Unmasking and redistribution of lysosomal sulfated glycoconjugates in phagocytic polymorphonuclear leukocytes

Rabbit heterophil and human neutrophil primary granules contain sulfated glycosaminoglycans (GAGs) and acid phosphatase, which can be readily stained in immature but not mature lysosomes. To determine whether this loss of staining represents masking of reactive components or removal of these components, we examined rabbit heterophils to see if high-iron diamine (HID)-reactive sulfate and acid phosphatase staining reappears in phagocytic vacuoles. Rabbit heterophils, obtained by peritoneal lavage, were incubated in vitro with latex beads or Pseudomonas aeruginosa for 15-60 min. Pre-embedment HID staining was enhanced in thin sections of unosmicated specimens with thiocarbohydrazide and silver proteinate (TCH-SP). Phagocytosis of latex beads or bacteria was progressively more prominent with time. Primary granules that were degranulated or in the process of degranulating into phagocytic vacuoles demonstrated intense sulfate staining with large (13 +/- 7 nm) HID-TCH-SP stain deposits. Smaller (6 +/- 1 nm) HID-TCH-SP stain deposits were present in tertiary granules, which were less frequently observed degranulating into phagosomes. Acid phosphatase staining was most intense during early phagolysosome formation. HID-TCH-SP staining was also observed in extracellular degranulated lysosomal matrices and on the surface of many peritoneal heterophils. These results indicate that loss of sulfate staining in mature heterophil granules is the result of masking by intragranular substances rather than of removal, and that these components may be unmasked during phagocytosis and/or redistributed to the cell surface after exocytosis.     

The influence of embedding media and fixation on the postembedment ultrastructural demonstration of complex carbohydrates. III. High iron diamine staining for sulfated glycoconjugates

The high iron diamine (HID) method for detection of sulfated complex carbohydrate has been applied directly on thin sections of variably fixed tissues embedded in epoxy and nonepoxy resins. Results with postembedment HID staining in mouse intestinal epithelium are compared to those previously obtained using preembedment methods. Sections from epoxy-embedded tissues have been found to exhibit the weakest staining intensity. Intense, specific staining was obtained in tissues not postfixed with osmium tetroxide and embedded in polystyrene, polyester resins, styrene-methacrylate, and especially the styrene-Vestopal W embedding mixture. Postosmication of tissues abolished HID staining in epoxy resins and the styrene-Spurr's resin embedding mixture, but only reduced the staining intensity in tissues embedded in nonepoxy resins.     

Ultrastructural localization of acidic glycoconjugates with the low iron diamine method

The present study has applied the low iron diamine (LID) method at the ultrastructural level to demonstrate acid glycoconjugates. We have examined rat epiphyseal cartilage, human bone marrow, rat tracheal glands, and mouse sublingual glands stained with LID prior to embedment. The LID staining appeared to require postosmication for adequate visualization at the electron microscope level. Thiocarbohydrazide-silver proteinate (TCH-SP) staining of thin sections variably enhanced LID reactive sites. LID-TCH-SP stained carboxyl and sulfate groups of glycosaminoglycans in the extracellular cartilage matrix, secretory granules, and expanded Golgi saccules of chondrocytes. In human bone marrow, LID-TCH-SP variably stained the cytoplasmic granules, known to contain sulfated glycosaminoglycans, and the external surface of the plasma membrane of leukocytes. Moderately strong LID staining was observed in secretory granules in mucous tubules of rat tracheal glands, known to contain sulfated glycoproteins, and in acinar cells of mouse sublingual glands, known to contain a sialoglycoprotein. The lack of sulfated glycoconjugates in acinar cells of the mouse sublingual gland was confirmed by their failure to stain with the high iron diamine method. Thus these studies indicate that the LID and LID-TCH-SP methods are useful for the ultrastructural localization of carboxylated and sulfated glycoconjugates in extracellular and intracellular sites.     

Post-embedding staining with high-iron diamine-thiocarbohydrazide-silver proteinate and its application to visualizing sulfated glycoconjugates in cryofixed kidney and cartilage.

Cryofixation is generally believed to provide optimal tissue preservation. However, certain post-embedding cytochemical reactions, such as high-iron diamine (HID) staining for sulfated glycoconjugates, are not applicable to cryofixed and freeze-substituted tissues. In the present study, the HID technique was therefore adapted for post-embedding staining. HID staining was performed on thin sections of chemically and cryofixed kidney and growth plate cartilage, embedded in Epon and various acrylic-based resins. All resins and most tissue preparation conditions allowed post-embedding staining with HID, albeit to variable degrees. However, no significant cytochemical reaction was obtained with tissue sections of osmicated kidney embedded in Epon. Profile views of re-embedded sections showed that large stain deposits were usually restricted to the surface, whereas small ones were observed throughout the entire thickness of the section. The staining pattern was essentially similar between chemically fixed and cryofixed specimens. In the glomerulus, stain deposits were mainly seen over the free surface of podocyte foot processes and over the lamina rara externa. The pericellular cartilage matrix of chemically fixed specimens often appeared as condensed elements, usually stained with large deposits. In cryofixed tissues this matrix formed a meshwork composed of thin, extended filamentous structures, many of which showed linear arrays of smaller stain deposits. The data presented here indicate that post-embedding HID-TCH-SP staining can be successfully performed on thin sections of tissues embedded in various resins and, as a result, can be further adapted to cryo-prepared specimens to give a high resolution localization of sulfated glycoconjugates in tissues with optimal molecular preservation.     

Ultrastructural localization of complex carbohydrates in odontoblasts, predentin, and dentin

Glycosaminoglycans (GAGs) and glycoproteins (GPs) are essential components for dentinogenesis. We have examined rat odontoblasts, predentin, and dentin decalcified with EDTA and stained with: 1) Spicer's hig-iron diamine-thiocarbohydrazide-silver proteinate (HID-TCH-SP) method for sulfated glycoconjugates, and 2) Thiéry's periodate-thiocarbohydrazide-silver proteinate (PA-TCH-SP) method for vicinal glycol-containing glycoconjugates. HIS-TCH-SP stained distended portions of Golgi saccules and secretory granules. The predentin contained three times the number of HID-TCH-SP stain precipitates when compared to the mineralization front of the dentin matrix. PA-TCH-SP weakly stained membranes of Golgi saccules and cisternae of rough endoplasmic reticulum (RER), whereas stronger staining was observed in secretory granules, lysosomes, and multivesicular bodies (MVBs). Collagen fibrils in predentin demonstrated moderate PA-TCH-SP staining. In contrast, strong PA-TCH-SP staining was observed on and between collagen fibrils in the mineralization front of the dentin matrix. TCH-SP controls of unosmicated specimens lacked significant staining, however, osmicated control specimens did contain some TCH-SP stain deposits in the mineralization front. These results indicate that sulfated and vicinal glycol-containing glycoconjugates are packaged in the same type of secretory granule and released into the extracellular matrix; subsequently vicinal glycol-containing glycoconjugates concentrate in the calcification front, whereas sulfated glycoconjugates accumulate in the predentin and are either removed or masked to staining in the dentin.     

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.     

Mitochondrial matrix granules in soft tissues : I. Elemental composition by X-ray microanalysis

The elemental composition of individual matrix granules in mitochondria of rat brown fat, mouse gall bladder and guinea pig kidney has been examined by X-ray microanalysis. The matrix granules showed a similar elemental composition that was strongly dependent upon the method of sample preparation. Low-temperature oxygen plasma microincineration or wavelength-dispersive X-ray spectrometers were used to demonstrate the presence of phosphorus in matrix granules of osmium tetroxide-fixed specimens. Matrix granule osmiophilia was retained in glutaraldehyde-fixed brown fat only if exposure to polar organic solvents was avoided during subsequent steps, e.g. by cryosectioning. As normally prepared, matrix granules lack detectable calcium but had bound this at detectable levels after fixation in osmium tetroxide, but not glutaraldehyde, supplemented with 5 mM Ca²⁺. The results demonstrate that mitochondrial matrix granules of normal soft tissues are not calcium phosphate deposits and contain phospholipids, apparently as a major constituent. Thus they provide evidence against the hypothesis that matrix granules are primarily involved in mitochondrial calcium sequestration and, indirectly, for the hypothesis that the granules may be related to inner membrane assembly.     

A simple impregnation technique for thin and semithin enterochromaffin cells in sections

Summary Constant, intense and precise impregnation of enterochromaffin (EC) cells was achieved simply by floating thin or semithin sections of gut mucosa, fixed in osmium tetroxide or in glutaraldehyde with postfixation in osmium, on a silver nitrate or proteinate solution. EC cells alone showed impregnation in the light microscope. In the electron microscope, impregnation affected not only the secretory granules of EC cells but also, although much more faintly, those of other, non-EC cells (D, X, D₁, G and other cells). Lysosomes also showed partial or total reactivity. Oxidation reduced but did not entirely suppress EC cell staining and had no effect on non-EC endocrine cell staining. Since the reaction did not occur with glutaraldehyde alone, osmium appeared to be a crucial component of the process. These findings should be borne in mind in applying Thiery's method for vicinal glycol groups to the type of study material used in these experiments.     

A simple impregnation technique for thin and semithin enterochromaffin cells in sections

Constant, intense and precise impregnation of enterochromaffin (EC) cells was achieved simply by floating thin or semithin sections of gut mucosa, fixed in osmium tetroxide or in glutaraldehyde with postfixation in osmium, on a silver nitrate or proteinate solution. EC cells alone showed impregnation in the light microscope. In the electron microscope, impregnation affected not only the secretory granules of EC cells but also, although much more faintly, those of other, non-EC cells (D, X, D1, G and other cells). Lysosomes also showed partial or total reactivity. Oxidation reduced but did not entirely suppress EC cell staining and had no effect on non-EC endocrine cell staining. Since the reaction did not occur with glutaraldehyde alone, osmium appeared to be a crucial component of the process. These findings should be borne in mind in applying Thiery's method for vicinal glycol groups to the type of study material used in these experiments.     

Ultrastructural silver enhancement of Prussian blue-reactive iron in hematopoietic and intestinal cells

Prussian blue has been widely used to localize iron in a variety of tissues at the light and electron microscopic level. In the present study, thin sections of human marrow and blood cells and rat duodenal cells were exposed to silver proteinate (SP) after staining en bloc with acid ferrocyanide (AF), with and without prior iron saturation using iron nitrilotriacetate (FeNTA). Silver deposition was observed over Prussian blue-reactive sites and significantly enhanced sites of minimal AF and FeNTA-AF staining. AF-SP stain deposits were present in the cytoplasmic matrix, granules, and occasionally on the surfaces of macrophages, monocytes, and erythroblasts. FeNTA-AF-SP stained additional cytoplasmic and surface sites in erythroblasts and stained neutrophil granules intensely. Duodenal epithelium from iron-loaded rats demonstrated strong AF-SP staining of ferric iron in microvilli, apical cytoplasmic matrix, and lateral membranes. Similar preparations from iron-replete rats stained sparsely; however, intense AF-SP staining was observed after iron saturation with FeNTA. SP similarly enhanced luminal ferrous iron deposits stained with acid ferricyanide in rats given intraluminal ferrous iron. AF-SP stain deposits were removed by exposure of thin sections to NH4OH, KCN, or HNO3 but were not affected by prior exposure to HIO4 or NaBH4, consistent with a silver cyanide or complex stain precipitate rather than reduced silver or silver ferriferrocyanide. SP enhancement of Prussian blue allows identification of reactive sites not readily visualized with AF or FeNTA-AF alone, and offers the potential for differentiating AF staining from other deposits or organelles of comparable density.     

Tridimensional architecture of the Golgi apparatus and its components in mucous cells of Brunner's glands of the mouse

The three-dimensional structure of the Golgi apparatus and its components has been analyzed in thin and thick sections of mucous cells of mouse Brunner's glands by using low- and high-voltage electron microscopes and a stereoscopic approach. In thick sections of glands impregnated with osmium or treated to detect nicotinamide adenine dinucleotide phosphatase (NADPase) or thiamine pyrophosphatase (TPPase) activity, the Golgi apparatus appeared, at low magnification, as a continuous network located in the supranuclear region. At higher magnifications and in thin sections of tissue postfixed with reduced osmium and stained with lead citrate or treated to demonstrate phosphatase activity, the following components were observed: on the cis-face of the Golgi stacks, an osmiophilic tubular network referred to as the cis-element; a cis-saccular-compartment composed of a distended porous saccule slightly reactive for NADPase and three or four underlying NADPase-positive, flattened, poorly fenestrated saccules; a trans-saccular-compartment consisting of four to six TPPase-positive saccules or sacculo-tubular elements, prosecretory granules, and "peeling off" trans-tubular networks. The saccules of the cis-compartment were often perforated by large pores in register. The cavities thus formed in the stacks were called wells and were pan-shaped with a mouth directed toward the cis-face of the stacks and a bottom closed by TPPase-positive saccules. The wells always contained 80-nm vesicles. The saccules of the trans-compartment were involved in the formation of secretory granules according to the following proposed sequence of transformation. The secretion product appeared initially as a granular material evenly distributed throughout a slightly distended, poorly fenestrated saccule. These saccules appeared to transform into fenestrated elements with irregular pores and with parts of them taking on the appearance of a tubular network; they were thus referred to as sacculotubular elements. The secretory material initially distributed throughout these elements accumulated in nodular dilatations randomly distributed along the tubular portions of the elements. The dilatations, considered as prosecretory granules, increased in size as they drained the secretory material from the rest of the sacculotubular elements. Such prosecretory granules, large and irregular in shape, "peeled off" from the stacks of saccules with residual saccular or tubular structures still attached to them, some of the latter forming trans-tubular networks. The prosecretory granules detached from such membranous residues, condensed, and finally transformed into spherical secretion granules.     

Ultrastructural cytochemistry of the secretory granules of the hamster submandibular gland

Summary The ultrastructure and cytochemistry of the secretory granules of the male hamster submandibular salivary gland were studied. After fixation in glutaraldehyde followed by osmium tetroxide the granules exhibit a characteristic bipartite substructure, with an electron lucid crescenteric rim and a more dense central core. A differentiation into two regions of the granules could also be visualized in specimens primarily fixed in Millonig's osmium tetroxide or in potassium permanganate. The electron lucid peripheral portion of the membrane bounded secretory granules further displays a strong positive reaction after staining of ultrathin sections with the periodic acid-chromic acid-(PA-CrA)-silver technique. The strong periodate reactivity of the rim relative to the core, suggests a difference in mucin composition of the two granule regions. With the PA-CrA-silver staining technique a positive reaction was also observed within the Golgi apparatus of the acinar cells.     

Application of silver stains to gastric endocrine cells in plastic sections

Summary A simultaneous light and electron microscopic study of mouse gastric mucosa was made to determine whether the silver nitrate methenamine stain of Duk-Ho Lee could be used to stain gastric endocrine-like cells in plastic embedded tissue. Examination of consecutive thick and thin sections showed that this stain blackened the granules of the predominant type of endocrine-like cell present. Blackening of the granules with silver occured in tissue fixed in osmium tetroxide solution with or without dichromate salt or in tissue fixed in glutaraldehyde then treated with osmium. The intensity of staining was deepest in the osmium-dichromate fixed tissue, but the glutaraldehyde-osmium procedure gave less interference from diffuse silver impregnation and better preservation of detail for electron microscopy.     

Ultrastructural visualization of complex carbohydrates in epiphyseal cartilage with the tannic acid-metal salt methods

The present study has ultrastructurally applied the tannic acid-ferric chloride (TA-Fe) and the TA-uranyl acetate (TA-UA) methods to thin sections of glutaraldehyde-fixed, unosmicated embedded epiphyseal cartilage from rat tibiae to demonstrate complex carbohydrates. The strongest TA-Fe and TA-UA staining was observed after fixation of the specimens in glutaraldehyde containing TA. TA-Fe (pH 1.5) strongly stained matrix granules presumed to be proteoglycan monomers and chondrocyte secretory granules at various maturational stages but did not stain collagen fibrils and glycogen. TA-UA (pH 4.2) strongly stained matrix granules, intracellular glycogen, and chondrocyte secretory granules, and moderately stained collagen fibrils in the cartilage matrix. Ribosomes and nuclei were not stained above background staining with UA alone. In alpha-amylase-digested specimens, all TA-UA-reactive cytoplasmic glycogen was selectively removed. Testicular hyaluronidase digestion of specimens selectively removed TA-UA staining in matrix granules and all TA-Fe staining. When the pH of the UA solution was reduced to 1.5, TA-UA staining of glycogen and collagen was markedly decreased or absent, whereas staining of anionic sites was unaltered and significantly greater than with UA staining alone. Thus the TA-metal salt methods are pH dependent and allow differential intracellular and extracellular localization of complex carbohydrates in cartilage tissues at the electron microscope level.     

Ultrastructural visualization of sulphated complex carbohydrates in blood and epithelial cells with the high iron diamine procedure

Synopsis The high iron diamine (HID) method has been found to impart density at the ultrastructural level selectively to sites known to contain sulphated complex carbohydrates. Thus, immature primary granules in rabbit heterophils, immature précrystalloid granules in rabbit eosinophils, all granules of rabbit basophils, mouse and rat mast cells and the nucleoids of -granules of rabbit platelets were stained by HID. Granules of mast cells in rat cervical lymph node varied in the distribution pattern of the HID-reactive component. Mucous droplets within goblets of mouse colonic epithelial cells varied in HID reactivity. Sites known to contain sialomucin but no sulphates, such as mucous cells and apical plasmalemmae in mouse rectosigmoid colon, failed to stain with HID in contrast to their reactivity for dialysed iron at the ultrastructural level. The surface of mast cells and blood cells lacked affinity for HID, indicating that the dialysed iron binding at the surfaces can be attributed to neuraminic acid. HID proved more effective than dialysed iron in visualizing acid mucosubstance in precursor forms of the crystalloid granules in the eosinophil and in mast cell granules. Inclusion of 0.5% glycerol in the HID solution enhanced staining in mouse colon.     

Mitochondrial calcium of intact and mechanically damaged bone and cartilage cells studied with K-pyroantimonate.

Cellular cation was localized with K-pyroantimonate osmium fixation in whole fetal mouse metatarsal bones and in deliberately mechanically damaged specimens. X-ray microprobe analysis of ultrathin sections showed a positive correlation between the concentration of Ca (and Sb) and the amount of electron-dense precipitate. In non-damaged osteoblasts and growth-plate chondrocytes dense precipitate had accumulated along the plasmalemma and the mitochondrial membranes, whereas damaged cells showed the precipitate on round granules in the mitochondrial matrix but not on membranes. Intermediate stages between these two patterns were also found. In a non-calcifying tissue such as liver no membrane-bound precipitate was found in intact cells. However, damaged liver cells showed precipitate-containing mitochondrial granules similar to those in damaged bone cells, but only after incubation of the damaged tissue for 1 h in a Ca-containing balanced salt solution. Freezing of fresh whole bones in liquid N2 before fixation in K-pyroantimonate osmium did not change the precipitate pattern in the damaged cells, but in intact cells it produced a random distribution of precipitate unrelated to membranes. The results are compared with those obtained in other studies on the subcellular localization of calcium and in biochemical studies on membrane versus matrix loading in calcium-accumulating isolated mitochondria.     

The effect of chemical fixatives on cell walls of Bacillus subtilis.

Cell walls of Bacillus subtilis were treated with several chemical fixatives which are commonly used preparatory to electron microscopy; i.e., osmium tetroxide, formaldehyde, acrolein, crotonaldehyde, and glutaraldehyde. Dimensional analysis was performed on thin sections of fixed walls from plastic embeddings and, by means of the statistical technique of multiple comparisons, significant differences were found between wall thicknesses from the various fixations. These differences varied with the fixation time and the type of fixative used in the reaction. When compared to embedded walls which had been stained before fixation, the overall effect was a reduction in wall thickness which was attributed to fixative action and not to the embedding or staining processes. The reduction of wall thickness was even more apparent when dimensions of fixed walls were compared to published dimensions of both frozen sections and freeze-etch profiles. Since these fixatives bind to reactive sites within the wall fabric, a change in electrochemical charge density is effected which can be monitored in terms of heavy-metal-binding capacity. Most monoaldehyde fixatives and osmium tetroxide render the wall as reactive, or less reactive, to uranyl acetate as unfixed walls, whereas glutaraldehyde can significantly increase the binding capacity.     

Tannic acid-metal salt sequences for light and electron microscopic localization of complex carbohydrates.

Use of tannic acid (TA), in sequence with ferric chloride, uranyl acetate or gold chloride resulted in staining of selective but sometimes different sites in paraffin sections. TA-uranyl acetate of TA-ferric chloride stained sites rich in complex carbohydrates, wherease TA-gold chloride stained the collagen of various connective tissues different shades of red-purple to gray-black. Applied to epoxy-embedded thin sections of tissues fixed with glutaraldehyde and not post-osmicated, TA-uranyl acetate and TA-ferric chloride imparted density to subcellular sites known to contain a high concentration of mucosubstances, such as secretory granules and cisternae of the Golgi complex of certain cells. TA-gold chloride proved unsatisfactory for ultracytochemistry because of its tendency to form globular precipitates on thin sections. The effect of blockage procedures at the light microscopic level indicated that vicinal glycols are not required for binding of TA to tissue sites. Electrostatic forces were shown to be of minimal significance, whereas hydrogen bonding appeared to play a part in both TA-tissue and TA-metal binding mechanisms.     

Morphology and histochemistry of the adrenal medulla. I. Various types of primary catecholamine-storing cells in the mouse adrenal medulla.

Semithin sections (Araldite) of mouse adreno-medullary tissue were examined in the light microscope after perfusion fixation with glutaraldehyde, glutaraldehyde/formaldehyde or after freeze-drying followed by a treatment with hot formaldehyde gas. The following methods were employed: (i) aldehyde-induced fluorescence of catecholamines, (ii) Schmorl's ferric ferricyanide reaction, (iii) argentaffin reaction, and (iiii) staining with alkaline lead citrate followed by Timm's silver sulphide reaction. The correspondence of results obtained by the various methods was proven in consecutive sections or by successively applying different methods to identical sections. Four types of primary catecholamine-storing cells were identified. NA1 cells contain cytoplasmic granules up to 0.3 mum in diameter which stain black with ammoniacal silver and display a bright white to yellow fluorescence. NA2 cells show smaller cytoplasmic granules which stain brown with the argentaffin method and give white catecholamine fluorescence. NA3 cells appear yellow-earth after applying the argentaffin reaction and show greenish fluorescence. NA4 cells are hardly identified in the light microscope. These cells are significantly smaller than the above mentioned cells and characterized by a high nucleo-cytoplasmic ratio. They become straw coloured with ammoniacal silver and show greenish fluorescence. The argentaffin reaction was also used to identify these cells in semithin sections of glutaraldehyde/osmium tetroxide fixed material. The fine structure of the various noradrenalin-storing cells was studied in consecutive thin sections. NA1 cells were found to contain two populations of granules, the larger ones measuring between 300 and 350 nm, the smaller ones about 175 nm. The granules in NA2 cells correspond to this latter population (175 nm). NA3 cells contain an uniform granule population with a main diameter of 120 nm. The smallest granules are seen in NA4 cells being in the dimension of 80 nm. Granules in NA1 and NA2 cells show uniformly high density, whereas those in NA3 and NA4 cells display cores of varying density. Granules with moderately dense cores in NA3 and NA4 cells may represent partially emptied sites of noradrenalin storage or dopamin containing particles.     

Mitochondrial calcium of intact and mechanically damaged bone and cartilage cells studied with K-pyroantimonate

Summary Cellular cation was localized with K-pyroantimonate osmium fixation in whole fetal mouse metatarsal bones and in deliberately mechanically damaged specimens. X-ray microprobe analysis of ultrathin sections showed a positive correlation between the concentration of Ca (and Sb) and the amount of electron-dense precipitate.In non-damaged osteoblasts and growth-plate chondrocytes dense precipitate had accumulated along the plasmalemma and the mitochondrial membranes, whereas damaged cells showed the precipitate on round granules in the mitochondrial matrix but not on membranes. Intermediate stages between these two patterns were also found.In a non-calcifying tissue such as liver no membrane-bound precipitate was found in intact cells. However, damaged liver cells showed precipitate-containing mitochondrial granules similar to those in damaged bone cells, but only after incubation of the damaged tissue for l h in a Ca-containing balanced salt solution.Freezing of fresh whole bones in liquid N₂ before fixation in K-pyroantimonate osmium did not change the precipitate pattern in the damaged cells, but in intact cells it produced a random distribution of precipitate unrelated to membranes.The results are compared with those obtained in other studies on the subcellular localization of calcium and in biochemical studies on membrane versus matrix loading in calcium-accumulating isolated mitochondria.