Abstract
The conserved target of rapamycin (TOR) kinase is a central regulator of cell growth in response to nutrient availability. TOR forms 2 structurally and functionally distinct complexes, TORC1 and TORC2, and negatively regulates autophagy via TORC1. Here we demonstrate TOR also operates independently through the TORC2 signaling pathway to promote autophagy upon amino acid limitation. Under these conditions, TORC2, through its downstream target kinase Ypk1, inhibits the Ca2+- and Cmd1/calmodulin-dependent phosphatase, calcineurin, to enable the activation of the amino acid-sensing EIF2S1/eIF2α kinase, Gcn2, and promote autophagy. Thus TORC2 signaling regulates autophagy in a pathway distinct from TORC1 to provide a tunable response to the cellular metabolic state.
Keywords: amino acid, calcineurin, GAAC, Gcn2, Gcn4, nitrogen, TOR, TORC1, TORC2, Ypk1
The TOR kinase plays a central role in a conserved signaling network that couples both intra- and extracellular nutritional status to growth in eukaryotic organisms. TOR assembles with a partially overlapping set of proteins to form 2 structurally and functionally distinct complexes, TORC1 and TORC2, which collectively regulate spatial and temporal aspects of cell growth and metabolism. TORC1 plays a central role in balancing anabolic and catabolic processes, and is a well-established negative regulator of autophagy. The sensitivity of TORC1 to nitrogen starvation or the macrolide antibiotic rapamycin has greatly facilitated an understanding of the mechanisms behind autophagy regulation. However, unlike TORC1, TORC2 is relatively insensitive to rapamycin and as a consequence its cellular functions have been less well characterized. In our recent publication, we identified a positive role for TORC2 in mediating autophagy during amino acid starvation in a pathway that is distinct from TORC1. In particular, we demonstrate that an intimate link exists between TORC2 signaling and activation of the general amino acid control (GAAC) response, which is essential for the induction of autophagy following amino acid, but not nitrogen, starvation.
We utilized the model yeast S. cerevisiae to investigate a role for TORC2 in regulating autophagy under various nutrient-limiting conditions. To monitor autophagy, we utilized a common assay that takes advantage of the GFP-Atg8 chimeric protein that associates with autophagosomes and is degraded in the vacuole to generate free GFP. This system allows for the quantification of autophagy-mediated protein turnover, or autophagy flux. We first examined the effect of inhibiting TORC2 signaling following nitrogen starvation or treatment with rapamycin, conditions that induce autophagy by inhibition of TORC1 activity. We observed that strains harboring either a temperature-sensitive allele of TORC2 (TORC2-ts) or deletion of the downstream target kinase, Ypk1 (ypk1Δ), produce a wild type-like autophagy response. Thus, TORC2 signaling is not required for autophagy flux following TORC1 inhibition and does not negatively regulate autophagy. Accordingly, we next examined the role of TORC2 signaling in mediating the autophagy response during amino acid starvation, which induces autophagy via a mechanism that is independent of TORC1. Significantly, in contrast to wild-type cells, both TORC2-ts and ypk1Δ cells exhibit greatly reduced autophagy flux during amino acid starvation. Thus, our data support a model wherein TORC2-Ypk1 function to promote autophagy specifically following amino acid limitation (Fig. 1). As inhibition of TORC1 activity by treatment with rapamycin or following nitrogen starvation induces an autophagy response comparable to wild type in both TORC2-ts and ypk1Δ cells, this rules out the possibility that the autophagy machinery is impaired by these mutations and further indicates that TORC2 promotes autophagy in a pathway of its own that is distinct from TORC1.
Figure 1.

A model for TORC2 regulation of autophagy and the GAAC response during amino acid starvation. TORC2-Ypk1 promote autophagy during amino acid starvation in a pathway distinct from TORC1. Specifically, TORC2-Ypk1 promote autophagy by negatively regulating Cmd1/calmodulin-calcineurin, whose activity inhibits the GAAC response (Gcn2-Gcn4) required for amino acid starvation-induced autophagy.
We extended our findings by examining a role for the Ca2+- and Cmd1/calmodulin-dependent phosphatase calcineurin as a regulator of autophagy downstream of TORC2-Ypk1. We and others have shown previously that disruption of TORC2 signaling increases the level of calcineurin activity. Remarkably, we found that this increase in calcineurin activity is responsible for the block in autophagy induction in ypk1Δ cells, as deleting the gene encoding the calcineurin regulatory subunit, CNB1, results in a wild-type-like autophagy response during amino acid starvation. Consistently, inhibition of the upstream calcineurin activator, Cmd1/calmodulin, also significantly increases autophagy flux in ypk1Δ cells. These results support a model wherein TORC2/Ypk1 signaling function to promote autophagy through the suppression of Cmd1/calmodulin-calcineurin, whose activity in turn suppresses autophagy during amino acid starvation (Fig. 1).
Amino acid starvation leads to activation of the GAAC response to reduce translation, promote amino acid biosynthesis, and activate adaptive cellular responses. Specifically, increased levels of uncharged tRNAs bind to directly, and activate, the conserved EIF2S1 kinase, Gcn2, which indirectly activates the transcription factor Gcn4. Interestingly, the GAAC response is required for autophagy flux upon amino acid but not nitrogen starvation. As TORC2 and Ypk1 mutants exhibit a similar autophagy defect to that of GAAC mutants, we hypothesized that increased calcineurin activity in TORC2 signaling mutants might block induction of the GAAC response upon amino acid starvation. To test this, we first examined the activity of Gcn2 in wild-type, ypk1Δ, and ypk1Δ cnb1Δ cells by monitoring Gcn2-dependent phosphorylation of EIF2S1 at serine 51. While wild-type cells exhibited a 4-fold increase in EIF2S1 phosphorylation, ypk1Δ cells failed to induce Gcn2 activity following amino acid starvation. However, reducing calcineurin activity in ypk1Δ cells restored Gcn2 activity to that of wild type. Consistently, similar results were observed when we monitored the translational activation of Gcn4 during amino acid starvation in these cells. We then asked whether activation of the GAAC response was sufficient to promote autophagy in TORC2 signaling mutants. Indeed, expression of a partially constitutively active allele of Gcn2 (Gcn2S577A) in ypk1Δ cells significantly increased autophagy flux. Taken together, our findings support a model wherein TORC2-Ypk1 promote the GAAC response and autophagy through inhibition of calcineurin, which we demonstrate is a negative regulator of Gcn2 activity (Fig. 1). Additional studies will be required to explore this new link between calcineurin and Gcn2 and to determine whether calcineurin phosphatase activity impairs activation of Gcn2 directly or acts indirectly via upstream regulatory components.
Although TOR kinase has been largely attributed as a negative regulator of autophagy through TORC1, our study establishes TORC2 as an independent positive regulator of autophagy during amino acid starvation. Thus TOR functions in 2 distinct complexes to induce an autophagy response that is in tune with the cellular metabolic state. Our study further identifies crucial mechanistic differences between nitrogen and amino acid starvation, and sheds light on the inhibitory effects of calcium stress signaling on the GAAC response during amino acid starvation. Given the link between TOR signaling and autophagy in the onset and progression of metabolic diseases, our proposed model involving TORC2 as a positive regulator of autophagy may shed light on the mechanisms related to such diseases and offer novel avenues for the development of more effective therapeutic strategies in humans.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Funding
This work was supported by National Institutes of Health (NIH) Grant GM086387 and NIH T-32 Training Grant in Molecular and Cellular Biology.
