Cross-Linking Mass Spectrometry Reveals How Influenza A Virus Rewires Host Cells at Unprecedented Structural Resolution

In a study published in Nature Microbiology, researchers at EMBL Hamburg (search)Search term, the Leibniz Research Institute for Molecular Pharmacology (search)Search term (FMP), and collaborators have produced the first large-scale map of direct protein-protein contacts between the influenza A virus (search)Search disease and its human host cells—captured inside intact, infected cells with sufficient structural detail to model how the proteins physically fit together.

Seasonal influenza kills up to 650,000 people globally each year and causes serious illness in 3–5 million individuals. The influenza A virus (search)Search disease has been responsible for several pandemics, including the 1918 Spanish Flu. When the virus infects a cell, it releases its RNA, which encodes a handful of proteins that spread throughout the host cell and repurpose its molecular machinery to produce more virus. Understanding exactly how viral proteins interact with host proteins has long been a central goal in virology, but previous methods required breaking cells open before measuring interactions—introducing artifacts and losing fragile or location-specific contacts.

"Our work provides a new way to study flu-host interactions in their native context and with structural insight," said Jan Kosinski, Group Leader at EMBL Hamburg (search)Search term and the Centre for Structural Systems Biology (CSSB). "The current results are a snapshot of a moment during infection, and it opens the door to studying flu-host interactions across the entire infection cycle."

A Critical Methodological Breakthrough

The key advance came from a specialized version of cross-linking mass spectrometry (XL-MS) developed by Boris Bogdanow and Fan Liu at FMP Berlin, tailored specifically to virus-infected cells. Unlike conventional biochemical approaches, XL-MS captures protein-protein interactions directly inside intact cells while simultaneously providing structural information about the interaction interfaces.

"XL-MS allows us to capture protein-protein interactions directly in infected intact cells, while also providing structural information about how these interactions are happening," explained Bogdanow, now a Junior Research Group Leader at the Institute of Virology, Charité – Universitätsmedizin BerlinView company profile. "This gives us insight into the interface between the virus and the human cell and may, through structural modelling, help identify actionable targets for future pharmaceutical interventions."

Structural Modeling with Modified AlphaFold

By feeding the experimental cross-linking data into a modified version of AlphaFold—the Nobel Prize-winning protein structure prediction algorithm—the team could not only identify which viral and human proteins interact but also predict their three-dimensional binding configurations.

"The key advantage of the modified AlphaFold approach is that it allowed us to feed our experimental cross-linking data directly into the structural modelling," Kosinski said. "This tells the model which parts of the viral and host proteins are close to each other inside infected cells. This was especially useful for virus-host complexes, which are often difficult to predict reliably."

Two Key Hijacking Mechanisms Uncovered

The study revealed two important ways the virus commandeers host cell machinery.

First, the researchers traced how haemagglutinin (search)Search target—the surface protein influenza A uses to bind and enter host cells—moves through the cell's internal transport and processing network. This revealed how host proteins, some with previously unknown functions, assist the virus in correctly folding and modifying haemagglutinin during infection.

Second, and most strikingly, the team discovered that influenza A infection causes paraspeckles—small, droplet-like compartments in the nucleus—to dissolve. This releases RNA-binding proteins normally sequestered within these organelles, which the virus then exploits for its own replication.

"What surprised us most was the paraspeckles," said Iuliia Kotova, former predoctoral fellow in the Kosinski Group at EMBL Hamburg (search)Search term, currently at ETH Zurich, and first author of the publication. "Watching these tiny organelles in the nucleus dissolve, consistently across every cell line and every flu strain we tested, told us this isn't a side effect of infection – it might be a strategy."

Kosinski added that there may be a second benefit for the virus: "Some evidence suggests paraspeckles contribute to cellular stress responses and antiviral gene regulation, so disrupting them could also weaken parts of the cell's defence response."

A Platform for Broader Application

The work relied on shared infrastructure across three institutions: cross-linking mass spectrometry at Charité in Berlin, glycoproteomics analyses at the EMBL Proteomics Core Facility, AlphaFold modeling on the EMBL Compute Cluster, and microscopy imaging at CSSB's Advanced Light and Fluorescence Microscopy (ALFM) Facility.

The researchers believe this "mapping in context" approach can be applied broadly to other viruses. "While the exact host factors and mechanisms often differ from virus to virus, we think our overall approach—combining in-cell cross-linking, structural modelling, and targeted cell-biology follow-up to map native virus-host interactions at specific stages of infection—remains broadly applicable," Kosinski said.

Bogdanow concurred: "Although this study has focused on a lab-adapted strain, this study lays the groundwork to apply the methodology to viruses of potential pandemic relevance, such as H5N1, and for uncovering the interaction networks that support their multiplication in human cells."