Glycoproteins associated with myelin in the central nervous system

Purified myelin fractions from the central nervous system contain one major myelin-associated glycoprotein and approximately 16 minor glycoproteins. While the genuine association of the major myelin-associated glycoprotein with the oligodendroglial myelin unit is demonstrated, the possibility exists that several of the minor glycoproteins have their origin in contaminating membranes not related to myelin. The major myelin-associated glycoprotein is probably not present in compacted myelin, but immunocytochemical and subfractionation studies indicate that it is confined to the periaxonal and paranodal region of the myelin sheath. In experimental demyelination and multiple sclerosis, the major glycoprotein is the first myelin constituent to be affected. Its localization on the membrane surface where myelin and axolemma are in close contact, and other indirect evidence indicate that the major glycoprotein, and possibly other myelin-associated glycoproteins, could play a role in the process of myelination and myelin maintenance.     

The PO glycoprotein of peripheral nerve myelin

The PO glycoprotein, the major protein of peripheral nerve myelin, is a hydrophobic glycoprotein which can be isolated in soluble and insoluble forms from rabbit sciatic nerve myelin following extensive defatting and mid acidic extraction. The PO glycoprotein was localized exclusively in peripheral nervous system (PNS) myelin of sciatic nerve and rootlets by the immunofluorescent technique using goat anti-PO serum which showed a single precipitin band in double diffusion and did not cross-react with the myelin basic protein or P2 protein. Central nervous system (CNS) myelin from brain and spinal cord was negative by the immunofluorescent procedure. The major glycoprotein bands in PNS myelin, in addition to the PO glycoprotein at 28K, exist at 23K and 19K, as shown by gel electrophoresis in dodecyl sulfate. These glycoproteins, isolated by gel filtration in 2% dodecyl sulfate, show identity to the PO glycoprotein in their monosaccharide profile and overlapping tryptic peptides on peptide mapping. We conclude that both the 23K and 19K glycoproteins are derived from the PO glycoprotein by in situ proteolysis; the 23K glycoprotein has the identical amino terminal sequence. The 19K glycoprotein, beginning with amino-terminal methionine, is identical with the TPO glycoprotein, shown previously to originate from tryptic hydrolysis of the PO glycoprotein in isolated myelin. A tryptic glycopeptide containing 27 amino acids was isolated from the PO glycoprotein and sequenced. It contained a relatively high proportion of aspartic acid (four residues) and glutamic acid (two residues), thus exhibiting a high negative charge. We conclude that the total carbohydrate of the PO, 23K, and 19K glycoproteins does indeed exist as a single nonasaccharide moiety linked through N-acetylglucosamine to Asp-14 of the glycopeptide in a N-glycosidic linkage. These results further support the role of the PO glycoprotein as a typical amphipathic membrane protein.     

Glycopeptide fractions prepared from purified central and peripheral rat myelin

Myelin was purified from rat brain and sciatic nerve after invivo labeling with [³H]fucose and [¹⁴C]glucosamine to provide a radioactive marker for glycoproteins. The glycoproteins in the isolated myelin were digested exhaustively with pronase, and glycopeptides were isolated from the digest by gel filtration on Bio-Gel P-10. The glycopeptides from brain myelin separated into large and small molecular weight fractions, whereas the glycopeptides of sciatic nerve myelin eluted as a single symmetrical peak. The large and small glycopeptide fractions from central myelin and the single glycopeptide fraction from peripheral myelin were analyzed for carbohydrate by colorimetric and gas liquid chromatographic techniques. The glycopeptides from brain myelin contained 2.4 μg of neutral sugar and 0.59 μg of sialic acid per mg total myelin protein, whereas sciatic nerve myelin glycopeptides contained 10 μg of neutral sugar and 3.8 μg of sialic acid per mg total protein. Similarly, the gas-liquid chromatographic analyses showed that the glycopeptides from peripheral myelin contained 4- to 7-fold more of each individual per mg total myelin protein than those from central myelin. Most of the sialic acid and galactose in the glycopeptides from central myelin were in the large molecular weight fraction, and the small molecular weight glycopeptides contained primarily mannose and $\text{N-}\text{acetylglucosamine}$. The considerably higher content of glycoprotein-carbohydrate in peripheral myelin supports the results of gel electrophoretic studies, which indicate that the major protein in peripheral myelin in glycosylated while the glycoproteins in purified central myelin are quantitatevely minor components.     

External labeling of galactose in surface membrane glycoproteins of the intact myelin sheath.

The molecular organization of surface galactose residues in glycoproteins of the intact myelin sheath was investigated using the enzymatic membrane probe, galactose oxidase. Rat spinal cords treated under physiological conditions with this nonpermanent probe were labeled specifically in galactose residues by reduction with tritiated sodium borohydride. The enzymatically modified proteins from isolated myelin were analyzed electrophoretically and their specific radioactivities determined. Results indicated tritium label associated with a surprising variety of high molecular weight proteins. The most extensively labeled peak corresponded to the major myelin glycoprotein as indicated by the coincidence of tritium label with that of [14C]fucose used as an internal marker for the glycoproteins. The radioactivity associated with this protein was 1.1 to 2.7 times higher after treatment with galactose oxidase when compared to reduction in the absence of the enzyme and 1.4 to 4.8 times higher when oxidized and reduced after prior treatment with neuraminidase. The results suggest a complex heterogeneity of minor glycoproteins associated with isolated myelin. It is concluded that from this complexity of glycoproteins, a major glycoprotein is at least partially localized on the external surface of either the intact myelin sheath or the closely associated oligodendroglial plasma membrane. Such a localization of this glycoprotein and the probable localization of the other glycoproteins enhances their potential role in specific interactions in the process of mpyelination or myelin maintenance.     

FRACTIONATION OF GLYCOPROTEINS ASSOCIATED TO ADULT RAT BRAIN MYELIN FRACTIONS

Abstract— The chloroform-methanol insoluble residue of adult rat brain myelin fractions (My-CMI) contains only 20% of protein but all myelin-associated glycoproteins (Z anetta et al ., 1977a). After solubilization in sodium dodecyl sulphate, these glycoproteins were separated by sequential affinity chromatography on 4 immobilized lectins. Ten fractions (9 of which contained only glycoproteins) were obtained. Glycoproteins added up to at least 25% of My-CMI proteins. Many minor glycoproteins were detected in the different fractions. However most of them appeared not to be intrinsic to myelin. On the contrary a major glycoprotein electrophoretic band, component A, appeared to be intrinsic to myelin although its presence also on oligodendrocyte plasma membrane cannot be excluded. Component A was tentatively identified with the'major myelin associated glycoprotein'described by QUARLES (1972, 1973 a, b ). It accounted for less than 0.4% of proteins and 8% of glycoproteins of myelin fractions and consisted of at least two'glycopolypeptides'which, both, bind to concanavalin A and to the Ulex europeus lectin. The other major glycoprotein, component B, did not bind to any of the lectins used and, thus, must have N -acetyl neuraminic acid as only terminal sugar. The separation of myelin-associated glycoproteins according to their affinity for lectins allowed a tentative identification of the lectin binding sites of myelin sheath.     

Lectin-binding proteins in central-nervous-system myelin. Binding of glycoproteins in purified myelin to immobilized lectins

The capacities of immature and mature rat brain myelin, bovine myelin and human myelin to be agglutinated by soya-bean agglutinin, Ricinus communis agglutinin, wheatgerm agglutinin, and Lotus tetragonolobus agglutinin were examined. The first two lectins, which are specific for galactose and N-acetylgalactosamine, strongly agglutinated immature and mature rat myelin, weakly agglutinated bovine myelin, but did not affect human myelin. The other myelin and lectin combinations resulted in very weak or no agglutination. [(3)H]Fucose-labelled glycoproteins of purified adult rat brain myelin were solubilized with sodium dodecyl sulphate and allowed to bind to concanavalin A-Sepharose and each of the other lectins mentioned above, which had been immobilized on agarose. About 60% of the radioactive fucose was in glycoproteins that bound to concanavalin A-Sepharose and these glycoproteins could be eluted with solutions containing methyl alpha-d-mannoside and sodium dodecyl sulphate. Periodate/Schiff staining or radioactive counting of analytical gels showed that most of the major myelin-associated glycoprotein (apparent mol.wt. approx. 100000) bound to the concanavalin A, whereas the glycoproteins that did not bind were mostly of lower molecular weight. Preparative polyacrylamide-gel electrophoresis of the glycoprotein fraction that was eluted with methyl alpha-d-mannoside yielded a relatively pure preparation of the myelin-associated glycoprotein. Similar results were obtained with each of the other lectins, i.e. the myelin-associated glycoprotein was in the fraction that bound to the immobilized lectin. Double-labelling experiments utilizing [(3)H]fucose-labelled glycoproteins from adult myelin and [(14)C]fucose-labelled glycoproteins from 14-day-old rat brain myelin did not reveal any difference in the binding of the mature and immature glycoproteins to any of the immobilized lectins. The results in this and the preceding paper [McIntyre, Quarles & Brady (1979) Biochem. J.183, 205-212] suggest that the myelin-associated glycoprotein is one of the principal receptors for concanavalin A and other lectins in myelin, and that this property can be utilized for the purification of this glycoprotein.     

Purification and partial characterization of two glycoproteins in bovine peripheral nerve myelin membrane

Two major glycoproteins of bovine peripheral nerve myelin were isolated from the acid-insoluble residue of the myelin by a procedure involving delipidation with chloroform/methanol (2:1, v/v) and chromatography on Sephadex G-200 column with a buffer containing sodium dodecyl sulfate. The separation patterns of the proteins on the gel were affected considerably by the dodecyl sulfate concentration in the elution buffer. At above 2% dodecyl sulfate concentration in the elution buffer, the glycoproteins could be separated clearly on the gel and were purified. The purified proteins, the BR protein (mol. wt. 28 000) and the PAS-II protein (mol. wt. 13 000), were homogeneous on dodecyl sulfate-polyacrylamide gel electrophoresis. The NH₂-terminal amino acids of the BR and the PAS-II proteins were isoleucine and methionine, respectively. The BR protein contained glucosamine, mannose, galactose, fucose and sialic acids and the PAS-II protein contained glucosamine, mannose, galactose, fucose and glucose. Neither the BR protein nor the PAS-II were a glycosylated derivative of a basic protein of bovine peripheral nerve myelin, a deduction based on the results of amino acid analysis. The two major glycoproteins were observed commonly in the peripheral nerve myelin of cows, pigs, rabbits and guinea pigs, using dodecyl sulfate-polyacrylamide gel electrophoresis.     

Morphological and biochemical characterization of light and heavy myelin isolated from developing rat brain.

Myelin from developing rat brains was separated on a discontinuous sucrose gradient into subfractions of two different densities, i.e. light and heavy myelin. Electron photomicrographs showed that heavy myelin consisted primarily of large compacted multilamellar structures with a distinct intraperiod line characteristic of myelin in situ. Light myelin, on the other hand, was composed of small vesicles having a unilamellar structure. Similar to whole myelin, both membrane subfractions were highly enriched in 2',3'-cyclic nucleotide-3'-phosphohydrolase. The specific activity of the enzyme, however, showed no developmental trend. Both subfractions contained all the four major proteins characteristic of the whole myelin membrane. There were, however, quantitative differences in the relative distribution of these proteins between light and heavy myelin. Basic protein accounted for 55% and proteolipid protein for 46% of the total myelin proteins of light and heavy myelin, respectively. DM-20 (Agrawal, H.C., Burton, R. M., Fishman, M.A., Mitchell, R.F. and Prensky, A.L. (1972) J. Neurochem. 19, 2083-2089) exhibited a developmental "switch" between light and heavy myelin. Light myelin appeared to contain more DM-20 in 15- to 20-day-old rat brain, whereas the concentration of this protein was higher in heavy myelin at subsequent ages studied.     

Lectin-binding proteins in central-nervous-system myelin. Detection of glycoproteins of purified myelin on polyacrylamide gels by [3h]concanavalin A binding.

Concanavalin A strongly agglutinates purified fragments of immature and mature rat brain myelin, but only weakly agglutinates mature bovine and human myelin fragments. A sensitive method involving [3H]concanavalin binding to sodium dodecyl sulphate/polyacrylamide gels was used to detect the concanavalin A-binding proteins in purified myelin. When applied to mature rat brain myelin proteins that had been labelled in vivo with [14C]fucose, the distribution of the [3H]concanavalin A on the gel was very similar to that of [14C]fucose with the major peak corresponding to the major myelin-associated glycoprotein. The technique revealed that the immature form of the myelin-associated glycoprotein with a slightly larger apparent molecular weight also bound concanavalin A, and that in purified immature rat myelin the quantitative importance of some of the other glycoproteins in binding concanavalin A was increased relative to the myelin-associated glycoprotein. The separated proteins of bovine and human myelin bound more [3H]-concanavalin A than those of rat myelin. In these species, the myelin-associated glycoprotein was a major concanavalin A-binding protein, although two higher-molecular-weight glycoproteins also bound significant quantities of [3H]concanavalin A. The results indicate that there are receptors for concanavalin A on the surface of rat, bovine and human myelin membranes and suggest that the myelin-associated glycoprotein is one of the principal receptors.     

Morphological and biochemical characterization of light and heavy myelin isolated from developing rat brain

Myelin from developing rat brains was separated on a discontinuous sucrose gradient into subfractions of two different densities, i.e. light and heavy myelin. Electron photomicrographs showed that heavy myelin consisted primarily of large compacted multilamellar structures with a distinct intraperiod line characteristic of myelin in situ. Light myelin, on the other hand, was composed of small vesicles having a unilamellar structure. Similar to whole myelin, both membrane subfractions were highly enriched in 2′,3′-cyclic nucleotide-3′-phosphohydrolase. The specific activity of the enzyme, however, showed no developmental trend. Both subfractions contained all of the four major proteins characteristic of the whole myelin membrane. There were, however, quantitative differences in the relative distribution of these proteins between light and heavy myelin. Basic protein accounted for 55 % and proteolipid protein for 46 % of the total myelin proteins of light and heavy myelin, respectively. DM-20 (Agrawal, H. C., Burton, R. M., Fishman, M. A., Mitchell, R. F. and Prensky, A. L. (1972) J. Neurochem. 19, 2083–2089) exhibited a developmental “switch” between light and heavy myelin. Light myelin appeared to contain more DM-20 in 15- to 20-day-old rat brain, whereas the concentration of this protein was higher in heavy myelin at subsequent ages studied.     

A biochemical comparison of Xenopus laevis and mammalian myelin from the central and peripheral nervous systems.

Myelin purified from the central nervous system of Xenopus laevis contained the same major lipid and protein components as human myelin. However, some minor differences in the myelin proteins were noted. The Xenopus basic protein had a higher apparent mol wt. on sodium dodecyl sulfate gels than the corresponding mammalian protein. The absolute specific activity of 2',3'-cyclic nucleotide 3'-phosphohydrolase in the Xenopus myelin was considerably higher than in mammals. There were differences in the high mol wt. proteins, and the glycoproteins in Xenopus myelin were more heterogeneous than those in mammals. Peripheral myelin from Xenopus sciatic nerve was compared with that from the rat. The lipids in the two types of myelin were similar. There was a major glycoprotein in the Xenopus myelin corresponding to the P0 protein and a basic protein of slightly larger mol wt. than the P1 protein of rat myelin.     

Lens glycoproteins: biosynthesis in cultured epithelial cells of bovine lens.

1. Radioactivity from [3H]glucosamine is rapidly incorporated into cellular fractions of lens epithelial cells cultured in vitro. The incorporated isotope appears largely in glycoproteins of the cell surface that are exposed to trypsin and are released into a soluble form by proteolysis of intact cells. Glycoproteins are also secreted by cultured cells and can be recovered in the culture fluids. Sodium dodecysulphate-polyacrylamide gell electrophoresis shows that a range of glycoproteins with apparent molecular weights from approximately 14000 to 120000 are present. The relationships of these glycoproteins to collagen and the non-collagenous glycoproteins of lens basement membranes are discussed. 2. A plasma membrane fraction obtained from non-trypsinised lens epithelial cells contains one major glycoprotein of apparent molecular weight 120000. A major non-glycosylated polypeptide of molecular weight about 38000 detectable by Bloemendal et al. (1972) in plasma membranes of differentiated lens fibre cells was not prominent in lens epithelial cell membranes. 3. Examination of lens basement membranes extracted in various ways failed to reveal major glycoproteins of low molecular weight. Higher molecular weight glycoproteins, some of them related to collagen, were present.     

EVIDENCE FOR THE CLOSE ASSOCIATION OF A GLYCOPROTEIN WITH MYELIN IN RAT BRAIN

Abstract— Myelin was purified from rats which had been injected intracerebrally with radioactive fucose in order to label specifically the glycoproteins. Myelin contained a small amount of fucose-labelled glycoproteins in comparison to that in other subcellular fractions, but polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate revealed a unique pattern of radioactive glycoproteins dominated by a major peak. The same glycoprotein was not prominent in the other subcellular fractions which were examined. This major glycoprotein in the myelin fraction was also labelled after injection with [³H]glucosamine or N -[³H]acetylmannosamine. It was the most intensely staining myelin protein when gels were treated with periodic acid-Schiff reagents, an indication that, in terms of protein-bound carbohydrate, it is the major glycoprotein in the myelin fraction. The glycoprotein was present in myelin purified from rats ranging in age from 14 days to 14 months. Extensive recycling of the myelin through the purification procedures did not significantly reduce the amount of glycoprotein in the myelin. Double label experiments with [³H]fucose and [¹⁴C]fucose were used to compare glycoproteins in myelin purified from white and grey matter, respectively, and from mixed homogenates of myelinated and unmyelinated brain. The results obtained from these experiments suggested that the glycoprotein is closely associated with myelin and that it is not in an unrelated contaminating structure. Possible locations of the glycoprotein are discussed. They include the myelin membrane itself, the oligodendroglial plasma membrane, and the axolemma of myelinated axons.     

Developmental regulation of the carbohydrate composition of glycoproteins associated with central nervous system myelin

Abstract: Glycoproteins from central nervous system myelin were evaluated for developmental alterations in their carbohydrate composition by autoradiographic analysis of radioiodinated lectin binding after separation by high-resolution sodium dodecyl sulfate-pore gradient slab gel electrophoresis (SDS-PGE). Sixteen lectin-binding components were assessed in highly purified myelin preparations from 15-day, 18-day, and adult rat brains, using the lectins Triticum vulgaris (wheat germ agglutinin) and Ulex europeus (gorse agglutinin I). Developmental changes in lectin binding for individual glycoproteins were evaluated semiquantitatively by comparing densitometric scans of the auto radiographs. Both increases and decreases in lectin binding for individual components were observed as a consequence of development, as well as the appearance and disappearance of lectin binding to three low-molecular-weight components. No changes in electrophoretic mobility and hence glycoprotein molecular weight were observed in any components when using these lectins. These developmental changes in lectin binding suggest that increases in glycoprotein (receptor) density occur, as well as an elaboration of oligosaccharide branching for individual glycoproteins. In addition, the appearance of a new glycoprotein in the adult myelin membrane could imply a new functional role not present in the immature membrane. These observations suggest that dynamic alterations of myelin-associated glycoproteins occur during development. Such developmental regulation of membrane glycoproteins increases the significance of their potential role in myelination and myelin maintenance.     

Glycosylation of myelin glycoproteins in peripheral nerve via lipid intermediates

Membrane preparations from chick peripheral nervous system (PNS) catalyzed the transfer of [³H]glucose from UDP-[³H]glucose into glucosylphosphoryl dolichol. The initial rate of glucosylphosphoryl dolichol formation in a non-myelin membrane fraction from actively myelinating chick PNS was 11 fold higher than that from adult. Exogenous dolichyl monophosphate stimulated glucosylphosphoryl dolichol synthesis in both fractions. The higher level of glucosylphosphoryl dolichol synthesis corresponded to the onset of myelination in chick PNS. Exogenous dolichyl monophosphate also stimulated the labeling of glucosylated oligosaccharide lipids and glycoproteins in the fraction. On SDS polyacrylamide gel electrophoresis, the relative mobility of the major and minor radioactive glycoprotein corresponded with that of the P0 and PASII glycoprotein in PNS myelin, respectively. The results suggest that myelin glycoproteins in PNS are glycosylated via lipid intermediates.     

Hypotheses regarding myelination derived from comparisons of myelin subfractions.

Patterns of myelin-associated proteins and glycoproteins in subfractions of central nervous system myelin and changes in these during development have been reviewed. Several hypotheses are put forward regarding classification of these proteins as structural components of myelin, as components of the membranes from which myelin is derived or as molecules present to play a role in myelin formation rather than as true components of compact myelin. 2′, 3′-Cyclic nucleotide 3′-phosphohydrolase is considered to fit into this last category and two different hypothesis are proposed for its role. The major myelin glycoprotein is postulated to serve a function in recognition of axons by oligodendrocytes, a function not needed in the peripheral nervous system where Schwann cells and axons develop in close contact.     

A biochemical comparison of Xenopus laevis and mammalian myelin from the central and peripheral nervous systems

Myelin purified from the central nervous system of Xenopus laevis contained the same major lipid and protein components as human myelin. However, some minor differences in the myelin proteins were noted. The Xenopus basic protein had a higher apparent mol wt. on sodium dodecyl sulfate gels than the corresponding mammalian protein. The absolute specific activity of 2′,3′-cyclic nucleotide 3′-phosphohydrolase in the Xenopus myelin was considerably higher than in mammals. There were differences in the high mol wt. proteins, and the glycoproteins in Xenopus myelin were more heterogeneous than those in mammals. Peripheral myelin from Xenopus sciatic nerve was compared with that from the rat. The lipids in the two types of myelin were similar. There was a major glycoprotein in the Xenopus myelin corresponding to the P₀ protein and a basic protein of slightly larger mol wt. than the P₁ protein of rat myelin.     

[3 5S]Sulfate incorporation into myelin glycoproteins I. Central nervous system

The in vivo incorporation of [^{3 5}S]sulfate and [³H]fucose into rat brain myelin was investigated. Most of the ^{3 5}S in the myelin was in sulfatide, but about 4% was associated with the residual proteins after chloroform/methanol extraction. Polyacrylamide gel electrophoresis of these proteins indicated that the major ^{3 5}S-labeled component corresponded to the major fucose-labeled glycoprotein. The labeling of this predominant glycoprotein with sulfate was more selective than with fucose, since there was relatively little incorporation of sulfate into some of the minor fucose-labeled glycoproteins. There was little or no ^{3 5}S associated with proteolipid or basic protein on polyacrylamide gels. The fucose-labeled glycoproteins were converted to glycopeptides by pronase digestion and separated into two major classes by gel filtration on Sephadex-G50. Only the higher molecular weight class contained significant amounts of ^{3 5}S. The association of ^{3 5}S with the glycopeptides was not due to binding of sulfatide or free inorganic sulfate. The results indicate that the predominant myelin-associated glycoprotein in rat brain is sulfated.     

Towards crystallization of hydrophobic myelin glycoproteins: P0 and PASII/PMP22

The preparation of a pure and homogeneous protein sample at proper concentration is a prerequisite for success when attempting their crystallization for structural determination. The detergents suitable for solubilization particularly of membrane proteins are not always the best for crystallization. Myelin of the peripheral nervous system of vertebrates is the example of a membrane for which neutral or "gentle" detergents are not even strong enough to solubilize its proteins. In contrast, sodium- or lithium-dodecyl sulfate is very effective. We solubilized myelin membrane in 2%(w/v) sodium dodecyl sulfate, followed by chromatographic purification of the hydrophobic myelin glycoproteins P0 and PASII/PMP22, and finally, we have exchanged the sodium dodecyl sulfate bound to protein for other neutral detergents using ceramic hydroxyapatite column. Theoretically, we should easily exchange sodium dodecyl sulfate for any neutral detergent, but for some of them, the solubility of myelin glycoproteins is low. To monitor the potential variability in the secondary structure of glycoproteins, we have used circular dichroism. Sodium dodecyl sulfate seems to be the appropriate detergent for the purpose of purification of very hydrophobic glycoproteins, since it can be easily exchanged for another neutral detergent.     

Association and release of the major intrinsic membrane glycoprotein from peripheral nerve myelin.

Hypo-osmotic homogenization of the endoneurium from the adult-rat sciatic nerve and subsequent evaluation of the 197 000 g aqueous supernatant by sodium dodecyl sulphate pore-gradient electrophoresis (SDS-p.g.e.) revealed a release of the major glycoprotein (P0) (29 000 Mr) from peripheral nerve myelin. Immunological verification of the presence of this asparagine-linked glycoprotein in the aqueous supernatant was obtained by immune overlay after SDS-p.g.e. and electrophoretic transfer to nitrocellulose using anti-P0 gamma-globulin followed by autoradiographic detection with 125I-protein A. A comparison of successive hypo- and iso-osmotic extractions of the endoneurium revealed that the hypo-osmotic extraction released increasing amounts of P0 into the supernatant fraction, whereas the iso-osmotic treatment revealed lower levels of P0 extracted from the myelin and lesser amounts with each successive extraction. Three successive hypo-osmotic extractions resulted in a 2.0-, 2.9-, and 9.5-fold increase in the amount of P0 released compared with the successive iso-osmotic extractions. Although these results suggest that this major myelin glycoprotein has properties similar to those of extrinsic membrane proteins, temperature-dependent phase-partitioning experiments with Triton X-114 revealed that this glycoprotein is recovered in the detergent-enriched lower phase. These results indicate that this major myelin glycoprotein is an amphipathic integral membrane protein with a distinct hydrophobic domain and yet has solubility characteristics typical of an extrinsic membrane protein. P0 labelled in vitro with [3H]mannose could be immunoprecipitated from the aqueous supernatant with anti-P0 gamma-globulin by centrifugation at 197000g without the addition of second antibody or protein A. Analysis of such an immune precipitate after incorporation in vitro with [14C]acetate to label endoneurial lipids revealed that all major endoneurial lipid classes contained radioactive label, as determined by fluorography after high-performance t.l.c. The mechanism of release of this intrinsic glycoprotein from the myelin membrane, therefore, involves the osmotic-dependent formation of mixed micelles or membrane vesicles with endogenous membrane lipids.