Computational design has immense potential to create new protein functions with applications in biotechnology, biology, and medicine. However, despite exciting progress in designing proteins with de novo structures, our ability to design proteins with new functions lags behind. A key reason for this discrepancy is that function typically requires protein geometries that deviate from the ?idealized? folds of de novo designed structures and that are hence more difficult to design. The long-term objective of our work is to advance computational design to make predictive design of more complex functions possible. The specific objective of this proposal is to address the generally unsolved problem of designing proteins that bind new small molecule ligands. A particular application is the design of new sensor/actuators: proteins that can detect a user-defined small molecule signal and trigger a biological response (such as protein signaling or gene expression). Significant applications of such sensor/actuators include maximizing production of industrially valuable chemicals in metabolic engineering, creating precise tools for dissecting biological processes in cell signaling, and achieving tight regulation in emerging cancer therapies. Our work in the prior project period has advanced methods for binding site design and applied them to engineer the first computationally designed chemically-induced protein dimerization system, which senses and responds to a new ligand in living cells; a crystal structure confirmed the accuracy of the de novo designed binding site. Despite this key progress, there are significant barriers to generalize the approach. The first step in engineering new ligand binding sites is generally to identify desired binding site geometries (constellations of amino acid side chains coordinating the ligand). The second step is then to place those geometries into a suitable protein termed ?scaffold?. This approach is critically limited by available geometries, both for binding sites and scaffolds to accommodate them. To address these problems, we propose two key methodological innovations:
Aim 1 will establish and experimentally test a new computational method to generate millions of possible binding site geometries de novo that can be built into proteins.
Aim 2 will develop and test a new computational approach to build ?de novo fold families? (sets of custom-shaped de novo designed proteins) by systematically varying the geometries of structural elements within a given fold topology, to be used as scaffolds. Feasibility is supported by preliminary results for both aims; we have designed new binding sites (prior period), and have solved structures of 3 de novo designed proteins with the same fold but distinct geometries. The proposed studies innovate in creating both new methods and new molecules that expand designable structures and functions and overcome problems with current approaches limited by available geometries. Ultimately, these studies will lead to advanced computational design methods that we will make freely available, new knowledge on strengths and limitations of these methods to drive further developments, and new tools to control cellular behavior in biological engineering and to probe basic and disease biology.

Public Health Relevance

This research advances computational methods to engineer proteins with new structures and functions. These technologies are used to engineer protein sensors that can detect and respond to molecular signals in living cells. Such new biosensors have many practical applications in biomedical research and biological engineering, and will also help to advance our understanding of fundamental cellular processes by probing health and disease states.

National Institute of Health (NIH)
National Institute of General Medical Sciences (NIGMS)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Lyster, Peter
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of California San Francisco
Schools of Pharmacy
San Francisco
United States
Zip Code
Mavor, David; Barlow, Kyle A; Asarnow, Daniel et al. (2018) Extending chemical perturbations of the ubiquitin fitness landscape in a classroom setting reveals new constraints on sequence tolerance. Biol Open 7:
Barlow, Kyle A; Ó Conchúir, Shane; Thompson, Samuel et al. (2018) Flex ddG: Rosetta Ensemble-Based Estimation of Changes in Protein-Protein Binding Affinity upon Mutation. J Phys Chem B 122:5389-5399
Alford, Rebecca F; Leaver-Fay, Andrew; Jeliazkov, Jeliazko R et al. (2017) The Rosetta All-Atom Energy Function for Macromolecular Modeling and Design. J Chem Theory Comput 13:3031-3048
Mavor, David; Barlow, Kyle; Thompson, Samuel et al. (2016) Determination of ubiquitin fitness landscapes under different chemical stresses in a classroom setting. Elife 5:
Ó Conchúir, Shane; Barlow, Kyle A; Pache, Roland A et al. (2015) A Web Resource for Standardized Benchmark Datasets, Metrics, and Rosetta Protocols for Macromolecular Modeling and Design. PLoS One 10:e0130433
Ollikainen, Noah; de Jong, René M; Kortemme, Tanja (2015) Coupling Protein Side-Chain and Backbone Flexibility Improves the Re-design of Protein-Ligand Specificity. PLoS Comput Biol 11:e1004335