High Gravity Fermentation of Sugarcane Bagasse Hydrolysate by Saccharomyces pastorianus to Produce Economically Distillable Ethanol Concentrations: Necessity of Medium Components Examined

: A major economic obstacle in lignocellulosic ethanol production is the low sugar concentrations in the hydrolysate and subsequent fermentation to economically distillable ethanol concentrations. We have previously demonstrated a two-stage fermentation process that recycles xylose with xylose isomerase to increase ethanol productivity, where the low sugar concentrations in the hydrolysate limit the ﬁnal ethanol concentrations. In this study, three approaches are combined to increase ethanol concentrations. First, the medium-additive requirements were investigated to reduce ethanol dilution. Second, methods to increase the sugar concentrations in the sugarcane bagasse hydrolysate were undertaken. Third, the two-stage fermentation process was recharacterized with high gravity hydrolysate. It was determined that phosphate and magnesium sulfate are essential to the ethanol fermentation. Additionally, the Escherichia coli extract and xylose isomerase additions were shown to signiﬁcantly increase ethanol productivity. Finally, the fermentation on hydrolysate had only slightly lower productivity than the reagent-grade sugar fermentation; however, both fermentations had similar ﬁnal ethanol concentrations. The present work demonstrates the capability to produce ethanol from high gravity sugarcane bagasse hydrolysate using Saccharomyces pastorianus with low yeast inoculum in minimal medium. Moreover, ethanol productivities were on par with pilot-scale commercial starch-based facilities, even when the yeast biomass production stage was included.


Introduction
The two major economic obstacles for ethanol production from lignocellulosic feedstock are the conversion of cellulose and hemicellulose into soluble sugars, glucose and xylose, and the subsequent conversion of xylose to ethanol [1][2][3][4]. Glucose is readily fermented to ethanol by the Saccharomyces species, but most native Saccharomyces species cannot metabolize xylose directly to ethanol [5]. Several non-Saccharomyces yeast species can metabolize xylose to ethanol, except the conversion rates are very slow compared to Saccharomyces glucose-ethanol fermentation rates [6]. Moreover, most acidic-and alkaline-based pretreatment processes yield lignocellulosic hydrolysates that contain high amounts of furan and other phenolic compounds, which are often inhibitory for ethanol production by Saccharomyces cerevisiae [7]. In contrast, S. pastorianus has demonstrated a higher tolerance to hydrolysate inhibitory components [8,9] and wider growth temperatures than reported for S. cerevisiae [10][11][12]. Furthermore, the glucose and fructose transporters in S. pastorianus appear to be unique [13] and do not exhibit diauxic behavior [14]. This allows for glucose and fructose (III) citrate (1 g/L) solution, where the final medium was pH 5.8. The carbon sources were glucose, xylose, or hydrolysate, as specified. When S. pastorianus cultures were fermented on reagent-grade sugars, a rich-medium addition was required in the minimal medium to provide micronutrients. This rich-medium addition (final concentration of 1% YPD medium by volume) was adapted from the yeast-peptone-dextrose (YPD) medium, as described in Gong et al. (1981) and contained 10 g/L yeast extract and 20 g/L peptone without glucose [39]. The chemicals KH 2 PO 4 , (NH 4 ) 2 HPO 4 , MnCl 2 ·4 H 2 O, Zn (CH 3 COO) 2 ·2 H 2 O, H 3 BO 3 , glucose, and citric acid were purchased from ThermoFisher; magnesium sulfate, CoCl 2 ·6 H 2 O, iron (III) citrate, and peptone were purchased from Sigma; yeast extract, xylose, Na 2 MoO 4 ·2 H 2 O, EDTA, and CuCl 2 ·2 H 2 O were purchased from Acros. Xylose isomerase was obtained from Sigma (G4116). All chemicals were ACS certified or ≥99% pure, except the yeast extract and peptone, which were microbiological certified.

E. coli Extract (ECE)
The E. coli cell extract (ECE) production process has been described previously [8]. Briefly, E. coli were grown aerobically a 5-L BioStatB fermenter (Sartorius) at 37 • C in minimal medium containing xylose. An exponential feed profile was used to control the growth rate by maintaining a xylose limited culture [38,40,41]. At the end of the fermentation, the E. coli biomass was killed by heating to 65 • C for 60 min. The heat-killed E. coli cell extract biomass slurries were stored at −20 • C and are referred to as ECE [8,14]. The E. coli fermentations on xylose as the sole carbon source were further characterized and optimized and this information can be found in the Supplemental Materials, Section S1. Additionally, there are several literature reports that describe the chemical composition and components found in E. coli cells [42][43][44][45][46].

Sugarcane Bagasse and the Pre-Treatment Process
Sugarcane bagasse was obtained from the Audubon Sugar Institute at Louisiana State University, St. Gabriel, LA, USA (donated by Dr. Donal F. Day). Table 1 lists the compositional analysis of the sugarcane bagasse [47]. The sugarcane bagasse was pre-treated by the ethanol-potassium hydroxide-hydrogen peroxide (EHO) method and enzymatically hydrolyzed as described in the patent application [48] and Jain et al. (2016) [47]. Briefly, the pretreatment used 200 g bagasse (dry basis, 606 g total, which contained 406 g moisture), 1400 mL of 95% ethanol, 144 mL 35% hydrogen peroxide, and 50 g potassium hydroxide. The bagasse was manually mixed prior to sealing the vessel. The reaction incubated for 1-h at approximately 77 • C in a Parr floor-stand pressure reactor system (Model 4553) with a Parr controller (Model 4843). Mixing was maintained at approximately a revolution per second during the reaction. Four consecutive cold tap water washes were used (7-L each). The pre-treated bagasse was pressed dry to a final moisture content of approximately 76%. The pre-treated bagasse was frozen at −20 • C. Table 1 lists the composition of pretreated bagasse.

Enzyme Hydrolysis Conditions
The standard hydrolysate was prepared as described in Jain and Walker [47]. The hydrolysate with 123 g/L glucose and 44 g/L xylose was prepared using the "improved batch" process that reduced water additions to the reaction mixture. See Supplemental Materials Section S2 for more details. The high gravity hydrolysates (glucose concentration greater than 160 g/L) were prepared using a fed-batch reaction process with an initial addition of 90 g (total mass) pre-treated bagasse (68% or 70% moisture depending on the batch). Each batch contained 12 mL "low activity" CTec2 and 9.5 mL citrate buffer and was incubated at 50 • C, pH 4.8, and 200 rpm. At 24 and 48 h, an additional 90 g of pre-treated bagasse was added for a total of 270 g bagasse per batch. The entire batch was incubated for 72 h.

S. Pastorianus Fermentations
Overnight cultures of S. pastorianus grown aerobically on glucose minimal medium were used to inoculate the shake flasks fermentations on reagent-grade sugars (glucose and xylose) or the sugarcane bagasse hydrolysate (also containing both glucose and xylose). Fermentations were performed at 30 • C in 125 mL shake-flasks with 25 mL working volumes sealed with rubber stoppers that contained a one-way air valve and a sampling syringe, which maintained the anaerobic environment and allowed the carbon dioxide to exit the shake-flask. Cultures were grown in an orbital shaker and agitated at 130 rpm. The initial cell densities in the fermentations were approximately 0.5 OD, corresponding to a dry cell weight (dcw) of 0.25 g dcw/L.
Prior to fermentations, the hydrolysate pH was adjusted to 5.65 with 10 M sodium hydroxide. The hydrolysate replaced the water and sugars in the minimal medium. The basic medium formulation for a 25-mL fermentation is given in Table 2, as also described in Miller et al. (2012) and Gowtham et al. (2014) [8,14]. Reagent-grade sugar control fermentations were run in parallel, where the hydrolysate was replaced by water, glucose, xylose, and 0.25 mL YPD medium. When required, water was used to normalize the initial sugar concentrations between cultures with different media additions. Table 2. Component additions for a 25-mL shake flask fermentation. The water, glucose, and xylose solution volumes required for a reagent-grade sugar fermentation depend on the sugar concentrations in the hydrolysate. For illustration, the target concentrations for the reagent fermentation broth were 186 g/L glucose and 51 g/L xylose.

Analytical Methods
Cell concentrations were measured as cell densities (OD) by absorbance at 600 nm with a Spectronic 20 Genesys instrument (Spectronic Instruments, UK). Sugars and ethanol samples were taken periodically for analysis throughout the fermentations (1.5 mL samples). These samples were centrifuged to separate the pellet and supernatant. The supernatant was collected and frozen for later analysis. Glucose and xylose concentrations were determined using a Yellow Springs Instruments 2900 biochemistry analyzer [8]. An Agilent 7890A GC system was used to measure ethanol concentrations [14]. The supernatant and calibration standards were filtered through a 0.2 µm syringe filter prior to analysis.

Statistical Analyses
Statistical analysis was performed using the software JMP Pro 10 (SAS Institute Inc., Cary, NC), where replicate cultures or reactions were conducted. Data were analyzed using the GLM procedure (p ≤ 0.05) and least squares mean differences with Student's t-tests. Statistical analysis of responses included glucose consumption rates, final ethanol concentrations, ethanol yields, xylose conversions, and ethanol productivities. The effectors were media additives, ECE, and xylose isomerase. Standard errors are reported initially for concentrations, consumption rates, yields, and productivities. Error bars on graphs represent standard deviations, as it was determined that error bars that represented standard error were too small to see in many cases.

Results
Our previous work has demonstrated that S. pastorianus can grow under high gravity fermentation conditions and produce high concentrations of ethanol on reagent-grade sugars (>150 g/L) [14,49] as part of a two-stage fermentation process. An updated schematic of the two-stage fermentation process is included in the Supplemental Materials ( Figure S3). Additionally, S. pastorianus are able to consume both fructose and glucose at enhanced rates when xylose isomerase is present in the medium, which resulted in the near theoretical conversion of glucose to ethanol [8,14]. Further, the recycling of the xylose was demonstrated as ECE increased the glucose consumption and ethanol productivity for low sugar concentration corn stover hydrolysate by S. pastorianus [8]. However, due to low glucose concentration in the corn stover hydrolysate (less than 30 g/L glucose), high gravity fermentation conditions were not evaluated at that time, nor the potential inhibitor effects. The Supplemental Materials describe the screen experiments used to develop the fed-batch hydrolysis reaction to generate hydrolysate with high gravity sugar concentrations ( Figure S2, Section S2). As these screen experiments were conducted in parallel with the medium component analysis, the glucose and xylose concentrations in the hydrolysate are clearly indicated.

Medium Component Effects on Ethanol Productivity
The minimal medium used in the yeast fermentations was originally developed for E. coli high cell density fermentations, and likely contains components that are not necessary for yeast cultured on hydrolysates [8,14,37]. Consequently, the first part of this study was to determine which minimal medium components were necessary for ethanol production using sugarcane bagasse hydrolysates that were pre-treated using the EHO method. One of the goals of this component analysis is to reduce the dilution of the fermentation medium by the water that accompanies these components, as all of the media components are prepared separately and used from stock solutions.
Sources of water in the original minimal medium recipe comes from the four major component stock solutions that are prepared separately due to issues of precipitation if autoclaved together [37]. Table 2 lists these four major component stock solutions and volumes added per shake flask culture. For the yeast ethanol fermentations, the hydrolysate and ECE replace the water and sugars in the standard medium formula. Therefore, reducing or eliminating one of these four medium components would allow for more hydrolysate to be added to the medium, and thus potentially increase the final ethanol concentration.
To assess the elimination of these dilution sources, S. pastorianus were cultured in media formulated using sugarcane bagasse hydrolysate containing 65 g/L glucose and 20 g/L xylose. The yeast-peptone-dextrose (YPD) medium additions were not included, as the hydrolysate and ECE contain sufficient micronutrients to support yeast growth and ethanol production. The glucose and ethanol profiles for the medium component effect experiments are shown in Figure 1. The initial glucose concentration in the fermentation broth was 51.5 ± 0.02 g/L. By calculation, the xylose concentration was about 16 g/L. The matrix effect for the xylose measurement due to the ECE resulted in non-meaningful values, which for this study, xylose quantification was determined to be non-critical. The glucose consumption rate was fastest in the complete medium, where the culture with only the Mg solution addition was the second-fastest. The ethanol concentration profiles also showed the highest productivity for the complete medium, where the culture with the Mg solution had only a slightly lower ethanol productivity. Interestingly, the iron and trace solution addition cultures proceeded no better than the culture without any additions (None) with respect to glucose consumption. Yet these three cultures (None, trace, and iron) generated similar levels of ethanol as the other cultures, although much slower, indicating S. pastorianus have the capability to generate ethanol without additives from hydrolysate when ECE and xylose isomerase are supplemented.  Figure 1. The initial glucose concentration in the fermentation broth was 51.5 ± 0.02 g/L. By calculation, the xylose concentration was about 16 g/L. The matrix effect for the xylose measurement due to the ECE resulted in non-meaningful values, which for this study, xylose quantification was determined to be noncritical. The glucose consumption rate was fastest in the complete medium, where the culture with only the Mg solution addition was the second-fastest. The ethanol concentration profiles also showed the highest productivity for the complete medium, where the culture with the Mg solution had only a slightly lower ethanol productivity. Interestingly, the iron and trace solution addition cultures proceeded no better than the culture without any additions (None) with respect to glucose consumption. Yet these three cultures (None, trace, and iron) generated similar levels of ethanol as the other cultures, although much slower, indicating S. pastorianus have the capability to generate ethanol without additives from hydrolysate when ECE and xylose isomerase are supplemented. Since the Mg solution addition culture performed well and was only slightly slower than the complete medium culture, the effect of the phosphate buffer solution was investigated. A second set of fermentations was conducted where the phosphate buffer concentration was varied while holding the Mg concentration constant. The glucose profiles are shown in Figure 2 for the phosphate buffer ranging experiment. In this case, the initial glucose concentration in the fermentation broth was 45 g/L glucose, as this was a different batch of sugarcane hydrolysate. The glucose consumption rate was highest in the complete medium, where the Mg solution only culture had the lowest glucose consumption rate. The culture with the Mg solution plus 1× phosphate buffer had glucose consumption rates equal to the culture in the complete medium. Likewise, the ethanol production rates were observed to be the same for the complete medium and Mg solution plus 1× Buffer cultures, whereas the Mg solution only and the Mg solution plus 0.1× Buffer were significantly lower. Interestingly, the xylose profiles for the four fermentations are not significantly different. The matrix effect caused by the hydrolysate on the xylose measurement makes precise xylose concentration measurement difficult; however, the trends are clear, and the xylose was consumed only slightly during the fermentation, as expected, due to minimal xylose isomerase activity at the fermentation temperature and pH conditions. Thus, the Mg solution plus 1× Buffer was determined to be sufficient for high glucose consumption and ethanol productivities on sugarcane bagasse hydrolysate. Since Since the Mg solution addition culture performed well and was only slightly slower than the complete medium culture, the effect of the phosphate buffer solution was investigated. A second set of fermentations was conducted where the phosphate buffer concentration was varied while holding the Mg concentration constant. The glucose profiles are shown in Figure 2 for the phosphate buffer ranging experiment. In this case, the initial glucose concentration in the fermentation broth was 45 g/L glucose, as this was a different batch of sugarcane hydrolysate. The glucose consumption rate was highest in the complete medium, where the Mg solution only culture had the lowest glucose consumption rate. The culture with the Mg solution plus 1× phosphate buffer had glucose consumption rates equal to the culture in the complete medium. Likewise, the ethanol production rates were observed to be the same for the complete medium and Mg solution plus 1× Buffer cultures, whereas the Mg solution only and the Mg solution plus 0.1× Buffer were significantly lower. Interestingly, the xylose profiles for the four fermentations are not significantly different. The matrix effect caused by the hydrolysate on the xylose measurement makes precise xylose concentration measurement difficult; however, the trends are clear, and the xylose was consumed only slightly during the fermentation, as expected, due to minimal xylose isomerase activity at the fermentation temperature and pH conditions. Thus, the Mg solution plus 1× Buffer was determined to be sufficient for high glucose consumption and ethanol productivities on sugarcane bagasse hydrolysate. Since the hydrolysate sugar concentrations used in these medium addition studies are too low to support fermentations that would reach over 70 g/L ethanol (high gravity), parallel work was conducted in-house to increase the sugar concentrations in the hydrolysate (details in the Supplemental Materials, Section S2).
Fermentation 2020, 6, x FOR PEER REVIEW 7 of 15 the hydrolysate sugar concentrations used in these medium addition studies are too low to support fermentations that would reach over 70 g/L ethanol (high gravity), parallel work was conducted inhouse to increase the sugar concentrations in the hydrolysate (details in the Supplemental Materials, Section S2). As shown in Table 2, the addition of ECE to the fermentation medium has a significant dilution effect. Thus, increasing the biomass concentration of the E. coli in the ECE would reduce the volume of ECE required (work described in the Supplemental Materials, Section S1). In parallel, the necessity for ECE in the more concentrated hydrolysate was assess for the on the hydrolysate containing 65 g/L glucose and 20 g/L xylose. This resulted in a fermentation broth with only 50 g/L glucose. Figure 3 shows the glucose and ethanol profiles for S. pastorianus on sugarcane bagasse hydroxylate with and without the ECE addition. Xylose isomerase was included as it is not considered to be a nutrient source. Despite the significant dilution of the sugars by the ECE addition, the ethanol productivity was significantly better for the culture with ECE and xylose isomerase additions compared to the cultures without ECE (fermentation times less than 48 h versus 300 h). As has been previously observed in the low sugar and reagent-grade sugar fermentations [8,14], the final ethanol yields are approximately equal. Moreover, ethanol fermentation optimization studies for the amount of ECE and xylose isomerase, conducted in reagent-grade sugars, can be found in the Supplemental Materials (Section S4). These data together indicate that significant optimization of the fermentation process with ECE and xylose isomerase is still possible.

High Gravity Fermentations Using Sugarcane Bagasse Hydrolysate
The first high gravity hydrolysate was prepared using an improved hydrolysis batch reaction process. This hydrolysate contained 123 g/L glucose and 44 g/L xylose, which resulted in fermentation media of approximately 94 g/L glucose and 31 g/L xylose. Due to the limited size of the batch hydrolysate reaction (<17 mL), the hydrolysate culture was conducted as a singleton, as a single shake flask culture can hold up to 21.7 mL hydrolysate (see the Supplemental Materials for the reaction conditions; Table S3). As this hydrolysate had significantly higher sugar concentrations than had been previously obtained, the effects of inhibitors was the main concern. Single cultures of each condition were conducted in parallel using the hydrolysate and reagent-grade sugars. Figure 4 shows the glucose, xylose, and ethanol profiles for the "improved" hydrolysate. The glucose and ethanol profiles are similar for the hydrolysate and reagent-grade sugar fermentations, which indicates that inhibitory effects were minimal even at these higher sugar concentrations. The xylose conversion was higher in the reagent-grade sugar fermentations, which has been previously observed [8]. However, the glucose and xylose concentrations were insufficient to achieve 70 g/L ethanol, even if theoretical yields were realized [50]. As shown in Table 2, the addition of ECE to the fermentation medium has a significant dilution effect. Thus, increasing the biomass concentration of the E. coli in the ECE would reduce the volume of ECE required (work described in the Supplemental Materials, Section S1). In parallel, the necessity for ECE in the more concentrated hydrolysate was assess for the on the hydrolysate containing 65 g/L glucose and 20 g/L xylose. This resulted in a fermentation broth with only 50 g/L glucose. Figure 3 shows the glucose and ethanol profiles for S. pastorianus on sugarcane bagasse hydroxylate with and without the ECE addition. Xylose isomerase was included as it is not considered to be a nutrient source. Despite the significant dilution of the sugars by the ECE addition, the ethanol productivity was significantly better for the culture with ECE and xylose isomerase additions compared to the cultures without ECE (fermentation times less than 48 h versus 300 h). As has been previously observed in the low sugar and reagent-grade sugar fermentations [8,14], the final ethanol yields are approximately equal. Moreover, ethanol fermentation optimization studies for the amount of ECE and xylose isomerase, conducted in reagent-grade sugars, can be found in the Supplemental Materials (Section S4). These data together indicate that significant optimization of the fermentation process with ECE and xylose isomerase is still possible.

High Gravity Fermentations Using Sugarcane Bagasse Hydrolysate
The first high gravity hydrolysate was prepared using an improved hydrolysis batch reaction process. This hydrolysate contained 123 g/L glucose and 44 g/L xylose, which resulted in fermentation media of approximately 94 g/L glucose and 31 g/L xylose. Due to the limited size of the batch hydrolysate reaction (<17 mL), the hydrolysate culture was conducted as a singleton, as a single shake flask culture can hold up to 21.7 mL hydrolysate (see the Supplemental Materials for the reaction conditions; Table S3). As this hydrolysate had significantly higher sugar concentrations than had been previously obtained, the effects of inhibitors was the main concern. Single cultures of each condition were conducted in parallel using the hydrolysate and reagent-grade sugars. Figure 4 shows the glucose, xylose, and ethanol profiles for the "improved" hydrolysate. The glucose and ethanol profiles are similar for the hydrolysate and reagent-grade sugar fermentations, which indicates that inhibitory effects were minimal even at these higher sugar concentrations. The xylose conversion was higher in the reagent-grade sugar fermentations, which has been previously observed [8]. However, the glucose and xylose concentrations were insufficient to achieve 70 g/L ethanol, even if theoretical yields were realized [50].  Using the hydrolysate prepared by the fed-batch hydrolysis reaction process, four replicate fermentations for hydrolysate and reagent-grade sugars were conducted using a common inocula for each pair. The sugar concentrations in the hydrolysate was 201 and 55 g/L xylose (the Supplemental Materials describe the steps used to develop this process in greater detail). The hydrolysate sugar concentrations in this material have the potential to generate over 70 g/L ethanol in fermentation [50]. The initial glucose concentrations in the fermentation broths were 186 ± 1.3 and 183 ± 12 g/L for the hydrolysate and reagent-grade sugars, respectively. Figure 5 shows the glucose and ethanol profiles for these fermentations. The final ethanol concentrations were 89.2 ± 0.5 and 84.3 ± 1.4 g/L from the hydrolysate and reagent-grade sugars, respectively. The ethanol yields calculated based on glucose   Using the hydrolysate prepared by the fed-batch hydrolysis reaction process, four replicate fermentations for hydrolysate and reagent-grade sugars were conducted using a common inocula for each pair. The sugar concentrations in the hydrolysate was 201 and 55 g/L xylose (the Supplemental Materials describe the steps used to develop this process in greater detail). The hydrolysate sugar concentrations in this material have the potential to generate over 70 g/L ethanol in fermentation [50]. The initial glucose concentrations in the fermentation broths were 186 ± 1.3 and 183 ± 12 g/L for the hydrolysate and reagent-grade sugars, respectively. Figure 5 shows the glucose and ethanol profiles for these fermentations. The final ethanol concentrations were 89.2 ± 0.5 and 84.3 ± 1.4 g/L from the hydrolysate and reagent-grade sugars, respectively. The ethanol yields calculated based on glucose Using the hydrolysate prepared by the fed-batch hydrolysis reaction process, four replicate fermentations for hydrolysate and reagent-grade sugars were conducted using a common inocula for each pair. The sugar concentrations in the hydrolysate was 201 and 55 g/L xylose (the Supplemental Materials describe the steps used to develop this process in greater detail). The hydrolysate sugar concentrations in this material have the potential to generate over 70 g/L ethanol in fermentation [50]. The initial glucose concentrations in the fermentation broths were 186 ± 1.3 and 183 ± 12 g/L for the hydrolysate and reagent-grade sugars, respectively. Figure 5 shows the glucose and ethanol profiles for these fermentations. The final ethanol concentrations were 89.2 ± 0.5 and 84.3 ± 1.4 g/L from the hydrolysate and reagent-grade sugars, respectively. The ethanol yields calculated based on glucose were approximately 48% and 46% from the hydrolysate and reagent-grade sugar, respectively. These yields includes the glucose used to generate the yeast biomass as the inocula were 0.25 g dcw/L. The ethanol productivity from the hydrolysate was 1.62 ± 0.01 g/L·h, which also includes the time required to grow the cells. were approximately 48% and 46% from the hydrolysate and reagent-grade sugar, respectively. These yields includes the glucose used to generate the yeast biomass as the inocula were 0.25 g dcw/L. The ethanol productivity from the hydrolysate was 1.62 ± 0.01 g/L·h, which also includes the time required to grow the cells. To obtain a uniform hydrolysate, the hydrolysates from four fed-batch reactions were combined. The reagent-grade sugar fermentations were run, pairwise in parallel, with the same inoculum. The legend for both panels is shown in panel (A). All data points are shown for the four replicates. The glucose consumption rates and ethanol productivities are significantly different (p ≤ 0.05); however, the ethanol yields are not significantly different (p > 0.05).

Effect of ECE and Xylose Isomerase on the Hydrolysate Fermentation Kinetics
The need for the ECE and xylose isomerase supplements was re-examined for the high gravity sugarcane bagasse hydrolysate fermentations. In particular, the goal of this set of fermentations was to confirm that the ECE and xylose isomerase additions still increased the glucose consumption rate and the ethanol productivity under high gravity fermentation conditions. Previous studies on reagent-grade sugars have demonstrated that both the ECE and xylose isomerase have significant effects on ethanol productivity in high gravity fermentations (Supplemental Materials, Section S4), but this effect has not be demonstrated on high gravity hydroxylates. The enzyme-treated hydrolysate used for this comparison was generated by the fed-batch process; however, from a different batch than used in Figure 5. The hydrolysate used to examine the effects of the ECE and xylose isomerase had sugar concentrations of 162 g/L glucose and 45 g/L xylose, which resulted in a fermentation media with 148.2 ± 0.14 g/L glucose. Since the goal of these fermentations was only to demonstrate if a significance difference existed due to the ECE and xylose isomerase for the glucose consumption rate and ethanol productivity, the sample analysis was streamlined to two times. The time profiles from Figures 4 and 5 were used to selected these two time points, with the objective of capturing the fermentation prior to completion. The sample times selected accounted for the lower initial sugar concentrations (148 vs. 186 g/L), and the times 36 and 45 h were selected. Figure 6 shows the glucose and ethanol profiles for these cultures, and clearly demonstrate that the ECE and xylose isomerase supplements significantly increased the glucose consumption rate and ethanol productivity (p ≤ 0.05). The ethanol concentration reached 82.4 ± 0.33 g/L ethanol in less than 45 h when ECE and xylose isomerase were present versus 59.3 ± 0.02 g/L for the cultures without additions.The ethanol productivity for the cultures with ECE and xylose isomerase were 1.89 ± 0.01 g/L·h with a yield of 56.2% ± 0.3% from glucose, which includes the time required to grow the cells ). Data represent the combined data from four replicate cultures per culture condition; all data points shown. To obtain a uniform hydrolysate, the hydrolysates from four fed-batch reactions were combined. The reagent-grade sugar fermentations were run, pairwise in parallel, with the same inoculum. The legend for both panels is shown in panel (A). All data points are shown for the four replicates. The glucose consumption rates and ethanol productivities are significantly different (p ≤ 0.05); however, the ethanol yields are not significantly different (p > 0.05).

Effect of ECE and Xylose Isomerase on the Hydrolysate Fermentation Kinetics
The need for the ECE and xylose isomerase supplements was re-examined for the high gravity sugarcane bagasse hydrolysate fermentations. In particular, the goal of this set of fermentations was to confirm that the ECE and xylose isomerase additions still increased the glucose consumption rate and the ethanol productivity under high gravity fermentation conditions. Previous studies on reagent-grade sugars have demonstrated that both the ECE and xylose isomerase have significant effects on ethanol productivity in high gravity fermentations (Supplemental Materials, Section S4), but this effect has not be demonstrated on high gravity hydroxylates. The enzyme-treated hydrolysate used for this comparison was generated by the fed-batch process; however, from a different batch than used in Figure 5. The hydrolysate used to examine the effects of the ECE and xylose isomerase had sugar concentrations of 162 g/L glucose and 45 g/L xylose, which resulted in a fermentation media with 148.2 ± 0.14 g/L glucose. Since the goal of these fermentations was only to demonstrate if a significance difference existed due to the ECE and xylose isomerase for the glucose consumption rate and ethanol productivity, the sample analysis was streamlined to two times. The time profiles from Figures 4 and 5 were used to selected these two time points, with the objective of capturing the fermentation prior to completion. The sample times selected accounted for the lower initial sugar concentrations (148 vs. 186 g/L), and the times 36 and 45 h were selected. Figure 6 shows the glucose and ethanol profiles for these cultures, and clearly demonstrate that the ECE and xylose isomerase supplements significantly increased the glucose consumption rate and ethanol productivity (p ≤ 0.05). The ethanol concentration reached 82.4 ± 0.33 g/L ethanol in less than 45 h when ECE and xylose isomerase were present versus 59.3 ± 0.02 g/L for the cultures without additions.The ethanol productivity for the cultures with ECE and xylose isomerase were 1.89 ± 0.01 g/L·h with a yield of 56.2% ± 0.3% from glucose, which includes the time required to grow the cells from the 0.05 g dcw/L inoculum. In comparison, the cultures without the additives have ethanol yields from glucose of 51.2% ± 0.04% and ethanol productivities of 1.29 ± 0.001 g/L·h. The greatest contribution the ECE and xylose isomerase provides is towards improved ethanol productivity.
Fermentation 2020, 6, x FOR PEER REVIEW 10 of 15 from the 0.05 g dcw/L inoculum. In comparison, the cultures without the additives have ethanol yields from glucose of 51.2% ± 0.04% and ethanol productivities of 1.29 ± 0.001 g/L·h. The greatest contribution the ECE and xylose isomerase provides is towards improved ethanol productivity.

Discussion
As we have demonstrated previously, on reagent-grade sugars, the ECE and xylose isomerase significantly increase glucose consumption rates and ethanol productivities [8,14]. As expected, the reagent-grade sugar fermentations consumed the glucose and produced ethanol within about 35 h ( Figure 5), while the hydrolysate fermentations took approximately 55 h to completely consume the glucose and reach maximum ethanol levels. These data indicate that the inhibitor effects were minimal and S. pastorianus in the system was robust; whereas for native S. cerevisiae, tolerance is a sufficient concern that genomic approaches have been taken to try to solve the lack of inhibitor tolerance issue [23,[27][28][29]51]. Another positive attribute of this process is that the required inoculum was only 0.25 g dcw/L (0.5 OD), which represents a very low inoculum within the biofuels community [29,52]. Further, the ethanol productivity reported in this study includes the growth phase as well as the production phase without the addition of yeast extract or peptone. Moreover, this process was competitive with starch-based processes [53]. Further, several recombinant yeast studies that have reported using high density (20 OD) inoculum [25,26] did not include in the ethanol yields or productivities the material and time required to generate the inoculum. In many other ethanol-from-lignocellulosic studies, it is common to use high inoculum, supplementing the medium with yeast extract and peptone [29,32,51,52], corn steep liquor [35], or other nutrients [54]. It is not always clear if the inocula or these added carbon/nutrient sources are included in the total carbon balances when yields are reported. It was surprising, in this work, that ethanol yield calculated based solely on glucose exceeds the theoretical yield of 51% [57]; however, it is very likely other carbon sources are being used for ethanol production derived from the hydrolysate or the ECE. Even though the final ethanol yields for the S. pastorianus were cultured on hydrolysate with and without the ECE and xylose isomerase to determine if the species was significant to high gravity hydrolysate fermentation outcomes. Both ECE and xylose isomerase (ECE & XI, ); No ECE or xylose isomerase (None, ). Duplicate cultures for both conditions were conducted. Error bars represent the standard deviation, which have been included on all data points, but are difficult to see in many cases due to the relative size of the bars to symbols.

Discussion
As we have demonstrated previously, on reagent-grade sugars, the ECE and xylose isomerase significantly increase glucose consumption rates and ethanol productivities [8,14]. As expected, the reagent-grade sugar fermentations consumed the glucose and produced ethanol within about 35 h ( Figure 5), while the hydrolysate fermentations took approximately 55 h to completely consume the glucose and reach maximum ethanol levels. These data indicate that the inhibitor effects were minimal and S. pastorianus in the system was robust; whereas for native S. cerevisiae, tolerance is a sufficient concern that genomic approaches have been taken to try to solve the lack of inhibitor tolerance issue [23,[27][28][29]51]. Another positive attribute of this process is that the required inoculum was only 0.25 g dcw/L (0.5 OD), which represents a very low inoculum within the biofuels community [29,52]. Further, the ethanol productivity reported in this study includes the growth phase as well as the production phase without the addition of yeast extract or peptone. Moreover, this process was competitive with starch-based processes [53]. Further, several recombinant yeast studies that have reported using high density (20 OD) inoculum [25,26] did not include in the ethanol yields or productivities the material and time required to generate the inoculum. In many other ethanol-from-lignocellulosic studies, it is common to use high inoculum, supplementing the medium with yeast extract and peptone [29,32,51,52], corn steep liquor [35], or other nutrients [54]. It is not always clear if the inocula or these added carbon/nutrient sources are included in the total carbon balances when yields are reported. The final ethanol concentrations and the yields reached in this work using ECE and xylose isomerase represent extremely competitive values, surpassing any values reported for recombinant yeast cultures on hydrolysates, evolutionarily adapted, and combined saccharification/fermentation systems [7,23,26,28,30,31,51,[54][55][56].
It was surprising, in this work, that ethanol yield calculated based solely on glucose exceeds the theoretical yield of 51% [57]; however, it is very likely other carbon sources are being used for ethanol production derived from the hydrolysate or the ECE. Even though the final ethanol yields for the cultures with and without ECE and xylose isomerase at the lower hydrolysate concentrations were not different (Figure 3), it is possible that differences in ethanol productivities could be observed in a high gravity fermentations. For example, in this study, in the fermentation without ECE and xylose isomerase on the high-gravity hydrolysate (Figure 6), the yield on glucose to ethanol was significant lower. Had these culture gone to completion, it might have been possible to detect the ethanol final yield differences. Moreover, if the xylose accounted for in the hydrolysate is included in the yield calculation, the conversion of total sugars (190 g/L) was converted to 82 g/L ethanol, which is a yield of 43% or approximately 85% of theoretical yield. In comparison to pilot-scale commercial rice starch fermentations, which use very high gravity fermentation conditions, the high gravity sugarcane bagasse hydrolysate fermentations reported here had comparable ethanol productivities (between 1.6 and 1.8 g/L·h). Additionally, the yields reported for rice starch fermentations are around 83% [18]. It is not unexpected that reported final ethanol concentrations for rice starch fermentations are higher since these are very high gravity fermentations, and not limited by the available sugars in the solution. Thus, the focus of future hydrolysate ethanol fermentation work should emphasis increasing the sugar concentrations in the hydrolysate.
In this work, native S. pastorianus has demonstrated the capability to produce ethanol from sugarcane bagasse hydrolysate under high gravity fermentation conditions. The specifics of this processes allow for very low inoculums to be used, corresponding to 0.25 g dcw/L. This has several implications on the economics. First, the yeast biomass production was accounted for in the conversion of glucose to ethanol. Conversely, most reported ethanol fermentations use yeast inoculums on the order of 10 to 50 g dcw/L (or 1.75 to 3.5 × 10 7 cell/mL) [18,30,32,[34][35][36]58,59]. In these high inoculum cases, it is not apparent that the yield from sugar to ethanol accounts for the sugar used to generate the biomass, which for the 10 g/L case would be at least 20 g/L of glucose. Second, the capability to generate the yeast cells in-fermenter within the normal fermentation timeframe eliminates the need to purchase yeast, which a cost-savings, and allows for process to be located remotely and not rely on yeast biomass delivery. Third, this process did not include the common media additions of yeast extract or peptone [21,22,31,52]. This further reduces the reliance of the process on outside sources for materials and closes the carbon balance better than when these nutrient sources are added to the fermentation but not accounted for in the carbon balance. Ultimately, recycling the xylose back to the yeast as a nutrient-rich supplement allows for a full accounting of the sugars and other carbon sources that enter the system to be directly compared to the ethanol generated. Finally, the process described here represents an economically competitive process that could still be improved by increasing the sugar content in the hydrolysate and examining the ECE and xylose isomerase levels required in the hydrolysate fermentation for optimal performance.

Conclusions
Native S. pastorianus have the capability to produce ethanol from sugarcane bagasse hydrolysate under high gravity fermentation conditions when cultured with ECE and xylose isomerase. The ethanol fermentation is rapid and leads to economically distillable concentrations. The biomass generated can be used for cattle feed as no recombinant organisms are utilized. The self-sufficient generating aspects of this process make it amenable to remote locates prevalent to sugarcane plantations. It would be anticipated that the capability to generate liquid fuel in a remote location could displace the currently common practice of simply burning the waste sugarcane bagasse.