| dc.description.abstracteng | Pyruvate dehydrogenase multi-enzyme complex (PDHc) is an assembly of multiple copies of four different proteins. Together they carry out the oxidative decarboxylation of pyruvate and generate acetyl-CoA and NADH, which are components of Krebs cycle, energy production and fatty acid biosynthesis in cells. Out of the four different subunits of PDHc, three are known to have distinct active sites and are found in PDHc from all organisms. The three catalytically important components are termed E1 (pyruvate dehydrogenase), E2 (dihydrolipoamide acetyltransferase) and E3 (dihydrolipoamide dehydrogenase). Exclusively in eukaryotic PDHc, an additional component, E3BP (E3 binding protein), is present whose role has been proposed to be for structural support of the PDHc assembly
This enzyme complex, which has already been studied for half a century, is a textbook prototype for substrate channelling between remotely located active sites in a multi-enzyme system. The individual active sites are spatially separated by at least few nanometers, and are coupled by highly flexible lipoyl arms of E2 and E3BP. The lipoyl arms consist of one or more lipoyl domains, each carrying covalently linked lipoamide groups. These lipoamide “swinging arms” need to visit all three active sites at E1, E2 and E3 in a sequential manner in order to complete a reaction cycle of PDHc. By structural design, the E2 core in prokaryotes and E2/E3BP core in eukaryotes make 24meric cube or 60meric pentagonal dodecahedron from which lipoyl arms and binding domains of E1 and E3 emanate outward.
In this PhD thesis, we focussed our attention on human PDHc and could elucidate the structural architecture of the core in details not achieved before. Firstly, we were able to calculate structural models for human E2/E3BP (hE2/E3BP) core from the cryo-EM density map at a resolution of ~ 6 Å. Our data revealed that the published pseudo-atomic model of human E2 was in part erroneous. By integrating crosslinking MS data in our E2/E3BP structural model, we predict a hitherto unknown mode of structural dynamics that act along the length of core subunits, at least for E3BP. This mode is different from the ‘breathing motion that acts at the inter-trimer bridges orthogonal to the length of the subunits.
We proved that hE2/E3BP core most likely consists 40 hE2 and 20 E3BP subunits, which can bind to 20 human E3 dimers (hE3). Furthermore, only hE2 can bind to substrate coenzyme A (CoA). We observed that, unlike in prokaryotic PDHc, in hPDHc, only hE1 appear to form outer shell while hE3 can fluctuate between the outer shell and the core cavities. Also, in native MS experiments, we detected different trimer arrangements of core subunits in isolated hE2/E3BP core versus fully assembled hPDHc, namely 2 hE2 - 1 E3BP type and 1 hE2 – 2 E3BP type, respectively. All of these observations indicate that in hPDHc, hE2 and E3BP are not equally distributed in the core but rather with local patches. Wherever hE2-hE1 is in excess, E1/E2 reactions might be preferred. And at the patches where E3BP-hE3 is dominant, E3-catalyzed regeneration of the lipoamide cofactor mostly occur.
Another key discovery made during this thesis work was the large conformational changes in the lipoyl arms when CoA binds to hE2. By quantifying crosslinks detected in crosslinking MS, we could show that when CoA substrate is bound, the preferred destination of lipoyl domains are the core surface and hE3. Together with increased crosstalk between lipoyl domain of E3BP and hE2, it appears that CoA binding primes the hPDHc for specifically E2 and E3 reactions. To our knowledge, this is the first instance where substrate binding in any PDHc could be shown to impact the conformational landscape of lipoyl arm dynamics. In addition, the orientation of lipoyl domains while they come into proximity to each other seems to be conserved. This was an unexpected observation due to the very high flexibility of the lipoyl arms and the distances they have to travel in order to visit all the various active sites. These findings all indicate that movement of lipoyl arms to couple active sites in PDHc is not a multiple random coupling mechanism alone but is also impacted by substrate binding and catalysis synchronizing their movements for subsequent steps in the multi-step PDHc reaction. | de |