, the Pygopus
gene is a key mediator of canonical Wnt signaling. In one study, 12 distinct measures of Wnt signaling in Pygo
mutants were performed, including analysis of leg, wing, and eye imaginal discs. In 2 cases there was a significant reduction of Wnt signaling and in 10 cases a complete block [10
]. A second study in Drosophila
examined the effects of Pygo
mutation on cuticle patterning, midgut constriction, central nervous system, and cardiac development, and concluded that "Pygo
is an essential component in the Wg signal transduction pathway". [8
Given these results in Drosophila
, the relatively mild phenotypes of mice with targeted Pygo1
, or double Pygo1
mutations were striking. Pygo2
mutants developed to birth and showed limited abnormalities, while Pygo1
homozygous mutants were normal and fertile. Furthermore, there was no detected synergism in the phenotype of the double Pygo1
mutant, although in several tissues the BAT-gal Wnt reporter did show more a severe reduction in expression. One possible explanation of these unexpected results is a failure of the gene targeting to eliminate functioning of the Pygo1
genes. The Pygopus
deletion alleles described in this report, however, are almost certainly functional nulls. In Drosophila
, it has been shown that the PHD domain is absolutely necessary for Pygopus function in Wg signaling. The PHD domain is 60 amino acids with seven cysteines and a histidine, predicted to chelate two zinc ions. PHD domains are found in diverse proteins, including transcription factors, and have been implicated in chromatin remodeling and protein-protein interactions [8
]. In Drosophila
allele, with a single missense mutation converting amino acid 802 in the PHD domain from cysteine to tyrosine, loses Wnt signaling function in both embryogenesis and imaginal disc development [8
]. The Pygo1
mutant alleles made in this report carried deletions of the entire PHD domains, as well as most other coding sequences. For the Pygo1
gene, the coding region for 372 of 417 total amino acids was deleted, and for the Pygo2
gene, we deleted coding for 354 of 405 amino acids. It is therefore very unlikely that the relatively mild phenotypes observed were the result of residual function of the targeted alleles.
We focused our analysis on the developing kidney, in which Wnt signaling has been shown to be of critical importance in several stages of nephrogenesis. Wnt9b is made by the ureteric bud and induces the metanephric mesenchyme to undergo nephrogenesis [2
]. Downstream of Wnt9b is Wnt4 [2
], which is made by the metanephric mesenchyme, and is also required for nephrogenesis [3
]. In addition, Wnt11 is produced by the ureteric bud tips and induces GDNF expression in the metanephric mesenchyme [4
In this report we describe a novel Wnt function in kidney development. The BAT-gal transgene reporter indicated the presence of canonical Wnt signaling in the ureteric bud and its derivatives in the developing kidney. Further, in the Pygo2 mutants this signal was lost, suggesting significant reduction of Wnt signaling. In addition, we observed a resulting decrease in ureteric tip density, reduced kidney size and altered morphology of the ureteric tree in mutants, indicating a role for canonical Wnt signaling in branching morphogenesis of the ureteric bud. While the simplest interpretation is direct Wnt signaling to the ureteric bud, it remains possible that the observed abnormalities are the result of indirect effects, with altered Wnt signaling to the metanephric mesenchyme then affecting mesenchyme to ureteric bud signaling.
mutant phenotype of reduced branching morphogenesis of the ureteric bud is surprisingly similar to that previously reported for Wnt11
]. The underlying mechanisms, however, are likely to be distinct. Wnt11 is generally thought to act through a noncanonical pathway, (although exceptions have been noted [31
]), whereas the Pygopus
genes promote canonical Wnt signaling. Further, the Wnt11
mutants showed an altered feedback loop between Wnt signaling from the ureteric bud to the mesenchyme and GDNF signaling from the mesenchyme to the bud, whereas in the Pygo1
mutants, we observed disrupted Wnt signaling in the ureteric bud.
mutants also showed an expansion of the zone of thickened mesenchyme that caps the ureteric bud. Nevertheless, the mutant metanephric mesenchyme formed nephrons normally. This was particularly interesting, as Wnt9b signaling from the ureteric bud to the metanephric mesenchyme has been shown to induce nephrogenesis via canonical Wnt signaling [2
The results in this report indicate a striking and surprising evolutionary divergence of Pygopus function between Drosophila and mammals. In Drosophila, the Pygopus gene is often required for canonical Wnt signaling, while in mammals the Pygo1/Pygo2 genes appear to play a smaller role in canonical Wnt signaling. The BAT-gal transgene reporter of canonical Wnt signaling showed reduced but generally not absent signal in Pygo1/Pygo2 mutant embryos, with tissue-specific variation in level of diminution. In addition, the kidneys were not unique in showing surprisingly normal development in Pygo1/Pygo2 mutants. Indeed, organogenesis generally proceeded without detectable abnormality, with few exceptions. These results suggest that the proteins encoded by mammalian Pygopus genes are often mere modulators of canonical Wnt signaling intensity, and not essential components.
The microarray results further support this conclusion. Whereas the BAT-gal transgene reporter monitors the expression level of only one Wnt responsive promoter, the microarray allows the analysis of the activity levels of promoters of all genes. Interestingly, the mutant kidneys showed gene-expression profiles surprisingly similar to wild type. Some genes did, however, show expression differences, and a high percentage of these differences could be validated by real-time PCR with independent biological samples. Assuming that the Pygopus genes in mammals are dedicated to canonical Wnt signaling, as has been previously shown in Drosophila, the genes with expression differences represent candidate Wnt targets (direct or indirect) in kidney development. We would predict these genes to show greater changes in expression level in a developing kidney with a more complete removal of canonical Wnt signaling.