Identification of evolutionarily conserved segments of homology within DNA

niaid.nih.gov Identification of evolutionarily conserved segments of homology within DNA regulatory regions, ‘phylogenetic footprints’, provides a means for discovering functionally relevant site-specific DNAbinding proteins. This approach has its drawbacks, however, because it depends on the conservation of sequences in pivotal cis regulatory elements. In some instances, functionally analogous DNA regions, such as replication origins, centromeres and gene promoters, are similarly regulated in different species despite high sequence divergence. For example, although the noncoding regulatory regions of the avian and mammalian MYC genes show little sequence identity, they are nevertheless similarly regulated during cell proliferation and differentiation. One simple solution to this puzzle was suggested by one of us more than ten years ago1–4. It assumed that an evolutionarily conserved factor with the ability to recognize highly diverged sequences might exist, giving similar functions to dissimilar sequences. Identification, cloning and characterization of a nuclear protein, CTCF (for ‘CCCTC-binding factor’), that binds to a number of different sequences in the human, mouse and avian MYC promoters3–6, and negatively regulates MYC in both mammals and birds4,6,7 provided evidence for the existence of regulatory factor(s) with multiple DNA sequence specificities. In addition, a silencer protein (NeP1) was discovered independently, which binds to the chicken lysozyme silencer8,9. Purification and protein sequencing of NeP1 showed it to be CTCF (Ref. 10) and thus identified a new AT-rich CTCF target site in the lysozyme silencer that is markedly dissimilar with GC-rich target sites in MYC. In two subsequent, independent studies, proteins binding to the APPβ site of the human amyloid precursor protein (APP) gene promoter11 and to the FII site within the HS4 enhancer-blocking region of the chicken β-globin locus12 were again purified by their ability to interact specifically with the cognate DNA targets. Once sequenced, they both turned out to be CTCF (Refs 13,14). To highlight the unusual ability of this factor to recognize multiple target sites, we suggested the term ‘multivalent’CTCF (Refs 6,10). In keeping with the multivalent character of CTCF, a growing number of different target sites are now implicated in a variety of regulatory functions, ranging from promoter repression and activation, to the creation of hormone-responsive silencers and enhancer-blocking and/or boundary elements. This article presents the first overview of CTCF structure and function, and discusses recent results that highlight links between CTCF, evolution, epigenetics and disease.