A tiny change—just one mutation—appears to have boosted the modern Zika virus’s ability to attack fetal brain cells, fueling the wave of birth defects involving microcephaly (small head size) that recently swept across the Americas. The findings are reported Thursday in Science.

Researchers in China found that a single swap of amino acids—from serine to asparagine—on a structural protein of the Zika virus occurred a few months before the pathogen first took off in French Polynesia in 2013.

The team’s results may begin to answer an outstanding question from the Zika epidemic: Why have Zika-related microcephaly and other brain abnormalities been seen in areas hard-hit by outbreaks in the past few years but not in the decades following the virus’s discovery in 1947? One theory is that the Zika–microcephaly connection previously flew under the radar because there were too few cases to see the link. Another leading theory is that something about the modern virus has changed, allowing it to infect brain cells more efficiently than its ancestors could. The new work suggests the latter is true. “This is a very good study and it gives a plausible explanation that is scientifically based,” says Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases at the U.S. National Institutes of Health. He adds that the results will be further strengthened if other groups replicate them.

Not all changes in a pathogen are significant. Viruses continually mutate as they replicate, which has made identifying functionally important changes difficult. But the Chinese research team catalogued the differences between modern Zika and an ancestral strain isolated from a patient in Cambodia in 2010, and then employed computer modeling software that suggested a singular amino acid mutation—called an S139N substitution—was likely important.

The researchers tested that idea by infecting newborn mice (which developmentally resemble a human fetus) with different lab-made versions of Zika. They found virus with the S139N mutation caused the most damage to the animals’ brain cells. Next the scientists confirmed their findings using reverse genetics—swapping a single replacement mutation for S139N into an otherwise identical Zika virus. The team infected newborn mice with either of the two versions of Zika, and found the S139N-free version was less harmful to the animals. They also replicated some of their testing in human neural stem cells in the lab, and noted the modern Zika virus killed more cells than an ancestral strain.

Exactly how the S139N mutation strengthens Zika’s ability to infect brain cells remains unknown. Because the mutation is within a protein that helps form the virus’s structure, it may have something to do with binding—perhaps allowing the virus to bind to cells with greater affinity, Fauci says.

Even with this new finding, the Chinese researchers still concede their study is likely not the final word on what causes Zika-linked microcephaly. They discovered the S139N swap caused the most severe clinical outcomes but they also note other modern Zika virus strains without this mutation (some occurring in the wild and others that were lab-made) could also cause mild microcephaly and other cellular harm to mice. “Besides host factors [such as low immunity to the virus in affected communities], there are definitely some other unknown viral proteins or amino acids that may contribute to the complex pathogenesis of microcephaly—independently or synergistically,” says lead author Chen-Feng Qin, chair of the Department of Virology at the Beijing Institute of Microbiology and Epidemiology.

“Our study identified a unique genetic determinant that links to severe microcephaly,” Qin adds. The work may have other implications for Zika control, too. Qin says subsequent experiments for testing vaccine or antiviral drug efficacy should use contemporary strains with the S139N mutation. And Qin also cautions: Any future Zika vaccine that includes a live but crippled form of the virus—like one currently in development at the NIH—should not contain this harmful mutation, even though the virus is attenuated (altered to become less virulent).* The NIH says its candidate vaccine—which contains the S139N mutation—did not damage brain tissue in earlier monkey tests. NIH senior associate scientist Stephen Whitehead, who led those experiments, says the new findings (that involved mice with the virus administered into their brains) may not reflect how primates would respond to viruses injected through their skin.

Still, removing potentially problematic mutations is something Fauci says may be worth considering. “If you have something with a neurotrophic (nervous system–related) mutation and are making a live-attenuated vaccine,” he explains, “it would make sense to delete that and not have a mutation there.”

*Editor's Note (9/28/17): This sentence was edited after posting. The original erroneously stated this live-virus vaccine was being tested by the NIH in phase I trials in South America.