click to enlarge Rigaku’s CrystalMation helps researcher’s find effective ways to grow a needed crystal. (Image: Rigaku)

Identifying a protein’s structure offers many benefits to drug discovery and development, and improving technology makes this easier to do. “Classically, there are two general ways to look at protein structure,” says Douglas Meinhart, PhD, analytical products manager at JEOL in Peabody, Mass. “These are crystallographic and NMR [nuclear magnetic resonance] approaches.” He adds, “Each has its own set of challenges.”

For crystallographic approaches, the name itself reveals a key challenge—crystallizing the protein. For NMR on the other hand, a sample must be isotopically labeled, which requires an expression system.

Despite these challenges, researchers expect to use these approaches on more projects. In addition, advances in instruments help researchers gather more data in less time—sometimes even with less hands-on time.

Heating up cool NMR
“For drug discovery,” explains Meinhart, “NMR offers the benefit of working in solution, or closer to the native state of the protein.” In addition, NMR allows some dynamic measuring of protein structure. Nonetheless, NMR imposes a size limit on how large a protein it can handle. But as Meinhart adds, “Every year the size of proteins accessible to NMR goes up, mostly due to new NMR pulse techniques or improved isotopic-labeling strategies.”

click to enlarge To study protein structures in solution, researchers use nuclear magnetic resonance devices, such as JEOL’s ECA NMR. (Image: JEOL)

To make the most of identifying protein structures in the pharmaceutical business, it must happen fast. Japanese research institute RIKEN put together a few dozen NMR devices to form a high-throughput approach to studying protein structure.

Pharma started using NMR to better understand protein structures some time ago. In the mid-1990s, Abbott Laboratories in Abbott Park, Ill., developed what it called “SAR by NMR,” or using this technology to reveal the structure-activity relationship (SAR) of proteins.

Many proteins, though, are bound to membranes, which has made them difficult to study with NMR. “These are not well-behaved in solution, but sit in cell walls, mitochondrial walls, and so forth,” says Meinhart. Such proteins can be isolated, but they don’t function in solution. Nonetheless, Meinhart and his colleagues are working on a few prototypes that could allow NMR to study membrane-bound proteins. For example, the company’s 1-millimeter magic-angle spinning NMR probe can sharpen the signal even for bound proteins. This probe spins at 80 kilohertz. “It’s good for large molecules,” Meinhart says. The company is also testing a 2.5-millimeter probe that spins at 35 kilohertz.

Meinhart also points out the general goal of increasing sensitivity of NMR. “For about 10 years, cryogenic probes—with the electronics cooled to liquid-helium temperatures—have provided about three- to four-fold improvement in sensitivity. Now, we have a cryogenic probe that keeps the sample at room temperature and can be combined with isotopic labeling and allows spin-angle adjustment.” With that probe, says Meinhart, researchers “can extract additional parameters, like bond angles and lengths.”

Once a researcher uses an NMR device to collect data about a protein, Meinhart explains, they load that into software that reveals more about the structure. “Most people use variations of public-domain packages,” Meinhart says, “unless they are pharma, which often have their own proprietary software.”

Cracking crystallization
When using an x-ray approach to identifying the structure of a protein, growing the crystals is a challenge from the beginning. A researcher may face hundreds of stacks of multi-well plates, each well consisting of a different condition for growing protein crystals, such as variation in pH. Then, the scientist wants to find the best crystals. Maybe a technician looks through each well, looking for something crystalline.

click to enlarge Finding the conditions that grow the best crystals poses a challenge when using an x-ray technique to study protein structures, but that can be simplified with Varian’s PX Scanner, which can test the diffraction of suspected crystals in multi-well plates. (Image: Varian)

“Something might not look crystalline even when it is or it might look crystalline and just be salt,” explains Leigh Rees, PhD, director of XRD products at Varian in Oxfordshire, UK. Traditionally, someone would remove a suspected crystal and put it on a diffractometer. Instead, Varian’s PX Scanner takes a photograph of each well; if something looks like a crystal, an x-ray can be shot up through the bottom of that well. “This shows if it diffracts,” says Rees. “You can see if it’s protein or salt and how well it diffracts.” He adds, “You can even collect up to six degrees of data on a crystal, and we are studying whether this might be enough to get structural information.”

If a crystal looks promising on the PX Scanner, it can be taken to Varian’s SuperNova, which uses a microsource for x-rays. “This is a sealed tube,” Rees says, which can be easily handled and quickly changed when needed. He adds, “It only uses 50 watts of power, but compares well with rotating-anode generators.” This instrument can also be fitted with two sources in one device, so that it can diffract small molecules or proteins.

The SuperNova also helps researchers collect data. For example, it will automatically “assess the crystal quality and decide how to best collect data,” says Rees. “If you want to change the data collection, though, you can.” He adds that the data come out in a format that is made for easy input to a variety of protein-structure software packages, such as the CCP4 shareware suite.

Adding automation
In drug discovery, says Paul Swepston, PhD, manager of life science at Rigaku in The Woodlands, Texas, “researchers have a target they’ve identified, and their protein group has already figured out how to make the protein of interest in large quantities.” If the process produces large enough crystals, the protein can be examined with in-lab x-ray instruments. In some cases, though, protein crystals are so small that it takes a synchrotron to study the structure. “Some pharmas, though, do not let their compounds off the property,” says Swepston, “so they prefer to do as much in-house as possible. Some other customers might be happy to go to a synchrotron.”

If needed, researchers can use Rigaku’s CrystalMation to find effective ways to grow a protein crystal. “This is mainly used at facilities doing structural genomics,” according to Swepston.

Once a researcher has crystals of a protein of interest, Rigaku can help them probe it with x-rays. “Our strength is in x-ray sources,” says Swepston. “We have the highest flux—the largest number of x-ray photons—that you can create in-house.” He adds that synchrotrons always create the absolute highest flux. Rigaku also makes detectors to measure the diffraction created when its x-ray photons encounter a protein crystal.

To automate the process, Rigaku licenses a robotic sample changer—the automated crystal transport orientation and retrieval technology (ACTOR)—from Abbott Laboratories. “In pharma,” says Swepston, “you look at crystals over and over, and the robot lets researchers freeze samples in advance and then the x-ray system runs constantly.”

These general techniques will delve even deeper into protein structures with modified approaches already being developed. For instance, Swepston mentions the pixel-array detector for x-ray technology. “With this detector, there’s no dead time between images, and it can collect data almost in real time.” So far, Rigaku is only testing this detector, but Swepston says, “these new detectors are the future.”

In addition, Swepston points to small-angle x-ray scattering (SAXS) as a growing technology. “It will be hot in the next three to five years,” he says. “It can be used to determine the shape of proteins in solution, and it can be used on complexes of multiple proteins, such as molecular machines that assemble at different places in the cell.”
As these techniques move ahead, researchers will determine the structures of proteins that were inaccessible with past technology.

About the Author
Mike May is a publishing consultant for science and technology based in Houston, Texas.