RIKEN Brain Science Institute (RIKEN BSI) RIKEN BSI News No. 17 (Aug. 2002)




Molecular Mechanisms of Learning and Memory

Yasunori Hayashi MD., PhD.
RIKEN-MIT Neuroscience Research Center
Center for Learning and Memory
Department of Brain and Cognitive Science
Massachusetts Institute of Technology


Since our laboratory started in September 2000, we have been working on several different projects ranging from molecular to disease. But all are united in two key words: synaptic plasticity and excitatory transmission.

IMAGE
magnified scene by clicking image
Fig.1 LTP in hippocampal CA1 region.

IMAGE
magnified scene by clicking image
Fig.2 Dendritic spines of CA1 pyramidal cells. Tiny puncta are spines.
Molecular Mechanism of Synaptic Plasticity
In the mammalian hippocampus, a brief tetanic stimulation of input fibers leads to a long-term potentiation (LTP) of excitatory synaptic transmission (Figure 1). LTP has been widely studied as a cellular and synaptic model for learning and memory. The LTP induction in hippocampal CA1 region requires post-synaptic NMDA receptor activation and a resultant influx of Ca2+ ions. One consequence of this rise in postsynaptic Ca2+ concentration is to trigger an increase of the transmission mediated by AMPA receptors via an activation of various Ca2+-dependent protein kinases. However, it has not been clear how this increase in transmission is attained. We suggested that an activity-induced translocation of the AMPA receptor from the extrasynaptic location to the postsynaptic site explains an increase in AMPA receptor transmission after LTP induction. We demonstrated this by using GFP and electrophysiologically tagged AMPA receptor molecules.
On the other hand, recent application of the yeast two-hybrid screening method disclosed a complex network of protein-protein interaction underlying excitatory synapse. This structural meshwork is not rigid but is dynamically regulated by various stimuli. It is therefore highly likely that the dynamics of AMPA receptors which themselves are embedded in this meshwork are regulated as a part of this dynamism of the meshwork. The principal aim of our research project is to understand the protein-protein interactions regulating AMPA receptor dynamics during LTP. For this purpose, we employ multidisciplinary approaches including biochemical isolation of glutamate receptor binding protein, construction of various mutants of glutamate receptor itself and binding proteins, gene introduction into neurons, and electrophysiological recording of those neurons.


Molecular mechanism spine formation
Spines are major site of excitatory synaptic transmission in hippocampal pyramidal cells (Figure 2). Their peculiar shape was first described as early as 19th century but the knowledge how it is formed and how its number and shapes are regulated are still limited. This is mainly due to a technical difficulty in accessing this structure. However, recent advance in imaging technique, mainly two-photon microscope, enabled direct visualization of spines in living neurons. The number and shape of the spines is regulated during development as well as by synaptic activity. Hippocampal CA1 pyramidal cells are virtually aspiny during the first two postnatal weeks except for filopodial structures, which are much longer and thinner than typical spine structure. After this period, typical spines with head and neck structure start emerging. In addition, tetanic stimulation induces new spines. In order to understand the mechanism involved in spine formation, we express various proteins found in spines and to see if they can form spine by themselves or change the morphology of spines. The spines are visualized with green fluorescent protein (GFP) under two-photon microscope. When GFP is expressed in neuron, it fills cytosol and thus depict the tiny protrusions of the cells including spines as if Golgi staining. Electrophysiological recording and Ca2+ imaging will be combined with GFP imaging to assess functional aspect of spines.

IMAGE
magnified scene by clicking image
Fig.3 Generation of knockout mice with tissue-specific deficiency using a Cre/LoxP recombination system
Structural approach to understand postsynapse
To understand the biophysical properties of synaptic protein from different aspect, we started a project to reveal 3D structures of those proteins. We will express various postsynaptic proteins in bacteria in large quantity, purify them, and make crystals. The resultant protein crystal will be analyzed with X-ray defraction. This approach is auxiliary the functional approach described above. For example, once functional unit of a protein is identified the physiological experiment, the relevant fragment will be used for crystallization trials. On the other hand, once we solved the structure and find a motif which seems to have functional significance, we can make mutants for further physiological analyses.

Role of adult neurogenesis in learning and memory
One interesting feature of hippocampus is a continual regeneration of neurons that takes place in adulthood. Placing an animal in a rich environment (for example, provided with a place to hide etc.) increases neurogenesis. The neurogenesis also takes place in olfactory bulb. However, there are no direct proof that the newly generated neurons are actually involved in formation of neuronal circuit and hence in a process of learning and memory. In order to test this, we are currently making a transgenic animal model where neuronal stem cell can be removed by infusion of cytotoxic substance. The resulting animal will be tested for memory paradigms such as Morris water maze task or pregnancy block.


Application of our knowledge to clinical medicine
We have recently identified a novel glutamate receptor subunit, NR3B, in the human and mouse genomes. NR3B shows very restricted expression in somatic motoneurons of the brainstem and spinal cord (Figure 3(. Due to this restricted expression, we are interested if this gene is involved in any of motoneuron diseases exemplified by amyotrophic lateral sclerosis (ALS).
ALS and motor neuron disease are neurological diseases that affect over 350,000 of the world's population, and kill over 100,000 every year. In the patients, spinal motor neurons degenerate resulting in a loss of control of muscles and, eventually, in a tragic death due to inability of respiration. Of the total ALS cases, 90% is sporadic and 10% has family history. Some populations of familial ALS cases have mutation in cytosolic Cu/Zn superoxide dismutase, an enzyme that converts free radical - O2 to H2O2 and O2. However, it accounts for only 15-20% of the familial cases and the cause the rest of familial and also most of sporadic cases is not known. We are interested in the possibility that a mutation in NR3B explains such familial ALS cases which cannot be attributed to SOD mutation and now determining the sequence of DNA samples obtained from ALS patients. It will help to understand the pathogenesis of ALS and motor neuron diseases, or even neurodegenerative diseases in general.

 
RIKEN BSI NEWS

RIKEN BSI HOME
RIKEN Brain Science Institute (RIKEN BSI)
Copyright All Rights Reserved.
s