J. Cell Set. 21, 437-448 (1976) 43-7 Printed in Great Britain A COMPARISON OF MEMBRANE FRACTURE FACES OF FIXED AND UNFIXED GLYCERINATED TISSUE A. S. BREATHNACH, M. GROSS, B. MARTIN AND C. STOLINSKI Department of Anatomy, St Mary's Hospital Medical School (University of London), London, Wz, England SUMMARY Fixed (glutaraldehyde, 3 %) and unfixed specimens of rat buccal epithelium, striated muscle, and liver, were cryoprotected with glycerol, freeze-fractured, and replicated without sublimation. A comparison of fracture faces of general plasma membranes, nuclear membranes, mitochondrial membranes, and membranes of rough endoplasmic reticulum revealed no significant differences as between fixed and unfixed material. Apart from some membranes of liver endoplasmic reticulum, there was no evidence of aggregation or redistribution of intramembranous particles in the unfixed material. The results demonstrate that chemical prefixation of tissues for freeze-fracture is not always necessary, or even desirable, and that glycerol may not be as deeply or directly implicated in particle aggregation as previously thought. Fixation with glutaraldehyde alters the cleaving behaviour of plasma membrane at desmosomes and tight junctions, but not at gap junctions. INTRODUCTION Aggregation, or redistribution of particles within membranes as seen in freezefracture replicas, has been attributed to treatment of tissues with glycerol as cryoprotectant, and the phenomenon is reported to be inhibited by previous fixation with glutaraldehyde (Gilula, Eger & Rifkin, 1975). This would appear to provide clear-cut grounds for advocating routine chemical fixation of glycerinated tissues. However, the matter is not quite as simple as this. Apart altogether from the facts that glutaraldehyde itself can alter the cleavage plane of some membranes, affect the distribution of particles as between fracture faces and even cause some degree of aggregation (Dempsey, Bullivant & Watkins, 1973; Staehelin, 1973; Parish, 1975), there are certain observations which suggest that glycerol may not be as deeply or directly implicated in particle aggregation, as has been thought. Experience indicates that not all membranes of an individual tissue are affected (Breathnach, Stolinski & Gross, 1972a) and that there may be tissue and other differences in susceptibility (Gilula et al. 1975); while variations in the degree of prominence of the phenomenon throughout a block of tissue further suggest that the local concentration of glycerol may be a prime factor. There is clear-cut evidence (Speth & Wunderlich, 1973; Wunderlich, Speth, Batz & Kleinig, 1973; Wunderlich, Wallach, Speth & Fischer, 1974) that temperature can Correspondence to: Professor A. S. Breathnach, Dept. of Anatomy, St Mary's Hospital Medical School, London, W2, England.
438 A. S. Breathnach, M. Gross, B. Martin and C. Stolinski influence the phenomenon, and since it is customary to infiltrate tissues with glycerol at low temperatures, it is possible that some, or all, of the effect is due to this. These considerations have led to a closer examination of the issues involved, with the longterm aim of devising a satisfactory standard method of processing tissues for freezefracture. Essential to such a project is the initial establishment of the appearance of tissues which have not been exposed to any of the agents under investigation, so as to obtain a reference base-line. This has recently been achieved (Stolinski & Breathnach, 1976), and this report is concerned with the next stage - a direct comparison of fixed and unfixed glycerinated tissues. MATERIALS AND METHODS Observations were made on tissues from 2 sources, namely, rat buccal epithelium with underlying muscle tissue, and rat liver. Some specimens of each tissue were fixed immediately after removal in 3 % buffered glutaraldehyde for 1 h, after which they were placed for 0-5 h in a 10 % glycerol solution in TC 199, and for a further 1 h in. a similar solution containing 25 % glycerol, both at 4 C. The unfixed specimens were placed directly in similar glycerol solutions after removal. Specimens were then loaded into thin-walled stainless steel tubes of 1-5 mm diameter to be rapidly frozen in liquid propane at 180 C C. Part of each specimen was left protruding by about 0-5 mm above the rim of the tubular holder in. order to allow for subsequent fracturing, which was performed in a module incorporating a cleaver mechanism mounted in a cold stage (Stolinski, 1975). Surfaces revealed by fracturing were not sublimated, but were directly replicated with platinum-carbon, and the tissue was separated (or digested away) from the replica according to the following procedure: (1) With the replica and tissue still in the holder, the tissue was slowly thawed, then placed in 25 % glycerol/199 solution and the tissue removed from the holder. (2) Replica and tissue were washed in distilled water. (3) Replica and tissue were placed for up to 12 h in a solution containing 5 % HNO a and 5 % H.aSO 4. (3 a) (optional for tough tissues). Replica and tissue were washed in distilled water, placed for 2-3 h in 40 % NaOH, and washed again in distilled water. (4) Replica and tissue were placed in a solution of equal parts of concentrated nitric and sulphuric acids and heated at 60 C for 0-5 h. By this time, all of the tissue was usually digested away, and the replica, after 3 washings in distilled water, was picked up on an uncoated grid for examination in the microscope. In all micrographs, direction of platinum shadowing is indicated by encircled arrow at bottom left. Fig. 1. Replica of complementary fracture faces of apposed plasma membranes of spinous layer cells of unfixed rat buccal epithelium. Ext, fracture face of membrane directed towards exterior of cell; Int, fracture face of plasma membrane directed towards interior of cell. Note higher concentration of general membrane-associated particles on Ext-face, and rounded or oval particle-free areas at sites of desmosomes (de). On the Int-face, desmosomes are represented by aggregates of closely packed particles (pa), ic, intercellular material. See also Fig. 9. x 33 600. Fig. 2. Replica of similar membrane fracture faces to those in Fig. 1 from buccal epithelium prefixed with glutaraldehyde. Labelling as in. Fig. 1. Note that distribution of general membrane-associated particles as between the fracture faces is similar to that in unfixed material, but that desmosomes are represented by particle aggregates (pa) on both faces. See also Fig. 10. x 33600
Membranes of fixed and unfixed tissue 43 Q 29-2
44 A. S. Breathnach, M. Gross, B. Martin and C. Stolinski RESULTS The illustrations are so arranged and selected as to permit direct comparison between similar features of fixed and unfixed tissues. The terminology used to describe membrane fracture faces (plasma and intracellular) is that suggested elsewhere (Stolinski & Breathnach, 1975), namely: 'Ext-face' is that fracture face of a membrane directed towards the material that is external to the substance enclosed by that membrane (the ' A-face' of other authors), and ' Int-face' is the fracture face of a membrane directed towards the material that is internal to, or enclosed by that membrane (the ' B- face' of other authors). General plasma membrane Examination of Figs. i-6, which illustrate fracture faces of general plasma membrane, reveals no significant differences in particle numbers or distribution as between fixed and unfixed glycerinated tissues. In each instance, particle density is greater on Ext-faces as compared with Int-faces, and there is no evidence of particle aggregation or redistribution within plasma membranes of unfixed glycerinated tissue. In general, individual particles appear to be more prominent with unfixed membranes, but this impression may to some extent be due to differences in angles of shadowing. Specialized contacts With unfixed material, as previously reported (Breathnach, Stolinski & Gross, 19726), desmosomes are revealed as localized aggregates of closely packed 8-nm particles in Int-faces of the plasma membrane, with particle-free areas on corresponding parts of complementary Ext-faces (Figs. 1,9). With fixed material, the appearance is totally different, the desmosomal particles being more or less equally distributed between the 2 fracture faces (Figs. 2, 10). It has already been reported that the appearance of the tight junction depends upon whether or not the tissue was chemically fixed before freezing (Staehelin, 1973) and the present observations confirm this. With unfixed material, almost continuous rows of particles are present in furrows on the Int-face (Fig. 11), and only the occasional particle on the complementary Ext-face. In fixed material the Ext-face exhibits con- Figs. 3-6. Fracture faces of plasma membranes of striated muscle cells from rat cheek for comparison of unfixed (Figs. 3, 5) and fixed (Figs. 4, 6) material. Many pinocytotic vesicles are visible on the fixed membranes. All x 82500. Fig. 3. Ext-face, unfixed. Particles are randomly distributed with no evidence of aggregation, and their numbers are approximately equal to those present on Ext-faces of fixed material (Fig. 4). Fig. 4. Ext-face of fixed plasma membrane. Fig. 5. Int-face, unfixed. Particles are fewer in number than on Ext face (Fig. 3), and the general appearance is very similar to that of corresponding fracture face of fixed material (Fig. 6). Fig. 6. Int-face of fixed plasma membrane.
Membranes of fixed and unfixed tissue 441
44* A. S. Breathnach, M. Gross, B. Martin and C. Stolinski
Membranes of fixed and unfixed tissue 443 tinuous ridges of semi-particulate material (Fig. 12), and the Int-face, furrows devoid of particles. No differences were observed between the gap junctions of fixed and unfixed specimens of the present series (Figs. 7, 8). The size, spacing, and numbers of pits and particles on the appropriate fracture faces were similar. Nuclear membrane Since the nuclear membrane consists of 2 leaflets - outer and inner - complementary fracture faces of both leaflets, i.e. 4 faces in all, are revealed in replicas. With the present material, the arrangement and distribution of particles on the faces were similar for fixed and unfixed membranes; thus, in both instances there were fewer particles on the Ext-face of the outer leaflet, and more on the Ext-face of the inner leaflet (Figs. 13, 14), and the complementary Int-faces (Figs. 15, 16) in each case exhibited a reverse numerical distribution. Rough endoplasmic reticulum Lamellar membranes of the rough endoplasmic reticulum are revealed in replicas as complementary (Ext and Int) fracture faces separated by a step which includes the cisternal lumen. In both fixed and unfixed specimens the numerical distribution of particles as between the 2 faces was similar, i.e. there were significantly fewer on the Ext-face as compared with the Int-face (Figs. 17, 18). In the unfixed liver material, a definite tendency towards aggregation of particles to produce a 'mosaic' pattern was observed in connexion with many Int-faces, but not nearly to the same degree as previously reported and illustrated (Breathnach et al. 1972a). Fig. 7. Gap junction from unfixed rat liver, exhibiting characteristic pits [pi) on the Intface of the plasma membrane, and particles (pa) on the complementary Ext-face. x 175000. Fig. 8. Gap junction from fixed rat liver, Appearance essentially similar to that of unfixed material, x 175000. Fig. 9. Deamosome from unfixed buccal epithelium showing aggregation of closely packed particles (pa) on Int-face, and almost complete absence of particles on desmosomal area (de) of Ext-face. See also Fig. 1. x 122000. Fig. 10. Desmo8ome from fixed buccal epithelium. Note presence of particles (pa) on both (Ext, Int) fracture faces. See also Fig. 2. x 122000. Fig. 11. Linear arrays of tight junction particles on Int-face of unfixed liver cell plasma membrane. On the complementary Ext-face, a very occasional particle would be seen, x 150000. Fig. 12. Tight junction 'ridges' on Ext-face of fixed liver cell plasma membrane. In fixed material the major feature is on this face in contrast to unfixed material (Fig. 1 ij. x 150000.
A. S. Breathnach, M. Gross, B. Martin and C. Stolinski
Membranes of fixed and unfixed tissue 445 Mitochondria Complementary (Ext and Int) fracture faces of mitochondrial membranes, each representing a combination of faces of the outer and inner membranes, exhibited essentially similar features in both fixed and unfixed specimens (Figs. 19-21). DISCUSSION The only significant differences observed between membranes of the present fixed and unfixed glycerinated specimens of organized tissue relate to two of the specialized contacts-the tight junction and the desmosome, Otherwise, comparable membranes exhibited essentially similar features and, with the exception of some lamellar membranes of rough endoplasmic reticulum of liver, there was no evidence of the type of redistribution or aggregation of particles usually attributed to glycerination. The results indicate that chemical prefixation of tissue cryoprotected by infiltration with glycerol is not always essential. The fact that particle redistribution was much more evident in an earlier study of unfixed material (Breathnach et al. 1972 a) in which the tissue was exposed to higher concentrations of glycerol for longer periods than in the present instance, suggests that local concentration of glycerol, and overall time of permeation may be important factors in causing the phenomenon. In this connexion, therefore, it would seem desirable to institute experiments aimed at establishing the rate of penetration of glycerol into standard size blocks of representative tissues under controlled conditions of temperature, etc., and the time required to reach various levels of concentration up to saturation. This would provide essential background information for investigating the possibility that there may be a critical level of local concentration of glycerol which is adequate for cryoprotection, but which does not produce redistribution of particles in those membranes which are susceptible to the phenomenon. The different appearances of fractured tight junctions and desmosomes, as between fixed and unfixed specimens, raises the question as to which image should be regarded Fig. 13. Fracture faces directed towards the exterior (i.e. towards the cytoplasm (cy) of outer (Ext 0) and inner (Ext 1) membranes of nucleus from unfixed buccal epithelium. The face of the outer membrane has significantly fewer particles, x 127000. Fig. 14. Fracture faces directed towards the exterior (i.e. towards the cytoplasm (cy)) of outer (Ext 0) and inner (Ext 1) membranes of nucleus from fixed buccal epithelium. The distribution of particles between the faces, and the overall general appearance closely resembles that of unfixed material (Fig. 13). x 127000. Fig. 15. Fracture faces directed towards the interior (i.e. towards the nucleoplasm) of outer (Int o) and inner (Int 1) membranes of nucleus from unfixed buccal epithelium. cy, cytoplasm. The face of the outer membrane has a dense population of particles as compared with that of the inner membrane, x 92000. Fig. 16. Fracture faces directed towards the interior (i.e. towards the nucleoplasm) of outer (Int 0) and inner (Int 1) membranes of nucleus from fixed buccal epithelium. cy, cytoplasm. The distribution of particles between the faces, and the overall general appearance closely resemble those of unfixed material (Fig. 15). x 92000.
A. S. Breatftnach, M. Gross, B. Martin and C. Stolinski
Membranes of fixed and unfixed tissue 447 as the more 'normal'. In material rapidly frozen without either prior chemical fixation or infiltration with glycerol, and exhibiting no damage due to ice crystallization, the cleaving behaviour and appearance of the tight junction is identical with that seen in the present unfixed glycerinated material (Stolinski & Breathnach, 1976). This certainly indicates that glutaraldehyde fixation causes a re-arrangement of internal structural components of the membrane at the tight junction. Unfortunately, to date, no micrographs are available of desmosomes from unfixed, non-glycerinated tissue, for comparison with the present specimens, However, in unfixed glycerinated epidermis a distinct difference in cleaving behaviour and appearance of the desmosome has been demonstrated as between stratum corneum and underlying strata (Breathnach, Goodman, Stolinski & Gross, 1973). In the non-cornified strata of epidermis the desmosome appears as in the present unfixed specimens, i.e. close-packed particles are present only on the Int-fracture face. In the stratum corneum of unfixed epidermis, however, desmosomal particles are present on both fracture faces, i.e. the appearance is identical with that of desmosomes of lower strata of the present fixed buccal epithelium. The difference in cleaving behaviour of the desmosome of stratum corneum of unfixed epidermis has been attributed to an alteration in the internal structure of the plasma membrane associated with keratinization (Breathnach et al. 1973). The fact that a similar alteration is produced by fixation in desmosomes of lower strata of the present material again indicates a modification of membrane internal structure produced by glutaraldehyde, and suggests that the appearances in unfixed material are more likely to represent the 'normal'. In marked contrast to these differences relating to tight junctions and desmosomes, is the identical appearance of the gap junction in both fixed and unfixed glycerinated material, and in material exposed only to rapid freezing (Stolinski & Breathnach, 1976). One might conclude from the above discussion that it is probably advisable as a routine to examine unfixed as well as fixed specimens of all cryoprotected material being investigated by the freeze-fracture replication technique. It seems evident that Fig. 17. Replica of lamellar membranes of rough endoplasmic reticulum of unfixed rat liver cell. Ext, fracture face of lamellar membrane directed towards the exterior, i.e. towards the cytoplasm, cy. Int, fracture face of lamellar membrane directed towards the interior i.e. towards the cisternal lumen, lu. The Int-face has a denser population of particles, and they exhibit a slight tendency towards aggregation with the production of a ' mosaic pattern' of smooth areas outlined by rows of particles, x 98 000. Fig. 18. Replica of lamellar membranes of rough endoplasmic reticulum of fixed rat liver cell. Labelling as in Fig. 17. The distribution of particles as between the 2 fracture faces, and the overall general appearance closely resemble those of unfixed material (Fig. 17). Fig. 19. Ext and Int fracture faces of mitochondrial membranes from unfixed rat buccal epithelium, x 92 500. Fig. 20. Ext fracture face of mitochondrial membrane from fixed rat buccal epithelium, x92500. Fig. 21. Int fracture face of mitochondrial membrane from fixed rat buccal epithelium, x92500.
448 A. S. Breathnach, M. Gross, B. Martin and C. Stolinski both fixatives and cryoprotectants may each induce alterations in membrane structure which affect appearances in micrographs, and it is important that these should be documented for individual membranes and tissues, since it seems there may be cell, tissue, and species differences in susceptibility to their effects. REFERENCES BREATHNACH, A. S., GOODMAN, T., STOLINSKI, C. & GROSS, M. (1973). Freeze-fracture replication of cells of stratum corneum of human epidermis. J. Anat. 114, 65-81. BREATHNACH, A. S., STOLINSKI, C. & GROSS, M. (1972 a). Freeze-fracture replication of rough endoplasmic reticulum of mouse liver cells. J. Cell Sci. 11, 477-489. BREATHNACH, A. S., STOLINSKI, C. & GROSS, M. (19726). Ultrastructure of fetal and post-natal human skin as revealed by the freeze-fracture replication technique. Micron 3, 287-304. DEMPSEY, G. P., BULLIVANT, S. & WATKINS, W. B. (1973). Endothelial cell membranes: polarity of particles as seen by freeze-fracturing. Science, N.Y. 179, 190 192. GILULA, N. B., EGER, R. R. &RIFKIN, D. B. (1975). Plasma membrane alteration associated with malignant transformation, in culture. Proc. natn. Acad. Sci. U.S.A. 72, 3594-3598. PARISH, G. R. (1975). Changes of particle frequency in freeze-etched erythrocyte membranes after fixation. J. Microscopy 104, 245-256. SPETH, V. & WUNDERLICH, F. (1973). Membranes of Tetrahymena. II. Direct visualization of reversible transitions in biomembrane structure induced by temperature. Biochim. biophys. Acta 291, 621-628. STAEHELIN, L. A. (1973). Analysis and critical evaluation of the information contained in freeze-etch micrographs. In Freeze-etching, Techniques and Applications (ed. E. L. Benedetti & P. Favard), pp. 113-134. Paris: Soci6t6 Francaise de Microscopie filectronique. STOLINSKI, C. (1975). A freeze-fracture replication apparatus for biological specimens. J. Microscopy 104, 235-244. STOLINSKI, C. & BREATHNACH, A. S. (1975). Freeze-fracture Replication of Biological Tissues. London: Academic Press. STOLINSKI, C. & BREATHNACH, A. S. (1976). Freeze-fracture replication of organized tissue without cryoprotection. J. Anat. (in Press). WUNDERLICH, F., SPETH, V., BATZ, W. & KLEINIG, H. (1973). Membranes of Tetrahymena. III. The effects of temperature on membrane core structures and fatty acid composition of Tetrahymena cells. Biochim. biophys. Acta 298, 39 49. WUNDERLICH, F., WALLACH, D. F. H., SPETH, V. & FISCHER, H. (1974). Differential effects of temperature on the nuclear and plasma membranes of lymphoid cells. A study by freezefracture electron microscopy. Biochim. biophys. Acta 373, 34-43. (Received 5 January 1976)