developmentally regulated masking of an intracellular epitope of the 180 kda isoform of the neural...

Developmentally Regulated Masking of anIntracellular Epitope of the 180 kDa Isoformof the Neural Cell Adhesion Molecule NCAMIris Kramer, 1Heike Hall,1Ulrike Bleistein,1 and Melitta Schachner2*1Department of Neurobiology, Swiss Federal Institute of Technology, Hoenggerberg,Zurich, Switzerland2Zentrum fur Molekulare Neurobiologie Hamburg, Universita¨t Hamburg, Hamburg, Germany

The neural cell adhesion molecule NCAM is a cellsurface glycoprotein that occurs in several isoforms. Itwas previously shown that the largest 180-kDa iso-form of NCAM (NCAM 180) accumulates at sites ofcell contact and in postsynaptic densities and may beresponsible for the stabilization of cell contacts by itsinteraction with the membrane-cytoskeleton linkerprotein brain spectrin. In immunohistochemical stud-ies on the expression of the NCAM 180, we noticedthat two NCAM 180 specific monoclonal antibodies,termed 481 and D3, showed different patterns ofimmunoreactivity in sections of fresh-frozen adultmouse brain. Here we show that the D3-specific, butnot the 481-specific epitope becomes inaccessible tothe antibody during development of the hippocampalformation, coincident with the establishment of stablecell-cell contacts. In contrast, in the olfactory bulbwith its continually regenerating olfactory nerve fi-bers, both NCAM 180 antibodies remain fully immu-noreactive throughout development and adulthood.We also show that the D3-specific epitope becomesinaccessible in primary cerebellar neuron cultureswith time in culture. Electrophoretic separation ofhippocampal membrane proteins under nondenatur-ating conditions showed NCAM to be present inprotein complexes of different molecular weights atdifferent developmental stages. We propose thatNCAM is involved in the formation of developmen-tally regulated, noncovalent complexes with as yetunknown partner molecules that could be responsiblefor the masking of the D3-specific epitope. J. Neurosci.Res. 49:161–175, 1997.r 1997 Wiley-Liss, Inc.

Key words: adhesion molecule; cell surface glycopro-tein; development; hippocampus; NCAM; NCAM180

INTRODUCTIONThe neural cell adhesion molecule NCAM is a cell

surface glycoprotein that may play important roles in theformation, maintenance, and regeneration of the nervous

system. It is expressed on almost all neural cells from thetime of neural induction and is involved in neuron-neuron, neuron-glia, and glia-glia adhesion, both in thecentral and peripheral nervous system (for reviews, seeEdelman et al., 1990; Schachner, 1991). NCAMmediatesneurite outgrowth, fasciculation, and branching (Rut-ishauser and Edelman, 1980; Bixby et al., 1987; Land-messer et al., 1988, 1990; Doherty et al., 1990; Me`ge etal., 1991; Frei et al., 1992) and is involved in retinal andtectal histogenesis and formation of the neuromuscularjunction (Buskirk et al., 1980; Fraser et al., 1984; Thanoset al., 1984; Covault and Sanes, 1986). It has also beenimplicated in synaptic plasticity (Lu¨thi et al., 1994; Ronnet al., 1995; Scholey et al., 1993). It acts via calcium-independent homophilic and heterophilic binding mecha-nisms, involving binding sites on the extracellular part ofthe molecule (for reviews, see Cunningham et al., 1987;Rutishauser and Jessell, 1988). The resulting cell surfacetrigger appears to be transduced from the cell surface tothe cell interior by activation of second messengersystems (Schuch et al., 1989; Doherty et al., 1991; vonBohlen und Halbach et al., 1992; Williams et al., 1994).

NCAM belongs to the immunoglobulin superfamilyand appears in various, closely related isoforms that are

Abbreviations: CA, Ammon’s horn of the hippocampal formation;CNS, central nervous system; EDTA, ethylenediaminetetraaceticacid;EGTA, ethyleneglycol-bis(beta-aminoethylether)-tetraacetic acid; FITC,fluorescein isothiocyanate; IgG, immunoglobulin G; NCAM, neuralcell adhesion molecule; NCAM 180, 180 kDa isoform of NCAM;NCAM 140, 140 kDa isoform of NCAM; NCAM 120, 120 kDaisoform of NCAM; PAGE, polyacrylamide gel electrophoresis; PBS,phosphate-buffered saline, pH 7.3; pEX 180, fusion protein containingthe sequence of the NCAM 180-specific domain; SDS, sodiumdodecylsulfate.

*Correspondence to: Melitta Schachner, Zentrum fu¨r MolekulareNeurobiologie Hamburg, Universita¨t Hamburg, Martinistraße 52,D-20246, Hamburg, Germany. E-mail: [email protected] [email protected]

Received 7 October 1996; Revised 25 February 1997; Accepted 26February 1997

Journal of Neuroscience Research 49:161–175 (1997)

r 1997 Wiley-Liss, Inc.

generated from a single gene by alternative splicing(Cunningham et al., 1987). It consists of three majorisoforms of 120, 140, and 180 kDa apparent molecularweights (hereafter designated NCAM 120, NCAM 140,and NCAM 180, respectively), which are characterizedby identical extracellular domains (Fig. 1; Owens et al.,1987; Santoni et al., 1987) containing five immunoglobu-lin-like domains and two fibronectin type III homologousrepeats (Barthels et al., 1987; Cunningham et al., 1987).NCAM 120 lacks an intracellular domain and is anchoredto the membrane via a phosphatidylinositol residue(Sadoul et al., 1986; He et al., 1987). NCAM 140 and 180share identical membrane-spanning and cytoplasmic do-mains, but the NCAM 180 molecule contains an addi-tional cytoplasmic domain of approximately 30 kDa(Murray et al., 1986) encoded by exon 18 in the mouse(Barbas et al., 1988). The prevalence of the isoformsdiffers during neural development (for review, see Edel-man et al., 1990). NCAM 180, for instance, seems to beexclusively expressed on neurons. It shows a reduction oflateral mobility in the cell membrane (Pollerberg et al.,1986) and a tendency to concentrate at sites of cellcontact (Pollerberg et al., 1985, 1987, 1991) and atpostsynaptic densities (Persohn et al., 1989). Theseobservations led to the suggestion that NCAM 180 is ableto stabilize cell contacts and influence the communicationof pre- and postsynaptic partners of a synapse.

In addition, the association of NCAM with itself orother adhesion molecules in the cell membrane wassuggested to modulate the function of the involvedmolecules. For instance, a close molecular associationbetween NCAM and L1 mediated by extracellular carbo-hydrate residues is thought to increase L1-dependent,homophilic cell binding (Kadmon et al., 1990a,b; Simonet al., 1991; Horstkorte et al., 1993). Also, clusters ofNCAM containing NCAM 140 and 180 but not NCAM120 have been described in electron microscopic studies(Hall and Rutishauser, 1987; Becker et al., 1989). It hasbeen speculated that a series of cysteine residues on thecytoplasmic side of NCAM 180 and NCAM 140, close tothe membrane-spanning region, may form covalent bondswith one another (Sorkin et al., 1984, 1985). However, itis still far from being understood how the differentisoforms and posttranslational modifications of NCAMas well as the association of the molecule with itself orother membrane proteins contribute to its functionalproperties and role in the formation of neural connec-tions.

We decided to investigate further the molecularproperties of the cytoplasmic domain of NCAM 180 withrespect to its interaction with other molecules. With thehelp of two antibodies specific for this domain, we foundthat the two, nonoverlapping NCAM 180-specific epi-topes are differentially accessible to the antibodies atdifferent developmental stages in the mouse centralnervous system in vivo and differentiating neurons invitro. Furthermore, we show that NCAM containingprotein complexes from mouse hippocampus, separatedby gel electrophoresis under nondenaturing conditions,change during development.

MATERIALS AND METHODSAnimals

For all experiments, NMRI mice obtained fromZentrale Tierzuchtanlage, Hannover, and ICR mice ob-tained from Institut fiir Labortierkunde of the UniversityZurich were used.Animals were taken at 0, 5, 6, 7, 10, 15,20, and 30 days of age. Adult animals were at least 10weeks old.

AntibodiesNCAM 180-specific monoclonal antibodies D3

from mouse and 481 from rat have been described(Pollerberg et al., 1985; Schlosshauer, 1989). In thepresent study, NCAM 180 specificity in mouse wascharacterized by Western blot analysis of membraneproteins of mouse hippocampi as well as by using fusionproteins of the NCAM 180-specific domain expressed bybacteria. The monoclonal antibody H28.123 from rat

Fig. 1. Schematic presentation of the molecular structure of thethree major isoforms of NCAM, NCAM 180, NCAM 140, andNCAM 120, characterized by identical extracellular (ex) butdifferent membrane anchoring and intracellular (in) domains.NCAM 180 contains a cytoplasmic domain encoded by exon 18in the mouse, which is recognized by the two monoclonalantibodies D3 and 481. The arrows indicate the position of theepitopes recognized by these antibodies.

162 Kramer et al.

against NCAM and the polyclonal NCAM antibodiesfrom rabbit have been shown to react with the three majorisoforms of NCAM (Faissner et al., 1984; Gennarini etal., 1984). The polyclonal antibodies against L1 (Faissneret al., 1985) and brain spectrin (240/23SE; Riederer et al.,1986) were used as described. Secondary antibodiescoupled to fluorescein isothiocyanate (FITC), tetrameth-ylrhodamine isothiocyanate (TRITC), or alkaline phospha-tase were obtained from Jackson ImmunoResearch (Di-anova, Hamburg, Germany).

Generation of Fusion ProteinsFusion proteins were obtained either as the entire

NCAM 180-specific domain encoded by exon 18 of themouse (designated pEX 180) or two complementary partsof this domain (designated pEX 180 sp and pEX 180 pp)in conjunction withb-galactosidase (Stanley and Lutzio,1984). To generate pEX 180, the NCAM 180-specificsequence Sph I (2092)–Pst I (2958) (Barbas et al., 1988)was subcloned into the Sma I/Pst I site of the vector pEX1after repair of the Sph I site. To generate pEX 180 sp thefragment Sph I (2092)–Pvu II (2628) was subcloned intothe SmaI site of the vector pEX1 after repair of the Sph Isite. For the generation of pEX 180 pp the fragment PvuII (2628)–Pst I (2958) was subcloned into the Sma I/Pst Isite of the pEX1 vector. The calculated molecular weightsof the resultant fusion proteins are 148 kDa (pEX 180),136 kDa (pEX 180 sp), and 128 kDa (pEX 180 pp).

ImmunohistochemistryFor indirect immunofluorescence microscopy,

freshly dissected hippocampus, cerebral cortex, cerebel-lum, and olfactory bulb of mouse were embedded inTissue-Tek (Miles), frozen in liquid nitrogen, and cryosec-tioned (10 µm). Cryosections were stained as described(Goridis et al., 1983). Detergent- or sucrose-treatedcryosections were prepared as follows. Prior to staining,sections were washed in phosphate-buffered saline (PBS),pH 7.3, transferred to solutions (all in PBS) of eithersucrose (5, 7.5, or 10%), n-decylsucrose (3 mg/ml;Calbiochem, La Jolla, California), n-dodecylmaltoside (5mg/ml), n-dodecylglucoside (0.1 mg/ml), n-octylgluco-side (5 mg/ml), Triton X-100 (0.1%), or CHAPS (6mg/ml) (all from Boehringer Mannheim) for 10 min andwashed three times with PBS. Alternatively, the speci-mens were fixed in a solution of 4% paraformaldehyde inPBS for 20 min prior to detergent treatment. Nonspecificcross-reaction of the secondary antibodies was controlledfor by incubation of sections with 1% bovine serumalbumin instead of primary antibody.

Double and single immunofluorescence labeling ofcultured cells using monoclonal and polyclonal antibod-

ies was conducted according to Schnitzer and Schachner(1981) with the following modifications: NCAM 180-specific intracellular epitopes were made accessible to theantibodies D3 and 481 by treatment of live cultures withethanol for 5 min at215°C followed by a rinse with PBSand fixation in 4% paraformaldehyde.

Cell Culture

Suspensions of single cerebellar cells from 6- to7-day-old mice were obtained by a combination ofproteolytic and mechanical disruption of small pieces oftissue (Schnitzer and Schachner, 1981). Populations en-riched in small cerebellar neurons were obtained from thesingle cell suspension by centrifugation through Percoll(Keilhauer et al., 1985). Cells were plated at a density of13 106 cells per glass coverslip (15 mm in diameter) andmaintained in culture for 1, 2, 3, 5, 7, or 10 days.Coverslips were either coated with poly-L-lysine (Sch-nitzer and Schachner, 1981), laminin (20 µg/ml; Boeh-ringer Mannheim), or rat tail collagen as described byBornstein (1958).

Western Blot Analysis

Freshly dissected hippocampi from 5-, 15-, and30-day-old mice were homogenized with a Teflon homog-enizer in 1 mM NaHCO3, pH 7.9, containing 0.2 mMCaCl2, 0.2 mMMgCl2, and 1 mM spermidine (homogeni-zation buffer) at 4°C. The homogenates were centrifugedat 200 g and 4°C for 20 min, the supernatant removed andfurther centrifuged at 30,000g and 4°C for 30 min. Theresultant pellet, containing membrane fragments, wassolubilized in 0.5% Triton X-100 in 20 mM Tris-HClbuffer, pH 7.4, containing 0.15 M NaCl, 1 mM EDTA, 1mM EGTA, and the protease inhibitors aprotinin (5µg/ml), soybean trypsin inhibitor (10 µg/ml), phenylmeth-ylsulfonylfluoride (1 mM), and iodoacetic acid (0.5 mM)(all from Boehringer Mannheim) and stirred for 2 hr at4°C followed by centrifugation at 100,000g for 45 min at4°C. The resultant supernatant contained the detergentextracts of membrane proteins.

For separation of proteins by electrophoresis, 50 µgof proteins were loaded in each lane and separated bySDS-PAGE (7% gels) as described (Laemmli, 1970).Western blot analysis was conducted according to Kyhse-Andersen (1984). After transfer of proteins to nitrocellu-lose filters, they were incubated with primary antibodies,washed several times, incubated with alkaline phospha-tase-conjugated secondary antibodies, and developedwith paranitrobluetetrazolium chloride (50 mg/ml in 70%dimethyl formamide) and 5-bromo-4-chloro-3-in-dolylphosphate-toluidine salt (50 mg/ml dimethyl for-

Developmental Masking of NCAM 180-Specific Domain 163

mamide) in 0.1 M Tris-HCl buffer, pH 9.6, containing 0.1M NaCl and 5 mMMgCl2.

Membrane preparations of cultured cells were ob-tained according to Goridis et al. (1983) with the follow-ing modifications: Plated cells were washed with PBS,suspended in ice-cold 50 mM Tris-HCl buffer, pH 8.0,containing 1 mM CaCl2 and the protease inhibitors usinga rubber policeman and homogenized in a small glasshomogenizer for 15 min on ice. NP-40 (Sigma) wasadded to a final concentration of 1%. Homogenates werecentrifuged for 10 min at 500g and 4°C, and thesupernatant was taken for Western blot analysis. Fiftymicrograms of protein were separated by SDS-PAGE(6% gels). Subsequent Western blot analysis was per-formed as described above.

Polyacrylamide Gel Electrophoresis UnderNondenaturing Conditions

Membrane proteins from tissue homogenates of5-day-old and adult mouse hippocampi were extracted in0.5% Triton X-100 in 20 mM Tris-HCl buffer, pH 7.4,containing 0.15 M NaCl, 1 mM EDTA, 1 mM EGTA, andprotease inhibitors (as above). Electrophoresis was per-formed on polyacrylamide slab gels (1103 2003 1.5mm) in a vertical slab gel apparatus (Bio-Rad) for 24 hr at4°C, applying a constant current of 40 mA. For thepreparation of the gels either 4, 4.5, 5, or 5.5% acrylamidesolutions weremixed with a 0.8%N,N-methylenebisacryl-amide solution in 0.3 M Tris-HCl, pH 8.8, and 0.1%Triton X-100. To determine the molecular weights of theproteins or protein complexes separated, four gels withdifferent polyacrylamide concentrations (4, 4.5, 5, or5.5%) were used. Samples containing either membraneextracts (50 µg protein) or the molecular weight standardsthyroglobulin (669 kDa), urease (dimer of 272 kDa andtrimer of 545 kDa),b-amylase (200 kDa), bovine serumalbumin (monomer 66 kDa and dimer 132 kDa), carbonicanhydrase (29 kDa) or apoferritin (monomer of 443 kDaand dimer of 900 kDa) (4–10 µg) were loaded onto thegel. After electrophoresis, proteins were transferred toProBlot membranes (Applied Biosystems, Frankfurt, Ger-many), and Western blot analysis was performed underthe same conditions as described in the previous para-graph for nitrocellulose filters. The molecular weights ofthe complexes were estimated according to the technicalbulletin MKR-137 (10-86) of Sigma Chemical Company.In brief, the relative mobility of each standard protein wasplotted against the gel concentration (4, 4.5, 5, or 5.5%acrylamide) and the negative slopes of these graphs wereplotted against the known molecular weights of thestandards on a bilogarithmic scale.

RESULTSCharacterization of the Antibodies D3 and 481 UsingRecombinant Protein Fragments of NCAM 180

Themonoclonal antibodies D3 and 481 react specifi-cally with the 180-kDa component of mouse NCAM byWestern blot analysis (Figs. 2 and Fig. 3B,C). With theuse of two complementary fragments (pEX 180 sp andpEX 180 pp) of the NCAM 180-specific domain (pEX180), antibody D3 recognizes an epitope on pEX 180 spand pEX 180. Antibody 481 recognizes an epitope onpEX 180 pp and pEX 180 (Fig. 2). The epitope of the D3antibody on pEX 180 sp is situated more proximal to thecell membrane spanning region (Fig. 1). The lowermolecular weight bands stained by the antibodies areproteolytic degradation products of the fusion proteins.

Western blot analysis of hippocampal membranepreparations using polyclonal NCAM antibodies showedthe three major isoforms of NCAM in 5- and 15-day-oldand adult mice (Fig. 3A). Because of the higher polysialicacid content of NCAM in young animals, a diffuselystained zone in the range of 180–250 kDa was visible atthese ages. Western blot analysis of 5-day-old and adultmouse hippocampus using D3 and 481 showed one bandof 180 kDa apparent molecular weight (Fig. 3B). Western

Fig. 2. Immunochemical characterization of the antibodies D3(A) and 481 (B) on recombinant protein fragments of mouseNCAM 180 by Western blot analysis. In (A) the antibody D3reacts with the pEX 180 sp (lane 2) and pEX 180 (lane 3)fragments but not with the pEX 180 pp (lane 1) fragment ofNCAM 180. Bands of 140 kDa (pEX 180 sp) and 150 kDa(pEX 180) are visible. In (B) the antibody 481 reacts with thepEX 180 pp (lane 1) and the pEX 180 (lane 3) fragments,whereas the pEX 180 sp fragment is not detected (lane 2).Bands of 130 kDa (pEX pp) and 150 kDa (pEX 180) are visible.Multiple bands with lower molecular weights than the calcu-lated molecular weights of the fusion proteins are stained by theantibodies and are most likely proteolytic degradation productsof the fusion proteins. Molecular weight standards of 180 and116 kDa are marked.

164 Kramer et al.

blot analysis of detergent extracts of membrane prepara-tions of small cerebellar neurons after 2 and 10 days ofculture (Fig. 3C) showed NCAM 180 using D3 and 481.As in the hippocampal membrane preparations, the highmolecular weight band in 2-day-old cell cultures is likelydue to a higher content of polysialylated NCAM 180 (Fig.3B,C).

Immunohistochemical Detection of NCAM 180in the Hippocampal Formation by the AntibodiesD3 and 481

NCAM 180 was detected by indirect immunofluo-rescence in cryosections of 5-, 15-, and 30-day-old mousehippocampus by the antibodies D3 and 481. D3 showedage-dependent differences in fluorescence intensity andpattern of immunoreactivity of NCAM 180 in the hippo-campus (Fig. 4). In 5-day-old mice (Fig. 4A–C), D3uniformly stained all layers of the hippocampus properand dentate gyrus with a surface membrane-associatedstaining of cell bodies and neuropil. In 15-day-old mice

(Fig. 4D–F), the overall fluorescence intensity was slightlyreduced and areas, such as the molecular layer and hilusregion of the dentate gyrus as well as the mossy fibers andparts of the stratum radiatum and lacunosum of thehippocampus, were more prominently stained than theother layers. At higher magnifications the staining patternappeared punctate. In 30-day-old mice (Figs. 4G–I andFig. 5A), fluorescence intensities were reduced to almostbackground levels. The pyramidal cell layer, the stratumoriens, and the stratum radiatum of the hippocampus weremore affected by this loss of immunoreactivity than thegranular and molecular layers of the dentate gyrus or thehilus and the stratum lacunosum of the hippocampus (Fig.5A). In 30-day-old mice, the cell bodies of the pyramidalcells appeared no longer positive, whereas the neuropilstill showed immunoreactivity. The staining pattern re-mained punctate, but the number and intensity of pointswere markedly decreased (Fig. 4I). The 481 revealed thesame staining pattern as D3 in hippocampus and dentategyrus of 5- and 15-day-old mice (not shown). However,

Fig. 3. Immunochemical characterization of NCAM isoformsin mouse hippocampus (A, B) and cultured cerebellar neurons(C) by Western blot analysis. In (A) detergent extracts ofmembrane preparations of hippocampi of 5- (lane 1), 15- (lane2), and 30-day-old (lane 3) mice, the three isoforms of NCAMwith 120, 140, and 180 kDa are detected by using polyclonalNCAM antibodies (6% gels (1003 1 mm)). In (B) detergentextracts of membrane preparations of hippocampi of adult(lanes 1and3) and 5-day-old (lanes 2and4) mice stained with

the monoclonal antibodies D3 (lanes 1 and 2) and 481 (lanes 3and 4) show one band of 180 kDa in all lanes. In (C) membranepreparations of small cerebellar neurons cultured for 10 (lanes1 and3) and 2 days (lanes 2and4) are stained with D3 (lanes 1and 2) and 481 (lanes 3 and 4). In all lanes the antibodies detecta band of 180 kDa. In (B) and (C) 7% gels (503 0.5 mm) areshown. Molecular weight standards of 180 and 116 kDa aremarked by arrowheads.


Fig. 4. Immunofluorescence labeling of cross sections of 5-(A–C), 15- (D–F), and 30-day-old (G–P) mouse hippocampalformation with D3 (A–M) and 481 (N–P). Micrographs of thedentate gyrus (A, C, D, F, G, I, K, M, N, P) and the CA3 regionof the hippocampus (B, E, H, L, O) are presented. In 5-day-oldanimals, NCAM 180 is uniformly distributed in all layers of thedentate gyrus (A, C) and CA3 region (B). The immunofluores-cence shows surface membrane-associated staining of the cellbodies and neuropil (C). In 15-day-old animals (D–F), immuno-fluorescence is also detectable in all layers. The molecular layer(m) and the hilus region (h) of the dentate gyrus (D) as well asthe mossy fibers (mf) of the CA3 region (E) are moreprominently stained than the other layers. The pyramidal cells(p) of the CA3 region (E) are less immunoreactive at this stage.At higher magnification, the staining pattern appears punctate

(F). In 30-day-old animals immunofluorescence with the anti-body D3 (G–I) is reduced to almost background levels. Theresidual staining pattern is still punctate (I). (K–M) and (N–P)show immunolabeling of cross sections of 30-day-old animalswith 481 on a fresh-frozen section (K–IVf) and D3 aftern-decylsucrose treatment of the sections (N–P) in comparison.Staining patterns are identical and show immunoreactivity inthe dentate gyrus around the granule cell bodies (g), in themolecular layer (m), and in the hilus (h) (K, N). In the CA3region prominent staining of the mossy fibers (mf) is seen (L,O). At higher magnification (M, P) the staining pattern appearspunctate. All micrographs were taken with the same exposuretime. Scale bars5 100 µm in O (for A, B, D, E, G, H, K, L, andN) and 10 µm in P (for C, F, I, and M).

in 30-day-old mice 481 stained more strongly than D3(Figs. 4K–M and Fig. 5E) with an intensity comparablewith that of 15-day-old mice.

In summary, the immunohistochemical and immu-nochemical observations indicate that NCAM 180 re-mains present in the hippocampus at later developmentalstages and can be detected by Western blot analysis aswell as by immunohistochemistry using 481. However,the D3-specific epitope becomes less accessible to theantibody as development proceeds.

Immunohistochemical Detection of NCAM 180 in theHippocampal Formation After Detergent Treatment

To unravel the discrepancy between the ability ofthe antibody D3 to recognize NCAM 180 byWestern blotanalysis but not by immunohistochemistry, cryosectionsof 30-day-old mouse hippocampus were treated withdifferent detergents prior to incubation with antibody. Incontrast to the untreated cryosections (Figs. 4G–I and5A), D3 stained the sucrose (Fig. 5B)- and n-decylsu-crose (Figs. 4N–P and 5C)-treated cryosections withintense immunofluorescence in all layers of the hippo-campus. Staining pattern and fluorescence intensitywere identical to those obtained with antibody 481 inuntreated and n-decylsucrose treated cryosections (Figs.4K–M and 5E,F). Treatment with n-dodecylmaltoside,n-dodecylglucoside, and n-octylglucoside yielded slightlyreduced immunofluorescence intensities with D3 com-pared with those obtained with sucrose or n-decylsucrose(Table I). This could be due to a lower efficacy of thedetergent, stronger solubilization of tissue in detergent, ordamage of the antigen. The nonionic detergent TritonX-100 and the zwitter-ionic detergent CHAPS didnot increase the ability of D3 to react with its epitope(Table I).

Cryosections that had been fixed with 4% para-formaldehyde prior to n-decylsucrose treatment alsoshowed no specific immunolabeling with D3 (Fig. 5D).Nonspecific reactivity of the secondary antibodieswas not seen in the untreated or in the detergent-treated cryosections (Fig. 5G). However, after fixa-tion in 4% paraformaldehyde, the secondary antibodiesagainst mouse IgG, in contrast to the secondary anti-bodies against rat or rabbit IgG, showed some non-specific reactivity with blood vessels and meninges (Fig.5H).

Immunohistochemical Detection of NCAM 180 inOther Parts of the CNS Using the AntibodiesD3 and 481

The loss of immunolabeling of NCAM 180 by D3in the adult animal was not only observed in thehippocampus but also in other brain regions, such as the

cerebral cortex, the thalamus, and the cerebellum (TableII) and most remarkably between postnatal days 20 and30. In all three regions, treatment of the cryosections withsucrose or detergent resulted in strong immunoreactivitycomparable with that of 481 (not shown).

The olfactory bulb was an exception to the generalphenomenon of masking of the D3-specific epitope withincreasing age (Table II). In the olfactory nerve of anadult mouse the NCAM isoforms have a differentialspatial distribution (Miragall et al., 1988, 1989): Using481 NCAM 180 is almost exclusively expressed on theincoming olfactory nerve fibers and the glomerular layer,whereas NCAM 140 and 120 are detectable in all layers.When fresh-frozen cryosections of the adult olfactorybulb were stained with D3 (Fig. 6A) and 481, identicalstaining patterns were obtained (not shown), with intenseimmunolabeling of the incoming olfactory nerve fibers,as well as the glomerular layer. The staining patternsobtained with n-decylsucrose-treated cryosections (Fig.6B) were identical to the untreated cryosection using D3.

Immunocytological Detection of NCAM 180 inCultured Neurons

Cultures of small cerebellar neurons from 6- to7-day-old mice were used to investigate the expressionand distribution of NCAM 180 after 1, 3, and 7 days inculture by double immunostaining with D3 and poly-clonal NCAM antibodies (Fig. 7A,C,E). After 1 day inculture on poly-L-lysine, cells adhered well to the sub-strate and neurites had extended (Fig. 7A9). The immuno-reactivity obtained with D3 (Fig. 7A) and polyclonalNCAM antibodies (Fig. 7A8) was equally distributed onthe surface of all cell bodies, neurites, and growth cones.After 3 and 7 days in culture, a dense network of neuritesbetween cell bodies had developed (Fig. 7C9,E9). Thestaining with the polyclonal NCAM antibodies showedno difference in 3- and 7-day-old cell cultures comparedwith 1-day-old cultures (Fig. 7A8,C8,E8), whereas theimmunofluorescence intensity given by the antibody D3decreased to almost background levels after 3 days inculture and remained undetectable after 7 days in culture(Fig. 7A,C,E). Immunoreactivity with 481 in 5- and10-day-old cultures remained as intense as after 1 day inculture (Fig. 7B,D,F and B8,D8,F8). Immunoreactivityalways appeared associated with the cell surface. Similarresults were obtained when cerebellar neurons wereplated on laminin or rat tail collagen (not shown).

These observations confirm the results found invivo in that NCAM 180 is expressed by neurons at anytime in culture but that the D3-specific epitope, incontrast to the 481-specific epitope, becomes less acces-sible to the antibody with time in culture.


Figure 5

168 Kramer et al.

Electrophoretic Separation of Protein ComplexesUnder Nondenaturing Conditions and Identificationof NCAM by Western Blot Analysis

It has been shown previously that membrane pro-teins can be extracted in the absence of SDS under

so-called nondenaturing conditions, preserving noncova-lent interactions among proteins (Ku¨hn et al., 1983;Schubert et al., 1983; Barbero et al., 1984; Hoessli andRungger-Brandle, 1985). We, therefore, performed poly-acrylamide gel electrophoresis in the presence of thenonionic, nondenaturing detergent Triton X-100 insteadof SDS followed by Western blot analysis, using proteinextracts of 5-day-old and adult mouse hippocampi. ForWestern blot analysis, proteins were transferred to ProB-lot membranes, which show a high capacity for bindingof high molecular weight protein complexes. PolyclonalNCAM antibodies showed developmentally regulateddifferences in the apparent molecular weights of theimmunoreactive complexes (Fig. 8A). In blots of 5-day-old animals two clearly demarcated bands, a faintlystained band of molecular weight 140 kDa and a moreprominent, broad band of 250–280 kDa were seen, whichare consistent with the molecular weights of the NCAMisoforms present at this age. Blots of adult hippocampishowed one broad band with an estimated molecularweight of 400 to 600 kDa.

These results indicate that in the adult hippocam-pus, NCAM engages in molecular associations withapparent molecular weights of at least 400–600 kDa. Itseems unlikely that the complexes found in this studyrepresent an artificial association of proteins in the TritonX-100 solution because the complexes are stable in arange of Triton X-100 concentrations from 0.1 to 0.5%. Itis possible that some complexes are not stable under theextraction conditions chosen, so that we isolated only asubclass of the existing complexes.

Western blot analysis of 5-day-old and adult hippo-campal membrane protein fractions using polyclonal L1antibodies showed one band at both ages different fromthe one revealed by the NCAM antibodies (Fig. 8B),indicating that L1 is not isolated as part of the proteincomplexes containing NCAM. Brain spectrin could notbe detected in these blots (not shown), which is consistentwith the notion that it is not released by detergent under

TABLE I. Intensity of Immunoreactivity Obtained With theAntibody D3 by Indirect Immunocytochemistry of Fresh-FrozenSections of Adult Mouse Hippocampus and Olfactory Bulb Withor Without Treatment With Different Detergents *

Treatment ofCryosection Priorto Staining Hippocampus Olfactory Bulb

None 1/2 111Sucrose 111 111n-Decylsucrose 111 111n-Dodecylmaltoside 11 11n-Dodecylglucoside 11 11n-Octylglucoside 11 11Triton X-100 1/2 111CHAPS 1/2 111

*Relative intensity of fluorescence is indicated by1/2 for minimaland111 for maximal immunoreactivity.

TABLE II. Intensity of Immunoreactivity Obtained With theAntibodies D3 and 481 by Indirect Immunocytochemistry ofFresh-Frozen Sections of Different Brain Regions*

Brain Region Antibody5-day-oldMouse

20-day-oldMouse

AdultMouse

Hippocampus D3 111 11 1/2481 111 11 11

Cortex D3 111 11 1481 111 11 11

Cerebellum D3 111 1 1/2481 111 11 11

Thalamus D3 111 1/2 2481 111 11 1

Olfactory bulb D3 111 111 111481 111 111 111

*Relative intensity of fluorescence is indicated by1/2 for minimaland111 for maximal immunoreactivity. A minus sign indicates nodetectable staining.

Fig. 5. Immunofluorescence labeling of cross sections of30-day-old mouse hippocampal formation using D3 (A-D) and481 (E, F) on nontreated (A, E,G, H) and sucrose (B) orn-decylsucrose treated (C, D, F) sections. Labeling with theantibody D3 on fresh-frozen cryosections (A) shows a generallyfaint immunoreactivity with higher reactivity in the molecularlayer (m) and hilus (h) of the dentate gyrus and stratumlacunosum (I ) of the hippocampus, whereas the strata oriens(a), pyramidale (p) and radiatum (r) are almost not immunoreac-tive. In contrast, the sucrose (B) and n-decylsucrose (C)-treatedsections stained with D3 and the fresh-rozen (E) and n-decylsucrose (F)-treated sections stained with 481 show intenseimmunofluorescence over the whole section with a slightly-

more intense fluorescence of the stratum lacunosum (I) and themossy fibers (mf) of the hippocampus and the molecular layer(m) and the hilus (h) of the dentate gyrus. In (D) the cryosectionwas fixed in formaldehyde prior to n-decylsucrose treatmentand staining with D3. The staining pattern is similar to that ofD3 in fresh-rozen sections (A) but shows nonspecific labelingof blood vessels and meninges. Differences in fluorescenceintensities in micrographs (A) and (D) are not significant.Control sections show no nonspecific binding of secondaryantibody to fresh-frozen tissue (E), but on sections pretreatedwith paraformaldehyde blood vessels and meninges (arrow-heads) are nonspecifically labeled (H). Scale bar5 200 µm.


these conditions but remains cytoskeleton associated inthe membrane pellet.

DISCUSSIONOur results show that during early postnatal devel-

opment of the mouse central nervous system with theexception of the olfactory bulb, the epitope on NCAM180 recognized by monoclonal antibody D3 undergoeschanges in a way that it is no longer accessible to theantibody at later developmental stages and in the adult.The same phenomenon was also seen when epitopeaccessibility was investigated in small cerebellar neuronsin vitro with increasing time in culture. However,Westernblot analysis and immunohistochemical detection of themolecule with polyclonal NCAMantibodies andmonoclo-nal antibody 481, another NCAM 180-specific antibody,revealed that NCAM 180 is still expressed at laterdevelopmental stages, although the amount of NCAM180 decreases during development (see also Gennarini etal., 1986; Goldowitz et al., 1990). Furthermore, the

D3-specific epitope can be made accessible to the anti-body after pretreatment of cryosections with sucrose ornonionic detergents consisting of a mono- or disaccharidecoupled to an alkane. These observations indicate that

Fig. 6. Immunofluorescence labeling of cross sections of adult mouse olfactory bulb with D3 innontreated (A) and n-decylsucrose treated (B) cryosections. Both sections show the samepattern of immunoreactivity. Incoming olfactory nerve fibers (ON) and the glomerular layer(GL) are intensely labeled, whereas the external plexiform layer (EP), the layer of the mitral cellbodies (M) as indicated by broken lines, and the granular cell layer (GR) are almost negative.Scale bar5 100 µm.

Fig. 7. Double-immunofluorescence labeling of cultured smallcerebellar with D3 (A, C, E) and polyclonal NCAM antibodies(A8, C9, E8) as well as labeling of NCAM 180 with 481 (B, D,F). Neurons were maintained in culture for 1 (A, A8, A9), 3 (C,C8, C9), and 7 (E, E8, E9) days or 1 (B, B8), 5 (D, D8), and 10 (F,F8) days. After 1 day in culture (A, A8, A9, B, B8) NCAM 180and NCAM as detected by D3 (A), 481 (B), and polyclonalNCAM antibodies (A8) are present on all neurons with auniform cell surface-associated distribution, including neuritesand growth cones. After either 3 (C, C8, C9) and 7 (E, E8, E9)days or 5 (D, D8) and 10 (F, F8) days in culture, theimmunofluorescence intensity obtained by labeling with D3decreases to almost background levels (C, E), whereas labelingwith polyclonal NCAM antibodies (C8, E8) and 481 (D, F)remains intense. A9, C9, E9, B8, D8, F8 show correspondingcontrast micrographs tofluorescence images. Scale bar5 20 µm.

170 Kramer et al.

Figure7


NCAM 180 contains an intracellular domain that be-comes masked during development, possibly by associa-tion of NCAM 180 with other molecules.

Other cell surface molecules (NILE/L1, TAG-1,and the antigen recognized by the antibody 69A1) havebeen described to show discrepancies between detectabil-ity in immunohistochemistry, Western, and Northern blotanalysis (Pigott and Kelly, 1984; Furley et al., 1990;Prince et al., 1991). The absence of immunoreactivitywas proposed to be due to either a problem of sensitivityof the immunocytochemical methods, posttranslationalmodifications of the antigen, release of the molecule fromthe membrane, or masking of the epitope recognized bythe antibody by complex formation with other proteins.However, none of these alternatives was investigatedfurther.

It could be argued that the ability of antibody D3 todetect its epitope after detergent treatment is due to theremoval of penetration barriers in the tissue (Williams,1977). However, this argument cannot explain the differ-ential immunoreactivies of the D3- and 481-specificepitopes because both antibodies are of the IgG class andshowed no penetration problems under the conditionschosen. In addition, Triton X-100 and CHAPS, detergentsthat readily permeabilize cells, were not able to enhanceimmunoreactivity for the two antibodies. A covalent,posttranslational modification of the epitope or a yetunknown, developmentally regulated, alternatively splicedform of the NCAM 180-specific intracellular domain alsocannot explain the differential immunoreactivity of anti-body D3 in cryosections vs. Western blot analysis,because the antibody would then be expected to failrecognizing its epitope in Western blots.

Our results suggest a change in the presentation ofthe epitope to the D3 antibody, which is noncovalent andunmasked by SDS, which separates and unfolds theproteins and thereby leads to free access of the antibodyto the epitope by Western blotting. The same unmaskingeffect could be achieved by treatment of cryosectionswith sucrose and mono- or dissacharide-related alkanes.Sucrose and alkyl mono- or dissacharides most probablyaffect the water structure of the NCAM molecule’ssurface by amphipathic interactions. The nonionic deter-gent Triton X-100 and the zwitter-ionic detergent CHAPSwere ineffective.

The molecular mechanisms underlying the maskingof the D3-specific epitope could be several. To supportthe notion that NCAM is involved in protein complexformation in the cell membrane and to investigate theexistence of possible complex partners, an attempt wasmade to separate such complexes by electrophoresisunder nondenaturing conditions using Triton X-100.Triton X-100 is known for its ability to preserve func-

tional membrane protein complexes (Ku¨hn et al., 1983;Schubert et al., 1983; Barbero et al., 1984; Hoessli andRungger-Brandle, 1985). The present immunochemicalinvestigations indeed confirm that Triton X-100 is able toprotect the noncovalent interactions of NCAM with otherproteins or possibly with itself. Our results indicate thatNCAM engages in protein complexes of at least 400–600kDa in the adult mouse hippocampus, whereas in thehippocampus of 5-day-old mice, NCAM does not appearto be involved in complex formation. The broad NCAMimmunoreactive smear with an apparent molecular weightof 400–600 kDa, seen in Western blots of adult mouse

Fig. 8. Immunochemical detection of NCAM in protein com-plexes by Western blot analysis after electrophoretic separationunder nondenaturing conditions.A: Membrane protein frac-tions of 5-day-old (lanes 1and2) and adult (lanes 3and4)mouse by using polyclonal NCAM antibodies. Blots wereobtained from either 4.5% (lanes 1 and 3) or 5.5% (lanes 2 and4) gels. Lanes 1 and 2 each show two distinct bands. A faintband of 140 kDa is marked by arrowheads and a prominent,broad band of 250–280 kDa is marked by arrows. Lanes 3 and 4show a broad band with diffuse labeling between 400 and 600kDa.B: Membrane protein fractions of 5-day-old (lane 1) andadult (lane 2) hippocampi stained with polyclonal L1 antibod-ies from a 4.5% gel. In 5-day-old and adult hippocampi oneband distinct from the NCAM containing bands is seen(compare with (A) lanes 1 and 3).

172 Kramer et al.

hippocampus can be taken as an indication for theexistence of either a set of NCAM-containing complexesor a variety of smaller complexes derived from onecomplex with a molecular weight of at least 600 kDa.These complexes could consist of self-assembled NCAMmolecules (Hall and Rutishauser, 1987; Becker et al.,1989) or of a set of different molecules closely associatedwith each other. Brain spectrin was not found in suchcomplexes, most probably because this cytoskeletal ele-ment may not have been solubilized in Triton X-100,remaining instead associated with cytoskeletal elementsin the membrane fraction. The neural cell adhesionmolecule L1 was also not found in the complexesisolated, although a close association of L1 with NCAMvia extracellular carbohydrate residues has been pro-posed. Most probably, these interactions are labile towardTriton X-100.

The data presented here confirm the existence ofNCAM complexes in the membrane and imply thatcomplex formation is developmentally regulated andbrain region-specific. Complex formation of NCAM 180,as indicated by the masking of the D3-specific epitope,occurs in the cerebral cortex, cerebellum, thalamus, andhippocampus during development, whereas the olfactorybulb shows no epitope masking. In the cerebral cortex,cerebellum, and hippocampus, the time course of epitopemasking correlates with neural differentiation. The epi-tope becomes masked at a time when synaptic connec-tions are formed. In the hippocampal formation, thematuration of the CA regions precede the development ofthe dentate gyrus (Angevine, 1975). This developmentalpattern is reflected by the earlier and more completemasking of the D3 epitope in the pyramidal cell layer aswell as the stratum oriens and radiatum, compared withthe granular cell layer, the stratum moleculare, and thehilus of the dentate gyrus. In the continuously regenerat-ing olfactory nerve, the D3-specific epitope does notappear to engage in complex formation. This is alsoconfirmed by a description of NCAM 180 expression inthe developing and regenerating axons of fish optic nerve(Bastmeyer et al., 1990), where the D3 antibody was usedto detect the molecule. In a study examining the expres-sion of NCAM 180 in the cat visual cortex, the D3-specific epitope of NCAM 180 also became maskedduring development, correlating with the maturation ofsynaptic connections during the critical period (Delius etal., 1993).

At present, there is some evidence from previousstudies that the association of several membrane proteins,via weak molecular interactions, may play an importantrole in triggering certain cell functions (Bourguignon etal., 1990; Volarevic et al., 1990; Stefanova and Horejsi,1991; Jacobson et al., 1995; for reviews, see Parton andSimons, 1995). The complexes observed appeared to

concentrate surface-glycoproteins and membrane-associ-ated molecules in clusters at specific locations in the cellmembrane. All of these proteins were involved in secondmessenger systems. The punctate staining pattern ob-tainedwith the antibodies D3 and 481 at later developmen-tal stages may well be a sign for the formation of cellmembrane domains in which complex formation ofNCAM with itself or other proteins changes the func-tional state of the molecule. We speculate that complexformation of adhesion molecules may induce conforma-tional changes that, in turn, may influence their adhesive-ness and their ability to undergo transmembrane signal-ing. Moreover, complex formation could allow all partnersin the complex to undergo functional modificationsrapidly and simultaneously, thus providing a powerfultool for the fine-tuning of functions as an immediatereaction to local changes.

ACKNOWLEDGMENTSThe authors are grateful to Dr. B. Schlosshauer for

the generous gift of D3 hybridoma cells, Dr. R. Martinifor his valuable comments on the manuscript, D. Schuh-macher and M. Keggenhoff for excellent technical assis-tance, Drs. N. Mitrovic and J. Taylor for helpful discus-sions, Dr. S. Goodman for antibodies against spectrin,and Deutsche Forschungsgemeinschaft for a fellowship(to I.K.).

REFERENCES

Angevine JB (1975): Development of the hippocampal region. InIsaacson RL, Pribram KH: The Hippocampus Vol. 1: Structureand Development. New York: Plenum Press, pp 61–94.

Barbas JA, Chaix J-C, Steinmetz M, Goridis C (1988): Differentialsplicing and alternative polyadenylation generates distinctNCAM transcripts and proteins in the mouse. EMBO J 7:625–632.

Barbero MC, Valpuesta JM, Rial E, Gustubay JI, Goni FM, MarcarullaJM (1984): Effect of the nonionic detergent Triton X-100 onmitochondrial succinate-oxidizing enzymes.Arch BiochemBio-phys 228:560–568.

Barthels D, Santoni M-J, Wille W, Ruppert C, Chaix L-C, Hirsch R,Fontecilla-Camps JC, Goridis C (1987): Isolation and nucleo-tide sequence of mouse NCAM cDNA that codes for a Mr79,000 polypeptide without a membrane-spanning region.EMBO J 6:907–914.

Bastmeyer M, Schlosshauer B, Stu¨rmer CA (1990): The spatio-temporal distribution of N-CAM in the retinotectal pathway ofadult goldfish detected by the monoclonal antibody D3. Devel-opment 108:299–311.

Becker JW, Erickson HP, Hoffman S, Cunningham BA, Edelman GM(1989): Topobiology of cell adhesion molecules. Proc Natl AcadSci USA86:1088–1092.

Bixby JL, Pratt RS, Lilien J, Reichardt LF (1987): Neurite outgrowthon muscle cell surface involves extracellular matrix receptors aswell as Ca21-dependent and -independent cell adhesion mol-ecules. Proc Natl Acad Sci USA84:2555–2559.


Bornstein MB (1958): Reconstituted rat-tail collagen used as substratefor tissue cultures on coverslips in Maximow slides and rollertubes. Lab Invest 7:134–137.

Bourguignon LY, Walker G, Huang HS (1990): Interactions between alymphoma membrane-associated guanosine 58-triphosphate-binding protein and the cytoskeleton during receptor patchingand capping. J Immunol 144:2242–2252.

Buskirk DR, Thiery JP, Rutishauser U, Edelman GM (1980): Antibod-ies to a neural cell adhesion molecule disrupt histogenesis incultured chick retina. Nature 285:488–489.

Covault J, Sanes JR (1986): Distribution of N-CAM in synaptic andextrasynaptic portions of developing and adult skeletal muscle.J Cell Biol 102:716–730.

Cunningham BA, Hemperly JJ, Murray BA, Prediger EA, BrackenburyR, Edelman GM (1987): Neural cell adhesion molecule: struc-ture, immunoglobulin-like domains, cell surface modulationand alternative RNA splicing. Science 236:799–806.

Delius JAM, Kramer I, Schachner M, SingerW (1993): N-CAM 180 inthe postnatal development of cat visual cortex: an immunohisto-chemical study, submitted.

Doherty P, Fruns M, Seaton P, Dickson G, Barton CH, Sears TA,WalshFS (1990): A threshhold effect of the major isoforms of NCAMon neurite outgrowth. Nature 343:464–466.

Doherty P, Rowett LH, Moore SE, Mann DA, Walsh FS (1991):Neurite outgrowth in response to transfected N-CAM andN-cadherin reveals fundamental differences in neuronal respon-siveness to CAMs. Neuron 6:247–258.

Edelman GM, Cunningham, BA, Thiery JP (1990): ‘‘Morphoregula-tory Molecules.’’ New York: JohnWiley & Sons.

Faissner A, Kruse J, Goridis C, Bock E, Schachner M (1984): Theneural cell adhesion molecule L1 is distinct from the N-CAMrelated group of surface antigens BSP-2 and D2. EMBO J3:733–737.

Faissner A, Teplow DB, Ku¨bler D, Keilhauer G, Kinzel V, SchachnerM (1985): Biosynthesis and membrane topography of the neuralcell adhesion molecule L1. EMBO J 4:3105–3113.

Fraser SE, Murray BA, Chuong C-M, Edelman GM (1984): Alterationof the retinotectal map in Xenopus by antibodies to neural celladhesion molecule. Proc Natl Acad Sci USA81:4222–4226.

Frei T, von Bohlen und Halbach F, Wille W, Schachner M (1992):Different extracellular domains of the neural cell adhesionmolecule NCAM are involved in different functions. J Cell Biol118:177–194.

Furley AJ, Morton SB, Manalo D, Karagogeos D, Dodd J, Jessell TK(1990): The axonal glycoprotein TAG-1 is an immunoglobulinsuperfamily member with neurite outgrowth-promoting activity.Cell 61:157–170.

Gennarini G, Rougon G, Deagostini-Bazin H, Goridis C (1984):Studies on the membrane disposition of the neural cell adhesionmolecule N-CAM. Eur J Biochem 142:57–64.

Gennarini G, Hirsch MR, He HT, Him M, Finne J, Goridis C (1986):Differential expression of mouse neural cell adhesion molecule(N-CAM) mRNA species during brain development and inneural cell lines. J Neurosci 6:1983–1990.

Goldowitz D, Barthels D, Lorenzon N, Jungblut A, Wille W (1990):NCAM gene expression during the development of cerebellumand dentate gyrus in the mouse. Dev Brain Res 52:151–160.

Goridis C, Deagostini-Bazin H, HimM, Hirsch MR, Rougon G, SadoulR, Langley OK, Gombos G, Finne J (1983): Neural surfaceantigens during nervous system development. Cold Spring HarbSymp Quant Biol 48:527–538.

Hall AK, Rutishauser U (1987): Visualization of neural cell adhesionmolecule by electron microscopy. J Cell Biol 104:1579–1586.

He HT, Finne J, Goridis C (1987): Biosynthesis, membrane associa-tion, and release of N-CAM-120, a phosphatidylinositol-linked

form of the neural cell adhesionmolecule. J Cell Biol 105:2489–2500.

Hoessli D, Rungger-Brandle E (1985): Association of specific cell-surface glycoproteins with a Triton X-100-resistant complex ofplasma membrane proteins isolated from T lymphoma cells(P1798). Exp Cell Res 156:239–250.

Horstkorte R, Schachner M, Magyar JP, Vorherr T, Schmitz B (1993):The fourth immunoglobulin-like domain of NCAM contains acarbohydrate recognition domain for oligomannosidic glycansimplicated in association with L1 and neurite outgrowth. J CellBiol 121:1409–1421.

Jacobson K, Sheets ED, Simson R (1995): Revisiting the fluid mosaicmodel of membranes. Science 268:1441–1442.

Kadmon G, Kowitz A, Altevogt P, Schachner M (1990a): The neuralcell adhesion molecule N-CAM enhances L1-dependent cell-cell interactions. J Cell Biol 110:193–208.

Kadmon G, Kowitz A, Altevogt P, Schachner M (1990b): Functionalcooperation between the neural cell adhesion molecules L1 andN-CAM is carbohydrate dependent. J Cell Biol 110:209–218.

Keilhauer G, Faissner A, Schachner M (1985): Differential inhibitionof neuron-neuron, neuron-astrocyte and astrocyte-astrocyte ad-hesion by L1, L2 and N-CAM antibodies. Nature 316:728–730.

Kuhn L, Meyer H, Rutschmann M, Thamm P (1983): Identification ofphotoaffinity labeled insulin receptor proteins by linear polyacryl-amide gradient gel electrophoresis under non-denaturatingconditions. FEBS Lett 113:189–192.

Kyhse-Andersen K (1984): Electroblotting of multiple gels: a simpleapparatus without buffer tank for rapid transfer of proteins frompolyacrylamide to nitrocellulose. J Biochem Biophys Methods10:203–209.

Laemmli UK (1970): Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 227:680–685.

Landmesser L, Dahm L, Schultz K, Rutishauser U (1988): Distinctroles for adhesion molecules during innervation of embryonicchick muscle. Dev Biol 130:645–670.

Landmesser L, Dahm L, Tang JC, Rutishauser U (1990): Polysialicacid as a regulator of intramuscular nerve branching duringembryonic development. Neuron 4:655–667.

Luthi A, Laurent JP, Figurov A, Muller D, Schachner M (1994):Hippocampal long-term potentiation and neural cell adhesionmolecules L1 and NCAM. Nature 372:777–779.

Mege RM, Nicolet M, Murawsky M, Goudou D, Tassin AM, Pinc¸on-Raymond M, Rieger F (1991): Role of neural cell adhesionmolecule in myogenesis, nerve regeneration andmuscle reinner-vation. In Wernig, A (ed): ‘‘Plasticity of Motoneuronal Connec-tions.’’ Amsterdam: Elsevier Science Publishers BV, pp 257–266.

Miragall F, Kadmon G, Husmann K, Schachner M (1988): Expressionof cell adhesion molecules in the olfactory system of the adultmouse: presence of the embryonic form of N-CAM. Dev Biol129:516–531.

Miragall F, Kadmon G, Schachner M (1989): Expression of L1 andN-CAM adhesion molecules during development of the mouseolfactory system. Dev Biol 135:272–286.

Murray BA, Owens GC, Prediger EA, Crossin KL, Cunningham BA,Edelman GM (1986): Cell surface modulation of the neural celladhesion molecule resulting from alternative mRNA splicing ina tissue-specific developmental sequence. J Cell Biol 103:1431–1439.

Owens GC, Edelman GM, Cunningham BA (1987): Organization ofthe neural cell adhesion molecule (N-CAM) gene: Alternativeexon usage as the basis for different membrane-associateddomains. Proc Natl Acad Sci USA84:294–298.

Parton RG, Simons K (1995): Digging into caveolae. Science 269:1398–1399.

174 Kramer et al.

Persohn E, Pollerberg GE, Schachner M (1989): Immunoelectron-microscopic localization of the 180 kD component of the neuralcell adhesion molecule N-CAM in postsynaptic membranes. JComp Neurol 288:92–100.

Pigott R, Kelly JS (1984): A cell surface antigen present on culturedcerebellar neurones appears to be transiently expressed duringcerebellar development in the rat. Neurosci Lett 49:105–110.

Pollerberg GE, Sadoul R, Goridis C, Schachner M (1985): Selectiveexpression of the 180-kD component of the neural cell adhesionmolecule N-CAM during development. J Cell Biol 101:1921–1929.

Pollerberg GE, Schachner M, Davoust J (1986): Differentiationstate-dependent surface mobilities of two forms of the neuralcell adhesion molecule. Nature 324:462–465.

Pollerberg GE, Burridge K, Krebs KE, Goodman SR, Schachner M(1987): The 180-kD component of the neural cell adhesionmolecule N-CAM is involved in cell-cell contacts and cytoskel-eton-membrane interactions. Cell Tissue Res 250:227–236.

Pollerberg GE, Nolte C, Schachner M (1991): Accumulation ofN-CAM 180 at contact sites between neuroblastoma cells andplastic beads coated with extracellular matrix molecules. Eur JNeurosci 2:879–887.

Prince JT, Alberti L, Healy PA, Nauman SJ, Stallcup WB (1991):Molecular cloning of NILE glycoprotein and evidence for itscontinued expression in mature rat CNS. J Neurosci Res30:567–581.

Riederer BM, Zagon IS, Goodman SR (1986): Brain spectrin (240/235)and brain spectrin (240/235 E): Two distinct spectrin subtypeswith different locations within mammalian neural cells. J CellBiol 102:2088–2097.

Ronn LC, Bock E, Linnemann D, Jahnsen H (1995): NCAM-antibodies modulate induction of long-term potentiation in rathippocampal CA1. Brain Res 677:145–151.

Rutishauser U, Edelman GM (1980): Effects of fasciculation on theoutgrowth of neurites from spinal ganglia in culture. J Cell Biol87:370–378.

Rutishauser U, Jessell TM (1988): Cell adhesion molecules in verte-brate neural development. Physiol Rev 68:819–887.

Sadoul K, MeyerA, LowMG, Schachner M (1986): Release of the 120kDa component of the mouse neural cell adhesion moleculeN-CAM from the cell surface by phosphatidylinositol-specificphospholipase C. Neurosci Lett 72:341–346.

Santoni M-J, Barthels D, Barbas JA, Hirsch M-R, Steinmetz M,Goridis C, Wille W (1987): Analysis of cDNA clones that codefor the transmembrane forms of the mouse neural cell adhesionmolecule (NCAM) and are generated by alternative RNAsplicing. Nucleic Acids Res 15:8621–8641.

Schachner M (1991): Neural recognition molecules and their influenceon cellular functions. In Letourneau PC, Kater SB, Macagno ER(eds): ‘‘The Nerve Growth Cone.’’ New York: Raven Press, pp237–254.

Schlosshauer B (1989): Purification of neuronal cell surface proteinsand generation of epitope-specificmonoclonal antibodies againstcell adhesion molecules. J Neurochem 52:82–92.

Simon H, Klinz S, Fahrig T, Schachner M (1991): Molecular associa-tion of the neural adhesion molecules L1 and N-CAM in thesurface membrane of neuroblastoma cells is shown by chemicalcross linking. Eur J Neurosci 3:634–640.

Schnitzer J, Schachner M (1981): Characterization of isolated mousecerebellar cell populations in vitro. J Neuroimmunol 1:457–470.

Scholey AB, Rose SPR, Zamani MR, Bock E, Schachner M (1993): Arole for the neural cell adhesionmolecule in a late, consolidatingphase of glycoprotein synthesis 6 hours following passiveavoidance training of the young chick. Neuroscience 55:499–509.

Schubert D, Boss K, Dorst HJ, Flossdorf J, Pappert G (1983): Thenature of the stable noncovalent dimers of band 3 protein fromerythrocyte membranes in solution of Triton X-100. FEBS Lett163:81–84.

Schuch U, Lohse MJ, Schachner M (1989): Neural cell adhesionmolecules influence second messenger systems. Neuron 3:13–20.

Sorkin BC, Hoffman S, Edelman GM, Cunningham BA (1984):Sulfation and phosphorylation of the neural cell adhesionmolecule NCAM. Science 225:1476–1478.

Sorkin BC, Grumet M, Cunningham BA, Edelman GM (1985):Structures of two neural cell adhesion molecules. Soc NeurosciAbstr 1 1:1138.

Stanley KK, Lutzio JP (1984): Construction of a new family of highefficiency bacterial expression vectors: Identification of cDNAclones coding for human liver proteins. EMBO J 3:1429–1434.

Stefanova L, Horejsi V (1991): Association of the CD55 cell surfaceglycoproteins with other membrane molecules. J Immunol147:1587–1592.

Thanos S, Bonhoeffer F, Rutishauser U (1984): Fiber-fiber interactionsand tectal cues influence the development of the chickenretinotectal projection. Proc Natl Acad Sci USA 81:1906–1910.

Volarevic S, Burns CM, Sussman JJ, Ashwell JD (1990): Intimateassociation of Thy-1 and the T-cell antigen receptor with theCD45 tyrosine phosphatase. Proc Natl Acad Sci USA 87:7085–7089.

von Bohlen und Halbach F, Taylor J, Schachner M (1992): Celltype-specific effect of the neural adhesion molecules L1 andNCAM on diverse second messenger systems. Eur J Neurosci4:896–909.

Williams MA (1977): Immunocytochemistry at the EM level: Stainingantigens with electron-dense reagents. In Williams MA (ed):‘‘Autoradiography and Immunocytochemistry.’’ Amsterdam:North-Holland Publishing Company, pp 41–76.

Williams EJ, Furness J,Walsh FS, Doherty P (1994):Activation of FGFreceptor underlies neurite outgrowth stimulated by L1, N-CAMand N-cadherin. Neuron 13:583–594.


developmentally regulated masking of an intracellular epitope of the 180 kda isoform of the neural...

Documents