13058
Biochemistry 2003, 42, 13058-13065
NMR Structure and Dynamics of the Second Transmembrane Domain of the Neuronal Acetylcholine Receptor β2 Subunit† Victor E. Yushmanov,‡ Yan Xu,‡,§ and Pei Tang*,‡,§ Department of Anesthesiology and Department of Pharmacology, UniVersity of Pittsburgh School of Medicine, Pittsburgh, PennsylVania 15261 ReceiVed June 17, 2003; ReVised Manuscript ReceiVed September 15, 2003
Structure and backbone dynamics of a selectively [15N]Leu-labeled 28-residue segment of the extended second transmembrane domain (TM2e) of the human neuronal nicotinic acetylcholine receptor (nAChR) β2 subunit were studied by 1H and 15N solution-state NMR in dodecylphosphocholine micelles. The TM2e structure was determined on the basis of the nuclear Overhauser effects (NOEs) and the hydrogen bond restraints, which were inferred from the presence of HRi-HNi+3, HRi-Hβi+3, and HRi-HNi+4 NOE connectivity and from the slow amide hydrogen exchange with D2O. The TM2e structure of the nAChR β2 subunit contains a helical region between T4 and K22. Backbone dynamics were calculated using the model-free approach based on the 15N relaxation rate constants, R1 and R2, and on the 15N-{1H} NOE. The data acquired at 9.4 and 14.1 T and calculations using different dynamic models demonstrated no conformational exchange and internal motions on the nanosecond time scale. The global tumbling time of TM2e in micelles was 14.4 ( 0.2 ns; the NOE values were greater than 0.63 at 9.4 T, and the order parameter, S2, was 0.83-0.96 for all 15N-labeled leucine residues, suggesting a restricted internal motion. This is the first report of NMR structure and backbone dynamics of the second transmembrane domain of the human nAChR β2 subunit in a membrane-mimetic environment, providing the basis for subsequent studies of subunit interactions in the transmembrane domain complex of the neuronal nAChR. ABSTRACT:
The nicotinic acetylcholine receptor (nAChR)1 is one of the most extensively studied members of a superfamily of neurotransmitter-gated ion channels (1, 2) that are responsible for the fast synaptic transmission in the central and peripheral nervous systems. It is believed that each receptor is composed of five subunits and that each subunit consists of an extended extracellular N-terminus, four transmembrane domains (TM1TM4), and a short C-terminus. On the basis of the studies of photoaffinity labeling, substituted cysteine accessibility, mutagenesis, and electrophysiology analyses, it was suggested that the TM2 domains of five subunits line the pore of the channel (2, 3). The electron microscopic image of a whole Torpedo nAChR channel was obtained at a resolution of 4-9 Å (4, 5). Recently, the crystal structure of the molluscan acetylcholine-binding protein has reached a †
This work was supported by NIH Grant R01 GM56257 to P.T. * To whom correspondence should be addressed: W1357 Biomedical Science Tower, University of Pittsburgh, Pittsburgh, PA 15261. Telephone: (412)383-9798.Fax: (412)648-9587.E-mail:
[email protected]. ‡ Department of Anesthesiology. § Department of Pharmacology. 1 Abbreviations: nAChR, nicotinic acetylcholine receptor; NMR, nuclear magnetic resonance; DPC, dodecylphosphocholine; SDS, sodium dodecyl sulfate; R1 and R2, longitudinal and transverse relaxation rate constants, respectively; NOE, nuclear Overhauser effect; GlyR, glycine receptor; CD, circular dichroism; TM2e, extended TM2 segment of the human neuronal nAChR β2 subunit; 2D, two-dimensional; 5-DSA, 5-DOXYL-stearic acid; DMPC, dimyristoylphosphatidylcholine; NOESY, NOE spectroscopy; TOCSY, total correlation spectroscopy; HSQC, heteronuclear single-quantum coherence; τm, global rotational correlation time; S2, squared order parameter; Rex, exchange rate constant; τe, effective correlation time for fast internal motions; TM1-TM4, transmembrane domains 1-4, respectively.
resolution of 2.7 Å (6). Because of the homology of its sequence to those of the N-terminal domains of the nAChR subunits, this very high resolution structure provides a good structural template for the extracellular domain of nAChRs. High-resolution structures of transmembrane domains of nAChRs, however, are still elusive. To achieve atomic resolution for the structures of transmembrane domains, one of the feasible approaches is to determine the structure of one or two domains at a time by high-resolution nuclear magnetic resonance (NMR) spectroscopy (7). Using this approach, we have recently obtained TM2 structures of the human glycine receptor (GlyR) R1 subunit in membranemimetic dodecylphosphocholine (DPC) and sodium dodecyl sulfate (SDS) micelles (8, 9). The nAChRs comprise a large number of subtypes with different combinations of various subunits. Most of the studies related to nAChRs so far have been focused on the muscle-type receptors derived either from the skeletal muscle or from the Torpedo electric organ. This is mainly due to the relatively widespread availability of this type of nAChR. Isolated transmembrane domains TM2 and TM3 of the Torpedo nAChR R1 subunit in a chloroform/methanol mixture (10, 11) and TM2 of the muscle-type δ subunit in DPC micelles and lipid bilayers (7, 12) have been characterized by NMR. The detailed structural data for transmembrane domains of neuronal nAChRs, however, are currently unavailable. Although it is reasonable to assume that domain structures of muscular and neuronal nAChRs are similar because of their sequence homology, resolving the subtle structural difference between subtypes may lead to a better
10.1021/bi0350396 CCC: $25.00 © 2003 American Chemical Society Published on Web 10/18/2003
Structure and Dynamics of β2 TM2 of the nAChR understanding of the difference in their biochemical and physiological functions. The neuronal nAChRs consist of only two types of subunits (13): an R-type or agonist binding subunit and a β-type or structural subunit. The naming of the R-type neuronal nAChR subunits is based on their substantial homology with the R subunit of the muscle nAChR, including the presence of vicinal cysteine residues (Cys192 and Cys193; R1 numbering) as a flag for the ACh-binding site of muscle-type nAChRs. The muscle R subunit is denoted as R1, and R2-R10 are distinct neuronal R subunits (13, 14). Three members of the neuronal nAChR β-type subunit group, β2-β4, have to date been identified. Unlike the R subunit, the neuronal β subunits lack equivalents of Cys192 and Cys193 and are not highly homologous with the muscle β subunit. Among all neuronal nAChRs, 2R43β2 is a predominant subtype of nAChRs in the central nervous system. The presence of the β2 subunit seems to be required for the anatomical and functional development of the visual system (15) and for the protective responses to stress (16). Overexpression and certain mutations of the β2 subunit have been associated with the development of Down syndrome (17) and nocturnal frontal lobe epilepsy (18). It has also been shown that neuronal heteromeric nAChRs containing β2 subunits are particularly sensitive to volatile general anesthetics (19-21). Site-directed mutagenesis within the TM2 domain has identified a specific amino acid as being critical for the sensitivity of the nAChR to anesthetics (21, 22). Highresolution domain structures of the β2 subunits seem to be essential for gaining insights into the molecular details of these mutation-induced functional changes. It has been generally realized that protein motions are important determinants of the stability and function of proteins (23). Our recent results suggest that low-affinity drugs, such as volatile general anesthetics, may exert their action by altering protein dynamics instead of changing protein structures (24, 25). The theory and techniques for determining protein dynamics using solution NMR have been well developed (23, 26). Overall protein backbone dynamics are often probed by the motions of backbone N-H vectors and detected through parameters that are sensitive to the strength of 15N-1H dipolar interactions and 15N chemical shift anisotropy, including the longitudinal and transverse relaxation rate constants (R1 and R2, respectively) and the 15N-{1H} steady-state nuclear Overhauser effect (NOE) of an amide 15N. We have previously determined the structure and backbone dynamics of the TM2 domain of the human GlyR R1 subunit in SDS and DPC micelles using highresolution NMR (8, 9). We found restricted internal motions in the helical region and a relatively high degree of motional freedom of the TM2 domain of the GlyR R1 subunit in micelles. The dynamics of the nAChR, however, have not been studied previously. Reconstitution of membrane-associated proteins and peptides into micelles of surfactants, such as DPC, is a common approach to determining their structure and dynamics using high-resolution solution-state NMR techniques. The channellining TM2 segments of the muscle nAChR and of the NMDA receptor were studied in DPC micelles and lipid bilayers by solution and solid-state NMR (7). It was shown that, when reconstituted in bilayers, the TM2 segments of
Biochemistry, Vol. 42, No. 44, 2003 13059 the nAChR δ subunit and of the neuronal nAChR R4 subunit alone form homopentameric channels with conductance characteristics very similar to those of a natural receptor (7). It has also been demonstrated recently that pentameric bundles of TM2 segments of the GlyR could form in micelles (8), suggesting that micelles may provide an adequate membrane-mimicking environment for transmembrane channel assembly. Hence, the study of isolated functional segments of the TM2 domains of nAChR in micelles is warranted. In this work, we report structure and backbone dynamics of a peptide segment corresponding to the extended TM2 chain of the human neuronal nAChR β2 subunit (TM2e) using solution-state NMR spectroscopy in DPC micelles. TM2e contains the putative transmembrane segment TM2 and seven residues at the C-terminus that belong to the linker region between TM2 and TM3. Seven leucine residues in TM2e, widely spread along the segment, were selectively labeled with 15N for the 15N NMR measurements of backbone dynamics. The TM2e orientation within membrane bilayers was probed by solid-state NMR spectroscopy. MATERIALS AND METHODS Sample Preparation. The selectively [15N]Leu-labeled, extended TM2 domain of the human neuronal nAChR β2 subunit, ΤΜ2e (EKMTLCISVLLALTVFLLLISKIVPPTS, molecular mass of 3.05 kDa), was obtained by solid-phase synthesis. For reconstitution of the TM2e peptide into DPC micelles, a solution of 4.6-7.7 mg of TM2e in trifluoroethanol was dried into a thin film under a stream of nitrogen gas. An aqueous solution of DPC (Avanti Polar Lipids, Alabaster, AL) was added to reach a DPC/peptide molar ratio of 150. After being vigorously mixed, the sample was lyophilized and rehydrated to final peptide and DPC concentrations of 3 and 450 mM, respectively, with 250 mM NaCl at pH 5 in water containing 10% D2O for deuterium lock in the NMR measurements. For homonuclear twodimensional (2D) 1H NMR, perdeuterated DPC-d38 and a final peptide concentration of 5 mM was used. DPC-d38 and D2O were obtained from Cambridge Isotope Laboratories (Andover, MA). To introduce the lipid nitroxyl spin probe 5-DOXYL-stearic acid (5-DSA, from Aldrich, Milwaukee, WI) into the micelles, an aliquot of a 5-DSA solution in ethanol was evaporated under nitrogen gas before mixing it with TM2e in DPC. The maximum concentration of 5-DSA reached 7.8 mM, i.e., approximately one 5-DSA molecule per micelle. For the samples of TM2e in aligned phospholipid bilayers (7, 27), 7 mg of TM2e, 64 mg of dimyristoylphosphatidylcholine (DMPC), and 8 mg of DMPC-d54 (Avanti Polar Lipids) were dissolved in trifluoroethanol, mixed, sonicated for 5 min, and stored overnight at -20 °C. This solution (with a DMPC:peptide molar ratio of approximately 45) was evenly distributed over 52 glass slides (5.7 mm × 12 mm each), air-dried, and vacuum-dried overnight. MilliQ water was added to each slide (3 µL per slide); the slides were stacked into a parallelepiped glass tube, and equilibrated in a chamber with a saturated solution of ammonium phosphate at 37 °C for at least 15 h. When the desirable degree of sample hydration and alignment was achieved (as monitored by 2H and 31P NMR), the glass tube was sealed.
13060 Biochemistry, Vol. 42, No. 44, 2003 CD Spectroscopy. CD spectra in the wavelength range from 185 to 290 nm were obtained at 25 °C on an Aviv CD spectrometer (model 202, Aviv Instruments, Lakewood, NJ) with a 1 mm cuvette using a wavelength step of 1 nm and averaging time of 1 s, and analyzed using the Web-based CD analysis software DICHROWEB at www.cryst.bbk.ac.uk/ cdweb/html/home.html (28). NMR Spectroscopy. Solution NMR spectra were recorded at 30 °C on a Bruker Avance-600 NMR spectrometer (Bruker Instruments, Billerica, MA) equipped with a triple-resonance, triple-axis gradient TBI probe, and a Chemagnetics CMX400SLI spectrometer (Varian NMR, Inc., Fort Collins, CO) equipped with a Z-gradient broadband Z-SPEC probe (Nalorac Cryogenics Corp., Martinez, CA). Operating frequencies were 600.83 and 401.10 MHz, respectively, for 1H and 60.88 and 40.64 MHz, respectively, for 15N. 15N-decoupled 1H homonuclear 2D NOE spectroscopy (NOESY) and total correlation spectroscopy (TOCSY) data were typically acquired at 600 MHz as 2048 t2 and 640 t1 data points with a spectral width of 10 ppm in each dimension, with 72-80 transients per increment. TOCSY and NOESY mixing times were 60 and 100-150 ms, respectively. All spectra were obtained in the phase-sensitive (States-TPPI) mode. For suppression of the solvent peak, the WATERGATE pulse scheme was applied. Gradient-selected, sensitivity-enhanced 1 H-15N heteronuclear single-quantum correlation (HSQC) data were typically acquired as 2048 t2 and 80 t1 data points, with a spectral width of 12-14 ppm for 1H and 25 ppm for 15 N. To assess the rates of exchange of amide protons, a series of HSQC spectra was acquired 20 min after rehydration of the lyophilized sample with D2O. In 20 h, a NOESY spectrum was acquired to observe the amide protons having very slow 1H-2D exchange rates. Spin-lattice (R1) and spin-spin (R2) 15N relaxation rate constants and 15N-{1H} NOE values were repeatedly measured for each of the 15N amides using standard pulse sequences with Echo-Antiecho gradient selection (29). In relaxation measurements, 80 t1 data points were used with intervals of 1.5 s between scans, and nine variable delays ranging from 20 to 1200 ms at 401 MHz or 1500 ms at 600 MHz for R1, and between 17 and 160 ms for R2. In NOE experiments, 80 indirect data points were acquired with or without proton saturation in an interleaved fashion. Saturation was achieved by a train of 120° pulses at 5 ms intervals for 3 s. Solid-state 15N NMR spectra were obtained at 40.6 MHz on a 89 mm bore Oxford magnet interfaced with a Bruker console at 30 °C using cross polarization with a contact time of 1 ms and a recycle delay of 7 s (27). Glass slides were positioned in the magnet with the glass plate normal parallel to the direction of the static magnetic field. 15N chemical shifts were referenced to 15NH4NO3 at 0 ppm. To monitor phospholipid alignment, solid-state 2H and 31P spectra of the same samples were recorded at 35 °C on a Chemagnetics CMX-400SLI spectrometer at 61.6 and 162.4 MHz, respectively. Data Processing and Analysis. Data were processed using NMRPipe 4.1 and NMRDraw 1.8 (30), and analyzed using PIPP (31) and Sparky 3.100 (32) software. For structure calculations, distance restraints, derived from NOESY experiments, were classified into four groups, i.e.,
