Latest / U.S. Right to Know articles / NIH files reveal broader coronavirus engineering research before COVID-19

Transcript

NIH files reveal broader coronavirus engineering research before COVID-19 · U.S. Right to Know articles

Years before the first known Covid-19 cases in Wuhan, a loose network of scientists — backed by U.S. government grants and linked through recurring collaborations — was already running experiments to answer a deceptively simple question: what would it take for newly-discovered bat coronaviruses to infect humans?The question surfaced repeatedly in grant applications, emails and internal NIH reviews. Researchers proposed modifying viral spike proteins — the part that latches onto host cells  — to test whether bat viruses could bind more tightly to human receptors. They also explored changes to the viruses' cleavage sites, molecular switches that allow them to unlock and enter cells more efficiently.Those kinds of viral modifications would soon become entangled in the pandemic’s most contentious question: Did SARS-CoV-2 emerge through natural spillover from animals to humans, or through a laboratory incident tied to research intended to anticipate the next outbreak? A flashpoint in that debate has been <a href="https://drasticresearch.org/wp-content/uploads/2021/09/main-document-preempt-volume-1-no-ess-hr00118s0017-ecohealth-alliance.pdf">DEFUSE</a> — a 2018 grant proposal submitted to DARPA, the Defense Department’s advanced research agency.Led by the U.S. nonprofit EcoHealth Alliance alongside collaborators including University of North Carolina virologist Ralph Baric, a pioneer in coronavirus reverse genetics, and Zhengli Shi of the Wuhan Institute of Virology, who led China’s largest bat coronavirus sampling program, DEFUSE outlined plans to test spike-protein swaps and cleavage-site insertions in bat coronaviruses. The project was never funded.After details of the proposal surfaced in the wake of the pandemic, some proponents of the lab-leak hypothesis argue that DEFUSE reads like a blueprint for SARS-CoV-2. They point to the virus’s distinctive furin cleavage site — a spike feature absent from its closest known relatives that can boost human infectivity and transmissibility — and to DEFUSE’s aim to insert similar human-specific cleavage sites into bat coronavirus spikes. Critics counter that the proposal is irrelevant because DARPA rejected it.But a more fundamental question is whether DEFUSE was an isolated idea or part of a broader, ongoing line of research already underway.Newly obtained NIH records suggest that the experimental concepts later spotlighted in <a href="https://usrtk.org/covid-19-origins/american-scientists-misled-pentagon-on-wuhan-research/">DEFUSE</a> — tuning bat coronavirus infectivity through spike swaps, receptor-binding changes, and cleavage-site insertions— were already embedded in multiple U.S.-funded coronavirus research projects years before the pandemic.Late last month, members of the World Health Organization’s Scientific Advisory Group for the Origins of Novel Pathogens (SAGO) addressed DEFUSE directly in a <a href="https://www.nature.com/articles/d41586-026-00530-y">Nature comment</a>.“Even if the DEFUSE grant application had been approved,” SAGO wrote, “it is scientifically implausible for SARS-CoV-2 to have been derived from the genome elements in the chimeric vaccine backbone or proposed spike protein.” The group stressed that the broader origins investigation remains open and that critical data gaps persist.The <a href="https://usrtk.org/wp-content/uploads/2026/03/12-5-2025-Partial-Production-for-USRTK-v.-NIH-Case-No.-24-cv-2785-JEB-NIH-FOIA-Request-63076.pdf">records</a>, recently obtained by U.S. Right To Know, do not contradict SAGO’s narrow conclusion about whether SARS-CoV-2 could have arisen directly from the DEFUSE proposal as written. But they show that multiple research proposals were already exploring similar approaches to altering how bat coronaviruses enter cells — experiments aimed at testing whether genetic tweaks could expand the viruses’ ability to infect new hosts, including humans.Internal NIH reviews also show that agency scientists recognized the potential hazards. As early as 2016, reviewers warned that modifying spike binding or cleavage sites in recombinant coronaviruses could produce “novel and unexpected” viral traits, even as similar work continued under other federally funded grants.The documents show that projects with clear conceptual overlaps were proposed, debated, revised, rejected and approved in various combinations in the years leading up to 2019 — with the same researchers repeatedly involved.“That’s absolutely true,” said Dr. Stanley Perlman, a University of Iowa coronavirus researcher, when asked if the broader scientific ecosystem pushed to explore how bat coronaviruses could become more infectious. “There’s no question about that.”The newly released records include internal <a href="https://usrtk.org/wp-content/uploads/2026/03/12-5-2025-Partial-Production-for-USRTK-v.-NIH-Case-No.-24-cv-2785-JEB-NIH-FOIA-Request-63076.pdf">NIH correspondence</a> from the Obama-era pause on funding certain <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4996883/">gain-of-function</a> experiments, as well as previously blacked out University of Minnesota emails now <a href="https://usrtk.org/wp-content/uploads/2026/03/12-5-2025-Partial-Production-for-USRTK-v.-NIH-Case-No.-24-cv-2785-JEB-NIH-FOIA-Request-63076.pdf">unredacted</a> after five years. Together, they reconstruct how U.S.-Chinese coronavirus collaborations operated and how federal officials assessed the risks. At the center of many of those exchanges is <a href="https://med.umn.edu/bio/fang-li">Fang Li</a>, a University of Minnesota virologist whose laboratory became a key structural biology hub connecting multiple research groups. Li’s expertise in mapping how spike proteins interact with receptors and antibodies positioned his team to identify key viral features while collaborators constructed recombinant viruses, swapped spikes, and ran animal-infection studies.One of Li’s collaborations with Baric in 2016 drew scrutiny inside the NIH. Records show agency reviewers concluded that a proposed experiment altering receptor binding in SARS-like bat coronaviruses could generate a virus with enhanced risk — and ultimately blocked the work under the federal gain-of-function funding pause.<h2 class="wp-block-heading" id="h-novel-and-unexpected-risks">“Novel and unexpected” risks</h2>In spring 2016, a White House-ordered <a href="https://grants.nih.gov/grants/guide/notice-files/not-od-15-011.html">pause</a> on funding certain gain-of-function research was reshaping virology, and NIH officials were pressing grant applicants to clarify whether their coronavirus experiments might fall within the restrictions.The moratorium applied to influenza, SARS, and MERS viruses. Closely related bat coronaviruses sometimes fell outside the pause because they had not yet been shown to infect humans. That distinction became central in early 2016 when NIH reviewed a renewal <a href="https://usrtk.org/wp-content/uploads/2026/03/AI089728.pdf">application</a> from Li and Baric titled “Receptor recognition and cell entry of coronaviruses.”Baric, a University of North Carolina virologist widely known for pioneering coronavirus reverse-genetics systems, had collaborated for years with Li and other international researchers studying how coronaviruses adapt to new hosts.On March 31, 2016, NIH <a href="https://usrtk.org/wp-content/uploads/2026/03/NIH-GoF-3.31.2016.pdf">notified</a> the University of Minnesota that the Li-Baric grant might include research subject to the funding pause and asked the scientists to clarify whether any proposed work could result in “enhanced pathogenicity and/or transmissibility in mammals via the respiratory route.”A few days later, Li and Baric sent their <a href="https://usrtk.org/wp-content/uploads/2026/03/Pages-from-Li-Baric-NIH-grant-records.pdf">response</a>.Most of the grant, they wrote, involved structural biology and pseudovirus systems. But one component — “Experiment 4” — proposed creating live recombinant SARS-like viruses with mutations designed to test how efficiently the viruses could infect different species. In particular, the experiments would test how efficiently the engineered virus could use ACE2 receptors — proteins on the surface of cells that some coronaviruses use as a doorway to infect them.The scientists argued that stronger receptor binding – which their experiment was designed for – did not necessarily translate into increased “pathogenicity,” or the virus’s ability to cause disease. They also promised to halt the work if engineered viruses showed substantial increases in replication, meaning the virus multiplied more rapidly in infected cells.Inside NIH, the proposal’s reviewers saw potential risk.In an internal “<a href="https://usrtk.org/wp-content/uploads/2026/03/Li-Baric-Biohazard-Comment.pdf">biohazard comment</a>,” a grants manager warned that recombinant coronaviruses engineered to enhance spike cleavage or strengthen ACE2 binding “may have novel and unexpected virulence phenotypes” – or, new and unpredictable traits that could make the virus more dangerous.The comment recommended allowing the work under <a href="https://www.niaid.nih.gov/research/biodefense-biosafety-labs">biosafety level-3</a> precautions before the proposal went to the agency’s internal gain-of-function oversight committee.On May 18, 2016, NIH officials made a <a href="https://usrtk.org/wp-content/uploads/2026/03/NIH-determination-letter-5.18.2016.pdf">decision</a>.While allowing other parts of the grant to proceed, the agency blocked Experiment 4. Engineering SARS-like viruses with enhanced receptor binding, NIH concluded, fell under the federal gain-of-function funding pause and “may not be conducted under this grant,” according to a letter signed by NIH program officer Dr. Erik Stemmy, who oversaw coronavirus grants at the agency.The determination marked an early instance in which NIH formally concluded that altering receptor binding in bat SARS-like viruses could plausibly create a more dangerous pathogen.It also exposed inconsistencies in how the pause was applied. While NIH halted Li and Baric’s proposed “enhanced affinity” experiments, the agency <a href="https://usrtk.org/covid-19-origins/nih-committee-green-lighted-wuhan-coronavirus-experiments-despite-concerns/">allowed</a> closely related work to proceed under an EcoHealth Alliance grant involving the Wuhan Institute. In that case, reviewers — including Stemmy — concluded that the bat coronaviruses under study had not yet been shown to infect humans and therefore fell outside the pause’s scope, despite internal <a href="https://drive.google.com/file/d/1GEbpzab90GJjM7ZT02ssbR7m4Vbj3JLY/view?pli=1">questions</a> about the experiments.The decision to block Li and Baric’s experiment did not settle the broader scientific debate — or stop the scientists from proposing similar experiments.<h2 class="wp-block-heading" id="h-a-regulatory-gray-zone">A regulatory gray zone</h2>Despite NIH’s 2016 rejection, Li and Baric continued exploring related ideas as federal policy evolved.In March 2017 — two months after the first Trump administration took office — Li <a href="https://usrtk.org/wp-content/uploads/2026/03/Li-email-to-Stemmy-April-2017.pdf">contacted</a> Stemmy about a new proposal.Baric, Li wrote, was “considering making a synthetic construct of a chimeric bat SARS-like coronavirus.”Li forwarded a <a href="https://usrtk.org/wp-content/uploads/2026/03/Baric-letter-new-SARS-Uganda-chimeras-.pdf">letter</a> signed by Baric formally proposing the idea: create a hybrid virus by combining most of one bat SARS-like coronavirus, SHC014, with the spike protein from another bat virus recently discovered in Uganda. <a href="https://usrtk.org/wp-content/uploads/2026/03/Li-Baric-emails-about-Uganda-sequence.pdf">Emails</a> show Baric had obtained the full sequence of the Uganda virus a month earlier from Columbia University virologist Simon Anthony, who at the time was collaborating with the EcoHealth Alliance and others on the <a href="https://whc.vetmed.ucdavis.edu/predict-project">PREDICT</a> project — a 10-year USAID-funded effort led by researchers at the University of California–Davis to catalog emerging viruses in wildlife.The new Baric and Li proposal also called for modifying the spike region that binds to ACE2 receptors — essentially adjusting how tightly the Uganda hybrid could attach to cells of different species.The team planned to test whether the engineered virus could infect cells carrying human, mouse, bat or civet receptors. If the modified virus replicated efficiently, the researchers proposed constructing full versions of the Uganda virus incorporating those receptor-enhancing mutations.In their letter, the scientists argued the experiment would not meet the federal definition of a potential pandemic pathogen, because neither of the original bat viruses had been shown to cause disease in humans.They also promised to stop the work if any engineered virus replicated more than ten times better than the SARS virus used as a benchmark.Two weeks later, Stemmy <a href="https://usrtk.org/wp-content/uploads/2026/03/Li-email-to-Stemmy-April-2017.pdf">replied</a> that NIH’s internal committee had not yet reached a decision.The proposal, he wrote, had landed in a regulatory transition.“Nothing to report yet on this,” Stemmy wrote. “Our internal committee hasn’t met yet. It’s a little bit of a gray area at the moment since the GoF research funding pause is still technically in effect while the department implements the P3CO policy that will replace it.”The available records do not indicate what decision NIH ultimately reached on the Uganda chimera proposal.Neither Li nor Baric responded to requests for comment. The NIH and Stemmy also did not respond to questions about its grant decisions.The exchange illustrates how researchers continued to explore new combinations of bat coronavirus genomes and spike proteins even as federal officials struggled to define where the gain-of-function boundary lay.<h2 class="wp-block-heading" id="h-green-lights-amid-red-flags">Green lights amid red flags</h2>The Li-Baric grant was not unique. Across multiple projects during the same period, NIH reviewed — and frequently approved — experiments designed to alter receptor-binding domains, swap spike proteins between viruses, or modify cleavage sites that influence how coronaviruses infect cells.Under a grant led by Baric and Vanderbilt University virologist Mark Denison, researchers proposed generating mouse-adapted versions of the SARS-related bat viruses WIV1 and SHC014. Internal reviewers noted the resulting strains were likely to show enhanced pathogenicity or transmissibility in animals. An early <a href="https://usrtk.org/wp-content/uploads/2026/03/NIH-draft-to-Denison-and-Baric-project.pdf">draft</a> response questioned whether the work fit the grant’s aims, but a later <a href="https://usrtk.org/wp-content/uploads/2026/03/NIH-April-authoritization-draft-Denison-Baric.pdf">version</a> prepared to approve the experiments.Another Baric <a href="https://usrtk.org/wp-content/uploads/2026/03/May-2015-determination-BARIC-SARS-MERS.pdf">project</a> constructed SARS chimeras bearing SHC014 or WIV1 spike proteins. NIH concluded those experiments did not meet its gain-of-function definition at the time, even though the purpose of such spike swaps was to test whether bat viruses could infect new hosts.Earlier <a href="https://usrtk.org/wp-content/uploads/2026/03/NIH-Determination-Letter-Baric-MERS-HKU-1.pdf">approvals</a> in 2015 included experiments inserting MERS cleavage-site sequences into the related bat virus HKU4 and altering receptor-binding regions to increase how efficiently the virus could attach to cells. Alterations to MERS cleavage sites — modifications that can influence viral entry efficiency – also were allowed under the same grant.Taken together, the records show NIH blocking certain receptor-binding experiments under the gain-of-function pause while allowing closely related spike-swap and cleavage-site research to proceed under other grants.By early 2018, that regulatory landscape formed the backdrop for a more ambitious proposal: <a href="https://drasticresearch.org/wp-content/uploads/2021/09/main-document-preempt-volume-1-no-ess-hr00118s0017-ecohealth-alliance.pdf">DEFUSE.</a>Informal scientific exchangesThe day-to-day emails also show how ideas and genetic sequences moved informally among collaborators.In one February 2018 <a href="https://usrtk.org/wp-content/uploads/2026/03/Li-email-to-Shi-about-sharing-422-spike-sequence.pdf">exchange</a>, Li told Shi, the renowned WIV coronavirus researcher, that Baric wanted access to her unpublished sequence of a spike protein from a MERS-like bat coronavirus known as “422.”Li wrote that he had “mentioned” Shi’s recent work with the virus to Baric, who asked if he could “get the sequence of the 422 spike protein.”“I said that I would need to check with you,” Li added. Shi replied that she saw “no problem” sharing the information if Baric’s work didn’t overlap too much with her own. She also described experiments her lab already had attempted.“We have already done this swapping on the MERS-CoV backbone,” she wrote.The recombinant virus, Shi said, could be rescued after transfection, but “couldn’t grow in the following passage.”“We have stop (sic) here for the moment,” Shi added. “I would encourage him to have a try.”The exchange offers a look inside the collaborative scientific ecosystem reflected throughout the records, one in which American and Chinese researchers shared sequences, experimental ideas and preliminary findings in real time.The following month, Baric and Shi joined EcoHealth Alliance in submitting the DEFUSE proposal to DARPA.<h2 class="wp-block-heading" id="h-scientists-debate-the-risks">Scientists debate the risks</h2>Stanley Perlman, the University of Iowa coronavirus researcher who collaborates with Li, said the broader research ecosystem described in the records was real, though loosely organized.Researchers, he said, were trying to determine which bat coronaviruses “could infect human cells” and therefore posed pandemic risk.At the same time, Perlman said he believes some of the scientific questions driving chimera experiments could often be addressed using simpler approaches.“It wasn’t necessary to make chimeric viruses to get some information,” said Perlman, who added he believes SARS-CoV-2 originated naturally. “Chimeric viruses are not my favorite way.”Two other scientists who separately reviewed the newly surfaced NIH files were more critical.Simon Wain-Hobson, a British-French virologist who has long opposed gain-of-function research, called one of Baric’s proposed recombinant virus concepts “Bonkers.”“After the GoF flu virus controversy,” he said, referencing a decade-long debate over whether researchers should intentionally make deadly viruses more contagious to study them, “this shows that Baric has learned nothing.”Steve Massey, a bioinformatics professor who also examined the records, said they reveal what he sees as a recurring pattern: researchers — particularly Baric — pushing experimental boundaries, using technical language to “bamboozle” reviewers and persuading them to approve studies he believes constituted gain-of-function research.Massey also pointed to what he sees as a throughline: proposals to alter “human protease cleavage sites” in MERS, which he said resemble later cleavage-site engineering debates around SARS-CoV-2.“Such experiments could easily enhance pathogenicity or transmissibility,” Massey said. “This is playing with fire.”<h2 class="wp-block-heading" id="h-post-outbreak-testing-the-furin-site">Post outbreak: Testing the furin site</h2>The records also include a <a href="https://usrtk.org/wp-content/uploads/2026/03/Post-outbreak-RaTG-SARS2-proposal.pdf">proposal</a> drafted early in the pandemic that focused on one of SARS-CoV-2’s most debated features: its furin cleavage site.In early 2020, Baric and Li proposed experiments inserting that cleavage site into RaTG13 — the bat coronavirus most closely related to SARS-CoV-2 — along with additional mutations affecting spike binding and viral entry.The goal was to test whether those changes could allow the virus to infect new species or make it infect cells more easily.The researchers acknowledged the possibility that such work might require additional review under the Potential Pandemic Pathogen Care and Oversight (P3CO) rules, the federal framework implemented after the gain-of-function funding pause.They proposed conducting the experiments under strict biosafety conditions while also pursuing loss-of-function studies designed to weaken the virus.But the proposal also made clear what the scientists expected the mutations might do.“We anticipate,” the researchers wrote, that inserting the furin cleavage site into RaTG13 may increase the virus’s ability to infect and cause disease in living organisms.<h2 class="wp-block-heading" id="h-what-the-records-show">What the records show</h2>The newly surfaced documents do not prove that SARS-CoV-2 was engineered or escaped from a laboratory.But they provide contemporaneous evidence of how researchers and federal officials were thinking about coronavirus engineering years before the outbreak began.NIH reviewers warned that modifying spike proteins could create “novel and unexpected” viral traits. Scientists debated how far those experiments should go. And proposals to reshape receptor binding or cleavage sites appeared across multiple grants.By the time the pandemic began, and even before DEFUSE was rejected, the tools and scientific concepts for tuning how coronaviruses enter human cells were no longer speculative. They had already been proposed, debated in federal oversight letters and pursued across an international network of collaborating laboratories.Lewis Kamb is an investigative reporter at U.S. Right to Know.  He was the first national FOIA Reporter for NBC News and spent nearly a decade reporting for the Seattle Times, where he was part of the reporting team awarded the Pulitzer Prize for exposing the failures behind the Boeing 737 MAX crashes.