Towards Mapping Domain Boundaries of Proteins

Abstract

GFP fluorescence from bacterial colonies expressing a polypeptide upstream of GFP depends on the solubility of the polypeptide. Inspired by this idea, we developed a method to test for the folding of any polypeptide longer than about 66 amino acids. This method bypasses the need for purification of the polypeptide to test for folding. We employed this technique to identify the independently folded domains of Mid1, a fission yeast protein of 920 amino acids that participate in cytokinesis. A random fragment library of Mid1 cDNA was generated using tagged random primer PCR (tPCR). The fragment library was cloned upstream of a monomeric enhanced GFP (mEGFP) and expressed in E.coli. The GFP fluorescence of the bacterial colonies expressing the fusion constructs were compared to a threshold level of fluorescence to test if the Mid1 fragment fused to mEGFP is folded. Colonies expressing folded fragments of Mid1 are brighter than control colonies or colonies expressing an unfolded protein fragment fused upstream of mEGFP. Approximately 27000 colonies were screened to identify five soluble folded domains and one insoluble region, accounting for the entire length of Mid1 protein. This large scale approach will be useful in rapidly mapping domain boundaries of proteins in absence of any prior knowledge about the domain organization of the proteins.