A UC Berkeley and LBNL team shows that a focused continuous wave laser inside a Titan Krios sharpens phase contrast for the protein class that has long been cryo EM's blind spot, with a dual laser successor already in development at the Chan Zuckerberg Biohub.
A laser-driven phase plate installed on a cryo-electron microscope has pushed the technique past one of its oldest limits: imaging small proteins, a class that makes up roughly 90 percent of the human proteome and includes many drug targets that were structurally invisible until now. In a paper published in Science, physicist Holger Mueller and colleagues at UC Berkeley and Lawrence Berkeley National Laboratory show that driving a continuous-wave laser directly into a transmission electron microscope sharpens phase contrast for proteins that conventional cryo-EM struggles to resolve.
The advance matters because cryo-EM's signal-to-noise floor has kept proteins below roughly 70 kilodaltons mostly out of reach. Mueller's group, as reported in Genetic Engineering & Biotechnology News, describes a custom laser phase plate, or LPP, installed on a Thermo Fisher Titan Krios. The phase plate uses a beam the team characterizes as roughly 75 kilowatts of optical power focused to a few microns, intense enough to phase-shift the electrons themselves rather than relying on conventional defocus-based contrast.
Demonstrations on apoferritin and hemoglobin, two reference proteins routinely used to benchmark cryo-EM hardware, produced sharper structures, better recovery of information from early frames, improved particle picking, and cleaner 3D classification, all within standard defocus ranges and reconstruction pipelines. Those are the workflows structural-biology labs already run, so the gain arrives as a hardware insert rather than a methodological overhaul.
The practical reach is what gives the result more than a microscope tweak. Mueller's team is pushing toward imaging proteins as small as 17 kilodaltons, a size class that includes peptides, many hormones, and fragments of larger drug targets, and that has historically been cryo-EM's blind spot. The paper's claim is bounded: the demonstration set is two proteins, not a full proteome survey, and the gain shows up most clearly on small particles and noisy specimens. For large, well-behaved proteins, conventional cryo-EM may still be the right tool.
Independent reaction has been cautiously optimistic. At the Chan Zuckerberg Biohub, founding technical director of imaging Bridget Carragher and VP of imaging science Stephani Otte told GEN the laser-driven approach is "really exciting," particularly for cryo-electron tomography in crowded cellular contexts where contrast is hardest to recover. The Biohub group is already developing a dual-laser variant of the LPP, designed to reduce component wear and minimize optical aberrations, a sign that the technique is moving from one lab's prototype toward broader use, though the dual-laser system remains in development rather than deployed.
Mueller's own framing is the most useful guide to where the LPP fits. The phase plate matters most where cryo-EM has historically been weakest: small particles, and bad specimens, the ones that look blurry under standard imaging conditions. That is a precise, falsifiable claim. It also leaves open the larger question of throughput. The paper does not address how a high-power continuous-wave laser behaves in a facility that runs multiple samples per day, how often the phase plate needs realignment, or whether the technique transfers cleanly to microscopes other than the Titan Krios. Those are the watch items for the next round of work.
For structural biology, the immediate consequence is access. Proteins that were structurally invisible because they were too small, or too rare, or because the sample was too messy to defocus into legibility, now sit in a tool's working range. Whether that translates into new drug leads is a downstream question, but the upstream step, getting a clean structure in the first place, has just gotten easier for a specific and large class of targets.