A binding site for homeodomain and Pax proteins is necessary for L1 cell adhesion molecule gene expression by Pax-6 and bone morphogenetic proteins.

The cell adhesion molecule L1 regulates axonal guidance and fasciculation during development. We previously identified the regulatory region of the L1 gene and showed that it was sufficient for establishing the neural pattern of L1 expression in transgenic mice. In the present study, we characterize a DNA element within this region called the HPD that contains binding motifs for both homeodomain and Pax proteins and responds to signals from bone morphogenetic proteins (BMPs). An ATTA sequence within the core of the HPD was required for binding to the homeodomain protein Barx2 while a separate paired domain recognition motif was necessary for binding to Pax-6. In cellular transfection experiments, L1-luciferase reporter constructs containing the HPD were activated an average of 4-fold by Pax-6 in N2A cells and 5-fold by BMP-2 and BMP-4 in Ng108 cells. Both of these responses were eliminated on deletion of the HPD from L1 constructs. In transgenic mice, deletion of the HPD from an L1-lacZ reporter resulted in a loss of beta-galactosidase expression in the telencephalon and mesencephalon. Collectively, our experiments indicate that the HPD regulates L1 expression in neural tissues via homeodomain and Pax proteins and is likely to be a target of BMP signaling during development.

The neural restrictive silencer element can act as both a repressor and enhancer of L1 cell adhesion molecule gene expression during postnatal development.

The cell adhesion molecule L1 mediates axonal guidance during neural development and mutations in its gene result in severe neurological defects. In previous studies, we identified the promoter for the L1 gene and showed that a neural restrictive silencer element (NRSE) was critical for preventing ectopic expression of L1 during early embryonic development. In the present study, we have investigated the role of the NRSE in the regulation of L1 expression during postnatal development. In gel mobility shift experiments, the NRSE formed DNA-protein complexes with nuclear extracts prepared from the brains of postnatal mice. To examine the influence of the NRSE on postnatal patterns of L1 expression in vivo, we compared the expression of two lacZ transgene constructs, one containing the native L1 gene regulatory sequences (L1lacZ) and another (L1lacZDeltaN) lacking the NRSE. Newborn mice carrying the L1lacZDeltaN showed enhanced beta-galactosidase expression relative to L1lacZ in the brain and ectopic expression in nonneural tissues. In contrast to L1lacZ mice, however, L1lacZDeltaN mice showed an unexpected loss, during postnatal development and in the adult, of beta-galactosidase expression in several neural structures, including the neural retina, cerebellum, cortex, striatum, and hippocampus. These data support the conclusion that the NRSE not only plays a role in the silencing of L1 expression in nonneural tissues during early development but also can function as a silencer and an enhancer of L1 expression in the nervous system of postnatal and adult animals.

Tissue-specific expression of the L1 cell adhesion molecule is modulated by the neural restrictive silencer element.

The cell adhesion molecule L1 mediates neurite outgrowth and fasciculation during embryogenesis and mutations in its gene have been linked to a number of human congenital syndromes. To identify DNA sequences that restrict expression of L1 to the nervous system, we isolated a previously unidentified segment of the mouse L1 gene containing the promoter, the first exon, and the first intron and examined its activity in vitro and in vivo. We found that a neural restrictive silencer element (NRSE) within the second intron prevented expression of L1 gene constructs in nonneural cells. For optimal silencing of L1 gene expression by the NRSE-binding factor RE-1-silencing transcription factor (REST)/NRSF, both the NRSE and sequences in the first intron were required. In transgenic mice, an L1lacZ gene construct with the NRSE generated a neurally restricted expression pattern consistent with the known pattern of L1 expression in postmitotic neurons and peripheral glia. In contrast, a similar construct lacking the NRSE produced precocious expression in the peripheral nervous system and ectopic expression in mesenchymal derivatives of the neural crest and in mesodermal and ectodermal cells. These experiments show that the NRSE and REST/NRSF are important components in restricting L1 expression to the embryonic nervous system.

Barx2, a new homeobox gene of the Bar class, is expressed in neural and craniofacial structures during development.

Homeobox genes are regulators of place-dependent morphogenesis and play important roles in controlling the expression patterns of cell adhesion molecules (CAMs). To identify proteins that bind to a regulatory element common to the genes for two neural CAMs, Ng-CAM and L1, we screened a mouse cDNA expression library with a concatamer of the sequence CCATTAGPyGA and found a new homeobox gene, which we have 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. Barx1 and Barx2 also encode an identical stretch of 17 residues downstream of the homeobox; otherwise, they share no appreciable homology. In vitro, Barx2 stimulated activity of an L1 promoter construct containing the CCATTAGPyGA motif but repressed activity when this sequence was deleted. Localization studies showed that expression of Barx1 and Barx2 overlap in the nervous system, particularly in the telencephalon, spinal cord, and dorsal root ganglia. Barx2 was also prominently expressed in the floor plate and in Rathke's pouch. During craniofacial development, Barx1 and Barx2 showed complementary patterns of expression: whereas Barx1 appeared in the mesenchyme of the mandibular and maxillary processes, Barx2 was observed in the ectodermal lining of these tissues. Intense expression of Barx2 was observed in small groups of cells undergoing tissue remodeling, such as ectodermal cells within indentations surrounding the eye and maxillo-nasal groove and in the first branchial pouch, lung buds, precartilagenous condensations, and mesenchyme of the limb. The localization data, combined with Barx2's dual function as activator and repressor, suggest that Barx2 may differentially control the expression of L1 and other target genes during embryonic development.

Structure of the human laminin gamma 2 chain gene (LAMC2): alternative splicing with different tissue distribution of two transcripts.

We have determined the structure of the human laminin gamma 2 chain gene (LAMC2), which is mutated in some patients with junctional epidermolysis bullosa. Eight lambda phage clones isolated from a genomic library and three subgenomic lambda phage clones made from a plasmid artificial chromosome clone spanned 75 kb, including the 55-kb gene. The LAMC2 gene contains 23 exons and is structurally highly homologous with the 28-exon LAMC1 gene (Kallunki et al., 1991, J. Biol. Chem. 266: 221-228), with 16 exons having identical sizes in the two genes. The gene analysis demonstrated that two previously described different size gamma 2 chain cDNAs (Kallunki et al., 1992, J. Cell Biol. 119: 679-693) are the result of alternative splicing. The longer gamma 2 chain is formed by using the coding sequence of the last exon 23, while the shorter gamma 2* chain is formed by using only 22 exons, together with part of the 5' end of intron 22. The two mRNAs were shown to have different expression patterns in 17-week-old human embryonic tissues, with the longer gamma 2 chain transcript strongly expressed in epithelia of all tissues studied, while distinct expression of the shorter gamma 2* chain mRNA was observed only in the cerebral cortex, in lung, and in distal tubules of the kidney.

Silencer elements modulate the expression of the gene for the neuron-glia cell adhesion molecule, Ng-CAM.

The combined factors that regulate the expression of cell adhesion molecules (CAMs) during development of the nervous system are largely unknown. To identify such factors for Ng-CAM, the neuron-glia CAM, constructs containing portions of the 5' end of the Ng-CAM gene were examined for activity after transfection into N2A neuroblastoma and NIH3T3 cells. Positive regulatory elements active in both cell types included an Ng-CAM proximal promoter with SP1 and cAMP response element motifs extending 447 base pairs upstream of a single RNA start site and a region within the first exon corresponding to 5'-untranslated sequences. Negative regulatory elements included five neuron-restrictive silencer elements (NRSEs) and a binding site for Pax gene products in a 305-base pair segment of the first intron. Constructs containing the promoter together with the entire first intron were active in N2A cells but were silenced in NIH3T3 cells. This silencer activity was mapped to the NRSEs. In contrast, the Pax motif inhibited activity of Ng-CAM constructs in both cell types. The DNA elements defined in these transfection experiments were examined for their ability to bind nuclear factors. The region within the first exon formed a DNA-protein complex after exposure to nuclear extracts prepared from both NIH3T3 and N2A cells. The NRSE region formed a more prominent complex with proteins prepared from NIH3T3 cells than it did with extracts from N2A cells. A member of the Pax protein family, Pax-3 bound to the Pax motif. Mutations introduced within the Pax motif in its ATTA sequence eliminated this binding whereas mutations in its GTTCC sequence did not, suggesting that paired homeodomain interactions are important for the recognition of Pax-3 by this DNA target sequence. The combined data suggest that negative regulation by NRSEs and Pax proteins may play a key role in the place-dependent expression patterns of Ng-CAM during development.

The gamma 2 chain of kalinin/laminin 5 is preferentially expressed in invading malignant cells in human cancers.

All known laminin isoforms are cross-shaped heterotrimeric molecules, consisting of one heavy alpha chain and two light beta and gamma chains. Recently, a cDNA encoding a new gamma chain from laminin 5 (also known as kalinin) was sequenced. This chain, named gamma 2, showed extended homology to the classical gamma 1 chain but differed from this by lacking the terminal globular domain. Recent data, indicating an important role of the gamma 2 chain gene in establishing adhesion contacts between epithelial cells and basement membranes, prompted us to investigate whether the gamma 2 chain gene is aberrantly expressed in cancer tissue, and if so whether its localization could provide clues to its possible role in cancer dissemination. Routinely processed tissue specimens from 36 cases of human cancer were investigated, including 16 cases of colon adenocarcinoma, 7 ductal mammary carcinomas, 4 squamous cell carcinomas, 3 malignant melanomas and 6 sarcomas. In situ hybridization for the detection of mRNAs for the gamma 2 chain and for the classical laminin chains alpha 1, beta 1, and gamma 1 was performed using S-35 labeled antisense RNA probes. As positive control of the specificity of the gamma 2 chain mRNA detection, two different anti-sense probes derived from two nonoverlapping cDNA clones were used. Malignant cells were found to express the gamma 2 chain in 29 of the 30 carcinomas studied and the expression was particularly high in cancer cells located at the invasion front. In contrast, mesenchymally derived cancer cells in three different types of sarcomas did not express the gamma 2 chain. In colon cancer there was a clear histological correlation between the expression of gamma 2 chain by cancer cells and their engagement in tumor budding processes. Laminin chains alpha 1, beta 1, and gamma 1 were weakly expressed throughout cancerous areas with no apparent correlation to sites of invasion. The aberrant expression of the gamma 2 chain gene seen in invasively growing cancer cells point to a role of this molecule in establishing focal adhesions of cancer cells to the extracellular matrix during their migration through surrounding normal tissue.

A truncated laminin chain homologous to the B2 chain: structure, spatial expression, and chromosomal assignment.

We describe the identification of a novel laminin chain. Overlapping clones were isolated from a human fibrosarcoma HT1080 cell cDNA library spanning a total of 5,200 bp. A second set of clones contained an alternative 3' end sequence giving a total of 4,316 bp. The longer sequence contained an open reading frame for a 1,193-residue-long polypeptide. The alternative sequence was shortened at the carboxyl-terminal end coding for a 1,111-residue-long polypeptide. The amino acid sequence contained 21 amino acids of a putative signal peptide and 1,172 residues or alternatively 1,090 residues of a sequence with five distinct domains homologous to domains I-V in laminin chains. Comparison of the amino acid sequences showed that the novel laminin chain is homologous to the laminin B2 chain. However, the structure of the novel laminin chain isolated here differs significantly from that of the B2 chain in that it has no domain VI and domains V, IV, and III are shorter, resulting in a truncated laminin chain. The alternative sequence had a shortened domain I/II. In accordance with the current nomenclature, the chain characterized here is termed B2t. Calculation of possible chain interactions of laminin chains with the B2t chain domain I/II indicated that the B2t chain can replace the B2 chain in some laminin molecules. The gene for the laminin B2t chain (LAMB2T) was localized to chromosome 1q25-q31 in close proximity to the laminin B2 chain gene. Northern analysis showed that the B2t chain is expressed in several human fetal tissues but differently from the laminin B1 and B2 chains. By in situ hybridization expression of the B2t chain was localized to specific epithelial cells in skin, lung, and kidney as opposed to a general epithelial and endothelial cell expression of the laminin B2 chain in the same tissues.

Human basement membrane heparan sulfate proteoglycan core protein: a 467-kD protein containing multiple domains resembling elements of the low density lipoprotein receptor, laminin, neural cell adhesion molecules, and epidermal growth factor.

The primary structure of the large human basement membrane heparan sulfate proteoglycan (HSPG) core protein was determined from cDNA clones. The cDNA sequence codes for a 467-kD protein with a 21-residue signal peptide. Analysis of the amino acid sequence showed that the protein consists of five domains. The amino-terminal domain I contains three putative heparan sulfate attachment sites; domain II has four LDL receptor-like repeats; domain III contains repeats similar to those in the short arms of laminin; domain IV has lg-like repeats resembling those in neural cell adhesion molecules; and domain V contains sequences resembling repeats in the G domain of the laminin A chain and repeats in the EGF. The domain structure of the human basement membrane HSPG core protein suggests that this mosaic protein has evolved through shuffling of at least four different functional elements previously identified in other proteins and through duplication of these elements to form the functional domains. Comparison of the human amino acid sequence with a partial amino acid sequence from the corresponding mouse protein (Noonan, D. M., E. A. Horigan, S. R. Ledbetter, G. Vogeli, M. Sasaki, Y. Yamada, and J. R. Hassell. 1988. J. Biol. Chem. 263:16379-16387) shows a major difference between the species in domain IV, which contains the Ig repeats: seven additional repeats are found in the human protein inserted in the middle of the second repeat in the mouse sequence. This suggests either alternative splicing or a very recent duplication event in evolution. The multidomain structure of the basement membrane HSPG implies a versatile role for this protein. The heparan sulfate chains presumably participate in the selective permeability of basement membranes and, additionally, the core protein may be involved in a number of biological functions such as cell binding, LDL-metabolism, basement membrane assembly, calcium binding, and growth- and neurite-promoting activities.

Cloning of human heparan sulfate proteoglycan core protein, assignment of the gene (HSPG2) to 1p36.1----p35 and identification of a BamHI restriction fragment length polymorphism.

We have isolated a cDNA coding for the core protein of the large basement membrane heparan sulfate proteoglycan (HSPG) from a human fibrosarcoma cell (HT1080) library. The library was screened with a mouse cDNA probe and one clone obtained, with a 1.5-kb insert, was isolated and sequenced. The sequence contained an open reading frame coding for 507 amino acid residues with a 84% identity to the corresponding mouse sequence. This amino acid sequence contained several cysteine-rich internal repeats similar to those found in component chains of laminin. The HSPG cDNA clone was used to assign the gene (HSPG2) to the p36.1----p35 region of chromosome 1 using both somatic cell hybrid and in situ hybridization. In the study of the polymorphisms of the locus, a BamHI restriction fragment length polymorphism was identified in the gene. This polymorphism displayed bands of 23 and 12 kb with allele frequencies of 76 and 24%, respectively.

Primary structure of the human laminin A chain. Limited expression in human tissues.

cDNA clones for the human laminin A chain were isolated from libraries prepared from human gestational choriocarcinoma cell line (JAR) RNA. They cover approx. 8 kb from the 5'-end of the 9.5 kb mRNA coding for this protein. Our clones contain 94 nucleotide residues for the 5'-end untranslated region and 7885 nucleotide residues of coding sequence. The complete human laminin A chain contains a 17-amino acid-residue signal peptide and a 3058-residue A chain proper. The human laminin A chain has a distinct domain structure with numerous internal cysteine-rich repeats. The large globular domain G has five repeats, which have several conserved glycine and cysteine residues. Furthermore the A chain contains 20 internal cysteine-rich repeats present in tandem arrays in three separate clusters (domains IIIa, IIIb and V). Domain I + II has a predicted continuous alpha-helical structure characterized by heptad repeats and three domains (IVa, IVb and VI) are predicted to contain a number of beta-sheets and coiled-coil structures. Northern-blot analysis was used to study the laminin A chain expression in the JAR cell line, full-term placenta and newborn-human tissues (kidney, spleen, lung, heart muscle, psoas muscle and diaphragm muscle). The expression was detectable in newborn-human kidney and JAR cell line only. The overall amino acid sequence identity between human and mouse is 76%. The human chain has only one Arg-Gly-Asp (RGD) sequence, which is located in the long arm within domain G, whereas the single RGD sequence in the mouse chain is located in the short arm in domain IIIb. The degree of identity between the human laminin A chain sequence and the sequence available for merosin [Ehrig, Leivo, Argraves, Ruoslahti & Engvall (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 3264-3268] is about 41% and when conservative substitutions are included the degree of similarity is 54%.

Structure of the human laminin B2 chain gene reveals extensive divergence from the laminin B1 chain gene.

The exon-intron structure of the human laminin B2 chain gene was elucidated from genomic lambda phage clones spanning 2 kilobase pairs (kb) of the 5'-flanking region, 58 kb of the structural gene and 10 kb of the 3'-flanking region. The entire gene was shown to contain 28 exons. The promoter region has no TATA or CAAT boxes whereas it contains five GC boxes and three AP-2-like binding sites. Comparison with the promoter region of the mouse gene revealed six highly conserved sequences of 14 to 42 base pairs in length. Sequencing of the last exon of the gene showed that the 3'-untranslated region of the mRNA can be up to 2797 nucleotides with five AATAAA potential polyadenylation signals. The similarity of the human 3'-untranslated sequence with that of mouse was shown to be 68.8%. The exon-intron structure of the laminin B2 chain gene demonstrated extensive divergence from the human laminin B1 chain gene, which has 34 exons. Only three intron locations are conserved in these two genes. The overall exon profile of the laminin B2 chain gene correlates only marginally with the pattern of structural domains and internal cysteine-rich repeats in the laminin B2 polypeptide chain.