Axonin-1/TAG-1 mediates cell-cell adhesion by a cis-assisted trans-interaction.

The neural cell adhesion molecule axonin-1/TAG-1 mediates cell-cell interactions via homophilic and heterophilic contacts. It consists of six Ig and four fibronectin type III domains anchored to the membrane by glycosylphosphatidylinositol. The recently solved crystal structure indicates a module composed of the four N-terminal Ig domains as the contact site between trans-interacting axonin-1 molecules from apposed membranes. Here, we have tested domain-specific monoclonal antibodies for their capacity to interfere with homophilic binding in a cell aggregation assay. The results confirmed the existence of a binding region within the N-terminal Ig domains and identified a second region contributing to homophilic binding on the third and fourth fibronectin domains near the C terminus. The perturbation of each region alone resulted in a complete loss of cell aggregation, suggesting that axonin-1-mediated cell-cell contact results from a cooperative action of two homophilic binding regions. The data support that axonin-1-mediated cell-cell contact is formed by cis-assisted trans-binding. The N-terminal binding regions of axonin-1 establish a linear zipper-like string of trans-interacting axonin-1 molecules alternately provided by the two apposed membranes. The C-terminal binding regions strengthen the cell-cell contact by enhancing the expansion of the linear string into a two-dimensional array via cis-interactions. Cis-assisted trans-binding may be a basic binding mechanism common to many cell adhesion molecules.     

Neurite Fasciculation Mediated by Complexes of Axonin-1 and Ng Cell Adhesion Molecule

Neural cell adhesion molecules composed of immunoglobulin and fibronectin type III-like domains have been implicated in cell adhesion, neurite outgrowth, and fasciculation. Axonin-1 and Ng cell adhesion molecule (NgCAM), two molecules with predominantly axonal expression exhibit homophilic interactions across the extracellular space (axonin- 1/axonin-1 and NgCAM/NgCAM) and a heterophilic interaction (axonin-1–NgCAM) that occurs exclusively in the plane of the same membrane (cis-interaction). Using domain deletion mutants we localized the NgCAM homophilic binding in the Ig domains 1-4 whereas heterophilic binding to axonin-1 was localized in the Ig domains 2-4 and the third FnIII domain. The NgCAM–NgCAM interaction could be established simultaneously with the axonin-1–NgCAM interaction. In contrast, the axonin-1–NgCAM interaction excluded axonin-1/axonin-1 binding. These results and the examination of the coclustering of axonin-1 and NgCAM at cell contacts, suggest that intercellular contact is mediated by a symmetric axonin-1₂/NgCAM₂ tetramer, in which homophilic NgCAM binding across the extracellular space occurs simultaneously with a cis-heterophilic interaction of axonin-1 and NgCAM. The enhanced neurite fasciculation after overexpression of NgCAM by adenoviral vectors indicates that NgCAM is the limiting component for the formation of the axonin-1₂/NgCAM₂ complexes and, thus, neurite fasciculation in DRG neurons.     

Binding between the neural cell adhesion molecules axonin-1 and Nr- CAM/Bravo is involved in neuron-glia interaction

Neural cell adhesion molecules of the immunoglobulin superfamily mediate cellular interactions via homophilic binding to identical molecules and heterophilic binding to other family members or structurally unrelated cell-surface glycoproteins. Here we report on an interaction between axonin-1 and Nr-CAM/Bravo. In search for novel ligands of axonin-1, fluorescent polystyrene microspheres conjugated with axonin-1 were found to bind to peripheral glial cells from dorsal root ganglia. By antibody blockage experiments an axonin-1 receptor on the glial cells was identified as Nr-CAM. The specificity of the interaction was confirmed with binding studies using purified axonin-1 and Nr-CAM. In cultures of dissociated dorsal root ganglia antibodies against axonin-1 and Nr-CAM perturbed the formation of contacts between neurites and peripheral glial cells. Together, these results implicate a binding between axonin-1 of the neuritic and Nr-CAM of the glial cell membrane in the early phase of axon ensheathment in the peripheral nervous system.     

Implications for the domain arrangement of axonin-1 derived from the mapping of its NgCAM binding site.

The neuronal cell adhesion molecule axonin-1 is composed of six immunoglobulin and four fibronectin type III domains. Axonin-1 promotes neurite outgrowth, when presented as a substratum for neurons in vitro, via a neuronal receptor that has been identified as the neuron-glia cell adhesion molecule, NgCAM, based on the blocking effect of polyclonal antibodies directed to NgCAM. Here we report the identification of axonin-1 domains involved in NgCAM binding. NgCAM-conjugated microspheres were tested for binding to COS cells expressing domain deletion mutants of axonin-1. In addition, monoclonal antibodies directed to axonin-1 were assessed for their ability to block the axonin-1-NgCAM interaction, and their epitopes were mapped using the domain deletion mutants. The results suggest that the four amino-terminal immunoglobulin domains of axonin-1 form a domain conglomerate which is necessary and sufficient for NgCAM binding. Surprisingly, NgCAM binding to membrane-bound axonin-1 was increased strongly by deletion of the fifth or sixth immunoglobulin domains of axonin-1. Based on these results and on negative staining electron microscopy, we propose a horseshoe-shaped domain arrangement of axonin-1 that obscures the NgCAM binding site. Neurite outgrowth studies with truncated forms of axonin-1 show that axonin-1 is a neurite outgrowth-promoting substratum in the absence of the NgCAM binding site.     

Cell adhesion molecules NgCAM and axonin-1 form heterodimers in the neuronal membrane and cooperate in neurite outgrowth promotion

The axonal surface glycoproteins neuronglia cell adhesion molecule (NgCAM) and axonin-1 promote cell-cell adhesion, neurite outgrowth and fasciculation, and are involved in growth cone guidance. A direct binding between NgCAM and axonin-1 has been demonstrated using isolated molecules conjugated to the surface of fluorescent microspheres. By expressing NgCAM and axonin-1 in myeloma cells and performing cell aggregation assays, we found that NgCAM and axonin-1 cannot bind when present on the surface of different cells. In contrast, the cocapping of axonin-1 upon antibody-induced capping of NgCAM on the surface of CV- 1 cells coexpressing NgCAM and axonin-1 and the selective chemical cross-linking of the two molecules in low density cultures of dorsal root ganglia neurons indicated a specific and direct binding of axonin- 1 and Ng-CAM in the plane of the same membrane. Suppression of the axonin-1 translation by antisense oligonucleotides prevented neurite outgrowth in dissociated dorsal root ganglia neurons cultured on an NgCAM substratum, indicating that neurite outgrowth on NgCAM substratum requires axonin-1. Based on these and previous results, which implicated NgCAM as the neuronal receptor involved in neurite outgrowth on NgCAM substratum, we concluded that neurite outgrowth on an NgCAM substratum depends on two essential interactions of growth cone NgCAM: a trans-interaction with substratum NgCAM and a cis-interaction with axonin-1 residing in the same growth cone membrane.     

Critical and optimal Ig domains for promotion of neurite outgrowth by L1/Ng-CAM

Mammalian L1 and avian Ng-CAM are homologous neural cell adhesion molecules (CAMs) that promote neurite outgrowth and cell adhesion in most neurons. Previous attempts to map these activities to discrete regions in the CAMs have suggested the involvement of a variety of different domains. However, these studies mainly used bacterially expressed proteins that were much less active on a molar basis than the native molecules. To define regions that are critical for maximal neurite outgrowth, we constructed and tested a panel of eukaryotically expressed proteins containing various extracellular segments of human L1 (hL1) or Ng-CAM. Our results indicate that Ig domains 1-4 of hL1 are critical for homophilic binding and neurite outgrowth; however this segment is less potent than the entire extracellular region. Optimal neurite outgrowth activity was seen with proteins containing all six Ig domains of hL1 or Ng-CAM. The adhesive properties of hL1 fragments correlated tightly with their neurite outgrowth activities, suggesting that these two processes are closely linked. These results suggest that Ig domains 1-4 form a structural cassette responsible for hL1 homophilic binding, while Ig domains 1-6 represent a functional region for optimal promotion of neurite outgrowth in vitro and possibly in vivo.     

Nr-CAM promotes neurite outgrowth from peripheral ganglia by a mechanism involving axonin-1 as a neuronal receptor.

Nr-CAM is a neuronal cell adhesion molecule (CAM) belonging to the immunoglobulin superfamily that has been implicated as a ligand for another CAM, axonin-1, in guidance of commissural axons across the floor plate in the spinal cord. Nr-CAM also serves as a neuronal receptor for several other cell surface molecules, but its role as a ligand in neurite outgrowth is poorly understood. We studied this problem using a chimeric Fc-fusion protein of the extracellular region of Nr-CAM (Nr-Fc) and investigated potential neuronal receptors in the developing peripheral nervous system. A recombinant Nr-CAM-Fc fusion protein, containing all six Ig domains and the first two fibronectin type III repeats of the extracellular region of Nr-CAM, retains cellular and molecular binding activities of the native protein. Injection of Nr-Fc into the central canal of the developing chick spinal cord in ovo resulted in guidance errors for commissural axons in the vicinity of the floor plate. This effect is similar to that resulting from treatment with antibodies against axonin-1, confirming that axonin-1/Nr-CAM interactions are important for guidance of commissural axons through a spatially and temporally restricted Nr-CAM positive domain in the ventral spinal cord. When tested as a substrate, Nr-Fc induced robust neurite outgrowth from dorsal root ganglion and sympathetic ganglion neurons, but it was not effective for tectal and forebrain neurons. The peripheral but not the central neurons expressed high levels of axonin-1 both in vitro and in vivo. Moreover, antibodies against axonin-1 inhibited Nr-Fc-induced neurite outgrowth, indicating that axonin-1 is a neuronal receptor for Nr-CAM on these peripheral ganglion neurons. The results demonstrate a role for Nr-CAM as a ligand in axon growth by a mechanism involving axonin-1 as a neuronal receptor and suggest that dynamic changes in Nr-CAM expression can modulate axonal growth and guidance during development.     

Intracellular signaling is changed after clustering of the neural cell adhesion molecules axonin-1 and NgCAM during neurite fasciculation

Neural cell adhesion molecules of the immunoglobulin/fibronectin type III family on axons have been implicated in promotion of neurite outgrowth, fasciculation, and the mediation of specific cell adhesion. The present study demonstrates that two of these molecules on dorsal root ganglion neurons are associated with distinct protein kinases, axonin-1 with the src-related nonreceptor tyrosine kinase fyn and NgCAM with a casein kinase II-related activity and a serine/ threonine kinase related to S6 kinase. When neurites grew without contacts involving axonin-1 and NgCAM, strong fyn kinase activity was associated with axonin-1, whereas the NgCAM-associated kinase activities were low. Clustering of axonin-1 with NgCAM induced by the formation of cell-cell contacts correlated with a reduction of the axonin-1-associated fyn activity and an increased phosphorylation of NgCAM by the associated casein kinase II-related activity. Thus, axonin-1 and NgCAM trigger distinctive intracellular signals during in vitro differentiation depending on their state of association.     

The CEACAM1 N-terminal Ig domain mediates cis- and trans-binding and is essential for allosteric rearrangements of CEACAM1 microclusters

Cell adhesion molecules (CAMs) sense the extracellular microenvironment and transmit signals to the intracellular compartment. In this investigation, we addressed the mechanism of signal generation by ectodomains of single-pass transmembrane homophilic CAMs. We analyzed the structure and homophilic interactions of carcinoembryonic antigen (CEA)–related CAM 1 (CEACAM1), which regulates cell proliferation, apoptosis, motility, morphogenesis, and microbial responses. Soluble and membrane-attached CEACAM1 ectodomains were investigated by surface plasmon resonance–based biosensor analysis, molecular electron tomography, and chemical cross-linking. The CEACAM1 ectodomain, which is composed of four glycosylated immunoglobulin-like (Ig) domains, is highly flexible and participates in both antiparallel (trans) and parallel (cis) homophilic binding. Membrane-attached CEACAM1 ectodomains form microclusters in which all four Ig domains participate. Trans-binding between the N-terminal Ig domains increases formation of CEACAM1 cis-dimers and changes CEACAM1 interactions within the microclusters. These data suggest that CEACAM1 transmembrane signaling is initiated by adhesion-regulated changes of cis-interactions that are transmitted to the inner phase of the plasma membrane.     

Pathological missense mutations of neural cell adhesion molecule L1 affect homophilic and heterophilic binding activities.

Mutations in the gene for neural cell adhesion molecule L1 (L1CAM) result in a debilitating X-linked congenital disorder of brain development. At the neuronal cell surface L1 may interact with a variety of different molecules including itself and two other CAMs of the immunoglobulin superfamily, axonin-1 and F11. However, whether all of these interactions are relevant to normal or abnormal development has not been determined. Over one-third of patient mutations are single amino acid changes distributed across 10 extracellular L1 domains. We have studied the effects of 12 missense mutations on binding to L1, axonin-1 and F11 and shown for the first time that whereas many mutations affect all three interactions, others affect homophilic or heterophilic binding alone. Patient pathology is therefore due to different types of L1 malfunction. The nature and functional consequence of mutation is also reflected in the severity of the resultant phenotype with structural mutations likely to affect more than one binding activity and result in early mortality. Moreover, the data indicate that several extracellular domains of L1 are required for homophilic and heterophilic interactions.     

Critical and optimal Ig domains for promotion of neurite outgrowth by L1/Ng‐CAM

Mammalian L1 and avian Ng‐CAM are homologous neural cell adhesion molecules (CAMs) that promote neurite outgrowth and cell adhesion in most neurons. Previous attempts to map these activities to discrete regions in the CAMs have suggested the involvement of a variety of different domains. However, these studies mainly used bacterially expressed proteins that were much less active on a molar basis than the native molecules. To define regions that are critical for maximal neurite outgrowth, we constructed and tested a panel of eukaryotically expressed proteins containing various extracellular segments of human L1 (hL1) or Ng‐CAM. Our results indicate that Ig domains 1–4 of hL1 are critical for homophilic binding and neurite outgrowth; however this segment is less potent than the entire extracellular region. Optimal neurite outgrowth activity was seen with proteins containing all six Ig domains of hL1 or Ng‐CAM. The adhesive properties of hL1 fragments correlated tightly with their neurite outgrowth activities, suggesting that these two processes are closely linked. These results suggest that Ig domains 1–4 form a structural cassette responsible for hL1 homophilic binding, while Ig domains 1–6 represent a functional region for optimal promotion of neurite outgrowth in vitro and possibly in vivo. © 2000 John Wiley & Sons, Inc. J Neurobiol 42: 287–302, 2000     

Neurite outgrowth on immobilized axonin-1 is mediated by a heterophilic interaction with L1(G4)

Axonin-1 is an axon-associated cell adhesion molecule with dualistic expression, one form being glycophosphatidylinositol-anchored to the axonal membrane, the other secreted from axons in a soluble form. When presented as a substratum for neuronal cultures it strongly promotes neurite outgrowth from chicken embryonic dorsal root ganglia neurons. In this study, the axon-associated cell adhesion molecule G4, which is identical with Ng-CAM and 8D9, and homologous or closely related to L1 of the mouse and NILE of the rat, was investigated with respect to a receptor function for axonin-1. Using fluorescent microspheres with covalently coupled axonin-1 or L1(G4) at their surface we showed that these proteins bind to each other. Within the sensitivity of this microsphere assay, no interaction of axonin-1 with itself could be detected. Axonin-1-coated microspheres also bound to the neurites of cultured dorsal root ganglia neurons. This interaction was exclusively mediated by L1(G4), as indicated by complete binding suppression by monovalent anti-L1(G4) antibodies. The interaction between neuritic L1(G4) and immobilized axonin-1 was found to mediate the promotion of neurite growth on axonin-1, as evidenced by the virtually complete arrest of neurite outgrowth in the presence of anti-L1(G4) antibodies. Convincing evidence has recently been presented that neurite growth on L1(8D9) is mediated by the homophilic binding of neuritic L1(G4) (1989. Neuron. 2: 1597-1603). Thus, both L1(G4)- and axonin-1-expressing axons may serve as "substrate pathways" for the guidance of following axons expressing L1(G4) into their target area. Conceivably, differences in the concentration of axonin-1 and L1(G4), and/or modulatory influences on their specific binding parameters in leading pathways and following axons could represent elements in the control of axonal pathway selection.     

Crystal structure of the Ig1 domain of the neural cell adhesion molecule NCAM2 displays domain swapping

The crystal structure of the first immunoglobulin (Ig1) domain of neural cell adhesion molecule 2 (NCAM2/OCAM/RNCAM) is presented at a resolution of 2.7 Å. NCAM2 is a member of the immunoglobulin superfamily of cell adhesion molecules (IgCAMs). In the structure, two Ig domains interact by domain swapping, as the two N-terminal β-strands are interchanged. β-Strand swapping at the terminal domain is the accepted mechanism of homophilic interactions amongst the cadherins, another class of CAMs, but it has not been observed within the IgCAM superfamily. Gel-filtration chromatography demonstrated the ability of NCAM2 Ig1 to form dimers in solution. Taken together, these observations suggest that β-strand swapping could have a role in the molecular mechanism of homophilic binding for NCAM2.     

Platelet endothelial cell adhesion molecule 1 (PECAM-1) and its interactions with glycosaminoglycans: 1. Molecular modeling studies

Platelet endothelial cell adhesion molecule 1 (PECAM-1) has many functions, including its roles in leukocyte extravasation as part of the inflammatory response and in the maintenance of vascular integrity through its contribution to endothelial cell-cell adhesion. PECAM-1 has been shown to mediate cell-cell adhesion through homophilic binding events that involve interactions between domain 1 of PECAM-1 molecules on adjacent cells. However, various heterophilic ligands of PECAM-1 have also been proposed. The possible interaction of PECAM-1 with glycosaminoglycans (GAGs) is the focus of this study. The three-dimensional structure of the extracellular immunoglobulin (Ig) domains of PECAM-1 were constructed using homology modeling and threading methods. Potential heparin/heparan sulfate-binding sites were predicted on the basis of their amino acid consensus sequences and a comparison with known structures of sulfate-binding proteins. Heparin and other GAG fragments have been docked to investigate the structural determinants of their protein-binding specificity and selectivity. The modeling has predicted two regions in PECAM-1 that appear to bind heparin oligosaccharides. A high-affinity binding site was located in Ig domains 2 and 3, and evidence for a low-affinity site in Ig domains 5 and 6 was obtained. These GAG-binding regions were distinct from regions involved in PECAM-1 homophilic interactions.     

Residues within a conserved amino acid motif of domains 1 and 4 of VCAM- 1 are required for binding to VLA-4

Vascular cell adhesion molecule 1 (VCAM-1), a member of the Ig superfamily originally identified on activated endothelium, binds to the integrin very late antigen-4 (VLA-4), also known as alpha 4 beta 1 or CD49d/CD29, to support cell-cell adhesion. Studies based on cell adhesion to two alternatively spliced forms of VCAM-1 or to chimeric molecules generated from them and intercellular adhesion molecule-1 (ICAM-1) have demonstrated two VLA-4 binding sites on the predominate form of VCAM-1. Here, we studied VLA-4-dependent adhesion of the lymphoid tumor cell line Ramos to cells expressing wild type and mutant forms of VCAM-1. Results based on domain deletion mutants demonstrated the existence and independence of two VLA-4-binding sites located in the first and fourth domains of VCAM-1. Results based on amino acid substitution mutants demonstrated that residues within a linear sequence of six amino acids found in both domain 1 and 4 were required for VLA-4 binding to either domain. Five of these amino acids represent a conserved motif also found in ICAM domains. We propose that integrin binding to these Ig-like domains depends on residues within this conserved motif. Specificity of integrin binding to Ig-like domains may be regulated by a set of nonconserved residues distinct from the conserved motif.     

Characterization of the L1-neurocan-binding site. Implications for L1-L1 homophilic binding

The L1 adhesion molecule is a 200-220-kDa membrane glycoprotein of the Ig superfamily implicated in important neural processes including neuronal cell migration, axon outgrowth, learning, and memory formation. L1 supports homophilic L1-L1 binding that involves several Ig domains but can also bind with high affinity to the proteoglycan neurocan. It has been reported that neurocan can block homophilic binding; however, the mechanism of inhibition and the precise binding sites in both molecules have not been determined. By using fusion proteins, site-directed mutagenesis, and peptide blocking experiments, we have characterized the neurocan-binding site in the first Ig-like domain of human L1. Results from molecular modeling suggest that the sequences involved in neurocan binding are localized on the surface of the first Ig domain and largely overlap with the G-F-C beta-strands proposed to interact with the fourth Ig domain during homophilic binding. This suggests that neurocan may sterically hinder a proper alignment of L1 domains. We find that the C-terminal portion of neurocan is sufficient to mediate binding to the first Ig domain of L1, and we suggest that the sushi domain cooperates with a glycosaminoglycan side chain in forming the binding site for L1.     

Distinct subpopulations of sensory afferents require F11 or axonin-1 for growth to their target layers within the spinal cord of the chick.

Dorsal root ganglion neurons project axons to specific target layers in the gray matter of the spinal cord, according to their sensory modality. Using an in vivo approach, we demonstrate an involvement of the two immunoglobulin superfamily cell adhesion molecules axonin-1/TAG-1 and F11/F3/contactin in subpopulation-specific sensory axon guidance. Proprioceptive neurons, which establish connections with motoneurons in the ventral horn, depend on F11 interactions. Nociceptive fibers, which target to layers in the dorsal horn, require axonin-1 for pathfinding. In vitro NgCAM and NrCAM were shown to bind to both axonin-1 and F11. However, despite this fact and despite their ubiquitous expression in the spinal cord, NgCAM and NrCAM are selective binding partners for axonin-1 and F11 in sensory axon guidance. Whereas nociceptive pathfinding depends on NgCAM and axonin-1, proprioceptive fibers require NrCAM and F11.     

Trimerization of cell adhesion molecule L1 mimics clustered L1 expression on the cell surface: influence on L1-ligand interactions and on promotion of neurite outgrowth

Several studies indicate that cell adhesion molecules have to be clustered on the cell surface to engage in adhesive functions. We investigated adhesive functions of clustered versus monomeric L1 extracellular parts in vitro to distinguish how clustering affects ligand binding and promotion of neurite outgrowth. Trimeric L1 was recombinantly expressed and covalently assembled by the cartilage matrix protein's coiled-coil domain. Trimeric L1 has an apparent molecular mass of approximately 380 kDa in the nonreduced form and approximately 130 kDa in the reduced form. Rotary shadowing electron micrographs of trimeric L1 revealed a rod-like shape terminating in three globular domains. Monomeric L1 assumes a horseshoe shape of domains Ig I-IV followed by a rod-like structure consisting of Ig V and VI and fibronectin type III 1-5. Circular dichroism measurements showed that the secondary structure consists of beta-sheets. Trimeric L1 binds to itself, to monomeric L1, to laminin-1, and to alpha5beta1 integrin in a concentration-dependent manner. In contrast, binding of monomeric L1 could only be saturated with itself but not with laminin-1 and with alpha5beta1 integrin. Promotion of neurite outgrowth from PC12 cells cultured on adsorbed trimeric L1 was increased by 100%, whereas on monomeric L1 the increase was only 50% over the control value. Promotion of neurite outgrowth from PC12 cells was specifically inhibited in a concentration-dependent manner by a polyclonal antibody against L1. These findings show that clustering of only three extracellular domains increases considerably L1's binding affinity to different ligands and enhances neurite outgrowth, suggesting that adhesive functions of L1 on the cell surface depend on cluster formation.     

Structure and interactions of NCAM Ig1-2-3 suggest a novel zipper mechanism for homophilic adhesion

The neural cell adhesion molecule, NCAM, mediates Ca(2+)-independent cell-cell and cell-substratum adhesion via homophilic (NCAM-NCAM) and heterophilic (NCAM-non-NCAM molecules) binding. NCAM plays a key role in neural development, regeneration, and synaptic plasticity, including learning and memory consolidation. The crystal structure of a fragment comprising the three N-terminal Ig modules of rat NCAM has been determined to 2.0 A resolution. Based on crystallographic data and biological experiments we present a novel model for NCAM homophilic binding. The Ig1 and Ig2 modules mediate dimerization of NCAM molecules situated on the same cell surface (cis interactions), whereas the Ig3 module mediates interactions between NCAM molecules expressed on the surface of opposing cells (trans interactions) through simultaneous binding to the Ig1 and Ig2 modules. This arrangement results in two perpendicular zippers forming a double zipper-like NCAM adhesion complex.     

A homologue of the axonally secreted protein axonin-1 is an integral membrane protein of nerve fiber tracts involved in neurite fasciculation

Axonin-1 is a glycoprotein that is released from axons of cultured neurons (Stoeckli, E. T., P. F. Lemkin, T. B. Kuhn, M. A. Ruegg, M. Heller, and P. Sonderegger. 1989. Eur. J. Biochem. 180:249-258). It has recently been purified from the ocular vitreous fluid of the chicken embryo (Ruegg, M. A., E. T. Stoeckli, T. B. Kuhn, M. Heller, R. Zuellig, and P. Sonderegger. 1989. EMBO (Eur. Mol. Biol. Organ.) J. 8:55-63). Immunohistochemistry localized axonin-1 prevalently in developing nerve fiber tracts. The presence of anti-axonin-1 Fab fragments during axon growth in vitro resulted in antibody binding to the axonal surfaces and in a marked perturbation of the fasciculation pattern. Hence, a fraction of axonin-1 is associated with axonal membranes and, by operational criteria, qualifies as a cell adhesion molecule. The major proportion of membrane-associated axonin-1 co-solubilized with the integral membrane proteins. By physico-chemical, immunological, and protein-chemical criteria, the integral membrane form was found to be highly similar to soluble axonin-1. In common with a number of other cell adhesion molecules, both soluble and membrane-bound axonin-1 express the L2/HNK-1 and the L3 epitopes. Radioactive pulse-chase and double-labeling experiments revealed that the released form was not derived from the membrane-bound form by shedding from the membrane surface, but directly secreted from an intracellular pool. Due to its high degree of similarity to the membrane-associated form and the presence of the L2/HNK-1 and L3 epitopes, reported to be ligands in adhesive cell interactions, adhesive properties are postulated for secreted axonin-1. As a soluble adhesive protein, it may function as a regulator of cell adhesion around its most likely site of secretion, the growth cone.