Search

Certain junctions between ependymal cells, between astrocytes, and between some electrically coupled neurons have heretofore been regarded as tight, pentalaminar occlusions of the intercellular cleft. These junctions are now redefined in terms of their configuration after treatment of brain tissue in uranyl acetate before dehydration. Instead of a median dense lamina, they are bisected by a median gap 20–30 A wide which is continuous with the rest of the interspace. The patency of these "gap junctions" is further demonstrated by the penetration of horseradish peroxidase or lanthanum into the median gap, the latter tracer delineating there a polygonal substructure. However, either tracer can circumvent gap junctions because they are plaque-shaped rather than complete, circumferential belts. Tight junctions, which retain a pentalaminar appearance after uranyl acetate block treatment, are restricted primarily to the endothelium of parenchymal capillaries and the epithelium of the choroid plexus. They form rows of extensive, overlapping occlusions of the interspace and are neither circumvented nor penetrated by peroxidase and lanthanum. These junctions are morphologically distinguishable from the "labile" pentalaminar appositions which appear or disappear according to the preparative method and which do not interfere with the intercellular movement of tracers. Therefore, the interspaces of the brain are generally patent, allowing intercellular movement of colloidal materials. Endothelial and epithelial tight junctions occlude the interspaces between blood and parenchyma or cerebral ventricles, thereby constituting a structural basis for the blood-brain and blood-cerebrospinal fluid barriers.

Summary Background Trichoplax adhaerens is the best-known member of the phylum Placozoa, one of the earliest-diverging metazoan phyla. It is a small disk-shaped animal that glides on surfaces in warm oceans to feed on algae. Prior anatomical studies of Trichoplax revealed that it has a simple three-layered organization with four somatic cell types. Results We reinvestigate the cellular organization of Trichoplax using advanced freezing and microscopy techniques to identify localize and count cells. Six somatic cell types are deployed in stereotyped positions. A thick ventral plate, comprising the majority of the cells, includes ciliated epithelial cells, newly identified lipophil cells packed with large lipid granules, and gland cells. Lipophils project deep into the interior, where they alternate with regularly spaced fiber cells whose branches contact all other cell types, including cells of the dorsal and ventral epithelium. Crystal cells, each containing a birefringent crystal, are arrayed around the rim. Gland cells express several proteins typical of neurosecretory cells, and a subset of them, around the rim, also expresses an FMRFamide-like neuropeptide. Conclusions Structural analysis of Trichoplax with significantly improved techniques provides an advance in understanding its cell types and their distributions. We find two previously undetected cell types, lipohil and crystal cells, and an organized body plan in which different cell types are arranged in distinct patterns. The composition of gland cells suggests that they are neurosecretory cells and could control locomotor and feeding behavior.

The reciprocal dendro-dendritic synapse between granule and mitral or tufted dendrites in the external plexiform layer of the olfactory bulb consists of an excitatory mitral-to-granule synaptic contact and an adjacent inhibitory granule-to-mitral synaptic contact. The pre- and postsynaptic membranes of both synaptic contacts were identified in replicas of freeze-fractured external plexiform layer in rabbits, mice, and chinchillas. At the excitatory synaptic contact there is a prominent specialization in the postsynaptic memberane, represented by an aggregate of homogeneous particles associated with the external half of the membrane. In contrast, the postsynaptic membrane at the inhibitory granule-to-mitral synaptic contact lacks evident internal specializations, and the distribution of particles on both fracture faces resembles that at non-synaptic regions. Less marked differences in particle distribution characterized the cytoplasmic half of the presynaptic membranes. These differences probably reflect diversity in the nature or distribution of membrane proteins at excitatory and inhibitory synapses. Protuberances on the external half of the presynaptic membrane, possibly sites of vesicle interaction with the plasma membrane, surrounded but were not coextensive with both types of synaptic contact. A few gap junctions connected proximal dendrites of mitral or tufted cells with granule cell dendrites.

Tight junction strands occur at three atypical locations in slices of rat ventral prostate exposed to conditions promoting rapid tight junction assembly: (1) in the basal plasma membranes of the columnar epithelial cells, over 40 microns from the native apical tight junction band (misplaced tight junctions); (2) in the plasma membranes of basal epithelial cells, which never have tight junctions in the native state; and (3) between processes of the same cell (reflexive or autocellular tight junctions) at the basal or lateral portions of the columnar epithelial cell. These findings suggest that tight junction formation is not limited to specific parts of the plasma membrane, even in highly polarized cells such as those in prostate epithelium. Taken together with other new evidence, they also suggest that tight junctions may be very labile.

The ctenophoran integument consists basically of a single-layered epidermis which covers the entire body, including appendages such as the tentacular apparatus, the lobes and auricles, and Unes the stomodeal cavity (generally referred to as the gastric cavity). This integument is always devoid of any cuticle or hard secretion, but is permanently covered by a film of mucus. It rests on a gelatinous mesoglea, which is an unusual connective tissue devoid of collagen and elastin fibers (Franc et al., 1976), and harboring mesenchymal cells and numerous true muscle cells. The mesoglea may be considered as the internal milieu of the ctenophore; it is fed and oxygenated by a system of gastrovascular channels, which may be very elaborate in large species. Franc (1972) demonstrated that specialized ciliated cells, located in the canal walls and grouped in ciliated rosettes, regulated the relative ionic and osmotic composition of the mesoglea.

We have used a scanning electron microscope to study the morphology of the postsynaptic folds in frog muscle. These folds were exposed by digesting muscles in collagenase or osmium tetroxide. During digestion the connective tissue and basement lamina were removed and the presynaptic nerve terminal pulled away from the muscle surface to reveal one or more concave junctional grooves, within which the folds in the postsynaptic membrane could be clearly seen. In places the folding patterns were more complex than could be inferred from single thin-section electron micrographs. The ease with which the postsynaptic folds of vertebrate muscles can be directly observed using this technique now makes it possible to investigate subtle changes in their structure during synaptogenesis, denervation or disease.

Direct freezing procedures have enabled us to visualize distinctive intramembrane particle ring structures in the cytoplasmic membranes of peritrichously flagellated bacteria by means of freeze-fracture electron microscopy. These structures were identified as flagellar motor components because their distribution matched that of flagella, and because they were absent in non-flagellated mutants of Escherichia coli. Particle rings were present in both the Gram-positive Streptococcus and the Gram-negative E. coli. In E. coli, a non-functional mocha operon produced flagellated but immotile cells lacking the particle rings. Simultaneous introduction of the motA and motB genes, led to recovery of both motility and the ring structures but neither gene alone was sufficient. The concomitant loss of the rings and motility is consistent with the ring particles having a central role in the flagellar motor.

Nonjunctional and nonsynaptic membranes of hair cells in the chinchilla organ of Corti were examined using the freeze-fracture technique. Cy-toplasmic leaflets of the apical membranes of hair cells have particles, 6–12 nm in diameter, but many more particles are found on apical membranes of outer hair cells than on inner hair cells. Cytoplasmic leaflets of the lateral membranes of outer hair cells are covered with large particles, but the corresponding regions of the inner hair cell membrane have fewer particles and these are small or medium-sized. Two types of particle aggregate also distinguish this region of the inner hair cell. The first consists of patches of particles closely spaced in rec tilinear arrays. The second consists of parallel strands of widely spaced large particles similar in size, but not in distribution, to the large particles on the lateral membranes of the outer hair cells. The basal membrane of outer hair cells is distinguished from that of inner hair cells by plaques consisting of cross-hatched in-cisures in the external membrane leaflet. While the significance of these anatomical features is not yet apparent, they give the hair cell plasmalemma a richness of regional specialization found in few other cells and suggest that there are important functional differences between inner and outer hair cells.

Crayfish abdominal stretch receptor organs are innervated by inhibitory (GABA) and excitatory (glutamate) synapses. Previous studies with aldehyde fixation showed that synaptic vesicles in the inhibitory synapse are flat and small, whereas those in the excitatory synapse are rounder and larger. We have reexamined these inhibitory and excitatory synapses by using direct rapid-freezing and freeze-substitution in order to preserve synaptic structure closer to its living state. Fine details of synaptic structure appear to be better preserved by this method. Synaptic vesicles in inhibitory as well as excitatory synapses are round, so the conventional flattened shape of vesicles in the inhibitory synapse must depend on some aspect of aldehyde processing. However, the average size of vesicles in the inhibitory synapse is significantly smaller than that of vesicles in the excitatory synapse, so synaptic vesicle size is regarded as having functional significance.

Several synapses of known physiological action are antomically segregated in the cerebellar cortex and are readily identified in freeze-fracture preparations. Excitatory synapses, such as the parallel fiber-to-Purkinje spine synapse, climbing fiber-to-Purkinje spine synapse, and mossy or climbing fiber-to-granule cell dendrite synapse, were characterized by small aggregates of large particles on the cytoplasmic half of the presynaptic membrane, by a distinctly widened synaptic cleft, and by a large aggregate of particles on the external half of the postsynaptic membrane. Inhibitory synapses, such as the stellate cell axon-to-Purkinje dendrite synapse and the basket cell axon-to-Purkinje soma synapse, had no comparable specialization of either the pre- and postsynaptic membrane. The striking contrast in membrane structure at excitatory and inhibitory synaptic contacts presumably reflects differences in either the composition or organization of membrane proteins integral to synaptic function. Puncta adhaerentia between granule cell dendrites in cerebellar glomeruli were characterized by particles aggregated on the external half of both apposed membranes and were further differentiated from synaptic contacts by the smaller size of the particles. Protuberances on the external half of the presynaptic membrane were either small and coextensive with the synaptic contact or were larger and surrounded it; it is suggested that the small protuberances are synaptic vesicle sites whereas the large ones are coated vesicle sites.

A reconstituted system for examining directed organelle movements along purified microtubules has been developed. Axoplasm from the squid giant axon was separated into soluble supernatant and organelle-enriched fractions. Movement of axoplasmic organelles along MAP-free microtubules occurred consistently only after addition of axoplasmic supernatant and ATP. The velocity of such organelle movement (1.6 micron/sec) was the same as in dissociated axoplasm. The axoplasmic supernatant also supported movement of microtubules along a glass surface and movement of carboxylated latex beads along microtubules at 0.5 micron/sec. The direction of microtubule movement on glass was opposite to that of organelle and bead movement on microtubules. The factors supporting movements of microtubules, beads, and organelles were sensitive to heat, trypsin, AMP-PNP and 100 microM vanadate. All of these movements may be driven by a single, soluble ATPase that binds reversibly to organelles, beads, or glass and generates a translocating force on a microtubule.

Turtle optic nerves were rapid-frozen from the living state, fractured, etched, and rotary shadowed. Stereo views of fractured axons show that axoplasm consists of three types of longitudinally oriented domains. One type consists of neurofilament bundles in which individual filaments are interconnected by a cross-bridging network. Contiguous to neurofilament domains are domains containing microtubules suspended in a loose, granular matrix. A third domain is confined to a zone, 80-100 nm wide, next to the axonal membrane and consists of a dense filamentous network connecting the longitudinal elements of the axonal cytoskeleton to particles on the inner surface of the axolemma. Three classes of membrane-limited organelles are distinguished: axoplasmic reticulum, mitochondria, and discrete vesicular organelles. The vesicular organelles must include lysosomes, multivesicular bodies, and vesicles which are retrogradely transported in axons, though some vesicular organelles may be components of the axoplasmic reticulum. Organelles in each class have a characteristic relationship to the axonal cytoskeleton. The axoplasmic reticulum enters all three domains of axoplasm, but mitochondria and vesicular organelles are excluded from the neurofilament bundles, a distribution confirmed in thin sections of cryoembedded axons. Vesicular organelles differ from mitochondria in at least three ways with respect to their relationships to adjacent axoplasm: (a) one, or sometimes both, of their ends are associated with a gap in the surrounding granular axoplasm; (b) an appendage is typically associated with one of their ends; and (c) they are not attached or closely apposed to microtubules. Mitochondria, on the other hand, are only rarely associated with gaps in the axoplasm, do not have an appendage, and are virtually always attached to one or more microtubules by an irregular array of side-arms. We propose that the longitudinally oriented microtubule domains are channels within which organelles are transported. We also propose that the granular material in these channels may constitute the myriad enzymes and other nonfibrous components that slowly move down the axon.

The outer plexiform layer (OPL) of the developing chick retina from 11 embryonic days to 11/2 weeks posthatching has been studied by freeze-fracture to characterize changes in the membrane structure of photoreceptor terminals during synaptogenesis. At early stages, the undifferentiated photoreceptor synaptic base is characterized by a sparse distribution of intramembrane particles on the inner leaflet (P-face). Later, as the synaptic base begins to differentiate by extending filopodia into the OPL, numerous small aggregates of large particles appear between and on filopodial surfaces. Many of the aggregates occupy crater-like depressions, which are seen in cross-fractures through the underlying cytoplasm to be associated with vesicular invaginations of the presynaptic membrane. Corresponding thin sections through these regions at this time reveal immature arciform densities and coated vesicles fusing with the presynaptic membrane adjacent to these densities. At later stages, many of the particle aggregates on the photoreceptor membrane appear to have coalesced into longer arrays overlying ridges surrounded by numerous vesicle fusion sites. These intramembrane changes correlate with the formation of the mature arciform density-synaptic ribbon specialization in the photoreceptor presynaptic terminal and with physiological maturation of the chick retina.

Single filaments, dissociated from the extruded axoplasm of the squid giant axon and visualized by videoenhanced differential interference contrast microscopy, transport organelles bidirectionally. Organelles moving in the same or opposite directions along the same filament can pass each other without colliding, indicating that each transport filament has several tracks for organelle movement. In order to characterize transport filaments, organelle movements were first examined by video microscopy, and then the same filaments were examined by electron microscopy after rapid-freezing, freeze-drying, and rotaryshadowing. Transport filaments that supported bidirectional movement of organelles are 22 nm to 27 nm in diameter and have a substructure indicative of a single microtubule. Immunofluorescence showed that virtually all transport filaments contain tubulin. These results show that single microtubules can serve as a substratum for organelle movement, and suggest that an interaction between organelles and microtubules is the basis of fast axonal transport.

Cytoplasmic filaments, separated from the axoplasm of the squid giant axon and visualized by video-enhanced differential interference contrast microscopy, support the directed movement of organelles in the presence of ATP. All organelles, regardless of size, move continuously along isolated transport filaments at 2.2 +/- 0.2 micron/sec. In the intact axoplasm, however, movements of the larger organelles are slow and saltatory. These movements may reflect a resistance to movement imposed by the intact axoplasm. The uniform rate of all organelles along isolated transport filaments suggests that a single type of molecular motor powers fast axonal transport. Organelles can attach to and move along more than one filament at a time, suggesting that organelles have multiple binding sites for this motor.

The meninges of various mammals were prepared for examination with the electronmicroscope by thin sectioning or freeze-fracturing. Particular attention was given to the distribution of tight junctions in order to determine the basis for the meningeal barrier between the blood circulating in dural vessels and the cerebrospinal fluid in the subarachnoid space. While some dural blood vessels are fenestrated, those in the subarachnoid space are not and their component endothelial cells are joined by an extensive system of tight junctions. An extensive and continuous system of tight junctions was also found in a layer of specialized cells at the border of the arachnoid with the dura. This arachnoid barrier layer is apparently the only basis of the meningeal barrier because often cellular layers in the dura and arachnoid lack tight junctions although they are linked by gap junctions and desmosomes. In particular, tight junctions are lacking at the border of the "subdural space" which is actually a fascial plane within the dura. Tight junctions are also lacking between astrocytes at the surface of the brain but these cells are linked by gap junctions and a new type of intercellular junction. The distribution of these junctions, as well as assemblies of intramembranous particles at the astrocytic border, raises the question whether this layer might have a role in the exchange of certain substances between the brain and cerebrospinal fluid.

In tissue prepared by rapid freezing, freeze fracture, and shallow etching, the postsynaptic density of Purkinje cell dendritic spines has a substructure consisting of fine filaments and irregular, globular adherent proteins. The number and packing density of the globular proteins vary from region to region within a single density and are even more variable when different junctions are compared. Whereas actinlike microfilaments and spectrinlike filaments are juxtaposed to the postsynaptic density, they do not appear to be continuous with the constituent filaments of the density. We suggest that the postsynaptic density at this class of synapse is composed of fine filamentous proteins that insert on the postsynaptic membrane and serve as a supporting framework for a variety of globular proteins. The globular proteins may vary qualitatively and quantitatively from junction to junction, and are positioned in the region of the spine that has the greatest concentration of ionized calcium entering with the synaptic current, and the greatest extent of postsynaptic depolarization.

Axons were loaded with calcium, rapidly frozen, and freeze-substituted. The endoplasmic reticulum, in addition to mitochondria, contained calcium deposits, as indicated by electron probe x-ray microanalysis. Oxalate injected into living axons helped to preserve calcium-containing deposits during preparation for microscopy. It is concluded that the endoplasmic reticulum is a calcium-sequestering compartment in the squid giant axon.

The three-dimensional structure and membrane specializations of the moth fast excitatory neuromuscular junction are qualitatively similar to those in fast vertebrate muscles. Both types of neuromuscular junction have linear, repeating presynaptic active zones, each of which lies over an aggregate of membrane particles on a process or fold of the muscle. The aggregates are congruent with the presumed sites of the postsynaptic receptors. Glial processes alternate with the muscle processes and curve around the nerve between the active zones. However, this presumed glutamate junction and the cholinergic neuromuscular junctions of vertebrates differ in two major ways. In the moth, there is an elaborate interdigitation of muscle and glial processes which is lacking in vertebrate synapses. Further, in the postsynaptic membrane of the moth the intramembrouns particles cleave with the external membrane leaflet instead of with the inner one. This latter difference, in conjunction with the particle distributions observed at other types of chemical synapse, is consistent with the hypothesis that there are morphological characteristics in membrane structure that reflect specific types of receptor-ion channel complexes.

Many cytoplasmic vesicles are observed to move along microtubules. Often, bidirectional movement of particles is observed on a single microtubule. We have isolated one cytoplasmic motor, kinesin, and defined another, the axoplasmic retrograde factor, which are capable of powering anionic latex beads toward the plus and minus ends of microtubules, respectively. Observations of vesicle movements show that vesicles have a defined direction of movement and that vesicles copurify with a kinesin motor activity. Current evidence suggests the hypothesis that kinesin and the retrograde motors power vesicle movements in vivo by attachment to the appropriate vesicle.

Smooth septate junctions from the midgut of the cricket, Acheta, and the horseshoe crab, Limulus, as well as Hydra-type septate junctions from the epidermis of Hydra have been studied by freeze-fracture after direct freezing using the liquid helium-cooled copper block/slam freezing method. The exoplasmic fracture face at both types of septate junction exhibits rows of closely packed but irregularly shaped intramembrane particles. Complementary to these particle rows, on the protoplasmic fracture face, are sharply defined grooves with a periodic variation in depth and width that was conspicuous in Hydra but less well defined in arthropods. The closely packed, irregular particles on the exoplasmic faces could represent plastically deformed portions of transmembrane proteins pulled through the bilayer during freeze-fracture. On the basis of this interpretation, the grooves on the protoplasmic faces represent a confluence of the bilayer disruptions occurring during fracturing. The structures observed here are different from those reported in replicas of glutaraldehyde-fixed and glycerol-cryoprotected tissue, in which the intramembrane junctional components partition with the protoplasmic face and often assume the appearance of continuous cylinders. This comparison illustrates some of the artifacts associated with freeze-fracturing and shadowing. On the basis of a comparison of freeze-fracture replicas and sections of lanthanum-infiltrated tissues, the relationship between intramembrane junctional components and intercellular septal elements is analysed.

The number of large intramembrane particles associated with sites of synaptic vesicle release at the squid giant synapse was determined and compared to the average maximal presynaptic calcium current in order to derive an estimate of the conductance each particle would have if it were a calcium channel. This value, 0.21 pS, compares favorably with conductances of calcium channels in other preparations, substantiating the idea that the large intramembrane particles, which are concentrated at "active zones," represent calcium channels.

Axoplasm from the squid giant axon contains a soluble protein translocator that induces movement of microtubules on glass, latex beads on microtubules, and axoplasmic organelles on microtubules. We now report the partial purification of a protein from squid giant axons and optic lobes that induces these microtubule-based movements and show that there is a homologous protein in bovine brain. The purification of the translocator protein depended primarily on its unusual property of forming a high affinity complex with microtubules in the presence of a nonhydrolyzable ATP analog, adenylyl imidodiphosphate. The protein, once released from microtubules with ATP, migrates on gel filtration columns with an apparent molecular weight of 600 kilodaltons and contains 110-120 and 60-70 kilodalton polypeptides. This protein is distinct in molecular weight and enzymatic behavior from myosin or dynein, which suggests that it belongs to a novel class of force-generating molecules, for which we propose the name kinesin.

The end bulb of Held in the rostral ventral cochlear nucleus of the chinchilla and guinea pig was studied with the freeze-fracture technique. The end bulb has multiple, small active zones which are uniformly distributed within the calyceal portion of this terminal. Single or small groups of active zones are surrounded by enlarged channels of extracellular space often containing processes of astrocytes. Small plasmalemmal deformations occur at these active zones. The number of these deformations is thought to be indicative of exocytotic transmitter release because they are more frequent in animals fixed in a noisy environment compared to animals fixed in a quiet environment. Thus, our study provides a basis for the quantitative study of changes in transmitter secretion at a central nervous system synapse driven by a controllable natural stimulus. The postsynaptic active zone at end bulbs resembles other excitatory synapses in the central nervous system in having an aggregate of large particles on the external membrane leaflet. This junctional aggregate of particles is coextensive with the presynaptic active zone and with the postsynaptic density seen in thin sections. Several perisynaptic aggregates of particles are deployed around each active zone on the external membrane leaflet. These irregularly-shaped aggregates occur preferentially opposite the channels of enlarged extracellular space and along the edge of the end bulb and are not components of intercellular junctions or plasmalemmal contacts with cytoplasmic organelles. Although the function of the different particle aggregates on the postsynaptic membrane is not clear, our findings provide a basis for studying the factors controlling and maintaining their structure as well as more evidence that a consistent relationship exists between types of synaptic action and structure of the postsynaptic membrane.

Striated muscles from Rana pipiens have been exposed for variable periods of time to Ringer solutions made hypertonic by addition of either sucrose or sodium chloride. The muscles have been rapid-frozen and then prepared for electron microscopy by either freeze-substitution, freeze-fracture or cryoultramicrotomy. The only compartment greatly affected by hypertonicity is the transverse tubular system, which is visibly swollen. None of the elements of the sarcoplasmic reticulum increase in size.

Assemblies are a specialization of intramembrane structure in astrocytes which is concentrated in astrocytic processes at the interface with vascular and cerebrospinal fluid compartments. When astrocytic processes are rapidly frozen and then freeze-fractured at very low stage temperature, the constituent particles of assemblies fracture in at least two planes in the lipid bilayer. The true outer surface of the astrocytic membrane can be exposed by etching rapidly frozen tissue and in such preparations the assemblies are seen to extend through the extracellular half of the membrane to be exposed on the surface to the extracellular fluid. The dimensions of the particles, their tendency to fracture at several levels and their exposure to the extracellular space all indicate that they are not composed solely of lipid. We conclude that assemblies represent a protein which partially or entirely bridges the membrane and which may serve a transport function.

Aldehyde fixation continues to be useful to prepare synapses for freeze fracture, but it may increase the rate of transmitter release. The effects of different aldehyde fixatives on spontaneous quantal release (m.e.p.p.s), and on the corresponding synaptic vesicle exocytosis at frog nerve-muscle synapses were investigated with the hope of finding a way to minimize side effects of fixation. Increases in m.e.p.p.s of up to 50 s−1 occurred during fixation, despite the species of aldehyde used in the fixative, and this fixative effect decreased only slightly as aldehyde concentration was increased. Increases in m.e.p.p. frequency were not blocked by tetrodotoxin, by lowering external calcium and raising external magnesium concentration, or by lowering the total osmotic strength of the fixative. The smallest increase in m.e.p.p. frequency was in 3% glutaraldehyde and corresponded to the lowest levèl of synaptic vesicle exocytosis seen by freeze-fracture, 0·15 per μm of active zone. The effects of aldehyde fixation on m.e.p.p. frequency and synaptic vesicle exocytosis could not be avoided, but this study suggests how its effect on morphological changes in synapses might be minimized.

Storage-body diameter and volume, and the number of molecules of serotonin contained in a storage body, were estimated for blood platelets. In the human, 5.23 x 10(5) molecules of serotonin are contained in a storage body 198 nanometers in diameter, while in the cat, 31.2 x 10(5) molecules of this amine are contained in a storage body 298 nanometers in diameter.

The internal organization of plasmalemmal membrane, as revealed by freeze-fracture techniques, varies dramatically and predictably over the surface of astrocytes in mouse cerebellar cortex. Assemblies of uniform, small intramembrane particles packed in orthogonal order into square or rectangular aggregates are specialized distributions of intramembrane particles which, in the cerebellar cortex, are found only in astrocytes. The concentration of assemblies is greatest in astrocytic membrane juxtaposed to vascular structures or facing the cerebrospinal fluid at the glial limitans. Many fewer are present in regions of astrocytic membrane apposed to neural structures and virtually none are present on the astrocyte cell body. Corresponding structures have not yet been found in thin-sectioned preparations. While the distribution of assemblies in membranes facing blood and cerebrospinal fluid compartments suggests that they may have a role in transport of some material into or out of those compartments, their function is unknown. A second, distinct specialization of intramembrane structure appears to represent a junction between apposed astrocytic processes. We have provisionally described this as a ‘polygonal particle junction’, since it appears as large, irregular particles densely packed without obvious order in co-extensive regions of two astrocytic membranes. This junction is regularly present just below the cerebellar surface in the processes of the glial limitans as well as between large, more proximal radial Bergmann fibers, and also occurs occasionally throughout the molecular layer. With tannic acid mordant after aldehyde-osmium fixation or rapid freezing and freeze-substitution, it is possible to demonstrate subtle electron-dense specializations of the astrocytic membranes and extracellular matrix in thin-sections which correspond to the sites of polygonal particle junctions. The function of this astrocytic specialization is also unknown. Cerebellar astrocytes manifest numerous gap junctions as well, whose structure in freeze-fractured and thin-sectioned preparations is quite distinct from that of assemblies or of polygonal particle junctions.

Publisher Summary This chapter discusses the sequence of membrane relationships underlying exocytosis at the surface of Limulus amebocytes. The scheme is used for peripheral granule exocytosis as a basis of comparison with the compound exocytosis of more internal granules reported in this chapter. One striking difference between exocytosis at the surface of amebocytes and internal, compound exocytosis is in the formation of an apposition among the fusing membranes. Appositions of secretory granules with the plasmalemma are rare in unstimulated cells but common among secretory granules in the cell interior. An array of filaments connects the plasmalemma with some peripheral granules and may be responsible for pulling the plasmalemma in contact with the underlying granule. The intramembrane particles are laterally displaced from appositions among secretory granules and from contacts among isolated secretory granules forced together by centrifugation. Limulus amebocytes offer the necessary synchronous secretion and rapid freezing offers the necessary time resolution to determine the role of osmotic forces in initiating compound exocytosis.

Hyperosmotic solutions of 3 M urea, either infused into one internal carotid artery or applied topically to the pia mater of rabbits, results in the opening of endothelial tight junctions through which horseradish peroxidase passes from blood to extracellular fluid of the brain. The evidence for this opening of the blood-brain barrier to protein is the entry of peroxidase into the extracellular pools between successive tight junctions. In animals not receiving 3 M urea, the interjunctional pools are inaccessible to proteins. Having passed through the endothelial junctions, the peroxidase spreads along the extracellular channels of the perivascular neuropil for approximately 100 μ in 90 seconds. Most of the affected vessels are capillaries, though larger vessels are rendered leaky as well. Calyciform cisterns, that lie beneath shallow notches in the endothelium of untreated rabbits, appear to be enlarged after the administration of 3 m urea. It is undetermined whether these few endothelial cisterns and vesicles are involved in carrying protein from blood to the cerebral extracellular fluid.

Anatomical and physiological evidence based on independent studies of the mammalian olfactory bulb points to synaptic interactions between dendrites. A theoretical analysis of electric potentials in the rabbit olfactory bulb led originally to the conclusion that mitral dendrites synaptically excite granule dendrites and granule dendrites then synaptically inhibit mitral dendrites. In an independent electron micrographic study of the rat olfactory bulb, synaptic contacts were found between granule and mitral dendrites. An unusual feature was the occurrence of more than one synaptic contact per single granule ending on a mitral dendrite; as inferred from the morphology of these synaptic contacts, a single granule ending was often presynaptic at one point and postsynaptic at an adjacent point with respect to the contiguous mitral dendrite. We postulate that these synaptic contacts mediate mitral-to-granule excitation and granule-to-mitral inhibition. These dendrodendritic synapses could provide a pathway for both lateral and self inhibition.

Bilayers incorporating phospholipids that are capable, in isolation, of forming non-bilayer structures show ‘pits and particles’ or ‘cusps’ by freeze-fracture electron microscopy1–6. These points of very high curvature in the bilayer have been interpreted either as inverted micelles within a single bilayer2,3 or intersecting bilayers4,5, or as intermembrane sites of attachment1,6. We have used extremely rapid freezing and cleavage at very low temperature to look at the cardiolipin–phosphatidylcholine system and report here a high incidence of cross-fractured liposomes, which shows that even at lower CaCl2 concentrations they are filled with membranous structures. More interestingly, the arrays of pits and particles are closely aligned and associated with two further kinds of asymmetric bilayer deformation. Flat areas in the bilayer seem to represent sites of particularly adherent interbilayer contact and deep imaginations appear to represent sites where fusion with internal bilayers has occurred.

The junctions between adjacent endothelial cells of the cerebral capillaries in normal vertebrates determine whether or not proteins can leave the capillary lumen to move into the parenchymal interspaces. In rodents (mice and hamsters), these junctions are of the tight variety, such as that shown in Fig.2(TJ), and are typical of endothelium in normal brain tissue. Such tight junctions appear to form circumferential belts of occlusion between contiguous endothelial cells. A protein such as horseradish peroxidase, when injected into the blood, is thereby prevented from leaving the capillary lumen to enter the extracellular clefts of the neuropil. The same tight junctions also stop peroxidase from entering the capillary lumen after this protein has been injected into the cerebral ventricles and has moved throughout the extracellular clefts.

Given reports of raft domains in rodent brain synaptosomes, we probed for their conservation in a poikilotherm, the Woods Hole squid Loligo. Because rafts are described as liquid-ordered phase-separated domains, we expected phase transition temperatures above body temperature if raft function is important to neuronal activity. We tested this hypothesis by comparing synaptosomes, intact nerve endings, from animals that live at two very different temperatures: mouse (body temperature-37°C) and squid (body temperature-20°C). We measured the temperature-dependence of the lipid phase of intact synaptosomes in the absence of exogenous probes by using line-width and spinning sideband intensities of lipid hydrocarbon chain resonances, using proton magic angle spinning NMR spectra as a function of temperature between 0 and 40°C. We also determined the phase behavior of total synaptosomal lipids by repeating the NMR protocol on extracted lipids. To our surprise, there was no detectable signal for liquid-ordered lipid phases at the body temperature of either species. However, there was a difference in the temperature of the onset of order as synaptosomes were cooled below body temperature: the phase state of mouse synaptosomal membranes changes drastically below 24°C, whereas this change occurs below 8°C for the squid synaptosomal membranes. We then measured the composition of synaptosomes in terms of total lipid heads and tails, and the main difference arises from high concentrations of omega-3 poly-unsaturated fatty acids and cholesterol in the squid. Fluorescence microscopy images of squid synaptosomes stained with lipid dyes confirmed the formation of domains below, but not above the phase transition temperatures obtained from NMR measurements. Thus although the membranes of synaptosomes contain lipids that can phase-separate, these lipids remain in the liquid-disordered state at the usual physiological temperatures for squid and mouse.

The antimitotic drug Cryptophycin-1 (Cr) arrests cell growth by destabilizing microtubules. When added in vitro to either dimeric tubulin or preformed microtubules, Cr induces the formation of single protofilament rings that are highly monodisperse, having a diameter of ∼28nm and thickness ∼4 nm, and composed of eight tubulin dimers. Here we apply negative stain tomography [Fera, et al, J. Comp. Neurol., 2012] to Cr-tubulin rings. This new technique, which utilizes a beam stable organo-tungsten negative stain, allows construction of tomograms resolving the individual components of the protofilament rings without averaging. Molecular structure of Cr-tubulin rings is revealed in series of en face virtual sections ∼4 A thick positioned parallel to the plane of Cr-tubulin rings. We repeated the observations with rings obtained from tubulin-S, which lack the C-terminal acidic tail filaments that face the center of the Cr-tubulin rings. These tail filaments, normally on the outside of the microtubule and important for MAP and motor protein binding, have a diameter of ∼1 nm and significant conformational freedom, making them impossible to detect by averaging methods. Thus, comparison of tomograms from rings of tubulin and tubulin-S will gather further information about these filaments. We also investigate ring-ring interactions in samples of higher concentration to probe their material properties.