Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 13 (2), 132-41

AMPK and mTOR Regulate Autophagy Through Direct Phosphorylation of Ulk1

Affiliations

AMPK and mTOR Regulate Autophagy Through Direct Phosphorylation of Ulk1

Joungmok Kim et al. Nat Cell Biol.

Abstract

Autophagy is a process by which components of the cell are degraded to maintain essential activity and viability in response to nutrient limitation. Extensive genetic studies have shown that the yeast ATG1 kinase has an essential role in autophagy induction. Furthermore, autophagy is promoted by AMP activated protein kinase (AMPK), which is a key energy sensor and regulates cellular metabolism to maintain energy homeostasis. Conversely, autophagy is inhibited by the mammalian target of rapamycin (mTOR), a central cell-growth regulator that integrates growth factor and nutrient signals. Here we demonstrate a molecular mechanism for regulation of the mammalian autophagy-initiating kinase Ulk1, a homologue of yeast ATG1. Under glucose starvation, AMPK promotes autophagy by directly activating Ulk1 through phosphorylation of Ser 317 and Ser 777. Under nutrient sufficiency, high mTOR activity prevents Ulk1 activation by phosphorylating Ulk1 Ser 757 and disrupting the interaction between Ulk1 and AMPK. This coordinated phosphorylation is important for Ulk1 in autophagy induction. Our study has revealed a signalling mechanism for Ulk1 regulation and autophagy induction in response to nutrient signalling.

Figures

Figure 1
Figure 1
Glucose starvation activates Ulk1 protein kinase through AMPK-dependent phosphorylation. (a) HEK293 cells were starved of glucose (4 h) as indicated, endogenous Ulk1 was immunoprecipitated and an autophosphorylation assay was performed. Proteins were resolved by SDS–PAGE and visualized with autoradiography (top) or western blotting (WB; bottom). (b) Cells were incubated in glucose-free medium (4 h) as indicated and lysed. Lysates were incubated with lambda phosphatase (λ PPase) as indicated. Endogenous Ulk1 mobility was examined by western blotting. (c) HA–Ulk1 was transfected into HEK293 cells together with wild-type (WT) AMPKα1 or a kinase-dead (DN) mutant. Cells were starved of glucose (4 h; Glu) or amino acids (−A.A) and treated with compound C (20 µM, C.C) as indicated. Ulk1 mobility as well as phosphorylation levels of ACC and S6K were determined by western blotting. (d) HA–Ulk1 proteins were immunopurified from transfected HEK293 cells, which had undergone glucose starvation (4 h) as indicated. The HA–Ulk1 proteins were treated with λ PPase, and in vitro kinase assays were performed in the presence of GST–ATG13. Proteins were resolved by SDS–PAGE; phosphorylated proteins were visualized with autoradiography, HA–Ulk1 by western blotting and GST–Atg13 by Coomassie staining. (e) HA–Ulk1 was immunopurified from transfected HEK293 cells under glucose-rich media and treated with AMPK in the presence of cold ATP for 15 min, followed by kinase assays as described in d. (f) AMPK wild-type (WT) and α1/α2 double knockout (DKO) MEFs were incubated with or without glucose (4 h). Endogenous Ulk1 was immunoprecipitated and autophosphorylation was measured (mean ± s.d., n = 3). Autophosphorylation activity was normalized to Ulk1 protein level; relative activity is calculated by normalization to Ulk1 activity from AMPK wild-type MEFs in glucose-rich conditions. (g) HA–Ulk1 was transfected into HEK293 cells together with vector (Vec) or an AMPKα1 kinase-dead mutant (DN). The cells were starved of glucose (−Glu) or amino acids (−A.A), or treated with 50 nM rapamycin (Rapa) for 3 h before lysis. Left: autophosphorylation activity was assessed and normalized as in f (mean ± s.d., n = 3). Right: fold induction in Ulk1 autophosphorylation, compared with Ulk1 autophosphorylation from cells under nutrient-rich conditions. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 2
Figure 2
AMPK directly phosphorylates Ulk1 at Ser 317 and Ser 777. (a) AMPK phosphorylates the Ulk1 S/T domain in vitro. Top: schematic representation of Ulk1 domain structure and deletion constructs used to map phosphorylation sites. The mouse Ulk1 protein consists of an N-terminal kinase domain (KD; 1–278), serine/threonine-rich domain (S/T domain, 279–828), and C-terminal domain (CTD, 829–1051). Bottom: the indicated Flag–Ulk1 deletion mutants were immunopurified from transfected HEK293 cells and were used for in vitro AMPK assay as a substrate. Phosphorylation was examined by 32P-autoradiogram and protein level was determined by western blot. (b) Determination of AMPK phosphorylation sites in Ulk1. The indicated recombinant GST–Ulk1 mutants were expressed and purified from Escherichia coli, and used as substrates for in vitro phosphorylation by AMPK. Deletion analyses indicated that two Ulk1 fragments in the S/T domain, 279–425 and 769–782, were highly phosphorylated by AMPK in vitro. Mutation of Ser 317 abolished the majority of phosphorylation in the Ulk1 fragment 279–425. Within the fragment 769–782, mutations of five serine residues (Ser 774, Ser 777, Ser 778, Ser 779 and Ser 780) to alanine, denoted as (769–782) 5SA, completely abolished the phosphorylation by AMPK. Reconstitution of Ser 777 in this mutation background, (769–782) 4SA-S777, but not any of the other four residues, restored the phosphorylation by AMPK. GST and GST–TSC2F (TSC2 fragment 1300–1367 containing AMPK phosphorylation site at Ser 1345) were used as negative and positive controls for AMPK reaction, respectively. Phosphorylation was determined by 32P-autoradiograph and the protein levels were detected by Coomassie staining. (c) Ser 317/Ser 777 are required for glucose-starvation induced Ulk1 phosphorylation in vivo. HA–Ulk1 and mutants were transfected into HEK293 cells. Cells were starved for glucose for 4 h as indicated. HA–Ulk1 was immunoprecipitated and examined by western blot for mobility. (d) Phosphorylation of Ulk1 Ser 317 and Ser 777 are induced by AMPK. Wild-type HA–Ulk1 or S317/777A mutant were co-transfected with AMPK into HEK293 cells as indicated. HA–Ulk1 was immunoprecipitated (IP) and phosphorylation of Ser 317 and Ser 777 were determined by western blotting. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 3
Figure 3
AMPK-dependent Ulk1 Ser 317 and Ser 777 phosphorylation is required for Ulk1 activation in response to glucose starvation. (a) AMPK wild-type or DKO MEFs were starved of glucose (4 h) as indicated. Total cell lysates were probed for Ulk1 protein and phosphorylation. (b) Time course of Ulk1 Ser 317 and Ser 777 phosphorylation in response to glucose starvation/re-addition. MEFs were starved of glucose (−Glu) for the indicated times. After 3 h starvation, the culture was switched to glucose-containing (25 mM) medium and samples were harvested (Re-Glu). In parallel, cells were treated with amino-acid-free (−A.A) medium or 50 nM rapamycin (Rapa) for 3 h. (c) Phosphorylation of Ulk1 Ser 317 and Ser 777 correlates with AMPK activity. MEFs were starved of glucose (4 h) as indicated in the presence or absence of 20 µM compound C (C.C). In parallel, cells were treated with 2 mM Metformin (Met, 2 h) in glucose-rich medium. Phosphorylation of ACC S79 was tested as a positive control for AMPK activation. (d) Ulk1 is highly phosphorylated at Ser 317 and Ser 777 by glucose starvation in vivo. To determine the Ulk1 phosphorylation level in vivo, immunopurified HA–Ulk1 protein was phosphorylated by AMPK in vitro (100% represents full phosphorylation of Ulk1 by AMPK). In vitro phosphorylated HA–Ulk1 was diluted as indicated, and was immunoblotted along with the immunoprecipitated HA–Ulk1 from cells grown in either glucose-rich (+ Glu) or glucose-free (− Glu, 4 h) medium. The density of the bands was then quantified. By this measurement, approximately 50% of Ulk1 isolated from glucose-starved cells was phosphorylated on Ser 317 and Ser 777. (e) The indicated HA–Ulk1 proteins were immunopurified from transfected HEK293 cells grown in high-glucose medium, and then incubated with AMPK in the presence of cold ATP for 15 min in vitro. After the reaction, AMPK was removed by extensive washing, the resulting Ulk1 immuno-complexes were assayed for kinase activity in the presence of 32P-ATP. (f) HA–Ulk1 proteins (wild type or S317/777A mutant) were immunoprecipitated from the transfected HEK293 cells, which were incubated with or without glucose (4 h) before lysis. An in vitro kinase reaction was performed in the presence of GST–ATG13 and FIP200. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 4
Figure 4
mTORC1 disrupts the Ulk1–AMPK interaction. (a) AMPK interacts with Ulk1. HEK293 cells were transfected with the various Flag–Ulk1 deletion mutants together with AMPK α/β/γ, Atg13 and FIP200. Flag–Ulk1 protein (indicated by white arrows) was immunoprecipitated and co-immunoprecipitation of AMPK α/β/γ, Atg13 and FIP200 were examined by western blots. (b) Deletion analysis of Ulk1 regions responsible for AMPK interaction. The indicated Flag–Ulk1 truncation mutants were immunoprecipitated from transfected HEK293 cells co-expressing AMPK complex (α/β/γ). Co-immunoprecipitation of AMPK subunits was determined by western blots. (c) Rheb inhibits the Ulk1–AMPK interaction. HA–AMPKα, Flag–Ulk1 and Myc–Rheb were co-transfected into HEK293 cells as indicated. Cells were treated with or without rapamycin (50 nM Rapa) for 1 h before lysis. Flag–Ulk1 was immunoprecipitated and co-immunoprecipitates of AMPKα were determined by western blot. (d) Rapamycin treatment enhances the interaction of endogenous Ulk1 and AMPK. Endogenous Ulk1 proteins were immunoprecipitated from either Ulk1 or AMPK wild-type and knockout (single-knockout; KO or double-knockout; DKO) MEFs. Treatment with 50 nM rapamycin for 1 h is indicated (Rapa). Co-immunoprecipitation of endogenous AMPKα protein was determined by western blot. The arrow indicates AMPKα protein. (e) Phosphorylation by mTORC1 inhibits the ability of Ulk1 to bind AMPK in vitro. CBP/SBP–Ulk1 was purified from transfected HEK293 cells by streptavidin beads and the Ulk1–bead complex was incubated with mTORC1, which was prepared by Raptor immunoprecipitation, in the presence of cold ATP, as indicated. The resulting Ulk1 complex was incubated with the cell lysates containing AMPK, then extensively washed. The Ulk1 and associated AMPKα were detected by western blot. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 5
Figure 5
mTORC1 phosphorylates Ulk1 at Ser 757. (a) mTORC1 phosphorylates the Ulk1 S/T domain. Ulk1 deletion mutants were prepared from the transfected HEK293 cells and used for in vitro mTORC1 assay. Phosphorylation was examined by 32P-autoradiogram (top) and protein level was determined by western blot (bottom). (b) Ser 757 is phosphorylated by mTORC1. Left: the indicated recombinant GST–mUlk1 mutants were purified from E. coli and used for in vitro mTORC1 assay as substrates. Deletion analyses isolated the fragment (753–771) as a target for mTORC1. The Ulk1 (753–771) fragment contains five conserved serine/threonine residues, Thr 754, Ser 757, Ser 760, Thr 763 and Thr 770. Right: mutation of Ser 757 abolished Ulk1 phosphorylation by mTORC1 in vitro. GST was used as negative control for mTORC1 phosphorylation reaction. Phosphorylation was determined by 32P-autoradiograph (top), whereas protein levels were detected by Coomassie staining (bottom). (c) Rheb increases Ulk1 Ser 757 phosphorylation. HA–Ulk1 wild type and the S757A mutant were immunoprecipitated from transfected HEK293 cells. Co-transfection with Rheb and rapamycin (Rapa) treatment are indicated. Ulk1 Ser 757 phosphorylation was determined by western blot. (d) Rheb induces a mobility shift in wild-type Ulk1, but not the Ulk1S757A mutant. HA–Ulk1 was transfected with or without Rheb into HEK293 cells. HA–Ulk1 was immunoprecipitated from the cells under nutrient-rich medium and Ulk1 mobility was examined by western blot. (e) Endogenous Ulk1 Ser 757 phosphorylation is elevated in Tsc1−/− MEFs. Tsc1+/+ (WT) and Tsc1−/− (KO) MEFs were starved of glucose (4 h), or treated with 50 nM rapamycin (Rapa, 1 h). Ser 757 phosphorylation of endogenous Ulk1 was detected by a phospho-Ulk1 Ser 757 antibody. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 6
Figure 6
Phosphorylation of Ulk1 Ser 757 by mTORC1 inhibits the Ulk1–AMPK interaction. (a) Ulk1 Ser 757 is required for mTORC1 to regulate the interaction of Ulk1 with AMPK in vivo. CBP/SBP tagged Ulk1 (wild type or S757C) was co-transfected with HA–AMPKα and Rheb into HEK293 cells as indicated. Ulk1 was purified by streptavidin beads and the co-precipitated HA–AMPKα was examined by western blot (Rapa, 50 nM rapamycin treatment for 1 h before cell lysis). (b) Ulk1 Ser 757 is required for rapamycin to enhance the Ulk1–AMPK interaction in vitro. CBP/SBP Ulk1 proteins (wild type or S757C) were prepared from transfected HEK293 cells, which were pre-incubated with 50 nM rapamycin (Rapa, 1h) as indicated. The Ulk1 proteins were purified by streptavidin beads and the resulting Ulk1–bead was incubated with the bacterial purified AMPKα/β/γ complex. AMPKα protein levels in the in vitro pulldown assays were examined by western blot using AMPKα antibody. L.E.; long exposure. (c) Phosphorylation of AMPK sites Ser 317 and Ser 777 in Ulk1 are decreased in Tsc1−/− MEFs. Tsc1+/+ (WT) and Tsc1−/− (KO) MEFs were starved of glucose (4 h), or treated with 50 nM rapamycin (Rapa, 1 h). Ser 317 and Ser 777 phosphorylation of endogenous Ulk1 was examined by western blotting with antibodies against Ulk1 phosphorylated at Ser 317 or Ser 777. (d) Rheb suppresses Ulk1 Ser 317 and Ser 777 phosphorylation in a manner dependent on mTORC1. HA–Ulk1, AMPKα kinase-dead mutant (DN), and Myc–Rheb were co-transfected into HEK293 cells as indicated. The cells were incubated with glucose-free medium (−Glu, 4 h), in which either 20 µM compound C (C.C.) or 50 nM Rapamycin (Rapa) was added. Total cell lysates were probed with antibodies against Ulk1 phosphorylated at Ser 317, Ser 777, Ser 757, and HA, as indicated. (e) Rheb inhibits glucose starvation-induced Ulk1 activation. HA–Ulk1 and Myc–Rheb was transfected into HEK293 cells, which were incubated with glucose-free (−Glu), amino-acid-free (−A.A) medium, or 50 nM rapamycin (Rapa) for 4 h before lysis. HA–Ulk1 was immunoprecipitated and kinase assays were performed. Ulk1 activity was measured by 32P-autoradiogram and the protein level of HA–Ulk1 and GST–Atg13 used in the assay was determined by western blot and by Coomassie staining, respectively. Uncropped images of blots are shown in Supplementary Fig. S5.
Figure 7
Figure 7
AMPK phosphorylation is required for Ulk1 function in autophagy on glucose starvation. (a) Ser 317/Ser 777 is required for Ulk1 to protect cells from glucose starvation. Viability (24 h, mean ± s.d., n = 4; top) and PARP cleavage (8 h; western blot, middle; quantification, n = 2, bottom) was examined in Ulk1+/+ (WT), Ulk1−/− (KO), Ulk1−/− re-expressing wild-type Ulk1 (KO-WT), and Ulk1−/− re-expressing Ulk1 S317/777A mutant (KO-S317/777A) MEFs. Arrows in western blot indicate non-cleaved and cleaved PARP. (b) The Ulk1 S317/777A mutant is compromised in LC3 lipidation in response to glucose starvation. ULK1 MEFs were cultured in glucose-free medium for the indicated times. LC3-II level was determined by western blotting and the LC3-II accumulation was normalized by α-tubulin and quantified (bottom, n = 3, mean ± s.d.). A representative western blot was shown. The LC3 antibody used in this experiment seemed to preferentially recognise the lipid-modified form of LC3-II, which migrated faster on the gel. (c) The Ulk1 S317/777A mutant is defective in autophagosome formation. The indicated MEFs were starved of glucose (4 h) and the formation of GFP–LC3-positive autophagosomes was examined by confocal microscopy. GFP–LC3; green and DAPI; blue. Scale bar, 20 µm. (d) Autophagy vacuole analysis by electron microscopy. Low-magnification images of Ulk1−/− (KO, upper left panel), Ulk1−/− reconstituted with wild-type Ulk1 (KO-WT, two middle panels with accompanying higher magnification images), and Ulk1−/− reconstituted with Ulk1 S317/777A (KO-S317/777A, lower left panel) are shown. High-magnification images of autophagosomes from KO-WT are shown in upper right and lower right panels. Autophagosome/autolysosome-like structures indicated by arrowheads on the lower-magnification images and arrows in higher-magnification images. Scale bars; lower-magnification, 1 µm; higher-magnification, 200 nm.
Figure 8
Figure 8
Model of Ulk1 regulation by AMPK and mTORC1 in response to glucose signals. Left: when glucose is sufficient, AMPK is inactive and mTORC1 is active. The active mTORC1 phosphorylates Ulk1 on Ser 757 to prevent Ulk1 interaction with and activation by AMPK. Right: when cellular energy level is limited, AMPK is activated and mTORC1 is inhibited by AMPK through the phosphorylation of TSC2 and Raptor. Phosphorylation of Ser 757 is decreased, and subsequently Ulk1 can interact with and be phosphorylated by AMPK on Ser 317 and Ser 777. The AMPK-phosphorylated Ulk1 is active and then initiates autophagy.

Comment in

Similar articles

See all similar articles

Cited by 1,672 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback