Co-Production of Cellulose Nanocrystals and Fermentable Sugars Assisted by Endoglucanase Treatment of Wood Pulp

In this study, fermentable sugars and cellulose nanocrystals (CNCs) were co-produced from endoglucanase treatment of wood pulp, followed by acid hydrolysis. Enzymatic hydrolysis was performed using two endoglucanases differentiated by the presence or absence of a cellulose-binding domain (CBD). The enzyme with an intact CBD gave the higher glucan conversion (up to 14.1 ± 1.2 wt %) and improved the degree of crystallinity of the recovered wood pulp fiber (up to 83.0 ± 1.0%). Thus, this endoglucanase-assisted treatment successfully removed amorphous content from the original cellulosic feedstock. CNC recovery (16.9 ± 0.7 wt %) from the feedstock going into the acid hydrolysis was improved relative to untreated pulp (13.2 ± 0.6 wt %). The mass loss from enzymatic treatment did not cause a decrease in the CNC yield from the starting material. The characteristics of CNCs obtained through acid hydrolysis (with or without enzyme treatment of pulp) were analyzed using X-ray diffraction, transmission electron microscopy, dynamic light scattering, Fourier transform infrared spectroscopy, and differential scanning calorimetry as characterization techniques. The CNCs generated through acid hydrolysis of endoglucanase-treated wood pulp displayed comparable properties relative to those generated using untreated pulp. Thus, endoglucanase treatment can enable co-production of CNCs and sugars for biofuel fermentation.


Introduction
Over the last decade, the potential for the development of nanomaterials from the forest sector has become apparent and achievable. Additionally, the market demand for some traditional forest products, such as pulp and paper, has greatly declined [1]. The production of high-performance nanomaterials provides opportunities for the development of more valuable forestry products to sustain the forest-based economy [2]. Cellulose nanocrystals (CNCs) are nanomaterials that are derived from the most abundant renewable polymer resource-cellulose [3]. Cellulose can be isolated from wood by a kraft pulping process [4]. In its natural form, cellulose is primarily comprised of two main forms: loosely-packed (amorphous) cellulose fibers and highly structured (crystalline) cellulose fibers. CNCs are formed from the elementary building blocks of highly ordered crystalline cellulose with lengths and widths in the nanoscale range [5]. Particles and matter at these dimensions display unique and novel properties that are unlike those displayed by particles at the larger dimensions of their precursors [2]. CNCs have very high tensile strength, axial Young's modulus, and are potentially

Feedstock
Northern bleached hardwood kraft (NBHK) pulp was oven-dried and cut into square pieces (roughly 8 × 6 × 1 mm, in length, width and thickness) by a pulp chopper (Cutting Machine G45L1, Pierret, Corbion, Belgium). The pulp, which was kindly provided by Alberta Pacific Forest Industries Inc. or Al-Pac (Edmonton, AB, Canada), had a moisture content of 5.9 ± 0.1%. A two-step acid digestion procedure was carried out in quadruplicate, according to Jain et al., for pulp compositional analysis [33]. NBHK pulp had 78.0 ± 0.8% glucan, 15.7 ± 1.0% xylan, 0.3 ± 0.1% extractives, and 0.8 ± 0.2% acid soluble lignin. It should be noted that routine compositional analysis following the standard TAPPI protocol (T 203) based on a two-step alkaline extraction, as determined by InnoTech Alberta's Biomass Conversion and Processing Technologies Unit for quality control purposes, gives 92.5% alpha cellulose, 1.8% acid insoluble lignin, 0.6% ash, and 0.4% extractives.

Enzymes and Specific Activity
Enzymatic hydrolysis was performed using two proprietary, non-commercially purified endoglucanase preparations: an endo-1,4-beta-glucanase (Endo-CBD+: NS 51137), and a mutant form of the same enzyme in which the cellulose-binding domain had been removed (Endo-CBD−: NS 51172). These enzymes were graciously provided by Novozymes ® A/S (Bagsvaerd, Denmark). The optimum pH of the enzyme treatment systems employed for Endo-CBD+ and Endo-CBD− were established to be 7.0 and 4.0, respectively. Endoglucanase activities were determined following the carboxymethyl cellulose (CMC) assay [34], which facilitated equalization of Endo-CDB+ and Endo-CDB− loading. Based on the results from the CMC assay, a range of enzyme loadings (0.2, 0.4, 0.6, 0.8, 1.0, and 5.0 U/g of pulp) were used for the experiments described below.
Since endoglucanases release long chain polysaccharides rather than mono-or disaccharides, total sugar release was measured following complete hydrolysis of sugars in the liquid phase recovered after endoglucanase treatment. This was achieved using a highly-processive proprietary cellulase cocktail solution (NS 51129) provided by Novozymes ® A/S (Bagsvaerd, Denmark), which was the same enzyme employed by Beyene et al. [32].

Enzymatic Hydrolysis
A citrate-phosphate buffer was used for enzyme hydrolysis of the wood pulp. The pH of the buffer was adjusted to 7.0 for Endo-CBD+ and 4.0 for Endo-CBD− by formulating different proportions of stock solutions of 0.1 M citric acid (99.5%; Sigma Aldrich, St. Louis, MO, USA) and 0.2 M dibasic sodium phosphate (99%; Fisher Scientific, Hampton, NH, USA).
Wood pulp (10 g, dry weight) was hydrolyzed with enzymes in citrate-phosphate buffer (100 mL) at 50 • C for 72 h in a shaking (100 rpm) water bath. The reaction was terminated by heating in boiling water for 10 min. Pulp samples that were suspended in buffer without addition of enzymes and treated under the same condition served as the negative control. All experiments were done in triplicates.
Following endoglucanase or mock treatment, the resulting slurries were centrifuged at 33,700× g for 10 min. The solid fraction was recovered and subsequently subjected to standard acid hydrolysis for production of CNCs (as described in Section 2.5). An aliquot of the liquid fraction (5 mL), recovered as enzyme hydrolysate, was subjected to an additional step of enzymatic hydrolysis using NS 51129 to promote complete saccharification of the released long and/or short cellulose chains. This subsequent enzymatic hydrolysis was conducted at 50 • C for 72 h in a shaking water bath (100 rpm). The measured sugar concentrations were used to calculate the recoverable sugar yield from the endoglucanase treatment.

High Performance Liquid Chromatography (HPLC)
Monosaccharides were quantified using an HPLC system (HPLC, 1200 series, Agilent Technologies, Mississauga, ON, Canada) equipped with a refractive index detector (1100 series, Agilent Technologies, Mississauga, ON, Canada) and an Aminex HPX-87P column (Bio-Rad, Hercules, CA, USA). The analyses were carried out using HPLC water (Sigma Aldrich, St. Louis, MO, USA) as the mobile phase at a temperature of 65 • C and a flow rate of 0.5 mL/min for 40 min. Glucan conversion (wt %), total sugar yield (wt % pulp), and enzyme-treated solids recovered (wt % pulp) were calculated using Equations (1)-(3), respectively.
where Glucose (wt) = mass of glucose calculated from HPLC analysis (g), W 1 = mass of glucose in the pulp (determined from compositional analysis) used for enzyme hydrolysis (g) where AHC = the anhydro correction factor to convert mass of glucose to cellulose (0.9), W 2 = mass of pulp used for enzyme hydrolysis (10 g) Enzyme-treated solids recovered (wt % pulp) = 100 − Total mass loss (wt % pulp)

Sample Preparation and Acid Hydrolysis
The undigested solids were centrifuged and washed three times with Milli-Q water at a centrifugation speed of 33,700× g for 10 min for each wash. The sugar-free solids were then freeze-dried and subjected to acid hydrolysis. The method used for bench scale acid hydrolysis was previously described by Beyene et al. [32] and was directly adopted from InnoTech Alberta's Biomass Conversion and Processing Unit's protocol (Edmonton, Alberta). The endoglucanase-treated solids (8 g, dry weight) were mixed with 100 mL of 64 wt % sulfuric acid (95-98%, Sigma-Aldrich, St. Louis, MO, USA) in 500 mL shake flasks to get a pulp to acid ratio of 8% (w/v). The flasks were placed in a water bath set at 45 • C, and were subjected to overhead stirring at 200 rpm. After a reaction time of 2 h, the reaction was quenched through addition of 10-fold (v/v) cold water. The suspension was then centrifuged for 10 min at 6400× g and some of the dilute acid was decanted. The pellet was resuspended in water to a final volume of 100-150 mL, and then neutralized with 4 M sodium hydroxide (98.8% Fisher Scientific; Hampton, NH, USA) to a final pH of 7.0. To help minimize the salt content, the suspension was centrifuged at 3700× g for 10 min, and the supernatant was decanted. The pellet was resuspended in water to a final volume of 20-70 mL, and placed into regenerated cellulose membrane tubing (Spectrum TM Spectra/Por TM , Rancho Dominguez, CA, USA, 12-14 kDa molecular weight cut off) to facilitate dialysis. A conductivity meter was used to monitor the ionic strength of the liquid used for dialysis for 3-5 days. Once the conductivity was measured at 100-150 µS/cm, the suspension was centrifuged at 8900× g for 10 min, which facilitated precipitation of oversize particles. The colloidal CNC suspensions were collected, and the pellet was washed with water in order to lower the ionic strength and extract any remaining CNCs. An aliquot was taken from the pooled supernatants and oven-dried overnight at 103 • C. CNC recovery (wt % acid hydrolyzed feedstock) and CNC yield (wt % initial feedstock) were calculated using Equations (4) and (5), respectively, as shown below: where W 3 = mass of feedstock used in the acid hydrolysis (8 g), W 4 = mass of oven-dried aliquot of CNCs (g), R = ratio of the aliquot to the total mass of the colloid CNC yield (wt % initial feedstock) = CNC recovery (wt % acid hydrolyzed feedstock)×Enzyme-treated solids recovered (wt % pulp) 100 (5) 2.6. Characterization 2.6.1. X-ray Diffraction (XRD) The crystalline structures of endoglucanase-treated wood pulp and CNCs were analyzed using a Rigaku Ultima IV X-ray diffractometer (Rigaku, Scottsdale, AZ, USA) with CuKα radiation (wavelength of 0.154 nm). The operating voltage and current were 40 kV and 44 mA, respectively. The samples were scanned at 2 • /min with a 2θ angle ranging from 5 • to 40 • . The crystallinity index (CrI) value was calculated according to Segal et al. [35] and is shown in Equation (6) below.
where I t = the highest total crystalline peak at a diffraction angle around 2θ = 22.5 • , I am = the amount of amorphous cellulose around at a diffraction angle around 2θ = 18 • .

Fourier Transform Infrared Spectroscopy (FTIR)
FTIR spectra of freeze-dried endoglucanase-treated pulp and CNCs were measured on a Thermo Nicolet iS50 FTIR spectrometer (Thermo Fisher Scientific, Madison, WI, USA) in attenuated total reflectance (ATR) mode and equipped with a diamond probe. The scan range was from 4000 to 400 cm −1 , with a resolution of 4 cm −1 . The spectra were averaged from 64 repeating scans per sample and the raw spectra were treated with auto baseline correction and advanced ATR correction. Three replicated measurements were recorded for each sample. The crystalline index (CI) was taken from the ratio of absorbance (%) from 1431 to 900 cm −1 , which is attributed to the crystallinity index of cellulose I [36]. When FTIR experiments were performed on endoglucanase-treated samples, solid wood pulp fiber residues and CNCs obtained from 1.0 U/g of pulp treatment were used.

Dynamic Light Scattering (DLS) and Zeta Potential Analyzer
CNC colloids were first diluted to 0.1 wt % in Milli-Q water. The 0.1 wt % dispersions were then diluted 1:1 with 10 mM NaCl to obtain a 0.05 wt % CNC dispersion in a 5 mM NaCl solution. The prepared CNC suspensions were filtered through a 0.45 µm syringe filter into sample cuvettes. The samples were analyzed on a Malvern Zetasizer Nano-ZS (Malvern Instruments Inc., Montreal, QC, Canada) for DLS and zeta potential analysis. The Zetasizer software (Zetasizer Software 7.11 provided by Malvern Instruments Inc., Montreal, QC, Canada) generates triplicate data from each experimental replicate (triplicate).

Transmission Electron Microscopy (TEM)
TEM images were captured using a Philips/FEI, Morgagni 268 TEM (Hillsboro, OR, USA) with 80 kV accelerating voltage. CNC colloids were diluted to 0.1 wt %, and a droplet was mounted on a formvar-coated 300-mesh copper grid (Ted Pella Inc., Redding, CA, USA). After 15 min, filter paper was used to carefully remove the excess liquid. Staining was achieved by exposing the specimen to phosphotungstic acid (2%, w/v) for 15 s. The excess stain was carefully removed using a filter paper before placing the sample onto the TEM sample holder.

Differential Scanning Calorimetry (DSC)
Freeze-dried CNCs and endoglucanase-treated pulp samples were analyzed on a TA Q1000 DSC (TA Instruments, New Castle, DE, USA). Samples (4-8 mg) were weighted into aluminum pans.
The temperature profile ranged from 30 • C to 350 • C with a heating rate of 10 • C/min. The nitrogen flow rate was set to 60 mL/min.

Statistical Analysis
Statistical analysis was done using SPSS Version 25 (IBM, Armonk, NY, USA). One-way ANOVA with Tukey's HSD test was conducted for determining the significant level per experimental conditions with a confidence interval of 95%. All experiments were done in triplicates, unless stated otherwise.

Enzymatic Hydrolysis and Sugar Yield
Endoglucanases were used to selectively digest amorphous regions of wood pulp without damaging the crystalline cellulose domains. Endoglucanases with (Endo-CBD+) and without (Endo-CBD−) a cellulose-binding domain were selected for enzymatic hydrolysis. Since endoglucanases only cleave cellulose into fractions with different chain lengths, limited monosaccharides are found in the hydrolysates. Therefore, to evaluate the recoverable sugars released from wood pulp through endoglucanase activity, a second enzymatic hydrolysis was conducted using a cellulase cocktail with high activity. Glucose was the only sugar detected in the enzymatic hydrolysates, which indicates that the two endoglucanase cocktails lack activity on the xylan chains. Endo-CBD+ generally released more polysaccharide chains (that were subsequently hydrolyzed to glucose through a secondary enzyme hydrolysis) than Endo-CBD−, when the two endoglucanases were compared at constant enzyme loading (Figure 1). Karita et al. compared endoglucanase activity with and without a CBD, and reported that, in the latter case, the enzyme showed poor adsorption, and at low cellulose concentration, the enzymes exhibited low levels of binding [29].

Statistical Analysis
Statistical analysis was done using SPSS Version 25 (IBM, Armonk, NY, USA). One-way ANOVA with Tukey's HSD test was conducted for determining the significant level per experimental conditions with a confidence interval of 95%. All experiments were done in triplicates, unless stated otherwise.

Enzymatic Hydrolysis and Sugar Yield
Endoglucanases were used to selectively digest amorphous regions of wood pulp without damaging the crystalline cellulose domains. Endoglucanases with (Endo-CBD+) and without (Endo-CBD−) a cellulose-binding domain were selected for enzymatic hydrolysis. Since endoglucanases only cleave cellulose into fractions with different chain lengths, limited monosaccharides are found in the hydrolysates. Therefore, to evaluate the recoverable sugars released from wood pulp through endoglucanase activity, a second enzymatic hydrolysis was conducted using a cellulase cocktail with high activity. Glucose was the only sugar detected in the enzymatic hydrolysates, which indicates that the two endoglucanase cocktails lack activity on the xylan chains. Endo-CBD+ generally released more polysaccharide chains (that were subsequently hydrolyzed to glucose through a secondary enzyme hydrolysis) than Endo-CBD−, when the two endoglucanases were compared at constant enzyme loading (Figure 1). Karita et al. compared endoglucanase activity with and without a CBD, and reported that, in the latter case, the enzyme showed poor adsorption, and at low cellulose concentration, the enzymes exhibited low levels of binding [29].

Enzyme Loading (U/g of pulp)
Endo-CBD-Endo-CBD+ Figure 1. Glucan conversion from hydrolysis of wood pulp with endoglucanases. Wood pulp was subjected to enzyme hydrolysis at various loadings of endoglucanases with (Endo-CBD+) or without (Endo-CBD−) a cellulose-binding domain. Since endoglucanases generate long/short chain sugars rather than monosaccharides (i.e., glucose), quantification of the sugars released from wood pulp into the supernatant through endoglucanase action was achieved through a secondary enzymatic hydrolysis that enabled complete hydrolysis of sugars to glucose monomers. Columns labelled with different letters indicate means that are statistically different at a 95% confidence level.
The control group with no enzyme addition did not show any glucose release, which indicated that no free sugars were associated with the original wood pulp. Thus, any glucose released likely resulted from the addition of endoglucanase. Glucose yield from the Endo-CBD+ enzyme hydrolysis was quite low (less than 4%) until the 5.0 U/g of pulp enzyme loading was employed, where 14.1 ± 1.2 wt % glucan conversion was observed. Conversely, when the Endo-CBD− endoglucases was used, virtually no glucose (≤0.5 wt %) was released from the wood pulp, except for the 5.0 U/g of pulp loading where 4.5 ± 0.3 glucan conversion (wt %) was observed. This yield was statistically similar to the amount of glucose released using the 0.8 U/g of pulp and 1.0 U/g of pulp Endo-CBD+ loadings (Figure 1). It should be noted that the levels of glucose yield using the two endoglucanases were significantly lower than the extents (20-40%) observed by Beyene et al. from wood pulp hydrolysis using a highly processive cellulase cocktail. In addition, 6-12% xylose release was reported in these experiments, while the endoglucanases used in the present study were unable to hydrolyze hemicellulose [32]. Zhu et al. also observed that endoglucanase (Fibercare ® ) hydrolysis of softwood pulp was very poor, with nearly 100% solid yield recovered [37]. A cocktail enzyme composed of all three classes of cellulases (Celluclast 1.5 L), on the other hand, nearly resulted in 40% degradation. It was suggested that the endoglucanase activity was limited to chain cutting, while the cocktail led to complete hydrolysis of cellulose to monosaccharides. Nevertheless, it is possible that the action of endoglucanases can partially hydrolyze amorphous regions, making them more amenable to acid hydrolysis, and that future experiments can incorporate a small amount of cellobiohydrolase and/or xylanase.

CNC Yields from Acid Hydrolysis
The solid residues remaining after the endoglucanase treatments were then subjected to standard acid hydrolysis for production of CNCs. The CNC recovery (wt % acid hydrolyzed feedstock) was then calculated based on the aliquot of endoglucanase-treated wood pulp going into the acid hydrolysis reactor. Incorporation of the mock-treated wood pulp (no enzyme) resulted in a CNC recovery of 13.2 ± 0.6%. This increased to maxima of 16.9 ± 0.7% and 15.2 ± 0.8% when acid hydrolyses were performed using the solid residues obtained through Endo-CBD+ and Endo-CBD− treatment at 5.0 U/g of pulp loading, respectively ( Figure 2). When Endo-CBD+ was used, a 1.0 U/g of pulp loading was the minimum enzyme loading that resulted in an increased CNC recovery following acid hydrolysis, relative to the no enzyme control. Conversely, for Endo-CDB-, a higher enzyme loading of 5.0 U/g of pulp was required before the CNC recovery obtained through acid hydrolysis was significantly higher than the mock-treated sample. It is worth noting that although increasing the enzyme loadings from 1.0 to 5.0 U/g of pulp resulted in a dramatic increase in the amount of recoverable sugar for both enzymes (Figure 1), there was no difference in the CNC recovery observed. With higher sugar yields, the endoglucanase treatment is expected to improve the concentration of crystalline cellulose as these enzymes have specific activity towards amorphous cellulose. However, it can be suggested that the extent of degradation obtained in the present study may not generate a significant concentration of crystals with the level of perfection/ordering of CNC precursor materials. Hence, the highly concentrated acid will likely dissolve these crystals as the lower level of ordering permits penetration and subsequent hydrolysis, resulting in a similar CNC recovery.
Overall, the improvements (1.2× and 1.3× from Endo-CBD− and Endo-CBD+ treatments, respectively) in CNC recovery from the acid hydrolysis process may have a notable impact in reducing the amount of acid and reagents, energy, and time requirements of subsequent CNC purification processes [32]. The CNC yield (wt % initial feedstock) was calculated to evaluate the impact of mass loss conferred by enzymatic digestion on the yield. There was no significant change in the overall yield from both endoglucanase enzymatically-mediated processes, even with 15.3 ± 1.3 wt % glucan conversion, relative to the mock-treated wood pulp ( Figure 3). Hence, cellulosic precursors for CNCs were not degraded by the enzymes, implying that the sugar release from the endoglucanases was from selective degradation of amorphous cellulose chains. The widely accepted understanding of cellulase activity is that endoglucanases attack amorphous cellulose, unlike cellobiohydrolases, which can degrade crystalline cellulose. It is postulated that endoglucanases have an open groovelike active site to allow binding of internal chains of amorphous regions and decrease the degree of polymerization of cellulose. On the other hand, cellobiohydrolases have a tunnel-shaped active site for processive catalysis of crystalline domains [26,38]. We have previously reported similar results on the overall CNC yields from acid hydrolysis of wood pulp treated with a cellulase cocktail composed of all three classes of enzymes for 2-8 h [32]. However, longer enzymatic treatment, for 10 h, reduced yields, suggesting that the presence of cellobiohydrolases is likely to confer damage to cellulosic precursors for CNCs.
In summary, these results show the potential of the Endo-CBD+ mediated acid hydrolysis process in recovering fermentable glucose and improving the CNC recovery process without compromising the yield. However, further investigations are required to improve the efficiency of endoglucanase hydrolysis to obtain sufficient sugar release for an economically feasible integrated ethanol co-production system. A feasibility study could shed light into how much sugar release and improvement in CNC recovery is required to incorporate an enzymatic treatment with meaningful impact on the process economics of acid hydrolysis.

Enzyme Loading (U/g of pulp)
Endo-CBD-Endo-CBD+ The CNC yield (wt % initial feedstock) was calculated to evaluate the impact of mass loss conferred by enzymatic digestion on the yield. There was no significant change in the overall yield from both endoglucanase enzymatically-mediated processes, even with 15.3 ± 1.3 wt % glucan conversion, relative to the mock-treated wood pulp ( Figure 3). Hence, cellulosic precursors for CNCs were not degraded by the enzymes, implying that the sugar release from the endoglucanases was from selective degradation of amorphous cellulose chains. The widely accepted understanding of cellulase activity is that endoglucanases attack amorphous cellulose, unlike cellobiohydrolases, which can degrade crystalline cellulose. It is postulated that endoglucanases have an open groove-like active site to allow binding of internal chains of amorphous regions and decrease the degree of polymerization of cellulose. On the other hand, cellobiohydrolases have a tunnel-shaped active site for processive catalysis of crystalline domains [26,38]. We have previously reported similar results on the overall CNC yields from acid hydrolysis of wood pulp treated with a cellulase cocktail composed of all three classes of enzymes for 2-8 h [32]. However, longer enzymatic treatment, for 10 h, reduced yields, suggesting that the presence of cellobiohydrolases is likely to confer damage to cellulosic precursors for CNCs.
In summary, these results show the potential of the Endo-CBD+ mediated acid hydrolysis process in recovering fermentable glucose and improving the CNC recovery process without compromising the yield. However, further investigations are required to improve the efficiency of endoglucanase hydrolysis to obtain sufficient sugar release for an economically feasible integrated ethanol co-production system. A feasibility study could shed light into how much sugar release and improvement in CNC recovery is required to incorporate an enzymatic treatment with meaningful impact on the process economics of acid hydrolysis. The yield from the initial feedstock was calculated by taking into account the mass loss due to endoglucanase enzymatic treatment. Columns labelled with different letters indicate means that are statistically different at a 95% confidence level.

Characterization of Endoglucanase-Treated Wood Pulp and Resulting CNCs
The CNCs generated through acid hydrolysis of the mock-treated or endoglucanase-treated wood pulp were then subjected to several analyses to assess their characteristics. In almost all cases, X-ray diffraction, dynamic light scattering, zeta potential, transmission electron microscopy, and differential scanning calorimetry analyses demonstrated that the CNCs generated from the solid residue remaining after treatment with endoglucanases (both Endo-CBD+ and Endo-CBD−) exhibited qualities that were similar to those generated from mock-treated wood pulp. A more detailed discussion of these results is found below.

Crystallinity Index of Endoglucanase-Treated Wood Pulp
•

X-ray diffraction (XRD) analysis
To determine whether the endoglucanase treatment affected the crystallinity index of the wood pulp, XRD analysis was performed (Figure 4). In the case of Endo-CBD+ treatment, a significant increase in crystallinity index, from 77% up to 83%, was observed when the wood pulp was treated with ≥0.8 U/g of pulp enzyme loading. Conversely, for the treatment of wood pulp with Endo-CBD−, the crystallinity index of CNCs derived from all enzyme loadings were identical to the mock-treated control. XRD analysis of Eucalyptus holocellulose, hydrolyzed using a combination of endoglucanase and β-glucosidase, showed an increase in crystallinity index from 72% to 81% [22], which was similar to the results achieved using Endo-CBD+. The findings from our experiments provide further evidence that endoglucanases are able to degrade amorphous regions of cellulose in wood pulp to promote increased crystallinity and CNC yields. However, it should be noted that higher dosages of enzymes are desired to achieve both high oligosaccharide/sugar release and CNC yields.

CNCs Yield (wt % initial feedstock) Enzyme Loading (U/g of pulp)
Endo-CBD-Endo-CBD+ Figure 3. CNC yield of the overall endoglucanase enzymatically-mediated acid hydrolysis process. The yield from the initial feedstock was calculated by taking into account the mass loss due to endoglucanase enzymatic treatment. Columns labelled with different letters indicate means that are statistically different at a 95% confidence level.

Characterization of Endoglucanase-Treated Wood Pulp and Resulting CNCs
The CNCs generated through acid hydrolysis of the mock-treated or endoglucanase-treated wood pulp were then subjected to several analyses to assess their characteristics. In almost all cases, X-ray diffraction, dynamic light scattering, zeta potential, transmission electron microscopy, and differential scanning calorimetry analyses demonstrated that the CNCs generated from the solid residue remaining after treatment with endoglucanases (both Endo-CBD+ and Endo-CBD−) exhibited qualities that were similar to those generated from mock-treated wood pulp. A more detailed discussion of these results is found below.

• X-ray diffraction (XRD) analysis
To determine whether the endoglucanase treatment affected the crystallinity index of the wood pulp, XRD analysis was performed (Figure 4). In the case of Endo-CBD+ treatment, a significant increase in crystallinity index, from 77% up to 83%, was observed when the wood pulp was treated with ≥0.8 U/g of pulp enzyme loading. Conversely, for the treatment of wood pulp with Endo-CBD−, the crystallinity index of CNCs derived from all enzyme loadings were identical to the mock-treated control. XRD analysis of Eucalyptus holocellulose, hydrolyzed using a combination of endoglucanase and β-glucosidase, showed an increase in crystallinity index from 72% to 81% [22], which was similar to the results achieved using Endo-CBD+. The findings from our experiments provide further evidence that endoglucanases are able to degrade amorphous regions of cellulose in wood pulp to promote increased crystallinity and CNC yields. However, it should be noted that higher dosages of enzymes are desired to achieve both high oligosaccharide/sugar release and CNC yields. •

Fourier transformed infra-red (FTIR) analysis
In order to confirm the crystallinity index results obtained through XRD, we also employed a secondary approach based on ATR-FTIR to analyze fibers generated from enzymatic treatment with 1.0 U/g of pulp loading, and the resulting CNCs from acid hydrolysis. The cellulose I crystallinity index was calculated from the ratio of absorbances at 1431 to 900 cm −1 , according to a method proposed by Kuo and Lee [39]. For this calculation, the absorption band at 1430 cm −1 is assigned to CH2 bending in crystalline cellulose; this band shifts to 1420 cm −1 in cellulose II and amorphous cellulose. Conversely, the absorption band at 900 cm −1 , which has been assigned to C-O-C stretching of glycosidic bonds, is a marker of cellulose II and amorphous cellulose, as this band becomes weak and broad in cellulose I [36]. Based on the cellulose I crystallinity index as determined through FTIR, the solid fiber residue remaining after endoglucanase treatment displayed improved crystallinity index compared to the mock-treated wood pulp (Table 1). On the other hand, after acid hydrolysis, the crystallinity index of the CNCs generated from mock-treated (1.6 ± 0.1) or Endo-CBD+-treated (1.0 U/g of pulp) wood-pulp (1.4 ± 0.1) was statistically similar. Interestingly, while XRD analysis indicated that the Endo-CBD− treatment had no effect on crystallinity index of CNCs, ATR-FTIR suggested that the Endo-CBD− treatment actually resulted in a lower crystallinity index. These results highlight the need for a more reliable method of crystallinity index determination. Khatri et al. recently described the use of fluorescent markers to quantify the amount of crystalline and amorphous celluloses, which may serve as a better method for determination of crystallinity index [40].

Enzyme Loading (U/g of pulp)
Endo-CBD-Endo-CBD+ Figure 4. Crystallinity index as determined by XRD analysis. The solid residue recovered after treatment (or mock-treatment) of wood pulp with endoglucanases with (Endo-CBD+) or without (Endo-CBD−) cellulose-binding domains were subjected to XRD analysis to determine crystallinity. Columns labelled with different letters indicate means that are statistically different at a 95% confidence level.

• Fourier transformed infra-red (FTIR) analysis
In order to confirm the crystallinity index results obtained through XRD, we also employed a secondary approach based on ATR-FTIR to analyze fibers generated from enzymatic treatment with 1.0 U/g of pulp loading, and the resulting CNCs from acid hydrolysis. The cellulose I crystallinity index was calculated from the ratio of absorbances at 1431 to 900 cm −1 , according to a method proposed by Kuo and Lee [39]. For this calculation, the absorption band at 1430 cm −1 is assigned to CH 2 bending in crystalline cellulose; this band shifts to 1420 cm −1 in cellulose II and amorphous cellulose. Conversely, the absorption band at 900 cm −1 , which has been assigned to C-O-C stretching of glycosidic bonds, is a marker of cellulose II and amorphous cellulose, as this band becomes weak and broad in cellulose I [36]. Based on the cellulose I crystallinity index as determined through FTIR, the solid fiber residue remaining after endoglucanase treatment displayed improved crystallinity index compared to the mock-treated wood pulp (Table 1). On the other hand, after acid hydrolysis, the crystallinity index of the CNCs generated from mock-treated (1.6 ± 0.1) or Endo-CBD+-treated (1.0 U/g of pulp) wood-pulp (1.4 ± 0.1) was statistically similar. Interestingly, while XRD analysis indicated that the Endo-CBD− treatment had no effect on crystallinity index of CNCs, ATR-FTIR suggested that the Endo-CBD− treatment actually resulted in a lower crystallinity index. These results highlight the need for a more reliable method of crystallinity index determination. Khatri et al. recently described the use of fluorescent markers to quantify the amount of crystalline and amorphous celluloses, which may serve as a better method for determination of crystallinity index [40].

Dynamic Light Scattering (DLS) and Zeta Potential (ZP) Analysis of CNCs Suspensions
Additional analyses were then performed to characterize the suspensions of CNCs produced following acid hydrolysis. First, particle sizes and size distributions were assessed using dynamic light scattering analysis (DLS). For these experiments, CNCs generated through acid hydrolysis of the solid residue derived from treatment of wood pulp with endoglucanases (1.0 U/g of pulp loading) were compared to the mock-treated control. The average hydrodynamic diameter of CNCs from mock-treated wood pulp was 144.6 ± 2.3 nm, with 92.9 ± 5.5% particles in the same distribution ( Table 2). The average diameter of CNCs isolated from wood pulp treated with Endo-CBD− (147.7 ± 7.8 nm) was comparable to the mock-treated samples. This was not observed for CNCs obtained through acid hydrolysis of solid residue generated from Endo-CBD+ treatment, which displayed a smaller average diameter of 125.2 ± 7.4 nm. Bondeson et al. obtained CNCs with an average diameter of 200 to 400 nm from acid hydrolysis of microcrystalline cellulose derived from Norway Spruce [41]. The average hydrodynamic diameter in the present study seems relatively narrow, most likely because the colloid was passed through a 0.45 µm filter. Zeta potential measurement evaluates the density and nature of surface charges. Higher values of zeta potential indicate a higher capacity of dispersion in water, while lower values indicate low dispersion stability. Based on the literature, an absolute value greater than 40 mV reflects good stability of a colloidal suspension [42]. As shown in Table 2, CNCs obtained from acid hydrolysis of endoglucanase-treated wood pulp had zeta potential values ≤ −40 mV, which confirms their colloidal stability. The stability of CNCs dispersions is due to the electrostatic repulsion between the negative sulphate groups (-OSO 3 − ) on the surface of the cellulose [43], which are derived from esterification of the hydroxyl groups present on the cellulose surface. Vasconcelos et al. [42] reported that higher H 2 SO 4 concentrations and prolonged reaction times increased the absolute zeta potential value, favoring the stability of such CNC suspensions.

Transmission Electron Microscopy (TEM) Images of CNCs
TEM micrographs of CNCs derived from acid hydrolyzed mock-treated and endoglucanase-treated (Endo-CBD+ or Endo-CBD−) wood pulp are shown in Figure 5. The CNC particles generated from the three different acid hydrolysis feedstocks appear similar in size, with all particles adopting an expected needle shape form, as reported in other studies [5,42].

Transmission Electron Microscopy (TEM) Images of CNCs
TEM micrographs of CNCs derived from acid hydrolyzed mock-treated and endoglucanasetreated (Endo-CBD+ or Endo-CBD−) wood pulp are shown in Figure 5. The CNC particles generated from the three different acid hydrolysis feedstocks appear similar in size, with all particles adopting an expected needle shape form, as reported in other studies [5,42].

Differential Scanning Calorimetry (DSC)
The heat transition curves generated from DSC study of the CNCs samples are shown in Figure  6. All the curves showed an initial drop between 50 °C and 150 °C, which was likely due to moisture loss. For all three CNCs samples, degradation commenced at roughly 290 °C to 300 °C, as evidenced by the large thermal event observed in the DSC thermograms at this temperature range. Similar DSC profiles were also reported in other studies in the literature [42,44,45]. There was no difference in the thermal transition temperatures between the CNCs generated from the mock and enzyme-treated wood pulp. This further confirms that the endoglucanase treatment in the present study did not result in changes to the CNCs generated through subsequent acid hydrolysis.

Differential Scanning Calorimetry (DSC)
The heat transition curves generated from DSC study of the CNCs samples are shown in Figure 6. All the curves showed an initial drop between 50 • C and 150 • C, which was likely due to moisture loss. For all three CNCs samples, degradation commenced at roughly 290 • C to 300 • C, as evidenced by the large thermal event observed in the DSC thermograms at this temperature range. Similar DSC profiles were also reported in other studies in the literature [42,44,45]. There was no difference in the thermal transition temperatures between the CNCs generated from the mock and enzyme-treated wood pulp. This further confirms that the endoglucanase treatment in the present study did not result in changes to the CNCs generated through subsequent acid hydrolysis.

Conclusions
Enzymatic hydrolysis of wood pulp using endoglucanases has shown potential to improve the yields of CNCs and produce a fermentable monosaccharide by-product stream. In general, the use of endoglucanases with a cellulose-binding domain (Endo-CBD+) produced higher sugar yields and improved crystallinity of wood pulp when compared to the use of endoglucanases without a cellulose-binding domain (Endo-CBD−). The CNCs produced from the solid residue remaining after endoglucanase treatment showed similar characteristics compared to CNCs produced from traditional acid hydrolysis. However, given the relatively low amounts of sugar released from endoglucanase treatment, future research will examine a combination of enzymes, such as endoglucanase, xylanase, and β-glucosidase, to remove both the hemicellulose and completely hydrolyze the amorphous cellulose in the wood pulp to monosaccharides prior to acid hydrolysis. An advanced enzyme system such as this would be expected to improve both the CNC and monosaccharide yield. Nevertheless, this study serves as an important benchmark for future studies, demonstrating that endoglucanase treatment of wood pulp does not negatively impact the quality of CNCs produced from subsequent acid hydrolysis.

Acknowledgments:
The authors would also like to thank Novozymes for providing enzymes and Alberta Pacific Forest Industries Inc. (Al-Pac) for providing wood pulp for this project.

Conflicts of Interest:
The authors declare no conflict of interest.

Conclusions
Enzymatic hydrolysis of wood pulp using endoglucanases has shown potential to improve the yields of CNCs and produce a fermentable monosaccharide by-product stream. In general, the use of endoglucanases with a cellulose-binding domain (Endo-CBD+) produced higher sugar yields and improved crystallinity of wood pulp when compared to the use of endoglucanases without a cellulose-binding domain (Endo-CBD−). The CNCs produced from the solid residue remaining after endoglucanase treatment showed similar characteristics compared to CNCs produced from traditional acid hydrolysis. However, given the relatively low amounts of sugar released from endoglucanase treatment, future research will examine a combination of enzymes, such as endoglucanase, xylanase, and β-glucosidase, to remove both the hemicellulose and completely hydrolyze the amorphous cellulose in the wood pulp to monosaccharides prior to acid hydrolysis. An advanced enzyme system such as this would be expected to improve both the CNC and monosaccharide yield. Nevertheless, this study serves as an important benchmark for future studies, demonstrating that endoglucanase treatment of wood pulp does not negatively impact the quality of CNCs produced from subsequent acid hydrolysis.