High-resolution crystal structure of spin labelled (T21R1) azurin from Pseudomonas aeruginosa: a challenging structural benchmark for in silico spin labelling algorithms
© Florin et al.; licensee BioMed Central Ltd. 2014
Received: 7 March 2014
Accepted: 8 May 2014
Published: 29 May 2014
EPR-based distance measurements between spin labels in proteins have become a valuable tool in structural biology. The direct translation of the experimental distances into structural information is however often impaired by the intrinsic flexibility of the spin labelled side chains. Different algorithms exist that predict the approximate conformation of the spin label either by using pre-computed rotamer libraries of the labelled side chain (rotamer approach) or by simply determining its accessible volume (accessible volume approach). Surprisingly, comparisons with many experimental distances have shown that both approaches deliver the same distance prediction accuracy of about 3 Å.
Here, instead of comparing predicted and experimental distances, we test the ability of both approaches to predict the actual conformations of spin labels found in a new high-resolution crystal structure of spin labelled azurin (T21R1). Inside the crystal, the label is found in two very different environments which serve as a challenging test for the in silico approaches.
Our results illustrate why simple and more sophisticated programs lead to the same prediciton error. Thus, a more precise treatment of the complete environment of the label and also its interactions with the environment will be needed to increase the accuracy of in silico spin labelling algorithms.
The structural characterization of proteins by pulsed EPR methods such as pulsed electron electron double resonance (PELDOR, also known as DEER) has become increasingly popular in recent years [1–3]. A prerequisite for the PELDOR experiment is the presence of at least two paramagnetic centres in the protein. Since most proteins are diamagnetic, spin labels such as MTSSL  are routinely attached to the surface of proteins via site-directed spin labelling techniques [5–7]. The PELDOR experiment measures distances between such spin labels. The obtained distance information can then be used to analyse conformational changes e.g. of membrane proteins [8–10] or to reconstruct macromolecular complexes [11, 12]. Although spin labels, such as MTSSL, are small compared to FRET labels (MTSSL has roughly the size of an arginine residue), they still act as a flexible link between the protein and the spin-centre itself (usually the nitroxide group of MTSSL). Thus, the problem arises, that on the one hand, the PELDOR experiment delivers an accurate distance between the two spin centres, but on the other hand, the exact position of the distance vector with respect to the protein is unknown. Many studies have investigated this problem. For example, the crystallographic analysis of multiple spin labelled T4-lysozyme mutants provided insights into the interactions of spin labels with proteins and preferred rotameric states of the labels [13–16]. The problem also led to the development of in silico spin labelling programs, which aim to predict the conformation of the spin label in the local environment of the attachment site [17–21]. These programs attach a model of the spin label to a macromolecular structure and generate an ensemble of possible rotamers of the label. The programs either allow all possible rotamers (e.g. for MTSSL: the five dihedral angles (χ1- χ5) are randomly set, ) or the angles are set based on calculated rotamer libraries and/or crystal structures [18, 19, 22]. With each method, distances between the generated ensembles can be measured and compared to experimental distances. Surprisingly, extensive benchmarks have revealed that independent of the approach, the average error of the predicted distances is around 3 Å, [20, 22, 23]. This indicates that current in silico spin labelling programs do not accurately enough model the complex interactions between a spin-label and its molecular environment. It is therefore important to gain further experimental insight into label-environment interactions. This will help to further extend available rotamer libraries with experimental rotamer data and provide clues concerning possible improvements of in silico spin labelling algorithms.
We present here the high-resolution (1.5 Å) structure of azurin from Pseudomonas aeruginosa PAO1 spin labelled with MTSSL at position T21. Due to the crystal-packing environment in the triclinic crystals, the label is observed in two different but fully ordered states. The excellent quality of the electron density allows precise measurements of all five dihedral angles in both conformations. Whereas one of the conformations fits to calculated rotamer libraries  the second conformation does not, illustrating how the environment of the label can force it to adopt a conformation which is less favourable for the free label. The implications for in silico spin labelling are discussed. Further, we present a newly developed, affinity chromatography based purification and spin labelling protocol for azurin.
Protein production, spin labelling, crystallisation and structure solution
Data collection and refinement statistics
Unit cell (Å, °)
37.0, 53.7, 73.2, 74.4, 89.3, 83.4
Matthews coefficient (Å3 Da-1)
Solvent content (%)
Molecules per ASU
Resolution range (Å)
Resolution range (Å)
RMSD bonds (Å)/angles (°)
Ramachandran plot, (% favoured/allowed/disallowed)
MOLPROBITY score/clash score
Overall structure of azurin T21R1
Structure of the R1 side chain
Mutual influence of R1 side chain and protein environment
In silico spin labelling programs are important tools to translate EPR-derived distances into structural information, despite rather unsatisfying deviations between experimental and predicted distances [20, 22, 23]. To improve the algorithms, it is important to understand the reasons for these deviations, and to identify the most promising points for improvements. In published benchmark studies, the performance of individual programs is usually compared based on a comparison of predicted and experimental distances. But, the primary results of in silico spin labelling programs are predicted ensembles of spin labels. The two well-defined conformations of MTSSL in our crystal structure give rise to a number of inter-spin distances in the crystal packing and we thought this to be an interesting way to investigate how well predicted ensembles correlate with predicted distances. It should be noted that spin label conformations found in a crystal structure can be biased by interactions with the crystalline protein environment. However, in contrast to crystals of small molecules, protein crystals are interspersed by large solvent channels and usually consist of around 50% solvent. Also, before X-ray data are collected at low temperature, the crystals are typically cryo-protected by soaking them in e.g. 35% glycerol (even higher concentrations are used for cryo-protection in PELDOR samples). This prevents the formation of crystalline ice in the solvent channels, which would otherwise destroy the crystal or severly degrade its diffraction quality . Thus, a spin label, which points into a solvent channel of a protein crystal is surrounded by the protein lattice, some ordered solvent molecules interacting with the spin-labelled protein and, a glassy, frozen solution of the solvent. Nevertheless, because of the flash-cooling process and interactions with the protein lattice, the spin label conformations observed in a crystal might be different from those that are found at room temperature and/or in liquid solution.
The table in Figure 5B compares the inter-label distances taken from this experimental structure to the equivalent distances obtained from mtsslWizard and MMM. Details of the MMM analysis (partition function, number of rotamers) are shown in Additional file 2: Figure S2. Further, a comparison of the predicted mean spin label positions is shown in Additional file 3: Figure S3. As found in larger distance-based benchmarks [20, 22, 23], both programs predict some of the distances quite accurately (e.g. II↔III), whereas large deviations are found for other distances (e.g. III↔IV’). This indicates that errors stemming from the generation of the spin label ensembles are sometimes compensated for by the relative geometric arrangement of a pair of ensembles. A similar observation has also been made in the crystal structure of the Spa15 chaperone . Figure 5C illustrates this for the II-III distance: Whereas the absolute values of the distance vectors are very similar between X-ray structure, MMM and mtsslWizard, their directions differ considerably. This pair of labels is analysed in more detail below.
Clearly, the influence of the protein environment will be more pronounced at tight labelling sites, such as the one on monomer II (Figure 5C, left). As a result, deviations from the rotamer libraries become more likely, and we observe this in the crystal structure (Figure 3B). Consequentially, the R1-II/IV conformation cannot be correctly predicted when the rotamer library shown in Figure 3 is used (Figure 6A, left). It should be noted that this limitation of the rotamer approach at tight sites was pointed out by the authors of the MMM software . In such cases, the rotamer approach effectively boils down to an accessible volume approach and would again deliver results that are very similar to the latter if the occupancies for all rotamers were set to the same value (Figure 6, left).
In the examples above, the spin labelled X-ray structure was used as a basis for in silico spin labelling. Usually, the problem is even more difficult, since it is unknown, how the protein will structurally react to the addition of the spin label. The K128 side chain in our structure is an example for a structural response of the protein (Figure 4). In silico spin labelling programs try to deal with this problem by allowing a certain number of clashes between protein and label. However, by doing this, the ensemble of created rotamers will simply grow in size whereas in reality the ensemble might merely change its shape, not necessarily its size. In the end this again leads to an increased uncertainty of the prediction.
Our observations vividly illustrate why in our test case (and possibly also in general), the accessible volume approach and the more sophisticated rotamer approach often deliver very similar results. In essence, employing rotamer libraries will only increase the accuracy, when not only the rotamers but also their occupancy can be correctly predicted. The occupancy prediction is crucially dependent on the interaction of the label with its environment. The current software programs use only relatively simple (but fast) descriptions of the protein environment, whereas solvent molecules such as the bridging water shown in Figure 2 are completely ignored. Recently, the formation of hydrogen bonds has been found to be a very common type of interaction for nitroxide spin labels on proteins . Also ignored are cryo protectants such as ethylene glycole or glycerol which are used at high percentages in PELDOR samples. These can bind close to the labelling site and thereby potentially influence the label dynamics (see above and Additional file 1: Figure S1). For labelling sites close to lipid bilayers or detergent micelles it is also important to consider label-lipid or label-detergent interactions since these will likely have profound effects on the conformation of the label. Thus, to increase the prediction accuracy, more sophisticated algorithms, which accurately account for label-environment interactions will have to be employed. It has been tried multiple times to use MD simulations for this purpose, but so far, the increased effort does not seem to pay off in terms of better prediction accuracy . Promising ways to alleviate the described difficulties on the experimental side would be the use of spin labels with shorter or conformationally restrained linkers, such as the RX side chain  or the recently published V1 side chain .
Cloning, protein expression, purification and spin labelling
The gene for azurin (azu, PA4922) was PCR amplified from genomic Pseudomonas aeruginosa DNA using the PCR primers 5′-TTATAACCATGGCCGAGTGCTCGGTGG-3′ and 5′-CACCCTGACCCTGAAGTGAGAGCTCTTATAA-3′. The resulting PCR product did not contain the coding region for the N-terminal signal peptide of azurin (residues - 20 - 0), so that the target protein (residues 1 - 128 of azurin) could be expressed intracellularly in E. coli. The PCR product was then cloned into the vector pEHISGSTTEV (Huanting Liu, Biomedical Sciences Research Center, University of St Andrews, UK) via restriction enzymes NcoI and SacI, resulting in an expression construct with a TEV cleavable N-terminal His6-GST (glutathione S-transferase) tag. The T21C mutant was introduced into this construct using PCR. The resulting construct was transformed into E. coli Rosetta cells. A single colony was picked and grown over night in 50 ml of 2xYeast-Trypton (2YT) media supplemented with 100 μg/ml ampicillin and 17 μg/ml chloramphenicol with shaking at 37°C. On the following day, 1 l of 2YT medium with 50 μg/ml kanamycin and 17 μg/ml chloramphenicol were inoculated with 20 ml of the overnight culture and grown to an OD600 of 1.0. Protein expression was then induced by addition of 0.3 mM iso-propoyl-beta-thiogalactoside (IPTG). The protein expression was allowed to proceed for 3 h at 37°C with shaking at 200 rpm. The cells were then harvested by centrifugation at 2800 g, resuspended in 100 ml of lysis buffer (20 mM Tris-HCl pH 7.5, 500 mM NaCl, 30 mM Imidazol) and lysed with a cell disrupter at 30 kPsi (Constant Systems). Cell debris and insoluble proteins were spun down at 32.000 g for 15 min at 4°C. The soluble fraction was mixed with 1.5 ml Ni-NTA resin (Quiagen, pre-equilibrated in lysis buffer) and incubated for 1 h at 4°C with shaking. The resin was washed with 100 ml of lysis buffer, followed by 50 ml of lysis buffer supplemented with 1 mM tris(2-carboxyethyl)phosphine (TCEP) to reduce the introduced cysteine residue. The reducing agent was then quickly removed by washing the column with 50 ml of lysis buffer, directly followed by addition of 15 ml elution buffer (20 mM Tris-HCl pH 7.5, 500 mM NaCl, 1 M Imidazol) containing 0.7 mM of MTSSL. A large excess (~20×) of MTSSL was used, since the GST-tag of the expression construct also contained four cysteine residues. The labeling reaction was transferred to dialysis tubing and dialyzed over night against 5 l of dialysis buffer (20 mM Tris-HCl pH 7.5, 500 mM NaCl). On the next day, 4 mg TEV protease were added to the sample to cleave the GST-tag. The cleavage reaction was incubated for 3 h at room temperature. The sample was then concentrated to a volume of 2 ml, supplemented with 1 mM CuCl2 and loaded onto a Superdex200 16/60 column (GE) equilibrated with gel filtration buffer (10 mM Tris-HCl pH 8.0, 150 mM NaCl). Labelled monomeric azurin eluted at a volume of ~100 ml and had an intense blue colour.
Crystallisation, data collection and refinement
Purified and MTSSL-labelled azurin T21R1 was concentrated to 10 mg/ml for crystallization and sitting drop crystallization setups were prepared with the commercial JCSG + screen (Molecular Dimensions) in MRC plates (Molecular Dimensions). Blue, plate shaped azurin crystals grew within 2-3 days at room temperature in conditions A2, B7, C4 and C11. The components of these conditions where then used for a stochastic optimization and led to the final condition: 2.14 M ammonium sulfate, 0.28 M ammonium nitrate, 0.1 M sodium cacodylate pH 6.0. For data collection, the crystals were harvested and cryo-protected with 35% glycerol prior to flash cooling in liquid nitrogen. A diffraction data set was collected at BESSYII, BL14.1 (Berlin) using a PILATUS 6M detector. The data were indexed (iMOSFLM ) in space group P1 and processed with iMOSFLM, POINTLESS and SCALA . The structure of Zn2+-bound azurin (PDB-ID: 1E67) was used as model for molecular replacement with PHASER . The program located all four azurin molecules in the asymmetric unit. The monomers (I-IV) are related by a two-fold non-crystallographic symmetry between I, II and III, IV. The model was refined automatically using PHENIX.REFINE  and by hand using COOT . Data collection and refinement statistics are listed in Table 1.
We thank Drs Uwe Müller and Manfred Weiss for granting us beam time on BL14.1, BESSY II, Berlin. We would like to thank Dr Hartmut Niemann for access to his X-ray source. Funding from DFG (SFB 813, project A6) is gratefully acknowledged. The pEHisGSTTEV vector was a gift from Huanting Liu from the University of St Andrews, UK.
- Reginsson GW, Schiemann O: Studying biomolecular complexes with pulsed electron-electron double resonance spectroscopy. Biochem Soc Trans 2011, 39: 128–139. 10.1042/BST0390128View ArticlePubMedGoogle Scholar
- Jeschke G: DEER distance measurements on proteins. Annu Rev Phys Chem 2012, 63: 419–446. 10.1146/annurev-physchem-032511-143716View ArticlePubMedGoogle Scholar
- Schiemann O, Prisner TF: Long-range distance determinations in biomacromolecules by EPR spectroscopy. Q Rev Biophys 2007, 40: 1–53. 10.1017/S003358350700460XView ArticlePubMedGoogle Scholar
- Berliner LJ, Grunwald J, Hankovszky HO, Hideg K: A novel reversible thiol-specific spin label: papain active site labeling and inhibition. Anal Biochem 1982, 119: 450–455. 10.1016/0003-2697(82)90612-1View ArticlePubMedGoogle Scholar
- Klare JP, Steinhoff HJ: Spin labeling EPR. Photosynth Res 2009, 102: 377–390. 10.1007/s11120-009-9490-7View ArticlePubMedGoogle Scholar
- Altenbach C, Marti T, Khorana HG, Hubbell WL: Transmembrane protein structure: spin labeling of bacteriorhodopsin mutants. Science 1990, 248: 1088–1092. 10.1126/science.2160734View ArticlePubMedGoogle Scholar
- Hubbell WL, López CJ, Altenbach C, Yang Z: Technological advances in site-directed spin labeling of proteins. Curr Opin Struct Biol 2013, 23(5):725–733. doi:10.1016/j.sbi.2013.06.008 10.1016/j.sbi.2013.06.008PubMed CentralView ArticlePubMedGoogle Scholar
- Smirnova I, Kasho V, Choe J-Y, Altenbach C, Hubbell WL, Kaback HR: Sugar binding induces an outward facing conformation of LacY. Proc Natl Acad Sci U S A 2007, 104: 16504–16509. 10.1073/pnas.0708258104PubMed CentralView ArticlePubMedGoogle Scholar
- Altenbach C, Kusnetzow AK, Ernst OP, Hofmann KP, Hubbell WL: High-resolution distance mapping in rhodopsin reveals the pattern of helix movement due to activation. Proc Natl Acad Sci U S A 2008, 105: 7439–7444. 10.1073/pnas.0802515105PubMed CentralView ArticlePubMedGoogle Scholar
- Ward R, Pliotas C, Branigan E, Hacker C, Rasmussen A, Hagelueken G, Booth IR, Miller S, Lucocq J, Naismith JH, Schiemann O: Probing the structure of the mechanosensitive channel of small conductance in lipid bilayers with pulsed electron-electron double resonance. Biophys J 2014, 106: 834–842. 10.1016/j.bpj.2014.01.008PubMed CentralView ArticlePubMedGoogle Scholar
- Hilger D, Polyhach Y, Padan E, Jung H, Jeschke G: High-resolution structure of a Na+/H + antiporter dimer obtained by pulsed electron paramagnetic resonance distance measurements. Biophys J 2007, 93: 3675–3683. 10.1529/biophysj.107.109769PubMed CentralView ArticlePubMedGoogle Scholar
- Park S-Y, Borbat PP, Gonzalez-Bonet G, Bhatnagar J, Pollard AM, Freed JH, Bilwes AM, Crane BR: Reconstruction of the chemotaxis receptor-kinase assembly. Nat Struct Mol Biol 2006, 13: 400–407. 10.1038/nsmb1085View ArticlePubMedGoogle Scholar
- Langen R, Oh KJ, Cascio D, Hubbell WL: Crystal structures of spin labeled T4 lysozyme mutants: implications for the interpretation of EPR spectra in terms of structure. Biochemistry 2000, 39: 8396–8405. 10.1021/bi000604fView ArticlePubMedGoogle Scholar
- Guo Z, Cascio D, Hideg K, Hubbell W: Structural determinants of nitroxide motion in spin-labeled proteins: solvent-exposed sites in helix B of T4 lysozyme. Protein Sci 2008, 17: 228–239. 10.1110/ps.073174008PubMed CentralView ArticlePubMedGoogle Scholar
- Guo Z, Cascio D, Hideg K, Kálái T, Hubbell WL: Structural determinants of nitroxide motion in spin-labeled proteins: tertiary contact and solvent-inaccessible sites in helix G of T4 lysozyme. Protein Sci 2007, 16: 1069–1086. 10.1110/ps.062739107PubMed CentralView ArticlePubMedGoogle Scholar
- Fleissner MR, Cascio D, Hubbell WL: Structural origin of weakly ordered nitroxide motion in spin-labeled proteins. Protein Sci 2009, 18: 893–908. 10.1002/pro.96PubMed CentralView ArticlePubMedGoogle Scholar
- Sale K, Song L, Liu Y-S, Perozo E, Fajer PG: Explicit treatment of spin labels in modeling of distance constraints from dipolar EPR and DEER. J Am Chem Soc 2005, 127: 9334–9335. 10.1021/ja051652wView ArticlePubMedGoogle Scholar
- Polyhach Y, Bordignon E, Jeschke G: Rotamer libraries of spin labelled cysteines for protein studies. Phys Chem Chem Phys 2011, 13: 2356–2366. 10.1039/c0cp01865aView ArticlePubMedGoogle Scholar
- Hatmal MM, Li Y, Hegde BG, Hegde PB, Jao CC, Langen R, Haworth IS: Computer modeling of nitroxide spin labels on proteins. Biopolymers 2011, 97: 35–44.PubMed CentralView ArticlePubMedGoogle Scholar
- Hagelueken G, Ward R, Naismith JH, Schiemann O: MtsslWizard: in silico spin-labeling and generation of distance distributions in PyMOL. Appl Magn Reson 2012, 42: 377–391. 10.1007/s00723-012-0314-0PubMed CentralView ArticlePubMedGoogle Scholar
- Hagelueken G, Abdullin D, Ward R: mtsslSuite: In silico spin labelling, trilateration and distance-constrained rigid body docking in PyMOL. Mol Phys 2013, 111: 2757–2766. doi:10.1080/00268976.2013.809804 10.1080/00268976.2013.809804PubMed CentralView ArticlePubMedGoogle Scholar
- Alexander NS, Stein RA, Koteiche HA, Kaufmann KW, Mchaourab HS, Meiler J: RosettaEPR: rotamer library for spin label structure and dynamics. PLoS One 2013, 8: e72851. 10.1371/journal.pone.0072851PubMed CentralView ArticlePubMedGoogle Scholar
- Jeschke G: Conformational dynamics and distribution of nitroxide spin labels. Prog Nucl Magn Reson Spectrosc 2013, 72: 42–60.View ArticlePubMedGoogle Scholar
- Karlsson BG, Pascher T, Nordling M, Arvidsson RH, Lundberg LG: Expression of the blue copper protein azurin from Pseudomonas aeruginosa in Escherichia coli. FEBS Lett 1989, 246: 211–217. 10.1016/0014-5793(89)80285-6View ArticlePubMedGoogle Scholar
- Liu H, Naismith JH: An efficient one-step site-directed deletion, insertion, single and multiple-site plasmid mutagenesis protocol. BMC Biotechnol 2008, 8: 91. 10.1186/1472-6750-8-91PubMed CentralView ArticlePubMedGoogle Scholar
- Nar H, Huber R, Messerschmidt A, Filippou AC, Barth M, Jaquinod M, van de Kamp M, Canters GW: Characterization and crystal structure of zinc azurin, a by-product of heterologous expression in Escherichia coli of Pseudomonas aeruginosa copper azurin. Eur J Biochem 1992, 205: 1123–1129. 10.1111/j.1432-1033.1992.tb16881.xView ArticlePubMedGoogle Scholar
- Sauter NK, Grosse-Kunstleve RW, Adams PD: Improved statistics for determining the Patterson symmetry from unmerged diffraction intensities. J Appl Crystallogr 2006, 39: 158–168. 10.1107/S0021889805042299View ArticleGoogle Scholar
- Krissinel E, Henrick K: Inference of macromolecular assemblies from crystalline state. J Mol Biol 2007, 372: 774–797. 10.1016/j.jmb.2007.05.022View ArticlePubMedGoogle Scholar
- Laskowski RA, Swindells MB: LigPlot+: multiple ligand-protein interaction diagrams for drug discovery. J Chem Inf Model 2011, 51: 2778–2786. 10.1021/ci200227uView ArticlePubMedGoogle Scholar
- Garman E: Cool data: quantity AND quality. Acta Crystallogr D Biol Crystallogr 1999, 55: 1641–1653. 10.1107/S0907444999008653View ArticlePubMedGoogle Scholar
- Lillington JED, Lovett JE, Johnson S, Roversi P, Timmel CR, Lea SM: Shigella flexneri Spa15 crystal structure verified in solution by double electron electron resonance. J Mol Biol 2011, 405: 427–435. 10.1016/j.jmb.2010.10.053PubMed CentralView ArticlePubMedGoogle Scholar
- Polyhach Y, Jeschke G: Prediction of favourable sites for spin labelling of proteins. J Spectroscopy 2010, 24: 651–659. 10.1155/2010/706498View ArticleGoogle Scholar
- Urban L, Steinhoff HJ: Hydrogen bonding to the nitroxide of protein bound spin labels. Mol Phys 2013, 111: 2873–2881. doi:10.1080/00268976.2013.804217 10.1080/00268976.2013.804217View ArticleGoogle Scholar
- Klose D, Klare JP, Grohmann D, Kay CWM, Werner F, Steinhoff HJ: Simulation vs. reality: a comparison of in silico distance predictions with DEER and FRET measurements. PLoS One 2012, 7: e39492. 10.1371/journal.pone.0039492PubMed CentralView ArticlePubMedGoogle Scholar
- Fleissner MR, Bridges MD, Brooks EK, Cascio D, Kálái T, Hideg K, Hubbell WL: Structure and dynamics of a conformationally constrained nitroxide side chain and applications in EPR spectroscopy. Proc Natl Acad Sci U S A 2011, 108: 16241–16246. 10.1073/pnas.1111420108PubMed CentralView ArticlePubMedGoogle Scholar
- Toledo Warshaviak D, Khramtsov VV, Cascio D, Altenbach C, Hubbell WL: Structure and dynamics of an imidazoline nitroxide side chain with strongly hindered internal motion in proteins. J Magn Reson 2013, 232: 53–61.View ArticlePubMedGoogle Scholar
- Battye TGG, Kontogiannis L, Johnson O, Powell HR, Leslie AGW: iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM. Acta Crystallogr D Biol Crystallogr 2011, 67: 271–281. 10.1107/S0907444910048675PubMed CentralView ArticlePubMedGoogle Scholar
- Evans P: Scaling and assessment of data quality. Acta Crystallogr D Biol Crystallogr 2005, 62: 72–82.View ArticlePubMedGoogle Scholar
- McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC, Read RJ: Phasercrystallographic software. J Appl Crystallogr 2007, 40: 658–674. 10.1107/S0021889807021206PubMed CentralView ArticlePubMedGoogle Scholar
- Adams P, Grosse-Kunstleve R, Hung L, Ioerger T, McCoy A, Moriarty N, Read R, Sacchettini J, Sauter N, Terwilliger T: PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr D Biol Crystallogr 2002, 58: 1948–1954. 10.1107/S0907444902016657View ArticlePubMedGoogle Scholar
- Emsley P, Cowtan K: Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 2004, 60: 2126–2132. 10.1107/S0907444904019158View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.