Research‎ > ‎

Diffraction Imaging

(Major Center-Driven Research Project)

LLNL: HN Chapman, S Marchesini, A Barty, MJ Bogan, S Bajt, U Rohner, M Frank, WH Benner, B Woods, R London, S Hau-Riege; SLAC: S Boutet, J Hajdu; 

LBNL: D Shapiro; ASU: U Weierstall, D Starodub, M Hunter, D Deponte, R Kirian, G Hembree, 
RB Doak, K Schmidt, P Fromme, JCH Spence

Virtually all the known information about the molecular structure of biomolecules has been obtained by x-ray crystallography. Unfortunately, only 2% of the human proteome structures have yet been determined because of the extreme difficulty in producing high-quality crystals. Worse yet, while most drugs target membrane proteins, these proteins are exceedingly difficult to crystallize. Only 80 of the 47,000 or so known protein structures are membrane proteins.

It was suggested by Janos Hajdu (University of Uppsula and CBST member) that an x-ray source of sufficient brightness and with a sufficiently short pulse could be used to produce a diffraction pattern using scattered x-rays from a single molecule before it is destroyed by the Coulomb explosion following rapid photoionization. The required brightness of ~1033 photons / (s ⋅ mm-2 ⋅ mrad2 ⋅ (0.1% bandwidth)), x-ray energy of 5 keV, and pulse width of 10 femtoseconds are expected to be achieved by the Linac Coherent Light Source (LCLS) 4th generation x-ray source funded by DOE and scheduled to be in operation by 2009. Successful implementation of this approach would represent a quantum leap in structural biology and proteomics research. The goal of this project is, as a member of the international team, address the key challenges of recording ultrafast single shot coherent diffraction patterns of injected particles with low noise, and developing robust image reconstruction algorithms.

A second approach to single molecule diffraction imaging is simultaneously being pursued in a collaboration of Center researchers at LLNL, Arizona State University, and LBNL. Here, intense polarized laser light will be used to orient single biomolecules. This work will not only benefit the LCLS approach by learning how to align molecules, but could also lead to a near-term approach to crystal-less molecular structure determinations called serial crystallography. Serial crystallography uses an established method to produce a stream of ice pellets, each containing an identical biomolecule (or alternatively a larger particle such as a single virus). The molecules are then laser-aligned and passed through a synchrotron beam. X-rays diffracted from the individual molecules will be collected and after contributions from ~104 to 106 molecules have accumulated, a diffraction pattern with sufficient signal-to-noise to make structure determinations will have been acquired. If sufficient alignment can be obtained, this method will lead to a new high-resolution method for obtaining structure information for proteins, protein complexes, and larger biological samples such as viral particles. Thus the goal is to demonstrate the feasibility of single molecule structure determinations using x-ray diffraction measurements on existing sychrotron sources.

Using the free-electron laser, FLASH, at Deutsches Elektronen-Synchrotron (DESY) in Hamburg, scientists co-sponsored by CBST, as part of an international collaboration led by CBST members Henry Chapman and Janos Hajdu of Uppsala University, a demonstration of nanometer-resolution X-ray diffraction imaging of freely moving particles in a vacuum was performed. In these experiments, 30 nm particles were injected into the beam line of the FLASH free-electron X-ray laser where <30 fs X-ray pulses produced diffraction patterns from individual particles which were reflected by a multilayer X-ray mirror and recorded on a CCD camera.  Real images were recovered from the diffraction pattern using iterative techniques developed by a former CBST post doctoral researcher. 

A second set of experiments at the FLASH facility demonstrated femtosecond time-delay X-ray holography. Here a single X-ray pulse (<30 fs) illuminated a single polystyrene bead mounted on a 20-nm thick silicon nitride membrane. Some of the X-rays are scattered from the bead and are reflected by a multilayer X-ray mirror back towards the bead where additional X-rays are scattered. X-rays scattered from the first pass interfere with X-rays from the second pass and are then reflected by the detector mirror to record the resultant X-ray hologram on a CCD camera (see figure). The time delay, which can be controlled to fs accuracy, is used to record the motion of the exploding bead.

In the area of serial crystallography, the viability of a flow-focusing nozzle for use in the serial crystallography of proteins has been investigated. The procedure involves sending a beam of aligned water-encapsulated protein molecules to intercept a beam of either X-rays or electrons, and recording the patterns over a large time. The viability of the nozzle
was determined by measuring the size of the produced jet as well as the ability to see diffraction data from species in a jet produced by the nozzle. The jet size was determined to be less than 5 microns, although highly dependent on the nozzle, using an environmental scanning electron microscope. The diffraction patterns of Photosystem I crystals and gold nanoparticles, ran through the jet, were collected using an X-ray source with wavelength of 2.34 nm. Definitive evidence of flow alignment was determined from the asymmetric diffraction pattern of gold nanorods in solution. The data collected showed that the flow-focusing nozzle could be used for the serial crystallography experiment, and also showed that serial crystallography has the potential to compliment X-ray diffraction of proteins for structure determination.