Transgenerational inheritance of epigenetic states at the murine Axin(Fu) allele occurs after maternal and paternal transmission

Abstract

Phenotypic variation that cannot be explained by genetic or environmental heterogeneity has intrigued geneticists for decades. The molecular basis of this phenomenon, however, is largely a mystery. Axin-fused (Axin(Fu)), first identified in 1937, is a classic example of a mammalian allele displaying extremely variable expression states. Here we demonstrate that the presence or absence of its characteristic phenotype, a kinked tail, correlates with differential DNA methylation at a retrotransposon within Axin(Fu) and identify mutant transcripts arising adjacent to the retrotransposon LTR that are likely to be causative of the phenotype. Furthermore, the epigenetic state at Axin(Fu) can be inherited transgenerationally after both maternal and paternal transmission. This is in contrast to epigenetic inheritance at the murine agouti-viable yellow (A(vy)) allele, which occurs through the female only. Unlike the egg, the sperm contributes very little (if any) cytoplasm to the zygote, and therefore paternal inheritance at Axin(Fu) argues against the possibility that the effects are due to cytoplasmic or metabolic influences. Consistent with the idea of transgenerational inheritance of epigenetic marks, we find that the methylation state of Axin(Fu) in mature sperm reflects the methylation state of the allele in the somatic tissue of the animal, suggesting that it does not undergo epigenetic reprogramming during gametogenesis. Finally, we show that epigenetic inheritance is influenced by strain background. These findings enable us to propose a model for transgenerational epigenetic inheritance in mammals.

Figure 1

The tail-kink phenotype correlates with the methylation state at the Axin^Fu allele. (A) An Axin^Fu/+ mouse of the 129P4/RrRk strain with a kinked tail. (B Upper) The intron 6 probe (checkered box) and the PstI (P) and HhaI (H) sites in the region of the IAP insertion within intron 6 of Axin^Fu. (B Lower) Tail DNA was digested with PstI alone or in combination with HhaI (sensitive to CpG methylation) and then transferred and hybridized with the intron 6 probe. Digestion with PstI alone produces a 917-bp fragment at the wild-type axin allele and a 660-bp fragment at the Axin^Fu. In penetrant Axin^Fu/+ mice this band is digested by HhaI, indicating hypomethylation. In silent Axin^Fu/+ mice the 660-bp band is not digested by HhaI, indicating hypermethylation. The data shown are representative of seven penetrant and eight silent mice. The membrane was rehybridized with a probe to the mouse α-globin gene to ensure equal digestion of all samples. (C) Methylation profiles of 13 CpG dinucleotides at the IAP–intron 6 junction of the Axin^Fu allele. The methylation state of each CpG was obtained by sequencing PCR clones from bisulfite-treated genomic DNA. Open and filled circles represent unmethylated and methylated CpGs, respectively. The numbers in the parentheses represent the proportion of methylated CpG sites relative to all CpG sites for which sequence was obtained. Each line represents the sequence of one clone, and each block of clones represents the data from one mouse. The data shown are from four penetrant (P) and four silent (S) mice but are representative of those obtained for six penetrant and five silent mice.

Figure 2

Mutant transcripts expressed in penetrant mice. (A) Northern blot of adult kidney poly(A)⁺ RNA from wild-type (+/+) silent and penetrant Axin^Fu/+ adult mice. The blot was probed with the intron 6 probe described in Fig. 1. The GAPDH control of the same blot is also shown. The data presented are representative of that obtained for three wild-type, four penetrant, and four silent mice. (B) RLM-RACE (data not shown) revealed several mutant transcripts (≈2 kb) in penetrant mice only that initiate within intron 6 (arrows), +33, +64, and +97 bp downstream of the LTR at the 3′ end of the IAP insertion. ATG₁ and ATG₃ are out-of-frame with the axin-coding sequence. Translation initiated from ATG₂ would result in a 39-aa peptide only. ATG₄ is in-frame and could be used to generate a truncated form of the wild-type Axin protein that lacks the amino-terminal domain encoded by exons 1–6. (C) An in vitro transcription/translation assay was performed by using a construct containing a 1,109-bp cDNA insert derived from the Axin^Fu allele. The insert encompasses a region starting in intron 6 (marked with * in B), 40 bp downstream of LTR, and ending in exon 10 at the wild-type axin stop codon. The in vitro-translated protein that results from initiation of translation at ATG₄ is indicated. The experiment was repeated after site-directed mutagenesis of the ATG₄ site to CTC. Transcription/translation of the luciferase gene was used as a positive control.

Figure 3

Inheritance of parental phenotype. Axin^Fu/+ 129P4/RrRk mice of the indicated phenotypes were crossed with congenic +/+ mice, and the percentage of offspring of each phenotype was scored. The number of total Axin^Fu/+ progeny of each cross is indicated (n); +/+ offspring have been omitted from the pedigrees. (A) The proportions of offspring phenotypes arising from penetrant and silent sires differ significantly (P < 0.05). The data were collated from 18 different matings. (B) The proportions of offspring phenotypes arising from penetrant and silent dams differ significantly (P < 0.05). The data were collated from 29 different matings. (C) The data presented in A were reanalyzed according to the severity of the penetrant phenotype. The severity of the phenotype of the sire influences the severity of the phenotypes in the penetrant offspring (P < 0.05). Silent Axin^Fu/+ and wild-type mice have been omitted from the pedigree.

Figure 4

The methylation state of Axin^Fu and A^vy in mature sperm reflects the methylation state of the alleles in the somatic tissue of the animal. (A) DNA from mature sperm of penetrant (P) and silent (S) 8-week-old Axin^Fu/+ males was subjected to bisulfite sequencing. One line represents the sequence from one clone and the methylation profile of the LTR/intron 6 region (refer to Fig. 1D) in one mature sperm. Each block of lines represents the methylation data of sperm DNA obtained from one adult male mouse. (B) DNA from mature sperm of yellow (Y) and agouti (A) 8-week-old A^vy/a males was subjected to bisulfite sequencing. All the CpG sites shown for the A^vy allele are contained within the LTR.

Figure 5

Transgenerational epigenetic inheritance is influenced by strain background. (A) A^vy/a C57BL/6J males of the indicated phenotypes (eight yellow sires and six agouti sires) were mated to 129P4/RrRk females (129P4/RrRk mice are A^w/A^w at the agouti locus; A^vy is dominant over A^w), and the offspring's coat color was recorded. The number of total A^vy/A^w progeny of each cross is indicated (n); mice that did not carry the A^vy allele have been omitted from the pedigree. The phenotype of the sire influences the range of phenotypes in the offspring (P < 0.05). (B) Axin^Fu/+ 129P4/RrRk sires of the indicated phenotypes (five penetrant sires and five silent sires) were mated to C57BL/6J dams (+/+ at the axin locus), and the F₁ offspring's tail phenotype was recorded. The number of total Axin^Fu/+ progeny from each cross is indicated (n); +/+ mice have been omitted from the pedigree. The proportions of phenotypes arising from penetrant and silent sires do not differ significantly (P = 0.48).

Figure 6

A model for transgenerational epigenetic inheritance. We can consider a hypothetical allele that is marked at a single site by methylation (m). The silent phenotype corresponds to a methylated state, and the penetrant phenotype corresponds to the nonmethylated state. During gametogenesis (i), the marks escape demethylation, and therefore the overall epigenetic profile of the allele in mature sperm (five gametes shown for each male) correlates with the epigenetic state of the allele in somatic tissue of that animal. After fertilization (ii), the inability of the cell to erase marks at initially unmethylated alleles is inconsequential, but some alleles that were methylated initially will not be cleared completely. Ultimately, this will result in some memory of the epigenetic state of the parent's allele. This transgenerational epigenetic inheritance will not occur if the epigenetic marks are cleared completely postfertilization (not indicated in the figure), e.g., paternal copies of Axin^Fu and A^vy in the C57BL/6J-fertilized egg.