Figure 2: Cytoplasmic domains of Atg9 interact with Atg17 to promote autophagy.

(a) The scheme shows the topology of Atg9 in membranes (transmembrane domains, pink cylinders; membrane bilayer in grey; cD, cytoplasmic domain). The grey bar represents Atg9^WT, the pink bars illustrate Atg9-variants with corresponding domain borders indicated by numbers. (b) α-myc immunoblot of co-floatation assays of Atg9-PL, containing Atg9^core or its variants, with Atg17. Large unilamellar vesicles (LUVs) lacking Atg9 served as a control for unspecific membrane binding. Proteins, recombinant Atg17 and Atg9–PLs used for experiments. (c) Immunoprecipitation (IP) of Atg17^myc from ATG9^WT, ATG9^ΔN, ATG9^ΔcD, ATG9^ΔNΔcD and ATG9Δ lysates under vegetative conditions and after 2 h of starvation. Atg9-variants contained the very N- (1–280) and C-terminal (779–997) unstructured regions, which were not present in Atg9^core (indicated by the prefix fl), because these regions are essential for proper trafficking of Atg9 in vivo^10,49,50. Atg17 and co-immunoprecipitated Atg9-variants were detected by α-myc and α-HA immunoblots as indicated. Atg17 strongly interacts with wild-type Atg9, but not with Atg9 lacking N-, cD- or both domains. Bottom panel: total protein amounts in cell lysates. (d) Pho8Δ60 assay of vegetative (white bars) and starved (black bars) ATG9^WT, ATG9^ΔN, ATG9^ΔcD, ATG9^ΔNΔcD and ATG9Δ. The Pho8Δ60 activity of starved ATG9^WT was set to 100% and used for normalization. Mean values±s.d. of N=3 independent experiments are shown.