|Home | About | Journals | Submit | Contact Us | Français|
The influence of membrane environment on human cannabinoid 1 (hCB1) receptor transmembrane helix (TMH) conformational dynamics was investigated by solid-state NMR and site-directed spin labeling/EPR with a synthetic peptide, hCB1(T377-E416), corresponding to the receptor’s C-terminal component, i.e., TMH7 and its intracellular α-helical extension (H8) (TMH7/H8). Solid-state NMR experiments with mechanically aligned hCB1(T377-E416) specifically 2H- or 15N-labeled at Ala380 and reconstituted in membrane-mimetic dimyristoylphosphocholine (DMPC) or 1-palmitoyl-2-oleoyl-sn-glycerophosphocholine (POPC) bilayers demonstrate that the conformation of the TMH7/H8 peptide is more heterogeneous in the thinner DMPC bilayer than in the thicker POPC bilayer. As revealed by EPR studies on hCB1(T377-E416) spin-labeled at Cys382 and reconstituted into the phospholipid bilayers, the spin label partitions actively between hydrophobic and hydrophilic environments. In the DMPC bilayer, the hydrophobic component dominates, regardless of temperature. Mobility parameters (ΔH0-1) are 0.3 and 0.73 gauss for the peptide in the DMPC or POPC bilayer environment, respectively. Interspin distances of doubly-labeled hCB1(T377-E416) peptide reconstituted into a TFE/H20 mixture or a POPC or DMPC bilayer were estimated to be 10.6 ± 0.5, 16.8 ± 1, and 11.6 ± 0.8 Å, respectively. The extent of coupling (≥ 50%) between spin labels located at i and i+4 in a TFE/H20 mixture or a POPC bilayer is indicative of an α-helical TMH conformation, whereas the much lower coupling (14%) when the peptide is in a DMPC bilayer suggests a high degree of peptide conformational heterogeneity. These data demonstrate that hCB1(T377-E416) backbone dynamics as well as spin-label rotameric freedom are sensitive to and altered by the peptide’s phospholipid bilayer environment, which exerts a dynamic influence on the conformation of a TMH critical to signal transmission by the hCB1 receptor.
Signal-transducing G protein-coupled receptors (GPCRs)1 are proteins that share as a common defining structural signature an intracellular carboxyl terminus, seven hydrophobic transmembrane helices (TMHs) connected by intra- and extracellular loops, and an extracellular amino terminus (1). Agonist binding is believed to induce ligand-specific GPCR structural changes that allow the receptor to assume, likely through a continuum of intermediate conformations, a functionally active state (2-4). GPCR conformation, activity, and intracellular trafficking may also be modulated by local membrane properties (e.g., tension, fluidity, lateral pressure) in the absence of agonist binding (5, 6). Such ligand-independent GPCR responses imply that the membrane phospholipid bilayer itself may modulate GPCR signal transmission, congruent with the general importance of lipid-protein interactions to membrane protein function (7). Regardless of the agonist- or membrane-related influences involved, however, the molecular mechanisms responsible for GPCR conformational transition, the natures of the resulting receptor structural states, and the relationships between GPCR conformation and the triggering of distinct effector circuitry are largely ill-defined (1, 2). Given that GPCRs represent prime pharmaceutical targets and some 40% of known drugs exert their therapeutic action by interacting with them (1), GPCR conformational plasticity has great biomedical and translational relevance.
Among the GPCRs so far identified, the two main cannabinoid (CB) receptor subtypes, designated CB1 and CB2, have gained intense interest as potential therapeutic targets (8). The CB1 receptor is expressed principally in brain regions that influence mood and reward-related behaviors. The CB2 receptor is largely peripheral, found mainly in immune and white blood cells (9). CB1 and CB2 receptors are activated by a variety of exogenous natural cannabinoids (e.g., the plant cannabinoid Δ9-tetrahydrocannabinol), endogenous cannabinoid agonists (endocannabinoids) (e.g., anandamide and 2-arachidonoylglycerol), and synthetic cannabimimetics, the latter often designed with high selectivity for one particular CB-receptor subtype (8, 10, 11). Depending somewhat upon the activating ligand, these CB receptors primarily couple to inhibitory G proteins (Gi/o) during signal transduction (9). The psychotrophic effects of Δ9-tetrahydrocannabinol are largely mediated through CB1 receptors in the central nervous system (12), and the first-in-class ethical pharmaceutical designed to modulate endocannabinoid-system activity, rimonabant, is a CB1-receptor antagonist (13). Consequently, the CB1 receptor has been targeted by designer small-molecule ligands to treat such pressing medical problems as overweight/obesity and associated cardiometabolic risk, metabolic syndrome, drug addiction, and substance abuse (8, 13, 14).
In order to exploit the CB1 receptor safely and effectively for therapeutic purposes, an operational understanding of the determinants of CB1-receptor conformation is essential. Although rhodopsin’s X-ray crystal structure has enabled homology modeling of CB receptors in lieu of their crystal structures being solved (11, 15-19), the inherent extrapolations involved (16, 20) and the many distinguishing features of CB receptors, including relatively long N- and C-termini implicated in their activity and stability (16-18), mandate direct experimental analysis of CB-receptor structure and dynamics. The CB1-receptor C-terminal region appears to be particularly critical to the conformational transitions leading to receptor activation and the docking of specific G-protein subtypes (15, 19). Here, we describe the conformation and dynamics of a synthetic peptide [hCB1(T377-E416)]2 that corresponds, by rhodopsin homology modeling, to the hCB1 receptor’s C-terminal component, i.e., the hCB1 receptor segment consisting of TMH7 (T377-L399) and its intracellular α-helical extension (H8) (L404-E416) (17-21). TMH7 includes Ser383, a ligand-interaction site that plays a crucial role in hCB1 receptor agonist recognition and activation (17). We have reconstituted hCB1(T377-E416) into mechanically oriented (22, 23) dimyristoylphosphocholine (DMPC) and 1-palmitoyl-2-oleoyl-sn-glycero-phosphocholine (POPC) phospholipid bilayer environments prior to solid-state nuclear magnetic resonance (NMR) and site-directed spin labeling (SDSL)/electron paramagnetic resonance (EPR) spectroscopic analyses. Whereas solid-state NMR spectroscopy has been applied to investigate CB-receptor TMH orientation (22), to the authors’ knowledge, SDSL/EPR has not heretofore been used to probe CB-receptor structure. Using these experimental approaches, we define in this report specific conformational aspects of a CB1-receptor TMH7/H8 peptide sensitive to and dynamically modulated by its membrane phospholipid matrix.
POPC and DMPC in chloroform (Avanti Polar Lipids, Alabaster, AL) were stored at -20 °C prior to use. 2,2,2-Trifluoroethanol (TFE), N-[2-hydroxyethylpiperazine-N’-2-ethane] sulfonic acid (HEPES), EDTA, deuterium-depleted water, and most other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). The spin label, [3-(iodomethyl)-2,2,5,5-tetramethyl-2,5-dihydropyrrol-1-oxyl], was from Toronto Research Chemicals (Toronto, Ontario, Canada). CD3- and 15N-Ala were purchased from Cambridge Isotope Labs (Andover, MA).
The TMH7/H8 peptide 40-mer, hCB1(T377-E416), was produced by standard methodology using an Applied Biosystems synthesizer equipped for 9H-fluoren-9-ylmethoxycarbonyl (Fmoc) chemistry (GenScript Corp., Piscataway, NJ). A CD3- or 15N-Ala was incorporated at position 380 during peptide synthesis. All amino acids were single coupled with a total of 9.5 min for the coupling and monitoring module. The crude peptide was isolated to 95% purity by LC. MALDI-TOF mass spectrometry using an α-cyano-4-hydroxycinnamic acid calibration matrix revealed a peptide m/z of 4583.36.
Samples were prepared for SDSL by resuspending lyophilized hCB1(T377-E416) peptide (95% purity) in a solution of 4 M guanidine-HCl and adding an appropriate amount of 0.1 M stock solution of 3-(iodomethyl)-2,2,5,5-tetramethyl-2,5-dihydropyrrol-1-oxyl to achieve a final molar ratio of 1:3, peptide to spin label. The mixture was incubated overnight at 4 °C. Unreacted spin label was removed by dialysis with a 500-Da cut-off membrane prior to reverse-phase LC peptide purification to 95%. Mass spectrometry confirmed the singly and doubly spin-labeled hCB1(T377-E416) to have m/z = 4636 and 4789, respectively. Nitroxide activity was verified by EPR. The labeling efficiency was estimated by LC-MS to be ~ 90 %. All experiments were conducted from the same stock of labeled peptide.
Samples of hCB1(T377-E416) incorporated in DMPC and POPC bilayers were prepared according to literature protocol, as slightly modified (22-25). Appropriate amounts of hCB1(T377-E416) peptide dissolved in minimal TFE and either DMPC or POPC dissolved in chloroform (20 mg/mL) were mixed to give a 1:100 peptide-to-lipid molar ratio (26). Solvent was removed under a nitrogen stream at ambient temperature, and the dried samples were held overnight in a vacuum desiccator. For NMR, the 2H/15N-labeled powder sample was then packed into a 4-mm ZrO2 rotor. A hydrated sample was prepared by incubating the rotor containing the powder sample in a sealed chamber with saturated ammonium monophosphate (relative humidity ~ 93%) for 6-12 h at 45 °C.
A mechanically aligned sample was prepared by mixing hCB1(T377-E416) with either DMPC or POPC to achieve a 1:100 peptide-to-lipid molar ratio (as above) in a pear-shaped flask. Nitrogen gas was passed through the resulting mixture until the chloroform volume was reduced by two-thirds. The sample was spread onto 25 glass plates (8.5 mm × 14 mm), which were dried overnight in a desiccator (27). Deuterium-depleted water was introduced onto each dried peptide-lipid mixture, and the glass plates were stacked on top of one other. The stack was then placed into an ammonium monophosphate chamber (relative humidity ~ 93%) at 42 °C until the sample became transparent, indicating complete incorporation of hCB1(T377-E416) peptide into either phospholipid bilayer.
All NMR spectra were routinely collected at 35 °C (308 K). Quadrupolar echo pulse sequence (28) was used for the 2H NMR studies with quadrature detection capabilities and complete phase cycling of the pulse pairs. A 3.0 μs 90° pulse, 100 kHz sweep width, 400 ms recycle delay, and 30 μs interpulse delay were used to accumulate 200,000 transients. Prior to Fourier transformation, an exponential multiplication of 200 Hz line broadening was performed on the spectra. 15N NMR spectra were collected utilizing a standard cross polarization pulse sequence with 1H decoupling (29). For the 15N-labeled aligned samples, the following pulse sequence parameters were used: 4.7 s 1H 90°, 500 ppm sweep width, 1.5 ms contact time, and 4 sec recycle delay with 1H decoupling (30-32). The 15N NMR spectra were referenced to an external standard of (15NH4)2SO4 (27 ppm). For 2H- and 15N-labeled, mechanically aligned samples, a double resonance flat coil probe was used operating at the required temperature in an Utrashield Plus AVANCE 700 MHz spectrometer (Bruker Daltonics, Billerica, MA).
EPR spectroscopy was carried out with a Bruker EMX instrument equipped with a high-sensitivity cylindrical cavity and a variable-temperature module. For aqueous samples at 35 °C to 80 °C (308 K to 353 K), spectra were acquired at a microwave frequency of 9.37 GHz, 2.0 mW microwave power, and 0.5 G 100 kHz field modulation amplitude. Rigid limit spectra were acquired under the same conditions, but at 123 °C (150 K), 0.2 mW microwave power, and 1.0 G 100 kHz field modulation amplitude.
Homology modeling of the CB1 receptor has localized Ala3802 to the receptor’s TMH7, close to the phospholipid bilayer/water interface in situ (17, 33). We conducted 2H-NMR studies of hCB1(T377-E416) site-specifically CD3-labeled at Ala380 with the peptide either mechanically aligned normal to the bilayer and parallel to the static magnetic field or unoriented. In the fully hydrated, aligned samples, the quadrupolar splitting of Ala380 was 31.0 ± 0.1 kHz in DMPC bilayers (Figure 1A) and 32.5 ± 0.1 kHz in POPC bilayers (Figure 1B) (means ± variance, n = 3 trials), a mean difference well beyond experimental error (±0.1). For unoriented samples, the rotational motion about the Cα-Cβ Ala bond axis ideally would yield a single Pake doublet characterized by jumps among three sites with a quadrupolar splitting of about 40 kHz. In accord with this expectation, the 2H quadrupole splitting of Ala380 in unoriented, partially hydrated samples of hCB1(T377-E416) in POPC bilayers is indeed 40 kHz (Figure 1C), suggestive of a peptide with restricted side-chain mobility. The resolved Pake doublets of the oriented samples clearly show that the peptide is well aligned in both POPC and DMPC bilayers as compared to the powder spectrum and also indicate that the peptide is embedded within the phospholipid matrix (Figure 1A-1C). The smaller quadrupolar splitting in the hydrated, aligned hCB1(T377-E416) sample is most likely due to increased Ala380 mobility as compared to that in the powder sample. For the TMH7/H8 peptide in the DMPC bilayer, along with the 31-kHz Pake doublet, we also observed peaks from the evidencing quadrupolar splittings of 10.3 and 14.1 kHz not present when the peptide was in a POPC bilayer. We attribute these inner Pake doublets to conformational heterogeneity of the hCB1(T377-E416) peptide in the DMPC membrane environment.
Although the methyl group’s mobility and the three chemically-equivalent deuterons of the CD3-Ala label afford a high signal intensity relative to solid-state 15N NMR (34-36), NMR of site-specifically 15N-Ala-labeled peptides can provide information on their backbone secondary structure and topology (23). For this purpose, we studied by solid-state NMR mechanically aligned, 15N-Ala380-labeled hCB1(T377-E416). The spectrum of the 15N-labeled TMH7/H8 CB1-receptor peptide when aligned in a DMPC bilayer exhibits a rather broad resonance with a chemical shift at 180 ppm (Figure 1D). The spectrum in Figure 1E is that of an aligned hCB1(T377-E416) sample in a POPC bilayer. The relatively sharp resonance peak at 200 ppm is due to the 15N-labeled Ala380 residue. The NMR spectrum in Figure 1F was obtained from an unoriented, 15N-Ala380-labeled TMH7/H8 peptide in a POPC bilayer. The spectrum of unoriented hCB1(T377-E416) spans 215 to 70 ppm, indicating that, in the NMR time scale, the peptide backbone within the membrane is quite rigid. In either a POPC (Figure 1F) or DMPC (not shown) membrane environment, the 15N powder pattern spectra of hCB1(T377-E416) displayed the same widths/chemical shift spans, implying that backbone mobility is identical on the NMR time scale in both environments. The validity of this conclusion is supported by the identical peptide concentrations, molar peptide/lipid ratios, and number of transients for both the DMPC and POPC samples analyzed. The carbonyl-to-carbonyl thickness of the membrane mimetics differed, being about 23 Å for the DMPC bilayer and 27 Å for the POPC bilayer, with a further 10 Å to be added to these figures to account for the choline head groups (37). The effective hydrophobic peptide length of the hCB1(T377-E416) transmembrane region (i.e., T377-L399) (18, 19, 21) was estimated to be 34.5 Å, assuming a perfect α-helical structure (38).
With respect to CB1-receptor TMH7/H8 topology, the 15N-Ala380 resonance peaks of the peptide in either the POPC or DMPC bilayer correspond to the 0° edge of the powder pattern (downfield). This indicates that the 15N-labeled peptide backbone is parallel to the bilayer normal and is embedded within the bilayer. The difference in chemical shift between POPC and DMPC bilayers reflects a change in peptide orientation from one phospholipid environment to the other, a potential response suggested by our previous molecular dynamics simulation and modeling data (38). The envelope of spectral overlap in the narrower DMPC bilayer implies a greater peptide conformational heterogeneity than in the POPC bilayer, as was also indicated by the 2H-NMR data presented above. The same 2H Pake quadrupolar splittings and 15N chemical shifts as shown in Figure 1 for the mechanically aligned samples were obtained at temperatures up to 50 °C (323 K) (data not shown), indicating that the peptide retains its transmembrane alignment at higher temperatures.
In conjunction with SDSL, EPR spectroscopy has emerged as a sensitive method to probe local protein dynamics and conformation (39, 40). Both overall molecular tumbling and local motion of a nitroxide spin label affect EPR lineshape (41). Detailed information about the local dynamics of spin-labeled peptides may be obtained by least-squares analysis of the EPR spectral lineshape of singly-labeled peptides (42). We investigated the local conformational dynamics of the TMH7/H8 peptide within discrete membrane-mimetic phospholipid bilayers by applying SDSL/EPR to hCB1(T377-E416) spin-labeled at a cysteine residue (Cys382) localized by homology modeling (11, 15, 18, 19, 21) near the phospholipid bilayer/water interface. The nitroxide spin label utilized reacts covalently and selectively with cysteine residues by forming a stable thioether bond (Figure 2) (43). Figure 3 shows the EPR spectra at 308 K of the hCB1-receptor TMH7/H8 peptide labeled at Cys382 and reconstituted in a POPC or DMPC bilayer or a TFE/H2O mixture. The spectral lineshapes observed for hCB1(T377-E416) in either TFE/H2O or the POPC bilayer are similar, indicating that the nitroxide spin label is experiencing similar structural environments, consistent with fast tumbling of the polypeptide and reorientation of the flexible nitroxide tether. In contrast, the nitroxide EPR spectrum of the TMH7/H8 peptide in the DMPC bilayer is markedly different, with broadening of the resonance lines and lineshape features, suggestive of a nitroxide tether having comparatively less mobility when hCB1(T377-E416) is in the DMPC rather than the POPC bilayer or in TFE/H2O.
Non-linear least-squares lineshape analysis was used to estimate rotational diffusion parameters with respect to the nitroxide molecular frame (R and R,), isotropic hyperfine spitting (Aiso), and mobility parameter (ΔH0-1) for the TMH7/H8 polypeptide reconstituted in TFE/H2O or in a bilayer of either POPC or DMPC. Two-site fitting was chosen based on spectral features (e.g., peak asymmetry, shoulder features) suggestive of the presence of multiple components (Table 1). The rotational (and more qualitative) mobility parameters support the conclusion from the EPR spectra (above) that the nitroxide spin label experiences similar dynamic environments in TFE/H2O and POPC bilayers; in the DMPC bilayer, the peptide’s mobility is considerably less. In the TFE/H2O mixture, the two sites have equivalent Aiso, as expected from the single hydrophilic phase, whereas the TMH7/H8 peptide in either a POPC or DMPC bilayer evidences two phases with marked differences between the two phospholipid environments. Parameters calculated for spin-labeled hCB1(T377-E416) in the TFE/H2O mixture also support conclusion that the nitroxide is experiencing a low-viscosity hydrophilic solvent, in contrast to the peptide in a POPC bilayer, where it experiences both hydrophilic and hydrophobic environments. The hydrophilic environment likely reflects the proximity of the nitroxide spin label to the phospholipid headgroup/water interface. Conversely, the smaller Aiso values derived from the DMPC sample imply that, in this spectral component, the nitroxide is embedded within the phospholipid bilayer, a disposition restricting both peptide backbone dynamics and spin label rotameric freedom.
The effect of temperature on the dynamics of the nitroxide spin label at position Cys382 of hCB1(T377-E416) in membrane-mimetic phospholipid bilayers was next examined. The EPR spectrum of the TMH7/H8 peptide reconstituted in POPC bilayers shows two temperature-dependent components in the 318-343 K range which are particularly prominent at higher temperatures (Figure 4). We used a two-site fitting to estimate Aiso and extract the relative contribution of each component to the spectral intensity (42). As summarized in Table 2, the results confirm the presence of a hydrophilic phase with an Aiso of 45.1 MHz (“site 1”) and a hydrophobic phase with an Aiso of 43.0 MHz (“site 2”). These data suggest that the local environment of the nitroxide spin label in the hCB1-receptor TMH7/H8 peptide is distributed between the phospholipid/water interface and the hydrophobic POPC lipid core. At lower temperatures, the hydrophilic phase component accounts for 74% of the spectral intensity. As the temperature increases, the fractional hydrophobic component increases, to 85% of the spectral intensity at 343 K. This temperature-dependent behavior is reversible upon sample cooling (data not shown). The EPR spectrum of hCB1(T377-E416) in TFE/H2O evidences minor temperature-induced changes attributable to increased peptide mobility and lower solvent viscosity at higher temperatures (Figure 5A). By comparison, reconstitution of the TMH7/H8 peptide in DMPC bilayers significantly increases spin label mobility at higher temperatures (Figure 5B).
The equilibrium constant for the temperature-dependent hydrophilic/hydrophobic partitioning of the hCB1(T377-E416) TMH7/H8 peptide in POPC bilayers was estimated from the relative populations calculated from the least-squares fits. Figure 6 presents a van’t Hoff plot from which enthalpy (ΔH = +62.7 kJ/mol) and entropy (TΔS = +60 kJ/mol) changes were derived for the hydrophobic-to-hydrophilic environment transition. The positive signs for ΔH and ΔS are consistent with a thermally driven immersion of the polar spin label into the hydrophobic membrane interior, where the value of ΔH reflects the energy required to localize the polar probe in the hydrophobic environment (44, 45)
Spin-spin dipolar interaction in peptides labeled with two spin probes allows accurate extraction from the continuous wave EPR spectrum of interspin separations in the 5 to 20 Å range (46). Spectral reconstruction offers a means to evaluate the broadening from dipolar interaction as well as the percentage of non-interacting component caused by incomplete labeling or misfolding of the labeled peptide, leading to interspin separations beyond the 20 Å detection limit (47). hCB1(T377-E416) contains two unique cysteine residues, Cys382 and Cys386, positioned within a complete turn of i to i + 4 in TMH7 (16, 22, 33). Thus, SDSL/EPR analysis of this peptide double-labeled at these cysteines would allow determination of dipolar interactions between i to i + 4, from which the distance between the two spin labels could be estimated using a dipolar broadening deconvolution method (47). Figure 7 shows the rigid limit EPR spectra for the doubly-labeled CB1-receptor TMH7/H8 peptide in each phospholipid environment studied. hCB1(T377-E416) reconstituted in either TFE/H2O solution or a POPC bilayer exhibits a significantly greater fraction of coupled spins than the peptide reconstituted in a DMPC bilayer (Table 3). Since all samples examined were prepared in parallel from the same labeled TMH7/H8 peptide stock, the drastic reduction in dipolar coupling in the DMPC sample most likely reflects a conformational heterogeneity of the peptide in DMPC that brings the paramagnetic centers out of the 20 Å range observable by the dipolar broadening technique.
The membrane phospholipid bilayer has been shown to influence rhodopsin’s overall helical content and the orientation of its TMHs (4, 48). However, the many distinguishing features of CB receptors preclude general extrapolation of rhodopsin TMH properties and dynamics to them and support direct experimental analysis of CB-receptor structure and conformational properties (15, 16). We have used the well-validated experimental approaches of solid-state NMR and SDSL/EPR to define the conformational responsiveness of hCB1(T377-E416) to its membrane phospholipid environment. Representing the hCB1 receptor segment consisting of TMH7 and its intracellular α-helical extension (H8), the hCB1(T377-E416) peptide is a domain strongly linked to this GPCR’s activity and stability (16, 19, 21).
Differences observed in the 2H and 15N NMR spectra of hCB1(T377-E416) with change of phospholipid (i.e., POPC or DMPC) bilayer matrix suggest that membrane thickness dynamically affects hCB1(T377-E416) structural ordering, the TMH7/H8 peptide displaying greater disorder in the thinner DMPC bilayer than in the thicker POPC bilayer. As evidenced by our double-label EPR data, the high degree of coupling between spin labels located at i and i+4 is strongly indicative of an α-helical structure for hCB1(T377-E416) in TFE/H2O or in a POPC bilayer, whereas the peptide undergoes a conformational adjustment in the thinner DMPC bilayer that causes the interspin distances to exceed the 20 Å detection limit of the dipolar interaction technique. The conformational adjustment is indicative of the sensitivity and adaptability of a critical helical region of the hCB1 receptor to at least one physical property of its phospholipid matrix. Within TMH7, Pro394 is found in the highly-concerved GPCR NP(X)nY motif considered important to GPCR conformational flexibility (1, 2). Molecular dynamics simulations have indeed suggested that a decrease in phospholipid chain-length (i.e., a thinner membrane bilayer) may induce a kink at Pro394 in TMH7 to avoid energetically unfavorable hydrophobic mismatch (38). Other factors including distinctive peptide-lipid interactions, decreased wobbling of the side-chain Ala380, and/or a change in peptide tilt might have contributed to the hCB1(T377-E416) structural differences suggested by the 2H and 15N NMR spectra when the peptide is in the POPC versus the DMPC environment.
In order to characterize further the structural changes in hCB1(T377-E416) in response to its associated phospholipid bilayer that were suggested by our NMR data, we applied the more sensitive, complementary SDSL/EPR technique. EPR spectra of hCB1(T377-E416) spin-labeled at Cys382 are indicative of marked differences in the local dynamics of the TMH7/H8 peptide when reconstituted in TFE/H2O solvent or either POPC or DMPC bilayers. hCB1(T377-E416) evidences fast motional spectral lineshapes in TFE/H2O and POPC which contrast with the more complex slow motional spectrum observed for the peptide in DMPC. EPR analysis of hCB1(T377-E416) reconstituted in a POPC bilayer over the temperature range 308 K to 343 K revealed the presence of two spectral components attributable to a hydrophilic phase predominant at lower temperatures and a hydrophobic phase more pronounced at higher temperatures. Partitioning between hydrophobic and hydrophilic environments suggests that the nitroxide spin label resides near the phospholipid bilayer/water interface. As the temperature is increased, the TMH7/H8 peptide appears to undergo conformational change that positions the spin label in a more hydrophobic environment. This interpretation is supported by the positive ΔH and ΔS values extracted from van’t Hoff analysis of the relative populations. In contrast, the peptide’s nitroxide reporter shows a distinct orientation and dynamic in a DMPC bilayer and does not undergo the same temperature-dependent partitioning that it did in the POPC lipid environment. Analogous EPR experiments have demonstrated that the influence of temperature on the conformational stability of various peptides associated with phospholipid membranes varies widely, from slight (44) to significant (45).
Figure 8 presents a conceptual schematic of hCB1(T377-E416) derived from a previously published molecular dynamics simulation of the peptide in POPC/DMPC bilayers (38) into which the nitroxide spin label at Cys382 has now been placed. In the prior work, Cys382 was placed close to the phospholipid headgroup-water interface, and the tilt angle of TMH7 was determined to be 13 ± 3° in the POPC bilayer and 28 ± 5° in the DMPC bilayer. The angle between TMH7 and H8 was determined to be 89 ± 5° in the POPC bilayer and 119 ± 7° in the DMPC bilayer, the increased tilt proposed to reflect a Pro394 kink in TMH7 when the peptide resides in the thinner DMPC bilayer. The measured 16.8 Å interspin distance between Cys382 and Cys386 (i to i+4) when the hCB1 peptide is in a POPC bilayer is indicative of an extended helix, which justifies positioning the nitroxide reporter group close to the headgroup/water interface. This interspin distance is somewhat greater than that estimated for an ideal α-helical structure in solution (49). The difference may reflect the orientation of the spin label tether in the membrane as compared to the tether orientation in solution, the latter representing the model system from which the estimate was derived. Furthermore, the comparator interspin-distance estimates on doubly-labeled model peptides (49) employed as spin label methanethiosulfonate, which may become ordered on the protein surface as a result of interactions between the δ-sulfur and the protein’s backbone obviated by our use of the [3-(iodomethyl)-2,2,5,5-tetramethyl-2,5-dihydropyrrol-1-oxyl] spin label (41). As the temperature is elevated, the probe more frequently samples the hydrophobic environment due to increased peptide motion and membrane dynamics. In DMPC, the majority of the interspin distances observed for hCB1(T377-E416) are outside the range of the dipolar-broadening technique applied here. However, the measured distance of 11.6 Å, which corresponds to only 14% of peptide population, places the spin probe well within the hydrophobic bilayer. The apparent conformational heterogeneity of hCB1(T377-E416) in DMPC contrasts with the situation when the peptide is in POPC, where more than 60% of the peptide population produces dipolar coupling. We take this as further evidence that the phospholipid bilayer directly affects the conformation of the hCB1(T377-E416) TMH, which in turn appears sensitive and adaptable to its membrane environement. Diverse physical and biochemical characteristics of phospholipid membranes can influence transmembrane protein (including GPCR) structure and function (50). Accordingly, factors other than bilayer thickness/hydrophobic matching may have played a part in the hCB1(T377-E416) structural differences noted between DMPC and POPC, perhaps interactively. For example, the monounsaturated nature of the oleoyl moiety of POPC vs. the fully saturated DMPC fatty-acyl groups means that bilayers composed of these lipids have intrinsically distinct molecular ordering, expansiveness, and fluidity properties that could affect transmembrane protein conformation, lateral diffusion/mobility, interaction, and distribution (50, 51). Although the occurrence of peptide aggregation on the DMPC bilayer surface cannot be excluded as contributing to the relatively lower coupling in that membrane environment (Table 3), the decrease in dipolar interaction observed in DMPC suggests that peptide aggregation is not a significant contributor to the hCB1(T377-E416) structural dynamics documented.
Given the importance of its C-terminal component to the activity and stability of the hCB1 receptor and other GPCRs (2, 16, 19, 21, 52), the results presented have implications for the biological function of the CB1 receptor and, perhaps, GPCRs in general. Conserved proline residues have been implicated in the mobility and signal-transducing capability of various GPCR TMHs (1, 2). Consequently, it may be hypothesized that a kink at Pro394 in the NP(X)nY microdomain of TMH7 (38), which is highly conserved across class A GPCRs (1, 2), could be crucial to the conformational plasticity of the C-terminal segment of the hCB1 receptor we have documented. In this manner, Pro394 may be a component of a sensitive response mechanism that allows the CB1 receptor to adjust its conformation to varied hydrophobic membrane domains during its internalization, intracellular transit, and activation (5). Our observations support the proposition that peptide conformational responses to membrane bilayer properties are a hallmark of GPCRs important to their activation and consequent downstream signaling (6). The potential of mutual structural accommodation between GPCRs (including the CB1 receptor) and their membrane environment to affect GPCR transmission has implications for GPCR-targeted drug discovery (1, 6).
†This study was supported by National Institutes of Health/National Institute on Drug Abuse grants DA009158-10S2 (EKT) and DA3801 (AM) and a National Science Foundation instrumentation grant CHE-0466616 (DEB).
2The Ballesteros and Weinstein universal numbering scheme for class A GPCRs is used herein (20). Amino acid positions designated are those in the full-length, wild-type human CB1 receptor.