Abstract

The combinatorial DNA binding of proteins from the Sox and Oct transcription factor (TF) families are known to be crucial for the pluripotency of stem cells. Experiments have revealed that the Sox/Oct TF pairs bind to different DNA sequences with divergent cooperativities. However, an underlying mechanism of this binding specificity is lacking. Here we use coarse-grained molecular dynamics simulations to study the heterodimerization of Sox2 and Oct4 on designed DNA sequences with a different length of the gap between the individual consensus binding sites. Our results provided a structure-based model to explain the DNA sequence specificity of the Sox/Oct TF pair. With the same model, our results revealed a possible stepwise recognition mechanism for the Sox/Oct dimer.