Making movies of molecular machines
They never wavered from the belief that the technique would work, despite many people telling them it was simply not possible. How can you subject a tiny biological molecule to temperatures of more than 1 million degrees Celsius—equal to that of the solar corona—in less than a trillionth of a second and expect to get any useful information from it? The extremely bright light alone from the laser can punch a hole through steel. The biomolecule doesn’t stand a chance.
But John Spence and Petra Fromme, two Arizona State University professors, persisted in their beliefs. That determination has led to a new, exciting technique that is building a track record of advances and gaining widespread acceptance in scientific circles.
The technique is called Serial Femtosecond Crystallography (SFX), and it has already unveiled the structural details of several proteins not seen before. SFX has been used to map out the structure of proteins involved in sleeping sickness, a fatal disease caused by protozoan parasites. It has been used to reveal the fine details of how an experimental drug, called an angiotensin II receptor blocker, works to regulate blood pressure, paving the way to developing better hypertension drugs. And it has provided the first snapshots of photosynthesis in action as it uses sunlight to split water into protons and electrons to make food for plants and oxygen for all of us.
Time-resolved SFX was lauded as one of the “top 10 breakthroughs of 2012” by the journal Science.
Research teams from all over the world have published hundreds of scientific papers describing in new detail the inner workings of biological molecules that were revealed by SFX.
“It’s the birth of a new field of science,” says Spence, one of the pioneers of the technique. Spence is a Regents’ Professor of physics at ASU and director of science for the National Science Foundation’s BioXFEL Science and Technology Center, a consortium of institutions devoted to the use of X-ray lasers for biology.
Most of the SFX work currently is carried out at the Stanford Linear Accelerator Center, a national facility and home to the world’s first hard X-ray laser, the Linear Coherent Light Source (LCLS).
In a delicate dance of target molecule and high energy X-rays, researchers have been successful in shining the X-ray laser on molecules and grabbing the data from the diffraction pattern it produces, all before the sample is fried by the laser. In SFX, everything takes place in femtoseconds (10-15 seconds). For scale, a femtosecond is to a second what a second is to 32 million years.
The “diffract then destroy” method of SFX operates on an ultrafast time scale to obtain pictures of the protein before the laser obliterates the sample. With tens of thousands of “pictures” taken for each protein, researchers can construct movies of molecular machines at work or interacting with specific molecules, providing a level of understanding that was previously unimaginable in structural biology.
“SFX will revolutionize the field of structural biology,” says Fromme, a key contributor to both the SFX method and the science it reveals. Fromme is a Regents’ Professor in ASU's School of Molecular Sciences and director of the Center for Applied Structural Discovery in ASU’s Biodesign Institute.
ASU researchers laid the foundation for development of the SFX method. They played key roles in developing much of the instrumentation essential for the SFX experiments at the LCLS, they have developed growth and detection techniques of nanocrystals from huge molecular complexes, they were involved in the first time-resolved SFX experiments, and they developed the methods of analyzing the huge amount of data from diffract-then-destroy experiments. For ASU students, the technique has provided an unparalleled educational experience.
Capturing elusive proteins
As scientists developed methods to use X-rays to unravel the structure of biomolecules, they were dismayed to find that the very high radiation dose needed at the highest magnification destroys the sample before a useful image can be recorded. But following earlier suggestions, in 2000 Janos Hajdu and Richard Neutze, biophysicists at Uppsala University in Sweden, created a simulation proposing that on a femtosecond timescale, even molecular explosions may unfold slowly.
They proposed that it would take roughly 10 femtoseconds for atoms to start moving significantly from their original positions. So if one could take a snapshot with a pulse of X-rays briefer than 10 femtoseconds, it should be possible to “outrun” radiation damage.
The X-ray laser, which amplifies the intensity of an X-ray beam exponentially, can pack a thousand billion photons of X-rays into each pulse. The LCLS reads out 120 of these scattering patterns from biomolecules every second, so that the scattering from tiny crystals running across the pulsed X-ray beam (each destroyed after scattering) can be detected.
In December 2009, just two months after the LCLS began operation for users, an international group including the ASU team, Hajdu’s team and Henry Chapman’s team from the German DESY lab in Hamburg, focused SFX on protein nanocrystals of Photosystem I made in Fromme’s ASU lab. Photosystem I is a membrane protein that functions as a biosolar energy converter that catalyzes the first step of photosynthesis.
Conventional protein crystallography, mostly undertaken at synchrotrons, determines the structure of molecules from large crystals containing trillions of molecules. For biological samples, such as proteins, the difficult part is to coax the biomolecules into a crystalline structure and then freeze the structure in place to minimize radiation damage. In the past, this meant trying to grow large crystals and mounting them at a precise angle to the X-ray beam, then rotating the crystals in the beam until diffraction from all crystal planes is recorded.
However, many of the most scientifically interesting biomolecules, including membrane-bound protein complexes, have long evaded structure determination. Membrane proteins are extremely important biomolecules, playing key roles in photosynthesis, respiration, cell transport, cell communication, nerve function, vision and hearing.
Some 60 percent of all current drugs are targeted to membrane proteins. However, knowledge of the structures and dynamics of these biomolecules lags far behind. While more than 100,000 structures of soluble proteins have been determined, fewer than 600 membrane protein structures are known to date.
SFX turned this practice on its head. Rather than using large crystals held to position, SFX uses nanocrystals free-floating in solution. Because they don't need to be frozen, SFX can get X-ray snapshots of the molecules in the crystal changing shape, and make molecular movies to reveal their function.
“Small is beautiful!” exclaims Fromme. Large crystals containing trillions of molecules can take years or decades to grow but often suffer from long-range defects. Nanocrystals, on the other hand, often contain only a few hundred molecules and are much better ordered.
Refining the process
But using nanocrystals created new challenges. How do you detect nanocrystals too small to be seen under a microscope, characterize them and position them in front of X-ray pulses that will make them explode, and do it consistently 120 times each second?
New methods had to be applied to detect the nanocrystals. SONICC (Second Order Nonlinear Imaging of Chiral Crystals) does this through second harmonics generation—when two infrared light pulses hit a chiral crystal with less than 10 femtoseconds between the pulses, frequency doubling occurs and the crystal emits a green photon. It can then be detected like fireflies in the night sky.
Spence worked with physicists Uwe Weierstall and Bruce Doak on the sample delivery challenge. They came up with a device that functions much like an ink-jet printer, firing tiny droplets of solution across the X-ray beam in a continuous stream with the nanocrystals in suspension.
A key component of this system is a sample injector that delivers a “fresh” crystal between each of the X-ray shots. Early versions of the injector had one major problem—the nozzle clogged. Weierstall, an ASU research professor, began work on a new clog-free design, which was further developed by ASU physics professor Doak and given to SLAC for all SFX users.
Weierstall designed a nozzle with a capillary-like tube surrounded by an opening through which helium gas is forced. As the solution exits the tube, the gas pushes it, thinning the solution and speeding up its motion.
“This gives you focusing of the beam without a physical hole,” Spence says. “We can have a big hole where the solution exits that doesn’t clog and still have a very fine jet.”
A further refinement was the delivery of the sensitive membrane proteins in their native “lipid-like” environment. Scientists doubted that this would be possible, as this medium has the consistency of grease. However, the new injector not only delivered the crystals in this “toothpaste” jet but also improved the method, so that the crystals could be delivered with 100 times less consumption of the precious membrane protein crystals.
“Uwe showed us how he could make toothpaste fly,” Fromme says.
Setting it in motion
Until now, most structures determined with conventional methods provide a static picture of the molecule. But complex biological processes like photosynthesis are very dynamic. To understand them, one would need a molecular movie of the biomolecules at work.
The Fromme group led the first time-resolved femtosecond crystallography studies on the water splitting process in plants. It is based on nanocrystals of Photosystem II, which are fully active. When they are exposed to multiple pulses of light, oxygen bubbles can be seen coming out of the crystals. A future goal is to unravel all conformational states and determine the full “movie” of the water splitting process.
SFX can potentially yield 120 snapshots per second, resulting in a huge amount of data. The team developed special software to handle the influx of data from the SFX method and to take the data and recreate the molecules it detected. The software helps the researchers build an accurate image of the molecule from the large number of snapshots taken during the experiment.
With the SFX technique and tools continuing to prove their usefulness in structural biology, ASU researchers are turning their attention to what’s next.
Fromme is working with her collaborators at DESY to build a much smaller version of a free-electron X-ray laser. This one will operate on even shorter time frames, at attoseconds (10-18 seconds), and is combined with X-ray and optical spectroscopy to unravel the secret of photosynthesis. The group will build a second compact X-ray source at ASU in collaboration with groups from MIT and DESY.
To aid in that effort, ASU recently hired physicist William Graves from MIT. Graves is in charge of designing and eventually building a compact X-ray free electron laser on the ASU campus that would shrink the two-mile long accelerator at the LCLS to a laboratory-sized device. This would greatly reduce the price of such a machine and its operating costs.
If successful, ASU will be the first public university with a compact “synchrotron” and attosecond free-electron laser.