The Grigorieff laboratory is studying the three-dimensional (3D) structure of membrane proteins, particularly voltage-gated ion channels. We are developing new methods in electron microscopy (EM) to examine non-crystalline material. Traditionally, the techniques of X-ray crystallography and NMR have been used to determine atomic structure, but many membrane proteins--such as receptors, channels, and transporters--do not form crystals suitable for X-ray diffraction and are too large for NMR analysis. Under EM, however, proteins and protein complexes with molecular weights that range from a few thousand to several million Daltons can be visualized. If a protein's molecular weight is much less than approximately 100,000 Daltons, then two-dimensional (2D) crystals are needed in order to solve the protein's structure. Many membrane proteins form 2D crystals more readily than the 3D crystals that are needed for X-ray diffraction, whereas larger molecules and complexes that do not crystallize can be visualized under EM even when they occur as single particles. For example, many of the ion channel complexes that underlie the generation of electric impulses in nerve and muscle cells have a combined molecular weight of 250,000 Daltons or more.
New Methods in Single Particle Electron Microscopy
Electron crystallography of 2D crystals has been successfully used to determine some atomic structures (such as the proton pump bacteriorhodopsin and a plant light-harvesting complex). Electron microscopy of single particles still falls short of atomic resolution, although theoretical predictions and signal-to-noise estimates, based on real images of 2D crystals, clearly show that atomic structures can be determined using images of single particles. The current resolution limit of about 7 Å (as demonstrated for highly symmetrical viruses) is the result of experimental factors, such as the blurring of images by electron beam-induced sample movement and electrostatic charging, as well as imperfect image processing, such as particle misalignment and failure to extract a high-resolution signal from the images. Our laboratory is refining experimental methods for obtaining high quality images and improving image-processing algorithms in order to completely retrieve the high-resolution signal in the images.
Voltage Gated Ion Channels
Existing and newly developed EM techniques are being applied to voltage-gated ion channels that play a central role in the function and physiology of excitable cells. These channels may be divided, according to their ion specificity, into Na+, K+, Ca2+, and Cl- channels, each performing unique cellular functions. Apart from the propagation of excitatory impulses along cell membranes, accomplished mainly by Na+ and K+ channels, voltage-gated ion channels are also engaged in the coupling between cell excitation and physiological function. Ca2+ channels, for example, realize excitation-contraction coupling in cardiac and smooth muscle. Cl- channels are involved in the control of the cell volume and the stabilization of the membrane potential. The structure of a bacterial K+ channel (MacKinnon et al.) has shown, for the first time, how such a channel could achieve ion selectivity and conduction. Other structural elements required for voltage-dependent activation and inactivation of more complex channels still remain to be fully understood. The recently solved crystal structure of a bacterial voltage-gated potassium channel (MacKinnon et al.) indicates a possible arrangement of the voltage sensor and suggests a paddle-like motion that moves the sensor through the membrane in response to a change in the membrane potential. To locate the C-terminus in our structure, for example, we looked at a channel mutant that is lacking the C-terminus. A 3D structure obtained from this mutant revealed where the missing density, corresponding to the deleted C-terminus, is located.
The Spliceosome
The second line of research in our laboratory is directed at the determination of the 3D structure of the spliceosome. The work is carried out in collaboration with the Moore laboratory (Brandeis) which provides outstanding expertise in the expression and purification of homogeneous splicing complexes which are ideally suited for single particle EM. The spliceosome mediates removal of introns from nascent transcripts, an essential step in eukaryotic gene expression. Most introns interrupt precursors to messenger RNAs (pre-mRNAs) and their precise excision is required to create readable mRNAs. Spliceosomes are ribosome-sized (50 - 60 S) complexes composed of pre-mRNA, four small nuclear ribonucleoprotein (snRNP) particles and a host of associated protein factors. The snRNPs (U1, U2, U4/6, and U5) are in turn multicomponent complexes, each containing at least one small stable RNA molecule (snRNA) and five or more tightly bound polypeptides. In all, it has been estimated that nuclear pre-mRNA splicing requires the action of over 100 different gene products. Images of individual snRNPs embedded in negative stain have been obtained by EM (Berthold Kastner and co-workers) and provide a first notion of how structural elements of the entire spliceosome look like. However, the 3D structure of one or more of the spliceosomal complexes, at a resolution of about 20 Å or higher, will be invaluable for a better understanding of the inner workings of the spliceosome. The catalytically competent C complex stands at the end of an ordered pathway by which the snRNPs assemble to form spliceosomes. To better understand this assembly, and how splice sites are recognized, we are also working on earlier splicing complexes. Finally, together with the Moore laboratory, we study the exon junction complex (EJC), a post-splicing complex that remains on the spliced mRNA substrate. The EJC targets the spliced mRNA for nuclear export and is involved in determining its fate in subsequent processing, such as translation by the ribosome.
N-ethyl Maleimide Sensitive Factor (NSF)
NSF belongs to the family of AAA ATPases and is an essential component of the protein machinery that regulates vesicle fusion with target membranes, for example at synaptic terminals. NSF associates with a-SNAP (Soluble NSF Attachment Protein) to disassemble SNARE (Soluble NSF Attachment Protein REceptor) complexes. SNAREs, together with other proteins, facilitate docking and fusion of vesicles, and they are recycled and reactivated through disassembly by NSF. NSF functions as a homo-hexamer and each protomer contains three domains. The N-terminal domain of NSF is essential for the binding of a-SNAP and is followed by ATPase domains D1 and D2. Binding and hydrolysis of ATP by the D1 domain induces conformational changes in NSF leading to disassembly of the SNARE complex. The Brunger laboratory determined the crystal structures of the N and D2 domains, and of a-SNAP and a SNARE complex. Together with the Brunger laboratory, we recently obtained a structure at 11 Å resolution of NSF bound to a-SNAP and a SNARE that revealed the arrangement of the D1 and D2 domains within the NSF hexamer. Other parts of the structure, including the N domain and a-SNAP/SNARE complex, appeared to be disordered and were not resolved at the same level of detail. Our goal is to visualize these parts of the structure at higher resolution using improved preparations of the complex, and novel image processing techniques that can accommodate sample heterogeneity.
Chloride Ion Channels
Cl- channels play a multitude of roles in biological membranes. In contrast to cation-conducting channels, which service ions with fixed, defined gradients, Cl- channels handle a biologically ambidextrous ion whose cytoplasmic concentration, and hence equilibrium potential, varies greatly with cellular context. The ClC family of Cl- channels is the only family identified so far. ClC channels come in different functional flavors (voltage-gated, osmosensitive, and inwardly or outwardly rectifying), but all eukaryotic ClCs are built from polypeptides of ~100 kD with a characteristic transmembrane topology signature. ClC channels are homodimers with one independent pore per monomer. In addition, they have a slow common gate.
Together with the Miller laboratory at Brandeis University, we obtained 2D crystals of a ClC homologue from E. coli called EriC. Using cryo-EM, these crystals diffract to 6 Angstrom resolution when embedded in glucose, which is sufficient to resolve a-helices. A projection structure was calculated from images of these 2D crystals, revealing a dimer with at least two water-filled pores. Subsequently, Rod MacKinnon and co-workers used X-ray crystallography to solve the 3D structure of EriC, providing a much more detailed picture of the channel. Recently, however, the Miller laboratory found that EriC is not a channel but a chloride-proton antiporter. Despite the high-resolution X-ray structure, it remains unclear how this antiporter transports ions and protons. We are conducting experiments to detect conformational changes in the channel by subjecting our 2D crystals to varying environmental conditions, such as chloride concentration and pH.
Amyloid Fibrils
Amyloid fibrils are peptide or protein aggregates that form under certain conditions in vitro or in vivo. For example, the amyloid fibril plaques found in brain tissue of Alzheimer patients are formed from the peptide Ab and are associated with neurodegeneration. Amyloid formation is also observed with other diseases, such as type II Diabetes and Creutzfeldt-Jakob. Amyloid structures represent an alternative to the native folding pattern of many peptides and proteins. A characteristic motif of this folding pattern is the cross-b structure in which the peptides or proteins associate by b-sheet formation within protofilaments making up a fibril. In collaboration with Marcus Fändrich (Leibniz-Institute for age research, Jena, Germany) we study the molecular architecture of amyloid fibrils associated with human disease. Our goal is to identify fundamental principles of amyloid formation, and potential targets for disease treatment.