Steroid recognition and engineering of catalysis in mammalian aldo-keto reductases

Of the steroid transforming aldo-keto reductases (AKRs), rat liver 3$\alpha$-hydroxysteroid dehydrogenase (3$\alpha$-HSD) is the best studied and provides a model for understanding protein-steroid interactions. Initially the location of the steroid binding site in rat liver 3$\alpha$-HSD was unknown. Through site-directed mutagenesis of the tryptophans in 3$\alpha$-HSD and analysis of the mutants by steady-state kinetics and protein fluorescence, we proposed that Trp86 and Trp227 form opposite sides of the binding site. Based on this study, a model for steroid binding in 3$\alpha$-HSD was proposed and subsequently confirmed when the structure of the 3$\alpha$-HSD$\cdot$NADP$\sp+\cdot$testosterone ternary complex was determined. This model may be applicable to other steroid transforming AKRs, including other 3$\alpha$-, 17$\beta$-, and 20$\alpha$-HSDs and the $\Delta\sp4$-3-ketosteroid-5$\beta$-reductases. Using both structural and sequence comparisons as a guide, $\Delta\sp4$-3-ketosteroid-5$\beta$-reductase activity was engineered into rat liver 3$\alpha$-HSD by mutating His117 of the catalytic tetrad into a glutamic acid. Our results demonstrate that the AKRs use the same general acid (Tyr55) for catalyzing the reduction of both ketones and carbon-carbon double bonds. The identity of amino acid 117 can determine if an AKR functions as an oxidoreductase and/or a double bond reductase. This result has implications for how Nature evolves new enzymatic activities. In endocrine target tissues, 3$\alpha$-HSD may function as a molecular switch by interconverting active steroid hormones with their cognate inactive metabolites. We cloned, expressed, and characterized the type 2 3$\alpha$-HSD isoform from a human prostate cDNA library. The identification of dual 3$\alpha$-/17$\beta$-HSD activity in this enzyme demonstrates that the steroid binding site accommodates substrate in two orientations. Also, the bifunctional activity allows this enzyme to remove the potent androgen 5$\alpha$-DHT from the prostatic androgen pool. However, it cannot catalyze the formation of 5$\alpha$-DHT and does not act as a molecular switch. This work suggests that different 3$\alpha$-HSD isoforms may have distinct roles in controlling androgen levels in the prostate