PerspectiveCell Biology

mTORC1 Phosphorylates the ULK1-mAtg13-FIP200 Autophagy Regulatory Complex

See allHide authors and affiliations

Science Signaling  18 Aug 2009:
Vol. 2, Issue 84, pp. pe51
DOI: 10.1126/scisignal.284pe51

Abstract

High nutrient availability stimulates the mammalian target of rapamycin complex 1 (mTORC1) to coordinately activate anabolic processes, such as protein synthesis, while inhibiting the cellular catabolism of autophagy. Positive regulation of protein synthesis through the mTORC1 substrates p70 ribosomal S6 kinase (p70S6K) and eukaryotic initiation factor 4E binding protein 1 (4E-BP1) has been well characterized. The complementary inhibitory mechanism in which mTORC1 phosphorylates the autophagy regulatory complex containing unc-51-like kinase 1 (ULK1), the mammalian Atg13 protein, and focal adhesion kinase interacting protein of 200 kD (FIP200) has also been elucidated.

Macroautophagy (herein referred to as autophagy) is an intracellular degradative pathway that is a key metabolic homeostatic response to nutrient starvation. During this process, cytoplasmic proteins and organelles are engulfed by expanding membranes to eventually produce enclosed specialized vesicles called autophagosomes, which subsequently fuse with late endosomal and lysosomal hydrolytic compartments. Not only does autophagy recycle basic biomolecular building blocks to support the survival of starving cells, but basal autophagy is also critical for the constitutive removal of damaged proteins and organelles that otherwise accumulate with toxic consequences. Reflecting these essential functions, autophagy affects many biomedical processes, prompting intense interest in understanding its fundamental regulation (1, 2). A series of recent reports advances our understanding of the regulation of autophagy by providing a comprehensive characterization of the mammalian Atg1 complex and, more strikingly, by defining a direct mechanism linking the Atg1 complex to TOR (target of rapamycin) (36).

TOR is a serine-threonine kinase that plays a conserved role, from yeast to humans, in balancing the production and removal of cell mass (7, 8). The mammalian TOR (mTOR) protein is ~280 kD in size and exists in one of two complexes. mTORC1 (mTOR complex 1) contains mTOR along with the subunits Raptor and mLST8, and this complex is sensitive to nutrient (amino acid) abundance and the drug rapamycin. mTORC1 activation stimulates cell growth by promoting protein translation through its substrates p70 ribosomal protein S6 kinase (p70S6 kinase) and eukaryotic initiation factor 4E (eIF-4E) binding protein (4E-BP1, an inhibitor of the elongation factor eIF-4E). mTORC1 can also promote ribosome biogenesis, through transcriptional mechanisms. The second mTOR complex, mTORC2, which is not responsive to nutrients or rapamycin, contains mTOR along with the subunits Rictor, Sin1, and mLST8. mTORC2 regulates the actin cytoskeleton through a distinct set of substrates, including the serine-threonine kinase Akt, serum glucocorticoid–induced protein kinase (SGK), and protein kinase C (PKC).

Increasing concentrations of amino acids activate mTORC1 by promoting its binding to the Rag and Rheb GTPases (guanosine triphosphatases). The regulation by amino acids can be further modulated by insulin and insulin-like growth factor 1, which act through phosphoinositide 3-kinase and Akt, and cellular energy status, which is communicated through AMPK (adenosine 5′-monophosphate–activated protein kinase). Under amino acid–rich conditions, activated mTORC1 promotes protein synthesis while coordinately inhibiting cellular degradation by autophagy. Amino acid starvation or rapamycin treatment relieves this negative regulation of autophagy.

The serine-threonine kinase Atg1 was first identified in yeast screens for essential Autophagy (Atg) genes (9). Further work solidified Atg1 as a dynamic signaling protein that functions downstream of TOR (1012). Inhibition of yeast TORC1 induces formation of an activated Atg1 complex containing the cofactors Atg13 and Atg17, both of which are required for maximal Atg1 catalytic activity. TORC1 activation leads to Atg13 phosphorylation, destabilizing the complex and effectively inactivating Atg1. However, the kinase(s) that phosphorylates Atg13 and the substrates of Atg1 are not yet identified in yeast.

Insight into the mechanism by which TOR regulates the Atg1-Atg13-Atg17 complex has been provided by studying autophagy in mammalian cells. Mammalian Atg13 (herein referred to as mAtg13), which displays loose sequence conservation with its yeast homolog, performs a similar function by binding, stabilizing, and promoting catalytic activity of the mammalian Atg1 homolog, ULK1 (unc-51-like kinase 1) (4, 5). FIP200 (focal adhesion kinase interacting protein of 200 kD) is another essential component of the mammalian complex that is required for stability and full catalytic activity of ULK1 (5, 13). Although there is no sequence homology, FIP200 may be the functional analog of yeast Atg17 (13). ULK1 directly phosphorylates both mAtg13 and FIP200, thereby providing the best evidence to date of Atg1 substrates with clear roles in autophagy (3). ULK1, mAtg13, and FIP200 form a large, stable complex with an apparent molecular mass of >3 MD (as determined by gel filtration) (4). However, the stability of this ULK1-mAtg13-FIP200 complex is insensitive to nutrient conditions, which would not have been predicted based on the behavior of the yeast Atg1-Atg13-Atg17 complex (11, 12).

mTORC1 associated with the ULK1-mAtg13-FIP200 complex by a direct interaction between Raptor and ULK (4). ULK did not bind the mTORC2 component Rictor, so this direct link to the autophagic machinery is limited to mTORC1. Atg1 also binds TOR in Drosophila melanogaster, indicating that this direct binding is conserved throughout the metazoans (6). Nutrients robustly promoted a substantial fraction of mTORC1 to join the >3-MD complex. mTORC1 phosphorylates mAtg13 and ULK1, both in vitro and in intact cells (35). As in yeast, mTORC1 activation and inhibition of autophagy correlate with ULK1 inhibition and reduced phosphorylation of ULK1 substrates, such as FIP200 (3).

A working model can be proposed wherein amino acid–rich conditions activate and translocate mTORC1 to the ULK1-mAtg13-FIP200 complex (Fig. 1). mTORC1 then phosphorylates ULK1 and mAtg13 at inhibitory sites to repress the kinase activity of ULK1, thereby suppressing the initiation of autophagy. Although this scheme has a straightforward appeal, critical details need to be ironed out. The heterogeneous mobility of ULK1 on protein gels, the large number of potential phosphorylation sites within the serine-rich ULK1 spacer region, and the fact that ULK1 readily autophosphorylates all suggest that multiple phosphorylation events likely occur in vivo (14). The inhibitory sites in ULK1 that are phosphorylated by mTORC1 and, furthermore, the ULK1 autophosphorylation sites (presumably activating), need to be mapped. Similarly, mAtg13 is phosphorylated by mTORC1 under nutrient-rich conditions (35) and by ULK1 (most likely under starvation conditions) (3, 4, 15), and these sites need to be determined. The phosphorylation state of these specific sites on ULK1 and mAtg13 then need to be monitored under different nutrient conditions and their effect on the regulation of autophagy determined. Disruption of the Raptor binding site in ULK1 might also create a constitutively active autophagy activator that is insensitive to nutrient concentrations.

Fig. 1

A model of how mTORC1 regulates catabolic and anabolic processes. (A) Under amino acid–rich conditions, mTORC1 inhibits autophagy and promotes protein synthesis through its action on different substrates. mTORC1 associates with the ULK1-mAtg13-FIP200 complex and phosphorylates ULK1 and mAtg13 on sites that inhibit their activity. (B) Under starvation conditions, mTORC1 dissociates from the ULK1-mAtg13-FIP200 complex, causing ULK1 to become active, autophosphorylate, and phosphorylate mAtg13 and FIP200, thus promoting the autophagy complex to translocate to sites of autophagosome formation. Molecules that are less active in each condition are shown with faded colors.

It is currently unclear what signals are generated downstream of ULK1 to trigger autophagy. One possibility is that subcellular redistribution of the ULK1-mAtg13-FIP200 complex constitutes this molecular signal. As in yeast, autophagy induction in mammalian cells requires the recruitment of protein factors to pre-autophagosomal sites, where they direct autophagosomal membrane formation. Several specific marker proteins have been validated for visualizing forming autophagosomes, such as LC3 (a mammalian homolog of the yeast Atg8 protein), Atg5, and Atg16. Following the activation of autophagy, ULK1, mAtg13, and FIP200 all rapidly translocate from a diffuse cytoplasmic pool to autophagosomal structures (4, 13). Because ULK1, mAtg13, and FIP200 remain in a stable high molecular weight complex that is insensitive to nutrients, they likely translocate together to autophagosomal sites. It is possible that FIP200 phosphorylation by activated ULK1 might control the subcellular redistribution of the entire complex. Consistent with this idea, kinase-dead ULK1 mutants, which are dominant inhibitors of autophagy, do not localize properly after cells are exposed to conditions that should trigger autophagy (13, 15). Alternatively, the decrease in mAtg13 phosphorylation as a consequence of mTORC1 inhibition might direct translocalization of the complex. Future experiments should determine if FIP200 with phosphomimetic mutations or mAtg13 lacking its key phosphorylation sites appropriately localize ULK complexes and activate autophagy.

It is also possible that another ULK substrate transmits the autophagy activation signal. The actin-associated signaling protein paxillin regulates autophagy, possibly as an Atg1 substrate (16). In addition, several other Atg1 substrates have been identified in Caenorhabditis elegans (UNC-14 and VAB-8) and Drosophila (Unc-76), but these are so far only understood to regulate endocytic trafficking and outgrowth in neurons (17, 18). Mammalian ULK proteins also play roles in neuronal vesicular transport and development (1921). At present, it is a still a mystery how Atg1 proteins coordinately regulate vesicular trafficking and autophagy. Furthermore, it is unknown if regulation of ULK1 through TORC1 might balance these two functions.

We now appreciate that mTORC1 activated by high nutrient concentrations has a distinct set of substrates that target the ULK1 pathway to inhibit autophagy. Because the molecular characterization of the multifunctional ULK proteins is ongoing, our understanding of the mTORC1-ULK pathway will need to incorporate additional complexities. ULK1 is the mammalian Atg1 homolog best characterized in autophagy, but genome data suggest at least four more potential mammalian Atg1 homologs (22). ULK2 and ULK3 have roles in autophagy, although they are not functionally equivalent to ULK1 (15, 23, 24). From limited data, it is clear that some of the mechanisms involving mAtg13, FIP200, and mTORC1 are conserved for ULK2 (3, 13). In addition, at least two mAtg13 splice variants each bind to both ULK1 and 2, increasing the total number of possible complexes (3). ULK1 also has other binding partners, such as Atg101, synaptic GTPase-activating protein (SynGAP), and Syntenin (20, 25). ULK proteins can be polyubiquitinated and subsequently associated with sequestosome-1 (also known as p62/sequestosome-1), a polyubiquitin binding protein (19). These all present potential mechanisms that could further modulate signals through the mTORC1-ULK pathway.

Acknowledgments

26. E.Y.C. is supported by funding from the University of Strathclyde.

References and Notes

View Abstract

Navigate This Article