The post synaptic density (PSD) at excitatory glutamatergic synapses is a complex molecular machine which appears to be a key site of information storage. New methods to probe its structure show that a lattice-like backbone labeling for PSD-95 forms its core, while other structural components-such as the kinase CaMKII-occupy various locations in the lattice. The PSDs in intact neurons, however, change size rapidly and reversibly during activity. Immunolabeling shows that CaMKII is a major component of the added mass, involving up to 100 CaMKII holoenzymes. Comparison of PSD dynamics in the cerebellum and forebrain indicate a correlation between cellular alpha-CaMKII abundance and degree of PSD thickening. CaMKII is aggregated at the PSD is in its phosphorylated form, and it appears that addition of CaMKII depends on phosphorylation. We have also shown that reversibility is blocked by chemical LTP (long term potentiation, thought to be a key step in establishing memory). Both the thickening and addition of CaMKII persists in LTP, suggesting that the induced association with the PSD is a key step leading to LTP. In order to explore the detailed molecular organization of the PSD, we have developed a method to freeze-substitute hippocampal cultures and then examine them by tomography of thin sections. Tomography reveals a filamentous meshwork as the core structure of the PSD. We have developed new methods for visualizing the structure of these filaments and associated molecules. Several classes of filaments are recognized in the PSD, and their sizes, numbers, and orientations have been catalogued. The next step is to identify eachclass of filament. Ongoing complementary studies with isolated PSDs, using our rotary shadowing technique, are aimed at mapping the distributions of PSD-specific scaffolding molecules, including PSD-93 and SAP-97 within the PSD. We have also developed a method to affinity- purify PSDs from other components of the PSD fraction, thereby allowing independent measurement of CaMKII content, as well as proteomic analysis by mass spectroscopy (with S. Markey). This study represents the first proteomic analysis of purified PSDs, thereby eliminating the contributions from abundant contaminating cytoskeletal elements and other proteins in the PSD fraction. Initial comparison of mass spectrometric data for the conventional and affinity-purified PSD fractions reveals enrichment following affinity purification of specialized scaffolding molecules and glutamate receptors, especially of the AMPA type, accompanied by a notable decrease in certain cytoskeletal and presynaptic proteins. Ultimately we plan to determine the number of copies of each major component of the PSD using new mass analysis techniques we recently described, as well as by EM analysis of immunogold labeled replicas. Another new series of experiments, using an antibody that selectively recognizes phosphorylated Ser/Thr residues followed by prolines, suggests that PSDs contain the kinase(s) and phosphatases responsible for the regulation of the phosphorylation of several endogenous proteins at Ser(Thr)-Pro motifs. Preliminary results indicate involvement of phosphatases of type 1 or 2A in dephosphorylation and involvement of p38-gamma in the phosphorylation of PSD-95. These studies will ultimately help clarify some of the molecular mechanisms involved in activity-induced synaptic plasticity.
Showing the most recent 10 out of 16 publications
