Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Contents lists available at SciVerse ScienceDirect

The International Journal of Biochemistry & Cell Biology journal homepage: www.elsevier.com/locate/biocel

Review

Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump Marina Verkhovskaya ∗ , Dmitry A. Bloch Helsinki Bioenergetics Group, Institute of Biotechnology, PO Box 65 (Viikinkaari 1) FIN-00014 University of Helsinki, Finland

a r t i c l e

i n f o

Article history: Available online xxx This work is dedicated to the memory of Michael Verkhovsky, whose ideas and studies it is based upon. Keywords: Bioenergetics NADH:ubiquinone oxidoreductase Electron transfer Proton translocation Proton-pump

a b s t r a c t In respiring organisms the major energy transduction flux employs the transmembrane electrochemical proton gradient as a physical link between exergonic redox reactions and endergonic ADP phosphorylation. Establishing the gradient involves electrogenic, transmembrane H+ translocation by the membrane-embedded respiratory complexes. Among others, Complex I (NADH:ubiquinone oxidoreductase) is the most structurally complex and functionally enigmatic respiratory enzyme; its molecular mechanism is as yet unknown. Here we highlight recent progress and discuss the catalytic events during Complex I turnover in relation to their role in energy conversion and to the enzyme structure. © 2012 Elsevier Ltd. All rights reserved.

Contents 1. 2. 3.

4.

5.

6.

Introduction: mitochondrial and bacterial respiratory chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Function and overall structure of Complex I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reactions upon the catalytic turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Bimolecular reaction of Complex I with the substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Hydride ion transfer from NADH to FMN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. FMN oxidation and the unpairing of electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron transfer along the intramolecular redox chain of FeS clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. EPR signals of FeS clusters and their assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Thermodynamic properties of FeS clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Real-time electron transfer along the FeS cluster chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Spatial organization of the FeS clusters chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ubiquinone reduction coupled to the transmembrane proton translocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Ubiquinone in Complex I: one or two binding sites? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Mapping of the quinone-binding site(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Tightly bound ubiquinone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Models of the UQ redox transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Long-distance energy transduction coupled to proton translocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. The mechanical “piston” model of Complex I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Can the electron transfer and proton pumping be uncoupled? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Abbreviations: cyt, cytochrome; Eh , apparent redox potential; Em , midpoint redox potential; EPR, electron paramagnetic resonance; eT, electron transfer; N, negatively charged membrane side, facing bacterial cytoplasm or mitochondrial matrix; P, positively charged membrane side, facing bacterial periplasm or mitochondrial intermembrane space; pmf, proton-motive force; pT, proton transport; Q, quinone; QH2 , quinol; ROS, reactive oxygen species; TMS, transmembrane segment; UQ, ubiquinone; UQH2 , ˜ H + , transmembrane electrochemical potential of protons;  ˜ Na+ , transmembrane ubiquinol; H + , standard transmembrane electrochemical potential of protons;  electrochemical potential of sodium ions;  , transmembrane electric potential; ε, protein dielectric constant; , apparent time constant. ∗ Corresponding author. Tel.: +358 9 191 59749; fax: +358 9 191 58003. E-mail addresses: marina.verkhovskaya@helsinki.fi (M. Verkhovskaya), Dmitry.bloch@helsinki.fil (D.A. Bloch). 1357-2725/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biocel.2012.08.024

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 2

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

6.3. Rendering of protonatable residues in the Complex I membrane domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4. Electrostatically driven “wave-spring” model of Complex I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction: mitochondrial and bacterial respiratory chains Cell respiration is the most efficient energy-transforming system in mitochondria and in many respiring bacteria. The process couples the highly exergonic electron transfer (oxidation of respiratory substrates) to the highly endergonic formation of ATP from ADP and phosphate (often referred to as “ATP synthesis”; oxidative phosphorylation) and some other types of cellular work. The two general principles of the process are the following: (i) The source of energy for metabolism is the redox reaction between the electron donor and acceptor. The vast majority of electron donors (respiratory substrates) are derived from photosynthetically formed reduced organic compounds, e.g. carbohydrates; for the aerobic life, the almost universal electron acceptor is molecular oxygen; (ii) The physical link between redox and phosphorylation parts of the process is the transmembrane electrochemical potential of protons, ␮ ˜ H+ (proton-motive force, or pmf; in certain special cases in bacteria, also the gradient of Na+ ions, ␮ ˜ Na+ ). The ␮ ˜ H+ is established as a result of transmembrane translocation of H+ by the respiratory complexes (Complexes I, III, and IV) and consumed by H+ - (or Na+ -) motive ATP synthase. Since the electrogenic nature of proton translocation, both electric ( [mV]) and chemical (pH gradient, 59 × pH [mV] at +20 ◦ C) ˜ H+ . Although the two forms are components contribute to ␮ thermodynamically convertible, the primary form is, however, the electric potential established between the two aqueous compartments separated by the lipid bilayer membrane; the formation of pH requires work against buffering capacity of both transmembrane compartments.

(pH 7, 1 atm O2 ). The respective standard free energy drop for this reaction is given by G◦ (kJ/mol) = −n × 0.0961 × Em (mV) or G◦ (meV) = −n × Em (mV) where n is the stoichiometric number of electrons participating in the reaction. For the oxidation of one NADH molecule by O2 , G◦ = −119 kJ/mol (Table 1). Taking into account the experimentally measured value of ATP production in mitochondria (∼2.27 ATP per NADH Hinkle, 2005), this gives 52.5 kJ/mol per ATP molecule (a typical literature value for the formation of 1 ATP molecule from ADP and Pi under “average” cellular condition is 57 kJ/mol). For the lossless (ideal) electrochemical coupling, the electrogenic, transmembrane H+ translocation driven by a redox reaction with standard free energy G◦ causes standard electrochemical potential: |H+ ,max |(mV) =

Em ≈ (+820 mV) − (−320 mV) = +1140 mV

1 All redox potentials quoted are expressed versus normal hydrogen electrode (NHE) scale.

|G◦ | (meV) ≡ |Em |(mV) n

giving 1.14 V for the oxidation of one NADH molecule by O2 . At physiological conditions the value may differ due to the concentration ratio factors: G = G◦ − 59 × lg

× lg

[electron donorreduced ] [electron donoroxidized ]

[electron acceptorreduced ] [electron acceptoroxidized ]

+ 59

,

or | ˜ H+ ,max |(mV) =

About 90% of ATP production in the cell comes from oxidative phosphorylation coupled to cell respiration. Textbook values for oxidation of 1 molecule of glucose in mammalian mitochondria give 38 ATP molecules, 34 of which are formed by oxidative phosphorylation. (The real values may be slightly less due to the revised mechanistic H+ /P or P/e− ratios for H+ -ATPase Watt et al., 2010; Wikström and Hummer, 2012.) Note that in respiring bacteria the stoichiometries may differ significantly due to the variation in the mechanistic H+ /P ratio of H+ -ATPase (Steigmiller et al., 2008; Lau and Rubinstein, 2012). Most of the redox energy entering aerobic metabolism is supplied via NADH produced by the Krebs cycle (but also, to a lesser degree, from fatty acid oxidation, protein degradation, and glycolysis). The midpoint redox potential (Em ) values for glucose and NADH differ only slightly (see Table 1), so almost all energy is preserved in the form of NADH (contribution from FADH2 and succinate is negligible, in comparison, from the point of view of energy conservation). Energy provided from the oxidation of NADH by oxygen is given by the difference between the respective Em of the electron acceptor and donor1 :

00 00 00 00

|G| (meV) ≡ |Eh |(mV) n

and for the oxidation of NADH it is less than the standard value; however, keeping in mind that the electrical contribution in ␮ ˜ H+ predominates, the transmembrane electric potential is still rather large. Pure lipid is excellent electrical insulator; the resistance of a typical artificial bilayer lipid membrane is ∼1 G. Being directly applied across the lipid bilayer membrane (average thickness, ˚ defined as the distance between the acyl carbonyl groups, is 36 A) such  supplies electric field of ∼290 kV × cm−1 . However, in biological membranes containing many membrane-embedded or membrane-associated proteins, the tolerance of the bilayer to electric field, though dependent on the source of lipid, is typically much weaker. For example, in mitochondria or Escherichia coli cells, a typical maximum ␺ value sustained by the membrane ranges from 220 to 250 mV. At higher voltage the membrane loses its integrity and its conductivity sharply rises leading to a short-circuit and electrical damage. This poses a problem of efficient generation and use of the electric field coupled to the respiratory redox process; an obvious solution is to split the whole voltage span (1.14 V) into several, mechanistically coupled voltage generators, each of which can only generate smaller field. The latter principle is indeed realized in the real respiratory chain (Rich and Maréchal, 2010; Jastroch et al., 2010; Nicholls, 2010). Mitochondrial respiratory electron transfer chain (Fig. 1) consists of three enzymes catalyzing linear, sequential electron transfer (eT) from NADH through ubiquinol (UQH2 ) and cytochrome (cyt) c

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

ARTICLE IN PRESS

G Model BC-3836; No. of Pages 21

M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

3

Table 1 Estimated standard free energy drops (G◦ ) in mitochondrial respiratory electron-transfer chain (pH 7). Redox pair

Em (mV)

n

G◦ of oxidation by O2 (per substrate molecule)

Enzyme and G◦ per 2e−

−

Glucose + 6H2 O/6CO2 + 24H + 24e NADH/NAD+ + H+ + 2e− Succinate/fumarate + 2H+ + 2e− FADH2 /FAD + 2H+ + 2e− UQH2 /UQ + 2H+ + 2e−

−420 −320 +30 +50

cyt cRed /cyt cOx + e− 2H2 O/O2 + 4H+ + 4e−

+260 +820e

+

+90d

24 2 2 2 2 1 4

∼ −30 eV (−2880 kJ/mol) −2280 meV (−119 kJ/mol) −1580 meV (−152 kJ/mol) −1540 meV (−148 kJ/mol) −1460 meV (−140.3 kJ/mol) −560 meV (−53.8 kJ/mol) –

per H+ translocated Complex I

−820 meVa (−78.8 kJ/mol)

−205 meVb (−19.7 kJ/mol) −273 meVc (−26.3 kJ/mol) Cytochrome bc1 complex −170 meV (−16.3 kJ/mol) −340 meV (−32.7 kJ/mol) Cytochrome c oxidase −1120 meV (−107.6 kJ/mol) −280 meV (−26.9 kJ/mol)

Experimentally measured values in mitochondria, −718 to −728 meV (−69 to −70 kJ/mol) (Hinkle et al., 1991; Muraoka and Slater, 1969). Assuming 4H+ per NADH. c Assuming 3H+ per NADH. d Determined in the chromatophore membrane pool (Takamiya and Dutton, 1979). The aqueous EtOH solution value is +110 mV. e At 1 atm O2 . Although real [O2 ] in tissues or in bacterial habitats may vary, the [O2 ]-dependence of Em of the H2 O/O2 pair is weak: ten-fold change in [O2 ] will shift the Em only for ∼15 mV, which value is close to a typical experimental error of the Em determination in biological systems. a

b

to O2 . All three are ␮ ˜ H+ -generators; despite their similar function, they possess very different coupling mechanisms between electron transfer (eT) and proton transport (pT) providing different efficiency and proton pumping stoichiometries. The proton translocation mechanism is best known for cytochrome bc1 complex (Complex III; UQH2 :ferricytochrome c oxidoreductase) and cytochrome c oxidase (Complex IV). The former is strictly saying not a proton pump but a Mitchellian redox-loop; the latter is a proton pump with a direct coupling, charge compensation (for review, see e.g. Crofts et al., 2008; Wikström and Verkhovsky, 2007; Belevich and Verkhovsky, 2008). The bc1 complex and cytochrome c oxidase translocate 2 and 4 electrogenic protons (or charge equivalents) per 2e− , respectively. From the point of view of efficiency, these stoichiometries are consistent with the respective energy drops across each of the enzymes: −33 kJ/mol and −108 kJ/mol per 2e− ,

Fig. 1. Mitochondrial respiratory complexes. The respiratory complexes are arranged against the Em scale (left axis) according to the respective Em values of their substrates (NADH, UQ, cytochrome c and O2 ). The PDB entries used: 319V (Complex I), 2A06 (bc1 complex), 2DYR (cytochrome c oxidase), and 2B4Z (cytochrome c).

respectively (Table 1). For Complex I, the experimentally measured proton translocation stoichiometry is 4 per 2e− (or less, ranging from 3.4 to ∼4, but mechanistically interpreted as 4) (Reynafarje et al., 1978; Pozzan et al., 1979; Hinkle et al., 1991; Wikström, 1984; Galkin et al., 1999, 2006), except that in Bogachev et al. (1996), it was estimated 3. However, a recent reevaluation of the literature data based on the structure-derived H/P ratios for mitochondrial H+ -ATPase and careful measurements of P/O ratios in mitochondria proposed 3H+ /2e− as the likely value (Wikström and Hummer, 2012). Both values, 3 and 4, make Complex I a highly efficient proton pump, taking into account its standard energy drop (−79 kJ/mol per 2e− , Table 1). Despite that the latter value is rather high, Complex I can operate in reverse mode under conditions close to equilibrium, when the thermodynamics of the catalyzed redox reaction is counterbalanced by the steady-state pmf (see e.g. Scholes and Hinkle, 1984). However, the molecular mechanism of proton translocation by Complex I remains yet unknown. Besides Complex I, there are other “entry points” for the electrons into the respiratory chain at the level of quinone. In mitochondria these are succinate dehydrogenase (respiratory Complex II, the membrane-bound component of the Krebs cycle) and two FADH2 dehydrogenases. Bacterial respiratory chains present a variety of other, functionally similar enzymes, including NADH dehydrogenases. Often these enzymes are less energyefficient with comparison to Complex I (compare e.g. the Em values for succinate/fumarate and FADH2 /FAD pairs with the one for NADH, Table 1) and are used to feed electrons into the high redoxpotential segment of the respiratory chain. Despite the fact that soluble NADH dehydrogenases are numerous and ubiquitous, only 3 membrane-bound NADH-oxidizing enzymes participating in energy metabolism are known. One is Complex I found in the inner mitochondrial membrane and in many bacteria; the other two are non-coupled NDH-2-type and the bacterial NQR (Kerscher et al., 2008). Note that NQR, is not H+ -, but in fact ˜ Na+ (Verkhovsky Na+ -motive, contributing to the formation of  and Bogachev, 2010; Verkhovsky et al., 2012b). Mitochondrial Complex I is an important source of cellular generation of superoxide (O2 •− ), which is a significant part of the reactive oxygen species (ROS) generated in intact mammalian mitochondria in vitro (see for review Murphy, 2009), although production of hydrogen peroxide by Complex I was also demonstrated experimentally (Pryde and Hirst, 2011). Intracellular ROS production is implicated in many mitochondrial pathologies, neurodegenerative diseases, oxidative stress, aging, and Parkinson’s disease; there exists a vast amount of recent literature on the medical aspects of the function of mitochondrial Complex I and we leave the subject outside the scope of the present review.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 4

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

2. Function and overall structure of Complex I Although Complex I was first isolated from mitochondria 50 years ago (Hatefi et al., 1962), progress in its studies has been strongly hampered by its extraordinary size (>900 kDa) and complexity (up to 46 different subunits) (Carroll et al., 2002). A first breakthrough in the Complex I studies was made when bacterial homologues of mitochondrial Complex I performing the same function were accepted as the model system. Bacterial Complex I is about half the size of the mitochondrial enzyme and is composed of only 14 (13, in E. coli, see Table 2) subunits representing a “minimum” functional version of the mitochondrial enzyme and performing the same function of coupling redox reaction to proton translocation. Basic properties of Complex I and its functional models described below are equally applicable both to prokaryotic and eukaryotic variants. Besides the relative simplicity, bacterial Complex I presents another advantage of its relatively easy genetic modification. Among the 14 core subunits essential for energy transduction and highly similar in pro- and eukaryotes, seven hydrophilic are nucleus-encoded, the other seven hydrophobic membrane subunits are encoded by mitochondrial genome; the genes are arranged into two clusters consisting of 2 and 5 genes, respectively, separated by tRNAencoding genes (all eukaryotic auxiliary subunits are encoded in the nucleus). Mutations in the core subunits often result in a number of severe neurodegenerative diseases (see for the review e.g. DiMauro, 2001; Petruzzella et al., 2012), however the information of how Complex I structure and function are affected is rather limited. Complex I catalyzes two-electron NADH oxidation and ubiquinone (UQ) reduction coupled to the transmembrane translocation of 3 or 4H+ (see above) from negatively charged side (N-side, cytoplasm or mitochondrial matrix) to positively charged side (P-side, periplasm or mitochondrial intermembrane space) of the membrane per 2 electrons. Two domains, hydrophilic and hydrophobic, constitute Complex I. The hydrophilic domain of Complex I contains noncovalently bound FMN and 8–9 FeS clusters, 8 of which are organized as a continuous eT chain connecting FMN and a UQ binding site. One or two UQ-binding sites are located at the interface between the hydrophilic and membrane Complex I domains or in the membrane domain close to the interface area. The hydrophilic domain is composed of 6 or 7 core subunits and protrudes to cytoplasm or mitochondrial matrix at approx. 130 A˚ (Efremov et al., 2010) (see Table 2, for Complex I subunit numbering). It is generally accepted that the eT through the FeS cluster chain occurs without the energy drop (see below), which chain, therefore, functions as an “electron wire”. Membrane domain of bacterial Complex I consists of 7 subunits equivalent to core subunits of mitochondrial enzyme (see Table 2). Very little is still known about the molecular mechanism of energy conversion by Complex I, in contrast to the two other redoxdriven membrane proton-pumps (␮ ˜ H+ -generators) comprising the mitochondrial respiratory chain: cytochrome bc1 -complex and cytochrome c oxidase (see e.g. Crofts et al., 2008; Yu et al., 2008; Belevich and Verkhovsky, 2008; Wikström and Verkhovsky, 2007; Kaila et al., 2010; Osyczka et al., 2004). The second breakthrough on the way of solving the Complex I mechanism was achieved with the resolution of X-ray crystal structures of hydrophilic (Berrisford ˚ and and Sazanov, 2009; Sazanov and Hinchliffe, 2006) (3.1–3.3 A) ˚ sephydrophobic domains (Efremov and Sazanov, 2011b) (3.0 A) arately, and also of the whole Complex I with lower resolution ˚ and (Hunte et al., 2010) (6.3 A). ˚ Since (Efremov et al., 2010) (4.5 A) the catalyzed chemical reaction (NADH oxidation and the electron transfer to UQ) occurs in the hydrophilic domain, whereas the proton translocation is performed by membrane subunits (see

Fig. 2. Schematic representation of Complex I from Thermus thermophilus (PDB entries 2FUG and 3M9S). Red lines indicate the electron transfer path from FMN to ubiquinone through the FeS cluster chain. Upon the oxidation of one NADH molecule 4H+ are translocated across the membrane from N-side (cytoplasm, equivalent to the mitochondrial matrix) to P-side (periplasm, equivalent to the mitochondrial intermembrane space).

Fig. 2), the two functions are spatially separated by a distance of ˚ nevertheless, both functions are mechanistically more than 100 A; well-coupled. Therefore, the enzyme’s molecular mechanism differs from the better studied mechanisms both in cytochrome c oxidase and bc1 complex. 3. Reactions upon the catalytic turnover The reactions in Complex I can be divided into the several types. These catalytic events are discussed in relationship with their role in energy conversion and the enzyme structure. 3.1. Bimolecular reaction of Complex I with the substrate The substrate binding site is located in the open cleft on the surface of protein (Berrisford and Sazanov, 2009). The conserved residues aligning this solvent-accessible cavity form an unusual Rossmann fold, which provides tight packing of the substrate, ensures the planar condensed system of the nicotinamide and the FMN isoalloxazine rings and therefore determines high affinity to NADH, substrate specificity and high rate of hydride transfer to FMN (Berrisford and Sazanov, 2009). During the forward reaction catalysis, the Complex I affinity to the substrate NADH and product NAD+ differ for more than 2 orders of magnitude: apparent NADH falls in micromolar range, e.g. 1.4 ␮M for bovine Complex Km I (Grivennikova et al., 2003), 5.1 ␮M for Paracoccus denitrificans (Kotlyar and Borovok, 2002) and 10.4 ␮M for E. coli (Euro et al., + 2009a) enzymes, whereas KiNAD is close to 1 mM for both bovine (Sled and Vinogradov, 1993) and E. coli (Euro et al., 2009a) enzymes. However, upon the reverse reaction, NAD+ reduction by Complex I supported by succinate oxidation in submitochondrial or subbacterial vesicles, the NAD+ affinity is significantly higher: apparent NAD+ is 7 ␮M for the bovine enzyme (Grivennikova et al., 2003) Km and 19.6 ␮M for the P. denitrificans (Kotlyar and Borovok, 2002) enzyme. The observed phenomenon is not explained yet, although the studies using highly potent and specific Complex I inhibitor, NADH-OH (Kotlyar et al., 2005), allowed to suggest that there are redox-dependent changes in nucleotide binding site of Complex I affecting the nucleotide affinity (Grivennikova et al., 2007). We have previously reported the bimolecular rate constant for NADH oxidation to be 8.3 × 107 M−1 s−1 in the E. coli enzyme,

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

ARTICLE IN PRESS

G Model BC-3836; No. of Pages 21

M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

5

Table 2 Numbering of the Complex I core subunits and cofactors from different sources a . Subunits E. coli

Hydrophilic domain NuoE

Cofactors Paracoccus denitrificans, Thermus thermophilus

a

c d

EPR signal assignment, consensus

EPR signal assignment according to Yakovlev et al. (2007) and Bridges et al. (2012)

Nqo2

24 kDa

[2Fe–2S]

N1a

N1a

NuoF

Nqo1

51 kDa

FMN [4Fe–4S]

– N3

– N3

NuoG

Nqo3

75 kDa

[2Fe–2S] [4Fe–4S] [4Fe–4S] [4Fe–4S]b

N1b N4 N5 N7b

N1b N5 Not detected –

NuoCDc (C) NuoCDc (D)

Nqo5 Nqo4

30 kDa 49 kDa

– –

NuoI

Nqo9

TYKY

[4Fe–4S] [4Fe–4S]

N6a N6b

N4 Not detected

PSST

[4Fe–4S]

N2

N2

Nqo6 NuoB Hydrophobic (membrane) domain Nqo7 NuoA NuoJ Nqo10 NuoK Nqo11 Nqo8 NuoH Nqo12 NuoL Nqo13 NuoM NuoN Nqo14 b

Bos taurus

d

ND3 ND6d ND4Ld ND1d ND5d ND4d ND2d

– – – – – – –

Only those subunits homologous between the bacterial and mitochondrial Complex I (“core” subunits) are shown. Only shown for E. coli and T. thermophilus. Genes for subunits C and D are fused in E. coli. Coded by mitochondrial DNA (mtDNA).

whereas the dissociation of the product NAD+ is much slower ( ≈ 1–2 ms) and is likely to be rate-limiting for the forward reaction (Verkhovskaya et al., 2008). The latter fact can be functionally important to prevent O2 •− production by the enzyme-bound FMN. Multiple NADH oxidation can lead to the accumulation of electrons in the FeS cluster chain in the situation when the electron-acceptor side (UQ) works slower than the electron-donor side (NADH) or the electron transfer chain. At least some of the redox transitions of UQ are supposed to be connected with (highly activated) conformational changes, whereas both NADH oxidation and eT through the FeS cluster chain are largely activation-less. Slow NAD+ release should prevent overreduction of the FeS chain and lower the risk of O2 •− production.

3.2. Hydride ion transfer from NADH to FMN Although the mechanism of the Complex I-bound FMN reduction by NADH is not fully understood, the possibility of one-electron oxidation of NADH can be excluded since the one-electron reduced form of NAD+ (a highly unstable radical NAD• ) was never observed. There is a consensus in modern literature that in a majority of redox-active flavoenzymes, the flavin is reduced in a single twoelectron transfer reaction as a concerted hydride ion transfer (tunneling) from a substrate molecule, rather than through a flavin semiquinone (Fl•− ) as a kinetic intermediate (Fig. 3) (see, e.g. Kohen and Klinman, 1999; Meijers and Cedergren-Zeppezauer, 2009; Sumner and Matthews, 1992). Hydride transfer was described, e.g.

Fig. 3. Two-electron oxidation of NADH by the enzyme-bound FMN. Modified from the reductive half-reaction mechanism known in methylenetetrahydrofolate reductase (Elizabeth E. Trimmer, personal communication; see also e.g. Trimmer et al., 2005 and references therein).

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 6

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

for glucose oxidation by FAD in glucose oxidase as early as in 1969 (Bright and Appleby, 1969). The possibility of hydride transfer from nicotinamide to flavin was first shown in a chemical system in solution (Powell and Bruice, 1983). When the flavin is bound to the protein, such mechanism requires that the hydride donor and acceptor are located in a close proximity; this is achieved by stacking the nicotinamide ring against the flavin isoalloxazine ring system, indeed crystallographically shown for Complex I (Berrisford and Sazanov, 2009; Sazanov and Hinchliffe, 2006). The hydride transfer tested with different substrates is fully reversible (see e.g. Basran et al., 2003) making the mechanism also likely for the reversed catalysis by Complex I. It has been shown for many flavin-reduction systems (see e.g. Leskovac et al., 2005) that the hydride can be mobilized and the hydride transfer becomes faster by polarizing the respective C H bond; therefore one can also expect such effect by abstraction a positive charge from the nicotinamide N–R group. Since in the NADH binding site in Complex I, both ribose and the ADP adenine residues are tightly stacked to the protein structure (Berrisford and Sazanov, 2009; Sazanov and Hinchliffe, 2006), such effect cannot be ruled out. The midpoint redox potential (Em ) for two-electron reduction of free FMN ranges from −207 mV (Mayhew, 1999) to −221 mV (pH 7) (Alberty, 1998); Em for the NADH/NAD+ pair is −320 mV (pH 7). In Complex I, however, the acidic surrounding of bound FMN shifts its Em to the more negative values, −340 (Sled et al., 1994) or −350 (Euro et al., 2008b) mV (Fig. 4). We have found that the replacement of one of these residues, Glu95, for neutral Gln in the NuoF subunit of E. coli Complex I located at a distance of ∼5 A˚ from the FMN isoalloxazine ring results in a positive shift of FMN Em of 40 mV (Euro et al., 2009a). These acidic residues most probably are the target for lanthanide ions, which strongly inhibit oxidation of 3+ NADH (KiLa = 1 ␮M). We have shown that the sensitivity to La3+ is strongly decreased in the mutant E95Q enzyme (Euro et al., 2009b). 3.3. FMN oxidation and the unpairing of electrons Two-electron redox centers with aromatic (homo- or heterocyclic) structure are typically capable of one-electron transitions

Fig. 4. Catalytic (FMN-binding) site in Nqo1 (NuoF in E. coli) subunit of Complex I from T. thermophilus (RDB ID: 2FUG). Only conserved acidic amino acid residues surrounding FMN are shown. These residues are proposed to lower the flavin Em from −210 mV for free FMN in aqueous solution to −350 mV for Complex I-bound FMN. Replacement of Glu95 for Gln in NuoF subunit in E. coli (counterpart of Glu95 in T. thermophilus) resulted in a positive Em shift of FMN for 40 mV (Euro et al., 2009a).

via formation of semiquinone species; among these compounds flavin is a rather unique one since its two- and one-electron transitions have close Em values at physiological pH. Therefore, flavin can act as an efficient converter between two- and one-electron eT steps. In Complex I FMN couples eT between an obligatory twoelectron donor, NADH, and a strictly one-electron acceptors, FeS clusters: the two sequential eT steps both have low Em values, −414 and −336 mV for the 1st and 2nd electron, respectively, at pH 7.5 (Sled et al., 1994). The intermediate, flavosemiquinone radical, was indeed observed, although its population was only 2% upon the reduction of Complex I by NADH (Sled et al., 1994). Since there is a very small difference between Em values for both FMN transitions and the one of NADH, clearly, no energy drop occurs upon NADH oxidation by FMN reduction, so this catalytic event cannot be directly involved in the transmembrane proton translocation.

4. Electron transfer along the intramolecular redox chain of FeS clusters 4.1. EPR signals of FeS clusters and their assignment The redox chain of FeS clusters bound to the hydrophilic domain of Complex I transfers 2 electrons from FMN to UQ. Since FeS clusters are one-electron carriers, both loading of the chain at the FMN end and release at the UQ end can only proceed sequentially, 1 electron at a time. The FeS chain from mitochondria and bacteria consists of 6 tetranuclear and 2 binuclear clusters, although some bacterial enzymes (from E. coli and Thermus thermophilus) contain an additional tetranuclear cluster (Fig. 5A, see Table 2, for the FeS cluster numbering). The additional cluster, N7, is in fact not a member of the chain and it is located more than 20 A˚ away from its closest neighbor within the chain. The latter distance is too large for the efficient electron transfer. It is unlikely that the N7 cluster takes part in the enzyme turnover; however, it does play a role in the enzyme stability (Pohl et al., 2007). The N7 cluster has been suggested as an evolutionary remnant from one of the Complex I ancestors, soluble Ni–Fe hydrogenase (Efremov and Sazanov, 2012). The other 7 clusters (with the exception of one binuclear cluster, N1a, see below) form a continuous electron transfer chain, ˚ some 130 A-long, connecting the FMN residue at its cytoplasmic electron-acceptor end with the enzyme’s membrane domain at its electron-donor end (Fig. 2). FeS clusters from Complex I have been studied by lowtemperature electron paramagnetic resonance (cryo-EPR) spectroscopy for many years, starting from pioneering works by Beinert et al. (1963) and Orme-Johnson et al. (1971). EPR spectra of Complex I taken at various conditions (temperature, microwave power, and ambient redox potential (Eh )) provide a rich set of information about the electronic structure and redox state of the FeS clusters. However, there is no consensus in the literature yet about the assignment of EPR spectral signatures to every individual FeS cluster (see for discussion Ohnishi and NakamaruOgiso, 2008; Yakovlev et al., 2007). NADH- or dithionite-reduced Complex I (when all or at least most of the FeS clusters are reduced) is often used to obtain “sample-characteristic” EPR spectra, which, however, differ significantly depending on the source of the enzyme; there is variation in the g-values for individual FeS clusters even among the eukaryotic mitochondrial Complex I from fungi, Neurospora crassa (Schulte et al., 1999) and Yarrowia lipolytica (Djafarzadeh et al., 2000), and mammal, Bos taurus (Reda et al., 2008; Roessler et al., 2010). The difference with bacterial complexes is even stronger; for example, it is not yet solved why there is no obvious EPR signal from one of the binuclear cluster, N1a, observed in the mitochondrial enzyme, although it is well resolved in the bacterial enzyme and can also be observed when the a small subunit

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

7

Fig. 5. FeS clusters of Complex I and their EPR spectra. (A) Intramolecular redox chain in the hydrophilic domain of Complex I from T. thermophilus (RDB ID: 319V). Round center-to-center distances, Å, are shown by numbers. (B) Unsaturated X-band EPR spectrum of purified Complex I from E. coli (10 mg/ml) anaerobically reduced by NADH. Conditions: temperature, 12 K, microwave power, 40 ␮W; other conditions, as in (Euro et al., 2008b). (C) Simulated individual spectra for the binuclear clusters (N1a and N1b) and tetranuclear clusters (N2, N3, N6b, and unidentified cluster Nx). Modified from Euro et al. (2008b) with permission (Copyright (2008) American Chemical Society).

24 kDa containing this cluster is isolated (Reda et al., 2008). Two possible explanations for these facts can be suggested, as following: (i) in the intact mitochondrial Complex I the g-values of N1a cluster are shifted in comparison with the isolated subunit 24 kDa so that its spectrum is completely overlapped with the other [2Fe–2S] cluster, N1b, as it was suggested by Albracht et al. (1977) based on the EPR spectra quantification; (ii) the Em for N1a in the intact mitochondrial Complex I becomes so low that it cannot be reduced even by such low-potential reductant as Eu-DTPA (Em −1.14 V) (Reda et al., 2008). Prokaryotic Complex I is much more fragile than the eukaryotic mitochondrial enzyme: to the best of the authors’ knowledge, prokaryotic Complex I has been successfully purified only from five eubacteria (E. coli Leif et al., 1995; Sinegina et al., 2005; Sazanov et al., 2003), P. denitrificans (Yip et al., 2011), two thermophiles, Rhodothermus marinus (Fernandes et al., 2006) and T. thermophilus (Efremov et al., 2010), a hyperthermophile, Aquifex aeolicus (Scheide et al., 2002) and a cyanobacterium, Thermosynechococcus elongatus (Zhang et al., 2005). EPR spectra of the entire reduced bacterial Complex I were obtained only for E. coli (Leif et al., 1995; Sinegina et al., 2005; Belevich et al., 2007; Euro et al., 2008b; Verkhovskaya et al., 2008) and R. marinus (Fernandes et al., 2006) and they are quite different. All these differences most probably arise from the g-tensor sensitivity to the immediate surroundings of the FeS clusters variable in different species and well seen by EPR indicating that the electron-transport function of the FeS chain is rather robust regardless differences in the fine individual properties of its components. Of a primary importance are, however, the thermodynamic characteristics and distances between the FeS clusters, the factors determining the eT rates (see below). Fig. 5B shows a typical unsaturated X-band EPR spectrum at 12 K, of the NADH-reduced, purified Complex I from E. coli obtained in our group. Considered formally, the spectrum should represent a combination of up to eight overlapping individual signals, one from each of Complex I’s FeS clusters (Fig. 5A); however, the exact number of the individual signals cannot be determined a priori because

(i) it is not clear whether the “non-functional” cluster N7 can be reduced at all; (ii) not all clusters can be fully reduced; and (iii) some very fast-relaxing clusters can only be detectable at lower temperatures. Indeed, at 5.5 K an additional signal is observed, deriving most probably from one of the “inner” clusters, N4 or N5. Therefore, to decompose such complex spectra we used selective temperature and microwave conditions as described in (Ohnishi, 1998) as well as Eh ; as the result, 6 typical individual signals can be identified (Fig. 4C). From these 6 signals the signals belonging to the five clusters located on the flanking regions of the FeS-chain (“external” clusters N1a, N3, N1b, N6b, N2) are well-attributed. The two slowrelaxing binuclear clusters, N1a and N1b, well described previously (Ohnishi, 1998), have different type of symmetry; their signals are rhombic and axial, respectively (Fig. 5C). Among the fast-relaxing tetranuclear clusters, the terminal one, N2, was shown to have the slowest relaxation rate, its signal was simulated (Leif et al., 1995) and its assignment was proved by the site-directed mutagenesis when the replacement of amino acid residues in the vicinity of N2 strongly changed its properties (Ahlers et al., 2000; Kashani-Poor et al., 2001; Zwicker et al., 2006; Belevich et al., 2007; Flemming et al., 2003a,b). The properties of a neighboring cluster, N6b, were also altered by an amino acid substitution in its vicinity (Belevich et al., 2007); its g-values obtained in our group in the whole enzyme (Belevich et al., 2007; Verkhovskaya et al., 2008) well correlate with those obtained with “connection fragment” of Complex I from E. coli containing only N6a, N6b and N2 clusters (Rasmussen et al., 2001) or Nqo9 subunit of Complex I from Paracoccus denitficans containing clusters N6a and N6b (Yano et al., 1999) (see Table 1, for Complex I subunit numbering). Note that Roessler et al. (2010 PNAS) obtained different assignment for the EPR signals from bovine Complex I (see Table 2). The [4Fe–4S] cluster N3, the immediate electron acceptor from FMN, has the least variable g-tensor among all clusters in bacteria and there is practically consensus on its EPR signal assignment (Leif et al., 1995; Velazquez et al., 2005; Yano et al., 1996; Euro et al., 2008b). The attribution of the characteristic EPR signals to the clusters N4, N5 and N6a located in the middle part of

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 8

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

the chain (“inner” clusters) is still under discussion (Yakovlev et al., 2007; Ohnishi and Nakamaru-Ogiso, 2008). Sixth simulated signal, Nx, (Fig. 5C) derives from one of them or from “non-functional” N7 cluster. 4.2. Thermodynamic properties of FeS clusters To establish whether there is an energy conversion step linked to eT along the FeS chain the redox properties of the individual centers were studied using submitochondrial particles (Ingledew and Ohnishi, 1980; Zwicker et al., 2006), bacterial membranes (Meinhardt et al., 1987b; Leif et al., 1995; Yano et al., 2003), or isolated subunits expressed from plasmids (Yano et al., 2003; Zu et al., 2002). The first two systems have the advantage of possessing the intact Complex I, but the disadvantage that the relatively small EPR signals from the FeS clusters are overlapped with other signals from redox-active, membrane-bound FeS proteins, such as succinate dehydrogenase. The isolated subunits have the advantage of well-detected, clear signals; however, the obtained g-values and Em values may differ from those in the intact protein due to possible perturbation of the cluster environment. As a result, there is no consensus in the contemporary literature on the Em values of the FeS clusters of Complex I. It was widely accepted previously that most of the clusters are equipotential with Em ≈ −270 mV (Ohnishi, 1998; Hirst, 2005) with the two following exceptions: (i) cluster N2 has the highest Em among all clusters, although varying results were reported ranging from −30 (Krishnamoorthy and Hinkle, 1988) to −220 mV (Leif et al., 1995); (ii) cluster N1a has an uncertain Em varying with the source of the enzyme ranging from −150 (Meinhardt et al., 1987a) to −400 mV (Zu et al., 2002). Recently we have accomplished an EPR spectroelectrochemical redox titration of FeS clusters in purified Complex I from E. coli. The redox spectra were decomposed using the simulated individual signals shown in Fig. 5C and for each of them the redox titration curves were obtained. We found that (i) the centers are not equipotential and (ii) for the three clusters of seven the titration curves could not be fitted by one-electron Nernstian curves but had more complex nature as a combination of at least two one-electron Nernstian curves (Euro et al., 2008b). The latter phenomenon evidently comes from the electrostatic interaction between the neighboring clusters in the FeS chain. Center-to-center distances within the pairs in the chain N3–N1b–N4–N5–N6a–N6b–N2 are all within the range of 12–17 A˚ (Sazanov and Hinchliffe, 2006), indicating significant electrostatic interaction. Such interaction would shift the Em of neighbor cluster to more negative values on arrival of an electron to a particular FeS cluster, due to electrostatic repulsion between two negative charges (for theory of redox interaction in multicenter proteins, see Ullmann, 2000; Hill, 1985 and references therein). It is obvious that such a shift of Em in a chain of redox centers will be stronger for the “inner” chain elements than for “external”, flanking clusters. Taking only four redox centers we developed a model to illustrate such behavior of the system, which showed that the redox titration curve should have complicated shape depending on the position of the cluster in the chain and Em of its neighbors and the negative Em shift due to electrostatic interactions could be over 100 mV (Euro et al., 2008b). The results of the titration are summarized in Fig. 6. Cluster N1a has relatively high Em = −235 mV and behaves as one-electron carrier, its marginal location explains lack of the electrostatic interaction. The titration curve of immediate electron acceptor from FMN, N3 cluster, can also be fitted with one-electron curve with Em = −315 mV. For each of the other three clusters (N1b, N6b, and N2) a pair of apparent Em values (intrinsic and those with interaction, respectively) was obtained: −245 and −320 mV for N1b, −235 and −315 mV for N6b, and higher values of −200 mV and −300 mV for the terminal cluster N2. The Em of the other, “inner”

Fig. 6. Potential energy profile for the Complex I substrates and cofactors. Double bars indicate the splitted redox potentials for clusters thermodynamically interacting with their neighbors: the higher values are close to the intrinsic Em values (those without interaction); the lower values include interaction. Broad bars show redox potentials known only approximately. Redox titration data are taken from Euro et al. (2008b).

clusters (N4, N5, and N6a) cannot be unambiguously assigned but all observed redox transitions happened at Eh between −330 and −365 mV (Euro et al., 2008b). Much higher Em (∼ −250 mV) of the clusters N3, N4 and N6a in bovine mitochondrial Complex I were suggested by Hirst (2010) and Bridges et al. (2012). However, there is no direct experimental evidence (e.g. provided by redox titration) for such redox transitions of these clusters. The electrostatic interaction energies between the FeS clusters were theoretically calculated by Stuchebryukhov and colleagues (Couch et al., 2009) based on the structure of the hydrophilic domain of Complex I (Sazanov and Hinchliffe, 2006) and experimental redox titration data (Euro et al., 2008b). The interaction energies were subsequently used by the same group for the developing of the model considering a variety of possible assignments of the redox titration data to specific clusters. Using statistical analysis the best fit of calculated and observed titration curves was found (Medvedev et al., 2010). The model let the authors to calculate intrinsic (without interaction) Em values for all redox centers of the chain. The obtained values for clusters N1a, N2, N6b were in good agreement with the experimental, upper Em values reported in Euro et al. (2008b); the calculated Em for cluster N3 were also similar to the experimental value, and for cluster N4, slightly higher (∼ −280). The cluster N6a remained unassigned; one reason for the latter can be its unusually strong interaction leading to an extraordinary low Em , although we have never observed any transition lower than −365 mV. Bridges et al. (Bridges et al., 2012) also reported clusters N6b as well as N1a, N5 are oxidized (and thus, EPR-silent) in the NADH-reduced B. taurus enzyme due to their low Em . Note that such low intrinsic Em values for the inner clusters may impose a kinetic problem (see below) for the eT rates, particularly if the slowest step, N5 → N6a, becomes a considerably uphill reaction due to a very low Em for N6a. The discrepancy between the calculated and experimental redox potentials for the inner clusters can be also accounted for the value of protein dielectric constant (ε) used for the calculations (ε = 20): in reality ε can be distributed unevenly within the hydrophilic Complex I domain accommodating the FeS chain. The thermodynamic approach when the redox properties of FeS clusters are determined at equilibrium has many difficulties such as complicated EPR spectra decomposition, assignment of

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

individual signals to particular clusters and electrostatic interactions that change Em as well as saturation properties of the clusters. Nevertheless this approach clearly shows the lack of significant energy drop upon the eT through the FeS chain, except for certain drop between the two last clusters, N6b and N2. Due to the latter finding and also the observed pH-dependence of Em for cluster N2 in mitochondrial Complex I (Ingledew and Ohnishi, 1980) it was suggested that this cluster could participate in UQ protonation (Berrisford and Sazanov, 2009; Treberg and Brand, 2011). However, the replacement of histidine residue, which proved to be the N2 redox-Bohr group, by methionine resulted in a negative shift of N2 Em and the complete loss of its pH dependence; meanwhile, the mutation did not significantly affect the catalytic activity and yielded in unchanged pumping stoichiometry (Zwicker et al., 2006). Therefore, a specific role of cluster N2 in the proton translocation seems unlikely. 4.3. Real-time electron transfer along the FeS cluster chain Electron transfer in Complex I was studied by the ultrafast freeze-quench approach. Complex I was mixed with NADH by means of a stopped-flow mixing system, and then at certain time points the sample was transferred to the freeze-quench apparatus and rapidly frozen. Freezing at the liquid nitrogen temperature was proven efficient to trap the nonequilibrium electron distribution among the FeS clusters of Complex I without letting the reduced clusters anneal on much longer timescales (Belevich et al., 2009), despite the theoretical possibility of such annealing (for theory and discussion, see e.g. Venturoli et al., 1998; Wang et al., 2009, and references therein). The redox state of FeS clusters was analyzed in the frozen samples by cryo-EPR at temperatures from 4 to 45 K (Verkhovskaya et al., 2008; Belevich et al., 2009). The data show that after binding of one NADH molecule the first pair of electrons is transferred simultaneously to the two FeS clusters N1a and N2 spatially separated by ∼100 A˚ and both having the highest Em in the FeS chain with an apparent time constant  ≈ 100 ␮s. The fast reduction of N1a and N2 is followed by a slow component ( ≈ 1–2 ms) consisting of the reduction of other FeS clusters (Fig. 7). The reduction rate of the terminal cluster N2 is close to the one predicted by electron transfer theory, which claims for the absence of coupled proton transfer or conformational changes during eT along the whole chain from FMN to N2 (Moser et al., 2006; Page et al., 1999). For the eT step across the largest distance between any two adjacent FeS clusters in the chain (N5 and N6a, edge-to-edge N6a − ˚ and the free energy drop, G = −|e− | × (Em distance, 14.1 A, N5 ) ≈ 0 meV) the predicted time constant () value ranges from Em 5 to 20 ␮s depending on the reorganization energy () value ranging from 0.5 to 0.7 eV (being largely unknown,  has a typical value of 0.5 for tight membrane proteins such as photosynthetic reaction centers or of 0.7 for soluble globular proteins (Chamorovsky et al., 2007; Page et al., 1999). Since the exact Em values are not known for N5 and N6a, the  will slightly vary if G = / 0 and it may become slower for the uphill reaction. For example,  of 250–1000 ␮s, which is way beyond the experimentally determined rates, is expected at G ≈ +150 meV. The retardation effect can be even stronger if the redox chain was reduced in steady state when the electrostatic interactions shift the Em values of these clusters even lower. Anyway, N5 → N6a is the slowest step along the chain, which is most likely the rate-limiting for the eT along the whole FeS chain. For ˚ G = 0) gives 2–10 ns comparison, the fastest step (N4–N5, 8.5 A, ˚ downhill reac( = 0.5–0.7). The reduction of N1a by FMN (12.3 A, tion with G ≈ −100 meV) gives 0.1–0.5 ␮s ( = 0.5–0.7), although ˚ G = 0) time constant is for the reduction of N3 by FMN (7.6 A, almost 100 times faster. The experimentally measured apparent rate of the N1a reduction is almost the same as for N2 (Fig. 7), despite the ∼50 times

9

difference between the intrinsic eT rate constants for FMN → N1a and FMN → · · · → N2 steps. Therefore, the two electrons from FMN are most likely loaded to the FeS chain through N3 and then redistributed according to the redox potentials of the clusters. Since all elementary rate constants of eT between the clusters are much higher than the slowest, experimentally observed reduction of the remaining clusters by NADH (∼1 ms, Fig. 7), the latter is limited by a single process: the dissociation of the product NAD+ is the only possible candidate. The important observation made in this work (Verkhovskaya et al., 2008; Belevich et al., 2009) is that no semiquinone radical signal can be detected at Eh as low as −400 mV. The sensitivity of the technique used allowed to resolve as low population of EPR radical signal as 1–2%; since the purified Complex I used in this study contained one molecule of ubiquinone per protein such low population of semiquinone suggested a very low Em (≤−300 mV) for the couple UQ/UQ•− . Since the latter value is close to Em of the NAD+ /NADH couple it means that transfer of the first electron from NADH to UQ is not associated with any energy drop and that it is the second electron transfer step which is coupled with a large potential drop driving the translocation of all four protons across the membrane (Verkhovskaya et al., 2008). 4.4. Spatial organization of the FeS clusters chain Since the recent studies clearly indicate that no energy is converted upon the electron transfer from NADH to the last FeS cluster the reasonable question arises: why such complicated, long intramolecular redox chain was formed in the evolution. There could be several explanations: (a) Evolutionary aspect. The subunits of the hydrophilic Complex I domain are similar to the two different types of hydrogenases: some subunits are homologous to soluble Fe-containing NAD+ reducing hydrogenases and some to NiFe hydrogenases (Albracht, 1994; Pilkington et al., 1991). These findings gave rise to so called modular evolution hypothesis suggesting that Complex I emerged by combination and merging particular proteins, (Brandt, 2006; Efremov and Sazanov, 2012; Hedderich, 2004; Moparthi and Hägerhäll, 2011; Friedrich and Weiss, 1997). The hydrophilic domain evolved from NAD+ -reducing NiFe hydrogenases, lost NiFe site but preserved all FeS centers to connect nucleotide binding site to the membrane domain. (b) Acceleration of NADH/NAD+ exchange. Evacuation of the NADH binding site to the cytoplasm may have advantage of leaving the unstirred water layer adjacent to membrane layer with limited diffusion. 5. Ubiquinone reduction coupled to the transmembrane proton translocation 5.1. Ubiquinone in Complex I: one or two binding sites? The reduction of ubiquinone (UQ) in Complex I necessarily comprises several elementary steps of electron and proton transfer and possibly conformational change connected with a mechanical transduction system in the enzyme’s membrane domain. According to present knowledge, it is during the reduction of UQ where most redox energy is converted into a form of mechanical work leading to transmembrane proton translocation. This process is however the least studied in the Complex I catalytic mechanism. Despite the availability of several recently obtained X-ray crystal structures of Complex I, the precise location of UQ binding site(s) still remains unknown. None of the structures contains bound UQ and the atomic resolution of neither isolated hydrophilic and hydrophobic

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 10

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Fig. 7. Transient kinetics of FeS cluster reduction upon Complex I reaction with NADH. Fast, ∼100 ␮s, component represents reduction of two clusters, N1a and N2. Slow, ∼1 ms, component represents reduction of other clusters. The intramolecular redox chain is drawn on the basis on the structure of Complex I from T. thermophilus (RDB ID: 2FUG). The red circles indicate the destination of the first two electrons from reduced FMN. Modified from Verkhovskaya et al. (2008) with permission (Copyright (2008) National Academy of Sciences, USA).

domains of Complex I (3.1 A˚ Sazanov and Hinchliffe, 2006 and 3.0 A˚ Efremov and Sazanov, 2011b; Efremov et al., 2010, respectively) nor the whole enzyme (4.5 A˚ Efremov et al., 2010 or 6.3 A˚ Hunte et al., 2010) allowed clear identification of the UQ binding sites. In the latter two structures the interface between the hydrophilic and hydrophobic domains is resolved very poorly and certain parts of the structure, such as some amino acid residues in N-terminus of hydrophilic Nqo4 subunits and the loops of the membrane subunits are completely missing; however the hydrophilic–hydrophobic interface is the most likely location of the UQ binding site(s). Even the arrangement of the two domains raised questions regarding the UQ interaction with its immediate electron donor, the terminal FeS cluster, N2. This cluster resides at the distance of 20–25 A˚ (Efremov et al., 2010) or 30 A˚ (Hunte et al., 2010) from the enzyme’s hydrophobic membrane core (see Fig. 2). According to the electron transfer theory (Page et al., 1999; Moser et al., 2006) the distance of the efficient eT should not exceed 14–15 A˚ to match the experimentally observed rates of Complex I turnover (over 300 s−1 Verkhovskaya et al., 2011). Indeed, a fast-relaxing semiquinone radical species, SQNf , was found at a distance 8–11 A˚ from N2 based on the EPR signal splitting data (Ohnishi et al., 1998; Yano et al., 2005) and at a distance of 10–13 A˚ from N2 based on the microwave power saturation profiles (Ohnishi, 1998) in mitochondrial Complex I bound to the natural closed membrane vesicles (bovine sub-mitochondrial particles). Consequently, to approach its binding site the UQ headgroup should move out from the membrane domain for at least ∼10 A˚ leaving at least one hydrophobic isoprenoid tail segment exposed to the rather hydrophilic protein milieu between the membrane and hydrophilic domains. Such a location of a UQ-binding site is in a striking contrast to that in other quinone-binding enzymes: in most of them, quinone headgroup is bound at the membrane-water interface close to the position where the lipids’ polar groups are located, leaving the tail within the membrane hydrophobic core (see e.g. Lancaster, 1999; Lancaster and Kröger, 2000). Keeping in mind that during the enzyme’s turnover UQ (or UQH2 ) should leave the binding site

and exchange with another UQ (or UQH2 ) molecule, shuttling of UQ between its “membrane-exchangeable” position and the unusually located “eT” position would be highly energetically unfavorable. To overcome this problem Brandt and co-authors suggested that in the yet unresolved interface area, a dielectric “ramp” is formed by the loops of hydrophilic (NuoB and CD) and membrane (NuoA, K, J and H, E. coli numbering) subunits enclosing a local hydrophobic zone that accommodates the UQ tail (Tocilescu et al., 2007, 2010b; Zickermann et al., 2003). However, such hydrophobic zone cannot be clearly predicted from the primary amino acid sequences. Note that a chain of Met residues in the hydrophilic subunit PSST (NuoB, E. coli numbering) involved in the access and binding of the hydrophobic parts of ubiquinone has recently been proposed (Angerer et al., 2012). The other way to elude energetically unflavored UQ movement is to suggest two UQ-binding sites: one, “eT” site in the vicinity of N2 containing a tightly bound UQ, and the other, “exchangeable” site residing at the surface of membrane core and accommodating a second molecule of membrane poolexchangeable UQ. Except for the abovementioned semiquinone SQNf positioned at the distance of ∼10–12 A˚ from N2, another, slow-relaxing, semiquinone species (SQNs ) was detected in bovine Complex I. This radical was separated from N2 by more than 30 A˚ and it was suggested to be distinct from the SQNf (Ohnishi et al., 1998; Yano et al., 2000). The latter statement was argued (Tocilescu et al., 2010b) based on the fact that the sum of the SQNf and SQNs spin concentrations did not exceed one per enzyme (van Belzen et al., 1997); this argument, however, is not solid because the yield of each given semiquinone state can depend on many factors and in practice, never approaches 100%. A so-called reductant-induced oxidation mechanism, proposed by Dutton et al. (1998) suggests three UQ-binding sites: one for the tightly bound UQ and two for the pool-exchangeable quinones at the opposite sides of the membrane. Upon reduction/oxidation the tightly bound ubiquinone alternately shuttles between two proton half-channels accessible from the N and P sides of the membrane

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

11

and therefore translocates protons. Such direct coupling mechanism based on the electrostatic interaction between the reduced quinone and proton was strongly criticized later (see below); however, its modifications considering translocation of only 1 or 2H+ from 4 per turnover were reported recently (Ohnishi et al., 2010a,b). Other, more recent data provide stronger support in favor of two UQ binding sites in Complex I (see below), although formally speaking, for the lack of structural information, the question of the number of quinone binding sites is still open. Note that there are no selective inhibitors for the two UQ binding sites and strictly speaking, for the most of the inhibitors no precise binding site is known: the inhibitors are identified primarily functionally for their block of the UQ reductase activity and competition with UQ (see for recent review Tocilescu et al., 2010b). 5.2. Mapping of the quinone-binding site(s) The extended study aimed to localize the quinone-binding site(s) by insertions of point mutations affecting UQ-reductase activity of Y. lipolytica Complex I and its affinity to UQ and competitive and structurally related inhibitors revealed the functionally important region in the interface between hydrophilic and membrane domains (Kashani-Poor et al., 2001; Ahlers et al., 2000; Tocilescu et al., 2007, 2010b; Zickermann et al., 2008; Angerer et al., 2012). Two subunits of hydrophilic domain, 49 kDa and PSST, one or both of which are in contact with the membrane domain were studied. The presumed quinone binding pocket is rather spacious: functionally critical residues are located in close vicinity ˚ and remotely (such of N2 (such as Tyr144, at a distance of ∼7 A) ˚ (Tocilescu et al., 2010b). Due to unique catalytic as His95, 16 A) properties, i.e. strong difference of Y144F mutant in comparison to wild type in activity with miscellaneous ubiquinones, the conserved Tyr144 was claimed to interact directly to ubiquinone headgroup (Tocilescu et al., 2010a) and the only one Q-binding site was declared (Tocilescu et al., 2010b; Brandt, 2011). Another approach for Q binding site search employed photoaffinity labeling with derivatives of photoreactive quinone and Complex I inhibitors. However, this method has strong requirements to the label: the bulky indicator tag introduced improperly could decrease the affinity and result in an artificial labeling. This was probably the reason why besides the UQ-binding site in the interfacial area several sites were reported previously in the membrane domain (Gong et al., 2003; Nakamaru-Ogiso et al., 2003, 2010a). However, these results are not supported by structural data. Recent studies by Miyoshi’s group with the derivatives of specific Complex I inhibitors, acetogenins, revealed one binding site in the membrane ND1 (corresponding to Nqo8 in Fig. 8) subunit, presumably in the region which covers 4th and 5th or 5th and 6th transmembrane segments (TMS) depending on the acetogenin type (Nakanishi et al., 2011; Sekiguchi et al., 2009; Kakutani et al., 2010). These authors also reported that binding of another Complex I inhibitor, fenperoxymate, occurs in two hydrophilic subunits, 49 kDa and PSST. They were able to identify the particular region in these subunits responsible to the interaction with the label (Shiraishi et al., 2012). Homologous amino acid residues in the resolved structure of T. thermophilus enzyme are shown in Fig. 8. These residues from two subunits cover a vast funnel-shape area, the sharp end of which points to N2; the whole area could hardly be specific for ubiquinone binding but could rather indicate importance of particular structure for catalysis or/and reflect the accessibility of the inhibitor to the enzyme. Other binding sites, in 49 kDa and ND1 subunits were found for the photoreactive derivative of the most structurally similar to ubiquinone inhibitor quinazoline. This label is compact; the photo labile azido group was introduced to the toxophoric quinazoline ring without loss of inhibitory effect what minimized the artificial binding (Murai et al.,

Fig. 8. Localization of the inhibitor-labeled sites in Complex I. The amino acid residues situated within the inhibitor binding sites are indicated by red, blue and cyan colors. Quinazoline-labeled area of Nqo4 (Murai et al., 2011) is shown in red. Fenperoxymate-labeled areas in Nqo4 and Nqo6 (Shiraishi et al., 2012) are shown in blue and cyan, respectively. Second quinazoline binding site is resided in the subunit Nqo8, shown in orange, particularly 5th and 6th TMSs and loop 3 (Murai et al., 2011), exact location of those is not yet resolved. The inhibitor-labeled sites studied in the bovine enzyme (Murai et al., 2009, 2011) are shown on the Complex I structure from T. thermophilus (RDB ID: 3M9S).

2009). The identified region of the binding site in 49 kDa subunit, Asp41–Arg63, (Murai et al., 2009) does not overlap with that for fenperoxymate (homologous amino acid residues in T. thermophilus enzyme are shown in red in Fig. 8). The other site in ND1 subunit was shown to situate in the region of 5th and 6th TMSs which includes long and highly conserved cytoplasmic loop 3 (Murai et al., 2011). The authors suggested that this loop interacts with the Nterminus of 49 kDa; therefore the both subunits could participate in formation of the common binding site. However, the resolved structure indicates that rather the N-terminus of 49 kDa interacts with membrane ND3 subunit (Nqo7 in T. thermophilus (Fig. 7); the latter is supported by Kao et al. (2004b) who proved the proximity of these two subunits by zero-length cross-linking. Noteworthy, the acetogenin binding on membrane ND1 subunit could be strongly prevented by exogenous ubiquinones (Sekiguchi et al., 2009; Kakutani et al., 2010). This is an indication to the site where UQ can be exchangeable with the pool. The quinazoline label binding in 49 kDa and ND1 subunits is insignificantly affected by added ubiquinone although it is strongly reduced by other ubiquinone-like inhibitors (Murai et al., 2009, 2011) what could indicate that this spot contains a site that binds inhibitors but not necessarily the ubiquinone. It is worth performing juxtaposition of the mutagenesis and labeling data. The mutations in Y. lipolytica corresponding proximal (close to N2) end of the quinazoline labeled area, at positions 95–99 (49 kDa subunit, counterpart of Nqo4 in Fig. 7) resulted in strong decrease or complete loss of the quinone reductase activity but insignificant changes in the NADH dehydrogenase activity (Tocilescu et al., 2007). The same effect was observed upon the replacement of residues 458–461 and Tyr87 in this subunit (Tocilescu et al., 2007) located in the vicinity of this proximal end. On the other hand the replacement of two histidines 224 and 228 which reside within the quinazoline-labeled area in E. coli enzyme did not yield in drastic change in activity although some increase in superoxide production was observed (Belevich et al., 2007). No mutations have been reported yet in the fenperoxymatelabeled area in subunits corresponding to Nqo4 (49 kDa) in other organisms. The site responsible to fenperoxymate labeling in Nqo6 (PSST) subunit includes 4 cysteines liganding N2 cluster; therefore, mutations of those obviously resulted in loss of quinone

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

ARTICLE IN PRESS

G Model BC-3836; No. of Pages 21 12

M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

reductase activity (Flemming et al., 2003b). The residues Trp77, Met90 and Met91 in PSST subunit located in this site are reported to be essential for Q-reductase activity of Y. lipolytica Complex I (Angerer et al., 2012). In E. coli the replacement E67Q (NuoB subunit, within the area homologous to fenperoxymate binding site in bovine enzyme) resulted in strong decrease of the quinone reductase activity (Flemming et al., 2006), which can also be explained by the mutation effect on N2 due to close proximity of the residue to this cluster. The two essential acidic residues, E89 and D99 in Y. lipolytica PSST (Ahlers et al., 2000), are situated outside of this area. 5.3. Tightly bound ubiquinone The presence of an equimolar amount of UQ in detergentsolubilized, purified Complex I can itself indicate its tight binding since the loosely bound UQ can be easily washed away by the excess of detergent during the isolation. The data on the UQ content in purified Complex I is contradictory. Isolated Y. lipolytica enzyme was reported to contain substoichiometric amount of UQ, 0.2–0.4 molecules per enzyme (Dröse et al., 2002; Marshall et al., 2006). Recent studies of bovine Complex I by Yoshikawa and co-workers yielded in the ratio exactly 1 UQ/FMN (mol/mol) (Shinzawa-Itoh et al., 2010). In our group the same (Euro et al., 2008a) or slightly higher, 1.3 (Verkhovsky et al., 2012a), values of UQ/FMN ratio were obtained in the purified Complex I from E. coli. However, these results can also be explained by contamination due to hydrophobicity of the enzyme and retained lipids. To exclude the latter we performed studies of kinetic, by means of stopped flow techniques, and thermodynamic, by redox titration, properties of tightly bound ubiquinone. Rapid mixing of Complex I with NADH, traced optically, demonstrated that both reduction and re-oxidation kinetics of ubiquinone coincide with the respective kinetics of the majority of Fe–S clusters, indicating the kinetic competence of the detected ubiquinone. The redox titration showed that the transitions in ubiquinone UV band occur at ambient potentials lower than −200 mV indicating tight binding (Verkhovsky et al., 2012a). Participation of tightly bound ubiquinone in the enzyme turnover necessarily implies existence of loosely bound one, exchangeable with the pool. Whether there are two different binding sites or the only one, which undergoes the changes drastically altering its affinity, remains still unrevealed. Summing up one can safely assert that there is one UQ-binding site in close (eT-competent) proximity to the FeS cluster N2. Most likely the UQ is tightly bound there. However, no clear answer is yet obtained whether there is a second site present in Complex I: the extended studies discussed above can neither rule it out, nor support it. 5.4. Models of the UQ redox transitions So far it remains unclear which particularly quinone state(s) is coupled to the proton transfer. Four models of redox transitions of ubiquinone(s) coupled to the energy conversion in Complex I have been suggested in the literature; they are briefly summarized below. The model suggested by Ohnishi et al. (2010a,b) and Ohnishi (2010) combines direct and indirect coupling steps (Model A, Fig. 9). Direct coupling employs immediate participation of tightly bound UQ species in proton translocation: the first electron arrived from N2 results in formation of semiquinone anion, which attracts two protons from the negative (N) side of the membrane. Arrival of the second electron results in the movement of the bound quinone to the proton channel leading to the positive (P) membrane side. Two protons are released there and the two electrons are simultaneously transferred to the loosely bound UQ. Charging of the

latter induces the conformational change in the membrane subunits resulting in the proton translocation. Thus, Model A postulates the existence of two enzyme-bound UQ molecules and two different roles for the tightly bound and loosely bound UQ: the former acts as a two-electron chemical gate, whereas the latter acts as a spring-loaded latch for a mechanical transmission to the membrane-bound, gated proton channels. Switching of the tightly bound UQ between the two positions, itself quite possible, automatically requires two proton channels connecting the quinone-binding site with the input (N) and output (P) for protons. The existence of the latter, the proton channel spanning the whole membrane dielectric from the hypothetical UQ location to the P-side of membrane, has indeed been recently predicted by Efremov and Sazanov (2011b) based on the structure of the membrane subunits NuoN, J and K of E. coli Complex I. Keeping in mind the likely location of the tightly bound UQ at the N-side membrane interface, the input channel is not supposed to be long and it can be rather hydrophilic. However, it is very difficult to explain twoproton coupling with the 1-electron reduction of UQ: the solution pKa of the product QH2 •+ is rather low (Rich, 1984) and its value for the bound UQ species can hardly differ much from the solution value. Another problem arising from the model is the highly improbable simultaneous transfer of two electrons from the fully reduced tightly bound UQ to the oxidized, exchangeable UQ. A third problem arises from the stoichiometry of the pumped protons: only 2H+ are translocated through the three homologous antiporter-like membrane subunits (gated H+ -channels), which is not supported by structural and mutagenesis data. A modified version of the same Model A was suggested by Treberg and Brand (2011) based on the studies of ROS production by Complex I (Vinogradov and Grivennikova, 2005; Grivennikova and Vinogradov, 2006; Lambert and Brand, 2004a). Besides a well-known ROS production site at the enzyme-bound FMN, the additional superoxide-generating site active upon the reverse electron transfer, was proposed to be at Q-binding site (Lambert and Brand, 2004a,b; Hirst et al., 2008). The increase in the ROS production during the reverse electron transfer in the presence of pH across the membrane (Lambert and Brand, 2004b) let Lambert and Brand conclude that the semiquinone radical is responsible for the superoxide generation (Treberg et al., 2011). The model proposed by Treberg and Brand involves formation of neutral semiquinone (one e− and one H+ are transferred to UQ from N2). Arrival of the second electron to UQ results in one H+ translocated across the membrane. The other three H+ are translocated by membrane subunits via conformational changes driven by the energy released upon Q2− protonation (Treberg and Brand, 2011). In this model, only 1H+ is translocated by a chemical gate mechanism, leaving the other three for the mechanical transmission mechanism. This stoichiometry fits better to the number of membrane antiporter-like subunits; however, in contrast to original Model A the authors do not specify the possible role of the two UQ binding sites. The “two-state stabilization-change” model (Model B, Fig. 9) presented by Brandt (2011) suggests two sequential steps of oneelectron UQ reduction involving the quinone molecule situated in the same site (Fig. 8). Both steps induce conformational changes in the membrane subunits yielding in 2H+ translocation. At the first step the quinone is tightly bound but the arrival of the second electron and partial quinone protonation change the quinone binding properties, it becomes loosely bound and could be released to the pool. As in original Ohnishi’s Model A, the perturbed stoichiometry of two pumped protons per three antiporter-like subunits raises questions. On the basis of the real-time eT data we have earlier suggested that there is only one stroke when the energy of the reduced quinone is transmitted to the proton-pumping membrane subunits

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

13

Fig. 9. Mechanistic models of the coupling between UQ redox transitions and transmembrane H+ translocation in Complex I. Model A: combined direct and indirect coupling (Ohnishi, 2010; Ohnishi et al., 2010a,b). Model B: two-stroke mechanism, assuming two sequential proton translocation steps. A single UQ-binding site with alternating affinity is considered (Brandt, 2011). Model C: one-stroke mechanism, assuming two UQ-binding sites (Verkhovskaya et al., 2008) and (Verkhovsky et al., 2012a). Model D: one-stroke mechanism, assuming one UQ-binding site (Efremov and Sazanov, 2012). Boxes denote spatially distinct redox centers (binding sites); different Q states within each box show the same Q molecule after (but not before) the 1st or 2nd electron transfer. High-energy states of UQ shown in red are proton translocation-competent. Relaxed states of UQ shown in blue are proton translocation-silent. Gray arrows marked by arch lines designate the proton transfer mediated by conformational changes in the membrane subunits. Short gray arrows in Model A designate the proton translocation in half proton channel equilibrated with bulk on N or P side of the membrane.

(Model C, Fig. 9) (see also: Fig. 2 in Wikström and Hummer, 2012). Considering our studies of tightly bound ubiquinone (see above) we proposed that the tightly bound UQ only acts as one-electron carrier. Since the Em of the UQ/UQ•− pair is low and probably close to the value for NADH/NAH+ (which explains low population of semiquinone) semiquinone formation is not associated with any energy drop and it is the transfer of the second electron to the semiquinone that occurs across a large potential drop and therefore drives translocation of all four protons across the membrane (Verkhovskaya et al., 2008; Wikström and Hummer, 2012). However, the stoichiometry question, i.e. how three antiporter-like membrane subunits translocate 4H+ , is still open. Recently Efremov and Sazanov (2012) suggested the model also presuming the one-stroke mechanism but with single Qbinding site (Model D, Fig. 9). Delivery of the second electron to ubiquinone via electrostatic interactions provokes conformational changes in the membrane subunits (see for details Efremov et al., 2010; Efremov and Sazanov, 2011a,b); three antiporter-like subunits translocate 3H+ , the fourth H+ was predicted to be transported via additional proton channel formed between one large, NuoN and two small membrane subunits, NuoJ and K (Efremov and Sazanov, 2011b). Note that the models postulating only one UQ-binding site (Models B and D) require energetically unfavorable, long-distance quinone movement from the membrane pool to load the binding site, whereas the two-binding site models (Models A and C) lack this difficulty. It is important to note also that according to the semiquinone studies in bovine Complex I in native submitochondrial vesicles, the “proximal” to cluster N2 semiquinone radical signal SQNf was ␮ ˜ H+ -sensitive, whereas the “distal” signal, SQNs , was ␮ ˜ H+ -insensitive (Ohnishi et al., 1998; Yano et al., 2000). This may point to a possible role of a tightly bound, rather than loosely bound UQ as a mechanical coupling element, although such possibility has not been yet analyzed in the literature. The important observation that the point mutation(s) of large membrane subunits, even in the distal one, NuoL, could completely block the ubiquinone reduction (see below) remains without explanation.

Thus it is largely accepted now that the energy-coupled step occurs upon the reduction/protonation of the ubiquinone species: its negative charge does the job. The energy, partially or completely, is transmitted to the membrane subunits via a UQ conformational change driving a mechanical transmission to the three membranebound gated proton channels. 6. Long-distance energy transduction coupled to proton translocation 6.1. The mechanical “piston” model of Complex I The three larger membrane subunits of Complex I, NuoL, M and N (eukaryotic subunits ND5, ND4 and ND2, respectively) (Fig. 10; see also Table 2 for subunit numbering), are homologous to certain subunits of multicomponent Cat/H+ antiporters (Hamamoto et al., 1994; Mathiesen and Hägerhäll, 2002). They share a pattern of conserved amino acid residues, the replacement of which inactivates the antiporter (Morino et al., 2010) as well as Complex I (Euro et al., 2008a; Nakamaru-Ogiso et al., 2010b; Torres-Bacete et al., 2007; Amarneh and Vik, 2003). Based on the X-ray crystal structure of the E. coli Complex I membrane domain, presumable proton channels have been proposed in these subunits (Efremov and Sazanov, 2011b). All this makes it certain that the transmembrane proton translocation occurs in Complex I via these three subunits. For many years the most intriguing question about the Complex I catalytic mechanism has remained how energy released upon UQ reduction at the interface between the hydrophilic and hydrophobic domains can be transmitted over the distance of ∼100 A˚ up to the distal NuoL subunit. Indeed, the structure of Complex I reveals a unique feature not found in any other membrane protein: long amphipathic ␣-helix (AH) running along the membrane domain (Efremov et al., 2010; Efremov and Sazanov, 2011b; Hunte et al., 2010). This helix belonging to the subunit NuoL (at its C-terminus), starts from the distal end of the membrane domain at the 15th TMS, spans almost the entire length of the membrane domain and ends with a transmembrane helix, the 16th TMS, located close to the hydrophilic domain (Efremov et al., 2010; Efremov and Sazanov,

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 14

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Fig. 10. Proton-translocating membrane subunits of Complex I. The “piston” model of Complex I function (Efremov et al., 2010). The amphipathic ␣-helix (AH) derived from subunit NuoL anchored with 15th and 16th TMS at its distal and proximal (with regard to the UQ binding site) positions is in a tight contact with the discontinuous TM ␣helices forming H+ channels in all three antiporters-like subunits, NuoL, M and N. Reduction of UQ initiates conformational changes that pushes AH and causes reorganization of the proton channels and synchronous release of 3H+ at the periplasmic side of the membrane. The fourth proton is suggested to be pumped through the proton channel formed by the small membrane subunits, NuoA, J, K and subunit N.

2011b) (Fig. 10A). The AH forms contacts with the two other antiporter-like subunits, NuoM and NuoN (Efremov and Sazanov, 2011b). According to a “piston model” (Fig. 10B), the AH acts as a mechanical transmission element to provide a long-distance (over ˚ energy transfer by a conformational change along the mem100 A) brane domain. Delivery of a pair of electrons from NADH to UQ results in the conformational changes in the proximal membrane subunits pushing AH toward the other membrane subunits. The latter tilts the TM helices in three antiporters-like subunits NuoL, NuoM and NuoN causing reorientation of certain residues in the proton channels and passing protons across the membrane (Efremov et al., 2010; Ohnishi, 2010). Such model based on a mechanical transmission necessarily has two general requirements: (i) the AH should contain certain amino acid residues to form tight contact with “matching” residues located on the channel-forming subunits NuoM and N, and both former and latter residues are expected to be conserved and (ii) the AH should be highly rigid to minimize energy loss and provide the reversibility of the enzyme. Although AH is rich in charged residues in all studied organisms, none of them is fully conserved. In accordance to the structure of membrane domain the most tight contact between AH and NuoM is formed by the amino acid residues 563–570 in AH and residues M Lys173, M Asp246, M Tyr317 and M His241 in NuoM (E. coli enzyme) (Efremov and Sazanov, 2011b). The shortest distance is between M His241 and L Asp563 (2.6 Å); however, none of the replacements M H241 M (Euro et al., 2008a), M H241A (TorresBacete et al., 2007), M H241E (Michel et al., 2011; Efremov and Sazanov, 2011a) did strongly affect the enzyme activity, and only the replacements of His241 for the positively charged residues (Lys and Arg) diminished the activity to 40–45% (Michel et al., 2011). The replacement of a weakly conserved Pro552 in the middle of AH (closest distance to Trp303 in NuoM 8.5 Å) did not significantly decrease the enzyme activity and left the proton pumping stoichiometry (H+ /e− ) unchanged (Belevich et al., 2011). The most tight contact of AH with NuoN subunit is between conserved N His224 and L Glu587 (4.45 Å). The substitution of N His224 to Ala or Tyr did not affect the activity of Complex I (Amarneh and Vik, 2003), to Glu it reduced the activity to 70% and, again, the replacement of His224 for positively charged residues depressed the activity more significantly, to 32–37% (Michel et al., 2011). These results indicate that the contact areas between AH and the antiporter-like

Fig. 11. Location of the insertion and substitution sites predicted to increase AH flexibility in the membrane domain of E. coli Complex I (RDB ID: 3RKO). Upper panel: top view from cytoplasm; lower panel: side view. Insertions at Gly532 and Pro564 and substitutions at Gly532 yielded wild type phenotype. Insertion at Pro552 resulted in significant loss of activity and decrease in H+ pumping efficiency. Substitution at Pro564 and insertion at Pro572 yielded in highly unstable Complex I (Belevich et al., 2011).

membrane subunits, though important for the enzyme activity, are not highly specific and therefore can hardly be involved in the pumping mechanism requiring a very tight coupling. To test the rigidity of AH we generated a series of mutations introducing insertions and substitutions, 6–7 amino acid residues long, at the different positions of AH2 (Fig. 11). The introduced fragments contained ␣-helix-disrupting sequences and were predicted to break helical structure and form disordered loops, thus making AH more flexible, or both more flexible and longer. Such insertions made at two positions (Gly532 and Pro564) did not yield

2 Substitutions or insertions were introduced after the amino acid residues indicated.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

15

Table 3 Replacements of amino acid residues in the membrane domain of E. coli Complex I lowering or abolishing the ubiquinone reductase activity. Subunit

Quinone reductase activity (%)

Ref.

NuoA

E81A D79N

96 99

36 37

Kao et al. (2004a)

NuoJ

V65G V65G

99 82

12 2

Kao et al. (2005a) Pätsi et al. (2008)

NuoK

E36A E72Q

95 99

7 26

Kao et al. (2005b)

NuoH

G134V R148A D213A E220Q

54 49 53 45

3 4 12 1

Sinha et al. (2009)

NuoN

K217C K247C K395C

0 7 5

n.d. n.d. n.d.

Amarneh and Vik (2003)

NuoM

E144A K234A K265R E144Q K234A K265A

100 100 100 98 91 96

E144A K229E K399A R431A

∼90 ∼105 ∼120 ∼60

NuoL

Mutation

NADH dehydrogenase activity (%)

16 13 50 2 16 30 ∼20 ∼15 ∼29 ∼7

Euro et al. (2008a)

Torres-Bacete et al. (2007)

Nakamaru-Ogiso et al. (2010b)

n.d., not determined.

in any significant effect on the enzyme activity or H+ pumping efficiency. The substitution at Pro564 resulted in highly fragile enzyme with decreased activity indicating that the replaced amino acid residues are important for the enzyme integrity. Only the insertion at Pro552 caused a significant decrease in both activity and proton pumping (H+ :e− ) stoichiometry (30% and ∼25% from the wild type level, respectively), although according to the Complex I structure, the mutated area is in contact with a rather non-conserved C-terminal fragment of NuoM, located on the border between NuoL and NuoM. If AH was the mechanical transmission element and the contact was broken by the insertion at Pro552 then the maximum expected effect of the mutation would be inactivation of both NuoL and M resulting in 50% loss in the H+ :e− stoichiometry. On the other hand, the observed phenomenon of apparent decoupling can be explained by proton leakage: Complex I destabilized by the mutation can become leaky upon turnover. (Note that we did not observe any increase in membrane permeability when the subbacterial membrane vesicles were energized by the other proton pump, bo3 oxidase, Euro et al., 2008a.) The number of other mutations in subunit NuoL, i.e. truncations at its N- and C-termini and insertion at the position Pro572 in AH resulted in poorly expressed, highly fragile Complex I (Belevich et al., 2011). Torres-Bacete et al. (2011) also showed that the C-terminus of subunit NuoL (TMS 16th alone or together with AH) is critical for Complex I stability. The above results allowed us to propose that the AH is in fact not a piston providing mechanical transmission to the membrane proton channels, but rather a clamp keeping the three subunits NuoL, M, and N together during the catalytic turnover. To perform mechanical function of a “piston” the direct contact of AH with main body is a necessary, but not a sufficient condition. Shuttle movement requires rigidity, which can be a property only of non-disturbed ␣-helix. Any increase in elasticity (“softening”) of the ␣-helix by insertion of disordered fragments would lead to the energy dissipation and inevitably to the drop in proton pumping stoichiometry and uncoupling between the eT and proton translocation activity in Complex I, which was not supported by abovementioned studies. This allowed us to conclude that AH works as mechanical connection but not transmitting energy, it

only stabilizes the enzyme as a clamp. This hypothesis, however, does not specify any further what is the energy-coupling mechanism, which is still unsolved (see below). 6.2. Can the electron transfer and proton pumping be uncoupled? To understand the molecular mechanism of the enzyme it is important to know whether particular parts of it could perform the partial proton translocation (in other words, whether the eT and proton translocation can be partially uncoupled) or it is only the entire complex which is functionally competent. This is a contradictory and argued issue for the time being. It has been recently reported by Brandt and co-workers (Dröse et al., 2011) that Complex I from Y. lipolytica lacking its two distal membrane subunits ND4 and ND5 (counterparts of NuoM and NuoL in E. coli) is capable of proton pumping at half the stoichiometry of the wild-type enzyme. The same 50% decrease of pumping efficiency was found by Friedrich et al. (Steimle et al., 2011) in E. coli enzyme lacking either the entire subunit NuoL or parts of its C-terminal domain (containing AH or only the last 16th TMS). These findings strongly contradict the results of other groups that the deletion of nuoL gene resulted in the loss of assembled Complex I (Nakamaru-Ogiso et al., 2010b; Belevich et al., 2011; Torres-Bacete et al., 2011). Such conflicting data could be partially attributed to the difference between mitochondrial and bacterial enzyme. However, the mutagenesis studies strongly indicated that at least the bacterial enzyme functions as one whole. The reported mutations in subunit NuoH led to a decrease in both quinone reductase and NADH dehydrogenase activity, which can be explained by the importance of this subunit as a mechanical link between hydrophilic and membrane domains: the mutations in subunit NuoH could make the enzyme unstable or/and not fully assembled (Sinha et al., 2009). The replacement of intramembrane Val65 in NuoJ (Kao et al., 2005a; Pätsi et al., 2008) and acidic residues in NuoA (Kao et al., 2004a) and NuoK (Kao et al., 2005b) significantly decrease quinone reductase activity without much effect on NADH dehydrogenase activity, which can be explained by the indirect effect on UQ binding site although it is not supposed to locate in close proximity (see e.g.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 16

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Fig. 12. Rendering of protonatable (chargeable) residues in the membrane domain of E. coli Complex I (RDB ID: 3RKO). Positions of Asp and Glu (red), Lys and Arg (blue), and His (purple) residues in the middle of the hydrophobic domain are shown in the side view (upper plate) and top view from the cytoplasm (lower plate).

Efremov and Sazanov, 2012). However, in the remote membrane subunits NuoL (Nakamaru-Ogiso et al., 2010b) and NuoM (Euro et al., 2008a; Torres-Bacete et al., 2007) the particular mutations of amino acid residues located within the TMSs, which significantly or completely block the reduction of ubiquinone, but leaving NADH oxidation activity almost unchanged, can be generated (Table 3). As it follows from Table 3, even the replacement of a single remote residue in the membrane part of the distal subunit NuoL can almost completely prevent the UQ reduction in the interface between the hydrophilic and membrane Complex I domains located ∼100 A˚ away from the mutation. Despite that the latter fact might seem surprising from a mechanistic point of view, it is not at all surprising from the point of view of the process thermodynamics: Complex I is a fully reversible enzyme capa˜ H+ ble of functioning not far from equilibrium (under the full ␮ load). Thus to minimize energy dissipation there should be a tight coupling between the spatially remote sites of the two catalytic half-reactions: redox transition of UQ and proton translocation. In other words, each proton translocating element should provide a feedback to the ubiquinone binding site in order to allow further turnover after accomplished H+ transfer. Solving a mechanism of such feedback, i.e. how remote proton channel operation can control the ubiquinone reduction, is a prerequisite for any model of the Complex I molecular mechanism. 6.3. Rendering of protonatable residues in the Complex I membrane domain Analysis of the E. coli Complex I membrane domain structure (RDB ID: 3RKO) shows that the long helix AH is not the only structure located parallel to the membrane plane and spanning the whole membrane domain. Surprisingly a number of protonatable amino acid residues are found in the hydrophobic core of all membrane subunits (except NuoH whose structure is not sufficiently resolved) forming a relatively continuous chain, with the neighboring residues 3.5–13 A˚ apart from each other, running all along the Complex I membrane domain (Fig. 12). The location of charged (or chargeable) amino acid residues within the low dielectric area in membrane proteins, when occurs, serves either for keeping structural integrity (Sahin-Toth et al., 1992; Kumari et al., 2006; Liu and Matherly, 2001) or assembly (Call et al., 2010) (by formation

Fig. 13. Rendering of protonatable (chargeable) residues in the membrane domain of E. coli Complex I: membrane subunits are shown separately. All protonatable residues comprising the midmembrane chain are shown in licorice, the other protonatable residues are shown in red (acidic), blue (basic) and purple (histidine). Left panels: side view, right panels: view along the long axes of the membrane domain (RDB ID: 3RKO).

of ionic pairs), or as parts of the transmembrane charge translocation mechanism. A classical example of the latter is the key midmembrane glutamate residue in subunit c of FO F1 -ATP synthase (Junge, 2004); another example is a lysine residue in organic anion-transporting polypeptide 1B3 (Mandery et al., 2011). In case of Complex I two pairs of neighboring residues with the opposite charges and within the distance of 7–8 Å (and one unpaired Glu) are indeed found at the interface between the neighboring subunits NuoL, M, N, and K (L R175/M E407, M D144/N K395, and N E133). Two acidic residues E401 and D400 are located in NuoL close to its interface with lipid phase of the membrane and five unpaired acidic residues reside in small membrane subunits NuoA (E102, E81 and D79) and NuoK (E36 and K72). (Note that all other residues

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

17

Fig. 14. Electrostatically driven “wave-spring” model of Complex I catalytic cycle. For simplicity the fast and energy non-coupled transformation step of electron transfer from NADH to ubiquinone via intramolecular FeS chain is omitted from the figure; only events in the membrane domain are considered. (A) Oxidized state; the lysines in NuoN and L are deprotonated and connected with proton half-channel open to periplasm; the lysine in NuoM is protonated. (B) Two electrons arrive from N2 to ubiquinone. (C) Negatively charged ubiquinol species moves to inaccessible for H+ position and repulses neighbor acidic residues. It results in pKa shift of lysine in NuoN and attraction of H+ from the H+ half channel open to cytoplasmic side. The wave of electrostatic interactions spreads further causing deprotonation of NuoM lysine via H+ half-channel open to the periplasmic side and protonation of lysine in NuoL from cytoplasm. One H+ could be released through the additional H+ channel between NuoN, J and K (Efremov and Sazanov, 2011a,b). (D) Full polarization and re-orientation of the chain of protonatable amino acid residues allow neutralization of negatively charged semiquinone/ubiquinol species and hydroquinole can be released to the pool leaving the ubiquinone binding site yet not accessible. Half of energy is still saved in stressed position in the chain of protonatable residues. (E) Relaxation of the chain yields in reverse wave of electrostatic interaction and release of the H+ from NuoL and N and protonation of the lysine in NuoM, thus the cycle is complete with translocation 3 or 4H+ across the membrane. Thin arrows indicate proton translocation and quinone binding/dissociation events that precede each state shown. The chain of protonatable residues is designated by green symbols.

in the chain are positively charged/chargeable (Arg, Lys and His) (Fig. 12) and no other protonatable residues could be found within the membrane dielectric (Fig. 13). Importantly, all residues in the chain are conserved except Glu102 in subunit NuoA. The chain includes all residues the replacement of which results in loss of ubiquinone reductase activity in mutants (Table 3). Some of these residues, particularly lysines, have been suggested to directly participate in proton translocation on the basis of mutagenesis (Euro et al., 2008a) and structural (Efremov and Sazanov, 2011b) studies. Despite the large residue-to-residue distances, the inter-subunit pairs of oppositely charged residues may in fact play a stabilization role: it can be seen from the structures of the membrane domain of Complex I and the whole complex (PDB entries: 3RKO and 3M9S, respectively) that the position of the subunits NuoL, M, N, and K allows subtle reorientation with respect to each other, which can be well enough to allow these residues to form ion pairs providing tight contact between the subunits. Such minor structural changes were proposed to play a role in the energy transduction mechanism (Efremov and Sazanov, 2011b), but they can also play a structural role. The other positively charged members of the chain are located

in a very hydrophobic interior well isolated from both bulk water and the polar surface of the enzyme. This is a unique structure and it can be a functional element of redox-proton coupling mechanism in Complex I (see below). 6.4. Electrostatically driven “wave-spring” model of Complex I Based on the studies of the conserved midmembrane lysines in subunits NuoL, M, and N, we have proposed the mechanism of the long-range coupling between electron transfer and proton pumping in Complex I (Fig. 14). The mechanism implies spreading the cascade of electrostatic interactions along the membrane domain of Complex I with the result of alternating protonation and deprotonation of these lysines switched between the proton input and proton output conformations (Euro et al., 2008a). The coupling between a redox reaction and proton pumping has been extensively studied in cytochrome c oxidase (Krab and Wikström, 1987; Kim et al., 2009). From the basic principles, the proton-translocating element of a proton pump should necessarily possess two properties: (i) eT-linked pKa shift of the coupled

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

18

protonation/deprotonation reaction; (ii) a gate, i.e. switching between proton input and proton output conformations (Wikström and Hummer, 2012). In Complex I, the source of the driving force for proton pumping (redox reaction) and the proton translocation itself are spatially remote; therefore a transmission between these two sites is also required. In the present model, the gate elements directly involved in proton translocation are the conserved residues Lys229, Lys234, and Lys217 in subunits NuoL, M and N, respectively (E. coli numbering), whereas the transmission element is the chain of the positively charged (chargeable) residues located in the hydrophobic domain core. Changes in the electric field in the membrane domain caused by the displacement of charged UQ species (either UQH− or UQ2− ) are transmitted by the reorientation of the positively charged “chain” residues. The long-range propagation of electrostatic interactions in proteins is well described in the literature (so-called “electrostatic domino”; see e.g. (Maróti et al., 1995; Sebban et al., 1995); the efficiency of the latter, though, depends both on distance and dielectric properties of the medium. For a hydrophobic protein domain, dielectric constant (ε) is typically ranging from 4 to 10; a simple Coulombic interaction (Eq. (1)):

  pKa  ≈

244 r(Å ) · ε

(1)

predicts that for ε = 5, introduction of a charged group will shift pKa value of another charged (protonatable) group for e.g. 3.3 pH units ˚ Thus, electrostatic at a distance r = 15 A˚ and 1 pH unit, at r = 50 A. interactions can propagate at quite large distance when ε is low. We further postulate that in oxidized Complex I the lysines in NuoL and N are deprotonated and the lysine in NuoM, protonated (Fig. 14A). The energy conversion cycle starts from the arrival of two electrons from the FeS cluster N2 to UQ and possibly attracting one proton from the cytoplasm side (Fig. 14B). Upon next step UQ is stabilized in the state unavailable for protons, trapped in the anionic form, either UQH− or UQ2− . Making the quinone isolated from the proton channel(s) most probably requires the displacement of the former; we propose that it is the displacement of the charged species which causes re-orientation of the neighboring charged residues resulting in a pKa shift of channelforming lysine in NuoN and its protonation from the cytoplasmic side. The same displacement can also cause the pumping of the “additional” proton not connected with the antiporter-like subunits to make the H+ :e− stoichiometry of 4. The protonation of the NuoN-subunit lysine results in the repulsion of neighboring chain residues and release the proton from pre-loaded lysine in NuoM. This again provokes the re-orientation of following residues in the chain and a pKa shift of the lysine in subunit NuoL, which attract H+ from the cytoplasmic side (Fig. 14C). When the entire chain is polarized the anionic ubiquinone can move back to the input channel to attract 1 or 2 protons from cytoplasmic site. However, due to the re-orientation of the chain residues QH2 cannot diffuse to the membrane quinone pool (or it can diffuse but a ubiquinone from the pool cannot occupy the site). At this state half of energy is saved in the stressed polarized chain (Fig. 14D). Upon its relaxation the “wave” of electrostatic interactions turns back and results in loading H+ from cytoplasm to the lysine in NuoM, release H+ to periplasm from the lysines in NuoN and in NuoL (Fig. 14E). When the relaxation is finished the ubiquinone from the pool occupies its binding site and the cycle with translocation of 3 or 4H+ across the membrane is complete (Fig. 14A). Note that the proposed out-of-phase protonation behavior of the membrane subunit lysines is not an absolute prerequisite of the model. It is obvious that charging of three groups simultaneously in a hydrophobic medium requires additional energy cost and leads to the energy losses, although the actual distribution of charged

and uncharged lysines may depend on the dielectric properties and the electric field distribution within the Complex I membrane domain. We should also assume that the transmission between the redox and proton translocation parts of the Complex I pump is not purely electrostatic, but is a combination of the electrostatic and mechanical elements. Apparently, a conformation change is necessary within each antiporter-like subunit to drive the gating mechanism allowing transient switching of a lysine residue between the input and output proton channels, as it has recently been proposed by Wikström and Hummer (2012). The fact that the membrane domain of Complex I seems slightly distorted in the outof-plane direction with respect to lipid bilayer and also twisted in the in-plane direction (see e.g. Fig. S4 in Efremov and Sazanov, 2011b) may provide a hint that in the crystal structure, the enzyme is trapped in a particular, non-relaxed state. We can envisage that the force responsible for such distortion actually drives the proton channel gates in the situation when (a) out-of-plane distortion of Complex I is restrained by the lipid bilayer and/or (b) in-plane distortion in restrained by the AH helix. This force is transmitted along the Complex I membrane domain, at least partially, as a “wave” of electrostatic interactions caused by the charged quinone species displacement directly coupled to one of its redox transitions. The latter mechanism does not in fact require a piston-like, shuttling mechanics and does not require any massive structural change, but rather a cascade of subtle conformational changes in the membrane subunits coupled to the electrostatic field propagation along the membrane domain.

Acknowledgments This work was supported by grants from Biocentrum Helsinki, the Sigrid Jusélius Foundation, and the Academy of Finland. Many fruitful discussions with Prof. MÅrten Wikström, critical comments and his support of the work are greatly acknowledged. We thank Dr. Alexander Bogachev and Galina Belevich for critical reading of the manuscript and helpful comments.

References Ahlers PM, Zwicker K, Kerscher S, Brandt U. Function of conserved acidic residues in the PSST homologue of complex I (NADH:ubiquinone oxidoreductase) from Yarrowia lipolytica. J Biol Chem 2000;275:23577–82. Alberty RA. Calculation of standard transformed formation properties of biochemical reactants and standard apparent reduction potentials of half reactions. Arch Biochem Biophys 1998;358:25–39. Albracht SP. Nickel hydrogenases: in search of the active site. Biochim Biophys Acta 1994;1188:167–204. Albracht SP, Dooijewaard G, Leeuwerik FJ, Swol BV. EPR signals of NADH:Q oxidoreductase. Shape and intensity. Biochim Biophys Acta 1977;459:300–17. Amarneh B, Vik SB. Mutagenesis of subunit N of the Escherichia coli complex I. Identification of the initiation codon and the sensitivity of mutants to decylubiquinone. Biochemistry 2003;42:4800–8. Angerer H, Nasiri HR, Niedergesäß V, Kerscher S, Schwalbe H, Brandt U. Tracing the tail of ubiquinone in mitochondrial complex I. Biochim Biophys Acta 2012;1817:1776–84. Basran J, Harris RJ, Sutcliffe MJ, Scrutton NS. H-tunneling in the multiple H-transfers of the catalytic cycle of morphinone reductase and in the reductive halfreaction of the homologous pentaerythritol tetranitrate reductase. J Biol Chem 2003;278:43973–82. Beinert H, Palmer G, Cremona T, Singer TP. Correlation of enzymatic activity and the appearance of the EPR signal at g = 1.94 in NADH dehydrogenase and its thermal breakdown products. Biochem Biophys Res Commun 1963;12:432–8. Belevich G, Euro L, Wikström M, Verkhovskaya M. Role of the conserved arginine 274 and histidine 224 and 228 residues in the NuoCD subunit of complex I from Escherichia coli. Biochemistry 2007;46:526–33. Belevich G, Knuuti J, Verkhovsky MI, Wikström M, Verkhovskaya M. Probing the mechanistic role of the long alpha-helix in subunit L of respiratory complex I from Escherichia coli by site-directed mutagenesis. Mol Microbiol 2011;82:1086–95. Belevich I, Verkhovsky MI. Molecular mechanism of proton translocation by cytochrome c oxidase. Antioxid Redox Signal 2008;10:1–29.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Belevich NP, Verkhovskaya ML, Verkhovsky MI. Chapter 4. Electron transfer in respiratory complexes resolved by an ultra-fast freeze-quench approach. Methods Enzymol 2009;456:75–93. Berrisford JM, Sazanov LA. Structural basis for the mechanism of respiratory complex I. J Biol Chem 2009;284:29773–83. Bogachev AV, Murtazina RA, Skulachev VP. H+ /e− stoichiometry for NADH dehydrogenase I and dimethyl sulfoxide reductase in anaerobically grown Escherichia coli cells. J Bacteriol 1996;178:6233–7. Brandt U. Energy converting NADH:quinone oxidoreductase (complex I). Annu Rev Biochem 2006;75:69–92. Brandt U. A two-state stabilization-change mechanism for proton-pumping complex I. Biochim Biophys Acta 2011;1807:1364–9. Bridges HR, Bill E, Hirst J. Mössbauer spectroscopy on respiratory complex I: the iron–sulfur cluster ensemble in the NADH-reduced enzyme is partially oxidized. Biochemistry 2012;51:149–58. Bright HJ, Appleby M. The pH dependence of the individual steps in the glucose oxidase reaction. J Biol Chem 1969;244:3625–34. Call ME, Wucherpfennig KW, Chou JJ. The structural basis for intramembrane assembly of an activating immunoreceptor complex. Nat Immunol 2010;11: 1023–9. Carroll J, Shannon RJ, Fearnley IM, Walker JE, Hirst J. Definition of the nuclear encoded protein composition of bovine heart mitochondrial complex I. Identification of two new subunits. J Biol Chem 2002;277:50311–7. Chamorovsky SK, Cherepanov DA, Chamorovsky CS, Semenov AY. Correlation of electron transfer rate in photosynthetic reaction centers with intraprotein dielectric properties. Biochim Biophys Acta 2007;1767:441–8. Couch VA, Medvedev ES, Stuchebrukhov AA. Electrostatics of the FeS clusters in respiratory complex I. Biochim Biophys Acta 2009;1787:1266–71. Crofts AR, Holland JT, Victoria D, Kolling DR, Dikanov SA, Gilbreth R, et al. The Q-cycle reviewed: How well does a monomeric mechanism of the bc1 complex account for the function of a dimeric complex. Biochim Biophys Acta 2008;1777:1001–19. DiMauro S. Lessons from mitochondrial DNA mutations. Semin Cell Dev Biol 2001;12:397–405. Djafarzadeh R, Kerscher S, Zwicker K, Radermacher M, Lindahl M, Schägger H, et al. Biophysical and structural characterization of proton-translocating NADHdehydrogenase (complex I) from the strictly aerobic yeast Yarrowia lipolytica. Biochim Biophys Acta 2000;1459:230–8. Dröse S, Krack S, Sokolova L, Zwicker K, Barth HD, Morgner N, et al. Functional dissection of the proton pumping modules of mitochondrial complex I. PLoS Biol 2011;9:e1001128. Dröse S, Zwicker K, Brandt U. Full recovery of the NADH:ubiquinone activity of complex I (NADH:ubiquinone oxidoreductase) from Yarrowia lipolytica by the addition of phospholipids. Biochim Biophys Acta 2002;1556:65–72. Dutton PL, Moser CC, Sled VD, Daldal F, Ohnishi T. A reductant-induced oxidation mechanism for Complex I. Biochim Biophys Acta 1998;1364:245–57. Efremov RG, Baradaran R, Sazanov LA. The architecture of respiratory complex I. Nature 2010;465:441–5. Efremov RG, Sazanov LA. Respiratory complex I: ‘steam engine’ of the cell. Curr Opin Struct Biol 2011a;21:532–40. Efremov RG, Sazanov LA. Structure of the membrane domain of respiratory complex I. Nature 2011b;476:414–20. Efremov RG, Sazanov LA. The coupling mechanism of respiratory complex I – a structural and evolutionary perspective. Biochim Biophys Acta 2012. Euro L, Belevich G, Bloch DA, Verkhovsky MI, Wikström M, Verkhovskaya M. The role of the invariant glutamate 95 in the catalytic site of complex I from Escherichia coli. Biochim Biophys Acta 2009a;1787:68–73. Euro L, Belevich G, Verkhovsky MI, Wikström M, Verkhovskaya M. Conserved lysine residues of the membrane subunit NuoM are involved in energy conversion by the proton-pumping NADH:ubiquinone oxidoreductase (complex I). Biochim Biophys Acta 2008a;1777:1166–72. Euro L, Belevich G, Wikström M, Verkhovskaya M. High affinity cation-binding sites in complex I from Escherichia coli. Biochim Biophys Acta 2009b;1787:1024–8. Euro L, Bloch DA, Wikström M, Verkhovsky MI, Verkhovskaya M. Electrostatic interactions between FeS clusters in NADH:ubiquinone oxidoreductase (complex I) from Escherichia coli. Biochemistry 2008b;47:3185–93. Fernandes AS, Sousa FL, Teixeira M, Pereira MM. Electron paramagnetic resonance studies of the iron–sulfur centers from complex I of Rhodothermus marinus. Biochemistry 2006;45:1002–8. Flemming D, Hellwig P, Friedrich T. Involvement of tyrosines 114 and 139 of subunit NuoB in the proton pathway around cluster N2 in Escherichia coli NADH:ubiquinone oxidoreductase. J Biol Chem 2003a;278:3055–62. Flemming D, Hellwig P, Lepper S, Kloer DP, Friedrich T. Catalytic importance of acidic amino acids on subunit NuoB of the Escherichia coli NADH:ubiquinone oxidoreductase (complex I). J Biol Chem 2006;281:24781–9. Flemming D, Schlitt A, Spehr V, Bischof T, Friedrich T. Iron–sulfur cluster N2 of the Escherichia coli NADH:ubiquinone oxidoreductase (complex I) is located on subunit NuoB. J Biol Chem 2003b;278:47602–9. Friedrich T, Weiss H. Modular evolution of the respiratory NADH:ubiquinone oxidoreductase and the origin of its modules. J Theor Biol 1997;187:529–40. Galkin A, Dröse S, Brandt U. The proton pumping stoichiometry of purified mitochondrial complex I reconstituted into proteoliposomes. Biochim Biophys Acta 2006;1757:1575–81. Galkin AS, Grivennikova VG, Vinogradov AD. (H+ /2e− stoichiometry in NADH–quinone reductase reactions catalyzed by bovine heart submitochondrial particles. FEBS Lett 1999;451:157–61.

19

Gong X, Xie T, Yu L, Hesterberg M, Scheide D, Friedrich T, et al. The ubiquinonebinding site in NADH:ubiquinone oxidoreductase from Escherichia coli. J Biol Chem 2003;278:25731–7. Grivennikova VG, Kotlyar AB, Karliner JS, Cecchini G, Vinogradov AD. Redoxdependent change of nucleotide affinity to the active site of the mammalian complex I. Biochemistry 2007;46:10971–8. Grivennikova VG, Roth R, Zakharova NV, Hägerhäll C, Vinogradov AD. The mitochondrial and prokaryotic proton-translocating NADH:ubiquinone oxidoreductases: similarities and dissimilarities of the quinone-junction sites. Biochim Biophys Acta 2003;1607:79–90. Grivennikova VG, Vinogradov AD. Generation of superoxide by the mitochondrial complex I. Biochim Biophys Acta 2006;1757:553–61. Hamamoto T, Hashimoto M, Hino M, Kitada M, Seto Y, Kudo T, et al. Characterization of a gene responsible for the Na+ /H+ antiporter system of alkalophilic Bacillus species strain C-125. Mol Microbiol 1994;14:939–46. Hatefi Y, Haavik AG, Griffiths DE. Studies on the electron transfer system. XL. Preparation and properties of mitochondrial DPNH-coenzyme Q reductase. J Biol Chem 1962;237:1676–80. Hedderich R. Energy-converting [NiFe] hydrogenases from archaea and extremophiles: ancestors of complex I. J Bioenerg Biomembr 2004;36:65–75. Hill TL. Cooperativity theory in biochemistry. Steady-state and equilibrium systems. NY: Springer-Verlag; 1985. Hinkle PC. P/O ratios of mitochondrial oxidative phosphorylation. Biochim Biophys Acta 2005;1706:1–11. Hinkle PC, Kumar MA, Resetar A, Harris DL. Mechanistic stoichiometry of mitochondrial oxidative phosphorylation. Biochemistry 1991;30:3576–82. Hirst J. Energy transduction by respiratory complex I – an evaluation of current knowledge. Biochem Soc Trans 2005;33:525–9. Hirst J. Towards the molecular mechanism of respiratory complex I. Biochem J 2010;425:327–39. Hirst J, King MS, Pryde KR. The production of reactive oxygen species by complex I. Biochem Soc Trans 2008;36:976–80. Hunte C, Zickermann V, Brandt U. Functional modules and structural basis of conformational coupling in mitochondrial complex I. Science 2010;329:448–51. Ingledew WJ, Ohnishi T. An analysis of some thermodynamic properties of iron–sulphur centres in site I of mitochondria. Biochem J 1980;186:111–7. Jastroch M, Divakaruni AS, Mookerjee S, Treberg JR, Brand MD. Mitochondrial proton and electron leaks. Essays Biochem 2010;47:53–67. Junge W. Protons, proteins and ATP. Photosynth Res 2004;80:197–221. Kaila VRI, Verkhovsky MI, Wikström M. Proton-coupled electron transfer in cytochrome oxidase. Chem Rev 2010;110:7062–81. Kakutani N, Murai M, Sakiyama N, Miyoshi H. Exploring the binding site of delta(lac)acetogenin in bovine heart mitochondrial NADH–ubiquinone oxidoreductase. Biochemistry 2010;49:4794–803. Kao MC, Di Bernardo S, Nakamaru-Ogiso E, Miyoshi H, Matsuno-Yagi A, Yagi T. Characterization of the membrane domain subunit NuoJ (ND6) of the NADH–quinone oxidoreductase from Escherichia coli by chromosomal DNA manipulation. Biochemistry 2005a;44:3562–71. Kao MC, Di Bernardo S, Perego M, Nakamaru-Ogiso E, Matsuno-Yagi A, Yagi T. Functional roles of four conserved charged residues in the membrane domain subunit NuoA of the proton-translocating NADH–quinone oxidoreductase from Escherichia coli. J Biol Chem 2004a;279:32360–6. Kao MC, Matsuno-Yagi A, Yagi T. Subunit proximity in the H+ -translocating NADH–quinone oxidoreductase probed by zero-length cross-linking. Biochemistry 2004b;43:3750–5. Kao MC, Nakamaru-Ogiso E, Matsuno-Yagi A, Yagi T. Characterization of the membrane domain subunit NuoK (ND4L) of the NADH–quinone oxidoreductase from Escherichia coli. Biochemistry 2005b;44:9545–54. Kashani-Poor N, Zwicker K, Kerscher S, Brandt U. A central functional role for the 49-kDa subunit within the catalytic core of mitochondrial complex I. J Biol Chem 2001;276:24082–7. Kerscher S, Drose S, Zickermann V, Brandt U. The three families of respiratory NADH dehydrogenases. Results Probl Cell Differ 2008;45:185–222. Kim YC, Wikström M, Hummer G. Kinetic gating of the proton pump in cytochrome c oxidase. Proc Natl Acad Sci U S A 2009;106:13707–12. Kohen A, Klinman JP. Hydrogen tunneling in biology. Chem Biol 1999;6:R191–8. Kotlyar AB, Borovok N. NADH oxidation and NAD+ reduction catalysed by tightly coupled inside-out vesicles from Paracoccus denitrificans. Eur J Biochem 2002;269:4020–4. Kotlyar AB, Karliner JS, Cecchini G. A novel strong competitive inhibitor of complex I. FEBS Lett 2005;579:4861–6. Krab K, Wikström M. Principles of coupling between electron transfer and proton translocation with special reference to proton translocation mechanisms in cytochrome-oxidase. Biochim Biophys Acta 1987;895:25–39. Krishnamoorthy G, Hinkle PC. Studies on the electron transfer pathway, topography of iron–sulfur centers, and site of coupling in NADH–Q oxidoreductase. J Biol Chem 1988;263:17566–75. Kumari K, Baggenstoss BA, Parker AL, Weigel PH. Mutation of two intramembrane polar residues conserved within the hyaluronan synthase family alters hyaluronan product size. J Biol Chem 2006;281:11755–60. Lambert AJ, Brand MD. Inhibitors of the quinone-binding site allow rapid superoxide production from mitochondrial NADH:ubiquinone oxidoreductase (complex I). J Biol Chem 2004a;279:39414–20. Lambert AJ, Brand MD. Superoxide production by NADH:ubiquinone oxidoreductase (complex I) depends on the pH gradient across the mitochondrial inner membrane. Biochem J 2004b;382:511–7.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21 20

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

Lancaster CRD. Quinone-binding sites in membrane proteins: what can we learn from the Rhodopseudomonas viridis reaction centre. Biochem Soc Trans 1999;27:591–6. Lancaster CRD, Kröger A. Succinate:quinone oxidoreductases: new insights from X-ray crystal structures. Biochim Biophys Acta 2000;1459:422–31. Lau WCY, Rubinstein JL. Subnanometre-resolution structure of the intact Thermus thermophilus H+ -driven ATP synthase. Nature 2012;481:214–9. Leif H, Sled VD, Ohnishi T, Weiss H, Friedrich T. Isolation and characterization of the proton translocating NADH ubiquinone oxidoreductase from Escherichia coli. Eur J Biochem 1995;230:538–48. Leskovac V, Trivic S, Wohlfahrt G, Kandrac J, Pericin D. Glucose oxidase from Aspergillus niger: the mechanism of action with molecular oxygen, quinones, and one-electron acceptors. Int J Biochem Cell Biol 2005;37:731–50. Liu XY, Matherly LH. Functional interactions between arginine-133 and aspartate88 in the human reduced folate carrier: evidence for a charge-pair association. Biochem J 2001;358:511–6. Mandery K, Sticht H, Bujok K, Schmidt I, Fahrmayr C, Balk B, et al. Functional and structural relevance of conserved positively charged lysine residues in organic anion transporting polypeptide 1B3. Mol Pharmacol 2011;80:400–6. Maróti P, Hanson DK, Schiffer M, Sebban P. Long-range electrostatic interaction in the bacterial photosynthetic reaction centre. Nat Struct Biol 1995;2:1057–9. Marshall D, Fisher N, Grigic L, Zickermann V, Brandt U, Shannon RJ, et al. ATRFTIR redox difference spectroscopy of Yarrowia lipolytica and bovine complex I. Biochemistry 2006;45:5458–67. Mathiesen C, Hägerhäll C. Transmembrane topology of the NuoL, M and N subunits of NADH:quinone oxidoreductase and their homologues among membrane-bound hydrogenases and bona fide antiporters. Biochim Biophys Acta 2002;1556:121–32. Mayhew SG. The effects of pH and semiquinone formation on the oxidation–reduction potentials of flavin mononucleotide. A reappraisal. Eur J Biochem 1999;265:698–702. Medvedev ES, Couch VA, Stuchebrukhov AA. Determination of the intrinsic redox potentials of FeS centers of respiratory complex I from experimental titration curves. Biochim Biophys Acta 2010;1797:1665–71. Meijers R, Cedergren-Zeppezauer E. A variety of electrostatic interactions and adducts can activate NAD(P) cofactors for hydride transfer. Chem Biol Interact 2009;178:24–8. Meinhardt SW, Kula T, Yagi T, Lillich T, Ohnishi T. Electron paramagnetic resonance characterization of the iron–sulfur clusters in the NADH–ubiquinone oxidoreductase segment of the respiratory chain in Paracoccus denitrificans. J Biol Chem 1987a;262:9147–53. Meinhardt SW, Kula T, Yagi T, Lillich T, Ohnishi T. EPR characterization of the iron–sulfur clusters in the NADH:ubiquinone oxidoreductase segment of the respiratory chain in Paracoccus denitrificans. J Biol Chem 1987b;262:9147–53. Michel J, DeLeon-Rangel J, Zhu S, Van Ree K, Vik SB. Mutagenesis of the L, M, and N subunits of complex I from Escherichia coli indicates a common role in function. PLoS One 2011;6:e17420. Moparthi VK, Hägerhäll C. The evolution of respiratory chain complex I from a smaller last common ancestor consisting of 11 protein subunits. J Mol Evol 2011;72:484–97. Morino M, Natsui S, Ono T, Swartz TH, Krulwich TA, Ito M. Single site mutations in the hetero-oligomeric Mrp antiporter from alkaliphilic Bacillus pseudofirmus OF4 that affect Na+ /H+ antiport activity, sodium exclusion, individual Mrp protein levels, or Mrp complex formation. J Biol Chem 2010;285:30942–50. Moser CC, Farid TA, Chobot SE, Dutton PL. Electron tunneling chains of mitochondria. Biochim Biophys Acta 2006;1757:1096–109. Murai M, Mashimo Y, Hirst J, Miyoshi H. Exploring interactions between the 49 kDa and ND1 subunits in mitochondrial NADH–ubiquinone oxidoreductase (complex I) by photoaffinity labeling. Biochemistry 2011;50:6901–8. Murai M, Sekiguchi K, Nishioka T, Miyoshi H. Characterization of the inhibitor binding site in mitochondrial NADH–ubiquinone oxidoreductase by photoaffinity labeling using a quinazoline-type inhibitor. Biochemistry 2009;48:688–98. Muraoka S, Slater EC. The redox states of respiratory-chain components in rat-liver mitochondria. I. Effect of varying substrate concentration and of azide. Biochim Biophys Acta 1969;180:221–6. Murphy MP. How mitochondria produce reactive oxygen species. Biochem J 2009;417:1–13. Nakamaru-Ogiso E, Han H, Matsuno-Yagi A, Keinan E, Sinha SC, Yagi T, et al. The ND2 subunit is labeled by a photoaffinity analogue of asimicin, a potent complex I inhibitor. FEBS Lett 2010a;584:883–8. Nakamaru-Ogiso E, Kao MC, Chen H, Sinha SC, Yagi T, Ohnishi T. The membrane subunit NuoL(ND5) is involved in the indirect proton pumping mechanism of Escherichia coli complex I. J Biol Chem 2010b;285:39070–8. Nakamaru-Ogiso E, Sakamoto K, Matsuno-Yagi A, Miyoshi H, Yagi T. The ND5 subunit was labeled by a photoaffinity analogue of fenpyroximate in bovine mitochondrial complex I. Biochemistry 2003;42:746–54. Nakanishi S, Abe M, Yamamoto S, Murai M, Miyoshi H. Bis-THF motif of acetogenin binds to the third matrix-side loop of ND1 subunit in mitochondrial NADH–ubiquinone oxidoreductase. Biochim Biophys Acta 2011;1807:1170–6. Nicholls DG. Mitochondrial ion circuits. Essays Biochem 2010;47:25–35. Ohnishi ST, Salerno JC, Ohnishi T. Possible roles of two quinone molecules in direct and indirect proton pumps of bovine heart NADH–quinone oxidoreductase (complex I). Biochim Biophys Acta 2010a;1797:1891–3. Ohnishi T. Iron–sulfur clusters/semiquinones in complex I. Biochim Biophys Acta 1998;1364:186–206. Ohnishi T. Structural biology: piston drives a proton pump. Nature 2010;465:428–9.

Ohnishi T, Nakamaru-Ogiso E. Were there any misassignments among iron–sulfur clusters N4, N5 and N6b in NADH–quinone oxidoreductase (complex I). Biochim Biophys Acta 2008;1777:703–10. Ohnishi T, Nakamaru-Ogiso E, Ohnishi ST. A new hypothesis on the simultaneous direct and indirect proton pump mechanisms in NADH–quinone oxidoreductase (complex I). FEBS Lett 2010b;584:4131–7. Ohnishi T, Sled VD, Yano T, Yagi T, Burbaev DS, Vinogradov AD. Structure–function studies of iron–sulfur clusters and semiquinones in the NADH–Q oxidoreductase segment of the respiratory chain. Biochim Biophys Acta 1998;1365:301–8. Orme-Johnson NR, Orme-Johnson WH, Hansen RE, Beinert H, Hatefi Y. EPR detectable electron acceptors in submitochondrial particles from beef heart with special reference to the iron–sulfur components of DPNH–ubiquinone reductase. Biochem Biophys Res Commun 1971;44:446–52. Osyczka A, Moser CC, Daldal F, Dutton PL. Reversible redox energy coupling in electron transfer chains. Nature 2004;427:607–12. Page CC, Moser CC, Chen XX, Dutton PL. Natural engineering principles of electron tunnelling in biological oxidation–reduction. Nature 1999;402:47–52. Petruzzella V, Sardanelli AM, Scacco S, Panelli D, Papa F, Trentadue R, et al. Dysfunction of mitochondrial respiratory chain complex I in neurological disorders: genetics and pathogenetic mechanisms. Adv Exp Med Biol 2012;942:371–84. Pilkington SJ, Skehel JM, Gennis RB, Walker JE. Relationship between mitochondrial NADH–ubiquinone reductase and a bacterial NAD-reducing hydrogenase. Biochemistry 1991;30:2166–75. Pohl T, Bauer T, Dörner K, Stolpe S, Sell P, Zocher G, et al. Iron–sulfur cluster N7 of the NADH:ubiquinone oxidoreductase (complex I) is essential for stability but not involved in electron transfer. Biochemistry 2007;46:6588–96. Powell MF, Bruice TC. Hydride vs electron-transfer in the reduction of flavin and flavin radical by 1,4-dihydropyridines. J Am Chem Soc 1983;105:1014–21. Pozzan T, Miconi V, Di Virgilio F, Azzone GF. H+ /site, charge/site, and ATP/site ratios at coupling sites I and II in mitochondrial e− transport. J Biol Chem 1979;254:10200–5. Pryde KR, Hirst J. Superoxide is produced by the reduced flavin in mitochondrial complex I: a single, unified mechanism that applies during both forward and reverse electron transfer. J Biol Chem 2011;286:18056–65. Pätsi J, Kervinen M, Finel M, Hassinen IE. Leber hereditary optic neuropathy mutations in the ND6 subunit of mitochondrial complex I affect ubiquinone reduction kinetics in a bacterial model of the enzyme. Biochem J 2008;409:129–37. Rasmussen T, Scheide D, Brors B, Kintscher L, Weiss H, Friedrich T. Identification of two tetranuclear FeS clusters on the ferredoxin-type subunit of NADH:ubiquinone oxidoreductase (complex I). Biochemistry 2001;40: 6124–31. Reda T, Barker CD, Hirst J. Reduction of the iron–sulfur clusters in mitochondrial NADH:ubiquinone oxidoreductase (complex I) by EuII-DTPA, a very low potential reductant. Biochemistry 2008;47:8885–93. Reynafarje B, Lehninger AL. The K+ /site H+ /site stoichiometry of mitochondrial electron transport. J Biol Chem 1978;253:6331–4. Rich PR. Electron and proton transfers through quinones and cytochrome bc complexes. Biochim Biophys Acta 1984;768:53–79. Rich PR, Maréchal A. The mitochondrial respiratory chain. Essays Biochem 2010;47:1–23. Roessler MM, King MS, Robinson AJ, Armstrong FA, Harmer J, Hirst J. Direct assignment of EPR spectra to structurally defined iron–sulfur clusters in complex I by double electron–electron resonance. Proc Natl Acad Sci U S A 2010;107:1930–5. Sahin-Toth M, Dunten RL, Gonzalez A, Kaback HR. Functional interactions between putative intramembrane charged residues in the lactose permease of Escherichia coli. Proc Natl Acad Sci U S A 1992;89:10547–51. Sazanov LA, Carroll J, Holt P, Toime L, Fearnley IM. A role for native lipids in the stabilization and two-dimensional crystallization of the Escherichia coli NADH–ubiquinone oxidoreductase (complex I). J Biol Chem 2003;278:19483–91. Sazanov LA, Hinchliffe P. Structure of the hydrophilic domain of respiratory complex I from Thermus thermophilus. Science 2006;311:1430–6. Scheide D, Huber R, Friedrich T. The proton-pumping NADH:ubiquinone oxidoreductase (complex I) of Aquifex aeolicus. FEBS Lett 2002;512:80–4. Schulte U, Haupt V, Abelmann A, Fecke W, Brors B, Rasmussen T, et al. A reductase/isomerase subunit of mitochondrial NADH:ubiquinone oxidoreductase (complex I) carries an NADPH and is involved in the biogenesis of the complex. J Mol Biol 1999;292:569–80. Sebban P, Maroti P, Schiffer M, Hanson DK. Electrostatic dominoes: long distance propagation of mutational effects in photosynthetic reaction centers of Rhodobacter capsulatus. Biochemistry 1995;34:8390–7. Sekiguchi K, Murai M, Miyoshi H. Exploring the binding site of acetogenin in the ND1 subunit of bovine mitochondrial complex I. Biochim Biophys Acta 2009;1787:1106–11. Shinzawa-Itoh K, Seiyama J, Terada H, Nakatsubo R, Naoki K, Nakashima Y, et al. Bovine heart NADH–ubiquinone oxidoreductase contains one molecule of ubiquinone with ten isoprene units as one of the cofactors. Biochemistry 2010;49:487–92. Shiraishi Y, Murai M, Sakiyama N, Ifuku K, Miyoshi H. Fenpyroximate binds to the interface between PSST and 49 kDa subunits in mitochondrial NADH–ubiquinone oxidoreductase. Biochemistry 2012;51:1953–63. Scholes TA, Hinkle PC. Energetics of ATP-driven reverse electron transfer from cytochrome c to fumarate and from succinate to NAD in submitochondrial particles. Biochemistry 1984;23:3341–5. Sinegina L, Wikström M, Verkhovsky MI, Verkhovskaya ML. Activation of isolated NADH:ubiquinone reductase I (complex I) from Escherichia coli by detergent

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024

G Model BC-3836; No. of Pages 21

ARTICLE IN PRESS M. Verkhovskaya, D.A. Bloch / The International Journal of Biochemistry & Cell Biology xxx (2012) xxx–xxx

and phospholipids. Recovery of ubiquinone reductase activity and changes in EPR signals of iron–sulfur clusters. Biochemistry 2005;44:8500–6. Sinha PK, Torres-Bacete J, Nakamaru-Ogiso E, Castro-Guerrero N, Matsuno-Yagi A, Yagi T. Critical roles of subunit NuoH (ND1) in the assembly of peripheral subunits with the membrane domain of Escherichia coli NDH-1. J Biol Chem 2009;284:9814–23. Sled VD, Rudnitzky NI, Hatefi Y, Ohnishi T. Thermodynamic analysis of flavin in mitochondrial NADH ubiquinone oxidoreductase (complex-I). Biochemistry 1994;33:10069–75. Sled VD, Vinogradov AD. Kinetics of the mitochondrial NADH–ubiquinone oxidoreductase interaction with hexammineruthenium(III). Biochim Biophys Acta 1993;1141:262–8. Steigmiller S, Turina P, Gräber P. The thermodynamic H+ /ATP ratios of the H+ ATP synthases from chloroplasts and Escherichia coli. Proc Natl Acad Sci U S A 2008;105:3745–50. Steimle S, Bajzath C, Dorner K, Schulte M, Bothe V, Friedrich T. Role of subunit NuoL for proton translocation by respiratory complex I. Biochemistry 2011;50:3386–93. Sumner JS, Matthews RG. Stereochemistry and mechanism of hydrogen transfer between NADPH and methylenetetrahydrofolate in the reaction catalyzed by methylenetetrahydrofolate reductase from pig liver. J Am Chem Soc 1992;114:6949–56. Takamiya KI, Dutton PL. Ubiquinone in Rhodopseudomonas sphaeroidea. Some thermodynamic properties. Biochim Biophys Acta 1979;546:1–16. Tocilescu MA, Fendel U, Zwicker K, Dröse S, Kerscher S, Brandt U. The role of a conserved tyrosine in the 49-kDa subunit of complex I for ubiquinone binding and reduction. Biochim Biophys Acta 2010a;1797:625–32. Tocilescu MA, Fendel U, Zwicker K, Kerscher S, Brandt U. Exploring the ubiquinone binding cavity of respiratory complex I. J Biol Chem 2007;282:29514–20. Tocilescu MA, Zickermann V, Zwicker K, Brandt U. Quinone binding and reduction by respiratory complex I. Biochim Biophys Acta 2010b;1797:1883–90. Torres-Bacete J, Nakamaru-Ogiso E, Matsuno-Yagi A, Yagi T. Characterization of the NuoM (ND4) subunit in Escherichia coli NDH-1: conserved charged residues essential for energy-coupled activities. J Biol Chem 2007;282:36914–22. Torres-Bacete J, Sinha PK, Matsuno-Yagi A, Yagi T. Structural contribution of Cterminal segments of NuoL (ND5) and NuoM (ND4) subunits of complex I from Escherichia coli. J Biol Chem 2011;286:34007–14. Treberg JR, Brand MD. A model of the proton translocation mechanism of complex I. J Biol Chem 2011;286:17579–84. Treberg JR, Quinlan CL, Brand MD. Evidence for two sites of superoxide production by mitochondrial NADH–ubiquinone oxidoreductase (complex I). J Biol Chem 2011;286:27103–10. Trimmer EE, Ballou DP, Galloway LJ, Scannell SA, Brinker DR, Casas KR. Aspartate 120 of Escherichia coli methylenetetrahydrofolate reductase: evidence for major roles in folate binding and catalysis and a minor role in flavin reactivity. Biochemistry 2005;44:6809–22. Ullmann GM. The coupling of protonation and reduction in proteins with multiple redox centers: theory, computational method, and application to cytochrome c3 . J Phys Chem B 2000;104:6293–301. van Belzen R, Kotlyar AB, Moon N, Dunham WR, Albracht SP. The iron–sulfur clusters 2 and ubisemiquinone radicals of NADH:ubiquinone oxidoreductase are involved in energy coupling in submitochondrial particles. Biochemistry 1997;36:886–93. Velazquez I, Nakamaru-Ogiso E, Yano T, Ohnishi T, Yagi T. Amino acid residues associated with cluster N3 in the NuoF subunit of the protontranslocating NADH–quinone oxidoreductase from Escherichia coli. FEBS Lett 2005;579:3164–8. Venturoli G, Drepper F, Williams JC, Allen JP, Lin X, Mathis P. Effects of temperature and G◦ on electron transfer from cytochrome c2 to the photosynthetic reaction center of the purple bacterium Rhodobacter sphaeroides. Biophys J 1998;74:3226–40. Verkhovskaya M, Knuuti J, Wikström M. Role of Ca2+ in structure and function of complex I from Escherichia coli. Biochim Biophys Acta 2011;1807:36–41. Verkhovskaya ML, Belevich N, Euro L, Wikström M, Verkhovsky MI. Real-time electron transfer in respiratory complex I. Proc Natl Acad Sci U S A 2008;105:3763–7. Verkhovsky M, Bloch DA, Verkhovskaya M. Tightly-bound ubiquinone in the Escherichia coli respiratory complex I. Biochim Biophys Acta 2012a;1817:1550–6.

21

Verkhovsky MI, Bogachev AV. Sodium-translocating NADH:quinone oxidoreductase as a redox-driven ion pump. Biochem Biophys Acta 2010;1797:738–46. Verkhovsky MI, Bogachev AV, Pivtsov AV, Bertsova YV, Fedin MV, Bloch DA, Kulik LV. Sodium-dependent movement of covalently bound FMN residue(s) in Na+-translocating NADH:quinone oxidoreductase. Biochemistry 2012b;1817:1550–6. Vinogradov AD, Grivennikova VG. Generation of superoxide-radical by the NADH:ubiquinone oxidoreductase of heart mitochondria. Biochemistry (Moscow) 2005;70:120–7. Wang H, Lin S, Katilius E, Laser C, Allen JP, Williams JC, et al. Unusual temperature dependence of photosynthetic electron transfer due to protein dynamics. J Phys Chem B 2009;113:818–24. Watt IN, Montgomery MG, Runswick MJ, Leslie AGW, Walker JE. Bioenergetic cost of making an adenosine triphosphate molecule in animal mitochondria. Proc Natl Acad Sci USA 2010;107:16823–7. Wikström M. Two protons are pumped from the mitochondrial matrix per electron transferred between NADH and ubiquinone. FEBS Lett 1984;169:300–4. Wikström M, Hummer G. Stoichiometry of proton translocation by respiratory complex I and its mechanistic implications. Proc Natl Acad Sci U S A 2012;109:4431–6. Wikström M, Verkhovsky MI. Mechanism and energetics of proton translocation by the respiratory heme-copper oxidases. Biochim Biophys Acta 2007;1767:1200–14. Yakovlev G, Reda T, Hirst J. Reevaluating the relationship between EPR spectra and enzyme structure for the iron sulfur clusters in NADH:quinone oxidoreductase. Proc Natl Acad Sci U S A 2007;104:12720–5. Yano T, Dunham WR, Ohnishi T. Characterization of the ␮H+ -sensitive ubisemiquinone species (SQNf ) and the interaction with cluster N2: new insight into the energy-coupled electron transfer in complex I. Biochemistry 2005;44:1744–54. Yano T, Magnitsky S, Ohnishi T. Characterization of the complex I-associated ubisemiquinone species: toward the understanding of their functional roles in the electron/proton transfer reaction. Biochim Biophys Acta 2000;1459:299–304. Yano T, Magnitsky S, Sled VD, Ohnishi T, Yagi T. Characterization of the putative 2x[4Fe–4S]-binding NQO9 subunit of the proton-translocating NADH–quinone oxidoreductase (NDH-1) of Paracoccus denitrificans. Expression, reconstitution, and EPR characterization. J Biol Chem 1999;274:28598–605. Yano T, Sklar J, Nakamaru-Ogiso E, Takahashi Y, Yagi T, Ohnishi T. Characterization of cluster N5 as a fast-relaxing [4Fe–4S] cluster in the Nqo3 subunit of the protontranslocating NADH–ubiquinone oxidoreductase from Paracoccus denitrificans. J Biol Chem 2003;278:15514–22. Yano T, Sled VD, Ohnishi T, Yagi T. Expression and characterization of the flavoprotein subcomplex composed of 50-kDa (NQO1) and 25-kDa (NQO2) subunits of the proton-translocating NADH–quinone oxidoreductase of Paracoccus denitrificans. J Biol Chem 1996;271:5907–13. Yip CY, Harbour ME, Jayawardena K, Fearnley IM, Sazanov LA. Evolution of respiratory complex I: supernumerary subunits are present in the alphaproteobacterial enzyme. J Biol Chem 2011;286:5023–33. Yu CA, Cen X, Ma HW, Yin Y, Yu L, Esser L, et al. Domain conformational switch of the iron–sulfur protein in cytochrome bc1 complex is induced by the electron transfer from cytochrome bL to bH. Biochim Biophys Acta 2008;1777:1038–43. Zhang P, Battchikova N, Paakkarinen V, Katoh H, Iwai M, Ikeuchi M, et al. Isolation, subunit composition and interaction of the NDH-1 complexes from Thermosynechococcus elongatus BP-1. Biochem J 2005;390:513–20. Zickermann V, Bostina M, Hunte C, Ruiz T, Radermacher M, Brandt U. Functional implications from an unexpected position of the 49-kDa subunit of NADH:ubiquinone oxidoreductase. J Biol Chem 2003;278:29072–8. Zickermann V, Dröse S, Tocilescu MA, Zwicker K, Kerscher S, Brandt U. Challenges in elucidating structure and mechanism of proton pumping NADH:ubiquinone oxidoreductase (complex I). J Bioenerg Biomembr 2008;40:475–83. Zu Y, Di Bernardo S, Yagi T, Hirst J. Redox properties of the [2Fe–2S] center in the 24 kDa (NQO2) subunit of NADH:ubiquinone oxidoreductase (complex I). Biochemistry 2002;41:10056–69. Zwicker K, Galkin A, Dröse S, Grgic L, Kerscher S, Brandt U. The Redox-Bohr group associated with iron–sulfur cluster N2 of complex I. J Biol Chem 2006;281:23013–7.

Please cite this article in press as: Verkhovskaya M, Bloch DA. Energy-converting respiratory Complex I: On the way to the molecular mechanism of the proton pump. Int J Biochem Cell Biol (2012), http://dx.doi.org/10.1016/j.biocel.2012.08.024