HIF-independent role of prolyl hydroxylases in the cellular response to amino acids

Abstract

Hypoxia-inducible factor (HIF) prolyl hydroxylases (PHDs) are α-ketoglutarate (αKG)-dependent dioxygenases that function as cellular oxygen sensors. However, PHD activity also depends on factors other than oxygen, especially αKG, a key metabolic compound closely linked to amino-acid metabolism. We examined the connection between amino-acid availability and PHD activity. We found that amino-acid starvation leads to αKG depletion and to PHD inactivation but not to HIF stabilization. Furthermore, pharmacologic or genetic inhibition of PHDs induced autophagy and prevented mammalian target of rapamycin complex 1 (mTORC1) activation by amino acids in a HIF-independent manner. Therefore, PHDs sense not only oxygen but also respond to amino acids, constituting a broad intracellular nutrient-sensing network.

Figure 1

Amino-acid starvation decreases αKG levels and inhibit PHD activity. (a) Intracellular αKG levels of U2OS cells after incubation for 3 h under the indicated condition. (b) GFP-ODD accumulation in HCT116 under the indicated condition. (c) Intracellular αKG content of starved cells treated with TαKG as indicated. Non-starved cells were used as control. (d, e) GFP-ODD accumulation examined microscopically (d) or by western blot (e) in fed and starved cells treated with TαKG and rapamycin (RAP) as indicated. (f) Phosphorylation levels of AMPK in U2OS cells incubated in either the presence or absence of amino acids and glucose. Results represent the average and s.d. of three independent experiments (*P<0.05).

Figure 2

Inactivation of PHDs during amino-acid starvation does not activate HIF. (a) Protein levels of HIF1α, REDD1, Glut1 and BNIP3 in U2OS cells incubated in the presence or absence of amino acids for the indicated time. A positive control of cells incubated under hypoxic conditions for 12 h (Hyp) was included. (b) U2OS cells expressing a construct containing either a wild-type HRE-luciferase reporter (black bars) or a mutant HRE-luciferase reporter unable to respond to HIF1α (white bars) were incubated in the presence or absence of amino acids for the indicated time and the luciferase intensity was analyzed using a luminometer (values represent average and s.e. of three independent experiments). A positive control of cells incubated in hypoxic conditions for 12 h was included (U=untransfected). (c) Protein levels of HIF1α and REDD1 in U2OS cells incubated in fed or starving conditions for 2 h followed by 30 min of incubation in the presence or absence of DMOG. (d) U2OS cells where incubated in hypoxia (1% O₂) in starving conditions for 2 h followed by 1 extra hour of incubation in the presence or absence of amino acids. HIF1α accumulation and S6 phosphorylation were then analyzed by western blot. (e, f) U2OS cells stably expressing NF-κB-reporter construct pGL4.32[luc2P/NF-kB-RE/Hygro] were incubated for the indicated time in KRBB either in the presence or absence of all amino acids or tumor necrosis factor-α (10 ng/ml). Activation of NF-κB was analyzed by luminescence analysis of the NF-κB-luciferase reporter (e) and by phosphorylation of IκB kinase (IKK) and IKβ-α by western blot (f). Graph depicts mean and s.d. of three independent experiments and represents fold activation compared with untreated samples.

Figure 3

Inhibition of PHD by DMOG induces autophagy independently of HIF. U2OS cells stably expressing GFP-LC3 were transfected with a plasmid encoding control, nontargeting shRNA (shCntr) or transfected with a plasmid encoding shRNA for HIF1b (clone no. 4; shHIF1b no. 4). After 48 h, cells were incubated for 5 h in the presence of amino acids with or without DMOG as indicated. Aggregation of GFP-LC3 was analyzed by fluorescence microscopy (a) and quantified (b). Results represent the average and s.d. of three independent experiments.

Figure 4

Reduction of intracellular αKG during amino-acid starvation correlates with mTORC1 inhibition. (a, b) Intracellular αKG levels (a) and phosphorylation status of S6K and S6 proteins (b) in U2OS cells at different time points after incubation without amino acids. (c, d) Intracellular αKG levels (c) and phosphorylation status of S6K and S6 proteins (d) in U2OS cells under normal (non-starved) conditions, under starvation for amino acids for 3 h (−AA) or under starvation for amino acid for 2 h followed by re-addition of amino acids for 1 h (+AA).

Figure 5

Inhibition of PHDs blocks the response of mTORC1 to amino acids. (a, b) Phosphorylation state of S6K, S6 and 4EBP1 in cells starved for 2 h and then re-fed with amino acids for one more hour in the presence or absence of DMOG, rapamycin (Rap) or cobalt chloride (CC). (c) U2OS cells were transfected with a plasmid encoding control, nontargeting shRNA (shCntr) or cotransfected with three different plasmids encoding shRNAs for each member of the PHD family (shPHDs). After 48 h, cells were starved for amino acids for 2 h and then incubated for 20 min with (+ AA) or without (− AA) amino acids. The phosphorylation state of S6 was analyzed by western blot. (d) Cells were transfected as in (c) and starved for amino acids for 2 h and then incubated for the indicated time with amino acids (+ AA). The phosphorylation state of S6K was analyzed by western blot.

Figure 6

PHDs signal to mTORC1 in a HIF/TSC-independent manner. (a) U2OS cells were transfected with a plasmid encoding control, nontargeting shRNA (shCntr) or transfected with a plasmid encoding shRNA for HIF1b (clone no. 4; shHIF1b no. 4). After 48 h, cells were starved for amino acids for 2 h and then incubated for 20 min with (+ AA) or without (− AA) amino acids either in the presence or absence of DMOG. Accumulation of REDD1 and phosphorylation state of S6K and 4EBP1 were analyzed by western blot. (b) Phosphorylation status of S6K in TSC2^+/+ and TSC2^−/− mouse embryonic fibroblasts starved for 2 h and then re-fed with amino acids for one more hour in the presence or absence of DMOG. (c) U2OS cells were transfected with either empty vector or Flag-Rheb-GTP, as indicated. After 48 h, cells were amino-acid starved for 2 h and re-stimulated for 1 h with amino acids (AA) either in the presence or absence of DMOG, as indicated. Phosphorylation of S6K and S6 was then measured by immunoblotting. (d) U2OS cells were transfected with either empty vector or with HA-RagB-GTP plus HA-RagC-GDP, as indicated. After 48 h, cells were amino-acid starved for 1 h where indicated, DMOG was added 30 min before the collection of the cells. Phosphorylation of S6K and S6 was then measured by immunoblotting.

Figure 7

Model for oxygen- and nutrient-sensing mechanisms by PHDs. Left side: PHDs mediate the inhibition of HIF in the presence of oxygen. Under hypoxia, lack of oxygen inhibits PHD activity, resulting in the stabilization of HIF, which inhibits mTORC1 via REDD1 induction. Right side: PHDs are required for the activation of mTORC1 in the presence of amino acids. Under amino-acid starvation, lack of αKG inhibits PHD activity, resulting in the inhibition of mTORC1, hence a reduction in HIF expression. Loss of PHD activity can also mediate the induction of autophagy under amino-acid starvation conditions in a HIF-independent manner.