| || |
| The Molecular Basis of Learning and Memory |
It is clear that an understanding of mechanisms at the level of the synapse explains changes in our behaviors, like movements. But what about longer-term changes associated with learning and memory? Can they be understood in molecular terms, too? Memory and, thus, learning involves molecular changes in the brain. During the last few decades, researchers have started to map the molecular processes involved in memory formation. They have been increasingly able to link the ability to remember with physical changes in the structure of neurons.
One important change that occurs in memory formation is long-term potentiation (LTP). This phenomenon involves the long-term modification of the synaptic communication. Under normal circumstances the rate at which a postsynaptic neuron fires depends on how much stimulation it receives from presynaptic neurons. Once the increased stimulation has stopped, the postsynaptic neuron will return to its normal rate of firing. In LTP, however, the postsynaptic neuron will continue to fire at an elevated rate, even after the increased stimulation has subsided. It seems to become more sensitive -- or gives a bigger reaction by firing more action potentials -- to a given stimulus. How does this happen?
Glutamate is the neurotransmitter involved in LTP. Glutamate can bind to several different types of ionotropic receptors, including the NMDA- (N-methyl-D-aspartate) and AMPA- (amino-3-hydroxy-5-methyl-4-isoxazolepropionate) type glutamate receptors, each of which opens a specific type of channel within the receptor proteins. Both channels are involved in memory formation. The NMDA channel requires both glutamate and depolarization from another source to open. Why? The molecular mechanism is as follows. Normally, at negative potentials, positively-charged magnesium ions plug the pore of the NMDA channel. While glutamate may "open" the pore, the ions cannot travel through the channel due to the magnesium block. When the membrane is depolarized, however, the inside of the cell becomes more positive, and the magnesium ions are no longer driven into the channel. Thus, the block is relieved, allowing sodium and calcium ions to flow in.
So, this mechanism allows the NMDA-type glutamate receptor to act as a "coincidence detector." When the neuron receives input from only one source - another neuron - glutamate binds to and opens both NMDA- and AMPA-type receptors (Fig. 6). Because the neurotransmitter arrives at a resting, negatively charged, postsynaptic membrane, magnesium ions prevent flow through NMDA channels. When, however, stimulation of a neuron occurs simultaneously from more than one source -- say several other neurons -- some glutamate will bind NMDA receptors in parts of the neuron that are already depolarized, or less negatively charged.
Where does this voltage change come from? Recall that once an action potential has started, it spreads from its source throughout the entire membrane of the neuron in a wave-like fashion; thus, other dendrites may be "pre-depolarized" before glutamate binds. In this case, the block by magnesium is relieved and the NMDA channel also passes ions. While AMPA channels can pass only sodium ions in, NMDA channels also pass calcium. This calcium permeability gives the NMDA channel its ability to trigger LTP.
Now that we have examined the requirements for LTP, what is the effect? When calcium ions rush in, they set off an intracellular signaling cascade that can involve dozens of molecules. Speculation about the identity and functions of these molecules has been the subject of intense scientific inquiry since the early 1990s - it was perhaps the most studied aspect of neuroscience during that "decade of the brain."
So how could this intricate electrical mechanism act to form new memories? LTP, like learning, is not just dependent on increased stimulation from one particular neuron, but on a repeated stimulus from several sources. It is thought that when a particular stimulus is repeatedly presented, so is a particular circuit of neurons. With repetition, the activation of that circuit results in learning. Recall that the brain is intricately complicated. Rather than a one-to-one line of stimulating neurons, it involves a very complex web of interacting neurons. But it is the molecular changes occurring between these neurons that appear to have global effects. LTP can lead to strengthened synapses in a variety of ways. One such way, as discussed in the video, is by the phosphorylation of glutamate receptor channels, which is accomplished by a calcium-triggered signaling cascade. This results in those channels passing more ions with subsequent stimulation, strengthening the signal to and from the neuron.
But more permanent changes - long-term memory - require the synthesis of new proteins. In a variety of organisms, including flies (Drosophila) and humans, one enzyme, CREB (cyclic-AMP response element binding protein), seems to be involved in the steps that facilitate this new protein expression. When calcium flows in through NMDA channels, one of the molecules it activates is CREB. In turn, activated CREB acts as a transcription factor (see the Genetics of Development unit) that activates the expression of other genes. This gene expression can lead to the production of more ion channel receptors, as well as structural proteins like actin, which cement the synaptic connection between two repeatedly communicating neurons.
Mutant mice lacking the NMDA receptors show severe deficiencies in memory tasks. On the other hand, researchers have genetically engineered (see the Genetically Modified Organisms unit) mice that have more of the NMDA receptors. These mice, dubbed "smart mice" by the popular press, are substantially better at several memory tasks than are normal mice.