News and Publications

Regulation, Function, and Signaling Mechanisms of Cell Adhesion Molecules

The fundamental processes of embryogenesis have evolved to translate the linear genetic code reproducibly a into 3-dimensional organism. Early in development, dividing cells derived from the fertilized egg adhere into specific aggregates. The gene program of each aggregate is then altered, moving the cells down differentiation pathways leading to specific tissues and organs. My colleagues and I have identified and characterized adhesion proteins on the surfaces of cells that allow aggregation, and we have shown that different tissues have different combinations of these cell adhesion molecules or CAMs. One of our objectives is to describe the signals that regulate the expression of gene programs when cells aggregate via specific CAMs.

In addition to forming tissues and organs, development must proceed in such a way that each collective or aggregate of cells is in its proper place in the overall body plan. Transcription factors such as the protein products of genes called Hox and Pax are among the regulators of this process. These factors act by turning sets of other genes on or off at critical places along the axis of the embryo. We showed that CAM genes are targets for proteins encoded by Hox and Pax, indicating that regulation of CAM gene expression may play a role in tissue morphogenesis in specific regions of the embryo. Our second major goal is to define the cellular processes and cellular components in addition to products of Hox and Pax that are responsible for changing the expression of CAM genes during developmental processes.

Genes are controlled through regions of the DNA called promoters, and transcription factors can bind to specific parts of the promoter DNA to activate or repress gene expression. In previous studies, we found that several proteins encoded by Hox and Pax alter the expression of a number of CAM genes. We discovered that in addition to being regulated by Hox and Pax proteins, 2 related CAMs restricted to the nervous system, Ng-CAM and L1, contain a 21-bp negative regulatory element known as the neural restrictive silencer element. This element and its binding factor, REST/NRSF, are responsible for silencing the expression of these CAMs in nonneural tissues.

In our most recent studies, we identified an important positive regulatory element, the HPD, within the gene for L1 that binds to both homeobox and Pax proteins. The HPD carries out a number of important regulatory functions for the gene. It is required for binding to and activation of the L1 promoter by the transcription factor Pax-6, is necessary for the proper pattern of L1 expression in the forebrain and midbrain, and is critical for the response of the gene to bone morphogenetic proteins. Bone morphogenetic proteins are signaling proteins secreted by cells during embryonic induction and are important regulators of morphogenesis in both neural and nonneural tissues. We are investigating the contexts by which these proteins regulate the expression of CAMs and are identifying the DNA elements and transcription factors that participate in these responses.

In previous studies, we showed that the neural CAM (N-CAM) inhibits astrocyte proliferation in vitro and in vivo after a lesion. Furthermore, N-CAM binding affected mitogen-activated protein kinase activity in astrocytes and altered gene expression in both astrocytes and neurons. These functions of N-CAM depended on activation of the glucocorticoid receptor, a transcription factor. These findings led us to consider effects of N-CAM on other transcription factors, such as AP-1 and NF-κB, that are also influenced by the glucocorticoid receptor.

Recently, we analyzed DNA binding of and transcriptional activation by NF-κB after N-CAM binding to the cell surface. Addition of purified N-CAM, the recombinant third immunoglobulin (Ig) domain of N-CAM, or N-CAM antibodies to neonatal rat forebrain astrocytes or to cerebellar granule neurons increased NF-κB­DNA binding activity in nuclear extracts, as measured by electrophoretic mobility shift assays. NF-κB­mediated transcription was increased after N-CAM binding to astrocytes and neurons, as shown by the activation of 2 different luciferase reporter constructs containing NF-κB binding sites. N-CAM binding also resulted in degradation of IκB-α protein, which inhibits the translocation of NF-κB into the nucleus, and a subsequent increase in the levels of IκB-α mRNA and protein.

These results indicate that N-CAM homophilic binding at the cell membrane leads to activation of NF-κB binding to DNA and transcriptional activation in both neurons and astrocytes. Activation of NF-κB, however, did not influence the ability of N-CAM to inhibit astrocyte proliferation. These observations together support the hypothesis that N-CAM binding activates multiple pathways leading to changes in gene expression in astrocytes and neurons. Ongoing studies are aimed at elucidating intracellular signaling pathways and genes that are regulated by cell adhesion. Moreover, we are pursuing the hypothesis, based on our earlier work, that N-CAM synthesis is regulated in part by synaptic activity.

Despite extensive studies of CAMs, little is known about the mechanism of binding of these molecules or how this binding leads to intracellular signals that eventually lead to changes in gene expression. The goal of our studies is to describe the molecular events involved in these processes, with a focus on N-CAM. The mechanism of homophilic N-CAM binding is being investigated initially by looking for specific interactions between the extracellular Ig domains (Ig I­Ig V) of the protein. In collaborative studies with J. Dyson and P. Wright, The Scripps Research Institute, the distal Ig domain (Ig I), was uniformly labeled with carbon 13 and nitrogen 15, and the backbone and side chain assignments were determined on the basis of triple-resonance nuclear magnetic resonance experiments. Knowledge of the assignments of Ig I allowed us to probe potential interactions with other domains through chemical-shift perturbations induced by protein-protein binding. Using this method, we detected a weak but specific interaction between Ig I and Ig II, centered on a hydrophobic surface on Ig I. Under similar experimental conditions, Ig I did not bind to itself and did not interact with Ig III.

Further support for an interaction between the first and second Ig domains was obtained from biophysical studies in which the protein corresponding to these 2 domains was expressed as a single polypeptide chain. In analytical ultracentrifugation experiments performed by H. Lashuel and J. Kelly, the Skaggs Institute, the Ig I­Ig II concatemer dimerized with a significantly stronger association constant than that of the individual Ig I and Ig II domains, consistent with an antiparallel orientation of the molecules. These results suggest that an interaction can occur between these domains and provide new views on how an N-CAM molecule on one cell may interact with an N-CAM molecule on the same cell or on an apposing cell.

To assess the structural features of the cytoplasmic domain that are involved in N-CAM signaling events, we transfected cells with constructs that express the native and mutated forms of the cytoplasmic domain. Initial studies focused on the 4 cysteines immediately adjacent to the internal face of the plasma membrane that were assumed to be the sites where N-CAM is acylated with fatty acids. Site-directed mutagenesis verified that these are the sites of palmitoylation and indicated that the cysteines are differentially acylated. The results also indicated that the fatty acids help direct the protein to the membrane and help anchor it there. The cytoplasmic domain without the transmembrane region was directed to the membrane fraction but only if at least 2 of the cysteines were present.

These results suggest that fatty acid acylation influences the association of N-CAM with the membrane and may be important in localization of the protein or in organizing the structure of the cytoplasmic segment. These studies are being extended to show that the cytoplasmic domain can have a dominant effect on N-CAM signaling and to assess the role of the various phosphorylation sites on N-CAM function. For the latter studies, the cytoplasmic domain is expressed with the membrane-spanning region to ensure proper orientation in the cell and to maximize expression.

Cell differentiation during development can be influenced by controlling translation of mRNA and by gene transcription. In earlier studies, we observed that large numbers of eukaryotic mRNAs contain sequences that resemble segments contained within the 18S or 28S rRNAs. We suggested that these rRNA-like sequences would interact with ribosomes. The sequences in the sense orientation might mimic rRNA and bind to ribosomal proteins, whereas those in the antisense or complementary orientation might form base pairs with rRNA. Because ribosomes are the cellular structures that translate mRNAs into protein, we proposed that these interactions would directly affect translation.

In more recent studies, photochemical cross-linking was used to show that within intact ribosomes various segments of the 18S rRNA are accessible and can form base pairs with complementary mRNA sequences. In addition, cell-free translation and transfection studies indicated that these interactions directly affect the translation of the mRNAs encoding ribosomal protein S15 and the homeodomain protein Gtx. These mRNAs contain segments complementary to different regions of the 18S rRNA. Samples of mRNAs prepared from S15-luciferase or Gtx-luciferase fusion constructs with various degrees of complementarity to the 18S rRNA were translated in cell-free lysates. For both mRNAs, we found a strong correlation between the degree of complementarity to the 18S rRNA and inhibition of luciferase activity. As the degree of complementarity to the 18S rRNA increased, translation of the S15-luciferase mRNA decreased 2-fold; translation of the Gtx-luciferase mRNA decreased 9-fold.

In addition to examining complementary sequence matches, we are determining whether rRNA-like sequences in the sense orientation allow some mRNAs to bind ribosomal proteins and affect translation. We are generating reagents for this analysis, including ribosomal proteins expressed as recombinant proteins and antibodies to individual ribosomal proteins.

Our studies provide some of the first examples indicating that the translation of some eukaryotic mRNAs is directly controlled by base pairing to the 18S rRNA. One of our goals is to determine the actual extent to which base pairing to rRNA affects the overall pattern of gene expression.

Publications

Atkins, A.R., Osborne, M.J., Lashuel, H.A., Edelman, G.M., Wright, P.E., Cunningham, B.A., Dyson, H.J. Association between the first two immunoglobulin-like domains of the neural cell adhesion molecule N-CAM. FEBS Lett. 451:162, 1999.

Hu, M.C.-Y., Tranque, P., Edelman, G.M., Mauro, V.P. rRNA-complementarity in the 5´ untranslated region of mRNA specifying the Gtx homeodomain protein: Evidence that base-pairing to 18S rRNA affects translational efficiency. Proc. Natl. Acad. Sci. U. S. A. 96:1339, 1999.

Krushel, L.A., Cunningham, B.A., Edelman, G.M., Crossin, K.L. NF-κB activity is induced by neural cell adhesion molecule binding to neurons and astrocytes. J. Biol. Chem. 274:2432, 1999.

Little, E.B., Edelman, G.M., Cunningham, B.A. Palmitoylation of the cytoplasmic domain of the neural cell adhesion molecule N-CAM serves as an anchor to cellular membranes. Cell Adhes. Commun. 6:415, 1998.

Meech, R., Kallunki, P., Edelman, G.M., Jones, F.S. A binding site for homeodomain and Pax proteins is necessary for L1 cell adhesion molecule gene expression by Pax-6 and bone morphogenetic proteins. Proc. Natl. Acad. Sci. U. S. A. 96:2420, 1999.

Tranque, P., Hu, M.C.-Y., Edelman, G.M., Mauro, V.P. rRNA complementarity within mRNAs: A possible basis for mRNA-ribosome interactions and translational control. Proc. Natl. Acad. Sci. U. S. A. 95:12238, 1998.