Gaining an advantage in the cat-and-mouse game with the COVID-19 virus

In the study, Bloom's team tested the impact of the mutations on functions important for the virus' evolution, such as the ability of the spike protein to grab and bind to a region on the surface of human cells called ACE2.

They also tested for mutations that may affect a virus's ability to evade antibodies generated by vaccines or past exposure to the virus, which often try to disrupt ACE2 binding to prevent infection.

To create the pseudovirus, they stripped a live virus commonly used in research and gene therapy down to its skeleton, removing its ability to replicate, while adding some features that make it more efficient for use in deep mutation scans.

The team then created thousands of slightly different pseudoviruses, each genetically engineered to express a particular variant of the spike protein, allowing them to measure the effects of 7,000 mutations in a single experiment.

To track the mutations throughout the process, the researchers tagged each one with a unique genetic barcode — think of it like the barcodes libraries use to track books as they are checked out, returned, and rearranged.

They then cultured a stock of millions of pseudovirus particles, each with a unique barcode, to create a viral library.

The researchers dropped the variant into small dishes containing human cells and antibodies from people with different immune status and infection histories.

Rather than exposing cells to one mutation at a time, they exposed cells to the entire library of 7,000 barcoded mutations in one experiment.

The next day, the team collected the infected cells from the dishes and analysed their genome sequences to see which pseudovirus particles had evaded the antibodies and infected the cells.

The team then analysed that massive dataset, using computer analysis to search for barcodes to see which of the 7,000 mutations made a difference.

Finding important mutations

The team was able to directly measure which mutations affected antibody evasion and made a surprising discovery.

Many antibodies focus on preventing the capture part of the virus (the receptor binding domain, or RBD) from being captured by the ACE2 region on the surface of human cells.

As expected, many relevant mutations occur within or close to the RBD to evade these antibodies.

But the team also found that there are important mutations outside the RBD. The strongest of these distant mutations appear to move the RBD to an up or down position.

The RBD needs to be in the upward position to grab and bind ACE2, but it is also vulnerable to attack in the upward position. Moving down the RBD helps the virus escape neutralization by antibodies, but it also means it can't bind to ACE2 and infect cells.

“A mutation that's very good for receptor binding can be very bad for neutralization, and vice versa,” said lead author Bernadette Dadonaite, PhD, a staff scientist in the Bloom lab. “Evolution will work out the precise balance of how much up-body posture you need and how much down-body posture you need.”

Somehow, this balance of mutations causes the RBD to act like a submarine's periscope, rising during cellular attacks and falling during stealthy escapes.

Research suggests that this escape mechanism occurs frequently and should be monitored.

“It's always a game of cat and mouse,” Dadonite said. “We can try different things, but eventually, through evolutionary forces, the virus will find a way to evade whatever we do.”

RBD periscope-like mutations that help pseudoviruses evade antibodies are also found in the most successful variants of SARS-CoV-2, partially explaining why these variants have thrived despite improving immunity and multiple vaccinations.

Bloom Lab sees this as a positive sign that its scanning methods are mature enough to produce useful predictions about what a real virus will do next, allowing it to get ahead of it.

This could provide a watchlist of potentially dangerous mutations for other researchers, such as computational biologists tracking the spread of SARS-CoV-2 and other viruses. Dr. Trevor Bedford in Fred Hutch's Vaccines and Infectious Diseases Department and is a co-author of the study.

“Predicting how a virus will evolve is a complex problem, so I don't want to overstate it,” Bloom says. “The key is to have something that helps us interpret what's happening as it's happening.”

This research was supported by grants from the National Institutes of Health, a Pew Biomedical Scholars Award, a Burroughs Wellcome Fund Infectious Disease Pathogenesis Investigator Award, the University of Washington, the Arnold and Mabel Beckman Center for Cryo-Electron Microscopy, the Fred Hutch/University of Washington/Seattle Children's Cancer Consortium Shared Resources, SciLife Labs' Pandemic Laboratory Readiness Program, and the Erling Persson Foundation.