Towards rational computational design of self-assembling peptides
Abstract
This proposal will result in the development of enhanced computer simulation methods to explore how flexible molecules interact and identify the most stable thermodynamic states. These methods will be applied to the rational design of biomaterials. Biomaterials are macromolecules that can interact with human cells in order to induce certain behavior and has powerful implications for tissue regeneration, self-healing and improved prosthetics. Bioengineering has led the field of biomaterial design, discovering several successful molecular strategies. The molecules are first synthesized and characterized for their ability to create macrocellular assemblies with the right mechanical properties where cells can survive. At a later stage, they are characterized for their biocompatibility through in vivo cell assays. Computational approaches will accelerate the design principles by reducing the cost and time to filter plausible molecular strategies. In this proposal we focus on peptides that self-assemble into fibers upon sensing an environmental change (such as temperature or pH). The organization into fibers upon self-assembly allows these peptides to interact with proteins on the surface of the cell named integrins. These peptide-integrin interactions trigger cascades of events inside the cell leading to changes in cell behavior. By incorporating specific sequences of amino acids in the peptides, we create functional motifs that induce the cell to migrate, adhere, proliferate, differentiate or die. For example, to close an open wound we would design biomaterials that induce the cell to move to the region where the wound is and then adhere to the peptides to start closing the wound. The challenge is how to determine which peptides will self-assemble, what structures they generate and whether the functional motifs will be accessible to the cell. Molecular modeling is challenged by the large flexibility in these systems, allowing them to explore many possible arrangements. The methods we will derive use powerful atomic simulation methods, running on supercomputers powered by GPUs to explore those arrangements. We use statistical mechanics and Bayesian inference tools to identify those arrangements that will be more thermodynamically stable. Through this process, we can narrow from millions of possible peptide sequences to those that are good candidates to test experimentally. A second problem these methods will tackle is increasing the repertoire of known functional motifs. These functional motifs expand between three and eight amino acids -- leading to over 25 billion possible binding motifs. Currently the number of known functional motifs is close to 50 -- a very small fraction of the possible binding domains. Integrins recognize not only the amino acid sequence of the functional motifs, but the conformation they adopt -- which is dictated by the self-assembling process. The lack of atomistic models has limited the discovery of new motifs. Methods based on Bayesian inference can accelerate the binding process of flexible molecules and provide atomistic detail of the geometries that lead to binding. However, these methods are too computationally demanding to assess over 25 billion sequences. The known geometries from Bayesian simulations from the current known motifs can be used to guide docking simulations allowing the filtering of possible sequences. Finally, relative binding affinities will be calculated on the filtered sequences to identify the best sequence motifs. On the short term, this proposal will enable the development of computational simulation methods grounded on physics to tackle molecular binding of flexible molecules. Success in this endeavor will open the field of biomaterial design to computational design Ð parallel to the first stages in the drug discovery process that are now routinely done via computational approaches.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 02, 2022
- Source ID
- W911NF2210142
Entities
People
- Alberto Perez
Organizations
- Army Contracting Command
- United States Army
- University of Florida