Key elements for protein foldability revealed by a combinatorial approach among similarly folded but distantly...

We have determined the key regions for protein foldability by creating multiple crossover libraries from two proteins that share similar fold but have low sequence identity and differ significantly in stability. One protein is the propeptide of a serine protease, subtilisin BPN', and the other is Pleurotus ostreatus proteinase A inhibitor 1 (POIA1). The propeptide has a compact structure when complexed with subtilisin but is unstructured when isolated, whereas POIA1 takes a stable structure. We selected four of the conserved amino acid residues for the boundaries of crossover sites and utilized these residues to make same cohesive-ends to assemble synthetic DNA fragments. Each segment has one or two secondary structure units, and the interchange of these structural elements produces 32 (= 2(5)) combinations, including the propeptide and POIA1. The stability of these mutants was first screened by formation of turbid zones on skim milk plates containing subtilisin BPN'. It was shown that six variants were foldable and structural units necessary for folding were identified. Further fragmentation and recombination of these mutants (the "multisection" method) revealed that two interactions between secondary structures are important; one is interaction between the loop-alpha1 and beta2-turn-beta3, and the other is hydrophobic interaction between the adjoining beta1 and beta4 strands. We were also able to specify the significant amino acid combinations for tolerance to proteolysis. These combinatorial methods not only elucidate how domains can be interchanged to make the whole protein foldable but also extract essential regions for the function, which is correlated with the instability of the molecule.