Minireview 1265

The multi-talented b-catenin makes its first appearanceLawrence Shapiro

b-catenin plays a central part in cell adhesion as astructural component of the cadherin complex. In aseemingly disparate role, it is also important in embryopatterning, and now has emerged as a leading actor incarcinogenesis. b-catenin achieves its diverse functions byinteracting with many partners. The recent structure of thecore domain from b-catenin suggests how this talentedmolecule can achieve its many functions.

Address: Structural Biology Program, Mount Sinai School ofMedicine, 1425 Madison Avenue, New York, NY 10029.

E-mail: shapiro@convex.hhmi.columbia.edu

Structure 15 October 1997, 5:1265–1268http://biomednet.com/elecref/0969212600501265

© Current Biology Ltd ISSN 0969-2126

β-catenin was first identified as a structural component ofvertebrate adherens junctions, where it functions in acomplex of proteins that link cadherins, the transmem-brane adhesion receptors, to the cytoskeleton [1]. Theprimary structure of β-catenin revealed ~70% sequenceidentity to the armadillo protein from Drosophila, which isknown to function in embryo patterning and cell-fatedecisions [2]. Armadillo and β-catenin turn out to be truecross-species orthologues that serve the same functions inflies and vertebrates — both have a function in cell adhe-sion as structural components of adherens junctions, andboth proteins are important in signal transduction path-ways that powerfully mediate cell-fate decisions. It hasnow been shown that mutations affecting the cellular levelof β-catenin are associated with two of the most commonforms of human cancer, melanoma [3] and colorectalcancer [4,5]. The crystal structure of the core domain fromβ-catenin recently reported by Huber et al. gives clues onhow β-catenin functions in these diverse roles [6].

In its cell-adhesion role, β-catenin helps link cadherins tothe actin cytoskeleton [7]; it binds directly to the cytoplas-mic tail of cadherin molecules, and simultaneously bindsto the unrelated protein α-catenin, which is in turn linkedto actin filaments (Figure 1). Catenin-mediated linkage tothe cytoskeleton appears to be crucial to the function ofadherens junctions; cadherins cannot mediate cell adhe-sion in cells that lack either α- or β-catenin.

In its signaling role [8,9], β-catenin/armadillo is involved inthe Wnt/wingless growth factor signaling pathway (wing-less is the Drosophila homologue of Wnt). In Drosophila,this pathway mediates anterior/posterior patterning withineach segment of the fly embryo [2]. In vertebrates, the

Wnt pathway is important in the induction of thedorsal/ventral and anterior/posterior embryonal axes. Overor under expression of β-catenin can produce stunningphenotypes in vertebrate embryos [10], for example, axisduplication — two headed frogs can result when Xenopuslaevis embryos are injected with β-catenin or its corre-sponding mRNA (Figure 2).

Elements of the Wnt pathway are summarized in Figure 1.Normally, β-catenin is targeted for degradation by the ade-nomatous polyposis coli protein (APC; [4]). APC, of course,is the infamous ‘colon cancer gene’. The interactionbetween β-catenin and APC is mediated by glycogen syn-thase kinase-3β (GSK-3; Zest-white-3 in Drosophila) [11]. Itis thought that phosphorylation of APC by GSK-3 enhancesbinding to β-catenin, activating its degradation [4]. Theactivity of GSK-3 is modulated by the Wnt signal, which istransmitted through the seven-transmembrane helix recep-tor frizzled [12] and another intermediary of less wellunderstood mechanism, disheveled. The take-homemessage is that Wnt signaling is effected through the stabi-lization of a cytoplasmic pool of β-catenin. When a cytoplas-mic pool of β-catenin has formed, free β-catenin moleculescan bind to members of the Tcf/LEF-1 family of transcrip-tion factors [13], and enter the nucleus to activate transcrip-tion of target genes. Tcf proteins are ‘architectural’transcription factors that activate transcription by inducingbending of DNA; binding of β-catenin to LEF-1 has beenshown to substantially alter the ability of this transcriptionfactor to bend DNA [13].

So now, β-catenin’s connection with colon cancer seemsclear. APC functions to down-regulate β-catenin by target-ing it for degradation. When APC is missing or nullified bymutation, β-catenin levels rise, and downstream elementsof the Wnt pathway are activated — out of context, andthis can lead to malignant transformation. β-cateninappears to play the same role in some melanomas [3], andtruncated β-catenin transcripts have also been observed inthe cells of some stomach cancers [14]. Thus, it seemspossible that β-catenin may have a broad role in carcino-genesis, and may act in other types of cancer as well.

β-catenin and armadillo are each composed of an N-termi-nal head domain of about 130 amino acids followed bytwelve tandem imperfect sequence repeats, each of about42 amino acids (called armadillo or arm repeats, also foundin many other cytoplasmic proteins) and a C-terminal tailof about 100 amino acids [1]. The three regions have dis-tinctive charge distributions; the head and tail regions areacidic, whereas the arm repeat region is highly basic.

Huber et al. [6] subjected β-catenin to limited proteolysisand found that a stable nicked fragment was producedthat approximately corresponds to the arm repeat region.An analogous recombinant protein termed β59 was con-structed, and the crystal structure of this 538 residueprotein was determined by MAD analysis of the seleme-thionine-substituted molecule.

It is this region of the molecule composed of 12 armadillorepeats, encompassed by β59, that mediates the inter-action of β-catenin with most of its known binding part-ners. Deletion mutagenesis has mapped the binding sitesfor E-cadherin, LEF-1, and APC to β-catenin’s corearmadillo repeat region. Notably, binding to E-cadherinand APC seems to be mutually exclusive, and β-cateninbound to cadherins in adherens junctions should be there-fore inaccessible to APC-mediated degradation.

β59, which includes the twelve arm repeats, forms a rod-like structure composed entirely of α helices with short

1266 Structure 1997, Vol 5 No 10

Figure 2

The morphogenic power of β-catenin is illustrated by the effect ofinjecting its mRNA into Xenopus embryos. This embryo was injectedventrally, earlier in development during early cleavage stage. β-catenin can cause dorsalization of cells that would have otherwiseassumed a ventral fate. The result here is duplication of theembryonic axis — the most striking feature being the growth of twoheads.

Figure 1

β-catenin function in adhesion and signaling.β-catenin is a structural component of thecadherin cell adhesion complex, and alsofunctions in the Wnt growth factor signalingpathway. In cell adhesion, β-catenin helps linkcadherin adhesion molecules to cytoskeletalactin filaments. In its signal transduction role,β-catenin functions as a transcriptionalco-activator of target genes involved in celldifferentiation and proliferation. The Wnt signalregulates cytoplasmic levels of β-catenin; onlywhen levels are high can β-catenin translocateto the nucleus to activate downstreameffectors. See text for details.

Frizzled

Wnt

Cadherins

β-catenin

α-catenin

Actin fibers

β-catenin

Destruction of β-catenin(in the absence of Wnt)

APC

GSK

Cytosol

Nucleus

β-catenin

LEF Transcription

Dsh

connecting loops between them (Figure 3). As was sug-gested by the proteolysis studies, the arm repeats are not aseries of domains like beads on a string. Rather, they forma single domain with a continuous hydrophobic coreformed by interactions between adjacent arm repeats.Each arm repeat is made up of three helices, H1, H2 andH3. The arm repeats pack together to form a remarkably

regular right-handed superhelix of helices. This regularstructure is created by a canonical ‘3 in 4’ knobs into holespacking, in which every third sidechain of one helixinserts into the grooves created by every fourth sidechainof another. This leads to an average 30° rotation and 10 Åtranslation per arm repeat. Although deviations in thecanonical topology are found in repeats 7 and 10, the

Minireview b-catenin structure Shapiro 1267

Figure 3

The β59 structure determined by Huber et al.[6], which includes the entire β-cateninarmadillo repeat region, is shown as atopology (a) and a ribbon diagram (b).Helices H1, H2 and H3 are colored blue, redand green, respectively. The twelve armadillorepeat regions are numbered accordingly. TheH1 helix from the first arm repeat isdisordered in the β59 crystal structure. TheH3 helices (green) of the first ten arm repeatsdefine a putative binding groove for theβ-catenin binding partners, APC, Tcf familytranscription factors and cadherins.(Reproduced from [6] with kind permission ofthe authors.)

1 2 3 4 5 6 7 8 9 10 11 12

COOH

NH2

Figure 4

Electrostatic surface of β59. The three viewsare related by successive 90° rotations aboutthe vertical axis. The long positively charged(blue) groove is the proposed binding site formany β-catenin binding partners. Mutagenesisstudies have implicated negatively chargedregions of these molecules in their interactionwith β-catenin. (Reproduced from [6] withkind permission of the authors.)

superhelix is substantially diverted only in the vicinity ofrepeats 8 and 9, where a 50° bend creates a kink in themolecule.

H2 helices from adjacent repeats pack in parallel with oneanother, as do H3 helices, thus forming a two-sided spiral-ing ribbon with each side of the ribbon connected byorthogonal short H1 helices. The H3 side of the spiralingribbon creates a long shallow groove, about 95 Å long and20 Å wide, which runs along the first ten arm repeats. Thisgroove contains a large band of positively charged residues(Figure 4). Huber et al. propose that this positivelycharged groove might contain the binding sites for cad-herins, APC and the Tcf family transcription factors.

Several lines of evidence converge to suggest that chargecomplementarity is important in the interaction ofβ-catenin with its partners. Although the β-catenin-binding sites on cadherins, APC and Tcf proteins areunrelated in sequence, all are highly acidic, and deletionmutagenesis studies show that the number of arm repeatsroughly correlates with the strength of β-catenin bindingto these molecules (see references in [6]). Finally, phos-phorylation of the β-catenin-binding site of APC, whichwould increase its negative charge, enhances its inter-action with β-catenin [4]. All of these observations arguethe importance of charge complementarity over a largesurface. This is also quite satisfying as an explanation ofhow β-catenin can bind to its multitude of structurallydiverse partners.

The 95 Å H3 groove could accommodate an extendedpolypeptide of 25–30 amino acids, or larger polypeptidesin other conformations. Ultimately, co-crystal structures ofβ-catenin in complex with its binding partners will beneeded to understand the atomic-level function ofβ-catenin in its many roles with its many partners.

AcknowledgementsThanks are due to François Fagotto, Ursula Gluck and Barry Gumbiner forproviding the Xenopus embryo figure, and to Andrew Huber and Bill Weisfor the figures of the β-catenin structure.

References1. McCrea, P.D., Turck, C.W. & Gumbiner, B. (1991). A homolog of the

armadillo protein in Drosophila (plakoglobin) associated withE-cadherin. Science 254, 1359–1361.

2. Wieschaus, E. & Riggleman, R. (1987). Autonomous requirements forthe segment polarity gene armadillo during Drosophilaembryogenesis. Cell 49, 177–184.

3. Rubinfeld, B., Robbins, P., El-Gamil, M., Albert, I., Porfiri, E. & Polakis,P. (1997). Stabilization of beta-catenin by genetic defects inmelanoma cell lines. Science 275, 1790–1792.

4. Rubinfeld, B., Robbins, P., El-Gamil, M., Albert, I., Porfiri, E. & Polakis,P. (1996). Binding of gsk3β to the APC—catenin complex andregulation of complex assembly. Science 272, 1023–1026.

5. Su, L.-K., Vogelstein, B. & Kinzler, K.W. (1993). Association of theAPC tumor suppressor protein with catenins. Science 262,1734–1737.

6. Huber, A.H., Nelson, W.J. & Weis, W.I. (1997). Three-dimensionalstructure of the armadillo repeat region of β-catenin. Cell 90,871–881

7. Aberle, H., Butz, S., Stappert, J., Weissig, H., Kemler, R. &Hoschuetzky, H. (1994). Assembly of the cadherin-catenin complex invitro with recombinant proteins. J. Cell Sci. 107, 3655–3663.

8. Gumbiner, B.M. (1995). Signal transduction of beta-catenin. Curr.Opin. Cell Biol. 7, 634–640.

9. Miller, J.R. & Moon, R.T. (1996). Signal transduction through beta-catenin and specification of cell fate during embryogenesis. GenesDev. 10, 2527–2539.

10. McCrea, P.D., Brieher, W.M. & Gumbiner, B.M. (1993). Induction of asecondary body axis in Xenopus by antibodies to beta-catenin. J. CellBiol. 123, 477–484.

11. Peifer, M., Sweeton, D., Casey, M. & Wieschaus, E. (1994). winglesssignal and Zeste-white 3 kinase trigger opposing changes in theintracellular distribution of Armadillo. Development 120, 369–380.

12. Bhanot, P., et al., & Nusse, R. (1996). A new member of the frizzledfamily from Drosophila functions as a wingless receptor. Nature 382,225–230.

13. Behrens, J., et al., & Birchmeier, W. (1996). Functional interaction ofβ-catenin with the transcription factor LEF-1. Nature 382, 638–642.

14. Oyama, T., et. al., Hirohashi, S. (1994). A truncated beta-catenindisrupts the interaction between E-cadherin and alpha-catenin: acause of loss of intercellular adhesiveness in human cancer cell lines.Cancer Res. 54, 6282–6287.

1268 Structure 1997, Vol 5 No 10