Thermophilic Degradation of Hemicellulose, a Critical Feedstock in the Production of Bioenergy and Other Value-Added Products

HEMICELLULOSES

After cellulose, hemicellulose is the most abundant renewable energy source (16). The name hemicellulose is ascribed to polysaccharides of heterogeneous composition. Therefore, they require sets of diverse enzymes for complete depolymerization. The major hemicelluloses include xylans, mannans, and xyloglucans. The structural compositions of these polysaccharides are well known; however, depending on the plant source, the plant age, and the plant tissue, the structural composition of these polymers may greatly vary in their fine structures. For further reading on hemicellulose structure, the reader is referred to pertinent reviews (17–19). In addition, we provide a schematic representation of the major hemicelluloses, including xylans, mannans, and xyloglucans, in Fig. S1 in the supplemental material.

THERMOPHILIC BACTERIA EQUIPPED WITH EXTENSIVE HEMICELLULOSE-DEGRADING ENZYMES

Interest in the bioconversion of hemicellulose to value-added products (16) has resulted in detailed analysis of thermophilic bacteria with potential for complete hydrolysis of these abundant polysaccharides. The enzymes in two bacteria, in particular, have been extensively characterized, and in this review, we will focus on their hemicellulose degrading machineries. The two bacteria of interest are Thermoanaerobacterium polysaccharolyticum and Anaerocellum thermophilum, which have been renamed Caldanaerobius polysaccharolyticus and Caldicellulosiruptor bescii, respectively (20–22).

The bacterium C. polysaccharolyticus was isolated from a waste pile of a corn-canning factory and grows under anaerobic conditions, although it is aerotolerant. The optimum growth temperature is 65 to 68°C, and the optimum pH is close to neutral (6.8 to 7.0). Growth was not seen above 70°C or below 45°C (20). Thus, C. polysaccharolyticus is classified as a thermophile. In contrast, C. bescii is a hyperthermophile, since it grows above 80°C. Similar to C. polysaccharolyticus, C. bescii is an anaerobe, and it grows in the temperature range of 42 to 90°C, with an optimum temperature of 78 to 80°C. Growth was neither observed at 37°C nor above 92°C. Cells of C. bescii also grow the best at pH values close to neutral (22).

The sequencing of the genomes of C. polysaccharolyticus and C. bescii and the extensive biochemical analyses of their plant cell wall-degrading enzymes provide the opportunity to explore how thermophilic bacteria hydrolyze hemicelluloses. Thus, in this review, we delve into the published literature to unravel the hemicellulose degradation strategies of the two bacteria and compare the properties of their enzymes to those of other thermophilic bacterial and fungal enzymes of similar glycoside hydrolase (GH) families.

THERMOPHILIC ENDOXYLANASES

Endoxylanases are required for initiation of xylan/arabinoxylan degradation. C. polysaccharolyticus encodes a large endoxylanase of GH family 10, and the enzyme has been designated CpXyn10A. The molecular mass of CpXyn10A is ∼170 kDa, making it one of the largest endoxylanases reported to date. The CpXyn10A endoxylanase has a modular architecture composed of an N-terminal signal peptide, and its GH10 catalytic module is flanked by two family 22 carbohydrate binding modules (CBM22) at the N-terminal and 2 CBM9 at the C-terminal regions, respectively. At the C terminus of the polypeptide are two surface layer homology (SLH) sequences that suggest that CpXy10A is anchored to the outer cell membrane of the bacterium (23) (see also Fig. 1).

• Open in new tab
• Download powerpoint

FIG 1

Diagram showing the modular organization of the major endoxylanases found in the thermophilic bacteria under review. (A) C. polysaccharolyticus endoxylanase CpXyn10A with its modular architecture. (B) C. bescii endoxylanases CbXyn10A and CbXyn10B. The signal peptides (SP) in CpXyn10A and CbXyn10A suggest that the two proteins are secreted outside the cell; however, the SLH in CpXyn10A likely anchors the protein to the cell surface. CbXyn10B lacks a signal peptide and is predicted to be intracellularly located.

Since the genome sequence of C. polysaccharolyticus is not closed, we are unable to determine whether other genes encoding xylan degradation enzymes are associated with xyn10A on the genome. In contrast, the genome of C. bescii harbors a region with many genes that target xylan/arabinoxylan degradation. In this cluster are located two genes encoding GH10 endoxylanases designated CbXyn10A and CbXyn10B. Similar to CpXyn10A, the Xyn10A of C. bescii is composed of a N-terminal signal peptide, followed by two CBM22 and then the GH10 catalytic module. However, the CbXyn10A of the hyperthermophile is far smaller in size (∼70 kDa) compared to CpXyn10A (∼170 kDa). The CbXyn10B endoxylanase is even smaller in molecular mass compared to CbXyn10A. This endoxylanase is predicted to be composed of only a GH10 module and lacks a discernible signal peptide. Thus, in C. bescii, it appears that one of the two GH10 endoxylanases encoded in its xylan degradation locus (CbXyn10A) is extracellularly located, while the other enzyme (CbXyn10B) is intracellular in localization. Endoxylanases act on high-molecular-weight polysaccharides and therefore are usually secreted. Thus, it is unclear why C. bescii has an intracellularly located endoxylanase.

The recombinant forms of the two endoxylanases of C. bescii, expressed in Escherichia coli, have similar pH optima for activity (pH 6.0 for CbXyn10A and pH 6.5 for CbXyn10B); however, the larger CbXyn10A has a higher temperature optimum (85°C) than CbXyn10B (80°C). Deleting one or both CBMs of CbXyn10A led to optimum temperatures of 75 and 70°C for the truncated derivatives, respectively, suggesting that the CBM increases the thermostability of the larger endoxylanase. Appending the two CBM22 N-terminally to CbXyn10B to mimic the modular architecture of CbXyn10A led to an enzyme with an optimal temperature of 55°C, suggesting that linking the CBMs to Xyn10B adversely impacted its polypeptide fold (24). The properties of the C. polysaccharolyticus Xyn10A have not been carefully studied, although it has been demonstrated to release oligosaccharides from birchwood xylan. Long-term incubations led to the release of monosaccharides in the end products of the enzymatic reaction (23). In contrast, the two endoxylanases of C. bescii have been shown to release oligosaccharides from soluble wheat arabinoxylan (sWAX), oat spelt xylan (OSX), and birchwood xylan (BWX), and glucose-configured and mannose-configured substrates were not hydrolyzed by the two enzymes. Using the three model xylan substrates, the catalytic efficiencies of CbXyn10A were estimated as 562, 95, and 29 s⁻¹ml/mg for sWAX, BWX, and OSX, respectively. The estimated catalytic efficiencies for CbXyn10B were 779, 50, and 30 s⁻¹ml/mg for sWAX, BWX, and OSX, respectively (24).

The pH and temperature optima and the catalytic efficiencies of the foregoing enzymes were compared to those of thermophilic GH10 enzymes of other bacterial and fungal sources, and as shown in Table 1 , the temperature optima of the C. bescii endoxylanases are close to the reported upper limits of bacterial and fungal endoxylanases of the same family (90°C). In the case of the optimal pH, the C. bescii enzymes are very similar to the fungal enzymes; however, they still fall within the range reported for thermophilic bacteria. It is worth noting that the available data for fungal thermophilic GH10 enzymes show a very narrow pH range of 4.8 to 6.0, whereas the bacterial counterparts range widely from acidic to alkaline. A possible reason is the availability of data, since the available fungal data are far less than the bacterial data. Although it is not feasible to compare the catalytic efficiencies and specific activities compiled in Table 1 due to the differences in the substrates used for determining these parameters, it is clear that the C. bescii enzymes have great potential for application in the hydrolysis of xylan polysaccharides.

THERMOSTABLE β-XYLOSIDASES

Endoxylanases attack xylan polysaccharides to generate oligosaccharides that are either further degraded extracellularly or transported into the cell for hydrolysis. The monosaccharide end products can then enter a fermentation pathway, such as the pentose phosphate pathway or the Emden-Meyerhof-Parnas pathway (23). From the incomplete genome sequence of C. polysaccharolyticus, there is an 18-kb locus that contains the genes encoding the enzymes needed for further degradation of the oligosaccharides released by the endoxylanase. One of the C. polysaccharolyticus genes identified for further hydrolysis of the oligosaccharides encodes a GH3 polypeptide. In addition to the GH3 N-terminal and C-terminal domains, this polypeptide harbors a fibronectin type 3 (FN3)-like domain. The enzyme, designated CpXyl3A, has been expressed as a recombinant protein of molecular mass ∼85 kDa. Based on size exclusion chromatography, the enzyme exists as a homodimer in solution. Using artificial substrates to screen for its catalytic activity, the enzyme was confirmed to be most active on pNP-β-d-xylopyranoside. The enzyme exhibited activity in the temperature range of 40 to 75°C and the pH range of 4.0 to 6.5. The optimum temperature and pH with this artificial substrate were 65°C and 5.5, respectively (Table 2). An overnight incubation of CpXyl3A with xylo-oligosaccharides ranging from X2 (xylobiose) to X6 (xylohexaose) yielded only xylose, indicating complete hydrolysis of these oligosaccharides. CpXyl3A also cleaved cello-oligosaccharides ranging from cellobiose (G2) to cellohexaose (G6) to mostly glucose. Thus, the enzyme appears to be a bi-functional enzyme, cleaving both β-1,4-linked xylo- and cello-oligosaccharides. The highest catalytic efficiency was for xylotriose (80 mM⁻¹ s⁻¹), although similar values were also obtained for xylotetraose and xylopentaose. The catalytic efficiencies for the cello-oligosaccharides were 3 to 5 orders of magnitude lower than that of the xylo-oligosaccharides (23).

To completely hydrolyze xylan, C. bescii also encodes a β-xylosidase (CbXyl3A), and it is of interest that the modular organization of the enzyme of the hyperthermophile is the same as that observed for the CpXyl3A of the thermophile. Incubating CbXyl3A with X2 to X6 led to complete hydrolysis to xylose, as observed with CpXyl3A. The C. bescii enzyme, however, appears to prefer substrates with degree of polymerization (DP) lower than xylohexaose (X6). The CbXyl3A showed a catalytic efficiency of 78 mM⁻¹ s⁻¹ on pNP-xylopyranoside. Although the recombinant enzyme was stable at 70°C, maintaining 70% of its activity after 24 h of incubation, at 80°C it rapidly lost its activity. The subunit organization in solution has not been determined. However, it is predicted that the enzyme exists as a homodimer. Preliminary studies indicate that CbXyl3A does not cleave β-1,4-linked cello-oligosaccharides. Examination of critical residues in the active site may provide insights into this major difference between the CbXyl3A and CpXyl3A. However, this will likely require three-dimensional structures or homology modeling based on related enzymes, molecular dynamics simulations, and site-directed mutational analyses.

The GH3 family β-xylosidases of C. polysaccharolyticus and C. bescii were compared to thermophilic members reported for other bacteria and fungi. The optimal temperature for the C. polysaccharolyticus enzyme is closer to the upper limit reported for the counterparts derived from thermophilic fungi (70°C), whereas the C. bescii value matches the upper limit for the bacterial enzymes (90°C) (Table 2).

Thus, the lower and upper limits of the optimum temperature for thermophilic bacteria are, respectively, about 10 and 20°C higher than the corresponding fungal temperatures. The pH optima of the two bacterial β-xylosidases were consistent with those reported for other thermophilic bacteria, which is a narrow range of 5.5 to 6.5. In contrast, the values for the fungal enzymes ranged from highly acidic (2.5) to neutral (7.0) pH. While both the C. polysaccharolyticus enzyme and the C. bescii enzymes do not have reported specific activities on pNP-xylopyranoside, their catalytic efficiencies have been reported. The activities, although within the ranges reported for both thermophilic bacteria and fungi, are, however, far below the reported upper limits.

THERMOSTABLE α-GLUCURONIDASE

Xylan/arabinoxylan polysaccharides may be decorated with 4-O-methylglucuronyl groups linked to the xylose backbone. Complete hydrolysis and utilization of the xylose units depend on the cleavage of this side chain, since its presence may limit the enzymatic activities of endoxylanases and β-xylosidases through steric hindrance. The importance of this enzyme in the degradation of xylan substrates is underscored by the presence of a GH67 encoding gene in the 18-kb xylan degradation locus on the C. polysaccharolyticus genome. The encoded polypeptide contains three domains conserved for this family of polypeptides, i.e., an N-terminal GH67, a central GH67, and a C-terminal GH67 domains. The C. polysaccharolyticus polypeptide was expressed and its estimated molecular mass in solution was determined to be ∼158 kDa, which is twice its calculated molecular mass. The polypeptide was thus predicted to exist as a homodimer in solution. Using several substrates, it was observed that CpAgu67A has significant activity on alduronic acids, which are xylo-oligosaccharides with 4-O-methyl-glucuronyl side chains (Fig. 2). Incubating the alduronic acids with CpAgu67A led to release of xylo-oligosaccharides into the reaction mixture, indicating that the enzyme is an α-glucuronidase (23). Importantly, the solute binding protein of a transporter encoded within the xylan degradation locus of C. polysaccharolyticus recognizes both xylo-oligosaccharides and alduronic acids. Such a transporter will be essential for the transport of the products of the endoxylanase (extracellular) into the cell, where Agu67A is predicted to function. The specific activity of CpAgu67A for alduronic acids was estimated to be 154 IU/mg.

• Open in new tab
• Download powerpoint

FIG 2

Schematic representation of alduronic acid showing short β-1,4-linked xylose chains with methyl-glucuronyl side chain. The α-glucuronidases, such as the Agu67A of C. bescii and C. polysaccharolyticus, cleave the side chain as shown in the diagram, and in combination with a β-xylosidase will completely degrade the branched oligosaccharides to glucuronic acid and xylose.

Within the C. bescii genome is a gene encoding a polypeptide with a similar domain architecture as CpAgu67A. The GH67 encoding gene of C. bescii was expressed in E. coli, and the recombinant protein was purified. The optimal pH and temperature for CbAgu67A were 5.5 to 6.0 and 70 to 75°C, respectively. Incubating CbAgu67A with alduronic acids leads to cleavage of the 4-O-methyl-glucuronyl groups, with the release of straight chain xylo-oligosaccharides. Furthermore, adding CbXyl3A to the CbAgu67A/alduronic acids reaction led to complete hydrolysis of the xylo-oligosaccharides to xylose, demonstrating that the β-xylosidase (CbXyl3A) and the α-glucuronidase (CbAgu67A) work together to completely hydrolyze the methyl-glucuronic acid decorated xylo-oligosaccharides (alduronic acids) (Fig. 2 and Table 3). Kinetic parameters have not been determined for either CbAgu67A or CpAgu67A due to the heterogeneity of the substrate used for the enzymatic assay. The CbAgu67A lost most of its activity at 65°C and 70°C with only 20% and 8% activities, respectively, remaining after incubation for 24 h. Both enzymes (CpAgu67A and CbAgu67A) lack signal peptides, suggesting that they are intracellularly located. The hyperthermophilic C. bescii grows at very high temperatures, and the very low thermostability of CbAgu67A suggests that the enzyme in the cell is stabilized, perhaps through posttranslational modification, such as glycosylation. It is also possible that the enzyme attains more thermostability when it binds to substrate, a hypothesis amenable to testing in the future.

The recombinant enzymes from both bacteria have temperature optima within the range determined for Agu67 of thermophilic bacteria, i.e., 50 to 85°C, although the hyperthermophilic enzyme exhibited a value about 10°C lower than the reported upper limit in thermophilic bacteria. The pH optima of the two enzymes under review were also not very different from those reported for other thermophilic bacteria, i.e., ∼6.0, whereas the fungal enzymes have pH optima ranging from 4.5 to 6.5. The specific activity of CpAgu67A on alduronic acids is 4- and 38-fold higher than the reported data for bacteria and fungi, respectively. However, the data sets for both the bacteria and the fungi are very small and may not be a true reflection of the properties of these enzymes in the two groups of organisms.

THERMOSTABLE ARABINOFURANOSIDASE

The bacterium C. polysaccharolyticus has a large gene cluster encoding arabinan-degrading enzymes, and we predicted that the genes within the arabinan degradation cluster are deployed both in the degradation of hemicelluloses and pectin. The size of the gene cluster (17 kb) is similar to the xylan degradation gene cluster described above for this bacterium. Within the arabinan targeting locus are genes encoding two GH51 polypeptides that function as α-l-arabinofuranosidases (CpAbf51A and CpAbf51B). The temperatures and pH optima of CpAbf51A and CpAbf51B are 70 and 65°C and 6.5 and 5.5, respectively. Both enzymes were demonstrated to completely hydrolyze α-1,5-linked linear arabino-oligosaccharides of chain length six or less and also oligosaccharides with side chains at position O-3, O-2, or O-3 and O-2 to arabinose. Within the gene cluster is also the gene encoding CpAbn43A, which was shown to hydrolyze both arabinan and debranched arabinan to mostly arabinobiose. Coupling the reaction of CpAbn43A with the GH51 enzymes led to the conversion of both arabinan and debranched arabinan to the unit sugar arabinose. It was also demonstrated that the two GH127 polypeptides (CpAbf127A and CpAbf127B) encoded in the arabinan degradation locus have no activity on α-linked arabinose substrates but rather have robust activity on β-arabinofuranose residues. Thus, C. polysaccharolyticus is highly adapted to depolymerizing arabinose-configured substrates with the capacity to cleave diverse linkages. The enzymes were stable at 65°C, with 60% of activity retained by CpAbn43A, CpAbf51A, and CpAbf51B after 24 h of incubation. Using pNP-α-l-arabinofuranoside as the substrate, CpAbf51A and CpAbf51B exhibited catalytic efficiencies of 79.5 and 69.4 mM⁻¹ s⁻¹, respectively (67). The arabinan-degrading enzymes should therefore aid C. polysaccharolyticus in completely depolymerizing arabinoxylan or arabinose-linked polysaccharides.

An arabinofuranosidase, designated CbAra51A, has been characterized from C. bescii. The protein is ∼75 kDa in molecular mass as determined by SDS-PAGE, which is slightly larger than the arabinofuranosidases CpAbf51A and CpAbf51B of C. polysaccharolyticus, which migrated below 66 kDa (67). The C. bescii enzyme is also highly versatile, since it shows hydrolytic activities on arabinan, debranched arabinan, rye arabinoxylan, wheat arabinoxylan, and oat-spelt xylan (24). The optimal pH and temperature of CbAra51A were determined to be 6.0 and 90°C, respectively. The enzyme lacks a signal peptide, as also observed for the C. polysaccharolyticus GH51 arabinofuranosidases. Therefore, it is predicted to function in the intracellular environment. We hypothesize that these enzymes act on arabino-oligosaccharides and xylo-oligosaccharides decorated with arabinose side chains that are transported into the cell, after initial degradation of the polysaccharides in the extracellular environment. The catalytic efficiency of CbAra51A with pNP-α-l-arabinofuranoside as the substrate was 1,121.5 mM⁻¹ s⁻¹. Therefore, this enzyme has a far superior activity on the artificial substrate compared to its C. polysaccharolyticus counterparts (CpAbf51A and CpAbf51B) (24).

Based on the artificial substrate pNP-α-l-arabinofuranoside, CbAra51A has a higher optimal temperature (90°C) than its counterparts reported in both bacteria (a maximum of 85°C) and fungi (maximum of 70°C). The temperature optima of CpAra51A and CpAra51B, on the other hand, were within the range reported for both bacteria and fungi (Table 4). The thermophilic fungal enzymes appear to function best under acidic pH conditions. In contrast, the bacterial enzymes had pH optima ranging from acidic to alkaline. The pH values of CbAra51A, CpAbf51A, and CpAbf51B from C. bescii and C. polysaccharolyticus fell within the bacterial range. In the case of catalysis, the C. bescii enzyme CbAra51A with a catalytic efficiency of ∼1,120 mM⁻¹ s⁻¹ is far more active than other thermophilic bacterial Ara51 proteins, although this is within the range reported for thermophilic fungi (84 to 1,904 mM⁻¹ s⁻¹). The catalytic efficiencies of the two Ara51 enzymes of C. polysaccharolyticus were, however, very low compared to the C. bescii enzyme, and the values were also far lower than the 700 mM⁻¹ s⁻¹ reported as the highest value in the thermophilic bacteria (Table 4).

THERMOSTABLE ACETYL-XYLAN ESTERASES

Hemicelluloses such as xylan, arabinoxylan, and galactoglucomannan carry acetyl groups linked to the C-2 and/or the C-3 positions. In order to completely hydrolyze hemicelluloses, bacteria encode acetyl xylan esterases that cleave these side chains to enhance the activities of other enzymes that act in concert to depolymerize the polysaccharides. In the xylan degradation locus of C. polysaccharolyticus is a gene encoding a protein designated CpAxe4A. The enzyme has not been biochemically characterized. However, its presence in the xylan degradation gene cluster emphasizes its importance in xylan or hemicellulose depolymerization (23). An acetyl xylan esterase has, however, been characterized from the hyperthermophilic bacterium C. bescii. The enzyme is designated CbAxe1A, indicating that its polypeptide sequence belongs to the carbohydrate esterase (CE) family 1 (85), and is <45 kDa in molecular mass. These esterases are, however, known to exist in solution or function in different oligomerization states. The oligomerization state of CpAxe1A is currently unknown. The enzyme is, however, very active on pNP-acetate at 75°C from pH 6.0 to 8.0 (24). Due to the instability of the substrate, analysis beyond these conditions were not performed. The catalytic efficiency of CbAxe1A on pNP-acetate was 566.6 mM⁻¹ s⁻¹. The enzyme did not lose any activity after 24 h of incubation at 60°C. About 60% of activity remained after incubation at 65°C for 24 h, and almost all activity was lost after 19 h of incubation at 70°C. We have been unable to find a significant number of research reports on acetylxylan esterases from thermophilic bacteria; however, reports on two enzymes from Thermoanaerobacterium sp. suggest that similar esterases from this bacterium have a neutral pH and temperature optima of around 80°C. The substrate used for the Thermoanaerobacterium sp. esterase activity was 4-methylumbelliferyl acetate (86). Therefore, the data could not be compared to those for CbAxe1A, where pNP-acetate was used as the substrate. There are also only few biochemically characterized fungal acetylxylan esterases, and the reported enzymes have optimal temperatures and pH values of ∼60°C and 8.0, respectively (87, 88).

THERMOSTABLE MANNANASES

Mannans are an important source of hemicellulose, and different forms of mannan occur in nature. A critical enzyme for the depolymerization of mannan is β-mannanase, which cleaves the β-1,4-linked mannose backbone commonly found in mannans. The β-mannanase or CpMan5A of C. polysaccharolyticus was originally cloned as an endoglucanase. Further analyses, however, showed that the enzyme is far more active on mannans. Similar to the endoxylanase CpXyn10A, the mannanase is a large polypeptide, composed of several modules. An N-terminal signal peptide, suggesting that the enzyme is secreted, followed by a GH5 catalytic domain, a large region of unassigned function, a tandem repeat of CBM16, and finally three SLH sequences at the C terminus that suggest that the ∼120 kDa enzyme is also anchored to the cell surface. CpMan5A was most active on glucomannan, and its specific activity of close to 2,500 IU/mg was about twice that on locust bean gum, a galactomannan. The specific activity on β-mannan was 800 IU/mg, whereas the specific activity on guar gum was 250 IU/mg. In addition to hydrolyzing carboxymethylcellulose, a model substrate for cellulose (specific activity of about 100 IU/mg), the enzyme also exhibited strong activity on lichenin (a polysaccharide composed of repeating units of glucose in β-1,3 and β-1,4 linkages). Both locust bean gum and guar gum contain α-linked galactose units, with guar gum containing larger amounts of galactose. It was postulated that the lower specific activity of CpMan5A on guar gum compared to locust bean gum might be explained by this difference in galactose content. CpMan5A was also shown to hydrolyze longer (DP >4) chains of β-1,4-linked glucose and mannose chains. The catalytic efficiencies on mannopentaose and mannohexaose were 33 and 120 mM⁻¹ s⁻¹, respectively, and on cellopentaose and cellohexaose the values were 9.3 and 20 mM⁻¹ s⁻¹, respectively. These observations further confirmed the preference of mannose-configured compared to glucose-configured substrates by the enzyme. The structures of the two CBM16 from the C. polysaccharolyticus Man5A are the only ones published from this family (89), and deleting them from the polypeptide led to dramatic reduction of activity.

The hyperthermophilic bacterium C. bescii is also equipped with mannanase activity, and here we consider the properties of two of these enzymes. The first mannanase activity described for this bacterium is a large polypeptide composed of a signal peptide, a GH9 module, followed by three CBM3 modules and then a GH5 catalytic module at the C terminus. The enzyme was shown to exhibit endoglucanase activity associated with the GH9 module and a mannanase activity associated with the GH5 module. The optimum pH and temperature of the CbCel9B/Man5A, using phosphoric acid swollen cellulose as the substrate, were 5.0 to 5.5 and 85°C, respectively. Surprisingly, the thermostability of this enzyme, as an endoglucanase, at 80 and 85°C was quite low, with <20% of activity remaining after 24 h of incubation. The pH and temperature optima of CbCel9B/Man5A on mannan were 5.5 to 6.5 and 90°C, respectively. The thermostability of the enzymes on mannan substrates have not been determined. The mannanase module or the GH5 and also the wild-type protein showed a preference for mannose-configured substrates with a degree of polymerization (DP) of ≥3. The catalytic efficiencies on locust bean gum, guar gum, and konjac glucomannan were 2,290, 308, and 581 s⁻¹ml/mg, respectively. When the GH9 module was deleted leaving the CBMs and the GH5 module, a reduction in catalytic efficiencies was observed. Hence, the values decreased to 1,893, 162, and 460 s⁻¹ml/mg for locust bean gum, guar gum, and konjac glucomannan, respectively (90).

A second polypeptide composed of a signal peptide, a GH5 module, two CBM3, and a GH44 module at the C terminus has been characterized from C. bescii. The GH5 module in this polypeptide shares 99% polypeptide sequence identity to the one present in CbCel9B/Man5A, and the enzyme was designated CbMan5B/Cel44A. On phosphoric acid swollen cellulose, the pH and temperature optima were not very different (5.0 and 85°C) from that of CbCel9B/Man5A. The catalytic efficiencies of CbMan5B/Cel44A for konjac glucomannan, guar gum, and lichenan were 50.0, 333.6, and 172.1 s⁻¹ml/mg, respectively. Interestingly, the GH44 module also harbored some activity toward mannose-configured substrates (91). However, even with the GH44 contributing to the mannan degrading activity of this polypeptide, its overall performance on different mannans was less than the hydrolytic activity of CbCel9A/Man5A, an observation that may be attributable to the presence of only two CBMs in CbMan5B/Cel44A, especially since the identity of the Man5A and Man5B modules are close to 100%. In contrast, the differences in the number of CBMs and their positions (N-terminal or C-terminal to the GH5 module) may impact the overall affinity for the mannan substrates and hence the catalytic activity. Importantly, the polypeptide was incapable of degrading mannobiose, suggesting that the GH5 mannan-degrading modules of C. bescii and the Man5A of C. polysaccharolyticus will both require accessory enzymes to cleave mannan polysaccharides completely to mannose for fermentation.

The optimal temperatures of the mannanases from the two organisms fall within the range reported for thermophilic bacteria (50 to 90°C) (Table 5). The enzymes under discussion also performed the best at pH 5.5, which coincided with the lower limit of the range reported for thermophilic bacteria (5.5 to 8.0). Thus, while the range reported for the thermophilic bacteria span acidic to alkaline pH, hitherto-reported thermophilic mannanases from fungal GH5 family exhibited pH optima in the acidic range (1.0 to 6.0). Note, however, that the diversity of the mannan substrates used in the determination of the specific activities and the catalytic efficiencies makes it difficult to provide a meaningful comparison.

THERMOSTABLE MANNOSIDASES

C. polysaccharolyticus has other genes encoding enzymes that aid in complete mannan degradation, including the man5B gene encoding CpMan5B. The man5B gene clusters with other genes that are required for further processing of manno-oligosaccharides derived from the catalytic action of endomannanases or mannosidases on mannan polysaccharides. Located in the mannan degradation gene cluster are genes coding for a GH130, a transcriptional regulator, and also a transporter system for which the solute-binding protein CpMBP1 has been characterized and the three-dimensional structure determined. Interestingly, the gene for CpMan5A also clusters with other mannan processing genes, including transporters and a GH130 encoding gene (110). The enzyme CpMan5B is a 39.2-kDa polypeptide that lacks a signal peptide, suggesting that it functions inside the cell. The enzyme hydrolyzed both pNP-β-d-mannopyranoside and pNP-β-d-cellobioside with specific activities of 14.2 and 11.3 IU/mg, respectively. Enzymatic activity was detected with the two substrates at 45 to 80°C. On pNP-cellobioside, the optimal temperature was 65 to 70°C, and on pNP-mannopyranoside, it was about 60 to 65°C. The pH range at which activity was detected was 4.5 to 7.0, with an optimum at 5.5. The CpMan5B cleaved manno-oligosaccharides ranging from mannobiose to mannohexaose with mannose as the end product, although in each case some mannobiose remained in the reaction mixture, suggesting that the disaccharide is not a good substrate for the enzyme. Although the enzyme also hydrolyzed cello-oligosaccharides, the hydrolysis was inefficient, and cellobiose was also resistant to hydrolysis by CpMan5B. Thus, CpMan5B had catalytic efficiencies of 25, 1.6, 220, 270, 17, and 43 mM⁻¹ s⁻¹, respectively, for pNP-cellobioside, pNP-β-d-mannopyranoside, mannopentaose, mannohexaose, cellopentaose, and cellohexaose. Although CpMan5A and its truncational derivatives showed no catalytic activity on the branched oligosaccharides 6-galactosyl-mannobiose and 6-galactosyl-mannotriose, CpMan5B cleaved these substrates with catalytic efficiencies of 1.5 and 30 mM⁻¹ s⁻¹, respectively. The enzyme was also shown to release large amounts of reducing ends from locust bean gum, guar gum, β-mannan, and glucomannan, with the highest activity seen on glucomannan, as observed also for CpMan5A. Unlike CpMan5A, however, very minimal activity was seen with CMC and lichenan as the substrates (110). The three-dimensional structure of CpMan5B has been determined at a 1.6-Å resolution. In the structure, several active site residues that are not conserved in those of other GH5 enzymes with known structures were observed in CpMan5B. It was determined that residue N92 is important for enzymatic activity and forms a novel bridge over the active site. Importantly, this structural feature is absent in other GH5 family structures (111).

Using the reported structure of CpMan5B and molecular dynamics simulations, the molecular basis for the more efficient hydrolysis of mannose-configured oligosaccharides, compared to the cello-oligosaccharides, has been explained. Briefly, the cello-oligosaccharides slow down the movement (opening and closing) of the active-site pocket, i.e., the substrate takes a longer time to leave the pocket and thus slows down new substrates from entering the active site. On the other hand, binding of manno-oligosaccharides to the active site does not elicit this inhibitory effect, leading to enhanced hydrolysis that translates to an order of magnitude more rapid catalysis (112).

Extensive analyses have not been carried out for the subsequent hydrolysis of the manno-oligosaccharides released by the bi-functional enzymes of C. bescii (CbCel9A/Man5A and CbMan5B/Cel44A); however, an enzyme CbMan5D with similar polypeptide sequence to CpMan5B has been expressed and characterized from this hyperthermophile (111). The enzyme has an optimal pH and temperature of 5.0 to 5.5 and 75 to 80°C, respectively. Similar to CpMan5B, this enzyme also hydrolyzes cello-oligosaccharides (DP of ≥4) and manno-oligosaccharides (DP of ≥2), and the latter was hydrolyzed with far higher efficiency than the glucose-configured substrates. Thus, CbMan5D is a mannosidase with some capacity to degrade cello-oligosaccharides and, combined with the C. bescii endomannanases, should increase mannan polysaccharides hydrolysis.

OTHER MANNAN-DEGRADING ENZYMES

The analysis of the end products of hydrolysis of CpMan5B shows that this enzyme is incapable of completely hydrolyzing manno-oligosaccharides. The monosaccharide mannose is released in each reaction mixture; however, the end products always contained mannobiose. This observation suggested that cellular metabolism of the mannose, released by CpMan5B from the oligosaccharides generated by CpMan5A, will drive the reaction forward, leading to extensive hydrolysis and consumption of the manno-oligosaccharides. Alternatively, C. polysaccharolyticus may have other accessory enzymes that enable the bacterium to completely hydrolyze manno-oligosaccharides. A search of the incomplete genome of C. polysaccharolyticus showed that it encodes five GH130 polypeptides that likely aid in hydrolyzing manno-oligosaccharides, especially the accumulating mannobiose. The GH130 may thus hydrolyze the mannobiose into mannose and mannose-1-phosphate, which will lead to energy conservation when the mannose-1-phosphate subsequently enters the glycolytic pathway. There is a GH130 associated with the gene clusters in which the genes encoding CpMan5A and CpMan5B occur, respectively, thus emphasizing the importance of the gene products (GH130) in mannan degradation. The other three GH130 enzymes are encoded elsewhere in the genome, and the five gene products have been designated CpGH130A, CpGH130B, CpGH130C, CpGH130D, and CpGH130E, respectively. It appears, therefore, that C. polysaccharolyticus utilizes at least two GH5 enzymes and multiple GH130 enzymes for hydrolysis of mannan hemicelluloses. It is important to note that an α-galactosidase, that may be important for a more complete hydrolysis of mannan substrates, has been purified from this thermophilic bacterium and its properties published. The enzyme, which functions as a dimer, has a monomeric molecular weight of 80 kDa and exhibited pH and temperature optima of 8.0 and 77.5°C, respectively. The enzyme, however, functioned within a broad pH range (5.0 to 9.0) (113).

To our knowledge, GH130 enzymes have not been characterized from C. bescii. However, for an organism as versatile as this hyperthermophilic bacterium in the degradation of complex polysaccharides, we suspected that such enzymes are present in its genome, and by searching the CAZy database (www.cazy.org) we located three predicted GH130 polypeptides (ACM59798.1, ACM59662.1, and ACM61623.1). Furthermore, Lee et al. have characterized a GH36 α-galactosidase with an optimal pH of 5.0 and an optimal temperature of 70°C from C. bescii (114). Based on these findings, one can hypothesize that the two thermophilic bacteria use similar enzymes to degrade mannan hemicelluloses.