Ca²⁺/calmodulin-dependent protein kinase II (CaMKII), which is present in high concentrations in the brain, contributes to many forms of synaptic plasticity. The induction of synaptic plasticity by CaMKII involves an intracellular signalling cascade that links neuronal Ca²⁺ signals with the phosphorylation of neurotransmitter receptors; an important step in this biochemical cascade is the autophosphorylation of CaMKII after binding of Ca²⁺/calmodulin (Ca₄-CaM).

The dependence of this autophosphorylation reaction on the temporal structure of Ca₄-CaM signals has been investigated in previous experiments [1] and computer simulations [2]. These experimental and theoretical studies have indicated that the autophosphorylation of CaMKII is sensitive to the frequency of repetitive Ca²⁺ pulses, and it has been concluded that CaMKII can decode oscillatory Ca²⁺ signals [1, 2].

Here, we apply a simplified version of the commonly used CaMKII activation model by Dupont and collaborators [2] to investigate the mechanism that underlies the dependence of the overall autophosphorylation kinetics on the frequency of Ca²⁺ oscillations. In the simulations by Dupont et al., CaMKII was subjected to different average, or 'effective', Ca₄-CaM concentrations, which in turn affected the average concentration of the CaMKII subunits, and the autophosphorylation kinetics.

We first replicate the simulation results presented in [2] with our simplified model (Figure 1A). To identify the mechanism that underlies the observed frequency dependence, we then rescale the Ca₄-CaM concentrations to an equal effective concentration, and compare the phosphorylation kinetics (Figure 1B). We demonstrate that in our model the overall phosphorylation rate under sustained application of Ca₄-CaM pulses depends on the average (‘effective’) concentration of Ca₄-CaM in the system, rather than on the pulse frequency itself. Moreover, we show that the application of a constant level of Ca₄-CaM with the same mean concentration as in the pulsed protocol results in the same level of CaMKII phosphorylation.

Our simulation results indicate that the notion of CaMKII as a decoder of Ca²⁺ oscillations is misleading and suggest experimental tests with rescaled Ca₄-CaM concentrations.