Differential distribution of salivary agglutinin and amylase in the Golgi apparatus and secretory granules of human salivary gland acinar cells.код для вставкиСкачать
THE ANATOMICAL RECORD 230:307-318 (1991) Differential Distribution of Salivary Agglutinin and Amylase in the Golgi Apparatus and Secretory Granules of Human Salivary Gland Acinar Cells KUNIO TAKANO, MEREDITH BOGERT, DANIEL MALAMUD, EDWARD LALLY, AND ARTHUR R. HAND Department of Oral Histology, Nagasaki University School of Dentistry, Nagasaki 852, Japan (K.T.);Center for Oral Health Research, University of Pennsylvania School of Dental Medicine, Philadelphia, Pennsylvania 19104 (M.B., D.M., E.L.); Clinical Investigations and Patient Care Branch, National Institute of Dental Research, NIH, Bethesda, Maryland 20892 (A.R.H.) ABSTRACT The secretory granules of salivary glands often display complex internal substructures, yet little is known of the molecular organization of their contents or the mechanisms involved in packaging of the secretory proteins. We used post-embedding immunogold labeling with antibodies to two secretory proteins, agglutinin and a-amylase, to determine their distribution in the Golgi apparatus and secretory granules of the human submandibular gland acinar cells. With monoclonal antibodies specific for carbohydrate epitopes of the agglutinin, reactivity was found in the trans Golgi saccules, trans Golgi network, and immature and mature secretory granules. In the granules, labeling was seen in regions of low and medium electron density, but not in the dense cores. Reactivity seen on the apical and basolateral membranes of acinar and duct cells was attributed to a shared epitope on a membrane glycoprotein. Labeling with a polyclonal antibody to amylase was found in the Golgi saccules, immature and mature secretory granules, but not in the trans Golgi network. In the granules, amylase was present in the dense cores and in areas of medium density, but not in the regions of low density. These results indicate that these two proteins are distributed differently within the secretory granules, and suggest that they follow separate pathways between the Golgi apparatus and forming secretory granules. Small vesicles and tubular structures that labeled only with the antibodies to the agglutinin were observed on both faces of the Golgi apparatus and in the vicinity of the cell membrane. These structures may represent constitutive secretion vesicles involved in transport of the putative membrane glycoprotein to the cell membrane. Abundant evidence has recently been obtained demonstrating the differential regulation of protein transport and secretion in eukaryotic cells. Key regulatory points in this process are a t the level of ER-Golgi apparatus transfer (Rose and Doms, 1988), transport and post-translational modifications in the Golgi apparatus (Kornfeld and Kornfeld, 1985; Farquhar, 1985; Pfeffer and Rothman, 1987; Rothman and Orci, 1990), and sorting and packaging into vesicles or granules destined for constitutive or regulated discharge (Kelly, 1985; Burgess and Kelly, 1987). In spite of considerable recent attention (Chung et al., 1989; Stoller and Shields, 1989), relatively little is known a s yet of the mechanisms involved in the sorting and packaging of proteins into the characteristic granules found in regulated secretory cells (Burgess and Kelly, 1987). The acinar cells of salivary glands have served a s useful models for studying various aspects of the secretory process. Unlike other exocrine glands, salivary acinar cells show a remarkable diversity of morphological appearances among glands and among species (Young and van Lennep, 1978). In particular, the struc0 1991 WILEY-LISS, INC ture of secretory granules is highly variable. They range in appearance from electron-lucent to highly electron-dense, and their content may be homogeneous or display regions of variable density, often arranged in complex patterns (e.g., Tandler and MacCallum, 1972). The granules of acinar cells of human parotid and submandibular glands typically contain dense cores, crescent-shaped structures, andlor strands of dense material in a skein-like pattern (Tandler and Erlandson, 1972; Riva and Riva-Testa, 1973; Riva et al., 1974). Received April 16, 1990; accepted December 5, 1990. Dr. Arthur R. Hand’s present address is Department of Pediatric Dentistry, University of Connecticut Health Center, Farmington, CT 06030. Abbreviations used: BSA, bovine serum albumin; ELISA, enzymelinked immunosorbent assay; GERL, Golgi-endoplasmic reticulumlysosome; IgG, immunoglobulin G; IgM, immunoglobulin M; mAb, monoclonal antibody; Mr, relative mobility; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; SDS, sodium dodecyl sulfate; TBS, Tris-buffered saline. 308 K. TAKANO ET AL. A B C D E F k Da - 1 1 0 - a 4 - 7 7 - 4 7 - 3 3 - 2 4 Fig. 1. Analysis of saliva and agglutinin by electrophoresis and Western blotting. Samples were separated by SDS-PAGE (10% acrylamide) and either silver stained (A,B) or blotted onto nitrocellulose (C-F). The relative mobility of molecular weight markers is shown on the right. A, parotid saliva, silver stain; B, purified agglutinin, silver stain; C , Western blot of parotid saliva reacted with mAb 303; D, Western blot of saliva reacted with polyclonal antibody to denatured agglutinin; E, Western blot of saliva reacted with polyclonal antibody to amylase; F, Western blot of purified agglutinin reacted with mAb 303. The appearance of the granules is clearly dependent upon their composition, but the biochemical basis for their complex structural organization is currently poorly understood. The acinar secretory granules are the source of most of the more than 50 proteins present in human saliva, which include digestive enzymes, antimicrobial factors, and proteins which adhere to and coat the surfaces of the oral cavity (Ellison, 1979). We have selected two of these proteins, a-amylase and agglutinin, to investigate specific aspects of protein transport and packaging in human salivary acinar cells. a-Amylase represents a family of proteins, some of which are glycosylated, t h a t are secreted by the salivary glands of most species and cleave starch to limit dextrins (Merritt and Karn, 1977). The agglutinin is a high molecular weight acidic sialoglycoprotein that binds to and aggregates specific oral streptococci (Hay et al., 1971; A GS ICS IG L M SD SG A bbreuiations acinar cells Golgi saccules lateral cell surface immature secretory granule lumen mucous cells striated duct secretory granule Kashket and Donaldson, 1972; Malamud and Golub, 1981).This agglutinin has been isolated from both parotid and submandibular saliva of humans, and shown to be distinct from secretory IgA and mucins (Williams and Gibbons, 1975; Malamud, 1985). The agglutinin has been purified by both column chromatography (Demuth et al., 1988) and bacterial affinity (Rundegren and Arnold, 1987). The isolated glycoprotein is approximately 40% carbohydrate and 60% protein (Ericson and Rundegren, 1983). Glycoproteins with similar functions have been described in the saliva of rats (Brack and Reynolds, 1987). We have used mouse polyclonal and monoclonal antibodies (mAbs) directed against the agglutinin, and a rabbit polyclonal antibody to human salivary amylase, to localize these secretory proteins by post-embedding immunogold labeling procedures. Our results demonstrate t h a t agglutinin and amylase are differentially distributed within the secretory granules, and suggest that distinct Golgi subcompartments may be involved in the transport and sorting of these two proteins. MATERIALS AND METHODS Agglutinin was purified from human parotid saliva by column chromatography on DEAE-Sephadex and Sepharose 4B a s previously described (Demuth et al., 1988) and shown to migrate a s a single band on SDS- 309 SALIVARY AGGLUTININ AND AMYLASE DISTRIBUTION TABLE 1. Characterization of anti-agglutinin monoclonal antibodies mAb Isotype no, = 1 td i n h i b i t 143 116 303 IeG, IgG; IgM 0% at 20 mM 80% at 0.5 mM 80% at 0 . 5 m M Titer (Western blot) 1:lOO 1:1,000 1:500 prepared from the spleens of these mice, and fused (polyethylene glycol) with SP 2/0 Ag14 myeloma cells. Approximately 350 clones were screened using 1)a bioassay in which culture media was mixed with saliva and tested for inhibition of bacterial aggregation (Malamud et al., 1983),and 2) a n ELISA using PVC plates coated with purified agglutinin and exposed to culture media and then a second antibody conjugated to peroxidase. Of the original 350 clones, 62 were positive in both assays. Eight of these have been recloned three times and seven of these have been characterized with respect to isotype, titer €or inhibition and ELISA, and effects of sodium periodate and neuraminidase treatments. Additionally, a polyclonal antibody to denatured agglutinin eluted from SDS electrophoretic gels was prepared in mice. Light Microscopic lmmunohistochemistry Parafin-embedded formalin or Bouin’s fixed tissues from the Surgical Pathology file of either The Hospital of the University of Pennsylvania or the American Oncologic Hospital were used. Immunoreactivity was detected using the preformed avidin-biotin-peroxidase complex procedure (Vectastain ABC Kit, Vector Laboratories). Sections were incubated for 15 min in 0.3% H,Oz in methanol to block endogenous peroxidases. Following a wash in PBS, slides were incubated sequentially in normal horse serum for 20 min, with appropriately diluted mAbs for 30 min at room temperature, biotinylated anti-mouse immunoglobulin (1:200) and avidin-biotin-peroxidase complex. Sections were incubated with 0.05% diaminobenzidine (Sigma Chemical Co.) with 0.01% H20, for 5 min and counterstained with hematoxylin. Mouse myeloma-P3 supernatant was used as a negative control. Tissue Preparation for Electron Microscopy Fig. 2. Light microscopic immunolocalization of agglutinin in human salivary glands with mAb 303. A-C: Submandibular gland. D: parotid gland. Serous acinar (A) and demilune cells (arrowheads in A and B) are labeled, but no reaction is seen in mucous cells (M). The luminal and basal portions of the striated ducts (SD) are also stained. x 435. PAGE with a subunit molecular weight of approximately 400,000 Daltons (Fig. 1). Preparation of Antibodies to Salivary Agglutinin BALBic mice were immunized with purified agglutinin in Freund’s adjuvant. A single cell suspension was Normal human salivary gland tissues were obtained from surgical specimens (Brown et al., 1985).Promptly after surgical removal, the tissue specimens were cut into small pieces (1mm3) in cold fixative solution. Fixatives used were 3% (para)formaldehyde plus 0.1% glutaraldehyde, 1% glutaraldehyde, and 4% (para)formaldehyde in 0.05 or 0.1 M phosphate or cacodylate buffer, pH 7.4. Total fixation time was approximately 1 hour for glutaraldehyde-containing fixatives, and 4 to 6 hours for formaldehyde-containing fixatives. The tissue samples were rinsed in 0.1 M phosphate or cacodylate buffer, pH 7.4, with 5-7% sucrose, then dehydrated and embedded in LR White resin (London Resin Co.) or EMbed 812 (Electron Microscopy Sciences) without osmium postfixation. Some tissue pieces from samples fixed in 1%glutaraldehyde were postfixed in 1% OsO, and processed for conventional morphologic evaluation. Additionally, some fixed and embedded tissue samples were generously provided by Professor A. Riva, University of Cagliari, Italy. Semithin sections 310 K. TAKANO ET AL. Fig. 3. Acinar secretory granules of human submandibular gland, immunolabeled with mAb 303 and 15 nm protein A-gold. In A, the agglutinin is restricted to the peripheral electron-lucent region of the granules. Only a few particles are present over the dense cores and the areas of medium density. x 28,000. In B, labeling of the medium density areas of the granules with mAb 303 is heavy, but the electrondense cores are unlabeled. Labeling is also seen on the cell membrane a t the luminal (L) and lateral (ICS) surface. x 21,000. Fig. 4. Submandibular acinar cell granules immunolabeled with polyclonal anti-agglutinin antibody and 15 nm protein A-gold. The labeling pattern is identical to that seen with mAb 303: gold particles are concentrated over regions of low and medium electron density in the secretory granules, and are present on the luminal (L) and lateral (ICS) cell surfaces. x 30,000. were stained with toluidine blue and appropriate areas were selected for thin sectioning. Ultrathin sections were collected on Formvar-coated or bare nickel grids for immunostaining, or on copper grids for routine observation. Electron Microscopic lmmunocytochemical Procedures The thin sections were immunolabeled using the protein A-gold method, modified from Roth et al. (1978) and Bendayan et al. (1980). The grids were floated se- SALIVARY AGGLUTININ AND AMYLASE DISTRIBUTION 311 lase was visualized with 5 nm gold particles and agglutinin with 15 nm particles. Immunocytochemical controls included 1)incubation of sections with primary antibody which had been preabsorbed by incubation with a n excess of the antigen overnight a t 4°C; 2) incubation of the sections with non-immune immunoglobulins or a n irrelevant mAb of the IgG or IgM class instead of the primary antibody; and 3) omission of the primary antibody andior the secondary antibody from the staining sequence. RESULTS Characterizationof Antibodies The characterization of the three mAbs utilized for immunolabeling is shown in Figure 1 and Table 1. Western blots of human parotid saliva and purified agglutinin (Fig. 1)demonstrated staining of a single band migrating with a n approximate Mr of 400,000 Daltons. Some diffuse staining of smaller sized components was also noted. Monoclonal antibodies 303 and 116 appeared to recognize carbohydrate epitopes based on the sensitivity of the ELISA to inhibition by sodium periodate (Table l).None of the mAbs required sialic acid for recognition, as neuraminidase treatment did not decrease reactivity (data not shown). In addition, isotyping revealed that mAbs 143 and 116 are IgGs, whereas mAb 303 is a n IgM. Western blots of saliva incubated with the polyclonal antibody to the denatured agglutinin revealed a single reactive band (Fig. 1). Fig. 5. Submandibular acinar granules immunolabeled with antiTo test whether any of the three mAbs recognize a amylase antibody and 15 nm protein A-gold. Amylase is present in common epitope, the technique of Friguet et al. (1983) the dense core and the areas of medium density, whereas the regions was used. Each antibody was titered in a n ELISA using of low electron density are unlabeled. x 22,000. a fixed concentration of antigen (agglutinin). Selecting the antibody concentration giving one-half maximum optical density, mAbs were then tested in combination. quentially on small drops (25-30 PI) of PBS for 5 min; These experiments indicated that mAbs 303 and 116 1% BSA in TBS for 30 min; and either the polyclonal recognized the same epitope. In an attempt to identify the carbohydrate epitope anti-agglutinin or one of the monoclonal anti-agglutinins diluted in 0.02 M Tris-HC1, pH 7.4, with 0.5 M recognized by mAbs 303 and 116, a series of known NaCl and 0.1% Tween 20 for 60 min at room temper- glycolipids was separated on TLC plates and reacted ature. In some cases incubation with the mAb was done with the mAbs and also with a panel of mAbs with overnight at 4°C. After rinsing with the Tris-NaC1- known specificities (Magnani et al., 1982). Results of Tween buffer, the grids were incubated with the appro- these studies indicated that mAbs 303 and 116 recogpriate unlabeled secondary antibody (affinity-purified nized the Y epitope (Blaszczyk-Thurin et al., 1987). rabbit anti-mouse IgG or goat anti-mouse IgM, Orga- This was confirmed by demonstrating that these mAbs non Teknika-Cappel) diluted in TBS for 60 min; rinsed agglunitated latex beads containing Y-BSA conjugates in TBS; and incubated with 10 or 15 nm diameter pro- but not beads with X, Lewis a, or Lewis b conjugates. Figure 1also shows the reaction of anti-amylase antein A-gold complex (Janssen Life Science Products) diluted 1 : l O in TBS for 30 min. The grids were rinsed tibodies (Sigma) on Western blots of human parotid with TBS and finally with distilled water before being saliva. A double band migrating with a n Mr of approxdried and stained with uranyl acetate and lead citrate. imately 60,000 Daltons was labeled. Presumably, the Immunolabeling for a-amylase was performed in a two bands correspond to the glycosylated and non-glysimilar manner, using a polyclonal antibody to human cosylated forms of human salivary amylase (Keller et salivary amylase (Sigma) diluted in TBS, followed by al., 1971). rinsing with TBS and visualization with protein ALight Microscopic lmmunohistochemicalLocalization of gold. Salivary Agglutinin Double labeling for a-amylase and the agglutinin Sections of five human submandibular glands were was performed by the two-sided method of Bendayan (1982). One side of the section was incubated for 60 min incubated with varying concentrations of the three with anti-amylase antibodies, and the bound IgG was mAbs described above. In all cases, staining was seen localized with 15 nm protein A-gold complex. The grid in serous acinar (Fig. 2A-C) and demilune cells (Fig. was rinsed and dried, and the other side was incubated 2A,B), intercalated duct cells, and striated duct cells with mAb 303, goat anti-mouse IgM, and finally 5 nm (Fig. 2C). In the striated ducts, the labeling appeared protein A-gold (Janssen). In some experiments, amy- strongest at the luminal surface and in the infranu- 312 K. TAKANO ET AL. Fig. 6. Double labeling of acinar secretory granules with mAb 303 and anti-amylase antibody. Agglutinin (5 nm gold particles) is present in areas of low electron density, and amylase (15 nm gold particles) is present mainly in the dense regions. Some reactivity for amylase is also seen along with agglutinin in areas of low density (arrowheads). ~46,000. clear region. No labeling of mucous acinar cells occurred. In the parotid gland (Fig. 2D), which is a pure serous gland, a similar distribution was observed for the agglutinin in the acini and ducts. al., 1974). In the present study, two or three regions of differing electron density were observed in the granules of most cells. Labeling for the agglutinin with either mAb 303 (Fig. 3) or the polyclonal anti-agglutinin antibody (Fig. 4) was consistently observed in the regions of lowest electron density, which usually were located peripherally in the granules. The dense corelike structures, often positioned eccentrically, were unlabeled, whereas regions of medium density showed a variable labeling intensity. With anti-amylase antibody, the opposite labeling pattern was observed (Fig. 5). The dense cores labeled most intensely, whereas regions of lowest density were unlabeled. Again, areas of moderate density showed a variable labeling intensity. Sections labeled with both anti-amylase antibody followed by 15 nm protein A-gold, and mAb 303 followed by 5 nm protein A-gold, to simultaneously localize both proteins (Fig. 61, showed a labeling pattern similar to that seen in singly labeled preparations. Low density regions were labeled almost exclusively with 5 nm gold particles, and the dense cores were labeled only with 15 nm gold particles, indicating the presence of agglutinin and amylase, respectively, in these two regions. In some granules, both small and large gold particles, in variable proportions, were present over low to moderately dense regions (arrowheads, Fig. 6), showing a mixture of the two proteins in these areas. An identical lmmunogold Labeling of Submandibular Gland Acinar Cells In acinar cells of the human submandibular gland, immunolabeling for the agglutinin and for amylase was present in the Golgi region and over the secretory granules. Additionally, some of the anti-agglutinin mAbs also labeled the apical and basolateral membranes. Labeling with the mAbs was most intense with mAb 303; labeling was less intense with mAb 116, but the distribution was similar. With mAb 143, which recognizes a polypeptide epitope, labeling of the granules and Golgi apparatus was weak and no reactivity was observed on the cell membranes. The present report focuses on the labeling of the Golgi apparatus and secretory granules of the acinar cells; a complete description of the labeling patterns of the various cell types present in the human salivary glands obtained with the three mAbs will be communicated separately. Secretory granules The secretory granules of human submandibular acinar cells typically contain internal substructures of variable morphology and electron density (Tandler and Erlandson, 1972; Riva and Riva-Testa, 1973; Riva et SALIVARY AGGLUTININ AND AMYLASE DISTRIBUTION 313 Fig. 7. Double labeling of Golgi apparatus with mAb 303 (5 nm particles) and anti-amylase antibody (15 nm particles). Large gold particles demonstrating amylase are present over the Golgi saccules (GS) and the dense core of a small forming granule (arrow).Only a few small particles representing agglutinin are seen over the electronlucent area of the forming granule and in the trans saccules (arrowheads). x 63,000. Fig. 8. Double labeling of Golgi apparatus. Agglutinin reactivity (5 nm particles) is seen in small tubules (arrowheads) and immature secretory granules (IG) a t the trans face, in regions of low density in the mature secretory granule (SG), and in a large tubular structure (arrow) at the left of the micrograph. Amylase reactivity (15 nm particles) is low but gold particles are seen over the Golgi saccules (GS), immature granules and the dense regions of the mature granule. x 52,000. distribution of the two proteins was seen in preparations in which the sizes of the gold particles were reversed (data not shown). with mAb 303 was restricted to structures in the trans Golgi region. These included immature secretory granules (Figs. 7-9); vesicles, tubules, and cisternae with morphological characteristics of the trans Golgi network (Griffths and Simons, 1986) or GERL (Novikoff, 1976; Hand and Oliver, 1977a,b) (Figs. 8-11); and occasionally the trans Golgi saccule (Fig. 7). The remainder of the Golgi saccules were unlabeled. Occasional Golgi apparatus The distribution of the two secretory proteins in the Golgi region was most clearly demonstrated in doublylabeled sections. Immunolabeling for the agglutinin 314 K. TAKANO ET AL. Fig. 9. Double labeling of Golgi apparatus. Amylase (15 nm particles) is seen in the Golgi saccules (GS) and the dense core of one forming granule. Agglutinin reactivity (5 nm particles) is seen in the same forming granule, in elements of the trans Golgi network (GERL) (arrow), and in a small tubular or vesicular structure a t the cis face (arrowhead). x 53,000. tubular and/or small spherical bodies which labeled only with mAb 303 were also present. These structures were often seen in the Golgi region (Fig. 8), where they were associated with both faces of the Golgi apparatus, and also in peripheral regions of the cytoplasm, sometimes near the basolateral membrane (Fig. 12A,B). Similar vesicular structures near the basolateral membrane could be identified in unincubated sections of specimens postfixed in OsO, (Fig. 12C,D). In the absence of labeling with mAb 303, however, it was not possible to determine if these vesicles were the same structures as those seen in immunolabeled sections. Immunolabeling for amylase, in contrast, was seen over the Golgi saccules (Figs. 7-10), but only rarely over structures in the trans Golgi region labeled with mAb 303. Immature secretory granules, especially those showing condensations of dense material (Figs. 7-9), were labeled with the antibody to amylase. When these forming granules also were labeled with mAb 303, small gold particles indicating the presence of the agglutinin were located over the electron-lucent regions, whereas large particles representing amylase were present over the dense core (Figs. 7, 8). lmmunocytochemical controls Omission of the primary antibodies from the staining sequence, or substitution of non-immune immunoglobulins for the primary antibodies, essentially eliminated the binding of gold to the sections for both agglutinin and amylase. Preabsorption of the antibodies with their respective antigens also greatly reduced labeling of the secretory granules. DISCUSSION Amylase and agglutinin, two secretory proteins of human salivary acinar cells, are both packaged in the same secretory granules. The double labeling results clearly demonstrate, however, that within individual secretory granules, these two content proteins are differentially distributed. Amylase is concentrated in the denser portions of the granule content, whereas agglu- tinin is found in the less dense regions. These findings are consistent with the results of previous immunolabeling studies of salivary glands of various species showing preferential localization of secretory proteins in specific regions of the granules. For example, amylase, lysozyme and proline-rich proteins are concentrated in the dense regions of the granules of serous cells (Ichikawa et al., 1989; Machino et al., 1986; Kousvelari et al., 1982), neonatal and adult secretory proteins show a differential distribution in certain granules of the postnatal rat submandibular gland (Moreira et al., 19911, and mucin is located in the electron-lucent regions of adult rat submandibular mucous cells (Moreira et al., 1989). Differential localization of secretory proteins has also been observed in granules of some endocrine cells (Ravazzola and Orci, 1980; Fumagalli and Zanini, 1985; Bendayan, 1989). Although the distribution of amylase and agglutinin was correlated with the electron density of the granule content, some variability was noted. Thus, both proteins were often found in regions of moderate density. Also, the size, shape, and relative electron density of the different intragranular regions was found to vary among granules. These factors may be responsible for the difference in our results and those of Machino et al. (1986), who reported a relatively homogeneous distribution of amylase in human parotid and submandibular acinar cell granules. In contrast to the distribution of the A, B, and H blood group antigens in human salivary glands (Nakajima et al., 1987), amylase and agglutinin reactivities were observed in the Golgi apparatus and immature secretory granules, as well as the mature granules. Several underlying questions remain unanswered. For example, why do secretory granules of some cells, such as rat parotid (Ball e t al., 1988; Hand and Ball, 1988) and pancreatic acinar cells (Bendayan et al., 1980), exhibit a relatively homogeneous distribution of several secretory proteins, whereas those of the human salivary glands display such a heterogeneous distribution of their content? What is the molecular basis for the segregation of different proteins within the granule? Is it due to the specific physical-chemical properties of individual content proteins, resulting in self- or coaggregation (Burgess and Kelly, 1987; Stoller and Shields, 1989),or to other components involved in condensation and packaging, such as sulfated macromolecules (Palade, 1975), calcium (Clemente and Meldolesi, 19751, specific binding proteins (Chung et al., 1989), or perhaps a n internal reticular component involved in organization of the granule content (Ermak and Rothman, 1978)? I t is conceivable that the segregation of the granule contents is a n artifact of chemical fixation, and t h a t a homogeneous distribution of secretory proteins exists in vivo. Several observations argue against such a simple explanation, however. The structure of the granules appears relatively similar in several studies of human salivary glands (Tandler and Erlandson, 1972; Riva and Riva-Testa, 1973; Riva et al., 1974), despite varied fixative formulations and fixation protocols. It seems unlikely that the extremely elaborate substructural organization found in granules of many species (Tandler and MacCallum, 1972) could be a result of fixation-induced molecular rearrangements, without SALIVARY AGGLUTININ AND AMYLASE DISTRIBUTION Figs. 10, 11 : Double labeling of Golgi apparatus. Amylase (15 nm particles) is present in the Golgi saccules (GS) and secretory granules (SG). In Figure 10, agglutinin reactivity (5 nm particles) is seen in several tubular elements of the trans Golgi network (GERL) (arrowheads). x 69,000. In Figure 11, an extensive tubular network a t the trans face contains agglutinin reactivity (5 nm particles). Agglutinin is also seen in the secretory granules. x 90,000. 315 316 K. TAKANO ET AL. Fig. 12. A , B Double labeling for agglutinin and amylase. Small vesicles (arrowheads) labeled only for agglutinin (5 nm particles) are present near the lateral membrane. In A, agglutinin reactivity is also seen on the cell surf-ace (arrow). C , D Small vesicles (arrowheads) are seen near the lateral membrane of acinar cells in Os0,-postfixed tissue. A, x 63,000; B, x 67,000; C, x 86,000; D, x 58,000. the presence of a variety of intermediate forms where stabilization occurred before the rearrangements were completed. Finally, secretory granules exhibiting complex intragranular substructures andlor regions of variable density, a s well as a heterogeneous distribution of amylase have been described in rapidly frozen, freeze-substituted gerbil sublingual and parotid glands (Ichikawa and Ichikawa, 1987; Ichikawa e t al., 1989). It is possible that the differential distribution of amylase and agglutinin within the granules results from the apparently separate pathways these two proteins travel in reaching the forming secretory granules. Agglutinin appears to reach the granules through the trans Golgi network (GERL); agglutinin reactivity was present in the trans saccules and in elements of the trans Golgi network. Although agglutinin reactivity was not seen in the cis saccules, this is consistent with the specificity of mAb 303 for a fucose-containing carbohydrate epitope, because fucose is incorporated into glycoproteins in the trans Golgi saccules (Bennett, 1984; Roth, 1987). In contrast, the route followed by amylase is presently unclear. Amylase reactivity was found in the Golgi saccules, but not in structures identifiable as part of the trans Golgi network (GERL). Both amylase and agglutinin were present in the earliest identifiable granules forming at the trans face, and amylase appeared to be associated with the dense core a s i t initially began to condense. The trans Golgi network is generally regarded as the site of sorting of secretory, lysosomal, and membrane proteins (Griffiths and Simons, 1986) and the point of divergence of the regulated and constitutive pathways (Burgess and Kelly, 1987). In some cells, enzyme- and immunocytochemical studies have failed to demonstrate secretory proteins (including amylase) in the trans Golgi network (Hand and Oliver, 1977a; Geuze et al., 1979; Bendayan, 1984,1989),despite its continuity with and apparent involvement in secretory granule formation. In other cells, a s in the present study, secretory proteins are clearly present in elements of the trans Golgi network (Geuze et al., 1984, 1985; Tooze and Tooze, 1986). These observations suggest t h a t in different cells, or even within a single cell, different routes may exist between the trans Golgi saccule and the forming granules, and/or that sorting of regulated proteins may occur at more than one site in the secretory pathway. A complicating factor in assessing the labeling of the trans Golgi compartments with mAb 303 is the apparent reactivity of this antibody with a shared carbohydrate epitope on one or more membrane glycoproteins located on the basolateral and luminal surfaces of the acinar cells. Consistent with this interpretation is the fact that mAb 116, which recognizes the same carbohydrate epitope, also labels the cell membrane, whereas mAb 143, with specificity for a polypeptide epitope, does not. The trans Golgi region probably is the site of sorting of the secretory agglutinin from these putative membrane glycoproteins, and at least part of the label in the trans Golgi region may be attributed to the latter. However, the abundance of agglutinin in the secretory granules, and the presumably higher turnover rate of the granules compared with that of the cell surface components recognized by mAb 303, suggest that most of the reactivity seen in the trans Golgi region must be agglutinin. Interestingly, small vesicles and tubules labeled with mAb 303, but not with anti-amylase antibody, were observed at both faces of the Golgi apparatus and frequently near the cell membrane. It is tempting to suggest that these structures are the morphological correlates of the constitutive secretion pathway in these cells, carrying membrane glycoproteins to the cell surface. On the other hand, i t is equally possible that these vesicles and tubules are part of the endocytic apparatus of these cells, and contain cell membrane glycoproteins destined for degradation in the lysosomal system. We did not observe any morphologically identifiable lysosomes labeled with mAb 303, but this may indicate that the epitope had already been degraded. Studies with additional markers, such a s antibodies to other cell membrane glycoproteins or to lysosomal membrane proteins such as lgp 120 (Lewis et al., 1985) or endolyn 78 (Croze et al., 1989), may help to resolve the nature of these vesicles and tubules. SALIVARY AGGLUTININ AND AMYLASE DISTRIBUTION 317 berg 1983 A convenient ELISA for testing whether monoclonal antibodies recognize the same antigenic site. J . Immunol. MethThe excellent technical assistance provided by Ms. ods, 60:351-358. Cheryl Davis and Mr. Alexis Kladakis is gratefully Fumagalli, G., and A. Zanini 1985 In cow anterior pituitary, growth hormone and prolactin can be packed in separate granules of the acknowledged. We thank Ms. Olevia Ambrose for help same cell. J. Cell Biol., 1OOt2019-2024. in preparing the photographs. Partial support for these Geuze, H.J., J.W. Slot, G.J.A.M. Strous, A. Hasilik, and K. von Figura studies was provided by NIH grants DE02623 and 1984 Ultrastructural localization of the mannose-6-phosphate receptor in rat liver. J . Cell Biol., 98.2047-2054, RR01224 to D.M., and T.K. was the recipient of a travel H.J., J.W. Slot, G.J.A.M. Strous, A. Hasilik, and K. von Figura grant from the Japanese government. We also thank Geuze, 1985 Possible pathways for lysosomal enzyme delivery. J . Cell Dr. J a n Thurin, Wistar Institute, for his advice and Biol., 101:2253-2262. assistance in demonstrating reactivity with Y and Geuze, J.J., J.W. Slot, and K.T. Tokuyasu 1979 Immunocytochemical localization of amylase and chymotrypsinogen in the exocrine Lewis epitopes. These studies were supported by the pancreatic cells with special attention to the Golgi complex. J . W.W. Smith Charitable Trust, Philadelphia, PennsylCell Biol., 82:697-707. vania. Grifiths, G., and K. Simons 1986 The trans Golgi network: sorting at the exit site of the Golgi complex. Science, 234:438-443. Hand, A.R., and W.D. Ball 1988 Ultrastructural immunocytochemical LITERATURE CITED localization of secretory proteins in autophagic vacuoles of paBall, W.D., A.R. Hand, and A.O. Johnson 1988 Secretory proteins as rotid acinar cells of starved rats. J . Oral Pathol., 17:279-286. markers for cellular phenotypes in rat salivary glands. Dev. Biol., Hand, A.R., and C. Oliver 1977a Relationship between the Golgi ap125t265-279. paratus, GERL, and secretory granules in acinar cells of the rat Bendayan, M. 1982 Double immunocytochemical labeling applying exorbital lacrimal gland. J . Cell Biol., 74r399-413. the protein A-gold technique. J. Histochem. Cytochem., 30r81Hand, A.R., and C. Oliver 1977b Cytochemical studies of GERL and 85. its role in secretory granule formation in exocrine cells. HisBendayan, M. 1984 Concentration of amylase along its secretory tochem. J., 9r375-392. pathway in the pancreatic acinar cell as revealed by high reso- Hay, D.I., R.J. Gibbons, and D.M. Spinell 1971 Characteristics of lution immunocytochemistry. Histochem. J., 16t85-108. some high molecular weight constituents with bacterial aggreBendayan, M. 1989 Ultrastructural localization of insulin and C-pepgating activity from whole saliva and dental plaque. Caries Res., tide antigenic sites in rat pancreatic B cell obtained by applying 6:111-123. the quantitative high-resolution protein A-gold approach. Am. J . Ichikawa, M., and A. Ichikawa 1987 The fine structure of sublingual Anat., 185:205-216. gland acinar cells of the Mongolian gerbil, Meriones unguiculaBendavan, M.. J. Roth, A. Perrelet, and L. Orci 1980 Quantitative tus, processed by rapid freezing followed by freeze-substitution immunocytochemical localization of pancreatic secretory proteins fixation. Cell Tissue Res., 25Or305-314. in subcellular compartments of the rat acinar cell. J. Histochem. Ichikawa, M., K. Sasaki, and A. Ichikawa 1989 Immunocytochemical Cytochem., 28:149-160. localization of amylase in gerbil salivary gland acinar cells proBennett, G. 1984 The role of the Golgi apparatus in secretion. In: Cell cessed by rapid freezing and freeze-substitution fixation. J. HisBiology of the Secretory Process. M. Cantin, ed. Karger, Basel, tochem. Cytochem., 37t185-194. pp. 102-147. Kashket, S., and C.G. Donaldson 1972 Saliva-induced aggregation of Blaszczyk-Thurin, M., J . Thurin, 0. Hindsgaul, K-A. Karlsson, Z. oral streptococci. J . Bacteriol., 112r1127-1133. Steplewski, and H. Koprowski 1987 Y and blood group B type 2 Keller, P.J., D.L. Kauffman, B.J. Allan, and B.L. Williams 1971 Furglycolipid antigens accumulate in a human gastric carcinoma cell ther studies on the structural differences between the isoenzymes line as detected by monoclonal antibody. Isolation and characterof human parotid a-amylase. Biochemistry, 1Ot4867-4874. ization by mass spectrometry and NMR spectroscopy. J. Biol. Kelly, R.B. 1985 Pathways of protein secretion in eukaryotes. Science, Chem., 262:372-379. 230:25-32. Brack, C.M., and E.C. Reynolds 1987 Characterization of a rat sali- Kornfeld, R., and S. Kornfeld 1985 Assembly of asparagine-linked vary sialoglycoprotein complex which agglutinates Streptococcus oligosaccharides. Annu. Rev. Biochem., 543531-664. mutans. Infect. Immun., 55t1264-1273. Kousvelari, E., F.G. Oppenheim, and L.S. Cutler 1982 UltrastrucBrown, AS., J. Silverman, S. Greenberg, D. Malamud, M. Album, tural localization of salivary acidic proline-rich proteins from R.W. Lloyd, and M. Sarshik 1985 A team approach to drool conMacaca fascicularzs. J. Histochem. Cytochem., 30:274-278. trol in cerebral palsy. Ann. Plastic Surg., 15:423-430. Lewis, V., S.A. Green, M. Marsh, P. Vihko, A. Helenius, and I. MellBurgess, T.L., and R.B. Kelly 1987 Constitutive and regulated secreman 1985 Glycoproteins of the lysosomal membrane. J . Cell Biol., tion of proteins. Annu. Rev. Cell Biol., 3t243-293. 100:1839-1847. Chung, K-N., P. Walter, G.W. Aponte, and H-P. H. Moore 1989 Mo- Machino, M., H. Morioka, and M. Tachibana 1986 Amylase and lecular sorting in the secretory pathway. Science, 243r192-197. lysozyme differentiate their localization within the serous secreClemente, F., and J. Meldolesi 1975 Calcium and pancreatic secretory granule of the human salivary gland. Acta Histochem. Cytion. I. Subcellular distribution of calcium and magnesium in the tochem., I9:329-332. exocrine pancreas of the guinea pig. J. Cell Biol., 65238-102. Magnani, J.L., M. Brockhaus, D.F. Smith, and V. Ginsburg 1982 DeCroze, E., I.E. Ivanov, G. Kreibich, M. Adesnik, D.D. Sabatini, and tection of d v c o h i d liaands bv direct bindinn of carbohvdrateM.G. Rosenfeld 1989 Endolyn-78, a membrane glycoprotein binding pgieins'to thjn-layer chromatogrank Methods" Enzypresent in morphologically diverse components of the endosomal mol., 83t235-240. and lysosomal compartments: implications for lysosome biogene- Malamud, D. 1985 Influence of salivary proteins on the fate of oral sis. J. Cell Biol., 108t1597-1613. bacteria. In: Molecular Basis of Oral Microbial Adhesion. S.E. Demuth, D.R., C.A. Davis, A.M. Comer, R.J. Lamont, P.S. Leboq, and Mergenhagen and B. Rosan, eds. Am. SOC. Microbiology, WashD. Malamud 1988 Cloning and expression of a Streptococcus sanington, DC, pp. 117-124. guis surface antigen that interacts with a human salivary agglu- Malamud, D., and E. Golub 1981 A model for saliva-mediated bactetinin. Infect. Immun., 56:2484-2490. rial aggregation. Adv. Physiol. Sci., 28:277-282. Ellison, S.A. 1979 The identification of salivary components. In: Sa- Malamud, D., R. Goldman, and N.S. Taichman 1983 Modulation of liva and Dental Caries. I. Kleinberg, S.A. Ellison and I.D. Manbacterial aggregation by PMN and platelet extracts. Inflammadel, eds. Sp. Suppl. Microbiology Abstracts. Information Retrieval tion, 7t133-144. Inc., New York, pp. 12-29. Merritt, A.D., and R.C. Karn 1977 The human a-amylases. Adv. HuEricson, T., and J . Rundegren 1983 Characterization of a salivary man Genetics, 8r135-234. agglutinin reacting with a serotype c strain of Streptococcus m u - Moreira, J.E., W.D. Ball, L. Mirels, and A.R. Hand 1991 Accumulatans. Eur. J. Biochem., 133t255-261. tion and localization of two adult acinar cell secretory proteins Ermak, T.H., and S.S. Rothman 1978 Internal organization of the during development of the rat submandibular gland. Am. J . zymogen granule: formation of reticular structures in uitro. J. Anat., 191:167-184. Ultrastruct. Res., 64r98-113. Moreira, J.E., L.A. Tabak, G.S. Bedi, D.J. Culp, and A.R. Hand 1989 Farquhar, M.G. 1985 Progress in unraveling pathways in Golgi trafLight and electron microscopic immunolocalization of rat subfic. Annu. Rev. Cell Biol., 1:447-488. mandibular gland mucin glycoprotein and glutamineiglutamic acid-rich proteins. J. Histochem. Cytochem., 37515-528. Friguet, B., L. Djavadi-ohaniance, J. Pages, A. Bussard, and M. Gold- ACKNOWLEDGMENTS 318 K. TAKANO ET AL. Nakajima, M., N. Ito, K. Nishi, Y. Okamura, and T. Hirota 1987 Immunogold labeling of blood-group antigens in human salivary glands using monoclonal antibodies and the streptavidin-biotin technique. Histochemistry, 87:539-543. Novikoff, A.B. 1976 The endoplasmic reticulum: a cytochemist’s view (a review). Proc. Natl. Acad. Sci. USA, 73t2781-2787. Palade, G. 1975 Intracellular aspects of the process of protein secretion. Science, 189:347-358. Pfeffer, S.R., and J.E. Rothman 1987 Biosynthetic protein transport and sorting by the endoplasmic reticulum and Golgi. Annu. Rev. Biochem., 56:829-852. Ravazzola, M., and L. Orci 1980 Glucagon and glicentin immunoreactivity are topologically segregated in the alpha granule of the human pancreatic A cell. Nature, 284:66-67. Riva, A., and F. Riva-Testa 1973 Fine structure of acinar cells of human parotid gland. Anat. Rec., 176~149-166. Riva, A,, G. Motta, and F. Riva-Testa 1974 Ultrastructural diversity in secretory granules of human major salivary glands. Am. J. Anat., 139r293-298. Rose, J.K., and R.W. Dams 1988 Regulation of protein export from the endoplasmic reticulum. Annu. Rev. Cell Biol., 4257-288. Roth, J. i987 Subcellular organization of glycosylation in mammalian cells. Biochim. Biophys. Acta, 906:405-436. Roth, J., M. Bendayan, and L. Orci 1978 Ultrastructural localization of intracellular antigens by the use of protein A-gold complex. J. Histochem. Cytochem., 26: 1074-108 1. Rothman, J.E., and L. Orci 1990 Movement of proteins through the Golgi stack: a molecular dissection of vesicular transport. FASEB J., 4:1460-1468. Rundegren, J., and R.R. Arnold 1987 Differentiation and interaction of secretory immunoglobulin A and a calcium-dependent parotid agglutinin for several bacterial strains. Infect. Immun., 55.288292. Stoller, T.J., and D. Shields 1989 The propeptide of preprosomatostatin mediates intracellular transport and secretion of a-globin from mammalian cells. J . Cell Biol., 108~1647-1655. Tandler, B., and R.A. Erlandson 1972 Ultrastructure of the human submaxillary gland. IV. Serous granules. Am. J. Anat., 135t419434. Tandler, B., and D.K. MacCallum 1972 Ultrastructure and histochemistry of the submandibular gland of the European hedgehog, Erinaceous europaeus L. I. Acinar secretory cells. J . Ultrastruct. Res., 39:186-204. Tooze, J., and S.A. Tooze 1986 Clathrin-coated vesicular transport of secretory proteins during the formation of ACTH-containing secretory granules in AtT-20 cells. J. Cell Biol., 1035339-850. Williams, R.C., and R.J. Gibbons 1975 Inhibition of streptococcal attachment to receptors on human buccal epithelial cells by antigenically similar salivary glycoproteins. Infect. Immun., 11:711718. Young, J.A., and E.W. van Lennep 1978 The Morphology of Salivary Glands. Academic Press, London.