MIE review


Published on

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

MIE review

  1. 1. REVIEW pubs.acs.org/CRChemistry of Enzymatic ATP Synthesis: An Insight through theIsotope WindowAnatoly L. Buchachenko,*,† Dmitry A. Kuznetsov,†,‡ and Natalia N. Breslavskaya§† N. N. Semenov Institute of Chemical Physics, Russian Academy of Sciences, 4 Kosygin Street, 119991 Moscow, Russian Federation and Institute of Problems of Chemical Physics, Russian Academy of Sciences, 142432 Chernogolovka, Russian Federation‡ Department of Medicinal Nanobiotechnologies, N. I. Pirogov Russian State Medical University, 1 Ostrovityanov Street, 117997 Moscow, Russian Federation§ N. S. Kurnakov Institute of General and Inorganic Chemistry, Russian Academy of Sciences, 119991 Moscow, Russian FederationCONTENTS 1. INTRODUCTION1. Introduction A Among enzymes that are well-known for being perfectly2. Isotope Effects in ATP Synthesis B arranged molecular devices of paramount importance are those3. Isotope Effects in Creatine Kinase Reaction B generating the energy carrier adenosine triphosphate (ATP). A tremendous success has been reached in understanding the 3.1. Magnesium Isotopes B mechanical functioning of these enzymes. New technologies of 3.2. Calcium and Zinc Isotopes C single-molecule optics and mechanics are developed to visualize4. Are Isotope Effects Trustworthy? C the mechanics of the most beautiful molecular motor, ATP5. Pyruvate Kinase: Magnesium and Zinc Isotopes E synthase.1À4 It has been shown that the direction of its rotor6. Phosphoglycerate Kinase: Magnesium Isotopes E circulation—clockwise or counterclockwise—controls chemical7. Mitochondria: Magnesium and Zinc Isotopes E function of ATP synthase and changes it reversibly from synthe-8. Nuclear Spin-Dependent Chemistry sis to hydrolysis.5,6 Moreover, macroscopic reversibility in ATP of ATP Synthesis F synthesis does not mean microscopic reversibility, i.e., the path-9. In Vivo Isotope Effects and Medicine H way for ATP hydrolysis is not the same as for ATP synthesis.2,7À9 Despite the great progress in the knowledge of structure and10. Magnetic Field Effects H understanding of molecular dynamics and functioning of ATP11. Energetics of ATP Synthesis I synthesizing enzymes,10 far less is known about the chemical 11.1. Complexes of Metal Ions and their Reactions I mechanism of ATP synthesis.2,4 How does mechanical motion in 11.2. Energy-Forbidden and Energy-Allowed molecular motors result in the chemical reaction of PÀO bond Electron Transfer J formation? How does mechanical energy transform into the12. How Do ATP-Synthesizing Enzymes Function? K energy of the newly born chemical PÀO bond in ATP? What is13. Magnetic Parameters and Spin Dynamics the limiting step—entering of reactants into the catalytic site, release of products, or the chemical reaction itself? All these of the Ion-Radical Pair K questions are under discussion.14. Kinetics and Quantitative Measure of Isotope A universal property of the phosphorylating enzymes is their Effects L magnesium dependence. A generally accepted mechanism im-15. Rate-Limiting Step of ATP Synthesis M plies a nucleophilic addition of inorganic phosphate (provided by16. Efficiency of ATP Synthase as a ATP synthase) or phosphate group of substrate (kinases) to Molecular Machine M ADP. In terms of this mechanism, the main function of the Mg2+17. Do Enzymatic Motors Function Reversibly? N ion was traditionally thought to coordinate phosphate reactants, to keep them in a position precisely within the reaction pathway18. Comparative Analysis of the Two Mechanisms of of nucleophilic attack, and perhaps to slightly modify their ATP Synthesis N chemical reactivity by redistribution of charges in a Mg(ADP) Energy N complex. The Mg2+ ion was always considered as an assistant and Ion Concentration N never assumed to participate in the ATP synthesis as a reactant.19. Is ATP Synthesis a Nucleophilic or an Ion-Radical This simplistic nucleophilic mechanism was accepted as a para- Process? N digm, so that the chemistry of ATP synthesis itself did not attractAuthor Information O much attention because there was nothing to discuss.Biographies OAcknowledgment P Received: April 28, 2011References P r XXXX American Chemical Society A dx.doi.org/10.1021/cr200142a | Chem. Rev. XXXX, XXX, 000–000
  2. 2. Chemical Reviews REVIEW The nucleophilic mechanism does not imply participation of Table 1. Magnesium, Calcium, and Zinc Isotopes26any paramagnetic species in the ATP synthesis, so that the isotope spin magnetic moment, mB natural abundance, %discovery of magnetic isotope and magnetic field effects11,12 inenzymatic ATP synthesis catalyzed by magnesium was unprece- 24 Mg 0 0 79dented and seemed to be an unbelievable result. The goal of this 25 Mg 5/2 0.85 10paper is to present the chemistry and physics of these unexpected 26 Mg 0 0 11effects, to formulate a new, ion-radical mechanism of ATP 40 Ca 0 0 96.94synthesis, to consider its energy and kinetic arguments, and to 43 Ca 7/2 1.317 0.135illustrate its biomedical applications based on the first-time 64 Zn 0 0 48.6observed in vivo isotope effect in living organisms. Finally we 66will show that the discovery of the ion-radical mechanism has also Zn 0 0 27.9 67provided a background for the generally accepted but so far Zn 5/2 0.875 4.1 68hypothetical nucleophilic mechanism of ATP synthesis. Zn 0 0 18.82. ISOTOPE EFFECTS IN ATP SYNTHESIS For the first time, heavy isotopes were used for studying ATP-generating enzymes in 1953.13 The 18O/16O isotope exchangebetween 18O inorganic phosphate and 16O water in mitochon-dria was shown to occur only during the ATP functioning; noisotope exchange was observed when the ATP synthesis wassuppressed, particularly by omission of Mg2+ ions. No certainconclusions have been derived from this observation. Whenmitochondria were incubated with Pi labeled with 18O, 33P,and 32P and unlabeled ATP, a fast exchange of Pi oxygen withwater and rapid PiÀATP exchange were detected.14,15 Moreover,the terminal bridge oxygen in ATP formed by oxidative phos-phorylation was shown to come from ADP, not Pi16 (for a reviewof the early studies of isotopes in ATP synthesis, see ref 17).Measuring 18O isotope effects, it was shown that the dissociationof PÀO bond is the rate-limiting step in the Ras-catalyzed Figure 1. Rates of ATP synthesis by creatine kinases as a function of magnesium isotopes. The yield Y is given as radioactivity of [32P]ATPhydrolysis of GTP.18,19 measured as a number of scintillations/min/mg of total amounts of The first example of the magnetic (mass-independent) isotope protein (pure CK); concentration of MgCl2 in all experiments waseffect operating in a biochemical system was the reaction of 15 mM. The lines are drawn for guiding the eyes. Reprinted withmethyl mercury chloride with creatine kinase.20 The effect permission from ref 25. Copyright 2005 Springer.manifests itself both in the nuclear spin dependence of enzymaticactivity and the fractionation of magnetic and nonmagnetic unbelievable result indicated that the presence of magneticmercury isotopes. These observations demonstrate that the isotope nuclei 25Mg in the catalytic site of the enzyme somehowinteraction of methyl mercury chloride, a hazardous environment promotes ATP synthesis. In a series of later studies, this new andpollutant, with creatine kinase is a radical spin-selective process; unpredictable phenomenon was investigated in detail, and thebeing electron spin-selective, it inevitably results in the nuclear spin results are summarized and discussed later.selectivity. However, methyl mercury chloride does not intervenein enzymatic transfer of phosphate; it just chemically modifies the 3.1. Magnesium Isotopesstructure of the site and suppresses the catalytic activity of kinase. Because isotope effects in ATP synthesis are unexpected,This magnetic (mass-independent) isotope effect with particular unusual, and even unbelievable phenomena, it is appropriate toemphasis of its importance in environmental chemistry was later describe in brief all key materials and technologies used in theobserved in many studies (for review, see ref 21). biochemical experiments.24,25 Isotope-containing MgCl2, CaCl2, and ZnCl2 samples (Table 1) were obtained using the treatment3. ISOTOPE EFFECTS IN CREATINE KINASE REACTION of magnesium, calcium, and zinc oxides with analytically pure As has been shown by Ivanov, the reaction of photosynthetic HCl. Monomeric creatine kinase was purified from ViperaCO2 fixation, which is catalyzed by rubiloso-5-bisphosphate xanthia venom, phosphoglycerate kinase was purified from pigcarboxylase/oxidase (RuBisCo), is strongly activated by the skeletal muscle, and pyruvate kinase was isolated and purifiedpresence of 25MgCl2.22 This result stimulated Kouznetsov et al.23 from rabbit muscles.to develop a preparative electrophoretic technique to substitute The catalytic activities of enzymes were measured conven-Mg2+ ions with natural isotope composition by 25Mg2+ ions. By tionally as the amounts of labeled ATP formed in the presence ofusing this technique, the samples of creatine kinase from the [32P]phosphocreatine, [32P]phosphoglycerate, or [32P]phospho-Vipera xanthia venom with high content of 25Mg2+ ions (86% pyruvate in 1 min at optimal incubation conditions and corrected toversus 10% of natural abundance) were obtained. In the experi- 1 mg of pure enzyme tested. For loading or substitution ofments, these samples were shown to exhibit enormously high magnesium, calcium, and zinc ions into the enzyme active sites,ATP-generating activity of enzyme: an 8-fold increase in the the original electrophoretic technique was employed. The proce-share of 25Mg2+ ions in a total pool of magnesium ions was dure of ion substitution as such does not change the enzyme activity;accompanied by a 2.4-fold increase in the ATP yield. This this has been proven in special experiments, when magnesium ions B dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  3. 3. Chemical Reviews REVIEW*Mg2+ were substituted by the *Mg2+ ions of the same composition(*Mg means magnesium with natural abundance of the threeisotopes; see Table 1). Such a replacement has no influence onthe enzymatic activity. For monitoring total ion contents, atomicabsorption spectrometry has been employed. For isotope composi-tion control, isotope mass spectrometry was applied. For analyticalpurposes, some routine high-performance liquid chromatography(HPLC) procedures were used. The enzyme samples were incu-bated at 30 °C for 40 min; this time was required for the ATP yieldto reach an equilibrium mode, i.e., a limiting magnitude in all cases(for all metal ions and all isotopes). Further, the enzymatic activityvalues were determined as mentioned above (for details, see ref 25). Creatine kinases (CKs) loaded with different ions and iso-topes will be further specified as 24Mg-CK, 25Mg-CK, 26Mg-CK,and *Mg-CK, respectively. The first question the authors25 werefaced with was the nature of the observed isotope effect. Toanswer it, a special series of experiments were carried out inwhich CK was loaded with almost pure magnesium isotope ions24 Figure 2. Yield of ATP synthesized by 25Mg-CK (filled circles) and Mg2+ (isotope content 99.6%), 25Mg2+ (95.75%), and 26Mg2+ *Mg-CK (open circles) as a function of MgCl2 concentration. The(96.26%) using identical techniques and conditions of the fraction of 25Mg isotope in 25Mg-CK is 78%. Y is the amount of ATPloading. The results are shown in Figure 1. They unambiguously produced for 40 min and expressed in mM/g of CK.demonstrate that the kinases in which magnesium are substitutedby nonmagnetic isotopic nuclei 24Mg and 26Mg were shown to beidentical in their activity whereas 25Mg-CK is almost twice asefficient. This is convincing evidence that magnetic isotopeeffects do occur in ATP synthesis27À31 while the classical,mass-dependent effect may be ignored. Figure 2 demonstrates the yield of ATP synthesized by 25Mg-CK and *Mg-CK as a function of MgCl2 concentration. Atconcentrations around 10 mM, it exhibits a large isotope effectsimilar to that shown in Figure 1, although experimental techni-ques were different. The loss of activity at high concentrations ofMg2+ ions does not mean that the enzyme is paralyzed. Its activitymay be recovered by removing the metal ions and loading theenzyme repeatedly with the new portions of the ions (seeSupporting Information). Figure 3 shows a linear relationshipbetween the rate of ATP synthesis and the fraction of 25Mg in atotal pool of magnesium ions in CK;32 a similar relationshiop isalso valid for the phosphoglycerate kinase (PGK),33 as will bespecified later.3.2. Calcium and Zinc Isotopes We will further specify CK loaded with calcium and zincisotopes as 43Ca-CK and 40Ca-CK, 64Zn-CK, and 67Zn-CK.Concentration dependences of the ATP yield for 40Ca-CK and43 Ca-CK are shown in Figure 4; they are similar to those for Mg-CK (Figure 2). Again, at a low concentration of CaCl2 there is nodifference in the ATP yield; however, at [CaCl2] g 40 mM, 43Ca- Figure 3. Catalytic activity Y of 25Mg-CK and 25Mg-PGK referred toCK produces ATP more efficiently than 40Ca-CK. In maximum, Y0l the activity of 24Mg-CK and 24Mg-PGK, respectively, as a function ofat [CaCl2] ≈ 120 mM, the isotope effect, i.e., the ratio of the the fraction of 25Mg in a total magnesium pool. Reprinted withyields, is 1.8 ( 0.2. Taking into account that the content of 43Ca permission from ref 33. Copyright 2005 PNAS.in 43Ca-CK is 86.7% only, it is easy to estimate a net, referred to as100% of 43Ca, isotope effect as 2.0 ( 0.2.34 to those for magnesium and calcium CK. First, Zn2+ ions actually ATP synthesis by CK is accompanied by generation of catalyze ATP synthesis, like Ca2+ ions, with the efficiencycreatine and creatinine; they originate from the substrate, comparable with that of Mg2+ ions; second, ATP yield increasesphosphocreatine. The yield of creatine is shown in Figure 5; as concentration of ZnCl2 increases, and then reaches maximumevidently, both the concentration dependence and isotope effect and gradually decreases.35on the creatine yield satisfactorily (not perfectly precise becausethe creatinine is not taken into account) reproduce those for theATP yields (Figure 4).34 4. ARE ISOTOPE EFFECTS TRUSTWORTHY? Figure 6 demonstrates concentration dependence of the ATP Being unexpected and unprecedented, isotope effects meetyield for 64Zn-CK and 67Zn-CK. It exhibits features very similar with distrust (note that the use of Mg, Ca, and Zn isotopes in C dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  4. 4. Chemical Reviews REVIEWFigure 4. Yields of ATP produced by 40Ca-CK (1) and 43Ca-CK (2). Ais the radioactivity of 32P-ATP (in scintillations/min/mg CK). Rep-rinted with permission from ref 34. Copyright 2011 Elsevier. Figure 6. Yield of ATP produced by Zn-CK as a function of 64ZnCl2 (open circles) and 67ZnCl2 (filled circles) concentration. A is the radioactivity of [32P]ATP (in scintillations/min/mg CK).Figure 5. Yield Y of creatine (in mM/min/g of enzyme) produced by40 Ca-CK (1, filled circles) and 43Ca-CK (2, open circles) as a function ofCaCl2 concentration. Reprinted with permission from ref 34. Copyright2011 Elsevier.enzymatic ATP synthesis also has no precedent). To get con-vinced of the reality of the detected isotope effects, it is notenough to claim that all experimental conditions and procedureswere identical. Independent proofs are required. The argumentslisted below will serve this aim. (1) Using different sources of magnesium isotopes, no differ- Figure 7. Rate of ATP synthesis by Mg-PK as a function of MgCl2 ences in enzymatic activity values measured in the pre- concentration. Y is the radioactivity of 32P-ATP (in scintillations/min/ sence of 24Mg and 26Mg were ever found—in contrast to mg PK). The contents of 25Mg2+ are shown in the right upper corner. the experiments involving 25Mg. Reprinted with permission from ref 36. Copyright 2008 American (2) Impurities of metals, other than magnesium, determined Chemical Society. by atomic absorption spectroscopy and electron spectros- copy for chemical analysis (ESCA) in 24Mg-CK, 25Mg-CK, The yields of ATP in identical conditions were estimated as 26 Mg-CK, and *Mg-CK, did not exceed 10À30 ppmm 9 100 ( 200, 20 100 ( 200, 8 900 ( 200, and 12 600 ( and their influence on the ATP yield independently tested 200 scintillations/min/mg for 24Mg-CK, 25Mg-CK, 26Mg- was found to be negligible.36 Among the impurities, the CK, and *Mg-CK, respectively. (The yield of ATP was most suspicious was manganese. It was in fact shown to measured as radioactivity of [32P]ATP generated from the catalyze ATP synthesis (see Supporting Information, [32P]creatine phosphate precursor.) The total (summarized) Table S1) with the efficiency much lower than that of activity of the first three kinases, taken in fractions equivalent magnesium, so that its contribution into the total produc- to those in natural abundance (9 100 Â 0.79 + 20 100 Â tion of ATP is absolutely ignorable when it is presented in 0.10 + 8 900 Â 0.11 = 10 200 ( 400 scintillations/min/mg) tracer amounts. However, even detailed and careful would be expected to coincide with that of *Mg-CK. In fact, it analysis of traces of impurities does not guarantee the is very close to the measured activity of *Mg-CK (12 600 ( absence of unexpected, accidental impurities. 400 scintillations/min/mg).25 (3) To exclude accidental impurities and unpredictable fac- (4) Then the experiment was inverted. In another series, a tors, an independent series of experiments was carried out. mixture of 24MgCl2, 25MgCl2, and 26MgCl2 was taken in D dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  5. 5. Chemical Reviews REVIEW Figure 9. Activity of Mg-PGK as a function of MgCl2 concentration. A stands for radioactivity of [32P]ATP (in scintillations/min/mg PGK). The contents of 25Mg in series 1À6 are equal to 0, 9, 25, 50, 75, and 98%, respectively. The lines are drawn to guide the eyes. Reprinted withFigure 8. Yield of ATP produced by PK as a function of 64ZnCl2 and permission from ref 33. Copyright 2005 American Chemical Society.67 ZnCl2 concentrations. A is the radioactivity of 32P-ATP (in scintilla-tions/min/mg CK). 67Zn-PK contains 78.4% of 67Zn; 64Zn-PK has a and its stability and functioning may be increased by magneticnatural abundance of 67Zn. isotope 25Mg. The functioning of PK at high concentrations of MgCl2 certainly has no importance in terms of biochemistry and proportion of 0.79:0.10:0.11 identical to that in natural physiology; however, it deserves attention as a purely chemical *MgCl2. The yield of ATP produced by CK loaded with phenomenon in which magnetic isotope effect functions. this mixture was measured to be 11 400 ( 400 scintillations/ Figure 8 demonstrates the anomalous behavior of Zn-PK. min/mg, whereas the yield of ATP produced by *Mg-CK Again a nuclear spin dependence of the ATP synthesis exists only was 12 100 ( 400 scintillations/min/mg. These experi- at high concentrations of Zn2+ ions (at 50 mM of ZnCl2 isotope ments leave no doubts in the validity of the detected effect is 2.2 ( 0.3).35 magnesium isotope effect and exclude any possible artifacts. 6. PHOSPHOGLYCERATE KINASE: MAGNESIUM5. PYRUVATE KINASE: MAGNESIUM AND ZINC ISOTOPESISOTOPES Figure 9 shows the dependence of the ATP yield produced by Pyruvate kinase (PK) transfers a phosphate group from phosphoglycerate kinase (PGK) on the MgCl2 concentration.phosphoenolpyruvate to ADP producing ATP and pyruvate The activity of PGK depends on the isotope composition: theaccording to the reaction limiting yield of ATP synthesized by 25Mg-PGK at [MgCl2] g 20 mM is by 1.7 ( 0.2 times higher than that synthesized by 24 Mg-PGK. There is also a linear correlation between the yield of ATP and the share of 25Mg in the total magnesium pool (Figure 3).33 Like in the case of CK, the transfer of the phosphate group ismediated by a magnesium ion. The rate of ATP synthesis by PK 7. MITOCHONDRIA: MAGNESIUM AND ZINC ISOTOPESas a function of magnesium concentration is shown in Figure 7.36 Myocardial mitochondria were tested in vitro. To estimate the It exhibits two remarkable features. First, its dependence on contribution of oxidative phosphorylation to the total ATP yield,the concentration reveals two maxima; second, it demonstrates methyl nicotine amide (MNA) or KCN were used. Actingan unusual dependence on the magnesium isotope composition. differently, both inhibitors are nonetheless known to suppressAt low concentrations of Mg2+ ions (10À50 mM), there is no the activity of ATP synthase. The rates of ATP production bydependence on the isotopes. At high concentrations (100À isolated mitochondria (Figure 10) incubated in media containing300 mM), Mg-PK demonstrates an enormously strong nuclear separately ions 24Mg2+ (99.6% of 24Mg), 25Mg2+ (95.8% of 25spin dependence of the ATP yield. Step-by-step replacement of Mg), and 26Mg2+ (96.3% of 26Mg) strongly depend on thespinless 24,26Mg2+ ions in the catalytic sites by nuclear magnetic nuclear spin and magnetic moment: mitochondria with magnetic25 Mg2+ ions gradually increases the ATP yield (Figure 7). Among nuclei 25Mg produce twice as much ATP as mitochondria withother enzymes, PK demonstrates a very specific behavior. It is not spinless, nonmagnetic nuclei 24,26Mg.25,37 When oxidative ATPexcluded that, at high concentration of MgCl2, its structure is synthesis is selectively blocked by treatment with MNA, the yieldsomehow disturbed; nevertheless, it continues to function. of ATP strongly decreases. The remaining part refers to the ATPMoreover, the higher the content of 25Mg, the higher is the produced by mitochondrial creatine kinase (m-CK), which alsoextent of its survival. It means that PK is a strongly sustained enzyme demonstrates an isotope effect: 25 Mg2+ ions are about twice more E dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  6. 6. Chemical Reviews REVIEW Table 2. Rate of ATP Synthesis Catalyzed by MgCl2 in Mitochondriaa MgCl2 Y Â 103 MgCl2 (20 mM, 10% 25Mg)a 6.7 25 MgCl2 (20 mM, 92% 25Mg) 9.8 25 MgCl2 (20 mM, 95.8% 25Mg) 11.1 a Y is the yield of ATP in μm/min/g of protein; the accuracy is 10%.Figure 10. Rates of ATP synthesis by mitochondria as a function ofmagnesium isotope. 1, intact mitochondria; 2, mitochondria subjectedto selective blockade of oxidative phosphorylation by MNA. The yield ofATP is given in μmol/min/g of total amounts of protein; concentrationof MgCl2 in all experiments was 15 mM, and temperature was 37 °C.The lines are drawn to guide the eyes. Reprinted with permission fromref 25. Copyright 2005 Springer.active than 24Mg2+ or 26Mg2+ ions. There is no difference between24 Mg2+ and 26Mg2+ ions in both oxidative and substrate ATPsynthesis. Moreover, the activity of m-CK was shown to be almostidentical to that of creatine kinase isolated from Vipera xanthiavenom.25 These observations demonstrate that enzymatic ATPsynthesis in mitochondria is a spin-selective process exhibitingmagnetic isotope effect. Figure 11. Rates of ATP synthesis by m-CK in mitochondria as a Additional experiments were carried out to compare the ATP function of zinc isotopes: 64ZnCl2 (open circles) and 67ZnCl2 (filledformation capabilities of mitochondria with natural magnesium circles). A is the radioactivity of ATP (in scintillations/min/mg of(*MgCl2) and with magnesium enriched with 25Mg (25MgCl2) protein).(Table 2). They demonstrate high efficiency of mitochondrial25 Mg-ATP synthase, proportional to the content of 25MgCl2. magnetic moments of the metal ions: ATP synthase and kinases, Rat heart muscle mitochondrial m-CK was tested with in which metal ions have magnetic nuclei, are found to be by 2À3[32P]phosphocreatine and ZnCl2 of different isotope composi- times more efficient than enzymes with spinless, nonmagnetictions. After 60 min incubation of mitochondria, the total ATP isotopic nuclei. This phenomenon clearly and unambiguouslyyield and oxygen consumption were measured. The ATP yield as manifests that the ATP synthesis is a process in which the mass-a function of ZnCl2 concentration is shown in Figure 11. It exhibits independent, magnetic isotope effect operates rather than thethe following features: First, Zn2+ ions were indeed found to classical, mass-dependent one.31catalyze mitochondrial ATP synthesis with the efficiency compar- The nuclear spin-dependence of ATP synthesis is a new,able with that of Mg2+ ions. Second, the ATP yield increases as the paradigm-shifting phenomenon. It reliably indicates that theconcentration of ZnCl2 increases, reaches maximum at ∼5 mM, mechanism of the ATP synthesis includes paramagnetic inter-and then decreases. Third, even at low concentration of ZnCl2 mediates and that, besides the generally accepted nucleophilic(∼1 mM), there is a large isotope effect in the ATP synthesis. mechanism, there is also another mechanism that should be inFourth, there is no isotope effect in the oxygen consumption of accordance with the following physical postulates:mitochondria. The latter means that m-CK in mitochondria (1) ATP synthesis is a radical or, more exactly, an ion-radicalfunctions independently of the ATP synthase.33 process. It is necessary to note that the concentration of Mg2+ and Zn2+ions in the catalytic sites of mitochondria does not correspond to (2) The starting, key reaction of the process is an electronthat in the reaction media because the distribution of ions between transfer from ADP to the M2+ ion, which generates anthe reaction media and mitochondria is not known. However, the ion-radical pair composed of an M+ ion and phosphatedistribution of isotopes in catalytic sites perfectly corresponds to that radical-anion of ADP as the partners; here M is Mg,in the reaction media because neither binding of ions nor their Ca, Zn.diffusion is supposed to depend on the isotopes. At least they cannot (3) Because of spin conservation, the chemical reactivity ofaccount for the experimentally observed isotope effects. triplet and singlet spin states of the ion-radical pair is different. This leads to a different yield of ATP along the singlet and triplet phosphorylation channels.8. NUCLEAR SPIN-DEPENDENT CHEMISTRY OF ATP (4) The relative contribution of these spin channels to theSYNTHESIS ATP yield is controlled by electronÀnuclear (hyperfine) To summarize what has been presented so far, the rate of magnetic coupling of unpaired electrons with the mag-enzymatic ATP synthesis strongly depends on the nuclear netic nucleus 25Mg (43Ca, 67Zn) in the Mg+ (Ca+, Zn+) F dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  7. 7. Chemical Reviews REVIEWScheme 1. Mechanism of ATP Synthesis by ATP Synthase Scheme 2. Mechanism of ATP Synthesis by Ca-CK ion and with 31P in the ADP phosphate radical. It induces singletÀtriplet spin conversion and results in a nuclear fast tripletÀsinglet conversion due to electron spin relaxation (in spin-dependence of the ATP synthesis. OH the spin relaxation time is ∼10À11 s) and again regenerates A reaction scheme was suggested to account for this unusual Mg2+, Ca2+, and Zn2+ ions in the reactions identical to (I).but firmly established phenomenon.36 Schemes 1 and 2 illustrate Schemes 1 and 2 elucidate nuclear spin control of the ATPsuch a mechanism for particular cases of ATP synthase (oxidative yield as the main feature of ATP synthesis. However, they are notphosphorylation) and CK (substrate phosphorylation). The adequate to real enzymatic chemistry as long as they do notschemes are exemplified by magnesium and calcium ions but reflect the role of metal ions as catalysts. The following will beare valid also for zinc as well as for other enzymes. As a first step, exemplified by the case of magnesium ions. Numerous studiesScheme 1 implies transfer of an electron from the terminal convincingly demonstrate that the functioning of protein kinasephosphate group of ADP to Mg2+ ion, which generates the requires at least two Mg2+ ions in the catalytic site.8,38À40 PGK wasprimary ion-radical pair, composed of the radical-cation Mg+ and shown to function only if two Mg2+ ions are presented in theoxyradical of ADP (reaction 1). Because of the total spin catalytic site; one of them is tightly bound in the MgADP complex,conservation in the process, this pair is in a singlet spin state. the other is “free”, not bound with phosphate groups, but solvatedThe next step is the phosphorylation itself in which the ADP by water molecules and protein groups.41 Even RNA synthesis byoxyradical attacks the PdO chemical bond of inorganic phos- RNA polymerase requires assistance of the two Mg2+ ions.42phate (reaction 2). The oxyradical that has been additionally Moreover, detailed kinetic studies of ATP synthesis by ATPgenerated in this reaction decomposes via β-scission of the PÀO synthase convincingly prove that the presence of at least twochemical bond (reaction 3), generating ATP and a final ion- magnesium ions is indeed required for ATP synthesis to occur: oneradical pair (HO Mg+); the latter regenerates Mg2+ in the is tightly bound to ADP, the other is present as a free but hydratedreaction of back electron transfer: complex Mg(H2O)n2+.43 Thus, in the modified Scheme 1, the Hþ starting electron transfer reaction 1 should be written as follows: ðHO Mgþ Þ s H2 O þ Mg2þ f ðIÞ MgðH2 OÞn 2þ þ MgðH2 OÞm 2þ ðADP3À Þ f MgðH2 OÞn þ The rate of phosphorylation along the singlet channel(reactions 1-3 in Schemes 1and 2) is suppressed by spin-allowed þ MgðH2 OÞm 2þ ðADP2À Þ ð1Þreverse electron transfer in the primary ion-radical pair, whichregenerates the starting reactants and decreases the ATP yield.However, in the presence of 25Mg2+, 43Ca2+, and 67Zn2+ ions, MgðH2 OÞn 2þ þ MgðH2 OÞm 2þ ðADP2À Þ f MgðH2 OÞn þhyperfine coupling of the unpaired electron with the magnetic þ MgðH2 OÞm 2þ ðADPÀ Þ ð2Þnuclei 25Mg, 43Ca, and 67Zn in the Mg+, Ca+, and Zn+ ionsstimulates singletÀtriplet spin conversion of the primary ion- They imply that the MgADP complexes are also hydrated by mradical pair and transforms it into a triplet pair, in which back water molecules (their structures and energies will be discussedelectron transfer is spin-forbidden. This new triplet channel of later). Reaction 1 refers to fully deprotonated ADP, and reactionphosphorylation (reactions 20 and 30 ) provides an additional 2 refers to partly protonated ADP because at physiological pHyield of ATP that increases the total production of ATP by 2À3 in cells and mitochondria both forms of ADP are supposed totimes. The final ion-radical pair in the triplet channel undergoes be present almost equally. These two reactions generate an G dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  8. 8. Chemical Reviews REVIEWion-radical pair composed of Mg(H2O)n+ ion and the phosphateradical anion of ADP as partners. Another important consequence in the case of CK is theformation of creatine and creatinine from the substrate residue.The yield of ATP was shown to coincide with the total yield ofcreatine and creatinine.32 It means that every chemical act ofADP phosphorylation quantitatively transforms the creatineresidue into either creatine or creatinine and the yield of creatineis expected to exhibit nuclear spin dependence identical to thatfor the ATP yield. This prediction is in a perfect agreement withexperiment (see Figures 4 and 5). Note that only the ion-radicalmechanism is able to properly explain isotope as well as magneticfield effects.9. IN VIVO ISOTOPE EFFECTS AND MEDICINE The great advantage of the ion-radical mechanism is that it canbe switched on artificially by insertion of MgCl2 (or, even better,of 25MgCl2) in excess to stimulate ATP synthesis and hence to Figure 12. Recovery degree RD of ATP production in living rats as aprevent some pathological disorders related to the deficiency of function of delivered amounts of Mg2+ ions.ATP in vivo such as ischemia caused local tissue hypoxia cases,heart muscle metabolic disbalance of all sorts, drug and toxinscardiotropic side effects, electrical trauma consequences, inhala-tion deplete hypoxia, etc. For these specific purposes, a magne-sium ion carrier (nanocationite) based on some new porphyrinadduct of fullerene-C60 (PMC16) (see Supporting Information,Figure 1S) has been designed.44 This novel product was found tobe suitable for the targeted delivery of Mg2+ nanoamountstoward the heart muscle. This nanocarrier gets accumulatedpredominantly in the heart muscle because myocardial cells havea high-affinity protein receptor to porphyrin residue of PMC16situated in the external membrane of myocardiocyte mitochon-dria. Another attractive property of PMC16 as a potential drug isits day-long retention in the heart tissue, particularly inside theheart mitochondria membranes.45,46 Being membranotropiccationites, these “smart” nanocontainers release Mg2+ ions onlyin response to the metabolic acidosis, i.e., to the slight acidic pH Figure 13. Rate W of ATP synthesis by 24Mg-CK (1) and by 25Mg-CKshift known as a direct and inevitable consequence of a tissue (2) as a function of the magnetic field. The rate was measured ashypoxia of any sort, but take them back once the cell metabolism radioactivity of [32P]ATP. The lines are drawn to guide the eye.gets recovered. Reprinted with permission from ref 12. Copyright 2008 American The ability to stimulate ATP synthesis in the heart muscle of Chemical Society.living rats was demonstrated by in vivo experiments. Injection ofdoxorubicin was to promote a significant damage to the local of cell/tissue hypoxia and some heart pathologies related to ATPmyocardial ATP synthesis, suppressing it by ∼70%. Then the deficiency. PMC16 is shown to be a safe, easy-to-eliminate, andPCM16 loaded with 24MgCl2 or 25MgCl2 was injected the same efficient medicinal agent with a prolonged pharmacologicalway, and after that a recovery of the ATP production up to the activity and with no sign of harmful metabolites formed.initial, predoxorubicin, level was observed. The extent of recov-ery as a function of magnesium concentration is shown inFigure 12. Evidently, there is a large isotope effect: delivery of 10. MAGNETIC FIELD EFFECTS[25Mg]PMC stimulates ATP synthesis by 2À3 times more According to the ion-radical mechanism, the enzyme catalyticefficiently than [24Mg]PMC. This is the first report ever on the site is a nuclear spin-dependent nanoreactor with the twoisotope effect manifesting itself in a living organism and hence competing reaction channels, singlet and triplet ones. Theirshowing the potential of this specific effect to be applied as an relative contributions to the ATP synthesis are controlled byefficient remedy for the treatment of heart disease—particularly hyperfine coupling of unpaired electrons with magnetic nucleifor minimizing some known drug cardiotoxic side effects.47À52 25 Mg, 43Ca, 67Zn, and 31P. However, spin conversion of the ion- The [25Mg2+]PMC16 nanocationites were carefully tested radical pair is controlled not only by internal magnetic fieldsin vivo to estimate all major pharmacokinetics and pharmacody- (hyperfine coupling) of the magnetic nuclei but also by annamics patterns in a variety of mammals including mice, rats, external magnetic field. To verify this statement, the yield ofrabbits, dogs, and goats. One may conclude that the stimulation ATP synthesized by Mg-CK as a function of an external magneticof ATP synthesis by a magnetic isotope of magnesium is a new field was studied.12 The yield of ATP produced by enzyme withphenomenon of fundamental biomedical importance; it repre- 24 Mg2+ ions was shown to decrease by ∼10% in magnetic fieldsents a breakthrough in the design of new remedies for treatment 80 mT (Figure 13), whereas for enzyme with 25Mg2+ ions it H dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  9. 9. Chemical Reviews REVIEWChart 1. Magnesium Complexes of the Pyrophosphates Chart 2. Magnesium Complexes of Protonated Pyrophosphatesincreases by 50 and 70% in the fields 55 and 80 mT, respectively. Chart 2, also have been calculated, as well as their ion-radicals 3aIn the Earth field, the rate of ATP synthesis by enzyme with and 4a generated from 3 and 4 by electron detachment.5825 Mg2+ ions is 2.5 times higher than that by enzymes with Typical structures for selected values of m are shown innonmagnetic nuclei 24Mg or 26Mg. Both magnetic field and Figure 2S (Supporting Information). The withdrawal of elec-magnetic isotope effects are in perfect agreement with the tron from complexes 1À4 results in redistribution of thephysical postulate12 and demonstrate that the ATP synthesis is electron density and slightly changes interatomic distances inan ion-radical process affected by Zeeman interaction and complexes 1aÀ4a with respect to those in 1À4. The structureshyperfine coupling in the intermediate ion-radical pair. The of calcium and zinc complexes were shown to be identical toexperimental results are in perfect agreement with theoretical those of magnesium34,59 (see Figures 3S and 4S). The energiespredictions.53,54 of the hydrated complexes M(H2O)n2+ and M(H2O)n+ as a function of n, the number of water molecules in coordination11. ENERGETICS OF ATP SYNTHESIS sphere of ions, also have been calculated58,59 at the B3LYP/6- 31G* level of DFT theory55,56 (see Figure 5S in Supporting Schemes 1 and 2 explain isotope and magnetic field effects in Information).ATP synthesis. However, the intriguing problem is that the key Knowing the energies of individual complexes, one canreaction 1 of ATP synthesis shown in Schemes 1 and 2 occurs calculate energies of the electron transfer from hydrated pyro-only in enzymes and does not take place in water. phosphate complexes 1À4 shown in Charts 1 and 2 to theTo answer this question, it is necessary to know the structure complexes Mg(H2O)n2+ as a function of n and m:of the metal complexes and the energies of their reactions. Theywere calculated using density functional theory (DFT).55À5711.1. Complexes of Metal Ions and their Reactions MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðOPO2 OPO3 CH3 Þ3À In the catalytic site, ADP is present as complex M2+(ADP); we f MðH2 OÞn þ þ MðH2 OÞm 2þ ðOPO2 OPO3 CH3 Þ2Àwill specify the metal as M where M implies Mg, Ca, or Zn.Further, its structure is modeled by the hydrated pyrophos- ð1aÞphate complexes M(H2O)m2+(HP2O73À) and M(H2O)m2+-(CH3P2O73À) with a hydrogen atom and a methyl group, MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðOPO2 OPO3 CH3 Þ3Àrespectively, instead of an adenosine residue. Chart 1 illustratestheir structure by examples of magnesium complexes 1 and 2 in f MðH2 OÞn þ þ MðH2 OÞm 2þ ðOPO2 OPO3 CH3 Þ2Àwhich a Mg2+ ion is supposed to be coordinated to the oxygen ð2aÞatoms of the pyrophosphate anion and accepts two coordinatebonds. The other coordinate bonds may be used for addition ofm water molecules, with m being in the range from 0 to 4. The MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðHOPO2 OPO3 CH3 Þ2Àvalue m = 4 corresponds to the fully completed six-coordinated f MðH2 OÞn þ þ MðH2 OÞm 2þ ðHOPO2 OPO3 CH3 ÞÀshell of the Mg2+ ion. Besides complexes 1 and 2, ion-radicals 1aand 2a (Chart 1) generated from 1 and 2 by electron detachment ð3Þwere also calculated. In complexes 1 and 2, 1a and 2a thepyrophosphate anions are deprotonated. However, in cells and MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðHOPO2 OPO3 CH3 Þ2Àmitochondria ADP is presented in both forms—deprotonated f MðH2 OÞn þ þ MðH2 OÞm 2þ ðHOPO2 OPO3 CH3 ÞÀand partly protonated—almost equally, on a par so that thestructures of partly protonated complexes 3 and 4, presented in ð4Þ I dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  10. 10. Chemical Reviews REVIEW Figure 15. Energies of reactions 3 and 4 as a function of the number of water molecules n in the Ca(H2O)2+n ions for different m, the number of water molecules in Ca(H2O)m2+(HOPO2OPO3CH3)2À (curve 1) and Ca(H2O)m2+(OPO2OPO3CH3)3À (curve 2). Reprinted with permis-Figure 14. Energies of reactions 1 (curve 1) and 2 (curve 2) as a sion from ref 59. Copyright 2010 Russian Academy of Science.function of the number of water molecules n in hydrated complexMg(H2O)2+n. Reprinted with permission from ref 58. Copyright 2010 that these energies may be certainly attributed to theAmerican Chemical Society. reactions of Mg2+(ADP) in native enzymes. These reactions model reaction 1 in Schemes 1 and 2 as (2) The energies depend only slightly on m (m = 0À4), thestarting reaction of electron transfer in the ATP synthesis: number of water molecules in the first coordination sphere of ions M(H2O)2+m(pyrophosphate) and M- MðH2 OÞn 2þ þ M2þ ðADPÞ3À f MðH2 OÞn þ (H2O)+m(pyrophosphate); therefore, they are almost independent of the hydration of the Mg2+(ADP) complex þ M2þ ðADPÞ2À ð5Þ in the catalytic site. (3) The energies of all reactions are strongly dependent on n, 2þ 2À þ the number of water molecules in the hydration shell of MðH2 OÞn þ M ðADPÞ 2þ f MðH2 OÞn M(H2O)n2+ and M(H2O)n+ (see Figures 14À16). þ M2þ ðADPÞÀ ð6Þ (4) No reaction occurs in water, when the hydrate shells of The total energy for reactions 1À4 is defined as the difference the reactants are fully completed (n = ∞). This result is inbetween the energy of products and reactants. The reaction is perfect agreement with the fact that ATP synthesis fromexoergic if the energy of reactants exceeds that of products. On ADP and substrate does not take place in water.the contrary, the reaction is endoergic and energy-forbidden if (5) For ATP synthesis to proceed in the catalytic site, it isthe energy of products is higher than that of reactants. Such an necessary to partly remove water from the site andapproach has an important advantage because it cancels any hydration shell of M(H2O)n2+. Only under this conditionerrors and compensates possible inaccuracy in energy calcula- the key reaction—electron transfer from M2+(ADP) totions of individual reactants and products. M(H2O)n2+, reaction 1 in Schemes 1 and 2—becomes The energies of reactions 1À4 for magnesium are shown in energy-allowed.Figure 14 as a function of n. For n from 0 to 6, they were (6) The reaction of complexes 1 and 2 with Mg(H2O)n2+ iscalculated by DFT theory; the energies of reactions in water, i.e., exoergic at n = 0À6 and even at n > 6 (Figure 14). Theunder conditions of the total hydration of magnesium ion (we switching from exoergic to endoergic reaction takes placewill assume for this case n = ∞), were derived from the at n*, which lies somewhere between n = 6 and n = ∞.thermodynamics of the Mg(H2O)n2+ and Mg(H2O)n+ ions.36 This magnitude 6 < n* < ∞ functions as a trigger; itFigures 15 and 16 demonstrate energy of electron transfer as a determines the boundary between energy-allowed andfunction of n for calcium and zinc, respectively.59 energy-forbidden regimes, and the boundary is overcome11.2. Energy-Forbidden and Energy-Allowed Electron as the compression of the catalytic site squeezes waterTransfer molecules out of the site and partly removes it from the The total energies of the electron transfer reactions presented Mg(H2O)n2+ ion. The same is valid for calcium and zincin Figures 14À16 as functions of m and n result in the following ions (Figures 15 and 16).conclusions. (7) For the reactions of complexes 3 and 4, which include (1) The energies do not depend on whether an adenosine protonated pyrophosphate anions HP2O7H2À and residue is replaced by hydrogen atom (in complexes 1 and CH3P2O7H2À, with Mg(H2O)n2+ the trigger value of n* 3) or by a methyl group (in complexes 2 and 4). It means is estimated quantitatively as n* = 4 (Figure 11); at n > 4 J dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  11. 11. Chemical Reviews REVIEW However, the remarkable property of the enzymes is that, in the reactive state, when the enzyme domains are drawn together to unite substrate and ADP, they squeeze water molecules out of the catalytic site60 and partly dehydrate the M(H2O)n2+ ion.61 Now the question is what complexes of ADP are responsible for the ATP synthesis, fully deprotonated Mg2+(ADP)3À or partly protonated Mg2+(ADP)2À. As follows from the energy consideration, it depends on the state of the partners, the Mg(H2O)n2+ complexes. The binding energies of the outer, sixth water molecules in the complexes Mg(H2O)62+, Ca- (H2O)62+, and Zn(H2O)62+ were calculated to be 24.5, 24.7, and 21.8 kcal/mol, respectively.62 The energies of releasing the fifth water molecules are slightly larger, À28.0, 27.7, and 24.0 kcal/mol for the same complexes, respectively. To activate M(H2O)n2+ complexes, i.e., to increase their electron affinity and to reach the threshold values of n*, it is necessary to release at least two water molecules, the sixth and fifth. The total energy cost of this “undressing” process is ∼53À46 kcal/mol for M = Mg, Ca, Zn.62 These values are rather high, so that the escape ofFigure 16. Energies of reactions 3 and 4 as a function of the number of coordinated water from the hydrated shell of ions seems to be thewater molecules n in the Zn(H2O)2+n ion for different m, the number of most energy-expensive process in the ion-radical mechanism. Itwater molecules in Zn(H2O)m2+(HOPO2OPO3CH3)2À (curve 1) and means that the protonated complexes Mg2+(ADP)2À are hardlyZn(H2O)m2+(OPO2OPO3CH3)3À (curve 2). Reprinted with permis- involved in the ATP synthesis. The main sources of ATP seem tosion from ref 59. Copyright 2010 Russian Academy of Science. be complexes Mg2+(ADP)3À, i.e., complexes of the deprotonated ADP, for which n* . 6. For these complexes to react there is no the reactions are endoergic, at n < 4 they are energy- need to remove tightly bound water from the first coordination allowed. For the calcium ions the trigger value is n* = 3; shell of Mg(H2O)62+. The release of the outer, weakly bound for zinc ions it is n* = 4 (Figures 15 and 16). water of the third or even fourth coordination sphere is enough to activate the complex Mg(H2O)n2+ as an electron acceptor and12. HOW DO ATP-SYNTHESIZING ENZYMES stimulate ATP synthesis.FUNCTION? The removal of water from the coordination sphere of a metal Despite the difference in enzyme mechanics (ATP synthase is ion activates it by increasing its electron affinity, so that at somea rotary molecular motor, kinases are molecular pumps), a threshold value n* electron transfer becomes exoergic and energygeneral principle of the functioning of the ATP-generating allowed. The water molecule with number n* in complex M-enzymes can be formulated as follows. Theoretical inspection (H2O)n2+ functions as a trigger; it switches the reaction betweenof the reactions between pyrophosphate and hydrated complexes endoergic and exoergic regimes. At n > n* electron transfer isreveals their general properties. These reactions perfectly model endoergic; at n < n* it is energy-allowed and exoergic (seethose in catalytic sites of enzymes: Figures 14À16). Note that energy of electron transfer is inde- pendent of the m in the M(H2O)2+m(pyrophosphate) complex; MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðADP3À Þ f MðH2 OÞn þ it is almost identical for m = 0À4. A huge isotope effect in the rate of ATP synthesis is a direct þ MðH2 OÞm 2þ ðADP2À Þ and irrefutable proof of the proposed mechanism. It is evidence that the M2+ ion is a central reactant, which functions reversibly MðH2 OÞn 2þ þ MðH2 OÞm 2þ ðADP2À Þ f MðH2 OÞn þ and transforms mechanical energy of the enzyme into the energy of the PÀO bond in ATP.58,61 It is a point where mechanics þ MðH2 OÞm 2þ ðADPÀ Þ meets chemistry, the key point of the functioning of enzymes as mechano-chemical molecular motors, where energy of confor-Here M = Mg, Ca, or Zn. These reactions generate ion-radical mational mechanics of the enzyme macromolecule transformspairs in which subsequent reactions 20 and 30 in Schemes 1 and 2, into the energy of a chemical bond.similar to reactions 2 and 3, accomplish ATP synthesis. The modified scheme implies the presence of at least twomagnesium ions in the catalytic site: one is tightly bound to ADP, 13. MAGNETIC PARAMETERS AND SPIN DYNAMICS OFthe other is present as a hydrated complex M(H2O)n2+. Detailed THE ION-RADICAL PAIRkinetic studies of the ATP synthesis by ATP synthase convin- Electron transfer generates an ion-radical pair composed ofcingly prove that the presence of two magnesium ions is indeed paramagnetic complexes M2+(ADP)2À and M(H2O)n+. Therequired for the ATP synthesis to occur: one of them is bound contributions of singlet and triplet spin states of the pair toADP, the other belongs to hydrated ion.37À43 This experimental ATP production are different because the reverse electronresult is in perfect agreement with our theoretically derived transfer regenerates the starting reactants; it is spin-allowed inconclusions. the singlet state but spin-forbidden in the triplet state. The The ion-radical mechanism presented by Schemes 1 and 2 populations of these states and the rates of singletÀtriplet spinseems to be unbelievable, because electron transfer (reaction 1) conversion are controlled by magnetic parameters, g-factors, anddoes not occur in water where metal ions are highly hydrated. hyperfine coupling (HFC) of the unpaired electrons with K dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  12. 12. Chemical Reviews REVIEW Scheme 3. Kinetic Scheme of Spin-Dependent ATP SynthesisFigure 17. Spin densities (green color) in pyrophosphate radical(HOPO2OPO3)2À (a) and Mg(H2O)42+(HOPO2OPO3)2À complex(b). Reprinted with permission from ref 63. Copyright 2009 Elsevier. nuclei 43Ca. No doubt, the same conclusions are valid for the zinc-induced ATP synthesis.magnetic nuclei 31P and 25 Mg, 43Ca, 67Zn in the paramagneticcomplexes, the partners of the ion-radical pair. 14. KINETICS AND QUANTITATIVE MEASURE OF ISO- The magnetic parameters (g-factors and HFC constants TOPE EFFECTSa(25Mg), a(31P)) of the paramagnetic pyrophosphate complexes In order to quantitatively determine the isotope effect, oneMg(H2O)m2+(HOPO2OPO3)2À, which model those in the would need to analyze a detailed kinetic scheme of reactions incatalytic site of enzymes, are identical to the parameters of the catalytic site equivalent to chemical Schemes 1 and 2, as seen inpyrophosphate radical.63 It means that the pyrophosphate ligand Scheme 3:in the complex is present as a pyrophosphate radical, in which It implies generation of the ion-radical pair S in the singletspin density is almost completely localized on the terminal state with a rate constant k1; this pair may be annihilated via backoxygen atom and only slightly propagates to the Mg(H2O)m2+ electron transfer (rate constant kÀ1) or produce ATP with a ratefragment. This conclusion is illustrated in Figure 17 by direct constant k. In the presence of a 25Mg2+ ion, it may experiencecomparison of spin densities calculated for isolated pyrophos- reversible spin conversion to the triplet state T and back with thephate radical and Mg(H2O)42+(HOPO2OPO3)2À complex. rate constants kST and kTS, respectively. Triplet state T is The dominating contribution to the spin conversion of the supposed to generate ATP with the same rate constant k as theion-radical pair results from the Mg(H2O)n+ ions. The HFC singlet pair because reactions 2, 20 and 3, 30 in Schemes 1 and 2constants for these ions are rather large; they are in the limits of are not spin-dependent.180À40 G for n in the range from 0 to 463 (see Figure 6S in It is easy to show that kST = kTS despite the fact that the tripletSupporting Information) and responsible for the fast spin con- state is composed of the three spin substates, T+, T0, and TÀ.version. The hyperfine coupling constant for 31P in the oxyradical Taking into account these substates, the rate constants kST andis ∼45 G so that the dominating contribution of the triplet kTS may be presented aschannel into the ATP production is provided by catalytic siteswith 25Mg. kST ¼ k0 þ þ k0 0 þ k0 À ST ST ST Similar conclusions were derived for the calcium pyrophos- kTS ¼ k0 þ S þ k0 0 S þ k0 À S T T Tphate complexes Ca(H2O)2+m(HOPO2OPO3CH3)À.32 Theirg-factors are almost independent the number of water molecules Here the rate constants with superscripts zero characterize them in the coordination sphere of the calcium ion; their magnitudes rate constants of SÀT and TÀS spin conversion along theare in the limits 2.013À2.015 and identical to the g-factor of the individual spin channels. Because k0 +S = kST+, k0 0S = kST0 , T 0 T 0pyrophosphate radical HOPO2OPO3CH3À. The HFC constants 0 0 kTÀS = kSTÀ, it immediately follows that kST = kTS. The physicala(43Ca) were calculated to be small: 1.1 and 1.0 G for m = 0 and reason is that the spin conversion occurs along the three channels1, respectively, and negligibly small for m = 2À4 (see Figure 7S in with equal rates in both directions.Supporting Information). No doubts that they contribute almost The kinetic scheme is described by a system of kineticnothing into the spin conversion of the ion-radical pair as well as equations:the isotope effect. The g-factors as well as the HFC constantsa(43Ca) in the Ca(H2O)2+m(HOPO2OPO3CH3)À complexes d½AŠ ¼ k1 ½AŠ À kÀ1 ½SŠ ð1aaÞ(Table 2S in Supporting Information) unambiguously indicate dtthat the pyrophosphate ligand in the complex is present as apyrophosphate radical, in which spin density is almost comple- d½SŠ ¼ k1 ½AŠ þ kST ½TŠ À ðkÀ1 þ k þ kST Þ½SŠ ¼ 0 ð2aaÞtely localized on the terminal oxygen atom and only slightly dtpropagates to the Ca(H2O)m2+ fragment, like in the similarparamagnetic complexes Mg(H2O)m2+(HOPO2OPO3)2À. It d½TŠunambiguously demonstrates that the terminal oxygen atom of ¼ kST ½SŠ À ðk þ kST Þ½TŠ ¼ 0 ð3aÞ dtthe pyrophosphate ligand in the complex Ca(H2O)2+m-(HOPO2OPO3CH3)2À donates the electron from its lone pair d½ATPŠto the Ca(H2O)n2+ ion. ¼ kð½SŠ þ ½TŠÞ ð4aÞ In contrast to the pyrophosphate complexes, the HFC con- dtstants a(43Ca) in Ca(H2O)n+ ions are in the limits of 270À50 G where [A], [S], [T], and [ATP] are concentrations of thefor n in the range from 0 to 5,32 i.e., they ensure fast singletÀ catalytic sites, singlet and triplet pairs, and ATP, respectively.triplet spin conversion of the ion-radical pairs with magnetic For the short-lived S and T pairs, the well-known method of L dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  13. 13. Chemical Reviews REVIEWquasi-stationary concentrations may be used, which implies that Here α is the fraction of catalytic sites occupied by 25Mg2+. If only 24,26stationary concentrations of the pairs do not change, so that their Mg2+ ions are presented in catalytic sites (α = 0), then Y0 =derivatives with respect to time can be set at zero. This is a usual, kn [A] andwell-verified, and reliable approximation that gives accurate results for both chemical and physical kinetics. Υ km ¼α À1 þ 1 ð11Þ The solution of the system of eqs 1 À 4 results in the following Υ0 knrate of ATP synthesis: Comparing this ratio with the experimentally estimated d½ATPŠ kk1 ð2kST þ kÞ ¼ ½AŠ dependence Y/Y0 = f(α) (Figure 3), the ratio km/kn for PGK dt m ðkST þ kÀ1 þ kÞðkST þ kÞ À kST 2 is found to be equal to 1.5, i.e., the rate of ATP synthesis is 1.5 ð5aÞ times higher in 25Mg-PGK enzyme with respect to the same enzyme with a nonmagnetic isotope 24Mg. For CK it is evenwith the rate constant higher and equal to 2.6. kk1 ð2kST þ kÞ km ¼ ð6aÞ 15. RATE-LIMITING STEP OF ATP SYNTHESIS ðkÀ1 þ k þ kST ÞðkST þ kÞ À kST 2 The isotope effect seems to be the most convincing evidence It is specified as the rate constant in the site with magnetic that the limiting step in ATP synthesis is the reaction itselfnucleus 25Mg (subscript m). For the catalytic sites with non- because neither entering of reactants into the catalytic site normagnetic nuclei 24Mg or 26Mg, kST = 0 and the rate constant kn release of ATP are expected to be dependent on the isotope(subscript n) of ATP synthesis in these sites follows from eq 6: substitution. However this at a first glance quite evident conclu- kk1 sion is not correct. The catalytic site as a dynamic system may kn ¼ ð7Þ be subdivided into two subsystems, which can be conventionally kÀ1 þ k specified as a fast and a slow one. The former stimulates theTheir ratio represents the isotope effect chemical transformations shown in Schemes 1 and 2 and requires short-range shifts and displacements of reactants to unite them km ðkÀ1 þ kÞð2kST þ kÞ and accomplish chemical synthesis. The latter includes long- IE ¼ ¼ ð8Þ kn ðkÀ1 þ k þ kST ÞðkST þ kÞ À kST 2 range displacements of protein domains, which maintain enter- ing of reactants and release of ATP. In contrast to the short-range Its magnitude is controlled by competition of the three subsystem that functions on the time scale of nanoseconds, theelementary reactions in catalytic site: annihilation (kÀ1), ATP long-range subsystem operates on the time scale of micro- orsynthesis (k), and spin conversion (kST). even possibly of milliseconds. The former may be denoted the For a quantitative estimation of the rate constants, the nanosystem, the latter the microsystem.dependence of IE on the dimensionless ratios kÀ1/k and kST/k If the reaction, controlled by the nanosystem, is irreversiblewas analyzed.64 For the value IE = 1.7, which was observed (later we will show that this is the case), the isotope effect arisingexperimentally in mitochondria, the relation k/kST/kÀ1 = 1:5:10 in it may be detected even if the limiting step of the total ATPwas shown to reproduce this magnitude of IE. As shown in a synthesis is the release of ATP from the catalytic site. The latter asprevious section, ATP synthesis by the ion-radical mechanism well as the entering of reactants into the catalytic site arestarts with the complex Mg(H2O)6+ (if its partner is Mg2+ controlled by the microsecond molecular dynamics whereas(ADP)3À) or Mg(H2O)4+ (if its partner is protonated Mg2+ ATP synthesis itself and, accompanying it, the isotope effect(ADP)2À). As was shown before, a reaction with participation of are controlled by the nanodynamics. The widely used parametersthe intermediate complex Mg(H2O)6+ is the most probable. For kcat and KM of enzymatic kinetics may be related to the micro-this radical cation the hyperfine coupling constant a(25Mg) was dynamics of the catalytic site. The Michaelis constant KM seemscalculated to be 40 MHz. Taking into account the hyperfine to be almost independent of the isotope substitution; however,coupling constant for 31P in the oxyradical, the total hyperfine the isotope effect resulting from the reaction in the nanosystemcoupling constant in the catalytic site with 25Mg is expected to be may exhibit itself in the total rate constant kcat.∼60 G (∼170 MHz) so that in the catalytic site the approximaterate constants k, kST, and kÀ1 are 3  107, 1.7  108, and 3.4 Â108 sÀ1, respectively. 16. EFFICIENCY OF ATP SYNTHASE AS A MOLECULAR The fastest step in the kinetic scheme is the back electron MACHINEtransfer, i.e., kÀ1 . k, kST. Then from eq 8 follows Under the efficiency we will imply the probability that the ADP molecule, entering into the catalytic site, transforms into 2kST þ k IE ¼ ð9Þ the ATP molecule. In terms of the rate constants, it is the ratio kST þ k k/(k + kST + kÀ1); see section 14. On the basis of the If singletÀtriplet conversion is faster than ATP synthesis (kST k), experimentally measured 25Mg isotope effect on the rate ofthen the limiting value of IE corresponds to doubling of the ATP synthesis by ATP synthase, the rate constants k, kST, andATP yield in sites with 25Mg2+ ions with respect to those with kÀ1 of the reactions in the catalytic site were estimated (see24,26 Mg2+ ions. section 14). The limiting chemical step of the synthesis is shown Now the total rate of ATP synthesis and ATP yield is to consist of an addition of the phosphoryl anion-radical of ADPdetermined as a sum of ATP produced independently by to the PdO bond of phosphate with a rate constant of 3  107enzymatic sites with 25Mg2+ and 24,26Mg2+ ions: sÀ1. From the relationship between the rate constants, the efficiency of ATP synthase with 24Mg2+ ions in which only Y ¼ d½ATPŠ=dt ¼ ½akm þ ð1 À aÞkn Š½AŠ ð10Þ hyperfine coupling with 31P operates is estimated to be ∼8%. M dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  14. 14. Chemical Reviews REVIEWIt means that only one molecule of ATP is formed from the 12ADP molecules sequentially entering into the catalytic site. Forthe synthase with 25Mg2+ ions in the catalytic site, the efficiency isalmost doubled.17. DO ENZYMATIC MOTORS FUNCTION REVERSIBLY? It was shown that macroscopic reversibility in ATP synthesisdoes not mean microscopic reversibility, i.e., the pathway forATP hydrolysis is not that for ATP synthesis.8,9 A completelyindependent proof for this statement is seen through the isotopewindow. The authors36 measured the rate of ATP hydrolysis byMg-CK where magnesium ions were substituted by pure isotopicions 24Mg2+, 25Mg2+, and 26Mg2+. Figure 18 demonstrates thedecay of ATP (curve 1, black points) and accumulation of ADP(curve 2, empty points) as a function of incubation time. Thesum of ATP and ADP remains constant, that is, ATP quantita-tively transforms into ADP. It is remarkable that the ATP decay(and, therefore, ADP accumulation) is identical for all kinasesand does not depend on the isotopes. It is a direct indication thatthe ATP f ADP enzymatic reaction is not spin-selective; no Figure 18. ATP decay (curve 1) and ADP formation (curve 2) induced by Mg-CK (concentration of MgCl2 = 20 mM) as a function ofparamagnetic species are formed in this reaction. In this reaction incubation time. Crosses and dotted line refer to the sum (ATP +there is no magnetic isotope effect, but the classical, mass- ADP); the amounts Y of ATP and ADP are given in mM per g of enzyme.dependent isotope effect is negligibly small, so that it is not Reprinted with permission from ref 36. Copyright 2008 Americandetected in enzymatic activity. It means that the reaction Chemical Society.trajectories for the direct and reverse reactions are not equivalent. nucleophilic channel of ATP synthesis. This is convincingly18. COMPARATIVE ANALYSIS OF THE TWO MECHANISMS proved by direct kinetic studies of ATP synthesis by ATPOF ATP SYNTHESIS synthase.43 Namely, in this transition region a switching over between two mechanisms occurs: the nucleophilic ATP synthesisEnergy is replaced by the ion-radical one. It provides an additional source Nucleophilic addition of a phosphate group to ADP is a rather of ATP, which is much more efficient than the nucleophilic oneenergy-consuming route. Even at long distances, the reaction and is accompanied by the appearance of a magnetic isotopepartners as negatively charged ions repel each other—despite the effect. Pyruvate kinase is an evident illustration of this statementfact that the charged amino acid residues, constituting the cata- (Figure 7): here nucleophilic and ion-radical mechanisms arelytic site, slightly compensate the Coulomb repulsion of the clearly expressed and separated on the magnesium ion concen-reactants. However, at short distances and in the transition state, tration scale. Thus, the discovery of isotope effects simulta-the repulsion of the closed electron shells is rather strong and neously discloses a new, ion-radical mechanism and proveshigh energy is required to overcome the repulsion for the nucle- classical, nucleophilic mechanism.ophilic addition to occur. Indeed, the activation energy of phos- The intracellular concentration of metal ions is rather low,60 sophorylation by protein kinase was calculated to be 15 kcal/mol.39 that the dominating source of ATP in living organisms is thoughtThe energy barrier for the nucleophilic reaction of ATP hydro- to be a nucleophilic reaction. The ion-radical mechanism func-lysis is even higher, ∼39 kcal/mol, and relates to the nucleophilic tions under conditions when the concentration of metal ions isattack of HOÀ on the phosphate group of ATP.65 rather high. However, due to local concentrations induced by Electron transfer as a starting reaction of the ion-radical fluctuations of the distribution or inhomogeneous adsorption ofmechanism is exoergic. There is no calculation of the energy metal ions in cells and mitochondria, the contribution of the ion-profile for this reaction; nevertheless, one may assume that the radical mechanism of ATP synthesis in living organisms is notenergy barrier for electron transfer is rather low like in typical negligible; it was tested by in vivo experiments (Figure 12).exoergic electron transfer reactions. This statement is in accor- It demonstrates that both mechanisms, nucleophilic and ion-dance with experimental results: at concentrations of Mg2+, Ca2+, radical, coexist and function independently, producing comparableZn2+ ions where the ion-radical mechanism works, the yield of amounts of ATP even at low concentrations of the metal ions.ATP is strongly increased and considerably exceeds the con-tribution coming from nucleophilic mechanism.Ion Concentration 19. IS ATP SYNTHESIS A NUCLEOPHILIC OR AN ION- The efficiency of nucleophilic mechanism depends on the RADICAL PROCESS?concentration of ions:43 even a single ion in the catalytic site The generally accepted nucleophilic mechanism of ATPseems to be enough to do that. However, for the efficient synthesis had no alternatives until magnetic isotope effects offunctioning two magnesium ions are required: one is bound Mg, Ca, and Zn were observed in enzymatic ATP synthesiswith ADP, and the other is considered as “free”, not bound with driven by ATP synthase and kinases. To explain these effectsphosphate groups but solvated by water molecules and protein as well as magnetic field effects, an ion-radical mechanismresidues.8 The invasion of the additional ions in the catalytic site was formulated in which the starting reaction was suggestedresults in complexes with the substrate; the latter suppresses the to be an electron transfer from the Mg(ADP) complex to N dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000
  15. 15. Chemical Reviews REVIEWMg(H2O)n2+ ion. It generates an ion-radical pair as an electron for medicinal purposes to prevent and/or compensate some cell/and nuclear spin-selective nanoreactor. By analysis of magnetic tissue ATP losses caused be variable pathogenic factors and,parameters (g-factors and hyperfine coupling constants a(31P) therefore, to either treat or prevent a number of essentialand a(25Mg)) of the ion-radicals, it was shown that a terminal, cardiology-related syndromes.negatively charged oxygen atom of ADP in the complex Mg-(ADP) donates an electron to the Mg(H2O)n2+ ion. The same ASSOCIATED CONTENTwas proved to be valid for the calcium- and zinc-directed ATP S b Supporting Informationsynthesis. This mechanism is unprecedented but firmly established; it is Tables and figures of the optimized structures, energies, andthe only one that explains the whole body of experimental magnetic parameters. This material is available free of charge viaresults. the Internet at http://pubs.acs.org. (1) In the ion-radical pair, four reaction channels appear: the singlet one and three triplet ones. Their relative contribu- AUTHOR INFORMATION tion to the ATP yield is controlled by the rate of Corresponding Author singletÀtriplet spin conversion induced by hyperfine *Phone: +7-495-9397128. Fax: +7-495-9382484. E-mail: abuchach@ electronÀnuclear interaction. In pairs with nonmagnetic, chph.ras.ru. spinless nuclei, spin conversion is slow and the contribu- tion of triplet channels into the ATP production is not too significant. However, the magnetic nuclei 25Mg, 43Ca, and BIOGRAPHIES 67 Zn strongly stimulate spin conversion, resulting in an increase of ATP yield by 2À4 times. (2) The magnetic field effects on ATP synthesis are another predictable and experimentally detected irrefutable argu- ment in favor of the ion-radical mechanism. (3) Theoretical calculations and thermodynamic data on the energy of electron transfer reactions between M(H2O)n2+ ions and M(ADP) complexes reliably demonstrate that they are energy allowed only for the partly dehydrated ions M(H2O)n2+. As shown in section 12, there is no need to destroy the first coordination sphere and release the sixth and fifth tightly bound water molecules. This important effect is in perfect accordance with the well- known observation that ATP synthesis does not occur in water (it is energy-forbidden by ∼4 eV) from the same reactants and under the same conditions. ATP Anatoly L. Buchachenko graduated from Nizhny Novgorod synthesis takes place only in specially designed molec- University in 1958. Currently he is professor of N. N. Semenov ular motors, enzymes that are known to squeeze water Institute of Chemical Physics and M. V. Lomonosov Moscow out of the catalytic sites when protein domains are University. His interests are in spin chemistry, ESR and NMR approaching to unite phosphorylating substrate and spectroscopy, molecular ferromagnets, and isotope effects in ADP. The removal of water from the outer coordination chemical and biochemical reactions. sphere of M(H2O)n2+ ion increases its positive charge and electron affinity, thereby activating the electron transfer reaction. (4) In terms of the ion-radical mechanism, it is explicable why the efficiency of Ca2+ and Zn2+ ions in ATP production is almost the same as that of Mg2+ ions. Moreover, recent calculations of the electron transfer reactions convincibly predict that other ions, such as Ba2+, Cd2+, and Sn2+, should be also efficient catalysts of ATP synthesis and show a magnetic isotope effect59 similar to that of Mg2+, Ca2+, and Zn2+ ions. The novel ion-radical mechanism of ATP synthesis is uni-versal, but it is not the only one that is observed through theisotope window. The other outstanding result is the foundationof the classical, generally accepted, but so far hypotheticalnucleophilic mechanism. A general property of ATP synthesis Dmitry A. Kuznetsov is a leading research fellow in theis that, at low concentration of magnesium, calcium, and zinc N. N. Semenov Institute of Chemical Physics and professor ofions, there is no isotope effect in all enzymes. It unambiguously N. I. Pirogov Russian National Research Medical University.indicates that ATP synthesis does not involve paramagnetic His scientific interests are enzyme biochemistry, medicinalintermediates and occurs as a nucleophilic reaction. The great nanotechnologies, and magnetic isotope effects in enzymaticadvantage of the ion-radical mechanism is that it might be used catalysis. O dx.doi.org/10.1021/cr200142a |Chem. Rev. XXXX, XXX, 000–000