TY - JOUR
AU - Shiga, Kiyoshi
AB - Abstract Electron-transferring flavoprotein (ETF) from the anaerobic bacterium Megasphaera elsdenii is a heterodimer containing two FAD cofactors. Isolated ETF contains only one FAD molecule, FAD-1, because the other, FAD-2, is lost during purification. FAD-2 is recovered by adding FAD to the isolated ETF. The two FAD molecules in holoETF were characterized using NADH. Spectrophotometric titration of isolated ETF with NADH showed a two-electron reduction of FAD-1 according to a monophasic profile indicating that FAD-1 receives electrons from NADH without involvement of FAD-2. When holoETF was titrated with NADH, FAD-2 was reduced to an anionic semiquinone and then was fully reduced before the reduction of FAD-1. The midpoint potential values at pH 7 were +81, −136 and –279 mV for the reduction of oxidized FAD-2 to semiquinone, semiquinone to the fully reduced FAD-2 and the two-electron reduction of FAD-1, respectively. Both FAD-1 and FAD-2 in holoETF were reduced by excess NADH very rapidly. The reduction of FAD-2 was slowed by replacement of FAD-1 with 8-cyano-FAD indicating that FAD-2 receives electrons from FAD-1 but not from NADH directly. The present results suggest that FAD-2 is the counterpart of the FAD in human ETF, which contains one FAD and one AMP. 8-CN-FAD, electron-transferring flavoprotein, Megasphaera elsdenii, midpoint potential, NADH In the rumen, many species of microbe produce lactate as a metabolite. The anaerobic bacterium Megasphaera elsdenii, formerly called Peptostreptococcus elsdenii (1), is the main utilizer of lactate in rumen (2). Megasphaera elsdenii uptakes lactate as the energy source and produces short-chain fatty acids, mainly acetate and propionate (3). Figure 1 shows the main metabolic pathway of M. elsdenii when utilizing lactate. Lactate exists in the ruminal fluid as a mixture of d- and l-stereoisomers, and they are interconverted by lactate racemase in the M. elsdenii cell (4). d-Lactate is oxidized by the flavoprotein d-lactate dehydrogenase (d-LDH) to pyruvate (5). Phosphoroclastic reaction of pyruvate produces acetate, CO2 and ATP (6). Although not all the enzymes participating in this reaction have been isolated, it was suggested that the reducing equivalents derived from pyruvate are used to produce H2 or NADH, where ferredoxin plays some role (7). Electrons kept by d-LDH flavin and NADH are consumed by propionate production from lactate, which is called the acrylate pathway (shown on the right side in Fig. 1). CoA transferase (7,8), lactyl-CoA dehydrase (7) and butyryl-CoA dehydrogenase (BCD) (9) have been isolated. Fig. 1 View largeDownload slide Main metabolic pathway of lactate-utilizing Megasphaera elsdenii. Solid arrows show the reaction and dashed arrows show the electron flow. Squares denote the enzymes: LR (lactate racemase), d-LDH (d-lactate dehydrogenase), CoAT (CoA transferase), LCD (lactyl-CoA dehydrase), BCD (butyryl-CoA dehydrogenase) and ETF (electron-transferring flavoprotein). Fig. 1 View largeDownload slide Main metabolic pathway of lactate-utilizing Megasphaera elsdenii. Solid arrows show the reaction and dashed arrows show the electron flow. Squares denote the enzymes: LR (lactate racemase), d-LDH (d-lactate dehydrogenase), CoAT (CoA transferase), LCD (lactyl-CoA dehydrase), BCD (butyryl-CoA dehydrogenase) and ETF (electron-transferring flavoprotein). Electron-transferring flavoprotein (ETF) accepts electrons from d-LDH and NADH and donates the electrons to BCD to reduce acrylyl-CoA to propionyl-CoA (9). ETF is a heterodimer containing two FAD molecules. One FAD molecule is lost during purification (10), thus the isolated ETF contains only one FAD. The isolated ETF becomes the holoprotein by adding exogenous FAD (10). In this article, the FAD contained in the isolated ETF is referred to as FAD-1, and the other FAD as FAD-2. The function of each FAD has not been solved, thus, it is unknown which FAD interacts with NADH, d-LDH and BCD. In this study, we investigated the reduction of ETF by NADH and revealed that NADH interacts with only FAD-1 and that FAD-2 accepts electrons from reduced FAD-1. Materials and Methods Chemicals Thionin acetate and NADH were from Sigma. Phenosafranin was from Aldrich. Indigo disulfonate was from Nacalai Tesque. FAD purchased from Nacalai Tesque was further purified as described (10). 8-Cyano-flavin adenine dinucleotide (8-CN-FAD) (11) and ferricenium hexafluorophosphate (Fc+) (12) were prepared as described. Spectrophotometer Spectrophotometric titration was carried out using a Hitachi U-3210 spectrophotometer. Time-resolved spectra were measured by a Shimadzu MultiSpec-1500 photodiode array spectrometer. Anaerobic experiments Anaerobic titration was performed using a custom-made sealed-up cuvette combined with a Hamilton gas-tight syringe. Anaerobiosis was achieved by cycles of gentle evacuation and Ar introduction. Anaerobic titrant was added from the gas-tight syringe. An oxygen scrubber (13) was used during anaerobic titration. Other anaerobic spectroscopic experiments were carried out with cuvettes in which the sample solution contacts the gas phase of the Ar stream (14). Determination of midpoint redox potentials Standard redox potentials and absorption spectra of oxidized and reduced forms of redox indicator dyes were obtained by anaerobic potentiometric spectroscopy at each pH value, 6.0, 6.5 and 7.0. A custom-made cell similar to that illustrated previously [Fig. 2 in Cammack’s article (15) or Fig. 4A in Dutton’s paper (16)] combined with a HORIBA 6860-10C combined Pt/Ag-AgCl electrode was used. The electrode was calibrated with 6 mg/ml quinhydrone in 50 mM K-Pi, pH 7.0 at 25°C (286 mV versus standard hydrogen electrode) (15). A 7-ml sample was continuously bubbled with Ar gas. A sample of ∼10 µM dye was titrated with Na-dithionite for reduction and with O2 or K3Fe(CN)6 for reoxidation. The midpoint potential (), the number of electrons associated with the redox reaction (ndye) and the absorption spectra of oxidized and reduced forms of the indicator dye were determined by analysing the data set of the absorption spectra at various redox potentials according to the equation,   (1) where E is the redox potential of the system, R is the gas constant, T is the absolute temperature and F is the Faraday constant. Fig. 2 View largeDownload slide Anaerobic titration of isolated ETF with NADH. Isolated ETF (10 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. (A) Spectra of isolated ETF at titration points. The arrow indicates the spectrum before the addition of NADH. The spectra were corrected for dilution by the addition of titrant. (B) Titration profiles of absorbance values at 450 and 370 nm. Fig. 2 View largeDownload slide Anaerobic titration of isolated ETF with NADH. Isolated ETF (10 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. (A) Spectra of isolated ETF at titration points. The arrow indicates the spectrum before the addition of NADH. The spectra were corrected for dilution by the addition of titrant. (B) Titration profiles of absorbance values at 450 and 370 nm. Midpoint potentials of ETF were determined by anaerobic titration of ETF with NADH in the presence of an indicator dye. The absorption spectrum was measured at the equilibrium after each addition of NADH. Each spectrum was resolved into the spectra of the four redox states of ETF, oxidized and reduced dye, and NADH by a linear least-squares method using the predetermined spectra of the components. Thus, the concentrations of each redox state of ETF and the dye were determined. The midpoint potential of the redox couple in question () was determined according to the equation,   (2) Protein preparation ETF was purified from M. elsdenii as described (10). HoloETF was prepared by adding excess FAD to isolated ETF (∼2 : 1), followed by washing [repeated cycles of ultrafiltration/dilution using a Microcon YM-30 (Millipore) at 5°C] with 50 mM potassium phosphate buffer, pH 6.0, containing 5 µM FAD. The FAD in the washing buffer prevented loss of bound FAD-2. Finally, the protein was concentrated to >1 mM and then diluted with buffer to give an appropriate stock solution where the free FAD content was <0.5% of ETF. The molar ratio of FAD-2/FAD-1 was 0.75–0.8 due to the loss of FAD-2-binding ability during purification and storage (10). Replacement of FAD-1 with 8-CN-FAD A solution of denatured subunits free from FAD was prepared by washing isolated ETF with 50 mM potassium phosphate buffer, pH 6.0, containing 6 M guanidine hydrochloride and 10 mM dithiothreitol (DTT). The solution of denatured subunits (∼200 µM of each subunit) was then quickly diluted to 2 µM in the buffer containing 10 µM 8-CN-FAD and 1 mM DTT, and incubated at 25°C with light shielding for 2 h. The refolded ETF contained oxidized 8-CN-FAD in the FAD-1-binding site and semiquinoid 8-CN-FAD in the FAD-2-binding site. The reduction to the semiquinone of 8-CN-FAD was because of the presence of DTT (reducer) and the high midpoint potential of 8-CN-FAD in the FAD-2-binding site (see ‘Results’ section). The incubated solution was washed with the buffer containing 1 M KBr and 0.5 mM Fc+ for the purpose of removing the 8-CN-FAD bound to the FAD-2-binding site and then with the buffer to remove KBr and Fc+. Here, Fc+ was used for oxidation of the semiquinoid 8-CN-FAD: the semiquinoid 8-CN-FAD was not released even in the presence of 1 M KBr. At this time, the reconstituted ETF contained one 8-CN-FAD molecule in the FAD-1-binding site (denoted by apo-CN(1)-ETF). To this protein, 1.5 mol FAD was added per mol ETF and then washed with the buffer containing 10 µM FAD. Finally, the protein was concentrated to >1 mM and then diluted with buffer to give an appropriate stock solution. The resulting protein was reconstituted ETF containing 8-CN-FAD in the FAD-1-binding site and FAD in the FAD-2-binding site (denoted by holo-CN(1)-ETF) with free FAD <1% of ETF. Extinction coefficients Concentrations were determined using the following molar extinction coefficient values: ε340 = 6.22 mM–1 cm–1 for NADH, ε450 = 11.3 mM–1 cm–1 for FAD, ε450 = 11 mM–1 cm–1 for 8-CN-FAD (17), ε450 = 11.3 mM–1 cm–1 for isolated ETF (10) and ε462 = 11.2 mM–1 cm–1 for apo-CN(1)-ETF. The ε-value of apo-CN(1)-ETF was determined as described (18). Results Titration of ETF with NADH When it was titrated with NADH, isolated ETF was reduced monophasically (Fig. 2), thus no semiquinone was observed (two-electron manner). On the other hand, holoETF was reduced triphasically (Fig. 3): one FAD was reduced in a one-electron manner. It was first reduced to the red (anionic) semiquinone form (a–b in Fig. 3), second to the fully reduced form (b–c), followed by a monophasic reduction of the other FAD (c–d). Because FAD-1 is reduced monophasically (Fig. 2), the FAD which is reduced in a one-electron manner must be FAD-2. In both titrations, the FAD/NADH stoichiometry was 1 : 1 and no further NADH oxidation was observed after the reduction of the two FADs judging by the absorption at 340 nm, where NADH shows maximum absorbance. These results indicate that ETF contains no other redox active centres which contribute to the oxidation of NADH. Fig. 3 View largeDownload slide Anaerobic titration of holoETF with NADH. HoloETF (9.1 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. Spectra were corrected for dilution by the addition of titrant. (A) Spectra at the breaking points (a–d) on the titration profile shown in panel B were obtained by the intersections of the lines at each wavelength. (B) Titration profile of absorbance values at 450 and 375 nm. Fig. 3 View largeDownload slide Anaerobic titration of holoETF with NADH. HoloETF (9.1 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. Spectra were corrected for dilution by the addition of titrant. (A) Spectra at the breaking points (a–d) on the titration profile shown in panel B were obtained by the intersections of the lines at each wavelength. (B) Titration profile of absorbance values at 450 and 375 nm. Because NADH donates two electrons simultaneously, there remains the possibility that the equilibrium between the oxidized and two-electron reduced forms of FAD-1 observed during the NADH titration (c–d in Fig. 3) is only the kinetically stable state and that the semiquinoid form of FAD-1 is thermodynamically stable. Titration of holoETF with dithionite, a one-electron reagent, showed a triphasic titration profile, a similar result to that of NADH titration (data not shown). This eliminates the possibility that FAD-1 stabilizes the semiquinoid form thermodynamically. Midpoint potentials of redox couples associated with three steps The midpoint potential of the first step of holoETF titration, or oxidized to semiquinoid reduction of FAD-2, was determined by NADH titration of holoETF in the presence of thionin as the redox indicator. Figure 4 shows the relationship between the FAD-2 and thionin redox states at pH 6. The redox parameters of thionin was determined in advance to be ndye = 2 and = 97 mV at pH 6. The slope of the plot in Fig. 4, nETF/ndye = 0.5 [see Equation (2)], means nETF = 1, which is consistent with the oxidized to semiquinoid reduction of FAD (one electron reduction). From the -intercept and the value of thionin, the value for the oxidized/semiquinone couple of FAD-2 at pH 6 was determined to be 85 mV. Other redox couples were also investigated using similar procedures. Figure 5 summarizes the results of the midpoint potentials of all the steps under different pH conditions. The midpoint potential of FAD-2 (oxidized/semiquinoid) is independent of pH, consistent with the reduction of the neutral oxidized form to the anionic semiquinoid form, where the electron transfer is not accompanied by protonation. Dependence of the midpoint potential of FAD-2 (semiquinoid/reduced) on pH agrees well with the slope of –59.2 mV/pH, which means a one electron transfer is accompanied by a single protonation. This suggests that the two-electron reduced form of FAD-2 is anionic. Dependence of the midpoint potential of FAD-1 (oxidized/reduced) agrees well with the slope of –29.6 mV/pH, suggesting that the fully reduced form of FAD-1 is anionic. The midpoint potential values at pH 7 were +81 (ox/sq of FAD-2), –136 (sq/red of FAD-2) and –279 mV (ox/red of FAD-1) according to the lines in Fig. 5. Fig. 4 View largeDownload slide Equilibrium between oxidized/semiquinoid couple of FAD-2 and oxidized/reduced couple of thionin. An anaerobic solution of holoETF (6 µM) and thionin acetate (2 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. The plotted values were obtained at different titration points. Fig. 4 View largeDownload slide Equilibrium between oxidized/semiquinoid couple of FAD-2 and oxidized/reduced couple of thionin. An anaerobic solution of holoETF (6 µM) and thionin acetate (2 µM) in 50 mM K-Pi (pH 6.0) was titrated with NADH at 25°C. The plotted values were obtained at different titration points. Fig. 5 View largeDownload slide Standard redox potentials of the three steps of oxidation/reduction of holoETF. The values were obtained in 50 mM K-Pi buffer at different pH values at 25°C. Oxidized, semiquinoid and fully reduced states were designated as ox, sq and red, respectively. ETF was titrated with NADH in the presence of a redox indicator dye, thionin acetate for FAD-2 (ox/sq), indigo disulfonate for FAD-2 (sq/red) or phenosafranin for FAD-1 (ox/red). Fig. 5 View largeDownload slide Standard redox potentials of the three steps of oxidation/reduction of holoETF. The values were obtained in 50 mM K-Pi buffer at different pH values at 25°C. Oxidized, semiquinoid and fully reduced states were designated as ox, sq and red, respectively. ETF was titrated with NADH in the presence of a redox indicator dye, thionin acetate for FAD-2 (ox/sq), indigo disulfonate for FAD-2 (sq/red) or phenosafranin for FAD-1 (ox/red). The equilibration between ETF flavin and the redox indicator dye required several tens of minutes for each titration point. Therefore, measurement of the midpoint potential of FAD-1 of isolated ETF was not possible because prolonged reduction of isolated ETF causes the release of FAD-1 and its rebinding to the FAD-2-binding site with a half completion time of ∼30 min (10). Kinetics of ETF reduction by NADH The results of NADH titration of isolated ETF and holoETF showed that FAD-1 accepts electrons from NADH directly, thus without mediacy of FAD-2. On the other hand, it remains unclear whether FAD-2 receives electrons from NADH directly like FAD-1 or whether FAD-2 receives electrons from reduced FAD-1. Further kinetic studies on the ETF reduction by NADH are necessary to clarify this. However, the reduction of ETF by excess NADH was too rapid to measure the time course even by using a stopped flow apparatus (dead time of <0.1 s). For the purpose of measuring the time course, we used an FAD analogue as described below. Replacement of FAD-1 with 8-CN-FAD 8-CN-FAD has a cyano group at the 8-position of the isoalloxazine ring instead of the methyl group of normal FAD. The midpoint potential of 8-CN-FAD is much more positive than normal FAD by 160 mV (17). If FAD-2 receives electrons from FAD-1, the reduction of FAD-2 should be inhibited by replacing FAD-1 with 8-CN-FAD. Curve a in Fig. 6 shows the spectrum of apo-CN(1)-ETF, which has 8-CN-FAD in the FAD-1-binding site and no FAD in the FAD-2-binding site. The large shift of the second absorption band to a shorter wavelength (from ∼370 nm for normal flavin to ∼340 nm) is a characteristic feature of 8-CN-flavin (17). Titration of apo-CN(1)-ETF with FAD (inset to Fig. 6) indicated 1 mol apo-CN(1)-ETF bound 0.74 mol FAD, reflecting a slight loss of FAD-2-binding ability during the preparation of apo-CN(1)-ETF. Curve c in Fig. 6 shows the difference spectrum between the FAD-saturated form (holo-CN(1)-ETF, cure b) and apo-CN(1)-ETF (curve a). This unusual flavin spectrum is almost the same as the difference spectrum of holoETF minus isolated ETF [curve a in Fig. 5 in our previous paper (10)], indicating that the absorption spectra of FAD-1 and FAD-2 are not perturbed significantly by each other. The dissociation constant of FAD-2 from holo-CN(1)-ETF was 0.23 µM, which is comparable with that of FAD-2 from holoETF (0.07 µM) (10). Fig. 6 View largeDownload slide Titration of apo-CN(1)-ETF with FAD. Apo-CN(1)-ETF (8.5 µM) in 50 mM K-Pi (pH 6.0) was titrated with FAD at 25°C. The inset shows the titration profile of the 400-nm absorption with the horizontal and perpendicular lines indicating a 1 : 1 point. The spectra of apo-CN(1)-ETF (a), FAD-saturated form (b) and the difference between a and b (c) were calculated as described previously (10). Fig. 6 View largeDownload slide Titration of apo-CN(1)-ETF with FAD. Apo-CN(1)-ETF (8.5 µM) in 50 mM K-Pi (pH 6.0) was titrated with FAD at 25°C. The inset shows the titration profile of the 400-nm absorption with the horizontal and perpendicular lines indicating a 1 : 1 point. The spectra of apo-CN(1)-ETF (a), FAD-saturated form (b) and the difference between a and b (c) were calculated as described previously (10). Spectrophotometric titration of holo-CN(1)-ETF with NADH was unsuccessful because it took a long time (several hours) to attain the equilibrium after each addition. Apo-CN(1)-ETF was titrated with NADH with rapid equilibration after each addition as was the case for isolated ETF, and the spectra of oxidized and reduced forms of 8-CN-FAD-1, which are used below, were obtained (data not shown). Kinetics of reduction of holo-CN(1)-ETF by NADH Figure 7A shows the absorption spectra of holo-CN(1)-ETF before and after the addition of excess NADH. Immediately after the addition of NADH, a large part of the flavin molecules were reduced and then the residual reduction proceeded slowly. This is quite different from the reduction of holoETF with excess NADH, where all the flavin molecules were reduced immediately. The time-resolved spectra were analysed with the spectra obtained above to calculate the concentrations of oxidized, semiquinoid and reduced FAD-2, oxidized and reduced 8-CN-FAD-1 and NADH. Figure 7B shows the time course. Immediately after the addition of NADH, 8-CN-FAD-1 was fully reduced and FAD-2 was one electron-reduced. Then, the semiquinoid FAD-2 was slowly reduced to the fully reduced form. Namely, the reduction of FAD-2 was obviously inhibited by the replacement of FAD-1 by 8-CN-FAD, indicating that FAD-2 receives electrons from FAD-1, but not from NADH directly. The reduction of FAD-2 via two steps of one-electron transfer also removes the possibility for direct reduction by NADH because NADH donates two electrons simultaneously. FAD-1 in holoETF and 8-CN-FAD-1 in holo-CN(1)-ETF receive two electrons from NADH simultaneously and donate the electrons to FAD-2 one by one. The reduction rate (inverse of the half completion time) of the semiquinone to the reduced form of FAD-2 was non-linearly raised by increasing the NADH concentration with constant holo-CN(1)-ETF concentration, and also non-linearly raised by increasing the holo-CN(1)-ETF concentration with constant NADH concentration (data not shown). Although the mechanism for the electron transfer from FAD-1 or 8-CN-FAD-1 to FAD-2 is presently unclear, the dependence of the rate on holo-CN(1)-ETF concentration suggests an intermolecular interaction, i.e. electron transfer from reduced FAD-1 of one ETF molecule to the semiquinoid FAD-2 of another ETF molecule. Fig. 7 View largeDownload slide Reduction of holo-CN(1)-ETF by excess NADH. (A) The time-resolved spectra of anaerobic holo-CN(1)-ETF (4.3 µM) in 50 mM K-Pi (pH 6.0) at 25°C before and after the addition of 40 µM NADH. Only five spectra are shown for clarity. (B) The time courses of molecular species of different redox states. The concentrations of NADH and oxidized forms of 8-CN-FAD-1 and FAD-2 are not shown for clarity. Fig. 7 View largeDownload slide Reduction of holo-CN(1)-ETF by excess NADH. (A) The time-resolved spectra of anaerobic holo-CN(1)-ETF (4.3 µM) in 50 mM K-Pi (pH 6.0) at 25°C before and after the addition of 40 µM NADH. Only five spectra are shown for clarity. (B) The time courses of molecular species of different redox states. The concentrations of NADH and oxidized forms of 8-CN-FAD-1 and FAD-2 are not shown for clarity. Discussion NADH-titration (19) and measurement of midpoint potentials (20) of M. elsdenii ETF were reported previously by other researchers. They used the protein purified by the method reported in 1973 (19) without addition of exogenous FAD. The isolated ETF they used contained ∼0.03–0.4 mol FAD-2 per mol ETF and 1–50% of the bound FAD was modified as 6-OH-FAD or 8-OH-FAD. The content of FAD and the percentage of modified flavins were varied from preparation to preparation. In addition, if the FAD-2-binding site is vacant, prolonged reduction of FAD-1 results in the liberation of the FAD from the FAD-1-binding site and its rebinding to the FAD-2-binding site of another ETF molecule having FAD-1 to form holoETF (10,19). These facts made the interpretation of the experimental results unclear. Our improved purification method reported in 2003 (10) enabled the preparation of ETF with only FAD-1 (isolated ETF) and ETF with both FAD-1 and FAD-2 (holoETF) resulting in an experimental system with defined components. In this study, we clarified that NADH interacts with FAD-1 not FAD-2. As shown in Fig. 8, the more positive midpoint potential of FAD-1 (ox/red) than NAD+/NADH is advantageous for the reduction of FAD-1 by NADH. FAD-1 donates electrons to FAD-2, which has more positive midpoint potentials. Our unpublished data indicate that d-LDH and BCD interacts with FAD-2 not FAD-1. The more positive midpoint potentials of FAD-2, for both the ox/sq and sq/red redox pairs, than pyruvate/lactate and d-LDH are advantageous for the reduction of FAD-2 by the lactate-reduced d-LDH. The midpoint potentials of acrylyl-CoA/propionyl-CoA and BCD fall between the ox/sq and sq/red of FAD-2, suggesting that the semiquinoid/reduced couple rather than the oxidized/semiquinoid couple of FAD-2 works preferably for the reduction of BCD. Fig. 8 View largeDownload slide Midpoint potentials of ETF-related redox couples in M. elsdenii at pH 7. The values of FAD-1 and FAD-2 are the present results. The values of acrylyl-CoA/propionyl-CoA (14), BCD (21), pyruvate/lactate (22), d-LDH (22) and NAD+/NADH (23) are from the literature. Fig. 8 View largeDownload slide Midpoint potentials of ETF-related redox couples in M. elsdenii at pH 7. The values of FAD-1 and FAD-2 are the present results. The values of acrylyl-CoA/propionyl-CoA (14), BCD (21), pyruvate/lactate (22), d-LDH (22) and NAD+/NADH (23) are from the literature. ETFs have been isolated and characterized from other sources such as human (24,25), pig (13,26), Methylophilus methylotrophus (27,28) and Paracoccus denitrificans (29, 30). These ETFs are also heterodimeric but contain only one FAD and additionally one AMP (24,26,30,31). AMP has no effect on enzymatic activity but is important in the assembly process of the four components to form the holoprotein (32). Similar to M. elsdenii ETF, these ETFs catalyze the electron transfer between other flavoproteins such as acyl-CoA dehydrogenases, sarcosine dehydrogenase, ETF-CoQ oxidoreductase and trimethylamine dehydrogenase, depending on the source (33). However, unlike M. elsdenii ETF, these ETFs do not interact with NADH. Crystal structures of ETFs of human (PDB ID: 1efv) (24), M. methylotrophus (1o96, 1o97) (34) and P. denitrificans (1efp) (35) have been solved, whereas the structure of M. elsdenii ETF has not yet been solved. Therefore, it is uncertain which of FAD-1, FAD-2 and NADH of M. elsdenii ETF corresponds to the FAD or AMP in the other ETFs. We previously suggested that FAD-1 and FAD-2 correspond to AMP and FAD of other ETFs, respectively (10). This was based on the fact that (i) in KBr solution, FAD-2 of M. elsdenii ETF and FAD of other ETFs are liberated whereas FAD-1 and AMP remain bound tightly to the apoprotein and (ii) an unusual absorption spectrum similar to that of FAD-2 (spectrum c in Fig. 6) is obtained for pig ETF under a certain condition. All the other ETFs also stabilize the anionic semiquinone during redox titration, similar to FAD-2 of M. elsdenii ETF. Midpoint potentials at pH 7 of the oxidized/semiquinoid (Eox/sq) and semiquinoid/reduced (Esq/red) couples of the other ETFs have been reported as (+37, +25 mV) (25) and (+22, –42 mV) (36) for human ETF; (–14, –30 mV) (37) and (+4, –20 mV) (38) for pig ETF; (–36, –6 mV) for P. denitrificans ETF (30, 35) and (+196, –197 mV) (39), (+141, ND mV) (40) and (+153, < –250 mV) (28) for M. methylotrophus ETF. [Here, if the value had been reported for another pH value than pH 7, it was corrected to pH 7 using the pH dependence shown in Fig. 5. The values for P. denitrificans ETF were calculated by combining the results of the two papers. ND means not determined. The reduction of the semiquinone form to the fully reduced form of M. methylotrophus ETF is very slow, thus it is difficult to determine the Esq/red value]. The midpoint potential of the two-electron reduction, which is the average of Eox/sq and Esq/red, is near 0 mV for all ETFs including FAD-2 of M. elsdenii ETF. The difference between the two midpoint potentials of one-electron reductions (Eox/sq – Esq/red) is small for human, pig, and P. denitrificans ETFs, while it is very large for M. methylotrophus ETF. [The Eox/sq value of M. methylotrophus ETF is the most positive of the reported midpoint potentials of flavoproteins]. Thus, the extent of the thermodynamic stabilization of the semiquinone is in the order: P. denitrificans < human ≈ pig << M. methylotrophus. FAD-2 of M. elsdenii ETF shows features which are intermediate between human/pig and M. methylotrophus. These similarities in the thermodynamic redox features support the prediction that FAD-2 of M. elsdenii ETF corresponds to the FAD of other ETFs. Our unpublished result that d-LDH and BCD interact with FAD-2 but not with FAD-1 is also consistent with this prediction. As a summary, the electron transfer pathway for the reduction of acrylyl-CoA by NADH in M. elsdenii (Fig. 1) is represented as NADH → FAD-1 → FAD-2 → BCD → acrylyl-CoA. FAD-1 is replaced with AMP in ETFs from other sources so that these ETFs do not interact with NADH. The electron transfer pathway from d-LDH to BCD is represented as d-LDH → FAD-2 → BCD. The FAD in other ETFs, the counterpart of FAD-2, also participates in the electron transfer among the other flavoproteins. In the last decade, ETFs were isolated from the anaerobic bacteria, Clostridium propionicum and Clostridium kluyveri, as complexes with acyl-CoA dehydrogenases (41,42). Unfortunately, the FAD contents were not reported for both these complexes. The C. propionicum complex does not contain AMP. Whether the C. kluyveri complex contains AMP was unmentioned. However, both complexes oxidize NADH. Therefore, it is conceivable that these ETFs are similar to M. elsdenii ETF (with two FADs) rather than the other ETFs (with FAD and AMP). It is interesting that the C. kluyveri complex couples the two reactions: (i) reduction of crotonyl-CoA to butyryl-CoA by NADH and (ii) reduction of ferredoxin by NADH. Thus, the reduction of crotonyl-CoA by NADH must accompany the reduction of ferredoxin (42). The mechanisms, such as the electron transfer pathway among the FADs in the complex, are presently unclear, although a hypothesis for a possible mechanism was previously presented (43). Our present results will contribute to further investigations into these acyl-CoA dehydrogenase–ETF complexes. Conflict of Interest None declared. Abbreviations Abbreviations BCD butyryl-CoA dehydrogenase 8-CN-FAD 8-cyano-flavin adenine dinucleotide d-LDH d-lactate dehydrogenase ETF electron-transferring flavoprotein Fc+PF6− ferricenium hexafluorophosphate References 1 Rogosa M.  Transfer of Peptostreptococcus elsdenii Gutierrez et al. to a new genus, Megasphaera [M. elsdenii (Gutierrez et al.) cmb. Nov.],  Int. J. Sys. Bacteriol. ,  1971, vol.  21 (pg.  187- 189) Google Scholar CrossRef Search ADS   2 Counotte GHM,  Prins RA,  Janssen AM,  deBie MJA.  Role of Megasphaera elsdenii in the fermentation of dl-[2-13C]lactate in the rumen of dairy cattle,  Appl. Environ. Microbiol. ,  1981, vol.  42 (pg.  649- 655) Google Scholar PubMed  3 Marounek M,  Fliegrova K,  Bartos S.  Metabolism and some characteristics of ruminal strains of Megasphaera elsdenii,  Appl. Environ. Microbiol. ,  1989, vol.  55 (pg.  1570- 1573) Google Scholar PubMed  4 Hino T,  Kuroda S.  Presence of lactate dehydrogenase and lactate racemase in Megasphaera elsdenii grown on glucose or lactate,  Appl. Environ. Microbiol. ,  1993, vol.  59 (pg.  225- 259) 5 Brockman HL,  Wood WA.  d-Lactate dehydrogenase of Peptostreptococcus elsdenii,  J. Bacteriol. ,  1975, vol.  124 (pg.  1454- 1461) Google Scholar PubMed  6 Peel JL.  The breakdown of pyruvate by cell-free extracts of the rumen micro-organism LG,  Biochem. J. ,  1960, vol.  74 (pg.  525- 541) Google Scholar CrossRef Search ADS PubMed  7 Baldwin RL,  Milligan LP.  Electron transport in Peptostreptococcus elsdenii,  Biochim. Biophys. Acta ,  1964, vol.  92 (pg.  421- 432) Google Scholar PubMed  8 Tung KK,  Wood WA.  Purification, new assay, and properties of coenzyme A transferase from Peptostreptococcus elsdenii,  J. Bacteriol. ,  1975, vol.  124 (pg.  1462- 1474) Google Scholar PubMed  9 Brockman HL,  Wood WA.  Electron-transferring flavoprotein of Peptostreptococcus elsdenii that functions in the reduction of acrylyl-coenzyme A,  J. Bacteriol. ,  1975, vol.  124 (pg.  1447- 1453) Google Scholar PubMed  10 Sato K,  Nishina Y,  Shiga K.  Purification of electron-transferring flavoprotein from Megasphaera elsdenii and binding of additional FAD with an unusual absorption spectrum,  J. Biochem. ,  2003, vol.  134 (pg.  719- 729) Google Scholar CrossRef Search ADS PubMed  11 Nishina Y,  Sato K,  Setoyama C,  Tamaoki H,  Miura R,  Shiga K.  Intramolecular and intermolecular perturbation on electronic state of FAD free in solution and bound to flavoproteins: FTIR spectroscopic study by using the C=O stretching vibrations as probes,  J. Biochem. ,  2007, vol.  142 (pg.  265- 272) Google Scholar CrossRef Search ADS PubMed  12 Lehman TC,  Hale DE,  Bhala A,  Thorpe C.  An acyl-coenzyme A dehydrogenase assay utilizing the ferricenium ion,  Anal. Biochem. ,  1990, vol.  186 (pg.  280- 284) Google Scholar CrossRef Search ADS PubMed  13 Gorelick RJ,  Mizzer JP,  Thorpe C.  Purification and properties of electron-transferring flavoprotein form pig kidney,  Biochemistry ,  1982, vol.  21 (pg.  6936- 6942) Google Scholar CrossRef Search ADS PubMed  14 Sato K,  Nishina Y,  Setoyama C,  Miura R,  Shiga K.  Unusually high standard redox potential of acrylyl-CoA/propionyl-CoA couples among enoyl-CoA/acyl-CoA couples: a reason for the distinct metabolic pathway of propionyl-CoA from longer acyl-CoAs,  J. Biochem. ,  1999, vol.  126 (pg.  668- 675) Google Scholar CrossRef Search ADS PubMed  15 Cammack R.  Brown GC,  Cooper CE.  Redox state and potentials,  Bioenergetics: A Practical Approach ,  1995 Oxford Oxford University Press(pg.  85- 109) 16 Dutton PL.  Redox potentiometry: determination of midpoint potentials of oxidation-reduction components of biological electron-transfer systems,  Methods Enzymol. ,  1978, vol.  54 (pg.  411- 435) Google Scholar PubMed  17 Murthy YVSN,  Massey V.  Synthesis and properties of 8-CN-flavin nucleotide analogs and studies with flavoproteins,  J. Biol. Chem. ,  1998, vol.  273 (pg.  8975- 8982) Google Scholar CrossRef Search ADS PubMed  18 Thorpe C,  Matthews RG,  Williams CH.  Acyl-coenzyme A dehydrogenase from pig kidney. Purification and properties,  Biochemistry ,  1979, vol.  18 (pg.  331- 337) Google Scholar CrossRef Search ADS PubMed  19 Whitfield CD,  Mayhew SG.  Purification and properties of electron-transferring flavoprotein from Peptostreptococcus elsdenii,  J. Biol. Chem. ,  1974, vol.  249 (pg.  2801- 2810) Google Scholar PubMed  20 Pace CP,  Stankovich MT.  Redox properties of electron-transferring flavoprotein from Megasphaera elsdenii,  Biochim. Biophys. Acta ,  1987, vol.  911 (pg.  267- 276) Google Scholar CrossRef Search ADS PubMed  21 Stankovich MT,  Soltysik S.  Regulation of the butyryl-CoA dehydrogenase by substrate and product binding,  Biochemistry ,  1987, vol.  26 (pg.  2627- 2632) Google Scholar CrossRef Search ADS PubMed  22 Olson ST,  Massey V.  Purification and properties of flavoenzyme d-lactate dehydrogenase from Megasphaera elsdenii,  Biochemistry ,  1979, vol.  18 (pg.  4714- 4724) Google Scholar CrossRef Search ADS PubMed  23 Thauer RK,  Jungermann K,  Decher K.  Energy conservation in chemotrophic anaerobic bacteria,  Bacteriol. Rev. ,  1977, vol.  41 (pg.  100- 180) Google Scholar PubMed  24 Roberts DL,  Frerman FE,  Kim JJP.  Three-dimensional structure of human electron transfer flavoprotein to 2.1-Å resolution,  Proc. Natl. Acad. Sci. U. S. A. ,  1996, vol.  93 (pg.  14355- 14360) Google Scholar CrossRef Search ADS PubMed  25 Salazar D,  Zhang L,  deGala GD,  Frerman FE.  Expression and characterization of two pathogenic mutations in human electron transfer flavoprotein,  J. Biol. Chem. ,  1997, vol.  272 (pg.  26425- 26433) Google Scholar CrossRef Search ADS PubMed  26 Sato K,  Nishina Y,  Shiga K.  Electron-transferring flavoprotein has an AMP-binding site in addition to the FAD-binding site,  J. Biochem. ,  1993, vol.  114 (pg.  215- 222) Google Scholar PubMed  27 Steenkamp DJ,  Gallup M.  The natural flavoprotein electron acceptor of trimethylamine dehydrogenase,  J. Biol. Chem. ,  1978, vol.  253 (pg.  4086- 4089) Google Scholar PubMed  28 Talfournier F,  Munro AW,  Basran J,  Sutcliffe MJ,  Daff S,  Chapman SK,  Scrutton NS.  αArg-237 in Methylophilus methylotrophus (sp. W3A1) electron-transferring flavoprotein affords ∼200-millivolt stabilization of the FAD anionic semiquinone and a kinetic block on full reduction to the dihydroquinone,  J. Biol. Chem. ,  2001, vol.  276 (pg.  20190- 20196) Google Scholar CrossRef Search ADS PubMed  29 Husain M,  Steenkamp DJ.  Partial purification and characterization of glutaryl-coenzyme A dehydrogenase, electron transfer flavoprotein, and electron transfer flavoprotein-Q oxidoreductase from Paracoccus denitrificans,  J. Bacteriol. ,  1985, vol.  163 (pg.  709- 715) Google Scholar PubMed  30 Griffin KJ,  Dwyer TM,  Manning MC,  Meyer JD,  Carpenter JF,  Frerman FE.  αT244M mutation affects the redox, kinetic, and in vitro folding properties of Paracoccus denitrificans electron transfer flavoprotein,  Biochemistry ,  1997, vol.  36 (pg.  4194- 4202) Google Scholar CrossRef Search ADS PubMed  31 DuPlessis ER,  Rohlfs RJ,  Hille R,  Thorpe C.  Electron-transferring flavoprotein from pig and the methylotrophic bacterium W3A1 contains AMP as well as FAD,  Biochem. Mol. Biol. Int. ,  1994, vol.  32 (pg.  195- 199) Google Scholar PubMed  32 Sato K,  Nishina Y,  Shiga K.  In vitro assembly of FAD, AMP, and the two subunits of electron-transferring flavoprotein: an important role of AMP related with the conformational change of the apoprotein,  J. Biochem. ,  1997, vol.  121 (pg.  477- 486) Google Scholar CrossRef Search ADS PubMed  33 Thorpe C.  Müller F.  Electron-transferring flavoprotein,  Chemistry and Biochemistry of Flavoenzymes ,  1991, vol.  Vol. 2  Boston, London CRC Press, Boca Raton, Ann Arbor(pg.  471- 486) 34 Leys D,  Basran J,  Talfournier F,  Sutcliffe MJ,  Scrutton NS.  Extensive conformational sampling in a ternary electron transfer complex,  Nat. Struct. Biol. ,  2003, vol.  10 (pg.  219- 225) Google Scholar CrossRef Search ADS PubMed  35 Roberts DL,  Salazar D,  Fulmer JP,  Frerman FE,  Kim J-JP.  Crystal structure of Paracoccus denitrificans electron transfer flavoprotein: structural and electrostatic analysis of a conserved flavin binding domain,  Biochemistry ,  1999, vol.  38 (pg.  1977- 1989) Google Scholar CrossRef Search ADS PubMed  36 Dwyer TM,  Zhang L,  Muller M,  Marrugo F,  Frerman F.  The functions of the flavin contact residues, αArg249 and βTyr16, in human electron transfer flavoprotein,  Biochim. Biophys. Acta ,  1999, vol.  1433 (pg.  139- 152) Google Scholar CrossRef Search ADS PubMed  37 Gustafson WG,  Feinberg BA,  McFarland JT.  Energetics of β-oxidation,  J. Biol. Chem. ,  1986, vol.  261 (pg.  7733- 7741) Google Scholar PubMed  38 Husain M,  Stankovich MT,  Fox BG.  Measurement of the oxidation-reduction potentials for one-electron and two-electron reduction of electron-transfer flavoprotein from pig liver,  Biochem. J. ,  1984, vol.  219 (pg.  1043- 1047) Google Scholar CrossRef Search ADS PubMed  39 Byron CM,  Stankovich MT,  Husain M,  Davidson VL.  Unusual redox properties of electron-transfer flavoprotein from Methylophilus methylotrophus,  Biochemistry ,  1989, vol.  28 (pg.  8582- 8587) Google Scholar CrossRef Search ADS PubMed  40 Wilson EK,  Huang L,  Sutcliffe MJ,  Mathews FS,  Hille R,  Scrutton NS.  An exposed tyrosine on the surface of trimethylamine dehydrogenase facilitates electron transfer to electron transferring flavoprotein: kinetics of transfer in wild-type and mutant complexes,  Biochemistry ,  1997, vol.  36 (pg.  41- 48) Google Scholar CrossRef Search ADS PubMed  41 Hetzel M,  Brock M,  Selmer T,  Pierik AJ,  Golding BT,  Buckel W.  Acryloyl-CoA reductase from Clostridium propionicum: an enzyme complex of propionyl-CoA dehydrogenase and electron-transferring flavoprotein,  Eur. J. Biochem. ,  2003, vol.  270 (pg.  902- 910) Google Scholar CrossRef Search ADS PubMed  42 Li F,  Hinderberger J,  Seedorf H,  Zhang J,  Buckel W,  Thauer RK.  Coupled ferredoxin and crotonyl coenzyme A (CoA) reduction with NADH catalyzed by the butyryl-CoA dehydrogenase/Etf complex from Clostridium kluyveri,  J. Bacteriol. ,  2008, vol.  190 (pg.  843- 850) Google Scholar CrossRef Search ADS PubMed  43 Buckel W,  Thauer RK.  Energy conservation via electron bifurcating ferredoxin reduction and proton/Na+ translocating ferredoxin oxidation,  Biochim. Biophys. Acta ,  2013, vol.  1827 (pg.  94- 113) Google Scholar CrossRef Search ADS PubMed  © The Authors 2013. Published by Oxford University Press on behalf of the Japanese Biochemical Society. All rights reserved
TI - Interaction between NADH and electron-transferring flavoprotein from Megasphaera elsdenii
JF - The Journal of Biochemistry
DO - 10.1093/jb/mvt026
DA - 2013-03-29
UR - https://www.deepdyve.com/lp/oxford-university-press/interaction-between-nadh-and-electron-transferring-flavoprotein-from-8H030fsDbA
SP - 565
EP - 572
VL - 153
IS - 6
DP - DeepDyve
ER -