CYSTEINE-RICH REPEATS: WHAT ARE THEY GOOD FOR?

Conserved motifs found in proteins often help elucidate the function and structure of the protein. However, equivalent structural motifs in different proteins can have similar yet distinct functions. This seems to be the case with cysteine-rich repeat (CR) domains associated with Chordin. Chordin CR domains are typically 60-80 amino acids in length and characterized by ten cysteines with a conserved spacing pattern.¹ A typical CR repeat contains a conserved glycine and an aromatic residue, usually tryptophan, found between the first pair of cysteine residues, followed by CXXCXC motif in the middle and a CCXXC motif at the C-terminal end (with X representing any amino acid).^1,2 CR domains are also known as von Willebrand factor-C domains, identified in collagens, thrombospondins, CCN proteins, such as cTGF, and von Willebrand factor.³

Figure 1. Chordin-like CR domains are present in many extracellular proteins. This illustration is a schematic representation of the protein domains present in Chordin, Sog, Chordin-like proteins, Procollagen IIA, CTGF, Crossveinless-2, Kielin, and KCP (Kielin/Chordin-like protein). Red and blue boxes correspond to CR domains. Xenopus Kielin contains a total of 27 CR domains, while its mammalian homolog KCP contains 18 CR domains. CTGF has an N-terminal insulin-like growth factor binding domain and a C-terminal cysteine knot (CT). CTGF and Kielin also have a thrombospondin-1 domain (TSP). Crossveinless-2, Kielin, and KCP also contain a von Willebrand factor D domain (vWF-D). [Note: this figure is adapted from Abreu, J.G. et al. (2002) Gene 287:39.]

Chordin in vertebrates and Sog, a homolog in Drosophilia, share four highly conserved CR domains (Figure 1). These domains are necessary for Chordin/Sog biological activity—that is to bind to and inhibit BMP (bone morphogenetic protein) activity by reducing access to its receptor. In isolation these domains, especially CR1 and CR3 in Chordin and CR4 in Sog, have the same BMP binding activity, albeit at a lower level than the full-length protein.^4,5 Two related Chordin-like molecules have recently been identified: Chordin-like 1 (also known as ventroptin and neuralin) and Chordin-like ². Each of these contains 3 CR domains, where CR1 and 3 are most closely related to CR3 in Chordin.^6,7 These molecules have distinct expression patterns, yet also serve to bind and inhibit BMP activity.^6,7 Even CR domains in unrelated proteins can inhibit BMP activity. For example, procollagen IIA containing one CR domain, has dorsalizing activity in Xenopus, demonstrating its ability to inhibit endogenous BMP signaling.⁴ In addition, CTGF (connective tissue growth factor), which contains one CR domain in combination with an insulin-like growth factor domain (IGFB) and a thrombospondin type 1 repeat, also binds and inhibits BMP activity.⁸ Thus, the CR domains seem to define a modular BMP-binding domain that provides a basis for understanding the function of extracellular proteins that contain this domain.⁴

Of course, biology is never that simple. It turns out that CR domains are likely multi-functional, with distinct activities dependent upon other nearby motifs or sequences. Using a series of deletions, point mutations, and CR domain swaps in Drosophila, Yu et al. show that a construct containing only the first 2 CR domains in Sog actually promotes BMP activity rather than antagonizing it.⁵ Further analysis indicates that CR domains are partially redundant in function and that surrounding sequences are the predominant determinants of the type of Sog activity.⁵ Likewise, a different splice form of Chordin-like 2, in which CR domains remain unaffected, does not bind BMP, but does interact with Activin A, another TGF-ß superfamily member.⁹ Interestingly, CTGF can also bind and enhance TGF-ß signaling, and TGF-ß can compete with BMP-4 for binding.⁸ These results indicate that a single CR motif may have multiple functions, or that through the participation of other domains, CR-containing proteins can have multiple roles in the same cellular context. In the case of other CR domain-containing proteins, homologous proteins in different species have distinct effects. While CV-2 (crossveinless-2) in vertebrates acts as a BMP antagonist, CV-2 in Drosophila enhances BMP signaling, even though both forms are structurally similar with five CR domains and a vWF-D domain.^10,11 Kielin, a 27 CR domain protein in Xenopus exhibits dorsalizing activity, (i.e. BMP blocking activity), while its homolog in mammals, KCP, contains 18 CR domains and elevates BMP activity through enhanced binding to the type I receptor.^12,13 Thus, CR domains in extracellular proteins contribute to differentially modulating the activity of several TGF-ß superfamily members.

References

1. Abreu, J.G. et al. (2002) Gene 287:39.
2. O'Leary, J.M. et al. (2004) J. Biol. Chem. 279:53857.
3. Bork, P. (1993) FEBS Lett. 327:125.
4. Larrain, J. et al. (2000) Development 127:821.
5. Yu, K. et al. (2004) Genetics 166:1323.
6. Nakayama, N. et al. (2001) Dev. Biol. 232:372.
7. Nakayama, N. et al. (2004) Development 131:299.
8. Abreu, J.G. et al. (2002) Nat. Cell Biol. 4:599.
9. Oren, A. et al. (2004) Gene 331:17.
10. Kamimura, M. et al. (2004) Dev. Dyn. 230:434.
11. Binnerts, M.E. et al. (2004) Biochem. Biophys. Res. Commun. 315:272.
12. Matsui, M. et al. (2000) Proc. Natl. Acad. Sci. USA 97:5291.
13. Lin, J. et al. (2005) Nat. Med. 11:387.