ATG9 vesicles are incorporated into nascent autophagosome membranes

The autophagosome is a double-membrane organelle that traps cytoplasmic cargo and traffics it to the lysosome for degradation. How the autophagosome forms is uncertain, but a prevailing model suggests lipids are moved from the ER through the lipid transporter ATG2 to ATG9 vesicles, which then expand to comprise the growing autophagosomal membrane. However, evidence that ATG9 is ever resident within the autophagosome is scant; detection of this putative autophagosome-resident protein is made challenging both because most ATG9 vesicles in the cell are not involved in the biogenesis at any given time and because the dilution of one or a few vesicle membranes by potentially millions of transported lipids would result in a very low density of ATG9 on the mature autophagosome. Here we develop approaches to address each of these limitations. First, we show that in genetic knockouts of ATG2, ATG9 vesicles accumulate to very high numbers at sites of aborted autophagosome formation. Focused-ion beam scanning electron microscopy reveals that without ATG2, these putative autophagosome seed vesicles do not expand, but instead accumulate within a large vesicle cluster surrounded by ER. By fluorescence microscopy, we also detect downstream modifiers of the autophagosome membrane at these sites, including the lipid-anchored form of the LC3 proteins, which suggests that a biochemically competent seed membrane is present. To establish whether ATG9A is found on the same membrane as LC3B, we use styrene maleic acid (SMA) copolymer nanodiscs, to isolate very small intact sections of autophagosome membrane away from all other potential contaminants. Through rigorous combinations of isolation and purification, we reveal that ATG9A and LC3B are co-resident within the same 10 nm diameter membrane segment and likely engage in a protein-protein complex. We then apply the same SMA-based approach to show that even in wildtype autophagosomes, ATG9 and LC3 are co-resident. Thus, we assert that ATG9 vesicles are the seed membrane for the autophagosome.