Tissue-specific reactivation of gene expression at an imprinted locus

Genomic imprinting is the phenomenon where the expression pattern of an allele at a locus differs depending on the allele's parent of origin. In most cases, one of the two alleles is transcriptionally silent. Recent empirical work has shown some genes to be imprinted in a tissue-specific manner, where the silenced allele becomes reactivated in particular cell lineages during development. Here I describe an evolutionary model of tissue-specific transcriptional reactivation. The model describes the relationships among various inclusive fitness functions and phenotypic effects necessary for natural selection to favor the epigenetic reprogramming required for this sort of reactivation. and makes predictions regarding the nature and magnitude of phenotypic and fitness consequences of mutations in particular somatic tissues. In particular, if an imprinted gene is reactivated in one of two tissues that interact in producing a particular phenotype, expression of the gene in those two tissues is expected to have opposite phenotypic effects. The model predicts that in some cases, mutations affecting the silenced allele at all imprinted locus may be phenotypically more severe than those affecting the expressed allele. These predictions are contrasted with those of an alternative explanation for reactivation: protection against deleterious recessive somatic mutations. The inclusive-fitness model of reactivation indicates that the intragenomic conflicts present in the parental germ lines and developing embryo persist though adult life, and can have complex effects on phenotypes and patterns of gene expression in somatic tissues. (c) 2005 Elsevier Ltd. All rights reserved.