Structure and Mechanism of Respiratory III-IV Supercomplexes in Bioenergetic Membranes

Abstract

In the final steps of energy conservation in aerobic organisms, free energy from electron transfer through the respiratory chain is transduced into a proton electrochemical gradient across a membrane. In mitochondria and many bacteria, reduction of the dioxygen electron acceptor is catalyzed by cytochrome c oxidase (complex IV), which receives electrons from cytochrome bc₁ (complex III), via membrane-bound or water-soluble cytochrome c. These complexes function independently, but in many organisms they associate to form supercomplexes. Here, we review the structural features and the functional significance of the nonobligate III₂IV_1/2 Saccharomyces cerevisiae mitochondrial supercomplex as well as the obligate III₂IV₂ supercomplex from actinobacteria. The analysis is centered around the Q-cycle of complex III, proton uptake by CytcO, as well as mechanistic and structural solutions to the electronic link between complexes III and IV.

Figure 1

The mitochondrial respiratory chain. (A) Complex I of mammalian mitochondria is not present in S. cerevisiae. Instead, the external (Nde1, Nde2) and internal (Ndi1) membrane-associated NADH dehydrogenases catalyze the same NADH-oxidation:Q reduction reaction as complex I. All these enzymes are shown here in the same membrane only to illustrate the different pathways of NADH oxidation. The structures originate from different organisms: T. thermophilus complex I (PDB 3M9S), S. cerevisiae Ndi1 (PDB 4G9K), S. scrofa (pig) complex II (PDB 1ZOY), S. cerevisiae complex III and IV (PDB 6HU9), S. cerevisiae complex V (PDB 6CP6), and S. cerevisiae cyt. c (PDB 1YCC). (B) The respiratory chain is found in protrusions of the inner membrane that are called cristae. Here, the respiratory chain components I–IV (only complexes III and IV are shown) are located in the flat regions, while the ATP synthase (complex V) is restricted to the bent end regions. Approximate dimension and average distance are from refs (,−52). The cyt. c:CytcO ratio in S. cerevisiae is 2–4, which is equivalent to an average concentration of ∼100 μM cyt. c in the intercristae space.^,

Figure 2

Structures of III₂IV₂supercomplexes. (A) S. cerevisiae supercomplex (PDB 6HU9). Catalytically important subunits of complexes III are cytb, the Rieske iron–sulfur protein (also called Rip1 in S. cerevisiae) and cyt1, while those of complex IV are cox1–3 (also called SU I–III). (B) M. smegmatis supercomplex (PDB 6HWH, SodC is 1PZS). Catalytically important subunits of complexes III and IV are QcrA-C and CtaC-F (equivalent of SU I–III), respectively. The equivalent of canonical SU III is composed of two parts, CtaE and CtaF. Unidentified subunits are shown in gray. The SodC-type superoxide dismutase dimer subunit (PDB 1PZS) was identified in the structure.^, It was less resolved in ref (44), which did not allow identification of a connection between the subunit and the rest of the supercomplex (illustrated by the dashed line).

Figure 3

Distances between cofactors. (A) S. cerevisiae (PDB 6HU9) and (B) M. smegmatis (PDB 6HWH) supercomplexes. In (A), distances for the FeS center in the C (FeS_C) and B (FeS_B) positions, respectively (see inset), are indicated in the two halves of the complex III₂ dimer. Note that the arrangement shown in (A) is a fusion of two different structures where the FeS center is either in FeS_B (left monomer) or FeS_C (right monomer) (B position PDB is 1KYO, C position is PDB 3H1H). The positions of cyt. c bound to cyt. bc₁ or CytcO are indicated (cyt. c at complex III is PDB 1KYO, cyt. c at complex IV is PDB 5IY5), see also ref (37). In (B), the open and closed conformations of the cyt. cc domain, observed in a single supercomplex, are shown (SodC is PDB 1PZS).

Figure 4

Complex III. (A) The catalytically important subunits of one monomer of complex III₂ (cyt. bc₁) from S. cerevisiae (PDB 6HU9) and the catalyzed reaction. The electron-transfer paths along the B and C branches are indicated with dashed lines, while proton uptake and release are shown with blue arrows. Note that the total stoichiometry of electron and proton transfer is indicated for oxidation of two QH₂. Upon oxidation of each QH₂ in the Q_P site, two electrons are transferred, one electron along each of the B and C branches, respectively. One H⁺ is transferred to His161 (His181 in S. cerevisiae) ligand of the FeS center (shown in B) and is transferred to the p side upon movement of the FeS-domain from the B position (FeS_B in the right-hand side inset to A) to the C position (FeS_C). The second H⁺ is transferred via protonatable residues of the cyt. b subunit (see text). The same sequence of electron and H⁺ transfer is repeated upon binding of the second QH₂ in the Q_P site. The inset on the lower left shows all subunits of the dimer, including accessory subunits in gray and bound cyt. c (PDB 1KYO). In main panel A, the FeS center is found in an intermediate B/C position. (B) The Q_P binding site of O. aries (sheep, PDB 6Q9E) with a bound ubiquinone (UQ), the only structure of a mitochondrial cyt. bc₁ in which the Q_P site is occupied by Q. The Q_P site and all functionally important residues are conserved in the S. cerevisiae cyt. bc₁. (C) The Q_P site of M. smegmatis complex III (PDB 6ADQ).

Figure 5

Complex IV. (A) The core subunits of the S. cerevisiae CytcO (complex IV, PDB 6HU9) and the catalyzed reaction. The inset shows all subunits of the S. cerevisiae CytcO, including accessory subunits in gray and bound cyt. c (based on the cyt. c position in the bovine CytcO, PDB 5IY5, which displays the same geometry as the S. cerevisiae cyt. c-CytcO cocomplex). The D and K proton pathways of the S. cerevisiae (B) and M. smegmatis (PDB 6HWH) (C) CytcOs. In (B), water molecules seen in the crystal structures of bacterial and mammalian CytcOs are included. They were not resolved in the cryo-EM structures of the S. cerevisiae CytcO. (C) The QcrB “lid” of complex III, which covers the D pathway of CytcO in the M. smegmatis supercomplex. Amino acid residue side chains of QcrB that provide an alternative entry pathway to D115 are shown (along the blue arrow below the D pathway).

Figure 6

Reduction of O₂ at the catalytic site of CytcO. The first electron (e^–) from cyt. c to the oxidized CytcO (state O) is transferred to Cu_B to form state E. It is accompanied by proton uptake from the n side solution though the K pathway (H_K⁺) to Tyr245 (S. cerevisiae CytcO numbering, Tyr in the figure). Transfer of the second electron to heme a₃ and a proton through the K pathway to a hydroxide at heme a₃ leads to formation of state R, where the catalytic site is reduced by two electrons. Next, O₂ binds to heme a₃ forming state A. After transfer of one electron and one proton from the Tyr residue, a ferryl state is formed, called P (“peroxy”, for historical reasons). Transfer of the third electron is accompanied by proton uptake through the D pathway (H_D⁺) and formation of the ferryl state, F. After transfer of the fourth electron and another proton through the D pathway to the catalytic site, the oxidized state O is formed again. The four transitions P → F, F → O, O → E, and E → R are each associated with pumping of one proton across the membrane. These protons are taken up through the D pathway (H_D⁺, each proton released to the p side is indicated as H⁺).

Figure 7

Arrangement of supercomplexes with known structures that contain complexes III and IV in different species. The B. taurus (cow) CytcO dimer is also shown (PDB 1OCC), other references are given in Table 1. The main panel in the middle shows the alignment of complex III₂ relative to the position of complex IV with its subunits at the interface of complex III₂ indicated in bold text and in different colors. All interacting subunits for all supercomplexes are marked in specific colors for reference in order to indicate their relative positions in all supercomplexes. The prefixes Bt (B. taurus), Sc (S. cerevisiae), and Vr (V. radiata) are added because of the different subunit numbering used for the equivalent subunits in different organisms. The mitochondrial supercomplexes are shown on the left, while the obligate M. smegmatis III₂IV₂ and the alternative complex III–IV supercomplexes are shown on the right. The S. cerevisiae supercomplex is encircled by a blue line; it is shown as reference for both the mitochondrial and bacterial supercomplexes. Top views and side views with approximate positions of the membrane with black lines. The inset shows the same supercomplexes but aligned to the complex III₂ dimer (alternative complex III in F. johnsoniae is not shown here).

Figure 8

Cardiolipin in complex III–IV supercomplexes. All cardiolipin (shown in red) head groups face the n side. The boundaries of complexes III and IV are indicated by solid lines. The dashed lines indicate boundaries on the opposite side of each supercomplex. (A) The S. cerevisiae supercomplex. Subunits are colored as in Figure 2A. (B) The M. smegmatis supercomplex. Subunits are colored as in Figure 2B.

Figure 9

Surface representation of the electrostatic potential in III–IV supercomplexes. The S. cerevisiae (PDB 6HU9) (A), B. taurus (cow) (PDB 5LUF) (B), and M. smegmatis (PDB 6HWH) (C) supercomplexes are shown. Cyt. c is from either S. cerevisiae (A, PBD 1YCC) or B. taurus (B, PDB 2B4Z). For M. smegmatis (C), the cyt. cc head domain of QcrC in the closed conformation was separated from the supercomplex and the electrostatic potentials were calculated separately for the supercomplex and cyt. cc domain, respectively. The original position of the cyt. cc domain at the top of the supercomplex is encircled by a black line in (C). Color range from red to blue for an electrostatic potential from −5 to +5 k_BT/q, where k_B is the Boltzmann constant, T is the absolute temperature, and q is a the unit charge. The figure was prepared using the APBS tool with standard settings of the PyMOL software (Molecular Graphics System, version 2.4; Schrödinger, LLC).

Figure 10

Model for electron transfer from cyt.bc₁ to CytcO in the S. cerevisiae supercomplex. (A) Electron transfer via 3D diffusion of cyt. c. (B) Electron transfer via 2D diffusion of cyt. c. Note that the surface-attached cyt. c is assumed to be in equilibrium with the cyt. c pool, but the time constant for equilibration of the surface-attached cyt. c with the pool cyt. c (as well as electron transfer between the surface-attached cyt. c and pool cyt. c) is assumed to be slower than diffusion between the binding sites at cyt. bc₁ and CytcO (modeled after ref (37)). S. cerevisiae supercomplex and cyt. c are PDBs 6HU9 and 1YCC, respectively.