Proteins are made up of linear sequences of amino acids but understanding how these amino acids fold to form a three-dimensional structure is notoriously difficult. Knowing what a protein looks like in 3D is often necessary for understanding how it functions and how it can be manipulated. For instance, understanding how proteins such as antibodies bind to viruses like the flu would enable scientists to develop antibodies that bind tighter and work better. 

Two independent teams – one led by Debora Marks at Harvard Medical School and the other led by Ben Lehner at the Barcelona Institute of Science and Technology, are hoping to expand the speed with which protein structures can be solved. Their strategy involves systematically mutating every amino acid in a protein to every other amino acid (there are 22!) to create thousands of single mutant proteins. They then do the same for each pair of amino acids, creating double mutant proteins. After manufacturing each one of these single mutant and double mutant proteins in bacteria, they screen for the protein’s function, such as how tightly it binds its target. Marks and Lehner realized that occasionally they would get mutants in which the functional effect of a double mutant protein would be much more severe than what would be expected from each single mutant alone – a phenomenon known as epistasis. 

When a sequence of amino acids bends to form the protein’s final shape, amino acids that are far apart can interact with each other. If this interaction is critical to the protein’s function, then messing with both of these amino acids will heavily impact the protein’s functional performance (thereby, epistasis). By feeding information about these epistatic interactions into a modeling algorithm, scientists are able to predict three-dimensional structure with accuracies similar to traditional methods.

This new combinatoric approach might open up the possibilities of decoding more protein structures and of doing so in many more labs that do not have the high-tech equipment usually required for the traditional processes. Going through all possible sequence combinations remains laborious, which has led some laboratories to try and crowdsource these efforts. Though this method may not be the Rosetta Stone to decoding every protein’s structure, it certainly holds promise.

Popular news article:

Original scientific article:

Managing Correspondent: Radhika Agarwal

Image Credit: “Myoglobin Protein Structure PDB 1MBN” by Eduardo Beltrame is licensed under CC CC0 1.0

Leave a Reply

Your email address will not be published. Required fields are marked *