2.1.1. Degradation of Carbohydrates
Thermoanaerobacter AK15 degraded cellobiose and all tested monosaccharides except for the two deoxysugars, L-fucose and L-rhamnose (
Figure 1). The main end product in all substrates that were degraded was ethanol (
Figure 1) which resulted in optical densities between 0.35 to 0.48 and a decrease of pH for about 1.0 pH unit compared with the control (
Supplementary Table S1). Some thermophiles have been shown to degrade deoxysugars like
Caldicellulosiruptor species of which more than half of the type strains within the genus degraded L-rhamnose to 1,2-propanediol and three strains degraded L-fucose to 1,2-propanediol [
16]. Other thermophilic anaerobes known to produce 1,2-propanediol are moderately thermophilic strains of
Clostridium [
30,
31].
The strain degraded mannitol to ethanol mainly, as has also been shown for
T. uzonensis, the closest relative to strain AK15 [
32]. Mannitol utilization was not tested in the original characterization study of Wagner and coworkers describing
T. uzonensis [
33].
The original data on ethanol production from glucose for
T. uzonensis showed 1.15 mol glucose per mol glucose degraded [
33] while
Thermoanaerobacter strain AK15 produces 1.10 ethanol mol glucose
−1 (
Figure 1). Other well-known species producing high amounts of ethanol from sugars within the genus are i.e.
T. ethanolicus and
T. pseudethanolicus [
8,
27]. The other end products are mainly acetate with traces of lactate; this was also the case for other carbohydrates tested. It is a well-known phenomenon that the ratio of end products is highly influenced by culture conditions, particularly hydrogen accumulation. Therefore, a simple experiment was performed on the strain by cultivating it in three different liquid–gas phase ratios: 0.09, 1.00, and 5.62. As shown in
Figure 2, the end product profile changes dramatically during growth under different liquid-gas phase ratios for the strain.
At high p
H2, the main end product is ethanol, or 33.3 mM which corresponds to 1.57 mol/mol glucose degraded (controls subtracted) but drops to 0.44 mM at the lowest p
H2. Similarly, acetate and hydrogen decrease at high p
H2 conditions. The higher acetate concentrations at low liquid – gas phase rations resulted in lower pH, but the optical density was similar in all conditions (
Supplementary Table S2). This highlights the importance of the culture conditions on end-product formation and has been reported for various thermophilic bacteria in recent studies [
29,
34]. Thus, the strain is highly ethanologenic and can be manipulated to convert more than 75% of the glucose to ethanol as has been reported for several other
Thermoanaerobacter strains [
35]. Also, being versatile concerning the utilization of a wide variety of carbohydrates, strain AK15 may be considered as a good candidate for bioethanol production from lignocellulosic biomass. To investigate this further the strains were cultivated on several lignocellulosic and algal biomass pretreated with dilute acid and cellulases.
2.1.3. Degradation of Biomass Hydrolysates
The strain was tested on several complex biomass types, both lignocellulosic biomass (grass, newspaper, rhubarb leaves) and macroalgae types (
Ascophyllum nodusum, Palmaria palmata, Laminaria digitata, Ulva lactuca) as shown in
Figure 4. Whatman paper was used as a reference in the same concentration as for the biomass samples.
As for the end product formation from individual sugars, the main end product from the biomass hydrolysates used was ethanol. The strain was positive for end-product formation (
Figure 4) and growth (
Supplementary Table S4). Glucose degradation under the same conditions resulted in the formation of 23.8 mM of ethanol, or 1.15 mol/mol glucose degraded. Theoretical yields of glucose from 2.5% w/v Whatman paper that is completely hydrolysed is 15.4 mM. The amount of ethanol produced from Whatman paper hydrolysates is 15.4 mM, or 1.0 mol ethanol/mol glucose which corresponds to 5.6 mM/g glucose, assuming all glucose is released during hydrolysis (control ethanol production subtracted). Not surprisingly, ethanol yields were lower in lignocellulosic biomasses (newspaper, grass, rhubarb) because its more complex nature. Yields on newspaper, grass, and rhubarb were 3.7, 4.4, and 3.9 mM g
-1 dw). The highest yields of ethanol from complex biomass by thermophilic bacteria is by
Thermoanaerobacter strain BG1L1 on wheat straw and corn stover, or 8.5-9.2 mM ethanol g
-1 dw [
8,
37]. Yields from grass are relatively good compared with other similar species.
Thermoanaerobacter AK5 produces 4.1 mM g
-1 dw grass [
29], and
Thermoanaerobacterium strain AK17 produces 5.5 mM g
-1 dw [
38]. In all cases the grass hydrolysates were pretreated with acid, as well as with enzymatic pretreatment as performed in the present study. Many other reports on thermophiles capable of good ethanol yields from lignocellulose have also been presented in recent years [
39,
40,
41,
42,
43]. Thus,
Thermoanaerobacter strain AK15 seems to be suitable for bioethanol production from complex biomass with good ethanol yields.
Macroalgae is another type of biomass that may be of future use as a potential biomass for bioethanol and bioactive compound production [
44,
45,
46,
47]. In the present study, four types of macroalgae were used, two brown algae (
Ascophyllum nodusum,
Laminaria digitata), one red algae (
Palmaria palmata), and one green algae (
Ulva lactuca). Ethanol yields for these macroalgae were in most cases lower compared with lignocellulosic biomass. The highest yields were from
Laminaria digitata, or 2.8 mM g
-1 dw which corresponds to 25% of theoretical yields. Most data on ethanol production from various macroalgae species in literature originate from yeast fermentation of hydrolysates produced by a wide variety of pretreatment methods. Hou and coworkers reported the production 14.7 gL
-1 of ethanol using an SSF approach involving cellulase and alginate lyase of
L. digitata (10 g/L) using
S. cerevisiae as a fermenting microorganism [
48]. Similarly, Li and co-workers achieved an ethanol titer of 6.65 gL
-1 of ethanol from a cellulase hydrolysate of
S. japonica (60 g/L) using
S. cerevisiae strain DK 410362 [
49] while others have achieved a 14.0 gL
-1 ethanol titer from S. japonica waste [
50]. A report by Wang et al used
Gracilaria salicornia, an invasive brown algae, to achieve a modest ethanol titer of 1.57 g/L using
E. coli strain KO11 (ATCC 55124) [
51]. A recent investigation of the fermentation of
Laminaria digitata hydrolysates, pretreated with enzymatic hydrolysis showed 0.30 gg
-1 of consumed substrate (59% of theoretical yields) [
52].
2.1.4. Degradation of Amino Acids
The is a huge amount of data concerning sugar metabolism of thermophilic bacteria in general, with most focus on ethanol production from complex biomass. Much less attention has been on the capacity of thermophiles to degrade amino acids. Most of the information on amino acid catabolism among Clostridia has been gained from well-known proteolytic members of
Clostridium sporogenes [
53],
Clostridium botulinum [
54] and
Clostridium sticklandii [
55]. The degradation of amino acids is a complex process involving several oxidation and reduction steps, and only possible under specific conditions and unlike carbohydrates, many amino acids are not degraded as single substrates because of the thermodynamics involved in the reactions [
56]. There are various pathways to degrade the 20 amino acids present in proteins, some of which are highly reduced and are not degraded unless the electrons produced in the oxidative step are scavenged [
57,
58,
59]. The most common way to degrade amino acids is to use a two-step mechanism, first an oxidative deamination of the amino acid, yielding a corresponding keto acid, which is oxidatively decarboxylated to give one carbon shorter fatty acids [
56]. This is possible under anaerobic conditions only for amino acids with high oxidation stages [
56]. The so-called reduced amino acids, e.g. the branched-chain amino acids (leucine, isoleucine, valine) and alanine. In the 1990s, several investigations showed that the reduced amino acids could only be degraded when the amino acid-degrading bacteria could dispose of the electrons produced during the oxidation of these amino acids to an external electron acceptor. This could be done either by co-cultivating the amino acid degrading bacterium with a hydrogenotrophic bacterium, i.e. methanogens, via interspecies hydrogen transfer, or by using thiosulfate as a chemical electron acceptor. This was shown by
Thermoanaerobacter brockii during growth on the branched-chain amino acids [
60] where leucine, isoleucine, and valine were degraded to 3-methylbutyrate, 2-methylbutyrate and 2-methylpropionate, in the presence of thiosulfate, respectively. Later investigations revealed that the branched-chain amino acids were not only degraded to their corresponding fatty acid but to a mixture of their corresponding fatty acid and alcohol by
Thermoanaerobacter brockii, and
Caldanaerobacter subterraneous subsp.
yonseiensis, when thiosulfate was added into the culture medium [
61]. In the present investigation, serine and threonine were the only amino acids that were degraded without the addition of thiosulfate, mainly to ethanol and acetate. In the presence of thiosulfate, these two amino acids were also degraded with a change in the proportion of the end product formation (see equations 1 and 2 below). In the case of serine degradation, acetate became the main end product and ethanol minor product. Similarly, acetate became the main end product in the presence of thiosulfate during threonine degradation, but ethanol increased as well.
Equations 1A and 1B show the stoichiometry of serine degradation without and with thiosulfate, and equations 2A and 2B for threonine, respectively.
20 mM Serine → 6.91 mM Ethanol + 11.33 mM Acetate (Equation 1A)
20 mM Serine + 20 mM S2O3 → 1.21 mM Ethanol + 21.41 mM Acetate (Equation 1B)
20 mM Threonine → 3.42 mM Ethanol + 6.15 mM Acetate (Equation 2A)
20 mM Threonine + 20 mM S2O3 →10.68 mM Ethanol + 16.70 mM Acetate (Equation 2A)
Similarly,
Thermoanaerobacter brockii shifts its end product formation during growth on serine in the presence of electron acceptor, pushing end product towards more oxidized end product like acetate and less reduced end products, like ethanol, were produced [
49].
In present study, the thiosulfate addition resulted in the degradation of the branched-chain amino acids (leucine, isoleucine, and valine) (
Figure 6). As previously reported for other
Thermoanaerobacter strains [
18],
Thermoanaerobacter strain AK15 produced a mixture of branched-chain fatty acids (BCFA) and branched-chain alcohols (BCOH) from branched-chain amino acids (BCAA) and the concentration of the acid was always higher than the alcohol. When the strain was cultivated without the addition of thiosulfate, no growth was observed (
Supplementary Table S5).
Figure 5.
End product formation from leucine, isoleucine, and valine (all individually at 20 mM concentrations) in the presence of thiosulfate (40 mM). For simplicity formation of acetate, ethanol and lactate is not shown but was in similar concentrations as in control bottles. Standard deviations are presented as error bars.
Figure 5.
End product formation from leucine, isoleucine, and valine (all individually at 20 mM concentrations) in the presence of thiosulfate (40 mM). For simplicity formation of acetate, ethanol and lactate is not shown but was in similar concentrations as in control bottles. Standard deviations are presented as error bars.
In the presence of thiosulfate leucine was degraded to 3-methylbutyrate and 3-methylbutanol, isoleucine to 2-methylbutyrate and 2-methylbutanol, and valine to 2-methylpropionate and 2-methylpropanol. BCAA have been reported to be degraded to BCFA and BCOH under anaerobic conditions, mostly by
Lactobacillus and yeasts that use the Ehrlich pathway. Usually, the concentration of these compounds is of importance for the flavor of foods and beverages [
62]. Early studies of thermophilic anaerobic bacteria were performed in several investigations on
Thermoanaerobacter brockii. This bacterium was described to degrade the BCAAs to their corresponding BCFAs, but only when thiosulfate was added to the culture medium, acting as a hydrogen scavenger. The thermodynamics in the deamination of leucine to its corresponding keto acid is ΔG + 51.5 kJ/mol whereas the ΔG of the amino acid to 2-methybutyrate is + 4.2 kJ/ Recent work in our laboratory has since then shown that indeed the main bottleneck for the degradation of these reduced amino acids is the first energy-demanding deamination step. Original investigations also showed that
both Thermoanaerobacter brockii and
Caldanaerobacter subterraneous subsp.
yonseiensis degraded the branched-chain amino acids to only their corresponding BCFAs when the strains were co-cultivated in the presence of a hydrogenotrophic methanogen, but to a mixture of their BCFA and BCOH when cultivated alone in the presence of thiosulfate [
61]. Thus, it seems to be important how effective the electron acceptor is in determining the product distribution during the degradation of BCAAs. Later studies on other strains within the genera of
Thermoanaerobacter and
Caldanaerobacter showed that this ability to produce a mixture of alcohols and acids from BCAA was common among both genera [
18]. Investigations to understand in more detail the reaction pathway these bacteria use to produce both the acid and the alcohol have been done with Thermoanaerobacter strain AK85 which is closely related to
Thermoanaerobacter uzonensis. This study showed that indeed the partial pressure of hydrogen was of great importance for the ratio of end products formed [
63]. Finally, it was demonstrated by NMR studies that these bacteria first produce the BCFA which in turn is converted to their corresponding BCOH, both for
Thermoanaerobacter strain AK85 and
Thermoanaerobacter pseudethanolicus [
63,
64].
2.1.6. Conversion of fatty acids to alcohols
Recent investigations have shown that bacteria within the genera of
Thermoanaerobacter and
Caldanaerobacter can dispose their electrons produced during glucose (and other sugars) oxidations not only to pyruvate to produce ethanol or lactate but may also use other electron acceptors like fatty acids which are converted to their corresponding alcohols [
28,
65,
66]. This was tested for strain AK15 by cultivating the strain on glucose only and on glucose in the presence of butyrate. The strain degraded glucose to a mixture of ethanol, acetate, and lactate according to equation 3:
1 Glucose → 1.24 Ethanol + 0.48 Acetate + 0.05 Lactate + 0.54 H2 (eq. 3)
When the strain was cultivated on glucose with the addition of 20 mM of butyrate the reaction stoichiometry changed according to equation 4.
1 Glucose + 1 Butyrate → 0.78 Ethanol + 0.87 Acetate + 0.58 Butyrate + 0.36 Butanol (eq. 4)
Thus, as expected, the strain produces less ethanol and more acetate in the presence of butyrate as an electron acceptor, and the fatty acid is partially converted to its corresponding alcohol, butanol. No lactate and hydrogen were observed as end products at end of fermentation in the presence of butyrate This was also investigated for other volatile fatty acids, like propionate, branched chain fatty acids and pentanol, with similar results of the conversion of the fatty acid to their corresponding alcohol (results not shown).
Thermoanaerobacter pseudethanolicus has recently been shown to convert fatty acids to alcohols, during sugar degradation [
66]. Production of high carbon alcohols from complex biomass by adding cheap volatile fatty acids to the fermentation broth of the hydrolysates is indeed a new way of biofuel production that may be of great importance in the near future.