1. Introduction

Lignocellulosic feedstocks have enormous economic potential as a source of value-added products (VAPs) and biofuels ^[1][2][3]. Cellulose is the most abundant polymer on earth and a predominant component of lignocellulosic biomass. It is a source of fermentable sugars (glucose and cello-oligosaccharides) that serve as precursors for VAPs. Cellulose is a linear polymer that consists of β-D-glucose molecules linked by glycosidic bonds, sourced from plant material (e.g., wood pulp, cotton, or cereals such as wheat, sugarcane and rice bagasse) ^[1][4][5][6]. Many cellulose chains bundle via hydrogen bonding to constitute cellulose microfibrils, which consist of crystalline regions intersected by amorphous regions.

It is established that glycoside hydrolases (GHs) from different families act synergistically to produce fermentable sugars from the lignocellulose biomass ^{[3][7][8][9][10]}. The cellulolytic enzyme synergism is based on the simultaneous use of the cellobiohydrolases; CBHI (EC 3.2.1.176) and CBHII (EC 3.2.1.91) which attack the biomass from the reducing and non-reducing ends of cellulose chains to produce cellobiose, respectively, endoglucanase (EG, EC 3.2.1.4) which randomly cleave the amorphous regions of the to produce oligosaccharides with a DP between 2 and 10 or create new chain ends in the bulk of biomass, and β-glucosidase (β-gl, EC 3.2.1.21) which degrade oligosaccharides to glucose ^[9].  Figure 1 shows the well-established mode of action of cellulases during the synergistic hydrolysis of cellulose.

Figure 1. A schematic representation of cellulose and corresponding cellulolytic enzymes that degrade it. The enzyme cleavage sites on the cellulose are represented by red arrows. Microcrystalline and amorphous (green color) regions of cellulose consist of β-D-glucose linked by β-1,4-glycosidic bonds. Where β-gl, CBHI, CBHII and EG represent β-glucosidase, cellobiohydrolase I, cellobiohydrolase II and endo–glucanase.

The cellulolytic enzyme synergism can be grouped into two classes; (1) synergism between non-catalytic and catalytic cellulolytic enzymes (e.g., expansins and CBHs), and (2) synergism between catalytic cellulolytic enzymes; i.e., Exo-Exo, Exo-Endo, Exo-Endo-β-gl or cellulase and lytic polysaccharide monooxygenase (LPMO) synergy ^{[11][12][13][14][15][16]}. A few studies have documented Endo-Endo synergy, even though EGs constitute most diverse cellulase group, classified under GH families 5, 6, 7, 8, 9, 10, 12, 44, 45, 48, 51, 74 and not classified sequences (NC) (http://www.cazy.org/Glycoside-Hydrolases.html). Also, it is well documented that CBHs from different GH families that hydrolyze cellulose on the same cleavage sites do not synergize together, but compete for occupancy on the same binding sites ^[17][12].

Two other enzyme systems play a significant role in cellulose saccharification, these being LPMOs and non-hydrolytic enzymes (expansins and swollenins (SWOs)). According to the CAZy database, the enzymes classified under CBM33 were later assigned to auxiliary activities family 10 (AA10), while those in GH61 were assigned to AA9 [http://www.cazy.org/Auxiliary-Activities.html, ^[18]. The LPMOs in family AA9 and AA10 are copper-dependent enzymes that cleave the glycosidic bonds of crystalline cellulose regions by oxidizing the C1 (EC 1.14.99.54) or C4 (EC 1.14.99.56) and C1/C4 of the glucose residues (^[18][19], http://www.cazy.org/Auxiliary-Activities.html]. The LPMOs in family AA16 are newly classified cellulose-specific enzymes with a C1 regiospecificity ^[20]. The catalytic domain of LPMOs can also be attached to CBMs by a flexible linker ^[19][21].

Expansins and expansin-like proteins are nonhydrolytic or nonoxidative enzymes that can bind to the crystalline cellulose and weaken the degree of crystallinity of the polymer by breaking the inter- or intramolecular hydrogen bonding network  ^[22][23]. Kim et al. ^[24] demonstrated that Bacillus subtilis derived expansin sequences were homologous to maize plant expansins, and their 3D structures were similar. The role of expansins in plants is well established, and plant-derived expansins are classified into two groups, α- and β-expansins ^[4]. The α- and β-expansins only share 20% amino acid sequence similarity. Interestingly, expansins generally have about 30% sequence similarity to endoglucanases classified under GH45 ^[4]. Several studies have shown that the expansins synergize with cellulases that hydrolyze crystalline microfibrils of cellulose ^{[4][22][23][25]}. Fungi also possess another nonhydrolytic protein called swollenins, similar to the expansins ^[25]. Swollenins modify the chemical and structural features of microcrystalline cellulose resulting in a reduced degree of crystallinity, which allow the cellulases to hydrolyze cellulose effectively. Additionally, studies have demonstrated that cellulases and swollenins synergize to improve the saccharification of cellulose ^[12][25]. Herein, hydrolytic cellulolytic enzyme to non-hydrolytic enzyme synergism and the degrees of synergism (DS) thresholds to develop a model that could assist in the improvement of cellulose hydrolysis by cocktails will be explained.

Lack of investigation of the DS during synergism is attributed to a focus on hydrolysis yields by researchers. It appears that some enzyme combinations do not display a positive DS, yet these combinations demonstrate higher yields in terms of soluble sugar production compared to when the individual enzymes are used to hydrolyze the substrate. However, it is proposed that investigating synergism between enzymes should not only be limited to yield improvement but should also include DS. This would enable researchers and companies to design better enzyme cocktails for specific feedstocks. It will be addressed how can it could be achieved that higher yields and DS by reevaluating the established synergisms and setting the DS-threshold benchmarks, which can lead to higher yields.

2. Defining DS and Its Impact on Substrate Hydrolysis

It is well accepted that a DS of 1 arbitrary unit (AU) demonstrates no synergism between the enzymes used to formulate a specific enzyme cocktail ^{[9][10][17][26]}. These studies suggest that the enzymes do not cooperate during substrate hydrolysis leading to no improvement in the rate of substrate deconstruction compared to when the mono-components act individually. For example, endoglucanase and exoglucanase enzymes do not synergize and generally result in a DS of 1 AU, during the hydrolysis of the model amorphous cellulose, carboxyl-methylcellulose (CMC), because of their uncomplimentary modes of action on the substrate. A value below 1 AU shows that there is no synergism between the enzymes during the hydrolysis of the substrate. This observation suggests that one or more of the enzymes cannot hydrolyze the substrate efficiently due to competitive binding of two or more enzymes on the same substrate cleavage site. Instances which resulted in a DS below 1 AU have been documented between CBHII and EG, processive EGs from GH9. In all these cases, substrate hydrolysis was impeded, and product formation significantly reduced.  Generally, DS values above one are achieved when the enzymes in the cocktails hydrolyze different cleaving sites of the substrate, e.g., CBHI, CBHII, EG and β-gl degrade substrates with different cleaving sites, resulting in significantly higher product formation (Figure 1).

3. Synergism between catalytic enzymes

It was hypothesized that to formulate effective enzyme cocktails researchers should not only focus on the yield of soluble sugars as conventional synergy studies have done for the past several decades but should focus on both the yield of soluble sugars and the DS between enzymes. This approach will generate holistic information about the reaction with regards to the interactions between the enzymes in the cocktails, as well as the products formed. Hence, it is attempted to establish the DS threshold which results in high yields of soluble sugars. A detailed discussion of various cellulolytic enzyme synergy studies that were conducted with different enzyme combinations are presented below and summarized in Table 1.

Bp: Bacillus pumilus, Bs: Bacillus subtilis, Xc: Xanthomonas campestris, Cm: Micromonospora aurantiaca ATCC 27029, Tr-SWO: Trichoderma reesei-swollenin, Th: Trichoderma harzianum, Tl: Trichoderma longibrachiatum, Po: Penicillium oxalicum, Hj: Hypocrea jecorina, Hi: Humicola insolens, Cc: Clostridium cellulosi, Ba: Bacillus amyloliquefaciens mg: derived from metagenome, Aa: Aspergillus aculeatus, Mt: Myceliophthora thermophila, Tt: Thermothelomyces thermophilus. ^# Predicted DS values were estimated directly from published data. * Theoretical DS value thresholds which generally result in higher degrees of synergy. ^$ Solubility studies which were equivalent to synergistic activity. ^a Conversional exo–exo synergy can be substituted with processive EG in a new approach of exo-processive endo. ^b Endo–exo synergy is generally supplemented with β-gl to prevent product inhibition. ^c The enzymes were immobilized on synthetized Fe₃O₄ nanoparticles.

3.1. Synergism between processive cellulolytic enzymes; Exo-Exo or Exo-processive Endo synergy

There is strong evidence that supports the hypothesis that some backbone-cleaving enzymes always interact synergistically, resulting in a DS that is higher than 1 AU and a soluble sugar yield which is higher than that produced by the sum of the individual enzymes during the hydrolysis of biomass ^[9]. For instance, cellulase enzymes that cleave the microcrystalline cellulose, such as Cel6A and Cel7A (exo-exo synergy) or Cel6A, Cel7A and EG (exo-endo synergy) generally synergize with a DS that is higher than 1 AU ^[9][17][15]. Exo-exo synergy between cellulases has been demonstrated by Boisset et al. ^[27] using bacterial microcrystalline cellulose (BMC), while other researchers from the laboratory have used Avicel [phosphoric acid swollen cellulose (PASC) or NaOH-treated Avicel also called regenerated amorphous cellulose (RAC)] as a suitable substrate ^[10][17][28]. Badino et al. ^[15] showed that the DS between GH7 (Cel7A) and GH6 (Cel6A) enzymes sourced from Hypocrea jecorina was above 2.2 AU, while the glucose yield produced by synergistic actions was 4-fold higher than that which was produced by individual enzymes. A study by Boisset et al. ^[29] also found similar results using Humicola insolens GH7 Cel7A and GH6 Cel6A which showed a DS of about 4 AU between enzyme combinations and converted more than 50% of the bacterial microcrystalline cellulose compared to 15% by Cel6A and 28% by Cel7A alone. Based on the literature and the experience, the fungal sourced Cel7A and Cel6A synergism generally display a DS above 2 AU and will always result in a higher yield of soluble sugars than individual enzymes.

In some bacterial systems such as Clostridium phytofermentans ISDg, there is a different cellulase system that degrades the crystalline regions of cellulose ^[30]. The CBH enzymes employed by C. phytofermentans ISDg are Cel48 and Cel9. Zhang and co-workers suggested that Cel48 attaches to the substrate from the reducing end, while Cel9 (processive EG) attaches from the non-reducing end. In addition, CBHs that belong to GH 48, such as Cel48, usually contain a family 3 carbohydrate-binding domain (CBM) ^[30]. Anaerocellum thermophilum is another bacterium which has demonstrated that the GH48 and GH9 enzyme system is generally employed to hydrolyse microcrystalline cellulose via exo-exo synergy ^[31]. Additionally, Irwin and co-workers showed that there was higher synergism between Cel48A and Cel9A sourced from Thermobifida fusca compared to other enzymes from different GH families like GH5 and GH6 ^[30]. Cel9 from Bacillus licheniformis was shown to have a CBM3c linked to the catalytic domain and displayed the highest activity on PASC, BC and filter paper. Cel9 (GH 9) from Clostridium cellulosi possesses five CBMs, namely; CBM3c, three CBMX2s and CBM3b [48]. The exo-exo synergy between the Cel48 and Cel9A on the filter paper displayed a DS value of 1.5 AU ^[32]. Another study similarly reported a high DS of about 1.5 AU for Cel48 (GH48) and Cel9 (GH9), while the soluble sugars produced from Avicel was 4-fold higher than those produced by the individual enzymes ^[30]. These observations support the hypothesis that there are two or more distinct exo-cellulase systems that result in high DS and yields of soluble sugars; (1) one employed by both bacterial and fungal cellulolytic systems (mostly Cel7A and Cel6A), and (2) another unique to the bacterial cellulolytic system (Cel 48 and Cel 9A).

3.2. Exo-Endo and/or β-glucosidase synergy (cellulolytic enzyme cocktail)

Based on both theoretical and empirical evidence, CBHs, EGs and β-gls can synergize during the hydrolysis of biomass (Figure 1). Boisset et al. ^[33] demonstrated that CBHI (Cel7A), CBHII (Cel6A) and EG (Cel45A) sourced from Humicola insolens displayed synergism during the hydrolysis of bacterial cellulose ribbons (micro-crystalline cellulose). The composition of Cel6A and Cel7A were varied between 0 and 100%, while Cel45A was fixed at 1.2%. The cocktail composed of the three enzymes converted more than 60% of the bacterial cellulose ribbons and displayed a DS of about 4.5 AU, while the combination of Cel6A and Cel7A converted only 30% and had a DS of 2 AU  ^[29]. Some of the commercial cocktails such as Celluclast® 1.5L consist of Trichoderma reesei cellulases in the following combinations 55% Cel7A, 10% Cel6A, 10% Cel7B, 10%, Cel5A and 1% β-gl . The highest biomass conversion that the authors could achieve was about 50% yield, only when they added an extra 10% β-gl – but their experimental design did not factor in the DS, which shows the molecular level interactions between the enzymes. Addition of 50 U/g substrate of a GH12 EG sourced from Gloeophyllum trabeum with 10 U/g substrate of Celluclast® 1.5L resulted in an increased synergistic effect of about 14.5%, 16.1%, 29.0% and 13.4% on filter paper, hydrogen peroxide–acetic acid-pretreated pine, corn stover and rice straw, respectively ^[34].

The above information relates to the fungal cellulolytic enzyme cocktails; however, it is clear that exoglucanase systems of bacterial organisms are different to those of fungal systems (as it was proposed in the Exo-Exo synergy section). For instance, Cel6A, Cel48A, Cel6B and Cel9A sourced from Cellulomonas fimi were all processive on PASC, Avicel and crystalline celluloses Iα or IIII from green algae ^[18]. It is important to note that Cel6B and Cel9A are classified as endoglucanases, but Uchiyama et al. ^[18]  demonstrated that these enzymes are arguably highly processive and used high-speed atomic force microscopy (HS-AFM) to show that they were able to move on crystalline cellulose IIII. Interestingly, CfCel6B and CfCel5B demonstrated a DS of 2.16 AU and an improved yield of soluble sugar at a ratio of 3:1 when the enzyme loading was at 2.5 mg/g cellulose concentration ^[35]. However, when the enzyme concentration of CfCel6B and CfCel5B was increased to 10 mg/g cellulose, the DS values were approximately decreased by 1 AU for all combinations; 3:1, 1:1 and 1:3. The processive and non-processive EGs from GH family 9 and a CBHI from GH family 48 sourced from C. cellulosi CS-4-4 displayed synergy on filter paper when they were supplemented with β-gl from Caldicellulosiruptor sp ^[34]. Various molar ratios were tested in this study and the best combination for the formulated cellulose cocktail was 25:25:10:18 for CcCel9A: CcCel9B: CcCel48A: BlgA, respectively, which resulted in a DS of 2.6 and 1.86 mg/ml of glucose release. The literature contains more information regarding fungal cellulolytic enzyme cocktails, but so far, only a few bacterial cellulolytic enzyme cocktails have been recorded. These suggest that there is still a lot that needs to be done with regard to bacterial cellulolytic enzymes, to establish if their synergy can be superior to those of fungi or even commercial cocktails. It is propose that a DS-threshold above 1.5 AU can result in higher yields of soluble sugars for bacterial cellulolytic cocktails, but more studies are required to establish this claim.

4. Cellulase and LPMO synergism

Synergism between an LPMO and CBHI was demonstrated by core expression of the genes that encode for these proteins in the fungus Penicillium funiculosum that produced the LPMO naturally and five engineered strains ^[16]. It was evident that the secretome from the engineered strains (referred to as PfOAO1 and PfOA3) contained higher Avicelase activity, filter paper activity and production of hydrogen-peroxide, which suggest the presence of cellulases and an LPMO. In addition, the use of PfOAO1 and PfOA3 for the saccharification of acid pretreated wheat straw increased conversion to 80% and 75%, respectively. These findings suggest that overexpression of CBHI and LPMO in PfOAO1 and PfOA3 increased synergy between these enzymes. Interestingly, the individual application of three LPMOs from Thermoascus aurantiacus (TaAA9A), Lentinus similis (LsAA9A) and Thielavia terrestris (TtAA9E) on Avicel and PASC resulted in no release of detectable reducing sugars according to Tokin et al ^[36]. In contrast, the LPMOs synergized with TrCel7A and TrCel6A on Avicel and PASC, except for TrCel7A and TtAA9E which displayed anti-synergy. The binary synergy between TrCel6A and TtAA9E resulted in the highest DS of about 2.5 and 2 AU during PASC and Avicel degradation, respectively. The high synergism between these enzymes can be explained by the fact that TtAA9E cleaves cellulose substrates by oxidizing the C1 of sugar moieties, creating new binding sites for TrCel6A. TrCel6A also synergized with TaAA9A or LsAA9A as their DS values were 2.0 and 1.8 AU on PASC, and 1.5 and 1.3 AU on Avicel ^[36]. The CBHI and TaAA9A or LsAA9A displayed a DS of 2.25 and 1.4 AU on PASC or 1.8 and 1.5 AU, respectively ^[36]. EG (Cel7B) and MtLPMO9A synergy improved hydrolysis of Avicel, bacterial cellulose and sugarcane bagasse (SCB), which resulted in higher glucose production in the presence of Anβ-gl. The DS values between cellulases (TrCel7A, TrCel6A and TrCel7B) and MtLPMO9A during the degradation of Avicel, BC and SCB were 2.8, 2.5 and 2.6 AU, respectively. WeIt is proposed that DS values greater than 1.5 AU should be considered as a threshold which result in higher yields of soluble sugars during synergism between the cellulases and LPMOs from AA9 (LPMO9A) during the degradation of amorphous substrates such as PASC. However, DS values greater than 2 AU between cellulases and LPMO9A should be considered as a threshold which results in the improved degradation of crystalline substrates like Avicel.

5. Conclusion

Cellulolytic enzyme synergism is essential for the deconstruction or modification of cellulose in the bio-refinery sector. Hence, research on synergistic interactions has attracted extensive interest from all over the world over the past few decades. Cellulases are the main enzymes that degrade or modify cellulose substrates; however, expansins, SWO and LPMO demonstrate different modes of degradation action on the cellulose substrates. The LPMOs and cellulases or expansins/swollenins and cellulases synergism results in an enhanced conversion rate of the cellulose substrates to soluble sugars, aldoronic and gem-diol acids. Interestingly, to date, there have been no studies that have investigated the synergy between expansins/swollenins and cellulolytic LPMOs. This suggests that knowledge gaps exist in how expansins or swollenins and LPMOs (AA9, AA10 and AA16) interact during biomass deconstruction.  Perhaps the enzyme cocktail(s) containing the expansins or swollenins, cellulases and LPMOs could be superior to the currently available commercial cellulolytic enzyme cocktails.  However, it is vital to understand different forms of synergy between the enzymes before formulating enzyme cocktails containing these proteins. Lastly, the theorized DS value thresholds can be applied to understand the level of synergism between cellulolytic enzymes during the formulation of enzyme cocktails.