By Karolina Corin and Emily Kopp
COVID-19’s genome shows signs consistent with genetic manipulation, according to a new preprint.
The preprint, released last week, argues that SARS-CoV-2, the virus that causes COVID-19, may have originated as a synthetic virus in the lab. It has not yet been peer reviewed or published in a scientific journal.
Like much of U.S. Right to Know’s work, the preprint examines available evidence in order to get a glimpse into what virology experiments might have been undertaken before the pandemic.
Two competing theories about the origin of COVID-19 have taken center stage. One theory holds that the virus spilled over from live mammals sold at a seafood market, while the other posits that the virus was released from a laboratory famous for collecting and studying coronaviruses.
To determine whether SARS-CoV-2 had characteristics of genetic manipulation, the authors looked at the pattern of markers called restriction sites in the SARS-CoV-2 genome. Restriction sites are locations where DNA is cut during genetic engineering.
These restriction sites occur naturally in viruses, but can also be added at key locations in the lab to enable genetic manipulation.
The authors found that two restriction sites commonly used for genetic engineering — BsaI and BsmBI — were evenly spaced along the genome. The longest fragment between these restriction sites is unusually short.
Both clues are typical for synthetic viruses, according to the authors. Both restriction sites are recognized by type IIS restriction enzymes, which are handy for engineering. The authors estimate that the chances that a natural virus would have a similar pattern of restriction sites for this enzyme family is less than 0.07 percent.
“Our preprint found strong evidence suggesting SARS-CoV-2 was assembled with a known lab protocol, adding weight to the theory SARS-CoV-2 originated as a lab construct,” said Alex Washburne, one of the study co-authors and a computational biologist who has developed models of zoonotic spillover.
Criticism and concern from some virologists
The preprint raised eyebrows among some virologists.
The main concern voiced by experts is that a classic approach used to assemble coronavirus genomes — so called “Golden Gate Assembly” — usually removes restriction sites or “scars” from the final product.
“I’ve even joked … in the future we will easily be able to tell real vs. synthetic DNA just by seeing if there are BsaI and BsmBI sites present, because most of the synthetic methods to make DNA will leave the final product absent of these sites,” said Tom Ellis, a professor of synthetic genome engineering at Imperial College London.
Sylvestre Marillonnet, who pioneered Golden Gate Assembly, told The Economist that the pattern of restriction sites indeed appeared to be typical of engineering, but that he was struck by the fact that the restriction sites remained in the genome.
Stanley Perlman, a University of Iowa coronavirologist with expertise in reverse genetic systems, also thought it was unusual that sites used to generate fragments for in vitro assembly would be left in the assembled genome. As a result, he found the preprint to be “not very credible.”
However, the preprint authors said that they analyzed synthetic coronaviruses assembled before the pandemic, and eight out of ten retained their restriction sites.
Golden Gate Assembly is also not the only way to clone viral genomes. Other methods keep the restriction sites in the final genome.
Indeed, the authors point out methods that keep the restriction sites were popular at the Wuhan Institute of Virology and a partnering lab in the U.S. at the University of North Carolina.
“We showed that the by far most common method of synthesizing coronaviruses in the lab pre-COVID retained the modified restriction sites. This method was the most popular method of synthesis at the WIV and UNC,” said Washburne.
“The main takeaway is that leaving them is beneficial if you are planning on doing a lot of follow-up experiments, for instance testing a library of receptor binding domain mutations,” said Tony VanDongen, a coauthor of the preprint and an associate professor of pharmacology and cancer biology at Duke University. “In that case, it also becomes economically smarter to not have to assemble the entire genome from scratch every time.”
Perlman also conceded that restriction sites can be left in the genome if scientists want to run multiple experiments. But he cautioned against drawing strong conclusions, saying there is still no evidence the virus was under study at the Wuhan Institute of Virology.
Justin Kinney, an associate professor at Cold Spring Harbor Laboratory who provided feedback and criticism on the preprint before it was posted, echoed the study authors.
“For many studies, it makes sense not to put restriction sites in the final DNA product. But for some studies, leaving those sites in the final DNA product makes a lot of sense,” Kinney said. “For example, if researchers want to exchange one segment of the final DNA product for different versions of that segment, having IIS binding sites strategically positioned within the final DNA product makes a lot of sense.”
“Notably, the SARS-CoV-2 genome has two BsaI binding sites well-positioned to allow the efficient swapping of a small part of the genome that we know was of primary interest to researchers: the part containing the furin cleavage site and the receptor binding domain,” he added.
The receptor binding domain, and in particular the furin cleavage, are of interest to virologists because they aid the virus’ entry into human cells. The furin cleavage site, present in SARS-CoV-2 but absent in other SARS-like viruses, has also spurred concerns about possible engineering.
Those concerns were amplified by a leaked grant proposal that demonstrated that the Baric Lab and Wuhan Institute of Virology were interested in inserting furin cleavage sites into coronaviruses.
“The DEFUSE proposal proves that some virologists wanted to test numerous receptor binding domain and furin cleavage site micromutations and combinations,” said Bruttel, a co-author and molecular immunologist at the University of Würzburg. “For that, one needs 2 unique restriction sites flanking the S1 — exactly what we see in [SARS-CoV-2].”
Potential bias and chance
Some critics of the recent preprint say that choosing to analyze BsaI and BsmBI and not other restriction sites was biased, and would lead to skewed results.
However, Washburne said that the authors chose BsaI and BsmBI because their regular pattern stood out right away. Restriction sites ideal for engineering are evenly spaced along the genome to provide genomic fragments of similar lengths, while naturally occuring sites are usually randomly scattered throughout the genome.
Bruttel added that “most other type IIS enzymes just cannot be combined or produce shorter sticky ends.”
Longer ends can make assembly easier.
BsaI and BsmBI are also two of the most commonly used restriction enzymes for the assembling viral genomes.
Two labs that employed BsaI and BsmI to cut coronaviruses include the Wuhan Institute of Virology, which used the restriction sites to swap spike proteins, and the Baric Lab at the University of North Carolina, which used the restriction sites to reconstruct SARS-CoV-2 in 2020.
Critics of the preprint also say that the seemingly unusual restriction site map observed by the authors could have occurred naturally, either by silent mutations or recombination. Silent mutations alter DNA without changing any encoded proteins, while recombination happens when two different genomes swap segments of DNA.
The authors concede that recombination could explain the unusual restriction map, but that it could also make their results more significant and that more research on this is needed.
The authors also argue that silent mutations are an unlikely explanation for SARS-CoV-2’s unusual restriction map.
“We find exceptionally many (and only silent) mutations occur in the tiny part of the genome that make up these restriction sites,” says Bruttel.
Such a high rate of silent mutations in the restriction sites is unlikely to occur naturally, and “indicates that they were manipulated,” he added.
Backlash and ridicule
The preprint attracted ridicule from prominent proponents of the zoonosis theory.
“Pure unadulterated nonsense. … This paper will never be published,” Vincent Racaniello, a Columbia University virologist, said in an email.
Racienello, who has defended the Baric Lab’s work and promoted the zoonotic spillover theory on his podcast “This Week In Virology,” did not reply to a request for a more specific critique.
Scripps Research Institute virologist Kristian Andersen excoriated the preprint as failing “kindergarten molecular biology” in a viral tweet thread.
After the preprint was released, Andersen said that he too had examined SARS-CoV-2 restriction sites in late January and early February of 2020 — around the time he presented his suspicions that the virus was engineered to a group of top virologists and leaders of the NIH.
Andersen shared the results of his analysis on Twitter, which he claimed indicated that BsaI and BsmI restriction sites appear to occur naturally along similar places in the genome of related viruses.
Zoonosis proponents’ worries about the Baric Lab
Concerns about whether in vitro genetic assembly methods on coronaviruses employed by the Baric Lab and the Wuhan Institute of Virology could have factored into the origin of COVID-19 are not new.
As the pandemic first picked up speed in January and February 2020, National Institutes of Health (NIH) leaders and top virologists, including Andersen, fretted over a controversial gain-of-function experiment conducted by the Baric Lab and the Wuhan Institute of Virology which assembled chimeric SARS-like coronavirus genomes using different restriction sites.
The scientists were concerned that the methods described in the paper outlined “a how-to manual for building the Wuhan coronavirus in a laboratory.”