Figure 1.
Starting structures used for the MD simulations.
A, canine ATP-bound Grp94 structure, PDB entry 2O1U, with ATP lid and missing loops modeled as described in Materials and Methods. B, yeast ATP-bound Hsp90 structure, PDB entry 2CG9, with charged linker and disordered loops modeled as described in [35]. C HtpG structure, PDB entry 2IOP.pdb, corresponding to the ADP-bound form resolved in [17]. The structural domains are colored as follows: blue, N-terminal domains; Red, M-large domains; Orange, M-small domains; Yellow, C-terminal domains. The domains span the following sequence intervals: Grp94 (NTD: 85–285; M-large: 330–511; M-small: 512–602; CTD: 603–749); Hsp90 (NTD: 2–215; M-large: 264–426; M-small: 427–526; CTD: 527–676); HtpG (NTD: 1–215; M-large: 216–400; M-small: 401–490; CTD: 491–624).
Figure 2.
Representative structures of Grp94.
A, Labels ATP0 and ATP1 show the representative structures of the two most populated cluster in the Grp94-ATP simulation. Labels ADP0 and ADP2 show the representative structures of the first and third most populated cluster in the Grp94-ADP simulation. The second cluster (not shown) represents the intermediate between ADP0 and ADP2, characterized by a slight rotation of the NTDs and an initial opening of the M-domains. B, trajectory snapshots showing the time evolution of the interaction between the nucleotide and Arg448 in the ATP simulation (top) and in the ADP simulation (bottom). Snapshots are color coded according to time evolution: from blue (0 ns) to red (100 ns).
Figure 3.
Representative structures of HtpG.
Labels ATP0 and ATP1 show the representative structures of the two most populated clusters in the HtpG-ATP simulation. Labels ADP0 and ADP1 show the representative structures of the two most populated clusters in the HtpG-ADP simulation.
Table 1.
Final and Average Distances between ATP or ADP and residues important for binding and catalysis in Grp94, Hsp90 and HtpG.
Figure 4.
Distance fluctuation matrices for Grp94, Hsp90 and HtpG protomers in all different ligand states.
The magnitude of pairwise distance fluctuations is color coded from white (small fluctuations) to black (large fluctuations). The different absolute scale of pairwise fluctuations between Grp94, Hsp90 and HtpG reflects the different internal mobility of the three molecules.
Figure 5.
Left panel, the residue based profile of average strain calculated over the time interval 20–100 ns for the three proteins in ATP-bound state (blue line) and in the ADP-bound state (red line). Right panel, protein monomers with average strain maxima shown in blue (ATP case) and red (ADP case).
Figure 6.
Fitting of the meta-trajectory with only rigid-body movements.
Two orthogonal views of the template protomer used to fit all configurations of the Grp94, Hsp90 and HtpG meta-trajectory using only rigid-body movements of the two-terminal quasi rigid domains (black for N-term and red for C-term) respect to the middle (blue) one. The protomer is 525 amino acids long and is the structure closest to the average configuration of the full meta-trajectory. The optimal rotation axes are shown as cyan cylinders. The position and orientation of these axes are identified with an iterative procedure to allow the N-term and the C-term domains to perform those movements which best fit, on average, all frames in the meta-trajectory (see also Figure 7 for further details). It is worth noting that the axes are not forced a priori to pass through one of the two boundary residues at the interface between two domains.
Figure 7.
Different views of the optimal rotation axes identified for each part of the meta-trajectory corresponding to a specific chaperone, separately.
The template structures have different configurations, each one being the closest to the corresponding chaperone's protomer average structure.