Towards protein-crystal centering using second-harmonic generation (SHG) microscopy

Deep learning applications in protein crystallography

Deep learning techniques can recognize complex patterns in noisy, multidimensional data. In recent years, researchers have started to explore the potential of deep learning in the field of structural biology, including protein crystallography. This field has some significant challenges, in particular producing high-quality and well ordered protein crystals. Additionally, collecting diffraction data with high completeness and quality, and determining and refining protein structures can be problematic. Protein crystallographic data are often high-dimensional, noisy and incomplete. Deep learning algorithms can extract relevant features from these data and learn to recognize patterns, which can improve the success rate of crystallization and the quality of crystal structures. This paper reviews progress in this field.

Advancements in macromolecular crystallography: from past to present

Protein Crystallography or Macromolecular Crystallography (MX) started as a new discipline of science with the pioneering work on the determination of the protein crystal structures by John Kendrew in 1958 and Max Perutz in 1960. The incredible achievements in MX are attributed to the development of advanced tools, methodologies, and automation in every aspect of the structure determination process, which have reduced the time required for solving protein structures from years to a few days, as evident from the tens of thousands of crystal structures of macromolecules available in PDB. The advent of brilliant synchrotron sources, fast detectors, and novel sample delivery methods has shifted the paradigm from static structures to understanding the dynamic picture of macromolecules; further propelled by X-ray Free Electron Lasers (XFELs) that explore the femtosecond regime. The revival of the Laue diffraction has also enabled the understanding of macromolecules through time-resolved crystallography. In this review, we present some of the astonishing method-related and technological advancements that have contributed to the progress of MX. Even with the rapid evolution of several methods for structure determination, the developments in MX will keep this technique relevant and it will continue to play a pivotal role in gaining unprecedented atomic-level details as well as revealing the dynamics of biological macromolecules. With many exciting developments awaiting in the upcoming years, MX has the potential to contribute significantly to the growth of modern biology by unraveling the mechanisms of complex biological processes as well as impacting the area of drug designing.

Low-dose in situ prelocation of protein microcrystals by 2D X-ray phase-contrast imaging for serial crystallography

Serial protein crystallography has emerged as a powerful method of data collection on small crystals from challenging targets, such as membrane proteins. Multiple microcrystals need to be located on large and often flat mounts while exposing them to an X-ray dose that is as low as possible. A crystal-prelocation method is demonstrated here using low-dose 2D full-field propagation-based X-ray phase-contrast imaging at the X-ray imaging beamline TOMCAT at the Swiss Light Source (SLS). This imaging step provides microcrystal coordinates for automated serial data collection at a microfocus macromolecular crystallography beamline on samples with an essentially flat geometry. This prelocation method was applied to microcrystals of a soluble protein and a membrane protein, grown in a commonly used double-sandwich in situ crystallization plate. The inner sandwiches of thin plastic film enclosing the microcrystals in lipid cubic phase were flash cooled and imaged at TOMCAT. Based on the obtained crystal coordinates, both still and rotation wedge serial data were collected automatically at the SLS PXI beamline, yielding in both cases a high indexing rate. This workflow can be easily implemented at many synchrotron facilities using existing equipment, or potentially integrated as an online technique in the next-generation macromolecular crystallography beamline, and thus benefit a number of dose-sensitive challenging protein targets.

Visualization of protein crystals by high-energy phase-contrast X-ray imaging

High-energy phase-contrast X-ray microscopy and tomography of protein crystals in an optically opaque matrix is demonstrated with micrometre resolution on the macromolecular crystallography beamline P14 at PETRA III.

For the extraction of the best possible X-ray diffraction data from macromolecular crystals, accurate positioning of the crystals with respect to the X-ray beam is crucial. In addition, information about the shape and internal defects of crystals allows the optimization of data-collection strategies. Here, it is demonstrated that the X-ray beam available on the macromolecular crystallo­graphy beamline P14 at the high-brilliance synchrotron-radiation source PETRA III at DESY, Hamburg, Germany can be used for high-energy phase-contrast microtomography of protein crystals mounted in an optically opaque lipidic cubic phase matrix. Three-dimensional tomograms have been obtained at X-ray doses that are substantially smaller and on time scales that are substantially shorter than those used for diffraction-scanning approaches that display protein crystals at micrometre resolution. Adding a compound refractive lens as an objective to the imaging setup, two-dimensional imaging at sub-micrometre resolution has been achieved. All experiments were performed on a standard macromolecular crystallography beamline and are compatible with standard diffraction data-collection workflows and apparatus. Phase-contrast X-ray imaging of macromolecular crystals could find wide application at existing and upcoming low-emittance synchrotron-radiation sources.

DeepCentering: fully automated crystal centering using deep learning for macromolecular crystallography

High-throughput protein crystallography using a synchrotron light source is an important method used in drug discovery. Beamline components for automated experiments including automatic sample changers have been utilized to accelerate the measurement of a number of macromolecular crystals. However, unlike cryo-loop centering, crystal centering involving automated crystal detection is a difficult process to automate fully. Here, DeepCentering, a new automated crystal centering system, is presented. DeepCentering works using a convolutional neural network, which is a deep learning operation. This system achieves fully automated accurate crystal centering without using X-ray irradiation of crystals, and can be used for fully automated data collection in high-throughput macromolecular crystallography.

The effects of solution additives and gas-phase modifiers on the molecular environment and conformational space of common heme proteins

RATIONALE

The molecular environment is known to impact the secondary and tertiary structures of biomolecules both in solution and in the gas phase, shifting the equilibrium between different conformational and oligomerization states. However, there is a lack of studies monitoring the impacts of solution additives and gas-phase modifiers on biomolecules characterized using ion mobility techniques.

METHODS

The effect of solution additives and gas-phase modifiers on the molecular environment of two common heme proteins, bovine cytochrome c and equine myoglobin, is investigated as a function of the time after desolvation (e.g., 100-500 ms) using nanoelectrospray ionization coupled to trapped ion mobility spectrometry with detection by time-of-flight mass spectrometry. Organic compounds used as additives/modifiers (methanol, acetonitrile, acetone) were either added to the aqueous protein solution before ionization or added to the ion mobility bath gas by nebulization.

RESULTS

Changes in the mobility profiles are observed depending on the starting solution composition (i.e., in aqueous solution at neutral pH or in the presence of organic content: methanol, acetone, or acetonitrile) and the protein. In the presence of gas-phase modifiers (i.e., N₂ doped with methanol, acetone, or acetonitrile), a shift in the mobility profiles driven by the gas-modifier mass and size and changes in the relative abundances and number of IMS bands are observed.

CONCLUSIONS

We attribute the observed changes in the mobility profiles in the presence of gas-phase modifiers to a clustering/declustering mechanism by which organic molecules adsorb to the protein ion surface and lower energetic barriers for interconversion between conformational states, thus redefining the free energy landscape and equilibria between conformers. These structural biology experiments open new avenues for manipulation and interrogation of biomolecules in the gas phase with the potential to emulate a large suite of solution conditions, ultimately including conditions that more accurately reflect a variety of intracellular environments.

Strategies for sample delivery for femtosecond crystallography

Highly efficient data-collection methods are required for successful macromolecular crystallography (MX) experiments at X-ray free-electron lasers (XFELs). XFEL beamtime is scarce, and the high peak brightness of each XFEL pulse destroys the exposed crystal volume. It is therefore necessary to combine diffraction images from a large number of crystals (hundreds to hundreds of thousands) to obtain a final data set, bringing about sample-refreshment challenges that have previously been unknown to the MX synchrotron community. In view of this experimental complexity, a number of sample delivery methods have emerged, each with specific requirements, drawbacks and advantages. To provide useful selection criteria for future experiments, this review summarizes the currently available sample delivery methods, emphasising the basic principles and the specific sample requirements. Two main approaches to sample delivery are first covered: (i) injector methods with liquid or viscous media and (ii) fixed-target methods using large crystals or using microcrystals inside multi-crystal holders or chips. Additionally, hybrid methods such as acoustic droplet ejection and crystal extraction are covered, which combine the advantages of both fixed-target and injector approaches.

Morphology and structure of ZIF-8 during crystallisation measured by dynamic angle-resolved second harmonic scattering

Recent developments in nonlinear optical light scattering techniques have opened a window into morphological and structural characteristics for a variety of supramolecular systems. However, for the study of dynamic processes, the current way of measuring is often too slow. Here we present an alternative measurement scheme suitable for following dynamic processes. Fast acquisition times are achieved through Fourier imaging, allowing simultaneous detection at multiple scattering angles for different polarization combinations. This allows us to follow the crystal growth of the metal organic framework ZIF-8 in solution. The angle dependence of the signal provides insight into the growth mechanism by probing the evolution of size, shape and concentration, while polarization analysis yields structural information in terms of point group symmetry. Our findings highlight the potential of dynamic angle-resolved harmonic light scattering to probe crystal growth processes, assembly–disassembly of biological systems, adsorption, transport through membranes and myriad other applications.

Angle-resolved monitoring of structure parameters during crystal growth is often slow owing to mechanical movements. Here, the authors use second harmonic scattering and Fourier-plane imaging to dynamically monitor size, shape and concentration of ZIF-8 in situ during the growth process.

The complex analysis of X-ray mesh scans for macromolecular crystallography

A method and software program, MeshBest, for the detection of individual crystals based on two-dimensional X-ray mesh scans are presented.

In macromolecular crystallography, mesh (raster) scans are carried out either as part of X-ray-based crystal-centring routines or to identify positions on the sample holder from which diffraction images can be collected. Here, the methods used in MeshBest, software which automatically analyses diffraction images collected during a mesh scan and produces a two-dimensional crystal map showing estimates of the dimensions, centre positions and diffraction qualities of each crystal contained in the mesh area, are presented. Sample regions producing diffraction images resulting from the superposition of more than one crystal are also distinguished from regions with single-crystal diffraction. The applicability of the method is demonstrated using several cases.

Locating and Visualizing Crystals for X-Ray Diffraction Experiments

Macromolecular crystallography has advanced from using macroscopic crystals, which might be >1 mm on a side, to crystals that are essentially invisible to the naked eye, or even under a standard laboratory microscope. As crystallography requires recognizing crystals when they are produced, and then placing them in an X-ray, electron, or neutron beam, this provides challenges, particularly in the case of advanced X-ray sources, where beams have very small cross sections and crystals may be vanishingly small. Methods for visualizing crystals are reviewed here, and examples of different types of cases are presented, including: standard crystals, crystals grown in mesophase, in situ crystallography, and crystals grown for X-ray Free Electron Laser or Micro Electron Diffraction experiments. As most techniques have limitations, it is desirable to have a range of complementary techniques available to identify and locate crystals. Ideally, a given technique should not cause sample damage, but sometimes it is necessary to use techniques where damage can only be minimized. For extreme circumstances, the act of probing location may be coincident with collecting X-ray diffraction data. Future challenges and directions are also discussed.

Crystallization of Membrane Proteins: An Overview

Membrane proteins are crucial components of cellular membranes and are responsible for a variety of physiological functions. The advent of new tools and technologies for structural biology of membrane proteins has led to a significant increase in the number of structures deposited to the Protein Data Bank during the past decade. This new knowledge has expanded our fundamental understanding of their mechanism of function and contributed to the drug-design efforts. In this chapter we discuss current approaches for membrane protein expression, solubilization, crystallization, and data collection. Additionally, we describe the protein quality-control assays that are often instrumental as a guideline for a shorter path toward the structure.

Automatic crystal centring procedure at the SSRF macromolecular crystallography beamline

X-ray diffraction is a common technique for determining crystal structures. The average time needed for the solution of a protein structure has been drastically reduced by a number of recent experimental and theoretical developments. Since high-throughput protein crystallography benefits from full automation of all steps that are carried out on a synchrotron beamline, an automatic crystal centring procedure is important for crystallographic beamlines. Fully automatic crystal alignment involves the application of optical methods to identify the crystal and move it onto the rotation axis and into the X-ray beam. Crystal recognition has complex dependencies on the illumination, crystal size and viewing angles due to effects such as local shading, inter-reflections and the presence of antifreezing elements. Here, a rapid procedure for crystal centring with multiple cameras using region segment thresholding is reported. Firstly, a simple illumination-invariant loop recognition and classification model is used by slicing a low-magnification loop image into small region segments, then classifying the loop into different types and aligning it to the beam position using feature vectors of the region segments. Secondly, an edge detection algorithm is used to find the crystal sample in a high-magnification image using region segment thresholding. Results show that this crystal centring method is extremely successful under fluctuating light states as well as for poorly frozen and opaque samples. Moreover, this crystal centring procedure is successfully integrated into the enhanced Blu-Ice data collection system at beamline BL17U1 at the Shanghai Synchrotron Radiation Facility as a routine method for an automatic crystal screening procedure.

Second harmonic generation correlation spectroscopy for characterizing translationally diffusing protein nanocrystals

Second harmonic generation correlation spectroscopy (SHG-CS) is demonstrated as a new approach to protein nanocrystal characterization. A novel line-scanning approach was performed to enable autocorrelation analysis without sample damage from the intense incident beam. An analytical model for autocorrelation was developed, which includes a correction for the optical scattering forces arising when focusing intense, infrared beams. SHG-CS was applied to the analysis of BaTiO3 nanoparticles ranging from 200 to ∼500 nm and of photosystem I nanocrystals. A size distribution was recovered for each sample and compared with the size histogram measured by scanning electron microscopy (SEM). Good agreement was observed between the two independent measurements. The intrinsic selectivity of the second-order nonlinear optical process provides SHG-CS with the ability to distinguish well ordered nanocrystals from conglomerates and amorphous aggregates. Combining the recovered distribution of particle diameters with the histogram of measured SHG intensities provides the inherent hyperpolarizability per unit volume of the SHG-active nanoparticles. Simulations suggest that the SHG activity per unit volume is likely to exhibit relatively low sensitivity to the subtle distortions within the lattice that contribute to resolution loss in X-ray diffraction, but high sensitivity to the presence of multi-domain crystals.

Exploiting Microbeams for Membrane Protein Structure Determination

A reproducible, and sample independent means of predictably obtaining large, well-ordered crystals has proven elusive in macromolecular crystallography. In the structure determination pipeline, crystallisation often proves to be a rate-limiting step, and the process of obtaining even small or badly ordered crystals can prove time-consuming and laborious. This is particularly true in the field of membrane protein crystallography and this is reflected in the limited number of unique membrane protein structures deposited in the protein data bank (less than 650 by June 2016 - http://blanco.biomol.uci.edu/mpstruc ). Over recent years the requirement for, and time and cost associated with obtaining, large crystals has been partially alleviated through the development of beamline instrumentation allowing data collection, and structure solution, from ever-smaller crystals. Advances in several areas have led to a step change in what might be considered achievable during a synchrotron trip over the last decade. This chapter will briefly review the current status of the field, the tools available to ease data collection and processing, and give some examples of exploitation of these for membrane protein microfocus macromolecular crystallography.

A visible-light-excited fluorescence method for imaging protein crystals without added dyes

Fluorescence microscopy methods have seen an increase in popularity in recent years for detecting protein crystals in screening trays. The fluorescence-based crystal detection methods have thus far relied on intrinsic UV-inducible tryptophan fluorescence, nonlinear optics or fluorescence in the visible light range dependent on crystals soaked with fluorescent dyes. In this paper data are presented on a novel visible-light-inducible autofluorescence arising from protein crystals as a result of general stabilization of conjugated double-bond systems and increased charge delocalization due to crystal packing. The visible-light-inducible autofluorescence serves as a complementary method to bright-field microscopy in beamline applications where accurate crystal centering about the rotation axis is essential. Owing to temperature-dependent chromophore stabilization, protein crystals exhibit tenfold higher fluorescence intensity at cryogenic temperatures, making the method ideal for experiments where crystals are cooled to 100 K with a cryostream. In addition to the non-damaging excitation wavelength and low laser power required for imaging, the method can also serve a useful role for differentiating protein crystals from salt crystals in screening trays.

Intercalating dyes for enhanced contrast in second-harmonic generation imaging of protein crystals

The second-harmonic generation (SHG) activity of protein crystals was found to be enhanced by up to ∼1000-fold by the intercalation of SHG phores within the crystal lattice. Unlike the intercalation of fluorophores, the SHG phores produced no significant background SHG from solvated dye or from dye intercalated into amorphous aggregates. The polarization-dependent SHG is consistent with the chromophores adopting the symmetry of the crystal lattice. In addition, the degree of enhancement for different symmetries of dyes is consistent with theoretical predictions based on the molecular nonlinear optical response. Kinetics studies indicate that intercalation arises over a timeframe of several minutes in lysozyme, with detectable enhancements within seconds. These results provide a potential means to increase the overall diversity of protein crystals and crystal sizes amenable to characterization by SHG microscopy.

Interactions of arene ruthenium metallaprisms with human proteins

Interactions between three hexacationic arene ruthenium metallaprisms and human proteins have been studied using NMR spectroscopy, mass spectrometry and circular dichroism spectroscopy, showing that proteins are potential biological targets for these metallaprisms.

Identifying, studying and making good use of macromolecular crystals

As technology advances, the crystal volume that can be used to collect useful X-ray diffraction data decreases. The technologies available to detect and study growing crystals beyond the optical resolution limit and methods to successfully place the crystal into the X-ray beam are discussed.

Structural biology has contributed tremendous knowledge to the understanding of life on the molecular scale. The Protein Data Bank, a depository of this structural knowledge, currently contains over 100 000 protein structures, with the majority stemming from X-ray crystallography. As the name might suggest, crystallography requires crystals. As detectors become more sensitive and X-ray sources more intense, the notion of a crystal is gradually changing from one large enough to embellish expensive jewellery to objects that have external dimensions of the order of the wavelength of visible light. Identifying these crystals is a prerequisite to their study. This paper discusses developments in identifying these crystals during crystallization screening and distinguishing them from other potential outcomes. The practical aspects of ensuring that once a crystal is identified it can then be positioned in the X-ray beam for data collection are also addressed.

Kinetic trapping of metastable amino acid polymorphs

Second harmonic generation (SHG) microscopy measurements indicate that inkjet-printed racemic solutions of amino acids can produce nanocrystals trapped in metastable polymorph forms upon rapid solvent evaporation. Polymorphism impacts the composition, distribution, and physico-kinetic properties of organic solids, with energetic arguments favoring the most stable polymorph. In this study, unfavored noncentrosymmetric crystal forms were observed by SHG microscopy. Polarization-dependent SHG measurement and synchrotron X-ray microdiffraction analysis of individual printed drops are consistent with formation of homochiral crystal production. Fundamentally, these results provide evidence supporting the ubiquity of Ostwald’s Rule of Stages, describing the hypothesized transitioning of crystals between metastable polymorphic forms in the early stages of crystal formation. Practically, the presence of homochiral metastable forms has implications on chiral resolution and on solid form preparations relying on rapid solvent evaporation.

Approaches to automated protein crystal harvesting

The harvesting of protein crystals is almost always a necessary step in the determination of a protein structure using X-ray crystallographic techniques. However, protein crystals are usually fragile and susceptible to damage during the harvesting process. For this reason, protein crystal harvesting is the single step that remains entirely dependent on skilled human intervention. Automation has been implemented in the majority of other stages of the structure-determination pipeline, including cloning, expression, purification, crystallization and data collection. The gap in automation between crystallization and data collection results in a bottleneck in throughput and presents unfortunate opportunities for crystal damage. Several automated protein crystal harvesting systems have been developed, including systems utilizing microcapillaries, microtools, microgrippers, acoustic droplet ejection and optical traps. However, these systems have yet to be commonly deployed in the majority of crystallography laboratories owing to a variety of technical and cost-related issues. Automation of protein crystal harvesting remains essential for harnessing the full benefits of fourth-generation synchrotrons, free-electron lasers and microfocus beamlines. Furthermore, automation of protein crystal harvesting offers several benefits when compared with traditional manual approaches, including the ability to harvest microcrystals, improved flash-cooling procedures and increased throughput.

Integrated nonlinear optical imaging microscope for on-axis crystal detection and centering at a synchrotron beamline

Nonlinear optical (NLO) instrumentation has been integrated with synchrotron X-ray diffraction (XRD) for combined single-platform analysis, initially targeting applications for automated crystal centering. Second-harmonic-generation microscopy and two-photon-excited ultraviolet fluorescence microscopy were evaluated for crystal detection and assessed by X-ray raster scanning. Two optical designs were constructed and characterized; one positioned downstream of the sample and one integrated into the upstream optical path of the diffractometer. Both instruments enabled protein crystal identification with integration times between 80 and 150 µs per pixel, representing a ∼10(3)-10(4)-fold reduction in the per-pixel exposure time relative to X-ray raster scanning. Quantitative centering and analysis of phenylalanine hydroxylase from Chromobacterium violaceum cPAH, Trichinella spiralis deubiquitinating enzyme TsUCH37, human κ-opioid receptor complex kOR-T4L produced in lipidic cubic phase (LCP), intimin prepared in LCP, and α-cellulose samples were performed by collecting multiple NLO images. The crystalline samples were characterized by single-crystal diffraction patterns, while α-cellulose was characterized by fiber diffraction. Good agreement was observed between the sample positions identified by NLO and XRD raster measurements for all samples studied.