Similar protocols

Protocol publication

[…] Freshly prepared samples were adjusted to 1.5 A260 (50 nM 40S ribosomes) and applied to Quantifoil R3/3 holey grids pre-coated with 2 nm carbon. Data was collected on a Titan Krios TEM (FEI Company) equipped with a Falcon II direct electron detector at 300 keV under low dose conditions of about 2.4 e-/Å2 per frame for 10 frames in total using the software EM-TOOLS (TVIPS) and a defocus range of −0.8 to −2.4 μm at a pixel size of 1.08 Å. Original image stacks were summed up and corrected for drift and beam-induced motion at micrograph level using MotionCor2 (). The contrast transfer function parameters and resolution range of each micrograph were estimated by GCTF (). All 2D and 3D classifications and refinements were performed with RELION-2 () after automated particle picking by Gautomatch (http://www.mrc-lmb.cam.ac.uk/kzhang/). Two-dimensional reference-free classification was performed to screen for particle quality (), non-ribosomal particles as well as poorly resolved 2D classes were discarded. 266.800 particles from good classes were selected for 3D refinement using a mature 40S ribosome as reference. We performed two subsequent rounds of 3D classification in order to enrich pre-40S complexes (). First, the whole dataset was classified into 7 classes: class 1 and 2 contained orientation biased 40S ribosomes whereas classes 3 to 5 showed well-resolved 40S ribosomes with strong extra densities for the RBFs. In addition, class 6 showed poorly resolved pre-40S ribosomes and class 7 displayed a pre-40S ribosome with a very flexible head domain. The classes 3 to 5 were joined for movie refinement and a second round of 3D classification (six classes). Here, class 1 displayed distorted density due to orientation bias, while class 2 and 3 showed a very strong density for the majority of RBFs. Class 4 displayed a less distorted volume than class 1 but showed an extra density emanating from the platform to the head, which is likely Nob1 (). In addition Class 5 showed weak densities for RACK1, uS3 and the Dim1(). The most promising classes 2 and 3 were joined for final refinement and used for further interpretation. This final volume contained 84.100 particles was refined to 3.6 Å (FCS = 0.143) according to the ‘gold standard’ criterion, corrected for the modulation transfer function of the Falcon two detector and sharpened by applying a negative B-factor automatically estimated by RELION-2. Local resolution was calculated from 3.5 to 8.0 Å in steps of 0.5 Å using ResMap (). [...] For model building of the pre-40S subunit the structure of the mature S. cerevisiae 40S post splitting complex was used as a template (PDB 5LL6 []). Available structures of the biogenesis factors Rio2 (PDB 4GYG []), Tsr1 (PDB 5IW7 []), Pno1 (PDB 5WYJ []) and Enp1 (PDB 5WYJ []) were first fitted as rigid bodies into the isolated and appropriately low-pass filtered electron densities using UCSF Chimera (). After rough docking, and manual adjustments where needed, flexible fitting and jiggle fitting was applied in Coot (; ). Regions not present in the available structures were modelled de novo, where the local resolution of the map allowed it (for example Tsr1, Pno1). rRNA which could, due to flexibility not be modelled with sufficient reliability, was removed from the model. In order to identify the extra density that may correspond to Ltv1, we fitted the structure of Enp1 found in the 90S pre-ribosome into the density. This fit left two rod-like densities unexplained, which we designated as extra helices 1 and 2 (EH1 and EH2). We speculated that these extra densities are either Ltv1 or an as yet unidentified part of Enp1. All models were subsequently combined and subjected to real-space refinement using PHENIX (). After PHENIX refinement, the model was further subjected to reciprocal space refinement in REFMAC v5.8 () using restraints generated by ProSMART and LIBG as previously shown (; ). Because of the difference in local resolution and to avoid overfitting, h34, Enp1/Ltv1 and Rio2 were not subjected to REFMAC refinement. The final model was validated using MolProbity (), summarized statistics are displayed in . Cross-validation against overfitting was performed as previously described (; ) for both model refinements separately. Figures were created with the PyMOL Molecular Graphics System (Version 1.7.4, Schrödinger, LLC) and with UCSF Chimera. […]

Pipeline specifications

Software tools MotionCor2, Gctf, RELION, Gautomatch, ResMap, UCSF Chimera, Coot, PHENIX, MolProbity, PyMOL
Applications cryo-EM, Protein structure analysis
Organisms Saccharomyces cerevisiae