Characteristics of the CArG-SRF binding context in mammalian genomes

Serum response factor (SRF), a member of the MADS family, binds a 10-bp cis element known as the CArG box. However, despite our extensive knowledge of SRF and the CArG box, limited information is available on the CArG-SRF binding context or how CArG flanking sequences are defined. We used statistical tests and computer simulation to find characteristics of the CArG-SRF binding context. Based on the combination of published literature and a search of DBTSS, 150 and 136 functional CArG boxes together with 10 bp flanking on each side were found in mouse and human genomes, respectively. By statistical analysis of the 30 positions we found some new conserved positions of interest (P < 0.005) such as -15, -8, and +8, in addition to the ten highly conserved positions of the CArG box. Intriguingly, studies comparing the flanking positions between consensus CArG boxes and CArG-like boxes showed that there are more conserved positions in the latter. Moreover, CpG content within the CArG-SRF binding context is much higher than that within introns. Collectively, these results suggest that there are some special pre-existing features in the flanking sequences of functional CArG boxes probably contributing to SRF selectively recognizing and binding to the functional CArG from millions of functionless CArG boxes in mammalian genomes. This is a significant step toward understanding the mechanism of transcriptional regulation of SRF-dependent genes.