A molecular dynamics study of Ca(2+)-calmodulin: evidence of interdomain coupling and structural collapse on the nanosecond timescale

Abstract

A 20-ns molecular dynamics simulation of Ca(2+)-calmodulin (CaM) in explicit solvent is described. Within 5 ns, the extended crystal structure adopts a compact shape similar in dimension to complexes of CaM and target peptides but with a substantially different orientation between the N- and C-terminal domains. Significant interactions are observed between the terminal domains in this compact state, which are mediated through the same regions of CaM that bind to target peptides derived from protein kinases and most other target proteins. The process of compaction is driven by the loss of helical structure in two separate regions between residues 75-79 and 82-86, the latter being driven by unfavorable electrostatic interactions between acidic residues. In the first 5 ns of the simulation, a substantial number of contacts are observed between the first helix of the N-terminal domain and residues 74-77 of the central linker. These contacts are correlated with the closing of the second EF-hand, indicating a mechanism by which they can lower calcium affinity in the N-terminal domain.

FIGURE 1

(A) Cα RMSD of calmodulin in the simulation. The thick solid and shaded lines show the RMSD with respect to the extended crystal structure (Babu et al., 1988) and calmodulin in the CaM-smMLCKp crystal structure (Meador et al., 1992), respectively. Thin solid and shaded lines show the RMSD of the N-terminal and C-terminal domains relative to the extended crystal structures using independent fits to residues 5–77 and 78–148, respectively. (B) Calcium virtual dihedral angle as a function of simulation time. (Inset) Schematic showing the definition of the VDA.

FIGURE 2

Structures from the simulation. α-Helices are shown in magenta, turns in blue, β-sheet in yellow, and coil in white. Calcium ions are shown in green. The structure of the CaM-smMLCKp complex (Meador et al., 1992) is shown for comparison, with the smMLCKp shown in dark blue. The figure was made with the program VMD (Humphrey et al., 1996).

FIGURE 3

Secondary structure of the central linker region (residues 65–92) during the simulation. (A) Secondary structure as a function of simulation time. Residue numbers and time are shown along the x and y axes, respectively. The secondary structure of each residue was calculated every 40 ps of the simulation using the DSSP program (Kabsch and Sander, 1983) and colored according to the legend. (B) Secondary structure of the central linker region as an average over the entire simulation. The percentage of total simulation time spent by each residue in each of the secondary structure types is shown. Secondary structure was calculated as in A. Colors are according to the legend, with the exception of coil, which is shown as a dashed line.

FIGURE 4

Role of selected side-chain–side-chain interactions in the compaction of CaM. The minimal distance between any two atoms of the relevant side chains is plotted as a function of simulation time. Also shown is the helicity of the central linker region (residues 65–92), based on the number of residues in an α-helical conformation. Distance and helicity scales are shown along the right and left axes, respectively. Lines are displayed according to the legend.

FIGURE 5

(Top) Radius of gyration of CaM as a function of simulation time. (Bottom) Vector length distribution of CaM in the extended crystal structure (Babu et al., 1988; PDB ID: 3CLN), the CaM-smMLCKp complex (Meador et al., 1992; PDB ID: 1CLN), and at various times during the simulation. Values in brackets give the radius of gyration in angstroms.

FIGURE 6

Side-chain–side-chain contact maps. The average minimal distance between atoms of each side-chain pair is plotted. All distances ≤5 Å are considered contacts are colored according to the legend. The large matrix shows the contacts between side chains in the extended crystal structure (upper diagonal) and as an average over the entire simulation (lower diagonal). The two smaller matrices along either axis are identical to each other and show the contacts between CaM and the peptide in the CaM-smMLCKp complex (PDB ID: 1CDL). The secondary structure of CaM is shown along both axes, with the helices and calcium binding sites labeled.

FIGURE 7

Solvent accessibility of methionine side chains in CaM from the extended crystal structure (solid bars), as an average between 15 and 20 ns of the simulation (shaded bars), and in the CaM-smMLCKp complex (open bars). Solvent accessibilities were calculated using the Double Cubic Lattice Method (Eisenhaber et al., 1995).

FIGURE 8

VGM analysis. (A) Schematic of the vector geometry mapping method. The entering, or first helix, of the EF-hand is superimposed on a reference EF-hand on the +z axis, and the corresponding position of the exiting, or second helix, is evaluated using the angles θ, φ, and ω, as shown. Reproduced with permission from Yap et al. (1999). (B) Example of an open (blue) and a closed (red) state from EF-hand II in the Ca²⁺-CaM (Babu et al., 1988) and apo-CaM (Zhang et al., 1995) structures, respectively. (C) Analysis for EF-hand I of CaM, with structures taken every nanosecond of the simulation. (Blue cylinders) 0–10 ns and (red helices) 10–20 ns. (D) Analysis for EF-hand II. (Blue cylinders) 0–5 ns and (red cylinders) 6–20 ns.

FIGURE 9

Side-chain–side-chain contacts between residues 74–80 and the rest of CaM as a function of simulation time. Contacts were calculated as in Fig. 6 and averaged over 2-ns windows of the simulation. Contact maps for successive windows proceed from left to right in the figure. The secondary structure of CaM is shown along the y axis, with the helices and calcium binding sites labeled.

FIGURE 10

Correlation of side-chain–side-chain contacts and the closing of EF-hand II. (Top) Interhelical angles from the VGM analysis (see Fig. 8). (Bottom) Number of contacts between specific sets of residues (see legend) as a function of simulation time. Distances between all atoms of the relevant side chains were calculated, and any distance ≤5 Å was considered a contact. The number of contacts was multiplied by a factor of two in some cases for ease of presentation (see legend).