Redox bifurcations: Mechanisms and importance to life now, and at its origin

Crucial electron transfer steps in hydrogenotrophic methanogens – anaerobic Archaea that reduce CO2 with hydrogen to create a proton gradient – and in Clostridia go steeply uphill against a substantially unfavorable redox midpoint potential difference. A solution to this thermodynamic conundrum was proposed over the course of the last three years in the form of “energy conserving electron bifurcation” 1-4. The electrochemical principle of electron bifurcation is, in fact, an old but almost forgotten friend of research in biological redox reactions. It finally put an end to the sleepless nights of pioneers studying mitochondria who did not believe their eyes when they witnessed low potential mitochondrial cytochrome b becoming reduced upon adding a strong oxidant (ferricyanide). Is the aforementioned resurfacing of the bifurcation mechanism in a completely different system a coincidental phenomenon or does it provide a first glimpse into a largely uncharted world of energy conserving redox reactions, the existence of which we have previously suggested 5? The hypothesis put forward by Buckel, Thauer, and colleagues in essence stipulates that the 2-electron carrier, flavin, bifurcates its two electrons towards acceptors with much lower and much higher redox potentials than that of the average (2-electron) transition of flavin. In the case of methanogens, the high potential acceptor is the CoM-SS-CoB heterodisulphide, while the low potential acceptor is ferredoxin. The highly exergonic electron transfer to the high potential acceptor is argued to compensate the endergonic, half-reaction, i.e. reduction of an acceptor substantially more negative than the flavin. Energy conservation via such electron bifurcation is certainly not an outlandish idea; after all it is what allows us to write and you to read this article. The efficiency of mitochondria crucially depends on such a mechanism and, without it, eukaryotic life as we know it would clearly be impossible. The first step in Peter Mitchell's Q-cycle 6 in mitochondrial complex III exploits this principle and only the full Q-cycle allows mitochondria to squeeze the last drop of free energy out of the electrochemical difference between heterotrophic substrates and O2. In the case of mitochondrial complex III (and, actually, the entire family of prokaryotic and eukaryotic Rieske/cytb complexes), the bifurcating agent is a quinone. Quinones, similarly to flavins, are 2-electron carriers and proposing an analogous electron bifurcating function for flavins thus appears both tempting and promising. The hypothesis put forward by Buckel and colleagues for Clostridia 1, 2, and subsequently by Thauer and colleagues for methanogens 3, 4, takes the “peculiarity” out of quinone-based electron bifurcation in the Q-cycle while, at the same time, elevating electron bifurcation to the status of a general and likely ancient principle for achieving reduction of very low potential redox components and, in the process, for transferring energy. We have noted in the past 5 that a number of 2-electron redox centers are, in principle, able to perform this electron bifurcation. In addition to quinones and flavins, molybdenum, tungsten, NAD, methanophenazine, and probably several others can, in principle, achieve analogous redox reactions. Could all 2-electron redox substances bifurcate electrons? Certainly not, and the reason why leads us to discuss details of the bifurcation principle which, to our minds, have so far been neglected in the articles stipulating flavin-mediated uphill redox reactions. The salient property of redox compounds that are able to bifurcate is their ability to feature what is called in electrochemistry “crossed-over” individual redox transitions. The theory and mathematics of crossed-over redox transitions in 2-electron redox compounds is as old as electrochemistry and has been dealt with in the respective textbooks as far back as the 1960s. Intuitively, one might assume the complete oxidation of a fully reduced 2-electron component to at first proceed at a relatively low potential towards the semi-oxidized state and, eventually, at a higher potential to the fully oxidized form (Fig. 1, dashed curves). This is not how a typical reduced quinone (e.g. ubiquinol) behaves in aqueous solution. Instead, it will lose both electrons simultaneously, i.e. in a 2-electron oxidation, yielding a titration curve similar to that indicated by the green continuous line in Fig. 1 (note the steeper slope of this so-called “2-electron Nernst curve”). Quinones in certain binding sites (on proteins), however, do go through two individual reduction steps (such as, e.g., the QB site of photosystem II). The reason for this difference is that the electrostatic environment modifies the midpoint potentials of the two individual redox waves, as visualized in Fig. 1 by the dashed and dotted redox waves. Changes in pH, for example, can bring about such shifts in redox potential and, within proteins, exchanges of amino acids in close vicinity to the quinones can also produce these kinds of effect. This can even induce situations where the nominal first oxidation potential is at more positive values than the second, as indicated by the continuous blue and red curves in Fig. 1. It is, of course, physically impossible for a quinone to titrate in this way. Imagine that the first oxidation proceeded along the blue continuous line. Once in the semi-oxidized state, these quinones would suddenly be overdue full oxidation since their semi-oxidized/oxidized redox potential is at much lower values than the ambient potential they actually experience. The bulk of such “red hot” semiquinones would thus instantly become fully oxidized, without bothering to proceed down another Nernst-slope. Such behavior obviously sounds bizarre and, indeed, does not occur in titrations. Instead, the bulk of quinones take the 2-electron route mentioned previously and indicated by the green line in Fig. 1. Only in this case can energy conserving electron bifurcation occur as we will detail below. Theoretical titration curves of 2-electron redox compounds, such as those mentioned in the text. The blue parts of these “Nernst-curves” denote oxidation from the fully to the semi-oxidized state, whereas the red portions indicate the second oxidation step towards the fully oxidized compound. Dashed, dotted, and continuous lines represent different situations with the two half-reactions being shifted in redox midpoint potential as discussed in the text. The dashed and the dotted curves correspond to the un-crossed situation, whereas the continuous blue and red waves represent an example of a crossed-over case. In this latter case, the individual redox transitions are “virtual”, whereas the actual bulk redox titration of such a compound would follow the green continuous (n = 2 Nernst-) curve. It appears that this basic electrochemical fact is not fully appreciated in the articles proposing flavin-mediated redox bifurcation in Clostridia and methanogens 1-4. Kaster et al., for example, state that “The proposed mechanism of flavin-based electron bifurcation is based on the findings that flavoproteins (FP) can exhibit three different redox potentials, namely an for the FP/FPH2 couple (n = 2), an for the FP/FPH couple (n = 1), and an for the FPH/FPH2 couple (n = 1). (FP/FPH) is generally more positive and (FPH/FPH2) more negative than (FP/FPH2). Oxidation of FPH2 by two one-electron acceptors with different redox potentials thus leads to a bifurcation of the two electrons in FPH2” 4. In Fig. 2, we schematically represent the redox midpoint potentials of individual waves of a hypothetical flavin as a function of the stability constant (KS) of the semi-oxidized form; a scheme often referred to as the “redox see-saw” 7. The stability of the flavosemiquinone, of course, is a direct measure of the mutual positions of the two redox half reactions, as indicated by the titration curves shown at the top of Fig. 2. As long as the two half-reactions are un-crossed, a flavosemiquinone can be measured in the indicated redox range (orange curve) whereas, in the crossed-over case, both electrons go together and no stable flavosemiquinone is observable. The situation invoked by Kaster et al. corresponds to the individual potentials indicated to the right of the cross-over point. The reaction sequences proposed (2–4) invoke a sequential reduction of the low potential ferredoxin by the fully reduced flavin and, then, a subsequential redox transition to oxidized flavin-reducing crotonyl-CoA (in the case of Clostridia) or the CoA-S-S-CoB heterodisulphide (in methanogens), as indicated in Fig. 2. The “redox seesaw”. Schematic representation of the dependence of midpoint potentials of the two individual redox transitions in a hypothetical flavoprotein, as invoked by the scenarios of various studies 1-4 on the stability constant of the intermediate (flavosemiquinone) state. The individual potentials in the normal and the crossed-over (or inverted) region are supposed to be 400 mV apart, with an average potential for the 2-electron transition of −400 mV. The two individual transitions would, under these conditions, be more reducing than ferredoxin and the heterodisulphide, respectively. The redox waves indicated above the redox seesaw scheme represent the theoretical Nernst curves for the 1st (blue) and 2nd (red) oxidation waves on both sides of the cross-over point. The green curve and the green rectangle represent the 2-electron titration wave and the concomitant redox midpoint potential, respectively. The orange dotted line denotes the amount of flavosemiquinone present in the corresponding redox range if the flavin is in the regime of the right hand part of the scheme. KS represents [semi]2/[ox][red]. As is apparent from this figure, energy conserving electron bifurcation can only occur left of the cross-over point, whereas reduction of ferredoxin from the stable flavosemiquinone to the right of the cross-over point is thermodynamically excluded. This is not how redox bifurcations work. In their articles the authors argue that the endergonic reduction of ferredoxin is driven by the exergonic reduction of the high potential acceptor. But why should the fully reduced flavin need the second oxidation step to perform the first one? It is already in the regime of separated 1-electron transfers and will possibly reduce ferredoxin but, even much more readily, will reduce the high potential acceptor due to the much stronger driving force for electron transfer to the latter (see Fig. 2) and then just remain as a flavosemiquinone. Of course, one might invoke some conformational events that might ensure that the ferredoxin gets close enough to the reduced flavin before the high potential acceptor does. However, this does not solve the energetic problem, it only aggravates it. Because the fully reduced flavin on the right hand side of the scheme first had to be generated via reduction of the semi-oxidized form, it requires another even more reducing compound than the low potential ferredoxin, which is exactly the contrary of what is needed to rationalize the energetics of the overall reactions in Clostridia and methanogens. In contrast, the proposed energy conserving bifurcation works perfectly well in the left hand, crossed-over region. As shown in Fig. 1, the 2-electron transition of the flavin is, in general (for thermodynamic reasons), halfway between the two formal redox potentials of the individual 1-electron steps. H2, in the case of methanogens (or NADH in Clostridia), is thus sufficiently reducing to pass its electrons to flavin and thus convert it into its fully reduced state. If both the high potential acceptor and the ferredoxin are simultaneously close to the reduced flavin (corresponding to the high potential Rieske center and the low potential b-hemes in the Q-cycle enzymes), the first oxidation reaction (occurring at about −200 mV in Fig. 2) cannot reduce the ferredoxin but will be driven by the high potential acceptor. This, however, leaves the flavin in its “red hot” flavosemiquinone state, mentioned above, that can reduce the low potential ferredoxin. We have now come full circle to the oxidation of ubiquinol in the Q-cycle and the “magic” oxidant-induced-reduction of b-hemes in mitochondria. These mitochondria were, due to the presence of succinate, in a state where ubiquinone and the so-called high potential components (the Rieske center and heme c1) of the bc1 complex are reduced, whereas the low potential b-hemes remained oxidized. Added ferricyanide oxidized the Rieske center, which in turn removed the first electron from ubiquinol at a relatively high potential and the strongly reducing ubisemiquinone thus transferred its electron to b-heme. Thus, an astonishing experimental result finds its natural explanation. In short, energy conserving redox bifurcation absolutely requires (i) the redox compound to be in its 2-electron (redox crossed-over) regime, (ii) the initiation of the reaction by a high potential acceptor that is able to commandeer an electron from the formally high potential first oxidation transition of the 2-electron redox center, and (iii) spatial proximity of the low potential acceptor at the moment all this happens. The latter point is due to the fact that the “red hot” semi-oxidized state will proceed to reduce practically anything that is reducible within close proximity. It is noteworthy that the above facts provide powerful experimental tools to (hopefully) further consolidate the existence of the proposed process: equilibrium redox titrations of the flavin should demonstrate an n = 2 Nernst transition and a concomitant absence of a flavosemiquinone species. Since, as mentioned above, the ability to redox bifurcate electrons is not a unique property of quinones but a basic electrochemical characteristic of certain 2-electron redox compounds, it would have been more than astonishing if life had taken advantage of this formidable redox mechanism exclusively in the Q-cycle. The discovery of an analogous reaction mediated by flavins in Clostridia and methanogens is therefore reassuring. We would argue that further similar bifurcation-based energy conversion schemes will be discovered in future research and the first tentative candidates have already been proposed 8, 9. Commensurate with its physico-chemical generality, redox bifurcation via crossed-over potentials should be recognized as one of the basic principles of life to convert energy. This raises the question as to when did life first recruit this process to help meet its energetic needs. Quinones and flavins are non-proteinaceous small organic molecules and proposing their presence in a hypothetical RNA-world would probably not cause uproar amongst proponents of this scenario. However, as we have pointed out in the past 5, crossed-over potentials in 2-electron redox compounds are not limited to organic chemistry. Molybdenum and tungsten are further likely candidates. The molybdenum center in the molybdopterin enzyme arsenite oxidase, for example, has been shown by electrochemistry to feature fully crossed-over redox transitions 10. Molybdopterin enzymes are indeed key players in the first steps of Wood–Ljungdahl-related reaction pathways, frequently put forward as primordial ways to conserve energy 11. Oxidants such as nitrogen oxides or Fe3+ may have cooperated with molybdenum and/or tungsten to achieve reduction of otherwise unattainably electronegative redox carriers using, for example, H2 as the ultimate electron donor 5, 12. Even if details of such pathways still admittedly need to be worked out and critically assayed, it seems evident to us that the general physico-chemical nature of electron bifurcation via crossed-over redox potentials shows that this process is older than proteins and even of organic molecules and it consequently has very likely contributed to the earliest stages of emerging life. As we have outlined, basic electrochemical principles under particular conditions allow 2-electron redox compounds to achieve counterintuitive, but biologically advantageous, redox reactions via bifurcation of their two reducing equivalents into an endergonic and an exergonic pathway. We consider that, thus far, the importance and generality of this mechanism to biological redox reactions has been underestimated and that it may have facilitated entry to the first metabolic pathway of life. MJR works at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (NASA), supported by the NASA Exobiology and Evolutionary Biology award (NNH06ZDA001N) and by the NASA Astrobiology Institute (Icy Worlds). The authors would like to thank Bill Martin (Duesseldorf/Germany) and Frauke Baymann (Marseille/France) for stimulating discussions.