High-throughput discovery of energy landscapes in natural and designed nanobodies

• Grant number: 20/14421-1
• Support type: Scholarships abroad - Research Internship - Post-doctor
• Effective date (Start): April 01, 2021
• Effective date (End): March 31, 2022
• Field of knowledge: Biological Sciences - Biochemistry - Chemistry of Macromolecules
• Principal researcher: Munir Salomao Skaf
• Grantee: Allan Jhonathan Ramos Ferrari
• Supervisor abroad: Gabriel Rocklin
• Home Institution: Instituto de Química (IQ). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Research place: Northwestern University, Chicago, United States
• Associated to the scholarship: 19/17007-4 - Protein modeling of enzymes and complexes associated with the degradation of cellulose, BP.PD

Abstract

Evolution has sampled just a tiny fraction of the protein sequence space. De novo protein design aims to explore the uncharted sequence space searching for customized proteins with diverse applications in biology and material sciences. For many years, the field relied in experimentally testing one protein at a time with a very low success rate. Recently, Dr. Rocklin has developed a high-throughput protein stability assay that allows testing and classifying thousands of proteins according to their folding stability. This assay also allowed to interrogate the sequence and structure features determinants of stability and build predictive machine learning models that significantly increased the success rate of the designed proteins. However, in addition to being stable and well folded, proteins exhibit dynamics and exist as an ensemble of states that define their energy landscape with direct impact to their function. These states are commonly hidden from most experiments and have never been studied in high-throughput. In this project, we propose the development and application of a new high-throughput method based on pooled protein expression and hydrogen deuterium exchange mass spectrometry to quantitatively derive the energy landscapes for nanobodies. This class of proteins finds important applications in research, diagnosis, and therapeutics but presents low success rate due to unfavorable biophysical properties such as being prone to aggregation. Here, we aim to build predictive models trained on a large experimental dataset collected on few thousands nanobodies that will enable us to learn how to engineer nanobodies with optimized energy landscapes for diverse applications.