Cellulose is the most abundant polymer in lignocellulosic biomass, and typically constitutes 40%–60% of its dry weight (Sharma et al., 2017). Cellulose is predominantly present in the cellular walls and is composed of β-D-glucopyranose units linked via β-(1,4) glycosidic bonds. Each glucopyranosyl unit of cellulose contains three hydroxyl functional groups, which impart hydrophilic properties and reactivity to various agents. Long, unbranched chains of several hundred cellulose units form structures termed nanofibrils. Parallel chains of nanofibrils are tightly bound to each via intermolecular hydrogen bonding to form microfibrils (Robak and Balcerek, 2018). In bulk, these microfibrils constitute the basis of cellular walls. A hierarchical chart depicting the relative length scales of these constituent structures is shown in Figure 2. Within each fibrils there can be intervaled sections of amorphous cellulose structures (Ling et al., 2018). The extent in which cellulose within a material forms crystalline structures vs amorphous structures is known as the index of crystallinity, and is well known to affect the rate of enzymatic hydrolysis by cellulases (Yoshida et al., 2008). A significant negative correlation between enzymatic hydrolysis performances of various biomass sources and their crystallinity has been reported among literature as more crystalline regions have stronger hydrogen bonding networks that make them less accessible to enzymes (Li et al., 2014). Ionic liquids have been found to more easily penetrate biomass in less crystalline regions, likely because of the reduced energy to break inter-structural bonding (Ling et al., 2017). Cellulose chains with higher degrees of polymerization (DP) constrain more hydrogen bonds that increase the difficulty of enzymatic hydrolysis. Conversely, shorter chains of cellulose have weaker hydrogen-bond systems making these chains more accessible for enzymes (Hallac and Ragauskas, 2011; Meng et al., 2017). Naturally, cellulose fibrils tend to become embedded into a lignocellulosic matrix that provides additional barriers for enzymatic hydrolysis, as this matrix is structurally diverse and requires several different synergistic enzyme systems or high energy inputs to decompose (Zoghlami and Paës, 2019). The proportion of cellulose that depolymerizes into glucose is a key metric to determine the effectiveness of a deconstruction process.

Hemicelluloses are heterogenous biopolymers that contain various monosaccharide pentose and hexose subunits, most notably, xylose. A section of an exemplar xylan polymer is shown in Figure 3 along with common constituent hexose and pentose subunits. Hemicellulose usually constitutes 20%–35% of total dry weight of biomass, and tends to have the highest fractional composition in wood substrates compared to other biomass (Chandel et al., 2018). In general hemicelluloses have much lower degrees of polymerization than cellulose (Rodrigues Mota et al., 2018). The vast majority of hemicellulose is amorphous and forms mechanically weak structures. These weak bonding systems allow hemicelluloses to be readily hydrolyzed by dilute acids or bases, as well as hemicellulase enzymes (Isikgor and Becer, 2015). Hemicelluloses cross-link with cellulose fibrils and lignin via ferulic acid residues and in doing so create further barriers limiting the accessibility of hydrolytic enzymes inside of the lignocellulosic matrix (Balakshin et al., 2011). Often the addition of acids or some species of ILs can fragment a significant portion of hemicellulose polymers without the addition of specialized enzymes. The proportion of hemicellulose that decomposes into monomeric sugars is also a key metric to determine the effectiveness of a deconstruction process.

Lignin is the second most abundant lignocellulosic polymer, constituting roughly 15%–40% of the dry weight of biomass (Ragauskas et al., 2014). It forms an amorphous structure out of phenylpropanoid building units including p-coumaryl, coniferyl, and sinapyl alcohols, which are shown in Figure 4 (Agbor et al., 2011). Lignin is capable of binding hemicelluloses to cellulose in the cell wall and in doing so adds rigidity to the structure of cell walls. The high degree of heterogeneity of lignin structures typically makes it difficult for targeted enzymatic hydrolysis although the extent that lignin resists decay varies with species and tissue type. For example, lignin polymers with fewer aryl-aryl bonds such as syringyl (S) lignol are found to be more susceptible to decay vs guaiacyl (G) units, and thus the S/G ratio is a common metric of interest when determining the contents of biomass (Vane et al., 2006). Overall, the presence of lignin is well known to inhibit the conversion of cellulose as it is prevents enzyme access to cellulose, can non-specifically bind to cellulases, and is highly recalcitrant to decomposition itself, so it is often in the interest of biofuel processors to remove lignin from biomass prior to introducing enzymes (Santos et al., 2012).

Extractives are the nonstructural components of lignocellulosic matter and include fats, proteins, phenolics, resins, waxes, along with many other trace organic compounds (Pecha and Garcia-Perez, 2020). Inorganic trace metals are known to decrease the yield of sugars from cellulose and change the properties of lignin products (Pecha and Garcia-Perez, 2020). Typically extractives can be removed by solvents, water or weak acids. They are also typically inhibitors of biofuel processing but can be used to generate valuable byproducts such as terpenes, proteins, triacylglycerides, and fatty acids including stearic acid, and linoleic acid.

Cellulases serve to break down cellulose molecules into monosaccharides and shorter oligosaccharides. Specifically, cellulases hydrolyze 1,4-beta-D-glycosidic linkages in cellulose and some hemicelluloses. The hydrolysis of cellulose is typically less thermodynamically favorable than that of starches because of the strong hydrogen bonding systems between cellulose units. There are several broad categories of cellulases-each differing structurally and mechanistically. A combination of these cellulases working synergistically can theoretically fully reduce cellulose into monosaccharides. Endoglucanases (EC 3.2.1.176) randomly cleave the β-1,4-glycosidic linkages of cellulose and xylan to create a reducing and non-reducing end (Chen and Wang, 2017). Reducing ends being characterized by the presence of aldehyde groups in place of a hydroxyl group. Exoglucanases (EC 3.2.1.91), which are also known as cellobiohydrolases and exocellulases, cleave off cellobiose units from the reducing or nonreducing end of a cellulose chain. β-glucosidases (EC 3.2.1.21), also known as cellobiases, hydrolyze the released cellobiose to glucose. Figure 6 shows a simplistic overview of the reduction of a cellulose polymer chain to glucose via these three enzymes. There are discrepancies in the specificity of specific enzymes in each category with regards to crystallinity, polymer length, and other ligno-carbohydrate interactions.

More recently, other cellulases such as oxidative cellulases, that use radical reactions to depolymerize cellulose, and cellulose phosphorylases that use phosphates rather than water as a reactant, have attracted interest in the academic community. These include (lytic) polysaccharide monooxygenases (PMOs/LPMOs) (EC 1.14.99.56) and cellobiose dehydrogenases (CDHs) (EC 1.1.99.18), which use oxidative cleaving to break apart cellulose. PMOs/LPMOs are enzymes that depend on copper ions as metal cofactors to oxidize the carbon-hydrogen bonds of polysaccharides (Hemsworth et al., 2014; Hedegård and Ryde, 2018). The enzyme reacts with a hydrocarbon substrate by employing O₂ as a co-substrate and accepting two protons and two electrons. The copper ion aids in accepting electrons and oxygen atoms within the mechanism to ultimately produce a water molecule and an alcohol group attached to the previous hydrocarbon. The state of the copper ion is dependent on the stage of the mechanism. PMOs have varying degrees of substrate specificity which are dependent on their primary structure (Dimarogona et al., 2012). They are useful when using crystalline cellulose. The crystalline cellulose can help to produce new cellulose chain ends that can then be accessed by other enzymes. In non-crystalline cellulose, the endoglucanases can effectively produce enough new chain ends for the exo-glucanases.

Hemicellulases are a broad category of enzymes that participate in the breakdown and hydrolysis of galactans, xylans, mannans, and arabinans. Some of the most important hemicellulase categories are xylanase (EC 3.2.1.8) which hydrolyzes β-d xylano pyranosyl linkages of xylan to form xylo-oligosaccharides, β-mannanase (EC 3.2.1.78), arabinofuranosidase (EC 3.2.1.55), and β-xylosidases (EC 3.2.1.37) which also known as xylobiase, that catalyzes the hydrolysis of xylo-oligosaccharides into D-xylose sugars, although this is by no means an exhaustive list (Zhang et al., 2019).

Lignin modifying enzymes (LMEs) are a class of enzymes that catalyze lignin polymer deconstruction. Unlike most cellulases and hemicellulases, the most well-studied LMEs operate using oxidative processes that involve the formation of radical species rather than hydrolytic mechanisms. However, alternative reduction pathways in LMEs do exist (Lee et al., 2019; Chan et al., 2020). Radical chain reactions can cleave ether and carbon-carbon bonds-fracturing lignin as well as breaking down aromatic rings (Peng et al., 2002; Lange et al., 2013). A number of microorganisms, mainly white rot fungi (e.g., Trametes, Phanerochaete) and bacteria (e.g., Streptomycete, Pseudomonas, Rhodococcus) naturally produce enzymes capable of degrading and/or modifying lignin. Since lignin forms complex heterogeneous macromolecular structures, it is highly recalcitrant to degradation. In particular, its heterologous nature and large combination of structural bonds makes targeted reduction pathways ineffective. Among the most prominent categories of LMEs are laccases (EC 1.10.3.2), which can both polymerize and depolymerize lignin compounds via C-C cleavage, C oxidation, and alkyl-aryl cleavage, lignin peroxidases (EC 1.11.1.14), which targets the non-phenolic components of lignin (which can constitute up to 90% of lignin), manganese peroxidase (EC 1.11.1.13) which specifically targets phenolic components of lignin by catalyzing the reaction 2 Mn(II) + 2 H⁺ + H₂O₂ = 2 Mn(III) + 2 H₂O, and versatile peroxidase (EC 1.11.1.16) which, as their name suggests, can target both phenolic and non-phenolic components of lignin (Youn et al., 1995; Martínez et al., 2005; Falade et al., 2017).

The cation and anion structural components influence the properties of ILs, which can have significant impacts on the activity and stability of enzymes when utilized as a solvent medium. Several factors instigate different reactions from the enzymes including but not limited to structural dynamics, hydrophilicity/hydrophobicity, alkyl chain length of the cation, the hydrogen bond forming capabilities of the anion, rheological properties, co-solvents, and experimental conditions. With numerous variations in experimental, pretreatment conditions, and enzyme origin, it may be difficult to draw linear relationships between ionic liquid concentration and enzyme activity. Furthermore, trends in enzyme activity are investigated by classifying ionic liquids that share similar physicochemical properties into comparable groups where their deconstruction metrics are discussed below.

Changes in temperature can have a direct impact on kinetic energy, which can alter the activation energy and affect the likelihood of molecular collisions occurring—both factors that can be engineered to improve the likelihood of enzyme binding (Peterson et al., 2007). However, enzymes are also susceptible to structural changes induced by variations in temperature (Peterson et al., 2007). These effects occur regardless of whether ILs are introduced (Brogan et al., 2018). Heating enzymes to a sufficiently high temperature can permanently unfold enzymes and conversely, at too low of temperatures, the likelihood of enzyme-substrate interactions with sufficient activation energy for catalysis can be low enough so that virtually no activity occurs. This range of temperatures wherein the enzyme is active is unique to each enzyme. Improvements in enzyme activity and stability from changes in temperature compete with deconstructive effects of ionic liquids on the biomass substrate. Therefore, the reaction temperature needs to be adequately optimized to balance these effects and produce the desired enzymatic activity and product yields.

One important method concerning biofuel production is the temperature at which enzymes of interest have the highest activity. Most procedures to measure the optimal temperature and pH of an enzyme occur before adding ILs-effectively assuming that these parameters are decoupled from IL concentration. Kodou et al. found that within the range of 0–500 mM of added [BMIM][OAc], the optimal temperature of β-glucosidase from Themotoga maritima remained constant, which substantiates this assumption for small concentrations of ILs (Kudou et al., 2014). However, results from Campen et al., which includes data at much higher IL concentrations, shows a small negative correlation of IL concentration on optimal activity temperature of A. niger A5IL97 between 0% and 30% (v/v) [EMIM][OAc] (Amaike Campen et al., 2017). Finding an optimal temperature becomes less straightforward in multi-enzyme systems where each enzyme can have different temperature-reaction rate dependencies.

Sufficient evidence does exist to suggest that the presence of even small amounts of ILs can substantially modify the unfolding temperature of different cellulases (Turner et al., 2003). For the majority of studied IL-enzyme pairs, the addition of ILs decreases the unfolding temperature of the enzyme. Table 6 shows several examples of cellulases across multiple studies that directly measured the effects of increasing temperature of denaturing enzymes in IL solutions. A clear negative correlation (.25°C–1°C/% concentration (v/v) IL) between the denaturing temperature and IL concentration is observed for the listed enzymes.

Notable counterexamples do exist. In one study, Bose et al. found that GC 220 cellulase from Tricoderma reesei was stabilized in solutions containing tris-(2-hydroxyethyl)methylammonium methylsulfate (HEMA) at temperatures close to 100°C, whereas it denatured in solutions without HEMA at temperatures approximately 50°C (Bose et al., 2010). Brogan et al. found that β-Glucosidase with surface modifications could be stabilized in 1-butyl-1-methylpyrrolidinium ([BMPL]) based ionic liquids ([BMPL][OAc] [BMPL][OTf] [BMPL][NTf₂]) to temperatures up exceeding 130°C- a temperature nearly 80°C warmer than in aqueous media alone (Brogan et al., 2018). In doing so, the enzymatic capability of glucosidase, which is indispensable in biomass processing, increased by over 30-fold.

While seemingly promising, HEMA has shown relatively low potential for dissolving cellulose and its high viscosity reduces diffusion of enzymes to substrate making it a poor candidate for practical biomass systems by itself (Bose et al., 2012). Combined with another IL with a better ability to solubilize biomass components, a multi-IL cocktail might be able to best utilize the stabilization properties of HEMA (or another IL with analogous stabilization properties) while being able to effectively dissolve and degrade cellulose and hemicellulose. Indeed, Bose et al. found that the combination of HEMA and [BMIM][Cl] enhanced the process of glucose production from cellulose (Bose et al., 2012). The exact nature of the effect of ILs on temperature related parameters is variable on the specific IL-enzyme system. For instance, HEMA proved to be detrimental for other enzymes, such as penicillin G amidase (Cantone et al., 2007), and Cal B (Madeira Lau et al., 2004), but for future experimentation, beneficial properties of multicomponent systems might be able to be exploited to create more cost effective biorefinery production processes.

Temperature also has an effect on swelling and dissolution rates of lignocellulose in ILs (Sun et al., 2009). In part this is attributed to the destabilizing effects of higher temperatures on hydrogen bonds in the 3-D structure of cellulose (Zavrel et al., 2009). Swelling can fragment cellulose fibers as well as free cellulose fibers from the lignocellulosic matrix via applied inter and intra structural stresses. As solvating molecules have increased accessibility within the fiber structure after fragmentation they are able to better approach cellulose chains and dissolve them into solution (Cuissinat et al., 2008). Dissolved cellulose is in turn easier for enzymes to access, allowing for increased enzymatic activity. Wang et al. found that in wood chips treated with [AMIM][Cl], swelling did not begin until 70°C and cellulose did not begin being substantially dissolved until 100°C (Wang et al., 2011a). These temperatures are above temperatures that some enzymes are active, potentially limiting the practical application of these temperature induced effects. Hemicellulose is found to dissolve at lower temperatures in ionic liquids than cellulose. Labbé et al. reported that at temperatures 60°C–80°C [EMIM][OAc] is capable of cleaving the acetyl groups covalently attached to hemicellulose components in poplar wood (Labbé et al., 2012). This result suggests that at these more moderate temperatures ILs could more effectively disrupt the carbohydrate–lignin linkages and in turn, release hemicellulose from the lignocellulosic matrix and make it more accessible to xylanases. Another interesting effect of increasing temperature during the IL pretreatment process is that lignin polymers are susceptible to self-condensation reactions. Self-condensation results in increased molecular weight of lignin which can precipitate lignin. By disabling precipitated lignin from reentering lignocellulosic structures, it decreases the heterogeneity of the structure, and limits the recalcitrant effects that lignin usually applies (El Hage et al., 2010). However, these effects are observed at temperatures usually >150°C- well beyond the unfolding temperature of known lignocellulose degrading enzymes.

Some physicochemical properties of ILs such as viscosity are temperature dependent and have the ability to affect the reaction kinetics by causing changes in mass transfer and diffusivity, which plays a role in both the activity and stability of enzymes (Naushad et al., 2012). By extension, one might expect that viscosity of the IL solvent may play a deterministic role in process efficiency and cost. Indeed, some authors, such as Bose et al. attributed lower observed activities of cellulase in HEMA to the solvent’s high viscosity (Bose et al., 2010). Other authors, such as Zhao et al., investigated over 25 ILs and the relation between their viscosity and enzymatic reaction and suggested that IL viscosity played only a minor factor in enzymatic reaction rates (Zhao et al., 2009). For instance, some highly viscous ILs such as quaternary ammonium chloride, which has a viscosity of 512mPas at 25°C (roughly 500 times more viscous than water), did not have significantly lower reaction rates than other tested ILs. It should also be noted that ILs often appear at low concentrations mixed with water or organic solvents. These results, although by no means exhaustive, suggest that temperature dependent effects on IL viscosity do not play a completely deterministic role on enzyme function.

The acidity of the medium containing an enzyme not only affects the activity and shape of the enzyme, but also the shape and charge of the substrate. Specifically, pH can affect the ionization state of both acidic and alkaline amino acids that are present in the enzyme. This includes carboxyl functional groups on the side chains of acidic amino acids and amine-containing functional groups of basic amino acids. Changes in the chemical state of these amino acids can substantially alter the 3-dimensional geometry of an enzyme, thus changing the shape of the active site to be more or less effective at binding with a substrate of interest, or altogether inactivating the enzyme. Additionally, for the enzymatic reaction to occur, these functional groups often must be in a particular protonation state (Koshland, 1953). Typically, there is a range of pH at which enzymatic activity with a particular substrate is optimal. However, broad evidence suggests that this optimal pH range is a function of other factors such as temperature, salinity, IL concentration, and is highly dependent on enzymes and substrates in the system. It should be noted that ILs can be mildly acidic or basic and are able to alter the acidity of their local environment. The acidity of non-aqueous systems can be measured in many different ways, but perhaps the most well-established of these is the Hammett method which uses a range of closely related UV-vis probes to generate the Hammett acidity (Gräsvik et al., 2014). In practice, most ILs are mixed with water whereby more conventional methods of measuring their effects on pH can be utilized.

Differences in pH caused by the addition of ILs may have several different effects on enzymes. Most cellulases and some xylanases are glycoside hydrolases (GH) which function using acid-base mechanisms. Most GHs use a two-step double displacement mechanism whereby acid/base residue on a carbohydrate and a nucleophilic residue on the enzyme are produced (Knott et al., 2014). The first step requires proton transfer from the acid base residue. This mechanism is illustrated in Figure 9. Changes in pH caused by addition of ILs should in theory have some effect on this step by affecting the availability of protons, although their availability competes with the structural changes they induce.

The pH of a pretreatment process can also affect the way in which a substrate decomposes. In more acidic pretreatment environments, cellulose polymers have a tendency to stay intact, while hemicellulose hydrolyzes into monomeric sugars can form biotoxic compounds such as furfurals, HMF, and levulinic acid that inhibit downstream fermentation processes (Carvalheiro et al., 2008; Binder and Raines, 2010). At more neutral pHs, there tends to only be partial hydrolysis of the hemicelluloses. Often this is a result of autohydrolysis from organic acids within the lignocellulosic material. Since these conditions are not acidic enough, hemicelluloses tend to stay in polymeric or oligomeric form. In this regime, solubility of hemicelluloses is proportional to the concentration of organic acids available in biomass-meaning feedstocks such as softwoods typically have low solubilisation of hemicellulose. In more alkaline conditions lignin polymers tend to fractionalize and dissolve in solution, while hemicelluloses tend to remain in solid states. It may be of potential interest to factor inhibitory effects of biotoxic compound production during pretreatment on fermentation yields and cost.

Recent work by Pham et al. has reported that pH affects the ratio of products from the decomposition of lignin by lignin peroxidases isozyme H8 (Pham et al., 2021). This author suggests that pH is a key factor for efficient and selective depolymerization of lignin dimers (specifically guaiacyl glycerol beta-guaiacyl ether) and that engineering processes by altering pH could lead to improvements in bioprocessing in the future. How valuable this effect is for industrial processes is still to be determined.

Varying origin or expression methods of the enzymes are also factors in potential IL-tolerance. The expression system of cell cultures has had an impact on enzyme quality, functionality, production speed, as well as yield. Natively expressed enzymes, enzymes produced in the organism of its origin, are often limited by demanding growth conditions to replicate the organisms’ native environment. For instance, fungal organisms are commonly cultured for cellulase production, but are limited by their need for special culturing and induction conditions. Additionally, enzyme yields are often limited and non-tailorable to different substrates due to substrate specificity of their metabolism. Heterologously produced enzymes currently suffer from their own drawbacks including low yields, poor secretion, and high costs (Lambertz et al., 2014). Specific modifications have been made to overcome these limitations in order to lower production costs and optimize expression systems, which differ among research labs and protocols.

At present a standard procedure across literature to quantify lignocellulosic enzyme activity in ionic liquids has not been fully realized. This comes despite the Commission on Biotechnology of Pure and Applied chemistry proposed a number of standard procedures for measuring cellulase activity in 1984 and the National Renewable Energy Laboratory (NREL) developing standardized procedures for different measurements in biomass processing in the early 2000s (Dowe, 2001; Dashtban et al., 2010). Variability in factors such as enzyme type, conditions in which they are active, substrate of interest, desired product, access to equipment, improvement of sensor technology and computational abilities, and even the deconstruction process itself, have all engendered the need to conduct alternative methods of analyzing deconstruction metrics (Fapyane and Ferapontova, 2017; Kwon et al., 2018).

Some studies opt to use “real” substrates, as opposed to model compounds for processing. In this case, real substrates are naturally occurring samples of organic compounds that are taken from biomass. Model compounds by contrast are typically synthetically produced chemicals or polymers of high purity. Oftentimes carboxymethyl cellulose (CMC), p-nitrophenol based compounds, or varieties of highly crystalline cellulose such as Avicel^® are used to measure cellulase activity (Dashtban et al., 2010). Model compounds provide a more chemically simple system that can be specific to a single type of enzyme and give better insight on individual enzyme reaction pathway kinetics (Dashtban et al., 2010). Model compounds also lack impurities such as lignin that would typically be present at different concentrations in different biomass sources and create variations in enzyme activity across substrates (Ladeira Ázar et al., 2020). However, their use does not come without valid concerns. For one, model compounds do not chemically represent real substrates, and their reaction kinetics are not always directly comparable (Dashtban et al., 2011; Monschein et al., 2013). Additionally, the results from some model compounds have been reported to be notoriously non-reproducible for a variety of reasons specific to each compound (Percival Zhang et al., 2006). Despite their drawbacks, they remain popular options as measuring individual reaction pathways in enzyme assays from more complex substrates has proven difficult.

Due to the complexity of cellulose–cellulase systems and differences between kinetic characteristics of hydrolysis reaction and time, cellulase activity assays are often expressed by either their initial hydrolysis rate or by their end point hydrolysis yield. Initial hydrolysis rate measurements are preferred when measuring individual cellulase activity over a short time, whereas hydrolysis yields are used to measure total enzyme activity over a longer duration (Wu et al., 2006). Initial hydrolysis reactions are relatively free from complications such as back-reaction and enzyme degradation and product inhibition. Specific to cellulose-cellulase systems it should be noted that the rate of enzymatic hydrolysis of amorphous regions of cellulose are significantly higher than crystalline regions (Hall et al., 2010). Initial hydrolysis reaction rates are measured when the bulk of amorphous cellulose is still present and before their availability significantly alters enzyme velocity. End point measurements on the other hand account for enzyme deactivation caused by ILs, and feedback inhibition that occurs temporally. Reactions will significantly slow down by the end of the experiment as amorphous regions become scarcer, and cellulase activity becomes primarily a function of crystalline cellulose hydrolysis. It is sometimes indeterminate to what extent IL based deactivation occurs relative to feedback inhibition, so half-life studies in ILs are also conducted and first order deactivation models are applied (Schindl et al., 2019). As such, enzyme studies utilizing different assay methods are not always easily comparable and care must be taken when building datasets from multiple papers.