[Frontiers in Bioscience 8, s1210-1225, September 1, 2003]

THE L1CAM EXTRACELLULAR REGION: A MULTI-DOMAIN PROTEIN WITH MODULAR AND COOPERATIVE BINDING MODES

Jeffrey Haspel and Martin Grumet

W.M. Keck Center for Collaborative Neuroscience and Dept. of Cell Biology and Neuroscience, Rutgers University, Piscataway, NJ 08854-8082, USA

FIGURES

Figure 1. Domain structure of the L1CAM extracellular region A: Schematic of L1CAM. Blue ovals (Ig1-Ig6) represent I-set Ig domains, and orange rectangles represent (Fn1-Fn5) represent S-set domains. The transmembrane domain is denoted by a vertical bar. The cytoplasmic region is located at the C-terminus of the molecule following the transmembrane domain. The location of proteolytic cleavage sites recognized by plasmin and ADAM10 are denoted by arrows. The N-terminal exon 2 sequence and the 7 residue linker sequence between Ig2 and Ig3 are depicted above the cartoon. Below the schematic, the approximate molecular weights of L1CAM and its proteolytic fragments are shown. B: Ribbon diagram of an I-set Ig domain, which is representative of Ig1-Ig6. b-strands that belong to Sheet I are shaded green and those that belong to Sheet II are shaded orange. C: Ribbon diagram of an S-set Ig domain, which is representative of Fn1-Fn5.

Figure 2. Map of binding activities in the L1CAM extracellular region. Summary of structure-function studies that identified critical L1CAM extracellular domains required for particular molecular interactions. The left hand column lists names of L1CAM binding partners (Protein). Domains shown to be critical for binding are shaded. Domains that are not critical but nonetheless may enhance binding affinity are shaded lightly. In some cases, different species homologues of a given protein were utilized in different studies (for example human TAX-1 and chick axonin-1). Results obtained using different species homologues are grouped together. Note that for L1CAM homophilic binding in trans, several studies arrived at divergent conclusions about which domains are sufficient for binding. It should also be noted that all proteins were generated in eukaryotic systems except for the work by Zhao and Siu (49) and the work with integrins (46-48,71), which were generated in bacteria.

Figure 3. The L1CAM horseshoe structure and its proposed impact on activity A: The L1CAM extracellular region in folded (horseshoe) and extended conformations. The horseshoe structure is depicted schematically on the left side of the panel, and the extended form of the extracellular region is shown on the right. Note that the cartoon of L1CAM is shortened to highlight Ig1-Ig6. Dashed lines within the horseshoe represent hydrogen bonds that are proposed to link Ig1 to Ig4 and Ig2 to Ig3, and thereby stabilize the horseshoe structure. The horseshoe structure of L1CAM may reversibly open into its extended conformation, in which Ig1-Ig6 assume a linear, rod-like shape. B: Two proposed mechanisms of L1CAM homophilic binding in trans. Both theories are compatible with the idea that Ig1-Ig4 exist in a dynamic equilibrium between folded (horseshoe) and extended conformations (shown in the center of the panel). The first mechanism (pathway 1 on the left side of the panel), is suggested by the crystal structure of axonin-1 (58). In this model, horseshoe structures on opposing L1CAM do not open up, but rather interact with one another in trans. The interaction is mediated by non-covalent bonds between Ig3 of one protein with the horseshoe fold of the other L1CAM molecule (represented by a red dashed line). Specifically, a peptide loop from Ig3 is thought to contact a donut hole-shaped binding pocket (red star) within the opposing horseshoe fold. The second proposed mechanism (shown as pathway 2 on the right side of the panel), is suggested by the crystal structure of hemolin (56). In this proposed mechanism, Ig1-Ig4 first converts into its extended conformation, and then binds to Ig1-Ig4 of an opposing L1CAM protein. Non-covalent interactions that stabilize the horseshoe within individual L1CAM molecules (dashed lines) also mediate the intermolecular interaction between two L1CAM proteins.

Figure 4. The L1CAM extracellular region utilizes two distinct strategies for protein-protein interactions Ligands that L1CAM interacts with via the "modular mode" are listed to the left of the cartoon. Modular mode interactions are defined as interactions mediated by single L1CAM domains that do not require contributions from neighboring domains to mediate binding. Extracellular domains of L1CAM that support modular mode binding are colored red. To the right of the cartoon is a list of ligands that interact with L1CAM via the "cooperative mode" (i.e. requiring the contribution of multiple L1CAM domains for binding). Cooperative mode binding requires contributions from Ig1-Ig4, which make up the horseshoe structure (highlighted in green).

Figure 5. A model for L1CAM molecular interactions and protein folding A: At the cell surface, Ig1-Ig4 adopt a horseshoe-shaped conformation (highlighted in green) that can convert reversibly into an extended conformation. In this model, the horseshoe structure represents the critical functional unit for cooperative mode interactions (such as trans homophilic binding). The activity of the horseshoe is regulated by a "regulator" region (highlighted in red) that abuts the horseshoe from either side. The regulator region includes sequences from Ig5, Ig6 and exon 2, and acts by stabilizing the horseshoe structure. When Ig1-Ig4 adopt the extended conformation, cooperative mode binding is disabled but L1CAM can still engage in modular mode binding (such as binding to neurocan or integrins). In this model, L1CAM molecules may be clustered more easily in the extended form (perhaps through the action of Fn3), and can then be internalized by clathrin-mediated endocytosis (79). B: Nascent L1CAM proteins are synthesized on ribosomes (green ovals) and co-translationally inserted into the ER membrane. In this model, interactions among domains early in the folding pathway lead to a horseshoe-shaped scaffold (center, highlighted in green), which precedes the establishment of mature tertiary structures in Ig1-Ig4. Interference with this process (for example by pathological mutations) leads to misfolded Ig domains. Misfolding results in aberrant disulfide-linked dimers formed by mutant molecules that do not form stable horseshoes (right). L1CAM extracellular domains also make individual contributions to folding, in that misfolding in any domain can lead to retention in the ER (50, 75).