News & Views | Published:

Structural biology

Piston drives a proton pump

Nature volume 465, pages 428429 (27 May 2010) | Download Citation

The membrane-spanning enzyme known as complex I couples the movement of electrons to that of protons as a way of converting energy. Crystal structures suggest how electron transfer drives proton pumping from afar.

Complex I is one of the energy-converting enzyme complexes found in the membranes of the cell's fuel factories, the mitochondria, and was the last such complex without a structural portrait. But in an epoch-making paper in this issue, Sazanov and colleagues1 describe X-ray structures of bacterial complex I, and report that it has an unusual 'piston' mechanism for controlling proton movement across mitochondrial membranes (see page 441).

Humans and other animals derive most of their energy from a highly efficient process called oxidative phosphorylation. During this process, the part of the food we eat that is used as fuel is completely oxidized in mitochondria. The oxidation energy is used to synthesize adenosine triphosphate (ATP), the molecular energy carrier for almost all life processes.

During oxidative phosphorylation, electrons from the metabolic intermediate NADH are transported to oxygen through a set of biochemical reactions known as the aerobic respiratory chain. The chain involves three transmembrane energy-converting enzyme complexes arranged in sequence, of which complex I (also known as NADH–quinone oxidoreductase) is the first. Electron transfer through these complexes provides the energy to establish an electrochemical gradient of hydrogen ions (protons, H+) across the mitochondrial membrane. This, in turn, drives the enzyme that makes ATP. The general mechanistic principle connecting proton transport to ATP synthesis, known as the chemiosmotic theory2, was proposed by the biochemist Peter Mitchell, and won him the 1978 Nobel Prize in Chemistry.

Mitochondrial complex I was first isolated from bovine-heart mitochondria about 50 years ago. It has a molecular mass of about 1,000 kilodaltons and is composed of 45 different subunits3. Bacterial complex I, which is found in bacterial-cell internal membranes, is much simpler, with just 14 subunits and a molecular mass of about 500 kilodaltons (ref. 4). Despite these differences, the cofactors in mitochondrial and bacterial complex I are the same, generally consisting of one flavin mononucleotide (FMN), eight clusters of iron and sulphur atoms (Fe–S clusters), and perhaps two kinds of protein-associated quinone molecule5,6. Bacterial complex I also performs the same function as the mitochondrial enzyme, transporting two electrons from NADH to quinone molecules. For these reasons, bacterial complex I serves as a minimal model of mitochondrial complex I.

Structural information about complex I has come from many sources. An electron-microscope analysis7 of complex I from the mould Neurospora crassa revealed a bipartite structure, in which almost equally sized intra- and extramembrane parts form an L-shape. A spectroscopic technique known as electron paramagnetic resonance (EPR) also provided valuable insight by predicting the number of flavins, Fe–S clusters and quinones, and their physicochemical properties5,6.

In 2006, Sazanov and Hinchliffe8 reported a high-resolution X-ray crystal structure of the extramembrane part of complex I from the bacterium Thermus thermophilus. This confirmed the total number of flavin and Fe–S clusters predicted by EPR, and revealed a long (approximately 95 Å) pathway along which electrons from the FMN cofactor are passed, through a chain of seven Fe–S clusters. Another Fe–S cluster at the side of FMN can also act as part of the electron-transfer pathway. The structure8 suggested that no cofactors other than the protein-associated quinone are located in the intramembrane part of bacterial complex I. A recent study9 revealed that bovine-heart complex I retains one tightly bound quinone molecule, which acts as a cofactor that is essential for the complex's activity.

Of course, the electron-transfer pathway and the cofactors are only part of the story. The other crucial aspect of complex I is its ability to pump several protons — four is the current consensus — across a membrane for each pair of electrons passed along the chain of cofactors10,11. Accordingly, the intramembrane part of complex I must contain one or more proton-pumping 'devices' that are driven by the energy released during electron transfer. The molecular mechanism of this proton pumping, however, has remained obscure.

Sazanov and colleagues1 now report the long-awaited X-ray structures of the intramembrane part of complex I from the bacterium Escherichia coli, and of the entire complex I from T. thermophilus. This is another great achievement by Sazanov's group. Their findings reveal that the three largest transmembrane subunits of complex I (NuoL, NuoM and NuoN) are very similar in structure, and resemble Na+/H+ antiporter proteins12 that transfer sodium ions and protons in opposite directions across membranes. Each of these subunits contains transmembrane α-helices that contain a 'helix–peptide–helix' motif, known to be important for ion translocation in an E. coli Na+/H+ antiporter13 and other transporters.

Strikingly, the authors found that the NuoL subunit differs in one important respect from NuoM and NuoN: it has an unusually long (110 Å) α-helix that is aligned parallel to the membrane. This 'horizontal' helix extends to a transmembrane α-helix that lies close to the bundle of subunits (NuoA/J/K/H) next to the bound quinone and the electron-transfer pathway. The authors therefore propose a unique 'piston-like' mechanism for proton pumping by complex I, in which the long, horizontal α-helix acts as a connecting rod that can simultaneously open and close antiporter-like channels in NuoL, NuoM and NuoN (Fig. 1).

Figure 1: Indirect coupling of electron transfer to proton pumping in complex I.
Figure 1

Complex I, an enzyme found in mitochondrial and bacterial membranes, converts energy by coupling electron transfer to proton pumping. Sazanov and colleagues' crystal structures1 of bacterial complex I reveal that the transmembrane NuoL subunit of the enzyme projects a long α-helix through the adjacent NuoM and NuoN subunits. They suggest the following mechanism to explain how electron transfer drives proton pumping. a, Pairs of electrons from the metabolic intermediate NADH are transferred to a cofactor (flavin mononucleotide, FMN) and then passed along a chain of iron–sulphur clusters in the extramembrane region of complex I, eventually reaching a quinone cofactor; blue arrows indicate the electron-transfer pathway. This allows a proton (H+) to pass through complex I at the interface of the extra- and intramembrane regions. Protons can also enter channels in NuoL, NuoM and NuoN from the cytoplasm, but cannot pass through. White circles with minus signs represent negatively charged amino acids, which are key to proton transport. b, Conformational changes in the NuoA/J/K/H subunits push the long α-helix towards the other transmembrane subunits. This tilts three other helices in NuoL, NuoM and NuoN, causing the reorientation of certain residues in the subunits' channels. These local conformational changes allow protons in the channels to pass through the channels and enter the periplasm (the space between the inner and outer bacterial membranes).

Sazanov's group had previously reported14 that NuoL (and its neighbour, NuoM) are too far from the electron-transfer pathway to be directly influenced by it. The idea that proton pumping must be driven indirectly by long-range conformational changes has therefore been widely discussed14,15,16,17, but no experimental evidence for a molecular mechanism had been found that could cause such indirect coupling. The beautiful new crystal structure1 finally provides a plausible mechanism.

Some important information about the structure of complex I remains missing. It has been suggested that two quinone binding sites reside in the loops on one side of the NuoH (ref. 18) and NuoN (ref. 19) subunits, but in Sazanov and colleagues' structure1, these loop regions are disordered. The quinone binding sites may move dramatically during the enzyme's reactions, and so the conformational flexibility of these functionally important regions might be the source of this disorder. An interesting future avenue of research would be to obtain crystal structures of complex I in which the loop conformations have been fixed by the binding of quinone analogues that inhibit electron transfer from NADH to quinone.

Sazanov and colleagues' complete structure1 of complex I was obtained at relatively low resolution but, even so, will provide significant impetus to move research in this field forward again. For example, it will guide the selection of specific mutations that can be made in complex I to interrogate the effects of structural changes on the function of the enzyme. Both of the structures reported in this work1 also raise several big questions for the field. What is the direct driving force for the piston? And what are the possible roles of the quinone molecules? One intriguing possibility is that they might directly drive another mechanism for proton pumping that is connected with the NuoA/J/K/H subunits. Could it be that a direct mechanism of this sort pumps two protons, whereas the indirect proton pump in NuoL and NuoM translocates another two protons?


  1. 1.

    , & Nature 465, 441–445 (2010).

  2. 2.

    Science 206, 1148–1159 (1979).

  3. 3.

    , , , & Mol. Cell. Proteomics 2, 117–126 (2003).

  4. 4.

    , , & Biochim. Biophys. Acta 1364, 125–133 (1998).

  5. 5.

    Biochim. Biophys. Acta 1364, 186–206 (1998).

  6. 6.

    & FEBS Lett. 579, 4555–4561 (2005).

  7. 7.

    , & J. Mol. Biol. 221, 1027–1043 (1991).

  8. 8.

    & Science 311, 1430–1436 (2006).

  9. 9.

    et al. Biochemistry 49, 487–492 (2010).

  10. 10.

    FEBS Lett. 169, 300–304 (1984).

  11. 11.

    , & FEBS Lett. 451, 157–161 (1999).

  12. 12.

    & J. Theor. Biol. 187, 529–540 (1997).

  13. 13.

    et al. Nature 435, 1197–1202 (2005).

  14. 14.

    , , & Biochemistry 39, 7229–7235 (2000).

  15. 15.

    J. Bioenerg. Biomembr. 33, 169–177 (2001).

  16. 16.

    & Biochemistry 42, 2266–2274 (2003).

  17. 17.

    Annu. Rev. Biochem. 75, 69–92 (2006).

  18. 18.

    & Biochim. Biophys. Acta 1504, 352–362 (2001).

  19. 19.

    et al. FEBS Lett. 584, 883–888 (2010).

Download references

Author information


  1. Tomoko Ohnishi is in the Johnson Foundation, Department of Biochemistry and Biophysics, School of Medicine, University of Pennsylvania, Pennsylvania 19104, USA.

    • Tomoko Ohnishi


  1. Search for Tomoko Ohnishi in:

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing