Morgan Sheng

The Molecular Basis of Synaptic Plasticity




 

Morgan Sheng, M.B.B.S., Ph.D.

Associate Investigator,

Massachusetts Institute of Technology




 

Summary: Morgan Sheng seeks to understand the molecular and cellular mechanisms underlying the ability of the brain to change in response to experience, such as during development and learning.



The brain is a massive network of electrically active cells (neurons) that communicate with each other via specialized cell junctions (synapses). Throughout development and adult life, this neuronal network responds to experience by adjusting the strength of communication at individual synapses and by changing the physical pattern of synaptic connections between neurons. In this way, information can be stored by the nervous system in the form of altered structure and chemistry of synapses and/or by the formation of new synapses and the elimination of old ones. This “plasticity” of synapses is believed to be the basis of learning and memory in the brain. Because of their central importance in information processing and storage, it is important to understand the molecular architecture of synapses and the cellular processes that govern synapse formation and elimination. By using biochemical, genetic, and imaging approaches to study synaptic junctions between neurons, we hope to reveal the fundamental molecular mechanisms of brain plasticity.



In a “bottoms-up” approach, we are systematically characterizing the protein components of synapses and elucidating how these proteins interact with each other to construct the synaptic junction. Our focus is on the postsynaptic side of the synapse, which receives the neurotransmitter signal released from the presynaptic axon terminal. Attached to the postsynaptic membrane is a specialized microscopic structure, the postsynaptic density (PSD). The PSD contains the receptors for the neurotransmitter glutamate (glutamate receptors), of which there are three major classes: NMDA receptors, AMPA receptors, and metabotropic glutamate receptors. Starting from these glutamate receptors, we have unraveled the network of protein-protein interactions and the signal transduction pathways that connect the postsynaptic membrane to the interior of the neuron. By differential activation of postsynaptic signaling pathways, specific patterns of synaptic stimulation can lead to either strengthening or weakening of synaptic transmission (synaptic plasticity).



One way to change synaptic strength is by the dynamic regulation of glutamate receptors. In particular, we are investigating how the abundance and distribution of AMPA-type glutamate receptors on the neuronal surface is controlled by the coordinated movement (trafficking) of receptor proteins. Remarkably, AMPA-type glutamate receptors shuttle rapidly (a time-scale of minutes) in and out of the postsynaptic membrane. Delivery of AMPA receptors to the synapse results in strengthening (potentiation) of synapses, while removal of AMPA receptors from the synapse leads to weakening (depression) of synaptic transmission. We and others have discovered that specific proteins control AMPA receptor trafficking to and from synapses by binding to the cytoplasmic tails of AMPA receptor subunits. AMPA receptors contain at least two kinds of subunit: GluR1 and GluR2. The GluR1 subunit and its associated proteins control the synaptic delivery of AMPA receptors, while GluR2 and its interacting proteins regulate receptor internalization from the postsynaptic membrane and subsequent recycling to the surface. (A grant from the National Institutes of Health provided support for this project.)



Are other proteins also delivered to the synapse along with AMPA receptors? Our hypothesis is that the delivery package that reinforces postsynaptic AMPA receptors also contains other proteins important for synaptic growth. Using genetic and biochemical methods, we are isolating the AMPA receptor protein complex in order to purify and identify the specific associated proteins. So far we have found cell adhesion molecules (required for holding pre- and postsynaptic membranes together), motor proteins (presumably involved in moving AMPA receptors), and adaptor proteins (that mediate interactions between AMPA receptors and other components of the complex).



Compared with AMPA receptors, NMDA-type glutamate receptors are rather stably incorporated in synapses. When activated by glutamate, NMDA receptors allow the influx of calcium and trigger several signaling cascades in the postsynaptic neuron that ultimately lead to functional and morphological changes in synapses. NMDA receptors are anchored in the PSD by binding to PSD-95, a major scaffold protein of the PSD that in turn associates with many signaling proteins. NMDA receptors are thereby linked to a large complex of proteins that serve to transmit biochemical signals from the NMDA receptor to the interior of the neuron. A large number of PSD proteins are likely to play a role in NMDA receptor signaling and in the ensuing synaptic modifications, and this has motivated us to identify and characterize the individual components of the PSD.



Our initial insights into the molecular composition of the PSD were based on genetic approaches (the yeast two-hybrid system of identification of protein-protein interactions). More recently, we have exploited mass spectrometry (a physicochemical method that measures the precise mass of protein fragments) to analyze the components of the PSD. Approximately 500 proteins were identified, including all the previously discovered PSD proteins, plus many hitherto unsuspected components. Since it is important to understand quantitatively the composition of the PSD, we are using quantitative mass spectrometry to investigate the stoichiometric ratios of the major PSD proteins. In addition, we are using electron microscopy to look at the size and shape of individual molecules of the PSD.



What are the specific functions of the individual components of the PSD? We have chosen to focus on a subset of PSD proteins that are either abundant or likely to be involved in NMDA receptor signaling. We are taking several approaches to address protein function. First, we can overexpress the protein in cells and see what this does to neuronal morphology, synapse development, and function. Second, we can devise mutant variants of the protein (“dominant-negative” mutants) that when introduced into cells should poison the activity of the endogenous normal protein. Additionally, we can disrupt the gene in the animal (mouse) by genetic recombination and create a knockout mouse lacking the specific protein.



Using such approaches, we are investigating the function of a subset of PSD proteins in the regulation of dendritic spines, which are tiny protrusions found on the branches of many neurons. Dendritic spines are specialized morphological compartments on which excitatory synapses are formed, and these fascinating structures change in size and shape, depending on a wide variety of factors such as brain activity, neurological disease, hormonal cycles, and aging. It is believed that changes in dendritic spine number and morphology reflect synaptic plasticity, particularly changes in synaptic connections between neurons. We are interested in the specific molecules that control the formation, morphology, and motility of dendritic spines, and have recently identified several PSD proteins that regulate the structure of dendritic spines. Two of these (Shank and Homer) cooperate to enlarge the mushroom-like head of the spine, while another (SPAR) appears to stimulate elaboration and branching of spines. Dominant negative mutants of Shank, Homer, and SPAR cause loss of spines, implying that these proteins are important in spine formation or stabilization.



We wish to examine the physiological significance of such changes in spine morphology by deleting the genes for Shank and SPAR in mice and assessing the impact of these genetic manipulations on brain function and organismal behavior. We recently generated a knockout mouse lacking the scaffold protein Shank1, which links NMDA receptor and metabotropic glutamate receptor complexes with the actin cytoskeleton. The homozygous mutant mice are viable and appear outwardly normal, and we are studying the anatomical, electrophysiological, and behavioral (particularly learning and memory) consequences of Shank1 deletion.