A single glycine-alanine exchange directs ligand specificity of the elephant progestin receptor

Abstract

The primary gestagen of elephants is 5α-dihydroprogesterone (DHP), which is unlike all other mammals studied until now. The level of DHP in elephants equals that of progesterone in other mammals, and elephants are able to bind DHP with similar affinity to progesterone indicating a unique ligand-binding specificity of the elephant progestin receptor (PR). Using site-directed mutagenesis in combination with in vitro binding studies we here report that this change in specificity is due to a single glycine to alanine exchange at position 722 (G722A) of PR, which specifically increases DHP affinity while not affecting binding of progesterone. By conducting molecular dynamics simulations comparing human and elephant PR ligand-binding domains (LBD), we observed that the alanine methyl group at position 722 is able to push the DHP A-ring into a position similar to progesterone. In the human PR, the DHP A-ring position is twisted towards helix 3 of PR thereby disturbing the hydrogen bond pattern around the C3-keto group, resulting in a lower binding affinity. Furthermore, we observed that the elephant PR ligand-binding pocket is more rigid than the human analogue, which probably explains the higher affinity towards both progesterone and DHP. Interestingly, the G722A substitution is not elephant-specific, rather it is also present in five independent lineages of mammalian evolution, suggesting a special role of the substitution for the development of distinct mammalian gestagen systems.

Figure 1. Progesterone and dihydroprogesterone differ in the reduction state of C4–C5.

(A) Carbon atom and ring numbering system of steroids (B) Chemical structures of progesterone (P4) and 5α-dihydroprogesterone (DHP).

Figure 2. The G722A exchange alters receptor specificity of the PR.

(A) The sequence of human PR LBD was aligned with the corresponding translated genomic DNA sequence of the African elephant (Loxodonta africana). Amino acids making van der Waals contacts with bound ligands are indicated in bold type, amino acids making hydrogen bonds to bound ligands are bold and italicized according to Williams et al. . Secondary-structural elements of the PR LBD are indicated above the sequences. α-helices are pale blue, β-sheets and turns dark blue. Shaded residues indicate elephant specific amino acid exchanges. Dots resemble identical amino acids. (B) Elephant specific amino acid substitutions (+), were consecutively introduced into recombinant human PR LBD and relative binding affinity (RBA) of DHP compared to progesterone measured by competitive binding assays. (C) Competitive binding assays for progesterone and DHP with recombinant human (hPR) and elephant (elePR) PR LBDs. 1 nM [³H]-progesterone was displaced by increasing amounts of progesterone (P4) and DHP. (D) G722A and S796P exchanges were introduced into hPR, while A722G was introduced into elePR. IC50 values were measured as in (C). Data are presented as average IC50 values+SEM of at least three independent experiments.

Figure 3. In elePR alanine 722 orientates the DHP A-ring into a position similar to progesterone.

(A,B) Structures of progesterone (P4) and DHP bound to the human and elephant PR LBD compared to the X-ray structure (PDB code 1a28) (green) generated by molecular-dynamics simulations. The frames of each trajectory were fitted on the start frame using the Cα atoms of the helical parts and resulting averaged coordinates were used for a further fit. Represented are the frames with the least relative deviation to the averaged coordinates. The contour lines represent the solvent/ligand-accessible parts of residue 722. (A) hPR-P4 (orange), elePR-P4 (violet). (B) hPR-DHP (yellow), elePR-DHP (red). (C) Depiction of the four polymorphisms G722A, S796P, V698M, and S902C color-coded as in (A, B). Top panels: Depiction of G722A and S796P: Residue 722 and L797 form an axis through the binding pocket. Center panels: V698M is located in a flexible hinge and rotates freely towards and away from the binding pocket. Met but not Val can reach the crucial binding residue Q725. Bottom panels: Depiction of S902C, the hinge between H11 and H12, and the binding residues in the hinge and in H12.

Figure 4. Different positioning of the A-ring affects hydrogen bond network.

Hydrogen bond populations of the individual receptor ligand interactions were determined by molecular dynamics calculations for hPR-P4 (orange), hPR-DHP (yellow), elePR-DHP (red) and elePR-P4 (violet).

Figure 5. Gain of ligand-affinity in the elephant PR is accompanied by a loss of flexibility in the binding pocket.

Flexibilities of the Cα atoms for the regions of the receptors that are involved in ligand binding, calculated as mean root mean square deviations from the averaged structure (rmsf/A) and difference curves between the hPR and elePR complexes of either given ligand DHP or P4. Residues involved in ligand binding are indicated with black dots.

Figure 6. Increased binding affinity of the elephant PR cannot be generalized to synthetic progestins.

(A) Chemical structures of melengestrol acetate (MGA). (B) IC50 values of MGA and progesterone (P4) binding to human and elephant PR were determined by competitive binding assays.

Figure 7. The G722A exchange evolved 5 times during mammalian evolution.

(A) Sequence alignment of residue 722 and surrounding amino acids of the PR LBD. (B) Phylogenetic tree of mammalian evolution deduced from Murphy et al. (16) and Killian et al. (17). Blue arrows indicate the substitution of G722 to alanine, green arrows the substitution to cysteine. (C) PR LBD DNA sequences from mammals listed in (B) were aligned and a codon analysis for positive/purifying selection performed based on phylogenetic relationships depicted in (B). Residues of elephant PR are color-coded according to their selective pressure during mammalian evolution.