Phosphatidylinositol 3-Kinase/Akt Pathway Regulates Tuberous Sclerosis Tumor Suppressor Complex by Phosphorylation of Tuberin*

Normal cellular functions of hamartin and tuberin, encoded by the TSC1 and TSC2tumor suppressor genes, are closely related to their direct interactions. However, the regulation of the hamartin-tuberin complex in the context of the physiologic role as tumor suppressor genes has not been documented. Here we show that insulin or insulin growth factor (IGF) 1 stimulates phosphorylation of tuberin, which is inhibited by the phosphatidylinositol 3-kinase (PI3K) inhibitor LY294002 but not by the mitogen-activated protein kinase inhibitor PD98059. Expression of constitutively active PI3K or active Akt, including Akt1 and Akt2, induces tuberin phosphorylation. We further demonstrate that Akt/PKB associates with hamartin-tuberin complexes, promoting phosphorylation of tuberin and increased degradation of hamartin-tuberin complexes. The ability to form complexes, however, is not blocked. Akt also inhibits tuberin-mediated degradation of p27kip1, thereby promoting CDK2 activity and cellular proliferation. Our results indicate that tuberin is a direct physiological substrate of Akt and that phosphorylation of tuberin by PI3K/Akt is a major mechanism controlling hamartin-tuberin function.

Tuberous sclerosis complex (TSC) 1 is an autosomal dominant disorder and is characterized by the presence of hamartomas in many organs such as brain, skin, heart, lung, and kidney (1). It is caused by mutation of either the TSC1 or TSC2 tumor suppressor gene (2)(3)(4)(5). TSC1 encodes a protein, hamartin, containing two coiled-coil domains that have been shown to mediate binding to hamartin (6). The TSC2 gene codes for tuberin, which contains a small region of homology to the rap1GTPase-activating protein, rap1GAP (7). These two proteins function within the same pathway(s) regulating cell cycle, cell growth, adhesion, and vesicular trafficking (4,5). However, the regulation of hamartin and tuberin in the context of physiologic role as tumor suppressor genes has not been documented.
Among the various properties of these two proteins, the ability to interact and to form stable complex has been the most consistent finding. This led to the hypothesis that hamartin and tuberin function as a complex and that factors regulating their interaction are important in understanding physiologic roles. There is evidence to suggest that phosphorylation of tuberin may be a major mechanism of regulation of the hamartin-tuberin complex (8,9). However, the kinases that are responsible for phosphorylation of this complex are currently unknown. Recent Drosophila genetic studies showed that dTsc1 and dTsc2 play an important role in the insulin/dPI3K/ dakt signal transduction pathway by demonstrating that reduced cell size and cell proliferation caused by either mutations in dINR and dakt or by overexpression of dPTEN are overridden by homozygous mutants of dTsc1 or dTsc2. This implies that dTsc1 and dTsc2 are either direct downstream targets of dakt or on a parallel pathway of the insulin cascade downstream from dakt (10 -13). Akt, also known as protein kinase B (PKB), represents a subfamily of the serine/threonine protein kinase. Three isoforms of Akt have been identified including Akt1/PKB␣, Akt2/PKB␤, and Akt3/PKB␥, all of which are activated by growth factors and insulin in a PI3K-dependent manner and are inhibited by PTEN tumor suppressor (14). Akt regulates a wide spectrum of cell functions, including cell survival, cell growth, differentiation, angiogenesis, and glucose metabolism, through phosphorylation of a number of proteins that contain the RXRXXS/T motif (14 -16).
Here we show that Akt physically interacts with and phosphorylates tuberin, leading to degradation of the hamartintuberin complex and p27 kip1 without interfering with hamartin-tuberin complex formation. Moreover, IGF1 and insulin induce tuberin phosphorylation, which is mediated by the PI3K/Akt pathway but not by the MAPK pathway. As a result, cyclin-dependent kinase (CDK) 2 activity, DNA synthesis, and S phase of the cell cycle are elevated. We thus have identified Akt as a major tuberin kinase to negatively regulate hamartintuberin tumor suppressor function by inducing degradation.
Cell Culture, Transfection, and Flow Cytometry-Human embryonic kidney (HEK) 293 and HeLa cells were obtained from the American Type Culture Collection. EEF4 (TSC2-positive) and EEF8 (TSC2-negative) cells were derived from Eker rat embryos homozygous for the wild type and the Eker-mutant TSC2 gene, respectively (8). All cells were grown either in Dulbecco's modified Eagle's medium or in RPMI 1640 medium, both supplemented with 10% calf serum and antibiotics. Cell transfections were performed using LipofectAMINE Plus. For cytofluorometric analyses, cells were harvested by trypsinization, fixed, and analyzed on a FACScan.
Immunoprecipitation, Immunoblotting, and in Vitro Kinase Assay-For immunoprecipitation, lysates were incubated with the appropriate antibody (as noted in the figure legends) in the presence of protein A-protein G (2:1)-agarose beads. The beads were washed with lysis buffer. The immunoprecipitates were subjected to in vitro kinase assay or Western blotting analysis. Detection of antigen-bound antibody was carried out with the ECL System. Protein kinase assays were performed as described previously (18).
Pulse-chase Experiments-Prior to radioactive labeling normal culture medium was removed, and cells were washed twice with phosphate-buffered saline and refed with minimum Eagle's medium lacking methionine but supplemented with 10% dialyzed fetal bovine serum and 300 Ci of Tran 35 S-label per plate. After 60 min of labeling, cells were lysed and immunoprecipitated with anti-TSC1, -TSC2, or -p27 antibody. The immunoprecipitates were separated by SDS-PAGE gel. Gels were dried and autoradiographed. Quantification of bands was performed with a PhosphorImager.
In Vivo [ 32 P]orthophosphate Cell Labeling-COS7 cells were transfected with pcDNA3-TSC2 together with or without constitutively active Akt and labeled with [ 32 P]orthophosphate (0.5 mCi/ml) in minimum Eagle's medium without phosphate for 4 h. Tuberin was immunoprecipitated with anti-TSC2 antibody. The immunoprecipitates were separated on SDS-PAGE and transferred to membranes. Phosphorylated tuberin was detected by autoradiography.

RESULTS
Tuberin Is a Physiological Substrate of Akt-Recent studies demonstrated that tuberin is phosphorylated at serine and tyrosine residues in response to serum, phosphatase inhibitors, and anisomycin and that the phosphorylated tuberin regulates its interaction with hamartin (8,9). However, the kinases that are responsible for phosphorylation of tuberin are currently unknown. Because tuberin contains seven Akt phosphorylation consensus sites that are very conserved between human, rat, and mouse as well as four that are also found in Drosophila (Fig. 1a), we examined the possibility of Akt phosphorylation of tuberin. In vivo [ 32 P]orthophosphate cell-labeling experiments revealed that constitutively active Akt and IGF1-induced Akt significantly phosphorylate tuberin (Fig. 1b). To explore which sites on tuberin are potentially phosphorylated by Akt, in vitro kinase assay was carried out using wild type and mutant (converting S/T to alanine) GST fusion proteins for each of seven Akt putative phosphorylation sites as substrate. As shown in Fig. 1c, Akt can highly phosphorylate fusion proteins containing all seven serine and threonine sites of tuberin but not their mutants. We therefore conclude that tuberin is a physiological substrate of Akt.
The PI3K/Akt Pathway, but Not the MAPK Pathway, Mediates Insulin, IGF1, and Serum-induced Tuberin Phosphorylation-Because genetic studies of the dTsc complex in Drosophila have demonstrated that dTsc1/dTsc2 antagonize insulin signaling in cell growth (10 -13), we next examined whether insulin and IGF1 induce hamartin-tuberin phosphorylation and whether Akt mediates this action. Western blotting analyses showed that tuberin, but not hamartin, was phosphorylated upon insulin, IGF1, or serum stimulation in HeLa cells as demonstrated by gel mobility shift (Fig. 1d). The phosphorylation was abrogated by treatment with phosphatase PP2A or PI3K inhibitors, LY294002, and wortmannin, but not by MAPK inhibitor PD98059 (Fig. 1e). Furthermore, expression of constitutively active PI3K (p110*) and active Akt significantly induced the phosphorylation of tuberin (Fig. 1f). These data indicate that the function of TSC tumor suppressors is regulated by mitogenic growth factor IGF1 and insulin through the PI3K/ Akt, but not the MAPK, pathway.
Akt Interacts with Tuberin and Hamartin-To assess whether Akt physically associates with the hamartin-tuberin complex, coimmunoprecipitation experiments were performed in HeLa cells. As seen in Fig. 2a, Akt directly bound to tuberin. Because hamartin and tuberin function as a complex and Akt associates with tuberin, we assumed that Akt could indirectly interact with hamartin. To test this hypothesis, HEK293 cells were transfected with Myc-TSC1, HA-Akt, and/or TSC2-Xpress. Coimmunoprecipitation experiments revealed that interaction between Akt and hamartin was only detected in the cells transfected with Myc-TSC1/TSC2-Xpress/HA-Akt but not with Myc-TSC1/Akt (Fig. 2b), indicating that Akt binding to hamartin is mediated by tuberin. Because tuberin binds to Akt and is phosphorylated by Akt, we conclude that tuberin is a direct downstream target of Akt.
The Hamartin-Tuberin Complex Is Not Disrupted by Akt Phosphorylation of Tuberin-Because previous studies have suggested that phosphorylation of tuberin regulates its interaction with hamartin (8,9), we next examined whether Akt interferes with hamartin-tuberin complex formation. Coimmunoprecipitation revealed that expression of wild type and constitutively active Akt in HeLa cells did not disrupt the interaction between hamartin and tuberin (Fig. 2c), despite the fact that hamartin and tuberin function as a complex. Moreover, phosphomimic TSC2-7D-Xpress and nonphosphorylatable TSC2-7A-Xpress, prepared by converting seven Akt phospho-rylation sites of tuberin into aspartic acid and alanine, respectively, were transfected into HeLa cells. Immunoblotting analyses of TSC2-7D-Xpress and TSC2-7A-Xpress immunoprecipitates showed that both mutant forms of tuberin still bound to hamartin (Fig. 2d), indicating that Akt phosphorylation of tuberin did not hamper the interaction between hamartin and tuberin.
Akt Phosphorylation of Tuberin Induces Degradation of Hamartin and Tuberin-Strikingly, we observed that expression of constitutively active Akt significantly down-regulated hamartin and tuberin in a dose-dependent manner, i.e. protein levels of hamartin and tuberin progressively declined when the cells were transfected with increasing amounts of constitutively active Akt. Accordingly, the protein amount of hamartin and tuberin in the complex was also decreased (Fig. 3a). To exclude the possibility of Akt down-regulation of hamartin and tuberin resulting from inhibition of the TSC1 and TSC2 gene transcription, Northern blot analyses were performed and showed that mRNA levels of TSC1 and TSC2 did not change in HeLa cells transfected with constitutively active Akt as compared with the cells transfected with pcDNA3 vector alone (Fig.  3e). Because Akt has been shown to activate rather than inhibit translation initiation through regulation of FRAP/mTOR/ 4E-BP (14,15), it is unlikely that Akt regulates hamartin and tuberin at a translational level. Thus, we assumed Akt downregulation of hamartin and tuberin occurred through protein degradation. Pulse-chase experiments revealed that expression of constitutively active Akt considerably induced hamartin and tuberin degradation (Fig. 3b). Moreover, the proteosome inhibitor MG132 attenuated Akt-induced hamartin and tuberin degradation (Fig. 3c). These data suggest that Akt down-regulation of hamartin and tuberin is mediated by a post-transla- tional modification mechanism in which the proteosome pathway is involved.
To examine whether Akt-induced hamartin-tuberin degradation depends upon Akt phosphorylation of tuberin, EEF8 TSC2-deficient cells were transfected with wild type TSC2, phosphomimic TSC2-7D, or nonphosphorylatable TSC2-7A. Western blotting and pulse-chase analyses revealed that TSC2-7D was degraded more rapidly than wild type TSC2, whereas TSC2-7A became more stable. Expression of TSC2-7D promoted hamartin degradation, whereas TSC2-7A stabilized hamartin and inhibited Akt-induced hamartin degradation (Fig. 3d), indicating that Akt phosphorylation of tuberin is required for degradation of the hamartin-tuberin complex. Pre-vious studies have shown that tuberin functions as a cytosolic chaperone protein to prevent hamartin self-aggregation and maintain the tuberin-hamartin complex in a soluble form (9,19). However, we did not observe that Akt phosphorylation of tuberin affected its chaperone function (data not shown).
Akt Phosphorylation of Tuberin Leads to Down-regulation of p27 kip1 and Cell Proliferation-The results from studying altered expression of either TSC1 or TSC2 have demonstrated that both hamartin and tuberin inhibit cell growth and cell size in mammalian (4,5) and Drosophila (10 -12), respectively. The mechanism that has been characterized so far is that overexpression of hamartin and tuberin induces the expression of the cyclin-dependent kinase inhibitor p27 kip1 through inhibition of its degradation (20,21). To examine the effects of Akt phosphorylation of tuberin on p27 kip1 expression, pulse-chase analyses were performed with TSC2-deficient EEF8 cells that were transfected with TSC2, TSC2/Mry-Akt, TSC2-7D, or TSC2-7A. As shown in Fig. 4a, expression of constitutively active Akt abrogated the ability of stabilization of p27 kip1 by tuberin. P27 kip1 degraded rapidly in phosphomimic TSC2-7D-transfected cells as compared with the cells expressing wild type TSC2/constitutively active Akt. In contrast, the cells expressing TSC2-7A exhibited similar degradation rate of p27 kip1 to wild type TSC2-transfected cells. Moreover, expression of TSC2-7A abrogated constitutively active Akt-induced p27 kip1 degradation (Fig. 4a). These data indicate that degradation of p27 kip1 is regulated by Akt phosphorylation of tuberin.
Because CDK2 is a major regulator of cell growth and G 1 /S transition of the cell cycle, we further examined the effects of Akt phosphorylation of tuberin on cell proliferation measured by cell growth and thymidine incorporation. As shown in Fig. 4, c and d, expression of TSC2 or TSC2-7A in TSC2-deficient EEF8 cells inhibited cell growth and DNA synthesis as compared with the cells transfected with vector alone. However, cells expressing constitutively active Akt or phosphomimic TSC2-7D significantly enhanced cell growth and thymidine incorporation. Consistent with previous findings, the number of cells at the G 0 -G 1 phase of the cell cycle was increased in the cells expressing TSC2. Expression of TSC2-7A displayed a similar effect on cell cycle. In contrast, constitutively active Akt overrode wild type TSC2 action. The cell number of the S phase was increased in constitutively active Akt/TSC2-or phosphomimic TSC2-7D-transfected cells (Fig. 4e). These data indicate that Akt phosphorylated tuberin lost its tumor suppressor function at least in part by inducing p27 kip1 degradation. DISCUSSION Recent studies have demonstrated that phosphorylation of hamartin and/or tuberin may play an important role in the formation of the tuberin-hamartin complex. Tuberin is phosphorylated at serine and tyrosine residues, and a diseaserelated TSC2 tyrosine 1571 mutation (Y1571H) nearly abolishes tuberin tyrosine phosphorylation and disrupts tuberinhamartin binding, implying that the phosphorylation of tyrosine 1571 of TSC2 is required for tuberin-hamartin complex formation (8,9). Our study, however, shows that phosphorylation of tuberin by Akt and mitogenic factors (insulin and IGF1) abrogates hamartin-tuberin tumor suppressor activity without interfering with binding but by inducing degradation of both proteins through the proteosome pathway. Therefore, we provide a new paradigm for regulation of the TSC1/TSC2 tumor suppressor pathway.
In addition to the Forkhead transcription factor family (16,22), tuberin is the second Akt downstream target that has been uncovered by genetic studies so far. In this study, we present molecular evidence that tuberin is a direct physiological substrate of Akt by demonstrating that Akt binds to and phosphorylates tuberin. It has been documented that Akt induces cell cycle progression and cell proliferation through transcription repression and degradation of p27 kip1 (23,24). Akt inhibition of p27 kip1 transcription is achieved by Akt phosphorylation of a Forkhead transcription factor, AFX, leading to the decrease of p27 kip1 promoter activity (24). However, the mechanism of Akt degradation of p27 kip1 is unclear. Tuberin was revealed to stabilize p27 kip1 by maintaining p27 kip1 in the nucleus (20). We observed in this study that Akt attenuates the tuberin action but does not induce translocation of p27 kip1 from nuclear to cytoplasm (data not shown). Previous studies have shown that three isoforms of Akt share almost the same upstream regulators and downstream targets. Similarly, we have observed that Akt1, Akt2, and Akt3 all phosphorylate and interact with tuberin, even though Akt2 displays a slightly higher binding affinity to tuberin. The model in Fig. 5 illustrates the mechanism through which the PI3K/Akt pathway mediates insulin and IGF1 signals to down-regulate hamartin-tuberin function by phosphorylation of tuberin. Our results define a possible new mechanism through which Akt induces cell proliferation and transformation by inhibiting TSC1/TSC2 tumor suppressor functions.