A new NSF grant awarded to a team of engineers from Duke University and Washington University in St. Louis will allow them to further explore new self-assembling materials, which can be designed using machine-learning approaches. The design of these biological polymers can then be refined using numerical simulations and tested. (Credit: Pappu Lab)

Engineering new materials holds enormous potential to improve and advance the global community. Breakthroughs in medicine, defense and clean energy could be achieved by designing polymeric materials with a whole host of abilities and properties.

To push this emerging field forward, the National Science Foundation (NSF) set up an initiative called Designing Materials to Revolutionize and Engineer our Future (DMREF).  In August, DMREF awarded a four-year, $1.4 million grant to a team consisting of researchers from the engineering schools of  Duke University and Washington University in St. Louis. The initiative awards grants to researchers at the forefront of materials advancement, enabling them to push science, stretch their imaginations in the quest to streamline the development of new soft materials, and predict and tune their properties for both existing and novel applications. 

Researchers Ashutosh Chilkoti, the Alan L. Kaganov Professor of Biomedical Engineering and chair of the Department of Biomedical Engineering and the principal investigator for the project, and Stefan Zauscher, the Sternberg Family Professor of Mechanical Engineering & Materials Science, comprise the Duke team, and will collaborate with Rohit Pappu, the Edwin H. Murty Professor of Engineering at Washington University’s School of Engineering & Applied Science.

“You can imagine making an adhesive that will also have the strength of steel,” said Pappu. “Or perhaps something that will flow like toothpaste but also have the potential to be used as a miniature bioreactor. We could use new materials for drug delivery, drug storage, artificial tissues, and other applications we haven’t thought of yet.”

“Nature offers an enormous design space for the design of new materials, but most past efforts, including ours, have simply used a very narrow range of parameters to create bioinspired — protein-based — materials,” Chilkoti said. “This project will attempt to cover a large terrain of the sequence space of nature by combining fast computer screening techniques developed at Washington University by Professor Pappu with high-throughput synthesis and characterization at Duke University to identify peptide polymers with new or improved function.”

Zauscher and Chilkoti are already working to develop approaches to assemble these polypeptides into hierarchically structured materials.

“These materials can function as templates to guide the self-assembly of nanoscale objects and thus enable a broad range of biocatalytic, bioelectronic/photonic, or assay devices,” Zauscher said.

The next step: Scientists must be able to zero in on the exact amino acid sequences within the enormous sequence space of the 20 amino acids offered by nature that need to be manipulated to create materials with unprecedented physical properties. It’s a process combining mathematical and physical science, and, with an infinite number of sequences to explore, takes the concept of “needle in a haystack” to an extreme degree. That’s where Pappu and his team come in.

“Even with all the high-throughput technologies, the design space is huge,” Pappu said. “What we need are principles that will pave the way to rational design strategies, whereby we prescribe desired material properties and arrive at a set of sequences via computational design that are likely to achieve the properties of interest. This requires an understanding of the connection between a sequence and the design criteria, although this understanding doesn’t have to be perfect. It needs to be good enough to guide the search process through an astronomically large sequence space.”

Physics-based computer simulations of phase transitions of protein sequences combined with mathematical models developed in the Pappu lab will aid the design process. Pappu and his team plan to combine knowledge of the relationships between sequence-encoded physical properties of various block copolymeric amino acid sequences and measured material characteristics, along with prior data culled by the Duke researchers. Then they intend to design algorithms that should enable the customized design of materials with desired profiles such as response to external stimuli, specific flow and mechanical properties, and the ability to serve as scaffolds for biochemical reactions. 

“Through our research collaboration, we will converge on a generalized simulation and experimental toolkit that enables the efficient design and fabrication of protein material nanostructures using thermally responsive polypeptides,” Zauscher said.

The advances that Pappu and his team hope to realize in collaboration with Chilkoti and Zauscher, who are leaders in materials science and engineering, should enable the expansion of materials design that goes beyond the current scaffold, which is based on elastin-like polypeptides.

“The algorithms we develop are a blend of physics-based approaches that span multiple scales and data-driven approaches that enable the joint navigation of sequence space, conformational space, and the space of material morphologies in order to achieve the specified design criteria,” Pappu said. “The opportunity to work with experts like Chilkoti and Zauscher is very exciting and we hope to achieve some success in fulfilling the NSF’s goal of expanding the design space of protein-based polymeric materials.