Uncovering the Molecular Basis of Learning and Memory

Newly synthesized proteins (green) traveled to mushroom-shaped spines in mouse hippocampus neurons. Researchers believe this process is key to strengthening a memory. Images after 1 (left), and 6 (right) hours of fear conditioning.Drs. Mark Mayford and Naoki Matsuo, Scripps Research Institute

Researchers have developed a way to pinpoint the molecules involved in forming a specific memory. The finding, in genetically engineered mice, gives scientists new insight into how memories are formed.

For a memory to last long-term, the neural connections holding it need to be strengthened by incorporating new proteins. Some researchers have proposed that an experience creates a molecular “tag” at activated synapses, the connections between neurons. The tag allows synapses to capture newly made proteins and thus solidify a memory. But the molecular details of the process have been a mystery.

Previous studies suggested that proteins called AMPA-type glutamate receptors (AMPARs) strengthen memories by becoming part of the synapses that encode new memories. Drs. Mark Mayford and Naoki Matsuo of the Scripps Research Institute, supported in part by NIH's National Institute of Mental Health (NIMH), decided to explore the proteins further. They genetically engineered a strain of mice to make AMPARs that could be traced by their green glow.

The transgenic mice were taught to associate a specific environment with a foot shock, a process known as fear conditioning. This brief training produces a long-lasting memory that requires the brain’s hippocampus region. After fear conditioning had triggered new AMPARs deep in the neuron's nucleus, the researchers tracked where the newly made proteins went.

In the February 22, 2008, issue of the journal Science, the researchers reported that the newly synthesized AMPARs travel to and become captured by only certain hippocampus synapses—presumably the ones holding the new memory—within hours.

Synaptic connections are made onto small nubs on the neuron called spines. These spines come in 3 different shapes called thin, stubby and mushroom. The researchers found that the synapses receiving the new AMPARs were limited to the mushroom type.

The mushroom spines also figured prominently in the same neurons when fear conditioning was reversed by repeatedly exposing the animals to the feared situation without getting shocked—a procedure called extinction learning. This result shows that the same neurons activated when a fear is learned are also deactivated when it is lost.

The surge of receptors in mushroom spines appeared within hours of learning, suggesting that when mice learn something new, there are changes in some mushroom spines that allow them to capture newly synthesized AMPARs. The receptor surge was gone within 3 days, however, so other changes likely solidify the memory for the long term.

"Remarkably, this research demonstrates a way to untangle precisely which cells and connections are activated by a particular memory," said NIMH Director Dr. Thomas Insel. "We are actually learning the molecular basis of learning and memory."