Cryoelectron microscopy (cryoEM for short) involves freezing identical viruses or large biomolecules in a thin sheet of vitreous ice. The ice serves three purposes. First, it provides an aqueous environment that resembles the molecule's natural setting. Second, it immobilizes the molecules, thereby forestalling their eventual destruction by the energetic electrons used to image them.
The third purpose is perhaps more subtle. Even when molecules are immobilized in ice, the electron dose that any single one of them receives before it's destroyed is too low to yield a crisp diffraction pattern. X-ray crystallography faces a similar problem. It's not fatal because each crystal contains myriad copies of the same identically arrayed molecule. Collectively, the molecules share the photon dose. X-ray diffraction patterns tend to be crisp because they are averages.
Thanks to the ice's glassy structure, the individual molecules in a cryoEM experiment are frozen in random directions. Although that arrangement might seem disadvantageous, it means that the individual images correspond to different views. Under the assumption that the molecules in those views are identical, a computer algorithm can virtually align all the views and take their average to yield a three-dimensional image. For viruses and large molecules, which defy crystallization, avoiding the need to make a crystalline sample is a welcome advantage.
CryoEM surface map of the 70-nm-wide ε15 bacteriophage determined at a resolution of 4.5 Å. Two types of capsid proteins are rendered: stapling protein (red) and coat protein (cyan). The icosohedral capsid is made up of 60 hexons and 12 pentons. The yellow line outlines one full hexon and 1/5 of a penton. CREDIT: Corey F. Hryc, Baylor College of Medicine
I first learnt about cryoEM in 1999, when I wrote a news story for Physics Today about what was then a nascent field. One of the researchers I interviewed for the story, Wah Chiu of Baylor College of Medicine in Houston, was developing cryoEM to study viruses. At the time, he and his collaborators had recently determined the structure of the rice dwarf virus, which measures 700 Å across, with a spatial resolution of 25 Å.
Now Chiu and his team routinely resolve viruses near atomic resolution (3.0-4.5 Å). Their 2013 study of the Salmonella bacteriophage ε15 could even pick out the side chains of the amino acids that make up the virus's protein coat. (For more on cryoEM, read Robert Glaeser's Physics Today article, "Cryo-electron microscopy of biological nanostructures," which appeared in January 2008.)
Cold atoms and lasers
Although I had been keeping an eye on progress being made in cryoEM, I was totally surprised by a recent paper in Structural Dynamics by Jom Luiten of Eindhoven University of Technology and his collaborators.
If you want to image a static biomolecule, the limiting factor is the total number of electrons. Depending on the molecule's size, you need around 106 to 107. But if you want to see how that biomolecule changes shape, you need to deliver those electrons in short, bright bunches. That's difficult with conventional electron sources because the electrons that emerge from them, being hot and charged, resist being concentrated.
In their paper, Luiten and company describe a different approach to electron sources. First they optically trap and laser-cool a cloud of rubidium-85 atoms. After exciting the atoms' valence electrons from the 5s state to the 5p state, they photoionize the atoms with a 100-femtosecond pulse.
When high-voltage coils pull the electrons out of the trap, the electrons have an effective temperature of just 10 K. Thanks to its low temperature, the bunch of electrons remains tight and coherent as it makes its way through magnetic lenses to converge on the target.
Luiten and his collaborators demonstrated the feasibility of their ultracold electron source using a single-crystal graphite target. They successfully obtained the telltale six-spot diffraction pattern expected from the hexagonal crystal.
In their experiment, each 100-fs bunch contained a few hundred electrons. To obtain a distinct diffraction pattern, Luiten and his collaborators needed to combine the patterns from 1000 bunches. Reaching the point where a single bunch contains 107 electrons will entail developing methods to compensate for the electrons' mutual electrostatic repulsion.
But those methods are already on the drawing board. Pump–probe experiments in which the structure of a complex biomolecule is determined as it's induced to change shape are on the horizon.