Ultrastructural localization of concanavalin A-binding sites in the Golgi apparatus of various types of neurons in rat dorsal root ganglia Functional implications.код для вставкиСкачать
THE AMERICAN JOURNAL OF ANATOMY 177%-95 (1986) Ultrastructural Localization of Concanavalin A-Binding Sites in the Golgi Apparatus of Various Types of Neurons in Rat Dorsal Root Ganglia: Functional Implications F. MALCHIODI, A. RAMBOURG, Y. CLERMONT, AND A. CAROFF Dkpartment de Bwlogie du CEA, 91191 Ckdex, Saclay, France F: M., A.R., A. C.); and Department of Anatomy, McGill University, Montreal, Canada H3A ZBZ (Y C.) ABSTRACT The localization of concanavalin A (con A) binding sites has been determined at the electron-microscopic level in the six types of neurons (Al, Az, As, B1, B2, C) of rat dorsal root ganglia. In all ganglion cells, con A stained the plasma membrane, the nuclear envelope, the cisternae of the rough endoplasmic reticulum, and the matrix of some multivesicular bodies. In contrast, the con A reactivity of the Golgi apparatus varied according to cell type. In type B1 and Bz cells and possibly in type A3 cells, the lectin was exclusively located in three or four saccules on the cis side of the Golgi stacks, whereas the TPPase-positive saccules and the trans sacculotubular elements were unstained with con A. In type Al, A2, and C neurons, all Golgi saccules as well as the trans sacculotubular elements were stained with the lectin. These results suggest that different types of glycoproteins were produced in these two groups of neurons. In the type Al, Az, and C cells, the persistence of the lectin reactivity in the WPase-positive saccules or sacculotubular elements on the trans side of the Golgi stacks suggests the presence of glycoproteins with oligosaccharide side chains rich in a-D-mannosyl residues in terminal positions. In contrast, the disappearance of the con A reactivity in equivalent elements of the Golgi stacks in type B1, B2, and A3 cells suggests the addition a t this level of other sugar residues characteristic of complex oligosaccharide side chains. The majority of the vesicular elements associated with the Golgi apparatus, as well as lysosomes, were unstained with con A. An analysis with the electron microscope of the tridimensional structure of the components of the Golgi apparatus and the use of phosphatase cytochemistry has permitted the characterization of several distinct compartments within this organelle. A comparative study of the Golgi apparatus of various cell types also revealed that, although the general structure of this organelle was usually similar, the relative development and cytochemical properties of its various compartments varied between cell types (Rambourg et al., 1974, 1979, 1981, 1984; Hermo et al., 1980; Clermont et al., 1981; Lalli, 1983; Rambourg and Clermont, 1986). In dorsal root ganglia, differences in the localization of phosphatases (Novikoff, 1967; Boutry and Novikoff, 1975)or in the incorporation of sugars in the side chains of glycoproteins (Droz, 0 1986 ALAN R. LISS, INC. 1967) have been reported in the Golgi apparatus of the two main classes of neurons found therein, i.e., the large and light type A neurons and the small and dark type B neurons. Rambourg et al. (1983), on the basis of the ultrastructural appearance and distribution of organelles within the perikarya of neurons, proposed a classification of ganglion cells into six types (Al, A2, As, B1, Bz, C). The characteristic features of the various types of ganglion cells are the following: Type Al are large neurons in which Nissl bodies, distributed throughout the cytoplasm, are Dr. Fiorella Malchiodi’s present address is Istituto Superiore di Sanit6, Viale Regina Elena 299,00161 Roma, Italia. Address reprint requests to Dr. A. Rambourg, Departement de Biologie du CEA, Gif-sur-Yvette,91191 CBdex, Saclay, France. Received December 16, 1985. Accepted May 19, 1986. 82 F.MALCHIODI ET AL. CON A-BINDING SITES IN NEURON GOLGI APPARATUS separated from each other by narrow cytoplasmic spaces containing small stacks of Golgi saccules and mitochondria. Type A2 are large neurons with Nissl bodies also distributed throughout the cytoplasm but separated from each other by wider strands of neuroplasm containing Golgi stacks and mitochondria. Type A3 are the smallest of type A cells and display densely packed Nissl bodies and long stacks of Golgi saccules which form a perinuclear ring in the midportion of the perikaryon. Type B1 are small neurons with Nissl bodies showing parallel cisternae located mainly in the outer cytoplasmic zone and large, curved stacks of Golgi saccules which form a broad perinuclear net. Type B2 are small neurons showing a cortical zone composed mainly of Nissl substance, and a Golgi apparatus forming a ring separated from the nucleus by a zone of cytoplasm containing mitochondria and smooth ER. Type C are the smallest of the ganglion cells and contain small, poorly demarcated Nissl bodies and a juxtanuclear Golgi apparatus (Rambourg et al., 1983). It thus became of interest to analyze at the ultrastructural level the binding of concana- CM ER G L m MVB NE v Abbreviations cytoplasmic membrane endoplasmic reticulum Golgi apparatus lysosome mitochondria multivesicular body nuclear envelope vesicles Fig. 1. Low-power EM photograph of a type C cell. Con A is seen along the cell-membrane surface of the satellite cells (S)and of the neuron. In the perikaryon itself, the lectin is located in cisternae of the rough endoplasmic reticulum and in the Golgi apparatus. In the latter, all saccules or sacculotubular elements are reactive with the lectin. The nuclear envelope is also positively stained. The rest of the nucleus (N) and mitochondria are unstained. x 18,000, Fig. 2. EM photograph of the perikaryon of a type Bz cell. The cytoplasmic membrane, the cisternae of rough endoplasmic reticulum forming a Nissl body, as well as the nuclear envelope are reactive to con A. Some saccules of a Golgi stack are also stained with the lectin, but the cis (osmiophilic) element (arrowhead) and trans saccular structures (arrow) are unreactive with the lectin. Vesicles in the Golgi region (asterisk), a multivesicular body, a dense granule identified as a lysosome, as well as mitochondria, are unlabeled with the lectin. ~39,000. 83 valin A (con A) to the various types of neurons, to study their Golgi apparatus in particular, using a con A-horseradish peroxidase complex to detect glycoconjugates containing a-D-mannopyranosyl residues (Gold stein and Hayes, 1978). The results obtained revealed marked differences in the distribution of the lectin in the Golgi apparatus of the various types of neurons, the functional implication of which will be examined closely. MATERIALS AND METHODS Adult rats were perfused through the left ventricle with a 2.5% solution of glutaraldehyde in 0.1 M sodium cacodylate buffer. After 15 min of perfusion, the L4 and L5 dorsal root ganglia were removed and postfixed for a n additonal hour in the same fixative. After a rinse in cacodylate buffer, they were impregnated in 30% glycerol and frozen in partly solidified freon 22 cooled by liquid nitrogen. The pieces of ganglia were transferred into liquid nitrogen and 20- to 30-pm-thick sections were cut on a cryostat. The sections were thawed in 30% glycerol and immersed in a phosphate-buffered saline (PBS) solution containing 0.05% saponin for 40 min at room temperature. They were incubated overnight a t room temperature in a solution of con A (IBF, 200 pg/ml) in PBS, rinsed with PBS, and treated with peroxidase (type I1 Sigma, 200 pg per ml in PBS) for 3 hr at room temperature. The bound peroxidase was then revealed by incubating sections €or 15 min a t room temperature in a PBS solution containing diaminobenzidine (DAB, grade 11, Sigma, 1 mg per ml) and 0.5% H202. Control sections were incubated in a con A solution containing 0.1 M methyl mannoside, or were treated with peroxidase or DAB without the lectin pretreatment. After a last rinse in PBS, treated or control sections were postfixed €or 1 hr at room temperature in a 1:l mixture of 2% aqueous osmium tetroxide and 3% aqueous potassium ferrocyanide (Karnovsky, 19711, dehydrated in ethanol, and embedded in Epon. Thin sections were prepared and mounted on copper grids. They were counterstained for 2 min in lead citrate and examined at 80 kV with a Philips 400 electron microscope. For the demonstration of the thiamine pyrophosphatase (TPPase) activity the glutaraldehyde-fixed ganglia were kept overnight at 4°C in a solution of 0.1 M sodium cacodylate containing 4% sucrose and 0.05% CaClz a t pH 7.2. They were then sectioned a t a 84 F. MALCHIODI ET AL. Figs. 3, 4. Sections treated with con A but not counterstained with lead acetate. Fig. 3. Perikaryon of a type B1 neuron. Cisternae of the rough endoplasmic reticulum as well as some elements of the Golgi stacks are stained with the lectin. Other Golgi elements are unstained with con A (arrowheads). In the concavity of a Golgi stack, some trans tubular elements also show a reactivity to con A (aster- isk). Other cytoplasmic organelles such as mitochondria are unstained. x 19,000. Fig. 4. Perikaryon of a type A2 cell showing con A binding to the rough endoplasmic reticulum and multivesicular bodies. The Golgi apparatus is also stained with the lectin; but in this cell and in contrast to the type B1 cell above (Fig. 3), all saccules or sacculotubular elements of the stack are reactive. ~ 2 3 , 5 0 0 . CON A-BINDING SITES IN NEURON GOLGI APPARATUS thickness of 80-100 pm with an Oxford vibratome. The sections were incubated for 2 hr at 37°C in the medium (pH 7.2) described by Novikoff and Goldfkcher (1961). After incubation the sections were washed in buffer, rinsed in sodium cacodylate containing sucrose (4%)and CaClz (0.05%),postfixed and stained with K-ferrocyanide-reduced osmium, and embedded in Epon. Thin sections were counterstained with lead citrate before examination with the electron microscope. RESULTS Reactivity of the various types of ganglion cells to con A In sections treated with con A and examined at the low magnifications of the EM, several structures were sharply outlined in the various types of ganglion cells. As illustrated in a C cell (Fig. 1) and a B2 cell (Fig. 2), the plasma membrane, nuclear envelope, and cisternae of the rough endoplasmic reticulum (ER) were well stained with con A. In addition several saccules and sacculotubular elements of the Golgi apparatus were also stained (Figs. 1, 2). In the type B2 cell, the outer cortical zone of the perikaryon containing the con A-positive cisternae of Nissl bodies was easily distinguished from the ER-free juxtanuclear region in which con A-positive Golgi elements contrasted sharply with unstained mitochondria and profiles of the smooth endoplasmic reticulum (Fig. 2). In control sections no con A binding was observed. The reactivity of cellular structures with con A was emphasized in sections that were not counter-stained with lead citrate (Figs. 3, 4). In such sections, the selective staining of ER cisternae and of some components of the Golgi apparatus was evident; but it also became apparent that, while in type B1 and B2 cells and also possibly in the rare type A3 cells only some of the Golgi saccules were reactive with con A (Fig. 3), in type Al, A2, and C neurons all saccules of the Golgi stacks were reactive (Fig. 4). However, a better understanding of the con A reactivity of the Golgi apparatus of the various types of neurons was obtained only when sections counterstained with lead citrate were examined at higher magnifications of the electron microscope. Con A reactivity of the Golgi apparatus of type BI, B2,and A3 neurons In type B1 and B2 cells, three or four lectinpositive elements were observed on the cis 85 side of the Golgi stacks (Figs. 5-7). The first one, which corresponded to the so-called cis (osmiophilic) element observed in the Golgi apparatus of various cell types, consisted of a network of tightly anastomosed tubules which appeared as a series of interconnected membranous profiles when the plane of section was perpendicular to the stack (Fig. 5). This element reacted either positively or negatively with the lectin (Figs. 5-7). Subjacent to it there were three parallel lectin-positive saccules which showed a nodular appearance when seen in transverse sections (Fig. 5). This was due to the presence of small pores in the saccules (Fig. 11). In addition, these labeled saccules were interrupted by perforations in register referred to as “wells” (Figs. 5-7, see terminology in Rambourg and Clermont, 1986). Such wells in the stacks of saccules were pan shaped with their mouth opening on the cis face and their bottom closed by the underlying saccules which were unreactive with the lectin but thiamine-pyrophosphatase positive (see below). Such wells, which involved only the lectin-positive saccules, contained small vesicles (80 nm in diameter) which were unstained with con A (Fig. 5-7, 10). On the trans side of the Golgi stacks, under the lectin-positivesaccules, two closely apposed unstained saccules were present (Figs. 5,6,12). These saccules did not participate in the formation of wells. They showed relatively few, small pores centrally and were continuous at their edges with small anastomotic tubules (Fig. 10). Lastly, on the trans face of the curved or cup-shaped stacks, two or three sacculotubular elements, showing a peeling off configuration, were usually unreactive with con A (Figs. 5, 7). The transmost element, however, which often assumed the configuration of a network of anastomotic tubules with budding extremities, occasionally showed reactivity to the lectin (Figs. 3,5,6). This reactivity was not uniform and frequently involved only the tubular and budding extremities of the element (Figs. 3, 5, 6). The large vesicles (120 nm) seen between the peeling off saccules or small (80 nm) vesicles at the edge of the saccules or in the trans region of the Golgi apparatus were usually unreactive with con A (Figs. 5-7, 10, 11). Some circular or vesicular profiles were reactive in the trans region (Fig. 7, lo), but they were likely to represent cross sections through the con A-positive trans tubular network or ER cisternae. Indeed, terminal tubular extensions of con A-positive ER 86 F. MALCHIODI ET AL. ~~ Fig. 5. Perikaryon of a type B1 cell in a section of tissue treated with con A-peroxidase complex and counterstained with lead citrate (as in the following Figures 6-13). Two stacks of Golgi saccules cut transversely are seen in this field. While the cis (osmiophilic) element in the Golgi stack on the left is stained with the lectin, the cis element of the Golgi stack on the right is unstained (vertical arrows). The underlying three saccules are strongly reactive, while the next trans saccules are unstained with the lectin (arrowheads) in both stacks. The trans sacculotubular element is partly stained with con A in the concave region of the Golgi stack on the left (asterisk). The vesicles associated with Golgi stacks are generally unreactive to con A. Lysosomes and mitochon- dria are negative while the rough ER cisternae are positive. ~32,500. Fig. 6. Perikaryon of a type B1 cell showing transverse (left) and oblique (lower right) sections through Golgi stacks. In the Golgi stack on the left, four elements including the cis (osmiophilic) elements are positively stained with con A. Other saccules on the trans side are negative. In the obliquely cut Golgi stack, all small (80 nm) vesicles in a well are negative to the lectin (arrowhead) as well as most small and large vesicles associated with the Golgi stacks (star). Some positively stained rough ER cisternae approach the Golgi apparatus on all sides. Mitochondria and lysosomes are unstained with con A. ~ 4 4 , 2 0 0 . CON A-BINDING SITES IN NEURON GOLGI APPARATUS cisternae often entered or penetrated the trans region of the Golgi stacks and were sometimes seen in close apposition to unstained trans elements (Figs. 10-12). Con A reactivity of the Golgi apparatus of type Ab A% and C cells In type Al, A2, and C cells, all the saccules or the sacculotubular elements of the Golgi stacks as well as the cis (osmiophilic) elements were reactive with con A (Figs. 1,4,8, 9). Occasionally, however, this cis element was found to be unreactive (Fig. 13).In these three cell types as for the type B1 and Bz neurons, the small (80 nm) vesicles seen in wells, at the edges of the Golgi stacks, or in the trans region as well as the larger (120 nm) vesicles seen between the lectin-positive peeling off saccules were all negative to the lectin (Figs. 8,9,13). Con A reactivity of Golgi-associated vesicles and multivesicular bodies In all cell types examined, clusters of small vesicles seen in close contact with or close to ER cisternae (Figs. 8,131or in the vicinity of wells (Fig. 6) were always unreactive with con A. Multivesicular bodies were present in all ganglion cells; and, while some of them were unstained with con A, others reacted positively with the lectin (Figs. 2, 4, 12, 13). In the latter, the matrix was reactive while the enclosed vesicles were not (Fig. 12). TPPase activity in the Golgi stacks of neurons In sections of spinal ganglia used to demonstrate TPPase activity, it was evident that in type B1 and B2 cells the reactive saccules were the two con A-negative saccules seen on the trans side of the Golgi stack. These two saccules were sandwiched between the con A-positive saccules and the trans sacculotubular elements showing a peeling-off configuration (Fig. 14). In type A cells, these two saccules were already shown to be TPPasepositive (Rambourg and Clermont, 1986)but, as mentioned above, these two saccules were also stained with con A. Fig. 7. Photograph showing a Golgi stack of a type Bz neuron. The cis (osmiophilic)element is not stained with con A (arrow) while the three underlying saccules are positively stained. The other trans saccules or “peeling off’ trans sacculotubular elements are unreactive. The small vesicles seen within the wells and most small and large vesicles seen in the trans region are unreactive with the lectin. A few large vesicular profiles are positive, however (arrowheads). X 35,750. 87 DISCUSSION Staining properties of sugars with con A Goldstein et al. (1965), using the technique of hapten inhibition, established that the binding sites of con A were most complementary to terminal a-D-mannopyranosyl residues but could also bind to a-D-glucopyranosy1 groups and their 2-acetamido-2-deoxy derivatives. Since unmodified hydroxyl groups at c3, c4, and c6 of the a-D-hexapyranosyl configuration were considered to be the minimal structural features required for the binding of con A to saccharides, internal 1-2-linked a-D-mannopyranosyl residues were also shown to interact with the lectin in conditions in which these residues were sterically available Qoyoshima et al., 1972; Goldstein et al., 1973; 1974). Finally, the presence of at least two a-D-mannopyranosyl residues with free hydroxyl groups at C3, C4, c6 was found to be a prerequisite to obtain a significant reaction with con A (Ogata et al., 1975; Kornfeld and Ferris, 1975; Krusius et al., 1976; Baenziger and Fiete, 19791, to retain glycopeptides in a con A sepharose column (Ogata et al., 1975), and to inhibit con A binding to erythrocytes (Kornfeld and Ferris, 1975). Various types of glycoproteins have been isolated from the central nervous system and characterized biochemically. Most of these glycoproteins contain asparagine-linked oligosaccharide side chains (Margolis et al., 1972; Krusius and Finne, 1978),although an important fraction of them possess alkali-labile side chains 0-glycosidically linked to the hydroxyl groups of serine or threonine (Margolis et al., 1972; Margolis and Margolis, 1973; Finne, 1975; Finne and Krusius, 1976). Con A binding to glycoproteins of the brain (Susz et al., 1973; Javaid et al., 1975;Zanetta et al., 1975; Krusius and Finne, 1978; McIntyre et al., 1979; Gurd and Fu, 1982) has been attributed to the presence of neutral mannose-rich asparagine-linked oligosaccharides (Javaid et al., 1975)in which up to eight mannose residues may be associated with a single N acetylglucosamine (Krusius et al., 1974; Gurd and Fu, 1982). A biochemical analysis of glycoproteins in sensory ganglia is not yet available. Nevertheless, the reactivity of ganglion cells to con A described in the present study suggests that the same types of oligosaccharides, particularly N-asparagine-linked carbohydrates, are also present in rat dorsal root ganglia. 88 F. MALCHIODI ET AL. CON A-BINDING SITES IN NEURON GOLGI APPARATUS Biosynthetic steps of glycoproteins side chains us. con A reactivity of cytoplasmic elements of ganglion cells In the rough endoplasmic reticulum Biochemical studies on the synthesis of asparagine-linked complex oligosaccharides (see review by Kornfeld and Kornfeld, 1985) have indicated that the processing of these macromolecules takes place in various compartments of the cell. Thus, oligosaccharide side chains containing terminal mannosyl and glycosyl residues are added to the protein core in a dense (heavy) membrane fraction which corresponds to the rough ER. Following the removal of terminal glycosyl residues by glycosidase I and II in the same fraction (Fig. 161, the oligosaccharide chains will be processed in lighter membrane fractions containing elements of the Golgi apparatus (see below). Since the ER oligosaccharides are rich in glycosyl and mannosyl terminal groups as well as in internal 2-0-substituted a-mannosyl residues, as expected and stated above, con A was found to bind consistently to the luminal side of rough ER cisternae in various cell types (Wood et al., 1974; Pinto da Silva et al., 1981; Grifiths et a]., 1982; Roth, 1983; Tartakoff and Vassali, 1983; Pavelka and Ellinger, 1985). This binding of con A to the rough ER cisternae was confirmed on the present study for all types of ganglion cells (Fig. 15). In the Golgi apparatus Labeling of the Golgi apparatus with con A has been reported to occur, usually, in one or two saccules located on the cis side of the Golgi stacks moth, 1983; Tartakoff and Vassali, 1983; Pavelka and Ellinger, 1985), although Griffiths et al. (1982) observed a binding of the lectin to all Golgi saccules in frozen thin sections of BHK cells using an antilectin antibody revealed by protein Agold complex. In ganglion cells, two main Fig. 8. Stacks of Golgi saccules of a type A1 cell. All elements of the Golgi stack react positively with con A except the small or large vesicles (arrowheads) associated with the stack. Clusters of small vesicles (arrow) seen between the con A-positive cisternae of the rough ER and stack of saccules (on the right) are equally negative. x42,500. Fig. 9. Stack of Golgi saccule of a type A2 cell. With the exception of the vesicles (arrowheads)associatedwith the stack, all saccular structures are positively stained with con A. A positive ER cisterna (arrow) is closely apposed to a trans saccule. ER, Cisternae of rough endoplasmic reticulum forming a Nissl body. ~ 4 2 , 8 0 0 . 89 types of reactions were observed with con A: in type Al, A2, and C cells, all Golgi saccules or sacculotubular elements were usually reactive to con A, whereas in type B1 and Bz cells possibly in type A3 cells, only the cis side of stacks stained positively (Fig. 15). In the following discussion the latter group of neurons will be considered first. Golgi apparatus of type B1, Bz and C cells The oligosaccharide side chains of glycoproteins elaborated in the ER are further processed in the Golgi apparatus (Fig. 16). Enzymes responsible for the trimming of CY 12-linked terminal mannosyl residues (Golgi mannosidases I and IT) and the addition of terminal P-N-acetylglucosaminyl residues (glucosaminyl transferases I and IT) have been isolated in a membrane fraction corresponding to the cis face elements of the Golgi apparatus of Chinese hamster ovary cells (Dunphy and Rothman, 1983) as well as in mouse lymphoma cells and macrophages (Goldberg and Kornfeld, 1983). In the latter cells and in rat liver cells, the enzymes Nacetylglucosamine-1-phosphotransferaseand N-acetylglucosamine-l-phosphodiester-a-Nacetyl-glucosaminidase (phosphodiesterglycosidase) involved in the phosphorylation of high mannose oligosaccharides of lysosomal enzymes were also localized in the same cellular fraction (Pohlman et al., 1982;Goldberg and Kornfeld 1983; Fig. 16). As a result, in this Golgi fraction there should be a loss of con A-reactive terminal and 2-0-substituted mannosyl residues and an almost concomitant addition of p 1-2-linkedN-acetylglucosaminyl residues to terminal mannosyl groups of the core, a structural configuration which is known to interact strongly with con A (Toyoshima et al. 1972; Kornfeld and Ferris, 1975; Debray et al., 1981; see Fig. 16). In type B cells, the erratic staining of the cis (osmiophilic)element contrasted with the strong and constant binding of con A to the three subjacent saccules (Fig. 15). Since the mannose-6-phosphate receptor for lysosomal enzymes has been detected immunocytochemically in the cis element of several cell types (Brown and Farquhar, 1984) and in Golgi saccules of hepatocytes (Geuze et al., 19841, the occasional lack of staining of the cis element with con A may be attributed partly to the phosphorylation at C6 of one or several terminal or preterminal CY 1,2-linked mannosyl residues in lysosomal oligosaccharide side chains and partly to the removal 90 F. MALCHIODI ET AL. 91 CON A-BINDING SITES IN NEURON GOLGI APPARATUS REACTIVITY OF E R AND GOLGI ELEMENTS TO CONCANAVALIN A Type Trans sacculo - tubular elements + + + + Rough E R C I S element ( O s + ) 3 Sublocent saccules ( Pose - ) 2 TPPase+ soccules A3 Type A,. A,.C B,. B., + ) or - + or + + + + - + + +- + - Fig. 15. The structure of a Golgi stack is shown schematically on the left. The con A reactivity of the various components of the stacks of saccules for the two groups of ganglion cells, i.e., types B1, Bz, and A3 and types Al, An, and C, is summarized in the table on the right. of terminal mannosyl groups by mannosidase I in nonlysosomal oligosaccharides(Fig. 16; Kornfeld and Kornfeld, 1985).In contrast in the subjacent con A-positive elements, which in rabbit hepatocytes have been shown to contain the N-acetylglucosaminetransferase I (Dunphy et al., 1984; Dunphy and Rothman, 1985),the removal of reactive mannosyl residues by mannosidase I and 11was likely to be counterbalanced by the addition at C2 of the terminal mannoses of N-acetylglucosaminyl residues (Fig. 16),thus explaining the strong con A staining of these three saccules (Fig. 15). The lack of lectin reactivity in B cells of the next two saccules of the stack (Fig. 15) was attributed to the linkage of another terminal sugar residue to 0-N-acetylglucosamine (Fig. 16). Radioautographic studies by Droz (1967), and the use of lectins binding specifically to galactose (Streit et al., 19851, revealed that terminal galactose residues are present in the Golgi apparatus of type B neurons in rat spinal and trigeminal ganglia. Furthermore, in ganglion cells as in other cell types (Tartakoff and Vassali, 1983),these two lectin-negative saccules show TPPase activity (Novikoff et al., 1971; Rambourg and Clermont, 1986) which is usually colocated within the same saccules as galactosyl transferase CRoth and Berger, 1982) and galactose residues (Roth et al., 1983; Berger, 1985; Pavelka and Ellinger, 1985). It is likely, therefore, as indicated in Figure 16, that galactose is the additional sugar on the oligosaccharide side chains that prevents the binding of con A to these two saccules. Lastly, in the TPPase-positive saccules and in the trans Fig. 10. Oblique section through a Golgi stack of a type B1 cell. The saccules on the cis side of the stack stain positively with con A (arrowhead)while the underlying saccules and the trans sacculotubular elements (asterisk) are unstained with the lectin. One positive ER cisterna penetrates the cis region of the stack and approaches the trans sacculotublar element. ~ 3 5 , 1 0 0 . ledin. ER cisternae are reactive, and a cisterna is seen in the trans region of stacks on the right (arrow). Multivesicular bodies are either positively or negatively stained with con A. x 16,800. Fig. 11. Face view of the lectin-positive elements of a type B1 cell. Face views of wells containing unstained small vesicles are indicated (arrowheads) in the lectinreactive saccules. A few positive large vesicular profiles (arrows) are visible amongst unstained vesicles seen in the trans region of the Golgi stack. Lectin-positiverough ER cisternae are also visible. ~ 3 5 , 1 0 0 . Fig. 12. Low magnification view of the perikaryon of a type B1 cell showing stacks of saccules with a cupshaped arrangement. On the cis side or convex aspect of the stacks, con A-stained elements are visible, while the saccules and sacculotubular elements on the trans side and concave aspect of the stacks are not stained with the Fig. 13. Small portion of the perikaryon of a type Al cell. In a stack of Golgi saccules at the upper left while the saccules are positively stained with con A, a cis element is negative (small arrowhead).Small vesicles in the proximity of the Golgi stacks are unstained. A cluster of small vesicles (asterisk) seen next to ER cisternae a t a distance from the Golgi stacks are also unstained with con A. A multivesicular body is also unstained. X35,750. Fig. 14. Section of a type B1 cell stained to show TPPase activity in a Golgi stack. Two saccules seen on the trans side of the Gulgi stack do show the lead precipitate indicative of phosphatase activity. The other saccular elements and the Golgi-associatedvesicles do not show TPPase activity. A similar reaction was observed in type A neurons. ~ 3 5 , 0 0 0 . 92 F. MALCHIODI ET AL. BlOSYNTHESlS OF N -ASPARAGIN€ -LINKED OLlGOSACCHARlDE SlDE CHAINS Rough ER CIS-Element ( os+ 1 3 Su bjocent Soccules (Pase-1 2 TPPase + Saccules and Trans Sacculotubular Elements OReactive with Con A OStrongly reactlve with Con A ::::Possibly reactive with Con A Fig. 16. The steps in the biosynthesis of N-asparagine- indicates the biosynthetic steps and suggested localizalinked oligosaccharide side chains and their suggested tion of complex oligosaccharide side chains in type B1, localization within the rough ER and the various com- Ba, and A, neurons. As indicated on the diagram, the ponents of the Golgi stacks (&r Kornfeld and Kornfeld, circles around sugar residues indicate their degree of 1985). The shaded area indicates the steps in oligosac- reactivity with con A. See text for discussion. G, glucose; charide processing and their intracellular location which M, mannose; GlcNAc, N-acetylglucosamine; Gal, galactake place in all neurons as suggested by the con A tose; SA, sialic acid; P, phosphorus; R, refers to the stem reactivity. The stippled area on the lower right indicates of the oIigosaccharide side chain composed of two molethe presumed localization of mannose-rich oligosacchar- cules of GlcNAc N-linked to asparagine of the protein ide side chains in type Al, AP, and C cells. The nonstip- chain. pled and nonshaded area of the rest of the diagram sacculotubular elements, sialic acid is added to the galactose residues through the action of a sialyl transferase (Bergeron et al., 1985; Roth et al., 1985). There is no explanation as yet for the slight con A reactivity observed in the sacculotubular element seen in the trans region of the Golgi stacks. It has been proposed by Rambourg and Clermont (1986) that trans ele- ments peeling off from the trans face of the Golgi stacks of type A neurons could eventually break up into vesicular or tubular elements. Such structural alterations must possibly induce changes in the steric configuration of the membrane-bound oligosaccharides and thereby facilitate the binding of mannocon A with the two ~~-2-O-substituted syl residues of the core (Fig. 16). CON A-BINDING SITES IN NEURON GOLGI APPARATUS Golgi apparatus of type Al, A2, and C cells In this group of ganglion cells, all saccules and sacculotubular elements reacted with con A (Fig. 15). A similar response to con A was also observed in the Golgi apparatus of Purkinje cells, i.e., in neurons of the central nervous system, a tissue in which mannose-rich glycoproteins have been isolated (Gombos et al., 1972; Javaid et al., 1975; Gurd and Fu, 1982). Since the uptake of galactose (Droz, 1967) and the demonstration of this sugar with lectins (Streit et al., 1985) was limited to the small type B neurons, the relative amount of galactose within the oligosaccharide side chains of glycoprotein should be reduced in the Golgi apparatus of type A neurons. As discussed above, preterminal galactose residues of complex oligosaccharides appear to be responsible for the reduction of con A binding to the TPPase-positive elements of the Golgi apparatus of type B cells; and thus the reactivity observed in the same elements of the Golgi apparatus of type A cells is likely due to the persistence of con A-reactive mannose residues and to the presence of mannose-rich glycoproteins in this class of cells (Figs. 15, 16). In general, therefore, our results suggest that the two groups of neurons in the spinal ganglion, i.e., types Al, A2, and C vs. types B1,B2, and A3, are possibly involved in the biosynthesis of different types of glycoproteins. 93 tent that their detection would be beyond the sensitivity of the con A-staining technique. Tartakoff and Vassali (1983) also observed that in myeloma cells a considerable number of vesicles located between the rough ER and Golgi cisternae were not stained with con A. They concluded that “although the permeabilization of such vesicles could not be directly evaluated, they may simply lack or contain limited amounts of reactive oligosaccharides.” In fact, techniques which detect material present in both endoplasmic reticulum and the cis (osmiophilic) element of the Golgi apparatus, e.g., osmication method (unpublished observations) and glucose-6-phosphatase method (Broadwell and Cataldo, 19831, fail to stain these vesicles. It has recently been postulated (Rambourg and Clermont, 1986)that the small (80 nm) vesicles seen in wells of the Golgi apparatus of type A ganglion cells must form by budding from the saccules seen on the cis side of the stacks to return to the endoplasmic reticulum. During the process of vesiculation, conformational andor biochemical changes might occur within the vesicular membranes which would then lose the characteristic properties of the saccular membrane from which they originate. Such a possiblity, which has to be confirmed by further studies, could also account for the lack of reactivity of these vesicles with con A. ACKNOWLEDGMENTS Absence of con A binding to vesicles associated with the Golgi stacks Small vesicles, also called intermediate or transfer vesicles, seen in the proximity of ER cisternae have usually been assumed to bud from the cisternae and carry proteins to the cis face of the Golgi apparatus (Palade, 1975; Farquhar and Palade, 1981; Goldfscher, 1982). In the present study, con A-stained vesicles, if any, were located in the trans region of the Golgi stacks in all neurons. In contrast, the small vesicles in the “wells” and the clusters of vesicles observed close to ER cisternae were never found to bind con A. Several conditions may explain this lack of reactivity. One possibility is that the steric configuration of the oligosaccharide chains in the restricted space of such small vesicles did not allow the binding of the lectin. Another possibility is that the trimming of terminal mannosyl residues at this level reduces the number of reactive sites to such a n ex- The work done at McGill University was supported by the Medical Research Council of Canada. The help of Dr. L. Hermo in the preparation of the manuscript is acknowledged. The drawings were prepared by Margo Oeltzschner, McGill University. LITERATURE CITED Baenziger, J.U., and D. Fiete 1979 Structural determinations of concanavalin A specificity for oligosaccharides. J. Biol. Chem., 254.2400-2407. Berger, E.G. 1985 Mini reviews: How Golgi associated glycosylation works. Cell Biology International Reports, 9:407-457. Bergeron, J.J.M., J. Paiement, M.N. Khan, and C.E. Smith 1985 Terminal glycosylation in rat hepatic Golgi fractions: Heterogeneous locations for sialic acid and galactose acceptors and their transferases. Biochim. Biophys. Acta, 821:393-403. Boutry, J.M., and A.B. Novikoff 1975 Cytochemical studies on Golgi apparatus, GERL and lysosomes in neurons of dorsal root ganglia in mice. Proc. Natl. Acad. Sci. U.S.A., 72508-512. Broadwell, R.D., and A.M. Cataldo 1983 The neuronal endoplasmic reticulum: Its cytochemistry and contri- 94 F. MALCHIODI ET AL. bution to the endomembrane system. I. Cell bodies and dendrites. J. Histochem. Cytochem., 31r1077-1088. Brown, W.J., and M.G. Farquhar 1984 The mannose-6phosphate receptor for lysosomal enzymes is concentrated in cis Golgi cisternae. Cell, 36t295-307. Clermont, Y., M.F. Lalli, and A. Rambourg 1981 Ultrastructural localization of nicotinamide adenine dinucleotide phosphatase (NADPase), thiamine pyrophosphatase (TPPase), and cytidine monophosphatase (CMPase) in the Golgi apparatus of early spermatids of the rat. Anat. Rec., 201:613-622. Debray, H., D. Decout, G. Strecker, G. Spik, and J. Montreuil 1981 Specificity of twelve lectins towards oligosaccharides and glycopeptides related t o N-glycosylproteins. Eur. J. Biochem., 117:41-55. Droz, B. 1967 L’appareil de Golgi comme site d’incorporation du g a l a c t ~ s e - ~dans H les neurones ganglionnaires spinaux chez le rat. J. Microscopie (Paris), 6:419424. Dunphy, W.G., and J.E. Rothman 1983 Compartmentation of asparagine-linked oligosaccharides processing in the Golgi apparatus. J. Cell Biol., 97:270-275. Dunphy, W.G., and J.E. Rothman 1985 Compartmental organization of the Golgi stack. Cell 4213-21. Dunphy, W.G., R. Brands, and J.E. Rothman 1984 Electron microscopic visualization of N-acetyglucosaminyl transferase I in central (medial) Golgi cisternae. J. Cell Biol., 99:229a. Farauhar. M.G., and G. Palade 1981 The Golpi apparatuH (complex)-(1954-19~4)from artefact to centerstage. J. Cell Biol., 91:775-1035. Finne, J. 1975 Structure of the 0-glycosidically linked carbohydrate units of rat brain glycoproteins. Biochim. Biophys. Acta, 412317-325. Finne, J., and T. Krusius 1976 0-glycosidic-carbohydrate units from glycoproteins of different tissues: Demonstration of a brain specific disaccharide, 01 galactosyl(1--t 3)-N-acetylgalactosamine. FEBS Lett., 66t94-97. Geuze, H.J., J.W. Slot, G.J.A.M. Strous, A. Hasilik, and K. Von Figura 1984 Ultrastructural localization of the mannose-6-phosphate receptor in rat liver. J. Cell Biol., 98:2047-2054. Goldberg, D.E., and S. Kornfeld 1983 Evidence for extensive subcellular organization of asparagine-linked oligosaccharide processing and lysosomal enzyme phosphorylation. J. Biol. Chem., 258t3159-3165. Goldfscher, S. 1982 The internal reticular apparatus of Camillo Golgi: A complex, heterogeneous organelle, enriched in acid, neutral and alkaline phosphatase and involved in glycosylation, secretion, membrane flow, lysosome formation and intracellular digestion. J. Histochem. Cytochem., 3Ot717-733. Goldstein, I.J., and C.E. Hayes 1978 The lectins: Carbohydrate binding proteins of plants and animals. Adv. Carbohydr. Chem. Biochem., 35:127-340. Goldstein, I.J., C.E. Hollerman, and E.E. Smith 1965 Protein carbohydrate interaction. II-Inhibition studies on the interaction of concanavalin A with polysaccharides. Biochemistry, 4:876-883. Goldstein, I.J., Ch.M. Reichert, A. Misaki, and P.A.J. Gorin 1973 An “extension” of the carbohydrate binding specificity of concanavalin A. Biochim. Biophys. Acta, 317t500-504. Goldstein, I.J., Ch.M. Reichert, and A. Misaki 1974 Interaction of concanavalin A with model substrates. Ann. N.Y. Acad. Sci., 234:283-296. Gombos, G., J.C. Hermetet, A. Reeber, J.P. Zanetta, and J. Treska-Ciesielsky 1972 The composition of glycopeptides derived from neural membranes which affect neurite growth in vitro. FEBS Lett., 24~247-250. Griffths, G., R. Brands, B. Burke, D. Louvard, and G. Warren 1982 Viral membrane proteins acquire galactose on trans Golgi cisternae during intracellular transport. J. Cell Biol., 95:781-792. Gurd, J.W., and S.C. Fu 1982 Concanavalin A receptors associated with rat brain synaptic junctions are high mannose type oligosaccharides. J. Neurochem., 39t719725. Hermo, L., Y. Clermont, and A. Rambourg 1980 Threedimensional architecture of the cortical region of the Golgi apparatus in rat spermatids. Am. J. Anat., 157:357-373. Javaid, I., H. Hof, and E.G. Brunngraber 1975 Preparation and properties of concanavalin A binding glycopeptides derived from rat brain glycoproteins. Biochim. Biophys. Acta, 404~74-82. Karnovsky, M.J. 1971 Use of ferrocyanide-reduced osmium tetroxide in electron microscopy. Proc. 11th Am. Soc. Cell Biol., New Orleans, Louisiana, Abstract 284, p. 146. Kornfeld, R., and C. Ferris 1975 Interaction of immunoglobulin glycopeptides with concanavalin A. J. Biol. Chem., 250:2614-2619. Kornfeld, R., and S. Kornfeld 1985 Assembly of asparagine-linked oligosaccharides. Annu. Rev. Biochem., 54:631-664. Krusius, T., and J. Finne 1978 Characterization of a novel sugar sequence from rat brain glycoproteins containing fucose and sialic acid. Eur. J. Biochem., 84t395403. Krusius, T., J. Finne, J. Kiirkainen, and J. Jiirnefelt 1974 Neutral and acidic glycopeptides in adult and developing rat brain. Biochim. Biophys. Acta, 365:80-92. Krusius, T, J. Finne, and H. Rauvala 1976 The structural basis of the different affinities of two types of acidic N-glycosidic glycopeptides for concanavalin Asepharose. FEBS Lett., 71:117-120. Lalli, M.F. 1983Ultrastructural localization of phosphatase and aryl sulfatase activities in the Golgi apparatus and lysosomes of Sertoli cells in the rat. Anat. Rec., 2053105A (abstr.). Margolis, R.K., and R.U. Margolis 1973 Alkali labile oligosaccharides of brain glycoproteins. Biochim. Biophys. Acta, 304t421-429. Margolis, R.U., R.K. Margolis, and D.M. Atherton 1972 Carbohydrate peptide linkages in glycoproteins and mucopolysaccharides from brain. J. Neurochem., 19:2317-2324. McIntvre, L.J., R.H. Quarles, and R.O. Bradv 1979 Lectin binding proteins i n central nervous systkm myelin. Biochem. J., 183:205-212. Novikoff, A.B. 1967 Enzyme localization and ultrastructure of neurons. In: The Neuron. H. Hyden, ed. American Elsevier, New York, pp. 255-318. Novikoff, A.B., and S. Goldfscher 1961 Nucleoside diphosphatase activity in the Golgi apparatus and its usefulness for cytological studies. Proc. Natl. Acad. Sci. U S A . , 47:802-810. Novikoff, P.M., A.B. Novikoff, N. Quintana, and J.J. Hauw 1971 Golgi apparatus, GERL and lysosomes of neurons in rat dorsal root ganglia, studies by thick section and thin section cytochemistry. J. Cell Biol., 502359-886. Ogata, S., T. Muramatsu, and A. Kobata 1975 Fractionation of glycopeptides by affinity column chromatog raphy on concanavalin A, sepharose. J. Biochem. 7Xe37-696 . - . .- . - - -. Palade, G. 1975 Ultracellular aspects of the process of protein secretion. Science, 189: 347-357. Pavelka, M., and A. Ellinger 1985 Localization of bind- CON A-BINDING SITES IN NEURON GOLGI APPARATUS ing sites for concanavalin A, Ricinus communis I and HeZix pomatia lectin in the Golgi apparatus of rat small intestinal absorptive cells. J. Histochem. Cytochem., 33:905-914. Pinto da Silva, P., M.R. Torrisi, and B. Kachar 1981 Freeze fracture cytochemistry: Localization of wheat germ agglutinin and concanavalin A binding sites on freeze-fractured pancreatic cells. J. Cell. Biol., 91:361372. Pohlmann, R., A. Waheed, A. Hasilik, and K. Von Figura 1982 Synthesis of phosphorylated-recognition marker in lysosomal enzymes is located in the cis-part of the Golgi apparatus. J. Biol. Chem., 257:5323-5325. Rambourg, A., and Y. Clermont 1986 Tridimensional structure of the Golgi apparatus in type A ganglion cells of the rat. Am. J. Anat., 176:393-409. Rambourg, A., Y.Clermont, and A. Beaudet 1983 Ultrastructural features of six types of neurons in rat dorsal root ganglia. J. Neurocytol., 1247-66. Rambourg, A., Y. Clermont, and L. Hermo 1979 Threedimensional architecture of the Golgi apparatus in Sertoli cells of the rat. Am. J. Anat., 154:455-476. Rambourg, A., Y. Clermont, and L. Hermo 1981 Threedimensional structure of the Golgi apparatus. In: Basic Mechanisms of Cellular Secretion. A.A. Hand and C. Oliver, eds. Methods in Cell Biology, 23:155-166. Rambourg A., Y. Clermont, and A. Marraud 1974 Threedimensional structure of the osmium impregnated Golgi apparatus as seen in the high voltage electron microscope. Am. J. Anat., 140:2746. Rambourg, A., D. Segretain, and Y. Clermont 1984 Tridimensional architecture of the Golgi apparatus in the atrial muscle cell of the rat. Am. J. Anat., 170:163179. Roth, J. 1983 Application of lectin-gold complexes for electron microscopic Iocalization of glycoconjugates on thin sections. J. Histochem. Cytochem., 31t987-999. 95 Roth, J., and E.G. Berger 1982 Immunocytochemical localization of galactosyl transferase in Hela cells: Codistribution with thiamine pyrophosphatase in trans Golgi cisternae. J. Cell Biol., 93:223-229. Roth, J., M.J. Lentze, and E.J. Berger 1983 Localization of galactosyl transferase (GT) and galactose (gal) residues in human duodenal mucosa, demonstration of ectogalactosyl transferase. J. Histochem. Cytochem., 31:1072 (abstr.). Roth, J., D.J. Taatjes, J.M. Lucocq, J. Weinstein, and J.C. Paulson 1985 Demonstration of a n extensive transtubular network continuous with the Golgi apparatus stack that may function in glycosylation. Cell, 43t287295. Streit, W.J., B.A. Schulte, J.D. Balentine, and S.S. Spicer 1985 Histochemical localization of galactose containing glycoconjugates in sensory neurons and their processes in the central and peripheral nervous system of the rat. J. Histochem. Cytochem., 33:1042-1052. Susz, J.P., H.I., Hof, and E.G. Brunngraber 1973 Isolation of concanavalin A-binding glycoproteins from rat brain. FEBS Lett., 32289-291. Tartakoff, A.M., and P. Vassali 1983 Lectin binding sites as markers of Golgi subcompartments Proximal to distal maturation of oligosaccharides. J. Cell Biol., 97:1243-1248. Tovoshima, S., M. Fukuda, and T. Osawa 1972 Chemical nature of the receptor site for various phytomitogens. Biochemistry, 11:40004005. Wood, J.G., B.J. McLaughlin, and R.P. Barber 1974 The visualization of concanavalin A binding sites in Purkinje cell somata and dendrites of rat cerebellum. J. Cell Biol., 63:541-549. Zanetta, J.P., I.G. Morgan, and G. Gombos 1975 Synaptosomal plasma membrane glycoproteins: Fractionation by affinity chromatography on concanavalin A. Brain Res., 83:337-348.