Development and application of solid-state NMR spectroscopy to protein,
lipid and small molecule structure and dynamics

Overview

The Rienstra group has pioneered solid-state NMR methods and technologies that matured the discipline to solve high-resolution structures of small proteins. We have also contributed innovative approaches to enable larger structures of assemblies, fibrils and membrane proteins to be addressed and have extended these technologies to increasingly significant biomedical problems. These results are having a major impact on understanding Parkinson's disease, fungicidal drug action and protein-lipid interactions in blood coagulation.

Biological Applications

(1) Neurodegeneration and amyloid diseases. We have leveraged our capability to determine fibril structures by solid-state NMR in collaboration with Paul Kotzbauer and James Fitzpatrick (Washington U. St. Louis), experts in synucleinopathies and cryo-EM respectively, which aims to determine the structures of patient-derived fibrils with unique disease phenotypes by a combination of solid-state NMR and cryo-EM, and to structurally fingerprint a larger set of patient tissues utilizing a combination of cryo-ET and high-throughput NMR chemical shift analysis. We have developed methods specifically for evaluating the accuracy and similarity of structural models using minimal NMR data sets, such as ¹³C chemical shifts, enabling us to leverage structural models and to compare structures of various tissue samples utilizing minimal data sets and sample quantities. This project benefits from and motivates methodological projects to enhance NMR sensitivity and to accelerate data analysis (vide infra) and also includes collaborations with Virginia Lee (Penn Medical Center) and Marina Ramirez-Alvarado (Mayo Clinic, Rochester, MN).

(2) Small molecule drug assemblies and membrane interactions. Amphotericin B (AmB) serves as the prototype of a small molecule that acts by binding to another small molecule. Both in terms of fundamental biophysical chemistry and structural biology, and for its impact on broader potential applications in drug discovery and development, the specificity of binding of AmB with sterols and the cooperativity with aggregate assembly are fascinating topics. For half a century, pharmaceutical companies pursued the alternative and less efficacious mode of action (ion channel formation), and analogs of AmB that had a greater propensity to form ion channels failed to achieve greater in vivo fungicidal activity or reduced toxicity. Our studies with Marty Burke have changed this paradigm and conventional wisdom, and the sterol sponge model of AmB action correctly accounts for both the biophysical and structural observations as well as clinical outcomes. For example, the sponge model correctly predicts that AmB preloaded with cholesterol (i.e., the commercial Ambisome formulation) will result in reduced toxicity to renal epithetial cells. The sponge model also provides actionable intelligence to rationally develop better AmB analogs that retain ergosterol binding (and therefore fungicidal activity) yet reduce cholesterol binding (and therefore toxicity). These results illustrate for the first time a direct link between the structure and function of a macroscopically heterogeneous, high molecular weight, membrane-associated complex of small molecules.

(3) Allosteric regulation of blood coagulation. With Jim Morrissey (U. of Michigan) we have investigated the molecular basis of blood coagulation, focusing on membrane interactions of phosphatidylserine with gamma-carboxy glutamic acid-rich (GLA) domains of clotting factors. We have developed methods for semi-synthetically isotopically labeling phospholipid headgroups and identified early examples of lipid headgroups ordering with defined conformations in the presence of calcium ions and GLA domains. We have also assigned the spectra of tissue factor in several forms and are examining its membrane and factor VII interactions and dynamics in order to understand the mechanisms of allosteric regulation in this fascinating enzyme system, which when activated increases its activity by six orders of magnitude. These complexes are paradigmatic of where NMR dynamics studies (and solid-state NMR in particular) can provide essential data to elucidate allosteric networks that are coupled to sidechain conformational entropy.

Technology Development

(1) Sensitivity and resolution of solid-state NMR. High magnetic field, dynamic nuclear polarization (DNP), fast magic-angle spinning with indirect proton detection, and cryogenic probes are the four major approaches to enhance sensitivity of solid-state NMR spectra. Most of these methods are compatible with sample preparation conditions that yield high-resolution spectra. We are developing new magic-angle NMR probes for spectrometers at NMRFAM.

(2) Accelerated computational analysis of NMR data to determine structures. We have developed approaches to computational modeling of NMR structures (i.e., COMPASS) and will advance this in combination with molecular dynamics (Emad Tajkhorshid, Illinois) and simulated annealing (Charles Schwieters, NIH-Bethesda) approaches. COMPASS is a paradigmatic example of a suite of programs that maximally leverages the NMRBox.

(3) Sample preparation and isotopic labeling. A major thrust of the Rienstra group over the years has been preparation of samples with sparse isotopic labeling (for proteins, sterols, etc.) and/or site-specific isotopic labels incorporated by semi-synthesis methods (for lipids). We also have contributed a number of approaches to preparing auxotrophs and improving the yield of expression on isotopically labeled media, transferring samples into rotors, and preparing microcrystals and 2D crystals of membrane proteins for optimal spectral resolution.