CO Metabolism in the Acetogen Acetobacterium woodii

Abstract

The Wood-Ljungdahl pathway allows acetogenic bacteria to grow on a number of one-carbon substrates, such as carbon dioxide, formate, methyl groups, or even carbon monoxide. Since carbon monoxide alone or in combination with hydrogen and carbon dioxide (synthesis gas) is an increasingly important feedstock for third-generation biotechnology, we studied CO metabolism in the model acetogen Acetobacterium woodii. When cells grew on H₂-CO₂, addition of 5 to 15% CO led to higher final optical densities, indicating the utilization of CO as a cosubstrate. However, the growth rate was decreased by the presence of small amounts of CO, which correlated with an inhibition of H₂ consumption. Experiments with resting cells revealed that the degree of inhibition of H₂ consumption was a function of the CO concentration. Since the hydrogen-dependent CO₂ reductase (HDCR) of A. woodii is known to be very sensitive to CO, we speculated that cells may be more tolerant toward CO when growing on formate, the product of the HDCR reaction. Indeed, addition of up to 25% CO did not influence growth rates on formate, while the final optical densities and the production of acetate increased. Higher concentrations (75 and 100%) led to a slight inhibition of growth and to decreasing rates of formate and CO consumption. Experiments with resting cells revealed that the HDCR is a site of CO inhibition. In contrast, A. woodii was not able to grow on CO as a sole carbon and energy source, and growth on fructose-CO or methanol-CO was not observed.

INTRODUCTION

There is a still-growing demand for sustainable biotechnological processes to cope with increasing needs for food, water, and energy for humankind. First-generation biotechnological processes all use food-based crops that compete for scarce cropland, fresh water, and fertilizers (1). Second-generation processes rely on biomass residues from forestry and agriculture that are used through lignocellulose fermentation. This is considered to play a major role in meeting the increasing demand, but it is still in its infancy (2). However, the process does produce the greenhouse gas CO₂ along with nonfermentable waste products. In an effort to combine reduction of greenhouse gases and global warming with the production of biocommodities, acetogenic bacteria have come into focus. They can convert CO₂, H₂, and CO to acetyl coenzyme A (acetyl-CoA), which is a starter molecule for various chemicals such as ethanol, acetate, 2,3 butanediol, and butanol (3–6). CO₂, H₂, and CO are the main components of synthesis gas (syngas), but the amounts of CO₂, H₂, and CO differ considerably, depending on the source. By gasification, virtually any organic matter can be converted to syngas. In addition, industrial waste gases contain syngas. Fermentation of syngas to ethanol is already used on a precommercial scale (7).

Acetogenic bacteria have in common that they reduce CO₂ to acetic acid via the Wood-Ljungdahl pathway (WLP) (8). CO₂ reduction to acetyl-CoA proceeds via formate, formyl-tetrahydrofolic acid (THF), and methenyl-, methylene-, and methyl-THF (9). In a second branch, CO₂ is reduced to CO, and acetyl-CoA is then formed by the key enzyme of the pathway, acetyl-CoA synthase/CO dehydrogenase (CODH/ACS) from CO, methyl-THF, and CoA (10–14). This pathway is used not only for CO₂ fixation but also for energy conservation (15–19). Acetogenesis according to equation 1

$$4\hspace{0.17em}{\hspace{0.17em}\text{H}}_2+2\hspace{0.17em}{\hspace{0.17em}\text{CO}}_2\to {\text{CH}}_3\text{COOH}+2\hspace{0.17em}{\hspace{0.17em}\text{H}}_2\text{O}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}({\mathrm{\Delta }G}_{0^{\prime }}=-95\hspace{0.17em}\hspace{0.17em}\text{kJ}/\text{mol})$$ (1)

is coupled to ATP synthesis by a chemiosmotic process that involves a ferredoxin-dependent ion (Na⁺, H⁺) pump (Rnf, Ech) and an Na⁺,H⁺ ATP synthase (20).

Acetogenesis according to equation 1 was already discovered in 1932 (21). Forty-five years later, the growth of anaerobic phototrophic bacteria (22) and methanogenic archaea (23) using CO as the sole energy and carbon source was reported. The standard redox potential (E₀′) of −520 mV for the CO/CO₂ couple (24) is very favorable for acetogenesis, and in 1982 chemolithoautotrophic growth of the acetogens Butyribacterium methylotrophum (CO strain) and Eubacterium limosum on CO-CO₂ was reported (25, 26). CO was also discovered as an intermediate in acetogenesis from H₂-CO₂ (27). Since then, a number of acetogens, such as Blautia producta (formerly Peptostreptococcus productus), Moorella thermoautotrophica (formerly Clostridium thermoautotrophicum), Moorella thermoacetica (formerly Clostridium thermoaceticum), Clostridium ljungdahlii, and E. limosum KIST612, were reported to grow on or convert CO to acetate (28–32). Products produced from CO are different in different acetogens and include acetate, ethanol, butyrate, and 2,3-butanediol (5, 33). C. ljungdahlii and Clostridium autoethanogenum are well known for their ability to convert syngas into ethanol, and these two strains are used in industrial applications.

The establishment of acetogenic bacteria as production platforms for biocommodities from syngas has just begun (34–36). Genomic sequences are available for a few, and genetic tools will be developed further. However, although the principles of the Wood-Ljungdahl pathway used to convert H₂/CO-CO₂ were elucidated some time ago, the energetics of the pathway have been elucidated in detail for only one species, Acetobacterium woodii (17, 19). This acetogen uses a sodium ion current across its membrane for ATP synthesis (37–39). The sodium ion potential across the cytoplasmic membrane is established by the Rnf complex, which couples exergonic electron transfer from reduced ferredoxin to NAD⁺ to the export of Na⁺ from the cytoplasm, thus establishing an electrochemical sodium ion gradient across the membrane (20, 40, 41). Ferredoxin reduction with hydrogen as the reductant is endergonic, and the energy barrier is overcome by an electron-bifurcating, NAD⁺- and ferredoxin-reducing hydrogenase (42). Molecular hydrogen is used directly for the reduction of CO₂ to formate (43). The reduction of methenyl-THF to methylene-THF requires NADH as an electron donor (44). The further reduction to methyl-THF, catalyzed by the methylene-THF reductase, also requires NADH (45), whereas the CODH/ACS uses reduced ferredoxin as a reductant (14, 41, 46). In sum, the entire electron transfer pathway and its coupling to ATP synthesis are known. This makes it possible to calculate redox and ATP balances for any metabolic scenario, a prerequisite for metabolic engineering. A. woodii has great potential for production of acetic acid from H₂-CO₂ (4), but there are contradictory reports on its ability to use carbon monoxide (27, 47, 48). Therefore, we have studied its relationship to CO and its capability to grow on and convert H₂-CO₂-CO to acetate.

MATERIALS AND METHODS

Organism and cultivation.

A. woodii DSMZ 1030 was routinely cultivated on a rotary shaker at 30°C. The media were prepared using the anaerobic techniques described previously (49, 50). All growth experiments were performed in 120-ml serum flasks containing 20 ml of the medium previously described (51), except that KHCO₃ was omitted and trace element and vitamin solutions were increased to 10 ml/liter and 20 ml/liter, respectively. Fructose (20 mM), methanol (60 mM), or formate (100 mM) was added from sterile anaerobic stock solutions. Growth was monitored by measuring the optical density at 600 nm (OD₆₀₀).

For the growth experiments on H₂-CO₂-CO, the gas phase consisted of 64% H₂, 16% CO₂, and 0 to 15% CO at a pressure of 1 × 10⁵ Pa; N₂ was used as a makeup gas. For growth experiments with formate-CO, the gas atmosphere contained 0 to 100% CO at a pressure of 1 × 10⁵ Pa, using N₂ as a makeup gas.

A gas phase of N₂-CO₂-CO (60:15:25) at a pressure of 1 × 10⁵ Pa was used for the growth experiments on CO with or without a cosubstrate. When H₂ was used as an energy source, N₂ was replaced with H₂. For the controls without CO, a gas phase of N₂-CO₂ (85:15) was used.

Preparation of resting cells.

For preparation of cell suspensions, A. woodii was cultivated in 1.2-liter flasks (Müller-Krempel, Bülach, Switzerland) with either fructose or formate to an OD₆₀₀ of 1.8 or 0.4, respectively (16, 37). All the following steps were done under strictly anaerobic conditions in an anaerobic chamber (Coy, Grass Lake, MI). The cells were harvested and washed twice with imidazole buffer (50 mM imidazole, 20 mM MgSO₄, 20 mM KCl, 20 mM NaCl, 4 μM resazurin, 4 mM dithioerythritol [DTE], pH 7.0). The cells were resuspended in the same buffer to a protein concentration of 20 to 40 mg/ml and kept under an N₂-CO₂ atmosphere (80:20, vol/vol) on ice. The protein concentration of the resting cells was determined as described previously (52).

Cell suspension experiments.

All cell suspension experiments were performed in 120-ml flasks containing 10 ml of imidazole buffer at 30°C in a shaking water bath. For the experiments with H₂-CO₂-CO, the gas phase consisted of 40% H₂, 10% CO₂, and 0 to 50% CO at a pressure of 1 × 10⁵ Pa with N₂ as a makeup gas. For the experiments with formate, 100 mM formate was added to the buffer, and the gas phase consisted of 0 to 50% CO at a pressure of 1 × 10⁵ Pa with N₂ as a makeup gas. The experiments were started by adding the concentrated resting cells to a final protein concentration of 1 mg/ml.

Determination of acetate, ethanol, and formate concentrations.

Samples were withdrawn with a syringe and cells removed by centrifugation. The acetate and ethanol concentrations in the supernatant were determined using a gas chromatograph (Clarus 580 GC; PerkinElmer, Waltham, MA, USA). The samples were injected at 250°C and separated on an Elite FFAP column (30 m by 0.25 mm; PerkinElmer, Waltham, MA, USA) with helium as the carrier gas with a flow rate of 30 cm/s and a split of 1:50. The oven was kept at 60°C for 3 min, followed by a temperature gradient to 150°C at 10°C/min. The samples were analyzed with a flame ionization detector at 250°C. 1-Propanol (10 mM) was used as an internal standard.

The concentration of formate was determined using formate dehydrogenase of Candida boidinii (Sigma-Aldrich, St. Louis, MO, USA).

Determination of H₂ and CO concentrations.

The concentrations of H₂ and CO were determined using a gas chromatograph (Clarus 580 GC' PerkinElmer, Waltham, MA, USA). The samples were injected at 100°C and separated on a ShinCarbon ST 80/100 column (2 m by 0.53 mm; Restek Corporation, Bellefonte, PA, USA). Nitrogen and helium were used as carrier gases for the determination of H₂ and CO, respectively, with a head pressure of 400 kPa and a split flow of 30 ml/s. The oven was kept at 40°C, and the samples were analyzed with a thermal conductivity detector at 100°C.

Enzymatic activities.

For measurements of enzymatic activities, A. woodii was cultivated in 1.2-liter flasks (Müller-Krempel, Bülach, Switzerland) with either fructose or formate to an OD₆₀₀ of 1.8 or 0.4, respectively (16, 37). All the following steps were done under strictly anaerobic conditions in an anaerobic chamber (Coy, Grass Lake, MI). The cells were harvested and washed with 25 mM Tris-HCl (pH 7.5) containing 420 mM sucrose and 2 mM DTE. The cells were resuspended in 25 mM Tris-HCl (pH 7.5), 20 mM MgSO₄, 20% glycerol, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), and 0.1 mg/ml DNase I and disrupted with an ultrasonifier (Bandelin Electronic GmbH & Co. KG, Berlin, Germany). After removal of cell debris, the enzymatic activities in the cell extract were determined. The activity of the CODH was measured by following the CO-dependent reduction of methyl viologen as previously described (41). The activity of the hydrogen-dependent CO₂ reductase (HDCR) was measured by following the formate-dependent reduction of methyl viologen as previously described (43).

Chemicals.

All chemicals were purchased from AppliChem GmbH (Darmstadt, Germany), Merck KGaA (Darmstadt, Germany), Roche Diagnostics GmbH (Mannheim, Germany), Carl Roth GmbH & Co. KG (Karlsruhe, Germany), and Sigma-Aldrich Chemie GmbH (Steinheim, Germany). Gases were purchased from Praxair Deutschland GmbH (Düsseldorf, Germany).

RESULTS

Growth of A. woodii on H₂-CO₂-CO.

With H₂-CO₂ as the substrate, A. woodii grew at a rate of 0.112 h⁻¹ (Fig. 1A). Addition of CO led to a reduction of the initial growth rate in a concentration-dependent manner: with 5% CO in the gas phase, the growth rate was reduced by 70% (0.028 h⁻¹), and with 10 and 15% CO it was reduced by 80 and 90%, respectively. Higher CO concentrations led to further reductions of the growth rates. Growth in the presence of CO was biphasic, and the length of the first phase increased with increasing CO concentrations. The growth rate in the second phase was comparable to that of cultures growing on H₂-CO₂ without CO. The slow growth at the beginning could result from a metabolic adaptation, for example, a time period required to induce an enzyme(s) required for CO utilization. However, when these growth experiments were repeated with a preculture which was transferred five times on H₂-CO₂-CO (using 15% CO), the biphasic growth was still observed, although the initial growth rate was 3-fold higher than that of the nonadapted cultures growing on H₂-CO₂-CO.

FIG 1.

CO inhibits growth and H₂ consumption. The gas phase consisted of 64% H₂, 16% CO₂, and 5% (■), 10% (▲), or 15% (▼) CO. The control did not receive any CO (●). N₂ was used as a makeup gas at a pressure of 1 × 10⁵ Pa. All cultures where grown in phosphate-buffered medium free of bicarbonate and inoculated from the same preculture adapted to growth on H₂-CO₂. (A) Growth was measured by following the optical density at 600 nm. (B) The concentrations of the gases CO (open symbols) and H₂ (closed symbols) in the gas phase were determined via gas chromatography. The dashed lines mark the time points when the CO concentration was reduced to 3%. The data shown are representative of three independent experiments.

Since KHCO₃ was not present in the medium, the only carbon sources present were the gases added (CO₂ and CO). Under these carbon-limited conditions, the final optical density increased with increasing amounts of CO, from 0.5 (in the absence of CO) to 1.3 in the presence of 15% CO, indicating that CO was used as a carbon source. To analyze the effect of CO in more detail, we monitored the concentrations of H₂ and CO in the growing cultures. The control growing on H₂-CO₂ started to consume H₂ immediately at a rate of 130 μmol/h (Fig. 1B). Addition of CO led to a strong inhibition of H₂ consumption; the more CO was present, the slower was the initial consumption of H₂. With 5, 10, and 15% CO present in the gas phase, initial H₂ consumption was only 15, 7, and 2 μmol/h, respectively, while the CO concentration decreased in all cultures at a rate of 2 μmol/h from the beginning of growth. After CO was reduced to below 3% in the gas phase (Fig. 1B, dashed lines), H₂ consumption dramatically increased in all cultures, and the rates were comparable to that of H₂ consumption in the control. CO had been completely consumed when the cultures entered the stationary phase, which correlates well with the higher growth yields obtained in the presence of CO.

Thus, addition of CO to the gas phase led to initial inhibition of H₂ consumption and growth. The inhibition of H₂ consumption was removed when the CO concentration was reduced below 3%. The correlation between gas consumption and growth inhibition was not as explicit, since the transition from the first (slow) to the second (fast) growth phase was chronologically before the onset of fast H₂ consumption. The solubility of CO decreases with increasing cell densities (53), which will reduce the availability of CO in the medium. Thus, the inhibiting effect of CO will change with increasing cell densities, which might explain these results. To avoid this, we used resting cells of A. woodii to analyze the effect of CO on H₂ consumption under defined conditions.

Acetate formation from H₂-CO₂-CO in cell suspensions of A. woodii.

In contrast to growing cells, resting cells (grown on fructose) converted CO to acetate, demonstrating that in principle CO can serve as precursor for the carbonyl as well as the methyl group of acetate. However, the rate of 4 ± 1 nmol/mg · min was only 2% of that for acetate formation from H₂-CO₂, which had a rate of 215 ± 16 nmol/mg · min. The addition of CO to the H₂-CO₂ atmosphere led to a decrease of the acetate production rate (Fig. 2A). With 10% CO in the gas phase, the acetate production rate was only 6% of that for cells converting H₂-CO₂ to acetate in the absence of CO. When the CO concentration was further increased, the acetate production rate was in the same range as with CO alone. This indicated that at above 10% CO, acetate formation was only from CO. Indeed, H₂ consumption was inhibited with increasing CO concentrations, while CO consumption started right away, with identical rates irrespective of the CO concentration (Fig. 2B). The more CO was present, the slower was the H₂ consumption and the less H₂ was consumed in total. Thus, the presence of CO led to an inhibition of H₂ oxidation which resulted in the inhibition of acetate production from H₂-CO₂.

FIG 2.

CO inhibits acetate formation and H₂ consumption in resting cells of A. woodii. A. woodii was grown on fructose and harvested, and cells were washed and resuspended to a protein concentration of 1 mg/ml in 10 ml of buffer in 120-ml serum bottles. The cell suspension experiments were carried out at 30°C in a shaking water bath. The gas phase contained 40% H₂ and 10% CO₂; the remaining 50% was composed of different CO concentrations and N₂ as a makeup gas at a pressure of 1 × 10⁵ Pa. (A) The acetate production rate was determined in the first 80 min of the experiment with increasing amounts of CO. The data shown are the averages of duplicate measurements (means ± standard errors of the means [SEM]). (B) The concentrations of H₂ (closed symbols) and CO (open symbols) were measured with 0% (◆), 5% (■), 7.5% (●), and 10% (▲) CO. The data shown are representative of two independent experiments.

Growth of A. woodii on formate-CO.

Two H₂-oxidizing enzymes are known in A. woodii, and both enzymes are strongly inhibited by CO. A soluble electron-bifurcating hydrogenase reduces ferredoxin and NAD⁺ (42). However, inhibition of this enzyme seems unlikely to cause the inhibitory effect observed, since CO oxidation yields reduced ferredoxin that can reduce NAD⁺ via the Rnf complex (40). The second hydrogenase is part of the HDCR, which reduces CO₂ to formate using H₂ (43). The HDCR can also use reduced ferredoxin as an electron donor, but this activity is 20-fold lower than CO₂ reduction with H₂. Thus, it is expected that the presence of CO would inhibit formate production and create a bottleneck at the first step of the methyl branch of the Wood-Ljungdahl pathway. To address this hypothesis, we replaced H₂-CO₂ (the substrate of the HDCR) with formate (the product of the HDCR) and grew the cells in the presence of CO. Indeed, growth of A. woodii on formate-CO was not impaired with CO concentrations up to 25%, and the cultures grew with a growth rate of 0.125 h⁻¹, comparable to that for cultures growing on formate as the sole substrate (Fig. 3A). Up to an initial CO concentration of 50%, the final cell yield even increased, indicating that CO was beneficial for growth. With 75 and 100% CO (in combination with formate), the cultures grew slower and the final cell yield was reduced, but even with 100% CO growth was observed, and the growth rate of 0.04 h⁻¹ was still 30% of that of cultures growing on formate. When A. woodii was transferred five times on 100 mM formate with 50% CO in the gas phase, these adapted cultures were able to grow faster on higher CO concentrations. Using 75% CO in combination with formate, growth was 3-fold faster than that of nonadapted cultures.

FIG 3.

Growth of A. woodii on formate-CO. All cultures contained 100 mM formate and increasing CO concentrations (■, 0%; ▲, 10%; ▼, 25%; ◆, 50%; □, 75%; △, 100%), and the optical density at 600 nm (A), the concentration of formate (B), and the concentration of CO (C) were measured over time. N₂ was used as a makeup gas at a pressure of 1 × 10⁵ Pa. All cultures where inoculated with the same preculture adapted to growth on formate. The data shown are representative of two independent experiments.

All cultures consumed formate from the beginning on, and up to an initial CO concentration of 50% CO, the formate had been completely consumed once the cultures entered the stationary phase (Fig. 3B). At higher CO concentrations where growth was inhibited, the consumption of formate was also significantly slower, and when these cultures entered the stationary phase, only 30 mM formate had been consumed.

All cultures consumed CO simultaneously with formate (Fig. 3C). With low initial concentrations of CO (10 and 25%), the CO had been completely consumed when the cultures entered the stationary phase. With a 50% initial CO concentration, the consumption of CO immediately stopped once all the formate had been consumed and the cultures entered the stationary phase. With 75 and 100% CO, the consumption also stopped when the cultures entered the stationary phase; however, at this point formate was still available in the medium.

The formation of end products was also monitored. The control growing solely on formate produced 24.7 ± 0.9 mM acetate (Table 1). Addition of CO led to an increase of the final acetate concentration: cultures growing on formate in combination with 10, 25, and 50% CO produced 31.4 ± 0.2, 40.7 ± 0.2, and 53.8 ± 0.7 mM acetate, respectively. In addition, ethanol was produced as an end product, with a maximum of 3.8 ± 0.6 mM in the cultures growing on formate–50% CO. With higher CO concentrations (75 and 100% CO), the final acetate concentrations decreased, consistent with slower growth and formate and CO consumption.

TABLE 1.

End products present in the stationary phase during growth of A. woodii on formate-CO^a

CO (%)	Concn (mM)b
	Acetate	Ethanol
0	24.7 ± 0.9	0
10	31.4 ± 0.1	0
25	40.7 ± 0.2	0.8 ± 0.1
50	53.8 ± 0.7	3.8 ± 0.6
75	33.5 ± 1.2	1.0 ± 0
100	23.4 ± 2.6	0.4 ± 0

^aThe cells were grown with 100 mM formate and 0 to 100% CO; N₂ was used as a makeup gas. Samples were taken in the stationary phase and analyzed via gas chromatography.

^bThe data are means ± SEM from two independent experiments.

Acetate production from formate-CO in cell suspensions of A. woodii.

The above-mentioned growth experiments demonstrated a much higher tolerance toward CO of acetogenesis from formate compared to H₂-CO₂. This was also seen in resting cells. In cells grown on fructose, the acetate production from H₂-CO₂ was inhibited down to 6% when 10% CO was present (Fig. 4, black bars). In contrast, the acetate production from formate in the presence of 10% CO was inhibited down to only 54%, demonstrating that acetogenesis from formate is more tolerant toward CO inhibition (Fig. 4, white bars). However, a further increase of the CO concentration to 40% inhibited acetate formation from formate down to 10%. When cells grown on formate were used, the inhibition by CO was even smaller; 42% residual activity remained at 40% CO (Fig. 4, gray bars). Thus, the formate-grown cells were more tolerant against CO than fructose-grown cells. Since formate and CO are the only available electron donors, the rate of acetate formation is dependent on formate and/or CO oxidation by the HDCR and the CODH. Therefore, we compared the activities of the corresponding enzymes in fructose-grown and formate-grown cells. When the cells were grown with fructose as the substrate, the HDCR and the CODH had activities of 2.4 ± 0.2 and 9.1 ± 2.4 U/mg, measured as the formate- and CO-dependent reduction of methyl viologen, respectively. These activities where two times higher in cell extracts from formate-grown cells (HDCR, 5.1 ± 0.2 U/mg; CODH, 20.5 ± 5 U/mg), which would accelerate acetate production and support growth with formate-CO.

FIG 4.

CO inhibition of acetate production from H₂-CO₂ or formate in resting cells of A. woodii. A. woodii was grown on fructose (black and white bars) or formate (gray bars) and harvested. The cells were washed and resuspended to a protein concentration of 1 mg/ml in 10 ml of buffer in 120-ml serum bottles. The cell suspension experiments were carried out at 30°C in a shaking water bath. The acetate production from H₂-CO₂ (black bars, fructose-grown cells) and from 100 mM formate (white bars, fructose-grown cells; gray bars, formate-grown cells) was determined in the presence of increasing CO concentrations. The acetate production rate in the absence of CO (fructose-grown cells, 215 ± 16 nmol/min · mg with H₂-CO₂ and 211 ± 18 nmol/min · mg with formate; formate-grown cells, 252 ± 6 nmol/min · mg with formate) was set to 100% for the calculation of relative acetate production rates. The gas phase for the experiments with hydrogen contained 40% H₂, 10% CO₂, and 0 to 40% CO. For the experiments with formate, 100 mM formate was added in combination with 0 to 40% CO. N₂ was used as a makeup gas at a pressure of 1 × 10⁵ Pa. All data points are means ± SEM from two independent experiments.

A. woodii does not grow on fructose-CO, methanol-CO, or CO alone.

As demonstrated here, A. woodii can cometabolize CO when H₂-CO₂ or formate is present, but growth on H₂-CO₂ was strongly inhibited, while formate-dependent growth was not impaired (Fig. 5A). In contrast, growth on methanol or fructose was completely inhibited when 25% CO was present in the gas phase (Fig. 5B).

FIG 5.

Growth of A. woodii on CO with a cosubstrate. (A) A preculture adapted to growth on formate was used to inoculate cultures with 100 mM formate (●), 100 mM formate–25% CO (▲), or 25% CO alone (◆) as the substrate. A preculture adapted to growth on H₂-CO₂ was used to inoculate cultures with H₂-CO₂ (○), H₂–CO₂–25% CO (△), or 25% CO alone (♢) as the substrate. (B) A preculture adapted to growth on fructose was used to inoculate cultures with 20 mM fructose (○), 20 mM fructose–25% CO (△), or 25% CO alone (♢) as the substrate. A preculture adapted to growth on methanol was used to inoculate cultures with 60 mM methanol (●), 60 mM methanol–25% CO (▲), or 25% CO alone (◆) as the substrate. The medium used was phosphate buffered and free of bicarbonate. For all experiments, a gas phase of N₂-CO₂-CO (60:15:25) was used with a pressure of 1 × 10⁵ Pa. When hydrogen was the substrate, N₂ was replaced by H₂. For the controls without CO, a gas phase of N₂-CO₂ (85:15) was used. The data shown are representative of three independent experiments.

Since A. woodii has been reported to grow on CO as a sole energy source (47), cells were adapted to growth on fructose, methanol, H₂–CO₂, or formate. Upon transfer to fresh medium with CO as the sole substrate (N₂-CO₂-CO, 60:15:25), growth was not observed in any case (monitoring was for 1,100 h) (Fig. 5). A reduction of the CO concentration or addition of bicarbonate to the medium also did not lead to growth of A. woodii. Cultures transferred five times on H₂-CO₂-CO or formate-CO were also not able to grow on CO as the sole carbon and energy source. In cultures growing on formate-CO, the CO consumption immediately stopped once the formate had been used up. These results demonstrate that in A. woodii, CO is metabolized only as long as formate or H₂-CO₂ is available as a cosubstrate.

DISCUSSION

The energy metabolism of A. woodii and the biochemical basis for growth on a wide variety of different substrates are understood in much detail (19, 54–59). When it comes to CO as the substrate, however, the reports in the literature are scattered and often contradictory. A. woodii was shown to be capable of incorporating CO into acetate as methyl as well as carboxyl groups, using ¹³C nuclear magnetic resonance (NMR) (60). Three years later, it was shown that cell suspensions of A. woodii could use CO for the synthesis of ATP and for energy-dependent histidine uptake (27), but the cells did not produce acetate. In the following year, Genthner and Bryant reported that A. woodii was able to grow with CO as the sole energy source without adaptation by producing acetate as an end product (47), but Ma et al. were unable to reproduce growth of A. woodii with CO as the sole energy source (48).

We also did not observe growth of A. woodii on CO as the sole carbon and energy source; it was observed only on CO in combination with H₂-CO₂ or formate as a cosubstrate. How can the contradictory results be explained? The medium used by Genthner and Bryant (47) contained 5% of rumen fluid, which contains large amounts of volatile fatty acids such as acetate, propionate, and butyrate (61). Thus, it may be possible that formate also was present and served as a cosubstrate in analogy to our experiments. Our observation that resting cells of A. woodii were able to synthesize acetate from CO alone is in agreement with the results from Kerby et al. (60) but in contrast to those from Diekert et al. (27). However, the latter data were published 3 years before the discovery that A. woodii is strictly dependent on the presence of Na⁺ (37); thus, the missing Na⁺ in the buffer might have led to the fact that acetate was not produced.

When A. woodii is growing with formate as the sole energy and carbon source, 3 formate have to be oxidized (equation 2a) to reduce another formate and 1 CO₂ to acetate (equation 2b); thus, 4 formate are required for the synthesis of 1 acetate (equation 2):

$$3\hspace{0.17em}\text{\hspace{0.17em}HCOOH}\to 3\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2+6\hspace{0.17em}\hspace{0.17em}{e}^{-}+6\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}$$ (2a)

$$1\mathrm{HCOOH}+1{\text{CO}}_2+6\hspace{0.17em}\hspace{0.17em}{e}^{-}+6\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}\to 1\hspace{0.17em}{\text{CH}}_3\text{COOH}+2\hspace{0.17em}\hspace{0.17em}{\text{H}}_2\text{O}$$ (2b)

$$4\hspace{0.17em}\text{\hspace{0.17em}HCOOH}\to 1\hspace{0.17em}\hspace{0.17em}{\text{CH}}_3\text{COOH}+2\hspace{0.17em}\hspace{0.17em}{\text{H}}_2\text{O}+2\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2$$ (2)

When in addition to formate, CO is present as a cosubstrate, it can be fed into the WLP in two different ways. First, it can be bound to the CODH/ACS and further be incorporated as a carbonyl group into acetyl-CoA, serving only as a carbon source. Under these conditions, only 2 formate would have to be oxidized (equation 3a) to reduce 1 formate with 1 CO to acetate (equation 3b); thus, 3 formate and 1 CO would be required for the synthesis of 1 acetate (equation 3):

$$2\hspace{0.17em}\hspace{0.17em}\text{HCOOH}\to 2\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2+4\hspace{0.17em}\hspace{0.17em}{e}^{-}+4\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}$$ (3a)

$$1\hspace{0.17em}\hspace{0.17em}\text{HCOOH}+1\mathrm{CO}+4\hspace{0.17em}\hspace{0.17em}{e}^{-}+4\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}\to 1{\text{CH}}_3\text{COOH}+1\hspace{0.17em}\hspace{0.17em}{\text{H}}_2\text{O}$$ (3b)

$$3\mathrm{HCOOH}+1\hspace{0.17em}\hspace{0.17em}\text{CO}\to 1\hspace{0.17em}\hspace{0.17em}{\text{CH}}_3\text{COOH}+1\hspace{0.17em}\hspace{0.17em}{\text{H}}_2\text{O}+2\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2$$ (3)

Second, CO can also be oxidized and therefore be used as a carbon and electron source. If CO replaced formate as an electron donor, oxidation of 2 CO would be required (equation 4a) for the reduction of 1 formate with 1 CO to 1 acetate (equation 4b). Therefore, 1 formate and 3 CO would be required for the synthesis of 1 acetate (equation 4).

$$2\hspace{0.17em}\hspace{0.17em}\text{CO}+2{\text{H}}_2\text{O}\to 2\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2+4\hspace{0.17em}\hspace{0.17em}{e}^{-}+4\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}$$ (4a)

$$1\hspace{0.17em}\text{\hspace{0.17em}HCOOH}+1\hspace{0.17em}\hspace{0.17em}\text{CO}+4e^{-}+4\hspace{0.17em}\hspace{0.17em}{\text{H}}^{+}\to 1{\text{CH}}_3\text{COOH}+1{\text{H}}_2\text{O}$$ (4b)

$$1\mathrm{HCOOH}+3\hspace{0.17em}\hspace{0.17em}\text{CO}+{\text{H}}_2\text{O}\to 1\hspace{0.17em}\hspace{0.17em}{\text{CH}}_3\text{COOH}+2\hspace{0.17em}\hspace{0.17em}{\text{CO}}_2$$ (4)

All cultures received 100 mM formate and different concentrations of CO. The control received only formate, and at the end of growth, formate had been completely consumed and 24.7 ± 0.9 mM acetate had been produced, which is in agreement with equation 2. In cultures which grew on formate in the presence of 50% CO, at the end of growth, formate (100 mM) had been completely consumed, and in addition, 184 ± 8 mM CO had been consumed. Thus, the ratio of formate to CO (1:1.8) is close to the ratio in equation 4, demonstrating that CO was oxidized. Also, 53.8 ± 0.7 mM acetate and 3.8 ± 0.6 mM ethanol where found, which is much more than would be possible using only formate as an electron donor (as in equation 3). CO clearly had to be oxidized to provide the reducing equivalents to synthesize the large amounts of acetate and the even more reduced end product ethanol. Thus, in combination with formate, CO is used as carbon and energy source by A. woodii and therefore serves as a cosubstrate.

The cells growing on H₂-CO₂-CO started to consume CO and H₂ very slowly. When the CO concentration was below 3%, the H₂ consumption also increased. How can these results be explained? From the CO inhibition measurements of the purified enzymes, it can be estimated that the bifurcating hydrogenase is two times more sensitive to CO than the HDCR (42, 43). Thus, at a certain CO concentration, the HDCR should be able to produce formate (from H₂-CO₂), while the hydrogenase is still inhibited. Having CO-derived reduced ferredoxin which yields NADH as an electron donor for the further reduction of formate via methenyl- and methylene-THF to methyl-THF, acetate formation and growth should no longer be inhibited, which explains the increase of growth and of CO oxidation. When the concentration of CO is further reduced, the soluble hydrogenase is activated and starts to oxidize H₂, which explains the increase of H₂ consumption.

Synthesis gas contains CO and H₂ in substantial amounts; both gases are “energy-rich” substrates for acetogenic bacteria. In A. woodii the presence of CO inhibits the ability to use H₂ as the substrate, which is most likely due to the inhibition of the CO-sensitive hydrogenase and the HDCR (42, 43). Since most hydrogenases are inhibited by CO, the efficient simultaneous utilization of CO and H₂ in a syngas fermenting process is possible only if the organism has a CO-insensitive hydrogenase. This is known mainly in carboxydotrophic bacteria such as Carboxydothermus hydrogenoformans, which use the water gas shift reaction to oxidize CO and reduce H⁺ to hydrogen (62). E. limosum and C. ljungdahlii are both acetogenic bacteria capable of growing with CO as the sole energy source, with doubling times of 7 and 3.8 h, respectively (31, 47). For both organisms it was shown in batch cultures that when growth was on H₂-CO₂, H₂ consumption did not start before the CO was nearly removed (26, 63). For C. ljungdahlii it was demonstrated in a continuously stirred tank reactor that in the first 6 days of the experiment, CO was the only substrate oxidized. H₂ consumption did not start before cell densities which used 90% of the CO in the inlet gas stream were reached (64). Also, oxidation of CO is always thermodynamically more favorable than oxidation of H₂ (65). Thus, the simultaneous bacterial oxidation of CO and H₂ in an unlimited system is questionable.

When A. woodii was grown with formate-CO, growth was not impaired up to a CO concentration of 25%, but with higher concentrations, an inhibitory effect was observed. CO can inhibit metalloenzymes and electron transport chains in bacteria (66), and the inhibitory effects on the growth of acetogenic bacteria have also been reported in the literature. E. limosum is capable of growing with CO as the sole energy source, with a doubling time of 7 h. However, with CO concentrations above 50% an inhibitory effect was observed, and the doubling time increased 2.6-fold and the final optical density decreased by 75% (26). When C. ljungdahlii was cultivated with CO, CO₂, H₂, and Ar (30%, 30%, 30, and 10%, respectively) in batch experiments, pressures above 1.0 atm led to inhibitory effects, which are most likely to be caused by CO (67).