Computational Design of β Sheet Proteins Public Deposited
Downloadable ContentDownload PDF
- Last Modified
- March 20, 2019
- Affiliation: School of Medicine, Department of Biochemistry and Biophysics
- Computational protein design has become a very powerful approach to test our understanding of the forces and energetics of macromolecular systems. The ability to design proteins that have specific structures and functions will be very valuable to future protein drug discovery. Protein design technology has been successfully applied to stabilize proteins, increase protein-protein binding affinity and create new protein structures. However, de novo design remains very challenging, especially for β-sheet proteins. Most de novo designed β-sheet proteins tested to date either misfold or aggregate. In this thesis, we use a hierarchical approach to search for the bottleneck in β-sheet design. First, we tested our ability to redesign the sequence of a naturally occurring β-sheet protein. The molecular modeling program Rosetta was used to design new sequences for the β-sheet protein tenascin. The redesigned proteins are well-folded and have thermal melting temperatures that are 40 °C higher than the wild type. These results indicate that given a designable backbone we can create a well-folded β-sheet protein. To move towards complete de novo design we next asked if we could design a portion of a β-sheet protein from scratch. We tested our ability to design loops by removing a ten-residue loop from tenascin and rebuilding it to have a new but specific conformation. These studies involved the simultaneous search of conformational and sequence space. Two of the designed loops were crystallized, and one of them adopts a structure that is very similar to the design model. Lastly, we have explored designing whole β-sheet proteins from scratch. Four generations of designs have been tested to date, and unfortunately, none of the designs appear to be well folded. To lay the groundwork for future success, we have been comparing the design models to naturally occurring β- sheet proteins to identify structural features that may be missing from the designs. We find that naturally occurring proteins include fewer voids accessible to small probes (~ 0.7 Å ) than our design models. It remains to be seen if more conformational sampling is need to remove these voids or if the energy function requires changes.
- Date of publication
- December 2008
- Resource type
- Rights statement
- In Copyright
- Kuhlman, Brian
- Open access
This work has no parents.
|Computational design of [beta] sheet proteins||2019-04-09||Public||