Expression of transgenes in mice, when examined with assays that can distinguish individual cells, is often found to be heterocellular, or variegated. Line‐to‐line variations in expression of a transgene may be due largely to differences in the proportion of cells in which it is expressed. Variegated silencing by centromeric heterochromatin is well described, but other factors may also affect transgene silencing in mice. Tandem arrays of transgenes themselves form heterochromatin, and some cell lineages may tend to silence transgenes because of (...) extensive facultative heterochromatin in their nuclei. The cis‐acting transcriptional control elements within a transgene inhibit silencing, and strainspecific differences in chromatin proteins may strongly influence the extent of variegation. The accessibility of multiple differentiated cell lineages in mice suggests that they may provide a tool for dissecting the role of chromatin‐mediated silencing in cell differentiation and tissue‐specific gene expression. (shrink)

The early discovery of cis‐regulatory elements able to promote transcription of genes over large distances led to the postulate that elements, termed insulators, should also exist that would limit the action of enhancers, LCRs and silencers to defined domains. Such insulators were indeed found during the past fifteen years in a wide range of organisms, from yeast to humans. Recent advances point to an important role of transcription factors in insulator activity and demonstrate that the operational observation of an insulator (...) effect relies on a delicate balance between the “efficiency” of the insulator and that of the element to be counteracted. In addition, genuine insulator elements now appear less common than initially envisaged, and they are only found at loci displaying a high density of coding or regulatory information. Where this is not the case, chromatin domains of opposing properties are thought to confront each other at “fuzzy” boundaries. In this article, we propose models for both fixed and fuzzy boundaries that incorporate probabilistic and dynamic parameters. BioEssays 26:523–532, 2004. © 2004 Wiley Periodicals, Inc. (shrink)

Nucleosomes “talk” to each other about their modification state to form extended domains of modified histones independently of the underlying DNA sequence. At the same time, DNA elements promote modification of nucleosomes in their vicinity. How do these site-specific and histone-based activities act together to regulate spreading of histone modifications along the genome? How do they enable epigenetic memory to preserve cell identity? Many models for the dynamics of repressive histone modifications emphasize the role of strong positive feedback loops, which (...) reinforce histone modifications by recruiting histone modifiers to preexisting modifications. Recent experiments question that repressive histone modifications are self-sustained independently of their genomic context, thereby indicating that histone-based feedback is relatively weak. In the present review, current models for the dynamics of histone modifications are compared and it is suggested that limitation of histone-based feedback is key to intrinsic confinement of spreading and coexistence of short- and long-term memory at different genomic loci. See also the video abstract here: https://youtu.be/3bxr_xDEZfQ. Communication between nucleosomes and site-specific recruitment of histone modifiers act together to enable spreading and epigenetic memory of histone modifications. This review essay compares current models that describe these phenomena and outline a mechanism for how epigenetic short- and long-term memory can coexist in the same cell. (shrink)