Production of sugars from mixed hardwoods for use in the synthesis of sugar fatty acid esters catalyzed by immobilized‐stabilized derivatives of Candida antarctica lipase B

The synthesis of sugar fatty acid esters (SFAEs) from lignocellulosic biomass and oleic acid (C18:1) was catalyzed by immobilized‐stabilized derivatives of Candida antarctica lipase B in a methyl ethyl ketone medium. After steam‐explosion pretreatment of mixed hardwoods and enzymatic hydrolysis at 15%wt solids, xylose and glucose were purified/concentrated to a mass ratio of ~3 to 1. These lignocellulosic sugars were superior to commercial sugars as the carbohydrate source for the esterification reaction in terms of sugar conversions. The highest conversions were obtained using 1.5% w/v of Novozyme 435 (N435, uncoated) as the biocatalyst for the synthesis of SFAEs. Coating the N435 with polyethyleneimine (PEI) prevented enzyme leakage into the reaction medium and produced 35% and 50% higher xylose and glucose conversions to SFAEs, respectively, at the same enzyme loading. After six 24 h reuse cycles with the PEI‐coated N435, xylose conversion decreased by 44%, while a 65% reduction in xylose conversion was observed with the uncoated lipase. Mass spectrometry analysis confirmed the production of xylose and glucose mono‐ and di‐esters. Our purified product presented an emulsion capacity (EC) close to that of a commercial sugar ester and the ECs of the xylose oleate, laurate, and palmitate synthesized in previous studies. © 2023 The Authors. Biofuels, Bioproducts and Biorefining published by Society of Industrial Chemistry and John Wiley & Sons Ltd.


Introduction
L ignocellulosic substrates represent an alternative renewable carbohydrate source for synthesizing several biofuels and bioproducts. 1,2 The successful commercialization of these products depends strongly on a conversion technology that can transform biomass into monomeric sugars efficiently. This procedure may comprise lignocellulosic biomass (LB) pretreatment followed by highsolids enzymatic hydrolysis (HSEH), which is a promising strategy that can generate high sugar yields, operate under mild conditions, and reduce water usage, translating into smaller reactors and lower operating costs. 3,4 Such approaches are generally undertaken in conjunction with downstream purification and concentration steps to provide the target product attributes. 5 In some instances, the HSEH process can be designed to maximize the production of one sugar over the others, i.e. this process can be adapted selectively to produce a larger amount of a specific monomer. 6 For this, it is necessary to use a higher dose of specific enzymes that promote the hydrolysis of a certain LB component (e.g. hemicellulose) in comparison with the enzyme loading for the decomposition of another biomass constituent (e.g. cellulose). Preferentially producing more of a target product during hydrolysis can simplify the downstream purification methods that would otherwise contribute to high operational costs. Moreover, this strategy may even eliminate the necessity for the purification of hydrolysates, which supports the conclusions drawn by Gonçalves et al. 7 who investigated the production and recovery of sugars from lignocellulosics for use in the synthesis of bioproducts. The authors showed that the majority of studies processed crude extracts directly without purification of the hydrolysates. However, in some cases, enzymatic hydrolysates still need to be purified to remove high molecular weight and/or low molecular weight impurities using, for example, ion exchange, chromatographic separation, or ultrafiltration/nanofiltration membranes with targeted molecular weight cut-offs (MWCOs). Operating pressures, temperatures, and the use of diafiltration may all be adapted to recover sugars selectively while removing key contaminants. Purified extracts can be further concentrated as needed, using, for example, reverse osmosis, freeze or spray drying, and evaporative methods such as multiple-effect evaporators.
After the conversion of the LB into monomeric sugars and purification/concentration of extracts (when needed), the resulting sugars are typically used to synthesize other products. The most common goods obtained from a particular monomeric sugar, xylose, but not necessarily exclusively from this sugar, include ethanol, xylooligosaccharides, xylitol, lactic acid, lipids, biohydrogen, and butanol. 7 In this study, we evaluated the use of xylose as a substrate in the enzymatic synthesis of high-value biosurfactants, so-called sugar fatty acid esters (SFAEs). Sugar fatty acid esters have a market value that is at least two orders of magnitude greater than ethanol. [7][8][9] To date, SFAEs have mainly been produced using glucose; 10-23 the use of xylose thus represents a novel use of this sugar.
Sugar fatty acid esters are non-ionic surfactants, 24-26 synthesized through an esterification reaction between a sugar moiety and a fatty acid, 27 where the sugar acts as the acceptor for an acyl group, while the free fatty acid (FFA) or vinyl ester of fatty acid is the donor of the acyl group. 13,28 Gonçalves et al. 29 showed that FFAs were used in 63% of the studies that performed the Candida antarctica lipase B (CALB)-catalyzed synthesis of sugar esters, with lauric, palmitic, and oleic acids accounting for nearly 43% of the occurrences. By using FFAs instead of activated acyl donors (e.g. an ester) the only by-product formed is water, which is safe and can be easily removed from the reaction mixture by using water-adsorbent reagents. 30 This esterification reaction usually takes place in an organic solvent, producing esters with very good surfactant and emulsifier characteristics. 27 Sugar fatty acid esters are odorless, tasteless, and nonirritating surfactants, which are very interesting features for applications in detergents, oral hygiene products, food, pharmaceutical or cosmetic formulations. 31,32 They are also biodegradable and non-toxic to the environment. In addition, many sugar esters suppress bacterial growth, suggesting other potential applications. 11

MCP Gonçalves et al.
Original Article: Production of sugars to synthesize sugar fatty acid esters Chemical or enzymatic catalysis may be used to produce SFAEs. The enzymatic route may represent a more sustainable alternative, as it uses lower operating temperatures/ pressures, achieves high yields, and provides high specificity and selectivity to target products, facilitating product separation. 10,25,33 Nonetheless, the enzymatic synthesis of sugar esters has not been widely researched so far, although sustainable surfactants are in high demand. 11 The enzymatic synthesis of SFAEs plus the use of lignocellulosic sugars as substrates of the esterification reaction would therefore be appropriate, for example, for the exploration of synergies between the production processes of bioethanol and biodiesel for an integrated biorefinery process. This would occur through the use of by-products from these production processes, such as the hemicellulosic fraction of lignocellulosic biomass (a source of xylose), as reagents in the synthesis of other molecules with higher added value (e.g. biosurfactants -SFAEs). By doing so, 'bottlenecks' in the production of these biofuels may be addressed through solutions that are more effective, technically and financially.
Lipases are commonly used as biocatalysts for the synthesis of SFAEs. In particular, microbial lipases from Candida antartica, Candida rugosa, Rhizomucor miehei, Burkholderia sp., or Pseudomonas sp. are commonly used. 27 Candida antarctica produces two different lipases: A and B. Fraction B (CALB) is probably the most widely used hydrolase in the field of biocatalysis, due to its high activity and yields under mild conditions. 34 The commercial CALB Novozyme 435 (N435) was reported in ~80% of the scientific publications investigating the synthesis of SFAEs from 2010 to 2019. 29 N435 is a CALB immobilized by non-covalent linkage and interfacial activation on Lewatit VP OC 1600, a poly-methacrylic acid crosslinked with divinylbenzene. 35 Despite the extensive use of this enzyme preparation, it is difficult to use it multiple times because of enzyme leaching from the support in the presence of substrates/products with surfactant properties. 36 The non-covalent bonds between lipase and the support are disrupted by the presence of fatty acids and/or SFAEs in the reaction medium, leading to lipase loss to the medium. 37 As an alternative, coating the N435 surface with polyethyleneimine (PEI) can promote physical intermolecular crosslinking, with the potential to prevent enzyme release while tailoring some of the biocatalyst properties (e.g. enzyme specificity). 30,[38][39][40] Accordingly, in this research, our main goal is to synthesize SFAEs from renewable resources by using the N435 as catalyst of the esterification reaction. For this, monomeric sugars (mostly xylose) were first produced by HSEH of steampretreated mixed hardwoods. Mixed hardwoods were chosen to be used in this research because they are widely available, low cost, and well-characterized substrates. They have been studied extensively and used to produce lignocellulosic sugars and the properties of mixed hardwoods are similar to those of poplar. A particular advantage is that they have a fairly high xylan content -which is needed to produce the xylose required for the synthesis of the target sugar esters. 41 These carbohydrates were then purified and concentrated to be used as substrates in the synthesis of xylose and glucose oleate. The esterification reaction was catalyzed by the enzyme N435 with and without coating with PEI, in methyl ethyl ketone (MEK) medium. Sugar conversion profiles during the esterification were evaluated, comparing lignocellulosic sugars versus commercial sugars. The effects of using uncoated versus PEI-coated N435 and the effects of N435 loading were also investigated. The operational stability of the enzyme preparations was assessed, along with the emulsion capacities of the synthesized SFAEs, in comparison to a commercial surfactant. The formation of esters was evaluated by mass spectrometry.

Steam-explosion pretreatment of hardwood
Mixed hardwoods were pretreated in a system that includes a customized feeder designed to feed wood chips continuously into a high-pressure hydrolyzer, a discharge screw conveyor, and a blow valve plus cyclone to collect and isolate the pretreated product. Wood chips with a length of 3-5 cm were fed to the system using a conveyor at a rate of 1-3 tons h −1 . Pretreatment conditions were set at 200-208 °C (15-18 bar saturated steam) for 5-7 min. Previous investigations demonstrated that these conditions had less sugar degradation compared to the operation at temperatures >210 °C, thus reducing the formation of degradation products with potential inhibitory effects on fermentation organisms and hemicellulolytic/ cellulolytic enzymes. The steam-exploded wood fiber was used as it was (without washing) and stored in sealed plastic drums or bags until the enzymatic hydrolysis was conducted.

Enzymatic hydrolysis of steam-exploded hardwood
The enzymatic hydrolysis was conducted at 50 °C for 4 h in a 4 L reactor at 15 wt% total solids, with continuous agitation at 60 rpm. The reactor was prefilled with water and steamexploded hardwood, which were mixed and preheated to 50 °C for approximately 30 min before the addition of the enzymes. The amounts of water and biomass added were calculated based on the dry matter of the pretreated hardwood and the target dry matter (solids) content of the mixture. The reaction was initiated by adding 1.75 wt% Cellic HTec3 and 1.0 wt% Cellic CTec3 to the reactor (enzyme doses are relative to the dry weight of the biomass).
Over the course of the hydrolysis, the pH was adjusted to 5.5 by adding 4 mol L −1 NaOH solution. Samples were taken hourly. On completion of the enzymatic hydrolysis, the reactor contents were rapidly cooled to room temperature. The liquid and solid portions were recovered by filtration and refrigerated before further acid hydrolysis (to characterize oligomers of xylose and glucose; see below) and sugar analysis using HPLC (to characterize monomers -soluble xylose and glucose; see below).

Reactor configuration
The stainless-steel 4 L reactor is equipped with a DC motor, and the motor rotation was set clockwise. An external controller (Mixer Clone -Oracle VM Virtual Box, SunOpta Bioprocess Inc., Brampton, Ontario, Canada) was used to control the direction and rotational speed of the motor. Radial-flow impellers (10.16 cm in outside diameter) with four rectangular blades (90° angle) were mounted on a stainlesssteel shaft inside the reactor. The reactor was connected to a recirculating water bath (VWR model 1130A, VWR Scientific Products, McGaw Park, IL, USA) to control the temperature.

Purification and concentration of the crude enzymatic hydrolysis extract
The enzymatic hydrolysate (liquid fraction) containing xylose and glucose was first purified using a particulate filter (5 μm) to remove any of the suspended solids. The filtered extract was further purified to remove soluble lignin-derived compounds using a PES 4 kDa ultrafiltration membrane (Synder filtration VT-2B-6338, Vacaville, CA, USA). The RO Mini membrane system (Tangent Membranes Inc., Grass Valley, CA, USA) was operated at 150 psi, between 10 and 25 °C. The retentate from the ultrafiltration was considered waste and the permeate was the feed material of the subsequent concentration stage using a reverse osmosis membrane (Trisep Turboclean RO-8038-ACM2-46; Wiesbaden, Germany).
The concentration of the extract was initially at the lowest temperature possible (10-20 °C), with the retentate recirculated to the feed tank. The process continued until the feed to the system was exhausted. The permeate of this process was considered waste but the retentate was used in the following drying stage. A spray dryer (Mini B-290; Buchi, Flawil, Sankt Gallen, Switzerland) was used with two fluid nozzles to spray feed solution into the nitrogen carrier gas (45 mmHg). The equipment was set to an inlet temperature of 180 °C and a feed rate of 1 mL min −1 , and 20%wt of maltodextrin was added to the feed material to reduce the stickiness of the spray dried product. The resulting material was then cooled to room temperature and prepared for further acid hydrolysis and sugar analysis (see below).

Enzymatic synthesis of SFAEs
The spray dried product was used as an acyl acceptor in the synthesis of SFAEs (mainly xylose oleate but also glucose oleate). The esterification reaction was conducted between oleic acid (C18:1) and xylose/glucose (at a mass ratio of 3), using MEK as solvent and N435 as the biocatalyst (before and after coating with PEI -see the next subsection). Commercial xylose and glucose were also used as model compounds to compare esterification performance.
The reaction was conducted in the same reactor configuration described above under the following conditions: 60 °C; 350 rpm; 1%, 1.5%, and 1.75% w/v of uncoated Novozyme 435 (10 000 ± 589 TBU g −1 ). The PEI-coated Novozyme 435 (6042 ± 295 TBU g −1 ) activity was adjusted to the same activity provided when the uncoated N435 was used (see below). Samples were collected after 1, 3, 6, and 24 h of reaction, and analyzed according to the methodology presented below. Molecular sieves (3 Å) with an adsorption capacity of 22%wt were used for water removal during the reaction.

N435 PEI-coating
N435 was coated with PEI by adding 100 μL of PEI solution (100 mg mL −1 ) (2 kDa) to a suspension of N435 prepared in 5 mmol mL −1 sodium phosphate buffer, pH 7.0 (biocatalyst/ buffer ratio of 1/10 v/v). The reaction medium was incubated at 25 °C under 150 rpm stirring for 60 min. The PEI-coated N435 was then recovered by vacuum filtration. 42

MCP Gonçalves et al.
Original Article: Production of sugars to synthesize sugar fatty acid esters N435 operational stability in the synthesis of sugar fatty acid esters The synthesis of SFAEs catalyzed by N435 (before and after its coating with PEI) was performed in six successive 24 h batches. Reactions were conducted under the following conditions: 60 °C; continuous stirring (350 rpm); xylose: glucose mass ratio of 3; 1.5% w/v N435 (adjusting to the same provided activity when using the PEI-coated biocatalyst). After each batch, the biocatalyst and molecular sieves were recovered, washed with ethanol and MEK, and subjected to vacuum filtration followed by drying in an oven at 40 °C to remove the remaining solvent. Molecular sieves were removed using a sieve shaker (AS 200; 40 amplitude; Retsch, Haan, Germany) with sieves of 20 and 100 mesh to retain the molecular sieves and the enzyme, respectively. Biocatalysts were then reused in a new esterification reaction. Xylose and glucose conversions were calculated based on HPLC measurements, as described below.

Liquid fraction -acid hydrolysis
Liquid samples collected over the course of the enzymatic hydrolysis of steam-exploded hardwood were centrifuged at 14 000 × g for 10 min (Galaxy 14D Centrifuge 37 001-599; VWR, Radnor, PA). Ten milliliters of the supernatant of each sample and 348 μL of 72% sulfuric acid was added to pressure tubes. The content of the tubes was homogenized and put in the autoclave (Napco 9000-D) for 1 h. After cooling the samples, they were prepared for measurements of the sugar content as described below.

Solid fraction
Solids obtained after (1) steam-explosion pretreatment and (2) enzymatic hydrolysis of pretreated hardwood were analyzed in terms of carbohydrate and insoluble lignin content according to the National Renewable Energy Laboratory (NREL) Laboratory Analytical Procedure (LAP) #42618. 43 The soluble lignin content was measured by a UV-Vis spectrophotometer (Genesys 10-s, Thermo-Scientific). The ash content was determined according to the NREL LAP #42622 44 and the dry matter content (% DM) was assessed according to the NREL LAP #42620. 45

Chromatographic analysis
Samples from the enzymatic hydrolysis of pretreated hardwood were neutralized with calcium carbonate, diluted with deionized water, and filtered through 0.22 μm syringe filters (Millex Syringe filters; Merck Millipore, Tullagreen, Cork, Ireland) into 2 mL HPLC vials (Agilent Technologies, Santa Clara, CA, USA). For samples from the esterification reactions, 2 mL of the supernatant was evaporated, 0.5 mL of water was added to each sample, they were homogenized, and syringe filtered into 2 mL HPLC vials. Samples and calibration standards were analyzed using an Agilent HPLC system (Model series 1200, Agilent Technologies Canada Inc., Mississauga, Ontario, Canada) equipped with a refractive index detector (RID) and a Bio-Rad HPX-87P column (1300 × 7.8 mm × 9 μm, Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) maintained at 80 °C. Deionized water was used as the eluent at 0.6 mL min −1 . The injection volume was 20 μL, with a run time of 30 min. Highperformance liquid chromatography results were tabulated using ChemStation software and the concentrations of xylose and glucose were calculated based on the peak heights (accounting for prior dilutions).

Mass spectrometry analysis
The purification of the crude esterification reaction product was conducted according to the protocol described in Gonçalves et al. 30 The purified product was then characterized by mass spectrometry using a Thermo Scientific Q Exactive Orbitrap MS/MS coupled to a nanoelectrospray ionization source operated in negative ionization mode, across the mass range of 400-1200 m z −1 . Samples were dissolved in methanol, loaded into a borosilicate nano emitter (Thermo Scientific), and analyzed by direct infusion onto the mass spectrometer.

Emulsion capacity assays
Ten milligrams of the purified esterification reaction product (see Gonçalves et al. 30 for details about the purification methodology) was dissolved in 1 mL of water and mixed with 2 mL of kerosene in a test tube. The content of the tube was continuously homogenized for 2 min at room temperature and then left standing for 24 h. After this period, the height of the emulsified region and the height of the total column was measured, and the ECs were calculated according to Equation 1. 46 Sucrose monolaurate was used as a commercial standard in this experiment. Tests were performed in duplicate, and intrasample variations were < 1%.

N435 hydrolytic activity measurements
The protocol to measure the N435 hydrolytic activity in tributyrin is an adaptation of the methodology described by Beisson et al. 47 The method is based on the hydrolysis of a mixture of 1.5 mL of tributyrin, 6 mL of 100 mmol L −1 phosphate buffer, pH 7.3, and 16.5 mL of distilled water at 37 °C for 5 min. The released butyric acid was titrated with 0.02 mol L −1 KOH solution. A unit of tributyrin hydrolysis activity (TBU) is defined as the amount of enzyme that releases 1 μmol of butyric acid per minute under the conditions described above.

Results and discussion
Enzymatic hydrolysis of steam-exploded hardwood and purification/concentration of the hydrolysate extract Table 1 presents the chemical composition of the hardwood (solid fraction) after (1) steam-explosion pretreatment (steam-exploded hardwood -SEH) and (2) enzymatic hydrolysis (post hydrolysis hardwood -PHH). Table 2 summarizes the composition of the crude hydrolysate extract -CHE (liquid fraction) after 4 h of enzymatic hydrolysis and the extract composition after (1) particulate filtration (PF), (2) ultrafiltration (UF), and (3) reverse osmosis (RO). Table 3 shows the powder composition after spray drying (SD) of the RO retentate. Images of the material obtained after each of these steps are shown in Fig. 1.
A primary goal was to obtain a higher concentration of xylose than glucose by using a higher dose of HTec3 than CTec3 in the enzymatic hydrolysis. Although we aimed to produce mainly xylose, CTec3, a cellulase and hemicellulase complex, was also used to increase xylan solubilization, which increased the xylose yield, albeit with the production of a small amount of glucose. Using this enzyme complex, it was possible to obtain a xylose : glucose mass ratio of ~3 in the RO retentate, which was used for spray drying. This method   thus enhanced xylose production while relatively low levels of glucose production were ensured by adjusting the enzyme dose and the hydrolysis time.
The HSEH at 15%wt solids was essential to produce significant quantities of xylose; hydrolysis yields of ~48 and 5% were obtained for xylose and glucose, respectively. Purification methods were used to remove soluble lignin and high molecular weight polyphenols, as shown by the significantly lighter color of the extract after UF (#2 in Fig. 1). Reverse osmosis removed some organic acids and concentrated the extract -with a corresponding increase in sugar and polyphenol concentrations -thus increasing the color intensity of the extract (#3 in Fig. 1). A ~ 54% yield of the total mass of carbohydrates recovered was obtained after spray drying.
It was important to remove as much water as possible from the purified extract because residual water would adversely affect the subsequent use of this material in the synthesis of SFAEs. The esterification reaction is supposed to occur in a nonaqueous environment, 24 and molecular sieves must be added to remove water present in substrates and generated as a by-product of the esterification reaction. Excess water could force the reaction equilibrium towards hydrolysis instead of esterification. 36 Water also influences the activity and selectivity of the enzyme and may lead to its inhibition or inactivation. 48,49 For these reasons, we tried to keep the water at the lowest possible level (a minimal amount of water is necessary to ensure the hydration, stability, and catalytic activity of the enzyme). 50 The spray-dried powder contained about 3% moisture, and roughly 70% xylose (Table 3). Figure 2 illustrates the process of LB conversion into monomeric sugars, the purification and concentration of the carbohydrates produced, and their subsequent use to obtain purified SFAEs through enzymatic esterification.

Comparison of spray-dried lignocellulosic sugars versus commercial sugars as substrates for the esterification
The consumption of xylose and glucose during the esterification reaction is presented in Fig. 3. Here, we compared the use of spray-dried lignocellulosic sugars (SDLS) and commercial sugars (CS) for this reaction, using the uncoated N435 as biocatalyst. After 24 h, the xylose conversion obtained using SDLS was approximately twofold the conversion resulting from the use of CS. The glucose conversion was also higher with SDLS after the same period (1.4-fold). These results confirmed the better performance of the lignocellulosic sugars in comparison with CS as acyl acceptors in the enzymatic synthesis of SFAEs.
When producing SFAEs from beechwood carbohydrates (glucose and xylose), Siebenhaller et al. 11 observed that roughly 0.7% of the glucose was converted to glucoseoctanoate and less than 0.1% of the xylose was used to form xylose-octanoate. This was much less than in the conversions we observed in this study. The authors used the same commercial enzyme (N435), and the reaction was conducted at 50 °C in a rotator with a vortex mixer at 50 rpm for 72 h. The lower temperature and level of mixing may have contributed to the lower conversion.
In another paper, Siebenhaller et al. 10 synthesized SFAEs from beechwood carbohydrates (glucose and xylose) using iCalB (lipase acrylic resin from Candida antarctica, Sigma-Aldrich) as the biocatalyst. The reaction was performed at 50 °C for 70 h in a rotator with a vortex mixer at 50 rpm. In this study, the authors confirmed the formation of glucose and xylose esters with thin-layer chromatography, but sugar conversions to esters were not calculated. Accordingly, our results demonstrated the potential to use lignocellulosic biomass as the sole sugar source to produce high-value SFAEs. Moreover, SDLS appears to be a better substrate than CS in the esterification reaction (in terms of sugar conversions to esters). Nonetheless, we found low conversions for both cases (xylose and glucose; CS and SDLS). The low solubility of sugars in the organic solvent (MEK), the size of the reactor (4 L), the type of agitation used (radial-flow impellers), compounds that may adsorb to the enzyme surface during the esterification reaction and block the enzyme pores (e.g. oligomers of xylan and glucan or the maltodextrin used during the SD stage), among other factors, may help explain these low sugar conversions. On the other hand, such conversions were significantly improved by adjusting the enzyme dose (see below).
Still using N435 but different acyl acceptors and acyl donors, Méline et al. 51 synthesized l-arabinose lauryl mono-and diesters with an l-arabinose conversion of 49%. Reactions were performed at 50 °C over 48 h with 50 mM arabinose, 150 mM vinyl laurate, and 1% w/v N435. Shin et al. 52 synthesized glucose laurate using a supersaturated glucose solution and found a 55% glucose conversion. Reactions were conducted at 40 °C for 12 h with 81 mM glucose, 163 mM vinyl laurate, and 50 mg N435 in 1 mL ionic liquid mixture. Although these conversions are higher than the ones found in this research, the reaction time in the first study was double our reaction time, and in the second study the enzyme dose was much higher than the dose used in this study.
Using a different enzyme source, Teng et al. 53 produced 6-O-palmitoylsucrose at 40 °C, obtaining a greater than 95% conversion of sucrose. The reaction was catalyzed by Lipozyme TL IM in a continuous flow micro-reactor with  Comparison of uncoated N435 versus PEI-coated N435 as biocatalysts of the esterification Figure 4 shows the effects of using uncoated and PEIcoated N435 on xylose and glucose conversions to SFAEs.
Here, we only used SDLS as the carbohydrate source of the esterification reaction. Xylose conversions were ~ 35% higher when using the PEI-coated biocatalyst in comparison with the uncoated N435 after 24 h-reaction. Glucose conversions were also higher with the PEI-coated N435 (by ~50%) after the same period. Until 6 h of esterification, the conversion with the PEI-coated N435 was not statistically different from the conversion obtained with the uncoated N435, according to a Tukey test performed at a 95% confidence interval, except for the glucose conversion after a 1 h reaction.
Uncoated lipases can be desorbed from the hydrophobic support under certain circumstances, resulting in product contamination by the enzyme and biocatalyst inactivation. 30 Coating the N435 with PEI may prevent/ reduce enzyme dissociation from the support because the resulting intermolecularly crosslinked proteins (linked by physical intermolecular adsorption) must now be released simultaneously from their matrix, which is much more difficult. 55,56 This means that coating the enzyme with this cationic polymer can reduce enzyme loss/leakage into the medium.
Another consideration is the generation of a hydrophilic environment when a PEI layer is added to the enzyme/ support surface, which may cause partitioning of some detrimental hydrophobic compounds (e.g. hydrophobic organic solvents) from the enzyme surroundings and thus, improve enzyme performance and support stability. 38 Such consequences might help to explain the higher sugar conversions observed when using the PEI-coated biocatalyst. The use of PEI for coating immobilized enzymes has scarcely been explored so far, although it is a simple way to protect enzymes from hydrophobic compounds, among other advantages mentioned above.

Effect of N435 loading on sugar conversions to sugar fatty acid esters
The effect of N435 loading on glucose and xylose conversions to SFAEs is presented in Fig. 5. The highest conversions (for both sugars) were obtained when using 1.5% w/v N435 (uncoated). Clearly, increasing the amount of enzyme added to the reaction can enhance conversion rates and address the low yields noted earlier when using an enzymatic load of 1% w/v N435.
At the highest enzyme loading tested (1.75% w/v), there was little additional impact on conversions; the difference between conversions at 1.5% and 1.75% is not statistically significant according to a Tukey test performed at a 95% confidence interval. Accordingly, for subsequent studies, we used 1.5% w/v N435.
Despite the benefits provided by the use of a higher enzyme dosage to the esterification reaction medium, the main challenges for enhancing SFAEs' productivity still arise from the low solubility of sugars in organic solvents and the low enzymatic activity in these media, which lead to low reaction rates and low product concentrations. 54 The high cost of enzymes can affect the cost-effectiveness of SFAEs' biosynthesis. However, enzyme costs can be reduced by improving the biocatalyst's operational stability, as discussed in the next section. Using a fed-batch substrate feeding procedure is expected to alleviate some of the limitations with substrate solubility. Adding the substrate and/or enzymes gradually could reduce the capital cost due to a smaller reactor volume, which would translate into lower operating and downstream processing costs with higher product concentrations. 57 Furthermore, using an organic solvent with higher polarity can increase the solubility of the sugar but, at the same time, it might decrease the enzyme activity and produce unwanted hydrolysis side reactions. 58,59 Similarly, increasing the temperature will increase carbohydrate solubility while negatively affecting the enzyme. 60 Further research is therefore needed to establish approaches to address these problems and facilitate the large-scale synthesis of SFAEs.

N435 operational stability in the synthesis of sugar fatty acid esters
The repeated use of N435 (uncoated and PEI coated) for the conversion of sugars (xylose and glucose) to SFAEs was investigated in six successive batches (Fig. 6). When using the uncoated N435, a reduction in conversion of roughly 65% and 60% for xylose and glucose was observed, respectively, after six 24 h cycles. With the PEI-coated N435, a 44% decrease in xylose conversion and a 62% reduction in glucose conversion were observed after six cycles.
Results from Fig. 6 showed a 22% greater decrease in xylose conversion after six cycles with the uncoated N435 in comparison with the reduction observed with the PEIcoated enzyme. On the other hand, in cycles 4 and 5, we found lower xylose conversions when using the PEI-coated N435. However, in these cycles, according to a Tukey test performed at a 95% confidence interval, we identified that the PEI-coated conversions were not statistically significantly different from those found with the uncoated N435. Some factors might help to explain the lower (but not significant) values obtained with the PEI-coated enzyme in these cycles, such as the incomplete removal of compounds adsorbed to the N435 surface during the reuse (e.g. water) or the support solubility in MEK. 61,62 To corroborate our results, for further research, it might be relevant to conduct trials with additional cycles. Overall, the higher reduction in the xylose conversion after six 24 h batches with the uncoated N435 is likely due to the CALB desorption from the matrix of the commercial lipase, as discussed in our previous article. 36 In particular, residual oleic acid and/or SFAEs can lead to CALB desorption from the support due to their surfactant properties. Indeed, at the end of the reuse cycles, we identified a 56 ± 0.4% drop in the N435 hydrolytic activity (A H ) (see above) after incubation in a crude reaction product that also contained unconverted oleic acid, which confirmed N435 leakage from the matrix. Conversely, the A H of the PEI-coated preparation did not decrease after the same process, indicating that the N435 PEI-coating reduced the desorption that has otherwise been reported when using the uncoated biocatalyst. 16,29,40,53,[63][64][65] In a previous study, by the end of five 6 h cycles of xylose laurate synthesis, the A H of the uncoated N435 presented a reduction twofold higher than the A H of the PEI-coated derivative after the biocatalysts' incubation in the crude reaction product. 30 Results from both studies have confirmed that the addition of the PEI layer to the N435 surface reduced enzyme leakage and therefore improved N435 operational stability, facilitating its industrial implementation.
Villalba et al. 66 also coated the N435 surface with PEI and reused this enzyme preparation in the alcoholysis of Camelina sativa oil to ethyl and methyl fatty acid esters after 24 h-reactions. The authors used a 3-1 alcohol to fatty acid molar ratio, 10% w/w biocatalyst, orbital agitation at 200 rpm, 50 °C, and the reaction was conducted in the presence of 40% w/w t-butanol as co-solvent. They found that the conversions obtained with the PEI-coated N435 increased from cycle 1 to cycle 4 and remained constant (~100%) in successive cycles (4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14). Fernandez-Lopez et al. 56 reused the commercial lipase from Rhizomucor miehei (RML) immobilized on octyl-agarose and octyl-glyoxylic beads and coated with PEI and dextran sulfate in the hydrolysis of triacetin. After 10 reaction cycles, there was no change in the enzyme performance (initial rate and yield of diacetine). Furthermore, RML desorption from the support was greatly decreased after the enzyme coating with PEI, a result that is consistent with what was found in this research.
Finally, regarding glucose conversions, there is only a statistically significant difference between PEI-coated and uncoated N435 results in the first cycle (according to a Tukey test performed at a 95% confidence interval), which may indicate that the coating of the N435 surface with PEI did not alter the operational stability of the enzyme for the consumption of this sugar. Nevertheless, as we are using a xylose/glucose mass ratio of 3, the amount of glucose in the reaction medium is so low that the effects of the enzyme PEIcoating on the glucose consumption over the reuse may be impossible to identify with certainty.

Mass spectrometry analysis
Samples of the crude reaction product from the esterification of SDLS with oleic acid were purified (after evaporation of the MEK through distillation) according to the methodology described by Gonçalves et al. 30 Most of the unreacted sugars were removed after this process (91% of xylose and 100% of glucose). The color of the purified product was visually lighter, indicating the likely removal of phenolic compounds.
The purified samples were then characterized by mass spectrometry. Mass spectra of the [M-H] − ion with m/z = 413.29 and 677.53 revealed that we produced mainly xylose mono-and dioleate, as shown in Figs S1 and S2 in the supplementary material. To a lesser extent, we also detected the formation of glucose mono-and dioleate (m/z = 443.30 and 707.54 - Figs S3 and S4). Although N435 has been reported to selectively modify only one hydroxyl group of the sugar 67 (i.e. synthesis of monoesters), the literature has shown that the solvent can change this behavior, with MEK preferentially leading to the formation of diesters instead of monoesters. 68 Here, we found an outcome similar to those previously described by Gonçalves et al. 30,36 when the formation of a Original Article: Production of sugars to synthesize sugar fatty acid esters mixture of SFAEs with different degrees of modification was confirmed. However, such results reported the production of mono, di, and tri-esters, and used CS as the carbohydrate source of the esterification reaction. In this study, using SDLS, we only synthesized mono and di-esters, which may indicate the production of a slightly more purified product, now using a sustainable sugar source in an upscaled environment (from an 80 mL to a 4 L-reactor).

Emulsion capacity assays
The emulsion capacity (EC) of SFAEs synthesized and purified in this study (7.5%; using SDLS in a 4 L-reactor) was comparable with the EC of (1) a commercial sugar ester (10%; sucrose monolaurate) and (2) the xylose oleate synthesized in our previous study 36 (6.25%; using CS in an 80 mL-reactor). These results are also comparable with the ECs of the xylose laurate (12.5%) and xylose palmitate (11.8%) described by Gonçalves et al. 30 As mentioned previously, SFAEs have hydrophobic characteristics due to their fatty acyl chain and hydrophilic hydroxyl groups due to their sugar moiety. A higher proportion of hydrophobic moieties of the molecule leads to a lower EC. 53,69,70 The esters that we synthesized here are composed of a blend of mono-and diesters (as confirmed by mass spectrometry), which means that we may have more than one fatty acid residue per sugar molecule, enhancing the product hydrophobicity and so, reducing its EC. Furthermore, when tri-esters were also produced, 36 the EC was a bit lower than the EC found when only mono-and diesters were synthesized. This outcome confirms that the presence of more oleic acid residues decreases the EC and that the use of SDLS rather than CS in an upscaled environment provided a product with slightly better EC properties.
Other parameters still need to be assessed to confirm the commercial viability of the esters produced in this research. Nonetheless, the EC results are an important starting point towards the future use of the purified xylose oleate (our main product) as an industrial surfactant for a wide range of practical applications, such as skin hydration products in the cosmetic and pharmaceutical sectors, oral-care products, detergents, and foods (ice cream, milk beverages, cake batter, sauces, or dressings). 32,33,71 Conclusions Our primary goal in this research was to maximize xylose production yields and rates relative to those for glucose. By adjusting the enzyme dose, HSEH time, and purification and concentration conditions, we produced xylose and glucose in a mass ratio of ~3. These sugars were then used as substrates in the enzymatic synthesis of SFAEs and presented better esterification performance compared to commercial sugars in terms of sugars conversions to esters.
Xylose and glucose conversions obtained after 24 h esterification using the PEI-coated biocatalyst were increased by ~35 and 50%, respectively, in comparison with the uncoated N435. The improved conversions may be related to: (1) a greater resistance of the coated enzyme to activity reduction after contact with organic solvents and (2) improved support stability due to the partitioning of detrimental hydrophobic compounds from the enzyme surroundings. The addition of the PEI layer to the N435 surface also reduced enzyme desorption and therefore improved the N435 operational stability, facilitating its industrial implementation.
The formation of mono and di-esters using SDLS may indicate a more purified product than the one previously obtained using CS (a blend of mono, di, and tri-esters). The emulsion capacity of the purified SFAEs was close to that of a commercial sugar ester. The presence of fewer oleic acid residues and the use of SDLS in an upscaled environment provided a product with slightly better EC properties than the one produced with CS. Overall, this research represents a pathway towards the sustainable synthesis of high-value SFAEs using lignocellulosic biomass as the sole sugar source and PEI-coated N435 as an enhanced biocatalyst.