In continuing studies on the architecture of pyruvate dehydrogenase complexes, we have systematically addressed the arrangement of the E1, E2, E3 and linker domains in a variety of multi-enzyme icosahedral pyruvate dehydrogenase complexes. Two-dimensional projection images of the complex reveal the general structural arrangement of the central core decorated by peripheral subunits, but the three-dimensional reconstruction of complexes from bacteria, yeast and mammalian complexes that we and others have carried out indicate a surprising diversity in detailed structural arrangements of each type of complex. These structural studies thus allow the development of an integrated understanding of the connections between the enzymology and structural biology of these complexes. In PDH subcomplexes of B. stearothermophilus in which the E2 core is fully decorated with E1 or E3, the outer shell of peripheral subunits localize 75-100 above the E2 catalytic core. This structural architecture may increase productive substrate channeling because the active sites of E1, E2 and E3 each face the annular gap. The swinging lipoyl domain may also preferentially localize to the gap, increasing its effective concentration, at least when the number of peripheral subunits bound to the complex is high. Three-dimensional snapshots of individual complexes obtained using cryo-electron tomography indicate that the peripheral enzymes localize well above the core, even in lightly decorated complexes although their arrangement deviates from the strict icosahedral symmetry present in the inner core. In B. stearothermophilus, lateral movements of the subunits above the core can extend the effective reach of the swinging arm since a lipoyl domain associated with a specific E2 monomer potentially can contact up to 12 peripheral subunits. Radial movements of the outer enzymes toward the core, possibly through the transient formation of a type VI-turn in the inner semi-flexible linker may maintain the peripheral subunits in close proximity to each other and the core to facilitate catalysis. This is particularly relevant for E3 given that few enzymes populate each complex. The studies we have carried out with B. stearothermophilus can be viewed in the context of work on other model systems to develop a mechanistic understanding of how these enzyme complexes work in different organisms. Mammalian pyruvate dehydrogenases have different constraints than their bacterial counterparts: while E3 is more plentiful and potentially more symmetrically distributed around the E2 core, the kinases and phosphatases that bind to the inner lipoyl domain prior to interaction with E1 must have clear accessibility to the interior of the complex. The inner linker that joins the inner E2 domain and the peripheral subunit-binding domain is more ordered in the region adjacent to the core. Upon E1 binding, the linker is thought to become even more rigid, since density for this region can be visualized in bovine PDH complexes. The E2 monomers also integrate into a relatively dynamic central icosahedron that exhibits variation in size. These modifications likely enhance active site coupling in annular gaps that are more crowded than those found in bacteria, and promote interactions between E1 and the regulatory kinases and phosphatases. In addition, while both bacterial and eukaryotic enzymes are thought to transfer acetyl groups between adjacent lipoyl domains, the eukaryotic complexes additionally could require the movement of the regulatory enzymes accomplished through coordinated interactions with both E1 and the lipoyl domains of E2 in a hand-over-hand transfer mechanism. In related studies, and in collaboration with Dr. Subramaniam and colleagues, we have worked extensively on understanding structure and structural changes of HIV envelope glcoproteins. Our collaboration has resulted in the determination of the 3D structures of envelope glycoproteins on a large set of HIV and SIV strains. We have contributed to methodological advances for automation of image processing and structure refinement, and these have led to a 10-fold increase in the speed of determining structures of these membrane proteins displayed on intact viruses using cryo-electron tomography. Our collaboration has also extended to structural studies of trimeric gp140 immunogens using cryo-electron microscopy. These studies have produced the first structures of trimeric Env immunogens and will be invaluable for rational vaccine design.
Showing the most recent 10 out of 17 publications
