Regulation of tuberous sclerosis complex (TSC) function by proteins

Transcription

1 TSC Genes Novel Players in the Growth Regulation Network 587 Regulation of tuberous sclerosis complex (TSC) function by proteins M. Nellist 1,M.A. Goedbloed and D.J.J. Halley Department of Clinical Genetics, Erasmus MC, Dr. Molewaterplein 50, 3015 GE Rotterdam, The Netherlands Abstract Tuberous sclerosis complex (TSC) is a genetic disorder characterized by seizures, mental disability, renal dysfunction and dermatological abnormalities. The disease is caused by inactivation of either hamartin or tuberin, the products of the TSC1 and TSC2 tumour-suppressor genes. Hamartin and tuberin form a complex and antagonise phosphoinositide 3-kinase/protein kinase B/target of rapamycin signal transduction by inhibiting p70 S6 kinase, an activator of translation, and activating 4E-binding protein 1, an inhibitor of translation initiation. Phosphorylation-dependent binding between tuberin and members of the protein family indicates how the tuberin hamartin complex may interact with upstream and downstream effectors, and suggests how phosphorylation-dependent regulation of the complex may be controlled. Introduction Tuberous sclerosis complex (TSC) is an autosomal dominant disorder caused by inactivating mutations to either the TSC1 gene on chromosome 9q34 or the TSC2 gene on chromosome 16p13 [1]. The cloning of the TSC1 and TSC2 genes, and the identification of the Eker rat [2,3] and Drosophila gigas [4] models of the disease has facilitated progress towards understanding the molecular pathology of TSC. TSC1 encodes a 130 kda protein called hamartin, and TSC2 encodes a 200 kda protein called tuberin [5,6]. Hamartin and tuberin do not share any amino acid sequence identity. Alternative splicing of the TSC2 gene combined with multiple phosphorylation events explains the existence of multiple tuberin isoforms [7]. Tuberin and hamartin form a complex and it has been suggested that mutations to either protein inactivate the complex, causing many of the lesions seen in TSC patients [8,9]. Experiments in Drosophila demonstrated that tuberin and hamartin inhibit cell growth downstream of phosphoinositide 3-kinase (PI3K) and protein kinase B (PKB; Akt), but upstream of p70 S6 kinase (S6K) [10 12]. Upon stimulation of the PI3K/PKB/target of rapamycin (TOR) signalling pathway, tuberin is phosphorylated by PKB [13]. This inactivates the tuberin hamartin complex, allowing phosphorylation of S6K and 4E-binding protein 1 (4E-BP1) and increased cell growth [14 16]. Activated phosphorylated S6K phosphorylates the ribosomal protein S6, stimulating translation, while phosphorylation of 4E-BP1 prevents 4E-BP1 from binding the elongation factor elf4e and thereby promotes the initiation of translation [17]. Phosphorylation of S6K and 4E-BP1 is Key words: protein, hamartin, phosphoinositide 3-kinase (PI3K)/protein linase B (PKB)/target of rapamycin (TOR) signalling, tuberin, tuberous sclerosis complex (TSC). Abbreviations used: TSC, tuberous sclerosis complex; PI3K, phosphoinositide 3-kinase; S6K, p70 S6 kinase; 4E-BP1, 4E-binding protein 1; PKB, protein kinase B; TOR, target of rapamycin; FKHRL1, forkhead in rhabdosarcoma L1. 1 To whom correspondence should be addressed ( m.nellist@erasmusmc.nl). aconsequence of the activation of the protein kinase TOR. Initial studies indicated that the tuberin hamartin complex antagonizes TOR signalling, possibly through a direct interaction [15,18,19]. However, other work suggests that the tuberin hamartin complex inhibits S6K independently of TOR [20]. The exact mechanism for tuberin hamartin-dependent inhibition of S6K activity has not yet been determined. Therefore, whether the tuberin hamartin complex acts through TOR, or with other downstream effectors to regulate cell growth independently of TOR, the identification of proteins that interact with tuberin and hamartin will further our understanding of cell-growth control mechanisms, as well as providing insight into the molecular pathology of TSC. Tuberin interacts with proteins In the last year, four studies have been published describing the interactions between tuberin and members of the protein family [21 24]. Yeast two-hybrid screens using tuberin as bait identified multiple clones, encoding β and ζ [21,24]. Two-hybrid analysis of tuberin and β mutants provided confirmation that the interactions were specific. In addition, the hybridization of tuberin-derived peptides to an array of known protein protein-binding domains identified an interaction between a phosphorylated tuberin peptide and both γ and ζ [23]. The tuberin interactions were confirmed in co-immunoprecipitation experiments and by in vitro assays using recombinant tuberin and/or proteins [21 24] proteins are small (28 33 kda) acidic polypeptides that are expressed in all eukaryotic cells and show a considerable degree of conservation of both structure and function [25]. In vertebrates, the protein family consists of nine closely related isoforms, α, β, γ, δ, ε, ζ, η, σ and τ [26]. Although the screening experiments detected binding between tuberin and ζ, γ and β, inin vitro

2 588 Biochemical Society Transactions (2003) Volume 31, part 3 assays tuberin interacted with all the isoforms tested [21,23] proteinsconsistofnineanti-parallel α-helices (A I) that assume a cup-shaped conformation [27]. Conserved hydrophobic and polar residues in helices A, C and D mediate the formation of stable homo- and heterodimers, whereas polar and charged residues in helices E, G and I form a characteristic phosphopeptide-binding pocket. This pocket has a high affinity for two consensus phosphopeptide sequences, Arg-Ser-Xaa-pSer/Thr-Xaa-Pro (mode I; where p indicates a phosphorylated residue) and Arg-Xaa-Xaa-Xaa-pSer-/Thr-Xaa-Pro (mode II) which are often present in the proteins bound by [28]. The binding between proteins and specific phosphoserine/ threonine sequences allows putative binding partners and their specific recognition domains to be identified using Scansite ( [29] proteins are members of a group of proteins that specifically interact with phosphorylated proteins, facilitating the phosphorylation-dependent control of protein activity [30] proteins are stiff molecules [27,31]. Binding between a protein and a Arg-Ser-Xaa-pSer/Thr-Xaa- Pro or Arg-Xaa-Xaa-Xaa-pSer-/Thr-Xaa-Pro recognition site results in deformation of the target protein and changes in protein activity [27]. For example, binding between phosphorylated BAD, a death-promoting Bcl-2 family protein, and proteins prevents BAD from interacting with Bcl-2 and results in the inhibition of apoptosis [25] binding can therefore disrupt protein complexes. However, proteins can also promote protein protein interactions, acting as phosphorylation-dependent bridge molecules between, for example, Raf and the Raf-interacting protein Bcr [32]. In addition, a general function of binding is to inhibit nuclear import [30]. For example, the transcription factor forkhead in rhabdosarcoma L1 (FKHRL1) is phosphorylated by PKB and is then bound by proteins which prevent it from crossing the nuclear membrane, activating transcription [33]. Finally, binding can alter the activity of a target protein, either by direct up- or down-regulation of activity, or by targeting the protein for dephosphorylation and/or proteolysis [25]. The different functions of proteins are indicated in Figure proteins regulate signal transduction [34] and have been shown to be downstream effectors of TOR signalling [35], conferring rapamycin resistance on yeast cells by helping retain TOR-dependent transcription factors in the cytoplasm [36]. The interaction between tuberin and proteins may provide additional insight into how the tuberin hamartin complex functions in PI3K/PKB/TOR signal transduction and the control of cell growth. Tuberin binding is phosphorylation-dependent The recognition motif is similar to the consensus motifs required for protein phosphorylation by the protein Figure protein functions Schematic diagram to illustrate the different roles of the protein family in the cell, based on [25]. (A) Disruption of the BAD Bcl2 protein complex [25]; (B) promotion of Raf Bcr binding [32]; (C) retention of FKHRL1 in the cytoplasm [33]; (D) activation of tryptophan hydroxylase (TPH) [43] and inactivation of Ask-1 kinase [44,45]. kinases A, B and C. It has been suggested that these motifs have co-evolved so that phosphorylated proteins will subsequently interact with proteins [26]. As shown in Table 1, tuberin contains both putative (mode I)- recognition motifs and PKB-phosphorylation sites. Motifs surrounding tuberin amino acids Ser-939, Ser-981 and Ser correspond to both PKB- and recognition sites. Phosphorylation of tuberin has been described [37,38] and, more specifically, tuberin amino acids Ser-939 and Thr-1462 have been shown to be phosphorylated directly by PKB [13]. The domain hybridization experiments suggested that Ser- 939 phosphorylation may be necessary for tuberin binding [23]. Consistent with these findings, tuberin ζ binding was inhibited by over-expression of a dominant negative isoform of PKB [21]. In addition, the tuberin interaction was inhibited by phosphatase activity [21,22] and phosphatase inhibitors stabilized tuberin binding [21]. These experiments demonstrated that the

3 TSC Genes Novel Players in the Growth Regulation Network 589 Table 1 Conservation of putative sites of PKB-dependent phosphorylation and recognition motifs in human, rat and mouse tuberin The best eight predicted recognition motifs and best eight predicted sites of PKB phosphorylation in human tuberin were identified using Scansite ( [29] and compared with the corresponding sequences from rat and mouse. The amino acid position of the phosphorylated serine (Ser) or threonine (Thr) is indicated and the figures in parentheses indicate the percentile ranking of the candidate motifs calculated with respect to all the potential motifs in a protein database. Sites predicted to be phosphorylated by PKB that also correspond to recognition motifs are highlighted in bold. Note that only the best predicted human sites are presented. The Scansite ranking of the same sites in rat and mouse tuberin is, in each case, slightly different. For example, no recognition motif is predicted in the region of mouse tuberin corresponding to Ser-1254 in the human and rat proteins. Tuberin amino acids are numbered according to the appropriate full-length protein, as submitted to GenBank r.forthisreason there are consistent differences with the numbering used in this review and in, for example, [22] and [24]. Phosphorylated Ser or Thr and percentile ranking of candidate motifs Predicted site Human tuberin Rat tuberin Mouse tuberin Ser-540 (0.392) Ser-540 (0.392) Ser-540 (0.392) Thr-603 (0.149) Ser-603 (0.985) Ser-603 (0.985) PKB Ser-683 (0.365) /PKB Ser-939 (0.392/0.002) Ser-939 (0.392/0.002) Ser-939 (0.392/0.002) /PKB Ser-981 (0.527/0.030) Ser-981 (0.527/0.030) Ser-981 (0.527/0.030) PKB Ser-1130 (0.026) Ser-1130 (0.026) Ser-1130 (0.026) PKB Ser-1132 (0.166) Ser-1132 (0.166) Ser-1132 (0.166) Ser-1254 (0.297) Ser-1254 (0.297) /PKB Ser-1338 (0.297/0.244) Ser-1341 (0.214/0.617) Ser-1340 (0.149/0.617) Ser-1387 (0.297) Ser-1389 (0.297) Ser-1388 (0.297) PKB Thr-1462 (0.006) Thr-1466 (0.006) Thr-1465 (0.006) Ser-1730 (0.297) Ser-1732 (0.297) Ser-1738 (0.297) PKB Ser-1798 (0.283) Ser-1800 (0.218) Ser-1805 (0.218) tuberin interaction is dependent on tuberin phosphorylation, possibly by PKB. Interaction between proteins and the tuberin hamartin complex Tuberin interacts directly with hamartin to form a functional complex [8 12]. Detection of a ternary complex of tuberin, hamartin and suggests that the tuberin interaction is compatible with tuberin hamartin binding and that proteins interact with the tuberin hamartin complex [21,24]. In the absence of tuberin, hamartin is insoluble and unstable [39], making it difficult to test whether hamartin binds proteins directly in either co-immunoprecipitation experiments or pull-down assays [21,24]. However, there was no evidence for a direct interaction between hamartin and proteins in yeast two-hybrid experiments [21,24], making it most likely that hamartin and proteins are indirectly associated through the interactions of both proteins with tuberin. Tuberin phosphorylation by PKB promotes tuberin binding [23] but may disrupt the tuberin hamartin complex [14,15]. How is it possible, therefore, that tuberin and tuberin hamartin binding are compatible? Clearly, additional work is required to reconcile these somewhat contradictory findings. One possibility is that phosphorylation of tuberin by PKB is not sufficient to disrupt the tuberin hamartin complex [13]. Alternatively, proteins may bind tuberin phosphorylated independently of PKB. It will be interesting to determine how the tuberin interaction influences the tuberin hamartin complex. Are proteins necessary to disrupt the tuberin hamartin complex, or do they hold the complex together, antagonizing phosphorylation-dependent disruption of tuberin hamartin binding? To characterize the effect of the tuberin interaction on the tuberin hamartin complex in more detail, the recognition sites on tuberin were mapped [21,22]. Mapping the binding site(s) on tuberin Tuberin is subject to multiple phosphorylation events [13,22]. The tuberin interaction is phosphorylationdependent [21 24] and two studies concluded that the binding was dependent on PKB [21,23]. However, mutation of six putative PKB-phosphorylation sites did not prevent tuberin from binding β [22], suggesting that the tuberin interaction may be independent of PKB. Co-immunoprecipitation experiments indicated that proteins bind between tuberin amino acids 1145 and 1364 and that the binding motif surrounding Ser-1254 is essential for binding [22]. However, in contrast to the co-immunoprecipitation experiments, pull-down assays suggested that tuberin contains multiple binding sites [21]. Using this assay, a S1254A substitution did not affect the tuberin interaction, while a R611Q substitution, that inhibited tuberin phosphorylation [38], and a S939A

4 590 Biochemical Society Transactions (2003) Volume 31, part 3 Figure 2 Amino acid substitutions affect the interaction between tuberin and ζ COS-1 cell lysates overexpressing wild-type tuberin (TSC2) or the R611Q, S939A and S1254A variants were incubated with glutathione S-transferase ζ beads as described previously [21]. Wild-type tuberin and the S1254A variant were retained by the beads more efficiently than the S939A and R611Q variants. substitution both reduced tuberin binding, as shown in Figure 2. Why are the data so different? Technical differences between the assays may explain the different outcomes, or the fact that proteins from different species (rat and human) were studied. However, perhaps the most interesting possibility is that different isoforms of tuberin exhibit different binding behaviours. A S1254A substitution, but not a S939A substitution, eliminated binding between β and a tuberin isoform lacking amino acids [22]. In a tuberin isoform containing amino acids , the situation is reversed. A S939A substitution, but not a S1254A substitution, reduced tuberin binding (Figure 2). Tuberin amino acids correspond to the alternatively spliced exon 25 of the TSC2 gene and contain a potential site of PKB-dependent phosphorylation (Ser-981) that also matches the consensus recognition motif [13] (Table 1). Interactions between proteins and their targets are often stabilized by dimerization of the protein and interaction between the dimer and two adjacent phosphorylated recognition motifs on the target protein [40]. Absence of the Ser-981 motif may destabilize the tuberin Ser interaction, leaving Ser-1254 as the primary site of tuberin binding, and stabilization of tuberin binding in the presence of both the Ser-939 and Ser-981 motifs may compensate for the absence of the Ser-1254 motif. Consistent with proposals that the tuberin interaction is phosphorylation-dependent, Ser-1254 is phosphorylated [22]. However, it has not yet been determined whether Ser is phosphorylated by PKB. The Ser-1254 motif does not correspond to a consensus PKB phosphorylation site. An alternative explanation for the conflicting results is that deletion of tuberin amino acids may cause a conformational change, exposing or obscuring recognition motifs. This would be similar to the effect of the R611Q mutation on tuberin binding. Although the R611Q substitution does not alter a recognized phosphorylation or binding site directly, it still inhibits tuberin binding, tuberin phosphorylation and tuberin hamartin binding, presumably by causing a significant conformational change to tuberin [38]. Another possibility is that in the short tuberin isoform the S1254A substitution induces a conformational change that prevents phosphorylation or binding at Ser-939. Amino acids , on the other hand, may maintain access to the binding site around Ser-939, despite changes at the Ser-1254 motif. Function of tuberin binding Why do proteins interact with tuberin, possibly at multiple, phosphorylated sites? Phosphorylation of both S6K and 4E-BP1 was increased in cells over-expressing β, indicating that proteins antagonize the inhibition of S6K and 4E-BP1 phosphorylation by tuberin and hamartin [22,24] proteins are downstream effectors of PI3K/PKB/TOR signalling themselves [35,36]. To demonstrate that the dependent antagonism of tuberin hamartin function was due to the direct interaction between tuberin and proteins, the effect of β expression on S6K phosphorylation by a short tuberin isoform containing the S1254A substitution was investigated [24]. This isoform inhibited S6K phosphorylation, but did not interact with β. The tuberin S1254A-dependent inhibition of S6K phosphorylation was not antagonized by co-expression of β, indicating that proteins are direct antagonists of tuberin hamartin function. Inhibition of tuberin hamartin function by the interaction between proteins and phosphorylated tuberin is consistent with current models of both protein function [25] and PI3K/PKB/TOR signalling [17]. PKBdependent phosphorylation of tuberin at Ser-939 and Thr antagonises the tuberin-dependent inhibition of S6K [13,14], while phosphorylation at Ser-1254 has a similar effect, but requires interaction with proteins [24]. It will be important to establish whether these two mechanisms are independent. Figure 2 indicates that tuberin Ser-1254 phosphorylation does not necessarily influence the tuberin interaction. Instead, Ser-939 phosphorylation appears to be more important for the interaction. The effect of coexpression of proteins and tuberin isoforms containing both the Ser-939 and Ser recognition motifs, normal or inactivated, should help determine how different tuberin isoforms interact with proteins to regulate S6K and 4E-BP1 phosphorylation. Perspectives The identification of the role of tuberin and hamartin in PI3K/PKB/TOR signalling has facilitated progress towards possible therapeutic treatments for TSC [41]. However, the mechanisms whereby the tuberin hamartin complex inhibits

5 TSC Genes Novel Players in the Growth Regulation Network 591 phosphorylation of both S6K and 4E-BP1, either through TOR or independently of TOR, are not known. It is likely that there are additional factors involved in the regulation of tuberin and hamartin function. Identification of these factors will further our understanding of TSC and tell us more about the control of cell growth through PI3K, PKB and TOR. Phosphorylation of tuberin, followed by the binding of a protein results in inactivation of the tuberin hamartin complex and activation of S6K and 4E-BP1 [24]. How do proteins achieve this regulation? Tuberin binding may inhibit tuberin dephosphorylation, thereby stabilizing inactive phosphorylated tuberin, or may recruit other protein components to inactivate the tuberin hamartin complex. In addition, tuberin binding may inhibit the import of phosphorylated tuberin into the nucleus [42]. In yeast, proteins bind transcription factors that are phosphorylated via TORdependent signalling, and prevent import of these factors into the nucleus [35]. The identification and characterization of the interaction between tuberin and proteins has provided new insight into how the tuberin hamartin compex regulates PI3K/PKB/TOR signalling proteins antagonize tuberin hamartin function through their interaction with phosphorylated tuberin [21 24]. However, many questions are still unresolved. In particular it will be essential to determine the exact nature of tuberin binding. Are multiple phosphorylation sites involved and is the interaction of tuberin with, and consequently regulation of tuberin by, proteins dependent on the tuberin isoform that is expressed? Answering these questions will help determine how regulation of the tuberin hamartin complex is coordinated and indicate why alternative splicing of the TSC2 gene may be an important mechanism controlling cell growth in specific cells and tissues. Ultimately, full characterization of the interaction between tuberin and proteins of the family will help explain how tuberin inactivation causes the devastating symptoms associated with TSC. Ann Hunt is thanked for providing the title for this review. Financial support for the authors is provided by the TS Alliance (U.S.A.), the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (The Netherlands), Noortman B.V. (Maastricht, The Netherlands) and the Michelle Foundation (The Netherlands). References 1 Gomez, M.R., Sampson, J.R. and Holets-Whittemore, V. (eds.) (1999) Tuberous Sclerosis Complex, 3rd edn, Oxford University Press, Oxford 2Yeung, R.S., Xiao, G.H., Jin, F., Lee, W.C., Testa, J.R. and Knudson, A.G. (1994) Proc. Natl. Acad. Sci. U.S.A. 91, Kobayashi, T., Hirayama, Y., Kobayashi, E., Kubo, Y. and Hino, O. (1995) Nat. Genet. 9, Ito, N. and Rubin, G.M. (1999) Cell 96, The European Chromosome 16 Tuberous Sclerosis Consortium (1993) Cell 75, van Slegtenhorst, M., de Hoogt, R., Hermans, C., Nellist, M., Janssen, B., Verhoef, S., Lindhout, D., van den Ouweland, A., Halley, D., Young, J. et al. (1997) Science 277, Xu, L., Sterner, C., Maheshwar, M.M., Wilson, P.J., Nellist, M., Short, P.M., Haines, J.L., Sampson, J.R. and Ramesh, V. (1995) Genomics 27, van Slegtenhorst, M., Nellist, M., Nagelkerken, B., Cheadle, J., Snell, R., van den Ouweland, A., Reuser, A., Sampson, J., Halley, D. and van der Sluijs, P. (1998) Hum. Mol. Genet. 7, Plank, T.L., Yeung, R.S. and Henske, E.P. (1998) Cancer Res. 58, Tapon, N., Ito, N., Dickson, B.J., Treisman, J.E. and Hariharan, I.K. (2001) Cell 105, Potter, C.J., Huanh, H. and Xu, T. (2001) Cell 105, Gao, X. and Pan, D. (2001) Genes Dev. 15, Manning, B.D., Tee, A.R., Logsdon, M.N., Blenis, J. and Cantley, L.C. (2002) Mol. Cell 10, Potter, C.J., Pedraza, L.G. and Xu, T. (2002) Nat. Cell Biol. 4, Inoki, K., Li, Y., Zhu, T., Wu, J. and Guan, K.-L. (2002) Nat. Cell Biol. 4, Dan, H.C., Sun, M., Yang, L., Feldman, R.I., Sui, X.M., Ou, C.C., Nellist, M., Yeung, R.S., Halley, D.J., Nicosia, S.V., Pledger, W.J. and Cheng, J.Q. (2002) J. Biol. Chem. 277, McManus, E.J. and Alessi, D.R. (2002) Nat. Cell Biol. 4, Gao, X., Zhang, Y., Arrazola, P., Hino, O., Kobayashi, T., Yeung, R.S., Ru, B. and Pan, D. (2002) Nat. Cell Biol. 4, Tee, A.R., Fingar, D.C., Manning, B.D., Kwiatkowski, D.J., Cantley, L.C. and Blenis, J. (2002) Proc. Natl. Acad. Sci. U.S.A. 99, Jaeschke, A., Hartkamp, J., Saitoh, M., Roworth, W., Nobukuni, T., Hodges, A., Sampson, J., Thomas, G. and Lamb, R. (2002) J. Cell Biol. 159, Nellist, M., Goedbloed, M.A., de Winter, C., Verhaaf, B., Jankie, A., Reuser, A.J.J., van den Ouweland, A.M.W., van der Sluijs, P. and Halley, D.J.J. (2002) J. Biol. Chem. 277, Li, Y., Inoki, K., Yeung, R. and Guan, K.-L. (2002) J. Biol. Chem. 277, Liu, M.Y., Cai, S., Espejo, A., Bedford, M.T. and Walker, C.L. (2002) Cancer Res. 62, Shumway, S.D., Li, Y. and Xiong, Y. (2003) J. Biol. Chem. 278, Tzivion, G. and Avruch, J. (2002) J. Biol. Chem. 277, Yaffe, M.B., Rittinger, K., Volinia, S., Caron, P.R., Aitken, A., Leffers, H., Gamblin, S.J., Smerdon, S.J. and Cantley, L.C. (1997) Cell 91, Yaffe, M.B. (2002) FEBS Lett. 513, Muslin, A.J., Tanner, J.W., Allen, P.M. and Shaw, A.S. (1996) Cell 84, Yaffe, M.B., Leparc, G.G., Lai, J., Obata, T., Volinia, S. and Cantley, L.C. (2001) Nat. Biotechnol. 19, Yaffe, M.B. and Cantley, L.C. (1999) Nature (London) 402, Rittinger, K., Budman, J., Xu, J., Volinia, S., Cantley, L.C., Smerdon, S.J., Gamblin, S.J. and Yaffe, M.B. (1999) Mol. Cell 4, Braselmann, S. and McCormick, F. (1995) EMBO J. 14, Brunet, A., Bonni, A., Zigmond, M.J., Lin, M.Z., Juo, P., Hu, L.S., Anderson, M.J., Arden, K.C., Blenis, J. and Greenberg, M.E. (1999) Cell 96, Baldin, V. (2000) Prog. Cell Cycle Res. 4, Beck, T. and Hall, M.N. (1999) Nature (London) 402, Bertram, P.G., Zeng, C., Thorson, J., Shaw, A.S. and Zheng, X.F. (1998) Curr. Biol. 8, Aicher, L.D., Campbell, J.S. and Yeung, R.S. (2001) J. Biol. Chem. 276, Nellist, M., Verhaaf, B., Goedbloed, M.A., Reuser, A.J.J., van den Ouweland, A.M.W. and Halley, D.J.J. (2001) Hum. Mol. Genet. 10, Nellist, M., van Slegtenhorst, M.A., Goedbloed, M., van den Ouweland, A.M.W., Halley, D.J.J. and van der Sluijs, P. (1999) J. Biol. Chem. 274, Tzivion, G., Shen, Y.H. and Zhu, J. (2001) Oncogene 20, Kenerson, H.L., Aicher, L.D., True, L.D. and Yeung, R.S. (2002) Cancer Res. 62, Lou, D., Griffith, N. and Noonan, D.J. (2001) Mol. Cell Biol. Res. Commun. 4, Ichimura, T., Isobe, T., Okuyama, T., Takahashi, N., Araki, K., Kuwano, R. and Takahashi, Y. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, Zhang, L., Chen, J. and Fu, H. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, Liu, Y., Yin, G., Surapisitchat, J., Berk, B.C. and Min, W. (2001) J. Clin. Invest. 107, Received 13 January 2003