• basolateral sorting;
  • cell polarity;
  • immunoglobulin superfamily;
  • membrane targeting;
  • multiple sclerosis


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Oligodendrocytes possess two distinct membrane compartments – uncompacted plasma membrane (cell body, processes) and compact myelin. Specific targeting mechanisms must exist to establish and maintain these membrane domains. Polarized epithelial cells have the best characterized system for targeting components to apical and basolateral compartments. Since oligodendrocytes arise from neuroepithelial cells, we investigated whether they might utilize targeting paradigms similar to polarized epithelial cells. Myelin/oligodendrocyte glycoprotein (MOG) is a transmembrane Ig-like molecule restricted to uncompacted oligodendroglial plasma membrane. We stably expressed MOG in Madin–Darby canine kidney (MDCK) Type II epithelial cells, which have been extensively used in protein-targeting studies. Data from surface biotinylation assays and confocal microscopy revealed that MOG sorts exclusively to the basolateral membrane of MDCK cells. Expression vectors containing progressive truncations of MOG from the cytoplasmic C-terminus were expressed in MDCK cells to localize basolateral sorting signals. A loss of only four C-terminal residues results in some MOG expression at the apical surface. More strikingly, removal of the C-terminal membrane associated hydrophobic domain from MOG results in complete loss of basolateral sorting and specific targeting to the apical membrane. These data suggest that myelinating oligodendrocytes may utilize a sorting mechanism similar to that of polarized epithelia.

Abbreviations used

amino acid(s)






Dulbecco's modified Eagle's medium


Hank's balanced salt solution




monoclonal antibody


myelin basic protein(s)


Madin–Darby canine kidney


myelin/oligodendrocyte glycoprotein




polyacrylamide gel electrophoresis


phosphate buffered saline


polymerase chain reaction


proteolipid protein


sodium dodecyl sulfate.

A myelinating oligodendrocyte elaborates a complex array of cellular architectures with numerous slender processes that project outward from its cell body to terminate in highly specialized membrane sheets of compact myelin. While the cytoplasmic compartment is contiguous from the cell body to the myelin sheath, distinct membrane domains are evident with clear asymmetric regional localizations of lipids and proteins. As compared with membranes of the cell body and processes, compact myelin is particularly enriched in glycosphingolipids and cholesterol (Norton and Cammer 1984; Morell and Quarles 1999). Moreover, specific membrane proteins show restricted expression patterns in periaxonal areas (i.e. myelin-associated glycoprotein, MAG), compact myelin (i.e. proteolipid protein, PLP), or at the oligodendrocyte cell body (i.e. myelin/oligodendrocyte glycoprotein, MOG)(Eng et al. 1968; Webster et al. 1983; Brunner et al. 1989).

MOG is a CNS-specific integral membrane protein that is localized to oligodendrocyte cell bodies, oligodendroglial processes and the outermost wraps of myelin sheaths; it is virtually excluded from compact myelin layers and the periaxonal space (Linington et al. 1988; Brunner et al. 1989). A consideration of the asymmetric distribution of MOG in oligodendrocytes led us to study its membrane targeting. It was originally identified by a mouse monoclonal antibody (mAb) 8–18C5, which was raised against rat cerebellar glycoproteins (Linington et al. 1984). Our analysis of rat MOG cDNAs revealed a 1.6-kb mRNA transcript that encodes a mature 218 amino acid peptide (Gardinier et al. 1992). The sequence information predicted a 25-kDa protein with one immunoglobulin (Ig)-like domain, an N-linked glycosylation site, and two potential membrane spanning domains. By immunoblot analysis of sodium dodecyl sulfate (SDS) polyacrylamide gels, MOG protein migrates as a 26–28-kDa doublet exhibiting differential glycosylation, and deglycosylation results in the predicted 25-kDa moiety (Matthieu and Amiguet 1990; Amiguet et al. 1992). Developmentally, MOG is a late marker for mature myelinating oligodendrocytes (Scolding et al. 1989; Solly et al. 1996). With its two extensive hydrophobic domains, MOG is a unique Ig-like molecule. The more N-terminal domain spans the lipid bilayer, and the more C-terminal domain likely serves as a re-entrant loop on the cytoplasmic side of the plasma membrane (Kroepfl et al. 1996; della Gaspera et al. 1998). MOG residues 151–218 comprise this re-entrant loop region with its short, flanking hydrophilic elements, and it shows 94–97% identity based on amino acid sequence comparisons between baboon (Ballenthin and Gardinier, unpublished data), cow, human, mouse and rat MOG cDNAs (Gardinier et al. 1992; Pham-Dinh et al. 1993, 1994). These data suggest that this distinctive intracellular aspect of MOG plays a key functional role for this protein.

Within cells, a polarized phenotype is defined by a plasma membrane that has been organized into at least two distinguishable regions. Cell polarity is a universal and essential aspect in development, tissue architecture and maintenance of functional domains. In order to establish a polarized cell, a sorting mechanism is required to effect the distribution of lipids and proteins to specified regions and then to maintain these distinct asymmetric profiles. In epithelial cells, newly synthesized proteins destined for the apical or basolateral domain are sorted into distinct transport vesicles at the trans-Golgi network, followed by vesicular transport to the appropriate surface, where vesicles then fuse with the plasma membrane (Wandinger-Ness et al. 1990). Tight junctions also form a physical barrier that separates apical and basolateral surfaces at lateral interfaces near the apical surface. Thus lateral diffusion of lipids and proteins between apical and basolateral domains is blocked, as well as the flow of molecules and ions between adjacent epithelial cells (Rodriguez-Boulan and Nelson 1989).

Precise mechanisms must exist for the asymmetrical targeting of oligodendroglial membrane proteins. Both classical electron microscopy (Dermietzel 1974; Mugnaini and Schnapp 1974) and recent molecular genetic studies (Gow et al. 1999; Morita et al. 1999) have demonstrated that parallel tight junction strands exist at the interface between compact myelin and uncompacted plasma membrane of oligodendrocytes. Due to the architectural complexity of these cells, it is difficult to study in vivo vesicular targeting in oligodendrocytes and myelin. In vitro culture systems remain inadequate, because oligodendrocytes produce extensive membrane sheets without normal tight junction structures that would restrict lipids and proteins in a manner equivalent to that found within myelinating oligodendrocytes in vivo. Oligodendrocytes fail to form compact myelin when cultured as isolated cells. As they originate embryologically from neuroepithelial cells, we believe that these cells may establish their distinct domains by utilizing protein sorting mechanisms similar to those found in polarized epithelial cells. Numerous studies of lipid/protein sorting in epithelial cells have been done with epithelial Madin–Darby canine kidney Type II (MDCKII) cells (McRoberts et al. 1981), and these cells have been used extensively as a model for studying the sorting of viral proteins, endogenous proteins and heterologous eukaryotic proteins. The periaxonal myelin glycoprotein, MAG, has been studied using this paradigm (Minuk and Braun 1996). Indeed, many of the underlying mechanisms necessary for polarized sorting have been elucidated using these cells as a model. Although polarized cell lines may sort some proteins differently, in general, conservation of sorting signals seems to predominate. While apical targeting signals appear to reside in the extracellular or transmembrane domain, basolateral sorting determinants seem to lie within cytoplasmic domains (for review –Ikonen and Simons 1998).

MOG was stably transfected into MDCKII cells to determine if it would show preferential targeting to apical or basolateral surfaces in these cells. Surface biotinylation of membrane proteins and confocal microscopy revealed that MOG protein is specifically targeted to the basolateral membrane. Site-directed mutagenesis was used to truncate the highly conserved cytoplasmic domain of MOG, and a basolateral targeting domain was identified.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

MOG cDNAs and stable expression in MDCK cells

A SacI-ApaI digest of pMOG39 (Gardinier et al. 1997; GenBank accession #U64572), containing a full-length mouse MOG cDNA insert, was subcloned into a eukaryotic expression vector, pSRPBα (Takebe et al. 1988). This vector utilizes a pSP65 plasmid backbone, the SRα promoter, the SV40 polyadenylation signal, an antibiotic resistance cassette (geneticin,G418; or hygromycin), and a pBluescript® multicloning region. It provides good heterologous expression in a wide range of eukaryotic cells (G. Kansas, personal communication). Briefly, MDCKII cells were grown in Dulbecco's modified Eagle's medium (DMEM, Cat. no. 23700–040; Gibco-BRL, Grand Island, NY, USA) with 10% heat-inactivated fetal bovine serum (Harlan, Indianapolis, IN, USA) in the presence of antibiotics (penicillin/streptomycin; Gibco-BRL) and stably transfected by calcium phosphate precipitation of MOG cDNA on to subconfluent cells. Cells (∼5.0 × 105) were plated in 100-mm tissue culture dishes ∼24 h prior to transfection with 20 μg purified MOG cDNA (Mammalian Transfection Kit; Stratagene, La Jolla, CA, USA). About 24 h post-transfection, cells were split 1 : 15 following trypsinization, and antibiotic selection was begun with 400 μg/mL G418 (Gibco-BRL). Culture medium was changed at 3-day intervals with incremental G418 selection to 600 μg/mL and 800 μg/mL at each cell feeding with clonal selection at 800 μg/mL. Isolated clonal lines were then maintained in complete medium containing 200 μg/mL G418.

Site-directed mutagenesis was used to introduce premature termination codons that would effectively truncate MOG and block translation of specific cytoplasmic domains. Each exon within the MOG gene encodes a modular aspect of the protein structure of MOG (Fig. 1). Briefly, the single Ig-like loop of MOG is encoded by exon 2, the two hydrophobic domains are encoded by exons 3 and 6, and the hydrophilic cytoplasmic domains are encoded by exons 4, 5, 7 and 8. Mutagenic antisense oligonucleotides were prepared with internal stop codons (antisense sequence, tca), which would introduce a premature termination site at or near three specific exon junctions – exon 7/8 (MOG214), exon 6/7 (MOG207), and exon 5/6 (MOG173):


Figure 1. MOG membrane topology and site directed mutagenesis of MOG cDNA. (a) Model for the topology of MOG in the plasma membrane. MOG is an integral membrane protein with a single extracellular Ig-like domain containing a cysteine-cysteine disulfide linkage and an N-linked carbohydrate (CHO) site. MOG has one transmembrane domain (residues 121–151) and a hydrophobic membrane associated domain (residues 174–201). (b) Truncations of MOG cDNAs by site directed mutagenesis. Premature stop codons within the coding region of MOG were introduced to progressively truncate expression of the C-terminal cytoplasmic domain of MOG following the transmembrane region. MOG214 terminates at the end of exon 7. MOG207 terminates at the end of exon 6. MOG173 terminates at the beginning of exon 6 and lacks the hydrophobic re-entrant loop domain.

Download figure to PowerPoint

MOG214 (encodes residues 1–214)

tcaaaaggggtttcatagctcttcaaga …Leu214-Stop;

MOG207 (encodes residues 1–207)

tcaagaaactgtcatgccagtcttcggtg …Ala207-Stop;

MOG173 (encodes residues 1–173)

ccagcagggcactcacaggaagtgaggat …Leu173-Stop.

The integrity of the open reading frames as well as the specific MOG mutations were confirmed by DNA sequence analyses. Each mutant MOG cDNA (MOG214, MOG207 and MOG173) was transfected into MDCKII cells as described above for full-length MOG218 cDNA. Stably expressing cell lines were isolated, and MOG specific protein expression was determined by immunoblot analysis, surface biotinylation of membrane proteins and immunocytochemistry.

Immunoblot analysis

MOG-expressing MDCKII cell lines (MOG218, MOG214, MOG207 and MOG173) were screened for MOG protein expression. Untransfected wild-type MDCKII cells do not express MOG endogenously. Cells at ∼80–90% confluency were trypsinized, rinsed in Hank's balanced salt solution (HBSS, Gibco-BRL), and lysed in RIPA buffer (150 mm NaCl, 1% NP40, 0.5% deoxycholic acid, 0.1% SDS, 50 mm Tris pH 8.1, 5 mm EDTA, 2 mm EGTA) in the presence of Complete™ protease inhibitors (Roche-Boehringer-Mannheim, Indianapolis, IN, USA). A modified Lowry assay (Lees and Paxman 1972) was used to determine total protein concentration, and cell lysates (∼25 μg) were separated by SDS polyacrylamide gel electrophoresis (SDS–PAGE, 12% separating, 3.75% stacking gels). Proteins were transferred electrophoretically (0.5 A, 16 h, 4°C) to Immobilon P (Millipore, Bedford, MA, USA) polyvinyldifluoride (PVDF) membrane, and MOG was identified using 8–18C5 MOG mAb hybridoma supernatant (1 : 40 dilution in Tris-buffered saline, TBS). Horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG Ab (1 mg/mL; Promega, Madison, WI, USA) was diluted 1 : 20 000 and detected by chemiluminescence (LumiGLO; Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA). Immobilon filters were washed in TBS/0.05% Tween-20 (TBST). Blocking solution and Ab dilutions were made using TBST with 5% Carnation dry milk.

Surface biotinylation of membrane proteins

MDCKII cells adopt a columnar polarized phenotype when grown on Transwell™ tissue culture inserts (Costar, Corning, NY, USA). Biotinylation of membrane proteins at apical or basolateral surfaces of MDCKII cells was performed using a protocol from Dr R. Bacallao (Indiana University School of Medicine, Indianapolis, IN, USA). Briefly, ∼2.5 × 105 cells were seeded on semipermeable polycarbonate filters of 24 mm Transwell™ insert chambers, and culture medium was changed after 3 days. On day 4, culture wells were rinsed briefly in cold phosphate-buffered saline containing Mg2+ and Ca2+ (PBS2+), and cold PBS2+ with 1 mg/mL sulfosuccinimidyl-6-(biotinamido)hexanoate (EZ-Link™ Sulfo-NHS-LC-Biotin; Pierce, Rockford, IL, USA) was added to either the upper apical (1 mL) or the lower basolateral (2 mL) side of the Transwell™ filter chamber. Cold PBS2+ was also added to the side without biotinylation agent. Cells were protected from light and incubated at 4°C for 30 min with agitation. The biotin/PBS2+ solution was then removed, and biotinylation reactions were quenched by adding cold PBS2+ with 50 mm NH4Cl to both sides of the chamber (30 min with agitation, 4°C). Cells were lysed with 1% SDS, 4 mm EGTA, 10 mm Tris pH 8.0, and Complete™ protease inhibitors (∼100 μL per filter). Cell lysates were boiled (∼7 min), and RIPA buffer (∼900 μL) was then added. Lysates were incubated with 75 μL ImmunoPure™ immobilized streptavidin (Pierce) to precipitate biotinylated proteins (∼1 h, 4°C). Biotin–streptavidin complexes were washed three times sequentially in low salt (0.15 m NaCl), high salt (0.5 m NaCl), and no salt wash buffers (0.5% Triton X-100, 15 mm Tris pH 8.0, 4 mm EGTA, Complete™ protease inhibitors). Between washes, complexes were collected by centrifugation (10 000 g, 20 s, 4°C). The final material was resuspended in 2 × Laemmli buffer and boiled (∼5 min) prior to loading for SDS–PAGE.

Confocal microscopy

Confocal microscopy of MOG expressing polarized MDCKII cells was performed using a protocol from Dr R. Bacallao (Indiana University School of Medicine, Indianapolis, IN, USA). Briefly, ∼2.5 × 105 cells were seeded on semipermeable polycarbonate filters of 24-mm Transwell™ insert chambers, and the medium was changed after 3 days. On day 4, the filters were washed briefly (∼5 s) with PBS containing 80 mm KPIPES pH 6.8, 5 mm EGTA, 2 mm MgCl2 prewarmed to 37°C. Cells were then fixed in 0.25% glutaraldehyde, 0.1% Triton X-100, 80 mm KPIPES pH 6.8, 5 mm EGTA, and 2 mm MgCl2 (∼10 min, room temperature, 20°C). The fixative was removed, and reactions were quenched by washing the cells three times in PBS containing 3 mm NaBH4 (10 min each, room temp). Filters were washed three times in PBS containing 0.1% Triton X-100, 0.2% fish skin gelatin (PBS/TX100/FSG)(15 min each, room temp). Filters were cut into pieces (as needed) and incubated in 8–18C5 MOG mAb hybridoma supernatant diluted 1 : 20 in PBS/TX100/FSG solution (∼16–20 h, 4°C). EF21D8 mouse IgG mAb supernatant (1 : 3 dilution) was used to identify an apical membrane marker, GP135 (Ojakian and Schwimmer 1988). E-cadherin mouse IgG mAb supernatant (1 : 10 dilution) was used as a marker for basolateral surfaces (Gumbiner and Simons 1986). After incubation with primary antibody (∼16 h, 4°C), filter pieces were washed three times in the PBS/FSG solution (15 min each, room temp) and then in PBS alone (5 min, room temp). Filters were then incubated with fluorescein-labeled goat anti-mouse IgG secondary antibody (Calbiochem, La Jolla, CA, USA) diluted 1 : 100 in PBS/TX100/FSG (∼3 h, room temp). Filters were again washed as described above. Cells were postfixed in PBS containing 4% paraformaldehyde (30 min, room temp) and then quenched by incubation in PBS containing 50 mm NH4Cl (30 min, room temp). Filters were mounted on slides with Gelvatol containing 2.5% DABCO (Sigma Chemical, St Louis, MO, USA), and coverslips were applied atop nailpolish posts to preserve columnar cell architecture. Photomicroscopy was performed using a BioRad MRC-1024 laser scanning confocal microscope equipped with filters for immunofluorescence. Total magnification for all photographs was 630×.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

MOG expression in MDCK cells

Full-length and truncated MOG cDNA constructs were stably expressed in MDCKII cells. These cDNAs are shown in Fig. 1, and the relationship of the truncations is shown relative to the membrane topology of MOG (Fig. 1a) and the MOG cDNA/genomic structure (Fig. 1b). G418-resistant MDCKII cell lines were tested for MOG reactivity with 8–18C5 MOG mAb, and a panel of MDCK-MOG lines was selected from immunoblot analyses. MOG218, MOG214, MOG207 and MOG173 migrated in a manner consistent with their calculated molecular weights – 25.1, 24.6, 23.8 and 19.8 kDa, respectively (Fig. 2).


Figure 2. Immunoblot detection of full-length and truncated MOG in MDCKII cells. MDCKII cells were stably transfected with a eukaryotic expression vector encoding full-length MOG (218) or C-terminally truncated MOG (214, 207, or 173). Purified myelin (My, 2 μg) and cell lysates (5 μg/lane) of MOG218 (lane 1), MOG214 (lane 2), MOG207 (lane 3), or MOG173 (lane 4) were separated in 12% SDS–PAGE gels and electrophoretically transferred to Immobilon™ membrane filters. MOG proteins were identified with MOG mAb 8–18C5, horseradish peroxidase conjugated rabbit anti-mouse IgG, and chemiluminescent detection.

Download figure to PowerPoint

A preliminary study was performed to reaffirm that all MOG+ cell lines were setting up polarized epithelial layers in the same manner as untransfected MDCKII cultures. MDCK-MOG cells were cultured in Transwell™ chambers so that the cells would assume their polarized columnar phenotype. Upon reaching confluence, the cells set up tight junctions that separate their apical and basolateral compartments. This physical barrier restricts free diffusion of ions and macromolecules between cells, and these molecules must be transported through cells. This separation results in a measurable electrical resistance between apical and basolateral chambers of the tissue culture insert, which was assessed with an EVOM transepithelial resistance meter. Ions flow freely between the apical and basolateral chambers in cells lacking a network of tight junctions, and the resistance measured in this state is only that due to the semipermeable tissue culture insert membrane itself. Polarized MDCKII cells typically exhibit a resistance of 100–200 ohm/cm2 (Simons and Virta 1994). All MOG cell lines had resistance values in this range, indicating that functional tight junctions had been established. Finally, two well-characterized compartmental markers, E-cadherin and GP135, were used as endogenous internal positive controls for basolateral and apical sorting, respectively (Ojakian and Schwimmer 1988; Wollner et al. 1992). Confocal microscopy was used to show that these proteins were sorted appropriately in MOG+ MDCKII cells and demonstrated that their distribution was not perturbed by heterologous expression of MOG peptides (Fig. 3a). Surface biotinylation of E-cadherin confirmed its predominant localization at the basolateral surface (Fig. 3c); GP135 mAb is ineffective in immunoblotting and could not be studied in this manner. Multiple cell lines expressing each construct – both full-length and truncated MOG cDNAs – were tested in this manner, and all lines showed appropriate targeting of E-cadherin and GP135 proteins, as well as consistent targeting of each MOG construct.


Figure 3. Full-length MOG targets to the basolateral compartment of MDCK cells. (a) Laser scanning confocal microscopy was performed on MDCKII cells expressing GP135, E-cadherin, and full-length MOG. Confluent polarized cells were fixed with 0.25% glutaraldehyde, and immunocytochemistry was performed using antibodies specific for GP135 (apical marker), E-cadherin (basolateral marker), or MOG. Localization of each protein was examined by laser scanning confocal microscopy of 0.7 μm optical sections from the apical (Ap) surface through the basolateral (Bl) surface. (b) Surface biotinylation of MOG protein in MDCK cells. A biotinylating agent was added to either the apical (Ap) or basolateral (Bl) side of culture chambers containing confluent polarized cells. Biotinylated surface proteins were precipitated from cell lysates, separated in 12% SDS–PAGE gels, and electrophoretically transferred to Immobilon™ membrane filters. MOG protein was identified with MOG mAb 8–18C5, horseradish peroxidase conjugated rabbit anti-mouse IgG, and chemiluminescent detection. (c) Biotinylated E-cadherin was detected as described in (b) with E-cadherin mAb rr1, horseradish peroxidase-conjugated rabbit anti-mouse IgG, and chemiluminescent detection.

Download figure to PowerPoint

Full-length MOG is targeted to the basolateral surface

Polarization of MDCKII cells expressing full-length MOG218 was verified by transepithelial resistance measurements, and these cells were then used for immunocytochemical analysis of the localization of MOG in the plasma membrane. Portions of each Transwell™ membrane filter with the fixed cell layers were stained for MOG, GP135, and E-cadherin reactivity. The localization of these proteins was then visualized by laser scanning confocal microscopy. Optical sections were collected starting with the apical surface and proceeding at 0.70 μm sections down through the cells to the basal surface. As mentioned above, strong positive GP-135 and E-cadherin staining was apparent on apical and basolateral surfaces, respectively (Fig. 3a). MOG218 was localized exclusively to the basolateral surface and showed the cobblestone staining pattern that is typical of basolateral membrane proteins (Fig. 3a). MOG staining was not detected at the apical surface. Reminiscent of restricted expression in oligodendrocytes, the sorting machinery of MDCKII cells likely recognizes a basolateral sorting determinant present in MOG and targets it accordingly.

In order to get a finer assessment of the distribution of MOG at the two surfaces, biotinylation and immunoblot assays were performed. Cell layers were again grown to confluency, and transepithelial resistance measurements were taken to verify the establishment of tight junctions. The biotinylating agent, Sulfo-NHS-LC-Biotin, was added to either the apical or basolateral compartment of the Transwell™ chamber. Since Sulfo-NHS-LC-Biotin is not membrane soluble and cannot cross the lipid bilayer, it only biotinylates membrane proteins exposed at the cell surface. MOG and E-cadherin proteins were identified among the biotinylated proteins by immunoblot analysis (Figs 3b and c, respectively). As previously reported, E-cadherin sorts predominantly to the basolateral compartment with minor expression on the apical side (Wollner et al. 1992)(Fig. 3c). Consistent with the confocal microscopy data, MOG218 was expressed almost exclusively at the basolateral surface of the cells (Fig. 3b).

Basolateral targeting is lost upon removal of the hydrophobic re-entrant loop domain of MOG

Restricted targeting of membrane proteins typically requires sequence information contained within the protein itself. Basolateral delivery of transmembrane proteins is usually dependent upon a peptide sequence found within a discrete cytoplasmic domain (Casanova et al. 1991; Hunziker et al. 1991; LeBivic et al. 1991; Matter et al. 1992). Thus, the constructs described in Fig. 1 were designed to truncate MOG in an incremental fashion from its C-terminus on the cytoplasmic side of the plasma membrane. MDCKII cells stably expressing MOG214, MOG207 and MOG173 were used to localize the basolateral targeting domain of MOG.

Again, once cells established a polarized epithelium with high transepithelial resistance, immunocytochemistry and confocal microscopy were used to identify the localization of MOG in these cells. In contrast to the highly restricted basolateral sorting observed in MOG218 cells, any perturbation of the cytoplasmic domains of MOG resulted in disruption of MOG membrane targeting (Fig. 4). While MOG reactivity remained primarily at the basolateral surface, a significant amount of MOG reactivity was observed at the apical surface of MOG214 and MOG207 cells (Fig. 4a). Unexpectedly and in sharp contrast to the other MOG+ cells, MOG173 cells sorted this truncated MOG peptide almost exclusively to the apical surface with MOG barely detectable at the basolateral surface (Fig. 4a). These results suggest that a basolateral targeting signal lies between residues 173 and 207 – within or adjacent to the highly conserved hydrophobic re-entrant loop domain of MOG.


Figure 4. MOG173 reveals loss of basolateral targeting and unmasks an apical targeting domain in MOG. (a) Laser scanning confocal microscopy was performed on MDCKII cells expressing full length MOG (218) or truncated MOG (214, 207, or 173). Confluent polarized cells were fixed with 0.25% glutaraldehyde, and immunocytochemistry was performed using MOG mAb 8–18C5. Protein localization was examined by laser scanning confocal microscopy of 0.7 μm optical sections from the apical (Ap) through the basolateral (Bl) surface. (b) Surface biotinylation of full-length and truncated MOG proteins in MDCK cells. A biotinylating agent was added to either the apical (Ap) or basolateral (Bl) side of the culture chambers containing confluent polarized cells. Biotinylated surface proteins were precipitated from cell lysates, separated in 12% SDS–PAGE gels, and electrophoretically transferred to Immobilon™ membrane filters. MOG protein was identified with MOG mAb 8–18C5, horseradish peroxidase-conjugated rabbit anti-mouse IgG, and chemiluminescent detection.

Download figure to PowerPoint

Surface biotinylation of membrane proteins was also performed to provide a more sensitive assay for the distribution of MOG in the plasma membrane. While MOG214 and MOG207 cells express MOG that lacks only four and 11 C-terminal residues, respectively, recovery of biotinylated proteins from polarized cell layers revealed that significant amounts of MOG reactivity were found in the apical compartment (Fig. 4b). However, sorting of MOG in these two mutant lines still remained predominantly in the basolateral compartment, correlating with the confocal microscopy localization. As expected in the context of the confocal microscopy study, MOG173 cells showed a profound redistribution of MOG to the apical membrane surface with only a minor amount remaining at the basolateral surface (Fig. 4b). Indeed, the basolateral : apical targeting ratio for MOG173 appears to be the inverse of that observed for full-length MOG218. Besides a loss of basolateral targeting, removal of the hydrophobic re-entrant loop domain of MOG appears to unmask a potential apical targeting signal.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In recent years, studies have been undertaken to explore the asymmetric distribution of oligodendroglial membrane proteins. These investigations have used a variety of approaches, including biochemical membrane fractionations (Krämer et al. 1997; Kim and Pfeiffer 1999; Simons et al. 2000), expression of apical and basolateral markers in oligodendroglial cells (de Vries et al. 1998) and transfections of two oligodendroglial membrane proteins (MAG and MOG) into polarized MDCK epithelial cells (Minuk and Braun 1996; present study). Differential detergent extractions and flotation gradients have been used extensively to characterize lipid : protein associations in detergent insoluble glycosphingolipid-enriched complexes known as lipid rafts. An early study reported that MOG was not distributed in lipid rafts based on its solubility in Triton X-100 and its identification in high density membrane fractions (Krämer et al. 1997); however, subsequent work from the same laboratory reported that MOG distributed to lipid raft domains upon solubilization with CHAPS detergent (Simons et al. 2000). A more extensive study of the distribution of myelin membrane proteins in lipid microdomains reported that only MOG and CNP fulfilled all criteria for targeting to these areas, and only ∼40% was sequestered in this manner (Kim and Pfeiffer 1999). It has been suggested that MOG and CNP may act as signaling molecules upon their translocation into these raft compartments. It was also noted by these investigators that distinct populations of lipid rafts likely exist, which accounts for differential partitioning of oligodendroglial proteins depending upon the detergent used.

While lipid raft domains were originally postulated to function primarily in apical targeting, it cannot be assumed that all proteins associated with detergent-insoluble lipid microdomains must be sorted to apical domains. In particular, the formation of lipid rafts at the cell surface are increasingly implicated in serving as organizing platforms for signal transduction processes (Brown and London 1998). Previous studies have shown that basolateral proteins can indeed be found within these lipid microdomains (Sargiacomo et al. 1993; Melkonian et al. 1995). Several studies have shown that basolateral sorting determinants are most often found in discrete cytoplasmic regions of the targeted protein, and we have identified such a domain that preferentially targets MOG to the basolateral membrane surface of MDCKII cells. Deletion of only four (or 11) C-terminal amino acids from MOG resulted in significant quantities of MOG at the apical membrane surface, yet both mutated proteins were still sorted predominantly to the basolateral compartment. This aberrant sorting may be due to a decreased ability for signal recognition if the targeting signal also requires a noncontiguous determinant (Saunders et al. 1998). It may also result from a partial loss in retention at the basolateral membrane caused by transcytosis of MOG214 (or MOG207) to the apical surface. Future metabolic labeling studies will address this question. In either case, the near exclusivity to which full length MOG partitions to the basolateral compartment is both striking and consistent with its highly restricted expression in CNS oligodendrocytes.

Quite surprisingly, expression of MOG173, which lacks the hydrophobic re-entrant loop domain, revealed a specific apical targeting pattern. Given the dramatic difference in membrane targeting observed between MOG173 and MOG207, the virtually exclusive targeting to the apical compartment for MOG173 suggests that a basolateral sorting signal lies between residues 173 and 207 and that a default apical sorting signal exists in the absence of the basolateral signal. The cytoplasmic domain encoded by residues 175–218 is 100% conserved across all species studied to date. Within this region, a tyrosine-based motif that is recognized as a classical basolateral sorting signal is located. MOG residues 199–202 (Tyr-Asn-Trp-Leu) match a YXXφ motif, which is minimally defined by a critical tyrosine residue, any two amino acids, and a residue with a bulky hydrophobic side chain. It is one of a degenerate family of basolateral sorting signals that have been identified (for review – Bonifacino and Dell'Angelica 1999). Future experiments will more finely dissect this potential targeting domain, as well as consider its role more directly in the targeting of MOG within oligodendrocytes. In epithelial cells, N-linked carbohydrate attachment sites have been implicated in targeting some proteins to the apical compartment (Gut et al. 1998). MOG is differentially glycosylated at a conserved, single N-linked glycosylation site (Asn31), and we will determine if mutation of this carbohydrate attachment site results in a perturbation of the apical targeting found for MOG173.

The existence of a basolateral sorting signal at or near the more C-terminal hydrophobic domain of MOG is especially relevant with regard to the human MOG gene. We have identified eight alternatively spliced transcripts that are expressed by the human MOG gene during development (Ballenthin and Gardinier 1996). This complex splicing program involves alternative usage of exons that encode the cytoplasmic domains and, in particular, those exons that would encode the second hydrophobic domain and the hydrophilic C-terminal residues of MOG following this re-entrant loop domain. Four of eight MOG isoforms specifically lack the hydrophobic region (encoded by exon 6) that we have implicated in the basolateral targeting of MOG. Moreover, five of eight MOG isoforms replace the four C-terminal amino acids of MOG with nine different residues. As observed with MOG214, this alteration of the same four C-terminal residues may affect the membrane targeting of these MOG isoforms. We have reported that some of these alternatively spliced mRNAs may be more prevalent during early fetal development, and it is possible that altered membrane targeting of MOG gene products is required during this period. In addition to investigating the translation products from these alternatively spliced MOG mRNAs, we are now interested in assessing their membrane targeting specificities.

In the present study, we have shown that full-length mouse MOG is targeted to the basolateral compartment in MDCKII cells. In oligodendrocytes, MOG is predominantly found in the cell body, processes and outermost surface of myelin sheaths. More simply, MOG is found excluded from regions of compact myelin. We propose that the localization of MOG defines a domain within oligodendrocytes that is analogous to the basolateral domain of MDCKII cells. This hypothesis would conflict with an earlier report that a basolateral pathway is used by oligodendrocytes in vitro for myelin membrane formation (de Vries et al. 1998). Their conclusion was based upon immunocytochemical detection of vesicular stomatitis virus G protein and influenza virus hemagglutinin protein in primary oligodendrocyte cultures with these proteins defining basolateral and apical compartments, based upon their Triton X-100 solubility or insolubility, respectively. Similar studies by another laboratory failed to demonstrate differential targeting of these viral markers in cultured oligodendrocytes, and they proposed that their cultures were less mature and had not yet terminally differentiated (Simons et al. 2000). As discussed in the Introduction, we believe that primary cultures of purified oligodendrocyte do not establish a true polarized phenotype as compared with their in vivo environment.

As the restricted membrane targeting of MOG is observed in both oligodendrocytes and MDCKII cells, it is reasonable to consider that common sorting mechanisms may be utilized to achieve this localization. Indeed, oligodendrocytes arise from neuroepithelial cells, and it is attractive to propose that oligodendrocytes can establish and maintain their regional profiles for specific membrane proteins by using elements that also sort membrane proteins in epithelial cells. In order to postulate that common machinery may be used by both epithelial and oligodendrocytes, it is also necessary to consider other myelin membrane domains. How will proteins within compact myelin (e.g. PLP, P0 protein) sort using the MDCKII targeting model? Moreover, we can use this model and assess how MOG is transported to the basolateral compartment in MDCKII cells. We propose that this model may provide critical clues that will shed light on how membrane domains are established and maintained in myelinating oligodendrocytes.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We would like to thank our colleagues who generously provided us with various reagents for these studies: Dr Angela Wandinger-Ness (University of New Mexico, Albuquerque, NM, USA) for the MDCK Type II cells; Dr Robert Bacallao (University School of Medicine, Indianapolis, IN, USA) for protocols and helpful discussions; Dr George Ojakian (New York University, New York, NY, USA) for GP135 mAb; and Dr Geoffrey Kansas (Northwestern University Medical School, Chicago, IL, USA) for the pSRPBα eukaryotic expression vector. This work was supported by NIH-NINDS (NS35319).


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Amiguet P., Gardinier M. V., Zanetta J.-P., Matthieu J.-M. (1992) Purification and partial structural and functional characterization of mouse myelin/oligodendrocyte glycoprotein. J. Neurochem. 58, 16761682.
  • Ballenthin P. A. & Gardinier M. V. (1996) Myelin/oligodendrocyte glycoprotein is alternatively spliced in humans but not mice. J. Neurosci. Res. 46, 271281.
  • Bonifacino J. S. & Dell'Angelica E. C. (1999) Molecular bases for the recognition of tyrosine-based sorting signals. J. Cell Biol. 145, 923926.
  • Brown D. A. & London E. (1998) Functions of lipid rafts in biological membranes. Annu. Rev. Cell Dev. Biol. 14, 111136.
  • Brunner C., Lassmann H., Waehneldt T. V., Matthieu J.-M., Linington C. (1989) Differential ultrastructural localization of myelin basic protein, myelin/oligodendrocyte glycoprotein, and 2′,3′-cyclic nucleotide 3′ phosphodiesterase in the CNS of adult rats. J. Neurochem. 52, 296304.
  • Casanova J. E., Apodaca G., Mostov K. E. (1991) An autonomous signal for basolateral sorting in the cytoplasmic domain of the polymeric immunoglobulin receptor. Cell 66, 6575.
  • Dermietzel R. (1974) Junctions in the central nervous system of the cat. I. Membrane fusion in central myelin. Cell Tissue Res. 148, 565576.
  • Eng L. F., Chao T. C., Gerstl B., Pratt D., Tavastjerna M. G. (1968) The maturation of human white matter myelin. Fractionation of the myelin membrane proteins. Biochemistry 7, 44554465.
  • Gardinier M. V., Amiguet P., Linington C., Matthieu J.-M. (1992) Myelin/oligodendrocyte glycoprotein is a unique member of the immunoglobulin superfamily. J. Neurosci. Res. 33, 177187.
  • Gardinier M. V., Ballenthin P. A., Kroepfl J. F., Viise L. R. (1997) Myelin/oligodendrocyte glycoprotein: a molecular analysis, in Cell Biology and Pathology of Myelin ( JuurlinkB. H. J., DevonR. M., DoucetteJ. R., NazaraliA. J., SchreyerD. J. and VergeV. M. K., eds), pp. 3749. Plenum Press, New York.
  • Della Gaspera B., Pham-Dinh D., Roussel G., Nussbaum J. L., Dautigny A. (1998) Membrane topology of the myelin/oligodendrocyte glycoprotein. Eur. J. Biochem. 258, 478484.
  • Gow A., Southwood C. M., Li J. S., Pariali M., Riordan G. P., Brodie S. E., Danias J., Bronstein J. M., Kachar B., Lazzarini R. A. (1999) CNS myelin and sertoli tight junction strands are absent in Osp/Claudin-11 null mice. Cell 99, 649659.
  • Gumbiner B. & Simons K. (1986) A functional assay for proteins involved in establishing an epithelial occluding barrier: identification of a uvomorulin-like polypeptide. J. Cell Biol. 102, 457468.
  • Gut A., Kappeler F., Hyka N., Balda M. S., Hauri H. P., Matter K. (1998) Carbohydrate-mediated Golgi to cell surface transport and apical targeting of membrane proteins. EMBO J. 17, 19191929.
  • Hunziker W., Harter C., Matter K., Mellman I. (1991) Basolateral sorting in MDCK cells required a distinct cytoplasmic domain determinant. Cell 66, 907920.
  • Ikonen E. & Simons K. (1998) Protein and lipid sorting from the trans-Golgi network to the plasma membrane in polarized cells. Semin. Cell Devel. Biol. 9, 503509.
  • Kim T. & Pfeiffer S. E. (1999) Myelin glycosphingolipid/cholesterol-enriched microdomains selectively sequester the non-compact myelin proteins CNP and MOG. J. Neurocytol. 28, 281293.
  • Krämer E.-M., Koch T., Niehaus A., Trotter J. (1997) Oligodendrocytes direct glycosyl phosphatidylinositol-anchored proteins to the myelin sheath in glycosphingolipid-rich complexes. J. Biol. Chem. 272, 89378945.
  • Kroepfl J. F., Viise L. R., Charron A. J., Linington C., Gardinier M. V. (1996) Investigation of myelin/oligodendrocyte glycoprotein membrane topology. J. Neurochem. 67, 22192222.
  • LeBivic A., Sambuy Y., Patzak A., Patil N., Chao M., Rodriguez-Boulan E. (1991) An internal deletion in the cytoplasmic tail reverses the apical localization of human NGF receptor in transfected MDCK cells. J. Cell Biol. 115, 607618.
  • Lees M. B. & Paxman S. A. (1972) Modification of the Lowry procedure for the analysis of proteolipid protein. Anal. Biochem. 47, 184192.
  • Linington C., Webb M., Woodhams P. L. (1984) A novel myelin-associated glycoprotein defined by a mouse monoclonal antibody. J. Neuroimmunol. L 6, 387396.
  • Linington C., Bradl M., Lassmann H., Brunner C., Vass K. (1988) Augmentation of demyelination in rat acute allergic encephalomyelitis by circulating mouse monoclonal antibodies directed against a myelin/oligodendrocyte glycoprotein. Am. J. Pathol. 130, 443454.
  • McRoberts J. A., Taub M., Saler M. H. (1981) The Madin–Darby canine kidney (MDCK) cell line, in Functionally Differentiated Cell Lines ( SataG., ed.), pp. 117139. Alan Liss, New York.
  • Matter K., Hunziker W., Mellman I. (1992) Basolateral sorting of LDL receptor in MDCK cells: the cytoplasmic domain contains two tyrosine dependent targeting determinants. Cell 71, 741753.
  • Matthieu J.-M. & Amiguet P. (1990) Myelin/oligodendrocyte glycoprotein expression during development in normal and myelin-deficient mice. Dev. Neurosci. 12, 293302.
  • Melkonian K. A., Chu T., Tortorella L. B., Brown D. A. (1995) Characterization of proteins in detergent resistant membrane complexes from Madin–Darby canine kidney epithelial cells. Biochemistry 34, 1616116170.
  • Minuk J. & Braun P. E. (1996) Differential intracellular sorting of the myelin-associated glycoprotein isoforms. J. Neurosci. Res. 44, 411420.
  • Morell P. & Quarles R. H. (1999) Myelin formation, structure and biochemistry, in Basic Neurochemistry: Molecular, Cellular and Medical Aspects ( SiegelG. J., AgranoffB. M., AlbersR. W., FisherS. K. and UhlerM. D., eds), pp. 6993. Lippincott-Raven Publishers, Philadelphia.
  • Morita K., Sasaki H., Fujimoto K., Furuse M., Tsukita S. (1999) Claudin-11/Osp-based tight junctions of myelin sheaths in brain and Sertoli cells in testis. J. Cell Biol. 145, 579588.
  • Mugnaini E. & Schnapp B. (1974) Possible role of zonula occludens of the myelin sheath in demyelinating conditions. Nature 251, 725727.
  • Norton W. T. & Cammer W. (1984) Isolation and characterization of myelin. In: Myelin ( P.Morell, ed.), pp. 147196. Plenum Press, New York.
  • Ojakian G. K. & Schwimmer R. (1988) The polarized distribution of an apical cell surface glycoprotein is maintained by interactions with the cytoskeleton of Madin–Darby canine kidney cells. J. Cell Biol. 107, 23772387.
  • Pham-Dinh D., Mattei M.-G., Nussbaum J.-L., Roussel G., Pontarotti P., Roeckel N., Mather I. H., Artzt K., Lindahl K. F., Dautigny A. (1993) Myelin/oligodendrocyte glycoprotein is a member of a subset of the immunoglobulin superfamily encoded within the major histocompatibility complex. Proc. Natl Acad. Sci. USA 90, 79907994.
  • Pham-Dinh D., Allinquant B., Ruberg M., Gaspera B. D., Nussbaum J.-L., Dautigny A. (1994) Characterization and expression of the cDNA coding for the human myelin/oligodendrocyte glycoprotein. J. Neurochem. 63, 23532356.
  • Rodriguez-Boulan E. & Nelson W. J. (1989) Morphogenesis of the polarized epithelial cell phenotype. Science 245, 718725.
  • Sargiacomo M., Sudol M., Tang Z., Lisanti M. P. (1993) Signal transducing molecules and glycosyl-phosphatidylinositol-linked proteins form a caveolin-rich insoluble complex in MDCK cells. J. Cell Biol. 122, 789807.
  • Saunders C., Keefer J. R., Bonner C. A., Limbird L. E. (1998) Targeting of G protein coupled receptors to the basolateral surface of polarized renal epithelial cells involves multiple, non-contiguous structural signals. J. Biol. Chem. 273, 2419624206.
  • Scolding N. J., Frith S., Linington C., Morgan B. P., Campbell A. K., Compston D. A. S. (1989) Myelin/oligodendrocyte glycoprotein (MOG) is a surface marker of oligodendrocyte maturation. J. Neuroimmunol. 22, 169176.
  • Simons M., Krämer E.-M., Thiele C., Stoffel W., Trotter J. (2000) Assembly of myelin by association of proteolipid protein with cholesterol- and galactosylceramide-rich membrane domains. J. Cell Biol. 151, 143153.
  • Simons K. & Virta H. (1994) Growing Madin–Darby canine kidney cells for studying epithelial cell biology, in Cell Biology: a Laboratory Handbook ( CelisJ., ed.), pp. 225231. Academic Press, San Diego.
  • Solly S. K., Thomas J.-L., Monge M., Demerens C., Lubetzki C., Gardinier M. V., Matthieu J.-M., Zalc B. (1996) Myelin/oligodendrocyte glycoprotein (MOG) expression is associated with myelin deposition. Glia 18, 3948.
  • Takebe Y., Seiji M., Fujisawa J., Hoy P., Yokata K., Arai K., Yoshida M., Arai N. (1988) SR alpha promoter: an efficient and versatile mammalian cDNA expression system composed of the simian virus 40 early promoter and the R-U5 segment of the human T-cell leukemia virus type 1 long terminal repeat. Mol. Cell Biol. 8, 466472.
  • De Vries H., Schrage C., Hoekstra D. (1998) An apical-type trafficking pathway is present in cultured oligodendrocytes but the sphingolipid-enriched myelin membrane is the target of a basolateral-type pathway. Mol. Biol. Cell. 9, 599609.
  • Wandinger-Ness A., Bennett M. K., Anthony C., Simons K. (1990) Distinct transport vesicles mediate the delivery of plasma membrane proteins to the apical and basolateral domains of MDCK cells. J. Cell Biol. 111, 9871000.
  • Webster H. D., Palkovits C. G., Stoner G. L., Favilla J. T., Frail D. E., Braun P. E. (1983) Myelin-associated glycoprotein: electron microscopic immunocytochemical localization in compact developing and central nervous system myelin. J. Neurochem. 41, 14691479.
  • Wollner D. A., Krzeminski K. A., Nelson W. J. (1992) Remodeling of the cell surface distribution of membrane proteins during the development of epithelial cell polarity. J. Cell Biol. 116, 889899.