Enzymatic and electron paramagnetic resonance studies of anabolic pyruvate synthesis by pyruvate: ferredoxin oxidoreductase from Hydrogenobacter thermophilus Takeshi Ikeda 1, *, Masahiro Yamamoto 1, , Hiroyuki Arai 1 , Daijiro Ohmori 2 , Masaharu Ishii 1 and Yasuo Igarashi 1 1 Department of Biotechnology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Japan 2 Department of Chemistry, School of Medicine, Juntendo University, Chiba, Japan Introduction Pyruvate: ferredoxin oxidoreductase (POR; pyruvate synthase, EC 1.2.7.1) catalyzes the thiamine pyrophos- phate (TPP)-dependent oxidative decarboxylation of pyruvate to form acetyl-CoA and CO 2 . POR contains one or multiple iron-sulfur clusters in addition to TPP [1]; the two electrons that arise during oxidation of pyruvate at the TPP site are sequentially transferred via the iron-sulfur cluster(s) to external electron accep- tors. The physiological electron acceptor is a small iron-sulfur protein ferredoxin or FMN-containing Keywords Hydrogenobacter thermophilus; iron-sulfur cluster; pyruvate: ferredoxin oxidoreductase; reductive tricarboxylic acid cycle; thiamine pyrophosphate Correspondence M. Ishii, Department of Biotechnology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan Fax: +81 3 5841 5272 Tel: +81 3 5841 5143 E-mail: amishii@mail.ecc.u-tokyo.ac.jp Present address *Research Institute for Nanodevice and Bio Systems, Hiroshima University, Japan Institute of Biogeoscience, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Kanagawa, Japan (Received 11 September 2009, revised 17 November 2009, accepted 19 November 2009) doi:10.1111/j.1742-4658.2009.07506.x Pyruvate: ferredoxin oxidoreductase (POR; EC 1.2.7.1) catalyzes the thia- mine pyrophosphate-dependent oxidative decarboxylation of pyruvate to form acetyl-CoA and CO 2 . The thermophilic, obligate chemolithoauto- trophic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus TK-6, assimilates CO 2 via the reductive tricarboxylic acid cycle. In this cycle, POR acts as pyruvate synthase catalyzing the reverse reaction (i.e. reduc- tive carboxylation of acetyl-CoA) to form pyruvate. The pyruvate synthesis reaction catalyzed by POR is an energetically unfavorable reaction and requires a strong reductant. Moreover, the reducing equivalents must be supplied via its physiological electron mediator, a small iron-sulfur protein ferredoxin. Therefore, the reaction is difficult to demonstrate in vitro and the reaction mechanism has been poorly understood. In the present study, we coupled the decarboxylation of 2-oxoglutarate catalyzed by 2-oxogluta- rate: ferredoxin oxidoreductase (EC 1.2.7.3), which generates sufficiently low-potential electrons to reduce ferredoxin, to drive the energy-demanding pyruvate synthesis by POR. We demonstrate that H. thermophilus POR catalyzes pyruvate synthesis from acetyl-CoA and CO 2 , confirming the operation of the reductive tricarboxylic acid cycle in this bacterium. We also measured the electron paramagnetic resonance spectra of the POR intermediates in both the forward and reverse reactions, and demonstrate the intermediacy of a 2-(1-hydroxyethyl)- or 2-(1-hydroxyethylidene)-thia- mine pyrophosphate radical in both reactions. The reaction mechanism of the reductive carboxylation of acetyl-CoA is also discussed. Abbreviations DTNB, 5,5¢-dithiobis-(2-nitrobenzoic acid); EPR, electron paramagnetic resonance; HE-TPP, 2-(1-hydroxyethyl)- or 2-(1-hydroxyethylidene)- thiamine pyrophosphate; LDH, lactate dehydrogenase; OGOR, 2-oxoglutarate: ferredoxin oxidoreductase; OR, 2-oxoacid oxidoreductase; POR, pyruvate: ferredoxin oxidoreductase; TCA, tricarboxylic acid; TPP, thiamine pyrophosphate. FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 501 flavodoxin. By contrast to pyruvate dehydrogenase multienzyme complex, which irreversibly catalyzes the same reaction utilizing NAD + as an electron acceptor in mitochondria and respiratory bacteria, POR can also catalyze the reverse reaction (i.e. the reductive car- boxylation of acetyl-CoA) provided that a sufficiently low-potential electron donor is available. The reverse reaction (pyruvate synthesis) is a central step for some autotrophic bacteria because it serves to assimilate CO 2 into cell carbon [1]. Hydrogenobacter thermophilus TK-6 is a faculta- tive aerobic, thermophilic, obligate chemolithoauto- trophic hydrogen-oxidizing bacterium [2]. The optimum growth temperature range for H. thermophilus TK-6 is 70–75 °C. Phylogenetic analyses of 16S ribosomal RNA sequences have shown that the genus Hydroge- nobacter is a member of the deepest branching order in the domain Bacteria [3]. H. thermophilus assimilates CO 2 via the reductive tricarboxylic acid (TCA) cycle [4], which is one of the microbial CO 2 fixation path- ways [5]. The reductive TCA cycle is a reversal of the oxidative TCA cycle, and is an endergonic anabolic pathway that requires reducing equivalents to complete the cycle [6,7]. POR is one of the key enzymes of the reductive TCA cycle, and catalyzes the anabolic reduc- tive carboxylation of acetyl-CoA. Two [4Fe-4S] ferre- doxins, Fd1 and Fd2, from this bacterium are considered to serve as low-potential electron donors for this key reaction [8]. POR is distributed among archaea, bacteria and anaerobic protozoa, and is a member of the 2-oxoacid oxidoreductase (OR) family, which catalyzes the oxida- tive decarboxylation of 2-oxoacids to their acyl- or aryl-CoA derivatives [9]. OR enzymes can be homodi- meric [10,11], heterodimeric [12,13] or heterotetrameric [14], depending on the organism. These three types of OR are phylogenetically related and the heterotetra- meric enzyme has been proposed to be the common ancestor that underwent gene rearrangement and fusion to generate homo- and heterodimeric ORs [9,13,15]. We recently found novel heteropentameric ORs [POR and 2-oxoglutarate: ferredoxin oxidoreduc- tase (OGOR; 2-oxoglutarate synthase, EC 1.2.7.3)] in H. thermophilus and its close relatives [16,17]; in these organisms, the heteropentameric POR and OGOR function as the key components of the reductive TCA cycle and catalyze the anabolic reductive carboxylation of acetyl-CoA and succinyl-CoA, respectively [17,18]. Four of the five subunits correspond to those of the heterotetrameric ORs, suggesting that the heteropenta- meric ORs might have evolved from an ancestral het- erotetrameric enzyme by the acquisition of a unique fifth polypeptide of unknown function. Sequence align- ments suggest that H. thermophilus POR contains one TPP and three [4Fe-4S] 2+ ⁄ 1+ clusters per catalytic unit [17]; the enzyme contains all of the motifs required for cofactor binding that were identified in the crystal structure of the homodimeric Desulfovibrio africanus POR [19]. In most cases, the enzyme activity of POR is assayed by monitoring the reduction of an artificial electron carrier, methyl viologen, during the oxidative decar- boxylation of pyruvate. However, when the anabolic role of the novel heteropentameric POR in H. thermo- philus is considered, the reverse reaction [i.e. the reduc- tive carboxylation of acetyl-CoA (pyruvate synthesis)] needs to be assayed. However, this reverse reaction has proven more difficult to study because a strong reduc- tant is required to drive the reaction. Hence, the cata- lytic mechanism of the reductive carboxylation catalyzed by POR is poorly understood, whereas that of the oxidative decarboxylation has been intensively investigated [20,21]. In the present study, we developed an assay system to demonstrate the reductive carboxyl- ation catalyzed by H. thermophilus POR with reduced ferredoxin as an electron donor. Specifically, we uti- lized another OR-family enzyme, OGOR from H. ther- mophilus, to supply reduced ferredoxin for anabolic pyruvate synthesis mediated by POR. OGOR also cat- alyzes the oxidative decarboxylation of 2-oxoglutarate using (oxidized) ferredoxin as an electron acceptor. By coupling the OGOR decarboxylation, we demonstrate that H. thermophilus POR catalyzes the anabolic reductive carboxylation of acetyl-CoA to form pyru- vate. We also investigated the inter- and intramolecu- lar electron transfer during the reductive carboxylation by electron paramagnetic resonance (EPR) spectros- copy to clarify its catalytic mechanism. Results In vitro assay for pyruvate synthesis by POR: the coupled assay with OGOR and lactate dehydrogenase (LDH; EC 1.1.1.27) Because the synthesis of pyruvate from acetyl-CoA and CO 2 is an energetically unfavorable reaction with a reduction potential of )540 mV [22], this reaction requires a strong reductant. Pyruvate dehydrogenase multienzyme complex cannot catalyze the react- ion because the requisite electron donor, NADH (E 0 ¢ = )320 mV), is much too weak an electron source to drive the transformation. For the enzyme assay, the reducing power must be supplied in vitro by the physiological electron donor for POR, ferredoxin. A possible strategy is to couple the POR reaction to a Pyruvate synthesis by pyruvate oxidoreductase T. Ikeda et al. 502 FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS ferredoxin-reducing enzyme. For example, POR from Moorella thermoacetica biosynthesizes pyruvate using ferredoxin reduced by CO dehydrogenase [23]; Chloro- bium tepidum POR has been shown to catalyze pyru- vate synthesis mediated by ferredoxin reduced by the light-driven reactions of spinach chloroplasts or Chloro- bium reaction centers [24,25]. However, H. thermophilus does not possess these ferredoxin-reducing systems. In this bacterium, reducing equivalents are derived from hydrogen oxidization catalyzed by multiple hydrogenases [26], although the physiological electron transfer pathway(s) from hydrogen to ferredoxin has not yet been clarified. Instead, we utilized OGOR from this bacterium to reduce the low-potential ferredoxin (Fig. 1). The oxidative decarboxylation of 2-oxogluta- rate catalyzed by OGOR generates low-potential electrons that reduce ferredoxins as follows: 2-oxoglut- arate + CoA fi succinyl-CoA + CO 2 +H + +2e ) , E 0 ¢ = )520 mV [27]. The reactions catalyzed by POR and OGOR were further coupled to the LDH reaction to detect pyruvate formation spectrophotometrically (Fig. 1). Pyruvate generated by coupling the POR and OGOR reactions was reductively converted to lactate with NADH as an electron donor. Thus, the rate of pyru- vate formation was monitored as the decrease in A 340 as a result of NADH oxidation. The thermostable LDH from a thermophilic bacterium Thermus caldophilus [28] was used for the assay because the reaction was performed at 70 °C, which is the optimum temperature for H. thermo- philus. Table 1 shows the overall reaction of this coupled assay. Using the coupled assay with OGOR and LDH, H. thermophilus POR was found to catalyze the reduc- tive carboxylation of acetyl-CoA. Indeed, acetyl- CoA-dependent NADH oxidation was observed with either reduced Fd1 or Fd2 as an electron donor (Fig. 2). It was confirmed that pyruvate synthesis by POR was rate-limiting in this coupled system. A slight decrease in A 340 in the absence of acetyl-CoA was a result of the spontaneous thermal degradation of NADH [29]. The reductive carboxylation depended on the presence of POR, OGOR, LDH, ferredoxin, 2-oxo- glutarate and acetyl-CoA (data not shown), indicating that the coupled assay shown in Fig. 1 proceeded as expected. However, this reaction did not depend on the presence of NaHCO 3 (CO 2 ) and CoA (Fig. 2B). Because CO 2 was produced from 2-oxoglutarate by OGOR in this coupled assay, the addition of NaHCO 3 was not necessarily required for the total reaction (Fig. 1, dashed arrow). By contrast to CO 2 , CoA was an essential substrate to initiate this coupled reaction; nevertheless, the reaction actually proceeded without the addition of CoA with a higher reaction rate than Fig. 1. Schematic representation of the coupled enzyme assay. The reductive carboxylation catalyzed by POR was coupled with the OGOR and LDH reactions. Fd ox , oxidized ferredoxin; Fd red , reduced ferredoxin. Table 1. Enzymatic reactions. Enzyme Reaction catalyzed by the enzyme a POR (reductive carboxylation) Acetyl-CoA + CO 2 +2· Fd red fi pyruvate + CoA + 2 · Fd ox OGOR (oxidative decarboxylation) 2-Oxoglutarate + CoA + 2 · Fd ox fi succinyl-CoA + CO 2 +2· Fd red LDH Pyruvate + NADH fi lactate + NAD + Total Acetyl-CoA + 2-oxoglutarate + NADH fi lactate + succinyl-CoA + NAD + a Protons are omitted from the reactions for simplicity. Fd ox , oxidized ferredoxin; Fd red , reduced ferredoxin. A B Fig. 2. Reductive carboxylation catalyzed by POR in the coupled enzyme assay. The assay mixture contained 1 m M acetyl-CoA, 10 m M NaHCO 3 ,10mM 2-oxoglutarate, 0.5 mM CoA, 0.2 mM NADH, 1 mM fructose 1,6-bisphosphate, 10 mM MgCl 2 ,1mM di- thiothreitol, 0.5 m M TPP, 0.03 U of OGOR, 0.2 U of LDH and 10 l M Fd1 (A) or Fd2 (B) in 100 mM Hepes buffer (pH 8.0). Open circles, complete reaction; filled circles, acetyl-CoA omitted; open squares, NaHCO 3 omitted; filled squares, CoA omitted. T. Ikeda et al. Pyruvate synthesis by pyruvate oxidoreductase FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 503 that of the complete reaction (Fig. 2B, filled square). This was the result of a trace amount of CoA in the assay mixture. CoA quantification by the 5,5¢- dithiobis-(2-nitrobenzoic acid) (DTNB) assay showed that 100 mm acetyl-CoA stock solution contained 2.4 mm CoA, corresponding to 24 lm CoA in the stan- dard assay mixture (i.e. without the addition of CoA solution). The K m value for CoA of the OGOR enzyme is reported to be 80 lm [30]. Thus, decarboxyl- ation of 2-oxoglutarate can proceed as a result of CoA contamination of the acetyl-CoA solution. When the reaction commences, CoA is regenerated by the reduc- tive carboxylation of acetyl-CoA (Fig. 1, dashed arrow). The reaction rate of this coupled assay was significantly affected by the concentration of CoA (Fig. 2B). Although CoA was an essential substrate of this assay, pyruvate synthesis by POR was inhibited in the presence of excess CoA, which caused the reverse reaction (oxidative decarboxylation of pyruvate). This impasse prevented any further kinetic analysis of the reaction. In this assay, the reductive carboxylation activity of POR was determined to be 0.23 UÆmg )1 with 10 lm Fd1, or 0.19 UÆmg )1 with 10 lm Fd2. These values were comparable to those of the oxidative decarboxyl- ation of pyruvate with Fd1 or Fd2 as an electron acceptor (0.55 UÆmg )1 or 0.43 UÆmg )1 , respectively; data not shown), suggesting that H. thermophilus POR functions as an active pyruvate synthase. EPR measurements of POR during the oxidative decarboxylation The purified H. thermophilus POR showed an EPR sig- nal (g 1,2,3 = 1.973, 2.012 and 2.024) attributed to the oxidized S =1⁄ 2 [3Fe-4S] 1+ cluster [31] (Fig. 3A). In the dithionite-reduced state, the [3Fe-4S] signal disap- peared and a new signal attributed to the reduced S =1⁄ 2 [4Fe-4S] 1+ clusters was observed (Fig. 3B). This new signal is an overlap of a major signal with g-values of 1.910, 1.922 and 2.040 and a minor signal (approximately 4% of the major signal; determined by spectral simulation) with g-values of 1.880, 2.003 and 2.020 (note that the minor feature around g = 2.00 in Fig. 3B was a result of this minor signal and not a TPP radical intermediate described later). The relative amount of the [3Fe-4S] 1+ cluster in Fig. 3A was < 10% that of the [4Fe-4S] 1+ clusters in Fig. 3B (determined by comparing the double integrals of the EPR spectra), indicating that a substoichiometric amount of the [3Fe-4S] cluster was derived from the oxygen-sensitive [4Fe-4S] by partial oxidative damage, as reported for other ORs [14,32–34]. In the presence of 20 mm pyruvate, POR showed a strong sharp signal centered at g = 2.0040 (Fig. 3C), whereas 0.5 mm CoA did not affect the signal of POR (data not shown), indicating that the oxidative decarboxylation reaction begins with the binding of pyruvate, but not A B C D E F Fig. 3. EPR spectra of H. thermophilus POR. The purified POR was incubated with the components: (A) no substrate (as purified); (B) dithionite, (C) pyruvate; (D) pyruvate and CoA; (E) Fd1, OGOR, 2-oxoglutarate and CoA; (F) acetyl-CoA, Fd1, OGOR, 2-oxoglutarate and CoA. Instrument settings were: temperature, 10 °K; microwave power, 100 lW for (A), 250 lW for (B, D–F) or 1 lW for (C); micro- wave frequency, 9.024 GHz; modulation frequency, 100 kHz; mod- ulation amplitude, 0.2 mT. The arrow indicates the signal of the TPP radical intermediate generated during the reductive carboxyla- tion of acetyl-CoA. Pyruvate synthesis by pyruvate oxidoreductase T. Ikeda et al. 504 FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS CoA. These results are consistent with a ping-pong cat- alytic mechanism with pyruvate as the primary sub- strate [35]. (Note that the microwave power for Fig. 3C was 1 lW, 250-fold lower than that for the others; the intensity of EPR signals is proportional to the square root of the microwave power under nonsat- urating conditions.) Although the [3Fe-4S] 1+ signal in Fig. 3A disappeared at temperatures exceeding 30 °K, the g = 2.0040 signal remained even at 70 °K (data not shown), indicating that this signal was the result of a TPP-radical intermediate. Indeed, this radical is pro- posed to be the common intermediate in pyruvate decarboxylation catalyzed by all PORs [36], which is generated by the binding of pyruvate to TPP and the resultant decarboxylation. Although the chemical structure of this intermediate is still controversial [37– 39], it is often referred to as a 2-(1-hydroxyethyl)- or 2-(1-hydroxyethylidene)-TPP (HE-TPP) radical. The hyperfine structure of the radical was determined in detail at 70 °K, at which the EPR signals of iron-sulfur clusters are not detectable (Fig. 4A). The hyperfine splitting pattern is essentially the same as reported in other studies [11,36–38,40,41]. The HE-TPP radical is generated by one-electron transfer; one of the two elec- trons that are generated during decarboxylation of pyruvate remains on the TPP intermediate, and the other electron moves to the intramolecular iron-sulfur cluster [42]. Consistent with this mechanism, the oxi- dized [3Fe-4S] 1+ signal in Fig. 3A disappeared in Fig. 3C, indicating that the cluster was reduced by this electron. However, no reduced [4Fe-4S] signal was detectable in Fig. 3C. This is probably because a reduced [4Fe-4S] cluster(s) can be reoxidized by a trace amount of oxygen [42]. Because the redox potential of [3Fe-4S] clusters is generally much higher than that of [4Fe-4S] clusters ()150  )100 mV versus )650  )250 mV) [43], the [3Fe-4S] cluster in the enzyme was not reoxidized. Upon further addition of CoA, the signal of the HE-TPP radical markedly decreased, accompanied by concomitant formation of a reduced [4Fe-4S] signal, which was similar to that of the dithionite-reduced POR (Fig. 3D), indicating the second electron transfer from the radical to the [4Fe- 4S] cluster(s). The presence of both pyruvate and CoA allows catalysis to proceed until all the oxygen is con- sumed [40], preventing reoxidation of the reduced [4Fe-4S] cluster(s). Because iron-sulfur clusters can receive only one electron at a time, multiple clusters should be reduced in this state. These electrons are then readily released to external electron mediators. Indeed, in the presence of Fd1, the rhombic S =1⁄ 2 [4Fe-4S] 1+ signal of the reduced Fd1 [8] was clearly observed (data not shown). EPR measurements of POR during the reductive carboxylation Addition of 1 mm acetyl-CoA did not affect the EPR signal of POR (data not shown), indicating that elec- tron transfer from external electron donors is a key step to initiate the carboxylation reaction. To supply reducing equivalents to POR via ferredoxin, we utilized OGOR as described above. In the presence of 2-oxoglutarate, CoA, OGOR and Fd1, the reduced [4Fe-4S] signal of POR was observed, overlapping with the rhombic [4Fe-4S] 1+ signal of Fd1 (g z,y,x = 2.08, 1.94 and 1.92) [8] (Fig. 3E). Upon further addition of acetyl-CoA, a signal with g = 2.0040 was observed (Fig. 3F, arrow). (It was confirmed that this signal was not a result of the OGOR intermediate.) The hyperfine structure of this signal shows essentially the same hyperfine splitting pattern as that of the HE-TPP radi- cal intermediate observed during the decarboxylation of pyruvate (Fig. 4), indicating that the HE-TPP radi- cal was a common intermediate in both the oxidative and reductive reactions. The HE-TPP radical was formed only after the reduction of the iron-sulfur clus- ters of the enzyme, suggesting that the electrons sup- plied via external ferredoxin molecules played an important role in forming the radical intermediate from TPP with bound acetyl-CoA. A B Fig. 4. Hyperfine structures of the EPR signal of the TPP radical intermediates. (A) incubated with pyruvate and CoA (three scans); (B) incubated with acetyl-CoA, Fd1, OGOR, 2-oxoglutarate and CoA (five scans). Instrument settings were: temperature, 70 °K; micro- wave power, 1 lW for (A) or 100 lW for (B); modulation amplitude, 0.02 mT for (A) or 0.2 mT for (B); other settings were as described in Fig. 3. The higher power and wider modulation were used for (B) to increase sensitivity because the amount of the radical intermedi- ate was much less than for (A). T. Ikeda et al. Pyruvate synthesis by pyruvate oxidoreductase FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 505 Discussion In the present study, we demonstrate that H. thermo- philus POR catalyzes pyruvate synthesis from acetyl- CoA and CO 2 , by the coupled assay with OGOR and LDH (Fig. 1). Although carboxylation activity is gen- erally determined by monitoring the incorporation of 14 CO 2 to form [ 14 C] pyruvate, OGOR catalyzes the exchange reaction between CO 2 and the carboxyl group of 2-oxoglutarate [44], and therefore interferes with the detection of [ 14 C] pyruvate. Instead, the rate of pyruvate formation was determined by monitoring the LDH-coupled oxidation of NADH to NAD + . The coupled assay also demonstrated that Fd1 and Fd2 function as electron mediators for POR (and also for OGOR) [45] in both the oxidative and reductive reac- tions. These results corroborate the operation of the reductive TCA cycle in H. thermophilus. Specifically, two irreversible reactions in the oxidative TCA cycle, oxidative decarboxylation of pyruvate and 2-oxogluta- rate, are anabolically reversed by POR and OGOR, respectively, as suggested by our early work [4], with Fd1 and Fd2 acting as physiological electron donors. However, because this assay is a complex system involving four proteins, kinetic analysis was not possi- ble. The substrates, CoA and CO 2 , were involved in the two reactions catalyzed by POR and OGOR (Fig. 1, dashed arrows). In particular, CoA was a sub- strate of OGOR as well as a product of POR, and sig- nificantly affected the reaction rate of pyruvate synthesis. These problems were derived from the fact that POR and OGOR are similar OR-family enzymes, both reversibly catalyzing the CoA-dependent oxidative decarboxylation of 2-oxoacids. For further analysis, an alternative enzyme that can generate low- potential electrons to reduce ferredoxin is required. Thus far, two other enzymes that utilize ferredoxin as an electron mediator have been purified from H. ther- mophilus: ferredoxin-NADP + reductase (EC 1.18.1.2) [46] and ferredoxin-dependent glutamate synthase (EC 1.4.7.1) [47]. However, the midpoint potentials of the half reactions catalyzed by these enzymes are higher than that mediated by OGOR, and are there- fore unsuitable for the reduction of ferredoxin. Thus, the identification and characterization of enzymes that transfer electrons to ferredoxins in vivo is of particular importance for the improvement of this coupled assay and also with respect to obtaining a deeper under- standing of the metabolism of H. thermophilus. Indeed, this would enable the kinetic analysis of the POR reac- tions in both directions. In particular, the reaction rate under physiological intracellular concentrations of sub- strates needs to be determined to demonstrate that H. thermophilus POR functions toward pyruvate syn- thesis in vivo. To investigate the reaction mechanism of H. thermo- philus POR, we measured the EPR spectra of the enzyme in the presence of various combinations of sub- strates. Intra- and intermolecular electron transfer dur- ing the oxidative decarboxylation was essentially consistent with the catalytic cycle proposed by Menon and Ragsdale [36]. We further measured the EPR spec- tra during the reductive carboxylation of acetyl-CoA, using OGOR to reduce ferredoxin as in the coupled assay. In the presence of the reduced Fd1 and acetyl- CoA, the HE-TPP radical intermediate was formed (Figs 3F and 4B), indicating the intermediacy of the HE-TPP radical in both the oxidative and reductive reactions. The results obtained also indicate that elec- tron transfer from external ferredoxin to the enzyme is an indispensable step to form the radical in the reductive reaction. From the data obtained in the present study, along with evidence available from the literature, we are able to propose the catalytic mechanism of the reductive carboxylation of acetyl-CoA (Fig. 5). (1) The TPP carb- anion is generated by proton extraction from C2 carbon atom of the thiazolium ring by the tautomeric 4¢ imino group of the 4¢-aminopyrimidine ring [48], as is the case for all TPP-dependent enzymes; this process is also com- mon to the oxidative decarboxylation catalyzed by this enzyme. (2) The nucleophilic TPP C2-carbanion attacks the carbonyl carbon of acetyl-CoA (as it attacks the car- bonyl carbon of pyruvate in the oxidative decarboxyl- ation) to form a transient tetrahedral intermediate. (3) The tetrahedral intermediate undergoes CoA release and one-electron transfer to the adduct of TPP to form the HE-TPP radical intermediate. It is not known whether CoA release and one-electron reduction occur in a stepwise manner [possibly forming the acetyl-TPP as an intermediate (3¢-a)] or simultaneously. However, we believe the latter process is more likely because dur- ing the oxidative decarboxylation the binding of CoA appears to be tightly coupled to electron transfer from the HE-TPP radical [49]. (4) The generated HE-TPP radical is reduced to the HE-TPP C2a carbanion by a second electron transfer and then (5 and 6) the resultant carbanion attacks CO 2 , which might be tightly bound to the active site of the enzyme [37], to form pyruvate. These latter steps (4, 5 and 6) correspond to the exchange reaction between CO 2 and the carboxyl group of pyruvate catalyzed by this enzyme [50]. Further studies are being planed to confirm the above catalytic mechanism. Moreover, the investigation of the reductive reaction using the coupled system developed in this study is not only highly important itself, but also would provide further insights into the Pyruvate synthesis by pyruvate oxidoreductase T. Ikeda et al. 506 FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS reverse, oxidative reaction and vice versa. Thus, further studies on the POR reactions in both directions would lead to a deeper understanding of the overall reaction mechanism of this enzyme. Materials and methods Bacterial strains and growth conditions Escherichia coli JM109 and BL21(DE3) were used as hosts for derivatives of pUC19 and pET21c, respectively. E. coli MV1184 was used as a host for the expression of T. caldo- philus LDH. E. coli strains were grown in tryptic soy broth or LB medium at 37 °C. When necessary, ampicillin (100 lgÆmL )1 ) was added to the medium for plasmid selection. Heterologous expression and purification of POR, OGOR and ferredoxins Because H. thermophilus POR (UniProt accession numbers Q9LBF7–Q9LBG1) is oxygen-sensitive, as is the case for other ORs [51], the recombinant POR was expressed under microaerobic conditions and purified under anaerobic con- ditions as described previously [17]. In preparation for EPR spectroscopy, dithionite was removed from the purification buffers. H. thermophilus has two isozymes of OGOR, het- erodimeric Kor (UniProt accession numbers Q9AJL9 and Q9AJM0) and heteropentameric For (UniProt accession numbers Q93RA0–Q93RA4) [16,30]. Because the former is much more active than the latter, we utilized the recombi- nant Kor in the present study. Kor, Fd1 (UniProt accession number Q75VV9) and Fd2 (UniProt accession number Q4R2T6) were heterologously expressed and purified as described previously [8,52]. Heterologous expression and purification of LDH The plasmid, p8T4, carrying the gene encoding T. caldophi- lus LDH (UniProt accession number P06150) [53] was a kind gift from Professor Hayao Taguchi (Tokyo University of Science). LDH was heterologously expressed and purified to apparent homogeneity (M. Aoshima, A. Nishiyama and Y. Igarashi, unpublished results). The enzyme activity of the recombinant LDH was assayed at 70 ° C by monitoring the lactate-dependent NADH oxidation as the decrease in A 340 . The standard assay mixture contained 1 mm lactate, 0.2 mm NADH and 1 mm fructose 1,6-bisphosphate (an allosteric effector of T. caldophilus LDH) [28] in 100 mm Hepes buffer (pH 8.0 at 20 °C). The oxidation of NADH NS H H N H 3 C N N NS H 3 C O SCoA TPP C2-carbanion NS OHH 3 CH 3 C NS HO SCoA HE-TPP radical e – CoASH OHH 3 C NS CO 2 H 3 C NS HO COO – H 3 C O COO – e – HE-TPP C2 α -carbanion 2 3 4 56 1 NS OH 3 C CoASH 3 ′ -a 3 ′ -b e – Fig. 5. Proposed catalytic mechanism for the reductive carboxylation of acetyl-CoA catalyzed by POR. The HE-TPP radical is illustrated on the basis of the model proposed by Barletta et al. [57] with the unpaired electron on the C2a carbon, although its chemical structure is still controversial [37–39]. T. Ikeda et al. Pyruvate synthesis by pyruvate oxidoreductase FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 507 was calculated using an extinction coefficient of 6200 m )1 Æcm )1 . One unit of enzyme activity was defined as the oxidation of 1 lmolÆmin )1 of NADH. POR enzyme assays The oxidative decarboxylation activity of POR was assayed at 70 °C by monitoring the ferredoxin-mediated reduction of metronidazole [54]. The standard assay mixture con- tained 20 mm pyruvate, 0.5 mm CoA, 10 lm ferredoxin, 0.1 mm metronidazole, 10 mm MgCl 2 ,1mm dithiothreitol and 0.5 mm TPP in 100 mm Hepes buffer (pH 8.0 at 20 °C). The decrease in A 320 was measured under an argon atmosphere. The reduction of metronidazole was calculated using an extinction coefficient of 9300 m )1 Æcm )1 . One unit of enzyme activity was defined as the reduction of 2 lmolÆ- min )1 of metronidazole (corresponding to the decarboxyl- ation of 1 lmolÆmin )1 of pyruvate on the assumption that the bleaching of the chromophore is a one-electron process) [55]. The reductive carboxylation activity of POR was determined at 70 ° C by the coupled assay with OGOR and LDH (see Results). The standard assay mixture contained 1mm acetyl-CoA, 10 mm NaHCO 3 ,10mm 2-oxoglutarate, 0.5 mm CoA, 0.2 mm NADH, 1 mm fructose 1,6-bisphos- phate, 10 mm MgCl 2 ,1mm dithiothreitol, 0.5 mm TPP, 0.03 U of OGOR, 0.2 U of LDH and 10 lm ferredoxin (Fd1 or Fd2) in 100 mm Hepes buffer (pH 8.0). The assay mixture without NADH and acetyl-CoA was incubated at 70 °C under an argon atmosphere. The reaction was started by adding the NADH, acetyl-CoA and enzyme solutions to the mixture, and the decrease in A 340 as a result of NADH oxidation was measured. One unit of enzyme activity was defined as the reduction of 1 lmolÆmin )1 of NADH (corresponding to the carboxylation of 1 lmolÆmin )1 of acetyl-CoA). Quantification of CoA The concentration of CoA was quantified using DTNB, which reacts with free thiol groups (e.g. CoA-SH) to pro- duce 2-nitro-5-thiobenzoate with an extinction coefficient of 13 600 m )1 Æcm )1 at 412 nm [56]. The assay mixture contained 0.1 mm DTNB in 100 mm Tris–HCl buffer (pH 8.0). Measurement of A 412 was performed after the addition of the sample solution. EPR measurements The enzyme solution was incubated with a substrate(s) in an EPR sample tube at 70 °C for 5–10 min under a gentle argon flow that had passed through a deoxidizing column (Gasclean GC-RP; Nikka Seiko, Tokyo, Japan). The reac- tion was stopped by immersion of the tube in liquid nitro- gen. EPR spectra were measured on a JES-FA300 spectrometer (JEOL, Tokyo, Japan) using a cylindrical cavity (TE 101 mode). The measurement temperature was controlled with a JEOL ES-CT470 cryostat system and a digital temperature indicator ⁄ controller model 9650 (Scientific Instruments, West Palm Beach, FL, USA). The magnetic field was calibrated with a JEOL NMR field meter ES-FC5. The g-values were determined by spectral simulation using JEOL anisimu ⁄ fa software, version 2.0.0. Acknowledgements The authors thank Professor Hayao Taguchi (Tokyo University of Science) for the gift of the plasmid carry- ing the LDH gene from T. caldophilus; Dr Miho Aoshima and Ms Ayako Nishiyama for the preparation of the recombinant LDH; and Dr Ki-Seok Yoon (Iba- raki University) for helpful discussions. This research was supported in part by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. References 1 Kerscher L & Oesterhelt D (1982) Pyruvate: ferredoxin oxidoreductase – new findings on an ancient enzyme. Trends Biochem Sci 7, 371–374. 2 Kawasumi T, Igarashi Y, Kodama T & Minoda Y (1984) Hydrogenobacter thermophilus gen. nov., sp. nov., an extremely thermophilic, aerobic, hydrogen-oxi- dizing bacterium. Int J Syst Bacteriol 34, 5–10. 3 Pitulle C, Yang Y, Marchiani M, Moore ERB, Siefert JL, Aragno M, Jurtshuk P Jr & Fox GE (1994) Phylo- genetic position of the genus Hydrogenobacter. Int J Syst Bacteriol 44, 620–626. 4 Shiba H, Kawasumi T, Igarashi Y, Kodama T & Minoda Y (1985) The CO 2 assimilation via the reduc- tive tricarboxylic acid cycle in an obligately autotrophic, aerobic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus. Arch Microbiol 141, 198–203. 5Hu ¨ gler M, Huber H, Stetter KO & Fuchs G (2003) Autotrophic CO 2 fixation pathways in archaea (Crenarchaeota). Arch Microbiol 179, 160–173. 6 Evans MCW, Buchanan BB & Arnon DI (1966) A new ferredoxin-dependent carbon reduction cycle in a photo- synthetic bacterium. Proc Natl Acad Sci USA 55, 928–934. 7 Buchanan BB & Arnon DI (1990) A reverse KREBS cycle in photosynthesis: consensus at last. Photosynth Res 24, 47–53. 8 Ikeda T, Yamamoto M, Arai H, Ohmori D, Ishii M & Igarashi Y (2005) Two tandemly arranged ferredoxin genes in the Hydrogenobacter thermophilus genome: comparative characterization of the recombinant [4Fe-4S] ferredoxins. Biosci Biotechnol Biochem 69, 1172–1177. Pyruvate synthesis by pyruvate oxidoreductase T. Ikeda et al. 508 FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 9 Adams MWW & Kletzin A (1996) Oxidoreductase-type enzymes and redox proteins involved in fermentative metabolisms of hyperthermophilic archaea. Adv Protein Chem 48, 101–180. 10 Drake HL, Hu S-I & Wood HG (1981) Purification of five components from Clostridium thermoaceticum which catalyze synthesis of acetate from pyruvate and methyl- tetrahydrofolate. Properties of phosphotransacetylase. J Biol Chem 256, 11137–11144. 11 Pieulle L, Guigliarelli B, Asso M, Dole F, Bernadac A & Hatchikian EC (1995) Isolation and characterization of the pyruvate-ferredoxin oxidoreductase from the sulfate-reducing bacterium Desulfovibrio africanus. Biochim Biophys Acta 1250, 49–59. 12 Kerscher L & Oesterhelt D (1981) Purification and proper- ties of two 2-oxoacid: ferredoxin oxidoreductases from Halobacterium halobium. Eur J Biochem 116, 587–594. 13 Zhang Q, Iwasaki T, Wakagi T & Oshima T (1996) 2-oxoacid: ferredoxin oxidoreductase from the thermo- acidophilic archaeon, Sulfolobus sp. strain 7. J Biochem 120, 587–599. 14 Blamey JM & Adams MWW (1994) Characterization of an ancestral type of pyruvate ferredoxin oxidoreductase from the hyperthermophilic bacterium, Thermotoga maritima. Biochemistry 33, 1000–1007. 15 Kletzin A & Adams MWW (1996) Molecular and phy- logenetic characterization of pyruvate and 2-ketoisoval- erate ferredoxin oxidoreductases from Pyrococcus furiosus and pyruvate ferredoxin oxidoreductase from Thermotoga maritima. J Bacteriol 178, 248–257. 16 Yun N-R, Yamamoto M, Arai H, Ishii M & Igarashi Y (2002) A novel five-subunit-type 2-oxoglutarate: ferre- doxin oxidoreductase from Hydrogenobacter thermophi- lus TK-6. Biochem Biophys Res Commun 292, 280–286. 17 Ikeda T, Ochiai T, Morita S, Nishiyama A, Yamada E, Arai H, Ishii M & Igarashi Y (2006) Anabolic five sub- unit-type pyruvate: ferredoxin oxidoreductase from Hydrogenobacter thermophilus TK-6. Biochem Biophys Res Commun 340, 76–82. 18 Deckert G, Warren PV, Gaasterland T, Young WG, Lenox AL, Graham DE, Overbeek R, Snead MA, Keller M, Aujay M et al. (1998) The complete genome of the hyperthermophilic bacterium Aquifex aeolicus. Nature 392, 353–358. 19 Chabrie ` re E, Charon M-H, Volbeda A, Pieulle L, Hatchikian EC & Fontecilla-Camps J-C (1999) Crystal structures of the key anaerobic enzyme pyruvate: ferredoxin oxidoreductase, free and in complex with pyruvate. Nat Struct Biol 6, 182–190. 20 Ragsdale SW (2003) Pyruvate ferredoxin oxidoreductase and its radical intermediate. Chem Rev 103, 2333–2346. 21 Tittmann K (2009) Reaction mechanisms of thiamin diphosphate enzymes: redox reactions. FEBS J 276, 2454–2468. 22 Thauer RK, Jungermann K & Decker K (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacteriol Rev 41, 100–180. 23 Furdui C & Ragsdale SW (2000) The role of pyruvate ferredoxin oxidoreductase in pyruvate synthesis during autotrophic growth by the Wood-Ljungdahl pathway. J Biol Chem 275, 28494–28499. 24 Yoon K-S, Hille R, Hemann C & Tabita FR (1999) Rubredoxin from the green sulfur bacterium Chloro- bium tepidum functions as an electron acceptor for pyruvate ferredoxin oxidoreductase. J Biol Chem 274, 29772–29778. 25 Yoon K-S, Bobst C, Hemann CF, Hille R & Tabita FR (2001) Spectroscopic and functional properties of novel 2[4Fe-4S] cluster-containing ferredoxins from the green sulfur bacterium Chlorobium tepidum. J Biol Chem 276, 44027–44036. 26 Ueda Y, Yamamoto M, Urasaki T, Arai H, Ishii M & Igarashi Y (2007) Sequencing and reverse transcription- polymerase chain reaction (RT-PCR) analysis of four hydrogenase gene clusters from an obligately autotroph- ic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus TK-6. J Biosci Bioeng 104, 470–475. 27 Tersteegen A & Hedderich R (1999) Methanobacterium thermoautotrophicum encodes two multisubunit mem- brane-bound [NiFe] hydrogenases. Transcription of the operons and sequence analysis of the deduced proteins. Eur J Biochem 264, 930–943. 28 Taguchi H, Yamashita M, Matsuzawa H & Ohta T (1982) Heat-stable and fructose 1,6-bisphosphate- activated l-lactate dehydrogenase from an extremely thermophilic bacterium. J Biochem 91, 1343–1348. 29 Robb FT, Park J-B & Adams MWW (1992) Character- ization of an extremely thermostable glutamate dehy- drogenase: a key enzyme in the primary metabolism of the hyperthermophilic archaebacterium, Pyrococcus furiosus. Biochim Biophys Acta 1120, 267–272. 30 Yoon K-S, Ishii M, Igarashi Y & Kodama T (1996) Purification and characterization of 2-oxoglutarate: ferredoxin oxidoreductase from a thermophilic, obli- gately chemolithoautotrophic bacterium, Hydrogeno- bacter thermophilus TK-6. J Bacteriol 178, 3365–3368. 31 Beinert H & Thomson AJ (1983) Three-iron clusters in iron-sulfur proteins. Arch Biochem Biophys 222, 333–361. 32 Blamey JM & Adams MWW (1993) Purification and characterization of pyruvate ferredoxin oxidoreductase from the hyperthermophilic archaeon Pyrococcus furiosus. Biochim Biophys Acta 1161, 19–27. 33 Mai X & Adams MWW (1994) Indolepyruvate ferredoxin oxidoreductase from the hyperthermophilic archaeon Pyrococcus furiosus. A new enzyme involved in peptide fermentation. J Biol Chem 269, 16726–16732. T. Ikeda et al. Pyruvate synthesis by pyruvate oxidoreductase FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS 509 34 Ozawa Y, Nakamura T, Kamata N, Yasujima D, Urushiyama A, Yamakura F, Ohmori D & Imai T (2005) Thermococcus profundus 2-ketoisovalerate ferredoxin oxidoreductase, a key enzyme in the archaeal energy-producing amino acid metabolic pathway. J Biochem 137, 101–107. 35 Uyeda K & Rabinowitz JC (1971) Pyruvate-ferredoxin oxidoreductase. IV. Studies on the reaction mechanism. J Biol Chem 246, 3120–3125. 36 Menon S & Ragsdale SW (1997) Mechanism of the Clostridium thermoaceticum pyruvate: ferredoxin oxido- reductase: evidence for the common catalytic intermedi- acy of the hydroxyethylthiamine pyrophosphate radical. Biochemistry 36, 8484–8494. 37 Chabrie ` re E, Verne ` de X, Guigliarelli B, Charon M-H, Hatchikian EC & Fontecilla-Camps JC (2001) Crystal structure of the free radical intermediate of pyruvate: ferredoxin oxidoreductase. Science 294 , 2559–2563. 38 Mansoorabadi SO, Seravalli J, Furdui C, Krymov V, Gerfen GJ, Begley TP, Melnick J, Ragsdale SW & Reed GH (2006) EPR spectroscopic and computational char- acterization of the hydroxyethylidene-thiamine pyro- phosphate radical intermediate of pyruvate: ferredoxin oxidoreductase. Biochemistry 45, 7122–7131. 39 Frank RAW, Kay CWM, Hirst J & Luisi BF (2008) Off-pathway, oxygen-dependent thiamine radical in the Krebs cycle. J Am Chem Soc 130, 1662–1668. 40 Cammack R, Kerscher L & Oesterhelt D (1980) A sta- ble free radical intermediate in the reaction of 2-oxo- acid: ferredoxin oxidoreductases of Halobacterium halobium. FEBS Lett 118 , 271–273. 41 Bock A-K, Scho ¨ nheit P & Teixeira M (1997) The iron- sulfur centers of the pyruvate: ferredoxin oxidoreduc- tase from Methanosarcina barkeri (Fusaro). FEBS Lett 414, 209–212. 42 Kerscher L & Oesterhelt D (1981) The catalytic mecha- nism of 2-oxoacid: ferredoxin oxidoreductases from Halobacterium halobium. One-electron transfer at two distinct steps of the catalytic cycle. Eur J Biochem 116, 595–600. 43 Sticht H & Ro ¨ sch P (1998) The structure of iron-sulfur proteins. Prog Biophys Mol Biol 70, 95–136. 44 Gehring U & Arnon DI (1972) Purification and proper- ties of a-ketoglutarate synthase from a photosynthetic bacterium. J Biol Chem 247, 6963–6969. 45 Yamamoto M, Ikeda T, Arai H, Ishii M & Igarashi Y (2009) Carboxylation reaction catalyzed by 2-oxogluta- rate: ferredoxin oxidoreductases from Hydrogenobacter thermophlius. Extremophiles doi: 10.1007/s00792-009- 0289-4. 46 Ikeda T, Nakamura M, Arai H, Ishii M & Igarashi Y (2009) Ferredoxin-NADP + reductase from the thermo- philic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus TK-6. FEMS Microbiol Lett 297, 124–130. 47 Kameya M, Ikeda T, Nakamura M, Arai H, Ishii M & Igarashi Y (2007) A novel ferredoxin-dependent gluta- mate synthase from the hydrogen-oxidizing chemoauto- trophic bacterium Hydrogenobacter thermophilus TK-6. J Bacteriol 189, 2805–2812. 48 Schneider G & Lindqvist Y (1998) Crystallography and mutagenesis of transketolase: mechanistic implications for enzymatic thiamin catalysis. Biochim Biophys Acta 1385, 387–398. 49 Furdui C & Ragsdale SW (2002) The roles of coenzyme A in the pyruvate: ferredoxin oxidoreductase reaction mechanism: rate enhancement of electron transfer from a radical intermediate to an iron-sulfur cluster. Biochemistry 41, 9921–9937. 50 Yoon K-S, Ishii M, Kodama T & Igarashi Y (1997) Carboxylation reactions of pyruvate: ferredoxin oxido- reductase and 2-oxoglutarate: ferredoxin oxidoreductase from Hydrogenobacter thermophilus TK-6. Biosci Biotechnol Biochem 61, 510–513. 51 Yoon K-S, Ishii M, Kodama T & Igarashi Y (1997) Purification and characterization of pyruvate: ferredoxin oxidoreductase from Hydrogenobacter thermophilus TK-6. Arch Microbiol 167, 275–279. 52 Yamamoto M, Arai H, Ishii M & Igarashi Y (2003) Char- acterization of two different 2-oxoglutarate: ferredoxin oxidoreductases from Hydrogenobacter thermophilus TK-6. Biochem Biophys Res Commun 312, 1297–1302. 53 Koide S, Iwata S, Matsuzawa H & Ohta T (1991) Crys- tallization of allosteric l-lactate dehydrogenase from Thermus caldophilus and preliminary crystallographic data. J Biochem 109, 6–7. 54 Chen J-S & Blanchard DK (1979) A simple hydroge- nase-linked assay for ferredoxin and flavodoxin. Anal Biochem 93, 216–222. 55 Moreno SNJ, Mason RP, Muniz RPA, Cruz FS & Docampo R (1983) Generation of free radicals from metronidazole and other nitroimidazoles by Tritricho- monas foetus. J Biol Chem 258, 4051–4054. 56 Ellman GL (1959) Tissue sulfhydryl groups. Arch Biochem Biophys 82, 70–77. 57 Barletta G, Chung AC, Rios CB, Jordan F & Schlegel JM (1990) Electrochemical oxidation of enamines related to the key intermediate on thiamin diphosphate dependent enzymatic pathways: evidence for one-elec- tron oxidation via a thiazolium cation radical. JAm Chem Soc 112, 8144–8149. Pyruvate synthesis by pyruvate oxidoreductase T. Ikeda et al. 510 FEBS Journal 277 (2010) 501–510 ª 2009 The Authors Journal compilation ª 2009 FEBS . Enzymatic and electron paramagnetic resonance studies of anabolic pyruvate synthesis by pyruvate: ferredoxin oxidoreductase from Hydrogenobacter thermophilus Takeshi. Molecular and phy- logenetic characterization of pyruvate and 2-ketoisoval- erate ferredoxin oxidoreductases from Pyrococcus furiosus and pyruvate ferredoxin oxidoreductase