The molecular mechanisms of autophagy have been best characterized in the yeast Saccharomyces cerevisiae, where a number of proteins have been identified to be essential for this degradative pathway. ATG (autophagy-related) proteins(1) localize to a unique compartment, the pre-autophagosomal structure (PAS). Isolation membranes are suggested to originate from the PAS, enwrapping cytoplasmic components to form a double membrane autophagosome, which then fuses with the vacuole. Although many Atg proteins have been identified, the source of the PAS membrane in yeast is unknown. Identification of the source of the PAS in yeast has been hindered due to the transient association of Atg proteins with forming autophagosomes.(2) Likewise, in mammalian cells, it is not known if a PAS equivalent exists or if the formation of autophagosomes occurs from numerous membrane sources. The identification of stably associated markers would allow us to address this question further. Thus, characterization of the only transmembrane autophagy protein so far identified, Atg9, may aid in the search for the source of the PAS. Recent data from our lab suggests that mammalian Atg9 (mAtg9) traffics between the Golgi and endosomes, and suggests an involvement of the Golgi complex in the autophagic pathway.(3) Here we address the implications of our model with regard to membrane trafficking events in mammalian cells after starvation.