Tau is an intrinsically disordered microtubule-associated protein that self-aggregates in the brains of many neurodegenerative diseases, including the Alzheimer's disease. High-resolution structures of tau amyloid fibrils have been determined for AD and several other tauopathies using cryo-electron microscopy. However, the mechanisms of tau aggregation from its microtubule-bound state to the pathological fibrillar state remain poorly understood. The brain tau fibril cores encompass only a small fraction of the full protein, thus the impact of the dynamic domains on fibril formation is also unknown. I will present my group's latest efforts to deconstruct the mechanism of tau aggregation by determining the structures and dynamics of tau fibrils assembled in vitro. We carried out 2D and 3D solid-state NMR experiments and complemented these by cryo-EM. The fibrils are produced from both full-length and truncated tau, both wild-type and pseudo-phosphorylated tau, and fibrils are formed both with and without heparin. Despite the large size of full-length tau, we obtained well-resolved and high-sensitivity 2D and 3D correlation ssNMR spectra. We have now determined five high-resolution structural models of these tau fibrils and two low-resolution models. The structures indicate that truncation alone does not reconstruct brain tau structures, but the introduction of negative charges by phospho-mimetic mutations lead to conformations that may be on pathway to brain tau structures. Measurement of the mobility of these fibrils by ssNMR provide further insight into how the fuzzy coat domains interact with and stabilize the rigid core. These results represent our ongoing major efforts to elucidate the chemical code of tau aggregation in diseases and to reconstitute AD and other neurodegenerative tau in vitro for therapeutic intervention.