Our laboratory is using 3-D electron microscopy (3DEM) to determine the structures of proteins and macromolecular assemblages in muscle and the cytoskeleton. Although we use most 3-D imaging methods, our primary imaging method is electron tomography. Our approaches are somewhat from the norm in that we use exclusively marker-free image alignment. We are also at the forefront in the application of correspondence analysis to volume data from electron tomography.

Our longest running 3-D reconstruction project investigates the structure of myosin crossbridges in different states using the highly ordered filament lattice of insect flight muscle (IFM). In concert with the structural studies, we are developing 3-D reconstruction algorithms uniquely suitable for studying muscle structure.

We are also developing methods for analyzing the structure of paracrystalline specimens. This research is an outgrowth of our studies on IFM structure but the technology is applicable to many paracrystalline specimens. One of these techniques is a unique tomographic reconstruction method that uses crosscorrelation methods to align the images in a tilt series. To deal with the specimen disorder, which is manifest as variations in crossbridge structure, we are extending the widely used methods of 2-D correspondence analysis to 3-D motifs obtained by tomography. Finally, the knowledge of the atomic structure of the two important proteins in muscle contraction, myosin and actin, provides a unique opportunity to extend the low resolution information obtained by 3DEM to atomic resolution. The first and most important step in electron crystallography is the formation of 2-D crystalline arrays, which are the most suitable specimen for this technique. The first protein that we have successfully crystallized by this method is α-actinin. We have also formed 2-D arrays of smooth muscle HMM, smooth muscle myosin, and myosin-V.

As well as 2D protein crystallization, we utilize lipid monolayers to assemble multiprotein complexes in 2-D paracrystalline arrays. These arrays make structural analysis easier because they remove superposition problems that complicate image interpretation. The methodology for formation of what we call 2-D bundles open a number of avenues for research into the structure of the cytoskeleton.

Read More