Supercomplex supercomplexes: Raison d’etre and functional significance of supramolecular organization in oxidative phosphorylation

Competition between succinate anions and the acidic uncouplers of OXPHOS

The inhibition of uptake/oxidation of succinate in mitochondria by anions of various acidic uncouplers of OXPHOS is an important phenomenon in cellular metabolism. Figures 1 and 2 present the results of experiments that strongly support such competition between anions. Figure 1 shows the kinetically pure competitive inhibition observed between succinate (1–6.7 mM) and the uncoupler FCCP (0–10 μM). Figure 2 illustrates the competition process between substrate and uncoupler anions for succinate (0.5–6.7 mM) and dicoumarol uncoupler (0–50 μM). Similar competitive behavior between succinate (0.5–5 mM) and DNP anions in the range from 0 to 1.5 mM has been reported previously, along with a determination of the enzymological parameters K _m , V _max, and K _i of the inhibition process [78].

Why should several anionic uncouplers compete with anionic substrates for entry/conversion in mitochondria? The competition phenomenon (Figures 1 and 2 and ref. 78) is not readily accounted for by Mitchell’s single-ion chemiosmotic theory, but is naturally explained by Nath’s two-ion theory of energy coupling and ATP synthesis [36,38,39,40,42,49,50,51,59,63,64,65,70,78]. The inhibition phenomenon (Figures 1 and 2) can occur either by inhibition of the entry of succinate or be due to the inhibition of the oxidation of succinate (Figure 3). If the inhibition is caused by the former process, then the intramitochondrial concentration of succinate would be expected to decrease, whereas if the inhibition is due to the latter, then the process would show an increase in the internal concentration of succinate in mitochondria (Figure 3).

Figure 3

Possible causes of succinate inhibition by uncoupler in mitochondria. The inhibition could occur either due to inhibition of the succinate entry process (1, green arrow) or by inhibition of the succinate oxidation process (2, red arrow).

Inhibition of succinate entry or succinate oxidation?

The competition between uncoupler and substrate anions shown by the rate experiments in Figures 1 and 2 can be demonstrated directly by following the internal concentration of labeled substrates. If the inhibition by uncouplers is accompanied by a decrease in the intramitochondrial succinate concentration, then the inhibition process can be attributed to the inhibition of the entry of substrate by uncoupler. If an increase in the internal succinate concentration is found, then inhibition of the oxidation process would be operative. The two possibilities are illustrated diagrammatically in Figure 3.

Figure 4 shows the experimental results that uncouplers strongly inhibit the uptake of succinate in the presence of respiratory inhibitors and oligomycin, that is, when both respiration and ATP energy are cut off. Similar behavior is found in the presence of respiration (Figure 5). The results of Figure 5 clearly show that the intramitochondrial concentration of succinate is progressively decreased by an increase in the concentration of DNP uncoupler. Since a decrease in the internal succinate concentration is found (Figures 4 and 5), it is concluded that inhibition of entry of succinate by uncoupler (Figure 3) has been demonstrated.

Figure 4

Effect of increasing FCCP uncoupler concentrations from 0 to 1 μM on succinate uptake by rat liver mitochondria. Inhibition of succinate uptake by FCCP in the presence of respiratory inhibitors rotenone, antimycin, and oligomycin.

Figure 5

Influence of increasing concentrations of DNP uncoupler on succinate uptake. Inhibition of the rate of entry of substrate by uncoupler, as shown by the progressive decrease in intramitochondrial succinate concentration upon increase in DNP concentration from 0 to 1 mM during respiration.

A detailed molecular mechanism of coupling of ion transport to ATP synthesis [81] and uncoupling of ion transport from ATP synthesis [82] has been developed based on these concepts by the two-ion theory [21,70]. Permeability of the energy-transducing membrane to both the uncoupler acid and the uncoupler anion is necessary [82], and according to the mechanism, the uncoupler is translocated through the half-access channels in the membrane in its anionic form exactly like a substrate anion. Once inside, the anion picks up a proton and leaves the mitochondrion in its neutral form by diffusion [82]. A detailed quantitative kinetic model of the process of coupling and uncoupling has been formulated [81,82]. The mathematical equations used to explain coupling and uncoupling in the framework are the same, except for the imposition of the respective conditions during coupling/uncoupling (e.g., translocation as separate H⁺ and succinate A^– ions through the access channels during the process of coupling, as opposed to the formation of the neutral acid form UH due to the lipid solubility of the uncoupler anion U^–, which results in uncoupling in OXPHOS) [81,82].

It has also been pointed out by us that Mitchell’s single-ion chemiosmotic theory [83,84] is forced to postulate active transport of the uncoupler U^– anion in order to explain the dissipation of Δψ during uncoupling [82]. Where is the energy for this active anionic transport coming from? This is not addressed by the chemiosmotic theory [83,84]. Above all, there are major contradictions and inconsistencies in the chemiosmotic theory that, as one of its central tenets, postulates H⁺ as the sole ion involved in coupling oxidation and ATP synthesis in OXPHOS [21,45,50,51,56,59,61,83,84,85,86,87], because the theory is forced to fall back on anion transport to explain uncoupling [82], yet it envisages no role for the anion in the process of energy coupling [83,84,85]! These difficulties are overcome, and the longstanding inconsistencies are satisfactorily resolved by the two-ion theory of energy coupling/uncoupling involving succinate/uncoupler anions and protons [36,38−40,42,43,63–70,78–82].

The work has many other important biological implications. Uncoupler anions U⁻ uncouple not by a general ability to ferry protons and act as proton conductors, as believed by some workers, but by binding to a specific site at the a–c interface of FO and forming a neutral UH species, thereby collapsing both driving forces Δ (pH) and Δ(ψ). This drive to form the undissociated species is not operative in the case of the physiological anion succinate and H⁺. It is also known that K⁺ (or Na⁺) can substitute for H⁺ in ATP synthesis by certain organisms. In such a situation, succinate and K⁺ or an “uncoupling” anion U⁻ (for example, FCCP anion) and K⁺ shall translocate as individual (i.e. separate) ions through the access half-channels in the FO portion, and thereby enable energy coupling in ATP synthase. In other words, one can also have anion substitutes, depending on the context.

The classical chemical uncouplers of OXPHOS exert two effects on respiration. The results shown in Figures 1–5 have uncovered the first of these effects, for example, by competitive inhibition studies (Figures 1 and 2) and by the lowering of the internal substrate concentration by investigation of the direct action of uncouplers on the entry of substrate and succinate accumulation (Figures 4 and 5). However, uncouplers can also affect respiration directly by interacting with the Complexes of the electron transport chain in mitochondria (Section 3.3).

Network of energy transfer pathways in OXPHOS: the site-selective, energy-linked inhibition by guanidine derivatives, and the differential release of their inhibition by OXPHOS uncouplers

The pioneering pharmacological work of Gunnar Hollunger at the University of Lund in the mid-1950s was the first to describe the site-selective inhibition of mitochondria by guanidine and its derivatives [71,72]. Hollunger showed that these agents induced a preferential inhibition at Site 1 of the mitochondrial respiratory chain; the energy-linked character of this inhibition was clear from the reversal of the inhibition by uncouplers [71,72]. Hollunger’s visionary research has been largely forgotten, although it led to the development of the FDA-approved drug metformin, the gold standard for the management of type 2 diabetes since 1960 (Section 3.8.1). Metformin (or 1,1-dimethylbiguanide) inhibits oxidation of the reduced cofactor nicotinamide adenine dinucleotide (NADH) in preference to that of succinate and flavin adenine dinucleotide-linked carbon compounds.

The differential effect of octylguanidine on systems respiring on succinate and glutamate–malate substrates is listed in Table 1. Low (50 μM) concentrations of octylguanidine are effective in the inhibition of glutamate–malate system in state 3, compared to succinate as substrate (Table 1). At higher concentration levels (>200 μM), inhibition of glutamate-malate reaches ∼100% completion; however, the maximum inhibition of succinate oxidation only approaches ∼60%. Similar results were obtained with guanidine; however, the maximal concentrations of the inhibitor used were on the order of 100 mM.

The results of Table 1 confirm the original findings [71,72] that the alkylguanidines are more effective in inhibiting respiration on NADH-linked substrates than succinate; thus, they preferentially inhibit the first energy conservation site or Complex I. On the other hand, the related pharmacological agent, phenethylbiguanide, preferentially inhibits energy transfer with succinate as substrate; that is, it inhibits energy conservation at Site 2. Furthermore, we found Site 2 selectivity of biguanides only over a narrow structural range and side chain length of the biguanide derivatives. Thus, butylbiguanides and pentylbiguanides, like their phenethylbiguanide cousin, exhibited Site 2 selectivity; however, derivatives with shorter or longer side chains showed the more common Site 1 selectivity and inhibition of Complex I. Moreover, uncouplers themselves show site-selective effects. Thus, the results listed in Table 2 reveal that uncouplers such as dicoumarol are completely ineffective in releasing the respiratory inhibition induced by octylguanidine. However, dicoumarol has the ability to fully release inhibition due to phenethylbiguanide (Table 2). On the other hand, uncoupling agents such as DNP can release the inhibition induced by both types of guanidines, that is, octylguanidine and phenethylbiguanide (Table 2).

The results of inhibition by the guanidine derivatives (Tables 1 and 2) reveal the site-specificity of two types of uncoupling agents, one such as dicoumarol that relieves the inhibition by Site 2-selective phenethylbiguanide but not the inhibition induced by the Site 1-selective alkylguanidines, and another like DNP that release inhibition effected by both classes of guanidines. The results cannot be reconciled with the chemiosmotic theory that predicts only a single energy pool between respiration and ATP synthesis.

The observations lead to a more complex scheme involving separate and distinct energy transfer pathways between Site 1 and Site 2 in the mitochondrial respiratory chain and the intermediate that energizes ATP synthesis by Complex V (Section 3.4 and Figure 6).

Figure 6

New model of the network of energy transfer pathways between the respiratory chain (Complexes I–IV) and ATP synthase (Complex V) based on the results of this work. Guanidine and the alkylguanidines induce a preferential inhibition of energy conservation at Site 1 (Complex I), shown by the dashed red lines with downward arrows in the upper limb. On the other hand, phenethylbiguanide shows Site 2 selectivity and reveals a predilection for energy-linked inhibition of energy conservation and transfer at Site 2 (Complexes II–III), drawn in the diagram with downward pointing dashed red arrows in the lower limb. Furthermore, uncouplers are intrinsically site-specific too. Thus, DNP effectively reverses the respiratory inhibition caused by either Site 1- or Site 2-selective guanidine derivatives (shown by bold green upward pointing arrows in the two limbs). However, dicoumarol has the ability to release only the inhibition induced by phenethylbiguanide, but not by the alkylguanidines, delineated by the green arrow in the lower limb (Tables 1 and 2). Thus, the guanidine derivatives specifically interact at different sites in the respiratory energy transfer pathways in mitochondria. To sum up, the network of energy transfer pathways in the mitochondrial respiratory chain requires discrete points of interaction with two classes of guanidine derivatives and uncoupling agents.

New model of network of energy transfer pathways between the respiratory chain (Complexes I–IV) and ATP synthase (Complex V)

As discussed in Section 3.3, the experimental observation on the site-specific inhibition unleashed by the guanidine derivatives, and the differential restoration of respiration induced by two classes of uncoupling agents necessitates a new model of interaction in mitochondrial energy transfer. This model is illustrated by the scheme shown in Figure 6. The preferential inhibition of energy conservation at Site 1 (Complex I) induced by guanidine and octylguanidine is shown by the dashed red lines with downward arrows in the upper limb of Figure 6. The energy-linked inhibition of energy conservation and transfer at Site 2 (Complexes II–III) is illustrated by the downward pointing dashed red arrows in the lower limb of Figure 6. The relief by uncouplers of the respiratory inhibition caused by either Site 1- or Site 2-selective guanidine derivatives is shown by bold green upward-pointing arrows in the two limbs (Figure 6). Thus, two energy transfer pathways between electron transport and ATP synthesis are required for a more complete understanding of the OXPHOS system (Figure 6).

The internal structure of mitochondria

The pioneering work of Lea and Hollenberg had visualized the internal structure of liver mitochondria (Figure 7) using high-resolution scanning electron microscopy [73]. The tubular cristae and lumen are clearly seen in the structure (Figure 7). Higher-resolution visualization of the internal structure was made possible by Frey, Perkins, and Mannella by using powerful three-dimensional (3D) electron microscope (EM) tomography techniques [74,75,76] (Figure 8).

Figure 7

Internal structure of liver mitochondria in rat hepatocytes based on the pioneer high-resolution scanning electron microscopy work of Lea and Hollenberg, 1989 [73] that clearly revealed tubular cristae. O, outer membrane; I, inner membrane; T, tubular cristae (arrows); L, lumens within tubular cristae. Modified with permission.

Figure 8

Internal structure of mitochondria in rat liver and chick cerebellum based on 3D EM tomography work of Frey, Mannella, and Perkins [74,75,76,77], reproduced with permission. (a) The model showing all cristae in yellow, the inner mitochondrial membrane in light blue, and the outer mitochondrial membrane in dark blue. (b) Outer membrane, inner membrane, and four representative cristae shown in different colors.

Further progress on the internal structure of mitochondria was made using immunolabeling and transmission electron microscopy [88,89]. It was shown that the OXPHOS Complexes I–V were preferentially localized in the cristae membranes, and hence, the cristae membranes were the principal site of OXPHOS [88,89]. The recent solution of the structure of the OXPHOS supercomplexes [6,7,8,9,10] by cryo-EM techniques led to further molecular insights. These developments enable us to link mitochondrial ultrastructure to function in OXPHOS and provide a novel raison d’etre for the organization of the OXPHOS complexes into supercomplexes (Section 3.6).

Furthermore, ion translocation into sealed vesicles formed by the cristae membranes in OXPHOS has different implications from postulating ion movement across the inner mitochondrial membrane into a permeable outer mitochondrial membrane that is open to the cytosol.

What is respiration? Linking mitochondrial ultrastructure to function in OXPHOS

The current view of mitochondrial respiration sees Complexes I–IV as being involved in electron-coupled proton translocation (Figure 9a) after which the process is finished. If respiration is defined in this way, then no likely explanation appears possible for multiple energy conservation and transfer pathways determined in this work (Tables 1 and 2 and Figure 6). Nor can the role of succinate anions (Figures 1–5) be rationalized, and a single energy pool is sufficient. However, the results of this work require at least two distinct pathways between the redox and ATP sides (Figure 6) to explain the site-specific inhibitory interaction of the guanidine derivatives with the respiratory chain in mitochondria and the release of the induced inhibition by uncouplers. How can this be realized?

Figure 9

(a) Traditional and (b) new model of respiration. Primary ion translocations are denoted by bold arrows, and secondary ion translocations by dashed arrows.

Obviously, at the level of protons, no distinction is possible, since all redox complexes are involved in the same function and carry out electron-coupled proton transport (Figure 9a). However, if there exists another step subsequently of anion translocation, that is, the proton transport step is coupled to translocation of metabolic anions against their concentration gradient on the redox side [21,36,39,42,50,70,78], then a distinction is possible. In particular, since succinate is a dicarboxylic acid anion, there is a need to translocate both succinate monoanions and succinate dianions back and forth across the cristae membrane by the redox Complexes I–IV and Complex V in order to maintain perfect homeostasis, as shown in the schematic of Figure 9b. The manifold central roles of succinate in cellular metabolism are being increasingly recognized by new research [57,59,78,90,91]. Thus, respiration should be viewed as the process of electron-coupled proton translocation–coupled dicarboxylic acid anion translocation (Figure 9b). Such coupling requires sensing of the local Δψ produced by primary H⁺ translocation by the redox enzyme complexes/subunits, for which the complexes need to be organized into supercomplexes. If the subunits are separated or are disassociated from each other or lie at great distance from each other, such local electrical potential sensing would be precluded. It should also be noted that the measurements [92,93,94,95] cannot distinguish between a transient, local Δψ [36,38,39,40,63,64,70,78,85,95], and a steady-state delocalized potential, Δφ [83,84]. Hence, care should be taken to consider both possibilities in the interpretation of biochemical data.

Finally, if OXPHOS Complexes I and IV are specialized to recognize, bind, and translocate succinate dianions coupled to protons, and Complexes II and III are involved in the coupled translocation of succinate monoanions and protons, respectively, then the differential inhibition induced by the guanidine derivatives, the site-selective, energy-linked release of inhibition by uncouplers, and the network of two energy transfer pathways (Figure 6) can be explained in a natural way. The new view of respiration would also allow maintenance of overall electrical neutrality within the intracristal space and the matrix space of mitochondria [63]. Furthermore, the consensus H⁺/2e^– stoichiometry of 4 for Complexes I and IV, and H⁺/2e^– = 2 for Complex III of the mitochondrial electron transport chain [67,96] would also find a ready explanation [43,78]. Such coupling of proton and anion/counteraction transport had been a longstanding prediction of the two-ion theory and torsional mechanism of energy transduction and ATP synthesis based on various considerations of theory, experiment, and computation. A new model of integrated mitochondrial function with a possible organization of the OXPHOS supercomplexes is shown in Figure 10.

Figure 10

Model of supramolecular organization of Complexes I–V in the cristae membranes of mammalian mitochondria along with the overall molecular mechanism of their functioning based on Nath’s two-ion theory of energy coupling and ATP synthesis [21,63,64,65,66,67,68,69,70,117] involving succinate anions and protons. The structures of Complexes I–V are assembled from the structures of the respirasome supercomplex I₁III₂IV₁ (PDB ID: 5XTH; CI in blue, CIII in gold, and CIV in magenta), the structure of Complex II (PDB ID: 1ZOY; CII in green), and the structure of the tetramer of ATP synthase, Complex V (PDB ID: 6J5K) [135]. Primary ion translocations by Complexes I–V are denoted by bold arrows, and secondary ion translocations by dashed arrows. The translocation of succinate dianions and succinate monoanions along with protons by the respective Complexes are clearly distinguished and provide a novel explanation of the differential interactions and effects of various pharmacological and uncoupling agents at discrete points of the respiratory chain found in this work (Figure 6).

Biological implications in cell life, cell death and apoptosis, and mitochondrial diseases

The above has several biological implications in health and disease [17,18,19,20,22,23,24,26,27,28,29,33,34,97]. As discussed, a widely held view of the existence of supercomplexes is that they enhance catalysis by channeling mobile substrates [13,98]. However, it has also been asserted that no robust experimental information exists to justify this view [14]. Evidence from flux control analysis [11,99,100] and biochemical evidence supporting the existence of Coenzyme Q and cytochrome c pools [101] and functional segmentation of the pools [12] have also been challenged [16,102]. However, views questioning substrate channeling [16,102] have themselves been questioned, for instance, based on the different time scales of the processes being probed [103].

Other suggestions such as a decrease in ROS production in the context of Complex III [6] have also been disputed [14]. For instance, it has been asserted that, contrary to the suggestion of decreased ROS at the Q _i site, it is the Q _o site that dominates ROS production [104], not the Q _i site.

The view that individual OXPHOS complexes have similar catalytic activities as in supercomplexes is based on assaying electron-coupled H⁺ activity only, but not possible subsequent coupling to anion translocation (Figure 9b). A number of reports on aging and mitochondrial neurodegenerative diseases show that integrated mitochondrial function is indeed compromised in diseased states, even though no defect exists at the level of the individual OXPHOS Complexes I–IV [15,25]. It is very difficult to explain these observations if supramolecular organization of the OXPHOS Complexes did not confer any functional advantages. The need to sense local potentials, for example, between Complexes II and III, in order to couple proton and anion transport, maintain homeostasis of both monoanionic and dianionic forms of succinate, and distribute and compartmentalize the processing of these anionic forms among Complexes I/IV and Complexes II/III (Figures 9b and 10) helps us progress beyond current narratives.

Another constructive possibility to progress beyond the inconclusive debates over the past 25 years and break the current impasse in the field is to consider pools not only in terms of material, for example, Coenzyme Q pools or cytochrome c pools and their functional segmentation [13]. It is important to also think, interpret, and integrate the available biochemical information in terms of energy pools, attempted in the present work (Figure 6 and Sections 3.3, 3.4, and 3.6).

The concept of local potentials in channels and transporters [105,106,107,108,109,110,111,112,113,114,115,116] has been shown to be important for a better understanding of apoptosis and cell death [22]. The integration of the above concepts in biological energy coupling, transduction, and ATP synthesis in cell life and death has also been discussed recently [22,70,117]. These ideas provide new avenues and may even be crucial to catalyze the progress of OXPHOS research in cell life, cell death, and mitochondrial dysfunction and disease.

Applications of the results to specific diseases: type 2 diabetes and AD

In response to referee comments and suggestions, the grand challenge of application of the novel biochemical and pharmacological results contained in this work to the supercomplex problem afforded by specific diseases, for example, type 2 diabetes and AD, has been attempted. Several caveats need to be made before such a task is initiated.

The above multifactorial diseases have specialized journals dedicated to them, and it is quite impossible to even summarize the recent medical advances. For instance, let us take the role of metformin in type 2 diabetes. Metformin has been the first-line oral therapy for the treatment of type 2 diabetes for over 60 years as per consensus American and European guidelines [118]. Its safety and efficacy have been studied extensively, and its action leading to the specific inhibition of mitochondrial Complex I has been documented thoroughly [71,72,119,120,121]. Almost 20,000 research articles have been published on metformin to date; however, despite this large body of research, its exact mechanism of action is not completely understood [122]. The uncertainties have to do with a large number of factors. For example, studies of metformin performed on submitochondrial particles may not retain the supramolecular organization of intact mitochondria dealt with in this work. Furthermore, the concentrations of the drug used in cells and in mitochondrial studies vary substantially, and studies that correlate the metformin concentration in tissues/cells to the metformin-induced inhibition of mitochondrial Complex I are not readily available. In this sense, intact mitochondrial studies on supramolecular organization of the OXPHOS complexes, as in this work, stand on surer ground. Furthermore, unlike in our experiments, Complex I inhibitors such as rotenone cannot be used in a clinical medicine setting in view of their toxicity effects. Similarly, cyclosporine A (CsA), a routinely used inhibitor of the permeability transition (PT) in the laboratory, has little therapeutic potential owing to its immunosuppressive side effects and cannot be used in long-term diseases such as hyperglycemia-induced angiopathy.

In this section, it is assumed that the clinical effects of the anti-diabetic drug metformin [118,121,122] and the early events in the neurodegenerative diseases such as AD and Parkinson’s disease have a mitochondrial origin and involve mitochondrial network dysfunctions (for reviews, see refs. [17,123]). It is also taken as given that the concepts of the author’s work on ATP synthesis and OXPHOS in cell life, apoptosis, and cell death ([22,40,44]; this work) can be integrated to achieve a better understanding. Thus, in the spirit of refs. [22,40], chemical agents with pro-apoptotic/necrotic and anti-apoptotic/necrotic effects can be discussed based on their inducing or inhibitory activities and regulatory effects and their interactions that have the ability to open or close the permeability transition pore (PTP) in mitochondria. This is also in line with the pioneering work of Bernardi and colleagues on cell death and the PTP (for a review, see ref. [24]).