Simultaneous ultrastructural demonstration of retrogradely transported horseradish peroxidase and choline acetyltransferase immunoreactivity

In order to study the synaptic connections of neurons identified by their projection target and neurotransmitter content, we have adapted a method of combining retrograde tracing of horseradish peroxidase (HRP) and immunocytochemistry at the electron microscopic level. HRP was injected into the rat amygdala. Sections from the rostral forebrain were processed according to the 3,3'-diaminobenzidine/glucose oxidase reaction followed by choline acetyltransferase (ChAT) localization. Neurons in the ventral pallidum which contained both the diffuse immunoperoxidase reaction product (ChAT) and large electron dense bodies characteristic of retrogradely transported HRP were defined as double labeled, i.e. cholinergic neurons that project to the amygdaloid body.     

Transneuronal uptake of horseradish peroxidase in the central nervous system of dipterous insects

Summary Horseradish peroxidase (HRP) applied to lesioned neurons in the retina and thoracic ganglia of the flies Musca, Calliphora and Drosophila labeled axon terminals, dendrites and perikarya of the severed neurons after anterograde or retrograde passage. In addition, HRP reaction product secondarily labeled intact neurons that are contiguous with injured nerve cells. In many cases labeling of optic lobe neurons remote from primarily filled ones was also seen (here called tertiary labeling). HRP labeling was extensive and both primarily and transneuronally filled neurons could be resolved in almost as much detail as Golgi-impregnated or cobalt-silver-labeled cells. Electron microscopy showed that in both primarily and secondarily filled neurons, reaction product was distributed diffusely in the cytoplasm.Transneuronal uptake of HRP was specific to certain types of neurons in the brain and thus displayed certain pathways. The pathways resolved by transneuronal labeling with HRP extend from the optic lobes to the thoracic ganglia and include visual neurons previously identified electrophysiologically and anatomically.Transneuronal HRP uptake, although believed to occur in vivo, could not be shown to be dependent on synaptic activity. Three other heme peptides tested were taken up by injured neurons, but showed no transneuronal labeling: lactoperoxidase, cytochrome c, and microperoxidase.     

Expression of the cholinergic gene locus in the rat placenta

High amounts of acetylcholine (ACh) and its synthesising enzyme choline acetyltransferase (ChAT) have been detected in the placenta. Since the placenta is not innervated by extrinsic or intrinsic cholinergic neurons, placental ACh and ChAT originate from non-neuronal sources. In neurons, cytoplasmic ACh is imported into synaptic vesicles by the vesicular acetylcholine transporter (VAChT), and released through vesicular exocytosis. In view of the coordinate expression of VAChT and ChAT from the cholinergic gene locus in neurons, we asked whether VAChT is coexpressed with ChAT in rat placenta, and investigated this issue by means of RT-PCR, in situ hybridisation, western blot and immunohistochemistry. Messenger RNA and protein of the common type of ChAT (cChAT), its splice variant peripheral ChAT (pChAT), and VAChT were detected in rat placenta with RT-PCR and western blot. ChAT in situ hybridisation signal and immunoreactivity for cChAT and pChAT were observed in nearly all placental cell types, while VAChT mRNA and immunolabelling were detected in the trophoblast, mesenchymal cells and the visceral yolk sac epithelial cells. While ChAT is nearly ubiquitously expressed in rat placenta, VAChT immunoreactivity is localised cell type specifically, implying that both vesicular and non-vesicular ACh release machineries prevail in placental cell types.     

Cytosolic Choline Acetyltransferase Binds Specifically to Cholinergic Plasma Membrane of Rat Brain Synaptosomes to Generate Membrane-Bound Enzyme

Uncovering the way membrane-bound choline acetyltransferase (ChAT) interacts with membranes and with which membrane in cholinergic neurons may help in understanding its role in acetylcholine metabolism. Subfractionation of rat hippocampal synaptosomes aiming to separate synaptic vesicles from plasma membranes shows that membrane-bound ChAT is bound to plasma membrane. Either detergents or urea and alkali can solubilize membrane-bound enzyme. Detergent-solubilized enzyme has a higher sedimentation rate than urea-alkali solubilized or cytosolic ChAT. Once dissociated, membrane-bound ChAT reassociates specifically with cholinergic plasma membranes, a process that was abolished by previous treatment of membranes with trypsin. Cytosolic ChAT behaves similarly. Thus, in cholinergic synaptosomes, ChAT exists as cytosolic and peripheral activity. Cytosolic ChAT generates peripheral enzyme most probably by interacting with a protein of plasma membrane of cholinergic nerve terminals. This receptor protein might regulate the amount of membrane-bound ChAT in cholinergic neurons.     

Developmental regulatory elements in the 5′ flanking DNA of the Drosophila choline acetyltransferase gene

Summary Choline acetyltransferase (ChAT, EC 2.3.1.6) catalyzes the production of the neurotransmitter acetylcholine, and is an essential factor for neurons to be cholinergic. We have analyzed regulation of the Drosophila ChAT gene during development by examining the -galactosidase expression pattern in transformed lines carrying different lengths of 5 flanking DNA fused to a lacZ reporter gene. The largest fragment tested, 7.4 kb, resulted in the most extensive expression pattern in embryonic and larval nervous system and likely reflects all the cis-regulatory elements necessary for ChAT expression. We also found that 5 flanking DNA located between 3.3 kb and 1.2 kb is essential for the reporter gene expression in most of the segmentally arranged embryonic sensory neurons as well as other distinct cells in the CNS. The existence of negative regulatory elements was suggested by the observation that differentiating photoreceptor cells in eye imaginal discs showed the reporter gene expression in several 1.2 kb and 3.3 kb transformants but not in 7.4 kb transformants. Furthermore, we have fused the 5 flanking DNA fragments to a wild type ChAT cDNA and used these constructs to transform Drosophila with a Cha mutant background. Surprisingly, even though different amounts of 5 flanking DNA resulted in different spatial expression patterns, all of the positively expressing cDNA transformed lines were rescued from lethality. Our results suggest that developmental expression of the ChAT gene is regulated both positively and negatively by the combined action of several elements located in the 7.4 kb upstream region, and that the more distal 5 flanking DNA is not necessary for embryonic survival and development to adult flies. Correspondence to: P.M. Salvaterra     

Effects of Sex Hormones,Forskolin, and Nicotine on Choline Acetyltransferase Activity in Human Isolated Placenta

The activity of choline acetyltransferase (ChAT) was investigated in the human placenta before and after long-term incubation (24 h) to test the effects of sex hormones, nicotine and forskolin. ChAT activity differed considerably between the amnion (0.03 mol/mg protein/h) and the villus (0.56). After long-term incubation, ChAT activity persisted in the latter but declined in the amnion. Neither sex hormones (-estradiol, testosterone, progesterone; 10 or 100 nM each) nor follicle stimulating hormone and luteinizing hormone (FSH/LH; 8.4 U/ml each) modified ChAT activity. Also nicotine (1 nM–100 M) did not affect ChAT activity. Forskolin, an activitor of adenylyl cyclase, reduced ChAT activity in the villus but not in amnion. The present model offers the possibility to investigate ChAT regulation in intact tissue under long-term incubation. The risks of maternal smoking during pregnancy cannot be attributed to an effect of nicotine on placental ChAT activity. Differences in the regulation of ChAT appear to exist between neuronal and nonneuronal cells.     

Neurochemically similar myenteric and submucous neurons directly traced to the mucosa of the small intestine

Summary Antisera to neuropeptide Y (NPY) gave an intense immunohistochemical reaction of certain nerve cells in the myenteric and submucous plexuses of the guinea-pig small intestine. Each nerve cell had up to 20 branching, tapering processes that were less than 50 m long and a long process that could be followed for a considerable distance. This morphology corresponds to that of the type-III cells of Dogiel. The long process of each myenteric cell ran through the circular muscle to the submucosa, and in most cases the process could be traced to the mucosa. The submucous nerve cell bodies also had processes that extended to the mucosa. These cell bodies, in both plexuses, also stained with antisera raised against calcitonin generelated peptide (CGRP), cholecystokinin (CCK), choline acetyltransferase (ChAT) and somatostatin (SOM), but did not stain with antibodies against enkephalin, substance P or vasoactive intestinal peptide. Thus, it has been possible for the first time to trace the processes of chemically specified neurons through the layers of the intestinal wall and to show by a direct method that CGRP/CCK/ChAT/NPY/ SOM myenteric and submucous nerves cells provide terminals in the mucosa.     

Functional implications of the projections of neurons from individual labellar sensillum of Drosophila melanogaster as revealed by neuronal-marker horseradish peroxidase

Summary Earlier studies using Golgi silver impregnations from the labellar sensilla of adult Drosophila melanogaster revealed seven types of sensory axons projecting into the suboesophageal ganglion of the brain. These sensory terminals were designated as coiled fibres (type-I), shrubby fibres (type-II), ipsilateral ventral fibres (type-III), ipsilateral dorsal fibres (type-IV), contralateral ventral fibres (type-V), contralateral dorsal fibres (type-VI), and central fibres (type-VII). The present study identifies the projections of sensory neurons present in a single labellar taste-sensillum, using the neuronal marker horseradish peroxidase (HRP). Although the taste sensillum in question has five neurons, in a given experiment only one or at the most two neurons are labelled. The type of neuron labelled was usually specific to the stimulant solute (sucrose, sodium chloride or potassium chloride) present in the HRP solution. Although type-II fibres get labelled most of the time, irrespective of the stimulant present in HRP solution, type-IV fibres are labelled when attractants (0.1 M sucrose or 0.1 M sodium chloride) are used as stimulants in HRP solution. Type-VI fibres are labelled when the stimulant is 0.1 M potassium chloride, a repellent. HRP dissolved in distilled water revealed type-I coiled fibres. Besides revealing projections of sensillar neurons to the brain the present technique also inferred their possible function. Incubation of whole-brain tissue with 0.04% 3,3-diaminobenzidine tetrahydrochloride in presence of 0.06% hydrogen peroxide suggested that the glomerular organization is also present in the taste-sensory region as it is in olfactory neuropile.     

Apical membrane endocytosis via coated pits is stimulated by removal of antidiuretic hormone from isolated,perfused rabbit cortical collecting tubule

Summary Antidiuretic hormone increases the water permeability of the cortical collecting tubule and causes the appearance of intramembrane particle aggregates in the apical plasma membrane of principal cells. Particle aggregates are located in apical membrane coated pits during stimulation of collecting ducts with ADHin situ. Removal of ADH causes a rapid decline in water permeability. We evaluated apical membrane retrieval associated with removal of ADH by studying the endocytosis of horseradish peroxidase (HRP) from an isotonic solution in the lumen. HRP uptake was quantified enzymatically and its intracellular distribution examined by electron microscopy. When tubules were perfused with HRP for 20 min in the absence of ADH, HRP uptake was 0.5±0.3 pg/min/m tubule length (n=6). The uptake of HRP in tubules exposed continuously to ADH during the 20-min HRP perfusion period was 1.3±0.8 pg/min/m (n=8). HPR uptake increased markedly to 3.2±1.1 pg/min/m (n=14), when the 20-min period of perfusion with HRP began immediately after removal of ADH from the peritubular bath. Endocytosis of HRP occurred in both principal and intercalated cells via apical membrane coated pits. We suggest that the rapid decline in cortical collecting duct water permeability which occurs following removal of ADH is mediated by retrieval of water permeable membrane via coated pits.     

Horseradish peroxidase C

Summary Horseradish peroxidase C (HRP; ferric) reacts with H₂O₂ to form Compound I, with an equilibrium constant of about 10¹⁴ M^–1. Two-step reduction of Compound I to Compound II and further to the ferric enzyme occurs reversibly at E_o values of 0.90 and 0.93 V (pH 7.0), respectively. The pH dependence of E_o values for each one-electron step, ferrous ferric Compound II Compound I indicates the presence of redox-linked ionization at pK_a values of 7.3 in the ferrous state, 11.0 in the ferric and 8.6 in Compound II. Zinc-substituted HRP C is oxidized to its free-radical form at an E_o value of 0.74 (pH 6.0). Comparison of oxidized zinc HRP C with Compound I shows that Compound I contains a porphyrin -cation radical. The flash photolysis study on the NO-ferric HRP C complex clearly indicates that the iron is pentacoordinated in HRP C while it is hexacoordinated in metmyoglobin. From the kinetic analysis of the acid-alkaline conversion of HRP C, the second-order rate constants of the reactions with H⁺ and HO^– are estimated to be 1.5 × 10¹⁰ and 6.7 × 10⁴ M^–1s^–1, respectively. The latter rate constant greatly varies with the kind of hemoproteins. In the presence of HRP C and O₂, indole-3-acetate is oxidized to its hydroperoxide form, which reacts effectively with HRP C to form Compound I and further converts Compound I to a verdohemoprotein.Abbreviations HRP horseradish peroxidase (HRP without subgroup letter denotes a classical preparation consisting of HRP B and HRP C) - EPR electron paramagnetic resonance - NMR nuclear magnetic resonance     

Cytochemical detection of mannose-specific receptors for glycoproteins with horseradish peroxidase as a ligand

Summary Horseradish peroxidase (HRP), a glycoprotein rich in mannose groups, was used as a ligand to detect receptors for glycoproteins in formalinfixed, frozen sections of rat liver. Specific binding of HRP occurred to surface membranes of sinusoidal cells but not to those of parenchymal cells. The binding sites were visualized after the peroxidatic reaction in erythrocytes had been suppressed by methanol-H₂O₂ and phenylhydrazine, the latter reagent also decreasing the nonspecific background adsorption of HRP. Several factors influencing the reaction were studied systematically. The specific binding of HRP to sinusoidal cells was greatly decreased or abolished when tissue blocks were fixed for longer than 1–2 h in a cold 4% formaldehyde solution and the frozen sections subsequently treated for 30 min in cold methanol. The specific binding of HRP increased when the concentration of HRP in the medium was increased from 10 g/ml to 40 g/ml, when the time of incubation with HRP was increased from 1 h to 4 h, or when the temperature of incubation with HRP was increased from 4°C to 22°C, or from 22°C to 37°C. The specific binding of HRP also increased when the pH of the incubation medium was increased from 7.0 to 10.0. Little or no specific binding of HRP was observed in the absence of added Ca⁺⁺. The binding of HRP was suppressed by 10 mM mannose or 0.004% mannan whereas the suppression of the binding reaction by galactose or galactan required 30–40 times higher concentrations.This work was supported by the Morris A. Kaplan Fund     

Unusual binding sites for horseradish peroxidase on the surface of cultured and isolated mammalian cells

Summary Binding sites for horseradish peroxidase (HRP), with unusual properties, were detected on the surface of cultured and isolated cells after the cells (on cover slips) had been quickly dried, fixed in cold methanol, and postfixed in a paraformaldehyde solution. The reaction for surface-bound HRP was suppressed by micromolar concentrations of glycoproteins such as invertase, equine luteinizing hormone (eLH) or human chorionic gonadotropin (hCG). The reaction was also suppressed by 20 mM CDP, UDP, GTP, NAD, and ribose 5-phosphate. Two to six times higher concentrations of GMP, fructose 1-phosphate, galactose 6 phosphate, mannose 6-phosphate, fructose 6-phosphate, and glucose 6-phosphate were required to suppress the binding eaction. AMP, ATP, heparin, mannan, and eight non-phosphorylated sugars showed relatively low competing potencies but fucoidin and -lactalbumin were strong inhibitors. No addition of Ca²⁺ was required for the binding of HRP to the cell surface. However, calcium-depleted, inactive HRP did not compete with the binding of native (calcium-containing) HRP whereas H₂O₂-inactivated HRP suppressed the binding. GTP, NAD, ribose 5-phosphate, and EGTA accelerated the release of previously-bound HRP from the cell surface whereas glycoproteins (invertase, cLH, and hCG) did not do se. Addition of Ca²⁺ to GTP, NAD, ribose 5-phosphate or to EGTA prevented the accelerated release of HRP from the cell surface. It is suggested that calciam, present either in the surface membrane or in HRP itself, is involved in the binding of HRP to the cell surface and in the inhibition of binding by GTP, NAD, and ribose 5-phosphate. It is also suggested that -lactalbumin, GTP, UDP, and CDP compete with the binding of HRP to a glycosyltransferase on the cell surface.     

Permeability of submandibular salivary glands in dogs to blood-borne horseradish peroxidase (HRP)

Summary Horseradish peroxidase (HRP) was administered to the submandibular glands of dogs by close-arterial bolus-type injections, and its localisation was examined histochemically by light and electron microscopy. The HRP became widespread in the interstices of the glands and reached many central acinar lumina via scattered localised parts of their tight junctional complexes. Reaction product was less often found in the lumina of demilunes, which suggested that the intercellular junctions there were less leaky. HRP was often found in sizeable spaces between myoepithelial cells and the underlying parenchymal cells; such large spaces have not been observed in this situation in other species. The possibility that permeability pathways may arise intermittently at different sites in the adhering mechanisms between the acinar cells is discussed.It is concluded that potential paracellular permeability pathways for macromolecules exist in these glands and, if the concentration gradient is sufficiently high, molecules even as large as those of HRP can to some extent permeate passively from the interstices to the saliva. In resting glands the principal permeability site is between the central acinar cells.Supported by Grants from the M.R.C. and the V.R.T. King's College HospitalWe wish to acknowledge the technical help of Mr. K.J. Davies and Mr. P.S.A. Rowley     

Exploring the Regulation of the Expression of ChAT and VAChT Genes in NG108-15 Cells: Implication of PKA and PI3K Signaling Pathways

Involvement of different protein kinases regulated by cAMP and implication of muscarinic receptors in the regulation of choline acetyltransferase (ChAT) and vesicular acetylcholine transporter (VAChT) mRNA levels and ChAT activity has been studied in NG108-15 cells. Dibutyryl cAMP enhanced both ChAT and VAChT mRNA levels and stimulated ChAT activity. Muscarinic stimulation or inhibition did not change ChAT activity or the receptor subtype mRNA pattern. MEK1/2 did not affect the regulation of ChAT and VAChT mRNA levels. However, PKA plays a major role in regulating ChAT and VAChT mRNA levels, because H89 decreased both. Strikingly, inhibition of PI3K by LY294002 had two opposite effects: ChAT mRNA level was decreased and VAChT mRNA level was increased. Such a result consolidates the observation that ChAT and VAChT genes, despite their unusual organization in a single cholinergic locus, can be differentially or synergistically regulated, depending on the activated signaling pathways.     

Localization of choline acetyltransferase-expresing neurons in the larval vissual system of Drosophila melanogaster

Choline acetyltransferease (ChAT) is the enzyme catalyzing the biosynthesis of acetylcholine and is considered to be a phenotypically specific marker for cholinergic neurons. We have examined the distribution of ChAT-expressing neurons in the larval nervous system of Drosophila melanogaster by three different but complementary techniques: in situ hybridization with a cRNA probe to ChAT messenger RNA, immunocytochemistry using a monoclonal anti-ChAT antibody, and X-gal staining of transformed animals carrying a reporter gene composed of 7.4 kb of 5 flanking DNA from the ChAT gene fused to a lacZ reporter gene. All three techniques demonstrated ChAT-expressing neurons in the larval visual system. In embryos, the photoreceptor organ (Bolwig's organ) exhibited strong cRNA hybridization signals. The optic lobe of late third-instar larvae displayed ChAT immunoreactivity in Bolwig's nerve and a neuron close to the insertion site of the optic stalk. This neuron's axon ran in parallel with Bolwig's nerve to the larval optic neuropil. This neuron is likely to be a first-order interneuron of the larval visual system. Expression of the lacZ reporter gene was also detected in Bolwig's organ and the neuron stained by anti-ChAT antibody. Our observations indicate that acetylcholine may be a neurotransmitter in the larval photoreceptor cells as well as in a first-order interneuron in the larval visual system of Drosophila melanogaster.This work was supported by a grant from the National Institute of Neurological Disorders and Stroke.     

Localization of NADPH-diaphorase and choline acetyltransferase in the central nervous system of the bivalve mollusk Mactra sulcatoria

The presence of NADPH-diaphorase and choline acetyltransferase (ChAT) in all ganglia of the Mactra sulcatoria was demonstrated by histochemical and electron histochemical methods. Pecularities of cholinergic and nitrergic neurons localization were revealed in nervous ganglia, and their relative content there was estimated. It was established that in reaction to ChAT only large neurons were marked. Ultrastructural localization of NADPH-diaphorase and ChAT was determined in neurons and neuropile. The data obtained testify that NADPH-diaphorase and ChAT are located in different types of nervous cells. The opportunity of functional cooperation in activity of cholinergic and nitrergic systems in mollusks is discussed.     

Developmental changes in choline acetyltransferase and glutamate decarboxylase activity in various regions of the brain of the male,female, and neonatally androgenized female rat

In an attempt to discern effects of sex hormones on the development of neurotransmitter systems in the rat brain, choline acetyltransferase (ChAT) and glutamate decarboxylase (GAD) have been measured at postnatal days 8, 12, 25, and 60 in five regions (amygdala, anterior hypothalamus, hippocampus, olfactory bulbs, and cerebral cortex) of the brains of normal male, normal female, and neonatally androgen-treated female rats. Essentially no associations between sex or of neonatal androgenization on either enzyme were found. The data, however, provide new information on the relative rates of development of ChAT and GAD in five regions of the rat brain which supplement the limited information already available in the literature. ChAT activity was highest in amygdala and hypothalamus, but developed most rapidly in hippocampus and cerebral cortex. The relative activities and patterns of development of GAD activity were similar to those of ChAT.     

Binding sites for horseradish peroxidase on the cell surface

Summary The cytochemical reaction for surface-bound horseradish peroxidase (HRP) on cultured HeLa cells, GH₃ cells, and isolated rat liver cells was suppressed by 30 M monosialoganglioside, by 30 M trisialoganglioside, or by 5 mM CMP-neurminic acid. The reaction was also suppressed by 10 mM chitotriose or by 10 mM UDP-galactose, a galactose acceptor and donor, respectively, for galactosyltransferase. The addition of 2 mM Mn²⁺ to the incubation medium with HRP suppressed the reaction for surfacebound HRP, and the addition of 10–20 mM Ca²⁺ intensified the reaction. The addition of 2 mM Zn²⁺ caused less inhibition than that of 2 mM Mn²⁺, and the addition of 2 mM Co²⁺ caused either a slight inhibition, or no inhibition. These observations support the hypothesis that HRP may be bound to a glycosyltransferase at the cell surface.     

Morphology and central synaptic connections of the efferent neurons innervating the crayfish hindgut

Summary The distribution, morphology and synaptic connections of the hindgut efferent neurons in the last (sixth) abdominal ganglion of the crayfish, Orconectes limosus, have been investigated using light and electron microscopy in conjunction with retrograde cobalt/nickel and HRP labeling through the intestinal nerve. The hindgut efferent neurons occur singly and in clusters, and are unipolar. Their axonal projections are uniform and consist of a thick primary neurite with typical lateral projections and limited arborization of varicose fibers in the ganglionic neuropil. They also send lower order axon processes to the ganglionic neural sheath, where they arborize profusely, forming a network of varicose fibers. The majority of the efferent neurons project to the anterior part of the hindgut. HRP-labeled axon profiles are found in both pre- and postsynaptic position in the neuropil of the ganglion. HRP-labeled axon profiles also establish pre- and postsynaptic contacts in the intestinal nerve root. All hindgut efferent terminals contain similar synaptic vesicle populations: ovoid agranular vesicles (50–60 nm) and a few large granular vesicles (100–200 nm). It is suggested that the hindgut efferent neurons in the last abdominal ganglion are involved in: (1) innervation of the hindgut; (2) central integrative processes; (3) en route synaptic modification of efferent and afferent signals in the intestinal nerve; (4) neurohumoral modulation of peripheral physiological processes.Fellow of the Alexander von Humboldt Stiftung     

Fine structure of the adhesive pads of Gonionemus vertens

Summary The axonal transport of HRP in both the peripheral and central branches of dorsal root ganglion cells was studied in rats.For studying axonal transport in the peripheral branch HRP as a dry substance was applied to the peroneal nerve injured either by teasing, by cutting or crushing. After a short survival time (22 h) mainly small spinal ganglion cells of the corresponding segments were labelled, while after a prolonged survival time (70 h) mainly large cells were labelled. These labelling differences are referred to different transport rates or to differences in the process of accumulation of HRP in neurons of various sizes. No evidence could be found for HRP transport from the peripheral into the central branch.Injection of HRP into the spinal cord (survival time 22 h) or into the dorsal column nuclei (survival time 46 h) was followed by labelling of numerous spinal ganglion cell perikarya of all sizes. Reaction product was found also within the prebifurcation segment of spinal ganglion cell processes.On the basis of light microscopic exploration only somatopetal transport could be detected.This investigation was supported by the Fonds zur Förderung der wissenschaftlichen Forschung in ÖsterreichThe authors wish to thank Prof. Dr. H. Holländer and his coworkers (Neuroanatom. Abteilung, Max Planck Institut für Psychiatrie, München) for many helpful suggestions to improve the technique. Thanks are also due to Dr. E. Krammer and Dr. H. Gruber for stimulating criticism and to Miss F. Schramm for technical assistanceDedicated to Prof. H. Ferner with best wishes on his 65th birthday