Biochemical reconstitution of human autophagy initiation

Abstract

Autophagy is a tightly regulated process that eukaryotic cells use as a major survival mechanism to reallocate nutrients to essential processes in adverse conditions such as nutrient or energy deprivation. Autophagy is characterized by the formation of a phagophore, a double membrane organelle that matures into an autophagosome to capture damaged or surplus materials in the cytosol and deliver them to the lysosome for degradation and recycling. Yet, how the autophagosome is generated de novo remains a long-standing question in biology. Research in the last decades has suggested that autophagosome biogenesis requires the transfer of lipids from the endoplasmic reticulum (ER) to the nascent autophagosome (phagophore), which happens at the membrane contact site (MCS) between the two organelles. Several core autophagy initiation complexes and a lipid transfer machinery are recruited to the MCS to modulate autophagic membrane formation and elongation. However, profound questions remain: How does the MCS assemble at the right place and time? Is there a regulatory mechanism? How is lipid transfer modulated to support phagophore elongation? Answering these questions would provide fundamental insight into the mechanisms of autophagosome biogenesis. In this thesis, biochemical reconstitution is used as a reductionist approach to address these questions. One of the main challenges of this approach is the production of recombinant proteins. Here, we meet this challenge by purifying almost all full-length proteins of the core autophagy initiation complexes (ULK1 complex, PI3K complex 1, ATG9) and the lipid transfer unit (ATG2-WIPI4). Interestingly, we were able to reconstitute a seven-subunit autophagy initiation super-complex and found that a three-subunit complex of ATG9-ATG13-ATG101 serves as a core complex for the assembly of other four subunits, including ULK1, FIP200, ATG14L, and BECN1. Data from our lab also shows that ATG13 and ATG101 are metamorphic proteins, and their metamorphoses result in an incredibly slow self-assembly of the core complex. The slow assembly of the core complex thus acts as a rate-limiting step in the assembly of the super-complex and raises the possibility of a regulatory mechanism for on-demand assembly of the super-complex upon autophagy induction. Moreover, I found that the core complex also interacts with the lipid transfer unit, ATG2-WIPI4, to form a five-protein subcomplex. ATG2-WIPI4 was previously found to tether membranes and mediate lipid transfer at the MCS. Surprisingly, the lipid transfer efficiency of the lipid transfer unit can be significantly enhanced by both ATG9 and ATG13-ATG101 of the core complex. In summary, our findings pave the way for mechanistic models that explain how autophagosome biogenesis is regulated in space and time and how the co-incidence of the different functional complexes supports autophagosome expansion.