Molecular packing in type I collagen fibrils. A model with neighbouring collagen molecules aligned in axial register

A detailed stereochemical analysis of intermolecular interactions of collagens made with molecular models and summarized experimental data resulted in a new three-dimensional structural model for collagen fibrils. In this model collagen molecules aligned in axial register form a bunch. The bunches are aligned head to tail and penetrate by 300 A into each other, forming microfibrils; these in turn assemble into fibrils. The new model differs from all the others in that its characteristic axial regularity, with a period of 670 A, results from staggering of the adjacent microfibrils formed by unstaggered molecules rather than from the axial staggering of neighbouring collagen molecules.     

A procollagen C-proteinase inhibitor diminishes collagen and lysyl oxidase processing but not collagen cross-linking in osteoblastic cultures

The deposition of insoluble functional collagen occurs following extracellular proteolytic processing of procollagens by procollagen N- and C-proteinases, fibril formation, and lysyl oxidase dependent cross-linking. Procollagen C-proteinases in addition process and activate lysyl oxidase. The present study evaluates a possible role for procollagen C-proteinases in controlling different aspects of collagen deposition in vitro. Studies determine whether inhibition of procollagen C-proteinase activity with a specific BMP-1 inhibitor results in perturbations in lysyl oxidase activation, and in collagen processing, deposition, and cross-linking in phenotypically normal cultured murine MC3T3-E1 cells. Data show that BMP-1 Inhibitor dose dependently inhibits lysyl oxidase activation by up to 50% in undifferentiated proliferating cells. In differentiating cultures, BMP-1 inhibitor decreased collagen processing but did not inhibit the accumulation of mature collagen cross-links. Finally, electron microscopy studies show that collagen fibril diameter increased. Thus, inhibition of procollagen C-proteinases results in perturbed collagen deposition primarily via decreased collagen processing.     

A fibronectin-binding protein from Streptococcus equi binds collagen and modulates cell-mediated collagen gel contraction

The N-terminal fragment (FNZN) of the fibronectin-binding protein FNZ from Streptococcus equi subspecies zooepidemicus was investigated as to effects on murine cell interactions with extracellular matrix proteins. FNZN bound to immobilized fibronectin (FN) and native, but not denatured, collagen type I. FNZN had no effect on primary adhesion of cells from the murine myoblastic C2C12 cell line to immobilized fibronectin. C2C12 cells adhered to immobilized FNZN, a process that was not inhibited by anti-human FN IgG or by an inhibitor of integrin alphaVbeta3. C2C12 cells lack collagen-binding beta1 integrins and neither adhere to native collagen nor mediate contraction of three-dimensional collagen gels. FNZN stimulated collagen gel contraction by C2C12 cells but not adhesion of C2C12 cells to collagen. Experiments with an alphaVbeta3-inhibitor suggested that FNZN promoted contraction by a process requiring alphaVbeta3. Our data suggest that FNZN by binding to cells, collagen, and FN modulate complex adhesive processes mediated by the alphaVbeta3 integrin. Since alphaVbeta3-mediated contractile events function to counteract edema formation during inflammation, it is possible that FNZN and its secreted homologue FNE modulate edema responses in infected tissues.     

A new look at vitreous-humour collagen.

The collagens of bovine vitreous-humour and nasal-septum cartilage have been extracted, fractionated and compared. Both tissues show the same heterogeneity of collagen types, consisting of type II, 1 alpha, 2 alpha, 3 alpha and C-PS collagens. The type II collagen of the vitreous humour was significantly more hydroxylated both in the lysine and proline residues than was that of cartilage. C-PS1 collagen, together with higher-Mr forms were present in the vitreous humour, but the higher-Mr forms were not seen in cartilage. Both C-PS1 and C-PS2 were present in vitreous humour and cartilage, but vitreous humour contained three times more of these collagens than did cartilage. Despite the difference in amount, the molar ratio C-PS1/C-PS2 was approx. 1 in both tissues, suggesting that they are components of a larger molecule. The 1 alpha, 2 alpha, 3 alpha collagens were present in the same concentration in both tissues. These three chains co-precipitated on dialysis against phosphate-buffered saline, pH 7.2, in a manner analogous to type V collagen.     

A spectroscopic collagenase assay using peroxidase-labeled collagen

A quantitative collagenase assay detecting soluble collagen fragments is described in this paper. Using the reagent N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP) type I collagen was conjugated with horseradish peroxidase (POD) which was employed as a reporter enzyme. POD was preferentially linked to the TC B fragment in a ratio of 1.4 mol POD/mol collagen. The conjugation product was immobilized on AH-Sepharose via carbodiimide coupling to form the final collagenase substrate used in the assay. POD activity in the supernatants caused by liberated TC B fragments exhibited a linear relationship for collagenase concentrations up to 100 micrograms/ml bacterial collagenase. Over an incubation period of 4 h the lowest detection limits found were 20 ng/100 microliters for bacterial collagenase and 60 ng/100 microliters for human leukocyte collagenase. Incubation of the assay mixture with 5 micrograms trypsin resulted in 3.8% of the activity released by the equivalent amount of leukocyte collagenase. The assay developed here has been shown to be sensitive and specific for collagenase, with the additional advantage that this method is suited for simple and economic handling.     

A second fibronectin-binding region is present in collagen alpha chains

The interactions of plasma fibronectin with alpha chains or cyanogen bromide fragments of collagen types I and II have been studied using a variety of techniques. Affinity chromatography of cyanogen bromide-cleaved type II collagen on immobilized fibronectin revealed the binding of cyanogen bromide fragment CB12 in addition to the previously characterized CB10. Using fluorescence polarization, we analyzed the interaction between the collagen peptides and fluorescein isothiocyanate-labeled 42-kDa gelatin-binding fragment of fibronectin in solution. Dissociation constants for the binding of CB10 and CB12 to the fibronectin fragment were calculated as 0.38 and 0.94 microM, respectively, indicating a lower affinity for the uncharacterized site. However, as with CB10, CB12 was able to compete effectively with the intact alpha chain for bindinng to fibronectin. Additionally, both CB10 and CB12 absorbed to tissue culture surfaces were each able to support fibronectin-dependent cell adhesion. Finally, the regions of alpha 2(I) homologous to CB12 and CB10 were found to be active in fibronectin binding, demonstrating the presence of two fibronectin-binding regions in this collagen chain.     

Dysfunctional tendon collagen fibrillogenesis in collagen VI null mice

Tendons are composed of fibroblasts and collagen fibrils. The fibrils are organized uniaxially and grouped together into fibers. Collagen VI is a non-fibrillar collagen expressed in developing and adult tendons. Human collagen VI mutations result in muscular dystrophy, joint hyperlaxity and contractures. The purpose of this study is to determine the functional roles of collagen VI in tendon matrix assembly. During tendon development, collagen VI was expressed throughout the extracellular matrix, but enriched around fibroblasts and their processes. To analyze the functional roles of collagen VI a mouse model with a targeted inactivation of Col6a1 gene was utilized. Ultrastructural analysis of Col6a1−/− versus wild type tendons demonstrated disorganized extracellular micro-domains and associated collagen fibers in the Col6a1−/− tendon. In Col6a1−/− tendons, fibril structure and diameter distribution were abnormal compared to wild type controls. The diameter distributions were shifted significantly toward the smaller diameters in Col6a1−/− tendons compared to controls. An analysis of fibril density (number/μm²) demonstrated a ~ 2.5 fold increase in the Col6a1−/− versus wild type tendons. In addition, the fibril arrangement and structure were aberrant in the peri-cellular regions of Col6a1−/− tendons with frequent very large fibrils and twisted fibrils observed restricted to this region. The biomechanical properties were analyzed in mature tendons. A significant decrease in cross-sectional area was observed. The percent relaxation, maximum load, maximum stress, stiffness and modulus were analyzed and Col6a1−/− tendons demonstrated a significant reduction in maximum load and stiffness compared to wild type tendons. An increase in matrix metalloproteinase activity was suggested in the absence of collagen VI. This suggests alterations in tenocyte expression due to disruption of cell-matrix interactions. The changes in expression may result in alterations in the peri-cellular environment. In addition, the absence of collagen VI may alter the sequestering of regulatory molecules such as leucine rich proteoglycans. These changes would result in dysfunctional regulation of tendon fibrillogenesis indirectly mediated by collagen VI.     

Type XII collagen. A large multidomain molecule with partial homology to type IX collagen

Three overlapping cDNAs encoding alpha 1 (XII) collagen have been isolated and sequenced. The DNAs define five sequence domains within the chain. Three domains are nontriple-helical; two are relatively short triple-helical regions. The amino acid sequences of tryptic peptides derived from 16- and 10-kDa pepsin-resistant fragments isolated from tendon extracts are in full agreement with the deduced sequences of the triple-helical regions. Two of the five sequence domains in alpha 1 (XII), one triple-helical and one nontriple-helical, show a high degree of similarity to regions in type IX collagen chains. In addition, examination of seven exons in the alpha 1 (XII) gene shows that the gene is, in part, similar to the structure of type IX collagen genes. Therefore, collagen types IX and XII are partially homologous. The alpha 1 (XII) sequence data predict an asymmetric structure for type XII collagen molecules, fully consistent with the rotary shadowing images. These images show a triple-helical 75-nm tail attached through a central globule to three finger-like structures, each 60 nm long (Dublet, B., Oh, S., Sugrue, S. P., Gordon, M. K., Gerecke, D. R., Olsen, B. R., and van der Rest, M. (1989) J. Biol. Chem. 264, 13150-13156).     

A sensitive microassay for protein in cells cultured on collagen

A microassay for protein that is linear from 0.1 to 5 μg of protein and does not detect collagen has been developed. The assay is based on the ability of bromosulphophthalein (BSP) to form BSP-protein complexes which precipitate at low pH. Maximum precipitation occurs when 30 or more BSP molecules are bound per albumin molecule. Collagen is not detected because too little BSP binds to this protein to precipitate it. This assay should be of great value to those who grow dispersed cell cultures on a collagen substrate.     

Cyclosporin A reduces skin collagen content in renal graft recipients

Summary Stimulated by the observation of a direct cytopathic effect of cyclosporin A on dermal fibroblasts we have examined total skin collagen content and collagenase activity in three groups of patients. Group 1 (controls) consisted of 16 patients without internal diseases, group 2 of 12 patients with renal transplantation on cyclosporin A therapy and group 3 of six patients with renal transplantation on corticosteroid/azathioprine therapy.Total skin collagen was measured by hydroxyproline/protein determination, collagenase activity according to the principle of Wünsch. SDS page was employed in order to show collagen split products.Mean skin collagen content (expressed by hydroxyproline/protein) was significantly lower in patients on cyclosporin A treatment (42.4 ± 12.2µg/mg) compared to controls (78.6 ± 14.2µg/mg) and patients on corticosteroid/azathioprine therapy (73.7 ± 11.2µg/mg). Mean collagenase activity was significantly higher in patients on cyclosporin A treatment (0.59 ± 0.16 IU) compared to controls (0.21 ± 0.09 IU) and patients on corticosteroid/azathioprine treatment (0.25 ± 0.11 IU). Total skin collagen content and collagenase activity were significantly inversely correlated in patients on cyclosporin A treatment (r = –0.82,p < 0.01,y = –0.011x + 1.053). Patients on cyclosporin A treatment showed remarkable reduction of alpha 1 and alpha 2 collagen chains and significantly prominent split products.The results of our study could be explained either by the activation of collagenase or as a consequence of cyclophilin (peptidyl-prolyl cis-trans-isomerase) inhibition.     

Liquid crystallinity in condensed type I collagen solutions. A clue to the packing of collagen in extracellular matrices.

We recently described a new type of assembly of collagen molecules, forming typical liquid crystalline phases in highly concentrated solutions after sonication. The present work shows that intact 300 nm long collagen molecules also form cholesteric liquid crystalline domains, but the time required is much longer, several weeks instead of several days. Differential calorimetry and X-ray diffraction show that sonication does not alter the triple-helical structure of the collagen fragments. In the viscous solutions, observed between crossed polars in optical microscopy, the textures vary as a function of the concentration. Molecules first align near the air interface at the coverslip edge, then as the concentration increases by slow evaporation of the solvent, the birefringence extends inwards and liquid crystalline domains progressively appear. For concentrations estimated to be above 100 mg/ml, typical textures and defects of cholesteric phases are obtained, at lower concentrations zig-zag extinction patterns and banded patterns are observed; all these textures are described and interpreted. The cholesteric packing of collagen fibrils in various extracellular matrices is known, and the relationship that can be made between the ordered phases obtained with collagen molecules in vitro and the related geometrical structures observed between fibrils in vivo is thoroughly discussed.     

A role for glycosaminoglycans in the development of collagen fibrils

Extensive data on the glycosaminoglycan (GAG) composition and the collagen fibril diameter distribution have been collected for a diverse range of connective tissues. It is shown that tissues with the smallest diameter collagen fibrils (mass-average diameter less than 60 nm) have high concentrations of hyaluronic acid and that tissues with the largest diameter collagen fibrils (mass-average diameter approximately 200 nm) have high concentrations of dermatan sulphate. It is suggested that the lateral growth of fibrils beyond a diameter of about 60 nm is inhibited by the presence of an excess of hyaluronic acid but that this inhibitory effect may be removed by an increasing concentration of chondroitin sulphate and/or dermatan sulphate. It is also postulated that high concentrations of chondroitin sulphate will inhibit fibril growth beyond a mass-average diameter of approximately 150 nm. Such an inhibition may in turn be removed by an increasing concentration of dermatan sulphate such that it becomes the dominant GAG present in the tissue.     

A microassay to quantitate collagen synthesis by cells in culture

A method to quantitate collagen synthesis, total protein synthesis, and DNA in 24-well culture plates is presented. Collagen-producing cells such as human intestinal smooth muscle cells and dermal fibroblasts were pulse-labeled with [3H]proline. After incubation, the plates were heated to 90 degrees C to stop isotope incorporation and sonicated to lyse the cells and an aliquot was removed for DNA quantitation. Carrier protein was added, all protein was precipitated by trichloroacetic acid, and unbound isotope was removed by repeated precipitations. After incubation with purified bacterial collagenase, both the soluble 3H-labeled collagen-derived peptides and the remaining insoluble 3H-labeled noncollagen protein were quantified. Results were expressed as the amount of radioactivity incorporated into collagen and noncollagen protein per nanogram DNA and also as the percentage of collagen synthesis per total protein synthesized. The advantage of this technique over previous attempts to scale down the assay is that the entire assay for DNA, collagen, and non-collagen protein can be carried out in the same well without any transfer of material. This technique also provides a significant savings of culture medium, serum, growth factors, and cell material.