Energy-linked Mitochondrial Transhydrogenation from NADPH to NADP Analogs*

The mitochondrial energy-linked transhydrogenase enzyme catalyzes hydride ion transfer between NAD and NADP, of which the reaction NADH + NADP is slow in the absence of energy and is accelerated 10-fold or more when the mitochondrial membrane is ener- gized by ATP hydrolysis or respiration. The enzyme is a proton pump and effects proton translocation coupled to hydride ion transfer from NADPH to NAD (Earle, S. R., and Fisher, R. R. (1980) J. Biol. Chem 256,827-830). The present studies have shown that submitochondrial particles also catalyze transhydrogenation from NADPH to two NADP analogs, namely 3-acetyipyri-dine adenine dinucleotide phosphate (AcPyADP) and thionicotinamide adenine dinucleotide phosphate (thioNADP). Both reaction rates are greatly accelerated when the system is energized by ATP hydrolysis (inhibitable by uncouplers or rutamycin) or succinate oxidation (inhibitable by uncouplers or antimycin A). As in the case of NAD(H) e NADP(H) reactions, the transhydrogenations from NADPH to AcPyADP and thioNADP are inhibited by treatment of submitochondrial particles with trypsin or the arginyl residue mod- ifier, butanedione. The K,,, values of the above substrates and the V,, values under energy-linked con- ditions have been determined. The finding that the mitochondrial energy-linked enzyme catalyzes

The mitochondrial energy-linked transhydrogenase enzyme catalyzes hydride ion transfer between NAD and NADP, of which the reaction NADH + NADP is slow in the absence of energy and is accelerated 10-fold or more when the mitochondrial membrane is energized by ATP hydrolysis or respiration. The enzyme is a proton pump and effects proton translocation coupled to hydride ion transfer from NADPH to NAD ( The present studies have shown that submitochondrial particles also catalyze transhydrogenation from NADPH to two NADP analogs, namely 3-acetyipyridine adenine dinucleotide phosphate (AcPyADP) and thionicotinamide adenine dinucleotide phosphate (thioNADP). Both reaction rates are greatly accelerated when the system is energized by ATP hydrolysis (inhibitable by uncouplers or rutamycin) or succinate oxidation (inhibitable by uncouplers or antimycin A).
As in the case of NAD(H) e NADP(H) reactions, the transhydrogenations from NADPH to AcPyADP and thioNADP are inhibited by treatment of submitochondrial particles with trypsin or the arginyl residue modifier, butanedione. The K,,, values of the above substrates and the V , , values under energy-linked conditions have been determined.
The finding that the mitochondrial energy-linked transhydrogenase enzyme catalyzes transhydrogenation from NADPH to NADP analogs has revealed features regarding substrate site specificities and the effect of substrates on the directionality of proton translocation by the enzyme.
The mitochondrial energy-linked transhydrogenase enzyme (EC 1.6.1.1) catalyzes the transfer of hydride ion between NAD and NADP. Analogs of these substrates, in which the carbamyl group on position 3 of the pyridine ring is replaced with acetyl (AcPyAD' and AcPyADP) or thiocarbamyl (thioNAD and thioNADP) are also utilized by the enzyme, especially as hydride ion acceptors (1). It was discovered by Ernster and his colleagues that transhydrogenation from P To whom inquiries should be addressed.
NADH to NADP is greatly accelerated when the enzyme system (submitochondrial particles) is energized in the presence of ATP or an oxidizable substrate, while others showed that the reduction of NAD by NADPH was rapid and resulted in membrane energization (for reviews, see Refs. 2-4). Recent studies with the purified enzyme incorporated into liposomes have shown that transhydrogenase is a proton pump, NADPH + NAD transhydrogenation is coupled to proton translocation, and that the reverse reaction is driven by the membrane electrochemical potential of protons (5-7).
Previous studies, briefly reported elsewhere (€9, had shown that submitochondrial particles catalyzed NADPH + AcPyADP transhydrogenation, and that this reaction appeared to be energy-linked. The present manuscript documents in greater detail energy-driven transhydrogenation from NADPH to two NADP analogs ( i e . AcPyADP and thioNADP), as catalyzed by submitochondrial particles, and provides data indicating that these reaction are also catalyzed by the mitochondrial energy-linked transhydrogenase enzyme discussed above. The results have led to mechanistic clarifications regarding binding site specificities and substrate effects on the directionality of proton translocation by the transhydrogenase enzyme. Other important results obtained from study of energy-linked and non-energy-linked transhydrogenation from NADPH to AcPyADP and thioNADP are presented in the accompanying communication (1).

MATERIALS AND METHODS
SMP was prepared from beef heart essentially according to Low and Vallin (9) as described (l), and protein concentration was measured by the biuret method (10) in the presence of 1 mg of potassium deoxycholate/ml. Spectrophotometric studies were carried out with the Aminco DW-2a, the Aminco-Chance dual wavelength, and Cary 118 spectrophotometers.
Nicotinamide nucleotides were obtained from P-L-Biochemicals; ATP was from Boehringer; sodium succinate, trypsin inhibitor, and alcohol dehydrogenase were from Sigma; trypsin (grade B) was from CalBiochem; butanedione was from Aldrich Chemical Co.; rotenone was from S . B. Penick & Co.; and rutamycin was a gift from Eli Lilly.
Other chemicals used were reagent grade. Fig. 1 shows the reduction of AcPyADP by NADPH as catalyzed by SMP and energized by ATP hydrolysis (lefthand trace) or succinate oxidation (right-hand trace). The figure also shows inhibition of the ATP-driven transhydrogenation by rutamycin or CCCP, and the inhibition of the respiration-driven transhydrogenation by antimycin A or CCCP. Similar data were obtained for transhydrogenation from NADPH to thioNADP. Fig. 2, A and B, show slope and ordinate intercept replots of double reciprocal Lineweaver-Burk plots against the reciprocal concentrations of the fixed substrate (NADPH, Fig. 2A; AcPyADP, Fig. 2 B )  NADPH in the presence of 2 n m AcPyADP, 0.33 to 2 nm AcPyADP in the presence of 2 mM NADPH, 25 p~ rotenone, and 3.5 pg of rutamycin. The reactions were started by addition of 10 p1 of 2 M sodium succinate, and monitored at 400 minus 450 nm. Specific activities were calculated as nmol/min/mg of protein using a difference absorbance (AcPyADPH minus NADPH) value at the above wavelengths of 2.3 mM-l .cm". Data were then plotted in double reciprocal (Lineweaver-Burk) form, and slopes and ordinate intercepts were replotted as shown above. The ordinates shown are in nmol" . min. mg of protein for ordinate intercept values or the same divided by the reciprocal variable substrate concentration in mM for the slope values. results, K , and V,,, values for the transhydrogenation reaction NADPH + AcPyADP at 37°C were calculated to be as follows: KmNADrH = 0.4 mM, KmACPyADP = 1.0 mM, and V,,, = 37 nmol/min/mg of protein. ThioNADP was found to inhibit at high concentrations. Therefore, only apparent K , values for the NADPH + thioNADP transhydrogenation were calculated, as shown by the double reciprocal plots of Fig. 3.

RESULTS
These values were KmNADPH = 1.4 mM, Km'h'"NA"P = 38 PM, and V,,, = 10 nmol/min/mg of protein. Different preparations of SMP differ somewhat in their energy-linked transhydro-genase activity for the above reactions. Therefore, the Vmax values given above would be expected to change somewhat from one SMP preparation to another. Also, as will be seen in the accompanying communication (l), the rate of energylinked transhydrogenation from NADPH to AcPyADP and thioNADP is highly pH-dependent. In the pH range 6.5 to 8.5, the rate is highest at pH 7.5 and diminishes considerably on both the acid and the alkaline sides of pH 7.5. However, at pH 5 6.0, the rate increases sharply and does not seem to require membrane energization. The accompanying communication (1) also shows that, in the transhydrogenation reactions NADH + AcPyADP and NADH -+ thioNADP at pH 7.5, the K , values for both substrate pairs decrease severalfold in going from non-energy-linked to energy-linked conditions.
We have checked this point for the reaction NADPH * thioNADP, and have found that the apparent KmihioNADP under non-energy-linked conditions at pH 7.5 is about 0.5 mM, Le. more than 10-fold greater than the K , of this substrate at pH 7.5 under energy-linked conditions. These results are in excellent agreement with the interpretations offered in the accompanying communication (1) regarding the mechanisms of energy-linked and non-energy-linked transhydrogenations.
In genpral, submitochondrial particles catalyze not only transhydrogenations from NADPH to NADP analogs as shown above, but also the following types of transhydrogenation reactions (Table I): NAD(H) e NADP(H) and analogs, and NADH + NAD and analogs (Table I, Reactions 6 to 8).
Two separate enzymes are responsible for these reactions. The energy-linked transhydrogenase enzyme has been known to catalyze NAD(H) NADP(H) transhydrogenation, of which the rate of NADP reduction is accelerated 10-fold or more by membrane energization. The other enzyme is NADH dehydrogenase, which catalyzes NADH -+ NAD transhydrogenation at a rapid rate, and NADPH -+ NAD transhydrogenation at pH 5 6.0 and at a relatively slow rate. Our finding that submitochondrial particles catalyzed NADPH + AcPyADP and thioNADP transhydrogenations posed two important questions that needed to be clarified. These questions concerned (a) the validity of the assays for NADPH + AcPyADP and thioNADP transhydrogenations, and ( b ) the nature of the enzyme which catalyzed these reactions.  " Specific activity is expressed as nanomoles/min/mg of protein at 30°C. Reductions of AcPyAD(P), thioNAD(P), and PyrAldAD were measured, respectively, at 375, 400, and 365 nm.
Validity of the Assay-Regarding ( a ) , a serious complication would be the presence of NAD(H) either in the nucleotides used as substrate or bound to the added submitochondrial particles. The presence of NAD would result in the following transhydrogenation reactions, i.e. NADPH + NAD followed by the energy-requiring reaction NADH + NADP analogs. Both reactions are known to be catalyzed by the mitochondrial energy-linked transhydrogenase, and the sum would appear as though one is studying an energy-requiring transhydrogenation from NADPH to AcPyADP and thioNADP. That NADPH, NADP, and AcPyADP obtained from the source mentioned under "Materials and Methods" do not contain detectable amounts of NAD(H) was carefully tested and ascertained as in our previous studies (11). Both AcPyADP and thioNADP were also checked by thin layer chromatography and found not to contain detectable amounts of NAD. Submitochondrial particle preparations can contain, however, up to about 0.2 nmol of bound NAD/mg of protein (11, 12), and in the assays employed we could have added up to 0.11 PM NAD together with the particle preparations. This point and the possible presence of enzymically effective levels of NAD in the samples of AcPyADP and thioNADP were checked, therefore, under the assay conditions used. As seen in Fig. 4, left-handed trace, AcPyADP is reduced in a reactian mixture containing rotenone-treated SMP, succinate as an oxidizable substrate to provide energy, and NADPH as reductant. When succinate oxidation was interrupted by antimycin A, or an uncoupler was added, the reduction of AcPyADP was inhibited. The right-hand truce of Fig. 4 shows a similar experiment, except that the reductant NADPH was replaced by fl-hydroxybutyrate, the substrate of the NAD-linked D(-)-P-hydroxybutyrate dehydrogenase enzyme, which is bound to SMP. It is seen that there was no reduction of AcPyADP until NAD (2.67 PM) was also added. As seen in Fig. 4, this concentration of added NAD resulted in a rate of AcPyADP reduction close to that in the left-hand trace. Higher amounts of added NAD resulted in somewhat higher rates of AcPyADP reduction, and the rates at lower levels of added NAD are shown in Fig. 5. In this figure, the highest point shown represents the rate (AA/min) of the right-hand trace of Fig. 4, the lower rates are for smaller amounts of 9649 added NAD as shown, and the rate shown by the arrow at the lower left of the figure is at a concentration of added NAD equivalent to the amount of bound NAD that might have been added together with SMP (i.e. 0.2 nmol of bound NAD/ mg of SMP) in this experiment. These results indicated, therefore, that any contamination by NAD in the assay system used would have had to be less than 0.13 PM (the lowest level of added NAD in Fig. 5 5 (left). Energy-linked reduction of AcPyADP by DL-& hydroxybutyrate and increasing concentrations of added NAD as shown. The reaction conditions were the same as in Fig. 4. FIG. 6 (right). Energy-linked reduction of thioNADP by SMP in the presence of alcohol dehydrogenase and ethanol f added NAD. The reaction mixtures in 3 ml at 37'C contained 0.25 M sucrose, 50 mM potassium phosphate (pH 7.51, 5 p~ rotenone, 8 pg of rutamycin, 20 mM sodium succinate, 0.9 mg of SMP, 60 pg of crystalline alcohol dehydrogenase, 140 m~ ethanol, and 6 m~ hydrazine. The reaction mixtures represented by Traces 1 to 5 contained the following: Trace I , 0.5 m~ thioNADP and 50 p~ NAD Trace 2, same as Trace 1 in the absence of succinate; Trace 3, same as Trace 1 in the absence of NAD or NAD and succinate; Trace 4, same as Trace 1 in the absence of thioNADP; Trace 5, the reaction mixture in the absence of NAD and thioNADP. The reaction mixtures were incubated for 15 min at 37"C, and vigorously shaken at frequent intervals to reaerate. They were then placed in a boiling water bath for 2 min, and centrifuged in a clinical centrifuge at top speed to sediment the denatured protein. The clear supernatants were withdrawn, diluted 6-fold with water, and their absorbance spectra were recorded in a Cary 118 spectrophotometer. NAD elicited a measurable rate of AcPyADP reduction even though the K,,, for NAD in the P-hydroxybutyrate dehydrogenase reaction is about 1300 times higher (13). The next experiment (Fig. 6) was carried out with thioNADP in the presence of ethanol + alcohol dehydrogenase 2 added NAD as the source of reducing power. The amounts of ethanol and alcohol dehydrogenase were the same as those used by Teueira da Cruz et al. (14) in their NADH -+ NADP transhydrogenation experiments, and hydrazine was also added as done by these authors to trap the acetaldehyde produced. Trace I of Fig. 6 is the absorbance of the deproteinized reaction mixture after 15 min of incubation showing the energy-linked reduction of thioNADP (Amax at 395 n m ) in the presence of rotenone-treated SMP, succinate, ethanol, alcohol dehydrogenase, and NAD. Trace 2 is a parallel experiment in the absence of added succinate, which shows considerably less reduction of thioNADP a t 395 nm due to non-energy-linked transhydrogenation from NADH. Trace 3 is for the complete system in the absence of added NAD f succinate showing no  Fig. 1 in the presence of 0.9 mM NADPH and 0.35 m thioNADP. Each SMP addition was 5 p1 (125 detectable reduction of thioNADP, Trace 4 is the absorbance of the complete system in the absence of thioNADP showing the level of reduced NAD at 340 n m , and Trace 5 is the system in the absence of both NAD and thioNADP. These results show also that neither the particles nor thioNADP could have contained amounts of NAD large enough to effect appreciable reduction of thioNADP. Therefore, the data of Figs 4 to 6 allow the conclusion that submitochondrial particles do catalyze energy-linked NADPH + AcPyADP and thioNADP transhydrogenation directly and without intervention by NAD. The Nature of the Enzyme Catalyzing Transhydrogenation from NADPH to NADP Analogs-It was shown earlier that the mitochondrial energy-linked transhydrogenase contains an essential arginyl residue, highly susceptible to treatment of SMP with trypsin or the specific arginyl modifier, butanedione (15, see also Ref. 16). By contrast, electron transfer from NADH, NADPH, and succinate to oxygen were not so affected (15, 17). As seen in Table  11, SMP-catalyzed transhydrogenations involving NADP(H) and analogs are all strongly inhibited by treatment of the particles with trypsin, whereas transhydrogenations involving only NAD(H) and analogs are unaffected by this treatment. This is because the energy-linked transhydrogenase enzyme does not catalyze NADH + NAD transhydrogenation, and NADH dehydrogenase, which catalyzes this reaction (Table 11, Reactions 6 to 8), is not inhibited by treatment of SMP with trypsin or butanedione (see also Ref. 18). It might be added, also, that as far as is known these are the only two enzymes in SMP which catalyze nicotinamide nucleotide transhydrogenation.
Using trypsin-or butanedione-treated SMP, it was found that both energy-linked transhydrogenase reactions NADPH "+ AcPyADP and NADPH + thioNADP were inhibited (Figs. 7 and 8), while addition of untreated SMP to the same reaction mixtures elicited normal transhydrogenation.' In ad-As regards the effects of trypsin and butanedione on the ATPdriven reactions, the following points should be considered. Trypsin does not inhibit the ATPase activity of SMP (16) but butanedione does (19). The transhydrogenase enzyme is more sensitive to inhibition by butanedione than is ATPase (15, 19). Nevertheless, in Fig. 8, the inhibition of ATP-driven transhydrogenation in the presence of butanedione-treated SMP should be considered to be due to inhibition of both the transhvdroeenase and the ATPase. activity was not inhibited. , " dition, we have shown elsewhere (18) that NADH dehydrogenase purified from mitochondria does not catalyze NADPH -+ AcPyADP transhydrogenation at detectable rates. These results suggest, therefore, that the above reactions are catalyzed by the same mitochondrial enzyme which catalyzes energy-linked NADH + NADP transhydrogenation. In confirmation of this conclusion, Anderson and Fisher3 have shown that antibody to the purified transhydrogenase enzyme inhibits the NADPH -+ AcPyADP transhydrogenase activity of submitochondrial particles.