Resolution of Enzymes Catalyzing Energy-linked Tramhydrogenation

1. A highly purified transhydrogenase factor (THi) has been isolated from Rhodospirillum rubrum chromatophores by washing the membranes first with buffer containing 10 pnn NADP+ under conditions in which TH1 remains bound, then with buffer lacking NADPf resulting in the dissociation of TH1 from the membrane. 2. Purified TH1 stimulated the energy-linked reduction of NADP+ by NADH, the nonenergy-linked reduction of AcPyAD+ by NADPH, and the reduction of AcPyADf by NADH in the presence of chromatophores resolved with respect to these activities. THi did not catalyze these reactions independently of these resolved chromatophores. 3. The energy-linked reduction of NADP+ by NADH has been demonstrated to occur by the transfer of hydrogen without prior exchange to the medium from the 4A (nicotinamide) locus of NADH to the 4B (nicotinamide) locus of NADPH. This stereospecificity is identical with that found for the analogous mitochondrial reaction. 4. It is concluded that THi functions in the transfer of hydrogen between NAD(H) and NADP(H) in both the energyand nonenergy-linked transhydrogenase reactions.

the transhydrogenase factor (THJ, and the insoluble chromatophore membrane (2,3). The complex formed between THI and the membrane is stabilized by the presence of low concentrations of either NADPf or NADPH (3). This complex is dissociated in the presence of high concentrations of NADPH or NADH.
In addition to the essentially irreversible energy-linked transhydrogenation, both mitochondria (4) and chromatophores (1) catalyze a nonenergy-linked transhydrogenation which can be described by Equation 1.
Kaufman and Kaplan (5) have described the solubilization of a lipoprotein by detergent treatment of ox heart mitochondria which catalyzed this reaction.
Antibodies produced against the "soluble" enzyme inhibited both the energy-linked and nonenergy-linked transhydrogenase reactions of submitochondrial particles.
The membrane-bound enzyme (6) as well as the solubilized enzyme (7) transfers hydrogen in a stereospecific manner from the 4A locus of NADH to the 4B locus of NADPH.
When it was discovered that crude THI was able to catalyze the reversible nonenergy-linked reduction of NADP+ and NAD+ analogues by NADH (2), the possibility presented itself that THI might itself be a transhydrogenase enzyme.
In order to examine this possibility, it was felt necessary to prepare a highly purified TH,.
It is the purpose of this paper to describe a unique method for the isolation of a purified preparation of TH1 from chromatophore membranes and to describe some of its properties.
In addition, the stereospecificity of the energy-linked R. rubrum transhydrogenase was studied and was found to occur by direct hydrogen transfer and with the same stereospecificity as the mitochondrial reaction.
The rate of AcPyAD+ reduction was calculated from the increase in absorption at 375 nm assuming a millimolar extinction coefficient of 5.1 (8). The reaction was initiated by the addition of chromatophores or Crparticles containing about 20 pg of BChl.
The rate of AcPyAD+ reduction was measured as outlined above following the addition of chromatophores or CT-particles. Assay for Tritium Incorporation into [4B-aH]-NADPH-The concentration and specific activity of [4B-3H]-NADPH was determined after oxidation with glutathione and glutathione reductase according to the procedure of Lee et al. (6). In this method, hydrogen from the 4B locus of NADPH exchanges with the medium water.
The quantity of tritium recovered in the medium water following treatment of the sample with glutathione reductase minus the quantity of tritium in the water prior to the oxidation of NADPH was taken as a measure of the amount of tritium in [4B-aH]-NADPH.
Water was isolated from reaction mixtures by the following method.
Of the reaction mixture, 0.4 ml was transferred to the side bulb of a Thunberg tube and frozen in a methyl Cellosolve bath cooled with Dry Ice. The tube was evacuated and sealed, and its body was immersed in the Dry Ice bath. When the water had distilled from the side bulb into the body of the tube its tritium content was determined (0.3.ml samples) by liquid scintillation counting. Preparation of [4A-aH]-NADH-Of the 4-aH-labeled NADf (specific activity 1.24 Ci per mmole), 5.36 pg were reduced in a medium (1.28 ml) containing 31 mM Tris-HCl buffer, pH 8, 88 InM sucrose, 7.8 mu unlabeled NAD+, 47 mM n-sodium glutamate, 280 InM hydrazine hydrate, and 1 mg of glutamate dehydrogenase (ammonium sulfate free). The reaction mixture was incubated at 20" for 25 min and the reaction terminated by heating for 3 min in a boiling water bath.
Distilled water was added to bring the volume to 5 ml and the mixture was centrifuged to remove denaturated protein. The reaction mixture was incubated at 21" for 15 min and the reaction terminated by heating for 3 min in a boiling water bath. The specific activity of [4B-3H]-NADH was evaluated as described for [4A-aH]-NADH.
Separation of Pyridine Nucleotides-A neutralized solution containing a mixture of pyridine nucleotides (12 nmoles total) was placed on a column (0.65 X 10 cm) of DEAE-Sephadex A-25. The column was washed with 50 ml of distilled water and NADf, NADH, NADP+, and NADPH were eluted stepwise with 0.04 M, 0.09 M, 0.13 M, and 0.25 M Tris-HCl buffer, pH 7, respectively (9).
Liquid Scintiillation Counting--This was done with a Nuclear Chicago Mark I counter.
The counting efficiency was estimated to be 35yc. Defcnition of Units-A unit of TH1 activity is defined as the amount of protein required to stimulate the ATP-dependent transhydrogenase reaction rate in CT-particles by 1 pmole of NADP+ reduced per mg of BChl per min. Specific activity is expressed as units per mg of protein.
ReagentsSources of reagents not given previously (

Preparation of Purijied TH, from Chromatophores
The finding that NADF promoted the binding of TH, to Crparticles and stabilized the TH1 membrane complex of chromatophores during washing (3) led to the development of a procedure for the preparation of THI of high purity.
This method involves (a) removal of loosely bound membrane proteins excluding TH1, by washing chromatophores with buffer containing NADP+, (b) solubilization of THI by washing with buffer lacking NADPt, and (c) concentration of THI by precipitation with ammonium sulfate. All steps were carried out at O-4".
Step 1: Disruption of Cells-R. rubrum cells (40 g, wet wt) were suspended in 115 ml of 0.1 M Tris-HCl buffer, pH 8, 10% sucrose containing 10 pM NADP+, 2 mM MgC&, and transferred to a 200-ml stainless steel beaker chilled in ice. The cells were sonicated for 1 min with a Bronson sonic oscillator (J-32 probe) and the sonic extract was centrifuged for 10 min at 18,000 x g. The supernatant solution was decanted and centrifuged a second time for 30 min at 150,000 x g. The chromatophore pellets from this second centrifugation were suspended by homogenization in 15 ml of the above buffer and the supernatant solution was saved for the preparation of crude TH1 (see below).
Step d: Pyridine Nucleotide Wash-The 15 ml of suspended chromatophores (Step 1) were diluted to 165 ml with the same buffer, slowly stirred for 5 min, and then sedimented by centrifugation at 150,000 x g for 30 min. The supernatant solution was discarded.
Step 3: Solubilization oj TH1--The pellets (washed chromatophores) obtained from Step 2 were suspended in 95 ml of 0.1 M Tris-HCl buffer, pH 8, 10% sucrose containing 0.001 M dithiothreitol and stirred for 5 min prior to sedimentation by centrifugation at 150,000 x g for 30 min. The pellets were suspended again in 95 ml of the above buffer to solubilize the remaining membrane-bound TH,. After repeating the stirring and centrifugation procedure, the supernatant solutions from the first and second washes were combined and treated as described in step 4.
Step 4: Concentration oj Solubilized TH, by Ammonium Sul-fate Precipitation--To the combined wash supernatant solutions, solid ammonium sulfate was added with stirring to a final concentration of 70% of saturation.
Stirring was continued for 30 min after the salt completely dissolved.
The precipitate was sedimented by centrifugation for 15 min at 150,000 x g. The film-like precipitate was dissolved in 3 ml of 0.01 M Tris-HCl buffer, pH 8, 1% sucrose containing 0.15 mg of dithiothreitol per ml. This solution was centrifuged for 60 min at 150,000 x g. The colorless supernatant solution containing purified THr was stable for at least a week when stored at CM".
Notes on Preparation of Purified TH, During the development of the above procedure, it was found that TH, was very unstable in dilute protein solutions.
Dithiothreitol was found to stabilize the factor and for this reason it is included in the wash medium of Step 3. However, dithiothreitol was not able to reactivate TH1 which had been inactivated by dilution.
Bovine serum albumin (at 0.5 mg per ml) has been successfully employed in place of dithiothreitol; however, its use prevents the accurate determination of THr specific activity. The ease with which THr is dissociated from the membrane in Step 3 was dependent upon the concentration of NADP+ in the pyridine nucleotide wash step. When NADP+ concentrations in excess of 10 PM were used in Step 2, solubilization of the factor during Step 3 required as many as five or six washings to effect complete solubiliaation.
Thus, the procedure became unduly laborious and the subsequent precipitation of THr by ammonium sulfate in Step 4 was hampered.
If less than 10 PM NADP+ was used in Step 2, significant amounts of membrane-bound TH, could be solubilized prior to Step 3.
Preparation of Crude THr--The supernatant solution from Step 1 was diluted with an equal volume of 0.1 M Tris-HCl buffer, pH 8, 1% sucrose containing 0.001 M dithiothreitol and centrifuged at 150,000 x g for 60 min. THr was precipitated by bringing the high speed supernatant solution to 70% saturation with solid ammonium sulfate. The solution was stirred for an additional 30 min after the salt had dissolved and then centrifuged at 18,800 x g for 10 min to sediment the THr.
The supernatant solution was decanted and the interior of the centrifuge tube wiped free of excess ammonium sulfate solution.
The pellets were then suspended in a minimal volume of the above buffer and centrifuged at 150,000 x g for 60 min. The deep orange supernatant solution is termed crude THr.
It was shown in the previous paper (3) that THr solubilized directly from unwashed chromatophore membranes had a specific activity only 5-fold greater than that of THr prepared from the cell-free extract.
The relative specific activities of the crude THi and the purified TH,, prepared as described above, were determined by their ability to stimulate the ATP-dependent transhydrogenase reaction in C&-particles.
As can be seen in Table I, the specific activity of purified THr was over 1200 times that found in the crude supernatant fraction.
Precipitation of THr from the crude supernatant fraction with ammonium sulfate resulted in a preparation with a specific activity 'I-fold greater than that of the crude supernatant fraction.
The observed increase in total units obtained by this treatment was as great as the increase in specific activity.
This activation can be explained if the extract, unlike the ammonium sulfate fraction, contained a component which inhibits the interaction of THr with the chromatophore membrane.
It is obvious that the washing procedure which allows for the  isolation of THI directly from washed chromatophores resulted in a factor preparation of considerable purity. Two preliminary experiments with polyacrylamide disc gel electrophoresis showed the purified THr to be composed of components giving four protein bands.

Transhydrogenase
Activity with Purified TH, Purified THi reconstituted the light-dependent as well as the ATP-dependent transhydrogenase reactions in C&-particles (Fig.  1). Both reactions were stimulated to approximately the same percentage of their maximal rate at each titer of THr.
Unlike reconstitution with less pure factor preparations (2, 3, lo), the maximal rate obtainable with the light-dependent reaction was identical with that of the ATP-dependent reaction.
The relative abilities of purified and crude THi to restore ATP-and lightdependent transhydrogenase activity in CT-particles is shown in  The maximal rate of the nonenergy-linked reaction was 17.7 mpmoles of AcPyAD+ reduced per min, while that of the reconstituted ATP dependent transhydrogenase was 9.5 mpmoles of NADPf reduced per min.
lat,ed the ATP-driven reaction to an identical maximal rate. However, crude THI maximally stimulated the light-dependent reaction to a rate 1.6 times that obtained with purified THI. These results are in line with the observations described in the previous paper (3) indicating that crude THI contained a component, apparently not present in purified TN,, necessary for the light-but not the ATP-dependent transhydrogenase reaction.

Reconstitution of Nonenergy-linked
Transhydrogenase Activity with Purified TH, As can be seen from Fig. 2, the ATP-dependent reduction of NADP+ by NADH and the nonenergy-linked reduction of AcPyAD+ by NADPH were stimulated in C&particles to an identical percentage of their respective maximal rates at each concentration of purified TH,.
Unlike crude factor preparations (2), purified TH1 did not catalyze a nonenergy-linked transhydrogenation in the absence of G-particles.
Attempts to elicit transhydrogenase activity in purified THI by variation of pH or addition of bovine serum albumin or soy bean phospholipid micelles were unsuccessful. See the text for the method used for the washing of the chromatophores.
Since NADP+ was known to prevent the loss of THI from chromatophore membranes during washing (3)) it was of interest to compare the relative levels of the nonenergy-and energy-linked transhydrogenase reactions during washing. Therefore, chromatophores containing 7.5 mg of BChl were washed three times with 90 ml of either 0.1 M Tris-HCl buffer, pH 8, 10% sucrose or with 0.1 M Tris-HCl buffer, pH 8, 10% sucrose containing 10 PM NADPf and 2 mM MgC&. Following each washing, the chromatophore pellets were sedimented by centrifugation at 150,000 X g for 30 min and resuspended in 3 ml of Tris-HCI buffer. The rates of ATP-dependent and nonenergy-linked (TD) transhydrogenases were determined following each wash. It is apparent from Fig. 3 that the presence of NADP+ in the wash buffer resulted in the retention of both transhydrogenase activities to the same extent, while washing in the absence of NADP+ led to depletion of both activities.

Xtimulation of DD Transhydrogenation in &-particles by PuriJied THI
Enzymes which catalyze transhydrogenation between NADPH and NADf (TD transhydrogenase) have been isolated from a variety of sources, and a number have been reported to catalyze at least a trace of transhydrogenation between NADH and NAD+ analogues (DD transhydrogenase) (8,11,12). Since many dehydrogenases are known to catalyze DD transhydrogenation (13), it is difficult to conclude whether the DD activity represents an inherent property of the TD transhydrogenase enzyme or whether it represents a contaminating dehydrogenase activity.
A homogeneous flavoprotein pyridine nucleotide transhydrogenase from Pseudomonas aeruginosa has been found to catalyze both DD and TD transhydrogenase reactions (12). Purified THi was tested for DD transhydrogenase activity (reduction of AcPyAD+ by NADH) in the presence and absence of C&.-particles. Fig. 4 shows that THI did not manifest DD transhydrogenase activity in the absence of C&particles. However, the factor stimulated the endogenous DD transhydrogenase of C-r-particles maximally about 3-fold at the same titers which maximally stimulated the nonenergy-linked TD transhydrogenase more than &fold.
These observations suggest that TH1 is required for both kinds of nonenergy-linked transhydrogenase reactions of chromatophores.

StereospeciJicity of Hydrogen Transfer during Transhydrogenation
The energy-linked reduction of NADPf by NADH catalyzed by R. rubrum chromatophores possesses properties similar to the mitochondrial reaction (1). On the other hand, the mode of hydrogen transfer for the chromatophore reaction has not been investigated.
It was felt necessary to determine whether the over-all process occurs by a direct hydrogen transfer mechanism similar to the energy-linked transhydrogenases of mitochondria (7,9) and of Rhodosperiomonas spheroides (14).
The chromatophore ATP-dependent NADH reduction of NADP+ was studied in the presence of NADH tritiated in position 4 of the nicotinamide ring (see "Experimental Procedure"). In order to remove soluble proteins unrelated to the transhydrogenase reaction, but which nevertheless might catalyze an exchange of tritium to the medium, chromatophores (2 mg of BChl) were first washed with 50 ml of 0.01 M Tris-HCl buffer, pH 8, 1 y0 sucrose containing 50 PM NADP+. At prescribed times, 1.5-ml aliquots were transferred to centrifuge tubes and the transhydrogenation reaction was terminated by heating the centrifuge tube in a boiling water bath for 1 min. The tube was cooled in an ice bath and then centrifuged to remove denatured protein.
Tritium incorporation into NADPH was measured as described under "Experimental Procedure." Nevertheless, on the basis of the low specific activity of tritium in the water following the reaction (prior to oxidation of the NADPH with glutathione reductase), it is assumed that the tritium transfer between the pyridine nucleotides could not involve a prior exchange to water.

Mode sf THl Action-We
have previously demonstrated that THI has no influence on chromatophore light-induced phosphorylation (2) or the energy-linked reduction of NAD+ by succinate (15). On the basis of these experiments, it is concluded that TH1 acts neither as a coupling factor for energy conservation nor as an electron carrier in the electron transport chain. Preliminary studies with crude THI suggested that it could be a soluble transhydrogenase enzyme (2) ; however, as documented in this paper, no such activity could be detected in the most highly purified factor preparations (Fig. 2), The dependence of nonenergy-linked as well as energy-linked transhydrogenase reactions of C&-particles on TH1 indicates that the factor may act at the level of hydrogen transfer in both reactions. The view that a single factor reconstitutes both energylinked and nonenergy-linked transhydrogenase reactions is supported by the following facts: (a) energy-and nonenergy-linked transhydrogenase reactions are stimulated to an identical percentage of their maximal rates at each titer of THI and (b) both reactions required a factor which is membrane-bound in the presence of NADPf, but which is easily dissociated from the membrane in the absence of NADP+ (Fig. 3).
Assuming that THi functions solely in hydrogen transfer, a number of possibilities for its specific mode of action can be visualized.
1. THr may be a transhydrogenase enzyme which is nonfunctional until activated by being bound to the chromatophore membrane. The term "allotopy" has been used to define a change in properties, i.e. activation, inactivation, or sensitivity to inhibitors, which a number of enzymes undergo upon binding to membranes (16). This phenomenon is manifested not only by enzymes of a completely particulate nature, but also by a number of normally soluble enzymes which can be reversibly bound to membranes (17)(18)(19).
In addition, such membrane-enzyme interactions may involve specific lipid and or lipoprotein components within the membrane (18,20).
By way of example, treatment of submitochondrial particles (21) or of a detergent-solubilized mitochondrial transhydrogenase enzyme (5) with phospholipase inactivated the TD transhydrogenase activity of these preparations. Such findings indicate that for the mitochondrial system transhydrogenase activity is dependent on the integrity of a lipoprotein complex. Although it has not yet been possible to stimulate nonenergy-linked transhydrogenase activity in THI by incuba-tion with artificial phospholipid micelles, the possibility that a specific lipid-protein interaction is responsible for the activation and binding (3) of TH1 to chromatophores cannot be ruled out. Alternatively, the membrane may provide a cofactor which mediates hydrogen transfer between active sites for NAD(H) and NADP(H) on TH,. 2. An additional possibility is that THJ operates as terminal enzyme, of a multienzyme transhydrogenase system, catalyzing the dehydrogenation of either NADH or NADPH. If such is the case, a membrane component (or components) could function catalytically rather than structurally in hydrogen transfer. Griffiths and Robertson (9) have hypothesized that mitochondrial transhydrogenation may be mediated by two dehydrogenases, acting in series, one of which is A-specific for NADH and the other B-specific for NADPH.
This mode of hydrogen transfer is also consistent with the mechanism for the reaction proposed by Mitchell (22).
Path of Hydrogen Transfer-The demonstration that the energy-dependent transhydrogenase reaction of R. rubrum chromatophores occurs by a direct hydrogen transfer from the 4A position of NADH to the 4B position of NADPH supports the possibility that the mechanism of this reaction is similar to the reaction found in mitochondria.
Although it remains to be shown that chromatophore DD transhydrogenation also occurs by direct hydrogen transfer, the stimulation of both DD and TD transhydrogenase activities in C&-particles by purified TH, suggests that the former activity may be a partial reaction of the latter activity.
DD transhydrogenation would be expected to occur if the mechanism of TD transhydrogenation involves the transfer of hydrogen from NADH to an intermediate reactions has not been ruled out. The fact that the process in mitochondria and chromatophores takes place without exchange of hydrogen to medium water indicates that an easily accessible hydrogen acceptor is an unlikely participant in the reaction.
On the other hand, a direct transhydrogenation between NADH and NADP+ catalyzed by an enzyme from Pseudomonas aeruginosa has been reported to incorporate a reduced flavin intermediate (12). As of yet no such cofactor requirement has been found for THI.
Orlando (23) has recently shown that chromatophores of Rhodopseudomonas spheroides subjected to a discontinuous sucrose gradient lose the ability to catalyze the light-and ATPdependent transhydrogenation.
The light-dependent, but not ATP-dependent, activity can be reactivated by the addition of a factor isolated from lyophilized powders of the same cells. Both the light-and ATP-dependent reactions can as well be reactivated by the dithiols, dithiothreitol, dithioerythritol, or reduced thioctic acid, in the absence of this factor.
The factor from R. .apheroides does not influence the nonenergy-linked transhydrogenation which is still active in particles inactivated with respect to the energy-linked reaction.
The above characteristics represent major differences in comparison with the characteristics of the transhydrogenase factor isolated from R. rubrum and described in this paper.
(a) The TH, factor reconstitutes the light-driven, the ATP-driven, and the nonenergy-linked reaction in resolved chromatophores of R. rubrum, whereas the factor from R. spheroides is only active in reconstituting the light-driven reaction.
(b) In contrast to the R. spheroides factor, dithiols do not substitute for the TH1 factor in R. rubrum.