Tip-mediated fusion involving unipolar collagen fibrils accounts for rapid fibril elongation, the occurrence of fibrillar branched networks in skin and the paucity of collagen fibril ends in vertebrates.

Collagen fibrils are the principal source of mechanical strength of connective tissues such as tendon, skin, cornea, cartilage and bone. The ability of these tissues to withstand tensile forces is directly attributable to the length and diameter of the fibrils, and to interactions between individual fibrils. Although electron microscopy studies have provided information on fibril diameters, little is known about the length of fibrils in tissue and how fibrils interact with each other. The question of fibril length has been difficult to address because fibril ends are rarely observed in cross-sections of tissue. The paucity of fibril ends, or tips, has led to controversy about how long individual fibrils might be and how the fibrils grow in length and diameter. This review describes recent discoveries that are relevant to these questions. We now know that vertebrate collagen fibrils are synthesised as short (1-3 microm) early fibrils that fuse end-to-end in young tissues to generate very long fibrils. The diameter of the final fibril is determined by the diameter of the collagen early fibrils. During a late stage of tissue assembly fibril tips fuse to fibril shafts to generate branched networks. Of direct relevance to fibril fusion is the fact that collagen fibrils can be unipolar or bipolar, depending on the orientation of collagen molecules in the fibril. Fusion relies on: (1) specific molecular interactions at the carboxyl terminal ends of unipolar collagen fibrils; and (2) the insulator function of small proteoglycans to shield the surfaces of fibrils from inappropriate fusion reactions. The fusion of tips to shafts to produce branched networks of collagen fibrils is an elegant mechanism to increase the mechanical strength of tissues and provides an explanation for the paucity of fibril tips in older tissue.     

Control of collagen fibril diameters in tissues.

It is proposed that radial growth of collagen fibrils, which takes place in all connective tissues to varying extents, according to the tensile stresses exerted on them, proceeds mainly by aggregation of protofibrils (approximately 10 nm) and existing fibrils. In young tissues, fibrils are prevented from making frequent intimate contacts which would lead to aggregation by abundant interfibrillar proteoglycan, that keeps the fibrils apart. Collagen fibrils are probably unable to fuse except when the molecules within them are packed in the same sense, i.e. fusing fibrils are parallel. The roughly equal numbers of parallel and antiparallel fibrils seen in several tissues must limit radial fibril growth in older tissues, where proteoglycan is usually less abundant. Possible origins of the balance of fibril polarities, which must be conserved after fibril nucleation on cell or non-cell templates, are analysed. The two controlling factors, ambient proteoglycan and fibril polarity, working against the tendency of fibrils to fuse, account for many features of the observed distributions of collagen fibril diameters in diverse tissues and at different ages.     

Glycosaminoglycans show a specific periodic interaction with type I collagen fibrils

Current wisdom on intermolecular interactions in the extracellular matrix assumes that small proteoglycans bind collagen fibrils on highly specific sites via their protein core, while their carbohydrate chains interact with each other in the interfibrillar space. The present study used high-resolution scanning electron microscopy to analyse the interaction of two small leucine-rich proteoglycans and several glycosaminoglycan chains with type I collagen fibrils obtained in vitro in a controlled, cell-free environment. Our results show that most ligands directly influence the collagen fibril size and shape, and their aggregation into thicker bundles. All chondroitin sulphate/dermatan sulphate glycosaminoglycans we tested, except chondroitin 4-sulphate, bound to the fibril surface in a highly specific way and, even in the absence of any protein core, formed regular, periodic interfibrillar links resembling those of the intact proteoglycan. Only intact decorin, however, was able to organize collagen fibrils into fibres compact enough to mimic in vitro the superfibrillar organization of natural tissues. Our data indicate that multiple interaction patterns may exist in vivo, may explain why decorin- or biglycan-knockout organisms show milder effects than can be expected, and may lead to the development of better, simpler engineered biomaterials.     

Proteoglycan: collagen interactions in connective tissues. Ultrastructural, biochemical, functional and evolutionary aspects

Electron histochemical investigations of mammalian and echinoderm tissues, using cupromeronic blue to stain proteoglycans (PGs) specifically in critical electrolyte concentration methods, showed that collagen fibrils are associated with keratan sulphate and chondroitin (dermatan) sulphate ('tadpole') PGs at the a, c, d and e bands on the fibril surface, giving rise to the 'one proteoglycan: one binding site' hypothesis. Intra-fibrillar PGs have been observed, distributed in a regular way which suggests that collagen fibrils are aggregates of 'protofibrils', some of which carry PGs at their surfaces. A scheme for remodelling of collagen fibrils, based on recycling of these protofibrils, is outlined. The choice of which tadpole PG to use to carry out a given function is decided to a considerable extent by the availability of oxygen to the relevant tissue element.     

Amianthoid (asbestoid) transformation: electron microscopical studies on aging human costal cartilage

The present study reports on the fine structure of human costal cartilage at different ages in order to obtain information on the morphogenesis of amianthoid fibers. Our results reveal an overall increase of collagen fibril diameter with increasing age, even in areas with no signs of amianthoid transformation. Ultrastructural evidence is presented that this increase in diameter is due to a gathering of the preexisting collagen fibrils. The age-related change in collagen fibril diameter is paralleled by changes in the composition and ultrastructural appearance of cartilage proteoglycans (as revealed by acridine orange staining). Acridine-orange-positive filaments indicative for proteoglycans are markedly reduced in size with advancing age in centrally located regions of costal cartilage. Treatment with testicular hyaluronidase previous to acridine-orange staining leaves these small proteoglycan filaments unaffected. By contrast, the filaments visible after acridine-orange staining in the extracellular matrix near to the perichondrium are susceptible to hyaluronidase treatment. Infrequently, a sharp increase in collagen fibril diameter can be observed in territorial matrix areas of degenerating chondrocytes. This observation is conspicuous at ages of 10 and 20 years. Amianthoid transformation is characterized by the appearance of collagen fibrils strictly arranged in parallel. These amianthoid fibers are embedded in a matrix rich in small acridine-orange-positive filaments similar to the proteoglycan filaments observed in centrally located matrix regions. It can be concluded that extensive remodelling not only of the collagen fibrils but also of the cartilage proteoglycans is involved in the development of amianthoid transformation.     

beta-D xyloside alters dermatan sulfate proteoglycan synthesis and the organization of the developing avian corneal stroma.

Corneal transparency is dependent upon the development of an organized extracellular matrix containing small diameter collagen fibrils with regular spacing, organized as orthogonal lamellae. Proteoglycan-collagen interactions have been implicated in the regulation of collagen fibrillogenesis and matrix assembly. To determine the role of dermatan sulfate proteoglycan in the development and organization of the secondary corneal stroma, its synthesis was disrupted using beta-D xyloside. The secondary corneal stroma contains two different proteoglycans, dermatan sulfate and keratan sulfate proteoglycan. beta-D xyloside interferes with xylose-mediated O-linked proteoglycan synthesis, and thus disrupts dermatan sulfate proteoglycan synthesis. Corneal keratan sulfate proteoglycan, a mannose-mediated N-linked proteoglycan, should not be altered. Biochemical analysis of corneas treated both in vitro and in ovo revealed a reduced synthesis of normally glycosylated dermatan sulfate proteoglycans and an increased synthesis of free xyloside-dermatan sulfate glycosaminoglycans. Keratan sulfate proteoglycan synthesis was unaltered in both cases. Corneal stromas were studied using histochemistry and electron microscopy after in ovo treatment with beta-D xyloside. The observed biochemical alterations in dermatan sulfate proteoglycans translated into disruptions in the organization of beta-D xyloside-treated stromas. There was a reduction in the histochemical staining of proteoglycans, but no alteration in collagen fibril diameter. In addition, focal alterations in collagen fibril packing, and a disruption of lamellar organization were observed in beta-D xyloside-treated corneas. These data suggest that dermatan sulfate proteoglycans are not involved in the regulation of corneal collagen fibril diameter, but are important in the fibril-fibril spacing as well as in lamellar organization, and cohesiveness.     

Collagen fibril morphology and organization: implications for force transmission in ligament and tendon.

Connective tissue mechanical behavior is primarily determined by the composition and organization of collagen. In ligaments and tendons, type I collagen is the principal structural element of the extracellular matrix, which acts to transmit force between bones or bone and muscle, respectively. Therefore, characterization of collagen fibril morphology and organization in fetal and skeletally mature animals is essential to understanding how tissues develop and obtain their mechanical attributes. In this study, tendons and ligaments from fetal rat, bovine, and feline, and mature rat were examined with scanning electron microscopy. At early fetal developmental stages, collagen fibrils show fibril overlap and interweaving, apparent fibril ends, and numerous bifurcating/fusing fibrils. Late in fetal development, collagen fibril ends are still present and fibril bundles (fibers) are clearly visible. Examination of collagen fibrils from skeletally mature tissues, reveals highly organized regions but still include fibril interweaving, and regions that are more randomly organized. Fibril bifurcations/fusions are still present in mature tissues but are less numerous than in fetal tissue. To address the continuity of fibrils in mature tissues, fibrils were examined in individual micrographs and consecutive overlaid micrographs. Extensive microscopic analysis of mature tendons and ligaments detected no fibril ends. These data strongly suggest that fibrils in mature ligament and tendon are either continuous or functionally continuous. Based upon this information and published data, we conclude that force within these tissues is directly transferred through collagen fibrils and not through an interfibrillar coupling, such as a proteoglycan bridge.     

Heparan sulfate proteoglycans from mouse mammary epithelial cells. Basal extracellular proteoglycan binds specifically to native type I collagen fibrils

Mouse mammary epithelial cells (NMuMG cells) deposit at their basal surfaces an extracellular heparan sulfate-rich proteoglycan that binds to type I collagen. The binding of the purified proteoglycan to collagen was studied by (i) a solid phase assay, (ii) a suspension assay using preformed collagen fibrils, and (iii) a collagen fibril affinity column. The binding interaction occurs at physiological pH and ionic strength and can be inhibited only by salt concentrations that greatly exceed those found physiologically. Binding requires the intact proteoglycan since the protein-free glycosaminoglycan chains will not bind under the conditions of these assays. However, binding is mediated through the heparan sulfate chains as it can be inhibited by block-sulfated polysaccharides, including heparin. Binding requires native collagen structure which may be optimal when the collagen is in a fibrillar configuration. Binding sites on collagen fibrils are saturable, high affinity (Kd approximately 10(-10) M), and selective for heparin-like glycosaminoglycans. Because a culture substratum of type I collagen fibrils causes NMuMG cells to accumulate heparan sulfate proteoglycan into a basal lamina-like layer, binding of heparan sulfate proteoglycans to type I collagen may lead to the formation of a basal lamina and may link the basal lamina to the connective tissue matrix, an association found in basement membranes.     

Ultrastructural localization of proteoglycans in tissue using Cuprolinic Blue according to the critical electrolyte concentration method: Comparison with biochemical data from the literature

Summary Several connective tissues were stained for proteoglycans using the cationic dye Cuprolinic Blue according to the critical electrolyte concentration method. With this method, proteoglycans are visualized as electron-dense filaments. In most tissues, two types of proteoglycan filaments are present: a small (maximum length 60 nm), thin, collagen fibril-associated filament, and a thick, heavily-staining filament which is predominantly localized between bundles of collagen fibrils. Cartilage contains very large (about 300 nm) proteoglycan filaments while in cornea they are very small. Comparison with biochemical data from the literature suggests that the appearance of the proteoglycan filaments may be indicative for the glycosaminoglycan—protein ratio and for the molecular weight of the part of the protein core to which glycosaminoglycans are attached. The data thus obtained on the localization and structure of a proteoglycan may be useful when planning a strategy for its isolation.     

Proteoglycan and collagen morphology in superficially scarred rabbit cornea

Summary We have examined the changes in collagen and proteoglycan morphology in superficial lamellar keratectomy wounds produced in rabbit corneas. The ultrastructural location within the tisse of keratan sulphate and chondroitin sulphate proteoglycans was demonstrated using the cationic dye Cuprolinic Blue under critical electrolyte conditions. Large proteoglycan filaments (up to 500 nm long) appeared in the early stages of wound healing; these were most common after two weeks' wound healing, after which they decreased both in number and size. At these early stages of scar formation, spaces containing proteoglycans were present amongst bundles of collagen fibrils. As proteoglycans play an important role in controlling corneal hydration, the presence of the large proteoglycan-filled spaces would result in an abnormally high water content which is found in early early scar tissue.     

Echinoderm collagen fibrils grow by surface-nucleation-and-propagation from both centers and ends

Collagen fibrils from sea cucumber (class Holothuroidea) dermis were previously found to grow by coordinated monomer addition at both centers and ends. This analysis of sea urchin (class Echinoidea) collagen fibrils was undertaken to compare the growth characteristics of fibrils from two classes of echinoderms, and to determine whether a single growth model could account for the main features of fibrils from these two taxa. Native collagen fibrils (37-431 micrometer long) from the spine ligaments of the sea urchin Eucidaris tribuloides were studied by scanning transmission electron microscopy and image analysis. The analyses revealed the mass per unit length, and hence the number of molecules in cross-section, along the entire length of each fibril. The fibrils were symmetrically spindle shaped. The maximum mass per unit length occurred in the center of each fibril, where the fibril contains anti-parallel molecules in equal numbers. The two pointed tips of each fibril showed similar linear axial mass distributions, indicating that the two tips retain shape and size similarity throughout growth. The linear axial mass distributions showed that the tips were paraboloidal, similar to those of vertebrate and sea cucumber fibrils. The computed maximum diameters of the fibrils increased linearly with fibril length. The overall shapes of the fibrils showed that they retain geometric similarity throughout growth. Computer modeling showed that the simplest self-assembly mechanism that can account for the features of these fibrils, and of the sea cucumber fibrils that have been described, is one in which the fibril tips produce independent axial growth, while lateral growth takes place through a surface nucleation and propagation mechanism. This mechanism produces coordinated growth in length and diameter as well as geometric similarity, characteristic features of echinoderm collagen fibrils.     

Proteoglycan changes during restoration of transparency in corneal scars

Corneal scars generated in rabbits by penetrating wounds are initially opaque but become transparent within a year. Previous studies have shown that the corneal stroma consists of proteoglycans and collagen fibrils spaced at regular intervals and that the interfibrillar spaces, the presumed location of proteoglycans, are abnormally large in opaque scars. In the present study, the size and glycosaminoglycan composition of the corneal stromal proteoglycans were determined in corneal scars during the restoration of transparency. The results showed that initially opaque scars which contained the large interfibrillar spaces also contained unusually large chondroitin sulfate proteoglycans with glycosaminoglycan side chains of normal size. These opaque scars also lacked the keratan sulfate proteoglycan but did contain hyaluronic acid. In the 1-year-old scars there was a restoration of normal interfibrillar spacing, and a return to corneal stromal proteoglycans of normal size and composition. These correlations suggest that the corneal stromal proteoglycans may play a fundamental role in regulating corneal collagen fibril spacing.     

Proteoglycan-collagen associations in the non-lactating human breast connective tissue during the menstrual cycle

The human mammary gland undergoes a sequence of histological changes in both epithelial and stromal compartments during the menstrual cycle. Swelling and unswelling of the breast stromal tissue is a characteristic feature of the two phases of the cycle and is mediated by changes in the water content of sulfated proteoglycans in the matrix between the fibrils. In an ultrastructural study we investigated the distribution of sulfated proteoglycans identified as cupromeronic blue-positive needle-like structures and measured the distance between the dermatan sulfate-proteoglycan attachment sites at the d-bands of the collagen fibrils in the loose intralobular connective tissue and in the dense interlobular connective tissue. We characterized the dermatan sulfate proteoglycan by enzyme digestion and by immunogold-labeled antibody. In the follicular phase a relatively constant distance of 46 nm between neighboring proteoglycan attachment sites was found, while in the luteal phase the measured distances are strikingly variable and exceed the follicular value by up to 9 nm. This difference of the two cycle phases is more evident in the loose than in the dense connective tissue. Possibly the changes of the fibril-attached proteoglycans in the luteal phase reflect an influence of the higher water content of the matrix leading to a probably torsional swelling of the collagen fibril.     

Collagen fibril assembly by corneal fibroblasts in three-dimensional collagen gel cultures: small-diameter heterotypic fibrils are deposited in the absence of keratan sulfate proteoglycan.

Extracellular matrix assembly is a multistep process and the various steps in collagen fibrillogenesis are thought to be influenced by a number of factors, including other noncollagenous matrix molecules. The synthesis and deposition of extracellular matrix by corneal fibroblasts grown within three-dimensional collagen gel cultures were examined to elucidate the factors important in the establishment of tissue-specific matrix architecture. Corneal fibroblasts in collagen gel cultures form layers and deposit small-diameter collagen fibrils (approximately 25 nm) typical of the mature corneal stroma. The matrix synthesized contains type VI collagen in a filamentous network and type I and type V collagen assembled as heterotypic fibrils. The amount of type V collagen synthesized is relatively high and comparable to that seen in the corneal stroma. This matrix is deposited between cell layers in a manner reminiscent of the secondary corneal stroma, but is not deposited as densely or as organized as would be found in situ. No keratan sulfate proteoglycan, a proteoglycan found only in the corneal stroma, was synthesized by the fibroblasts in the collagen gel cultures. The assembly and deposition of small-diameter fibrils with a collagen composition and structure identical to that seen in the corneal stroma in the absence of proteoglycans typical of the secondary corneal stroma imply that although proteoglycan-collagen interactions may function in the establishment of interfibrillar spacing and lamellar organization, collagen-collagen interactions are the major parameter in the regulation of fibril diameter.     

The precision of lateral size control in the assembly of corneal collagen fibrils

Collagen fibrils in the corneal stroma have been recognised to have a high degree of uniformity of diameter and spatial arrangement compared with those in other mature connective tissues. The precision of this lateral size control has been determined in this study by mass per unit length measurements on fibrils isolated from adult bovine corneal stroma. At the molecular level, however, there are substantial variations in lateral size, both between fibrils and along individual fibrils. The mean mass per unit length was measured to be 304 kDa nm(-1), equivalent to 347 collagen molecules in transverse section and had a standard deviation of 8.3%. The variation of lateral size along individual fibrils was measured as a mass slope over approximately 7 microm lengths (100 D-periods) and had a mean mass slope equivalent to 0.56 molecules per D-period. Smoothly tapered tips of length approximately 7 microm were also observed with a mass slope of about approximately three molecules per D-period. The frequency of these tips was used to estimate a mean fibril length of approximately 940 microm in the sample tissue. Observations of molecular polarity within the fibril shafts and tips were used to consider possible models of fibril assembly.     

Type V collagen controls the initiation of collagen fibril assembly

Vertebrate collagen fibrils are heterotypically composed of a quantitatively major and minor fibril collagen. In non-cartilaginous tissues, type I collagen accounts for the majority of the collagen mass, and collagen type V, the functions of which are poorly understood, is a minor component. Type V collagen has been implicated in the regulation of fibril diameter, and we reported recently preliminary evidence that type V collagen is required for collagen fibril nucleation (Wenstrup, R. J., Florer, J. B., Cole, W. G., Willing, M. C., and Birk, D. E. (2004) J. Cell. Biochem. 92, 113-124). The purpose of this study was to define the roles of type V collagen in the regulation of collagen fibrillogenesis and matrix assembly. Mouse embryos completely deficient in pro-alpha1(V) chains were created by homologous recombination. The col5a1-/- animals die in early embryogenesis, at approximately embryonic day 10. The type V collagen-deficient mice demonstrate a virtual lack of collagen fibril formation. In contrast, the col5a1+/- animals are viable. The reduced type V collagen content is associated with a 50% reduction in fibril number and dermal collagen content. In addition, relatively normal, cylindrical fibrils are assembled with a second population of large, structurally abnormal collagen fibrils. The structural properties of the abnormal matrix are decreased relative to the wild type control animals. These data indicate a central role for the evolutionary, ancient type V collagen in the regulation of fibrillogenesis. The complete dependence of fibril formation on type V collagen is indicative of the critical role of the latter in early fibril initiation. In addition, this fibril collagen is important in the determination of fibril structure and matrix organization.     

Tendon proteoglycans: biochemistry and function

Tendon remodeling occurs in response to changes in loading and mobilization. Though the normal mechanical function depends on precise alignment of collagen fibrils, it is proteoglycans that regulate collagen fibrillogenesis and thus, indirectly, tendon function. In this paper we discuss the basic biochemical structure of several members of two proteoglycans families. Decorin, biglycan, fibromodulin and lumican, all members of the small leucine-rich proteoglycans family, bind to collagen fibrils and are active participants in fibrillogenesis. Aggrecan and versican, two members of large modular proteoglycans or lecticans, and their partner hyaluronan likely provide tendon tissues with a high capacity to resist high compressive and tensile forces associated with loading and mobilization. We present data from our laboratory showing that proteoglycans and glycosaminoglycan content increases not only with growth but also with loading of young avian gastrocnemius tendons. Specifically, an increase in the content of keratan sulfate, chondroitin sulfate and hyaluronan was observed. Moderate exercise for several weeks led not only to a further increase in total proteoglycans content but also to qualitative changes in proteoglycan make up.     

Two type XII-like collagens localize to the surface of banded collagen fibrils

Two recently identified collagen molecules, termed twelve-like A and twelve-like B (TL-A and TL-B) have properties similar to type XII collagen. These molecules have been localized in human and calf tissues by immunoelectron microscopy. The observations strongly suggest that both molecules are located along the surface of banded collagen fibers. The epitopes recognized by the antibodies are contained in large, nontriple-helical domains at one end of the collagen helix. The epitopes are visualized at a distance from the surface of the banded fibers roughly equal to the length of the nonhelical domains, suggesting that the nonhelical domains extend from the fibril, while the triple-helical domains are likely to bind directly to the fibril surface. Occasionally, both TL-A and TL-B demonstrate periodic distribution along the fibril surface. The period corresponds to the primary interband distance of the banded fibrils. Not all fibrils in a fiber bundle are labeled, nor is the labeling continuous along the length of labeled fibrils. Simultaneous labeling of TL-A and type VI collagen only rarely shows colocalization, suggesting that TL-A and TL-B do not mediate interactions between the type VI collagen beaded filaments and banded collagen fibrils. Also, interfibrillar distances are approximately equivalent in the presence and absence of these type XII-like molecules. While the results do not directly indicate a specific function for these molecules, the localization at the fibril surface suggests that they mediate interactions between the fibrils and other matrix macromolecules or with cells.     

Conformation and sequence evidence for two-fold symmetry in left-handed beta-helix fold

The glycosaminoglycan (GAG) side-chains of small leucine-rich proteoglycans have been postulated to mechanically cross-link adjacent collagen fibrils and contribute to tendon mechanics. Enzymatic depletion of tendon GAGs (chondroitin and dermatan sulfate) has emerged as a preferred method to experimentally assess this role. However, GAG removal is typically incomplete and the possibility remains that extant GAGs may remain mechanically functional. The current study specifically investigated the potential mechanical effect of the remaining GAGs after partial enzymatic digestion.A three-dimensional finite element model of tendon was created based upon the concept of proteoglycan mediated inter-fibril load sharing. Approximately 250 interacting, discontinuous collagen fibrils were modeled as having a length of 400 μm, being composed of rod elements of length 67 nm and E-modulus 1 GPa connected in series. Spatial distribution and diameters of these idealized fibrils were derived from a representative cross-sectional electron micrograph of tendon. Rod element lengths corresponded to the collagen fibril D-Period, widely accepted to act as a binding site for decorin and biglycan, the most abundant proteoglycans in tendon. Each element node was connected to nodes of any neighboring fibrils within a radius of 100 nm, the slack length of unstretched chondroitin sulfate. These GAG cross-links were the sole mechanism for lateral load sharing among the discontinuous fibrils, and were modeled as bilinear spring elements. Simulation of tensile testing of tendon with complete cross-linking closely reproduced corresponding experiments on rat tail tendons. Random reduction of 80% of GAG cross-links (matched to a conservative estimate of enzymatic depletion efficacy) predicted a drop of 14% in tendon modulus. Corresponding mechanical properties derived from experiments on rat tail tendons treated in buffer with and without chondroitinase ABC were apparently unaffected, regardless of GAG depletion. Further tests for equivalence, conservatively based on effect size limits predicted by the model, confirmed equivalent stiffness between enzymatically depleted tendons and their native controls.Although the model predicts that relatively small quantities of GAGs acting as primary collagen cross-linking elements could provide mechanical integrity to the tendon, partial enzymatic depletion of GAGs should result in mechanical changes that are not reflected in analogous experimental testing. We thus conclude that GAG side chains of small leucine-rich proteoglycans are not a primary determinant of tensile mechanical behavior in mature rat tail tendons.     

Fibromodulin-null mice have abnormal collagen fibrils, tissue organization, and altered lumican deposition in tendon

Fibromodulin is a member of a family of connective tissue glycoproteins/proteoglycans containing leucine-rich repeat motifs. Several members of this gene family bind to fibrillar collagens and are believed to function in the assembly of the collagen network in connective tissues. Here we show that mice lacking a functional fibromodulin gene exhibit an altered morphological phenotype in tail tendon with fewer and abnormal collagen fiber bundles. In fibromodulin-null animals virtually all collagen fiber bundles are disorganized and have an abnormal morphology. Also 10-20% of the bundles in heterozygous mice are similar to the abnormal bundles in fibromodulin-null tail tendon. Ultrastructural analysis of Achilles tendon from fibromodulin-null mice show collagen fibrils with irregular and rough outlines in cross-section. Morphometric analysis show that fibromodulin-null mice have on the average thinner fibrils than wild type animals as a result of a larger preponderance of very thin fibrils in an overall similar range of fibril diameters. Protein and RNA analyses show an approximately 4-fold increase in the content of lumican in fibromodulin-null as compared with wild type tail tendon, despite a decrease in lumican mRNA. These results demonstrate a role for fibromodulin in collagen fibrillogenesis and suggest that the orchestrated action of several leucine-rich repeat glycoproteins/proteoglycans influence the architecture of collagen matrices.