Paul Talalay† and H. Richard Levy ‡

The Ben May Laboratory for Cancer Research and Department of Biochemistry, University of Chicago

Absolute steric specificity is an intrinsic and almost universal property of enzymic reactions. The highly asymmetric nature of the catalytic proteins is believed to govern this stereospecificity. Recent formulations of the detailed mechanism of certain enzyme reactions have specified the requirement for sterically directed group displacements at the active centre (Koshland, 1954, 1956).

Perhaps in no other branch of chemistry have stereochemical problems commanded such intense interest as in the field of steroids. The immense versatility of pure cultures of microorganisms has in recent years been exploited to achieve many highly selective and stereospecific transformations of steroid hormones (Eppstein et al, 1956; Talalay, 1957a; Vischer and Wettstein, 1958).

In this paper, we propose to examine the enzymic mechanisms of steroid dehydrogenations by microbial systems and to analyse the chemical and stereochemical specificities of these reactions. Certain features of these systems are especially favourable for such studies. Many natural and synthetic modifications of the steroid structure are available, and there is widespread interest in the relation of structure and steric configuration to biological activity. The immense affinity and rigid structural requirements for steroid-enzyme interactions have provided information on the nature of the complementary groups participating in these combinations. Our interest in these reactions originated from a search for enzymic methods of steroid analysis (Hurlock and Talalay, 1957; 1958a) and for model systems for the study of steroid-protein interactions (Marcus and Talalay, 1955; Talalay, 19576). Stereochemical considerations have been prominent in many phases of this work.

* These studies were supported by the American Cancer Society, † Supported by a permanent faculty level grant from the American Cancer Society.

‡ Postdoctoral Fellow of the U.S. Public Health Service.

Hydroxysteroid Dehydrogenases

Although metabolic interconversions of various hydroxy-and ketosteroids were recognized many years ago, their enzymic mechanism has only recently been clarified. The hydroxysteroid dehydrogenases (Talalay, 1957a) are widely distributed, pyridine nucleotide-linked enzymes which interconvert hydroxyl and ketone functions of steroids, and exhibit a high degree of specificity with respect to the position and steric course of the reactions which they catalyse*:

Steroid alcohol+DPN+ (TPN+) <-> steroid ketone+DPNH(TPNH)+H+.

Several hydroxysteroid dehydrogenases have been obtained in varying degrees of purity and characterized with respect to steroid and pyridine nucleotide specificity. The most detailed information is available from studies with two highly purified, soluble hydroxysteroid dehydrogenases isolated as adaptive enzymes from Pseudomonas testosteroni, a micro-organism which may utilize various steroids as its only carbon source (Talalay, Dobson and Tapley, 1952).

3a-Hydroxysteroid dehydrogenase (Talalay and Marcus, 1956; Marcus and Talalay, 1956) promotes the reversible intercon-version of 3<x-hydroxy- and 3-ketosteroids of the C19, Cw and CM series with DPN as the specific hydrogen carrier. Characteristic reactions catalysed by this enzyme are:

Androsterone + DPN+ <-> androstane-3,17-dione+DPNH+H+

3a, 17a, 21-Trihydroxy-5p-pregnane-ll,20-dione+DPN+<-> 17a, 21-dihydroxy-5(3-pregnane-3,ll,20-trione + DPNH+H+

* The following abbreviations are used: DPN+, DPNH, TPN+ and TPNH for the oxidized and reduced forms of di- and triphosphopyridine nucleotides respectively. PMS, phenazine methosulphate.

3a, 7a, 12a-Trihydroxy-5β-cholan-24-oic acid+DPN+ <-> 7a, 12a-dihydroxy-5β-cholan-3-one-24-oic acid+DPNH+H+

Similar enzymes have been isolated from other sources. A DPN-linked 3a-hydroxysteroid dehydrogenase from Escherichia freundii reacts preferentially with bile acids and does not attack steroid hormones (Hayaishi et al, 1955). In the livers of a number of species, a soluble (Tomkins, 1956a and b) and a microsomal (Hurlock and Talalay, 1959) 3a-hydroxysteroid dehydrogenase have been described. The liver enzymes react with DPN and TPN at comparable rates whereas the bacterial enzymes require DPN specifically.

An adaptive, DPN-specific p-hydroxysteroid dehydrogenase from Ps. testosteroni (Marcus and Talalay, 1956; Talalay and Marcus, 1956) catalyses the reversible oxidation of 3β-hydroxyl groups of C19 and C21 steroids with A :B cis or trans ring fusions. The same highly purified enzyme also interconverts 17β-hydroxyl and 17-keto groups of C18 and C19 steroids.

3β-Hydroxy-5a-androstan-17-one+DPN+ <-> 5a-androstane-3,17-dione+DPNH+H+

Testosterone+DPN+ <->4-androstene-3,17-dione+DPNH+H+

Oestradiol-17β+DPN+ <-> oestrone+DPNH+H+

17β-Hydroxy-4-oestren-3-one+DPN+ <->4-oestrene-3,17-dione+DPNH+H+

A 17β-hydroxysteroid dehydrogenase which reacts preferentially with oestradiol-17β and other ring A aromatic steroids has been purified from human placenta (Langer and Engel, 1958). This enzyme, which has dual nucleotide specificity, reacts at approximately twice the rate with DPN (H) as with TPN (H). The hydroxysteroid dehydrogenases are sulphydryl enzymes which are sensitive to heavy metals and are profoundly stabilized by certain steroids and by pyridine nucleotides.

The known hydroxysteroid dehydrogenases of animal origin appear to possess dual pyridine nucleotide specificity. Consequently, these enzymes may promote not only stoichiometric reactions between steroids and pyridine nucleotides, but also the transfer of hydrogen between the two forms of pyridine nucleotides, in the following manner:

In these transhydrogenation reactions, specific steroids function catalytically and assume the r61e of hydrogen carriers or coenzymes. This catalytic function of steroids has been demonstrated with extremely low concentrations of oestradiol-17p and the placental 17β-hydroxysteroid dehydrogenase (Talalay and Williams-Ashman, 1958; Talalay, Hurlock and Williams-Ashman, 1958), and for the soluble 3<x-hydroxysteroid dehydrogenase of rat liver (Hurlock and Talalay, 19586). It has been suggested that the r61e of mammalian hydroxysteroid dehydrogenases with dual pyridine nucleotide specificity is to function as transhydrogenases. The ease of reversal and high affinity for steroids appear to favour this function.

Reaction kinetics. Kinetic studies have been carried out with purified 3a- and p-hydroxysteroid dehydrogenases to obtain detailed information on the relation between the velocity of oxidation and the concentration of substrates (Marcus and Talalay, 1955; Talalay and Marcus, 1956; Talalay, 19576). The most purified enzyme preparations catalyse the oxidation of 20-50 µmoles of steroid per minute per mg. of protein at 25° and pH 9.0. Both enzymes show extremely high affinities for certain steroids, as evidenced by the very small Michaelis constants (10^-6 to 10^-7 M or lower) which are among the lowest reported for any enzyme-substrate interaction.

The oxidation velocity of substrates catalysed by 3a-hydroxy-steroid dehydrogenase followed the well known postulates of Michaelis and Menten: Vm/V=1+KM/S where V is the initial reaction velocity at a substrate concentration S. The velocity rises asymptotically to a maximum (Vm) as 5 is increased; KM is the Michaelis constant. In contrast, the reaction kinetics of P-hydroxysteroid dehydrogenase with most substrates were found to be anomalous in that the velocity did not approach a maximum with increasing substrate concentration, but exhibited a sharp optimum velocity (V0) at a critical substrate concentration where V and S signify velocity and substrate concentration as previously, K1 and K2 are analogous to the Michaelis constant for the complexes containing respectively one and two molecules of substrate per active site of enzyme, and U is the hypothetical maximum reaction velocity which would have been obtained in the absence of inhibition by excess substrate. Useful graphical solutions for this relation are available (Friedenwald and Maeng-wyn-Davies, 1954).

(50). A second order theory to account for this inhibition by excess substrate was first proposed by Haldane (1930), and gave very satisfactory agreement with experimental observations (Marcus and Talalay, 1955). The kinetics were derived on the assumption that apart from the usual complex ES containing one molecule of substrate (5) per active site of enzyme (E), there occurred an additional enzyme substrate complex (ES2), which was catalytically inactive, and which contained two molecules of substrate per active site of enzyme. This condition is satisfied by the equation:

Specificity of substrate binding. The affinity constants and maximum velocities of oxidation of a large number of related substrates by 3a- and β-hydroxysteroid dehydrogenases have been reported (Marcus and Talalay, 1955, Talalay and Marcus, 1956; Talalay, 19576). These measurements indicate that the areas of the steroid molecules involved in the interaction with both enzymes extend over practically the entire steroid skeleton, and are not confined to the reacting group or to other oxygenated substituents. Furthermore, relatively minor structural modifications at a point distant from the centre of reaction may profoundly influence the affinity between substrate and enzyme.

No simple aliphatic or cyclic alcohols have been found to be substrates for these enzymes, whereas certain methyl decalols which are related to rings A and B of steroids, can undergo slow reaction. A comparison of the affinities for β-hydroxysteroid dehydrogenase of 3β-hydroxy-5a-androstan-17-one with those of the monofunctional steroid 3β-hydroxy-5a-androstane and cis-10-methyl-2-trans-decalol reveals that the polar oxygenated group at C(17)contributes very much less to the binding process than the non-polar hydrocarbon skeleton of the C and D rings (Table I). Similarly, 3a-hydroxysteroid dehydrogenase has a much higher affinity and velocity of oxidation for androsterone than for trans-10-methyl-2-trans decalol (Table I).

Table I. Velocity And Affinity Constants*

* Symbols are explained in the text. The velocities are expressed as percentages of the maximum oxidation velocity of androsterone (3a-hydroxysteroid dehydrogenase) and testosterone (3-hydroxysteroid dehydrogenase).

The stereochemistry of the A :B ring fusion at C(6) is of the utmost importance in controlling the affinity of both 3a- and β-hydroxysteroid dehydrogenases for their substrates. The relatively planar trans compounds are much more tightly bound and readily oxidized than the puckered A:B cis steroids. These observations do not consistently follow the predictions of conformation theory (Barton and Cookson, 1956). Thus, in agreement with the conformational analysis, 3a-hydroxy-5a-androstan-17-one (axial OH group) is more rapidly oxidized and at a lower concentration than 3a-hydroxy-5β-androstan-17-one (equatorial OH group) by 3a-hydroxysteroid dehydrogenase. However, contrary to conformational prediction, 3β-hydroxy-5a-androstan-17-one (equatorial OH group) is much more readily oxidized by p-hydroxysteroid dehydrogenase than 3β-hydroxy-5β-androstan-17-one (axial OH group) (Marcus, and Talalay, 1955; Talalay, 19576).

p-Hydroxysteroid dehydrogenase appears to have two reactive sites which are concerned respectively with the oxidation of 3β- and 17p-hydroxysteroids. The influences of the steric nature of the A :B ring fusion at C(5) on the affinity and oxidation velocity at the neighbouring 3β-hydroxyl group and at the much more distant 17p-hydroxyl group have been compared. Irrespective of whether the oxidation at C(3) or at C(17) is measured, the planar A :B trans compounds are much more readily oxidized than the puckered A:B cis compounds. Since changes in the stereochemistry of A:B ring fusion profoundly affect the geometry of the molecule, it would appear that the binding of the entire molecule is far more important in controlling enzyme affinity and oxidation rate than the conformation and hence the thermodynamic stability of the reacting group. This is borne out by the contrasting finding that the soluble 3a-hydroxysteroid dehydrogenase of liver reacts more rapidly with A:B cis compounds than with A:B trans compounds (Tomkins, 1956a and b; Hurlock and Talalay, 19586). It is also necessary to consider the possibility that the conformational forms which exist free in solution may undergo considerable modification upon attachment to such enzyme surfaces.