AINEWS
A 15-year method push lifts cryo-EM's small-protein floor
Berkeley's Theia microscope and a custom laser phase plate make visible the class of proteins that has blocked structure based drug design for a decade.
Berkeley's Theia microscope and a custom laser phase plate make visible the class of proteins that has blocked structure based drug design for a decade.
For more than a decade, cryo-electron microscopy could resolve large proteins but missed the small ones that make up the bulk of human biology. A class of targets, often too faint and too small for the technique's resolution floor, has been structurally invisible to the field. Now, a 15-year engineering effort at UC Berkeley and Lawrence Berkeley National Laboratory, published in Science, has built a custom laser phase plate that lifts that floor, imaging proteins down to roughly one-third the previous size threshold, according to Lawrence Berkeley National Laboratory's News Center.
The proof point is hemoglobin, a small, low-contrast protein near today's lower size limit for cryo-EM. Standard instruments struggle to render it cleanly. The new system, called Theia, renders it well, the Berkeley team reports. The team also tested aldolase, an easier benchmark, and saw only a small improvement, which is what the physics predicts when the bottleneck is sample prep rather than instrument contrast. That asymmetry, not the easier case, is the honest signal that the method does what it claims.
The mechanism is a phase plate, adapted from the light-microscopy technique that won Frits Zernike the 1953 Nobel Prize in Physics. Earlier attempts to bring phase contrast to electron beams failed because the plates themselves were unstable or absorbed too much of the beam. The Berkeley design sidesteps that by trapping an intense laser in a mirrored cavity less than four inches across, intensifying it to roughly 350 to 400 gigawatts per square centimeter, then focusing it on a spot about one-thousandth the width of a human hair, the lab explains. The electron beam passes through this laser interaction region and picks up a phase shift that turns invisible contrast into visible contrast.
The optical cavity sits inside a 14-foot Thermo Fisher Krios column, the standard cryo-EM platform, modified with extra electron optics that give Theia better resolution than a stock Krios even before the laser is engaged, per the lab's announcement. Co-first authors Jessie Zhang and Petar Petrov, both postdocs in Müller's group, tested the system on six biological samples of varying size and prep quality. The largest gains came on the smaller, harder-to-image targets. The full method appears in Science, DOI 10.1126/science.aeh0665, with the detailed resolution metrics behind a paywall.
The idea traces back to a 2010 collaboration between Müller and Robert Glaeser, a Berkeley professor emeritus of molecular and cell biology whose foundational work on cryo-EM was credited in the 2017 Nobel Prize in Chemistry. Roughly fifteen years of theory and prototyping followed, partly funded by the NIH, with a 2021 Chan Zuckerberg Biohub grant enabling the Krios purchase and ramp-up, the Biohub notes. The 15-year arc matters because it situates this as a method breakthrough built on iteration, not a single clever trick.
The structural-biology stakes are concrete. Researchers estimate that more than 90 percent of proteins inside human cells are too small for current cryo-EM, a figure cited in the LBNL release as a researcher's estimate rather than a measured literature survey. That estimate describes a class of targets, including many drug-binding proteins, whose structures have been out of reach for structure-based design. Lifting the floor does not solve drug discovery on its own, but it converts a structural-biology blind spot into a tractable problem.
Independent outside voices frame the significance in cellular terms. David Agard and Bridget Carragher, founding directors of imaging at the Biohub, characterized the technique as meaningful for cellular imaging in the release accompanying the paper. Müller's group plans next to extend the system from single-particle cryo-EM to cryo-electron tomography, which assembles 3D views of molecules in their native cellular context rather than as isolated particles, the lab says.
A parallel effort is already underway. The Biohub is independently building a dual-laser variant, called xLPP, described in a bioRxiv preprint that has not yet been peer reviewed. The dual-laser theory behind it was published last week in Nature Communications. Both groups collaborate with Thermo Fisher, and the two systems are distinct: Theia is the single-laser instrument at Berkeley whose result appears in Science; xLPP is the follow-on, dual-laser system at the Biohub.
The current scope is narrow. Theia is a prototype on a custom instrument, not a turnkey product. Access is limited to Müller's collaborators, and only benchmark targets have been demonstrated. The press release leans on phrases like "Formula 1 of microscopes" and "the world's best instrument overall," but those are Müller's own framing, not independent benchmarking, and the marketing language should be read as aspirational. The honest claim is narrower: a credible, demonstrated path to imaging proteins that standard cryo-EM has had to leave in the dark, on a single instrument, with hemoglobin as the proof.
What to watch next is whether the hemoglobin result generalizes to other small, low-contrast proteins, and whether the planned cryo-electron tomography extension preserves the contrast gain when targets are imaged inside cells rather than as isolated particles. If it does, structural biology will have a tool for the class of targets that has resisted it for the duration of the cryo-EM era.