Members of the Order Thermotogales: From Microbiology to Hydrogen Production

Summary

Members of the deep-branching order Thermotogales are widespread in various terrestrial, submarine and subterrestrial extreme environments. This bacterial order included both thermophilic and hyperthermophilic anaerobic microorganisms so far pertaining to ten genera. It is only recently (2011) that cultivation of a mesophilic member of this order belonging to a novel genus, Mesotoga, has been successful. All members, with the exception of Mesotoga spp., are recognized as high hydrogen producers having possible applications in biotechnology with a peculiar emphasis for members of the genus Thermotoga (e.g. T. maritima and T. neapolitana). The ecology, phylogeny and metabolism linked to hydrogen production of these bacteria, are reviewed.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

I. Introduction

Members of the order Thermotogales (Fig. 9.1) were represented as a deep-branching lineage within the phylogenetic tree (Huber et al. 1986; Reysenbach et al. 2001; Huber and Hannig 2006) thus suggesting that representatives of this order might have arisen early during the first steps of bacterial evolution. However the phylogenetic position of Thermotogales, as the nature (hyperthermophile or mesophile) of the ancestor of Bacteria, still remains a matter of debate (Brochier and Philippe 2002; Zhaxybayeva et al. 2009). Most of these Gram-negative anaerobic bacteria possess an outer sheath like structure called a “toga” ballooning over the ends of the cell (e.g. Thermotoga and Thermosipho spp.) (Huber et al. 1986; Antoine et al. 1997). Terminal protuberances on one end of the cells and single sphere containing several cells have also been observed in particular in Fervidobacterium spp. (Patel et al. 1985). Thermotogales range from thermophiles to hyperthermophiles having optimum temperature for growth above 80 °C with possible growth up to 90 °C (e.g. Thermotoga maritima, T. neapolitana, and T. hypogea). They are recognized as non-sporing rods occurring singly, in pairs or in chains with the absence of meso-diaminopimelic acid in the peptidoglycan (Reysenbach et al. 2001; Huber and Hannig 2006). This order comprises ten genera: Thermotoga, Thermosipho, Fervidobacterium, Geotoga, Petrotoga, Marinitoga, Thermococcoides and the recently described genera Kosmotoga, Oceanotoga and Defluviitoga (Di Pippo et al. 2009; Jayasinghearachchi and Lal 2011; Ben Hania et al. 2012). Because of the almost identical 16S rRNA gene sequences of K. olearia and Thermoccoides shengliensis, and their many shared phenotypic features, T. shengliensis has been proposed to be reclassified within the genus Kosmotoga and named K. shengliensis (Feng et al. 2010; Nunoura et al. 2010). Recently, a mesophilic lineage (Mesotoga) within the Thermotogales has been evidenced by the detection of Thermotogales 16S rRNA gene sequences in many mesothermic environments (Nesbø et al. 2006, 2010). The mesophilic nature of such microorganisms has been established by their isolation and cultivation in 2011 (“Mesotoga sulfurireducens” strain PhosAc3, Ben Hania et al. 2011), in 2012 (Mesotoga prima strain MesG1.Ag.4.2, Nesbø et al. 2012) and in 2013 (Mesotoga infera strain VNs100T, Ben Hania et al. 2013) with Mesotoga prima being the first described representative and type species of genus Mesotoga (Nesbø et al. 2012). All isolated Mesotoga species were confirmed to grow optimally at mesothermic conditions (40 °C for “Mesotoga sulfurireducens”, 45 °C for M. infera and 37 °C for M. prima) (Ben Hania et al. 2011, 2013; Nesbø et al. 2012) making them microorganisms of notable interest to understand bacterial evolution from mesophily to thermophily or vice versa (Nesbø et al. 2006; Ben Hania et al. 2011).

Fig. 9.1.

Phylogenetic tree based on 16 rRNA gene sequences representing the position of oilfield microorganisms (bold characters) within the order Thermotogales. Neighbor joining method was used, calculation from 1,190 aligned pb, bootstrap from 1,000 replicates. Aquifex aeolicus (AJ309733) and Bacillus subtilis (K00637) were used as outgroup (not shown). Bar scale, 0.02 substitution per nucleotide.

II. Habitat

Most of the thermophilic Thermotogales have been detected by molecular approaches and/or cultural approaches from geothermally heated environments (Table 9.1) (Huber and Hannig 2006). Most Fervidobacterium spp. have been isolated from low saline terrestrial hot springs (Patel et al. 1985; Huber et al. 1990; Andrews and Patel 1996; Friedrich and Antranikian 1996), which is also the case of Thermotoga thermarum (Windberger et al. 1989). Numerous slightly halophilic bacteria pertaining to the genera Marinitoga (Wery et al. 2001; Alain et al. 2002; Postec et al. 2005, 2010; Nunoura et al. 2007) and Thermosipho originated from marine ecosystems (e.g. deep-sea hydrothermal vents) (Huber et al. 1989; Antoine et al. 1997; Takai and Horikoshi 2000; Urios et al. 2004). Many thermophilic Thermotogales have been recovered from oilfield waters and facilities (Ollivier and Cayol 2005). They include Thermotoga spp. (T. elfii, T. petrophila, T. hypogea) (Magot et al. 2000; Ollivier and Cayol 2005), and Kosmotoga olearia (Di Pippo et al. 2009); species of both these genera were also isolated from submarine thermal vents (e.g. T. neapolitana and K. arenicorallina) (Windberger et al. 1989; Nunoura et al. 2010). Interestingly Petrotoga, Geotoga, and Oceanotoga spp. have representatives which originated only from oil reservoirs thus suggesting that such bacteria might be indigenous to these extreme environments and might possibly be relevant for enhancing oil recovery in the petroleum industry (Magot et al. 2000; Ollivier and Cayol 2005; Jayasinghearachchi and Lal 2011). However the indigenous character of Thermotogales to the oil field ecosystems should be considered with caution as their presence may result from anthropogenic activities after drilling operations or water injections during oil exploration (Magot et al. 2000). Besides geothermally heated sediments, only a few thermophilic Thermotogales have been isolated from anaerobic digesters. These include Thermotoga lettingae (Balk et al. 2002) and the recently described Defluviitoga tunisiensis (Ben Hania et al. 2012). Thermotoga lettingae was isolated from a thermophilic sulfate-reducing bioreactor operating at 65 °C with methanol as the sole substrate. D. tunisiensis was isolated from a mesothermic bioreactor (37 °C) treating lactoserum and phosphogypsum (Balk et al. 2002; Ben Hania et al. 2012). As mentioned above, there is now evidence for the presence of Thermotogales (Mesotoga) in mesothermic environments such as enrichment cultures degrading chlorinated compounds, temperate hydrocarbon–impacted sites, oil reservoirs, oil gas storage, anaerobic bioreactors treating solvent-containing pharmaceutical wastewater, and a gas-fed bioreactor treating sulfate- and zinc-rich wastewater (Ben Hania et al. 2011, 2013; Nesbø et al. 2006, 2012). In this respect, we may expect Mesotoga spp. to be of ecological relevance in the bioremediation of polluted sites (Ben Hania et al. 2011).

III. Metabolic Features

A. Electron Donors

Members of the Thermotogales are considered as heterotrophic fermentative microorganisms able to ferment sugars (Table 9.1) (Reysenbach et al. 2001; Huber and Hannig 2006). All thermophilic Thermotogales have been demonstrated to grow on complex organic substrates such as peptone and yeast extract, the latter being required to ferment sugars (Huber and Hannig 2006). Besides monosaccharides (e.g. glucose, fructose, xylose), di- and trisaccharides (e.g. sucrose, lactose, cellobiose, raffinose), Thermotogales can also ferment polysaccharides (Huber and Hannig 2006). Their use of cellulose has been reported, in particular for Thermotoga maritima, T. neapolitana, Fervidobacterium islandicum, and Marinitoga camini (Huber et al. 1986, 1990; Windberger et al. 1989; Wery et al. 2001; Huber and Hannig 2006). Thermotoga maritima, T. neapolitana, and M. camini were also reported to ferment glycogen (Huber et al. 1986; Windberger et al. 1989; Wery et al. 2001; Huber and Hannig 2006). The use of starch has been reported many times within the Thermotogales (e.g. Thermotoga, Geotoga, Petrotoga, Thermosipho, Fervidobacterium, Marinitoga, Kosmotoga spp.) (Huber and Hannig 2006). Xylanolytic activity has been detected in T. maritima, T. hypogea, and Petrotoga spp. (e.g. P. mobilis and P. olearia) (Huber et al. 1986; Davey et al. 1993; Fardeau et al. 1997; Lien et al. 1998). Marinitoga camini was shown to degrade chitin and F. pennovorans to degrade keratin (Friedrich and Antranikian 1996; Wery et al. 2001).

The alcohols possibly fermented by Thermotogales include mannitol (e.g. T. naphtophila), and glycerol (e.g. T. lettingae, T. neapolitana and F. nodosum) (Patel et al. 1985; Takahata et al. 2001; Van Ooteghem et al. 2004; Huber and Hannig 2006). Methanol was poorly fermented by T. lettingae, however, in the presence of thiosulfate or elemental sulfur as terminal electron acceptors, or methanoarchaea as hydrogenotrophic partners, methanol was more quickly used (12 days for the oxidative process instead of 30 days for the fermentative process) (Balk et al. 2002). Methanol utilization has also been demonstrated for T. subterranea, T. elfii, T. maritima and T. neapolitana by the same authors (Balk et al. 2002).

Regarding organic acids utilization, pyruvate served as energy source for some Thermotoga and Petrotoga spp. (Huber and Hannig 2006), and there is one report on formate utilization by Thermotoga lettingae (Balk et al. 2002). Besides formate, which was used only in the presence of thiosulfate as electron acceptor, T. lettingae was the first Thermotogales reported to ferment lactate (Balk et al. 2002). Thereafter, the use of lactate has been also evidenced in “Mesotoga sulfurireducens” and in M. infera where the presence of elemental sulfur as terminal electron acceptor was required thus suggesting that lactate was oxidized, but not fermented by this bacterium (Ben Hania et al. 2011, 2013). T. lettingae was also shown to oxidize acetate in the presence of thiosulfate or hydrogenotrophic methanoarchaea as terminal electron acceptors in agreement with results for previously described thermophilic or mesophilic acetate-degrading bacteria coupled to a methanogenic partner or an electron acceptor (Balk et al. 2002).

B. Electron Acceptors

Thermotogales do not only use a wide range of organic substrates as energy sources but have also the ability to reduce sulfur-containing compounds (e.g. elemental sulfur, and thiosulfate, but not sulfate) into sulfide (Table 9.1) (Huber and Hannig 2006). Amongst Thermotogales genera, Geotoga is the only genus in which no thiosulfate-reducing species was reported (Davey et al. 1993). The ability of Thermotogales to reduce elemental sulfur in the presence of sugars was first reported for T. maritima and was found stimulatory for its growth (Huber et al. 1986). However, this reductive process was suggested to be a detoxifying process preventing H₂ accumulation rather than an energy-yielding electron sink reaction (Huber et al. 1986; Huber and Hannig 2006). A similar conclusion could have been drawn regarding thiosulfate reduction by Thermotogales. However growth experiments with T. maritima and T. neapolitana in the presence of thiosulfate suggested that thiosulfate reduction could be regarded as an energy-yielding reaction through an oxidative phosphorylation process from sugars (Ravot et al. 1995). Indeed, for both bacteria, despite important improvements of growth observed in the presence of thiosulfate, there was no significant change in the end-products of sugar metabolism (Ravot et al. 1996). Besides thiosulfate and elemental sulfur, Fe(III) was also used as electron acceptor by T. maritima and in its presence, hydrogen oxidation was demonstrated (Vargas et al. 1998). While one study indicated that T. maritima may gain energy by iron respiration (Vargas et al. 1998), other reports suggested Fe(III) as an additional electron sink together with thiosulfate and elemental sulfur to prevent the inhibitory effect of hydrogen on growth (Huber et al. 1986; Schröder et al. 1994; Huber and Stetter 2001). The ability of T. maritima to reduce Fe(III) raises questions on its possible involvement in reducing heavy metals and should merit further attention. In the presence of thiosulfate as terminal electron acceptor, T. lettingae was shown to oxidize hydrogen (Balk et al. 2002). While the possible use of cystine as electron acceptor has been established for Marinitoga, Thermotoga, and Thermosipho species (Huber and Hannig 2006), only T. lettingae was reported to use anthraquinone −2,6-disulfonate as electron acceptor (Balk et al. 2002). Results obtained with Thermotogales regarding the possible use of sulfur compounds as electron acceptors, indicate they may play a crucial ecological role in mineralizing organic matter in hot ecosystems. This is particularly true for Thermotogales originating from shallow or deep sea hydrothermal vents (e.g. Thermosipho and Marinitoga spp.) where different oxidized forms of sulfur compounds, including elemental sulfur and thiosulfate, are not limiting.

C. End-Products of Metabolism

The major end-products of sugar metabolism by thermophilic Thermotogales are acetate, hydrogen, and CO₂ (Table 9.1) (Reysenbach et al. 2001; Huber and Hannig 2006). Surprisingly, while the recent isolated mesophiles “Mesotoga sulfurireducens”, M. prima and M. infera were shown to produce mainly/only acetate from sugar utilization, there is no report on H₂ production by these microorganisms (Ben Hania et al. 2011, 2013; Nesbø et al. 2012). Moreover, in contrast to M. prima which was described as a fermentative bacterium (Nesbø et al. 2012), M. infera was shown to rather oxidize its substrates in the presence of elemental sulfur as terminal electron acceptor (Ben Hania et al. 2013). Further researches on the physiological traits of Mesotoga spp. will be therefore of interest to understand the overall metabolic capabilities within the Thermotogales. It was found, for example, that acetate was also produced during methanol fermentation by Thermotoga lettingae (Balk et al. 2002). Lactate production was detected in particular by T. maritima and was dependent on culture conditions (e.g. H₂ partial pressure) (Janssen and Morgan 1992; Schröder et al. 1994). It was also produced by Marinitoga camini when growing on sugars (Wery et al. 2001). Ethanol has been measured at many occasions (e.g. Geotoga, Petrotoga, Kosmotoga, and Oceanotoga spp.) together with isovalerate, isobutyrate, and/or propionate (e.g. M. camini, K. olearia), but also alpha-aminobutyrate, hydroxyphenyl-acetate or phenylacetate (Huber and Hannig 2006) as end-products of sugar metabolism. Studies with T. maritima indicated that this bacterium used glucose via the Embden-Meyerhof glycolytic pathway and, to a lesser extent, via the Entner-Doudouroff pathway (Schröder et al. 1994; Selig et al. 1997; Huber and Hannig 2006). However the importance of the use of both pathways in other Thermotogales when fermenting sugars is still poorly documented and deserves further investigation. Notably, it was demonstrated in T. neapolitana that glucose was taken up via an active transport system that was energized by an ion gradient generated by ATP derived from substrate-level phosphorylation (Galperin et al. 1996; Huber and Hannig 2006). It is established, that several Thermotogales have a high ratio of acetate produced versus sugar consumed, thus indicating that they are good candidates for H₂ production from the biomass approaching the theoretical maximum yield of 4 mol H₂ per mol glucose consumed (Schröder et al. 1994; Van Ooteghem et al. 2004; Eriksen et al. 2011). This is emphasized by the capacities of many species to use various carbohydrates including cellulose, hemicelluloses and starch together with proteinaceous compounds. Besides acetate, lactate and hydrogen, L-alanine was also found as a significant end-product of sugar fermentation by Thermotoga elfii, Fervidobacterium islandicum, F. nodosum, F. gondwanense, and Thermosipho africanus with up to 0.52 mol L-alanine produced per mol glucose consumed (T. africanus). In contrast, T. maritima and T. neapolitana were found to be poor L-alanine producers (Ravot et al. 1996). In the presence of thiosulfate, a decrease of the L-alanine/acetate ratio was observed for F. islandicum, T. africanus, T. elfii, Thermotoga SEBR 7054, and Thermotoga lettingae (Ravot et al. 1996; Balk et al. 2002). For all these bacteria, the presence of thiosulfate caused a shift of metabolism with more acetate and less L-alanine being produced from sugars thus enabling them to obtain more ATP from substrate level phosphorylation via the formation of acetyl-CoA. It was hypothesized that similarly to the hyperthermophilic archaeon, Pyrococcus furiosus, L-alanine production from glucose fermentation resulted from alanine transferase activity coupled with glutamate dehydrogenase activity (Ravot et al. 1996). Therefore, because of a similar type of sugar metabolism by members of the Thermococcales (e.g. Pyrococcus furiosus and Thermococcus profundus), domain Archaea, placed as a deep-branching lineage within the phylogenetic tree, L-alanine production by Thermotogales has been interpreted as a remnant of an ancestral metabolism (Ravot et al. 1996). Interestingly, L-alanine was also produced when T. lettingae grew on methanol as energy source in the presence of thiosulfate or elemental sulfur and this was the first report of L-alanine formation from a C1 substrate. In contrast, in the presence of a methanogenic partner (e.g. Methanothermobacter thermoautotrophicus) methanol was completely oxidized to CO₂ (Balk et al. 2002). A complete oxidation of acetate was also observed when T. lettingae was cocultured with an hydrogenotrophic methanoarchaea, while L-alanine was produced from acetate in the presence of thiosulfate (Balk et al. 2002).

Table 9.1. Characteristics of ten genera of Thermotogales order.

D. Oxygen Tolerance

Although Thermotogales are recognized as strict anaerobes, there are evidences that they may cope with limited amount of oxygen during their growth (Table 9.1). T. maritima may grow in the presence of 0.5 % of oxygen (Le Fourn et al. 2008), T. neapolitana was shown to grow under oxygen concentrations ranging from 1 to 6 % (Van Ooteghem et al. 2004) (see Sect. IV.A). Kosmotoga olearia, an oilfield isolate, was reported to grow in the presence of 15 % oxygen (Di Pippo et al. 2009). This indicates that Thermotogales may use biochemical strategy(ies) to face oxidative stress as already reported for other strict anaerobes (e.g. sulfate-reducing bacteria) (Krekeler et al. 1998; Teske et al. 1998; Cypionka 2000; Dolla et al. 2006). Thermotoga maritima has been studied in recent years with regard to the strategy(ies) possibly used to deal with O₂ and reactive oxygen species (ROS) such as peroxides. Yang and Ma (2005) have purified and characterized a heterodimeric NADH oxidase catalyzing oxygen to hydrogen peroxide. They proposed this enzyme together with other hydrogen peroxide scavenging enzymes as acting in an oxygen-removing system. Le Fourn and collaborators (2008) using differential proteomics analyses identified a flavoprotein, homologous to the rubredoxin oxygen reductase (FprA) of Desulfovibrio sp. that was overproduced when T. maritima was cultivated under oxic conditions. They provided evidence that by reducing oxygen to water, this enzyme had a crucial role in protecting the bacterium against oxygen and suggested that NADH oxidase and rubredoxin oxygen reductase were involved in this process (Le Fourn et al. 2008). However, they outlined that the direct reduction of oxygen to water by FprA might be a preferred system as compared to the production of hydrogen peroxide, known to be highly toxic to cells. Later on, Le Fourn et al. (2011) showed that T. maritima cells could consume oxygen at a rate of 41.5 nmol min⁻¹ per mg of total protein and demonstrated that this bacterium reduced oxygen via a three-partner chain involving an NADH oxidoreductase (NRO), a rubredoxin (Rd) together with the rubredoxin oxidoreductase mentioned above (FprA), known as a flavor-diiron protein. They concluded that the genes coding for the three components O₂ reduction system were acquired by the Thermococcales, domain Archaea, through a single horizontal gene transfer event (Le Fourn et al. 2011). Because of the position of Thermotogales and Thermococcales within the phylogenetic tree, it has been suggested that such mechanism has been important for the first anaerobes to adjust to the presence of traces of oxygen on the “primordial” Earth. Oxygen uptake has also been observed when T. maritima was grown in a 2.3-L bioreactor under controlled oxygen exposure (Lakhal et al. 2010). Transcriptomic analysis, indicated that when exposed to oxygen for a short time, T. maritima had to deal with oxygen but also with the peroxides produced. The oxygen reductase FprA appeared as primary consumer of oxygen, followed by alkyl hydroperoxide reductase and peroxiredoxin-encoding genes as main ROS-scavenging systems when higher concentrations of O₂ were reached (Lakhal et al. 2011). It is noteworthy that the expression of the gene hyd that encodes the single hydrogenase of T. maritima was drastically affected by the presence of oxygen (Lakhal et al. 2011). These data are in accordance with the known sensitivity of hydrogenases towards oxygen (Vincent et al. 2005). Regarding the anaerobic metabolism of T. maritima, batch cultures indicated that it significantly decreased the redox potential (E_h) of the culture medium to about −480 mV (Lakhal et al. 2010) similarly to what was observed for the strict anaerobic methanoarchaea (Fetzer and Conrad 1993). Finally, under oxidative conditions, Lakhal et al. (2010) observed that glucose consumption rate by T. maritima decreased with a concomitant shift in glucose metabolism towards lactate production.

From these results, we may conclude that despite the high sensitivity of T. maritima to oxygen, this bacterium adapted an adequate strategy to face exposures to this gas (see Sect. IV.B for more details). This may explain why this hyperthermophilic bacterium can survive and even grow in shallow hydrothermal vents where partially oxygenated conditions cannot be precluded.

E. Hydrogen Sensitivity

Hydrogen which accumulates during the fermentation processes of carbohydrates by Thermotogales, with the exception of Mesotoga spp. (see also § III.D) is known to inhibit the growth of most of them (Huber and Hannig 2006). No growth occurred when cultures of T. maritima were pressurized with hydrogen-containing gas (H₂/CO₂ = 80:20; 300 kPa) (Huber et al. 1986). Similar observations were done with Thermotoga thermarum and T. neapolitana (Windberger et al. 1989), Petrotoga miotherma, Geotoga petraea, and Thermosipho melaniensis (Davey et al. 1993; Antoine et al. 1997). However for all these microorganisms, growth inhibition could be overcome by gassing the headspace with N₂ or N₂/CO₂ or by the addition of S° or thiosulfate (e.g. Petrotoga mobilis) which served as an electron acceptor in the culture medium; concomitantly H₂S was formed. Depending on microorganisms (see comments above), the addition of elemental sulfur and/or thiosulfate may or may not have a stimulatory effect on growth. H₂ removal in the gas phase may also result from co-cultures of Thermotogales with thermophilic to hyperthermophilic methanoarchaea (e.g. Methanococcus, Methanopyrus spp.) or other hydrogen oxidizing archaeons (e.g. Archaeoglobus or Ferroglobus spp.) (Huber and Hannig 2006). Some genera are hydrogen tolerant. Geotoga subterranea was not inhibited by the hydrogen concentration mentioned above (Davey et al. 1993). Marinitoga camini and M. piezophila reached maximum cell concentrations with 0 % H₂ while no growth occurred with 80 % H₂ (Wery et al. 2001; Alain et al. 2002), M. hydrogenotolerans can grow in the presence of 100 % H₂ in the gas phase (Postec et al. 2005) and this makes it the most hydrogen tolerant member of the Thermotogales. This confirms that Thermotogales, depending on species, have different H₂ sensitivities as reported earlier by Ravot et al. (1996).

IV. Hydrogen Production by Thermotogales spp.

A. Thermodynamic Features

The organisms described in this chapter are mainly thermophilic or hyperthermophilic anaerobes with temperature optima above 65 °C. As stated in the preceding paragraphs, many of these, such as Thermotoga maritima and T. neapolitana are capable of performing fermentative hydrogen production from a wide variety of substrates. Indeed, thermophilic hydrogen production benefits from some general advantages of performing such a process at elevated temperatures thank to a lower viscosity, better mixing, less risk of contamination, higher reaction rates and no need for reactor cooling (Wiegel et al. 1985). As already stated in Chap. 1, to make hydrogen production economically sustainable, organisms are needed that can generate hydrogen or directly or indirectly from biomass. As cellulose and hemicellulose are the most abundant polysaccharides in nature, xylose and glucose are the predominant monomeric sugars available (Kapdan and Kargi 2006). Further, starch and sucrose can be abundantly present in plants as storage material. Interestingly, the bacterial species belonging to Thermotoga have the capacity to hydrolyze most of the substrates derived from biomass. For example, as also reported in Chap. 8, both Caldicellulosiruptor and Thermotoga spp. contain a variety of glycoside hydrolases and transferases stating their metabolic capacity (Vanfossen et al. 2008).

A few thermodynamic considerations clearly indicate that under standard conditions (reactants concentration equal to 1 M, at 25 °C and pH 7.0), glucose oxidation to CO₂ and H₂ has a positive Gibbs energy change (reaction 9.1). This means that H₂ production requires an input of extra energy.

$$ \begin{array}{l}\mathrm{Glucose}+12{\mathrm{H}}_2\mathrm{O}\to 6{{\mathrm{H}\mathrm{CO}}_3}^{-}\\ {}\kern4.8em +6{\mathrm{H}}^{+}+12{\mathrm{H}}_2\end{array} $$

(9.1)

$$ \varDelta {\mathrm{G}}^{0^{,}}=+3.2\mathrm{KJ}/\mathrm{mol} $$

Table 9.2. H₂ production rates and H₂ yields from various sugars conversion by Thermotoga spp.

As shown in Table 9.2, most of the available literature reports that under optimal growth conditions, the oxidation of one hexose molecule will result in the formation of a variable number of hydrogen molecules (>2 ≤ 4) in addition to acetate and CO₂. The maximum amount of ATP is obtained through production of acetate but this can only occur if all the reducing equivalents are disposed in the form of hydrogen molecules. Apparently, the a priori requirement to get a significant hydrogen production yield is to keep a low hydrogen partial pressure through the use of a hydrogen-consuming system (Schink and Stams 2006). Under this latter condition, as shown in reaction 9.2, glucose oxidation to acetate, CO₂ and H₂ has a significantly negative Gibbs energy change.

$$ \begin{array}{l}\mathrm{Glucose}+4{\mathrm{H}}_2\mathrm{O}\to \mathrm{Acetate}\hbox{-} +2{{\mathrm{H}\mathrm{CO}}_3}^{-}\\ {}\kern4.56em +4{\mathrm{H}}^{+}+4{\mathrm{H}}_2\end{array} $$

(9.2)

$$ \varDelta {\mathrm{G}}^{0^{,}}=-206.1\mathrm{KJ}/\mathrm{mol} $$

In most fermentative hydrogen producers, catabolism via the Embden-Meyerhof pathway generates reducing equivalents in the form of NADH at the level of the glyceraldehyde-3-phosphate dehydrogenase (GAP deh) and reduced ferredoxin by the pryruvate:ferredoxin oxidoreductase reaction. Under standard conditions, the mid-point potentials of NAD⁺/NADH and ferredoxin^ox/ferredoxin^red are −320 mV and −380 mV, respectively (Thauer et al. 1977). Recycling of these redox couples can be accomplished by various reactions such as, for example, the production of lactate though reduction of pyruvate by NADH; however, the feasibility of these reactions is a priori determined by the standard Gibbs free energy change (ΔG⁰’) of each specific conversion step. This latter consideration predicts that the production of H₂ by reduction of H⁺ with NADH is a thermodynamically unfavorable process as expected by the low mid-point potential (E⁰’ = −414 mV) of the H⁺/H₂ redox couple. Although the situation is more favorable in the case of ferredoxin (E⁰’ = −380 mV) the formation of other products such as ethanol and lactate is thermodynamically more feasible. Thus, although the microbial reduction of protons to lead H₂ generation is a metabolic unexpected process, for certain thermophiles (see Table 9.1) the amount of hydrogen produced is close to four molecules per molecule of glucose oxidized, suggesting that in these microorganisms the thermodynamic constraints are somehow overcome as the growth conditions are quite far from the standard physiological values. Indeed, the actual Gibbs energy change as a function of both the substrates and products concentrations (see Eq. 9.3) predicts that even the reduction of H⁺ by NADH becomes exergonic (−4.7 kJ/mol) if the partial H₂ concentration [P(H₂)] is kept as low as 10⁻² kPa.

$$ \varDelta \mathrm{G}'=\varDelta \mathrm{G}'+\mathrm{RT}\; \ln \left(\left[\mathrm{C}\right]\;\left[\mathrm{D}\right]/\left[\mathrm{A}\right]\;\left[\mathrm{B}\right]\right) $$

(9.3)

Further, the thermodynamic of the process is affected by temperature at which the reaction takes place according to Eq. (9.4)

$$ \varDelta {\mathrm{G}}^0=\varDelta \mathrm{H}-\mathrm{T}\varDelta {\mathrm{S}}^0 $$

(9.4)

which says that at temperatures higher than standard conditions (25 °C), the Gibbs free energy change for the overall reaction from glucose to acetate (Eq. 9.1) is more favorable. A second consideration that might explain why thermophiles show such unexpected H₂ yields, is based on the fact that as previously shown by Thauer et al. (1977) and Amend and Plyasunov (2001), the hydrogen partial pressure needed to make reaction (9.2) feasible varies from 0.022 kPa at 25 °C to 2.2 kPa at 100 °C. Thus, at room temperature, hydrogen must be rapidly removed to avoid the inhibition of reaction (9.2) while the presence of 10² higher hydrogen concentration is tolerated at 100 °C.

Another possible explanation for the hydrogen formation from redox couples having a mid-point potential higher than −414 mV (E⁰’ of H⁺/H₂) might be the presence of a reversed electron transport (RET) mechanism linked to membrane bound NAD-dependent hydrogenases. Although this mechanism has never been described in thermophiles, its presence in the genus Clostridium (Cl. tetanomorphum), where a sodium gradient is used to drive the reduction first of ferredoxin and then for hydrogen production (Boiangiu et al. 2005), does not exclude a priori that other fermenting bacteria support the uphill reduction of H⁺ by NADH through the use of RET. Alternatively, it has clearly been shown by Schut and Adams (2009) that in cells of Thermotoga maritima (T. maritima) ferredoxin is a more suitable reducing agent for hydrogen production than NADH.

B. The Hydrogenases of Thermotoga spp.

As overviewed in Chaps. 2, 3, 4, and 8, the enzymes responsible for hydrogen production (H₂) combining hydrogen protons and reducing equivalents (2H⁺ + 2e⁻) are the hydrogenases (EC 1.12.99.6 and EC 1.12.7.2) also catalyzing the reversible oxidation of molecular hydrogen. In anaerobic thermophiles, two main types of hydrogenases, based on their metal content, are found: [Fe-Fe] and [Ni-Fe] hydrogenases. Further, hydrogenases can use different types of electron carriers, e.g. NAD, NADP, FAD and ferredoxin (Fd), which are reduced in the glycolytic pathway and in particular during the conversions of both glyceraldehyde-3-P to 3-P-glycerate and pyruvate to acetyl-CoA. In most fermentative hydrogen producers, reduced electron carriers generated in these steps (NADH and Fd_red) need to be re-oxidized to keep the glycolytic pathway functioning and this disposal mechanism can be different among the different thermophilic hydrogen producers (Jenney and Adams 2008).

As many other bacterial species, Thermotoga maritima uses the Embden-Meyerhof pathways for glycolysis resulting in acetate, lactate, ethanol, CO₂ and H₂. However, recycling of reducing equivalents is performed by a trimeric [Fe-Fe] hydrogenase which uses both NADH and Fd_red in a 1:1 ratio to generate hydrogen (see Eq. 9.5).

$$ \begin{array}{l}\mathrm{NADH}+2\;{\mathrm{Fd}}_{\mathrm{red}}+3{\mathrm{H}}^{+}\to {\mathrm{NAD}}^{+}\\ {}\kern1.92em +2\;{\mathrm{Fd}}_{\mathrm{ox}}+2{\mathrm{H}}_2\end{array} $$

(9.5)

This so-called “flavin-based bifurcating enzyme” is coupling the exergonic oxidation of Fd_red (E⁰’ of Fd_ox/Fe_red = −453 mV) to drive the unfavorable oxidation of NADH (E⁰’ of NAD⁺/NADH + H⁺ = −320 mV) to produce H₂ (E⁰’ = −420 mV) (Schut and Adams 2009). As this mechanism is favored by low hydrogen pressures, in the case of higher H₂ pressures, a switch from acetate to lactate production is seen (Huber et al. 1986). This however does not seem to affect the bifurcating mechanism of the hydrogenase, which presumably remains the same.

The anaerobically purified “bifurcating” hydrogenase of T. maritima, the enzyme being inactivated in the presence of even trace amounts of oxygen, is composed of three subunits – HydA (73 kDa), HydB (68 kDa) and HydC (18 kDa) – in a 1:1:1 ratio stoichiometry. The holoenzyme showed an apparent molecular mass of 500 kDa at pH 7.0 and one of 150 kDa at pH 10.0. The enzyme contained loosely bound FMN along with more than 30 Fe per heterotrimer in line with sequence analysis prediction (Buckel and Thauer 2012). Based on this latter approach, HydA subunit should harbor the active site interacting with hydrogen (H₂). This prediction was confirmed by showing that HydA alone, after dissociation of the trimeric-complex with urea, can catalyze the reduction of viologen dyes with H₂. Besides the hydrogen interacting site, HydA contains 3x[4Fe-4S] and 2x[2Fe-2S] iron-sulfur clusters with a 43 % sequence similarity to the monomeric [Fe-Fe] hydrogenase of Clostridium pasteurianum. HydA of T. maritima differs however in having a C-terminal extension with a [2Fe-2S] binding site which is lacking in the monomeric enzyme from C. pasteurianum.

The HydB subunit shows a 70 % similarity to the gene product HndC of the NADP⁺-reducing [Fe-Fe] hydrogenase from Desulfovibrio fructosovorans and 60 % similarity to NuoF of the NADH:ubiquinone oxidoreductase from E. coli. Within the sequence, there are two highly conserved stretches, one featuring NAD⁺ binding sites and the other recalling orthodox FMN binding sites. The C-terminal part contains Cys motifs that could bind three [4Fe-4S] clusters. At its N-terminus HydB contains four Cys residues that are suggested to be involved in binding a [2Fe-2S] cluster (Verhagen et al. 1999).

The smaller subunit of this trimeric bifurcating hydrogenase, HydC, contains fours Cys residues arranged in a motif which is highly similar (58 %) to motifs in E. coli Nuo and D. fructosovorans HndA, which are supposed to bind a [2Fe-2S] cluster.

The above reported information, gives a picture of the HydABC complex that is tentatively shown in Fig. 9.2 where a series of assumptions were made, namely: (a) the binding site for reduced ferredoxin (Fd_red) at HydC; (b) the presence of a second FMN loosely bound to HydB; (c) the [Fe-Fe] center plus [4Fe-4S] cluster as part of the active site of the hydrogenase in subunit HydA.

Fig. 9.2.

Tentative representation of the structure and function of the HydABC complex from Thermotoga maritimae (See text for details).

It has been proposed that the three genes encoding HydABC in T. maritima are arranged in a cluster hydCBA that is most likely a transcription unit (Verhagen et al. 1999). Clustered genes for proteins with sequence similarity to HydABC product are also found in many other anaerobic bacteria such as for example Clostridium ljungdahlii (Kopke et al. 2010), Acetobacterium woodii (Poehlein et al. 2012), and Moorella thermoacetica (Pierce et al. 2008), although these gene products have not been characterized. Interestingly, the enzyme complex from the anaerobe Thermoanaerobacter tengcongensis is composed of four subunits rather than three. Most likely, this is due to the fact that in this latter species the HydA homolog lacks the C-terminal extension with the [2Fe-2S] cluster so that a fourth subunit, homologous to the C-terminal extension, is required (Soboh et al. 2004).

Another interesting observation recently done by Thauer et al. (2010) is that electron-bifurcating [Fe-Fe] hydrogenases are not found in the Archaea domain which appear to only contain [Ni-Fe]- and/or [Fe]-hydrogenases (Thauer et al. 2010).

As mentioned in Sect. III.D, several strains of Thermotogales, such as Petrotoga miotherma, Thermosipho africanus, Thermotoga elfii, Fervidobacterium pennavorans and Thermotoga neapolitana, are able to tolerate microaerophilic growth conditions and efficiently produce H₂ as a by-product of their metabolism (Van Ooteghem et al. 2002). In particular, T. neapolitana showed the highest H₂ production (25–30 % v/v) in these conditions (Van Ooteghem et al. 2004). Through the use of a bioinformatics approach it has been shown that the operon structure of T. neapolitana is the same as T. maritima in both the ordering and spacing of the ORFs (open reading frames) (Tosatto et al. 2008). In details, a high sequence conservation is preserved from a minimum of 75 % to a maximum of 91 % for all gene products with a sequence identity attributed to the [Fe-Fe] hydrogenase subunits of 85–91 %. Notably, the HydA subunits of both species share a 91 % sequence identity, are of the same length, and can be aligned without gaps. At the DNA level, the two sequences share 82 % identity, for a total of 375 mutated nucleotides, comprising three fully mutated codons corresponding to mutations R363E (GAA → AGG), E475S (GAG → TCC) and T539L (ACA → GTG). Taking into account the sequences of T. petrophila and T. maritima showing that R363 is not conserved between the two species (GAA → AAA), it has been concluded that only two residues, E475S and T539L of the T. neapolitana HydA subunit appear to be subjected to strong selection (Tosatto et al. 2008). Apparently, the functional differences between T. neapolitana and T. maritima reside in these latter subtle changes possibly involved in the conformation of the active site (H-cluster) near the surface of the protein. As suggested by Cohen et al. (2005) in the C. pasteurianum [FeFe] hydrogenase crystal structure, there might be two alternative gas diffusion pathways defining a hydrophobic channel toward the H-cluster active site. As suggested in Fig. 9.3, the possibility of a selective effect on gas accessibility based on the size of side-chains and charge distribution on the protein surface at the entrance of the hypothetical gas channels, B (Threonine 539) and A (Serine 475) has been discussed in detail by Tosatto et al. (2008).

Fig. 9.3.

Structural model of Thermotoga neapolitana HydA protein. Panel A. Model is shown as cartoon beneath a semi-transparent surface. Iron–sulfur clusters are shown as spheres. Panel B. Close-up of upper half of model. Residues forming part of hydrophobic channel pathways A and B are shown as sticks. Residues mutating between T. neapolitana and Thermotoga maritima HydA proteins are shown as blue lines. Two mutated residues forming part of hydrophobic channel pathways A (E475S) and B (T539L), are in red and indicated by arrows. Panel C. Same model as in panel B, rotated by 90° around the x-axis to show a top view of the molecule, where hydrophobic channel entrances are located (Tosatto et al. 2008).

C. Hydrogen Production as a Function of Variable Substrate

As reported above, Thermotoga spp. are able to grow on a wide array of simple sugars and polysaccharides, including starch, β-1,4 glucan (cellulose), and hemicellulose (xylan, laminarin and mannan) (Conners et al. 2006). This capacity is consistent with the production of a diverse set of proteins and enzymes that are devoted to the uptake and processing of carbohydrates (Vanfossen et al. 2008). The use of functional genomics-based approaches has provided important insights into the various mechanisms employed by these microorganisms to assimilate and metabolize carbohydrates, and has helped to identify the specific genes and operons involved (Nguyen et al. 2001, 2004). T. maritima genome encodes for the largest number of glycoside hydroxylases of any sequenced thermophile (Chhabra et al. 2003). These enzymes specifically hydrolyze the glycosidic bond between two or more carbohydrates or between a carbohydrate and a non-carbohydrate moiety and function during the mobilization of complex carbohydrates for subsequent metabolism. T. neapolitana produces enzymes such as α-galactosidase, a laminarinase, and two cellulases (endo-1,4-β-glucanases) that have orthologs in T. maritima (Zverlov et al. 1997; Bok et al. 1998; King et al. 1998), but this microorganism also produces several unique glycoside hydrolases. One example is a 1,4-β-D-glucan glucohydrolase (GghA), which hydrolyzes cellotetraose, cellotriose, cellobiose, and lactose (McCarthy et al. 2004). T. petrophila slightly differs from T. maritima and T. neapolitana with respect to the utilization of certain sugars, growing weakly on cellulose and, in the presence of a mixture of monosaccharides, utilizing glucose to a significantly lesser extent than the other two species (Takahata et al. 2001; Frock et al. 2012). Unlike T. maritima and T. neapolitana, T. elfii failed to grow on sucrose and carboxy methyl cellulose (CMC) (Van Niel et al. 2002). T. maritima was shown to have a preference for complex carbohydrates, as growth in the presence of monosaccharides was slower than growth in the presence of oligo/polysaccharides (Chhabra et al. 2003).

Use of Simple Sugars and Polysaccharides

Thermotoga spp. produce hydrogen from a wide range of organic materials including complex carbohydrates and wastes/biomass rich in sugars. Simple sugars such as glucose and xylose are readily biodegradable and thus preferred as reference substrates for studying biochemical and physiological aspects of the hydrogen production by these bacteria. A wide variety exists among Thermotogales with respect to the utilization of sugars for growth and H₂ production that is consistent with the genetic diversity between the strains and the degree of optimization of the process for H₂ production (Frock et al. 2012). Table 9.2 summarizes the results of hydrogen production by Thermotoga spp. obtained with different types of simple and complex carbohydrates and different conditions of growth. Under optimal conditions, the oxidation of glucose and xylose will at best result in the formation of 4 mol of H₂ per mole of hexose and 3.33 mol⁻¹ mol of H₂ per mole of pentose, respectively. Maximum hydrogen yields, both from hexoses or pentoses, are obtained with acetate as fermentation product. Lower yields are associated to the formation of more reduced end products compared to acetate, such as butyrate, propionate and alcohols (ethanol, butanol) and lactic acid. It would be therefore important to establish the actual bacterial metabolism resulting in acetate as end product (Kaushik and Debabrata 2004).

T. neapolitana converted effectively sucrose to H₂ with H₂ yield of 4.95 mol_H2/mol_sucrose (Table 9.2) despite the poor fructose based metabolism reported for this strain (de Vrije et al. 2009, 2010). Woodward et al. (2002) reported the list of sugars that were used by T. maritima for H₂ production. This strain was shown to preferentially use glucose, fructose, and galactose, whereas mannose and lactose metabolisms mainly produced carbon dioxide (Woodward et al. 2002). T. neapolitana and T. maritima were able to use the glucose-based complex carbohydrates, starch and cellulose, for hydrogen fermentation, although H₂ yields obtained with raw cellulose (30 mL_H2/g_cellulose) were significantly lower than with starch (180 mL_H2/g_starch) (Nguyen et al. 2008a). The pre-treatment of cellulose with chemical agents (e.g. ionic liquids) that could disrupt the hydrogen-bonding network of the polysaccharide significantly increased its degradability (Nguyen et al. 2008b).

Use of Carbon Sources from Varies Waste-Residues

Most studies on hydrogen production in Thermotoga have used glucose as carbon source although hydrogen production at a large scale should be based on cheap and renewable substrates, such as industrial/agricultural wastes and residues. These materials have often high contents of hexose and pentoses stored in carbohydrate polymers that are ideal conversion substrates for H₂ generation (Ntaikou et al. 2010). The major criteria that have to be met for the selection of substrates suitable for fermentative biohydrogen production are availability, cost, carbohydrate content, biodegradability and concentration of inhibitory compounds (Hawkes et al. 2002). As shown in Table 9.3, T. neapolitana was widely used in technical processes with variable feedstock sources owing to its many advantageous properties including tolerance of moderate oxygen amount (6–12 %) and resistance to high H₂ partial pressure (Van Ooteghem et al. 2002, 2004). This strain was shown to successfully produce H₂ from lignocellulosic materials such as rice straw, wheat straw and Mischantus (de Vrije et al. 2002, 2009; Nguyen et al. 2010b; Eriksen et al. 2011), algal biomasses (Nguyen et al. 2010c; Dipasquale et al. 2012) and food waste materials, such carrot pulp, cheese whey and molasses (Table 9.3) (de Vrije et al. 2010; Cappelletti et al. 2012). Many physical and chemical along with structural and compositional factors in complex feedstock sources often hinder the biological digestibility (Chang et al. 2001). Because of that reason, pre-treatment methods including wet oxidation under alkaline conditions, mechanical pre-treatment, mild and concentrated acid hydrolysis and solvent extractions were required for an efficient utilization of these feedstock sources promoting the accessibility of polysaccharides in the substrates for enzymatic hydrolysis (Ntaikou et al. 2010). Molasses and cheese whey were effectively converted into H₂ by both attached- and suspended-cells of T. neapolitana without any pre-treatment needed (Cappelletti et al. 2012). Besides high yields in terms of H₂ production, the Thermotoga ability of utilizing different sugars in complex mixtures is essential in making H₂ production from biomass successful. Simultaneous utilization of glucose and xylose has also been observed in T. neapolitana (de Vrije et al. 2009), while in this microorganism, catabolite repression of lactose has been demonstrated in the presence of glucose (Vargas and Noll 1996). A preference for glucose was also shown by T. neapolitana when a mixture of glucose and fructose was present in the medium although both sugars were consumed at the same time (de Vrije et al. 2010). No mechanism of carbon catabolite repression has yet been defined for T. maritima.

In addition to carbohydrate-rich residues, T. neapolitana was also shown to produce H₂ by fermenting waste glycerol that is the main byproduct of the large-scale productions of bio-diesel (Ngo et al. 2011a; Ngo and Sim 2011). As compared with mesophilic Enterobacter aerogenes fermentation of glycerol, higher H₂ yield was obtained with T. neapolitana (2.7 mol_H2 mol⁻¹ _glycerol instead of 0.9 mol_H2 mol⁻¹ _glycerol) (Ito et al. 2005) (Table 9.3).

Non-sugar substrates, such as yeast extract and trypticase that are components of the typical Thermotoga growth media, were shown to contribute to 9–12 % of the total H₂ production (d’Ippolito et al. 2010; Cappelletti et al. 2012). These media components represent undefined sources of nitrogen and carbon for bacteria that were shown to increase cell biomass and H₂ production in T. maritima and T. neapolitana cultures growing on glucose, glycerol, cheese whey and molasses (Nguyen et al. 2008a; d’Ippolito et al. 2010; Ngo and Sim 2011; Cappelletti et al. 2012). van Niel et al. (2002) reported that growth of T. elfii was completely dependent on yeast extract, while in the absence of tryptone, lower H₂ yields were obtained. Alternative nitrogen sources such as soybean meal or canola meal alone supported growth but H₂ production rates were reduced (Drapcho et al. 2008).

Because of the complexity and richness of some industrial/agricultural wastes, the utilization of these complex feedstock sources for H₂ production by Thermotoga could allow the reduction of the process-associated cost by simplifying the culture medium. A growth medium composed only by NH₄Cl, K₂HPO₄, NaCl, buffer and cysteine-HCl (see Sect. IV.D.b) led to an efficient H₂ production from molasses with T. neapolitana (Cappelletti et al. 2012). The omission of vitamin and trace elements solutions, some inorganic elements and nitrogen sources reduced the fermentation cost without a significant lost in H₂ production. Further cost reductions were achieved by replacing the lab-grade NaCl with non-refined sea salt and the cysteine-HCl with metabisulfite (see Sect. IV.D.b) (Cappelletti et al. 2012).

Table 9.3. H₂ production rates and H₂ yields from various industrial/agricultural wastes conversion by Thermotoga spp. described in the literature.

D. H₂ Production Process and Culture Parameters

Environmental parameters such as pH, hydrogen partial pressure, media components and temperature, are key factors as they influence the metabolism and therefore the fermentation end products. Thus, optimization of these processes and culture parameters is required to give enhanced H₂ yields.

Product Inhibition

Strategies for growth under H₂ inhibition conditions have been developed in hypethermophiles including T. maritima and T. neapolitana. H₂-producing archea and bacteria can use sulfur compounds such as elemental sulfur, polysulfides, and cystine as alternative electron acceptors (Adams 1990; Drapcho et al. 2008). The use of Fe(III) as electron acceptor was also observed in T. maritima when H₂ levels became inhibitory (Vargas et al. 1998). Nevertheless, the metabolic pathways of hydrogen formation are sensitive to H₂ concentrations and are subject to end-product inhibition. Therefore, the H₂ partial pressure is an extremely important factor for hydrogen synthesis. H₂ production is a means by which bacteria re-oxidise reduced ferredoxin and hydrogen-carrying coenzymes, and these reactions are less favourable as the H₂ concentration in the liquid rises (Hawkes et al. 2002). Consequently, H₂ production decreases and the metabolic pathways shift to production of more reduced substrates such as lactate, ethanol, acetone, butanol, or alanine (Levin et al. 2004). Several strategies have been developed to avoid the negative effect of H₂ accumulation. These include vigorous mixing to avoid super-saturation (Lay 2000), utilization of H₂ permeable membrane to remove dissolved H₂ from mixed liquor (Liang et al. 2002) and sparging with inert nitrogen. The application of the latter technique to T. neapolitana growing on either glucose or waste glycerol increased the H₂ production by 50–80 % (Nguyen et al. 2010a; Ngo et al. 2011b). However, d’Ippolito et al. (2010) showed that sparging with nitrogen had a little influence on H₂ yield at low ratio between the volumes of culture and headspace. Gas sparging became more important with the increase of the liquid fraction. They observed the highest H₂ yield with a culture/headspace volume ratio of 1:3. An increased ratio between gas and liquid phase volumes was also associated to a decreased synthesis of lactic acid in cultures of Thermotoga maritima (Schröder et al. 1994).

In addition to culture/headspace volume ratio, the stirring regime and the substrate concentration were shown to influence H₂ production most probably because of their correlation to the dissolved H₂ concentration (Hawkes et al. 2002). A moderate agitation of the cultures (at 75 rpm) was shown to almost double the H₂ production by T. neapolitana cultures in 20 h of growth on glucose (Van Ooteghem et al. 2002). Using the same type of culture and carbon source, H₂ production rate was improved by increasing the stirring speed from 300 to 400 rpm. However, speeds of 500 and 600 rpm did negatively affect this fermentative process (Ngo et al. 2011b).

Considering the influence of the substrate concentration on the H₂ production process, Nguyen et al. (2008a) reported that concentrations of glucose over 15 g/L had inhibitory effects on cell growth and H₂ production in T. neapolitana and T. maritima strains batch cultures. The growth and H₂ content showed the highest values at the initial glucose concentration of 7.5 g/L for both strains. In T. neapolitana cultures growing on xylose, the maximal values of biomass and cumulative H₂ production were obtained at an initial substrate concentration of 5.0 g/L, while higher concentrations of xylose were not favorable for T. neapolitana growth and H₂ formation (Ngo et al. 2012). However, at substrate concentration of 5.0 g/L, the converted H₂ yield from xylose was lower than at the initial xylose concentration of 2 g/L suggesting that the change in xylose concentration remarkably affected not only H₂ production itself but also the substrate utilisation (Ngo et al. 2012). The best performing initial concentration of waste glycerol (3.0 g/L) was less than half than the pure glycerol (7.0 g/L) despite they resulted in comparable H₂ productions (Ngo and Sim 2011). This indicates different potentials of the two types of glycerol sources to release a specific amount of H₂.

pH Buffering System

A rapid decrease in pH is observed in cultures of Thermotoga spp. fermenting sugars to H₂ causing, in some cases, the process to stop before all the substrate is consumed (Eriksen et al. 2008). Experiments with pH adjustment showed that when cultures were neutralized with either increased initial buffer or injection of NaHCO₃ or NaOH, glucose was completely consumed and H₂/acetic acid productions increased proportionally. Both the type of buffer and the initial pH have significant effect on H₂ production process. Organic and inorganic buffer systems have been tested in literature resulting in different H₂ productions depending on the growth conditions and buffers concentration in the culture medium. HEPES resulted to be the best performing buffer when compared to HPO₄:H₂PO₄, Tris-HCl, Mops, Pipes buffers in experiments using T. neapolitana batch cultures growing on glucose (Cappelletti et al. 2012). Effective H₂ productions were also obtained with HEPES-buffered medium containing raw material feedstocks such as cheese whey, molasses and glycerol waste as carbon sources for T. neapolitana (Ngo et al. 2011a; Cappelletti et al. 2012). The good buffering properties of HEPES might be due to its pK (7.55) that is near the optimum for the growth of T. neapolitana. Itaconic acid was also successfully used to overcome pH-induced limitations of H₂-producing T. neapolitana cultures growing glucose. The buffering capacity of this carbohydrate was tested after that it was found to be poorly metabolized by this strain (Van Ooteghem et al. 2004). Its applicability for H₂ productions from industrial residues was demonstrated by Ngo and Sim (2011). The addition of itaconic acid into the culture medium of T. neapolitana growing on waste glycerol increased the process performance by almost 40 % (Ngo and Sim 2011).

In addition to the buffer, the initial pH value was shown to have a significant effect on growth and H₂ production of Thermotoga spp. Anna et al. (1991) indicated that the pH control is crucial to the H₂ formation pathway because of the effects of pH on the hydrogenase activity. The optimum initial pH for T. neapolitana suspended-cells ranged from 6.5 to 7.5 depending on both the substrate and the conditions of growth (Ngo et al. 2011b). Interestingly, an higher pH range (7.7–8.5) yielded the highest hydrogen production in experiments with glucose-grown T. neapolitana attached-cells on ceramic supports (Cappelletti et al. 2012). An initial pH of 7.0 provided the most promising results in terms of H₂ and acetic acid productions in T. neapolitana growing on xylose (Ngo et al. 2012). The optimum initial pH values for growth and hydrogen production from glucose were 6.5–7.0 for T. maritima and 6.5–7.5 for T. neapolitana, respectively. The best performing initial pH for H₂ production from glycerol by T. neapolitana was 7.0–7.5 (Ngo and Sim 2011). The application of initial pH above 8 resulted in a decrease of cumulative H₂ production as well as cell concentration suggesting the influence of pH over the metabolism pathway of the bacteria (Ngo and Sim 2011).

Oxygen Exposure

Hydrogen production by Thermotoga strains is a hyperthermophilic anaerobic process, thus, high temperatures (70–90 °C) and strictly anaerobic conditions must be initiated and maintained in the reactor vessel during production.

Some researchers have reported that low concentrations of oxygen are tolerated by both T. neapolitana (Tosatto et al. 2008) and T. maritima (Le Fourn et al. 2008) and an O₂ insensitive hydrogenase have been described in T. neapolitana (Käslin et al. 1998). Van Ooteghem et al. (2002, 2004) reported that microaerobic metabolism increased the H₂ yield from T. neapolitana up to values higher than the theoretical 4 mol_H2/mol_glucose possible from fermentative metabolism (Table 9.2); however, this result was not confirmed by other researchers who observed H₂ production after the injection of O₂, but with rate and extent that were lower than those found in cultures without O₂ (Eriksen et al. 2008; Munro et al. 2009).

Prevention of O₂ exposure could be difficult on an industrial scale and requires expensive reducing agents (cysteine-HCl). Anaerobic conditions can be initiated, maintained, and monitored in the reactor vessels by (1) flushing media with nitrogen gas, (2) heating or boiling of the media to remove dissolved oxygen, (3) adding chemical agents such as sodium sulfite or cysteine-HCl to consume residual O₂ in the liquid, (4) adding resazurin to act as visible redox indicator, and (5) maintaining positive pressure in headspace to prevent air contamination (Drapcho et al. 2008). Cysteine-HCl at concentration of 0.5–1 g/L was commonly added to provide reducing conditions in media and consume residual oxygen (de Vrije et al. 2002; Van Ooteghem et al. 2002, 2004). The addition of a reducing agent to H₂ production resulted to be fundamental even when complex feedstock sources were supplemented to the culture medium. However, to reduce the cost of the process, the utilization of cheaper reducing agents was attempted. For example, the replacement of cysteine-HCl with metabisulfite (MBS) gave promising results in terms of both medium cost (90 % reduction) and H₂ yield in T. neapolitana cultures growing on molasses (Cappelletti et al. 2012). Conversely, in the cheese whey tests, the attempt to replace cysteine with MBS led to poor performances (Cappelletti et al. 2012).

Growth-Temperature and H₂ Production

Optimum temperature of growth differs among Thermotoga species and has to be considered in the bioreactor operating process to maximize H₂ production rates (Munro et al. 2009; Nguyen et al. 2008a). Optimum temperature of growth and H₂ production is 77 °C for T. maritima, 88 °C for T. petrophila and T. naphthophila and only 66 °C for T. elfii (Jannasch et al. 1988; Huber and Hannig 2006). Cultures of T. neapolitana grown at 77 and 85 °C exhibited the greatest rate and extent of H₂ production (Munro et al. 2009).

Advantages to using high temperatures for fermentation process include (1) to reduce the likelihood of contamination by H₂-consuming organisms that lessens the need for sterilization of media and equipment, (2) to favour the catalytic activity of hydrogenase of evolving H₂ (Adams 1990) (3) to directly utilize industrial organic wastewaters that are often discharged at elevated temperatures; (4) to avoid cooling down processes usually required by mesophilic fermentations that in large scale generate excess heat (Drapcho et al. 2008).

V. Future Perspectives

This Chapter summarizes our present knowledge on the microbiology, physiology, and biochemistry of the Thermotogales order with the aim to better define problems and challenges linked to H₂ production. In this respect, it is worth noting that important improvements have been recently made through optimization of both the bioprocess parameters and Thermotoga spp. to be used. Although the identification of suitable feed-stocks for fermentative hydrogen production was done, more research work to improve hydrogen production rates and yields, is required. The use, for example, of Thermotoga strains metabolically engineered and the development of a “two stage process” is likely to improve H₂ production. Indeed, this latter approach involves the fermentation of the selected substrate to both hydrogen and organic acids by Thermotoga spp. in the first stage and, in a second stage, either an additional energy extraction or the generation of highly-valuable products by exploiting the effluent of the first stage reactor. An alternative approach to improve H₂ production might also be achieved through a specific bioreactor configuration ameliorating both biomass concentration and substrate conversion efficiency by employing biomass retention systems such as granules, flocs or biofilm-formation supports.