Abstract.A new protein, fibulin-2, was predicted from sequence analysis of cDNA clones obtained from a mouse fibroblast library. This protein consists of a l195-residue polypeptide preceded by a 26-residue signal peptide. The COOH-terminal region of 787 amino acids contained three anaphylatoxin-related segments (domain I), 11 EGF-like repeats (domain II), 10 of which had a consensus motif for calcium-binding, and a ll5-residue globular domain HI. Except for two additional EGF-like repeats, this COOH-terminal region showed 43 % sequence identity with the previously described fibulin-1 (BM-90). The NH2-terminal 408 residues, unique to fibulin-2, showed no sequence homology to other known proteins and presumably form two additional domains that differ in their cysteine content. Recombinant fibulin-2 was produced and secreted by human cell clones as a disulfide-bonded trimer. Rotary shadowing visualized the protein as three 40-45 nm long rods which are connected at one end in a globe-like structure. No significant immunological cross-reaction could be detected between fibulin-1 and fibulin-2. Production of the fibulin-2 was demonstrated by Northern blots and radioimmunoassay in fibroblasts but not in several tumor cell lines. Together with the observation that the serum level of fibulin-2 is 1,000-fold lower than that of fibulin-1, the data indicate that these two isoforms are not always coordinately expressed. This is also suggested by Northern blots of tissue mRNAs and by immunofluorescence localizations using mouse tissues. The latter studies also demonstrated an extracellular localization for fibulin-2 in basement membranes and other connective tissue compartments.

Recessive mutations in two of the three collagen VI genes, COL6A2 and COL6A3, have recently been shown to cause Ullrich congenital muscular dystrophy (UCMD), a frequently severe disorder characterized by congenital muscle weakness with joint contractures and coexisting distal joint hyperlaxity. Dominant mutations in all three collagen VI genes had previously been associated with the considerably milder Bethlem myopathy. Here we report that a de novo heterozygous deletion of the COL6A1 gene can also result in a severe phenotype of classical UCMD precluding ambulation. The internal gene deletion occurs near a minisatellite DNA sequence in intron 8 that removes 1.1 kb of genomic DNA encompassing exons 9 and 10. The resulting mutant chain contains a 33-amino acid deletion near the amino-terminus of the triple-helical domain but preserves a unique cysteine in the triple-helical domain important for dimer formation prior to secretion. Thus, dimer formation and secretion of abnormal tetramers can occur and exert a strong dominant negative effect on microfibrillar assembly, leading to a loss of normal localization of collagen VI in the basement membrane surrounding muscle fibers. Consistent with this mechanism was our analysis of a patient with a much milder phenotype, in whom we identified a previously described Bethlem myopathy heterozygous in-frame deletion of 18 amino acids somewhat downstream in the triple-helical domain, a result of exon 14 skipping in the COL6A1 gene. This deletion removes the crucial cysteine, so that dimer formation cannot occur and the abnormal molecule is not secreted, preventing the strong dominant negative effect. Our studies provide a biochemical insight into genotype-phenotype correlations in this group of disorders and establish that UCMD can be caused by dominantly acting mutations.

Partial sequence comparisons have recently indicated that two extracellular components, fibulin from human placenta and BM-90 from a basement-membrane-producing mouse tumor, are either identical or closely related proteins. In this study, a complete sequence analysis of mouse BM-90 cDNA showed a 539-residue N-terminal core structure (domains I and II), which was 85% identical with the same core structure of human fibulin. A 137-residue C-terminal sequence (domain 111) was unique for BM-90 and could also be identified by Edman degradation. This suggested a novel splice product, variant D, which is characteristic for the mouse tumor. A second 117-residue C-terminal sequence (domain 111) was identified in additional mouse cDNA clones and showed 91% identity with the region specific for variant C of fibulin. Northern blots using mouse cells demonstrated two mRNA species, 2.7 kb and 2.3 kb, which encoded the variants D and C, respectively. The sequence of BM-90/fibulin indicates the presence of nine epidermal-growth-factor-like repeats in the core domain-I1 structure, eight of which contain consensus motifs for calcium binding. This binding is apparently important for the interaction of BM-90 with laminin and nidogen and for some weaker interactions with collagen IV. Further binding of BM-90 was demonstrated to fibronectin and BM-90 itself, but did not depend on calcium. Major binding sites for BM-90 were identified at a C-terminal segment of laminin A chain and at the N-terminus of nidogen. The broad interaction repertoire of BM-90 is comparable to that of nidogen and both proteins may have similar roles as connecting elements in the extracellular matrix.Various isoforms of laminin, collagen IV, heparan sulfate proteoglycans and nidogen are major constituents of basement membranes and have been extensively characterized. They form large supramolecular structures that interact with cells, and thus control the growth and differentiation of tissues (Paulsson, 1992;Rohrbach and Timpl, 1993). Several other constituents exist and their comprehensive analysis will be required to understand their integration into the extracellular matrix and their functional contributions. A novel component BM-90, consisting of a single 90-kDa polypeptide, was recently isolated from the basement-membrane-producing mouse Engelbreth-Holm-Swarm (EHS) tumor and was also shown to be synthesized by fibroblasts and teratocarcinoma cells . Approxiniately 120 amino acids of BM-90 were determined by Edman degradation and showed the presence of epidermal-growth-factor-like (EGF-like) repeats and approximately 50% sequence identity with the Nterminus of fibulin. This protein had been isolated from human placenta extracts by using its binding to the cytoplasmic Correspondence to R. Timpl, Max-Planck-Institut fur BioFax: t 49 89 8578 2422. Abbreviations. EGF, epidermal growth factor; EHS, EngelbrethHolm-Swarm ; PCR, polymerase chain reaction.Note. The sequences published here have been submitted to the EMBL data bank under the accession numbers X70853 (c...

Full-length al(V1) and a2(VI) cDNAs in an eukaryotic expression vector were used to obtain stably transfected human kidney cell clones and to purify these collagen-VI chains in substantial quantities from the culture medium. Both chains appeared mainly as monomers together with some dimers that were disulfide linked through their C-terminal globular domains. Despite sufficient hydroxylation of proline and lysine residues, the chains did not form a triple-helix, as shown by electronmicroscopy, CD spectra and pepsin sensitivity. Digestion of the chains with bacterial collagenase released the N-terminal and C-terminal globular domains, which were identified by their size and partial sequences. They showed a substantial content of a-helical conformation and a distinct globular structure after rotary shadowing. Antibodies could be raised that distinguished between the two chains and reacted with the globular domains. The a2(VI) but not the al(V1) chain showed binding to a heparan sulfate proteoglycan (perlecan), fibronectin and pepsin-solubilized collagen VI. Purified globular domains did not bind these ligands indicating the localization of binding sites within the triple-helical domain. Both chains showed a distinct affinity for heparin but failed to bind to various collagen types.

We have characterized cDNA clones that encode a newly discovered collagenous polypeptide. A 4-kilobase (kb) cDNA clone was initially isolated by screening a human fibroblast cDNA library with a probe encoding the collagenous domain of the human a3(VI) collagen. Subsequent screening of another fibroblast cDNA library yielded overlapping clones having a total length of 5.4 kb, which contained an open reading frame of 1603 amino acids including a 21-amino acid signal peptide. The predicted polypeptide consists of 10 collagenous domains 15-422 amino acids long, which were interspersed with 11 noncollagenous (NC) domains. Except for a large N-terminal NC11 domain of 312 residues, most of the NC domains were short (11-39 residues) and cysteine-rich. The overall structural arrangement differed significantly from other known collagen chains. Further analysis indicated that the deduced polypeptide exhibited several structural features characteristically seen in members of the fibril-associated collagen, types IX, XII, and XIV. In addition, the cysteine-rich motifs in the NC domains resembled those found in the cuticle collagen of Caenorhabdiis elegans. Northern blot analyses showed hybridization of the cDNA to a 5.5-kb mRNA in human fibroblasts and keratinocytes. The gene was localized by in situ hybridization to band p34-35 of human chromosome 1. The data clearly support the conclusion that the cDNA encodes a collagen chain that has not been previously described. We suggest that the cDNA clones encode the al chain of type XVI collagen.Collagens, the major constituents of connective tissues, represent a large family of structurally related proteins with distinct tissue distributions and functions (see refs. 1 and 2 for recent reviews). To date, 14 different types of collagen have been described in vertebrates and have been assigned Roman numerals. Based on their primary structure and supramolecular assembly, collagens can be divided into two major classes: the fibril-forming collagens and the non-fibrilforming collagens. A long central triple-helical domain, without Gly-Xaa-Xaa interruptions, is the hallmark of the former class. Types I, II, III, V, and XI, which form highly organized fibrils in a quarter-staggered fashion, are members of this class. The remaining types belong to the latter class. These collagens are very heterogeneous in size, but a common feature is the presence of imperfections in the Gly-Xaa-Xaa repeating pattern. Within the latter class, types IX, XII, and XIV collagens form a subgroup named the fibril-associated collagens with interrupted triple helices (FACIT) (3). These collagens are associated with type I or II collagen fibrils,

Fibulin-2 is an extracellular matrix protein belonging to the five-member fibulin family, of which two members have been shown to play essential roles in elastic fiber formation during development. Fibulin-2 interacts with two major constituents of elastic fibers, tropoelastin and fibrillin-1, in vitro and localizes to elastic fibers in many tissues in vivo. The protein is prominently expressed during morphogenesis of the heart and aortic arch vessels and at early stages of cartilage development. To examine its role in vivo, we generated mice that do not express the fibulin-2 gene (Fbln2) through homologous recombination of embryonic stem cells. Unexpectedly, the fibulin-2-null mice were viable and fertile and did not display gross and anatomical abnormalities. Histological and ultrastructural analyses revealed that elastic fibers assembled normally in the absence of fibulin-2. No compensatory up-regulation of mRNAs for other fibulin members was detected in the aorta and skin tissue. However, in the fibulin-2 null aortae, fibulin-1 immunostaining was increased in the inner elastic lamina, where fibulin-2 preferentially localizes. The results demonstrate that fibulin-2 is not required for mouse development and elastic fiber formation and suggest possible functional redundancy between fibulin-1 and fibulin-2.

No abstract

Introduction Pancreatic ductal adenocarcinoma (PDA) is a highly lethal disease in which a prominent desmoplastic reaction is a defining characteristic. Fibrillar collagens, such as collagen I and to a lesser extent, collagen III and V comprise the majority of this stromal fibrosis. Type VI collagen (COL6) forms a microfibrillar network associated with type I collagen fibrils. The expression of COL6 has been linked to inflammation and survival. Importantly, tumor-specific alternative splicing in COL6A3 has been identified in several cancers by genome exon arrays. We evaluated the expression and localization of COL6A3 in PDA and premalignant lesions and explored the presence of alternative splicing events. Methods We analyzed paired PDA-normal (n=18), IPMN (n=5), pancreatic cystadenoma (n=5), and eight PDA cell lines with RT-PCR, using unique primers that identify total COL6A3 gene and alternative splicing sites in several of its exons. Western blot analysis and immunohistochemistry were used to analyze the expression levels and localization of COL6A3 protein in the different lesions, and in two animal models of PDA. Results COL6A3 protein levels were significantly upregulated in 77% of the paired PDA-adjacent tissue examined. COL6A3 was mainly present in the desmoplastic stroma of PDA, with high deposition around the malignant ducts and in between the sites of stromal fatty infiltration. Analysis of the COL6A3 splice variants showed tumor-specific consistent inclusion of exons 3 and 6 in 17 of the 18 (94%) paired PDA-adjacent tissues. Inclusion of exon 4 was exclusively tumor-specific, with barely detectable expression in the adjacent tissues. IPMN and pancreatic cystadenomas showed no expression of any of the examined exons. Total COL6A3 mRNA and exon 6 were identified in six PDA cell lines, but only two cell lines (MIA PACA-2 and ASPC-1) expressed exons 3 and 4. In both the xenograft and transgenic models of PDA, COL6A3 immunoreactivity was present in the stroma and some PDA cells. Conclusions We describe, for the first time, a dynamic process of tumor-specific alternative splicing in several exons of stromal COL6A3. Alternatively spliced proteins may contribute to the etiology or progression of cancer and may serve as markers for cancer diagnosis. Identification of COL6A3 isoforms as PDA-specific provides the basis for future studies to explore the oncogenic and diagnostic potential of these alternative splicing events.