Proc. Nati. Acad. Sci. USA Vol. 84, pp. 2808-2812, May 1987 Developmental Biology Sequence analysis of a cDNA clone encoding the liver cell adhesion molecule, L-CAM (intrinsic membrane protein/epithelial glycoprotein/calcium-dependent adhesion/gene duplication) WARREN J. GALLIN, BARBARA C. SORKIN, GERALD M. EDELMAN, AND BRUCE A. CUNNINGHAM Department of Developmental and Molecular Biology, The Rockefeller University, 1230 York Avenue, New York, NY 10021 Contributed by Gerald M. Edelman, January 5, 1987 ABSTRACT The liver cell adhesion molecule (L-CAM) the formation of tight junctions in MDCK cells react with appears on non-neural epithelial tissues and mediates calcium- L-CAM rather than components of the tight junctions them- dependent adhesion in these tissues both in the embryo and in selves, indicating that the formation of some junctional the adult. It appears on cell surfaces as a glycoprotein of Mr complexes is dependent on cell-cell adhesion via L-CAM 124,000 but is synthesized as a precursor of Mr 135,000. We (12). Antibodies to L-CAM have recently been shown to have isolated and determined the nucleic acid sequence of a perturb the inductive interactions between epidermis and cDNA clone (XL320) encoding chicken L-CAM. The 5' end of dermis and to alter pattern formation during the development this clone has an open reading frame extending for 2520 base of feathers (13). Inasmuch as L-CAM is expressed in epider- pairs, followed by an 850-base-pair untranslated region ter- mis but not in dermis, this effect is apparently due to minating with a polyadenylylation site at its 3' end. Protein disruption of cooperative cellular interactions within the sequence analysis of intact L-CAM and of cyanogen bromide epidermal cell collectives leading to consequent alterations of fragments of the protein confirmed the reading frame and signals between these cells and mesodermal cells. In addi- indicated that XL320 encodes the complete sequence ofL-CAM tion, L-CAM and N-CAM (neural cell adhesion molecule) are as it is expressed on the cell surface as well as the bulk of the often expressed in adjacent populations of early embryonic precursor. The sequence includes a hydrophobic segment of 31 cells; their coordinate expression throughout development, amino acids, supporting our earlier conclusion that L-CAM is particularly at sites of induction, suggests that their mutual an intrinsic membrane protein. There are five potential aspar- activity is crucial for the formation oftissue boundaries (1, 2, agine glycosylation sites on the extracellular part of the 10, 14). molecule and an intracellular domain that is phosphorylated in Because of the widespread distribution of L-CAM and its vivo. The mature L-CAM polypeptide consists of 727 amino apparent role in promoting cooperative interactions between acids, with a calculated Mr of 79,900 for the carbohydrate-free cells in collectives, particularly during embryogenesis, we protein. The L-CAM sequence is not homologous to other have undertaken a detailed analysis of this molecule and its known protein sequences, including those of the neural cell genes. These studies provide a structural basis for analyzing adhesion molecule (N-CAM) and other members of the immu- L-CAM function and the control of its expression during noglobulin superfamily, but the L-CAM molecule does contain development and should provide probes for analyzing the three contiguous segments (113 amino acids each) that are role of L-CAM in cellular mechanisms of development. We homologous to each other. The similarities among these seg- have previously described the overall structure of the L- ments suggest that at least part of the L-CAM molecule arose CAM molecule (9) and used a small cDNA probe, pEC301 by gene duplication. (15), and specific antibodies to show that the mRNA [4 kilobases (kb)] and the protein are the same size in all tissues The liver cell adhesion molecule (L-CAM) is a primary CAM expressing L-CAM. We report here the sequence of a cDNA that appears in a distinct pattern at a variety of inductive clone that encompasses almost the entire L-CAM mRNA, embryonic sites as well as in adult tissues (1, 2). It was including the complete amino acid sequence of the protein as initially isolated on the basis of its ability to mediate calcium- it is detected on the cell surface. dependent adhesion between cells of the chicken liver epi- thelium (3, 4). Several other calcium-dependent adhesion MATERIALS AND METHODS molecules, including uvomorulin (5), E-cadherin (6), cell- CAM 120/80 (7), and Arc-1 (8) have been isolated from Isolation of a L-CAM cDNA Clone. cDNA libraries in Xgtll different tissues or cell and these molecules were prepared as described (15). Bacteriophage were plated epithelial lines, at a density of 105 plaque-forming units per plate, and replicas have biochemical properties (9) and tissue distributions (2, of each plate on nitrocellulose were probed with pEC301 10) so similar to those of L-CAM that it is likely that all of insert. A positive bacteriophage, designated XL320, was them are its mammalian homologues. cloned to purity. DNA from this bacteriophage was digested The importance of L-CAM in embryonic development and with EcoRI and ligated with EcoRI-digested calf intestinal in the formation of epithelium is suggested by studies in alkaline phosphatase-treated pBR328 (15). A plasmid con- which L-CAM-mediated adhesion was blocked with antibod- taining an insert that hybridized to pEC301 insert was ies. The formation of epithelial colonies in primary cultures isolated and designated pEC320. of embryonic chicken hepatocytes is reversibly blocked by DNA Sequencing. For M13 sequencing, insert DNA from antibodies to L-CAM (3, 4). In addition, antibodies to pEC320 was digested with restriction enzymes and subcloned uvomorulin block the compaction of mouse embryos and into M13mpl8 or M13mpl9 (16). For deletion subcloning, the hence the polarization of blastomeres that is essential to insert from pEC320 was cloned into the EcoRI site of normal development (11). Similarly, antibodies that prevent Bluescript-KS vectors (17), and deletion subclones were prepared by digestion with exonuclease III and mung bean The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: L-CAM, liver cell adhesion molecule; N-CAM, in accordance with 18 U.S.C. §1734 solely to indicate this fact. neural cell adhesion molecule. Downloaded by guest on October 2, 2021 2808 Developmental Biology: Gallin et al. Proc. Natl. Acad. Sci. USA 84 (1987) 2809 nuclease and blunt-end ligation to recircularize the plasmids Analysis of the hydrophobicity of the amino acid sequence (17). Deleted DNA was transformed into competent JM109 (27) (Fig. 2) indicated a single membrane-spanning region cells and plasmid was prepared from single colonies for consisting of 31 consecutive hydrophobic amino acid resi- sequencing. DNA sequence analysis was performed by the dues flanked on the amino-terminal side by a Gly-Arg-Arg- dideoxynucleotide chain-termination method. Fragments sub- Ser-Tyr sequence, and on the carboxyl-terminal side by an cloned into M13 vectors were sequenced using the Klenow Arg-Arg-Arg-Lys sequence. The plot also shows a smaller fragment of DNA polymerase I (18, 19) and fragments peak at residues 303-335, but this region has hydrophilic subcloned into Bluescript vectors were sequenced using residues dispersed throughout. either avian myeloblastosis virus reverse transcriptase or Digestion of liver membranes with trypsin in the presence Klenow fragment (20). Sequence data were compiled using of Ca2+ ions releases an extracellular Mr 81,000 glycoprotein the Staden ANALYSEQ programs (21). fragment (Ftl) of L-CAM (4, 9). This fragment contains four Peptide Purification and Sequencing. The Mr 81,000 frag- asparagine-linked oligosaccharides (9). Consistent with these ment (Ftl) of L-CAM was isolated (9) and treated with CNBr results, the extracellular portion of the sequence shown in [10% (wt/vol) in 70% formic acid] for 4 hr at room temper- Fig. 1 contains five potential asparagine glycosylation se- ature. Fragments were separated by gel filtration on Sepha- quences (Asn-X-Thr). cryl S-300 (Pharmacia) in 0.1 M NH4HCO3, followed by The relative molecular weights ofFtl and intact L-CAM as preparative NaDodSO4/PAGE on 15% acrylamide gels. Pep- measured by NaDodSO4/PAGE in the Laemmli system (28) tides were detected by Coomassie blue staining, electroe- are significantly larger than the molecular weights predicted luted (22), and subjected to Edman degradation (23) in the from the sequence of L-CAM (Mr 70,000 vs. 59,000 and Mr Rockefeller University sequencing facility. 116,000 vs. 79,000, respectively). These discrepancies appear RNA Transfer Blotting. Liver or brain total RNA (17 pug) to be due in part to anomalous migration of the glycoprotein was resolved by electrophoresis on formaldehyde gels (24), and Ftl on NaDodSO4/PAGE. For transferred to nitrocellulose, and probed with 32PO4-labeled example, although intact DNA probes (107 cpm in 10 ml of buffer) (15). After glycosylated L-CAM migrates with a M, of 124,000 on a hybridization overnight at 420C, blots were washed (15) and NaDodSO4/PAGE system in Tris buffers, it has a Mr of exposed to XAR-5 film at -70'C, with enhancing screens for 105,000 in a phosphate buffer (29), which is much closer to the autoradiography. value predicted from the cDNA sequence. Thrombospondin has also been reported to migrate with a lower apparent molecular weight in the phosphate system than in the RESULTS Laemmli system and the lower molecular weight is closer to the value predicted from the cDNA sequence (30). In addi- A Xgtll cDNA library was screened by nucleic acid hybrid- tion, a number of other proteins have been reported to have ization using the L-CAM-specific cDNA clone XL301 (15).