Zheng Liu and Diannan Lu. Department of Chemical Engineering, Tsinghua University, Beijing, 100084, China
One challenge to protein refolding is to dissociate the non-native disulfide bonds and promote the formation of the native ones. Here we proposed a coarse-grained off-lattice model with disulfide bonds and simulated the shuffling of the disulfide bonds. It was shown that the introduction of disulfide bonds enhances the conformational stability while reduces the foldability of the model protein, as compared to its counterpart without the disulfide bonds. The shuffling of disulfide bonds occurs at a suitable redox environment, in which the non-native disulfide bonds can rearrange to the native ones. The folding trajectory suggested that disulfide bonds located in the hydrophobic core form before protein collapsing while those at protein surface form after protein rearrangement. A reductive environment at an early stage of folding favors the shuffling that gives the native disulfide bonds in the hydrophobic core. For those disulfide bonds at protein surface, however, an oxidative environment at late stage of folding is requested. Thus a dynamic redox environment, i.e., reductive environment to oxidative one, can intensify the shuffling of disulfide bonds and result in an improved yield of native protein. A dynamic redox environment can facilitate disulfide shuffling to native state and thus improve the folding yield. To validate the simulation results, we have also conducted protein refolding experiments using lysozyme as model protein. It is shown that there is optimal ratio of GSSG to GSH, which corresponds to optimal redox environment, for the effective protein folding. We have also conducted lysozyme refolding under a dynamic redox environment, i.e., form reductive environment at early folding stage to oxidative one at late folding stage. It is shown that the folding yield under this dynamic redox environment can effectively improve the folding yield than those at constant redox environment, agreeing well with above predictions by molecular dynamic simulations.