SP100B, a repressor of gene expression preferentially binds to DNA with unmethylated CpGs

SP100A and SP100B are mammalian nuclear proteins encoded by alternatively-spliced transcripts from the SP100 gene. The N-terminal portion of SP100B (aa 1-476) is identical to SP100A and contains an HP1 interaction domain. The C-terminal portion of SP100B (aa 477-688) contains an HMG2 interaction domain and a SAND domain. The SAND domain is a DNA-binding domain identified in several nuclear proteins involved in transcriptional regulation. We have previously reported that SP100B represses expression of genes present on transfected DNA in a SAND domain-dependent manner. The goal of the present study was to characterize the DNA binding properties of full-length SP100B expressed in mammalian cells. SP100B associated with DNA whereas SP100A did not. The SP100B SAND domain was essential for DNA binding. Deletion of the HP1- or HMG2-binding domain had no effect on DNA binding. SP100B preferentially associated with sequences containing CpG dinucleotides. Our results did not reveal any preference of SP100B for bases flanking CpG dinucleotides. The number of CpGs in a DNA sequence and spacing between CpGs influenced SP100B binding, suggesting that multimers of SP100B bind DNA in a cooperative manner. Binding of SP100B was abrogated by methylation of the cytosine residue within the context of the CpG dinucleotide. We propose that the preference of SP100B for non-methylated CpGs provides a mechanism to target SP100B to foreign DNA, including plasmid DNA or viral DNA genomes, most of which are hypomethylated.