Dynamic Structural Properties of Synapses

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications, trials & patents

Abstract

Progress Summary: The postsynaptic density (PSD) at excitatory glutamatergic synapses is known to be a key site of memory, information processing, and storage. To map the molecular organization of the PSD, we freeze-substitute hippocampal cultures and examine them in plastic-embedded sections by EM tomography. This reveals individual protein complexes within the PSD. Our early tomography work revealed that the core of the PSD is an array of membrane-associated, vertically-oriented filaments like PSD-95. This finding provided insight into the overall organization of the PSD. For instance, scaffolding proteins like PSD-95 have multiple common binding sites arrayed along their length, such that, regular arrays of vertically oriented PSD-95 filaments impose order on other PSD proteins, including the glutamate receptors, and provide an overall plan for the core structure of the PSD. The projects outlined below use and refine our insight into the organization and dynamics of synaptic structures within the PSD. We have several lines of ongoing research and collaboration to investigate the organization of specific synaptic proteins. Recently, we collaborated with the Rumbaugh Lab to characterize the structural role of SynGAP, which negatively regulates the glutamate AMPA receptor binding to the PDZ domain of PSD-95 at the PSD. We have used immunoEM to map the location, orientation, and conformation of SynGAP at the PSD to arrive at a structural model of how SynGAP might regulate and control synaptic excitability. We are finishing up work to identify NMDARs in the PSD using EM tomography in intact hippocampal synapses by using the CRISPR-Cas9 construct developed in the Nicoll Lab. The knockout eliminates the required GluN1 subunit of NMDARs thus eliminating NMDARs. We made 3D reconstructions of the resulting PSDs with dark-field scanning EM tomography. By comparing structures within wildtype and knockout, we can identify NMDARs structurally in tomograms and better characterize their organization as well as their connections with other molecules. Membrane-associated AKAPs are known to interact with PSD-95 MAGUKs and anchor several classes of kinases (PKA and PKC) important for synaptic plasticity (LTP and LTD). We are extending our study with M. DellAcqua's lab at the University of Colorado on the conformations and distribution of Anchoring Proteins (AKAPs) in the hippocampal synapse (see publications below). Our current investigation focuses on the potential role of AKAP in organizing PKA and PKC in inhibitory synapses. 3D-EM tomography delineates the organization of subsynaptic organelles, key synaptic proteins, and macromolecular complexes at synapses. Reconstructions of tomography data provide the morphology and location of structures at 2-4 nm resolution but cannot guarantee unambiguous molecular identification of the individual structures. While we had success using immunogold to label endogenous and overexpressed GFP-tagged PSD-95, the large antibody complexes that also manifest as filamentous structures in tomograms confound the identification of other key PSD proteins. Now we have developed two new independent but complementary methods using APEX2 tagging and nanobody labeling in conjunction with EM tomography to identify proteins in intact synapses. We are using the APEX2 genetic tag to localize CaMKII holoenzymes. The CaMKII-APEX2 construct in the presence of DAB and peroxide has revealed individual CaMKII holoenzymes. We are studying CaMKII distribution in the spine and membrane in basal and high potassium (high K+) stimulated conditions by EM tomography. This work is now being prepared for publication. In addition, we are expanding the APEX2 tagging approach to Shank and Homer. We are using nanobodies with EM tomography to swiftly and directly identify PSD proteins in spines. In collaboration with Shahid Khan, we are examining the higher-order association of CaMKII, stabilized with a -hydroxybutyrate analog, with isolated NMDAR NR2B subunits and PSDs. While -hydroxybutyrate analogs are known to act as neuroprotective agents, it has only been recently established that the analogs bind specifically to the hub of the CaMKII isoform. The hub-stabilized CaMKII offers a novel opportunity to explore the formation of CaMKII-mediated condensates with the NMDA NR2B subunits and scaffolds with the PSDs respectively. We are using two-color fluorescence and cryo-EM for this study. Also, Khan used the NIH Biowulf supercomputer to show that, among the hubs of the mammalian CaMKII isoforms, the phylogenetics of the CaMKII hub track the evolution of memory with the greatest fidelity (see publication list). Khan has subsequently found, upon study of the complete CaMKIIs from six model organisms ranging from humans to the placozoan, Trichoplax adhaerans, that the CaMKII kinase domain calmodulin, autophosphorylation and substrate (e.g., NR2B) binding sites are remarkably well-conserved. Since T. adherents do not have a nervous system, the bioinformatics implies that CaMKII architecture developed for a more primitive sensory or stress response but was then repurposed for memory storage. Experimental validation will require proof of T. adherents CaMKII holoenzyme formation, calcium-calmodulin-triggered autophosphorylation, and NR2B binding affinity.