Synaptic Development Never Stops
Synapse (from Greek σύναψις sunapsis, ‘together, joining’.) n. A junction between two nerve cells, consisting of a minute gap across which impulses pass by diffusion of a neurotransmitter (English Oxford Dictionary).
The synapse is the site for neuronal transmission. Based on the structural characteristics of the pre- and postsynaptic terminals observed by electron microscopy, two major types of chemical synapses were described in the brain: type I (asymmetric) and type II (symmetric) synapses (Figure 1A and B, respectively; (Gray, 1959)). The type I synapse underlies excitatory transmission whereas the type II is inhibitory. Excitatory neurons, through type I synapses, stimulate other neurons to respond and transmit electrical messages while inhibitory neurons through type II suppress responsiveness, preventing excessive firing. Both types of synapses can be made at the site of contact between two nerve cells that occurs between an axon and a dendrite (axo-dendritic). In addition, inhibitory synapses are also made between the axon and the soma (axo-somatic) or between the axon and the axonal initial segment (axo-axonic, (Knott et al., 2002)) of the receiving neuron. A schematic representation of both types of synapses is presented in Figure 1C and D.
The presynaptic compartment of the synapse can be divided into three components: the axonal boutons, the vesicles and the active zone. The axonal boutons are distal terminations of the branches of an axon that contain vesicles, which are secretory organelles whose role is to deliver neurotransmitters. In the excitatory synapse, the form of the vesicles is described as spherical, while the vesicles look more flattened or elongated in the inhibitory synapses. Three forms of boutons can be found: single synaptic boutons with a single postsynaptic partner, multi-synaptic boutons that present multiple postsynaptic partners and non-synaptic boutons with no partner (Bourne et al., 2013). The active zone is the area for the unloading of the vesicles, and is a specialized region on the presynaptic plasma membrane positioned in alignment with the postsynaptic density (Landis, 1988).
The synaptic cleft is the extracellular part of the synapse and is a conglomeration of secreted proteins and external regions of pre- and postsynaptic transmembrane proteins that can be visualized by electron microscopy, for example, on frozen hippocampal slices from Rattus norvegicus (Zuber et al., 2005).
The postsynaptic compartment is constituted by the postsynaptic membrane and the postsynaptic density (PSD), a complex scaffold of proteins. The postsynaptic membrane is covered with neurotransmitter receptors and other types of transmembrane proteins. The PSD of the excitatory synapse is thick and often found on dendritic spines. In the inhibitory synapse, the PSD is thinner, and most of the time there is no membrane protuberance at the postsynaptic side. The molecular complexity of the excitatory PSD is very high in comparison with the inhibitory one. A proteomic analysis of purified PSD fractions by liquid chromatography coupled with mass spectrometry identified 374 PSD proteins. These ~400 proteins were categorized into the following 13 functional groups: actin cytoskeleton (12%), translation (11%), GTPases and regulators (10%), cell adhesion (9%), other cytoskeleton (8%), scaffolds (7%), receptors and channels (7%), membrane trafficking (6%), mitochondria (6%), kinases/phosphatases, and regulators (5%), motor proteins (3%), metabolism (2%) and chaperone (1%; (Peng et al., 2004)).
The neurotransmitter receptors can be divided into two main groups depending on their architecture and type of activity. The ionotropic receptors are ligand-gated ion channels, their action is fast. The receptor/channel pore opens after the conformational change that happens when the ligand interacts with the binding site. The channel allows the ions flux, that can be translated into fast activity and is either excitatory or inhibitory. The excitatory receptors are mainly sensitive to glutamate and its variants. The ionotropic glutamate receptors (iGluRs) are grouped and named according to their preferential agonist, such as: N-Methyl-D-aspartic acid (NMDA) receptors (NMDARs), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (AMPARs) and kainate receptors (KARs). There are four different AMPAR subunits (GluA1-GluA4), six NMDAR subunits (NR1, NR2A-D and NR3) and five KAR subunits (GluA5-7 and KA1-2). There are other sub-groups within the iGluRs, such as the Delta receptors, that comprise of two members GluRδ1 and GluRδ2. GluRδ2 is called an orphan receptor because it does not bind to glutamate analogs (Yuzaki, 2003). The second group of receptors are metabotropic, which are G-protein coupled receptors. In this case, the binding of the ligand initiates a molecular signaling pathway. The metabotropic glutamate receptors (mGluRs) can be divided into three different Groups (I to III), containing a total of eight known or well-described subtypes of receptors. Interestingly, the mGluRs from the group I (mGluR1 and mGluR5) are mainly expressed at the postsynaptic side, while the mGluRs from the groups II (mGluR2 and mGluR3) and III (mGluR4, mGluR6, mGluR7 and mGluR8) are mainly expressed at the presynaptic side.
Synapse formation, or synaptogenesis, is a highly precise and tightly regulated process. It begins with the first contact between two neurons and requires the coordination of many cellular and molecular biological events, including cytoskeletal rearrangements and recruitment of proteins, to form the scaffolding machineries at both the pre- and the postsynaptic sides. Synapse maturation then allows the recruitment of the right neurotransmitter receptors. The mature synapse can further be changed by mechanisms of plasticity. The main steps during the life of the excitatory synapse are represented in Figure 2 there is a molecular aggregation to form the scaffolding machineries of the pre- and the postsynaptic terminals. Finally, neurotransmitter receptors are recruited. In general, immature excitatory synapses present only NMDA-type glutamate receptors. The maturation implicates the recruitment of AMPA-type glutamate receptors to the postsynaptic surface. The mature synapse still changes through mechanisms of plasticity such as long-term potentiation (LTP) and long-term depression (LTD).
The Formation and Maturation of Excitatory Synapses
Synaptogenesis was first described at the neuromuscular junction (NMJ). In this system, the contact of axonal growth cones with myotubes triggers the molecular cascades that allow the specialization of both the presynaptic bouton and the postsynaptic receptor apparatus (the PSD of the NMJ). The organization of the presynaptic side requires the clustering of the synaptic vesicles and the creation of the specialized cytoskeletal matrix at the active zone. Similarly, postsynaptic differentiation involves the assembly of the specialized cytoskeletal matrix, necessary for the clustering of the receptors. The newly formed NMJ now has the ability to release important factors, such as the protein AGRIN (a secreted heparan sulfate proteoglycan; named based on its involvement in the aggregation of acetylcholine receptors) and neurotransmitters, which are thought to promote the differentiation of the complex postsynaptic junction (for review, see (Sanes and Lichtman, 1999)). Synaptogenesis at the CNS follows the same main steps. First, the contact between the two neurons is mediated by several classes of cell adhesion molecules, such as the prototypic Neurexin and Neuroligin, the Ephrins and its receptors or the immunoglobulin domain-containing proteins (Tallafuss et al., 2010). Both neurons present molecules in its terminals that interact to form trans-synaptic complexes. The trans-synaptic interactions allow the physical approach between the two membranes and the communication between the two neurons. Once the two membranes are closer to each other, the molecular aggregation of the scaffolding machineries occurs and is orchestrated by specific molecules, such as the presynaptic Bassoon (Friedman et al., 2000) or the postsynaptic PSD95 (Chen et al., 2011). Finally, the neurotransmitter receptors are recruited. In general, it is thought that glutamate receptors are recruited to nascent synapses hours after initial axodendritic contact (Friedman et al., 2000).
H.V. Friedman and colleagues proposed a time course for excitatory synapse formation in cultured hippocampal neurons. They used a fluorescent endocytic dye (FM 4-64) that labels the synaptic vesicles which display the capacity to recycle after presynaptic activation, the “newly formed vesicles”. In parallel, they followed the time course of different synaptic proteins. Fixing the t0 (in minutes) at the axodendritic contact, they observed that the presynaptic side is assembled first, prior to the post synaptic side. The precise timing was found to be t10 – t25 for Bassoon, t25 – t35 for the “newly formed vesicles”, t30 – t55 for the accumulation of the synaptic vesicles, t75 – t110 for the PSD proteins SAP90/PSD95 and t85 – t105 for the glutamate receptors GluA1 and NR1 (Figure 3; (Friedman et al., 2000)). They observed that only two hours are needed for the synaptogenesis to be completed from the first contact to the ‘mature synapse’. Interestingly, the insertion of the postsynaptic receptors (including NMDAR and AMPAR) overlaps with the aggregation of the PSD scaffolding proteins, suggesting that synaptogenesis and synapse maturation (in terms of AMPAR recruitment) can occur simultaneously.
The synapse is considered mature when all the types of molecules that will be maintained in adulthood, are already recruited. However, the number of some of these synaptic molecules changes with time and experience. During synapse maturation, there are changes in the subunit composition of the receptors. NMDARs formed exclusively by NR1 and NR2B dominate during synaptogenesis. There is a switch between the subunits NR2B and NR2A during synapse maturation. The subunit NR2B inhibits the insertion of AMPAR at the membrane during synaptogenesis (Hall et al., 2007), ensuring that immature synapses do not recruit AMPARs. The time course in the recruitment of the different glutamate receptors have been studied in cultured cortical neurons. Interestingly, both types of iGluRs (NMDARs and AMPARs) form clusters that are present in the somatodendritic compartment before synaptogenesis; these clusters increase in number and in size with synaptogenesis. It has been shown that the NR1 clusters travel at the rate of ~4 µm/min while GluA1 clusters are slower, ~2 µm/min (Washbourne et al., 2002). The receptor clusters can travel in both directions, from the soma to the neurites and vice versa. Still-developing excitatory synapses contain only NMDARs (Isaac et al., 1995; Liao et al., 1995). AMPAR-lacking synapses are functionally dormant and are considered postsynaptically “silent synapses”. The unsilencing occurs through a process of experience-dependent insertion of AMPARs (reviewed in (Hanse et al., 2013)). Silent synapses are the substrate for activity-dependent strengthening of synaptic transmission at excitatory synapses in the hippocampus and cortex (Isaac et al., 1995; Liao et al., 1995).
The synaptic strength changes constantly in a use-dependent manner: this remarkable ability of the nervous system is referred to as synaptic plasticity. Synaptic plasticity affects the activity of neuronal networks and, ultimately, the behavior of the whole animal. There are many types of synaptic plasticity mechanisms that can be grouped depending on different criteria. Depending on the type of change in the strength, synaptic plasticity can either strengthen (synaptic potentiation or facilitation) or weaken (synaptic depression) a synapse. Synaptic plasticity can be divided into two groups depending on the timescale. Short-term plasticity occurs on a scale of milliseconds to few minutes, while long-term plasticity is on a longer timescale ranging from ~30 minutes to hours. Depending on the synaptic structures, synaptic plasticity can also be classified as homosynaptic plasticity, when it happens in only one type of synapses, or heterosynaptic plasticity, when there are different synaptic types implicated. For example, there are cases in which a modulatory neuron will form an axoaxonic synapse with the presynaptic element. In this case, heterosynaptic plasticity can take place: the modulatory neuron can produce plastic changes in the presynaptic terminal, such as facilitation or inhibition. Depending on the origin of the induction, synaptic plasticity can occur constitutively or be triggered by neuronal activity. Development, stress and hormones, are examples of chronic activity and can elicit a form of synaptic plasticity that is called homeostatic synaptic plasticity. The basic mechanism for activity-dependent synaptic plasticity was proposed by D.O. Hebb in 1949 in his postulate: when two neurons are active together, the efficacy of the synaptic transmission between them improves. Therefore, activity-dependent long-term potentiation is also referred as Hebbian plasticity. Two molecular mechanisms may explain activity-dependent synaptic plasticity. The first is due to changes in the presynaptic side such as changes in the amount of released neurotransmitter or the release efficacy. The second is due to changes at the postsynaptic side, such as changes in the number or in the function of the neurotransmitter receptors. After briefly introducing short-term forms of synaptic plasticity, I will present the two most studied forms of activity-dependent homosynaptic long-term synaptic plasticity: long-term potentiation and depression.
Short-Term Synaptic Plasticity
Short-term synaptic enhancement represents an increased probability of presynaptic neurotransmitter release in response to presynaptic action potentials and short-term synaptic depression is usually attributed to the depletion of readily releasable presynaptic vesicles (reviewed in (Regehr, 2012)). There are different forms of short-term synaptic plasticity, lasting in the order of hundreds to thousands of milliseconds. Paired pulse facilitation (PPF) or paired pulse depression (PPD) occur when there are two stimuli not much separated in time (milliseconds), that produce an increase or a decrease in the amplitude of the second postsynaptic potential/current compared to the first. Synaptic augmentation is the increase in the presynaptic neurotransmitter release probability during and after repetitive stimulation occurring in a timespan between milliseconds and several minutes. The post-tetanic potentiation consists of an increase in the frequency of miniature postsynaptic potentials (with no changes in the amplitude) or currents due to a high frequency train of stimulations (tetanic stimulation) lasting from ~0.2 to 5 seconds.
Long-Term Synaptic Plasticity
Long-term potentiation (LTP) and depression (LTD) are the two most studied forms of long-term synaptic plasticity. In mature neurons, the PSD composition undergoes continual molecular turnover under basal conditions and shows larger changes in response to activity (Inoue and Okabe, 2003). Proteins of the PSD turn over in large part by continuous exchange with counterparts outside the PSD. PSD95, for example, is dynamically exchanged between neighboring PSDs in cortical neurons in vivo (Gray et al., 2006). Among the most dynamic proteins in the PSDs are the AMPARs, which show rapid diffusion in and out of the postsynaptic membrane to regulate the synaptic strength. Regulated AMPAR insertion into and removal from the PSD are major mechanisms underlying the strengthening and weakening of synaptic transmission and thus synaptic plasticity. While there are different pathways to change the strength of synapses to induce LTD and LTP, such as changes in the conductivity of the pore of the AMPAR, here I will focus on the mechanisms for LTD and LTP caused postsynaptically by the removal or insertion of AMPARs, respectively (for review, see (Fleming and England, 2010; Malinow and Malenka, 2002)). These changes in the amount of AMPARs are schematically represented in Figure 4.
The key factor in the induction of LTD or LTP is the frequency of the stimulation: high-frequency stimulation triggers the insertion of AMPARs and low-frequency stimulation leads to the removal of AMPARs in hippocampal Schaffer collateral/CA1 synapses (Dudek and Bear, 1993). Indeed, long-term synaptic plasticity can be induced bidirectionally at the same synapse. When LTP or LTD are triggered by the NMDARs, it is called NMDAR-dependent. Under normal conditions (resting hyperpolarized membrane potentials), the NMDARs that have been activated by the glutamate binding are blocked with Mg++. Depolarization of the membrane allows the removal of the Mg++ block and, once the NMDAR pore is opened, Ca++ flows inside the neuron. The role of the different AMPAR subunits in hippocampal NMDAR-dependent long-term synaptic plasticity has been recently confirmed using a knock-in (KI) strategy by Zhou et al. Using a GluA1C2KI mouse (in which both GluA1 and GluA2 subunits have the intracellular domain of GluA2) and a GluA2C1KI mouse (in which both subunits contain the GluA1 internal tail), they demonstrated that the carboxyl-terminus domain of GluA1 is necessary for NMDAR-dependent LTP while the carboxyl-terminus domain of GluA2 is required for NMDAR-dependent LTD. Importantly, both phenotypes were restored in the double KI mice, in which both carboxyl-terminus domains are expressed but switched (Zhou et al., 2018).
Table of contents :
LIST OF FIGURES AND TABLES
LIST OF ABBREVIATIONS
PART 1: SYNAPTIC DEVELOPMENT NEVER STOPS
1. The Formation and Maturation of Excitatory Synapses
2. Synaptic Plasticity
2.1. Short-Term Synaptic Plasticity
2.2. Long-Term Synaptic Plasticity
3. AMPA-type Receptor Turnover
3.1. AMPA-type Receptors
3.2. AMPA-type Receptor Turnover
3.2.1. Delivery of AMPARs to the Synapse
3.2.2. AMPARs Membrane Insertion
3.2.3. Surface AMPARs: Stabilization, Anchoring and Clustering
3.2.4. Lateral Diffusion of AMPARs
3.2.5. Endocytosis of the AMPAR: Internalization
3.2.6. AMPARs Recycling
3.2.7. Degradation of AMPARs
PART 2: THE OLIVOCEREBELLAR NETWORK AS A MODEL TO STUDY EXCITATORY SYNAPSE FORMATION AND FUNCTION
1. Organization of the Olivocerebellar Network
1.1. Global Structure of the Cerebellum and Inferior Olive
1.2. Histology of the Cerebellar Cortex
1.3. Cerebellar Efferences and Afferences
1.4. Afferences and Efferences of the Inferior Olive
2. The Life of Purkinje Cell Excitatory Synapses
2.1. The Parallel Fiber/Purkinje Cell Synapse
2.2. The Climbing Fiber/Purkinje Cell Synapse
3. Cerebellar Functions
3.1. Sensorimotor Cerebellar Functions
3.2. Cognitive Cerebellar Functions
PART 3: SUSHI DOMAIN PROTEINS IN THE NERVOUS SYSTEM
1. The Sushi Domain
2. Sushi Domain-containing Proteins are Evolutionarily Conserved in the Nervous System
2.1. Caenorhabditis elegans: LEV-9
2.2. Drosophila melanogaster: Hig and Hasp
2.3.1. GABAB Receptors
2.3.4. SUSD Family
2.3.5. CSMD Family
PART 1: ROLE OF THE SUSHI DOMAIN PROTEIN SUSD4 IN THE BRAIN
1. Article in Preparation
2. Effect of SUSD4 on Transmission and GluA2 Content in Climbing Fiber Synapses Changes with Maturation
3. Susd4 Deletion Leads to a Defect on a Hippocampal-dependent Behavioral Task
4. Materials and Methods
PART 2: MASP1/3 IS EXPRESSED IN THE OLIVOCEREBELLAR SYSTEM
5. Experimental Procedures
1. Sushi Domains in the Nervous System
2. SUSD4 Controls Activity-dependent AMPAR Degradation
3. Conclusion and Future Directions.
ANNEXE 1: THE SECRETED PROTEIN C1QL1 AND ITS RECEPTOR BAI3 CONTROL THE SYNAPTIC CONNECTIVITY OF EXCITATORY INPUTS CONVERGING ON CEREBELLAR PURKINJE CELLS
ANNEXE 2: EXPRESSION AND ROLE OF GALECTIN-3 IN THE POSTNATAL DEVELOPMENT OF THE CEREBELLUM