9.1.1 Cellulose

Cellulose is the major carbohydrate component of plant cell walls, and thus is one of the most abundant biomass resources on earth. The cellulose molecule is an insoluble linear homopolymer of β-D-glucopyranose units linked together by the 1 and 4 carbon atoms. Because the adjacent glucose moieties are rotated 180° about the axis of the cellulose backbone chain as shown in Figure 9.1, the basic structural repeating unit of cellulose is cellobiose rather than glucose (Blackwell, 1982; Atalla, 1983). The ends of the cellulose chain contain different types of β-anhydroglucose units, i.e. the reducing end with a free hemiacetal (or aldehyde) group at C-1, and the non-reducing end with a free hydroxyl at C-4. The hydroxyl groups in cellulose are positioned in the ring plane, whereas the hydrogen atoms are in the vertical position. Therefore, a hydrophobic surface is formed in the cellulose chain due to the nature of successive β-anhydroglucose units.

The cellulose molecule does not exist as an isolated individual molecule. It forms a crystalline structure, known as cellulose microfibril (Nishiyama, 2009), through the tight packing of chains by intra- and intermolecular bonds such as hydrogen bonds and van der Waals forces. The crystalline structure gives cellulose high tensile strength and resistance to chemical or enzymatic attack. In nature, therefore, cellulose can be digested only by a limited number of organisms, most of which are bacteria and fungi.

9.1.2 General Properties of Cellulases

The word “cellulase” is a generic term used to refer to the enzymes that hydrolyze glucan chains in cellulose polymer into cellooligosaccharides or ultimately glucose. Traditionally, cellulases have been divided into two types, i.e. cellobiohydrolases (CBHs; EC 3.2.1.91) and endoglucanases (EGs; EC 3.2.1.4), based on their activity profiles. The former hydrolyzes cellulose molecules from the end of the chain with the release of cellobiose, whereas the latter randomly cleaves the internal β-1,4-glucosidic bond of cellulose. EGs preferentially act on amorphous cellulose and soluble cellulose derivatives such as carboxymethyl cellulose (CMC), resulting in a rapid decrease of the degree of polymerization in cellulose; they lack hydrolytic activity against crystalline cellulose. On the other hand, CBHs possess significant activity against the crystalline region of cellulose, although the majority of CBH-type enzymes show poor activity against CMC. These enzymes act concertedly to degrade cellulose biologically (Streamer et al., 1975; Wood and McCrae, 1978; Henrissat et al., 1985) in three sequential reactions: (1) EGs initially hydrolyze the cellulose chain in the amorphous region; (2) CBHs recognize the newly generated ends of cellulose chains and start the catalytic reaction with the release of cellobiose; and (3) successive hydrolysis of cellobiose by CBHs finally results in degradation of the crystalline region of cellulose. The sequence of these reactions is called the endo–exo mechanism (Figure 9.2).

Recently, the endo–exo concept has been considered as insufficient to understand cellulose biodegradation, so that the concept of a processivity mechanism has been suggested to comprehend the behavior of cellulase’s action (Davies and Henrissat, 1995). This concept explains the action of cellulases after the catalytic reaction (Figure 9.3), i.e. a high processive enzyme slides to the next cellobiose unit following hydrolysis of the β-1,4-glucosidic bond of cellulose, whereas a low processive enzyme easily detaches itself from the cellulose chain after the catalytic reaction. A correlation between structural features and processivity has been shown based on three-dimensional structures of cellulases. In particular, a tunnel structure around the active site is critical for expressing processive functions. This is covered in detail in Section 9.2.

Cellulase generally consists of two domains: (1) a catalytic domain, which catalyzes hydrolysis of the β-1,4-glucosidic bond of cellulose, and (2) a cellulose-binding domain (CBD), the binding domain from fungal enzymes, which is currently called family 1 carbohydrate-binding module (CBM1) and is one of the best characterized CBMs. This domain lacks any catalytic activity; it binds the surface of cellulose via three aromatic amino acids, as shown in Figure 9.4. Removal of this domain causes a significant decrease in both binding and catalytic activities against crystalline cellulose but shows no effects on the hydrolysis of amorphous cellulose (Tomme et al., 1988; Ståhlberg et al., 1993; Linder and Teeri, 1997). These observations suggest that CBM1 is related to the degradation of crystalline cellulose.

9.2.1 The CAZy Database: A Classification System Based on Structures

A database designated as CAZy (Carbohydrate-Active enZymes; http://www.cazy.org/; Cantarel et al., 2009) is indispensable for studying the structure and function of cellulases. The CAZy database was developed by Bernard Henrissat and colleagues, and is now used worldwide as the standard classification system of carbohydrate-acting enzymes. In the CAZy database, carbohydrate-acting enzymes are classified based on their structural similarity into four major groups, i.e. glycoside hydrolase (GH), glycoside transferase (GT), polysaccharide lyase (PL), and carbohydrate esterase (CE). Each group is divided into families, which are serially numbered. For example, one of the cellobiohydrolases is classified and designated as glycoside hydrolase (GH) family 6, and the abbreviated name of the family, GH6, is normally used.

The nomenclature of cellulases has recently been developed based on the CAZy database. As described in Section 9.1.2, cellulases are divided into EG (endoglucanase) and CBH (cellobiohydrolase) in terms of their enzymatic properties. The name of cellulases is conventionally given as a combination of EG/CBH and numbers; for example, cellulases from Trichoderma reesei are designated as “EG I” or “CBH II”. According to the recent nomenclature, the name of cellulases is a combination of “Cel” (which means cellulase), the number of the glycoside hydrolase family, and a capital letter. For example, the cellulases “EG I” and “CBH II” from Trichoderma reesei as described above are designated as “Cel7B” and “Cel6A” respectively.

Many carbohydrate-acting enzymes including cellulases consist of a catalytic domain and a carbohydrate-binding domain; the latter plays a crucial role in the enzymatic activity. In the CAZy database, these non-catalytic domains are classified into carbohydrate-binding modules (abbreviated as CBM).

9.2.2 Cellulases of Two Well-Studied Organisms, Trichoderma reesei and Clostridium thermocellum

Cellulases are classified into glycoside hydrolase families, GH5, GH6, GH7, GH8, GH9, GH12, GH44, GH45, GH48, GH51, GH74, and GH124. Enzymes belonging to GH61 also show activities similar to cellulases, but the GH61 enzymes have been reported not to be identified as cellulases; they are identified as copper-dependent monooxygenases (Quinlan et al., 2011). Although cellulases are found in many glycoside hydrolase families, none of the cellulase-producing organisms possesses genes encoding all of these GH members.

A filamentous fungus, Trichoderma reesei, is the most used microorganism in industry as a source of cellulolytic enzymes (Gusakov, 2011), and therefore numerous studies of enzymes derived from this organism have been carried out. The organism has been shown to be the anamorph of the ascomycete Hypocrea jecorina, but is still better known by its former name, T. reesei. The complete genome of T. reesei, which was sequenced in 2008, indicates that the organism possesses genes for GH5 (two genes), GH6 (one gene), GH7 (two genes), GH12 (one gene), GH45 (one gene), and GH61 (three genes) (Martinez et al., 2008). Among these enzymes, the most predominant cellulases secreted by T. reesei are Cel7A (CBH I) and Cel6A (CBH II), which account for 50–60% and ∼20% respectively of the total cellulases. Likewise, a basidiomycete, Phanerochaete chrysosporium, which is also well known as a producer of cellulolytic enzymes, secretes cellobiohydrolase (Cel7D) of up to ∼10% of the total secreted protein in liquid culture (Ubhayasekera et al., 2005).

Bacteria and fungi are important players in degrading the cellulolytic materials in nature. Although bacterial enzymes are not utilized by the industry, numerous studies on bacterial enzymes have been conducted. In particular, several anaerobic bacteria produce a multienzyme complex termed cellulosome, which is an attractive target for the structural biology. The cellulosome was first observed to exist in Clostridium thermocellum, and now several Clostridium species and Ruminococcus species as well as some other organisms have been reported to produce this multienzyme complex (Demain et al., 2005; Doi, 2008). The cellulosome is composed of a non-catalytic subunit called scafoldin that consists of multiple repeats of cohesin domains and cellulosomal enzymes, which typically comprise a catalytic domain and a dockerin domain. Each cohesin domain interacts with the dockerin domain of the cellulosomal enzymes, resulting in the formation of a multienzyme complex, or cellulosome (Figure 9.5C). Cellulases consisting of the cellulosome belong to GH5, GH8, GH9, and GH48, and the majority of the enzymes are classified as either GH9 or GH48. The catalytic domains are composed of an (α/α)₆ barrel (Figure 9.5A, B) for GH9 and GH48, and a (β/α)₇ barrel and β-sandwich (Figure 9.6A, B) for GH6 and GH7. Thus, the structural folds of cellulosomal enzymes are completely different from those of cellulases from T. reesei.

9.2.3 The GH6 and GH7 Enzymes from Trichoderma reesei

Trichoderma reesei mainly produces the GH6 and GH7 CBHs, Cel6A and Cel7A, as described above. The GH6 CBH releases cellobiose units from the non-reducing end of cellulose (EC 3.2.1.91), while the GH7 CBH removes cellobiose units from the reducing end of cellulose (EC 3.2.1.176) (Igarashi et al., 2011). GH7 CBH and GH48 CBH have a unique enzymatic character because there are only a few glycosidases that hydrolyze polysaccharides from the reducing end in an exo manner. In contrast, all the exo-acting amylases degrade starch from the non-reducing end.

A cellulose fiber contains crystalline and amorphous regions. For the T. reesei cellulase system, attacking the crystalline region by GH6 and GH7 CBHs, Cel6A, from the non-reducing end and by Cel7A from the reducing end has been proposed, whereas the amorphous region is accessible to endoglucanase action (Figure 9.6C) (Teeri, 1997). Although no structural homology is found between the catalytic domains of GH6 and GH7, an enclosed tunnel is present at the active site of both GH6 and GH7 CBHs (Figure 9.7A, B). It is noteworthy that the length of the tunnel of T. reesei Cel7A is ∼50 Å (Divne et al., 1998), which is an unusual structure among glycosidases. Some EGs are also classified as GH6 or GH7, but no enclosed tunnel is found in the structures of GH6 and GH7 EGs. Trichoderma reesei Cel6A and Cel7A possess a cellulose-binding domain that is grouped in the CAZy family, CBM1. The cellulose-binding domain of Cel6A is located on the N-terminal side, whereas the cellulose-binding domain of Cel7A is found on the C-terminal side (Linder et al., 1995; Rabinovich et al., 2002). The structural features of CBM1 are described in Section 9.1.2; the CBM1 cellulose-binding domain, which is smaller than the catalytic domain, consists of ∼40 amino acid residues.

9.2.4 Subsite and Catalytic Residues

The term subsite refers to the structural relationship between an enzyme and a polymer substrate. In the case of glycosidase and polysaccharide substrates, the glycosidase has individual regions called subsites, each of which binds to a specific monosaccharide unit of the polysaccharide substrate. The subsites that bind to the non-reducing end are labeled “minus” and those that bind to the reducing end of the substrate are labeled “plus”; the glycosidase cleaves between the −1 and +1 subsites (Davies et al., 1997). An enclosed tunnel is present at the active site of the GH6 and GH7 CBHs (Figure 9.7A, B). Therefore, the substrate cellulose is initially considered to bind the plus subsites of the GH6 CBH, and then slides through the active center tunnel until it occupies subsites −2 and −1. In the case of GH7 CBH, the plus and minus subsites have opposite signs. For GH6 and GH7 CBHs, subsite −3 and subsite +3 are presumed unnecessary to produce the respective cellobioses, but T. reesei Cel7A has been reported to have 10 subsites from −7 to +3 (Figure 9.7C) (Divne et al., 1998). In T. reesei Cel6A, the existence of eight subsites from −4 to +4 has also been proposed (Figure 9.7D) (Varrot et al., 2003).

Knowledge of the catalytic residues and the concept of subsites is of great importance in understanding the structure–function relationship of enzymes. Reaction mechanisms of glycosidases are divided into two groups: inverting mechanism and retaining mechanism. In the inverting mechanism, the enzymatic reaction is accompanied by an inversion of the anomeric configuration, whereas in the retaining mechanism the anomeric configuration of the substrate is retained. For instance, the “inverting” CBH and the “retaining” CBH produce α-cellobiose and β-cellobiose respectively from cellulose. GH6 and GH7 enzymes act through the inverting mechanism and retaining mechanism respectively. In the inverting mechanism, two amino acid residues that act as a general acid and a general base are found to be the catalytic residues, whereas in the retaining mechanism a nucleophile residue and a general acid/base residue are identified as the catalytic residues. In both mechanisms, aspartic acid and/or glutamic acid residues normally function as the catalytic residues of glycosidases (Rye and Withers, 2000). In T. reesei Cel7A, the nucleophile residue is Glu212 and the general acid/base residue is Glu217 (Divne et al., 1994). In GH6 enzymes, the catalytic general acid residue is identified as a conserved aspartic acid, Asp221 in T. reesei Cel6A, but it is still debatable concerning which residue acts as the catalytic base of the GH6 enzymes. Recent studies show that Asp175 in T. reesei Cel6A is located relatively far from the cleavage site of the substrate, but has been proposed to be the catalytic base. It acts indirectly through the hydrogen bond network of water molecules in what has been called a Grotthus mechanism (Koivula et al., 2002).

In conclusion, the structures and functions of the GH6 and GH7 cellulases are unusual compared with other glycosidases such as amylases. GH6 and GH7 CBHs are characterized by an enclosed tunnel at the active site. GH7 CBHs release cellobiose units from the reducing end of cellulose, and the catalytic base of the GH6 enzymes has been proposed to act indirectly in the so-called Grotthus mechanism. It is likely that these unusual features of cellulases make them suitable for the hydrolysis of cellulose. For more detailed information see Box 9.1.

9.3.1 β-Glucosidases

β-Glucosidases (BGLs; EC 3.2.1.21) catalyze the hydrolysis of β-glucosidic bonds in various β-glucosidic compounds, and thus are involved in various metabolic pathways such as degradation of polysaccharides, cellular signaling, oncogenesis, and host–pathogen interactions. Among these pathways, the enzymes important for cellulose degradation, i.e. cellobiases, are focused in this section. These enzymes are indispensable for the production of glucose, which is then fermented into ethanol by yeast in the subsequent process of bioethanol production.

In nature, various cellulolytic microorganisms produce BGLs during the course of cellulose degradation. Fungal enzymes are especially regarded as the most promising BGL in bioethanol production. Filamentous fungi mainly produce three types of BGLs, i.e. extracellular, intracellular, and cell wall-associated enzymes (Lynd et al., 2002); extracellular and intracellular enzymes have been well characterized so far. These enzymes, which are retaining enzymes, catalyze the hydrolysis of cellobiose and cellooligosaccharides with the release of glucose. In addition to their hydrolytic activities, the enzymes often show transglycosylation activities. Because the activity of BGL decreases at higher concentrations of glucose, improving the glucose tolerance is one of the many important targets for conducting BGL research.

The extracellular BGLs belong to the GH3 family in the CAZy database, and some of the enzymes possess CBM1 beside the catalytic domain. Because the enzymes catalyze the hydrolysis of cellobiose, which is the main product of cellulose hydrolysis by cellulases, the cellobiases are considered to have an important role in the final step of extracellular cellulose degradation (Sternberg, 1977; Eriksson, 1978). Previous studies on extracellular BGL obtained from Phanerochaete chrysosporium reveal interesting results. The enzyme prefers laminaribiose (disaccharide with β-1,3-glucosidic bond) instead of cellobiose, and the reaction efficiency against β-1,3-oligosaccharides increases with higher degree of polymerization of the substrates. This observation suggests that the enzyme is β-1,3-glucanase (Igarashi et al., 2003). Hence, not all extracellular BGLs from cellulolytic fungi are cellobiases. On the other hand, intracellular BGLs belong to the GH1 family, and the intracellular localization of the BGLs requires the uptake of cellobiose into the cell. In this regard, T. reesei is expected to be a putative diglucoside permease involved in the transport of cellobiose into the cell (Kubicek et al., 1993). In addition, intracellular BGL from T. reesei is reported to produce sophorose, which is a very strong inducer of cellulases (Mandels et al., 1962), via transglycosylation activities (Saloheimo et al., 2002), although the induction of cellulases by sophorose is observed only in a few limited fungal species, such as T. reesei.

9.3.2 Hemicellulases

Hemicelluloses, which represent about 20–35% of the plant cell wall, are heterogeneous polymers of various sugars such as xylose, arabinose, glucose, mannose, and galactose. The composition of hemicellulose depends on the source and type of plant biomass. Hardwood and graminaceous plants contain 4-O-methylglucuronoxylan and arabinoxylan as the major hemicellulose respectively, whereas hemicellulose of softwood mainly consists of glucomannan and glucuronoxylan. Xylans, such as 4-O-methylglucuronoxylan, arabinoxylan, and glucuronoxylan, have a backbone of 1,4-linked β-D-xylopyranose units. In addition to xylopyranoside residues, xylans contain various substituents, i.e. arabinoside residues, O-acetyl groups, ferulic acid, p-coumaric acid, and 4-O-methylglucuronic acid (Wilkie and Woo, 1977). Glucomannan is a mainly linear heteropolymer consisting of β-1,4-linked D-mannose and D-glucose (Timell, 1967). Xyloglucan is a highly branched heteropolymer with D-glucose and D-xylose. Xylan and glucomannan are located in the secondary wall, whereas xyloglucan is contained in the primary wall. Therefore, xyloglucan accounts for only a small fraction of the plant cell wall.

The word “hemicellulase” is a generic term used to refer to enzymes that have a role in hemicellulose degradation, and thus includes not only glycosidases but also esterases (de Vries and Visser, 2001).

Endo-1,4-β-D-xylanases (EC 3.2.1.8) are enzymes that randomly hydrolyze β-1,4-xylosidic bonds of xylan backbone (Polizeli et al., 2005). They belong to GH5, GF7, GH8, GH 10, GH11, and GH43 in the CAZy database; GH10 and GH11 have been well characterized. The GH10 enzymes act preferentially on soluble substrates and can readily hydrolyze small xylooligosaccharides such as xylotriose. Unlike GH10 enzymes, GH11 enzymes are most active in the hydrolysis of insoluble xylans, although the activity is more likely to be inhibited by the presence of arabinose decorations. Because xylanase enhances the hydrolysis of cellulose by cellulases, it has been considered as a major accessory enzyme in the saccharification of plant biomass.

α-L-Arabinofuranosidase (EC 3.2.1.55) catalyzes the hydrolysis of α-1,2-, α-1,3- or α-1,5-L-arabinofuranosidic bonds in arabinose-containing polysaccharides (Numan and Bhosle, 2006). It has been classified as GH family 3, 43, 51, 54, and 62 in the CAZy database. Some α-L-arabinofuranosidases demonstrate broad substrate specificity, i.e. acting on arabinofuranosyl residues at O-2, O-5 and/or O-3 positions as a single substituent. On the other hand, α-L-arabinofuranosidases that specifically catalyze the release of αα-1,3-L-arabinofuranosyl substituents from doubly substituted xylose residues have also been found. These enzymes work synergistically on arabinoxylan degradation with xylanases.

Acetylxylan esterases (EC 3.1.1.72) release acetate from xylan and xylooligosaccharides. The enzymes have been classified as carbohydrate esterase (CE) families 1, 2, 3, 4, 5, 6, 7, and 12 in the CAZy database. CE16 is also reported to contain acetyl esterase that releases acetate from 4-nitrophenyl-β-D-xylopyranoside monoacetate (Li et al., 2008). The removal of acetyl groups by acetylxylan esterases improves the access of xylanases to the xylan backbone and facilitates the degradation of xylans.

In addition to the enzymes described above, various other enzymes such as α-glucuronidases, xyloglucanases, galactosidases, glucuronoyl esterases, feruroyl esterases, mannanases, endoglucanases, and xylosidases are required to degrade hemicellulose (de Vries and Visser, 2001). The synergistic action of these enzymes leads to the hydrolysis of the complicated structure of hemicelluloses.

9.3.3 Starch-Hydrolyzing Enzymes: Amylases

Although producing ethanol from starch adversely affects food supplies, the process is still important to convert non-edible biomass such as crop waste into ethanol. The term “amylase” is defined as an enzyme that hydrolyzes starch, and this section includes a synopsis of amylases.

Starch is a polymer of glucose units linked predominantly via α-1,4-glucosidic linkages and some α-1,6-glucosidic linkages to form branch points. Amylases are classified according to their enzymatic reactions into the following four groups (Figure 9.9A):

In the CAZy database, numerous amylases are members of GH13, which was originally known as the α-amylase family. There are four conserved amino acid sequences in the original definition of the α-amylase family (Kuriki and Imanaka, 1999), and three catalytic residues (i.e. aspartic acid, glutamic acid, and aspartic acid residues) are located in the second, third, and fourth conserved regions. Now the α-amylase family is defined as a clan GH-H composed of three glycoside hydrolase families, including GH13, GH70, and GH77 (MacGregor et al., 2001). The essential structure of GH13 is composed of domains A, B, and C; domain A consists of a (β/α)₈ barrel with a small domain, i.e. domain B, protruding from it. Domain C is located at the C-terminal side of domains A and B, and it forms a β-sandwich (Figure 9.9B) (Kuriki and Imanaka, 1999). Isoamlase, pullulanase, and many other GH13 enzymes have some extra domains other than A, B, and C.