NADH/NAD⁺ interaction with NADH:ubiquinone oxidoreductase (complex I)

Summary

The quantitative data on the binding affinity of NADH, NAD⁺, and their analogues for complex I as emerged from the steady-state kinetics data and from more direct studies under equilibrium conditions are summarized and discussed. The redox-dependency of the nucleotide binding and the reductant-induced change of FMN affinity to its tight non-covalent binding site indicate that binding (dissociation) of the substrate (product) may energetically contribute to the proton-translocating activity of complex I.

1. Introduction

Mitochondrial NADH:ubiquinone oxidoreductase (complex I) and its prokaryotic homologues (NDH-1) are intrinsic components of the respiratory chains that catalyze the first step of intramitochondrial or intracellular NADH oxidation (or NAD⁺ reduction) and vectorial translocation of protons across the coupling membrane. Under “optimal” conditions (uncoupled NADH oxidase activity of inside-out bovine heart submitochondrial particles (SMP) at saturating NADH concentration, pH 8.0, 25^°C) the enzyme turnover is about 150 s^-1 [1]. Approximately the same rotenone-sensitive steady-state turnover is seen with exogenous ubiquinone homologue, Q₁, as electron acceptor. The multiple iron-sulfur clusters [2] participating in intramolecular electron transfer from FMN, presumably the primary oxidant (reductant) of NADH (NAD⁺), are almost completely reduced during the steady-state NADH-supported coupled or uncoupled respiration [3, 4]. Reoxidation of the terminal iron-sulfur cluster N-2 thus seems to be the rate limiting step of the overall NADH oxidase or NADH:ubiquinone reductase reactions. The enzyme turnover in the rotenone-insensitive “diaphorase” activities of complex I as revealed with a number of artificial electron acceptors such as ferricyanide, hexaammineruthenium (III) (HAR), DCIP, menadione, cytochrome c, acetyl-NAD⁺ (transhydrogenase reaction), oxygen (generation of superoxide radical) varies from less than 0.1 (superoxide generation) to 1000 per cent of that measured in the natural NADH oxidase reaction [1].

NADH oxidation by ubiquinone is energetically coupled with translocation of 4 protons per pair of electrons transferred [5-7]. This stoichiometry eliminates the single Mitchellian redox loop [8] as the only mechanism of pmf generation in the NADH:ubiquinone region of the respiratory chain. Other possibilities such as energy-coupled redox-dependent substrate/product binding/release conformational change mechanism are worthy of consideration besides a number of hypothetical models that have been proposed for a direct redox-coupled proton-translocating loop mechanism at the N-2–ubiquinone junction site [9-16]. In fact, the NADH-induced conformational change of complex I has been qualitatively demonstrated [17-19]. To create any testable model for the conformationally-driven energy transduction model quantitative information on the parameters of the substrate (product) binding (release) is needed. Here the current state of knowledge of NADH, NAD⁺, and their analogues interaction with complex I as related to the flavin redox redox state is briefly summarized and discussed.

2. NADH/NAD⁺ binding affinities as emerged from the steady-state kinetics

Numerous reports have been published on NADH/NAD⁺ affinities to complex I or to other enzymatically active preparations derived from the intact complex. These reports have used different approaches including the steady-state kinetics. Here only the data on the membrane bound and purified dispersed complex I or so-called “high molecular weight” NADH dehydrogenase are discussed. The redox potentials of the iron-sulfur centers [20] and likely FMN are significantly modified in simpler “low molecular weight” fragments of complex I (such as FP, flavoprotein fragment of Complex I. Although FP certainly bears the NADH binding site, these redox modifications are expected to modify also the apparent affinities to NADH/NAD⁺ (see below). Before discussing quantitative data summarized in Table 1, some comments on the steady-state kinetics of NADH oxidation seem to be helpful.

Table 1.

Apparent affinities of the substrate/product nucleotides to the active site(s) of the NADH-dehydrogenase (ubiquinone reductase) as derived from enzyme steady-state kinetics

Enzymea and assay conditions		NADH oxidation				NAD+-reduction (succinate-supported reverse electron transfer		Ref
	Activity	Kinetic mechanism Suggested	KiNADH	$K_i^{{\mathit{NAD}}^{+}}$	KiNADH (μM)	$K_m^{{\mathit{NAD}}^{+}}$	KmNADH
SNDH, 30°C, pH 7.8	Ferricyanide reductase		100					21
BHC-I, 25°C, pH 7.5	Ferricyanide reductase	ping-pong	100		50			22
B-SMP, 27°C, pH 8,0	HAR reductase	ordered	>40b (100)		no inhibition			23
B-SMP, 30°C, pH 8.0	oxidase		2	1 000; 1 600	no inhibition	7; 25	40; 80	24, 25
P-SBP, 25°C, pH 7.0	oxidase		5		no inhibition	20		26
P-SBP, 25°C, pH 8.0	oxidase		7			270		27
BHC-I, 20°C, pH 8.0	Q1 reductase	ordered, ternary complex	2		no inhibition			28
BHC-I, 32°C, pH 7.5	Acetyl-NAD+ transhydrogenase	ping-pong	100		160-260			29
	HAR reductase		90					29
	Superoxide generation		0.05c	Strong inhibition				30

^aAbbreviations: SNDH, soluble high molecular weight NADH dehydrogenase derived from bovine heart submitochondrial particles; BHC-I, purified bovine heart complex I; B-SMP, bovine heart inside-out submitochondrial particles; P-SBP, inside-out plasma membrane vesicles from anaerobically grown P. denitrificans.

^bAt 0,5 mM HAR. Significantly higher K_m^NADH (100 μM) is seen at high (2 mM) concentration of HAR.

^cOnly 5-times molar excess of NADH over the enzyme concentration in the presence of NADH-regenerating system.

The simplest (single NADH binding site, i.e., single site for the enzyme reoxidation) scheme describing NADH oxidation catalyzed by complex I is depicted in Fig. 1. Because two substrates (NADH and electron acceptor) are involved at least two alternative reaction pathways are conceivable: after NADH binding and intramolecular oxidoreduction (intermediates 1, 2, and 3) an electron acceptor may oxidize the enzyme before dissociation of NAD⁺ (the overall reaction proceeds via intermediates 1, 2, 3, and 4); alternatively, an electron acceptor may reoxidize product-free reduced enzyme only (intermediates 1, 2, 3, and 5). According to textbook steady-state kinetics the first alternative is characterized by a series of simple hyperbolic initial rate versus NADH concentration dependencies with different apparent K_m^NADH values, which are extrapolated to K_S^NADH when electron acceptor concentration is extrapolated to zero. For redox enzymes the K_S^NADH value thus obtained is not a measure of primary NADH binding to the active site (K₁, Fig. 1) but it is always a product, K₁·K₂, where K₂ is the redox equilibrium constant for the intramolecular electron transfer between bound NADH and primary enzyme-bound electron acceptor, presumably FMN. The alternative mechanism (reoxidation of product-free reduced enzyme, intermediates 1, 2, 3, and 5) gives a series of parallel lines in 1/v versus 1/[NADH] graphs at different acceptor concentrations with apparent K_m^NADH extrapolating to zero when an electron acceptor concentration extrapolates to zero (ping-pong mechanism). It should be noted that if K₂≪1, i.e. the intramolecular oxidoreduction is greatly shifted to the intermediate 3 (e.g. midpoint redox potential of FMN/FMNH₂ is significantly more positive than that of bound NADH/NAD⁺), two alternative mechanisms become kinetically undistinguishable. Indeed, if so, the interconversion 2→3 becomes “irreversible” just like interconversion 3→5 is in the ping-pong mechanism where “irreversibility” is due to the initial rate limitation ([NAD⁺]=0). It is also possible that the reaction proceeds both via 1, 2, 3, 4, and 1, 2, 3, 5 intermediates with relative contributions of two pathways that depend on particular electron acceptor concentration. In this case deviations from simple hyperbolic rate-reactant concentrations dependencies are expected.

Fig. 1.

Simplified scheme describing the steady-state NADH oxidation and NAD⁺ reduction by complex I. Shadowed numbered circles are designation for distinct intermediates (see text for further explanation).

Binding of NADH

The apparent K_m^NADH values published in the literature for the reactions with different electron acceptors (Table 1) can, at first approximation, be divided into “high K_m” (rotenone-insensitive artificial acceptor reductases) and “low K_m” (rotenone-sensitive quinone reductases) activities. The difference between apparent K_m^NADH in these groups are evidently too large (about 50 times) to be accounted for slight variations in pH, temperature and assay buffer composition. Rather, it is related to different species designated as E_red in the reaction pathway in Fig. 1. Complex I is a multiple redox component enzyme and since ferricyanide [31] and HAR [32] are efficient electron acceptors for FP it is safe to assume that they reoxidize flavin, the primary electron acceptor for bound NADH. The E_red species potentially reactive with these acceptors in 1, 2, 3, 4 pathway is NAD⁺·E·Fl_red, and apparent K_m^NADH is, as noted above, equal to K₁·K₂, where K₂ is the equilibrium constant for bound NADH-FMN oxidoreduction. When the enzyme is reoxidized by quinone (rotenone-sensitive reaction) the corresponding E_red specie is NAD⁺·E·(N-2)_red and a series of electron transfer steps from FMNH₂ to iron-sulfur cluster N-2 contribute to K₂. Because the midpoint potential of N-2 is much more positive than that of flavin (at least 200 mV difference [33, 34]) K₂ is ≪1 and low K_m^NADH values are related to thermodynamics of intramolecular electron transfer, not to the true affinity of NADH to its binding site. Therefore, true affinity of NADH to the oxidized enzyme (K₁, Fig. 1) is less or equal to ~10^-4 M.

It was noted many years ago that the reaction with ferricyanide is strongly inhibited by NADH [21, 22]. NADH-supported and succinate-supported superoxide generation [35], as well as, NADH:AcetylNAD⁺ transhydrogenase activity [29] are also strongly inhibited by NADH. This inhibition was interpreted as evidence for the ping-pong mechanism with double substrate inhibition [22]. The recently solved atomic structure of the hydrophilic fragment of NDH-1 from Thermus thermophilus [36] agrees with the earlier proposal that ferricyanide reoxidizes NAD⁺-free reduced enzyme only: the reactive component (FMN) in the enzyme from this species is located in a deep cleft being accessible for either negatively charged hydrophilic ferricyanide (-3) or negatively charged hydrophilic NADH (-2), not for both. However, no inhibition by NADH is seen in the NADH:HAR reductase activities catalyzed by the mammalian membrane-bound or solubilized purified complex I [23]. Also the kinetic mechanism of the NADH:HAR reductase reaction is not a ping-pong type [23]. These facts strongly suggest that some enzyme-bound redox component of complex I is easily accessible for hydrophilic positively charged HAR (+3) when the substrate binding “cleft” is occupied by the substrate (product) nucleotide. This component is most likely flavin because HAR is a quite efficient electron acceptor for FP which harbors only modified low midpoint potential N-3 and N-1 iron-sulfur centers [20]. It would be of great interest to see whether the T. thermophilus NDH-1 fragment is active in NADH:HAR reductase and, if so, what is the kinetic mechanism of this reaction.

Interaction of NADH with the intermediate 5 (equilibrium between intermediates 5 and 6, Fig. 1) can be considered as true measure of NADH affinity to the reduced enzyme. The estimated value of K₅ is about 5·10^-5 M which is close to the K_i^NADH in the energy-linked reverse electron transfer reaction ((4-8)·10^-5 M, Table 1) and to true K_s for NADH-reduced enzyme interaction determined by the more direct approach (2·10^-5 M, Table 2). These values when compared with the upper limit for K₁ (~1·10^-4M) as discussed above suggest that the reduction of the enzyme (flavin) increases its binding site affinity to NADH.

Table 2.

Relative affinities of the nucleotides to the oxidized and reduced Complex I [39].

(pH 8.0, 25°C)
Nucleotide	KS, μM
	Eoxa	Eredb
NADH-OH	3·10-4	7·10-3
NADH	(100)c	20b (50)c
NAD+	800a (1000)c	(7; 25)c
ADP-ribose	25 (30)c	400b

^aAerobically incubated coupled SMP in the presence of uncoupler.

^bComplex I in coupled SMP reduced by the reverse electron transfer (coupled succinate oxidation).

^cFigures in brackets are representative values taken from Table 1.

The lowest K_m^NADH (5·10^-8 M half-saturating concentration) have been reported for the NADH-supported superoxide generation catalyzed by purified complex I in the presence of an NADH regenerating system [30]. This value can only be explained if superoxide production proceeds by a ping-pong mechanism exclusively via intermediates 1, 2, 3, 5 with extremely slow interconversion at step 5→1 when oxygen serves as the electron acceptor.

Binding of NAD⁺

NAD⁺ is a weak competitive inhibitor of NADH oxidation by either natural (NADH oxidase) or artificial electron acceptors. $K_i^{{\mathit{NAD}}^{+}}$ corresponding to K₄ in Fig. 1 is (1-2)·10^-3 M, i.e. at least an order of magnitude higher than K₁. It is also almost two orders of magnitude higher than $K_m^{{\mathit{NAD}}^{+}}$ in the energy-dependent reverse reaction which under “optimal” conditions proceeds at the rate of approximately 1/5 of that of uncoupled NADH oxidase. A conceivable possibility to explain this difference between $K_m^{{\mathit{NAD}}^{+}}$ and $K_i^{{\mathit{NAD}}^{+}}$ in the forward and reverse reactions in a framework of the scheme in Fig. 1 (with an assumption that NAD⁺ reduction is a simple reversal of NADH oxidation) is to propose that the energy-linked NAD⁺ reduction proceeds exclusively by ping-pong mechanism, i.e. via intermediates 1, 5, 3, 2. It is worth emphasizing that any apparent K_m can be determined (from zero to infinity) for a ping-pong mechanism when maximal activity is varied by variation of the second substrate concentration. The second substrate for the reverse electron transfer is reduced ubiquinone and it is difficult, if not impossible, to design experimental conditions where the reductant, QH₂ could be varied at a constant level of pmf in order to verify the kinetic mechanism of this reaction.

Binding of ADP-ribose

The only redox inactive competitive inhibitor of NADH oxidation by complex I is ADP-ribose originally discovered as a contaminant in NAD⁺ samples [37]. The inhibitor shows a K_i of 3·10^-5 M (mammalian membrane-bound complex I) or 8·10^-5 M (membrane-bound NDH-1 of Paracoccus denitrificans) [27], the affinity to presumably oxidized enzyme which is significantly higher than that of NAD⁺. From the structural point of view this suggests that the nicotinamide moiety of NAD⁺ does not contribute to its binding affinity (in fact weakens its primary binding). Interestingly, ADP-ribose does not inhibit the energy-dependent reverse electron transfer in coupled SMP or P. denitrificans vesicles [26, 37]. Moreover, the “inhibitor” increases the steady-state NADH/NAD⁺ ratio reached when particles respire with succinate in the presence of low concentration of NAD⁺ (aerobic reverse electron transfer) [37].

3. Nucleotide binding without enzyme turnover

As it is evident from the “minimal” kinetic scheme (Fig. 1) the interactions of nucleotides with complex I during the steady-state turnover give rise to at least six intermediates and depending on the particular pathway and specific activity the equilibrium constants K₁–K₅ may or may not be significantly influenced by the kinetic terms, such as the rate constants for irreversible steps (k_cat’s). It seemed desirable to use other approaches which would allow determination of true binding constants for nucleotides under equilibrium conditions. The protective effect of the substrate (product) against the binding site-directed irreversible inhibitor is the method widely used for this purpose. Until recently no such inhibitor for complex I was available. In 2005 Kotlyar et al. reported that prolonged incubation of NADH under alkaline aerobic conditions results in a formation of an NADH derivative (operationally named NADH-OH) which specifically interacts with complex I or FP in close to 1 stoichiometric ratio [38]. Although the exact chemical structure of NADH-OH remains to be established the inhibitor seems to be a very useful tool for several aspects of complex I studies particularly for evaluation of NADH/NAD⁺ binding parameters. A short summary of these studies [39] are shown in Table 2 and briefly discussed below. Reduction of the enzyme (by succinate and pmf) results in significant (~10 times) decrease of NADH-OH reactivity towards its binding site. Under the conditions where NADH-OH interacts with the enzyme in the first-order fashion NAD⁺, NADH and ADP-ribose fully protect complex I against irreversible inhibition and their binding constants thus determined are redox-dependent (Table 2). The representative figures for corresponding K’s derived from the steady-state kinetics (Table 1) are given in brackets for comparison. The data shown in Table 2 suggest that binding of nucleotides per se alters the redox potential of the enzyme (flavin) shifting it to lower E_m value (note the more than 10 times difference in the binding constants of redox inactive ADP-ribose to the reduced and oxidized enzyme).

Another interesting fact of NADH-OH interaction is the non-trivial inhibition pattern observed when forward (NADH oxidase) and reverse (succinate-supported NAD⁺ reduction) are assayed after preincubation of SMP with gradually increasing concentrations of the inhibitor. NADH oxidase titration appears as a straight line with the final titration point similar to that for piericidin (~0.1 nmol/mg of protein [27]). In contrast, titration of the energy-linked NAD⁺ reduction appears as a curve with substantial lag in inhibition and the same final titration point. In other words, a region exists where NAD⁺ reduction is not inhibited whereas NADH oxidation is substantially decreased.

Several reasons for this type of titration merit consideration. It is conceivable that the potential activity of Complex I in the reverse reaction (NAD⁺ reduction) greatly exceeds the pmf-generating capacity of succinate oxidase. This model if considered as two homogeneous enzymes: QH₂ (and pmf)-producing and QH₂ (and pmf)-utilizing system, would result in the convex titration curve as observed. However, since pmf is involved as an intermediate, operation of succinate oxidase and coupled Complex I-mediated NAD⁺ reductase should be regarded as a gear wheel transmission and the rate-limiting enzyme model seems hard to apply. Another possible explanation is that the catalytic activity of Complex I in the reverse direction increases as the driving force (pmf) increases. In other words the turnover number of the residual, not inhibited Complex I increases upon decrease of the load in the electrochemically coupled unit thus resulting in pmf increase. Such a model has been suggested for different “modality” of F_o·F₁ operation in ATP synthesis [40, 41]. Both, rate-limit and different “modality” models predict that any “irreversible” inhibitor of Complex I would give convex titration curves for reverse electron transfer which is, however, not the case for example, for piericidin [27]. A modified model with two cooperative nucleotide binding sites [35] could explain the non-trivial titration pattern. A two nucleotide binding site model does not necessarily contradict the atomic structure of hydrophilic fragment of T. thermophilus NDH-1 [36] where a single tunnel to deeply buried FMN is seen.

4. Redox-dependent flavin-protein interaction

Redox-dependency of nucleotide affinity to the enzyme active sites suggests that conformations of the reduced and oxidized protein are different at least in the area where primary binding of the nucleotides takes place. Several observations point to FMN, the primary oxidant as the component responsible (involved) for structural rearrangements that occur upon its oxidoreduction. Tightly bound FMN located in 51 kDa (Nqo-1) subunit is not covalently bound [44] and oxidoreduction of the isoalloxazine ring is expected to change its binding affinity to the protein. This phenomenon is well documented for a number of other flavoproteins (see, for example Refs. [45, 46]). The first direct evidence for the redox-dependent change of FMN binding affinity operationally called “reductive inactivation” has been reported for the hydrophilic three-subunits FP [47]. When diluted samples of FP were incubated with NADH or dithionite the catalytic activity (menadione reductase) gradually decreased due to FMN dissociation. The enzyme was protected against NADH-induced inactivation by FMN and (or) NAD⁺. The midpoint FMN potential of − 325 mV (pH 8.0) was estimated from the dependency of inactivation rate on NADH/NAD⁺ ratio. The reductant-induced inactivation was irreversible thus suggesting that in addition to its redox cofactor function FMN plays an important role in maintaining a stable conformation of FP. These findings together with data on the redox-dependent nucleotide affinity change [39] prompted us to search for conditions where reversible dissociation of FMN could be demonstrated in a system where the flavin binding Nqo-1 subunit is in close contact with other subunits within native membrane-bound complex I. Recently we were able to demonstrate the reversible dissociation of FMN from complex I as illustrated in Fig. 2. Three factors happen to be of crucial importance for reversible dissociation of FMN: strong dilution of the enzyme samples (expected from Ostwald’s dilution law), alkaline pH, and the presence of NADH. The NADH-induced inactivation is seen for both rotenone-insensitive HAR-reductase and rotenone-sensitive Q₁ reductase activities, and either activity is protected and specifically restored by FMN. Remarkably, the rate of NADH-induced FMN dissociation is not diminished at high NADH concentration, i.e. under the conditions where the “tunnel” leading to buried FMN (according to the atomic structure of T. thermophilus NDH-1 fragment [36]) is expected to be occupied by NADH. This observation points to an existence of an “other door” for FMNH^- dissociation different from the cleft where ferricyanide and NADH compete for accessibility to the flavin (see previous sections). The data described above can be briefly and somehow hypothetically summarized as follows. (i) Binding of FMN to the 51 kDa subunit within complex I significantly shifts its midpoint redox potential compared to the free FMN/FMNH^- couple [48] to the negative side (see also Refs. [34] and [30] for E_m values of FMN/FMNH^- couple in complex I as emerged from flavin semiquinone EPR studies and from the data on superoxide generation, respectively). (ii) The reduction of FMN results in weakened binding which may lead to its dislocation within the binding cavity coupled with conformational change of the protein. (iii) Accessibility of the flavin for hydrophilic substances including artificial electron acceptors and nucleotides seems to be dynamic and changes upon FMN oxidoreduction during enzyme turnover.

Fig. 2.

Reversible dissociation of FMN from the membrane-bound complex I. SMP were preincubated for 1 h at pH 10, 20^°C. The additions: 1, none (control); 2, 100 μM NADH; 3, 100 μM NADH and 1 mM NAD⁺; 4, 100 μM NADH and 10 μM FMN; 5, 10 μM FMN added after inactivation (sample 2) has been completed.

5. Concluding remarks

After brief discussion of the kinetic and equilibrium studies on NADH/NAD⁺ interactions with complex I natural question arises: does the simplified scheme shown in Fig. 1 adequately describe NADH (NAD⁺) oxidation (reduction) by complex I? It is doubtful that we can answer this question with our present state of knowledge. Some kinetic data substantiated by more direct quantitation of binding constants in the absence of the enzyme turnover (Table 2) agree with the scheme. The others, such as the unidirectional effect of ADP-ribose on forward and reverse electron transfer [26, 37], peculiar titration of NADH oxidation and NAD⁺ reduction by the site-directed inhibitor, significant apparent discrepancy between only weak inhibitory effect of NAD⁺ on NADH oxidation and low $K_m^{{\mathit{NAD}}^{+}}$ in reverse electron transfer are difficult to explain within this deliberately simplified scheme. The two nucleotide binding site model [35] which has been suggested for mammalian complex I to overcome these difficulties received no support from the structural data on T. thermophilus NDH-1 [36]. It is pertinent to emphasize that this model was originated from kinetic studies and the key question which should be answered to accept or discharge it is whether catalytically competent complex exists that contains two simultaneously bound nucleotides at two or single site(s) during catalysis of any particular reaction. It should also be noted that the two-site model was based on data obtained for a system where complex I operates as an intrinsic component of the mitochondrial respiratory chain and where natural heterogeneity of complex I within the inner mitochondrial membrane seems likely to exist. This heterogeneity may arise from the permanent or rapidly equilibrating interaction of a monomer-dimer complex I entity or from complex I-complex III [49] and/or complex I-complex II temporary or stable associations. The possibility that these interactions would affect the nucleotide binding properties of complex I can not be excluded.

Most of the kinetic data relevant to the nucleotide binding reported in the literature have been obtained for the preparations incapable of the major enzyme function, i.e. that is energy transduction. On the other hand, binding and dissociation of NADH/NAD⁺ coupled with FMN redox potential alteration accompanied by conformational change of the protein may be an essential energetic contributor to the overall proton-translocating mechanism. Detailed studies on nucleotide binding parameters for complex I operating as an energy-coupled device are evidently needed. Particularly, energy-linked reverse electron transfer, the reaction discovered almost fifty years ago [50] is still only superficially understood in terms of its kinetic parameters. Whether complex I operates in the forward and reverse directions via a series of the same intermediates (microreversibility principle) or the energy-transducing NADH:ubiquinone reductase and ubiquinol:NAD⁺ reductase activities are catalyzed by rapidly or slowly equilibrating “isoforms” of complex I is an open question. In my opinion, which was advocated for another energy-transducing enzyme, F_o·F₁-ATP synthase [51], complex I is a macromolecular electro-mechano-chemical machine and as any other man-made or nature-created machine it operates in forward and reverse directions in different modes.

The final note concerns the significance of the kinetic studies of complex I. The next step following recent breakthroughs in the field, i.e. x-ray resolution of T. thermophilus NDH-1 fragment [36], resolution of the atomic structures of complete prokaryotic and eukaryotic complexes is expected. However valuable the knowledge of these static structures are, they will not tell reveal dynamic interactions between the subunits during catalysis. To understand how a machine works we need to see it at work. For enzymes work is their turnover. Thus, quantitative studies of the enzyme turnover, i.e. enzyme kinetics were, are, and continue to be a powerful tool for bioenergetics.

Abbreviations

Acetyl-NAD⁺

3-Acetylnicotinamide adenine dinucleotide

DCIP

2,6-dichlorophenol indophenol

FMN

flavin mononucleotide

FP

2-3 subunit soluble fragment of complex I

HAR

hexaammineruthenium(III)

NDH-1

prokaryotic homologue of the mitochondrial NADH:ubiquinone oxidoreductase (complex I)

pmf

proton-motive force

SMP

submitochondrial particles

Q₁

2,3-Dimethoxy-5-methyl-6-[3-methyl-2-butenyl]-1,4-benzoquinone

The iron-sulfur centers are designated according to Ohnishi’s nomenclature.