Tuberous Sclerosis Causing Mutants of the TSC2 Gene Product A€Ect

Tuberous sclerosis causing mutants of the TSC2 gene product aect proliferation and p27 expression

Thomas Soucek1,3, Margit Rosner1, Angelina Miloloza1, Marion Kubista1, Jeremy P Cheadle2, Julian R Sampson2 and Markus HengstschlaÈ ger*,1

1Obstetrics and Gynecology, University of Vienna, Prenatal Diagnosis and Therapy, WaÈhringer GuÈrtel 18-20, A-1090 Vienna, Austria; 2Institute of Medical Genetics, University Hospital of Wales, Health Park, Cardi, Wales, CF14 4XW, UK

The autosomal dominant disease tuberous sclerosis hamartomas, of the brain, skin, kidneys, heart and (TSC) is caused by mutations in either TSC1 on other organs. Common clinical features are facial chromosome 9q34, encoding hamartin, or TSC2 on angio®bromas, renal angiomyolipomas, hypopigmented chromosome 16p13.3, encoding tuberin. TSC is char- macules, cardiac rhabdomyomas, and cortical tubers acterized by hamartomas that occur in many organs of and subependymal glial nodules in the brain. The aected patients and these have been considered to likely greatest source of morbidity is the brain tumors, which result from defects in proliferation control. Although the cause seizures in 80 ± 90% of aected individuals, true biochemical functions of the two TSC proteins have mental retardation in about half of aected individuals, not been clari®ed, a series of independent investigations and behavioral abnormalities (mostly autism) in over demonstrated that modulated hamartin or tuberin half of aected individuals (Gomez et al., 1999). Two expression cause deregulation of proliferation/cell cycle TSC genes have been identi®ed, TSC1 on chromosome in human, rodent and Drosophila cells. In support of 9q34 (The TSC1 Consortium, 1997) encoding a tuberin acting as a tumor suppressor, ectopic over- 130 kDa protein named hamartin and TSC2 (The expression of TSC2 has been shown to decrease European Chromosome 16 Tuberous Sclerosis Con- proliferation rates of mammalian cells. Furthermore, sortium, 1993) on chromosome 16p13.3 encoding a overexpression of TSC2 has been demonstrated to 200 kDa protein named tuberin. Published TSC trigger upregulation of the cyclin-dependent kinase mutations include large deletions/rearrangements and inhibitor p27. We report that three dierent naturally missense mutations for TSC2 and insertions, deletions, occurring and TSC causing mutations within the TSC2 nonsense and splicing mutations in both genes (recently gene elliminate neither the anti-proliferative capacity of reviewed in Young and Povey, 1998; Cheadle et al., tuberin nor tuberin's eects on p27 expression. For the 2000). Hamartin and tuberin can associate physically in ®rst time these data provide strong evidence that vivo (Van Slegtenhorst et al., 1998; Plank et al., 1998; deregulation of proliferation and/or upregulation of p27 Nellist et al., 1999). Tuberin has been reported to are not likely to be the primary/only mechanisms of possess GTPase-accelerating protein (GAP) activity for hamartoma development in TSC. These results demand the small molecular weight GTPase Rap1a (Wienecke reassessment of previous hypotheses of the pathogenesis et al., 1995) and for Rab5 aecting the control of of TSC. Oncogene (2001) 20, 4904 ± 4909. endocytosis (Xiao et al., 1997), to be involved in transcriptional regulation (Tsuchiya et al., 1996; Henry Keywords: tuberous sclerosis; TSC2; tuberin; prolifera- et al., 1998), and to aect the control of neuronal tion; p27 dierentiation (Soucek et al., 1998a). Hamartin has been demonstrated to bind ezrin/radixin/moesin protein family members and to activate the Rho GTPase and actin ®bre assembly (Lamb et al., 2000). Tuberous sclerosis (TSC) is an autosomal dominant Furthermore, both tuberin and hamartin are involved phakomatosis with an estimated prevalence of about 1 in the regulation of cellular proliferation, the only in 6000 newborns. It is characterized by the occurence function so far demonstrated for both proteins. of mental retardation, epilepsy, and by the develop- Overexpression and antisense-experiments in mamma- ment of a spectrum of tumor-like growths, named lian cells revealed that high levels of tuberin have negative eects on cell proliferation, whereas down- regulation/loss of tuberin stimulates proliferation (Jin et al., 1996; Orimoto et al., 1996; Soucek et al., 1997, *Correspondence: M HengstschlaÈ ger; 1998b; Pajak et al., 1997; Pasumarthi et al., 2000). E-mail: [email protected] Gigas,aDrosophila homolog of TSC2, was reported to 3 Current address: The Salk Institute for Biological Studies, Cellular be involved in cell cycle control (Ito and Rubin, 1999). Neurobiology La Jolla, CA 92037-1099, USA Received 19 March 2001; revised 26 April 2001; accepted 10 May High ectopic levels of hamartin have also been shown 2001 to negatively regulate cellular proliferation (Miloloza et Tuberous sclerosis causing TSC2 mutants maintain anti-proliferative capacity T Soucek et al 4905 al., 2000; Benvenuto et al., 2000). However, mutations was 0.0078 for the dierences in S phase cells in the TSC genes that are causally involved in the P=0.0117. Surprisingly, overexpression of the three development of TSC in patients have not been dierent mutant clones also triggered eects in investigated in relation to eects of hamartin or logarithmically growing cells (Figure 1B,C). In no case tuberin on cellular proliferation. Answering this did the control cells show an increase in the G1 question is required to clarify the relevance of the fraction or a decline in the S phase fraction relative to functions of the proteins for the development of the overexpression of TSC2 cDNAs (the P values for the disease. dierences in G0/G1 cells compared to the control Tuberin has been proposed to harbor a leucin zipper experiments were 0.0039, 0.0039, 0.0039 for the region at amino acids 75 ± 107, coiled-coil structures at mutants 6 2/25, 6 3/5 and 6 5/2, respectively. For the amino acids 346 ± 371 and 1008 ± 1021, a rap1GAP dierences in S phase cells P(62/25)=0.0273, P(63/ homology domain spanning amino acids 1499 ± 1689 5)=0.0117, P(65/2)=0.0078). Assuming that the main and a rabaptin-binding domain at amino acids 1668 ± targets of overexpressed tuberin are the G0/G1 phase 1726 (The European Chromosome 16 Tuberous and the S phase of the mammalian cell cycle one would Sclerosis Consortium, 1993; Xiao et al., 1997; expect that in cell populations harboring already higher Maheshwar et al., 1997). As mentioned, tuberin possess numbers of G0/G1 cells the potential of tuberin to GAP activity for the GTPase Rap1a (Wienecke et al., increase the number of G0/G1 cells would be lower. 1995). Furthermore, mutation analysis of the rap1GAP Six independent transfections of Rat-1 cells grown related domain in tuberous sclerosis patients led to the under density-restricted growth conditions harboring characterization of disease associated mutations in- about 53% G0/G1 cells demonstrated that TSC2 and cluding missense mutations that were shown to occur the three mutant clones induce weaker eects on the as de novo mutations in sporadic TSC cases (Mahesh- G0/G1-S phase distribution (data not shown). war et al., 1997). Accordingly, we considered missense To exclude the possibility that overexpressed TSC2 mutations within the GAP homologous domain to be mutants mediate their anti-proliferative eects via optimal to address the question whether naturally overactivation of the endogenous wild-type tuberin occurring mutations would disrupt the anti-prolifera- we analysed tuberin-negative ®broblasts. Rat ®bro- tive capacity of tuberin. Three of such reported blasts derived from Eker rat embryos homozygous for mutations, N1643K, N1651S and N1681K, were the Eker mutation (a rearrangement in the rat homolog introduced into full-length wild-type human TSC2 of TSC2; Yeung et al., 1994; Kobayashi et al., 1995) cDNA by site directed mutagenesis, cloned into the were transiently transfected as described above for Rat- mammalian expression vector pSG5 and designated 6 1 cells. As shown by immunoblot analysis these cells do 2/25, 6 3/5 and 6 5/2, respectively. We transiently not express endogenous tuberin. Overexpression levels transfected Rat-1 ®broblasts with wild-type TSC2 in of wild-type TSC2 and of the three mutants were pSG5, with the three missense mutant clones or with comparable (Figure 2A). Investigation of DNA the empty expression vector as a negative control. Two distributions on the ¯ow cytometer revealed that days after transfection immunoblot analyses revealed transient transfection of wild-type TSC2 triggers an that the wild-type and the dierent mutant forms of increase in the proportion of G0/G1 cells from TSC2 were overexpressed to similar levels. For all 41.6+9.4% to 54.9+5.1%, accompanied by a decrease presented ®gures Western blots were chosen for which of S phase cells from 50.6+11.5% to 36.1+3.7%. Very the Ponceau-S staining revealed that the amount of similar eects could be detected after overexpression of protein loaded on each lane was about equal. This the missense TSC2 mutants as concluded from 10 together with analysing and presenting a non-speci®c independent experiments (Figure 2B) (for the dier- band was chosen to check loading (Figure 1A). After ences in G0/G1 phase cells compared to the control co-transfection with the pCMV-EGFP-Spectrin vector, P(wild-type)=0.0029, P(6 2/25)=0.001, P(6 3/5)= that expresses a membrane-associated form of the 0.0029, P(6 5/2)=0.001; for the dierences in the green ¯uorescence protein (GFP) under the control of amount of S phase cells P(wild-type)=0.001, P(6 the CMV promoter, the green emission enables 2/25)=0.001, P(6 3/5)=0.0033, P(6 5/2)=0.0049). transfected from untransfected cells to be separated Again, in no case did the control cells show an on a ¯ow cytometer. Cell cycle distributions of the increase in the G1 fraction or a decline in the S phase transfected cells can be investigated after DNA fraction relative to overexpression of TSC2 cDNAs. staining. Analysing eight independent experiments, we Twelve independent transfections of tuberin negative found that transient overexpression of wild-type cells under density-restricted growth conditions demon- tuberin in logarithmically growing Rat-1 cells lead to strated that all analysed TSC2 cDNAs mediate weaker an increase of the proportion of G0/G1 cells (from eects on the G0/G1-S phase distribution (e.g. for 40.0+7.4% to 50.1+7.7%), accompanied by a decline wild-type TSC2: from 69.2+7.3 to 77.8+8.6% G0/G1 in S phase cells (from 48.4+10.1% to 35.4+6.1%) cells and from 27.8+7.8 to 18.7+8.5% S phase cells) (Figure 1B,C). To evaluate the signi®cance of these (data not shown). Although the immortalized Rat-1 dierences statistical analysis using the Wilcoxon cell line, which is tuberin positive but harbors other Signed Rank Test was performed. In this test P values genetic alterations responsible for immortalization, and lower than 0.04 re¯ect statistical signi®cance. The P the tuberin negative Eker rat cells have comparable value for the dierences in the amount of G0/G1 cells DNA distributions when transfected with the empty

Oncogene Tuberous sclerosis causing TSC2 mutants maintain anti-proliferative capacity T Soucek et al 4906 control vector, the eects of wild-type TSC2 and the three TSC2 mutants appeared to be more pronounced in Eker TSC27/7cells (compare Figures 1 and 2). This could be due to the totally dierent genetic background of these two cell lines. But one could also speculate that ectopic tuberin and the mutants can mediate stronger eects in cells depleted of normal cellular tuberin functions. In conclusion, the eects of overexpression of TSC2 cDNAs on the G0/G1-S phase distribution of logarithmically cycling cells are con-

Figure 1 Transient overexpression of TSC2 cDNAs in Rat-1 cells. Transfections were performed using the empty mammalian expression vector pSG5, pSG5 harboring full-length wild-type Figure 2 Transient overexpression of TSC2 cDNAs in tuberin- human TSC2, pSG5 harboring the TSC2 GAP-domain missense negative ®broblasts. The tuberin negative cells used in this study mutant 6 2/25 (C to G at nucleotide 4947=Asn to Lys at amino were rat ®broblasts derived from Eker rat embryos homozygous acid 1643), the TSC2 missense mutant 6 3/5 (A to G at nucleotide for the Eker mutation (Yeung et al., 1994; Kobayashi et al., 1995; 4970=Asn to Ser at amino acid 1651), and the TSC2 missense see also Soucek et al., 1998b). Mammalian expression vectors, mutant 6 5/2 (C to G at nucleotide 5061=Asn to Lys at amino empty as negative control or harboring either full-length wild-type acid 1681). To analyse transient transfections on the ¯ow TSC2 cDNA or the missense mutant TSC2 cDNA clones 6 2/25, cytometer cells were co-transfected with the GFP-spectrin- 6 3/5, or 6 5/2, all derived from tuberous sclerosis patients, were expression vector described in (Kalejta et al., 1997). Cell co-transfected with a GFP expression vector. (A) Immunoblot transfections were performed either using the LipofectAMINE analysis of tuberin expression in the transfected cells. NB, non- reagent obtained from Life Technologies (Gibco BRL, Lofer, speci®c bands, which are presented to visualize the amounts of Austria) following the transfection protocol provided by the protein loaded. (B) Logarithmically growing cells were transfected manufacturer or following the standard protocol for calcium and GFP-positive cells were cyto¯uorometrically analysed for phosphate-mediated transfection of adherent cells (Sambrook et distribution of cells in dierent cell cycle phases (means+stan- al., 1989). (A) Immunoblot analysis of tuberin expression in Rat-1 dard deviations of 10 independent experiments are shown). For transfectants. NB, non-speci®c bands, which are presented to details of the experimental procedures see Figure 1 visualize the amounts of protein loaded. Protein extraction and Western blot analyses were performed as described earlier (Miloloza et al., 2000). Blots were stained with Ponceau-S to visualize the loaded protein. Immunodetection was performed 1 ml of staining solution (0.25 mg/ml propidium iodide, 0.05 mg/ using the anti-tuberin antibody 5063 (Wienecke et al., 1995). ml RNase, 0.1% Triton X-100 in citrate buer, pH 7.8). Green Signals were detected using the enhanced chemiluminescence ¯uorescent protein (GFP)-positive cells were selectively gated and method (Amersham, Little Chalfont, UK). (B) FACS-DNA DNA distribution was analysed on a Beckton Dickinson pro®les of GFP-positive Rat-1 cells from a representative FACScan (Beckton Dickinson, San Jose, CA). (C) Logarithmi- transfection experiment performed as described above. For cally growing cells were transfected and GFP-positive cells were cyto¯uorometric analyses, cells were harvested by trypsinization cyto¯uorometrically analysed for distribution of cells in dierent and ®xed by rapid submersion in ice-cold 85% ethanol. After 1 h cell cycle phases (means+standard deviations of eight indepen- ®xation at 7208C, cells were pelleted and DNA was stained in dent experiments are shown)

Oncogene Tuberous sclerosis causing TSC2 mutants maintain anti-proliferative capacity T Soucek et al 4907 sistent with the previously reported eects seen after modulation of TSC gene expression (Soucek et al., 1997, 1998b; Miloloza et al., 2000). Furthermore, these eects are comparable to those seen after overexpression of other tumor suppressor genes involved in familial cancer syndromes. For example, 3 days of transient overexpression of BRCA1 in logarithmically growing cells triggers an increase of G0/G1 cells from about 55 to 68% accompanied by a decrease of S phase cells from about 18 to 10% (Aprelikova et al., 1999). Modulated expression of cyclin E, that is known to directly regulate the cell cycle machinery, also mediates similar eects: 2 days of transient overexpression of cyclin E in logarithmically growing Rat-1 cells induces a 16% decrease of G0/G1 cells accompanied by a 13% increase of S phase cells (Ohtsubo and Roberts 1993). We next sought con®rmation of these ®ndings using Figure 3 Selection for ectopic overexpression of TSC2 cDNAs a dierent approach. The anti-proliferative eects of in Rat-1 cells. Mammalian expression vectors, empty as negative the TSC2 cDNAs were therefore tested by selection for control or harboring either full length wild-type TSC2 cDNA or the missense mutant TSC2 cDNA clones 6 3/5 or 6 5/2, were ectopic overexpression followed by cell counting. This transfected into Rat-1 cells. Selection for transfected cells was approach has previously demonstrated that high levels enabled by co-transfection with the G418-selectable vector of ectopic TSC1 and TSC2 attenuate cell proliferation/ pcDNA3. Selection was started 24 h after transfection. During cell cycling (Jin et al., 1996; Orimoto et al., 1996; the ®rst 2 days of selection the G418 concentration was 700 mg/ Soucek et al., 1998b; Miloloza et al., 2000; Benvenuto ml, thereafter the G418 concentration was set at 1500 mg/ml. The selected cell pools were then analysed for cell doubling rates by et al., 2000). We transfected Rat-1 cells with the empty performing cell counting using a hemocytometer. Cell numbers expression vector, with wild-type TSC2 and with the determined at day 8 of such experiments are given as a percentage GAP-domain missense TSC2 mutant clones 6 3/5 and relative to the highest value, set as 100% (means+standard 6 5/2 and co-transfected with the empty pcDNA3 deviations of two independent experiments are shown) vector to enable G418 selection for transfected cells. Levels of overexpression of TSC2 proteins were always comparable (compare also Figure 1A). Starting with equal cell numbers the cell pools were further kept under selection and cell numbers were determined at day 8 by counting using a hemocytometer. We found Rat-1 cells expressing wild-type TSC2 to grow signi®cantly more slowly than those containing only the empty vector. In agreement with the data described above, the TSC associated mutations did not elliminate the anti-proliferative eects of tuberin (Figure 3). One level of regulation of cyclin-dependent kinases, which are key regulators of the mammalian cell cycle, is provided by inhibitors. Two families of these inhibitors have been described, those that interact with a wide range of cyclin/Cdk complexes, including p21, p27 and p57, and those that only inhibit Cdk4 and Cdk6, including p15, p16, p18 and p19 (reviewed in e.p. HengstschlaÈ ger et al., 1999). It has been shown that TSC2 aects the expression of the cyclin- dependent kinase inhibitor p27 (Soucek et al., 1998b; Pasumarthi et al., 2000). Ectopic overexpression of TSC2 in tuberin-positive and in tuberin-negative cells triggers upregulation of endogenous p27 protein expression (Soucek et al., 1998b). Western blot Figure 4 Overexpression of the naturally occurring TSC2 analyses after transient transfections revealed that mutants triggers upregulation of p27 protein expression. Empty introducing TSC associated mutations into wild type expression vector as control, wild-type TSC2 and the TSC2 TSC2 cDNA does not disrupt the capacity to mutants were transiently overexpressed in Rat-1 ®broblasts (A) upregulate p27 expression in Rat1 cells as well as in and in tuberin-negative cells (B). Two days after transfection p27 Eker rat derived tuberin negative cells (Figure 4). protein expression was studied by Western blot analyses using the anti-p27 antibody C-19 (Santa Cruz Biotechnology). NB, non- In conclusion, we demonstrate that overexpression of speci®c bands, which are presented to visualize the amounts of three independent missense mutations within TSC2,all protein loaded

Oncogene Tuberous sclerosis causing TSC2 mutants maintain anti-proliferative capacity T Soucek et al 4908 proven to be disease causing, neither eliminate the anti- for the ®rst time, that tuberin is phosphorylated at proliferative eects of tuberin in speci®c mammalian serine and tyrosine residues. This process is altered by cells nor does it disrupt tuberin's capacity to induce the Y1571H mutation, what is suggested to be the p27 expression. Data were obtained using two dierent reason for the signi®cant reduction in anti-proliferative experimental approaches and analysing tuberin positive capacity (Aicher et al., 2001). Whether the mutations and tuberin negative cells. Earlier reported overexpres- analysed in the here presented study alter tuberin's sion experiments have demonstrated that other muta- phosphorylation status is currently not tested. It was tions within TSC genes can disrupt their anti- earlier reported that a TSC2 cDNA clone encoding proliferative capacity (Miloloza et al., 2000). Accord- only the C-terminus of tuberin (amino-acids 1049 ± ingly, overexpression experiments as performed in this 1809) was capable of reducing proliferation measured study are suitable to test mutations for their capacity in a colony formation assay and in vivo tumorigenicity to disrupt the anti-proliferative eects of tuberin. when overexpressed in Eker rat tumor cells (Jin et al., However, a diculty is that in vitro assays always 1996). These data suggest that only speci®c regions of represent a only simpli®ed model for proliferation tuberin are necessary for its eects on proliferation. It control in vivo. Therefore, we recognize, that negative will be of great interest to investigate other regions of in vitro data can not totally rule out a role for tuberin to enable a better understanding of how this proliferation control in the TSC patient. It is important protein in¯uences cell proliferation control. However, to analyse cells or cell lines derived from hamartomas since all mutations analysed in this study were derived of patients. Since the TSC2 mutations investigated in from TSC patients and were proven to be disease the presented study all map within the GAP homo- causing, our data provide strong evidence that neither logous domain our data suggest that this region is not deregulation of proliferation nor deregulated p27 prominently involved in the regulation of proliferation. expression represent the primary/only mechanism of During the review process of this manuscript another hamartoma development in TSC. report described the eects of overexpression of another natural occurring GAP-mutant of TSC2 in proliferation assays. Only 24 h after transient overexpression of a Y1571H mutant, the number of Cos1 Acknowledgments cells and protein concentration (compaired to control The authors wish to thank AJ Beavis, J DeClue, M Eilers vector transfections) was reduced 20 and 27%, and R Yeung for the generous gift of cell lines and respectively. Although under the conditions of these reagents, N Maderner for help with statistics, E Hengsts- experiments wild-type TSC2 induced a signi®cantly chlaÈ ger-Ottnad for helpful discussion and M Schreiber for stronger eect (about 50%), these data also demon- critically reading the manuscript. Work in M Hengsts- chlaÈ ger's laboratory is funded by the Austrian Federal strate that the analysed natural occurring mutation can Ministry of Science and Transport (GZ 70.037/2-Pr/4/98). not eliminate tuberin's capacity to negatively aect cell M Kubista is supported by the Austrian Ludwig Boltz- proliferation. In this very important report, the authors mann Institute for Clinical and Experimental Oncology. JR found phosphorylation to be an additional level of Sampson and JP Cheadle are supported by the Tuberous regulation of tuberin's functions. They demonstrated Sclerosis Alliance and the Tuberous Sclerosis Association.

References

Aicher LD, Campbell JS and Yeung RS. (2001). J. Biol. Kalejta RF, Shenk T and Beavis AJ. (1997). Cytometry, 29, Chem., in press. 286 ± 291. AprelikovaON,FangBS,MeissnerEG,CotterS,Campbell Kobayashi T, Hirayama Y, Kobayashi E, Kubo Y and Hino M, Kuthiala A, Bessho M, Jensen RA and Liu ET. (1999). O. (1995). Nature Genet., 9, 70 ± 74. Proc. Natl. Acad. Sci. USA, 96, 11866 ± 11871. LambRF,RoyC,DiefenbachTJ,VintersHV,Johnson BenvenutoG,LiS,BrownSJ,BravermanR,VassWC, MW, Jay DG and Hall A. (2000). Nature Cell Biol., 2, Cheadle JP, Halley DJJ, Sampson JR, Wienecke R and 281 ± 287. DeClue JE. (2000). Oncogene, 19, 6306 ± 6316. Maheshwar MM, Cheadle JP, Jones AC, Myring J, Fryer Cheadle JP, Reeve MP, Sampson JR and Kwiatkowski DJ. AE, Harris PC and Sampson JR. (1997). Hum. Mol. (2000). Hum. Genet., 107, 97 ± 114. Genet., 6, 1991 ± 1996. Gomez MR, Sampson JR and Whittemore VH. (1999). Miloloza A, Rosner M, Nellist M, Halley D, Bernaschek G Tuberous Sclerosis Complex, 3rd edn. Oxford University and HengstschlaÈ ger M. (2000). Hum. Mol. Genet., 9, Press, New York, NY. 1721 ± 1727. HengstschlaÈ ger M, Braun K, Soucek T, Miloloza A and Nellist M, Van Slegtenhorst MA, Goedbloed M, Ouweland HengstschlaÈ ger-Ottnad E. (1999). Mutat. Res., 436, 1±9. AMW. van den, Halley DJ and Sluijs P. van der. (1999). J. Henry KW, Yuan X, Koszewski NJ, Onda DJ, Kwiatkowski Biol. Chem., 275, 35647 ± 35652. DJ and Noonan DJ. (1998). J. Biol. Chem., 273, 20535 ± Ohtsubo M and Roberts JM. (1993). Science, 259, 1908 ± 20539. 1912. Ito N and Rubin GM. (1999). Cell, 96, 529 ± 539. Orimoto K, Tsuchiya H, Kobayashi T, Matsuda T and Hino JinF,WieneckeR,XiaoG-H,MaizeJC,DeClueJEand O. (1996). Biochem. Biophys. Res. Com., 219, 70 ± 75. Yeung RS. (1996). Proc. Natl. Acad. Sci. USA, 93, 9154 ± Pajak L, Jin F, Xiao GH, Soonpaa MH, Field LJ and Yeung 9159. RS. (1997). Am. J. Physiol., 273, 1619 ± 1627.

Oncogene Tuberous sclerosis causing TSC2 mutants maintain anti-proliferative capacity T Soucek et al 4909 Pasumarthi KBS, Nakajima H, Nakaji HO, Jing S and Field Tsuchiya H, Orimoto K, Kobayashi T and Hino O. (1996). LJ. (2000). Circ. Res., 86, 1069 ± 1077. Cancer Res., 56, 429 ± 433. Plank TL, Yeung RS and Henske EP. (1998). Cancer Res., Van Slegtenhorst M, Nellist M, Nagelkerken B, Cheadle J, 58, 4766 ± 4770. Snell R, Ouweland A, van den, Reuser A, Sampson J, Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular Halley D and Sluijs P. van der. (1998). Hum. Mol. Genet., cloning-a laboratory manual. Cold Spring Harbor Labora- 7, 1053 ± 1057. tory Press, USA. Wienecke R, KoÈ nig A and DeClue JE. (1995). J. Biol. Chem., Soucek T, Pusch O, Wienecke R, DeClue JE and Hengsts- 270, 16409 ± 16414. chlaÈ ger M. (1997). J. Biol. Chem., 272, 29301 ± 29308. Xiao G-H, Shoarinejad F, Jin F, Golemis EA and Yeung RS. Soucek T, HoÈ lzlG,BernaschekGandHengstschlaÈ ger M. (1997). J. Biol. Chem., 272, 6097 ± 6100. (1998a). Oncogene, 16, 2197 ± 2204. YeungRS,XiaoG-H,JinF,LeeW-C,TestaJRand Soucek T, Yeung RS and HengstschlaÈ ger M. (1998b). Proc. Knudson AG. (1994). Proc.Natl.Acad.Sci.USA,91, Natl. Acad. Sci. USA, 95, 15653 ± 15658. 11413 ± 11416. The European Chromosome 16 Tuberous Sclerosis Con- Young J and Povey S. (1998). Mol. Med. Today, July, 313 ± sortium. (1993). Cell, 75, 1305 ± 1315. 319. The TSC1 Consortium. (1997). Science, 277, 805 ± 808.

Oncogene