Proteins are incredibly adaptive molecules. For example, when exposed to a stimulus, such as an environmental change in temperature, pH or light, some proteins undergo conformational (shape) changes that can lead to useful material properties, such as a phase transition, turning from a solid to liquid, or a liquid to solid. Understanding and predicting how these changes occur on the molecular level could lead to the creation of an entire class of soft materials that can respond to environmental cues. This research will develop and apply computer algorithms to quickly go through large amounts of protein sequence data to predict as yet undiscovered temperature-sensitive peptide sequences. These sequences can then be subjected to computer-based modeling to predict conformational changes that ensue from a phase transition. Finally, experiments will be conducted to predict and determine the 2D and 3D materials architectures that can be created by combining stimulus-responsive peptide sequences. If successful, these methods could create a toolkit for the efficient design and fabrication of a large variety of materials with custom-designed properties.
Stimulus responsiveness is a striking feature of proteins in Nature, whereby responses to chemical stimuli such as ligand binding, phosphorylation, and methylation, and physical stimuli such as changes in temperature, pH, light, and salt concentration lead to sharp conformational or phase transitions. Unlike proteins, which encode diverse responses to numerous stimuli by richly sampling amino acid sequence space, current bioinspired designs of repetitive polypeptides have focused on a tiny fraction of the vast conceivable expanse of sequence space. The primary goal of the proposed research is thus to develop generalized materials design rules, by combining experiments, fast and accurate physics-based computer simulations, and data science, to accelerate the discovery and development of a potentially huge class of thermally-responsive polypeptide materials by a systematic exploration of sequence space. This research will "for the first time" provide a complete atomistic understanding of the determinants of the lower critical solution temperature (LCST) and upper critical solution temperature (UCST) phase behavior, enable de novo molecular design of LCST and UCST peptide polymers and identify rules on how to combine them to create hierarchically-ordered, nanostructured polypeptide materials that exhibit unique morphologies that can be tuned as a function of their stimulus responsiveness. These materials could serve as nanostructured scaffolds and templates and enable a broad range of biocatalytic, bioelectronic, or assay devices. The PIs also plan to release the PIMMS modeling package, a set of tools for performing lattice-based simulations of polymers. PIMMS will provide support for the machine learning algorithms that enable the design of responsive protein-based polymers. The PIMMS codebase will be released as open source. A user community will be coalesced around the language by ensuring that interested researchers are able to contribute modules to or implement application-specific algorithms within the codebase. This is expected to allow a wider growth of the project. This aspect is of special interest to the software cluster in the Office of Advanced Cyberinfrastructure, which has provided co-funding for this award.