Glial cells express N-CAM/D2-CAM-like polypeptides in vitro

The joining together of neurites to form fascicles and the growth of axons along glial surfaces during early development suggest that neurone-neurone and neurone-glial adhesion interactions are of considerable importance for defining nerve tracts. In vitro studies have indicated that adhesion between neurones involves a glycoprotein that has been independently studied under the names of N-CAM (for neural cell adhesion molecule), D2-CAM and BSP-2 (refs 10, 11). As N-CAM/D2-CAM appears to be a homophilic ligand that binds to N-CAM/D2-CAM polypeptide on adjacent cells, this glycoprotein is potentially important in adhesion interactions between any two N-CAM/D2-CAM-expressing cells. While it has been suggested that neurone-glial adhesion involves molecules other than N-CAM/D2-CAM, it is known that N-CAM/D2-CAM antigenic determinants are expressed by glial cells in vivo and that injection of anti-N-CAM antibodies into the eye-cup of chick embryos disrupts normal patterns of neuritic apposition to glial endfeet in the developing optic stalk. Do the molecules expressed by glia share restricted antigenic determinants, or binding domains, with N-CAM/D2-CAM, or are N-CAM/D2-CAM polypeptides expressed by glia? Here we present immunocytochemical evidence which suggests that all classes of macroglia express N-CAM/D2-CAM antigenic determinants on their surfaces and immunochemical analyses which indicate that the molecules expressed by purified astrocytes are closely similar, or identical, to at least some forms of N-CAM/D2-CAM obtained from whole brain or purified neurones. However, our results also suggest that different N-CAM/D2-CAM polypeptides may be separately expressed by neurones and astrocytes.     

The J1 glycoprotein--a novel nervous system cell adhesion molecule of the L2/HNK-1 family

The neural cell adhesion molecules L1 and N-CAM share a common carbohydrate epitope that is recognized by the monoclonal antibodies L2 and HNK-1. The L2/HNK-1 epitope is also present on the myelin-associated glycoprotein (MAG) which is thought to mediate surface interactions between the axon and myelinating cell. Other, as yet unidentified, cell-surface glycoproteins are recognized by the two antibodies and are believed to belong to a family of neural cell adhesion molecules. To test this hypothesis, we have prepared polyclonal antibodies to a prominent member of the L2/HNK-1 family, the 160K (relative molecular mass (Mr)160,000) glycoprotein. Here we report that these antibodies, designated J1 antibodies, react with astrocytes and oligodendrocytes and interfere with neurone-astrocyte adhesion, but not with neurone-neurone or astrocyte-astrocyte adhesion. This result suggests the involvement of the J1 antigen in cell-cell interactions.     

Neural adhesion molecule L1 as a member of the immunoglobulin superfamily with binding domains similar to fibronectin

Diverse glycoproteins of cell surfaces and extracellular matrices operationally termed 'adhesion molecules' are important in the specification of cell interactions during development, maintenance and regeneration of the nervous system. These adhesion molecules have distinct functions involving different cells at different developmental stages, but may cooperate when expressed together. Families of adhesion molecules which share common carbohydrate domains do exist, despite the structural and functional diversity of these glycoproteins. These include the Ca2+-independent neural adhesion molecules: N-CAM, myelin associated glycoprotein (MAG) and L1. L1 is involved in neuron-neuron adhesion, neurite fasciculation, outgrowth of neurites, cerebellar granule cell migration, neurite outgrowth on Schwann cells and interactions among epithelial cells of intestinal crypts. We show here that in addition to sharing carbohydrate epitopes with N-CAM and MAG, L1 is also a member of the immunoglobulin superfamily. It contains six C2 domains and also shares three type III domains with the extracellular matrix adhesion molecule fibronectin.     

Differential inhibition of neurone-neurone, neurone-astrocyte and astrocyte-astrocyte adhesion by L1, L2 and N-CAM antibodies

The cell adhesion molecules L1, N-CAM and Ng-CAM have been implicated in cell-cell interactions among developing neural cells. L1 and N-CAM are structurally and functionally distinct molecular entities and act synergistically in mediating Ca2+-independent adhesion between re-aggregating early postnatal cerebellar cells. N-CAM has been reported to be neurone-specific in the chicken and to mediate fasciculation of neurites and of nerve-muscle interactions. L1, which in the central nervous system has been found only on post-mitotic neurones, mediates migration of granule cell neurones in the mouse cerebellar cortex. In view of the molecules' distinct effects on cell interactions, we wondered whether different neural cell types are involved in the actions of each molecule. Here we report that L1 antigen promotes neurone-neurone adhesion. N-CAM, which is expressed on both neurones and glia, mediates neurone-neurone, neurone-astrocyte and astrocyte-astrocyte adhesion. The L2 carbohydrate epitope shared between the two adhesion molecules seems to be involved in neurone-astrocyte and astrocyte-astrocyte adhesion and acts in a more than additive manner in N-CAM-mediated neurone-neurone adhesion.     

Differentiation state-dependent surface mobilities of two forms of the neural cell adhesion molecule

The neural cell adhesion molecule (N-CAM) has been implicated in morphogenetic events during formation of the nervous system. Three forms of N-CAM exist, all glycoprotein chains, of relative molecular masses 180,000 (180K), 140K and 120K (N-CAM180, N-CAM140 and N-CAM120) which are differentially expressed on neural cell types and during development. The three chains are thought to carry similar if not identical amino-acid sequences on their extracellular amino-terminal domains, but differ in the length of their carboxy-terminal cytoplasmic region. They occur in highly sialylated embryonic and less sialylated adult forms. N-CAM180 is selectively expressed in more differentiated neural cells and may play a role in the stabilization of cell contacts. To investigate this, we have studied in the surface membrane of a mouse neuroblastoma cell line N2A the lateral mobility of the two predominant forms of N-CAM, N-CAM180 and N-CAM140, as a function of differentiation. Here we report that as judged by fringe pattern photobleaching, the surface mobility of N-CAM140 is higher than that of N-CAM180, suggesting an association of N-CAM180 with the cytoskeleton or other stabilizing factors. We also show that brain spectrin, a membrane-cytoskeleton linker protein, binds only to N-CAM180. The immobilization of N-CAM in differentiated N2A cells is achieved by a shift in expression from N-CAM140 to N-CAM180.     

Neural cell adhesion molecules and myelin-associated glycoprotein share a common carbohydrate moiety recognized by monoclonal antibodies L2 and HNK-1

Cell surface molecules have been implicated in cell interactions which underlie formation of the nervous system. The analysis of the functional properties of such molecules has profited from the combined use of antibodies and cell culture systems. It has been suggested that the interplay between these molecules modulates cell-to-cell interaction at critical developmental stages. In the mouse, N-CAM and L1 antigen have been shown to mediate Ca2+-independent adhesion among neural cells. N-CAM plays a role in fasciculation of neurites and formation of neuromuscular junction. L1 is apparently not involved in synaptogenesis, but in migration of granule cell neurones in the developing mouse cerebellar cortex. The two antigens are distinct molecular and functional entities which act synergistically in aggregation of neuroblastoma and early postnatal cerebellar cells. In view of a certain similarity in function between the two groups of molecules, it was not surprising to find that structural similarities are detectable by the monoclonal antibody L2. We show here that a carbohydrate moiety recognized by L2 and HNK-1 monoclonal antibodies, is present in mouse N-CAM and L1. The L2 epitope appears on all major neural cell types but not all N-CAM molecules express it. This heterogeneity points to a previously undetected molecular diversity which may have functional implications for modulating cell adhesion during development.     

Targeting of transmembrane and GPI-anchored forms of N-CAM to opposite domains of a polarized epithelial cell.

The calcium-independent neural cell adhesion molecule N-CAM is expressed transiently during development in many tissues, including epithelia. The three naturally occurring principal isoforms of N-CAM differ in the way in which they associate with the membrane and in their cytoplasmic domains. These isoforms are generated by developmentally regulated alternative splicing of a single gene: the large cytoplasmic domain (ld) form (relative molecular mass 180,000 (Mr 180K] is specific for post-mitotic neurons; the 120K small cytoplasmic domain (ssd) and 140K small surface domain (sd) forms also occur on other cell types. One function of the different isoforms could be to specify cellular localization; for example, glycosyl phosphatidyl inositol (GPI)-membrane anchoring acts as a targeting signal for expression on the apical surface of polarized epithelial cells. Neurons and epithelial cells may use similar mechanisms for polarizing their plasma membrane proteins. We have therefore investigated the targeting of GPI-anchored (ssd N-CAM, 120K) and transmembrane forms of N-CAM (sd N-CAM, 140K; ld N-CAM, 180K) by comparing the expression of each after transfection of the appropriate complementary DNAs into polarized epithelial cells. We find that isoforms with alternative modes of membrane association are targeted to different surfaces of polarized epithelial cells: ssd N-CAM is expressed on the apical surface, whereas sd and ld N-CAM are expressed on the basolateral surface. These results suggest that the different isoforms of N-CAM determine their own diverse cellular destinations. They also support the hypothesis that the GPI anchor acts as an apical targeting signal in epithelia.     

Cell-surface regulation of beta 1-integrin activity on developing retinal neurons.

Integrins are a family of alpha beta heterodimeric receptors that mediate cell-cell and cell-substratum interactions. Integrin binding to extracellular ligands regulates cell adhesion, shape, motility, intracellular signalling and gene expression. Mechanisms that regulate integrin function are, therefore, central to the participation of integrins in a diverse set of cellular events. Here we report the identification of TASC, a monoclonal antibody to a novel epitope on the integrin beta 1 subunit, which inhibits cell adhesion to vitronectin but promotes adhesion to laminin and collagen types I and IV. We show that developing retinal neurons that have lost responsiveness to laminin regain the ability to bind laminin in the presence of TASC. Thus, beta 1-class integrins are likely to occupy multiple affinity states that can be modulated at the cell surface.     

An LFA-3 cDNA encodes a phospholipid-linked membrane protein homologous to its receptor CD2

Recently the human T cell erythrocyte receptor CD2 has been shown to bind human erythrocytes through LFA-3, a heavily glycosylated surface protein of broad tissue distribution. CD2-LFA-3 interactions are important for cytolytic conjugate formation, for thymocyte adhesion, and for T cell activation. A complementary DNA clone encoding LFA-3 was isolated using a complementary DNA clone encoding LFA-3 was isolated using a novel transient expression system of mouse cells. The cDNA encodes a phospholipid-linked membrane protein whose extracellular domain shares significant homology with CD2. As CD2 is homologous with the neural cell adhesion molecule NCAM in immunoglobulin-like domains, cellular adhesion molecules in both neural and lymphoid tissues could have a common ancestor.     

Engineering galactose-binding activity into a C-type mannose-binding protein.

Calcium-dependent or C-type carbohydrate-recognition domains are homologous protein modules found in a variety of animal lectins. Selective binding of sugars by these domains is essential for glycoprotein clearance, cell-cell adhesion and pathogen neutralization. Although various C-type carbohydrate-recognition domains share sequence identity ranging from 20 to 55%, their sugar-binding characteristics vary widely. The structure of a mannose-binding carbohydrate-recognition domain in complex with a saccharide ligand suggests that two glutamic acid-asparagine pairs are essential determinants of ligand binding by this domain. In C-type lectins that bind galactose with higher affinity than mannose, one of these pairs is replaced by glutamine-aspartic acid. Here we shift the sequence of the mannose-binding protein to correspond to that found in galactose-binding domains in order to test the importance of these residues in sugar-binding selectivity. This simple switch in the position of a single amide group alters the binding activity of the domain so that galactose becomes the preferred ligand.     

Regulation of repulsion versus adhesion by different splice forms of an Eph receptor

Eph tyrosine kinase receptors and their membrane-bound ephrin ligands mediate cell interactions and participate in several developmental processes. Ligand binding to an Eph receptor results in tyrosine phosphorylation of the kinase domain, and repulsion of axonal growth cones and migrating cells. Here we report that a subpopulation of ephrin-A5 null mice display neural tube defects resembling anencephaly in man. This is caused by the failure of the neural folds to fuse in the dorsal midline, suggesting that ephrin-A5, in addition to its involvement in cell repulsion, can participate in cell adhesion. During neurulation, ephrin-A5 is co-expressed with its cognate receptor EphA7 in cells at the edges of the dorsal neural folds. Three different EphA7 splice variants, a full-length form and two truncated versions lacking kinase domains, are expressed in the neural folds. Co-expression of an endogenously expressed truncated form of EphA7 suppresses tyrosine phosphorylation of the full-length EphA7 receptor and shifts the cellular response from repulsion to adhesion in vitro. We conclude that alternative usage of different splice forms of a tyrosine kinase receptor can mediate cellular adhesion or repulsion during embryonic development.     

A two-tiered mechanism for stabilization and immobilization of E-cadherin

Epithelial tissues maintain a robust architecture which is important for their barrier function, but they are also remodelled through the reorganization of cell-cell contacts. Tissue stability requires intercellular adhesion mediated by E-cadherin, in particular its trans-association in homophilic complexes supported by actin filaments through beta- and alpha-catenin. How alpha-catenin dynamic interactions between E-cadherin/beta-catenin and cortical actin control both stability and remodelling of adhesion is unclear. Here we focus on Drosophila homophilic E-cadherin complexes rather than total E-cadherin, including diffusing 'free' E-cadherin, because these complexes are a better proxy for adhesion. We find that E-cadherin complexes partition in very stable microdomains (that is, bona fide adhesive foci which are more stable than remodelling contacts). Furthermore, we find that stability and mobility of these microdomains depend on two actin populations: small, stable actin patches concentrate at homophilic E-cadherin clusters, whereas a rapidly turning over, contractile network constrains their lateral movement by a tethering mechanism. alpha-Catenin controls epithelial architecture mainly through regulation of the mobility of homophilic clusters and it is largely dispensable for their stability. Uncoupling stability and mobility of E-cadherin complexes suggests that stable epithelia may remodel through the regulated mobility of very stable adhesive foci.     

Structural basis of Dscam isoform specificity

The Dscam gene gives rise to thousands of diverse cell surface receptors thought to provide homophilic and heterophilic recognition specificity for neuronal wiring and immune responses. Mutually exclusive splicing allows for the generation of sequence variability in three immunoglobulin ecto-domains, D2, D3 and D7. We report X-ray structures of the amino-terminal four immunoglobulin domains (D1-D4) of two distinct Dscam isoforms. The structures reveal a horseshoe configuration, with variable residues of D2 and D3 constituting two independent surface epitopes on either side of the receptor. Both isoforms engage in homo-dimerization coupling variable domain D2 with D2, and D3 with D3. These interactions involve symmetric, antiparallel pairing of identical peptide segments from epitope I that are unique to each isoform. Structure-guided mutagenesis and swapping of peptide segments confirm that epitope I, but not epitope II, confers homophilic binding specificity of full-length Dscam receptors. Phylogenetic analysis shows strong selection of matching peptide sequences only for epitope I. We propose that peptide complementarity of variable residues in epitope I of Dscam is essential for homophilic binding specificity.     

Structural basis for vinculin activation at sites of cell adhesion

Vinculin is a highly conserved intracellular protein with a crucial role in the maintenance and regulation of cell adhesion and migration. In the cytosol, vinculin adopts a default autoinhibited conformation. On recruitment to cell-cell and cell-matrix adherens-type junctions, vinculin becomes activated and mediates various protein-protein interactions that regulate the links between F-actin and the cadherin and integrin families of cell-adhesion molecules. Here we describe the crystal structure of the full-length vinculin molecule (1,066 amino acids), which shows a five-domain autoinhibited conformation in which the carboxy-terminal tail domain is held pincer-like by the vinculin head, and ligand binding is regulated both sterically and allosterically. We show that conformational changes in the head, tail and proline-rich domains are linked structurally and thermodynamically, and propose a combinatorial pathway to activation that ensures that vinculin is activated only at sites of cell adhesion when two or more of its binding partners are brought into apposition.     

Protocadherins mediate dendritic self-avoidance in the mammalian nervous system

Dendritic arborizations of many neurons are patterned by a process called self-avoidance, in which branches arising from a single neuron repel each other. By minimizing gaps and overlaps within the arborization, self-avoidance facilitates complete coverage of a neuron’s territory by its neurites. Remarkably, some neurons that display self-avoidance interact freely with other neurons of the same subtype, implying that they discriminate self from non-self. Here we demonstrate roles for the clustered protocadherins (Pcdhs) in dendritic self-avoidance and self/non-self discrimination. The Pcdh locus encodes 58 related cadherin-like transmembrane proteins, at least some of which exhibit isoform-specific homophilic adhesion in heterologous cells and are expressed stochastically and combinatorially in single neurons. Deletion of all 22 Pcdh genes in the mouse γ-subcluster (Pcdhg genes) disrupts self-avoidance of dendrites in retinal starburst amacrine cells (SACs) and cerebellar Purkinje cells. Further genetic analysis of SACs showed that Pcdhg proteins act cell-autonomously during development, and that replacement of the 22 Pcdhg proteins with a single isoform restores self-avoidance. Moreover, expression of the same single isoform in all SACs decreases interactions among dendrites of neighbouring SACs (heteroneuronal interactions). These results suggest that homophilic Pcdhg interactions between sibling neurites (isoneuronal interactions) generate a repulsive signal that leads to self-avoidance. In this model, heteroneuronal interactions are normally permitted because dendrites seldom encounter a matched set of Pcdhg proteins unless they emanate from the same soma. In many respects, our results mirror those reported for Dscam1 (Down syndrome cell adhesion molecule) in Drosophila: this complex gene encodes thousands of recognition molecules that exhibit stochastic expression and isoform-specific interactions, and mediate both self-avoidance and self/non-self discrimination. Thus, although insect Dscam and vertebrate Pcdh proteins share no sequence homology, they seem to underlie similar strategies for endowing neurons with distinct molecular identities and patterning their arborizations.     

Interaction between CD4 and class II MHC molecules mediates cell adhesion

The CD4 glycoprotein is expressed on T-helper and cytotoxic lymphocytes which are restricted to class II major histocompatibility complex (MHC) antigens on target cells. Antibody inhibition studies imply that CD4 acts to increase the avidity of effector-target cell interactions. These observations have led to the speculation that CD4 binds to a monomorphic class II antigen determinant, thereby augmenting low affinity T-cell receptor-antigen interactions. However, no direct evidence has been presented indicating that CD4 and class II molecules interact. To address this issue, we have used a vector derived from simian virus 40 (SV40) to express a complementary DNA (cDNA) encoding the human CD4 glycoprotein. When CV1 cells expressing large amounts of the CD4 protein at the cell surface are incubated with human B cells bearing MHC-encoded class II molecules, they are bound tightly to the infected monolayer, whereas mutant B cells which lack class II molecules fail to bind. Furthermore, the binding reaction is specifically inhibited by anti-class II and anti-CD4 antibodies. Thus, the CD4 protein, even in the absence of T-cell receptor-antigen interactions, can interact directly with class II antigens to function as a cell surface adhesion molecule.     

Dscam and Sidekick proteins direct lamina-specific synaptic connections in vertebrate retina

Synaptic circuits in the retina transform visual input gathered by photoreceptors into messages that retinal ganglion cells (RGCs) send to the brain. Processes of retinal interneurons (amacrine and bipolar cells) form synapses on dendrites of RGCs in the inner plexiform layer (IPL). The IPL is divided into at least 10 parallel sublaminae; subsets of interneurons and RGCs arborize and form synapses in just one or a few of them. These lamina-specific circuits determine the visual features to which RGC subtypes respond. Here we show that four closely related immunoglobulin superfamily (IgSF) adhesion molecules--Dscam (Down's syndrome cell adhesion molecule), DscamL (refs 6-9), Sidekick-1 and Sidekick-2 (ref. 10)--are expressed in chick by non-overlapping subsets of interneurons and RGCs that form synapses in distinct IPL sublaminae. Moreover, each protein is concentrated within the appropriate sublaminae and each mediates homophilic adhesion. Loss- and gain-of-function studies in vivo indicate that these IgSF members participate in determining the IPL sublaminae in which synaptic partners arborize and connect. Thus, vertebrate Dscams, like Drosophila Dscams, play roles in neural connectivity. Together, our results on Dscams and Sidekicks suggest the existence of an IgSF code for laminar specificity in retina and, by implication, in other parts of the central nervous system.     

Transformation of cell adhesion properties by exogenously introduced E-cadherin cDNA

E-cadherin is a cell surface glycoprotein responsible for Ca2+-dependent intercellular adhesion between epithelial cells; it is also called uvomorulin, L-CAM (ref. 3), cell-CAM 120/80 (ref.4) or Arc-1 (ref. 5). Because blocking the action of E-cadherin by monoclonal antibodies causes dispersion of compact cell colonies, this molecule is thought to be an important factor for maintenance of multicellular systems. To demonstrate directly that E-cadherin is involved in cell-cell adhesion, we cloned full-length cDNA encoding E-cadherin from F9 cells and introduced it into L fibroblasts deficient in E-cadherin. These L cells acquire strong Ca2+-dependent aggregating activity by expressing the E-cadherin derived from the introduced cDNA and were morphologically transformed so as to form colonies in which cells were tightly connected to each other.     

Crystal structure of fibroblast growth factor receptor ectodomain bound to ligand and heparin

Fibroblast growth factors (FGFs) are a large family of structurally related proteins with a wide range of physiological and pathological activities. Signal transduction requires association of FGF with its receptor tyrosine kinase (FGFR) and heparan sulphate proteoglycan in a specific complex on the cell surface. Direct involvement of the heparan sulphate glycosaminoglycan polysaccharide in the molecular association between FGF and its receptor is essential for biological activity. Although crystal structures of binary complexes of FGF-heparin and FGF-FGFR have been described, the molecular architecture of the FGF signalling complex has not been elucidated. Here we report the crystal structure of the FGFR2 ectodomain in a dimeric form that is induced by simultaneous binding to FGF1 and a heparin decasaccharide. The complex is assembled around a central heparin molecule linking two FGF1 ligands into a dimer that bridges between two receptor chains. The asymmetric heparin binding involves contacts with both FGF1 molecules but only one receptor chain. The structure of the FGF1-FGFR2-heparin ternary complex provides a structural basis for the essential role of heparan sulphate in FGF signalling.     

Role of myelin P0 protein as a homophilic adhesion molecule.

Peripheral nervous system myelin is an extension of the Schwann cell's plasma membrane that tightly enwraps axons in many layers and permits nerve impulses to be rapidly conducted. It is not known how these multiple membrane layers are held together in this compact form. Here we present evidence supporting the hypothesis that the extracellular leaflets of myelin are held together by the most abundant protein of myelin of the peripheral nervous system, P0, by homophilic interaction of its extracellular domains. Transfected Chinese hamster ovary cells expressing P0 protein adhere to each other in suspension, to form large aggregates, whereas cells that are identical but which do not express P0 do not. We also show that this aggregation is mediated by homophilic binding between P0-expressing cells and that the apposing plasma membranes of these cells specifically form desmosomes, whereas control transfected cells do not. As the only difference between the two cell populations is the expression of P0, this protein is apparently responsible for the changes in morphology and adhesion in the cells that express it. The idea that P0 is a homophilic adhesion molecule is supported by its inclusion in the immunoglobulin supergene family, all members of which are involved in recognition and/or adhesion.