Biovalorization of Lignocellulosic Materials for Xylitol Production by the Yeast Komagataella pastoris

: The main goal of this study was to screen different lignocellulosic materials for their ability to support the cell growth of the yeast Komagataella pastoris and the production of xylitol. Several lignocellulosic materials, namely banana peels, brewer’s spent grains (BSGs), corncobs, grape pomace, grape stalks, and sawdust, were subjected to dilute acid hydrolysis to obtain sugar rich solutions that were tested as feedstocks for the cultivation of K. pastoris . Although the culture was able to grow in all the tested hydrolysates, a higher biomass concentration was obtained for banana peels (15.18 ± 0.33 g/L) and grape stalks (14.58 ± 0.19 g/L), while the highest xylitol production (1.51 ± 0.07 g/L) was reached for the BSG hydrolysate with a xylitol yield of 0.66 ± 0.39 g/g. Cell growth and xylitol production from BSG were improved by detoxifying the hydrolysate using activated charcoal, resulting in a fourfold increase of the biomass production, while xylitol production was improved to 3.97 ± 0.10 g/L. Moreover, concomitant with arabinose consumption, arabitol synthesis was noticed, reaching a maximum concentration of 0.82 ± 0.05 g/L with a yield on arabinose of 0.60 ± 0.11 g/g. These results demonstrate the feasibility of using lignocellulosic waste, especially BSG, as feedstock for the cultivation of K. pastoris and the coproduction of xylitol and arabitol. Additionally, it demonstrates the use of K. pastoris as a suitable microorganism to integrate a zero-waste bioreﬁnery, transforming lignocellulosic waste into two high-value specialty chemicals with high market demand.


Introduction
Xylitol is a naturally occurring pentahydroxy sugar alcohol used as a natural food sweetener [1,2]. It can be found in a wide variety of fruits, vegetables, berries, oats, and mushrooms, in amounts of less than 1 wt% [3]. This polyol has sweetness equivalent to sucrose but with a lower glycemic index, which makes it suitable for diabetes patients [4,5]. Due to its anticariogenic properties, xylitol has also been used in the prevention of caries; it is a common component of sugar-free chewing gums [6,7]. Additionally, the nutritional benefits of xylitol promote its use in several dietary supplements [8,9], nutraceuticals [10,11], and confectionary products [12]. Owing to its trend across food and pharmaceutical industries, in 2019 the xylitol global market exceeded 880 million USD, and it is further expected to expand to over 1 billion USD by 2026 [13].
On an industrial scale, xylitol is produced though the catalytic hydrogenation of xylose from xylose-enriched biomass, such as birch wood, in a high energy-consuming and a nonenvironmentally friendly process [3]. Over the last decades, the biotechnological pathway has emerged as an ecological alternative for xylitol production because it involves milder conditions, low carbon emissions, and lower environmental disposal of chemicals [14]. Moreover, the use of lignocellulosic waste as a feedstock for microbial cultivation makes biotechnological processes economically sustainable [15,16].

Characterization of the Materials
The moisture content of the raw materials was determined by subjecting samples (1 g) to a temperature of 105 • C, until constant weight was attained. Dried samples were further subjected to pyrolysis at 575 • C, for 24 h, to evaluate their ash content. The particle size distribution of BSG, grape pomace, grape stalks, and sawdust, after drying at 70 • C until constant weight and then milling, was determined by submitting samples (100 g) to a mechanical sieving shaker, using sieves with pores sizes in the range of 106 to 2360 µm, for 15 min. The fraction collected in each sieve was weighed separately. The protein content of the raw materials was evaluated by elemental analysis, using a Flash EA 1112 CHNS analyzer (Thermo Scientific, Waltham, MA, USA). The protein content was estimated by the following equation: Protein = N × 6.25, (1) where N represents the total nitrogen and 6.25 is a nitrogen-to-protein conversion factor [34].
To determine the hemicellulose, cellulose, and lignin contents of the materials, a wellestablished procedure as described by Sluiter et al. [35] was used. Briefly, samples of dried raw materials (300 mg) were hydrolyzed with sulfuric acid (3 mL H 2 SO 4 72%) at 30 • C, for 1 h, under constant stirring. Afterwards, the samples were diluted with deionized water (84 mL) and further hydrolyzed in an autoclave (121 • C, 1 bar, 20 min).
After hydrolysis, the solutions were vacuum filtered, and the hydrolysis liquors obtained were used for quantification of the released monosaccharides and the acid soluble lignin (ASL). The remaining solid residues were used to quantify acid insoluble lignin (AIL), after being dried (105 • C, 4 h) and subjected to pyrolysis at 575 • C, for 24 h. The AIL (%) content was determined by the following equation: AIL = ((weight dried − weight pyrolyzed )/ODW sample ) × 100 (2) where ODW sample represents the oven-dry weight (g) of each sample calculated by using the moisture content (%) described above. For determination of ASL, hydrolysis liquors samples were diluted in deionized water, and the absorbance was measured at 240 nm. Each sample was analyzed in duplicate, and the ASL (%) content was given by the following equation: where abs is the average of UV-Vis absorbance for each sample, V filtrate corresponds to the volume of hydrolysis liquor (L), ε represents the absorptivity of biomass at 240 nm (L/g cm), and L is the pathlength of the UV-Vis cell (cm). Depending on the raw material, the ε values used were 12 L/g cm for sawdust and grape stalks; 25 L/g cm for grape pomace, banana peels, and BSG; and 30 L/g cm for corncobs [35]. The total amount of lignin (%) was given by the sum of AIL and ASL. For identification and quantification of the released monosaccharides, the pH of hydrolysis liquors was adjusted to values between 5 and 6 by addition of calcium hydroxide. After removal of the calcium sulphate precipitate by centrifugation (13,F, 20 min), the supernatants were used for quantification of monosaccharides by high performance liquid chromatography (HPLC), using a CarboPac PA10 and AminoTrap columns (Dionex, Thermo Scientific, Sunnyvale, CA, USA), equipped with a pulsed amperometric detector (PAD). The samples were analyzed at 25 • C, using NaOH and CH 3 COONa as the eluent, at a flow rate of 1 mL/min. Glucose (Merck KGaA, Darmstadt, Germany), xylose (Sigma-Aldrich, Darmstadt, Germany), galactose (Fluka, Buchs, Switzerland), arabinose (AppliChem Panreac, Germany), mannose (Sigma-Aldrich, Darmstadt, Germany), fructose (Scharlau, Barcelona, Spain), sucrose (Fluka, Buchs, Switzerland), and maltose (VWR, Lisboa, Portugal) were used as standards in a concentration range from 5 to 100 ppm.

Preparation of Hydrolysates
BSG, grape pomace, grape stalks, and sawdust were dried and milled as described above, while banana peels were mashed into a puree, and corncobs were cut into smaller pieces. The materials were then subjected to acid hydrolysis using sulfuric acid in an autoclave (121 • C, 1 bar). The hydrolysis conditions applied to each material are summarized in Table 1, which were defined based on the literature. The resulting mixtures were centrifuged (13,131× g, 4 • C, 30 min), and the solid fraction was discarded. Prior to fermentation, the pH value of the hydrolysates was adjusted to 5 by an addition of Ca(OH) 2 or CaCO 3 , and the formed precipitate was removed by centrifugation (13,131× g, 4 • C, for 20 min). The liquors that were obtained, referred as raw hydrolysates, were used as cultivation media for testing their suitability in supporting cell growth and xylitol synthesis. The raw BSG hydrolysate obtained as described above was concentrated (twofold) by placing the solution in an oven at a temperature of 70 • C, for 24 h, for solvent evaporation.

Detoxification Treatment
Detoxification of the concentrated hydrolysate was performed as described by Kamal et al. [42], with slight modifications. Briefly, powdered activated charcoal (Sigma-Aldrich, Darmstadt, Germany) was mixed with BSG concentrated hydrolysates at a concentration of 2.5% (w/v) and magnetic stirred (800 rpm), for 1 h, at room temperature. The activated charcoal was removed by centrifugation (13,131× g, 30 min at 4 • C), and the detoxified BSG hydrolysates were used as cultivation medium.

Characterization of Hydrolysates
The sugar composition of the raw hydrolysates and the treated BSG hydrolysates was determined by HPLC, as described above.
The hydrolysis efficiency, i.e., the conversion of cellulose, xylan, and arabinan into glucose, xylose, and arabinose (represented as GluR, XylR and AraR, respectively), were obtained using the following equations (adapted from [37]): XylR or AraR (%) = 0.88 × ((C sugar × V)/(xylan or arabinan content × M)) × 100 (8) where 0.90 and 0.88 are the stoichiometric factors for the conversion of glucose to equivalent cellulose, and xylose/arabinose to equivalent xylan/arabinan, respectively. The furfural, 5-hydroxymethylfurfural (HMF), and acetic acid contents in the hydrolysates were also determined by HPLC, using an Aminex HPX-87H (Bio-Rad, Hercules, CA, USA) ion exclusion column, operated at 30 • C, equipped with a UV-Vis detector. The mobile phase was 10 mM H 2 SO 4 and the flow rate 0.6 mL/min. Acetic acid detection was performed at 210 nm, and furfural and HMF were detected at 280 nm.
The total phenolics present in the raw hydrolysates was determined as described by Mussatto et al. [43]. Briefly, the pH of the hydrolysates was raised to pH 12 by an addition of NaOH (6 mol/L), and the optical density was measured at 280 nm in a UV-Vis spectrophotometer. The total phenolics concentration (g/L), TP, was given by the Equations (9) and (10): where Abs LIG280 is the absorbance at 280 nm; C F and C HMF are the concentrations (g/L) of furfural and HMF, respectively, determined by HPLC; and ε F and ε HMF are the extinction coefficients (L/g cm) of furfural (146.85) and HMF (114.00). All measurements were performed in duplicate.

Yeast Strain and Medium
All experiments were performed with Komagataella pastoris DSM 70877 (purchased from DSMZ, Deutsche Sammlung von Mikroorganismen und Zellkulturen). For inocula preparation, K. pastoris was cultivated in medium K, as described by Farinha et al. [44]. For the shake flasks experiments, the different hydrolysates were used as cultivation media, after being filter-sterilized (0.2 µm) and supplemented with ammonium sulfate (13.55 g/L).

Inocula Preparation
The inocula were prepared in 500 mL baffled shake flasks by inoculating the cryopreserved culture (1 mL) in 150 mL medium K supplemented with glucose (20 g/L) and pH adjustment to 5.0 by an addition of ammonium hydroxide (25%, v/v), followed by incubation in an orbital shaker, at 30 • C and 200 rpm, until an optical density of 2.5-3.0 (measured at 600 nm) was reached (around 40 h of cultivation).

Shake Flask Experiments
The yeast cultivation experiments were performed in 250 mL baffled shake flasks with 80 mL of raw hydrolysate/BSG detoxified hydrolysate, supplemented with ammonium sulfate (10 mL, 13.6 g/L) and a 10% (v/v) inoculum prepared as described above. Cultivation was performed in an orbital shaker, at 30 • C and 200 rpm, for 120-168 h. Samples (5 mL) were collected daily for cell growth and pH monitorization. At the end of the experiments, the cultivation broth was used for quantification of biomass, sugars, alcohols, and organic acids. All assays were performed in duplicate.

Analytical Techniques
Culture broth samples were centrifuged at 13,131× g, for 20 min at 4 • C, in order to separate yeast cells. The cell pellet was used to gravimetrically determine the cell dry weight (CDW), after being washed twice with deionized water and lyophilized (−110 • C, 48 h). The cell-free supernatant was used for the quantification of sugars, furfural, HMF, and acetic acid by HPLC, as described above. Xylitol concentration in the cell-free supernatant was also quantified by HPLC, with a CarboPac MA1 column (Dionex), equipped with an electrochemical detector (ED). The analysis was performed at 25 • C using NaOH (612 mM) as the eluent, at a flow rate of 0.4 mL/min. Xylitol (Sigma-Aldrich, Darmstadt, Germany) standard solutions were used in the concentration range of 5-100 ppm.
The yields of xylitol (Y xyl/S ) and arabitol (Y arab/S ) on a substrate basis were calculated using the following equations: Y arab/S = ∆ arabitol /∆ arabinose (12) where ∆ xylitol and ∆ arabitol (g/L) are the concentrations of xylitol and arabitol produced, respectively, and ∆ xylose and ∆ arabinose are the concentrations of xylose and arabinose consumed, respectively.

Statistical Analysis
All experiments were performed at least in duplicate, and results were expressed as mean ± standard deviation (SD).

Characterization of the Raw Materials
The composition of the lignocellulosic materials ( Figure 1) in terms of moisture, ash content, hemicellulose (xylan and arabinan), cellulose, lignin (ASL and AIL), and protein is presented in Table 2. weight (CDW), after being washed twice with deionized water and lyophilized (−110 °C 48 h). The cell-free supernatant was used for the quantification of sugars, furfural, HMF and acetic acid by HPLC, as described above. Xylitol concentration in the cell-free supe natant was also quantified by HPLC, with a CarboPac MA1 column (Dionex), equippe with an electrochemical detector (ED). The analysis was performed at 25 °C using NaOH (612 mM) as the eluent, at a flow rate of 0.4 mL/min. Xylitol (Sigma-Aldrich, Darmstad Germany) standard solutions were used in the concentration range of 5-100 ppm.
The yields of xylitol (Yxyl/S) and arabitol (Yarab/S) on a substrate basis were calculate using the following equations: where Δxylitol and Δarabitol (g/L) are the concentrations of xylitol and arabitol produced, re spectively, and Δxylose and Δarabinose are the concentrations of xylose and arabinose con sumed, respectively.

Statistical Analysis
All experiments were performed at least in duplicate, and results were expressed a mean ± standard deviation (SD).

Characterization of the Raw Materials
The composition of the lignocellulosic materials ( Figure 1) in terms of moisture, as content, hemicellulose (xylan and arabinan), cellulose, lignin (ASL and AIL), and protei is presented in Table 2.  As shown in Table 2, the materials' moisture content was highly variable, according to their origin and nature. The highest contents were found for banana peels (89.7 ± 0.4%), BSG (72.1 ± 0.8%), and grape pomace (58.9 ± 0.6%). The remaining materials had moisture contents below 10%.
Hemicellulose, the most relevant component for xylitol production, was present at higher contents in corncobs (34.1%), sawdust (18.0%), and BSG (17.6%). On the other hand, banana peels, grape pomace, and grape stalks contained only 7.3-7.8% hemicellulose ( Table 2). Cellulose was the major component of lignocellulosic polymers in sawdust (42.8 ± 1.1%), but for the other tested raw materials it represented considerably lower of their dry weight (12.2-28.4%) ( Table 2). Lignin content was similar for all the raw materials, representing 14.1% to 18.6% of the raw materials' dry weight. These values are in accordance with those reported for the same type of lignocellulosic biomass ( Table 2).
Other components found in the lignocellulosic materials included proteins and ash. Low ash contents (<10%) were found for all lignocellulosic materials, except for banana peels, which had a content of 13.4 ± 1.7%. BSG had the highest protein content (27.5 ± 0.5%), while the other lignocellulosic materials had considerably lower contents (<10%) ( Table 2).
The particle size distribution of BSG, grape stalks, grape pomace, and sawdust is represented in Figure 2. Corncobs and banana peels were excluded from this analysis due to their physical characteristics, since corncobs were hard to mill and banana peels were too soft, needing to be reduced to puree. As shown in Figure 2, BSG and sawdust were mainly composed of particles smaller than 1.0 mm (90.4 and 85.0 wt%, respectively). Grape stalks, on the other hand, mostly contained particles larger than 2.36 mm (69.8 wt%). In grape pomace, approximately 63.7 wt% of the particles' size was in the range of 1.0-2.36 mm (Figure 2).
Smaller particle sizes improved hydrolysis efficacy by increasing the surface area to volume ratio. Barakat et al. [45] reported dry fractionation of lignocellulose as an important step and concluded that lignocellulosic materials must have particle sizes between 0.5 and 2 mm to reach a well-accepted level of digestibility. Thus, particle size distribution of BSG, grape pomace, and sawdust are within the values reported, meaning that the effectiveness of dilute acid hydrolysis is increased.

Sugar Composition and Hydrolysis Efficiency
The sugar composition of the raw hydrolysates obtained by the dilute acid hydrolysis of the lignocellulosic materials is shown in Figure 3. Sci. 2021, 11, x FOR PEER REVIEW 8 of 17 Figure 2. Particle size distribution of brewer's spent grain (BSG), grape pomace, grape stalks, and sawdust.
Smaller particle sizes improved hydrolysis efficacy by increasing the surface area to volume ratio. Barakat et al. [45] reported dry fractionation of lignocellulose as an important step and concluded that lignocellulosic materials must have particle sizes between 0.5 and 2 mm to reach a well-accepted level of digestibility. Thus, particle size distribution of BSG, grape pomace, and sawdust are within the values reported, meaning that the effectiveness of dilute acid hydrolysis is increased.

Sugar Composition and Hydrolysis Efficiency
The sugar composition of the raw hydrolysates obtained by the dilute acid hydrolysis of the lignocellulosic materials is shown in Figure 3.  The highest total monosaccharides concentration was found for BSG hydrolysate (21.5 g/L) with similar contents of glucose and arabinose (6.4 ± 2.3 and 6.2 ± 2.1 g/L, respectively) and slightly higher xylose content (8.9 ± 2.6 g/L) ( Figure 3B). On the other hand, only 5.66 g/L monosaccharides were attained from grape pomace, with low concentrations The highest total monosaccharides concentration was found for BSG hydrolysate (21.5 g/L) with similar contents of glucose and arabinose (6.4 ± 2.3 and 6.2 ± 2.1 g/L, respectively) and slightly higher xylose content (8.9 ± 2.6 g/L) ( Figure 3B). On the other hand, only 5.66 g/L monosaccharides were attained from grape pomace, with low concentrations The highest total monosaccharides concentration was found for BSG hydrolysate (21.5 g/L) with similar contents of glucose and arabinose (6.4 ± 2.3 and 6.2 ± 2.1 g/L, respectively) and slightly higher xylose content (8.9 ± 2.6 g/L) ( Figure 3B). On the other hand, only 5.66 g/L monosaccharides were attained from grape pomace, with low concentrations The highest total monosaccharides concentration was found for BSG hydrolysate (21.5 g/L) with similar contents of glucose and arabinose (6.4 ± 2.3 and 6.2 ± 2.1 g/L, respectively) and slightly higher xylose content (8.9 ± 2.6 g/L) ( Figure 3B). On the other hand, only 5.66 g/L monosaccharides were attained from grape pomace, with low concentrations of glucose, xylose, and arabinose (1.7 ± 0.5, 3.1 ± 1.3, and 0.9 ± 0.3 g/L, respectively) ( Figure 3D).
Higher glucose concentrations were achieved from banana peels and grape stalks (8.2 ± 0.6 and 8.0 ± 0.7 g/L, respectively) ( Figure 3A,E), while corncobs and sawdust hydrolysates contained the highest concentrations of xylose (8.2 ± 1.5 and 10.8 ± 1.9 g/L, respectively), which is consistent with the high content of xylans present in these raw materials (Table 2). Moreover, apart from BSG hydrolysate, all the hydrolysates had low contents in arabinose (0.4-1.9 g/L) (Figure 3).
Compared with the literature, the xylose concentration obtained in the different raw hydrolysates was lower than the reported values. For example, Rivera and coworkers [38] reported that 7.2 g/L of xylose was obtained by hydrolysis of grape pomace under similar hydrolysis conditions but with a higher hydrolysis time (125 min). Furthermore, corncob hydrolysate's content in xylose (8.2 ± 1.5 g/L) was considerably lower than the values reported in several studies (21.7-31.2 g/L) [17,46,47], which is probably related to the fact that no mill treatment was applied to corncobs due to their hard structure.
Given the mild conditions employed for hydrolyzing the lignocellulosic materials (dilute acid), the hydrolysis efficiency of cellulose (GluR) was below 30% for all materials (Figure 3). This result is explained by the high crystallinity and polymerization degree of cellulose that requires harsher conditions for an efficient hydrolysis [48]. In contrast, the hydrolysis of hemicellulose was considerably more efficient for most materials. The highest hemicellulose hydrolysis efficiency was reached for BSG, where the xylan and arabinan fractions were almost completely converted into xylose and arabinose, respectively, with XylR and AraR values of 91.0 ± 1.8% and 98.7 ± 2.5% ( Figure 3B). These values are similar to the ones reported by Carvalheiro et al. [49], who demonstrated the recovery of 95% of the BSG's xylan as xylose and 96% of the arabinan as arabinose, under similar hydrolysis conditions.

Content of Hydrolysis Degradation Products
Raw hydrolysates also comprised degradation products, namely furan derivates (furfural and HMF) and aliphatic acids (acetic acid). Furfural was found in the hydrolysates of sawdust, BSG, and corncobs, while HMF was present in all hydrolysates, except grape pomace ( Figure 4).
Nevertheless, the overall concentration of furan derivates was low in all the hydrolysates (<0.5 g/L). These values are in line with the reported values (0.03-0.55 g/L) and is related to the pH adjustment of the raw hydrolysates that probably induced the precipitation and instability of the compounds [39,50]. The highest HMF concentration (0.51 g/L) was found in banana peel hydrolysate, which is probably due to the high glucose content of the material that suffered degradation into furfural during the hydrolysis procedure. On the other hand, sawdust hydrolysate revealed the highest furfural concentration (0.32 g/L), resulting from the degradation of xylose (Figure 3).

Content of Hydrolysis Degradation Products
Raw hydrolysates also comprised degradation products, namely furan derivates (furfural and HMF) and aliphatic acids (acetic acid). Furfural was found in the hydrolysates of sawdust, BSG, and corncobs, while HMF was present in all hydrolysates, except grape pomace (Figure 4). Nevertheless, the overall concentration of furan derivates was low in all the hydrolysates (<0.5 g/L). These values are in line with the reported values (0.03-0.55 g/L) and is related to the pH adjustment of the raw hydrolysates that probably induced the precipitation and instability of the compounds [39,50]. The highest HMF concentration (0.51 g/L) was found in banana peel hydrolysate, which is probably due to the high glucose content of the material that suffered degradation into furfural during the hydrolysis procedure. On the other hand, sawdust hydrolysate revealed the highest furfural concentration (0.32 g/L), resulting from the degradation of xylose (Figure 3). The most relevant degradation product was the acetic acid that was present in all raw hydrolysates at concentrations ranging from 0.47 to 2.90 g/L (Figure 4). Corncobs and sawdust hydrolysates revealed the highest concentrations of acetic acid (2.90 and 2.75 g/L, respectively), which is consistent with the high xylans fraction of these raw materials ( Table 2). In fact, acetic acid is derived from acetyl groups of hemicellulose, and increased concentrations of this degradation product are found in hardwoods [47,51].

Phenolic Compounds Content
Phenolic compounds are also present in all raw hydrolysates at concentrations between 0.19 and 1.85 g/L. Due to grape composition [52], grape pomace and grape stalk hydrolysates had the highest contents of total phenolics (1.85 and 1.62 g/L, respectively). Similarly, banana peel hydrolysate also revealed a total phenolics concentration of 1.23 g/L, which is in accordance with the reported values (1.35 g/L) [36]. Additionally to their natural origin, phenolic compounds can be obtained as a result of the lignin degradation [16]. In fact, the highest lignin fraction was found in the grape pomace and grape stalk materials, which is concomitant with the highest contents of total phenolics.

Cultivation of K. pastoris on Raw Hydrolysates
The raw hydrolysates, after pH adjustment to pH 5, were used as media for the cultivation of K. pastoris to evaluate their suitability to support cell growth and xylitol synthesis ( Table 3). The results show that the highest biomass production was obtained for banana peels (15.18 ± 0.33 g/L) and grape stalks (14.58 ± 0.19 g/L), concomitant with the high glucose consumption (8.20 and 6.68 g/L, respectively) as observed for both assays. The obtained CDW values are slightly higher than the ones reported for the same strain (12.15 and 12.08 g/L) using glucose/xylose mixtures with similar glucose:xylose ratios (15% and 30% xylose) [33]. Table 3. Substrate consumption, cell dry weight (CDW), xylitol and arabitol production, after 120 h of cultivation.
K. pastoris was able to consume xylose from all the raw hydrolysates and converting it into xylitol, although with differing efficiencies ( Table 3). Despite of the high xylose content in the sawdust hydrolysate, only traces of xylitol were detected because of the poor cell growth in the assay. The highest xylose consumption was observed in the assay performed with grape stalks (3.10 ± 1.2 g/L), but only 0.78 ± 0.04 g/L of xylitol were produced, with a Y P/S of 0.25 g/g on a xylose basis. For banana peel and corncob hydrolysates, even though high cell growth was attained (15.18 ± 0.33 and 13.70 ± 0.15 g/L, respectively), xylitol synthesis was low. This might be explained by the low xylose concentration available in banana peel hydrolysate (Table 3) and by the presence of high contents of inhibitory byproducts in corncob hydrolysate (Figure 4).
Despite its low biomass production (5.70 ± 0.94 g/L), BSG hydrolysate resulted in the highest production of xylitol (1.51 ± 0.07 g/L) with a xylitol yield of 0.66 g/g (Table 3). Analogous result was reported by Mussatto et al. [54] where a Y P/S of 0.70 g/g was achieved by the yeast Candida guilliermondii using BSG hydrolysate.
Arabitol was detected at low concentrations (<1.0 g/L) in the experiments carried out with BSG, corncob, grape pomace, and grape stalk hydrolysates, which is concomitant with the consumption of arabinose [55]. The highest arabitol synthesis was noticed for BSG hydrolysate (0.90 ± 0.09 g/L), which had the highest arabinose content (6.2 ± 2.1 g/L). These results revealed that K. pastoris can simultaneously synthesize both polyols from the same feedstock.

Detoxification of BSG Hydrolysate
Aiming to improve cell growth and xylitol production from BSG, several treatments were applied to the raw hydrolysate ( Figure S1). As mentioned above, the pH of the initial hydrolysate ( Figure S1A) was adjusted to pH 5 using Ca(OH) 2 ( Figure S15B), which is the optimum pH for K. pastoris cultivation [56]. Afterwards, the raw hydrolysate was concentrated ( Figure S1C) and detoxified ( Figure S1D) to increase its sugars concentration and decrease the content in toxic hydrolysis degradation products, respectively. The BSG hydrolysates presented a brown coloration, with the color being intensified after the concentration treatment.
As shown in Table 4, adjusting the pH to 5 had no significant effect on the hydrolysate's composition, since only a slight reduction in the concentration of the different components was noticed, namely the loss of xylose and arabinose. As expected, upon concentrating, the concentrations of glucose, xylose, and arabinose increased (11.71 ± 2.16, 15.22 ± 5.55, and 10.30 ± 2.96 g/L, respectively), but as a side effect, the concentration of degradation products also increased (up to 1.14 g/L) ( Table 4). Therefore, to reduce the concentration of toxic compounds, a detoxification treatment using activated charcoal was applied to the concentrated hydrolysate. Activated charcoal was reported as the most effective method for the removal of furan derivates [49]. As shown in Table 4, the detoxification treatment allowed HMF to be completely removed, and furfural concentration was reduced by 96%. Nevertheless, the removal rate of acetic acid was significantly lower (37%), with the concentration obtained (0.72 ± 0.05 g/L) being similar to the value present in the raw hydrolysate (0.78 ± 0.06 g/L). In addition, the detoxification treatment also resulted in losses of sugars, which were reduced by 14-18% (Table 4). These results are comparable to those reported by Carvalheiro et al. [57], namely a high removal of furfural and HMF (92% and 68%, respectively), along with a reduction of 13% of the sugars concentrations. Moreover, the removal of acetic acid was inefficient with a removal rate of 17%.

Cultivation of K. pastoris on Detoxified BSG Hydrolysate
The concentrated BSG hydrolysate promoted biomass production (13.73 ± 0.25 g/L), but not xylitol synthesis (Table 5). In fact, the production of xylitol (1.56 ± 0.07 g/L) was similar to that obtained with the raw hydrolysate (1.51 ± 0.07 g/L), which could be explained by the presence of high concentration of degradation products, especially HMF, which could affect xylitol synthesis [58]. Table 5. Substrate consumption, cell dry weight (CDW), xylitol and arabitol production, yield of xylitol (Y xyl / S ) and yield of arabitol (Y arab / S ) on a substrate basis, after 168 h of cultivation.

Type of BSG Hydrolysate Concentrated Detoxified
Initial glucose (g/L) 11.90 ± 0.19 11.68 ± 2.10 Glucose consumed (g/L) 11 The detoxification treatment, on the other hand, resulted in a significant improvement of the production of biomass and xylitol. Due to the decreased concentration of toxic compounds (Table 4), the biomass production increased to 20.26 ± 0.25 g/L, while xylitol production was improved to 3.97 ± 0.10 g/L. These results are within a range of xylitol production (1.40 and 5.31 g/L) as reported by Araújo et al. [33] for the same K. pastoris strain in batch bioreactor experiments using glucose (35-60 g/L) and xylose (10-35 g/L) mixtures. Analogous improvement was also reported for Candida guilliermondii, where the detoxification of BSG concentrated hydrolysate revealed an increase of the xylitol/xylose yield from 0.37 to 0.55 g/g [54].
The production of arabitol was also noticed, which explains the consumption of arabinose (Table 5). As for xylitol production, the detoxification treatment applied to the concentrated hydrolysate improved arabitol production from 0.43 ± 0.05 to 0.82 ± 0.05 g/L.
Although the production of xylitol was lower than the values reported for processes based on the use of brewer's spent grain as feedstock for microbial cultivation (Table 6), the yield of xylitol on a substrate basis was within the range of those reported for D. hansenii CCMI 941 (0.50 g/g) or C. guilliermondii FTI 20037 (0.55-0.79 g/g). These results show that K. pastoris is a promising candidate for the development of a bioprocess for xylitol production.

Techno-Economic Analysis (TEA)
A techno-economic analysis was performed based on the literature [60] to demonstrate the feasibility of xylitol production from BSG. It is known that BSG represents around 85% of the total by-products generated by the brewing industry and is commonly disposed in landfill. Several studies [54,57,59] have reported the valorization of BSG mainly with a focus on the production of health-promoting food products such as xylitol. In fact, xylitol represents a high-value product with large profit margins, even at a low production scale, with values >5 USD·kg −1 and a market with a compound annual growth rate (CAGR) >6%. Techno economic evaluations revealed that BSG valorization into xylitol presented the lowest production costs (<0.35 USD·kg −1 ) when compared to other chemicals production such as lactic acid or polyhydroxybutyrate.
The present study demonstrated that BSG has the potential to be used as a feedstock for the production of xylitol by K. pastoris. However, several improvements are required to make the process more attractive and economically feasible, namely optimizing the hydrolysis conditions to increase the xylose concentration in the acid hydrolysate, thereby maintaining the inhibitory compounds at low contents. In addition, the process scale-up is mandatory to investigate yield and productivity on a larger scale.

Conclusions
K. pastoris, a nonpathogenic GRAS yeast, demonstrated the ability to synthesize the sugar alcohol xylitol upon cultivation using sugar rich hydrolysates obtained from different lignocellulosic materials. Among the tested hydrolysates, brewer's spent grain hydrolysate resulted in the highest conversion of xylose into xylitol. Additionally, K. pastoris was able to synthesize arabitol based on the consumption of arabinose present in the hydrolysate, a feature that had not been reported previously. These findings show the suitability of using K. pastoris to develop a bioprocess for the valorization of lignocellulosic materials by their conversion into value-added sugar alcohols.