Protein Interactions

The yeast two-hybrid system and GST pull-down assays were used to examine the possibility that Lgs physically interacts with Armadillo (Arm). Indeed, the N-terminal half of Lgs binds to the Arm protein. The Arm binding domain of Lgs was subsequently fine-mapped to the HD2 region and the Lgs binding domain of Arm to armadillo repeats 1-4. Consistent with these results, BCL9 also binds to Arm, as well as to ß-catenin, and the domain required for the interactions with ß-catenin again maps precisely to the conserved HD2 sequence (Kramps, 2002).

Since lgs17P and lgs17E encode amino acid substitutions in HD2, whether their protein products can bind to Arm protein in vitro was tested. The binding of Lgs17P and Lgs17E to Arm is reduced at least 10-fold compared to wild-type Lgs protein. This finding, which reinforces the observations of genetic interactions between arm and lgs alleles, is interpreted as evidence that Wnt/Wg signal transduction normally depends on molecular interactions between the Lgs/BCL9 and Arm/ß-catenin proteins (Kramps, 2002).

A mutant form of Lgs protein from which HD2 was deleted exhibits a strong dominant-negative effect on Wg-dependent patterning processes when expressed from a transgene in wild-type larvae. This observation was taken as an indication that Lgs normally interacts not only with Arm, but also with at least one additional component. In an effort to identify such components, yeast two-hybrid screens were carried out for interacting proteins. In two independent screens in which either the entire protein or the N-terminal half of Lgs was used as a bait, a novel PHD finger protein was identified as a Lgs binding protein. The 815 amino acid residue Pygo protein carries a C-terminal domain of 60 amino acids that shows extensive homologies to the PHD (plant homology domain) finger, also known as LAP (leukemia-associated protein) domain. This domain comprises a cysteine-rich Zn binding motif, which has been associated with proteins involved in chromatin-mediated regulation of transcription. The PHD finger of Pygo is necessary and sufficient to mediate the interaction with Lgs. The region of Lgs responsible for Pygo binding to the HD1 sequence was mapped. Moreover, two human homologs of the Drosophila pygo gene were identified and isolated. The protein products of both human genes, hPYGO1 and hPYGO2, possess a highly conserved PHD finger that interacts with the HD1 of BCL9. The only other domain in Drosophila Pygo and hPYGO1/hPYGO2 that shows significant sequence homology is a 50 amino acid stretch in the N-terminal region, referred to as 'N-terminal homology domain' (NHD) (Kramps, 2002).

The interaction with Pygo appears to be relevant for the in vivo function of Lgs, since a mutant form of Lgs with a deletion of HD1 was unable to rescue lgs20F mutant animals. Consistent with this finding, epitope-tagged forms of Pygo were observed to localize to the nuclei of imaginal disc cells (Kramps, 2002).

The physical association of Pygo and Lgs suggests that Pygo, like Lgs, may be required for Wg signaling in vivo. To explore this hypothesis, a collection of suppressors of the sev-wg phenotype was searched for mutations that map to the tip of the right arm of chromosome 3, the position of the pygo gene. One such suppressor, Sup130, mapped to this position and intriguingly, it shows dominant lethality in combination with the lgs allele lgs17E (Sup130/+ lgs17E/+ transheterozygous animals do not survive). The pygo-coding region was sequenced using genomic DNA from homozygous Sup130 mutant larvae and a 14 bp deletion was identified starting at amino acid position 751. Hence, this allele is referred to as pygo130 and it encodes a truncated Pygo protein lacking the C-terminal PHD finger. The lethality caused by the homozygous pygo130 genotype can be fully rescued by a tubulinalpha1 promoter-driven transgene that contains either the coding region of the Drosophila pygo gene or that of one of its two human homologs: hPYGO1 or hPYGO2 (Kramps, 2002).

To assay the possible role of Pygo in Wg signal transduction during development, embryos homozyous for the pygo130 mutation were generated that derived from female germ cells equally mutant for pygo. Such embryos are devoid of any wild-type Pygo activity and die with a severe segment polarity phenotype. Mutant individuals lacking only zygotic function survive until early pupal stages and exhibit imaginal discs that are abnormally small. The Hh target gene ptc was expressed at wild-type levels in these discs; however, no expression of the Wg target Dll could be detected. These discs appear to lack the presumptive wing blade field and possess two primordia for the notum. From these results it is concluded that pygo, like arm, pan, and lgs, is required for Wg signal transduction in vivo (Kramps, 2002).

It can be concluded that Lgs and Pygo are required for the signaling activity of Arm, and that this function depends on the ability of Lgs to interact molecularly with Arm and on the ability of Pygo to molecularly interact with Lgs. Based on their subcellular localization and epistatic relationship with ArmS10, Lgs and Pygo are unlikely to exert their function by impeding proteasome-mediated degradation of Arm. They could play a role in permitting nuclear import or preventing nuclear export of Arm. However, this is thought to also be unlikely, because no difference is detected in subcellular localization of Arm in lgs mutant embryos. An alternative possibility is that Lgs and BCL9, respectively, function to tie Pygo to the ß-catenin-TCF complex, perhaps to allow Pygo to activate and sustain the expression of Wnt target genes. This hypothesis raises several predictions. (1) This model implies that the main task of Lgs/BCL9 is to serve as an adaptor to tether Pygo to Arm/ß-catenin. Thus, most of the Lgs/BCL9 protein should be dispensable, as long as Lgs HD1 is covalently linked to HD2, allowing the formation of a bridge between Arm/ß-catenin and Pygo. Lgs might even be entirely superfluous if Pygo is endowed with the ability to directly bind ß-catenin. (2) This model would require that Arm/ß-catenin is able to bind simultaneously, and stably, with both Pan/TCF and Lgs/BCL9. This in turn would necessitate separate binding sites on Arm/ß-catenin for Pan/TCF and Lgs/BCL9. (3) Pygo proteins would have to possess the ability to stimulate transcription when recruited to promoters of Wnt target genes (Kramps, 2002).

Evidence is provided that Lgs function critically depends on the integrity of its HD1 and HD2 domains. However, this study does not address the role of HD3, nor that of any other part of the protein. A key result was obtained with a transgene encoding a truncated Lgs protein (amino acids 1-583) that lacks HD3 and all sequences C-terminal to HD3. This transgene fully rescues homozygous lgs20F animals. It was next tested if HD1 and HD2 are sufficient for Lgs function and a 150 amino acid stretch of the unrelated nuclear yeast protein Gal11 was used to join HD1 and HD2. Surprisingly, expression of this hybrid protein fully rescues patterning and growth of lgs20F null mutant animals. Although some of these animals failed to hatch from their pupal cases, more than 50% did hatch, and hence represent fully rescued animals. An even smaller protein, consisting only of HD1 and HD2 connected with three copies of the HA tag, could also substitute for Lgs protein at a rescuing efficiency only slightly lower than that of the HD1-Gal11-HD2 protein. Finally, it was asked whether even HD1 might be dispensable if Pygo was directly equipped with HD2 and hence with the ability to bind to Arm. A tubulinalpha1 promoter-driven transgene was constructed in which the PHD finger of Pygo was replaced by the HD2 domain of Lgs. Remarkably, this fusion protein rescues both lgs20F as well as pygo130 mutant animals. Thus, it is concluded that the PHD finger has no other essential role besides binding Pygo to Lgs, and that the primary, and possibly sole, function of Lgs is the recruitment of Pygo to Arm (Kramps, 2002).

The primary structure of Arm and its mammalian homolog ß-catenin consists of an N-terminal and a C-terminal tail flanking a central domain of ~500 residues composed of 12 armadillo repeats. These repeats pack against one another to form a superhelix that features a positively charged groove. The armadillo repeat domain mediates the binding of ß-catenin to cadherins, APC, Axin, and TCF. Despite their lack of significant sequence homologies, these proteins bind competitively to ß-catenin, presumably because they contact the same surface area of ß-catenin. If Lgs/BCL9 also binds to this surface, it would be expected to compete with Pan/TCF for the interaction with ß-catenin and could not be recruited to Wnt target genes. This issue was addressed by using peptide competition and coimmunoprecipitation experiments (Kramps, 2002).

Biotinylated peptides representing the N-terminal domains of hTCF4 (or Pan) were used to pull down labeled ß-catenin (or Arm protein) with avidin beads. This peptide-protein interaction was effectively disrupted by an excess of nonbiotinylated TCF or Pan peptides, but not by an excess of HD2 peptides. GST-BCL9 protein was used to pull down labeled hTCF4 in the presence of ß-catenin protein. hTCF4 is efficiently retained on BCL9-charged glutathion beads in the presence, but not absence, of ß-catenin, which apparently can function as a bridge between the two proteins. These results indicate that Lgs/BCL9 and Pan/TCF do not compete for their interaction with Arm/ß-catenin, but rather bind it simultaneously. This in turn suggests that Lgs/BCL9 is recruited to TCF binding sites of Wnt target genes (Kramps, 2002).

To address the role of Pygo in ß-catenin-mediated transcription, a TCF reporter gene (TOP-Flash) was used in immortalized human embryo kidney cells (HEK 293 cells). Low levels of a stable mutant form of ß-catenin (DeltaN-ß-catenin) were introduced into these cells to partially stimulate the pathway. The additional expression of hPYGO1 or hPYGO2 leads to a large increase in luciferase activity (30-fold). These levels are significantly higher than the sum of those produced by either treatment alone. This potentiation of ß-catenin activity by hPYGO1 and 2 appears to be mediated by the interaction of endogenous TCF protein with its DNA target sites, as it is only observed with TOP-Flash, which contains five optimal TCF binding sites, but not with the control reporter FOP-Flash, which contains five mutated sites (Kramps, 2002).

Although less powerful per se than the genetic arguments, this experiment adds supportive evidence to the notion that Pygo proteins transduce Wnt signals by activating TCF target genes in a ß-catenin-dependent manner (Kramps, 2002).

Identification and in vivo role of the Armadillo-Legless interaction

The Wnt signalling system controls many fundamental processes during animal development and its deregulation has been causally linked to colorectal cancer. Transduction of Wnt signals entails the association of ß-catenin with nuclear TCF DNA-binding factors and the subsequent activation of target genes. Using genetic assays in Drosophila, a presumptive adaptor protein, Legless (Lgs), has been identifed that binds to ß-catenin and mediates signalling activity by recruiting the transcriptional activator Pygopus (Pygo). This study characterizes the β-catenin/Lgs interaction and shows that: (1) it is critically dependent on two acidic amino acid residues in the first Armadillo repeat of ß-catenin; (2) it is spatially and functionally separable from the binding sites for TCF factors, APC and E-cadherin; (3) it is required in endogenous as well as constitutively active forms of ß-catenin for Wingless signalling output in Drosophila, and (4) in its absence animals develop with the same phenotypic consequences as animals lacking Lgs altogether. Based on these findings, and because Lgs and Pygo have human homologues that can substitute for their Drosophila counterparts, it is inferred that the ß-catenin/Lgs binding site may thus serve as an attractive drug target for therapeutic intervention in ß-catenin-dependent cancer progression (Hoffmans, 2004).

This study is concerned with the question of how ß-catenin and Lgs interact molecularly with each other. The analysis addressed three issues: localization of the binding site on ß-catenin, specificity of this site vis-a-vis other partners of ß-catenin and in vivo significance of this interaction for Wg signal transduction. By means of site-directed mutagenesis the role of conspicuous ß-catenin residues in the binding to human LGS1 was examined. Two amino acids, D162 and D164, were identified that are both necessary for human LGS1 binding. Because substitutions of these residues with other amino acids did not affect the binding of several other proteins to ß-catenin, the role of these amino acids is interpreted as contact sites for human LGS1, rather than a structural function enhancing stability and/or three-dimensional conformation of ß-catenin. This conclusion, however, will need to be confirmed by determining the crystal structure of the ß-catenin/human LGS1 complex (Hoffmans, 2004).

Neither D162 nor D164 is required for binding to APC, E-cadherin or TCF4. Substitutions of these amino acids reduce binding to alpha-catenin twofold, but in vivo data suggest that this reduction does not prevent the assembly of adherens junctions. The specificity of the ß-catenin/human LGS1 interaction vis-a-vis that of ß-catenin and APC, E-cadherin or TCF4 is consistent with their respective locations on the surface of ß-catenin. While crystallographic studies show that APC, E-cadherin and TCF4 all bind to a common, extended surface within the groove of ß-catenin formed by Arm repeats 3-10, this analysis indicates that human LGS1 binds an acidic knob in Arm repeat 1. This knob is not only located more N terminally, it is also situated on the side of ß-catenin, which is opposite the groove. The spatial separation of these binding sites is in agreement with their separable functions observed in yeast binding assays, as well as with previous GST pull-down assays, in which simultaneous binding of TCF4 and human LGS1 to ß-catenin is observed (Kramps et al., 2002).

Thus, to assess the role of D162 and D164 in Wg transduction, mutant forms of Arm were subjected to various assays designed to reveal their in vivo function. Simple rescue and overexpression experiments have shown that transgenic Arm-D164A cannot substitute for endogenous Arm, and that the D164A mutation significantly reduces the constitutive signalling activity associated with N-terminal deletions of Arm. When tested in more advanced assays, it was found that D164 is required by wing disc cells to maintain Wg target gene expression and by developing embryos for segmentation. Together, these experiments support the conclusion that Arm signalling function relies on its capability to bind to Lgs throughout development (Hoffmans, 2004).

Although it is straightforward to interpret the results as a qualitative indication for the significance of the Arm/Lgs interaction, it is more difficult to assess their outcome in a quantitative manner. For example, the apparent residual expression of Dll in Arm-D164A cells may reflect perdurance of wild-type Arm or Dll proteins, but it could also indicate that a fraction of the Wg signal can be transmitted despite the D164A mutation. This latter scenario could in turn be attributed to some residual binding of Arm to Lgs, but it could also be explained by a partial redundancy of Lgs function. Lgs may be required for efficient Arm-mediated activation of Wg targets, but some activation may also occur in its absence. Consistent with this latter view, it was observed that animals lacking maternal and zygotic lgs product exhibit phenotypes equivalent to animals in which the sole source of Arm is the D164A transgene, yet neither of the two phenotypes are quite as severe as that of wg-null mutants (Hoffmans, 2004).

The Wnt pathway is highly conserved between Drosophila and vertebrates. The human homologues of Lgs (LGS1/BCL9) and Pygo (PYGO1 and PYGO2) can rescue lgs and pygo mutant flies, respectively. This suggests that these proteins have the same function in vertebrates and in Drosophila. It is possible therefore, that in vivo data can be extrapolated to Wnt signalling in mammals. Mutations in APC occur in more than 80% of inherited and sporadic colorectal cancers. These mutations lead to accumulation of free ß-catenin and as a result to overexpression of Wnt target genes. A chemical compound that interferes with the formation of the nuclear TCF/ß-catenin/Lgs/Pygo complex should in theory halt the progression of cancer. Such an anti-cancer drug must be highly specific though, since it should only disrupt the nuclear ß-catenin complex, but should not disrupt either the cytoplasmic ß-catenin/APC/Axin complex or the ß-catenin/E-cadherin complex at the cell membrane. APC, Axin and E-cadherin functions should not be compromised, since all three of them have tumour suppressor roles. This is not the case, however, for TCF and Lgs. Crystal structure data indicates that APC, Axin, E-cadherin and TCF4 partly use the same contact sites of ß-catenin for their binding. Therefore, designing an inhibitor that specifically disrupts the ß-catenin/TCF interaction is a difficult task. On the contrary, mapping and specificity results indicate that the ß-catenin/Lgs interaction site could be targeted without interfering with the binding of ß-catenin to APC and E-cadherin. Moreover, this analysis shows that genetic disruption of the Arm/Lgs interaction leads to severely reduced Wg signalling, suggesting that the protein-protein interaction between ß-catenin and Lgs may provide an attractive target for therapeutic intervention (Hoffmans, 2004).

A complex of Armadillo, Legless, and Pygopus coactivates dTCF to activate Wingless target genes

Upon receiving a Wnt signal, cells accumulate ß-catenin (Armadillo in Drosophila), which binds directly to TCF transcription factors, leading to the transcription of Wnt target genes. It is generally thought that ß-catenin/Armadillo is a transcriptional coactivator when bound to TCF in the nucleus and that this function is mediated by its C terminus. However, recent findings in Drosophila indicated that Armadillo may activate dTCF in the cytoplasm. This study reexamines the mechanism of Armadillo's signaling function in light of Legless and Pygopus, two nuclear factors recently discovered to be essential for this function. Armadillo, in order to activate dTCF, must enter the nucleus and form a complex with Legless and Pygopus. The ability of this complex to stimulate TCF-mediated transcription can be altered by linkage of a strong transcriptional activator or repressor to Armadillo. Furthermore, Armadillo is a strong transcriptional activator when fused to the yeast GAL4 DNA binding domain -- an activity that depends on regions of the Armadillo repeat domain that mediates binding to Legless and to chromatin modifying and remodeling factors. Finally, linkage of the N-terminal region of Pygopus, but not the C terminus of Armadillo, to dominant-negative dTCF can restore its signaling activity in transgenic flies. This evidence argues in favor of a revised coactivator factor model in which Armadillo's coactivator function depends on regions within its Armadillo repeat domain to which Legless/Pygopus and other transcriptional coactivators can bind. In contrast, the C terminus of Armadillo plays a less direct role in this function (Thompson, 2004).

The model that Arm functions in the nucleus as a transcriptional activator of dTCF clearly predicts that exclusion of Arm from the nucleus by tethering to membranes should render it unable to signal. Two such nuclear-excluded, membrane-tethered forms of Arm have been examined in Drosophila: Sev-Arm, a fusion of the extracellular and transmembrane domains of Sevenless to Arm's N terminus, and Arm-CAAX, which features a CAAX-type palmitoylation sequence at its C terminus. The signaling activity of Arm transgenes can be measured by examining their ability to rescue Drosophila embryos that are maternally and zygotically mutant (henceforth: mutant) for arm. A severe impediment to this analysis is that arm null mutants (eg: armXP33 and arm4 also called armYD35) have adhesion defects in addition to defective Wingless signaling and, consequently, do not develop beyond oogenesis. Thus, mutant conditions that affect signaling, but not adhesion, must be used. The most commonly used signaling-mutant (but adhesion-competent) allele is armXM19, a truncation of the Arm C terminus that generates embryos with defective Wingless signaling (Thompson, 2004).

Surprisingly, both Sev-Arm and Arm-CAAX were reported to substantially rescue Wingless signaling in armXM19 mutants. The two possible interpretations of these results are (1) that these proteins signal independently of endogenous Arm and (2) that the ArmXM19 mutant protein can be induced to signal in the presence of these transgenes. Discrimination between these two possibilities requires examination of these transgenes in alternative arm mutant backgrounds. In the case of Arm-CAAX it was possible to use an effectively null mutant, armXP33 (which does not express detectable Arm protein), because Arm-CAAX is able to function in adhesion. Arm-CAAX was found to rescue the adhesion, but not the signaling defect of armXP33. In the case of Sev-Arm, analysis in a null mutant background is not possible because this transgene is not competent to rescue the adhesion defect. Attempts were made with arm043A01, an allele that produces both signaling and mild adhesion defects, but the results are unclear, because mutant embryos do not secrete a cuticle. Therefore, alternative mutant conditions were generated by expressing signaling-mutant (but adhesion-competent) Arm transgenes, ArmS6 and ArmS12, in an arm4 null-mutant background. These conditions (henceforth: ArmS6 and ArmS12 mutants) generated embryos whose cuticle phenotype was a lawn of denticles, indicating that Wingless signaling was inactive. Ubiquitous expression of Sev-Arm with the Gal4-UAS system was unable to rescue the Wingless-signaling defect of these embryos, whereas similar ubiquitous expression of Sev-Arm was able to rescue the cuticular phenotype of armXM19 mutants considerably. Similarly, as a control, an activated form of Armadillo, ArmS10, was able to rescue all three signaling-mutant conditions. It is concluded that Sev-Arm, like Arm-CAAX, is unable to signal in the absence of functional endogenous Arm and that the C-terminally truncated ArmXM19 protein retains significant signaling activity that is revealed by the expression of membrane-tethered forms of Arm (Thompson, 2004).

In addition to dTCF, two other ubiquitous factors, Legless (Lgs) and Pygopus (Pygo), are essential for Arm's signaling activity in Drosophila. In lgs or pygo mutants, Arm is unable to signal, even when it accumulates at unusually high levels throughout the cell. The localization of these proteins (either the endogenous protein or epitope-tagged versions expressed from a transgene) were examined in the embryonic epidermis where high levels of Wingless induce accumulation of Arm in stripes of cells. dTCF, Lgs, and Pygo are predominantly nuclear in all cells regardless of their state of signaling. Notably, no evidence was found for nuclear export of tagged, expressed dTCF in response to Wingless in the embryonic epidermis (Thompson, 2004).

Although genetic analysis of Lgs and Pygo has demonstrated that they are essential for Arm's signaling activity, it remains possible that these proteins simply provide an essential function for dTCF. No evidence was found that dTCF stability or localization are affected in pygo mutants. Note that both Lgs and Pygo function are compromised in pygo mutants, because Lgs depends on Pygo for its nuclear localization (F.M. Townsley, A. Cliffe, and M. Bienz, unpublished data, cited in Thompson, 2004). If Lgs and Pygo provide an essential function for Arm rather than dTCF, then providing dTCF with a strong transcriptional activator should bypass the requirement for Lgs and Pygo. A fusion protein of dTCF with the VP16 transcriptional activation domain (dTCF-VP16) that had been shown to rescue armXM19 mutants was therefore expressed in wild-type and pygo mutant embryos with the GAL4-UAS system. Unfortunately, expression of dTCF-VP16 arrests embryogenesis prior to cuticular differentiation. Therefore the expression of the engrailed gene, a target of Wingless signaling in the embryo that is downregulated in pygo mutants was examined. Expression of dTCF-VP16 is able to restore engrailed expression in these embryos. It is concluded that Lgs and Pygo are not required for dTCF's stability, localization, or DNA binding activity but, rather, for activation of dTCF by Arm (Thompson, 2004).

Consistent with this view, in vitro binding experiments have shown that the Lgs HD2 domain binds directly to the first four Armadillo repeats of Arm, while the Lgs HD1 domain binds to the PHD domain of Pygo. On this basis, it was proposed that Arm, Lgs, and Pygo may form a complex in vivo. To test this proposal, an HA-tagged version of Pygo was expressed in Drosophila embryos that also expressed Wingless to activate signaling in all cells. The tagged Pygo was immunoprecipitated with αHA antibodies. Both Arm and Lgs were found to be readily coimmunoprecipitated from embryos expressing HAPygo, but not from control embryos. It is concluded that Arm, Lgs, and Pygo form a nuclear complex in Wingless-stimulated cells in vivo. These findings strongly support the view that Arm activates dTCF in the nucleus, since Lgs and Pygo, two binding partners for Arm that are essential for this process, are nuclear proteins (Thompson, 2004).

The the Arm/dTCF transcription factor model of Wingless signal transduction need to be reconsidered in light of the discovery of Legless and Pygopus. This model was originally prompted by the findings that (1) activation of dTCF depends on a direct binding interaction with Arm; (2) TCF transcription factors are constitutively localized to the nucleus, whereas Arm enters the nucleus only upon signaling; and (3) the C terminus of Arm, which is absent in armXM19 mutants, can function as a transcriptional activator when tethered to DNA (Thompson, 2004).

The model for Arm function predicts that Arm must enter the nucleus in order to form an active transcription factor with dTCF on DNA. The results of this study show that membrane-tethered forms of Arm cannot directly activate dTCF, supporting the notion that Arm must enter the nucleus to do so. The ability of membrane-tethered Arm to signal in an armXM19 mutant background must therefore reflect that the armXM19 mutation is not a null and must retain some signaling activity that is enhanced by the presence of membrane-tethered Arm. A plausible explanation for this phenomenon is that membrane-tethered Arm recruits negative regulators of Arm, thereby stabilizing and/or promoting nuclear translocation of endogenous Arm. In support of this explanation, effects of this kind have, in fact, been observed with several different types of membrane-targeted Arm and β-catenin. Consideration of these results reveals a point of conflict with the original form of the Arm/dTCF transcription factor model, which proposes that the Arm C terminus is necessary and sufficient for Arm's coactivator function. The armXM19 mutation encodes an Arm protein that lacks its C terminus. If this truncated protein retains some signaling activity, then the C terminus cannot be the sole mediator of Arm's coactivator function. In support of this view, several different C-terminally truncated Arm and β-catenins appear to retain significant signaling activity under conditions of overexpression. Furthermore, Arm's C terminus can be substituted without loss of function by the C terminus of a different Armadillo repeat domain protein, Pendulin. Unlike the Arm C terminus, the Pendulin C terminus lacks transactivating activity when fused to the GAL4 DNA binding domain. It is concluded that the C terminus is not sufficient to mediate Arm's coactivator function but instead, is likely to be required in some way for the stability or activity of the Armadillo repeat domain. These findings undermine one block of evidence upon which the Arm coactivator model was originally founded (Thompson, 2004).

Evidence was therefore sought that Arm functions as a transcriptional activator. Arm's ability to activate TCF-mediated transcription, as measured in the Topflash assay, is enhanced by addition of a strong transcriptional activator and reduced by addition of a strong transcriptional repressor. Tethering of Arm to DNA with the GAL4 DNA binding domain reveals that Arm functions as a strong transcriptional activator. Furthermore, this activity of Arm was suppressed by mutations in the Armadillo repeat domain (S6 and S12) that prevent Arm from transducing Wingless signals in vivo. The results indicate that Arm indeed functions as a coactivator and that this function depends on regions in the Armadillo repeat domain that may recruit additional coactivating factors (Thompson, 2004).

Two candidates that may mediate Arm's coactivator function are Lgs and Pygo. Lgs and Pygo are constitutively nuclear proteins that bind to the Armadillo repeat domain upon signaling and are essential for Arm to activate dTCF. Furthermore, Lgs and Pygo appear to be present in the coactivator complex. The N terminus of Pygo (PygoΔPHD) is sufficient to mediate the function of Lgs and Pygo in Wingless signaling when targeted to Arm by fusion to the Lgs HD2 domain. The same region of Pygo has the capacity to function as a transcriptional activator and, when fused to dTCF, can partially bypass the requirement for Armadillo in Wingless signal transduction (Thompson, 2004).

The results argue that Lgs and Pygo directly contribute to transcriptional activation of the Arm/dTCF transcription factor. Although the Pygo N terminus has been defined as a transactivator, it is possible that other regions of Lgs and Pygo may also possess this activity. It is further possible that Lgs and Pygo may contribute indirectly to Arm's coactivator activity: for example, by facilitating nuclear import or retention of Arm (F.M. Townsley, A. Cliffe, and M. Bienz, unpublished data cited in Thompson, 2004).

In any case, it is unlikely that Lgs and Pygo are the sole mediators of Arm's coactivator activity. For example, while the Arm S6 mutation (in repeat 1) might be predicted to affect Lgs binding, the Arm S12 mutation affects the C-terminal repeats (repeats 10 and 11). It is inferred that an additional, essential coactivating factor(s) is recruited to the C-terminal Arm repeats. Two obvious candidate factors are the histone acetyltransferase CBP/p300 and the chromatin remodeling enzyme Brahma, both of which have been found to bind to C-terminal regions of the Armadillo repeat domain (Thompson, 2004 and references therein). The evidence presented in this study argues in favor of an extended Arm/dTCF transcription factor model in which Arm coactivates dTCF by recruiting Lgs, Pygo, and other factors to its Armadillo repeat domain. (Thompson, 2004).

Pygopus residues required for its binding to legless are critical for transcription and development

Pygopus and Legless/Bcl-9 are recently discovered core components of the Wnt signaling pathway that are required for the transcriptional activity of Armadillo/beta-catenin and T cell factors. It has been proposed that they are part of a tri-partite adaptor chain (Armadillo>Legless>Pygopus) that recruits transcriptional co-activator complexes to DNA-bound T cell factor. Four conserved residues have been identified at the putative PHD domain surface of Drosophila and mouse Pygopus that are required for Pygopus binding to Legless in vitro and in vivo. The same residues are also critical for the transactivation potential of DNA-tethered Pygopus in transfected mammalian cells and for rescue activity of pygopus mutant embryos. These residues at the Legless>Pygopus interface thus define a specific molecular target for blocking Wnt signaling during development and cancer (Townsley, 2004a).

How Pygo and Lgs function molecularly in the transcription of Wnt target genes is still unclear. It has been suggested that Pygo is recruited to DNA-bound TCF through Lgs/Bcl-9 and Armadillo/ß-catenin and that the N terminus of Pygo recruits a co-activator complex or components of the general transcription machinery. The Pygo PHD domain has been implicated in the former, and this work demonstrates the importance of specific Loop2 residues in this recruitment step (Townsley, 2004a).

As yet, there is no direct evidence for the functional importance of the N terminus of Pygo. This portion of the protein is not required for transactivation by DNA-tethered Pygo. No function of the conserved NPFXD motif within the N-box has been detected. However, it is possible that available mutations have not inactivated the N terminus sufficiently to reveal its function. Also, the N terminus may act redundantly with other Pygo sequences (Townsley, 2004a).

Perhaps the most puzzling result of this work is the striking difference in the transactivation potential of untethered Pygo versus DNA-tethered Pygo: the latter is highly active in stimulating transcription, whereas the former is essentially inactive in transfected 293T cells. Similar results were obtained by other investigators who also found that overexpressed untethered Pygo proteins do not stimulate TOPFLASH transcription either in transfected 293 or 293T cells, even in the presence of co-expressed Bcl-9. Therefore, the suggestion that Pygo may act as a transcriptional co-activator may need revisiting. But how can the difference in activity between untethered and DNA-tethered Pygo be explained? Because the transcription assays measured the effects of overexpressed protein in conjunction with endogenous factors, a stimulatory effect would only be observed if none of these were limiting. Given that the direct linkage of Pygo to a DNA-binding domain confers on it high transactivation potential, this implies that the rate-limiting step in these assays is the recruitment of Pygo to DNA. Recall that this apparently involves the three proteins TCF>ß-catenin>Lgs, but none of these are likely to limit the function of exogenous (untethered) Pygo; its transactivation potential cannot be revealed even if these factors are co-overexpressed. However, one possible explanation is that the direct tethering of Pygo to DNA by means of a DNA-binding domain is likely to be far more efficient than indirect recruitment by means of a three-protein adaptor chain. Bypassing this chain may allow for the detection of the intrinsic transactivation potential of Pygo (Townsley, 2004a).

Transcription assays have identified conserved Loop1 residues that contribute to the transactivation potential of DNA-tethered Pygo. These Loop1 residues are predicted to be at a PHD domain surface opposite its putative Lgs-interacting surface. They do not seem to be involved in intramolecular interactions nor in homo-dimerization, nor is this surface of the PHD domain predicted to be involved in the putative binding to phosphoinositides, so they are likely to mediate binding to an unknown ligand. Rescue assays in embryos have not mirrored their functional importance in mammalian transcription assays for a number of possible reasons. The ultimate proof of their functional significance will have to await the identification of their cognate Loop1-binding ligand (Townsley, 2004a).

Pygopus and Legless target Armadillo/beta-catenin to the nucleus to enable its transcriptional co-activator function

Wnt signalling controls the transcription of genes that function during normal and malignant development. Stimulation by canonical Wnt ligands activates beta-catenin (or Drosophila melanogaster Armadillo) by blocking its phosphorylation, resulting in its stabilization and translocation to the nucleus. Armadillo/beta-catenin binds to TCF/LEF transcription factors and recruits chromatin-modifying and -remodelling complexes to transcribe Wnt target genes. The transcriptional activity of Armadillo/beta-catenin depends on two conserved nuclear proteins recently discovered in Drosophila, Pygopus (Pygo) and Legless/BCL-9 (Lgs). Lgs functions as an adaptor between Pygo and Armadillo/beta-catenin, but how Armadillo/beta-catenin is controlled by Pygo and Lgs is not known. This study shows that the nuclear localization of Lgs entirely depends on Pygo, which itself is constitutively localized to the nucleus; thus, Pygo functions as a nuclear anchor. Pygo is also required for high nuclear Armadillo levels during Wingless signalling, and together with Lgs increases the transcriptional activity of beta-catenin in APC mutant cancer cells. Notably, linking Armadillo to a nuclear localization sequence rescues pygo and lgs mutant fly embryos. This indicates that Pygo and Lgs function in targeting Armadillo/beta-catenin to the nucleus, thus ensuring its availability to TCF during Wnt signalling (Townsley, 2004b).

Dissecting nuclear Wingless signalling: Recruitment of the transcriptional co-activator Pygopus by a chain of adaptor proteins including Arm, Pan, and Legless

Members of the Wingless (Wg)/Wnt family of secreted glycoproteins control cell fate during embryonic development and adult homeostasis. Wnt signals regulate the expression of target genes by activating a conserved signal transduction pathway. Upon receptor activation, the signal is transmitted intracellularly by stabilization of Armadillo (Arm)/beta-catenin. Arm/beta-catenin translocates to the nucleus, interacts with DNA-binding factors of the Pangolin (Pan)/TCF/LEF class and activates transcription of target genes in cooperation with the recently identified proteins Legless/BCL9 (Lgs) and Pygopus (Pygo). This study analyses the mode of action of Pan, Arm, Lgs, and Pygo in Drosophila cultured cells. Evidence is provided that together these four proteins form a chain of adaptors linking the NH2-terminal homology domain (NHD) of Pygo to the DNA-binding domain of Pan. NHD has potent transcriptional activation capacity, which differs from that of acidic activator domains and depends on a conserved NPF tripeptide. A single point mutation within this NPF motif abolishes the transcriptional activity of the Pygo NHD in vitro and strongly reduces Wg signalling in vivo. Together, these results suggest that the transcriptional output of Wg pathway activity largely relies on a chain of adaptors designed to direct the Pygo NHD to Wg target promoters in an Arm-dependent manner (Stadeli, 2005).

The identification of β-catenin as the mediator of Wnt signals has spurred investigations and discussions of how a cell adhesion molecule (relocated to the nucleus) can function as a transcriptional activator. One step towards a molecular understanding of nuclear β-catenin function was the discovery of Lgs. Genetic and biochemical experiments indicated that both proteins assist β-catenin in its transactivator role and led to the hypothesis that Lgs functions to recruit Pygo to the β-catenin/TCF complex and hence to the regulatory regions of Wnt target genes. By a series of transcriptional assays in cultured Drosophila cells, this study shows that the four dedicated nuclear components of the Wg signalling pathway, Pan/TCF, Arm/β-catenin, Lgs, and Pygo, act together in a linear manner in order to activate target gene expression. The results have led to the formulation of a 'chain of adaptors' model for nuclear Wg signalling, assigning each of these four proteins a linker function connecting a proximal and a distal component, culminating in the recruitment of the NH2-terminal homology domain of Pygo (PygoNHD) to the promoter of Wg targets. In S2 cells, PygoNHD stimulates reporter gene expression when directly tethered to DNA, indicating that this domain is capable of bestowing transcriptional activity on Arm/β-catenin.

Arm/β-catenin is composed of three domains, each of which bears transcriptional activity when tethered to DNA: the NH2-terminal region. The activity of the NH2-terminus of Arm might not be relevant in the context of the full-length protein, since only an isolated NH2-terminal fragment of Arm[wt] shows transcriptional activity when tethered to DNA. In contrast, an NH2-terminal fragment of ArmS10 or ArmN-R4[D164A] does not show transcriptional activity in this assay. It is thus proposed that Arm's co-activator capacity maps primarily to the COOH-terminal domain and to the Lgs/Pygo recruiting region located in Arm repeats 1-4. While the COOH-terminus of Arm might activate gene expression by recruiting cofactors like the histone acetyltransferase CBP/p300 or the chromatin remodelling enzyme Brg-1/Brahma, the findings indicate that the activity of ArmR1-4 strictly relies on Lgs and Pygo (Stadeli, 2005).

The role of Lgs serving as an adaptor protein to link Pygo with Arm is widely accepted. In contrast, the role of Pygo is somewhat controversial. The finding that the replacement of PygoPHD by LgsHD2 or the fusion of LgsHD2 with PygoNHD results in chimeric proteins that can substitute for both Lgs and Pygo function implies that the LgsHD1 as well as the PygoPHD are dispensable if the PygoNHD is brought to Arm by more direct means. Interestingly, it has been shown in human 293T cells that the activity of DNA-tethered Pygo depends on its PHD. When similar assays were carried out using DNA-tethered constructs in Drosophila S2 cells, it was found that PygoΔPHD is about half as active as full-length Pygo. However, deletion of the NHD leads to a more pronounced decrease in activity, while the isolated NHD is almost as active as full-length Pygo, indicating that the NHD is the key activating domain of Pygo. Consistent with this view, the above-mentioned chimeric proteins lose their rescuing activity if the PygoNHD is mutated. Collectively, these results strongly suggest that Pygo, by means of its NHD, acts as a transcriptional co-activator in Wg signalling (Stadeli, 2005).

Analysis of DNA-tethered PygoNHD in S2 cells reveals that this domain does not act as a classical acidic activator domain. Rather, the conserved tripeptide NPF plays a crucial role in the activity of PygoNHD. Pygo is one out of 791 NPF motif-containing proteins in Drosophila. In yeast, a NPF motif has been found to serve as a recognition motif for proteins bearing Eps15 homology (EH) domains. Since EH domain-containing proteins might work as integrators of signals controlling cellular pathways as diverse as endocytosis, cell proliferation, or nucleocytosolic export, a possible functional link between such proteins and Pygo was examined. However, RNAi-mediated knock-down of all four Drosophila EH domain-containing proteins (CG1099, CG6148, CG6192, CG16932) in the UAS-luc/G4DBD-PygoNHD assay did not cause a decrease in reporter activity, suggesting that the PygoNHD, despite its NPF motif, does not rely on such EH domain-containing proteins for its capacity to activate transcription (Stadeli, 2005).

As an extension of this 'chain of adaptors' model, it is tempting to assume that Pygo, through its NHD, might recruit factors with enzymatic activities, such as histone acetyl transferases (HAT) or chromatin remodelling proteins. Such interactions may depend on an intact Pygo NPF motif. It is important to mention that mutations in this motif, although almost abolishing activity of the isolated DNA-tethered PygoNHD, retain some activity in the context of the full-length protein. It is possible that these mutations do not completely prevent presumptive interactions with positive regulators of transcription; some other part of Pygo may act in a partially redundant manner together with the NHD. Confirmation of such explanations will have to await the identification of PygoNHD interacting proteins (Stadeli, 2005).

Pygopus and legless provide essential transcriptional coactivator functions to armadillo/β-catenin

Wnt signaling controls important aspects of animal development, and its deregulation has been causally linked to cancer. Transduction of Wnt signals entails the association of β-catenin with nuclear TCF DNA binding proteins and the subsequent activation of target genes. The transcriptional activity of Armadillo (Arm, the Drosophila β-catenin homolog) largely depends on two recently discovered components, Legless (Lgs) and Pygopus (Pygo). Lgs functions as an adaptor between Arm/β-catenin and Pygo, but different mechanisms have been proposed as to how Arm/β-catenin is controlled by Lgs and Pygo. Although Lgs and Pygo were originally thought to serve as nuclear cofactors for Arm/β-catenin to enhance its transactivation capacity, a recent analysis argued that they function instead to target Arm/β-catenin to the nucleus. This study used genetic assays in cultured cells and in vivo to discriminate between the two paradigms. Regardless of the measures taken to maintain the nuclear presence of Arm/β-catenin, a transcriptional-activation function of Pygo could not be bypassed. These findings therefore indicate that Arm/β-catenin depends on Lgs and Pygo primarily for its transcriptional output rather than for its nuclear import (Hoffmans, 2005).

Wingless signals are secreted glycoproteins controlling many fundamental processes during animal development. Whereas several responses to Wnt ligands appear to entail direct cytoplasmic responses organizing planar cell polarity and organ morphogenesis, a significant fraction of Wnt responses concern transcriptional changes in the nucleus. This latter aspect of Wnt-signal transduction is mediated by β-catenin and is often referred to as “canonical” or β-catenin-dependent Wnt signaling. The canonical Wnt pathway plays important roles in embryonic-cell-fate determination, and its constitutive activation is oncogenic in several adult mammalian tissues, most notably in the intestinal epithelium. Hence, it is of prime interest to understand how β-catenin activity can upregulate transcription of Wnt target genes. Although cytoplasmic β-catenin was originally discovered through its role in cell adhesion, a large body of evidence indicates that it is degraded in the absence of a Wnt signal but stabilized in its presence. As a consequence, β-catenin can sufficiently accumulate, translocate to the nucleus, and be directed to Wnt target genes by associating with DNA-binding TCF/LEF proteins. However, it is less clear how a cell-adhesion component, relocated to the nucleus, can promote and sustain the transcriptional activity of these targets (Hoffmans, 2005).

Using genetic assays in Drosophila, a presumptive adaptor protein, Legless (Lgs) has been identified, that binds to β-catenin and its Drosophila homolog, Armadillo (Arm), as well as to the nuclear protein Pygopus (Pygo). On the basis of biochemical and phenotypic analysis, it is proposed that nuclear β-catenin/Arm assembles a quaternary complex, consisting of TCF, β-catenin, Lgs, and Pygo, in which Pygo serves as a transcriptional activator to induce and/or maintain the transcription of Wnt/Wg target genes. Alternatively, however, the requirement for Lgs and Pygo in Wnt/Wg signaling could be attributed to a role in targeting and retaining β-catenin in the nucleus, increasing its net nuclear concentration and, hence, its activity. This latter view has recently gained recognition and experimental support by a cell-biological analysis of these components. This study set out to address the mechanistic role of Pygo by subjecting the two models to three different tests; each case comes to the conclusion, that Pygo functions mainly in the transcriptional output of β-catenin (Hoffmans, 2005).

In the first approach, the consequences of disrupting the molecular interaction between β-catenin and Lgs was examined. β-catenin/Arm amino acid residues required for Lgs binding have been identified and it was observed that mutant β-catenin forms lacking these residues are severely compromised in their signaling activity. This reduction in activity could be caused either by a failure of β-catenin/Arm to recruit the “transcriptional mediator” Pygo or by a reduced (as a result of diminished nuclear anchoring) nuclear-cytoplasmic ratio of β-catenin/Arm. An experiment was repeated in which N-terminally truncated and therefore constitutively active forms of Arm, ArmS10-wt and ArmS10-D164A (differing solely in one critical amino acid residue necessary for Lgs binding), were expressed in the embryonic epidermis of Drosophila. Expression of ArmS10-wt suppresses denticle formation—a read-out for a gain of Wg signaling activity—whereas the D164A mutation, which impairs binding to Lgs, efficiently abolished this gain-of-function activity. When subjected to an immunohistochemical analysis, however, the two genotypes visually differed neither in amount nor subcellular localization of the ArmS10 proteins. The D164A mutation was further used in a cellular assay in which a constitutively active form of β-catenin (S33Y) was tethered to the enhancer of a reporter gene by the DNA binding domain of Gal4. Whereas β-catenin caused strong transcriptional activation, the D164A form lost this activity almost completely. Importantly, however, both forms were expressed at equivalent levels in human cells and did not differ in their ability to localize in nuclei. Because Lgs mediates the binding of β-catenin to Pygo, these results are interpreted as evidence that a failure of Arm/β-catenin to recruit Pygo impedes the transcriptional activity of the former despite the fact that it is nuclearly localized (Hoffmans, 2005).

A second test was devised on the assumption that Lgs appears to function merely as an adaptor between Arm/β-catenin and Pygo, thereby linking Arm/β-catenin either to a transcriptional activator or a nuclear anchor. Such a passive role for Lgs can be inferred from the observations that Lgs is dependent on Pygo both for its signaling activity and for its nuclear localization. If the main role of Lgs would be to link Arm/β-catenin to the constitutively nuclear anchor Pygo, it should gain functional independence of Pygo when bestowed with a nuclear-localization signal (NLS). Lgs was therefore modified by replacing a C-terminal portion with sequences of a green fluorescent protein (LgsN-eGFP) and adding the NLS of SV40 large T-antigen N-terminally (NLS-LgsN-eGFP). These altered forms of Lgs were examined for their subcellular distribution and signaling function. The addition of a single NLS effectively conferred nuclear localization, as assessed in transfected cells. When tested for their signaling capacity in Drosophila S2 cells, LgsN and NLS-LgsN were found to be equally active in rescuing the RNAi-mediated knockdown of endogenous Lgs. However, these two forms of Lgs were equally inactive in rescuing the knockdown of endogenous Pygo. Consistent with this result, it was also found that the Lgs-rescuing activity of NLS-LgsN still depends on the HD1 domain, through which it binds Pygo. Together, these results indicate that constitutive nuclear targeting of Lgs does not bypass the requirement for Pygo in Wg signaling, suggesting that Pygo must provide a function beyond ensuring availability of Lgs and β-catenin in the nucleus of Wg-transducing cells (Hoffmans, 2005).

The third test aimed at assessing the role of the N-terminal homology domain (NHD) of Pygo. Drosophila Pygo and its two mammalian homologs, Pygo1 and Pygo2, share—in addition to their C-terminal plant homology domain (PHD) finger domain, through which they bind Lgs—a short N-terminally located sequence of amino acids. On the basis of the conservation of Pygo function and absence of further common domains, the NHD was proposed to serve as transactivation domain. It was first confirmed that the NHD core domain (amino acids 91 to 101) is not required for nuclear localization of Pygo because neither the deletion of the core nor the change of a conserved and functionally required amino acid (F99A) affected the nuclear localization of Pygo in cultured cells. Importantly, these alterations also had no discernible effect on the capacity of Pygo to bind Lgs. If Pygo and Lgs primarily function to target Arm/β-catenin to the nucleus, then NHD mutations should not seriously affect Wnt/Wg signaling. It was found, however, that Pygo-F99A—in contrast to wild-type Pygo—failed to rescue Pygo function in cultured cells and in vivo. The endogenous pygo gene was replaced with a genomic pygo-F99A transgene in vivo, and it was observed that both mutant and wild-type Pygo proteins are expressed at comparable levels without detectable differences in nuclear-cytoplasmic distribution. The most explicit argument for a role of the NHD in transactivation was obtained by analyzing mutant clones of imaginal cells in which either the pygo-wt or the pygo-F99A transgenes were the only source of full-length Pygo protein. Both transgenes rescued Lgs nuclear localization in the mutant clones to a similar extent; however, pygo-F99A—but not pygo-wt—showed severely reduced transcription of the Wg target gene senseless. Because Pygo protein bearing a mutant NHD retains the capacity to localize Lgs (and, by inference, Arm), it is inferred that the key function of the Pygo NHD is to confer transcriptional activity to Arm (Hoffmans, 2005).

In summary, the functions of Lgs and Pygo were tested in β-catenin-dependent Wnt/Wg signaling by devising experiments that separate a role in transcriptional activation of targets from a role in nuclear targeting or retention of Arm/β-catenin. In all three situations examined, the transcriptional output of Arm/β-catenin depended on Pygo activity despite measures to grant Arm/β-catenin such alleged nuclear retention. When Arm/β-catenin was tethered directly to DNA via the Gal4 DNA binding domain, or when Lgs was endowed with an NLS of its own, Arm/β-catenin activity was still dependent on the recruitment of Pygo. Likewise, in vivo, when the nuclear retention activity of Pygo was left intact, Arm was not able to transduce Wg and activate target genes without the Pygo NHD. Although it cannot be rule out that Lgs and Pygo function as a nuclear anchor for β-catenin, the results collectively argue that the primary requirement for the two Arm/β-catenin partners must be attributed to a transcriptional role that allows Arm/β-catenin to activate and/or sustain the expression of Wnt/Wg target genes. Although information is lacking on the biochemical nature of this transactivation activity, it is tempting to assume that it involves the NHD-mediated recruitment of a chromatin-modification complex or of factors mediating transcription initiation or elongation (Hoffmans, 2005).

BCL9-2 binds Arm/β-catenin in a Tyr142-independent manner and requires Pygopus for its function in Wg/Wnt signaling

The Wingless (Wg)/Wnt signal transduction pathway controls fundamental processes during animal development. Deregulation of the Wg/Wnt pathway has been causally linked to several forms of cancer, most notably to colorectal cancer. In response to Wg/Wnt signaling, Armadillo/β-catenin associates in the nucleus with DNA bound TCF and several co-factors, among them Legless/BCL9, which provides a link to Pygopus. Recently, the second vertebrate homologue of Legless, BCL9-2 (or B9L), was characterized and proposed to mediate Wnt signaling in a Pygopus-independent manner, by binding to a Tyrosine-142-phosphorylated form of β-catenin. This study examined the role of Tyrosine-142 phosphorylation in several assays; it is neither important for the recruitment of BCL9-2, nor for the transcriptional activity of β-catenin in cultured mammalian cells, nor in Drosophila for Wg signaling activity in vivo. Furthermore, it is demonstrated that BCL9-2 can functionally replace Lgs both in cultured cells as well as in vivo and that this rescue activity depends on the ability of BCL9-2 to bind Pygo. These results do not show a significant functional difference between BCL9-2 and BCL9 but rather suggest that the two proteins represent evolutionary duplicates of Legless, which have acquired distinct expression patterns while acting in a largely redundant manner (Hoffmans, 2007).

Previous studies have ascribed unique properties and functions to BCL9 and BCL9-2, namely that BCL9-2 functions independently of Pygo and depends on phosphorylation of tyrosine 142 of β-catenin to be able to interact with it. Initially attempts were made determine if residue D164 of β-catenin, crucial for BCL9 binding, was also important for BCL9-2 binding. This study was extended to compare the signaling properties of the two proteins in other available assays. Although the repertoire of these assays has some limitations (e.g. a deficit in vertebrate genetics), these experiments did not reveal any significant functional differences between the two human homologs. Instead, it appears that BCL9 and BCL9-2 represent evolutionary duplicates of Lgs that can function in a largely redundant manner (Hoffmans, 2007).

Evidence was first provided that binding of BCL9, BCL9-2 and Lgs to Arm/β-catenin does not depend on tyrosine residue 142 of Arm/β-catenin. Single amino acid substitutions of Y142 had no effect on the transcriptional activity of Arm/β-catenin, neither in vivo nor in tissue culture cells. It was next shown that BCL9-2 can functionally replace Lgs in vivo and in cultured cells, and that this activity of BCL9-2 depends on its ability to bind Pygo. Finally, it was found that BCL9 and BCL9-2 can interact with the same partners and form similar complexes in the nucleus of stimulated cells (Hoffmans, 2007).

It was previously reported that BCL9-2 interacts efficiently with β-catenin only if the latter is phosphorylated on tyrosine 142. In the absence of tyrosine kinase activity no interaction of the two proteins was detected in a special yeast two-hybrid system. Furthermore, a mutation of β-catenin's tyrosine 142 to alanine abrogated binding to BCL9-2. In contrast, the current binding studies indicate that the binding of Arm/β-catenin to BCL9-2 does not dependent on Y142 phosphorylation. In agreement with these results, BCL9-2 was also identified as a β-catenin binding partner in a standard yeast two-hybrid system that lacks protein tyrosine kinase activity. One explanation to account for the discrepancy between the two results may be attributed to the use of different lengths of β-catenin proteins in the binding assays. While full-length β-catenin was used in the present study, a previous study used a shortened protein containing only the Armadillo repeats, and another study used a form that in addition to the Arm repeats contained small parts of the N- and C-termini (Hoffmans, 2007).

The arm2a9 allele is the strongest arm allele available, has a frameshift mutation in Arm repeat 3 and fails to provide both adhesion function as well as Wg transduction activity. A transgene encoding Arm-Y142A was able to rescue arm2a9 males, but less efficiently as the control transgene encoding wild-type Arm. This difference was attributed to a slight deficit in adhesion function, a hypothesis supported by the observations that Arm-Y142A exhibited reduced α-catenin binding while it fully rescued a signaling-defective but adhesion-competent allele of arm (armS7X-2). Furthermore, no difference could be observed in Wg signaling activity when Arm-Y142A and Arm-wt were interrogated in a quantitative signaling assay in S2 cells. Together, these experiments indicate that the transduction of the Wg signal through Arm does not depend on residue Y142 (Hoffmans, 2007).

To compare the BCL9-2 and BCL9 proteins their ability to substitute for Lgs in Wg signaling was tested. Both human homologs are able to provide Lgs signaling function. Interestingly, previous studies reported that the signaling activity of BCL9-2 does not depend on Pygo binding, but on the C-terminus of BCL9-2. However, in the current replacement assays in Drosophila, found that both human proteins depended equally on their Pygo-binding domain (HD1). At the amino acid level BCL9 and BCL9-2 share three homologous regions of 20-30 amino acids in their C-termini. These short clusters are not found in Lgs; furthermore, it was previously found that the C-terminal half of Lgs is dispensable for Lgs function in Drosophila. Future studies will have to clarify a potential contribution of these C-termini to Wnt signaling in vertebrates. The conservation of these clusters in both BCL9 and BCL9-2 suggests that whatever the function of the C-terminus it is conserved between the two proteins. One approach that might shed light on this issue is to generate mice lacking BCL9 and BCL9-2 and see if the phenotypes of these animals can be rescued by a transgene ubiquitously expressing either only BCL9, BCL9-2 or Lgs. Less stringent signaling assays based on RNAi in cultured mammalian cells were uninformative, as the simultaneous addition of two siRNAs (against both bcl9 and bcl9-2) decreased their knock-down efficiency, complicating the interpretation of the results (Hoffmans, 2007).

Phylogenetic sequence comparisons indicate that a relatively recent event of gene duplication created the two BCL9 forms in vertebrates. The current results indicate that these proteins are essentially identical, at least regarding their interaction with β-catenin as well as regarding their capacity to transduce the Wg signal and to form higher order complexes containing β-catenin and Parafibromin. It is thus suggested that the BCL9 proteins function in a largely redundant manner, as adaptor proteins for the recruitment of Pygo to the β-catenin-TCF complex (Hoffmans, 2007).

The PHD domain is required to link Drosophila Pygopus to Legless/β-catenin and not to histone H3

In Drosophila Pygopus (Pygo) and Legless (Lgs)/BCL9 are integral components of the nuclear Wnt/Wg signaling machine. Despite intense research, ideas that account for their mode of action remain speculative. One proposition, based on a recently discovered function of PHD fingers, is that Pygo, through its PHD, may decipher the histone code. This study found that human, but not Drosophila, Pygo robustly interacts with a histone-H3 peptide methylated at lysine-4. The different binding behavior is due to a single amino acid change that appears unique to Drosophilidae Pygo proteins. Rescue experiments with predicted histone binding mutants showed that in Drosophila the ability to bind histones is not essential. Further experiments with Pygo-Lgs fusions instead demonstrated that the crucial role of the PHD is to provide an interaction motif to bind Lgs. The results reveal an interesting evolutionary dichotomy in Pygo structure-function, as well as evidence underpinning the chain of adaptors model (Kessler, 2009).

This study has carefully tested the idea that histone binding contributes to the function of Pygo. The results demonstrate that in Drosophila the ability to bind H3K4me is not required for Pygo function. Instead the results suggest that the sole crucial function of the dPygoPHD finger is as a link in the chain of adaptors, which recruits the transcriptional activator potential of the NHD to Wg target genes (Kessler, 2009).

The findings do not exclude that in other species histone binding by the PygoPHD has a functional importance. Indeed it is tempting to speculate that the different behavior of mammalian and Drosophila Pygo with respect to histone binding accounts for their different requirement in vivo. In mice, where histone binding appears convincing, Pygo loss of function has a 'mild' phenotype. In Drosophila, where histone binding is apparently absent, Pygo has a positive role that is reflected in a strong loss of function phenotype. Since the effect of Pygo loss of function in other organisms has not been documented, it remains unclear if mice represent the rule and Drosophila the exception, or vice versa. If the conservation of the W in the K4 cage is an indication then Drosophila is the exception. One paradigm that would account for the observations is that Pygo has both positive and negative modes of action; histone binding providing the negative role. A possible mechanism for this model was suggested by Mosimann (2009): Pygo binds H3K4me via its PHD finger. By using BCL9 as adaptor it can retain β-catenin in the proximity of a Wingless response independently of TCF. Pygo-dependent tethering of β-catenin to methylated histones frees TCF to recruit co-repressors (e.g. Groucho and CtBP), which then counteract the β-catenin-nucleated activating chromatin remodeling processes. However, this mechanism relies on the existence of a Histone/Pygo/BCL9 complex, which the results suggest may not form (Kessler, 2009).

Whatever the case, the demonstration that dPygo’s role in Wg signaling can be reduced to the N-terminal Homology Domain (NHD) of Pygopus suggests that its primary positive function lies in this domain. The list of transcriptional complexes which interact with the NHD is small but growing, a trend which should continue. Via the interaction of the PHD with HD1 of Lgs this polygamy of interactions is targeted to Wnt target genes. In this way Pygo may act co-operatively with the C-terminus of Arm/β-catenin -- recruiting complexes required for Wg/Wnt target gene activation (Kessler, 2009).


Effects of Mutation or Deletion

Animals transheterozygous for a strong and a weak allele of lgs die as pharate adults with a striking phenotype: they lack legs and antennae, appendages whose growth and pattern are critically dependent on Wg activity. These animals also show occasional wing -to-notum transformations, the founding phenotype of the wg gene that reflects its role in defining the wing blade primordium. Strong lgs alleles cause larval lethality: homozygotes rarely reach the third instar stage, and the few that do so exhibit miniature imaginal discs that fail to express the Wg target gene Distalless (Dll) but show normal expression levels of the Hedgehog (Hh) target gene patched (ptc. To explore whether lgs may also be required for Wg signaling at embryonic stages, attempts were made to remove a potential maternal contribution of lgs. The location of lgs on chromosome 4 necessitated the use of pole cell transplantation to generate mosaic females with a wild-type soma and a lgs mutant germline. Homozygous mutant embryos derived from such a female display a segment polarity phenotype in which the larval epidermis forms a lawn of ventral denticles and lacks naked cuticle between the segmental denticle belts. The similarities between the phenotypes of lgs and wg further support the conclusion that lgs encodes a component of the Wg signaling pathway (Kramps, 2002).

To determine whether lgs is required for the generation of active Wg signal or for transduction of this signal, Wg target gene expression was examined in lgs genetic mosaics. Clones of cells lacking lgs function were generated with a tubulinalpha1>lgs>Gal4 transgene in a lgs20F homozygous background. Excision of the >lgs> Flp-out cassette generates lgs mutant cells that are marked by the gain of Gal4 expression, which can be monitored by the activity of a UAS-GFP reporter gene. Typically, clones of lgs mutant cells undergo only a few divisions after they are generated in the presumptive wing blade; then, the mutant cells stop proliferating and either die or are actively eliminated from the disc epithelium. When stained for the presence of Dll expression 48 hr after mitotic recombination is induced, lgs mutant cells exhibit reduced levels of Dll protein. This cell-autonomous behavior indicates that Lgs acts in Wg-receiving cells, permitting their survival and the expression of target genes (Kramps, 2002).

To map the position of Lgs action along the Wg transduction pathway, Shaggy/Zw3 (Sgg) was used as a reference for an epistasis analysis. The sgg gene encodes the Drosophila homolog of GSK3, which plays a critical role in the ß-catenin/Arm destruction complex. Loss of sgg function causes the constitutive activation of the Wg pathway, which results in an embryonic phenotype opposite that of wg or lgs mutations. Since sgg, as well as lgs, provides some of its function maternally, attempts were made to remove the maternal and zygotic products of both genes simultaneously. A tubulinalpha1>lgs;sgg+>Gal4 transgene was introduced into a lgs;sgg null mutant background to generate double-mutant female germ cells. The sgg;lgs double-mutant embryos derived from these cells were identified by the use of three marker genes. Such embryos displayed a 'lawn of denticles' phenotype, indistinguishable from that of lgs single-mutant embryos, but in direct contrast to the 'naked' phenotype of sgg single-mutant animals. This result is interpreted to indicate that lgs acts downstream of sgg in the Wg pathway (Kramps, 2002).

To further narrow down the point of lgs action, it was asked whether the constitutive signaling activity of N-terminally truncated Arm is dependent on lgs function. The tubulinalpha1>lgs>Gal4 system described above was used to generate embryos devoid of wild-type lgs product, yet instead ubiquitously expressing a UAS-armS10 transgene, which encodes a mutant Arm protein with a 54 amino acid deletion in the N-terminal domain. Expression of armS10 results in a 'naked cuticle' phenotype in the presence of lgs but not in its absence. Therefore, lgs function is required even for the activity of a stabilized form of Arm (Kramps, 2002).

Finally, the lgs gene product was localized in cells with an antiserum directed against the N-terminal half of Lgs and it was found that wild-type cells, but not lgs20F mutant cells, express Lgs protein in the nuclei. Both subcellular localization and Lgs protein expression levels are invariant throughout the disc, suggesting that neither of these properties is regulated by the activity of the Wg signaling pathway. Nuclear localization of Lgs appears to be essential for its signaling activity, since a membrane-localized form of Lgs (CD2-Lgs), which does not have access to the nucleus, cannot rescue lgs mutant animals (Kramps, 2002).

Together, these results indicate that Lgs acts at the very bottom of the Wg signal transduction cascade, in conjunction with, or downstream of, nuclear Arm (Kramps, 2002).

The role of Pygopus in the differentiation of intracardiac valves in Drosophila

Cardiac valves serve an important function; they support unidirectional blood flow and prevent blood regurgitation. Wnt signaling plays an important role in the formation of mouse cardiac valves and cardiac valve proliferation in Zebrafish, but identification of the specific signaling components involved has not been addressed systematically. Of the components involved in Wnt signal transduction, pygopus (pygo), first identified as a core component of Wnt signaling in Drosophila, has not been investigated with respect to valve development and differentiation. This study took advantage of the Drosophila heart model to study the role of pygo in formation of valves between the cardiac chambers. Cardiac-specific pygo knockdown in the Drosophila heart was found to cause dilation in the region of these cardiac valves, and their characteristic dense mesh of myofibrils does not form and resembles that of neighboring cardiomyocytes. In contrast, heart-specific knockdown of the transcription factors, arm/beta-Cat, lgs/BCL9, or pan/TCF, which mediate canonical Wnt signal transduction, shows a much weaker valve differentiation defect. Double-heterozygous combinations of mutants for pygo and the Wnt-signaling components have no additional effect on heart function compared with pygo heterozygotes alone. These results are consistent with the idea that pygo functions independently of canonical Wnt signaling in the differentiation of the adult interchamber cardiac valves (Tang, 2014).


Adachi, S., et al. (2004). Role of a BCL9-related beta-catenin-binding protein, B9L, in tumorigenesis induced by aberrant activation of Wnt signaling. Cancer Res. 64(23): 8496-501. 15574752

Bauer, A., Huber, O. and Kemler, R. (1998). Pontin52, an interaction partner of ß-catenin, binds to the TATA box binding protein. Proc. Natl. Acad. Sci. 95: 14787-14792. 9843967

Brembeck, F. H., et al. (2004). Essential role of BCL9-2 in the switch between ß-catenin's adhesive and transcriptional functions. Genes Dev. 18(18): 2225-30. 15371335

Chodaparambil, J. V., Pate, K. T., Hepler, M. R., Tsai, B. P., Muthurajan, U. M., Luger, K., Waterman, M. L. and Weis, W. I. (2014). Molecular functions of the TLE tetramerization domain in Wnt target gene repression. EMBO J 33(7): 719-731. PubMed ID: 24596249

Fiedler, M., Sanchez-Barrena, M. J., Nekrasov, M., Mieszczanek, J., Rybin, V., Muller, J., Evans, P. and Bienz, M. (2008). Decoding of methylated histone H3 tail by the Pygo-BCL9 Wnt signaling complex. Mol Cell 30(4): 507-518. PubMed ID: 18498752

Fiedler, M., Graeb, M., Mieszczanek, J., Rutherford, T. J., Johnson, C. M. and Bienz, M. (2015). An ancient Pygo-dependent Wnt enhanceosome integrated by Chip/LDB-SSDP. Elife 4. PubMed ID: 26312500

Hecht, A., Litterst, C. M., Huber, O. and Kemler, R. (1999). Functional characterization of multiple transactivating elements in ß-catenin, some of which interact with the TATA-binding protein in vitro. J. Biol. Chem., 274: 18017-18025. 10364252

Hoffmans, R. and Basler, K. (2004). Identification and in vivo role of the Armadillo-Legless interaction. Development 131: 4393-4400. 15294866

Hoffmans, R., Stadeli, R. and Basler, K. (2005). Pygopus and legless provide essential transcriptional coactivator functions to armadillo/β-catenin. Curr. Biol. 15(13): 1207-11. 16005293

Hoffmans, R. and Basler, K. (2007). BCL9-2 binds Arm/β-catenin in a Tyr142-independent manner and requires Pygopus for its function in Wg/Wnt signaling. Mech. Dev. 124(1): 59-67. 17113272

Kessler, R., Hausmann, G. and Basler, K. (2009). The PHD domain is required to link Drosophila Pygopus to Legless/β-catenin and not to histone H3. Mech. Dev. 126(8-9): 752-9. PubMed Citation: 19493659

Kadoch, C. and Crabtree, G. R. (2015). Mammalian SWI/SNF chromatin remodeling complexes and cancer: Mechanistic insights gained from human genomics. Sci Adv 1(5): e1500447. PubMed ID: 26601204

Kramps, T., et al. (2002). Wnt/Wingless signaling requires BCL9/Legless-mediated recruitment of Pygopus to the nuclear ß-Catenin-TCF complex Cell 109: 47-60. 11955446

Li, J., Chen, X., Ding, X., Cheng, Y., Zhao, B., Lai, Z. C., Al Hezaimi, K., Hakem, R., Guan, K. L. and Wang, C. Y. (2013). LATS2 suppresses oncogenic Wnt signaling by disrupting beta-catenin/BCL9 interaction. Cell Rep 5: 1650-1663. PubMed ID: 24360964; Graphical Abstract

Mosimann, C., Hausmann, G. and Basler, K. (2009). Beta-catenin hits chromatin: regulation of Wnt target gene activation. Nat. Rev. Mol. Cell Biol. 10: 276-286. PubMed Citation: 19305417

Sierra, J., Yoshida, T., Joazeiro, C. A. and Jones, K. A. (2006). The APC tumor suppressor counteracts beta-catenin activation and H3K4 methylation at Wnt target genes. Genes Dev. 20(5): 586-600. 16510874

Stadeli, R. and Basler, K. (2005). Dissecting nuclear Wingless signalling: Recruitment of the transcriptional co-activator Pygopus by a chain of adaptor proteins. Mech. Dev. 122(11): 1171-82. 16169192

Tang, M., Yuan, W., Bodmer, R., Wu, X. and Ocorr, K. (2014). The role of Pygopus in the differentiation of intracardiac valves in Drosophila. Genesis 52: 19-28. PubMed ID: 24265259

Thompson, B. J., et al. (2004). A complex of Armadillo, Legless, and Pygopus coactivates dTCF to activate Wingless target genes. Curr. Biol. 14: 458-466. 15043810

Townsley, F. M., Thompson, B. and Bienz, M. (2004a). Pygopus residues required for its binding to legless are critical for transcription and development. J. Biol. Chem. 279: 5177-5183. 14612447

Townsley, F. M., Cliffe, A. and Bienz, M. (2004b). Pygopus and Legless target Armadillo/beta-catenin to the nucleus to enable its transcriptional co-activator function. Nat. Cell Biol. 6: 626-633. 15208637

van Tienen, L. M., Mieszczanek, J., Fiedler, M., Rutherford, T. J. and Bienz, M. (2017). Constitutive scaffolding of multiple Wnt enhanceosome components by Legless/BCL9. Elife 6: e20882. PubMed ID: 28296634

Willis, T. G., et al. (1998). Molecular cloning of translocation t(1;14)(q21;q32) defines a novel gene (BCL9) at chromosome 1q21. Blood 91: 1873-1881. 9490669

legless: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation

date revised: 30 April 2017

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

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