J Biol Chem, Vol. 274, Issue 41, 29042-29049, October 8, 1999

Compartmentation of Fyn Kinase with Glycosylphosphatidylinositol-anchored Molecules in Oligodendrocytes Facilitates Kinase Activation during Myelination^*

Eva-Maria Krämer, Corinna Klein, Thomas Koch, Monica Boytinck, and Jacqueline Trotter^§

From the Department of Neurobiology, University of Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In many cell types, glycosylphosphatidylinositol (GPI)-anchored proteins are sequestered in detergent-resistant membrane rafts. These are plasma membrane microdomains enriched in glycosphingolipids and cholesterol and are suggested to be platforms for cell signaling. Concomitant with the synthesis of myelin glycosphingolipids, maturing oligodendrocytes progressively associate GPI-anchored proteins, including the adhesion molecules NCAM 120 and F3, in rafts. Here we show that these microdomains include Fyn and Lyn kinases. Both kinases are maximally active in myelin prepared from young animals, correlating with early stages of myelination. In the rafts, Fyn kinase is tightly associated with NCAM 120 and F3. In contrast, in oligodendrocyte progenitor cells lacking rafts or in raft-free membrane domains of more mature cells, F3 does not associate with Fyn. The addition of anti-F3 antibodies to oligodendrocytes results in stimulation of Fyn kinase specifically in rafts. Compartmentation of oligodendrocyte GPI-anchored proteins in rafts is thus a prerequisite for association with Fyn, permitting kinase activation. Interaction of oligodendrocyte F3 with axonal ligands such as L1 and ensuing kinase activation may play a crucial role in initiating myelination.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The formation of a myelin sheath is essential for the rapid saltatory propagation of action potentials in the vertebrate nervous system (1). Myelination in the central nervous system involves sequential stages of interaction between the myelinating glial cell, the oligodendrocyte, and the neuronal process, the axon. Initial recognition and adhesion results in wrapping of the glial process around the axon (ensheathment) followed by the laying down of the multilamellar sheath. Although progenitor cells have the intrinsic potential to differentiate into oligodendrocytes in vitro in the absence of axons (2-4), myelination in vivo demonstrates a high degree of specificity and requires axonal signals. For example, oligodendrocytes do not normally myelinate dendrites (5, 6). Cell adhesion molecules expressed by both axons and glial cells are known to play a crucial role in the establishment of axon-glial contact and subsequent signaling to the oligodendrocyte, driving the process of myelination (7, 8). In turn, the myelinating glial cell triggers an axonal reaction, culminating in an increased phosphorylation of neurofilaments and regulation of the axonal diameter (9, 10).

Members of the cadherin and the Ig superfamily such as L1, NCAM,¹ and especially MAG are candidate molecules mediating axon-glial interactions (7, 11). The formation of morphologically normal myelin sheaths in MAG knockout and even in MAG/NCAM double knockout mice suggests that either these molecules are not involved in myelin formation or additional molecules participate in the early events of myelination (12-14). Our knowledge about the signals exchanged between axons and oligodendrocytes during myelination is also incomplete. A role for oligodendroglial Fyn kinase (a nonreceptor tyrosine kinase of the Src family) has been proposed as Fyn phosphorylates MAG after co-transfection in COS cells and Fyn knockout mice are hypomyelinated (15). The integrity of the axon-glial unit requires continual bidirectional signaling between the oligodendrocyte and the axon.

GPI-anchored proteins expressed by oligodendrocytes and their precursor cells are candidates for recognition molecules dictating the initial interactions between axon and oligodendrocyte and acting as receptors mediating axon-glial signal transduction. Oligodendrocyte precursor cells express a distinct pattern of GPI-anchored molecules, which is retained as the cells mature and is present in preparations of adult myelin. Although both oligodendrocyte precursor cells and mature oligodendrocytes express a similar pattern of GPI-anchored proteins, in oligodendrocytes and myelin, but not in precursor cells, these proteins are associated with the major myelin lipids galactocerebroside, sulfatide, and cholesterol in membrane domains. These domains can be isolated as detergent-insoluble glycosphingolipid-rich microdomains (DIGs) by sucrose density gradient centrifugation (16). DIGs can be isolated from several cell types and are thought to represent raft-like microdomains within the plasma membrane, resulting from the lateral assembly of sphingolipids and cholesterol in the exoplasmic leaflet of the lipid bilayer due to weak interactions between their polar headgroups (17, 18). In polarized epithelial cells, where rafts emerge in the trans-Golgi network, they are thought to be responsible for the apical sorting of glycosphingolipids, GPI-anchored proteins, and other apical marker molecules (17-19). We postulated that in oligodendrocytes, as in epithelial cells, DIGs represent raftlike membrane domains in which oligodendrocyte GPI-anchored proteins together with the major myelin lipids are sorted into the forming myelin sheath (16). Mice in which the enzyme UDP-galactose:ceramide galactosyl transferase catalyzing the synthesis of the myelin lipids galactocerebroside and sulfatide has been knocked out exhibit tremors and hind limb paralysis and die prematurely (20, 21). They show deficits in nerve conduction despite the formation of compact myelin. However, the nodal/paranodal structure in these mice is severely perturbed (22, 23). Whether these defects result from a loss of functional properties of these lipids such as insulation or from the loss of lipid-associated targeting of specific proteins including signaling molecules to areas of the forming myelin sheath is unelucidated.

As was first shown for T-lymphocytes, in many cell types DIG microdomains include nonreceptor tyrosine kinases of the Src family (24, 25). In this paper, we investigated the association of such kinases with GPI-anchored proteins in oligodendrocytes and myelin, since they could be involved in signal transduction between the wrapping glial cell process and the axon. Our results show that the two major oligodendroglial GPI-anchored proteins, the 120-kDa isoform of NCAM (26, 27) and F3 (28-30), both members of the Ig superfamily of adhesion molecules, colocalize with the Src family kinases Fyn and Lyn in oligodendrocyte and myelin DIGs. The activity of both kinases within the DIGs is developmentally regulated, being most active at the beginning of myelination. Fyn but not Lyn kinase activity is stably associated with both NCAM 120 and F3. Furthermore, antibody-mediated cross-linking of F3 results in stimulation of the Fyn kinase activity localized to oligodendrocyte DIGs. The association of NCAM 120 and F3 with Fyn kinase to a functional signaling complex within raftlike glycosphingolipid-rich microdomains during oligodendrocyte maturation may be critical for signal transduction between axon and glial cell in the early phases of myelination.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials and Antibodies-- Radiochemicals ([-³²P]ATP, L-[³⁵S]Met/Cys in vitro labeling mix) and ECL reagents were from Amersham Pharmacia Biotech (Braunschweig, Germany); human recombinant platelet-derived growth factor (AA) and basic fibroblast growth factor were from TEBU (Frankfurt, Germany); dibutyryl cyclic AMP (dbcAMP), Triton X-100, Nonidet P-40, and sodium deoxycholate were from Sigma (Deisenhofen, Germany); Protein A-Sepharose CL4B was from Amersham Pharmacia Biotech (Freiburg, Germany); Bradford reagent for protein assays was from Bio-Rad (München, Germany); polyvinylidene difluoride membrane was from Millipore (Bedford, MA). The amino-terminal peptide of F3 (KGFGPIFEEQPINT) was synthesized by Dr. R. Frank (ZMBH, University of Heidelberg, Germany).

The following rabbit polyclonal antibodies were used: antibodies recognizing NCAM (27), F3 (Ig fraction of a serum raised against the N-terminal F3 peptide; Ref. 29), the AN2 antigen (31), Fyn, and Lyn (Santa Cruz Biotechnology, Inc., Heidelberg). The following monoclonal antibodies were used: murine monoclonal antibody 27-11-111 (mAb F11 No. 8) made against chick F11 (32), which cross-reacts with mouse F3 (56), kindly provided by Dr. F. Rathjen (Berlin, Germany); rat monoclonal antibody AN2, which recognizes a surface epitope of a 330-kDa glycoprotein expressed by the cell line Oli-neu and primary oligodendrocyte progenitors (31); murine monoclonal antibody 4G10 against phosphotyrosine from Upstate Biotechnology, Inc. (Lake Placid, NY); and mouse monoclonal antibody against Fyn (Pharmingen/Transduction Laboratories). Secondary antibodies were from Dianova (Hamburg, Germany).

Cell Cultures and Metabolic Labeling-- Primary cultures of oligodendrocytes were prepared from embryonic day 14-16 mice as described (27, 33). Oligodendrocytes growing on top of astrocyte monolayers were shaken off and plated in modified Sato medium (27) containing 1% horse serum on poly-L-lysine-coated dishes. To increase the proportion of precursor cells as well as to promote survival, 10 ng/ml human recombinant platelet-derived growth factor (AA), and 5 ng/ml basic fibroblast growth factor were added immediately after the shake and after 24 h in vitro. Oligodendrocytes were kept for 5 days in vitro without further growth factor additions before they were harvested. The resulting population, which was used for all experiments with primary cultures, is enriched for differentiated oligodendrocytes but contains a fraction of progenitor cells (27). The cell line Oli-neu (34) was cultured in Sato medium containing 1% horse serum. To induce differentiation of the precursor-like Oli-neu cells, cultures were treated with 1 mM dbcAMP for 3-4 days (daily additions to the culture medium).

For metabolic labeling primary oligodendrocytes and Oli-neu cells were starved for 1 h in SO₄/Met/Cys-free DMEM and incubated for 4 h with 100 µCi/ml L-[³⁵S]Met/Cys labeling mix.

Myelin Preparation-- Myelin was isolated from the brains of young postnatal (postnatal days (P)9/10, P12, P16, P20, P30, or P45) and adult NMRI mice of both sexes according to standard procedures (35, 36). Initially, brains were homogenized in ice-cold 10.5% sucrose using the Ultra-Turrax T25 (IKA, Staufen, Germany). Myelin was removed from the interface between 10.5 and 30% sucrose step gradients and subjected to two rounds of hypoosmotic shock by resuspension in a large volume of ice-cold water and reisolation on the step gradient; this separates myelin membranes from axolemma. Purified myelin was collected from the final sucrose interface, washed twice with cold water, resuspended in a small volume of water, and immediately used or frozen in small aliquots at 80 °C. The protein content of the myelin preparations was determined by using the Bio-Rad protein assay with bovine serum albumin as a protein standard.

Preparation of Detergent Extracts and Sucrose Density Gradient Centrifugation-- Detergent extracts were prepared as described (19, 37). In brief, primary oligodendrocytes (2-3 × 10⁷) or sonicated myelin (300 µg of total protein) were solubilized at 4 °C in 1 ml of extraction buffer containing 10 mM Tris/HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 2% Triton X-100 (TNE/Triton X-100). The extracts were shaken for 30 min at 4 °C. Detergent extracts were adjusted to 40% sucrose by adding equal volumes of 80% sucrose in TNE without Triton X-100 and placed into an ultracentrifuge tube. A linear gradient from 5 to 30% sucrose (in TNE without Triton X-100) was layered over the lysate. Gradients were centrifuged for 12 h at 35,000 rpm at 4 °C in a Beckmann SW 40 TI rotor (218,000 × g). 1-ml fractions were harvested, and the density was determined by measurement of the refractive index. Proteins in each fraction were analyzed by SDS-PAGE followed by Western blot. Light gradient fractions containing floating GPI-anchored proteins, glycosphingolipids, and cholesterol (DIGs) were collected, diluted with double distilled H₂O, and pelleted for 1 h at 218,000 × g and 4 °C. Isolated DIGs were subjected to an in vitro kinase assay and immunoprecipitation. The protein concentration of the DIGs was evaluated with the Bio-Rad protein assay kit using bovine serum albumin as a protein standard.

Immunoblotting-- Proteins blotted onto polyvinylidene difluoride membrane were detected by incubation with primary antibodies overnight at 4 °C. The blots were incubated with a second anti-species antibody conjugated with HRP for 30-60 min at room temperature. The blots were developed with ECL reagents according to the manufacturer's instructions.

Membranes were stripped with 100 mM glycine, pH 2, for 30 min, blocked, and reprobed with antibodies.

Immunoprecipitation from DIGs and Bottom Fractions of the Gradients-- Isolated DIGs were resuspended in 0.5 ml of lysis buffer (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1 mM Na₃VO₄, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, and 1% Nonidet P-40), incubated with antibodies and Protein A-Sepharose and washed under stringent conditions in radioimmune precipitation buffer (50 mM Tris/HCl, pH 7.4, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM dithioerythritol, 100 µM Na₃VO₄, 10 mM NaF) and once in 50 mM Tris/HCl, pH 7.4, 100 µM Na₃VO₄, 10 mM NaF. The immunoprecipitate was subjected to a [-³²P]ATP in vitro kinase assay. To identify associated kinases, some samples were subjected to a second round of immunoprecipitation; immune complexes were denatured and dissociated by two sequential incubations in 50 µl of 50 mM Tris/HCl, pH 7.4, 0.5% SDS, and 1% -mercaptoethanol for 10 min at 95 °C followed by centrifugation. Supernatants eluted from the Sepharose beads were diluted in 500 µl of lysis buffer and incubated with Fyn antibodies and Protein A-Sepharose for 1 h at room temperature. All samples were analyzed by SDS-PAGE, and radioactive protein bands were detected with a phosphoimager or autoradiography.

DIGs that had been subjected to a [-³²P]ATP kinase assay were immunoprecipitated with antibodies against Fyn and Lyn according to the same procedure, but SDS-PAGE analysis was performed directly after the precipitation.

For immunoprecipitation from bottom fractions of the gradients, 250 µl of the fraction was diluted with an equal volume of lysis buffer and processed similarly to the DIG fractions above.

In Vitro Kinase Assay-- Immune complexes were resuspended in 20 µl of kinase buffer (20 mM HEPES, pH 7.4, 5 mM MgCl₂, 1 mM MnCl₂, 100 µM Na₃VO₄) and incubated with 5 µCi of [-³²P]ATP for 20 min at room temperature. The samples were washed and subjected either to SDS-PAGE or a second round of immunoprecipitation to identify associated kinases.

DIGs were isolated from sucrose density gradients, resuspended in 50 µl of kinase buffer, and incubated with 10 µCi of [-³²P]ATP for 20 min at room temperature. Kinase assays were analyzed by SDS-PAGE and autoradiography or subjected to immunoprecipitation.

Densitometry-- Optical densities from linear exposures of autoradiograms and Western blots developed using ECL were measured using the UltraScan XL Laser Densitometer (Amersham Pharmacia Biotech) and the GelScan XL software. Kinase activities were expressed as a function of tyrosine phosphorylation in relation to the total amount of the kinase.

Antibody-mediated Cross-linking-- Oli-neu cells (1-2 × 10⁷) were induced to differentiate by incubation with 1 mM dbcAMP for 3-4 days (34). The cells were washed twice with ice-cold Tris-buffered saline and incubated for 1 h at 4 °C with the monoclonal antibody 27-11-111, which reacts with F3 (32). The cells were washed three times with ice-cold Tris-buffered saline and incubated with rabbit anti-mouse IgG for 30 min at 4 °C. In control samples, cells were incubated with the monoclonal AN2 antibody (31), followed by rabbit anti-rat IgG, or with primary or secondary antibodies alone. Subsequently, the dishes were transferred to 37 °C for 0, 3, 5, and 10 min (controls were left for 5 min at 37 °C). Following incubation at 37 °C, the cells were immediately lysed in 1.5 ml of extraction buffer, and the extracts were subjected to sucrose density gradient centrifugation as described above. DIGs were isolated from sucrose density gradients, subjected to SDS-PAGE, and analyzed by immunoblotting for tyrosine-phosphorylated Fyn and total Fyn protein with specific antibodies.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

GPI-anchored Proteins and Src Family Kinases Colocalize in Detergent-insoluble Glycosphingolipid-rich Microdomains during Development of the Myelin Sheath-- To examine the localization of GPI-anchored proteins in DIGs during myelination, primary oligodendrocytes or myelin prepared from mice at defined developmental stages between postnatal day 9 (P9) and adult were analyzed by detergent extraction and sucrose density gradient centrifugation. Equal amounts of total myelin protein from each age examined were applied to the gradients. Western blot analysis of the gradient fractions with antibodies recognizing NCAM or F3 showed that a fraction of both GPI-anchored NCAM 120 and F3 float at low densities in the gradient (Fig. 1A). Transmembrane isoforms of NCAM (NCAM 140 and 180) localize exclusively at the bottom of the gradient in high density fractions. Lipid analysis showed that the low density fractions are enriched in the typical myelin glycosphingolipids galactocerebroside and sulfatide as well as cholesterol (16). Flotation of GPI-anchored proteins in the gradient shows their specific association with DIGs. DIGs from cultured oligodendrocytes and myelin of P9/10 and P12 mice are distributed in gradient fractions between 18 and 25% sucrose (fractions 4-9). With increasing age of the animals and thus progressing myelination, the myelin DIGs focus at a density of 17-18% sucrose (fractions 8 and 9). The GPI-anchored proteins are enriched in DIGs and in myelin from P45, and they are localized exclusively in DIGs.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Colocalization of GPI-anchored proteins with Src family kinases in detergent-insoluble complexes from oligodendrocytes and myelin during development. Triton X-100 extracts prepared from oligodendrocytes and myelin from mice of different ages were prepared at 4 °C, adjusted to 40% sucrose, and loaded on a 5-30% linear sucrose density gradient. Equal amounts of total protein (300 µg) were loaded on the gradients of the different myelin preparations. Gradient fractions were collected, subjected to SDS-PAGE, and analyzed by Western blot for the distribution of the GPI-anchored proteins NCAM 120 and F3 (A) and for the Src family tyrosine kinases Fyn and Lyn (B). Fraction 1 is of high density from the bottom of the gradient. DIGs from oligodendrocytes and myelin are heterogeneous in density, whereas DIGs from mature myelin are quite homogeneous. The Src family kinase Fyn colocalizes with GPI-anchored proteins in DIGs from oligodendrocytes and myelin of all developmental stages. Lyn kinase is present in DIGs from oligodendrocytes and myelin from P9/10, and P12 mice and is down-regulated in DIGs from myelin from older mice.

Western blot analysis of the gradient fractions with antibodies against the Src family tyrosine kinases Fyn (p59) and Lyn (p53/56) showed that both kinases co-localize with NCAM 120 and F3 in DIG fractions (Fig. 1B). In gradients from cultured primary oligodendrocytes and myelin from P9/10 and P12 animals both kinases were found in bottom gradient fractions as well as in DIG fractions, in a distribution similar to that of the GPI-anchored proteins. The expression of Lyn kinase in myelin declined after P12, whereas Fyn was still expressed in adult myelin. With progressing myelination, Fyn kinase became progressively associated with DIG fractions, and in myelin from P30 up to adult it localized exclusively to DIGs, in a similar fashion to the shift in localization of the GPI-anchored proteins.

Fyn and Lyn Kinase Activity in Oligodendrocyte and Myelin DIGs Peaks at the Onset of Myelination-- Immunoprecipitation of phosphoproteins from the [-³²P]ATP kinase assays of DIGs with antibodies directed against the Src family kinases Fyn and Lyn showed that both participate in the kinase reactions and are phosphorylated. The phosphorylation of both these kinases in DIGs is down-regulated with ongoing myelination (Fig. 2A). The strongest signal for ³²P-phosphorylated Fyn was obtained in precipitates of Fyn from oligodendrocyte DIGs. The ³²P phosphorylation of Fyn in DIGs from myelin continuously declined between P9/10 and P30 and was virtually absent in DIGs from myelin of P45 and adult mice. High amounts of ³²P-phosphorylated Lyn were precipitated from DIGs from oligodendrocytes, whereas only very weak ³²P signals were obtained from DIGs of myelin from P9/10 and P12 mice. To investigate whether the reduction in phosphorylation of the kinases is due to a reduction in expression of the respective proteins, the samples were analyzed by Western blotting. Significant levels of Fyn kinase were present in myelin DIGs throughout development until adult (Fig. 2B). In contrast, expression of Lyn kinase was observed in myelin DIGs up to P12 and was thereafter absent. The phosphorylation of Src family kinases is in most cases due to autophosphorylation and can be taken as a measure of the enzymatic activity (38). We measured the optical densities of the autoradiograms showing the ³²P phosphorylation as well as the optical densities of the signals from the Western blots showing the total amount of the kinases and expressed the activity of each kinase as relative phosphorylation per unit of protein (Fig. 2C). The kinase activity of Fyn is highest in DIGs from oligodendrocytes and in DIGs from myelin of P9/10 mice, at a time point where in vivo myelination is commencing. With ongoing myelination, the Fyn activity localized in DIGs is rapidly down-regulated. Lyn kinase activity is maximal in DIGs from cultured oligodendrocytes and weak in DIGs from myelin of P9/10 and P12 mice.

View larger version (45K): [in this window] [in a new window]   Fig. 2.   Analysis of Fyn and Lyn kinase activity in DIGs from oligodendrocytes and myelin. Isolated DIGs from oligodendrocytes and myelin from mice of different ages were subjected to a [-32P]ATP kinase assay. Phosphorylated Fyn (left) and Lyn (right) were analyzed by immunoprecipitation (A) and compared with total Fyn and Lyn protein in the same samples analyzed by Western blot (B). The relative phosphorylation per unit of protein (relative kinase activity) of Fyn and Lyn was calculated by densitometric quantification of the immunoprecipitation in A and the Western blot in B (C). The kinase activity of Fyn in myelin DIGs is maximal at P9/10. Lyn activity is most prominent in DIGs from cultured oligodendrocytes.

Fyn Co-immunoprecipitates with the GPI-anchored Proteins NCAM 120 and F3 from Oligodendrocyte DIGs-- We next asked whether the co-localization of the GPI-anchored adhesion receptors and the Src kinases on sucrose density gradients reflects an association between these molecules. We isolated DIGs from oligodendrocyte extracts, used them as a source for immunoprecipitation of NCAM or F3, and performed a [-³²P]-ATP in vitro kinase assay on the immunoprecipitate. The bottom fractions of the gradient were also isolated and similarly subjected to immunoprecipitation and kinase assay. Phosphorylated proteins associated with the immunoprecipitate were separated by SDS-PAGE and visualized by autoradiography.

In contrast to the multiple signals seen when the entire DIG proteins were subjected to a kinase assay (data not shown), a dominant phosphorylated protein of 59 kDa was associated with the immunoprecipitated NCAM 120 and F3 in DIGs (Fig. 3A, lane 5, and Fig. 3B, lane 5). In both cases, this 59-kDa phosphoprotein was identified as Fyn kinase by reprecipitation with specific antibodies (Fig. 3, A, lane 6, and B, lane 6). No association with Lyn kinase was observed (data not shown). Additional signals of higher molecular weight were observed in the autoradiograph of the kinase assay on the NCAM 120 immunoprecipitate, but these were weak in comparison with the 59-kDa Fyn signal. In the case of the kinase assay on the F3 immunoprecipitate, additional signals were hardly visible. The precipitation and washing conditions were thus stringent enough (0.1% SDS in the washing buffer) to destroy the structural integrity of DIGs but did not perturb the association between Fyn and NCAM 120 or Fyn and F3. Similar analysis of the bottom fractions from the gradient showed that Fyn kinase activity was lacking in the NCAM or F3 immunoprecipitates (Fig. 3, A, lanes 2 and 3, and B, lanes 2 and 3).

View larger version (33K): [in this window] [in a new window]   Fig. 3.   Co-immunoprecipitation of Fyn kinase with NCAM 120 and F3 from isolated DIGs. Oligodendrocyte DIGs were isolated from sucrose density gradients and subjected to immunoprecipitation using specific antibodies recognizing NCAM (A) or F3 (B). Immunoprecipitates were subjected to a [-32P]ATP in vitro kinase assay and analyzed by SDS-PAGE followed by autoradiographic detection of phosphorylated proteins. Alternatively, the bottom gradient fractions were taken and similarly treated. The 59-kDa phosphoprotein co-precipitating with the GPI-anchored proteins NCAM 120 and F3 in the DIG fractions was identified by reprecipitation with antibodies specific for Fyn kinase (lane 6, A and B). As an alternative to the kinase assays, the same fractions (DIGs and bottom) were collected from gradients of cells that had been incubated with 35S translabel, subjected to immunoprecipitation with polyclonal antibodies against NCAM and F3 and analyzed by SDS-PAGE and phosphoimager detection (lanes 1 and 4, A and B). C, to further demonstrate the presence of Fyn, F3 and NCAM in the bottom fractions, 75-µl aliquots of the bottom fraction of the gradient were subjected to Western blotting with polyclonal antibodies against F3, NCAM, and Fyn. wblot with ab, Western blot with antibody.

The specificity of the first immunoprecipitation was shown by precipitating from ³⁵S-labeled cells. Immunoprecipitation with polyclonal antibodies against NCAM demonstrated the presence of NCAM 140 and some NCAM 120 in bottom gradient fractions (Fig. 3A, lane 1) and exclusively NCAM 120 in DIG fractions (Fig. 3A, lane 4). In the DIG fraction, an additional signal at 59 kDa was observed and represents ³⁵S-labeled Fyn protein. Immunoprecipitation with polyclonal antibodies against F3 yielded a weak signal from bottom fractions of the gradient (Fig. 3B, lane 1), but a strong F3 signal was observed in the DIG fraction (Fig. 3B, lane 4). As in the case for NCAM 120, an additional signal at 59 kDa of ³⁵S-labeled Fyn protein was also seen in the F3 precipitate from the DIG fraction. To ensure that Fyn was indeed present in these bottom gradient fractions, Western blots were performed on the fractions. These results (Fig. 3C, lane 2) confirmed the presence of Fyn in the bottom fractions, as is also shown in Fig. 1. Additionally, Western blots with antibodies against NCAM and F3 demonstrated the presence of NCAM (lane 1) and F3 (lane 3) in these fractions.

This high affinity association of NCAM 120 and F3 with Fyn kinase is thus specific to the DIGs and identifies these adhesion molecules as potential effectors regulating the activity of Fyn in DIGs.

Antibody-mediated Cross-linking of Oligodendroglial F3 Stimulates Fyn Activity in DIGs-- To simulate ligation of F3 by e.g. axonal ligands, we examined the effect of antibody-mediated cross-linking of F3 on living cells on the tyrosine phosphorylation of Fyn in oligodendrocytes. To ensure high yields of material and thus examination of several time points, we used the oligodendroglial cell line Oli-neu, after incubation in the presence of 1 mM dbcAMP to induce differentiation (34). This treatment also up-regulates the DIGs. Differentiated Oli-neu cells were incubated at 4 °C with a monoclonal antibody binding to the Ig domains of the F3 molecule, followed by a species-specific secondary antibody. Subsequently, the cells were transferred to 37 °C and incubated further for different periods of time. Cells were then immediately lysed and subjected to sucrose density gradient centrifugation. DIGs as well as bottom fractions were recovered from the gradient and analyzed by Western blot for the presence of phosphorylation on tyrosine of Fyn as a measure of kinase activation and of total Fyn protein. Tyrosine-phosphorylated Fyn in the DIG fractions increased with time of incubation of the cells at 37 °C, while the total amount of Fyn protein in DIGs remained unchanged (Fig. 4A, right panels a and b). The Fyn activity reached a transient maximum between 3 and 5 min of incubation at 37 °C and then subsequently declined. Interestingly, incubation with the primary antibody alone at 37 °C activated Fyn, while control incubations with secondary antibody alone had no effect on Fyn activity (data not shown). In contrast to the changes in Fyn phosphorylation seen in the DIG fraction, the tyrosine phosphorylation of proteins including Fyn present in bottom gradient fractions is not affected by cross-linking F3 (Fig. 4A, left panels a and b). The densitometric quantification of the increase in tyrosine phosphorylation of Fyn related to the total Fyn protein levels in DIGs showed variations in the degree of stimulation of Fyn activity between different experiments, with up to 8-fold increase of activity. Nevertheless, all experiments showed that ligation of F3 on the oligodendroglial cell surface results in an activation of Fyn kinase strictly localized to DIGs.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Cross-linking F3 molecules up-regulates Fyn activity. A, Oli-neu cells, after differentiation in the presence of 1 mM dbcAMP, were incubated at 4 °C with the monoclonal antibody 27-11-111 recognizing F3 followed by a secondary anti-mouse antibody. The cultures were then incubated at 37 °C for 0, 3, 5, and 10 min and lysed. The DIGs and bottom fractions were isolated from gradients and analyzed by Western blot with antibodies against phosphotyrosine (a), Fyn (b), and F3 (c). The band shown in a is the signal of tyrosine phosphorylated Fyn protein. While the total amount of Fyn protein is unchanged, the tyrosine phosphorylation of Fyn in DIGs increases to a maximum after 5 min at 37 °C. The asterisk marks a signal due to reactivity with the heavy chain of the secondary anti-mouse antibody used in the cross-link. The experimental design is shown schematically. B, as a control, Oli-neu cells, after differentiation in the presence of 1 mM dbcAMP, were incubated at 4 °C with the monoclonal antibody AN2 followed by secondary anti-rat antibody. The cells were lysed immediately or after incubation at 37 °C for 5 min, and the DIG and bottom sucrose density gradient fractions blotted with antibodies to phosphotyrosine (a) or Fyn protein (b). The band shown in a is the signal of tyrosine phosphorylated Fyn protein. C, differentiated Oli-neu cells (as above) were subjected to metabolic radiolabeling and immunoprecipitation with polyclonal AN2 antibody. In lane 1, the immunoprecipitated AN2 at around 330 kDa is seen. Lanes 2 and 3 show Western blots of these immunoprecipitates with monoclonal AN2 antibody (lane 2) and with monoclonal antibody to Fyn (lane 3). The total lysate of the differentiated Oli-neu cells prior to immunoprecipitation was blotted with monoclonal antibody to Fyn (lane 4). wblot with ab, Western blot with antibody.

As an additional control, monoclonal antibody to the 330-kDa cell surface glycoprotein AN2, which is expressed by these cells (31), was used in similar cross-linking experiments (Fig. 4B). In contrast to the results seen with antibody to F3, no stimulation of Fyn kinase activity was seen in the DIG fraction or the bottom fractions of the gradient. Immunoprecipitation experiments with AN2 antibody from ³⁵S-labeled cells further show that Fyn and AN2 do not associate (Ref. 31; Fig. 4C); no Fyn protein is seen in the immunoprecipitate of AN2 from differentiated Oli-neu cells. These experiments demonstrate that the activation of Fyn in DIGs by cross-linking F3 is specific and not due to cross-linking just any cell surface protein.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Glycosphingolipid-rich Membrane Microdomains as a Signaling Platform in Oligodendrocytes-- We have shown that in DIGs from both oligodendrocytes and myelin the Src family kinases Fyn and Lyn colocalize with the GPI-anchored proteins NCAM 120 and F3, suggesting their association with raftlike microdomains during development of the myelin sheath.

DIGs containing Src kinases were first isolated from lymphocytes by Cinek and Horejsi (39) and were subsequently isolated from many cell types (25, 40). It was postulated that these detergent-isolated complexes correspond to functional microdomains or rafts in the plasma membrane important for the compartmentation of distinct cellular functions including signal transduction pathways (17, 18, 41-47). The existence in vivo of raft-like microdomains has been a subject of discussion (48). However, recent studies have supported the existence of small domains in living cells (49-51). The detailed cellular functions and in particular the downstream targets of these signaling domains remain unelucidated in most cell types. Recent papers have shown that for efficient T-cell activation, a striking compartmentation in DIGs of activated T-cell receptor and signal-transducing molecules is essential, stressing the role of rafts as functional signaling domains in lymphocytes (46, 47).

A Role of Rafts in Fyn and Lyn Signaling during the Initiation Phase of Myelination-- We have shown that the total kinase activity in DIGs as measured by a kinase assay, as well as the specific kinase activity of Fyn and Lyn is developmentally regulated parallel to the in vivo myelination process. The clear peak of Fyn and Lyn activity in DIGs isolated from oligodendrocytes and myelin of young postnatal animals (P9/10 mice) suggests its role in the initiation of the myelination program.

The expression of Fyn in oligodendrocytes and myelin and its early postnatal activity have been described before (15, 52). Fyn knockout mice are hypomyelinated; the myelin content normalized for brain weight is 40-50% reduced compared with wild-type mice. Ligation of MAG, a cell adhesion molecule thought to play a role in early axon-glia recognition (6), was reported to stimulate Fyn activity and was a substrate for Fyn phosphorylation (15). Since MAG is not present in oligodendroglial DIGs,² it can be excluded as a substrate for Fyn activity in DIGs. Our results thus argue for the existence of an additional MAG-independent signaling pathway of Fyn using raftlike microdomains as signaling compartments.

The Stable Association of F3 and NCAM 120 with Fyn in Oligodendrocytes Is Exclusive to Rafts-- We find Fyn kinase activity associated with both NCAM 120 and F3 in immunoprecipitates from isolated DIGs. The complex is stable in SDS-containing buffer under stringent conditions that destroy the DIG complexes, stressing the affinity of the interactions between these molecules. F3, NCAM 120, and Fyn also are present outside the DIGs. However, Fyn is not associated with immunoprecipitates of F3 or NCAM from non-DIG fractions of oligodendrocytes (e.g. immunoprecipitates from bottom gradient fractions or DIG-free precursor cells). The stable association between these molecules is thus exclusive to DIGs. Staining of cultured oligodendrocytes shows co-localization of a fraction of F3 and Fyn with galactocerebroside (29),³ the main DIG lipid (16).

An interaction between NCAM and Fyn has been reported recently in growth cones of neurons, but Fyn was associated exclusively with the transmembrane isoform NCAM 140 and not with the GPI-anchored isoform NCAM 120 (53). The interaction between F3 or F11 (the chicken homolog) and Fyn has also been described for mouse cerebellar tissue and chick embryonal brain cells, respectively (54, 55).

Local Fyn Signaling in Rafts Is Evoked by Ligation of Oligodendroglial F3-- Ligation of cell surface receptors including GPI-anchored proteins with specific antibodies has been used to mimic the interaction with unknown ligands and to induce cellular responses, especially in lymphocytes (24, 39). Ligation of F3 by antibody-mediated cross-linking in differentiated Oli-neu cells induces a transient increase of Fyn activity (up to 8.5-fold), which is strictly localized to DIGs. The activation is rapid, and the Fyn kinase activity is rapidly down-regulated. Other proteins phosphorylated on tyrosine are also found in the DIG and the bottom fractions of the gradient; however, a change in their phosphorylation after cross-linking F3 was not detected in this assay. A stimulation of kinase activity after antibody-mediated cross-linking of F3 was shown in two other studies using primary chicken neuroblasts (54) and transfected cells (55). In these studies, an in vitro kinase assay was performed on immunoprecipitates. In contrast, we show the in situ stimulation of Fyn kinase in intact oligodendroglial cells after ligation of cell surface F3. We also show that this transmembrane signaling event is localized to a specified membrane subdomain, the raft-like microdomains.

Mechanisms of Kinase Activation and Axonal Ligands-- It is still an open question how exoplasmic GPI-anchored proteins interact with cytoplasmic membrane-associated Src kinases, leading to signal transmission across the membrane. Harder et al. (18, 51) suggested that the clustering of GPI-anchored proteins by ligands may aggregate individual small rafts into larger domains, leading to a focal concentration of kinase activity at the cytoplasmic face of the membrane. Alternatively, a transmembrane protein may link F3 and Fyn. Such a protein could be developmentally regulated and may also regulate the activity of the kinase (24, 25). The stability of the F3 and NCAM 120 associations with Fyn in our experiments argues for the presence of a transmembrane linker protein. Contactin-associated protein (Caspr/paranodin; Refs. 57 and 58), a member of the neurexin superfamily, has been described as a cis interaction partner of F3 in neurons (59). In oligodendrocytes that do not appear to express Caspr (30), other candidates may link F3 and Fyn. These may include receptor protein tyrosine phosphatase  and members of the L1 family, some of which are expressed by oligodendrocytes (60, 61). Axonal interaction partners of NCAM 120 and F3 serving as putative ligands to induce the Fyn signaling cascade are the axonally expressed adhesion molecules NCAM itself (homophilic interaction) and L1 (heterophilic interaction with F3). L1 has been found in a complex together with F3 and Fyn in isolates from cerebellar tissue (62). Furthermore, F3-positive Oli-neu cells adhere to an L1-positive substrate, and this adhesion can be substantially reduced by the addition of antibodies to F3 (data not shown).

Distinct Signaling Pathways Associate with NCAM 120 and F3 during Oligodendrocyte Differentiation-- In oligodendrocyte precursor cells, DIGs and consequently the F3-Fyn complex are lacking due to the low expression of glycosphingolipids (see as a model Fig. 5). With maturation to myelination-competent oligodendrocytes and the accompanying up-regulation of glycosphingolipids, the F3-Fyn complex forms as DIGs are up-regulated. A similar maturation-dependent signaling-concept has been suggested for the axonal adhesion molecule axonin-1, which is also a GPI-anchored protein (63). Axonin-1 associates with Fyn kinase in neurons of the chick dorsal root ganglion, while neurites are growing. During fasciculation of the neurites, which is mediated by interactions of axonin-1 and NgCAM, axonin-Fyn complexes dissociate, and axonin predominantly associates with NgCAM in large macromolecular complexes containing casein kinase activity. Our study thus underscores the premise that the coordinated association of cell surface receptors with distinctive signaling molecules in specialized membrane compartments allows these molecules to adjust their function to changing requirements during development. Coupling F3 and NCAM 120 to the Fyn signaling complex may initiate progression of a premyelinating oligodendrocyte to a myelinating oligodendrocyte. The lateral mobility of GPI-anchored proteins in the membrane may support their flexibility in interacting with different cis-binding partners. Targets of activated Fyn could ultimately include cytoskeletal proteins, since wrapping of the oligodendrocyte processes around the axon must involve dramatic changes in the cell cytoskeleton (64). In support of this concept, a recent publication (52) has shown that inhibiting Fyn prevents process outgrowth by oligodendrocytes.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Model of NCAM 120 and F3 signaling complexes during oligodendrocyte development. 1, oligodendrocyte precursor cells express F3, NCAM 120 (and NCAM 140) and Fyn, but these molecules are not stably associated. 2, in maturing oligodendrocytes, NCAM 120 and F3 are associated with Fyn in the DIG fraction, which can be isolated from these cells as oligodendrocyte glycosphingolipids (GSL) are up-regulated. NCAM 140 is not included in these complexes. The ligation of F3 (and possibly NCAM 120) in these complexes by an axonal ligand activates oligodendroglial Fyn. This is reflected in an increased autophosphorylation of Fyn. 3, in mature oligodendrocytes and myelin, the kinase activity of Fyn is reduced.

Our findings yield new insights into the molecular basis of oligodendrocyte-axon interaction and subsequent glial signaling transduction cascades that may instruct the oligodendroglial myelination program.

	ACKNOWLEDGEMENTS

We thank Doris Kendel and Iris Bünzli-Ehret for excellent technical help. Fritz Rathjen is thanked for the valuable gift of the 27-11-111 anti-F3/F11 monoclonal antibody. Our colleagues, especially Kai Simons, Klaus-Armin Nave, Nicole Schaeren-Wiemers, Christoph Thiele, Jeremy Garwood, and Bruno Kyewski, are thanked for critically reading and commenting on the manuscript. J. T. thanks Wieland Huttner for stimulating discussions and support.

	FOOTNOTES

^* This work was supported by Deutsche Forschungsgemeinschaft Grant SFB 317 (to J. T.) and Graduiertenkolleg "Molekulare und Zelluläre Neurobiologie" Stipendia (to C. K. and T. K.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Present address: Proctor and Gamble European Service GmbH, Industriestr. 30-34, 65760 Eschborn, Germany.

^§ To whom correspondence should be addressed. Tel.: 49-6221-548319; Fax: 49-6221-548301; E-mail: jtrotter@sun0.urz.uni-heidelberg.de.

² E.-M. Krämer and J. Trotter, unpublished results.

³ E.-M. Krämer and J. Trotter, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: NCAM, neural cell adhesion molecule; DIGs, detergent-insoluble glycosphingolipid-rich microdomains; GPI, glycosylphosphatidylinositol; MAG, myelin-associated glycoprotein; Pn, postnatal day n; dbcAMP, dibutyryl cyclic AMP; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES