Main

Protein crystallization has always been a bit more of an art than a science. It is nearly impossible to predict exactly which conditions will cause a protein to form a large, high-quality crystal, so crystallographers typically often screen thousands of different conditions in an empirical fashion. “Although we have gotten much better at collecting data and solving crystal structures (as beamlines and computation have improved), actually growing the crystals has remained an obstinate bottleneck,” says University of British Columbia professor Carl Hansen.

Hansen is a former member of Stephen Quake's group at Stanford University, undoubtedly a leader in developing microfluidic methods to address the needs of protein crystallographers. The Quake laboratory quite successfully has been able to miniaturize and automate the crystallization screening process, greatly reducing the quantity of protein required and allowing many more screening conditions to be tested in parallel. These innovations have been widely embraced by those with large structural pipelines (such as pharmaceutical companies), who use the commercial TOPAZ System (Fluidigm) invented by the Quake lab.

As recently reported in Journal of the American Chemical Society, the Quake lab has now put forth a slightly unconventional alternative to the crystallization screening problem, by constructing a free interface diffusion (FID)-based microfluidic device that uses 'kinetic optimization' instead of varying chemical conditions. “On this device the chemistry is completely fixed; instead we control the mixing and dehydration rates to screen different trajectories through chemical space,” explains Hansen. Each FID crystallization chamber on the chip is accessed by fluid channels of different lengths, which control the mixing rates of the protein and precipitant. The chambers are separated from a fluid reservoir by a permeable poly(dimethylsiloxane) (PDMS) membrane, which regulates the hydration.

In addition, this device provides a very practical benefit to protein crystallographers. Crystals grown inside the PDMS membranes in the chip can be punched out of the device and mounted directly for diffraction studies, such that direct handling is never needed. High-resolution structures could be solved from the in situ diffraction data. “In all other crystallization formats this is really problematic, to the point where you almost always cannot tell whether the crystal was a good crystal that was damaged during harvesting, or if it was a poor crystal right from the start,” says Hansen. This valuable feature could save protein crystallographers a lot of time, not to mention headaches.