Computational protocol: Structure–function insights reveal the human ribosome as a cancer target for antibiotics

Similar protocols

Protocol publication

[…] Human 80S ribosomes were prepared from HeLa cells as described earlier. A volume of 2.5 μl of freshly prepared human 80S ribosomes containing CHX (4-[(2R)-2-[(1S,3S,5S)-3,5-dimethyl-2-oxocyclohexyl]-2-hydroxyethyl]piperidine-2,6-dione; 10-fold molar excess of CHX in H2O, incubated 10 min on ice), diluted from 2 mg ml−1 to 0.5 mg ml−1, was applied to 300 mesh holey carbon Quantifoil 2/2 grids (Quantifoil Micro Tools, Jena, Germany) and flash-frozen as described. Data were collected on the in-house spherical aberration (Cs) corrected Titan Krios S-FEG instrument (FEI, Eindhoven, Netherlands) operating at 300 kV acceleration voltage and at a nominal underfocus of Δz=−0.4 to −3.0 μm at a magnification of × 79,000, corresponding to 0.88 Å per pixels on the specimen level. Images were recorded using a back-thinned direct electron detector (Falcon II) 4096 × 4096 camera with dose fractionation (seven individual frames were collected, starting from the second one). Total exposure time was 1 s with a dose of 60 ē Å−2 (or 3.5 ē Å−2 per frame). Images in the stack were aligned using the whole image motion correction method described in ref. . Particles (159,000) were picked automatically using the in-house software gEMpicker and the contrast transfer function of every image was determined using CTFFIND4 (ref. ) in the RELION workflow. For the first steps of refinement, images were coarsened by 2 (C2 images) and 4 (C4 images) using EMAN2. First, we applied 3D classification to remove bad particles (119,386 particles left), followed by two-dimensional (2D) classification to remove images with ice or noise (95,917 good particles; for 2D and 3D classifications see also ref. ). After 3D refinement, we performed one more 3D classification (see workflow in ) to split rotated (24,714) and non-rotated (71,236 particles) 80S ribosome states. Both complexes contain E-site tRNA but the rotated 80S particles have eEF2, while the non-rotated particles contain no factor. 3D sorting of non-rotated states showed (after exclusion of a bad 3D class containing 1,899 particles) a class without E-site tRNA (19,026 particles) with CHX bound and the rest contained E-site tRNA (50,311 particles). To address whether the eEF2-containing complexes could have CHX bound we did an additional 3D classifications of subpopulations: we found that when eEF2 is present, the E-site tRNA is also present but there is no CHX bound; when eEF2 is absent but the E-site tRNA is present, there is also no CHX bound; there is no state with eEF2 present and tRNA absent. Thus, the CHX ligand is only present in the subpopulation that has neither E-site tRNA nor eEF2 bound. To obtain the best possible resolution on the CHX-containing 3D sub-class comprising 19,026 particles we performed 3D refinement and movie processing using uncoarsened data with a box size of 640 pixels. The post-processing procedure implemented in RELION 1.4 (ref. ) was applied to the final maps for appropriate masking, B-factor sharpening and resolution validation to avoid over-fitting; the appropriate B-factor was determined according to the procedure described. The final map of the 80S/CHX complex was fine-scaled to the previous structure and the resolution was estimated in Relion and IMAGIC at 0.143 FSC and half-bit criteria, indicating an average resolution of 3.6 Å (). Local resolution estimation with ResMap shows that many regions reach ∼3.1 Å resolution in the 80S/CHX complex.The cryo-EM maps were interpreted using Chimera, COOT to derive an atomic model of the human ribosome obtained by model building and structure refinement using Phenix. For this, the atomic model of the human ribosome was used as starting point and refined including the CHX ligand. The atomic model was refined against the experimental cryo-EM map by iterative manual model building and restrained parameter refinement protocols (real space refinement, positional refinement, grouped B-factor refinement and simulated annealing as described in Khatter et al. and Natchiar et al. (manuscript in preparation). Feature-enhanced maps were used occasionally to facilitate assignment of side-chain conformations and the atomic model was then refined against the cryo-EM map as previously described. The refinement process was monitored with R-factor values (25.1%/28.2% Rfree; initial R 40%) to avoid over-fitting, and the entire 80S structure was refined at once by simulated annealing using parallel computing. The final atomic model comprises ∼220,000 atoms (excluding hydrogens) across the 5,866 nucleotide residues and ∼11,590 amino acids of the 80 proteins and the four rRNA's (28S, 5S, 5.8S and 18S; excluding certain ES rRNA which are only partially visible at the periphery of the structure probably due to conformational heterogeneity). In addition, 226 Mg2+ ions, 32 water molecules and one CHX ligand (20 atoms) were included in the atomic model. Protein residues show well-refined geometrical parameters (allowed regions 11.8%, preferred regions 87.4% in Ramachandran plots and 0.8% of outliers). Figures were prepared using the software Chimera and Pymol (DeLano, 2006). […]

Pipeline specifications

Software tools gEMpicker, CTFFIND, RELION, EMAN, IMAGIC, ResMap, Coot, PyMOL
Applications cryo-EM, Protein structure analysis
Organisms Homo sapiens
Diseases Neoplasms