Cdk5 Controls Memory Switch in Neurons

Young brains soak up experiences like sponges, but with age new material becomes harder to learn. Is there a way to turn back the clock? Perhaps so, according to a paper in the March 5 Neuron. Researchers led by James Bibb at the University of Texas Southwestern Medical Center, Dallas, reported that cyclin-dependent kinase 5 (Cdk5) regulates trafficking of the glutamate receptor subunit NR2B, which predominates in developing brains and plays a role in learning and memory. Cdk5 traps NR2B within the cell, preventing it from reaching synapses. The researchers designed a synthetic peptide that interfered with this interaction and allowed more NR2B to populate synapses, as it does in young brains. When administered to rodents, the peptide improved learning in both young and old animals. Drugs that disrupt the interaction between Cdk5 and NR2B might therefore have the potential to enhance cognition in diseased or damaged brains, Bibb suggested.

“This paper provides a precise biochemical definition of how Cdk5 affects synaptic activity. The data are strong, including everything from physiology to biochemistry to behavior,” said Karl Herrup at Hong Kong University of Science and Technology. He was not involved in the work. Other scientists could not be reached for comment. 

Previous work demonstrated that transient activation of Cdk5 benefits neurons by growing the number of synapses and sharpening learning and memory (see Dec 2005 conference story; Dec 2006 conference story; Dec 2007 news story). The kinase also has its bad side, however. Hyperactive Cdk5 bound to its cofactor p25 promotes Alzheimer’s pathology, including Aβ production, tau phosphorylation, synapse loss, and neuronal death (see Jun 2004 conference story; Dec 2006 news story; Mar 2008 news story). 

For its part, NR2B has also been linked to memory. Overexpression of the subunit in adult mice increases synaptic plasticity and improves learning ability (see Tang et al., 1999), while a variant of GRIN2B, the gene that encodes NR2B, is more common in people with mild cognitive impairment than in healthy elderly (see May 2010 news story).

Bibb and colleagues previously connected the two when they reported that Cdk5 promoted degradation of NR2B by the protease calpain. Cdk5 conditional knockout mice had more NR2B and learned more readily than controls (see Jun 2007 news story). However, calpain was not the whole story, Bibb told Alzforum. He noticed that while the overall level of NR2B in these mice inched up by only 25 percent, twice as much of the subunit made it to the cell surface. This suggested that Cdk5 somehow regulates the trafficking of NR2B. 

In the new paper, Bibb and colleagues dissect how this happens. First author Florian Plattner identified serine 1116 on NR2B as a phosphorylation site for Cdk5. Intriguingly, this site sits next to an amino acid sequence that typically helps retain proteins in the endoplasmic reticulum (ER), suggesting that phosphorylation might influence the protein’s location. When Plattner and colleagues treated hippocampal neurons expressing tagged NR2B with a Cdk5 inhibitor for one hour, phosphorylation at the Ser1116 site dropped, and the amount of NR2B at the cell surface jumped up 30 percent. Moving to a more physiological paradigm, hippocampal slices treated with the inhibitor had more than threefold endogenous NR2B at the surface and developed stronger post-synaptic currents in response to stimulation. 

To test the role of Cdk5 in vivo, the authors synthesized a peptide of eight amino acids that bound to the kinase and blocked its interaction with NR2B. They infused the peptide into the hippocampi of 3-month-old, wild-type rats for 72 hours using osmotic pumps. As in slices, the amount of NR2B at the cell surface shot up threefold. Treated rats retained memories better than controls, freezing nearly twice as often in a contextual fear-conditioning test. The authors saw the same improvement when they administered the peptide acutely to 3-month-old mice. Fourteen-month-old mice responded as well as the youngsters to the treatment, demonstrating that the mechanism works in the aged brain as well. 

Intriguingly, the authors found that stimulating hippocampal slices either electrically or with the glutamate agonist NMDA lessened phosphorylation of NR2B by Cdk5. This presented a puzzle, since scientists know that neuronal activity pumps up Cdk5 activity in neurons by promoting the cleavage of Cdk5’s cofactor p35 to its constitutively active form, p25. How to reconcile this? Bibb speculated that the switch from p35 to p25 might cause Cdk5 to abandon NR2B in favor of other substrates. Herrup agreed that this is plausible, noting that previous work has suggested such substrate switching. Therefore, even though neuronal activity turns up Cdk5, phosphorylation of NR2B drops, Bibb suggested. Other evidence supported the idea that the kinase was more active overall in the stimulated neurons. Bibb saw a drop in total NR2B in these cells, in keeping with his previous findings that activated Cdk5 promotes the degradation of the subunit.

What to make of the finding that activated Cdk5 allows NR2B to reach synapses, but also reduces total levels of the subunit? This suggests a sort of feedback loop in which neuronal activity causes NR2B phosphorylation to drop so that the subunit can travel to the cell surface and participate in synaptic remodeling and memory formation, and then active Cdk5 later prunes the subunit out of the receptor, Bibb said. To Bibb, the data fit with the known transient role for NR2B, which is replaced in synapses by the NR2A subunit during development. “NR2B is invoked, used, and degraded, and the neuron moves over to a more hard-wired synapse that stores memory,” Bibb suggested. “NR2B is like a mayfly that is there for just one day and then burns out.”

If NR2B helps memories form, then pumping up the amount of the subunit in synapses should restore them to a more plastic state, like that in developing brains. In ongoing work, Bibb is testing the inhibitory peptide in mouse models of AD and traumatic brain injury to see if it can sharpen memory in diseased or damaged brains. He plans to screen for small-molecule drugs that do the same and could enter the brain. Such a molecule might have potential as a drug to enhance cognition in aged or damaged brains, Bibb suggested.—Madolyn Bowman Rogers.

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. SfN: P25 at Synapses—A Bite Peps Up, A Binge Crashes the System
  2. Madrid: Beyond Aβ—Learning, Age, DNA Damage, and Pai-1
  3. Synapse Formation: Another Piece in Cdk5’s Repertoire?
  4. Tangles, Neurodegeneration, Plaques—p25 Does it All
  5. What Drives Dendritic Spine Loss? Study Taps Cdk5
  6. New Role for p25/Cdk5 in Regulation of BACE Expression
  7. Obesity Gene Depletes Brain Reserves, May Raise Alzheimer’s Risk
  8. The Multi-talented Cdk5—Role in Plasticity and Learning

Paper Citations

  1. Tang YP, Shimizu E, Dube GR, Rampon C, Kerchner GA, Zhuo M, Liu G, Tsien JZ. Genetic enhancement of learning and memory in mice. Nature. 1999 Sep 2;401(6748):63-9. PubMed.

Further Reading

Primary Papers

  1. Plattner F, Hernández A, Kistler TM, Pozo K, Zhong P, Yuen EY, Tan C, Hawasli AH, Cooke SF, Nishi A, Guo A, Wiederhold T, Yan Z, Bibb JA. Memory enhancement by targeting Cdk5 regulation of NR2B. Neuron. 2014 Mar 5;81(5):1070-83. PubMed.

Annotate

To make an annotation you must Login or Register.