Crystal structure of LGR ligand α2/β5 from Caenorhabditis elegans with implications for the evolution of glycoprotein hormones

Abstract

A family of leucine-rich-repeat-containing G-protein-coupled receptors (LGRs) mediate diverse physiological responses when complexed with their cognate ligands. LGRs are present in all metazoan animals. In humans, the LGR ligands include glycoprotein hormones (GPHs) chorionic gonadotropin (hCG), luteinizing hormone, follicle-stimulating hormone (hFSH), and thyroid-stimulating hormone (hTSH). These hormones are αβ heterodimers of cystine-knot protein chains. LGRs and their ligand chains have coevolved. Ancestral hormone homologs, present in both bilaterian animals and chordates, are identified as α2β5. We have used single-wavelength anomalous diffraction and molecular replacement to determine structures of the α2β5 hormone from Caenorhabditis elegans (Ceα2β5). Ceα2β5 is unglycosylated, as are many other α2β5 hormones. Both Hsα2β5, the human homolog of Ceα2β5, and hTSH activate the same receptor (hTSHR). Despite having little sequence similarity to vertebrate GPHs, apart from the cysteine patterns from core disulfide bridges, Ceα2β5 is generally similar in structure to these counterparts; however, its α2 and β5 subunits are more symmetric as compared with α and β of hCG and hFSH. This quasisymmetry suggests a hypothetical homodimeric antecedent of the α2β5 and αβ heterodimers. Known structures together with AlphaFold models from the sequences for other LGR ligands provide representatives for the molecular evolution of LGR ligands from early metazoans through the present-day GPHs. The experimental Ceα2β5 structure validates its AlphaFold model, and thus also that for Hsα2β5; and interfacial characteristics in a model for the Hsα2β5:hTSHR complex are similar to those found in an experimental hTSH:hTSHR structure.

Keywords: cystine-knot hormone (CKH); evolution; glycoprotein hormone (GPH); leucine-rich-repeat-containing G-protein-coupled receptor (LGR); thyrostimulin.

Fig. 1.

Overall structure of Ceα2β5. (A) Ribbon diagram of Ceα2β5. The α2 subunit is in red, and the β5 subunit is in sky blue. Disulfide bridges are represented as yellow sticks and structural elements are labeled. (B) Stereo C_α trace of Ceα2β5 in the same orientation as (A), emphasizing the heterodimer interface. Residues contributing to interchain hydrogen bonds are in stick representation for the main chain. Hydrogen bonds between Ceα2 and Ceβ5 subunits are shown as dotted lines. Every tenth residue is labeled and marked by a filled circle on the C_α position.

Fig. 2.

Topology and sequence comparisons. (A) Schematic drawing of the Ceα2β5 heterodimer topology. The α2 subunit is red, and β5 is sky blue. The secondary structural elements are indicated by cylinders for α helices and arrows for β strands. Cysteine residues are labeled. (B) Schematic drawing of the hCG heterodimer topology. The α subunit is magenta, and β is blue. The N-linked glycosylation sites are defined by large black “Y” symbols. (C) Sequence alignment of α2 and GPHα subunits. (D) Sequence alignment of β5 and GPHβ subunits. The secondary structural elements are those assigned to the Ceα2β5 structure, with cylinders for α helices and arrows for β strands. Cysteine residues are highlighted with yellow background, and disulfide pairs are indicated by orange connections. The cysteine residues conserved in cystine knots are underscored by asterisks (*). Ce, Caenorhabditis elegans; Dm, Drosophila melanogaster; Hs/h, Homo sapiens/human.

Fig. 3.

Comparison of the Ceα2β5 and hCG heterodimers. (A) Ribbon diagram of Ceα2β5-1 (α2 in red and β5 in sky blue). (B) Ribbon diagram of hCG (α in magenta and β in blue). Disulfide bridges are shown as yellow sticks for each. (C) Superimposition of the α2 and β5 subunit in Ceα2β5. (D) Superimposition of the α and β subunits of hCG. Arrow-headed black lines indicate the quasidyad symmetry axes that relate α2 and β5 in Ceα2β5 and α and β in hCG. The N- and C-termini are labeled in each.

Fig. 4.

Comparison of the two Ceα2β5 molecules in the form 2 (P2₁) asymmetric unit. (A) Superimposition of Ceα2β5-1 (α2-1 in red and β5-1 in skyblue) with Ceα2β5-2 (α2-2 in pink and β5-2 in cyan). Disulfide bridges are shown as yellow sticks. (B) Superimposition of Ceα2-1 (red) with Ceα2-2 (pink). The superimposed pair has been rotated clockwise by 35° about the x-axis from the view in (A) to afford a better view of the cystine-knot motif. (C) Superimposition of Ceβ5-1 (skyblue) with Ceβ5-2 (cyan). The superimposed pair has been rotated anticlockwise by 35° about the x-axis from the view in (A) to afford a better view of the cystine-knot motif. (D) Elaboration of the L1^β5 and L3^β5 loops from (C). The C_α atoms of Asn23 at the tip of L1^β5 and of Asp77 and Gly78 at the tip of L3^β5 in Ceβ5-2 have shifted inward by 7.5 Å, 8.1 Å, and 8.3 Å, respectively, as compared with Ceβ5-1. The side chains of β5 Asn23, Asp77, and Gly78 are shown as sticks with C_α atoms highlighted as filled spheres. Yellow dotted lines show the C_α shifts.

Fig. 5.

Electrostatic potential surfaces from ligand:receptor complexes: (A) Ceα2β5:CeLGR, (B) Hsα2β5:hTSHR and (C) hTSH:hTSHR complexes. Ligands (Left) are shown after rotation from the receptor (Right), oriented as in (D). The views are as after the opening of a book. Solvent-accessible surfaces are shown as colored by the electrostatic potential at this surface. Negative potentials are shown in gradations of red, and positive potentials are in gradations of blue. The boundaries of buried interface within each complex are drawn as black curves. (D) A ribbon diagram of the hTSH:hTSHR complex (PDB:7T9I) provides a key to the orientations of ligands and receptors in (A–C).

Fig. 6.

Sequence-aligned disposition of cysteines and disulfide linkages: (A) TGFβ and (B) LGR ligands. The disulfide bridges in common to all chains are represented by black connecting arrows, and the cysteine positions of the cystine-knots are marked by asterisks (*). Characteristic disulfide bridges of TGFβ, LGRα, and LGRβ subunits are shown as gray, red, and blue connecting arrows, respectively. Interchain disulfide bridges, designated as known from crystal structures or assigned based on our AlphaFold models, are shown as green connections. Hs, Homo sapiens (human); Ml, Mnemiopsis leidyi (comb jelly); Nv, Nematostella vectensis (sea anemone); Dm, Drosophila melanogaster (fruit fly); Ce, Caenorhabditis elegans (Nematode); Pm, Petromyzon marinus (lamprey); Ea, Eptatretus atami (hagfish). Adapted from Fig. 1A of Roch and Sherwood (19).

Fig. 7.

Implications for the evolution of glycoproteins. (A) Hypothetical course for the evolutionary divergence of LGR ligand genes. (B) Present-day representatives of LGR ligand structures at points of evolutionary divergence in (A), with comb jelly CKH a1:a1 deriving from (α_ancestral)₂, nematode α2β5 deriving from an original α2β5 dimer, hagfish αβ0, and lamprey α2β5 deriving from pre-WGD cyclostomes, and hTSH and Hsα2β5 deriving from post-WGD progenitors. Surfaces are drawn in sphere representation (nonhydrogen atoms at van der Waals radii). The α2, β5, GPHα, and GPHβ subunits are in red, sky blue, magenta, and blue, respectively. Except for our experimental structure of Ceα2β5 and a recently published structure of hTSH in the hTSH-hTSHR complex (12), all of these are AlphaFold models.