In the first phase of the project, we sought to describe how fatty acid synthesis initiates in fungal FAS. In the first step of the reaction cycle an acetyl group is transferred from the acetyl transferase domain to ACP. This step coincides with a rotation of the β-subunit. However, it is unknown if the conformational change is required for the formation of the initiation complex or occurs after binding of the ACP. To test this, we performed MD simulations where we selectively removed elements hypothesized to stabilize the rotated state of the β-subunit.
However, initial efforts focused on the construction of an all-atom model of the FAS. Although the overall structure of the complex is well defined, there are several gaps where structural information is missing. These gaps correspond to structurally heterogeneous regions. Since these regions can interconvert between several structurally distinct conformations, they play an essential role in the dynamics of the complex. To incorporate these regions, we combined MD simulations, Alphafold2 predictions, protein design and experimental structural characterization. In total we modeled and validated more than 1000 missing residues. In addition, we performed a comprehensive benchmark to determine the most suitable force field to simulate this multidomain complex. We also developed force field parameters for the phosphopantetheine group, which links ACP to the reaction intermediates.
Once we had an all-atom model, we performed a series of simulations of the FAS in the rotated conformation. In a subset of these simulations, we selectively removed either the γ-subunit or modeled the ACP bound to another enzymatic domain. We hypothesized that removal of either element would allow us to determine which elements are required for β-subunit rotation and, more importantly, initiation complex formation. More than 50 microseconds of simulations were generated for the system of about 2.5 million atoms. Based on the simulated systems, we were able to conclude that the γ-subunit is critical for stabilizing the rotated conformation. This supports the theory that the initiation complex is regulated by the global conformational dynamics of the FAS.
Subsequently, we aimed to simulate the transition of the ACP between distinct enzymatic domains. We adapted the workflow developed in the previous section, to reconstruct models of FAS in different catalytic states. Initial simulations were started with the FAS in different catalytic states. Moreover, we applied the methods developed to study the FAS to investigate other biomolecular systems. This work is still ongoing.
Motivated by the initial work on the FAS, we applied also applied the methods and workflows to other biomolecular systems. We were interested in the substrate specificity of the ubiquitin conjugating enzyme Ubc6. Unlike other ubiquitin conjugating enzymes, Ubc6 can transfer ubiquitin to serine amino acids. However, structural studies are difficult because ubiquitin is covalently linked to Ubc6 through by a labile thioester. To stabilize the complex the thioester is commonly replaced with an isopeptide bond. However, this makes interpretation of the structures data ambiguous. We created a model of the Ubc6-ubiquitin complex with the native thioester bond. Modeling and subsequent MD simulations showed that a histidine in the active site facilitates the transfer of ubiquitin to target proteins.
