Thomas Pühringer spent five years chasing a molecular ghost—a fleeting assembly of proteins that scientists had theorized for three decades but never seen. Now, in a breakthrough at the Francis Crick Institute, he’s captured the first image of the pre-initiation complex, the long-missing trigger that launches DNA replication with precision and fidelity.
Every time a cell divides, it must copy its DNA exactly once. Too little, and the cell dies. Too much, and genomic instability can lead to cancer. To prevent disaster, cells use a molecular checkpoint centered on the replicative helicase—an enzyme that unwinds the DNA double helix like a zipper. But turning on this helicase at the right time and place requires a precise sequence of events. For 30 years, one critical step remained invisible: the moment when three activator proteins converge on the helicase to form the pre-initiation complex, setting the stage for replication.
Pühringer, working in Alessandro Costa’s Macromolecular Machines Laboratory, recreated this process in yeast using a test-tube system originally developed by Crick collaborator John Diffley. Using cryo-electron microscopy, he froze the complex mid-formation, capturing its structure at near-atomic resolution. The challenge was immense—the complex exists for only a fleeting moment before springing into action. But with the help of AI tool AlphaFold, the team could model the most flexible regions of the complex that imaging alone couldn’t resolve. "This took us five years: without AI to complement the experimental work, it would’ve taken a lot longer," Costa says.
The resulting image revealed not just the architecture of the pre-initiation complex, but the dual roles of key "firing factors"—proteins that both assemble the machinery and help separate DNA strands. One such factor, once thought to have evolved different functions in yeast and humans, was shown to play the same role in both, suggesting a deeply conserved mechanism across species. This discovery resolves a longstanding mystery and strengthens yeast as a model for human DNA replication.
The implications extend beyond basic biology. Faults in this initiation process are linked to cancer and developmental disorders. By finally seeing how replication begins, scientists are better equipped to understand what goes wrong in disease. The next frontier, Costa says, is to capture the same process in multicellular organisms—"a richer sequel to the original film." With advances in imaging and AI, and strong collaborations across institutions, that chapter may not be far off. For now, one thing is clear: a molecular movie 30 years in the making has just gained its most crucial frame.