Ubiquitin-like (UBL) protein conjugations, such as ubiquitination and sumoylation, alter protein fate, localization, function and transcriptional activity. Two UBL protein conjugation systems, the ATG12 and ATG8 systems, are required for macroautophagy. The early steps of autophagosome formation require that these two conjugation systems covalently attach: (1) the UBL ATG12 to the target protein ATG5 (Fig. 1A); and (2) the UBL ATG8 to the lipid phosphotidylethanolamine (PE) (Fig. 1B). Interestingly, since the initial discovery of the ATG12 conjugation system, ATG5 has remained as the only known target of ATG12. Recently, we demonstrated that ATG12 can modify multiple protein targets in mammalian cells, and identified ATG3, the E2-like enzyme that conjugates ATG8 to PE, as a second ATG12 conjugation target.
Although ATG12 and ATG3 have been previously reported to physically interact, to our knowledge, we provide the first evidence for a covalent complex between these two essential components of the autophagy conjugation machineries. Similar to ATG12-ATG5, formation of the ATG12-ATG3 complex requires activation of ATG12 by the E1-like enzyme ATG7. Subsequently, ATG3 autocatalyzes the conjugation of ATG12 onto a single conserved lysine residue (K243) of this E2-like enzyme (Fig. 1A, see gray box). Based on immunoprecipitation studies, both free ATG3 and the ATG12-ATG3 complex are concomitantly detected in mammalian cells; notably, only a fraction of the total pool of endogenous ATG3 is in a complex with ATG12. This contrasts with ATG5, which appears to be completely conjugated to ATG12.
To gain insight into the cellular function of the ATG12-ATG3 complex, we complemented atg3-deficient mouse fibroblasts with either wild-type ATG3 (WTATG3) or a mutant version of ATG3 that is unable to be conjugated to ATG12 due to loss of a single lysine required for isopeptide bond linkage (ATG3KR). Because ATG12, in complex with ATG5, and ATG3 both play critical roles in the biogenesis and elongation of the autophagosomal membrane, we hypothesized that the ATG12-ATG3 complex would likewise regulate the early steps of autophagy. However, upon disrupting ATG12 conjugation to ATG3, we observed no effects on the lipidation of LC3 or other mammalian ATG8 orthologues. We also found no significant changes in nonselective autophagy in response to multiple stresses, including nutrient starvation, rapamycin-mediated mTOR inhibition and carbonyl cyanide m-chlorophenylhydrazone (CCCP)-induced mitochondrial depolarization.
Although nonselective autophagy appears intact, our results indicate that the disruption of ATG12 conjugation to ATG3 in cells does have profound effects on mitochondria. ATG3KR cells possess increased mitochondrial mass relative to WTATG3 cells, both at baseline and in response to mitochondrial depolarization with CCCP. Upon mitochondrial depolarization, mitochondria targeting to autophagosomes in ATG3KR cells is significantly reduced relative to WTATG3 cells, suggesting a defect in mitochondrial autophagy or mitophagy. To date, the prevailing view has been that specific adapter molecules target mitochondria for autophagic degradation in mammalian cells. For example, the two related BH3 family proteins, BNIP3 and NIX/BNIP3L, promote mitochondrial clearance during reticulocyte development, whereas the PINK1-Parkin pathway directs mitophagy in response to depolarization. In contrast, the ATGs comprising the conjugation systems have not been previously implicated in the selective targeting of mitochondria; rather, they are proposed to contribute more generally by forming the autophagosome. Our results now broach the hypothesis that this newly identified complex between two conjugation ATGs mediates a unique function for selectively targeting mitochondria for autophagic degradation.
In addition, we discovered that the mitochondrial network of ATG3KR cells is highly fragmented. Mitochondrial fragmentation is similarly present in atg3-deficient cells, whereas cells reconstituted with WTATG3 exhibit tubular mitochondrial networks. Mitochondria undergo extensive remodeling through continuous cycles of mitochondrial fusion and fission; appropriate balance of these two opposing processes is necessary for maintenance of overall mitochondrial structure and function. Mitochondrial fragmentation often indicates an imbalance in mitochondrial fission and fusion; in support of this prediction, cells lacking the ATG12-ATG3 complex display reduced mitochondrial fusion activity. Nonetheless, further studies are required to clarify if ATG12-ATG3 serves as a novel link between autophagy and the mitochondrial fusion machinery. Importantly, one can alternatively hypothesize that the effects of ATG12-ATG3 on mitochondrial fusion are secondary to its effects on mitophagy, since previous work indicates that fragmented mitochondria are preferentially degraded by mitophagy. In other words, due to reduced mitophagy, disruption of ATG12 conjugation to ATG3 results in the build-up of fusion defective mitochondria.
Finally, ATG12 conjugation to ATG3 sensitizes cells to apoptosis mediated by mitochondrial pathways, but has no effect on death-receptor mediated apoptosis. The resistance to intrinsic apoptosis in ATG3KR cells correlates with an increase in the basal protein levels of antiapoptotic Bcl-2 proteins, namely Bcl-xL. Accordingly, specific antagonism of antiapoptotic Bcl-2 family members with obatoclax, a pharmacological BH3 mimetic, causes equivalent levels of cell death between WTATG3 and ATG3KR cells. An important question for future study is whether these increases in Bcl-XL also contribute functionally to the changes in mitochondrial mass and morphology found in ATG3KR cells.
Dissecting the exact mechanisms through which the ATG12-ATG3 complex mediates these diverse effects on mitochondrial homeostasis, morphology and cell death remains an important topic for future investigation. Since a fraction of total ATG3 is in complex with ATG12, this motivates the hypothesis that ATG12 conjugation to ATG3 promotes the relocation of this complex to mitochondria. Alternatively, ATG12 conjugation may change ATG3 substrate specificity, alter its enzymatic activity, or facilitate new protein-protein interactions. Importantly, our results support the idea that the effects of the ATG12-ATG3 complex on mitochondrial homeostasis and cell death are unique functions that can be completely separated from the established roles of each individual ATG in autophagosome formation. Although this is a novel finding for ATG12 and ATG3, other studies demonstrate that ATGs may control cellular functions distinct from autophagy. For example, during phagocytosis, LC3 becomes recruited to, and incorporated into, the plasma membrane by a macroautophagy-independent mechanism that requires both ATG5 and Beclin 1. Hence, we propose that ATG12, like ubiquitin and SUMO, may serve as a broadbased protein modification that mediates diverse cellular functions in addition to autophagy.