Review The CCN family of proteins: structure– function relationships Kenneth P. Holbourn1, K. Ravi Acharya1 and Bernard Perbal2,3 1 Department of Biology and Biochemistry, University of Bath, Claverton Down, Bath BA2 7AY, UK 2 Department of Dermatology, University of Michigan, Ann Arbor, MI 48109, USA 3 Present address: Research and Development, L’Ore´ al USA, 111 Terminal Avenue, Clark, NJ 07066, USA The CCN proteins are key signalling and regulatory module (CT) (Figure 1). Apart from CCN5, which lacks a molecules involved in many vital biological functions, CT module, all CCN proteins exhibit the same type of including cell proliferation, angiogenesis, tumourigen- organization and share a closely related primary structure, esis and wound healing. How these proteins influence which includes a series of 38 cysteine residues that are such a range of functions remains incompletely under- strictly conserved in position and number. However, CCN6 stood but is probably related to their discrete modular lacks four cysteine residues in domain II (VWC). nature and a complex array of intra- and inter-molecular The 38 cysteine residues that spread across the four interactions with a variety of regulatory proteins and modules represent almost 10% of the CCN molecule by ligands. Although certain aspects of their biology can be mass. A short sequence that varies greatly in both compo- attributed to the four individual modules that constitute sition and length among the CCN family members is the CCN proteins, recent results suggest that some of located directly after the VWC domain. It was proposed their biological functions require cooperation between to act as a hinge between the first and second half of the modules. Indeed, the modular structure of CCN proteins molecule [1,3]. Recent data have established that each of provides important insight into their structure–function the linker regions that separate the four CCN modules is relationships. susceptible to proteolysis [6]. Cleavage at these sites might be responsible for the production of truncated molecules The CCN family of proteins and individual modules [4] that display distinct biological The CCN family of proteins is a complex family of multi- properties. Moreover, these events might constitute an functional proteins containing six members designated additional process to regulate the biological activity of CCN1 to CCN6. The CCN acronym was introduced from the CCN proteins [5,7]. the names of the first three members of the family to be The six CCN family members share 30–50% identity discovered: Cyr61 (cysteine-rich protein 61), CTGF (con- in their primary structure (40–60% similarity) [6,8]. nective tissue growth factor) and NOV (nephroblastoma overexpressed gene) [1]. Owing to the multifunctional nature of these proteins, they have been identified in Glossary several other biological studies and have been assigned ADAMTS: a disintegrin and metalloproteinase that harbours a thrombospon- multiple names; however, in this review we will use the din motif. official nomenclature [2] (Table 1). (See also http://ccnso- CCN: the CCN family of proteins is a complex family of multifunctional proteins ciety.com). This family of secreted extracellular matrix with six members designated CCN1 to CCN6. The CCN acronym was introduced in reference to the names of the first three members of the family (ECM)-associated proteins is involved in a wide range of to be discovered: Cyr61 (cysteine-rich protein 61), CTGF (connective tissue important functional pathways, including adhesion, mito- growth factor) and NOV (nephroblastoma overexpressed gene). genesis, migration and chemotaxis, cell survival, differen- CT: C-terminal cysteine knot domain. This domain has a conserved pattern of cysteines consistent with a six membered cysteine knot similar to several tiation, angiogenesis, chondrogenesis, tumourigenesis and members of the TGF-b growth factor family. wound healing. They have also been implicated in many HSPGs: heparin sulphated proteoglycans – multifunctional molecules involved in cell and protein adhesion and fine tuning regulatory proteins. human diseases (for some comprehensive reviews, see IGFBP: the insulin-like growth factor binding protein (IGFBP) family comprises Refs [3–5]). six IGFBPs (1–6) that bind insulin-like growth factors (IGFs) with high affinity The unifying feature of this family of proteins is their (KD 0.1 nM). IGFBPs exert their biological effects by modulating IGF behaviour. mosaic structure, consisting of discrete modules that share SMAD: small mothers against decapentaplegic – a family of transcription identity with functional domains of other regulatory mol- factors that can modulate TGF-b behaviour. ecules. A prototypical CCN protein contains an N-terminal TSP: the thrombospondin repeat domain is a 55 residue consensus sequence named after the extracellular matrix glycoprotein thrombospondin-1 (TSP-1) secretory signal peptide and four functional domains: (i) an that contains repeats of three distinct domains within a primarily linear insulin-like growth factor binding protein-like module structure. The TSP-1 domain binds a wide range of extracellular targets and (IGFBP); (ii) a von Willebrand factor type C repeat important signalling molecules including collagen V, fibronectin, TGF-b and heparin. module (VWC); (iii) a thrombospondin type-1 repeat VWC: von Willebrand factor C repeat domains, also referred to as chordin-like module (TSP-1); and (iv) a cysteine-knot-containing cysteine-rich (CR) repeats, are an extremely common motif found in >500 extracellular matrix proteins. This repeat typically comprises 70–100 amino acids with ten conserved cysteine residues and a pair of cysteine-containing Corresponding authors: Acharya, K.R. ([email protected]); Perbal, B. motifs located in the sequence. ([email protected]). 0968-0004/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2008.07.006 Available online 11 September 2008 461 Review Trends in Biochemical Sciences Vol.33 No.10 Table 1. Nomenclature of the CCN family of proteinsa adhesion of normal melanocytes to collagen type IV and CCN Family member Alternative names controls discoidin domain receptor 1 (DDR1)-mediated CCN1 Cyr61, CTGF-2, IGFBP10, IGFBP-rP4 three-dimensional localization of melanocytes in skin [13]. CCN2 CTGF, IGFBP8, IGFBP-rP2, HBGF-0.8 ecogenin CCN3 NOV, NOVH, IGFBP9, IGFBP-rP3 The ability of CCN proteins to promote adhesion CCN4 Wisp-1, Elm-1 involves cell survival because interactions with several CCN5 Wisp-2, CTGF-L, CTGF-3, HICP, Cop-1 surface integrins and transforming growth factor b CCN6 Wisp-3 (TGF-b) can prevent cell apoptosis [6,14]. In addition to aAbbreviations: Cyr61, cysteine rich 61; CTGF-2, connective tissue growth factor 2; interacting directly with ECM components, including fibro- IGFBP-rP2, IGFBP-related protein 2; NOV, nephroblastoma overexpressed gene; HBGF-0.8, heparin-binding growth factor 0.8; Wisp, Wnt-inducible secreted protein; nectin and heparin [6,12], CCN1 and CCN2 can induce Elm-1, expressed in low metastatic cells; HICP, haparin-induced CCN-like protein; ECM synthesis and TGF-b-mediated fibronectin and col- Cop-1 card-only protein 1. lagen expression via the induction of the SMAD (see Glos- sary) family of transcription factors [6]. However, recent In addition to their highly conserved sequence alignment data indicate that CCN2-dependent stimulation of col- (Figure 1), the CCN proteins share a common intron/exon lagen and extracellular matrix synthesis in murine pattern, and each CCN family member contains five exons. embryonic fibroblasts does not involve the SMAD signal- The first exon corresponds to the signal sequence, and the ling pathway [12,15]. remaining four exons correspond to one of the discrete Recombinant CCN3 induces adhesion of vascular protein modules. This exon arrangement suggests a cer- smooth muscle cells, endothelial cells and fibroblasts tain amount of ‘evolutionary shuffling’ of functional through interactions with integrin cell surface receptors domains that is also seen in other matricellular modular and heparin sulphate proteoglycans [12]. Despite the proteins [1,6,9]. absence of an Arg-Gly-Asp (RGD) tripeptide cell adhesion Although some CCN protein functions are directly motif, which is often required for physical interaction with related to an individual functional module, many functions integrins, CCN3 associates with the vitronectin receptor are thought to result from modules acting in concert. avb3, integrins a5b1 and a6b1, possibly through its C- Truncated CCN2 molecules, each comprising two modules, terminal domain, and with integrins avb3, avb5, a2b1, can function independently to stimulate differentiation or a3b1 and a7b1, which is consistent with the involvement of proliferation of fibroblast and to increase collagen syn- a cell-type-dependent subset of adhesion molecules. thesis [10]. Production of truncated proteins showing altered biological properties, or proteins without some Skeletal development and chondrogenesis internal modules, has been associated with specific devel- Because integrins and other adhesive signalling molecules opmental stages and pathological situations [11]. perform vital roles in tissue remodelling [16], it is not The make-up of a single protein into four modules that surprising that the CCN proteins are also involved in share identity with four large classes of regulatory skeletal formation and development [6,17]. All CCN family proteins, which play fundamental roles in cellular biology,