ABSTRACT
The coactivators CBP (CREBBP) and its paralog p300 (EP300), two conserved multi-domain proteins in eukaryotic organisms, regulate gene expression in part by binding DNA-binding transcription factors. It was previously reported that the CBP/p300 KIX domain mutant (Y650A, A654Q, and Y658A) altered both c-Myb-dependent gene activation and repression, and that mice with these three point mutations had reduced numbers of platelets, B cells, T cells, and red blood cells. Here, our transient transfection assays demonstrated that mouse embryonic fibroblast cells containing the same mutations in the KIX domain and without a wild-type allele of either CBP or p300, showed decreased c-Myb-mediated transcription. Dr. Wright's group solved a 3-D structure of the mouse CBP:c-Myb complex using NMR. To take advantage of the experimental structure and function data and improved theoretical calculation methods, we performed MD simulations of CBP KIX, CBP KIX with the mutations, and c-Myb, as well as binding energy analysis for both the wild-type and mutant complexes. The binding between CBP and c-Myb is mainly mediated by a shallow hydrophobic groove in the center where the side-chain of Leu302 of c-Myb plays an essential role and two salt bridges at the two ends. We found that the KIX mutations slightly decreased stability of the CBP:c-Myb complex as demonstrated by higher binding energy calculated using either MM/PBSA or MM/GBSA methods. More specifically, the KIX mutations affected the two salt bridges between CBP and c-Myb (CBP-R646 and c-Myb-E306; CBP-E665 and c-Myb-R294). Our studies also revealed differing dynamics of the hydrogen bonds between CBP-R646 and c-Myb-E306 and between CBP-E665 and c-Myb-R294 caused by the CBP KIX mutations. In the wild-type CBP:c-Myb complex, both of the hydrogen bonds stayed relatively stable. In contrast, in the mutant CBP:c-Myb complex, hydrogen bonds between R646 and E306 showed an increasing trend followed by a decreasing trend, and hydrogen bonds of the E665:R294 pair exhibited a fast decreasing trend over time during MD simulations. In addition, our data showed that the KIX mutations attenuate CBP's hydrophobic interaction with Leu302 of c-Myb. Furthermore, our 500-ns MD simulations showed that CBP KIX with the mutations has a slightly lower potential energy than wild-type CBP. The CBP KIX structures with or without its interacting protein c-Myb are different for both wild-type and mutant CBP KIX, and this is likewise the case for c-Myb with or without CBP, suggesting that the presence of an interacting protein influences the structure of a protein. Taken together, these analyses will improve our understanding of the exact functions of CBP and its interaction with c-Myb.
