Pygopus is a transcriptional activator important for the Wnt signaling pathway. It binds to the beta-catenin transcriptional complex via the adaptor proteins Bcl9 and Bcl9l (Bcl9/9l). This complex is considered to be a suitable target for the treatment of tumors that display activated Wnt signaling. In the mouse, there are two Pygopus-encoding genes, Pygo1 and Pygo2 (Pygo1/2), with the latter playing a major role. Here we introduce a single amino acid substitution in Pygo2, which was previously shown to abrogate binding to Bcl9/9l, and cause lethality in Drosophila melanogaster. We confirm that mutant Pygo2 protein fails in interacting with Bcl9 but, unexpectedly, homozygous mice with this mutation are viable and fertile, even when this mutant allele is combined with a null mutation of the potentially redundant Pygo1. Based on this observation, we conjecture that the Pygo-Bcl9/9l interaction requires scant affinity in vivo to fulfill developmental functions and thrust forward the notion that this interaction surface could be targeted in cancer therapy without major consequences on homeostatic functions.

The Wnt signaling pathway drives virtually all aspects of embryonic development, and its over activation is a causative factor in several human cancers^[1]. Beta-catenin is the molecular fulcrum of canonical Wnt signaling^[2]. It binds to several transcriptional co-factors; among them, Bcl9/9l and Pygo1/2^[3][4][5]. Of note, while Pygo1-null mice are viable and fertile, mice deficient for Pygo2 die between 13.5 days post coitum (dpc) and birth, presenting a series of developmental defects. Double Pygo1/2 homozygous mutants do not display any synergy in the severity of the phenotype^[6][7]. Bcl9/9l act as “bridge” proteins by simultaneously binding beta-catenin (via the HD2 domain) and Pygo1/2 (via the HD1 domain). The relevance of this protein complex in the mouse is also supported by previous work: the deletion of the HD1 domain-which fully abrogates Bcl9/9l's ability to bind Pygo1/2-leads to embryonic lethality, recapitulating the complete loss of Pygo1/2^[8]. Earlier results identified the amino acids within the Plant Homology Domain (PHD) of Pygo proteins that are critical for binding to Bcl9/9l: these mutations (e.g., L789A) lead to embryonic lethality in Drosophila melanogaster^[9]. Here we create the corresponding and functionally analogous mutation in the mouse Pygo2: the substitution of the Leucine in position 368 (L368A).

Our purpose is to further test the relevance of the Pygo2-Bcl9/9l interaction surface and the effect of its abrogation on mouse development. This interaction is a candidate target for the treatment of Wnt signaling-driven tumors.

We generated a Pygo2 knock-in mutant allele by introducing a two-nucleotide change (CT>GC) so that the Leucine in position 368-within the plant homology domain (PHD) of Pygo2-is replaced by an Alanine (Fig. 1A). The homologous mutation in pygo (L789A) leads to lethality in Drosophila melanogaster^[9]. The mice are genotyped with primers spanning the remaining FRT sequence that follows the NEO-cassette excision (Fig. 1B, see Materials and Methods section for a detailed description). The presence of the mutation is confirmed by DNA sequencing (Fig. 1B). Initially, three independent breedings between heterozygous mice (Pygo2L368A/+ X Pygo2L368A/+) gave rise to mutant homozygous Pygo2L368A/L368A animals (6/44 [13.6%], representing half of the expected Mendelian ratio). The homozygous mutant mice obtained were then interbred, giving rise to healthy animals, thereby showing that homozygous Pygo2-L368A mutant mice reach adulthood without any malformation and are fertile. To rule out any compensatory mechanism due to potential redundancy with Pygo1, we brought the Pygo2-L368A mutant allele into a Pygo1-knockout background. By crossing Pygo2 heterozygous, Pygo1-knockout mice (Pygo1-/-; Pygo2L368A/+), we obtained Pygo1/2 double mutant animals (Pygo1-/-; Pygo2L368A/L368A) in a proportion of 4/15 (26% the expected Mendelian ratio is 1/4). A double mutant male mouse (Pygo1-/-; Pygo2L368A/368A) from this progeny was successively bred twice with a Pygo1-/-; Pygo2L368A/+ female and gave rise to a total of 10/23 double mutant mice (43.5% the expected Mendelian ratio is 1/2). Both male and female double mutant mice could breed, allowing us to establish a colony of healthy Pygo1-/-; Pygo2L368A/L368A animals.

The effect of the L368A mutation on the Pygo2-Bcl9 interaction was confirmed by in vitro GST pull-down experiments. We extracted proteins from adult wild type and mutant kidneys, an abundant source for Pygo2 protein^[10]. We incubated wild-type and mutant protein extracts with a recombinant GST-Bcl9 protein fragment (amino acids 1-372 of mouse Bcl9, that includes the region spanning two relevant domains: HD1 [Pygo-binding] and HD2 [beta-catenin-binding]). GST-GFP was used as negative control. With glutathione-conjugated sepharose beads, we pulled down the GST-proteins and performed Western blot analysis to detect Pygo2 in the pull-down reactions. Whereas the GST-Bcl9 can strongly interact with the wild-type Pygo2, it fails in significantly pulling down the Pygo2-L368A mutated protein (Fig. 1C, compare the bands indicated by the two white asterisks).

Pygo1/2 double knockout mice die during embryonic development between 13.5 dpc and birth due to a series of developmental defects^[7][8]. However, Pygo2-L368A embryos are, in their superficial appearance, indistinguishable from wild type littermates (Fig. 1D). Out of 25 embryos analyzed, we scored six mutants (24%, close to the expected Mendelian ratio of 1/4). Moreover, histological analyses of the tissues affected during development in Pygo1/2-knockout mice (mainly the lens, the lungs, and the kidney^[7][6]) reveal no obvious alterations at 15.5 dpc (Fig. 1E).

The complete abrogation of the interaction between Bcl9/9l and Pygo1/2-via the deletion of the HD1 domain in both Bcl9 and Bcl9l-leads to embryonic lethality at 13.5 dpc, with a striking “Pygo knockout” phenotype^[8]. These results appear contradictory (see also the Alternative Explanations paragraph). It is possible that the deletion of the full Pygo-interacting HD1 domain of Bcl9/9l completely abrogates their interaction, while the single amino acid substitution L368A in Pygo2 strongly reduces their binding, but leaves some residual interaction. Thus, the residual binding between Pygo2-L368A and Bcl9 would be capable of fulfilling all the developmental functions for which this interaction is required. It is possible, in fact, that this interaction is weak and dynamic in nature in vivo. The extent to which the mutation L368A decreases the affinity between Pygo2 and Bcl9 remains to be determined. In conclusion, we cannot observe any developmental or homeostatic defect in mutant Pygo1-/-; Pygo2L368A/L368A mice: they reach adulthood healthy and fertile. Because several reports have previously shown that the interaction between Pygo1/2 and Bcl9/9l is necessary for proper development^[9][11][8], we propose that only a weak interaction between these factors is required.

Pygo2-L368A mice are viable and fertile despite displaying strongly reduced Pygo2-Bcl9/9l binding.

The main limitation of this study consists in the fact that, by using an in vitro binding assay to demonstrate the decreased affinity between the mutant Pygo2 and Bcl9, we cannot exclude that this interaction is intact in the cell.

When considered under the scrutiny of genetics formalism, the mutation in Pygo2 we describe here (L368A) and the deletion of the HD1 domain in Bcl9/9l should lead to the same phenotypic consequences- they both abrogate the Pygo2–Bcl9/9l interaction. The deletion of HD1 leads to embryonic lethality associated with recognizable phenotypic features, without compromising the overall stability and expression of Bcl9/9l^[8]. The observation that Pygo2-L368A animals are viable and fertile rather than mimicking the effect of the HD1 deletion is consistent with the idea that there could be a novel yet-unknown HD1 interactor. Despite the appeal of this explanation, we do not find it very plausible. The main reason for this is the astonishing phenotypic similarities between Pygo2-knockout and the Bcl9/9l-DHD1 mice. For example, both Pygo2 loss and deletion of HD1 induce the arrest of lens development^[8].

Pygo2 overexpression has been shown to drive elevated Wnt signaling as potential causative factor in different types of tumors^{[11][12][13][14]}. It will be necessary to test whether the decreased affinity between Pygo2-L368A and Bcl9/9l is mirrored by an attenuated Wnt signaling. We conjecture that this mouse model could serve as an ideal tool to test, as proof of principle, if the Bcl9/9l-interacting surface of Pygo proteins could constitute the target of small compounds/inhibitors aimed at dampening the aberrantly activated Wnt signaling in these tumors.

Generation of Pygo2 knock-in mouse strain

A knock-in mutant of the Pygo2 locus was generated by standard techniques (inGenious Targeting Laboratory, USA). Briefly, the targeting vector was generated and electroporated into BA1 (C57BL/6 x 129/SvEv) hybrid embryonic stem cells. After selection with the antibiotic G418, surviving clones were expanded for PCR and Southern blotting analyses to confirm recombinant ES clones, followed by validation of the point mutation. mES cells harboring the knock-in allele were microinjected into C57BL/6 blastocysts. Resulting chimeras with a high percentage agouti coat color were mated to wild-type C57BL/6N mice to generate F1 heterozygous offspring. Neo-cassette excision in somatic cells was obtained, as schematized in figure 1A, by crossing heterozygous knock-in animals with mice expressing Flp-recombinase under the control of CMV promoter. Further details are available upon request.

Mouse experiments and genotyping

Mouse experiments were performed in accordance with Swiss guidelines and approved by the Veterinarian Office of the Canton of Zurich, Switzerland. Genotyping of the Pygo2-L368A allele was performed using sequence-specific primers (forward: 5’-CCAGAAACAGAGGTAGG-3’; reverse: 5’-CAGAGGCCAACAAGGAGAGC-3’) that allow the detection of the remaining FRT site after the Flp-mediated Neo-cassette excision (blue arrows in Fig. 1B). Note that the genotyping of the mutant allele in the heterozygous mice produces two very close upper bands; this never occurs when the template genomic DNA is obtained from homozygous mutant mice. We observe this phenomenon for many other transgenic strains for which we design primers spanning the genomic traces of the Neo-cassette deletion. Although we do not have a clear explanation for this, we interpret it as a PCR artifact caused by the simultaneous presence of two highly similar templates within the same reaction (i.e., the possible formation of heteroduplex composed by one wild-type and one mutant strand). Pygo1 knockout alleles were generated in the laboratory of Michel Aguet and previously published^[15].

In vitro pull-down

Adult kidneys were minced in cold PBS and derived cells treated with a hypotonic lysis buffer (20 mM Tris-HCl; 75 mM NaCl; 1.5 mM MgCl2; 1 mM EGTA; 0.5% NP-40; 5% Glycerol). The protein extracts obtained were incubated with a GST-Bcl9-HD1+HD2 recombinant protein (aa1-372 of mouse Bcl9 protein) and mixed with glutathion-conjugated sepharose beads (GE Healthcare). After 4 h of incubation at 4°C on a rotating wheel, the beads were spun down and washed 3 times in lysis buffer. All steps were performed on ice, and all buffers were supplemented with fresh protease inhibitors (Complete, Roche) and 1 mM PMSF. The beads were then treated with “Laemmli buffer," boiled at 85°C for 15 min. Immunoblotting was performed to detect the presence of Pygo2.

Western blot

Total protein extracts and pulled-down proteins were subjected to SDS-PAGE separation and blotting. The following antibody was used: anti-Pygo2 (Novus Biological). Antibody binding was detected by using an anti-rabbit horseradish peroxidase-conjugated IgG and revealed by ECL (GE Healthcare).

Histology

Briefly, 15.5 and 18.5 dpc embryos were extracted and genotyped by a small biopsy obtained from the tail. The embryos were then fixed overnight in PBS and 4% paraformaldehyde, dehydrated with subsequent incubation steps in increasing ethanol concentration (30%, 50% and 70%), and finally embedded in paraffin blocks. H&E staining was performed on 5–10 µm thick paraffin sections.

This work was supported by the Swiss National Science Foundation (SNF) and by grants from the Forschungskredit of the University of Zurich (to C.C.).

We thank George Hausmann, Tomas Valenta, and Bahar Degirmenci for the valuable discussions, and Eliane Escher for helping with sequencing and genotyping.

The authors declare no conflicts of interest.

The mouse experiments were performed in accordance with Swiss guidelines and approved by the Veterinarian Office of the Canton of Zurich, Switzerland.

No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.