Design of electron conductive protein-protein interfaces and redox-active multiprotein assemblies
Inorganic Molecular Chemistry - Synthesis and Characterisation
Biochemistry
Biological and Biomimetic Chemistry
Biomaterials
Final Report Abstract
The goal of this research project was the design of protein-protein-interfaces, over which electrons can be conducted. This goal was approached through several different avenues. A first approach to design a dimer of the four-helix bundle protein cytochrome cb562 by computational interface design with an in vitro supplied heme at the interface, did not result in ligation of the external heme nor in dimer formation without heme. A posteriori, several problematic points can be identified: Binding of heme to alpha-helices with histidine ligation brings the tetra-pyrrole ring in close proximity to the helix backbone. To avoid clashes, amino acids with no or very small sidechains, alanine or glycine, are necessary, preventing formation of more intricate heme binding motifs. Heme with only two external axial metal-ligands is not allowing for a particularly specific or strong binding to the protein while at the same time with its large flat ring-system covers a good part of the possible binding interface between the two proteins chosen for this project, which weakens pre-formation of a protein dimer before heme binding. In a different approach, [4Fe4S]-clusters were chosen as the interfacial redox cofactor with stronger coordination and smaller interface footprint. A rather flexible [4Fe4S]-cluster binding motif from ferredoxins was selected to supply three cysteines on one monomer, while the fourth cysteine would coordinate the cluster from a second monomer in a C2 symmetric arrangement, resulting in two thiol-iron interactions spanning a protein-protein interface, with a significant interface area to be designed by PyRosetta. Self-assembly of [4Fe4S]-clusters in vitro from FeCl3 and NaHS seemed successful as measured by UV/Vis spectroscopy, but could not be confirmed by massspectrometry, electro-chemistry, X-ray crystallography or electron paramagnetic resonance spectroscopy. These measurements were partly impaired by sample heterogeneity, which often results from in vitro assembly of [4Fe4S]-clusters. Efforts are underway to use a cleaner in vivo cluster assembly approach. Several recent publications in the field using [4Fe4S]-clusters in protein design confirm these bioinorganic clusters as the most likely candidate to create a more general design principle to conductively couple two redox proteins. The second attempt to use interfacial [4Fe4S]-clusters using a different scaffold protein, Azurin, in the hope to engineer a stable dimer first by metal mediated design, fell short of this first intermediate goal. In contrast to the many possible relative orientations of two binding alpha-helices, the specific conformational selectivity imposed especially by the polar atoms on the sides of beta-sheets, prohibited formation of the envisioned metal-mediated dimer despite best efforts. Protein-protein-interfaces encompassing the sides of beta-strands are still not easy targets for protein design and therefore not ideal as starting points for more complex projects. Finally, conductivity over longer lengths scales through a protein assembly should be constructed by first generating a one-dimensional protein assembly through simpler protein-protein interactions and later filling a connected inner cavity with conductive material. The significantly more limited interface design component, making use of simple disulfide bonds, proved to be successful in designing protein nanotubes forming an internal cavity with one of the highest reported aspect ratios of sub-nanometer widths and µm length. The focus is now on growing metallic nanoparticles within this cavity to reach conductive protein assemblies. Overall, despite many highly visible publications and impressive progress in reaching new milestones for the first time, what is reliably achievable in protein design still seems a long way from what is sometimes possible.
Publications
- (2020). Constructing protein polyhedra via orthogonal chemical interactions. Nature 578, 172–176
Golub, E., Subramanian, R.H., Esselborn, J., Alberstein, R.G., Bailey, J.B., Chiong, J.A., Yan, X., Booth, T., Baker, T.S., and Tezcan, F.A.
(See online at https://doi.org/10.1038/s41586-019-1928-2)
