HomeBlogThe Virus That Might Be Eating Your Gut Bacteria

The Virus That Might Be Eating Your Gut Bacteria

A science explainer on a contested hypothesis about SARS-CoV-2 and the gut: what is early-stage and disputed, and what is well supported.

In 1915, long before anyone had seen a virus, the English bacteriologist Frederick Twort noticed that his bacterial colonies were developing glassy, transparent holes. Something invisible was killing them from the inside. A few years later the French-Canadian Félix d’Hérelle named the culprits bacteriophages, literally “bacteria eaters,” and those clear holes in a lawn of bacteria, called plaques, became the classic signature of a virus that preys on microbes. Phages are everywhere, including your intestines, where billions of them keep your resident bacteria in check.

Coronaviruses are supposed to belong to a different world entirely. They infect animal cells, not bacteria. So it raised eyebrows when a small group of researchers reported seeing those same telltale plaques on plates of human gut bacteria inoculated with material from a COVID-19 patient. Their question is deliberately provocative: could SARS-CoV-2, under some conditions, behave like a phage, hiding and multiplying inside our gut microbes, and could that hidden population explain why some people never get better?

The hypothesis

The idea comes mainly from Carlo Brogna, Mauro Petrillo and colleagues, who published a series of papers in F1000Research, Microorganisms and related journals between 2022 and 2025. Their starting observation was that SARS-CoV-2 RNA appeared to increase in bacterial growth medium seeded with gut bacteria, rising and falling along with bacterial growth, and shifting when antibiotics were added. That pattern, they argued, looks less like a virus sitting inertly in a dish and more like one replicating at a host’s expense.

In their 2022 paper on bacteriophage potential, the group pushed further. They took supernatant from the stool microbiota of a COVID-positive person and spotted it onto agar plates covered with two common gut species: Faecalibacterium prausnitzii, one of the most abundant and most anti-inflammatory bacteria in a healthy colon, and Dorea formicigenerans. Plaques appeared. Sequencing RNA scraped from those plaques recovered SARS-CoV-2 genetic material, though only partial coverage of the viral genome, roughly 62% from one species and 48% from the other. Antibodies against the viral nucleocapsid protein labeled material both outside and apparently inside bacterial cells, and mass spectrometry suggested newly made spike protein rather than leftover input virus. A later experiment fed the cultures nitrogen-15, a heavy isotope, so that any protein built inside the dish would carry a detectable mass signature.

If any of this holds up, the implications for Long COVID are obvious. One leading explanation for the condition, which affects an estimated 6% of American adults, is a viral reservoir: a pocket of virus or viral material that survives after the acute infection clears and keeps provoking the immune system. Bacteria would make an unusually good hiding place, shielded from antibodies and from antivirals designed for human cells, and capable of releasing virus again whenever the gut is stressed. It would also offer a tidy explanation for why some patients shed viral RNA in stool for months.

Why most virologists do not buy it

The objections are structural, and they are serious.

A bacteriophage and a coronavirus are built for different jobs. Most phages are naked protein capsids that inject DNA through a bacterial cell wall using machinery honed for exactly that purpose. SARS-CoV-2 is an enveloped RNA virus: it is wrapped in a lipid membrane, it enters cells by binding the ACE2 receptor found on eukaryotic cells (cells with a nucleus and internal compartments, which bacteria lack), and it assembles new particles by hijacking the endoplasmic reticulum and the Golgi apparatus before budding out in vesicles. Bacteria have none of those compartments. Bacterial ribosomes also read messenger RNA differently from ours, relying on sequence signals that the coronavirus genome has no reason to carry. A coronavirus genome inside a bacterium should be, in the bluntest terms, unreadable junk.

Then there is the problem of artifacts. Detecting viral RNA or protein in a bacterial culture does not prove replication inside bacteria. The virus may simply be stuck to the cell surface, swallowed into vesicles, or carried along as contamination from the original stool sample, which is itself teeming with real bacteriophages that could produce plaques all on their own. Critics have pointed to thin negative controls, the difficulty of ruling out native phages, and electron microscopy images where particles are identified as virions without independent confirmation. Partial genome coverage cuts both ways: it is what you would expect from degraded fragments, not only from active replication. To the authors’ credit, they say the same thing, writing that their findings are “in contrast with the current knowledge of virology” and explicitly inviting other labs to try to reproduce them. Several years on, no independent group has published a clear confirmation.

The part nobody disputes

Strip away the phage question and a robust body of evidence remains, and it still points at the gut.

SARS-CoV-2 demonstrably reaches intestinal tissue and lingers there. A 2021 Nature study by Gaebler and colleagues found viral nucleocapsid protein in the small-bowel lining of 5 of 14 people roughly four months after diagnosis, with nasal swabs already negative. Work in Gastroenterology found viral RNA in gut mucosa about seven months after mild illness in 32 of 46 patients with inflammatory bowel disease, and post-COVID symptoms clustered in those with persistent antigen. More recently, PET imaging combined with rectal biopsies found persistent T-cell activation in the gut wall and, in four of five Long COVID patients, viral RNA, including double-stranded forms that hint at ongoing replication. Note what this is: persistence in human intestinal cells, exactly where coronavirus biology predicts, with no bacteria required.

Meanwhile the microbial community itself gets rearranged. This is dysbiosis, a shift in the balance of resident microbes toward a less healthy mix. In a prospective study in Gut, Liu and colleagues followed 106 patients and found that those with persistent symptoms at six months had more Ruminococcus gnavus and Bacteroides vulgatus and less Faecalibacterium prausnitzii, while people who recovered fully had gut communities that looked like those of uninfected controls. Depleted butyrate producers matter because butyrate, a short-chain fatty acid made when bacteria ferment fiber, feeds the cells lining the colon and helps keep the gut barrier and immune signaling calm. Lose it, and inflammatory molecules leak into circulation, a plausible route to the fatigue, brain fog and malaise patients describe. Similar disruption shows up in the mouth and airway, and the gut’s own phage and fungal populations shift alongside the bacteria. Mouse experiments add a provocative note: transplanting stool from post-COVID patients into germ-free mice produced worse lung inflammation and brain dysfunction.

Why the distinction matters

These two stories demand different medicine. If virus genuinely replicates inside gut bacteria, then the target is the bacterial reservoir, and the treatment conversation turns toward selectively clearing it, perhaps even with engineered phages of our own. If the real mechanism is persistent virus in human intestinal tissue plus a destabilized microbiome, the path runs through extended antiviral trials, now underway with drugs like nirmatrelvir, paired with microbiome repair: targeted probiotics, butyrate producers, fiber, or fecal transplantation.

For now, the honest summary is layered. That SARS-CoV-2 disturbs the gut microbiome and can persist in intestinal tissue is established. That it behaves like a bacteriophage is a minority hypothesis resting on a handful of papers from a single research network, unreplicated and at odds with basic virology. The hypothesis deserves what its own authors asked for, which is a serious attempt at replication, because the cost of ignoring a real reservoir is high and the cost of chasing a laboratory artifact is a detour. Twort’s glassy holes, after all, were dismissed as a curiosity for years before anyone understood what they meant.

This article is general education about published research, not medical advice. It does not describe a treatment offered at Venturis Clinic. Long COVID care should be discussed with your own physician.

If you have questions about your own health, you can book a free 15-minute call.

Written by Dr. Alvin Philipose, DC, ICCP (Ret.)

author avatar
Dr. Alvin Philipose, DC, ICCP (Ret.)
Dr. Alvin Philipose, DC, ICCP (Ret.), founder of Venturis Clinic. Care is provided by the Venturis clinical team.