We have been interested in the protein CTCF, which we first identified some years ago as having properties of an insulator, blocking interaction between enhancers and promoters when placed between them. We demonstrated that this activity plays an important role in regulating parent of origin allele-specific gene expression at the Igf2/H19 imprinted locus. Work in recent years has shown that a principal mode of action of CTCF is to stabilize interactions between CTCF binding sites on DNA, leading to formation of loop domains. Depending on the geometry of the interactions such loops can either exclude an enhancer leading to insulation, or bring enhancer and promoter closer together, leading to activation. DNA within the cell nucleus is packaged into chromatin, and further organized into topologically associated domains (TADs) separating active and inactive genomic regions. The establishment and maintenance of TADs requires the protein CTCF, and we are interested in identifying and studying the interactions of CTCF with the protein and nucleic acid partners recruited for insulator function. We have shown that the N- and C-terminal domains that flank the DNA binding 11 zinc fingers of CTCF appear to be intrinsically disordered explaining, in part, the large number of CTCF binding partners identified in other studies. Current work focuses on further characterizing the physical nature of these domains, identifying partners that bind with high affinity, and studying the complexes formed. In the course of searching for CTCF interacting proteins we identified a strong interaction, both in vitro and in vivo, between CTCF and the centromeric protein CENP-E. We were able to identify, within the repeat DNA sequences of the pericentromeric region, unusual CTCF sites that recruited CENP-E. Binding of CENP-E is dependent on binding of CTCF. These results suggest a novel function for CTCF in recruiting CENP-E to the centromere, a novel role for these unusual CTCF binding sites, and perhaps a novel and functionally important organized structure centered on CTCF. We suggest that the CTCF/CENP-E complex may be involved in structural organization of the pericentromeric regions. We have recently reported an interaction between CTCF and the DEAD box helicase, p68. We found that a long non-coding RNA, SRA, which forms a complex with p68, is essential for this interaction. We have investigated the nature of the p68-SRA interaction, and the role of the RNA in insulator activity of CTCF. We have now extended these studies to a detailed examination of SRA function. Many long non-coding RNAs recruit either the Polycomb complex, PRC2, which delivers histone modifications that are associated with silencing of gene expression, or one of the Trithorax group related complexes, which delivers activating modifications. We have evidence that a single SRA molecule may bind both complexes simultaneously, a novel function for long non-coding RNAs. This result suggests that sites on DNA at which SRA is bound could display both silencing and activating marks. Such bivalent sites are known to occur in pluripotent stem cells. We have mapped SRA binding sites across the genome in an induced pluripotent stem cell, and show that such sites do exist. Furthermore, we show that p68 binding to SRA alters the affinity for TrxG as opposed to PRC2 complexes. All of these results relate to the role of chromatin structure, histone modifications, and long range organization of the genome in cell function, and are in turn related to questions of normal and abnormal cell metabolism and cell division.
Showing the most recent 10 out of 12 publications