OBJECTIVETo establish a prediction method for the refolding of inclusion bodies and classify refolding types of different inclusion bodies directly from their primary structure to improve the efficiency of high throughput refolding process.

METHODSForty-three recombinant proteins performing important biological functions were expressed in E. coli. The probability of forming inclusion bodies of these proteins was predicted using Harrison's two parameter prediction model based on the proteins' amino acid composition. Subsequently, the proteins from the inclusion bodies were refolded using a double denaturation method that involved washing and denaturation in GdnHCl solution followed by denaturation in Urea solution and refolding through dilution.

RESULTSAll the proteins were detected in the form of inclusion bodies using SDS-PAGE method. The proteins were divided into two types according to the results of both solubility prediction and refolding experiments. Fourteen proteins were predicted to have the dependency of soluble expression. The refolding yields of these inclusion bodies were up to 70%. Twenty-nine proteins were predicted to have the high dependency of insoluble expression, and their refolding yields could be higher than 70% and lower than 60%. Comparison of the characteristics between the proteins with high and low refolding yields showed that the theoretical pI was significantly different (P<0.05).

CONCLUSIONSHarrison's two parameter prediction model has the value for potential application in classification of the inclusion bodies and prediction of solubility of proteins refolded from different inclusion bodies. This a novel method enhances the efficiency of high throughput refolding of inclusion bodies, and suggests that the theoretical pI of the proteins is an important parameter in the prediction of refolding yields.