In vivo protein crystallization in combination with highly brilliant radiation sources offers novel opportunities for the structural analysis of post-translationally modified eukaryotic proteins

Convolutional neural network approach for the automated identification of in cellulo crystals

In cellulo crystallization is a rare event in nature. Recent advances that have made use of heterologous overexpression can promote the intracellular formation of protein crystals, but new tools are required to detect and characterize these targets in the complex cell environment. The present work makes use of Mask R-CNN, a convolutional neural network (CNN)-based instance segmentation method, for the identification of either single or multi-shaped crystals growing in living insect cells, using conventional bright field images. The algorithm can be rapidly adapted to recognize different targets, with the aim of extracting relevant information to support a semi-automated screening pipeline, in order to aid the development of the intracellular protein crystallization approach.

Fixed-target serial femtosecond crystallography using in cellulo grown microcrystals

The crystallization of recombinant proteins in living cells is an exciting new approach in structural biology. Recent success has highlighted the need for fast and efficient diffraction data collection, optimally directly exposing intact crystal-containing cells to the X-ray beam, thus protecting the in cellulo crystals from environmental challenges. Serial femtosecond crystallography (SFX) at free-electron lasers (XFELs) allows the collection of detectable diffraction even from tiny protein crystals, but requires very fast sample exchange to utilize each XFEL pulse. Here, an efficient approach is presented for high-resolution structure elucidation using serial femtosecond in cellulo diffraction of micometre-sized crystals of the protein HEX-1 from the fungus Neurospora crassa on a fixed target. Employing the fast and highly accurate Roadrunner II translation-stage system allowed efficient raster scanning of the pores of micro-patterned, single-crystalline silicon chips loaded with living, crystal-containing insect cells. Compared with liquid-jet and LCP injection systems, the increased hit rates of up to 30% and reduced background scattering enabled elucidation of the HEX-1 structure. Using diffraction data from only a single chip collected within 12 min at the Linac Coherent Light Source, a 1.8 Å resolution structure was obtained with significantly reduced sample consumption compared with previous SFX experiments using liquid-jet injection. This HEX-1 structure is almost superimposable with that previously determined using synchrotron radiation from single HEX-1 crystals grown by sitting-drop vapour diffusion, validating the approach. This study demonstrates that fixed-target SFX using micro-patterned silicon chips is ideally suited for efficient in cellulo diffraction data collection using living, crystal-containing cells, and offers huge potential for the straightforward structure elucidation of proteins that form intracellular crystals at both XFELs and synchrotron sources.

Rapid screening of in cellulo grown protein crystals via a small-angle X-ray scattering/X-ray powder diffraction synergistic approach

A rapid and sensitive detection approach utilizing high-brilliance and low-background small-angle X-ray scattering and X-ray powder diffraction to detect protein microcrystals grown within living insect cells is described.

Crystallization of recombinant proteins in living cells is an exciting new approach for structural biology that provides an alternative to the time-consuming optimization of protein purification and extensive crystal screening steps. Exploiting the potential of this approach requires a more detailed understanding of the cellular processes involved and versatile screening strategies for crystals in a cell culture. Particularly if the target protein forms crystalline structures of unknown morphology only in a small fraction of cells, their detection by applying standard visualization techniques can be time consuming and difficult owing to the environmental challenges imposed by the living cells. In this study, a high-brilliance and low-background bioSAXS beamline is employed for rapid and sensitive detection of protein microcrystals grown within insect cells. On the basis of the presence of Bragg peaks in the recorded small-angle X-ray scattering profiles, it is possible to assess within seconds whether a cell culture contains microcrystals, even in a small percentage of cells. Since such information cannot be obtained by other established detection methods in this time frame, this screening approach has the potential to overcome one of the bottlenecks of intracellular crystal detection. Moreover, the association of the Bragg peak positions in the scattering curves with the unit-cell composition of the protein crystals raises the possibility of investigating the impact of environmental conditions on the crystal structure of the intracellular protein crystals. This information provides valuable insights helping to further understand the in cellulo crystallization process.

Construction of gateway-compatible baculovirus expression vectors for high-throughput protein expression and in vivo microcrystal screening

Baculovirus mediated-insect cell expression systems have been widely used for producing heterogeneous proteins. However, to date, there is still the lack of an easy-to-manipulate system that enables the high-throughput protein characterization in insect cells by taking advantage of large existing Gateway clone libraries. To resolve this limitation, we have constructed a suite of Gateway-compatible pIEx-derived baculovirus expression vectors that allow the rapid and cost-effective construction of expression clones for mass parallel protein expression in insect cells. This vector collection also supports the attachment of a variety of fusion tags to target proteins to meet the needs for different research applications. We first demonstrated the utility of these vectors for protein expression and purification using a set of 40 target proteins of various sizes, cellular localizations and host organisms. We then established a scalable pipeline coupled with the SONICC and TEM techniques to screen for microcrystal formation within living insect cells. Using this pipeline, we successfully identified microcrystals for ~ 16% of the tested protein set, which can be potentially used for structure elucidation by X-ray crystallography. In summary, we have established a versatile pipeline enabling parallel gene cloning, protein expression and purification, and in vivo microcrystal screening for structural studies.

Enhanced X-ray diffraction of in vivo-grown μNS crystals by viscous jets at XFELs

μNS is a 70 kDa major nonstructural protein of avian reoviruses, which cause significant economic losses in the poultry industry. They replicate inside viral factories in host cells, and the μNS protein has been suggested to be the minimal viral factor required for factory formation. Thus, determining the structure of μNS is of great importance for understanding its role in viral infection. In the study presented here, a fragment consisting of residues 448-605 of μNS was expressed as an EGFP fusion protein in Sf9 insect cells. EGFP-μNS(448-605) crystallization in Sf9 cells was monitored and verified by several imaging techniques. Cells infected with the EGFP-μNS(448-605) baculovirus formed rod-shaped microcrystals (5-15 µm in length) which were reconstituted in high-viscosity media (LCP and agarose) and investigated by serial femtosecond X-ray diffraction using viscous jets at an X-ray free-electron laser (XFEL). The crystals diffracted to 4.5 Å resolution. A total of 4227 diffraction snapshots were successfully indexed into a hexagonal lattice with unit-cell parameters a = 109.29, b = 110.29, c = 324.97 Å. The final data set was merged and refined to 7.0 Å resolution. Preliminary electron-density maps were obtained. While more diffraction data are required to solve the structure of μNS(448-605), the current experimental strategy, which couples high-viscosity crystal delivery at an XFEL with in cellulo crystallization, paves the way towards structure determination of the μNS protein.

What is the structural chemistry of the living organism at its temperature and pressure?

The three probes of the structure of matter (X-rays, neutrons and electrons) in biology have complementary properties and strengths. The balance between these three probes within their strengths and weaknesses is perceived to change, even dramatically so at times. For the study of combined states of order and disorder, NMR crystallography is also applicable. Of course, to understand biological systems the required perspectives are surely physiologically relevant temperatures and relevant chemical conditions, as well as a minimal perturbation owing to the needs of the probe itself. These remain very tough challenges because, for example, cryoEM by its very nature will never be performed at room temperature, crystallization often requires nonphysiological chemical conditions, and X-rays and electrons cause beam damage. However, integrated structural biology techniques and functional assays provide a package towards physiological relevance of any given study. Reporting of protein crystal structures, and their associated database entries, could usefully indicate how close to the biological situation they are, as discussed in detail in this feature article.

In cellulo crystallization of Trypanosoma brucei IMP dehydrogenase enables the identification of genuine co-factors

Sleeping sickness is a fatal disease caused by the protozoan parasite Trypanosoma brucei (Tb). Inosine-5’-monophosphate dehydrogenase (IMPDH) has been proposed as a potential drug target, since it maintains the balance between guanylate deoxynucleotide and ribonucleotide levels that is pivotal for the parasite. Here we report the structure of TbIMPDH at room temperature utilizing free-electron laser radiation on crystals grown in living insect cells. The 2.80 Å resolution structure reveals the presence of ATP and GMP at the canonical sites of the Bateman domains, the latter in a so far unknown coordination mode. Consistent with previously reported IMPDH complexes harboring guanosine nucleotides at the second canonical site, TbIMPDH forms a compact oligomer structure, supporting a nucleotide-controlled conformational switch that allosterically modulates the catalytic activity. The oligomeric TbIMPDH structure we present here reveals the potential of in cellulo crystallization to identify genuine allosteric co-factors from a natural reservoir of specific compounds.

Trypanosoma brucei inosine-5′-monophosphate dehydrogenase (IMPDH) is an enzyme in the guanine nucleotide biosynthesis pathway and of interest as a drug target. Here the authors present the 2.8 Å room temperature structure of TbIMPDH determined by utilizing X-ray free-electron laser radiation and crystals that were grown in insect cells and find that ATP and GMP are bound at the canonical sites of the Bateman domains.

Data-driven challenges and opportunities in crystallography

Structural biology is in the midst of a revolution fueled by faster and more powerful instruments capable of delivering orders of magnitude more data than their predecessors. This increased pace in data gathering introduces new experimental and computational challenges, frustrating real-time processing and interpretation of data and requiring long-term solutions for data archival and retrieval. This combination of challenges and opportunities is driving the exploration of new areas of structural biology, including studies of macromolecular dynamics and the investigation of molecular ensembles in search of a better understanding of conformational landscapes. The next generation of instruments promises to yield even greater data rates, requiring a concerted effort by institutions, centers and individuals to extract meaning from every bit and make data accessible to the community at large, facilitating data mining efforts by individuals or groups as analysis tools improve.

Protein crystallization in living cells

Protein crystallization in living cells has been observed surprisingly often as a native assembly process during the past decades, and emerging evidence indicates that this phenomenon is also accessible for recombinant proteins. But only recently the advent of high-brilliance synchrotron sources, X-ray free-electron lasers, and improved serial data collection strategies has allowed the use of these micrometer-sized crystals for structural biology. Thus, in cellulo crystallization could offer exciting new possibilities for proteins that do not crystallize applying conventional approaches. In this review, we comprehensively summarize the current knowledge of intracellular protein crystallization. This includes an overview of the cellular functions, the physical properties, and, if known, the mode of regulation of native in cellulo crystal formation, complemented with a discussion of the reported crystallization events of recombinant proteins and the current method developments to successfully collect X-ray diffraction data from in cellulo crystals. Although the intracellular protein self-assembly mechanisms are still poorly understood, regulatory differences between native in cellulo crystallization linked to a specific function and accidently crystallizing proteins, either disease associated or recombinantly introduced, become evident. These insights are important to systematically exploit living cells as protein crystallization chambers in the future.

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.

Design of a confined environment using protein cages and crystals for the development of biohybrid materials

There is growing interest in the design of protein assemblies for use in materials science and bionanotechnology. Protein assemblies, such as cages and crystalline protein structures, provide confined chemical environments that allow immobilization of metal complexes, nanomaterials, and proteins by metal coordination, assembly/disassembly reactions, genetic manipulation and crystallization methods. Protein assembly composites can be used to prepare hybrid materials with catalytic, magnetic and optical properties for cellular applications due to their high stability, solubility and biocompatibility. In this feature article, we focus on the recent development of ferritin as the most promising molecular template protein cage and in vivo and in vitro engineering of protein crystals as solid protein materials with functional properties.

Microcrystallography of Protein Crystals and In Cellulo Diffraction

The advent of high-quality microfocus beamlines at many synchrotron facilities has permitted the routine analysis of crystals smaller than 10 µm in their largest dimension, which used to represent a challenge. We present two alternative workflows for the structure determination of protein microcrystals by X-ray crystallography with a particular focus on crystals grown in vivo. The microcrystals are either extracted from cells by sonication and purified by differential centrifugation, or analyzed in cellulo after cell sorting by flow cytometry of crystal-containing cells. Optionally, purified crystals or crystal-containing cells are soaked in heavy atom solutions for experimental phasing. These samples are then prepared for diffraction experiments in a similar way by application onto a micromesh support and flash cooling in liquid nitrogen. We briefly describe and compare serial diffraction experiments of isolated microcrystals and crystal-containing cells using a microfocus synchrotron beamline to produce datasets suitable for phasing, model building and refinement. These workflows are exemplified with crystals of the Bombyx mori cypovirus 1 (BmCPV1) polyhedrin produced by infection of insect cells with a recombinant baculovirus. In this case study, in cellulo analysis is more efficient than analysis of purified crystals and yields a structure in ~8 days from expression to refinement.

A pipeline for structure determination of in vivo-grown crystals using in cellulo diffraction

While structure determination from micrometre-sized crystals used to represent a challenge, serial X-ray crystallography on microfocus beamlines at synchrotron and free-electron laser facilities greatly facilitates this process today for microcrystals and nanocrystals. In addition to typical microcrystals of purified recombinant protein, these advances have enabled the analysis of microcrystals produced inside living cells. Here, a pipeline where crystals are grown in insect cells, sorted by flow cytometry and directly analysed by X-ray diffraction is presented and applied to in vivo-grown crystals of the recombinant CPV1 polyhedrin. When compared with the analysis of purified crystals, in cellulo diffraction produces data of better quality and a gain of ∼0.35 Å in resolution for comparable beamtime usage. Importantly, crystals within cells are readily derivatized with gold and iodine compounds through the cellular membrane. Using the multiple isomorphous replacement method, a near-complete model was autobuilt from 2.7 Å resolution data. Thus, in favourable cases, an in cellulo pipeline can replace the complete workflow of structure determination without compromising the quality of the resulting model. In addition to its efficiency, this approach maintains the protein in a cellular context throughout the analysis, which reduces the risk of disrupting transient or labile interactions in protein-protein or protein-ligand complexes.