Postsynaptic scaffolding molecules modulate the localization of neuroligins.

Previous work has shown an important role for neuroligins in promoting the formation of synaptic connections in cultured cells. Although neuroligins enhance both excitatory and inhibitory synapse formation, individual neuroligin isoforms have been shown to preferentially localize to either glutamatergic or GABAergic synapses. Current evidence points to an important role for both the extracellular and intracellular domains of neuroligins in their synaptic localization. Although postsynaptic density protein 95 (PSD-95) has been shown to be involved in the recruitment of neuroligin 1 to excitatory synapses, the localization of neuroligin 2 (NL2) and neuroligin 3 (NL3) to excitatory and inhibitory synapses is less well defined. We assessed the roles of gephyrin and PSD-95, postsynaptic scaffolding molecules exclusively localized to inhibitory and excitatory synapses, respectively, in localizing NL2 and NL3 in primary neuronal cultures. We demonstrate that knockdown of gephyrin results in a significant shift of NL2 from inhibitory to excitatory synaptic contacts, while knockdown of PSD-95 leads to a partial shift of NL2 and NL3 from excitatory to inhibitory contacts. Furthermore, analysis of specific domain deletions within the C-terminal, intracellular domain of NL2 reveals that the region between amino acids 716 and 782 is required for the normal synaptic clustering of this protein. Together, these data suggest that intracellular mechanisms are involved in the targeting of different neuroligin family members to synapses (216).

Dystroglycan contributes to the formation of multiple dystrophin-like complexes in brain.

In muscle, dystrophin anchors a complex of proteins at the cell surface which includes alpha-dystroglycan, beta-dystroglycan, syntrophins and dystrobrevins. Mutations in the dystrophin gene lead to muscular dystrophy and mental retardation. In contrast to muscle, little is known about the localization and the molecular interactions of dystrophin and dystrophin associated proteins (DAPs) in brain. In the present study, we show that alpha-dystroglycan and dystrophin are localized to large neurones in cerebral cortex, hippocampus, cerebellum and spinal cord. Furthermore, we show that dystroglycan is a member of three distinct dystrophin-containing complexes. Two of these complexes contain syntrophin and both dystrophin and syntrophin are enriched in post-synaptic densities. These data suggest that dystrophin and DAPs may have a role in the organization of CNS synapses. Interestingly, the enrichment for syntrophin in post-synaptic densities is not affected in mice mutant for all dystrophin isoforms. Thus in the brain, unlike in muscle, the association of syntrophin with dystrophin is not crucial for the DAP complex which suggests that it may be associated with other proteins.

Neural regulation of alpha-dystroglycan biosynthesis and glycosylation in skeletal muscle.

Alpha-dystroglycan (alpha-DG) is part of a complex of cell surface proteins linked to dystrophin or utrophin, which is distributed over the myofiber surface and is concentrated at neuromuscular junctions. In laminin overlays of muscle extracts from developing chick hindlimb muscle, alpha-DG first appears at embryonic day (E) 10 with an apparent molecular mass of 120 kDa. By E15 it is replaced by smaller (approximately 100 kDa) and larger (approximately 150 kDa) isoforms. The larger form increases in amount and in molecular mass (>200 kDa) as the muscle is innervated and the postsynaptic membrane differentiates (E10-E20), and then decreases dramatically in amount after hatching. In myoblasts differentiating in culture the molecular mass of alpha-DG is not significantly increased by their replication, fusion, or differentiation into myotubes. Monoclonal antibody IIH6, which recognizes a carbohydrate epitope on alpha-DG, preferentially binds to the larger forms, suggesting that the core protein is differentially glycosylated beginning at E16. Consistent with prior observations implicating the IIH6 epitope in laminin binding, the smaller forms of alpha-DG bind more weakly to laminin affinity columns than the larger ones. In blots of adult rat skeletal muscle probed with radiolabeled laminin or monoclonal antibody IIH6, alpha-DG appears as a >200-kDa band that decreases in molecular mass but increases in intensity following denervation. Northern blots reveal a single mRNA transcript, indicating that the reduction in molecular mass of alpha-DG after denervation is not obviously a result of alternative splicing but is likely due to posttranslational modification of newly synthesized molecules. The regulation of alpha-DG by the nerve and its increased affinity for laminin suggest that glycosylation of this protein may be important in myofiber-basement membrane interactions during development and after denervation.

alpha-dystroglycan isoforms are differentially distributed in adult rat retina.

alpha-Dystroglycan (alpha -DG) is a laminin/agrin receptor expressed in skeletal muscle as well as in nervous system and other tissues. Glycosylation of the core protein of alpha-DG is extensive, variable from tissue to tissue, and functionally relevant. To address differential glycosylation of alpha-DG in the retina, we have investigated the distribution of this protein using two different antibodies: 1B7 directed against the core protein of alpha-dystroglycan, and IIH6 directed against a carbohydrate moiety (Ervasti and Campbell [1993] J Cell Biol 122:809-823). Monoclonal antibody 1B7 recognizes a broader band than IIH6, which seems to recognize only a subset of alpha-DG forms in retina. These data reflect the existence of differentially glycosylated isoforms of alpha-DG. Monoclonal antibody 1B7 shows an extensive staining for alpha-DG in the inner limiting membrane as well as in the ganglion cell and inner plexiform layers labeling Müller cell processes, whereas monoclonal antibody IIH6 staining is restricted to the inner limiting membrane and blood vessels. Our data indicate that there are distinct isoforms of alpha-DG that are localized in apposition to basal lamina in the inner limiting membrane and blood vessels or within the parenchyma of the retina along Müller glia. Both isoforms are expressed in a Müller cell line in culture and coimmunoprecipitate with beta-dystroglycan. These data suggest that DGs may participate in organizing synapses and basement membrane assembly in the retina.

Chimaeric mice deficient in dystroglycans develop muscular dystrophy and have disrupted myoneural synapses.

Mutations in the dystrophin gene (DMD) and in genes encoding several dystrophin-associated proteins result in Duchenne and other forms of muscular dystrophy. alpha-Dystroglycan (Dg) binds to laminins in the basement membrane surrounding each myofibre and docks with beta-Dg, a transmembrane protein, which in turn interacts with dystrophin or utrophin in the subplasmalemmal cytoskeleton. alpha- and beta-Dgs are thought to form the functional core of a larger complex of proteins extending from the basement membrane to the intracellular cytoskeleton, which serves as a superstructure necessary for sarcolemmal integrity. Dgs have also been implicated in the formation of synaptic densities of acetylcholine receptors (AChRs) on skeletal muscle. Here we report that chimaeric mice generated with ES cells targeted for both Dg alleles have skeletal muscles essentially devoid of Dgs and develop a progressive muscle pathology with changes emblematic of muscular dystrophies in humans. In addition, many neuromuscular junctions are disrupted in these mice. The ultrastructure of basement membranes and the deposition of laminin within them, however, appears unaffected in Dg-deficient muscles. We conclude that Dgs are necessary for myofibre survival and synapse differentiation or stability, but not for the formation of the muscle basement membrane, and that Dgs may have more than a purely structural function in maintaining muscle integrity.

Serotonin axons of the neostriatum show a higher affinity for striatal than for ventral mesencephalic transplants: a quantitative study in adult and immature recipient rats.

We previously showed that grafts of fetal ventral mesencephalic tissue are practically not innervated by host serotonin (5-HT) axons after implantation into the striatum of rats aged more than 14 days, at variance with transplants of cortical or striatal tissue into the adult striatum, which are well innervated by these axons. Using 5-HT immunohistochemistry and in vitro [3H]5-HT uptake/autoradiography, we have examined and quantified the innervation of ventral mesencephalic versus striatal grafts several months after implantation into the striatum of neonatal (postnatal day 5 or P5), juvenile (P15), and adult rats. Ventral mesencephalic grafts implanted in P5 rats received a moderate 5-HT innervation, while similar grafts implanted in P15 or adult recipients were almost free of any 5-HT fibers (-80%, compared to P5). The density of 5-HT innervation showed a tendency toward higher values in striatal than in ventral mesencephalic grafts (1.6-2 times higher in P5 and adult recipients; 4 times higher in P15 recipients). The difference was more striking, and significant, when only the true striatal portions of the striatal grafts were considered, i.e., DARPP-32-immunopositive areas (4-5 times higher in P5 and adult recipients; 10 times higher in P15 recipients). Accordingly, these DARPP-32-positive areas were also more densely innervated than the DARPP-32-negative zones of the same grafts (3 times higher at any age). The 5-HT innervation density also decreased with increasing age of the recipients in DARPP-32-positive, as well as DARPP-32-negative compartments of the striatal grafts (-75% in adults), but this decrease appeared more gradual (-50% in juveniles) than with mesencephalic grafts. It is concluded that the 5-HT axons innervating the neostriatum have a better affinity for striatal grafts than for ventral mesencephalic grafts or the nonstriatal portions of striatal grafts. In adulthood, the relative affinity of these axons for the different types of grafts is maintained, even though their growth capacity decreases irrespective of the target tissue considered. This experimental model may prove useful for the identification of the receptors and ligands that are responsible for target recognition by 5-HT axons and to test the possibility that the progressive decrease of axonal growth capacity from neonatal age to adulthood be related to a downregulation of such molecules.

Quantitative and morphometric data indicate precise cellular interactions between serotonin terminals and postsynaptic targets in rat substantia nigra.

We have quantified the density of serotonin axonal varicosities, their synaptic incidence and their distribution among potential targets in the pars reticulata and pars compacta of the rat substantia nigra. Serotonin axonal varicosities, counted at the light microscopic level following in vitro [3H]serotonin uptake and autoradiography, amounted to 9 x 10(6)/mm3 in the pars reticulata and 6 x 10(6)/mm3 in the pars compacta, among the densest serotonin innervations in brain. As determined at the electron microscopic level following immunolabelling for serotonin, virtually all serotonin varicosities in the pars reticulata and 50% of those in the pars compacta formed a synapse, essentially with dendrites. The combination of serotonin immunocytochemistry with tyrosine hydroxylase immunolabelling of dopamine neurons reveals that 20% of the serotonin synaptic contacts in the pars reticulata are on dopamine dendrites and 6% are on a type of unlabelled dendrite characterized by its peculiarly high cytoplasmic content of microtubules. The comparison of the diameter of the dendritic profiles that were in synaptic contact with serotonin-immunoreactive varicosities with the diameter of all other dendritic profiles of the same type suggests that serotoninergic varicosities innervate dopamine dendrites uniformly along their length, whereas they tend to contact microtubule-filled dendrites in more proximal regions and the other, unidentified dendrites in more distal regions. Furthermore, the size of the serotonin-immunoreactive varicosities and of their synaptic junctions is significantly smaller on dopamine dendrites and larger on microtubule-filled dendrites than on other, unidentified dendrites, indicating that the nature of the postsynaptic target is an important determinant of synaptic dimensions. These data should help to clarify the role of serotonin in the nigral control of motor functions. They indicate that this dense serotonin input to the substantia nigra is very precisely organized, acting through both "non-junctional" and "junctional" modes of neurotransmission in the pars compacta, which projects to the neostriatum and the limbic system, whereas the predominant mode of serotonin transmission appears to be of the "junctional" type in the pars reticulata, where serotonin can finely control the motor output of the basal ganglia by acting on the GABA projection neurons either directly or through the local release of dopamine by dopaminergic dendrites. The data also raise the possibility that the postsynaptic targets have trophic retrograde influences on serotoninergic terminals.

Efficient immunodetection of various protein antigens in glutaraldehyde-fixed brain tissue.

Optimal ultrastructural preservation of brain tissue for electron microscopy is best achieved with fixatives containing high concentrations of glutaraldehyde, which is generally considered detrimental to the immunogenicity of most protein antigens. We tested seventeen mono- or polyclonal antibodies against peptide or protein antigens, including a majority for which immunoreactivity had previously been reported to be sensitive to glutaraldehyde fixation. Forebrain sections of rats or mice fixed by perfusion with 3.5% glutaraldehyde were processed for pre-embedding immunocytochemistry by the avidin-biotin method. The resulting immunostaining was in most cases at least similar to that obtained in sections fixed with paraformaldehyde. Immunoreactivity against the mouse or human neurofilament protein NF-L was even improved, being similar to that previously reported for unfixed brain tissue. Of all antigens tested, only choline acetyltransferase, phenylethanolamine-N-methyl transferase, and neuropeptide Y were detected with lower sensitivity than after paraformaldehyde fixation, which was attributed to a rather restricted penetration of the primary antibody into glutaraldehyde-fixed tissue sections. These results indicate that glutaraldehyde may be envisaged as a possible fixative for optimal immunocytochemical detection of any tissue antigen at the electron microscopic level, including antigens which, on the basis of results obtained after fixation with paraformaldehyde-glutaraldehyde mixtures, were considered highly sensitive to glutaraldehyde fixation.

Complex deficits on reaction time performance following bilateral intrastriatal 6-OHDA infusion in the rat.

The present study examined the ability of rats subjected to bilateral 6-hydroxydopamine lesions of the terminal area of the nigrostriatal dopamine system to perform a prelearned reaction time task. This lesion model, the induction of a partial dopamine denervation of the striatum (74% depletion of dopamine striatal tissue content) with a retrograde degeneration of dopamine cell bodies in the substantia nigra, sparing the mesolimbic dopaminergic pathway, closely approximates the neuronal degeneration observed in human idiopathic Parkinson's disease. Rats were trained previously to release a lever, within a reaction time limit, after the presentation of a visual cue through reinforcement with food pellets. The onset of the light stimulus varied randomly after an unpredictable delay period of 0.25-1.0 s. Rats with dopaminergic lesions showed moderate to extensive performance deficits which were not compensated for the five postoperative weeks. More than half of the lesioned animals (64%) showed severe deficits, characterized by a concomitant increase in the number of anticipated (premature release of the lever before the visual cue) and delayed responses (lever release after the reaction time limit) with shortened reaction times in some cases. A smaller proportion (36%) of lesioned animals exhibited mild impairment of performance with a large increase in delayed responses and lengthening of reaction times but with no change in the number of anticipated responses. Asymmetric lesions had no effect on the reaction time performance. Examination of tyrosine hydroxylase immunostaining revealed that in the most impaired animals dopamine depletion was extensive in the medial striatum, whereas it was restricted to the dorsolateral striatum in the least impaired animals.(ABSTRACT TRUNCATED AT 250 WORDS)

Behavioural recovery of rats grafted with dopamine cells after partial striatal dopaminergic depletion in a conditioned reaction-time task.

The functional effects of grafts of dopamine-rich ventral mesencephalic suspension transplanted in a partially dopamine-depleted striatum were studied in rats performing a reaction-time motor task. The animals were trained to depress a lever, hold it down and release it within a limited period of time (700 ms) after the onset of a visual conditioned stimulus to obtain a food reward. The animals' performances were tested daily for up to two months after transplantation and for up to three months in the case of the animals with lesion only (bilateral striatal 6-hydroxydopamine injection). The baseline performances of the sham-operated control animals tended to improve, whereas the performances of the lesioned rats were significantly disrupted throughout the three months test. The majority of the animals (13/21) in the lesion group showed severe deficits mainly reflected in an increase in the number of the anticipated responses (premature release of the lever before the visual stimulus), and also in the number of the delayed responses (lever release after the time limit) recorded after dopamine depletion. The remaining animals (8/21) exhibited mild deficits (delayed responses only). These differences in the performance deficits appeared to be in relation to the extent of the dopamine denervation within the striatum assessed by the tyrosine hydroxylase immunostaining. Grafted animals showed a large number of dopamine fibers in the reinnervated striata and most of them (73%) significantly improved the reaction-time performance after transplantation. In the most severely impaired animals the number of anticipated errors was totally reversed within one month post-grafting, while the number of delayed responses remained high after transplantation. The performances of the less severely impaired animals returned more rapidly (within three weeks) to the pre-operative levels. The results show that intrastriatal ventral mesencephalic transplants are able to induce substantial or complete recovery in a complex reaction-time task. In the present model for partial dopamine depletion of the striatum, the mechanisms underlying the graft-induced recovery probably involve the participation of endogenous dopamine neurons acting in addition to, and/or in synergy with the dopamine-rich grafted tissue so that a functional level of dopaminergic transmission is restored in transplanted animals.

Ultrastructural analysis of graft-to-host connections, with special reference to dopamine-neuropeptide Y interactions in the rat striatum, after transplantation of fetal mesencephalon cells.

In a previous study we demonstrated that grafted dopamine (DA) neurons are able to induce an early and widespread normalization of DA-neuropeptide Y (NPY) interactions in the host striatum previously deprived of its DA input. Since similar recoveries were found to occur in striatal areas densely or poorly reinnervated by the graft, the question was raised as to what mechanisms (synaptic or volumic release) were involved in these functional effects. Ultrastructural analysis of graft-to-host relationships was performed using single--and double--immunolabelling techniques to detect neurons containing tyrosine hydroxylase (TH) and NPY, with a view to analysing the early establishment of synaptic connectivity in various areas of the host striatum. Within 1 month of the grafting, TH-immunoreactive (TH-IR) neurons showed most of the normal intrinsic morphological features characteristic of adult rat neurons and were found to have established direct relationships with various striatal neuronal populations. TH-NPY relationships were observed only in the area most densely reinnervated by the graft, and their relative frequency was found to be roughly the same as that determined in the intact striatum. Three months after the grafting, this percentage decreased, probably owing to the further elongation in TH-IR axons resulting in a wider distribution of the TH-NPY associations over the host striatum. In the zones distal from the graft, the reinnervation was far from complete and the few TH-IR fibres projected only to some unlabelled elements, mainly of the spiny type, which have been shown to interact normally with both DA afferents and NPY cells and therefore may relay the DA action over the whole striatum on the NPY population. It can be concluded from these data that the rapid and extensive functional normalization of the TH-NPY interactions previously found to occur in the entire striatum may depend on the restoration of direct and indirect synaptic relationships. A diffuse action of DA through non-synaptic mechanisms may also account for the fact that the amine has access to broader striatal populations than to those presumably reached by DA fibres arising from the graft.

Regulation of dopamine levels in intrastriatal grafts of fetal mesencephalic cell suspension: an in vivo voltammetric approach.

An in vivo voltammetric technique was used to monitor dopamine (DA) release in the 6-hydroxydopamine (6-OHDA)-lesioned rat striatum reinnervated by grafts of ventral mesencephalon containing DA neurons. Extracellular levels of DA were measured during the administration of D1 or D2 DA receptor antagonists. In addition, changes in DA levels induced by agonists and antagonists of excitatory amino acid (EAA) receptors were studied to verify the possible existence of a host glutamatergic control on the grafted DA cells in the 'transplanted' rats. Two months after the grafts were performed, the voltammetric signal measured under baseline conditions in the grafted striata was found to be almost similar to that recorded on the contralateral control side. Likewise, in another group of transplanted rats, the turnover of the amine, as expressed by the DO-PAC/DA tissue level ratio, was found to have become "normalized" after grafting, compared with the lesion-only group. The increase in the voltammetric signal observed after administering the D2 antagonist sulpiride (100 mg/kg i.p.) was lower in the grafted striata than on the contralateral side, however. This suggests that some D2 autoreceptor subsensitivity may have helped to maintain the baseline level of dopaminergic transmission. Adaptive processes of this kind might compensate for the partial DA reinnervation of the host striatum found to occur on the basis of the tyrosine hydroxylase immunostaining patterns. After administration of either the D1 antagonist SCH 23390 (0.1 mg/kg s.c.), or injection of EAA receptor agonists--1-glutamate, quisqualate and N-methyl-D-aspartate (all 10 nmol i.c.v.)--and antagonists--amino-phosphono-valeric acid (10 nmol i.c.v.) and dizocilpine (MK801, 0.2 mg/kg i.p.)--no significant differences between the two striata were detected in the voltammetric signals. These results suggest that, in the grafted rats, neurons belonging to the host population, such as the striatal cells bearing D1 receptors or the corticostriatal afferents presumed to contain glutamate, might modulate the DA levels, as was found to occur under normal conditions.

Early and widespread normalization of dopamine-neuropeptide Y interactions in the rat striatum after transplantation of fetal mesencephalon cells.

Graft-to-host interactions were examined at cellular level, by measuring changes in the immunoreactivity of striatal interneurons expressing neuropeptide Y after dopamine denervation and transplantation of fetal mesencephalon neurons into the striatum of adult rats. Mesencephalic cell suspensions were implanted unilaterally into the dorsal part of the striatum in rats two weeks after intranigral injection of 6-hydroxydopamine. One month and three to four months later, rats showing abolition of amphetamine-induced turning were perfused. Serial brain sections containing intrastriatal grafts were treated for tyrosine hydroxylase and neuropeptide Y immunocytochemistry, and neuropeptide Y-immunoreactive neurons were quantified in various parts of the striatal surface and compared with the striatum of controls and age-matched rats with lesions. Biochemical analyses of dopamine and dihydroxyphenyl acetic acid tissue levels and [3H]dopamine uptake were also performed on striatal samples from similar groups of normal, lesioned and transplanted rats. As early as one month post-grafting, a complete reversal of the increase in the number of neuropeptide Y-immunoreactive neurons occurring after 6-hydroxydopamine lesion was observed in dopamine-grafted animals, although a partial restoration of the tyrosine hydroxylase immunostaining and a recovery of 8% dopamine tissue level were observed in the striata of grafted as compared to normal rats. This effect on the host immunoreactivity was found to be specific to dopamine grafts, since no reversal was observed in sham-spinal cord-transplanted rats. Moreover, similar degrees of normalization were recorded either in the total striatum, or in the area immediately adjacent to the graft, or even in the zone most sensitive to dopamine denervation in terms of neuropeptide Y immunoreactivity. No more pronounced functional effects were observed three to four months after transplantation. These data suggest that grafted dopamine neurons are able to induce rapid and extensive host responsiveness, possibly by means of mechanisms involving synaptic and diffuse release of dopamine and adaptive changes in the host brain. These data may provide a cellular basis for interpreting larger behavioural recoveries than those expected to occur with dopamine grafts in view of the partial restoration of the dopaminergic innervation.