Structural analysis and stochastic modelling suggest a mechanism for calmodulin trapping by CaMKII

Abstract

Activation of CaMKII by calmodulin and the subsequent maintenance of constitutive activity through autophosphorylation at threonine residue 286 (Thr286) are thought to play a major role in synaptic plasticity. One of the effects of autophosphorylation at Thr286 is to increase the apparent affinity of CaMKII for calmodulin, a phenomenon known as "calmodulin trapping". It has previously been suggested that two binding sites for calmodulin exist on CaMKII, with high and low affinities, respectively. We built structural models of calmodulin bound to both of these sites. Molecular dynamics simulation showed that while binding of calmodulin to the supposed low-affinity binding site on CaMKII is compatible with closing (and hence, inactivation) of the kinase, and could even favour it, binding to the high-affinity site is not. Stochastic simulations of a biochemical model showed that the existence of two such binding sites, one of them accessible only in the active, open conformation, would be sufficient to explain calmodulin trapping by CaMKII. We can explain the effect of CaMKII autophosphorylation at Thr286 on calmodulin trapping: It stabilises the active state and therefore makes the high-affinity binding site accessible. Crucially, a model with only one binding site where calmodulin binding and CaMKII inactivation are strictly mutually exclusive cannot reproduce calmodulin trapping. One of the predictions of our study is that calmodulin binding in itself is not sufficient for CaMKII activation, although high-affinity binding of calmodulin is.

Figure 1. Opening of a CaMKII subunit.

Overlay of 100 model structures created with MODELLER, where structural information was omitted for the four residues linking the kinase domain of CaMKII with the inhibitory helix. These four residues are shown in yellow. The structure corresponding to the published structure of the kinase domain (with the linker region intact, PDB ID: 2BDW, chain A) is shown in red.

Figure 2. Opening of a CaMKII subunit: Mean RMSD per residue.

Average root-mean-square deviation (RMSD) per residue for the structures shown in figure 1. RMSD values were computed using Chimera .

Figure 3. Models of Calmodulin bound to the low-affinity site of a CaMKII subunit.

Left panel: Binding to a CaMKII subunit in a closed conformation. Right panel: Binding to a CaMKII subunit with flexible linker. Although the linker region between inhibitory helix and catalytic domain was flexible, the complex CaMKII-calmodulin still favoured a closed conformation of CaMKII. Calmodulin is shown in red, CaMKII in grey. The Thr286 autophosphorylation site is shown in teal, the region that links the catalytic domain to the autoinhibitory helix in yellow.

Figure 4. Interactions between calmodulin and its low-affinity binding site on the kinase domain of CaMKII.

Interacting residues are shown as sticks. Calmodulin is shown in red and CaMKII in grey.

Figure 5. High-affinity binding of calmodulin to the open state of CaMKII.

Calmodulin is shown in red, CaMKII in grey and the autophosphorylation site at Thr286 in teal.

Figure 6. Residues crucial for calmodulin trapping.

Left panel: In the high-affinity structure, residue Met124 on calmodulin (in red) makes contact both with residue Glu120, also on calmodulin and with residue Phe293 on CaMKII (in grey). Right panel: In the low-affinity structure, these contacts are missing.

Figure 7. Interaction surfaces for low-affinity and high-affinity binding.

The interaction surfaces for calmodulin binding are projected onto a CaMKII monomer. The low-affinity interaction surface is shown in blue, the high-affinity interaction surface in green, and the overlap in cyan. Interaction surfaces were computed using Chimera with a Å cutoff.

Figure 8. Model of calmodulin trapping by CaMKII.

The model is shown as an SBGN ER diagram . For clarity, only one monomeric subunit is shown. In the actual model, six such subunits form a ring, and autophosphorylation at Thr286 of one subunit is dependent on the neighbouring subunit being open.

Figure 9. Several trapping simulations on wildtype and mutant CaMKII.

Calmodulin is inactivated, mimicking calcium withdrawal after . The ratio of calmodulin to CaMKII concentration used in the simulation was the same as used in the experimental setup by Meyer et al. . The number of calmodulin-bound monomeric CaMKII subunits is plotted against time for each simulation run. The total number of CaMKII subunits in the simulation was 360. Wildtype is shown in black, T286A mutant in red. Ten simulation runs are shown for each.

Figure 10. Calmodulin trapping on the level of single subunits.

Each individual panel represents a single subunit chosen at random from the simulations from those subunits that bind calmodulin. The x-axis represents time, going from to . The three levels in the y-dimension represent calmodulin binding, with no binding (lowest level), low-affinity binding (middle level) and high-affinity binding (highest level). Events of calmodulin sliding back and forth between the high-affinity and the low-affinity binding sites appear as drops from the top level to the centre and back up. The colour of the trace represents subunit phosphorylation at Thr286, with unphosphorylated subunits shown in black, and phosphorylated subunits in red.

Figure 11. Comparison of the two-binding-site model with a one-binding site model.

For each of the models, we ran ten simulations with wildtype CaMKII and ten simulations with the autophosphorylation deficient T286A mutant. All CaMKII molecules are open and fully saturated with calmodulin to begin with, and calmodulin is withdrawn, mimicking calcium withdrawal, at the start of the simulation. The ratio of calmodulin to CaMKII concentrations was the same as used in the experimental setup by Meyer et al. . The number of calmodulin-bound monomeric CaMKII subunits is plotted against time. Wildtype is shown in black, T286A mutant in red. There is no difference in slope between mutant and wildtype in the one-binding-site model, whereas the two-binding-site model displays a clear difference between wildtype and mutant.