The evolution of the human mitochondrial bc1 complex- adaptation for reduced rate of superoxide production?

The mitochondrial bc1 complex is a major source of mitochondrial superoxide. While bc1-generated superoxide plays a beneficial signaling role, excess production of superoxide lead to aging and degenerative diseases. The catalytic core of bc1 comprises three peptides -cytochrome b, Fe-S protein, and cytochrome c1. All three core peptides exhibit accelerated evolution in anthropoid primates. It has been suggested that the evolution of cytochrome b in anthropoids was driven by a pressure to reduce the production of superoxide. In humans, the bc1 core peptides exhibit anthropoid-specific substitutions that are clustered near functionally critical sites that may affect the production of superoxide. Here we compare the high-resolution structures of bovine, mouse, sheep and human bc1 to identify structural changes that are associated with human-specific substitutions. Several cytochrome b substitutions in humans alter its interactions with other subunits. Most significantly, there is a cluster of seven substitutions, in cytochrome b, the Fe-S protein, and cytochrome c1 that affect the interactions between these proteins at the tether arm of the Fe-S protein and may alter the rate of ubiquinone oxidation and the rate of superoxide production. Another cluster of substitutions near heme bH and the ubiquinone reduction site, Qi, may affect the rate of ubiquinone reduction and thus alter the rate of superoxide production. These results are compatible with the hypothesis that cytochrome b in humans (and other anthropoid primates) evolve to reduce the rate of production of superoxide thus enabling the exceptional longevity and exceptional cognitive ability of humans.


Introduction
The mitochondrial bc1 complex, complex III of the mitochondrial electron transport chain, is a ubiquinol/cytochrome c oxidoreductase, i.e., it oxidizes ubiquinol to ubiquinone and reduces cytochrome c + 3 to cytochrome c + 2 . This reaction is coupled to the generation of a protonmotive force across the mitochondrial inner membrane that together with the other mitochondrial proton pumps, NADH dehydrogenase (complex I) and cytochrome c oxidase (complex IV) generate a protonmotive force sufficient to reverse ATP synthase (complex V) and drive the synthesis of ATP (Mitchell 1979). The bc1 complex in animal mitochondria is a dimer of 2 identical monomers each consisting of 11 cytochrome c1. Cytochrome c1 also have one TMH and a head group that contain the high potential heme c that is also located on the P-side above cytochrome b (Fig. 1).
The mechanism by which the 2 electrons oxidation of ubiquinol to ubiquinone is converted by bc1 into one electron reduction of cytochrome c and generate a protonmotive force was first explained by Mitchell's hypothetical Q cycle (Mitchell 1975). The modified Q cycle mechanism is now well established (Xia et al. 2013;Crofts 2021). Accordingly, the Q cycle begin by binding of ubiquinol at a Qo site where the high potential [2Fe-2 S] in the ISP-HD is in the "b" position, bound to ubiquinone. Then, a bifurcation of the ubiquinol reducing electrons occur where the [2Fe-2 S] accept one electron and the other electron is accepted by heme bL. At the same time one proton is taken by the ISP-HD and the other is released to the P-side, thus producing a ubiquinone at the Qo site (Fig. 2). The IDP-HD with the reduced [2Fe-2 S] center then move from the "b" position toward the "c" position, near the high potential heme c, of cytochrome c1, and transfer the electron to heme c, that later donate this electron to the mobile cytochrome c that bind to cytochrome c1. At the same time heme bL transfer the second electron to heme bH across the membrane thereby generating a membrane potential. Heme bH transfer the electron to a ubiquinone that is bound at the Qi site which then accept a proton from the N-side thus generating a protonmotive force while forming a semiquinone, completing half of the Q cycle. In the second half of the cycle, another ubiquinol enter the Qo site and the one electron reduction of the high potential chain is repeated, while the second electron is transferred to the low potential chain: from heme bL to heme bH to the semiquinone at the Qi site. After the semiquinone that is bound at the Qi site accept the second electron from heme bH it accepts another proton from the N-side completing the formation of a ubiquinol at the Qi Fig. 1 The core peptides of the bovine bc1 The figure shows the core peptides of the bovine bc1: cytochrome b (brown), cytochrome c1(green) and the Fe-S protein (grey). The Fe-S protein head group is in the "b" position with the [Fe2-S2] redox center bound to stigmatellin in the Qo site. The three hemes bL, bH (cytochrome b) and heme c (cytochrome c1) are also shown. Drown form [PDB: 1PPJ]. site. This ubiquinol diffuses across the membrane to the Qo site thereby completing the Q cycle ( Fig. 2) The net result of the complete cycle is: Ubiquinol (P-side) + 2 cyt c + 3 (-P-side) + 2 H + (N-side)--->Ubiquinone (P-side) + 2 cyt c + 2 (-P-side) + 2 H + (P-side). Therefore, the stoichiometry of proton pumping by the bc1 complex is only 1 H + /e − , which is only half as much as that of complexes I and IV (2 H + /e − ).
During the Q cycle a leak of electrons to oxygen occurs within bc1 complex from a semiquinone radical that interacts with oxygen to generate superoxide. The semiquinone that is bound strongly at the Qi site is stable, does not interact with oxygen, and does not generate significant amount of superoxide. However, the semiquinone that is formed transiently at the Qo site is not stabilized, can interact with oxygen, and enable the generation of superoxide by the bc1 complex. The highest rate of superoxide generation by bc1 is observed when electron transfer to ubiquinone at the Qi site is blocked. Thus, when antimycin is bound at the Qi site, preventing the binding of ubiquinone at this site, electron transport of the second electron to the low potential chain is completely inhibited, enabling the transfer of an electron, from the unstable ubiquinone radical at the Qo site to oxygen, generating superoxide. During the operation of the Q cycle the rate of superoxide generation depends, in a complicated and still not fully understood ways, on the kinetics and thermodynamic parameters of this complex system (cf. Dröse & and Brandt 2008, Quinlan et al. 2011, Bazil et al. 2013, Markevich and Hoek 2015, Guillaud et al. 2014, Bujnowicz et al. 2019, Pagacz et al. 2021. It is still not clear whether the semiquinone in the Qo site that interacts with oxygen to produce superoxide is formed transiently by the electron and proton transfer to the [2Fe-2 S] center or that the semiquinone that produce superoxide in the Qo site is formed by reverse electron transfer from heme bL to ubiquinone. Most likely both mechanisms contribute to superoxide generation depending on the kinetics of the reaction (Pagacz et al. 2021). Other than the intrinsic kinetic and thermodynamic parameters of the bc1 complex, two important extrinsic variables that can significantly affect the rate of superoxide generation by bc1 is the ratio of the ubiquinol/ubiquinone pool andBrandt 2008, Quninlan et al. 2011) and the mitochondrial membrane potential (Rottenberg et al. 2009), and changes in these parameters during metabolism are mostly responsible to the variations that are observed in superoxide production by bc1 under different metabolic conditions.
The mitochondrial electron transport chain, mostly complex I (NADH dehydrogenase) and complex III (bc1), is the major source of Reactive Oxygen Species (ROS) in most animal cells. The contribution of each of these sites vary greatly depending on the metabolic states of the system (Wong et al. 2017). Beside the important role that complex III generated superoxide play in signaling (cf. Weinberg et al. 2019, Cabello-Rivera et al. 2022, Homma et al. 2021 it is well established that excess mitochondrial ROS drive the aging process and thus accelerate neurodegeneration and other aging-driven diseases, thus limiting lifespan (Rottenberg and Hoek 2017) and healthspan (Rottenberg and Hoek 2021). In anthropoid primates, and humans in particular, the evolution of complex social lifestyle most likely drove the evolution of exceptionally long lifespan, exceptionally large brain and enhanced cognitive ability, that are necessary for complex social lifestyle (Street et al. 2017). We have previously suggested that the accelerated evolution of mtDNA-coded peptides that is observed in anthropoid primates, including human, and cytochrome b, in particular, was driven by a selection pressure to reduce the rate of superoxide production by mitochondria (Rottenberg 2014). We showed that the rate of evolution of cytochrome b was corelated with exceptional longevity both in mammals and song birds (Rottenberg 2007a,b). We showed that in primates the accelerated evolution of cytochrome b is associated with increased fraction of polar residues, decreased hydrophobicity, and modulation of the electrostatic interaction of heme Fig. 2 The Q cycle In the Q cycle two molecules of ubiquinol are oxidized at the Qo site on the P-side (facing the intermembrane space) of the mitochondrial inner membrane. The first electron is accepted by the high potential chain: the [2Fe-2 S] center of ISP, heme c of cytochrome c1 and heme c of the mobile cytochrome c, a proton is also accepted by ISP thus forming a semiquinone. The second electron is accepted by heme bL and the second proton is released on the P-side. Heme bL transfer the electron to the low potential chain: to heme bH across the membrane on the N-side (facing the mitochondrial matrix) and from there to the ubiquinone at the Qi, which accept a proton from the matrix forming a semiquinone. On the second oxidation of ubiquinol at the Qo site the second electron is transferred to the semiquinone at the Qi site that accept another proton from the matrix forming a ubiquinol that can diffuse across the membrane to the Qo site (see text for further details). the Cn3D 4.3 program. The same program was used for the preparation of the figures..

Results and discussion
1. The evolution of the core peptides of the bc1 complex in anthropoid primates.
To calculate the rate of evolution of the core peptides of the bc1 complex in placental mammals we aligned the amino acids sequences of anthropoid primates (all genera of apes, and genera representing all OWM and NWM families) with sequences of genera representing all major placental mammal orders: 98 sequence of cytochrome b, 81 sequences of ISP, and 94 sequences of cytochrome c1 (see methods and supplementary materials). We estimated the relative rate of evolution from the pairwise genetic distance between the amino acids sequences of each placental mammal and a reference nonplacental mammal (Opossum, see materials and methods). Although the rate of evolution was much faster in the mtDNA-coded cytochrome b than in the nuclear-coded ISP and cytochrome c1, all three core peptides evolved significantly faster in anthropoids primates than in other mammals: cytochrome b: 0.3447 + 0.040 in anthropoids compared to 0.2593 + 0.025 in other mammals (P < 0.001); cytochrome c1: 0.153 + 0.012 and 0.117 + 0.010 in anthropoids and other mammals respectively (P < 0.001); ISP: 0.128 + 0.016 and 0.104 + 0.014 in anthropoids and other mammals respectively (P < 0.001). In cytochrome b and cytochrome c1 the rates of evolution in the three branches of anthropoid primates were not significantly different from each other. However, for ISP the rate of evolution in apes (including human) was only slightly higher than that of other mammals but not significantly different (0.115 + 0.007); only in OWM and NWM the evolution of ISP was significantly accelerated compared to other mammals (0.135 + 0.015 compared to 0.104 + 0.014, P < 0.001).
The fact that all three catalytic core peptides of bc1 evolved faster in anthropoid primates than other mammals suggests that this evolution is adaptive evolution driven by a selection pressure to modulate the function of the enzyme. It was previously reported, on the basis of smaller databases, that cytochrome b (Rottenberg 2007a) and ISP (Doan et al. 2005) evolved faster in anthropoid primates than in other mammals; here we report, for the first time, that cytochrome c1 also evolved faster in anthropoid primates than in other mammals.
To identify significant anthropoid-specific substitutions in the core peptides of mammalian bc1 complex we bH, and that these changes affect mostly the environment of Qi and heme bH. We suggested that these changes are likely to reduce the redox potential of heme bH and most likely affect the rate of superoxide production (Rottenberg 2014).
Here we show that all three core-peptide of the bc1 complex -cytochrome b, ISP, and cytochrome c1, evolved faster in anthropoid primates than in other mammals. We identified anthropoid-specific substitutions where residues that are largely conserved in mammals are substituted with residues that may affect the structure and function of bc1, and examined the structures of bovine, mouse, sheep and human bc1 in an effort to predict the effect of these substitutions. We conclude that the results are compatible with the suggestion that the accelerated evolution of the core-peptides of bc1 in humans was driven by a pressure to reduce the production of superoxide.

Methods and materials
Complex bc1 core peptides amino acids alignments The reference sequences of cytochrome b, the Fe-S protein and cytochrome c1 were obtained from the Protein collection of NCBI. The sequences were aligned by the ClustalX2.1 software. For each alignment the sequences of each of the ape genera and representative genera for all the NWM families and subfamilies of OWM were included. In addition, we included several representative genera from all major placental mammal orders (the alignments are provided in Supplementary Materials).
Rates of evolution of the core peptides of the bc1 complex The relative rate of evolution of the bc1 core peptides was estimated from the genetic distances of each placental mammal amino acids sequence and a non-placental mammal sequence, Didelphis virginiana for mtDNA-coded peptides and Monodelphis domestica for nuclear-coded peptides. The pair-wise distances were calculated with the MEGA-X software using the JTT model with gamma parameter of 2. Excel was used for statistical analysis of the results. program, and the online program, at the NCBI structure/pdb site (iCn3D). All the measured bonds and interactions up to 8 Å were verified by all Interactions tables available at the online site. Longer distances were measured directly within also form hydrophobic contact with L71(FS). These interactions of F88(FS) in the human bc1 could affect the movement of the ISP-HD from the Qo site to the heme c site. There are also two anthropoid specific substitutions in the hinge region of the Fe-S protein and four more anthropoid specific substitutions in the "vise" region of cytochrome b that interact with "hinge" residues of ISP, and quite likely affect the movement of the ISP-HD, as discussed below.

Analysis of structural data and preparation of figures
Unlike other anthropoids, the evolution of ISP was not accelerated in apes including humans. Nevertheless, there are six anthropoid-specific substitution in the human ISP: K26R, A41G, T42V, M71L, S72A and K173L. Lysine K26 is in the N-terminal on the N-side of the membrane; in the mammalian structures (e.g., [PDB: 1PPJ],[PDB: 7O3H]) it forms a hydrogen bond and a salt bridge with aspartate D258 of the core protein 1 and a hydrophobic contact with a tyrosine Y414 of core protein 1. These interactions are retained in the human structure by arginine R26. The pair of mutated residues A41 and T42 are in the single Trans Membrane Helix (TMH) of ISP. In the mammalian structures the alanine A41 is in contact with F21 of subunit 9, and T42 is in contact with L218 of cytochrome c1, these contacts are retained by glycine G41 and Valine V42 in the human structure. Thus, these three anthropoid specific substitutions do not appear to have significant effect on the structure of the human bc1 complex. The anthropoid-specific substitutions M71L in apes and S72A that are observed in both apes and OWM, are located on the "hinge" section of ISP that crosses over and is located above cytochrome b of the second monomer. In the bovine structure ([PDB:1PPJ]) there is a close contact (3.6 Å) between the serine S72(FS) hydroxyl and the Nz nitrogen of K86(c1) of cytochrome c1. However, in the human structure the smaller alanine A72(FS) is not in contact with cytochrome c1 but form close contact with tyrosine Y168(b) of cytochrome b. This cytochrome b tyrosine is also an anthropoid-specific substitution replacing a phenylalanine F168(b) in the mammalian cytochrome c sequence. These two related anthropoid-specific substitutions are a part of a cluster of cytochrome b, ISP, and cytochrome c1, anthropoid-specific substitutions, and are predicted to have a significant effect on bc1 function as discussed further below.
Finally, the ISP anthropoid-specific substitution, K173L(FS), is located in ISP-HD in the vicinity of the redox center [2Fe-2 S]. In the bovine bc1 structure ([PDB:1PPJ]), K173(FS) is < 7.5 Å from the [2Fe-2 S] center and the coulombic interactions with K173(FS) could affect the redox potential of the [2Fe-2 S] center. The substitution K173L(FS) observed in both apes and OWM is expected to modulate the redox potential of ISP, most likely affecting the rate of the first step of ubiquinol oxidation. examine the alignments of the mammalian sequences and selected those substitutions that are common to one or more of the three branches of anthropoid primates and rarely or never observed in other mammals. Although any substitution could potentially modulate the function of the enzyme, we concentrate on substitutions that result in a significant change in the character of the residue side chain, e.g., charge, polarity, volume, or hydrogen bonding capacity. We identified 20 significant anthropoid specific substitutions in cytochrome b, 15 in cytochrome c1 and 11 in ISP.
There are a few substitutions where the same site is substituted in all anthropoids, but other substitutions are specific to one or two branches of anthropoids: Apes including humans, Old World Monkeys (OWM), and New World Monkeys (NWM). In our study we concentrate mostly on the human substitutions that can be observed in the human bc1 structure ([PDB:5XTE]). To find out if the substitutions affect the structure of the enzyme, we compare the human cryo EM bc1 structure with several crystal x-ray structures of the bovine enzyme (e.g. There are 8 significant anthropoid specific substitutions in cytochrome c1 in the human bc1 complex. Two of these G213A and A221T are on the TMH but only A221T appears to affect the structure of bc1 since unlike A221 in the mammalian protein, the larger T221 is in contact with residues of other subunits -V39(FS) of ISP and Y27(S8) of subunit 8. All the other significant mutations in cytochrome c1 (S38A, Y45F, A63E, E66A, S88F and P98S) are in the protein headgroup that carry the catalytic ligand heme c, on the P-side of the membrane. Only two of these mutations appear to have a significant effect on the structure and therefore likely also the function of the complex. In the bovine structure ([PDB:1PPJ]) the serine S38 is hydrogen bonded to a water molecule and is close to heme c (< 6.0 Å). Alanine A38 in the human structure is not a polar residue and cannot bind water and therefore this substitution would lower, slightly but possibly significantly, the dielectric environment of heme c and therefore may raise its redox potential. The other significant substitution, S88F, very likely affect the function of the human bc1 complex. In the bovine structure ([PDB:1PPJ]), serine S88 is hydrogen bonded by several water molecule. S88(c1) is also in contact with the "hinge" region of ISP. In the bovine ([PDB:1PPJ]) structure, where ISP-HD is in the "b" position, S88(c1) is in contact with Alanine A70(FS) of ISP. However, in the mouse bc1 structure [PDB:7O3H], in which ISP-HD is in the "c" position, S88(c1) is in contact with M71(FS). The human substitution of serine with the very hydrophobic and much larger phenylalanine F88, that cannot bind water, also modulate the interaction with ISP. In humans, F88(c1) forms a Pi-cation bond with K90(FS) and more on the periphery, largely substitutions that interact with adjacent subunits.
2. The interaction between the "hinge" region of ISP and the "vise" regions of cytochrome b and cytochrome c1 is modified in human bc1 by anthropoid specific substitutions in all three core peptides.
ISP is composed of a single TMH that is part of one monomer of the dimeric bc1 complex and a mobile head group (HD), on the p-side of the complex, that contain the [2Fe-2 S] redox center. The HD crosses over to the second monomer and is in contact with cytochrome b of the second monomer. The "hinge" or "tether" that connects the TMH and the HD is a short partially helical segment that is tethered to a "vise" that is composed of residues of the two b cytochromes of the dimer and cytochrome c1 of the first monomer. The "Hinge" allows the rotational and translational movement of the HD, that contain the [2Fe-2 S] redox center, from the Qo site of cytochrome b (the "b" position) to the redox center of cytochrome c1, heme c (the "c" position). During the 1 electron oxidation of ubiquinone at the Qo site, histidine H161(FS) of the ISP-HD, which is hydrogen bonded with ubiquinone, also mediate the release of a proton to the P-side resulting in the formation of a semiquinone radical; a movement of the HD to the "c" position, along the cytochrome b surface There are 20 significant anthropoid-specific substitutions in cytochrome b (18 in humans). Figure 3 shows the location of all the significant anthropoid-specific substitutions in cytochrome b -substitutions that alter the properties of the residues and are not present in most other mammals. insert Figure 1 here.
In addition to the anthropoid-specific substitutions, there are several human substitutions that are not observed in the other mammalian cytochrome b structures (bovine, mouse and sheep) that can be predicted to modulate the structure and function of the complex. Moreover, more recent human evolution resulted in several regional specific cytochrome b substitutions (haplotypes) that affect the properties of the bc1 complex (Ruiz-Pesini et al. 2004;Song et al. 2016, see below). Table 1 summarizes the distribution of the cytochrome b anthropoids-specific substitutions in the three branches of anthropoids, their location in the protein, their interactions, and possible effects on bc1 function.
It is observed that most of the substitutions appears in clusters, mostly on the two surfaces of the membrane, one cluster is on the p-side, in the cytochrome b region ("vise") that interact with the "hinge" of the Fe-S protein, and another, on the N-side, near the heme bH-Qi site. There are only few substitutions in the hydrophobic core of cytochrome b, mostly close to the heme bH-Qi site, and few   enable the transfer of the electron from the [2Fe-2 S] directly to heme c that form a salt bridge with H161(FS). Mutations of residues in the "hinge" and "vise" regions of the F-S protein, and cytochrome b greatly affect the rate of ubiquinol oxidation (Borek et al. 2018;Jafari et al. 2015;Song et al. 2015), as well as the rate of superoxide generation (Borek et al. 2015, Ekeirt et al. 2016, Pagacz et al. 2021. In highresolution structures of the bovine bc1, many with bound ligands or inhibitors, the ISP-HD is found in various locations along the path of this movement from the "b" position in the Qo site to the "c" position near heme c of cytochrome c1 (Berry et al. 2013). The interactions between the "hinge" and the "vise" are largely unchanged except for subtle changes in the many hydrophobic interaction between these regions. Most of the residues in these regions are highly conserved (> 95%) in all form of bc1 from bacteria to animals. The amino acids sequence of the ISP "hinge" region (V59 ( Table 1). The two cytochrome C1 residues that interact with the "hinge", R49(c1) that is hydrogen bonded to D67(FS) and S88(c1) that contact A70(FS) are conserved in all mammals except anthropoid primates where the substitution S88F(c1) is observed in apes and OWM. There are subtle difference in the hydrophobic interaction between these regions observed by X-ray diffraction of the numerous crystal structure of the bovine bc1 dimer (e.g.  complex E162(b) does not interact with the ISP. Of particular significance is the human S72A(FS) substitution in the ISP. The introduction of alanine(s) in this position was shown to enable the movement of the ISP in R. capsulatus mutants that inhibit the movement of ISP-HD (Borek et al. 2015). Similarly, experiments with the yeast bc1 demonstrated that the activity of cytochrome b mutants of the Qi site that inhibit the complex activity could be restored by compensating mutations in the hinge region of ISP (Song et al. 2015). Figure 4 show the position of the human-specific substitutions in the "vise" and "hinge" regions of human bc1. insert figure 4 here. Also, in the human bc1, several residues in the "hinge" region of ISP and W163(b) in cytochrome b are in contact with a cardiolipin which is not observed in this position in the mouse, sheep and cow bc1 (see below) and may also affect the movement of ISP-HD.
In humans, mutations of conserved cytochrome b residues in the "vise" regions (G166E(b), D171N(b)) are associated with mitochondrial diseases (Table 2). More remote human cytochrome b pathogenic mutations -S151P(b), G251D(b), and G290D(b) were modeled in the bacterial bc1 (R. capsulatus) and were shown to affect the position of ISP head group, inhibit the enzyme and increased superoxide production as described in Table 2. Moreover, human haplotypes carrying the beneficial substitutions -I153T(b) and N260D(b) -that were also modeled in R. capsulatus were also shown to modulate the interaction between cytochrome b and ISP-HD, accelerating electron transport and inhibiting superoxide production (Pagacz et al. 2021). It is apparent that mutations that modulate the interaction between cytochrome b and ISP can either enhance superoxide production, as was observed with the pathogenic mutations, or inhibit superoxide production, as observed in some human haploid groups. Therefore, the clustering of anthropoidspecific substitutions, in all three catalytic protein of bc1, at structures there is a strong salt bridge K172(b)-D58(b) that affect the position of "vise" sequence. This salt bridge does not exist in the human structure because of the substitution K172S(b). All the other interactions between the "hinge" and "vise" are hydrophobic-there are ~ 10 residue pair contacts (< 4 Å) between these two sequences both in the mouse, sheep and cow structures but the identities of these pairs are different in nearly every structure. It thus appears that the hydrophobic contacts between the hinge and vise regions change during the movement of the ISP-HD. In the human sequences of these interacting regions there are 7 substitutions that may affect these interactions: E162Q(b), F168Y(b), K172N(b), and S173A(b) in cytochrome b, M71L(FS) and S72A(FS) in the F-S protein, and S88F(c1) in cytochrome c1. In human bc1, F88(c1) form a pi-cation bond with K90(FS) and contact L71(FS). In the human cryo EM bc1 structure ([PDB:5XTE]) the cytochrome b substitutions appear to have a subtle but significant effect on the interactions between the "hinge and "vise'' segments. Interestingly, these interactions are not identical in the two monomers. While the W163(b)-V59(FS) bond that is observed in the cow and mouse structure, is conserved in the human structure in both monomers, there is a hydrogen bond R177(b) -S63(RS) in addition to the R177(b)-M62(RS) in one monomer, while in the other monomer there is additional hydrogen bond formed by the anthropoid-specific Y168(b) with L69(FS). Both monomers in the human structure show a pi-cation bond Y168(b)-R92(FS) similar to the pi-cation bond observed in the mouse structure F168(b)-R92(FS). Most significantly the two anthropoid primates substituted residues Y168(b) andA72(FS) strongly interact with each other (3.6 Å). Both monomers of the human bc1 show increased interactions between the hinge and vise section most likely as the result of the many substitutions in this region. For example, the anthropoid specific substitution Q162 (b) interact with S63(FS) but in the mammalian Fig. 4 The cluster of seven anthropoid specific substitutions in the human bc1 at the site of interactions between the "hinge" of the Fe-S protein and the "vise" (cytochrome b and cytochrome c1). The Fe-S protein "hinge" is in green, the peptide on the left is cytochrome c1 and the peptide on the right is cytochrome b. The anthropoid-specific substitutions form 4 salt bridges with arginine R80-4.4 Å average distance from propionate A and 4.9 Å average distance from propionate D. There is also a weak salt bridge of propionate D with histidine H83(b) (5.6 Å) but there are no other electrostatically interacting charges with heme bL propionates. Hydrogen bonding with the propionates also raises the heme redox potential (Kuleta et al. 2021) but there is only one hydrogen bond, propionate A with R80, in heme bL in the bovine structure. In contrast the propionates of heme bH form much stronger salt bridges with arginine R100 (3.0 Å average distance from propionate A and 4.5 Å average distance from propionate D). There are also four strong salt bridges between H97 and propionate A (4.7 Å average distance). In addition, arginine R313 is in close electrostatic interaction with the D propionate of heme bH, 4.8-6.4 Å distance in various bovine structures, and 6.0 Å in the mouse structure. Moreover, there are several hydrogen bonds with the propionates of heme bH in the bovine structure: W30 and N206 with propionate D, and G34 (backbone oxygen), and R100 with propionate A. This difference of the electrostatic interactions of arginines and histidines with propionates in cytochromes bL and bH, and the difference in the propionates' hydrogen bonding is a major contributor to the very large difference in redox potentials between bL and bH (~ 160mV). Mutations that affect the electrostatic interactions of cytochrome b hemes could have great effect on electron transport, proton pumping and superoxide production (cf. Qu et al. 2013).
The most significant substitution in the human bc1 complex is probably the cytochrome b substitution R313Q(b). The elimination of the coulombic interaction between R313(b) and propionate D is expected to lower the redox the "vise"-"hinge" interacting region, strongly suggest that there was a strong evolutionary pressure in anthropoids to modulate the interaction between cytochrome b and ISP to modulate ISP-HD movement and possibly the production of superoxide. It is impossible to predict the effect of each one of the anthropoid-specific substitutions on the movement of ISP-HD or on the generation of superoxide. Most likely some substitutions on their own have very strong effects which are antagonized by other substitutions so that the net effect is relatively subtle (Borek et al. 2018). While it is not possible to predict the net result of these substitutions, the fact that the evolution of the human bc1 resulted in a clustering of seven substitutions, in all three catalytic subunit of the enzyme at the "vise-hinge' contact sites, that are critical for the movement of the ISP-HD, and the generation of superoxide, suggests that there was a very strong evolutionary pressure to modulate the movement ISP-HD in anthropoid primates, and since this may affect superoxide production this finding is compatible with the hypothesis that evolution of the bc1 complex in anthropoid primates was driven by a pressure to reduce the rate of generation of superoxide.

A cluster of anthropoid specific substitutions near cytochrome bH and Qi probably affect its redox potential and ubiquinone reduction in the Qi site.
It is known that the electrostatic interaction within hemes of the positively charged Fe with the negatively charged propionates are a major determinate of the heme redox potentials (Das and Medhi 1998). In the bovine bc1 structure ([PDB:1PPJ]) cytochrome bL propionates A and B hydrogen bonded to heme bH, and glutamate E111 interact electrostatically with heme bH (< 7.5 Å) and the change of charge of this residue should affect the redox potential of heme bH. Moreover, in the mammalian structure E111 form a salt bridge with histidine H8(b2) of the second cytochrome b monomer, which would affect the charge of E111(b) and thus the redox potential of heme bH. This salt bridge is lost in the human cytochrome b because of the substitution H8N (Table 1) 194(b)). This cluster of human substitutions around heme bH and Qi is expected to increase the dielectric constant around heme bH and ubiquinone at the Qi site (Fig. 5). Indeed, substitution of residues near the Qi site were shown to affect the reduction of ubiquinone (Rotsaert 2008) and it is therefore possible that these human substitutions modulate ubiquinone reduction in the Qi site. This suggestion is compatible with the fact that the human allele T194M(b), which is a reversal to the more hydrophobic mammalian consensus, was found to be associated with slower rates of metabolism (Tranah et al. 2012). In addition to the anthropoid-specific substitutions in conserved mammalian residues there are other human variants (haplotypes) of residues in the heme bH -Qi region of cytochrome b potential of human heme bH significantly (Rottenberg 2014). Moreover, there are other significant changes in heme bH propionates interactions between the human and the bovine structure. Instead of the hydrogen bond with the backbone oxygen of W30, that is observed in the bovine structure, propionate D form a hydrogen bond with the sidechain NE1 of W31 in the human bc1 structure. Moreover, the hydrogen bond between R100 and Propionate A, that is observed in the bovine structure, is missing in the human structure which could lower heme bH redox potential further.
Although R313(b) is not conserved in all forms of bc1 it is highly conserved in mammals and is retained in many other forms of bc1, including some Fungi and bacteria. In mammals, the substitution R313Q(b) is observed only in primates and one genus of bats. In addition to R313Q(b) there are 3 more anthropoid-specific substitutions in the human cytochrome b that are within 6.5 Å of heme bH: I/V39A(b), Y107F(b) and V123T(b). Alanine 39 A(b) and threonine 123T(b), which are in the hydrophobic core of human cytochrome b, replace hydrophobic residues with hydrophilic residues and thus expected to increase the dielectric constant around heme bH, lowering its redox potential (Popović et al. 2001). In particular, valine V123(b) is very close to both hemes (3.9-4.4 Å) and the substitution V123T(b) may affect the coulombic interaction between these hemes. The human substitution Y107F eliminate a very strong hydrogen bond between Y107(b) and H308(b) which is observed both in the bovine and mouse structures. In NWM there are 2 substitution that may affect heme bH -N206S and E111N/K (also observed in some OWM). The asparagine N206 is Fig. 5 The cluster of residues in bovine cytochrome b near heme bH and Qi that are substituted in anthropoids. The figure is drawn from [PDB:1NTZ] showing heme bH, ubiquinone and residues within 6 Å of these sites that are substituted in anthropoid primates. of the Ubiquinone Binding Protein. E344(b) form a hydrogen bond and a salt bridge with lysine K70 (UBP) while histidine H345(b) form hydrogen bond and a salt bridge with E65(UQP). In Human bc1 Y345(b) forms strong hydrogen bonds with both E65(UBP) and S69(UBP), but S344(b) form no bonds with the ubiquinone binding protein except for contact with F67(UBP). Aspartate D214(b), which is located on the D-E loop on the N-side (Fig. 3b), forms a very strong hydrogen bond and a salt bridge with the Ubiquinone Binding Protein arginine R2(UBP) located at the N-terminal tail on the N-side of cytochrome b. In contrast in the human bc1 histidine H214(b) form no bond with the ubiquinone binding protein. The function of the Ubiquinone Binding Protein is not known so it is not possible to predict the effect of these anthropoid-specific substitutions on bc1 function. However, a Ubiquinone Binding Protein arginine, R40(UBP), anchor the two cardiolipin that also form a salt bridge with the cytochrome b lysine K227(b) at the ubiquinone reduction site, Qi, and most likely play a role in proton access to ubiquinone (see below).
There are two anthropoid-specific substitutions in the N-terminal of cytochrome b on the N-side-N3P(b) and H8N(b) (Fig. 3b). The N-terminal of cytochrome b interact mostly with core protein 1, but also other subunits including the cytochrome b of the second monomer. These interactions are different in the various bovine, mouse and human structures so it is difficult to predict the contribution of the anthropoid substitutions to these variations in structure. Nevertheless, it is clear that some difference could be the result of these substitutions. For example, asparagine N3(b) form hydrogen bond with aspartate D333(CP1) in the mouse structure but the proline P3(b) in the human cytochrome b cannot form hydrogen bonds. Similarly, histidine H8(b) form a slat bridge with E111(b2) on the second monomer in a bovine structure [PDB:1NTZ] but this salt bridge cannot exist in humans because of the substitution H8N(b). Alanine A180(b) also interact that could affect the reduction of Ubiquinone at the Qi site. A number of these variants in the bH-Qi region, that are associated with diseases, where shown to affect the biological properties of the bc1 complex by studies of homologous mutations in yeast (Song et al. 2016). Example of these mutations are described in Table 3. In contrast to the numerous anthropoid-specific substitutions in the vicinity of the bH-Qi site there are no anthropoid-specific substitutions close to the heme bL-Qo site, other than V123T(b) that is situated between bL and bH.
We have previously reported that the amino-acid composition of cytochrome b is significantly different between anthropoid primates and other mammals. Anthropoid's cytochrome b is much less hydrophobic than that of other mammals, with particular enrichment with the hydrogen-bonding residues threonine and serine (Rottenberg 2014). For example, while bovine cytochrome b has 22 serines, human cytochrome b has 29 serines. Similarly bovine cytochrome b has 27 threonines, while human has 31. This change may render cytochrome b more permeable to protons which may further reduce the magnitude of the protonmotive force across bc1 and hence inhibit superoxide production. It was recently reported that the fatty acid docosahexaenoic acid intrinsically uncouples the bc1 complex (Semenova et al. 2021) which suggest that there is a pathway for fatty acid mediated proton leak within the bc1 complex. This pathway may be enhanced by the enrichment of hydrogen bonding residue in the human cytochrome b.

Anthropoids-specific substitutions in cytochrome b
modulate its interactions with other subunits of the bc1 complex.
Three anthropoid-specific substitutions in human cytochrome b alter the interactions between the Ubiquinone Binding Protein, UQCRQ, (UBP) and cytochrome b -E344S(b), H345Y(b) and D214H(b). The pair glutamate E344(b), and Histidine H345(b), which are located on the G-H loop on the P-side (Fig. 3b), interact with the P-side end of the TMH ----------Increase sensitivity to CLOM, decrease activity Song et al. 2016Fauser et al. 2002Song et al. 2016Andreu et al. 2000Schuelke et al. 2002Tranah et al. 2012 form another hydrogen bond with UBP (T41(UBP)) and with cytochrome b serine S29(b). This serine is substituted in humans, S29A(b), but the backbone bond remains. This cardiolipin also form a salt bridge with K227(b) but not with the charges of cytochrome c1. Nevertheless, the interaction of the second cardiolipin with K227(b) suggests that it also affects proton transfer to the ubiquinone in the Qi site. In the bovine bc1, leucine L235(b) is in contact with cardiolipin. In humans, the substitution L235F(b) eliminate this contact. A third cardiolipin on N-side surface of the membrane is observed in several bovine bc1 structures as well as the mouse and human structure, but it does not appear to be associated with the Qi site. This cardiolipin is anchored by hydrogen bonds with residues of core protein 1 and the N-tail of cytochrome b as well as salt bridges of arginines on core protein 1 (R445, R436(cp1) and R5(b) on cytochrome b. In addition, it forms a salt bridge with H221(b) on the N-side of TMH E.
In addition to the six ubiquitous mammalian cardiolipins (three in each monomer), described above, there are three more cardiolipins observed in the human bc1 structure (Fig. 6a), one of them is also located on the N-side.insert Figure 6 here. This single cardiolipin is shared between the two cytochrome b monomers, and is located in the clef between the two cytochrome b monomers on the N-side of the protein. In the human bc1 structure one phosphate of the cardiolipin is anchored by salt bridges with two cytochrome b histidines, H16(b) and H201(b), of one monomer, and the second phosphate is anchored by salt bridges with H16(b) and H201(b) of the second monomer (Fig. 6b). The structure is completely symmetric, as all the interactions of one phosphate with residues of one cytochrome b monomer is mirrored with interactions of the second phosphate with the same residues of the second cytochrome b monomer. The only other bc1 structure that shows a cardiolipin is this position is the sheep structure. However, in the sheep as well as the bovine structure residue 16 is asparagine not histidine so this salt bridges cannot exist in the sheep structure or in most other mammals. Histidine H201(b) is located within the Qi site and was proposed to be the second proton donor for the reduction of ubiquinone at the Qi site (Rotsaert et al 2008, Pintscher et al 2020. A very strong hydrogen bond (2.5 Å) of histidine H201(b) with the natural ubiquinone10, is observed in the sheep bc1 structure. The salt bridges between the cardiolipin phosphates and the H201(b) residues of the two monomers may provide proton access to this donor site in both monomers, and the existence of additional salt bridges between these phosphates and H16(b) residues in the human bc1 may therefore affect the protonation of H201(b). It is also possible that the bridge that this cardiolipin provide between the two histidine H201(b) of the two monomers provide a pathway for proton sharing between with cytochrome b of the second monomer in the bovine structures ([PDB:1PPJ], [PDB: 1NTZ]) forming a strong contact with F183(b2) and the human substitution A180T eliminates this contact but retain the weaker contacts with residues A52(b2) and M953(b). Therefore, it appears that anthropoid-specific substitutions also change the interactions between the two cytochrome b monomers in the dimeric bc1 complex. Another anthropoid-specific substitution that modify the interaction of cytochrome b with other bc1 subunits is the cysteine C70T(b) substitution. C70(b) is located on helix ab (part of the AB loop) on the p-side (Fig. 3b). While C70(b) only interact with water in the bovine structure (PDB:1PPJ), threonine T70(b) in human cytochrome b form a strong hydrogen bond with tyrosine Y199(c1) of cytochrome c1. Other residues that interact heavily with water on the AB loop, S57(b) and T59(b), are substituted in humans with residues that cannot form hydrogen bonds with water S57P(b) and T59A(b) (Fig. 3b). Thus, three anthropoid-specific substitution on the AB loop reduce the interaction of this loop with water molecules.

Human-specific interactions with cardiolipin molecules may affect proton access to the ubiquinone reduction site Qi.
Cardiolipin was shown to be essential for the activity of the bc1 complex . All the structures of mammalian bc1 that contain phospholipids show cardiolipin molecules interacting with cytochrome b. The number of cardiolipin molecules resolved in the various mammalian bc1 structures vary from 2 (per monomer) to 5. All the mammalian bc1 structures that contain phospholipids show at least two adjacent cardiolipin molecules on the N-side of the membrane near the Qi site. Both molecules are anchored by strong hydrogen bonds and salt bridges to an arginine, R40(UBP) on the Ubiquinone Binding Protein (UBP). One of these cardiolipins also form another hydrogen bond with UBP (N36(UBP)), and a hydrogen bond with cytochrome C1 (Y220). Most significantly this cardiolipin form a salt bridge with a cytochrome b lysine, K227(b), at the Qi binding site, as well as K223(c1), R224(c1) and K231(c1) of cytochrome c1, near K227(b) in the Qi binding site. K227(b) and D228(b) were suggested to be the proton doners to ubiquinone during ubiquinone reduction , Song et al 2020 and the adjacent charged residues on cytochrome c1 were suggested to provide proton access to Qi (Klingen et al 2007). There is some evidence for cardiolipin participation in the reduction of ubiquinone at the Qi site . The second ubiquitous cardiolipin that is anchored by R40(UBP)also

Conclusion
The evolution of the three catalytic core peptides of the bc1 complex was accelerated in anthropoid primates including humans. Most of the anthropoid specific substitutions involve residues that are conserved in most other mammals and result in a significant change in the physical characteristic of the residue and are therefore expected to affect the structure and modulate the function of the bc1 complex. Moreover, many of these substitutions are clustered around highly conserved regions of the complex that are known to be critical for the complex function. We identified two such clusters, one is on the p-side of the complex that can be expected to affect the movement of ISP-HD between the "b" position at the Qo quinol binding site in cytochrome b and the "c" position near heme c on cytochrome c1 and therefore modulate the rate of ubiquinol oxidation. This cluster of seven human substitutions in the interacting "hinge" (ISP) and "vise" (cytochrome b and cytochrome c1) regions, four the two Qi sites in the bc1 dimer, similar to the electron sharing between the two heme bL that apparently play important role in the efficient function of bc1 . Asparagine N16(b) is not conserved in mammals, but most mammals could not form a salt bridge between cardiolipin and residue 16(b). Histidine, H16(b), is observed in all apes and most OWM, but also in several other species, including mice. Even if all mammals bind cardiolipin at this position, the effect of cardiolipin on the protonation of H201(b) could depend on the identity of residue 16(b). In some OWM species there is arginine in this position, R16(b), that can also form a salt bridge with cardiolipin. In the human structure there is an additional cardiolipin that is located on the P-side of the membrane, and is associated with the ISP, as described above.  figure); the two adjacent inner molecules are anchored by the Ubiquinone Binding Protein and interact with the Qi sites in cytochrome b. In each monomer there is additional cardiolipin on the p-side that interact mostly with the "hing" of the Fe-S protein. One additional cardiolipin is bound on the N-side in the clef between the two cytochrome b monomers (see below) 6b: A cardiolipin molecule connects the Qi sites of the two cytochrome b monomers through salt bridges with H201 (top on left corner and bottom on the right corner). Additional salt bridges with H16 in the human cytochrome b may modulate this interaction on the observation that the rates of cytochrome b evolution in anthropoid primates (particularly NWM) is corelated with exceptional longevity (Rottenberg 2007a(Rottenberg , 2014 and that longevity in mammals is corelated with the rate of production of superoxide (Lambert et al. 2007). Anthropoid primate evolution is driven by a selective pressure to enable complex sociality (Street et al. 2017). Reduction in the rate of ROS production, that delays aging and protects the brain from neurodegeneration Hoek 2017, 2021), would enable the evolution of complex sociality that require a longer lifespan and enhanced cognitive abilities (Street et al. 2017). A recent study of the accelerated evolution of human cytochrome c oxidase (Rottenberg 2022) that suggests selection for reduction in the protonmotive force, and thus selection for reduced generation of superoxide (that depend on the magnitude of the protonmotive force (Rottenberg et al. 2009)) is also compatible with the hypothesis.

Author's contribution Not applicable.
Funding No funding was received for conducting this study.
Availability of data and materials Data are available from the author on request.
Declarations in cytochrome b (E162Q, F168Y, K172S and A173P), one in cytochrome c1 (S88F) and two in ISP (M71L, S72A), suggest a selection pressure to modulate the rate of ubiquinol oxidation at the Qo site. Another human substitution in ISP, K173L, is predicted to reduce the redox potential of the [2Fe-2 S] redox center and modulate the rate of oxidation of ubiquinol by the human bc1 complex. Another cluster of human substitutions in cytochrome b surrounding heme bH and Qi is expected to modify the redox potentials of heme bH and ubiquinone and modulate the rate of proton uptake during ubiquinone reduction at the Qi site, suggesting a selection pressure to modulate the reduction of ubiquinone at site Qi. Moreover, the specific interactions of additional cardiolipin with critical Qi residues in the human bc1 further suggest that ubiquinone reduction is modulated in the human bc1. While it is very likely that the modulation of the rates of both ubiquinol oxidation and ubiquinol reduction will be associated with modulation of the rate of superoxide production it is not possible to predict the direction or extent that of the effects on the turnover rate of the human bc1, or the rate of production of superoxide. Only direct measurements of superoxide production by bc1 in human compared to other mammals (e.g., mouse, cow) could reveal if there is a significant difference between human and other mammals. There is plenty of evidence, from studies of hydrogen peroxide generation by isolated mitochondria, that there are big differences in superoxide production by the bc1 (complex III) of different species (cf. Herrero andBarja 1997, Lambert et al. 2007) but there are no data of such comparative studies with human mitochondria. More accurately the rate of superoxide production should be examined with the isolated and reconstituted bc1 complex (Rottenberg et al. 2009). Comparative studies with the isolated bc1 complex from humans and other mammalian species could also test the predicted differences in the rate of ubiquinol oxidation as well as the rate of ubiquinone reduction. Another possible experiments to test whether the substitution of hydrophobic residues in the Qi site in human bc1 resulted in modulation of ubiquinone reduction and superoxide production is by studies with cybrids. Since it has been shown that humans carrying the cytochrome b substitution A194M (M194(b) is common to most mammals) exhibit low metabolic rates (Tranah et al. 2012) it can be predicted that the rate of electron transport in cybrid containing the mtDNA allele M194(b) would be slower than wild type. Comparison of superoxide generation between the cybrid and wild type could test the hypothesis that the human substitutions in the Qi site were selected to lower production of superoxide.
The results of this study are compatible with the hypothesis that the accelerated evolution of bc1 in anthropoid primates is driven by a selection pressure to reduce the rate of ROS production (Rottenberg 2014). This hypothesis is based