The subcellular distribution of GABARAP and its ability to interact with NSF suggest a role for this protein in the intracellular transport of GABA(A) receptors.

GABA(A) receptors the major sites of fast synaptic inhibition in the brain are composed predominately of alpha, beta, and gamma2 subunits. The receptor gamma2 subunit interacts with a 17-kDa microtubule associated protein GABARAP, but the significance of this interaction remains unknown. Here we demonstrate that GABARAP, which immunoprecipitates with GABA(A) receptors, is not found at significant levels within inhibitory synapses, but is enriched within the Golgi apparatus and postsynaptic cisternae. We also demonstrate that GABARAP binds directly to N-ethylmaleimide-sensitive factor (NSF), a protein critical for intracellular membrane trafficking events. NSF and GABARAP complexes could be detected in neurons and these two proteins also colocalize within intracellular membrane compartments. Together our observations suggest that GABARAP may play a role in intracellular GABA(A) receptor transport but not synaptic anchoring, via its ability to interact with NSF. GABARAP may therefore have an important role in the production of GABAergic synapses.

Immunolocalization of tripeptidyl peptidase II, a cholecystokinin-inactivating enzyme, in rat brain.

Tripeptidyl peptidase II (EC 3.4.14.10) is a serine peptidase apparently involved in the inactivation of cholecystokinin octapeptide [Rose C. et al. (1996) Nature 380, 403-409]. We have compared its distribution with that of cholecystokinin in rat brain, using a polyclonal antibody raised against a highly purified preparation for immunohistochemistry at the photon and electron microscope levels. Tripeptidyl peptidase II-like immunoreactivity was mostly detected in neurons, and also in ependymal cells and choroid plexuses, localizations consistent with a possible participation of the peptidase in the inactivation of cholecystokinin circulating in the cerebrospinal fluid. Immunoreactivity was mostly detected in cell bodies, large processes and, to a lesser extent, axons of various neuronal populations. Their localization, relative to that of cholecystokinin terminals, appears to define three distinct situations. The first corresponds to neurons with high immunoreactivity in areas containing cholecystokinin terminals, as in the cerebral cortex or hippocampal formation, where pyramidal cell bodies and processes surrounded by cholecystokinin axons were immunoreactive. A similar situation was encountered in many other areas, namely along the pathways through which cholecystokinin controls satiety, i.e. in sensory vagal neurons, the nucleus tractus solitarius and hypothalamic nuclei. The second situation corresponds to cholecystokinin neuronal populations containing tripeptidyl peptidase II-like immunoreactivity, as in neurons of the supraoptic or paraventricular nuclei, axons in the median eminence or nigral neurons. In both situations, localization of tripeptidyl peptidase II-like immunoreactivity is consistent with a role in cholecystokinin inactivation. The third situation corresponds to areas with mismatches, such as the cerebellum, a region devoid of cholecystokinin, but in which Purkinje cells displayed high tripeptidyl peptidase II-like immunoreactivity, possibly related to a role in the inactivation of neuropeptides other than cholecystokinin. Also, some areas with cholecystokinin terminals, e.g., the molecular layer of the cerebral cortex, were devoid of tripeptidyl peptidase II-like immunoreactivity, suggesting that processes other than cleavage by tripeptidyl peptidase II may be involved in cholecystokinin inactivation. Tripeptidyl peptidase II-like immunoreactivity was also detected at the ultrastructural level in the cerebral cortex and hypothalamus using either immunoperoxidase or silver-enhanced immunogold detection. It was mainly associated with the cytoplasm of neuronal somata and dendrites, often in the vicinity of reticulum cisternae, Golgi apparatus or vesicles, and with the inner side of the dendritic plasma membrane. Hence, whereas a fraction of tripeptidyl peptidase II-like immunoreactivity localization at the cellular level is consistent with its alleged function in cholecystokinin octapeptide inactivation, its association with the outside plasma membrane of neurons remains to be confirmed.

Subcellular distribution of survival motor neuron (SMN) protein: possible involvement in nucleocytoplasmic and dendritic transport.

Spinal muscular atrophy (SMA) is among the most common recessive autosomal diseases and is characterized by the loss of spinal motor neurons. A gene termed 'Survival of Motor Neurons' (SMN) has been identified as the SMA-determining gene. Recent work indicates the involvement of the SMN protein and its associated protein SIP1 in spliceosomal snRNP biogenesis. However, the function of SMN remains unknown. Here, we have studied the subcellular localization of SMN in the rat spinal cord and more generally in the central nervous system (CNS), by light fluorescence and electron microscopy. SMN immunoreactivity (IR) was found in the different regions of the spinal cord but also in various regions of the CNS such as the brainstem, cerebellum, thalamus, cortex and hippocampus. In most neurons, we observed a speckled labelling of the cytoplasm and a discontinuous staining of the nuclear envelope. For some neurons (e.g. brainstem nuclei, dentate gyrus, cortex: layer V) and, in particular in motoneurons, SMN-IR was also present as prominent nuclear dot-like-structures. In these nuclear dots, SMN colocalized with SIP1 and with fibrillarin, a marker of coiled bodies. Ultrastructural studies in the anterior horn of the spinal cord confirmed the presence of SMN in the coiled bodies and also revealed the protein at the external side of nuclear pores complexes, in association with polyribosomes, and in dendrites, associated with microtubules. These localizations suggest that, in addition to its involvement in the spliceosome biogenesis, the SMN protein could also play a part in nucleocytoplasmic and dendritic transport.

Presence of the vesicular inhibitory amino acid transporter in GABAergic and glycinergic synaptic terminal boutons.

The characterization of the Caenorhabditis elegans unc-47 gene recently allowed the identification of a mammalian (gamma)-amino butyric acid (GABA) transporter, presumed to be located in the synaptic vesicle membrane. In situ hybridization data in rat brain suggested that it might also take up glycine and thus represent a general Vesicular Inhibitory Amino Acid Transporter (VIAAT). In the present study, we have investigated the localization of VIAAT in neurons by using a polyclonal antibody raised against the hydrophilic N-terminal domain of the protein. Light microscopy and immunocytochemistry in primary cultures or tissue sections of the rat spinal cord revealed that VIAAT was localized in a subset (63-65%) of synaptophysin-immunoreactive terminal boutons; among the VIAAT-positive terminals around motoneuronal somata, 32.9% of them were also immunoreactive for GAD65, a marker of GABAergic presynaptic endings. Labelling was also found apposed to clusters positive for the glycine receptor or for its associated protein gephyrin. At the ultrastructural level, VIAAT immunoreactivity was restricted to presynaptic boutons exhibiting classical inhibitory features and, within the boutons, concentrated over synaptic vesicle clusters. Pre-embedding detection of VIAAT followed by post-embedding detection of GABA or glycine on serial sections of the spinal cord or cerebellar cortex indicated that VIAAT was present in glycine-, GABA- or GABA- and glycine-containing boutons. Taken together, these data further support the view of a common vesicular transporter for these two inhibitory transmitters, which would be responsible for their costorage in the same synaptic vesicle and subsequent corelease at mixed GABA-and-glycine synapses.

Localization of components of glycinergic synapses during rat spinal cord development.

The sequence of events leading to the chemical matching of presynaptic neurotransmitters and postsynaptic transmitter receptors is investigated here in vivo for the spinal glycine receptor (GlyR) by using immunocytochemical methods. In the ventral horn of adult rat spinal cord, GlyRs are only present at glycinergic postsynaptic differentiations where they are stabilized by the associated protein gephyrin. With quantitative confocal microscopy, we found that gephyrin is detected before GlyRs at embryonic day (E)13-E14 and at E15, respectively, inside the cytoplasm and at plasmalemmal loci. Around the time of birth, the number of cell surface gephyrin-immunoreactive (-IR) spots exceeds that of GlyR. They first match 10 days after birth. The densities of postsynaptic gephyrin- and GlyR-IR were quantified between birth and the adult stage with post-embedding immunogold staining. Immunostaining for gephyrin and GlyR was not detected in the extrasynaptic membrane. The density of staining in postsynaptic membrane increased progressively with development. The inhibitory amino-acid content of the presynaptic terminal boutons opposed to gephyrin-IR sites was also analyzed. In the newborn, postnatal day 10, and adult, more than 90% of these boutons were immunostained for glycine. As seen with serial sections, 38% and 51.2% of the terminals also contained gamma-aminobutyric acid (GABA) in neonate and adult, respectively. These data indicate that around the time of birth, most glycine-containing boutons, some also containing GABA, are opposed to gephyrin-IR postsynaptic densities, whereas GlyRs are not present. Our results suggest that gephyrin determines subsynaptic loci on the plasma membrane where GlyR will subsequently accumulate.

Morphofunctional evidence for mature synaptic contacts on the Mauthner cell of 52-hour-old zebrafish larvae.

In a previous study, miniature inhibitory synaptic events recorded in the Mauthner cell of the 52-hour-old zebrafish larvae (Brachydanio rerio) were found to be mainly glycinergic. Their amplitude distribution was not Gaussian and it was proposed that their large amplitude variation might reflect the activation of immature synapses. However, ultrastructural studies of the synaptic contacts over the M-cell soma of 52 h larvae described here, revealed that numerous synaptic contacts on this neuron are already mature at this developmental stage and that most of them already contain a single active zone. As in the adult goldfish, immunohistochemistry indicates the presence of both glycine- and GABA-immunoreactive boutons which establish synaptic contacts. We also found that, in addition to the predominant glycinergic postsynaptic inhibitory currents, some postsynaptic currents are also GABAergic since they are specifically inhibited by bicuculline (20 microM). GABAergic miniature events (time to peak close to 0.8 ms and decay time-constant close to 45 ms) were only detected in the presence of 11.5 mM [KCl]o. Their amplitude distributions were well fitted by one, or at most two, Gaussian curves. Outside-out recordings showed one class of GABA receptors with a main conductance state of 23 pS. This indicates that the smallest GABAergic miniature inhibitory synaptic events correspond to the opening of 14-20 chloride channels Pre- and postsynaptic factors which contribute to the predominance of glycinergic synaptic currents over GABAergic ones in untreated preparations and to the striking differences between their frequencies and their respective amplitude distribution histograms are discussed with reference to the morphological characteristics of the mature synaptic endings impinging on this still developing neuron.

Gephyrin accumulates at specific plasmalemma loci during neuronal maturation in vitro.

The distribution of glycine receptor (GlyR)-associated gephyrin has been investigated in rat spinal cord neurons maintained in vitro by means of immunocytochemical techniques. Gephyrin, which is crucial for the stabilization of postsynaptic GlyR microdomains, is present in mature neurons at postsynaptic differentiations. With immunofluorescence, discontinuous patches of gephyrin were detected within the neuronal soma of spinal cord neurons on the 1st day after plating. Subsequently, gephyrin was present at membrane areas that correspond to points of contact between cells or with the culture dish. By the 5th day, gephrin was mostly associated with the MAP2-positive somatodendritic compartment. With immunoelectron microscopy, gephyrin blobs detected at the earliest stages (1-3 days after plating) were found within the cytoplasm or associated with the plasma membrane. Asymmetrically immunostained intercellular contacts were only detected after 5 days, and gephyrin was found in association with clearly differentiated postsynaptic membranes at 7 days. At later stages, we observed gephyrin immunoreactivity only at some synapses. Our results suggest that gephyrin accumulates initially at the locus of cell-to-cell contacts involved in adhesion processes. These localizations may define hot spots for later accumulation of the GlyR and possibly other receptors.