Thyroid-stimulating Hormone/cAMP and Glycogen Synthase Kinase 3β Elicit Opposing Effects on Rap1GAP Stability*

Beyond regulating Rap activity, little is known regarding the regulation and function of the Rap GTPase-activating protein Rap1GAP. Tuberin and E6TP1 protein levels are tightly regulated through ubiquitin-mediated proteolysis. A role for these RapGAPs, along with SPA-1, as tumor suppressors has been demonstrated. Whether Rap1GAP performs a similar role was investigated. We now report that Rap1GAP protein levels are dynamically regulated in thyroid-stimulating hormone (TSH)-dependent thyroid cells. Upon TSH withdrawal, Rap1GAP undergoes a net increase in phosphorylation followed by proteasome-mediated degradation. Sequence analysis identified two putative destruction boxes in the Rap1GAP C-terminal domain. Glycogen synthase kinase 3β (GSK3β) phosphorylated Rap1GAP immunoprecipitated from thyroid cells, and GSK3β inhibitors prevented phosphorylation and degradation of endogenous Rap1GAP. Co-expression of GSK3β and Rap1GAP in human embryonic kidney 293 cells stimulated proteasome-dependent Rap1GAP turnover. Mutational analysis established a role for serine 525 in the regulation of Rap1GAP stability. Overexpression of Rap1GAP in thyroid cells impaired TSH/cAMP-stimulated p70S6 kinase activity and cell proliferation. These data are the first to show that Rap1GAP protein levels are tightly regulated and are the first to support a role for Rap1GAP as a tumor suppressor.

Rap1 belongs to the Ras superfamily of small G-proteins (1,2). Unlike Ras where there is abundant evidence supporting its role in tumorigenesis, far less is known regarding the role of Rap1 in neoplastic transformation. Early studies by Altschuler and co-workers (3,4) demonstrated that overexpression of Rap1 stimulated morphological transformation and altered growth properties, albeit in a cell type-specific fashion. Mutations in RapGEFs 1 have been identified in tumors and cancer cell lines. A subset of myeloid leukemias in BXH-2 mice con-tains proviral insertions in the CalDAG-GEFI gene, resulting in its activation (5). Several human cancer cell lines exhibit mutations in DOCK4, a Rap-activating protein (6). Alterations in RapGAPs have also been associated with aberrant cellular proliferation. Tuberin (7) is inactivated in the tumor predisposition syndrome, tuberous sclerosis type 2 (TSC-2). E6TP1 is targeted for ubiquitin-mediated protein turnover by the human papilloma virus, an event that appears to be required for E6mediated transformation (8). Finally, mice lacking SPA-1 develop myelodysplastic disorders (9). Far less is known regarding the role of Rap1GAP, the first Rap1GAP to be identified (10 -12). Rap1GAP is expressed in a tissue-specific fashion (10). Interestingly, its expression is low in proliferating cells and increases upon differentiation (13).
We report that Rap1GAP is highly expressed in TSH-dependent thyroid cells. This is striking in that thyroid cells are one of very few cellular models where cAMP stimulates cell proliferation and differentiation (reviewed in Ref. 14), effects that are mediated at least partly through Rap1 (4,15). Moreover, Rap1GAP exhibits a number of unique features including phosphorylation by protein kinase A (16) and interaction with G␣ subunits (17)(18)(19) that could be particularly pertinent to TSH signaling. These findings prompted us to examine whether TSH regulates Rap1GAP. Our results demonstrate that endogenous Rap1GAP protein levels are dynamically regulated and that TSH and GSK3␤ elicit opposing effects on Rap1GAP protein stability. In addition, our studies revealed striking similarities in the regulation and biological effects of Rap1GAP with those of tuberin, E6TP1, and SPA-1.

MATERIALS AND METHODS
Reagents-pCMV2-FLAG-Rap1GAP and pMT2-HA-Rap1GAP expression vectors were kind gifts from Drs. Lawrence Quilliam (Indiana University) and Johannes L. Bos (Utrecht University), respectively. The GSK3␤ serine 9 to alanine mutation was generated in human Myc-tagged GSK3␤ in pCS2 vector (20) by site-directed mutagenesis using QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Anti-Rap1GAP serum was generously provided by Dr. Michiyuki Matsuda (Osaka University). FLAG and GSK3␤ (raised to rabbit GSK3␤) antibodies were purchased from Sigma and Transduction Laboratories (Lexington, KY), respectively. HA and phospho-S6 antibodies were generous gifts from Drs. Jeffrey Field and Morris Birnbaum (University of Pennsylvania), respectively. Rap1, actin, and p70S6K antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antiphospho-Thr-389-p70S6K antibody was obtained from Cell Signaling Technology (Beverly, MA). Thyroglobulin antibody was purchased from Dako Corporation (Carpinteria, CA). Alkaline phosphatase and LAR tyrosine phosphatase were from Promega (Madison, WI). Kenpaullone and MG132 were purchased from Calbiochem.
Cell Culture-WRT cells were propagated at 37°C in 5% C0 2 in Coon's modified Ham's F-12 medium containing 3H (TSH, insulin, transferrin, and 5% calf serum) as described previously (15). Cells were starved in basal medium (growth factor-free Coon's modified Ham's F-12 medium) as described previously (15). Stable cell lines expressing FLAG-Rap1GAP were generated by co-transfection with an expression vector encoding G418 resistance followed by selection in 300 g/ml and maintenance in 150 g/ml Geneticin (Invitrogen).
Transient Transfection-Human embryonic kidney 293T cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Transfection was carried out using LipofectAMINE 2000 (Invitrogen) according to manufacturer's recommendations. Cells were plated in 12-well plates 18 h prior to transfection. Just prior to transfection, cells were transferred into Opti-MEM I (Invitrogen) and exposed to 100 ng of FLAG-Rap1GAP DNA, 50 ng of GSK3␤S9A DNA (or empty vector), and 4 l of LipofectAMINE 2000 in 100 l of Opti-MEM I. Transfection was for 5 h, and the cells were harvested and analyzed 24 h later.
BrdUrd Labeling-WRT cells plated on glass coverslips in 35-mm dishes were transfected with 2 g of HA-Rap1GAP DNA and 3 l of FuGENE 6 (Roche Applied Science) in 100 l of Opti-MEM I. After 5 h, cells were transferred to 3H medium overnight. Cells were then starved in basal medium for 24 h and stimulated with 1 milliunit/ml TSH, 5% bovine calf serum for 24 h. BrdUrd was added for the last 4 h. Cells were fixed and stained for incorporated BrdUrd as described previously (21).
Pulse-Chase Analysis-Pulse-chase was performed essentially as described previously (23). WRT cells were incubated in cysteine-and methionine-free RPMI 1640 medium supplemented with TSH (1 milliunit/ml) for 30 min and then labeled with [ 35 S]methionine/cysteine (160 Ci/ml) (Trans-label, Fisher Scientific, Pittsburgh, PA) in the presence of TSH for 2 h. Duplicate plates were transferred to basal medium in the presence or absence of TSH and chased for various times. To examine the effects of proteasome inhibitors, cells in TSH-supplemented RPMI 1640 medium were pulse-labeled for 2 h and then transferred to basal medium in the presence or absence of 25 M MG132 for various times. Following lysis in RIPA buffer (15), cell extracts were incubated with anti-Rap1GAP serum for 1 h at 4°C followed by protein A-agarose for 1 h at 4°C. Immune complexes were washed with RIPA buffer, heated to 95°C in Laemmli sample buffer, and resolved by SDS-PAGE. Gels were fixed, soaked in Amplify solution (Amersham Biosciences), dried, and subjected to autoradiography at Ϫ80°C. All of the data were analyzed on a PhosphorImager Storm 840 (Molecular Dynamics, Amersham Biosciences) using ImageQuant software.
Phosphatase Treatment-For alkaline phosphatase experiments, conditions were derived where TSH-induced serine/threonine phosphorylation was abolished. 2 To verify the activity of LAR tyrosine phosphatase, conditions were derived where insulin-induced tyrosine phosphorylation was abolished. WRT cells were lysed in 50 mM Tris, pH 8.0, 0.1% Triton X-100, and 2 mM MgCl 2 . 60 g of cell extract were incubated with 30 units of alkaline or tyrosine phosphatase for 1 h according to the manufacturer's recommendations. The reaction was terminated by the addition of 6ϫ Laemmli sample buffer, and samples were heated at 95°C for 5 min and resolved on 8% SDS-PAGE.
GSK3␤ Assays-GSK3␤ assays were performed as described previously (24,25). Rap1GAP was immunoprecipitated from WRT cells, and the immune complexes were washed with RIPA and then kinase (20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 5 mM dithiothreitol) buffers. Kinase reactions were carried out in 30 l of kinase buffer containing 0.2 l of rabbit GSK3␤ (500,000 units/ml, New England Biolabs, Inc., Beverly, MA) and 5 Ci of [␥-32 P]ATP for 25 min at 30°C and stopped by the addition of Laemmli sample buffer and heating for 5 min at 95°C. Proteins were separated on 8% PAGE, transferred to nitrocellulose membrane, exposed to film, and blotted with anti-Rap1GAP serum.

TSH Regulates Rap1GAP
Protein Stability-Immunoblotting using a peptide-directed Rap1GAP serum demonstrated a prominent 95-kDa band and a spectrum of faster migrating protein species in lysates prepared from growing WRT cells (Fig. 1A, 3H). A similar pattern has been reported previously in other cells (10,11,26,27). When lysates prepared from cells transferred to growth factor-free basal medium were analyzed, the Rap1GAP serum detected reduced levels of a discrete upshifted protein species (Fig. 1A, basal). Remarkably, the inclusion of either TSH or forskolin in basal medium prevented the upshift and decrease in Rap1GAP expression.
A time course experiment was performed to examine the kinetics over which these changes took place. Cells growing in 3H were transferred to basal medium for times ranging from 15 min to 24 h, and Rap1GAP expression was analyzed by Western blotting (Fig. 1B). Within 1 h following transfer to basal medium, Rap1GAP migration was upshifted. This was followed by a rapid decline in the abundance of faster migrating Rap1GAP species, which were absent from cells transferred to basal medium for as little as 4 h. Thus, total Rap1GAP levels were significantly decreased upon TSH withdrawal for as little as 4 h.
Rap1GAP is a substrate for phosphorylation (10,11,26). To examine whether phosphorylation contributed to the Rap1GAP upshift seen upon TSH withdrawal, lysates prepared from cells in basal medium were subjected to in vitro alkaline phospha- In contrast, exposure to a protein tyrosine phosphatase had no effect (compare lanes 3 and 6). These results suggest that TSH withdrawal results in a net increase in serine/threonine phosphorylation of Rap1GAP followed by a decline in Rap1GAP expression. Similar results were obtained in two other rat thyroid cell lines, PC-Cl3 and FRTL-5, 2 indicating that regulation of Rap1GAP is a conserved feature in thyroid cells.
To explore the mechanism through which TSH regulates Rap1GAP expression, pulse-chase experiments were performed. Cells were labeled with [ 35 S]methionine/cysteine in TSH-supplemented medium for 2 h and then chased in the presence or absence of TSH for various times. Rap1GAP was immunoprecipitated, resolved by SDS-PAGE, and subjected to autoradiography. Rap1GAP levels decreased by 50% within 5 h of TSH withdrawal (Fig. 2, open circles). In the presence of TSH, Rap1GAP levels decreased by only 10% over the same time period (closed squares). Not only did TSH stabilize Rap1GAP, it also prevented the upshift in the migration of Rap1GAP (Fig. 2). Together, these results suggest that TSH regulates the phosphorylation status of Rap1GAP, resulting in the stabilization of Rap1GAP protein levels. As phosphorylation triggers the proteasomal degradation of many proteins (reviewed in Refs. 28 and 29), experiments to determine whether Rap1GAP is a target for proteasome-mediated turnover were performed.
Rap1GAP Is a Target for Proteasomal Degradation-Proteasome inhibitors were used in pulse-chase experiments to assess whether inhibition of the proteasome stabilized Rap1GAP. Cells were pulse-labeled for 2 h in TSH-supplemented basal medium in the absence or presence of MG132 (30,31) and then chased in basal medium without TSH in the presence or absence of proteasome inhibitor. As seen before (Fig. 1B), Rap1GAP was upshifted within 2 h following TSH withdrawal and its abundance declined over the next 4 h (Fig. 3A). Although inclusion of MG132 did not prevent the Rap1GAP upshift, it markedly delayed the decline in Rap1GAP (Fig. 3B,  open circles). Two additional proteasome inhibitors, MG115 and lactacystin, also increased Rap1GAP protein levels. 2 These results support the idea that Rap1GAP is a substrate for proteasome-mediated degradation.
An analysis of the primary structure of Rap1GAP revealed 56 potential serine/threonine phosphorylation sites. Within these sites, there were two DSGXXS destruction box motifs analogous to one found in ␤-catenin (Fig. 4A) (32). Phosphorylation of serine residues within this motif by GSK3␤ triggers the recognition of ␤-catenin by the F-box protein ␤TrCP, leading to its ubiquitination and degradation (reviewed in Ref. 33). To determine whether Rap1GAP was a GSK3␤ substrate, Rap1GAP was immunoprecipitated from growing WRT cells and used as a substrate for recombinant GSK3␤ in an immune complex kinase assay. The data shown in Fig. 4B demonstrate that GSK3␤ phosphorylates Rap1GAP in vitro.
To determine whether GSK3␤ phosphorylates endogenous Rap1GAP, WRT cells were transferred to basal medium in the presence or absence of LiCl, a well characterized inhibitor of GSK3␤ activity (24). Inclusion of 5-40 mM LiCl prevented the upshift and attenuated the loss of faster migrating Rap1GAP species (Fig. 4C). This was not due to changes in osmolarity, because the transfer to NaCl-supplemented basal medium did not alter Rap1GAP mobility or expression. Furthermore, the effects of LiCl on Rap1GAP were reproduced by kenpaullone, an alternative GSK3␤ inhibitor (Fig. 4D) (34). When used independently, these inhibitors are highly specific for GSK3␤ (35). Therefore, these data provide strong support for the role for GSK3␤ in targeting Rap1GAP for proteasome-mediated degradation.
GSK3␤ Regulates Rap1GAP Protein Levels-To prove that GSK3␤ can regulate Rap1GAP stability, transient overexpression studies were performed in human embryonic kidney 293T cells, which exhibit significantly higher transfection efficiencies than thyroid cells. FLAG-tagged Rap1GAP was co-expressed with GSK3␤ and Rap1GAP protein levels assessed by Western blotting 24 h post-transfection. We used human GSK3␤S9A, which contains a serine to alanine substitution at position 9 to prevent the inactivating phosphorylation of GSK3␤ at this site (36,37) that occurs constitutively in 293T cells. 3  GSK3␤S9A but not by empty vector (Fig. 5A). Mutation of a serine residue within the destruction box (S33Y) stabilizes ␤-catenin protein by preventing GSK3␤-mediated phosphorylation (38 -41). Therefore, Rap1GAP mutants were constructed in which the first serine in each destruction box motif was changed to isoleucine, N-terminal Rap1GAPS525I (N-Rap1GAP) and C-terminal Rap1GAPS606I (C-Rap1GAP). A double mutant, N-and C-terminal Rap1GAPS525/606I (NC-Rap1GAP), was also constructed. Unlike wild-type Rap1GAP, N-Rap1GAP was stable in the presence of GSK3␤S9A. Additionally, N-Rap1GAP was consistently expressed at higher levels than was C-Rap1GAP, whose expression was decreased by GSK3␤S9A. As expected, the mutation of both serine residues abolished down-regulation of Rap1GAP by GSK3␤S9A. The ability of GSK3␤S9A to decrease Rap1GAP expression was impaired by the proteasome inhibitor, MG132 (Fig. 5B). Col-lectively, these findings demonstrate that Rap1GAP can be targeted for proteasomal degradation by phosphorylation of serine 525 by GSK3␤.
Rap1GAP Impairs Cell Proliferation and p70S6 Kinase-The ability of TSH to stabilize Rap1GAP implied an important role for Rap1GAP in mediating TSH effects in thyroid cells. To explore this possibility, WRT cells stably overexpressing Rap1GAP were isolated. Of 12 cell lines screened, only two overexpressed Rap1GAP compared with parental cells and the level of overexpression was modest (Fig. 6A). Nonetheless, modest overexpression of Rap1GAP was sufficient to impair Rap1 activation by TSH (Fig. 6B) and forskolin. 2 Rap1GAP expression was regulated similarly in the overexpressing cells as in parental cells. Upon removal of TSH, Rap1GAP was upshifted and its expression decreased (Fig. 6A, basal or B). Intriguingly, Rap1GAP-overexpressing cells grew more slowly than parental cells (Fig. 6C). To ensure that this was not attributed to secondary changes associated with the isolation of stable cell lines, the effects of transient overexpression of Rap1GAP on DNA synthesis were investigated. Both basal and TSH/serum-stimulated DNA synthesis were impaired in Rap1GAP-transfected cells (Fig. 6D). These results are strikingly similar to those reported for tuberin where overexpression slowed cell proliferation (42,43). Indeed, as for tuberin (44,45), overexpression of Rap1GAP impaired activation of p70S6K kinase in WRT cells. However, in this case, inhibition was specific for cAMP-elevating agents. Fig. 7 demonstrates that   FIG. 4. Rap1GAP is phosphorylated by GSK3␤. A, the C-terminal region of human Rap1GAP (NCBI accession number P47736) contains two putative destruction box motifs (amino acids 524 -529 and 605-610) analogous to that found in mouse ␤-catenin (amino acids 32-37, NCBI accession number S35091). The N-and C-boxes in rat Rap1GAP (NCBI accession number XP_233609) include amino acids 555-560 and 636 -641, respectively. B, Rap1GAP was immunoprecipitated from growing cells and used as a substrate for purified GSK3␤ in vitro. Rap1GAP was phosphorylated in immune complexes prepared with the Rap1GAP serum (IP:Rap1GAP) and exposed to GSK3␤ (lane 1). No phosphorylation was detected in the absence of GSK3␤ (lane 2), although Rap1GAP was present in the precipitates or in immune complexes prepared using nonspecific rabbit IgG (lane 3). Autophosphorylation of GSK3␤ was observed in Rap1GAP and control immunoprecipitates (lanes 1 and 3). The data represent one of two experiments performed with similar results. C, cells were transferred from 3H to basal medium for 3 h in the presence or absence of LiCl or NaCl at the concentrations indicated, and Rap1GAP expression was examined by Western blotting. Four experiments were performed with similar results. D, cells were transferred to basal medium for 4 h in the presence or absence of 10 M kenpaullone (Kenp), and Rap1GAP expression was analyzed. The data represent one of two experiments performed with similar results.

FIG. 5. GSK3␤ targets Rap1GAP for proteasome-mediated degradation in 293T cells.
A, FLAG-tagged Rap1GAP was co-expressed with GSK3␤S9A in human embryonic kidney 293T cells. Rap1GAP expression was analyzed by Western blotting (WB) with anti-FLAG antibody, and GSK3␤S9A expression was detected by immunoblotting with anti-GSK3␤ antibody. Myc-tagged GSK3␤S9A appears as a minor species migrating more slowly than endogenous GSK3␤. Wild-type Rap1GAP (Rap1GAP) contains two putative destruction boxes. Rap1GAP mutants contain serine to isoleucine mutations in the first serine residue in the N-terminal (N-Rap1GAP), C-terminal (C-Rap1GAP), or both (NC-Rap1GAP) destruction boxes. The results shown are representative of three experiments performed with similar results. B, Rap1GAP was co-expressed with GSK3␤S9A, and the transfected cells were incubated in the presence or absence of 25 M MG132 for the last 3 h prior to lysis. Two experiments were performed with similar results. the ability of TSH and forskolin to stimulate the phosphorylation of ribosomal S6 protein, a substrate of p70S6K, was markedly impaired in Rap1GAP-overexpressing cells. p70S6K activity is regulated by multi-site phosphorylation including phosphorylation on threonine 389 by mTOR (46,47). Consistent with its effects on S6 phosphorylation, Rap1GAP overexpression attenuated the effects of TSH and forskolin on mTORdependent phosphorylation of p70S6K (Fig. 7, lower panel). Insulin-and serum-stimulated p70S6K activities assessed by S6 or p70S6K protein phosphorylation were only modestly decreased by Rap1GAP overexpression (Fig. 7). Because p70S6K activity is required for thyroid cell proliferation (48,49), the slower growth rate of the Rap1GAP-overexpressing cells could be a consequence of impaired p70S6K activity.
Alterations in RapGAP expression have been reported in myeloid cells (13). Immature bone marrow cells express high levels of SPA-1 but not Rap1GAP. Upon maturation, the expression of SPA-1 decreases and Rap1GAP expression is increased (9). As TSH is the primary regulator of differentiated gene expression in thyroid cells, we assessed whether increased Rap1GAP expression induced changes in differentiated gene expression. Overexpression of Rap1GAP markedly enhanced the effects of TSH and forskolin on thyroglobulin expression, a marker of thyroid differentiation (Fig. 8). The effects of TSH on Rap1GAP stability, together with the alterations in growth and differentiated gene expression in Rap1GAP-overexpressing cells, support an important role for Rap1GAP in glycoprotein hormone action. DISCUSSION Prior to this report, little was known regarding the regulation or function of Rap1GAP. We now present a model for the dynamic regulation of Rap1GAP and new insight into potential roles for Rap1GAP in the regulation of thyroid cell proliferation and differentiation.
Rap1GAP exists as multiple phosphorylated protein species in thyroid cells growing in the presence of TSH. Upon TSH withdrawal, Rap1GAP undergoes a phosphatase-sensitive upshift followed by a rapid decline in its expression. Treatment with cAMP-elevating agents including TSH or with GSK3␤ inhibitors prevented both the upshift and decline in Rap1GAP protein levels. On the other hand, proteasome inhibitors stabilized Rap1GAP but had no effect on the Rap1GAP upshift. Together, these data support a model wherein TSH withdrawal enhances phosphorylation of Rap1GAP by GSK3␤, an event that triggers the proteasomal degradation of Rap1GAP. In support of this model, we identified a destruction box in Rap1GAP similar to the motif in ␤-catenin (33). Co-expression of Rap1GAP with GSK3␤ in 293 cells decreased Rap1GAP levels in a proteasome-dependent manner, and mutation of serine 525 in the N-terminal destruction box (Rap1GAPS525I) rendered Rap1GAP insensitive to GSK3␤. These studies clearly reveal the potential for regulation of Rap1GAP by GSK3␤. Three lines of evidence support the similar regulation of endogenous Rap1GAP by GSK3␤ in thyroid cells. First, cellular Rap1GAP was stabilized by inclusion of lithium or kenpaullone. Because the inhibitory profiles of lithium and kenpaullone on other protein kinases do not overlap (35), this provides compelling evidence that cellular Rap1GAP is sensitive to GSK3␤ activity. Second, treatment with three different proteasome inhibitors stabilized endogenous Rap1GAP levels. Third, Rap1GAP immunoprecipitated from thyroid cells was a substrate for GSK3␤.
The mechanism through which TSH stabilizes Rap1GAP is not yet clear. Several kinases activated by TSH including protein kinase A, Akt, and p70S6K inhibit GSK3␤ activity via phosphorylation on serine 9 (36, 50 -54). However, inhibition of these kinases had no effect on the ability of TSH to stabilize Rap1GAP protein levels. 2 Treatment with okadaic acid stimulated a Rap1GAP upshift when growing cells were transferred to basal or TSH-supplemented basal medium. 2 Cumulatively, these findings suggest that TSH stabilizes Rap1GAP through effects on an okadaic acid-sensitive phosphatase, possibly protein phosphatase 2A.
Rap1GAP is not the first example of a Rap1GAP targeted for proteasomal turnover. E6TP1 interacts with the high risk human papillomavirus E6 oncoprotein, an interaction that targets E6TP1 for proteasomal degradation (8). Tuberin, the product of the tsc2 locus (55), is stabilized through its interaction FIG. 7. Overexpression of Rap1GAP impairs p70S6K activity. Rap1GAP-overexpressing and parental cells were starved in basal medium for 48 h and stimulated with 1 milliunit/ml TSH (T), 10 M forskolin (F), 10 g/ml insulin (I), or 5% bovine calf serum (S) for 45 min. Total cell lysates were prepared and analyzed by immunoblotting with antibodies raised to phosphorylated ribosomal S6 protein (S6-P) and phospho-Thr-389 p70S6K (p70K-P). Similar results were obtained in four experiments.
FIG. 6. Rap1GAP impairs thyroid cell proliferation. A, Rap1GAP expression was examined in Rap1GAP-overexpressing and parental cells growing in 3H or following transfer to basal medium (B) for 48 h. B, Rap1GAP-overexpressing and parental cells were starved in basal medium for 48 h and then stimulated with TSH for the indicated times (in minutes). Rap1 activation was assessed by interaction assay as described previously (15). Activated and total Rap1 expression was examined using a polyclonal Rap1 antibody. The results shown are representative of two experiments. C, 2 ϫ 10 5 Rap1GAP-overexpressing (open circles) and parental cells (filled squares) were plated overnight in 3H. Replicate plates were harvested, and cell number was determined at days 1, 2, 4, and 6 as described previously (15). Data  with the TSC1 gene product, hamartin (56). Phosphorylation of tuberin by Akt destabilizes this complex and results in the ubiquitination and proteasome-dependent degradation of tuberin (44,57). Although the mechanisms through which these RapGAPs are targeted for degradation differ from that for Rap1GAP in thyroid cells, the finding that their protein levels are dynamically regulated supports important regulatory roles for these proteins. Tuberin functions as a tumor suppressor and, when overexpressed, impairs cell proliferation (42,58). Growth inhibition is mediated through the ability of the tuberin-hamartin complex to inhibit p70S6K activity by blocking mTOR-mediated phosphorylation of threonine 389 (44,45). Even when modestly overexpressed, Rap1GAP impaired mTORdependent phosphorylation of p70S6K, an important mediator of proliferation in thyroid cells (48,49), and slowed thyroid cell proliferation. Collectively, these data reveal an interesting duality in the regulation of Rap1GAP and tuberin protein levels as well as in their ability to inhibit cell proliferation. The ability of Rap1GAP to impair cAMP-dependent proliferation in thyroid cells is intriguing in that thyroid cells are one of very few cellular models where cAMP stimulates proliferation (reviewed in Refs. 14 and 59) and where Rap1 has been proposed to function as an oncogene (4).
The stabilizing effects of TSH on Rap1GAP, a putative growth suppressor, would seem counter-intuitive as TSH is strictly required for thyroid cell proliferation. However, thyroid cell proliferation is regulated by the cooperative action of TSH, insulin, and serum (14). These factors may act in concert to dynamically regulate the stability and/or growth-suppressive activity of Rap1GAP. A primary role of TSH is to regulate the expression of genes involved in thyroid hormone biosynthesis including thyroglobulin. Overexpression of Rap1GAP enhanced the effects of TSH on thyroglobulin expression, suggesting that Rap1GAP plays a role in TSH-stimulated differentiation.
A critically important issue that remains to be addressed is whether Rap1GAP elicits effects distinct from the negative regulation of Rap1 activity. Although overexpression of either Rap1GAP or Rap1A17N impaired Rap1 activity, their other effects in thyroid cells were distinct. Rap1A17N-expressing cells exhibited an enhanced growth rate and impaired differentiated gene expression (15), effects opposite from those induced by Rap1GAP. Rap1GAP impaired TSH/cAMP-stimulated p70S6K activity, whereas dominant negative Rap1A had no effect. Although confounded by issues associated with stable overexpression, these results raise the intriguing possibility that the effects of Rap1GAP may not be limited to impaired Rap1 activity. Stable complexes among Gz, Rap1, and Rap1GAP have been reported previously (19), raising the interesting possibility that multi-protein complexes containing Rap1GAP1 may elicit activities of their own.