Regulation of TSC2 by 14-3-3 Binding*

Mutation in either the TSC1 or TSC2 tumor suppressor gene is responsible for the inherited genetic disease of tuberous sclerosis complex. TSC1 and TSC2 form a physical and functional complex to regulate cell growth. Recently, it has been demonstrated that TSC1 (cid:1) TSC2 functions to inhibit ribosomal S6 kinase and negatively regulate cell size. TSC2 is negatively regulated by Akt phosphorylation. Here, we report that TSC2, but not TSC1, associates with 14-3-3 in vivo . Phosphorylation of Ser 1210 in TSC2 is required for its association with 14-3-3. Our data indicate that 14-3-3 association may inhibit the function of TSC2 and represents a possible mechanism of TSC2 regulation. Tuberous sclerosis complex (TSC) 1 is a relatively common genetic disorder. TSC is caused by mutation in either the TSC1 (hamartin) or TSC2 (tuberin) gene, of which each contributes to (cid:1) 50% of the genetic defects (1, 2). Studies of TSC patients and animal models support the hypothesis that TSC1 and TSC2 are tumor suppressor genes. Homozygous deletion of either TSC1 or TSC2 in mice produces an embryonic lethal phenotype, suggesting an essential function in development. As predicted, heterozygous deletion of either TSC1 or TSC2 in mice results in a significant increase of carcinomas in many tissues with a 100% incidence of renal

Mutation in either the TSC1 or TSC2 tumor suppressor gene is responsible for the inherited genetic disease of tuberous sclerosis complex. TSC1 and TSC2 form a physical and functional complex to regulate cell growth.
Recently, it has been demonstrated that TSC1⅐TSC2 functions to inhibit ribosomal S6 kinase and negatively regulate cell size. TSC2 is negatively regulated by Akt phosphorylation. Here, we report that TSC2, but not TSC1, associates with 14-3-3 in vivo. Phosphorylation of Ser 1210 in TSC2 is required for its association with 14-3-3. Our data indicate that 14-3-3 association may inhibit the function of TSC2 and represents a possible mechanism of TSC2 regulation.
Tuberous sclerosis complex (TSC) 1 is a relatively common genetic disorder. TSC is caused by mutation in either the TSC1 (hamartin) or TSC2 (tuberin) gene, of which each contributes to ϳ50% of the genetic defects (1,2). Studies of TSC patients and animal models support the hypothesis that TSC1 and TSC2 are tumor suppressor genes. Homozygous deletion of either TSC1 or TSC2 in mice produces an embryonic lethal phenotype, suggesting an essential function in development. As predicted, heterozygous deletion of either TSC1 or TSC2 in mice results in a significant increase of carcinomas in many tissues with a 100% incidence of renal carcinomas (3,4).
Mutation of TSC1 or TSC2 results in similar phenotypes, suggesting that the two proteins function in the same pathway. Biochemical studies have shown that TSC1 and TSC2 form a stable complex (5). Genetic studies in Drosophila have demonstrated that TSC1⅐TSC2 plays a major negative role in the regulation of cell growth. Mutation of either TSC1 or TSC2 results in a significant increase of cell mass in Drosophila (6 -8). Overexpression of either TSC1 or TSC2 in Drosophila produces little phenotype, while co-expression of both TSC1 and TSC2 causes a significant reduction of cell size. Furthermore, genetic epistatic studies indicate that TSC1⅐TSC2 acts downstream of the insulin receptor (6 -8).
Recently, we and other groups have demonstrated that TSC1⅐TSC2 functions to inhibit S6K activation (9 -14). In TSC1Ϫ/Ϫ or TSC2Ϫ/Ϫ cells, S6K is highly activated. S6K activation requires phosphorylation of multiple sites. Interestingly, TSC1⅐TSC2 specifically inhibits the phosphorylation of Thr 389 , but not phosphorylation of Thr 421 and Ser 424 in S6K. Thr 389 is the primary site phosphorylated by mTOR (mammalian target of rapamycin) (15). Furthermore, TSC1⅐TSC2 also inhibits phosphorylation of 4E-BP1 (eukaryotic initiation factor 4E-binding protein 1), which is also an mTOR target. Both genetic data and biochemical data indicate that TSC1⅐TSC2 inhibits the function of mTOR (9,11). Several groups, including ours, have shown that TSC2 is directly phosphorylated and inhibited by Akt (11)(12)(13). These studies provide an important link between TSC2 and growth factor signaling.
In this report, we show that TSC2 interacts with 14-3-3, but not the binding-defective 14-3-3 mutant. This interaction is dependent on the phosphorylation of TSC2. We have identified that Ser 1210 of TSC2 is phosphorylated in vivo and is the primary binding site of 14-3-3. In contrast, mutation of all AKT phosphorylation sites in TSC2 had no effect on its interaction with 14-3-3. Overexpression of 14-3-3 enhanced phosphorylation of both S6K and 4E-BP1. Furthermore we demonstrated that interaction between TSC2 and 14-3-3 is also modulated by serum starvation.
Metabolic Labeling and Two-dimensional Phosphopeptide Mapping-HEK293 cells were co-transfected with the indicated plasmids. The serum-starved cells were washed twice with phosphate-free DMEM and incubated with 0.25 mCi/ml [ 32 P]orthophosphate (ICN) for 4 h. HA-tagged TSC2 was immunoprecipitated, resolved by SDS-PAGE, and transferred to a polyvinylidene difluoride membrane. Phosphorylated TSC2 was visualized by autoradiography. Phosphopeptide mapping was performed as described previously (11).
Stable Expression of TSC2 in LExF2 Cells by Retrovirus Infection-The TSC2 cDNA was subcloned to the retrovirial vector pPGS-CMV-CITE-Neo. The vectors containing TSC2 were transfected into the Phoenix retrovirus packaging cell line using the calcium phosphate method (Profection Kit, Promega). 48 h post-transfection, retrovirus produced by the Phoenix cells was used for infection of LExF2 cells. LExF2 cells stably expressing TSC2 were selected for and maintained by G418 (300 g/ml).
To directly demonstrate the phosphorylation status of Ser 1210 in TSC2, we performed in vivo 32 P labeling and two-dimensional phosphopeptide mapping of TSC2. The TSC2/S1210A mutant eliminated a single phosphopeptide spot depicted by the arrow in Fig. 4b, while the rest of phosphopeptides were unchanged (Fig.  4b). These results strongly indicate that Ser 1210 is an in vivo phosphorylation site in TSC2.
Binding of 14-3-3 may modulate the cellular function of TSC2. We have shown that one of the physiological functions of TSC1⅐TSC2 is to inhibit S6K activation. In TSC2Ϫ/Ϫ LExF2 cells, S6K is highly activated. The abilities of wild type and 14-3-3 binding-defective mutant TSC2 to inhibit S6K were tested in the TSC2Ϫ/Ϫ cells. We observed that both wild type FIG. 4. Phosphorylated Ser 1210 of TSC2 is the primary 14-3-3 binding site. a, phosphorylation of TSC2 is required for interaction with 14-3-3. GST-TSC2 was purified from transfected HEK293 cells and dephosphorylated with -phosphatase as indicated. The immunoprecipitated Myc-14-3-3␤ (from 300 g of transfected cell lysates) was incubated with the purified GST-TSC2 (ϳ10 ng). GST-TSC2 present in the anti-Myc-14-3-3 immunoprecipitates was detected by Western blot. b, Ser 1210 of TSC2 is phosphorylated. TSC2 and the TSC2/S1210A mutant was transfected into HEK293 cells and labeled with [ 32 P]phosphate. Two-dimensional phosphopeptide mapping was performed. The circle indicates the phosphopeptide absent in the mutant, but present in wild type TSC2. c, inhibition of S6K in the LExF2 cells by TSC2. The LExF2 TSC2Ϫ/Ϫ cells were infected with wild type TSC2 or TSC2 (S1210A) mutant, and stably infected cells were selected. The phosphorylation status of endogenous S6K was determined by phospho-specific antibodies (pS6K(T389)). d, serum starvation decreased interaction between TSC2 and 14-3-3. HA-TSC2transfected HEK293 cells were treated with PD98059 (50 M, 90 min), wortmannin (100 nM, 30 min), rottlerin (5 M, 60 min), or serum starvation (16 h). Co-immunoprecipitation of 14-3-3 was determined. e, expression of 14-3-3 enhanced phosphorylation of S6K and 4E-BP1. Increasing amounts of Myc-14-3-3␤ were transfected into HEK293 cells with HA-S6K or FLAG-4E-BP1. Basal phosphorylation of transfected S6K and 4E-BP1 was determined by anti-phospho-S6K and anti-phospho-4E-BP1. The expression of Myc-14-3-3␤ was also determined. IP, immunoprecipitation. and the 14-3-3 binding-defective TSC2 could inhibit S6K (Fig.  4c). These data indicate that 14-3-3 binding may not modulate the ability of TSC2 to inhibit S6K. However, the lack of a difference between the wild type and the mutant TSC2 could be due to the fact that the majority of the expressed TSC2 is not phosphorylated on Ser 1210 , therefore, free of 14-3-3 binding. Our two-dimensional phosphopeptide mapping data also indicates that the majority of TSC2 is not phosphorylated on Ser 1210 , because the intensity of this phosphopeptide is significantly weaker compared with the Akt phosphorylation site Ser 939 (Fig. 4b) (11).
Sequences surrounding Ser 1210 have limited resemblance to PKA and PKC-␦ phosphorylation sites. Inhibition of PKA (data not shown) or PKC-␦ by rottlerin had no effect on the interaction between TSC2 and 14-3-3 (Fig. 4d). We also observed that inhibition of the phosphatidylinositol 3-kinase-Akt pathway by wortmannin and the ERK pathway by PD90589 had no effect on the interaction (Fig. 4d), which suggests that the Akt phosphorylation sites are not involved. Interestingly, serum starvation resulted in a visible reduction of association between TSC2 and 14-3-3 (Fig. 4d). The above data indicate that the complex formation between TSC2 and 14-3-3 may be modulated by cell growth status.
The cellular functions of the TSC1⅐TSC2 tumor suppressor gene products have just begun to be elucidated. TSC1⅐TSC2 plays an important role in cell growth regulation and cell size control. Recent studies have demonstrated the TSC2 protein was phosphorylated and inhibited by Akt-dependent phosphorylation (11)(12)(13). In this report, we showed that TSC2 binds to 14-3-3 under physiological conditions. We have mapped a single site, Ser 1210 , in TSC2 responsible for binding with 14-3-3. The binding of 14-3-3 requires the phosphorylation of Ser 1210 . During the preparation of this manuscript, Nellist et al. (22) also reported that TSC2 interacts with 14-3-3. However, they showed that 14-3-3 binds to multiple sites in TSC2 and concluded that the Akt phosphorylation sites in TSC2 are responsible for 14-3-3 binding (22). We have no obvious explanation why the data by Nellist et al. (22) are dramatically different from ours. Nevertheless, our data clearly indicate that one major 14-3-3 binding site exists in TSC2, and Akt phosphorylation sites are not responsible for 14-3-3 binding.
Overexpression of 14-3-3 results in elevated phosphorylation of S6K and 4E-BP1. In contrast, overexpression of TSC1⅐TSC2 suppresses the phosphorylation of these two proteins, indicating 14-3-3 and TSC1⅐TSC2 have opposite effects on S6K acti-vation. Our results indicate that 14-3-3 binds to phosphorylated TSC2 and may suppress its activity. This interpretation is consistent with the fact that the interaction between TSC2 and 14-3-3 is decreased under serum-starved conditions. Serum starvation is predicted to activate TSC2 and suppress cell growth. The dissociation of 14-3-3 may partly contribute to TSC2 activation and S6K inhibition under conditions of serum starvation. However, we cannot exclude the possibility that the effect of 14-3-3 on S6K and 4E-BP1 may not be mediated by TSC2. It has been reported that 14-3-3 may positively modulate the function of TOR in yeast (23). 14-3-3 has also been shown to interact with mTOR (24). Therefore, 14-3-3 may regulate S6K and 4E-BP1 through multiple targets. Future studies to identify the kinase responsible for phosphorylation of Ser 1210 will provide new insights into the mechanism of TSC2 regulation.