JOURNAL

OF STRUCTURAL

BIOLOGY

lo?‘,

106-115

(19%)

Electron Microscopic Studies of the Endoplasmic Reticulum in Whole-Mount Cultured Cells Fixed with Potassium Permanganate JINDAN SONG,~ CHRISTOPHER Division

of Cellular

and

Molecular

Biology,

LEE,~ CHIEN-HUNG SYLVIA

Dana-Farber Boston,

Cancer Institute Massachusetts 02115

LIN, AND LAN

and Harvard

Medical

Bo CHEN

School,

44 Binney

Street,

Received October 10, 1990; and in revised form June 3, 1991

played an important role in refining our understanding of ER organization through direct visualization of complete ER structure in whole (and even living) cells (Terasaki et al., 1986; Lee and Chen, 1988; Lee et al., 1989). However, due to the substantially inferior resolution of optical microscopy, it would be desirable to develop a technique for visualizing ER in whole-mount specimens using conventional transmission EM, and we have therefore sought a suitable fixation and staining method. Fixation with potassium permanganate seemed like a plausible technique in that it preserves and stains membranes while digesting protein structures (Hayat, 1981). This reagent oxidizes proteins and coats membrane structures with an electron dense precipitate (Hayat, 1981). We were thus interested in seeing if this method might be used to visualize the ER by removing the dense matrix of cytosolic proteins, while staining and preserving intracellular membrane systems. However, previously described fixation methods using potassium permanganate proved unsatisfactory in whole mount, giving only incomplete digestion of cytosolic protein, with the ER obscured by thick background staining. To correct this problem, we have tested different formulations of potassium permanganate fixative, and examined their staining pattern with conventional 60-kV transmission EM. In this paper we present results describing a method which gives clear visualization of the ER and other membrane organelles both in whole-mount cultured cells and in thin and thick sections.

A method for visualizing the endoplasmic reticulum and other membrane organelles in whole-mount cells with a standard, 60-kV transmission electron microscope has been developed. By use of a new formulation of potassium permanganate as a fixative, intracellular membranes were preserved and stained, while cytosolic proteins were digested, giving a pattern of membranous organelles against a clear background, suitable for transmission EM of whole-mount cells at 60 kV. Mitochondria, lysosomes, and ER were clearly visible in whole-mount cells fixed by this method. We have employed this technique to examine the organization of the ER in a variety of different cell lines. This method also allowed visualization of the three-dimensional organization, relationships, and fine structure of mitochondria. With prolonged permanganate fixation, mitochondrial cristae were clearly visible in whole-mount cells. This method was also useful for fixation and staining of thin sections, and allowed examination of thicker sections than previously possible, thus giving improved imaging of organelle relationships and fine structure. Using this method, we have examined the ER, mitochondria, and Golgi in thin section. D 1991 Academic PRSS, IW. INTRODUCTION

The endoplasmic reticulum (ER) has long posed a difficult problem for microscopists. Because of the dispersed, fine structure of the ER membrane system, imaging of the ER in thin sections typically reveals only fragments, and often obscures its main form and organization. For this reason, optical microscopy (Buckley, 1964) and new methods for staining the ER with fluorescent lectin (Virtenen et al., 1980), phospholipid precursors (Pagan0 et al., 1981), antibodies (Franke et al., 1978; Louvard et al., 19821, and carbocyanine dyes (Terasaki et al., 1984) have 1 Present address: Dept. of Cell Biology, versity, Shenyang, China. a Present address: Dept. of Cell Biology, Medical Center, Stanford, CA 94305.

RESULTS

Using a new formulation of permanganate fixative, we have developed a method that gives complete digestion of cytosolic proteins while preserving and staining membranes, and have visualized the ER by standard TEM in whole mount cells fixed by this technique (Figs. 1, 2). Through use of sodium citrate as a buffer, and a higher concentration of

China Medical UniStanford

University 106

1047~8477/91

$3.00

Copyright 0 1991 by Academic Press, Inc. All rights

of reproduction

in any form

reserved.

EM

FIG.

llum;

1. MI,

ER in permanganate-fixed, mitochondria; LY, lysosomes.

STUDIES

whole-mount

OF ER IN WHOLE-MOUNT

CV-1.

(a) Magnification:

CELLS

x 4160;

(b) magnification

X 6500.

ER,

endoplasmic

retie

108

FIG. 2. fication: reticulum;

SONG

ER in permanganate-fixed, whole-mount fixed with permanganate MI, mitochondria; LY, lysosomes.

x 8000. (b) CV-1

ET

AL.

CV-1. (a) CV-1 prefixed in glutaraldehyde and visualized by phase-contrast microscopy.

permanganate than in previous formulations, this recipe did not produce the heavy background observed with previous permanganate fixatives and gave fixation suitable for examination of wholemount cells by 60-kV TEM. This method is rapid and simple. For these experiments, cultured cells grown on EM grids were fixed 7 min with the permanganate reagent, washed, dehydrated, and examined with a standard transmission electron microscope (see Materials and Methods for details). This method gave excellent staining of mitochondria and lysosomes, and also revealed a reticular structure of

prior to permanganate Magnification: x 4000.

fixation. MagniER, endoplasmic

interconnecting membrane tubules extending throughout the cytoplasm. This structure was identified as the ER by comparison with the pattern of staining in the same cells prior to fixation by the ER-specific fluorescent dye 3,3’-dihexyloxacarbocyanine iodide (DiOCc(3)). The reticular structure stained by permanganate fixation can be visualized by phase-contrast optical microscopy as well as by transmission EM (Fig. 2b). Since it was easier to relocate cells for comparison on the glass coverslips, we used phase contrast to compare this structure with the ER as stained by the

EM

STUDIES

OF ER IN

dye. The two methods stained the identical structure (Fig. 3). Since DiOC,(S) staining corresponds to the ER as visualized by immunofluorescence against the ER-localized protein BiP (Terasaki, personal communication), it is clear that the reticular structure preserved by potassium permanganate is the ER. Furthermore, close scrutiny of the phase-contrast and fluorescent images did not reveal any significant differences, indicating that permanganate fixation does not alter ER structure, at least at the level of resolution possible by optical microscopy. When cells were prefixed with 0.1% glutaraldehyde prior to fixation with permanganate, the same retic-

FIG. 3. Comparison (b) The same cell after

of DiOC,(3) permanganate

vs permanganate staining fixation. Magnification:

WHOLE-MOUNT

109

CELLS

ulate structure was observed, although the background was higher, and the diameter of ER tubules appeared more uniform (Fig. 2a). The most significant perturbation noted in these studies was that the digestion of cytosolic protein sometimes appeared to compact thick regions of cells, causing areas of three-dimensional structure to become superimposed (Fig. 4). In some cases, ER tubules appeared to have broken during fixation (see Fig. la). We have used this method to examine ER organization in a variety of whole-mount cultured cells (Fig. 5). With this technique the ER appeared to be a lattice of interconnecting tubules approximately

in the same x 4000.

cell.

(a) CV-1

stained

with

DiOCs(3).

Magnification:

x 4000.

110

SONG

FIG. 4.

Compaction

of three-dimensional

30-60 nm in diameter, arranged as a dispersed twodimensional network in the spread areas of cells, and as a dense three-dimensional matrix interspersed with mitochondria and other membrane structures in the region around the nucleus. The ER near the cell periphery was often observed to closely follow the line of the outer edge of the plasma membrane (see Figs. la and 2a). ER tubules and mitochondria were often found closely aligned, as is sometimes observed in thin sections (Terasaki et al.,

ET AL.

structure

of ER.

Magnification:

x 4000.

1984). Groupings of mitochondria were usually surrounded by a dense matrix of ER, more finely reticulated than in the surrounding cytoplasm (see Figs. la and 2a). In CV-1 and FS-2, ER cisternae were observed, as flat, sheet-like regions of membrane contiguous with the tubular network (Fig. 5d). These cisternae were typically 15 pm across, and at their corners were connected with tubules of the regular ER network. Other organizational features of the ER were ap-

EM

STUDIES

OF ER IN

WHOLE-MOUNT

CELLS

FIG. 5. ER in different cell lines fixed with permanganate. (a) CV-1, African green monkey kidney epithelial x 10 000. (b) PtK-2, Potoroo kangaroo kidney epithelial cell. Magnification: x 8000. (c) BALBlc-3T3, mouse embryo fication: x 4000. (d) FS-2, human foreskin fibroblast. Note the presence of ER cisternae (arrowheads). Magnification:

parent by this method. The ER was often observed to be structurally polarized. In extremely large cells, the ER was typically organized into arrays of long, ER tubules with short cross-connections (Fig. 6a). The main axis of these arrays oriented radially, pointing away from the nucleus. Short segments of ER tubule sometimes appeared branched from the network structure (see Fig. lb). In lamellapodia the ER was arranged as a three-dimensional lattice, with many interconnecting tubules present in multiple planes of focus, in contrast to the twodimensional network structure observed in the spread regions of cells (Fig. 6b). In filopodia the ER often forms lines of straight tubules converging at the outer point of the cell (see Fig. la, arrow). The ER seemed conspicuously distributed in every region of the cell. Some cells had thin processes extending away from the cell body; these often contain a single ER tubule traveling the length of the ex-

cell. Magnification: fibroblast. Magnix 4000.

tension and connecting to the ER network at its base (Fig. 6~). Other membranous organelles were also visualized in whole-mount cells by this method. Mitochondria appeared well preserved, and stood out against the background of ER tubules as darkly stained, thick rods (see Figs. 1, 2). Most mitochondria clustered around the nucleus, but individual mitochondria were also seen in the spread areas of the cells. Both filamentous and spherical morphologies of mitochondria were observed, although the former structure was fa.r more prevalent. With 7-min fixation, the mitochondria were heavily stained throughout, and no internal structure could be seen. However, by increasing the length of incubation in the fixative to 15-20 min, the uniformly dark staining of mitochondria was reduced to a light background against which the cristae were clearly visible in whole-mount cells (Fig. 7). This probably re-

112

FIG.

SONG

6.

Organization

ET

AL.

of ER in CV-1. (a) Polarized ER, arrays of long, parallel tubules with structure of ER in lamellapodia, compacted, giving the appearance tubule extending outward into cell process. Magnification: x 6000.

x 5000. (b) Three-dimensional tion: x 5555. (c) Single ER

sulted from digestion of proteins within the mitochondrial matrix, revealing the internal membrane structure. Because this method works in whole-mount cells, it allowed visualization of the three-dimensional relationships of mitochondria, which were often clustered together in the thick regions of the cells. Round vesicles, distributed throughout the cytoplasm, with the appearance of lysosomes were also stained (see Figs. 1,2). Overall, this method appeared to visualize most intracellular membrane structures in whole-mount cells, without apparent damage to or alteration in the structures. Permanganate fixation can also be applied to the visualization of intracellular membranes in thin or thick sections. For these experiments, CV-1 cells were grown in culture dishes, fixed with permanganate, embedded, and sectioned for microscopy. In both lOOO- and 2000-A sections, membranes of the ER, mitochondria, and Golgi were clearly visible (Fig. 8). This method appeared to fix and stain membrane structures with excellent clarity, while digesting and removing proteins. This permitted examination of thicker sections than is possible with stan-

short cross-connections. of a dense, random

Magnification: mesh. Magnifica-

dard fixation techniques. Thus, this technique visualized sections of the ER as regions of tubular reticulum (Figs. 8a, 8b), similar to that seen in whole-mount, rather than as fragmented tubules, the morphology typically seen in thin sections (Fig. 8C).

This method also obtained very clear images of the structure of other organelles, including mitochondria (Figs. 8a, 8d). The inner and outer membranes, as well as the cristae were visualized with good clarity and resolution. In 2000-A sections, ER tubules and mitochondria often appeared to be closely aligned, and in some cases seemed to contact each other (Fig. 8d). The Golgi apparatus was also clearly visible (Fig. 8e), and in thick sections (2000 A) the three-dimensional structure of its stacks could be seen, in contrast to the fragmented morphology of isolated laminae and vesicles obtained with typical thin sections. An unusual vesicular structure was also stained, which appeared to contain extensive layers of membrane infoldings reminiscent of the myelin layers around Schwann cells (Fig. 8d). Overall, the preservation, resolution, and

EM

FIG. 7.

STUDIES

Mitochondrial

OF ER IN WHOLE-MOUNT

cristae

in whole-mount

contrast of intracellular membrane systems in sections fixed by this method were excellent. DISCUSSION

Potassium permanganate seems to be a good fixative for the visualization of intracellular membrane systems. Prior to the introduction of glutaraldehyde fixation by Sabatini et al. (1963), potassium permanganate was suggested as an alternative fixative to osmium tetroxide (Luft, 1956). Since then, it has been used by investigators to fix and stain membranes in thin sections (Bradbury and Meek, 1960; Dimmock, 1970; Bloom et al., 1977; Chiba and Murata, 1981), especially in botanical specimens (Glauert, 1975). In this paper, a new formulation of this reagent was used as fixative for whole-mount cultured cells. Potassium permanganate destroys proteinaceous structures but preserves membrane lipids, including such organelles as ER, mitochondria, the Golgi apparatus, and lysosomes of the cells (Hayat, 1981). Fixation with this reagent extracts most of the cellular protein (Millonig and Marionozzi, 1968); many cell components, including the cytoskeleton, ribosomes, and soluble cytoplasmic proteins can no longer be seen after this treatment. The preservation of membranes if probably due to deposition of MnOa on the hydrophilic surface of the lipid bilayer which coats membranes with an electron-dense precipitate (Eddy and Johns, 1965). Combined with the fixative’s destruction of cytoplasmic protein, this yields a staining pattern of intracellular membranes with good, differential penetration of electrons even in whole cells at standard voltages (60 kV). Only in the thick nuclear region was the staining too dense to resolve individual structures. Using this technique, we have examined the organization of ER in whole-mount cultured cells. The

CV-1.

113

CELLS

Magnification:

x 6500.

network structure of ER visualized by this method is similar to that obtained by osmium tetroxide fixation of whole-mount cells (Porter, 1953) and by high voltage electron microscopy of critical-point dried cells (Buckley and Porter, 1975). The same pattern has also been observed, although at lower resolution, by fluorescence microscopy using fluorescently labeled phospholipid precursors (Pagan0 et al., 1981) and carbocyanine dyes in both fixed and live cells (Terasaki et al., 1984). Thus, these observations of ER structure do not seem to be artifactual. Indeed, staining of glutaraldehyde-fixed cells with DiOCs(3) followed by permanganate fixation, and comparison of the two images, has shown that the permanganate fixative does not perceptibly alter the structure of the ER. This method clearly demonstrates the continuity of the ER throughout the cytoplasm, in agreement with previous results (Palade and Porter, 1954). In this regard, it has proved to be useful for examining the organization of the ER, which is entirely obscured in thin sections. We now consider the advantages and disadvantages of this method, as well as possible applications for which it is likely to be useful. Clearly, the principal feature of t,his method-its destruction of protein-based structures-has both disadvantages and advantages. The utility of this method is limited entirely to the study of membranes, without reference to protein structures such as the cytoskeleton. However, by digesting cytosolic protein, permanganate allows the microscopist to work with much thicker sections and even whole-mount cells, permitting direct examination of the large-scale form, organization, and relationships of membrane organelles. This method is particularly useful for studying organelle distribution and organization, which are often difficult or impossible to perceive in thin sec-

SONG

ET AL.

FIG. 8. Thin sections of CV-1 cells. (a) Permanganate-fixed 2000-A section. Note the extensive ER network and mitochondria. Magnification: x 10 000. (b) Permanganate-fixed 1000-A section. Note the interconnected morphology of the ER still detectable. Magnification: x 5000. (c) Glutaraldehyde-fixed 600-A section. Note the fragmented morphology of the ER, and its alignment with microtubules. Magnification: x 20 000. (d) Permanganate-fixed 2000-A section. Note the mitochondria, ER, and vesicular structures. Magnification: x 15 000. (e) Permanganate-fixed 2000-A section. Note the compaction of Golgi. Magnification: x 18 000.

EM

tions. It thus allows one to apply imaging power of the electron study of membrane structure in simple methods and a standard tron microscope. MATERIALS

AND

STUDIES

OF ER

the resolution and microscope to the whole cells, using transmission elec-

METHODS

Cell Culture. CV-1 (African green monkey kidney epithelial cells) and PtK-2 (Potoroo kangaroo kidney epithelial cells) were obtained from the American Type Culture Collection. BALB/c3T3 mouse embryo fibroblasts were obtained from Dr. C. Stiles (Dana-Farber Cancer Institute). The human breast adenocarcinoma cell line MCF-7 was a gift from Dr. M. Rich (Michigan Cancer Foundation). The human foreskin fibroblast strain FS-2 was from Dr. R. Sager (Dana-Farber Cancer Institute), and chick embryo fibroblasts were prepared according to the protocol of Rein and Rubin (1968). Cells were grown in Dulbecco’s modified Eagle’s medium containing 5% calf serum (MA Bioproducts, Walkersville, MD) at 37” with 100% humidity and 5% CO,. Preparation ofpotassium permanganate fixative. After testing many new formulations, the following recipe was found to give the best membrane preservation and the lowest background: 60 m&f trisodium citrate, 25 mM potassium chloride, 35 mM magnesium chloride, and 125 mM potassium permanganate. The pH of freshly made fixative should be between 7.4 and 7.8. Since the fixative is not stable, it should be stored in a brown, glass-stopped bottle, and has a useful lifetime at 4°C of up to 8 weeks, or up to 2 weeks when stored at room temperature. Alternatively, one can prepare a stock solution of citrate buffer (60 mM trisodium citrate, 25 mM potassium chloride, and 35 mM magnesium chloride), which can be kept for several months at 4°C and add potassium permanganate (final concentration, 120 mM) to an aliquot of buffer on the day of the experiment. Electron microscopy. Cells were cultured on loo-mesh gold EM grids (Ladd Research Industries, Inc.) for examination by whole-mount electron microscopy. Grids were coated with 0.5% Formvar solution (Ladd Research Industries, Inc.), placed on sterile 12-mm2 glass coverslips (Bradford Scientific, Epping, NH), and plated with cells at low density. Cells grown on grids for several days were fixed in the potassium permanganate reagent described above for 7 min at room temperature. In some experiments, prefixation with glutaraldehyde was tested by incubating grids in 0.1% glutaraldehyde in citrate buffer for 7 min prior to the standard permanganate treatment. After fixation, the grids were removed from the glass coverslip, washed with distilled water, and dehydrated in a graded series of ethanol solutions (25,50, 75, and 100%). The whole-mount cells on the grids were examined with a standard transmission electron microscope (Zeiss OM-10) at 60 kV. For sections, cells were grown on 35-mm plastic dishes, fixed with either permanganate of 2.5% glutaraldehyde in citrate buffer, and dehydrated with an ethanol series as described above. They were then embedded in Epon 812, sectioned by an LKB microtome to thickness ranging from 1000 to 2000 A, and examined upon the Zeiss OM-10. The sections were prepared by Elizabeth Beaumont in Dr. Henry Slate& laboratory (Dana-Farber Cancer Institute). Detection of ER by fluorescent microscopy and phase-contrast microscopy. The ER was first visualized by the method of Terasaki et al. (1984). 3,3’-Dihexyloxacarbocyanine iodide, denoted DiOC,(S), (Eastman Organic Chemicals, Rochester, NY) was dissolved in ethanol to make a 0.5-mgiml stock solution. Cells grown at low density on 12-mm2 glass coverslips (Bradford Scientific, Eping, NH) were fixed in 0.25% glutaraldehyde in sucrose-cacodylate buffer (0.1 M sucrose, 0.1 M sodium cacodylate, pH 7.4) for 5 min at room temperature, and then stained with 2.5 mg/ml of DiOCs(3) in sucrose-cacodylate buffer for 10 set at room temperature. The coverslips were then washed with sucrose-

IN WHOLE-MOUNT

115

CELLS

cacodylate buffer and mounted in the same buffer in a live-cell observation chamber made of 0.7-mm-thick silicon rubber (N. A. Reiss, Belle Mead, NJ) as described by Johnson et al. (1980). Cells were photographed on a Zeiss Photomicroscope III using a Neofluar (100X, NA 1.2) objective, blue excitation barriers, and Kodak Tri-X film at E.I. 6300. The film was developed in Kodak HCllO (dilution B) for 15 min. To assist later relocation of the same cells, maps were drawn at low magnification of the neighboring regions around each photograph. The coverslips were then removed from the mount and fixed in potassium permanganate fixative for 7 min. The coverslips were then washed, remounted in PBS as above, and examined on a Zeiss Photomicroscope III using a Planapo 100X phase-contrast objective. Photographs were made on Kodak 2415 film at E.I. 200 and developed in Kodak Technidol. To relocate the same cells, coverslips were scanned at low magnification to find the mapped regions of cells. We are most grateful to Drs. Keith Porter and Mark Terasaki for valuable advice. This work has been supported by grants from the National Institute of Health (HD24926 and GM38318). REFERENCES BLOOM, (1977)

G. D., CARLSOO, Histochemistry

BRADBURY, 241-250.

S., AND

B., DANNIELSSON, 51, 261-268.

MEEK,

G. A. (1960)

A., AND

WAHLIN,

T.

Sci.

101,

Q. J. Microsc.

BUCKLEY, I. K. (1964) Protoplasma 59, 569-588. BUCKLEY, I. K., AND PORTER, K. R. (1975) J. Microsc. 104, 107-120. CHIBA,

T., AND MURATA,

Y. (1981)

J. Neurocytol.

10, 315-329.

DIMMOCK, E. (1970) J. Cell Sci. 7, 719-737. EDDY, A. A., AND JOHNS, B. (1965) SCZ Monog. FRANKE, W. W., FINKE, Rep. 2, 465-474.

A., AND SCHMIDT,

GLAUERT, A. M. (1975) copy. North-Holland, New York. pp. 50-57.

Practical Amsterdam;

HAYAT, M. A. (1981) Fixation 193, Academic Press, New

(Oxford)

19, 24-36.

E. (1978)

Cell Biol.

Methods in Electron Oxford/American

for Electron York.

MicrosElsevier,

Microscopy,

pp. 183-

JOHNSON, L. V., WALSH, M. L., AND CHEN, L. B. (1980) Natl. Acad. Sci. USA 77, 990-994. LEE, C., AND CHEN, L. B. (1988) Cell 54, 3746. LEE, C., FERGUSON, 2045-2055.

M., AND CHEN,

LOUVARD, D., REGGIO, 92, 92-107. LUFT,

T. H. (1956)

H.,

AND WARREN,

J. Biophys.

Biochem.

MILLONIG, G., AND MARINOZZI, LETT, V. E. (Eds.). Advances copy, Vol. 2, p. 251, Academic PAGANO, R. E., LONGMUIR, D. K. (1981) J. Cell Biol.

L. B. (1989)

Cytol.

109,

J. Cell Biol.

2, 799-807.

V. (1968) in BARER, R., AND Cossin Optical and Electron MicrosPress, New York.

K. J., MARTIN, 91, 872-877.

PALADE, 656.

G. E., AND PORTER,

K. R. (1954)

PORTER,

K. R. (1953)

Med.

J. Exp.

Proc.

J. Cell Biol.

G. (1982)

Znt.

0. C., AND J. Exp.

Med.

STRUCK, 100, 641-

97, 727-750.

REIN, A., AND RUBIN, H. (1968) Exp. Cell Res 49, 666-678. SABATINI, D. D., MILLER, K., AND BARNETT, R. J. (1963) Biol. 17, 19-58. TERASAKI, M., SONG;, J., WONG, L. B. (1984) Cell 38, 101-108.

J. R., WEISS,

M. J., AND

J. Cell CHEN,

TERASAKI, M., CHEN, L. B., AND FUJIWARA, K. (1986) J. Cell Biol. 103, 1557-1568. VIRTENEN, I., ELBLO~M, P., AND LAURILA, P. (1980) J. CeZZBiol. 85, 429434.