Computational protocol: Computational Exploration of Structural Hypotheses for an Additional Sequence in a Mammalian Mitochondrial Protein

Similar protocols

Protocol publication

[…] The cryo-EM density maps of the E. coli 70S initiation complex (EMD 1248) , the T. thermophilus 30S initiation complex (EMD 1523) , and IF2mt bound to the E. coli ribosome were used for generating the initial IF2mt models. The crystal structure of the archaeal IF2 (PDB 1G7T) from M. thermoautotrophicum was manually docked in corresponding IF2 electron densities in each map using Chimera . The crystal structure of the 30S subunit from T. Thermophilus with bound IF1 (PDB 1HRO) was similarly manually docked in the electron density map regions corresponding to the small subunits for both the E. coli and the T. thermophilus ribosomes to obtain the binding site and orientation of bacterial IF1 in both ribosomes. The crystal structure of the 70S E. coli ribosome was also manually docked into the cryo-EM map of both E. coli translation initiation complexes.Multiple sequence alignment of the representative eukaryotic IF2mt sequences near the insert region were generated using ClustalW2 . A pair-wise sequence alignment of the IF2mt sequence with the archaeal IF2 sequence was also generated using ClustalW2 and was manually adjusted to correct for any obvious misalignments. Since bovine IF2mt is 727 aa and archaeal IF2 is 594 aa, the first 175 aa residues in IF2mt, which represent the mitochondrial signal sequence (residues 1 to 77) and domain III (residues 78 to 175) that are absent in the archaeal IF2, were removed. In addition, by empirically removing additional residues on either side of the 37 aa insert region in IF2mt as compared to E. coli IF2, it was observed that removing an additional three amino acids on the N-terminal side and nine amino acids on the C-terminal side of the 37 aa insert improved the sequence alignment between IF2mt and archaeal IF2, yielding a slightly larger 49 aa insert region as compared to E. coli IF2. Alignment of the C1 and C2 sub-domains in domain VI of IF2mt to their corresponding C1 (PDB 1Z9B, ) and C2 (PDB 1D1N, ) regions in the B. stearothermophilus IF2 sequence yielded higher sequence homology as compared to archaeal IF2. Removal of the initial 35 amino acids from the N-terminal end of the C1 sub-domain of B. stearothermophilus IF2 yielded the best sequence alignment.These individual sequence alignments and the corresponding PDB files (1G7T,1Z9B and 1D1N) were used to build initial homology models for the respective bovine IF2mt sequence regions using the program MODELLER . The I-TASSER server was used to build a separate ab initio model for the 49 aa insert region which was then manually aligned to the crystal structure of IF1 bound to the T. thermophilus 30S subunit using the program VMD . The models for the VI-C1 and VI-C2 sub-domains were structurally aligned to the corresponding regions in the overall homology model based on the archaeal crystal structure using the program RAPIDO . A composite model of IF2mt was then generated by connecting the 49 aa insert, and the C1 and C2 sub-domains based on B. stearothermophilus NMR structures , to the rest of the IF2mt modeled on the archaeal crystal structure using the program LOOPY with the CHARMM22 protein parameters . For linker structure prediction using LOOPY, only the minimal number of residues at each junction required to get the different domains covalently connected were randomized. To connect the 49 aa insert, VI-C1, and VI-C2 sub-domains to the main body of IF2mt, the minimal regions that required randomization and linker structure prediction with LOOPY were residues 464–473, 595–600, and 615–626, respectively. This initial IF2mt model was then optimized using the program CHARMM by multiple rounds of 5,000 steps of steepest descent (SD) and adopted basis newton raphson (ABNR) minimizations followed by 5,000 steps of room temperature langevin dynamics with a high friction coefficient (60 ps−1) in the presence of gradually reducing harmonic restraints on all non-hydrogen atoms. The insert and its neighboring regions (residues 464–518) were then subjected to similar rounds of minimization and dynamics under center-of-mass restraints to allow them to relax their internal structure, while the rest of the IF2mt protein was kept under strong harmonic restraints. The alternate model with altered orientation of the α-helix formed by residues 446 to 460 was generated by manually positioning that α-helix and the ab initio I-TASSER insert sequence region model, and then connecting the linker regions to the rest of the structure using LOOPY and optimizing the structure as mentioned above. […]

Pipeline specifications

Software tools Clustal W, MODELLER, I-TASSER, VMD, RAPIDO, loopy, CHARMM
Applications cryo-EM, Protein structure analysis
Organisms Bos taurus, Homo sapiens