Neurodegenerative diseases are a major public health burden. There are only a few effective therapies because the underlying disease mechanisms are poorly understood. Tau aggregation is a hallmark of several neurodegenerative diseases, including Alzheimer’s disease and frontotemporal dementia. There are causal disease variants of the tau-encoding gene, MAPT, and the presence of tau aggregates is highly correlated with disease progression. The molecular mechanisms linking pathological tau to neuronal dysfunction are not well understood. A major challenge for the tau field is a lack of clarity around tau’s normal function in development and disease and how these processes change in the context of causal disease variants or amyloid beta plaques.

To address these questions in an unbiased manner, we conducted multi-omic characterization of iPSC-derived neurons harboring the MAPT V337M mutation, which revealed major changes in regulators of axonogenesis. MAPT V337M neurons have increased axon branching. Pathogenic tau mutations have generally been assumed to act through a gain of toxic function, leading to therapeutic approaches aiming to lower tau levels that are currently in clinical trials. Surprisingly, we found that tau knockdown did not rescue axon branching in MAPT V337M neurons, and tau knockdown induced axon branching in MAPT WT neurons, strongly supporting a tau loss of function effect. Intriguingly, knockdown the tau-interacting protein MYO1B also increases axon branching in wild-type neurons without modifying axon branching in MAPT V337M neurons. We conclude that the FTD-associated tau mutation MAPT V337M drives major changes in neuronal differentiation and maturation caused by loss of tau function.

Translation begins when initiation factor-2 (eIF2) delivers methionyl initiator tRNA (Met-tRNAi) to the ribosome. The exchange of GDP bound to eIF2 for GTP is a prerequisite to binding Met-tRNAi and is mediated by a second initiation factor, eIF2B. Regulation of mRNA translation is achieved through phosphorylation of eIF2 α at Ser51 which converts eIF2 from a substrate into a competitive inhibitor of eIF2B. Using the latest cryo-electron microscopy (cryoEM) technologies in both collection and data processing we were able to obtain three high resolution structures to interrogate the structural basis of the integrated stress response.

The family of dynamin fission proteins polymerize and use GTP to divide eukaryotic membranous structures including endocytic necks, peroxisomes, mitochondria, chloroplasts, and plastids. Current models for how fission dynamins constrict and divide membranes describe cleavage of budding endocytic vesicles. However, the mechanisms by which dynamins can divide the larger, low-curvature membranes of entire organelles remain poorly understood. Here we report cryo-EM structures of polymers of the human mitochondrial fission dynamin-related protein 1 (DRP1) stabilized in high- and low- curvature states at 3.5-4 Å resolution. Together these reveal the structural determinants by which DRP1 is adapted to the larger diameter of the mitochondria, and a unique mechanism of constriction that allows DRP1 to constrict low curvature membranes to the size of endocytic necks without GTP hydrolysis. These adaptations of metazoan DRP1 proteins are localized to the canonical dynamin interface 1 and the interface between the Stalk domain and the Bundle-Signaling Element. These results suggest that tuning of these two interfaces allow different members of the fission dynamin family to constrict not only endocytic necks, but also a diversity of eukaryotic membrane-bound organelles.

Membranes are fundamental to cellular life. They define the cellular border and establish biochemically specialized subcellular compartments. As cell grow, change, and divide, cellular membranes are remodeled accordingly: undergoing membrane fission and fusion to maintain cellular architecture. For example, the nuclear membrane that organizes and protects DNA is remodeled during cell division. In cells that undergo “open mitosis”, the nuclear membrane is fully disassembled to allow chromosomes to be segregated by the spindle apparatus and reassembled following chromosome segregation. Understanding how cells remodel nuclear membranes to rebuild nuclei with every cell division is critical to understanding the core principals of membrane biology that underlie cellular life.

In my dissertation, I describe how a membrane fusion complex is organized to seal holes in the reforming nuclear envelope that are occupied by spindle microtubules. Using biochemical approaches, I find that an inner nuclear membrane protein, called LEM2, uses multivalent, low-affinity binding interactions to condense in a liquid-like phase at the junction of the nuclear envelope, chromatin, and residual spindle microtubules. There, LEM2 activates the ESCRT protein, CHMP7, in a looping copolymer. Together, the fluid LEM2 toroid and structured LEM2-CHMP7 copolymer collaborate to serve as a molecular “O-ring” for early nuclear sealing. Furthermore, the LEM2-CHMP7 ring serves as a foundation for the membrane-remodeling ESCRT-III complex to execute coupled spindle disassembly and membrane fusion. This work establishes a new paradigm in membrane organization, in which a liquid-like protein phase can function as both a barrier within a discontinuous membrane and a scaffold for membrane fusion factors.