ZURICH. Researchers in Zurich have mapped the folding of a protein at a resolution fine enough to watch individual atoms settle into place, a feat long considered beyond the reach of structural biology. The work, carried out jointly at ETH Zurich and the University of Zurich, resolves the folding process into steps separated by less than a microsecond.
Proteins begin as floppy chains of amino acids and must fold into precise three-dimensional shapes to function; misfolding underlies diseases from Alzheimer's to cystic fibrosis. Until now, scientists could photograph the unfolded chain and the finished structure, but the transition between them was a blur that theory could only guess at.
The puzzle has a famous formulation known as Levinthal's paradox: a chain of 100 amino acids has more possible shapes than there are atoms in the universe, yet a real protein finds the correct one in thousandths of a second. Folding must therefore follow preferred pathways, like a river finding channels, and it is those channels that the Zurich team has now filmed directly.
The Zurich method combines two established techniques in a new way. Cryo-electron microscopy, which flash-freezes molecules and images them with electrons, provides the pictures, while a microfluidic chip mixes the protein with a chemical trigger so precisely that thousands of molecules begin folding within the same few microseconds. Staggering the freezing captures the process as a film of stills.
The result matters far beyond structural biology's specialist journals. Drug designers, who target pockets on protein surfaces, now have their first direct view of transient pockets that exist only during folding, openings that vanish in the finished structure and were therefore invisible to conventional drug screening.
The project drew on unusual depth across the Zurich research landscape. The microfluidic chips were fabricated at ETH Zurich's clean-room facility, the electron microscopes belong to the university's scientific imaging centre, and the computing to reconstruct the atom positions ran on the national supercomputer at CSCS in Lugano. Funding came principally from an ERC grant and the Swiss National Science Foundation.
The team is careful to note what the method cannot do. It works so far only on small proteins that fold in isolation, whereas many human proteins require helper molecules called chaperones or fold while embedded in membranes. Each additional layer of biological realism complicates the timing trick at the heart of the technique. The films also average over millions of molecules, so rare folding detours that cause disease may still escape the camera.
“We have, in effect, built a stop-motion camera for chemistry,” said the group's lead structural biologist. “Now we have to point it at proteins that matter for medicine.”
The path toward application runs through collaboration with Basel's pharmaceutical corridor, where two companies have already licensed access to the folding maps of disease-linked proteins. The immediate prize is a class of enzymes implicated in Parkinson's disease, whose misfolded intermediates have never been structurally characterised.
Work now under way will extend the method to chaperone-assisted folding within two years, using a modified chip that delivers the helper molecules at defined moments. The team also plans to deposit its folding films in a public archive so that other laboratories can test folding theories against the data.
Internationally, the result lands in a field transformed by artificial intelligence, which can predict final protein structures from sequence but says nothing about the route taken. The Zurich films offer exactly the intermediate data that the prediction algorithms lack, and several AI groups have requested access to benchmark their models.
Structural biology has spent 60 years photographing molecules that were persuaded to hold still. The Zurich work suggests the discipline's next era belongs to those that refuse to.