As shown in Fig. 3, the pretreatment of lignocellulose is generally followed by the hydrolysis of the pretreated carbohydrate fractions. Since most pretreatment techniques already hydrolyze hemicellulose [16, 19, 46], this section focuses on the enzymatic hydrolysis of cellulose – the main component of lignocellulose.
4.2.1 Cellulases: activities, structures, limiting factors, and non-hydrolyzing proteins
The hydrolysis of cellulose to glucose necessitates three different types of cellulases: endoglucanase (EG, EC 184.108.40.206), cellobiohydrolase (CBH, EC 220.127.116.11), and β-glucosidase (EC 18.104.22.168) [7, 8, 29, 30, 59]. As shown in Fig. 4A, EGs without cellulose-binding domains (CBD) hydrolyze cellulose chains at internal amorphous sites, generating oligo- or polysaccharides of various lengths and, therefore, new chain ends [8, 60, 61]. In contrast, CBD-exhibiting EGs (such as EG I and EG II) can additionally hydrolyze cellulose chains within ordered/crystalline regions . Cellobiohydrolases, also called exoglucanases, attack the ends of cellulose chains and act in a processive manner, thereby releasing cellobiose . Here, CBH I and CBH II hydrolyze from the reducing and non-reducing ends, respectively. In contrast to EGs without CBD, CBHs can hydrolyze amorphous as well as crystalline cellulose, presumably detaching cellulose chains from the crystalline structure [29, 62–64]. Finally, β-glucosidases convert cellobiose to glucose . This hydro-lytic activity is especially important, since CBHs and EGs are severely inhibited by cellobiose [65, 66]. Nevertheless, glucose is also a weak inhibitor for all cellulases and reduces the final yield of cellulose hydrolysis [46, 65, 67]. In general, cellulase systems consisting of all three cellulase types show a higher collective activity than the sum of the individual activities [68, 69]. This synergism can be explained by: (i) endo-exo synergy, since EGs produce new cellulose chain ends for CBHs; (ii) exo-exo synergy between CBH I and CBH II, since one chain can be attacked simultaneously from the reducing and non-reducing end; and (iii) synergy between β-glucosidase and the other cellulases, since β-glucosidase removes the strong inhibitor cellobiose [8, 62, 66, 68, 70].
Figure 4. (A) Scheme of the enzymatic hydrolysis of cellulose. (B) Overview and classification of established cellulase assays. A detailed description of the assays can be found in section 4.2.3. Abbreviations: cellulose-binding domain (CBD), dinitrosalicylic acid (DNS), glycosyl hydrolase (GH), high-performance liquid chromatography (HPLC), p-hydroxybenzoic acid hydrazide (PAHBAH).
Download figure to PowerPoint
Since CBHs and EGs hydrolyze insoluble cellulose, the adsorption of these cellulases onto cellulose is a prerequisite for hydrolysis . Consequently, most CBHs and EGs have a common structure and consist of a CBD connected by a linker region to a catalytic domain [29, 63, 71]. Whereas the CBD facilitates the efficient adsorption onto insoluble cellulose, the catalytic domain performs the hydrolysis of glycosidic bonds. The highly O-glycosylated linker region contributes to the stabilization and the conformation of cellulases .
In general, the hydrolysis of insoluble cellulose performed by CBHs and EGs is the rate-limiting step for the whole hydrolysis . This enzymatic step is primarily hindered by the recalcitrant structure of cellulose and its physical properties [64, 73–76]. Thereby, the degree of polymerization, accessibility, and crystallinity are the main limiting factors [13, 23, 55, 77]. Among these, cellulose accessibility is the most important factor for hydrolysis [13, 23, 24]. It reflects the total surface area available for direct physical contact between cellulase and cellulose and, therefore, determines cellulase adsorption as well as the rates and yields of cellulose hydrolysis [24, 78]. Furthermore, crystallinity is a relevant parameter affecting the reactivity of adsorbed cellulases . According to various authors [79, 80], crystallinity may also influence cellulase adsorption and, thus, cellulose accessibility [23–25]. To achieve high cellulose hydrolysis rates and yields, the prior pretreatment needs to increase accessibility and to reduce crystallinity [55, 76, 81–84].
In recent years, different types of non-hydrolyzing proteins have been investigated that significantly support the enzymatic hydrolysis of cellulose [77, 85–87]. Prominent examples are expansins or expansin-related proteins (e.g. swollenin from Trichoderma reesei) as well as single CBDs. During hydrolysis, these non-hydrolyzing proteins bind to the cellulose. As a result, cellulose microfibrils (Fig. 3; diameter: ca. 10 nm; [12, 88, 89]) are dispersed, and the thicker cellulose macrofibrils or fibers (Fig. 3; diameter: ca. 0.5-10 μm, consisting of microfibrils; [8, 88, 89]) swell, thereby decreasing crystallinity and increasing accessibility [85, 86, 90–92]. This phenomenon was named amorphogenesis (Fig. 4A, see swollenin) [85, 90]. Furthermore, cellulose-binding proteins can lead to deagglomeration of cellulose agglomerates (diameter: > 0.1 mm, consisting of cellulose fibers) [86, 93, 94], thereby separating cellulose fibers from each other and additionally increasing cellulose accessibility. Ultimately, amorphogenesis as well as deagglomeration promote cellulose hydrolysis . Recently, a systematic and quantitative analysis of the effects of swollenin on cellulosic substrates and their subsequent hydrolysis was published . Moreover, oxidative proteins of the glycosyl hydrolase 61 (GH61) family have been described as another type of non-hydrolyzing proteins [85, 95–97]. In the presence of ascorbate, glutathione or the enzyme cellobiose dehydrogenase, these GH61 proteins cleave cellulose by oxidation (Fig. 4A, see GH61) [97, 98]. The position of the oxidation has been reported as the reducing end, non-reducing end, or both, suggesting differences amongst GH61 proteins . Finally, this oxidative cleavage enhances the activity of cellulases and lowers the required cellulase concentration [97, 98]. Despite their great potential, the precise promoting mechanism of GH61 proteins is still unclear.
4.2.2 Expression of cellulases by Trichoderma reesei
The filamentous fungus T. reesei (synonym: Hypocrea jecorina) is an efficient cellulase producer achieving very high expression levels (> 40 g/L) and yields (> 0.2 g cellulase per g carbon source) of secreted cellulases [8, 99, 100]. Although other eukaryotes (e.g. Aspergillus spec.) and bacteria (e.g. Clostridium thermocellum) express cellulases, T. reesei has been the focus of cellulase research for over 60 years . T. reesei expresses two CBHs (CBH I and CBH II), at least five EGs (EG I, EG II, EG III, EG IV, and EG V), and two β-glucosidases (β-glucosidase I and β-glucosidase II) [99–107]. The relative amounts of the individual cellulases (g cellulase per g total cellulase) are as follows: CBH I (40-60%), CBH II (12-20%), EG I (5-10%), EG II (1-10%), EG III (< 1-5%), EG IV (< 1%), EG V (< 5%), β-glucosidase I (1-2%), and β-glucosidase II (< 1%) . Similar values of relative cellulase levels have also been reported by Goedegebuur et al. . By varying these relative cellulase levels, hydrolysis rates and yields can be optimized for different cellulosic substrates .
4.2.3 Screening of cellulase activities
In recent years, cellulases have been screened and engineered through directed evolution and rational design approaches leading to novel or improved cellulases [5, 108–110]. In order to screen and characterize these variants, cellulase assays are essential [5, 111]. Fig. 4B gives an overview of established cellulase assays. These assays can be categorized into three general approaches: quantification of substrate, quantification of product, and quantification of the physical properties of the substrate.
During hydrolysis, the decrease in cellulose concentration can be detected by weighing or chemical methods (Fig. 4B). Weighing is not suitable when using small amounts of cellulose, because limitations of sampling and instrumentation accuracy can lead to high coefficients of variation . Cellulose can be directly quantified by the K2Cr2O7-H2SO4 assay  or, after liquefaction, by total sugar assays such as the phenol-H2SO4  and the anthrone-H2SO4 assay .
The majority of cellulase assays are based on the detection of hydrolysis products (Fig. 4B). Reducing sugar assays, in particular the dinitrosalicylic acid assay, are the most common cellulase assays . Recently, enzymatic cellobiose assays were also described . Derivatized glycosides may also be used to quantify cellulase activities or inhibition constants [115, 116]. The released chromophores or fluorophores from these water soluble derivatized glycosides (such as p-nitrophenyl glycosides or methylumbelliferyl glycosides) can be easily detected and measured after hydrolysis . However, data based on soluble substrates are not entirely pertinent to hydrolysis of insoluble substrates , because enzymatic hydrolysis involves the CBD of cellulases in addition to the catalytic domain [8, 48, 49, 92].
Alternatively, cellulase activities can also be determined by measuring changes in the physical properties of cellulose: swelling factor, fiber strength, structure collapse, viscosity, absorbance, or scattered light. For example, viscometric measurements have been frequently used to determine the initial hydrolysis rate of EGs while using soluble carboxymethyl cellulose (CMC) [104, 117, 118].
A detailed description of all aforementioned cellulase assays and their associated substrates (Fig. 4B) is given in the review of Zhang et al. . Despite the apparently comprehensive number of substrates tested, cellulases are often screened with impractical model substrates that do not reflect the real cellulosic biomass in industrial processes. In particular, the application of soluble substrates should be avoided during screening experiments, because the adsorption of cellulases onto the cellulose is not accounted for [5, 119]. Over the last years, automated cellulase screening systems have been developed that are based on insoluble cellulosic substrates [120, 121]. Of particular note is the development of a cellulase assay that simultaneously combines high-throughput, online analysis, and insoluble cellulose in one simple system . This assay is based on the BioLector technique, which monitors the scattered light intensities of cellulose suspensions in a continuously shaken microtiter plate. In the future, automated and sophisticated cellulase assays will considerably accelerate the optimization of cellulose hydrolysis.