Introduction

The theoretical maximal yield of H2 from anaerobe microbial conversion of glucose is 4 mol of H2 mol−1 glucose, when glucose is metabolised into acetic acid and CO2. In microbial cultures where other fermentation products or biomass are also produced from glucose, the yield of H2 will be lower. In most cases, H2 yields from axenic as well as mixed microbial cultures have been in the range of 1–2.5 mol H2 mol−1 glucose (Rachman et al. 1998; Mizuno et al. 2000; Fabiano and Perego 2002; Oh et al. 2003; Lin and Chang 2004; Morimoto et al. 2004; Kotay and Das 2007). However, Thermotogales, a group of anaerobic thermophilic bacteria, can metabolise carbohydrates with H2 yields that may approach 4 mol H2 mol−1 glucose. For Thermotoga elfii a yield of 3.8 mol H2 mol−1 glucose (van Niel et al. 2002), and for T. petrophila and T. naphtophila, yields of 3.7 and 4.0 mol H2 mol−1 glucose, respectively (Takahata et al. 2001) have been reported. T. maritima converted 1 mol of glucose into 2 mol of acetic acid and 4 mol of H2 (Schröder et al. 1994) but since also lactic acid was formed, the overall yield of H2 produced per glucose consumed would have been lower.

Yields exceeding 4 mol of H2 produced per mol glucose consumed have been reported for Thermotoga neapolitana, but only when this bacterium was grown under microaerobic conditions (Van Ooteghem et al. 2002). These surprisingly high yields of H2 have led to the hypothesis that T. neapolitana can meet its energy requirements by oxidising a small fraction of glucose using molecular oxygen as terminal electron acceptor, while the remaining glucose is metabolised through pathways that result in an increased H2 yield (Van Ooteghem et al. 2002, 2004). The nature of these pathways has not been documented in the literature.

Growth and H2 production at low O2 partial pressures have also been observed in other members of the Thermotogales (Van Ooteghem et al. 2004; Van Ooteghem 2005; Yang and Ma 2005), although H2 production is generally believed to occur only under anaerobic conditions since hydrogenases are usually inhibited by O2. An O2 insensitive hydrogenase has been described in T. neapolitana (Käslin et al. 1998), which may explain why microaerobic H2 production could take place in this bacterium. Since high H2 yields are important for the development of microbial derived H2 as a sustainable source of ‘green’ energy, we have tested the hypothesis that microaerobic metabolism increase the yield of H2 from T. neapolitana (Van Ooteghem et al. 2002, 2004) by investigating the conversion of glucose and generation of H2 in anaerobic and microaerobic cultures of this bacterium.

Materials and methods

Strain and growth media

Thermotoga neapolitana DSM 4359 (ATCC 49049) was obtained from the German Collection of Microorganisms and Cell Cultures. The cells were grown in a modified version of the ATCC 1977 medium (see van Ooteghem 2002) containing 1.0 g cysteine · HCl l−1as reducing agent. Glucose was autoclaved separately, and the concentration of potassium phosphate was increased from 0.6 to 6.3 g l−1 to increase the buffering capacity of the growth medium. A few crystals of resazurin were added as redox indicator. In some cultures, 1.5 g l−1 (10 mM) malonic acid was added, or cysteine was omitted, as explained in the “Results and discussion” section.

Growth conditions

Cultures were grown in 160 ml serum bottles containing 50 ml growth medium. The bottles were placed in a shaking incubator at 0 or 200 rpm and 80°C. Before the cultures were inoculated, the bottles were sealed by buthyl rubber stoppers (Apodan, Denmark) and flushed with N2. In some cultures, air was added to the headspace to give 0.1–6% O2.

Sampling and analyses

Gas and liquid samples from the headspaces of the serum bottles were taken using syringes fitted with needles. Excess gas pressure in a serum bottle was released into a syringe and the amount of gas, which had been present in the headspace was estimated at ambient air pressure as the total volume of gas in the flask and in the syringe. H2, the headspace gas, was measured by GC: 0.3 ml headspace gas equilibrated to ambient temperature was injected at 110°C, and separated on a 2 m Hyacep Q column at 70°C. H2 was quantified by a thermal conductivity detector against known standards. N2 was used as carrier at a 14 ml min−1.

Growth was monitored by the OD700, at which wavelength no medium components absorbed, and by microscopic examination. Glucose and organic acid concentrations in culture supernatants were quantified using HPLC: 50 μl of 0.22 μm filtered culture supernatant was separated on an Aminex HPX−87H column (Bio-rad), eluted with 0.4 ml min−1 of 0.5 mM H2SO4 at 30°C. Detection was performed by a Knauer K-2300 refractive index detector.

Results and discussion

Thermotoga neapolitana DSM 4359 was grown in non-shaken cultures at conditions resembling former studies of H2 production in this bacterium (Van Ooteghem et al. 2002, 2004) and in cultures shaken at 200 rpm to prevent gradient formation in the liquid phase. When T. neapolitana was grown anaerobically on glucose, a mixture of acetic acid and lactic acid was produced. H2 accumulated in the headspace, and pH decreased from pH 8 to pH 5.5 (Fig. 1A and B). Accumulation of cells increased OD700 from ca. 0.1 to 1–1.5 within 40–60 h, but since cell morphology changed during the cultivation (Belkin et al. 1986), there was no direct relationship between OD700 and cell density.

Fig. 1
figure 1

Amounts of glucose (□), acetic acid (•), and lactic acid (◯), and pH (◊) in liquid phase and H2 (■) in gas phase in batch cultures of Thermotoga neapolitana DSM 4359 grown in medium containing 1.0 g L−1 cysteine HCl. (A) and (B) Cultures grown anaerobically, (C) and (D) Cultures initially grown anaerobically but 6% O2 added to the headspace at the time indicated by arrow. (A) and (C) Cultures shaken at 200 rpm. (B) and (D) Cultures not shaken

Full size image

There were no significant differences in the final yields of H2 released in shaken cultures (2.4 ± 0.3 mol H2 mol−1 glucose, n = 12) and non-shaken cultures (2.4 ± 0.4 mol H2 mol−1 glucose, n = 4), but the production of H2 occurred at a higher rate when the cultures were shaken. The H2 transfer rate from the liquid phase to the gas phase would have been lower in non-shaken cultures than in shaken cultures, and the non-shaken cultures would therefore have experienced relatively higher dissolved H2 pressures compared to the shaken cultures. H2 represses growth in T. maritima (Schröder et al. 1994) and the lower rates of glucose consumption and H2 production in the non-shaken cultures (Fig. 1B and D) compared to the shaken cultures (Fig. 1A and C) were probably the result of increased H2 repression in the former cultures.

The yield of H2 per amount of acetic acid formed was 1.8–2.1 mol mol−1, as would also be expected from stoichiometrical considerations when alternative electron acceptors are not present, but the ratio between formation of acetic acid and lactic acid was variable. Acetic acid, which is generated along with H2 by oxidation of pyruvate, started to accumulate prior to lactic acid, the product of pyruvate reduction. Probably, the dissolved H2 partial pressure was initially too low to repress the oxidation of pyruvate. After glucose was depleted, the headspace gas consisted of 30–40% H2. Since the pressure in the flasks increased above ambient pressure between samplings, the H2 partial pressure in the headspace gas increased occasionally to values above 0.5 atm. The dissolved H2 partial pressure in the liquid phase would have reached at least similar magnitudes, making lactic acid synthesis gradually more favourable. Because lactic acid was produced, the overall yield of H2 did not reach the theoretical maximal value. The yield of H2 actually obtained in these cultures would have been influenced by the ratio between gas phase to liquid phase volumes, since this ratio had a large influence on the H2 partial pressure. In cultures of Thermotoga maritima, an increased ratio between gas and liquid phase volumes resulted in decreased synthesis of lactic acid (Schröder et al. 1994). Furthermore, pH decreased in the culture supernatants, resulting in increasing pH stress. In some cultures, the process stopped before all glucose was consumed. This has previously been observed also in cultures of T. neapolitana (Van Ooteghem 2002) and T. maritima (Schröder et al. 1994).

Addition of 6% O2 to the headspace of growing cultures, which at 80°C corresponded to an oxygen partial pressure of 7.2 kPa, temporarily inhibited H2 production in Thermotoga neapolitana (Fig. 1C and D). H2 production resumed after approximately 4 h, at rates similar to those observed in anaerobic cultures. In contrast to previous reports (Van Ooteghem et al. 2002, 2004, Van Ooteghem 2005), we did not observe an increased yield of H2 after addition of O2. Because O2 is removed in cultures of T. neapolitana, it has been speculated that this bacterium has a low capacity for aerobic metabolism which somehow increases the yield of H2 (Van Ooteghem et al. 2002, 2004, Van Ooteghem 2005). If this is the case, the H2-producing pathway does not necessarily have to generate ATP if sufficient ATP is formed by aerobic metabolism. It is not clear which metabolic pathways would be responsible for the increased yield of H2, but the oxidative pentose phosphate pathway could theoretically be a mediator. Using a mixture of enzymes from the oxidative pentose phosphate cycle and an NADPH reducing hydrogenase, Woodward et al. (2000) measured the release of 11.6 mol H2 per mol of glucose-6-phosphate consumed. Coding sequences for the enzymes of the pentose phosphate cycle and a NADPH reducing hydrogenase have been identified in the genome of T. maritima (Nelson et al. 1999) and are probably also present in T. neapolitana.

However, aerobic metabolism and ATP formation are not supported in Thermotoga maritima, based on the complete genomic sequence (Nelson et al. 1999) which does not contain evidence for a complete citric acid cycle or the presence of cytochromes. We therefore measured the fate of O2 in sterile growth medium under conditions similar to the growth experiments (Fig. 2). In sterile serum bottles shaken at 200 rpm, all O2 disappeared from the gas phase in less than 25 h, while approximately 50% of the initial amount of O2 disappeared in non-shaken serum bottles. In the growth medium, O2 was reduced by cysteine, but sterile, cysteine free medium was also able to remove O2 from the gas phase although at a lower rate. The O2 concentration was also followed in the gas phase of sterile serum bottles, in which the growth medium was substituted by a solution of 0.5 M Na2SO3 and 5 μM CoS04. This solution reduces O2 instantaneously, and the rate of O2 removal will therefore be an estimate of the transfer capacity of O2 from the gas to the liquid phase. Since O2 was removed at a lower rate by the growth medium than by the Na2SO3 solution, O2 removal would have been reaction rate limited in the growth medium, and the dissolved O2 partial pressure would have been greater than zero in cultures where O2 was added to the headspace.

Fig. 2
figure 2

Oxygen content (quantified by gas chromatography) in headspace of sterile serum bottles containing different media: growth medium with cysteine (◯, •), growth medium without cysteine (□, ■), water (◊, ◆), or 0.5 M Na2SO3 and 5 μM of CoS04 (▽, ▼). Non-shaken bottles shown as open symbols, bottles shaken at 200 rpm shown as closed symbols. All data points are averages of measurements from 2 replicate bottles

Full size image

Addition of O2 to growing cultures, followed by manual shaking caused the liquid phases to become pink as the redox indicator resazurin was oxidised. After less than 2 h incubation, only the upper half of the liquid phase in non-shaken cultures with 6% O2 in the headspace remained pink. In the lower parts of these cultures, where most of the biomass had settled, anaerobic conditions had been re-established. It is therefore possible that the production of H2 in our non-shaken cultures (Fig. 1D) as well as in previous non-shaken Thermotoga neapolitana cultures (Van Ooteghem et al. 2002, 2004) has been sustained primarily in anaerobic zones. In the cultures shaken at 200 rpm, stratification of the liquid phase was not observed. However, the increased rates of O2 mass transfer and O2 reduction in the shaken cultures resulted in most of the added O2 being consumed within 4 h (Fig. 2). These cultures would therefore have been exposed to rapidly decreasing O2 partial pressures. Development of anaerobic zones and reduction of O2 by medium components may explain why the inhibitory effects of O2 were only temporary in both non-shaken and shaken cultures (Fig. 1C and D).

To obtain more representative data on the effect of dissolved O2 on the production of H2 in Thermotoga neapolitana, H2 production was also measured in cultures grown in cysteine free medium (Fig. 3). The rates of glucose consumption and H2 production and the yield of H2 (1.1–1.3 mol H2 mol−1 glucose) were considerably lowered when cysteine was omitted from the growth medium, although cysteine residues were present in the peptone used in the medium. In cultures shaken at 200 rpm, the addition of 6% O2 (7.2 kPa) to the headspace completely inhibited H2 production and the rate of glucose consumption decreased from 12 to 4 μmol h−1 (Fig. 3C). In a series of cultures grown in cysteine free medium, shaken at 200 rpm, and 0.1–4% O2 added to the headspace H2 production continued when the initial O2 contents did not exceed 1% (1.2 kPa, data not shown). In non-shaken, cysteine-free cultures glucose consumption and H2 production were not measurably affected by addition of up to 6% O2 to the headspace (Fig. 3D). Anaerobic zones in which resazurin remained reduced did not develop in these cultures. Since non-shaken and shaken cultures grown without cysteine produced H2 with 6% and 1% O2 in the gas phase, respectively, and shaken cultures grown with cysteine produced H2 with 6% O2 in the gas phase, our results confirm that T. neapolitana is able to tolerate exposure to low dissolved O2 partial pressures (Van Ooteghem et al. 2002, 2004). The related bacteria, T. hypogea and T. maritima contain NADH oxidases which may work in O2 detoxification (Yang and Ma 2005, 2007). Similar catalytic capacities may also explain why T. neapolitana can tolerate low O2 partial pressures, but do not explain the increased yield of H2 in microaerobic conditions observed by Van Ooteghem et al. (2002, 2004) and Van Ooteghem (2005). The reduction of O2 requires reducing equivalents and these would therefore not be available for the generation of H2.

Fig. 3
figure 3

Amounts of glucose (□), acetic acid (•), and lactic acid (◯), and pH (◊) in liquid phase and H2 (■) in gas phase in batch cultures of Thermotoga neapolitana DSM 4359 grown in medium without cysteine. (A) and (B) Cultures grown anaerobically, (C) and (D) Cultures initially grown anaerobically but 6% O2 added to the headspace at the time indicated by arrow. (A) and (C) Cultures shaken at 200 rpm. (B) and (D) Cultures not shaken

Full size image

We also grew Thermotoga neapolitana in the presence of malonic acid (Fig. 4). It has been suggested that malonic acid inhibits succinic dehydrogenase and thus O2 dependent metabolism in T. neapolitana, that malonic acid itself can be metabolised by T. neapolitana, and that the O2 dependent metabolism which resulted in increased yield of H2 re-established in T. neapolitana cultures after malonic acid was depleted (Van Ooteghem et al. 2002). However, a coding sequence for succinic dehydrogenase has not been identified in the T. maritima genome (Nelson et al. 1999), and malonic acid may cause unknown side effects resulting in the increased H2 yields observed by Van Ooteghem et al. (2002) in T. neapolitana grown with malonic acid. In anaerobic T. neapolitana cultures (Fig. 4), as well as in microaerobic cultures (data not shown), malonic acid was not metabolised, even though we used a lower malonic acid concentration (10 mM or 1.5 g l−1) than in the previous study (2.08 g l−1, Van Ooteghem et al. 2002). However, we did observe an increased yield of up to 3.5–3.8 mol H2 mol−1 glucose in cultures grown with malonic acid (Fig. 4). Lactic acid production was reduced by 50–100% compared to cultures grown without malonic acid. It therefore appears that malonic acid affected the metabolism in T. neapolitana. The production of lactic acid was decreased, and the yield of H2 increased.

Fig. 4
figure 4

Amounts of glucose (□), acetic acid (•), and lactic acid (◯), concentration of malonic acid (▽), and pH (◊) in liquid phase and H2 (■) in gas phase in batch cultures of Thermotoga neapolitana DSM 4359 grown in medium added 10 mM malonic acid. Data points indicate averages ± standard deviations from 2 replicate cultures shaken at 200 rpm

Full size image

The results presented in this paper confirm that Thermotoga neapolitana DSM 4359 can metabolise glucose anaerobically with a high yield of H2 close to the theoretical maximum of 4 mol H2 mol−1 glucose. We did not find evidence for an increased yield of H2 in microaerobic cultures of T. neapolitana, although we used the same bacterial strain and a similar experimental set-up in which microaerobic H2 production with a very high H2 yield has previously been described (Van Ooteghem et al. 2002, 2004, Van Ooteghem 2005). Our results actually showed that T. neapolitana is sensitive to dissolved O2, but complete inhibition of growth and H2 synthesis was only seen when cysteine was removed from the growth medium. It therefore seems that T. neapolitana metabolises glucose only through anaerobic metabolism, and acetic acid, lactic acid, CO2 and H2 are the main end products. The ratio between acetic acid and lactic acid formation was variable, and therefore the yield of H2 per glucose was also variable. To optimise the yield of H2 by T. neapolitana, the production of acetic acid has to be maximised at the expense of lactic acid production. This was observed in cultures exposed to malonic acid, which only produced a little lactic acid and gave the highest yields of H2.