The PSD-95/SAP90 family of proteins has recently been implicated in the organization of synaptic structure. Here, we describe the isolation of a novel Ras-GTPase activating protein, SynGAP, that interacts with the PDZ domains of PSD-95 and SAP102 in vitro and in vivo. SynGAP is selectively expressed in brain and is highly enriched at excitatory synapses, where it is present in a large macromolecular complex with PSD-95 and the NMDA receptor. SynGAP stimulates the GTPase activity of Ras, suggesting that it negatively regulates Ras activity at excitatory synapses. Ras signaling at the postsynaptic membrane may be involved in the modulation of excitatory synaptic transmission by NMDA receptors and neurotrophins. These results indicate that SynGAP may play an important role in the modulation of synaptic plasticity.

Tau aggregation is a common feature of neurodegenerative diseases such as Alzheimer's disease, and hyperphosphorylation of tau has been implicated as a fundamental pathogenic mechanism in this process. To examine the impact of cdk5 in tau aggregation and tangle formation, we crossed transgenic mice overexpressing the cdk5 activator p25, with transgenic mice overexpressing mutant (P301L) human tau. Tau was hyperphosphorylated at several sites in the double transgenics, and there was a highly significant accumulation of aggregated tau in brainstem and cortex. This was accompanied by increased numbers of silver-stained neurofibrillary tangles (NFTs). Insoluble tau was also associated with active GSK. Thus, cdk5 can initiate a major impact on tau pathology progression that probably involves several kinases. Kinase inhibitors may thus be beneficial therapeutically.

Synapse-associated proteins (SAPs) are constituents of the pre- and postsynaptic submembraneous cytomatrix. Here, we present SAP102, a novel 102kDa SAP detected in dendritic shafts and spines of asymmetric type 1 synapses. SAP102 is enriched in preparations of synaptic junctions, where it biochemically behaves as a component of the cortical cytoskeleton. Antibodies directed against NMDA receptors coimmunoprecipitate SAP102 from rat brain synaptosomes. Recombinant proteins containing the carboxy-terminal tail of NMDA receptor subunit NR2B interact with SAP102 from rat brain homogenates. All three PDZ domains in SAP102 bind the cytoplasmic tail of NR2B in vitro. These data represent direct evidence that in vivo SAP102 is involved in linking NMDA receptors to the submembraneous cytomatrix associated with postsynaptic densities at excitatory synapses.

An mRNA encoding a helix-oop-helix protein that we have named HLH462 is induced in mouse 3T3 cells as part of the immediate early transcriptional response to growth factors and other signaling agents. The RNA is present in a number of mouse tissues and in the developing mouse fetus. The HLH462 gene has been mapped by interspecific backcross analysis to the distal region of mouse chromosome 4. In this report we describe the induction in mouse 3T3 cells and the presence in the developing mouse fetus and in adult tissues of mRNA that codes for a protein (called HLH462) that belongs to a recently recognized class of transcription factors, the helix-loop-helix proteins (3). The distinguishing feature of this class is an amino acid sequence motifpredicted to form a helix-loop-helix structure through which the proteins dimerize (3). Most members of the helix-loop-helix protein family so far described also have a basic DNAbinding domain adjacent to the dimerizing region, but others do not (4-6). Among the latter is the protein Id, implicated in the suppression of mammalian muscle cell differentiation (5) and the Drosophila emc protein, a negative regulator of sensory organ development on the fly's body surface (4, 6). These proteins are thought to act by forming inactive dimers with positively acting helix-loop-helix transcription factors (5). HLH462 shows extensive amino acid sequence similarity to Id and the Drosophila emc protein in the helix-loop-helix domain, and like Id it lacks a basic DNA-binding region but can interact with other helix-loop-helix proteins. Based on this structural and functional similarity, we hypothesize that HLH462 may negatively regulate the activity of other helixloop-helix transcription factors during the cellular response to growth factors and during development. § MATERIALS AND METHODS Cell Culture. BALB/c 3T3 cells were maintained and stimulated with serum or other inducing agents as described (7) Primer Extension and S1 Nuclease Assay. These procedures were carried out as described (9). The primer used was a 32P-labeled oligonucleotide complementary to nucleotides 132-155 of the cDNA (Fig. 1). Partial sequence of the genomic fragment used to prepare the probe for S1 analysis is shown in Fig. 1.In Vitro Transcription and Translation. In vitro transcription and translation (10) used either T7 or T3 RNA polymerase for transcription. The HLH462 cDNA template was a cDNA clone containing nucleotides 44-969 (see Fig. 1 (MUT.3, ref. 12) were labeled as described (12). Binding to the MCK enhancer site was carried out as described (5) except that sonicated salmon sperm DNA was used at 50 ,ug/ml as nonspecific competitor, and reactions were carried out in a total volume of 25 ,l with 12.5 fmol of 32P-labeled DNA probe. Binding to the Zif268 site was carried out as described (12) except that EDTA was omitted from the binding buffer. Competition assays were carried out in a total volume of25 Al with 5 fmol oflabeled probe (12). Competitors used were double-stranded oligonucleotides corres...