Mol Cell: 西雅圖華盛頓大學(xué)許文清實驗室：ICAT結(jié)構(gòu) Molecular Cell, Vol 10, 563-571, September 2002 The Crystal Structure of the -Catenin/ICAT Complex Reveals the Inhibitory Mechan ism of ICAT Thomas A. Graham1,2, Wilson K. Clements3, David Kimelman3, and Wenqing Xu1 1 Department of Biological Structure, University of Washington, Seattle, WA 9819 5 USA 2 Biomolecular Structure and Design Program, University of Washington, Seattle, WA 98195 USA 3 Department of Biochemistry, University of Washington, Seattle, WA 98195 USA Correspondence: Wenqing Xu 206-221-5609 (phone) 206-543-1524 (fax) [email protected] Table of Contents Download as printable (PDF) file - 583K Search Medline for articles by: Thomas A. Graham||Wenqing Xu Download to Citation Manager Summary Summary Introduction Results Discussion Experimental Procedures References -catenin is a multifunctional protein involved in both cell adhesion and transcr iptional activation. Transcription mediated by the -catenin/Tcf complex is invol ved in embryological development and is upregulated in various cancers. We have determined the crystal structure at 2.5 Å resolution of a complex between -catenin and ICAT, a protein that prevents the interaction between -catenin and Tcf/Lef family transcription factors. ICAT contains a 3-helix bundle that binds armadillo repeats 10–12 and a C-terminal tail that, similar to Tcf and E-cadherin, binds in the groove formed by armadillo repeats 5?C9 of -catenin. We show that ICAT selectively inhibits -catenin/Tcf binding in v ivo, without disrupting -catenin/cadherin interactions. Thus, it should be possi ble to design cancer therapeutics that inhibit -catenin-mediated transcriptional activation without interfering with cell adhesion. Introduction Summary Introduction Results Discussion Experimental Procedures References The canonical Wnt pathway is critical for embryological development in vertebrat es (reviewed by Moon and Kimelman, 1998 ; Wodarz and Nusse, 1998 ; Peifer and Po lakis, 2000 ) as well as for the initiation and growth of certain cancers such a s colorectal cancer (reviewed by Kinzler and Vogelstein, 1996 ; Morin, 1999 ; Bi enz and Clevers, 2000 ; Polakis, 2000 ). A key step in the activation of target genes for this pathway is the formation of a complex between -catenin and the DN A binding factor Tcf. Tcf binds the promoters of Wnt responsive genes, keeping t hem in a repressed state in the absence of -catenin (reviewed by Roose and Cleve rs, 1999 ; Barker et al., 2000 ; Hecht and Kemler, 2000 ; Sharpe et al., 2001 ). The -catenin/Tcf complex acts to recruit transcriptional coactivators, such as p300/CBP (Hecht et al., 2000 ; Miyagishi et al., 2000 ; Sun et al., 2000 ; Takem aru and Moon, 2000 ) and Pygopus (Kramps et al., 2002 ; Parker et al., 2002 ; Th ompson et al., 2002 ) to the promoter. Genes activated by the -catenin/Tcf compl ex, including c-myc and cyclin D1, play critical roles in cell growth, prolifera tion, and differentiation (He et al., 1998 ; Shtutman et al., 1999 ; Tetsu and M cCormick, 1999 ). -catenin also plays an essential role in cell adhesion by simu ltaneously binding to the transmembrane protein cadherin and the cytoskeleton bi nding protein -catenin (reviewed by Provost and Rimm, 1999 ; Gottardi and Gumbin er, 2001 ; Nagafuchi, 2001 ; Pokutta and Weis, 2002 ). The release of -catenin f rom this complex promotes cell motility and causes the level of cytosolic -caten in to rise (reviewed by Conacci-Sorrell et al., 2002 ). The level of -catenin in the cytosol is normally regulated by the -catenin destr uction complex, which consists of the scaffold protein Axin, the product of the colon cancer gene APC (adenomatous polyposis coli), two Ser/Thr kinases CK1 and GSK-3, and -catenin itself (reviewed by Peifer and Polakis, 2000 ; Seidensticker and Behrens, 2000 ; Harwood, 2002 ). In colon cancer cells, components of this complex are frequently mutated, which results in a constitutive increase in -cat enin levels and enhanced formation of the -catenin/Tcf complex (reviewed by Mori n, 1999 ; Polakis, 2000 ; van Noort and Clevers, 2002 ). The importance of developing an inhibitor that is able to counteract this deregu lation of -catenin has made the -catenin/Tcf complex a drug target for cancer tr eatment (reviewed by Hecht and Kemler, 2000 ). However, it is critical that an i nhibitor of the -catenin/Tcf complex does not interfere with -catenin/cadherin i nteractions. The crystal structures of the armadillo repeat region of -catenin i n complex with the catenin binding domain (CBD) of Tcf and the cytoplasmic domai n of E-cadherin reveal striking similarities between the binding modes of Tcf an d cadherin along the positively charged groove of -catenin (Graham et al., 2000 , 2001 ; Huber and Weis, 2001 ; Poy et al., 2001 ). Specifically, the core -cate nin binding regions of Tcf and cadherin interact with -catenin through essential ly the same set of contacts, including two critical salt bridges. This similarit y has made it seem potentially difficult to design compounds that can inhibit th e formation of the -catenin/Tcf complex without simultaneously disrupting that o f the -catenin/cadherin complex. ICAT (inhibitor of -catenin and Tcf), a negatively charged 81 residue protein, i s a physiological inhibitor of the Wnt signaling pathway that prevents the bindi ng of Tcf to -catenin (Tago et al., 2000 ; Tutter et al., 2001 ). ICAT is requir ed for normal embryonic development in Xenopus (Tago et al., 2000 ), and ICAT an d an ICAT homologous gene, LZIC, are localized in a human chromosome region that is frequently rearranged or deleted in various cancers (Katoh, 2001 ). Reduced ICAT transcript levels were found in more than two-thirds of malignant melanomas tested (Reifenberger et al., 2002 ). It was puzzling how ICAT could prevent Tcf binding to -catenin. A truncational s tudy of -catenin showed that armadillo repeats 10–12 are required for ICAT binding (Tago et al., 2000 ), whereas the -catenin/Tcf structure showed that only armadillo repeats 3?C9 of -catenin are involved in the binding of Tcf to -catenin (Graham et al., 20 00 , 2001 ; Poy et al., 2001 ). In addition, since ICAT is found in the cytosol as well as the nucleus, and since cadherin and Tcf bind -catenin in very similar ways, ICAT might also be expected to regulate cell adhesion, although this was not studied in the original report (Tago et al., 2000 ). To reveal how ICAT inhibits Wnt signaling and to understand how it may affect ce ll adhesion, we determined the crystal structure of a -catenin/ICAT complex at 2 .5 Å resolution. The structure reveals the basis of the -catenin/ICAT inte raction and shows how ICAT is able to disrupt canonical Wnt pathway gene transac tivation. In combination with our structural studies, our in vivo analysis provi des insight into how drugs that disrupt the -catenin/Tcf complex without disturb ing the integrity of adherens junctions might be designed. Results Summary Introduction Results Discussion Experimental Procedures References Overall Structure The structure presented here consists of the armadillo repeat region of human -c atenin (residues 134–664) in complex with murine ICAT (81 residues). Only two residues differ between human and murine ICAT, residue 49 is asparagine in human and serine in mouse, and residue 57 is a proline in human and glutamine in mouse. From the structure, it is clear that neither residue plays a critical role in the binding of ICAT to -catenin or in its inhibitory function. The crystal used in this study has one complex of -catenin/ICAT per asymmetric unit (Table 1). Residues 150?C549 and 563–663 of -catenin and residues 5–75 of ICAT are visible in the elec tron density map. Table 1. Structure Determination and Refinement Statistics View this table: [in this window] [In new window] The model shows that the binding mode of ICAT extends along -catenin armadillo r epeats 5–12 with two distinct structural domains (Figure 1). The first ICAT domain (residues 5?C58) is a 3-helix bundle, with the helices denoted as H1 (residues 11–28), H2 (residues 36?C42), and H3 (residues 46–55). This bundle specifically interacts with -catenin armadillo repeats 10?C12. The second ICAT domain (residues 59–75) is an extended tail that binds in an antiparallel fashion along the positively charged groove of -catenin armadillo repeats 5?C9. Figure 1. Overall Structure of ICAT Bound to the Armadillo Repeat Region of -Catenin The core of -catenin is made up of the armadillo repeat region which consists of 12 repeats (labeled in the figure). Each repeat consists of three helices color ed in blue, green, and yellow, except for repeat 7, which only contains the seco nd and third helices. The third helices of repeats 5–12 form the platform upon which ICAT binds to -catenin in an anti-parallel fashion. ICAT binds to -catenin with two discrete modules, an N terminus 3-helix bundle (shown in fuchsia) and a C terminus extended strand-like tail (shown in red). The 3-helix bundle domain binds to -catenin armadillo repeats 10?C12 and the tail binds along repeats 5–9. This figure was generated with GRASP (Nicholls et al., 1991 ), Molscript (Kraulis, 1991 ), and RASTER3D (Merritt and Murphy, 1994 ). View larger version: [In this window] [In new window] Contacts between ICAT and -Catenin ICAT helices H1 and H2 sit atop the third helix of -catenin armadillo repeat 12, making hydrophobic and electrostatic interactions with -catenin residues Phe660 and Arg661. ICAT residues Tyr15, Lys19, and Val22 of H1 form a hydrophobic patc h that interacts with -catenin residue Phe660, while ICAT Glu37 of H2 forms two hydrogen bonds with -catenin Arg661 (Figures 2A and 2B). In addition, Met29 of I CAT stabilizes the C terminus of H1 by forming a hydrophobic contact with -caten in Tyr654 (Figures 2A and 2B). The residues in between H1 and H2 form a stretch that makes contacts with the third helix of the twelfth repeat of -catenin but a re otherwise exposed to the solvent. Figure 2. ICAT 3-Helix Bundle Domain Bound to -Catenin (A) Stereo 2Fo-Fc map of the 3-helix bundle domain of ICAT bound to -catenin. Th e ICAT residues are denoted in fuchsia, and the -catenin residues are denoted in yellow. The map is contoured at 1. (B) Ribbon and ball-and-stick diagram showing the specific interactions between the 3-helix bundle domain of ICAT and the third helix of -catenin armadillo repe at 12. The color scheme is the same as that described in Figure 1. (C) Superposition of the -catenin/E-cadherin and the -catenin/ICAT structures wa s carried out by superimposing -catenin repeats 10–12. ICAT is shown in fuchsia and E-cadherin in turquoise. View larger version: [In this window] [In new window] H3 does not form any contacts with -catenin; rather, it provides structural inte grity to the helical bundle. In particular, Val47 and Val48 of this helix are bu ried in the hydrophobic core, along with Gln17, Val20, and Leu24 of helix H1. Ty r44, a residue linking H2 and H3, also packs in the core of the helical bundle w ith Ile16 of H1. In the ICAT tail domain, Asp66 forms a salt bridge with -catenin Lys435, and Glu 75 forms a salt bridge with -catenin Lys312 (Figure 3A). We previously identifie d these two lysines in -catenin as charged buttons that are critical for anchori ng Tcf to -catenin (Graham et al., 2000 ). The ICAT residues between these two s alt bridges form hydrogen bonds and hydrophobic interactions with -catenin. For example, the side chain of ICAT Phe71 packs against that of -catenin Arg386. Figure 3. ICAT Tail Domain Bound to -Catenin (A) Molecular model of the ICAT tail domain bound to -catenin. ICAT is shown in the ball-and-stick format (residues labeled in black) and -catenin as cylinders. The two charged buttons of -catenin, residues Lys312 and Lys435, are also shown in the ball-and-stick format. The color scheme is the same as described in Figu re 1. (B) Structural alignment of the -catenin-bound forms of ICAT (red), XTcf3 (green ), and E-cadherin (turquoise). Three conserved residues are labeled and color co ded with the respective sequential positions given in the structures. The struct ures of these three complexes were superimposed using armadillo repeats 5–8 of -catenin. (C) Sequence alignment of the charged-button-recognition region of ICAT together with representative Tcf and cadherin family members. LZIC, a recently discovere d member of the ICAT family, is also shown. The conserved aspartic acid that has been shown to interact with -catenin Lys435 for ICAT, hTcf4, XTcf3, and E-cadhe rin is shown in red, and the conserved hydrophobic residue is shown in yellow. T he cluster of glutamic acids, denoted in green, can form a salt bridge with Lys3 12, using various conformations (Graham et al., 2001 ). View larger version: [In this window] [In new window] Mutational Analysis of the -Catenin/ICAT Interaction To test whether the 3-helix bundle might anchor ICAT to -catenin, we performed s ite-directed mutagenesis, changing -catenin residues Phe660 and Arg661 to alanin es. These two residues are predicted from our structure to be critical for the b inding of the 3-helix bundle to -catenin (Figure 2B). As shown in Figure 4, the conversion of residues 660 and 661 to alanine abolished the binding of ICAT to - catenin, demonstrating that the 3-helix bundle is necessary for binding ICAT to -catenin. In addition, we found that the helical bundle domain alone is sufficie nt for -catenin binding (data not shown). Figure 4. -Catenin Phe660 and Arg661 Are Required for ICAT Binding -catenin carrying the mutations F660A/R661A (mut Cat) was tested for its ability to bind ICAT in vitro. Wild-type (lane 1) but not mutant (lane 2) -catenin is a ble to coprecipitate ICAT. All of the proteins were 35S-labeled and mixed as des cribed. -catenin was immunoprecipitated via its HA-epitope tag. Lanes 4–6 show protein levels prior to immunoprecipitation. View larger version: [In this window] [In new window] To determine if the tail is the structural domain responsible for blocking the b inding of Tcf to -catenin, we constructed a truncated form of ICAT (ICATC) that consists of residues 1–58. While the addition of full-length ICAT prevented the binding of Tcf to -catenin, ICATC did not impair Tcf binding to -catenin ( Figure 5A, lanes 2 and 3). Thus we conclude that the 3-helix bundle provides the critical domain for binding -catenin, whereas the tail is necessary for excluding Tcf from -catenin. Figure 5. The ICAT Tail Domain Is Necessary to Block the Binding of Tcf to -Catenin -catenin, precipitated via its HA-epitope tag, was tested for its ability to cop recipitate XTcf3C (A) or C-cadherin (B) in the presence of ICAT (lane 2), ICATC (lane 3), or BSA (lane 4). 35S-labeled -catenin, XTcf3C, and C-cadherin were pro duced in vitro and mixed with 100 ng unlabeled ICAT, ICATC, or BSA. Lanes 5–8 show protein levels present prior to immunoprecipitation. XTcf3C was used instead of full-length XTcf3 for convenience (Graham et al., 2000 ). (A) Wild-type ICAT (lane 2) but not ICATC (lane 3) blocks coprecipitation of XTc f3C. (B) Either wild-type ICAT (lane 2) or ICATC (lane 3) blocks the coprecipitation of C-cadherin in vitro. View larger version: [In this window] [In new window] The Inhibitory Role of ICAT In Vitro and In Vivo The structures of the -catenin/Tcf-CBD (Graham et al., 2000 , 2001 ; Poy et al., 2001 ), -catenin/E-cadherin-CBD (Huber and Weis, 2001 ), and -catenin/ICAT comp lexes reveal that Tcf, cadherin, and ICAT bind to the positively charged groove of -catenin in a very similar fashion (Figure 3B). Since ICAT inhibits Tcf from binding to -catenin, we reasoned that it might inhibit cadherin from binding as well. In order to test this hypothesis, we tested in vitro the ability of ICAT t o compete off cadherin. Using full-length ICAT protein, we observed that ICAT was able to effectively bl ock the binding of C-cadherin to -catenin (Figure 5B). This result is consistent with our previous study demonstrating that the binding of cadherin to -catenin, like Tcf, requires the charged buttons Lys312 and Lys435 (Graham et al., 2000 ) , which are also bound by the ICAT tail (Figure 3). Intriguingly, ICATC also par tially inhibited the binding of C-cadherin to -catenin (Figure 5B), although it had no effect on Tcf binding (Figure 5A). This ability is consistent with the st ructure, which shows that while the 3-helix bundle of ICAT does not overlap with the -catenin binding region of Tcf, it would be predicted to interfere with cad herin binding. Thus the 3-helix bundle, together with the tail of ICAT, can act as a potential inhibitor of cadherin binding. Our in vitro results suggested that ICAT should be a potent inhibitor of -cateni n's interactions with both Tcf and cadherin in vivo. Indeed, Xenopus embryos inj ected with ICAT show a ventralized phenotype characteristic of an inhibition of the -catenin/Tcf interaction (Tago et al., 2000 ; our unpublished data). Surpris ingly, however, injected embryos did not appear to have defects in cell-cell adh esion as might be expected if -catenin's interaction with cadherin were inhibite d (Levine et al., 1994 ). We therefore tested whether ICAT inhibits cadherin bin ding to -catenin in vivo as well as in vitro. Xenopus embryos were injected with RNA encoding -catenin-FLAG, in combination with RNA encoding HA-epitope tagged Tcf or C-cadherin, and the -catenin-FLAG was immunoprecipitated by the FLAG-epit ope tag. In some samples, RNA encoding ICAT was also injected. ICAT effectively blocked the binding of Tcf to -catenin, whereas it did not block the binding of C-cadherin to -catenin (Figure 6). Thus, while ICAT has the ability to block bot h Tcf and C-cadherin in vitro, it can only block the binding of Tcf in vivo. Figure 6. ICAT Selectively Blocks Tcf Binding to -Catenin In Vivo The ability of -catenin to bind either XTcf3-HA or C-cadherin-HA in the presence or absence of ICAT was tested in Xenopus embryos. Equivalent levels of GFP RNA were injected into embryos that did not receive ICAT RNA. Embryo lysates were im munoprecipitated (lanes 6–10) for the FLAG epitope tag on -catenin and probed with an anti-HA antibody. Lanes 1?C5 correspond to lanes 6–10 and show the levels of proteins in total lysates prior to immunoprecipitation. Lanes 1 and 6 show results from uninjected embryos. ICAT prevents coprecipitation of XTcf3 (lane 8) but not C-cad (lane 10) with -catenin. View larger version: [In this window] [In new window] Discussion Summary Introduction Results Discussion Experimental Procedures References In the canonical Wnt signaling pathway, APC, Axin, and GSK-3 regulate the transd uction of Wnt signals by controlling the cytosolic -catenin level through the -c atenin destruction complex. In contrast, ICAT regulates the transcription of Wnt responsive genes by directly blocking the formation of the -catenin/Tcf complex (Tago et al., 2000 ; Tutter et al., 2001 ). Structural determination of the -ca tenin/ICAT complex is thus a key step in understanding how ICAT functions as an inhibitor of the canonical Wnt pathway. An ''Anchor-and-Kick'' Model of ICAT Only armadillo repeats 10–12 of -catenin are required for ICAT binding (Tago et al., 2000 ), whereas the residues critical for the binding of Tcf to -catenin have been structurally shown to lie along repeats 3?C9 of -catenin (Graham et al., 2000 , 2001 ; Poy et al., 2001 ). The regions of -catenin required for binding the inhibitor (ICAT) and the molecule being inhibi ted (Tcf) do not overlap, raising a puzzle as to how the inhibition occurs. The structure of the -catenin/ICAT complex, together with our in vitro studies, reso lves this conundrum. The 3-helix bundle provides the key interactions for ''anch oring'' ICAT to repeat 12 of -catenin whereas the tail binds in the positively c harged groove where it interacts with the two charged buttons, Lys312 and Lys435 . It is the binding of the ICAT tail to these lysines that allows it to ''kick'' Tcf off of -catenin. Elimination of the tail does not prevent ICAT from binding to -catenin but does prevent ICAT from inhibiting the binding of Tcf to -cateni n. Thus, the 3-helix bundle domain may function as a dominant-negative form of I CAT. The use of -catenin repeats 10–12 as the anchoring site for ICAT, rather than repeats 3?C9 that Tcf uses, may provide another level of regulation over ICAT's inhibition of -catenin. Proteins that bind to this region of -catenin may act to specifica lly regulate the inhibitory role of ICAT. For example, the -catenin binding site s of p300/CBP, TBP, and Brg-1 overlap with that of the 3-helix bundle of ICAT (H echt et al., 1999 , 2000 ; Miyagishi et al., 2000 ; Takemaru and Moon, 2000 ; Ba rker et al., 2001 ). Likewise, the 3-helix bundle may also affect the binding an d function of these proteins in vivo. In fact, Daniels and Weis have shown that the 3-helix bundle of ICAT prevents the binding of p300 to -catenin in vitro (Da niels and Weis, 2002 [this issue of Molecular Cell]). LZIC is a potential tumor suppressor that contains an N-terminal leucine zipper of unknown function and a C-terminal domain that is 38% identical to ICAT (Katoh , 2001 ). While the binding of LZIC to -catenin has not been tested, ICAT residu es involved in the -catenin interaction of the 3-helix bundle and the tail are c onserved in the ICAT homologous domain of LZIC. Thus, we predict the ICAT homolo gous domain of LZIC can also interact with -catenin and inhibit Tcf binding with an ''anchor-and-kick'' mechanism. A Comparison of ICAT and Cadherin Binding to -Catenin While cadherins bind the two charged buttons of -catenin, Lys312 and Lys435 (Hub er and Weis, 2001 ), they also contact the same region of -catenin as does the 3 -helix bundle of ICAT (Figure 2C). Residues 18–28 of ICAT helix H1 structurally correspond to cadherin residues 653?C663 in the -catenin/E-cadherin-CBD structure (Figure 2C), even though a compari son of the sequences of these two corresponding helices in ICAT and E-cadherin s hows no obvious homology. The helix H1 of ICAT is anchored in the middle by hydr ophobic interactions with -catenin Phe660 and a salt bridge between ICAT Glu37 a nd -catenin Arg661 (Figure 2B), whereas the corresponding helix in E-cadherin is stabilized at its C terminus by a hydrogen bond between its Asp665 and -catenin Tyr654 (Figure 2C). Phosphorylation of -catenin Tyr654 by c-Src causes a 6-fold reduction in the binding affinity between -catenin and E-cadherin (Roura et al. , 1999 ). Although Met29 of ICAT forms a van der Waals contact with -catenin Tyr 654 in the -catenin/ICAT structure, it is possible that the phosphorylation of - catenin Tyr654 might not disrupt the -catenin/ICAT interaction. ICAT Competes Specifically with Tcf ICAT and E-cadherin share an extensive binding region on -catenin (Figures 2C an d 3B) . The structures and sequence homology of the domains in these proteins th at bind the charged buttons suggested that ICAT would interfere with the binding of cadherin to -catenin just as it inhibits Tcf (Figures 3B and 3C). Our in vit ro data demonstrate that ICAT can indeed inhibit cadherin binding to -catenin (F igures 5A and 5B); however, our in vivo results make it clear that in the cell I CAT is able to interfere with Tcf binding to -catenin without disrupting the -ca tenin/cadherin complex (Figure 6). The difference between the in vitro and in vivo results may be due to the phosph orylation state of cadherin since phosphorylation can increase its affinity for -catenin by three orders of magnitude (W. Weis, personal communication). Since I CAT is found in both the nucleus and cytosol, the enhanced binding of cadherin t o -catenin in vivo may be critical for the ability of ICAT to act as a natural i nhibitor that blocks the Wnt signaling pathway without disrupting cell adhesion. Implications for a New Cancer Therapy A drug that inhibits the binding of Tcf to -catenin without interfering with the -catenin/cadherin complex could be very valuable in treating certain forms of c ancer. The prospects for such a drug seemed complicated by the structural observ ation that cadherin and Tcf bind the same critical region of -catenin (Graham et al., 2000 , 2001 ; Huber and Weis, 2001 ; Poy et al., 2001 ). However, our in v ivo studies with ICAT indicate that it may be possible to design a drug that can displace Tcf from -catenin without inhibiting the binding of cadherin. More spe cifically, our results suggest that this drug might not need to specifically bin d to a region of -catenin unique to the binding mode of Tcf. These findings suggest at least two possible approaches to the development of ne w cancer therapies. First, the -catenin groove where the core region of the Tcf- CBD binds, in particular the two charged buttons, might be targeted for drug des ign. Second, ICAT itself might be used as a gene therapy agent. The prospect of using this method to fight cancer has been recently studied in cultured cells (S ekiya et al., 2002 ). Furthermore, an adenovirus was recently engineered that ca n specifically target tumor cells in which there is constitutive activation of t he Wnt signaling pathway (Fuerer and Iggo, 2002 ). The advantage of using this a denovirus as the method of treatment is that the delivery of the inhibitor is se lective for tumor cells. Therefore, ICAT or an ICAT mimic, whose mRNA is deliver ed by this adenovirus, would only act upon -catenin that is unregulated in the d iseased cell and not upon -catenin that is properly regulated in healthy neighbo ring cells. Experimental Procedures Summary Introduction Results Discussion Experimental Procedures References Protein Expression, Purification, and Crystallization ICAT and the armadillo repeat region of -catenin (residues 134–664) were individually expressed in E. coli as GST fusion proteins with a TEV cleavage site for removing the tag. ICAT and -catenin were individually purified with a glutathione sepharose column. The GST tags were cleaved with TEV, and proteins were further purified with a H-trap Q column. The final stock solution for ICAT was 10 mM Tris (pH 8.5), 75mM NaCl, and 2mM dithiothreitol (DTT), and for -catenin it was 20 mM Tris (pH 8.5), 50 mM NaCl, and 2 mM DTT. ICAT was concentrated to 4 mg/ml and -catenin to 3.2 mg/ml. A 2:1 molar ratio of ICAT and -catenin was incubated at 4?鉉 for 30 min. The complex was separated from excess ICAT via centrifuge filtration resulting in a final concentration of 5 mg/ml. The hanging drop vapor diffusion method was used at room temperature to screen a nd grow crystals. Clusters of single crystals were obtained by mixing 1 µl of the protein complex stock solution together with 1 µl of the well solu tion, which contained 4% (w/v) PEG 8000, 20 mM MgCl2, 50 mM Tris (pH 8.0), and 5 mM DTT. Crystals with a size of 50 µm × 100 µm × 200 µm typ ically showed up in 3–4 days. Data Collection and Structure Determination A 2.5 Å data set was collected at the APS synchrotron beamline 19ID ( = 0. 9196 Å), and the data was processed with DENZO and SCALEPACK (Otwinowski a nd Minor, 1997 ). The data collection statistics are given in Table 1. The space group was P21212 with the unit cell dimensions a = 95.17 Å, b = 98.50 &Ar ing;, and c = 86.49 Å. Molecular replacement via AmoRe (Navaza, 1994 ) fou nd one solution in the asymmetric unit using the -catenin models in PDB entries 1G3J and 1JDH as the search models. The final correlation coefficient and R fact or were 60.2 and 39.0, respectively. Refinement and model building were carried out with CNS (Brunger et al., 1998 ) and XtalView (McRee, 1999 ). A test set of 10% of the reflections were set aside and a cutoff of 2 was used throughout the refinement. After an initial round of simulated annealing, the initial model was built in. This was followed by round s of positional minimization and individual B factor refinement and further mode l building. PROCHECK (Laskowski et al., 1993 ) was used to check the stereochemi stry of the model after the refinement step (Table 1). With anisotropic B factor and bulk solvent corrections, the final R factors are Rwork = 21.1% and Rfree = 25.5%. In Vitro Translation and Competition Assays Unlabeled ICAT was produced as above. A truncated form of ICAT, ICATC, was made by introducing a stop codon in place of ICAT Ser59 with the primer 5'-CTCAGCCAGC TGCCACAGCACTCCATCGACCAGGGTGCAGAGGAC-3'; site-directed mutagenesis was carried ou t using the QuickChange procedure (Stratagene). The TNT Quick Coupled SP6 Transc ription/Translation system (Promega) was used to produce 35S-labeled -catenin-HA (Farr et al., 2000 ), XTcf3C-Myc (Graham et al., 2001 ), and C-cadherin (Levine et al., 1994 ). Proteins were combined as described in the text and nutated at 4°C for 1 hr in a total volume of 24 µl using 1× PBS as a diluent. One h undred nanogram unlabeled competitor (ICAT, ICATC, or BSA) in 1 µl buffer (50% glycerol, 10mM Tris [pH 8.5], 75mM NaCl, 2mM DTT) was added and incubated 1 hr at 30°C. One microliter of sample was removed for analysis of input. Sample s were immunoprecipitated using mouse anti-HA antibodies (BABCO) preconjugated t o protein G sepharose beads (Pharmacia) in wash buffer (10 mM Tris [pH 8.0], 140 mM NaCl, 1 mM EDTA, 0.1% Triton X-100) for 1 hr at 4°C. Beads were washed thre e times in wash buffer, eluted with 2× SDS sample buffer, and analyzed by SDS-P AGE. Experiments examining the ability of -catenin carrying mutations Phe660Ala/ Arg661Ala (mut -catenin) to bind ICAT were performed the same way using 35S-labe led in vitro TNT-produced ICAT. Site-directed mutagenesis was done with the prim er 5'-GCAGCTGCTGTTCTGGCCGCTATGTCTGAGGACAAACCCCAGGAC-3', using the QuickChange pr ocedure. In Vivo Competition Assays Xenopus laevis embryos were obtained as described (Newport and Kirschner, 1982 ) . Embryos were microinjected (Moon and Christian, 1989 ) with in vitro synthesiz ed mRNA (mMessage mMachine, Ambion). An HA-epitope tag was added to the C termin us of XTcf3 using the primer 5'-CCTCTAGATTAAGCGTAATCTGGCACATCGTATGGGTAGTCACTGGAT TTGGTCACCAGAG-3', and the PCR product was subcloned into CS2+ (Turner and Weintr aub, 1994 ). Five hundred picograms of mRNA encoding both -catenin-FLAG (the Myc tags of -catenin-Myc in CS2+ [Yost et al., 1996] were replaced with a FLAG tag ), and 1 ng of XTcf3-HA or 500 pg C-cadherin-HA (generous gift of Pierre McRea) were coinjected along with 1 ng of mRNA encoding either ICAT (Tago et al., 2000 ) or GFP. Single injections were at the animal pole of two-cell embryos. Embryos were cultured for 4 hr, and embryos were lysed in Triton X-100 lysis buffer as described (Rubinfeld et al., 1993 ). A portion was retained for analysis of prot ein content in total lysates. Half the lysate was immunoprecipitated for 1 hr at 4°C using mouse anti-FLAG antibodies (Sigma) conjugated to protein G sepharose (Pharmacia) and processed as described above, except that proteins were transfe rred to nitrocellulose membranes for detection by Western blotting. Blots were p robed with mouse anti-HA primary antibodies (BABCO) and goat anti-rabbit/horsera dish peroxidase secondary antibodies (Zymed), and were visualized using Renaissa nce enhanced chemiluminescence (NEN Life Science Products, Inc.). Total lysate l anes represent the protein content of 1 embryo equivalent. Immunoprecipitation l anes represent precipitations from 10 embryo equivalents. Acknowledgments We are grateful to R. Zhang for assistance with synchrotron data collection at A PS 19ID and C. Weaver for help with protein binding assays. This work was suppor ted by NIH grants CA90351 to W.X. and HD27262 to D.K. T.A.G. was supported by NI H training grant GM07270. Accession Numbers The coordinates of the -catenin/ICAT structure have been deposited in the Protei n Data Bank with the ID code 1LUJ. References Summary Introduction Results Discussion Experimental Procedures References Barker N., Morin P.J. and Clevers H. (2000) The yin-yang of TCF/-catenin signali ng. Adv. Cancer Res., 77:1-24. [Medline] Barker N., Hurlstone A., Musisi H., Miles A., Bienz M. and Clevers H. (2001) The chromatin remodelling factor Brg-1 interacts with -catenin to promote target ge ne activation. EMBO J., 20:4935-4943. [Medline] Bienz M. and Clevers H. (2000) Linking colorectal cancer to Wnt signaling. Cell, 103:311-320. [Medline][Full Text] Brunger A.T., Adams P.D., Clore G.M., DeLano W.L., Gros P., Grosse-Kunstleve R.W ., Jiang J.S., Kuszewski J., Nilges M. and Pannu N.S. et al. (1998) Crystallogra phy & NMR system: a new software suite for macromolecular structure determinatio n. Acta Crystallogr. D Biol. Crystallogr., 54:905-921. Conacci-Sorrell M., Zhurinsky J. and Ben-Ze'ev A. (2002) The cadherin-catenin ad hesion system in signaling and cancer. J. Clin. Invest., 109:987-991. [Medline] Daniels D.L. and Weis W.I. (2002) ICAT inhibits -catenin binding to Tcf/Lef-fami ly transcription factors and the general coactivator p300 using independent stru ctural modules. Mol. Cell, 10:573-584.[Summary][Full Text] Farr G.H. III, Ferkey D.M., Yost C., Pierce S.B., Weaver C. and Kimelman D. (200 0) Interaction among GSK-3, GBP, axin, and APC in Xenopus axis specification. J. Cell Biol., 148:691-702. [Medline] Fuerer C. and Iggo R. (2002) Adenoviruses with Tcf binding sites in multiple ear ly promoters show enhanced selectivity for tumour cells with constitutive activa tion of the wnt signalling pathway. Gene Ther., 9:270-281. [Medline] Gottardi C.J. and Gumbiner B.M. (2001) Adhesion signaling: how -catenin interact s with its partners. Curr. Biol., 11:R792-R794. [Medline][Summary][Full Text] Graham T.A., Weaver C., Mao F., Kimelman D. and Xu W. (2000) Crystal structure o f a -catenin/Tcf complex. Cell, 103:885-896. [Medline][Summary][Full Text] Graham T.A., Ferkey D.M., Mao F., Kimelman D. and Xu W. (2001) Tcf4 can specific ally recognize -catenin using alternative conformations. Nat. Struct. Biol., 8:1 048-1052. [Medline] Harwood A.J. (2002) Signal transduction in development: holding the key. Dev. Ce ll, 2:384-385. He T.C., Sparks A.B., Rago C., Hermeking H., Zawel L., da Costa L.T., Morin P.J. , Vogelstein B. and Kinzler K.W. (1998) Identification of c-MYC as a target of t he APC pathway. Science, 281:1509-1512. [Medline] Hecht A. and Kemler R. (2000) Curbing the nuclear activities of -catenin. Contro l over Wnt target gene expression. EMBO Rep., 1:24-28. Hecht A., Litterst C.M., Huber O. and Kemler R. (1999) Functional characterizati on of multiple transactivating elements in -catenin, some of which interact with the TATA-binding protein in vitro. J. Biol. Chem., 274:18017-18025. [Medline] Hecht A., Vleminckx K., Stemmler M.P., van Roy F. and Kemler R. (2000) The p300/ CBP acetyltransferases function as transcriptional coactivators of -catenin in v ertebrates. EMBO J., 19:1839-1850. [Medline] Huber A.H. and Weis W.I. (2001) The structure of the -catenin/E-cadherin complex and the molecular basis of diverse ligand recognition by -catenin. Cell, 105:39 1-402. [Medline][Summary][Full Text] Katoh M. (2001) Molecular cloning and characterization of LZIC, a novel gene enc oding ICAT homologous protein with leucine zipper domain. Int. J. Mol. Med., 8:6 11-615. Kinzler K.W. and Vogelstein B. (1996) Lessons from hereditary colorectal cancer. Cell, 87:159-170. [Medline][Full Text] Kramps T., Peter O., Brunner E., Nellen D., Froesch B., Chatterjee S., Murone M. , Zullig S. and Basler K. (2002) Wnt/wingless signaling requires BCL9/legless-me diated recruitment of pygopus to the nuclear -catenin-TCF complex. Cell, 109:47- 60. [Medline][Summary][Full Text] Kraulis P.J. (1991) Molscript: a program to produce both detailed and schematic plots of protein structure. J. Appl. Crystallogr., 24:946-950. Laskowski R.A., MacArthur M.W., Moss D.S. and Thornton J.M. (1993) Procheck: a p rogram to check the stereochemical quality of protein structures. J. Appl. Cryst allogr., 26:283-291. Levine E., Lee C.H., Kintner C. and Gumbiner B.M. (1994) Selective disruption of E-cadherin function in early Xenopus embryos by a dominant negative mutant. Dev elopment, 120:901-909. [Medline] McRee D.E. (1999) XtalView/Xfit—a versatile program for manipulating atomic coo rdinates and electron density. J. Struct. Biol., 125:156-165. [Medline] Merritt E.A. and Murphy M.E.P. (1994) RASTER3D version 2.0. A program for photor ealistic molecular graphics. Acta Crystallogr. D, 50:869-873. Miyagishi M., Fujii R., Hatta M., Yoshida E., Araya N., Nagafuchi A., Ishihara S ., Nakajima T. and Fukamizu A. (2000) Regulation of Lef-mediated transcription a nd p53-dependent pathway by associating -catenin with CBP/p300. J. Biol. Chem., 275:35170-35175. [Medline] Moon R.T. and Christian J.L. (1989) Microinjection and expression of synthetic m RNAs in Xenopus embryos. Technique, 1:76-89. Moon R.T. and Kimelman D. (1998) From cortical rotation to organizer gene expres sion: toward a molecular explanation of axis specification in Xenopus Bioessays, 20:536-545. [Medline] Morin P.J. (1999) -catenin signaling and cancer. Bioessays, 21:1021-1030. [Medli ne] Nagafuchi A. (2001) Molecular architecture of adherens junctions. Curr. Opin. Ce ll Biol., 13:600-603. [Medline] Navaza J. (1994) AMoRe: an automated package for molecular replacement. Acta Cry stallogr. A, 50:157-163. [Medline] Newport J. and Kirschner M. (1982) A major developmental transition in early Xen opus embryos: I. Characterization and timing of cellular changes at the midblast ula stage. Cell, 30:675-686. [Medline][Summary] Nicholls A., Sharp K.A. and Honig B. (1991) Protein folding and association: ins ights from the interfacial and thermodynamic properties of hydrocarbons. Protein s, 11:281-296. [Medline] Otwinowski Z. and Minor W. (1997) Processing of X-ray diffraction data collected in oscillation mode. In: Carter C.W.J. and Sweet R.M. (Eds.) Methods in Enzymol ogy, Volume 276. New York: Academic Press Parker D.S., Jemison J. and Cadigan K.M. (2002) Pygopus, a nuclear PHD-finger pr otein required for Wingless signaling in Drosophila. Development, 129:2565-2576. [Medline] Peifer M. and Polakis P. (2000) Wnt signaling in oncogenesis and embryogenesis— a look outside the nucleus. Science, 287:1606-1609. [Medline] Pokutta S. and Weis W.I. (2002) The cytoplasmic face of cell contact sites. Curr . Opin. Struct. Biol., 12:255-262. [Medline] Polakis P. (2000) Wnt signaling and cancer. Genes Dev., 14:1837-1851. [Medline] Poy F., Lepourcelet M., Shivdasani R.A. and Eck M.J. (2001) Structure of a human Tcf4--catenin complex. Nat. Struct. Biol., 8:1053-1057. [Medline] Provost E. and Rimm D.L. (1999) Controversies at the cytoplasmic face of the cad herin-based adhesion complex. Curr. Opin. Cell Biol., 11:567-572. [Medline] Reifenberger J., Knobbe C.B., Wolter M., Blaschke B., Schulte K.W., Pietsch T., Ruzicka T. and Reifenberger G. (2002) Molecular genetic analysis of malignant me lanomas for aberrations of the WNT signaling pathway genes CTNNB1, APC, ICAT and BTRC. Int. J. Cancer, 100:549-556. [Medline] Roose J. and Clevers H. (1999) TCF transcription factors: molecular switches in carcinogenesis. Biochim. Biophys. Acta, 1424:M23-M37. [Medline] Roura S., Miravet S., Piedra J., Garcia de Herreros A. and Dunach M. (1999) Regu lation of E-cadherin/catenin association by tyrosine phosphorylation. J. Biol. C hem., 274:36734-36740. [Medline] Rubinfeld B., Souza B., Albert I., Muller O., Chamberlain S.H., Masiarz F.R., Mu nemitsu S. and Polakis P. (1993) Association of the APC gene product with -caten in. Science, 262:1731-1734. [Medline] Seidensticker M.J. and Behrens J. (2000) Biochemical interactions in the wnt pat hway. Biochim. Biophys. Acta, 1495:168-182. [Medline] Sekiya T., Nakamura T., Kazuki Y., Oshimura M., Kohu K., Tago K., Ohwada S. and Akiyama T. (2002) Overexpression of Icat induces G2 arrest and cell death in tum or cell mutants for adenomatous polyposis coli, -catenin, or Axin. Cancer Res., 62:3322-3326. [Medline] Sharpe C., Lawrence N. and Martinez Arias A. (2001) Wnt signalling: a theme with nuclear variations. Bioessays, 23:311-318. [Medline] Shtutman M., Zhurinsky J., Simcha I., Albanese C., D'Amico M., Pestell R. and Be n-Ze'ev A. (1999) The cyclin D1 gene is a target of the -catenin/LEF-1 pathway. Proc. Natl. Acad. Sci. USA, 96:5522-5527. [Medline] Sun Y., Kolligs F.T., Hottiger M.O., Mosavin R., Fearon E.R. and Nabel G.J. (200 0) Regulation of -catenin transformation by the p300 transcriptional coactivator . Proc. Natl. Acad. Sci. USA, 97:12613-12618. [Medline] Tago K., Nakamura T., Nishita M., Hyodo J., Nagai S., Murata Y., Adachi S., Ohwa da S., Morishita Y., Shibuya H. and Akiyama T. (2000) Inhibition of Wnt signalin g by ICAT, a novel -catenin-interacting protein. Genes Dev., 14:1741-1749. [Medl ine] Takemaru K.I. and Moon R.T. (2000) The transcriptional coactivator CBP interacts with -catenin to activate gene expression. J. Cell Biol., 149:249-254. [Medline ] Turner D.L. and Weintraub H. (1994) Expression of achaete-scute homolog 3 in Xen opus embryos converts ectodermal cells to a neural fate. Genes Dev., 8:1434-1447 . [Medline] Tetsu O. and McCormick F. (1999) -catenin regulates expression of cyclin D1 in c olon carcinoma cells. Nature, 398:422-426. [Medline] Thompson B., Townsley F., Rosin-Arbesfeld R., Musisi H. and Bienz M. (2002) A ne w nuclear component of the Wnt signalling pathway. Nat. Cell Biol., 4:367-373. Tutter A.V., Fryer C.J. and Jones K.A. (2001) Chromatin-specific regulation of L EF-1--catenin transcription activation and inhibition in vitro. Genes Dev., 15:3 342-3354. [Medline] van Noort M. and Clevers H. (2002) TCF transcription factors, mediators of Wnt-s ignaling in development and cancer. Dev. Biol., 244:1-8. [Medline] Wodarz A. and Nusse R. (1998) Mechanisms of Wnt signaling in development. Annu. Rev. Cell Dev. Biol., 14:59-88. Yost C., Torres M., Miller J.R., Huang E., Kimelman D. and Moon R.T. (1996) The axis-inducing activity, stability, and subcellular distribution of -catenin is r egulated in Xenopus embryos by glycogen synthase kinase 3. Genes Dev., 10:1443-1