On Tuesday, July 21, 2026, experts from the European Molecular Biology Laboratory (EMBL Hamburg) and the Leibniz Institute for Research in Molecular Pharmacology (FMP) released an in-depth study that describes how the influenza A virus alters the structure of human cells after infection. The innovative methodology allowed the team to observe interactions between proteins within intact cells, in contrast to previous methods that required fragmentation of samples.
Globally, seasonal flu is responsible for around 3 to 5 million serious cases and up to 650,000 deaths annually. Historically, the influenza A virus has also been behind major pandemics, such as the devastating Spanish flu of 1918, which highlights the importance of understanding its mechanism of action.
Once the virus invades the cell, it injects its RNA, which carries the instructions needed to synthesize viral proteins. These proteins disperse, reorganizing the host cell’s molecular systems to convert it into a factory of new viral particles.
Understanding the behavior of the flu virus in non-fragmented cells
Having deeper knowledge about these interactions is crucial so that the scientific community can create more potent vaccines and antivirals against influenza. To achieve this, it is essential to identify which viral proteins interact with human proteins, the locations of these interactions and how the virus manipulates them to replicate.
This new study represents a milestone, being the first to map direct contacts between influenza virus proteins and human proteins on a large scale, within infected cells that remain intact. The level of structural precision achieved allowed researchers to simulate how proteins interact and dock.
“Our work provides a new approach to investigating interactions between the influenza virus and its host in its natural environment, providing valuable structural information,” said Jan Kosinski, group leader at EMBL Hamburg and the Center for Structural Systems Biology (CSSB). He added that “the current findings are a glimpse into one stage of infection and pave the way for examining virus-host interactions throughout the infectious cycle.”
New methodology solves challenges in observing the virus
Monitoring interactions between proteins during an active infection is a very complex task. Previous research often used biochemical techniques that required cells to rupture before protein contacts could be measured.
Such a procedure can introduce distortions into the real scenario inside the living cell. When the internal compartments are disrupted, proteins that were previously isolated can artificially interact in the laboratory environment. Furthermore, more fragile, transient or localized connections may be lost, making it difficult to determine genuine interactions during infection.
“It was at this moment that we realized that our partners, Boris Bogdanow and Fan Liu from FMP Berlin, had created an improved version of cross-linking mass spectrometry (XL-MS), an established method for identifying protein contacts, but now specifically tuned for virus-infected cells,” explained Kosinski.
Learn more: Scientists discover that Earth is home to up to 20 million species of insects
This specialized methodology represented a crucial step forward for the group. It made it possible to record interactions that happen for very short periods or in specific areas of an infected cell.
“XL-MS allows us to record protein-protein interactions directly in intact infected cells, as well as offering structural data on the mechanics of these interactions”, detailed Bogdanow, now Junior Research Group Leader at the Charité Institute of Virology – Universitätsmedizin Berlin. He added that “this provides an understanding of the interface between the virus and the human cell and can, via structural modeling, assist in identifying intervention points for future drug development.”
Improved AlphaFold: Bridging Experiments and Structural Modeling
The research team integrated the data obtained with XL-MS with computational structural modeling techniques. This combination made it possible to identify viral and human proteins that interact, in addition to estimating the positioning of these proteins during connection.
In creating these structural models, the scientists used an adapted version of AlphaFold, the renowned Nobel Prize-winning protein structure prediction algorithm.
“The great advantage of the modified AlphaFold methodology was the possibility of directly incorporating our experimental cross-linking data into structural modeling,” clarified Kosinski. “This insertion guides the model as to which portions of the viral and host proteins are in close proximity in infected cells, which has proven particularly useful for virus-host complexes, which are often challenging to predict accurately.”
Flu virus uses two main tactics to control cells
The study’s findings, published in the journal Nature Microbiology, point to two notable strategies that the influenza A virus employs to take control of cells.
The first strategy involves hemagglutinin, a protein present on the surface of the virus. The flu virus uses it to connect and penetrate host cells. The researchers monitored this protein as it traveled through the complex internal cellular transport and processing network.
This network is made up of specialized compartments that fold, modify and prepare proteins before directing them to their final destinations. The analysis revealed that several human proteins contributed to the correct folding and modification of hemagglutinin during infection, with some of these host proteins having previously poorly elucidated functions.
Flu virus dissolves structures in the cell nucleus to replicate
The second major discovery focused on paraspeckles, small droplet-shaped compartments found inside the cell nucleus. The team observed that infection with the influenza A virus caused these structures to dissolve.
Full coverage: Latest News (EN)
With the rupture of the paraspeckles, proteins that bind to RNA and that were stored there were released. The virus can subsequently employ these proteins to drive its own replication.
“Paraspeckles were the most surprising revelation for us,” commented Iuliia Kotova, former Kosinski Group PhD Fellow at EMBL Hamburg, now at ETH Zurich, and lead author of the publication. She emphasized that “seeing these tiny organelles in the nucleus disintegrate, consistently across all cell lines and flu strains tested, suggests to us that this is not a mere side effect of the infection, but rather a possible viral strategy.”
“There may also be a second benefit for the virus: there is evidence that paraspeckles help in responses to cellular stress and in the modulation of antiviral genes. Therefore, disorganizing them could also weaken segments of the cell’s defense”, added Kosinski.
Collaboration between institutions drives innovative research
This project was the result of the joint expertise and technology of three important institutions. Crosslinked mass spectrometry was conducted by researchers at Charité, located in Berlin. Glycoproteomic analyzes were completed at the EMBL Proteomics Center.
Modeling using AlphaFold was performed by the team on the EMBL computing cluster, while microscopy images were generated at the CSSB’s Advanced Light and Fluorescence Microscopy (ALFM) Facility.
Methodology opens doors to the study of viruses with pandemic potential
The study’s conclusions highlight how investigating molecular contacts in intact infected cells can elucidate where and how a virus takes control of human cellular machinery. Experts indicate that this “mapping in context” has the potential to clarify the functioning of other viruses as well.
“Although host-specific factors and mechanisms often vary among viruses, our conviction is that the general approach—which integrates intracellular cross-linking, structural modeling, and targeted monitoring of cell biology to map innate virus-host interactions at distinct phases of infection—remains broadly applicable,” said Kosinski.
Although the research focused on a laboratory-adapted strain of influenza, those involved believe that the same methodological strategy could be applied in the future to investigate viruses with a higher pandemic risk.
Bogdanow corroborated: “While this study addressed a laboratory-adapted strain, it supports the application of the methodology to viruses of potential pandemic relevance, such as H5N1, and the identification of the interaction networks that support their proliferation in human cells.”