Mapping gold-labeled IgE receptors on mast cells by scanning electron microscopy: receptor distributions revealed by silver enhancement, backscattered electron imaging, and digital image analysis

Immunogold labeling and silver enhancement techniques are widely used to determine density and distribution of cell membrane receptors by light and transmission electron microscopy. However, these techniques have not been widely used for receptor detection by scanning electron microscopy. We used antigen- or protein A-conjugated colloidal gold particles, together with silver enhancement, sequential secondary and back-scattered electron imaging (SEI and BEI), and digital image processing, to explore cell surface distribution of IgE-receptor complexes on RBL-2H3 cells, a rat leukemia line that provides a model for the study of mucosal mast cells. Cells were first incubated with a monoclonal antidinitrophenol IgE (anti-DNP-IgE) that binds with high affinity to cell surface IgE receptors. The resulting IgE-receptor complexes were cross-linked either with the multivalent antigen, DNP-BSA-gold, or with a polyclonal anti-IgE antibody. Antibody-treated cells were labeled after fixation with protein A-gold. Fixed, gold-labeled cell monolayers were silver enhanced (or not), dehydrated, critical point-dried, and coated with gold-palladium (for SEI analysis) or carbon (for combined SEI/BEI analysis). They were observed in an Hitachi S800 SEM equipped with a field emission tip and a Robinson backscattered electron detector. An image processor (MegaVision 1024XM) digitized images directly from the S800 microscope at 500-1000 line resolution. Silver enhancement significantly improves detection of gold particles in both SEI and BEI modes of SEM. On gold-palladium-coated samples, 20-nm particles are resolved by SEI after enhancement. BEI resolves 15-nm particles without enhancement and 5- or 10-nm particles are resolved by BEI on silver-enhanced, carbon-coated samples. Neither BEI nor SEI alone can yield high resolution topographical maps of receptor distribution (BEI forms images on the basis of atomic number contrast which reveals gold but not surface features). Image analysis techniques were therefore introduced to digitize, enhance, and process BEI and SEI images of the same field of view. The resulting high-contrast, high-resolution images were superimposed, yielding well-resolved maps of the distribution of antigen-IgE-receptor complexes on the surface of RBL-2H3 mast cells. The maps are stored in digital form, as required for computer-based quantitative morphometric analyses. These techniques of silver enhancement, combined BEI/SEI imaging, and digital image analysis can be applied to analyze density and distribution of any gold-labeled ligand on its target cell.     

Binding and internalization of gold-labeled IFN-gamma by human Raji cells

Binding and internalization of gold-labeled IFN-gamma (IFN-gamma/Au) by human Raji cells was examined by scanning and transmission electron microscopy. For SEM, visualization of gold particles was enhanced by the silver enhancement technique and by backscattered electron imaging. Binding studies revealed distinct labeling of microvilli-bearing cells after incubation with at least 10 U/ml IFN-gamma/Au, whereas cells with a smooth surface showed substantially lower labeling. After application of higher IFN-gamma (greater than 200 U/ml) concentrations, labeling intensity remained constant, which is consistent with the concentration of radiolabeled IFN-gamma required for saturating receptors on Raji cells. The specificity of IFN-gamma/Au binding was demonstrated by complete displacement with unlabeled IFN-gamma and by partial inhibition of labeling with a monoclonal anti-IFN-gamma R antibody. Thus, colloidal gold represents a valuable tag for visualizing the interaction of IFN-gamma with its receptor. Internalization of IFN-gamma/Au was initiated by accumulation of gold particles in coated pits which occurred within 10 min after warming of Raji cells. Additional incubation at 37 degrees C (up to 2 h) led to the appearance of gold particles in endocytic vesicles and lysosomes. Thus, our studies indicate that IFN-gamma/Au enters the Raji cells via the typical endocytotic pathway.     

Immunogold-silver staining method for light and electron microscopic detection of lymphocyte cell surface antigens with monoclonal antibodies

We used the immunogold-silver staining method (IGSS) for detection of lymphocyte cell surface antigens with monoclonal antibodies in light and electron microscopy and compared this procedure with the immunogold staining method. Two different sizes of colloidal gold particles (5 nm and 15 nm) were used in this study. Immunolabeling on cell surfaces was visualized as fine granules only by IGSS in light microscopy. The labeling density (silver-gold complexes/cell) and diameters of silver-enhanced gold particles on cell surfaces were examined by electron microscopy. Labeling density was influenced not by the enhancement time of the physical developer but by the size of the gold particles. However, the development of shells of silver-enhanced gold particles correlated with the enhancement time of the physical developer rather than the size of the colloidal gold particles. Five-nm gold particles enhanced with the physical developer for 3 min were considered optimal for this IGSS method because of reduced background staining and high specific staining in the cell suspensions in sheep lymph. Moreover, this method may make it possible to show the ultrastructure of identical positive cells detected in 1-micron sections counterstained with toluidine blue by electron microscopy, in addition to the percentage of positive cells by light microscopy.     

High-resolution immunogold localization of Giardia cyst wall antigens using field emission SEM with secondary and backscatter electron imaging

We describe here the ultrastructural localization of Giardia cyst antigens in the filaments associated with the outer portion of intact cysts and on developing cyst wall filaments in encysting trophozoites. Post-embedding immunogold labeling of thin sections of intact Giardia cysts with polyclonal and monoclonal antibodies specific for cyst wall antigens (major protein bands of approximately 29, 75, 88, and 102 KD on Western blots) showed strong labeling of the filamentous cyst wall, whereas no labeling was seen on the membranous portion. High-resolution field emission scanning electron microscopy (FESEM) of Giardia cysts revealed that the cyst wall-specific polyclonal rabbit antisera and monoclonal mouse antibody produced gold labeling of 20-nm filaments in the cyst wall as detected with secondary electron imaging (SEI) and backscatter electron imaging (BEI) at 10 kV, despite coating of the cells with platinum by ion sputtering. FESEM studies of encysting Giardia trophozoites demonstrated that immunostaining with antibodies to cyst wall antigens produced colloidal gold labeling of developing cyst wall filaments on the cell surface; however, the intervening membrane domains were unlabeled. Substitution of normal serum for cyst wall-specific antibodies, or preabsorption of specific antibodies with Giardia cysts, eliminated immunolabeling of the filaments.     

Preembedding colloidal gold immunostaining of hypothalamic neurons: light and electron microscopic localization of beta-endorphin-immunoreactive perikarya

A preembedding immunogold staining (IGS) procedure was developed to identify beta-endorphin/adrenocorticotropic hormone immunoreactive neurons at the light and electron microscopic levels. Colchicine-treated rats were perfused with Nakane's periodate-lysine-paraformaldehyde fixative. Vibratome sections were incubated in primary antisera followed by goat anti-rabbit immunoglobulin G coupled to 16 nm colloidal gold, and, in some cases, rabbit immunoglobulin G coupled to gold. The appearance to pink to light red perikarya, corresponding to colloidal gold deposition at antigenic sites, was monitored under the light microscope. Positive cell bodies in the arcuate region sometimes extended lateral to the nucleus. Only proximal portions of neuronal processes were stained. At the ultrastructural level, colloidal gold labeled the periphery of 90-110 nm dense neurosecretory granules in the perikaryal cytoplasm and a few proximal axons. Clusters of gold particles, appearing free in the neuroplasm, actually labeled secretory granules in adjacent thin sections. Granules associated with the Golgi apparatus were not stained. Colloidal gold labeling of mature beta-endorphin granules, but not progranules, in rat hypothalamic neurons was confirmed using the peroxidase-antiperoxidase technique. The results correlate well with data on the intracellular processing of pro-opiomelanocortin in pituitary cells and prepropressophysin in the paraventricular nucleus. These data demonstrate the first application of the preembedding colloidal gold staining method for the identification of intracellular antigens within the central nervous system. The IGS method provides a definitive marker for single or double labeling of nervous tissue at both the light and electron microscopic levels.     

A new method of preparing gold probes for multiple-labeling cytochemistry

A new method is described for preparing colloidal gold particles in any size between approximately 3 and 17 nm for electron microscopy. The gold particles are homogeneous in size (homodisperse). When bound to various proteins (e.g. IgG and protein A), the complexes were stable for long periods and suitable for affinity cytochemistry. We demonstrated the usefulness of the new gold probes bound to protein A for multiple labeling in a current immunocytochemical study on receptor mediated transport of IgA in human duodenal crypt cells. Other gold-protein complexes were useful for studying macromolecular arrangements at high resolution.     

A colloidal gold labeling technique for the direct determination of the surface area of eukaryotic cells

We have developed a colloidal gold labeling technique for the direct quantitation of the cell surface area. The method is based on coating the cell surface with [195Au] colloidal gold-protein complexes followed by morphometric determination of the labeling density (gold particles/micron2 cell surface) and radiometric determination of the total number of gold particles bound per cell. The ratio of both values directly gives the cell surface area. The accuracy of the method was shown using Staphylococcus aureus cells as a model system, where the cell surface area determined with our assay (4.0 microns2) corresponded well to the value calculated from the radius of the cells (3.6 microns2). In a more complex model system J-774 mouse macrophages were labeled with different amounts of [195Au] gold-protein complexes to show that the assay is independent of the degree of saturation of the cell surface binding sites. Both high (135 Au/microns2) and low (65 Au/microns2) labeling densities resulted in a surface area of about 1200 microns2. The technique finally was applied to L-929 fibroblasts to determine the increase of the cell surface area when the cells change from a spherical to a flat monolayer state. We found that the cell surface area increased 3-fold during the spreading process. The results show that the colloidal gold labeling technique allows the direct determination of the surface area of complex eukaryotic cells. The technique is suitable for the quantitation of changes in the surface architecture known to occur in different functional states of eukaryotic cells.     

Preferential association of glycoproteins to the euchromatin regions of cross-fractured nuclei is revealed by fracture-label

We used fracture-label to establish ultrastructural localization of glycoproteins in cross-fractured nuclei of duodenal columnar and exocrine pancreatic cells. Mannose residues were detected in cell nuclei by labeling freeze-fractured tissues with concanavalin A-horseradish peroxidase X colloidal gold (Con A-HRP X CG) or direct concanavalin A X colloidal gold (Con A X CG); fucose residues were detected with Ulex Europaeus I X colloidal gold (UEA I X CG) markers. Areas of the three main intranuclear compartments (euchromatin, heterochromatin, and nucleolus) exposed by freeze-fracture were determined by automated image analysis. Colloidal gold particles bound to each nuclear subcompartment were counted and the results expressed in number of colloidal gold particles per square micrometer +/- SEM. Duodenal and pancreatic tissues fractured and labeled with Con A-HRP X CG complex or direct Con A X CG conjugates showed that the vast majority of Con A binding sites was confined to euchromatin regions with only sparse labeling of the heterochromatin and nucleolus. UEA I labeling of duodenal columnar cells showed that colloidal gold particles were almost exclusively confined to cross-fractured areas where euchromatin is exposed. Trypsinization of the fractured tissues before labeling with Con A and UEA I abolished 95-100% of the original label. Our results show that, within the nucleoplasm, mannose and fucose are residues of glycoproteins preferentially located within the regions of euchromatin.     

Enhancement of Antigenic Site Detection with Gold Labeled Secondary and Tertiary Antibodies Using the Immunogold-Silver Staining Method

We report a modification of the immunogold-silver staining method (IGSS) for localizing hepatic phosphoenolpyruvate carboxykinase (PEPCK) in tissue sections, and we compare the efficacy of localizing the primary antibody with either a 5 nm gold labeled secondary antibody or 5 nm gold labeled secondary and tertiary antibodies. Light microscope examination of 10 μm frozen sections demonstrated that the use of combined secondary and tertiary gold labeled antibodies was superior to using a secondary gold labeled antibody alone. The increased labeling density (number of colloidal gold particles/antigenic site/cell) achieved by combined gold labeled antibodies was confirmed by electron microscopy. The increased labeling density resulted in a two-thirds reduction in the time needed for the IGSS physical development of the silver shells and less background. We achieved intense specific staining of hepatocytes expressing PEPCK while minimizing background staining. The use of combined secondary and tertiary gold labeled antibodies enhances the signal-to-noise ratio, achieves high resolution and is a suitable method for use in both light and electron microscopy.     

Immunoelectron microscopic double labeling of alkaline phosphatase and penicillinase with colloidal gold in frozen thin sections of Bacillus licheniformis 749/C.

The subcellular distribution of alkaline phosphatase and penicillinase was determined by double labeling frozen thin sections of Bacillus licheniformis 749/C with colloidal gold-immunoglobulin G (IgG). Antipenicillinase and anti-alkaline phosphatase antibodies were used to prepare complexes with 5- and 15-nm colloidal gold particles, respectively. The character of the labeling of membrane-bound alkaline phosphatase and penicillinase was different: the immunolabels for alkaline phosphatase (15-nm particles) were bound to a few sites at the inner surface of the plasma membrane, and the gold particles formed clusters of various sizes at the binding sites; the immunolabels for penicillinase (5-nm particles), on the other hand, were bound to the plasma membrane in a dispersed and random fashion. In the cytoplasm, immunolabels for both proteins were distributed randomly, and the character of their binding was similar. The labeling was specific: pretreating the frozen thin sections with different concentrations of anti-alkaline phosphatase or penicillinase blocked the binding of the immunolabel prepared with the same antibody. Binding could be fully blocked by pretreatment with 800 micrograms of either antibody per ml.     

Intracellular location of the autolytic N-acetylmuramyl-L-alanine amidase in Bacillus subtilis 168 and in an autolysis-deficient mutant by immunoelectron microscopy.

Antisera against purified autolytic N-acetylmuramyl-L-alanine amidase from Bacillus subtilis 168 were prepared in rabbits. They neutralized the enzymatic action of the purified amidase acting on isolated sodium dodecyl sulfate (SDS)-treated walls from the same organism. They also inhibited the lysis of native walls, but only after the walls lysed partially. Amidase adsorbed to insoluble walls still combined with antibody. Antisera did not stop the lysis of whole cells. Lowicryl HM20 sections of both strain 168 and its autolytic mutant strain FJ6 were prepared by the progressive-lowering-of-temperature technique, immunolabeled with the antisera, and visualized with colloidal gold particles as markers. The highest concentration of gold particles seemed to be in the septa of dividing cells, followed by the side walls. There was some labeling of the cytoplasm. Adsorption of sera with SDS-treated walls reduced the overall labeling of sections considerably but did not alter the relative intracellular distribution of particles. The results for strains 168 and FJ6 were similar. Labeling of SDS-treated walls unexpectedly revealed the presence of a wall-bound amidase fraction.     

Immunoelectron microscopic localization of thiolases, beta-oxidation enzymes of an n-alkane-utilizable yeast, Candida tropicalis.

The location of acetoacetyl-CoA thiolase (T-I) and 3-ketoacyl-CoA thiolase (T-III), enzymes of the fatty acid beta-oxidation system, was studied in n-alkane-grown Candida tropicalis cells by immunoelectron microscopy using a post-embedding method with colloidal gold conjugated IgG. The deposition of gold particles for T-I was detected in the microbodies and cytoplasm and that of gold particles for T-III specifically in the microbodies. The double labeling technique confirmed that T-I and T-III occurred concurrently in a microbody and T-I also in cytoplasm. These results were consistent with the biochemical data based on subcellular fractionation and indicated that the yeast beta-oxidation system operates efficiently only in the microbodies.     

Nanovid tracking: a new automatic method for the study of mobility in living cells based on colloidal gold and video microscopy.

We describe a new automatic technique for the study of intracellular mobility. It is based on the visualization of colloidal gold particles by video-enhanced contrast light microscopy (nanometer video microscopy) combined with modern tracking algorithms and image processing hardware. The approach can be used for determining the complete statistics of saltatory motility of a large number of individual moving markers. Complete distributions of jump time, jump velocity, stop time, and orientation can be generated. We also show that this method allows one to study the characteristics of random motion in the cytoplasm of living cells or on cell membranes. The concept is illustrated by two studies. First we present the motility of colloidal gold in an in vitro system of microtubules and a protein extract containing a kinesin-like factor. The algorithm is thoroughly tested by manual tracking of the videotapes. The second study involves the motion of gold particles microinjected in the cytoplasm of PTK-2 cells. Here the results are compared to a study using the spreading of colloidal gold particles after microinjection.     

Ultrastructural localization of nuclei acids by the use of enzyme-gold complexes

A cytochemical technique for the ultrastructural localization of substrates using enzyme-gold complexes is reported. RNase A and DNase I have been labeled with gold particles. The RNase-gold and dNase-gold complexes obtained were applied on thin sections of glutaraldehyde-fixed and Epon-embedded tissues. Different cellular compartments were labeled by these enzyme-gold complexes. Using the RNase-gold complex the rough endoplasmic reticulum appeared decorated with gold particles. The gold marker was also present over the nucleus, especially over the nucleolus; mitochondria were weakly labeled. Using the DNase-gold complex, gold particles were concentrated over the euchromatin of the nucleus and the mitochondria. The heterochromatin and the nucleolus showed a less intense labeling. For both enzyme-gold complexes, the Golgi area, the secretory granules and the extracellular space appeared free of label. In those control conditions where the substrates were added to the enzyme-gold complexes a major reduction in the labeling was observed. A quantitative evaluation of the labeling was performed. This evaluation confirmed the qualitative observations and the marked reduction of labeling occurring under the control conditions. The combination of the specificity of the enzyme-substrate interactions with the size and electron density of the gold particles and the good ultrastructural preservation of the tissues resulted in a very specific labeling with high resolution. These results demonstrate the possibility of detecting substrates by means of enzyme-gold complexes at the electron microscope level.     

Intracellular transport of transferrin- and asialoorosomucoid-colloidal gold conjugates to lysosomes after receptor-mediated endocytosis

Proteins coupled to colloidal gold particles have been widely used to visualize the uptake and intracellular transport of specific ligands by receptor-mediated endocytosis. The intracellular route of lysosome-directed ligands such as asialoglycoproteins (ASGP) are apparently unaltered by conjugation to gold, but the pathway of transferrin, a ligand that normally recycles to the cell surface, was reported to be altered by conjugation to 15-20 nm gold. In this study, we sought to determine whether a smaller transferrin-gold probe would recycle, and whether it might enter the same endosomal and lysosomal compartments as does a larger, lysosome-directed ASGP gold probe by visualizing their simultaneous uptake in human hepatoma (HepG2) cells. In the same cells, endocytosis of fluid-phase protein was followed using the soluble tracer native ferritin; lysosomal compartments were identified by acid phosphatase cytochemistry; and cell surfaces were labeled with ruthenium red or cationized ferritin. During the first 10 min of uptake at 37 degrees C, specific receptor-bound ferrotransferrin (FeTf)-8 nm gold and asialoorosomucoid (ASOR)-20 nm gold were clustered together in coated pits and entered the same coated vesicles, smooth vesicles, and tubules in the peripheral cytoplasm. At later times, however, transferrin-gold did not return to the cell surface; unlike native transferrin, this gold probe accompanied ASOR-gold into multivesicular bodies (MVB). The MVBs that contained probes were at first devoid of acid phosphatase activity, but at 30 min, enzyme activity was detected in a few MVBs. Native ferritin was present, along with gold probes, in all compartments of the endocytic pathway. We conclude that the normal intracellular pathway of transferrin is altered by its association with a colloidal gold particle.     

Visualization of binding and receptor-mediated uptake of low density lipoproteins by human endothelial cells

Low density lipoproteins (LDL) were conjugated to colloidal gold for investigation of the ultrastructural aspects of binding and receptor-mediated internalization of LDL by cultured endothelial cells from the human umbilical artery and vein. The number of LDL receptors was increased by preincubation in lipoprotein-depleted serum. When the cells were incubated with LDL-gold particles for 2 h at 4 degrees C, the complexes were found in coated pits as well as in clusters attached to the plasma membrane. Small vesicles containing a few LDL-gold complexes appeared in the cytoplasm close to the plasma membrane when the cells were incubated with the conjugate for 5 min at 37 degrees C. After 15 min at 37 degrees C, larger vesicles with a pale matrix and membrane-orientated LDL-gold complexes were seen. After incubation for 30 min at 37 degrees C, colloidal gold particles were present in dense bodies. Quantification of the binding of LDL-gold complexes to the plasma membrane at 4 degrees C showed no differences between arterial and venous endothelial cells.     

Visualization and rapid quantification of autoradiographic labeling in scanning electron microscopy applied to localization of receptor sites on the surface of whole cells.

The use of backscattered electron imaging (BEI) as a routine procedure for examining autoradiographic reactions in scanning electron microscopy (SEM) is described. This technique allows the determination of the number of receptor sites occupied by 125I-epidermal growth factor (EGF) on whole cells. The effect of 1.25 dihydroxyvitamin D3 (1,25 (OH)2D3) on the number of epidermal growth factor receptors (EGF-R) in the BT 20 human mammary carcinoma cell line (which is known to possess a very high number of EGF-R) has been evaluated with this method. To compare the silver grain density over the cells (controls and 1,25 (OH)2D3-treated cells) we used an image analysis system Quantimet 900. The results were compared with those of a previous study using transmission electron microscopy (TEM). This study confirmed the results obtained with TEM and showed the even distribution of receptors sites on a single cell and a large difference in the number of receptor sites from one cell to another. The use of BEI to visualize the autoradiographic reaction in SEM allowed the examination of a large surface with good contrast and resolution and eliminated artefacts not corresponding to the silver grains. It gave new information not delivered by quantitative TEM autoradiography and was easier and faster to use. The efficient use of SEM autoradiography combined with BEI could facilitate whole area distribution mapping of radioactive labeling.     

Effective diameters of protein A-gold and goat anti-rabbit-gold conjugates visualized by field emission scanning electron microscopy.

High-voltage (15-30 kV) field emission scanning electron microscopy (FESEM) was used to evaluate the effects of gold particle size and protein concentration on the formation of protein-gold complexes. Six colloidal gold sols were prepared, ranging in diameter from 7.6 to 39.8 nm. The minimal protecting amounts (m.p.a.) of protein A and goat anti-rabbit antibody (GAR) were experimentally determined. Gold particles were conjugated at the m.p.a., one half the m.p.a., and ten times the m.p.a. for both proteins, and protein-gold complexes prepared for FESEM. The smallest colloidal gold particles required the most protein per milliliter of gold suspension for stabilization. Transmission electron microscopy was found to be the preferred method for accurate sizing of gold particles, whereas FESEM of protein-gold complexes permitted visualization of a protein halo around a spherical gold core. Protein halo width varied significantly with changes in gold particle size. Measurements of protein halos indicated that conjugation with the m.p.a. of protein A resulted in the thickest protein layers for all gold sizes. GAR conjugation with the m.p.a. again produced the thickest protein layers. However, GAR halos were significantly smaller than those obtained with protein A conjugation. The proteins used showed similar adsorption patterns for the larger gold particles. For smaller gold particles, proteins may act differently, and these complexes should be further characterized by low-voltage FESEM.     

Use of colloidal gold and ruthenium red in stereo high-voltage electron microscopic study of Con A-binding sites in mouse macrophages

Concanavalin A (Con A)-binding sites were labeled with colloidal gold (CG), stained with ruthenium red, and observed under a high-voltage electron microscope. Mouse peritoneal macrophages were labeled by the indirect Con A/CG labeling method at 0 degree C. After washing, some of the cells were incubated in phosphate-buffered saline (PBS) at 37 degrees C. The specimens were then stained with ruthenium red, to enhance the contrast of the cell surface, and embedded in Epon. Sections (0.3 approximately 3 micron thick) were cut and examined by high-voltage electron microscopy at accelerating voltages of 200 approximately 1,000 kV. Staining with ruthenium red provided a strong contrast of the cell surface and the invaginating tubules beneath it against the cytoplasm; in thick sections, both of them were clearly seen by stereomicroscopy. CG particles which represented Con A-binding sites were also sufficiently electron dense to be recognized by high-voltage electron microscopy of thick sections. The two- and three-dimensional distribution of CG particles on the ruthenium-red-positive cell surface was clearly visualized. At 0 degree C, Con A-binding sites were randomly distributed on the cell surface. The redistribution and endocytosis of Con A-binding sites were seen at 37 degrees C. The three-dimensional organization of membrane invagination, which represented the process of endocytosis, was clearly seen by stereomicroscopy. The combination of CG labeling and ruthenium red staining is a useful method for high-voltage electron microscopic analysis of the two- and three-dimensional distribution of CG-labeled ligands on the cell surface in thick sections.     

Nuclear exchange of the U1 and U2 snRNP-specific proteins

The snRNP particles include a set of common core snRNP proteins and snRNP specific proteins. In rodent cells the common core proteins are the B, D, D', E, F and G proteins in a suggested stoichiometry of B2D'2D2EFG. The additional U1- and U2-specific proteins are the 70-kD, A and C proteins and the A' and B" proteins, respectively. Previous cell fractionation and kinetic analysis demonstrated the snRNP core proteins are stored in the cytoplasm in large partially assembled snRNA-free intermediates that assemble with newly synthesized snRNAs during their transient appearance in the cytoplasm (Sauterer, R. A., R. J. Feeney, and G. W. Zieve. 1988. Exp. Cell Res. 176:344-359). This report investigates the assembly and intracellular distribution of the U1 and U2 snRNP-specific proteins. Cell enucleation and aqueous cell fractionation are used to prepare nuclear and cytoplasmic fractions and the U1- and U2-specific proteins are identified by isotopic labeling and immunoprecipitation or by immunoblotting with specific autoimmune antisera. The A, C, and A' proteins are found both assembled into mature nuclear snRNP particles and in unassembled pools in the nucleus that exchange with the assembled snRNP particles. The unassembled proteins leak from isolated nuclei prepared by detergent extraction. The unassembled A' protein sediments at 4S-6S in structures that may be multimers. The 70-kD and B" proteins are fully assembled with snRNP particles which do not leak from isolated nuclei. The kinetic studies suggest that the B" protein assembles with the U2 particle in the cytoplasm before it enters the nucleus.