The development of new technology and new methods has been central to our lab’s mission. One key development has been that of a complete framework for alignment, classification, and averaging of volumes derived by electron tomography that is computationally efficient and effectively accounts for the missing wedge that is inherent to limited angle electron tomography. This method blends sub-tomogram averaging, which uses information about individual molecules acquired from multiple angles, with single particle analysis, a technique that combines information from many individual molecules. This fusion compensates for the fact that tomographic data has low resolution and has the “missing wedge”, information that is lost due to the fact that images can’t be acquired from very high tilt angles. The development of these tools has proved to be critical for our successful effort in determining the structure of trimeric HIV-1 envelope glycoproteins, as well as our initial studies of membrane integral proteins such as Tsr and the glutamate receptor GluK2. The speed of data processing has increased dramatically over the last several years, without which we could not have undertaken the systematic effort of analyzing structural variations across a wide variety of HIV and SIV strains, as well as undertaking a number of new projects.
More recently, we have undertaken a project to streamline and automate single particle analysis of cryo-electron microscopy data. This project has brought together a number of tools for data analysis and structure validation, towards the goal of a real-time structural readout during data collection, as well as a streamlined, tested protocol for high-resolution structure determination of small protein complexes.
Our first efforts into atomic-resolution cryo-electron microscopy involved the 464 kDa soluble β-galactosidase complex. While the structure of this molecule has been solved to high resolution by X-ray crystallography, our cryo-electron microscopic work both proved that atomic resolution was possible with a small, soluble protein, and also revealed key information about the process of determining structures to this resolution. Using the latest methods for single particle analysis, we have recently determined the structure of β-galactosidase in complex with the inhibitor PETG to 2.2 Å resolution, revealing amino acid side chains, individual ions, and water molecules, as well as small but clear differences from published X-ray crystal structures. This data has been deposited in the EMPIAR database, and has subsequently been reanalyzed by a variety of groups around the world who, like us, are working to improve data analysis techniques.
Our very high resolution work with β-galactosidase – and subsequently, with the small metabolic complexes glutamate dehydrogenase (GDH), lactate dehydrogenase (LDH), and isocitrate dehydrogenase (IDH) – depended on a variety of factors, including new detector technologies that first became available in 2013. These direct detectors allow for much higher resolution data, and allow one to correct for the small movements of molecules that occur during imaging. While many groups are making use of these broadly available tools, this last year we published the first better-than-2Å resolution structure (of GDH) and the first better-than-4Å resolution structure of a protein smaller than 100 kDa (IDH).