research

The structures of membrane proteins and aggregates, important in human disease, can be solved with solid-state NMR methods.The Protein Data Bank (PDB) contains more than 45,000 biomolecular structures. Unfortunately, the PDB has a high redundancy of many water-soluble proteins and vastly understates the importance of membrane proteins: genomic analyses predict over 10,000 membrane proteins in humans, but relatively few of these structures have been determined. Membrane proteins with unknown structures include those responsible for immune recognition, signal and energy transduction and ion conduction; these are primary targets of most modern pharmaceuticals. Those membrane proteins whose structures have been determined are often at too low a resolution (~3 to 4 Å) to benefit rational drug design.

Protein aggregates, which are often found in “intrinsically unstructured” precursors states, present a second major structural challenge. Fibrils and their precursor states (oligomers and protofibrils) are directly implicated in diseases ranging from Parkinson's, Alzheimer's and Huntington's to amyotrophic lateral sclerosis and bovine spongiform encephalopathy (Mad Cow disease). (Merlini & Westermark, 2004) These pathologies - also implicated in cancer, type II diabetes, heart disease and cataracts - frequently derive from individual point mutations. (Polymeropoulos, 1997) Overall, more than half of the ~40,000 mutations in the Human Gene Mutation Database are missense or nonsense mutations, many directly implicated in specific diseases. In both fibrous and membrane proteins, the dramatic effects of changing a few atoms out of thousands highlight the precise balancing of forces that enable normal physiology and motivate atomic-resolution structural insight into these problems. No high-resolution solution NMR or crystallography structures have been determined in fibrils, because they are insoluble and do not form 3D crystals. Crystallization of membrane proteins is rarely successful, and solution NMR requires special solubilization techniques; these protocols often compromise physiological relevance.

In contrast, SSNMR samples require order only on the ~1 nm scale, a criterion common to microcrystalline, precipitated, frozen, fibrous and membrane-embedded samples. We obtain high-resolution spectra from this wide range of preparations, revealing structural details beyond the limit of diffraction. Our SSNMR techniques therefore drive both hypothesis-directed and discovery-driven studies of structure and function, under previously inaccessible conditions.