Femtosecond Nanocrystallography Research Articles

Fast reconstruction of single-shot wide-angle diffraction images through deep learning

Single-shot x-ray imaging of short-lived nanostructures such as clusters and nanoparticles near a phase transition or non-crystalizing objects such as large proteins and viruses is currently the most elegant method for characterizing their structure. Using hard x-ray radiation provides scattering images that encode two-dimensional projections, which can be combined to identify the full three-dimensional object structure from multiple identical samples. Wide-angle scattering using XUV or soft x-rays, despite yielding lower resolution, provides three-dimensional structural information in a single shot and has opened routes towards the characterization of non-reproducible objects in the gas phase. The retrieval of the structural information contained in wide-angle scattering images is highly non-trivial, and currently no efficient rigorous algorithm is known. Here we show that deep learning networks, trained with simulated scattering data, allow for fast and accurate reconstruction of shape and orientation of nanoparticles from experimental images. The gain in speed compared to conventional retrieval techniques opens the route for automated structure reconstruction algorithms capable of real-time discrimination and pre-identification of nanostructures in scattering experiments with high repetition rate—thus representing the enabling technology for fast femtosecond nanocrystallography.

Nanocrystallography measurements of early stage synthetic malaria pigment.

The recent availability of extremely intense, femtosecond X-ray free-electron laser (XFEL) sources has spurred the development of serial femtosecond nanocrystallography (SFX). Here, SFX is used to analyze nanoscale crystals of β-hematin, the synthetic form of hemozoin which is a waste by-product of the malaria parasite. This analysis reveals significant differences in β-hematin data collected during SFX and synchrotron crystallography experiments. To interpret these differences two possibilities are considered: structural differences between the nanocrystal and larger crystalline forms of β-hematin, and radiation damage. Simulation studies show that structural inhomogeneity appears at present to provide a better fit to the experimental data. If confirmed, these observations will have implications for designing compounds that inhibit hemozoin formation and suggest that, for some systems at least, additional information may be gained by comparing structures obtained from nanocrystals and macroscopic crystals of the same molecule.

Corrugated structure insertion for extending the SASE bandwidth up to 3% at the European XFEL

The usage of x-ray free electron laser (XFEL) in femtosecond nanocrystallography involves sequential illumination of many small crystals of arbitrary orientation. Hence a wide radiation bandwidth will be useful in order to obtain and to index a larger number of Bragg peaks used for determination of the crystal orientation.Considering the baseline configuration of the European XFEL in Hamburg, and based on beam dynamics simulations, we demonstrate here that the usage of corrugated structures allows for a considerable increase in radiation bandwidth. Data collection with a 3% bandwidth, a few microjoule radiation pulse energy, a few femtosecond pulse duration, and a photon energy of 5.4keV is possible.For this study we have developed an analytical modal representation of the short-range wake function of the flat corrugated structures for arbitrary offsets of the source and the witness particles.

Calculation of x-ray scattering patterns from nanocrystals at high x-ray intensity.

We present a generalized method to describe the x-ray scattering intensity of the Bragg spots in a diffraction pattern from nanocrystals exposed to intense x-ray pulses. Our method involves the subdivision of a crystal into smaller units. In order to calculate the dynamics within every unit, we employ a Monte-Carlo-molecular dynamics-ab-initio hybrid framework using real space periodic boundary conditions. By combining all the units, we simulate the diffraction pattern of a crystal larger than the transverse x-ray beam profile, a situation commonly encountered in femtosecond nanocrystallography experiments with focused x-ray free-electron laser radiation. Radiation damage is not spatially uniform and depends on the fluence associated with each specific region inside the crystal. To investigate the effects of uniform and non-uniform fluence distribution, we have used two different spatial beam profiles, Gaussian and flattop.

Protein crystal screening and characterization for serial femtosecond nanocrystallography

The recent development of X-ray free electron lasers (XFELs) has spurred the development of serial femtosecond nanocrystallography (SFX) which, for the first time, is enabling structure retrieval from sub-micron protein crystals. Although there are already a growing number of structures published using SFX, the technology is still very new and presents a number of unique challenges as well as opportunities for structural biologists. One of the biggest barriers to the success of SFX experiments is the preparation and selection of suitable protein crystal samples. Here we outline a protocol for preparing and screening for suitable XFEL targets.

Serial Femtosecond Crystallography Opens a New Era in Membrane Protein Structural Biology

Serial Femtosecond Crystallography (SFX) provides a novel concept for structure determination, where X‐ray diffraction “snapshots” are collected from a fully hydrated stream of nanocrystals, using femtosecond pulses from the world's first high energy X‐ray free‐electron laser, the Linac Coherent Light Source. The large membrane protein complex Photosystem I was used as the model system for the first serial femtosecond crystallography experiments that showed the proof of concept that diffraction of nanocrystals can be observed using femtosecond pulses that are 1012 stronger than 3rd generation synchrotron sources and destroy any material that is placed in its focus [1],[2]. By using femtosecond pulses briefer than the time‐scale of most damage processes, femtosecond nanocrystallography overcomes the problem of X‐ray damage in crystallography [3] and extends to atomic resolution [4],[5]. It is also applied to membrane proteins crystallized in lipidic environments [6],[7], and has recently led to the structure determination of the Angiotensin receptor which regulates the blood sugar level and is one of the major targets for drugs fighting hypertension [8] and the structure of the rhodopsin‐arrestin complex [9]. Femtosecond crystallography also opens a new avenue for determination of protein dynamics. First experiments on the proof of principle for time resolved serial femtosecond nanocrystallography have been performed on Photosystem I‐ferredoxin [10] and Photosystem II nano‐crystals [11] and conformational changes of the Mn4Ca cluster and its protein environment were observed for the first time in the transition from the dark to the double excited state [12]. Very recently we were able to show that TR‐SFX studies extend to atomic resolution using the photoactive yellow protein as a model system [13]. This pioneering work paves the way for the determination of molecular movies of the dynamics of membrane proteins “at work” in the future including the determination of molecular movies of water splitting.

Whole-pattern fitting technique in serial femtosecond nanocrystallography.

Serial femtosecond X-ray crystallography (SFX) has created new opportunities in the field of structural analysis of protein nanocrystals. The intensity and timescale characteristics of the X-ray free-electron laser sources used in SFX experiments necessitate the analysis of a large collection of individual crystals of variable shape and quality to ultimately solve a single, average crystal structure. Ensembles of crystals are commonly encountered in powder diffraction, but serial crystallography is different because each crystal is measured individually and can be oriented via indexing and merged into a three-dimensional data set, as is done for conventional crystallography data. In this way, serial femtosecond crystallography data lie in between conventional crystallography data and powder diffraction data, sharing features of both. The extremely small sizes of nanocrystals, as well as the possible imperfections of their crystallite structure, significantly affect the diffraction pattern and raise the question of how best to extract accurate structure-factor moduli from serial crystallography data. Here it is demonstrated that whole-pattern fitting techniques established for one-dimensional powder diffraction analysis can be feasibly extended to higher dimensions for the analysis of merged SFX diffraction data. It is shown that for very small crystals, whole-pattern fitting methods are more accurate than Monte Carlo integration methods that are currently used.

Expression, purification and crystallization of CTB-MPR, a candidate mucosal vaccine component against HIV-1.

CTB-MPR is a fusion protein between the B subunit of cholera toxin (CTB) and the membrane-proximal region of gp41 (MPR), the transmembrane envelope protein of Human immunodeficiency virus 1 (HIV-1), and has previously been shown to induce the production of anti-HIV-1 antibodies with antiviral functions. To further improve the design of this candidate vaccine, X-ray crystallography experiments were performed to obtain structural information about this fusion protein. Several variants of CTB-MPR were designed, constructed and recombinantly expressed in Escherichia coli. The first variant contained a flexible GPGP linker between CTB and MPR, and yielded crystals that diffracted to a resolution of 2.3 Å, but only the CTB region was detected in the electron-density map. A second variant, in which the CTB was directly attached to MPR, was shown to destabilize pentamer formation. A third construct containing a polyalanine linker between CTB and MPR proved to stabilize the pentameric form of the protein during purification. The purification procedure was shown to produce a homogeneously pure and monodisperse sample for crystallization. Initial crystallization experiments led to pseudo-crystals which were ordered in only two dimensions and were disordered in the third dimension. Nanocrystals obtained using the same precipitant showed promising X-ray diffraction to 5 Å resolution in femtosecond nanocrystallography experiments at the Linac Coherent Light Source at the SLAC National Accelerator Laboratory. The results demonstrate the utility of femtosecond X-ray crystallography to enable structural analysis based on nano/microcrystals of a protein for which no macroscopic crystals ordered in three dimensions have been observed before.

Femtosecond Nanocrystallography and Characterization of Photosystem II

Membrane proteins are extremely difficult to crystallize, however they are highly important proteins for cellular function. Photosystem I, one of the most complex membrane proteins solved to date took more than a decade to have a structure solved to molecular resolution. Large, well-ordered crystal growth is one of the major bottlenecks in structural determination by x-ray crystallography, due to the difficulty of making the "perfect" crystal. The development of femtosecond nanocrystallography, which uses a stream of fully hydrated nanocrystals to collect diffraction snapshots, effectively reduces this bottleneck[1] Photosystem II changed our biosphere via splitting water and evolving oxygen 2.5 billion years ago. Using femtosecond nanocrystallography we are developing a time-resolved femtosecond crystallography method [2] to unravel the mechanism of water splitting by determining the conformational changes that take place during the oxygen evolution process. Multiple crystallization techniques were originally developed in order to make the nanocrystals necessary for femtosecond nanocrystallography. For Photosystem II nano/microcrystals a free interface diffusion method, is used to increase yield over traditional methods. These crystals are then characterized by three different methods before being used for collecting diff raction data. The three methods currently used are optical microscopy, dynamic light scattering (DLS), and Second Order Nonlinear Imaging of Chiral Crystals (SONICC).

Aspects of direct phasing in femtosecond nanocrystallography

X-ray free-electron laser diffraction patterns from protein nanocrystals provide information on the diffracted amplitudes between the Bragg reflections, offering the possibility of direct phase retrieval without the use of ancillary experimental data. Proposals for implementing direct phase retrieval are reviewed. These approaches are limited by the signal-to-noise levels in the data and the presence of different and incomplete unit cells in the nanocrystals. The effects of low signal to noise can be ameliorated by appropriate selection of the intensity data samples that are used. The effects of incomplete unit cells may be small in some cases, and a unique solution is likely if there are four or fewer molecular orientations in the unit cell.

Diffraction by nanocrystals

X-ray femtosecond nanocrystallography is a new, potentially powerful technique for imaging biological macromolecules that uses ensemble-averaged measurements of diffraction of x-ray free-electron laser pulses from nanocrytalline specimens. Nanocrystals have some diffraction characteristics that are distinct from those of macroscopic crystals, due to the presence of different kinds of unit cell in the crystal and of truncated unit cells on the crystal surface. Expressions are derived for diffraction by nanocrystals with variable and incomplete unit cells, averaged over a distribution of crystal sizes and shapes. The diffraction contains differently modulated Bragg components that are due to interference effects within and between the full and incomplete unit cells. Estimates are obtained for the relative magnitudes of the components. The nature of the diffraction is illustrated by two-dimensional simulations. Implications for molecular imaging are discussed.

Crystallization of the Large Membrane Protein Complex Photosystem I in a Microfluidic Channel

Traditional macroscale protein crystallization is accomplished nontrivially by exploring a range of protein concentrations and buffers in solution until a suitable combination is attained. This methodology is time-consuming and resource-intensive, hindering protein structure determination. Even more difficulties arise when crystallizing large membrane protein complexes such as photosystem I (PSI) due to their large unit cells dominated by solvent and complex characteristics that call for even stricter buffer requirements. Structure determination techniques tailored for these "difficult to crystallize" proteins such as femtosecond nanocrystallography are being developed yet still need specific crystal characteristics. Here, we demonstrate a simple and robust method to screen protein crystallization conditions at low ionic strength in a microfluidic device. This is realized in one microfluidic experiment using low sample amounts, unlike traditional methods where each solution condition is set up separately. Second harmonic generation microscopy via second-order nonlinear imaging of chiral crystals (SONICC) was applied for the detection of nanometer- and micrometer-sized PSI crystals within microchannels. To develop a crystallization phase diagram, crystals imaged with SONICC at specific channel locations were correlated to protein and salt concentrations determined by numerical simulations of the time-dependent diffusion process along the channel. Our method demonstrated that a portion of the PSI crystallization phase diagram could be reconstructed in excellent agreement with crystallization conditions determined by traditional methods. We postulate that this approach could be utilized to efficiently study and optimize crystallization conditions for a wide range of proteins that are poorly understood to date.

Dielectrophoretic Sorting of Membrane Protein Nanocrystals

Structure elucidation of large membrane protein complexes is still a considerable challenge, yet is a key factor in drug development and disease combat. Femtosecond nanocrystallography is an emerging technique with which structural information of membrane proteins is obtained without the need to grow large crystals, thus overcoming the experimental riddle faced in traditional crystallography methods. Here, we demonstrate for the first time a microfluidic device capable of sorting membrane protein crystals based on size using dielectrophoresis. We demonstrate the excellent sorting power of this new approach with numerical simulations of selected submicrometer beads in excellent agreement with experimental observations. Crystals from batch crystallization broths of the huge membrane protein complex photosystem I were sorted without further treatment, resulting in a high degree of monodispersity and crystallinity in the ~100 nm size range. Microfluidic integration, continuous sorting, and nanometer-sized crystal fractions make this method ideal for direct coupling to femtosecond nanocrystallography.

Femtosecond nanocrystallography of membrane proteins opens a new era for structural biology

Femtosecond nanocrystallography of membrane proteins opens a new era for structural biology

New Avenues for Structure Determination of Membrane Proteins

Membrane proteins are of extreme relevance for all living cells. They are important for energy supply of the cells, transport across membranes, cell communication and play also vital roles in the pathogenesis of viruses and bacteria. Despite their important function, less than 300 membrane protein structures have been determined to date. One of the major limiting steps for the structure determination is the growth of well-ordered single crystals. New results of a rationale design of membrane protein crystallization are presented, that include screening for the reversibility of crystallization conditions, determination of phase diagrams, and determination of early events in nucleation and crystal growth by a combination of dynamic light scattering, SONICC, and tryptophan fluorescence microscopy. Nanocrystals detected, can either be used as seeds for growth of large single crystals or used for structure determination by femtosecond nanocrystallography. The latter method is based on a novel concept for structure determination, where X-ray diffraction “snapshots” are collected from a fully hydrated stream of nanocrystals, using femtosecond pulses from the world's first high energy X-ray free-electron laser, the Linac Coherent Light Source. The experiments show the proof of concept that diffraction of nanocrystals that contain only 100-10 000 membrane proteins can be observed and used for structure determination of large membrane protein complexes [1]. The femtosecond pulses are 1012 stronger than 3rd generation synchrotron sources and destroy any material that is placed in its focus, but as the femtosecond pulses are briefer than the time-scale of most damage processes, the method overcomes the problem of damage in crystallography. Femtosecond crystallography also opens new avenues for determination of molecular movies of protein dynamics.Reference: [1] Chapman et al, 2011 Nature, 470, 73-77.

Compressive auto-indexing in femtosecond nanocrystallography

Ultrafast nanocrystallography has the potential to revolutionize biology by enabling structural elucidation of proteins for which it is possible to grow crystals with 10 or fewer unit cells on the side. The success of nanocrystallography depends on robust orientation-determination procedures that allow us to average diffraction data from multiple nanocrystals to produce a three-dimensional (3D) diffraction data volume with a high signal-to-noise ratio. Such a 3D diffraction volume can then be phased using standard crystallographic techniques. "Indexing" algorithms used in crystallography enable orientation determination of diffraction data from a single crystal when a relatively large number of reflections are recorded. Here we show that it is possible to obtain the exact lattice geometry from a smaller number of measurements than standard approaches using a basis pursuit solver.