Introductory Review Cellscience Reviews Vol 1 No.4 ISSN 1742-8130

Is CaMKII Forgetful?

Johannes W. Hell

CaMKII - an unusual kinase involved in learning

Ca²⁺/calmodulin-dependent kinase II (CaMKII) is remarkable for a number of reasons. It is comprised of 12 homologous subunits encoded by four different genes (for review see Colbran and Brown, 2004; Hudmon and Schulman, 2002). CaMKIIα and CaMKIIβ are the prevailing two isoforms in the forebrain, where they form mixed α/β holoenzymes. The kinase itself is one of the most abundant proteins in the brain; in cortex and hippocampus it may compriseas much as 1-2% of the total protein (Erondu and Kennedy, 1985). When activated by Ca²⁺/calmodulin, CaMKIIα undergoes autophosphorylation on threonine 286 (T286; T287 in the β isoform, see Figure 1) which results in autonomous activity independent of Ca²⁺/calmodulin. This autophosphorylation serves as a switch, leading to perpetuated CaMKII activity which may outlast the initial activation by the Ca²⁺/calmodulin signal for an extended time period (Colbran and Brown, 2004; Hudmon and Schulman, 2002). It is this latter property in combination with CaMKII’s abundance which led to the initial hypothesis that CaMKII plays a central role in learning and memory (Lisman and Goldring, 1988).

This hypothesis has gain support from a number of observations. Long-term potentiation (LTP), an increase in synaptic transmission following high frequency stimulation patterns which may last for long periods, is specific for stimulated synapses and changes the properties of their network, and thus is thought to underlie learning and memory. LTP requires Ca²⁺ influx through the NMDA-type glutamate receptor and the subsequent activation of CaMKII (Collingridge et al., 2004; Malenka et al., 1989; Malenka and Nicoll, 1999; Malinow et al., 1989), as do several forms of learning (Martin et al., 2000; Silva et al., 1992).

The main support for a critical role of CaMKII in learning stems from mice with alteration in the gene for CaMKIIα. Especially insightful were experiments with knock-in mice in which T286 was mutated to alanine (T286A, Giese et al., 1998), which resulted in a deficiency in hippocampal LTP. These mice were also unable to learn the hippocampus-dependent water maze task, a test for spatial memory, which requires the subject to find a stationary platform that is hidden underneath the opaque water surface during consecutive trials. This key observation indicated that the specific capacity of CaMKIIα to autophosphorylate itself on T286 and thereby remain active beyond the initial stimulation by Ca²⁺/calmodulin is important in hippocampal learning mechanisms. However, the impaired capacity of T286A mice to learn the water maze task prevented the assessment of the role of memory retention with this learning paradigm.

Is CaMKII’s role limited to memory formation?

To circumvent this problem, Giese and coworkers recently made use of a passive avoidance task which involves both the amygdala and the hippocampus (Irvine et al., 2005). When given the choice, mice will move from a bright to a dark environment within a few seconds. Mice received a mild electric foot shock after entering the dark compartment, and when put back into the bright compartment either 30 min or 24 hrs after the initial trial, wild type mice now remained for several minutes within the light before returning to the dark compartment. T286D mice showed no such increase in latency in re-entering the dark compartment during a second trial, either 30 min or 24 hrs after the first trial. If the foot shock was repeated during multiple trials, a few minutes apart, T286A mice ultimately acquired the same reluctance to enter the dark compartment as their wild type littermates. This reluctance was preserved 24 hrs after learning trials. T286A mice also exhibited an impairment in cued fear conditioning, which requires the amygdala only, as well as in contextual fear conditioning, which depends upon both the hippocampus and the amygdala. When fear conditioning was repeated three (cued) or five (contextual) times rather than just once, the T286A mice learned, albeit less efficiently, and retained cued and contextual memory to the same extend as litter matched wild type control mice. The authors conclude that autophosphorylation of CaMKIIα on T286 facilitates learning in its early phases but is not essential for long term storage of the acquired information.

Similarly, carefully executed experiments which tested the effect of microinjecting a CaMKII inhibitory peptide after induction of LTP in one pathway did not impair the maintenance of LTP in this pathway, although it inhibited subsequent LTP induction within a second pathway (Chen et al., 2001). Accordingly, CaMKII activity does not seem to be critical for the maintenance of LTP, paralleling the findings in memory maintenance. However, it cannot be excluded that a small fraction of CaMKII, perhaps a population associated with the NMDA receptor subunit NR2B (see below), may not have been inhibited by the peptide in the LTP studies.

Binding of CaMKII to NR2N and eag induces autonomous activity

Do these findings preclude a role of sustained CaMKII activity in hippocampal and amygdala-dependent memory storage? Although T286 autophosphorylation might appear dispensable for memory storage, a more recently discovered mechanism of sustained CaMKII activation, which is induced by its interaction with NR2B, constitutes an attractive alternative and merits further discussion.

CaMKII binds to a number of postsynaptic proteins including the NMDA receptor subunits NR1, NR2A, and NR2B (Gardoni et al., 1998; Leonard et al., 1999; Strack and Colbran, 1998), although NR2A binding to CaMKII is of much lower affinity than its other interactions (Leonard and Hell, unpublished data; see also Strack et al., 2000a). Its association with the NMDA receptor is of special interest for several reasons. It places CaMKII at a strategically ideal location where it is most effectively activated by Ca²⁺ influx through the NMDA receptor, and puts it in close apposition to neighboring AMPA-type glutamate receptors. The AMPA receptor GluR1 subunit is phosphorylated on serine 831 by CaMKII upon Ca²⁺ influx through the NMDA receptor (Leonard et al., 1999), and during LTP (Barria et al., 1997; Lee et al., 2000). Furthermore, the interactions of CaMKII with NR1 and NR2B are activity-driven. NR1 binding requires T286 phosphorylation, whereas NR2B binding can be induced by either association of Ca²⁺/calmodulin with CaMKII, or via T286 autophosphorylation (Bayer et al., 2001; Leonard et al., 2002; Leonard et al., 1999; Strack and Colbran, 1998; Strack et al., 2000a). Autophosphorylation of CaMKII may also enhance binding to NR2A (Gardoni et al., 1999). These biochemical findings were supported by studies in hippocampal slices that demonstrated that Ca²⁺ influx through the NMDA receptor induces CaMKII binding to the NMDA receptor, presumably at postsynaptic sites (Leonard et al., 1999). Subsequent work with overexpressed GFP-tagged CaMKIIα confirmed that Ca²⁺ influx through the NMDA receptor leads to CaMKII clustering at postsynaptic sites in primary hippocampal cultures (Shen and Meyer, 1999; Shen et al., 2000). Collectively these studies suggest that activation-coupled interactions with the NMDA receptor may be responsible for the induction of CaMKII clustering upon postsynaptic Ca²⁺ influx, although final proof for this hypothesis is still missing. Together with NMDA receptors, densin 180, α-actinin, F-actin, the multi-PDZ domain protein MUPP1, and the cdk5 activators p35 and p39 are prominent components of dendritic spines, which constitute the postsynaptic sites of glutamatergic synapses. Autophosphorylation of CaMKIIα promotes its association with densin-180 although some binding is detectable in the non-phosphorylated state (Strack et al., 2000b; Walikonis et al., 2001). CaMKII binding to α-actinin, p35 and p39 is stimulated by addition of Ca²⁺ to brain lysates, which may involve endogenous calmodulin (Dhavan et al., 2002). Induction of Ca²⁺ influx through the NMDA receptor also upregulates coimmunoprecipitation of CaMKII with p35 and p39 in neuronal cultures; however the CaMKII - α-actinin interaction was, somewhat surprisingly, not increased under the same conditions (Dhavan et al., 2002). Accordingly, densin-180, p35, p39, and perhaps also α-actinin may contribute to activity-induced clustering of CaMKII at postsynaptic sites. In contrast, CaMKII binding to MUPP1 is inhibited by Ca²⁺/calmodulin (Krapivinsky et al., 2004). Similarly, CaMKIIα binds under resting conditions to F-actin (Shen et al., 1998); however, Ca²⁺/calmodulin binding to CaMKIIα or autophopshorylation of CaMKIIα on T287 leads to its dissociation from F-actin (Shen and Meyer, 1999). These latter two interactions are, therefore, not likely to play a critical role in activity-driven clustering of CaMKII at postsynaptic sites but could be critical for anchoring the kinase for subsequent activation and translocation closer to the NMDA receptor.

As stated above, association of CaMKII with NR2B requires either Ca²⁺/calmodulin or T286 autophosphorylation. However, after CaMKII binding to NR2B has been induced by Ca²⁺/calmodulin, it is stable even if calmodulin is completely removed from this complex by the chelation of Ca²⁺ (Bayer et al., 2001). Furthermore, after removal of Ca²⁺/calmodulin, NR2B-associated CaMKII showed autonomous activity similar to that caused by T286 autophosphorylation (Bayer et al., 2001), although no ATP was present throughout these experiments and thus T286 autophosphorylation could not have occurred. A similar constitutive activation mechanism has recently been described for CaMKII binding to the Drosophila eag K⁺ channel (Sun et al., 2004). To understand the molecular mechanism for the constitutive activation of CaMKII by NR2B, and also of eag binding, we have to consider how CaMKII is activated by Ca²⁺/calmodulin (for review see Colbran and Brown, 2004; Hudmon and Schulman, 2002). An autoinhibitory pseudosubstrate region just upstream of T286 interacts with the catalytic site of CaMKII, thereby occluding substrate access. This interaction is stabilized by the interaction of T286 with a separate hydrophobic pocket. Ca²⁺-calmodulin binds to the pseudosubstrate region, causing a conformational change which pulls it away from the catalytic site, thereby unsheathing T286 from its hydrophobic binding pocket. If Ca²⁺/calmodulin associates with two neighboring subunits within the dodecameric CaMKII complex, T286 from one subunit can be phosphorylated by the neighboring subunit. This autophosphorylation prevents the reassociation of T286 with its hydrophobic binding pocket, and thereby reassociation of the pseudosubstrate region with the catalytic site. Through these events the kinase gains autonomous activity. The CaMKII interaction sites on NR2B and eag bear substantial similarity with the T286 region. When Ca²⁺/calmodulin displaces T286 from its binding pocket, NR2B or eag can associate with this CaMKII site, so yielding the current model. This interaction remains intact even after dissociation of Ca²⁺/calmodulin from CaMKII, thereby occluding reassociation of T286 with its hydrophobic internal CaMKII binding pocket, and subsequent reassociation of the pseudosubstrate region with the catalytic site. The kinase will thus remain active. This mechanism provides an alternative way of endowing CaMKII with autonomous activity.

There is a potential caveat to the hypothesis that Ca²⁺/calmodulin-induced NR2B binding keeps CaMKII constitutively active for extended time periods. As mentioned above, GFP-tagged CaMKII clusters at postsynaptic sites upon NMDA receptor-mediated Ca²⁺ influx (Otmakhov et al., 2004; Shen and Meyer, 1999; Shen et al., 2000). This clustering phenomenon can also be observed for both the T286A and T286D mutation of CaMKII, which prevent and mimic autophosphorylation, respectively. One report finds that after Ca²⁺ influx is terminated, wild type GFP-CaMKII disperses within a few minutes whereas the T286A mutant does so within seconds; the T286D mutant did not show any appreciable redistribution away from the synapse within several minutes (Shen et al., 2000). These observations suggest that T286 phosphorylation is critical for prolonged CaMKII clustering at postsynaptic sites. They argue that T286 phosphorylation increases CaMKII binding to NR2B even in the presence of Ca²⁺/calmodulin, which is sufficient to induce full binding to monomeric NR2B in pull-down experiments (Bayer et al., 2001; Leonard et al., 2002). Perhaps T286-phosphorylated CaMKII may form multiple parallel interactions with NR2B and NR1 within the native complex, the latter depending on T286 phosphorylation (see above). However, it cannot be excluded that a small subpopulation of CaMKII even in its T286A mutant form remains at postsynaptic sites following cessation of the Ca²⁺ influx beyond the dwell times described for the bulk of wild type and mutant CaMKII. Furthermore, a more recent study in which LTP was induced by chemical stimulation found that the increase in accumulation of wild type GFP-CaMKII was stable for more than 60 min (Otmakhov et al., 2004). Perhaps under various conditions CaMKII may behave differently and may, in fact, in vivo be able to remain clustered for extended time periods.

Conclusions

CaMKII autophosphorylation is important for the acquisition of memory, but less so, if indeed at all, for its maintenance. The fact that T286A mutant mice can acquire several learning tasks also indicates that learning is not completely dependent upon the T286 phosphorylation event within CaMKII. Accordingly other signaling mechanisms must exist that can support learning and memory. These results suggest that a complex signaling network underlies memory formation and maintenance with both linear and parallel signaling cascades. Biochemical and cell biological studies indicate complex structural and functional interactions of CaMKII with its postsynaptic complex. We have only just begun to unravel the intricacies of CaMKII signaling at the postsynaptic site, and we need still more general and specific understanding of the molecular events which occur during the events which underlie synaptic plasticity and memory. We should perhaps keep an open mind for the role of CaMKII in these key events which underlie the essence of our being, as we may find that they may depend upon novel and as yet unrecognized properties of CaMKII.

References

Barria, A., D. Muller, V. Derkach, L.C. Griffith, and T.R. Soderling. 1997. Regulatory phosphorylation of AMPA-type glutamate receptors by CaM-KII during long-term potentiation. Science. 276:2042-2045.

Bayer, K.U., P. De Koninck, A.S. Leonard, J.W. Hell, and H. Schulman. 2001. Interaction with the NMDA receptor locks CaMKII in an active conformation. Nature. 411:801-5.

Chen, H.X., N. Otmakhov, S. Strack, R.J. Colbran, and J.E. Lisman. 2001. Is persistent activity of calcium/calmodulin-dependent kinase required for the maintenance of LTP? J Neurophysiol. 85:1368-76.

Colbran, R.J., and A.M. Brown. 2004. Calcium/calmodulin-dependent protein kinase II and synaptic plasticity. Curr Opin Neurobiol. 14:318-27.

Collingridge, G.L., J.T. Isaac, and Y.T. Wang. 2004. Receptor trafficking and synaptic plasticity. Nat Rev Neurosci. 5:952-62.

Derkach, V., A. Barria, and T.R. Soderling. 1999. Ca2+/calmodulin-kinase II enhances channel conductance of alpha-amino-3- hydroxy-5-methyl-4-isoxazolepropionate type glutamate receptors. Proc Natl Acad Sci U S A. 96:3269-74.

Dhavan, R., P.L. Greer, M.A. Morabito, L.R. Orlando, and L.H. Tsai. 2002. The cyclin-dependent kinase 5 activators p35 and p39 interact with the alpha-subunit of Ca2+/calmodulin-dependent protein kinase II and alpha-actinin-1 in a calcium-dependent manner. J Neurosci. 22:7879-91.

Erondu, N.E., and M.B. Kennedy. 1985. Regional distribution of type II Ca2+/calmodulin-dependent protein kinase in rat brain. J Neurosci. 5:3270-7.

Gardoni, F., A. Caputi, M. Cimino, L. Pastorino, F. Cattabeni, and M. Di Luca. 1998. Calcium/calmodulin-dependent protein kinase II is associated with NR2A/B subunits of NMDA receptor in postsynaptic densities. J. Neurochem. 71:1733-41.

Gardoni, F., L.H. Schrama, J.J. van Dalen, W.H. Gispen, F. Cattabeni, and M. Di Luca. 1999. AlphaCaMKII binding to the C-terminal tail of NMDA receptor subunit NR2A and its modulation by autophosphorylation. FEBS Lett. 456:394-8.

Giese, K.P., N.B. Fedorov, R.K. Filipkowski, and A.J. Silva. 1998. Autophosphorylation at Thr286 of the alpha calcium-calmodulin kinase II in LTP and learning. Science. 279:870-3.

Hudmon, A., and H. Schulman. 2002. Structure-function of the multifunctional Ca2+/calmodulin-dependent protein kinase II. Biochem J. 364:593-611.

Irvine, E.E., J. Vernon, and K.P. Giese. 2005. alphaCaMKII autophosphorylation contributes to rapid learning but is not necessary for memory. Nat Neurosci. 8:411-2.

Krapivinsky, G., I. Medina, L. Krapivinsky, S. Gapon, and D.E. Clapham. 2004. SynGAP-MUPP1-CaMKII synaptic complexes regulate p38 MAP kinase activity and NMDA receptor-dependent synaptic AMPA receptor potentiation. Neuron. 43:563-74.

Lee, H.K., M. Barbarosie, K. Kameyama, M.F. Bear, and R.L. Huganir. 2000. Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature. 405:955-9.

Leonard, A.S., K.-U. Bayer, M.A. Merrill, I.A. Lim, M.A. Shea, H. Schulman, and J.W. Hell. 2002. Regulation of calcium/calmodulin-dependent protein kinase II docking to N-methyl-D-aspartate receptors by calcium/calmodulin and a-actinin. J. Biol. Chem. 277:48441-48448.

Leonard, A.S., I.A. Lim, D.E. Hemsworth, M.C. Horne, and J.W. Hell. 1999. Calcium/calmodulin-dependent protein kinase II is associated with the N-methyl-D-aspartate receptor. Proc. Natl. Acad. Sci. USA. 96:3239-3244.

Lisman, J.E., and M.A. Goldring. 1988. Feasibility of long-term storage of graded information by the Ca2+/calmodulin-dependent protein kinase molecules of the postsynaptic density. Proc Natl Acad Sci U S A. 85:5320-4.

Malenka, R.C., J.A. Kauer, D.J. Perkel, M.D. Mauk, P.T. Kelly, R.A. Nicoll, and M.N. Waxham. 1989. An essential role for postsynaptic calmodulin and protein kinase activity in long-term potentiation. Nature. 340:554-7.

Malenka, R.C., and R.A. Nicoll. 1999. Long-term potentiation-a decade of progress? Science. 285:1870-1874.

Malinow, R., H. Schulman, and R.W. Tsien. 1989. Inhibition of postsynaptic PKC or CaMKII blocks induction but not expression of LTP. Science. 245:862-6.

Martin, S.J., P.D. Grimwood, and R.G. Morris. 2000. Synaptic plasticity and memory: an evaluation of the hypothesis. Annu Rev Neurosci. 23:649-711.

Shen, K., and T. Meyer. 1999. Dynamic control of CaMKII Translocation in hippocampal neurons by NMDA receptor stimulation. Science. 284:162-166.

Shen, K., M.N. Teruel, J.H. Connor, S. Shenolikar, and T. Meyer. 2000. Molecular memory by reversible translocation of calcium/calmodulin-dependent protein kinase II. Nature Neurosci. 3:881-6.

Silva, A.J., R. Paylor, J.M. Wehner, and S. Tonegawa. 1992. Impaired spatial learning in alpha-calcium-calmodulin kinase II mutant mice. Science. 257:206-11.

Strack, S., and R.J. Colbran. 1998. Autophosphorylation-dependent targeting of calcium/calmodulin-dependent protein kinase II by the NR2B subunit of the N-methyl-D-aspartate receptor. J Biol Chem. 273:20689-20692.

Strack, S., R.B. McNeill, and R.J. Colbran. 2000a. Mechanism and regulation of calcium/calmodulin-dependent protein kinase II targeting to the NR2B subunit of the N-methyl-D-aspartate receptor. J Biol Chem. 275:23798-806.

Strack, S., A.J. Robison, M.A. Bass, and R.J. Colbran. 2000b. Association of calcium/calmodulin-dependent kinase II with developmentally regulated splice variants of the postsynaptic density protein densin-180. J Biol Chem. 275:25061-4.

Sun, X.X., J.J. Hodge, Y. Zhou, M. Nguyen, and L.C. Griffith. 2004. The eag potassium channel binds and locally activates calcium/calmodulin-dependent protein kinase II. J Biol Chem. 279:10206-14.

Walikonis, R.S., A. Oguni, E.M. Khorosheva, C.J. Jeng, F.J. Asuncion, and M.B. Kennedy. 2001. Densin-180 forms a ternary complex with the (alpha)-subunit of Ca2+/calmodulin-dependent protein kinase II and (alpha)-actinin. J Neurosci. 21:423-33.