Every year, seasonal flu kills up to 650,000 people worldwide and sickens millions more. We’ve had vaccines and antivirals against it for decades, but there’s still a lot we don’t know about exactly how the influenza A virus takes over a human cell once it gets inside. A new study from EMBL Hamburg and the Leibniz Research Institute for Molecular Pharmacology (FMP) in Berlin just filled in some of that picture, in unusually fine detail, and without breaking the cells open to look.

That last part matters more than it sounds. Traditionally, scientists studying how viral and human proteins interact have had to lyse the cell first, essentially popping it open, to measure which proteins are touching. The problem is that once a cell’s internal compartments are destroyed, proteins that were never actually in contact can end up mixing together in the test tube, muddying the results. It becomes hard to know which interactions were real and which were test-tube artifacts.

The team, led by Jan Kosinski, sidestepped that problem using a specialized version of an established technique called cross-linking mass spectrometry, adapted specifically for infected cells by collaborators Boris Bogdanow and Fan Liu. It lets researchers chemically “freeze” proteins in place while they’re still touching inside an intact, living cell, then identify exactly which proteins were in contact, including fleeting or location-specific interactions that would normally be lost. They combined that data with a modified version of AlphaFold, the protein-structure-prediction tool that won its creators a Nobel Prize, to model not just which viral and human proteins interact, but how they physically fit together.

Two discoveries stood out. First, the researchers traced how hemagglutinin, the protein flu uses to latch onto and enter cells, travels through the cell’s internal protein-processing system, and found host proteins, some previously unstudied, that help fold and modify it correctly during infection.

The second finding was the more surprising one. Infection consistently caused paraspeckles small droplet-like structures in the cell nucleus, to dissolve, releasing RNA-binding proteins that the virus can then repurpose for its own replication. That dissolution happened reliably across every cell line and every flu strain the team tested, which is what convinced the researchers it isn’t incidental damage, it looks like a deliberate viral strategy. There’s a second possible payoff for the virus too: paraspeckles appear to play a role in the cell’s antiviral stress response, so dissolving them may also blunt part of the cell’s own defenses.

This is a snapshot from one point in infection with a lab-adapted flu strain, but the method itself is the bigger story, the researchers believe the same “mapping in context” approach could be turned on other viruses, including those with real pandemic potential, like H5N1.

Original paper: Kotova, I., et al. “Mapping in-cell protein contact sites reveals hijacking of paraspeckles during influenza A virus infection.” Nature Microbiology, 20 July 2026. https://www.nature.com/articles/s41564-026-02416-1