent electron density is visible for all the five subunits although the continuity was variable among subunits. The lipid molecule (PLC) is also shown in a ball-and-stick presentation. Clear density for the bound PLC was observed at only one subunit (bottom).
72 Figure 2.8
Figure 2. 8: R118A mutation reduces the effect of DHA on desensitization.
Sequence for the β6-β7 loop is shown, and the position Arg118 is highlighted in red. GLIC- R118A currents measured by TEVC in response to pH jumps (from 7.4 to 4.5), at -60 mV holding potential, in the presence or absence of 50 μM DHA (left). Ratio of the steady- state current (measured at 2.2 min) to the peak amplitude was plotted (n = 12 for GLIC wt and n = 7 for R118A) with s.d shown as error bars (right).
73 Figure 2.9
Figure 2. 9: Conformational changes in the GLIC pore.
(A) The ion-permeation pathway through the channel pore, as determined by the MOLE PyMOL plugin [199]. The M2 from two subunits are shown using a ribbon representation, with residues lining the pore represented as sticks (left). Pore radius along the channel axis in GLIC structures at pH 7.0 (PDB ID: 4NPQ), pH 4.0 (PDB ID: 4HFI), and at pH 4.0 in the presence of DHA calculated using HOLE software [199, 200] (right). The constricted region from Ile9 to Thr2 is highlighted by a grey box. (B) Fo-Fc omit electron density of six dodecyl-maltoside molecules (shown in stick representation) and water pentagon (shown as red spheres) at 3.0 σ-level in M2 for different GLIC structures. The PDB ID for the structures are: GLIC-pH7: 4NPQ; GLIC-pH4: 4HFI; GLIC-pH4 (lower resolution): 3UU8. The resolution for each of the structures is indicated below.
74 Figure 2.10
Figure 2. 10: EPA and AA modulation of GLIC function.
Co-application of EPA or AA results in inhibition of GLIC currents by increased desensitization (faster current decay and lower steady-state currents; highlighted by vertical blue arrow and dotted blue/red lines) and reduced peak currents. Currents were measured in the absence (marked by red lines) or presence of 50 μM EPA or AA (marked by blue lines). EPA shows a more pronounced effect by its faster kinetics of decay of GLIC currents compared to AA.
75 Figure 2.11
Figure 2. 11: DHA, EPA and AA effect on 5-HT3AR function.
Each trace shows a continuous recording of 5-HT3AR currents in oocytes measured by TEVC in response to multiple 5-HT pulses. The pulses were interspaced by perfusion with the pH 7.4 solution (for deactivation and recovery). Currents were measured in the absence (marked by red lines) or presence of 50 μM DHA, EPA (ω-3) or AA (ω-6) (marked by blue lines). The lipids seem to enhance desensitization of 5-HT3AR currents with an additional effect of reduced peak currents (smaller panels showing normalized pulse to determine effect on desensitization with and without lipids). The effect of DHA, EPA and AA on the current decay was fully reversible, as seen in the third pulses.
76
Figure 2.12
Figure 2. 12: Changes in M4 distance measured by DEER for GLIC in nanodiscs.
GLIC structure showing the positions investigated by DEER and the two expected distance distributions (from the adjacent and non-adjacent subunits). Background subtracted DEER- echo intensity is plotted against evolution time and fit using model-free Tikhonov regularization. The corresponding inter-spin distance distribution (right) for the closed (black, pH 7.0) and desensitized (red, pH 4.0) states for different spin-labeled positions. The arrows highlight the direction of change.
77 Chapter 3
Understanding allosteric and orthosteric inhibition mechanisms of the 5-HT3A receptor
Portions of this chapter were published in: Basak S, Gicheru Y, Kapoor A, Mayer M.L, Filizola M, Chakrapani S. Molecular mechanism of setron-mediated inhibition of full-length 5-HT3A receptor. (Under review). Structural studies were conducted by S.B. Functional studies were conducted by Y.G.
78 Introduction
In the previous chapter, the critical roles played by PUFA metabolites
(eicosanoids and autacoids) in neuroprotection, neurogenesis and neuroinflammation were discussed. Endogenous cannabinoids
(endocannabinoids, EC) are derivatives of PUFAs which play critical roles in signaling through the endocannabinoid system (ECS). The ECS is composed of endocannabinoids and enzymes that synthesize and degrade ECs and the cannabinoid receptors [204]. ECs derived from DHA and EPA are N-
Docosahexaenoylethanolamine (DHEA) and N-Eicosapentaenoylethanolamine
(EPEA) respectively. AA derived ECs are N-Arachidonoylethanolamine (AEA) and
2-Arachidonoylglycerol (2-AG). AEA and 2-AG are the major ECs in the ECS and the most studied (Figure 3.1A). ω-3 derived ECs are less well studied and are presumed to modulate AEA signaling and have anti-inflammatory properties. Their concentration is about two fold higher than AEA and thought to have low affinity to the cannabinoid receptors [170].
AEA and 2-AG are synthesized and degraded in two distinct pathways. The
AEA pathway involves NAT (N-acyltransferase enzyme) and NAPE-PLD (N-acyl- phosphatidylethanolamine phospholipase D) in synthesis and FAAH (fatty acid amide hydrolase) in degradation. 2-AG is produced by the actions of PLC-β and
DAGL (Diacylglycerol-lipase) and degraded mainly by MAGL (monoacylglycerol lipase). DHA and EPA ethanolamides share the AEA pathways while their glycerol counterparts are not well studied [170]. Synthesis of ECs happens on demand in part because DAGL and NAPE-PLD are activated by high Ca2+ concentrations
79 which then initiates AEA and 2-AG synthesis from their precursors. Studies have found that basal and activated levels of 2-AG are much higher than AEA [205,
206].
Cannabinoid receptors (CB) are divided into two classes: CB1 and CB2 receptors. ECs mediate their effects primarily through these receptors although we now know that there are CB receptor-independent effects of EC signaling [207].
CB receptors are G-protein coupled receptors (GPCRs) mainly to Gi or Go proteins where they inhibit adenylyl cyclase and voltage gated Ca2+ channels and activate inwardly rectifying K+ channels and MAP kinases. CB1 receptors are predominantly located at presynaptic sites in the central and peripheral nervous systems and some peripheral tissues while CB2 receptors are mainly in microglia and peripheral immune cells [204, 208]. ECs are retrograde signaling molecules which inhibit synaptic transmission inducing several forms of synaptic plasticity mainly through CB1 receptors. AEA binds to CB1 receptors with partial agonist affinity and to CB2 receptors with even lower affinity. 2-AG binds both CB1/CB2 receptors with equal high agonist efficacy. In general the ECS is involved in modulatory actions of the autonomic and sensory nervous systems in pain perception, gastrointestinal, respiratory and cardiovascular functions. In pathological brain conditions such as glutamate excitotoxicity, ischemia, starvation and painful stimulus, the ECS is activated [208]. Interestingly, there are reports of
AEA CB1-independent effects such as cataplexy, analgesia, decreased neuronal excitability and pain in CB1 knockout mice [95].
80 Plant cannabinoids (phytocannabinoids) mediate their actions in the human body through CB receptors. However, it is increasingly clear that phytocannabinoids can bind to and modulate the actions of many other receptors including pLGICs [96, 141, 209-212]. THC (∆9-tetrahydrocannabinol) and CBD
(cannabidiol) are two of close to 60 cannabinoids present in the cannabis plant
(Figure 3.1B). Of those, THC is the only known component to have psychoactive effects. THC and CBD are reported to have antiemetic, antitumor, anticonvulsive and analgesic effects [213]. In addition, CBD has been shown to have anti- inflammatory, antianxiety, antipsychotic, and anti-rheumatoid arthritic effects [214].
THC is a partial agonist at CB1 receptors and CB2 receptors. CBD has very low agonist affinity for CB1/CB2 receptors that in fact it antagonizes agonists for these receptors. Due to its low efficacy at these receptors, CBD does not display the psychoactive effects seen in THC [215].
CB receptor-independent therapeutic effects of the endogenous cannabinoids and the phytocannabinoids in the context of ligand gated ion channels is an active area of research. The interaction of the phytocannabinoids and AEA with TRP channels is being investigated with implications for pain and cancer therapies [211, 216]. CBD, THC and AEA have been shown to potentiate glycine-activated currents in the α1 and α1β receptors implying that CBD could be mediating some of its neuroprotective and anti-inflammatory effects through CB- independent mechanisms [141, 209]. Functional mutagenesis studies and NMR chemical shifts identified a potential THC binding site in the M3 helix (near Ser296) in the α1 and α3 GlyR. Additionally, THC derivatives with analgesic properties but
81 without psychoactive effects could provide a new avenue to designing new classes of analgesics [217]. The α7 nAChR currents are inhibited by 2-AG, R- methanandamide (stable AEA analog) and CBD. Interestingly, THC did not alter the function of the receptor [212, 218]. THC and CBD structures are fairly similar suggesting that specific interactions between the receptor and CBD must occur that are not present with THC.
The 5-HT3AR is also modulated by endo- and phyto- cannabinoids. 5-HT3AR in nodose ganglion neurons were inhibited by AEA. A separate study demonstrated that the inhibitory effect of AEA, THC and CBD on receptors expressed in oocytes was through CB1-independent mechanisms (IC50s of 3.7
μM, 1.2 μM, 0.6 μM respectively) [96, 219, 220]. Importantly, 5-HT3R are the main target for the management of nausea and vomiting caused by chemotherapy and radiation therapy. These symptoms drastically affect patient health and quality of life, and in turn impact chemotherapy compliance [221]. 5-HT3R antagonists, referred to as “setrons”, are the standard first-line of therapy. However, delayed and refractory emesis continue to be serious clinical concerns and adverse side effects limit clinical use [105]. Setrons also find use in chronic pain management, alcohol abuse, mood disorders and irritable bowel syndrome [98, 106, 222, 223].
Thus, it bears noting that 5-HT3R antagonists and cannabinoids bring about the same therapeutic effects such antiemetic effects, antianxiety, and analgesia. In fact, THC based medications (dronabinol, (Marinol), and nabilone) are FDA approved for the treatment of refractory emesis [224-226]. The lipophilicity of cannabinoids suggests that their mechanism of action in modulating receptor
82 function is similar to PUFAs acting at allosteric sites. However, there is no definitive structural mechanism to explain the inhibitory actions of endo- and phytocannabinoids. Further, THC’s psychoactive effects are a concern for some patients receiving it as treatment. CBD on the other hand is becoming increasingly sought-after because it has the same properties as THC without the adverse effects through the CNS [227, 228].
The aim of this chapter is to explore allosteric (cannabinoids) and orthosteric (setrons) mechanisms of action on the function of the 5-HT3R. An understanding of allosteric mechanisms of inhibition could allow receptor subtype specificity without altering the pattern of endogenous neurotransmitters. The legalization of medical marijuana with a focus on THC and CBD makes it all the more critical to understand the mechanism of action of these compounds.
Structural insight could also lead to the development of more drugs with fewer CNS effects.
Conventional belief in the field of pLGICs was that the conformation of a receptor in the resting state was structurally identical to that of the receptor in the antagonist-bound state. This is because both states are non-conductive and are not bound to activating ligands. However, some studies seemed to suggest that the resting state and the antagonist bound state displayed conformational differences [15, 229-231]. The gravity of the side-effects of cancer treatment necessitated an understanding of how these drugs mediate their functions in an effort to enhance their efficacy especially with delayed and refractory emesis [232].
Before high resolution structures became available, setron binding was based on
83 docking studies using homology models. Many of these predictions found several energetically favorable poses within the orthosteric site and residue interactions varied significantly among these studies due to uncertainties in homology models and the binding pocket conformation [40, 233-235]. A chimera of the AChBP containing 5-HT3AR binding site residues was solved by X-ray crystallography in the presence of granisetron showing the overall binding pose and important residues for ligand recognition [236]. A recent 5-HT3AR structure bound to tropisetron was solved to 4.5 Å showing density for the antagonist in the classic orthosteric site with an occluded pore [49]. However, at this resolution, the precise orientation of the ligand and surrounding side chains are somewhat ambiguous. In addition, the differences between the resting and antagonist-bound states of the
5-HT3AR remain unclear.
Therefore, we set out to understand the mechanism for modulation of cannabinoids and orthosteric inhibition of setrons in the 5-HT3AR using electrophysiology and cryo-electron microscopy. Electrophysiology experiments showed a reduction in 5-HT3AR peak currents by endo-/ phyto- cannabinoids and granisetron. Granisetron is clinically approved for the management of nausea and vomiting, and in combination with corticosteroids, is the recommended regimen for patients undergoing highly-emetogenic chemotherapy. We conducted structural studies to identify the potential allosteric binding site for THC on the 5-HT3AR and thus the structural basis for modulation. We elected to solve the structure of 5-
HT3AR in complex with THC first because its derivatives are currently in use for the treatment of refractory nausea and vomiting in cancer therapy. We then solved the
84 structure of 5-HT3AR in complex with granisetron to identify the precise binding of granisetron in the orthosteric site and the conformational changes resulting in channel inhibition.
Results
Cannabinoid inhibition of 5-HT3AR currents
To test the effect of cannabinoids (AEA, 2-AG, THC and CBD) on 5-HT3AR function, we expressed 5-HT3AR in Xenopus oocytes and measured currents by two-electrode voltage-clamp (TEVC). Application of 1 µM serotonin elicited transient activation and desensitization currents. The peak current was significantly inhibited when oocytes were pre-exposed to each of the cannabinoids for about 10 minutes with the exception of AEA (~3 min). Ligand concentrations used were chosen to ensure sufficient current density was evoked in response to serotonin and yet an inhibition was observed with each cannabinoid. In general, the effect of the cannabinoids was reversible after wash off and subsequent 5-HT application (Figure 3.2A and 3.2B).
Ligand binding sites in 5-HT3AR
To better understand the mechanism of action of THC we used single- particle Cryo-EM to solve the structure of the full length 5-HT3AR in complex with
THC and 5-HT. Detergent solubilized 5-HT3AR was incubated with 30 µM 5-HT and
85 30 µM THC and incubated for 1 hour prior to vitrification on Cryo-EM grids. Four separate sample preparations and data collection sessions were completed. Each data set was processed individually and the particles that resulted in the best 3D reconstruction were extracted and combined. Iterative classifications and refinements produced a final three dimensional reconstruction at a resolution of
3.6 Å with 82,000 particles. The final map contained density for the entire ECD,
TMD and ICD with the exception of the flexible MA loop and the post M3 loop and parts of the MX helix (Figure 3.3A). The non-protein densities in the map include three sets of protrusions at the ECD periphery corresponding to N-linked glycans
(Figure 3.3B) and a distinct strong density at each of the intersubunit interfaces in the ECD, corresponding to serotonin (Figure 3.3C). Particle polishing and further refinements and classifications will be done to improve the quality of the map and to obtain the best class for model building.
The THC-5HT-5HT3AR map was compared with the maps for the recently solved serotonin bound structures (state 1- pre-open or desensitized and state 2- open, conductive) [18] and it was clear that the ECD and most of the TMD resembled state 2 (Figure 3.4, left panel) suggesting that the conformational changes in state 2 in these domains are likely present in the THC-5HT-5HT3AR structure. This will be verified when model building is completed. Loop C appears closed in on 5-HT and the ECD is expected to have moved in an anticlockwise direction relative to Apo-5HT3AR. The TMD-ICD interface is somewhat less well resolved. Density for MX and part of the post-M3 loop is diffuse. The MA helices of the ICD in the THC-5HT-5HT3AR structure have continuous density which
86 appears a bit diffuse when approaching the membrane. The potential THC binding site is indicated in the red boxed region (Figure 3.4, right panel) but it is not possible to unambiguously model THC until further data processing. The density for the MX helix in all the data sets appears diffuse suggesting that the presence of THC in this vicinity possibly causes the helix be more dynamic compared to the other 5-
HT3AR structures. When we compare the relative “openness” of the lateral portal in the two structures, state 2 appears to have a more “open” surface area compared to THC-5HT-5HT3AR (Figure 3.4, right panel) possibly due to residues from the post-M3 loop and MA helices coming closer together in the latter structure. Further analyses on the structure as well as the potential THC binding site is currently underway. Validation of the THC binding site will be done by mutagenesis.
Granisetron inhibition of 5-HT3AR currents
In looking to understand the structural basis for inhibition of the 5-HT3AR, we conducted functional and structural experiments using the clinically approved first generation antagonist, granisetron. Granisetron competitively inhibits serotonin-activated currents at nanomolar concentrations. The effect of granisetron on 5-HT3AR function was measured by two electrode voltage clamp
(TEVC) in Xenopus oocyte. Full-length 5-HT3AR cRNA was injected into oocytes and measurements were recorded 2-5 days post-injection. Application of 10 µM serotonin elicited transient currents associated with rapid channel activation and desensitization (Figure 3.5, first pulse). The peak current was significantly inhibited
87 when oocytes were pre-exposed to 1 nM granisetron (Figure 3.5, second pulse).
The inhibition was fully reversible when granisetron was washed off and serotonin was subsequently applied (Figure. 3.5, third pulse).
Granisetron binding and conformational changes
5-HT3AR was incubated with 100 µM granisetron for 1 hour before vitrification on Cryo-EM grids. Iterative classifications and refinements produced a final three-dimensional reconstruction at a resolution of 2.92 Å using a total of 46,
757 particles. The overall architecture of granisetron-bound 5-HT3AR (5-HT3AR- granisetron) is similar to previously solved 5-HT3AR structures. In the present map, all the domains are well-resolved, with the exception of the unstructured MX-MA loop in the ICD (Figure 3.6A) making it possible to build a model of the full-length structure (Figure 3.6B). Non-protein densities present in the map include three sets of protrusions at the ECD periphery corresponding to N-linked glycans (Figure
3.6A, right) and a distinct strong density at each of the intersubunit interfaces in the ECD, corresponding to granisetron (Figure 3.6A red arrow, and 3.6C).
The high resolution of the density map enabled the modelling of the precise orientation of granisetron and its immediate environment especially the “aromatic cage” formed by surrounding residues at the canonical neurotransmitter-binding pocket (Figure 3.7A). In the 5-HT3AR-granisetron structure, residues from loops A,
B, and C on the principal (+) subunit and loops D, E, and F from the complementary
(-) subunit form an enclosure around granisetron. Residues within 4 Å of
88 granisetron include Trp156 in loop B, Phe199 and Tyr207 in loop C, Trp63 and
Arg65 in loop D, and Tyr126 in loop E. These residues are strictly conserved and perturbations at each of these positions impact binding of both granisetron and serotonin [40, 237, 238]. In comparison to the resting conformation of the 5-
HT3AR (5-HT3AR-apo) [36] and a serotonin-bound potentially pre-open or desensitized conformation (5-HT3AR-serotonin) [18], these residues have undergone rotameric reorientations (Figure 3.7B). The cationic center of granisetron is a tertiary ammonium in the bicyclic tropane ring (pKa 9.6) and is positioned at the deep-end of the pocket nestled by Trp156 (loop B), Tyr207 (loop
C), Trp63 (loop D) and Tyr126 (loop E). The close proximity of the tropane nitrogen and Trp156 is conducive for a cation-pi interaction, as predicted from the
AChBP-5-HT3 chimera structure. Trp156 is also involved in a similar interaction with the primary amine of serotonin. However, in comparison to serotonin, granisetron is larger in size and extends further outward into the subunit interface such that the aromatic indazole ring is oriented parallel to the membrane and makes a cation-pi interaction with the positively charged nitrogen in the guanidinium group of Arg65 (Figure 3.7A). Granisetron’s position also requires
Arg65 and Trp168 sidechains to reorient. Crystal structures of AChBP carrying mutations to mimic the serotonin-binding pocket of 5-HT3AR and bound to either granisetron or tropisetron reveal an identical binding pose for the two inhibitors.
However, comparison of 5-HT3AR-granisetron and 5-HT3AR-tropisetron structures
(Figure 3.8) shows that although the bicyclic tropane rings are in similar
89 orientations, the indole/indazole ring and the sidechains of Arg65 and Trp168 are positioned differently [49].
An alignment of 5-HT3AR-granisetron with the 5-HT3AR-apo structure shows a counter-clockwise twist of beta strands in the ECD leading to a small inward movement of loop C closing in on granisetron (Figure 3.7C and Figure
3.9A). The conformation of loop C has been correlated with the ligand occupancy in the binding site and the functional state of the channel. Studies with the
AChBP have shown that agonist binding induces a “closure” of loop C, capping the ligand-binding site. This conformational change may be linked to channel opening of pLGICs. On the other hand, antagonist-bound structures show loop
C further extended outward [15]. The loop C conformation in 5-HT3AR- granisetron, although open, is distinct from other antagonist loop C conformations because it lies between the open position of loop C in the 5-HT3AR-apo structure and the closed position in the 5-HT3AR-serotonin structure (Fig. 3.7C). The TMD of 5-HT3AR-granisetron is rotated clockwise compared to 5-HT3AR-apo, once again in a position between the apo and serotonin-bound structures (Fig. 3.9B). The M2-
M3 linker is bent and tilted upwards but positioned slightly lower than in the 5-
HT3AR-apo conformation. Although granisetron induces a rotational movement in the TMD, the pore radius is comparable to 5-HT3AR-apo and the permeation pathway presents barriers in all the three domains (Fig. 3.9C). The first barrier to ion permeation is in the middle of the ECD at Lys108 in the β4–β5 loop at ~2 Å.
The M2 helices appear straight with constrictions at 16′, 13′, 9′, 6′ and 2′ that are smaller than the radius of a hydrated Na+ [191]. Leu260 (9′) and Ser253 (2′) are
90 located in the middle of the pore and form the narrowest region of the pore. The intracellular end of the channel lined by Glu250 which has been implicated in governing charge selectivity is wide open [44]. The constriction at L9′ renders the channel non-conductive. In agreement, the recently solved tropisetron bound 5-
HT3AR [49] shows constrictions at the same locations of the pore.
Radioligand binding studies have revealed that mutating Trp156 and Trp63 drastically affects granisetron binding [239, 240]. In addition, eliminating the interaction between Arg65 and the indazole ring of granisetron significantly reduces granisetron affinity [235, 237]. In line with these studies, we found that the extent of granisetron inhibition is significantly reduced when residues in the 5-
HT3AR binding site are mutated (W63Y, R65A and W156Y) compared to the wild type receptor (Figure 3.10). For competitive antagonists, the extent of inhibition depends on agonist concentration as antagonist effects decrease with increasing agonist concentration. To make a meaningful comparison of granisetron inhibition across various mutations, the serotonin concentration in each case was kept close to the EC50 value for the mutant. This value was chosen to ensure sufficient current density was evoked in response to serotonin and yet an inhibition was observed with granisetron. The EC50 value for R65A was determined experimentally (Figure
3.11). In 5-HT3AR, serotonin and granisetron interact at the same binding pocket and interestingly, granisetron induces rotameric reorientations of the sidechains in the same direction as serotonin, albeit of smaller magnitude. In addition, the global motion of the granisetron-bound 5-HT3AR is in the direction of activation relative to the 5-HT3AR-apo structure but channel activation is not realized in the presence of
91 the antagonist. Binding of granisetron seems to increase the energy barrier required to reach activation thus resulting in the stabilization of an inhibited state.
This work highlights the precise orientation of granisetron’s binding pose in the orthosteric site and induces conformational changes different than the resting state that stabilize a non-conducting state.
Discussion
Endo- and phyto- cannabinoids inhibit serotonin activated currents when the channels are pre-exposed to the compounds. THC, CBD and 2-AG were pre- applied for about 10 minutes for inhibition to be evident upon co-application with serotonin. A 3 minute pre-application was enough for the effect of AEA to be seen.
Recovery for most of the channels required at least 5-10 minutes of wash off. Co- application of cannabinoids with 5-HT without pre-application did not have an effect on peak currents. It seems that PUFAs are either able to diffuse through the membrane very fast and get to the channel or they interact with the receptors at the protein-lipid interface as suggested by the GLIC-DHA structure. On the contrary, cannabinoids appear to take time to perfuse into the membrane before they can modulate channel function. The inhibitory effect of AEA happens much faster than the other cannabinoids suggesting that it partitions into the membrane faster or it interacts closer to the protein-lipid interface. A higher concentration of
2-AG was required to show inhibition because a 3 μM concentration did not have an effect on peak current. THC and CBD appear to be similar in effect. A detailed functional analysis of the cannabinoids would have informed on IC50s to better
92 indicate stronger effectors. The effects of the cannabinoids are reversible after considerable wash off. It is interesting that similarly to PUFAs, there seems to be no residual effect on the oocyte bilayer. It is possible that the exposed surface area is large enough to adsorb the cannabinoids without changing the polarity of the bilayer. It is also possible that over a period of time exposure to the cannabinoids and PUFAs may change the membrane properties thus affecting channel function differently. Cannabinoids have been shown to inhibit pLGIC function by enhancing desensitization [241]. Our functional studies did not show apparent enhanced desensitization although our analyses were limited. Additional data collection with varied ligand concentrations may provide additional insight.
The density map of 5-HT3AR bound to THC and 5-HT shows that the ECD and TMD is likely in an activated conformation because it resembles the serotonin activated state 2 map. However, the lateral portals in the ICD of the THC bound structure appear to be more closed in relative to the state 2 portals. This may be the first indication for how cannabinoids may mediate their inhibitory action. The potential THC binding site was initially guided by the structure of THC bound to 5-
HT3AR in the absence of serotonin. In this structure, new density was evident in an intersubunit cavity lined by residues from M4 in one subunit and M3 in the adjacent subunit. However, there were minimal conformational changes in the protein. This intersubunit binding pocket towards the bottom of the helices has been shown to bind potentiating neurosteroids in GABAA receptors while an intrasubunit pocket in the vicinity of the inter-subunit pocket has also been shown to bind inhibitory neurosteroids [242]. Other studies have suggested multiple neurosteroid sites
93 based on subunit composition [243]. In the THC-5-HT bound structure, there appears to be new density in the same general area. This is an exciting preliminary finding that will be validated with mutagenesis and functional studies.
In looking at orthosteric inhibition, granisetron binding induces conformational changes that are in the path toward channel activation but the energy barrier for activation is not realized in the presence of granisetron. This does not, however, suggest that granisetron has weak agonist activity because pre-exposure to the channels in the absence of serotonin does not activate the 5-
HT3AR. The overall conformation of loop C in 5-HT3AR-granisetron is similar to the bicuculline bound GABAAR conformation. Bicuculline binds at the orthosteric location and loop C flexes inward slightly to accommodate the antagonist but it maintains an overall extended conformation. The pore in both structures is constricted at the activation gate Leu9′ position, but the -2′ position in the 5-HT3AR- granisetron is open enough to allow ions to pass while in the bicuculline-GABAAR the same position is fully closed [19]. This is an interesting observation because the -2′ location is considered important for ion selectivity and desensitization and this variation may reflect differences between cationic and anionic pLGICs. The differences between the granisetron bound structure and the resting state structure indicate that the two structures are distinct despite both being non-conductive.
It is not surprising that granisetron occupies all five binding sites because it is a homomeric receptor with identical binding sites. In addition, all the 5-HT3R structures solved until now are homopentamers. Thus, the question of ligand occupancy cannot be addressed because data processing requires averaging of
94 all five ligand-binding sites. There is some indication that 5-HT3AR may be more abundant in the central nervous system and the 5-HT3ABR in the peripheral nervous system. Future meaningful studies would involve solving structures of the heteropentameric 5-HT3AB receptor in complex with orthosteric and allosteric ligands. Questions of ligand occupancy and the physiological relevance of the B subunit to channel activation and inhibition may be answered in this context.
Furthermore, a more complete understanding of ligand recognition leading to channel activation and inhibition will require the context of a membrane environment depicting a physiological state. Also, the physiological contribution of the other 5-HT3R subunits (C-E) is still unclear and structural studies may shed light on receptor assembly, modulation and function. Overall, strategies targeting both allosteric and orthosteric sites can pave the way for design of drugs that are subunit specific in the CNS or gut to treat psychiatric or GI tract related conditions with less off-target effects.
Methods
Electrophysiological measurements in oocytes The gene encoding mouse 5-
HT3AR (purchased from GenScript) was inserted into a Xenopus laevis expression vector (pTLN). DNA linearization was carried out with the Mlu1 restriction enzyme overnight at 37 °C. cRNA synthesis was done using the mMessage mMachine kit
(Ambion) as per the manufacturer’s instructions and purification was done with
RNAeasy kit (Qiagen). cRNA for studied mutants was prepared in the same way.
3–10 ng of mRNA was injected into X. laevis oocytes (stages V–VI) and
95 experiments were performed 2-5 days after injection. For control experiments to verify that no endogenous currents were present, oocytes were injected with the same volume of water. W. F. Boron kindly provided oocytes used in this study.
Female X. laevis were purchased from Nasco. Animal experimental procedures were approved by Institutional Animal Care and Use Committee (IACUC) of Case
Western Reserve University. Oocytes were maintained at 18 °C in OR3 medium
(GIBCO-BRL Leibovitz medium containing glutamate, 500 units each of penicillin and streptomycin, pH adjusted to 7.5, osmolarity adjusted to 197 mOsm). Two- electrode voltage-clamp experiments were performed on a Warner Instruments
Oocyte Clamp OC-725. Currents were sampled and digitized at 500 Hz with a
Digidata 1332A. Data were analyzed by Clampfit 10.2 (Molecular Devices).
Oocytes were clamped at a holding potential of −60 mV. Solutions were changed using a syringe pump perfusion system flowing at a rate of 6 ml/min. The electrophysiological solutions consisted of (in mM) 96 NaCl, 2 KCl, 1.8 CaCl2, 1
MgCl2, and 5 HEPES (pH 7.4, osmolarity adjusted to 195 mOsM). Chemical reagents (THC, CBD, AEA, 2-AG, serotonin hydrochloride, and granisetron hydrochloride) were purchased from Sigma-Aldrich. For wild type and mutants, the extent of inhibition was assessed by the ratio of peak current in the presence of granisetron over peak current in the absence of granisetron.
Full-length 5-HT3AR cloning and transfection Codon-optimized mouse 5-HT3AR
(NCBI Reference Sequence: NM_001099644.1) was purchased from GenScript.
The gene was subcloned into the pFastBac1 vector consisting of four strep-tags
(WSHPQFEK) at the N terminus, followed by a linker sequence
96 (GGGSGGGSGGGS) and a TEV-cleavage sequence (ENLYFQG) and a C- terminal 1D4-tag [244]. Spodoptera frugiperda (Sf9) cells (Expression System) were cultured in ESF921 medium (Expression Systems) without antibiotics and incubated at 28 °C without CO2 exchange. Transfection of sub-confluent cells was done with recombinant 5-HT3AR bacmid DNA using Cellfectin II transfection reagent (Invitrogen) according to manufacturer instructions. Cell culture supernatants were collected and centrifuged at 72 h post-transfection at 1,000g for 5 min to remove cell debris to obtain progeny 1 (P1) recombinant baculovirus.
Sf9 cells infected with P1 virus stock produced P2 virus and subsequently P3 viruses from the P2 virus stock. P3 viruses were used for recombinant protein production.
6 5-HT3AR expression and purification Approximately 2.5 × 10 per ml Sf9 cells were infected with P3 recombinant viruses. After 72 h post-infection, the cells were harvested and centrifuged at 8,000g for 20 min at 4 °C to separate the supernatant from the pellet. The cell pellet was resuspended in dilution buffer (20 mM Tris-HCl, pH 7.5, 36.5 mM sucrose) supplemented with 1% protease inhibitor cocktail
(Sigma-Aldrich). Cells were disrupted by sonication on ice and non-lysed cells were removed by centrifugation (3,000g for 15 min). The membrane fraction was separated by ultracentrifugation (167,000g for 1 h) and solubilized with 1% C12E9 in a buffer containing 500 mM NaCl, 50 mM Tris pH 7.4, 10% glycerol and 0.5% protease inhibitor by rotating for 2 h at 4 °C. Non-solubilized material was removed by ultracentrifugation (167,000 g for 15 min). The supernatant containing 5-HT3A receptors was collected and bound with 1D4 beads pre-equilibrated with 150 mM
97 NaCl, 20 mM HEPES pH 8.0 and 0.01% C12E9 for 2 h at 4 °C. The beads were then washed with 100 column volumes of 150 mM NaCl, 20 mM HEPES pH 8.0, and 0.01% C12E9 (buffer A). The protein was then eluted with buffer A supplemented with 3 mg/ml 1D4 peptide (TETSQVAPA). Eluted protein was concentrated and deglycosylated with PNGase F (NEB) by incubating 5 units of the enzyme per 1 μg of protein for 2 h at 37 °C under gentle agitation.
Deglycosylated protein was then applied to a Superose 6 column (GE healthcare) equilibrated with buffer A. Fractions containing the protein were collected and concentrated to 2–3 mg/ml using 50-kDa MWCO Millipore filters (Amicon) for cryo-
EM studies.
Cryo-EM sample preparation and data acquisition 5-HT3AR protein (~2.5 mg/ml) was filtered and incubated with 100 μM granisetron or 30 μM THC and 30
μM 5-HT for 1 hour. 3 mM fluorinated Fos-choline-8 (Anatrace) was added and the sample was incubated until blotting. The sample was blotted twice with 3.5 μl sample each time onto Cu 300 mesh Quantifoil 1.2/1.3 grids (Quantifoil Micro
Tools), and immediately the grid was plunge frozen into liquid ethane using a
Vitrobot (FEI). The grids were imaged using a 300 kV FEI Titan Krios microscope equipped with a Gatan K2-Summit direct detector camera. Movies containing 30 frames (granisetron) or 40 frames (THC) were collected at 130, 000× magnification
(set on microscope) in super-resolution mode with a physical pixel size of 0.532
Å/pixel, dose per frame 1.30 e-/Å2. Defocus values of the images ranged from −1.0 to −2.5 µm (input range setting for data collection) as per the automated imaging software EPU or Leginon.
98 Image processing MotionCor with a B-factor of 150 pixels2 was used to correct beam-induced motion. Super-resolution images were binned (2×2) in Fourier space, making a final pixel size of 1.064 Å. All subsequent data processing was conducted in RELION 3.0. The defocus values of the motion-corrected micrographs were estimated using Gctf software. For the granisetron dataset, approximately, ~243, 290 auto-picked particles from 1, 318 micrographs were subjected to 2D classification to remove suboptimal particles. An initial 3D model was generated from the 5-HT3AR-apo cryo-EM structure (RCSB Protein Data Bank code (PDB ID): 6BE1). A low-pass filter of 60 Å was applied using EMAN2. Multiple rounds of 3D auto-refinements and 3D classifications generated 2 good classes with 53, 029 particles and 7, 072 particles. To further improve the resolution, per- particle motion correction was performed using Bayesian polishing in RELION 3.0 which was followed by 3D-autorefinement and classification yielding one major class with 46, 757 particles. Per-particle contrast transfer function (CTF) refinement and beam tilt correction were applied followed by a final 3D- autorefinement. In the post-processing step in RELION, a soft mask was calculated and applied to the two half-maps before the Fourier shell coefficient
(FSC) was calculated resulting in an overall resolution of 2.92 Å (FSC = 0.143 criterion). The B-factor estimation and map sharpening were performed in the post- processing step. For the THC dataset, the best particles after multiple rounds of
3D refinements and classifications from 3 datasets were combined for a total of
98, 000 particles. Further refinements and classifications and postprocessing yielded one good class with 82, 000 particles at 3.62 Å resolution. To further
99 improve the resolution we will perform per-particle motion correction and CTF refinement/beam tilt followed by final 3D classification and refinements similar to what is described above in the granisetron data processing.
5-HT3AR model building THC-5HT bound 5-HT3AR dataset: Model building to be done once data processing is completed.
Granisetron dataset: The map for 5-HT3AR-granisetron contained density for the entire ECD, TMD and a large region of the ICD. The final refined models comprised of residues Thr7–Leu335 and Leu397–Ser462. The missing region (336–396) is of the unstructured MX loop that links the amphipathic MX helix6 and the MA helix21. The 5-HT3AR-apo Cryo-EM structure (PDB ID: 6BE1) was used as an initial model and aligned to the 5-HT3AR-granisetron Cryo-EM map calculated with
RELION 3.0. Cryo-EM map was converted to the mtz format using mapmask and sfall tools in CCP4i software. The mtz map was then used for manual model building in COOT. After initial model building, the 5-HT3AR-granisetron model was refined against its EM-derived map using the phenix.real_space_refinement tool from the PHENIX software package, using rigid body, local grid, NCS, and gradient minimization. The model was then subjected to additional rounds of manual model fitting and refinement. Stereochemical properties of the model was evaluated by
Molprobity. To compare the 5-HT3AR-apo to 5-HT3AR-granisetron all ligands, ions and water molecules were removed from the PDB files. The pore profile was calculated using the HOLE program. Figures were prepared using PyMOL v.2.0.4
(Schrödinger, LLC).
100 Figure 3.1
Figure 3. 1: Endocannabinoids and phytocannabinoids.
(A) The endocannabinoids anandamide (AEA) and 2-arachidonoylglycerol (2-AG) which are synthesized from arachidonic acid. (B) THC and CBD are two of the main plant cannabinoids that are currently in medicinal use (adapted from Fezza et al, Molecules 2014) [245] distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0)).
101 Figure 3.2
Figure 3. 2: Cannabinoid modulation of 5-HT3AR function.
Each trace shows a continuous recording of 5-HT3AR currents induced by serotonin (5- HT) in oocytes by TEVC. Currents were measured in the absence (marked by red lines) or presence of pre-applied modulators at the indicated concentration (marked by blue lines). The third pulse shows the cannabinoid effect is reversible. The arrows highlight the extent of inhibition. (A) Endocannabinoids (B) Phytocannabinoids. AEA was pre-applied for the shortest time (~3 min) while the other ligands were pre-applied for ~10 min. Concentrations lower than 15 μM of 2-AG did not show much inhibition. THC and CBD appeared to have a similar effect on inhibition. Wash-off times varied from at least 5-10 min.
102 Figure 3.3
Figure 3. 3: Cryo-EM reconstruction of THC-5HT-bound full-length 5-HT3AR.
Map of full-length 5-HT3AR bound to 5-HT and THC reconstructed from 82, 000 particles at 3.62 Å resolution. (A) Side-view parallel to the membrane and (B) extracellular view shown with glycans indicated by arrow. (C) Top-view of 5-HT3AR-THC-5HT ECD sliced at the binding site to show five serotonin molecules, each bound at the interface of two subunits (indicated by arrow).
103
Figure 3.4
Figure 3. 4: Comparison of serotonin-bound 5-HT3AR with 5-HT3AR-THC-5HT.
The map of serotonin-bound 5-HT3AR (grey) superimposed on the 5-HT3AR-THC-5HT map (green). The ECDs are identical and most of the TMD helices in state 2 matched the THC- 5HT bound structure (left panel). A close up of the ICD showing the potential THC binding site boxed in the red square. An outline of the surface area of the portals shows that state
2 appears to have a more “open” surface area than in the 5-HT3AR-THC-5HT (right panel).
104 Figure 3.5
Figure 3. 5: Granisetron inhibition of 5-HT3AR currents.
Trace showing a continuous recording of 5-HT3AR currents (-60mV) in oocytes measured by two electrode voltage-clamp (TEVC) in the presence of serotonin (marked by red line) and pre-applied granisetron (marked by blue line). The effect of granisetron inhibition was fully reversible as seen in the third pulse.
105 Figure 3.6
Figure 3. 6: Cryo-EM structure of granisetron-bound full-length 5-HT3AR.
(A) Map of full-length 5-HT3AR-granisetron reconstructed from 46, 757 particles at 2.92 Å resolution. Side-view parallel to the membrane and extracellular view are shown in left and right panels, respectively. Each monomer is shown in a different color for clarity. Density corresponding to granisetron (left panel, arrow) and glycans (right panel, arrow) are indicated. (B) Three-dimensional model of 5-HT3AR-granisetron structure generated from EM reconstruction (side-view). For each subunit, three sets of glycans are shown as stick representation. (C) Top-view of 5-HT3AR-granisetron map sliced at the binding site to show all five granisetron molecules, each bound at the interface of two subunits (indicated by arrows).
106
Figure 3.7
Figure 3. 7: The granisetron-binding site.
(A) The density map of granisetron contoured at 9σ (left) and map around the residues at the binding site located at the inter-subunit interface (right). The residues on the principal subunit are marked in black and those on the complementary subunit are marked in brown.
(B) A comparison of the 5-HT3AR-apo, 5-HT3AR-granisetron, and 5-HT3AR-serotonin structures shows that residues involved in ligand-binding undergo rotameric reorientation.
(C) Alignment of the three structures reveals an inward motion of loop C in 5-HT3AR- granisetron relative to 5-HT3AR-apo, which is in the direction toward activation as seen in the 5-HT3AR-serotonin structure.
107
Figure 3.8
Figure 3. 8: Differences between granisetron and tropisetron binding poses.
(A) The indazole ring in granisetron lies flat in the binding site allowing for potential cation- pi stacking interaction between Arg65, granisetron and Trp168 in loop F. (B) The indole ring in tropisetron is tilted upwards causing Arg65 to move outward and consequently Trp168 tilts away from the binding site.
108 Figure 3.9
Figure 3. 9: Conformational differences between the apo and ligand-bound states.
(A) A view of the ECD upon global alignment of 5-HT3AR-apo structure with 5-HT3AR- granisetron (left) and 5-HT3AR-serotonin (right). Only two ECD subunits are shown for clarity. A counter clockwise motion of the ECD is observed as indicated by the arrows.
109 The serotonin-induced motion is of larger magnitude compared to that of granisetron, highlighted by the solid and dotted arrows, respectively. (B) A comparison of the TMDs in
5-HT3AR-granisetron structure (left) and 5-HT3AR-serotonin (right) when aligned with respect to 5-HT3AR-apo. Only two TMD subunits are shown for clarity. In both panels a clockwise rotation of the TMD is observed with 5-HT3AR-serotonin revealing a larger change. (C) Pathway of ion permeation of 5-HT3AR-apo and 5-HT3AR-granisetron generated with HOLE [200]. The cartoon representation of two subunits are shown for clarity. The locations of pore constrictions are shown as sticks. The pore radius is plotted as a function of distance along the pore axis. The dotted line indicates the approximate radius of a hydrated Na+ ion which is estimated at 2.76 Å [191] (right).
110 Figure 3.10
Figure 3. 10: Effects of mutations at the ligand-binding pocket.
(A) TEVC recordings (at -60mV) of wild type 5-HT3A receptor and W63Y, R65A and W156Y mutants, expressed in oocytes. Currents were elicited in response to serotonin
(concentrations used near EC50 values of WT and mutants). The following concentrations of serotonin were used: WT 1 μM (reported EC50 ~ 2 μM) [240], W63Y- 10 μM (reported
EC50 ~ 8 μM) [240], R65A- 10 μM (EC50 ~ 14 μM); and W156Y- 200 μM (reported EC50 ~ 194 μM) [240]. Currents were measured in response to serotonin (marked by red line) and pre-application of granisetron (marked by red line). Dotted arrows show the extent of
111 granisetron inhibition. (B) A plot of the ratio of peak current in the presence of granisetron to peak current in the absence of granisetron is shown for WT and mutants. Data are shown as mean ± s.d (n=5). ***, statistically significant. Two sample t-test for wild type and mutants at 95% confidence levels. (C) Granisetron interactions with Trp156 from the principal subunit and Trp63 and Arg65 from the complementary subunit are depicted as stick representation.
112 Figure 3.11
Figure 3. 11: Functional characterization of R65A 5-HT3AR.
Currents were elicited in response to application of various concentration of serotonin with a holding membrane potential of -60 mV. The dose-response plot was generated by normalizing the peak current amplitude to that measured at 100 µM. The curve is a fit to the Hill equation yielding an EC50 of 13.8 + 0.5 µM and nH of 4.4 + 0.5 for n = 4. The error bars are standard deviations.
113
Chapter 4
A commentary on the relevance of a membrane environment for ion channel gating
Portions of this chapter were published in: Gicheru Y, and Chakrapani S. Direct visualization of ion-channel gating in a native environment. PNAS 2018 115 (41) 10198-10200. doi.org/10.1073/pnas.1814277115
114
Foreword
The lipid membrane environment has been shown experimentally to be critical for channel function. This chapter is a commentary highlighting conformational changes evident in a pLGIC in a membrane environment. Gating in this channel was imaged by high speed atomic force microscopy and some of the changes observed during gating have not been described previously in the canonical pLGIC gating cycle.
The commentary was written for the article: Structural titration of receptor ion channel GLIC gating by HS-AFM by Yi Ruan et al. PNAS 2018.
115 Introduction
Pentameric ligand-gated ion channels (pLGICs), also known as Cys- loop receptors, are localized primarily in the postsynaptic membranes, and mediate fast chemical transmission in the central and peripheral nervous systems. Binding of neurotransmitter activates these receptors, causing changes in postsynaptic membrane potential and consequently modulation of neuronal or muscle activity. pLGIC functions are altered by a variety of drugs, making them significant pharmaceutical targets. Indeed, the rise in pLGIC crystal and cryo-electron microscopy structures underscores the magnitude of research efforts that have gone into unraveling the molecular details of channel function. Current structure determination methods of pLGICs capture stationary images and, most often, in nonnative environments. As a result, gaps remain in our understanding of the dynamic properties, intermediate conformational states, and the energetics of gating transitions. In PNAS, Ruan et al. [246] seek to probe some of these missing links using high-speed atomic force micros- copy (HS-AFM) by directly visualizing the pLGIC gating process in a membrane environment and under conditions that mimic physiological buffer, temperature, and pressure.
The Cys-loop family in vertebrates includes the cationic acetylcholine receptor (nAChR) and serotonin receptor (5HT3AR) and the anionic γ- aminobutyric acid receptor (GABAAR) and glycine receptor (GlyR). Homologs of pLGICs include invertebrate members [20], as well as a growing number of prokaryotic members, including the Erwinia chrysanthemi ligand-gated ion
116 channel (ELIC) and Gloeobacter violaceus ligand-gated ion channel (GLIC) [6].
In general, pLGICs share a conserved architecture (despite low sequence identity) where five identical or homologous subunits are pseudosymmetrically arranged around an ion-conducting pore. Each subunit has an N-terminal extracellular domain (ECD) that binds neurotransmitters, a transmembrane domain (TMD) with four membrane-spanning helices (M1–M4) with M2 helices lining the central pore, and a cytosolic intracellular domain in eukaryotic channels formed by the M3–M4 loop [55].
Impact of membrane environment on channel gating and dynamics
Interestingly, high-resolution pLGIC structures solved thus far are in nonnative (detergent) environments, leading to speculations on the possible alteration of channel structures in the absence of membranes. Membrane composition has been shown to be critical for channel gating of the nAChR
[157] and possibly other eukaryotic channels. GLIC has served as an archetypal pLGIC because of the overall conservation of its architecture and pharmacological properties [11]. GLIC crystal structures in the open and resting conformations were the first high-resolution pLGIC structures available [8, 149, 247]. Importantly, the global concerted movements resulting in opening of the channel pore in GLIC are comparable to the gating mechanism emerging from recent eukaryotic pLGIC structures. This conservation is remarkable considering that GLIC is activated by protons rather than a neurotransmitter, and the gating kinetics are approximately two orders-of- magnitude slower in comparison.
117 The crystal structures of GLIC solved at pH 7.0 and pH 4.0 reveal distinct pore profiles at the level of pore- lining M2, clearly revealing changes leading to channel opening [149]. However, conformational changes at the
ECD and most notably at the peripheral TM helices are subtle. This is surprising because global structural movements are implicated in occurring that couple the conformational changes across the ECD and TMD to facilitate gating. Consistent with this idea, electron paramagnetic resonance spectroscopic measurements of membrane-reconstituted GLIC show that the lipid-sensing M4 undergoes an outward motion away from the fivefold axis in transition from the closed to the open state [156]. While GLIC may not be stringent on membrane composition [180] compared with eukaryotic channels, the importance of a membrane environment on channel dynamics cannot be overstated. Crystallographic constraints and cryofreezing in a detergent environment have the potential to mask dynamic properties of channels, making unambiguous mechanistic interpretations a challenging task in the absence of appropriate complementary approaches.
In light of these observations, Ruan et al. [246] employed HS-AFM to probe the conformational changes of GLIC in a near-native membrane environment and physiological buffer conditions. AFM is a tool for imaging biological samples in aqueous solution upon attachment to a substrate. A short cantilever with a probe oscillates vertically near its resonance frequency, briefly touching the surface of the sample, and the force between the sample and the probe is then measured [248]. GLIC was solubilized in detergent and
118 reconstituted at lipid–protein ratios, which produced densely packed vesicles where GLIC molecules were pre- dominantly in an outside-out orientation. The sample was then injected on to an HS-AFM sample support. Titration experiments were performed where the HS-AFM fluid chamber was coupled to a high-precision buffer exchange system starting with pH 3.4 buffer and then pH 7.5 buffer and back to pH 3.4 buffer using multiple syringes. HS-AFM continuously recorded conformational changes from the same ∼70 molecules during the real-time titration. A caveat with contact AFM is the potential for damage as the probe scans the topography of the sample. Ruan et al. are careful about this limitation and explicitly show that the topography of GLIC molecules at the beginning and end of the titration experiments at pH 3.4 was similar, and it is comparable to the open GLIC crystal structure.
Ruan et al. [246] report fascinating reversible supramolecular rearrangements and single-molecule conformational changes of GLIC channels during pH titration. It is important to note that HS-AFM images present only a top view of the channels and any implied changes in the
TMD can only be alluded to, perhaps, based on the dramatic modifications in the ECD. Nevertheless, it is very interesting that the supramolecular arrangement of reconstituted channels is akin to the acetylcholine receptor in the Torpedo postsynaptic membrane captured by cryo-electron tomography
[249] in the absence of ligand. The nAChRs are highly clustered at synaptic membranes due to the scaffolding protein rapsyn. However, it would seem that even in an artificial system with channel densities less than what is seen
119 in situ, the overall assembly is still similar, possibly due to the conserved architecture between the channels. GLIC at pH 3.4 showed dense packing with at least six interacting angles of five nearest neighbors arranged in higher- order patterns. Ruan et al. [246] suggest that this pattern of interaction reflects favorable protein–protein contacts denoting positive cooperative gating. A physiological relevance of positive cooperative gating would be in the amplification of postsynaptic signals. Furthermore, the inclination to think that channels gate independently may not be entirely correct, at least for some systems. In fact, intermolecular interactions in gating have been described in a number of ligand and voltage-gated ion-channel types [250-
253]. In the case of ligand-gated ion channels, an allosteric mechanism has been put forward where the conformational changes associated with channel opening of one channel will influence conformational changes in the neighboring channels perhaps by altering ligand-binding affinity [250].
Through HS-AFM Ruan et al. [246] show for the first time direct imaging of allosteric changes spatio-temporally. Videos are recorded for about an hour and in that time frame we see changes in the ECD and lateral diffusion of channels in titrations from low to high pH buffers.
New insights from noncanonical domain changes?
A question that remains to be answered, however, is whether the reversible conformational changes evident are functionally coupled to gating. At the single- molecule level, titrations with pH 3.4 buffer show channels with a diameter akin to the activated GLIC structure [8]. At this low pH, HS-AFM shows the ECDs of
120 individual GLIC subunits exhibiting fivefold symmetry with a visible pore that is similar to other active/desensitized pLGICs. As the titration changes to pH
7.5, the channels have collapsed ECDs that appear to shut the pore. These
ECD changes are accompanied by channels having a significantly larger diameter. Titration back to low pH shows a notable asymmetry of the channels, which Ruan et al. [246] suggest may correspond to an intermediate state where some of the subunits have undergone activating conformational changes and others have not. A pre-active intermediate conformation for GLIC has been described where the ECD undergoes activating changes while the
TMD remains closed to ion permeation [253]. The low resolution of HS-AFM does not allow us to decipher the molecular details of the striking ECD conformational changes and the possible coupling to the TMD, which continue to be of fundamental interest. The increase in channel diameter from the activated/desensitized state to the closed state is a curious observation, which Ruan et al. [246] propose is likely to correspond to an outward movement of the extracellular end of the TM helices. This result is opposite to the current view of pLGIC gating, where the TMD helices (mainly M2 but also
M4 in some structures) expand away from the pore axis from closed to activated conformations [8, 17]. The ECD changes at neutral pH also deviate from the GLIC crystal structure under similar conditions [149] where the
ECDs are upright, and display fivefold symmetry with an apparently visible pore. Overall, these results differ from the existing pLGIC structures, where transition from closed to open conformations shows ECDs that are
121 loosely packed with relatively few intersubunit contacts to more compact arrangements with increased subunit–interface contacts, respectively [92].
In summary, Ruan et al. [246] present several new insightful features of GLIC gating, some of which differ from mechanisms pro- posed based on pLGIC cryoelectron microscopy and crystal structures. A future direction of this study would be to functionally identify mutations and experimental conditions, such as changing lipid compositions, which uncouple agonist binding to gating as was done for the nAChR [254]. Visualization by HS-
AFM of such samples would then provide a clearer picture of what these physical changes mean in light of channel gating.
122
Chapter 5
Discussion and future directions
123 Discussion
The purpose of this work was to investigate the structural basis for the modulation pLGICs by lipids and lipid derivatives. The membrane environment influences a channel’s ability to transition into different states [41, 51, 162, 181,
246]. However, whether lipids directly interact with channels or whether channel modulation happens as a result of changing membrane properties remained an open question. In this thesis, we showed that lipids, specifically polyunsaturated fatty acids (PUFAs) can directly interact with ion channels to modulate function.
Cannabinoids exert their primary effects through cannabinoid (CB) receptors but numerous studies have revealed that cannabinoids mediate some of their effects through cannabinoid receptor-independent mechanisms [207]. We demonstrated inhibition by functional studies and a potential structural basis for this inhibition.
Our work also provided insight into orthosteric site inhibition of the 5-HT3AR by a commonly prescribed competitive antagonist thus laying the foundation for the design of more specific drugs. Taken together, this body of work provides evidence for the direct modulation of pLGICs by PUFAs and cannabinoids. It is possible to envision that free fatty acids released from phospholipids and their metabolites can directly modulate the function of synaptic ion channels.
Studies on PUFAs modulating channel function have been done extensively on the GABAAR and the nAChR. GABAAR peak currents are potentiated and desensitization is enhanced at GABA concentrations lower than 10−4 M and at
DHA concentrations lower than 10−6 M. At higher DHA concentrations the peak currents are significantly reduced with enhanced desensitization still evident. In
124 addition, the γ-subunit confers sensitivity to DHA [124]. In the nAChR, studies demonstrated that free PUFAs applied outside of a patched membrane area can diffuse and modulate the function of receptors under the patched area. In addition,
AA has been shown to reversibly inhibit nAChR currents likely through direct interaction. Further, the level of saturation of the PUFAs has been correlated with the strength of inhibition. Fatty acids with no or one double bond have little effect on ACh responses while two or three double bonds produced partial inhibition and four double bonds as in AA had the most pronounced effect [177]. Forster resonance transfer studies (FRET) have suggested that free fatty acids (FFAs) can displace phospholipids at annular and non-annular sites suggesting that direct ion channel modulation by lipids is possible [158, 165, 255]. In all of these studies, there is a strong indication of direct lipid-protein interaction with potential membrane alteration, but structural evidence for the former is lacking. Our functional studies suggest a fast mode of action of FFAs in enhancing GLIC desensitizing currents. Further, DHA did not have an effect on a non-desensitizing
GLIC mutant suggesting that DHA enhances agonist induced desensitization. The crystal structure of GLIC in complex with DHA is the first structural evidence of a
PUFA directly binding to and causing conformational changes in a pLGIC.
Functionally, AA appears to be similar in its effect to DHA but the higher potency of EPA has yet to be investigated further.
The relevance of this work also extends to our understanding of the previously elusive desensitized conformation. A number of pLGIC structures have been solved in the presence of saturating agonist concentrations and are likely to
125 represent desensitized conformations. The notable features in these structures are that the loop C conformations are ‘closed’ over the ligand binding site and the channel pores are constricted in the -1 ′ or -2′ position which is considered the desensitization gate [19, 25]. The activation gate at the 9′ position in these structures is open enough to allow ion conduction. However, the pore in the GLIC-
DHA desensitized state is constricted from the 9′ to 2′ while the -2′ position is open
[156]. An implication here is that an agonist induced desensitized conformation may be structurally distinct from a desensitized conformation induced through allosteric sites.
Cannabinoids inhibit currents in cationic pLGICs and potentiate anionic pLGIC currents [141, 241]. One study demonstrated that the inhibitory effect of cannabinoids on the 5-HT3AR was dependent on the extent of receptor desensitization [241]. In addition, cannabinoids did not affect the specific binding of radiolabeled GR65630, a 5-HT3AR antagonist indicating that cannabinoids modulate 5-HT3AR function at allosteric sites [256]. Cannabinoid effect was only evident after pre-incubation with the channels. We showed inhibition of 5-HT3AR currents by endo- and phyto- cannabinoids but the effect on desensitization was not apparent. A 3-dimensional reconstruction of the 5-HT3AR bound to THC and 5-
HT showed a potentially activated ECD and TMD and partially occluded lateral portals in the ICD. Density for 5-HT is evident while that for THC has yet to be clearly defined. Further data processing will provide more insight. Nevertheless, this preliminary structure provides a glimpse into the potential structural basis for cannabinoid modulation of pLGICs.
126 The docking predictions of antagonist binding in the orthosteric site of the
5-HT3AR showed binding poses [40, 235] that are significantly different from the recently published 5-HT3AR bound to tropisetron [49] and granisetron (under review). The limited resolution of the tropisetron-bound structure (4.5 Å) precludes the determination of the precise orientation of the ligand and the interacting residues. The granisetron-bound structure was solved at 2.92 Å and at this higher resolution the specific granisetron binding pose and the interacting residues could be modelled. The most important finding of this work is that the conformation of the antagonist bound structure is distinct from the resting conformation. Motions of the ECD and the TMD in the antagonist bound state are along the pathway to activation relative to the resting conformation, but the energy required to reach activation is not realized in the presence of the antagonist. The high resolution of the density map of the ligand and the residues surrounding the binding site can allow for the design of more specific orthosteric antagonists.
Overall this work is the first to provide insights into the possible structural basis for PUFA and cannabinoid modulation of pLGICs. These structures provide a model for the investigation of how allosteric sites can be targeted to modulate 5-
HT3AR and potentially other members of the pLGIC family. Further implications of this work are mainly in the area of drug design of more specific orthosteric inhibitors and new classes of subunit specific allosteric inhibitors that do not have psychoactive properties.
127 Future directions
Detailed functional analyses of the effects of cannabinoids and PUFAs on the function of the 5-HT3AR are needed. Preliminary data showed that peak currents in the presence of PUFAs are reduced with enhanced desensitization.
There does not seem to be a correlation between carbon chain length or number of double bonds with strength of PUFA effect because EPA is right between DHA and AA. A detailed functional characterization of the effects of PUFAs and cannabinoids on the 5-HT3AR may provide insight on IC50s (concentration- response curves) and on activation, deactivation and desensitization time constants. A limitation of two-electrode voltage clamp is temporal resolution.
Solution exchange is typically much slower than in a patch clamp set-up thus information on the kinetics of gating cannot be accurately determined using the
TEVC set-up. A patch-clamp set up would therefore be more applicable for these experiments to arrive at conclusive results.
Structural studies by Cryo-EM can be done on PUFAs and cannabinoids in complex with the 5-HT3AR and the GlyR. The binding site (TM proper or protein- lipid interface?) of PUFAs on eukaryotic pLGIC may provide information on the immediate action of PUFAs compared to cannabinoids. The results presented here suggest a potential THC binding site in the TM domain perhaps explaining why cannabinoids need to be pre-applied to exhibit their effect. The move to legalize medical marijuana makes it imperative for the scientific community to understand the varied interactions that cannabinoids have on other receptors other than CB receptors. Our work contributes to this understanding from a mechanistic
128 perspective on the 5-HT3AR but also the GlyR for future work. The GlyR is a major target for pain transmission and its potentiation by cannabinoids makes it an attractive target. These studies on the GlyR would be conducted similarly to the ones outlined in this work. Further, utilizing a membrane environment such as nanodiscs for these studies may reveal new conformational changes that may not be realized in detergent. Most of the protocols for these experiments are already established although some optimization may be required.
We have solved the structures of the 5-HT3AR in the resting, open and potentially pre-open or desensitized conformations and GLIC in a potentially desensitized conformation [18, 36, 156]. However, the definitive desensitized state of the 5-HT3AR has yet to be solved. Because PUFAs enhance desensitization, it is possible that addition of these modulators in the presence of serotonin could shift the equilibrium towards populating more of a desensitized conformation. It would be informative to see how PUFAs interact with the 5-HT3AR and the structural changes that result. Would the potentially desensitized conformation be similar to the GLIC-DHA pore or similar to the agonist desensitized states?
Alternatively, functional studies suggest that the open conformation of the 5-HT3AR is stabilized by an interaction between His309 in M3 and Glu250 in M2 and mutating these residues (H309S and E250D) resulted in enhanced desensitization
[18]. 5-HT3AR structures containing these mutations in the presence of 5-HT can be solved to capture a potential desensitized state.
Long term studies involve understanding the functional and structural relevance of other subunits of the 5-HT3R and possibly GlyR. Studies show that
129 the heteromeric 5-HT3ABR has higher single channel conductance, different desensitization kinetics, lower calcium permeability and possibly different pharmacology [50, 104]. Questions to be addressed here include: how does the heteromeric receptor assemble and how many ligands are needed to activate the receptor? What is the ligand occupancy of the heteromeric receptor? How do
PUFAs and cannabinoids modulate the 5-HT3ABR functionally? Where do these modulators bind on the 5-HT3ABR? Is there a subunit dependency similar to the
GABAAR? A model of the 5-HT3 B subunit is similar to the A subunit but its MX helix is relatively shorter than in the A subunit. A high resolution structure of the heteromeric channel would begin to provide insights into the differences between the homomeric and heteromeric channels. The homology in subunits for pLGICs makes it challenging to identify stoichiometry in heteromeric receptors during structure determination. To overcome this, antibodies specific to chosen subunits or to the interface between two different subunits are attached to the receptor before vitrification. Additionally, the antibodies allow for screening during expression to determine whether the heteromeric receptor has trafficked to the membrane. Because the A subunit assembles as a stable functional homomer, optimization with the ratios for infecting cells is necessary. A starting point would be to use a 3:1 ratio for the B:A subunit. The excess B subunit will increase the chances of forming AB heteromers and decrease the chances of forming A homomers.
Palonosetron is the only second generation setron that has showed remarkable improvement with delayed emesis because of its longer half-life and
130 possibly additional mechanisms of channel inhibition. Structural studies with palonosetron on the homomeric and heteromeric (AB) receptors would provide insight on the binding mode of this antagonist and potentially any additional binding sites. It is clear that structurally similar antagonists do not bind exactly the same in the orthosteric site as seen with granisetron and tropisetron. A recent study showed that 5-HT3AC receptors had significantly more serotonin-induced responses compared to the homomeric receptor (A). Additionally, ondansetron and palonosetron inhibited the 5-HT3ACR similar to the 5-HT3AR although a 10 min wash off regained more current in the 5-HT3ACR than the 5-HT3AR. Heteromeric receptors containing A,C,E subunits (5-HT3ACER) had a smaller activation response by serotonin and inhibition by ondansetron and palonosetron was complete without regaining of currents after wash-off [257]. The functional properties of the 5-HT3ACER beg similar questions raised earlier about receptor assembly, ligand occupancy, activation and inhibition. These results suggests that the availability of binding sites may influence the effect of setrons based on where different heteromers are expressed. Electrophysiology and structural characterization of these receptor combinations would provide answers to these questions. Expressing three subunits may require the use of stable inducible cell lines that have been optimized to express all three subunits and form functional heteromers. Subunit specific antibodies maybe used to check for expression.
Additionally, subunit specific antibodies may be used to help with randomizing particle orientation and differentiating between subunits. There is tissue-specific distribution of 5-HT3R subunits and because they require co-assembly with the A
131 subunit, studies with these variable receptors are necessary to provide insight into the function of these channels.
In summation, long term studies with eukaryotic pLGICs point to understanding homomeric and heteromeric receptors in a relevant environment with an aim to first understand functional modulation and second structural features that bring about conformational changes at both orthosteric and allosteric sites.
132 Appendix 1. Portions of Basak, Schmandt, Gicheru et al, eLife 2017;6:e23886 doi: 10.7554/eLife.23886, were reproduced in Chapter 2. License to Use Journal Articles and Related Content Unless otherwise indicated, the articles and journal content published by eLife on the eLife Sites are licensed under a Creative Commons Attribution license (also known as a CC-BY license). This means that you are free to use, reproduce and distribute the articles and related content (unless otherwise noted), for commercial and noncommercial purposes, subject to citation of the original source in accordance with the CC-BY license. https://creativecommons.org/licenses/by/4.0/ 2. Portions of Basak, Gicheru et al., (under review, Nature Communications)……were reproduced in Chapter 3.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
3. Portions of Gicheru and Chakrapani, 2018:115;(41):10198-10200; https://doi.org/10.1073/pnas.1814277115 The author(s) retains copyright to individual PNAS articles, and the National Academy of Sciences of the United States of America (NAS) holds copyright to the collective work and retains an exclusive License to Publish these articles, except for open access articles submitted beginning September 2017.
Authors retain the following rights: The right to include the article in the author’s thesis or dissertation. https://www.pnas.org/page/authors/licenses
133 References
1. Langley, J.N., On the reaction of cells and of nerve-endings to certain poisons, chiefly as regards the reaction of striated muscle to nicotine and to curari. J Physiol, 1905. 33(4-5): p. 374-413. 2. del Castillo, J. and B. Katz, On the localization of acetylcholine receptors. The Journal of Physiology, 1955. 128(1): p. 157-181. 3. Tanabe, T., et al., Primary structure of beta subunit precursor of calf muscle acetylcholine receptor deduced from cDNA sequence. Eur J Biochem, 1984. 144(1): p. 11-7. 4. Boulter, J., et al., Functional expression of two neuronal nicotinic acetylcholine receptors from cDNA clones identifies a gene family. Proc Natl Acad Sci U S A, 1987. 84(21): p. 7763-7. 5. Dent, J.A., The evolution of pentameric ligand-gated ion channels. Adv Exp Med Biol, 2010. 683: p. 11-23. 6. Tasneem, A., et al., Identification of the prokaryotic ligand-gated ion channels and their implications for the mechanisms and origins of animal Cys-loop ion channels. Genome Biol, 2005. 6(1): p. R4. 7. Hilf, R.J.C. and R. Dutzler, Structure of a potentially open state of a proton- activated pentameric ligand-gated ion channel. Nature, 2008. 457: p. 115. 8. Bocquet, N., et al., X-ray structure of a pentameric ligand-gated ion channel in an apparently open conformation. Nature, 2009. 457(7225): p. 111-4. 9. Zimmermann, I. and R. Dutzler, Ligand Activation of the Prokaryotic Pentameric Ligand-Gated Ion Channel ELIC. PLOS Biology, 2011. 9(6): p. e1001101. 10. Hilf, R.J. and R. Dutzler, X-ray structure of a prokaryotic pentameric ligand- gated ion channel. Nature, 2008. 452(7185): p. 375-9. 11. Corringer, P.J., et al., Structure and pharmacology of pentameric receptor channels: from bacteria to brain. Structure, 2012. 20(6): p. 941-56. 12. Unwin, N., Nicotinic acetylcholine receptor at 9 A resolution. J Mol Biol, 1993. 229(4): p. 1101-24. 13. Unwin, N., Refined structure of the nicotinic acetylcholine receptor at 4A resolution. J Mol Biol, 2005. 346(4): p. 967-89. 14. Brejc, K.a., et al., Crystal structure of an ACh-binding protein reveals the ligand-binding domain of nicotinic receptors. Nature, 2001. 411: p. 269. 15. Hansen, S.B., et al., Structures of Aplysia AChBP complexes with nicotinic agonists and antagonists reveal distinctive binding interfaces and conformations. EMBO J, 2005. 24(20): p. 3635-46. 16. Miller, P.S. and A.R. Aricescu, Crystal structure of a human GABAA receptor. Nature, 2014. 512(7514): p. 270-5. 17. Du, J., et al., Glycine receptor mechanism elucidated by electron cryo- microscopy. Nature, 2015. 526(7572): p. 224-9. 18. Basak, S., et al., Cryo-EM reveals two distinct serotonin-bound conformations of full-length 5-HT3A receptor. Nature, 2018. 563(7730): p. 270-274.
134 19. Masiulis, S., et al., GABAA receptor signalling mechanisms revealed by structural pharmacology. Nature, 2019. 20. Thompson, A.J., H.A. Lester, and S.C. Lummis, The structural basis of function in Cys-loop receptors. Q Rev Biophys, 2010. 43(4): p. 449-99. 21. Bartos, M., J. Corradi, and C. Bouzat, Structural basis of activation of cys- loop receptors: the extracellular-transmembrane interface as a coupling region. Mol Neurobiol, 2009. 40(3): p. 236-52. 22. Bouzat, C., et al., Coupling of agonist binding to channel gating in an ACh- binding protein linked to an ion channel. Nature, 2004. 430(7002): p. 896- 900. 23. Ortells, M.O. and G.G. Lunt, Evolutionary history of the ligand-gated ion- channel superfamily of receptors. Trends in Neurosciences, 1995. 18(3): p. 121-127. 24. Monk, S.A., et al., Identification and importance of N-glycosylation of the human 5-hydroxytryptamine3A receptor subunit. Biochemical Pharmacology, 2004. 68(9): p. 1787-1796. 25. Morales-Perez, C.L., C.M. Noviello, and R.E. Hibbs, X-ray structure of the human alpha4beta2 nicotinic receptor. Nature, 2016. 538(7625): p. 411- 415. 26. Hassaine, G., et al., X-ray structure of the mouse serotonin 5-HT3 receptor. Nature, 2014. 512(7514): p. 276-81. 27. Zhu, S., et al., Structure of a human synaptic GABAA receptor. Nature, 2018. 559(7712): p. 67-72. 28. Blount, P. and J.P. Merlie, Mutational analysis of muscle nicotinic acetylcholine receptor subunit assembly. J Cell Biol, 1990. 111(6 Pt 1): p. 2613-22. 29. Sine, S.M. and A.G. Engel, Recent advances in Cys-loop receptor structure and function. Nature, 2006. 440: p. 448. 30. Velisetty, P., S.V. Chalamalasetti, and S. Chakrapani, Structural Basis for Allosteric Coupling at the Membrane-Protein Interface in Gloeobacter violaceus Ligand-gated Ion Channel (GLIC). Journal of Biological Chemistry, 2014. 289(5): p. 3013-3025. 31. Miyazawa, A., Y. Fujiyoshi, and N. Unwin, Structure and gating mechanism of the acetylcholine receptor pore. Nature, 2003. 423: p. 949. 32. Wotring, V.E., T.S. Miller, and D.S. Weiss, Mutations at the GABA receptor selectivity filter: a possible role for effective charges. The Journal of physiology, 2003. 548(Pt 2): p. 527-540. 33. Gunthorpe, M.J. and S.C. Lummis, Conversion of the ion selectivity of the 5-HT(3a) receptor from cationic to anionic reveals a conserved feature of the ligand-gated ion channel superfamily. J Biol Chem, 2001. 276(24): p. 10977-83. 34. Keramidas, A., et al., Ligand-gated ion channels: mechanisms underlying ion selectivity. Prog Biophys Mol Biol, 2004. 86(2): p. 161-204. 35. Bocquet, N., et al., A prokaryotic proton-gated ion channel from the nicotinic acetylcholine receptor family. Nature, 2007. 445(7123): p. 116-9.
135 36. Basak, S., et al., Cryo-EM structure of 5-HT3A receptor in its resting conformation. Nat Commun, 2018. 9(1): p. 514. 37. Laverty, D., et al., Cryo-EM structure of the human α1β3γ2 GABAA receptor in a lipid bilayer. Nature, 2019. 565(7740): p. 516-520. 38. Keramidas, A. and J.W. Lynch, An outline of desensitization in pentameric ligand-gated ion channel receptors. Cell Mol Life Sci, 2013. 70(7): p. 1241- 53. 39. Jackson, M.B., Kinetics of unliganded acetylcholine receptor channel gating. Biophysical journal, 1986. 49(3): p. 663-672. 40. Thompson, A.J., et al., Locating an antagonist in the 5-HT3 receptor binding site using modeling and radioligand binding. J Biol Chem, 2005. 280(21): p. 20476-82. 41. daCosta, C.J., et al., A distinct mechanism for activating uncoupled nicotinic acetylcholine receptors. Nat Chem Biol, 2013. 9(11): p. 701-7. 42. Lee, W.Y., C.R. Free, and S.M. Sine, Binding to gating transduction in nicotinic receptors: Cys-loop energetically couples to pre-M1 and M2-M3 regions. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2009. 29(10): p. 3189-3199. 43. Lester, H.A., The Permeation Pathway of Neurotransmitter-Gated Ion Channels. Annual Review of Biophysics and Biomolecular Structure, 1992. 21(1): p. 267-292. 44. Thompson, A.J. and S.C. Lummis, A single ring of charged amino acids at one end of the pore can control ion selectivity in the 5-HT3 receptor. Br J Pharmacol, 2003. 140(2): p. 359-65. 45. Panicker, S., et al., Evidence for a centrally located gate in the pore of a serotonin-gated ion channel. J Neurosci, 2002. 22(5): p. 1629-39. 46. Filatov, G.N. and M.M. White, The role of conserved leucines in the M2 domain of the acetylcholine receptor in channel gating. Molecular Pharmacology, 1995. 48(3): p. 379. 47. Auerbach, A. and G. Akk, Desensitization of mouse nicotinic acetylcholine receptor channels. A two-gate mechanism. J Gen Physiol, 1998. 112(2): p. 181-97. 48. Monod, J., J. Wyman, and J.P. Changeux, ON THE NATURE OF ALLOSTERIC TRANSITIONS: A PLAUSIBLE MODEL. J Mol Biol, 1965. 12: p. 88-118. 49. Polovinkin, L., et al., Conformational transitions of the serotonin 5-HT3 receptor. Nature, 2018. 563(7730): p. 275-279. 50. Hapfelmeier, G., et al., Co-expression of the 5-HT3B serotonin receptor subunit alters the biophysics of the 5-HT3 receptor. Biophys J, 2003. 84(3): p. 1720-33. 51. daCosta, C.J. and J.E. Baenziger, A lipid-dependent uncoupled conformation of the acetylcholine receptor. J Biol Chem, 2009. 284(26): p. 17819-25. 52. Del Castillo, J. and B. Katz, Quantal components of the end-plate potential. J Physiol, 1954. 124(3): p. 560-73.
136 53. Changeux, J.-P., M. Kasai, and C.-Y. Lee, Use of a Snake Venom Toxin to Characterize the Cholinergic Receptor Protein. Proceedings of the National Academy of Sciences, 1970. 67(3): p. 1241. 54. Brisson, A. and P.N.T. Unwin, Quaternary structure of the acetylcholine receptor. Nature, 1985. 315(6019): p. 474-477. 55. Miyazawa, A., et al., Nicotinic acetylcholine receptor at 4.6 A resolution: transverse tunnels in the channel wall. J Mol Biol, 1999. 288(4): p. 765-86. 56. Walsh, R.M., et al., Structural principles of distinct assemblies of the human α4β2 nicotinic receptor. Nature, 2018. 557(7704): p. 261-265. 57. Cecchini, M. and J.-P. Changeux, The nicotinic acetylcholine receptor and its prokaryotic homologues: Structure, conformational transitions & allosteric modulation. Neuropharmacology, 2015. 96: p. 137-149. 58. Hogg, R.C., M. Raggenbass, and D. Bertrand, Nicotinic acetylcholine receptors: from structure to brain function. Rev Physiol Biochem Pharmacol, 2003. 147: p. 1-46. 59. Kalamida, D., et al., Muscle and neuronal nicotinic acetylcholine receptors. Structure, function and pathogenicity. Febs j, 2007. 274(15): p. 3799-845. 60. Zoli, M., F. Pistillo, and C. Gotti, Diversity of native nicotinic receptor subtypes in mammalian brain. Neuropharmacology, 2015. 96(Pt B): p. 302- 11. 61. Conti-Fine, B.M., et al., Neuronal nicotinic receptors in non-neuronal cells: new mediators of tobacco toxicity? Eur J Pharmacol, 2000. 393(1-3): p. 279- 94. 62. Dani, J.A., Neuronal Nicotinic Acetylcholine Receptor Structure and Function and Response to Nicotine. International review of neurobiology, 2015. 124: p. 3-19. 63. Albuquerque, E.X., et al., Mammalian nicotinic acetylcholine receptors: from structure to function. Physiological reviews, 2009. 89(1): p. 73-120. 64. Nelson, M.E., et al., Alternate stoichiometries of alpha4beta2 nicotinic acetylcholine receptors. Mol Pharmacol, 2003. 63(2): p. 332-41. 65. Lester, H.A., et al., Nicotine is a Selective Pharmacological Chaperone of Acetylcholine Receptor Number and Stoichiometry. Implications for Drug Discovery. The AAPS Journal, 2009. 11(1): p. 167-177. 66. Tapia, L., A. Kuryatov, and J. Lindstrom, Ca2+ Permeability of the (α4)3(β2)2 Stoichiometry Greatly Exceeds That of (α4)2(β2)3 Human Acetylcholine Receptors. Molecular Pharmacology, 2007. 71(3): p. 769-776. 67. Barrantes, F.J., Structural basis for lipid modulation of nicotinic acetylcholine receptor function. Brain Research Reviews, 2004. 47(1): p. 71-95. 68. Brannigan, G., et al., Embedded cholesterol in the nicotinic acetylcholine receptor. Proceedings of the National Academy of Sciences, 2008. 105(38): p. 14418-14423. 69. Hénin, J., et al., A Predicted Binding Site for Cholesterol on the GABAA Receptor. Biophysical Journal, 2014. 106(9): p. 1938-1949.
137 70. Bruno, M.J., R.E. Koeppe, 2nd, and O.S. Andersen, Docosahexaenoic acid alters bilayer elastic properties. Proc Natl Acad Sci U S A, 2007. 104(23): p. 9638-43. 71. Forman, S.A., D.C. Chiara, and K.W. Miller, Anesthetics target interfacial transmembrane sites in nicotinic acetylcholine receptors. Neuropharmacology, 2015. 96(Pt B): p. 169-77. 72. Hurst, R., H. Rollema, and D. Bertrand, Nicotinic acetylcholine receptors: From basic science to therapeutics. Pharmacology & Therapeutics, 2013. 137(1): p. 22-54. 73. Le Novere, N., P.J. Corringer, and J.P. Changeux, The diversity of subunit composition in nAChRs: evolutionary origins, physiologic and pharmacologic consequences. J Neurobiol, 2002. 53(4): p. 447-56. 74. Conti-Fine, B.M., M. Milani, and H.J. Kaminski, Myasthenia gravis: past, present, and future. J Clin Invest, 2006. 116(11): p. 2843-54. 75. Engel, A.G., K. Ohno, and S.M. Sine, Sleuthing molecular targets for neurological diseases at the neuromuscular junction. Nature Reviews Neuroscience, 2003. 4: p. 339. 76. Rapport, M.M., A.A. Green, and I.H. Page, Purification of the substance which is responsible for the vasoconstrictor activity of serum. Fed Proc, 1947. 6(1 Pt 2): p. 184. 77. Walstab, J., G. Rappold, and B. Niesler, 5-HT(3) receptors: role in disease and target of drugs. Pharmacol Ther, 2010. 128(1): p. 146-69. 78. Gaddum, J.H. and Z.P. Picarelli, Two kinds of tryptamine receptor. Br J Pharmacol Chemother, 1957. 12(3): p. 323-8. 79. Thompson, A.J. and S.C. Lummis, 5-HT3 receptors. Curr Pharm Des, 2006. 12(28): p. 3615-30. 80. Kilpatrick, G.J., B.J. Jones, and M.B. Tyers, Identification and distribution of 5-HT3 receptors in rat brain using radioligand binding. Nature, 1987. 330(6150): p. 746-8. 81. Derkach, V., A. Surprenant, and R.A. North, 5-HT3 receptors are membrane ion channels. Nature, 1989. 339(6227): p. 706-9. 82. Maricq, A.V., et al., Primary structure and functional expression of the 5HT3 receptor, a serotonin-gated ion channel. Science, 1991. 254(5030): p. 432. 83. Davies, P.A., et al., The 5-HT3B subunit is a major determinant of serotonin- receptor function. Nature, 1999. 397(6717): p. 359-63. 84. Dubin, A.E., et al., The pharmacological and functional characteristics of the serotonin 5-HT(3A) receptor are specifically modified by a 5-HT(3B) receptor subunit. J Biol Chem, 1999. 274(43): p. 30799-810. 85. Niesler, B., et al., Cloning, physical mapping and expression analysis of the human 5-HT3 serotonin receptor-like genes HTR3C, HTR3D and HTR3E. Gene, 2003. 310: p. 101-11. 86. Holbrook, J.D., et al., Characterisation of 5-HT3C, 5-HT3D and 5-HT3E receptor subunits: evolution, distribution and function. Journal of Neurochemistry, 2009. 108(2): p. 384-396.
138 87. Kapeller, J., et al., Serotonin receptor diversity in the human colon: Expression of serotonin type 3 receptor subunits 5-HT3C, 5-HT3D, and 5- HT3E. J Comp Neurol, 2011. 519(3): p. 420-32. 88. Hubbard, P.C., A.J. Thompson, and S.C.R. Lummis, Functional differences between splice variants of the murine 5-HT3A receptor: possible role for phosphorylation. Molecular Brain Research, 2000. 81(1): p. 101-108. 89. Barnes, N.M., et al., The 5-HT3 receptor--the relationship between structure and function. Neuropharmacology, 2009. 56(1): p. 273-84. 90. Das, P. and G.H. Dillon, The 5-HT3B subunit confers reduced sensitivity to picrotoxin when co-expressed with the 5-HT3A receptor. Molecular Brain Research, 2003. 119(2): p. 207-212. 91. Kelley, S.P., et al., A cytoplasmic region determines single-channel conductance in 5-HT3 receptors. Nature, 2003. 424(6946): p. 321-4. 92. Nemecz, Á., et al., Emerging Molecular Mechanisms of Signal Transduction in Pentameric Ligand-Gated Ion Channels. Neuron, 2016. 90(3): p. 452- 470. 93. Machu, T.K., Therapeutics of 5-HT3 receptor antagonists: current uses and future directions. Pharmacol Ther, 2011. 130(3): p. 338-47. 94. Davies, P.A., Allosteric modulation of the 5-HT3 receptor. Current Opinion in Pharmacology, 2011. 11(1): p. 75-80. 95. Xiong, W., et al., Anandamide inhibition of 5-HT3A receptors varies with receptor density and desensitization. Mol Pharmacol, 2008. 73(2): p. 314- 22. 96. Yang, K.H., et al., The nonpsychoactive cannabinoid cannabidiol inhibits 5- hydroxytryptamine3A receptor-mediated currents in Xenopus laevis oocytes. J Pharmacol Exp Ther, 2010. 333(2): p. 547-54. 97. Faerber, L., et al., The neuronal 5-HT3 receptor network after 20 years of research — Evolving concepts in management of pain and inflammation. European Journal of Pharmacology, 2007. 560(1): p. 1-8. 98. Thompson, A.J. and S.C. Lummis, The 5-HT3 receptor as a therapeutic target. Expert Opin Ther Targets, 2007. 11(4): p. 527-40. 99. Galligan, J.J., Ligand-gated ion channels in the enteric nervous system. Neurogastroenterol Motil, 2002. 14(6): p. 611-23. 100. Minami, M., et al., Pharmacological aspects of anticancer drug-induced emesis with emphasis on serotonin release and vagal nerve activity. Pharmacol Ther, 2003. 99(2): p. 149-65. 101. Higgins, G.A., et al., 5-HT3 receptor antagonists injected into the area postrema inhibit cisplatin-induced emesis in the ferret. British Journal of Pharmacology, 1989. 97(1): p. 247-255. 102. Smith, H.S., L.R. Cox, and E.J. Smith, 5-HT3 receptor antagonists for the treatment of nausea/vomiting. Ann Palliat Med, 2012. 1(2): p. 115-20. 103. Aapro, M.S., Palonosetron as an anti-emetic and anti-nausea agent in oncology. Therapeutics and clinical risk management, 2007. 3(6): p. 1009- 1020.
139 104. Lummis, S.C. and A.J. Thompson, Agonists and antagonists induce different palonosetron dissociation rates in 5-HT(3)A and 5-HT(3)AB receptors. Neuropharmacology, 2013. 73: p. 241-6. 105. Spiller, R.C., Targeting the 5-HT3 receptor in the treatment of irritable bowel syndrome. Current Opinion in Pharmacology, 2011. 11(1): p. 68-74. 106. Bétry, C., et al., Role of 5-HT(3) Receptors in the Antidepressant Response. Pharmaceuticals, 2011. 4(4): p. 603-629. 107. Niesler, B., 5-HT3 receptors: potential of individual isoforms for personalised therapy. Current Opinion in Pharmacology, 2011. 11(1): p. 81- 86. 108. Olsen, R.W. and W. Sieghart, International Union of Pharmacology. LXX. Subtypes of gamma-aminobutyric acid(A) receptors: classification on the basis of subunit composition, pharmacology, and function. Update. Pharmacol Rev, 2008. 60(3): p. 243-60. 109. Macdonald, R.L. and R.W. Olsen, GABAA receptor channels. Annu Rev Neurosci, 1994. 17: p. 569-602. 110. Curtis, D.R., et al., Bicuculline, an antagonist of GABA and synaptic inhibition in the spinal cord of the cat. Brain Research, 1971. 32(1): p. 69- 96. 111. Curtis, D.R. and G.A. Johnston, Amino acid transmitters in the mammalian central nervous system. Ergeb Physiol, 1974. 69(0): p. 97-188. 112. Hill, D.R. and N.G. Bowery, 3H-baclofen and 3H-GABA bind to bicuculline- insensitive GABAB sites in rat brain. Nature, 1981. 290(5802): p. 149-152. 113. Patel, B., M. Mortensen, and T.G. Smart, Stoichiometry of δ subunit containing GABA(A) receptors. Br J Pharmacol, 2014. 171(4): p. 985-94. 114. Farrant, M. and Z. Nusser, Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors. Nature Reviews Neuroscience, 2005. 6: p. 215. 115. Mody, I. and R.A. Pearce, Diversity of inhibitory neurotransmission through GABA(A) receptors. Trends Neurosci, 2004. 27(9): p. 569-75. 116. Sigel, E. and M.E. Steinmann, Structure, function, and modulation of GABA(A) receptors. J Biol Chem, 2012. 287(48): p. 40224-31. 117. Lobo, I.A. and R.A. Harris, GABAA receptors and alcohol. Pharmacology Biochemistry and Behavior, 2008. 90(1): p. 90-94. 118. Olsen, R.W., Allosteric ligands and their binding sites define gamma- aminobutyric acid (GABA) type A receptor subtypes. Adv Pharmacol, 2015. 73: p. 167-202. 119. Macdonald, R.L., J.Q. Kang, and M.J. Gallagher, GABAA Receptor Subunit Mutations and Genetic Epilepsies, in Jasper's Basic Mechanisms of the Epilepsies, th, et al., Editors. 2012, National Center for Biotechnology Information (US) Michael A Rogawski, Antonio V Delgado-Escueta, Jeffrey L Noebels, Massimo Avoli and Richard W Olsen.: Bethesda (MD). 120. Maguire, J.L., et al., Ovarian cycle-linked changes in GABA(A) receptors mediating tonic inhibition alter seizure susceptibility and anxiety. Nat Neurosci, 2005. 8(6): p. 797-804.
140 121. Verrotti, A., et al., Diagnosis and management of catamenial seizures: a review. Int J Womens Health, 2012. 4: p. 535-41. 122. Belelli, D. and J.J. Lambert, Neurosteroids: endogenous regulators of the GABAA receptor. Nature Reviews Neuroscience, 2005. 6: p. 565. 123. Sigel, E., et al., The major central endocannabinoid directly acts at GABA(A) receptors. Proc Natl Acad Sci U S A, 2011. 108(44): p. 18150-5. 124. Nabekura, J., et al., Functional modulation of human recombinant gamma- aminobutyric acid type A receptor by docosahexaenoic acid. J Biol Chem, 1998. 273(18): p. 11056-61. 125. Tan, K.R., U. Rudolph, and C. Luscher, Hooked on benzodiazepines: GABAA receptor subtypes and addiction. Trends Neurosci, 2011. 34(4): p. 188-97. 126. Young, A.B. and S.H. Snyder, Strychnine binding associated with glycine receptors of the central nervous system. Proc Natl Acad Sci U S A, 1973. 70(10): p. 2832-6. 127. Kirsch, J., Glycinergic transmission. Cell Tissue Res, 2006. 326(2): p. 535- 40. 128. Schmitt, B., et al., The Mr 93,000 polypeptide of the postsynaptic glycine receptor complex is a peripheral membrane protein. Biochemistry, 1987. 26(3): p. 805-811. 129. Pfeiffer, F., D. Graham, and H. Betz, Purification by affinity chromatography of the glycine receptor of rat spinal cord. J Biol Chem, 1982. 257(16): p. 9389-93. 130. Dutertre, S., C.-M. Becker, and H. Betz, Inhibitory glycine receptors: an update. The Journal of biological chemistry, 2012. 287(48): p. 40216- 40223. 131. Betz, H. and B. Laube, Glycine receptors: recent insights into their structural organization and functional diversity. J Neurochem, 2006. 97(6): p. 1600- 10. 132. McCracken, L.M., et al., Glycine receptor alpha3 and alpha2 subunits mediate tonic and exogenous agonist-induced currents in forebrain. Proc Natl Acad Sci U S A, 2017. 114(34): p. E7179-e7186. 133. Huang, X., et al., Crystal structure of human glycine receptor-α3 bound to antagonist strychnine. Nature, 2015. 526: p. 277. 134. Grudzinska, J., et al., The beta subunit determines the ligand binding properties of synaptic glycine receptors. Neuron, 2005. 45(5): p. 727-39. 135. Yang, Z., et al., Stoichiometry and subunit arrangement of alpha1beta glycine receptors as determined by atomic force microscopy. Biochemistry, 2012. 51(26): p. 5229-31. 136. Dutertre, S., et al., Probing the pharmacological properties of distinct subunit interfaces within heteromeric glycine receptors reveals a functional betabeta agonist-binding site. J Neurochem, 2012. 122(1): p. 38-47. 137. Laube, B., et al., Modulation of glycine receptor function: a novel approach for therapeutic intervention at inhibitory synapses? Trends Pharmacol Sci, 2002. 23(11): p. 519-27.
141 138. Wu, F.S., T.T. Gibbs, and D.H. Farb, Inverse modulation of gamma- aminobutyric acid- and glycine-induced currents by progesterone. Molecular Pharmacology, 1990. 37(5): p. 597. 139. Mascia, M.P., T.K. Machu, and R.A. Harris, Enhancement of homomeric glycine receptor function by long-chain alcohols and anaesthetics. Br J Pharmacol, 1996. 119(7): p. 1331-6. 140. Yévenes, G.E. and H.U. Zeilhofer, Molecular sites for the positive allosteric modulation of glycine receptors by endocannabinoids. PloS one, 2011. 6(8): p. e23886-e23886. 141. Hejazi, N., et al., Delta9-tetrahydrocannabinol and endogenous cannabinoid anandamide directly potentiate the function of glycine receptors. Mol Pharmacol, 2006. 69(3): p. 991-7. 142. Xiong, W., et al., A common molecular basis for exogenous and endogenous cannabinoid potentiation of glycine receptors. J Neurosci, 2012. 32(15): p. 5200-8. 143. Harvey, R.J., et al., GlyR alpha3: an essential target for spinal PGE2- mediated inflammatory pain sensitization. Science, 2004. 304(5672): p. 884-7. 144. Thompson, A.J., et al., The pharmacological profile of ELIC, a prokaryotic GABA-gated receptor. Neuropharmacology, 2012. 63(4): p. 761-7. 145. Corringer, P.J., et al., Atomic structure and dynamics of pentameric ligand- gated ion channels: new insight from bacterial homologues. J Physiol, 2010. 588(Pt 4): p. 565-72. 146. Hu, H., et al., Crystal structures of a pentameric ion channel gated by alkaline pH show a widely open pore and identify a cavity for modulation. Proc Natl Acad Sci U S A, 2018. 115(17): p. E3959-e3968. 147. Parikh, R.B., M. Bali, and M.H. Akabas, Structure of the M2 transmembrane segment of GLIC, a prokaryotic Cys loop receptor homologue from Gloeobacter violaceus, probed by substituted cysteine accessibility. J Biol Chem, 2011. 286(16): p. 14098-109. 148. Gonzalez-Gutierrez, G. and C. Grosman, Bridging the Gap between Structural Models of Nicotinic Receptor Superfamily Ion Channels and Their Corresponding Functional States. Journal of Molecular Biology, 2010. 403(5): p. 693-705. 149. Sauguet, L., et al., Crystal structures of a pentameric ligand-gated ion channel provide a mechanism for activation. Proc Natl Acad Sci U S A, 2014. 111(3): p. 966-71. 150. Spurny, R., et al., Pentameric ligand-gated ion channel ELIC is activated by GABA and modulated by benzodiazepines. Proc Natl Acad Sci U S A, 2012. 109(44): p. E3028-34. 151. Weng, Y., et al., Anesthetic Sensitivity of the Gloeobacter violaceus Proton- Gated Ion Channel. Anesthesia & Analgesia, 2010. 110(1): p. 59-63. 152. Nury, H., et al., X-ray structures of general anaesthetics bound to a pentameric ligand-gated ion channel. Nature, 2011. 469: p. 428. 153. Pan, J., et al., Structure of the pentameric ligand-gated ion channel GLIC bound with anesthetic ketamine. Structure, 2012. 20(9): p. 1463-9.
142 154. Sauguet, L., et al., Structural basis for potentiation by alcohols and anaesthetics in a ligand-gated ion channel. Nature Communications, 2013. 4: p. 1697. 155. Velisetty, P. and S. Chakrapani, Desensitization Mechanism in Prokaryotic Ligand-gated Ion Channel. Journal of Biological Chemistry, 2012. 287(22): p. 18467-18477. 156. Basak, S., et al., Crystal structure and dynamics of a lipid-induced potential desensitized-state of a pentameric ligand-gated channel. eLife, 2017. 6. 157. Sunshine, C. and M.G. McNamee, Lipid modulation of nicotinic acetylcholine receptor function: the role of neutral and negatively charged lipids. Biochim Biophys Acta, 1992. 1108(2): p. 240-6. 158. Antollini, S.S. and F.J. Barrantes, Unique Effects of Different Fatty Acid Species on the Physical Properties of the Torpedo Acetylcholine Receptor Membrane. Journal of Biological Chemistry, 2002. 277(2): p. 1249-1254. 159. van Meer, G., D.R. Voelker, and G.W. Feigenson, Membrane lipids: where they are and how they behave. Nature Reviews Molecular Cell Biology, 2008. 9: p. 112. 160. Harayama, T. and H. Riezman, Understanding the diversity of membrane lipid composition. Nature Reviews Molecular Cell Biology, 2018. 19: p. 281. 161. Criado, M., H. Eibl, and F.J. Barrantes, Effects of lipids on acetylcholine receptor. Essential need of cholesterol for maintenance of agonist-induced state transitions in lipid vesicles. Biochemistry, 1982. 21(15): p. 3622-3629. 162. daCosta, C.J.B., et al., Lipid-Protein Interactions at the Nicotinic Acetylcholine Receptor: A FUNCTIONAL COUPLING BETWEEN NICOTINIC RECEPTORS AND PHOSPHATIDIC ACID-CONTAINING LIPID BILAYERS. Journal of Biological Chemistry, 2002. 277(1): p. 201- 208. 163. Baenziger, J.E., et al., Effect of Membrane Lipid Composition on the Conformational Equilibria of the Nicotinic Acetylcholine Receptor. Journal of Biological Chemistry, 2000. 275(2): p. 777-784. 164. Ordway, R.W., J.J. Singer, and J.V. Walsh, Direct regulation of ion channels by fatty acids. Trends in Neurosciences, 1991. 14(3): p. 96-100. 165. Antollini, S.S. and F.J. Barrantes, Fatty Acid Regulation of Voltage- and Ligand-Gated Ion Channel Function. Front Physiol, 2016. 7: p. 573. 166. Schmitz, G. and J. Ecker, The opposing effects of n−3 and n−6 fatty acids. Progress in Lipid Research, 2008. 47(2): p. 147-155. 167. Janssen, C.I.F. and A.J. Kiliaan, Long-chain polyunsaturated fatty acids (LCPUFA) from genesis to senescence: The influence of LCPUFA on neural development, aging, and neurodegeneration. Progress in Lipid Research, 2014. 53: p. 1-17. 168. Jeyapal, S., et al., Substitution of linoleic acid with alpha-linolenic acid or long chain n-3 polyunsaturated fatty acid prevents Western diet induced nonalcoholic steatohepatitis. Sci Rep, 2018. 8(1): p. 10953. 169. Hofmanova, J., et al., Interaction of dietary fatty acids with tumour necrosis factor family cytokines during colon inflammation and cancer. Mediators Inflamm, 2014. 2014: p. 848632.
143 170. Bosch-Bouju, C. and S. Layé, Dietary Omega-6/Omega-3 and Endocannabinoids: Implications for Brain Health and Diseases, in Cannabinoids in Health and Disease. 2016. 171. Yoshida, K., et al., Bacterial Long-Chain Polyunsaturated Fatty Acids: Their Biosynthetic Genes, Functions, and Practical Use. Mar Drugs, 2016. 14(5). 172. Xie, X., D. Meesapyodsuk, and X. Qiu, Ketoacylsynthase Domains of a Polyunsaturated Fatty Acid Synthase in Thraustochytrium sp. Strain ATCC 26185 Can Effectively Function as Stand-Alone Enzymes in Escherichia coli. Appl Environ Microbiol, 2017. 83(9). 173. Metz, J.G., et al., Production of polyunsaturated fatty acids by polyketide synthases in both prokaryotes and eukaryotes. Science, 2001. 293(5528): p. 290-3. 174. Hamilton, M.L., et al., Towards the Industrial Production of Omega-3 Long Chain Polyunsaturated Fatty Acids from a Genetically Modified Diatom Phaeodactylum tricornutum. PLoS One, 2015. 10(12): p. e0144054. 175. Hoshi, T., et al., Mechanism of the modulation of BK potassium channel complexes with different auxiliary subunit compositions by the omega-3 fatty acid DHA. Proc Natl Acad Sci U S A, 2013. 110(12): p. 4822-7. 176. Miller, B., et al., Potentiation of NMDA receptor currents by arachidonic acid. Nature, 1992. 355(6362): p. 722-725. 177. Vijayaraghavan, S., et al., Arachidonic acid as a possible negative feedback inhibitor of nicotinic acetylcholine receptors on neurons. J Neurosci, 1995. 15(5 Pt 1): p. 3679-87. 178. Sogaard, R., et al., GABA(A) receptor function is regulated by lipid bilayer elasticity. Biochemistry, 2006. 45(43): p. 13118-29. 179. Bouzat, C., et al., Mutations at Lipid-Exposed Residues of the Acetylcholine Receptor Affect Its Gating Kinetics. Molecular Pharmacology, 1998. 54(1): p. 146. 180. Carswell, C.L., J. Sun, and J.E. Baenziger, Intramembrane aromatic interactions influence the lipid sensitivities of pentameric ligand-gated ion channels. J Biol Chem, 2015. 290(4): p. 2496-507. 181. Henault, C.M., et al., The role of the M4 lipid-sensor in the folding, trafficking, and allosteric modulation of nicotinic acetylcholine receptors. Neuropharmacology, 2015. 96(Pt B): p. 157-68. 182. Carswell, C.L., et al., Role of the Fourth Transmembrane alpha Helix in the Allosteric Modulation of Pentameric Ligand-Gated Ion Channels. Structure, 2015. 23(9): p. 1655-1664. 183. Labarca, C., et al., Channel gating governed symmetrically by conserved leucine residues in the M2 domain of nicotinic receptors. Nature, 1995. 376(6540): p. 514-6. 184. Revah, F., et al., Mutations in the channel domain alter desensitization of a neuronal nicotinic receptor. Nature, 1991. 353(6347): p. 846-9. 185. Yakel, J.L., et al., Single amino acid substitution affects desensitization of the 5-hydroxytryptamine type 3 receptor expressed in Xenopus oocytes. Proc Natl Acad Sci U S A, 1993. 90(11): p. 5030-3.
144 186. Gonzalez-Gutierrez, G., et al., Gating of the proton-gated ion channel from Gloeobacter violaceus at pH 4 as revealed by X-ray crystallography. Proc Natl Acad Sci U S A, 2013. 110(46): p. 18716-21. 187. Wang, H.L., X. Cheng, and S.M. Sine, Intramembrane proton binding site linked to activation of bacterial pentameric ion channel. J Biol Chem, 2012. 287(9): p. 6482-9. 188. Rienzo, M., S.C. Lummis, and D.A. Dougherty, Structural requirements in the transmembrane domain of GLIC revealed by incorporation of noncanonical histidine analogs. Chem Biol, 2014. 21(12): p. 1700-6. 189. Schmandt, N., et al., A chimeric prokaryotic pentameric ligand-gated channel reveals distinct pathways of activation. J Gen Physiol, 2015. 146(4): p. 323-40. 190. Sauguet, L., et al., Structural basis for ion permeation mechanism in pentameric ligand‐gated ion channels. The EMBO Journal, 2013. 32(5): p. 728-741. 191. Marcus, Y., Ionic radii in aqueous solutions. Chemical Reviews, 1988. 88(8): p. 1475-1498. 192. Gielen, M. and P.J. Corringer, The dual-gate model for pentameric ligand- gated ion channels activation and desensitization. J Physiol, 2018. 596(10): p. 1873-1902. 193. Nievas, G.A.F., F.J. Barrantes, and S.S. Antollini, Modulation of Nicotinic Acetylcholine Receptor Conformational State by Free Fatty Acids and Steroids. Journal of Biological Chemistry, 2008. 283(31): p. 21478-21486. 194. Battye, T.G., et al., iMOSFLM: a new graphical interface for diffraction- image processing with MOSFLM. Acta Crystallogr D Biol Crystallogr, 2011. 67(Pt 4): p. 271-81. 195. The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr, 1994. 50(Pt 5): p. 760-3. 196. McCoy, A.J., Solving structures of protein complexes by molecular replacement with Phaser. Acta Crystallogr D Biol Crystallogr, 2007. 63(Pt 1): p. 32-41. 197. Murshudov, G.N., A.A. Vagin, and E.J. Dodson, Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr, 1997. 53(Pt 3): p. 240-55. 198. Emsley, P., et al., Features and development of Coot. Acta Crystallogr D Biol Crystallogr, 2010. 66(Pt 4): p. 486-501. 199. Petrek, M., et al., MOLE: a Voronoi diagram-based explorer of molecular channels, pores, and tunnels. Structure, 2007. 15(11): p. 1357-63. 200. Smart, O.S., et al., HOLE: a program for the analysis of the pore dimensions of ion channel structural models. J Mol Graph, 1996. 14(6): p. 354-60, 376. 201. Mishra, S., et al., Conformational dynamics of the nucleotide binding domains and the power stroke of a heterodimeric ABC transporter. Elife, 2014. 3: p. e02740. 202. Denisov, I.G., et al., Cooperativity in cytochrome P450 3A4: linkages in substrate binding, spin state, uncoupling, and product formation. J Biol Chem, 2007. 282(10): p. 7066-76.
145 203. Sokola-Wysoczanska, E., et al., Polyunsaturated Fatty Acids and Their Potential Therapeutic Role in Cardiovascular System Disorders-A Review. Nutrients, 2018. 10(10). 204. Lu, H.C. and K. Mackie, An Introduction to the Endogenous Cannabinoid System. Biol Psychiatry, 2016. 79(7): p. 516-25. 205. Giuffrida, A., et al., Dopamine activation of endogenous cannabinoid signaling in dorsal striatum. Nature Neuroscience, 1999. 2: p. 358. 206. Stella, N., P. Schweitzer, and D. Piomelli, A second endogenous cannabinoid that modulates long-term potentiation. Nature, 1997. 388: p. 773. 207. Oz, M., Receptor-independent actions of cannabinoids on cell membranes: focus on endocannabinoids. Pharmacol Ther, 2006. 111(1): p. 114-44. 208. Di Marzo, V., M. Bifulco, and L. De Petrocellis, The endocannabinoid system and its therapeutic exploitation. Nat Rev Drug Discov, 2004. 3(9): p. 771-84. 209. Ahrens, J., et al., The nonpsychotropic cannabinoid cannabidiol modulates and directly activates alpha-1 and alpha-1-Beta glycine receptor function. Pharmacology, 2009. 83(4): p. 217-22. 210. Bisogno, T., et al., Molecular targets for cannabidiol and its synthetic analogues: effect on vanilloid VR1 receptors and on the cellular uptake and enzymatic hydrolysis of anandamide. Br J Pharmacol, 2001. 134(4): p. 845- 52. 211. De Petrocellis, L., et al., Plant-derived cannabinoids modulate the activity of transient receptor potential channels of ankyrin type-1 and melastatin type-8. J Pharmacol Exp Ther, 2008. 325(3): p. 1007-15. 212. Mahgoub, M., et al., Effects of cannabidiol on the function of alpha7- nicotinic acetylcholine receptors. Eur J Pharmacol, 2013. 720(1-3): p. 310- 9. 213. Pacher, P., S. Bátkai, and G. Kunos, The endocannabinoid system as an emerging target of pharmacotherapy. Pharmacol Rev, 2006. 58(3): p. 389- 462. 214. Mechoulam, R., L.A. Parker, and R. Gallily, Cannabidiol: an overview of some pharmacological aspects. J Clin Pharmacol, 2002. 42(S1): p. 11s- 19s. 215. Pertwee, R.G., The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: delta9-tetrahydrocannabinol, cannabidiol and delta9- tetrahydrocannabivarin. Br J Pharmacol, 2008. 153(2): p. 199-215. 216. De Petrocellis, L., et al., Effects of cannabinoids and cannabinoid-enriched Cannabis extracts on TRP channels and endocannabinoid metabolic enzymes. Br J Pharmacol, 2011. 163(7): p. 1479-94. 217. Xiong, W., et al., Cannabinoid potentiation of glycine receptors contributes to cannabis-induced analgesia. Nat Chem Biol, 2011. 7(5): p. 296-303. 218. Oz, M., et al., Differential Effects of Endogenous and Synthetic Cannabinoids on α7-Nicotinic Acetylcholine Receptor- Mediated Responses in Xenopus Oocytes. Journal of Pharmacology and Experimental Therapeutics, 2004. 310(3): p. 1152-1160.
146 219. Yang, K.H., et al., The effect of Delta9-tetrahydrocannabinol on 5-HT3 receptors depends on the current density. Neuroscience, 2010. 171(1): p. 40-9. 220. Oz, M., L. Zhang, and M. Morales, Endogenous cannabinoid, anandamide, acts as a noncompetitive inhibitor on 5-HT3 receptor-mediated responses in Xenopus oocytes. Synapse, 2002. 46(3): p. 150-156. 221. Gilmore, J., et al., Recent advances in antiemetics: new formulations of 5HT3-receptor antagonists. Cancer Manag Res, 2018. 10: p. 1827-1857. 222. Costall, B. and R.J. Naylor, 5-HT3 receptors. Curr Drug Targets CNS Neurol Disord, 2004. 3(1): p. 27-37. 223. Machu, T.K. and R.A. Harris, Alcohols and anesthetics enhance the function of 5-hydroxytryptamine3 receptors expressed in Xenopus laevis oocytes. Journal of Pharmacology and Experimental Therapeutics, 1994. 271(2): p. 898-905. 224. Tramer, M.R., et al., Cannabinoids for control of chemotherapy induced nausea and vomiting: quantitative systematic review. Bmj, 2001. 323(7303): p. 16-21. 225. Martin, B.R. and J.L. Wiley, Mechanism of action of cannabinoids: how it may lead to treatment of cachexia, emesis, and pain. J Support Oncol, 2004. 2(4): p. 305-14; discussion 314-6. 226. Slatkin, N.E., Cannabinoids in the treatment of chemotherapy-induced nausea and vomiting: beyond prevention of acute emesis. J Support Oncol, 2007. 5(5 Suppl 3): p. 1-9. 227. Scuderi, C., et al., Cannabidiol in medicine: a review of its therapeutic potential in CNS disorders. Phytother Res, 2009. 23(5): p. 597-602. 228. Mechoulam, R., et al., Cannabidiol--recent advances. Chem Biodivers, 2007. 4(8): p. 1678-92. 229. Pless, S.A. and J.W. Lynch, Ligand-specific Conformational Changes in the α1 Glycine Receptor Ligand-binding Domain. Journal of Biological Chemistry, 2009. 284(23): p. 15847-15856. 230. Chang, Y. and D.S. Weiss, Site-specific fluorescence reveals distinct structural changes with GABA receptor activation and antagonism. Nature Neuroscience, 2002. 5: p. 1163. 231. Yu, R., et al., Agonist and Antagonist Binding in Human Glycine Receptors. Biochemistry, 2014. 53(38): p. 6041-6051. 232. Gilmore, J., et al., Recent advances in antiemetics: new formulations of 5HT(3)-receptor antagonists. Cancer Manag Res, 2018. 10: p. 1827-1857. 233. Joshi, P.R., et al., Interactions of granisetron with an agonist-free 5-HT3A receptor model. Biochemistry, 2006. 45(4): p. 1099-105. 234. Maksay, G., Z. Bikádi, and M. Simonyi, Binding Interactions of Antagonists with 5‐Hydroxytryptamine3A Receptor Models. Journal of Receptors and Signal Transduction, 2003. 23(2-3): p. 255-270. 235. Yan, D. and M.M. White, Spatial orientation of the antagonist granisetron in the ligand-binding site of the 5-HT3 receptor. Mol Pharmacol, 2005. 68(2): p. 365-71.
147 236. Kesters, D., et al., Structural basis of ligand recognition in 5-HT3 receptors. EMBO Rep, 2013. 14(1): p. 49-56. 237. Yan, D., et al., Structural features of the ligand-binding domain of the serotonin 5HT3 receptor. J Biol Chem, 1999. 274(9): p. 5537-41. 238. Duffy, N.H., H.A. Lester, and D.A. Dougherty, Ondansetron and granisetron binding orientation in the 5-HT(3) receptor determined by unnatural amino acid mutagenesis. ACS Chem Biol, 2012. 7(10): p. 1738-45. 239. Ruepp, M.-D., et al., The binding orientations of structurally-related ligands can differ; A cautionary note. Neuropharmacology, 2017. 119: p. 48-61. 240. Spier, A.D. and S.C. Lummis, The role of tryptophan residues in the 5- Hydroxytryptamine(3) receptor ligand binding domain. J Biol Chem, 2000. 275(8): p. 5620-5. 241. Xiong, W., et al., Psychotropic and nonpsychotropic cannabis derivatives inhibit human 5-HT(3A) receptors through a receptor desensitization- dependent mechanism. Neuroscience, 2011. 184: p. 28-37. 242. Laverty, D., et al., Crystal structures of a GABAA-receptor chimera reveal new endogenous neurosteroid-binding sites. Nature Structural &Amp; Molecular Biology, 2017. 24: p. 977. 243. Chen, Z.W., et al., Multiple functional neurosteroid binding sites on GABAA receptors. PLoS Biol, 2019. 17(3): p. e3000157. 244. MacKenzie, D., et al., Localization of binding sites for carboxyl terminal specific anti-rhodopsin monoclonal antibodies using synthetic peptides. Biochemistry, 1984. 23(26): p. 6544-6549. 245. Fezza, F., et al., Endocannabinoids, related compounds and their metabolic routes. Molecules, 2014. 19(11): p. 17078-106. 246. Ruan, Y., et al., Structural titration of receptor ion channel GLIC gating by HS-AFM. Proceedings of the National Academy of Sciences, 2018. 115(41): p. 10333. 247. Hilf, R.J.C. and R. Dutzler, A prokaryotic perspective on pentameric ligand- gated ion channel structure. Current Opinion in Structural Biology, 2009. 19(4): p. 418-424. 248. Ando, T., et al., A high-speed atomic force microscope for studying biological macromolecules. Proceedings of the National Academy of Sciences, 2001. 98(22): p. 12468-12472. 249. Zuber, B. and N. Unwin, Structure and superorganization of acetylcholine receptor-rapsyn complexes. Proc Natl Acad Sci U S A, 2013. 110(26): p. 10622-7. 250. Choi, K.H., Cooperative gating between ion channels. Gen Physiol Biophys, 2014. 33(1): p. 1-12. 251. Keleshian, A.M., et al., Evidence for Cooperativity Between Nicotinic Acetylcholine Receptors in Patch Clamp Records. Biophysical Journal, 2000. 78(1): p. 1-12. 252. Ding, S. and F. Sachs, Evidence for non-independent gating of P2X2 receptors expressed in Xenopus oocytes. BMC Neurosci, 2002. 3: p. 17. 253. Menny, A., et al., Identification of a pre-active conformation of a pentameric channel receptor. eLife, 2017. 6: p. e23955.
148 254. Sun, J., J.F. Comeau, and J.E. Baenziger, Probing the structure of the uncoupled nicotinic acetylcholine receptor. Biochim Biophys Acta, 2017. 1859(2): p. 146-154. 255. Perillo, V.L., et al., The position of the double bond in monounsaturated free fatty acids is essential for the inhibition of the nicotinic acetylcholine receptor. Biochim Biophys Acta, 2012. 1818(11): p. 2511-20. 256. Barann, M., et al., Direct inhibition by cannabinoids of human 5-HT3A receptors: probable involvement of an allosteric modulatory site. Br J Pharmacol, 2002. 137(5): p. 589-96. 257. Yaakob, N.S., et al., The C and E subunits of the serotonin 5-HT3 receptor subtly modulate electrical properties of the receptor. Biomed Pharmacother, 2018. 97: p. 1701-1709.
149