Rheb

Protein Interactions

Mutations in the TSC1 or TSC2 genes cause tuberous sclerosis, a benign tumor syndrome in humans. Tsc2 possesses a domain that shares homology with the GTPase-activating protein (GAP) domain of Rap1-GAP2, suggesting that a GTPase might be the physiological target of Tsc2. The small GTPase Rheb (Ras homolog enriched in brain) has been shown to be a direct target of Tsc2 GAP activity both in vivo and in vitro. Point mutations in the GAP domain of Tsc2 disrupt its ability to regulate Rheb without affecting the ability of Tsc2 to form a complex with Tsc1. These studies identify Rheb as a molecular target of the TSC tumor suppressor genes (Zhang, 2003).

TSC1 and TSC2 were initially discovered as tumor suppressor genes mutated in tuberous sclerosis, a human syndrome characterized by the widespread development of benign tumors termed harmatomas. TSC2 encodes a putative GAP protein, whereas TSC1 encodes a novel protein containing two coiled-coil domains. Studies of Drosophila TSC1 and TSC2 homologs have identified a specific function for TSC1-TSC2 in the control of cell growth, with loss of TSC1-TSC2 resulting in increases in cell size. Recent studies further suggest that Tsc1-Tsc2 antagonizes the amino-acid-TOR signalling pathway, which normally couples amino-acid availability to S6 Kinase (S6K) activation, translation initiation and cell growth. Strikingly, loss of Drosophila TSC1-TSC2 results in a TOR-dependent increase of S6K activity that is resistant to amino-acid starvation (Zhang, 2003 and references therein).

Despite these new advances, the biochemical activity of the Tsc1-Tsc2 complex remains unknown. Tsc2 possesses a domain homologous to Rap1-GAP. The GAP homology domain of Tsc2 is important for its function, and mis-sense mutations of this domain have been identified in a high proportion of TSC patients. These observations suggest that an unknown small GTPase might be the direct target of Tsc2. This study set out to determine the target GTPase of Tsc2-GAP using an RNAi-based screen in Drosophila S2 cells. It was reasoned that this putative GTPase should be expressed in S2 cells and that RNAi of this GTPase should result in downregulation of S6K-Thr 398 phosphorylation, a phenotype opposite that caused by Tsc2 RNAi. During the course of the RNAi screen, genetic studies have implicated the small GTPase Rheb as a potential target of Tsc2. In S2 cells, RNAi inhibition of Rheb, but not any of the other 17 GTPases tested so far, abolished S6K-Thr 398 phosphorylation, as predicted for a Tsc2 GAP substrate. Among the 17 GTPases screened were Rab5 and Rap1, two proteins previously implicated as TSC2 GAP substrates from in vitro studies, suggesting that Rab5 and Rap1 are improbable physiological substrates of Tsc2. The highly specific effect of Rheb RNAi on S6K phosphorylation suggests that Rheb might be the physiological substrate of TSC2 GAP activity (Zhang, 2003).

Rheb is an evolutionarily conserved small GTPase found from yeast to mammals. Unlike Ras and most other Ras superfamily GTPases, Rheb has an arginine at the third residue of the G1 box (residue 15 of mammalian Rheb) instead of glycine. Rheb is unique, compared with many small GTPases, in that it exists in a highly activated state in mammalian cells. Studies of mammalian Rheb further implicated the existence of a Rheb-GAP that is normally present at relatively limiting concentrations, since overexpression of Rheb results in a progressive increase in the proportion of Rheb in the active GTP-bound state. Genetic analyses in Drosophila support a model in which Tsc2 functions as a Rheb-GAP. These studies also suggest that similarly to mammalian cells, Tsc2, the putative Rheb-GAP, is normally present in limiting concentrations in Drosophila, because overexpression of wild-type Rheb results in an activated phenotype and overexpression of Tsc2 (together with Tsc1) results in the opposite phenotype (Zhang, 2003).

To test directly whether Rheb is a physiological substrate of Tsc2 GAP activity, it was asked if Tsc2 could regulate Rheb in vivo. Rheb, similar to other small GTPases, cycles between an active GTP-bound form and an inactive GDP-bound form. Thus, the steady state GTP/GDP-loading status of Rheb can be used as a measurement of its in vivo activity. An in vivo labelling procedure was adapted to analyse the steady-state GTP/GDP-binding status of Rheb. Drosophila S2 cells expressing Myc-tagged Rheb were labelled with ³²P-orthophosphate. Rheb protein was then purified by immunoprecipitation and Rheb-associated GTP/GDP was analysed by thin-layer chromatography (TLC) on polyethyleneimine (PEI) cellulose plates. In wild-type S2 cells, Rheb binds preferentially to GTP, in agreement with studies of mammalian Rheb. In addition, co-overexpression of Tsc1 and Tsc2 results in a marked decrease (approximately eightfold) in the ratio of GTP to GDP bound on Rheb. Interestingly, overexpression of Tsc2 alone has much weaker effect on GTP:GDP ratio. This observation is consistent with previous studies in Drosophila, which show that co-overexpression of Tsc1 and Tsc2, but not either gene alone, results in growth inhibition. The weaker effect of Tsc2 alone on Rheb GTP loading is caused, at least in part, by the lower level of Tsc2 when expressed alone, as compared with Tsc1 co-expression. Mutual stabilization between Drosophila Tsc1 and Tsc2 has been documented previously (Zhang, 2003).

To demonstrate that the effect of Tsc1-Tsc2 overexpression on Rheb GTP loading was caused by the GAP activity of Tsc2, similar in vivo labelling experiments were performed with Tsc2 variants carrying point mutations in the GAP domain. The mutations Tsc2^K1693A and Tsc2^N1698K changed residues in the GAP domain that are conserved in Drosophila, human and a probable Schizosaccharomyces pombe Tsc2 homolog. In addition, a mutation analogous to Tsc2^K1693A has been shown to abolish Rap1-GAP activity, whereas Tsc2^N1698K mimics a disease-causing mutation in human TSC patients. The activity of Tsc2-N, a construct that contains just the amino-terminal half of Tsc2 and thus lacks the carboxy-terminal GAP domain, was also examined. Tsc2-N can associate with Tsc1 normally, but does not interact with Rheb in co-immunoprecipitation assays. Similar to Tsc2-N, neither Tsc2^K1693A nor Tsc2^N1698K affects the ability of Tsc2 to associate with Tsc1. Despite their ability to associate with Tsc1, these mutants all abolished the effect of Tsc1-Tsc2 overexpression on Rheb GTP loading. Complementary to the results from Tsc1-Tsc2 overexpression, RNAi of Tsc2 increases the ratio of GTP:GDP bound to Rheb. The smaller change in GTP:GDP ratio after Tsc2 RNAi, compared with Tsc1-Tsc2 overexpression, is not surprising given that Rheb is already at a relatively active state in wild-type cells. Taken together, these results provide strong evidence that Rheb is a physiological target of Tsc2 GAP activity (Zhang, 2003).

To test whether Rheb is a direct substrate of Tsc2 GAP in vitro, a fusion protein of glutathione S-transferase (GST) and the Tsc2 GAP domain against GTP-loaded Rheb protein was tested using a nitrocellulose filter assay. alpha-³²P-GTP- or gamma-³²P-GTP-loaded GST-Rheb was incubated with GST-Tsc2 and the remaining radioactive GTP bound on Rheb was measured at different time intervals. GST-Tsc2 results in a dramatic decrease of Rheb-associated radioactive counts when gamma-³²P-GTP, but not alpha-³²P-GTP, was used in the assay. Thus, Tsc2 functions as a Rheb GAP in vitro. This GAP activity is highly specific, and no activity was detected, using as a substrate Drosophila Ras1, the closest relative of Rheb among all GTPases. In addition, the K1693A or the N1698K point mutation abrogates the in vitro GAP activity of Tsc2 towards Rheb. These results provide further evidence that Tsc2 functions as a Rheb GAP (Zhang, 2003).

The data presented so far suggest a model in which the tuberous sclerosis tumor suppressor proteins negatively regulate Rheb through the Rheb GAP activity of Tsc2. To further substantiate this model, whether there are any genetic interactions between TSC1-TSC2 and Rheb was tested. Flies homozygous for a null allele of TSC1, TSC1²⁹, do not survive beyond the second-instar larval stage. Strikingly, the lethality of TSC1 null animals was partially rescued by removing one of the two copies of Rheb gene from the diploid genome: 61% of TSC1²⁹ homozygotes that were also heterozygous for a null allele of Rheb, Rheb^PDelta1, survived to third-instar larval stage, and 21% of the third-instar survivors continued development and arrested at the pupal stage. Such dose-sensitive interactions are reminiscent of those observed between TSC1-TSC2 and TOR, further supporting the model that Tsc1-Tsc2 negatively regulates Rheb during cell growth (Zhang, 2003).

Finally, how the Tsc-Rheb pathway interacts with the amino acid-TOR-S6K signalling network was investigated. Tsc and Rheb could either function as obligatory components between amino acids and TOR in a linear amino-acid sensing pathway, or in a parallel pathway that converges on TOR. The former (but not the latter) model predicts that the activity of Rheb is dependent on the presence of amino acids. The ratio of GTP:GDP bound to Rheb is not reduced after 5 h of amino-acid starvation. Thus, a model is favored in which TSC and Rheb function in a parallel pathway that converges on TOR. According to this model, loss of Tsc1-Tsc2 or ectopic activation of Rheb results in constitutive activation of TOR, which bypasses the requirement for amino acids and renders S6K activity resistant to amino-acid starvation. How Rheb signals to TOR will be an important question for future investigation (Zhang, 2003).

In summary, the small GTPase Rheb is a direct target of the tuberous sclerosis tumor suppressor proteins. Wild-type Tsc2, but not mutant Tsc2 carrying point mutations in the GAP domain, shows GAP activity towards Rheb both in vitro and in vivo. The importance of Tsc2's GAP activity is further supported by the high proportion of mis-sense mutations localized to the Tsc2 GAP domain among TSC patients. Thus, the Tsc2 tumor suppressor functions as a Rheb-GAP in an analogous way to the neurofibromin (NF1) tumor suppressor as a Ras-GAP. These studies suggest that Rheb represents a novel target for therapeutic intervention in the TSC disease. The identification of a small GTPase as the direct target of the TSC tumor suppressors further implicates the existence of activators of GTPases, such as guanine nucleotide-exchange factors (GEFs), as potential regulators of this disease pathway. Identification of the putative Rheb-GEF represents an important goal for the next phase of TSC research (Zhang, 2003).

Drosophila TCTP is essential for growth and proliferation through regulation of dRheb GTPase

Cellular growth and proliferation are coordinated during organogenesis. Misregulation of these processes leads to pathological conditions such as cancer. Tuberous sclerosis (TSC) is a benign tumour syndrome caused by mutations in either TSC1 or TSC2 tumour suppressor genes. Studies in Drosophila and other organisms have identified TSC signalling as a conserved pathway for growth control. Activation of the TSC pathway is mediated by Rheb (Ras homologue enriched in brain), a Ras superfamily GTPase. Rheb is a direct target of TSC2 Gigas in Drosophila) and is negatively regulated by its GTPase-activating protein activity. However, molecules required for positive regulation of Rheb have not been identified. This study shows that a conserved protein, translationally controlled tumour protein (TCTP), is an essential new component of the TSC-Rheb pathway. Reducing Drosophila TCTP (dTCTP) levels reduces cell size, cell number and organ size, which mimics Drosophila Rheb (dRheb) mutant phenotypes. dTCTP is genetically epistatic to Tsc1 and dRheb, but acts upstream of dS6k, a downstream target of dRheb. dTCTP directly associates with dRheb and displays guanine nucleotide exchange activity with it in vivo and in vitro. Human TCTP (hTCTP) shows similar biochemical properties compared to dTCTP and can rescue dTCTP mutant phenotypes, suggesting that the function of TCTP in the TSC pathway is evolutionarily conserved. These studies identify TCTP as a direct regulator of Rheb and a potential therapeutic target for TSC disease (Hsu, 2007).

TCTP is a highly conserved protein upregulated in various tumours. Despite studies on the biochemical and structural properties of TCTP, the physiological significance of these findings has not been determined. Thus, the function of TCTP in vivo was studied using Drosophila as a model organism (Hsu, 2007).

First dTCTP expression was knocked down in developing flies by targeted expression of double-stranded RNA (dsRNA) for RNA interference (RNAi). Expression of dTCTP RNAi depleted endogenous dTCTP to nearly undetectable levels by different GAL4 drivers. Tissue-specific expression of dTCTP RNAi reduced the size of the eye, wing, notum, or a specific region in the wing pouch, corresponding to the expression domains of various GAL4 lines. The size reduction was caused by a decrease in both cell size and cell number, a typical phenotype for mutations in the insulin or TSC pathways. Ubiquitous expression of UAS-dTCTP RNAi by actin-GAL4 (act>dTCTPi) caused lethality around the third instar larval stage. A portion of these larvae was able to survive to the pupal stage with reduced body size, consistent with the organ size reduction. The lethality and phenotypes caused by dTCTP RNAi were rescued by co-expression of a dTCTP complementary DNA, indicating that these defects were due to a reduction of dTCTP levels. It is therefore concluded that dTCTP RNAi suppresses organ growth by affecting both cell size and number (Hsu, 2007).

The phenotypes of dTCTP loss-of-function mutants were further investigated because RNAi may represent a hypomorphic situation. dTCTP^Ey09182 is a hypomorphic allele of dTCTP resulting from a P-element insertion in its 5' untranslated region. Rare homozygous dTCTP^Ey09182 flies that escaped from lethality showed smaller body size compared with their heterozygous siblings. To create null alleles, imprecise excisions were generated from dTCTP^Ey09182. One imprecise excision line, dTCTP^h59, showed a 1.1-kilobase deletion downstream of the insertion site that removes the entire dTCTP coding sequence. Western blot analysis showed no detectable dTCTP protein in dTCTP^h59 early first instar larvae. Both dTCTP^h59 homozygotes and dTCTP^h59 heterozygotes for a deficiency chromosome uncovering the dTCTP region (dTCTP^h59/Df(3R)M-Kx1) showed 100% lethality at the late first instar stage, indicating that this allele is a genetic null. Expression of dTCTP cDNA or a genomic DNA construct was able to rescue dTCTP^h59 mutants, indicating that the lethality is due to deletion of the dTCTP gene (Hsu, 2007).

To examine the phenotypes of dTCTP null mutant cells, dTCTP mutant clones were generated using mitotic recombination. dTCTP^h59 mutant clones showed growth disadvantage compared to their wild-type twin spots. The sizes of dTCTP^h59 clones were similar to those of the twin spots 24 h after heat shock. However, the sizes of the twin spots were much larger than dTCTP^h59 clones 48 h after heat shock, and most dTCTP^h59 clones were eliminated by 60 h after heat shock. Similarly, using the EGUF/Hid technique (Stowers, 1999) to remove most wild-type cells in dTCTP^h59 mosaic eyes resulted in either a no-eye or small-eye phenotype. Therefore, the null phenotypes were qualitatively comparable to the dTCTP RNAi phenotype, but more severe (Hsu, 2007).

The reduction in cell number caused by dTCTP RNAi and the behaviour of dTCTP mutant cells may result from a proliferation defect or abnormal cell death. These possibilities were tested using the MARCM (mosaic analysis with a repressible cell marker) technique. Similar to clones generated by traditional mitotic recombination, numerous small dTCTP^h59 green-fluorescent-protein-expressing (GFP⁺) clones were observed at 24 h after heat shock. By 72 h after heat shock, very few GFP⁺ clones remained on the discs. In contrast, Cyclin E (CycE) overexpression, via the MARCM technique, within dTCTP^h59 clones resulted in four times more dTCTP^h59 cells at 72 h after heat shock. Similarly, CycE overexpression also suppressed the dTCTP RNAi phenotypes. Next whether the dTCTP^h59 phenotypes can be attributed to abnormal cell death was tested. Expression of the P35 cell death inhibitor also significantly suppressed the dTCTP^h59 phenotypes, leading to the presence of four times more GFP⁺ cells at 72 h after heat shock. These data indicate that loss of dTCTP causes defects in cell proliferation and triggers cell death (Hsu, 2007).

Insulin and TSC signalling are two parallel but interacting pathways for growth control. Inactivation of positive regulatory components, such as Insulin receptor (InR), dRheb and Tor, leads to a decrease in organ size by affecting cell size and cell number. In contrast, overexpression of InR and dRheb, as well as inactivation of negative regulatory components such as Tsc1 and Pten, causes tissue overgrowth. Given the similar phenotypes between dTCTP mutants and mutants in the insulin/TSC pathways, genetic epistasis experiments were performed to test whether dTCTP has a role in these two pathways. Overexpression of InR by patched (ptc)-GAL4 caused weak but consistent expansion of the distance between L3 and L4 wing veins. In contrast, co-expression of InR and dTCTP RNAi reduced the L3-L4 distance, resembling the dTCTP RNAi phenotype. Therefore, dTCTP is epistatic to (acts downstream of) InR (Hsu, 2007).

Next, the relationship between Tsc1 and dTCTP was examined. Mutations in Tsc1 or Tsc2 cause similar phenotypes because they function as a complex. Mosaic eyes and heads consisting primarily of Tsc1 mutant cells were much larger than wild type. Expression of dTCTP RNAi in Tsc1 mutant cells suppressed this overgrowth both in the eye and head. Furthermore, when the eye was composed of Tsc1 and dTCTP double mutant cells, the flies displayed a small or no-eye phenotype indistinguishable from the dTCTP single mutant phenotype, suggesting that dTCTP acts downstream or in parallel to Tsc1 (Hsu, 2007).

The relationship between dTCTP and dRheb was tested. Ectopic expression of dRheb using ptc-GAL4 resulted in a 15% increase in the L3-L4 distance compared with the ptc>GFP control. However, co-expression of dTCTP RNAi and dRheb in the ptc expression region showed the dTCTP RNAi phenotype, suggesting that dTCTP is epistatic to dRheb (Hsu, 2007).

Finally, whether activation of dS6k, a downstream effector of the insulin/TSC pathway, is dependent on dTCTP was tested. The level of dS6k activation was measured using a phospho-specific antibody (dS6k p-Thr 398). Extracts from act>dTCTPi larvae showed a significantly lower amount of activated dS6k compared with the controls, indicating that dTCTP is required for dS6k activation. Consistent with this, the eyeless (ey)>dTCTPi phenotype was dominantly enhanced by heterozygosity for a null mutation of dS6k (dS6k^l-1). Removing a copy of dS6k caused an approximately 20% further reduction of eye size in ey>dTCTPi animals. Taken together, these data support a model wherein dTCTP functions either downstream or in parallel to Tsc and dRheb, but upstream of dS6k (Hsu, 2007).

Epistatic analysis suggests that dTCTP may be a new component in the TSC pathway. Because TCTP structurally resembles a small GTPase regulator, Mss4, it is proposed that dTCTP might directly associate with dRheb and positively regulate its activity. To test this, co-immunoprecipitation experiments were performed. Flag-tagged dTCTP immunoprecipitated together with Myc-tagged dRheb in 293T cell extracts, suggesting that dTCTP and dRheb form a complex in vivo. Furthermore, in vitro pull-down assays demonstrated direct binding of glutathione S-transferase (GST)-dTCTP to maltose binding protein (MBP)-dRheb. Notably, dTCTP showed preferential binding to nucleotide-free dRheb, a property shared among guanine nucleotide exchange factors (GEFs). To test whether dTCTP has GEF activity for dRheb, in vitro GDP release experiments were performed. MBP-dRheb alone showed weak intrinsic GDP dissociation. In contrast, addition of GST-dTCTP stimulated the GTP/GDP exchange on dRheb rapidly even when low amounts of dTCTP were used. The GEF-like activity is specific, as dTCTP did not accelerate the exchange reaction on dRas1, the closest GTPase to dRheb at sequence level. Moreover, a glutamic acid to valine mutation in the putative GTPase binding groove of dTCTP (dTCTP^E12V) abolished this GEF activity, even at a high concentration. Because dTCTP^E12V still retained binding activity to dRheb, this residue seems to have a critical function in catalytic reactions rather than binding between the two proteins. To determine whether E12 is critical for the function of dTCTP in vivo, genetic rescue experiments were performed. Whereas wild-type dTCTP was able to rescue fully the dTCTP RNAi phenotype, mutant dTCTP^E12V failed to rescue the RNAi defects even though the mutant protein was expressed at a high level. Therefore, the conserved E12 residue of dTCTP is essential for its normal function in development (Hsu, 2007).

To test whether this in vitro GEF activity has a physiological relevance, the in vivo level of GTP bound to dRheb was determined in Drosophila S2 cells. S2 cells treated with dTCTP dsRNA or a control EGFP dsRNA were transfected with Myc-tagged dRheb. dTCTP-dsRNA-treated cells consistently showed a lower percentage of GTP-bound dRheb, suggesting that dTCTP is required for dRheb activation in vivo. These GEF-like properties displayed by dTCTP towards dRheb are particularly intriguing, as no Rheb GEFs have been reported. Further kinetic and structural analysis will help to elucidate whether dTCTP is a bona fide GEF enzyme for dRheb (Hsu, 2007).

Human TCTP (hTCTP) and dTCTP are roughly 50% identical in their protein sequence. dTCTP RNAi and mutant phenotypes can be rescued by expression of hTCTP. Furthermore, hTCTP interacts most strongly with the nucleotide-free hRheb and stimulates the GDP/GTP exchange of hRheb in vitro. These data suggest that the function of TCTP in the TSC pathway is likely to be conserved throughout evolution (Hsu, 2007).

dTCTP controls cell growth and proliferation by positively regulating dRheb activity. The data suggest that dTCTP may function as a GEF or a related regulatory factor to activate dRheb. Given the strong epistatic effect of dTCTP to dRheb, it is also possible that dTCTP may have additional roles in the TSC pathway, acting downstream of dRheb but upstream of S6k (Hsu, 2007).

TCTP has been implicated in a wide range of cancers. Nevertheless, overgrowth phenotypes as a result of dTCTP overexpression were not observed, suggesting that dTCTP is not sufficient to induce growth. Notably, reduction of TCTP levels is sufficient for suppression of malignancy in tumour reversion models. This study provides a possible explanation for this phenomenon. It will be intriguing to learn whether lowering the insulin/TSC signalling output can be a general mechanism for tumour reversion (Hsu, 2007).

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.