Computational protocol: Crystal Structure of Interleukin-6 in Complex with a Modified Nucleic Acid Ligand

Similar protocols

Protocol publication

[…] Two native data sets were collected that provided the basis of the two refined IL-6·SOMAmer crystal structures. The form 1 crystal was grown from 31% PEG 3350, 180 mm LiNO3, 100 mm NaOAc (pH 5.5), and 2.5% (w/v) hexamine cobalt chloride, belonged to space group P32, and had unit cell dimensions of a = b = 50.23 Å and c = 103.63 Å. The form 2 crystal was grown from 31% PEG 3350, 90 mm MgSO4, 90 mm NaOAc (pH 5.5), and 100 mm LiCl, also belonged to space group P32 but had unit cell dimensions of a = b = 69.02 Å and c = 108.47 Å. The distinction between form 1 and form 2 crystals is that the larger form 2 unit cell accommodates two IL-6·SOMAmer complexes per asymmetric unit, whereas form 1 only contains one heterodimer per asymmetric unit. Both native data sets could readily be solved by molecular replacement using the program Phaser in the CCP4 software suite using IL-6 PDB coordinates (chain B of PDB code 1P9M) as a search model. However, the contribution of the IL-6 model alone was not sufficient to provide enough phasing power to elucidate interpretable maps of the electron density for the bound SOMAmer.To obtain additional phasing information, crystals were soaked with iodide and cesium salts for the purpose of conducting single wavelength anomalous dispersion (SAD) experiments. Iodide and cesium ions have nearly identical anomalous scattering properties, give strong anomalous signals (especially at the relatively low energy produced by in-house x-ray equipment), and can be readily soaked into most types of protein crystals. Full data sets on iodide-soaked and cesium-soaked crystals were collected to ∼3 Å and used to confirm that each ion could provide several partially occupied ion-binding sites and a sizable anomalous signal. For structure solution, a crystal soaked for 20 min in crystallization reservoir solution supplemented with 500 mm LiI and 500 mm CsCl was examined at ALS beamline 5.0.3 (Berkeley, CA) on December 21, 2010. The wavelength of the x-rays used to collect the data was 0.9765 Å. The data were scaled to a resolution of 2.4 Å. Anomalous difference maps were calculated from the data to locate the positions of bound heavy atoms. Twelve potential sites were identified by this method. The heavy atom positions were then input, along with the IL-6 molecular replacement search model, into the program Phaser, using the “SAD plus MR” mode. This program used phasing information calculated from heavy atom positions and phasing information from the molecular replacement model to calculate phases and electron density maps. The SAD plus MR maps provided a clearer view of a portion of the SOMAmer and allowed the building of the model of the SOMAmer to begin.Following the initial phasing of the structure, a process of “bootstrapping” was undertaken, whereby two or three nucleotide residues were built into the model, and then the larger, improved model was used as the input model for the SAD plus MR algorithms of Phaser. In each successive round, the improved MR model improved phases and thereby provided clearer electron density maps. This process continued until ∼15 nucleotide residues were built, at which time successive rounds of refinement and molecular replacement seemed to provide no improvement to the electron density maps (D). At this point, the partial IL-6·SOMAmer complex model was moved into the 2.4 Å native data set and used for simple molecular replacement. Because the 2.4 Å “form 1” native data set was higher quality data (with an Rmerge more than 2% lower than for the CsI soak), the resulting electron density maps had less noise and were easier to interpret. Model building continued with refinement against the native form 1 data until a final model was obtained. The form 1 model was then used for molecular replacement in the 2.55 Å native form 2 data set. Two residues of the SOMAmer, which were disordered in the form 1 structure, could be resolved in form 2 due to a distinct crystal packing interaction at the hairpin loop of the SOMAmer. Data collection and refinement statistics for both form 1 and form 2 structures are listed in . Structure analysis was performed using PyMol () and web3DNA (). […]

Pipeline specifications

Software tools CCP4, PyMOL
Application Protein structure analysis
Chemicals Deoxyuridine