sciencebriefs
13:00in productionCh. 1 · The crystallisation problem/ 13:00 · ceiling 15 min
Life sciences · Engineering

Cryo-electron microscopy

Freezing biological molecules in place fast enough to stop ice crystals forming let electron microscopes see proteins as they actually sit in solution, ending structural biology's dependence on crystallisation.

Richard Henderson spent decades pushing electron microscopy toward a goal many colleagues doubted was reachable: resolving the atomic structure of proteins without first growing them into crystals, which many biologically important molecules simply refuse to do. His early work on bacteriorhodopsin in the 1970s and a full atomic-resolution model by 1990 showed the approach could work in principle, using flash-frozen samples held in vitreous, non-crystalline ice. The technique, cryo-electron microscopy, only became broadly practical after 2012, when direct electron detectors and better computational methods produced what researchers called a resolution revolution, letting labs resolve viruses, ribosomes, and ion channels in near-native detail. Henderson shared the 2017 Nobel Prize in Chemistry with Jacques Dubochet and Joachim Frank for developing the method into a working structural biology tool.

Chapters & takeaways6
  1. 0:08
    The crystallisation problem

    X-ray crystallography, the standard method for seeing molecular structure, requires a protein to form a crystal first, and many will not.

  2. 2:10
    Freezing without ice crystals

    Cryo-EM flash-freezes samples fast enough to trap water in a glass-like, vitreous state rather than ordinary crystalline ice, preserving the molecule's natural shape.

  3. 4:20
    A membrane protein in 1975

    Henderson and Nigel Unwin's early electron microscopy work on bacteriorhodopsin showed membrane proteins had definite, resolvable structures.

  4. 6:30
    Atomic resolution by 1990

    Henderson's electron crystallography model of bacteriorhodopsin reached atomic detail, only the second such result for any membrane protein at the time.

  5. 8:40
    The resolution revolution

    Direct electron detectors and improved computation after 2012 let cryo-EM resolve far smaller and more varied structures than before.

  6. 10:50
    A shared Nobel Prize

    Henderson, Dubochet, and Frank were recognised jointly in 2017 for turning cryo-EM into a genuine alternative to crystallography.

Worth your time?

Yes. Study the whole thing.

4/ 5
What works
  • it shows how a technique widely doubted for decades became mainstream once a specific technical bottleneck was solved
  • Henderson's own multi-decade arc, from a 1975 low-resolution model to a 2017 Nobel Prize, gives the science a clear throughline
  • the contrast with X-ray crystallography makes clear what problem the method actually solves
What does not
  • the computational side of image processing that made the resolution revolution possible is only gestured at
  • the specific contributions of Dubochet and Frank alongside Henderson's are not developed in comparable depth
Study it if
  • anyone curious how scientists now see the shape of a protein without growing a crystal of it
  • readers interested in long, unglamorous methodological progress that eventually pays off
  • people who want the actual mechanics behind recent structural biology headlines
Skip it if
  • readers wanting worked technical detail on electron detector hardware
  • anyone looking for a single dramatic discovery moment rather than a decades-long development
The written brief4 min read

The crystallisation problem

For decades, the standard way to see a protein’s atomic structure was X-ray crystallography, which requires coaxing the molecule into a regular, repeating crystal lattice before firing X-rays through it. That step works well for some proteins and is nearly impossible for others, particularly the membrane proteins that sit half-embedded in a cell’s outer wall and tend not to hold a stable crystal shape at all. Richard Henderson, working at Cambridge’s Laboratory of Molecular Biology from the early 1970s, pursued an alternative: using electron microscopy directly on the protein instead, without requiring a crystal. His early collaboration with Nigel Unwin on bacteriorhodopsin, published in 1975, produced a low-resolution model showing the protein arranged as seven helices crossing the membrane, evidence that membrane proteins had definite, mappable structures after all.

Freezing without ice crystals

The core technical problem was that ordinary freezing turns water into crystalline ice, and ice crystals distort or destroy the delicate arrangement of a biological molecule before it can be imaged. The solution, developed and refined over subsequent years by Henderson and others, was to freeze samples fast enough that the water forms vitreous ice instead, a disordered, glass-like state that traps the molecule roughly as it existed in solution rather than crushing it into a crystal lattice. Henderson pushed his own bacteriorhodopsin work toward ever finer detail, and by 1990 he had produced an atomic-resolution model of the protein using electron crystallography, only the second atomic-level structure ever obtained for a membrane protein by any method, a result that argued the electron microscopy route could in principle match X-ray crystallography’s precision.

A membrane protein in 1975

The method Henderson championed held up, eventually, as a genuine and now widely used alternative for exactly the cases X-ray crystallography struggles with. The turning point came around 2012 and 2013, when direct electron detectors, cameras that record electrons hitting them far more efficiently than earlier film or indirect sensors, arrived alongside improved computational algorithms for combining thousands of noisy individual images into one sharp structure. Together these advances produced what researchers have called a resolution revolution, and cryo-EM structures reaching near-atomic detail, and in some cases resolutions as fine as roughly 1.2 Ångströms by 2020, became achievable for viruses, ribosomes, ion channels, and large enzyme complexes that had previously resisted structural study by any technique.

Atomic resolution by 1990

What the method still does not fully replace is X-ray crystallography’s usefulness for very small, well-behaved molecules that crystallise easily and can be resolved to extremely fine detail that way with comparatively modest equipment. Cryo-EM’s advantage grows with the size and structural awkwardness of the target, and it took a long stretch of methodological work, spanning Henderson’s own career from the 1970s through the 2010s, before the computational and hardware pieces caught up with the underlying idea. The gap between having a working concept and having a practical, widely deployable tool was measured in decades, not years, which is easy to lose sight of when the technique now appears as a routine part of structural biology papers.

The resolution revolution

The wider consequence is that structures once considered essentially unreachable, including many membrane proteins that are the targets of a large share of modern drugs, became visible in detail without needing to solve the separate, often intractable problem of growing them into crystals. That has fed directly into drug design and into basic understanding of how large molecular machines, ribosomes among them, actually assemble and function. The 2017 Nobel Prize in Chemistry, shared by Henderson with Jacques Dubochet, who worked out the vitrification freezing method, and Joachim Frank, who developed key image-processing approaches, recognised that the achievement was distributed across several complementary contributions rather than belonging to any one person’s single breakthrough.

A shared Nobel Prize

This is worth understanding because it is an unusually honest example of how a scientific method actually matures: not through one triumphant experiment but through a long sequence of partial results, from a rough 1975 helix map to a 1990 atomic structure to a 2012 hardware leap, each step necessary and none sufficient on its own. Readers drawn to dramatic single discoveries may find the pace slower than they expect, since the interesting part is the accumulation rather than any one moment. But for anyone who wants to understand why cryo-EM structures now appear so routinely in biology and medicine, and why it took until the 2010s rather than the 1970s for the promise Henderson saw early to actually pay off, it is time well spent.

Same field · Life sciences4 of 78
Up next in Science

CT scan

· 13:00

Godfrey Hounsfield built a scanner from an idea on a country walk; Allan Cormack had already worked out the mathematics years earlier without knowing anyone intended to build the machine.

13:00