The highest possible resolution we can get in a typical image is limited by the wavelength of the light we’re using. Although there are some clever ways around this limit, one alternative has been to use something with a smaller wavelength. That “something” turns out to be electrons, and the electron microscope has provided a glimpse of the details inside cells, showing us how their parts are ordered and structured.
But this year’s Nobel Prize in Chemistry went to a group of individuals who pushed the electron microscope to its very limit, figuring out how to use it to determine the position of every single atom in large, complex molecules. The award goes partly to a researcher who successfully used electron microscopes to image proteins. But it also goes to two people who developed some of the techniques to make the whole thing work: figuring out how to freeze water quickly enough that it formed a glass and developing an algorithm that could take a large collection of random data and convert it into a coherent picture.
For years, understanding the structure of a complex molecule like a protein or RNA involved a technique called X-ray crystallography. As its name implies, this works by shooting X-rays through crystals of the molecule in question. But there’s a pretty major limitation to this technique: your molecule has to form a crystal. Not all proteins do. In fact, some rather important classes of proteins completely refuse to do so, like the ones that are embedded in a cell’s membrane.
But X-rays aren’t the only things that bounce off the atoms in a molecule; electrons will as well. And if you can image the patterns the electrons form after they bounce off, you can infer where the atoms are. An electron microscope is nothing more than a system for focusing beams of electrons and recording where they end up. So it should, in principle, work to get you information on where all the atoms are, which tells you the structure of the molecule.
But the gap between principle and reality here is a canyon-sized chasm. To begin with, electrons tend to transfer energy to atoms they bump into, which will break chemical bonds and destroy the molecule. So you want as few electrons as possible to do the imaging. Yet most of the electrons will pass by a molecule without interacting with it, meaning the actual signal is tiny. To bring it up, you need a lot of electrons (which will, of course, damage it).
Electron microscopy is done in a vacuum. But lots of proteins have water molecules tightly linked to them, essentially acting as part of the structure. Put them in a vacuum, and the water goes away, taking the structure with it. You can freeze the water so it stays in place in a vacuum, but then the ice crystals also interact with the electron, creating a huge signal that swamps your protein.
Then there’s the issue of orientation. Without a crystal, each individual protein molecule you try to image will be oriented in a different direction. Since you don’t know which way any two molecules are pointing, you can’t combine the data from both of them to enhance the signal.
You might think that this would be enough to convince anyone that the use of electron microscopy for figuring out protein structures was an impossibility. But scientists can be an extremely stubborn bunch.
Very different contributions
Richard Henderson of the MRC Laboratory in Cambridge is cited for his early work in the field. He took a protein that normally forms arrays where all the proteins are oriented in the same direction. Using this, he was able to get a rough idea of the outline of the shape of the protein, though nothing close to the position of any atoms. It was an important validation of the approach, even though finding proteins that naturally formed regular arrays made it of limited utility at the time.
But the technology kept improving. Jacques Dubochet, the new Laureate at the University of Lausanne, figured out how to handle the problem of the water associated with proteins. With the right technique, the water could be frozen in a glass-like form, with all the molecules being randomly oriented. In this state, the water didn’t impart a pattern into the electrons, which allowed the signal from the proteins to stand out.
(Dubochet, now an honorary professor, has placed his CV online, and it lists his earliest scientific achievements as “Conceived by optimistic parents” and “No longer scared of the dark, because the sun comes back; it was Copernicus who explained this.”)
But the critical breakthrough was probably solving the orientation problem, where most proteins will land on a surface in a random orientation. The solution came courtesy of Joachim Frank, as did improvements in computational power. Frank figured out how to get computers to recognize when similar patterns of electrons indicated that the protein was in a similar orientation. This let researchers scan through large populations of proteins, taking brief images of each—brief enough that the electrons didn’t destroy the protein. The computer would then identify images with similar patterns, align the data precisely, and add the signals from multiple images to help them stand out from the noise.
Progress didn’t end with the laureates. Better microscope hardware, especially improved electron detectors, were critical to continuing to refine the techniques. But today, cryo electron microscopy has progressed to the point where it’s possible to get atomic-resolution images of proteins, telling us precisely how their constituent parts combine to create their structure and function—no crystals needed.
Powered by WPeMatico