The structures of transforming growth factor-β (TGF-β), nerve growth factor (NGF), platelet-derived growth factor (PDGF) and gonadotropin have been shown to be similar [1,2 and references therein]: these proteins are folded into two highly twisted antiparallel pairs of β-strands and contain three disulfide bonds, of which two form a cystine ring through which the third bond passes (see the schematic representation below). This structure is called cystine knot [3].

            +--<======b3======C---c------+
            |                 :   :      |
            +--==c====b4===>C-:-C-c-COOH |
                 :          : : :        |
            +--==c====b2===>C---C--------+
            |                 :
            +--<======b1======C--------------NH(2)

'=>': indicates the direction of the beta-strands 'b1' to 'b4'.
'C' : conserved cysteine in cystine knot.
'c' : additional conserved cysteine in C-terminal cystine knot.
':' : disulfide bridge.

Functional diverse modular proteins share a conserved domain of about 90 amino acids in their C-terminal cysteine-rich region, that has been proposed [4,5] to be structurally related to the cystine-knot family [3] and which is therefore called C-terminal cystine-knot (CTCK). Members of the C-terminal cystine knot family are listed below:

• von Willebrand factor (vWF), a multifunctional protein which is involved in maintaining homeostasis. It consists of 4 vWF type D domains, 3 vWF type A domains, 3 vWF type B domains, 2 vWF type C domains (see <PDOC00928>, a X domain and the C-terminal cystine knot.
• Mucins. Human mucin 2, a highly polymorphic multidomain molecule with a modular architecture similar to vWF. Xenopus mucin B.1 which contains a CCP domain, a vWF type C domain, a X domain and a CTCK. Other mucins that contain a CTCK are the human tracheobronchial mucin (gene MUC5), bovine submaxillary mucin-like protein, pig apomucin and rat intestinal mucin-like protein.
• CCN family (cef-10/cyr61/CTFG/fisp-12/nov protein family). These growth- factor inducible proteins are structurally related to the insulin-like growth factor binding proteins (see <PDOC00194>) and could also function as growth-factor binding proteins.
• Drosophila slit protein which is essential for development of midline glia and commissural axon pathways. It is composed of four leucine-rich repeats, seven EGF-like domains, a laminin G-like repeat and the CTCK.
• Norrie disease protein (NDP) which may be involved in neuroectodermal cell- cell interaction and in a pathway that regulates neural cell differentiation and proliferation.
• Silk moth hemocytin, an humoral lectin which is involved in a self-defence mechanism. It is composed of 2 FA58C domains (see <PDOC00988>), a C-type lectin domain (see <PDOC00537>), 2 VWFC domains (see <PDOC00928), and a CTCK.

The pattern we developed for CTCK correspond to the C-terminal half of the domain.