Myelin membrane structure and composition correlated: a phylogenetic study.

We have correlated myelin membrane structure with biochemical composition in the CNS and PNS of a phylogenetic series of animals, including elasmobranchs, teleosts, amphibians, and mammals. X-ray diffraction patterns were recorded from freshly dissected, unfixed tissue and used to determine the thicknesses of the liquid bilayer and the widths of the spaces between membranes at their cytoplasmic and extracellular appositions. The lipid and protein compositions of myelinated tissue from selected animals were determined by TLC and sodium dodecyl sulfate-polyacrylamide gel electrophoresis/immunoblotting, respectively. We found that (1) there were considerable differences in lipid (particularly glycolipid) composition, but no apparent phylogenetic trends; (2) the lipid composition did not seem to affect either the bilayer thickness, which was relatively constant, or the membrane separation; (3) the CNS of elasmobranch and teleost and the PNS of all four classes contained polypeptides that were recognized by antibodies against myelin P0 glycoprotein; (4) antibodies against proteolipid protein (PLP) were recognized only by amphibian and mammalian CNS; (5) wide extracellular spaces (ranging from 36 to 48 A) always correlated with the presence of P0-immunoreactive protein; (6) the narrowest extracellular spaces (approximately 31 A) were observed only in PLP-containing myelin; (7) the cytoplasmic space in PLP-containing myelin (approximately 31 A) averaged approximately 5 A less than that in P0-containing myelin; (8) even narrower cytoplasmic spaces (approximately 24 A) were measured when both P0 and 11-13-kilodalton basic protein were detected; (9) proteins immunoreactive to antibodies against myelin P2 basic protein were present in elasmobranch and teleost CNS and/or PNS, and in mammalian PNS, but not in amphibian tissues; and (10) among mammalian PNS myelins, the major difference in structure was a variation in membrane separation at the cytoplasmic apposition. These findings demonstrate which features of myelin structure have remained constant and which have become specifically altered as myelin composition changed during evolutionary development.

A survey of neurological mutant mice. II. Lipid composition of myelinated tissue in possible myelin mutants.

The lipids of white matter and peripheral nerve from neurological mutant mice with possible myelin abnormalities were analyzed by thin-layer chromatography and quantitated by densitometry. Eight mutants had major abnormalities in the central nervous system (CNS) and/or peripheral nervous system (PNS) tissues examined (optic nerve, and trigeminal and sciatic nerves). In the optic nerve of axJ/axJ, there were increases of 20-30% in the levels of the major phospholipids; peripheral nerve was normal. In bc3J/bc3J CNS, the major phospholipids and cholesterol were increased by 25-40%; the PNS was normal. In myd/myd CNS, there were increases of about 20% in the levels of both forms of cerebrosides and in the major phospholipids; in the PNS the lipids were normal. ot/ot CNS had 20-40% reductions of all the glycolipids and minor alterations in some of the phospholipids and cholesterol; the PNS had 20% losses of both forms of cerebrosides. In the PNS of ji/ji, there were decreases of 10-40% among the glycolipids and of 15-25% in three of the major phospholipids; the CNS was virtually normal. In the PNS of dtJ/dtJ, vb/vb and wr/wr, almost all lipids were significantly decreased. The CNS of dtJ/dtJ and vb/vb were normal; wr/wr had minor reductions of certain glycolipids and phospholipids. Six mutants had relatively minor lipid abnormalities in their myelinated tissues. In cr/cr PNS, there were elevated levels of the cerebrosides and major phospholipids; the CNS was virtually normal. In db/db CNS and PNS, there were reduced levels of the nonhydroxy forms of cerebroside and sulfatide. The major change in htr/htr was the elevation of all the glycolipids in the CNS. In the CNS of Lc/+, nonhydroxy cerebroside was reduced. In shm/shm PNS, nonhydroxy sulfatide was elevated and there were small decreases in some of the phospholipids. wl/wl CNS showed decreases among most of the glycolipids. Mutants homozygous for du, mto, spa and tg had virtually normal lipid levels in both the optic and peripheral nerves. Cholesterol ester, lysophospholipids and other unusual lipid species were not detected in any of the mutants. The plasmalogen forms of ethanolamine and choline phosphatides were at normal levels in all mutants that otherwise had significant alterations among their lipids. Although many alterations in lipid composition were found in these mutants, the changes were moderate compared to the classical myelin mutants and indicate that none of the mutants are severely myelin-deficient.(ABSTRACT TRUNCATED AT 400 WORDS)

A survey of neurological mutant mice. I. Lipid composition of myelinated tissue in known myelin mutants.

The lipids of white matter and peripheral nerve from mutant mice with known myelin deficiencies were analyzed by one- and two-dimensional high-performance thin-layer chromatography and quantitated by densitometry. In optic nerve, the mutants jp/Y, jpmsd/Y, qk/qk, shi/shi and shimld/shimld, which have severe central nervous system (CNS) myelin deficiency, had a common pattern of lipid loss: cerebrosides and sulfatides (hydroxy and nonhydroxy forms) were generally reduced by 70-95% or more; most phospholipids were diminished by 15-55%, and cholesterol was reduced by 35-60%. Only in the CNS of jp/Y and jpmsd/Y did cholesterol ester accumulate. In peripheral nerve, the lipid composition varied markedly among these mutants. In jp/Y there was no change, while in jpmsd/Y there was a 5-15% loss among the phospholipids and cholesterol. Homozygous qk had reductions of 75-85% in the nonhydroxy forms of cerebroside and sulfatide, a 130% increase in hydroxy sulfatide, and a 55% loss of sphingomyelin. In shi/shi and shimld/shimld homozygotes, the glycolipids were altered by +/- 20%, most phospholipids and cholesterol were reduced by 5-15%, and sphingomyelin was reduced by 40%. Tr and TrJ showed 35-90% reductions in most lipid classes of the peripheral nervous system; CNS lipid composition was normal. Homozygous twi had a uniform loss of most lipid classes in both optic (generally 10-20%) and trigeminal nerves (generally 40-55%); cerebrosides did not accumulate in these tissues. dy/dy had a 10-20% reduction of cerebrosides in trigeminal nerve trunk. The CNS of dy homozygotes had 10-35% increases in specific classes of glycolipids and phospholipids, and in cholesterol. None of the mutants showed detectable levels of lysophospholipids or other unusual lipid species. The fractions of ethanolamine and choline phosphatides in the plasmalogen form were close to normal in all mutants.

Spasmodic, a mutation on chromosome 11 in the mouse.

A new recessive mutation, spasmodic (spd), producing behavior that mimics that of the neurological mutation spastic (spa) with rapid tremors, stiff posture, and difficulty in righting, arose spontaneously in strain A/HeJ at the Jackson Laboratory in 1979. It is not an allele of spa and linkage tests show that this mutation is located close to vestigial tail (vt) near the center of chromosome 11. Additional genetic tests show that it is not an allele of trembler (Tr), shaker-2 (sh-2), nor vibrator (vb), all neurological mutations located in the same region of chromosome 11. No differences were observed in the levels of the major CNS and PNS myelin proteins or lipids of spd/spd mice versus littermate controls, suggesting that, unlike several closely linked mutations, the spd mutation does not affect myelination. Pharmacological studies reported here show that aminooxyacetic acid improves the behavioral abnormalities of affected spd/spd mice in the same way it improves the behavior of affected spa/spa mice. However, unlike the spa/spa mice, there are no changes in the postsynaptic receptors for glycine, GABA, or benzodiazepines in spd/spd mice.

Shiverer and normal peripheral myelin compared: basic protein localization, membrane interactions, and lipid composition.

We have correlated membrane structure and interactions in shiverer sciatic nerve myelin with its biochemical composition. Analysis of x-ray diffraction data from shiverer myelin swollen in water substantiates our previous localization of an electron density deficit in the cytoplasmic half of the membrane. The density loss correlates with the absence of the major myelin basic proteins and indicates that in normal myelin, the basic protein is localized to the cytoplasmic apposition. As in normal peripheral myelin, hypotonic swelling in the shiverer membrane arrays occurs in the extracellular space between membranes; the cytoplasmic surfaces remain closely apposed notwithstanding the absence of basic protein from this region. Surprisingly, we found that the interaction at the extracellular apposition of shiverer membranes is altered. The extracellular space swells to a greater extent than normal when nerves are incubated in distilled water, treated at a reduced ionic strength of 0.06 in the range of pH 4-9, or treated at constant pH (4 or 7) in the range of ionic strengths 0.02-0.20. To examine the biochemical basis of this difference in swelling, we compared the lipid composition of shiverer and normal myelin. We find that sulfatides, hydroxycerebroside, and phosphatidylcholine are 20-30% higher than normal; nonhydroxycerebroside and sphingomyelin are 15-20% lower than normal; and ethanolamine phosphatides, phosphatidylserine, and cholesterol show little or no change. A higher concentration of negatively charged sulfatides at the extracellular surface likely contributes to an increased electrostatic repulsion and greater swelling in shiverer. The cytoplasmic surfaces of the apposed membranes of normal and shiverer myelins did not swell apart appreciably in the pH and ionic strength ranges expected to produce electrostatic repulsion. This stability, then, clearly does not depend on basic protein. We propose that P0 glycoprotein molecules form the stable link between apposed cytoplasmic membrane surfaces in peripheral myelin.

The interaction of mercurials with myelin: comparison of in vitro and in vivo effects.

Our previous study on the in vitro interactions of mercurials with peripheral nerve had shown that HgCl2 labels phosphatidylethanolamine plasmalogen in the myelin membrane, and that both HgCl2 and CH3HgCl alter the packing of the membrane array (Kirschner and Ganser, 1982). Thin-layer chromatography shows that in vitro treatment of sciatic and optic nerve with HgCl2 causes the hydrolysis of phosphatidylethanolamine plasmalogen while treatment with CH3HgCl does not. The present study addresses the possibility that the interaction of mercurials with myelin phosphatidylethanolamine plasmalogen may underlie their neurotoxicity. HgCl2 was administered to different groups of mice by intravenous, intraperitoneal and subcutaneous injections, and perorally through their drinking water. CH3HgCl was given perorally. Elemental mercury (Hg degree) vapor was administered by inhalation. The mice were monitored for signs of neurotoxicity. Myelin labeling and structure in sciatic and optic nerves was examined using X-ray diffraction and histochemical electron microscopy. The levels of mercury in tissues were measured using atomic absorption spectrophotometry. Mice exposed to CH3HgCl or to Hg degree vapor developed neurological symptoms, while mice exposed to HgCl2 did not show dysfunction even after doses as high as 10-20 mg/kg/day for 14 months. Neither labeling of the myelin membrane nor changes in membrane packing were detected in nerves from mice treated with either mercurial or with Hg degree. These nerves did not show any histochemical evidence for mercury deposition in the myelin, whereas in vitro treated nerves did. The level of mercury in sciatic and optic nerves from mice intoxicated with CH3HgCl was measurable, but at least 30-40 times less than that after in vitro treatment. With HgCl2 intoxication, no measurable amount of mercury was detected in these nerves. Exposure to Hg degree vapor resulted in low but detectable levels of mercury in the nerves. We conclude from these results that the neurotoxicity of mercurials does not involve their interaction with lamellar myelin.

Differential expression of gangliosides on the surfaces of myelinated nerve fibers.

The binding of cholera and tetanus toxins to receptors on the surfaces of teased nerve fibers was used to localize GM1 and G1b-series gangliosides, respectively, by immunocytochemical methods. Native fibers and fibers treated with various hydrolytic enzymes to degrade specific surface components were studied. With native fibers, both toxins bound abundantly to nodes of Ranvier and poorly to the most external, internodal Schwann cell surfaces. Treatment of the fibers with proteases, hyaluronidase, and chondroitin ABC lyase neither eliminated receptors at the nodes nor unmasked receptors over the internodes. The axolemma underlying the paranodal or internodal myelin, exposed by extensive treatment with protease, bound both toxins in large amounts. Neuraminidase action induced cholera toxin receptors on the Schwann cell surface; these receptors were insensitive to protease. The results indicate that GM1 and G1b-series gangliosides are predominantly localized to axonal and glial structures of the node of Ranvier and to paranodal/internodal Axolemma, and that polysialogangliosides not of the G1b-series are present on the internodal Schwann cell surface.

Ganglioside localization on myelinated nerve fibres by cholera toxin binding.

GM1 ganglioside has been localized on the surfaces of myelinated, peripheral nerve fibres by using immunofluorescence to detect cholera toxin receptors. Unfixed, mouse sciatic nerves were teased into individual, intact fibres in order to expose their extracellular surfaces. Cholera toxin binding sites were abundant at all nodes of Ranvier; they were scarce on the internodal fibre surfaces. The nodal receptors were resistant to various degradative enzymes, including trypsin and proteinase K. Proteases did not unmask receptors on the internodal surfaces. Exogenous GM1 successfully competed for the toxin binding sites on the fibres. From this evidence and the specificity of cholera toxin binding, we conclude that GM1 ganglioside is abundantly present on the membrane surfaces of peripheral nodes of Ranvier and is not present on the internodal Schwann cell surfaces in an appreciable amount. The patterns of fluorescence within the node suggest that the axon and Schwann cell structures are sites where GM1 is localized. Treatment of the teased fibres with Vibrio cholerae neuraminidase, which is known to reduce polysialogangliosides to the monosialoganglioside GM1, induced cholera toxin binding on the internodal Schwann cell surfaces. The induced receptors, as well as their precursors, were resistant to trypsin and proteinase K. We conclude that the internodal Schwann cell surface is rich in an unidentified polysialoganglioside(s) that can be converted to GM1 by neuraminidase.