AI finds new target for monkeypox ^{CEFR B2}

The 2022 mpox outbreak prompted scientists to search for targeted and more practical interventions. Vaccines for smallpox were reused during the outbreak, but those vaccines are complex and costly to produce. Researchers in Italy first isolated twelve antibodies from the blood of people who had been infected with MPXV or vaccinated, yet the exact viral binding sites were unknown.

A team at The University of Texas at Austin then applied the AlphaFold 3 AI model to predict which of roughly 35 proteins on the virus surface the antibodies bind to. AlphaFold 3 predicted with high confidence that several antibodies target a surface protein called OPG153, and follow-up laboratory experiments verified that prediction.

When mice were injected with an engineered form of OPG153, they produced antibodies that neutralized MPXV. UT researcher Jason McLellan, a co-lead author, called the work unexpected because OPG153 had not been considered before as a vaccine or drug target. The team describes the method as reverse vaccinology: they began with survivor antibodies, identified the viral target, and engineered an antigen to elicit a similar response. The discovery could support a vaccine antigen or antibody therapies that are cheaper and easier to make than whole-virus vaccines. UT Austin and an Italian foundation have filed patent applications, and the work received partial funding from the Welch Foundation.