Following endocytosis and entry into the endosomal network, integral membrane proteins undergo sorting for lysosomal degradation or are alternatively retrieved and recycled back to the cell surface. Here we describe the discovery of an ancient and conserved multi-protein complex which orchestrates cargo retrieval and recycling and importantly, is biochemically and functionally distinct to the established retromer pathway. Composed of a heterotrimer of DSCR3, C16orf62 and VPS29, and bearing striking similarity with retromer, we have called this complex ‘retriever’. We establish that retriever associates with the cargo adaptor sorting nexin 17 (SNX17) and couples to the CCC and WASH complexes to prevent lysosomal degradation and promote cell surface recycling of α5β1-integrin. Through quantitative proteomic analysis we identify over 120 cell surface proteins, including numerous integrins, signalling receptors and solute transporters, which require SNX17-retriever to maintain their surface levels. Our identification of retriever establishes a major new endosomal retrieval and recycling pathway.

Kvainickas et al. show that the retromer cargo CI-MPR does not recycle from endosomes to the trans-Golgi network through interactions with the core retromer trimer. Instead, CI-MPR depends on cargo-selective SNX-BAR proteins, which function independently of the core retromer trimer.

Retromer is an endosomal multi‐protein complex that organizes the endocytic recycling of a vast range of integral membrane proteins. Here, we establish an additional retromer function in controlling the activity and localization of the late endosomal small GTPase RAB7. Surprisingly, we found that RAB7 not only decorates late endosomes or lysosomes, but is also present on the endoplasmic reticulum, trans‐Golgi network, and mitochondrial membranes, a localization that is maintained by retromer and the retromer‐associated RAB7‐specific GAP TBC1D5. In the absence of either TBC1D5 or retromer, RAB7 activity state and localization are no longer controlled and hyperactivated RAB7 expands over the entire lysosomal domain. This lysosomal accumulation of hyperactivated RAB7 results in a striking loss of RAB7 mobility and overall depletion of the inactive RAB7 pool on endomembranes. Functionally, we establish that this control of RAB7 activity is not required for the recycling of retromer‐dependent cargoes, but instead enables the correct sorting of the autophagy related transmembrane protein ATG9a and autophagosome formation around damaged mitochondria during Parkin‐mediated mitophagy.

Retromer is an evolutionarily conserved multiprotein complex that orchestrates the endocytic recycling of integral membrane proteins. Here, we demonstrate that retromer is also required to maintain lysosomal amino acid signaling through mTORC1 across species. Without retromer, amino acids no longer stimulate mTORC1 translocation to the lysosomal membrane, which leads to a loss of mTORC1 activity and increased induction of autophagy. Mechanistically, we show that its effect on mTORC1 activity is not linked to retromer’s role in the recycling of transmembrane proteins. Instead, retromer cooperates with the RAB7-GAP TBC1D5 to restrict late endosomal RAB7 into microdomains that are spatially separated from the amino acid–sensing domains. Upon loss of retromer, RAB7 expands into the ragulator-decorated amino acid–sensing domains and interferes with RAG-GTPase and mTORC1 recruitment. Depletion of retromer in Caenorhabditis elegans reduces mTORC1 signaling and extends the lifespan of the worms, confirming an evolutionarily conserved and unexpected role for retromer in the regulation of mTORC1 activity and longevity.

Retromer and the associated actin-polymerizing WASH complex are essential for the endocytic recycling of a wide range of integral membrane proteins. A hereditary Parkinson's-disease-causing point mutation (D620N) in the retromer subunit VPS35 perturbs retromer's association with the WASH complex and also with the uncharacterized protein ankyrin-repeat-domain-containing protein 50 (ANKRD50). Here, we firmly establish ANKRD50 as a new and essential component of the SNX27-retromer-WASH super complex. Depletion of ANKRD50 in HeLa or U2OS cells phenocopied the loss of endosome-to-cell-surface recycling of multiple transmembrane proteins seen upon suppression of SNX27, retromer or WASHcomplex components. Mass-spectrometry-based quantification of the cell surface proteome of ANKRD50-depleted cells identified amino acid transporters of the SLC1A family, among them SLC1A4, as additional cargo molecules that depend on ANKRD50 and retromer for their endocytic recycling. Mechanistically, we show that ANKRD50 simultaneously engages multiple parts of the SNX27-retromer-WASH complex machinery in a direct and co-operative interaction network that is needed to efficiently recycle the nutrient transporters GLUT1 (also known as SLC2A1) and SLC1A4, and potentially many other surface proteins.