Mammalian X-chromosome inactivation is an impressive example of epigenetic gene regulation, whereby the majority of genes on the approximately 160 Mb X chromosome are silenced in a strictly cis-limited fashion. In this review we will discuss the important players involved in the silencing process. The process is initiated by transcription and cis-localization of the non-coding XIST RNA, which then recruits many of the epigenetic features generally associated with heterochromatin, including histone modifications, histone variants and DNA methylation.

BackgroundThe transcriptional silencing of one X chromosome in eutherians requires transcription of the long non-coding RNA gene, XIST. Many regulatory elements have been identified downstream of the mouse Xist gene, including the antisense Tsix gene. However, these elements do not show sequence conservation with humans, and the human TSIX gene shows critical differences from the mouse. Thus we have undertaken an unbiased identification of regulatory elements both downstream and upstream of the human XIST gene using DNase I hypersensitivity mapping.ResultsDownstream of XIST a single DNase I hypersensitive site was identified in a mouse undifferentiated ES cell line containing an integration of the human XIC region. This site was not observed in somatic cells. Upstream of XIST, the distance to the flanking JPX gene is expanded in humans relative to mice, and we observe a hypersensitive site 65 kb upstream of XIST, in addition to hypersensitive sites near the XIST promoter. This -65 region has bi-directional promoter activity and shows sequence conservation in non-rodent eutheria.ConclusionsThe lack of regulatory elements corresponding to human TSIX lends further support to the argument that TSIX is not a regulator of XIST in humans. The upstream hypersensitive sites we identify show sequence conservation with other eutheria, but not with mice. Therefore the regulation of XIST seems to be different between mice and man, and regulatory sequences upstream of XIST may be important regulators of XIST in non-rodent eutheria instead of Tsix which is critical for Xist regulation in rodents.

The centromere is a complex structure required for equal segregation of newly synthesised sister chromatids at mitosis. One of the significant objectives in centromere research is to determine the complete repertoire of protein components that constitute the kinetochore. Here, we identify a novel centromere protein using a centromere-positive autoimmune serum from a patient with watermelon stomach disease. Western blot and screening of a lambda phage expression library revealed a 60-kDa protein, ZNF397. This protein belongs to the classical Cys(2)His(2) group of the zinc-finger protein superfamily and contains two conserved domains: a leucine-rich SCAN domain and nine Cys(2)His(2) zinc fingers. Bioinformatic analysis shows that ZNF397 is conserved in placental mammals. Stable GFP:ZNF397-expressing human cells show co-localisation of ZNF397 with the constitutive centromere protein CENP-A during interphase and early prophase. Deletion and domain-swap constructs indicate that the SCAN domain is necessary but not sufficient for centromere localisation. Gene-knockout studies in mice using the mouse orthologue (Zfp397) reveal that ZNF397 is a non-essential protein. These properties define ZNF397 as a member of a new class of interphase to early prophase-specific and SCAN domain-containing mammalian centromere protein. The possible role of this protein in transcription at the centromere is discussed.

The silent X chromosome in mammalian females is a classic example of facultative heterochromatin, the term highlighting the compacted and inactive nature of the chromosome. However, it is now clear that the heterochromatin of the inactive X is not homogeneous--as indeed, not all genes on the inactive X are silenced. We summarize known features and events of X inactivation in different mouse and human model systems, and highlight the heterogeneity of chromatin along the inactive X. Characterizing this heterogeneity is likely to provide insight into the cis-acting sequences involved in X chromosome inactivation.