The integrin alpha 7 cytoplasmic domain regulates cell migration, lamellipodia formation, and p130CAS/Crk coupling.

The integrin alpha(7)beta(1) is the major laminin-binding integrin in skeletal, heart, and smooth muscle and is a receptor for laminin-1 and -2. It mediates myoblast migration on laminin-1 and -2 and thus might be involved in muscle development and repair. Previously we have shown that alpha(7)B as well as the alpha(7)A and -C splice variants induce cell motility on laminin when transfected into nonmotile HEK293 cells. In this study we have investigated the role of the cytoplasmic domain of alpha(7) in the laminin-induced signal transduction of alpha(7)beta(1) integrin regulating cell adhesion and migration. Deletion of the cytoplasmic domain did not affect assembly of the mutated alpha(7)Deltacyt/beta(1) heterodimer on the cell surface or adhesion of alpha(7)Deltacyt-transfected cells to laminin. The motility of these cells on the laminin-1/E8 fragment, however, was significantly reduced to the level of mock-transfected cells; lamellipodia formation and polarization of the cells were also impaired. Adhesion to the laminin-1/E8 fragment induced tyrosine phosphorylation of the focal adhesion kinase, paxillin, and p130(CAS) as well as the formation of a p130(CAS)-Crk complex in wild-type alpha(7)B-transfected cells. In alpha(7)BDeltacyt cells, however, the extent of p130(CAS) tyrosine formation was reduced and formation of the p130(CAS)-Crk complex was impaired, with unaltered levels of p130(CAS) and Crk protein levels. These findings indicate adhesion-dependent regulation of p130(CAS)/Crk complex formation by the cytoplasmic domain of alpha(7)B integrin after cell adhesion to laminin-1/E8 and imply alpha(7)B-controlled lamellipodia formation and cell migration through the p130(CAS)/Crk protein complex.

Laminin alpha4 and integrin alpha6 are upregulated in regenerating dy/dy skeletal muscle: comparative expression of laminin and integrin isoforms in muscles regenerating after crush injury.

The expression of laminin isoforms and laminin-binding integrin receptors known to occur in muscle was investigated during myogenic regeneration after crush injury. Comparisons were made between dystrophic 129ReJ dy/dy mice, which have reduced laminin alpha2 expression, and their normal littermates. The overall histological pattern of regeneration after crush injury was similar in dy/dy and control muscle, but proceeded faster in dy/dy mice. In vitro studies revealed a greater yield of mononuclear cells extracted from dy/dy muscle and a reduced proportion of desmin-positive cells upon in vitro cultivation, reflecting the presence of inflammatory cells and "preactivated" myoblasts due to ongoing regenerative processes within the endogenous dystrophic lesions. Laminin alpha1 was not detectable in skeletal muscle. Laminin alpha2 was present in basement membranes of mature myofibers and newly formed myotubes in control and dy/dy muscles, albeit weaker in dy/dy. Laminin alpha2-negative myogenic cells were detected in dy/dy and control muscle, suggesting the involvement of other laminin alpha chains in early myogenic differentiation, such as laminin alpha4 and alpha5 which were both transiently expressed in basement membranes of newly formed myotubes of dy/dy and control mice. Integrin beta1 was expressed on endothelial cells, muscle fibers, and peripheral nerves in uninjured muscle and broadened after crush injury to the interstitium where it occurred on myogenic and nonmyogenic cells. Integrin alpha3 was not expressed in uninjured or regenerating muscle, while integrin alpha6 was expressed mainly on endothelial cells and peripheral nerves in uninjured muscle. Upon crush injury integrin alpha6 increased in the interstitium mainly on nonmyogenic cells, including infiltrating leukocytes, endothelial cells, and fibroblasts. In dy/dy muscle, integrin alpha6 occurred on some newly formed myotubes. Integrin alpha7 was expressed on muscle fibers at the myotendinous junction and showed weak and irregular expression on muscle fibers. After crush injury, integrin alpha7 expression extended to the newly formed myotubes and some myoblasts. However, many myoblasts and newly formed myotubes were integrin alpha7 negative. No marked difference was observed in integrin alpha7 expression between dy/dy and control muscle, either uninjured or after crush injury. Only laminin alpha4 and integrin alpha6 expression patterns were notably different between dy/dy and control muscle. Expression of both molecules was more extensive in dy/dy muscle, especially in the interstitium of regenerating areas and on newly formed myotubes. In view of the faster myogenic regeneration observed in dy/dy mice, the data suggest that laminin alpha4 and integrin alpha6 support myogenic regeneration. However, whether these accelerated myogenic effects are a direct consequence of the reduced laminin alpha2 expression in dy/dy mice, or an accentuation of the ongoing regenerative events in focal lesions in the muscle, requires further investigation.

The role of extracellular and cytoplasmic splice domains of alpha7-integrin in cell adhesion and migration on laminins.

The major laminin-binding integrin of skeletal, smooth, and heart muscle is alpha7beta1-integrin, which is structurally related to alpha6beta1. It occurs in three cytoplasmic splice variants (alpha7A, -B, and -C) and two extracellular forms (X1 and X2) which are developmentally regulated and differentially expressed in skeletal muscle. Previously, we have shown that ectopic expression of the alpha7beta-integrin splice variant in nonmotile HEK293 cells specifically induced cell locomotion on laminin-1 but not on fibronectin. To investigate the specificity and the mechanism of the alpha7-mediated cell motility, we expressed the three alpha7-chain cytoplasmic splice variants, as well as alpha6A- and alpha6B-integrin subunits in HEK293 cells. Here we show that all three alpha7 splice variants (containing the X2 domain), as well as alpha6A and alpha6B, promote cell attachment and stimulate cell motility on laminin-1 and its E8 fragment. Deletion of the cytoplasmic domain (excluding the GFFKR consensus sequence) from alpha7B resulted in a loss of the motility-enhancing effect. On laminin-2/4 (merosin), the predominant isoform in mature skeletal muscle, only alpha7-expressing cells showed enhanced motility, whereas cells transfected with alpha6A and alpha6B neither attached nor migrated on laminin-2. Adhesion of alpha7-expressing cells to both laminin-1 and laminin-2 was specifically inhibited by a new monoclonal antibody (6A11) specific for alpha7. Expression of the two extracellular splice variants alpha7X1 and alpha7X2 in HEK293 cells conferred different motilities on laminin isoforms: Whereas alpha7X2B promoted cell migration on both laminin-1 and laminin-2, alpha7X1B supported motility only on laminin-2 and not on laminin-1, although both X1 and X2 splice variants revealed similar adhesion rates to laminin-1 and -2. Fluorescence-activated cell sorter analysis revealed a dramatic reduction of surface expression of alpha6-integrin subunits after alpha7A or -B transfection; also, surface expression of alpha1-, alpha3-, and alpha5-integrins was significantly reduced. These results demonstrate selective responses of alpha6- and alpha7-integrins and of the alpha7 splice variants to laminin-1 and -2 and indicate differential roles in laminin-controlled cell adhesion and migration.

Absence of integrin alpha 7 causes a novel form of muscular dystrophy.

Integrin alpha 7 beta 1 is a specific cellular receptor for the basement membrane protein laminin-1 (refs 1,2), as well as for the laminin isoforms -2 and -4 (ref. 3). The alpha 7 subunit is expressed mainly in skeletal and cardiac muscle and has been suggested to be involved in differentiation and migration processes during myogenesis. Three cytoplasmic and two extracellular splice variants that have been described are developmentally regulated and expressed in different sites in the muscle. In adult muscle, the alpha 7A and alpha 7B subunits are concentrated in myotendinous junctions but can also be detected in neuromuscular junctions and along the sarcolemmal membrane. To study the potential involvement of alpha 7 integrin, during myogenesis and its role in muscle integrity and function, we generated a null allele of the alpha 7 gene (Itga7) in the germline of mice by homologous recombination in embryonic stem (ES) cells. Surprisingly, mice homozygous for the mutation are viable and fertile, indicating that the alpha 7 beta 1 integrin is not essential for myogenesis. However, histological analysis of skeletal muscle revealed typical symptoms of a progressive muscular dystrophy starting soon after birth, but with a distinct variability in different muscle types. The observed histopathological changes strongly indicate an impairment of function of the myotendinous junctions. These findings demonstrate that alpha 7 beta 1 integrin represents an indispensable linkage between the muscle fibre and the extracellular matrix that is independent of the dystrophin-dystroglycan complex-mediated interaction of the cytoskeleton with the muscle basement membrane.

Specific induction of cell motility on laminin by alpha 7 integrin.

Laminin, the major glycoprotein of basement membranes, actively supports cell migration in development, tissue repair, tumor growth, metastasis, and other pathological processes. Previously we have shown that the locomotion of murine skeletal myoblasts is specifically and significantly enhanced on laminin but not on other matrix proteins. One of the major laminin receptors of myoblasts is the alpha 7 beta 1 integrin, which was first described in human MeWo melanoma cells and Rugli glioblastoma cells. In order to investigate and directly test the role of the alpha 7 integrin in cell migration on laminin, we expressed the murine alpha 7B splice variant in human 293 kidney cells and 530 melanoma cells which cannot migrate on laminin and are devoid of endogenous alpha 7. Northern blotting of the transfected cells showed that the alpha 7 mRNA was expressed efficiently, and the protein was detected on the cell surface by immunofluorescence and fluorescence-activated cell sorter analysis. Cell motility measurements by computer-assisted time-lapse videomicroscopy of the alpha 7-transfected cells revealed an 8-10-fold increase in motility on laminin-1 and its E8 fragment, but not on fibronectin. Mock-transfected cells did not migrate significantly of alpha 7-transfected 293 cells through laminin-coated filters in a Boyden chamber assay was significantly enhanced in comparison to mock transfected cells. These findings prove that alpha 7 integrin expression confers a gain of function-motile phenotype to immobile cells and may be responsible for transduction of the laminin-induced cell motility.

Localization of beta 1-integrins in human cartilage and their role in chondrocyte adhesion to collagen and fibronectin.

In the past, proteins have been described that may be involved in chondrocyte interactions with extracellular collagen, but little is known about the role of integrins in chondrocyte-collagen interactions. Here we report on the analysis of beta 1-integrin distribution in human fetal cartilage and on the expression of integrins on fetal chondrocytes, using monoclonal and polyclonal antibodies to integrin alpha- and beta-chains. We show the presence of alpha 2-, alpha 5-, alpha 6-, alpha v-, and beta 1-chains on freshly isolated chondrocytes by surface immunofluorescence in the fluorescence-activated cell sorter and by surface iodination followed by immunoprecipitation. Affinity chromatography of bovine chondrocyte membrane proteins on a collagen-Sepharose column followed by immunoprecipitation confirmed the presence of the collagen-binding alpha 2 beta 1-integrin on chondrocytes. Chondrocyte adhesion on native collagens I and II, on fibronectin, and on laminin was completely blocked by anti-beta 1; anti-alpha 2 reduced chondrocyte binding to collagen by only 40-50%; similarly, anti-alpha 1-antibodies were also able to reduce chondrocyte binding to collagen, although alpha 1 could not be unequivocally identified on chondrocytes. Chondrocyte adhesion to fibronectin was Mg(2+)- and Ca(2+)-dependent and could be inhibited by anti-alpha 5 and by RGD peptides. Chondrocyte adhesion to native collagens is Mg(2+)-, but not Ca(2+)-dependent and RGD-independent. Interestingly, although these data point to a role of alpha 2 beta 1 in chondrocyte-collagen interactions in vitro, alpha 2 could not be visualized in sections of human fetal cartilage, in contrast to the beta 1-, alpha v-, and alpha 5-chains which were present. This suggests that alpha 2 beta 1-integrin may be involved in the assembly of a pericellular collagen matrix in vitro, but may not be required for chondrocyte-collagen interactions in intact cartilage.

Skeletal myoblasts utilize a novel beta 1-series integrin and not alpha 6 beta 1 for binding to the E8 and T8 fragments of laminin.

The E8 fragment of laminin stimulates myoblast attachment and locomotion. Myoblast attachment to laminin/E8 was blocked by anti-integrin antibodies against beta 1-chains but not by antibodies against alpha 6-chains. By contrast, other cell lines (e.g. B16, HT1080, P19, F9, Pys2, 3T3, and 3T6) were blocked both by anti-beta 1 and anti-alpha 6. All cells tested also bound to approximately 125-kDa C-terminal fragments of E8 (T8 and T8'). Immunoprecipitation of surface-iodinated myoblasts revealed beta 1-, alpha 3-, and alpha 5-integrin chains and a novel chain that co-precipitated with anti-beta 1 antibodies running at approximately 95 kDa (reduced). I125-alpha 6 beta 1 was immunoprecipitated from cells whose attachment to E8 was blocked by anti-alpha 6 antibodies. By contrast, little alpha 6 beta 1 could be immunoprecipitated from myoblasts. beta 1-Integrin and the novel alpha-chain (alpha'), Mr approximately 120,000/approximately 95.000 (nonreduced/reduced), from myoblast lysates were retained during affinity chromatography on Engelbreth-Holm-Swarm-laminin affinity columns. beta 1, alpha 1, and the novel alpha' were retained from Rugli cell lysates on Engelbreth-Holm-Swarm-laminin columns. alpha 3 was not bound. When E8 was used as affinity matrix, only beta 1 and alpha' were retained. The N-terminal sequence of Rugli alpha' was homologous to alpha-chains of beta 1-series integrins and was most similar to alpha 6 (9 identical residues out of 14). However, there were distinctive differences; in particular, 2 residues were deleted in comparison with alpha 6.

Multiple cell surface receptors for the short arms of laminin: alpha 1 beta 1 integrin and RGD-dependent proteins mediate cell attachment only to domains III in murine tumor laminin.

Cell surface molecules that interact with the cross formed by the three short arms of murine tumor laminin were studied using thermal perturbation, antibody and peptide blocking, and affinity chromatography. Several potential receptors for the laminin short arms were revealed that differed from those mediating cell attachment to the E8 (long arm) fragment. Two cell lines, Rugli and L8 attached well to E1-X (short arm) fragments of laminin. This attachment was blocked by antibodies against alpha 1 integrin chains. Other cells were unable to attach strongly to E1-X, but attached to P1. This attachment was unaffected by anti-beta 1 integrin antibodies, but specifically blocked by the peptide GRGDS. By contrast, binding of Rugli cells was RGD independent and blocked by anti-beta 1 integrin antibodies. G7 and C2C12 myoblasts were very sensitive to GRGDS (ID50 approximately 2 micrograms.ml-1) for attachment to P1 which implied that a non-beta 1 series integrin, possibly alpha v beta 3, was involved. On heat denaturation of P1(3) attachment remained sensitive to RGDS and ID50 was unchanged. On heat denaturation of E1-X, attachment remained sensitive to RGDS but the ID50 increased to approximately 200 micrograms.ml-1. Cellular beta 1 integrins were retained on laminin affinity columns. A beta 1 integrin with an approximately 190 kD alpha-chain could be isolated from Rugli cells whose attachment could be blocked by anti-alpha 1 antibodies and not from cells blocked by RGDS peptides. Anti-alpha 1 antibodies blocked Rugli attachment to native laminin, but only when the E8 cell binding sites on laminin were also blocked. Thus, a receptor related to alpha 1 beta 1 integrin can function simultaneously with a receptor for E8. Anti-alpha 1 also blocked attachment to heated laminin, suggesting that the heat-stable attachment activity in laminin involved the E1-X binding site. Thus, at least two putative receptors mediate attachment to the short arms of laminin. One, related to alpha 1 beta 1 integrin, recognizes RGDS-independent sites in E1-X defined by P1 (within domains III, IIIa, IIIb), and one is an RGD-dependent molecule recognizing sites in P1, and is not a beta 1 integrin.

Identification of a major antigenic epitope on CNBr-fragment 11 of type II collagen recognized by murine autoreactive B cells.

Immunization of certain strains of mice with native type II collagen (CII) induces both development of arthritis and an antibody response to autologous CII. The autoantibody response in a high-responder strain, the DBA/1 mouse, has been described earlier, and a number of monoclonal antibodies have been characterized for arthritogenicity and autoreactive binding to cartilage in vivo and in vitro. Here we map the antigenic epitope of one of these arthritogenic monoclonal antibodies (CII-C1). It belongs to a group of antibodies recognizing the CNBr fragment alpha 1(II)-CB11 of CII. Using the enzyme-linked immunosorbent assay technique, we show that the antibody reacts only with native, triplehelical CII, but not with other collagens. The antibody is able to stain specifically the CB11 fragment by immunoblotting, suggesting some partial renaturation of the CNBr fragment into triple-helical structures after blotting. The binding site of CII-C1 on CB11 was further focused by rotary shadowing of antibody-labeled CII to a site 89 +/- 8 nm from the amino end of CII, corresponding to the middle of CB11. This location was confirmed by cleavage of CB11 with trypsin, separation of the tryptic peptides by high-performance liquid chromatography and dot-blot analysis of the antigenic peptides with the CII-C1 antibody. Sequencing of the single positive peptide located the antigenic epitope within the sequence GFAGQAGPAGATGAPGRP (residues 316-333). Assuming 0.29 nm per residue, this corresponds to a position within 92-96.5 nm from NH2 terminal end of CII. Apart from glycine residues, which are not exposed on the triple-helical structure, only two amino acid residues (F-x-y-Q) are conserved in CII from different species but are not found in the triple-helix of other collagens except type IV collagen. Therefore, this structure is likely to be of critical importance for the binding of the CII-C1 antibody. Of potential importance is that this structure is also found in certain other arthritogenic proteins such as 65-kDa mycobacterial protein, in CMV and EBV.

Long-term physical exercise retards trabecular bone loss in lumbar vertebrae of aging female mice.

The present study examined the effect of long-term, moderate physical exercise on trabecular bone volume (TBV), calcium content, 3H-proline uptake, and the activities of alkaline and acid phosphatases in lumbar vertebrae of aging and senescent mice. It became apparent that if physical activity starts at an early stage of life, i.e., prior to middle age and is extended until old age, it exerts beneficial effects on trabecular bone mass and mineralization. Such a positive effect is not obtained if the training program is initiated after middle age. The training-induced reduction in bone loss was accompanied by a significant decrease in acid phosphatase activity whereas no changes took place with regard to the activity of alkaline phosphatase. Long-term physical exercise also enhanced the uptake of 3H-proline by lining cells along the bone trabecules. In spite of its moderate nature, the endured training program served as a stress factor for the involved animals, a fact that was manifested by an increase in the serum levels of corticosterone. Thus, it seems that whereas young animals respond favorably to such a stimulatory stress, older animals lose this ability of adaptation.

Cell attachment properties of collagen type VI and Arg-Gly-Asp dependent binding to its alpha 2(VI) and alpha 3(VI) chains.

Twelve of sixteen different cell types including fibroblasts and tumor cells were able to attach and spread on substrates of pepsin-solubilized or intact collagen VI, and on its triple helical domain. Attachment and spreading were independent of soluble mediator proteins (fibronectin, laminin) and collagen VI was distinct from collagens I, IV and V in the cells with which it interacted. Many of the same cells bound and spread on substrates prepared from unfolded alpha 2(VI) and alpha 3(VI) chains but not on the alpha 1(VI) chain. The interactions with the chains were inhibited by low concentrations (10-100 microM) of synthetic RGDS and RGDT but not RGES peptides while the binding of cells to pepsin-solubilized collagen VI was more than 20-fold less sensitive to these peptides. The data indicate that cells have the ability to bind to collagen VI in a specific manner suggesting a similar function for collagen VI in situ.

Antibodies to basement membrane protein nidogen in Chagas' disease and American cutaneous leishmaniasis.

About 50 to 70% of sera from patients with American cutaneous leishmaniasis and chronic Chagas' disease possessed antibodies which reacted in enzyme and radioimmunoassays with nidogen obtained from a tumor basement membrane. The antibodies were of the immunoglobulin M and G classes in acute American cutaneous leishmaniasis but mainly of the immunoglobulin G class in chronic Chagas' disease. Similar antibodies could not be detected in patients suffering from a variety of other infectious or inflammatory diseases when compared with healthy control groups. Inhibition and immunoadsorption studies indicated a close relationship of epitopes recognized by patients' antibodies on nidogen and on another basement membrane protein, laminin. Since rabbit antisera to both proteins do not cross-react, a special nature of the epitopes involved in the reaction with patient sera is suggested. Similar epitopes may exist on various forms of Leishmania or Trypanosoma protozoa.

Globular domain of basement membrane collagen induces autoimmune pulmonary lesions in mice resembling human Goodpasture disease.

A distinct circulating antibody response could be evoked in C57BL mice after immunization with the globular domain NC1 of basement membrane collagen IV obtained from the mouse Engelbreth-Holm-Swarm tumor when injected together with complete Freund's adjuvant. The antibodies reacted with various subunits of NC1, did not cross-react with other basement membrane proteins, and exhibited a tissue reactivity restricted to certain basement membranes. Tissue-bound antibodies could be detected by direct immunofluorescence and were distributed together with C3 in a linear pattern along glomerular and alveolar basement membranes. Pathological changes were mainly observed in lung and kidney and consisted of inflammatory infiltrates and massive hemorrhages with strong granulomatous fibrotic development in the lung. Kidney alterations were comparably weaker and of focal nature. A nephrotoxic serum model showed rapid binding of rabbit antibodies against mouse NC1 to lung, liver, and kidney basement membranes which was followed several weeks later by an autologous phase with anti-rabbit IgG antibodies bound to basement membranes as immune complexes. There was no fibrotic response but hemorrhagic and inflammatory lung and kidney changes similar to those after active immunization were observed after passive transfer. The experimental NC1 autoimmune model has several features such as anti-NC1 response, tissue restriction, lung hemorrhages, and glomerulonephritis in common with patients suffering from Goodpasture disease. The development of lung fibrosis appears to be unique for the animal model.

Size and domain structure of collagen VI produced by cultured fibroblasts.

Disulfide-bonded forms of collagen VI were analyzed by immunoblotting of fibroblast culture medium and cell extracts. The protein consists of pepsin and collagenase-resistant domains of about equal size indicating a molecular mass of 340 kDa for collagen VI monomers.

Immunochemical and autoantigenic properties of the globular domain of basement membrane collagen (type IV).

Polyclonal rabbit antibodies raised against the globular domain NC1 of collagen IV from human placenta and a mouse tumor react with conformational antigenic determinants present on the NC1 hexamers and also with the three major subunits obtained after dissociation. The antibodies recognized unique structures within basement membranes and showed a broad tissue reactivity but only limited species cross-reactivity. Using these antibodies, it was possible to detect small amounts of collagen IV antigens from cell cultures and in serum. Monoclonal rat antibodies against mouse NC1 revealed a similar reaction potential. Autoantibodies could be produced in mice against mouse NC1 which react with kidney and lung basement membranes in a pathological manner, mimicking Goodpasture syndrome.

Structure and biology of the globular domain of basement membrane type IV collagen.

A procedure was developed for purifying the globular domain NC1 of basement membrane collagen from collagenase digests of a variety of tissues. The globule (Mr = 170,000) is a hexameric structure originating from two collagen IV molecules that are cross-linked at their COOH-terminal ends. Dissociation into subunits derived from alpha 1(IV) and alpha 2(IV) chains occurs at a pH below 4 and after denaturation (8 M urea). The subunits obtained include monomers (Mr = 28,000) and two different dimers (Da,Db) which are connected by disulfide bonds (Db) and/or nonreducible bonds (Da). Almost perfect reconstitution to hexamers is obtained in neutral buffer with mixtures of the subunits or purified dimers but not with purified monomers. Stabilization by dimer formation and other physical data suggest conformationally distinct segments within the subunits, which is also supported by a repeating subdomain structure deduced from cDNA sequences. Monocline crystals of NC1 give a sufficiently detailed X-ray diffraction pattern that should permit elucidation of the three-dimensional structure of the hexamer. Antibodies raised against the globular domain react with all subunits and mainly recognize epitopes stabilized by internal disulfide bridges and/or the hexameric assembly. Immunoprecipitation tests with these antibodies demonstrated a slightly larger subunit size of NC1 in PYS-2 cell culture and the rapid release of precursor-specific segments prior to secretion from the cells. Autoantibodies against mouse tumor NC1 were produced in mice and were detected both in the blood and as tissue-bound forms (kidney, lung). The autoantibody response is accompanied by certain pathological alterations mimicking Goodpasture's syndrome. The possible relationship between the two diseases is substantiated by reaction of Goodpasture antisera with the globular domain obtained from various tissue sources.

Structure and macromolecular organization of type VI collagen.

Collagen VI is a large, disulfide-bonded protein complex which is widely distributed in connective tissue. The constituent polypeptide chains (Mr = 110,000-140,000) consist of collagenous and noncollagenous segments, are degraded to chains of about half the size when collagen VI is solubilized by pepsin, and assemble to a unique pattern of oligomers. As revealed by electron microscopy, the triple-stranded protomer consists of a triple helix 105 nm in length flanked on each side by globular domains of similar size (diameter about 7 nm). Protomers are assembled to dimers by an antiparallel staggered alignment of triple-helical segments. This leads to inner regions, 75 nm in length, of two slightly supercoiled triple helices flanked by globular domains. At both sides 30-nm-long outer triple-helical segments emerge that are terminated by globules. Tetramers are formed from laterally aligned dimers that cross with their outer triple-helical segments in a scissors-like fashion. The same structures, except with much smaller globular domains, are found in pepsin-treated collagen VI. Disulfide-linked collagen VI produced by cultured fibroblasts has a size similar to that of genuine collagen VI found in tissue extracts. Larger forms of collagen VI are assembled from tetramers by end-to-end aggregation which because of an overlap of the outer segments brings all globular domains close together. This arrangement predicts microfibrillar structures in tissues with a periodicity of 100-110 nm and a diameter of 5-10 nm. Structures consistent with this proposal were indeed found by immunoelectron microscopy of placenta and aorta using the ferritin technique. Large, lateral aggregates of collagen VI microfibrils may in addition exist in cell cultures and tissues ("zebra collagen," "Luse bodies") and are presumably maintained by contacts between globular domains.

Immunochemistry, genuine size and tissue localization of collagen VI.

Collagen VI was solubilized with pepsin from human placenta and used for preparing rabbit antisera. Major antigenic determinants were located in the central region of the antigen including triple-helical and globular structures. Antisera prepared against a constituent-chain showed preferential reactions with unfolded structures. Antibodies were purified by affinity chromatography and failed to cross-react with other collagen types I-V and with fibronectin. These antibodies demonstrated intracellular and extracellular collagen VI in fibroblast and smooth muscle cell cultures. Immunoblotting identified a disulfide-bonded constituent chain about twice as large as those of the pepsin fragments in both cell cultures and tissue extracts. Rotary shadowing electron microscopy indicated that the increase in mass is due to larger globular domains present at both ends of collagen VI monomers. Indirect immunofluorescence demonstrated a wide occurrence of collagen VI in connective tissue particularly of large vessels, kidney, skin, liver and muscle. Collagen VI is apparently not a typical constituent of cartilage or of basement membranes. Ultrastructural studies using the immunoferritin technique showed collagen VI along thin filaments or in amorphous regions of aortic media or placenta but not in association with thick, cross-striated collagen fibrils or elastin. This supports previous suggestions that collagen VI is a constituent of microfibrillar structures of the body.