Molecular Interactions of the Major Myelin Glycosphingolipids and Myelin Basic Protein in Model Membranes

The molecular organization, interactions, phase state and membrane-membrane interactions of model membranes containing cerebroside (GalCer), sulfatide (Sulf) and myelin basic protein (MBP) were investigated. Sulf shows a larger cross-sectional area than GalCer, in keeping with the lateral electrostatic repulsions in the negatively charged polar head group. The interactions of GalCer with different phospholipids are similar while those with Sulf depend on the phosphoryl choline moiety in the phospholipid. MBP induces a decrease of the phase transition temperature in both lipids but with Sulf this occurs at lower proportions of MBP. In mixtures of Sulf with phosphatidylcholine MBP induces phase separation among Sulf-rich and PC-rich domains. Extensive apposition of bilayers containing Sulf is induced by MBP while GalCer interferes with this process. Few membrane interactions proceed to bilayer merging or whole bilayer fusion and the glycosphingolipids help preserve the membrane integrity.     

Study of myelin purity in relation to axonal contaminants

Axonal remnants are considered a probable source of contamination of isolated myelin in view of the relatively tight axon-glial intercellular junction. Using the rabbit optic system to label specifically axonal components, we have found the levels of such contaminants to depend on the myelin isolation procedure, the tissue source, and the nature of the contaminant. A procedure employing repetitive treatments with EGTA was found to be highly effective in removing proline-labeled axonal proteins, the estimated upper limit of such contamination being approximately 0.6–1.2% of the myelin protein. The standard isolation procedure of Norton and Poduslo, supplemented with an additional discontinuous gradient step, proved equally effective in removing rapidly transported proteins from myelin isolated from the superior colliculus or lateral geniculate body. When the optic tract was the source, however, the EGTA procedure proved more effective in removing both rapidly and slowly transported proteins. Axonal gangliosides labeled with N-[ ³ H]acetylmannosamine were efficiently removed by both procedures, adding support to the proposition that gangliosides detected in isolated myelin are intrinsic to that membrane.     

An electronmicroscopic study of the distribution of myelin basic protein in multilayer dispersions of diacylphosphatidylserine

Although dispersions containing lipid and protein are widely used as model systems to explore the properties of biomembranes, the extent of mixing of the two components has generally not been determined. Here, the distribution of bovine myelin basic protein in dispersions with bovine brain L-alpha-diacylphosphatidylserine (PS) has been examined electronmicroscopically. Dispersions of PS were prepared by hydrating a known amount of dried lipid with buffer or with buffer containing an equal weight of myelin basic protein or lysozyme. The lipid-protein complexes were separated from unbound protein by centrifugation in 0-60% sucrose density gradients. In both systems only a few percent of the protein was unbound and the resultant recombinants, which gave single bands on the gradients, contained about 50% protein by weight. After removal of the sucrose by dialysis the dispersions were fixed in 2.5% glutaraldehyde and 1% osmium tetroxide, dehydrated and embedded in epoxy resin. Thin sections cut from these blocks were incubated, after removal of osmium tetroxide, with antiserum raised in rabbits against human myelin basic protein. Excess antiserum was removed and the antigen-antibody complexes on the thin sections were labelled with 13 nm diameter colloidal gold particles stabilized with protein A. The distributions of these gold particles were examined under an electronmicroscope. Comparison of the labelling patterns for PS, PS-lysozyme and PS-basic protein demonstrated specific labelling in the last, and showed the gold particles to be uniformly dispersed. It was concluded that in these dispersions the protein and lipid were intimately mixed at the molecular level.     

Epifluorescence Microscopy of Surface Domain Microheterogeneity in Myelin Monolayers at the Air-Water Interface

Myelin lipids form liquid-expanded monolayers at the air-water interface, with no evidence of surface pressure-induced two-dimensional phase transition. However, the film doped with 2 mole % of the fluorescent probe N-(7-nitro-2-1,3-benzoxadiazol-4-yl) Diacyl Phosphatidyl-ethanolamine (NBD-PE) shows an irregular pattern of coexisting laterally segregated surface domains with diffuse boundaries that change from smooth patterns to fractal-like structures depending on surface pressure. Successive expansion-recompression cycles lead to more defined domains, with a general reorganization occurring at surface pressures of about 20 mN/m. At least two coexisting phases occur over almost all the compression isotherms. The presence of proteins in whole myelin monolayers induces defined domain textures with relatively sharp boundaries. The patterns during compression and expansion are quite similar and, after the first cycle, little changes occur under recompression. The patterns observed provide topographical evidence for the existence of dynamic domain microheterogeneity in the surface of myelin interfaces.     

Triton X-100 extractions of central nervous system myelin indicate a possible role for the minor myelin proteins in the stability of lamellae

Isolated CNS myelin membranes were extracted with Triton X-100 under conditions previously established for the isolation of cytoskeletal proteins. Treated myelin retained much of its characteristic lamellar structure despite the removal of most of the major myelin basic protein (18.5 kDa) and the proteolipid protein, which together normally constitute 60% of the total myelin protein. The SDS-PAGE profile of this extract residue demonstrated an enrichment in proteins of Mr 30 to 60 kilodaltons (the Wolfgram group). The major myelin proteins were identified by antibodies on Western immunoblots, as were the 23-cyclic nucleotide 3-phosphodiesterase (CNP), actin, tubulin, myelin-associated glycoprotein (MGP) and the 21.5 kDa MBP. The overall behavior of CNP, the 21.5 kDa MBP, MGP and tubulin towards Triton extraction is reminiscent of the behavior of other membrane-skeletal complexes, supporting the idea that these and other minor myelin proteins might be part of heteromolecular complexes with interactions spanning several lamellae of the myelin sheath.     

Degradation products of myelin-oligodendrocyte-associated proteins in a light CNS subcellular fraction

The presence of degradation products of the myelin/oligodendrocyte glycoprotein (MOG) and a new myelin/oligodendrocyte associated protein, FD1, defined by a monoclonal antibody was established in a subfraction (the floating fraction, or FF) of adult rabbit CNS. The histochemical distribution of FD1 was determined by indirect immunofluorescense using conventional and confocal microscopy. FD1 was found to be present in oligodendrocytes, and at the outer rim of CNS myelin sheaths. Strong antibody reactivity was noted at nodes of Ranvier, as well as in regions with a high nodal density. No staining of compact myelin was seen. In the PNS, inner and outer cytoplasmic compartments of the Schwann cells as well as their cell bodies were stained, with no staining of compact myelin. The FF has previously been shown to be highly enriched in Marchi-positive bodies. These structures are situated paranodally in the CNS of myelinated nerve fibers, and their presence has been interpreted as reflections of myelin breakdown and turnover occurring in association with myelin sheath segments situated close to nodes at Ranvier in adult, normal vertebrate CNS. The present findings extend previous observations of partially degraded myelin-associated proteins in the FF, and give further results indicating that Marchi-positive bodies are aspects of intermediate stages in myelin catabolism.     

Basic proteins of rodent peripheral nerve myelin: immunochemical identification of the 21.5K, 18.5K, 17K, 14K, and P2 proteins

Abstract: Proteins in peripheral nervous system and central nervous system myelin and homogenates of sciatic nerve and brain from young and adult mice and rats were characterized with affinity-purified anti-P₂ and anti-myelin basic protein sera after electrophoretic transfer from sodium dodecyl sulfate-polyacrylamide gels to nitrocellulose sheets. Using this method we have identified a component of rodent peripheral nervous system myelin as P₂ protein. Peripheral nervous system myelin also showed the presence of four basic proteins in addition to P₂ protein. These were found to be analogous to the 14, 17, 18.5, and 21.5K species found in the central nervous system myelin. A number of high-molecular-weight proteins were also detected with anti-myelin basic protein serum in peripheral nervous system, as well as central nervous system myelin. In addition, we report the presence of a high-molecular-weight P₂ cross-reactive protein in rodent brain stem homogenates, but not in central nervous system myelin.     

Intraneuronal N-acetylaspartate supplies acetyl groups for myelin lipid synthesis: evidence for myelin-associated aspartoacylase

Despite its growing use as a radiological indicator of neuronal viability, the biological function of N-acetylaspartate (NAA) has remained elusive. This is due in part to its unusual metabolic compartmentalization wherein the synthetic enzyme occurs in neuronal mitochondria whereas the principal metabolizing enzyme, N-acetyl-L-aspartate amidohydrolase (aspartoacylase), is located primarily in white matter elements. This study demonstrates that within white matter, aspartoacylase is an integral component of the myelin sheath where it is ideally situated to produce acetyl groups for synthesis of myelin lipids. That it functions in this manner is suggested by the fact that myelin lipids of the rat optic system are well labeled following intraocular injection of [14C-acetyl]NAA. This is attributed to uptake of radiolabeled NAA by retinal ganglion cells followed by axonal transport and transaxonal transfer of NAA into myelin, a membrane previously shown to contain many lipid synthesizing enzymes. This study identifies a group of myelin lipids that are so labeled by neuronal [14C]NAA, and demonstrates a different labeling pattern from that produced by neuronal [14C]acetate. High performance liquid chromatographic analysis of the deproteinated soluble materials from the optic system following intraocular injection of [14C]NAA revealed only the latter substance and no radiolabeled acetate, suggesting little or no hydrolysis of NAA within mature neurons of the optic system. These results suggest a rationale for the unusual compartmentalization of NAA metabolism and point to NAA as a neuronal constituent that is essential for the formation and/or maintenance of myelin. The relevance of these findings to Canavan disease is discussed.     

Cerebroside synthesis as a measure of the rate of remyelination following cuprizone-induced demyelination in brain

We studied markers of myelin content and of the rate of myelination in brains of mice between 8 and 20 weeks of age. During the 12-week time-course, control animals showed slight increases in the content of oligodendroglial-specific cerebroside, as well as cholesterol (enriched in, but not specific to, myelin). In contrast, synthesis of these lipids, as assayed by in vivo incorporation of (3)H(2)O, was substantial, indicating turnover of 0.4% and 0.7% of total brain cerebroside and cholesterol, respectively, each day. We also studied mice exposed to a diet containing 0.2% of the copper chelator, cuprizone. After 6 weeks 20%, and by 12 weeks, over 30% of brain cerebroside was gone. Demyelination was accompanied by down-regulation of mRNA expression for enzymes controlling myelin lipid synthesis (ceramide galactosyl transferase for cerebroside; hydroxymethylglutaryl-CoA reductase for cholesterol), and for myelin basic protein. Synthesis of myelin lipids was also greatly depressed. The 20% cerebroside deficit consequent to 6 weeks of cuprizone exposure was restored 6 weeks after return to a control diet. During remyelination, expression of myelin-related mRNA species, as well as cerebroside and cholesterol synthesis were restored to normal. However, in contrast to the steady state metabolic turnover in the control situation, all the cerebroside and cholesterol made were accumulated. To the extent that accumulating cerebroside is targeted for eventual inclusion in myelin (discussed) the rate of its synthesis is proportional to remyelination. With our assay, in vivo rates of cerebroside synthesis can be determined for a time window of the order of hours. This offers greater temporal resolution and accuracy relative to classical methods assaying accumulation of myelin components at time intervals of several days. We propose this experimental design, and the reproducible cuprizone model, as appropriate for studies of how to promote remyelination.     

Bimodal occurrence of aspartoacylase in myelin and cytosol of brain

The growing use of N-acetylaspartate as an indicator of neuronal viability has fostered interest in the biological function(s) of this unusual amino acid derivative. In considering the various physiological roles that have been proposed for this relatively abundant molecule one is obliged to take into account its unusual metabolic compartmentalization, according to which synthesis and storage occur in the neuron and hydrolytic cleavage in the oligodendrocyte. The latter reaction, catalyzed by aspartoacylase (ASPA), produces acetyl groups plus aspartate and has been proposed to occur in both soluble and membranous subfractions of white matter. Our study supports such bimodal occurrence and we now present immunoblot, proteomic, and biochemical evidence that the membrane-bound form of ASPA is intrinsic to purified myelin membranes. This was supported by a novel TLC-based method for the assay of ASPA. That observation, together with previous demonstrations of numerous lipid-synthesizing enzymes in myelin, suggests utilization of acetyl groups liberated by myelin-localized ASPA for lipid synthesis within the myelin sheath. Such synthesis might be selective and could explain the deficit of myelin lipids in animals lacking ASPA.     

Identification of the Protein Target of Myelin-Binding Ligands by Immunohistochemistry and Biochemical Analyses

The ability to visualize myelin is important in the diagnosis of demyelinating disorders and the detection of myelin-containing nerves during surgery. The development of myelin-selective imaging agents requires that a defined target for these agents be identified and that a robust assay against the target be developed to allow for assessment of structure-activity relationships. We describe an immunohistochemical analysis and a fluorescence polarization binding assay using purified myelin basic protein (MBP) that provides quantitative evidence that MBP is the molecular binding partner of previously described myelin-selective fluorescent dyes such as BMB, GE3082, and GE3111.     

Structural Changes of Surfactant Protein A Induced by Cations Reorient the Protein on Lipid Bilayers

Surfactant protein A (SP-A) is an octadecameric hydrophilic glycoprotein and is the major protein component of pulmonary surfactant. This protein complex plays several roles in the body, such as regulation of surfactant secretion, recycling and adsorption of surfactant lipids, and non-serum-induced immune response. Many of SP-A's activities are dependent upon the presence of cations, especially calcium. Here, we have studiedin vitrothe effect of cations on the interaction of purified bovine SP-A with phospholipid vesicles made of dipalmitoylphosphatidylcholine and unsaturated phosphatidylcholine. We have found that SP-A octadecamers exist in an “opened-bouquet” conformation in the absence of cations and interact with lipid membranes via one or two globular headgroups. Calcium-induced structural changes in SP-A lead to the formation of a clearly identifiable stem in a “closed-bouquet” conformation. This change, in turn, seemingly results in all of SP-A's globular headgroups interacting with the lipid membrane surface and with the stem pointing away from the membrane surface. These results represent direct evidence that the headgroups of SP-A (comprising carbohydrate recognition domains), and not the stem (comprising the amino-terminus and collagen-like region), interact with lipid bilayers. Our data support models of tubular myelin in which the headgroups, not the tails, interact with the lipid walls of the lattice.     

Parameters related to lipid metabolism as markers of myelination in mouse brain

Myelination, during both normal development and with respect to disorders of myelination, is commonly studied by morphological and/or biochemical techniques that assay as their end-points the extent of myelination. The rate of myelination is potentially a more useful parameter, but it is difficult and time-consuming to establish, requiring a complete developmental study with labor-intensive methodology. We report herein development of methodology to assay the absolute rate of myelination at any desired time during development. This involves intraperitoneal injection of (3)H(2)O to label body water pools, followed by determination of label in the myelin-specific lipid, cerebroside. The absolute amount of cerebroside synthesized can then be calculated from the specific radioactivity of body water and knowledge of the number of hydrogens from water incorporated into cerebroside. During development, the rate of cerebroside synthesis correlated well with the rate of accumulation of the myelin-specific components, myelin basic protein and cerebroside. For purposes of control, we also tested other putative, albeit less quantitative, indices of the rate of myelination. Levels of mRNA for ceramide galactosyltransferase (rate-limiting enzyme in cerebroside synthesis) and for myelin basic protein did not closely correlate with myelination at all times. Cholesterol synthesis closely matched the rate of cholesterol accumulation but did not track well with myelination. Synthesis of fatty acids did not correlate well with accumulation of either fatty acids (phospholipids) or myelin markers. We conclude that measurement of cerebroside synthesis rates provides a good measure of the rate of myelination. This approach may be useful as an additional parameter for examining the effects of environmental or genetic alterations on the rate of myelination.     

Epigenetic factors up-regulate expression of myelin proteins in the dysmyelinating jimpy mutant mouse

Proteolipid protein (PLP) is a major structural component of central nervous system (CNS) myelin. Evidence exists that PLP or the related splice variant DM-20 protein may also play a role in early development of oligodendrocytes (OLs), the cells that form CNS myelin. There are several naturally occurring mutations of the PLP gene that have been used to study the roles of PLP both in myelination and in OL differentiation. The PLP mutation in the jimpy (jp) mouse has been extensively characterized. These mutants produce no detectable PLP and exhibit an almost total lack of CNS myelin. Additionally, most OLs in affected animals die prematurely, before producing myelin sheaths. We have studied cultures of jp CNS in order to understand whether OL survival and myelin formation require production of normal PLP. When grown in primary cultures, jp OLs mimic the relatively undifferentiated phenotype of jp OLs in vivo. They produce little myelin basic protein (MBP), never immunostain for PLP, and rarely elaborate myelin-like membranes. We report here that jp OLs grown in medium conditioned by normal astrocytes synthesize MBP and incorporate it into membrane expansions. Some jp OLs grown in this way stain with PLP antibodies, including an antibody to a peptide sequence specific for the mutant jp PLP. This study shows that: (1) an absence of PLP does not necessarily lead to dysmyelination or OL death; (2) OLs are capable of translating at least a portion of the predicted jp PLP; (3) the abnormal PLP made in the cultured jp cells is not toxic to OLs. These results also highlight the importance of environmental factors in controlling OL phenotype. © 1996 John Wiley & Sons, Inc.     

Soluble and bound acid protease activity of myelin from bovine cerebral white matter and spinal cord

Isolated myelin of bovine spinal cord was found to degrade exogenous myelin basic protein (MBP) at pH 4.4. Electrophoretic peptide patterns were consistent with limited proteolysis of MBP. Some of the proteolytic activity was soluble at increased ionic strength, some remained bound, withstanding extraction at 37°C for up to 12 hr. While being measurable with exogenous MBP, bound protease degraded neither bound MBP nor any other major intrinsic myelin protein. Both soluble and bound protease activity was completely inhibited by pepstatin A. The patterns of limited proteolysis of MBP they produced were identical. Myelin of cerebral white matter also exhibited soluble and bound acid protease activity which was likewise inhibited by pepstatin A. Protease activity of spinal cord and cerebral myelin is therefore suggested to be due to a cathepsin D-like endopeptidase, present in a loosely and tightly bound form. Both forms increased by 50 to 80% in activity when myelin was isolated from mixtures of white and cortical gray matter. While increased soluble activity of myelin is consistent with binding of cathepsin D of lysosomal origin during the isolation of myelin the tightly bound form might point to a principal mechanism through which exogenous proteins may become attached to the myelin sheath in vivo.     

Alterations in metabolism and gene expression in brain regions during cuprizone-induced demyelination and remyelination

Exposure of mice to the copper chelator, cuprizone, results in CNS demyelination. There is remyelination after removal of the metabolic insult. We present brain regional studies identifying corpus callosum as particularly severely affected; 65% of cerebroside is lost after 6 weeks of exposure. We examined recovery of cerebroside and ability to synthesize cerebroside and cholesterol following removal of the toxicant. The temporal pattern for concentration of myelin basic protein resembled that of cerebroside. We applied Affymetrix GeneChip technology to corpus callosum to identify temporal changes in levels of mRNAs during demyelination and remyelination. Genes coding for myelin structural components were greatly down-regulated during demyelination and up-regulated during remyelination. Genes related to microglia/macrophages appeared in a time-course (peaking at 6 weeks) correlating with phagocytosis of myelin and repair of lesions. mRNAs coding for many cytokines had peak expression at 4 weeks, compatible with intercellular signaling roles. Of interest were other genes with temporal patterns correlating with one of the three above patterns, but of function not obviously related to demyelination/remyelination. The ability to correlate gene expression with known pathophysiological events should help in elucidating further function of such genes as related to demyelination/remyelination.     

Increased susceptibility to degradation by trypsin and subtilisin of in vitro peroxidized myelin proteins

We examined the possibility that the peroxidative damage to central nervous system myelin produced by reactive oxygen species (ROS), could modify the susceptibility of its proteins to the proteolytic action of proteases such as trypsin and subtilisin. Purified myelin membranes obtained from adult rat brains were in vitro peroxidized by two non-enzymatic systems: Fe³⁺ plus ascorbic acid and Cu²⁺ plus hydrogen peroxide. Myelin proteins were severely affected by peroxidation. There was an increase in the amount of carbonyl groups (CO), accompanied by and enhanced susceptibility to degradation by trypsin and subtilisin of myelin basic proteins (MBP) and of the major proteolipid protein (PLP). The effect upon the degradation of myelin protein is a possible consequence of the appearance in the structure of myelin proteins of peroxidative modifications that contribute to the recognition by proteolytic enzymes. This hypothesis is supported by the fact that if peroxidation of myelin membranes is done in the presence of EDTA, both CO formation and increased sensitivity to enzymatic breakdown disappear. These results suggest that the appearance of abnormal post-translational modifications in the myelin membrane produced by peroxidation could constitute a putative mechanism of modulating the capacity of myelin proteins to be metabolized by proteases.     

Impaired proteasome activity and accumulation of ubiquitinated substrates in a hereditary neuropathy model

Accumulation of misfolded proteins and alterations in the ubiquitin-proteasome pathway are associated with various neurodegenerative conditions of the CNS and PNS. Aggregates containing ubiquitin and peripheral myelin protein 22 (PMP22) have been observed in the Trembler J mouse model of Charcot-Marie-Tooth disease type 1A demyelinating neuropathy. In these nerves, the turnover rate of the newly synthesized PMP22 is reduced, suggesting proteasome impairment. Here we show evidence of proteasome impairment in Trembler J neuropathy samples compared with wild-type, as measured by reduced degradation of substrate reporters. Proteasome impairment correlates with increased levels of polyubiquitinated proteins, including PMP22, and the recruitment of E1, 20S and 11S to aggresomes formed either spontaneously due to the Trembler J mutation or upon proteasome inhibition. Furthermore, myelin basic protein, an endogenous Schwann cell proteasome substrate, associates with PMP22 aggregates in affected nerves. Together, our data show that in neuropathy nerves, reduced proteasome activity is coupled with the accumulation of ubiquitinated substrates, and the recruitment of proteasomal pathway constituents to aggregates. These results provide novel insights into the mechanism by which altered degradation of Schwann cell proteins may contribute to the pathogenesis of certain PMP22 neuropathies.