Protein synthesis initiation can be regulated, either positively or negatively, by protein kinases. One such kinase, GCN2, results in general inhibition of protein synthesis, but there is one exception; it stimulates the synthesis of the protein ATF4. ATF4 in turn inhibits the CREB protein, which is responsible for synaptic plasticity and behavioural learning, so GCN2 indirectly leads to the inhibition of CREB. As GCN2 is an evolutionarily conserved kinase and is normally expressed in the hippocampal areas of mammalian brains, a major site for memory formation, Costa-Mattioli and co-workers investigated the effect of GCN2 inactivation upon synaptic plasticity and behavioural learning.

Wondering what effect the loss of GCN2 would have on synapse formation, the team decided to look at mice that had been modified to lack GCN2. They knew that GCN2 regulates expression of the CREB-inhibitor ATF4 and wondered if levels of the inhibitor would fall in the knockout mouse with a resulting rise in the levels of active CREB. Accordingly, the authors saw both a decrease in ATF4 levels and a 25-35% increase in genes regulated by CREB, suggesting that CREB activity had increased.

Next, the authors explored how long-term potentiation (LTP) induction, a model for memory formation, was affected in the hippocampus of knockout mice. LTP has long been used as an experimental tool to study the cellular mechanisms of memory formation. Electrical stimulation of neurons can result in an artificial increase, or potentiation, in synaptic strength. When the team tested a protocol that typically elicits only early-LTP (i.e. lasting 2-3 hours) in normal mice, they found that the protocol resulted in late-LTP (i.e. lasting longer than 2-3 hours) in the knockout mice. Amazingly, a protocol that normally produces early-LTP in normal mice had produced a completely different form of LTP in mice that lacked GCN2. Thus, less stimulation in the knockout mice resulted in longer potentiation of synaptic strength compared with normal mice. But what effect would it have on real memory formation?

The authors behaviourally tested the knockout mice to see if hippocampal-dependent memory was, in fact, similarly affected by the lack of GCN2 and subsequent rise in CREB levels. The popular Morris water maze was used to test for hippocampal-dependent spatial memory. In this task, mice are placed in a pool of opaque water and expected to swim to a hidden platform positioned in a particular spot. After mice learn where the platform is they will swim directly to it when put back in the maze. In the course of normal training (three times a day for five days) the performance of both normal and knockout mice improved, with the normal mice learning faster than the knockout mice. However, when mice were trained only once a day, knockout mice showed better spatial memory than normal mice during training and three days later. Enhanced spatial memory in knockout mice was only seen after weaker training.

Overall, the authors saw a decrease in the threshold for late-LTP in the hippocampus, which was associated with improved spatial memory of weaker conditioning in knockout mice. The authors propose that, in mice lacking the GCN2 gene, there is enhanced CREB function after weak stimulation,but with stronger stimulation an unknown inhibitory pathway may be activated. The results indicate that neurons might not only have a threshold for activating gene expression but also a second threshold at which too much gene expression blocks synaptic plasticity. So, in the end, the idea of everything in moderation still holds true; from exercise to red wine and now to CREB.

Costa-Mattioli, M., Gobert, D., Harding, H., Herdy, B., Azzi,M., Bruno, M., Bidinosti, M., Mamou, C. B., Marcinkiewicz, E., Yoshida, M. et al. (
2005
). Translational control of hippocampal synaptic plasticity and memory by the eIF2α kinase GCN2.
Nature
436
,
1166
-1170.