Computers take the lead in protein engineering

Designing protein molecules for the pharma sector often includes a ‘guess-and-check’ approach. However, researchers at Harvard University claim to have found a new approach using computer science.

Researchers at the Wyss Institute for Biologically Inspired Engineering at Harvard have been working on a method to rationally design and select protein molecules to create effective biologic drug therapies while reducing unintended side effects. The method has been published in the Biophysical Journal.

"I believe that biology is the technology of this century," said the study's senior author and Wyss Institute Core Faculty member Pamela Silver, Ph.D., who is also Professor in the Department of Systems Biology at Harvard Medical School. "But in order for that to be true in protein drug therapy, we must make drug discovery and development cheaper, faster and more predictable, with a higher potency for targets while eliminating side effects on healthy cells."

The new model finds that the drug efficacy of fusion–protein therapies depends on the geometric characteristics of a drug's molecular components. The researchers claim that use of the model has the potential replace the need to physically make and test new biologic drug designs.

The engineered fusion proteins are created by attaching a specific antibody to a specific therapeutic protein by a "linker" made of rigid DNA strands. The antibody, a protein itself used as a targeting tool, is selected based on what types of cells the therapeutic portion is intended to treat. As the antibody finds its target, such as a receptor on a cancerous or otherwise infected cell, the therapeutic protein simultaneously attaches to another receptor and triggers mechanisms to disrupt the cell's behavior. The efficacy of these new types of drugs depends on how well the two components of the fusion protein work together — that is, how well they each attach to their intended receptors at the same time.

The computational model reveals that the length of the DNA linker used to connect the parts of the fusion protein influences how successfully both components are able to reach their independent receptors, empowering researchers to rationally design linker lengths that will translate into the most effective results.

"We are treating protein engineering as we would nanotechnology," said Wyss Institute Senior Staff Scientist Jeffrey Way, who contributed key research to the study. "We do so by looking at biological molecules as if they are nano–machines, with programmable reactions and mechanisms."