Generation of a proton potential by succinate dehydrogenase of Bacillus subtilis functioning as a fumarate reductase


G. Unden, Institut für Mikrobiologie und Weinforschung, Johannes Gutenberg-Universität Mainz, Becherweg 15, 55099 Mainz, Germany. Fax: + 49 6131 3922695, Tel.: + 49 6131 3923550, E-mail: unden@mail.uni


The membrane fraction of Bacillus subtilis catalyzes the reduction of fumarate to succinate by NADH. The activity is inhibited by low concentrations of 2-(heptyl)-4-hydroxyquinoline-N-oxide (HOQNO), an inhibitor of succinate: quinone reductase. In sdh or aro mutant strains, which lack succinate dehydrogenase or menaquinone, respectively, the activity of fumarate reduction by NADH was missing. In resting cells fumarate reduction required glycerol or glucose as the electron donor, which presumably supply NADH for fumarate reduction. Thus in the bacteria, fumarate reduction by NADH is catalyzed by an electron transport chain consisting of NADH dehydrogenase (NADH:menaquinone reductase), menaquinone, and succinate dehydrogenase operating in the reverse direction (menaquinol:fumarate reductase). Poor anaerobic growth of Bsubtilis was observed when fumarate was present. The fumarate reduction catalyzed by the bacteria in the presence of glycerol or glucose was not inhibited by the protonophore carbonyl cyanide m-chlorophenyl hydrazone (CCCP) or by membrane disruption, in contrast to succinate oxidation by O2. Fumarate reduction caused the uptake by the bacteria of the tetraphenyphosphonium cation (TPP+) which was released after fumarate had been consumed. TPP+ uptake was prevented by the presence of CCCP or HOQNO, but not by N,N′-dicyclohexylcarbodiimide, an inhibitor of ATP synthase. From the TPP+ uptake the electrochemical potential generated by fumarate reduction was calculated (Δψ = −132 mV) which was comparable to that generated by glucose oxidation with O2 (Δψ = −120 mV). The Δψ generated by fumarate reduction is suggested to stem from menaquinol:fumarate reductase functioning in a redox half-loop.


carbonyl cyanide m-chlorophenyl hydrazone










tetraphenylphosphonium cation.

The succinate dehydrogenases of aerobic bacteria catalyze the oxidation of succinate by respiratory quinones (‘succinate:quinone reductase’), and the quinols are reoxidized by O2 (‘succinate oxidase’). Depending on the quinone type present in the bacterial membrane, succinate:ubiquinone or succinate:menaquinone reductases are found. Succinate:ubiquinone and succinate:menaquinone reductases share many properties, but show characteristic differences in the quinone reactive subunit(s) (reviewed in [1,2]).

In Gram-positive bacteria like Bacillus subtilis and other prokaryotes containing succinate:menaquinone reductases, the activities of succinate oxidase and of succinate:menaquinone reductase are lost when the proton potential across the membrane is degraded by disruption of the bacteria or by the addition of a protonophore [3,4]. Such inhibition is not observed with bacteria containing succinate:ubiquinone reductases. This difference has been explained by the fact that succinate (E0′ succinate/fumarate = +30 mV) oxidation by ubiquinone (E0′ ubiquinol/ubiquinone =+110 mV) is exergonic, whereas succinate oxidation by menaquinone (E0′ menaquinol/menaquinone = −80 mV) is endergonic. It has been suggested that the endergonic reduction of menaquinone by succinate is driven by the electrochemical proton potential (Δp) across the membrane in a process of reverse electron transport [4]. The protons consumed in the reduction of menaquinone were envisaged to be taken up from the bacterial outside (Eqn 1), while the protons liberated in the oxidation of succinate are released on the cytoplasmic side (Eqn 2).


The two reactions are linked by the transfer of two electrons.

If the endergonic succinate oxidation by menaquinone is driven at the expense of Δp, the reverse reaction (fumarate reduction by menaquinol) catalyzed by the same enzyme should generate a Δp across the membrane. This hypothesis will be tested in this communication. To this end the question was addressed whether the succinate dehydrogenase in the membrane of Bsubtilis can operate as a fumarate reductase. Furthermore it was tested whether electrons can be transported from NADH to fumarate via NADH:menaquinone reductase and menaquinone, and whether this process is coupled to the generation of a Δp.

Experimental procedures

Bacteria and growth

Bsubtilis strains W23 (wild-type) (DSMZ, Braunschweig, no. 1087), ICD2 (trpC2 met ade ΔsdhCAB::ble) [5], and 1A654 (aroD120 bgl-33 trpC2) [6] were grown in modified White medium [7,8] with glucose, glycerol or succinate (10 mm each). Aerobic growth was performed on a shaker at 200 r.p.m. in 2-L flasks with baffles filled with 200 mL medium. The aroD120 mutant 1A654 was grown with glucose and tested for lack of revertants by confirming lack of aerobic growth on succinate. Absence of menaquinone was tested as described [3].

Cell suspensions and membrane fraction

Bacteria in the late exponential growth phase at A578 of three were sedimented by centrifugation, washed twice with buffer A (50 mm potassium phosphate, pH 7.4), and suspended in the buffer at 0 °C (0.4 g wet cells·mL−1). The cell suspension was used for measurement of tetraphenylphosphonium cation (TPP+) accumulation, fumarate respiration, and oxygen uptake. Membranes were isolated from cell suspensions incubated with lysozyme followed by cell disruption with a French press [8]. The membrane fraction was resuspended in buffer A and stored at −80 °C, or used directly.

Determination of Δψ with the TPP+ electrode

The TPP+ concentrations in bacterial cell suspensions were measured with the TPP+ electrode [9,10]. The measurements were performed in a total volume of 6 mL anoxic buffer A with bacteria (2.8 mg dry weight·ml−1, corresponding to about 1.4 mg protein·ml−1) at 30 °C under an atmosphere of N2. Δψ was calculated from the internal (Tin) and external (Tex) TPP+ concentration using the Nernst equation. Tin was derived from Tex, the internal volume (Vin) and the amount of TPP+ that disappeared from the medium (Ts) according to Zaritzky et al. [11] using (Eqn 3) where K (5 mL·g protein−1) [11] represents the binding constant of TPP+ to the membrane. The internal volume of the bacteria was determined from


the distribution and concentrations of nonmembrane-permeable [14C]-taurine and permeable [3H]-H2O in the bacteria and in the solution by silicone oil centrifugation [11,12].

Enzyme activities

Enzyme activities were measured in anoxic cuvettes in the Zeiss S10 diode array photometer with the membrane fraction (0.05–0.2 mg protein·mL−1) in buffer A [3,4,13]. Oxidation of NADH (0.8 mm) by fumarate (10 mm) or 2,3-dimethoxybenzoquinone (QO, 0.2 mm) was measured at the wavelength pair 340–400 nm, oxidation of 2,3-dimethyl-1,4-naphthoquinol (DMNH2; 0.2 mm) by fumarate at 260–290 nm. The rate of fumarate reduction by intact bacteria in the presence of glucose or glycerol was obtained from the amounts of succinate formed. Succinate was measured by HPLC as described below. Oxygen uptake (succinate→Ο2) was measured with intact cells using a Clark-type electrode [4]. The activities (U·g protein−1) give the conversion of 1µatom O or 1 µmol of fumarate or succinate·min−1·g protein−1 at 30 °C. In Tables 1 and 2 all enzyme activities refer to cell protein for direct comparison of specific activities of cells and membranes. To this end the specific activities of the membranes (based on the protein in the membrane fraction) were divided by a factor of three, as the membranes comprise one third of the cellular proteins.

Table 1.  Electron transport activities of fumarate reduction and succinate oxidation in the membrane fraction of B. subtilis W23 (wild-type), the sdh mutant (ICD2), and the menaquinone deficient mutant (1A654, aro). The membranes were prepared from bacteria grown aerobically on succinate.
Enzyme activities
(µmol substrate·g cell protein−1·min−1)

W23 (wild-type)

ICD2 (sdh)

1A654 (aro)
Bacterial membranes
 NADH → fumarate 77≤ 5≤ 5
 NADH → QO217106181
 DMNH2 → fumarate 52≤ 4 44
 Glucose → fumarate 70≤ 5ND
 Succinate → O2280≤ 20≤ 5
Table 2.  Comparison of the effect of uncoupler (CCCP) and of cell disintegration (preparation of membrane fraction) on fumarate reduction versus succinate oxidation in Bsubtilis W23 (wild-type). Fumarate reduction was measured with glucose (bacteria) or NADH (membrane fraction) as the electron donor. The bacteria used for the cell suspension and for isolation of the membrane fraction were grown aerobically with succinate
 Fumarate reductionSuccinate → O2
 (µmol substrate·g cell protein−1·min−1)
  • a

    Glucose as electron donor.

  • b

    b NADH as electron donor.

Bacteria (cell suspension)70 a270
Bacteria + CCCP (10 µm)68 a< 10
Membrane fraction82 b< 10

Products of fumarate reduction

The substrates and products of fumarate reduction were measured by HPLC in cell suspensions of the bacteria in Mops mineral medium [13]. Bacteria grown aerobically on succinate or glucose were washed and incubated under anoxic conditions with glucose or glycerol plus fumarate at 30 °C at an A578 of five. After various time intervals samples were removed and centrifuged. The supernatants were analyzed by HPLC (Aminex HP × 87H, 300 × 7.8 mm; Bio-Rad, München) designed for organic acids, alcohols and sugars as described [13].

Other methods

Protein was determined by the method of Bradford [14] or by the Biuret method with KCN [15]. Stock solutions of N,N′-dicyclohexylcarbodiimide (DCCD; 100 mm), carbonyl cyanide m-chlorophenyl hydrazone (CCCP; 1 mm) and 2,3-dimethyl-1,4-naphthoquinol (HOQNO; 1 mm) were prepared in ethanol.


Fumarate reduction by resting cells of Bsubtilis

Bsubtilis did not show substantial growth under anaerobic conditions with glycerol or glucose. However, slight growth stimulation was observed after fumarate was added to the medium. The number of cells formed under these conditions amounted to 10 or 20% of that observed with the related Paenibacillus macerans, which is able to grow by fumarate respiration [8]. When Bsubtilis grown aerobically on succinate was incubated under anoxic conditions with fumarate and an electron donor like glycerol, succinate was formed as the major end product (0.5–0.8 mol per mol fumarate). Malate was initially formed from fumarate and then converted to succinate via fumarate. In addition, acetate and two other minor products were formed, presumably mostly from glycerol. The rate of succinate production by the resting bacteria amounted to 65 µmol·g protein−1·min−1, which is about one quarter of the rate of succinate oxidation by aerobically growing Bsubtilis. A similar rate of fumarate reduction was observed with a cell suspension containing glucose as electron donor. Under these conditions succinate, acetate and lactate were the major products. The rate of fumarate reduction was negligible (< 5 µmol·g protein−1·min−1) with other electron donors like formate or H2.

Fumarate reduction by NADH requires succinate dehydrogenase and menaquinone

The membrane fraction isolated from Bsubtilis grown with succinate or glucose catalyzed fumarate reduction by NADH at a high specific activity (232 U·g membrane protein−1). The specific activity given in Table 1 is three times lower, as it is based on total cell protein. This specific activity was similar to that of fumarate reduction by glucose in intact bacteria, and only by a factor of 3.6 lower than the succinate oxidase activity (Table 1). NADH dehydrogenase (NADH → Qo) and fumarate reductase (DMNH2 → fumarate) were present in activities compatible with the overall electron transport from NADH to fumarate in the membrane fraction, and with the rates of fumarate reduction in the bacteria. NADH-fumarate reductase was inhibited nearly completely by low amounts of HOQNO (not shown) which is an inhibitor of the succinate:quinone reductase of Bsubtilis[16]. The activities of fumarate reduction were negligible (< 5 U·g protein−1, not shown) with formate, H2, or glycerol 3-phosphate as electron donor.

The membrane fraction of a mutant lacking functional sdh genes catalyzed fumarate reduction by NADH or DMNH2 with < 7% of the wild-type activities (Table 1), whereas the NADH dehydrogenase activity was nearly 50%. The activities of fumarate reduction with glucose and of succinate respiration were negligible in the sdh mutant cells. The membrane fraction of an aro mutant which is deficient of menaquinone lacked the activity of fumarate reduction by NADH, whereas the activities of NADH dehydrogenase and of fumarate reduction by DMNH2 were close to those of the wild-type strain. The activity of succinate respiration was missing in cells of the aro mutant. These results are consistent with the view that succinate dehydrogenase and menaquinone are compounds of the electron transport chain catalyzing fumarate reduction by NADH.

Effect of a protonophore on fumarate reduction

The protonophore CCCP did not inhibit the activity of fumarate reduction catalyzed by cells of Bsubtilis in the presence of glucose (Table 2). The same amount of the protonophore abolished the succinate respiration of the cells. The specific activity (based on cell protein) of fumarate reduction with NADH by the membrane fraction was slightly higher than that of the corresponding cells with glucose. In contrast, succinate respiration was observed with cells, but was absent from the corresponding membrane fraction. The results concerning succinate oxidation are in agreement with earlier observations [3,4]. It is concluded that protonophores or cell disruption abolish succinate respiration, but do not affect fumarate reduction in Bsubtilis.

Δψ generation by fumarate reduction

In the experiment shown in Fig. 1, TTP+ was added to an anoxic cell suspension of Bsubtilis and the TTP+ concentration was recorded using a TPP+ electrode. The addition of glucose (Fig. 1A) or glycerol (not shown) had no effect on the external TPP+ concentration. Upon initiation of fumarate reduction by the addition of fumarate, the external TPP+ concentration decreased due to the uptake of TPP+ by the bacteria. After consumption of the added fumarate, TPP+ was slowly released into the medium. When the bacteria were pretreated with the protonophore CCCP which did not inhibit fumarate reduction, no TPP+ was taken up upon the addition of fumarate (Fig. 1B). Fumarate supplied to starved cells without glucose or glycerol, was converted only slowly to succinate, and TPP+ accumulation was low (not shown). Under these conditions the TPP+ accumulation could be further decreased by starving the cells for internal electron donors.

Figure 1.

Accumulation of TPP+ by cell suspensions of Bsubtilis W23 (2.8 mg dry weight·mL−1) by fumarate reduction under anoxic conditions. Bsubtilis W23 for the cell suspension was grown aerobically with succinate. The TPP+ electrode was calibrated by three additions of 0.85 µm TPP+. (A) Addition of fumarate (0.4 mm) and glucose (0.4 mm). (B) Addition of fumarate (0.4 mm) and glucose (0.4 mm) to bacteria incubated with CCCP (10 µm). (C) Addition of succinate (0.4 mm) to the cell suspension containing glucose (0.4 mm). The scale gives the TPP+ concentration in the buffer.

In an experiment similar to that of Fig. 1A, succinate was added instead of fumarate, and no TPP+ uptake by the cells was seen (Fig. 1C). This result argues against the view that the effect of fumarate addition (Fig. 1A) might be caused by dicarboxylate transport across the membrane of the bacteria. The results (Fig. 1) suggest that fumarate reduction is coupled to the generation of a Δp (negative inside) in Bsubtilis.

The electrical part of Δp (Δψ) was calculated from the maximum amount of TPP+ taken up by the cells (Fig. 1A) using the Nernst equation after correction for the amount of TPP+ which was bound to the bacterial membrane [9–11]. The value of Δψ so obtained was −132 mV with fumarate, and slightly lower with malate which is converted to fumarate (Table 3). Similar values were observed when glucose was replaced by glycerol, and for glucose oxidation in the presence of O2. Treatment of the cells with DCCD, an inhibitor of the ATP synthase of Bsubtilis[17], had only a small effect on the Δψ generated by fumarate reduction. The amount of DCCD applied in the experiment (Table 3) was sufficient to abolish aerobic growth of Bsubtilis with succinate (not shown). Therefore fumarate reduction forms a membrane potential directly, and secondary processes like H+ extrusion by ATPase due to ATP hydrolysis are not required for the potential generation. The menaquinone-deficient mutant as well as the sdh mutant did not accumulate TPP+ upon fumarate addition in experiments equivalent to that given in Fig. 1A (not shown). These results confirm the view that the Δψ is generated by fumarate reduction, and is not due to proton extrusion driven by ATP hydrolysis.

Table 3.  Membrane potentials of Bsubtilis W23 with fumarate (or other C4-dicarboxylates) or O2 as the electron acceptors.Δψ was determined with the TPP+ electrode under anoxic conditions in cell suspensions of the bacteria (substrates 0.4 mm) grown under aerobic conditions with succinate. The experiment with O2 was performed in air-saturated buffer. Δψ was calculated from experiments as in Fig. 1 after correction for nonspecifically bound TPP+[9–11].
Substrates (0.4 mm each)Δψ (mV)
Glucose≤ −10
 + fumarate−132
 + malate−120
 + succinate≤ −10
 + fumarate + CCCP (10 µm)≤ −10
 + fumarate + DCCD (100 µm)−121
Glycerol + fumarate−108
Glucose + O2−120


Composition of the fumarate reduction pathway in Bsubtilis

Resting cells of Bsubtilis are able to catalyze fumarate reduction with glucose or glycerol. The enzymatic system for the fumarate reduction was shown to be an electron transport chain comprising a NADH dehydrogenase, menaquinone and succinate dehydrogenase. Participation of menaquinone and succinate dehydrogenase could be directly demonstrated by the menaquinone-deficient and succinate dehydrogenase-deficient mutants, and there was no indication for the presence of an alternate fumarate reductase. Accordingly, the genome of Bsubtilis[18] contains only the sdhCAB operon encoding succinate dehydrogenase [19], but no further set of genes resembling the sdh/frd gene family. Therefore the menaquinol:fumarate reduction has to be performed by succinate:menaquinone reductase (SdhCAB protein). The succinate dehydrogenase of Bsubtilis has been described to operate in the reverse direction with quinols [20].

With the NADH dehydrogenase of Bsubtilis the situation is less clear-cut because this enzyme has not been characterized at the gene or protein level and no mutants are available. NADH is probably the only potential electron donor for fumarate reduction which is produced anaerobically from glucose in Bsubtilis[21]. Therefore, a membrane integrated NADH:menaquinone dehydrogenase is predicted to be the donor enzyme operating in fumarate reduction. The yilD gene of Bsubtilis is predicted to encode an enzyme which is similar to the noncoupling NADH dehydrogenase of Escherichia coli encoded by the ndh gene [18,22]. Other genes coding for respiratory NADH dehydrogenase were not found in the genome of Bsubtilis. Genes resembling nuoA-N of Ecoli which encode the subunits of the proton-pumping NADH dehydrogenase I are missing in the genome of Bsubtilis. This is consistent with the finding that the NADH dehydrogenase of Bsubtilis is not sensitive to rotenon or piericidin, which selectively inhibit proton-pumping NADH dehydrogenases. Furthermore, Bsubtilis NADH dehydrogenase reacts with deamino-NADH as well as with NADH, and this property is typical for NADH dehydrogenases of the noncoupling Ndh-type (unpublished results, Schnorfeil and Unden). Thus genetic and enzymatic data suggest that Bsubtilis can form only one NADH dehydrogenase. Therefore, this enzyme has to be a member of the electron transport chain catalyzing fumarate reduction by NADH. As this enzyme is unlikely to act as a proton pump, the Δp generated by fumarate reduction (Fig. 1 and Table 3) has to be formed in another way.

Mechanism of Δp generation

The Δp generated by fumarate reduction with NADH in Bsubtilis is proposed to be coupled to the redox reactions of menaquinone (Fig. 2). The protons consumed in menaquinone reduction at the NADH dehydrogenase are assumed to be taken up from the bacterial cytoplasm. The protons liberated by menaquinol oxidation at the succinate dehydrogenase are envisaged to be released at the bacterial outside. The H+/e ratio is predicted to be one. This mechanism is consistent with the view that NADH oxidation by menaquinone is electroneutral. Fumarate reduction by menaquinol as well as succinate oxidation by menaquinone are predicted to be electrogenic processes with an H+/e ratio of one. The reduction of fumarate should generate a Δp in the same direction (negative inside) as succinate respiration, in agreement with the experimental results (Table 3). In contrast, the oxidation of succinate by menaquinone is expected to generate a Δp with the opposite direction (negative outside). This prediction is in agreement with the finding that succinate respiration is dependent on the presence of a Δp (positive outside) (Table 3). Thus the endergonic oxidation of succinate by menaquinone is driven by the Δp generated by succinate (succinate → Ο2) or NADH respiration (NADH → Ο2).

Figure 2.

Schematic presentation of components of fumarate reduction by NADH in Bsubtilis. These comprise NADH dehydrogenase (Ndh, encoded by yilD), menaquinone (MK), and succinate dehydrogenase (SdhABC) with a distal and a proximal heme B (BD and BP in subunit SdhC). Ndh carries the active site for NADH at the cytoplasmic aspect of the membrane, and does not function as proton pump. Succinate dehydrogenase has the active site for fumarate/succinate in the cytoplasm, and for menaquinol (MKH2) in the cytoplasmic membrane close to the outside (positive) [4]. It is suggested that by this topology [23,24] a proton potential (outside positive) is generated by fumarate reduction (reverse succinate dehydrogenase reaction) due to the use of H+ in MK reduction (MK + 2e + 2H+ i → MKH2) from the cytoplasm, and the release of 2 H+ by MKH2 oxidation (MKH2 → MK + 2e + 2H+o) to the outside.

The structure of fumarate reductase (menaquinol:fumarate reductase) from Wolinella succinogenes has been determined at 2.2 Å resolution [25]. The membrane subunit of fumarate reductase has a structure very similar to that predicted for Bsubtilis succinate dehydrogenase with one proximal and one distal heme B. A cavity close to the distal heme B oriented to the periplasm has been identified which presumably represents the site for menaquinol oxidation [26]. This orientation would indicate the release of the protons from menaquinol oxidation to he periplasm and the generation of a transmembrane electrochemical potential. Experimental evidence, however, suggests an electroneutral operation of fumarate reductase [10]. Therefore it is not clear whether menaquinol oxidation by the fumarate reductase from Wsuccinogenes generates a proton potential as suggested from the structural similarities to succinate dehydrogenase from Bsubtilis.


We are grateful to Deutsche Forschungsgemeinschaft, the Naturwissen-schaftlich-Medizinische Forschungszentrum (Universität Mainz) and the Fonds der Chemischen Industrie for financial support.


  1. Enzymes: menaquinol:fumarate reductase (EC;NADH:menaquinone reductase (EC; succinate:quinone reductase (EC