Cryo-Neutron Protein Crystallography Produces Sharper Structures

A SHARPER FOCUS:

© 2004 National Academies of ScienceComparison of three water sites captured under various conditions demonstrates the power of cryo-neutron crystallography. Associated water molecules W1 (left), W6 (center), and W75 (right), were imaged by neutron crystallography at 15K (top) and 293K (middle) and by X-ray crystallography at 110K. Note the changing positions of hydrogen atoms in all three structures, and the extra D2O position near W6 revealed at 15K. (Reprinted with permission from M.P. Blakeley et al., Proc Natl Acad Sci, 101:16405–10, 2004.)X-ray crystallography has long been the method of choice for obtaining atomic-resolution protein structures, but it doesn't provide the whole picture. Hydrogen atoms, particularly labile hydrogens that interact with solvent, are often lost in the noise. "It is exceedingly difficult – usually impossible – to see a single hydrogen atom with X-rays," says Dean Myles, a crystallographer at the Oak Ridge National Laboratory (ORNL) in Tennessee and director of the facility's center for structural molecular biology.To overcome this problem, some protein crystallographers use neutron crystallography, which, as the name implies, relies on interactions between neutrons and atomic nuclei. Myles is a member of a research team that recently described the determination of a high-resolution, low-temperature neutron structure of the saccharine-binding lectin, concanavalin A.1X-ray diffraction studies rely on the interaction of X-rays with the electron distribution around a molecule, Myles explains. The intensity of that interaction at any one atom is roughly proportional to the square of the number of electrons associated with that atom. Thus carbon, with six electrons, has a scattering factor of around 36; nitrogen's is 49, and oxygen's is 64. Hydrogen, with its lone electron, has a scattering factor of 1. "When we have a hydrogen sitting beside a carbon, oxygen, or nitrogen, we're looking at 1/36th, 1/50th, or 1/64th of the signal from hydrogen, and basically that gets swamped," says Myles.Neutrons diffract with the nucleus, however, so relative scattering is more uniform. "If we collect neutron diffraction data on a protein crystal where we already know the carbon/nitrogen backbone, then we expect the hydrogen signal to show up at the same level as a carbon, nitrogen, or oxygen," Myles explains. Thus the positions of hydrogen atoms (which are replaced with the heavy hydrogen isotope, deuterium, in the experiment) within the protein structure, as well as those of hydrogens in water molecules bound to the protein, can readily be found, and the resulting data can be combined with X-ray diffraction data to obtain a complete, low-temperature protein structure.Running neutron crystallography at low temperatures (cryo-neutron crystallography) fixes atoms in as static a position as possible, boosting resolution. It also allows researchers to compare spectra with those obtained from X-ray diffraction experiments, which also are typically run at low temperatures to avoid radiation damage to the protein. "The extra clarity that you get of the structure that comes with that [approach] is very pleasing, but the second benefit that you get is that it then opens up the opportunity for making [comparisons of] the very-low-temperature to room-temperature structures," says John Helliwell of the University of Manchester, one of the study's authors.When Myles' team compared their 15K (-258°C) structure, providing 2.5-angstrom resolution, against a 2.4-angstrom, room-temperature neutron crystallography structure, they found they could more accurately pinpoint water molecules within it. The new structure includes 227 associated water sites, of which 167 are modeled as molecular water and 60 as oxygen; that's twice the number provided by the room-temperature (20°C) structure.At present the main drawback to neutron crystallography is the cost. According to Helliwell, a typical X-ray crystallography run is eight hours long; a neutron experiment can take 100 times longer. "So you've got something like a factor of 100 increase in the cost, because the unit price is similar," he says. But, he points out, "What you get at the end of it is as complete a structure as is possible, right down to the last atom that is to be determined."Furthermore, neutron sources are weaker than X-ray sources, and larger crystals are needed to compensate for the difference in flux. But freezing large crystals is challenging: Protein crystals can average 50% solvent content, and cooling them using standard laboratory freezing techniques can lead to ice formation, which destroys the protein crystals. Helliwell and colleagues flash-cooled their crystals in liquid nitrogen in the presence of a cryoprotectant, but this technique does not work for all proteins. Some may interact unfavorably with the cryoprotectant, says Leif Hanson, a neutron crystallographer at the University of Toledo in Ohio, who wrote an accompanying commentary about the research.2Efforts to improve crystal-cooling methods are ongoing. "We will be working on several projects, including protein crystal systems that don't survive room-temperature data collection, and improved flash-cooling protocols for large crystals," Hanson writes in an E-mail.Meanwhile, the development of more intense beam lines, such as the one currently under construction at the Spallation Neutron Source at ORNL, should reduce the run times and the need for large crystals. Myles says the new source promises a gain of two orders of magnitude in signal-to-noise over current sources.

Cryo-Neutron Protein Crystallography Produces Sharper Structures

X-ray crystallography has long been the method of choice for obtaining atomic-resolution protein structures, but it doesn't provide the whole picture.

Interested in reading more?

Become a Member of