Abstract
Ca(2+)/calmodulin-dependent protein kinase II (CaMKII) is essential for long-term potentiation (LTP) of excitatory synapses, a process fundamental to learning. CaMKII responds to Ca(2+) influx into postsynaptic spines by phosphorylating proteins and forming new protein interactions. The relative importance of these enzymatic and structural functions is debated. LTP induction triggers CaMKII docking to NMDA receptors, and recent evidence indicates that LTP can proceed without kinase activity after this event. Furthermore, interactions between CaMKII and α-actinin-2 that form following LTP induction are required for dendritic spine enlargement. CaMKII can autophosphorylate at T286, which enables autonomous activity after Ca(2+)/CaM dissociation. Experiments with CaMKII variants including a T305A/T306A ("AA") double substitution have led to a model whereby T305/T306 phosphorylation by autonomously active CaMKII prevents further Ca(2+)/CaM activation. However, this mechanism is not fully compatible with previous studies including a phosphoproteomic analysis of CaMKII and imaging using CaMKII activity reporters in live neurons. In this study, we show using rat hippocampal cultures that the AA substitution has an unintended gain-of-function property: elevated binding to α-actinin-2 in unstimulated neurons to a level only normally observed after induction of LTP. CaMKIIα AA also increases the proportion of enlarged spines in unstimulated neurons without altering synaptic currents. Calorimetric measurements with purified protein confirm that α-actinin-2 binds tightly to CaMKIIα AA with no requirement for kinase activation. Using x-ray crystallography, we show that the AA substitution enables α-actinin-2 to adopt a different tighter binding mode. Our findings reinforce the notion that CaMKII primarily fulfils a structural role in LTP.