CpG-binding protein is a nuclear matrix- and euchromatin-associated protein localized to nuclear speckles containing human trithorax. Identification of nuclear matrix targeting signals.

CpG-binding protein (CGBP) binds unmethylated CpG dinucleotides and is essential for mammalian development. CGBP exhibits a punctate nuclear localization correlated with 4,6-diamidino-2-phenylindole light regions and is excluded from metaphase chromosomes. The distribution of CGBP is distinct from the heterochromatin-associated proteins MBD1, methyl-CpG-binding protein 2, and HP1alpha. Some CGBP-containing nuclear speckles co-localize with splicing factor SC-35 and actively transcribed regions of the genome, whereas most CGBP co-localizes with acetylated histones, indicating that CGBP is localized to active chromatin. CGBP contains two nuclear localization signals that are insufficient to direct punctate subnuclear distribution. Instead, localization of CGBP to nuclear speckles requires signals within the acidic, basic, and coiled-coil domains. CGBP associates with the nuclear matrix, and fragments of CGBP that fail to associate with the nuclear matrix fail to localize to nuclear speckles and exhibit reduced transcriptional activation activity. Mutated versions of CGBP that lack DNA binding activity exhibit a normal nuclear distribution, suggesting that CGBP accumulates at nuclear speckles as a result of protein/protein interactions. Importantly, the subcellular distribution of CGBP is identical to human trithorax, suggesting that these proteins may be components of a multimeric complex analogous to the histone-methylating Set1 complex of Saccharomyces cerevisiae that contains CGBP and trithorax homologues.