Kinetic analysis of L1 homophilic interaction: role of the first four immunoglobulin domains and implications on binding mechanism

L1 is a cell adhesion molecule of the immunoglobulin (Ig) superfamily, critical for central nervous system development, and involved in several neuronal biological events. It is a type I membrane glycoprotein. The L1 ectodomain, composed of six Ig-like and five fibronectin (Fn) type-III domains, is involved in homophilic binding. Here, co-immunoprecipitation studies between recombinant truncated forms of human L1 expressed and purified from insect Spodoptera frugiperda Sf9 cells, and endogenous full-length L1 from human NT2N neurons, showed that the L1 ectodomain (L1/ECD) and L1/Ig1-4 interacted homophilically in trans, contrary to mutants L1/Ig1-3 and L1/Ig2-Fn5. All mutants were correctly folded as evaluated by combination of far-UV CD and fluorescence spectroscopy. Surface plasmon resonance analysis showed comparable dissociation constants of 116 +/- 2 and 130 +/- 6 nm for L1/ECD-L1/ECD and L1/ECD-L1/Ig1-4, respectively, whereas deletion mutants for Ig1 or Ig4 did not interact. Accordingly, in vivo, Sf9 cells stably expressing L1 were found to adhere only to L1/ECD- and L1/Ig1-4-coated surfaces. Furthermore, only these mutants bound to HEK293 cells overexpressing L1 at the cell surface. Enhancement of neurite outgrowth, which is the consequence of signaling events caused by L1 homophilic binding, was comparable between L1/ECD and L1/Ig1-4. Altogether, these results showed that domains Ig1 to Ig4 are necessary and sufficient for L1 homophilic binding in trans, and that the rest of the molecule does not contribute to the affinity under the conditions of the current study. Furthermore, they are compatible with a cooperative interaction between modules Ig1-Ig4 in a horseshoe conformation.     

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.     

Properties of Manduca sexta chitinase and its C-terminal deletions

Manduca sexta (tobacco hornworm) chitinase is a molting enzyme that contains several domains including a catalytic domain, a serine/threonine-rich region, and a C-terminal cysteine-rich domain. Previously we showed that this chitinase acts as a biopesticide in transgenic plants where it disrupts gut physiology. To delineate the role of these domains further and to identify and characterize some of the multiple forms produced in molting fluid and in transgenic plants, three different forms with variable lengths of C-terminal deletions were generated. Appropriately truncated forms of the M. sexta chitinase cDNA were generated, introduced into a baculovirus vector, and expressed in insect cells. Two of the truncated chitinases (Chi 1-407 and Chi 1-477) were secreted into the medium, whereas the one with the longest deletion (Chi 1-376) was retained inside the insect cells. The two larger truncated chitinases and the full-length enzyme (Chi 1-535) were purified and their properties were compared. Differences in carbohydrate compositions, pH–activity profiles, and kinetic constants were observed among the different forms of chitinases. All three of these chitinases had some affinity for chitin, and they also exhibited differences in their ability to hydrolyze colloidal chitin. The results support the hypothesis that multiple forms of this enzyme occur in vivo due to proteolytic processing at the C-terminal end and differential glycosylation.     

Integrin and neurocan binding to L1 involves distinct Ig domains.

The cell adhesion molecule L1, a 200-220-kDa type I membrane glycoprotein of the Ig superfamily, mediates many neuronal processes. Originally studied in the nervous system, L1 is expressed by hematopoietic and many epithelial cells, suggesting a more expanded role. L1 supports homophilic L1-L1 and integrin-mediated cell binding and can also bind with high affinity to the neural proteoglycan neurocan; however, the binding site is unknown. We have dissected the L1 molecule and investigated the cell binding ability of Ig domains 1 and 6. We report that RGD sites in domain 6 support alpha5beta1- or alphavbeta3-mediated integrin binding and that both RGD sites are essential. Cooperation of RGD sites with neighboring domains are necessary for alpha(5)beta(1). A T cell hybridoma and activated T cells could bind to L1 in the absence of RGDs. This binding was supported by Ig domain 1 and mediated by cell surface-exposed neurocan. Lymphoid and brain-derived neurocan were structurally similar. We also present evidence that a fusion protein of the Ig 1-like domain of L1 can bind to recombinant neurocan. Our results support the notion that L1 provides distinct cell binding sites that may serve in cell-cell or cell-matrix 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.     

Properties of Manduca sexta chitinase and its C-terminal deletions

Manduca sexta (tobacco hornworm) chitinase is a molting enzyme that contains several domains including a catalytic domain, a serine/threonine-rich region, and a C-terminal cysteine-rich domain. Previously we showed that this chitinase acts as a biopesticide in transgenic plants where it disrupts gut physiology. To delineate the role of these domains further and to identify and characterize some of the multiple forms produced in molting fluid and in transgenic plants, three different forms with variable lengths of C-terminal deletions were generated. Appropriately truncated forms of the M. sexta chitinase cDNA were generated, introduced into a baculovirus vector, and expressed in insect cells. Two of the truncated chitinases (Chi 1-407 and Chi 1-477) were secreted into the medium, whereas the one with the longest deletion (Chi 1-376) was retained inside the insect cells. The two larger truncated chitinases and the full-length enzyme (Chi 1-535) were purified and their properties were compared. Differences in carbohydrate compositions, pH–activity profiles, and kinetic constants were observed among the different forms of chitinases. All three of these chitinases had some affinity for chitin, and they also exhibited differences in their ability to hydrolyze colloidal chitin. The results support the hypothesis that multiple forms of this enzyme occur in vivo due to proteolytic processing at the C-terminal end and differential glycosylation.     

High and low molecular weight rabbit secretory components. Evidence for the deletion of the second and third domains in the smaller polypeptide

Rabbit secretory components exist in two forms which differ in apparent mass by about 25 kDa. Each of these two forms were reduced, carboxymethylated, and extensively digested with trypsin. The resulting peptides were purified by reverse-phase high performance liquid chromatography and characterized by NH2- and COOH-terminal sequence determination and/or amino acid analysis. They were aligned with the protein sequence predicted from the cDNA nucleotide sequence encoding the rabbit poly(Ig) receptor (Mostov, K. E., Friedlander, M., and Blobel, G. (1984) Nature 308, 37-43). All peptides belonging to the fourth and fifth domains except one (positions 488-496) were accounted for in both forms. In addition, limited tryptic proteolysis of the native low Mr secretory components produced the intact 18-kDa NH2-terminal domain (positions 1-117) and the 30-kDa fragment encompassing the fourth and fifth domains. These results suggest that the smaller polypeptide derives from the larger secretory component form by the deletion of the second and third domains.     

Novel insect cell line capable of complex N-glycosylation and sialylation of recombinant proteins

Paucimannose or oligomannose structures are usually attached to glycoproteins produced by insect cells, while mammalian glycoproteins usually have complex glycans. The lack of complex glycosylation has limited the use of the insect cell baculovirus expression vector system (BEVS), despite its high productivity and versatility. The availability of cell lines capable of complex glycosylation can overcome such a problem and potentially increase the utility of BEVS. In this work the capability of two novel cell lines, one from Pseudaletia unipuncta (A7S) and one from Danaus plexippus (DpN1), to produce and glycosylate a recombinant protein (secreted human placental alkaline phosphatase, SeAP) was assessed. SeAP produced by Tn5B1-4 cells at a low passage number (<200) was utilized for comparison. The optimal conditions for the production of SeAP by DpN1 cells were defined, and the glycosylation profiles of SeAP produced by the cell lines were quantitatively determined. Both the A7S and the DpN1 cells produced lower concentrations of SeAP than the Tn5B1-4 cells. Less than 5% of the glycans attached to SeAP produced by the Tn5B1-4 cells had complex forms. Glycans attached to SeAP from A7S cells contained 4% hybrid and 8% complex forms. Galactosylated biantennary structures were identified. Glycans attached to SeAP produced by the DpN1 cell line had 6% hybrid and 26% complex forms. Of the complex forms in SeAP from DpN1, 13% were identified as sialylated glycans. The galactosyltransferase activity of the three cell lines was measured and correlated to their ability to produce complex forms. Even though neither novel cell line produced as much recombinant protein as the Tn5B1-4 cells, the glycosylation of SeAP expressed by both cell lines was more complete. These novel cell lines represent interesting alternatives for the production of complex glycosylated proteins utilizing the BEVS.     

Cloning of an alternate form of vascular cell adhesion molecule-1 (VCAM1).

Vascular cell adhesion molecule-1 (VCAM1) of the Ig superfamily is induced by the inflammatory cytokines interleukin-1 and tumor necrosis factor on human umbilical vein endothelial cells (HUVECs). It binds to mononuclear leukocytes via the integrin VLA-4. We have cloned and expressed a cDNA encoding a new form of human VCAM1 containing an additional Ig homologous domain inserted between the third and fourth domains of the original six-domain protein. Characterization of mRNA from HUVECs from three individuals at various time points after induction by tumor necrosis factor indicates that both the long and short VCAM1 mRNAs are made by all three individuals, with the long form predominating quantitatively. Immunoprecipitation of VCAM1 protein from cos7 cells transfected with each cDNA and from cultured endothelial cells followed by deglycosylation suggests that the long form is the major form found on endothelium. The two forms may result from alternate splicing of a precursor mRNA. Both forms support adhesion of VLA-4-expressing cell lines.     

The LG1-3 tandem of laminin alpha5 harbors the binding sites of Lutheran/basal cell adhesion molecule and alpha3beta1/alpha6beta1 integrins

The laminin-type globular (LG) domains of laminin alpha chains have been implicated in various cellular interactions that are mediated through receptors such as integrins, alpha-dystroglycan, syndecans, and the Lutheran blood group glycoprotein (Lu). Lu, an Ig superfamily transmembrane receptor specific for laminin alpha5, is also known as basal cell adhesion molecule (B-CAM). Although Lu/B-CAM binds to the LG domain of laminin alpha5, the binding site has not been precisely defined. To better delineate this binding site, we produced a series of recombinant laminin trimers containing modified alpha chains, such that all or part of alpha5LG was replaced with analogous segments of human laminin alpha1LG. In solid phase binding assays using a soluble Lu (Lu-Fc) composed of the Lu extracellular domain and human IgG1 Fc, we found that Lu bound to Mr5G3, a recombinant laminin containing alpha5 domains LN through LG3 fused to human laminin alpha1LG4-5. However, Lu/B-CAM did not bind other recombinant laminins containing alpha5LG3 unless alpha5LG1-2 was also present. A recombinant alpha5LG1-3 tandem lacking the laminin coiled coil (LCC) domain did not reproduce the activity of Lu/B-CAM binding. Therefore, proper structure of the alpha5LG1-3 tandem with the LCC domain was essential for the binding of Lu/B-CAM to laminin alpha5. Our results also suggest that the binding site for Lu/B-CAM on laminin alpha5 may overlap with that of integrins alpha3beta1 and alpha6beta1.     

Antibody Fragments Directed against Different Portions of the Human Neural Cell Adhesion Molecule L1 Act as Inhibitors or Activators of L1 Function

The neural cell adhesion molecule L1 plays important roles in neuronal migration and survival, neuritogenesis and synaptogenesis. L1 has also been found in tumors of different origins, with levels of L1 expression correlating positively with the metastatic potential of tumors. To select antibodies targeting the varied functions of L1, we screened the Tomlinson library of recombinant human antibody fragments to identify antibodies binding to recombinant human L1 protein comprising the entire extracellular domain of human L1. We obtained four L1 binding single-chain variable fragment antibodies (scFvs), named I4, I6, I13, and I27 and showed by enzyme-linked immunosorbent assay (ELISA) that scFvs I4 and I6 have high affinity to the immunoglobulin-like (Ig) domains 1–4 of L1, while scFvs I13 and I27 bind strongly to the fibronectin type III homologous (Fn) domains 1–3 of L1. Application of scFvs I4 and I6 to human SK-N-SH neuroblastoma cells reduced proliferation and transmigration of these cells. Treatment of SK-N-SH cells with scFvs I13 and I27 enhanced cell proliferation and migration, neurite outgrowth, and protected against the toxic effects of H₂O₂ by increasing the ratio of Bcl-2/Bax. In addition, scFvs I4 and I6 inhibited and scFvs I13 and I27 promoted phosphorylation of src and Erk. Our findings indicate that scFvs reacting with the immunoglobulin-like domains 1–4 inhibit L1 functions, whereas scFvs interacting with the fibronectin type III domains 1–3 trigger L1 functions of cultured neuroblastoma cells.     

Improvement of glycosylation in insect cells with mammalian glycosyltransferases

The N-glycans of recombinant glycoproteins expressed in insect cells mainly contain high mannose or tri-mannose structures, which are truncated forms of the sialylated N-glycans found in mammalian cells. Because asialylated glycoproteins have a shorter half-life in blood circulation, we investigated if sialylated therapeutic glycoprotein can be produced from insect cells by enhancing the N-glycosylation machinery of the cells. We co-expressed in two insect cell lines, Sf9 and Ea4, the human alpha1-antitrypsin (halpha1AT) protein with a series of key glycosyltransferases, including GlcNAc transferase II (GnT2), beta1,4-galactosyltransferase (beta14GT), and alpha2,6-sialyltransferase (alpha26ST) by a single recombinant baculovirus. We demonstrated that the enhancement of N-glycosylation is cell type-dependent and is more efficient in Ea4 than Sf9 cells. Glycan analysis indicated that sialylated halpha1AT proteins were produced in Ea4 insect cells expressing the above-mentioned exogenous glycosyltransferases. Therefore, our expression strategy may simplify the production of humanized therapeutic glycoproteins by improving the N-glycosylation pathway in specific insect cells, with an ensemble of exogenous glycosyltransferases in a single recombinant baculovirus.     

Overexpression and functional characterization of the extracellular domain of the human alpha1 glycine receptor

A novel truncated form (residues 1-214, with a randomized C-terminal tail) of the ligand-binding extracellular domain (ECD) of the human alpha1 glycine receptor (GlyR), with amino acids from the corresponding sequence of an acetylcholine binding protein (AChBP) substituted for two relatively hydrophobic membrane-proximal loops, was overexpressed using a baculovirus expression system. The mutant GlyR ECD, named GlyBP, was present in both soluble and membrane-associated fractions after cell lysis, though only the latter appeared to be in a native-like conformation capable of binding strychnine, a GlyR specific antagonist. The membrane-associated GlyBP was solubilized, and detergent/lipid/protein micelles were affinity purified. After detergent removal, GlyBP may be isolated in either aqueous or vesicular form. Binding assays and spectroscopic studies using circular dichroism and FRET are consistent with both forms adopting equivalent native-like conformations. Thus, GlyBP may be isolated as a soluble or membrane-associated assembly that serves as a structural and functional homologue of the ECD of GlyR.     

Sushi domains in the B subunit of factor XIII responsible for oligomer assembly

Factor XIII (FXIII) is a heterotetramer composed of two catalytic A subunits (FXIII-A) and two B subunits (FXIII-B). FXIII-B has 10 Sushi domains. To explore the structure-function relationship of FXIII-B, we looked for domains in FXIII-B responsible for its homodimer and heterotetramer assembly with FXIII-A. Full-length recombinant human FXIII-B (rFXIII-B) and truncated rFXIII-Bs with various numbers of Sushi domains (rFXIII-B x- y ) were expressed in a baculovirus expression system. rFXIII-B was indistinguishable from purified human plasma FXIII-B, in terms of the molecular weight (after being deglycosylated by glycosidases) and the ability to form complexes between the two subunits. rFXIII-B was in dimer form and produced a heterotetramer complex with FXIII-A. Gel-filtration and FXIII-A binding analysis of the various truncated forms of rFXIII-B x- y revealed that the first Sushi domain was responsible for the binding of FXIII-B to FXIII-A and that the fourth and ninth Sushi domains were involved in the FXIII-B homodimer assembly. rFXIII-B and rFXIII-B 1-9, which formed a heterotetramer complex with FXIII-A, protected FXIII-A from proteolytic digestion. These findings suggest that only full-length or nearly full-length FXIII-B is large enough to cover the exposed surface of FXIII-A. In conclusion, at least 3 out of the 10 Sushi domains of FXIII-B have the distinct function of forming a homodimer and a heterotetramer, which should be ascribed to the differences in their amino acid sequences. The present studies, however, do not exclude the possibility that additional Sushi domains may also support either or both functions.     

Proteolytic activation of alternative CCR1 ligands in inflammation

Although chemokines CCL3/MIP-1alpha and CCL5/RANTES are considered to be primary CCR1 ligands in inflammatory responses, alternative CCR1 ligands have also been described. Indeed, four such chemokines, CCL6/C10/MIP-related protein-1, CCL9/MIP-1gamma/MIP-related protein-2, CCL15/MIP-1delta/hemofiltrate CC chemokine-2/leukotactin-1, and CCL23/CKbeta8/myeloid progenitor inhibitory factor-1, are unique in possessing a separately encoded N-terminal domain of 16-20 residues and two additional precisely positioned cysteines that form a third disulfide bridge. In vitro, these four chemokines are weak CCR1 agonists, but potency can be increased up to 1000-fold by engineered or expression-associated N-terminal truncations. We examined the ability of proinflammatory proteases, human cell supernatants, or physiological fluids to perform N-terminal truncations of these chemokines and thereby activate their functions. Remarkably, most of the proteases and fluids removed the N-terminal domains from all four chemokines, but were relatively unable to cleave the truncated forms further. The truncated chemokines exhibited up to 1000-fold increases in CCR1-mediated signaling and chemotaxis assays in vitro. In addition, N-terminally truncated CCL15/MIP-1delta and CCL23/CKbeta8, but not CCL3/MIP-1alpha or CCL5/RANTES, were detected at relatively high levels in synovial fluids from rheumatoid arthritis patients. These data suggest that alternative CCR1 ligands are converted into potent chemoattractants by proteases released during inflammatory responses in vivo.     

Identification of a Homophilic Binding Site in Immunoglobulin-Like Domain 2 of the Cell Adhesion Molecule L1

Abstract: The cell adhesion molecule L1 plays an important role in neural development, and mutations in human L1 have been implicated in X-linked hydrocephalus and related neurological diseases. We have previously demonstrated that recombinant proteins containing the second immunoglobulin-like domain (Ig2) of L1 contain both homophilic binding and neuritogenic activities. In this report, the involvement of L1 Ig2 in cell-cell adhesion and neuritogenesis was further evaluated in cell transfection studies. Transfectants expressing intact L1 were capable of undergoing L1-dependent self-aggregation and promoting neurite outgrowth from neural retinal cells. However, both activities were abolished in transfectants expressing L1Δ2, a mutant L1 with Ig2 deleted. In competition experiments, the wild-type Ig2 fusion protein inhibited L1-dependent cell aggregation, whereas an Ig2 fusion protein containing the hydrocephalus mutation R184Q did not. Oligopeptides flanking Arg¹⁸⁴ were therefore synthesized and assayed for their effects on L1-mediated cell-cell binding and neuritogenesis. The peptide L1-A, spanning the residues His¹⁷⁸ and Gly¹⁹¹, inhibited both L1- and Ig2 fusion protein-mediated homophilic binding. When neural retinal cells were cultured on substrate-coated Ig2 fusion protein, peptide L1-A also abolished L1-dependent neurite outgrowth. Substitutions of several charged residues and hydrophobic residues with alanine in peptide analogues led to the loss of inhibitory effects, suggesting that multiple amino acids might be involved in L1-L1 binding. Taken together, these results identify an L1 homophilic binding site within the sequence HIKQDERVTMGQNG of Ig2 and demonstrate the requirement of L1 homophilic binding in the promotion of neurite outgrowth.     

Analysis of human monoclonal antibodies derived from lymphocytes of patients with cancer

Human lymphocytes from tumor-bearing patients and normal individuals have been fused with the NS-1 mouse myeloma line or the LICR -LON- HMY2 ( LICR -2) or SK0 -007 human cell lines. For a given number of lymphocytes, fusions with NS-1 produced 8 times more clones than fusions with LICR -2 and greater than 20 times more clones than fusions with SK0 -007. The percentage of clones that secrete human immunoglobulin (Ig) and the range of Ig production were comparable for clones derived from the three myeloma/lymphoblastoid lines. Clones derived from fusions with LICR -2 and SK0 -007 were found to secrete new species of light and heavy Ig chains in addition to those of the myeloma/lymphoblastoid lines, and clones derived from fusions with NS-1 secreted human Ig and contained both mouse and human chromosomes, which indicates that true hybrid cells were derived from fusions with each of the myeloma/lymphoblastoid lines under study. The stability of Ig production was similar for clones derived from fusions with NS-1, LICR -2, or SK0 -007; these results were comparable to those obtained with standard mouse/mouse hybrids. Stable clones producing human monoclonal antibodies that react with cell surface, cytoplasmic, cytoskeletal, nuclear, or nucleolar antigens have been isolated from tumor-bearing patients and normal individuals. A number of human monoclonal antibodies reactive with cytoskeletal antigens appear to be directed against components of the intermediate filament family. Techniques for the production of human monoclonal antibody appear to be sufficiently advanced to initiate a serological dissection of the humoral immune response to cancer.