The energy-converting NADH:ubiquinone oxidoreductase, also known as respiratory complex I, couples the transfer of electrons from NADH to ubiquinone with the translocation of protons across the membrane. Electron microscopy revealed the two-part structure of the complex consisting of a peripheral and a membrane arm. The peripheral arm contains all known cofactors and the NADH-binding site, whereas the membrane arm has to be involved in proton translocation. Owing to this, a conformation-linked mechanism for redox-driven proton translocation is discussed. By means of electron microscopy, we show that both arms of the Escherichia coli complex I are widened after the addition of NADH but not of NADPH. NADH-induced conformational changes were also detected in solution: ATR-FTIR (attenuated total reflection Fourier-transform infrared) of the soluble NADH dehydrogenase fragment of the complex indicates protein re-arrangements induced by the addition of NADH. EPR spectroscopy of surface mutants of the complex containing a covalently bound spin label at distinct positions demonstrates NADH-dependent conformational changes in both arms of the complex.
The energy-converting NADH:ubiquinone oxidoreductase, also known as respiratory complex I, couples the transfer of electrons from NADH to ubiquinone with the translocation of protons across the membrane [1–7]. In doing so, it contributes to the generation of the protonmotive force required for energy-consuming processes. The bacterial complex consisting of 14 subunits named NuoA–NuoN (or Nqo1–Nqo14) constitutes the minimal structural and functional framework for redox-driven proton translocation [7,8]. Seven peripheral subunits bear the redox groups of the complex, namely one FMN and, depending on the species, up to ten Fe–S (iron–sulfur) clusters. The remaining seven subunits are mostly hydrophobic proteins predicted to fold into 61 α-helices across the membrane and seem to be involved in quinone binding and proton translocation [8–13]. In addition, two semiquinone radicals named SQNf and SQNs have been identified as reaction intermediates [14,15].
Electron microscopy revealed the two-part structure of the complex consisting of a peripheral and a membrane arm [16,17]. The peripheral arm is made up of the seven soluble subunits, whereas the membrane arm consists of the seven hydrophobic subunits. Recently, the structure of the Thermus thermophilus peripheral arm of complex I was resolved at molecular resolution revealing the electron pathway from NADH to the most distal Fe–S cluster N2 . The quinone reduction site is most probably located at the interface between the peripheral and the membrane arms [19–21]. Owing to the considerable distance between the electron pathway and the proposed proton pathway, a conformation-linked mechanism for proton translocation is discussed [4,6,22–24].
It was demonstrated previously that the addition of either NADH or NADPH to the Escherichia coli complex I led to conformational changes identified by an altered cross-link pattern . It was assumed that these changes are caused by the substrate binding and/or the reduction of the complex. In fact, electron microscopy of an NADH-treated sample revealed that both arms of the complex were widened. The addition of NAD+ affected neither the cross-link pattern nor the shape of the particles. It was discussed that either binding of NADH or the reduction of the enzyme is responsible for the observed changes . The results of our approach for detecting nucleotide-induced conformational changes of E. coli complex I by means of electron microscopy, ATR-FTIR (attenuated total reflection Fourier-transform infrared) spectroscopy and SDSL (site-directed spin labelling) are summarized below.
Enzyme kinetics revealed the Km value of complex I for NADH to be 10 μM, while the Km for NADPH was determined to be 1.8 mM. Nevertheless, all EPR-detectable Fe–S clusters of the complex were reduced by the same amount with both NADH and NADPH. Thus NADPH is able to reduce the Fe–S clusters of the complex although it exhibits a low affinity for the enzyme. To examine whether both nucleotides induce conformational changes, we used electron microscopy to determine the overall conformation of the complex as isolated, after reduction with NADH, and after reduction with NADPH. To keep the reduced state of the NAD(P)H-treated samples, these were washed on the grid several times with the buffer containing 2 mM NAD(P)H. All three samples were negatively stained with uranyl acetate. The samples were analysed by single-particle image processing techniques. Some of the typical class averages are shown in Figure 1. The class averages were enlarged and contour lines were drawn around areas of the same density. The size of both arms of each class average was measured and the averages were calculated. The difference between the shape of the complex as isolated and after treatment with NADH as described in  was also detectable in our preparations (Figure 1). The average length of both arms measures 197±14 Å (1 Å=0.1 nm) for the complex as isolated and did not change for the NADH-treated complex with 204±12 Å. The apparent thickness of both arms increased from 70±5 to 82±6 Å as a result of the NADH treatment. Mamedova et al.  reported an increase in the thickness by 20 Å. This slight difference from our results may be attributed to a partial re-oxidation of the complex during staining with uranyl acetate or to a different threshold used to draw the contour lines. However, there was no obvious difference between the shape of the complex as isolated and the shape of the complex when treated with NADPH (Figure 1). The length of both arms in the preparation treated with NADPH measures 197±10 Å and their thickness 70±5 Å. According to the Student's t test, the probability that the difference in thickness between the arms of complex I when treated with NADH, NADPH or used as isolated is significant is >99%. Thus the addition of NADH widened both arms of the complex significantly, whereas addition of NADPH had no effect on the size of the complex. As both nucleotides reduced the Fe–S clusters of the complex, the widening of both arms is not due to the reduction of the cofactors.
Averages of classified complex I particles in the air-oxidized state (A), after reduction with NADPH (B) and after reduction with NADH (C)
The structural changes induced by the addition of NADH were detected with negatively stained single particles. To investigate these changes for the enzyme in solution, spectroscopic methods have to be applied. IR spectroscopy allows the determination of the secondary-structure elements based on the analysis of the amide I vibration that essentially includes the ν(C=O) vibration of the protein backbone [25,26]. Owing to the influence of the hydrogen-bonding environment on the specific spectral position of this vibration, it is possible to determine the relative contribution of each secondary-structure element.
Figure 2 shows the amide I band obtained for a 250 µM solution of the soluble NADH dehydrogenase fragment of E. coli complex I [27,28]. This fragment is made up of the subunits NuoE, NuoF and NuoG and represents the electron input part of the complex. It contains the FMN and six Fe–S clusters as well as the NADH-binding site. From the structure of the peripheral arm of the T. thermophilus complex, it is clear that the distribution of secondary-structure elements of the NADH dehydrogenase fragment in the oxidized state is 30% α-helical, 17% β-sheet and 53% random . Figure 2 shows the spectrum of the fragment as isolated (Figure 2A) and after reduction with a 1 mM NADH (Figure 2B), as well as each of the decomposed bands. The exact position of each component was obtained from the fourth derivative. Subsequently, a band fitting with mixed Gaussian and Lorentzian band shapes was performed. The overall distribution of the secondary-structure elements of the fragment as isolated with 25% α-helical, 16% β-sheet, 19% turns and 40% random structures corresponds well to the results obtained from X-ray crystallography . Small differences between the numbers are expected for the fragments from different species. On addition of NADH, the amount of the highly organized α-helical and β-sheet structural elements decreased by a negligible 2%, which is within the error of the method. Variations of 4–5% were seen for turns and random-type elements. NADH binding induced an increase in signals corresponding to turns, whereas the random character of the structure decreased. Clearly, on the binding of NADH, a conformational change took place in the fragment: a change that will mostly affect the hydrogen-bonding interaction. Thus the NADH-induced conformational changes were also detectable in the enzyme in solution.
IR absorbance spectra of the amide I band of the soluble NADH dehydrogenase fragment of the E. coli complex I as isolated (A) and after reduction with NADH (B)
Another powerful tool for monitoring the structure and dynamics of soluble and membrane proteins is SDSL in combination with EPR spectroscopy . For SDSL, accessible cysteine residues are introduced at distinct positions in a protein by means of site-directed mutagenesis. The thiol group is modified by a specific nitroxide reagent. A commonly used reagent is MTSL [(1-oxyl-2,2,5,5-tetramethylpyrroline-3-methyl)-methanethiosulfonate]. The spectral properties of the nitroxide probe containing an abundance of information are measured by EPR spectroscopy. The use of this technique for E. coli complex I has been impaired so far by the lack of a genetic system to easily introduce cysteine residues at any given positions of the enzyme. Recently, an overexpression plasmid was constructed in our laboratory that contains all nuo genes under the control of an arabinose promoter . The exposed N-terminus of subunit NuoF of the peripheral arm was chosen for engineering the complex with a hexahistidine tag by λ-Red-mediated recombineering . The E. coli complex I was overproduced from this construct in a strain which is devoid of any membrane-bound NADH dehydrogenases. The entire NADH oxidase activity of the cytoplasmic membranes of this strain is derived from the overproduced complex. After solubilization with dodecyl maltoside, the engineered complex binds to an Ni2+-iminodiacetic acid matrix, allowing the purification of approx. 11 mg of complex I from 25 g of cells. The preparation is pure and monodisperse and comprises all known subunits and cofactors .
Incubation of the native E. coli complex I with MTSL did not lead to a binding of the probe. Thus the native enzyme does not contain any accessible cysteine residues. Using the information provided by the structure of the peripheral arm of T. thermophilus complex I, several positions at various subunits of the peripheral arm of E. coli complex I were selected and mutated to cysteine residues. The structure of the membrane arm is still unknown. Using secondary-structure predictions, we identified several positions in bona fide loop regions of the hydrophobic subunits constituting the membrane arm. Again, cysteine residues were introduced at these positions. The mutations had no effect on the NADH oxidase activity of the mutant strains. The cysteine variants were purified by affinity chromatography and incubated with MTSL. Excess label was removed by size-exclusion chromatography. The samples were reconstituted in phospholipids and concentrated to approx. 5 mg/ml and used for EPR spectroscopy at room temperature. The NADH:decyl-ubiquinone oxidoreductase activity of the labelled complex I variants remained unchanged compared with the native enzyme.
Figure 3 shows the EPR spectra of the free label, the label bound to subunit NuoB of the peripheral arm and to subunit NuoM of the membrane arm. The spectrum of the free label shows three absorption lines due to the coupling of the free electron with the magnetic moment of the nitrogen nucleus. A nearly perfect isotropic spectrum was obtained by the fast and unrestricted thermal motion of the label (Figure 3A). When bound to a cysteine residue located in a predicted loop region of subunit NuoM, a similar spectrum of the probe is obtained with a slightly broadened line width caused by its attachment to the large protein complex containing lipids (Figure 3B) . However, the EPR spectral line shape indicated a high mobility of the loop region where the label is bound. In contrast, a highly anisotropic spectrum was obtained with the label bound to subunit NuoB, indicating very restricted motion (Figure 3C). The label is located in close proximity to the proposed connection between the peripheral and the membrane arm. The restricted motion is most likely to be a result of contact with the lipid bilayer. The spectrum of the label attached to the same position of the complex in detergent solution clearly showed a more isotropic signal (results not shown). The addition of NADH to the sample resulted in a slight broadening of the EPR spectral line shape of the label (Figure 3D). The small change is best visualized in a difference spectrum of the oxidized sample minus the NADH-reduced sample (Figure 3E). This small spectral change indicates an alteration of the microenvironment of the probe due to the addition of NADH. As the label is attached in close proximity to N2 , which is located on NuoB , the alteration of the microenvironment might be due to the reduction of the Fe–S cluster. To examine this possibility, the sample was reduced by an addition of NADPH, which led to a similar change of the spectral properties. Thus the observed effect is most likely to be caused by the reduction of the enzyme and not by the binding of NADH.
EPR spectra of MTSL at room temperature
It is well accepted that the redox reaction of complex I is accompanied by conformational changes. This was experimenally shown in the mitochondrial complex I from bovine heart by cross-link experiments and treatment with trypsin [22,34,35]. Conformational changes were not detectable when the complex was treated with NADH and K3(FeCN)6. From this it was concluded that the reduction of the enzyme induces the conformational changes and not the binding of the substrate [22,34]. The NAD(P)H-induced conformational changes were also detected in the E. coli complex I by Mamedova et al.  using cross-link experiments. In addition, our groups have recently shown by means of FTIR spectroscopy that the electrochemical-induced redox reaction of complex I in the absence of NAD(P)H is accompanied by conformational changes [36–38]. Therefore the FTIR data (Figure 2), as well as the experiments involving the labelling of the protein with a nitroxide probe (Figure 3), indicated that the reduction of the Fe–S clusters results in local conformational changes.
Furthermore, it was reported that the addition of NADH leads to an expansion of both arms of the E. coli complex . It was assumed that the difference in the cross-link pattern obtained by the addition of either NADH or NADPH is derived from the same effect reflected in the conformational changes detected by electron microscopy. Here, we have shown that the addition of NADH but not NADPH leads to an extension of the overall structure of the complex (Figure 1), although all cofactors of the complex are reduced by NADPH. Therefore it seems likely that the change in the cross-link pattern and the expansion of the overall structure observed by electron microscopy reflect different molecular processes. According to our results, the binding of NADH might induce long-range global conformational changes that are possibly linked to the mechanism of the complex. This would imply that the addition of NADH to complex I has two different effects: first, it leads to local conformational changes caused by the reduction of the cofactors, and secondly, it leads to global conformational changes by its binding to the complex. It is unclear whether and how both processes contribute to proton pumping, although it is widely accepted in the field that energy transduction by complex I is mainly accomplished by means of conformational changes.
Integration of Structures, Spectroscopies and Mechanisms: Second Joint German/British Bioenergetics Conference, a Biochemical Society Focused Meeting held at University of Edinburgh, U.K., 2–4 April 2008. Organized by Ulrich Brandt (Frankfurt, Germany), Steve Chapman (Edinburgh, U.K.), Peter Heathcoate (Queen Mary, University of London, U.K.), John Ingledew (St Andrews, U.K.), Mike Jones (Bristol, U.K.), Bernd Ludwig (Frankfurt, Germany), Fraser MacMillan (University of East Anglia, Norwich, U.K.), Hartmut Michel (Max-Planck-Institute for Biophysics, Frankfurt am Main, Germany), Peter Rich (University College London, U.K.) and John Walker (MRC Dunn Human Nutrition Unit, Cambridge, U.K.). Edited by Ulrich Brandt and Peter Rich.
This work was supported by the Deutsche Forschungsgemeinschaft, the Volkswagen Stiftung and the Université Franco-Allemande. We are grateful to Linda Williams for her help in preparing the manuscript.