Improved visualization of the immunoreactive hypothalamo-neurohypophysial system by use of immuno-gold techniques

Summary Ultrastructural post-embedding immuno-gold techniques were applied to the supraoptic nucleus and the neurohypophysis of mice and rats. The primary antibodies were three different monoclonal antineurophysins, used in protein A-gold and immunoglobulin-gold procedures. Conventional plastic embedding as well as hydrophilic media (L.R. White) were used; non-osmicated and osmicated tissues were immunolabeled; sodium metaperiodate oxidation was used, but was not essential for immunolabeling.Vasopressinergic and oxytocinergic NSGs were identified by the specific immunoreactivity of their respective neurophysins on adjacent thin sections, and by sequential double labeling on the same thin section using two different antibodies associated with gold probes of different diameters. The immunoidentification indicates that vasopressin NSGs can additionally be differentiated as larger, with more electron-dense matrix, and susceptible to damage by sodium metaperiodate.The only organelles consistently labeled were neurosecretory granules (NSGs), either intact or within lysosomal configurations. Some lysosomal dense bodies were immunoreactive even when discrete NSGs were no longer morphologically recognisable within them. Labeled NSGs were located within neuronal cell bodies, along axonal shafts and within axonal swellings and endings; occasionally immunoreactive NSGs were observed within synaptic boutons. Labeling intensity was semi-quantitatively gauged by counting gold particles in relation to numbers of NSGs per axonal varicosity.The precise localisation achieved with particulate immunogold labeling surpasses that previously obtained with diffuse electron-dense immunoreaction products.     

Electron microscopic localization of chromogranin A in osmium-fixed neuroendocrine cells with a protein A-gold technique

An antibody (LK2H10) to chromogranin A has been recommended for use in ultrastructural identification of neuroendocrine secretory granules. Previous studies have demonstrated immunoreactive chromogranin A in specimens prepared for electron microscopy by glutaraldehyde fixation only. In this study, the effect of specimen post-fixation by osmium tetroxide on post-embedding localization of chromogranin A was evaluated. Human tissues from benign endocrine glands, neuroendocrine tumors, and non-neuroendocrine tumors were post-fixed in osmium, embedded in epoxy resin, and the sample thin sections immunolabeled using a protein A-gold technique. Chromogranin A-positive neurosecretory granules were detected in pancreatic islets, adrenal medulla, stomach, ileum, anterior pituitary, and parathyroid. Mid-gut carcinoids, bronchial carcinoids, pheochromocytomas, paragangliomas, carotid body tumors, and thyroid medullary carcinomas contained immunoreactive granules. Cytoplasmic granules in non-neuroendocrine tumors did not react for chromogranin A. Tissues post-fixed in osmium tetroxide had optimally preserved ultrastructural features, and use of this fixative is compatible with postembedding localization of chromogranin A in neurosecretory granules.     

Ultrastructural localization of neurohypophyseal hormones and their carrier proteins by the "triple bridge" immunoenzymatic technique]

High specific antibodies have permitted the ultrastructural localization of vasopressin, ocytocin and neurophysins 1 and 2. Hormones and carrier proteins are detected in neurosecretory granules of the hypothalamic and neurohypophyseal fibres and so at the granular and extra-granular level of supra-optic pericaryons.     

Light and electron microscopical demonstration of concanavalin A and wheat-germ agglutinin binding sites by use of antibodies against the lectin or its label (peroxidase)

Three straining protocols for the ultrastructural visualization of concanavalin A (ConA) and wheat germ agglutinin (WGA) binding sites were applied to samples of nervous tissue embedded in Lowicryl K4M. The hypothalamo-neurohypophysial neurosecretory system was chosen for this investigation because it has two major neuronal populations, one secreting vasopressin, whose precursor is glycosylated, and the other secreting oxytocin whose precursor form is not glycosylated. The series of incubations of the tissue sections for the three protocols were: Protocol 1: i) non labeled ConA or WGA; ii) ConA or WGA antibody; iii) protein A-gold; Protocol 2: i) pre-prepared WGA-anti-WGA complex; ii) protein A-gold; Protocol 3: i) peroxidase-labeled ConA or WGA; ii) anti-peroxidase; iii) protein A-gold. The three methods allowed to detect fine differences in the distribution of sugar residues. This, in turn, made it possible to distinguish vasopressin granules containing precursor forms from those containing processed precursor. At the light microscopic level the three methods were successfully applied to paraffin and 1-micron methacrylate sections by using a second antibody, PAP complex and the diaminobenzidine reaction.     

Improved immunoelectron microscopic method for localizing cytoskeletal proteins in Lowicryl K4M embedded tissues

We have modified the Lowicryl K4M low-temperature dehydration and embedding procedure for immunoelectron microscopy to provide improved ultrastructural detail and facilitate the localization of actin and tubulin in isolated rat adrenocortical cells, chick spinal cord with attached dorsal root ganglia (SC-DRG), and cultured dorsal root ganglia (DRG). Cells and tissues were fixed for immunocytochemistry either in a mixture of 2% paraformaldehyde and 0.25% glutaraldehyde (0.1 M PIPES buffer, pH 7.3) or in a mixture of 0.3% glutaraldehyde and 1.0% ethyldimethylaminopropylcarbodiimide (0.1 M phosphate buffered saline, pH 7.3). Dehydration was in ethanol at progressively lower temperatures to -35 degrees C. Infiltration at -35 degrees C was followed by ultraviolet polymerization at -20 degrees C. Comparable samples were fixed in glutaraldehyde and osmium tetroxide and embedded in Epon 812 or Epon-Araldite. Post-embedding immunostaining of thin sections utilized commercially available monoclonal antibodies to tubulin and actin followed by the protein A-gold technique (Roth et al., Endocrinology 108:247, 1981). Actin immunoreactivity was observed at the periphery of mitochondria and between mitochondria and lipid droplets in rat adrenocortical cells and at the periphery of neuronal cell processes of SC-DRG. Tubulin immunoreactivity was associated with microtubules throughout neurites of cultured DRG. Our modified technique allows preservation of ultrastructural details as well as localization of antigens by immunoelectron microscopy.     

Quantitative immunogold localization of protein phosphatase 2B (calcineurin) in Paramecium cells.

For immunogold EM labeling analysis, we fixed Paramecium cells in 4% formaldehyde and 0.125% glutaraldehyde, followed by low-temperature embedding in unicryl and UV polymerization. We first quantified some obvious but thus far neglected side effects of section staining on immunogold labeling, using mono- or polyclonal antibodies (Abs) against defined secretory and cell surface components, followed by F(ab)(2)- or protein A-gold conjugates. Use of alkaline lead staining resulted in considerable rearrangement and loss of label unless sections were postfixed by glutaraldehyde after gold labeling. This artifact is specific for section staining with lead. It can be avoided by staining sections with aqueous uranyl acetate only to achieve high-resolution immunogold localization of a protein phosphatase on unicryl sections. In general, phosphatases are assumed to be closely, although loosely, associated with their targets. Because the occurrence of protein phosphatase 2B (calcineurin) in Paramecium has been previously established by biochemical and immunological work, as well as by molecular biology, we have used Abs against mammalian CaN or its subunits, CaN-A and CaN-B, for antigen mapping in these cells by quantitative immunogold labeling analysis. Using ABs against whole CaN, four structures are selectively labeled (with slightly decreasing intensity), i.e., infraciliary lattice (centrin-containing contractile cortical filament network), parasomal sacs (coated pits), and outlines of alveolar sacs (subplasmalemmal calcium stores, tightly attached to the cell membrane), as well as rims of chromatin-containing nuclear domains. In other subcellular regions, gold granules reached densities three to four times above background outside the cell but there was no selective enrichment, e.g., in cilia, ciliary basal bodies, cytosol, mitochondria, trichocysts (dense-core secretory organelles), and non-chromatin nuclear domains. Their labeling density was 4- to 8.5-fold (average 6.5-fold) less than that on selectively labeled structures. Labeling tendency was about the same with Abs against either subunit. Our findings may facilitate the examination of molecular targets contained in the selectively labeled structures. (J Histochem Cytochem 48:1269-1281, 2000)     

Post-embedding methods for immunolocalization of elastin and related components in tissues

Elastic tissue is composed of amorphous-appearing elastin and 12-nm diameter microfibrils, one component of which has recently been isolated and characterized as the 31 KD microfibril-associated glycoprotein MAGP. Monospecific antibodies to each of these components have been developed in this laboratory. The parameters that determine optimal localization of colloidal gold probes for post-embedding immunolabeling of elastic tissue components have been systematically studied in a variety of normal and developing tissues in mammals and birds. Protein A-gold probes stabilized with dextran have been shown to provide complexes that remain stable after more than 2 years. Conditions have been defined that permit precise localization within the extracellular matrix of antibodies to MAGP and to elastin, singly and together. Best results were obtained with acrylic resins (Lowicryl K4M or LR White). Fixation in glutaraldehyde or other aldehydic fixatives, with or without osmium, did not affect the immunostaining of elastic tissue with affinity-purified antibodies to tropoelastin, or to anti-[alpha-elastin] or anti-[alkali-insoluble elastin]. Immunostaining with the anti-MAGP antibody was less robust and was possible in tissues which had been fixed only lightly before embedding in Lowicryl K4M or LR White. This staining was enhanced by metaperiodate oxidation of the sections as well as by reduction of the tissues with sodium borohydride en bloc, followed by hyaluronidase digestion of the sections. The effects on immunostaining of a range of enzyme digestions have also been examined. Conditions have thus been defined that make possible detailed study of the relationship between elastic tissue, elastin-associated microfibrils, and other microfibrillar structures in normal and abnormal tissues during development and aging.     

An improved immunocytochemical method for subcellular localization of serotonin in rat enterochromaffin cells

Serotonin-like immunoreactivity (5-HT-LI) has been localized at the ultrastructural level in enterochromaffin (EC) cells of rat gastrointestinal tract. Ultra-thin sections of tissues embedded in epoxy resin were incubated with 5-HT antisera and antibody binding sites were visualized with protein A-gold. Three different antisera were compared and were shown to require different fixation regimens for optimal preservation of 5-HT-LI. For one antiserum, tissues fixed in glutaraldehyde and osmium tetroxide could be used to demonstrate 5-HT-LI in EC cells. Immunocytochemical localization of 5-HT can thus be performed with good ultrastructural preservation of tissues. Quantitative evaluation of the intracellular distribution of 5-HT-LI was performed on EC cells from antrum, duodenum, and proximal colon, fixed in glutaraldehyde only. In all three locations, the majority of the gold particles (90%) in EC cells were localized over the dense core of the secretory granules, while a minor fraction (10%) were localized in parts of the cytoplasm devoid of granules. In EC cells fixed in glutaraldehyde and post-fixed in osmium tetroxide, 5-HT-LI was reduced by about 85%, although intracellular distribution was essentially the same as in cells fixed in glutaraldehyde alone. The results indicate that 5-HT in EC cells is stored mainly in secretory granules, with a small fraction of 5-HT being localized outside the granules.     

Ultrastructural localization of rat Clara cell 10 KD secretory protein by the immunogold technique using polyclonal and monoclonal antibodies

Using a monoclonal antibody and affinity-purified polyclonal antiserum against a 10 KD protein isolated from rat pulmonary lavage, we have localized the protein within Clara cells by a post-embedment protein A-gold technique. The gold particles were localized over the secretory granules of rat Clara cells. Ultrastructural immunolocalization was abolished when the primary antibodies were previously absorbed with purified 10 KD protein. Other pulmonary cells, including type II pneumocytes and ciliated cells, were negative with this technique. These results demonstrate the presence of the 10 KD protein in the secretory granules of the Clara cell and support the concept that this protein constitutes a specific and unique secretory product of Clara cells.     

Ultrastructural,cytochemical and immunocytochemical investigation of the interphase nucleus in the unicellular green alga Chlamydomonas reinhardtii

Summary— The ultrastructural organization of the interphase nucleus of the green alga Chlamydomonas reinhardtii was investigated and found to be largely dependent on the fixation conditions. In specimens stained with bismuth, densely contrasted granules ranging from 25 to 45 nm in diameter were localized throughout the interchromatin space and often formed clusters. These granules were labeled by RNase A-gold complexes and may represent the counterparts of animal and higher plant cll interchromatin granules. Within the nucleolus the Ag-NOR and pyroantimonate stains and, to a lesser extent, the bismuth stain reacted with the nucleolar dense fibrillar component (DFC). When cells were subjected to a heat shock at 42°C, the nucleolar DFC was found to progressively separate from the nucleolus and, after 3 h, appeared as a continuous meandering thread about 0.1 μm in width. Within the nucleolus, labeling on conventional preparations occurred as small clusters with antibodies to H3 histones or to DNA whereas RNase A-gold complexes labeled most of it including fibrillar centers. Improved ultrastructural preservation in cryofixed, cryosubstituted specimens gently fixed in glutaraldehyde permitted to localize nucleolar DNA predominantly at the outer edge of fibrillar centers and to a lesser extent within the neighbouring DFC. Our results indicate that the structure and composition of Chlamydomonas interphase nuclei are comparable, despite particularities, to those of animal and higher plant nuclei.     

Ultrastructural localization of antigenic sites on osmium-fixed tissues applying the protein A-gold technique

The protein A-gold immunocytochemical technique has been modified to allow labeling of cellular antigenic sites on osmium-fixed or postfixed tissues. Several strong oxidizing agents have been found able to restore protein antigenicity on osmicated tissue thin sections. According to the fine structural preservation and intensities of labeling, pretreatment with sodium metaperiodate gave optimal results. Pancreatic secretory proteins (and/or proproteins) as well as insulin (and/or proinsulin) were localized over perfectly preserved rough endoplasmic reticulum (rER), Golgi apparatus, and secretory granules of the corresponding pancreatic cells; carbamyl phosphate synthetase and catalase were revealed over liver mitochondria and peroxisomes, respectively. In addition to the higher resolution in the labeling obtained using osmium-fixed tissues, the present modification confers an additional advantage to the protein A-gold technique by allowing labeling on tissues processed for routine electron microscopy.     

Protein G-gold complex: comparative evaluation with protein A-gold for high-resolution immunocytochemistry

We combined the protein G-gold complex with several polyclonal and monoclonal antibodies for localization of various antigenic sites. The labelings were compared with those obtained using the protein A-gold complex. The results from either the immunodot experiment or immunoelectron microscopy have demonstrated that, for rabbit and guinea pig antibodies, both protein G-gold and protein A-gold complexes label several different specific antibodies with similar efficiency. However, with antibodies raised in goats or in mice, and particularly with mouse monoclonal antibodies, protein G-gold yielded intense and specific labeling, whereas protein A-gold yielded intense and specific labeling, whereas protein A-gold was very variable; it either gave weaker signals or failed to reveal any specific site or, as with one monoclonal, both protein G and protein A gave similar results. The higher affinity and versatility of protein G over protein A, established by the immunochemical approach, was confirmed by immunocytochemistry. Because of its enhanced reactivity with monoclonal antibodies and its broader affinity for polyclonal antibodies, protein G-gold complex appears to be a better and more versatile probe for high-resolution immunocytochemistry.     

Histology, histochemistry, and ultrastructure of neurosecretory cells in the optic lobe of the cockroach, Periplaneta americana

In the region of the distal optic chiasma of each optic lobe of Periplaneta americana, there is a group of about 120 monopolar neurosecretory cells. These cells do not stain with paraldehyde fuchsin but remain acidophilic after oxidation. They stain red or sometimes indigo with the azan technique. Histochemically, the neurosecretory material is positive for protein and the amino acids tryptophan and arginine but negative for 1, 2-glycols and strongly acidic groups. At the ultrastructural level, the cytoplasm of the cells contain many elementary neurosecretory granules 100 to 170 nm in dia. The cells also contain well-developed Golgi bodies and endoplasmic retieulum. The axons from these cells run toward the interior of the optic lobe. In this region, axons containing dense granules (mean diameter 70 nm) and synaptic vesicles synapse onto the axons from the neurosecretory cells. The neurosecretory axons then cross over to the anterior side of the optic lobe and run towards the brain. The function of these neurosecretory cells is unknown, but they may be involved with photoperiodically controlled activity rhythms.     

A Similar Calmodulin-Binding Protein Expressed in Chromaffin, Synaptic, and Neurohypophyseal Secretory Vesicles

The presence of calmodulin-binding proteins in three neurosecretory vesicles (bovine adrenal chromaffin granules, bovine posterior pituitary secretory granules, and rat brain synaptic vesicles) was investigated. When detergent-solubilized membrane proteins from each type of secretory organelle were applied to calmodulin-affinity columns in the presence of calcium, several calmodulin-binding proteins were retained and these were eluted by EGTA from the columns. In all three membranes, a 65-kilodalton (63 kilodaltons in rat brain synaptic vesicles) and a 53-kilodalton protein were found consistently in the EGTA eluate. 125I-Calmodulin overlay tests on nitrocellulose sheets containing transferred chromaffin and posterior pituitary secretory granule membrane proteins showed a similarity in the protein bands labeled with radioactive calmodulin. In the presence of 10(-4) M calcium, eight major protein bands (240, 180, 145, 125, 65, 60, 53, and 49 kilodaltons) were labeled with 125I-calmodulin. The presence of 10 microM trifluoperazine (a calmodulin antagonist) significantly reduced this labeling, while no labeling was seen in the presence of 1 mM EGTA. Two monoclonal antibodies (mAb 30, mAb 48), previously shown to react with a cholinergic synaptic vesicle membrane protein of approximate molecular mass of 65 kilodaltons, were tested on total membrane proteins from the three different secretory vesicles and on calmodulin-binding proteins isolated from these membranes using calmodulin-affinity chromatography. Both monoclonal antibodies reacted with a 65-kilodalton protein present in membranes from chromaffin and posterior pituitary secretory granules and with a 63-kilodalton protein present in rat brain synaptic vesicle membranes. When the immunoblotting was repeated on secretory vesicle membrane calmodulin-binding proteins isolated by calmodulin-affinity chromatography, an identical staining pattern was obtained. These results clearly indicate that an immunologically identical calmodulin-binding protein is expressed in at least three different neurosecretory vesicle types, thus suggesting a common role for this protein in secretory vesicle function.     

In cow anterior pituitary, growth hormone and prolactin can be packed in separate granules of the same cell

The ultrastructural localization of growth hormone and prolactin in cow anterior pituitary was studied by double immunocytochemical labeling using specific antibodies and protein A-gold particles of different sizes. The two hormones were found in specific somatotrophs and mammotrophs as well as in somatomammotropic cells which were multinucleated and predominantly arranged in clusters in the central area of the lobules. In these mixed cells the two hormones were packaged (a) in different granules of the same cell, (b) in the same granules where they were segregated in different portions of the granule content, or (c) in the same granules but evenly intermixed. The relative proportion of these three types of granules varied in somatomammotrophs of different animals. A single large Golgi complex was generally present in somatomammotrophs. Small, immature granules containing either growth hormone or prolactin or both hormones were found randomly distributed along Golgi stacks. This suggests that in these cells the two hormones are processed in the same Golgi cisternae and that mechanism(s) exist(s) to sort out the two hormones from each other.     

The isolation of neurophysin-I and -II from bovine pituitary neurosecretory granules separated on a large scale from other subcellular organelles. Demonstration of slow equilibration of neurosecretory granules during centrifugation in a sucrose density gradient

1. An improved procedure for the isolation of neurosecretory granules from the posterior lobe of the bovine pituitary gland is described. 2. Of the total oxytocic and pressor activities present in the original tissue 80% was sedimentable. 3. The granules were separated from mitochondria by prolonged centrifugation in a sucrose density gradient. During a sedimentation period of 5hr. the granules moved progressively into denser regions of the gradient and the mitochondria remained at the top. 4. The biological activities of the granules were measured: the oxytocic activity was 11·56±1·63 and the pressor activity was 15·60±3·91 units/mg. of protein. 5. A protein was isolated from a lysate of granules prepared from 40 pituitary glands. Amino acid analysis showed that it consisted of a mixture of neurophysin-I and neurophysin-II in equal proportions. It accounted for 60% of the soluble granule protein and for 50% of the total granule protein. 6. The neurophysins present in the granules are associated with 19·1 units of oxytocic and 21·1 units of pressor activity/mg. of protein. 7. Starch-gel electrophoresis revealed the presence of both neurophysins in extracts of 15 pituitary glands studied individually. 8. We conclude that the polypeptide hormones, oxytocin and [8-arginine]-vasopressin, are normally closely associated with the two neurophysins within neurosecretory granules of the pituitary gland.     

Immunoelectron microscopic labeling of immunoglobulin in plasma cells after osmium fixation and epoxy embedding

Previous studies have found that immunoglobulin cannot be immunolabeled in tissues prepared for electron microscopy by usual methods. To test this conclusion, we used a protein A-gold postembedding immunolabeling method on tissues that were fixed in glutaraldehyde, post-fixed in osmium tetroxide, and embedded in epoxy resin; sections were pretreated with sodium metaperiodate. A variety of common fixation protocols were also used and the most suitable conditions for immunolabeling were determined. This technique permitted the ultrastructural localization of immunoglobulin light chains in optimally preserved and contrasted plasma cells from human tonsil, lymph nodes, plasmacytomas, and a renal biopsy. We were able to demonstrate multiple antigens in the same tissue and label antigens in tissues that had been stored for many years in epoxy resin. The technique allows quantitation of the gold label over plasma cell organelles and therefore gives information about the immunoglobulin secretory pathway in these cells. We found that the protein A-gold procedure compares favorably in technical ease with the immunoperoxidase, avidin-biotin peroxidase, and immunoglobulin-colloidal gold immunolabeling methods, and has added advantages in allowing precise localization and quantitation of the labeled antigen.     

Ultrastructural staining with sodium metaperiodate and sodium borohydride.

This article describes new ultrastructural staining methods for osmicated tissues based on the incubation of sections with sodium metaperiodate and sodium borohydride solutions before uranyl/lead staining. Sections incubated with sodium metaperiodate and sodium borohydride, treated with Triton X-100, and stained with ethanolic uranyl acetate/lead citrate showed a good contrast for the nucleolus and the interchromatin region, whereas the chromatin masses were bleached. Chromatin bleaching depended on the incubation with these oxidizing (metaperiodate) and reducing (borohydride) agents. Other factors that influenced the staining of the chromatin masses were the en bloc staining with uranyl acetate, the incubation of sections with Triton X-100, and the staining with aqueous or ethanolic uranyl acetate. The combination of these factors on sections treated with metaperiodate/borohydride provided a different appearance to the chromatin, from bleached to highly contrasted. Most cytoplasmic organelles showed a similar appearance with these procedures than with conventional uranyl/lead staining. However, when sections were incubated with metaperiodate/borohydride and Triton X-100 before uranyl/lead staining, the collagen fibers, and the glycocalix and zymogen granules of pancreatic acinar cells, appeared bleached. The possible combination of these methods with the immunolocalization of the amino acid taurine was also analyzed. (J Histochem Cytochem 50:11-19, 2002)     

Light and electron microscopical demonstration of concanavalin A and wheat-germ agglutinin binding sites by use of antibodies against the lectin or its label (peroxidase)

Summary Three straining protocols for the ultrastructural visualization of concanavalin A (ConA) and wheat germ agglutinin (WGA) binding sites were applied to samples of nervous tissue embedded in Lowicryl K4M. The hypothalamo-neurohypophysial neurosecretory system was chosen for this investigation because it has two major neuronal populations, one secreting vasopressin, whose precursor is glycosylated, and the other secreting oxytocin whose precursor form is not glycosylated.The series of incubations of the tissue sections for the three protocols were: Protocol 1: i) non labeled ConA or WGA; ii) ConA or WGA antibody; iii) protein A-gold; Protocol 2: i) pre-prepared WGA-anti-WGA complex; ii) protein A-gold; Protocol 3: i) peroxidase-labeled ConA or WGA; ii) anti-peroxidase; iii) protein A-gold.The three methods allowed to detect fine differences in the distribution of sugar residues. This, in turn, made it possible to distinguish vasopressin granules containing precursor forms from those containing processed precursor.At the light microscopic level the three methods were successfully applied to paraffin and 1-m methacrylate sections by using a second antibody, PAP complex and the diaminobenzidine reaction.Supported by Grant I/63 476 from Stiftung Volkswagenwerk, Federal Republic of Germany and Grant S-85-39 from Dirección de Investigaciones, Universidad Austral de Chile. The authors wish to acknowledge the technical help of Genaro Alvial and Elizabeth     

Electron microscopic immunocytochemical localization of proline-rich proteins in normal mouse parotid salivary glands

Summary Rabbit polyclonal antibodies against isoproterenol-induced mouse proline-rich proteins (PRPs) were used to localize PRPs in the parotid salivary glands of normal adult BALB/c mice. The antibodies recognized both acidic-type and basic-type PRPs. Immunoblotting experiments revealed that the glands contained an acidic-type and a basic-type PRP. Parotid gland tissue was fixed with Karnosky's fixative and embedded in Lowicryl resin at low temperature. PRPs were localized at the electron microscope level using an indirect post-embedding staining technique with protein A-gold. The secretion granules of the acinar cells were strongly labelled. Pre-absorption of the antibody with purified acidic-type and basic-type PRPs indicated that the basic-type PRP is mainly located at the periphery of the granules but that the acidic-type PRP is more evenly distributed within the granules. Pre-absorption of the antibody with -amylase did not affect the staining pattern, suggesting minimal cross-reactivity. PRPs were also detected within the rough endoplasmic reticulum and the Golgi apparatus of acinar cells, within the granules of the proacinar cells and in the lumena of the ducts, but not within the intercalated or striated duct cell granules.