Computational protocol: Unique architecture of thermophilic archaeal virus APBV1 and its genome packaging

Similar protocols

Protocol publication

[…] A 3-μl aliquot of sample was applied to a glow-discharged grid (C-flat; Protochips, Raleigh, NC) at ambient temperature and ~80% relative humidity and vitrified by plunge-freezing into liquid ethane using a vitrification apparatus (CP3, Gatan, Pleasanton, CA). Data were acquired using a 300-kV transmission electron microscope (Tecnai F30 ‘Polara’; FEI) equipped with an energy filter (slit width 20 eV; GIF Quantum, Gatan) and a direct electron detector (K2 Summit, Gatan). Movies (22 frames, each frame 0.2 s) were collected in electron counting mode at dose rate of 8 e– pixel−1 s−1 at calibrated magnification of 37,037× resulting in a total dose of ~22 e– Å−2 and pixel size of 1.35 Å.Movie frames were aligned to account for drift using Motioncorr and contrast transfer function (CTF) parameters were estimated using CTFFIND3. A total of 400 drift-corrected micrographs were used to pick 3145 virus particles using e2helixboxer.py from EMAN2. Helical reconstruction was calculated in Spring. Particles were segmented into overlapping segments, which were treated as single particles. Initial helical parameters were derived by layer-line analysis of 2D class averages to assign layer-line heights and Bessel orders. The parameters were refined by calculating amplitude correlation for a set of different parameters around the original ones. The final map was calculated using iterative real-space gold-standard refinement.The ends of the virus particle were reconstructed using Relion. An initial estimate for the in-plane rotation (psi) was calculated from the original coordinates of the helix ends and all filament ends were subjected to 2D classification in a template-free manner. Centering of the helix ends was improved by manually defining the particle centers in the class averages, which allowed us to calculate more accurate particle coordinates in the micrographs and re-extract them for a second round of 2D classification. The 2D classification separated the filament ends into a set of pointed ends and a set of rounded ends. For 3D refinement of particles in both sets, initial models were calculated using relion_reconstruct. First, out-of-the-plane rotation (tilt) was set to 90° and the angle around the long axis of the filament (rot) was set to a random value. Gold standard local refinement was run for both sets restricting deviations from the initial psi and tilt angles by sigma value of 2° and 4°, respectively. Initial refinements run without applying symmetry confirmed the presence of C5 symmetry in both ends and C5 symmetry was applied in subsequent refinements.Resolution of the models was estimated by using Fourier shell correlation in Relion using 0.143-cutoff and phase randomization to account for masking effects. Local resolution was estimated using ResMap. Reconstruction statistics are summarized in Supplementary Table .A 2D class average of the complete particle was created in Relion. Extracted particles were downsampled by factor of 2× resulting in a box size of 700 pixels and pixel size of 2.7 Å pixel−1. The final average of particles in the best class was calculated in relion_reconstruct up to 1/8 Å−1 spatial frequency. [...] An atomic model for VP1 in the helical part of the filament was traced with Buccaneer, built with Coot, and refined in Refmac and Phenix applying secondary structure, rotamer, and Ramachandran plot restraints. The ends of the filament were modeled in COOT by jelly-body fitting several copies of the VP1 atomic model. The refinement and geometry of the models were validated with Refmac and MOLPROBITY. Model refinement statistics are summarized in Supplementary Table . […]

Pipeline specifications

Software tools MotionCorr, CTFFIND, EMAN, RELION, ResMap, Buccaneer, Coot, PHENIX, MolProbity
Applications cryo-EM, Protein structure analysis
Organisms Viruses, Human poliovirus 1 Mahoney, Aeropyrum pernix