6 June 2007. Kinases generally perform the quintessential job of substrate phosphorylation. But it appears that cyclin-dependent kinase 5 (Cdk5) has been hiding another ability—one that compromises learning and memory. In the May 27 Nature Neuroscience online, researchers led by James Bibb at the University of Texas Southwestern Medical Center, Dallas, report that Cdk5 promotes degradation of glutamate receptors in a manner apparently independent of phosphorylation. The finding may be of considerable interest to those eyeing Cdk5 as a potential therapeutic target for Alzheimer disease (AD).
Ckd5 has been implicated in the pathology of AD at several levels. Not only does it phosphorylate the neurofibrillary tangle protein tau (see ARF related news story), but there is evidence that it contributes to production of intraneuronal Aβ (see ARF related news story) and loss of dendritic spines (see ARF related news story). Furthermore, p25, a constitutively active fragment of the Cdk5 activator p35, is elevated in AD brain (see ARF related news story). But there’s also the softer side to Cdk5. It is absolutely required for neuronal development and may be necessary for spatial learning and memory in adult mammals. However, since knocking out Cdk5 compromises development, testing its role in the adult brain has been difficult.
Bibb and colleagues got around this dilemma by generating Cdk5 conditional knockout (cKO) mice. Joint first authors Ammar Hawasli and David Benavides and their colleagues used a modified Cre/loxP system to ablate Cdk5 genes in the adult brain. Based on a Cre/modified estrogen receptor chimera, the system can be induced by administering the estrogen receptor antagonist hydroxytamoxifen, which promotes translocation of the chimera to the nucleus, where it “floxes” the gene. Administering the antagonist reduced overall levels of Cdk5 mRNA in the CA1 layer of the hippocampus by about 70 percent, and totally removed it in about 90 percent of CA1 neurons.
To test the physiological effects, Hawasli and colleagues compared the performance of knockout and control mice in various learning tasks. Surprisingly, the cKO animals learned better in tasks that draw on hippocampal function. The knockout mice outperformed controls in contextual fear conditioning, in contextual fear extinction, in a water maze test of spatial learning and memory, and in a reversal task in the water maze. Because learning and memory are intimately tied up with neuronal plasticity, the researchers turned to electrophysiological measurements of the Schaffer collateral pathway to try to explain the learning and memory enhancements in the cKO animals. They found that the threshold for long-term potentiation was lower in Cdk5-negative tissue slices. While weak tetanic or theta-burst stimuli had no effect on control slices, it induced measurable LTP in cKO tissue. Hawasli and colleagues found that this was associated with increased NMDA receptor-mediated transmission—AMPA receptor transmission appeared normal.
In the hippocampus, NR1, NR2A, and NR2B subtypes of NMDA receptors are available for glutamate binding, but the researchers found that only NR2B subtypes seem altered in cKO tissue. NR2B receptor levels were 30 percent higher in cKO hippocampal tissue, and NR2B-mediated excitatory post-synaptic potentials were up 3.2-fold. Furthermore, when the researchers examined cell surface receptor load, NR2B levels were twofold higher than in control tissue, and the NR2B selective antagonist ifenprodil blocked LTP enhancement.
The data suggest that under normal conditions, Cdk5 somehow reins in NR2B receptors, How might that work? It is known that NMDA itself can stimulate NR2B degradation by calpain, but the authors found no evidence that Cdk5 phosphorylates either NR2B or the calcium-dependent protease. And yet, they found less calpain-mediated degradation of NR2B in cKO slices than controls. Bibb and colleagues also found that Cdk5/p25 stimulates degradation of NR2B in cell-free homogenates and that it activated calpain in the absence of Mg2+ and ATP, which are normally required for kinase activity. The data seem to suggest that Cdk5 stimulates calpain-mediated degradation of NR2B, but in a kinase-independent manner. In support of this idea, the authors found that Cdk5, its coactivator p35, calpain, and NR2B all co-immunoprecipitated with each other from control hippocampus and that the amount of NR2B associated with p35 was reduced by about 50 percent in Cdk5 knockout tissue. “Thus, elevated NR2B levels in the Cdk5-knockout hippocampus likely resulted from the loss of these critical protein-protein interactions,” conclude the authors.
The findings reveal a hitherto unknown role for Cdk5 in promoting NMDA receptor degradation. Whether this may be related to Alzheimer pathology is unclear at present, but it raises an interesting question about potential Cdk5-targeted therapeutics. That is because small molecules designed to block the kinase activity of Cdk5 may not necessarily prevent it from wreaking havoc on NMDA receptors.—Tom Fagan.
Hawasli AH, Benavides DR, Nguyen C, Kansy JW, Hayashi K, Chambon P, Greengard P, Powell CM, Cooper DC, Bibb JA. Cyclin-dependent kinase 5 governs learning and synaptic plasticity via control of NMDAR degradation. Nature Neuroscience. 2007 May 27, advanced online publication. Abstract