Visualization of the intact endoplasmic reticulum by immunofluorescence with antibodies to the major ER glycoprotein, endoplasmin

Transcription

1 Visualization of the intact endoplasmic reticulum by immunofluorescence with antibodies to the major ER glycoprotein, endoplasmin G. L. E. KOCH, D. R. J. MACER and M. J. SMITH Medical Research Council Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, England Summary Antibodies to endoplasmin were used to examine the morphology of the endoplasmic reticulum (ER) by immunofluorescence on permeabilized plasmacytoma and fibroblastoid cells. In unfixed cells, permeabilization led to a pronounced vesiculation of the ER. Therefore cells were first fixed lightly prior to permeabilization with detergent. Fibroblastoid cells gave a characteristic reticular pattern surrounding the nucleus with clear staining of the nuclear membrane. Plasmacytoma cells, in the conventional fluorescence microscope, gave a cisternae-like pattern. Optical sectioning with a confocal scanning laser microscope gave a distinct pattern of concentric cisternae similar to those obtained with transmission electron microscopy on cell sections. The overall morphology of the ER in such cells could be revealed by serial optical sectioning. Evidence was obtained that the ER does not undergo extensive vesiculation during mitosis in plasmacytoma cells. Using anti-endoplasmin immunofluorescence monitoring, conditions were developed for the retention of ER morphology in unfixed, permeabilized cells. These studies illustrate the value of endoplasmin as a general marker for the analysis of ER morphology in different types of cells by immunofluorescence microscopy. Key words: endoplasmic reticulum, ER visualization, ER immunofluorescence, endoplasmin antibodies. Introduction The endoplasmic reticulum (ER) consists of a complex set of membrane lamellae and tubules, which is the site at which many major biosynthetic processes such as protein secretion, membrane assembly and organelle biosynthesis are initiated (Porter et al. 1945; Palade & Porter, 1954; Porter, 1953; Krstic, 1979; Alberts et al. 1983). Although there have been striking advances in the understanding of the biochemistry of these processes, several fundamental questions remain about the relationship between ER structure and function. These include: (1) the possibility of segmentation beyond the morphologically distinct rough and smooth ER; (2) the distribution of resident luminal as well as secretory proteins within the ER; (3) identification of the sites from which material, presumably in a vesicular form, leaves the ER; (4) the fate of the ER itself during cell division and other perturbations. Examination of many of these questions remains difficult because they require visualization of the entire organelle in situ, rather Journal of Cell Science 87, (1987) Printed in Great Britain The Company of Biologists Limited 1987 than parts of it as is possible by electron microscopy of thin sections. As pointed out previously (Terasaki et al. 1984), the original whole-mount electron micrographs of Porter and co-workers (see above) remain among the best examples of ER visualization in situ. Since then, the most significant advance has probably been the studies of Terasaki et al. (1984), which have shown that the entire ER can be rendered visible by fluorescence microscopy with certain cationic dyes that appear to partition preferentially into the ER. However, these reagents are not completely specific for ER and can lead to artefactual staining of other membrane organelles. Louvard et al. (1982) have shown that it is possible to use antibodies against putative ER-specific antigens to visualize the ER by immunofluorescence microscopy. However, it is not possible to assess the general applicability of the approach used, since it employed a polyspecific antibody, directed against a large number of undefined protein antigens, which was dependent on 535

2 extensive absorption to render it specific towards the ER. A general approach to the visualization of the ER by immunofluorescence microscopy should employ markers that have been formally demonstrated to be specifically located in the ER, are known to be present in a wide variety of cell types, are sufficiently abundant to facilitate access in relatively intact cells and towards which strong monospecific antibodies can be produced. Recently it was shown that the ER of all vertebrate cells contains a major glycoprotein that is undetectable in any other organelle and towards which it is possible to obtain monospecific antibodies for immunolocalization (Koch et al. 1986). The glycoprotein, which is called endoplasmin, therefore satisfies all the criteria mentioned above as a general marker for the visualization of ER in cells by immunofluorescence microscopy. In this study, the suitability of endoplasmin as an ER marker in immunofluorescence microscopy on both flattened as well as large, rounded cells such as plasma cells was examined. The results indicate that endoplasmin is indeed suitable for this purpose, particularly when used in combination with a newly developed confocal scanning microscope. The morphology of the ER has been characterized under a variety of conditions by this approach, and conditions developed for the retention of ER morphology in a 'cell-free' system. Materials and methods Cells All cells were grown in RPM medium (Gibco) with 10 % foetal calf serum, 100 units per ml penicillin/streptomycin and 4mM-L-glutamine. NIH-3T3 cells were grown on tissue-culture plastic substrate and harvested with trypsin/edta (Todaro & Green, 1963). MOPC-315 plasmacytoma cells were grown in suspension. NIH-3T3 cells were prepared for microscopy by placing a droplet of a 1X10 7 cells ml" 1 suspension in growth medium on a clean glass slide. After 30min at 37 C for attachment, slides were incubated in growth medium at 37"C to permit spreading to occur. This was usually complete in about 2h. MOPC-315 cells were suspended in phosphate-buffered saline (PBS) (Koch & Smith, 1986) and a droplet was placed on a polylysine-coated clean glass slide. Cells became attached in about IS min. Permeabilizing of cells for immunofluorescence labelling Unfixed cells were treated with O'Ol % saponin in PBS for 2 min and then fixed with 3-5% formaldehyde in PBS for 15 min. Cells were prefixed with 3-5% formaldehyde in PBS for 15 min. Final permeabilization was effected with 0-2% saponin in PBS for 15 min. Antibodies and lectin Affinity-purified antibodies to endoplasmin were prepared as described previously (Kochef al. 1986). Fluorescein-labelled goat anti-rabbit immunoglobulin (Ig) was purchased from Sigma and used at 1:20 dilution. Fluorescein-labelled wheat-germ agglutinium and concanavalin A (ConA) were purchased from Sigma. Immunofluorescence labelling Cells that had been treated with 02% saponin (see above) after fixing were treated with antibody for 15 min at room temperature, washed extensively with 0-2% saponin/pbs and developed with fluorescent second antibody for 15 min at room temperature. After extensive washing, samples were mounted in 90% glycerol with 1 % phenylene diamine, and sealed with varnish. Controls were carried out without the first antibody layer. Staining with fluorescent wheat-germ agglutinin was as above. Immunofluorescence microscopy Samples were viewed in a Zeiss Universal epifluorescence microscope equipped with a narrow band fluorescein excitation filter. Photographs were taken with Kodak Tri-X film. Confocal laser scanning fluorescence microscopy This was carried out on an instrument (the MRC system confocal fluorescence microscope) designed by Dr J. White, details of which will be described elsewhere (White, Amos & Fordham, unpublished). The instrument employs a scanning beam of laser light and confocal optics to illuminate and analyse relatively thin regions of the specimen. This provides considerable improvements in resolution and permits serial optical sectioning. Results ER vesiculation in permeabilized cells Previous studies have shown that cells treated with the detergent saponin became depleted of their cytoplasmic contents but retain all the endoplasmin, suggesting that the ER is still intact (Koch et al. 1986). Such 'shells' could be ideal for analysis of the ER. However, immunofluorescence studies with anti-endoplasmin (Fig. 1) show that although there is considerable intracellular staining this is located in vesicular structures, suggesting that significant distortion of the ER has occurred upon permeabilization. Electron microscopy of the shells (Fig. 1) confirms that the ER, identified by the attached ribosomal particles, has indeed dilated and vesiculated in the permeabilized cells. Thus the permeabilized cells do not appear suitable for analysis of ER morphology. Anti-endoplasmin staining of fixed, permeabilized cells To overcome the distortion caused by direct permeabilization of cells, a light-fixation step was introduced. It was found that fixation with formaldehyde 536 G. L. E. Koch et al.

3 Fig. 1. Immunofluorescence with anti-endoplasmin on unfixed cells permeabilized with saponin. A. 3T3 cells were treated with O'Ol % saponin/pbs, fixed with formaldehyde and stained with anti-endoplasmin/fitc-goat anti-rabbit Ig as described in Materials and methods. Note the vesicular pattern of staining. Bar, S jlm. B. Cells prepared as above were prepared for electron microscopy as described in Materials and methods. Note the dilation of ribosome-lined rough ER and the apparent dispersal of the cisternae of the Golgi apparatus. Bar, yielded samples that possessed the correct balance of stability and permeability to permit access of antibodies to the interior of the ER. In such preparations, antiendoplasmin gave a characteristic reticular pattern around the nucleus, with clear staining of the nuclear membrane (Fig. 2). The latter is characteristic of endoplasmic reticulum (Krstic, 1979), and is a useful indicator of the permeabilization procedure. The staining pattern with anti-endoplasmin is clearly different from that obtained with the Golgi-indicator wheat-germ agglutinin (Fig. 2), which gives a tight polar cap at one pole of the nucleus and no staining of the nuclear membrane. Although pre-fixation was also useful for the examination of round cells such as plasmacytoma cells, which yielded a pattern suggesting an intracellular membranous reticulum (Fig. 3), the thickness of the cells interfered with the visualization of the overall pattern of fluorescence. However, examination of the same preparation by confocal laser scanning fluorescence microscopy (Fig. 3C) revealed the striking concentric pattern characteristic of the ER in plasma cells. The resolution of the technique was sufficient to permit the separation of individual cisternae. Because the optics of the system permit relatively thin (0-7/Urn) sections to be examined, there is a dramatic increase in signal to ER visualization by immunofluorescence 537

4 noise above that obtained with the conventional microscope. This also permits the examination of the general morphology of a complex organelle like ER by serial optical sectioning microscopy (Fig. 4). Although the analysis is not complete, it illustrates the feasibility of using this approach for the visualization of ER morphology by immunofluorescence microscopy using antibodies to an ER marker such as endoplasmin. Serial optical sectioning of stained plasmacytoma cells, which were identified as being in mid-division by their general morpholog}' and the absence of nucleus, revealed an apparently continuous structure with no signs of extensive vesiculation (Fig. 5). Preservation of ER morphology in permeabilized cells The dilation and vesiculation of the ER that occurs when cells are permeabilized with saponin (Fig. 1) Fig. 2. Immunofluorescence with anti-endoplasmin on fixed cells permeabilized with saponin. A,B. 3T3 cells were fixed with formaldehyde, permeabilized with 02% saponin/pbs and stained with anti-endoplasmin as described in Materials and methods. C. Cells prepared as above stained with FITC-wheat-germ agglutinin. Note the perinuclear cap typical of the Golgi apparatus. Bar, 5;um. Fig. 3. Immunofluorescence with anti-endoplasmin on plasmacytoma cells. MOPC-315 cells were fixed with formaldehyde followed by permeabilizing with 0-2% saponin/pbs. Labelling was with anti-endoplasmin and FITC-anti-rabbit Ig as described in Materials and methods. A,B. The same field of cells examined by phasecontrast and epifluorescence, respectively. Bar, 10 Jim. C. The same sample examined by confocal laser scanning fluorescence microscopy. Magnification was not measured directly for samples analysed by this technique. 538 G. L. E. Koch et al.

5 suggested that osmotic effects might be involved. Therefore permeabilization with saponin was carried out in the presence of a relatively impermeant sugar, sucrose, to overcome these putative osmotic effects. The results (Fig. 6) show that there is indeed a significant improvement, in that the ER in such cells retains the morphology observed with pre-fixed cells. A particular advantage of this procedure is that it improves the general accessibility of the ER to antibodies and thereby provides improved visualization by immunofluorescence. Comparison between anti-endoplasmin and other fluorescent probes for ER membranes A comparison was carried out between anti-endoplasmin, cyanine dyes (Terasaki et al. 1984) and fluorescent ConA for the visualization of the ER in plasmacytoma cells (Fig. 7). Although the latter do reveal the ER, they are clearly less specific, both also reacting with the plasma membrane and probably with other non-er membranes. Therefore, although useful reagents, cyanine dyes and fluorescent ConA are of less value than anti-endoplasmin for the specific visualization of the ER. Fig. 7 also shows examples of cells that were double-labelled with anti-endoplasmin and antitubulin. Cells that are clearly mitotic, as shown by the presence of well-developed spindles, show essentially intact ER, confirming the above-mentioned studies (Fig. 5), which showed no evidence for gross vesiculation in mitotic cells. Discussion Fig. 4. Serial optical sectioning of the endoplasmic reticulum by confocal laser scanning fluorescence microscopy. The field shown in Fig. 3C was sectioned optically at various levels from top to bottom. Each section is about 0'7/im thick. These studies have shown that antibodies to the major endoplasmic reticulum glycoprotein, endoplasmin, can be used to visualize this organelle in its entirety by immunofluorescence microscopy. In the standard protocol, cells were lightly fixed with formaldehyde to retain the integrity of the structure and then permeabilized with detergent. Remarkably, even under such circumstances access of the antibodies to endoplasmin, which is probably a luminal protein of the ER (unpublished data), is not precluded and a strong fluorescence signal is obtained. This is probably a reflection of the fact that endoplasmin is a very abundant luminal protein. Therefore even the limited permeability achieved by treating fixed cells with a mild detergent such as saponin seems sufficient. One of the objects of developing a procedure for visualizing the ER in its entirety was to monitor ER morphology in permeabilized cells. It is clear that even gentle treatment of cells with saponin, which generates preparations with a relatively intact ER since the endoplasmin is fully retained (KocheZ al. 1986), causes considerable distension and some vesiculation. Such preparations are inadequate for the study of the ER structure and function, since the organelle is considerably distorted. However, use of an osmotic stabilizer such as sucrose overcomes this problem and generates well-permeabilized cells without vesiculation of the ER. We expect that such preparations will prove 7? visualization by immunofluorescence 539

6 Fig. 5. Morphology of the endoplasmic reticulum in MOPC-315 mitotic cells. Three sections at differing levels (top to bottom) are shown for each mitotic cell. Note the appearance of an intact reticulum in both cells. particularly useful cell-free systems for the analysis of the ER. The amenability of the intact ER to visualization with anti-endoplasmin has permitted the examination of a concept that has intrigued cell biologists over a long period, i.e. what is the fate of the ER during cell division? The first direct attempt to examine the issue in HeLa cells (Robbins & Gonatas, 1964) indicated that the ER vesiculates and fragments away from being a single continuous structure in mitotic cells. This has subsequently been confirmed in other types of cells (see Zeligs & Wollman, 1979, and references therein), although there have been sporadic reports to the contrary (Melmed et al. 1973; Redman & Sreebny, 1970; Pictet et al. 1972; Jimbow et al. 1975). A major limitation of these studies is that they have examined a relatively few isolated thin sections and extrapolated to the overall morphology of the ER. The validity of such an extrapolation is highly questionable for such a complex organelle. Furthermore, no previous study has addressed the question of the continuity or otherwise of the ER across the contractile ring. In this study it was possible to examine the ER right across dividing cells, including the region of the contractile ring, and demonstrate the apparent continuity of the organelle across this region. It is not yet clear whether these observations are unique to cells such as plasmacytoma cells, which contain a large amount of ER, but further analyses on other types of cells with antibodies to endoplasmin should help to clarify the general question. One of our major objectives during these studies has been to obtain preparations of endoplasmic reticulum that are accessible to external reagents, but which retain the normal morphology of the ER in the cell. Fig. 7 shows that such preparations can be obtained provided suitable osmotic stabilizers are included during the permeabilization of the cells. Although extensive studies have been carried out on ER function with microsomal membrane preparations (De Pierre & 540 G. L. E. Koch et al.

7 Fig. 6. Stabilization of endoplasmic reticulum in permeabilized unfixed cells. MOPC-315 cells were equilibrated with 9% sucrose/pbs at 4 C. Saponin was added to 001% and, after 15min at 4 C, cells were developed with anti-endoplasmin as described in Materials and methods, and examined with the confocal laser scanning microscope. Note the reduction in ER vesiculation compared with that in Fig. 1. Fig. 7. Immunofluorescence labelling with different reagents for ER. Cells were prepared as described in Materials and methods, and stained with: A, D10C6 (Terasaki et al. 1984); B, FITC-Con A (Sigma, dil. 1:20); C, anti-endoplasmin ( + FITC goat anti-rabbit Ig); D, anti-endoplasmin and anti-tubulin (see Koch et al. 1986). ER visualization by immunofluorescence 541

8 Dallner, 1975), these are highly disorganized replicas of the in situ organelle and not suitable for detailed analyses of the relationship between ER structure and function. In contrast, it is expected that the abovementioned preparations from the plasmacytoma cell, which retain the essential morphology of the ER but which are accessible to external manipulation, will provide a more realistic model for the analysis of ER function in a cell-free system. We thank Dr J. White for performing the microscopy with the MRC system confocal fluorescence microscope, and Dr W. B. Amos for helpful discussions. References ALBERTS, B., BRAY, D., LEWIS, J., RAFF, M., ROBERTS, K. & WATSON, J. D. (1983). Molecular Biology of the Cell. New York, London: Garland Publishing Inc. DE PIERRE, J. W. & DALLNER, G. (1975). Structural aspects of the membrane of the endoplasmic reticulum. Biochim. biophys. Ada 415, JIMBOW, K., ROTH, S. I., FITZPATRICK, T. B. & SZABO, G. (1975). Mitotic activity in non-neoplastic melanocytes in vivo as determined by histochemical autoradiographic and electron microscope studies. J. Cell Biol. 66, KOCH, G. L. E. & SMITH, M. J. (1986). Specificity of antibodies to the purified Con A acceptor glycoproteins of cultured tumour cells. Br.J. Cancer S3, KOCH, G. L. E., SMITH, M., MACER, D., WEBSTER, P. & MORTARA, R. (1986). Endoplasmic reticulum contains a common abundant calcium-binding glycoprotein, endoplasmin. J. Cell Sci. 86, KRSTIC, R. V. (1979). Ultrastructure of the Mammalian Cell. Berlin: Springer-Verlag. LOUVARD, D., REGGIO, H. & WARREN, G. (1982). Antibodies to the Golgi complex and the rough endoplasmic reticulum. J. Cell Biol. 92, MELMED, R. N., BENITEZ, C. J. & HOLT, S. J. (1973). An ultrastructural study of the pancreatic acinar cell in mitosis, with special reference to changes in the Golgi complex.7. Cell Sci. 12, PALADE, G. E. & PORTER, K. R. (1954). Studies on the endoplasmic reticulum. I. Its identification in cells in situ.j. exp. Med. 100, PiCTET, R. L.,CLARK, W. R., WILLIAMS, R. H. & RUTTER, W. J. (1972). An ultrastructural analysis of the developing embryonic pancreas. Devi Biol. 29, PORTER, K. R. (1953). Observations on the submicroscopic basophilic component of the cytoplasm. J. exp. Med. 97, PORTER, K. R., CLAUDE, A. & FULLAM, E. (1945). A study of tissue culture cells by electron microscopy. J. exp. Med. 81, REDMAN, R. S. & SREEBNY, L. M. (1970). Proliferate behaviour of differentiating cells in the developing rat parotid gland. J. Cell Biol. 46, ROBBINS, E. & GONATAS, N. K. (1964). The ultrastructure of a mammalian cell during the mitotic cycle. J. Cell Biol. 21, TERASAKI, M., SONG, J., WONG, J. R., WEISS, M. J. & CHEN, L. B. (1984). Localisation of endoplasmic reticulum in living and glutaraldehyde-fixed cells with fluorescent dyes. Cell 38, TODARO, G. J. & GREEN, H. (1963). Quantitative studies of the growth of mouse embryo cells in culture and their development into established lines. J. Cell Biol. 17, ZELIGS, J. D. & WOLLMANN, S. H. (1979). Mitosis in rat thyroid epithelial cells in vivo. Ultrastructural changes in cytoplasmic organelles during the mitotic cycle. J. Ultrastruct. Res. 66, (Received 5 December Accepted 10 February! 1987) 542 G. L. E. Koch et al.