CTCF has been discovered as the first example of a true "multivalent" nuclear factor capable of specific recognition of varying -50-bp-long DNA sites with no single "consensus" sequence that could describe all of the binding sites [reviewed in (1)]. The ubiquitously expressed gene is conserved from Drosophila through the mammalian radiation showing, within the 11 Zn-finger DNA-binding domain, - 100% amino acid identity in all vertebrates. Remarkably, chicken CTCF can bind to all Drosophila DNA sequences identified as targets for Drosophila CTCF, but Drosophila CTCF cannot bind to all vertebrate CTCF targets. This suggests that during evolution, CTCF was co-opted to serve an increasing number of functions. In mammals, these functions have apparently converged because amino acid sequences of all mammalian CTCF proteins are near-identical throughout the entire length of the proteins.
Mammalian CTCF activities are now known to include transcriptional activation and repression, hormone-inducible gene silencing, creation of constitutive chromatin insulators or boundaries, and functional "reading" of imprinted states. In the H19 imprinting control region (ICR), the methylation-sensitive binding of CTCF to the H19 ICR insulator discriminates maternal versus paternal alleles in vivo (2). It was suggested that parent-of-origin-dependent, allele-specific DNA methylation that modulates interactions of dissimilar CpG-rich sites within diverged ICRs with the multivalent CTCF may be a common denominator in epigenetic regulation of different imprinted loci. Indeed, CTCF has been shown to interact with the Prader-Willi/Angelman syndrome ICR and with the differentially methylated CpG island of the murine Kcnql gene. Moreover, the choice/imprinting center of the XIST antisense that affects silencing of one X chromosome also depends on epigenetically controlled CTCF binding (3). Abnormal methylation patterns of some CTCF sites were found in cancer and in other diseases, including severe congenital forms of myotonic dystrophy.
Although the mechanism for reading at least some imprinting marks in somatic cells may involve CTCF, there was no explanation for the recognition and re-establishment of the imprinting marks during epigenetic reprogramming in the germ line. Recently, our lab and collaborators have identified a CTCF paralog, termed BORIS for brother of the regulator of imprinted sites, that is expressed only in the testis (4). BORIS has the same exons encoding the 11 Zn-finger domain as mammalian but not avian CTCF genes, and hence interacts with similar cis elements in vitro but with selected CTCF sites in vivo. Normally, CTCF and BORIS are expressed in a mutually exclusive pattern that correlates with the re-setting of methylation marks during male germ cell differentiation, suggesting that BORIS directs epigenetic reprogramming at selected CTCF targets. While CTCF has properties of a tumor suppressor at 16q22, expression of BORIS in normal somatic cells promotes cell growth which can lead to transformation. BORIS is not only abnormally activated in a range of human cancers, but also maps to the cancer-associated amplification region at 20q13. The sibling rivalry occasioned by aberrant expression of BORIS may interfere with normal functions of CTCF (including growth suppression) and contribute to epigenetic deregulation, which is a common feature in cancer.
Ohlsson, R., et al. Trends Genet. 17: 520-527, 2001.
Kanduri, C., et al. Curr. Biol. 10: 853-856, 2000.
Chao, W., et al. Science 295: 345-347, 2002.
Loukinov, D.I., et al. Proc. Natl. Acad. Sci. USA 99: 6806-6811, 2002.