Skip to main content
Acta Crystallographica Section F: Structural Biology and Crystallization Communications logoLink to Acta Crystallographica Section F: Structural Biology and Crystallization Communications
. 2012 Jan 21;68(Pt 2):119–123. doi: 10.1107/S1744309111051499

Room-temperature ultrahigh-resolution time-of-flight neutron and X-ray diffraction studies of H/D-exchanged crambin

Julian C-H Chen a,*, Zoë Fisher b, Andrey Y Kovalevsky b, Marat Mustyakimov b, B Leif Hanson c, Vladimir V Zhurov c, Paul Langan c,d
PMCID: PMC3274385  PMID: 22297981

The 0.85 Å room-temperature ultrahigh-resolution structure of H/D-exchanged crambin is reported. Preliminary 1.1 Å resolution neutron diffraction data have been collected at the neutron Protein Crystallography Station at LANSCE.

Keywords: crambin, neutron diffraction, ultrahigh resolution, H/D exchange

Abstract

The room-temperature (RT) X-ray structure of H/D-exchanged crambin is reported at 0.85 Å resolution. As one of the very few proteins refined with anisotropic atomic displacement parameters at two temperatures, the dynamics of atoms in the RT and 100 K structures are compared. Neutron diffraction data from an H/D-exchanged crambin crystal collected at the Protein Crystallo­graphy Station (PCS) showed diffraction beyond 1.1 Å resolution. This is the highest resolution neutron diffraction reported to date for a protein crystal and will reveal important details of the anisotropic motions of H and D atoms in protein structures.

1. Introduction  

H atoms make up approximately one-half of all atoms in a protein molecule. H atoms are involved in enzyme mechanisms and stabilizing hydrogen-bond interactions and are key indicators of the microenvironment within a protein structure. However, they are very difficult to locate within a protein structure using X-ray diffraction techniques owing to their weak scattering power and their proximity to heavier atoms.

Ultrahigh-resolution X-ray data have the potential to yield a significant amount of structural information about a protein. At resolutions beyond ∼0.70 Å, interatomic scattering from bonding electrons and even lone electron pairs can be resolved. Charge-density refinement has been applied to a number of macromolecules at this resolution (Jelsch et al., 2000; Guillot et al., 2008). H atoms may also be visible at ultrahigh resolution, but normally only well ordered H atoms, typically at the core of proteins. However, most of the mobile and labile H atoms of critical mechanistic interest remain difficult to identify. Even with the best-diffracting protein crystals, only around two-thirds of H atoms can be resolved (Gardberg et al., 2010). Furthermore, the vibrational characteristics of H atoms in proteins remain poorly understood, with little experimental information on their properties.

Neutron crystallography is the only method that can be used to directly and precisely determine the locations of H and D atoms in macromolecular structures, even at a medium resolution of ∼1.8–2.5 Å (Blakeley, 2009). Neutrons are scattered by atomic nuclei, in contrast to X-rays, which are scattered by electrons. The scattering power of X-rays is dependent on the atomic number; in contrast, the neutron scattering properties of atomic nuclei are relatively evenly spread across the periodic table. Notably, the heavy elements in proteins have similar neutron scattering lengths. More importantly, the neutron scattering length of deuterium (+6.671 fm) is comparable to those of oxygen (+5.803 fm), carbon (+6.646 fm) and nitrogen (+9.36 fm), while hydrogen has a negative scattering length (−3.7406 fm). This property has been utilized to accurately locate H atoms of critical interest in enzyme mechanisms (Kossiakoff & Spencer, 1981; Bennett et al., 2006; Blakeley et al., 2008; Kovalevsky et al., 2008, 2010) and also to obtain information on the dynamics of proteins through their backbone amide-exchange characteristics (Kossiakoff, 1982). Furthermore, neutron crystallography has been successful in resolving water species such as D2O, OD and D3O+, orienting hydrogen-bond donors and acceptors, and for orienting metal-coordinated solvent molecules (Blum et al., 2009; Chen et al., 2010; Fisher et al., 2010; Kovalevsky et al., 2011).

Crambin (46 amino acids, 4.7 kDa) is a small protein derived from the seed of the Abyssinian cabbage (Crambe abyssinica). Crambin contains both α-helical and β-strand secondary structures and is stabilized by three disulfide bridges (Fig. 1 a). Its overall fold is similar to that of a family of plant toxins, although it has no known biological function. The protein is hydrophobic, is soluble in polar organic solvents such as acetone and ethanol, and is insoluble in water and alkanes. Crambin forms the best-diffracting protein crystals known; synchrotron radiation has yielded data to a usable resolution of 0.48 Å, with observed reflections to 0.40 Å (Teeter & Hendrickson, 1979; Jelsch et al., 2000; Schmidt et al., 2011). With the possibility of obtaining ultrahigh-resolution neutron diffraction data, it is expected that it should be possible to resolve not only the locations of H and D atoms but also to refine their anisotropic vibrational characteristics. As such, crambin is an excellent target for neutron crystallography and the aim of this study is to obtain neutron diffraction to ultrahigh resolution as a first step towards understanding the motions of H and D atoms in proteins in the context of their heavier neighbors. These ultrahigh-resolution data may also help in improving the parameterization of neutron structure refinement techniques.

Figure 1.

Figure 1

(a) Stick and secondary-structure representations of crambin. Disulfide bridges are labeled in green. (b) X-ray structure of H/D-exchanged crambin with anisotropic ellipsoids, colored by B factor from low (blue) to high (red). The orientation of the molecule is the same in (a) and (b). (c) Representative electron density for residue Tyr44 of H/D-exchanged crambin. 2F oF c electron density contoured at 1.5σ is shown in blue, with F oF c density at 3.0σ in green. Difference density for the hydrogen of the tyrosine hydroxyl group allows clear orientation of the H atom.

The ultrahigh-resolution RT X-ray structure of H/D-exchanged crambin is reported at 0.85 Å, together with preliminary 1.1 Å neutron diffraction data collected at the PCS at the Los Alamos Neutron Science Center, demonstrating the exceptional order of crambin crystals. Future prospects for the use of ultrahigh-resolution neutron and X-ray data are discussed.

2. Experimental  

2.1. Protein preparation and crystallization  

Crambin was purified according to the procedures of Van Etten et al. (1965). Crystals were grown by vapor diffusion of an ∼20 mg ml−1 solution of crambin in 80% ethanol against a well solution of 55% ethanol at RT. For the ultrahigh-resolution X-ray data collection, a single prismatic crystal measuring approximately 1.6 × 0.4 × 0.7 mm was mounted in a quartz capillary and was exchanged for two weeks against mother liquor containing 55% perdeuterated ethanol (Cambridge Isotopes) in D2O. For neutron data collection, a larger crystal measuring 3.0 × 2.0 × 0.7 mm (4.2 mm3) was mounted in a quartz capillary and was exchanged against perdeuterated mother liquor by vapor diffusion for six weeks prior to data collection.

2.2. X-ray data collection and refinement  

Data collection was performed with a Rigaku diffractometer equipped with an UltraX-18 Mo rotating-anode generator (50 kV, 300 mA), a graphite monochromator and a RAPID cylindrical image-plate detector. Data were processed and scaled in d*TREK (Pflugrath, 1999) and the structure was solved by molecular replacement using the crambin structure 1crn (Teeter, 1984) as a search model. The structure was refined to 0.85 Å resolution in PHENIX using positional and anisotropic B-factor refinement. Riding H atoms were added in the later rounds of refinement and model building was performed in Coot. Solvent molecules were refined as O atoms only with anisotropic atomic displacement parameters (ADPs). Data-collection and refinement statistics are shown in Table 1.

Table 1. Data-collection and refinement statistics.

Values in parentheses are for the highest resolution shell.

  X-ray Neutron
Source Rotating anode PCS, LANSCE
Wavelength () 0.71073 [MoK] 0.66.0
Temperature (K) 293
Unit-cell parameters
a () 22.79
b () 18.83
c () 41.04
() 90.89
Space group P21 P21
Resolution () 300.85 (0.880.85) 201.10 (1.161.10)
Completeness (%) 99.8 (99.9) 77.4 (64.6)
Unique reflections 30963 (3067) 10930 (1302)
Multiplicity 3.41 (3.21) 2.6 (1.9)
R merge (%) 5.0 (34.2) 22.1 (29.6)
R p.i.m. (%) N/A 13.9 (21.8)
I/(I) 8.2 (1.9) 5.6 (1.9)
Refinement
No. of reflections 30659  
R free/R cryst (%) 13.60/11.94  
R.m.s.d. bonds () 0.008  
R.m.s.d. angles () 1.380  
B protein/solvent (2) 6.6/18.1  
Ramachandran outliers 0  

2.3. Neutron data collection  

Neutron time-of-flight Laue data were collected at room tem­perature on the Protein Crystallography Station at the Los Alamos Neutron Science Center (LANSCE; Schoenborn & Langan, 2004). The detector is a position-sensitive 3He-filled detector and the sample was mounted on a kappa-circle goniometer from Huber. Each exposure was 12 h and to date 44 crystal settings have been collected to a maximum usable resolution of 1.1 Å. With the detector positioned with 2θ = 0° the diffraction was observed to extend to the edge of the detector. One image was collected at 2θ = 70° for 20 h to assess the limit of diffraction (Fig. 4c). In order to collect a complete high-resolution data set the detector was set to 2θ = 30°. The images were processed with an in-house-modified version of d*TREK and wavelength-normalized using LAUENORM (Langan & Greene, 2004; Helliwell et al., 1989). Data were merged with SCALA from the CCP4 suite (Winn et al., 2011). Data-collection statistics are summarized in Table 1.

3. Results and discussion  

3.1. RT X-ray structure of H/D-exchanged crambin  

The RT X-ray structure of H/D-exchanged crambin was solved at 0.85 Å resolution. Labile H atoms were exchanged for D atoms through vapor exchange against perdeuterated mother liquor prior to data collection. H atoms in the structure were refined in their riding positions with isotropic ADPs, and a number of them could be identified as F oF c difference peaks throughout the structure (Fig. 1 c). Future refinement of a complete neutron data set will clarify which atoms have exchanged for deuteriums and the extent of the exchange.

The overall structure is identical to the 1.50 Å resolution RT structure 1crn, with an average 0.07 Å r.m.s.d. at the Cα positions, and closely resembles the 0.54 Å resolution (PDB entry 1ejg; Jelsch et al., 2000) and 0.48 Å resolution (PDB entry 3nir; Schmidt et al., 2011) structures solved at 100 K, with average r.m.s.d.s of 0.116 and 0.128 Å at Cα positions, respectively. A comparison of the hydration patterns showed little difference between the structures. Therefore, the H/D-exchange process did not affect the overall structure of the protein or its hydration. The X-ray structures of perdeuterated proteins, in which all H atoms have been replaced by deuteriums, also showed nearly identical structures to their H/D-exchanged and hydrogenated counterparts (Blum et al., 2010). The crystal is notable in having a low solvent content of 34.4%. Much of the solvent region between symmetry-related protein molecules is ordered, with relatively little ‘bulk’ solvent. Such an extensive network of hydrogen-bonded contacts may in part explain the very high degree of order in these crystals.

3.2. Anisotropic motions of crambin at RT versus 100 K  

This is one of the few ultrahigh-resolution protein structures reported at ambient temperature and provides a view of anisotropic atomic motions of a macromolecule at RT (Fig. 1 b). Furthermore, the RT structure and the anisotropic vibrational characteristics of the atoms can be compared with the set of structures solved at 100 K. Anisotropic atomic displacement parameters (ADPs) are expressed as a tensor consisting of six values relating to the atomic motions in three dimensions.

Data for 1ejg and 3nir were obtained at 100 K and the structures reported at 0.54 and 0.48 Å, respectively. As expected for a RT structure, the average B factor for the protein atoms in our current structure is 6.6 Å2, higher than 1ejg and 3nir, and the distribution of B factors is similar in all three structures (Fig. 2).

Figure 2.

Figure 2

Distribution of Cα B factors (top) and anisotropy (bottom) in RT H/D-exchanged crambin (green) and the 100 K 1ejg (gray) and 3nir (magenta) structures.

The RT structure is considerably more isotropic in nature, with a mean anisotropy of 0.578 for the protein atoms, compared with 0.404 for 1ejg, 0.420 for 3nir and an average of 0.45 for ∼1000 structures in the PDB (http://skuld.bmsc.washington.edu/parvati/parvati_survey.html). The anisotropy value is defined as the square of the ratio of the smallest and largest of the eigenvalues of the 3 × 3 matrix of the anisotropic displacement parameters. A higher anisotropy value represents a more isotropic atom. The distribution of ADPs and their anisotropy are similar in the three crambin structures discussed here, with a region of higher ADP values between residues 34 and 42 which correlates with lower anisotropy values (more anisotropic, less iso­tropic; Fig. 2). In the case of the two low-temperature structures 1ejg and 3nir, several discrepancies are seen in the anisotropy values which may be a consequence of incomplete modeling of alternate conformations. These figures are likely to be accurate, although they may in part reflect differences in the refinement programs used. The structure reported in this work was refined in PHENIX, while 1ejg and 3nir were refined using SHELX. Merritt compiled information about the anisotropy of structures refined using SHELX, REFMAC and PHENIX and noted a slight but statistically significant difference between the two programs (http://skuld.bmsc.washington.edu/parvati/parvati_survey.html). Nevertheless, the anisotropy of the 100 K structures is comparable to other protein structures in the PDB, while that of the RT structure is higher. At this time, it is difficult to make a comparison of the anisotropy of other RT structures in the PDB, as it is difficult to accurately assess the data-collection temperature in the PDB entries. Other possible reasons for this discrepancy may be the separation of static and dynamic disorder at lower temperatures.

3.3. Neutron diffraction of a H/D-exchanged crambin crystal  

To assess the diffraction limit currently possible for crystals of crambin using current neutron sources, 44 diffraction images have been collected from a single crystal (Fig. 3) at the PCS in Los Alamos, with a total coverage of 77% of reflections to 1.1 Å resolution and 65% of reflections in the outermost shell (1.16–1.10 Å resolution; Table 1, Figs. 4 a and 4 b).

Figure 3.

Figure 3

Photograph of the crambin crystal used for neutron diffraction data collection, measuring 3 × 2 × 0.7 mm (4.2 mm3).

Figure 4.

Figure 4

(a, b) Laue time-of-flight neutron diffraction images from two different settings of H/D-exchanged crambin collected at the PCS. The exposure time was 12 h, with the 2θ arm set at 30°. (c) For this image the 2θ arm was moved to 70° and the exposure time was 20 h. The acquired data are displayed as an overlay of all time bins, generating the Laue diffraction pattern in the image. (d) Diffraction slice at a wavelength of 1.35 Å between 57° and 82° in 2θ. A reflection at 1.05 Å resolution is circled.

A full data set is currently being collected. We expect complete data to ∼1.1 Å resolution using the current data-collection protocol, although it is clear that the crystal diffracts to beyond 1.05 Å from inspection of a diffraction image collected at a 2θ of 70° (Figs. 4 c and 4 d). A significant amount of beam time is necessary to collect data beyond 1.1 Å resolution and is currently not practical given the constraints on beam time and the geometry of the detector. The preliminary data show strong diffraction even in the highest resolution shell, with a 〈I/σ(I)〉 of 1.9 in the outermost shell, with very good merging statistics. The data are thus of excellent quality. As the data collection currently stands, there is sufficient completeness to 1.8 Å resolution (∼85%), which is a high enough resolution to determine the orientation of solvent molecules and amino-acid side chains and to determine the extent of H/D exchange in the backbone. A joint X-­ray and neutron refinement of the structure to 1.8 Å resolution is under way.

We expect that when the full data set is collected the data-to-parameter ratio for the neutron data alone will be sufficient to justify anisotropic refinement of all atoms, including H and D atoms, and solvent molecules. We can also exploit the advantages of joint refinement protocols, which in the case of crambin will combine ultrahigh-resolution X-ray data and atomic resolution neutron data, possibly yielding the most detailed and accurate protein structure to date.

3.4. Perspective  

This work may open up a number of new possibilities in macromolecular neutron crystallography. Of great benefit to our understanding of protein structure and dynamics is the possibility of refining anisotropic ADPs for H and D atoms. Although refinement of anisotropic ADPs is routinely performed for the heavier atoms in proteins (C, N, O and S) using atomic resolution X-ray data (typically to better than 1.2 Å resolution), such treatment is not possible for H atoms. Neutron crystallography, on the other hand, offers insight into H-atom positions owing to their scattering properties, and now with ultrahigh-resolution neutron diffraction data, both positions and dynamics can be determined with high accuracy.

To date, 11 structures in the PDB have been resolved to beyond 0.70 Å resolution, a number that will undoubtedly increase, especially with the development of new synchrotron sources and free electron lasers combined with new and more accurate detectors and hardware. Therefore, more accurate parameterization for crystallo­graphic refinement that includes H atoms will be needed in order to provide higher quality models at high resolutions.

At a software level, this is an ideal test case for joint XN refinement protocols. Joint XN refinement, described over 30 years ago by Wlodawer and Hendrickson, has become a practical tool for refining neutron structures and has been implemented in the commonly used refinement programs CNS (as nCNS) and PHENIX (Adams et al., 2009).

Synchrotron data have also been collected on crambin crystals at DESY to a nominal resolution of 0.48 Å, with observed diffraction spots out to 0.40 Å resolution (Schmidt et al., 2011). With a combination of synchrotron and neutron diffraction data, it should be possible to separate the static and dynamic contributions to disorder. These methods will afford the most detailed view of a macromolecule to date and may provide experimental information about bonding and electrostatics that can be tested against theoretical calculations.

The coordinates and structure factors for the RT X-ray structure of H/D-exchanged crambin have been deposited in the PDB under accession code 3u7t.

Supplementary Material

PDB reference: H/D-exchanged crambin, 3u7t

Acknowledgments

We thank Dr Blaine Schatz for providing crambe seed for this study, Dr Pavel Afonine for advice on PHENIX refinement and Mary Jo Waltman, Dr A. Alan Pinkerton and Dr Kristin Kirschbaum for technical assistance. JC-HC thanks Dr Justin and Cynthia Chen for their support. The PCS is funded by the US Department of Energy Office of Biological and Environmental Research (DOE-OBER).

References

  1. Adams, P. D., Mustyakimov, M., Afonine, P. V. & Langan, P. (2009). Acta Cryst. D65, 567–573. [DOI] [PMC free article] [PubMed]
  2. Bennett, B., Langan, P., Coates, L., Mustyakimov, M., Schoenborn, B., Howell, E. E. & Dealwis, C. (2006). Proc. Natl Acad. Sci. USA, 103, 18493–18498. [DOI] [PMC free article] [PubMed]
  3. Blakeley, M. P. (2009). Crystallogr. Rev. 15, 157–218.
  4. Blakeley, M. P., Ruiz, F., Cachau, R., Hazemann, I., Meilleur, F., Mitschler, A., Ginell, S., Afonine, P., Ventura, O. N., Cousido-Siah, A., Haertlein, M., Joachimiak, A., Myles, D. & Podjarny, A. (2008). Proc. Natl Acad. Sci. USA, 105, 1844–1848. [DOI] [PMC free article] [PubMed]
  5. Blum, M.-M., Mustyakimov, M., Rüterjans, H., Kehe, K., Schoenborn, B. P., Langan, P. & Chen, J. C.-H. (2009). Proc. Natl Acad. Sci. USA, 106, 713–718. [DOI] [PMC free article] [PubMed]
  6. Blum, M.-M., Tomanicek, S. J., John, H., Hanson, B. L., Rüterjans, H., Schoenborn, B. P., Langan, P. & Chen, J. C.-H. (2010). Acta Cryst. F66, 379–385. [DOI] [PMC free article] [PubMed]
  7. Chen, J. C.-H., Mustyakimov, M., Schoenborn, B. P., Langan, P. & Blum, M.-M. (2010). Acta Cryst. D66, 1131–1138. [DOI] [PMC free article] [PubMed]
  8. Fisher, S. Z., Kovalevsky, A. Y., Domsic, J. F., Mustyakimov, M., McKenna, R., Silverman, D. N. & Langan, P. A. (2010). Biochemistry, 49, 415–421.
  9. Gardberg, A. S., Del Castillo, A. R., Weiss, K. L., Meilleur, F., Blakeley, M. P. & Myles, D. A. A. (2010). Acta Cryst. D66, 558–567. [DOI] [PubMed]
  10. Guillot, B., Jelsch, C., Podjarny, A. & Lecomte, C. (2008). Acta Cryst. D64, 567–588. [DOI] [PubMed]
  11. Helliwell, J. R., Habash, J., Cruickshank, D. W. J., Harding, M. M., Greenhough, T. J., Campbell, J. W., Clifton, I. J., Elder, M., Machin, P. A., Papiz, M. Z. & Zurek, S. (1989). J. Appl. Cryst. 22, 483–497.
  12. Jelsch, C., Teeter, M. M., Lamzin, V., Pichon-Pesme, V., Blessing, R. H. & Lecomte, C. (2000). Proc. Natl Acad. Sci. USA, 97, 3171–3176. [DOI] [PMC free article] [PubMed]
  13. Kossiakoff, A. A. (1982). Nature (London), 296, 713–721. [DOI] [PubMed]
  14. Kossiakoff, A. A. & Spencer, S. A. (1981). Biochemistry, 20, 6462–6474. [DOI] [PubMed]
  15. Kovalevsky, A. Y., Hanson, L., Fisher, S. Z., Mustyakimov, M., Mason, S. A., Forsyth, V. T., Blakeley, M. P., Keen, D. A., Wagner, T., Carrell, H. L., Katz, A. K., Glusker, J. P. & Langan, P. (2010). Structure, 18, 688–699. [DOI] [PMC free article] [PubMed]
  16. Kovalevsky, A. Y., Hanson, B. L., Mason, S. A., Yoshida, T., Fisher, S. Z., Mustyakimov, M., Forsyth, V. T., Blakeley, M. P., Keen, D. A. & Langan, P. (2011). Angew. Chem. Int. Ed. Engl. 50, 7520–7523. [DOI] [PubMed]
  17. Kovalevsky, A. Y., Katz, A. K., Carrell, H. L., Hanson, L., Mustyakimov, M., Fisher, S. Z., Coates, L., Schoenborn, B. P., Bunick, G. J., Glusker, J. P. & Langan, P. (2008). Biochemistry, 47, 7595–7597. [DOI] [PMC free article] [PubMed]
  18. Langan, P. & Greene, G. (2004). J. Appl. Cryst. 37, 253–257.
  19. Pflugrath, J. W. (1999). Acta Cryst. D55, 1718–1725. [DOI] [PubMed]
  20. Schmidt, A., Teeter, M., Weckert, E. & Lamzin, V. S. (2011). Acta Cryst. F67, 424–428. [DOI] [PMC free article] [PubMed]
  21. Schoenborn, B. P. & Langan, P. (2004). J. Synchrotron Rad. 11, 80–82. [DOI] [PubMed]
  22. Teeter, M. M. (1984). Proc. Natl Acad. Sci. USA, 81, 6014–6018. [DOI] [PMC free article] [PubMed]
  23. Teeter, M. M. & Hendrickson, W. A. (1979). J. Mol. Biol. 127, 219–223. [DOI] [PubMed]
  24. Van Etten, C. H., Nielsen, H. C. & Peters, J. E. (1965). Phytochemistry, 4, 467–473.
  25. Winn, M. D. et al. (2011). Acta Cryst. D67, 235–242.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

PDB reference: H/D-exchanged crambin, 3u7t


Articles from Acta Crystallographica Section F are provided here courtesy of International Union of Crystallography

RESOURCES