Docking method

Previous experimental studies showed varying degrees of translocation defects of SRP-independent and SRP-dependent substrates in SEC62 mutant, SEC63 mutant and SEC72 deletion strains [14, 15, 17, 50]. The hydrophobic cores of the signal sequences of SRP-dependent substrates possess a substantially greater average hydrophobicity than SRP-independent substrates [17, 51]. Therefore, the aim of the docking part was to investigate the interaction of the hydrophobic cores of signal sequences with Sec61 in its conformation when bound to the Sec62-Sec63 complex (Sec62/Sec63/Sec71/Sec72), which is believed to reflect its conformation during post-translational protein translocation. We selected the hydrophobic cores of the signal sequences of the Sec61 substrates CPY, Gas1, PDI, ppα-factor, Dap2 and Pho8 (see S1 Table) for our study. CPY, Gas1, PDI and ppα-factor are SRP-independent substrates [17] with N’-terminal cleavable signal sequences. Dap2 and Pho8 are SRP-dependent substrates [17] and contain a noncleavable transmembrane signal sequence (signal-anchor sequence). As the signal-anchor sequences of Dap2 and Pho8 are longer than 50 amino acids docking of the full signal sequences is not feasible due to limitations of the AutoDock CrankPep (ADCP) software posed on the maximal length of the peptide [52].

The homology-based structural model of Sec61 described above was used for docking with ADCP. ADCP makes use of CRANKITE [53], a software package that samples the conformational space of proteins using a Metropolis Monte Carlo method. In ADCP, CRANKITE is combined with the grid-based AutoDock representation of a rigid receptor to optimize peptide conformations and peptide-receptor interactions at the same time. Thus, we adopted a relatively large grid box of dimensions 90 Å X 90 Å X 126 Å, centered at the center of mass of Sec61 to cover the channel pore. Three independent sets of docking runs were conducted that were started with different random seeds. All clustered pockets within the box with energy below 0 kcal/mol were considered as potential binding pockets. The number of replicas for each docking run was set to 100 with 1.75 x 10⁸ evaluations of the scoring function for each replica. The partition parameter was set to 20 meaning that 20% of the replicas were started with the peptide sequence modeled initially into a helical conformations and the remaining 80% were started from an extended conformation. All other docking parameters were kept at the default values of ADCP. Each independent docking run provided top ten final docking poses with ranking after clustering. The top ten solutions from three independent docking runs (N = 30) were considered for analyzing the interaction between signal sequences and Sec61. The minimum distance between heavy atoms of two residues was considered as residue-residue distance. S4 Fig compares docking solutions of the full-length signal sequence of CPY to docking solutions for its hydrophobic core. For 18 out of 30 docking poses, structural matches below 2.0 Å RMSD were obtained. Pose #2 has the lowest RMSD of 0.76 Å from pose #5 of the hydrophobic core. At least for the relatively short CPY signal sequence, this exercise showed that docking of the hydrophobic core identified poses that are representative of the full length signal sequence.