Search

X chromosome inactivation (XCI) is the phenomenon through which one of the two X chromosomes in female mammals is silenced to achieve dosage compensation with males. XCI is a highly complex, tightly controlled and developmentally regulated process. The mouse undergoes two forms of XCI: imprinted, which occurs in all cells of the preimplantation embryo and in the extraembryonic lineage, and random, which occurs in somatic cells after implantation. This review presents results and hypotheses that have recently been proposed concerning important aspects of both imprinted and random XCI in mice. We focus on how imprinted XCI occurs during preimplantation development, including a brief discussion of the debate as to when silencing initiates. We also discuss regulation of random XCI, focusing on the requirement for Tsix antisense transcription through the Xist locus, on the regulation of Xist chromatin structure by Tsix and on the effect of Tsix regulatory elements on choice and counting. Finally, we review exciting new data revealing that X chromosomes co-localize during random XCI. To conclude, we highlight other aspects of X-linked gene regulation that make it a suitable model for epigenetics at work.

Cloning by somatic cell nuclear transfer requires that epigenetic information possessed by the donor nucleus be reprogrammed to an embryonic state. Little is known, however, about this remodeling process, including when it occurs, its efficiency, and how well epigenetic markings characteristic of normal development are maintained. Examining the fate of epigenetic information associated with imprinted genes during clonal development offers one means of addressing these questions. We examined transcript abundance, allele specificity of imprinted gene expression, and parental allele-specific DNA methylation in cloned mouse blastocysts. Striking disruptions were seen in total transcript abundance and allele specificity of expression for five imprinted genes. Only 4% of clones recapitulated a blastocyst mode of expression for all five genes. Cloned embryos also exhibited extensive loss of allele-specific DNA methylation at the imprinting control regions of the H19 and Snprn genes. Thus, epigenetic errors arise very early in clonal development in the majority of embryos, indicating that reprogramming is inefficient and that some epigenetic information may be lost.

An intriguing characteristic of imprinted genes is that they often cluster in large chromosomal domains, raising the possibility that gene-specific and domain-specific mechanisms regulate imprinting. Several common features emerged from comparative analysis of four imprinted domains in mice and humans: (a) Certain genes appear to be imprinted by secondary events, possibly indicating a lack of gene-specific imprinting marks; (b) some genes appear to resist silencing, predicting the presence of cis-elements that oppose domain-specific imprinting control; (c) the nature of the imprinting mark remains incompletely understood. In addition, common silencing mechanisms are employed by the various imprinting domains, including silencer elements that nucleate and propagate a silent chromatin state, insulator elements that prevent promoter-enhancer interactions when hypomethylated on one parental allele, and antisense RNAs that function in silencing the overlapping sense gene and more distantly located genes. These commonalities are reminiscent of the behavior of genes subjected to, and the mechanisms employed in, dosage compensation.