Structure of human steroid 5α-reductase 2 with the anti-androgen drug finasteride

Evidence

Extracted from the full text; page numbers refer to the paper. The library’s own notes are labelled as such.

Key results

1. Structure solved at 2.8 Å; the enzyme has seven transmembrane helices (unlike the 10 of the bacterial MaSR1), with all 254 residues modelled except the first four and S39–A43; deposited as PDB 7BW1 · pp. 1–3, 9, Table 1 (p. 3)
2. The bound ligand is not finasteride itself but an NADP–dihydrofinasteride adduct, with the NADPH C-4 and finasteride C-2 atoms about 1.5 Å apart (a covalent bond), sealed in a cavity open only to the membrane; this matches earlier enzymology. Citing that work (Ki ≤ 3 × 10⁻¹³ M, koff = 2.74 × 10⁻⁷ s⁻¹), the paper gives a half-life of about 31 days for the enzyme–adduct complex; ln 2 / koff gives about 29 days · p. 3, Fig. 2 (p. 4)
3. Proposed mechanism: residues E57 and Y91 hold the steroid's C-3 carbonyl so NADPH can transfer a hydride; the testosterone pose comes from computer docking (hydride about 2.5 Å from the Δ4,5 bond). Y91F essentially abolished DHT formation and E57Q reduced it (Fig. 3e, read from the chart, DHT/T about 2.2% wild type, 1.0% E57Q, near 0 Y91F, and about 0.5% wild type with 0.5 mM finasteride) · p. 4, Fig. 3 (p. 5)
4. Selectivity: arginine R114 hydrogen-bonds finasteride's tert-butylacetamide tail; SRD5A1 has a methionine at this position, which may explain finasteride's preference for type 2 · pp. 3–4
5. Simulations showed the cytosolic loop L1 (and less so L5) to be highly flexible, opening the NADPH pocket; the authors propose L1 acts as a gate for cofactor exchange, consistent with its higher crystal B-factors · p. 5, Fig. 4 (p. 6)
6. Over 100 SRD5A2 gene mutations cause 5α-reductase deficiency; most mapped missense sites line the ligand cavity (e.g. R227Q, R171S, likely disrupting NADP binding), while others such as C133G and R246 changes likely affect folding or stability · p. 6, Fig. 5 (p. 7)