Biology of the Extracellular Matrix

Collagen triple helix

All collagens consist of three polypeptide chains, known as α-chains, which are folded into a triple helix. In some collagens, the α-chains are identical (homotrimers), while others contain two or three different α-chains (heterotrimers). In the collagenous repeat of each polypeptide chain, every third amino acid is glycine (Gly), and the sequence of an α-chain can be expressed as (Gly-X-Y) _n , where X and Y represent other amino acids and n varies according to the length of the repeat. A high number of proline (Pro) and hydroxyproline (Hyp) residues are in the X and Y positions, respectively, and hydrogen bonds between the hydroxyl groups of Hyp contribute to the stability of the helix. The prototype collagen (type I) has an uninterrupted Gly-X-Y repeat sequence that is almost 1000 amino acid residues in length; this forms a rigid, rodlike structure with a diameter of 1.5 nm and length of 300 nm. In some collagens, the (Gly-X-Y) _n repeats are interrupted by one or more amino acids. The interruptions may be numerous and longer than the (Gly-X-Y) _n repeats, and they provide the molecule with flexibility, which is important for the specific functions of a given collagen type .

Biosynthesis of collagens

Collagen biosynthesis involves a number of post-translational modifications ( Fig. 95.2 ). Some collagens are first synthesized as procollagens that have propeptide extensions at their N-terminus, C-terminus, or both. The main intracellular steps in collagen biosynthesis include the following:

• cleavage of signal peptides

• hydroxylation of certain Pro and lysine (Lys) residues to 4-Hyp, 3-Hyp, and hydroxylysine (Hyl)

• glycosylation of some of the Hyl residues to galactosyl-Hyl and glucosylgalactosyl-Hyl

• glycosylation of certain asparagine residues

• association of the α-chains in a specific manner

• formation of intra- and interchain disulfide bonds

• folding of the triple helix.

After the chains have become associated and sufficient Hyp residues have been formed in each chain, a nucleus of the triple helix forms (usually in the C-terminal region) and the triple helix propagates towards the other end of the molecule in a zipper-like fashion . The procollagen molecules are then transported from the endoplasmic reticulum across the Golgi apparatus without leaving the lumen of the Golgi cisternae. During this transport, the molecules begin to aggregate laterally and form early fibrils ready for secretion . Folding and transport of the large, rigid collagens require special machinery . Deficiencies in the proteins that interact with collagens during these events can lead to phenotypes similar to those of collagen deficiency disorders . The extracellular steps in biosynthesis include cleavage of the N- and/or C-terminal propeptides, assembly into suprastructures with other collagens and non-collagenous components, and formation of covalent cross-links.

The specific enzymes involved in the biosynthesis of collagens include prolyl-4-hydroxylase and prolyl-3-hydroxylase, which hydroxylate Pro residues to Hyp, and three lysyl hydroxylases, which hydroxylate Lys residues to Hyl. These enzymes require O ₂ , Fe ²⁺ , α-ketoglutarate and ascorbate as cofactors for the reactions. In the rough endoplasmic reticulum, glycosyltransferases add glucosylgalactosyl disaccharides onto the α-chains. The same intracellular enzymes modify all collagen chains , whereas extracellular processing enzymes have a higher substrate specificity. Procollagen I N-proteinase cleaves the N-propeptide of procollagens I and II. This enzyme is a member of the disintegrin and metalloproteinase (ADAM) proteinase family and is also designated as ADAM with t hrombo s pondin type 1 motif (ADAMTS)-2. Procollagen C-proteinases that cleave the C-propeptide of collagens I, II, III, V, and VII belong to the astacin family and include bone morphogenetic protein-1 (BMP-1) and tolloid-like proteins; meprins in this family can cleave both N- and C-termini of collagens .

Cross-linking between collagen molecules involves the ε-amino groups of Lys and Hyl and is catalyzed by lysyl oxidase, a copper-requiring enzyme ; this process can be regulated by proteoglycans . Another enzyme that catalyzes cross-linking of some collagens is tissue transglutaminase. Collagen VII-containing anchoring fibrils in the skin appear to be transaminated, and collagen VII serves as a substrate for tissue transglutaminase in vitro .

The collagen family of proteins

The length and continuity of the triple helical domains vary among the collagen types. For practical purposes, the collagens have been divided into groups according to their ability to form supramolecular aggregates, which are depicted in Fig. 95.3 . The types and tissue distributions of the collagens in each group are listed in Table 95.2 , and more detailed information is available in reviews .

Collagens of the skin

Collagens account for 75% of the dry weight and 20–30% of the volume of the dermis. Different collagens polymerize into distinct suprastructures and have specific functions in the dermis as well as in epidermal and vascular basement membranes. “Pure” collagen fibrils do not exist; these fibrils are always mixtures of several collagens and other molecules, e.g. proteoglycans . Classic, ultrastructurally recognizable, cross-banded fibrils in the dermis contain collagens I, III, V, XII, and XIV. The characteristic cross-banding ( Fig. 95.4 ) with a periodicity of 64 nm results from precise lateral packing of the different collagens within the fibrils (see Fig. 95.3 ). Collagen I is the major component of the fibrils, and the amount of other collagens varies. For example, during embryonic development and wound repair, the relative content of collagen III increases. Collagen VI, a highly glycosylated and disulfide-bonded collagen, is a component of almost all tissues, including the skin. In vitro , it polymerizes to form beaded filaments (see Fig. 95.3 ), but in vivo in the dermis, the ultrastructure of the collagen VI fibrils is reminiscent of microfibrils.

The collagen IV molecules in different basement membranes contain six genetically distinct but structurally homologous α-chains, which form three major networks – α ₁ /α ₂ , α ₃ /α ₄ /α ₅ , and α ₁ /α ₂ /α ₅ /α ₆ . The chain composition is determined by the carboxy-terminal non-collagenous (NC1) domains; covalent interactions between these domains link α-chains to each other. In the skin, the α ₁ /α ₂ -containing collagen IV network dominates within the dermal–epidermal junction (see Ch. 28 ), but the α ₁ /α ₂ /α ₅ /α ₆ -containing network is also thought to be present .

Two collagens are essential for the cohesion of the epidermis and dermis (see Ch. 28 ). Collagen VII is the major, if not sole, component of the anchoring fibrils that attach the basement membrane to the dermal ECM. Collagen XVII (also known as bullous pemphigoid antigen 2) is a component of the anchoring filaments that bind the basal keratinocytes to the lamina densa of the basement membrane. It is a transmembrane collagen in type II orientation with a long extracellular C-terminal region containing a multiply-interrupted triple helix referred to as collagenous domains 1–15 (see Fig. 31.9 ). The ectodomains can be shed from the cell surface by transmembrane proteases , a process important for regulation of cell adhesion and migration. Basal keratinocytes also express other transmembrane collagens such as type XIII, which is a component of focal contacts, and type XXIII.

The vascular basement membranes in the skin contain yet other collagens – namely, collagens VIII and XVIII. Collagen VIII builds hexagonal networks below the endothelial basement membranes (see Fig. 95.3 ) and, thus, structurally strengthens the vascular wall. Collagen XVIII is localized at the dermal side of vascular basement membranes. It is present in both epidermal and vascular basement membranes. Its C-terminal fragment, endostatin, is proteolytically released from the collagen molecule and has independent antiangiogenic and anti-scarring activities .

Most collagens in the skin are products of dermal fibroblasts. Exceptions include: (1) collagen XVII, a surface component of epidermal keratinocytes; (2) collagen VII, which can be synthesized by both keratinocytes and fibroblasts; and (3) collagens VIII and XVIII, which can also be produced by endothelial cells. Several genetic and acquired diseases are associated with abnormalities of skin collagens or the enzymes that process them (see Tables 95.5 & 95.6 and below).

Elastin

Elastin is a highly cross-linked protein . Its monomer, tropoelastin, exists in several tissue-specific splice variants. Ala- and Lys-rich repeats form critical cross-linking domains within a background of hydrophobic amino acid repeats. Lysyl oxidase, the same copper-dependent enzyme that catalyzes collagen cross-linking (see above), catalyzes the formation of desmosine cross-links between elastin molecules, which account for the elasticity and insolubility of the elastic fibers. Elastin makes up ~90% of mature elastic fibers. During fiber assembly, tropoelastin both self-assembles and interacts with fibulins 4 and 5, fibrillin 1, and microfibril-associated glycoprotein-1 (MAGP-1). A model for this spatially and temporally regulated process is shown in Fig. 95.5 .

Microfibrils

Microfibrils (10 to 12 nm in diameter) insert into the basement membrane in a perpendicular orientation (classically referred to as “oxytalan fibers”) and extend into the papillary dermis, where they gradually merge with the elastic fibers that form a plexus parallel to the dermal–epidermal junction (classically referred to as “elaunin fibers”) ( Fig. 95.6 ). These fibers appear to be continuous with the elastic fibers within the reticular dermis, which contain more abundant elastin. Microfibrils consist mainly of fibrillins but also contain and associate with other proteins, such as MAGPs, latent TGF-β binding proteins (LTBPs), fibulins, elastin microfibril interface (EMILIN)-1, and (in the papillary dermis) collagen XVI . Fibrillins 1–3 are large glycoproteins that contain both epidermal growth factor (EGF) and cysteine (Cys) repeats. The EGF repeats bind calcium, which is required for protein stabilization. Fibrillin 1 interacts with perlecan in the epidermal basement membrane to attach microfibrils to the dermal–epidermal junction. Fibrillin interactions with heparan sulfate proteoglycans and integrin receptors also contribute to microfibril assembly .

An important paradigm has emerged from investigations demonstrating that, in addition to guiding elastic fiber assembly and providing structural support, fibrillins function together with LTBPs to regulate cell signaling . The four members of the LTBP family are high-molecular-weight glycoproteins characterized by EGF-like and 8-Cys repeats, and they all associate with the small latent complex of TGF-β (see below; Fig. 95.7 ). The LTBP-1/latent TGF-β complex (referred to as the large latent complex) binds to fibrillin-containing microfibrils. Together, they control bioavailability of TGF-β in the skin and other tissues. LTBPs contain multiple proteinase-sensitive sites, providing the means to solubilize the large latent complex from ECM structures, and both soluble and ECM-associated forms are known to exist. Another group of signaling molecules that can bind to fibrillins are the bone morphogenetic proteins (BMPs) . These are secreted as cross-linked dimeric complexes that are associated through their prodomains, which interact directly with the N-terminus of fibrillin. Fibrillins and LTBPs therefore have important roles in targeting latent, activatable TGF-β and BMPs into different ECMs. Thus, the microfibrils serve not only as a force-bearing element and scaffolding for elastin deposition in the dermis, but also as an important repository for signaling molecules in the skin .