Development and Function of the Synapse
Séminaire FSER organisé par Pietro De Camilli et Yukiko Goda du 23 au 29 avril 2003
Tobias Bonhoeffer, Nils Brose, Graeme Davis, Pietro De Camilli, Paolo Di Fiore, Thierry Galli, Frank Gertler, Jean-Antoine Girault, Yukiko Goda, Eckart D. Gundelfinger, Reinhard Jahn, Eric Jorgensen, Josh Kaplan, Michela Matteoli, Andrew Matus, Harvey McMahon, Franck Polleux, Elena Porro, Anuradha Rao, Tim Ryan, Bernardo Sabatini, Christoph Schuster, Michele Solimena
Synaptic connections are essential elements of information processing in the nervous system. They are the specialized sites where signals are transmitted from one neuron to another with high speed and spatial precision via chemical intermediates called neurotransmitters. Over the last several year major progress has been made in the elucidation of the mechanisms underlying the traffic of neurotransmitter containing synaptic vesicles and their regulated fusion (exocytosis) with the plasma membrane of the presynaptic compartment. Likewise, much has been learned about the dynamics of postsynaptic receptors. Contrary to the widespread idea that neurotransmitter receptors are stable components of the postsynaptic membrane, it has now been established that such receptors are dynamically regulated by vesicular transport reactions mechanistically similar to those occurring presynaptically. In addition, there is growing evidence for an unexpected degree of structural plasticity of synaptic junctions. This progress has been facilitated by the rapid growth of information about the genomes and proteomes of entire organisms as well as by major technological advances. These include the development of powerful light microscopy imaging techniques for the observation of living cells and tissues, the introduction of new biophysical methods for the analysis of exo-endocytosis, the availability of widely applicable methods for gene disruption or gene silencing. However, important details as well as major fundamental questions, such as for example the mechanisms of synapse formation and maintenance remain poorly understood. From 23rd to 29th of April 2003, twenty-two scientists who specialize in various aspects of synaptic cell biology and who approach the problem from different perspectives, gathered at “Les Treilles” in the South of France to discuss the current state of understanding of synapse development and function. Proceedings of the conference are summarized below in the form of a brief outline of the individual presentations.
Josh Kaplan (Harvard Medical School) reported results of genetic and cell biological studies on synaptic function in C. elegans. He first summarized evidence for presynaptic signaling pathways that regulate synaptic vesicle priming (a late step in the fusion of neurotransmitter containing synaptic vesicles with the plasma membrane). Two opposite pathways, controlled by acetylcholine and serotonin respectively, converge on both Unc13 and an atypical PKC (PKCe) via the generation of diacylglycerol. Next he discussed two distinct modes of regulation of postsynaptic receptor abundance. One is the result of a homeostatic response to changes in the presynaptic release of glutamate. Mutations in the vesicular glutamate transporter, eat4, that impairs glutamate loading of synaptic vesicles and therefore glutamate release, results in a compensatory increase in postsynaptic glutamate receptor abundance. The other involves receptor ubiquitination. Ubiquitination of glutamate receptors promotes their endocytosis and subsequent trafficking to lysosomes for degradation. Worms carrying a point mutation of the lysines that function as acceptors for ubiquitination on the receptors, or those defective in the component of the E3 ligase complex, lin23, accumulate glutamate receptors at the cell surface.
Tobias Bonhoeffer studies molecules that can translate neuronal activity into morphological changes at synapses. BDNF is one such candidate. It is secreted in response to neural activity, and by activating TrkB receptors, it modulates long-term potentiation of synaptic strength (LTP) via signaling pathways that include PLCg activation. While BDNF is absolutely required for the late phase of LTP, it also plays a permissive role in promoting the early phase of LTP. In an effort to determine the consequences of neuronal activity on synaptic connectivity in vivo, his laboratory is making an elegant use of 2 photon microscopy to simultaneously visualize dendritic architecture and intrinsic signaling in the brain of live rodents.
Pietro De Camilli addressed the role of phosphorylation-dephosphorylation of phosphoinositides in the recycling of synaptic vesicles. PI(4, 5)P2, which is primarily restricted to the plasma membrane, functions as a membrane coreceptors in clathrin coat recruitment and may also play an important role in synaptic vesicle exocytosis. Dephosphorylation of PI(4, 5)P2 by synaptojanin is essential to ensure efficient recycling of synaptic vesicles and is one of the mechanisms facilitating clathrin uncoating. PI(4, 5)P2 must then be regenerated by the sequential action of a PI 4-kinase and PIP kinase type 1g, a PI(4)P kinase. PIP kinase type 1g is modulated by phosphorylation-dephosphorylation reactions and is stimulated by interactions with talin and GTPases, including Arf6. A preliminary analysis of mice lacking PIP kinase type 1g displays defects in synaptic vesicle recycling, which may include a defect in vesicle priming. Finally, electrospray ionization mass spectrometry was employed for sensitive and quantitative measurements of phosphoinositides.
Harvey MacMahon focused on proteins that mediate and support membrane deformations occurring during the budding of clathrin coated vesicles. He focused on two related but distinct proteins, AP180 and epsin, that function as accessory clathrin adaptors and help generate membrane curvature. Both proteins bind PI(4, 5)P2 via their NH2-terminal ANTH and ENTH domains, respectively, and to clathrin and the clathrin adaptor AP-2 via unfolded domains located downstream of these domains. Epsin deforms membranes via an amphipatic a-helix that is generated upon binding to PI(4, 5)P2 and phosphorylation of epsin at Ser5 prevents its ability to induce membrane curvature. The crystal structure of the BAR domain of Drosophila amphiphysin, another domain implicated in the generation of membrane curvature, was also presented.
Erik Jorgensen discussed studies of C. elegans mutants defective in components of synaptic vesicle endocytosis. Worms carrying synptojanin (Unc26) null alleles are viable, and their mutant nerve terminals display reduced synaptic vesicle number, accumulation of endocytic pits plus other endocytic intermediates, and presence of a filamentous matrix reminiscent of actin. In null mutants of endophilin (Unc57), a lipid binding protein that also binds synaptojanin, GFP-synaptojanin is mislocalized; thus lipid binding by endophilin is required for correct targeting of synpatojanin. In contrast, the subcellular localization of endophilin remains normal in the absence of synaptojanin. In unc57 mutants the synaptic vesicle protein VAMP partially accumulates in the plasma membrane, suggesting a defect in endocytosis. In these mutants, a reduced frequency of spontaneous vesicle fusion and a 50 % reduction of evoked synaptic responses is also observed. The phenotyupe of unc26/ unc57 double mutants is very similar to that of unc26, suggesting that the interaction with synaptojanin is the main function of endophilin.
Thierry Galli described a neuronal exocytic pathway in both axons and dendrites driven by the tetanus toxin insensitive V-SNARE TI-VAMP. RNAi-mediated knock-down of TI-VAMP attenuates neurite outgrowth whereas deletion of the N-terminal autoinhibitory domain of TI-VAMP promotes SNARE complex formation and facilitates neurite outgrowth. The cell adhesion molecule L1 localizes to TI-VAMP containing vesicles, implicating the delivery of L1 as a function of TI-VAMP-mediated vesicle traffic in neurite elongation. TI-VAMP is also present in the adult brain, where it is primarily enriched in the somatodendritic area. The significance of TI-VAMP-dependent traffic in mature neurons will be an interesting area for future investigations.
Michela Matteoli continued the discussion of exocytosis in growing axons. Axons are characterized by robust constitutive vesicle recycling that is down-regulated following synaptogenesis. Interestingly, such down-regulation does not occur in GABA-ergic neurons. Moreover, GABAergic synapses show lower threshold for Ca2+-activated synaptic vesicle recycling. This down-regulation of constitutive recycling and increased stringency of Ca2+-coupling in excitatory neurons appears to be contributed by SNAP25, which is surprisingly absent in the processes of inhibitory neurons. Exogenous expression of SNAP25 in GABAergic neurons confers the property of reducing constitutive vesicle recycling in conjunction with neuronal maturation. The presence of AMPA receptors in axonal growth cones was also reported. Pools of GluR1 and GluR2 subunits that are present in synaptic vesicles are recruited to the plasma membrane in response to depolarization. Such delivery of AMPA receptors may be central to the steering of growth cones through the AMPA-regulated motility of presynaptic filopodia.
Paolo Di Fiore discussed mono-ubiquitination as a general signaling mechanism and as a protein targeting mechanism in the endocytic pathway. Monoubiquitination can occur at multiple sites on a protein, leading to the so-called poly-monoubiquitination. For example, EGF and PDGF receptors undergo ligand-dependent poly-monoubiquitination. Recognition of the monoubiquitin signal is mediated by a network of proteins containing ubiquitin-binding motifs, such as epsin and Eps15, which contain UIM domains. The importance of the mono-ubiquitination signal for internalization and lysosomal targeting of receptor tyrosine kinases is supported by the constitutive internalization of a chimaeric EGF receptor-ubiquitin mutant that cannot be polyubiquitinated.
Jean-Antoine Girault addressed signal transduction mechanisms in neurons. Focal adhesion kinase (FAK) is a non-receptor tyorosine kinase that undergo regulation by sumoylation. Its expression in neuronal growth cones is thought to facilitate growth in concert with signaling downstream of GPCRs, integrin-dependent adhesion, and tyrosine kinase receptors. He also discussed the significance of endocannabinoid signaling in the adult brain. Activation of CB1 cannabinoid receptors by anandamide, arachidonyl-glycerol and tetrahydrocannabinol stimulates MAP kinase/ERK kinase in a mechanism requiring Fyn, as ERK activation is lost in Fyn knock-out mice. Surprisingly, the catalytic activity of Fyn may not be required as ERK activation is insensitive to inhibitors of Src family tyrosine kinases. ERK signaling participates in long-term synaptic plasticity suggesting that cannabinoid receptor modulation has a physiological role in regulating the activity of specific neuronal networks.
Reinhard Jahn gave an overview of SNARE-mediated fusion in membrane traffic. SNAREs are small membrane proteins whose interaction to form so-called SNARE complexes is a key step leading to membrane fusion. SNARE proteins are important determinants of the specificity of membrane fusions, although such specificity is further assisted by other factors, including the rab family of GTPases. For example, based on antibody inhibition, the SNAREs Vti1a and Vti1b have a differential role in early and late endosome fusion respectively. Mutations in a given SNARE can be suppressed by compensatory mutations in a partner SNARE so that complementarily of their interacting surfaces is maintained. To characterize the spatial organization of SNARE complex in an intact membrane, SNARE protein distribution was examined in plasma membrane lawns obtained by cracking cells attached to glass slides. Syntaxin and SNAP25 form separate clusters that require cholesterol; exocytosis of GFP-labeled vesicles occur preferentially on syntaxin clusters.
Bernardo Sabatini addressed postsynaptic mechanisms. The role of NMDA receptors in signaling pathways that regulate dendritic growth and spine formation was investigated by cell autonomous RNAi-mediated knock-down of NR1 subunit in slice cultures. Plasmid-based dsRNA specific for NR1 (delivered by microinjection) produced a substantial reduction in NR1 subunit levels in the somatodendritic area 7-10 days post transfection. Knock-down of NR1 was accompanied by a severe decrease in spine density and dendritic arbor size specifically in transfected neurons. In contrast, silencing of NMDA receptors in the entire culture did not produce pronounced effects on dendrites and spines. Paired-pulse facilitation, a parameter that reveals neurotransmitter release probability in the presynaptic terminal, was increased when synaptic responses were monitored in NR1 knock-down neurons. This indicates the occurrence of retrograde signaling from spines to the presynaptic neurotransmitter release machinery, which communicates the state of excitability of the postsynaptic compartment. The behavior and significance of compartmentalized Ca2+ signals in regulating synaptic transmission were also discussed. At aspiny interneuron synapses that form onto stellate cells in the cerebellum, Ca2+-diffusion is highly restricted; in the medium spiny neurons in the striatum, L-type somatic Ca2+ channels affect spike firing independently of Ca2+ signals in the spines. Future studies employing 2-photon uncaging of caged glutamate will allow to analyze the behavior of Ca2+-signals and postsynaptic receptor traffic at individual spines in response to localized stimulation.
Tim Ryan presented studies of the balance between exo- and endocytosis at presynaptic terminals using optical probes for vesicle recycling, specifically focusing on the speed and modes of endocytosis. At synapses formed between dissociated hippocampal neurons grown in culture, the average rate of endocytosis is slowed as the number of extracellular stimuli is increased. At 10 Hz stimulation, the endocytic rate is 2-3 times slower than that of exocytosis, whereas at 2 Hz stimulation, the rate of endocytosis is approximately ~87% of the rate of exocytosis. Measurements of the endocytic delay following exocytosis suggest a minimal delay time of at least 300 msec.
Nils Brose described the regulation of synaptic vesicle priming by Munc13. The C1 domain of Munc13 binds diacylglycerol and phorbol esters, and according to his studies Munc13 isoforms fully accounts for the DAG/phorbol ester-dependent regulation of neurotransmitter release. Munc13-1 and Munc13-2 are differentially expressed in distinct populations of hippocampal neurons. In contrast to Munc13-1-dependent synapses, Munc13-2-depedent synapses show pronounced and transient augmentation of synaptic amplitudes following high-frequency stimulation. This augmentation is caused by an increase in release probability and releasable vesicle pool size, and requires phospholipase C activity. Thus, differential expression of Munc13 isoforms at individual glutamatergic synapses represents a mechanism for modulating short-term plasticity and provides a basis for synapse heterogeneity. Interestingly, the two Munc13 isoforms have completely redundant functions in GABAergic cells.
Yukiko Goda addressed the role of actin dynamics in regulating neurotransmitter release. Acute depolymerization of actin filaments by latrunculin A enhances neurotransmitter release without a change in the size and the rate of refilling of the readily releasable vesicles, suggesting that actin might play a role in regulating late stages of synaptic vesicle exocytosis at the active zone. The involvement of actin in the regulation of neurotransmitter release is also supported by studies in which the function of native modulators of active dynamics, ADF/cofilins and gelsolin, are perturbed. How synaptic activity regulates the synaptic cytoskeleton was examined by photoconductive stimulation of neurons grown on silicon wafers. Stimuli that mimic the induction of a durable form of long-term synaptic plasticity triggered the formation of new presynaptic actin puncta that contain actively recycling synaptic vesicles.
Franck Polleux discussed the molecular basis for the patterning of cortical connectivity. He developed a slice overlay assay to identify the contribution of extracellular signals to the proper orientation and growth of neuronal processes. When dissociated labeled neurons are plated onto a slice of the cortical plate, the axonal growth is properly targeted towards the pia. The oriented outgrowth of axons, which express neuropilin 1 in their growth cones, require semaphorin 3A. In addition, semaphorin 3A facilitates the growth of apical dendrites through a mechanism requiring the asymmetric localization of soluble guanylate cyclase. Semaphorin 3A thus regulate the patterning of both developing axons and dendrites. The requirement of area-specific cues for the formation of cortical topography of thalamocortical projections is being examined. EphA4/ephrinA5 signaling is required for the correct rostral projection of ventrolateral neurons to the motor cortex.
Frank Gertler reported the mechanism by which Ena/Vasp family members regulate cell migration. Ena/Vasp proteins transiently accumulate at the tips of filopodia. Overexpression of Ena/Vasp increases the rate of lamellipodial protrusion, yet the speed of cell translocation is reduced. In contrast, the absence of Ena/Vasp results in faster cell migration while lamellipodial protrusion is slowed. Actin networks in Ena/VASP-deficient lamellipodia are composed of short, highly branched filaments. In lamellipodia with excess Ena/VASP, actin is present as longer, less branched filaments, due to the antagonistic effect of Ena/Vasp on barbed end capping proteins. Ena/VASP therefore regulate cell motility by controlling the geometry of actin filament networks within lamellipodia. Interestingly, mechanisms of motility may be quite distinct in neuronal growth cones and at the leading edge of fibroblasts. Not only the morphology of the two structures is dissimilar, but the pattern of Arp2/3 distribution is also very different.
Michele Solimena discussed mechanisms in the regulation of mobilization of peptide hormone containing secretory granules. Islet cell autoantigen (ICA) 512 is a receptor-tyrosine phosphatase-like protein present on the secretory granules of neuroendocrine cells. ICA512 also contains an inactive tyrosine phosphatase domain that binds bIV spectrin andb2-syntrophin. The ICA512/b2-syntrophin complex binds utrophin and is found on secretory granules that accumulate next to sites of cell-cell contact. Stimulation of insulin secretion from INS-1 cells triggers Ca2+-dependent dephosphorylation of b2-syntrophin, which in turn dissociates from ICA512. The cytoplasmic tail of ICA512 is the cleaved by m-calpain. Stimulus-dependent dissociation of ICA512/b2-syntrophin/utrophin complex and the cleavage of ICA512 may provide a positive feedback regulation that ensures the supply of available secretory granules. Potential nuclear functions of the cleaved ICA512 fragment and the role of post-transcriptional regulation in the stimulus-dependent upregulation of secretory granule biogenesis was also discussed.
Andrew Matus investigated the molecular basis for the regulation of dendritic spine morphology. Overexpression of drebrin, an actin binding protein, increases the relative number of dendritic filopodial protrusions. The expression of the actin-binding fragment of drebrin is sufficient to produce this effect. Profilin is a small regulatory protein that enhances actin polymerization. Prolonged glutamate stimulation of cultured neurons promotes the accumulation of GFP-tagged profilin II in the spine head. The redistribution of profilin II-GFP is not readily reversible, requires Ca2+-influx via the NMDA receptors, and is associated with cessation of spine motility. These experiments suggest a role of actin dynamics in the regulation of synaptic plasticity.
Eckart D. Gundelfinger discussed the assembly and organization of presynaptic “active zones” of secretion. These sites are characterized by the presence of electron-dense material associated with the presynaptic membrane, also termed the Cytomatrix assembled at the Active Zone (or CAZ). He focused on Bassoon and Piccolo, two giant proteins thought to serve as specific scaffolding components of the CAZ. Based on studies of rodent hippocampal neurons grown in vitro, Bassoon associates with a membranous compartment in the Golgi complex region and is then transported to the presynapse as part of transport vesicles which also carry other major components of the CAZ including Piccolo, Bassoon, RIM and Munc13 as well as integral membrane proteins such as N-type calcium channels and cadherins. Thus, parts of the active zone are likely to be transported along the axon as pre-assembled units. Mutant mice lacking functional Bassoon are viable but suffer from epilepsy. A significant fraction of hippocampal synapses of these mice are inactive although they have a normal ultrastructure. However, striking morphological differences are observed at ribbon synapses of photoreceptor cells in the retina, where ribbons are detached from the active zone. Synaptic ribbons are thought to represent a form of CAZ specialized for the highly efficient tonic release of glutamate from photoreceptor nerve terminals. Bassoon is thus essential for the proper assembly and/or function of central synapses, and its deficiency has variable effects on different types of synapses.
Christoph Schuster reported acute experience-dependent changes that occur at the Drosophila neuromuscular junction (NMJ). Removal of food instantaneously induces enhanced surface locomotion of larvae. 45 minutes of increased motility results in an increase in postsynaptic muscle responses that are initially sustained by an increase in the size of miniature events. By 2-3 hours after induction of enhanced locomotor activity, miniature event sizes are downregulated while evoked responses continue to be potentiated. These physiological alterations are accompanied by an experience-dependent increase in local subsynaptic protein synthesis as reflected by an enhanced occurrence of subsynaptic eIF4e polysome aggregates and within 4-6 hours by an increased perisynaptic accumulation of the postsynaptic glutamate receptor subunit DGluR-IIA. Transgenic expression of a Ca2+-sensor protein yellow Cameleon-2, whose stimulation-dependent fluorescence dynamics correlates with the position and the size of active zone harboring presynaptic boutons, revealed localized growth of boutons following chronic stimulation of locomotor activity at 29°C. These data show that Drosophila NMJs are capable of performing rapid and reversible experience-dependent alterations of glutamatergic signal transmission, which later on appear to consolidate by synapse addition and morphological growth. Christoph Schuster also discussed the synaptic role of the unconventional myosin VIIa Crinkled, which was recently identified in a mutant screen for genes interferring with synaptic plasticity at NMJs in Drosphila. At NMJs Crinkled is localized in synapse-associated areas of the subsynaptic reticulum, where it may be involved in vesicular transport processes that control the functional integrity of glutamatergic synapses. Crinkled encodes the Drosophila homologue of the human usher 1 gene, which when mutated leads to Usher syndrome (deafness and progressive blindness).
Graeme Davis addressed the molecular basis of synapse homeostasis required for maintaining neuronal secretion and muscle excitation within physiological ranges. Impairing Drosophila muscle excitability by the expression of the Kir2.1 potassium channel does not alter muscle depolarization in response to synaptic activation. This is due to a compensatory upregulation in presynaptic release via an increase in the number of presynaptic boutons: both quantal size and motor end plate ultrastructural morphology is unchanged. Thus, a modification in the level of muscle membrane depolarization is sufficient to initiate synaptic homeostatic compensation. A screen for genes that control synapse development led to the identification of spinster, which encodes a multipass transmembrane protein present in late endosomal compartment. Spinster mutant synapses show a two-fold increase in bouton number and a deficit in neurotransmitter release. Spinster is expressed in both nerve and muscle and is required both pre- and postsynaptically for normal synaptic growth. Consistent with the endosomal localization of spinster, spinster mutants have an enlarged endosomal compartment. TGFb signaling is needed for neuromuscular junction development and prolongation of TGFb signaling may be one of the mechanisms through which spinster mutations induce synaptic overgrowth.
The conference, which ended with a summary of the presentations by two scientific journal editors, Anuradha Rao and Elena Porro, was an opportunity to discuss current knowledge about synaptic development and function at different levels of analysis. The synergistic use of a variety of experimental approaches, together with major methodological advances in molecular biology, optical and electrophysiological techniques, have yielded a first molecular insights into some aspects of synaptic physiology such as the exo- and endo-cytotic cycle of neurotransmitter containing synaptic vesicles and the mechanism of postsynaptic transduction. Studies of synaptic function have generated fall-outs in a variety of fields of biology and medicine. Much less understood are the mechanisms by which synapses are formed and stabilized and those controlling molecular turn-over at synapses. Elucidating how the findings from in vitro systems compare to the behavior of synapses in a living brain remains a major challenge. The progress made from the Neurobiology conference held at Les Treilles little more than a decade ago has been impressive. The meeting closed with the expectation that the next decade will witness an at least equally impressive pace of discovery.
By overwhelming consensus the meeting was intellectually stimulating and productive. The format of the conference, particularly the hour-long presentations, the large time allocated for discussion, the required stay of all the participants throughout the conference, and the logistic arrangements were key factors in promoting interactions and facilitating free exchange and discussion of unpublished results. The atmosphere of Les Treilles, the warm hospitality of the staff and the beauty of the environment made the conference a most pleasant and truly unique experience.