Michael Rossmann is a veteran virus crystallographer, with a standard X-ray tube right down the hall from his lab at Purdue University. But when he needs to perform ab initio structure determination - which requires tunable X-rays many orders of magnitude brighter than his standard Purdue X-ray tube - he and his colleagues pile into the car with some toiletries, a change of clothes, and a dewar full of frozen protein crystals for the 120-mile drive across the Indiana state line to the Advanced Photon Source (APS) at Argonne National Laboratories, just outside Chicago. APS is one of the country's premier synchrotron facilities.
He can't just show up, however; the APS, like all synchrotrons, has a queue, and anyone wanting time must submit a proposal. There are 21 synchrotrons worldwide engaged in protein crystallography; six are in the United States. According to David Moncton, director of the nuclear reactor laboratory at the Massachusetts Institute of Technology and former associate director at Argonne, these facilities average between 5,000 and 6,000 hours per year up-time. To fill its time, the APS issues three calls for proposals each year, during which reviewers consider some 1,000 proposals and 2,000 beamline requests, according to deputy director Dennis Mills. About 60% of new proposals are approved annually, based on scientific merit. Successful applicants might therefore wait anywhere from a few weeks to five months for their turn.
In Rossmann's case, that delay has practical implications. "Virus crystals are notoriously unstable," he says, and sometimes they degrade before they can be placed in the beamline. Even if they don't degrade, the travel and scheduling issues still cause considerable inconvenience. Instead of running experiments whenever they please, crystallographers must hoard their crystals in advance of their beamline dates, and then run them all at once. That means there's little time for "what if" work - the research that stems from a postdoc's late-night eureka moment - not to mention on-the-fly experiments based on data collected earlier in the day.
Palo Alto-based Lyncean Technologies is working on a tunable minisynchrotron called the Compact Light Source (CLS) that could plug these gaps. Unlike the APS, with its 1,104 meter storage ring, the CLS sports a storage ring small enough for any lab, measuring just 1 x 2 meters. Its brilliance should fall somewhere between that of an X-ray tube and a synchrotron, says Ron Ruth, Lyncean president and chief scientist, and its cost will be on par with a 900-MHz NMR spectrometer - a few million dollars, give or take.
At that price the CLS could be a reasonable investment for universities and companies with large crystallography groups, if installations of high-field NMRs and electron microscopes are any guide. Lyncean plans to place its first beta instrument at the Scripps Research Institute in La Jolla, Calif., within a year, and Ruth says he hopes to have several orders by then. "We believe the market is quite substantial," he says. "We could produce about 10 systems per year." Scripps cell biologist Peter Kuhn, a consultant for Lyncean and co-principal investigator of the group that will run the pilot study, hasn't decided which protein he'll use to inaugurate the CLS, but "it will hopefully be something that will produce a Science paper," he says.
On March 2 Lyncean announced a major milestone in the instrument's development: the production of its first X-ray beams. The CLS's actual operational parameters are not yet known, but the underlying physics are sound, says Moncton. The CLS, he explains, is based on Compton scattering, in which the collision of a laser beam with an electron beam generates X-rays. "The question isn't, does the process work? We know it works. The question is, how intense and well focused is the electron beam, and how intense and well focused is the laser beam? Because that's what determines how many X-rays you're going to get out."
Moncton's team at MIT is developing what he calls a second-generation technology based on the same principle, but using a superconducting linear accelerator instead of an electron storage ring. "Electron beams in linear accelerators preserve a much higher degree of beam brightness, because they don't bend around corners," explains Moncton. Nevertheless, he says, "We're very excited about his work, and we're cheering him on."