These authors contributed equally to this work.
A possible linkage between AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling pathway
Article first published online: 8 JAN 2003
Genes to Cells
Volume 8, Issue 1, pages 65–79, January 2003
How to Cite
Kimura, N., Tokunaga, C., Dalal, S., Richardson, C., Yoshino, K.-i., Hara, K., Kemp, B. E., Witters, L. A., Mimura, O. and Yonezawa, K. (2003), A possible linkage between AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling pathway. Genes to Cells, 8: 65–79. doi: 10.1046/j.1365-2443.2003.00615.x
Communicated by: Kozo Kaibuchi
- Issue published online: 8 JAN 2003
- Article first published online: 8 JAN 2003
- Received: 6 September 2002 Accepted: 31 October 2002
Background: The mammalian target of rapamycin (mTOR) regulates multiple cellular functions including translation in response to nutrients, especially amino acids. AMP-activated protein kinase (AMPK) modulates metabolism in response to energy demand by responding to changes in AMP.
Results: The treatment of SV40-immortalized human corneal epithelial cells (HCE-T cells) with 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR), widely used as an AMPK activator, inhibits p70 S6k activities. Altered glucose availability, which regulates AMPK activity, also modulates the activity of p70 S6k. AICAR treatment also inhibits phosphorylation of Thr-412 in the p70 S6 kinase (p70 S6k), which is indispensable for the activity. Furthermore, over-expression of mutant AMPK subunits by stable expression in rabbit pulmonary fibroblast cell lines (PS120 cells) also modulates p70 S6k activity. The insensitivity of the rapamycin-resistant p70 S6k variant to AICAR treatment suggests that the inhibition of p70 S6k is mediated through a common effector, supporting a model whereby mTOR and its downstream effector are controlled by AMPK.
Conclusion: These results indicate that the AMPK and mTOR signalling pathways are possibly linked. In addition to the mTOR signal acting as a priming switch that modulates p70 S6k activation, AMPK appears to provide an overriding switch linking p70 S6k regulation to cellular energy metabolism.
The target of rapamycin proteins, TOR1 and TOR2, were first identified in Saccharomyces cerevisiae through mutants that confer resistance to growth inhibition induced by the immunosuppressive macrolide rapamycin (Heitman et al. 1991; Abraham & Wiederrecht 1996). Rapamycin binds to its intracellular receptor, the FK506-binding protein (FKBP12), and the FKBP12–rapamycin complex inhibits TOR function through direct binding to TOR in vivo (Stan et al. 1994; Zheng et al. 1995; Hall 1996).
Early findings in S. cerevisiae have indicated that TOR1 and TOR2 participate in the control of translational initiation and early G1 progression in response to nitrogen availability (Di Como & Arndt 1996; Schmelzle & Hall 2000). Additional functions of TOR include the regulation of transcription, amino acid uptake, cytoskeletal organization and protein degradation through autophagy (Schmelzle & Hall 2000; Schmidt et al. 1998). TOR orthologues have been identified in mammals (mTOR, also known as FRAP, RAFT1 or RAPT) (Avruch et al. 2001). The C-terminus of mTOR kinase is most closely related to the phosphatidylinositol kinase (PIK) subfamily, which plays important roles in controlling the cell cycle that includes ATM, ATR and DNA-PK (Keith & Schreiber 1995).
In cultured mammalian cells, rapamycin blocks the phosphorylation of several translational regulators, including eukaryotic initiation factor-4E binding protein 1 (4E-BP1) (Lawrence & Abraham 1997; Beretta et al. 1996) and p70 S6 kinase (p70 S6k) (Chung et al. 1992; Price et al. 1992). Furthermore, mTOR is an upstream regulator of these translational effectors (Lawrence & Abraham 1997; Hara et al. 1997; Brown et al. 1995).
The activity of p70 S6k dramatically increases through multisite phosphorylation in response to insulin or mitogens in vivo (Avruch et al. 2001). p70 S6k is thought to play an important role in the translation of a subclass of mRNAs containing a short oligopyrimidine sequence (Jefferies et al. 1997), although this has been recently challenged (Tang et al. 2001). In addition to regulation by insulin or mitogens, nutrients, especially amino acids, have been discovered to regulate the phosphorylation of p70 S6k and 4E-BP1 and to be necessary for the insulin or mitogen regulation of these targets by mTOR (Hara et al. 1998; Wang et al. 1998; Patti et al. 1998; Fox et al. 1998; Shigemitsu et al. 1999; Xu et al. 1998). Several studies of Drosophila melanogaster TOR have demonstrated that TOR proteins play a critical role in controlling cell growth by regulating several translational effectors in response to the nutritional environment (Oldham et al. 2000; Zhang et al. 2000; Montagne et al. 1999).
As well as nitrogen, carbon is the most basic nutrient source for cellular organisms. Carbon is used to produce energy and synthesize a wide range of biomolecules. In the yeast S. cerevisiae, in addition to the role of carbon in energy metabolism and metabolic biosynthesis, the presence of carbon, such as that in glucose, represses the expression of genes involved in the transport and utilization of other carbon sources. This is commonly known as glucose repression (Carlson 1999). Gene de-repression is mediated by activation of the Snf1 protein kinase. When glucose is limiting, Snf1 is activated and the transcription of many glucose-repressed genes is switched on (Hardie et al. 1998).
AMP-activated protein kinase (AMPK) is the mammalian orthologue of Snf1 (Hardie et al. 1998; Kemp et al. 1999) and exists as a heterotrimetric complex comprising a catalytic α subunit and non-catalytic β and γ subunits (Hardie et al. 1998; Kemp et al. 1999). Under conditions of hypoxia, exercise, ischaemia, heat shock and low glucose, AMPK is activated allosterically by rising cellular AMP and by phosphorylation of the catalytic α subunit by one or more AMPK kinases on Thr-172, located in the ‘T-loop’ of the catalytic domain (Hardie et al. 1998; Kemp et al. 1999). Once activated, AMPK phosphorylates multiple downstream substrates aimed at conserving existing ATP levels. AMPK reduces further ATP expenditure by inhibiting key enzymes in biosynthetic pathways such as acetyl-CoA carboxylace in fatty acid synthesis and 3-hydroxy-3-methyl-CoA reductase in cholesterol synthesis (Hardie et al. 1998). AMPK also increases the supply of ATP, for example by stimulating the rate of fatty acid oxidation and cellular glucose uptake (Hardie et al. 1998).
In the present study, in order to identify novel mTOR-interacting proteins, yeast two-hybrid screening was carried out using the catalytically inactive mutant of mTOR kinase domain as a bait, revealing the AMPK γ1 subunit as a putative mTOR-interacting protein. This led us to examine the possible relationship between the AMPK and mTOR/amino acid signalling pathways. The results demonstrate that modulation of AMPK activities using various reagents, including 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR), d-glucose and its analogues, and mutant AMPK subunits, affects the activity and phosphorylation state of p70 S6k, indicating a convergence of AMPK signalling and mTOR/amino acid signalling pathways on to p70 S6k.
Inhibition of endogenous p70 S6k activity by AICAR in different cell lines
To investigate mTOR-interacting proteins, yeast two-hybrid screening was carried out using mTOR kinase domain as a bait. We isolated several independent cDNA clones encoding the carboxyl terminal of the γ1 subunit of mammalian AMPK (amino acids 256–331). A portion of the γ1 subunit, which was isolated by yeast two-hybrid screening, was found to specifically bind to both the wild-type and the catalytically inactive mutant of mTOR kinase domain in HEK293 cells. However, we could not observe any interactions between the full length of mTOR and the γ1 subunit. Nevertheless, it was of interest to test for a possible relationship between the AMPK and mTOR/amino acid signalling pathways. In order to investigate whether AMPK is involved in the mTOR/amino acid signalling pathway, the effects of AICAR, an activator of AMPK, on p70 S6k activities were examined in several cell lines including SV40-immortalized human corneal epithelial cells (HCE-T cells), H-4-II-E rat hepatoma cells (H4IIE cells), Chinese Hamster Ovary cells over-expressing human insulin receptors (CHO-IR cells), and HEK293 cells (Fig. 1). Cells were deprived of serum for 16 h, followed by incubation with a fresh serum-free medium for 1 h and then treatment with 1 mm AICAR for 1 h. In HCE-T cells and H4IIE cells, treatment with AICAR caused an 80% inhibition of p70 S6k activities compared with the control (Fig. 1, top two panels, compare lanes 1 with 2), while in CHO-IR cells and HEK293 cells, no obvious effect of AICAR was observed (Fig. 1, bottom two panels, compare lanes 1 with 2). We investigated whether the p70 S6k activation induced by insulin was also inhibited by AICAR. Insulin-induced p70 S6k activation was partially inhibited in HCE-T cells and H4IIE cells (Fig. 1, top two panels, compare lanes 3 with 4) but not in CHO-IR cells and HEK293 cells (Fig. 1, bottom two panels, compare lanes 3 with 4). As expected, treatment of cells with rapamycin completely abolished p70 S6k activity in all cell lines (Fig. 1, lane 5). These results indicated that AMPK may be involved in p70 S6k regulation, at least in HCE-T cells and H4IIE cells. We therefore employed HCE-T cells for further studies to investigate the involvement of AMPK in mTOR/amino acid signalling.
Inhibition of p70 S6k activity by AICAR in HCE-T cells
HCE-T cells were deprived of serum for 16 h, followed by incubation with a fresh serum-free medium for 1 h and then treatment with the indicated concentration of AICAR for 1 h (Fig. 2A,B) or with 1 mm AICAR for the indicated time periods (Fig. 2C,D). The treatment with AICAR inhibited p70 S6k activities in a dose- and time-dependent manner. Approximate 70% and 80% reductions of the p70 S6k activities were observed following treatment with 2 mm AICAR for 1 h (Fig. 2A) or with 1 mm AICAR for 2 h (Fig. 2C), respectively. As expected (Fig. 2B,D), AMPK was activated by AICAR in a dose- and time-dependent manner. It should be noted that we report AMPK activities in immunoprecipitates of AMPKα1-containing heterotrimers; no AMPKα2 containing heterotrimers could be detected in these cells (data not shown). Three- to six-fold activation of AMPK was observed by the treatment with 2 mm AICAR for 1 h (Fig. 2B) or with 1 mm AICAR for 2 h (Fig. 2D), respectively. In H4IIE cells, the same inhibitory effects of AICAR on p70 S6k activities were observed (data not shown).
Modulation of p70 S6k activity by glucose availability in HCE-T cells
In rat pancreatic islet cells, glucose withdrawal leads to a large increase in cellular AMP/ATP ratio, thereby activating AMPK. Re-addition of glucose causes a reduction of AMPK activity (Salt et al. 1998; Hamilton et al. 2002). Thus, to further explore the involvement of AMPK in mTOR/amino acid signalling, we investigated the effects of glucose availability on p70 S6k activities in HCE-T cells. HCE-T cells were deprived of serum for 16 h, followed by incubation with a fresh serum-free medium for 1 h and then transferred to glucose-free medium. Glucose withdrawal inhibited p70 S6k activities for up to 30 min in time-dependent manner (Fig. 3A). As expected, AMPK was rapidly activated by glucose withdrawal and the activation was sustained for up to 1 h (Fig. 3B). Moreover, re-addition of 5.5 mm glucose to the glucose-free medium rapidly restored p70 S6k activity and simultaneously inhibited AMPK within 5 min (Fig. 3C,D). Taken together, the effects of AICAR and of glucose availability on p70 S6k activities are consistent with the interpretation that, either directly or indirectly, AMPK regulates p70 S6k activities. There is a slight temporal discrepancy between effects of glucose withdrawal and AICAR after 30 min (Fig. 3A,B). Although glucose withdrawal caused sustained activation of AMPK, p70 S6k activities slightly increased back toward initial values after 30 min.
Phosphorylation and activation of p70 S6k, phosphorylation of protein kinase B (PKB), and activation of AMPK under various nutritional conditions in the presence or absence of insulin and AICAR
p70 S6k activities in HCE-T cells were next examined under various nutritional conditions in the presence or absence of insulin and/or AICAR. HCE-T cells were deprived of serum for 16 h, followed by incubation with a fresh serum-free medium for 1 h, and then incubated with a fresh serum-free medium containing glucose and/or amino acids or lacking both nutrients. The activity of p70 S6k in serum-deprived HCE-T cells was still observed in the presence of both amino acids and glucose (Fig. 4A, lane 1). In contrast, it was reduced in the presence of glucose alone (Fig. 4A, lane 4) or in the absence of both amino acids and glucose (Fig. 4A, lane 7). The same inactivation of the activity of p70 S6k was also observed in the presence of amino acids alone (data not shown). The treatment of HCE-T cells with 100 nm insulin for 10 min (in the presence of amino acids and glucose) caused a four-fold activation of p70 S6k (Fig. 4A, compare lanes 1 with 2). Insulin treatment of HCE-T cells in the presence of glucose alone also led to p70 S6k activation (Fig. 4A, lane 5). The ability of insulin to stimulate p70 S6k activity was retained in the presence of amino acids alone (data not shown). However, the ability of insulin to stimulate p70 S6k activity was completely abolished in the absence of amino acids and glucose (Fig. 4A, lane 8).
The pre-treatment of HCE-T cells with 1 mm AICAR for 1 h caused an approximate 70% inhibition of insulin-induced activation of p70 S6k in the presence of amino acids and glucose (Fig. 4A, compare lanes 2 with 3) or in the presence of glucose alone (Fig. 4A, compare lanes 5 with 6). Among these culture conditions, both the addition of AICAR and the withdrawal of glucose caused the activation of AMPK (Fig. 4C, lanes 3, 6, 7 and 8). These findings are also consistent with the idea that the activation of AMPK negatively regulates p70 S6k activity.
p70 S6k is activated by insulin and mitogens in vivo through coordinate phosphorylation at multiple sites. Among multiple phosphorylation sites, phosphorylation of Thr-412 in the kinase extension domain of p70 S6kα is indispensable for p70 S6k activity (Dennis et al. 1996). Thus, in order to confirm the activation state of p70 S6k, immunoblotting of p70 S6k using anti-phosphothreonine 412 of p70 S6k antibodies was carried out (Fig. 4B, top panel). In agreement with the results shown in Fig. 4A, the immunoreactivity with the anti-p70 S6k (pT412)-specific antibody was decreased during AICAR treatment (Fig. 4B, top panel, compare lanes 2 with 3, or lanes 5 with 6), amino acid withdrawal (Fig. 4B, top panel, compare lanes 4 with 1) or withdrawal of amino acids and glucose in the absence or presence of insulin (Fig. 4B, top panel, compare lanes 7, 8 with lane 1). In contrast to these results, immunoreactivities of anti-phosphoserine 473 of protein kinase B (PKB) antibodies, the phosphorylation of which is critical for its activity (Alessi 2001), were increased after insulin stimulation (Fig. 4B, third panel, lanes 2, 5 and 8), even after the treatment with AICAR (Fig. 4B, third panel, lanes 3 and 6). These data indicate that effects of AICAR and withdrawal of nutrients target p70 S6k, but not PKB.
Resistance of rapamycin-resistant p70 S6k variant to inhibition caused by AICAR
The selective inhibition of p70 S6k that occurs consequent to AICAR treatment or glucose withdrawal corresponds closely to the cellular response seen following the addition of rapamycin. The similar selectivity of the cellular response to AICAR and glucose withdrawal, and rapamycin led us to examine the effects of AICAR on the activity of a rapamycin-resistant p70 S6k variant, p70Δ2-46/ΔCT104 (Weng et al. 1995). This variant was activated by insulin (Fig. 5A, compare lanes 5 with 6) as was the wild-type p70 S6k (Fig. 5A, compare lanes 1 with 2). Like the wild-type p70 S6k (Fig. 5A, compare lanes 2 with 4), this p70 S6k variant was markedly inhibited by treatment of cells by wortmannin (Fig. 5A, compare lanes 6 with 8), but was quite resistant to inhibition by rapamycin (Fig. 5A, compares lanes 2 with 3, lanes 6 with 7). Treatment with 1 mm AICAR did not inhibit, but activated the activity of p70Δ2-46/ΔCT104 (Fig. 5B, right panel) in marked contrast to the significant inhibition of recombinant wild-type p70 S6k engendered by 1 mm AICAR treatment (Fig. 5B, left panel). The unexpected activation of p70Δ2-46/ΔCT104 by AICAR treatment is reproducible. The mechanism of this phenomenon is, however, unknown. The resistance of p70Δ2-46/ΔCT104 to inhibition caused by both rapamycin and AICAR suggests that mTOR is required in order to mediate the inhibition of p70 S6k by these agents.
Effect of over-expression of mutant AMPK subunits, which alters AMPK activity, on p70 S6k activities
AICAR has been widely used as an activator of AMPK but is not necessarily specific (Kemp et al. 1999). Accordingly, we examined the effects of over-expression of mutant AMPK subunits, that alters AMPK activity, on p70 S6k activities. First we employed the constitutively active truncated form AMPKα1 (1–312) that does not require β and γ subunits for activity (Crute et al. 1998). As shown in Fig. 6, over-expression of GST-AMPKα1 (1–312) fragment in HCE-T cells resulted in 60% inhibition of recombinant p70 S6k activity compared with control (Fig. 6, top panel, compare lanes 2 with 1).
We next undertook to investigate the regulation of p70 S6k in cell lines stably over-expressing mutant AMPK subunits which either activate or inhibit endogenous AMPK activity. Mutation of the γ1 subunit (R70Qγ1) causes a four-fold increase in AMPK activity when stably expressed in PS120 cells (Hamilton et al. 2001). Stable over-expression of a mutant, inactive α1 subunit (K45Rα1) caused a marked inhibition of AMPK activity to < 2% of control values (data not shown) (Dyck et al. 1996). As shown in Fig. 7A, immunoblotting of cell extracts from two cell lines over-expressing either HA-tagged R70Qγ1 or GST-tagged K45Rα1 confirmed these alterations in AMPK activity under basal conditions and after stimulation with AICAR. As shown in Fig. 7A, AICAR slowed the migration of AMPKα in the control line (NHE) (compare lanes 1 with 2) due, in part, to altered an subunit phosphorylation on Thr-172 (Fig. 7C, compare lanes 1 with 2). In the HA-R70Qγ1-over-expressing cell line (Fig. 7B, lanes 3 and 4), AMPK mobility and phosphorylation were altered even in the absence of AICAR stimulation, and changes were exaggerated after AICAR treatment (Fig. 7A,C, lanes 3 and 4). In contrast, in the GST-K45Rα1 over-expressing cell line, expression of the endogenous AMPKα was markedly reduced (Fig. 7A, lanes 5 and 6). We have shown that this is due to an increased α turnover in the presence of this inactive subunit, which competes for binding of the endogenous β and γ subunits, increasing the degradation of native α (Crute et al. 1998). This phenomenon has also been noted by others in a transgenic mouse model with a dominant negative α construct (Mu et al. 2001). While the inactive GST-α fusion protein was phosphorylated on Thr-172 with a migration shift, active endogenous AMPK α, detected by migration shift and pThr-172, was markedly reduced (Fig. 7A,C, lanes 5 and 6), consistent with the marked decrease in measured enzyme activity (latter data not shown).
These cell lines were deprived of serum for 16 h, followed by incubation with a fresh serum-free medium for 1 h and then treatment with or without 1 mm AICAR for 1 h. In NHE cells, treatment with AICAR caused a 50% inhibition of p70 S6k activities (Fig. 7D, compare lanes 1 with 2). In HA-R70Qγ1-over-expressing cells, the basal activities of p70 S6k were reduced (Fig. 7D, compare lanes 1 with 3), and in the presence of AICAR this inhibition was strongly enhanced (Fig. 7D, compare lanes 3 with 4). These results are consistent with constitutive activation of AMPK by stable over-expression of R70Qγ1. In contrast, in the GST-K45Rα1-over-expressing cells, both the basal activities and activities in the presence of AICAR of p70 S6k were increased (Fig. 7D, lanes 5 and 6) in agreement with the dominant-negative effect of K45Rα on AMPK activities. These results strongly support a molecular link between AMPK and the mTOR/amino acid signalling pathways.
Effect of d-glucose withdrawal or substitution to d-glucose analogues on activities of AMPK and p70 S6k in HEK293 cells
As shown in Fig. 1 (bottom panel, compare lanes 2 with 1) and Fig. 8A (lower panel, compare lane 2 with 1), in HEK293 cells, treatment with AICAR failed to inhibit p70 S6k in contrast to its effects on HCE-T cells. The treatment with AICAR caused AMPK activation in HCE-T cells (Fig. 8B, upper panel, compare lanes 2 with 1) but not in HEK293 cells (Fig. 8B, lower panel, compare lanes 2 with 1). To gain some insight into a linkage between AMPK and p70 S6k in HEK293 cells, we explored whether d-glucose withdrawal or different d-glucose analogues could modulate the activities of AMPK and p70 S6k in HEK293 cells. The non-physiological stereoisomer l-glucose can be neither transported nor metabolized. In contrast, the d-glucose analogue 2-deoxyglucose (2-DG) can be transported, but cannot be metabolized fully, thus inhibiting glycolysis. Withdrawal of d-glucose (Fig. 8, lane 3) or substitution of d-glucose to l-glucose (Fig. 8, lane 4) was able to activate AMPK and inhibit p70 S6k in HCE-T cells, but not in HEK293 cells. In contrast, the substitution of d-glucose to 2-DG was able to induce activation of AMPK and inhibition of p70 S6k in both HCE-T cells and HEK293 cells (Fig. 8, lane 5).
The present study provides evidence that the AMPK and mTOR signalling pathways are possibly linked. First, the treatment of AICAR, widely used as an AMPK activator, inhibits p70 S6k activities in a dose- and time-dependent manner in HCE-T cells (Fig. 2). Second, the reagents or treatments that cause activation of AMPK, inhibit p70 S6k activities in many cell lines, even in HEK293 cells (Figs 1 and 8). Third, altered glucose availability leads to the change of AMPK activities that modulates p70 S6k activities (Fig. 3). Fourth, the treatment of AICAR also inhibit phosphorylation of Thr-412 in the kinase extension domain of p70 S6kα, which is indispensable for its kinase activities, but not Ser-473 of PKB (Fig. 4). Fifth, rapamycin-resistant p70 S6k variant is also insensitive to AICAR (Fig. 5). Sixth, over-expression of mutant AMPK subunits, which alter AMPK activities, modulates p70 S6k activities (Figs 6 and 7).
Recently, it was reported that AMPK suppresses protein synthesis in rat skeletal muscle through down-regulated PKB and mTOR signalling (Bolster et al. 2002). This study utilized AICAR to activate AMPK. Since AICAR may have multiple cellular effects other than the activation of AMPK (Kemp et al. 1999), more rigorous proof of a relationship is necessary to conclude that these effects are due to AMPK activity. In the present study, in addition to AICAR, d-glucose availability, d-glucose analogues, and mutant AMPK subunits were employed to activate or inhibit AMPK. The latter experiments support the idea that there is a possible molecular link between the mTOR and AMPK signalling pathways. Thus, while a truncated form of amino-terminal AMPKα1 or a mutant activating γ1 subunit (R70Qγ1), which causes a constitutive activation of AMPK activities, inhibits p70 S6k activities, an inactive α1 subunit mutant (K45Rα1), which induces a marked inhibition of AMPK activities, increases p70 S6k activities and blocks the inhibiting effects of AICAR (Figs 6 and 7).
In the present study, we observed that HCE-T and H4IIE cells are sensitive to AICAR to inhibit p70 S6k but not CHO-IR cells and HEK293 cells (Fig. 1). In CHO-IR cells, p70 S6k activity is sensitive to amino acid withdrawal and addition but not glucose (Hara et al. 1998). Prior studies in our laboratory have shown that AICAR is unable to activate AMPK in HEK293 cells (L.A.W., unpublished observation). In HCE-T cells, the activation of AMPK and subsequent p70 inhibition is caused by d-glucose withdrawal and treatment of d-glucose analogues, as well as AICAR treatment. In contrast, in HEK293 cells, only 2-DG is able to induce activation of AMPK and subsequent inhibition of p70 S6k (Fig. 8). The glycolytic inhibitor 2-DG is likely to be effective in reducing ATP concentrations, raising the AMP/ATP ratio, and subsequently activating AMPK in HEK293 cells, because those cells predominantly use anaerobic respiration to produce ATP (Dennis et al. 2001). Thus it is reasonable to suppose that the activation of AMPK leads to inhibition of p70 S6k even in HEK293 cells, suggesting that modulation of the mTOR signalling by AMPK signalling will be a generalized phenomena in cells, although the change of AMP/ATP ratio induced by various stimuli may vary among cell lines.
Recently, it has been reported that p70 S6k is inhibited by treatment with the mitochondrial inhibitors, azide, rotenone and carbonylcyanide m-chlorophenlyhydrazone (Dennis et al. 2001; Desai et al. 2002; Xu et al. 2001). These results are similar to our present observations of inhibition of p70 S6k activities induced by AICAR, glucose withdrawal, d-glucose analogues and over-expression of mutant AMPK subunits, indicating that mitochondrial dysfunction caused by the mitochondrial inhibitors may inhibit p70 S6k by activating AMPK. In hepatocytes, ATP depletion resulting from incubation with fructose or glycerol, in addition to AICAR treatment, likewise has been reported to inhibit p70 S6k (Dubbelhuis & Meijer 2002). Importantly, we have observed that rapamycin-resistant p70 S6k variant was insensitive to AICAR treatment. Previously the rapamycin-resistant p70 S6k variant was found to be insensitive to the depletion of ATP levels caused by either 2-DG or mitochondrial inhibitors (Dennis et al. 2001; Desai et al. 2002). Therefore it seems reasonable to conclude that AMPK acts as a common effector for the actions of AICAR, 2-DG, mitochondrial inhibitors and other ATP depletors by inhibiting the mTOR-mediated activation of p70 S6k.
Dennis et al. (2001) have proposed that mTOR could function as an ATP sensor in eukaryotic cells through a direct regulation of its activity by varying the concentrations of ATP. However, it would seem that such a mechanism would be likely to only operate at the extremes of cellular ATP depletion, in that they report an ATP Km of approximately 1 mm for mTOR. The relationship between the mTOR signalling system and that of AMPK, irrespective of the molecular mechanism involved, would make mTOR sensitive to even minimal ATP depletion. Even small decrements in ATP can result in elevations of cellular AMP through the actions of adenylate kinase, leading to AMPK activation and p70 S6k inhibition.
In contrast to Bolster et al. we have not been able to confirm an effect of AICAR in reducing the phosphorylation of mTOR on Ser-2448 (data not shown). In addition, the lack of inhibition of PKB by AICAR (Fig. 4) clearly indicates that AMPK does not signal through PKB to modulate mTOR and its downstream effector. The mechanism underlying the AMPK mediated inhibition of mTOR signalling is not known. To date, we could not observe the effect of AICAR on mTOR intrinsic kinase activity toward 4E-BP1 in vitro (data not shown). In addition, we could not observe the interaction between the full-length of mTOR and the γ1 subunit of AMPK, although the C-terminal portion of both proteins were found to bind each other in HEK293 cells. This could be due to the differences in three-dimensional structures between the full-length of mTOR and the C-terminal portion of mTOR. Recently, raptor, a binding partner of mTOR which mediates mTOR action, was identified (Hara et al. 2002; Kim et al. 2002; Loewith et al. 2002). AMPK also exists as a complex of α, β and γ subunits (Kemp et al. 1999). Thus, if the ability of the mTOR–raptor complex to bind to the complete AMPK complex were examined, the discrepancies in the binding experiments might be clarified.
In conclusion, mTOR may not directly sense the intracellular level of ATP, but the AMPK sensing the intracellular level of AMP/ATP ratio could signal to p70 S6k. While the amino acid and mTOR signals may act as a priming switch to modulate p70 S6k activation, AMPK signal could play a role as another overriding switch for p70 S6k regulation in the distinct pathway from PKB. This idea is consistent with a recent study in S. cerevisiae which showed that Snf1-glucose signalling and TOR-nitrogen signalling pathways converge on to the GATA-type transcription factor Gln3 (Bertram et al. 2002).
Antibodies and cDNA
The anti-HA antibody was purchased from Roche Molecular Biochemicals and the monoclonal ant-FLAG antibody (M2) was purchased from Sigma. The anti-GST antibody was from Upstate Biotechnology and the anti-p70 S6k antibody for Western blotting was purchased from Santa Cruz Biotechnology. The anti-PKB antibody was purchased from Transducion Laboratories. The anti-phosphothreonine 412 of p70 S6k antibodies and the anti-phosphoserine 473 of PKB antibodies were from Cell Signalling Technology Inc. The anti-AMPK α1 antibody for immunoprecipitation was prepared using a synthetic peptide corresponding to the amino acid sequences for residues 339–358 (DFYLATSPPDSFLDDHHLTR) from AMPK α1. The anti-AMPK α1 antibody for Western blotting and anti-phosphopeptides antibodies against Thr-172 of AMPK α1 were purchased from Cell Signalling Technology Inc. The anti-p70 S6k antibody for immunoprecipitation was as previously described (Hara et al. 1997).
The expression vectors of FLAG-tagged wild-type p70 S6kα (FLAG-p70 S6k, wild) and the rapamycin-resistant mutant of p70 S6kα (Δ2–46/ΔCT104), and GST-tagged constitutive active form of rat AMPK α1 subunits (GST-AMPK α1(1–312)) were as previously described (Crute et al. 1998; Weng et al. 1995). cDNA fragments encompassing nucleotides from 6441 to 7650 were amplified by PCR using wild-type mTOR (mTOR-WT) or kinase-negative mutant (mTOR-NK) as a template to create fragments encoding the wild-type (ΔmTOR-WT) or kinase-negative (ΔmTOR-NK) kinase domain of mTOR, respectively. Both cDNA fragments were ligated into the pCMV-FLAG vector to produce FLAG-tagged ΔmTOR-WT (FLAG-ΔmTOR-WT) or ΔmTOR-NK (FLAG-ΔmTOR-NK).
Yeast two-hybrid screen
The yeast two-hybrid screening was carried out using a fusion protein of the inactive mutant mTOR kinase domain (ΔmTOR-NK) and the DNA-binding domain of prokaryotic LexA protein as a bait (Clontech Laboratories Inc., Palo Alto, CA) (Gyuris et al. 1993). The yeast strain EGY48 expressing the bait was transformed with a human brain cDNA library (Clontech) to express fusion proteins of the 88-residue acidic E. coli peptide (B42) that activates transcription in yeast. Approximately 118 primary transformants were pooled and selected for growth in the absence of Leu and for β-galactosidase activity on X-gal plates. The sequence of cDNA fragments of positive clones were determined on both strands by using a Big Dye terminator cycle sequencing kit and a DNA sequencer, model 377 (Applied Biosystems).
Cell culture and stable cell lines
HCE-T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Sigma) with Nutrient Mixture F-12 (Sigma) supplemented with 15% foetal bovine serum (Life Technologies, Inc.), 5 µg/mL bovine insulin (Roche Diagnostics Corp), 0.1 mg/mL cholera toxin (Calbiochem), and 10 ng/mL human epidermal growth factor (Upstate Biotechnology) at 37 °C in 5% CO2. H4IIE cells were cultured in a modified Eagle's medium (αMEM) (Life Technologies Inc.) supplemented with 10% foetal bovine serum at 37 °C in 5% CO2. HEK293 cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) (Sigma) supplemented with 10% foetal bovine serum at 37 °C in 5% CO2. CHO-IR cells were cultured in Nutrient Mixture F-12 (Sigma) supplemented with 10% foetal bovine serum at 37 °C in 5% CO2. PS120 cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) (Sigma) supplemented with 10% foetal bovine serum at 37 °C in 5% CO2. PS120 cells were stably transfected with mutant AMPK subunits, either R70Qγ (AMPK activating) or GST-K45Rα1 (a dominant negative inhibitor), by a previously published method (Hamilton et al. 2001). Cell incubation, measurement of AMPK activity and immunoblotting of AMPK subunits was performed as indicated (Hamilton et al. 2001). 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR) were purchased from Sigma.
Complete Minimum Eagle's medium salt solution [referred to as MEM AA(+)/G(+)] contained 5.5 mm d-glucose and amino acids as follows (in mg/L): Arg, 126.4; Cys, 24.02; Glu, 292; Ile, 52.46; Leu, 52.46; Lys, 73.06; Met, 14.92; Phe, 33.02; Thr, 47.64; Trp, 10.2; Tyr, 36.22; Val, 46.86; Ala, 8.9; Asn, 13.2; Asp, 13.3; Gln, 14.7; Gly, 7.5; Pro, 11.5; and Ser, 10.5. In MEM AA(–)/G(+), all the amino acids were deprived from MEM AA(+)/G(+). In MEM AA(+)/G(–), d-glucose was deprived from MEM AA(+)/G (+). In MEM AA(–)/G(–), both total all the amino acids and d-glucose were deprived form MEM AA(+)/G(+).
Transient transfection was performed by the lipofection method using LipofectAMINE (Life Technologies Inc.) or SuperFect (Qiagen) according to the manufacturer's protocol.
Immunoprecipitation and p70 S6 kinase assay
p70 S6k activity was determined in immunoprecipitates using 40 S ribosomal subunits as substrate. Cells were lysed in ice-cold buffer A (Hara et al. 1998), and the extracts were centrifuged at 10 000 g for 20 min. Aliquots of the supernatants were subjected to immunoprecipitation with the anti-p70 S6kα antibody. The immunoprecipitates were washed twice with buffer A containing 0.5 m NaCl and twice with 20 mm MOPS, pH 7.2, 10 mmβ-glycerophosphate. The p70 S6 kinase assay was performed in the reaction mixture [50 mm MOPS, pH 7.2, 12 mm MgCl2, 2 mm EGTA, 1 mm DTT, 10 mmβ-glycerophosphate, 0.5 mm PKI, 0.5 A260 units of 40 S ribosomal subunits, and 10 mm ATP, 2 µCi of [γ-32P]ATP (Institute of Isotopes Co. Ltd, Hungary)]. The reaction was continued for 30 min at 30 °C. The p70 S6 kinase assay was terminated by adding SDS sample buffer followed by SDS-PAGE and autoradiography. The radioactivity was quantified with a BAS-2500 Bioimaging analyser (Fuji). Immunoblotting was performed using the anti-p70 S6k antibody as a primary antibody and Protein A-HRP (Bio-Rad) as a second probe, and visualized by the ECL method.
Immunoprecipitation and AMPK Assay
AMPK activity was determined in immunoprecipitates by using a synthetic peptide, HMRSAMSGLHLVKRR (referred to as SAMS peptide) as substrate (Hardie et al. 1998). The immunoprecipitates using the AMPK α1 antibodies were washed twice with lysis buffer containing 100 mm NaCl, 50 mm Tris/HCl, pH 7.4, 50 mm NaF, 5 mm NaPPi, 1% Na-deoxycholate, 1% Triton X-100, 1 mm EDTA, 1 mm EGTA, 20 mmβ-glycerophosphate, 1 mm DTT, 1 mm phenylmethylsulphonyl fluoride, 2 µg/mL aprotinin, 1 µm leupeptin. The immunoprecipitates were washed twice with lysis buffer, twice in 4× AMPK assay buffer containing 240 mm Na-HEPES, pH 7.0, 480 mm NaCl, 10 mmβ-glycerophosphate, 4 mm DTT. The AMPK kinase assay was performed in 1× AMPK assay buffer (5 mm MgCl2, 2 mm EGTA, 1 mm DTT, 10 mmβ-glycerophosphate, 0.2 mm AMP, 200 µm SAMS peptide, 20 µm ATP, 5 µCi of [γ-32P]ATP). The reaction was continued for 20 min at 30 °C. The SAMS peptide kinase assay was terminated by the addition of 20 mm EDTA and 1.0 mm ATP. The reaction mixtures were spotted on to P81 phosphocellulose paper, followed by washes in 75 mm phosphoric acid. 32P-labelled peptides were measured by Cerenkov counting.
We are grateful to Dr Y. Nishizuka for encouragement. We thank Dr J. Avruch for valuable advice and discussion, Dr S. Hidayat for helpful discussion, H. Miyamoto for technical assistance, and Dr K. Sasaki for providing HCE-T cells. The skilful secretarial assistance of R. Kato is cordially acknowledged. This work was supported in part by research grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to K.H. and K.Y.), Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists (to C.T.), the Charitable Trust Osaka Cancer Research-Fund (to K.-i.Y.), Suntory Institute for Bioorganic Research (to K.-i.Y.), and the National Institutes of Health (DK35712, to L.A.W.). B.E.K. is an NHMRC Fellow supported by the NHMRC, ARC and National Heart Foundation.
- 1996) Immunopharmacology of rapamycin. Annu. Rev. Immunol. 14, 483–510. & (
- 2001) Discovery of PDK1, one of the missing links in insulin signal transduction. Colworth Medal Lecture. Biochem. Soc. Trans. 29, 1–14. (
- 2001) The p70 S6 kinase integrates nutrient and growth signals to control translational capacity. In: Progress in Molecular and Subcellular Biology (ed. R.E.Rhoads), pp. 115–154. Berlin/Heiderberg: Springer-Verlag. , , , & (
- 1996) Rapamycin blocks the phosphorylation of 4E-BP1 and inhibits cap-dependent initiation of translation. EMBO J. 15, 658–664. , , , & (
- 2002) Convergence of TOR-nitrogen and Snf1-glucose signaling pathways onto Gln3. Mol. Cell. Biol. 22, 1246–1252. , , , , & (
- 2002) AMP-activated protein kinase suppresses protein synthesis in rat skeletal muscle through downregulated mTOR signaling. J. Biol. Chem. 277, 23977–23980. , , & (
- 1995) Control of p70 s6 kinase by kinase activity of FRAP in vivo. Nature 377, 441–446. , , , , & (
- 1999) Glucose repression in yeast. Curr. Opin. Microbiol. 2, 202–207. (
- 1992) Rapamycin-FKBP specifically blocks growth-dependent activation of and signaling by the 70 kd S6 protein kinases. Cell 69, 1227–1236. , , & (
- 1998) Functional domains of the alpha-1 catalytic subunit of the AMP-activated protein kinase. J. Biol. Chem. 273, 35347–35354. , , , & (
- 2001) Mammalian TOR: a homeostatic ATP sensor. Science 294, 1102–1105. , , , , & (
- 1996) The principal rapamycin-sensitive p70s6k phosphorylation sites, T-229 and T-389, are differentially regulated by rapamycin-insensitive kinase kinases. Mol. Cell. Biol. 16, 6242–6251. , , & (
- 2002) FKBP12-rapamycin associated protein associates with mitochondria and senses osmotic stress via mitochondrial dysfunction. Proc. Natl. Acad. Sci. USA 99, 4319–4324. , & (
- 1996) Nutrients, via the Tor proteins, stimulate the association of Tap42 with type2A phosphatases. Genes Dev. 10, 1904–1916. & (
- 2002) Hepatic amino acid-dependent signaling is under the control of AMP-dependent protein kinase. FEBS Lett. 521, 39–42. & (
- 1996) Regulation of 5′-AMP-activated protein kinase activity by the noncatalytic β and γ subunits. J. Biol. Chem. 271, 17798–17803. , , , et al. (
- 1998) Amino acids stimulate phosphorylation of p70S6k and organization of rat adipocytes into multicellular clusters. Am. J. Physiol. 274, C206–C213. , , & (
- 1993) Cdi1, a human G1 and S phase protein phosphatase that associates with Cdk2. Cell 75, 791–803. , , & (
- 1996) The TOR signaling pathway and growth control in yeast. Biochem. Soc. Trans. 24, 234–239. (
- 2002) AMP-activated protein kinase kinase: detection with recombinant AMPK α1 subunit. Biochem. Biophys. Res. Commun. 293, 892–898. , , , et al. (
- 2001) An activating mutation in the γ1 subunit of the AMP-activated protein kinase. FEBS Lett. 500, 163–168. , , , et al. (
- 2002) Raptor, a binding partner of target of rapamycin (TOR), mediates TOR action. Cell 110, 177–189. , , , et al. (
- 1997) Regulation of eIF-4E BP1 phosphorylation by mTOR. J. Biol. Chem. 272, 26457–26463. , , , et al. (
- 1998) Amino acid sufficiency and mTOR regulate p70 S6 kinase and eIF-4E BP1 through a common effector mechanism. J. Biol. Chem. 273, 14484–14494. , , , , & (
- 1998) The AMP-activated/SNF1 protein kinase subfamily: metabolic sensors of the eukaryotic cell? Annu. Rev. Biochem. 67, 821–855. , & (
- 1991) Targets for cell cycle arrest by the immunosuppressant rapamycin in yeast. Science 253, 905–909. , & (
- 1997) Rapamycin suppresses 5′TOP mRNA translation through inhibition of p70S6k. EMBO J. 16, 3693–3704. , , , , & (
- 1995) PIK-related kinases: DNA repair, recombination, and cell cycle checkpoints. Science 270, 50–51. & (
- 1999) Dealing with energy demand: the AMP-activated protein kinase. Trends. Biochem. Sci. 24, 22–25. , , , , & (
- 2002) mTOR interacts with raptor to form a nutrient-sensitive complex that signals to the cell growth machinery. Cell 110, 163–175. , , , et al. (
- 1997) PHAS/4E-BPs as regulators of mRNA translation and cell proliferation. Trends. Biochem. Sci. 22, 345–349. & (
- 2002) Two TOR complexes, only one of which is rapamycin sensitive, have distinct roles in cell growth control. Moll. Cell 10, 457–468. , , , et al. (
- 1999) Drosophila S6 kinase: a regulator of cell size. Science 285, 2126–2129. , , , , & (
- 2001) A role for AMP-activated protein kinase in contraction- and hypoxia-regulated transport in skeletal muscle. Mol. Cell 7, 1085–1094. , , , & (
- 2000) Genetic and biochemical characterization of dTOR, the Drosophila homolog of the target of rapamycin. Genes Dev. 14, 2689–2694. , , , & (
- 1998) Bidirectional modulation of insulin action by amino acids. J. Clin. Invest. 101, 1519–1529. , , , & (
- 1992) Rapamycin-induced inhibition of the 70-kilodalton S6 protein kinase. Science 257, 973–977. , , , & (
- 1998) AMP-activate protein kinase is activated by low glucose in cell lines derived from pancreatic β cells, and may regulate insulin release. Biochem. J. 335, 533–539. , , & (
- 2000) TOR, a central controller of cell growth. Cell 103, 253–262. & (
- 1998) The TOR nutrient signaling pathway phosphorylates NPR1 and inhibits turnover of the tryptophan permease. EMBO J. 17, 6924–6931. , , , & (
- 1999) Regulation of translational effectors by amino acid and mammalian target of rapamycin signaling pathways. Possible involvement of autophagy in cultured hepatoma cells. J. Biol. Chem. 274, 1058–1065. , , , , & (
- 1994) Interaction between FKBP12-rapamycin and TOR involves a conserved serine residue. J. Biol. Chem. 269, 32027–32030. , , , , & (
- 2001) Amino acid induced translation of TOP mRNAs is fully dependent on phosphatidyl inositol 3-kinase-mediated signaling, is partially inhibited by rapamycin acid is independent of S6K1 and rpS6 phosphorylation. Mol. Cell. Biol. 21, 8671–8683. , , , et al. (
- 1998) Amino acid availability regulates p70 S6 kinase and multiple translation factors. Biochem. J. 334, 261–267. , , & (
- 1995) Multiple independent inputs are required for activation of the p70 S6 kinase. Mol. Cell. Biol. 15, 2333–2340. , , , & (
- 2001) Metabolic regulation by leucine of translation initiation through the mTOR-signaling pathway by pancreatic beta-cells. Diabetes 50, 353–360. , , , & (
- 1998) Branched-chain amino acids are essential in the regulation of PHAS-I and p70 S6 kinase by pancreatic β-cells. A possible role in protein translation and mitogenic signaling. J. Biol. Chem. 273, 28178–28184. , , , , & (
- 2000) Regulation of cellular growth by the Drosophila target of rapamycin dTOR. Genes Dev. 14, 2712–2724. , , , & (
- 1995) TOR kinase domains are required for two distinct functions, only one of which is inhibited by rapamycin. Cell 82, 121–130. , , , & (