Vennesland, Westheimer and their colleagues have made the fundamental observation that pyridine nucleotide-linked dehydrogenases promote the " direct transfer" of hydrogen between substrate and position 4 of the pyridine moiety of the nucleotide (Vennesland and Westheimer, 1954; Vennesland, 1956, 1958a; Pullman, San Pietro and Colowick, 1954). Furthermore, this transfer is stereospecific with respect to both pyridine nucleotide and substrate, as was demonstrated with the aid of deuterium labelling. The experimental evidence for "direct transfer" relies upon the tracing of istotopically labelled hydrogen between substrate and pyridine nucleotide, and upon the immunity to exchange with the aqueous medium of the hydrogen undergoing transfer in the reaction. Mediation of the transfer through the enzyme is a possibility, but not a necessity, since even with certain model reactions which occur in homogeneous systems, direct hydrogen transfer has been observed (Abeles, Hutton and Westheimer, 1957; Abeles and Westheimer, 1958). Although the precise nature of the hydrogen species undergoing transfer has not been identified, it appears probable that a hydride group (H2-) is transferred in these two electron oxidations (Abeles, Hutton and Westheimer, 1957).

When one of the two hydrogens at position 4 of the nicotinamide ring of DPNH is labelled isotopically, two diastereomers of the reduced nucleotide may be distinguished. By tracer studies with deuterium, the stereospecificities of a large number of di-and triphosphopyridine nucleotide-linked dehydrogenases have been established (Vennesland, 1958a). They belong to one or the other of two groups, those which utilize the same side of the pyridine ring as alcohol dehydrogenase (side 1 or a)* and those which have opposite stereospecificity (side II or β).

The stereospecificity of addition and removal of hydrogen is not confined to the pyridine nucleotide but also extends to the substrate. Whereas the asymmetric handling of hydrogen is readily recognized in the reduction of pyruvate to D- or L-lactate by different enzymes (Kaplan, 1959, this volume, p. 37), it may also be detected in symmetrical products by appropriate isotopic labelling. Yeast alcohol dehydrogenase reduces acetaldehyde by stereospecific transfer of hydrogen from DPNH. In the course of the oxidation of ethanol by this enzyme, the same hydrogen is removed from the a-carbon atom. This has made it possible to prepare enzymically the two enantiomorphs of a-monodeutero-ethanol (Loewus, Westheimer and Vennesland, 1953; Levy, Loewus and Vennesland, 1957).

Enzymic discrimination of "identical" groups. The capacity of enzymes to discriminate between two apparently identical methylene hydrogens illustrates perhaps the simplest type of a more general aspect of enzyme stereospecificity. In certain enzyme-catalysed reactions involving symmetrical molecules, the stereospecificity extends to the discrimination of apparently identical groups. The classic example is citric acid, in which the two -CH2COOH groups which appear chemically identical are differentiated by their origin and reactivity in enzymic systems. This apparent paradox was examined by Ogston (1948), who pointed out that although citric acid was symmetrical in space, it could no longer be regarded as symmetrical when attached to an asymmetric enzyme surface by at least three points.

*The designations, side I and side II, for the diastereomers of DPN-nicotinamide-4-d are used in preference to a and β to avoid confusion with steroid nomenclature (cf. Vennesland, 1958a).

Schwartz and Carter (1954) reconsidered this problem in a more general and formal manner. They pointed out that a molecule containing a carbon atom bearing two similar and two dissimilar groups has certain unique properties not previously recognized. They suggested the term miio-carbon atom for this type of carbon atom. Whereas such molecules have a plane of symmetry, the half molecules so visualized are not superimposable and therefore are not identical. Hence, the apparently identical groups attached to a meso-carbon atom are intrinsically different for an asymmetrical reagent, such as an enzyme. Schwartz and Carter (1954) emphasized that this difference did not depend upon attachment to a surface, and demonstrated experimentally the relative stereospecificity of a chemical reaction between an asymmetrical molecule and one containing a miro-carbon atom.

Stereospecificity of hydroxysteroid dehydrogenases for pyridine nucleotides. Bacterial β-hydroxysteroid dehydrogenase was among the first enzymes found to exhibit direct transfer to and from side II of DPN (Talalay, Loewus and Vennesland, 1955). This conclusion was based upon studies of the reduction of 4-androstene-3,17-dione to testosterone by purified β-hydroxysteroid dehydrogenase under various conditions: (1) if the reaction was carried out with DPNH in D2O, no stably bound deuterium was found in the testosterone; (2) when DPND prepared by the reduction of DPN+ with 1,1-dideutero-ethanol and yeast alcohol dehydrogenase and hence bearing the D on side I was the reductant, likewise no isotope was found in the testosterone. These results suggested that β-hydroxysteroid dehydrogenase had stereospecificity for the opposite side of the pyridine nucleotide ring to that used by alcohol dehydrogenase. Accordingly, (3) DPND carrying D on side II was prepared by a somewhat indirect procedure (which alone was available at that time) and deuterium was incorporated into testosterone from this reductant during the conversion of 4-androstene-3,17-dione to testosterone.

More recently, simpler methods for studying steric specificity have been applied to 3a- and p-hydroxysteroid dehydrogenases. Nicotinamide-4-t-DPN+ has been prepared by decomposing the complex which forms between DPN+ and basic cyanide in T2O enriched water according to the procedure of San Pietro (1955). The purified DPN+-4-t was then used in the following conversions (Jarabak, 1959, unpublished): androsterone + DPN+ → androstane-3,17-dione + DPNH + H+ (3a-hydroxysteroid dehydrogenase) epi-androsterone + DPN+ → androstane-3,17-dione+DPNH+H+ (β-hydroxysteroid dehydrogenase)

An excess of steroid and limiting amounts of DPN were used in each case, and the reaction conditions were so arranged as to assure complete reduction of the pyridine nucleotide. The steroid dehydrogenases were then inactivated by heating, and the DPNH reoxidized completely by two procedures in separate experiments: (1) acetaldehyde and crystalline yeast alcohol dehydrogenase (side I stereospecificity) and (2) a-ketoglutarate, NH3 and liver glutamic dehydrogenase (side II stereospecificity). The oxidized DPN was in each case hydrolysed at the nicotin-amide-ribose linkage, and the nicotinamide isolated after appropriate dilution with carrier, and its radioactivity determined in a liquid scintillation counter. Similar results were obtained by Jarabak (1959, unpublished) in experiments with both 3a-and p-hydroxysteroid dehydrogenases in the oxidations at C(3) of the steroids. The nicotinamide obtained from reoxidation of the pyridine nucleotide with alcohol dehydrogenase retained less than 5 per cent of the radioactivity of the original DPN+-4-t. In contrast, the glutamic dehydrogenase oxidation product retained essentially all the original radioactivity (see Table II). It may be concluded that both 3a- and p-hydroxysteroid dehydrogenases have the same stereospecificity for DPN as glutamic dehydrogenase (side II).

The possible significance of the steric course of pyridine nucleotide-linked oxidations has been discussed by Levy and Vennesland (1957). No rational basis for the existence of two groups of enzymes with opposite stereospecificities has become apparent. It has been suggested that the metabolic coupling of enzymes of opposite pyridine nucleotide stereospecificities is facilitated and that this would provide a mechanism for the selection between alternative pathways involving protein-bound pyridine nucleotides (Levy and Vennesland, 1957). Since a- and p-hydroxysteroid dehydrogenases attack diastereomeric steroids, but utilize the same side of the pyridine nucleotide ring, it may be surmised that no relationship exists between the optical structure of the substrate and the configuration of the reacting hydrogen of the pyridine nucleotide. Much evidence points to the presence of multiple enzyme-binding sites on both steroid and pyridine nucleotide. For reasons of geometry, the attachment of androsterone and epi-androsterone to their respective dehydrogenases is presumably quite different in order to facilitate the introduction or removal of the 3β-or 3a-hydrogens respectively, yet these reactions utilize the same side of the pyridine nucleotide ring.

Table II. Pyridine Nucleotide Stereospecificity Op A- And P-Hydroxysteroid Dehydrogenases (Jarabak, 1959, Unpublished)

The DPN+-4-t had a specific activity of 4,270 c.p.m. per nmole.

ADH = crystalline yeast alcohol dehydrogenase; GDH = crystalline liver glutamic dehydrogenase.