Influenza Hemagglutinin (HA) is a homotrimeric viral fusion protein critical for the invasion of flu viruses. It is composed of two domains, a receptor-binding domain called HA1 and a viral fusion stem domain called HA2. HA assists in the invasion of viruses through an HA2 induced membrane fusion process under a lowered pH environment. The crystal structures of HA2 before and after the membrane fusion were solved previously. A comparison between them reveals a dramatic structural rearrangement of HA2 during the viral invasion process. In spite of the solved structures, dynamic details about how this structural transition happens are still missing. This thesis focuses on the investigation of the molecular mechanism underlying this structural transition and understanding how this transition induces the subsequent membrane fusion process.

In Chapter Two, we used a coarse-grained dual-basin structure-based model for investigating the overall structural transition of HA2. We find two disparate routes on this transition landscape and multiple metastable intermediates. Specifically, our simulations feature an early cracking'' process initializing the HA2 transition and a symmetry-breaking'' process leading to a functional bending structure of HA2.

In Chapter Three, we employed detailed explicit-solvent simulations with transferrable force fields to probe the initial phase of HA2 transition. Specifically, we focused on the role of a lowered pH in the release of fusion peptides. Our results indicate that a buried salt bridge locks the fusion peptides in the pre-fusion structure, and the breaking of it is crucial for releasing fusion peptides and the subsequent HA2 transition. Further, our detailed simulations reproduce the cracking and symmetry-breaking processes as we observed in the simulations with the structure-based model.

In Chapter Four, we focused on a loop-to-coiled-coil transition of the B-loop domain of HA2, which was presumed to be a critical step in the structural transition of HA2. We implemented explicit-solvent simulations together with enhanced sampling techniques and showed that the post-fusion state of the B-loop by itself is unstable. A buried hydrophilic residue, Thr59, is shown to cause the instability. A further study indicates that Thr59 is the only residue of the B-loop that strictly differentiates between two taxonomic groups of HAs. Our simulations support previous studies by showing that the functional motion of HA2 is dynamic. The slow transition of the B-loop allows for more degrees of freedom for choosing the transition pathways.

Overall, our simulations indicate a dynamic motion of HAs in its functional transition, which encourages different pathways HAs utilized to induce the membrane fusion process.