Using Your DNA to Remember
21 Mar, 2007 07:12 pm
In our article entitled ?Covalent modification of DNA regulates memory formation,? published in the current issue of the journal Neuron, Dr. David Sweatt and I report findings that suggest the adult nervous system has adopted an important developmental process that alters DNA for the storage of memories in the adult brain.
Surprisingly, the enzymes responsible for methylating DNA, DNA methyltransferases, are present in adult brain cells. If DNA methylation is static following development, why would the adult brain need such high levels of the enzymes responsible for this process?
We began investigating this tantalizing question in an earlier study published in the Journal of Biological Chemistry. In these experiments, we looked at the effect that inhibiting the DNA methylation process has on long-term potentiation or LTP. LTP is a representation of memory on the cellular level that involves the strengthening of connections between brain cells. We found that blocking DNA methylation in slices from the hippocampus, an important memory structure of the brain, prevented LTP from forming. This important finding made by Dr. Jonathan Levenson, a postdoctoral fellow in Sweatt’s lab at the time, suggested to us that DNA methyltransferases in the adult brain may, in fact, be performing a very important role—helping to form and store memories.
To investigate this possibility we trained a group of rats to form a memory in which they learned to associate the context of a novel place (the conditioning chamber) with a mild footshock, which is delivered through the grid floor the animals stand on in the conditioning chamber. The footshock is used as a stimulus for association because while mild, it is unpleasant enough for the animals to remember the experience. This learning paradigm is an excellent way to study long-term memory formation because you can easily test the animals for their memory by returning them to the conditioning chamber the following day and observing the strength of their fear response (freezing) in the absence of a footshock.
In this experiment, we trained animals to form the association between context and footshock and then administered a drug directly into the hippocampus to block the activity of the DNA methyltransferase enzymes during the memory formation period. We then measured how well they formed the memory by returning them to the conditioning chamber the following day. Similar to our earlier finding on LTP, animals that received the inhibitor drug did not freeze; demonstrating that blocking DNA methylation prevented the animals from remembering that they had been shocked in the conditioning chamber the day before.
This provided us with strong, but indirect evidence that the DNA methylation process may indeed be crucial for normal memory formation. To more directly test this, we looked for evidence of alterations in DNA methylation on specific genes that are important for memory formation. We chose to focus on two genes that are both important for memory formation, but in opposing ways. Reelin is a memory-promoting gene, while protein phosphatase 1 (PP1) is a memory suppressing gene. We hypothesized that the nervous system might be using DNA methylation as a way of activating the reelin gene while de-activating the PP1 gene to allow a long-term memory to form. As predicted, we found that when an animal is in the process of storing a memory after associative training, reelin’s methylation decreases, while PP1’s methylation increases. Recalling that gene methylation is associated with silencing a gene, it makes sense that PP1 would become more methylated in order to silence it, while the opposite would occur to reelin. So it appears that the process of DNA methylation serves to promote a memory promoting gene and silence a memory suppressing gene in order to control the activation of these genes during the window of long-term memory formation.
These findings have important implications not only for the further understanding of normal memory formation, but also for cognitive disorders, such as Rett syndrome and schizophrenia. Schizophrenia is associated with too little reelin, the opposite of what occurs during normal memory formation. These findings provide evidence that many researchers in the schizophrenia field have long suspected—that genes such as reelin can be dynamically regulated by the environment, and suggest that perhaps drug therapies can be developed to target hypermethylated genes like reelin that contribute to cognitive disorders.
Miller CA and Sweatt JD (2007). Covalent modification of DNA regulates memory formation. Neuron 53:857-869.