Purification and characterization of a 43-kDa rotenone-insensitive NADH dehydrogenase from plant mitochondria.

A 43-kDa NAD(P)H dehydrogenase was purified from red beetroot mitochondria. An antibody against this dehydrogenase was used in conjunction with the membrane-impermeable protein cross-linker 3,3'-dithiobis(sulfosuccinimidylpropionate) to localize the dehydrogenase on the matrix side of the inner membrane. Immunoblotting showed that the dehydrogenase was found in mitochondria isolated from several plant species but not from rat livers. Antibodies against the purified dehydrogenase partially inhibited rotenoneinsensitive internal NADH oxidation by inside-out submitochondrial particles. The level of rotenone-insensitive respiration with NAD-linked substrates correlated with the amount of 43-kDa NAD(P)H dehydrogenase present in mitochondria isolated from different soybean tissues. Based on these results, we conclude that the 43-kDa NAD(P)H dehydrogenase is responsible for rotenone-insensitive internal NADH oxidation in plant mitochondria.

A 43-kDa NAD(P)H dehydrogenase was purified from red beetroot mitochondria. An antibody against this dehydrogenase was used in conjunction with the membrane-impermeable protein cross-linker 3,3-dithiobis(sulfosuccinimidylpropionate) to localize the dehydrogenase on the matrix side of the inner membrane. Immunoblotting showed that the dehydrogenase was found in mitochondria isolated from several plant species but not from rat livers. Antibodies against the purified dehydrogenase partially inhibited rotenoneinsensitive internal NADH oxidation by inside-out submitochondrial particles. The level of rotenone-insensitive respiration with NAD-linked substrates correlated with the amount of 43-kDa NAD(P)H dehydrogenase present in mitochondria isolated from different soybean tissues. Based on these results, we conclude that the 43-kDa NAD(P)H dehydrogenase is responsible for rotenone-insensitive internal NADH oxidation in plant mitochondria.
It has been known for some time that plant mitochondria, unlike their mammalian counterparts, can oxidize NAD-linked substrates in the presence of the complex I inhibitor rotenone (1). However, rotenone lowered the ADP/O ratio for NADlinked substrates by one third (1), suggesting the presence of an internal NADH dehydrogenase that did not translocate protons and which was insensitive to rotenone. This was supported by the finding (2) that inside-out submitochondrial particles (SMP) 1 from Jerusalem artichoke exhibited two different K m values for NADH, depending on whether rotenone was present or not. These results suggested that the rotenoneinsensitive bypass was mediated by a distinct NADH dehydrogenase that had a high K m for NADH relative to complex I. This notion was strengthened further by the finding of Rasmusson and Møller (3), that respiration by potato submitochondrial particles oxidizing the complex I-specific, NADH analogue, deamino-NADH (4), was almost completely inhibited by rotenone, but was restored by adding NADH. This result demonstrated that the rotenone-insensitive bypass was mediated by an NADH dehydrogenase distinct from complex I. This enzyme has a low affinity for NADH, does not translocate protons, and probably underlies the rapid rates of resting state respiration seen with plant mitochondria oxidizing NAD-linked substrates (5).
The mitochondria of Saccharomyces cerevisiae also oxidize internal NADH via a rotenone-insensitive dehydrogenase (6). The yeast enzyme is nuclear-encoded (6), is composed of a single subunit with a molecular mass of 53 kDa and contains FAD as the sole prosthetic group (7).
There are several reports detailing the isolation of non-complex I NADH dehydrogenases from plant mitochondria (8 -16). However, only a few of these resolve single proteins and provide evidence for their possible function within the mitochondrion. Luethy et al. (14) isolated NAD(P)H dehydrogenases with molecular masses of 32, 58, and 42 kDa from red beetroot mitochondria. Then it was shown that the 32-and 58-kDa proteins are also present in maize mitochondria and are located either on the outside of the inner membrane or in the intermembrane space (13,15). Therefore, these proteins, are likely to be involved in oxidation of cytosolic NAD(P)H. In contrast, Rasmusson et al. (16) concluded by a process of elimination, rather than by direct evidence, that a 26-kDa NAD(P)H dehydrogenase from red beetroot mitochondria was responsible for rotenone-insensitive internal NADH oxidation.
In this paper, we report on the purification of a 43-kDa NAD(P)H dehydrogenase from red beetroot mitochondria and provide strong evidence that this enzyme is responsible for rotenone-insensitive internal NADH oxidation.

EXPERIMENTAL PROCEDURES
Materials-Red beetroots (Beta vulgaris L.) and Potatoes (Solanum tuberosum L.) were purchased from local markets. Soybean (Glycine max L. cv. Stevens) plants were glass house-grown and cotyledons and roots harvested at 7 days; nitrogen-fixing nodules were harvested at 8 weeks.
Preparation and Fractionation of Mitochondria-Beetroot mitochondria were prepared by the method of Menz et al. (17), rat liver mitochondria by the method of James et al. (18), potato tuber mitochondria by the method of Liden and Akerland (19), and soybean mitochondria by the method of Day et al. (20). The mitochondrial "soluble fraction" was prepared by resuspending the mitochondrial pellet in 20 mM Tris/HCl (pH 8.0) containing 1 mM phenylmethylsulfonyl fluoride and 5 M E64 protease inhibitor and sonicating for 5 ϫ 5-s bursts using an MSE (Sussex, United Kingdom) ultrasonic disintegrator at maximum power. The sonicate was centrifuged at 300,000 ϫ g for 45 min, and the resulting supernatant represented the soluble fraction. Submitochondrial particles were prepared as detailed previously (17).
Chromatographic Techniques-All chromatographic procedures were carried out using a Pharmacia (Sydney, Australia) fast protein liquid chromatography system. Anion exchange chromatography was performed on a Pharmacia resource Q (1 ml) column, while blue affinity chromatography was performed on a Pharmacia Hi-Trap blue column. * This work was supported by an Australian Research Council grant (to D. A. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Recipient of a graduate student assistantship from the Co-operative Research Center for Plant Science.
Electrophoretic Techniques-SDS-PAGE was carried out using the method of Laemmli (21) with 12% (w/v) polyacrylamide resolving gels. Polyclonal antibodies against the purified 43-kDa NAD(P)H dehydrogenase were raised in rats according to the method described by Harlow and Lane (22). Western blotting was performed according Towbin et al. (23) using the alkaline phosphatase/nitro blue tetrazolium detection system.
Other Assays-Noncovalently bound flavin was removed from the protein by boiling for 3 min and the flavin then determined fluorometrically by the method of Siegel (24). NAD(P)H:Q 0 reductase activity was measured in 10 mM BTP (bis-Tris-propane) buffer (pH 7.2) containing 200 M NAD(P)H and 200 M Q 0 . The reduction of NAD(P)H was monitored at 340 nm (extinction coefficient at 340 nm ϭ 6.22 mM Ϫ1 cm Ϫ1 ). NADH:FeCN reductase activity was measured in a similar fashion except 0.5 mM FeCN replaced Q 0 , and the reduction of FeCN was monitored at 420 nm (extinction coefficient at 420 nm ϭ 1.05 mM Ϫ1 cm Ϫ1 ).
Protein was estimated by the method of Bradford (25) using the Bio-Rad reagent and bovine serum albumin (fraction V) as a standard. (26) and others (14,16) have shown that rotenone-insensitive NADH dehydrogenase activity is released from mitochondrial membranes by sonication and can be recovered in the soluble fraction (see "Experimental Procedures"). Use of this fraction has the advantage of avoiding complications due to complex I, all of which remain in the membrane fraction. The soluble fraction generated from sonication of red beetroot mitochondria was subjected to anion exchange chromatography, and the resulting fractions assayed for NADH:Q 0 reductase activity (Fig. 1). Three activity peaks were resolved, a result consistent with previous work using mitochondria from red beetroot (14,16). Luethy et al. (15) and Knudten et al. (13) had shown previously that the proteins that give rise to activity peaks 2 and 3 of the anion exchange profile are located external to the mitochondrial matrix. Consequently we focused on the first activity peak, which eluted at an NaCl concentration of approximately 85 mM; fractions corresponding to this peak were collected, diluted 2-fold to reduce the concentration of NaCl, and subjected to further chromatography.

Purification of a 43-kDa NAD(P)H Dehydrogenase-We
The pooled fractions were applied to a blue affinity column. The column was washed with buffer containing 150 mM NaCl prior to elution with a gradient of 0 -10 M NADPH in elution buffer. The resulting fractions were assayed for NADH:Q 0 reductase activity. The majority of the NADH:Q 0 reductase activity eluted from the blue affinity column at an NADPH concentration of approximately 3 M (Fig. 2). When fractions corresponding to this peak were pooled and subjected to SDS-PAGE, a single major band with an apparent molecular mass of 43 kDa was detected (Fig. 3). When the purified protein sample was subjected to gel filtration chromatography, the NADH:Q 0 reductase activity was found to be associated with a protein species with a native molecular mass of 86 kDa (not shown). This result suggests that the active form of the enzyme was a dimer of two 43-kDa subunits.
Submitochondrial Location of the 43-kDa NAD(P)H Dehydrogenase-A polyclonal antibody against the 43-kDa protein was prepared and used in conjunction with the membraneimpermeable protein cross-linker DTSSP to determine the submitochondrial location of the enzyme (13). Intact mitochondria, which exclude DTSSP from proteins located on the inside of the inner membrane, and sonicated mitochondria, with a disrupted and exposed inner membrane that could not exclude DTSSP, were incubated with the cross-linker. Following cross-linking, the mitochondria were subjected to SDS-PAGE and Western blot analysis. In the absence of ␤-mercaptoethanol, the 43-kDa protein was only detected in the intact mitochondria and not in the sonicated mitochondria (Fig. 4). However, when SDS-PAGE was performed in the presence of ␤-mercaptoethanol, which cleaves the cross-linker, Western analysis detected the 43-kDa protein in both the intact and sonicated mitochondria (Fig. 4). These results demonstrate that the 43-kDa protein was protected from the cross-linker in intact mitochondria and therefore must be located within the mitochondrial inner membrane (DTSSP can pass through the outer membrane; Ref. 13). These results also suggest that DTSSP cross-links the 43-kDa protein to other proteins forming a supermolecular weight complex which, because of its large size, is not transferred to the blotting membrane.
Correlation between 43-kDa Protein and Matrix Rotenoneinsensitive NADH Oxidation-Mitochondria from several different plant species as well as from rat livers were subjected to Western analysis with the antibodies raised against the 43-kDa red beet protein. The antibodies were found to cross-react with a protein of approximately 43 kDa in mitochondria from potato tuber and soybean cotyledons but not in rat liver mitochondria (Fig. 3). The presence of the 43-kDa protein in a variety of plant species, but not in completely rotenone-sensitive mammalian mitochondria, is consistent with the 43-kDa protein being one of the unique NADH dehydrogenases found in plant mitochondria.  It has been shown that mitochondria isolated from different tissues of soybean plants exhibit different levels of the internal rotenone-insensitive bypass activity (17). Therefore, mitochondria were isolated from soybean cotyledons, roots, and nodules, and the level of rotenone-insensitive internal NADH oxidation measured. The level of this activity was found to vary, with that in cotyledons being greater than in roots which in turn were greater than in nodules (Fig. 5). These mitochondria were also subjected to Western analysis with the antibody against the 43-kDa NAD(P)H dehydrogenase. A good correlation between the level of this protein and the amount rotenone-insensitive internal NADH oxidation was found (Fig. 5). This correlation provides further evidence that the 43-kDa protein mediates rotenone-insensitive internal NADH oxidation.
Antibodies against the 43-kDa protein were tested for their ability to inhibit rotenone-insensitive internal NADH and NADPH oxidation, which are likely to be mediated by different proteins (27). Rotenone-insensitive internal NADH oxidation was measured in inside-out beetroot SMP that were incubated with either the 43-kDa NAD(P)H dehydrogenase antiserum or preimmune serum, for 2 h on ice. The 43-kDa protein antiserum inhibited rotenone-insensitive internal NADH oxidation by approximately 20% (not shown). In contrast, there was no significant inhibition of NADPH oxidation by SMP with the 43-kDa NAD(P)H dehydrogenase antiserum. This result supports the hypothesis that there are separate enzymes for NADH and NADPH oxidation (27) and that NADH oxidation is mediated by the 43-kDa NAD(P)H dehydrogenase.
Characterization of the 43-kDa NAD(P)H Dehydrogenase-The relative effects of different substrates and acceptors on the dehydrogenase activity are shown in Table I. The enzyme was capable of oxidizing both NADH and NADPH at an equal rate with Q 0 as acceptor. In this regard the purified enzyme appears similar to both the 42-kDa dehydrogenase purified by Luethy et al. (14) and the 26-kDa protein purified by Rasmusson et al. (16). The rate of oxidation by the enzyme with FeCN as accep-  tor was only 60% of that observed with Q 0 as acceptor and the purified dehydrogenase was only capable of oxidizing deamino-NADH at 45% of the rate observed with NADH (Table I). This preference for NADH over deamino-NADH is indicative of a type II NADH dehydrogenase (28). The purified dehydrogenase was also found to contain predominantly FAD (not shown), another feature of type II dehydrogenases (28). The kinetics of NADH oxidation were found to best fit a Michaelis-Menten equation for a bi-substrate reaction with a ping-pong mechanism (Fig. 6). The estimated apparent K m for NADH was 340 Ϯ 195 M and the apparent K m for Q 0 was 65 Ϯ 39 M (Fig. 6). The enzyme had an estimated V max of 53 Ϯ 22 mol min Ϫ1 mg Ϫ1 (Fig. 6). The pH optimum was found to be 6.5, but the enzyme maintained greater than 90% of its maximal activity between pH 5.5 and 7.5 (not shown).
The activity of the purified enzyme showed little sensitivity to calcium. In the presence of EGTA only a 5% reduction in activity was observed (Table I), while added calcium had no effect (Table I). The nucleotides NAD ϩ , NADP ϩ , ADP, and ATP had only small effects on activity (Table I). The inhibitors mersalyl and p-chloromercuribenzoic acid, on the other hand, inhibited NADH:Q 0 reductase activity by 32 and 26%, respectively (Table I). Dicumarol was found to be the most potent inhibitor, inhibiting by 88% (Table I). The other NADH dehydrogenase inhibitors, platanetin and flavone, had no effect on the NADH:Q 0 reductase activity of the purified enzyme (Table I). DISCUSSION We have purified a 43-kDa NAD(P)H dehydrogenase from red beet mitochondria. The enzyme was located on the matrix side of the inner membrane. It was present in mitochondria from several plant species, but not rat liver, and there was good correlation between the abundance of the 43-kDa NAD(P)H dehydrogenase and rotenone-insensitive oxidation by soybean mitochondria. Taken together, these results provide the strong-est evidence to date that the 43-kDa protein is the internal rotenone-insensitive bypass of complex I found in most plant mitochondria. The antibody inhibition of rotenone-insensitive internal NADH oxidation, but not NADPH oxidation, further supports this idea and suggests that a separate enzyme is responsible for rotenone-insensitive internal NADPH oxidation.
Comparison of the kinetic and inhibitor data of the purified dehydrogenase with those of rotenone-insensitive NADH oxidation in inside-out SMP from various plant tissues reveals several similarities. These include: a high K m for NADH (of the order of 100 M; Refs. 2 and 3); a pH optimum of approximately 6.5 (3); inhibition by dicumarol and mersalyl (3,26). A notable difference between the purified enzyme and that in SMP is the ability of the former to facilitate rapid rates of NADPH oxidation; this is most likely due to the removal of the enzyme from its hydrophobic environment (26).
The 43-kDa NAD(P)H dehydrogenase isolated here is probably the same as the 42-kDa enzyme previously isolated by Luethy et al. (14). Both activities have similar molecular masses and activities with different substrates, activators, and inhibitors. The most obvious difference between the two preparations is that the protein of Luethy et al. (14) appeared more sensitive to calcium and ADP. These differences could easily be explained by different ionic conditions in the assay. The identity of the 26-kDa protein isolated by Rasmusson et al. (16) remains unclear. However, it is interesting to note that the latter authors found a native molecular mass for their enzyme similar to that reported here, and their enzyme also had the ability to oxidize NADPH and reduce FeCN. It is possible, therefore, that the enzyme isolated by Rasmusson et al. (16) was identical to that reported here, but the protein was degraded to a lower apparent molecular mass during isolation or storage.
The enzyme purified here differs in molecular mass from the rotenone-insensitive internal NADH dehydrogenase of S. cerevisiae (7). However, similarities may become apparent when sequence data are available for the 43-kDa enzyme of plants.