The Scripps Research Institute
  News Room Contacts  
  Information for Journalists  
  Calendar of Events  



News and Publications



Gerald M. Edelman, M.D., Ph.D. Member and Chairman
Bruce A. Cunningham, Ph.D. Member
Kathryn L. Crossin, Ph.D. Associate Member
Frederick S. Jones, Ph.D. Associate Member
Vincent P. Mauro, Ph.D. Assistant Member


Donald W. Copertino, Ph.D.
Brent D. Holst, Ph.D.
Vesa Pekka Kallunki, Ph.D.
Pedro A. Tranque, Ph.D.


Jeffrey J. Essner, Ph.D.
Douglas W. Ethell, Ph.D.
Chia-Yuan Hu, Ph.D.
Chrissa Kioussi, Ph.D.
Edward B. Little, Ph.D.
Robyn Meech, Ph.D.
Todd S. Ranheim, Ph.D.
Peter W. Vanderklish, Ph.D.
D. Wade Walke, Ph.D.


Greg R. Phillips


Leslie A. Krushel, Ph.D. The Neurosciences Institute, La Jolla, CA
George L.G. Miklos, Ph.D. The Neurosciences Institute, La Jolla, CA

Top of Page

Chairman's Overview

Gerald M. Edelman, M.D., Ph.D.

The major focus of the Department of Neurobiology is vertebrate development, in particular, development of the nervous system. Because the molecular processes leading to the emergence of animal form over time are just beginning to be described, no adequate theory of development exists comparable to current theories of evolution and genetics. The unfolding of modern methods of cell biology in the 1950s and 1960s and of molecular biology in the 1970s and 1980s has led to reductionist views of embryonic development that center on the cell and the gene as the functional units of development. In most inductive and morphogenetic processes in the embryo, however, the functional units are not single cells. Rather, they are collectives of interacting cells that give rise to tissues and organs. Can we reconcile a molecular analysis with the fact that form arises epigenetically from the collective interactions of an increasing number of embryonic cells during development? The key question is, How does the one-dimensional genetic code specify a three-dimensional animal of a given species?

Our department seeks to answer these questions at several levels by attempting to link genetic regulation to the mechanochemical processes that coordinate cell division, cell movement, and cell death. Recent studies suggest that one such link is provided by cell adhesion molecules (CAMs), which mediate cell-cell binding, and by substrate adhesion molecules (SAMs), which affect movement of cells and transformations of cell states. CAMs are involved in defining cell collectives and the borders of these groups of cells as the collectives interact during inductive events in morphogenesis. Networks of SAMs are involved in patterned cell migration. Although CAMs cannot be considered the "cause" of induction, they play major constraining roles in complex chains of inductive interactions that involve hormones and growth factors, components of the extracellular matrix, and cellular receptors.

CAMs function in neurite guidance and fasciculation, cell migration, and regeneration. Each CAM studied so far is specified by a single gene, although the genomic structures of different CAMs vary widely. The neural CAM (N-CAM), the first CAM to be characterized, is an example of a calcium-independent CAM. Detailed investigation of its structure supports the hypothesis that an N-CAM--like gene was the evolutionary precursor of a family of neurally important adhesion molecules and of the entire immunoglobulin superfamily.

Perturbation of CAM binding can lead to changes in morphology. Moreover, these changes (e.g., in nerve-muscle regeneration) lead to alterations in the expression of CAMs, suggesting that a series of readjustments in signaling events controls the expression of CAMs. During morphogenesis and regeneration, CAMs are coregulated with SAMs to affect migration and positioning. An example is cytotactin/tenascin, an extracellular matrix protein that supports counteradhesive events. This molecule may be one of a series of molecules controlled by neural signals that affect migration of cells and neurites into particular sites.

Expression of CAMs and SAMs is regulated by place-dependent dynamic signaling that is essential to normal neural morphogenesis. Understanding how such morphoregulatory molecules are regulated at the level of the gene is essential. Our studies indicate that transcription factors critical in development (Hox and Pax proteins, in particular) are involved in regulating the expression of CAMs. We have also obtained direct evidence that CAM-mediated adhesion itself can influence gene expression.

The key accomplishments of the past year are summarized in the reports that follow. Highlights include proof that mutations in the paired domain binding sites within the N-CAM promoter alter the neural pattern of expression of N-CAM in transgenic mice; findings that the tissue specificity of certain neural CAMs is determined at least in part by neural restrictive silencer elements; identification of a signal to the glial cell nucleus after N-CAM binding that involves the glucocorticoid receptor; the observation that rRNA-like sequences are present in mRNA transcripts and that these sequences may regulate gene expression at the level of translation; information on the in vivo distribution of the sialyltransferases responsible for adding -2,8-polysialic acid to N-CAM; discovery of a homeodomain protein called Barx2 that binds to a regulatory element common to the genes for neuron-glia CAM and L1; and finally, use of Cre recombinase integrated into adenovirus vectors to allow conditional expression of genes in mice that are engineered to contain flanking LoxP elements.

All these findings raise a number of questions that continue to be addressed by members of the department. What gene products or combinations of gene products control the expression of CAMs and SAMs in embryogenesis and differentiation? When cells adhere through particular adhesion molecules, what signals are sent back to the genome to change the pattern of differentiation? Which domains of adhesion molecules carry out the molecules' various biological functions? Attempts to answer these questions are at the center of the research currently being done by the members of our department.

Top of Page

Investigators' Reports

Structure and Function of Cell Adhesion Molecules

B.A. Cunningham, K.L. Crossin, G.R. Phillips, L.A. Krushel, T. Ranheim, P. Vanderklish, G.M. Edelman

Cell adhesion molecules (CAMs) are classified structurally into three families: the N-CAM family, the calcium-dependent CAMs or cadherins, and the selectins. Members of the N-CAM family all contain multiple immunoglobulin-like (Ig) domains, and many contain repeats resembling the type III repeats of fibronectin. A variety of data indicate that the various domains contribute differentially to homophilic binding, to other binding properties, and to the other activities of the various molecules. The intracellular regions of the transmembrane proteins are involved in intracellular signaling and in interactions with the cytoskeleton. We are examining the functions of all these domains in the neural CAM (N-CAM) and in a subfamily of the N-CAM family represented by neuron-glia CAM (Ng-CAM) and Ng-CAM--related molecule (Nr-CAM).

N-CAM mediates homophilic binding between cells, but the mechanism by which N-CAM on one cell interacts with N-CAM on another cell is unknown. To help elucidate this mechanism, we generated recombinant proteins corresponding to each of the five Ig domains (I--V) and to the combined fibronectin-like repeats 1 and 2. The purified proteins and antibodies were used in aggregation experiments with fluorescent microspheres and chicken embryo brain cells to determine the contributions of each domain to homophilic adhesion. The results indicated that the Ig III domain could self-aggregate, whereas the Ig I domain bound to the Ig V domain, and the Ig II domain bound to the Ig IV domain. The combined fibronectin-like repeats 1 and 2 showed a slight ability to self-aggregate but did not bind to any of the Ig domains. These findings suggest that N-CAM--N-CAM binding involves all five Ig domains and that these domains may interact pairwise in an antiparallel orientation.

To extend these results, we are examining the three-dimensional structures of the individual domains and the ability of one domain, or a peptide derived from it, to influence the structure of another domain. We have also generated recombinant proteins corresponding to combinations of domains (e.g., I--III, III--V, I--V) for additional chemical studies and to obtain three-dimensional structural data that might reveal more directly the binding mechanisms. Similar analyses are being carried out with recombinant proteins corresponding to specific domains in Nr-CAM.

In addition to its ability to mediate interactions between cells, N-CAM can support the outgrowth of neurites and inhibit the proliferation of astrocytes. We found that only the Ig II domain supports neurite attachment when coated as a substrate on plastic, consistent with a variety of other data suggesting that cell-cell adhesion and neurite outgrowth involve separate pathways. By contrast, tests of the recombinant domains for their ability to inhibit glial proliferation showed that all the domains were inhibitory and that their efficacy was similar to their ability to influence N-CAM binding. These results are consistent with the notion that inhibition of glial proliferation by N-CAM depends on homophilic binding.

N-CAM is unusual in that it contains large amounts of -2,8-polysialic acid linked to complex carbohydrates in the Ig V domain. This polysialylation reduces the efficacy of N-CAM--mediated homophilic binding and enhances neural migration and neurite outgrowth during development and regeneration. A high level of polysialic acid is expressed in the nervous system during embryonic and early postnatal development, and then levels decrease, except in regions undergoing neurogenesis or remodeling in adults. The cDNAs encoding two closely related rat sialyltransferases, PST-1 and STX, have recently been isolated. We have shown that both enzymes can sialylate N-CAM when transfected into the neuro-2A neuroblastoma cell line. The expression patterns of PST-1 and STX were examined by using in situ hybridization with RNA probes specific for unique regions in each enzyme.

Both enzymes were expressed abundantly throughout the nervous system during prenatal and early postnatal development and were coexpressed in most tissues examined. In the brains in adults, expression of PST-1 and STX was restricted to subsets of cells in discrete areas. This restricted expression was consistent with previous immunohistologic data on the distribution of polysialic acid. Moreover, expression of both enzymes was upregulated in areas of reactive gliosis induced by injections of kainic acid. These results suggest that polysialylation of N-CAM is regulated by at least two distinct sialyltransferases, both of which may contribute to modulation of the adhesive activity of N-CAM during development.

A key factor in the influence of CAMs on cells is the ability of CAMs as transmembrane proteins to mediate signaling events and to interact with the cytoskeleton. N-CAM signaling, for example, appears to involve pathways linked to the glucocorticoid receptor, and members of the Ng-CAM and Nr-CAM subfamily can bind to ankyrin. To help delineate additional signal pathways and interactions with the cytoskeleton, we have produced recombinant proteins corresponding to each of the two variant cytoplasmic domains of N-CAM and to the cytoplasmic domains of Nr-CAM and Ng-CAM. These proteins are being used as affinity reagents to detect molecules with which Nr-CAM and Ng-CAM interact and to define structure-function relationships. The cytoplasmic regions have also been produced in mammalian expression vectors to assess the effects of overexpressing the regions in cells that express the intact molecules.

The overall goal of these studies is to delineate in detail the interactions involved in the homophilic binding that leads to cell-cell binding and to describe the mechanisms by which such binding stimulates intracellular signaling or leads to other changes in cell properties.

Top of Page

Signals Resulting From the Expression and Interaction of Morphoregulatory Molecules

K.L. Crossin, V.P. Mauro, L.A. Krushel, E.B. Little, G.R. Phillips, C.-Y. Hu, P. Tranque, G.M. Edelman

Understanding development requires knowledge of how genes control the patterned organization of tissues in the embryo and of how gene networks operate in individual cells as a consequence of the feedback the networks receive from cellular interactions. Specific interactions of cells with other cells and with the extracellular matrix depend on cell adhesion molecules (CAMs) and on substrate adhesion molecules and their receptors. Our recent work has focused on understanding the changes in morphology, physiology, and gene expression that result when cells interact through CAMs or with molecules of the extracellular matrix. Cell-cell adhesion molecules and proteins in the extracellular matrix regulate cellular proliferation, movement, and differentiation and, in the nervous system, neurite guidance and extension.

To understand the coordination of these primary cellular processes by adhesion molecules, we have concentrated on the neural CAM (N-CAM). N-CAM contains immunoglobulin-like domains and fibronectin type III repeats in its extracellular domain, and it can be linked to cells via a phosphatidylinositol linkage or a transmembrane domain. On the basis of our observation that N-CAM can decrease the proliferation of astrocytes both in vitro and in vivo, we are examining how N-CAM binding might affect subsequent gene expression. Astrocytes express N-CAM on their surface and upregulate N-CAM levels after injury, suggesting that regulation of N-CAM expression is important in nerve regeneration.

To identify the gene targets involved in this process, we used subtractive hybridization to examine changes in gene expression that occur in astrocytes after N-CAM binding. We found that the levels of a number of mRNAs were changed. N-CAM mRNA levels were decreased after N-CAM binding, a result similar to our previous findings in neurons. This decrease suggests that a feedback mechanism exists after N-CAM interactions at the cell surface that leads to decreased expression of N-CAM. Two mRNAs were increased upon N-CAM binding. One of these encoded the glial marker enzyme glutamine synthetase, the expression of which is regulated by the glucocorticoid receptor. The other was calreticulin, a calcium-binding protein that can bind to the glucocorticoid receptor. These observations prompted an examination of the effects of glucocorticoid receptor agonists and antagonists on N-CAM--mediated changes in astrocyte proliferation and gene expression.

Activation of the glucocorticoid receptor by corticosteroids inhibited astrocyte proliferation in a dose-dependent manner. RU-486, an antagonist of the glucocorticoid receptor, prevented the inhibition of astrocyte proliferation by both N-CAM and steroids. Moreover, N-CAM treatment increased the expression of a luciferase reporter gene under the control of a minimal promoter linked to a glucocorticoid response element, and this activity was also abolished in the presence of RU-486. The combined data suggest that N-CAM can inhibit astrocyte proliferation and alter gene expression through alterations in the activity of the glucocorticoid receptor. This may involve a novel signaling mechanism through which N-CAM binding at the cell surface influences nuclear gene expression.

One of our major goals is to understand the influence of CAMs on subsequent morphogenesis and gene expression. In earlier studies, we found that in addition to mRNAs for several CAM and transcription factors, the mRNAs for ribosomal proteins were changed when cells aggregated in a number of cell systems. This finding led us to examine the possibility that changes in rRNAs or ribosomal proteins might be a mechanism for regulating gene expression at the level of translation during development. To begin to address this idea, we did extensive database analyses and found an astonishing correlation.

A number of eukaryotic mRNAs contain sequences that resemble segments of rRNAs. The rRNAs are a critical component of ribosomes, the cellular structures that translate mRNAs into proteins. The extensive homology found in many mRNAs led to the hypotheses that rRNA-like sequences may have spread throughout the eukaryotic genome during evolution and that the presence of these sequences in mRNA may differentially affect gene expression. We are exploring the possibility that the presence of these rRNA-like sequences function as cis regulatory elements. This ongoing analysis may reveal a heretofore undetected mechanism by which protein expression can be regulated at the level of translation and thereby affect development.

The results of these and other ongoing studies should provide a deeper understanding of the modulation of gene expression by CAMs and related molecules. These studies should complement investigations on the regulation of the synthesis of tenascin and CAMs by the products of homeotic genes.

Top of Page

Regulation of the Expression of the Genes for Adhesion Molecules

F.S. Jones, D.W. Copertino, B.D. Holst, P. Kallunki, C. Kioussi, R. Meech, D.W. Walke, G.M. Edelman

We are interested in understanding the mechanisms by which cell adhesion molecules (CAMs) are expressed in defined temporal and spatial patterns during neural development and regeneration. Of special interest are molecules of the N-CAM family, calcium-dependent CAMs or cadherins, and extracellular matrix glycoproteins. Cellular contact, sorting, movement, differentiation, and proliferation appear to depend on specific combinations of adhesion molecules expressed by cells at a given time and place. The orchestration of these activities leads to the creation of tissues with specific shapes at particular locations in the animal body. Our approach has been to identify cis regulatory sequences and associated transcription factors that control expression of the genes for different adhesion molecules, especially those molecules that appear in dynamic and restricted patterns during morphogenesis.

We have discovered that transcription factors encoded by homeobox-containing genes, including those of the Hox gene network and the paired box (Pax) gene family regulate expression of the genes for adhesion molecules. In studies of the neural CAM (N-CAM) promoter, we identified two DNA elements, the homeodomain binding site (HBS) and the paired domain binding site (PBS), that are targets for several proteins encoded by homeobox and Pax genes.

To determine how these elements influence patterns of N-CAM expression, we generated transgenic mice containing native and mutated N-CAM promoter constructs linked to a ß-galactosidase reporter gene. The native N-CAM promoter directed expression of ß-galactosidase in the CNS in a pattern consistent with that of the endogenous N-CAM gene. In contrast, mice with N-CAM promoter constructs with mutations in either the HBS or the PBS had altered patterns of expression in the spinal cord. At embryonic day 11, expression of ß-galactosidase was detected only in the ventrolateral region of the spinal cord, and at embryonic day 14.5, ß-galactosidase was no longer detected in any cells of the cord. These experiments showed that both the HBS and the PBS are important in determining specific expression patterns of N-CAM along the dorsoventral axis in the developing spinal cord and further support the conclusion that N-CAM is an in vivo target of homeobox and Pax gene products in vertebrates.

What factors are critical for restricting the expression of certain CAMs to the nervous system? To answer this question, we have focused on identifying regulatory elements that govern the neural expression of neuron-glia CAM (Ng-CAM) and L1, two closely related CAMs with immunoglobulin-like domains. These CAMs are restricted to the nervous system and appear predominantly in neurons and peripheral glia. We have identified two regulatory elements within the genes for L1 and Ng-CAM: a neural restrictive silencer element (NRSE) and a binding site for homeodomain proteins. The gene for Ng-CAM contains five NRSEs, whereas the gene for L1 contains only one; both genes have a nearly identical sequence motif that binds to homeodomain proteins. To determine how these elements control expression of these CAMs, we have analyzed the activities of the elements in DNA binding, cellular transfection, and transgenic mice experiments.

We found that both the NRSEs and the HBS were required for silencing Ng-CAM expression in nonneural cells. Analyses of the regulation of the gene for L1 showed that an 18-kb fragment of the gene containing the promoter, the first four exons, and three introns was silent in NIH3T3 cells but active in N2A cells. The silencing in NIH3T3 cells could be eliminated by deleting the NRSE from the L1 construct. In transgenic mice that had an 18-kb L1/ß-galactosidase reporter construct, expression of ß-galactosidase was restricted to neurons and glia of the CNS and the peripheral nervous system. However, a comparable construct lacking the NRSE showed ectopic expression in nonneural derivatives of the neural crest and in several mesodermal and ectodermal cell populations. Collectively, these experiments show that the tissue-specific expression of L1 is modulated by the NRSE.

In other studies, we used the homeodomain binding site common to Ng-CAM and L1 to search for factors that regulate the genes for both of these CAMs. This analysis led to the discovery of a new homeobox gene called Barx2. The homeodomain encoded by Barx2 is 87% identical to that of Barx1, and both genes are related to genes at the Bar locus of Drosophila melanogaster. We compared patterns of expression of Barx1 and Barx2 during mouse embryogenesis and showed that both genes are expressed in neural and craniofacial structures during development. Barx2 mRNA was expressed in the nervous system in regions that overlapped those in which L1 was expressed. Moreover, we found that Barx2 stimulated the activity of the L1 promoter in cellular cotransfection experiments. Thus, these studies suggest that Barx2 is a transcription factor that may control the expression of L1 and other target genes during neural development.

Taken together, our results suggest that CAMs are major downstream targets for Hox and Pax transcription factors and that the NRSE is critical for silencing the expression of neurally expressed CAMs in nonneural tissues. Our hypothesis is that NRSE and its associated transcription factor, NRSF/REST, work in combination with Hox and Pax proteins leading to precise spatiotemporal expression of members of the N-CAM family during embryogenesis.


Copertino, D.W., Edelman, G.M., Jones, F.S. Multiple promoter elements differentially regulate the expression of the mouse tenascin gene. Proc. Natl. Acad. Sci. U.S.A. 94:1846, 1997.

Crossin, K.L., Tai, M.-H., Krushel, L.A., Mauro, V.P., Edelman, G.M.

Glucocorticoid receptor pathways are involved in the inhibition of astrocyte proliferation. Proc. Natl. Acad. Sci. U.S.A. 94:2687, 1997.

Edelman, G.M. Gene knockout by viral delivery. Nature Biotechnol. 14:1537, 1996.

Edelman, G.M., Jones, F.S. Gene regulation of cell adhesion: A key step in neural morphogenesis. Brain Res., in press.

Edelman, G.M., Jones, F.S. Gene regulation of cell adhesion molecules in neural morphogenesis. Acta Paediatr. 86:12, 1997.

Holst, B.D., Wang, Y., Jones, F.S., Edelman, G.M. A binding site for Pax proteins regulates expression of the gene for the neural cell adhesion molecule in the embryonic spinal cord. Proc. Natl. Acad. Sci. U.S.A. 94:1465, 1997.

Jones, F.S., Kioussi, C., Copertino, D.W., Kallunki, P., Holst, B.H., Edelman, G.M. Barx2, a new homeobox gene of the Bar class is expressed in neural and craniofacial structures during development. Proc. Natl. Acad. Sci. U.S.A. 94:2632, 1997.

Kallunki, P., Edelman, G.M., Jones, F.S. Tissue-specific expression of the L1 cell adhesion molecule is modulated by the neural restrictive silencer element. J. Cell Biol., in press.

Mauro, V.P., Edelman, G.M. rRNA-like sequences occur in diverse primary transcripts: Implications for the control of gene expression. Proc. Natl. Acad. Sci. U.S.A. 94:422, 1997.

Phillips, G.P., Krushel, L.A., Crossin, K.L. Developmental expression of two rat sialyltransferases that modify the neural cell adhesion molecule, N-CAM. Dev. Brain Res., in press.

Wang, Y., Krushel, L.A., Edelman, G.M. Targeted DNA recombination in vivo using an adenovirus carrying the Cre recombinase gene. Proc. Natl. Acad. Sci. U.S.A. 93:3932, 1996.

Top of Page

Top of Page



Copyright © 2004 TSRI.