No abstract

Abstract.A new connective tissue protein, which we call fibrillin, has been isolated from the medium of human fibroblast cell cultures. Electrophoresis of the disulfide bond-reduced protein gave a single band with an estimated molecular mass of 350,000 D. This 350-kD protein appeared to possess intrachain disulfide bonds. It could be stained with periodic acid-Schiff reagent, and after metabolic labeling, it contained [3H]glucosamine. It could not be labeled with [35S]sulfate. It was resistant to digestion by bacterial collagenase. Using mAbs specific for fibrillin, we demonstrated its widespread distribution in the connective tissue matrices of skin, lung, kidney, vasculature, cartilage, tendon, muscle, cornea, and ciliary zonule. Electron microscopic immunolocalization with colloidal gold conjugates specified its location to a class of extracellular structural elements described as microfibrils. These microfibrils possessed a characteristic appearance and averaged 10 nm in diameter. Microfibrils around the amorphous cores of the elastic fiber system as well as bundles of microfibrils without elastin cores were labeled equally well with antibody. Immunolocalization suggested that fibrillin is arrayed periodically along the individual microfibril and that individual microfibrils may be aligned within bundles. The periodicity of the epitope appeared to match the interstitial collagen band periodicity. In contrast, type VI collagen, which has been proposed as a possible microfibrillar component, was immunolocalized with a specific mAb to small diameter microfilaments that interweave among the large, banded collagen fibers; it was not associated with the system of microfibrils identified by the presence of fibrillin.

The scleraxis (Scx) gene, encoding a bHLH transcription factor, is expressed in the progenitors and cells of all tendon tissues. To determine Scx function, we produced a mutant null allele. Scx -/-mice were viable, but showed severe tendon defects, which manifested in a drastically limited use of all paws and back muscles and a complete inability to move the tail. Interestingly, although the differentiation of all force-transmitting and intermuscular tendons was disrupted, other categories of tendons, the function of which is mainly to anchor muscles to the skeleton, were less affected and remained functional, enabling the viability of Scx -/-mutants. The force-transmitting tendons of the limbs and tail varied in the severity to which they were affected, ranging from dramatic failure of progenitor differentiation resulting in the loss of segments or complete tendons, to the formation of small and poorly organized tendons. Tendon progenitors appeared normal in Scx -/-embryos and a phenotype resulting from a failure in the condensation of tendon progenitors to give rise to distinct tendons was first detected at embryonic day (E)13.5. In the tendons that persisted in Scx -/-mutants, we found a reduced and less organized tendon matrix and disorganization at the cellular level that led to intermixing of tenocytes and endotenon cells. The phenotype of Scx -/-mutants emphasizes the diversity of tendon tissues and represents the first molecular insight into the important process of tendon differentiation.

. Basal keratinocytes attach to the underlying dermal stroma through an ultrastructurally unique and complex basement membrane zone. Electron-dense plaques along the basal surface plasma membrane, termed hemidesmosomes, appear to attach directly to the lamina densa of the basement membrane through fine strands, called anchoring filaments . The lamina densa is secured to the stroma through a complex of type VII collagen containing anchoring fibrils and anchoring plaques . We have identified what we believe is a novel antigen unique to this tissue region . The mAbs to this antigen localize to the anchoring filaments, just below the basal-dense plate of the hemidesmosomes . In cell culture, the antigen is deposited upon the culture substate by growing and migrating human keratinocytes. Addition of mAb to the cultures causes the cells to round and detach, but does not impair them metabolically. Skin fragments incubated with antibody exten-HE basement membrane zones at epithelial-stromml interfaces of external tissues are unique and complex . The dermal-epidermal junction is one ofthe best stud ied (Palade and Farquhar, 1965 ;Rowlatt, 1969;Susi, 1967;Bruns, 1969;Briggaman and Wheeler, 1975a ;Kawanami et al., 1978) . When visualized by EM following standard conditions of fixation, dehydration, and staining, a typical basal lamina is seen, containing a lamina lucida and lamina densa extending -100 nm from the epithelial basal surface (see Fig . 2 a). In addition, electron-dense thickenings, termed hemidesmosomes (Weiss and Ferris, 1954;Kelly, 1966), are seen along the basal keratinocyte surface. Thin filaments, termed anchoring filaments (Ellison and Garrod, 1984), appear to transverse the lamina lucida, bridging the hemidesmosome and the lamina densa . Along the dermal margin ofthe lamina densa, anchoring fibrils extend fromthe basement membrane. The anchoring fibrils either loop back and reinsert into the lamina densa, or extend perpendicularly from the basement membrane and insert into anchoring plaques, which are electron-dense condensations of the ends of anchoring fibrils toDr. Rousselle'spresent address is Laboratoire desSubstituts Cutan6s, H6pi-tal Fdouard Herriot, Lyon, France .C The Rockefeller University Press, 0021-9525/91/08/567/10 $2 .00 The Journal of Cell Biology, Volume 114, Number 3, August 1991567-576 sively de-epithelialize. These findings strongly suggest that this antigen is intimately involved in attachment of keratinocytes to the basement membrane.This antigen was isolated from keratinocyte cultures by immunoaffinity chromatography. Two molecules are observed. The most intact species contains three nonidentical chains, 165, 155, and 140 kD linked by interchain disulfide bonds. The second and more abundant species contains the 165-and 140-kD chains, but the 155-kD chain has been proteolytically cleaved to 105 kD. Likewise, two rotary-shadowed images are observed . The larger ofthe two, presumably corresponding to the most intact form, appears as an asymmetric 107-run-long rod, with a si...

Latent transforming growth factor ␤-binding protein 1 (LTBP-1) targets latent complexes of transforming growth factor ␤ to the extracellular matrix, where the latent cytokine is subsequently activated by several different mechanisms. Fibrillins are extracellular matrix macromolecules whose primary function is architectural: fibrillins assemble into ultrastructurally distinct microfibrils that are ubiquitous in the connective tissue space. LTBPs and fibrillins are highly homologous molecules, and colocalization in the matrix of cultured cells has been reported. To address whether LTBP-1 functions architecturally like fibrillins, microfibrils were extracted from tissues and analyzed immunochemically. In addition, binding studies were conducted to determine whether LTBP-1 interacts with fibrillins. LTBP-1 was not detected in extracted beaded-string microfibrils, suggesting that LTBP-1 is not an integral structural component of microfibrils. However, binding studies demonstrated interactions between LTBP-1 and fibrillins. The binding site was within three domains of the LTBP-1 C terminus, and in fibrillin-1 the site was defined within four domains near the N terminus. Immunolocalization data were consistent with the hypothesis that LTBP-1 is a fibrillin-associated protein present in certain tissues but not in others. In tissues where LTBP-1 is not expressed, LTBP-4 may substitute for LTBP-1, because the C-terminal end of LTBP-4 binds equally well to fibrillin. A model depicting the relationship between LTBP-1 and fibrillin microfibrils is proposed.The fibrillins and latent transforming growth factor ␤-binding proteins (LTBPs) 1 are members of a family of homologous molecules. The fibrillins and LTBPs contain multiple calciumbinding epidermal growth factor-like modules interspersed by a domain module (the 8-Cys or TB module), so far found only in these two proteins. Fibrillin-1 (1-4) and fibrillin-2 (5, 6) share a highly similar overall structure. Both molecules are of equivalent size (ϳ350 kDa) and domain organization. In contrast, LTBP-1 (7, 8), LTBP-2 (9), LTBP-3 (10), and LTBP-4 (11, 12) are each smaller than the fibrillins and variable in size.Extensive immunolocalization data combined with structural analyses of the fibrillin-1 monomer and fibrillin-containing microfibrils (1, 13-15) have established that fibrillin-1 is a major structural component of connective tissue microfibrils. In addition, genetic evidence in humans (16, 17) and mice (18,19) has confirmed that fibrillin-1 performs a significant role in the maintenance of microfibrils and elastic fibers.Fibrillin-2, whose structure is predicted to be highly similar to fibrillin-1, has also been immunolocalized to microfibrils (20). However, in contrast to fibrillin-1, the contribution of fibrillin-2 to microfibril structure is temporally and spatially restricted. In situ hybridization studies in mice indicated that expression of the fbn2 gene is most prominent in the early developing fetus (20). Genetic evidence in humans (5, 21) suggests that fibrillin-...