CaMKII is an abundant synaptic protein strongly implicated in plasticity. Overexpression of autonomous (T286D) CaMKII in CA1 hippocampal cells enhances synaptic strength if T305/T306 sites are not phosphorylated, but decreases synaptic strength if they are phosphorylated. It has generally been thought that spine size and synaptic strength covary; however, the ability of CaMKII and its various phosphorylation states to control spine size has not been previously examined. Using a unique method that allows the effects of overexpressed protein to be monitored over time, we found that all autonomous forms of CaMKII increase spine size. Thus, for instance, the T286D/T305D/T306D form increases spine size but decreases synaptic strength. Further evidence for such dissociation is provided by experiments with the T286D form that has been made catalytically dead. This form fails to enhance synaptic strength but increases spine size, presumably by a structural process. Thus very different mechanisms govern how CaMKII affects spine structure and synaptic function. S tudies have shown that there are structural correlates of synaptic plasticity. Photolysis of caged glutamate at individual spines produces long-term potentiation (LTP) and is accompanied by a persistent increase in spine size (1, 2). Similarly, the induction of long-term depression (LTD) causes spine shrinkage (3). Moreover, the spines on a dendrite vary dramatically in size (4) in a manner that correlates with the postsynaptic strength of the synapse on that spine (5, 6).CaMKII is one of the most abundant proteins in the brain (7). It is activated during LTP (8, 9), an activation that is required for potentiation (10-12). This activation is dependent on phosphorylation at T286, which makes the enzyme autonomous of Ca 2+ /calmodulin (13). The instructive role of CaMKII in the control of synaptic strength has been examined by overexpression of various phosphorylation states of the kinase. Recent work (14) revealed that the effect of the autonomous form (T286D) has unexpected complexity that depends on additional phosphorylation sites (T305/T306). These sites have been previously implicated in metaplasticity (15). If these sites are not phosphorylated (T286D/T305A/T306A), potentiation is produced that occludes LTP (14). However, if the sites are pseudophosphorylated (T286D/T305D/T305D), depression is produced that occludes LTD (14). Given the correlation of synaptic strength and spine size noted above, we expected that the forms of CaMKII that increase synaptic strength would increase spine size, whereas those that decrease synaptic strength would decrease spine size. We tested this prediction and found that it is not correct. Autonomous CaMKII forms increase spine size regardless of whether synaptic strength is increased or decreased. The mechanisms by which CaMKII affects spine structure and synaptic strength thus have a different dependence on the phosphorylation state of the enzyme. This difference is underscored by a further finding: that a T286D form with an ...