The four mammalian CaMKII genes (alpha, beta, gamma, delta) encode more than 30 distinct CaMKII subunits due to extensive mRNA splicing. The N-terminal catalytic and regulatory domains are highly conserved across the subunit variants. CaMKII subunits assemble into dodecameric holoenzymes consisting of a stacked pair of hexameric rings due to homo- or hetero-meric interactions of their C-terminal association domains, which are more variable. Each CaMKII subunit in the holoenzyme is independently activated by binding calcium/calmodulin. Simultaneous activation of adjacent subunits results in the inter-subunit autophosphorylation at Thr286 within the regulatory domain. Thus, the level of Thr286 autophosphorylation depends on the amplitude, duration and frequency of the changes in intracellular calcium concentrations. Thr286 autophosphorylated subunits retain kinase activity even after dissociation of calcium/calmodulin, serving as an intracellular molecular “memory” of transient calcium signals. In the absence of bound calcium/ calmodulin, CaMKII autophosphorylates at Thr305 or Thr306 in the calmodulin-binding domain, which “desensitizes” the kinase to subsequent calcium signals by blocking the re-binding of calcium/calmodulin. These effects of autophosphorylation are opposed by cellular protein phosphatases. Thus, intracellular signaling via CaMKII depends on the dynamic balance between calcium concentrations, autophosphorylation and dephosphorylation. Our lab is pursuing structure-function analyses to understand the the molecular mechanisms that control CaMKII activity and subcellular targeting.

We are testing the over-arching hypothesis that the activity, localization and ultimately biological function of CaMKII are modulated by interactions with a diverse array of CaMKII-associated proteins (CaMKAPs). Our lab was the first to show that the targeting of CaMKII to neuronal postsynaptic densities is dynamically regulated by the activation and autophosphorylation of CaMKII, and we identified GluN2B subunits of the NMDA-type glutamate receptor (NMDAR) as the first CaMKAP and characterized the interactions in detail. This interaction significantly contributes to CaMKII targeting in dendritic spines, and is necessary for normal synaptic plasticity. As additional examples, we have shown that another CaMKAP, densin, can inhibit CaMKII phosphorylation of GluA1 AMPA-type glutamate receptor subunits, but not of GluN2B. In contrast, alpha-actinin can partially activate CaMKII binding to and phosphorylation of GluN2B, but not of GluA1. Our recent proteomics studies identifed over 100 proteins associated directly or indirectly with synaptic CaMKII holoenzymes, including several other proteins linked to autism spectrum disorders. Thus, we are working on the idea that numerous subpopulations of CaMKII holoenzymes, associated with distinct CaMKAPs play distinct physiological and pathophysiological roles.

Calcium entry via voltage-gated calcium channels drives numerous neuronal responses, including gene transcription and endocannabinoid synthesis. CaMKII is a key mediator of these downstream signals, but CaMKII also serves as a feedback regulator of several different types of calcium channel, generally facilitating calcium entry. We found that CaMKII directly interacts with and phosphorylates multiple types of the pore-forming alpha1 and the beta regulatory subunits in the calcium channel complexes. Ongoing studies are investigating the roles of individual phosphorylation sites and specific interactions in both feedback modulation of the channels and in downstream calcium signaling.

Calcium signaling drives many biochemical processes in dendritic spines, including synthesis of the major brain endocannabinoid, 2-arachidonylglycerol (2-AG), by diacylglycerol lipase-alpha (DGLa). We found that DGLa is another CaMKAP, and that CaMKII phosphorylates DGLa to inhibit 2-AG synthesis. Thus, in addition to activating DGLa, calcium recruits CaMKII to restrain synaptic 2-AG signaling, limiting the resulting short-term depression of synaptic transmission in the striatum, and modulating overall motor function. Ongoing studies are using lentiviral-based strategies and mouse genetics to manipulate DGLa and protein kinase activities to investigate the impact on striatal synaptic transmission and motor function.