Common structural traits for cystine knot domain of the TGFβ superfamily of proteins and three-fingered ectodomain of their cellular receptors

Abstract

The transforming growth factor-β (TGFβ) superfamily of proteins and their receptors are crucial developmental factors for all metazoan organisms. Cystine-knot (CK) motif is a spatial feature of the TGFβ superfamily of proteins whereas the extra-cellular domains (ectodomains) of their respective receptors form three-fingered protein domain (TFPD), both stabilized by tight cystine networks. Analyses of multiple sequence alignments of these two domains encoded in various genomes revealed that the cystines forming the CK and TFPD folds are conserved, whereas the remaining polypeptide patches are diversified. Orthologues of the human TGFβs and their respective receptors expressed in diverse vertebrates retain high sequence conservation. Examination of 3D structures of various TGFβ factors bound to their receptors have revealed that the CK and TFPD domains display several similar spatial traits suggesting that these two different protein folds might have been acquired from a common ancestor.

Fig. 1

CK, ICK, and TFPD spatial motifs extracted from the structure of TGFβ3 bound to TGFβ-IIR/TGFβ-IR ternary complex (2PYJ.pdb) [7], the human agouti-related protein (1HYK.pdb) [8], and the X-ray structure of the human TGFβ-RII ectodomain/TGF-β3 complex (1KTZ) [16], respectively. a The presented CK spatial motif is an eight-member cycle formed with C44 linked to C48 via S45-G46-P47 triad (red) whereas C48 forms an SS bond with C111 (yellow) that is linked to C109 via K110 (light blue) and which form an SS bond with C44 (yellow); the SS bond C15–C78 (violet) passes through the cycle (knotted SS bond). The additional SS bond (C7–C16) prolongs the ‘knotted SS bond’ (SS_KB) while C77 (blue) may form an intermolecular SS bond (homo- or heterodimer); b the ICK spatial motif in the structure of the human agouti-related protein forms a cystine stack made with the following three SS bonds: C1–C16 and C8–C22 (yellow) surrounding the threading SS bond formed with C15–C33 (violet); C1 is linked to C8 with the sequence VRLHES (wheat) whereas C8 is linked to C16 with the sequence LGQQV (orange) while the remaining two other SS bonds (C19–C43 and C24–C31 in blue) are outside the ICK motif; c the tight Cys network at the base of the palm is made of three SS bonds (C28–C61, C54–C78, and C115–C120, yellow sticks) and C113–C98 (blue sphere) which is at a distance longer than 7.5 Å from to the three SS bonds; C31–C48 (pink sticks) and C38–C44 (orange sticks are in Finger 1c and do not contribute to the stability of the principal Cys network; there is highly conserved N residue (N121, red sphere) in all the TFPDs [12]

Fig. 2

Schema illustrating analyses of sequence attributes in multiple sequence alignment (MSA)

Fig. 3

a Multiple alignment of 37 sequences of the CK domains of the human TGFβ superfamily of proteins and 8 of their antagonists. b Dendrogram of the human TGFβ superfamily of proteins and their antagonists; the dendrogram was made with dendroscope [31]

Fig. 4

a Distribution of the IDs in the MSA45 (brown bars) and MSA449 (violet bars). b Information entropy (I_e) calculated from the MSA45. One should note that I_e = 0 for fully conserved sequence positions whereas I_es > 1.0 signify that given sequence position display heterogeneous AAC. Negative values were automatically assigned to sequence positions having less than 50% residue occupancy

Fig. 5

Bi-dimensional map of BMP7 (1M4U.pdb) bound to Noggin [36] (upper panel) and human Artemin bound to its receptor GFRα3 (2GYZ.pdb) [37] (lower panel). The sequence similarity score for hBMP7/Artemin is ID = 17%. β-strands are shown as blue arrows whereas α-helices positioned near the diagonal are shown as yellow arrows

Fig. 6

2D map of intramolecular interaction clusters in BMP2 (upper panel) bound to BMP-RIA1 (lower panel) (2H64.pdb) [41]

Fig. 7

The structure of murine Sclerostin (upper panel, 2KD3.pdb) [53] and Activin receptor IIB (lower panel, 1S4Y.pdb) [54] showing spatial positioning of the fingers, a flexible ‘heel’ loop while the disulfide networks are shown in shaded ellipses