Filamentous oligomer structure reconstruction and a new filament picking method for Cryo-EM micrographs

Macromolecular protein complexes play important diverse roles in cells. Among various macromolecular complexes, the filamentous oligomer is a special category, which is assembled from identical subunits with helical symmetry. Protein filaments carry out unreplaceable roles in the cellular process. These complexes not only provide mechanical support to cytoskeletons but also contribute towards regulating critical signaling pathways that often determine the live and death fate of the cells. Thus, atomic-resolution structures of these protein filaments help to reveal biological insights of related cellular processes and may serve as druggable targets. As native state protein filaments are reluctant to crystallize, the most common method of structure determination of these filaments is to use Cryo-EM. Thanks to the rapid development of Cryo-EM instruments and reconstruction algorithms, it becomes a viable and efficient way to obtain an atomic-resolution filament structure nowadays. Recognizing the bottleneck step of this process, the filament picking, we developed a rapid and easy-to-use software. Also included in several examples in this thesis, we have demonstrated that we have established a robust workflow in optimizing filamentous protein constructs, preparing filamentous cryo-EM samples, and resolved their atomic structures using cryo-EM helical reconstruction methods. In one example, Uromodulin (UMOD) is a secretory protein that belongs to the zona pellucida-like domain family. It is the most abundant protein in normal human urine, which forms filament and captures uropathogens, principally serving as decoys to gram-negative bacteria. We solved UMOD native state filament structure with resolution 3.35 Å (EMDB-10553). This high-resolution structure revealed a unique interdomain linker that connects the ZP-N and ZP-C domain of UMOD. This inter- 19 locked chain architecture confers a tremendous resistance to a hazardous environment, particularly to proteases. Besides, as the first solved ZP domain filament, UMOD filament structure also provided insight on potential ZP4 ZP3 interactions, which is a critical biochemical event involved in the fertilization process. In another example, Caspase recruitment domains (CARD) also tend to form filaments. They belong to a sub-family of the Death Domain superfamily, whose members play critical roles in cellular apoptotic events and innate immune signaling. CARD complexes are usually stabilized by homotypic interaction among various domains. Atomic-resolution structures of CARD oligomers help to reveal details of signaling regulatory mechanisms, which are highly sought after for drug development and mending-related autoimmunity diseases. NLRC4 belongs to such CARD domain containing receptor family, and it detects bacterial-associated pathogen patterns. Facilitated by NAIP5 and activates downstream ASC-CASP1 inflammasome signaling axis, the CARD domain in NLRC4 relays the activation signal via a cooperative oligomerization process. NLRP1 and CARD8 are two similar innate immune receptors that undergo auto-proteolytic activation, when encountering the right triggering signals from cellular damage or pathogen infection, and release a filament-forming CARD domains. Here, as part of my Ph.D. research, I present structures of NLRC4-CARD, NLRP1-CARD, and CARD8-CARD filaments, at approximately 3.7 Å resolution. These results reveal signaling filament activation mechanisms conducted by specific CARD-CARD interactions.Doctor of Philosoph