Setting the Record Straight: A New Twist on the Chemiosmotic Mechanism of Oxidative Phosphorylation

Mitchell's chemiosmotic theory has been the cornerstone of the mechanistic understanding of mitochondrial energy metabolism for the past 7 decades.^1,2 Although aspects have remained controversial (in the minds of some) since its original description, many of its features and predictions have withstood the test of time and it has gained widespread acceptance; yet there have been certain areas that it clearly does not seem to adequately describe.^3–8 The original Mitchell view posited that the respiratory chain, harnessing redox energy, in turn pumps protons out of the matrix setting up an inward-directed H⁺ gradient with energy components stored as ΔΨ_m and ΔpH. These energies, particularly that stored as ΔΨ_m, are harnessed by driving H⁺ back into the matrix, exerting work by its specific translocation path in turning the F_o motor in ATP synthase, and in a mechanochemical energy transduction with the F₁ motor of ATP synthase, produces ATP from ADP + Pi.⁹ This constitutes the (electro)chemical energy transduction part of “chemiosmosis.” A particular area that probably isn't adequately accounted for by Mitchell's theory as originally described¹⁰ is the “osmotic” part of chemiosmosis (see below). Historically, this osmotic aspect has been considered to be a key mechanistic insight.^11,12

What we have discovered^13,14 is entirely compatible with Mitchell, but also builds upon this foundation and extends these mechanisms by revealing an additional “K⁺ uniport circuit” through ATP synthase (beyond the presently accepted H⁺ uniport circuit) whereby cytosolic K⁺ is also able to be driven by ΔΨ_m and, in an analogous manner as H⁺, transported without being exchanged by ATP synthase (and turning its F_o motor, etc.) to make ATP (Fig. 1A). In short, the electrochemical uniport of K⁺ in addition to that of H⁺ (in approximately the ratio of 3 K⁺ for every H⁺), to make ATP, and its accumulation in the matrix in proportion to workload, is a new discovery that advances the understanding of not only the electrochemical, but also the osmotic part of chemiosmosis. Specifically, the electrogenic uniport of K⁺ by ATP synthase that substantially contributes to the synthesis of ATP, accumulates matrix K⁺ at levels producing much higher concentrations (by many orders) than that of ATP synthase-uniported H⁺ levels, so that these K⁺ levels and, importantly, their changes in direct proportion to that of workload, produce relevant osmotic-drive signals that control matrix volume changes, which feeds back to cause changes in the activation state of the respiratory chain (by the so-called “volume-activation of respiration” mechanism), serving to facilitate matching energy supply with demand. The resulting matrix-accumulated K⁺ is then extruded via a separate K/H exchanger, which is (likely) kinetically phase-lagged, and this enables the dynamics of the process to produce transient/temporary, controlled “up” and “down” matrix-volume signals to new work-level-related physiological steady-state levels (i.e., by this process, preventing excess volume expansion or contraction). All the input energy involved in K⁺ extrusion still derives from the original respiratory chain-generated H⁺ gradient (Δµ_H), and thus the entire process is again entirely consistent with Mitchell.

Figure 1.

Comparison and functional implications of alternate models of ion-directed energy coupling in ATP synthase. (A) Chemiosmotic theory extended by the capacity of ATP synthase to utilize K⁺, H⁺ (and Na⁺) electrogenic uniport to drive ATP synthesis. The respiratory chain pumps protons out of the matrix setting up an inward-directed H⁺ gradient with energy components stored as ΔΨ_m and ΔpH. These energies (mainly ΔΨ_m) are used by driving H⁺, K⁺ (and Na⁺) back into the matrix via an electrogenic translocation path enabling them to perform work (by exerting torque on the F_o  c-ring) and turning the F_o rotary motor in ATP synthase, and in a mechanochemical energy transduction with the F₁ catalytic complex of ATP synthase, produces ATP from ADP + Pi in the mitochondrial matrix. K⁺ and H⁺ each occupy (but physically separately) the same ion-binding locations on the c-ring of the F_o motor in ATP synthase. In the standard “H⁺-only”-running model of chemiosmotic theory, all eight positions on the c-ring (in the mammalian F_o) would be H⁺ (shown as green balls, as in Panel B). In our model shown here, of the eight c-ring positions, on average approximately six will be occupied by K⁺ (red balls) and two will be occupied by H⁺ (in a net ratio of ∼3:1) in a random ordering, to produce three ATP for a full c-ring turn with the translocation of those eight ions (N.B., for clarity, 5 K⁺ and 3 H⁺ are shown on the c-ring in the illustration; since only ∼1 Na⁺ is translocated for every 24 cations driving ATP synthase, it is not shown; the direction of ion movement shown here in F_o is for convenient illustration purposes only—the actual direction of rotation is clockwise for ATP synthesis, with the ion entrance and exit “hemichannels” arranged respectively). The resulting matrix-accumulated K⁺ is then extruded via a separate (kinetically phase-lagged) K⁺/H⁺ exchanger. The electrogenic uniport of K⁺ by ATP synthase is adaptive and workload dependent, producing relevant osmotic-driven signals that control matrix volume changes that in turn result in activation of the respiratory chain (volume-activation of respiration mechanism), serving as feedback for matching energy supply with demand. See text and accompanying references for details. (B) Nath's electroneutral “two-ion” theory of energy coupling. According to Nath's electroneutral “two-ion” theory the primary redox energy-harnessing step at the respiratory chain is the net electroneutral pumping of H⁺ out of the matrix in exchange for K⁺, which secondarily generates ΔpK on the matrix side (together with ΔpH on the intermembrane space side). Then, harnessing the newly formed ΔpK, K⁺ is translocated first from the matrix down its concentration gradient by a postulated K/H antiport mechanism residing within the ATP synthase. The resulting K⁺ movement-related local charge separation generates a temporary, localized ΔΨ_m (within the ATP synthase complex itself) that electrostatically attracts H⁺, enabling translocation of H⁺ and binding to the c-ring (producing torque, etc.), and ultimately (together with ΔpH) driving H⁺ back into the matrix. Except for H⁺ using a putative, coupled ion-antiport/translocation activity (structurally uncorroborated and unverified by any physical measurement technique) hypothetically occurring at the a-c-subunits’ interface of F_o to access the c-ring (Panel B), rather than using the accepted, structurally identified a-subunit aqueous access half-channels (as depicted in Panel A; e.g., for bovine structure, see ¹⁶, and references therein for concordant structural evidence across taxonomic domains), the subsequent processes apparently involve the standard “H⁺-only”-running model to drive ATP synthase, with all positions on the F_o  c-ring bound exclusively by H⁺ (shown as green balls). While this latter K/H-exchanged energy (together with that of ΔpH) transferred back to H⁺ is finally used to drive ATP synthesis, another important distinction in this model (vs that depicted in Panel A) is that K⁺  neither binds to, nor has its energy directly harnessed to turn the c-ring. In fact, the vector direction of this latter ATP synthase-antiported K⁺ movement (resulting from the putative K/H exchange activity) is exactly the opposite of that depicted in the model shown in Panel A. This process would produce apparently paradoxically maladaptive reciprocal changes in matrix K⁺ and workload, by retaining higher matrix K⁺, volumes and respiratory activities at low workloads, and in the opposite case, causing a lower matrix K⁺, volumes and respiratory activities at high workloads, which would serve the opposite purposes required for appropriate energy supply-demand matching. Because all these ion exchange processes are net electroneutral, the bulk phase ΔΨ_m is functionally irrelevant in this model.

Prof. Sunil Nath has written an Opinion article related to our work in the previous issue of Function¹⁵ that we feel needs a response. From his published work, Prof. Nath is clearly a skilled and dedicated scientist. We appreciate his enthusiasm in pushing the scientific envelope, a maverick trait that we appreciate as kindred spirits in the pursuit of scientific understanding. Prof. Nath's current conjecture in his Opinion article¹⁵ suggests that the compendium of our 2-decades of research that we have put forth in the two-part series in the current issue of Function^13,14 is somehow the same idea and mechanism as his “two-ion theory of energy coupling” of ATP synthase. We respectfully point out that these 2 scientific stories, which might appear to be superficially similar because they happen to involve the same two cations (K⁺ and H⁺) in mitochondrial energy metabolism, are nevertheless actually entirely distinct and even incompatible models with each other regarding the way Nature works. In the spirit of collegiality, we infer this to be a simple misunderstanding on the part of Prof. Nath, as we will explain below.

We describe a process that involves the movement of univalent, alkali metal cations (K⁺ and Na⁺) in addition to H⁺, all in the same direction from the intermembrane space into the matrix, travelling along the same path in ATP synthase and occupying the same location on the c-ring (i.e., the monovalent cation-conducting, rotary component of ATP synthase's F_o motor). The electrochemical potential of each ion (K⁺, Na⁺ or H⁺) including ΔΨ_m is harnessed as it moves across the inner mitochondrial membrane to make ATP^13,14,17 (Fig. 1A). This is emphatically not a charge-neutral process, and intrinsically dissipates ΔΨ_m (and the respective ion-gradients, if present).

In stark contrast, Nath's “two-ion theory of energy coupling” model¹⁵ proposes an electroneutral H⁺/K⁺  antiport within the F₁F_o-ATP synthase itself to maintain electroneutrality¹⁵ (Fig. 1B). As recognized by Bertero and Maack¹⁸, our model ^13,14,19 defines electrogenic “uniport” of H⁺ and K⁺ in the same direction via the ATP synthase, where K⁺ extrusion is accounted for by a separate molecularly distinct K⁺/H⁺ exchanger (Fig. 1A).

The mechanism we describe is not only compatible but extends Mitchell's chemiosmotic theory: an electrogenic, ΔΨ_m (Δµ_H)-driven one-ion-uniport mechanism of H⁺, K⁺ (and some Na⁺) in ATP synthase to make ATP, explicitly leads to the adaptive workload-dependent changes in matrix K⁺ controlling osmotic drive and changes in matrix volume in direct proportion to regulate respiratory chain activity. Both H⁺ and K⁺ (as well as a lesser amount of Na⁺) are directly involved in making ATP by turning the F_o motor. Because of the electrogenic nature of the main processes described here, an obligatory feature is that the development of even small additional electrical charge imbalances due to the primary ion movements (for a sufficient time lag before the extrusion process “catches up”^13,14,19) will be accompanied by the movement of appropriate charge-balancing counterions. This produces the ability to develop changes in osmotic driving gradients that can regulate water and matrix volume dynamics, and in turn, the function of the respiratory chain (Fig. 1A).

In contrast, Nath's mechanism is completely contrary to this Mitchellian mechanism, as Nath has unambiguously stated.²⁰ Specifically, his mechanism describes the redox-driven electroneutral extrusion of matrix H⁺ by respiratory complexes with an obligatory, associated K/H exchange activity producing a “secondary translocation” of K⁺ at the respiratory complexes producing a matrix K⁺ gradient without changing bulk phase ΔΨ_m. This outwardly directed matrix K⁺ gradient (Δ[K⁺]) then apparently drives K⁺ back across the inner mitochondrial membrane through an entirely hypothetical path in mammalian ATP synthase to power an obligatory, net-electroneutral exchange process (putatively inside ATP synthase) of this K⁺ with H⁺, using the energy transferred and acquired in that exchange process to drive that H⁺ back through the usual path that turns the F_o motor enabling ATP synthase to make ATP. Only H⁺ is directly making ATP by turning the F_o motor (Fig. 1B).

So, the Nath mechanism is an electroneutral, ΔΨ_m-independent (ΔpH → ΔpK → ΔpH)-driven two-ion-antiport mechanism. It involves the two-ion antiport of K⁺ with H⁺ inside the ATP synthase molecule to make ATP, without a clear ability to develop significant, positively adaptive changes in osmotic drive (Fig. 1B). Even if, for some reason, an intrinsic K⁺-related osmotic change mechanism would yet be asserted for the Nath hypothesis, it would still seem to produce apparently paradoxically maladaptive reciprocal changes in matrix K⁺ and workload: it would seem this mechanism would serve the opposite purposes required for energy supply-demand matching by retaining higher matrix K⁺, volumes and respiratory activities at low workloads, and in the opposite case, causing a lower matrix K⁺, volumes and respiratory activities at high workloads.

Finally, and paradoxically, there are at least four lines of definitive and independent experimental evidence in our published work^13,14 that was cited by Nath as “confirming” or “revalidating” his theory¹⁵, each based on different and distinct experimental techniques, that instead entirely contradict and refute Nath's hypothesized mechanism. On the basis of our experimental evidence (elaborated below) ATP synthase behaves electrogenically, as demonstrated by, (1) the voltage-dependence of ion channel currents produced by ATP synthase, (2) the inability of single ATP synthase molecules—that are demonstrably competent to utilize a K⁺ gradient to synthesize ATP under certain defined conditions—to also achieve ΔpK-driven ATP synthesis at the system reversal potential, E_rev (i.e., the specific potential where the net flow of charge across the membrane is zero), (3) F₁F_o-ATP synthase reconstituted proteoliposomes in a K⁺ gradient (without an H⁺ gradient) can only make ATP when an exogenous protonophore is provided, and (4) a large K⁺ gradient creates a large, stable membrane (K⁺ diffusion) potential in F₁F_o-ATP synthase reconstituted proteoliposomes in the absence of an added protonophore.¹³

Importantly, regarding the first and second points raised above, when ATP synthase is examined in isolation, it behaves electrogenically—showing macroscopic currents that vary by voltage—rather than in an electroneutral fashion (¹³see Fig. 1D-G therein). Furthermore, these macroscopic currents in ATP synthase can be manifest by a K⁺ gradient itself in the absence of a H⁺ gradient at a holding potential of 0 mV (¹³see Fig. 4 therein). Under these latter conditions, these are pure K⁺ currents (driven by ΔpK), and when the vector of their charge movement is properly directed (with respect to the orientation of ATP synthase), they drive and enable ATP synthase to produce ATP. On the contrary, these same K⁺ gradient conditions (absent a H⁺ gradient) when set at a holding potential where there are no macroscopic currents (i.e., at the system's reversal potential, E_rev) result in that same molecule of ATP synthase producing no ATP (i.e., in the same experiments) (¹³see Fig. 4B therein). These are primary proof-of-principle results in the present research¹³. In contrast, if ATP synthase could function to synthesize ATP in an electroneutral fashion (i.e., driven by the K⁺ gradient and putatively enabled by an internal 1:1 exchange with H⁺ contained within ATP synthase, as Nath proposes), ATP synthesis would have occurred regardless of voltage in these experiments, but this does not happen¹³, and thus the Nath hypothesis is experimentally refuted.

Two additional, key experiments also contradict and refute the Nath hypothesis of an intrinsic K⁺/H⁺ exchange mechanism in ATP synthase to maintain electroneutrality: (1) F₁F_o-ATP synthase reconstituted proteoliposomes in a K⁺ gradient (without an H⁺ gradient) can only make ATP when a protonophore, FCCP, is provided, to enable the exchange of transported K⁺ with a positive counterion, H⁺ (i.e., in this case, together with the movement of K⁺ through ATP synthase to make ATP, the counter-movement of an equal amount of H⁺ constitutes a “K/H exchanger activity” and enables ATP synthesis) (¹³see Fig. 2 therein). Importantly, without adding FCCP (to allow a H⁺ leak pathway), there is no ATP synthesis in a K⁺ gradient, so there must be no functional K/H exchange activity present inside ATP synthase such as is central to the proposed Nath mechanism. (2) The membrane potential (K⁺ diffusion-potential) generated in F₁F_o-ATP synthase reconstituted proteoliposomes by a large K⁺ gradient is extremely stable (>> minutes) unless FCCP is added to enable the counter-movement of an equal amount of charge as H⁺ constituting a synthetic “K/H exchanger activity.” (¹³see Fig. 1A, C therein). In short, all of this evidence definitively refutes Nath.

In conclusion, our experiments^13,14  most certainly did not validate the Nath “Two-ion Theory of Energy Coupling and ATP Synthesis” hypothesis¹⁵. On the other hand, the ability of mitochondria to have a K/H (Na/H) exchanger activity separate from the ATP synthase (as we propose^13,14,19) produces a situation whereby even small kinetic differences between those cation fluxes, which are physically different and (potentially) spatially/diffusively separate entities (e.g., the effects of those component processes having different rise- and relaxation-times to a driving step-function) can yield small but important, regulatory accumulations or reductions in matrix K⁺ and Na⁺. This, in turn, controls changes in osmotic drive, matrix volume,^13,14 Ca²⁺ levels and redox status,²¹ with consequent actions on respiratory chain function, energy supply-demand matching and cardioprotection signaling. These are some of the important implications of our findings. Notably, while our principles, experimental findings and conclusions are fully consistent with Mitchell's chemiosmotic mechanism^13,14,19, Nath's “torsional mechanism of energy transduction” and ATP synthesis is stated to be incompatible with Mitchell's chemiosmotic theory and cannot be explained by it.^15,20

We confidently feel that the theoretical subject matter referenced in Prof. Nath's Opinion Article,¹⁵ while interesting, is not fundamentally or substantially relevant to, or consistent with, our novel body of experimental work.^13,14,19

Conflict of interest

The authors have declared that no conflict of interest exists.