The structure of a protein on cell surfaces may explain why certain animals such as mice and chickens have not contracted COVID-19, presenting clues for how scientists might develop a treatment for the disease and predict the next potential epidemic.

Research published on Thursday in PLOS Computational Biology examined the molecular structure of ACE2 proteins in humans and 25 other species using 3D models to determine how effectively each binds with the coronavirus. A handful of key amino acids in the proteins determined whether the virus successfully latches on, according to the study, which included researchers from the U.S., Canada and Germany.

“Our paper essentially gives a structural framework to explain why certain species are susceptible to the virus and why some are not,” said Joāo Rodrigues, a postdoctoral research fellow in structural biology at Stanford University and the paper’s lead author. 

Coronavirus infections of cells begin when a protruding “spike” of the virus binds with the ACE2 protein, which helps regulate blood pressure and is found on the membrane of most human cells. This initial interaction has been a research focus of many scientists trying to understand and combat the deadly virus.

The ACE2 proteins of non-susceptible species lack nearly all polar and charged regions that are crucial to binding with a coronavirus spike, according to the 3D models, although they sport different mutations that create the incompatibility.

To test the influence of these specific regions of ACE2, the researchers transposed coronavirus-friendly mutations to ACE2 models of non-susceptible species. They found that 16 of the 17 single-point mutations strengthened the protein-coronavirus interaction to some degree.

The authors also generated values that measured how well the coronavirus could bind to each ACE2 model, with a low score representing strong connection and stronger likelihood of infection. The results largely align with known COVID-19 susceptibilities: Infection-prone pangolins, sheep and tigers all scored low, while immune mice, ducks and rats had high scores.

But some results didn’t match with prior research findings, such as the guinea pig ACE2 protein earning a low score despite the rodent’s imperviousness to COVID-19. Rodrigues said his team’s research is insufficient to predict the COVID-19 susceptibility of different species, in part because the team only modeled the first step of the infection process.

Findings could, however, inform the development of ACE2-like molecules that easily bind to coronavirus spikes. In theory, they could be used as a COVID-19 treatment to slow infections by occupying the spikes on a virus before it binds with living cells. Researchers at the University of Illinois at Urbana-Champaign have already engineered proteins that bind to the spikes more effectively than the human ACE2 protein does.

The team is working with other researchers to develop better predictors of animal susceptibility to COVID-19, which could potentially identify endangered species that would be harmed by coronavirus infections, Rodrigues said.

A prediction tool could also be used to predict which species are most likely brewing the next highly contagious coronavirus following SARS, MERS and COVID-19, the researcher said.

“There are probably going to be others sooner or later, so it would be great if we could predict, which species are most likely reservoirs of these types of viruses?” Rodrigues said. “If humans keep interacting with these species, chances are, sooner or later, a fourth coronavirus is going to jump.”

The Stanford research fellow said the work benefited from the scientific community’s widespread practice of releasing coronavirus research on preprint websites such as BioRxiv, where papers have not been peer-reviewed but become publicly available much earlier than if they were held until journal publication. 

Rodrigues — who got the idea for the study after hearing about Bronx Zoo tigers contracting COVID-19 in April — said this cooperative attitude among researchers extended to his own team, which grew as he encountered other scientists investigating the ACE2 protein. They shared their preprint paper on BioRxiv in July, more than four months before it was published in a peer-reviewed publication.

“People realized that this was a larger issue than my career, your career, my paper, your paper — so they all deposited these papers on these preprint servers,” Rodrigues said. “In doing so, they essentially allowed other researchers to have a look at their work and start building from their work much earlier than if these things have been peer-reviewed.”

The article, “Insights on cross-species transmission of SARS-CoV-2 from structural modeling,” was published Dec. 3 inPLOS Computational Biology. The authors of the study were Joāo Rodrigues, Elizabeth Seckel and Michael Levitt, Stanford University; Susana Barrera-Vilarmau, Institute of Advanced Chemistry of Catalonia; Joāo Teixeira, The Hospital for Sick Children in Toronto; and Marija Sorokina and Panagiotis Kastritis, Martin Luther University Halle-Wittenberg. The lead author was Joāo Rodrigues.