Atomistic Insights into Human Cys-loop Receptors by Solution NMR

Abstract

Cys-loop receptors are pentameric ligand-gated ion channels (pLGICs) mediating fast neurotransmission in the central and peripheral nervous systems. They are important targets for many currently used clinical drugs, such as general anesthetics, and for allosteric modulators with potential therapeutic applications. Here, we provide an overview of advances in the use of solution NMR in structural and dynamic characterization of ion channels, particularly human Cys-loop receptors. We present challenges to overcome and realistic solutions for achieving high-resolution structural information for this family of receptors. We discuss how subtle structural differences among homologous channels define unique channel pharmacological properties and advocate the necessity to determine high-resolution structures for individual receptor subtypes. Finally, we describe drug binding to the Cys-loop receptors’ TMD identified by solution NMR and the associated dynamics changes relevant to channel functions.

1. Introduction

Cys-loop receptors, named after the signature 13-residue loop formed between two conserved cysteine residues, are membrane-spanning ion channels that mediate fast neurotransmission in the central and peripheral nervous systems. This family of receptors includes the nicotinic acetylcholine receptors (nAChRs) and serotonin type-3 receptors (5-HT₃RS) that conduct cations and mediate excitatory neurotransmission. It also includes gamma-aminobutyric acid type-A and type-C receptors (GABA_ARs and GABA_CRs) and glycine receptors (GlyRs) that conduct anions and mediate inhibitory neurotransmission. Malfunction of these receptors is often associated with various neurological disorders, such as epilepsy, depression, cognitive impairment, nicotine and alcohol addiction, congenital myasthenic syndromes, and startle disease [1-3]. Cys-loop receptors are important targets for many currently used clinical drugs, such as general anesthetics, and for potential therapeutics.

Cys-loop receptors are pentameric ligand-gated ion channels (pLGICs) comprised of five identical or homologous subunits arranged around a central channel axis (Fig. 1). Each subunit consists of a neurotransmitter-binding extracellular domain (ECD), a pore-forming transmembrane domain (TMD) containing four transmembrane helices (TM1-TM4), and a large intracellular domain (ICD) connecting TM3 and TM4 [4]. The ICD has been implicated in receptor assembly, trafficking, and localization [5-8]. It may influence channel conductance and desensitization [9-11], but replacing the Cys-loop receptor ICD with a short TM3-TM4 linker still produce functional channels [12, 13]. Neurotransmitter binding to the orthosteric site in the ECD triggers channel opening and allows ions to pass through the cell membrane. Channel activity can also be modulated allosterically by a variety of ligands bound to other regions of these receptors. There is emerging interest in the development of the Cys-loop receptors’ modulators for treating various neurological disorders [14-18].

Fig. 1. Topology of the Cys-loop receptor.

(a) Top view of a Cys-loop receptor showing the quaternary arrangement of the five subunits around the central chamia axis. (b) Side view of the Cys-loop receptor highlighting three domains: ECD (yellow), TMD (green) and ICD (pink). (c) Side view of a single subunit showing the labels for four transmembrane helices, TM1-TM4. The EM structure of the Torpedo marmorata nAChR (PDB ID: 2BG9) was used as the model Cys-loop receptor [4].

Given their biological and pharmacological importance, it is highly desirable to gain a better understanding of Cys-loop receptors’ structures and functions. Four sources of information have contributed to the overall structural understanding of these receptors. The first source consists of crystallographic studies of ECDs, including acetylcholine binding proteins (AChBP) [19-24], the mouse α1-nAChR ECD [25], and an α7nAChR-AChBP chimera [26]. These structures provide valuable information about ligand binding and the resulting conformational changes in the ECD. The second source includes crystal structures of prokaryotic homologues of Cys-loop receptors from Erwinia chrysanthemi (ELIC) in the absence and presence of ligands [27-29], Gloebacter violaceus (GLIC) in an open or closed conformation [30-32], and eukaryotic Caenorhabditis elegans glutamate-gated chloride channel α (GluCl) [33]. The prokaryotic homologues do not contain an ICD and have less than 35% sequence homology with Cys-loop receptors, but share a similar structural scaffold to the ECD and TMD of Cys-loop receptors. The third source of structural information for Cys-loop receptors results from EM data derived from the Torpedo marmorata nAChR, which provided a valuable structural model with a 4-Å resolution on an intact nAChR, but little structural information for the ICD [4], probably due to an intrinsically unstructured nature of the ICD. To date neither EM nor crystallography has been able to provide high-resolution structures for human Cys-loop receptors, even though encouraging results in protein expression and purification have been reported for some Cys-loop receptors in recent years [34-37]. Finally, NMR has provided high-resolution structures for the TMDs of human α1-GlyR [38], α4β2-nAChR [39], and α7-nAChR [40]. Moreover, NMR has provided insights into drug binding and drug-mediated changes in channel dynamics [16, 17, 40-43]. There is no doubt that NMR has established an important position in structure and dynamics determination of Cys-loop receptors and other channel proteins.

In this review, we provide an overview of advances in the use of solution NMR in structural and dynamic characterization of ion channels, particularly human Cys-loop receptors. We discuss challenges and realistic solutions for achieving high-resolution structural information for this family of receptors. We also describe drug binding onto the Cys-loop receptors’ TMD, which is a known target for many Cys-loop receptor modulators.

2. Current Frontier of Solution NMR for Channel Proteins

In the past decade, NMR has emerged as a powerful tool in investigating structures and dynamics of channel proteins at atomic resolution [44, 45]. Although X-ray crystallography has been successfully used to solve numerous channel structures, many systems have not been crystallized or well diffracted even after persistent efforts. Since each channel protein can sample multiple conformations, some discrepancies reflected in NMR and crystal structures from the same protein may be valuable for understanding channel plasticity. Furthermore, NMR can effectively characterize channel motions, in the absence and presence of drug modulation, which are relevant to channel functions. Altogether, NMR is highly valuable and complementary to other tools in providing structural and functional information for channels. This review will focus only on solution NMR’s applications to channel proteins (Table 1), particularly to Cys-loop receptors.

Table 1.

Summary of structures for channels and representative integral membrane proteins determined by solution NMR

Structural Motif	Protein	Oligomerization State	MW (kDa)	PDB Code	Membrane Mimetic	Reference
β-barrel	OmpX	•	16.4	1Q95	DHPC	[48]
	OmpA	•	19.1	1G90	DPC	[49]
	PagP	•	20.2	1MM4	DPC	[125]
	VDAC-1	•	31.9	2K4T	LDAO	[46]
	OmpG	•	32.8	2JQY	DPC	[47]
α-helix	KvaP	• • • •	16.2	2KYH	D7PC	[126]
	M2	• • • •	20.8	2RLF	DHPC	[50]
	PLN	• • • • •	30.5	2KYV	DPC	[61]*
	p7	• • • • • •	40.6	2M6X	DPC	[63]
	KcsA	• • • •	71.2	2A9H	DPC	[103]
	α7-nAChR	• • • • •	73.0	2MAW	LDAO	[40]
	α4-nAChR	• • • • •	75.0	2LLY	LDAO	[39]
	β2-nAChR	• • • • •	75.4	2LM2	LDAO	[39]
	α1-GlyR	• • • • •	86.5	2M6B	LPPG	[38]
α-helix (non- channel)	SRII	•	26.7	2KSY	DHPC	[65]
	PR	•	27.0	2L6X	D7PC	[127]
	DAGK	• • •	39.4	2KDC	DPC	[64]

^*Structure determined by a hybrid solution- and solid-state NMR method.

β-barrels in general have better structural stability and NMR spectral dispersion than α helical proteins. Solution NMR structures for β-barrel channels, including bacterial outer membrane proteins and the voltage dependent anion channel from human mitochondria (VDAC-1), were determined in micelles with protein molecular weights in a range of 16-33 kDa [46-49]. Progress has also been made over the last decade in solution NMR structural determination of α-helical channels. Intensive structural and functional characterization was conducted on the homotetrameric proton channels (M2 channels) of influenza A and influenza B in DHPC micelles, yielding NMR structures for AM2 (~21 kDa) and the BM2 transmembrane (~15 kDa) and cytoplasmic (~41 kDa) domains in different functional states [50, 51]. However, some structural differences of M2 derived from solid state NMR were reported [52, 53]. As a proof of principle, the NMR structure of a “soluble” version of the 46-kDa homo-tetrameric KcsA [54] was determined and proven to agree well with the KcsA crystal structure [55]. The structure of the monomeric subunit of KcsA was also probed in SDS micelles [51]. With the use of solution NMR, several studies revealed K⁺ and pH-dependent conformational and dynamics changes of KcsA [56-60]. The structural dynamics and topology of the homopentameric phospholamban (PLN) channel (~31 kDa) were determined using a combination of solution and solid-state NMR methods to reveal the role of PLN in the regulation of calcium transport [61, 62]. More recently, an unusual architecture of the 42-kDa homo-hexameric p7 channel, the hepatitis C virus viroporin, was revealed using the latest solution NMR technologies [63].

It is also noteworthy that structures of several relatively large non-channel multi-span transmembrane proteins have been determined in micelles using solution NMR, including the 40-kDa homotrimer DAGK [64] and the seven-helix TM receptor sensory rhodopsin II (~27 kDa) [65]. The reported structures for channels and non-channel transmembrane proteins highlight the advances made in using solution NMR for tackling atomic insights into complex proteins. Ironically, they also mark the current upward limit of protein molecular weights amenable for high-resolution NMR structural studies.

3. Use of NMR for Structures and Dynamics of Cys-loop Receptors

3.1 Challenges and solutions for NMR studies of Cys-loop receptors

A full-length Cys-loop receptor has a typical molecular weight of ~250 kDa, which is too large for structure determination using solution NMR. However, individual domains of Cys-loop receptors can be studied separately after proper truncation. This is particularly true for the TMD, which spontaneously assembles into pentameric channels in membrane mimetics. Moreover, one can evaluate whether a TMD channel retains functional and pharmacological characteristics of the authentic Cys-loop receptor before investing effort into structure determination.

3.1.1 Spontaneous pentameric assembly

Electron microscopy images [38] and size exclusion chromatography-multi-angle light scattering (SEC-MALS) [39, 40] revealed a spontaneous pentameric assembly of the Cys-loop receptor TMD in micelles. The ability to form channel assemblies was observed even for the isolated pore-lining TM2 helix of the GlyR [66], GABA_A receptor [67], and nAChR [68, 69]. In contrast, quaternary structural assembly of an isolated ECD is less certain. The isolated ECD of GLIC presents as a hexamer in crystal structure [70] or monomer in solution NMR [71], even though its tertiary structure remains the same as that observed in the pentameric structure of the full-length GLIC [30, 31]. Varying oligomerization states were also observed for the ECD α1-GlyR subunit [72]. The preservation of the native quaternary assembly for an isolated TMD and the destruction observed for the ECD alone suggests that the TMD drives the pentameric assembly of Cys-loop receptors.

3.1.2 Validation of channel functions

It is vital to assure that the channels used for structural studies are functional and retain essential pharmacological characteristics of Cys-loop receptors. Several approaches have been used for functional assessments. First, the NMR magnetization-inversion-transfer (MIT) experiment [73, 74] is convenient for tracking ion permeation through channels. It has been used to characterize Cl⁻ transport through the human α1-GlyR TMD channel reconstituted in vesicles. The MIT experiment demonstrates that the channel conductance could be inhibited by the pore-blocker picrotoxin, a signature for the native pore structure [38]. Second, for Na⁺-conducting channels, such as α4β2- and α7-nAChRs, the Na⁺ flux assay is useful to quickly assess whether the TMD channels are functional [39]. Finally, injection of the α7-nAChR TMD, purified in LDAO micelles, into Xenopus laevis oocytes has yielded functional channels that retain pharmacological features of the native α7-nAChR [40].

These measurements served not only for validation of functional structures, but also for the revelation of the interplay between Cys-loop receptor domains. The channels have a much higher opening probability in the absence of the ECD than in the intact Cys-loop receptors, where the channels are closed until agonist binds to the ECD [75-79]. Based on this observation, we hypothesize that the ECD sets restraints on the intrinsically opened TMD; agonist binding to the ECD relieves the restraints temporarily so that the TMD can rapidly “relax” into its preferred open conformation before transitioning into desensitized conformations.

3.1.3 Optimization of NMR spectra

For a chosen channel protein, the sample preparation and NMR experimental conditions should be optimized in order to obtain adequate quality spectra for structural studies. In particular, one needs to fine-tune protein stability. When ECD or TMD is isolated, the hydrophobic residues at the ECD-TMD interface, protected originally by Van der Waals contacts, become exposed to water and cause protein destabilization. Substituting two or three non-polar residues in the TM2-TM3 linker with hydrophilic residues, as shown in Fig. 2, greatly prevented aggregation of the α4β2- and α7-nAChR TMDs [39, 40]. Mutations of charged residues at the N- and C-termini were also beneficial for preventing aggregation of the nAChR TMDs at a relatively low pH suitable for NMR data collection [39, 40], but such mutations were found unnecessary for the α1-GlyR TMD [38]. In each case the large ICD was replaced with a glycine linker to limit the molecular size and number of peaks in the spectra. While the ICD plays roles in receptor assembly and trafficking, fully functional receptors could be obtained by replacing the Cys-loop receptor ICD with the short TM3-TM4 linker observed in GLIC [12, 13].

Fig. 2. Sequence comparison for α4, β2, and α7 nAChRs and their respective NMR constructs, α4′, β2′, and α7′.

A few point mutations stabilized the channels under the NMR experimental condition. Glutamate mutations (red) at the N- and C-termini lower the isoelectric point and prevent protein precipitation at pH 4.7. Serine mutations (green) introduced to the TM2-TM3 linker increase sample stability in the absence of the ECD. The sequence numbering corresponds to that of the α7 nAChR.

Selecting a proper membrane mimetic is another important step for sustaining stable channel folding for the purpose of acquiring good quality NMR spectra. A large number of membrane mimetic options for solution NMR studies of membrane proteins are available [80-82]. Although bicelles have been used in studies of transmembrane proteins [83, 84], micelles remain a primary choice for solution NMR. The ability of micelles to sustain stability and functions of membrane proteins has been documented [85]. A specific choice of detergent needs to be tested for each channel protein. For example, LDAO is a good choice for the α4β2- and α7-nAChR TMDs [39, 40], while LPPG is better for the α1-GlyR TMD [38]. A molar ratio of detergent to protein concentrations between 150 and 250 is optimal to maintain protein stability and to achieve high quality NMR spectra [38-40].

TROSY-based [86] triple-resonance experiments and TROSY-based NOESY experiments are well suited for the Cys-loop TMD, which typically has ~120 residues. For homo-pentamers, such as the α7-nAChR and α1-GlyR TMD [38, 40], the NMR spectra can benefit greatly from magnetic symmetry, in which five identical polypeptides assemble a channel. For hetero-pentamers, such as the α4β2 nAChR, the isotope labeled α4 subunit was mixed with the unlabeled β2 or vice versa to avoid spectral crowding [39]. Selective labeling is often needed to resolve ambiguities of chemical shift assignments [38]. In addition to chemical shift index [87, 88] and nuclear Overhauser effects (NOEs) [89-92], long-range spatial restraints from paramagnetic relaxation enhancement (PRE) [93], and orientation restraints from residual dipole couplings (RDCs) [83, 94] were also aided in determination of the structures of the TMDs [38-40].

3.2 NMR Structures of Cys-loop Receptors

NMR structures for the full-length TMD of the human α1-GlyR (PDB ID: 2M6I; 2M6B) [38], human α7- (PDB ID: 2MAW) [40], α4- (PDB ID: 2LLY), and β2-nAChRs (PDB ID: 2LM2, 2KSR) [39, 40, 95] represent the first few atomic resolution experimental TMD structures of mammalian Cys-loop receptors. Although a four-α-helical-bundle scaffold is common to TMDs of all Cys-loop receptors, NMR structures reveal atomic details unique to individual channels that can account for their distinct functional and pharmacological properties.

Take the human α1-GlyR as an example. Several structural and dynamic features of the open-channel α1-GlyR structures revealed by NMR [38] could not be observed via homology modeling [96, 97]. The pore-lining TM2 of α1GlyR has a shorter helix (Fig. 3). It is more than two helical turns (9 residues) shorter than TM2 shown in the cryo-EM structure of nAChR [4], and six residues shorter than observed in the NMR structures of α4β2- and α7-nAChRs [39, 40], and seven residues shorter than in crystal structures of ELIC [27, 28], GLIC [30, 31, 98] and GluCl [33]. Moreover, the C-terminal end of the α1-GlyR TM2 shows two conformations in NMR spectra and displays conformational flexibility that is likely related to channel functions [99, 100]. A kink at A288 changes the α1-GlyR TM3 helix axis direction by ~33° (Fig. 3). Such a kink has not been observed in structures for nAChRs or other homologous pLGICs [4, 27, 28, 30, 31, 33, 39, 40, 98]. The shorter TM2 helix with a flexible C-terminal end and the TM3 helical kink are intrinsic to the α1-GlyR. These structural features were observed in the studies of the α1-GlyR TMD using different membrane mimetics and varying numbers of TMD helices [38, 66, 83, 101, 102].

Fig. 3. Structural comparison of Cys-loop receptor TMDs.

Representative solution NMR structures for (a) the α1-GlyR TMD (PDB ID: 2M6I); (b) the α4- (PDB ID: 2LLY) and (c) β2-(PDB ID: 2LM2) nAChRs; and (d) the cryo-EM structure of the Torpedo marmorata nAChR (PDB ID: 2BG9) with the α1 subunit highlighted. In each colored subunit, the TM2 and TM3 helices are labeled 2 and 3, respectively. Note a shorter length of the TM2 helix in α1-GlyR (a) than those in nAChRs (b-d). Also note the unique helical kink in the α1-GlyR TM3 that is absent in nAChRs.

High-resolution NMR structures for human neuronal nAChRs provide insights into unique characteristics of individual subunits. While the NMR structures for α7, α4, and β2 nAChRs share a common scaffold, differences in the lengths and orientations of the TM helices among the nAChR subtypes exist. Such differences can affect intra-subunit cavities, both in size and shape, and ultimately affect drug binding. The α7- and α7β2-nAChRs are a good pair of Cys-loop receptors to underscore the importance of structural individuality in shaping the unique pharmacology for each Cys-loop receptor subtype [42]. Even though the α7 and β2 TMDs are ~50% identical in sequence, a variation in cavity-lining residues and a less than 10° orientation difference of TM2 helix yield disparity for the cavity at the EC end of the TMD: ~180 ų and ~120 ų for β2 and α7, respectively. Unlike the cavity in β2, a smaller cavity in α7 excludes the anesthetic isoflurane from binding. The structural insights offer an explanation as to why α7 is functionally insensitive to isoflurane while α7β2 is hypersensitive [42]. The subtle structural differences, such as those reflected in α7 and β2, are often diminished in homology modeling. Thus, comprehensive structure determination for individual subunits is necessary in order to design therapeutics targeting specific subtypes of Cys-loop receptors.

Chemical shift perturbation experiments identified interacting residues between the α4 and β2 subunits in the heteromeric α4β2-nAChR [39]. Such information was used to generate a pentameric assembly for the α4β2 TMD. We notice that leucine-leucine and leucine-isoleucine inter-subunit contacts make up 31% and 39% of the total inter-subunit contacts, respectively. They are the predominant driving force for the pentameric assembly of the α4β2 channel. In addition, other hydrophobic contacts involving valine or methionine make up 22% of the inter-subunit contacts. The role of leucine zippers in folding and/or oligomerization of other membrane-spanning proteins has been well documented [103, 104]. The importance of a leucine-isoleucine zipper, interlocking of alternating leucine and isoleucine residues between subunits, has also been demonstrated in structures of PLN [61].

Rich leucine-leucine and leucine-isoleucine contacts between subunits are associated with the abundance of leucine and isoleucine residues that constitute ~30% of the residues in the α4β2-nAChR TMD. In contrast, α1-GlyR has ~10% less leucine and isoleucine in its TMD. Thereby, only ~10% and ~5% inter-subunit leucine-leucine and leucine-isoleucine contacts are observed in the α1-GlyR structure [38]. Aromatics, however, play a more significant role in assembly of the α1-GlyR channel [38] than that of the α4β2-nAChR channel, where only one pair of aromatic contacts between α4 and β2 subunits was found [39].

4. Cys-loop Receptor TMDs as Drug Targets

The TMDs of Cys-loop receptors are targets of general anesthetics [41, 42, 105-112], neurosteroids [113, 114], and various positive allosteric modulators (PAMs) and negative allosteric modulators (NAMs) of therapeutic potential [11, 18, 43, 115]. Complementary to photoaffinity labeling [105-112, 116] and crystallography [29, 48, 98, 117] used for identifying drug binding sites in Cys-loop receptors and homologous pLGICs, solution NMR is another effective tool for providing atomic details of drug binding sites [17Xiong, 2011 #157, 40-42]. Furthermore, solution NMR can identify drug-induced changes in protein dynamics [118-120] and provide insight into potential functional consequences due to drug binding [17Xiong, 2011 #157, 42].

A novel binding site for cannabinoids was identified in human GlyRs (Fig. 4a) by using solution NMR titration and NOESY experiments, which demonstrated a direct contact between cannabinoids and the residue S296 in α1-GlyR [16] or in α3-GlyR [17]. In addition, the NMR experiments revealed two coexisting states of binding on a relatively slow exchange time scale or cannabinoid-induced conformation exchange at S296 [17]. When the binding information is mapped onto the NMR structure of the α1-GlyR channel [38], it becomes clear that S296 exposes to lipids and the cannabinoid-binding pocket is involved with both TM3 of GlyR and surrounding lipids. The atomic insight revealed by NMR adds a valuable structural basis in pursuing glycinergic cannabinoids for chronic pain treatment [16, 17]. The cannabinoid-binding pocket found in human α1- or α3-GlyRs has also documented the appropriateness of shallow and lipid-facing pockets for drug binding to Cys-loop receptors.

Fig 4. Drug binding sites in the human α1-GlyR and α4β2-nAChR TMDs.

(a) Left: top view of the GlyR-α1 TMD (PDB ID: 2M6I) highlighting a novel cannabinoid binding site (green) in each subunit; right: side view of the cannabinoid binding site in an α1-GlyR subunit, which involves residue S296 (green) in TM3 and the adjacent lipids (yellow) [16]. (b) NMR-identified halothane binding sites in the α4β2-nAChR TMD. The α4 (PDB ID: 2LLY) and (β2 (PDB ID: 2LM2) subunits are shown in pink and gray, respectively. Halothane binding was observed at both intra-subunit sites (yellow-green) and inter-subunit sites (cyan) between α4 and β2 subunits. Inter-subunit halothane sites were observed exclusively at the EC end of the TMD, while intra-subunit sites were observed at the IC end of the TMD for both α4 and β2. Note that the intra-subunit site at the EC end of the TMD was unique to β2.

Most drug-binding pockets found in the TMDs of Cys-loop receptors and the homologous pLGICs were found more deeply embedded into the proteins. For low affinity drugs, such as volatile general anesthetics, multiple binding sites often exist in one channel. For example, the anesthetic halothane binds to both intra- and inter-subunit sites in the α4β2-nAChR TMD (Fig. 4b). Similar binding sites were also found separately in the muscle type nAChR with the photo-labeled halothane [112], the fluorescence quenching of GLIC by halothane [121], crystal structures of GLIC bound with the anesthetics desflurane or propofol [117], crystal structures of the GLIC F14′A mutant bound with alcohols [48], and the crystal structure of ELIC bound with a derivative of chloroform [29]. The binding at both intra- and inter-subunit sites inhibited these cation-conducting channels except the GLIC F14′A mutant, in which ethanol binding to an inter-subunit site caused potentiation of the channel current [48]. Thus, the relation between binding sites and functional impact varies among individual channels. The binding to an inter-subunit site can potentiate function in one channel, but inhibit function of another.

Channel function is closely associated with channel motion. The drug binding relevant to a channel’s functional impact is often associated with the drug-induced changes in channel dynamics that can be observed in NMR spectra. Take the recent study on the α7- and α7β2-nAChRs as an example (Fig. 5). The α7-nAChR is functionally insensitive to volatile anesthetics [122, 123], such as isoflurane. The presence of β2 makes the α7β2-nAChR hypersensitive to isoflurane [42]. We identified different isoflurane binding sites in α7 and β2 by using NMR experiments [42]. More importantly, we observed that isoflurane induced a transition of several β2 residues from a single peak to double peaks in NMR spectra, which suggested a shift in the motional timescale or conformational equilibrium [124]. Such transitions were observed for both the V22′ residue lining the isoflurane-binding-site and the channel gate residue L9′ in β2. In contrast, similar dynamics changes were not detected in α7. Only isoflurane binding to the β2-specific site could affect the channel dynamics and produce a functional consequence.

Fig. 5. Dynamics as a critical link between drug binding and changes in channel function.

(a) NMR structures of β2 (top) and α7 (bottom) nAChRs show different size of cavities (transparent surface) at the EC end of their respective TMDs. The anesthetic isoflurane (green surface) binds to this cavity in β2, but to a different site in α7. The cavity-lining residue 22′ is shown as blue and red sticks for β2 and α7, respectively, (b) NMR spectra for residue 22′ of β2 (top) and α7 (bottom) in the presence of 0 mM (black), ~1.5 mM (cyan), and ~3.0 mM (magenta) isoflurane. Note that only in β2 was isoflurane-induced peak splitting observed, indicating a dynamics change for the 22′ residue on the μs time-scale. In contrast, the isoflurane-induced peak splitting was not observed in α7. (c) Isoflurane (Iso) can inhibit the current elicited by acetylcholine (ACh) in Xenopus laevis oocytes expressing α7β2 (top), but cannot affect currents in the case of α7 (bottom). The α7β2-V22′M (top) and α7-M22′V (bottom) mutants have reversed functional responses to isoflurane in comparison to their parent channels. The functional sensitivity to isoflurane (c) is correlated to the dynamics modulation by isoflurane (b) on these channels. The currents were recorded by two-electrode voltage clamp held at −60 mV. The vertical and horizontal scales represent 25 nA and 1 min, respectively.

5. Closing Remarks

There are many subtypes of Cys-loop receptors covering a wide range of functions and pharmacology. It remains a great challenge to determine atomic-resolution structures of these receptors. Recent technological advances in NMR, crystallography, and cryo-EM have rapidly improved our ability to study complex channel proteins. Integrated information generated by these tools will further accelerate our understanding of Cys-loop receptors. Solution NMR will continue to provide atomistic insights into channel structures and dynamics, drug binding (especially at low affinities), and allosteric modulation.

Highlights.

• Biological significance and therapeutic potential of Cys-loop receptors

• Challenges and strategies for structure and dynamics determination of Cys-loop receptors

• Structural differences among homologous channels differentiating channel pharmacology

• Drug interaction with Cys-loop receptors revealed by NMR