Adrenodoxin Reductase* PROPERTIES OF THE COMPLEXES OF REDUCED ENZYME WITH NADP+ AND NADPH

Anaerobic reduction of the flavoprotein adrenodoxin reductase with NADPH yields a spectrum with long wavelength absorbance, 750 nm and higher. No EPR signal is observed. This spectrum is produced by titration of oxidized adrenodoxin reductase with NADPH, or of dithionite-reduced adrenodoxin reductase with NADP+. Both titrations yield a sharp endpoint at 1 NADP(H) added per flavin. Reduction with other reductants, including dithionite, excess NADH, and catalytic NADP’ with an NADPH generating system, yields a typical fully reduced flavin spectrum, without long wavelength absorbance. The species formed on NADPH reduction appears to be a two-electron-containing complex, with a low dissociation constant, between reduced adrenodoxin reductase and NADP+, designated ARH,.NADP+. Titration of dithionite-reduced adrenodoxin reductase with NADPH also produces a distinctive spectrum, with a sharp endpoint at 1 NADPH added per reduced flavin, indicating formation of a four-electron-containing complex between reduced adrenodoxin reductase and NADPH. Titration of adrenodoxin reductase with NADH, instead of NADPH, provides a curved titration plot rather than the sharp break seen with NADPH, and permits calculation of a

reductase is calculated. It is concluded that the strong binding of NADP' to reduced adrenodoxin reductase provides the thermodynamic driving force for formation of a fully reduced flavoprotein form under conditions wherein incomplete reduction would otherwise be expected.
Stopped flow studies demonstrate that reduction of adrenodoxin reductase by equimolar NADPH to form the ARH,.NADP+ complex is first order (k = 28 s-l). When a large excess of NADPH is used, a second apparently first order process is observed (k = 4.25 s-l), which is interpreted as replacement of NADPH for NADP+ in the ARH,.NADP+ complex. Comparison of these rate constants to catalytic flavin turnover numbers for reduction of various oxidants by NADPH, suggests an ordered sequential mechanism in which reduction of oxidant is accomplished by the ARH,.NADP+ complex, followed by dissociation of NADP'.
The absolute dependence of NADPH-cytochrome c reduction on both adrenodoxin reductase and adrenodoxin is confirmed. Adrenodoxin reductase catalyzes 2,6'-dichlorophenolindophenol reduction by NADPH with two types of "intrinsic" rate constants: a relatively low one for adrenodoxin reductase itself, and a higher one for the previously reported 1:l complex between adrenodoxin reductase and adrenodoxin.
Adrenodoxin reductase (NADPH:adrenal ferredoxin reduc-consists of a single subunit, and contains one FAD per tase), (EC 1.6.7.1), from bovine adrenocortical mitochondria, molecule and no . It is part of the adrenal * A preliminary report of part of this work has been presented at the mitochondrial steroid hydroxylase system, which has been 1975 meeting of the Federation of American Societies for Experimental resolved into the flavoprotein (adrenodoxin reductase) an Biology (1) and at the Fifth International Conference on Flavins and Flavoproteins, San Francisco, 1975. This work was supported by iron-sulfur protein (adrenal ferredoxin or adrenodoxin), and National Institutes of Health Grant GM-21226. a heme protein or proteins (cytochrome P-450) (6-8). Cyto-*These studies were carried out in part during the tenure of an chrome P-450 has been shown to be the hydroxylating compo-Insurance Medical Scientist Scholarship from Prudential Insurance nent of this system, while flavoprotein and iron-sulfur protein co.
are considered to function in that sequence as an electron 4300 Adrenodoxin Reductase: Interaction with Pyridine Nucleotides transport chain, transferring reducing equivalents from 9).
The flavoprotein can accept two electrons (2), and has been shown to form a 1:l complex with the iron-sulfur protein, a one-electron acceptor (10,11). Adrenodoxin reductase is specific for NADPH rather than for NADH, with apparent K, values for these pyridine nucleotides of 1.82 FM and 5.56 mM, respectively (2, 12). Chu and Kimura (2) have described the spectrum of adrenodoxin reductase which had been anaerobically reduced with NADPH, and suggested that the low, broad absorbance extending from 505 to beyond 750 nm represents a "charge transfer complex" between NADP(H) and flavoprotein. Such a "complex" might be catalytically important and might contribute to the specificity for NADPH. This study was undertaken to determine the nature of the interaction between pyridine nucleotides and adrenodoxin reductase and to investigate the reduction of flavoprotein by this and other reductants. We have found that NADP+ forms a low dissociation constant complex with reduced adrenodoxin reductase, but binds relatively weakly to oxidized adrenodoxin reductase; and that this preferential binding shifts the potential of the flavoprotein by almost 100 mV.

Absorption
Spectra' of Reduced Adrenodoxin Reductase-We have confirmed the observation (2) that anaerobic reduction of adrenodoxin reductase with NADPH produces a spectrum with low, long wavelength absorbance (Fig. 1). The flavoprotein, reduced with 1.1 eq of NADPH, also shows residual absorbance at 450 and 377 nm, and isosbestic points for oxidized and NADPH-reduced enzyme at 507 and 352 nm. This spectrum does not resemble those of semiquinone forms hitherto described (27, 28), nor does it resemble a typical fully reduced spectrum (26,27).
'This calculation is valid for up to two electrons added because of the relatively positive potential of the ARH,.NADP+ complex (see 'Results"). Reduction of flavoprotein with reducing agents alternate to equimolar NADPH yielded a spectrum typical of fully reduced flavin, without long wavelength absorbance or residual peaks at 450 or 377 nm (Fig. 2). We designate this species the "fully reduced spectrum," or ARH,. Such a spectrum has been reported for the enzyme following dithionite reduction (I). We confirm this report, and have also observed the fully reduced spectrum on photoreduction by light-EDTA, by a 3-fold excess of NADH (not shown), and by using an NADPH-generating system in place of equimolar NADPH (Fig. 2). Reduction in the later experiment was accomplished using a catalytic amount of NADP' with isocitrate and isocitrate dehydrogenase to generate reduced pyridine nucleotide. No long wavelength absorbance or residual peaks at 450 or 377 nm were seen. An isosbestic point for oxidized and reduced enzyme was seen at 336 nm. It is thus apparent that the enzyme form produced by reduction with an NADPH-generating system is different than that obtained by reduction with NADPH.
No spectra resembling those of known semiquinone forms were observed at intermediate titration points of the experiment described in Fig. 2, nor were they seen during photoreduction of flavoprotein, titration of photoreduced flavoprotein with ferricyanide, or anaerobic reduction with puridine nucleotides. Because reduction and reoxidation in these experiments was by small increments, and the solutions were allowed several minutes between additions and measurement, detectable spectra of semiquinone forms would not be expected if equilibrium between oxidized and reduced forms and semiquinone form (adrenodoxin reductase,, + adrenodoxin reductase,,, = 2 adrenodoxin reductase.,), is far to the left, and if electron transfer between flavoprotein molecules occurs at an appreciable rate. A spectrum resembling that of neutral or "blue" semiquinone (26) was observed transiently, however, on reoxidation of NADPH-reduced flavoprotein by air (Fig. 3). The spectrum below 360 nm could not be evaluated because of the excess of NADPH.

EPR Spectra of Adrenodoxin
Reductase-Neither oxidized adrenodoxin reductase, nor adrenodoxin reductase reduced aerobically with excess dithionite or anaerobically with stoichiometric NADPH, showed EPR signals at g = 2.0 or in the entire EPR scan range, suggesting that semiquinone forms were absent in these experiments. Because of the close correla- Reoxidation of NADPH-reduced adrenodoxin reductase by air. 1, absorption spectrum of adrenodoxin reductase, reduced anaerobically with a 4-fold excess of NADPH. 2, absorption spectrum taken 3 min after opening the system to room air. 3, absorption spectrum taken 6 min after opening to room air. tion between EPR and visible spectra, no other concentrations of NADPH were tested, and the visible spectra were assumed to reflect the oxidation-reduction state of the flavoprotein.

Titrations of Adrenodoxin
Reductase-Anaerobic titration of oxidized adrenodoxin reductase with NADPH ( Fig. 4) showed decreasing absorbance at 450 nm and increasing absorbance at 700 nm, with a sharp break at 1 NADPH added per flavin, indicating a 1:l stoichiometry for formation of a product containing two electrons.
Results of titration of dithionite-reduced adrenodoxin reductase with NADP+ are shown in Fig. 5. Adrenodoxin reductase was first reduced with 1.1 eq of dithionite to produce the fully reduced spectrum (Fig. 5, curue 2). The reduced flavoprotein was then titrated with NADP', producing a spectrum identical to that obtained by reduction of oxidized flavoprotein with NADPH ( Fig. 5, curue 3). The course of this titration (inset, Fig. 5) again indicated a 1:l stoichiometry when observed at either 450 or 700 nm and again showed a sharp break point at 1 NADP' per FAD, indicating formation of a reduced species which had interacted with NADP'. It is thus clear that this form can be produced by treating either oxidized enzyme with NADPH, or reduced enzyme with NADP', and that this species represents a two-electron-containing complex between ,008 2 and pyridine nucleotide (see "Discussion"). We designate this form ARH,.NADP+.S Titration of reduced adrenodoxin reductase with NADPH was shown to yield an unusual spectrum (Fig. 6). Adrenodoxin reductase was first reduced with a slight excess of dithionite ( Fig. 6, curue 2) and then titrated with NADPH. At a 1:l stoichiometry spectral changes were observed over a broad wavelength range (Fig. 6, curue 3). Data from a titration followed at 570 nm are shown in the inset of Fig. 6. A sharp break in the titration plot at 1 NADPH per FADH, can be seen.
Thus, it appears that reduced adrenodoxin reductase can form a complex with NADPH as well as with NADP+.
Anaerobic titrations of oxidized adrenodoxin reductase with NADH yielded a curved titration plot (Fig. 7)  ful, since FAD was lost from the flavoprotein during the several days required to reach equilibrium.
However, the ultrafiltration method described under "Experimental Procedure" allowed determination of a dissociation constant within several hours with no appreciable loss of FAD. Binding data were plotted according to the method of Scatchard et al. (30), and are consistent with a &,. of 1.4 x lo-* M (Fig. 8). Approximately one binding site per flavin is indicated. This is in reasonable agreement with a previously published "K," for NADP+ of 5.32 x lo-' M (2), and in good agreement with a K, for NADP+ of 2 x lo-" M determined by us for ferricyanide reduction.' L ________ --__------------. Reductase-The titration of adrenodoxin reductase with NADH (Fig. 7) indicates that at 1 NADH added per flavin, there is only approximately a 70% reduction of the flavoprotein, i.e. an equilibrium mixture of oxidized and reduced forms. Fig. 9B shows the spectrum obtained by addition of equimolar NADH to flavoprotein, again showing no spectral evidence for an NAD(H).flavoprotein complex, and indicating incomplete reduction of flavoprotein. The lower curve (Fig. 9D) indicates the theoretical fully reduced spectrum calculated from the per cent residual absorbance after dithionite reduction. Addition of NADP+ to this equilibrium results in the spectrum of the ARH,.NADP+ species (Fig. 9C) by complex formation, the achievement of a fully reduced state.
Potential of Adrenodoxin Reductase-The curved titration plot obtained by reduction of adrenodoxin reductase by NADH (Fig. 7) indicates an equilibrium between oxidized and reduced forms of flavoprotein and NAD. From this equilibrium, the reduction potential for adrenodoxin reductase could be calculated from Equation 1, assuming an E, for NAD+/NADH of -0.316 V. The solid curve shown in Fig. 7, is a theoretical line generated by assuming a difference in reduction potential between NADH and adrenodoxin reductase of 0.025 V, i.e. a reduction potential for AR/ARH, of -0.291 V at pH 7.5. This is in reasonable agreement with a previously reported reduction potential of -0.274 V at pH 7.0, obtained by different methods (2).
The potential was also obtained for the adrenodoxin reductase . NADP + complex by titration of flavoprotein with NADPH in the presence of approximately equimolar safranine T as described under "Experimental Procedure." The titration observed at 520 nm was biphasic (Fig. lo), with a break at about 1 NADPH added per adrenodoxin reductase. Before this break there was only slight reduction of dye, indicating that flavin was being reduced to a large extent, followed by a more complete reduction of dye after the flavoprotein had been fully reduced. In two experiments, the potential of the flavoprotein in the presence of NADP+ was calculated from Equation 1 to be -0.198 (i.e. -0.198  of NADPH, confirming the interpretation of the reduction as a first order process. The second, slower phase showed decreasing absorbance at both 450 and 700 nm. This second phase can be interpreted as the replacement of NADPH for NADP+ in the reduced flavoprotein.NADP+ species to give the reduced flavoprotein' NADPH spectral species shown in Fig. 6, 3. Assuming dissociation of the ARH,.NADP+ species to be rate-limiting in this process, an apparent lz,,, of 4.25 s-1 was calculated. Catalytic Turnover of Adrenodoxin Reductase with NADPH-The rates of turnover of adrenodoxin reductase, calculated on a two-electron basis, obtained from a large number of catalytic experiments with various electron acceptors are summarized in Table 1. In the absence of adrenodoxin, reduction of K$FeCN, was most rapid while DPIP reduction was relatively slow. Cytochrome c reduction was intermediate in rate and was determined in the presente of adrenodoxin. The rate of DPIP reduction could be increased by adding adrenodoxin to the reaction mixture, as has been reported by Kimura and Chu (see Ref. 2). Fig. 11 shows the rate of DPIP reduction at various concentrations of adrenodoxin. The Sharp break at a 1:l ratio of adrenodoxin to adrenodoxin reductase indicates that adrenodoxin reductase can reduce DPIP with two types of "intrinsic" rate constants: a relatively low one for adrenodoxin reductase itself, and a higher one for the previously reported 1:l complex (ll) between adrenodoxin reductase and adrenodoxin.
Intermediate velocity points, then, represent variations in the molar ratio of adrenodoxin reductase to adrenodoxin reductase.adrenodoxin complex. In the case of cytochrome c reduction (where there is an absolute requirement for adrenodoxin (6)), the rate constant for reduction of the cytochrome by uncomplexed adrenodoxin reductase is essentially zero.

The interpretation
of the spectrum shown in Fig. 1 as a complex between NADP+ and reduced adrenodoxin reductase (ARH,) is supported by the following considerations. of adrenodoxin were added to 6.0 x lo-* M adrenodoxin reductase as indicated. Apparent turnover of flavin was calculated from initial rates of DPIP reduction as described under "Kesults." DPIP concentration was 3.9 x lOe' M.
1. The spectrum is not typical of any known form of free flavin (oxidized, reduced, or semiquinone), and the material does not provide an EPR signal. This suggests that it is not an unusual semiquinone form. The long wavelength absorbance, past 750 nm, supports the suggestion (2) that this species represents a charge transfer complex (31, 32). A similar spectrum has been observed as a transient intermediate ín stopped flow studies of reduction of ferredoxin NADP+-reductase with NADPH. 2. The spectrum is formed both by titration of oxidized flavoprotein with NADPH, and by titration of dithionitereduced flavoprotein with NADP+. Both titrations show a Sharp endpoint at 1 pyridine nucleotide added per flavin, indicating a complex with 1:l stoichiometry.
The sharpness of these titration endpoints suggests a "tight" complex, that is, a dissocia- complex formation. The possibility that NAD+ interacts with the flavoprotein without producing a spectral change cannot be rigorously ruled out. However, the lack of inhibition of enzymatic catalysis by NAD+ (2) makes this possibility unlikely.
4. The observed spectrum was due to complex formation with pyridine nucleotide, and did not arise as an inevitable mechanistic consequence of reduction of the enzyme by NADPH. This was shown by the absence of this species when reduction was accomplished by a catalytic quantity of NADP' continually reduced by an NADPH-generating system. 5. The spectrum of the complex is converted to that of fully reduced free flavoprotein by Neurospora NADase which acts catalytically to cleave the nicotinamide ring from NADP+. Thus, these studies clearly demonstrate a two-electron-containing complex (i.e. a "fully reduced" form) between pyridine nucleotide and flavoprotein.
The course of reduction of adrenodoxin reductase by NADPH and NADH is summarized in Fig. 12. Both pyridine nucleotides are assumed to have the same oxidation-reduction potential, -0.316 V (22, 23). The oxidation-reduction potential of adrenodoxin reductase (2), in the absence of complex formation, should dictate an equilibrium between oxidized and reduced forms of flavoprotein and pyridine nucleotide as is seen during the NADH titration (Fig. 7). Such an equilibrium is shown in Equation 2. NADH + AR=ARH, + NAD+ (2) At 1 NADH added per FAD, there is only approximately a 70% reduction of flavoprotein. However, at 1 NADPH added per FAD, there is 100% reduction. Since NADP+ binds only weakly to oxidized adrenodoxin reductase (K,,, = 1.4 x 10e5 M) (Fig.  8), this result is readily explained by the preferential complex formation of NADP+ with reduced adrenodoxin reductase, as is shown in Equations 3, 4, and 5. NADPH + AR = NADP+ + ARH, Data from Fig. 3 and Fig. 8 were replotted as percentage of fully reduced spectrum. binds tightly to reduced ARH,, effectively removing free reduced flavoprotein from the equilibrium (Equation 4). The net result is the sum of Equations 3 and 4 (Equation 5), in which addition of NADPH stoichiometric to adrenodoxin reductase results in essentially 100% formation of a fully reduced species. Thus, complex formation with NADP(H) provides the thermodynamic driving force for full reduction of the flavoprotein under conditions wherein the reduction potentials of the flavin and the free pyridine nucleotide would otherwise dictate an equilibrium between oxidized and reduced forms.
Equations 2 and 4 predict, moreover, that two electrons from NADH may be used to form a fully reduced flavoprotein species if NADP+ is added to bind ARHz and effectively remove it from the oxidation-reduction equilibrium.
The net result is the sum of Equations 2 and 4 shown in Equation 6. NADH + AR + NADP+ +NAD+ + ARH,.NADP+ The results of such an experiment, in which NADP+ was added to adrenodoxin reductase which had been incompletely reduced with 1 eq of NADH ( Fig. 9), confirm this prediction, and result in the paradoxical, but predictable situation in which addition of an oxidized substrate drives a reaction toward complete reduction.
The preferential binding of NADP+ to the reduced form of adrenodoxin reductase predicts that the oxidation-reduction potential of the flavoprotein will be made more positive in the presence of the ligand, NADP+ (34). This prediction is verified in these studies. A difference of approximately 94 mV is seen comparing NADP+-associated uersus unassociated flavoprotein. Since the dissociation constant for binding of NADP+ to oxidized adrenodoxin reductase has been established (K,,., = 1.4 x 10ms M), the dissociation constant for binding of NADP+ to reduced adrenodoxin reductase may be calculated from Equation 79 K, = K, x 1~-'wF"ml2.3Rn (7) where K, is the dissociation constant for binding of NADP+ to reduced adrenodoxin reductase, K,, that for binding to oxidized adrenodoxin reductase, and AE, the alteration in potential in the presence of NADP+. The K, calculated from this equation is 1.0 x 1Oe8 M. This tight binding of pyridine nucleotide to the reduced flavoprotein occurs not only with NADP', but also with NADPH, as indicated by the sharpness of the endpoint when titrating reduced reductase with NADPH (Fig. 6). The binding site of reduced adrenodoxin reductase appears to have specificity for 2'-monophosphopyridine nucleotides. Since the oxidized enzyme has a lower affinity for this grouping, one may conjecture that a conformational change in the binding site accompanies reduction. Adrenodoxin Reductase: Interaction with Z'yridine Nucleotides obtained for comparison with that of the ARH,.NADP+ complex.
Stopped flow studies indicate that the formation of the ARH,.NADP+ complex is rapid enough so that it could not be rate-limiting in the catalytic reduction of the three electron acceptors tested (Table I). Assuming complex formation as the rate-limiting step could account for a maximum flavin turnover of 28 to 30 s-l, well above observed catalytic turnovers. There was no evidence for the appearance of free ARH, during rapid reduction by NADPH, suggesting that the flavoprotein reacts with NADPH to form the ARH,.NADP+ complex as in Equation 5, rather than proceeding via the free reduced form. The lack of variation of reduction rate with a 30-fold alteration in NADPH concentration indicated, however, that reduction was proceeding through some intermediate(s), probably AR. NADPH, prior to formation of ARH,.NADP+.
Perhaps fluorescence quenching studies could demonstrate directly whether such a complex exists, as has been shown for microsomal NADPH .cytochrome c reductase and Escherichia coli NADPH . sulfite reductase (35).
The k,,, for dissociation of NADP+ from the ARH,.NADP+ complex, if rate-limiting, could account for a maximum turnover of 4.25 s-l. Dissociation of NADP+ from ARH,. NADP+ is, therefore, too slow to permit the observed rates for K,FeCN, and cytochrome c reduction, but is sufficiently rapid to permit DPIP reduction. These findings suggest that reduction of K,FeCN, and cytochrome c by adrenodoxin reductase (with adrenodoxin present in the later case) results from reaction of substrate with the ARH,.NADP+ form rather than with free ARH,. Reoxidation of the flavoprotein to a fully oxidized (or possibly a semiquinone) form should then allow the dissociation of NADP+' from flavoprotein, as suggested by the much higher dissociation constant for interaction of NADP+ and oxidized flavoprotein. Thus, these studies suggest an ordered sequential mechanism for reduction of K,FeCN, and cytochrome c as shown in Equation 8, where S represents either K,FeCN, or cytochrome c-adrenodoxin. The rate of DPIP reduction is sufficiently slow, however, so that DPIP reduction could proceed either by this proposed mechanism or by a ping-pong mechanism in which NADP' is dissociated from ARH, prior to reduction of DPIP. The enzyme has not yet been observed in a single catalytic turnover event, so the mechanism presented in Equation 8, although supported by present data, remains tentative. In addition, the present data do not deal directly with the function of the iron-sulfur protein, adrenodoxin, which probably serves as the physiological electron acceptor.
The present studies have shown that NADP+ forms a complex with reduced adrenodoxin reductase and modifies its potential by about 0.1 V, so that when NADPH is used as reductant complete reduction of flavoprotein occurs. NADH, having the same oxidation-reduction potential as NADPH, produces only partial reduction of flavoprotein. Complex formation with 2'-monophosphopyridine nucleotides could, therefore, represent a mechanism to insure efficient reduction of adrenodoxin reductase.