Comparison of three deep eutectic solvents and 1-ethyl-3-methylimidazolium acetate in the pretreatment of lignocellulose: effect on enzyme stability, lignocellulose digestibility and one-pot hydrolysis

Abstract

Certain ionic liquids (ILs) are well-known pretreatment chemicals for lignocellulosic substrates prior to enzymatic total hydrolysis. Deep eutectic solvents (DESs) are closely related to ILs in many properties, but are easier and on occasion cheaper to synthesize and have been claimed to be less inactivating to enzymes used in the hydrolysis, and less toxic for the environment and to micro-organisms used in fermentation. The use of DESs as lignocellulose pretreatment chemicals has not been studied to a similar extent as the use of ILs. In this study, the stability of three Trichoderma reesei cellulases (the endoglucanases Cel5A and Cel7B and the cellobiohydrolase Cel7A) and one T. reesei xylanase (Xyn11) was compared in concentrated solutions (85% w/w) of three DESs (choline chloride : boric acid in molar ratio 5 : 2, choline chloride : glycerol 1 : 1 and betaine : glycerol 1 : 1) and 1-ethyl-3-methylimidazolium acetate ([EMIM]AcO), a powerful lignocellulose-dissolving IL. The pretreatment efficiency of these chemicals was further compared in a mild pretreatment (90% w/w DES or [EMIM]AcO, 80 °C, 24 h, 5% (w/w) lignocellulose consistency) of four different substrates; microcrystalline cellulose, eucalyptus dissolving pulp, shredded wheat straw and spruce saw dust. After pretreatment, the enzymatic digestibility of the pretreated substrates was evaluated in the enzymatic total hydrolysis in three different setups, including hydrolysis of the washed pretreated substrates in buffer, and of the pretreated substrates in solutions containing 30% (w/w) and 80% (w/w) of DES or [EMIM]AcO. The stability analysis identified glycerol-containing DESs to be highly stabilizing for the cellulases, but their pretreatment efficiency was limited. [EMIM]AcO had a high pretreatment efficiency, but was highly inactivating for the used cellulases. The presence of DES or [EMIM]AcO led in all cases to decreased enzymatic hydrolysis yields. Thus, good enzymatic stability in a certain DES does not directly implicate good performance in the hydrolysis of solid lignocellulosic substrates in that DES.

---

Introduction

The increased global oil demand and decreasing oil reserves, combined with feedstock insecurity issues, price fluctuations and the need to reduce the emissions of fossil CO₂, act as strong drivers to develop alternative non-fossil routes for fuel and chemical production. In 1st generation biofuel production, ethanol and biodiesel produced from nutritionally important plant-based sources, such as corn, wheat or rapeseed oil, compete with their use as food. With a growing population and increasing food demand, the use of food stuff as a fuel and chemical source is not ethically acceptable. Lignocellulosic biomass, constituted from cellulose, hemicelluloses and lignin, provides an alternative and sustainable feedstock for the production of renewable biofuels and chemicals. Lignocellulose is available in abundance around the World and is currently not utilized to its full potential. When producing fuels and chemicals via the sugar route, the polysaccharides in the feedstock are hydrolysed to monosaccharides, which are further fermented biotechnically to the desired product. With starch as raw material, the hydrolysis of the polysaccharides is straightforward, but when using lignocellulose, a pretreatment step is needed to make the polysaccharides more accessible for hydrolysis.¹ The enzymatic total hydrolysis of plant cell wall polysaccharides is currently one of the major bottlenecks in designing techno-economically feasible industrial processes for the production of 2nd generation biofuels and other key chemical components from lignocellulose via the sugar route. The recalcitrance of lignocellulose to enzymatic hydrolysis is due to factors such as substrate insolubility, the shielding effect of lignin and hemicellulose, cellulose crystallinity, low porosity and generally low accessibility.^2–5 Different pretreatment techniques have been developed to increase the lignocellulose digestibility, of which steam explosion and hydrothermal treatments are the most commonly used in current demonstration and commercial scale plants.⁶

The use of certain ionic liquids (ILs), and more recently, deep eutectic solvents (DESs), has arisen as a new technology for the pretreatment of lignocellulose.^1,7 ILs are salts having low melting temperatures (<100 °C) and some classes of ILs have been shown to dissolve cellulose and lignocellulose, e.g. wood.^8,9 DESs are closely related to ILs but differ chemically from ILs, which are purely salts, by being composed of a hydrogen bond acceptor (often a salt) and a hydrogen bond donor, and DESs can also be prepared in different molar ratios of the components.¹⁰ The use of DESs as alternative solvents to ILs can be much acknowledged to the work of Abbott et al.,¹¹ though DESs have not, until recently, been subject of much study in biomass processing. DESs have been pointed out to generally possess some advantages over conventional ILs, including cheaper and potentially renewable starting materials, easier preparation and purification, lower general toxicity and tentatively a high biocompatibility (with enzymes and living cells).^12,13 This comparison is admittedly somewhat generalising, as e.g. cellulose-dissolving acid–base pair ILs have recently been reported which can be prepared by a simple neutralisation reaction.¹⁴ In a strictly scientific sense, the term “deep eutectic solvent” should be reserved for mixtures at the eutectic molar ratio of the components, at which the maximum freezing point depression is observed. However, in recent literature the DES term has been used in a rather inclusive manner, often for different low-melting mixtures of DES components in some other molar ratio than the actual eutectic mixture ratio. The eutectic molar ratio is generally not emphasised when DESs are introduced in many articles, and DESs have been presented e.g. as “mixtures of hydrogen bond donors with simple halide salts which produce liquids which have physical and solvent properties that are comparable with ionic liquids”,¹⁵ or as mixtures of a solid hydrogen-bond donor with a solid hydrogen-bond acceptor which, when mixed together, are capable of autoassociation to form a liquid phase characterized by a very large depression of the freezing point¹⁶ or even simply as mixtures of “a solid salt with a hydrogen-bond donor in different proportions”.¹² In this article we have chosen to use the term DES in the more inclusive manner, although we acknowledge that not all DES mixtures discussed in this article, either those in our own results or in the literature references, are necessarily true DESs by being composed of the eutectic molar ratio mixture of the components.

IL pretreatment has been shown effective for increasing the digestibility of cellulose¹⁷ and lignocellulose samples,¹⁸ but the low compatibility of biomass-dissolving ILs and hydrolytic enzymes (cellulases and hemicellulases) has prohibited the integrated use of ILs and enzymes in hydrolysis in one-pot processes.¹⁹ One-pot hydrolysis of cellulose in IL solution originating from the pretreatment step was originally proposed by Kamiya et al. in 2008,²⁰ and this concept would allow a more competitive water economy²¹ and fewer process unit operations than the traditional two-step hydrolysis in which the IL is removed between the pretreatment and hydrolysis.¹⁹ The washing step between IL pretreatment and enzymatic hydrolysis suffers from a very low water economy.²¹ In a technoeconomic evaluation, the one-pot hydrolysis was shown competitive with the separate IL pretreatment and hydrolysis with intermediary IL removal, provided that the feedstock loading is at a high 50% (w/w) level, and a suitable glucose separation method is available.²¹ Recently, a one-pot process using amino acids salts of choline was shown to be a very competitive alternative to the earlier ones using imidazolium-based ILs.²² The one-pot process is on a more experimental stage than the processes with IL separation and washing between the pretreatment and enzymatic hydrolysis, and further studies are needed to assess the compatibility of microbial fermentation with ILs or DESs, and to evaluate IL and DES recycling from the hydrolysate or fermented liquid. The current high price of ILs has also limited their use in industrial biomass processing.²¹ ILs have been observed to increase lignocellulose digestibility through several effects: removal of lignin and hemicellulose, partial hydrolysis of the polysaccharides, decreasing the crystallinity of cellulose and disrupting the lignin–carbohydrate complexes in the matrix.¹⁹ To date, DESs have not been reported to dissolve cellulose in considerable amounts analogously to cellulose-dissolving ILs, but lignin and starch have been found soluble in a number of DESs.²³

Although DESs have been studied to a comparatively lesser extent than ILs in lignocellulose pretreatment, they appear to be rather efficient in this use. Xia et al. found the pretreatment of microcrystalline cellulose (MCC) in neat and 1 and 2 M aqueous solution of a set of ILs and DESs to increase MCC digestibility, with choline chloride ([Chol]Cl) : glycerol (Gly) (1 : 2) being the most efficient DESs.²⁴ Procentese et al. pretreated corn cobs in three different DESs at different temperatures and found some of the DESs to efficiently remove lignin and some hemicellulose, depending on the treatment severity.²⁵ Cellulose was not considerably dissolved nor was its crystallinity changed, but the treatments were observed to increase the polysaccharide digestibility. Kumar et al. reported delignification of rice straw using lactic acid : betaine and lactic acid : [Chol]Cl DESs.²⁶ The subsequent saccharification efficiency was to not as good as with the used alkali and acid reference pretreatments, which was explained by the presence of residual DES in the substrate, as the used DESs were reported to be rather inactivating to cellulases.

The inhibiting effect of cellulose-dissolving ILs on cellulases was already established in a report by Turner et al. in 2003.²⁷ Much effort has thereafter been put into establishing the reasons for IL-induced cellulase inactivation and finding new enzymes and ILs with increased enzyme-compatibility, as recently reviewed.¹⁹ The compatibility of DESs and hydrolytic enzymes has not been studied to a nearly as high degree as in the case of ILs, and it is not currently known whether the proposed mechanisms for IL-induced enzyme inactivation and inhibition are applicable on DESs as such. Gunny et al. recently compared the inactivation of a cellulase mixture from Aspergillus niger in aqueous [Chol]Cl DESs with ethylene glycol (EG), Gly or malonic acid as hydrogen bond donors and found [Chol]Cl : Gly and [Chol]Cl : EG to have only minor inactivating effects on the cellulase mixture at up to 30% DES concentrations.²⁸ Lehmann et al. reported the DES [Chol]Cl : Gly (1 : 2) to suppress the activity of the cellulase CelA2, but several mutated variants had increased DES and IL tolerance.^7,29

To date, little data has been published about how DESs affect the stability of specific cellulase or hemicellulase monocomponent enzymes. In this article, we report the stability of four different glycosyl hydrolases from Trichoderma reesei, namely the endoglucanases Cel5A and Cel7B, the cellobiohydrolase Cel7A and the xylanase Xyn11, in three different concentrated (85% w/w) DES solutions: [Chol]Cl : boric acid (BA) (5 : 2), [Chol]Cl : Gly (1 : 1) and betaine (Bet) : Gly (1 : 1). For comparison, inactivation data in buffer and in 85% (w/w) of the cellulose-dissolving IL 1-ethyl-3-methylimidazolium acetate ([EMIM]AcO) is included. The DESs were chosen based on the previously reported stabilizing effect of glycerol to enzymes,³⁰ and the [Chol]Cl : BA (5 : 2) DES was chosen as an interesting DES for lignocellulose pretreatment based on a recent patent application, in which this type of DES was described to at least partially dissolve cellulose.³¹ [EMIM]AcO was chosen as reference treatment because of its powerful cellulose-dissolving capacity. Further, the effect of a mild pretreatment in these solvents is reported for four different substrates: MCC and eucalyptus dissolving pulp as cellulosic model substrates, and shredded wheat straw and spruce saw dust as more realistic lignocellulose substrates. The effect of pretreatment is investigated by following the change in substrate composition, digestibility and appearance under light microscopy. The pretreated and washed lignocellulose samples were subjected to enzymatic hydrolysis in buffer, as well as in the presence of the DESs or [EMIM]AcO from the pretreatment step in a one-pot hydrolysis procedure.

Experimental

Materials

Microcrystalline cellulose (MCC, research grade, particle size 0.020 mm) was acquired from Serva Electrophoresis GmbH (Heidelberg, Germany) and prehydrolysis kraft dissolving grade pulp from Eucalyptus urograndis from Specialty Cellulose (Brazil). The pulp was milled with a Fritsch Pulverisette 14 variable speed rotor mill to a final size of <1 mm. Spruce saw dust with a coarseness of 2 mm was acquired from a Finnish saw mill and shredded wheat straw from a farm in Jokioinen, Finland. The dry weight of the (ligno)cellulose samples was determined as the average mass loss for three parallel samples by keeping them at 105 °C overnight. The carbohydrate content and composition, as well as the content of acid soluble and acid insoluble lignin in the lignocellulose samples, were determined according to the NREL procedure,³² but with the saccharides analysed by a HPLC method as described by Tenkanen et al.³³ The monosaccharides were calculated as polysaccharides using a correction factor of 0.9 for the hexoses and 0.88 for the pentoses. Water used in this work was of Milli-Q grade. The amount of acid-soluble lignin was determined based on the UV absorption of the acid hydrolysate measured at 210 and 280 nm as described by Goldschmid.³⁴ The ash and extractives content were not measured in this study.

The DESs were prepared by mixing the components together, heating the mixture to 80 °C and adding a small amount of water (10% of the total weight) to lower the melting temperature to allow for easy handling at the enzymatic incubation temperature of 50 °C in the stability measurements. All 90% (w/w) DESs formed clear solutions in the course of 60 min when heated to 80 °C. [EMIM]AcO (purity > 98%) was acquired from Ionic Liquid Technologies (Heilbronn, Germany) and used as such. Water activity of aqueous DES and [EMIM]AcO solutions was measured by an AquaLab CX-2 Water Activity Measurement device (Pullman, WA).

Trichoderma reesei endoglucanases Cel5A and Cel7B, and cellobiohydrolase Cel7A, were produced, isolated and purified at VTT as described by Suurnäkki et al.³⁵ and the xylanase Xyn11 according to Tenkanen et al.³⁶

Residual enzymatic activity measurements

The residual enzyme activity measurements were done by incubating the enzyme preparations in 85% (w/w) DES or [EMIM]AcO solutions (aqueous part added as 0.050 M citrate buffer, pH 5.0), or in citrate buffer, at 50 °C under gentle magnetic agitation. At set time points (every hour for the first 6 h, then with 24 h intervals) samples were withdrawn, diluted and frozen to −18 °C. The frozen samples were later melted, appropriately further diluted and their activity was measured as indicated below. For the activity assay, blank samples were also produced containing the DES or the IL diluted in the same manner as the actual samples, to correct for any background absorption caused by the DES or IL during spectrophotometric detection.

The endoglucanase activity was measured using a 1% (w/v) carboxymethylcellulose (CMC) solution as substrate in citrate buffer (0.050 M, pH 5.0). The assay temperature was 50 °C, time 10 min and the residual activity was determined by measuring the released reducing sugars using a 3,5-dinitrosalicylic acid (DNS) assay with spectrophotometric detection at 540 nm.³⁷ Glucose was used as standard. Xylanase activity was measured with a similar protocol but using beach xylan (Carl Roth, Karlsruhe, Germany) as substrate, a reaction time of 5 min and xylose as standard. CBH1 activity was measured using 4-methylumbelliferyl-β-D-cellobioside as substrate and 4-methylumbelliferone as standard at 50 °C for 30 min. The enzyme activity was determined by fluorescence spectrometry using an excitation wavelength of 355 nm and an emission wavelength of 460 nm, with a Wallac VICTOR² 1420 Multilabel HTS Counter (PerkinElmer). β-Glucosidase activity was measured according to Bailey and Linko.³⁸ The protein content of the pure monocomponent enzyme preparations was measured using the Bio-rad DC protein assay with bovine serum albumin (BSA) as standard and the purity of the preparations was verified by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE).³⁹ For SDS-PAGE, samples were run in ready-made Tris–HCl gradient gels (4–20%, Bio-Rad) and the visualization was done with a Criterion stain-free imaging system (Bio-Rad) according to the manufacturer's instruction.

Enzymatic hydrolysis of DES-pretreated samples

Preparation of washed pretreated lignocellulose samples. The MCC, milled dissolving pulp, spruce saw dust and shredded wheat straw were subjected to a mild DES, [EMIM]AcO or buffer (reference) pretreatment, in which the samples were gravitationally mixed in 90% (w/w) DES or IL, or 100% buffer for 24 h at 80 °C at a consistency of 5% (w/w). After the treatment, the pretreatment mixtures were cooled to room temperature, mixed with water for 30 min and thereafter washed by centrifuging and mixing with clean water until visually deemed free of any DES or IL. The washed pretreated samples were examined by light microscopy (Olympus BX61 microscope) and digital image recording was performed with the Soft Imaging Systems analySIS® 3.2 software. The polysaccharide and lignin contents of the pretreated samples were determined as described in the “Materials” section.

Enzymatic hydrolysis. For enzymatic hydrolysis an enzyme mixture was constituted from the four monocomponent T. reesei enzymes Cel5A, Cel7A, Cel7B and Xyn11. The weight proportion between the enzymes was Cel5A 15%, Cel7A 60%, Cel7B 15% and Xyn11 10%. In addition, an excess of β-glucosidase (Novozym 188, Novozymes) was supplemented to the mixture. The final enzyme dosage in the hydrolysis experiments was 15 mg of protein per g of dry substrate + 200 nkat of β-glucosidase per g of dry substrate.

The hydrolysis of untreated or in DES, [EMIM]AcO or buffer pretreated samples was conducted in 2% (w/w) solid consistency (by dry mass) at 45 °C for 48 h in 0.050 M citrate buffer (pH 5.0). First, the substrate was weighed into a sealable test tube, the defined amount of buffer was added and finally the mixture was heated to 45 °C in a water bath. The hydrolysis was initiated with the addition of the enzyme mixture, and the test tubes were sealed and left under magnetic agitation for 48 h. The weight of the hydrolysis mixture was 3 g with a substrate mass of 0.060 g (dry) and the hydrolysis was carried out in triplicate and with one reference sample without added enzyme. The hydrolysis was stopped by boiling the hydrolysis mixtures for 600 s. The test tubes were centrifuged (2500 rpm for 10 min) and the liquid fraction was collected for sugar quantification by a DNS assay. The hydrolysis yield was estimated by dividing the amount of formed reducing saccharides with the amount of saccharides theoretically available in each pretreated lignocellulose sample as determined by composition analysis.

Enzymatic one-pot hydrolysis of pretreated lignocellulose samples in the presence of 30 and 80% (w/w) DES or [EMIM]AcO solutions. Untreated substrate (0.060 g by dry mass, aiming at 2% w/w substrate loading in the enzymatic hydrolysis) was weighed into a sealable test tube and 1 or 2.67 g of 90% (w/w) DES or [EMIM]AcO solution, or buffer, was added (1 g aiming at hydrolysis in 30% w/w DES or IL and 2.67 g aiming at hydrolysis in 80% w/w DES or IL). The sample was mixed to make sure all the substrate was homogenized and wetted by the added solvent, the sample tube was sealed and then kept at 80 °C for 24 h. After pretreatment, the sample temperature was equilibrated to 45 °C in a water bath and the sample was further diluted with buffer to yield a hydrolysis mixture with either 30 or 80% (w/w) of DES or IL (total mass of liquid was 3 g in hydrolysis). The hydrolysis was started by the addition of enzyme mixture. The hydrolysis conditions, hydrolysis ending, work-up and analysis was done as described in the “Enzymatic hydrolysis” section. Extra care was taken with respect to the possible background caused by the presence of DES or IL in the hydrolysate samples in the DNS assay, by correcting the hydrolysate absorbance by subtracting the reference sample background obtained with the corresponding dilution factor. In some of the 80% (w/w) DES/[EMIM]AcO hydrolyses, the interaction between substrate and DES or IL led to gelling. The gel was broken by adding 3 g of water and vigorously mixing the hydrolysis mixture, after which the sample was centrifuged and the hydrolysate was collected and treated as described in the “Enzymatic hydrolysis” section.

Results and discussion

Stability of T. reesei cellulases and xylanases in concentrated DES solutions

The stability of the studied T. reesei glycosyl hydrolases, i.e. cellobiohydrolase I (Cel7A), the endoglucanases Cel5A and Cel7B and xylanase Xyn11 was studied in 85% (w/w) aqueous solutions of three different DESs, choline chloride ([Chol]Cl) : boric acid (BA) (5 : 2), [Chol]Cl : glycerol (Gly) (1 : 1) and betaine (Bet) : Gly (1 : 1). For comparison, the stability of the enzymes was also studied in citrate buffer (pH 5) and in 85% (w/w) [EMIM]AcO, an ionic liquid which is a well-known and powerful (ligno)cellulose solvent. The DESs were chosen because of the known stabilizing effect of glycerol to enzymes,³⁰ and the DESs composed of [Chol]Cl : BA were recently described to at least partially dissolve or gel cellulose.³¹ Showing that the enzymes retain stability in these solvent systems is necessary for any one-pot hydrolysis in DES solution to be possible, although also other than stability factors may limit the enzymatic hydrolysis of lignocellulosic samples. The rather high DES concentration (85% w/w) used in the experiments represents a pretreatment system with a minimum subsequent dilution by water. The 85% (w/w) DES solution is concentrated enough to allow for significant amounts of lignocellulose components to be dissolved in it, and thus poses an interesting medium for conducting enzymatic reactions on dissolved biopolymers. It was chosen to work with monocomponent enzymes instead of commercial cellulase cocktails in order to pinpoint the individual inactivation and inhibition effects of the DESs and [EMIM]AcO on the glycosyl hydrolase components.

The measured residual activities clearly show the inactivating effect of both 85% (w/w) DES and [EMIM]AcO solutions (Fig. 1). In buffer solution at pH 5.0, which usually is considered as ideal conditions especially in comparison to the ILs and DESs, all the enzymes lost their activity within 120 h at 50 °C, with the exception of the cellobiohydrolase Cel7A, which retained 52% of its initial activity after 120 h. On the contrary, the xylanase Xyn11 lost its activity completely in less than 24 h in buffer (the residual activity in buffer is superposed by the residual activity curve in 85%, w/w, Bet : Gly 1 : 1 in Fig. 1). For all the enzymes, except Xyn11, the IL [EMIM]AcO was a very inactivating environment, well in agreement with previous studies.⁴⁰ Xyn11 was the most stable in [EMIM]AcO, by far more stable than in buffer, and clearly shows that inactivation in ILs depends to a high degree on the studied enzyme. [Chol]Cl : BA (5 : 2) was the most inactivating of the DESs for all of the studied enzymes, although the endoglucanase Cel7B showed 10% residual activity in this DES solution after 24 h and it took approximately 24 h for cellobiohydrolase Cel7A to be completely inactivated in this DES. Recent results in two separate studies with lactic acid DESs and malonic acid DESs, respectively, also indicated strong inactivation of cellulases,^26,28 which suggests DESs containing acid components and low pHs to be inactivating to cellulases in general. [Chol]Cl : Gly (1 : 1) was to a medium degree inactivating, suggesting the glycerol component of this DES to have a stabilizing effect on the enzymes. This is expected, as glycerol is known to stabilize enzymes.³⁰ In contrast to the other solutions, Bet : Gly (1 : 1) appeared to stabilise the endoglucanases after an initial sharp decrease in activity. The stabilisation of the endoglucanases in 85% (w/w) Bet : Gly (1 : 1) showed, that not only the glycerol component was affecting enzyme stability, but also the hydrogen bond acceptor had a significant contribution. Betaine was clearly more enzyme-compatible than [Chol]Cl. These results suggest that Bet : Gly (1 : 1) is a promising DES for designing enzymatic processes in this medium.

Fig. 1 Residual activity of Trichoderma reesei endoglucanase Cel5A, cellobiohydrolase Cel7A, endoglucanase Cel7B and xylanase Xyn11 in 0.050 M citrate buffer (pH 5.0), or in 85% (w/w) of the ionic liquid [EMIM]AcO or the deep eutectic solvents [Chol]Cl : BA (5 : 2), [Chol]Cl : Gly (1 : 1) or Bet : Gly (1 : 1) during incubation at 50 °C.

One of the most distinguishing differences between the five solution systems, in which enzyme stability was studied, is their widely differing pH values (Table 1). The aqueous 85% (w/w) [Chol]Cl : BA (5 : 2) formed a very acidic solution, whereas the citrate buffer was at pH 5.0, the two glycerol-containing DES solutions had pH values close to neutral and the 85% (w/w) solution of [EMIM]AcO was highly basic. It should, however, be kept in mind that the pH scale is defined for diluted aqueous solutions and the accuracy of the measurements may partly be compromised by the high DES and IL contents in the solutions, although the measured pH values likely give good approximations. These pH effects are likely to affect enzyme activity and stability. Interestingly Xyn11 has a pI of 9.0,³⁶ which is significantly higher compared to the values ranging between 4.3 and 5.6 of the other enzymes,⁴¹ which may explain the Xyn11 tolerance to the basic [EMIM]AcO. Xyn11 has previously been reported to be tolerant towards high pH environments.³⁶ The pI differences between the other glycosyl hydrolases are rather small and cannot as such be correlated to the differences in enzyme stability in the studied solutions. pI alone is not the only explicating factor of the stability in DES/IL, but it may be one factor to take into consideration when choosing enzymes for these systems. Also the ionic strength of IL-containing hydrolysis matrices has been shown to affect cellulase performance,⁴² but the effect of ionic strength was not studied in this work. The DESs used in this work contained one uncharged component (glycerol or boric acid, which may have been partially deprotonated depending on solution conditions) in addition to the salt component, meaning that their impact on the ionic strength may be significantly lower from that of [EMIM]AcO. In previous work, it has been proposed that halides, typically chloride anions, may have a significantly inactivating effect on cellulases.²⁷ This hypothesis is well supported by the glycosyl hydrolase stability results obtained in this study, according to which the chloride-containing DESs inactivate the cellulases at different rates, whereas the Bet : Gly (1 : 1) DES, lacking chlorine, clearly stabilised the endoglucanases.

Table 1 Measured pH values of the studied 85% (w/w) DES, [EMIM]AcO and reference buffer solutions (A) and the pI values of the studied Trichoderma reesei glycosyl hydrolases (B)

A
Solution	pH
[Chol]Cl : BA 5 : 2	2
[Chol]Cl : Gly 1 : 1	6.3
Bet : Gly 1 : 1	7.8
[EMIM]AcO	12.1
Buffer	5.0

---

B
Enzyme	pI
Cel5A	5.6
Cel7A	4.3
Cel7B	4.5
Xyn11	9.0

---

Effect of DES pretreatment on lignocellulose samples

The four starting materials, MCC, milled dissolving pulp, spruce saw dust and shredded wheat straw, were subjected to a mild 24 h pretreatment at 80 °C in 90% (w/w) DES or [EMIM]AcO, or buffer as reference. After the pretreatment, the samples were washed with water to remove free DES or [EMIM]AcO. Typically, lignocellulose pretreatments are done at harsher conditions (>120 °C) and with shorter treatment times (from minutes to a few hours), although the pretreatment severity is always dependent on the starting material.¹ The pretreatment was evaluated by light microscopy and composition analysis of the pretreated material, to assess the possible removal of lignin and hemicellulose. Some differences were visually observed between the substrate–solution combinations (ESI1†). MCC formed dispersions in buffer and DESs but dissolved (at least partially) in [EMIM]AcO as expected. The same was observed for the milled dissolving pulp, although the pulp strongly swelled in buffer and DES solutions. The shredded straw formed dark brown solutions in all treatments but colour formation was light in buffer, indicating dissolution or degradation of some biomass components in the DES and IL solutions. The saw dust formed a dark solution only in 90% (w/w) [EMIM]AcO, whereas differences were small between buffer and the 90% (w/w) DES treatments. By visual inspection of the pretreatment mixtures, it was clear that [EMIM]AcO was a more powerful solvent than the DESs and that the recalcitrance of saw dust was higher than that of shredded straw. Not even the 90% (w/w) [EMIM]AcO was able to dissolve these two lignocellulosic samples completely under the used mild conditions.

The pretreated samples were examined by light microscopy to detect changes in the physical structure caused by the pretreatment in 90% (w/w) DES or [EMIM]AcO. The only treatment to cause significant changes in the substrate structures was that with [EMIM]AcO. For MCC and milled dissolving pulp, the crystal or fibre structure was completely interrupted and the pretreated material contained small particles which had very little common structural motives with the starting material (ESI2†). The reference pretreatment in buffer did not cause any visible changes in any of the starting materials. The effect of the pretreatment in 90% (w/w) [EMIM]AcO was not very pronounced on the two lignocellulosic samples (shredded straw and saw dust), once again demonstrating their recalcitrance to disintegration, although some opening of the structure could be seen (ESI3†). The microscopy images of the shredded straw and saw dust also illustrated the heterogeneous nature of these starting materials. In the DES pretreatments, most changes observed to the starting materials were small and within the limits of interpretation, demonstrating that the DESs were not as efficient in dissolving lignocellulose as [EMIM]AcO under the used mild conditions.

The untreated and in 90% (w/w) DES, [EMIM]AcO or in buffer pretreated materials were subjected to a composition analysis in which the structural carbohydrates and lignin were analysed. The MCC and dissolving pulp were almost pure cellulose, containing only a few per cents of hemicellulose in the form of xylan or mannan (ESI4A†). Wheat straw and spruce saw dust, both representing real lignocellulosic substrates, contained similar amounts of lignin (29%) and polysaccharides (58%). The greatest difference between the straw and saw dust was in the type of hemicellulose, namely mannan, being the dominant form in the spruce saw dust and xylan in the wheat straw, as previously reported.^43,44 The portion of acid soluble lignin was greater in the straw than in the saw dust.

The saccharide distribution of the pretreated samples appeared practically unchanged and only the DES treatments were able to partly remove xylose and mannan from the dissolving pulp, but not from other materials (ESI4B†). The changes in sample composition were in general small between the different treatments, which is surprising taking into account the dissolving power of [EMIM]AcO. Pretreatments in which the lignin and hemicellulose content is significantly reduced by IL treatment have been reported,⁴⁵ but likely harsher conditions or a different choice of solvents would have been needed than those used in this study to achieve pronounced effects. Procentese et al. recently showed that temperatures of 150 °C may be needed to effectively remove lignin and hemicellulose from corn cob, but this effect was also much dependent on the employed DESs.²⁵

The [EMIM]AcO treatment of straw appeared to be the only one to remove significant amounts of acid insoluble lignin from the sample (ESI4A†). The uncertainty in case of acid soluble lignin analysis was rather large for samples treated in [EMIM]AcO. The amounts of acid soluble lignin were overestimated because [EMIM]AcO in the acid hydrolysate gave a response in UV-Vis spectrophotometry, which can be misleadingly interpreted as acid soluble lignin, because it has UV absorption in part on the same wavelengths as lignin. Likely some [EMIM]AcO had been attached to the samples during pretreatment and not easily washed off before composition analysis. The low content of polysaccharides in pulp and straw samples treated in DESs or [EMIM]AcO was also surprising, and likely explained by significant amounts of pretreatment chemical residues in the samples. This was only observed for the “soft” samples, i.e. pulp and straw. Thus, the entrapment of pretreatment solvent in the substrates, or chemical reactions between the solvent components and the substrates, could potentially be serious issues to be solved in future work. Entrapment of significant amounts of ILs in pretreated substrates has also been reported in earlier studies.²⁴

Enzymatic hydrolysis of washed DES- and IL-pretreated substrates

The materials pretreated in 90% (w/w) DES or [EMIM]AcO or buffer, together with the untreated substrates, were hydrolysed with an enzyme mixture reconstituted from T. reesei glycosyl hydrolases to evaluate the efficiency of the pretreatments. The enzyme mixture contained 60% of cellobiohydrolase Cel7A, 15% both of the endoglucanases Cel5A and Cel7B, 10% of xylanase Xyn11 and was supplemented with β-glucosidase to avoid product inhibition of Cel7A by cellobiose formed during the hydrolysis. Washing of the pretreated substrates should eliminate the inhibiting effect of the DESs and [EMIM]AcO, although the composition analysis suggested significant residual contents of these chemicals in the substrates. Because the hydrolysis was done at a substrate consistency of 2% (w/w) in buffer, the concentration of possibly dissolved DES or [EMIM]AcO from the substrate was low and not likely to significantly inactivate or inhibit the enzymes.

The hydrolysis yields of the untreated substrates showed large differences (Fig. 2). Dissolving pulp could be hydrolysed to the highest degree under the chosen conditions with a yield of 62%, followed by MCC with 49% and the lignocellulosic materials with expected lower yields of 18% for wheat straw and 8% for saw dust. In addition to structural and chemical differences between the wheat straw and saw dust, the higher hydrolysis yield of wheat straw might partly be explained by the fact that it contained xylan and the enzyme mixture contained xylanase, whereas the enzyme mixture did not contain any major mannanase component to hydrolyse the mannan in the saw dust. The pretreatments in DES generally showed small increases in the hydrolysis yields. MCC digestibility was increased by pretreatment with [Chol]Cl : Gly (1 : 1) and Bet : Gly (1 : 1). The dissolving pulp was clearly an exception, having >100% hydrolysis yields for all the samples treated in DES and [EMIM]AcO. Of the lignocellulosic materials, the most encouraging pretreatment result with DES was obtained for straw treated with [Chol]Cl : BA (5 : 2), which showed an increase in hydrolysis yield to 33% from 18% for the untreated material. The spruce saw dust was the most recalcitrant substrate and only small effects of the DES pretreatment could be observed. For all four substrates, close to 100% yields were obtained for the samples treated in [EMIM]AcO, clearly demonstrating that this powerful lignocellulose solvent was able to significantly improve the digestibility of both crystalline, pure cellulose samples as well as of lignocellulosic samples, under the fairly mild pretreatment conditions. The analysed hydrolysis yields of pretreated dissolving pulp samples were in some cases over 100%, which can be attributed to two reasons: the previously mentioned high content of residual pretreatment chemicals may have influenced the sugar quantification by DNS assay, and they have likely also influenced the composition analysis, by which the saccharide content of the pretreated dissolving pulp samples was analysed. Even with these analytical uncertainties in mind, the positive pretreatment effect is clear for the dissolving pulp samples as they all had significant increases in their digestibility. This conclusion is also supported by the fact that the dissolving pulp hydrolysates became completely clear during hydrolysis (ESI5†).

Fig. 2 Enzymatic hydrolysis yields (% of polysaccharides) of untreated and by 90% (w/w) DES or [EMIM]AcO pretreated (A) microcrystalline cellulose, (B) milled eucalyptus dissolving pulp, (C) shredded wheat straw and (D) spruce saw dust. The DES or IL from the pretreatment was washed off the samples before enzymatic hydrolysis. Error bars describe the standard deviation between three parallel samples. The hydrolysis yield of pretreated dissolving pulp samples (B) are overestimated due to high contents of entrapped pretreatment chemicals.

The reducing sugars analysis by the DNS assay was confirmed by comparing to HPLC saccharide analysis for selected samples. The DNS results were generally well in line with the chromatographic results, although an overestimation in the order a few %-points of the yield with DNS was noticed. Likely this is due to the different released hemicellulose saccharides having different response factors in the DNS assay. All yield calculations were based on the actual carbohydrate content in the substrates, as determined in the composition analysis.

Visual inspection of the hydrolysis mixtures also confirmed the complete hydrolysis of [EMIM]AcO-pretreated MCC and pulp, as the mixtures became completely clear during the hydrolysis, whereas the blank samples were turbid to the end (ESI5†). The wheat straw treated in [EMIM]AcO was still turbid after hydrolysis, but all the coarse particles had been disintegrated. The enzymatic treatment also caused an interesting effect on the in [Chol]Cl : Gly (1 : 1) and Bet : Gly (1 : 1) pretreated substrate sedimentation after the mixing was stopped for the MCC (to some extent also for pulp); the enzyme-treated materials stayed in dispersion to a much higher degree than the blank samples (ESI6†). For saw dust, the opposite effect was noticed; the samples pretreated in [Chol]Cl : Gly (1 : 1) and Bet : Gly (1 : 1) sedimented more heavily after enzymatic hydrolysis than the blank samples. Thus, the pretreatments and subsequent enzymatic hydrolysis apparently changed the dispersion behaviour of the substrates.

The DES pretreatments clearly improved the enzymatic hydrolysis of dissolving pulp and with certain DESs also the hydrolysis of MCC and wheat straw. However, the DES pretreatments were generally not powerful enough to significantly affect the sample digestibility of the lignocellulosic samples under the chosen mild pretreatment conditions. In contrast to our results, a recent report describes how rice straw could be effectively pretreated with DES combinations based on lactic acid and betaine or [Chol]Cl at a low temperature of 60 °C, by which especially lignin was removed.²⁶ The reported efficient pretreatment under these mild conditions may either have been due to the choice of DES, differences in the lignocellulose structure, or other differences in the experimental setup, as compared to the results in this study. It is also possible that the presence of 10% (w/w) of water in the DESs used during pretreatment in this study diluted the DES too much from efficiently interacting with the substrates. DES pretreatment efficiency may be adjusted by adding water to lower the viscosity but at the same time the DES hydrogen bonding interactions are gradually lost by the water competition; a recent study indicated that at 50% water content the DES internal hydrogen bonding cease to exist.⁴⁶ The dissolving pulp had a higher digestibility after DES treatment, which may be attributed to swelling of the fibre structures. In the case of wheat straw, the increased digestibility after treatment in [Chol]Cl : BA (5 : 2) could be due to this acidic DES altering the hemicellulose structures, although no hemicellulose or lignin removal was observed as a result of this treatment. Saw dust was obviously more recalcitrant to such effects. The high digestibility of [EMIM]AcO-treated MCC and pulp is expected based on previous reports,¹⁷ as the pretreatment in this IL in fact was a regeneration, in which at least most of the cellulose was dissolved and then precipitated presumably as amorphous cellulose or cellulose II or a mixture of both, with higher digestibility. In the case of the two lignocellulosic substrates, saw dust and wheat straw, the sharp increase in digestibility after the treatment with [EMIM]AcO could not be explained by observed changes in the sample compositions after the IL and DES treatments, and the samples also appeared rather intact as studied by light microscopy, but yet practically full yields were obtained in their hydrolysis. General sample swelling and partial reallocation of hemicellulose and lignin may partially explain this effect, but likely some other changes, which are not detected with the analytical methods used in this study, have occurred in the substrates during [EMIM]AcO treatment. Previously, reallocation of lignin has been suggested to happen e.g. during steam pretreatment.⁴⁷ Although speculative, a mechanism by which some of the lignin and hemicellulose would be dissolved during the pretreatment with DES or [EMIM]AcO, and then reprecipitated during the washing of the samples with water, cannot be excluded.

One-pot enzymatic hydrolysis of lignocellulose in the presence of DES

The one-pot or in situ IL pretreatment with subsequent enzymatic hydrolysis of lignocellulose in IL solution was introduced by Kamiya et al. in 2008 as an alternative approach to the separate pretreatment with washing and hydrolysis process.²⁰ The benefits of the one-pot procedure include avoiding extra processing steps and extensive water usage for substrate washing after the pretreatment. The high water consumption in washing off the IL from the substrate has been identified as a major cost when performing separate pretreatment and hydrolysis steps.²¹ In this work, two different one-pot systems were evaluated; one in which the final hydrolysis mixture contained 30% (w/w) of DES or [EMIM]AcO, while the other contained a high 80% (w/w) concentration of these solvents. Pretreatment was in both cases done in 90% (w/w) DES or [EMIM]AcO concentration at 80 °C for 24 h before the hydrolysis, with a similar pretreatment in citrate buffer (pH 5.0) as reference.

In 30% (w/w) DES or [EMIM]AcO one-pot hydrolysis, the beneficial effects of the DES or [EMIM]AcO pretreatment were greatly subdued by their inhibiting effect on the cellulase action (Fig. 3). For MCC and pulp, the pretreatment with DES had beneficial effects on the hydrolysis yield after washing (Fig. 2), but in the 30% (w/w) DES hydrolysis mixture this effect was overcome by cellulase inactivation or inhibition and the yields of the DES-treated substrates did not reach the same levels as for the untreated substrates in buffer. In the hydrolysis of the washed pretreated substrates, the [EMIM]AcO treatment clearly led to the highest hydrolysis yields (Fig. 2), but in the one-pot procedure the negative effect if this IL on enzymatic action became dominant and the hydrolysis yields were for all four substrates <3%, clearly demonstrating incompatibility with enzymes. The reference pretreatment in buffer did not change the yields as compared to those obtained from the untreated materials. It should, however, be noted that the hydrolysis yields in 30% (w/w) DES were significantly higher than in 30% (w/w) [EMIM]AcO, which is well in agreement with the presented enzyme stability results (Fig. 1) and means that the combined inactivating and inhibiting effect of the DESs is much smaller than that of [EMIM]AcO.

Fig. 3 Enzymatic hydrolysis yields (% of polysaccharides) of untreated and by 90% (w/w) DES or [EMIM]AcO pretreated (A) microcrystalline cellulose, (B) milled eucalyptus dissolving pulp, (C) shredded wheat straw and (D) spruce saw dust in the presence of 30% (w/w) of DES or [EMIM]AcO. Error bars describe the standard deviation between three parallel samples.

Some differences could be seen between the different DESs in one-pot hydrolysis. The difference between enzymatic hydrolysis yields for the two glycerol-containing DESs were generally small (Fig. 3). On the other hand, the effect of [Chol]Cl : BA (5 : 2) in the one-pot hydrolysis of spruce saw dust and shredded wheat straw was interesting. Although this DES was rather inactivating for the cellulases in the stability measurements (Fig. 1), enzymatic hydrolysis yields from the pretreated lignocellulose samples in 30% (w/w) of this DES were practically as high as for the untreated substrates in buffer and mostly higher than the yields in the Gly-containing DES systems (Fig. 3). As earlier stated, the microscopy images and composition analysis did not indicate significant changes caused by the [Chol]Cl : BA (5 : 2) treatment in these substrates and a clear reason to the comparatively high yields in one-pot hydrolysis with this DES cannot be identified with the used analytical methods. Possibly, the acidity of this DES had caused some hydrolysis of the cellulose and hemicellulose chains, which renders the sample more susceptible to enzymatic hydrolysis. Another alternative interpretation is that the DESs and ILs in addition to having inactivating effects also cause inhibition of the enzymes, which in part could explain the partial discrepancy between stability results and one-pot hydrolysis yields in 30% (w/w) DES or IL. Some ILs have in earlier studies been suggested to cause competitive inhibition of glycosyl hydrolases.⁴⁸

A one-pot hydrolysis series was also made in concentrated 80% (w/w) DES and [EMIM]AcO solutions. With all combinations of substrates and chemicals in the pretreatment and subsequent enzymatic hydrolysis, hydrolysis yields were low (typically <2%), showing that enzymatic hydrolysis in the presence of high concentrations of these solvents is not feasible with the chosen enzymes (results not shown). The combined one-pot hydrolysis and stability results demonstrate that retaining enzymatic activity in highly concentrated DES solutions is not enough for successfully hydrolysing solid substrates under these conditions. In many cases the substrate solutions in 80% (w/w) DES or [EMIM]AcO became very viscous and in 80% [EMIM]AcO the MCC and pulp were mostly dissolved. It is proposed that mass transfer limitations due to high viscosity and gelling play a role in the low hydrolysis levels as well as the actual cellulase inactivation and inhibition.

The water activity of the used aqueous DES and [EMIM]AcO solutions was also measured and lower water activity for water solutions of [EMIM]AcO than of the DESs was measured (Fig. 4). All the three DESs showed similar water activities. Although water activity may be one limiting factor for enzymatic hydrolysis,⁴⁹ the small differences in water activity of the three DESs and [EMIM]AcO does not in the current study provide an explanation to the differences in enzymatic hydrolysis yields in one-pot hydrolysis nor for the differences in enzyme stability in the DESs and [EMIM]AcO.

Fig. 4 Water activity values for aqueous DES and [EMIM]AcO solutions.

Conclusions

In this study, the enzymatic compatibility and pretreatment efficiency under mild conditions (80 °C) of three deep eutectic solvents (DESs) were compared to the ionic liquid (IL) [EMIM]AcO, a well-known and powerful lignocellulose solvent. Two DESs containing glycerol as hydrogen bond donor were identified to significantly stabilize cellulases, and are thus of great interest for further work in the development of enzyme-compatible DES systems for lignocellulose modification or deconstruction.

After DES and IL pretreatment of cellulosic substrates, MCC and dissolving pulp, a significant increase in enzymatic digestibility was observed. For the two lignocellulosic substrates, spruce saw dust and shredded wheat straw, the DES treatments led to small increases in enzymatic hydrolysis yield and the treatment with [Chol]Cl : boric acid (5 : 2) led to a significant increase in wheat straw hydrolysis. [EMIM]AcO was found to be a more powerful pretreatment chemical as measured by enzymatic hydrolysis yield, when the pretreated substrate was washed. Achieving high hydrolysis yields with [EMIM]AcO pretreatment was thus possible with relatively mild pretreatment conditions. To achieve comparable results with DESs, the pretreatment conditions would likely need to be more severe, or then DESs with better pretreatment efficiency would need to be found. Repeating the pretreatment in pure DES instead or in 90% (w/w) DES solution might also yield better pretreatment results. Biomass modification with DESs is currently a rising research topic, and thus it can be presumed that efficient DESs for lignocellulose pretreatment will be identified in the near future, a property which needs to be combined with fair enzyme compatibility. The entrapment of pretreatment chemicals in washed substrates was identified as a potential processing difficulty with both DESs and [EMIM]AcO.

In one-pot hydrolysis, with the DES or [EMIM]AcO present in the hydrolysis mixture, [EMIM]AcO was noticed to be extremely inactivating for the used cellulases. However, a lot of research has been directed towards finding compatible enzymatic hydrolysis and IL systems during recent years as recently reviewed,¹⁹ and the performance of e.g. more IL-tolerant or stabilised enzymes should be studied in the presence of [EMIM]AcO. Hydrolysis yields in the presence of DESs were significantly higher than in [EMIM]AcO solutions, correlating well with the stability measurements of the monocomponent cellulases used in the total hydrolysis experiments. The rather modest hydrolysis yields obtained in the presence of 30% (w/w) DES demonstrate that enzyme stability in DESs or ILs does not predict well the enzyme performance in the hydrolysis of real, solid substrates in these solutions. In addition to decreased enzyme stability, also enzyme inhibition, e.g. competitive inhibition by the IL or DES components may be a significant contributor to the low observed hydrolysis performance of the enzymes.

Acknowledgements

The Finnish Funding Agency for Innovation TEKES is thanked for funding this study through the “Deep eutectic solvents as irreplaceable tools towards sustainable chemistry” project (grant decision number 40329/13). VTT's Bioeconomy Transformation program is as well thanked for financial support for this study. Nina Vihersola, Jenni Lehtonen and Riitta Alander are acknowledged for skilful technical assistance.

References

1. A. Brandt, J. Gräsvik, J. P. Hallett and T. Welton, Green Chem., 2013, 15, 550–583.
2. R. P. Chandra, R. Bura, W. E. Mabee, A. Berlin, X. Pan and J. N. Saddler, Adv. Biochem. Eng./Biotechnol., 2007, 108, 67–93.
3. H. E. Grethlein, Nat. Biotechnol., 1985, 3, 155–160.
4. C. A. Mooney, S. D. Mansfield, M. G. Touhy and J. N. Saddler, Bioresour. Technol., 1998, 64, 113–119.
5. Y.-H. P. Zhang and L. R. Lynd, Biotechnol. Bioeng., 2004, 88, 797–824.
6. J. Larsen, M. T. Haven and L. Thirup, Biomass Bioenergy, 2012, 46, 36–45.
7. C. Lehmann, F. Sibilla, Z. Maugeri, W. R. Streit, M. de Dominguez, R. Martinez and U. Schwaneberg, Green Chem., 2012, 14, 2719–2726.
8. R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J. Am. Chem. Soc., 2002, 124, 4974–4975.
9. I. Kilpeläinen, H. Xie, A. King, M. Granström, S. Heikkinen and D. S. Argyropoulos, J. Agric. Food Chem., 2007, 55, 9142–9148.
10. M. Francisco, A. Van Den Bruinhorst and M. C. Kroon, Angew. Chem., Int. Ed., 2013, 52, 3074–3085.
11. A. P. Abbott, D. Boothby, G. Capper, D. L. Davies and R. K. Rasheed, J. Am. Chem. Soc., 2004, 126, 9142–9147.
12. P. Domínguez de María and Z. Maugeri, Curr. Opin. Chem. Biol., 2011, 15, 220–225.
13. H. Zhao and G. A. Baker, J. Chem. Technol. Biotechnol., 2013, 88, 3–12.
14. A. W. T. King, J. Asikkala, I. Mutikainen, P. Järvi and I. Kilpeläinen, Angew. Chem., Int. Ed., 2011, 50, 6301–6305.
15. A. Abbott, R. Harris, K. Ryder, C. D'Agostino, L. Gladden and M. Mantle, Green Chem., 2011, 13, 82–90.
16. K. D. O. Vigier, G. Chatel and F. Jérôme, ChemCatChem, 2015, 7, 1250–1260.
17. A. P. Dadi, S. Varanasi and C. A. Schall, Biotechnol. Bioeng., 2006, 95, 904–910.
18. G. Cheng, P. Varanasi, C. Li, H. Liu, Y. B. Melnichenko, B. A. Simmons, M. S. Kent and S. Singh, Biomacromolecules, 2011, 12, 933–941.
19. R. M. Wahlström and A. Suurnäkki, Green Chem., 2015, 17, 694–714.
20. N. Kamiya, Y. Matsushita, M. Hanaki, K. Nakashima, M. Narita, M. Goto and H. Takahashi, Biotechnol. Lett., 2008, 30, 1037–1040.
21. N. M. Konda, J. Shi, S. Singh, H. W. Blanch, B. A. Simmons and D. Klein-Marcuschamer, Biotechnol. Biofuels, 2014, 7, 86.
22. F. Xu, J. Sun, N. V. S. N. M. Konda, J. Shi, T. Dutta, C. D. Scown, B. A. Simmons and S. Singh, Energy Environ. Sci., 2016, 9, 1042–1049.
23. M. Francisco, A. Van Den Bruinhorst and M. C. Kroon, Green Chem., 2012, 14, 2153–2157.
24. S. Xia, G. A. Baker, H. Li, S. Ravula and H. Zhao, RSC Adv., 2014, 4, 10586–10596.
25. A. Procentese, E. Johnson, V. Orr, A. Garruto Campanile, J. A. Wood, A. Marzocchella and L. Rehmann, Bioresour. Technol., 2015, 192, 31–36.
26. A. K. Kumar, B. S. Parikh and M. Pravakar, Environ. Sci. Pollut. Res., 2016, 23, 9265–9275.
27. M. B. Turner, S. K. Spear, J. G. Huddleston, J. D. Holbrey and R. D. Rogers, Green Chem., 2003, 5, 443–447.
28. A. A. N. Gunny, D. Arbain, E. M. Nashef and P. Jamal, Bioresour. Technol., 2015, 181, 297–302.
29. C. Lehmann, M. Bocola, W. R. Streit, R. Martinez and U. Schwaneberg, Appl. Microbiol. Biotechnol., 2014, 98, 5775–5785.
30. S. L. Bradbury and W. N. Jakoby, Proc. Natl. Acad. Sci. U. S. A., 1972, 69, 2373–2376.
31. J. Hiltunen, S. Vuoti and L. Kuutti, Deep eutectic solvents and their use, International patent application, WO2015128550, 2015.
32. A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton and D. Crocker, Determination of Structural Carbohydrates and Lignin in Biomass. Laboratory Analytical Procedure (LAP), National Renewable Energy Laboratory, Golden, Colorado, 2008.
33. M. Tenkanen and M. Siika-aho, J. Biotechnol., 2000, 78, 149–161.
34. O. Goldschmid, in Lignins: Occurrence, Formation, Structure and Reactions, ed. K. V. Sarkanen and C. H. Ludwig, John Wiley & Sons, Ltd, New York, 1971, pp. 241–298.
35. A. Suurnäkki, M. Tenkanen, M. Siika-aho, M.-L. Niku-paavola, L. Viikari and J. Buchert, Cellulose, 2000, 7, 189–209.
36. M. Tenkanen, J. Puls and K. Poutanen, Enzyme Microb. Technol., 1992, 14, 566–574.
37. J. B. Sumner, J. Biol. Chem., 1924, 62, 287–290.
38. M. J. Bailey and M. Linko, J. Biotechnol., 1990, 16, 57–66.
39. U. K. Laemmli, Nature, 1970, 227, 680–685.
40. G. Ebner, P. Vejdovszky, R. Wahlström, A. Suurnäkki, M. Schrems, P. Kosma, T. Rosenau and A. Potthast, J. Mol. Catal. B: Enzym., 2014, 99, 121–129.
41. M. Sandgren, PhD Thesis, Uppsala University, 2003.
42. P. Engel, R. Mladenov, H. Wulfhorst, G. Jäger and A. C. Spiess, Green Chem., 2010, 12, 1959–1966.
43. E. Sjöström, Wood Chemistry. Fundamentals and Applications, Academic Press, Inc., San Diego, 2nd edn, 1993.
44. J. Vogel, Curr. Opin. Plant Biol., 2008, 11, 301–307.
45. K. Ninomiya, T. Yamauchi, M. Kobayashi, C. Ogino, N. Shimizu and K. Takahashi, Biochem. Eng. J., 2013, 71, 25–29.
46. Y. Dai, G.-J. Witkamp, R. Verpoorte and Y. H. Choi, Food Chem., 2015, 187, 14–19.
47. L. A. Donaldson, K. K. Y. Wong and K. L. Mackie, Wood Sci. Technol., 1988, 22, 103–114.
48. H. Li, A. Kankaanpää, H. Xiong, M. Hummel, H. Sixta, H. Ojamo and O. Turunen, Enzyme Microb. Technol., 2013, 53, 414–419.
49. A. C. Salvador, M. d. C. Santos and J. A. Saraiva, Green Chem., 2010, 12, 632–635.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11719h

---