Calcium influx and mitochondrial alterations at synapses exposed to snake neurotoxins or their phospholipid hydrolysis products

Snake presynaptic phospholipase A2 neurotoxins (SPANs) bind to the presynaptic membrane and hydrolyze phosphatidylcholine with generation of lysophosphatidylcholine (LysoPC) and fatty acid (FA). The LysoPC+FA mixture promotes membrane fusion, inducing the exocytosis of the ready-to-release synaptic vesicles. However, also the reserve pool of synaptic vesicles disappears from nerve terminals intoxicated with SPAN or LysoPC+FA. Here, we show that LysoPC+FA and SPANs cause a large influx of extracellular calcium into swollen nerve terminals, which accounts for the extensive synaptic vesicle release. This is paralleled by the change of morphology and the collapse of membrane potential of mitochondria within nerve bulges. These results complete the picture of events occurring at nerve terminals intoxicated by SPANs and define the LysoPC+FA lipid mixture as a novel and effective agonist of synaptic vesicle release.     

Serial synapses in Aplysia

Serial synapses occur between small profiles in the neuropil of Aplysia abdominal ganglion. Material was fixed in phosphate buffered OsO₄, embedded in epon, and sections were stained with uranyl acetate and lead citrate. A class of synapses had the following characteristics: (1) synaptic vesicles clustered against the presynaptic membrane, (2) a widened extracellular space of about 20 nm containing electron-dense material, (3) straightening of the pre- and postsynaptic membranes, and (4) no postsynaptic membrane specialization. Some density between the presynaptic membrane and the adjacent synaptic vesicles was occasionally observed. Synapses occurred between small profiles in the neuropil (typical profile diameters were 1–3 m?m). In this sample of approximately 100 synapses, four serial synapses were identified. The serial synaptic profiles were all small. In addition to the finding of serial synapses, 40% of the postsynaptic profiles contained vesicles similar to the synaptic vesicles seen in presynaptic profile. Serial synapses may be the anatomical substrate of presynaptic inhibition and facilitation and of dishabituation.     

Effective Mechanism for Synthesis of Neurotransmitter Glutamate and its Loading into Synaptic Vesicles

Glutamate accumulation into synaptic vesicles is a pivotal step in glutamate transmission. This process is achieved by a vesicular glutamate transporter (VGLUT) coupled to v-type proton ATPase. Normal synaptic transmission, in particular during intensive neuronal firing, would demand rapid transmitter re-filling of emptied synaptic vesicles. We have previously shown that isolated synaptic vesicles are capable of synthesizing glutamate from α-ketoglutarate (not from glutamine) by vesicle-bound aspartate aminotransferase for immediate uptake, in addition to ATP required for uptake by vesicle-bound glycolytic enzymes. This suggests that local synthesis of these substances, essential for glutamate transmission, could occur at the synaptic vesicle. Here we provide evidence that synaptosomes (pinched-off nerve terminals) also accumulate α-ketoglutarate-derived glutamate into synaptic vesicles within, at the expense of ATP generated through glycolysis. Glutamine-derived glutamate is also accumulated into synaptic vesicles in synaptosomes. The underlying mechanism is discussed. It is suggested that local synthesis of both glutamate and ATP at the presynaptic synaptic vesicle would represent an efficient mechanism for swift glutamate loading into synaptic vesicles, supporting maintenance of normal synaptic transmission.     

Light and electron microscopic immunocytochemical localization of clathrin in rat cerebellum and kidney

The precise cellular and subcellular locations of coated vesicle protein, clathrin, in rat kidney and cerebellum have been visualized by immunocytochemical techniques. In the renal tubular epithelia, clathrin-positive products were found on both free ribosomes and on those attached to rough endoplasmic reticulum (RER) and the nuclear envelope. No clathrin was observed in the cisternae of RER or the Golgi apparatus. Clathrin-positive reaction products could also be seen on coated pits, coated vesicles, Golgi-associated vesicles, basolateral cell membrane, the ground substance, and in the autophagic vacuoles. In cerebellar Purkinje and granule cell bodies, reaction products were seen localized on coated vesicles, on the budding areas from the Golgi-associated membrane and Golgi-associated vesicles. Furthermore, the membrane of the multivesicular body, the bound-ribosomes, and the ground substance were also stained. In the myelinated axon, the clathrin appeared to be concentrated on certain segments and seemed to fill in the space between neurotubules and some vesicles. In certain synaptic terminals clathrin was often seen attached to presynaptic vesicles, presynaptic membrane, and post-synaptic membrane. However, in most mossy fibers, some synaptic vesicles were not stained. These observations suggest that clathrin is synthesized on bound and free ribosomes and discharged into the cytosol where it becomes associated with a variety of ground substances and assembles on coated pits, coated vesicles, Golgi-associated vesicles, presynaptic vesicles, and pre- and postsynaptic membranes. Clathrin may be finally degraded in autophagic vacuoles.     

An ultrastructural study of the microfilaments in rat brain by means of E-PTA staining and heavy meromyosin labeling

To identify structures involved in the translocation of the synaptic vesicles towards the presynaptic membrane, an ultrastructural study has been undertaken by means of (1) the E-PTA stain and (2) the HMM-labeling procedure. Using serial sections of E-PTA stained nervous tissue, especially those made in transversal and tangential planes, the geometric order of the presynaptic grid and of its constituents has been described in detail. It consisted of dense projections having the shape of small truncated pyramids cut parallel to their hexagonal bases which rested on the electron-lucent presynaptic membrane. The dense projections were arranged at the points of equilateral triangles. Around each dense projection, six asymmetric hexagonal holes were seen to be arrayed in an hexagonal pattern, forming thus the presynaptic sieve. From the spiny tops of the dense projections, which appeared as specialized structures of the dense material coating the inner surface of the plasma membrane at the level of the synaptic cleft, fine filaments, 40--60 A in diameter, radiated and formed a three-dimensional meshwork pervading the presynaptic bag. The dense cytoplasmic coating delineating the plasma membrane served as anchor points for these microfilaments. Upon incubation with rabbit skeletal muscle HMM the microfilaments underwent specific structural changes, consisting of: (1) a striking increase in diameter; (2) the association of periodic and polarized substructures with their surfaces. The synaptic vesicles and mitochondria were seen to be attached to the numerous HMM-decorated filaments or enmeshed in the network formed by these filaments. The actin-like filaments were anchored to the plasma membrane at many points and to the presynaptic dense projections. Following incubation in the buffer alone or in buffer HMM solutions containing Na+ pyrophosphate or ATP, no arrowheaded structures were observed. Thus, a network consisting of actin-like filaments was demonstrated in the presynaptic bag. Of particular interest was the structural relation of the actin-like filaments with the occasional, tapered myosin-like filaments. The role of the presynaptic actin-like network in the transport of synaptic vesicles towards the presynaptic membrane by a mechanism of chemomechanical transduction is discussed. In the postsynaptic dendrite or dendritic spine, a filamentous network was observed to be attached to the subsynaptic web by means of the E-PTA stain and of the HMM-labeling procedure. The occurrence of an actin-like meshwork in the postsynaptic region is suggested to produce changes in the macromolecular configuration of the postsynaptic membrane by a "mechanoenzyme" system similar to that described in the mitochondrial membrane.     

Intracellular movements of fluorescently labeled synaptic vesicles in frog motor nerve terminals during nerve stimulation.

We stained synaptic vesicles in frog motor nerve terminals with FM1-43 and studied changes in the shape and position of vesicle clusters during nerve stimulation. Each stained vesicle cluster appeared as a fluorescent spot. During repetitive nerve stimulation the spots gradually dimmed, most without changing shape or position. Occasionally, however, a spot moved, appearing in some cases to stream toward and coalesce with a neighboring spot. This suggests the existence of translocation mechanisms that can actively move vesicles in a coordinated fashion between vesicle clusters. Within single clusters, we saw no signs of such directed vesicle movements. Fluorescent spots in terminals viewed from the side with a confocal microscope did not shrink toward the presynaptic membrane during nerve stimulation, but dimmed uniformly. This suggests that vesicles continuously mix within a cluster during destaining and provides no evidence of active vesicle translocators within single vesicle clusters for moving vesicles to the presynaptic membrane.     

Isolation of a presynaptic plasma membrane fraction from Torpedo cholinergic synaptosomes: evidence for a specific protein

Synaptosomal plasma membranes were isolated from Torpedo cholinergic synaptosomes which had been purified as previously described or repurified by equilibrium centrifugation. The synaptosomal plasma membrane could be distinguished from postsynaptic membranes by the absence of postsynaptic specific markers (nicotinic AChR) and by its low intramembrane particle complement after freeze fracture. In addition, the presynaptic membrane fraction contained acetylcholinesterase. Gel electrophoresis permitted the identification of a major protein component of the presynaptic membrane fraction which had a molecular weight of 67,000. This protein was not found in postsynaptic membrane or synaptic vesicle fractions. Thus it appeared to be specific to the nerve terminal plasma membrane.     

Polyhedral Protein Cages Encase Synaptic Vesicles and Participate in Their Attachment to the Active Zone

In an effort to elucidate the interactions between synaptic vesicles and the membrane of the active zone, we have investigated the structure of interneuronal asymmetric synapses in the neocortex of adult rats using thin-sectioning, freeze-fracture, and negative staining electron microscopy. We identified three subtypes of spherical synaptic vesicles. Type I were agranular vesicles of 47.5 ± 3.8 nm (mean SD,n = 24) in diameter usually seen aggregated in clusters in the presynaptic bouton. Type II synaptic vesicles were composed of a ∼45-nm-diameter lipid bilayer sphere encased in a cage 77 ± 4.6 nm (mean SD,n = 42) in diameter. The cage was composed of open-faced pentamers 20–22 nm/side arranged as a regular polyhedron. Type II caged vesicles were found in clusters at the boutons, adhered to the active zone, and were also present in axons. Type III synaptic vesicles appeared as electron-dense spheres 60–75 nm in diameter abutted to the membrane of the active zone. Clathrin-coated vesicles and pits of 116.6 ± 9 nm (mean SD,n = 14) in diameter were also present in both the pre- and postsynaptic sides. Freeze-fracture showed that some intrinsic membrane proteins in the active zone were arranged as pentamers exhibiting the same dimension of those forming cages (∼22 nm/side). From these data, we concluded that: (a) the presynaptic bouton contains a heterogeneous population of “caged” and “plain” synaptic vesicles and (b) type II synaptic vesicles bind to receptors in the active zone. Therefore, current models of transmitter release should take into account the substantial heterogeneity of the vesicle population and the binding of vesicular cages to the membrane of the active zone.     

Recycled Synaptic Vesicles Contain Vesicle but Not Plasma Membrane Marker, Newly Synthesized Acetylcholine, and a Sample of Extracellular Medium

To monitor the fate of the synaptic vesicle membrane compartment, synaptic vesicles were isolated under varying experimental conditions from blocks of perfused Torpedo electric organ. In accordance with previous results, after low-frequency stimulation (0.1 Hz, 1,800 pulses) of perfused blocks of electric organ, a population of vesicles (VP2 type) can be separated by density gradient centrifugation and chromatography on porous glass beads that is denser and smaller than resting vesicles (VP1 type). By simultaneous application of fluorescein isothiocyanate-dextran as extracellular volume marker and [3H]acetate as precursor of vesicular acetylcholine, and by identifying the vesicular membrane compartment with an antibody against the synaptic vesicle transmembrane glycoprotein SV2, we can show that the membrane compartment of part of the synaptic vesicles becomes recycled during the stimulation period. It then contains both newly synthesized acetylcholine and a sample of extracellular medium. Recycled vesicles have not incorporated the presynaptic plasma membrane marker acetylcholinesterase. Cisternae or vacuoles are presumably not involved in vesicle recycling. After a subsequent period of recovery (18 h), all vesicular membrane compartments behave like VP1 vesicles on subcellular fractionation and still retain both volume markers. Our results imply that on low-frequency stimulation, synaptic vesicles are directly recycled, equilibrating their luminal contents with the extracellular medium and retaining their membrane identity and capability to accumulate acetylcholine.     

Ganglioside composition of subcellular fractions,including pre- and postsynaptic membranes,fromTorpedo electric organ

Gangliosides were isolated from four subcellular fractions of the electric organ ofTorpedo marmorata: synaptosomes, presynaptic membranes, postsynaptic membranes, and synaptic vesicle membranes. This exploited a principal advantage offered by this tissue: facile separation of pre-and postyynaptic elements. Total ganglioside concentration in presynaptic membranes was approximately twice that of synaptosomes and 15 times that of postsynaptic membranes (47.7, 24.4, and 3.21 g of lipid sialic acid per mg protein, respectively). Synaptic vesicle membranes had the highest overall concentration (78.9) relative to protein, but a concentration approximately comparable to that of presynaptic membranes when expressed relative to phospholipid. The thin-layer patterns of these two fractions were similar, both in terms of total pattern and the specific pattern of gangliotetraose structures as revealed by overlay with cholera toxin B subunit; these were notable for the paucity of monosialo structures and the virtual absence of GM1. Postsynaptic membranes, on the other hand, had a significantly higher content of monosialogangliosides including the presence of GM1. The synaptosomal pattern resembled that of the presynaptic membranes and synaptic vesicles. Thus, a clear difference in ganglioside pattern could be discerned between the pre- and postsynaptic elements of the electric organ.Abbreviations SVs synaptic vesicles - TLC thin-layer chromatography - cholera B-HRP B subunit of cholera toxin linked to horseradish peroxidase     

The uptake of horseradish peroxidase by cortical synapses in rat brain

Summary Horseradish peroxidase (HRP) was introduced directly into the cerebral cortex of adult rats, which were allowed to survive for 60 min before perfusion fixation. After the tissue had been incubated to demonstrate HRP at the LM and EM levels, blocks of cortical tissue were taken at varying distances from the injection site. These eight blocks of tissue constituted a time sequence for HRP diffusion.Qualitative examination of the presynaptic terminals showed that the most commonly encountered profiles are the plain synaptic vesicles, many of which accumulate tracer. In some terminals labelled vesicles are lined-up in tubular fashion. Other profiles commonly labelled are coated vesicles, tubular and vacuolar cisternae, and plain and coated pinocytotic vesicles.Quantitative analyses based on the number of terminals containing labelled profiles demonstrate an early rise in the rate of labelling of both plain synaptic vesicles and coated vesicles, after which synaptic vesicle labelling rises slowly towards a plateau. By contrast, there is a late parallel increase in the rate of labelling of coated vesicles and cisternae. A more detailed analysis, based on the actual numbers of labelled and total profiles within each presynaptic terminal, highlight early and late periods of rapid labelling for plain synaptic vesicles, coated vesicles and cisternae. A further aspect of HRP incorporation studied, concerns its uptake into four delineated regions of the presynaptic terminal.Our data indicate that membrane uptake into the presynaptic terminal is accomplished mainly via coated vesicles, although plain synaptic vesicles may also be involved. Coated vesicles, in turn, appear to give rise directly to plain synaptic vesicles, with some coalescing to produce vacuolar cisternae. The latter are involved in a two-way interchange of membrane with tubular cisternae, plain synaptic vesicles and coated vesicles. An additional source of plain synaptic vesicles are the tubular cisternae. Exocytosis of plain synaptic vesicles constitutes the mechanism by which transmitter is released from the presynaptic terminal.Supported by the Nuffield Foundation. We are grateful to Mr. M. Austin for help with the photography     

The role of coated vesicles in recycling of synaptic vesicle membrane

The uptake of extracellular tracers into synaptic nerve terminals has been a phenomenon of persistent interest. Uptake is into synaptic vesicles, hence vesicles spend part of their life in continuity with the plasma membrane, as expected if exocytosis underlies the quantal discharge of neurotransmitters. However, exactly how or when synaptic vesicles acquire extracellular tracers has not been unambiguously determined. Two schools of thought have developed, one holding that vesicles acquire tracers directly via a reversible exo/endocytotic sequence in which they consistently maintain their biochemical identity during their transient continuity with the plasma membrane, the other holding that synaptic vesicles acquire tracers indirectly, via the formation of clathrin-coated vesicles which are spatially and temporally separate from exocytosis and reverse a temporary loss of the vesicles' individual identity upon merger with the plasma membrane. Efforts to distinguish between these two alternatives have generated an interesting diversity of electron microscopic experiments, many of which are reviewed here. However, definitive determination of which view is correct may ultimately require direct visualization of synaptic vesicle turnover in living nerve terminals. To this end, we here review the results of visualizing endocytosis in tissue cultured cells, where light microscopy can provide sufficient resolution to reveal membrane dynamics in living cells. This has allowed visual discrimination of two different types of endocytosis, one clathrin-mediated (coated vesicle formation) and the other actin-mediated (macropinocytosis). Current work is also reviewed which aims at determining experimental methods for inhibiting each type of endocytosis selectively. Hypertonicity and severe cytoplasmic acidification turn out to inhibit coated vesicle formation, while cytochalasin D and mild cytoplasmic acidification selectively inhibit macropinocytosis. Applied to nerves, these various treatments affect synaptic vesicle turnover in a manner that supports the notion that synaptic vesicle membrane recycles via the "indirect" route of coated vesicle formation.     

Axons containing neuropeptide Y innervate arginine vasopressin-containing neurons in the rat paraventricular nucleus. Dual electron microscopic immunolabeling

Synaptic connections between neurons immunoreactive for arginine vasopressin (AVP) and axon terminals immunoreactive for neuropeptide Y (NPY) were found in the magnocellular part of the paraventricular nucleus (PVN) in the rat hypothalamus. In pre-embedding double immunolabeling, NPY axon terminals labeled with diaminobenzidine (DAB) reaction product established synaptic junctions on the perikarya and neuronal processes of AVP neurons labeled with silver-gold particles. Ultrastructural morphology of the neurons was more suitably preserved by a combination of pre- and post-embedding procedures. The presynaptic NPY terminals contained many small clear vesicles and a few cored vesicles, and DAB chromogen (immunoreaction product) was located on the surface of the vesicular profiles and on the core. The postsynaptic AVP neurons possessed many large secretory granules labeled with gold particles. At the synaptic junctions, small clear vesicles were accumulated at the presynaptic membrane, and the postsynaptic membrane was coated with a dense accumulation of fine electron dense particles. The perikarya also received synapses made by immuno-negative axon terminals containing many small clear vesicles and a few cored vesicles. These terminals were found more frequently than those containing NPY.     

Rab2 regulates presynaptic precursor vesicle biogenesis at the trans-Golgi

Reliable delivery of presynaptic material, including active zone and synaptic vesicle proteins from neuronal somata to synaptic terminals, is prerequisite for successful synaptogenesis and neurotransmission. However, molecular mechanisms controlling the somatic assembly of presynaptic precursors remain insufficiently understood. We show here that in mutants of the small GTPase Rab2, both active zone and synaptic vesicle proteins accumulated in the neuronal cell body at the trans-Golgi and were, consequently, depleted at synaptic terminals, provoking neurotransmission deficits. Ectopic presynaptic material accumulations consisted of heterogeneous vesicles and short tubules of 40 × 60 nm, segregating in subfractions either positive for active zone or synaptic vesicle proteins and LAMP1, a lysosomal membrane protein. Genetically, Rab2 acts upstream of Arl8, a lysosomal adaptor controlling axonal export of precursors. Collectively, we identified a Golgi-associated assembly sequence of presynaptic precursor biogenesis dependent on a Rab2-regulated protein export and sorting step at the trans-Golgi.     

Diurnal changes in the fine structure of photoreceptors in an abalone,Nordotis discus

Summary The ultrastructure of the synapses in the brain of the monogenean Gastrocotyle trachuri (Platyhelminthes) is described. The synapses consist of one presynaptic terminal separated by a uniformly wide synaptic cleft, from one or more postsynaptic elements. The presynaptic terminals are characterized by the presence of paramembranous dense projections and associated synaptic vesicles. The postsynaptic elements while possessing membrane densities, are usually devoid of vesicles.The structure of the synapses in the brain of Gastrocotyle is compared to synapses from other platyhelminths.     

Axons containing neuropeptide Y innervate arginine vasopressin-containing neurons in the rat paraventricular nucleus

Summary Synaptic connections between neurons immunoreactive for arginine vasopressin (AVP) and axon terminals immunoreactive for neuropeptide Y (NPY) were found in the magnocellular part of the paraventricular nucleus (PVN) in the rat hypothalamus. In pre-embedding double immunolabeling, NPY axon terminals labeled with diamin-obenzidine (DAB) reaction product established synaptic junctions on the perikarya and neuronal processes of AVP neurons labeled with silver-gold particles. Ultrastructural morphology of the neurons was more suitably preserved by a combination of pre- and post-embedding procedures. The presynaptic NPY terminals contained many small clear vesicles and a few cored vesicles, and DAB chromogen (immunoreaction product) was located on the surface of the vesicular profiles and on the core. The postsynaptic AVP neurons possessed many large secretory granules labeled with gold particles. At the synaptic junctions, small clear vesicles were accumulated at the presynaptic membrane, and the postsynaptic membrane was coated with a dense accumulation of fine electron dense particles. The perikarya also received synapses made by immuno-negative axon terminals containing many small clear vesicles and a few cored vesicles. These terminals were found more frequently than those containing NPY.     

Quantitative characteristics of ultrastructural changes in axo-dendritic synapses under the influence of tetanus toxin

Changes in synaptic ultrastructure of the external geniculate body (EGB) were investigated in rats when a generator of pathologically intensified excitation (GPIE) was produced in this nucleus under the influence of tetanus toxin (TT). At the period of pronounced convulsive activity (24 h after TT injection), synaptic changes were estimated electronmicroscopically and with quantitative comparison of the materials from three groups. The first group included EGB synapses where TT was injected, the second group included contralateral EGB synapses and the third included EGB from the rats injected with inactivated toxin. By means of electron optic computer "Klassimat" average amount of round, flat, anomalous and adjacent to the presynaptic membrane vesicles was measured, average relative length of the active zone, average area of the presynaptic terminal, average relative section areas of pre- and postsynaptic cytoplasm condensation were estimated. In the area of GPIE formation, under the influence of TT, the increased amount of the vesicles related to the presynaptic membrane and that of flat vesicles were statistically significant. At the same time, the synaptic terminals, by the number of vesicles, have bimodal, while the control groups have unimodal distribution.     

Synaptophysin (p38) at the frog neuromuscular junction: its incorporation into the axolemma and recycling after intense quantal secretion

Recycling of synaptophysin (p38), a synaptic vesicle integral membrane protein, was studied by the use of antisera raised against the protein purified from frog brain. When frog cutaneous pectoris muscles were fixed at rest, a bright, specific immunofluorescent signal was observed in nerve-terminal regions only if their plasma membranes had been previously permeabilized. When muscles were fixed after they had been treated for 1 h with a low dose of alpha-latrotoxin in Ca2+-free medium, an equally intense fluorescence could be observed without previous permeabilization. Under this condition, alpha-latrotoxin depletes nerve terminals of their quantal store of acetylcholine and of synaptic vesicles. These results indicate that fusion of synaptic vesicles leads to the exposure of intravesicular antigenic determinants of synaptophysin on the outer surface of the axolemma, and provide direct support for the vesicle hypothesis of neurotransmitter release. After 1 h treatment with the same dose of alpha-latrotoxin in the presence of 1.8 mM extracellular Ca2+, immunofluorescent images were obtained only after permeabilization with detergents. Under this condition, the vesicle population was maintained by an active process of recycling and more than two times the initial store of quanta were secreted. Thus, despite the active turnover of synaptic vesicles and of quanta of neurotransmitter, no extensive intermixing occurs between components of the vesicle and presynaptic plasma membrane.     

An ultrastructural study of the development of afferent and efferent synapses on outer hair cells of the guinea pig organ of Corti

The guinea pig organ of Corti was studied using transmission electron microscopy, the second turn of the cochlea being examined at various ages between 20 days before birth and 30 days postnatal. Outer hair cells were examined at each of these ages. At all ages studied, the efferent (presynaptic) terminals are large and are packed with synaptic vesicles, whereas the afferent (postsynaptic) terminals are generally smaller, with a relatively small number of vesicles. During development, the subsynaptic cistern changes from a fragmented, diffuse profile extending over 50-70% of the length of the efferent contact zones, to a continuous, compact structure spanning neighbouring synapses. Synaptic vesicles in the efferent terminals are predominantly rounded in early development, flattened vesicles appearing postnatally. The synaptic bodies at afferent synapses do not change noticeably during development. Quantitative analysis revealed that the area of efferent terminals and the length of their active zone increase with increasing age, the same parameters decreasing in afferent terminals. Synaptic vesicles in the efferent terminals decrease in diameter, but remain constant in afferent terminals, with increasing age. The number of hair cell membrane invaginations decreases as development proceeds.     

Circulation and turnover of synaptic vesicle membrane in cultured fetal mammalian spinal cord neurons

Intact neurons in cultures of fetal rodent spinal cord explants show stimulation-dependent uptake of horseradish peroxidase (HRP) into many small vesicles and occasional tubules and multivesicular bodies (MVB) at presynaptic terminals. Presynaptic terminals were allowed to take up HRP during 1 h of strychnine-enhanced stimulation of synaptic transmitter release and then "chased" in tracer-free medium either with strychnine or with 10 mM Mg++ which depresses transmitter release. Tracer-containing vesicles are lost from terminals under both chase conditions; the loss is more rapid (4-8 h) with strychnine than with 10 mM Mg++ (8-16 h). There is a parallel decrease in the numbers of labeled MVB's at terminals. Loss of tracer with 10 mM Mg++ does not appear to be due to the membrane rearrangements (exocytosis coupled to endocytosis) that presumably lead to initial tracer uptake; terminals exposed to HRP and Mg++ for up to 16 h show little tracer uptake into vesicles. Nor is the decrease likely to the due to loss of HRP enzyme activity; HRP is very stable in solution. During the chases there is a striking accumulation of HRP in perikarya that is far more extensive in cultures initially exposed to tracer with strychnine than 10 mM Mg++ regardless of chase conditions. Much of the tracer ends up in large dense bodies. These findings suggest that synaptic vesicle membrane turnover involves retrograde axonal transport of membrane to neuronal perikarya for further processing, including lysosomal degradation. The more rapid (4-8 h) loss of tracer-containing vesicles with strychnine may reflect vesicle membrane reutilization for exocytosis.