Research ArticleBIOCHEMICAL EVOLUTION

Origin of an ancient hormone/receptor couple revealed by resurrection of an ancestral estrogen

See allHide authors and affiliations

Science Advances  31 Mar 2017:
Vol. 3, no. 3, e1601778
DOI: 10.1126/sciadv.1601778
  • Fig. 1 Estradiol synthesis pathway in vertebrates and distribution across metazoans of related biochemical features.

    Vertebrate estradiol synthesis from cholesterol is a multistep pathway involving side-chain cleavage of cholesterol to pregnenolone by the CYP11A enzyme (red), a complex suite of reactions from pregnenolone to testosterone (dotted arrow), and aromatization by the CYP19A enzyme (blue). CYP19A is already present in chordates (blue dot on tree), whereas CYP11A is vertebrate-specific (red dot on tree). Aromatization has been described in cnidarians, mollusks, and cephalochordates (blue ovals). In cephalochordates, CYP19A may catalyze the aromatization reaction, whereas in mollusks and cnidarians, this definitely has to be a different enzyme, possibly a paralogous CYP (blue dots on mollusk and cnidarian branches of the tree). Because there is no consistent evidence for side-chain cleavage activities in those organisms, the exact endogenous substrate of those aromatizing enzymes remains elusive. It could be a steroid (hence the dotted line from cholesterol) but different from the vertebrate ones (hence the incomplete molecule structures).

  • Fig. 2 Parsimony analysis of metazoan steroidogenic pathways.

    (A) Five examples of metazoan steroidogenic pathways. Some reactions are detailed to explain the character-coding procedure. Some pathways share identical reactions, defined here as type I homologies (hI, in red), such as side-chain cleavage of cholesterol to pregnenolone (reaction 31, in red) in the synthesis of estradiol and estrone (C-to-E1 and C-to-E2 pathways) or dehydrogenation of cholesterol to 7-dehydroxycholesterol (reaction 58, in red) in the C-to-ECDY and C-to-DELTA7 pathways. Some pathways also share similar reactions, defined as type II homologies (hII), with various degrees of similarity. Type IIa homologies (hIIa) describe the same chemical modification at the same place on two different substrates, such as aromatization (character 134, in blue) on androstenedione in the C-to-E1 pathway (reaction 54) or testosterone in the C-to-E2 pathway (reaction 55). Type IIc homologies (hIIc) describe the same chemical modification at different places on two different substrates, such as hydroxylation (character 147, in green) on carbon 25 from cholesterol (reaction 94) in the C-to-CHENO pathway or hydroxylation on carbon 20 from ecdysone in the C-to-ECDY pathway. C. elegans, Caenorhabditis elegans. (B) Extract of the data matrix corresponding to the pathways and steps described above. The presence or absence of characters is coded by 1 and 0, respectively (see complete matrix in fig. S1). (C) Nomenclature of carbon numbers and rings on the sterol skeleton.

  • Fig. 3 Relationships among animal steroidogenic pathways inferred from standard parsimony analysis.

    The 50% majority-rule consensus parsimony tree is rooted using eukaryotic sterol synthesis pathways. S1 and S2 indicate suboptimal nodes. Character numbers on branches refer to fig. S1B. All vertebrate sex and adrenal steroid synthesis pathways (highlighted in light red) are grouped together, sharing a side-chain cleavage reaction (31), whereas aromatized steroids are dispersed across the tree, indicating that aromatization of steroids without a side chain was recruited many times independently (blue ovals, 134). A shared trait of all analyzed animal steroid synthesis pathways is the occurrence of hydroxylation on some carbon residues (green dot, 147).

  • Fig. 4 Paraestrol A synthesis and binding assays.

    (A) Synthetic paraestrol A was produced in vitro using a six-step protocol, starting from cholesterol acetate (see the Supplementary Materials). (B) Two ancestral steroid receptors (AncSR_AC and AncSR_D) were reconstructed, taking into account phylogenetic uncertainty regarding the position of annelid and mollusk receptors (see sequence data set in fig. S3 and table S1). AR, androgen receptor; PR, progesterone; MR, mineralocorticoid receptor; GR, glucocorticoid receptor; SR, steroid receptor; ERR, estrogen-related receptor. (C to E) The ability of AncSR_AC and AncSR_D to activate the transcription of the luciferase reporter gene was tested in transfected human embryonic kidney (HEK) 293T cells with increasing concentrations of paraestrol A (10−6, 10−5, and 5 × 10−5 M) or estradiol (10−8 to 10−6 M). Human estrogen receptor α (ERα) was used as control (Ctrl). Activation folds are in arbitrary units, and their absolute values cannot be compared from one receptor to the other. (F to H) The binding ability of AncSR_AC and AncSR_D for paraestrol A and estradiol was tested by limited proteolytic assays. Lane 1, undigested protein; lanes 2 to 8, digested protein with ethanol as negative control (lane 2), paraestrol A at increasing concentrations (lanes 3 to 5 at 10−5 to 10−3 M, respectively), and estradiol (lanes 6 to 8 at 10−8 to 10−6 M, respectively).

  • Fig. 5 Sequence alignments of the ligand-binding pockets of ancestral receptors compared to human ERα.

    α-Helices that make up the LBD are boxed and numbered from 1 to 12 on the basis of the crystal structure from ERα (81). Amino acids from human ERα making direct hydrogen bonds with estradiol are highlighted in green. Amino acids making hydrophobic bonds with estradiol are highlighted in pink. Amino acids known to be involved in coactivator interaction are indicated with a star on top of each site. Differences with human ERα on ligand-binding or coactivator-binding sites are highlighted in yellow. Sequences of AncSR_AC and AncSR_D have been inferred in this study. The other ones come from previous analyses.

  • Fig. 6 Paraestrol A docking into the pocket of the ancestral steroid receptor.

    (A) Overall view of the homology model of AncSR LBD based on the crystal structure of ERα LBD in complex with 17β-estradiol (PDB code, 1ERE). Paraestrol A (green) is docked inside and binds to the receptor through the residues E50 and R91. Similar results are obtained using the crystal structure of the ancestral corticoid receptor in complex with desoxycorticosterone (PDB code, 2Q3Y) as a template (light orange trace). Both traces are from AncSR_D, with the traces from AncSR_AC being identical. (B) Detailed view on the binding pocket. Compared to the binding of 17β-estradiol (orange, inset), paraestrol A lacks the possibility of binding H524 from the receptor through its 17β-hydroxylated carbon. (C) Because of lack of binding to H210, the residue homologous to H524 in ERα (orange, inset), the aliphatic chain of paraestrol A wobbles inside the pocket. A few among many possible conformations are represented by different colors (encircled).

  • Fig. 7 Sequential ordering of animal steroid metabolic pathways.

    (A) Two theoretical models about metabolic pathway evolution: anabolic pathways evolving backward, with the more downstream reactions appearing first (47), and catabolic pathways evolving forward, with the more upstream reactions appearing first (50). (B) Sequential view of pathway appearance during successive periods in animal steroidogenesis. Colors indicate successive periods corresponding to the appearance of new types of enzymatic activities. Final compounds appearing at each period are indicated. All compounds appear in fig. S5. My B.P., million years before the present.

  • Fig. 8 Origin of vertebrate estrogens through opportunistic connections from ancient paraestrol synthesis pathways.

    Coevolution between metabolic pathways, in colors corresponding to different periods (A to C) and steroid receptors of the NR1 (gray) and NR3 (black) subfamilies. We infer that, in addition to the sequential appearance of actual metabolites at different time steps (periods 2 and 3, oxysterols; period 4, progestagens; period 5, androgens), the history of NR ligand synthesis pathways involved metabolites combining features that are separated in actual molecules, such as paraestrols in periods 2 to 4. Those intermediary metabolites may have facilitated a gradual shift in substrate specificity for ancient enzymes, such as the CYP19A aromatase, and hence the building of new pathways through opportunistic connections of old enzymes with new substrates during period 5.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/3/3/e1601778/DC1

    Supplementary Materials and Methods

    fig. S1. Data matrix containing 72 taxa (rows) and 151 characters (columns).

    fig. S2. Complements regarding the coding procedure.

    fig. S3. Phylogenetic relationships among the sequences that were used to reconstruct AncSR_AC and AncSR_D.

    fig. S4. Temporal relationships among steroid synthesis pathways, inferred from a 50% majority-rule consensus parsimony tree.

    fig. S5. Detailed view of animal steroid metabolic pathways.

    table S1. Accession numbers for the 142 sequences used for the predictions of the ancestral receptors.

    table S2. Sequences of the LBDs for the two reconstructed receptors.

    References (8286)

  • Supplementary Materials

    This PDF file includes:

    • Supplementary Materials and Methods
    • fig. S1. Data matrix containing 72 taxa (rows) and 151 characters (columns).
    • fig. S2. Complements regarding the coding procedure.
    • fig. S3. Phylogenetic relationships among the sequences that were used to reconstruct AncSR_AC and AncSR_D.
    • fig. S4. Temporal relationships among steroid synthesis pathways, inferred from a 50% majority-rule consensus parsimony tree.
    • fig. S5. Detailed view of animal steroid metabolic pathways.
    • Legends for tables S1 and S2
    • References (82–86)

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • table S1 (Microsoft Excel format). Accession numbers for the 142 sequences used for the predictions of the ancestral receptors.
    • table S2 (Microsoft Excel format). Sequences of the LBDs for the two reconstructed receptors.

    Download Tables S1 and S2

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article