A Dissertation
entitled
Optimizing Simultaneous-Isomerization-and-Reactive-Extraction (SIRE) Followed by
Back-Extraction (BE) Process for Efficient Fermentation of Ketose Sugars to Products
by
Peng Zhang
Submitted to the Graduate Faculty as partial fulfillment of the requirements for the
Doctor of Philosophy Degree in
Biomedical Engineering
______Dr. Patricia Relue, Co-Committee Chair
______Dr. Sasidhar Varanasi, Co-Committee Chair
______Dr. Sridhar Viamajala, Committee Member
______Dr. Stephen Callaway, Committee Member
______Dr. Randall Ruch, Committee Member
______Dr. Amanda Bryant-Friedrich, Dean College of Graduate Studies
The University of Toledo
May 2018
Copyright 2018, Peng Zhang
This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of
Optimizing Simultaneous-Isomerization-and-Reactive-Extraction (SIRE) Followed by Back-Extraction (BE) Process for Efficient Fermentation of Ketose Sugars to Products
by
Peng Zhang
Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Biomedical Engineering
The University of Toledo
May 2018
Lignocellulosic biomass is an abundant, inexpensive feedstock. It is mostly composed of cellulose (38-50% of the dry mass) and hemicellulose (23-32% of the dry mass). Cellulose and hemicellulose are polysaccharides that can be hydrolyzed to monosaccharides, mostly glucose and xylose. These sugars can eventually be fermented to many different products, such as ethanol and 2,3-butanediol.
Fuel ethanol, which is currently produced from food-based sugars, can also be produced via fermentation of sugars derived from lignocellulosic biomass. The fermentation of xylose is essential for the cost-effective bioconversion of lignocellulose to fuels and chemicals, but wild-type strains of Saccharomyces cerevisiae do not metabolize xylose because the metabolic pathways convert xylose to xylitol via an NADPH-linked xylose reductase. Fermentation of xylose to ethanol through xylulose does occur in organisms which possess an NADH-linked aldose reductase, indicating that a balanced supply of NADH and NADPH must be maintained to avoid xylitol production. Although
S. cerevisiae does not convert xylose to ethanol, it does have the metabolic pathway for the conversion of xylulose, the ketose isomer of xylose, to ethanol. Conversion of xylose to iii xylulose in high yield and at low cost from biomass hydrolysate has the potential to bypass the barrier to ethanol production from C5 and C6 sugars with native yeast.
2,3-Butanediol (2,3-BD) is a key building block and a promising bulk chemical due to its extensive industrial applications in making polymers, plastics, and hydrocarbon fuels.
For example, 2,3-BD can be readily converted to butenes, butadiene, and methyl ethyl ketone that are used in the production of hydrocarbon fuels. Enterobacter cloacae NRRL
B-23289, isolated from decaying wood/corn soil samples by the USDA Agricultural
Research Service (Peoria, IL), is a natural producer of 2,3-BD. Previous work at the USDA
ARS has shown that this strain is more efficient in converting ketose than aldose sugars to
2,3-BD. Particularly interesting is that fermentation of fructose showed higher 2,3-BD yield within a much shorter period of time as compared to glucose.
Converting aldoses to ketoses involves isomerization, typically conducted enzymatically. However, the isomerization does not have a favorable equilibrium with respect to ketose formation. We have previously developed a method of simultaneous isomerization and reactive extraction (SIRE) to produce the ketose isomer of xylose
(xylulose) in high yield and purity. SIRE-followed by back-extraction (BE) allows recovery of xylulose in nearly pure form. Although successfully implemented with low concentration xylose, SIRE has not been tested for high concentration sugars (C5 and C6 mixtures), which would be relevant for biomass hydrolysates. Optimization of SIRE-BE with high concentrations of both C5 (xylose/xylulose) and C6 (glucose/fructose) is one of the objectives of this dissertation.
Using this innovative and optimized method to pretreat the aldose sugars (glucose and xylose) and produce large quantities of nearly pure, concentrated ketose sugars
iv
(fructose and xylulose), the production of 2,3-butanediol was investigated. Ketose sugar fermentation yielded more 2,3-butanediol and in a shorter time than the aldose fermentations. Using this method to produce large quantities of nearly pure, concentrated xylulose, the production of ethanol was also investigated. The native yeast produced 0.44-
0.45 g ethanol/ g xylulose in xylulose fermentation.
v
This dissertation is dedicated to my father (1954-1999), my grandmother (1929-2003) and my grandfather (1929-1995).
Acknowledgements
I would like to express the deepest appreciation to my advisors: Dr. Patricia Relue and Dr. Sasidhar Varanasi. Without their guidance and persistent help, my research and this dissertation would not have been possible.
I would like to thank my committee members: Dr. Sridhar Viamajala, Dr. Stephen
Callaway, and Dr. Randall Ruch, for giving me advice on my research.
In addition, a thank you to Tammy Phares, our lab coordinator, who helped me with my experimental set-ups. I also thank my lab mates: Bin Li, Heng Shao, Kelly Marbaugh, and Jeremy Schreur, for their assistance and support. I would also like to thank to the
College of Engineering, especially the Department of Bioengineering, for giving me the opportunity and supporting my Ph.D. study here.
Last but not least, a special thank you to my family, especially my wife and my mother, for their endless love, encouragement and support during this journey. A sincere thank you to my best friends, I feel fortunate to have them at my back.
v
Table of Contents
Abstract ...... iii
Acknowledgements ...... v
Table of Contents ...... vi
List of Tables ...... x
List of Figures ...... xi
1 Overview ...... 1
1.1 Background ...... 1
1.1.1 Simultaneous-isomerization-and-reactive-extraction (SIRE),
followed by back-extraction (BE) ...... 4
1.1.2 Ethanol Production...... 8
1.1.3 2,3-Butanediol Production ...... 10
1.2 Objectives and Significance ...... 12
1.2.1 Optimize the SIRE-BE process with high concentrations of both C5
and C6 sugars ...... 13
1.2.2 Evaluate the ethanol yield of native S. cerevisiae on fermentation of
xylulose and xylulose/glucose mixed sugars from the SIRE-BE
process ...... 13
1.2.3 Investigate 2,3-butandiol production from single and mixed sugars
by Enterobacter cloacae NRRL B-23289 ...... 13 vi
1.3 Organization of the Dissertation ...... 13
2 Kinetics of Isomerization and SIRE-BE Process with High Concentration of
Single Sugar… ...... 15
2.1 Introduction ...... 15
2.2 Method and Material ...... 16
2.2.1 Chemicals and materials ...... 16
2.2.2 Sugar isomerization ...... 16
2.2.3 SIRE-BE with high concentration of single sugar ...... 17
2.2.4 Analytical techniques and data analysis ...... 18
2.3 Results and Discussion ...... 18
2.3.1 Kinetic of glucose isomerization ...... 18
2.3.2 Kinetic of xylose isomerization ...... 21
2.3.3 Kinetic of mixed sugars isomerization ...... 21
2.3.4 SIRE with 165 mM and 490 mM N2B in the organic phase for C6
sugars ...... 26
2.3.5 SIRE with 490 mM N2B in the organic phase for C5 sugars ...... 28
2.4 Conclusions ...... 29
3 Production of High Purity Xylulose and Fructose Sugars from Biomass
Hydrolysates for Flexible Downstream Processing ...... 31
3.1 Introduction ...... 31
3.2 Method and Material ...... 35
3.2.1 Chemicals and materials ...... 35
3.2.2 Simultaneous isomerization and reactive extraction (SIRE) ...... 36
vii
3.2.3 Back-extraction (BE) of sugars from the organic phase ...... 36
3.2.4 Analytical techniques and data analysis ...... 37
3.2.5 Criteria for evaluating experiments ...... 38
3.3 Results and Discussion ...... 38
3.3.1 Mixed sugar SIRE-BE ...... 38
3.3.2 Two stage BE to enhance xylulose purity ...... 43
3.3.3 Non-equilibrium SIRE to enhance xylulose purity ...... 47
3.3.4 Multiple cycles of transient SIRE ...... 50
3.3.5 Purity of C6 sugars in the aqueous phase ...... 53
3.4 Conclusions ...... 55
4 New Route of C5 Sugars Fermentation to Ethanol by Native Saccharomyces
cerevisiae...... 57
4.1 Introduction ...... 57
4.2 Method and Material ...... 59
4.2.1 Chemicals and yeast strain ...... 59
4.2.2 Xylulose production from xylose using SIRE-BE ...... 60
4.2.3 Fermentation media ...... 60
4.2.4 Analytical techniques and data analysis ...... 60
4.3 Results and Discussion ...... 61
4.3.1 Glucose Fermentation ...... 61
4.3.2 Xylulose Production and Fermentation ...... 62
4.3.3 Glucose/Xylulose Fermentation ...... 68
viii
4.3.4 Schematic route for separation of xylulose from C6 sugars, followed
by fermentation ...... 72
4.4 Conclusions ...... 73
5 Production of 2,3-butanediol from Single and Mixed Sugars by Enterobacter
cloacae NRRL B-23289 ...... 75
5.1 Introduction ...... 75
5.2 Method and Material ...... 76
5.2.1 Chemicals and bacteria strain ...... 76
5.2.2 Xylulose production by SIRE-BE ...... 77
5.2.3 Fermentation media ...... 77
5.2.4 Analytical techniques and data analysis ...... 78
5.3 Results and Discussion ...... 78
5.3.1 Single sugar fermentation ...... 78
5.3.2 Mixed sugars fermentation ...... 80
5.4 Conclusions ...... 82
6 Future Directions ...... 84
6.1 Validation of scaled SIRE system...... 84
6.2 Purification of glycerol from industrial waste using SIRE-BE ...... 86
References ...... 87
A Media and Solution Formulas ...... 95
B Analytical Information on HPLC...... 96
ix
List of Tables
1.1 Market prices for research-grade sugars ...... 4
3.1 Proportion of C6 sugars and ratio of fructose to glucose in aqueous phase after each cycle ...... 54
4.1 Summary of C5 and C5/C6 fermentations ...... 71
4.2 Ethanol production by different GMO strains from C5 sugar ...... 74
B.1 Elution peak times on two Shodex SH1011 columns ...... 96
B.2 Elution peak times on two Shodex SP0810 columns...... 96
x
List of Figures
1-1 Crude oil prices from 2010-2016 ...... 2
1-2 Structure of cellulose and its digestion to glucose...... 2
1-3 Structure of xylan and its digestion to xylose...... 3
1-4 Flow chart for biomass processing to useful products ...... 4
1-5 Schemes of (a) reactive-extraction and (b) sugar recovery by stripping ...... 6
1-6 Schematic of SIRE-BE in bench-scale ...... 8
1-7 Xylose metabolic pathway in S. cerevisiae yeasts ...... 10
1-8 The pathway of 2,3-butanediol formation in E. cloacae ...... 12
2-1 Transient glucose isomerization with 45 g/L enzyme for different initial sugar loadings (A) 17 g/L, (B) 33 g/L, and (C) 65 g/L at 60 °C ...... 19
2-2 Transient glucose isomerization with 18 g/L enzyme for initial sugar loadings (A)
17 g/L, (B) 33 g/L, and (C) 65 g/L at 60 °C...... 20
2-3 Transient xylose isomerization with 45 g/L enzyme for different initial sugar loadings (A) 12 g/L, (B) 24 g/L, and (C) 48 g/L at 50 °C...... 22
2-4 Transient xylose isomerization with 18 g/L enzyme for initial sugar loadings (A)
12 g/L, (B) 24 g/L, and (C) 48 g/L at 50 °C ...... 23
2-5 Time course data of mixed sugars isomerization at 50 °C ...... 24
2-6 Time course data of mixed sugars isomerization at 60 °C ...... 25
2-7 Effect of N2B to sugar mole ratio on equilibrium SIRE with glucose ...... 27 xi
2-8 Effect of N2B to sugar mole ratio on sugar extraction efficiency and xylulose extraction selectivity for SIRE ...... 29
3-1 Hypothetical mechanism of (a) SIRE and (b) BE ...... 34
3-2 Results for mixed sugar equilibrium SIRE-BE ...... 40
3-3 Ketose yield and purity with changing N2B:sugar mole ratio ...... 42
3-4 Results of two-stage BE ...... 44
3-5 Results of two-stage BE with two N2B:C5 sugar mole ratios ...... 46
3-6 Results of transient SIRE ...... 49
3-7 Results for six cycles of transient SIRE for mixed sugar SIRE at 50°C ...... 52
3-8 Aqueous phase sugar composition after multiple cycles of transient SIRE ...... 54
4-1 Results for glucose fermentation to ethanol ...... 62
4-2 Schematic representation of the three-step process producing xylulose (Y) in high yield from xylose (X) and its fermentation by native S. cerevisiae to ethanol ...... 63
4-3 Results for fermentation of a 25% xylulose/75% xylose solution with 30 g/L total
C5 sugars ...... 65
4-4 Results for fermentation of xylulose produced from SIRE for 90% and 99% xylulose purity ...... 67
4-5 Results for supplemental glucose (5 g/L, 10 g/L, 20 g/L, and 30 g/L) added to a fermentation media containing 30 g/l C5 sugar ...... 69
4-6 Results for xylulose consumption with supplemental glucose added to a fermentation containing 30 g/l C5 sugar (99% xylulose) ...... 70
4-7 Schematic route for separation of C5 and C6 sugars from biomass hydrolysate for fermentation ...... 72
xii
5-1 Results for fermentation of 50 g/l sugar (xylose, xylulose, glucose, or fructose) to
2,3-BD by Enterobacter cloacae NRRL B-23289 at pH 5.0, 30 oC ...... 79
5-2 Results of sugar consumption and 2,3-BD production for fermentation of 50 g/l
(A) xylose and glucose, and (B) xylulose and fructose by E. cloacae NRRL B-23289 at pH 5.0, 30 oC ...... 81
6-1 The (A) schematic and (B) photo of the one-liter SIRE system ...... 85
xiii
Chapter 1
Overview
1.1 Background
Reducing the use of fossil fuels by utilizing renewable resources for the production
of liquid fuels continues to draw interest as the price of crude oil is fluctuates-see Figure
1-1 [1, 2]. Lignocellulosic biomass feedstock, which includes all plant and plant-derived
material, forms a potential renewable source to produce alternative fuels and a variety of
useful chemicals [3, 4]. Thus, one promising method to achieve the goal of reducing the
use of fossil fuels is to use lignocellulosic biomass. Lignocellulosic biomass is an abundant,
cheap feedstock that does not compete with the food supply. It is composed of carbohydrate
polymers (cellulose and hemicellulose) and lignin [5]. The cellulose and hemicellulose are
polysaccharides that can be hydrolyzed to monosaccharides [6] and eventually fermented
to various products, such as ethanol and 2,3-butanediol.
1
140
120
100
80
60
40
20 Brent crude oil price (dollars per barrel)per (dollarspriceoil crude Brent
0 2010 2011 2012 2013 2014 2015 2016
Figure 1-1 Crude oil prices from 2010-2016. Data is taken from [2].
Cellulose (38–50% of the dry biomass) is a linear polymer of cellobiose (glucose– glucose dimer) (Figure 1-2) [7]. The orientation of the linkages and additional hydrogen bonding make the polysaccharide rigid and difficult to deconstruct. The polymer linkages can be broken down by hydrolysis to release glucose monomers; glucose is a six-carbon
(C6) sugar [8].
Figure 1-2 Structure of cellulose and its digestion to glucose. Taken from [7] 2
Hemicellulose (23–32% of the dry biomass) consists of short, highly branched chains of C5 and C6 sugars, mainly xylose (C5), but also arabinose (C5), galactose (C6), glucose (C6), and mannose (C6). Xylan polymers are a major constituent of the hemicellulose (Figure 1-3) [7]. Hemicellulose, because of its branched, amorphous nature, is relatively easy to hydrolyze [9].
Figure 1-3 Structure of xylan and its digestion to xylose. Taken from [7]
To efficiently release sugars, lignocellulosic biomass usually undergoes pretreatment to open the structure. Pretreatment can occur by various techniques, including dilute acid pretreatment, ionic-liquid pretreatment, ammonia fiber explosion, and steam explosion [10, 11]. The pretreated biomass is then enzymatically digested by saccharification enzymes to release monomeric sugars. In the sugar-rich mixture
(hydrolysate) that results there will be both C5 and C6 sugars as well as sugar and lignin degradation products, such as furans and aliphatic acids. This dissertation focuses on
3
optimizing the process of sugar isomerization and separation and converting the sugars to useful products by fermentation (Figure 1-4).
Enzymatic Biomass Pretreatment Sugars Products Hydrolysis
Figure 1-4 Flow chart for biomass processing to useful products.
1.1.1 Simultaneous-isomerization-and-reactive-extraction (SIRE), followed by back- extraction (BE)
The ketose sugars (fructose, xylulose) are preferred over their aldose counterparts
(glucose, xylose) for several reasons. First, ketose sugars tend to be value-added and the market prices for xylulose and fructose are higher than their aldose counterparts (see Table
1.1). Second, in the ketose form, some transformations to products have either better outcomes or higher yields. For example, xylulose, but not xylose, can be fermented to ethanol by native yeast [12-14]. Third, the structure of the ketose sugars also facilitates their separation from the hydrolysate.
Table 1.1 Market prices for research-grade sugars (October, 2017). Prices are from Sigma-
Aldrich Co.
Sugar Form $/g
Glucose Aldose $42.20/1000 g
Fructose Ketose $121.00/1000 g
Xylose Aldose $202.00/1000 g
Xylulose Ketose $1025.00/250 mg
4
Converting aldoses to ketoses involves isomerization, typically conducted enzymatically. However, the isomerization does not have a favorable equilibrium with respect to ketose formation. At equilibrium, the ratio of xylulose to xylose is about 1:3 to
1:4, and the ratio of fructose to glucose is about 1:1. A traditional engineering approach to increasing product formation in a reversible reaction is through continuous product removal. A process that our research group has developed to increase the yield of ketose sugars is simultaneous-isomerization-and-reactive-extraction, followed by back-extraction
(SIRE-BE). The detailed mechanisms for SIRE and BE have been described previously
(Figure 1-5) [15].
5
a)
R
organic phase HO B Q O O (ABA) K
2
aqueous phase, high pH XI isomerization b)
organic phase
2
aqueous phase, low pH
Figure 1-5 Schemes of (a) reactive-extraction and (b) sugar recovery by stripping.
Taken from [15]. (a) During reactive-extraction, an aryl boronic acid (ABA) species in contact with a high pH aqueous phase is readily converted to its conjugate base (ABA−) form. In the presence of sugar, the ABA− forms a tetragonal ester with the sugar (ABAS−).
The negatively charged tetragonal ester remains confined at the interface until ion-pair formation occurs in the organic phase with the cation of Q+Cl−. (b) Stripping of sugars from the organic phase occurs by contacting the sugar-laden organic phase with a low pH aqueous phase containing an anion for the regeneration of the Q+Cl−. The ABAS− in contact with the low pH solution transitions to the uncharged triagonal ester (ABAS) form. The triagonal ester is readily hydrolyzed; the uncharged lipophilic ABA remains in the organic phase while the sugar moves to the aqueous phase. In SIRE, XI-catalyzed isomerization of aldose (A) to ketose (K) transformation is occurring simultaneously with sugar extraction as shown in (a). 6
To overcome the unfavorable isomerization equilibrium, we employ SIRE to extract the ketose sugars as they are formed (Figure 1-6). The selective extraction of ketose sugar from the aqueous phase solution is facilitated by the addition of naphthalene-2- boronic acid (N2B) and Aliquat® 336 to the organic phase. N2B preferentially binds to ketose sugars, and ion-pair formation between Aliquat® 336 and the sugar-N2B complex confines the complex to the organic phase. The differential, pH-dependent affinity of the
N2B for aldose and ketose not only influences their selective extraction but also the relative ease with which they can be dissociated from N2B and concentrated in acid media through
BE. In SIRE, the aldose to ketose isomerization is catalyzed by commercially-available immobilized glucose/xylose isomerase (GXI) enzyme at 50-60 °C. This temperature is compatible with saccharification, the last step for production of biomass hydrolysate.
Starting with a low concentration of xylose (10 mM) in SIRE, it was shown that
BE under moderately acidic conditions would release xylose, while the more strongly bound xylulose required more acidic conditions for dissociation. By implementing a two staged BE process with mid and low pH stripping solutions, the xylulose was recovered in stage 2 as a nearly-pure, concentrated aqueous stream [15].
7
Organic phase removed, 50 °C added to stripping media 25 °C
Organic Organic w/sugar
Isomerization, pH 8.5 Stripping, low pH
Figure 1-6 Schematic of SIRE-BE in bench-scale. During SIRE, ketose sugars are selectively removed from the aqueous phase by binding with complexing agents (CAs) such as N2B in the organic phase. After SIRE, sugars are back-extracted from organic phase into a low pH strip solution.
Although successfully implemented with low concentration xylose, SIRE has not been tested for high sugar concentrations (up to 490 mM) or for both C5 and C6 sugars, which would be present in biomass hydrolysate.
1.1.2 Ethanol Production
Fuel ethanol, which is currently produced from food-based sugars, can also be produced via fermentation of sugars derived from lignocellulosic biomass. However, utilization of both C6 (glucose) and C5 (xylose) sugars is necessary to make ethanol production from biomass economically feasible [12-14, 16, 17].
Ethanol-producing bacteria have attracted much attention because their growth rates are substantially higher than yeast. The main problems in the use of bacteria for fermentation are low ethanol tolerance and reduced product selectivity toward ethanol [18-
8
20]. Ethanol production from cellulosic materials with these strains has not reached a level sufficient for commercial application.
Significant progress has been made in the development and testing of genetically modified organisms (GMO) capable of metabolizing both C5 and C6 sugars. However, the genetic stability, overall ethanol yield, and ability to survive under the conditions of industrial fermentation are unproven for GMOs [13, 21]; as such, approaches that employ native strains are highly desirable.
Wild-type Saccharomyces cerevisiae is used in large-scale fermentation of sucrose and starch-based glucose for fuel ethanol due to its near-theoretical ethanol yields from glucose and tolerance to elevated osmotic pressure and a wide spectrum of inhibitors. The metabolic pathways in native S. cerevisiae convert xylose to xylitol [22-24] via an
NADPH-linked xylose reductase (XR) (Figure 1-7). Although xylitol can be oxidized to xylulose by an NADH-linked xylitol dehydrogenase (XDH), a redox imbalance between the cofactor usage of XR and XDH has been attributed to xylitol accumulation and disruption of metabolic flux to xylulose in long-term fermentation [18, 19, 25-29]. A comparative study of various xylose-assimilating yeasts showed that significant fermentation of xylose to ethanol through xylulose only occurred in those organisms which possessed an NADH-linked aldose reductase [27]. Pichia stipitis, one of the best naturally occurring xylose-fermenting yeasts, can convert xylose to ethanol in high yield. However, this yeast has low ethanol and sugar tolerance, preventing its use as an industrial strain for large-scale ethanol production from lignocellulosic biomass sugars [16].
9
Xylose NADPH Xylose reductase (XR) NADP+ Xylose isomerase Xylitol (XI) NAD+ Xylitol dehydrogenase (XDH) NADH Xylulose
Xylulose kinase (XK)
Xylulose-5P Pentose Phosphate Pathway Glycolysis
Ethanol
Figure 1-7 Xylose metabolic pathway in S. cerevisiae yeasts. Taken from [24].
Fermentation of xylulose to ethanol by S. cerevisiae in high yield has been shown previously [30, 31], but the unfavorable equilibrium (~20% xylulose) for isomerization of xylose to xylulose limits the practical utility of xylulose fermentation. High-yield, low cost production of xylulose from xylose has the potential to eliminate the bottleneck currently encountered for ethanol production from C5 and C6 sugars with native yeast.
1.1.3 2,3-Butanediol Production
The global butanediol market in 2015 was over 6 billion dollars and is estimated to exceed 12.6 billion dollars by 2025, growing at a compound annual growth rate (CAGR)
10
of 7.7% from 2015 to 2025. The Asia-Pacific region is the market leader, accounting for more than 63% of total volume consumed in 2015. The market for butanediol is expected to advance at a rapid pace during the next few years, and Asia-Pacific will remain the fastest growing region for butanediol due to its economic growth. Europe and North
America constitute the second and third largest markets for butanediol [32, 33].
2,3-Butanediol (2,3-BD) is a promising bulk chemical due to its extensive industrial applications as a key building block in making polymers, plastics, and hydrocarbon fuels.
For example, 2,3-BD can be readily converted to butenes, butadiene, and methyl ethyl ketone that are used in the production of hydrocarbon fuels [34]. 2,3-BD is a colorless and nearly odorless viscous liquid at room temperature. With a heating value of 27.2 kJ/g, it compares favorably with ethanol and methanol for use as a liquid fuel and fuel additive. It is also an important intermediate chemical for production of tetrahydrofuran, polytetramethylene ether glycol (PTMEG), polybutylene terephthalate (PBT), gamma- butyrolactone (GBL), polyurethane (PU), and other solvents. These chemicals are widely used in fibers, engineering plastics, medicines, cosmetics, artificial leathers, pesticides, plasticizers, hardeners, solvents and rust removers [35].
There are many petroleum-based chemical methods to produce 2,3-BD today, but a bio-based process would have a favorable greenhouse gas balance and less pollution when compared to petroleum-based processes [36]. The unstable global oil price is another reason to focus on a bio-based process. 2,3-BD can be produced directly from sugars by fermentation using bacteria.
Enterobacter cloacae NRRL B-23289, isolated from decaying wood/corn soil samples by the USDA Agricultural Research Service (Peoria, IL), is a natural producer of 11
2,3-BD. Previous work at the USDA ARS has shown that this strain is more efficient in converting ketose than aldose sugars to 2,3-BD [37]. Particularly interesting is that fermentation of fructose showed higher 2,3-BD yield within a much shorter period of time as compared to glucose [37].
The pathways for sugars to 2,3-BD are shown in Figure 1-8. Interestingly, xylulose is an intermediate in the xylose-to-2,3-BD pathway. To our knowledge, no one has tested the effectiveness of this organism in converting the ketose sugar xylulose to 2,3-BD; this is likely due to the difficulty of producing high purity xylulose cost-effectively.
Glucose
D-xylose D-xylulose Glucose-6-P
2,3-BD Acetoin Pyruvate Fructose-6-P Fructose
Figure 1-8 The pathway of 2,3-butanediol formation in E. cloacae.
1.2 Objectives and Significance
The objectives of this dissertation are to produce xylulose from biomass in high purity and high yield and evaluate the conversion of the xylulose to useful products.
These objectives include:
12
1.2.1 Optimize the SIRE-BE process with high concentrations of both C5 and C6 sugars
The SIRE-BE process was developed previously to produce the ketose isomer of xylose (xylulose) in high yield and purity. However, this process was only evaluated with a low concentration (10 mM) of xylose. Sugar found in biomass hydrolysate are likely to be around 500 mM.
1.2.2 Evaluate the ethanol yield from fermentation of xylulose and xylulose/glucose mixed sugars from the SIRE-BE process using native S. cerevisiae
S. cerevisiae yeast does not have the metabolic pathway for the conversion of xylose to ethanol. Converting xylose to xylulose using SIRE-BE has the potential to bypass this barrier to ethanol production.
1.2.3 Investigate 2,3-butandiol production from single and mixed sugars by
Enterobacter cloacae NRRL B-23289
Previous work at the USDA ARS has shown that this particular strain is more efficient in converting ketose than aldose sugars to 2,3-BD [38]. No one has evaluated 2,3-
BD production from high purity xylulose with this strain because of xylulose’s limited availability and high cost.
1.3 Organization of the Dissertation
Chapter 2 discusses the conditions and parameters used in high sugar concentration
SIRE-BE. Two important molar ratios were changed during the SIRE: (1) the ratio of N2B to sugars and (2) the ratio of Aliquat® 336 to N2B. The kinetics of single and mixed sugar isomerization at different temperatures were also measured and are included in this chapter.
13
Based on the SIRE-BE results, strategies for fractionating mixed sugar streams, with the specific goal of preferentially extracting xylulose from multi-sugar mixtures produced in SIRE, are presented in Chapter 3.
In Chapter 4, using high purity xylulose and glucose/xylulose mixtures obtained from SIRE-BE, the feasibility of ethanol production from both C5 and C6 sugars with native yeast was assessed.
Chapter 5 provides the experimental results for the organism Enterobacter cloacae
NRRL B-23289 which was selected to ferment the ketose sugar xylulose to 2,3-BD.
Chapter 6 provides suggestions for potential future applications of the SIRE-BE process.
14
Chapter 2
Kinetics of Isomerization and SIRE-BE Process with High Concentration of Single Sugar
2.1 Introduction
Converting aldoses to ketoses involves isomerization, typically conducted
enzymatically. Glucose/xylose isomerase (GXI, such as Gensweet® IGI) is commercially
used in the isomerization reaction of glucose to fructose and xylose to xylulose,
respectively. The industrially used enzymes are produced by Streptomyces and Bacillus
species [39]. As the market for GXI is large in the food industry [40], GXI can be acquired
at low cost. Using this enzyme, kinetics of C6 or C5 sugar at different temperatures were
assessed. Results are presented in this chapter.
We previously developed the SIRE-BE method to produce the ketose isomer of
xylose (xylulose) in high yield and purity [15]. However, this process was only tested with
a low concentration (10 mM) of xylose. It was shown that 2-stage BE can separate low
concentrations of C5 isomers. This process has not been tested for high concentrations of
sugars likely to be found in biomass hydrolysates. SIRE-BE results discussed in this
chapter were generated using high concentrations of glucose or xylose (up to 490 mM).
15
2.2 Method and Material
2.2.1 Chemicals and materials
1-octanol and naphthalene-2-boronic acid (N2B) were purchased from Thermo
Fisher Scientific Inc. (Pittsburgh, PA, USA). Aliquat® 336 (an industrial extractant consisting of primarily tricaprylylmethylammonium chloride), D-glucose, D-xylose and
D-fructose were purchased from Sigma Aldrich Co. (St. Louis, MO, USA). D-xylulose used for HPLC standards was purchased from ZuChem (Chicago, IL, USA). All additional chemicals and solvents were purchased from Thermo Fisher Scientific Inc. (Pittsburgh, PA,
USA).
Immobilized xylose isomerase (Gensweet® IGI, XI) was a kind gift from Dupont
Industrial Biosciences (Wilmington, DE). It is produced from a genetically-modified strain of Streptomyces rubiginosus and catalyzes the isomerization of glucose to fructose and xylose to xylulose. The immobilized XI pellets were cylinder-shaped granules with a diameter of approximately 1-3 mm, and were stored at 4°C until use.
2.2.2 Sugar isomerization
Isomerization reactions with either pure glucose (C6) or xylose (C5) with initial concentrations of approximately 82.5 mM, 165 mM, and 330 mM in 50 mM phosphate buffer were conducted. Each sugar solution (50 mL) was placed into a capped shake flask and preheated to 60 °C (glucose) or 50 °C (xylose) in a water shaker bath. Two enzyme loadings were used: 45 g/L and 18 g/L.
Based on the information from the enzyme supplier, the sugar may bind non- specifically on the enzyme support, resulting in sugar loss. Sugar loss is expected to occur when dry enzyme is added to sugar solution, and no additional sugar loss should be 16
observed if the enzyme is immersed in the solution to avoid contacting with the air. To minimize this loss of sugar, all the isomerization experiments were conducted by first presoaking the enzyme in reaction media (cycle 1) and then replacing 35 mL of the reaction media with fresh sugar solution (cycle 2).
Transient data was collected from both cycle 1 and cycle 2. Each sample (0.2 mL) was stored on ice until analysis for aldose and ketose sugar concentrations.
2.2.3 SIRE-BE with high concentration of single sugar
SIRE was conducted at 50 °C (xylose) or 60 °C (glucose) by contacting the aqueous phase sugar solution (pH 8.5; 166.7 or 490 mM) in 50 mM sodium phosphate buffer containing Gensweet® IGI (immobilized glucose/xylose isomerase), with the organic phase of octanol containing N2B and Aliquat® 336. SIRE was conducted until the pH stopped changing, at which true equilibrium was assumed. The two phases were then separated by centrifuge. The sugars extracted into the organic phase were recovered by back-extraction into 500 mM HCl solution for 40 minutes with constant vortexing. The concentrations of
N2B used were approximately equal to the sugar concentration, 165 and 490 mM. The concentrations of Aliquat® 336 were varied to determine the best N2B:Aliquat® 336 ratio for extraction.
Results from the process are reported based on the selectivity of individual sugars recovered and the overall yield or efficiency of sugar extraction. Specifically,
Sugar Extraction Efficiency (SEE) = Sugar in the organic phase (mol) Initial sugar in the aqueous phase (mol)
Ketose Extraction Selectivity (KES) = Ketose in the organic phase (mol) Total sugar in the organic phase (mol)
17
Back Extraction Selectivity (BES) = Ketose in the stripping phase (mol) Total sugar in the stripping phase (mol)
2.2.4 Analytical techniques and data analysis
All sugar samples were diluted in ultrapure water and filtered through a 0.22 µm pore-size sterile filter. Calibration standards for sugars were prepared in a similar manner.
Samples and standards were analyzed using two Shodex SH1011 columns (300×8 mm, from Showa Denko K.K, Japan) in series on an Agilent 1100 HPLC system equipped with a refractive index detector (RID). A mobile phase of 0.05 M H2SO4 was run at 0.6 ml/min; a column temperature of 50 °C and detector temperature of 35 °C were used for optimal peak resolution and detection. For mixed sugars samples (glucose, fructose, xylose and xylulose), samples and standards were analyzed using two Shodex SP0810 columns
(300×8 mm, from Showa Denko K.K, Japan) in series on an Agilent 1100 HPLC system equipped with a refractive index detector (RID). A mobile phase of ultra-pure water was run at 0.6 ml/min; a column temperature of 80°C and detector temperature of 35 °C were used for optimal peak resolution and detection.
All the experiments were conducted in duplicate. In cases where error bars are too small to be visible, they are not included in the data plotted.
2.3 Results and Discussion
2.3.1 Kinetic of glucose isomerization
The results for the isomerization of glucose to fructose at 60 °C and pH 8.5 are shown in Figure 2-1 (45 g/L enzyme) and Figure 2-2 (18 g/L enzyme). As shown in the results, increasing the enzyme loading reduced the time to reach the isomerization equilibrium, over 400 minutes with 18 g enzyme/L and about 90 minutes with 45 g
18
18 Glucose Cycle 2 16 Fructose Cycle 2 Glucose Cycle 1 14 Fructose Cycle 1
12
10
8
6 Concentration (g/L) Concentration
4
2
0 0 20 40 60 80 100 120 140 160 Time (mins) (A)
35
30
25
20
15
Concentration (g/L) 10
5
0 0 20 40 60 80 100 120 140 160 Time (mins) (B)
70
60
50
40
30
Concentration (g/L) 20
10
0 0 20 40 60 80 100 120 140 160
Time (mins) (C)
Figure 2-1 Transient glucose isomerization with 45 g/L enzyme for different initial sugar loadings (A) 17 g/L, (B) 33 g/L, and (C) 65 g/L at 60 °C. 19
18 Glucose Cycle 2 16 Fructose Cycle 2 Glucose Cycle 1 14 Fructose Cycle 1
12
10
8
6 Concentration Concentration (g/L)
4
2
0 0 50 100 150 200 250 300 350 400 450 Time (mins) (A)
35
30
25
20
15
Concentration (g/L) Concentration 10
5
0 0 50 100 150 200 250 300 350 400 450 Time (mins) (B)
70
60
50
40
30
Concentration(g/L) 20
10
0 0 50 100 150 200 250 300 350 400 450 Time (mins) (C)
Figure 2-2 Transient glucose isomerization with 18 g/L enzyme for initial sugar loadings (A) 17 g/L, (B) 33 g/L, and (C) 65 g/L at 60 °C.
20
enzyme/L. The sugar loss was minimized by presoaking the enzyme as expected by comparing the mass balance of cycle 1 and cycle 2 (data not shown).
2.3.2 Kinetic of xylose isomerization
The results for the isomerization of xylose to xylulose at 50 °C and pH 8.5 are shown in Figure 2-3 (45 g/L enzyme) and Figure 2-4 (18 g/L enzyme). As shown in the results, increasing the enzyme loading reduced the time to reach the isomerization equilibrium: it took about 20 minutes to reach equilibrium with 45 g/L enzyme, while it took 50 minutes to reach equilibrium with 18 g/L enzyme. The sugar loss was minimized by presoaking the enzyme as expected.
2.3.3 Kinetic of mixed sugars isomerization
The kinetics of mixed sugars isomerization was also investigated at both 50 °C and
60 °C. Two initial sugar concentrations were used: (1) 30 g/L glucose and 15 g/L xylose; and (2) 30 g/L glucose and 10 g/L xylose. The enzyme loading used was 45 g/L. The results for the presoaked isomerization of mixed sugars at pH 8.5 are shown in Figure 2-5 (50 °C) and Figure 2-6 (60 °C). As shown in the results, increasing the temperature reduced the time to reach the equilibrium of isomerization for both xylose and glucose. The rate of xylose isomerization is much higher than that of glucose isomerization, especially at the lower temperature, 50 °C. This difference of isomerization rate is explored further in
Chapter 3 to separate C5 and C6 sugars.
21
15 Xylose Cycle 2 Xylulose Cycle 2 Xylose Cycle 1 12 Xylulose Cycle 1
9
6 Concentration Concentration (g/L)
3
0 0 20 40 60 80 100 120 140 160 (A) Time (mins)
25
20
15
10 Concentration (g/L) Concentration
5
0 0 20 40 60 80 100 120 140 160 Time (mins) (B)
50
40
30
20 Concentration (g/L)
10
0 0 20 40 60 80 100 120 140 160 (C) Time (mins)
Figure 2-3 Transient xylose isomerization with 45 g/L enzyme for different initial sugar loadings (A) 12 g/L, (B) 24 g/L, and (C) 48 g/L at 50 °C.
22
15 Xylose Cycle 2 Xylulose Cycle 2 Xylose Cycle 1 12 Xylulose Cycle 1
9
6 Concentration Concentration (g/L)
3
0 0 50 100 150 200 250 300 350 400 450
Time (mins) (A)
25
20
15
10 Concentration (g/L) Concentration
5
0 0 50 100 150 200 250 300 350 400 450
Time (mins) (B)
50
40
30
20 Concentration(g/L)
10
0 0 50 100 150 200 250 300 350 400 450
Time (mins) (C)
Figure 2-4 Transient xylose isomerization with 18 g/L enzyme for initial sugar loadings (A) 12 g/L, (B) 24 g/L, and (C) 48 g/L at 50 °C.
23
30 Glucose Fructose 25 Xylose Xylulose
20
15
Concentration (g/L) Concentration 10
5
0 0 50 100 150 200 250 300 350 400 450 (A) Time (min)
30
25
20
15
10 Concentration (g/L)
5
0 0 50 100 150 200 250 300 350 400 450 (B) Time (mins)
Figure 2-5 Time course data of mixed sugars isomerization at 50 °C. Initial sugar concentrations were 30 g/L glucose and (A) 15 g/L xylose or (B) 10 g/L xylose.
24
30 Glucose Fructose 25 Xylose Xylulose
20
15
Concentration (g/L) 10
5
0 0 50 100 150 200 250 300 350 400 450 (A) Time (min)
30
25
20
15
10 Concentration (g/L)
5
0 0 50 100 150 200 250 300 350 400 450 (B) Time (mins)
Figure 2-6 Time course data of mixed sugars isomerization at 60 °C. Initial sugar concentrations were 30 g/L glucose and (A) 15 g/L xylose or (B) 10 g/L xylose.
25
2.3.4 SIRE with 165 mM and 490 mM N2B in the organic phase for C6 sugars
Theoretically, increasing the concentration of N2B in the organic phase should allow increased sugar extraction efficiency and higher sugar concentration in the back- extraction phase. To determine if increasing the concentration of N2B and Aliquat® 336 in the organic phase is feasible for effective SIRE, the first step was to determine N2B solubility in the organic phase. The highest solubility of N2B in the organic phase that was practical for SIRE was found to be 490 mM. The next step was to determine if increasing the N2B and Aliquat® 336 concentrations in the organic phase would change the organic phase character and sugar extraction characteristics. To test this, 165 mM and 490 mM
N2B was used in the organic phase. The results of SIRE using 165 mM N2B are summarized in Figure 2-7 (A). Two important molar ratios govern the sugar extraction during the SIRE: (1) the ratio of N2B to sugars and (2) the ratio of Aliquat® 336 to N2B.
Both have a very significant impact on sugar extraction efficiency (solid symbols); their impact on fructose extraction selectivity (open symbols) is less pronounced. As shown in
Figure 2-7, increasing the mole ratio of N2B to sugars will increase the sugar extraction efficiency and slightly reduce the fructose extraction selectivity. For Aliquat® 336 to N2B mole ratios greater than 1.5, the sugar extraction efficiency and the fructose extraction selectivity are comparable.
26
100
90
80 A:N2B 0.5 70 A:N2B 1.5 A:N2B 2 60 A:N2B 2.5
50
40
30
Sugar extraction efficiency or or efficiency extraction Sugar 20 Fructose extraction selectivity, % selectivity, extraction Fructose 10
0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 N2B:Sugar mole ratio (A)
100
90
80 A:N2B 1
70 A:N2B 1.5 A:N2B 2 60
50
40
30
20 Sugar extraction efficiency or efficiency extraction Sugar Fructose extraction selectivity, % selectivity, extraction Fructose 10
0 0 0.5 1 1.5 2 2.5 3 N2B:Sugar mole ratio (B)
Figure 2-7 Effect of N2B to sugar mole ratio on equilibrium SIRE with glucose.
Results are shown for sugar extraction efficiency (solid symbols) and fructose extraction selectivity (open symbols) for SIRE. (A) The concentration of N2B was held constant at
165 mM and the concentration of glucose was held constant at 30 g/L (166.7 mM). The volume of the two phases was adjusted to produce different mole ratios. Volume ratios of organic to aqueous phase ranged from 0.4 to 1.5. (B) The concentration of N2B was held constant at 490 mM and the concentration of glucose was held constant at 88.2 g/L (490 mM). Volume ratios of organic to aqueous phase ranged from 0.6 to 2.5. SIRE was conducted at 60°C.
27
The results for SIRE with 490 mM N2B are shown in Figure 2-7 (B). The results are very similar to those obtained with 165 mM N2B, but were collected over a larger range of N2B:sugar ratios. Similar impact of Aliquat® 336:N2B ratio is seen for 165 mM and
490 mM N2B. For both 165 mM and 490 mM N2B with Aliquat® 336:N2B=2 and
N2B:sugar=1.5, the sugar extraction efficiencies are comparable (~75%). Due to the difference in N2B concentration in the organic phase, however, the concentration of sugars in the organic phase is 377 mM for 490 mM N2B compared to only 125 mM with 165 mM
N2B. We have also been able to demonstrate that the sugar extraction efficiency is
N2B:sugar ratio-dependent and not concentration-dependent. This is significant because the sugar concentration can be increased by a factor of 5 and the volume of sugar decreased by a factor of 5 with the same overall sugar extraction efficiency (data not shown). Thus, even a dilute initial sugar solution can be extracted effectively and concentrated significantly in the SIRE process.
2.3.5 SIRE with 490 mM N2B in the organic phase for C5 sugars
After assessing C6 sugar extraction for high N2B concentrations, xylulose SIRE was also independently assessed with 490 mM N2B. The results for SIRE with C5 sugars and 490 mM N2B are shown in Figure 2-8. The results follow the same trends as observed with C6 sugars. No significant difference was observed for Aliquat® 336 to N2B ratios of
1.5 or 2. Increasing the mole ratio of N2B to sugars increases the sugar extraction efficiency and reduces the xylulose extraction selectivity due to the presence of excess N2B. If C5 and C6 sugars are compared at the same condition, similar sugar extractions, but higher xylulose than fructose selectivity are observed.
28
100
90
80 A:N2B 1.5 70 A:N2B 2 60
50
40
30
20 Sugar extraction efficiency or efficiency extraction Sugar Xylulose extraction selectivity, % selectivity, extraction Xylulose 10
0 0 0.5 1 1.5 2 2.5
N2B:Sugar mole ratio
Figure 2-8 Effect of N2B to sugar mole ratio on sugar extraction efficiency (solid symbols) and xylulose extraction selectivity (open symbols) for SIRE. The concentration of N2B was held constant at 490 mM and the concentration of xylose was held constant at
73.5 g/L (490 mM). The volume of the two phases was adjusted to produce different mole ratios. SIRE was conducted at 50°C. Results reported are for equilibrium extraction.
2.4 Conclusions
In this chapter, isomerization kinetics were measured for single and mixed sugar solutions at different initial sugar concentrations, enzyme loadings and temperature.
Increasing the enzyme loading or temperature increase the rate of isomerization. The rate of xylose isomerization is much faster than that of glucose isomerization under the same conditions. Observed sugar loss due to the non-specific binding of sugar to the enzyme
29
pellets. This observed sugar loss can be reduced by pre-soaking the GXI in aqueous sugar media.
The highest solubility of N2B in the organic phase that was practical for SIRE was found to be 490 mM. Two important molar ratios govern the sugar-extraction during the
SIRE: (1) the ratio of N2B to sugars and (2) the ratio of Aliquat® 336 to N2B. The effects of these two molar ratios on sugar extraction efficiency and ketose extraction selectivity
(fructose or xylulose) for SIRE were evaluated for C6 (165 mM and 490 mM) and C5 sugars (490 mM). No significant differences were observed between the two concentrations for the C6 sugars. Increasing the mole ratio of N2B to sugars increases the sugar extraction efficiency but reduces the ketose extraction selectivity. For Aliquat® 336 to N2B mole ratios of 1.5 or greater, the sugar extraction efficiency and the ketose extraction selectivity are comparable.
Depending on the downstream process, the operating conditions for SIRE can be adjusted to maximize either sugar extraction efficiency or ketose extraction selectivity. For example, with fermentation of mixed C5 and C6 sugars to ethanol, xylulose (not xylose) is fermented to ethanol, but both fructose and glucose can be readily converted. Hence, conditions that favor high xylulose conversion and extraction are needed, but selectivity between glucose and fructose extraction is not necessary. In the following chapter, strategies for isolating a single isomer (xylulose) from a mixture of C5 and C6 sugars are explored, an approach that would benefit ethanol fermentations.
30
Chapter 3 Production of High Purity Xylulose and Fructose Sugars from Biomass Hydrolysates for Flexible Downstream Processing
3.1 Introduction
Reducing the use of fossil fuels by utilizing renewable resources for the production
of liquid fuels and renewable products continues to draw interest [1, 41, and 42]. One
promising method to achieve this goal is to use lignocellulosic biomass. Lignocellulosic
biomass is an abundant, cheap feedstock that does not compete with the food supply. It is
composed of carbohydrate polymers (cellulose and hemicellulose) and lignin [5]. The
cellulose and hemicellulose are polysaccharides that can be hydrolyzed to
monosaccharides. These sugars serve as platform molecules for fermentation or chemical
conversion to higher-value products, such as ethanol, and 2,3-butanediol and furans [6, 9,
43-45].
Cellulose (38–50% of the dry biomass) is a linear polymer of cellobiose, a glucose–
glucose dimer. The orientation of the linkages and additional hydrogen bonding make the
polysaccharide rigid and difficult to deconstruct, but pretreatment methods and
saccharification can be used to release glucose monomers [10]. Hemicellulose (23–32% of
the dry biomass) consists of short, highly branched chains of C5 and C6 sugars, mainly
31
xylose (C5), but also arabinose (C5), galactose (C6), glucose (C6), and mannose (C6).
Hemicellulose is relatively easy to hydrolyze because of its branched, amorphous nature
[46].
Utilization of not only glucose but also xylose sugars significantly improves the economic feasibility of downstream product production [12-14, 16, 17, 47, 48]. Glucose and xylose are naturally occurring in biomass; however, the ketose isomers fructose and xylulose have several advantages over their aldose counterparts (glucose, xylose) for production of products. In the ketose form, some transformations to products have better outcomes and/or higher yields. For example, xylulose, but not xylose, can be fermented to ethanol by native yeast [12, 13, 30, 49].
Converting aldoses to ketoses involves isomerization, typically conducted enzymatically. However, the isomerization does not have a favorable equilibrium with respect to ketose formation. At equilibrium, the ratio of xylulose to xylose is 1:4, and the ratio of fructose to glucose is about 1:1. To increase the yield and purity of ketose sugars from hydrolysates, our research group has employed a product removal strategy of simultaneous-isomerization-and-reactive-extraction, followed by back-extraction (SIRE-
BE). The detailed mechanisms for SIRE and BE have been described previously [15]. To overcome the unfavorable isomerization equilibrium, we employ SIRE to extract the ketose sugars as they are formed. Interestingly, the structure of the ketose sugars facilitates their separation from the aldose sugars and other compounds present in biomass hydrolysate.
The selective extraction of ketose sugar from the aqueous phase solution is achieved by the addition of naphthalene-2-boronic acid (N2B) and Aliquat® 336 to the contacting organic phase. N2B preferentially binds to ketose sugars, and ion-pair formation between Aliquat® 32
336 and the sugar-N2B complex stabilizes the complex in the organic phase. The differential, pH-dependent affinity of N2B for ketose over aldose not only influences their selective extraction but also the relative ease with which they can be dissociated from N2B and concentrated in acid media through BE. A summary of the reaction chemistries involved in the SIRE and BE processes are shown in Figure 3-1.
33
Organic Phase
Q+Cl-
N2B N2B- Q+N2B- N2B-S N2B--S Q+N2B--S
S S Cl- S H+
- OH H2O
(a) Aqueous Phase, high pH
Organic Phase
Q+Cl-
Q+N2B--S N2B--S N2B
+ - - H + Cl H2O H2OOH S
(b) Aqueous Phase, low pH
Figure 3-1 Hypothetical mechanism of (a) SIRE and (b) BE. Chemistries involved with movement of sugar, base, acid and chloride ion during (a) SIRE and (b) BE. S represents any sugar, Q+Cl- represents Aliquat® 336. (a) During SIRE, an organic phase containing N2B and Aliquat® 336 is in contact with aqueous sugar solution at pH 8.5.
Addition of 15 M NaOH is used to maintain the pH. (b) Back extraction of sugars from the organic phase occurs by contacting the sugar-laden organic phase with a low pH HCl aqueous phase. As sugar is released from the N2B-sugar ester, the aqueous phase pH rises.
34
Glucose, which is the predominant sugar in biomass, is of high interest as a precursor for HMF production. HMF is produced more readily and in higher yield from fructose than glucose, but a pathway for high yield production of fructose from biomass hydrolysate has not been demonstrated. Although SIRE has been successfully implemented with isomers of single sugars (glucose/fructose or xylose/xylulose [15]), SIRE has not been previously evaluated for high sugar concentrations (up to 220 g/l) or for mixtures of C5 and C6 sugars which would be present in biomass hydrolysate. Of the four sugars, xylulose has the highest affinity for N2B, followed by fructose, glucose, and xylose [15]. While xylose and xylulose are easily separated, xylulose and fructose will be more difficult to separate since their N2B binding affinities are less different.
In this paper we present experimental results for implementation of SIRE-BE on high concentration, mixed C5 and C6 sugar streams representative of biomass hydrolysates.
We have evaluated SIRE for separation of ketoses from aldoses to produce a high purity ketose stream. We have also evaluated strategies for fractionating mixed sugar streams with the specific goal of preferentially purifying xylulose (and consequently fructose) from this multi-sugar mixture. We have modified our SIRE approach to increase xylulose yield, and altered BE to increase xylulose purity. We have also implemented non-equilibrium
SIRE for this same purpose.
3.2 Method and Material
3.2.1 Chemicals and materials
Naphthalene-2-boronic acid (N2B) and 1-octanol were purchased from Thermo
Fisher Scientific Inc. (Pittsburgh, PA, USA). Aliquat® 336 (an industrial extractant consisting of primarily tricaprylylmethylammonium chloride), D-glucose, D-xylose and 35
D-fructose were purchased from Sigma Aldrich Co. (St. Louis, MO, USA). D-xylulose used for HPLC standards was purchased from ZuChem (Chicago, IL, USA). All additional chemicals and solvents were purchased from Thermo Fisher Scientific Inc. (Pittsburgh, PA,
USA). Immobilized xylose isomerase (Gensweet® IGI, XI) was a kind gift from Dupont
Industrial Biosciences (Wilmington, DE). The immobilized XI pellets were cylinder- shaped granules with a diameter of approximately 1-3 mm, XI pellets were stored at 4°C until use.
3.2.2 Simultaneous isomerization and reactive extraction (SIRE)
Briefly, the aqueous phase sugar solution containing 45 g/L XI was contacted with the organic extraction phase for ~6 hours. Two simulated hydrolysates in 50 mM sodium phosphate buffer were used: (1) 30 g/L (200 mM) xylose and 60 g/L (333 mM) glucose; and (2) 73.5 g/L xylose (490 mM) and 147 g/L (817 mM) glucose. The organic extraction phase consisted of 490 mM N2B and 980 mM Aliquat® 336 in octanol. The volume ratio of the organic phase to the sugar solution was set based on the desired molar ratio of N2B to sugar. SIRE was conducted on a water bath shaker at 200 rpm and 50°C. The aqueous sugar solution was maintained at pH 8.5 during SIRE by addition of 15 M NaOH as needed.
To collect samples, the organic phase was separated from the aqueous phase by centrifugation at 5000 rpm for 5 min at room temperature. Aqueous phase samples were analyzed by HPLC to determine the composition of the sugars remaining and the total sugar extracted into the organic phase.
In this paper, SIRE was conducted in two ways: (1) equilibrium SIRE, in which the aqueous and organic phases were left in contact until isomerization and reactive extraction
36
both reached equilibrium and no further pH changes occurred (~6 hr); and (2) transient
SIRE, where glucose isomerization had not yet reached equilibrium (up to 125 min).
3.2.3 Back-extraction (BE) of sugars from the organic phase
Sugars were stripped, or back-extracted, from the organic phase by contacting the organic phase with an HCl solution. Back-extraction was conducted in (a) one stage to extract all of the sugars from the organic phase or (b) two stages to further enhance isomer separation and purification. The concentration of HCl in the stripping solution ranged from
25 mM to 750 mM, depending on the experimental objective. To facilitate efficient stripping, the two-phase mixture was subjected to continuous vortexing for 40 min at room temperature. After BE of sugars, phase separation was achieved by centrifugation at 5000 rpm for 5 min at room temperature. The composition of sugar in the stripping solution was analyzed by high performance liquid chromatography (HPLC).
Sugar-rich organic phase samples produced by SIRE were processed with 500-750 mM HCl solution to extract all sugars in one stage of BE. For two stage BE, the acid concentration in the stripping solution ranged from 25-400 mM for the first stage of BE;
500 or 750 mM HCl was used in the second stage of BE to strip all of the remaining sugars.
The purity of xylulose obtained in the second stage of BE was used to determine the best
HCl concentration to use in the first stage of BE.
3.2.4 Analytical techniques and data analysis
Samples collected from the experiments were diluted as needed in ultrapure water and filtered through a 0.22 µm pore-size sterile filter. Calibration standards for sugars were prepared in a similar manner. All samples and standards were analyzed using two Shodex
SP0810 columns (300×8 mm, from Showa Denko K.K, Japan) in series on an Agilent 1100 37
HPLC system equipped with a refractive index detector (RID). A mobile phase of ultra- pure water was run at 0.6 ml/min; a column temperature of 80 °C and detector temperature of 35 °C were used for optimal peak resolution and detection.
The composition of sugars extracted into the organic phase was assumed to be equivalent to the composition of sugars recovered by BE with high HCl concentration (500 to 750 mM). All data presented are the average values of duplicate experiments.
3.2.5 Criteria for evaluating experiments
The following metrics were used to judge the outcomes of the SIRE-BE experiments: (1) the xylulose purity, defined as the mole fraction of xylulose to total sugars
(xylose, xylulose, glucose and fructose) in the stripping phase; and (2) the xylulose yield, defined as the fraction of xylose from the simulated hydrolysate recovered as xylulose in the BE phase.
3.3 Results and Discussion
3.3.1 Mixed sugar SIRE-BE
Biomass hydrolysates contain both C6 and C5 sugars (glucose and xylose). To evaluate the efficacy of the SIRE-BE process on high concentration sugar mixtures, we used a simulated mixed sugar hydrolysate containing 60 g/l (333 mM) glucose and 30 g/l
(200 mM) xylose.
As a starting point, an N2B:sugar molar ratio of 1:1 (corresponding to an N2B:C5 sugar molar ratio of 2.45:1) was used. Since isomerization is significantly slower than reactive extraction, sugars were pre-isomerized to equilibrium (6 hr, often conducted overnight) with XI. Results in Figure 3-2 shown that about 10% sugar loss was seen during the pre-isomerization due to sugar binding to the enzyme support (confirmed by enzyme 38
supplier). Sugar loss was eliminated in subsequent rounds of isomerization if the enzyme was kept fully-immersed in the sugar solution. SIRE was conducted for ~6 hours until equilibrium isomerization was achieved as indicated by no further change in pH of the aqueous phase. Sugars were recovered from the organic phase by BE into an equal volume of 500 mM HCl. Sugar extraction during SIRE depends on both sugar concentration in the aqueous phase and affinity of the sugar for N2B. As shown in figure, SIRE significantly increased the isomerization of xylose relative to isomerization alone, and the majority of xylulose was recovered in the strip solution. More than 70% of the sugars recovered in the stripping solution were in the ketose form, demonstrating that SIRE is effective in selectively extracting the ketose sugars. Since the affinity of N2B to xylulose is the highest of the four sugars, xylulose was significantly enriched in the stripping solution.
39
500 pre-isomerization Xylulose Fructose 400 Xylose Glucose
300 after SIRE-BE
200
SugarConcentration (mM) 100
0 Initial sugar Pre-isomerized Loss associated Sugar solution Stripping solution solution solution with XI after SIRE (500mM HCl)
Figure 3-2 Results for mixed sugar equilibrium SIRE-BE. Initial sugar composition was approximately 30 g/L (200 mM) xylose and 60 g/L (333 mM) glucose. Mole ratio of
N2B (organic extraction phase) to total initial sugar was 1:1; N2B to C5 sugar was 2.45:1.
Sugars were pre-isomerized for at least 6 hr, and SIRE was conducted until both extraction and isomerization reached equilibrium at pH 8.5. Sugars were recovered in the stripping solution by one-stage BE with 500 mM HCl. Equal volumes of initial sugar, organic phase, and stripping solution were used for this experiment.
We have shown previously that the molar ratio of N2B:sugar significantly impacts the sugar extraction efficiency and selectivity for sugar isomers [15]. To establish equilibrium operating curves (yield and selectivity) for SIRE with mixed sugars, we varied
40
the N2B:C5 sugar ratios between 0.25 and 2.45. The aqueous phase and organic phase compositions were fixed; the volume ratio of the two phases was changed to alter the N2B to C5 sugar mole ratio. We have verified that changing the N2B:sugar ratio in this fashion yields results comparable to fixing the volumes and changing the aqueous or organic phase composition (data not shown). The equilibrium xylulose purity and yield curves in Figure
3-3a show that at low molar ratios of N2B to C5 sugar, the xylulose extraction yield is low but the recovered xylulose purity is high due to the limited extraction of other sugars. As the N2B:C5 sugar molar ratio increases, xylulose extraction increases but selectivity, and hence purity, decreases. Fructose, which also has high affinity for N2B, is present in the isomer mixture at high concentration and is the primary competitor for N2B binding sites in the organic phase. As shown in Figure 3-3b, fructose yield increases over the entire range tested.
The operating conditions for SIRE can be adjusted to maximize either xylulose yield or purity, depending on the needs for downstream processing. In the following sections we describe strategies focused on maximizing xylulose purity by selecting appropriate SIRE operating conditions and by adopting modified BE strategies. We also address strategies that increase high-purity xylulose yield.
41
(a) xylulose N2B:sugar mole ratio 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 80 Purity 70 Yield 60
50
40
30
20
10 Xylulose yieldor purity (%)
0 0 0.5 1 1.5 2 2.5 N2B:C5 sugar mole ratio (b) fructose N2B:sugar mole ratio 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 80 Purity 70 Yield 60
50
40
30
20
10 Fructose oryield purity (%) 0 0 0.5 1 1.5 2 2.5 N2B:C5 sugar mole ratio
Figure 3-3 (a) Xylulose and (b) fructose yield and purity with changing N2B:sugar mole ratio. Equilibrium SIRE was followed by complete sugar BE into 500 mM or 750 mM HCl. Initial sugar concentrations were 73.5 g/L xylose (490 mM) and 147 g/L (817 mM) glucose except the one with N2B to C5 sugar mole ratio at 2.45 (30 g/L (200 mM) xylose and 60 g/L (333 mM) glucose). (a) Xylulose and (b) fructose yield and purity show that xylulose purity is compromised with increasing N2B:C5 molar ratio as additional fructose is extracted into the organic phase.
42
3.3.2 Two stage BE to enhance xylulose purity
Our previous experiments with SIRE-BE showed that xylulose purity in the stripping phase could be further enhanced by employing a two-stage back-extraction technique [15]. To determine if two-stage BE would be effective for isolating xylulose from a 4-sugar mixture, the concentration of acid in the first stage stripping solution was varied to determine if sugars other than xylulose could be preferentially extracted into the first stripping solution, resulting in a more xylulose-enriched extraction. SIRE was conducted until equilibrium isomerization and sugar extraction was achieved. The organic phase was then subjected to two stages of BE.
For the first stage, the acid concentration in the stripping solution ranged from 25-
400 mM. Figure 3-4A shows the sugar composition in the stripping solution after the first stage of BE. Xylulose shows a distinct range of acid concentration (≤ 200 mM) with limited extraction due to its high affinity binding with N2B. It is important to note that over this same range of acid concentration, fructose back-extracts to a much high degree, even though it is present in the organic phase at a concentration similar to xylulose (e.g. compare
Figure 3-4A 75 mM and 500mM HCl).
To recover the sugars remaining in the organic phase, 500 mM HCl was used for the second stage BE; results are shown in Figure 3-4B. For the first stage acid concentrations tested, 200 mM HCl resulted in the highest xylulose purity in the second stage stripping solution. The result, 76% xylulose and 90% ketose, is much higher than that achieved from one-stage stripping with 500 mM HCl (39% xylulose and 72% ketose).
43
200 (A) First stage BE Xylulose Fructose Xylose Glucose 150
100
50 Sugar concentration (mM)
0 1234567 25 mM 50 mM 75 mM 100 mM 200 mM 400 mM 500 mM
Y yield, % 0 1 2 3 15 35 43 K yield, % 2 4 6 7 18 26 30 Y purity, % 0 3 6 8 21 36 39 K purity, % 45 45 49 50 63 71 72
200 (B) Second stage BE Xylulose Fructose Xylose Glucose 150
100
50 Sugarconcentration (mM)
0 1234567 25 mM 50 mM 75 mM 100 mM 200 mM 400 mM 500 mM
Y yield, % 43 43 41 40 28 8 --- K yield, % 28 26 24 22 12 4 --- Y purity, % 45 49 52 57 76 63 --- K purity, % 76 79 81 85 90 78 ---
Figure 3-4 Results of two-stage BE. N2B to C5 mole ratio used in this experiment was 2.45. Equal volumes of initial sugar, organic phase, and stripping solution were used. 44
Implementing two-stage BE is inherently a trade-off between increasing xylulose purity and decreasing xylulose yield. We hypothesized that if the organic phase was initially more pure in xylulose after SIRE, two-stage BE would allow a further increase in purity with minimal loss in yield. To test this hypothesis, we performed two-stage BE starting from two extreme SIRE conditions shown in Figure 3-3a: (1) high yield/low purity
(N2B:C5 of 2.45, the conditions used for Figure 3-4) and (2) low yield/high purity
(N2B:C5 of 0.5). Two-stage BE conditions were established independently for each case to maximize xylulose purity in the second stage stripping solution. As shown in Figure 3-
5, for the low yield/high purity SIRE conditions, two-stage BE increased xylulose purity from near 60% to over 90%, a significant second stage purity compared to those obtained with an N2B to C5 mole ratio at 2.45. Interestingly, the drop in yield in moving from a one-stage to a two-stage BE is small under the high purity/low yield conditions, so the significant enhancement in purity comes with minimal cost to yield (20% compared to 16% xylulose yield). When comparing overall ketose purity for two-stage back-extraction under either condition shown in Figure 3-5, very high purity (>90%) is achieved. Thus, if the goal is to produce high purity mixed ketose sugars, two-stage BE also benefits this strategy.
45
350 N2B:C5 = 0.5 Xylulose Fructose 300 Xylose Glucose
250 N2B:C5 = 2.45
200
150
100 Sugar concentration Sugarconcentration (mM)
50
0 One-stage BE Two-stage BE One-stage BE Two-stage BE
Y yield, % 20 16 43 28 K yield, % 10 6 30 12 Y purity, % 61 93 39 76 K purity, % 84 98 72 90
Figure 3-5 Results of two-stage BE with two N2B:C5 sugar mole ratios.
Equilibrium SIRE was followed by complete sugar BE into stripping phase plus two-stage
BE data. Organic phase samples from N2B to C5 mole ratio at 0.5 and 2.45 were back- extracted in two stages, first with a lower concentration of HCl (100 mM or 200 mM) to remove lower affinity sugars followed by a second stage extraction with 500 mM or 750 mM HCl. Data of two stage BE is second stage result with highest xylulose purity. For experiment of N2B to C5 mole ratio at 2.45, initial sugar composition was ~200 mM xylose and 333 mM glucose, and equal volumes of initial sugar, organic phase, and stripping solution were used for this condition. For experiment of N2B to C5 mole ratio at 0.5, initial sugar composition was ~490 mM xylose and 817 mM glucose, and the volume ratio of organic solution to initial sugar solution was 0.5, stripping solution had the same volume as the organic solution. 46
Fructose, which has high affinity for N2B, is the most difficult of all sugars present to separate from xylulose during BE. Conducting multiple rounds of SIRE with two-stage
BE could be used to increase overall yield of xylulose. However, as the C5 sugar is depleted in successive rounds, the C6/C5 sugar ratio increases. At some point, xylulose will be extremely difficult, if not impossible, to separate from fructose due to the disparity in concentration. To address this challenge, we next explored ways to minimize fructose production and hence its extraction to the organic phase during SIRE.
3.3.3 Non-equilibrium SIRE to enhance xylulose purity
Due to the composition of biomass, C6 sugars are present in hydrolysate at a higher concentration than C5 sugars. The equilibrium ratio of ketose to aldose is also higher for
C6 than C5 sugars. Thus, fructose is readily extracted into the organic phase under equilibrium SIRE conditions even though it has a lower affinity than xylulose for N2B.
Since fructose is the most competitive sugar to xylulose for reactive extraction during SIRE, we hypothesized that conducting SIRE under conditions where glucose to fructose isomerization was lower than equilibrium levels could enhance our ability to recover xylulose from mixed sugars while also decreasing our process time. Interestingly, the glucose isomerization rate is more sensitive to temperature and at 50°C is significantly slower than the rate of xylose isomerization to xylulose.
We implemented transient SIRE (non-equilibrium for glucose isomerization) to determine if xylulose/fructose separation could be improved. SIRE was conducted with an
N2B:C5 molar ratio of 0.5 for up to 125 min. Sugars were recovered from the organic phase using one-stage BE with 750 mM HCl. Results are shown in Figure 3-6. Fructose yield was very low during the entire transient isomerization period (less than 2%) due to its very 47
low concentration in the aqueous phase solution. Xylulose yield increased rapidly over this same time-frame as reactive-extraction of xylulose resulted in increased isomerization of xylose. After 125 min, the yield of xylulose was comparable to that achieved with equilibrium SIRE conditions (20%), so this strategy does not compromise xylulose yield.
For comparison, results for 125 min of SIRE with an N2B:C5 sugar molar ratio of 2.45 are also shown. As expected, the xylulose yields are very similar since the extent of xylose isomerization is similar. Xylulose purity was slightly higher for both cases as compared to equilibrium SIRE (65-70% compared to ~60% and 53% compared to 39%), but more importantly, the majority of other sugars extracted were aldose (glucose/xylose) which are significantly easier to separate from xylulose with two-stage BE.
48
350 N2B:C5 = 0.5 Xylulose Fructose 300 Xylose Glucose
250 N2B:C5 = 2.45
200
150
100 Sugar concentration (mM)
50
0 Equilibrium SIRE Transient SIRE Equilibrium SIRE Transient SIRE
Y yield, % 20 20 43 24 K yield, % 10 8 30 11 Y purity, % 61 69 39 53 K purity, % 84 81 72 68
Figure 3-6 Results of transient SIRE. N2B to C5 mole ratios were 0.5 (left) and
2.45 (right). Transient SIRE was followed by complete sugar back-extraction into 500 mM or 750 mM HCl. The sugar compositions and volume ratios of different phases for these two conditions are those of Figure 3-5. Result of N2B to C5 sugar mole ratio at 0.5 is better than that of 2.45 because of higher purity with similar yield.
49
3.3.4 Multiple cycles of transient SIRE
To increase the total yield of xylulose, five additional 125 min cycles of transient
SIRE were conducted at 50°C on the sugars remaining in the aqueous phase. The concentration of N2B was held constant at 490 mM. Initial sugar composition was approximately 73.5 g/L xylose (490 mM) and 147 g/L (817 mM) glucose. The N2B:C5 ratio of 0.5 was used for these experiments because of the higher xylulose purity (69% versus 53%). After each cycle of SIRE, the residual aqueous phase composition was analyzed. For each new cycle, the volume ratio of the aqueous to organic phase was adjusted to maintain a fixed value of 0.5 for the mole ratio of N2B to C5 sugar to retain high purity extraction of xylulose into the organic phase. After each SIRE cycle, the organic phase sample was processed by either one-stage or two-stage back-extraction.
For one-stage BE, the organic phase from each cycle was back-extracted with 750 mM HCl. Cumulative yield and purity were calculated from the concentrations of sugars measured in the stripping phase for each cycle. To further enhance xylulose purity, two- stage back-extraction was conducted on the organic phases. The organic phases from sequential cycles were first combined and then two-stage BE was conducted to maximize xylulose purity in the second stage. For the first stage BE, the acid concentrations used in the stripping solution ranged from 25-400 mM; 750 mM HCl was used for all second stage back-extractions. Data of two-stage BE shown here is second stage result with the highest xylulose purity
The results for xylulose purity and yield with multiple transient (125 min) cycles of SIRE are shown in Figure 3-7. The cumulative purity achieved by adding cycles of transient SIRE with one-stage BE is higher than that achieved with equilibrium SIRE for 50
all six cycles. Yield was significantly improved by conducting multiple cycles of SIRE – after four cycles, yield increased from 20% to 45%.
After two cycles of transient SIRE with two-stage BE, xylulose purity was comparable to that seen with equilibrium SIRE, but overall xylulose yield was 50% higher
(compare to Figure 3-5, 16% versus 25% yield). For successive cycles of transient SIRE with two-stage BE, the xylulose purity was 88% or higher, which is 25-30% higher than that achieved with one-stage BE with only a 5% loss in yield relative to the one-stage BE results. This is because the concentration of fructose was less compared with equilibrium
SIRE and xylulose was easier to separate from aldose sugars. Therefore, to overcome the inverse trends of xylulose purity and yield, mixed sugars can be processed by four successive cycles of transient SIRE followed by two-stage BE to achieve high purity
(~90%) and relatively high yield (~40%) of xylulose.
51
25 (A) One-stage BE
Xylulose Fructose 20 Xylose Glucose
15
10 Cumulative Cumulative sugar,mmol 5
0 1 2 3 4 5 6
Y yield, % 20 30 39 45 49 53 K yield, % 8 14 20 24 28 31 Y purity, % 69 67 63 60 57 54 K purity, % 80 79 79 79 79 80 25 (B) Two-stage BE
20
15
10 Cumulative Cumulative sugar,mmol 5
0 1 2 3 4 5
Y yield, % 18 25 33 39 43 K yield, % 7 11 14 17 19 Y purity, % 94 92 90 89 88 K purity, % 97 96 96 96 96
Figure 3-7 Results for six cycles of transient SIRE for mixed sugar SIRE at 50°C.
The cumulative data was from all the finished cycles. 52
3.3.5 Purity of C6 sugars in the aqueous phase
As a consequence of removing high purity xylulose from the simulated hydrolysate, the aqueous phase becomes enriched in C6 sugars. After each 125 min SIRE cycle, the sugars composition in the aqueous phase was analyzed. As shown in Figure 3-8 and Table
3-1, as xylulose is removed from the sugar mixture, C6 sugars increase from 63% in the initial sugar solution to 80% after six cycles of transient SIRE. At this point, glucose and fructose ratios in the aqueous phase are still below equilibrium ratios, indicating that additional cycles could be conducted that would favor xylulose removal. For the transient
SIRE cycles, the cycle time of 125 min was held fixed for each cycle. However, as the concentration of C5 sugar decreases, the time of each subsequent cycle could be decreased, without inhibiting xylulose production, to further reduce fructose formation.
After C5 sugar is extracted to a sufficient extent, SIRE conditions could then be tailored to separate fructose from glucose, including increasing the temperature to 60°C to significantly increases the rate of glucose isomerization. Thus, the goal of producing separate high-purity streams of xylulose and fructose from mixed sugar hydrolysates can be achieved by modifying the SIRE process and BE conditions, first isolating xylulose in high purity and yield, followed by recovery of fructose from the glucose/fructose-rich residual.
53
1200 Xylulose Xylose Fructose 1000 Glucose
800
600
400
Sugar concentration (mM) 200
0 Initial Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6
Figure 3-8 Aqueous phase sugar composition after multiple cycles of transient
SIRE. The composition of the aqueous phase was analyzed prior to isomerization and after each 125-min cycle of SIRE. Experimental conditions are those of Figure 3-7.
Table 3.1 Proportion of C6 sugars and ratio of fructose to glucose in aqueous phase after each cycle.
54
3.4 Conclusions
When processing saccharified lignocellulosic biomass hydrolysates, isolation of C5 and C6 sugars into separate process streams may be beneficial, and conversion of the aldose sugars to their ketose isomers may confer additional benefits. Based on the key observations of differences in isomerization kinetics between xylose and glucose, and different sugar binding affinities with N2B, a novel process based on SIRE-BE is proposed here. In this process, high purities of xylulose and fructose, obtained as mixtures or separate sugar streams, are achieved from mixed aldose sugars representative of biomass hydrolysate.
We demonstrated that the conditions of SIRE can be customized to maximize xylulose purity or xylulose yield under equilibrium conditions by changing the N2B:C5 sugar mole ratio. Transient SIRE can be used to exploit the differential isomerization kinetics of glucose and xylose to further increase xylulose purity of the separation.
Implementing two-stage back extraction further increases xylulose purity and little loss of xylulose yield. Xylulose yield can be increased through repeating SIRE for multiple cycles.
Best overall results were achieved using 5 cycles of transient SIRE, resulting in 40% xylulose yield with ~90% xylulose purity. This is particularly significant considering that
C5 is a minor sugar, initially present in a mole ratio of xylose:glucose of 0.6, a ratio expected for biomass hydrolysate.
As C5 sugars are removed from the hydrolysate, C6 sugars are enriched in the aqueous phase. By applying SIRE to the predominantly glucose/fructose mixture, a fructose-rich stream can also be produced. Thus, strategic use of SIRE-BE conditions can
55
fractionate C5 and C6 sugars, providing high-purity ketose sugar streams for downstream processes.
56
Chapter 4
New Route of C5 Sugars Fermentation to Ethanol by Native Saccharomyces cerevisiae
4.1 Introduction
Ethanol is used primarily as fuel in industry in US [50]. Fuel ethanol, which is
currently produced from food-based sugars, can also be produced via fermentation of
sugars derived from lignocellulosic biomass. Lignocellulosic biomass is an abundant, less
expensive feedstock that does not compete with the food supply. However, utilization of
both C6 (glucose) and C5 (xylose) sugars is necessary to make ethanol production from
biomass economically feasible [12-14, 16, 17].
Ethanol-producing bacteria have attracted much attention because their growth
rates are substantially higher than yeast. The main problems in the use of bacteria for
fermentation are low ethanol tolerance and reduced product selectivity toward ethanol [18-
20].
Significant progress has been made in the development and testing of genetically
modified organisms (GMO) capable of metabolizing both C5 and C6 sugars. However, the
genetic stability, overall ethanol yield, and ability to survive under the conditions of
57
industrial fermentation are unproven for GMOs [13, 21]; as such, approaches that employ native strains are highly desirable.
Large-scale fermentation of sucrose and starch-based glucose for fuel ethanol use wild-type Saccharomyces cerevisiae due to its near-theoretical ethanol yields from glucose and tolerance to elevated osmotic pressure and a wide spectrum of inhibitors. The metabolic pathways in native S. cerevisiae convert xylose to xylitol [22-24] via an
NADPH-linked xylose reductase (XR). Although xylitol can be oxidized to xylulose by an
NADH-linked xylitol dehydrogenase (XDH), a redox imbalance between the cofactor usage of XR and XDH has been attributed to xylitol accumulation and disruption of metabolic flux to xylulose in long-term fermentation [18, 19, 25-29]. A comparative study of various xylose-assimilating yeast showed that significant fermentation of xylose to ethanol through xylulose only occurred in those organisms which possessed an NADH- linked aldose reductase [27]. Pichia stipitis, one of the best naturally occurring xylose- fermenting yeasts, can convert xylose to ethanol in high yield. However, this yeast has low ethanol and sugar tolerance, preventing its use as an industrial strain for large-scale ethanol production from lignocellulosic biomass sugars [16].
Fermentation of xylulose to ethanol by S. cerevisiae in high yield has been shown previously [30, 31], but the isomerization of xylose to xylulose exhibits an unfavorable equilibrium (~20% xylulose) that limits the practical utility of xylulose fermentation. High- yield, low cost production of xylulose from xylose has the potential to eliminate the bottleneck currently encountered for ethanol production from C5 and C6 sugars with native yeast.
58
We have previously developed a reactive-extraction based method for the high yield production of xylulose from xylose [15]. This method couples enzymatic isomerization of xylose to xylulose and selective reactive-extraction of the ketose isomer into an organic phase. Concentration and recovery of the extracted sugar is achieved by back-extraction into a low pH buffer, and xylulose purity ranges from 90 to nearly 100% based on the extraction pH employed. In this chapter, using high purity xylulose and glucose/xylulose mixtures obtained from SIRE-BE, the feasibility of ethanol production from both C5 and C6 sugars with native yeast will be discussed.
4.2 Method and Material
4.2.1 Chemicals and yeast strain
D-glucose, D-xylose and D-fructose were purchased from Sigma Aldrich Co. (St.
Louis, MO, USA); D-xylulose used for HPLC standards was purchased from ZuChem
(Chicago, IL, USA). High purity, concentrated D-xylulose, used for yeast fermentation and xylulose metabolism was prepared, in-house, from xylose by simultaneous isomerization- and-reactive-extraction (SIRE) followed by one or two-stage back-extraction (BE). All additional chemicals and solvents were purchased from Thermo Fisher Scientific Inc.
(Pittsburgh, PA, USA). Immobilized xylose isomerase (Gensweet® IGI, XI) was a kind gift from Dupont Industrial Biosciences (Wilmington, DE). Immobilized XI was stored at 4°C until use.
Baker’s yeast (S. cerevisiae, YSC2, type II, >90% active yeast) was purchased in dried pellet form from Sigma Aldrich Co. (St. Louis, MO). Yeast was stored at 4°C until use and was added to the fermentation media as dry yeast.
59
4.2.2 Xylulose production from xylose using SIRE-BE
The SIRE-BE process for xylulose produced from xylose has been described previously [15]. Briefly, SIRE was conducted here using a 10 mM solution of xylose in 50 mM sodium phosphate buffer at pH 8.5 and 50 °C with 4.5 g/l Gensweet® IGI (immobilized xylose isomerase); the aqueous sugar mixture was contacted with an equal volume of organic phase (octanol containing 33 mM naphthalene-2-boronic acid (N2B) and 82.5 mM
Aliquat® 336). The sugars extracted into the organic phase were back-extracted into 0.1 M
HCl solution. 90% purity of xylulose was produced from one-stage BE; 99% purity of xylulose was produced from two-stage BE. After SIRE-BE, xylulose was concentrated to
~30 g/L by evaporation.
4.2.3 Fermentation media
Fermentation media contained 30 g/l xylulose (from SIRE-BE). Sodium citrate was added to 0.05 M and the pH was adjusted to 4.5 using NaOH. Glucose at 30 g/l prepared in the same manner was used as the control. The concentration of dry yeast used for sugar fermentation was 50 g/l for all fermentations. N2 gas was flushed into the shake flask to maintain anaerobic conditions every time the flask was opened. Fermentation was performed at 34 °C in a 250 ml shake flask agitated at 130 rpm in a water bath shaker (C76,
New Brunswick, NJ, USA). Shake flasks were capped with rubber stoppers, and a one-way valve was used to release any pressure that built up during the experiment.
4.2.4 Analytical techniques and data analysis
During fermentation, samples periodically collected from the experiments were centrifuged, and the supernatants were analyzed for sugars, ethanol, glycerol, xylitol and acetic acid diluted in ultrapure water and filtered through a 0.22 µm pore-size sterile filter. 60
Calibration standards for sugars, ethanol and by-products were prepared in a similar manner. All samples and standards were analyzed using two Shodex SH1011 columns
(300×8 mm, from Showa Denko K.K, Japan) in series on an Agilent 1100 HPLC system equipped with a refractive index detector (RID). A mobile phase of 0.05 M H2SO4 was run at 0.6 ml/min; a column temperature of 50 °C and detector temperature of 35 °C were used for optimal peak resolution and detection.
4.3 Results and Discussion
Theoretically, 1 g of glucose (or xylulose) will produce 0.51 g of ethanol and 0.49 g of carbon dioxide. The stoichiometric fermentation of xylulose and glucose to ethanol and CO2 are:
3C5H10O5 → 5C2H5OH + 5CO2
C6H12O6 → 2C2H5OH + 2CO2
In practice, the microorganisms use some of the sugar for growth and the actual yield is less than 0.51g ethanol/ g sugar [18]. Other abundant products that were identified include xylitol (C5H12O5) produced from five carbon sugars and glycerol (C3H8O3) from both sugars.
4.3.1 Glucose Fermentation
Fermentation of glucose solution was used as the control to determine the sugar utilization and the ethanol yield by native S. cerevisiae. In our high density 50 g/l yeast fermentation (Figure 4-1), pure glucose was consumed quickly. Within 3 h, all the sugar was consumed and ~14 g/l of ethanol was produced. The ethanol yield was 0.46 g ethanol/g
61
sugar, or 90% of the theoretical value. The most abundant byproduct was glycerol, although acetic acid was also detected (data not shown).
30 A. 30 g/l glucose
25 glucose ethanol glycerol
20
15
10 Concentration Concentration (g/l)
5
0 0123456 Time (h)
Figure 4-1 Results for glucose fermentation to ethanol. 50 g/L native yeast and 30 g/L glucose were used. Fermentation was conducted at 34 °C with pH 4.5.
4.3.2 Xylulose Production and Fermentation
As shown in Figure 4-2, the overall process for xylulose fermentation to ethanol requires three steps: (1) simultaneous-isomerization-and-reactive extraction (SIRE) to generate large quantities of xylulose from xylose; (2) selective back-extraction (BE) of xylulose to recover the sugar into an aqueous solution compatible with fermentation; and
(3) xylulose fermentation to ethanol with native yeast. Following BE, the pH is adjusted to
4.5 for xylulose fermentation to ethanol with native yeast at 34 °C.
62
Step 2: Step 1: Back extraction of sugars from the organic phase Step 3: SIRE with extraction of Fermentation of xylulose sugars to organic phase to ethanol Xylose (X) into Xylulose (Y) mid-pH strip into low-pH Immobilized phase strip phase Organic phase – GXI column octanol, N2B & Q+Cl-
X X·N2B-Q+ - + X·N2B Q X - + - + Y·N2B Q Y Y Y·N2B Q Y·N2B-Q+
Xylulose-rich medium EtOH
X, Y HCl HCl Y EtOH pH 8.5 pH 4-6 pH 1-3 pH 4.5 T = 50 °C T = 34°C Xylose-rich medium acidic medium
Figure 4-2 Schematic representation of the three-step process producing xylulose
(Y) in high yield from xylose (X) and its fermentation by native S. cerevisiae to ethanol.
In step 1, the aldose to ketose isomerization is effected with commercially-available immobilized xylose isomerase (XI) enzyme. The temperature (50 °C) at which the enzyme effectively catalyzes the isomerization is compatible with saccharification, the last step for production of biomass hydrolysate. To overcome the unfavorable isomerization equilibrium (~20% xylulose), we employ SIRE to extract the xylulose as it is formed. The selective extraction of xylulose from the aqueous phase solution is facilitated by the addition of naphthalene-2-boronic acid (N2B) and Aliquat® 336 to the organic phase. N2B preferentially binds to xylulose, and ion-pair formation between Aliquat® 336 and the sugar-N2B complex confines the complex to the organic phase. The differential, pH- dependent affinity of the N2B for xylulose over xylose not only influences their selective extraction but also the relative ease with which they can be dissociated from N2B and concentrated in acid media through BE in step 2. The relatively low affinity of xylose for 63
the N2B allows its back-extraction under moderately-acidic conditions while the more strongly bound xylulose require more acidic conditions for dissociation. By implementing a 2-stage BE process with mid and low pH stripping solution, the xylulose can be recovered as a nearly-pure, concentrated aqueous stream in stage 2. The last step in the overall process is then a pH adjustment to 4.5 and the fermentation of xylulose to produce ethanol at 34 °C and pH 4.5 with native yeast.
High purity, concentrated xylulose was successfully produced from xylose by SIRE followed by one stage BE (90% xylulose purity) or two stage BE (99% xylulose purity)
[15]. As shown previously by Chiang et al. [30], high yeast densities can promote better utilization of C5 sugars. By using high yeast density for fermenting high concentration xylulose solutions, the fermentation time is reduced significantly and ethanol yield is improved.
For xylulose fermentation, xylulose/xylose solution produced from isomerization alone (25% xylulose/75% xylose) was used as a control. The results of this control fermentation are shown in Figure 4-3. The ethanol production stopped when xylulose was consumed. The native yeast cannot convert the xylose to ethanol, but is capable of converting xylose to xylitol as expected.
64
30 25% xylulose (control group)
xylose xylulose ethanol 25 xylitol glycerol
20
15
10
Concentration (g/l) Concentration 5
0 0 4 8 12 16 20 24 Time (h)
Figure 4-3 Results for fermentation of a 25% xylulose/75% xylose solution with 30 g/L total C5 sugars. Data presented are for duplicate experiments; error bars are too small to show. Xylitol, which is the predominant by-product, results from xylose reductase- xylitol dehydrogenase pathway.
A 25/75 mixture is the highest xylulose proportion possible without shifting the reaction by product removal, such as SIRE. Our next experiments utilized the very high purity xylulose produced via SIRE-BE. Fermentations were conducted using 90% xylulose
(one-stage BE) and 99% xylulose (two-stage BE) solutions. As shown in Figure 4-4, the ethanol produced from both 90% and 99% xylulose fermentations were very similar, a yield of 0.44-0.45 g ethanol/ g xylulose, which is 88% of the theoretical yield. Interestingly, the xylitol production was also similar for both cases, which means that xylitol was
65
produced from xylulose and xylose. In previous research on mixed xylulose/xylose fermentation by Chiang and Roman et al. [30, 31], xylitol was found to be a major by- product. Based on these results, increasing xylulose content to over 90% does not improve overall ethanol production. Thus, using one-stage BE to produce 90% xylulose is sufficient for maximum ethanol production from C5 sugars.
66
30 90% xylulose (SIRE with one-stage BE)
xylose xylulose ethanol 25 xylitol glycerol
20
15
10
Concentration (g/l) Concentration 5
0 0 4 8 12 16 20 24 (A) Time (h)
30 99% xylulose (SIRE with two-stage BE)
xylose xylulose ethanol 25 xylitol glycerol
20
15
10
Concentration (g/l) Concentration 5
0 0 4 8 12 16 20 24 (B) Time (h)
Figure 4-4 Results for fermentation of xylulose produced from SIRE for (A) 90%, and (B) 99% xylulose purity. Fermentation was conducted at 34 °C with pH 4.5. Data presented are based on duplicate experiments; error bars are too small to show. Xylitol is the predominant by-product of both fermentations. 67
4.3.3 Glucose/Xylulose Fermentation
For xylulose fermentation, it is important to note that the rate of xylulose utilization and ethanol production was 10 times slower than that of glucose fermentation, see Figure
4-1, 4-4. This is not a surprising result considering that yeast is not known to express a xylulose-specific transporter. Xylose is shuttled across the cell membrane by high and intermediate affinity hexose transporters, for which the affinity for xylose is usually two orders of magnitude lower than for glucose. Since xylulose is the keto-isomer of xylose, it may be transported similarly.
Since glucose and xylulose likely share a transporter, and low concentrations of glucose are known to up regulate the high affinity hexose transporter, we next fermented mixtures of glucose and xylulose to see if the presence of glucose aided or hindered xylulose uptake and fermentations. Varying amounts of glucose (5 g/L, 10 g/L, 20 g/L, and
30 g/L) were added to the xylulose fermentation media. As shown in Figure 4-5, the native yeast first consumed all the glucose, and then started to utilize xylulose. It took approximately one hour to consume all the glucose, and it took approximately twelve hours to consume all the xylulose. These results are the same as single sugar fermentations with only glucose or xylulose.
68
30 With 5 g/l glucose 30 With 10 g/l glucose
glucose xylulose ethanol glucose xylulose ethanol 25 25 xylitol glycerol xylitol glycerol
20 20
15 15
10 10 Concentration (g/l) Concentration 5 (g/l) Concentration 5
0 0 0 4 8 12 16 20 24 0 4 8 12 16 20 24 Time (h) Time (h)
30 With 20 g/l glucose 35 With 30 g/l glucose
glucose xylulose ethanol 25 glucose xylulose ethanol 30 xylitol glycerol xylitol glycerol 25 20 20 15 15 10 10 Concentration (g/l) Concentration Concentration (g/l) Concentration 5 5
0 0 0 4 8 12 16 20 24 0 4 8 12 16 20 24 Time (h) Time (h)
Figure 4-5 Results for supplemental glucose (5 g/L, 10 g/L, 20 g/L, and 30 g/L) added to a fermentation media containing 30 g/l C5 sugar (99% xylulose). Data presented are based on duplicate experiments; error bars are too small to show.
As shown in Figure 4-6 and summarized in Table 4.1, the addition of glucose to the fermentation neither enhanced the utilization of xylulose nor improved ethanol yield from total fermentable sugars (xylulose+glucose). Mass balances were based on carbon present in the initial sugars and in the reaction products (ethanol, xylitol, glycerol, acetic acid (data not shown), and carbon dioxide). Carbon dioxide was not measured, but was assumed based on the reaction stoichiometry given in Section 4.3.
69
100 90 80 70 60 50 40 Concentration of added glucose 30 0 g/l 5 g/l 10 g/l 20 20 g/l 30 g/l
xylulose consumption (%)consumption xylulose 10 0 0 4 8 12 16 20 24 Time (h)
Figure 4-6 Comparison of xylulose consumption with and without added glucose.
Data shown are a summary of fermentations shown in Figure 4-4B, and 4-5.
70
Table 4.1 Summary of C5 and C5/C6 fermentations. Results are analyzed at 24
hours. G= glucose; X= xylose; and Y= xylulose. Yield is defined as g ethanol/g sugar
consumed.
Concentrations at 24 hours Initial Sugar Ethanol Xylitol Mass Sugar consumption Sugars Ethanol Yield Xylitol Yield Glycerol Balance (g/l) rate (g/hr) (g/l) (g/l) (g/g) (g/l) (g/g) (g/l) (%) (%theoretical)
30 G 1.5 G 0 G 14.11 0.46 (90%) 0 0 3.85 98
3.2 X 2.68 X 24.6 Y 0.19 Y 10.95 0.45 (88%) 6.83 0.28 1.40 110 0 Y (90%) 0.5 X 0.41 X 27.2 Y 0.18 Y 11.65 0.44 (86%) 6.21 0.23 1.48 110 0.22 Y (99%)
5 G 0 G 0.18 Y 13.86 0.41 (80%) 6.23 0.22 2.10 107 28.6 Y 0.26 Y
10 G 0 G 0.16 Y 15.62 0.40 (78%) 5.88 0.21 2.65 102 28.6 Y 0.37 Y
20 G 0 G 0.15 Y 16.79 0.37 (73%) 4.76 0.17 3.71 96 28.6 Y 0.32 Y
30 G 0 G 0.12 Y 22.02 0.38 (75%) 4.47 0.16 4.85 96 28.6 Y 0.39 Y
Based on these results, the highest ethanol productivity will likely occur with
separate C5 and C6 fermentation. Hence, the strategy presented in Chapter 3 for producing
separate xylulose and fructose streams may be worth exploring for fermentation of biomass
sugars to ethanol.
71
4.3.4 Schematic route for separation of xylulose from C6 sugars, followed by fermentation
Both C5 and C6 sugars are present in lignocellulosic biomass. Based on the results of the last section, the best ethanol productivity occurs with separate C5 and C6 fermentation. SIRE can be used to separate C5 from C6 sugars due to differences in their rates of isomerization as shown in Chapter 3. Conversion of xylose to xylulose is much faster than that of glucose to fructose at 50 °C. High purity xylulose can be generated as a separate sugar stream from a 4-sugar mixture by using appropriate strategies of SIRE and
BE, as demonstrated in the last chapter. A proposed schematic route for isolating xylulose from C6 sugars is shown in Figure 4-7 below. With two rounds of SIRE followed by two- stage BE, nearly all xylose is converted to xylulose. The first round SIRE includes multiple cycles of transient SIRE. The first BE fraction (BE1-1) is feed for the second round of
SIRE. The first BE fraction from SIRE 2 (BE2-1) is nearly all C6 sugars while the BE1-2 and BE2-2 fractions are nearly pure xylulose.
multiple Xylulose Glucose cycles (Glucose) BE1-2 Xylulose Xylulose Xylose transient (Xylose) (Fructose) Fermentation SIRE Fructose
BE1-1 BE2-2
Glucose (Xylulose) SIRE BE2-1 Glucose C6 Fructose Glucose Fructose Fermentation (Xylose) Fructose
Figure 4-7 Schematic route for separation of C5 and C6 sugars from biomass hydrolysate for fermentation. 72
4.4 Conclusions
Fermentations were conducted from ~30 g/L C5 sugars with purities of 25%, 90%, and 99%. After 12 hours, the ethanol yield of 99% xylulose is 0.44g/g, which is 85.5% of the theoretical value; the ethanol yield of 90% xylulose is 0.45g/g, which is 87.9% of the theoretical value. Since lignocellulosic biomass is an inexpensive, nonfood source of fermentable sugars, ethanol has been produced from many biomass sources. Since xylose cannot be used directly by native yeast, GMO strains have been developed to be enable xylose conversion to ethanol. As summarized in Table 4.2, C5 sugar fermentations with sugar concentrations ranged from 10 g/L to 150 g/L have been conducted with varying degrees of success in producing ethanol. Our results for xylulose fermentations compare favorably with these GMO strains both for ethanol yield and fermentation time. However, fermentation of high purity of xylulose produced from xylose with SIRE avoids the controversy of using GMOs and utilizes a highly productive, industrially proven yeast.
The rate of xylulose utilization and ethanol production is 10 times slower than that of glucose fermentation. Xylulose transport and fermentation of xylulose in
Saccharomyces cerevisiae were not aided by the presence of glucose in the fermentation
(and upregulation of the high affinity hexose transporter). The addition of glucose neither enhanced the utilization of xylulose nor improved the total ethanol yield.
73
Table 4.2 Ethanol production by different GMO strains from C5 sugar.
Name of Strain C5 concentration Ethanol Yield Time Reference (g/L) (g/g) (hours) De-Acetylation/ Zymomonas mobilis 50 0.47-0.48 N/A [51] A7 P.stipitis NCIM 3499 38.50 0.42 72 [52] P.stipitis NCIM 3498 32.15 0.36 48 [53] Evolutionary engineering 10 0.51 25 [54] S.cerevisiae Recombinant yeast 150 0.37 100 [16] ScF2 MTCC174 31.6 0.33 36 [55] S.cerevisiae CBS 8066 S.cerevisiae 9.07 0.46 50 [56] ATCC 26602 23.6 0.43 6 [57] S.cerevisiae K.marxianus CCA 510 18.59 0.38 4 [58] TISTR 5048 240 (total sugars) 0.48 120 [59] S.cerevisiae Non-GMO Yeast 30 0.44-0.45 12 Ch. 4
We have concluded that for fermentation of C5 and C6 biomass sugars to ethanol with native yeast, separation of xylulose from the C6 sugars and separate fermentation of each allows for highest ethanol yield. The methods devised in Chapter 3 could be adapted to accomplish separation of xylulose from the C6 sugars for this purpose.
74
Chapter 5
Production of 2,3-butanediol from Single and Mixed Sugars by Enterobacter cloacae NRRL B-23289
5.1 Introduction
The global butanediol (BDO) market reached 3990.0 kilotons in 2016 and is
estimated to grow at a compounded annual growth rate (CAGR) of 5.2% to 2022. Asia-
Pacific is the market leader, accounting for more than 60% of total volume consumed in
2016. The market for BDO is expected to advance at a rapid pace during the next few years,
and Asia-Pacific will remain the fastest growing region for butanediol. Europe and North
America constitute the second and third largest markets for butanediol [60].
2,3-Butanediol (2,3-BD) is a promising bulk chemical due to its extensive industrial
applications as a key building block in making polymers and hydrocarbon fuels. For
example, 2,3-BD can be readily converted to butenes, butadiene, and methyl ethyl ketone
that are used in the production of hydrocarbon fuels [34]. 2,3-BD is a colorless and nearly
odorless viscous liquid at room temperature. With a heating value of 27.2 kJ/g, it compares
favorably with ethanol and methanol for use as a liquid fuel and fuel additive. It is also an
important intermediate chemical for production of tetrahydrofuran, polytetramethylene
ether glycol (PTMEG), polybutylene terephthalate (PBT), gamma-butyrolactone (GBL),
75
polyurethane (PU), and other solvents. These chemicals are widely used in fibers, engineering plastics, medicines, cosmetics, artificial leathers, pesticides, plasticizers, hardeners, solvents and rust removers [35].
There are many petroleum-based chemical methods to produce 2,3-BD today, but a bio-based process would have a favorable greenhouse gas balance and less pollution when compared to petroleum-based processes. The unstable global oil price as shown in
Chapter 1 is another reason to focus on a bio-based process. 2,3-BD can be produced directly from sugars by fermentation using bacteria.
Enterobacter cloacae NRRL B-23289, isolated from decaying wood/corn soil samples by the USDA Agricultural Research Service (Peoria, IL), is a natural producer of
2,3-BD. Previous work at the USDA ARS has shown that this strain is more efficient in converting ketose than aldose sugars to 2,3-BD [37]. Particularly noteworthy is that fermentation of fructose showed higher 2,3-BD yield within a much shorter period of time as compared to glucose [37].
To our knowledge, no one has tested the effectiveness of this organism in converting the ketose sugar xylulose to 2,3-BD; this is likely due to the difficulty of producing high purity xylulose cost-effectively. In this chapter, we present the results of experiments using NRRL B-23289 for 2,3-BD production from single and mixed sugars.
5.2 Method and Material
5.2.1 Chemicals and bacteria strain
D-glucose, D-xylose, D-fructose, N2B, and Aliquat® 336 were purchased from
Sigma Aldrich Co. (St. Louis, MO, USA); D-xylulose used for HPLC standards was purchased from ZuChem (Chicago, IL, USA). High purity, concentrated D-xylulose, used 76
for 2,3-BD fermentation was prepared from xylose by simultaneous isomerization-and- reactive-extraction (SIRE) at 50 °C followed by a two-stage back-extraction (BE) as described previously. All additional chemicals and solvents were purchased from Thermo
Fisher Scientific Inc. (Pittsburgh, PA, USA).
Enterobacter cloacae NRRL B-23289 was obtained from USDA ARS.
5.2.2 Xylulose production by SIRE-BE
The SIRE-BE process has been described previously [15]. Briefly, SIRE was conducted here using a 10 mM of xylose in 50 mM sodium phosphate buffer at pH 8.5 and
50 °C with 4.5 g/l Gensweet® IGI (immobilized xylose isomerase); the aqueous sugar mixture was contacted with an equal volume of organic phase (octanol containing 33 mM naphthalene-2-boronic acid (N2B) and 82.5 mM Aliquat® 336). The sugars extracted into the organic phase were back-extracted into 0.1 M HCl solution. After SIRE-BE, xylulose was concentrated to ~50 g/L by evaporation
5.2.3 Fermentation media
Pre-cultures of the E. cloacae NRRL B-23289 were produced by transferring a full loop of cells from agar plate to 50 ml of media formulated with yeast extract, micronutrients, and 50 g/l sugars [61]. The pre-culture was grown in a 250 ml shake flask at 30 °C for 20-
24 hours with continuous mixing. For production of 2,3-BD, 5 ml of the pre-culture was transferred to 50 ml of fermentation media containing 50 g/l of corresponding sugar; fermentations were conducted for 3 days. All media formulations were at pH 5.0 in
KH2PO4/Na2HPO4 buffer.
77
5.2.4 Analytical techniques and data analysis
During fermentation, samples periodically collected from the experiments were centrifuged, and the supernatants were diluted in ultrapure water, filtered through a 0.22
µm pore-size sterile filter, and analyzed for sugars and 2,3-BD. Calibration standards for sugars and 2,3-BD were prepared in a similar manner. All samples and standards were analyzed using two Shodex SH1011 columns (300×8 mm, from Showa Denko K.K, Japan) in series on an Agilent 1100 HPLC system equipped with a refractive index detector (RID).
A mobile phase of 0.05 M H2SO4 was run at 0.6 ml/min; a column temperature of 50 °C and detector temperature of 35 °C were used for optimal peak resolution and detection.
5.3 Results and Discussion
5.3.1 Single sugar fermentation
Four single sugars (glucose, fructose, xylose and xylulose) were fermented. The theoretical yield for fermentation of both C5 and C6 sugars is 0.5 g 2,3-BD per g sugar
(0.50 g/g). As shown in Figure 5-1, rates of sugar consumption for both ketose sugars
(fructose and xylulose) were faster than the aldose sugars. C6 aldose or ketose sugars were consumed faster than their C5 counterparts. Our results are consistent with those previously reported for fructose, glucose, and xylose [37]; however, no one has previously reported xylulose fermentation to 2,3-BD for this bacterium.
78
54 Xylose 48 Xylulose 42 Glucose Fructose 36 30 24 18 12 6
Sugar Concentration (g/L) Sugar Concentration 0 0 8 16 24 32 40 48 56 64 72 Time (hrs)
24
20
16
12
8 2,3-BD from Xylose 2,3-BD from Xylulose 4 2,3-BD from Glucose 2,3-BD from Fructose 0 2, 3-BD Concentration(g/L)2, 3-BD 0 8 16 24 32 40 48 56 64 72 Time (hrs)
Figure 5-1 Results for fermentation of 50 g/l sugar (xylose, xylulose, glucose, or fructose) to 2,3-BD by Enterobacter cloacae NRRL B-23289 at pH 5.0, 30 °C. Values reported are from experiments conducted in duplicate. Lines are intended to show the trend in the collected data.
79
For 2,3-BD production, a plateau was reached after 48 hours for the ketose sugars.
Fructose and xylulose fermentation yielded the highest concentrations of 2,3-BD (20.98 and 20.68 g/l), or 84% of the theoretical value. Glucose and xylose fermentations were less productive, with 18.60 and 18.50 g/L 2,3-BD, or ~74% of the theoretical value.
Both acetoin and ethanol have been reported as byproducts of sugar fermentation to 2,3-BD [37]. In the experiments conducted and presented here, no acetoin was detected, but low ethanol concentrations (data not shown) were detected. The aldose sugar fermentations produced more ethanol than the ketose sugar fermentations. The nearly pure xylulose produced from our SIRE-BE process can be used to produce 2,3-BD in higher yield than its aldose counterpart xylose and with fewer contaminating by-products.
5.3.2 Mixed sugars fermentation
To simulate a biomass hydrolysate mixed sugar fermentation, either non- isomerized glucose/xylose fermentations or simulated fully isomerized fructose/xylulose fermentations were conducted. The fermentation media contained 35 g/l of C6 sugars
(glucose or fructose) and 15 g/l of C5 sugars (xylose or xylulose). As shown in Figure 5-2,
C6 sugars (glucose or fructose) were preferentially consumed before C5 sugars (xylose or xylulose). The aldose sugars were consumed more slowly, taking between 48-64 hours for complete consumption, whereas the ketose sugars were all consumed within 48 hours.
These data are consistent with the results from the single sugar fermentations.
80
36 A Xylose 30 Glucose 2,3-BD 24
18
12
Concentration (g/L)Concentration 6
0 0 8 16 24 32 40 48 56 64 72 Time (hrs)
36 B Xylulose 30 Fructose 2,3-BD 24
18
12
Concentration (g/L)Concentration 6
0 0 8 16 24 32 40 48 56 64 72 Time (hrs)
Figure 5-2 Results of sugar consumption and 2,3-BD production for fermentation of 50 g/l (A) xylose and glucose, and (B) xylulose and fructose by E. cloacae NRRL B-
23289 at pH 5.0, 30 °C. Values reported are from experiments conducted in duplicate.
Lines are intended to show the trend in the collected data.
81
The mixed ketose sugar fermentation produced 20.26 g/l of 2,3-BD with a yield of
0.41 g/g, 82% of the theoretical yield. The mixed aldose sugar fermentation was less productive with 17.97 g/l of 2,3-BD (0.37g/g), 74% of the theoretical yield.
The 2,3-BD productivity for the single and mixed sugar fermentations were very comparable (single ketose versus mixed ketoses; single aldose versus mixed aldoses).
Ketose sugar fermentations were complete with higher 2,3-BD yield in a shorter time than the aldose fermentations.
5.4 Conclusions
Based on the results of single sugar fermentations (glucose, xylose, fructose or xylulose), the nearly pure xylulose produced from our SIRE-BE process can be used by
Enterobacter cloacae NRRL B-23289 to produce 2,3-BD in higher yield than its aldose counterpart xylose and with fewer contaminating by-products.
Since both fructose and xylulose fermentations were more productive and faster than their aldose counterparts glucose and xylose, mixed glucose/xylose and fructose/xylulose fermentations were conducted to simulate biomass hydrolysate containing mixed sugars in non-isomerized and fully isomerized forms. The mixed ketoses showed higher yield and faster fermentation than mixed aldoses, similar trends as observed with the single sugar fermentations. Ketose sugars are much more readily fermented, and an aldose-heavy mixture of the sugars would reduce the specific productivity of the fermentation. Hence, for 2,3-BD fermentation, high purity ketose sugars after SIRE-BE are desired, which can be achieved using the strategies in Chapter 3. In hindsight, since C6 sugars are fermented faster than the C5 counterparts, the results for a mixed glucose/xylulose fermentation would also be of interest. The additional 2,3-BD yield would 82
have to be evaluated in terms of the additional SIRE-BE processing required to get high yields of both xylulose and fructose.
83
Chapter 6
Future Directions
In this dissertation, strategies to optimize the SIRE-BE process the maximize
xylulose and fructose yield were developed, and the ketose-rich solutions were fermented
to several industrially relevant products. Our SIRE-BE method has potential to be scaled
up to meet industrial operation needs, and it could also be used to extract and purify
chemicals (polyols) other than sugars.
6.1 Validation of scaled SIRE system
A system to scale-up the SIRE process has been designed and built and is shown in
Figure 6-1. In the system, the flow rates of the two phases (organic and aqueous) for SIRE
are separately set using two gear pumps. Within the Liqui-Cel membrane contactor, the
sugars are transferred from the aqueous phase to the organic phase through the membrane
pores. The pH of the aqueous phase is controlled with a pH titrator; the pH-probe is
immersed in the aqueous phase reservoir. The larger-scale flow system shown in Figure 6-
1 will be used to gather data need for scale-up of the SIRE process. This system will also
be used to determine reusability of the organic phase.
84
(A)
(B)
Figure 6-1 The (A) schematic and (B) photo of the one-liter SIRE system. Blue loop with the associated numbers represents the path of the aqueous phase. Green loop with the associated numbers represents the path of the organic phase.
85
6.2 Purification of glycerol from industrial waste using SIRE-BE
Crude glycerol is a valuable byproduct of biodiesel production. Glycerol is a useful building block for many chemical synthesis reactions and has many applications [62].
Glycerol is used in the production of cosmetics, soaps, pharmaceuticals, polyglycerol esters and food [63]. Current glycerol purification technology includes sequential steps of neutralization, stripping, filtration/centrifugation and vacuum distillation [64]. However, these purification techniques have high energy requirements, and are not economically feasible for industrial scale [62].
In preliminary experiments, we have conducted a modified version of SIRE without isomerization (RE-BE) and have shown that N2B can bind to glycerol and extract it from an aqueous phase. Therefore, our RE-BE process should be explored further for assessing its potential to produce high purity and yield of glycerol from industrial biodiesel waste.
Optimizing RE-BE for glycerol purification will be essential to make it economically and operationally feasible. Identifying other polyols that can benefit from the use of either
SIRE-BE or RE-BE as an extraction and purification method will help to reduce energy costs and provide alternate, economically beneficial opportunities for industrial implementation.
86
References
1. Ho DP, Ngo HH, Guo WS, A mini review on renewable sources for biofuel.
Bioresource Technology, 2014. 169: 742-749.
2. U.S.A. Energy Information Administration: Today in energy, Crude oil prices increased in 2016, still below 2015 average.
3. Zabed, H., Sahu, J.N., Boyce, A.N., and Faruq, G., Fuel ethanol production from lignocellulosic biomass: An overview on feedstocks and technological approaches.
Renewable and Sustainable Energy Reviews, 2016. 66: 751-774.
4. Saini, J.K., Sani, R., and Tewari, L., Lignocellulosic agriculture wastes as biomass feedstocks for second-generation bioethanol production: concepts and recent developments. 3 Biotech, 2015. 5(4): 337-353.
5. Gray, K.A., L. Zhao, and M. Emptage, Bioethanol. Curr Opin Chem Biol, 2006.
10(2): p. 141-6.
6. Stefanidis SD, Kalogiannis KG, et al., A study of lignocellulosic biomass pyrolysis via the pyrolysis of cellulose, hemicellulose and lignin. Journal of Analytical and Applied
Pyrolysis, 2014. 105: 143-150.
7. BioTeK, Enzymatic Digestion of Polysaccharides (Part II).
87
8. Lynd, L.R., Overview and evaluation of fuel ethanol from cellulosic biomass:
Technology, Economics, the Environment, and Policy. Annual Review of Energy & the
Environment, 1996. 21(1): p. 403.
9. van Maris, A.J., et al., Alcoholic fermentation of carbon sources in biomass hydrolysates by Saccharomyces cerevisiae: current status. Antonie Van Leeuwenhoek,
2006. 90(4): p. 391-418.
10. Mood, S.H., et al., Lignocellulosic biomass to bioethanol, a comprehensive review with a focus on pretreatment. Renewable and Sustainable Energy Reviews, 2013. 27: 77-
93.
11. Sindhu, R., Binod, P., and Pandey, A., Biological pretreatment of lignocellulosic biomass-An overview. Bioresource Technology, 2016. 199: 76-82.
12. Yuan, D., et al., Fermentation of biomass sugars to ethanol using native industrial yeast strains. Bioresource Technology, 2011. 102: p. 3246-3253.
13. Yuan, D., et al., A viable method and configuration for fermenting biomass sugars to ethanol using native Saccharomyces cerevisiae. Bioresource Technology, 2012. 117: p.
92-988.
14. Rao, K., et al., A Novel Technique that Enables Efficient Conduct of Simultaneous
Isomerization and Fermentation (SIF) of Xylose. Bioresource Technology, 2008. 146: p.
101-117.
15. Li, B., P. Relue, and S. Varanasi, Simultaneous isomerization and reactive extraction of biomass sugars for high yield production of ketose sugars. Green Chemistry,
2012. 14(9): p. 2436-2444.
88
16. Zhang, W. and A. Geng, Improved ethanol production by a xylose-fermenting recombinant yeast strain constructed through a modified genome shuffling method.
Biotechnology for Biofuels, 2012. 5: p. 46-56.
17. Grzenia, D.L., S.R. Wickramasinghe, and D.J. Schell, Fermentation of Reactive-
Membrane-Extracted and Ammonium-Hydroxide-Conditioned Dilute-Acid-Pretreated
Corn Stover. Applied Biochemistry and Biotechnology, 2012. 166(2): p. 470-478.
18. Prasad, S., A. Singh, and H.C. Joshi, Ethanol as an alternative fuel from agricultural, industrial and urban residues. Resources, Conservation and Recycling, 2007.
50(1): p. 1-39.
19. Lachke, A., Biofuel from D-xylose — The second most abundant sugar. Resonance,
2002. 7(5): p. 50-58.
20. Jefferies, T.W. and Y.-s. Jin, Ethanol and Thermotolerance in the Bioconversion of
Xylose by Yeasts. Advances in applied microbiology, 2000. 47: p. 221-268.
21. Ito, T., et al., Intensification of Bio-ethanol Fermentation by Recombinant Yeast with Xylose Isomerase Pathway. Chemical Engineering Transaction, 2010. 20: p. 103-108.
22. Tantirungkij, M., et al., Construction of xylose-assimilating Saccharomyces cerevisiae. Journal of Fermentation and Bioengineering, 1993. 75(2): p. 83-88.
23. Wang, P.Y., C. Shopsis, and H. Schneider, Fermentation of a pentose by yeasts.
Biochemical and Biophysical Research Communications, 1980. 94(1): p. 248-254.
24. de Albuquerque, T. L., et al., Biotechnological production of xylitol from lignocellulosic wastes: A review. Process Biochemistry, 2014. 49(11): p. 1779-1789.
89
25. Lee, S.-H., et al., Effects of NADH-preferring xylose reductase expression on ethanol production from xylose in xylose-metabolizing recombinant Saccharomyces cerevisiae. Journal of Biotechnology, 2012. 158(4): p. 184-191.
26. Watanabe, S., et al., Ethanol production from xylose by recombinant
Saccharomyces cerevisiae expressing protein-engineered NADH-preferring xylose reductase from Pichia stipitis. Microbiology, 2007. 153(9): p. 3044-3054.
27. Bruinenberg, P.M., et al., NADH-linked aldose reductase: the key to anaerobic alcoholic fermentation of xylose by yeasts. Applied Microbiology and Biotechnology, 1984.
19: p. 256-260.
28. Toivari, M.H., et al., Conversion of Xylose to Ethanol by Recombinant
Saccharomyces cerevisiae: Importance of Xylulokinase (XKS1) and Oxygen Availability.
Metabolic Engineering, 2001. 3(3): p. 236-249.
29. Jeffries, T.W. and Y.S. Jin, Metabolic engineering for improved fermentation of pentoses by yeasts. Applied Microbiology and Biotechnology, 2004. 63(5): p. 495-509.
30. Chiang, L.C., et al., D-Xylulose Fermentation to Ethanol by Saccharomyces cerevisiae. Applied and Environmental Microbiology, 1981. 42(2): p. 284-289.
31. Roman, G.N., N.B. Jansen, and G.T. Tsao, Ethanol inhibition of D-xylulose fermentation by Schizosaccharomyces pombe. Biotechnology Letters, 1984. 6(1): p. 7-12.
32. Strategic Analysis of the Asia-Pacific Biorenewable. Frost & Sullivan, Octorber,
2012.
33. 1,4 BDO Industry Report 2014-2025. Market Research Report, February, 2017.
34. CEO 360 Green Fuels and Chemicals. Frost & Sullivan, June, 2009.
90
35. Xie, NZ., et al., Microbial Routes to 2,3-Butanediol: Recent Advances and Future
Prospects. Curr Top Med Chem, 2017. 17(21): 2433-2439.
36. Koutinas, A.A., et al., Techno-economic evaluation of a complete bioprocess for
2,3-butanediol production from renewable resources. Bioresource Technology, 2016. 204:
55-64.
37. Saha BC, Bothast RJ, Production of 2,3-butanediol by a newly isolated
Enterobacter cloacae. Appl Microbiol Biotechnol, 1999. 52: 321-326.
38. Saha BC, Hemicellulose Bioconversion. J Ind Microbiol Biotechnol, 2003. 30: 279-
291.
39. Asboth, B. and N.S.G., Mechanism of action of D-xylose isomerase. Current Protein
and Peptide Science, 2000. 1(3): 237-254.
40. Bhosale, S.H., Rao, M.B., and Deshpande, V.V., Molecular and industrial aspects
of glucose isomerase. Microbiological reviews, 1996. 60(2): 280-300.
41. Hosseini, S., et al., Hydrogen production from renewable and sustainable energy resources: Promising green energy carrier for clean development. Renewable and
Sustainable Energy Reviews, 2016. 57: 850-866.
42. Liao, J.C, et al., Fuelling the future: microbial engineering for the production of
sustainable biofuels. Nature Reviews Microbiology, 2016. 14: 288-304.
43. Li, B., S. Varanasi, and P. Relue, High yield aldose-ketose transformation for
isolation and facile conversion of biomass sugar to furan. Green Chemistry, 2013. 15(8):
p. 2149-2157.
91
44. Siamak, A., P. Relue, S. Viamajala and S. Varanasi, High yield 5-(hydroxymethyl) furfural production from biomass sugars under facile reaction conditions: a hybrid enzyme-and chemo-catalytic technology. Green Chemistry, 2017. 19(7): p. 1782.
45. H. Xia, S. Xu, L. Yang, Efficient conversion of wheat straw into furan compounds, bio-oils, and phosphate fertilizers by a combination of hydrolysis and catalytic pyrolysis.
RSC Advances, 2017. 7: p. 1200-1205.
46. Kristensen, B.J., et al., Cell-wall structural changes in wheat straw pretreated for bioethanol production. Biotechnology for Biofuels, 2008. 1 (5).
47. Gnansounou, Edgard. and A. Dauriat, Techno-economic analysis of lignocellulosic ethanol: A review. Bioresource Technology, 2010. 101(13): p. 4980-4991.
48. Aloysius J.J.E. Eerhart, et al., Fuels and plastics from lignocellulosic biomass via the furan pathway: an economic analysis. Biofuels, Bioproducts and Biorefining, 2015.
9(3): p. 307-325.
49. B. Hahn-Hägerdal, K. Skoog, S. Berner, Improved ethanol production from xylose with glucose isomerase and Saccharomyces cerevisiae using the respiratory inhibitor azide.
Appl. Microbiol. Biotechnol., 1986. 24(4): p. 287-293.
50. P. Owusu, S. Asumadu-Sarkodie, A review of renewable energy sources, sustainability issues and climate change mitigation. Cogent Engineering, 2016. 3(1).
51. T. Foust, and A. Bratis, Review of Recent Pilot Scale Cellulosic Ethanol
Demonstration. U.S. Department of Energy, July, 2013.
52. Chandel AK, et al., Bioconversion of de-oiled rice bran (DORB) hemicellulosic hydrolysate into ethanol by Pichia stipites NCIM3499 under optimized conditions. Int. J.
Food Eng., 2009. 2(4): p. 1-12. 92
53. Chandel AK, et al., Bioconversion of novel substrate Saccharum spontaneum, a weedy material, into ethanol by Pichia stipites NCIM3498. Bioresource Technology, 2011.
102(2): p. 1709-1714.
54. Vilela, et al., Enhanced xylose fermentation and ethanol production by engineered
Saccharomyces cerevisiae strain. AMB Express, 2015. 5(16).
55. A. Singh, et al., Comparative study on ethanol production from pretreated sugarcane bagasse using immobilized Saccharomyces cerevisiae on various matrices.
Renewable Energy, 2013. 50: 488-493.
56. M. Ishola and M. Taherzadeh, Effect of fungal and phosphoric acid pretreatment on ethanol production from oil palm empty fruit bunches (OPEFB). Bioresource
Thchnology, 2014. 165: 9-12.
57. P. Karagoz and M. Ozkan, Ethanol production from wheat straw by Saccharomyces cerevisiae and Scheffersomyces stipitis co-culture in batch and continuous system.
Bioresource Thchnology, 2014. 158: 286-293.
58. M. V. P. Rocha, et al., Evaluation of dilute acid pretreatment on cashew apple bagasse for ethanol and xylitol production. Chemical Engineering Journal, 2014. 243: 234-
243.
59. L. Laopaiboon and P. Laopaiboon, Ethanol production from sweet sorghum juice in repeated-batch fermentation by Saccharomyces cerevisiae immobilized on corncob.
World Journal of Microbiology and Biotechnology, 2012. 28 (2): 559-566.
60. Global and China BDO Industry Report, 2017-2021. Research In China, August,
2017.
93
61. Slininger PJ, Bothast RJ, Van Cauwenberge JE, Kurtzman CP, Conversion of D- xylose to ethanol by the yeast Pachysolen tannophilus. Biotechnol Bioeng, 1982. 23: 371-
384.
62. Ardi MS, Aroua MK, Hashim NA, Process, prospect and challenges in glycerol purification process: A review. Renewable and Sustainable Energy Reviews, 2015. 42:
1164-1173.
63. Bondioli P, Overview from oil seeds to industrial products: present and future oleochemistry. J Synth Lubr, 2005. 21: 331-343.
64. Sdrula N, A study using classical or membrane separation in the biodiesel process.
Desalination, 2010. 250 (3): 1070-1072.
94
Appendix A
Media and Solution Formulas
Chapter 2-5:
Preparation of organic phase for SIRE: The molar ratio of Aliquat® 336 to N2B was ranged
from 0.5 to 2.5. The molar ratio of N2B to sugar directly impacts the sugar extraction
selectivity (xylulose/xylose and fructose/glucose). Since Aliquat® 336 is very difficult to
handle due to its viscosity, tare the shake flash that will be used and weigh out Aliquat®
336 in the shake flask; based on its density (0.884 g/mL at 25 °C), the volume of it should
be calculated. N2B can be weighed out in a weigh dish and added to the shake flask.
Octanol is added to a final volume of desired value. The solution is stored tightly capped
at 4°C until needed for SIRE.
Chapter 5:
Media used for culturing E. cloacae NRRL B-23289 [61]: 10 mL of solution A, 10 mL of
solution B, 100 mL of solution C, 10 g of yeast extract, and 50 g of sugar (glucose, fructose,
xylose, xylulose)/L. Solution A: 1.10 g of CaO, 0.40 g of ZnO, 5.40 g of FeCl3-6H2O, 0.36
g of MgO, 0.25 g of CuSO4-5H2O, 0.24 g of CoCl2-6H2O, 0.06 g of H3BO3, and 13.0 mL
of concentrated HCl/L. Solution B: 10.1 g of MgO and 45.0 mL of concentrated HCl/L.
Solution C: 64.0 g of urea, 12 g of KH2PO4, and 1.8 g of Na2HPO4/L. 95
Appendix B
Analytical Information on HPLC
Range of concentrations for standards of most of the samples was 0.25-5.50 g/L.
Two different columns were used in this dissertation:
1) Shodex SH1011 columns (300×8 mm) (two in series). A mobile phase of 0.05 M H2SO4
was run at 0.6 ml/min; a column temperature of 50 °C was used. Elution peak times of
major compounds are shown in Table B.1:
Table B.1 Elution peak times on two Shodex SH1011 columns.
Glucose Fructose Xylose Xylulose Ethanol 2,3-BD
Time (min) 23 24 24 25 44 37
2) Shodex SP0810 columns (300×8 mm) (two in series). A mobile phase of ultra-pure water
was run at 0.6 ml/min; a column temperature of 80 °C was used.
Table B.2 Elution peak times on two Shodex SP0810 columns.
Glucose Fructose Xylose Xylulose
Time (min) 30 41 32 38
96