Dorsal Roots of the Rabbit Investigated by Freeze-substitution ' S. K. MALHOTRA AND A. VAN HARREVELD Kerckhoff Laboratories of the Biological Sciences, California Institute of Technology, Pasadena, California ABSTRACT Dorsal roots from adult rabbits prepared by a freeze-substitution technique and studied by electron microscopy have been compared with those obtained by standard fixation with 0 ~ 0 4 . The essential features of myelinated and nonmyelinated fibers and their associated Schwann cells are similar in appearance after both types of fixation. There are, however, slight differences. The extracellular clefts between the nonmyelinated axons and their associated Schwann cells are wider i n freeze-substituted roots (200 to 250 A versus 100 to 150 A ) . The major dense lines in the myelin sheath are wider (45 to 55 A versus about 30 A ) and their radial repeat period is larger (145 to 155 A versus 100 to 120 A) i n freeze - substituted material than i n routinely fixed nerves. Electron microscopy of cerebellar cortex processed by a freeze-substitution technique revealed considerably more extracellular space i n the molecular layer than is usually seen in electron micrographs obtained by routine chemical fixation (Van Harreveld, Crowell and Malhotra, '65). I n view of these findings it was of interest to investigate peripheral nervous tissue by the same technique of freeze-substitution and to compare the results with those obtained by standard chemical fixation. MATERIALS AND METHODS Dorsal roots from adult rabbits have been used which are similar to peripheral nerves, but contain less connective tissue and lack the epineurium found in peripheral nerves (Nathaniel and Pease, '63; Gamble, '64). The study of frozen material has to be restricted to the superficial layer of the tissue (about 10 thick) because the fine structure in the deeper parts is disorganized by ice crystals (Van Harreveld and Crowell, '64). The absence of a perineurium in the root is therefore a distinct advantage. Lumbar roots were taken from rabbits narcotized with pentobarbital (50 mg/kg) and frozen within 1 to 2 minutes after removal from the body. The procedure for freezing and the subsequent treatment of the tissue have been described previously (Van Harreveld and Crowell, '64; Van Harreveld, Crowell and ANAT. REC., 152: 283-292. Malhotra, '65). The tissue was frozen by bringing it in contact with a mirror smooth silver surface precooled to about -207°C by liquid nitrogen under reduced pressure. The root was placed on a small block of agar kept in a trough of aluminum foil attached to the apparatus that carried the tissue to the freezing surface. The frozen tissue was subjected to substitution fixation for two days at -85°C in acetone containing 2% OsO,. The material was subsequently kept for a few hours a t -25°C and then warmed to room temperature while still in the substitution medium. The excess osmium was removed by bathing the tissue for 3 to 4 hours in repeatedly changed acetone before the roots were embedded in Maraglas (Freeman and Spurlock, '62). For controls, freshly dissected roots were fixed ex situ in a 1% solution of OsOl in acetate-Verona1 buffer (pH 7 . 3 to 7.4) made isotonic by addition of indifferent salts (Zetterqvist, '56). The tissue blocks were sectioned on a n LKB Ultrotome in such a way that the nerve fibers were cut transversely. Sections about 1 thick were stained with methylene blue and Azure B (Richardson, Jarett and Finke, '60) for examination by light microscopy. Only those blocks of tissue that met the criteria of satisfactory freezing (absence of ghosts of ice crystals in 1 Supported by a grant from The National Science Foundation (GB 2055). 283 284 S. K . M A L H O T R A AND A . V A N HARREVELD the surface layer) were examined by electron microscopy. Ultrathin sections were stained with lead citrate (Reynolds, '63). RESULTS The structure of the dorsal root has been described previously and is essentially similar to that of peripheral nerves (Nathaniel and Pease, '63). This study will be restricted to the myelinated and nonmyelinated nerve fibers and the Schwann cells associated with them. Little attention has been paid to the connective tissue component, which is not very prominent in the present electron micrographs. The preservation of cellular structural detail by freeze-substitution is comparable with that observed in electron micrographs produced by standard fixation (Van Harreveld, Crowell and Malhotra, '65; Malhotra and Van Harreveld, '65). Nonmyelinated fibers. The axons of the nonmyelinated fibers and the Schwann cells are separated by extracellular spaces which are of more or less uniform width (figs. 2 and 3). Measurements at places where the membranes are sharp and parallel to each other give values of 200 to 250 A for these extracellular clefts. An extracellular space of similar width is found between the membranes forming the mesaxons in Schwann cells enclosing nonmyelinated fibers. Occasionally, however, the two apposing membranes come closer together and even touch each other over a small distance (fig. 3, arrows), where they may form five layered tight junctions (Farquhar and Palade, '63; Karlsson and Schultz, '64). A few nonmyelinated fibers can be seen in figures 1 and 2 which are not associated with Schwann cells. In the deeper regions of the root, ice crystals are formed which disorganize the fine structural details (fig. 8). Ghosts of ice crystals can be recognized as light (unstained) areas bounded by sharp dark outlines which represent concentrated protoplasm. The plasma membrane of the nonmyelinated fiber and the Schwann cell membrane are at places separated by a n intervening space which may be as wide as 300 A or more. At other locations the intervening space is completely obliterated and the two apposing membranes form a five layered tight junction over large distances. The prevalence of tight junctions in the deeper regions in contrast to their rare occurrence in the superficial layers differs from the findings of Elfvin ('63) who observed many tight junctions in the surface layers of frozen-dried preparations of splenic nerve. The tight junctions observed in the present material are most likely artifacts without functional significance as contrasted to e.g., electrical synapses (Bennett et al., '63; Dewey and Barr, '64). Myelinated fibers. The myelin sheath after freeze-substitution (figs. 5 and 6 ) shows essentially the same pattern as described for instance by Robertson ('58) and Sjostrand ('60) in material prepared by routine fixation in OsO, (fig. 7) or by freeze-drying (Elfvin, '63). The well defined major dense lines, which are formed by the intimate apposition of the dense layer of the Schwann cell membrane that faces the cytoplasm, measure between 45 A and 55 A. Each major dense line is separated from the next by a distance of 90 A to 100 A; and the distance from the middle of one major dense line to the middle of the next is about 150 A, which is the repeating period of the major dense lines in freeze-substituted material (figs. 5 and 6 ) . The about 100 A wide pale area between two major dense lines is bisected by a n inconspicuous intraperiod line, which is not very well preserved in tissues fixed by OsO+ In a few electron micrographs, some of the major dense lines are locally resolved into two thin dense lines separated by a lighter line (fig. 6 ) . Such a n appearance in the myelin sheath which has also been found by Robertson ('57) and Sandborn et al. ('64) could be due to incomplete fusion of the apposed cytoplasmic surfaces of the Schwann cell membranes. Differences between roots fixed by freezesubstitution and by standard methods. Although the preservation of myelinated and nonmyelinated fibers by freeze-substitution and direct Os04 fixation is essentially alike, there are a few differences. The clefts between nonmyelinated axons and Schwann cells in roots fixed e x situ by OsOl are in general 100 to 150 A wide (fig. 4 ) . Similar values have been reported in the literature for peripheral nerves (see e.g., Sjostrand, '60; Elfvin, '61, '63; Finean, '61; Gray, '64; Peters, '64; Robertson, '64). FREEZE-SUBSTITUTED DORSAL ROOTS The extracellular cleft between the nonmyelinated axon and the Schwann cell and the space in the mesaxons is generally 50 to 100 A wider in freeze-substituted material than in routinely fixed tissue. A direct comparison of this difference in the width of the clefts around the nonmyelinated fibers can be made from figures 3 and 4. The major dense lines in the myelin sheath are slightly wider in freeze-substituted nerve fibers (45 to 55 A versus 30 A ) . The repeat period of the major dense lines is 145 to 155 A after freeze-substitution which is a little larger than the 100 to 120 A period determined in material fixed with OsO,. Figures 6 and 7 which are at the same magnification demonstrate these differences in the myelin sheaths. This discrepancy can be explained by different degrees of shrinkage of the myelin sheath during freeze-substitution and routine fixation. The repeat period for the major dense lines obtained by freeze-substitution and by freeze-drying (Elfvin, '63) is closer to the radial repeat period of 180 A to 185 A determined by x-ray diffraction in fresh mammalian nerve (Schmitt, Bear and Palmer, '41 ; Fernandez-Moran and Finean, '57) than the 100 A to 120 A repeat period observed in standard electron micrographs (Finean, '61; Finean and Robertson, '58). DISCUSSION The clefts between the Schwann cell membranes and nonmyelinated fibers are somewhat wider in freeze-substituted than in routine OsO, fixed tissue. It h a s been suggested that the water distribution i n the living tissue may be better preserved by freeze-substitution than by routine chemical fixation (Van Harreveld, Crowell and Malhotra, '65). The larger radial repeat period in the myelin sheath in freezesubstituted material, which approaches the repeat period observed with x-ray diffraction in fresh nerve, seems to support this. I n the cerbellum the large extracellular space observed in freeze-substituted material was not uniformly distributed i n the tissue but was found especially between the axons of granular layer cells which are present in groups (Van Harreveld, Crowell and Malhotra, '65). These axons are not closely associated with glia cells. Dendritic elements and presynaptic endings showed 285 in general a close relationship with glia from which these elements were in most instances separated by the usual narrow (100 A ) clefts. In the dorsal root the nonmyelinated fibers are associated with the Schwann cells which can be compared with the glia of the central nervous system and from which the nerve fibers are separated by 200-250 A wide clefts. The relationship of associated cells and nervous elements is in the dorsal roots as well as in the cerebellar cortex characterized by a separation of the elements by relatively narrow clefts. ACKNOWLEDGMENTS We greatly appreciate the valuable technical help of Mrs. Josephine Pagano and Miss Judy Schwaffel. LITERATURE CITED Bennett, M. V. L., E. Aljure, Y . Nakajima and G. D. Pappas 1963 Electronic junctions between teleost spinal neurons: Electrophysiology and ultrastructure. Science, 141: 262-264. Dewey, M. M., and L. Barr 1964 A study of the structure and distribution of the nexus. J. Cell Biol., 23: 553-585. Elfvin, L. G . 1961 Electron microscopic investigation of the plasma membrane and myelin sheath of the autonomic nerve fibres in the cat. J. Ultrastr. Res., 5: 388-407. 1963 The ultrastructure of the plasma membrane and myelin sheath of peripheral nerve fibres after fixation by freeze-drying. J. Ultrastr. Res., 8: 283-304. Farquhar, M. G., and G . E. Palade 1963 Functional complexes i n various epithelia. J. Cell Biol., 18: 375-412. Fernandez-Moran, H., and J. B. Finean 1957 Electron microscopic and low angle x-ray diffraction studies of the nerve myelin sheath. J. Biophys. Biochem. Cytol., 3 : 725-748. Finean, J. B. 1961 Chemical ultrastructure in living tissues. Charles C Thomas, Springfield, Illinois. Finean, J. B., and J. D. Robertson 1958 Lipids and the structure of myelin. Brit. Med. Bull., 14: 267-273. Freeman, J. S . , and B. 0. Spurlock 1962 A new epoxy embedment for electron microscopy. J. Cell Biol., 13: 437-443. Gamble, H. J. 1964 Comparative electron microscopic observations on the connective tissues of a peripheral nerve and a spinal nerve root i n the rat. J. A.nat. London, 98: 17-25. Gray, E. G. 1964 Electron microscopy of cell surface. Endeavour, 23: 61-65. Karlsson, U., and R. Schultz 1964 Plasma membrane apposition in the central nervous system after aldehyde perfusion. Nature, 201 : 12301231. 286 S. K. MALHOTRA AND A. VAN HARREVELD Malhotra, S. K., and A. Van Harreveld 1965 Some structural features of mitochondria in tissues prepared by freeze-substitution. J. Ultrastr. Res., 12: 4 7 3 4 8 7 . Nathaniel, E. J. H., and D. C. Pease 1963 Degenerative changes i n rat dorsal roots during Wallerian degeneration. J. Ultrastr. Res., 9: 511-532. Peters, A. 1964 An electron microscope study of the peripheral nerves of the hag fish (Myxine glutinosa, L.). Quart. J. Exp. Physiol., 49: 35-42. Reynolds, E. S. 1963 The use of lead citrate at high pH as an electron opaque stain in electron microscopy, J. Cell Biol., 17: 208-213. Richardson, K. C., L. Jarett and E. H. Finke 1960 Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol., 35: 313-323. Robertson, J. D. 1957 The ultrastructure of frog muscle spindles, motor endings and nerve fibres. J. Physiol., 137: 6P-8P. - 1958 Structural alterations i n nerve fibres produced by hypotonic and hypertonic solutions. J. Biophys. Biochem. Cytol., 4: 349364. - 1964 Unit membranes: A review with recent new studies of experimental alterations and a new sub-unit structure in synaptic membranes. In: Cellular Membranes in Development. Edited by M. Locke. Academic Press, New York, pp. 1-81. Sandborn, E., P. F. Koen, J. D. McNabb and G. Moore 1964 Cytoplasmic microtubules i n mammalian cells. J. Ultrastr. Res., 11: 123-138. Schmitt, F. O . , R. S. Bear and K. J. Palmer 1941 X-ray diffraction studies o n the structure of the nerve myelin sheath. J. Cell. and Comp. Physiol., 18: 31-42. Sjostrand, F. S. 1960 Electron microscopy of myelin and nerve cells and tissues. In: Modern Scientific Aspects of Neurology. Edited by J. N. Cummings, Edward Arnolds, London, pp. 188-231. Van Harreveld, A., and J. Crowell 1964 Electron microscopy after rapid freezing on a metal surface and substitution fixation. Anat. Rec., 149: 381-386. Van Harreveld, A., J. Crowell and S. K. Malhotra 1965 A study of extracellular space in central nervous tissue by freeze-substitution. J. Cell Biol., 25: 117-137. Zetterqvist, H. 1956 The ultrastructural organization of the columnar absorbing cells of the mouse jejunum. Karolinska Institute, Stockholm. PLATE I EXPLANATION O F FIGURES 1 Dorsal root prepared by freeze-substitution. The surface of the root shows on the bottom of the figure. Nonmyelinated and most of the myelinated fibers are well preserved. A thin layer of Schwann cell cytoplasm surrounds the myelinated axons. The nonmyelinated axons are present i n bundles generally invested by Schwann cell cytoplasm. Calibration line 14. 2 Shows a group of nonmyelinated fibers prepared by freeze-substitution. Most of the fibers are associated with Schwann cell cytoplasm but a few in the upper half of the figure do not contact Schwann cells. But for a few small regions the spaces between axons and Schwann cells or between apposed Schwann cell membranes are of uniform width. In the upper corner of the micrograph part of a myelinated fiber is present. Calibration line 0.5 p . FREEZE-SUBSTITUTED DORSAL ROOTS S. K. Malhotra and A. Van Harreveld PLATE 1 PLATE 2 EXPLANATION O F FIGURES 288 3 A magnified view of a small group of nonmyelinated fibers enclosed by Schwann cell cytoplasm ( freeze-substitution). The cleft between nerve fiber and Schwann cell is of uniform width except for one small area where the two apposing membranes are separated by a very narrow space (arrow 1). Arrow 2 shows apposing Schwann cell membranes in close contact. The figure also shows a small part of a well preserved myelin sheath with a thin layer of Schwann cell cytoplasm. Calibration line 0.5 p. 4 Shows dorsal root prepared by ex situ fixation in isotonic osmium tetroxide. The results are essentially similar to those observed in well frozen tissue. The spaces between nonmyelinated axons and Schwann cell membranes and in the mesaxon are more or less uniformly wide. Part of a myelinated fiber is present on the right. Calibration 0.5 p. FREEZE-SUBSTITUTED DORSAL ROOTS S. K. Malhotra and A. Van Harreveld PLATE 2 289 PLATE 3 EXPLANATION O F FIGURES 290 5-6 Show typical appearance of the myelin sheath in freeze-substituted preparations. Intraperiod lines are faintly visible. The axonal membrane (arrow 1) appears slightly thicker than the Schwann cell membrane (arrow 2 i n fig. 5). The basement membrane is exceptionally well preserved i n figure 5. Arrow in figure 6 indicates split i n a major dense line. Calibration line indicates 0.1 ,u. 7 Typical appearance of the myelin sheath fixed ex situ in isotonic osmium tetroxide. Calibration line is 0.1 p . FREEZE-SUBSTITUTED DORSAL ROOTS S . K. Malhotra and A. Van Harreveld PLATE 3 291 FREEZE-SUBSTITUTED DORSAL ROOTS S. K. Malhotra and A. Van Harreveld 8 292 PLATE 4 Is taken at some distance from the surface of a dorsal root subjected to freeze-substitution. The ghosts of ice crystals are obvious. A comparison with figures 2 and 3 shows that the spaces between axons and Schwann cells have become irregular, wider at some places and showing no space over large areas at others (tight junctions). Calibration line is 0 . 2 ~ .