Abstract

The status and prospects of protein microcrystallography (MPX) at high brilliance synchrotron radiation sources are reviewed. We discuss emerging trends in miniaturizing sample environments for serial crystallography (SX) experiments allowing manipulation and positioning of biological objects down to nanoscale dimensions with low contact forces.

Keywords

Synchrotron radiation, X-ray free electron laser, Protein microcrystallography, Serial crystallography, Nanotechnology, Sample environment

Abbreviations

MPX: Protein Microcrystallography; SR: Synchrotron Radiation; SX: Serial Crystallography; SFX: Serial Femtosecond Crystallography; SSX: Synchrotron Serial Crystallography; XFEL: X-ray Free Electron Laser; TMV: Tobacco Mosaic Virus

Introduction

Protein microcrystallography (MPX) at 3^rd generation synchrotron radiation (SR) sources [1, 2] has provided access to structures which are difficult to crystallize such as amyloids [3] or membrane proteins [4, 5]. Technical and scientific progress is dominated by the development of more and more brilliant X-ray sources allowing recording diffraction patterns from increasingly smaller crystals in a shorter time. X-ray free electron laser sources (XFEL), such as the Linac Coherent Light Source (LCLS, Stanford, California, USA) have profoundly modified MPX [6, 7]. Indeed, femtosecond X-ray flashes from XFELs allow obtaining quasi radiation-damage-free diffraction data from nm-sized crystallites in a hitand- destroy mode [7-9]. This has resulted in serial femtosecond crystallography (SFX) techniques based on the merging of patterns due to hits from thousands of crystallites [10]. Indeed, a microjet injector containing photosystem I crystallites down to 200 nm size was used for the first SFX experiment (Figure 1) [9, 11]. For an overview on other SFX flow-injectors and fixed target supports see: [12].The limited availability of XFELs provides, however, a drive for optimizing SR-MPX beam lines and using both types of sources in a complementary way. Indeed, SR-MPX provides currently access to crystallite volumes down to a few μm³ [1, 2, 13] while nm³ volumes have become accessible at XFELs [9]. Radiation-free XFEL data collection is a dream-come-true for protein crystallographers. SRexperiments provide, however, more flexibility for in-situ sample environments such as crystallization plates or microfluidic devices. Special care has, however, to be taken to minimize radiation damage in SR-MPX experiments.

In the following we will first review several emerging trends for SR-MPX with beam sizes from the μm- to the nmrange and then highlight the potential of advanced sample environments. Rather than being exhaustive we will focus on selected on-chip sample environments and low contact-force manipulation derived from nanobiotechnology tools and techniques [14]. The term “MPX” will be used independently of the X-ray beam size.

Sources and Beamlines

A comparison of the brilliance (corresponding to the photon concentration; also called spectral brightness) of 3^rd generation SR sources shows the progress in maximum brilliance for the most recent SR sources (e.g. MAX IV, [15] NSLS II [16]) which are based on advanced magnet lattice designs (Figure 2A). The correlation of focal spot size and maximum flux density of a number of SR-MPX beam lines worldwide reveals that the most advanced SR sources aim for a particularly high (monochromatic) flux density of ~10¹³ photons/(s, μm²) for an ~1 μm² focus (Figure 2B); (FMX_{NSLS II} beamline). The drive towards higher SR brilliances is, however, continuing. Indeed, the brilliance depends on the horizontal and vertical emittance (ε_hor,vert) of the electron beam in a storage ring which is the product of its size (δ_hor,vert) and divergence (δ'_hor,vert). The emittance can be reduced by a compression of the electron beam up to its theoretical limit when electron diffraction effects set in. This limit defines a 4^th generation (“diffraction limited”) SR source [17]. Indeed, the ESRF storage ring, which has started operation in 1992, will be upgraded until 2020 into a close-to 4^th generation SR-source by a change of its magnetic lattice (ESRF II; dashed curve in (Figure 1A) [18]. Several other major 3^rd generation SR sources (e.g. APS and SPring8) will follow this example. This will favor in particular “brilliancehungry” experimental techniques, currently proposed only at few SR beamlines worldwide. Examples are sub-μm SR-MPX [19, 20] (e.g. ID13_ESRF); (Figure 2B) and time-resolved SRMPX based on picoseconds X-ray flashes using an undulator harmonics (“pink” beam), a technique currently developed only for larger SR beam foci [21]. One can estimate that a “pink” SR-MPX beam line would reach at ESRF-II the 10¹⁶ flux density range.

Synchrotron Radiation Serial Crystallography (SSX)

SSX on cryo frozen samples is becoming an automated data collection workflow at SR-MPX beam lines [22]. It has first been demonstrated for Trypanosoma brucei procathepsin B (TbCatB) crystallites of about 2 μm average size using helical raster-rotation scans to obtain “classical” rotation patterns and distribute the radiation dose (Figure 3A) [13]. The hits are displayed in a “heat map” scaled to the overall diffraction strength in each pattern (Figure 3B). The 0.3 nm resolution structure based on 426 patterns from 80 crystallites is of similar quality for the main-chain features as the 0.21 nm resolution SFX study based on 2.9x10⁵ patterns [23].

The statistical treatment of SSX data based on the hierarchical cluster analysis [24] allows readily matching complementary patterns from many different crystallites into groups [25]. A potential target for this approach is exploring experimentally the mosaic block (called here: microdomain) model (Figure 4A) [26]. We note that individual micro-domains in the bulk of a well-ordered protein crystal are not readily accessible to a diffraction experiment. The combination of laser-microdissection and solvent-induced micro-fragmentation allows, however, physically separating a crystal into micro-domains as shown for a lysozyme crystal generated by the Langmuir-Blodgett nanotemplate technique (Figure 4B, 4C) [19, 27]. A raster-scan heat map obtained with a 400 nm beam of a micro-fragmented lysozyme crystal in a cryogenic loop revealed two micro-domains [figure 4D]. Rotation patterns were collected from each micro-domain and their 3D structures refined [19].The results suggest that laser-induced microcavitation and solvent-induced microdomain separation do not introduce lattice distortions. SSX of a sufficiently large set of micro-domains from a microfragmented crystal could possibly provide information on the variability of crystallographic parameters or allow picking out crystalline micro-domains for highly mosaic crystals.

SSX at room temperature is a challenge for SR-MPX due to radiation damage effects although secondary radiation damage becomes negligible for crystallites in the μm-range [28, 29]. A number of room temperature sample environments tested at SR-MPX beam lines have been initially developed for SFX experiments [12, 30-33]. Such experiments are generally performed without crystal rotation. Probing a capillary filled with a lysozyme crystallite flush [34] would, however, not be possible at an XFEL due to an immediate destruction of the capillary walls. The SSX lysozyme structure was refined to 0.21 nm based on about 4x10⁴ patterns using about 250 mg protein. Increasing the flow-density by a liquid crystalline polymer (LCP) phase allows reducing the protein consumption to the 1 mg range. Indeed, an LCP matrix filled with about 30 μm size bacteriorhodopsin crystallites was extruded into air and probed by SSX [35].This membrane protein structure was refined to 0.24 nm resolution. A similar protein consumption has been obtained for randomly distributed crystallites on a fixed target by raster-scan SFX [32]. The same approach was used for SSX with micro- and nanobeams on a slurry of lysozyme crystallites contained between Si₃N₄ membranes (Figure 5A, 5B). The main-chain lysozyme structure was only slightly affected by radiation damage and there was practically no difference for the two beam sizes. The particularly high radiation dose for nanobeam SSX of 29 MGy/crystallite was tentatively attributed to a lag phase [20].The availability of up to KHz framing rate, zero readout noise pixel detectors -such as the EIGER detector from DECTRIS®- will allow further reducing protein consumption in SSX experiments.

On-Chip Sample Environments and Low Contact-Force Manipulation

Sample environments for SX experiments have to cope with the manipulation of μm- to nm-sized biological objects which suggests exploring the use of nanobiotechnology tools and techniques. Furthermore, deterministic positioning of objects in the beam allows reducing the sample consumption and therefore also the data collection time by increasing the crystallites hit rate. The fragility of many proteins, such as membrane proteins, also requires manipulation and positioning with low contact forces. Ideally, self-assembly into crystalline objects and their localization should be done in the same sample environment. The strategies discussed below are likewise of interest for SFX and SSX. A more detailed overview on different positioning & manipulation techniques will be provided elsewhere [36].

Sorting of objects on structured substrates

The aim is to preposition objects on a structured surface so that a high hit rate can be obtained by mesh-scans with a step size corresponding to the separation of the objects. An example for this is provided by a silicon chip with a holepattern etched into a 10 μm membrane by photolithography techniques (Figure 6A) [37]. Evaporating a droplet containing a slush of crystallites on the chip allows immobilizing individual crystallites in the down to 1 μm diameter holes by capillary action (Figure 6B).

Chips with nanofabricated pillared superhydrophobic surfaces are an alternative possibility for exploiting localization of objects during droplet evaporation (Figure 6C, 6D) [38]. The shrinkage of a water or dilute solute droplet during evaporation results in a wetting transition with liquid entering the gaps between the pillars [38]. At higher solute concentrations, a viscous protein layer is formed at the droplet rim by convective-flow mediated mass transport during evaporation. The reduced molecular mobility at the rim can result in self-assembly and the formation of 1D or 2D morphologies [39, 40]. Indeed, a solution of rod-like tobacco mosaic virus (TMV) particles of 300 nm length and 18 nm diameter forms a coffee-ring type layer on the pillars (Figure 7A) [40, 42]. Raster-diffraction with a 170x130 nm² beam reveals micro-domains on and between the pillars with the nanorods axes aligned more or less parallel to the rim and pillar surface (Figure 7B, 7C). Single crystalline features, already noted in a TMV microbeam study [43], are apparent in a pattern from the residue of the retracting layer on top of a pillar (Figure 7A, 7B). The best resolution of 1.6 nm obtained until now for a TMV coffee-ring residue [43] cannot yet compete with the 0.29 nm resolution obtained by X-ray fiber diffraction on much larger scattering volumes [44]. Improving the resolution of self-assembled TMV particles on artificially structured substrates will be an interesting challenge.

Manipulation and assembly of objects with ultralow contact forces

Objects from nm to several 10^th of μm dimensions can be trapped by pN optical forces generated by milliwatts of visible laser light focused through a microscope objective onto the object, such as crystallites in a capillary (Figure 8A) [45, 49]. The optical field reduces random motion due to Brownian forces, allowing raster-scans of trapped objects by microPX techniques [48, 49]. The laser beam can be split holographically into multiple traps [46]. Indeed, an about 30 μm insulin crystal could be kept in a fixed position by three holographic traps and its radiation damage explored in-situ using a dedicated optical tweezers setup (Figure 8B) [49]. Radiation damage appeared only beyond about 50 ms local exposure suggesting that SSX based on raster-rotation scans would be feasible for larger microcrystals provided that the separation of the scan points is a few μm, i.e. corresponding to the range of secondary radiation damage due to X-ray induced photoelectrons [50].

Materials used for sample supports or confinement

In order to cope with objects of very low scattering power, sample supports and confining walls have to have ultralow background scattering and X-ray absorption. Indeed, scattering from a 10 μm single-crystalline silicon membrane [37] is very low for a transmission of ~97 % at λ~0.1 nm wavelength. The transmission is > 99.8% for a 1 μm thick Si₃N₄ membrane used in a superhydrophobic chip [41]. This material is, however, very fragile and future sample supports should rather be based on few atomic layers thick graphene which has virtually 100 % transmission and no background scattering [51].

Conclusion

“Pushing the limits of protein crystallography to smaller crystals and smaller beam sizes requires the integration of more and more nanotechnology into a MPX beam line and annex laboratories” [1]. The examples reviewed in this article with an emphasis for on-chip sample environments support this statement and suggest R&D opportunities related to nanobiotechnology tools and techniques.

Acknowledgements

Helpful discussions with A. Accardo, F. De Angelis and G. Marinaro (IIT-Genova), E. Pechkova (Genova University), C. Nicolini (Fondazione ELBA), M. Burghammer, A. Popov, S.C. Santucci (ESRF-Grenoble), J.P. Colletier and N. Coquelle (IBS-Grenoble) are gratefully acknowledged.

References

1. Riekel C, Burghammer M, Popov D. 2011. Protein Micro- and Nanocrystallography using Synchrotron Radiation. In: Pechkova E and Riekel C (eds) Synchrotron Radiation and Structural Proteomics, Pan Stanford Series on Nanobiotechnology, CRC (Taylor & Francis), Boca Raton, FL, USA, pp: 3-30.

2. Smith JL, Fischetti RF, Yamamoto M. 2012. Micro-Crystallography Comes of Age. Curr Opin Struct Biol 22(5): 602-612. doi: 10.1016/j.sbi.2012.09.001

3. Nelson R, Sawaya MR, Balbirnie M, Riekel C, Grothe R, et al. 2005. Structure of the Cross-b Spine of Amyloid-Like Fibrils. Nature 435: 773-778. doi: 10.1038/nature03680

4. Cherezov V, Rosenbaum DM, Hanson MA, Thian FS, Kobilka TS, et al. 2007. High-Resolution Crystal Structure of an Engineered Human β2-Adrenergic G Protein–Coupled Receptor. Science 318(5854): 1258-1265. doi: 10.1126/science.1150577

5. Rasmussen SGF, Choi HJ, Rosenbaum DM, Kobilka TS, Thian FS, et al. 2007. Crystal Structure of the Human [bgr]: 2: Adrenergic G-Protein-Coupled Receptor. Nature 450(7168): 383-387. doi: 10.1038/nature06325

6. Boutet S, Williams GJ. 2010. The Coherent X-ray Imaging (CXI) instrument at the Linac Coherent Light Source (LCLS). New J Phys 12: 035024. doi: 10.1088/1367-2630/12/3/035024

7. Schlichting I. 2015. Serial Femtosecond Crystallography: the First Five Years. IUCrJ 2(2): 246-255. doi: 10.1107/S205225251402702X

8. Neutze R, van der Spoel D, Weckert E, Hadju J, Wouts R. 2000. Potential for Biomolecular Imaging with Femtosecond X-Ray Pulses. Nature 406: 753-757. doi: 10.1038/35021099

9. Chapman HN, Fromme P, Barty A, White TA, Kirian RA, et al. 2011. Femtosecond X-Ray Protein Nanocrystallography. Nature 470(7332): 73-77. doi: 10.1038/nature09750

10. White TA, Barty A, Stellato F, Holton JM, Kirian RA, et al. 2013. Crystallographic Data Processing for Free-Electron Laser Sources. Acta Crystallogr D Biol Crystallogr 69(7): 1231-1240. doi: 10.1107/S0907444913013620

11. DePonte DP, Weierstall U, Schmidt K, Starodub D, Starodub JCH, et al. 2008. Gas Dynamic Virtual Nozzle for Generation of Microscopic Droplet Streams. J Phys D Appl Phys 41: 195505.

12. Muniyappan S, Kim SO, Ihee H. 2015. Recent Advances and Future Prospects of Serial Crystallography using XFEL and Synchrotron X-Ray Sources. Bio Design 3(2): 98-110.

13. Gati C, Bourenkov G, Klinge M, Rehders D, Stellato, F, et al. 2014. Serial Crystallography on In Vivo Grown Microcrystals using Synchrotron Radiation. IUCrJ 1(2): 87-94. doi: 10.1107/S2052252513033939

14. Gazit E, Mitraki A. 2013. Plenty of Room for Biology at the Bottom. An Introduction into Bionanotechnology (World Scientific).

15. Tavares PF, Leemann SC, Sjoestroem M, Andersson A. 2014. The MAX IV Storage Ring Project. J Synchr Rad 21(5): 862-877. doi: 10.1107/S1600577514011503

16. National Synchrotron Light Source II: Wikipedia

17. Eriksson M, Van der Veen JF, Quitmann C. 2014. Diffraction-Limited Storage Rings – a Window to the Science of Tomorrow. J Synchr Rad 21(5): 837-842. doi: 10.1107/S1600577514019286

18. Dimper R, Reichert H, Raimondi P, Ortiz LS, Sette F, et al. 2015. The Orange Book: ESRF Upgrade Programme Phase II (2015-2022) - Technical Design Study. The European Synchrotron, Grenoble.

19. Pechkova E, Belmonte L, Riekel C, Popov D, Nicolini C, et al. 2013. Laser-Microdissection of Protein Crystals Down to Submicron Dimensions. J Nanomed Nanotech S15: 002. doi: 10.4172/2157-7439

20. Coquelle N, Brewster AS, Kapp U, Shilova A, Weinhausen B, et al. 2015. Raster-Scanning Serial Protein Crystallography using Micro and Nano-Focused Synchrotron Beams. Acta Crystallogr D Biol Crystallogr 71(8): 1184-1196. doi: 10.1107/S1399004715004514

21. Gentile F, Moretti M, Limongi T, Falqui A, Bertoni G, et al. 2012. Direct Imaging of DNA Fibers: The Visage of Double Helix. Nano Lett 12(12): 6453-6458. doi: 10.1021/nl3039162

22. Zander U, Bourenkov G, Popov AN, de Sanctis D, Svensson O, et al. 2015. Mesh and Collect: an Automated Multi-Crystal Data Collection Workflow for Synchrotron Macromolecular Crystallography Beamlines. Acta Crystallogr D Biol Crystallogr 71(Pt 11): 2328-2343. doi: 10.1107/S1399004715017927..

23. Redecke L, Nass K, DePonte DP, White TA, Rehders D,  et al. 2013. Natively Inhibited Trypanosoma Brucei Cathepsin B Structure Determined by Using an X-ray Laser. Science 339(6116): 227-230. doi: 10.1126/science.1229663

24. Giordano R, Leal RM, Bourenkov GP, McSweeney S, Popov AN. 2012. The Application of Hierarchical Cluster Analysis to the Selection of Isomorphous Crystals. Acta Crystallogr D Biol Crystallogr 68(6): 649-658.doi: 10.1107/S0907444912006841

25. Svensson O, Malbet-Monaco S, Popov A, Nurizzoa D, Bowler MW. 2015. Fully Automatic Characterization and Data Collection from Crystals of Biological Macromolecules. Acta Crystallogr D Biol Crystallogr 71(8): 1757-1767. doi: 10.1107/S1399004715011918

26. Nave C. 1999. Matching X-Ray Source, Optics and Detectors to Protein Crystallography Requirements. Acta Crystallogr D Biol Crystallogr 55: 1663-1668. doi: 10.1107/S0907444999008380

27. Pechkova E, Nicolini C. 2003. Proteomics and Nanocrystallography. Kluwer Academic Press, NY, USA. doi: 10.1007/978-1-4615-0041-4

28. Nave C, Hill MA. 2005. Will Reduced Radiation Damage Occur with Very Small Crystals? J Synchotron Rad 12: 299-303. doi: 10.1107/S0909049505003274

29. Cowan JA, Nave C. 2008. The Optimum Conditions to Collect X-Ray Data from Very Small Samples. J Synchrotron Rad 15(5): 458-462. doi: 10.1107/S0909049508014623

30. Liu W, Wacker D, Gati C, Han GW, James D, et al. 2013. Serial Femtosecond Crystallography of G Protein–Coupled Receptors. Science 342: 1521-1524. doi: 10.1126/science.1244142

31.Weierstall U, James D, Wang C, White TA, Wang D, et al. 2014. Lipidic Cubic Phase Injector Facilitates Membrane Protein Serial Femtosecond Crystallography. Nat Commun 5: 3309. doi: 10.1038/ncomms4309

32. Hunter MS, Segelke B, Messerschmidt M, Williams GJ, Zatsepin NA, et al. 2014. Fixed-Target Protein Serial Microcrystallography with an X-Ray Free Electron Laser. Sci Rep 4: 1-5. doi: 10.1038/srep06026

33. Caffrey M, Li D, Howe N, Shah STA. 2014. "Hit and Run" Serial Femtosecond Crystallography of a Membrane Kinase in the Lipid Cubic Phase. Phil Trans R Soc London B 369(1674): 20130621. doi: 10.1098/rstb.2013.0621

34. Stellato F, Oberthur D, Liang M, Bean R, Gati C, et al. 2014. Room-Temperature Macromolecular Serial Crystallography Using Synchrotron Radiation. IUCrJ 1(4): 204-212. doi: 10.1107/S2052252514010070

35. Nogly P, James D, Wang D, White TA, Zatsepin N, et al. 2015. Lipidic Cubic Phase Serial Millisecond Crystallography using Synchrotron Radiation. IUCrJ 2(2):168-176. doi: 10.1107/S2052252514026487

36. Accardo A, Graceffa R, Riekel C. Deterministic Delivery of Ultrasmall Volumes of Biological Materials (in preparation).

37. Roedig P, Vartiainen I, Duman R, Panneerselvam S, Lorbeer O, et al. 2015. A Micro-Patterned Silicon Chip as Sample Holder for Macromolecular Crystallography Experiments with Minimal Background Scattering. Sci Rep 5: 10451. doi: 10.1038/srep10451

38. Accardo A, DiFabrizio E, Limongia T, Marinaro G, Riekel C. 2014. Probing Droplets on Superhydrophobic Surfaces by Synchrotron Radiation Scattering Techniques. J Synchrotron Radiat 21(4): 643-653. doi: 10.1107/S1600577514009849

39. Gentile F, Moretti M, Limongi T, Falqui A, Bertoni G, et al. 2012. Direct Imaging of DNA Fibers: The Visage of Double Helix. Nano Lett 12(12): 6453-6458. doi: 10.1021/nl3039162

40. Marinaro G, Burghammer M, Costa L, Dane T, de Angelis F, et al. 2015. Directed Growth of Virus Nanofilaments on a Superhydrophobic Surface. ACS Appl Mater Interfaces 7(23): 12373–12379. doi: 10.1021/am507509z

41. Marinaro G, Accardo A, De Angelis F, Dane T, Weinhausen B, et al. 2014. A Superhydrophobic Chip based on SU-8 Photoresist Pillars Suspended on a Silicon Nitride Membrane. Lab Chip 14(19): 3705-3709. doi: 10.1039/c4lc00750f

42. Marinaro G. 2015. Contributions to Modeling and Applications of Superhydrophobic Surfaces for Self-Assembly of Biological Materials. PhD Thesis, University of Genova, Genova, Italy.

43. Gebhardt R, Teulon JM, Pellequer JL, Burghammer M, Colletier JP, et al. 2014. Virus Particle Assembly into Crystalline Domains Enabled by the Coffee Ring Effect. Soft Matter 10(30): 5458-5462. doi: 10.1039/c4sm00414k

44. Namba K, Pattanayek R, Stubbs G. 1989. Visualization of Protein-Nucleic Acid Interactions in a Virus. Refined Structure of Intact Tobacco Mosaic Virus at 2.9 A Resolution by X-Ray Fiber Diffraction. J Mol Biol 208(2): 307-325.

45. Ashkin A, Dziedzic JM. 1971. Optical Levitation by Radiation Pressure. Appl Phys Lett 19: 283-285. doi: 10.1063/1.1653919

46. Grier DG. 2003. A Revolution in Optical Manipulation. Nature 424: 810-816. doi: 10.1038/nature01935

47. Neuman KC, Block SM. 2004. Optical Trapping. Rev Sci Instrum 75(9): 2787-2809. doi: 10.1063/1.1785844

48. Santucci SC, Amenitsch H, Cojoc D, Riekel C. 2011. Optical Tweezers for Touchless Sample Manipulation in Synchrotron Radiation Experiments. In: Pechkova E and Riekel C (eds) Synchrotron Radiation and Structural Proteomics, Pan Stanford Series on Nanobiotechnology, CRC (Taylor & Francis), Boca Raton, FL, USA, pp 183–211.

49. Santucci SC, Cojoc D, Amenitsch H, Marmiroli B, Sartori B, et al. 2011. Optical Tweezers for Synchrotron Radiation Probing of Trapped Biological and Soft Matter Objects in Aqueous Environments. Analytical Chemistry 83(12): 4863-4870. doi: 10.1021/ac200515x

50. Cole A. 1969. Absorption of 20-eV to 50,000-eV Electron Beams in Air and Plastic. Radiat Res 38(1): 7-33.

51. Gruner SM, Lattman EE. 2015. Biostructural Science Inspired by Next-Generation X-Ray Sources. Annu Rev Biophys 44: 33-51. doi: 10.1146/annurev-biophys-060414-033813

52. Wallen E. 2011. Comparison of the Brilliance Limit between MAX IV 3 GeV Ring and NSLS II Low Beta Straights Using the Same Undulator Technique. in Internal Note 20111110 (MAX IV Lab).