The L1 Family of Neural Review Cell Adhesion Molecules: Old Proteins Performing New Tricks

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Neuron, Vol. 17, 587–593, October, 1996, Copyright 1996 by Cell Press The L1 Family of Neural Review Cell Adhesion Molecules: Old Proteins Performing New Tricks

Michael Hortsch chordate phyla (Figure 2). Subsequently, at some point Department of Anatomy and Cell Biology during early chordate evolution, a further separation into University of Michigan three different L1 subfamilies occurred. As exemplified Ann Arbor, Michigan 48109-0616 by the two L1-CAM-type molecules found in zebrafish (Tongiorgi et al., 1995), additional members of the L1 family might have arisen by gene duplication events and may well exist in other vertebrate species. Introduction L1-type molecules are predominantly, although not The discovery that a range of neurological disorders in exclusively, expressed in developing peripheral and humans are caused by mutations in the gene for the central nervous systems. In vertebrates, L1-CAM, Nr- neural cell adhesion molecule L1-CAM has heightened CAM, and neurofascin exhibit a differential, dynamic interest in the role of L1-type proteins in nervous system expression pattern during nervous system develop- development and neuronal function. Many different ment. At various stages during embryonic and postem- members of the L1 gene family (molecular designations bryonic development, many types of neurons and glial are indicated in parenthesis after the species name) cells express one or several of these L1-type molecules have now been found in a number of invertebrate and on their surface (Moscoso and Sanes, 1995). Experimen- vertebrate species. These include the grasshopper tal evidence indicates that these molecules are involved Schistocerca americana (neuroglian), the moth Man- in diverse cellular processes, such as myelination, neu- duca sexta (neuroglian/3B11), the fruitfly Drosophila rite outgrowth, growth cone morphology, axon fascicu- melanogaster (neuroglian), zebrafish (L1.1- and L1.2- lation and pathfinding, neuronal cell migration, and long- CAM), goldfish (L1-CAM/E587), chicken (Ng-CAM/G4/ term potentiation in the hippocampus. These functions 8D9, neurofascin, Nr-CAM/Bravo), mouse (L1-CAM, are mediated by a number of different protein–protein neurofascin, Nr-CAM), rat (NILE, neurofascin/ABGP, Nr- interactions. CAM), and humans (L1-CAM/5G3, neurofascin, Nr- Outside the nervous system, L1-CAM is expressed CAM). A potential nematode L1 homolog (L1-like mole- by leukocytes and epithelial cells of the intestine and cule) has recently been identified as a result of the urogenital tract (Kowitz et al., 1992; Kujat et al., 1995; Caenorhabditis elegans genome sequencing effort. Probstmeier et al., 1990). A number of transformed cells These molecules are involved in a multitude of molecular from a variety of tissues also express L1-CAM (Mujoo interactions with other extra- and intracellular proteins. et al., 1986), and L1-CAM expression inversely corre- In this review, I consider the relationship between the lates with the metastatic capacity of a lymphoma cell molecular interactions, the known cellular L1 functions, line in mice (Kowitz et al., 1993), leading to the specula- and the phenotypic expression of human L1-CAM mutation that L1-CAM may also play a role in metastatic tions and summarize in a unifying hypothesis, how these events. different levels of L1 activities may be connected. Molecular Interactions and In Vitro Functions of L1 Molecules Structural Features and Expression Over the last few years, a number of new ligands as Patterns of L1 Family Members well as novel functions of this divergent group of cell All members of this gene family share a basic structural adhesion molecules have been identified. All L1 family plan of six extracellular immunoglobulin domains, fol- members are strong homophilic, Ca2ϩ-independent cell lowed by five fibronectin type III (FNIII) domains and adhesion molecules. Attempts to map the homophilic a hydrophobic transmembrane segment with a short adhesive activity of L1 family members have implicated cytoplasmic domain of 85–147 amino acids (Figure 1). various sets of protein domains (Figure 1). In the human Differentially spliced cDNAs that generate several pro- L1-CAM molecule, this activity has now been pinpointed tein isoforms differing in their extracellular domain have to the second immunoglobulin domain (Zhao and Siu, been described for several L1 family members (Davis et 1995). In contrast with several other classes of adhesion al., 1993; Grumet et al., 1991; Kayyem et al., 1992; molecules, the homophilic adhesion function of mem- Takeda et al., 1996; Volkmer et al., 1992). It is not known, bers of the L1 family is relatively independent of the however, whether any of these differentially spliced cytoplasmic domain and its interaction with cytoskeletal polypeptide inserts conveys any unique function to the elements. L1-CAM and Drosophila neuroglianmolecules entire L1 protein. lacking their endogenous cytoplasmic domain are still So far, only a single neuroglian gene has been found fully active in cell adhesion and are able to stimulate in the several arthropod species analyzed. In contrast, neurite outgrowth (Doherty et al., 1995; Hortsch et al., up to three different L1-type genes have been identified 1995; Wong et al., 1995a). in higher vertebrates (usually referred to as L1-CAM, Nr- In neurons, homophilic L1 interactions trigger the acti- CAM, and neurofascin). A phylogenetic analysis of the vation of neuronal FGF receptors (Williams et al., 1994). currently available L1 sequences indicates that they all This initiates a second messenger cascade that ulti- probably originated from a common ancestral prototype mately results in the activation of Ca2ϩ channels and L1 molecule predating the split of the arthropod and the induction of neurite outgrowth. However, the exact Neuron 588

Figure 1. Protein Domain Model of a Generic L1 Molecule Functionally important domains that have been mapped in the chicken Nr-CAM (Mauro et al., 1992), the human L1-CAM (Zhao and Siu, 1995), and the Drosophila neuroglian molecule (A. Bieber, personal communication) as well as the positions of other important structural features, like RGD motifs, potential protease cleavage sites, homologies to FGF receptors, and the differentially spliced -RSLE- miniexon, are indicated.

mode of interaction of L1 family members with neuronal expressing the RPTP␨/␤ ligand. In support of this model, FGF receptors is still not known. an L1-dependent stimulationof protein phosphatase ac- L1-dependent neurite outgrowth can also be triggered tivity in growth cone–enriched membranes has been by a heterophilic interaction with the DM-GRASP cell reported by Klinz et al. (1995). Phosphacan as well as adhesion molecule (DeBernardo and Chang, 1996), the neurocan inhibit Ng-CAM- and rat L1-CAM-mediated GPI-anchored membrane proteins F3/F11 (Morales et neuronal adhesion and neurite outgrowth (Friedlander al., 1993), and axonin-1/TAG-1 (Felsenfeld et al., 1994; et al., 1994; Milev et al., 1994). Currently, there is no Kuhn et al., 1991), all of which are also members of the clear evidence that laminin is able to induce neurite immunoglobulin domain superfamily. Other ligands that outgrowth through its interaction with L1-CAM. interact with vertebrate L1 family members include the Recent evidence suggests that an RGD motif found extracellular matrix molecule laminin (Grumet et al., in the sixth immunoglobulin domain of several L1-CAM 1993b) and the two chondroitin sulfate proteoglycans, subgroup members (Figure 1) is recognized by several neurocan and phosphacan (Grumet et al., 1993a). The RGD-specific integrin heterodimers (Montgomery et al., phosphacan proteoglycan represents the extracellular 1996; Ruppert et al., 1995). All neurofascins and chicken domain of a larger membrane protein, called RPTP␨/␤, Ng-CAM also have a single RGD motif in their third that has an intracellular protein-tyrosine phosphatase FNIII domain (see Figure 2), but whether it can also be domain. This protein-tyrosine phosphatase gene is ex- recognized by and bind to RGD-specific integrins is pressed in many glial and other cell types, which sug- currently unknown. Whether or not Ng-CAM is the gests the possibility that interactions with L1 family chicken equivalent of L1-CAM is controversial. At the members may influence phosphotyrosine levels in cells sequence level, Ng-CAM is most closely related to Review 589

Figure 2. Evolutionary Tree of the L1 Gene Family L1 cDNA sequences encoding most of the cytoplasmic protein domain (starting with the first basic amino acid residue after the transmembrane segment and ending with the conserved tyrosine residue at the end of the putative ankyrin binding domain) were aligned using the multiple alignment option of the MacDNASIS Pro 3.0 program package (Higgins–Sharp algorithm). The differentially spliced -RSLE- miniexon was omitted from the analysis. A phylogenetic tree was con- structed from the aligned sequences by using the DNAMLK program (based on the maximum likelihood method described by Felsen- stein [1981]) from the PHYLIP program package. Boxed are L1 representatives with a differentially spliced cytoplasmic -RSLE- miniexon (the goldfish sequence contains a homologous -KSLE- motif instead). Indicated are also groups of L1 proteins sharing homologous RGD motifs in either the sixth immunoglobulin or the third FNIII protein domain. Most cDNA sequences were downloaded from GenBank. cDNA sequences for mouse Nr-CAM and neurofascin, as well as for human Nr-CAM were directly entered from the published records (Lane et al., 1996; Mos- coso and Sanes, 1995). Unpublished cDNA sequences were communicated by J. Rehm and C. Goodman (grasshopper neuroglian), C.-L. Chen and J. Nardi (Manduca neuroglian), S. Giordano and C. Stuermer (goldfish L1-CAM), and V. Bennett (rat Nr-CAM). the other vertebrate L1-CAMs, making it a member of the grasshopper or Manduca neuroglian protein se- the L1-CAM subgroup, but it contains an RGD motif in quences. It appears that the -RSLE- miniexon arose its third FNIII domain, as do the neurofascins, rather after the separation of the chordate from the arthropod than in its sixth immunoglobulin domain (Figure 2). This lineage (Figure 2). L1-CAMs with or without the RSLE might indicate that the RGD motif in the third FNIII do- insert in their cytoplasmic domain have very similar abili- main arose twice independently during evolution. ties to induce cell aggregation and neurite outgrowth. No RGD motif has been found in any of the cloned Nr- However, L1-CAM containing these four amino acids CAMs, fish L1-CAMs, or any of the insect neuroglians. supports cell migration on L1-CAM substrates much Therefore, the ability of L1 family members to interact better than the RSLE-minus isoform (Takeda et al., with RGD-specific integrins mustbe a fairly recent evolu- 1996). tionary acquisition (Figure 2). The functional significance All vertebrate as well as invertebrate members of the of L1-CAM recognition by integrins is not well under- L1 family share two very conserved amino acid seg- stood. Currently, there is no indication for a role of a ments in their cytoplasmic domain. Davis et al. provided direct heterophilic interaction between L1-CAM and evidence that these conserved amino acid segments RGD-specific integrin in the nervous system. Physiologi- constitute a binding site for the cytoskeletal linker pro- cal processes in which the L1-CAM interaction with inte- tein ankyrin, thereby anchoring L1-type molecules to grins could be important are the interactions of leuko- the submembranous actin–spectrin cytoskeleton (Davis cytes with fibroblasts, epithelial and endothelial cells, et al., 1993; Davis and Bennett, 1994). In Drosophila S2 and T lymphocytes and platelets. Upon leukocyte acti- cells, this interaction between the neuroglian cyto- vation, the L1-CAM level is rapidly down-regulated plasmic domain and ankyrin is completely dependent (Hubbe et al., 1993). on the extracellular homophilic adhesive function of the Portions of the L1 cytoplasmic domain are especially neuroglian molecule (Dubreuil et al., 1996). Moreover, well conserved, suggesting that they may be involved this interaction and its regulation by cell adhesion is in important cellular functions. One of these features is evolutionarily extremely well conserved, such that hu- a differentially spliced miniexon (-RSLE-) that has been man L1-CAM can interact with Drosophila ankyrin in S2 found in the middle of the cytoplasmic domain of almost cells in the same manner as Drosophila neuroglian all vertebrate L1 family members (Figures 1 and 2). This (M. H., unpublished data). Dubreuil et al. (1996) specu- miniexon appears to be included in L1-type molecules lated that L1 family members may be able to translate that are expressed by neuronal cells and is missing from cell–cell adhesion events into the recruitment of cy- L1 polypeptides of non-neuronal origin (Moscoso and toskeletal elements to these sites. This in turn may facili- Sanes, 1995; Takeda et al., 1996). The genomic structure tate the establishment or maintenance of neuronal cell of the Drosophila neuroglian gene rules out the exis- polarity by attracting other membrane cytoskeleton- tence of an -RSLE- miniexon in fruitflies (M. H., unpub- binding proteins such as ion pumps, ion channels, and lished data), and no such peptide motif was found in other signaling molecules to sites of cell–cell contact Neuron 590

and may thereby create specific subdomains within the several cell lines suggests that the alcohol-induced inhi- plasma membrane. bition of L1-CAM function during embryonic development could be an important factor in FAS (Ramanathan Neurological Disorders Caused by Human et al., 1996). L1-CAM Mutations As expected for recessive, X chromosome–linked mu- In humans, a set of three phenotypically overlapping, tations, the majority of reported cases with mutations hereditary, neurological syndromes (X-linked hydro- in the L1-CAM gene are affected males born to asymto- cephalus [HSAS], MASA syndrome, and X-linked spastic matic female carriers. However, a few cases of female paraplegia [SPG1]) have been linked to mutations in the carriers expressing some phenotypes, rangingfrom mild L1-CAM gene located on the human X chromosome mental retardation and adducted thumbs to neonatal (Wong et al., 1995b). X-linked hydrocephalus is the most hydrocephalus, have been reported (Kaepernick et al., common cause of hereditary hydrocephalus, affecting 1994). Most likely, these cases are due to skewed X about 1 in every 30,000 male births. It is also referred chromosome inactivation in these individuals. to as hydrocephalus due to stenosis of the aqueduct A large number of different mutations in the human of Sylvius and is abbreviated as HSAS. This historic L1-CAM gene have now been identified and character- designation is derived from a constriction in the aque- ized at the molecular level (see Figure 3). So far, each duct of Sylvius, which can be observed in some, but of these mutations is restricted to the one family in which not all patients suffering from this syndrome. In extreme it has been originally identified. The clinical diagnoses cases, L1-CAM-mediated hydrocephalus will result in of related individuals sharing the same mutation some- pre- or neonatal death, whereas milder forms have no times diverge considerably, with distinct cases of HSAS, significant effect on patient survival. Among the individ- MASA syndrome, and SPG1 found in the same family uals who do not exhibit a visible hydrocephalus, an (Fransen et al., 1994; Jouet et al., 1995; Ruiz et al., 1995; enlargement of brain ventricles can often be diagnosed S. Claes and E. Legius, personal communication). The using brain CAT scanning. Almost all surviving patients analysis of the gene structure and genomic DNA se- with L1-CAM mutations suffer from various degrees of quences of these L1-CAM mutations has revealed single mental retardation, ranging from slight learning disabili- base pair changes as well as deletions and duplications. ties to severe mental dysfunctions. Other common phe- These different types of mutations have very disparate notypes, which are observed in many but not all affected effects on the L1-CAM gene product. Most are very patients, are adducted thumbs, micro- or macrocephaly, subtle, resulting in single amino acid changes. Other delayed speech development, shuffling gait, and spas- mutations, however, cause more dramatic modifications ticity, especially of the lower limbs. Also, hypoplasia of of the L1-CAM protein, e.g., the creation of an in-frame the corpus callosum and the septum pellucidum have stop codon or changes in the L1-CAM mRNA splicing been found in a number of cases. The putative role of pattern, resulting in frameshift mutations or deletions of L1-CAM in developmental processes, such as neuronal entire amino acid segments. Some of these more severe cell migration, axonal growth and pathfinding, and my- mutations produce secreted, truncated L1-CAM mole- elination, may well account for some of these neurologi- cules that probably represent functional null mutations. cal phenotypes. Considering the wide range of pheno- In contrast, most of the single amino acid changes map types caused by L1-CAM mutations and the variable to domains within the L1-CAM molecule, which are not levels of phenotypic expression, it is somewhat surpris- involved in known L1 functions, like homophilic adhe- ing that no obvious linkage has been found between the sion and neurite outgrowth (Figures 1 and 3). Interest- occurrence or the severity of specific phenotypes. ingly, no correlation has been established between spe- A comparison of disease phenotypes and the analysis cific types of mutations and the severity of clinical of larger families with L1-CAM mutations confirmed that phenotypes. Furthermore, neither the location of L1- several other neurological genetic syndromes, which CAM mutations nor their resulting specific clinical phe- have been mapped to Xq28 on the long arm of the notypes appear to cluster in specific protein domains. X chromosome, are allelic to HSAS. These syndromes L1-CAM mutations have been identified in almost every include MASA syndrome (for mental retardation, ad- L1-CAM protein domain, and many of them give rise to ducted thumbs, spastic paraplegia, and aphasia), one type of X-linked complicated spastic paraplegia (SPG1), the full spectrum of clinical symptoms described above and certain forms of corpus callosum agenesis (ACC) (Figure 3). This suggests that the molecular function(s) or dysgenesis (DCC). Summarizing the various clinical that are disrupted by these various mutations are not manifestations of this group of syndromes, Fransen et al. concentrated in specific protein domains but rather re- (1995) coined the acronym CRASH (for corpus callosum quire the entire, intact L1-CAM polypeptide. hypoplasia, retardation, adducted thumbs, spastic para- The finding that the phenotypes caused by specific plegia, and hydrocephalus) to describe the range of phe- L1-CAM mutations and their severity can be extremely notypes caused by L1-CAM mutations. Interestingly, variable suggests that the L1-CAM gene locus or its several of the phenotypes linked to L1-CAM mutations, product may exhibit more complex genetic interactions like hydrocephalus, agenesis of the corpus callosum, with other genes and their products. Such genes could and mental retardation, overlap with someof the anoma- modify the L1-CAM mutant phenotypes by various lies found in infants and neonates suffering from fetal mechanisms. Since a number of proteins directly inter- alcohol syndrome (FAS). The finding that levels of etha- act with L1-CAM, these gene products could act as nol comparable to those observed after moderate alco- suppressors and may be able to compensate for a lack hol consumption inhibit L1-mediated cell adhesion in of certain L1 functions. Therefore, it will be interesting Review 591

Figure 3. Changes in the Human L1-CAM Polypeptide That Are Caused by Identified, Known Mutations in the Human L1-CAM Gene Deletions, nonsense mutations, and splice mutations resulting in deletions of amino acid segments or frameshift mutations are shown on the left side of the diagram. Missense mutations with their respective amino acid changes are shown on the right side. The amino acid residues altered by the different mutations are indicated following the nomen- clature for the human L1-CAM sequence published by Hlavin and Lemmon (1991). The pri- mary clinical diagnosis for each specific mutation is indicated as well (HSAS, X-linked hydrocephalus; MASA, MASA syndrome; SPG1, X-linked spastic paraplegia; ACC, agenesis of the corpus callosum). The infor- mation for some unpublished L1-CAM mutations was provided by S. Kenwrick, S. Claes, and E. Legius.

to analyze, in mice or fruitflies, double mutant pheno- can cause neurological malfunctions similar to those types of mutations for L1-CAM/neuroglian with muta- observed for L1-CAM in humans. tions for axonin-1/TAG-1, F3/F11, or any of the other L1-CAM-interacting proteins mentioned in the first half Correlation between In Vitro Functions of L1-CAM of this review. The expression or activity levels of these and Mutational Phenotypes in Humans proteins may have a significant influence on the expres- According to the molecular features of the mutated L1- sion the L1-CAM mutant phenotype. A different explana- CAM gene products, L1-CAM mutations can be subdi- tion for the variable L1-CAM phenotype could be that vided into three major classes, one class with point certain L1-CAM functions are redundant and can be mutations in the extracellular domain, one class re- carried out by other molecules in parallel pathways. A sulting in truncated, secreted molecules of various plethora of other cell surface molecules with adhesive lengths, and a third class with mutations (truncations, and neurite outgrowth promoting activities are ex- frameshift and point mutations) in the cytoplasmic L1 pressed in the developing nervous system. These mole- domain (Wong et al., 1995b). Since mutations in both cules may be able to compensate partially or substitute extra- as well as intracellular L1-CAM domains result in for mutant L1-CAM molecules. Premier candidates for the same range of syndromes, it is currently difficult to such redundant molecules are other members of the L1 correlate the observed phenotypes with any specific family, especially neurofascin and Nr-CAM. Moscoso characterized molecular L1-CAM function. It has been and Sanes (1995) demonstrated that the local and tem- shown that the cytoplasmic domain of L1 molecules is poral expression pattern of the three L1 family members dispensable for several if not all of their extracellular (L1-CAM, neurofascin, and Nr-CAM) is partially overlap- functions, such as homophilic adhesion and neurite out- ping in the developing mouse spinal cord. Both neuro- growth stimulation (Doherty et al., 1995; Hortsch et al., fascin and Nr-CAM have been recently mapped to differ- 1995; Wong et al., 1995a). The only known L1 function ent chromosomal locations in humans and mice that could be potentially affected by all three types of (Burmeister et al., 1996; Lane et al., 1996). Unfortunately, mutations is the interaction of L1-CAM with ankyrin. so far it is not known whether mutations in these genes Point mutations or deletions in the extracellular domain Neuron 592

may interfere with the adhesive activity of L1-CAM and, Fransen, E., Schrander-Stumpel, C., Vits, L., Coucke, P., Van Camp, as a result, may indirectly block ankyrin binding. Zhao G., and Willems, P.J. (1994). X-linked hydrocephalus and MASA and Siu (1996) recently analyzed two missense muta- syndrome present in one family are due to a single missense mutation in exon 28 of the L1CAM gene. Hum. Mol. Genet. 3, 2255–2256. tions in the second immunoglobulin domain of the hu- Fransen, E., Lemmon, V., Van Camp, G., Vits, L., Coucke, P., and man L1-CAM polypeptide (Arg184Gln and His210Gln; Willems, P.J. (1995). CRASH syndrome: clinical spectrum of corpus see Figure 3) for their ability to support homophilic, L1- callosum hypoplasia, retardation, adducted thumbs, spastic para- mediated cell adhesion and neurite outgrowth. One mu- paresis and hydrocephalus due to mutations in one single gene, L1. tation inactivated these L1-CAM functions completely, Eur. J. Hum. Genet. 3, 273–284. whereas the other mutant protein was still partially ac- Friedlander, D.R., Milev, P., Karthikeyan, L., Margolis, R.K., Margolis, tive. This raises the possibility that some of these muta- R.U., and Grumet, M. (1994). The neuronal chondroitin sulfate pro- tions may not inactivate L1-mediated adhesion at all, teoglycan neurocan binds to the neural cell adhesion molecules but rather disrupt the transmission of a signal to the Ng-CAM/L1/NILE and N-CAM, and inhibits neuronal adhesion and neurite outgrowth. J. Cell Biol. 125, 669–680. cytoplasmic domain, which otherwise would lead to an- Grumet, M., Mauro, V., Burgoon, M.P., Edelman, G.M., and Cunning- kyrin binding. Missing the cytoplasmic ankyrin binding ham, B.A. (1991). Structure of a new nervous system glycoprotein, domain, secreted L1-molecules are by definition not Nr-CAM, and relationship to subgroups of neural celladhesion mole- able to interact with the intracellular cytoskeleton. L1- cules. J. Cell Biol. 113, 1399–1412. CAM molecules with a mutation in their cytoplasmic Grumet, M., Flaccus, A., and Margolis, R.U. (1993a). Functional char- domain are expected to function properly in mediating acterization of chondroitin sulfate proteoglycans of brain: interac- extracellular adhesion and neurite outgrowth promotion; tions with neurons and neural cell adhesion molecules. J. Cell Biol. however, they would be unable to translate this activity 120, 815–824. into cytoskeleton binding and rearrangement. An analy- Grumet, M., Friedlander, D.R., and Edelman, G.M. (1993b). Evidence sis of the functional capabilities of different mutated for the binding of Ng-CAM to laminin. Cell Adhes. Commun. 1, 177–190. human L1-CAMs should shed light on the molecular defects leading to this wide range of phenotypes includ- Hlavin, M.L., and Lemmon, V. (1991). Molecular structure and functional testing of human L1CAM: an interspecies comparison. Geno- ing mental retardation, hydrocephalus, macrocephaly, mics 11, 416–423. the agenesis of the corpus callosum, spastic parapare- Hortsch, M., Wang, Y.M., Marikar, Y., and Bieber, A.J. (1995). The sis, aphasia, and adducted thumbs. cytoplasmic domain of theDrosophila celladhesion moleculeneuro- glian is not essential for its homophilic adhesive properties in S2 Acknowledgments cells. J. Biol. Chem. 270, 18809–18817. Hubbe, M., Kowitz, A., Schirrmacher, V., Schachner, M., and Alte- I would like to thank my many colleagues who communicated their vogt, P. (1993). L1 adhesion molecule on mouse leukocytes: regula- unpublished results and cDNA sequences to me or who suggested tion and involvement in endothelial cell binding. Eur. J. Immunol. many improvements to the manuscript. Work in my laboratory is 23, 2927–2931. supported by a grant from the National Institutes of Health Jouet, M., Moncla, A., Paterson, J., McKeown, C., Fryer, A., Carpen- (HD29388). ter, N., Holmberg, E., Wadelius, C., and Kenwrick, S. (1995). New domains of neural cell-adhesion molecule L1 implicated in X-linked References hydrocephalus and MASA syndrome. Am. J. Hum. Genet. 56, 1304– 1314. Burmeister, M., Ren, Q., Makris, G.J., Samson, D., and Bennett, V. Kaepernick, L., Legius, E., Higgins, J., and Kapur, S. (1994). Clinical (1996). Genes for the neuronal immunoglobulin domain celladhesion aspects of the MASA syndrome in a large family, including express- molecules neurofascin and Nr-CAM map to mouse chromosome 1 ing females. Clin. Genet. 45, 181–185. and 12 and homologous human chromosomes. Mamm. Genome 7, Kayyem, J.F., Roman, J.M., de la Rosa, E.J., Schwarz, U., and 558–559. Dreyer, W.J. (1992). Bravo/Nr-CAM is closely related to the cell Davis, J.Q., and Bennett, V. (1994). Ankyrin binding activity shared by adhesion molecules L1 and Ng-CAM and has a similar heterodimer the neurofascin/L1/NrCAM family of nervous system cell adhesion structure. J. Cell Biol. 118, 1259–1270. molecules. J. Biol. Chem. 269, 27163–27166. Klinz, S.G., Schachner, M., and Maness, P.F. (1995). L1 and N-CAM Davis, J.Q., McLaughlin, T., and Bennett, V. (1993). Ankyrin-binding antibodies trigger protein phosphatase activity in growth cone– proteins related to nervous system cell adhesion molecules: candi- enriched membranes. J. Neurochem. 65, 84–95. dates to provide transmembrane and intercellular connections in Kowitz, A., Kadmon, G., Eckert, M., Schirrmacher, V., Schachner, adult brain. J. Cell Biol. 121, 121–133. M., and Altevogt, P. (1992). Expression and function of the neural DeBernardo, A.P., and Chang, S. (1996). Heterophilic interactions cell adhesion molecule L1 in mouse leukocytes. Eur. J. Immunol. of DM-GRASP: GRASP–NgCAM interactions involved in neurite ex- 22, 1199–1205. tension. J. Cell Biol. 133, 657–666. Kowitz, A., Kadmon, G., Verschueren, H., Remels, L., De Baetselier, Doherty, P., Williams, E., and Walsh, F.S. (1995). A soluble chimeric P., Hubbe, M., Schachner, M., Schirrmacher, V., and Altevogt, P. form of the L1 glycoprotein stimulates neurite outgrowth. Neuron (1993). Expression of L1 cell adhesion molecule is associated with 14, 57–66. lymphoma growth and metastasis. Clin. Exp. Metastasis 11, Dubreuil, R.R., MacVicar, G., Dissanayake, S., Liu, C., Homer, D., 419–429. and Hortsch, M. (1996). Neuroglian-mediated cell adhesion induces Kuhn, T.B., Stoeckli, E.T., Condrau, M.A., Rathjen, F.G., and assembly of the membrane skeleton at cell contact sites. J. Cell Sonderegger, P. (1991). Neurite outgrowth on immobilized axonin-1 Biol. 133, 647–655. is mediated by a heterophilic interaction with L1(G4). J. Cell Biol. Felsenfeld, D.P., Hynes, M.A., Skoler, K.M., Furley, A.J., and Jessell, 115, 1113–1126. T.M. (1994). TAG-1 can mediate homophilic binding, but neurite Kujat, R., Miragall, F., Krause, D., Dermietzel, R., and Wrobel, K.H. outgrowth on TAG-1 requires an L1-like molecule and ␤1 integrins. (1995). Immunolocalization of the neural cell adhesion molecule L1 Neuron 12, 675–690. in non-proliferating epithelial cells of the male urogenitaltract. Histo- Felsenstein, J. (1981). Evolutionary trees from DNA sequences: a chemistry 103, 311–321. maximum likelihood approach. J. Mol. Evol. 17, 368–376. Lane, R.P., Chen, X.-N., Yamakawa, K., Vielmetter, J., Korenberg, Review 593

J.R., and Dreyer, W.J. (1996). Characterization of a highly conserved human homolog to the chicken neural cell surface protein Bravo/Nr- CAM that maps to chromosome band 7q31. Genomics 35, 456–465. Mauro, V.P., Krushel, L.A., Cunningham, B.A., and Edelman, G.M. (1992). Homophilic and heterophilic binding activities of Nr-CAM, a nervous system cell adhesion molecule. J. Cell Biol. 119, 191–202. Milev, P., Friedlander, D.R., Sakurai, T., Karthikeyan, L., Flad, M., Margolis, R.K., Grumet, M., and Margolis, R.U. (1994). Interactions of the chondroitin sulfate proteoglycan phosphacan, the extracellular domain of a receptor-type protein tyrosine phosphatase, with neurons, glia, and neural cell adhesion molecules. J. Cell Biol. 127, 1703–1715. Montgomery, A.M.P., Becker, J.C., Siu, C.H., Lemmon, V.P., Cher- esh, D.A., Pancook, J.D., Zhao, X.N., and Reisfeld, R.A. (1996). Hu- man neural cell-adhesion molecule L1 and rat homolog NILE are ligands for integrin ␣V␤3. J. Cell Biol. 132, 475–485. Morales, G., Hubert, M., Bru¨ mmendorf, T., Treubert, U., Tarnok, A., Schwarz, U., and Rathjen, F.G. (1993). Induction of axonal growth by heterophilic interactions between the cell surface recognition proteins F11 and Nr-CAM/Bravo. Neuron 11, 1113–1122. Moscoso, L.M., and Sanes, J.R. (1995). Expression of four immunoglobulin superfamily adhesion molecules (L1, Nr-CAM/Bravo, neurofascin/ABGP, and N-CAM) in the developing mouse spinal cord. J. Comp. Neurol. 352, 321–334. Mujoo, K., Spiro, R.C., and Reisfeld, R.A. (1986). Characterization of a unique glycoprotein antigen expressed on the surface of human neuroblastoma cells. J. Biol. Chem. 261, 10299–10305. Probstmeier, R., Martini, R., Tacke, R., and Schachner, M. (1990). Expression of the adhesion molecules L1, N-CAM and J1/tenascin during development of the murine small intestine. Differentiation 44, 42–55. Ramanathan, R., Wilkemeyer, M.F., Mittal, B., Perides, G., and Char- ness, M.E. (1996). Alcohol inhibits cell–cell adhesion mediated by human L1. J. Cell Biol. 133, 381–390. Ruiz, J.C., Cuppens, H., Legius, E., Fryns, J.P., Glover, T., Marynen, P., and Cassiman, J.J. (1995). Mutations in L1-CAM in two families with X linked complicated spastic paraplegia, MASA syndrome, and HSAS. J. Med. Genet. 32, 549–552. Ruppert, M., Aigner, S., Hubbe, M., Yagita, H., and Altevogt, P. (1995). The L1 adhesion molecule is a cellular ligand for VLA-5. J. Cell Biol. 1881–1891. Takeda, Y., Asou, H., Murakami, Y., Miura, M., Kobayashi, M., and Uyemura, K. (1996). A nonneuronalisoform of cell adhesion molecule L1: Tissue-specific expression and functional analysis. J. Neuro- chem. 66, 2338–2349. Tongiorgi, E., Bernhardt, R.R., and Schachner, M. (1995). Zebrafish neurons express two L1-related molecules during early axonogen- esis. J. Neurosci. Res. 42, 547–561. Volkmer, H., Hassel, B., Wolff, J.M., Frank, R., and Rathjen, F.G. (1992). Structure of the axonal surface recognition molecule neurofascin and its relationship to a neural subgroup of the immunoglobulin superfamily. J. Cell Biol. 118, 149–161. Williams, E.J., Furness, J., Walsh, F.S., and Doherty, P. (1994). Acti- vation of the FGF receptor underlies neurite outgrowth stimulated by L1, N-CAM, and N-cadherin. Neuron 13, 583–594. Wong, E.V., Cheng, G.H., Payne, H.R., and Lemmon, V. (1995a). The cytoplasmic domain of the cell-adhesion molecule L1 is not required for homophilic adhesion. Neurosci. Lett. 200, 155–158. Wong, E.V., Kenwrick, S., Willems, P., and Lemmon, V. (1995b). Mutations inthe cell adhesion molecule L1 cause mental retardation. Trends Neurosci. 18, 168–172. Zhao, X., and Siu, C.-H. (1995). Colocalization of the homophilic binding site and the neuritogenic activity of the cell adhesion molecule L1 to its second Ig-like domain. J. Biol. Chem. 270, 29413– 29421. Zhao, X., and Siu, C.-H. (1996). Differential effects of two hydrocephalus/MASA syndrome-related mutations on the homophilic binding and neuritogenic activities of the cell adhesion molecule L1. J. Biol. Chem. 271, 6563–6566.