Yeast Mating Redesign Offers New Path for Safer Drug Screening

Compiled by the editorial desk with reference to the original article published in The Conversation, which was authored by Ian Haydon, a doctoral student in biochemistry at the University of Washington.

Scientists at the University of Washington have turned yeast's natural mating process into a high-throughput screening tool that could make drug development safer. By reprogramming how the single-celled fungus recognizes sexual partners, the team has created a method to test thousands of potential drug interactions at once—a task that previously required separate experiments.

The work, led by graduate student David Younger, builds on a technique called yeast surface display, which was pioneered in the lab of Dane Wittrup. That method allows researchers to attach foreign proteins to the outside of yeast cells, making it easier to study how those proteins bind to other molecules. However, the original approach could only test one interaction at a time. Younger's redesign allows for parallel testing of many drug candidates against many potential targets simultaneously.

In yeast, sexual reproduction depends on surface proteins that determine mating type. When two cells of opposite mating types meet, their surface proteins interact and the cells fuse. The University of Washington team replaced the natural mating proteins with custom pairs, linking each to a fluorescent marker—red for one type, blue for the other. When cells mated, they produced purple offspring, allowing the researchers to measure mating efficiency simply by counting colors.

In lab-grown yeast, natural mating efficiency is about 60 percent. When the team swapped in a pair of weakly interacting proteins, efficiency dropped to 5.7 percent. A stronger pair raised it to 19 percent, and a high-affinity pair reached 51.6 percent—nearly matching natural levels. For proteins that should not interact at all, efficiency fell to just 0.2 percent. This correlation between protein affinity and mating efficiency forms the basis of the new screening tool.

To demonstrate its potential in drug development, the team generated 1,400 variants of an experimental anti-cancer drug called XCDP07. By mixing yeast displaying these variants with yeast displaying human proteins, they identified versions that bound only to the intended target, avoiding off-target interactions that could cause harmful side effects.

Off-target binding is a major concern in drug design. For example, Alnylam Pharmaceuticals' drug revusiran, an RNA-based treatment for a rare heart disease, failed a phase III clinical trial last year after 19 patient deaths. The company's stock lost $3 billion. Some experts attribute the failure to off-target interactions, highlighting the need for more comprehensive preclinical screening.

Younger has already shared his engineered yeast strains with researchers at Stanford, Yale, and UC San Diego. He is also founding a company, funded by scientific grants rather than investors, to commercialize the technology. His goal, he says, is to provide comprehensive preclinical drug screening, rather than the current practice of testing only a small subset of possible off-target interactions.

This new approach does not eliminate the need for animal testing or clinical trials, but it could make the early stages of drug development more efficient and safer. By catching dangerous interactions before a drug reaches humans, it may help prevent tragedies like the revusiran trial.