The precise specificity and multiplicity of the particulate transhydrogenases of animal tissues remains to be clarified through extensive purification of the individual proteins.

3. Transhydrogenations Catalyzed By Pyridine Nucleotide-Linked Dehydrogenases

It is obvious that equations for the transfer of hydrogen from one form of pyridine nucleotide to another can be written for any pyridine nucleotide-linked dehydrogenase (which reacts with the requisite nucleotides) if an appropriate substrate is present. However, Colowick et al. (20) were unable to mimic the action of the Pseudomonas and mitochondrial transhydrogenases with various dehydrogenases which reacted with both DPN and TPN either in the presence or absence of their substrates. Similar experiments by Stern (106), who studied a β-hydroxyacyl-CoA dehydrogenase with dual nucleotide specificity, gave the same type of negative results. Hagerman and Villee (39, 40) have stated that a mixture of DPN-specific and TPN-specific testosterone dehydrogenases isolated from guinea pig liver failed to catalyze the transfer of hydrogen from TPNH to DPN in the presence of testosterone. Recently, Holzer and Schneider (52) reported that hydrogen transfer from TPNH (generated in situ by the action of glucose-6-phosphate dehydrogenase) to DPN occurred in the presence of lactic and glutamic dehydrogenases and their appropriate substrates. The experimental conditions under which this transhydrogenation occurred were markedly different from those used for the study of pyridine nucleotide transhydrogenases, and for the steroid-mediated hydrogen transfers catalyzed by the hydroxysteroid dehydrogenases. The total amount of hydrogen transferred from TPNH to DPN in the experiments of Holzer and Schneider (52) was very small in comparison with the concentration of carrier molecules added. Thus, in the glutamic dehydrogenase system, 0.07 umole of TPN was added originally and reduced to completion by the addition of 20 umoles glucose-6-phosphate with glucose-6-phosphate dehydrogenase. a-Ketoglutarate (5 umoles) was then added (the enzyme preparation contained unspecified-but probably large-amounts of ammonium sulfate) and the reaction allowed to come into equilibrium. Following the further addition of 0.25 umole of DPN, only 0.02 umole of DPNH was formed when any further reaction ceased. In the lactic dehydrogenase reaction, 0.3 umole of TPNH (continuously reduced) transferred in 20 minutes, about 0.01 to 0.02 [imoles of hydrogen to DPN (8 umoles) in the presence of 3.4 umoles of pyruvate, and 0.4 mg. of crystalline lactic dehydrogenase. If we assume that this preparation has a turnover number of 36,000 moles per minute per 100,000 gm. of protein (8), it can be calculated that the rate of transhydrogenation under these conditions is 0.01% of the rate of reduction of pyruvate by DPNH.

It is noteworthy that the rate of reaction of lactic dehydrogenase with TPNH is very much less than with DPNH, especially at low pyruvate concentrations (62) and at hydrogen ion concentrations greater than pH 6.0. Navazio et al. (92) found that at pH 6.0, TPNH reduced pyruvate (3.3 X 10-3M) at 2.5%, and at pH 7.6 at less than 0.1% the rate of DPNH. Thus the extremely inefficient catalysis of hydrogen transfer from TPNH to DPN by lactic dehydrogenase results from: (a) an unfavorable equilibrium constant (KH = 4 X 10-12M), (b) the great discrepancy between the reaction rates with DPN and TPN, (c) the high Michaelis constant for pyruvate with TPNH (62).

Bacterial and mitochondrial transhydrogenases (which do not require an added carrier), and hydroxysteroid dehydrogenases (which transhydrogenate in the presence of as little as 10 -8M of an appropriate steroid), may transport relatively enormous amounts of hydrogen from one pyridine nucleotide to another. As we have emphasized previously (112, 117), the special features of hydroxysteroid dehydrogenases which make them efficient transhydrogenating catalysts at pH close to neutrality in the presence of low concentrations of steroids are: (a) high affinities for steroids, (b) favorable equilibrium constants for the oxidoreductions of the steroids by pyridine nucleotides, (c) rates of reaction with donor and acceptor nucleotides which are of the same order of magnitude.

The 17β-hydroxysteroid dehydrogenase of human placenta binds estradiol-176 so firmly that we have found it virtually impossible to remove this steroid from enzyme preparations which have been prepared in its presence. Such preparations catalyze transhydrogenation without the further addition of estradiol-17β. In this respect they resemble the transhydrogenases of bacteria and of animal mitochondria.

Kaplan and co-workers (64, 102) have visualized the transfer of hydrogen to occur between donor and acceptor nucleotides simultaneously bound to the enzyme system without the mediation of an additional carrier. The existence of auxiliary carrier molecules on these enzymes is still within the bounds of possibility. At least in the case of the Pseudomonas transhydrogenase, such carriers, if they exist, are unlikely to be steroidal in nature because of the absence of steroids from bacteria (109).

4. Transhydrogenases And Metabolic Regulation

Kaplan and his associates (67) have proposed a regulatory function for the pyridine nucleotide transhydrogenases of mitochondria which catalyze the transfer of hydrogen from TPNH to DPN. There is considerable evidence that the oxidation of TPNH by mitochondria (which is more sluggish than the oxidation of DPNH) is not coupled with the synthesis of adenosine triphosphate, unless hydrogen is first transferred to DPN by the action of transhydrogenases (67, 75, 124), although there is not unanimous agreement on this matter (61). In this way, mitochondrial transhydrogenases would facilitate the capture of energy from the oxidation of TPNH and could divert hydrogen flowing through this nucleotide from other metabolic pathways. There is a growing realization that there are many important biological reductive syntheses for which TPNH, but not DPNH, serves as a specific hydrogen donor. These include the fabrication of fatty acids and steroids, hydroxylations, reductive carboxylations, and the synthesis of ascorbic acid (25, 53). In most animal cells the ratio TPNH : TPN is large, and greater than the quotient DPNH : DPN (25). That TPN is present intracellularly in predominantly its reduced form possibly reflects that its rate of oxidation is much slower than its rate of formation, particularly by the dehydrogenases of the pentose phosphate cycle (25, 53). Indeed, there is rather convincing circumstantial evidence, obtained from studies with tissue slices, that the operation of the pentose phosphate cycle is steered by reactions which reoxidize TPNH (16, 44, 81). Thus the action of pyridine nucleotide transhydrogenases may affect many metabolic processes by controlling the disposal of TPNH. Another function of mitochondrial transhydrogenases may be to catalyze the transfer of hydrogen from extramitochondrial DPNH to an intramitochondrial form of DPN, which Lehninger and his co-workers (75) implicate as an intermediate in oxidative phosphorylation.

We have suggested (112, 117) that certain mammalian hydroxysteroid dehydrogenases may serve a regulatory function by catalyzing transhydrogena-tions between TPNH and DPN in the presence of very low concentrations of steroids. The mammalian hydroxysteroid dehydrogenases with dual pyridine nucleotide specificity are localized either in the soluble cell sap, or in the microsomal fraction of the cell. Whether extramitochondrial, steroid-mediated transhydrogenations are of regulatory significance cannot be decided at present. From the very meager experimental data available, it is virtually impossible to make meaningful comparisons between the rates of transhydrogenation catalyzed by steroid-hydroxysteroid dehydrogenase systems on the one hand, and by mitochondrial transhydrogenases and dual nucleotide specific dehydrogenases for nonsteroidal substrates on the other. The rather large changes in the oxidation of many substrates, and in the utilization of pyruvate, by slices of human placenta and endometrium which Villee and Hagerman (36, 37, 125) found to result from the in vitro addition of estradiol-176 suggest that transhydrogenations mediated by this steroid are rate-limiting steps in the oxidative metabolism of intact preparations of these cells. In a recent publication, Stein and Kaplan (104) have voiced strong objection to any regulatory function of transhydrogenations catalyzed by 3a-hydroxysteroid dehydrogenase in rat liver. They concluded from experiments with unfractionated preparations of liver cell sap, and with liver mitochondria suspended in digitonin, that the rates of hydrogen transfer from TPNH to the acetylpyridine analog of DPN by the cell sap are negligible compared with those catalyzed by the mitochondria. Stein and Kaplan were unable to demonstrate any influence of androsterone (a substrate for liver 3a-hydroxysteroid dehydrogenase) on the activity of soluble transhydrogenase(s), and conclude with the statement: "We suggest that the term pyridine nucleotide transhydrogenase be reserved for enzymes exhibiting these activities as quantitatively primary functions, and which, by virtue of concentration in unmodified tissue fractions or high purification, possess a reasonably high specific activity." It is difficult to see the relevance of this purely ad hoc, and questionable, definition of a pyridine nucleotide transhydrogenase to the unequivocal experimental demonstration of the transhydrogenase activity of liver 3a-hydroxysteroid dehydrogenase (55). Although this activity is admittedly quite feeble, the experiments of Stein and Kaplan do not, in our opinion, eliminate the possibility of a regulatory function of even this hydroxysteroid dehydrogenase in cellular metabolism, let alone that of other, more active, enzymes of this class. The negative findings of Stein and Kaplan were only to be expected, since they attempted to measure steroid-mediated hydrogen transfers from high concentrations of TPNH to the acetylpyridine analog of DPN. Technical considerations (55, 112) make it clear that this is not feasible for either the liver 3a- or the placental 17β-hydroxysteroid dehydrogeneses (Fig. 2). Furthermore, one may question their experiments with unfractionated tissue extracts on two main grounds. First, as we have shown experimentally, bound pyridine nucleotides and interfering enzymes in crude liver extracts make the demonstration of androsterone-sensitive transhydrogenations extremely difficult (55). Secondly, one may doubt whether crude, unfractioned preparations of the soluble fraction of liver, isolated from sexually mature animals, would be devoid of endogenous steroid hormones.