ATG5 interacts with the small coiled-coil protein ATG16L1 (Mizushima et al. 1999, Mizushima et al. 2003: mouse proteins were used; Otomo et al. 2013, Kim et al. 2015: human proteins were used). The N-terminus of ATG16L1 binds to the ubiquitin-folds of ATG5, facilitating the formation of the ATG12:ATG5:ATG16L1 complex (Otomo et al. 2013, Kim et al. 2015). The resulting ATG12:ATG5:ATG16L1 complex multimerizes through homotypic interactions of the coiled-coil domain of ATG16L1 (Mizushima et al. 1999, Mizushima et al. 2003: mouse proteins were used; Wible et al. 2019: human proteins were used). The molecular weight of this multimeric complex suggests that it probably represents a tetramer of ATG12:ATG5:ATG16L1 (Kuma et al. 2002: mouse proteins were used).

In multiple human cell lines, ATG16L1 orthologue, ATG16L2, was shown to compete with ATG16L1 for binding to the ATG5:ATG12 complex and to interfere with the formation of autophagosomes (Wible et al. 2019; reviewed in Don Wai Luu et al. 2022). Knockdown of ATG16L2 by siRNA in human neuroblastoma cell line H4 increases the autophagy flux (Lipinski et al. 2010). In mouse embryonic fibroblasts (MEFs), ATG16L2 was shown to bind to the ATG12:ATG5 complex through interaction with ATG5, with a similar affinity as ATG16L1 (Ishibashi et al. 2011). ATG16L2, however, is not required for starvation-induced autophagy in MEFs and cannot substitute for ATG16L1 (Ishibashi et al. 2011). While the loss of Atg16l1 is embryonic-lethal in mouse, loss of Atg16l2 is not (Khor et al. 2019). Conditional double knockout of Atg16l1 and Atg16l2 in mouse T lymphocytes shows that their functions are not redundant and that Atg16l2 loss does not exacerbate Atg16l1 knockout phenotype (Khor et al. 2019). In intestinal epithelium, loss of Atg16l2 reverses the phenotype of a conditional Atg16l1 knockout, suggesting that the role of Atg16l2 in modulating Atg16l1 function is tissue-specific (Khor et al. 2019). In mouse bone marrow-derived macrophages, Atg16l2 is needed to promote autophagy under stress conditions but not under resting conditions, and Atg16l2 may act by promoting binding of Atg16l1 to Atg5 through an unknown mechanism (Wang et al. 2022).