Cell adhesion and synapse formation

Although numerous molecules have been associated with synaptic cell adhesion, there has been little conclusive evidence linking them to specific effects on synapse formation and function. By searching vertebrate sequence databanks, a novel homophilic synaptic cell adhesion molecule called SynCAM has been found which is homologous to members of the immunoglobulin-like (Ig) superfamily and in vitro acts in the formation and functional differentiation of synapses. Before forming a chemical synapse, physical contact between a cell and its target must occur. Even though it is commonly accepted that extracellular matrix proteins in combination with cell surface molecules on both the preand postsynaptic membrane mediate this interaction, what the identity of these molecules is remains largely unknown. This is a significant question since the probable action of synaptic cell adhesion molecules is not limited to imparting mechanical stability to synapses. Their purported role is wider and includes specific target recognition, synaptic differentiation, and the regulation of synaptic function. Numerous synaptic proteins have been implicated in these processes, including the Eprhin/EphB receptor pathway, the neuroligin/_-neuroexin pathway, the cadherin protein family and the Ig domain proteins. The ephrins have been shown to influence synaptic assembly, even though their main activity is during development and they classically mediate repulsion signaling. The neuroligin/_-neuroexin heterotypic cell adhesion pair is involved in mammalian transsynaptic signaling. However, their interaction occurs in a calcium dependent fashion, whereas synaptic cell adhesion is not calcium dependent (Nguyen and Sudhof, 1997). Cadherins specify the formation of synaptic connections but are not part of established junctions and occur in non-synaptic areas of the cell as well. Protocadherins, also members of the cadherin protein family, are synaptic but not evolutionarily conserved. Currently under investigation are Ig domain proteins – they operate in homophilic extracellular interactions and are coupled to intracellular PDZ domain containing proteins. Invertebrate Ig domain proteins include the Drosophila protein fasciclin II Nicole Coufal, NEU200A 2 (Davis et al. 1997) which has been shown to have preand postsynaptic localization, is necessary for synaptic stabilization, and becomes concentrated in muscle at the site of synapses. Similarly, the Aplysia homolog apCAM is required for normal synapse formation (Mayford et al. 1992). Both of these proteins have intracellular PDZ domains which mediate their interaction with signaling proteins such as CASK. However, the mammalian homologs previously identified have not had the necessary characteristics to be synaptic cell adhesion molecules. NCAM, for example, functions in cell adhesion but lacks a PDZ domain. The Sidekicks, sdk 1 & 2, act as adhesion proteins, have been localized to synapses by confocal microscopy, possess PDZ domains, and function in lamina-specific connections. However, they are restricted to two thin sub-layers of the inner plexiform layer of the retina (Yamagata et al 2002). Therefore, despite having actions correlated with synaptic cell adhesion, each of these molecules presents some difficulty which prevents it from playing a central role in synaptic cell adhesion: not evolutionarily conserved, calcium dependent, restricted tissue expression, or lacking synaptic localization. In the current paper, Biederer et al have searched through vertebrate sequence databases for proteins with homology to fasciclin II and apCAM which contain both extracellular Ig domains and an intracellular PDZ-domain protein-interaction sequence. They identified the mouse protein SynCAM which is evolutionarily conserved and exhibits brain-restricted expression. Expression of SynCAM increases during the first three weeks postnatal, corresponding to the period of synaptogenesis in mice. Using immunoblots the authors showed that SynCAM is highly N-glycosylated and demonstrates variations in glycosylation both regionally and temporally. There is increased glycosylation of SynCAM in the cortex and striatum as opposed to the cerebellum and hindbrain and, over the course of the first two weeks of postnatal life, the pattern changes from one of high glycosylation to more core glycosylation. The expression pattern was confirmed with immunohistochemistry and immunogold labeling of SynCAM in the hippocampus and molecular layer of the cerebellum. Characterization of SynCAM indicates that it participates in strong homophilic interactions and this binding is salt resistant and calcium independent. Both binding assays and in vitro cell culture experiments indicated SynCAM’s three Ig domains were necessary to mediate homophilic binding. Transfection with full-length SynCAM induced cell aggregation, a phenotype which was abolished when a mutant SynCAM lacking Ig domains was used. The intracellular portion of SynCAM contains a PDZ Nicole Coufal, NEU200A 3 domain which the authors showed binds to CASK (a CaMK-, SH3-, and guanylate-kinase-domain containing protein) and recruits it to the plasma membrane. The PDZ domain is homologous to that found in neurexins and syndecans (Tomita et al. 2001). Lastly, SynCAM is localized to the synapse, to both preand postsynaptic membranes. The authors illustrated this through subcellular fractionation showing SynCAM is enriched in the synaptic plasma membrane fraction (along with CASK and neuroligin1), followed by immunoelectron microscopy to identify SynCAM both preand postsynaptically, and finally through double immunofluorescence to demonstrate colocalization in a punctuate pattern of SynCAM with synaptophysin. So SynCAM is found in the right place – at the synapse on both sides – and at the right time – postnatal weeks 1-3. It also has plausible homophilic extracellular and PDZ domain intracellular interactions. But does is have any functional impact on synapses? To address this question Biederer and colleagues elegantly investigated SynCAM’s action both preand postsynaptically. To look at postsynaptic activity they transfected hippocampal neuronal cultures with full length SynCAM and observed increased spontaneous synaptic activity in the form of a 2-3 fold increase in miniature synaptic currents (mini’s). This effect was not elicited with the Ig-domain-deletion SynCAM or control vector. The number of mini’s is purported to be proportional to the number of synapses made onto the cell, although it could also be due to enhanced presynaptic neurotransmitter release at existing synapses, a possibility the authors do not address. They saw no change in the mini amplitude, which is generally regarded as being proportional to the amount of receptor present at the synapse. To investigate presynaptic terminal formation in response to SynCAM the authors transfected hippocampal neurons with a dominant negative form of SynCAM (dnSynCAM) which contained the complete intracellular domain but lacked the extracellular domain? The authors theorized that expressing only the cytoplasmic tail of SynCAM in neurons would inhibit synapse assembly by out-competing endogenous SynCAM for CASK and other PDZ binding proteins. In response to this construct the authors observed decreased numbers of presynaptic terminals formed onto transfected neurons as measured by increased inter-puncta distances using FM dye. As a control they used a truncated cytoplasmic SynCAM with a deletion of the PDZ domain which did not disrupt the interaction of endogenous SynCAM with PDZ binding proteins. In addition to decreased synaptic density, the authors also observed that vesicle release was abnormal in dnSynCAM transfected neurons. Synaptic terminals Nicole Coufal, NEU200A 4 exhibited slower vesicle release, visualized as slower FM dye destaining in response to K+ depolarization. Additionally the amplitude of destaining was decreased by half, indicating a 50% decrease in the size of the vesicle pool. After characterizing the effect of increasing or decreasing SynCAM levels in hippocampal neurons, Biederer et al addressed the crucial question of whether expression could induce synapse formation onto non-neurons. This is a key inquiry as it indicates whether SynCAM is sufficient for synapse formation, above and beyond just influencing the number or efficiency of synapses onto a cell. To tackle this question the authors expressed SynCAM in HEK293 cells and cocultured them with hippocampal neurons to investigate if the neurons would form synapses onto 293 cells. This approach is similar to one used previously to study the _-neurexin –neuroligin interaction and its role in synapse formation (Scheiffele et al. 2000) and was also used by Fu and colleagues to cotransfect 293’s with neurexin and either NMDA or AMPA receptor subunits and study resulting synaptic currents (Fu et al. 2003). Biederer et al found that 293 cells expressing SynCAM formed reductionist synapses. They saw punctuate staining of synaptic terminals labeled with FM dye adjacent to SynCAM positive areas on 293 cells (visualized by immunohistochemistry). These synapses colocalized with synaptophysin and syntaxin. Although there were also synapse-like structures between 293 cells and neurons under control situations (Ig domain deleted SynCAM transfections) they were fewer in number and none of them were active as assayed with FM dye. Only synapses between wild type SynCAM transfected 293 cells and neurons actively took up FM dye with stimulation. When the authors compared these synapses to interneuron synapses in the same culture they saw no functional differences in the size of the vesicle pool or in the kinetics of vesicle release as analyzed by FM dye destaining. These data indicate de novo formation of presynaptic terminals in response to SynCAM. In their last experiment, Biederer et al coexpressed SynCAM and the glutamate receptor subunit GluR2 in 293 cells and cocultured them with hippocampal neurons. This gave them a “functional postsynaptic response” such that the 293 cells exhibited spontaneous currents in some cases (9 of 31 cells) which were abolished by CNQX treatment, and increased with glutamate application. Rise time and amplitudes of currents were comparable to those of interneuron synapses. Spontaneous activity of 293 cells was abolished by treating the culture with TTX to block action potentials, indicating that spontaneous Nicole Coufal, NEU200A 5 activity in the hippocampal cultures is likely being transmitted to the 293 cells. This surprising outcome indicates that SynCAM together with GluR2 meets the minimal requirements for synapse reconstitution. Based on the large number of molecules which are implicated in synaptic cell adhesion there is likely to be redundancy in the system such that SynCAM, ephrins, and cadherins all contribute in some way to synaptic specificity and postsynaptic organization. The identification of SynCAM is interesting because its action is not limited to adhesion, is calcium independent, and clearly also involves synapse function and differentiation. However, there are many questions that are still left unanswered. Firstly, in vivo data in transgenic mice is necessary to show that loss of SynCAM results in deficits of synapse formation. Secondly, there needs to be elucidation of the downstream cascades mediating SynCAM’s actions and how these mechanisms translate cell adhesion into formation of a functional synapse. Since both SynCAM and neurexin are likely to signal through PDZ domains, it begs the question of the functional significance of this interaction and how much of the effect of dnSynCAM is due to disruption of neurexin signaling. Thirdly, what role, if any, does SynCAM play in the specificity of the synapses formed, in their differentiation, and whether this process is dependent on neuronal activity. Recently, it was shown that Eprhin B2 and EphB modulate NMDA receptors and therefore link activity dependent and independent signals (Takasu et al. 2002), possibly SynCAM has similar actions. Lastly, given recent evidence that neuroligan-3 and -4 mutations play a role in mental retardation and autism (Laumonnier et al., 2004), are there SynCAM mutations in humans, and what phenotype ensues? Possibly one of the most fascinating conclusions of this study is that it is possible to functionally construct minimalist synapses in vitro. This is a powerful assay for isolating individual proposed synaptic adhesion molecules and studying their action in isolation. References: Biederer T., Sara Y., Mozhayeva M, Atasoy D., Liu X., Kavalali E.T., and Sudhof T.C., 2002. Science 297, pp.1525-1531. Davis G.W., Schuster C.M., and Goodman C.S. 1997. Neuron 19, pp.561-73 Fu Z., Washbourne P., Ortinski P., and Vincini S. 2003. J. Neurophysiol 90. pp. 3950-3957. Laumonnier F., Bonnet-Brillault F, Gomet M., Blanc R., David A., Moizard M.P., Raynaud M., Ronce N., Lemonnier E., Calvas P., Laudier B., Chelly J., Fryns J.P., Ropers H.H., Hammel B.C., Andres C., Barthelemy C., Moraine C., Briault S., 2004. Am J Hum Genet 74, pp. 552-557. Nicole Coufal, NEU200A 6 Mayford M., Barzilai A., Keller F., Schacher S., Kandel E.R. 1992. Science 256, pp. 638-644. Nguyen, T., and Sudhof, T.C., 1997. J. Biol. Chem. 272, pp.26032-26039. Scheiffele P., Fan J., Choih J., Fetter R., and Serafini T. 2000. Cell 101, pp.657-669. Takasu M.A., Dalva M.B., Zigmond R.E., and Greenberg M.E. 2002 Science 295, pp. 491-495. Tomita S., Nicoll R.A., and Bredt D.S. 2001 J Cell Biol. 153, pp.F19-24. Yamagata M., Weiner J., and Sanes J. 2002. Cell 110, pp. 649-660.