( Yournal of Microscopy, Vol. 111, Pt 2, November 1977, p p . 193-201. Revised paper accepted 11J u l . ~1977
Freeze-fracture for scanning electron microscopy
by G . H . H A G G I Sand B E V E R L Y P H I P P S - T O D DElectron , Microscope Centre, Chemistry and Biology Research Institute, Agriculture Canada, Ottawa
SUMMARY
Two different freeze-fracture methods are explored for preparation of biological material for scanning electron microscopy. In the simpler method the tissues are first fixed and dehydrated. They are then frozen and fractured, and after thawing, critical-point dried. This method has already been used in a number of studies of animal tissues (heart, liver, kidney). It is applied here to the examination of plant material (leaf mesophyll cells). In the second method tissues, or cells, are first infiltrated with cryoprotectant (dimethylsulphoxide) then frozen and fractured, and not fixed until after thawing. The fixed tissues are finally dehydrated and critical-point dried. This method also has previously been used in the study of animal tissues, and is applied here to carrot protoplasts, chicken erythrocytes, and leaf mesophyll cells. INTRODUCTION
We have been exploring two different freeze-fracture procedures for examination of cells and tissues by SEM. The first is the method of Sybers & Ashraf (1973) as modified by Humphreys et al. (1974). In this method the tissue is first fixed and dehydrated to absolute alcohol, then frozen, fractured, thawed in absolute alcohol and critical-point dried. We shall refer to this method as FDFf (fix, dehydrate, freeze-fracture). This method gives excellent results for the examination of the luminal surfaces of cells which line natural tissue spaces as, for example, liver sinusoids, or Bowman’s capsule in kidney (Humphreys et al., 1974). We show here that this method can be used for examination of the outer surface of plant cells, in situ in the leaf, the inner and outer surfaces of plant cell walls, also the inner surface of the tonoplast. This first method, in general, shows little detail with the cytoplasm or nucleoplasm of cells (Humphreys et al., 1974), and it is for this reason that we have pursued a second alternative. This is a procedure, previously described (Haggis et al., 1976), in which the material is first infiltrated with cryoprotectant, frozen and fractured, and then thawed into fixative solution, fixed, dehydrated and criticalpoint dried. We shall refer to this method as F f T F (freeze-fracture, thaw, fix). The previous paper (Haggis et al., 1976) describes results for animal tissues (heart, liver, kidney). We report here results for carrot protoplasts, chicken erythrocytes chosen for study as a simple cell of known ultrastructure, and for leaf mesophyll cells. The aim of this F f T F method is to ‘wash out’ soluble proteins from the fracture face at the thawing stage, and thereby reveal cytoplasmic and nucleoplasmic
193
G. H. Haggis and Beverly Phipps-Todd structure which remains hidden in the FDFf method. There is the possibility that certain aspects of internal membrane, cytoskeleton, or chromatin arrangement may be simply revealed in the three-dimensional view offered by the SEM, that would not be apparent in thin sections without recourse to the much more timeconsuming process of serial section reconstruction, or access to a high-voltage microscope. MATERIALS AND METHODS
The materials used were radish leaves (Scarlet Globe) grown from seed in the laboratory for 5 days on moist vermiculite, protoplasts prepared from carrot cells (line CA68 obtained from Dr I. Veliky, NRC, Ottawa) maintained in culture in 67V medium (Veliky et al., 1969), and erythrocytes from fresh healthy chicken blood. For the FDFf studies leaves were cut into thin slices (0.5 mm) with a razor blade under tap water and the slices were immediately fixed in 4 O A glutaraldehyde + 0.1 M sucrose in 0.1 M Na phosphate buffer at pH 7.0 at room temperature. They were sometimes post-fixed in 1;' osmium tetroxide in 0.1 M Na phosphate buffer (pH 7.0). Where post-fixation in osmium has been included, or where there has been modification of this 'standard' glutaraldehyde fixative solution this is noted in the appropriate Figure legend. After fixation the slices were dehydrated in increasing concentrations, to absolute ethanol, then frozen by plunging into liquid nitrogen, fractured under the liquid nitrogen, thawed into absolute ethanol, criticalpoint dried, and, after drying, mounted and gold-coated for SEM viewing. For the F f T F studies carrot protoplasts or chicken erythrocytes were spun down to a fairly thick suspension and mixed with an equal volume of 12(;4 fibrinogen (Calbiochem, Calif.). A small amount of thombin (Sigma, St Louis, Mo.) was then mixed in, to a final concentration of 5 N I H units/ml, and the cell-fibrinogenthrombin mixture spread as a thin layer (0.2 mm) in a moist chamber, and allowed to set for 15-30 min. Prior to addition of the fibrinogen and thrombin, the protoplasts were in 67V medium (Veliky et al., 1969) and the erythrocytes in 0.85:, NaCl. The 12", fibrinogen was made up, in each case, in the appropriate medium. The set gels were placed in the appropriate medium (67V medium for the protoplast gels and 0.S5°/0 NaCl for the erythrocyte gels) and dimethylsulphodide (DMSO) was added in 5Ob steps every 15 min to a final concentration of 25,;. The gels were then left for a further hour to equilibrate. They were then rapidly frozen by plunging into Freon 22, held at 123 K, transferred to a liquid nitrogen bath and fractured under the nitrogen, then thawed into 2O , glutaraldehyde 25", , DMSO in 0.1 M Na phosphate buffer. Again, glutaraldehyde fixation was sometimes followed by osmium fixation, and after fixation the material was dehydrated to absolute ethanol, critical-point dried, mounted and gold-coated for SEM viewing. Leaf slices were treated in the same way as the gels, DMSO being added in 5",, steps (final concentration 25",,) to a solution of 0.1 M Na phosphate buffer+O.l M sucrose in which the slices were suspended. The micrographs were taken on a Cambridge IIa Stereoscan. It may be noted that in the FDFf method rapid freezing is not essential, since no ice crystals form in the alcohol infiltrated tissues. The samples are frozen directly into liquid nitrogen and need not be very thin. I n the F f T F method we have tried to make freezing and thawing as rapid as possible (to limit ice crystal size) by using thin gels and freezing in Freon 22. The chosen final concentration of DMSO (20-250,,) is a compromise, found to be optimal by trial and error. A higher concentration would be desirable, to limit ice crystal size on freezing and thawing. A lower concentration would be desirable, to limit possible ultrastructure
+
194
Freeze-fracture for SEM change due to the effect of DMSO itself (as distinct from possible ultrastructure change occurring during freezing and thawing). RESULTS AND DISCUSSION
A. Fixed, dehydrated, freeze-fracture (FDFf) Part of the cytoplasm and vacuole of a leaf mesophyll cell prepared by this first method, with glutaraldehyde fixation only, are shown in Fig. 1. A limitation to this method, known from its application to animal cells (Humphreys et al., 1974) is that it reveals very little detail of the cell cytoplasm or nucleoplasm. Thus no detail is seen of chloroplast internal structure in Fig. 1 except for the starch grains (S in Fig. 1). At high magnification ( x 10,000) such fractured starch grains can be clearly identified from their location within chloroplasts and because they protrude a little from the fracture plane, possibly due to a slight relative shrinkage of the surrounding tissues in critical point drying. If glutaraldehyde fixation is followed by osmium fixation (Fig. 2) the bodies seen as spherical hollows in Fig. 1 are now seen as solid spheres (L in Fig. 2). These are probably lipid bodies whose content is retained after osmium post-fixation, but lost after glutaraldehyde fixation alone during dehydration or during ethanol thawing after fracture. In the FDFf method variation in fixation conditions can be exploited to reveal different ultrastructural features. In the preparation of Figs. 1 and 2 fixation conditions are a little different from the standard method given in the Materials and Methods section, with initial fixation in 49, glutaraldehyde+0.15 M sucrose in 0.1 M Na phosphate buffer. This fixative is slightly hypertonic and causes some shrinking of the cells away from their outer walls. This exposes the outer surface of the plasma membrane and the inner surface of the wall to SEM viewing. Careful examination of Fig. 1 (left edge) shows perhaps the remnants of plasmodesmata that have been torn out from the wall during this shrinking. Initial fixation in 4 O , , glutaraldehydet- 0.1 M sucrose in 0.1 M buffer (the standard methods given in the Materials and Methods section) is shown in Fig. 3 where the plasma membranes of two cells lie closely apposed to their outer walls. The outer surface of the wall is here exposed to SEM viewing (W in Fig. 3 ) looking down into the air space between two cell walls. The inner surface of the tonoplast can be examined ( T in Fig. 3 ) and is seen in these cells to be covered by club-shaped blebs. Fixation in hypotonic fixative can also be exploited. Figure 4 shows a cell which has been fixed in 47, glutaraldehyde in 0.1 M Na phosphate buffer, without added sucrose, and has broken up during fixation with some swelling of the chloroplasts. Where the fracture runs through these chloroplasts the internal membranes are seen. We are unlikely to see, in the SEM, any features of thylakoid membrane structure which have not been seen in transmission microscopy, but possibly high-resolution secondary-electron SEM can show up chloroplast DNA and its membrane attachments.
B. Freeze-fracture, thaw jix (FfTF) In the first preparative method (FDFf) the internal details of chloroplast structure are not seen in a well-fixed specimen (Fig. l), but only as a result of poor fixation (Fig. 4). The aim of the second method (FfTF) is to reveal such internal detail, within chloroplasts and other cytoplasmic organelles, and within nuclei, in a well-fixed specimen. In this method, at the moment the fractured surface thaws, the cells are unfixed. On thawing, soluble proteins diffuse out from the surface. However, since thawing takes place in the fixative, fixation of surface structure is very rapid. Tests with frozen chicken erythrocytes in a fibrin gel show individual
195
Figs. 1-4. Mesophyll cells of young radish leaf. Figures 1 and 2 compare glutaraldehyde fixation alone (Fig. 1) and glutaraldehyde + osmium fixation (Fig. 2). Both are fixed in slightly hypertonic solution, bringing the cytoplasm away from the cell wall. Figure 3 shows isotonic, and Fig. 4 hypotonic fixation (see discussion in text). V=vacuole, C = chloroplast, S = starch grain, L =lipid bodies, W = cell wall, P =outer surface of plasma membrane, T = inner surface of tonoplast, FDFf preparation. Magnifications : Fig. 1 x 9000, Figs. 2-4 x 11,000.
196
Freeze-fracture for SEM
cells cleaned of haemoglobin but, where a number of red cells are clumped together, the larger amount of haemoglobin becomes fixed before it can diffuse away (Hb in Fig. 10). The results obtained by this method with carrot protoplasts are shown in Figs. 6, 7, 8 and 9 and can be compared with the thin-section micrograph of a carrot protoplast in Fig. 5 . In the cell of Fig. 6 the fracture passed through a nucleus, in that of Fig. 7 through a large vacuole. In the preparation of Fig. 8 the fracture has passed through the cell above the nucleus, so that here the cytoplasmic surface of the Outer nuclear membrane is revealed. At the resolution of the present study the nuclear pores are not seen in Fig. 8, nor is the resolution adequate to show ribosomes on the nuclear surface. In the cytoplasm of these cells, the main features seen are mitochondria, lipid bodies organelles of comparable size in Fig. 5 ) and fibrous strands (Figs. 6, 7, 8 and 9). Figure 9 shows cytoplasmic detail observed on only one occasion, where beads or small VaCUOkS of relatively uniform size were seen along the strands. These cultured cells are in varying states of growth and activity, and we may, in the preparation of Fig. 9, have caught a particular growth phase. Typically the cytoplasm shows bodies of more varied size attached to the strands, as in Fig. 6. An immediate question arises as to whether these strands (seen running from vacuole to plasma membrane in Fig. 7 and from nucleus to plasma membrane in Fig. 8) represent genuine cytoplasmic components, or artefacts arising from DMSO treatment, freezing, thawing, or delayed fixation ? In other words, is it possible by this technique to show up the genuine cytoskeleton of cells, and if so, are our present methods of freezing and thawing sophisticated enough to achieve this result ? To attempt to answer this question we have begun a study of a simpler cell, the chicken erythrocyte. For turkey erythrocytes, thin-section micrographs of ghosts show fine strands running from nucleus to plasma membrane in the central region of the cell (Harris, 1971). Also a ring of microtubules runs around the cell periphery (Harris, 1971). F f T F micrographs of these cells are shown in Figs. 10, 11 and 12. Thin strands are indeed seen running from nucleus to plasma membrane (Fig. 11). Where the main part of a cell has fractured away, leaving a part of the rim behind, slightly thicker strands can be seen in the expected location of the microtubules (Fig. 12). There remains the possible criticism that the fibrin in which the cells are embedded might in some way penetrate into the cells to form the structures seen. However, identical fine fibrils, running from nucleus to plasma membrane, and the same thicker fibres running round the rim of the cell, which we believe to be the microtubules, are also seen in cells embedded in Agar. F f T F micrographs of leaf mesophyll cells are shown in Figs. 13 and 14. The tenuous cytoplasmic layers of leaf cells have proved difficult to preserve well, through the freezing and thawing which precedes fixation in the F f T F method. However, in the cell of Fig. 13 only a small tear is seen. Good structural preservation is achieved by using very thin slices of tissue to allow rapid infiltration, and rapid freezing and thawing. Thin-section controls show little ultrastructure change during DMSO infiltration, mainly a slight swelling of lipid bodies, provided the DMSO concentration is not brought above 257,. CONCLUSIONS A N D FUTURE PROSPECTS
The results for chicken erythrocytes (Figs. 10-12) allow the important conclusion that the F f T F method can, in appropriate circumstances, preserve cytoskeletal components, so that these are displayed in three dimensions for SEM
197
G. H. Haggis and Beverly Phipps- Todd
198
Freeze-fracture for SEM
Figs. 5-9. Carrot protoplasts. Figure 5 is a standard thin section, Figs. 6-9 FfTF preparations. In Fig. 6 the fracture passes through a nucleus, in Fig. 7 through a vacuole, and in Fig. 8 above a nucleus. Figure 9 shows cytoplasmic detail (see discussion in text). V = vacuole, F N = fracture through nucleus, N = cytoplasmic surface of outer nuclear membrane. Magnifications: Figs. 5-7 x 8000, Figs. 8 and 9 x 11,000.
14
199
Figs. 10-12. Chicken erythrocytes. Figure 10 shows haemoglobin diffusing out from the fracture face (see discussion in text). Figures 11 and 12 show ultrastructure details. H b = haemoglobin, N = cytoplasmic surface of outer nuclear membrane, SS = strands running from nucleus to plasma membrane, arrow = microtubules. FfTF preparation. Magnifications : Fig. 10 x 2000, Figs. 11 and 12 x 9000.
Figs. 13 and 14. Mesophyll cells of young radish leaf. FfTF preparation. Figure 13 shows only a small defect in preservation (arrowhead) arising during freezing and thawing. Vacuolar material is washed out in the FfTF method. C = fractured chloroplast. Magnifications: Fig. 13 x 5500, Fig. 14 x 11,000.
Freeze-fracture for SEM viewing. Fractures can be obtained through nuclei in these and other cells (Fig. 6) and through the cytoplasm, to show the outer nuclear surface (Fig. 8) and suggestive cytoplasmic detail (Fig. 9). Further detail should be seen in preparations of this type in a high-resolution SEM. We also have the possibility in the F f T F method, at the moment of thawing, of labelling intracellular structure with antibodies linked to markers, such as haemocyanin, which can be recognized in the SEM. ACKNOWLEDGMENTS
We are very grateful to S. Lesley for the supply of carrot protoplasts and to S. Becker and A. Greig for the supply of chicken erythrocytes.
References Haggis, G.H., Bond, E.F. & Phipps, B. (1976) Visualization of mitochondria1 cristae and nuclear chromatin by SEM. 9th I I T R I SEM Symposium, Chicago, p. 281: Harris, J.R. (1971) The ultrastructure of the erythrocyte. In: Physiology and Bzochemistry of the Domestic Fowl (Ed. by D. T. Bell and B. M. Freeman), Vol. 2, p. 835. Academic Press, New York. Humphreys, W.J., Spurlock, B.O. & Johnson J.S. (1974) Critical point drying of ethanolinfiltrated cryofractured biological specimens for scanning electron microscopy. 7th I I T R I SEM Symposium, Chicago, p. 275. Sybers, H.D. & Ashraf, M. (1973) Preparation of cardiac muscle for SEM. 6th IITRZ SEM Symposium, Chicago, p. 341. Veliky, I.A., Sandkvist, A. & Martin, S.M. (1969) Physiology of, and enzyme production by, plant cell cultures. Biotechnol. Bioeng. 11, 1247.
201
Freeze-fracture for scanning electron microscopy
by G . H . H A G G I Sand B E V E R L Y P H I P P S - T O D DElectron , Microscope Centre, Chemistry and Biology Research Institute, Agriculture Canada, Ottawa
SUMMARY
Two different freeze-fracture methods are explored for preparation of biological material for scanning electron microscopy. In the simpler method the tissues are first fixed and dehydrated. They are then frozen and fractured, and after thawing, critical-point dried. This method has already been used in a number of studies of animal tissues (heart, liver, kidney). It is applied here to the examination of plant material (leaf mesophyll cells). In the second method tissues, or cells, are first infiltrated with cryoprotectant (dimethylsulphoxide) then frozen and fractured, and not fixed until after thawing. The fixed tissues are finally dehydrated and critical-point dried. This method also has previously been used in the study of animal tissues, and is applied here to carrot protoplasts, chicken erythrocytes, and leaf mesophyll cells. INTRODUCTION
We have been exploring two different freeze-fracture procedures for examination of cells and tissues by SEM. The first is the method of Sybers & Ashraf (1973) as modified by Humphreys et al. (1974). In this method the tissue is first fixed and dehydrated to absolute alcohol, then frozen, fractured, thawed in absolute alcohol and critical-point dried. We shall refer to this method as FDFf (fix, dehydrate, freeze-fracture). This method gives excellent results for the examination of the luminal surfaces of cells which line natural tissue spaces as, for example, liver sinusoids, or Bowman’s capsule in kidney (Humphreys et al., 1974). We show here that this method can be used for examination of the outer surface of plant cells, in situ in the leaf, the inner and outer surfaces of plant cell walls, also the inner surface of the tonoplast. This first method, in general, shows little detail with the cytoplasm or nucleoplasm of cells (Humphreys et al., 1974), and it is for this reason that we have pursued a second alternative. This is a procedure, previously described (Haggis et al., 1976), in which the material is first infiltrated with cryoprotectant, frozen and fractured, and then thawed into fixative solution, fixed, dehydrated and criticalpoint dried. We shall refer to this method as F f T F (freeze-fracture, thaw, fix). The previous paper (Haggis et al., 1976) describes results for animal tissues (heart, liver, kidney). We report here results for carrot protoplasts, chicken erythrocytes chosen for study as a simple cell of known ultrastructure, and for leaf mesophyll cells. The aim of this F f T F method is to ‘wash out’ soluble proteins from the fracture face at the thawing stage, and thereby reveal cytoplasmic and nucleoplasmic
193
G. H. Haggis and Beverly Phipps-Todd structure which remains hidden in the FDFf method. There is the possibility that certain aspects of internal membrane, cytoskeleton, or chromatin arrangement may be simply revealed in the three-dimensional view offered by the SEM, that would not be apparent in thin sections without recourse to the much more timeconsuming process of serial section reconstruction, or access to a high-voltage microscope. MATERIALS AND METHODS
The materials used were radish leaves (Scarlet Globe) grown from seed in the laboratory for 5 days on moist vermiculite, protoplasts prepared from carrot cells (line CA68 obtained from Dr I. Veliky, NRC, Ottawa) maintained in culture in 67V medium (Veliky et al., 1969), and erythrocytes from fresh healthy chicken blood. For the FDFf studies leaves were cut into thin slices (0.5 mm) with a razor blade under tap water and the slices were immediately fixed in 4 O A glutaraldehyde + 0.1 M sucrose in 0.1 M Na phosphate buffer at pH 7.0 at room temperature. They were sometimes post-fixed in 1;' osmium tetroxide in 0.1 M Na phosphate buffer (pH 7.0). Where post-fixation in osmium has been included, or where there has been modification of this 'standard' glutaraldehyde fixative solution this is noted in the appropriate Figure legend. After fixation the slices were dehydrated in increasing concentrations, to absolute ethanol, then frozen by plunging into liquid nitrogen, fractured under the liquid nitrogen, thawed into absolute ethanol, criticalpoint dried, and, after drying, mounted and gold-coated for SEM viewing. For the F f T F studies carrot protoplasts or chicken erythrocytes were spun down to a fairly thick suspension and mixed with an equal volume of 12(;4 fibrinogen (Calbiochem, Calif.). A small amount of thombin (Sigma, St Louis, Mo.) was then mixed in, to a final concentration of 5 N I H units/ml, and the cell-fibrinogenthrombin mixture spread as a thin layer (0.2 mm) in a moist chamber, and allowed to set for 15-30 min. Prior to addition of the fibrinogen and thrombin, the protoplasts were in 67V medium (Veliky et al., 1969) and the erythrocytes in 0.85:, NaCl. The 12", fibrinogen was made up, in each case, in the appropriate medium. The set gels were placed in the appropriate medium (67V medium for the protoplast gels and 0.S5°/0 NaCl for the erythrocyte gels) and dimethylsulphodide (DMSO) was added in 5Ob steps every 15 min to a final concentration of 25,;. The gels were then left for a further hour to equilibrate. They were then rapidly frozen by plunging into Freon 22, held at 123 K, transferred to a liquid nitrogen bath and fractured under the nitrogen, then thawed into 2O , glutaraldehyde 25", , DMSO in 0.1 M Na phosphate buffer. Again, glutaraldehyde fixation was sometimes followed by osmium fixation, and after fixation the material was dehydrated to absolute ethanol, critical-point dried, mounted and gold-coated for SEM viewing. Leaf slices were treated in the same way as the gels, DMSO being added in 5",, steps (final concentration 25",,) to a solution of 0.1 M Na phosphate buffer+O.l M sucrose in which the slices were suspended. The micrographs were taken on a Cambridge IIa Stereoscan. It may be noted that in the FDFf method rapid freezing is not essential, since no ice crystals form in the alcohol infiltrated tissues. The samples are frozen directly into liquid nitrogen and need not be very thin. I n the F f T F method we have tried to make freezing and thawing as rapid as possible (to limit ice crystal size) by using thin gels and freezing in Freon 22. The chosen final concentration of DMSO (20-250,,) is a compromise, found to be optimal by trial and error. A higher concentration would be desirable, to limit ice crystal size on freezing and thawing. A lower concentration would be desirable, to limit possible ultrastructure
+
194
Freeze-fracture for SEM change due to the effect of DMSO itself (as distinct from possible ultrastructure change occurring during freezing and thawing). RESULTS AND DISCUSSION
A. Fixed, dehydrated, freeze-fracture (FDFf) Part of the cytoplasm and vacuole of a leaf mesophyll cell prepared by this first method, with glutaraldehyde fixation only, are shown in Fig. 1. A limitation to this method, known from its application to animal cells (Humphreys et al., 1974) is that it reveals very little detail of the cell cytoplasm or nucleoplasm. Thus no detail is seen of chloroplast internal structure in Fig. 1 except for the starch grains (S in Fig. 1). At high magnification ( x 10,000) such fractured starch grains can be clearly identified from their location within chloroplasts and because they protrude a little from the fracture plane, possibly due to a slight relative shrinkage of the surrounding tissues in critical point drying. If glutaraldehyde fixation is followed by osmium fixation (Fig. 2) the bodies seen as spherical hollows in Fig. 1 are now seen as solid spheres (L in Fig. 2). These are probably lipid bodies whose content is retained after osmium post-fixation, but lost after glutaraldehyde fixation alone during dehydration or during ethanol thawing after fracture. In the FDFf method variation in fixation conditions can be exploited to reveal different ultrastructural features. In the preparation of Figs. 1 and 2 fixation conditions are a little different from the standard method given in the Materials and Methods section, with initial fixation in 49, glutaraldehyde+0.15 M sucrose in 0.1 M Na phosphate buffer. This fixative is slightly hypertonic and causes some shrinking of the cells away from their outer walls. This exposes the outer surface of the plasma membrane and the inner surface of the wall to SEM viewing. Careful examination of Fig. 1 (left edge) shows perhaps the remnants of plasmodesmata that have been torn out from the wall during this shrinking. Initial fixation in 4 O , , glutaraldehydet- 0.1 M sucrose in 0.1 M buffer (the standard methods given in the Materials and Methods section) is shown in Fig. 3 where the plasma membranes of two cells lie closely apposed to their outer walls. The outer surface of the wall is here exposed to SEM viewing (W in Fig. 3 ) looking down into the air space between two cell walls. The inner surface of the tonoplast can be examined ( T in Fig. 3 ) and is seen in these cells to be covered by club-shaped blebs. Fixation in hypotonic fixative can also be exploited. Figure 4 shows a cell which has been fixed in 47, glutaraldehyde in 0.1 M Na phosphate buffer, without added sucrose, and has broken up during fixation with some swelling of the chloroplasts. Where the fracture runs through these chloroplasts the internal membranes are seen. We are unlikely to see, in the SEM, any features of thylakoid membrane structure which have not been seen in transmission microscopy, but possibly high-resolution secondary-electron SEM can show up chloroplast DNA and its membrane attachments.
B. Freeze-fracture, thaw jix (FfTF) In the first preparative method (FDFf) the internal details of chloroplast structure are not seen in a well-fixed specimen (Fig. l), but only as a result of poor fixation (Fig. 4). The aim of the second method (FfTF) is to reveal such internal detail, within chloroplasts and other cytoplasmic organelles, and within nuclei, in a well-fixed specimen. In this method, at the moment the fractured surface thaws, the cells are unfixed. On thawing, soluble proteins diffuse out from the surface. However, since thawing takes place in the fixative, fixation of surface structure is very rapid. Tests with frozen chicken erythrocytes in a fibrin gel show individual
195
Figs. 1-4. Mesophyll cells of young radish leaf. Figures 1 and 2 compare glutaraldehyde fixation alone (Fig. 1) and glutaraldehyde + osmium fixation (Fig. 2). Both are fixed in slightly hypertonic solution, bringing the cytoplasm away from the cell wall. Figure 3 shows isotonic, and Fig. 4 hypotonic fixation (see discussion in text). V=vacuole, C = chloroplast, S = starch grain, L =lipid bodies, W = cell wall, P =outer surface of plasma membrane, T = inner surface of tonoplast, FDFf preparation. Magnifications : Fig. 1 x 9000, Figs. 2-4 x 11,000.
196
Freeze-fracture for SEM
cells cleaned of haemoglobin but, where a number of red cells are clumped together, the larger amount of haemoglobin becomes fixed before it can diffuse away (Hb in Fig. 10). The results obtained by this method with carrot protoplasts are shown in Figs. 6, 7, 8 and 9 and can be compared with the thin-section micrograph of a carrot protoplast in Fig. 5 . In the cell of Fig. 6 the fracture passed through a nucleus, in that of Fig. 7 through a large vacuole. In the preparation of Fig. 8 the fracture has passed through the cell above the nucleus, so that here the cytoplasmic surface of the Outer nuclear membrane is revealed. At the resolution of the present study the nuclear pores are not seen in Fig. 8, nor is the resolution adequate to show ribosomes on the nuclear surface. In the cytoplasm of these cells, the main features seen are mitochondria, lipid bodies organelles of comparable size in Fig. 5 ) and fibrous strands (Figs. 6, 7, 8 and 9). Figure 9 shows cytoplasmic detail observed on only one occasion, where beads or small VaCUOkS of relatively uniform size were seen along the strands. These cultured cells are in varying states of growth and activity, and we may, in the preparation of Fig. 9, have caught a particular growth phase. Typically the cytoplasm shows bodies of more varied size attached to the strands, as in Fig. 6. An immediate question arises as to whether these strands (seen running from vacuole to plasma membrane in Fig. 7 and from nucleus to plasma membrane in Fig. 8) represent genuine cytoplasmic components, or artefacts arising from DMSO treatment, freezing, thawing, or delayed fixation ? In other words, is it possible by this technique to show up the genuine cytoskeleton of cells, and if so, are our present methods of freezing and thawing sophisticated enough to achieve this result ? To attempt to answer this question we have begun a study of a simpler cell, the chicken erythrocyte. For turkey erythrocytes, thin-section micrographs of ghosts show fine strands running from nucleus to plasma membrane in the central region of the cell (Harris, 1971). Also a ring of microtubules runs around the cell periphery (Harris, 1971). F f T F micrographs of these cells are shown in Figs. 10, 11 and 12. Thin strands are indeed seen running from nucleus to plasma membrane (Fig. 11). Where the main part of a cell has fractured away, leaving a part of the rim behind, slightly thicker strands can be seen in the expected location of the microtubules (Fig. 12). There remains the possible criticism that the fibrin in which the cells are embedded might in some way penetrate into the cells to form the structures seen. However, identical fine fibrils, running from nucleus to plasma membrane, and the same thicker fibres running round the rim of the cell, which we believe to be the microtubules, are also seen in cells embedded in Agar. F f T F micrographs of leaf mesophyll cells are shown in Figs. 13 and 14. The tenuous cytoplasmic layers of leaf cells have proved difficult to preserve well, through the freezing and thawing which precedes fixation in the F f T F method. However, in the cell of Fig. 13 only a small tear is seen. Good structural preservation is achieved by using very thin slices of tissue to allow rapid infiltration, and rapid freezing and thawing. Thin-section controls show little ultrastructure change during DMSO infiltration, mainly a slight swelling of lipid bodies, provided the DMSO concentration is not brought above 257,. CONCLUSIONS A N D FUTURE PROSPECTS
The results for chicken erythrocytes (Figs. 10-12) allow the important conclusion that the F f T F method can, in appropriate circumstances, preserve cytoskeletal components, so that these are displayed in three dimensions for SEM
197
G. H. Haggis and Beverly Phipps- Todd
198
Freeze-fracture for SEM
Figs. 5-9. Carrot protoplasts. Figure 5 is a standard thin section, Figs. 6-9 FfTF preparations. In Fig. 6 the fracture passes through a nucleus, in Fig. 7 through a vacuole, and in Fig. 8 above a nucleus. Figure 9 shows cytoplasmic detail (see discussion in text). V = vacuole, F N = fracture through nucleus, N = cytoplasmic surface of outer nuclear membrane. Magnifications: Figs. 5-7 x 8000, Figs. 8 and 9 x 11,000.
14
199
Figs. 10-12. Chicken erythrocytes. Figure 10 shows haemoglobin diffusing out from the fracture face (see discussion in text). Figures 11 and 12 show ultrastructure details. H b = haemoglobin, N = cytoplasmic surface of outer nuclear membrane, SS = strands running from nucleus to plasma membrane, arrow = microtubules. FfTF preparation. Magnifications : Fig. 10 x 2000, Figs. 11 and 12 x 9000.
Figs. 13 and 14. Mesophyll cells of young radish leaf. FfTF preparation. Figure 13 shows only a small defect in preservation (arrowhead) arising during freezing and thawing. Vacuolar material is washed out in the FfTF method. C = fractured chloroplast. Magnifications: Fig. 13 x 5500, Fig. 14 x 11,000.
Freeze-fracture for SEM viewing. Fractures can be obtained through nuclei in these and other cells (Fig. 6) and through the cytoplasm, to show the outer nuclear surface (Fig. 8) and suggestive cytoplasmic detail (Fig. 9). Further detail should be seen in preparations of this type in a high-resolution SEM. We also have the possibility in the F f T F method, at the moment of thawing, of labelling intracellular structure with antibodies linked to markers, such as haemocyanin, which can be recognized in the SEM. ACKNOWLEDGMENTS
We are very grateful to S. Lesley for the supply of carrot protoplasts and to S. Becker and A. Greig for the supply of chicken erythrocytes.
References Haggis, G.H., Bond, E.F. & Phipps, B. (1976) Visualization of mitochondria1 cristae and nuclear chromatin by SEM. 9th I I T R I SEM Symposium, Chicago, p. 281: Harris, J.R. (1971) The ultrastructure of the erythrocyte. In: Physiology and Bzochemistry of the Domestic Fowl (Ed. by D. T. Bell and B. M. Freeman), Vol. 2, p. 835. Academic Press, New York. Humphreys, W.J., Spurlock, B.O. & Johnson J.S. (1974) Critical point drying of ethanolinfiltrated cryofractured biological specimens for scanning electron microscopy. 7th I I T R I SEM Symposium, Chicago, p. 275. Sybers, H.D. & Ashraf, M. (1973) Preparation of cardiac muscle for SEM. 6th IITRZ SEM Symposium, Chicago, p. 341. Veliky, I.A., Sandkvist, A. & Martin, S.M. (1969) Physiology of, and enzyme production by, plant cell cultures. Biotechnol. Bioeng. 11, 1247.
201
