XB-ART-57473Nat Commun January 1, 2020; 11 (1): 5357.
Sclerostin inhibits Wnt signaling through tandem interaction with two LRP6 ectodomains.
Low-density lipoprotein receptor-related protein 6 (LRP6) is a coreceptor of the β-catenin-dependent Wnt signaling pathway. The LRP6 ectodomain binds Wnt proteins, as well as Wnt inhibitors such as sclerostin (SOST), which negatively regulates Wnt signaling in osteocytes. Although LRP6 ectodomain 1 (E1) is known to interact with SOST, several unresolved questions remain, such as the reason why SOST binds to LRP6 E1E2 with higher affinity than to the E1 domain alone. Here, we present the crystal structure of the LRP6 E1E2-SOST complex with two interaction sites in tandem. The unexpected additional binding site was identified between the C-terminus of SOST and the LRP6 E2 domain. This interaction was confirmed by in vitro binding and cell-based signaling assays. Its functional significance was further demonstrated in vivo using Xenopus laevis embryos. Our results provide insights into the inhibitory mechanism of SOST on Wnt signaling.
PubMed ID: 33097721
PMC ID: PMC7585440
Article link: Nat Commun
Genes referenced: dkk1 egf ldlr lrp6 nodal sia1 wnt1 wnt2 wnt3a wnt9b
GO keywords: Wnt signaling pathway
Article Images: [+] show captions
|Fig. 1: Construct design and the crystal structure of the LRP6 E1E2–SOSTtr177 complex. a Schematic representations of the LRP6 and SOST constructs. SP signal peptide, P β-propeller domain, EGF epidermal growth factor-like domain, LA LDLR type-A repeats, TM transmembrane domain, CTD C-terminal domain, CK cystine-knot domain. The amino acid residues interacting with LRP6 E1 in the SOST-loop 2 region are shown in dark green. The SOST C-terminal tail sequences including HNQS (orange) and the basic-residue clusters (blue) are also shown. Construct information is described in detail in Supplementary Table 1. b The crystal structure of the LRP6 E1E2–SOSTtr177 complex is represented with a ribbon diagram. For simplicity, only one complex out of two in an asymmetric unit is shown (Supplementary Fig. 4). SOST is colored in magenta, and disulfide bonds are represented as yellow sticks. SOST residues 122–127 are not resolved in this structure and are represented as a dashed line. The LRP6 E1 β-propeller domain is colored in green, and the E2 β-propeller is shown in cyan. The EGF domains are shown in orange.|
|Fig. 2: Canonical binding site of SOST loop 2 in LRP6 E1. a Key interactions of the SOST loop 2 region with LRP6 E1 (within 4 Å) are shown with dotted lines. Interacting residues are presented as sticks and are labeled black for LRP6 or magenta for SOST. The missing region of SOST (residues 122–127) in this structure is represented with a dashed line. b Comparison of the SOST loop 2 binding site in our structure with that of the LRP6 E1–SOST peptide complex structure (PDB ID: 3SOV) is presented. The SOST “LPNAIGR” peptide is shown in light gray, and LRP6 E1 from PDB 3SOV is omitted for clarity. The interactions of the NXI motif were found to be conserved in both structures, although the R121-mediated interactions differed. The interactions of R121 in the LRP6 E1-peptide structure are represented with orange dotted lines. c Structural comparison with the LRP6 E1–DKK1 peptide (NSNAIKN) complex (PDB ID: 3SOQ) is shown. The DKK1 peptide is colored in light gray. The interactions of K43 in the LRP6 E1–DKK1 peptide structure are represented in orange dotted lines.|
|Fig. 3: Interaction of SOST C-tail with LRP6 E2. a The HNQS motif of the SOST C-terminus fits into the binding pocket of the LRP6 E2 β-propeller domain. Interactions of SOST H174 with LRP6 residues are shown with black dotted lines. b The ligand-free state of LRP6 E2 (PDB ID: 3S94; light gray) was aligned to LRP6 E2 (cyan) in our SOST-bound structure. For clarity, SOST was omitted from the overall alignment of LRP6 E2, on the left. Structural changes near W465 and H534 in LRP6 E2 upon SOST binding are shown. H534 of LRP6 moved from the concave center of the β-propeller domain by binding to SOST H174 (colored in magenta in an enlarged box on the right). c The structure of HNQS-bound LRP6 E2 (cyan) was aligned with the structure of loop 2-bound LRP6 E1 (green). The HNQS region is not shown for clarity. Loop 2-interacting residues in LRP6 E1 and corresponding residues in LRP6 E2 are represented as sticks and E2 residues are labeled after the E1 residues.|
|Fig. 4: Effects of the SOST C-tail on LRP6 binding and Wnt1-signaling inhibition. a The affinities of various SOST mutants for LRP6 E1E2 or LRP6 E1 were measured by MST, and the relative KD values from those measurements are shown. The KD value of SOST WT for LRP6 E1E2 was used as a reference. The experimentally determined KD values are provided in Supplementary Table 2. b The inhibitory effects of SOST mutants on Wnt1 signaling were detected by TopFlash assays. Each luciferase signal observed after co-transfecting plasmids encoding Wnt1 and each SOST mutant was normalized to that found with the empty-vector control. c Differences in the inhibitory activities of SOST with WT LRP6 or the LRP6Δacidic mutant in LRP6-knockout HEK293T cells are shown in the bar graph. Data from three independent experiments (n = 3) were analyzed and expressed as mean ± SEM in b and c.|
|Fig. 5: Assessing the inhibitory effect of SOST on Wnt signaling in Xenopus embryos. a The transverse dark-brown line (blue arrow) indicates the normal position of the axis. The yellow arrow indicates the position of the second ectopic axis. Wnt1 induced formation of the second ectopic axis, and successfully inhibiting Wnt resulted in the formation of a single axis. Embryos at the four-cell stage were co-injected in a ventro-vegetal blastomere with 30 pg Wnt1 mRNA (n = 104), 25 pg WT SOST mRNA (n = 106), 25 pg SOSTtr169 mRNA (n = 98), 25 pg SOSTtr177 mRNA (n = 119), 25 pg SOSTtr204 mRNA (n = 104), and 25 pg SOST∆loop2 mRNA (n = 96) (n number of embryos). b The percentage of embryos with a second axis. The error bars represent the standard error of the mean (SEM) from three independent experiments (n = 3) and p value by two-tailed t-test is indicated. The number of injected embryos is described in a. c, d Real-time qPCR analyses of the expression of genes directly targeted by Wnt1, namely siamois (c) and nodal 3.1 (d) from animal cap explants isolated at stage 8 and grown to stage 11. The error bars represent SEMs from three independent experiments (n = 3) and p values by two-tailed t-test are indicated.|
|Fig. 6: Proposed mechanism whereby SOST inhibits Wnt signaling. The inhibitory effects of WT SOST and SOST mutants on Wnt signaling activated by Wnt1, Wnt2, Wnt9b, and Wnt3a were measured by TopFlash assays. Full-length DKK1 was also used in the activity assays for comparison with SOST. Data from three independent experiments (n = 3) were analyzed and expressed as mean ± SEM. b A schematic model of the mechanism whereby SOST inhibits the canonical Wnt signaling pathway. In a previous model (left), only the loop 2 region of SOST was suggested to interact with the LRP6 E1 domain, which appears to inefficiently inhibit the binding of various Wnt subtypes. The double-headed arrows indicate the wobbling motion of SOST, owing to flexibility in the loop 2 region. On the other hand, the structural and functional data presented here suggest that the C-terminal HNQS region of SOST binds to LRP6 E2, and that the basic cluster region in the far C-terminus of SOST may further strengthen the interaction with the E2 domain. Based on the SOST C-tail-mediated interaction with E2, we propose a new binding model for SOST (right). In this model, tandem interaction of SOST with LRP6 E1 and E2 leads to more efficient blockage of Wnt binding.|
References [+] :
Afonine, Towards automated crystallographic structure refinement with phenix.refine. 2012, Pubmed