Spectroscopic signatures of bilayer ordering in native biological membranes

M. van Beekveld R, Breukink E, Weingarth M, Wylie BJ. Coordination of bilayer properties by an inward-rectifier K+ channel is a cooperative process driven by protein-lipid interaction

Physical properties of biological membranes directly or indirectly govern biological processes. Yet, the interplay between membrane and integral membrane proteins is difficult to assess due to reciprocal effects between membrane proteins, individual lipids, and membrane architecture. Using solid-state NMR (SSNMR) we previously showed that KirBac1.1, a bacterial Inward-Rectifier K+ channel, nucleates bilayer ordering and microdomain formation through tethering anionic lipids. Conversely, these lipids cooperatively bind cationic residues to activate the channel and initiate K+ flux. The mechanistic details governing the relationship between cooperative lipid loading and bilayer ordering are, however, unknown. To investigate, we generated KirBac1.1 samples with different concentrations of 13C-lableded phosphatidyl glycerol (PG) lipids and acquired a full suite of SSNMR 1D temperature series experiments using the ordered all-trans (AT) and disordered trans-gauche (TG) acyl conformations as markers of bilayer dynamics. We observed increased AT ordered signal, decreased TG disordered signal, and increased bilayer melting temperature with increased PG concentration. Further, we identified cooperativity between ordering and direct binding of PG lipids, indicating KirBac1.1-driven bilayer ordering and microdomain formation is a classically cooperative Hill-type process driven by and predicated upon direct binding of PG lipids. Our results provide unique mechanistic insight into how proteins and lipids in tandem contribute to supramolecular bilayer heterogeneity in the lipid membrane.

Cooperative Gating of a K+ Channel by Unmodified Biological Anionic Lipids Viewed by Solid-State NMR Spectroscopy

Lipids adhere to membrane proteins to stimulate or suppress molecular and ionic transport and signal transduction. Yet, the molecular details of lipid-protein interaction and their functional impact are poorly characterized. Here we combine NMR, coarse-grained molecular dynamics (CGMD), and functional assays to reveal classic cooperativity in the binding and subsequent activation of a bacterial inward rectifier potassium (Kir) channel by phosphatidylglycerol (PG), a common component of many membranes. Past studies of lipid activation of Kir channels focused primarily on phosphatidylinositol bisphosphate, a relatively rare signaling lipid that is tightly regulated in space and time. We use solid-state NMR to quantify the binding of unmodified 13C-PG to the K+ channel KirBac1.1 in liposomes. This specific lipid-protein interaction has a dissociation constant (Kd) of ∼7 mol percentage PG (ΧPG) with positive cooperativity (n = 3.8) and approaches saturation near 20% ΧPG. Liposomal flux assays show that K+ flux also increases with PG in a cooperative manner with an EC50 of ∼20% ΧPG, within the physiological range. Further quantitative fitting of these data reveals that PG acts as a partial (80%) agonist with fivefold K+ flux amplification. Comparisons of NMR chemical shift perturbation and CGMD simulations at different ΧPG confirm the direct interaction of PG with key residues, several of which would not be accessible to lipid headgroups in the closed state of the channel. Allosteric regulation by a common lipid is directly relevant to the activation mechanisms of several human ion channels. This study highlights the role of concentration-dependent lipid-protein interactions and tightly controlled protein allostery in the activation and regulation of ion channels.

INEPT and CP transfer efficiencies of dynamic systems in MAS solid-state NMR

Hartmann-Hahn cross polarization and INEPT polarization transfer are the most popular sequences to increase the polarization of low-γ nuclei in magic-angle spinning solid-state NMR. It is well known that the two methods preferentially lead to polarization transfer in different parts of molecules. Cross polarization works best in rigid segments of the molecule while INEPT-based polarization transfer is efficient in highly mobile segments where (nearly) isotropic motion averages out the dipolar couplings. However, there have only been few attempts to define the time scales of motion that are compatible with cross polarization or INEPT transfer in a more quantitative way. We have used simple isotropic jump models in combination with simulations based on the stochastic Liouville equation to elucidate the time scales of motion that allow either cross polarization or INEPT-based polarization transfer. We investigate which motional time scales interfere with one or both polarization-transfer schemes. We have modeled isolated I-S two-spin systems, strongly-coupled I2S three-spin systems and more loosely coupled I-I-S three-spin systems as well as I3S groups. Such fragments can be used as models for typical environments in fully deuterated and back-exchanged molecules (I-S), for fully protonated molecules (I2S and I3S) or situations in between (I-I-S).

Bacterial lipid biophysics and membrane organization

The formation of lateral microdomains is emerging as a central organizing principle in bacterial membranes. These microdomains are targets of antibiotic development and have the potential to enhance natural product synthesis, but the rules governing their assembly are unclear. Previous studies have suggested that microdomain formation is promoted by lipid phase separation, particularly by cardiolipin (CL) and isoprenoid lipids, and there is strong evidence that CL biosynthesis is required for recruitment of membrane proteins to cell poles and division sites. New work demonstrates that additional bacterial lipids may mediate membrane protein localization and function, opening the field for mechanistic evaluation of lipid-driven membrane organization in vivo.

Strategies for acquisition of resonance assignment spectra of highly dynamic membrane proteins: a GPCR case study

In protein nuclear magnetic resonance (NMR), chemical shift assignment provides a wealth of information. However, acquisition of high-quality solid-state NMR spectra depends on protein-specific dynamics. For membrane proteins, bilayer heterogeneity further complicates this observation. Since the efficiency of cross-polarization transfer is strongly entwined with protein dynamics, optimal temperatures for spectral sensitivity and resolution will depend not only on inherent protein dynamics, but temperature-dependent phase properties of the bilayer environment. We acquired 1-, 2-, and 3D homo- and heteronuclear experiments of the chemokine receptor CCR3 in a 7:3 phosphatidylcholine:cholesterol lipid environment. 1D direct polarization, cross polarization (CP), and T2' experiments indicate sample temperatures below - 25 °C facilitate higher CP enhancement and longer-lived transverse relaxation times. T1rho experiments indicate intermediate timescales are minimized below a sample temperature of - 20 °C. 2D DCP NCA experiments indicated optimal CP efficiency and resolution at a sample temperature of - 30 °C, corroborated by linewidth analysis in 3D NCACX at - 30 °C compared to - 5 °C. This optimal temperature is concluded to be directly related the lipid phase transition, measured to be between - 20 and 15 °C based on rINEPT signal of all-trans and trans-gauche lipid acyl conformations. Our results have critical implications in acquisition of SSNMR membrane protein assignment spectra, as we hypothesize that different lipid compositions with different phase transition properties influence protein dynamics and therefore the optimal acquisition temperature.

Cholesterol Biases the Conformational Landscape of the Chemokine Receptor CCR3: A MAS SSNMR-Filtered Molecular Dynamics Study

Cholesterol directs the pathway of ligand-induced G protein-coupled receptor (GPCR) signal transduction. The GPCR C-C motif chemokine receptor 3 (CCR3) is the principal chemotactic receptor for eosinophils, with roles in cancer metastasis and autoinflammatory conditions. Recently, we discovered a direct correlation between bilayer cholesterol and increased agonist-triggered CCR3 signal transduction. However, the allosteric molecular mechanism escalating ligand affinity and G protein coupling is unknown. To study cholesterol-guided CCR3 conformational selection, we implement comparative, objective measurement of protein architectures by scoring shifts (COMPASS) to grade model structures from molecular dynamics simulations. In this workflow, we scored predicted chemical shifts against 2-dimensional solid-state NMR 13C-13C correlation spectra of U-15N,13C-CCR3 samples prepared with and without cholesterol. Our analysis of trajectory model structures uncovers that cholesterol induces site-specific conformational restraint of extracellular loop (ECL) 2 and conserved motion in transmembrane helices and ECL3 not observed in simulations of bilayers with only phosphatidylcholine lipids. PyLipID analysis implicates direct cholesterol agency in CCR3 conformational selection and dynamics. Residue-residue contact scoring shows that cholesterol biases the conformational selection of the orthosteric pocket involving Y411.39, Y1133.32, and E2877.39. Lastly, we observe contact remodeling in activation pathway residues centered on the initial transmission switch, Na+ pocket, and R3.50 in the DRY motif. Our observations have unique implications for understanding of CCR3 ligand recognition and specificity and provide mechanistic insight into how cholesterol functions as an allosteric regulator of CCR3 signal transduction.

Dynamic Sorting of Mobile and Rigid Molecules in Biomembranes by Magic-Angle Spinning 13C NMR

Understanding the membrane dynamics of complex systems is essential to follow their function. As molecules in membranes can be in a rigid or mobile state depending on external (temperature, pressure) or internal (pH, domains, etc.) conditions, we propose an in-depth examination of NMR methods to filter highly mobile molecular parts from others that are in more restricted environments. We have thus developed a quantitative magic-angle spinning (MAS) ¹³C NMR approach coupled with cross-polarization (CP) and/or Insensitive Nuclei Enhanced by Polarization Transfer (INEPT) on rigid and fluid unlabeled model membranes. We demonstrate that INEPT can detect only very mobile lipid headgroups in gel (solid-ordered) phases; the remaining rigid parts are only detected with CP. A direct correlation is established between the normalized line intensity as obtained by CP and the C-H (C-D) order parameters measured by wide-line ²H NMR or extracted from molecular dynamics: I_CP/I_DP^eq ≈ 5|S_CH|, indicating that when the order is greater than 0.2-0.3 (maximum value of 0.5 for chain CH₂), only rigid parts can be filtered and detected using CP techniques. In very fluid (liquid-disordered) membranes, where there are many more active motions, both INEPT and CP detect resonances, with, however, a clear propensity of each technique to detect mobile and restricted molecular parts, respectively. Interestingly, the ¹³C NMR chemical shift of lipid hydrocarbon chains can be used to monitor order-disorder phase transitions and calculate the fraction of chain defects (rotamers) and the part of the transition enthalpy due to bond rotations (6-7 kJ·mol^-1 for dimyristolphosphatidylcholine, DMPC). Cholesterol-containing membranes (liquid-ordered phases) can be dynamically contrasted as the rigid-body sterol is mainly detected by the CP technique, with a contact time of 1 ms, and the phospholipid by INEPT. Our work opens up a straightforward, robust, and cost-effective route for the determination of membrane dynamics by taking advantage of well-resolved conventional ¹³C NMR experiments without the need of isotopic labeling.