Jniirnal o/ Neurochemrslry Raven Press, Ltd., New York 0 1990 International Society for Neurochemistry
Differential Phosphorylation of Myelin-Associated Glycoprotein Isoforms in Cell Culture Daniel E. H. Afar, *James L. Salzer, ?John Roder, $Peter E. Braun, and John C. Bell Departments of Biochemistry and Medicine, University of Ottawa, Ottawa, and ?Department of Medical Genetics, University of Toronto, Toronto, Ontario, and $Department of Biochemistry, McGill University, Montreal, Quebec, Canada; and *Department of Cell Biology, New York University Medical Center, New York, New York, U.S.A.
Abstract: The alternative splicing of myelin-associated glycoprotein (MAG) mRNA generates two isoforms that harbor distinct potential phosphorylation sites in their cytoplasmic tails. Here we characterize the in vivo phosphorylation of MAG isoforms in NIH 3T3 cells transfected with the cDNAs encoding the two isoforms of MAG. Our results demonstrate that the longer isoform, L-MAG, is phosphorylated constitutively mainly on serine, but also on threonine and tyrosine residues. This phosphorylation is subject to change by 12-0tetradecanoylphorbol 13-acetate (TPA) and ammonium vanadate, but not by dibutyryl-cyclic AMP. The shorter isoform, S-MAG, is constitutivelyphosphorylated only on serine residues. While TPA and dibutyryl-cyclic AMP have no de-
tectable effect, ammonium vanadate induces tyrosine and threonine phosphorylation in S-MAG. 32Plabeling of v-srctransformed NIH 3T3 cells that express L-MAG also show that L-MAG is likely to be an in vivo substrate for pp60” tyrosine kinase activity. These results demonstrate that both MAG isoforms are phosphorylated in a heterologous cell system and that this phosphorylation is subject to pharmacological manipulation. Key Words: Myelin-associated glycoprotein-Tyrosine phosphorylation-Protein kinase C-Cell adhesion. Afar D. E. H. et al. Differential phosphorylation of myelin-associated glycoprotein isoforms in cell culture. J. Neurochem. 55, 14 18- 1426 ( 1990).
The process of selective cell adhesion is essential for the organization of tissues in higher organisms during development. The characterization of molecules involved in cell adhesion is therefore basic to the understanding of developmental events. Certain types of cell surface glycoproteins have been demonstrated to play a fundamental role in this process. One group of cell adhesion molecules has been included in the immunoglobulin supergene family by virtue of their sequence and structural features (Williams, 1987). Among these molecules is myelin-associated glycoprotein (MAG), a transmembrane glycoprotein involved in neuron-glia and glia-glia interactions in the developing nervous system (Poltorak et al., 1987). MAG is expressed on the surface of oligodendrocytes in the CNS and Schwann cells in the PNS (Sternberger et al., 1979). MAG mRNA is alternatively spliced into two developmentally regulated isoforms, suggesting that the
larger (L-MAG) and the smaller (S-MAG) isoforms perform functionally different roles in myelin formation (Frail and Braun, 1984; Tropak et al., 1988). The early appearance of L-MAG at the onset of the myelin program suggests a potential role for the cytoplasmic domain of this isoform in the initial processes of intercellular recognition and signal transduction. S-MAG mRNA, which is maximally expressed in mature myelin, may play a structural role in maintaining the integrity of the myelin sheath. L-MAG and S-MAG exhibit identical external and transmembrane domains, but differ in their cytoplasmic carboxyl termini ( h i et al., 1987; Salzer et al., 1987). Both isoforms contain several potential phosphorylation sites in the carboxylterminal region, among which is a region identified as a potential protein kinase C (PKC) phosphorylation site by Arquint et al. (1987). In addition, L-MAG contains a segment potentially phosphorylated by protein
Received December 2 I , 1989; revised manuscript received March 20, 1990; accepted March 22, 1990. Address correspondence and reprint requests to Dr. J. C. Bell at Department of Biochemistry, University of Ottawa, 45 1 Smyth Rd., Ottawa, Ontario, Canada KIH 8M5. Abbreviations used: CAMP, cyclic AMP; dbcAMP, dibutyrylCAMP;EGF, epidermal growth factor; IGF-I, insulin-like growth
factor 1; LM and SM cells, L-MAG- and S-MAG-expressing fibroblasts; L-MAG and S-MAG, larger and smaller isoforms of MAG, MAG, myelin-associated glycoprotein; MEM, minimum essential medium; PAGE, polyacrylamide gel electrophoresis; PKC, protein kinase C; PVsrcLM cells, L-MAG-expressing v-src-transformed fibroblasts; SDS, sodium dodecyl sulfate; TPA, 12-0-tetradecanoylphorbol 13-acetate.
1418
DIFFERENTIAL PHOSPHORYLA TION OF MAG ISOFORMS kinase A (Arquint et al., 1987) and PKC (Salzer et al., 1987),as well as a region homologous to the epidermal growth factor (EGF) receptor autophosphorylation site (Arquint et al., 1987; Salzer et al., 1987). We have previously detected phosphorylation of L-MAG in myeh a t i n g rodent brain in vivo as well as in myelin in vitro on serine, threonine, and tyrosine residues (Edwards et al., 1988). Phosphorylation of S-MAG was never detected in myelinating brain. Here we report the isolation and characterization of fibroblast cell lines expressing the MAG cDNAs. Our results demonstrate that L-MAG is phosphorylated on serine, threonine, and tyrosine residues in cell culture, whereas S-MAG is constitutively phosphorylated only on serine residues. While growth factors are unable to modulate MAG phosphorylation, the v-src kinase constitutively phosphorylates L-MAG on tyrosine residues. The results presented here are consistent with phosphorylation of the cytoplasmic domains of both molecules playing a role in the regulation of MAG activity in developing and mature myelin.
MATERIALS AND METHODS Cell lines LMAG cDNA was inserted into the retroviral vector pmv7 (Kirschmeieret al., 1988) and transfected into NIH 3T3 virus packaging cell lines (PA317 and Psiz) by calcium phosphate coprecipitation (Graham and van der Eb, 1973). Stable transfectants of L-MAG-expressing cells were selected by G4 18 resistance [600 pgml in a-minimum essential medium (aMEM), 10%fetal calf serum]. The transfectants, referred to as LM cells, were screened for MAG expression using immunofluorescent cytochemistryon live cells and immunoblot analysis of membrane extracts. The generation of stable S-MAG-expressing cell lines has previously been described (Johnson et al., 1989). The S-MAGexpressing clone used in our study (3TV2-9)will be referred to as “SM cells.” The v-src-transformed NIH 3T3 cell line, PVsrc, was infected by cocultivation with virus from the Psizcells expressing L-MAG. Psi2-L-MAG cells were treated with mitomycin C (200 pg/ml) for 2 h and then cocultured with PVsrc cells in a ratio of 1:2 for 24 h. After 24 hours stable infectants were selected by G4 18 resistance (600 pglml in aMEM, 10% fetal calf serum) and are referred to as PVsrcLM cells.
Phosphorylation of MAG LM and SM cells were labeled with 0.5 mCi [32P]orthophosphate/mlphosphate-free Dulbecco’s Modified Eagle Medium and 10%fetal calf serum for 4 h in the presence or absence of ammonium vanadate. The cells were harvested by lysis in 250 pl of hot ( 100°C) 2% sodium dodecyl sulfate (SDS). The cell lysate was passaged several times through a 25-gauge syringe needle and diluted to 1 ml with an immunoprecipitation buffer (10 mM Tris-HC1, pH 7.4, 5 mM EDTA, 150 mMNaCl,l%Triton, 500 mMNH,VO,, 2 mM NaF, 2 mM Na-pyrophosphate, 2 pg/ml aprotinin, 5 pglml leupeptin, 0.4 p&ml pepstatin, 200 pdml polymethylsulfonyl fluoride). Undissolved cellular debris was removed by centrifugation for 10 min at 14,000 rpm in a microfuge. Immunoprecipitation of MAG was performed by overnight incubation of the cell lysate with a mixture of anti-MAG
1419
monoclonal antibodies Gen 3 SI and Gen 3 S3 (generously provided by N. Latov). The following day the mixture was incubated with rabbit anti-mouse antibody (Jackson Immunoresearch Lab.) for 1 h prior to addition of protein Acoupled Sepharose beads (Pharmacia). After an additional hour of incubation, the beads were washed five times with the immunoprecipitation buffer and MAG was eluted using SDS sample buffer. The immunoprecipitated samples were resolved by SDS polyacrylamide gel electrophoresis (PAGE; 10%polyacrylamide) and detected by autoradiography.
Stoichiometry of L-MAG phosphorylation To determine the half-life of L-MAG in LM cells, LM cells were labeled for 18 h at 37OC with [35S]methionine(50 pCi/ ml; Amersham) in methionine-free minimal essential medium (MEM), supplemented with 10%fetal calf serum and 5% aMEM. Then the cells were washed two times with phosphate-buffered saline and were either harvested in 2% SDS or were incubated for an additional 4 and 9 h in aMEM. LMAG was immunoprecipitated as described above and resolved on SDS-PAGE. The half-life was calculated by extracting radioactive L-MAG out of excised gel slices and determining the amount of radioactivity in the extracts using liquid scintillation counting. To assess the extent of L-MAG phosphorylation, LM cells were labeled for 18 h at 37°C with either [32P]orthophosphate (200 pCi/ml) or [35S]methionine(50 pCi/ml) in aMEM containing 10%fetal calf serum. L-MAG was immunoprecipitated and resolved on SDS-PAGE. The number of moles of phosphate in L-MAG was determined by scintillation counting of the excised L-MAG bands. The number of moles of immunoprecipitated LMAG was estimated from the level of [35S]methionineincorporated, assuming that L-MAG contains nine methionine residues (Salzer et al., 1987). The specific activity of label in the medium was determined by scintillation counting of an aliquot of the medium and from the known concentrations of phosphate (10 m M in aMEM and 1.5 mM in fetal calf serum) and methionine (0.2 p M in aMEM and 0.03 m M in fetal calf serum).
Treatment of L-MAG-expressing cells with growth factors, dibutyryl-cyclic AMP, and phorbol esters To assess the effects of growth factor treatment on L-MAG phosphorylation, LM cells were fed aMEM without serum for 24 h. The next day the cells were 32P-labeledin phosphatefree medium without serum for 4 h. EGF (1 50 nglml; Boehringer Mannheim), insulin-like growth factor 1 (IGF- 1; Boehringer Mannheim), or insulin (0.5 pM, generously provided by Dr. M. Bernier) was added to the cells at the end ofthe labeling period for 10 min at 37°C. The effects of cyclic AMP (CAMP)-mediated phosphorylation of L-MAG and SMAG were assessed by incubating LM and SM cells with either 5 mM dibutyryl-CAMP (dbcAMP Sigma) or sodium butyrate (5 mM) during the labeling procedure. Sodium butyrate was used to control for non-CAMP-related effects of dbcAMP. To determine the effects of phorbol esters on MAG phosphorylation, LM and SM cells were treated with 100 nM 4a-phorbol or 100 nM 12-0-tetradmnoylphorbol 13-acetate (TPA) for 10 min at 37°C at the end of “P labeling of cells. At the end of the labeling period, the cells were harvested in hot 2% SDS and processed as described before.
Phosphoamino acid analysis Phosphorylated MAG isoforms were immunoprecipitated as described above, excised out of the polyacrylamide gel, J. Neurochem., Vol. 55, No. 4, 1990
1420
D. E. H. AFAR ET AL.
FIG. 1. lmmunofluorescent cytochemistry on MAGexpressing fibroblasts. Live LM (A), SM (B),and PVsrcLM (C) cells were stained with monoclonalanti-MAG antibody (513) and subsequently with fluoresceinoonjugatedgoat antimouse secondary antibody. No fluorescence was observed with cells stained with secondary antibody alone. The exposure times and magnification are the same for all pictures. Bar, 20 Pm.
J. Nerirochem.. VoI 55. No. 4, 1990
DIFFERENTIAL PHOSPHOR YLATION OF MAG ISOFORMS electroeluted using a BioRad electroeluter, and precipitated by the addition of trichloroacetate (20%). The trichloroacetate precipitates were washed twice in ice-cold acetone and hydrolyzed with 6M double-distilled HCl for 75 min at 1 10°C. The hydrolysate was lyophilized, mixed with authentic phosphoamino acids, spotted onto MN cell 400 thin-layer cellulose plates, and resolved by two-dimensional electrophoresis as described by Cooper et al. (1 983). Phosphoamino acid standards were detected by ninhydrin staining of the thin-layer plates. Radioactive phosphoamino acids were detected by autoradiography .
Tryptic digestion of MAG Phosphorylated MAG was analyzed by digestion with trypsin essentially as described previously (Edwards et al., 1989) with the modification that tryptic fragments were spotted onto MN cell 400 thin-layer cellulose plates for separation by ascending chromatography. The chromatograms were dried and rotated 90", and thin-layer electrophoresis was performed at pH 1.9 for 1 h at 1,000V. Phosphotryptic fragments were visualized by autoradiography.
Immunoblot analysis Cell membranes were prepared from LM and SM cells by harvesting the cells in 10 ml of ice-cold hypotonic HEPES buffer (20 mM HEPES, pH 7.2, 3 mM KCI, 1 mM EDTA, 500 p M NH4V04, 2 mM NaF, 2 pg/ml aprotinin, 5 pg/ml leupeptin, 0.4 p d m l pepstatin, 200 pg/ml polymethylsulfonyl fluoride). The cells were then lysed in a tight-fitting glass homogenizer using 15 strokes on ice. DNA and cellular debris were removed by a 1,000 g centrifugation for 10 min. A membrane-enriched fraction was obtained by subjecting the supernatant to centrifugation for 30 min at 100,000 g. The membrane pellet was resuspended in the hypotonic HEPES buffer with 40% glycerol. Protein content was determined using the BioRad reagent. Twenty micrograms of membrane protein was dissolved in sample buffer, while 150 p g of membrane protein was precipitated with 12%trichloroacetate and chemically deglycosylated using trifluoromethanesulfonic acid (Aldrich) essentially as described before (Horvath et al., 1989). The proteins were resolved on SDS-PAGE (10%polyacrylamide), transferred to nitrocellulose paper (Schleicher and Schuell), and probed with a polyclonal antibody directed against MAG or an anti-phosphotyrosine monoclonal antibody (PY20; ICN). Visualization ofthe blots was carried out using either an alkaline phosphatase-coupled goat anti-rabbit second antibody (for MAG) or a '251-labeledsheep anti-mouse second antibody (for phosphotyrosine).
1421
RESULTS Characterization of MAG expression in fibroblasts Live MAG-expressing cells were examined by immunofluorescent cytochemistry using a monoclonal anti-MAG antibody (5 13). The results indicate that LMAG and S-MAG are both expressed uniformly on the cell surface (Fig. 1). Immunoblot analysis of membrane extracts of these cells confirms that the antibody recognizes L-MAG and S-MAG in LM and SM cells, respectively (Fig. 2, lanes a and b). It should be noted that glycosylated S-MAG exhibits a slightly lower molecular weight than glycosylated L-MAG. Phosphorylation of MAG MAG phosphorylation was assessed by labeling the cells in phosphate-free medium containing [32P]orthophosphateand immunoprecipitating MAG from a cell extract with monoclonal anti-MAG antibodies (Gen 3 S1 and Gen 3 S3). The immunoprecipitations were performed on cell lysates from equal numbers of cells, and the results demonstrate that LMAG and S-MAG are both phosphorylated (Fig. 2). Using differential labeling with [35S]methionineand [32P]orthophosphate, L-MAG protein was shown to contain 10 mol phosphate/mol polypeptide. In parallel experiments we determined that the half-life of MAG is -4 h in NIH 3T3 cells. For phosphate stoichiometry determinations, cells were incubated overnight ( 18 h) with [32P]orthophosphate and [35S]methionine and therefore contain MAG molecules that have been uniformly labeled with both isotopes. Thus, our calculation of the extent of phosphorylation will not be biased by long-lived MAG molecules that do not contain
a b
c d e f
200-
9768 -
Immunofluorescence cytochemistry Cells were plated onto coverslips and incubated with antiMAG monoclonal antibody 5 13 ( 1 :1,000 in phosphate-buffered saline and 2% calf serum; generously provided by M. Schachner) for 1 h at 4°C. The coverslips were then washed three times in phosphate-buffered saline and 2% calf serum and subsequently incubated with a fluorescein-conjugated goat anti-mouse second antibody (Bio Can) for 1 h at 4°C. After three washes in phosphate-buffered saline and 2% calf serum, the cells were fixed in 4% formaldehyde for 20 min and then mounted onto slides. Visualization of the immunofluorescence was achieved using epifluorescent microscopy.
Phosphorylation of MAG isoforms in LM and S M cells. lmmunoblot analysis of membrane extracts from LM (lane a) and S M (lane b) cells was probed with a mixture of monoclonal antiMAG antibodies Gen 3 S1 and Gen 3 S3 and detected using an 'Z51-labeledsheep anti-mouse antibody. Cell extracts from 3'P-labeled LM (lane c), S M (lane d), and NIH 3T3 (lane e) cells were immunoprecipitated using anti-MAG antibodies as described in Materials and Methods. A sample containing a mixture of LM and S M cell extract was immunoprecipitated with nonimmune serum (lanef). The samples were analyzed using SDS-PAGE and detected by autoradiography FIG. 2.
J Neurocliem.. Vol. 55. No 4. 1990
D. E. H . AFAR ET AL
1422
[35S]methioninebut are posttranslationally modified with 32P.Furthermore, tryptic mapping of 32P-labeled MAG molecules revealed that there are at least seven tryptic phosphopeptides present in L-MAG (see Fig. 6). Taken together these results indicate that L-MAG is in fact a formidable protein kinase substrate. Although SM cells express high levels of S-MAG on the cell surface, the stoichiometry of S-MAG phosphorylation was determined to be approximately one-tenth the amount of L-MAG phosphorylation (Fig. 2). Phosphoamino acid analysis of L-MAG phosphorylated in metabolically labeled fibroblasts shows that it is constitutively phosphorylated mainly on serine, but also on threonine and tyrosine residues (Fig. 3B), thereby mimicking the phosphorylation pattern in brain. Ammonium vanadate, a phosphotyrosine phosphatase inhibitor, increases phosphotyrosine levels in cells (Klarlund, 1985). Although no quantitative change occurred in phosphorylated L-MAG protein from vanadate-treated cells (cf. lanes a and b, Fig. 3A), there was a dose-dependent increase in the amount of phosphotyrosine present in the protein (Fig. 4). Phosphothreonine and phosphoserine levels increased with 50 p M ammonium vanadate, but decreased with higher concentrations (Fig. 4). While ammonium vanadate did not change the amount of L-MAG in the mem-
A
B a
b
c
200-
a
97-
a
C
b
d
e
FIG. 4. Effect of increasing doses of ammonium vanadate on LMAG phosphorylation. LM cells were labeled with 32Pin the presence of increasing concentrations of ammonium vanadate. Phospho-amino acid analysis was performed on immunoprecipitated phosphorylated L-MAG protein using two-dimensional thin-layer electrophoresis as described in Materials and Methods. Samples are shown in order of increasing concentration of ammonium vanadate: 10 (a), 50 (b), 100 (c), 200 (d),and 500 (e)p M . The positions of phosphoserine(S),phosphothreonine(T), and phosphotyrosine (Y) are shown for reference in (b).
b
C
FIG. 3. The effect of vanadate and v-sfc transformation on phosphorylation of L-MAG in vivo. A Phosphorylated L-MAG was immunoprecipitatedfrom LM cells 32P-labeledin the absence (lane a) or presence (lane b) of 500 pM ammonium vanadate and from 32P-labeledPVsrcLM cells (lane c). The samples were resolved on SDS-PAGE and detected by autoradiography. 0: Phosphoamino acid analysis on electroeluted L-MAG protein was performed using two-dimensional thin-layerelectrophoresisas described in Materials and Methods: L-MAG from untreated LM cells (a); L-MAG from vanadate-treatedLM cells (b); L-MAG from PVsrcLM cells (c). The positions of phosphoserine (S), phosphothreonine (T), and phosphotyrosine (Y) are indicated for reference in (a).
J Nrunx.hi.m, L’ol 55. No 4, 1990
brane of these cells as detected by immunoblot analysis (Fig. 5A) and immunofluorescent cytochemistry (data not shown), it significantlyelevated its phosphotyrosine content (Fig. 5B). In fact, L-MAG was determined to be the major tyrosine phosphorylated protein in these cells. Control experiments determined that the faintappearing protein bands on the immunoblots are nonspecific proteins reacting with the secondary antibody (data not shown).
Effect of pharmacological agents on MAG phosphorylation The L-MAG cDNA sequence revealed the presence of potential phosphorylation sites for PKC, CAMP-dependent kinase, and an EGF receptor autophosphorylation site (Arquint et al., 1987; Salzer et al., 1987), while S-MAG exhibited only a possible PKC phosphorylation site (Arquint et al., 1987). We examined the effects of various kinase activators on L-MAG and
DZFFERENTZAL PHOSPHOR YLATION OF MAG ISOFORMS
B
A
abed 200
-
07-
a b c d
200-
07-
6868-
43-
43-
1423
ylated L-MAG protein from TPA- and dbcAMP-stimulated cells showed no qualitative change in the ratio of phosphoserine to phosphothreonine to phosphotyrosine (data not shown). Phosphoamino acid analysis of phosphorylated SMAG demonstrated that constitutive phosphorylation occurred only on serine residues. While TPA and dbcAMP had no detectable effect on S-MAG phosphorylation as determined by SDS-PAGE (Fig. 7A) and tryptic digest (data not shown) analysis, vanadate treatment induced low but detectable levels of phosphotyrosine and phosphothreonine in this isoform (Fig. 7B).
L-MAG phosphorylation in PVsrcLM cells Previously we demonstrated that partially purified MAG can be phosphorylated by pp 130gag-fpsand pp60"~"cin vitro, two cytoplasmic tyrosine kinases en-
FIG. 5. L-MAG is the major tyrosine phosphorylated membrane protein in LM cells. LM cells were left untreatedor were pretreated with 500 p.M vanadate for 3 h. Membrane extracts were prepared and were either left untreated or were subjected to chemical d e glycosylationwith trifluoromethanesulfonate as described in Materials and Methods. lmmunoblot analysis was performed using a polyclonalanti-MAG antibody (A) or using the PY20 anti-phosphotyrosine antibody (B). Samples were as follows: LM membranes from cells untreated(lane a) and deglycosylated (lane b), LM membranes from cells vanadate-pretreated(lane c) and deglycosylated (lane d).
S-MAG phosphorylation. While EGF, IGF- 1, and insulin had no detectable effect on the phosphorylation of L-MAG (data not shown), TPA, a known activator of the serine-threonine-specificPKC (Nishizuka, 1984), stimulated L-MAG phosphorylation as determined by tryptic digestion of L-MAG (Fig. 6d). The phosphopeptide in L-MAG, where the major stimulation of phosphorylation with TPA occurs, is also stimulated with vanadate treatment (Fig. 6b and d, arrows). The inactive phorbol ester 4a-phorbol served as a control for TPA stimulation and did not have an effect on constitutive L-MAG phosphorylation (Fig. 6, cf. a and c). Treatment of LM cells with dbcAMP, an activator of a different class of serine-threonine kinase, quantitatively increased the phosphorylation of L-MAG in some experiments (data not shown). However, dbcAMP did not alter L-MAG phosphorylation significantly in a qualitative sense (Fig. 6f). In this case sodium butyrate was used to control for non-CAMPrelated effects of dbcAMP and no effect was observed (Fig. 6e). Phosphoamino acid analysis of phosphor-
TLC)
TLE
FIG. 8. Tryptic digestion of L-MAG phosphorylated in LM cells. LM cells were 32P-labeledand were left untreated(a) or were treated with 500 p M ammonium vanadate (b), 4a-phorbol (c), TPA (d), sodium butyrate (e), or dbcAMP (f) as described in Materials and
Methods. Phosphorylated1-MAG was purifiedand digested with trypsin. The tryptic fragments were resolved by thin-layer chromatography (TLC) and electrophoresis (TLE) as described in Materials and Methods. The tryptic fragments were detected by autoradiography. Phosphopeptidesthat were stimulated during TPA and vanadate treatment are identified with arrows.
J. Neurachem.. Val. 55. No. 4, 1990
D.E. H . AFAR ET AL.
1424
A a b c d e 200
f
-
68 97
FIG. 7. The effect of various pharmacologicalagents on S-MAG phosphorylation. A Phosphorylated S-MAG was immunoprecipitated from 32P-labeledSM cells, resolved on SDS-PAGE, and d e tected by autoradqraphy. SM cells were either left untreated (lane a) or labeledin the presence of 500 p M ammonium vanadate (lane b), 4a-phOrbd (lane c), TPA (lane d). sodium butyrate (lane e), or dbcAMP (lane f). The radioactive material at the top of the gel in each lane is labeled nucleic acid that associates nonspecifically to protein A-Sepharose. The amount of this material varies randomly between samples in different experiments.B Phosphoamino acid analysison phosphorylatedS-MAG immunopurifiedfrom vanadate treated SM cells. S , phosphoserine; T. phosphothreonine; Y, phosphotyrosine.
coded by the v-ks and v-src oncogenes, respectively (Edwards et al., 1988). To investigate whether L-MAG can be a substrate for pp60"-srcin vivo, the L-MAG cDNA was introduced into v-src-transformed NIH 3T3 cells by retroviral infection. In PVsrcLM cells the cell surface distribution of L-MAG appears uniform (Fig. 1). Thus, v-src transformation does not appear to affect the transport of MAG to the cell surface, or its distribution on the cell membrane. Phosphoamino acid analysis of L-MAG protein immunoprecipitatedfrom "P-labeled PVsrcLM cells demonstrates that L-MAG is constitutively phosphorylated mainly on serine and tyrosine residues, with a small amount of phospho-
J. Neurochem.. Vol. 55, No. 4, 1990
threonine being present (Fig. 3). The high level of phosphotyrosine present in GMAG isolated from these cells is not due to pp60'-s" activity during the isolation procedure, since the cells were lysed in hot 2% SDS. Labeling PVsrcLM cells in the presence of ammonium vanadate did not change the constitutive phosphorylation of L-MAG (data not shown).
DISCUSSION In this study we demonstrate that both L-MAG and S-MAG are phosphorylated in a heterologous cell culture system. Our present results are obtained from cells expressing high amounts of MAG (Fig. 2) and demonstrate that the stoichiometry of GMAG phosphorylation is at least one order of magnitude greater than the phosphorylation of the S-MAG isoform. While previous observations had suggested that only GMAG is an in vivo protein kinase substrate (Edwards et al., 1988, 1989), it is likely that the inability to detect SMAG phosphorylation in brain reflects its relatively lower stoichiometry of phosphorylation. The protein kinases involved in MAG phosphorylation in brain are currently unknown. However, the phosphorylation of L-MAG in NIH 3T3 cells is comparable with the pattern observed in brain (Edwards et al., 1988),suggesting that common regulatory pathways are involved. TPA, a PKC activator, can augment the phosphorylation of L-MAG but not S-MAG in NIH 3T3 cells. Therefore, the potential PKC phosphorylation site that is common between L-MAG and S-MAG is unlikely to be the site stimulated by TPA in tissue culture cells. Since vanadate treatment stimulates the phosphorylation of the same tryptic fragment as TPA, it is likely that this phosphopeptide contains phosphotyrosine. This suggests that the potential PKC phosphorylation sites identified in L-MAG by Salzer et al. (1987) (amino acids 575-582 and 604-608) are not stimulated by TPA, since these sites reside in tryptic fragments that do not contain any tyrosine residues. The tryptic fragments spanning amino acids 588-604 or 609-624 are more likely to be candidates for TPAand vanadate-induced phosphorylation, since these fragments contain serine and tyrosine residues in the former and serine, tyrosine, and threonine residues in the latter peptide. It is intriguing that L-MAG, which is expressed concomitantly with the initiation of the myelin program, appears to be a PKC substrate, since it has been previously shown that adherence of oligodendrocytesto a substratum causes a simultaneous activation of myelinogenesisand a PKC-mediated phosphorylation of myelin basic protein (Vartanian et al., 1986). Protein tyrosine kinases have been implicated in regulatory processes during development and differentiation (Hafen et al., 1987; Haller et al., 1987; Mellstroem et al., 1987; Geissler et al., 1988). They can be divided into two major classes: (a) the growth factor
DIFFERENTIAL PHOSPHORYLATION OF MAG ISOFORMS receptor-type protein tyrosine kinases, such as the EGF and insulin receptors; (b) the cytoplasmic protein tyrosine kinase, such as the protein products of the c-SIC and c-abl protooncogenes (Hunter, 1987).While EGF, IGF- 1, and insulin had no effect on the phosphorylation of MAG, v-src transformation significantly increased L-MAG phosphotyrosine levels. Vanadate did not raise the constitutive tyrosine phosphorylation of L-MAG in v-src-transformed cells, indicating that pp60v-srclikely phosphorylates the same tyrosine residues as the endogenous NIH 3T3 cell MAG protein tyrosine kinase. It may be that L-MAG is phosphorylated by the cellular homologue of the protein product of v-src, pp60c-src,which has been detected in myelin (Hirano et al., 1988). A potential role for tyrosine-specific protein phosphorylation in the regulation of cell-cell and/or cellmatrix interactions has been demonstrated for the sevenless homeotic gene of Drosophifa(Hafen et al., 1987), the platelet integrin-glycoproteingpIIb-IIIa (Ferrell and Martin, 1989; Golden and Brugge, 1989), and the fibronectin receptor (Hirst et al., 1986). Recently, it was demonstrated that tyrosine phosphorylated integrins, isolated from Rous sarcoma virus-transformed chick embryo fibroblasts,exhibited a decreased ability to bind fibronectin and talin compared to nonphosphorylated receptor (Tapley et al., 1989). The critical integrin phosphorylation site has been mapped to tyrosine 788, which is located in a domain responsible for binding the cytoskeleton (Tapley et al., 1989). A homologous stretch of amino acids is found in both MAG isoforms (amino acids 55 1-573; Tamkun et al., 1986;see Salzer et al., 1987). This region in MAG contains two tyrosines, one of them (tyrosine 558) nestled among amino acids similar to those surrounding tyrosine 788 of the fibronectin receptor. Since there are only two tyrosine residues in the S-MAG cytoplasmic domain, both within the integrin homology region, it is possible that this conserved region may be a protein tyrosine kinase target in two distinct cell adhesion molecules. We speculate that phosphorylation of the cytoplasmic domain may modify the interaction between MAG and certain cytoplasmic components, thus regulating MAG-specific cell adhesion functions. Alternatively, L-MAG may function as a signal transducer as has been observed for other cell adhesion molecules that are members of the immunoglobulin gene superfamily (Veillette et al., 1988; Schuch et al., 1989). The phosphorylation sites found on the unique domain of L-MAG may represent docking sites for protein kinases. Oligomenzation of L-MAG molecules through ligand binding may result in kinase activation, generation of second messenger signals, and ultimately alteration in gene expression. S-MAG, which lacks this domain, may have more restricted activities and serve only as an adhesive molecule.
1425
and D.E.H.A. is a recipient of a n FRSQ scholarship. We appreciate the help given to us by Beth Mason in preparing this manuscript.
REFERENCES Arquint M., Roder J., Chia L.-S., Down J., Wilkinson D., Bayley H., Braun P. E., and Dunn R. (1987) Molecular cloning and primary structure of myelin associated glycoprotein. Proc. Natl. Acad. Sci. USA 84,600-604. Cooper J. A., Sefton B. M., and Hunter T. (1983) Detection and quantification of phosphotyrosinein proteins. Methods Enzymol. 99,387-405.
Edwards A. M.,Arquint M., Braun P. E., Roder J. C., Dunn R. J., Pawson T., and Bell J. C. (1988) Myelin-associated glycoprotein, a cell adhesion molecule of oligodendrocytes, is phosphorylated in brain. Mol. Ce[l. Biol. 8, 2655-2658. Edwards A. M., Braun P. E., and Bell J. C. (1989) Phosphorylation of myelin-associated glycoprotein in vivo and in vitro occurs only in the cytoplasmic domain of the large isoform. J. Neurochem. 52, 317-320. Ferrell J. E. and Martin G. S. (1989) Tyrosine-specificprotein phosphorylation is regulated by glycoprotein IIb-IIIa in platelets. Proc. Natl. Acad. Sci. USA 86,2234-2238. Frail D. E. and Braun P. E. (1984) Two developmentally regulated messenger RNAs differing in their coding region may exist for myelin-associated glycoprotein. J. Biol. Chern. 259, 1485714862.
Geissler E. N., Ryan M. A., and Housman D. E. (1 988) The dominantwhite spotting (W) locus of the mouse encodes the c-kit protooncogene. Cell 55, 185-192. Golden A. and Brugge J. S. (1989) Thrombin treatment induces rapid changes in tyrosine phosphorylation in platelets. Proc. Nut/.Acad. Sci. USA 86,901-905. Graham F. L. and van der Eb A. J. (1973) A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52, 456-467.
Hafen E., Basler K., Edstroem J.-E., and Rubin G. (1987) Sevenless, a cell-specific homeotic gene of Drosophila, encodes a putative transmembrane receptor with a tyrosine kinase domain. Science 236,55-63.
Haller J., Cote S., Broenner G., and Jaeckle H. (1987) Dorsal and neural expression of a tyrosine kinase-related Drosophila gene during embryonic development. Genes Dev. 1,862-867. Hirano A. A,, Greengard P., and Huganir R. L. (1988) Protein tyrosine kinase activity and its endogenous substrates in rat brain: a subcellular and regional survey. J. Neurochem. 50, 1447-1455. Hirst R., Horwitz A,, Buck C., and Rohrschneider L. (1986) Phosphorylation of the fibronectin receptor complex in cells transformed by the oncogenes that encode tyrosine kinases. Proc. Natl. Acad. Sci. USA 83,6470-6474. Horvath E., Edwards A. M., Bell J. C., and Braun P. E. (1 989) Chemical deglycosylationon a microscale of membrane glycoproteins with retention of phosphoryl-protein linkage. J. Neurosci. Rex (in press). Hunter T. (1987) A thousand and one protein kinases. CeN 50,823829.
Johnson P. W., Abramow-Newerly W., Seilheimer B., Sadoul R., Tropak M. B., Arquint M., Dunn R. J., Schachner M., and Roder J. C. (1989) Recombinant myelin-associated glycoprotein confers neural adhesion and neurite outgrowth function. Neuron 3,377-385.
Kirschmeier P. T., Housey G. M., Johnson M. D., Perkins A. S., and Weinstein I. B. (1988) Construction and characterization of a retroviral vector demonstrating efficient expression of cloned cDNA sequences. DNA 7,219-225. Klarlund J. K. (1985) Transformation of cells by an inhibitor of phosphatases acting on phosphotyrosineproteins. CeN41,707717.
Acknowledgment: This work was supported by NCI grants to J.C.B. and P.E.B. J.C.B. is an M R C of Canada scholar
Lai C., Brow M. A., Nave K.-A,, Noronha A. B., Quarles R. A., Bloom F. E., Milner R. J., and Sutcliffe J. G. (1987) Two forms
J. Nertrochem.. Val. 55. No. 4. 1990
1426
D. E. H. AFAR ET AL.
of IB236/myelin-associatedglycoprotein, a cell adhesion molecule for postnatal neural development, are produced by alternative splicing. Proc. Natl. Acad. Sci. USA 84,4337-4341. Mellstroem K., Bjelfman C., Hammerling U., and Pahlman S. (1987) Expression of c-src in cultured human neuroblastoma and smallcell lung carcinoma cell lines correlates with neurocrine differentiation. Mol. Cell. Biol. 7, 4 178-4 184. Nishizuka Y.(1984)The role ofprotein kinase C in cell surface signal transduction and tumor promotion. Nature 308, 693-698. Poltorak M., Sadoul R., Keilhauer G., Landa C., Fahrig T., and Schachner M. (1987) Myelin associated glycoprotein, a member of the L2/HNK- I family of neural cell adhesion molecules, is involved in neuron-ohgodendrocye and oligodendrocyte-oligodendrocyte adhesion. J. Cell Biol. 105, 1893-1899. Salzer J., Holmes P. W., and Colman D. R. (1987). The amino acid sequence of the myelin-associated glycoproteins: homology to the immunoglobulin gene superfamily. J. Cell Biol. 104, 957965. Schuch U., Lohse M. J., and Schachner M. (1989)Neural cell adhesion molecules influence second messenger systems. Neuron 3, 1320. Sternberger N. H., Quarles R. H., Itoyama Y., and deF. Webster H. ( 1979) Myelin-associated glycoprotein demonstrated immuno-
J. Neurochem., Vol. 55. No. 4, 1990
cytochemically in myelin and myelin-formingcells of developing rat. Proc. Natl. Acad. Sci. USA 76, 15 10- 15 14. Tamkun J. W., DeSimone D. W., Fonda D., Pate1 R. S., Buck C., Honvitz A. F., and Hynes R. 0. (1986) Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46, 271-282. Tapley P., Honvitz A. F., Buck C., Duggan K., and Rohrschneider L. (1989) Integrins isolated from rous sarcoma virus-transformed chicken embryo fibroblasts. Oncogene 4,325-333. Tropak M. B., Johnson P. W., Dunn R. J., and Roder J. C. (1988) Differential splicing of MAG transcripts during CNS and PNS development. Mol. Brain Res. 4, 143-155. Vartanian T. S., Szuchet S., Dawson G., and Campagnoni A. T. ( 1986) Oligodendrocyte adhesion activates protein kinase Cmediated phosphorylation of myelin basic protein. Science 234, 1395-1 398. Veillette A,, Bookman M. A,, Horak E. M., and Bolen J. B. (1988) The CD4 and CD8 T-cell surface antigens are associated with the internal membrane tyrosine-protein kinase ~ 5 6 ''~Cell . 55, 301-308. Williams A. F. (1987) A year in the life of the immunoglobulin superfamily. Immunol. Today 8,298-303.
Differential Phosphorylation of Myelin-Associated Glycoprotein Isoforms in Cell Culture Daniel E. H. Afar, *James L. Salzer, ?John Roder, $Peter E. Braun, and John C. Bell Departments of Biochemistry and Medicine, University of Ottawa, Ottawa, and ?Department of Medical Genetics, University of Toronto, Toronto, Ontario, and $Department of Biochemistry, McGill University, Montreal, Quebec, Canada; and *Department of Cell Biology, New York University Medical Center, New York, New York, U.S.A.
Abstract: The alternative splicing of myelin-associated glycoprotein (MAG) mRNA generates two isoforms that harbor distinct potential phosphorylation sites in their cytoplasmic tails. Here we characterize the in vivo phosphorylation of MAG isoforms in NIH 3T3 cells transfected with the cDNAs encoding the two isoforms of MAG. Our results demonstrate that the longer isoform, L-MAG, is phosphorylated constitutively mainly on serine, but also on threonine and tyrosine residues. This phosphorylation is subject to change by 12-0tetradecanoylphorbol 13-acetate (TPA) and ammonium vanadate, but not by dibutyryl-cyclic AMP. The shorter isoform, S-MAG, is constitutivelyphosphorylated only on serine residues. While TPA and dibutyryl-cyclic AMP have no de-
tectable effect, ammonium vanadate induces tyrosine and threonine phosphorylation in S-MAG. 32Plabeling of v-srctransformed NIH 3T3 cells that express L-MAG also show that L-MAG is likely to be an in vivo substrate for pp60” tyrosine kinase activity. These results demonstrate that both MAG isoforms are phosphorylated in a heterologous cell system and that this phosphorylation is subject to pharmacological manipulation. Key Words: Myelin-associated glycoprotein-Tyrosine phosphorylation-Protein kinase C-Cell adhesion. Afar D. E. H. et al. Differential phosphorylation of myelin-associated glycoprotein isoforms in cell culture. J. Neurochem. 55, 14 18- 1426 ( 1990).
The process of selective cell adhesion is essential for the organization of tissues in higher organisms during development. The characterization of molecules involved in cell adhesion is therefore basic to the understanding of developmental events. Certain types of cell surface glycoproteins have been demonstrated to play a fundamental role in this process. One group of cell adhesion molecules has been included in the immunoglobulin supergene family by virtue of their sequence and structural features (Williams, 1987). Among these molecules is myelin-associated glycoprotein (MAG), a transmembrane glycoprotein involved in neuron-glia and glia-glia interactions in the developing nervous system (Poltorak et al., 1987). MAG is expressed on the surface of oligodendrocytes in the CNS and Schwann cells in the PNS (Sternberger et al., 1979). MAG mRNA is alternatively spliced into two developmentally regulated isoforms, suggesting that the
larger (L-MAG) and the smaller (S-MAG) isoforms perform functionally different roles in myelin formation (Frail and Braun, 1984; Tropak et al., 1988). The early appearance of L-MAG at the onset of the myelin program suggests a potential role for the cytoplasmic domain of this isoform in the initial processes of intercellular recognition and signal transduction. S-MAG mRNA, which is maximally expressed in mature myelin, may play a structural role in maintaining the integrity of the myelin sheath. L-MAG and S-MAG exhibit identical external and transmembrane domains, but differ in their cytoplasmic carboxyl termini ( h i et al., 1987; Salzer et al., 1987). Both isoforms contain several potential phosphorylation sites in the carboxylterminal region, among which is a region identified as a potential protein kinase C (PKC) phosphorylation site by Arquint et al. (1987). In addition, L-MAG contains a segment potentially phosphorylated by protein
Received December 2 I , 1989; revised manuscript received March 20, 1990; accepted March 22, 1990. Address correspondence and reprint requests to Dr. J. C. Bell at Department of Biochemistry, University of Ottawa, 45 1 Smyth Rd., Ottawa, Ontario, Canada KIH 8M5. Abbreviations used: CAMP, cyclic AMP; dbcAMP, dibutyrylCAMP;EGF, epidermal growth factor; IGF-I, insulin-like growth
factor 1; LM and SM cells, L-MAG- and S-MAG-expressing fibroblasts; L-MAG and S-MAG, larger and smaller isoforms of MAG, MAG, myelin-associated glycoprotein; MEM, minimum essential medium; PAGE, polyacrylamide gel electrophoresis; PKC, protein kinase C; PVsrcLM cells, L-MAG-expressing v-src-transformed fibroblasts; SDS, sodium dodecyl sulfate; TPA, 12-0-tetradecanoylphorbol 13-acetate.
1418
DIFFERENTIAL PHOSPHORYLA TION OF MAG ISOFORMS kinase A (Arquint et al., 1987) and PKC (Salzer et al., 1987),as well as a region homologous to the epidermal growth factor (EGF) receptor autophosphorylation site (Arquint et al., 1987; Salzer et al., 1987). We have previously detected phosphorylation of L-MAG in myeh a t i n g rodent brain in vivo as well as in myelin in vitro on serine, threonine, and tyrosine residues (Edwards et al., 1988). Phosphorylation of S-MAG was never detected in myelinating brain. Here we report the isolation and characterization of fibroblast cell lines expressing the MAG cDNAs. Our results demonstrate that L-MAG is phosphorylated on serine, threonine, and tyrosine residues in cell culture, whereas S-MAG is constitutively phosphorylated only on serine residues. While growth factors are unable to modulate MAG phosphorylation, the v-src kinase constitutively phosphorylates L-MAG on tyrosine residues. The results presented here are consistent with phosphorylation of the cytoplasmic domains of both molecules playing a role in the regulation of MAG activity in developing and mature myelin.
MATERIALS AND METHODS Cell lines LMAG cDNA was inserted into the retroviral vector pmv7 (Kirschmeieret al., 1988) and transfected into NIH 3T3 virus packaging cell lines (PA317 and Psiz) by calcium phosphate coprecipitation (Graham and van der Eb, 1973). Stable transfectants of L-MAG-expressing cells were selected by G4 18 resistance [600 pgml in a-minimum essential medium (aMEM), 10%fetal calf serum]. The transfectants, referred to as LM cells, were screened for MAG expression using immunofluorescent cytochemistryon live cells and immunoblot analysis of membrane extracts. The generation of stable S-MAG-expressing cell lines has previously been described (Johnson et al., 1989). The S-MAGexpressing clone used in our study (3TV2-9)will be referred to as “SM cells.” The v-src-transformed NIH 3T3 cell line, PVsrc, was infected by cocultivation with virus from the Psizcells expressing L-MAG. Psi2-L-MAG cells were treated with mitomycin C (200 pg/ml) for 2 h and then cocultured with PVsrc cells in a ratio of 1:2 for 24 h. After 24 hours stable infectants were selected by G4 18 resistance (600 pglml in aMEM, 10% fetal calf serum) and are referred to as PVsrcLM cells.
Phosphorylation of MAG LM and SM cells were labeled with 0.5 mCi [32P]orthophosphate/mlphosphate-free Dulbecco’s Modified Eagle Medium and 10%fetal calf serum for 4 h in the presence or absence of ammonium vanadate. The cells were harvested by lysis in 250 pl of hot ( 100°C) 2% sodium dodecyl sulfate (SDS). The cell lysate was passaged several times through a 25-gauge syringe needle and diluted to 1 ml with an immunoprecipitation buffer (10 mM Tris-HC1, pH 7.4, 5 mM EDTA, 150 mMNaCl,l%Triton, 500 mMNH,VO,, 2 mM NaF, 2 mM Na-pyrophosphate, 2 pg/ml aprotinin, 5 pglml leupeptin, 0.4 p&ml pepstatin, 200 pdml polymethylsulfonyl fluoride). Undissolved cellular debris was removed by centrifugation for 10 min at 14,000 rpm in a microfuge. Immunoprecipitation of MAG was performed by overnight incubation of the cell lysate with a mixture of anti-MAG
1419
monoclonal antibodies Gen 3 SI and Gen 3 S3 (generously provided by N. Latov). The following day the mixture was incubated with rabbit anti-mouse antibody (Jackson Immunoresearch Lab.) for 1 h prior to addition of protein Acoupled Sepharose beads (Pharmacia). After an additional hour of incubation, the beads were washed five times with the immunoprecipitation buffer and MAG was eluted using SDS sample buffer. The immunoprecipitated samples were resolved by SDS polyacrylamide gel electrophoresis (PAGE; 10%polyacrylamide) and detected by autoradiography.
Stoichiometry of L-MAG phosphorylation To determine the half-life of L-MAG in LM cells, LM cells were labeled for 18 h at 37OC with [35S]methionine(50 pCi/ ml; Amersham) in methionine-free minimal essential medium (MEM), supplemented with 10%fetal calf serum and 5% aMEM. Then the cells were washed two times with phosphate-buffered saline and were either harvested in 2% SDS or were incubated for an additional 4 and 9 h in aMEM. LMAG was immunoprecipitated as described above and resolved on SDS-PAGE. The half-life was calculated by extracting radioactive L-MAG out of excised gel slices and determining the amount of radioactivity in the extracts using liquid scintillation counting. To assess the extent of L-MAG phosphorylation, LM cells were labeled for 18 h at 37°C with either [32P]orthophosphate (200 pCi/ml) or [35S]methionine(50 pCi/ml) in aMEM containing 10%fetal calf serum. L-MAG was immunoprecipitated and resolved on SDS-PAGE. The number of moles of phosphate in L-MAG was determined by scintillation counting of the excised L-MAG bands. The number of moles of immunoprecipitated LMAG was estimated from the level of [35S]methionineincorporated, assuming that L-MAG contains nine methionine residues (Salzer et al., 1987). The specific activity of label in the medium was determined by scintillation counting of an aliquot of the medium and from the known concentrations of phosphate (10 m M in aMEM and 1.5 mM in fetal calf serum) and methionine (0.2 p M in aMEM and 0.03 m M in fetal calf serum).
Treatment of L-MAG-expressing cells with growth factors, dibutyryl-cyclic AMP, and phorbol esters To assess the effects of growth factor treatment on L-MAG phosphorylation, LM cells were fed aMEM without serum for 24 h. The next day the cells were 32P-labeledin phosphatefree medium without serum for 4 h. EGF (1 50 nglml; Boehringer Mannheim), insulin-like growth factor 1 (IGF- 1; Boehringer Mannheim), or insulin (0.5 pM, generously provided by Dr. M. Bernier) was added to the cells at the end ofthe labeling period for 10 min at 37°C. The effects of cyclic AMP (CAMP)-mediated phosphorylation of L-MAG and SMAG were assessed by incubating LM and SM cells with either 5 mM dibutyryl-CAMP (dbcAMP Sigma) or sodium butyrate (5 mM) during the labeling procedure. Sodium butyrate was used to control for non-CAMP-related effects of dbcAMP. To determine the effects of phorbol esters on MAG phosphorylation, LM and SM cells were treated with 100 nM 4a-phorbol or 100 nM 12-0-tetradmnoylphorbol 13-acetate (TPA) for 10 min at 37°C at the end of “P labeling of cells. At the end of the labeling period, the cells were harvested in hot 2% SDS and processed as described before.
Phosphoamino acid analysis Phosphorylated MAG isoforms were immunoprecipitated as described above, excised out of the polyacrylamide gel, J. Neurochem., Vol. 55, No. 4, 1990
1420
D. E. H. AFAR ET AL.
FIG. 1. lmmunofluorescent cytochemistry on MAGexpressing fibroblasts. Live LM (A), SM (B),and PVsrcLM (C) cells were stained with monoclonalanti-MAG antibody (513) and subsequently with fluoresceinoonjugatedgoat antimouse secondary antibody. No fluorescence was observed with cells stained with secondary antibody alone. The exposure times and magnification are the same for all pictures. Bar, 20 Pm.
J. Nerirochem.. VoI 55. No. 4, 1990
DIFFERENTIAL PHOSPHOR YLATION OF MAG ISOFORMS electroeluted using a BioRad electroeluter, and precipitated by the addition of trichloroacetate (20%). The trichloroacetate precipitates were washed twice in ice-cold acetone and hydrolyzed with 6M double-distilled HCl for 75 min at 1 10°C. The hydrolysate was lyophilized, mixed with authentic phosphoamino acids, spotted onto MN cell 400 thin-layer cellulose plates, and resolved by two-dimensional electrophoresis as described by Cooper et al. (1 983). Phosphoamino acid standards were detected by ninhydrin staining of the thin-layer plates. Radioactive phosphoamino acids were detected by autoradiography .
Tryptic digestion of MAG Phosphorylated MAG was analyzed by digestion with trypsin essentially as described previously (Edwards et al., 1989) with the modification that tryptic fragments were spotted onto MN cell 400 thin-layer cellulose plates for separation by ascending chromatography. The chromatograms were dried and rotated 90", and thin-layer electrophoresis was performed at pH 1.9 for 1 h at 1,000V. Phosphotryptic fragments were visualized by autoradiography.
Immunoblot analysis Cell membranes were prepared from LM and SM cells by harvesting the cells in 10 ml of ice-cold hypotonic HEPES buffer (20 mM HEPES, pH 7.2, 3 mM KCI, 1 mM EDTA, 500 p M NH4V04, 2 mM NaF, 2 pg/ml aprotinin, 5 pg/ml leupeptin, 0.4 p d m l pepstatin, 200 pg/ml polymethylsulfonyl fluoride). The cells were then lysed in a tight-fitting glass homogenizer using 15 strokes on ice. DNA and cellular debris were removed by a 1,000 g centrifugation for 10 min. A membrane-enriched fraction was obtained by subjecting the supernatant to centrifugation for 30 min at 100,000 g. The membrane pellet was resuspended in the hypotonic HEPES buffer with 40% glycerol. Protein content was determined using the BioRad reagent. Twenty micrograms of membrane protein was dissolved in sample buffer, while 150 p g of membrane protein was precipitated with 12%trichloroacetate and chemically deglycosylated using trifluoromethanesulfonic acid (Aldrich) essentially as described before (Horvath et al., 1989). The proteins were resolved on SDS-PAGE (10%polyacrylamide), transferred to nitrocellulose paper (Schleicher and Schuell), and probed with a polyclonal antibody directed against MAG or an anti-phosphotyrosine monoclonal antibody (PY20; ICN). Visualization ofthe blots was carried out using either an alkaline phosphatase-coupled goat anti-rabbit second antibody (for MAG) or a '251-labeledsheep anti-mouse second antibody (for phosphotyrosine).
1421
RESULTS Characterization of MAG expression in fibroblasts Live MAG-expressing cells were examined by immunofluorescent cytochemistry using a monoclonal anti-MAG antibody (5 13). The results indicate that LMAG and S-MAG are both expressed uniformly on the cell surface (Fig. 1). Immunoblot analysis of membrane extracts of these cells confirms that the antibody recognizes L-MAG and S-MAG in LM and SM cells, respectively (Fig. 2, lanes a and b). It should be noted that glycosylated S-MAG exhibits a slightly lower molecular weight than glycosylated L-MAG. Phosphorylation of MAG MAG phosphorylation was assessed by labeling the cells in phosphate-free medium containing [32P]orthophosphateand immunoprecipitating MAG from a cell extract with monoclonal anti-MAG antibodies (Gen 3 S1 and Gen 3 S3). The immunoprecipitations were performed on cell lysates from equal numbers of cells, and the results demonstrate that LMAG and S-MAG are both phosphorylated (Fig. 2). Using differential labeling with [35S]methionineand [32P]orthophosphate, L-MAG protein was shown to contain 10 mol phosphate/mol polypeptide. In parallel experiments we determined that the half-life of MAG is -4 h in NIH 3T3 cells. For phosphate stoichiometry determinations, cells were incubated overnight ( 18 h) with [32P]orthophosphate and [35S]methionine and therefore contain MAG molecules that have been uniformly labeled with both isotopes. Thus, our calculation of the extent of phosphorylation will not be biased by long-lived MAG molecules that do not contain
a b
c d e f
200-
9768 -
Immunofluorescence cytochemistry Cells were plated onto coverslips and incubated with antiMAG monoclonal antibody 5 13 ( 1 :1,000 in phosphate-buffered saline and 2% calf serum; generously provided by M. Schachner) for 1 h at 4°C. The coverslips were then washed three times in phosphate-buffered saline and 2% calf serum and subsequently incubated with a fluorescein-conjugated goat anti-mouse second antibody (Bio Can) for 1 h at 4°C. After three washes in phosphate-buffered saline and 2% calf serum, the cells were fixed in 4% formaldehyde for 20 min and then mounted onto slides. Visualization of the immunofluorescence was achieved using epifluorescent microscopy.
Phosphorylation of MAG isoforms in LM and S M cells. lmmunoblot analysis of membrane extracts from LM (lane a) and S M (lane b) cells was probed with a mixture of monoclonal antiMAG antibodies Gen 3 S1 and Gen 3 S3 and detected using an 'Z51-labeledsheep anti-mouse antibody. Cell extracts from 3'P-labeled LM (lane c), S M (lane d), and NIH 3T3 (lane e) cells were immunoprecipitated using anti-MAG antibodies as described in Materials and Methods. A sample containing a mixture of LM and S M cell extract was immunoprecipitated with nonimmune serum (lanef). The samples were analyzed using SDS-PAGE and detected by autoradiography FIG. 2.
J Neurocliem.. Vol. 55. No 4. 1990
D. E. H . AFAR ET AL
1422
[35S]methioninebut are posttranslationally modified with 32P.Furthermore, tryptic mapping of 32P-labeled MAG molecules revealed that there are at least seven tryptic phosphopeptides present in L-MAG (see Fig. 6). Taken together these results indicate that L-MAG is in fact a formidable protein kinase substrate. Although SM cells express high levels of S-MAG on the cell surface, the stoichiometry of S-MAG phosphorylation was determined to be approximately one-tenth the amount of L-MAG phosphorylation (Fig. 2). Phosphoamino acid analysis of L-MAG phosphorylated in metabolically labeled fibroblasts shows that it is constitutively phosphorylated mainly on serine, but also on threonine and tyrosine residues (Fig. 3B), thereby mimicking the phosphorylation pattern in brain. Ammonium vanadate, a phosphotyrosine phosphatase inhibitor, increases phosphotyrosine levels in cells (Klarlund, 1985). Although no quantitative change occurred in phosphorylated L-MAG protein from vanadate-treated cells (cf. lanes a and b, Fig. 3A), there was a dose-dependent increase in the amount of phosphotyrosine present in the protein (Fig. 4). Phosphothreonine and phosphoserine levels increased with 50 p M ammonium vanadate, but decreased with higher concentrations (Fig. 4). While ammonium vanadate did not change the amount of L-MAG in the mem-
A
B a
b
c
200-
a
97-
a
C
b
d
e
FIG. 4. Effect of increasing doses of ammonium vanadate on LMAG phosphorylation. LM cells were labeled with 32Pin the presence of increasing concentrations of ammonium vanadate. Phospho-amino acid analysis was performed on immunoprecipitated phosphorylated L-MAG protein using two-dimensional thin-layer electrophoresis as described in Materials and Methods. Samples are shown in order of increasing concentration of ammonium vanadate: 10 (a), 50 (b), 100 (c), 200 (d),and 500 (e)p M . The positions of phosphoserine(S),phosphothreonine(T), and phosphotyrosine (Y) are shown for reference in (b).
b
C
FIG. 3. The effect of vanadate and v-sfc transformation on phosphorylation of L-MAG in vivo. A Phosphorylated L-MAG was immunoprecipitatedfrom LM cells 32P-labeledin the absence (lane a) or presence (lane b) of 500 pM ammonium vanadate and from 32P-labeledPVsrcLM cells (lane c). The samples were resolved on SDS-PAGE and detected by autoradiography. 0: Phosphoamino acid analysis on electroeluted L-MAG protein was performed using two-dimensional thin-layerelectrophoresisas described in Materials and Methods: L-MAG from untreated LM cells (a); L-MAG from vanadate-treatedLM cells (b); L-MAG from PVsrcLM cells (c). The positions of phosphoserine (S), phosphothreonine (T), and phosphotyrosine (Y) are indicated for reference in (a).
J Nrunx.hi.m, L’ol 55. No 4, 1990
brane of these cells as detected by immunoblot analysis (Fig. 5A) and immunofluorescent cytochemistry (data not shown), it significantlyelevated its phosphotyrosine content (Fig. 5B). In fact, L-MAG was determined to be the major tyrosine phosphorylated protein in these cells. Control experiments determined that the faintappearing protein bands on the immunoblots are nonspecific proteins reacting with the secondary antibody (data not shown).
Effect of pharmacological agents on MAG phosphorylation The L-MAG cDNA sequence revealed the presence of potential phosphorylation sites for PKC, CAMP-dependent kinase, and an EGF receptor autophosphorylation site (Arquint et al., 1987; Salzer et al., 1987), while S-MAG exhibited only a possible PKC phosphorylation site (Arquint et al., 1987). We examined the effects of various kinase activators on L-MAG and
DZFFERENTZAL PHOSPHOR YLATION OF MAG ISOFORMS
B
A
abed 200
-
07-
a b c d
200-
07-
6868-
43-
43-
1423
ylated L-MAG protein from TPA- and dbcAMP-stimulated cells showed no qualitative change in the ratio of phosphoserine to phosphothreonine to phosphotyrosine (data not shown). Phosphoamino acid analysis of phosphorylated SMAG demonstrated that constitutive phosphorylation occurred only on serine residues. While TPA and dbcAMP had no detectable effect on S-MAG phosphorylation as determined by SDS-PAGE (Fig. 7A) and tryptic digest (data not shown) analysis, vanadate treatment induced low but detectable levels of phosphotyrosine and phosphothreonine in this isoform (Fig. 7B).
L-MAG phosphorylation in PVsrcLM cells Previously we demonstrated that partially purified MAG can be phosphorylated by pp 130gag-fpsand pp60"~"cin vitro, two cytoplasmic tyrosine kinases en-
FIG. 5. L-MAG is the major tyrosine phosphorylated membrane protein in LM cells. LM cells were left untreatedor were pretreated with 500 p.M vanadate for 3 h. Membrane extracts were prepared and were either left untreated or were subjected to chemical d e glycosylationwith trifluoromethanesulfonate as described in Materials and Methods. lmmunoblot analysis was performed using a polyclonalanti-MAG antibody (A) or using the PY20 anti-phosphotyrosine antibody (B). Samples were as follows: LM membranes from cells untreated(lane a) and deglycosylated (lane b), LM membranes from cells vanadate-pretreated(lane c) and deglycosylated (lane d).
S-MAG phosphorylation. While EGF, IGF- 1, and insulin had no detectable effect on the phosphorylation of L-MAG (data not shown), TPA, a known activator of the serine-threonine-specificPKC (Nishizuka, 1984), stimulated L-MAG phosphorylation as determined by tryptic digestion of L-MAG (Fig. 6d). The phosphopeptide in L-MAG, where the major stimulation of phosphorylation with TPA occurs, is also stimulated with vanadate treatment (Fig. 6b and d, arrows). The inactive phorbol ester 4a-phorbol served as a control for TPA stimulation and did not have an effect on constitutive L-MAG phosphorylation (Fig. 6, cf. a and c). Treatment of LM cells with dbcAMP, an activator of a different class of serine-threonine kinase, quantitatively increased the phosphorylation of L-MAG in some experiments (data not shown). However, dbcAMP did not alter L-MAG phosphorylation significantly in a qualitative sense (Fig. 6f). In this case sodium butyrate was used to control for non-CAMPrelated effects of dbcAMP and no effect was observed (Fig. 6e). Phosphoamino acid analysis of phosphor-
TLC)
TLE
FIG. 8. Tryptic digestion of L-MAG phosphorylated in LM cells. LM cells were 32P-labeledand were left untreated(a) or were treated with 500 p M ammonium vanadate (b), 4a-phorbol (c), TPA (d), sodium butyrate (e), or dbcAMP (f) as described in Materials and
Methods. Phosphorylated1-MAG was purifiedand digested with trypsin. The tryptic fragments were resolved by thin-layer chromatography (TLC) and electrophoresis (TLE) as described in Materials and Methods. The tryptic fragments were detected by autoradiography. Phosphopeptidesthat were stimulated during TPA and vanadate treatment are identified with arrows.
J. Neurachem.. Val. 55. No. 4, 1990
D.E. H . AFAR ET AL.
1424
A a b c d e 200
f
-
68 97
FIG. 7. The effect of various pharmacologicalagents on S-MAG phosphorylation. A Phosphorylated S-MAG was immunoprecipitated from 32P-labeledSM cells, resolved on SDS-PAGE, and d e tected by autoradqraphy. SM cells were either left untreated (lane a) or labeledin the presence of 500 p M ammonium vanadate (lane b), 4a-phOrbd (lane c), TPA (lane d). sodium butyrate (lane e), or dbcAMP (lane f). The radioactive material at the top of the gel in each lane is labeled nucleic acid that associates nonspecifically to protein A-Sepharose. The amount of this material varies randomly between samples in different experiments.B Phosphoamino acid analysison phosphorylatedS-MAG immunopurifiedfrom vanadate treated SM cells. S , phosphoserine; T. phosphothreonine; Y, phosphotyrosine.
coded by the v-ks and v-src oncogenes, respectively (Edwards et al., 1988). To investigate whether L-MAG can be a substrate for pp60"-srcin vivo, the L-MAG cDNA was introduced into v-src-transformed NIH 3T3 cells by retroviral infection. In PVsrcLM cells the cell surface distribution of L-MAG appears uniform (Fig. 1). Thus, v-src transformation does not appear to affect the transport of MAG to the cell surface, or its distribution on the cell membrane. Phosphoamino acid analysis of L-MAG protein immunoprecipitatedfrom "P-labeled PVsrcLM cells demonstrates that L-MAG is constitutively phosphorylated mainly on serine and tyrosine residues, with a small amount of phospho-
J. Neurochem.. Vol. 55, No. 4, 1990
threonine being present (Fig. 3). The high level of phosphotyrosine present in GMAG isolated from these cells is not due to pp60'-s" activity during the isolation procedure, since the cells were lysed in hot 2% SDS. Labeling PVsrcLM cells in the presence of ammonium vanadate did not change the constitutive phosphorylation of L-MAG (data not shown).
DISCUSSION In this study we demonstrate that both L-MAG and S-MAG are phosphorylated in a heterologous cell culture system. Our present results are obtained from cells expressing high amounts of MAG (Fig. 2) and demonstrate that the stoichiometry of GMAG phosphorylation is at least one order of magnitude greater than the phosphorylation of the S-MAG isoform. While previous observations had suggested that only GMAG is an in vivo protein kinase substrate (Edwards et al., 1988, 1989), it is likely that the inability to detect SMAG phosphorylation in brain reflects its relatively lower stoichiometry of phosphorylation. The protein kinases involved in MAG phosphorylation in brain are currently unknown. However, the phosphorylation of L-MAG in NIH 3T3 cells is comparable with the pattern observed in brain (Edwards et al., 1988),suggesting that common regulatory pathways are involved. TPA, a PKC activator, can augment the phosphorylation of L-MAG but not S-MAG in NIH 3T3 cells. Therefore, the potential PKC phosphorylation site that is common between L-MAG and S-MAG is unlikely to be the site stimulated by TPA in tissue culture cells. Since vanadate treatment stimulates the phosphorylation of the same tryptic fragment as TPA, it is likely that this phosphopeptide contains phosphotyrosine. This suggests that the potential PKC phosphorylation sites identified in L-MAG by Salzer et al. (1987) (amino acids 575-582 and 604-608) are not stimulated by TPA, since these sites reside in tryptic fragments that do not contain any tyrosine residues. The tryptic fragments spanning amino acids 588-604 or 609-624 are more likely to be candidates for TPAand vanadate-induced phosphorylation, since these fragments contain serine and tyrosine residues in the former and serine, tyrosine, and threonine residues in the latter peptide. It is intriguing that L-MAG, which is expressed concomitantly with the initiation of the myelin program, appears to be a PKC substrate, since it has been previously shown that adherence of oligodendrocytesto a substratum causes a simultaneous activation of myelinogenesisand a PKC-mediated phosphorylation of myelin basic protein (Vartanian et al., 1986). Protein tyrosine kinases have been implicated in regulatory processes during development and differentiation (Hafen et al., 1987; Haller et al., 1987; Mellstroem et al., 1987; Geissler et al., 1988). They can be divided into two major classes: (a) the growth factor
DIFFERENTIAL PHOSPHORYLATION OF MAG ISOFORMS receptor-type protein tyrosine kinases, such as the EGF and insulin receptors; (b) the cytoplasmic protein tyrosine kinase, such as the protein products of the c-SIC and c-abl protooncogenes (Hunter, 1987).While EGF, IGF- 1, and insulin had no effect on the phosphorylation of MAG, v-src transformation significantly increased L-MAG phosphotyrosine levels. Vanadate did not raise the constitutive tyrosine phosphorylation of L-MAG in v-src-transformed cells, indicating that pp60v-srclikely phosphorylates the same tyrosine residues as the endogenous NIH 3T3 cell MAG protein tyrosine kinase. It may be that L-MAG is phosphorylated by the cellular homologue of the protein product of v-src, pp60c-src,which has been detected in myelin (Hirano et al., 1988). A potential role for tyrosine-specific protein phosphorylation in the regulation of cell-cell and/or cellmatrix interactions has been demonstrated for the sevenless homeotic gene of Drosophifa(Hafen et al., 1987), the platelet integrin-glycoproteingpIIb-IIIa (Ferrell and Martin, 1989; Golden and Brugge, 1989), and the fibronectin receptor (Hirst et al., 1986). Recently, it was demonstrated that tyrosine phosphorylated integrins, isolated from Rous sarcoma virus-transformed chick embryo fibroblasts,exhibited a decreased ability to bind fibronectin and talin compared to nonphosphorylated receptor (Tapley et al., 1989). The critical integrin phosphorylation site has been mapped to tyrosine 788, which is located in a domain responsible for binding the cytoskeleton (Tapley et al., 1989). A homologous stretch of amino acids is found in both MAG isoforms (amino acids 55 1-573; Tamkun et al., 1986;see Salzer et al., 1987). This region in MAG contains two tyrosines, one of them (tyrosine 558) nestled among amino acids similar to those surrounding tyrosine 788 of the fibronectin receptor. Since there are only two tyrosine residues in the S-MAG cytoplasmic domain, both within the integrin homology region, it is possible that this conserved region may be a protein tyrosine kinase target in two distinct cell adhesion molecules. We speculate that phosphorylation of the cytoplasmic domain may modify the interaction between MAG and certain cytoplasmic components, thus regulating MAG-specific cell adhesion functions. Alternatively, L-MAG may function as a signal transducer as has been observed for other cell adhesion molecules that are members of the immunoglobulin gene superfamily (Veillette et al., 1988; Schuch et al., 1989). The phosphorylation sites found on the unique domain of L-MAG may represent docking sites for protein kinases. Oligomenzation of L-MAG molecules through ligand binding may result in kinase activation, generation of second messenger signals, and ultimately alteration in gene expression. S-MAG, which lacks this domain, may have more restricted activities and serve only as an adhesive molecule.
1425
and D.E.H.A. is a recipient of a n FRSQ scholarship. We appreciate the help given to us by Beth Mason in preparing this manuscript.
REFERENCES Arquint M., Roder J., Chia L.-S., Down J., Wilkinson D., Bayley H., Braun P. E., and Dunn R. (1987) Molecular cloning and primary structure of myelin associated glycoprotein. Proc. Natl. Acad. Sci. USA 84,600-604. Cooper J. A., Sefton B. M., and Hunter T. (1983) Detection and quantification of phosphotyrosinein proteins. Methods Enzymol. 99,387-405.
Edwards A. M.,Arquint M., Braun P. E., Roder J. C., Dunn R. J., Pawson T., and Bell J. C. (1988) Myelin-associated glycoprotein, a cell adhesion molecule of oligodendrocytes, is phosphorylated in brain. Mol. Ce[l. Biol. 8, 2655-2658. Edwards A. M., Braun P. E., and Bell J. C. (1989) Phosphorylation of myelin-associated glycoprotein in vivo and in vitro occurs only in the cytoplasmic domain of the large isoform. J. Neurochem. 52, 317-320. Ferrell J. E. and Martin G. S. (1989) Tyrosine-specificprotein phosphorylation is regulated by glycoprotein IIb-IIIa in platelets. Proc. Natl. Acad. Sci. USA 86,2234-2238. Frail D. E. and Braun P. E. (1984) Two developmentally regulated messenger RNAs differing in their coding region may exist for myelin-associated glycoprotein. J. Biol. Chern. 259, 1485714862.
Geissler E. N., Ryan M. A., and Housman D. E. (1 988) The dominantwhite spotting (W) locus of the mouse encodes the c-kit protooncogene. Cell 55, 185-192. Golden A. and Brugge J. S. (1989) Thrombin treatment induces rapid changes in tyrosine phosphorylation in platelets. Proc. Nut/.Acad. Sci. USA 86,901-905. Graham F. L. and van der Eb A. J. (1973) A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52, 456-467.
Hafen E., Basler K., Edstroem J.-E., and Rubin G. (1987) Sevenless, a cell-specific homeotic gene of Drosophila, encodes a putative transmembrane receptor with a tyrosine kinase domain. Science 236,55-63.
Haller J., Cote S., Broenner G., and Jaeckle H. (1987) Dorsal and neural expression of a tyrosine kinase-related Drosophila gene during embryonic development. Genes Dev. 1,862-867. Hirano A. A,, Greengard P., and Huganir R. L. (1988) Protein tyrosine kinase activity and its endogenous substrates in rat brain: a subcellular and regional survey. J. Neurochem. 50, 1447-1455. Hirst R., Horwitz A,, Buck C., and Rohrschneider L. (1986) Phosphorylation of the fibronectin receptor complex in cells transformed by the oncogenes that encode tyrosine kinases. Proc. Natl. Acad. Sci. USA 83,6470-6474. Horvath E., Edwards A. M., Bell J. C., and Braun P. E. (1 989) Chemical deglycosylationon a microscale of membrane glycoproteins with retention of phosphoryl-protein linkage. J. Neurosci. Rex (in press). Hunter T. (1987) A thousand and one protein kinases. CeN 50,823829.
Johnson P. W., Abramow-Newerly W., Seilheimer B., Sadoul R., Tropak M. B., Arquint M., Dunn R. J., Schachner M., and Roder J. C. (1989) Recombinant myelin-associated glycoprotein confers neural adhesion and neurite outgrowth function. Neuron 3,377-385.
Kirschmeier P. T., Housey G. M., Johnson M. D., Perkins A. S., and Weinstein I. B. (1988) Construction and characterization of a retroviral vector demonstrating efficient expression of cloned cDNA sequences. DNA 7,219-225. Klarlund J. K. (1985) Transformation of cells by an inhibitor of phosphatases acting on phosphotyrosineproteins. CeN41,707717.
Acknowledgment: This work was supported by NCI grants to J.C.B. and P.E.B. J.C.B. is an M R C of Canada scholar
Lai C., Brow M. A., Nave K.-A,, Noronha A. B., Quarles R. A., Bloom F. E., Milner R. J., and Sutcliffe J. G. (1987) Two forms
J. Nertrochem.. Val. 55. No. 4. 1990
1426
D. E. H. AFAR ET AL.
of IB236/myelin-associatedglycoprotein, a cell adhesion molecule for postnatal neural development, are produced by alternative splicing. Proc. Natl. Acad. Sci. USA 84,4337-4341. Mellstroem K., Bjelfman C., Hammerling U., and Pahlman S. (1987) Expression of c-src in cultured human neuroblastoma and smallcell lung carcinoma cell lines correlates with neurocrine differentiation. Mol. Cell. Biol. 7, 4 178-4 184. Nishizuka Y.(1984)The role ofprotein kinase C in cell surface signal transduction and tumor promotion. Nature 308, 693-698. Poltorak M., Sadoul R., Keilhauer G., Landa C., Fahrig T., and Schachner M. (1987) Myelin associated glycoprotein, a member of the L2/HNK- I family of neural cell adhesion molecules, is involved in neuron-ohgodendrocye and oligodendrocyte-oligodendrocyte adhesion. J. Cell Biol. 105, 1893-1899. Salzer J., Holmes P. W., and Colman D. R. (1987). The amino acid sequence of the myelin-associated glycoproteins: homology to the immunoglobulin gene superfamily. J. Cell Biol. 104, 957965. Schuch U., Lohse M. J., and Schachner M. (1989)Neural cell adhesion molecules influence second messenger systems. Neuron 3, 1320. Sternberger N. H., Quarles R. H., Itoyama Y., and deF. Webster H. ( 1979) Myelin-associated glycoprotein demonstrated immuno-
J. Neurochem., Vol. 55. No. 4, 1990
cytochemically in myelin and myelin-formingcells of developing rat. Proc. Natl. Acad. Sci. USA 76, 15 10- 15 14. Tamkun J. W., DeSimone D. W., Fonda D., Pate1 R. S., Buck C., Honvitz A. F., and Hynes R. 0. (1986) Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46, 271-282. Tapley P., Honvitz A. F., Buck C., Duggan K., and Rohrschneider L. (1989) Integrins isolated from rous sarcoma virus-transformed chicken embryo fibroblasts. Oncogene 4,325-333. Tropak M. B., Johnson P. W., Dunn R. J., and Roder J. C. (1988) Differential splicing of MAG transcripts during CNS and PNS development. Mol. Brain Res. 4, 143-155. Vartanian T. S., Szuchet S., Dawson G., and Campagnoni A. T. ( 1986) Oligodendrocyte adhesion activates protein kinase Cmediated phosphorylation of myelin basic protein. Science 234, 1395-1 398. Veillette A,, Bookman M. A,, Horak E. M., and Bolen J. B. (1988) The CD4 and CD8 T-cell surface antigens are associated with the internal membrane tyrosine-protein kinase ~ 5 6 ''~Cell . 55, 301-308. Williams A. F. (1987) A year in the life of the immunoglobulin superfamily. Immunol. Today 8,298-303.
