CASK and CaMKII function in Drosophila memory

Abstract

Calcium (Ca(2+)) and Calmodulin (CaM)-dependent serine/threonine kinase II (CaMKII) plays a central role in synaptic plasticity and memory due to its ability to phosphorylate itself and regulate its own kinase activity. Autophosphorylation at threonine 287 (T287) switches CaMKII to a Ca(2+) independent and constitutively active state replicated by overexpression of a phosphomimetic CaMKII-T287D transgene or blocked by expression of a T287A transgene. A second pair of sites, T306 T307 in the CaM binding region once autophosphorylated, prevents CaM binding and inactivates the kinase during synaptic plasticity and memory, and can be blocked by a TT306/7AA transgene. Recently the synaptic scaffolding molecule called CASK (Ca(2+)/CaM-associated serine kinase) has been shown to control both sets of CaMKII autophosphorylation events during neuronal growth, Ca(2+) signaling and memory in Drosophila. Deletion of either full length CASK or just its CaMK-like and L27 domains removed middle-term memory (MTM) and long-term memory (LTM), with CASK function in the α'/ß' mushroom body neurons being required for memory. In a similar manner directly changing the levels of CaMKII autophosphorylation (T287D, T287A, or TT306/7AA) in the α'/ß' neurons also removed MTM and LTM. In the CASK null mutant expression of either the Drosophila or human CASK transgene in the α'/ß' neurons was found to completely rescue memory, confirming that CASK signaling in α'/β' neurons is necessary and sufficient for Drosophila memory formation and that the neuronal function of CASK is conserved between Drosophila and human. Expression of human CASK in Drosophila also rescued the effect of CASK deletion on the activity state of CaMKII, suggesting that human CASK may also regulate CaMKII autophosphorylation. Mutations in human CASK have recently been shown to result in intellectual disability and neurological defects suggesting a role in plasticity and learning possibly via regulation of CaMKII autophosphorylation.

Keywords: CASK; CaMKII; Drosophila; autophosphorylation; calcium imaging; disease model; memory; mushroom body.

Figure 1

A model of CASK’s regulation of CaMKII autophosphorylation during memory formation. (A) The large colored rectangle represents a hypothetical neuron in the middle of which is a cartoon of a single layer of a CaMKII dodecamer holoenzyme. On the right, under conditions of increased synaptic activity (high [Ca²⁺], in red) Ca²⁺/CaM binds CaMKII via the CaM binding site that contains the inhibitory T306 T307 sites hence blocking them from autophosphorylation. This also promotes T287 autophosphorylation (pT287) and the switch to persistently high kinase activity even after Ca²⁺ levels fall. On the left, under conditions of low synaptic activity and low [Ca²⁺] (in blue), there is low probability of CaM binding to CaMKII allowing CASK to promote autophosphorylation of the inhibitory T306 T307 (pT306 pT307) sites. This renders the kinase inactive and even if there is a subsequent increase in Ca²⁺/CaM, CaM binding is blocked by pT306 pT307 in the CaM site. Eventually phosphatases will act to remove phosphorylation events and return endogenous CaMKII to its basal state. Therefore, in the absence of CASK there is a decrease in inhibitory pT306 pT307 and an increase in pT287 constitutively active CaMKII, conversely increased CASK promotes inhibitory pT306 pT307 decreasing pT287 and endogenous CaMKII activity. Therefore, neurons expressing transgenic CaMKII with inhibitory phosphorylation sites mutated to blocking residues (TT306/7AA) or with too little CASK due to mutation results in a form of CaMKII that is unable to switch off. This causes abnormally high CaMKII activity that subsequently interferes with the physiology of the neuron disrupting memory. (B) Predicted domain structure of CASK isoforms, the short isoform CASK-α contains PDZ, SH3, and GUK domains while the long isoform CASK-β contains additional CaMK-like (CamK), Calmodulin binding domain (CaMBD), and L27 domains at its N-terminus. The CASK-β null contains a N-terminal deletion removing CaMK, CaMBD, and L27 domains but leaves the downstream promoter and whole of CASK-α intact (Slawson et al., 2011).

Figure 2

CASK and CaMKII autophosphorylation function in the mushroom body α′/ß′ neurons during middle and long-term memory formation. (A) A cartoon representation of a frontal section of the adult Drosophila brain showing subdivision of the memory circuit using Gal4 promoters that express in subsets of mushroom body neurons. Olfactory information (CS, conditioned stimulus) is relayed via the antennal nerve (AN) from the olfactory receptor neurons (the first order neurons) to the antennal lobe (AL, dark blue). This information is received in the glomeruli in the antennal lobe which represent the dendrites of the second order neurons, the projection neurons (PNs) send information to the higher brain centers: the mushroom body (the large lobed structures in the center) and the lateral horn (LH) neurons. The mushroom body consists of about 2000 neurons called Kenyon cells whose soma are depicted as small light blue circles. Drosophila mushroom bodies consist of three different classes of intrinsic neurons (α/β, α′/β′, and γ) that extend their axons into the five lobes of neuropil that are bilaterally symmetrically arranged in the center of the fly brain (Davis, 2011). (B) The OK107-Gal4 promoter expresses in all mushroom body neurons and a number of neuronal regions outside the mushroom body (in red, Connolly et al., 1996). (C) While c305a-Gal4 (in purple) promoter expresses in the mushroom body α′/β′ neurons as well as other regions (Krashes et al., ; Pech et al., 2013) and (D) MB247-Gal4 (in green) expresses in mushroom body α/β and γ neurons (Zars et al., 2000). (E) By measuring memory at different times after training, memory retention curves readily depict the effect of each genotype on memory performance (Performance index). Learning or initial (2 min) short-term memory (STM) was measured immediately after administering one cycle of shock-odor training, no statistical difference in learning was seen between CASK and CaMKII genotypes with wildtype (CSw- in black). MTM measured 3 h post-one cycle training was completely removed in CASK-β null (light blue dashed line) flies. Likewise targeted expression of CASK-RNAi (downward triangle), constitutively active CaMKII-T287D (black circle with cross in), Ca²⁺ dependent CaMKII-T287A (square) and uas-CaMKII-TT306/7AA (triangle, both inhibitory sites blocked) throughout the mushroom body (OK107-Gal4, red) or just the α′/β′ neurons (c305a-Gal4, purple) was sufficient to cause the reduction in MTM and LTM compared to control. Expression of CASK and CaMKII transgenes in the remaining mushroom body α/β and γ neurons (MB247-Gal4, green) had little effect. Flies null for CASK-β or overexpressing CASK-RNAi, CaMKII-T287D, CaMKII-T287A, or CaMKII-TT306/7AA throughout their mushroom body or just the α′/β′ neuron completely lacked LTM induced by five cycles of spaced training. Mushroom body α/β and γ neuron expression of CASK and CaMKII transgenes did not affect LTM.

Figure 3

CASK and CaMKII regulate dynamic changes in neural activity in mushroom body α′/ß′ neurons. (A) Color coded images of a fly brain showing GCaMP fluorescence in the mushroom body α′/ß′ lobes using c305a-Gal4 before and after application of depolarizing high [KCl]. (B) Traces showing averaged GCaMP fluorescence overtime in the α′/ß′ lobes (c305a-Gal4) co-expressing the different CASK and CaMKII transgenes or CASK-β null compared to the control c305a/+ expressing GCaMP (solid black line). GCaMP fluorescence is reduced in CASK-β null (dotted orange line) and when CASK-RNAi (purple line), CaMKII-RNAi (green line), or CaMKII-T287D (yellow line) were expressed in the α′/ß′ neurons, while CaMKII overexpression (blue line) increased the maximum response compared to control.