Assembly of a Photoactivatable Cofactor Triad within a Designed Protein

Abstract

Natural proteins hide the essential requirements for their function beneath layers of unnecessary complexity. To test our understanding of structure-function relationships and to develop proteins with novel properties, we engineer functional model proteins (maquettes) from scratch. Previously, we designed and demonstrated a 14.7 kDa 4-α-helix bundle that catalyzes light-activated electron transfer from zinc protoporphyrin IX to heme. However, without the use of a sacrificial electron-donor in solution, the electron and hole quickly recombine. In order to stabilize the charge-separated state for useful chemistry, we designed an extended 22.4 kDa version of this protein that incorporates a third cofactor, a di-metal center, to serve as an electron donor. A tyrosine residue was inserted between the zinc porphyrin and the di-metal site, similar to the redox-active YZ of photosystem II. The design of this protein was guided by analysis of natural cytochromes and carboxylate-bridged diiron and dimanganese proteins as well as previous designed proteins from the Dutton, DeGrado, Lombardi and Haehnel groups. The distances between cofactors were selected by applying the Moser-Dutton ruler to maximize electron transfer efficiency. Each cofactor binding site included multiple helical glycine residues to impart local flexibility to the protein backbone, minimize steric clashing between amino acid side chains and cofactors, and prevent cofactors from creating long-range structural distortions that could result in anticooperative binding. Ultraviolet/visible spectroscopy showed that all three cofactors bind stoichiometrically with dissociation constants less than 1 μM at pH 7.0, suggesting that modularity of these designed maquettes is extremely useful in constructing functional manmade enzymes. Ultimately, we would like to incorporate these functional maquettes in vivo to make them a part of natural redox cycles and to do useful catalysis.