© 2012 Srujana Koganti All Rights Reserved

SRUJANA KOGANTI

CONVERSION OF BIODIESEL BYPRODUCT GLYCEROL TO ARABITOL AND

SOPHOROLIPIDS THROUGH MICROBIAL FERMENTATION

A Dissertation

Presented to

The Graduate Faculty of The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Srujana Koganti

May, 2012

CONVERSION OF BIODIESEL BYPRODUCT GLYCEROL TO ARABITOL AND

SOPHOROLIPIDS THROUGH MICROBIAL FERMENTATION

Srujana Koganti

Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Lu-Kwang Ju Dr. Lu-Kwang Ju

______Committee Member Dean of the College Dr. George Chase Dr. George K. Haritos

______Committee Member Dean of the Graduate School Dr. Jie Zheng Dr. George R. Newkome

______Committee Member Date Dr. Narender P Reddy

______Committee Member Dr. John M Senko

ABSTRACT

Biodiesel is an attractive alternative fuel produced by renewable resources and glycerol is a major byproduct in biodiesel production. It is important to find commercial/industrial applications of glycerol for the sustainability and economics of biodiesel industry. We found the use of glycerol in producing the sugar alcohol, arabitol, using Debaryomyces hansenii and the group of biosurfactants, sophorolipids, using

Candida bombicola through microbial fermentation. Arabitol, a stereoisomer to xylitol, has the potential applications as a sweetener for diabetic patients and reducer of dental caries. About 217 strains were screened by our project partner, the late Dr. Tsung Min

Kuo, at the United States Department of Agriculture for their ability to produce arabitol from glycerol as carbon source. A D. hansenii strain was chosen from the screening process due to its better arabitol production compared to other strains and for being able to produce arabitol as the only polyol. We successfully achieved the optimal conditions that are required for maximum arabitol production from glycerol as a substrate using D. hansenii. The optimal conditions thus found were 5% dissolved oxygen, 3.5 pH, temperature of 30 °C and nitrogen-to-phosphorous ration of 9 to achieve a yield of 55% and a productivity of 0.2 g/L-h.

It was found in our study that high glycerol concentrations (at least 100 g/L) are favorable in improving the arabitol yield from 14% to 40% with 50 g/L and 100 g/L initial glycerol concentration respectively. However, higher glycerol concentrations needed longer fermentation run time for complete consumption of glycerol, particularly

iii when the rates of consumption of glycerol and production of arabitol slowed and even stopped as the yeast cells entered into an extended stationary phase. This problem was shown to be solvable by the addition of organic nitrogen source, to reinvigorate the cells, and xylose, as a cosubstrate. This method could not only improve the yield but also the productivity of arabitol from 0.1 g/L-h to 0.5 g/L-h. Such successful method of producing arabitol using glycerol reduces the production costs One of the ways of reducing the production cost in industrial scale is to use low cost raw materials and in this study it was achieved by the use of biodiesel glycerol. Another potential option found in our study was the potential use of lignocellulose materials as the raw materials for the production of arabitol. The ability of D. hansenii in utilizing glucose, xylose and their representing the lignocellulose hydrolysate was studied and good arabitol production was achieved.

Lignocellulose hydrolysate majorly contains mixtures of glucose, xylose, and other sugars in lower amounts.

Another use of biodiesel glycerol studied in this research was for the production of sophorolipids. Biosurfactants have many applications in food, cosmetic and drug industries. Sophorolipids are usually produced from a combination of hydrophobic carbon source and lipid precursor. Biodiesel waste product was used for sophorolipids production by others but the specific yield of sophorolipids achieved in our research was higher (2.5 g sophorolipids/ g of cells) which was achieved by the controlled addition of lipid precursor compared to others with 1.5 g sophorolipids/ g cells. Further, optimization studies could be done using biodiesel glycerol to improve the productivity and yield.

ACKNOWLEDGEMENTS

My deepest gratitude is to my advisor Dr. Lu-Kwang Ju for his invaluable guidance, constant inspiration, motivation and endless patience which made my research and dissertation possible.

I would like to thank my family for their unconditional love and constant support, especially my grandfather Mr. Tatineni Venkataramaiah for his motivation and encouragement throughout these years.

I am also grateful to Dr. George Chase, Dr. Jie Zheng, Dr. Narender Reddy, Dr.

Weiping Zhengh and Dr. John Senko for serving in my committee. I would like to thank all the members of Dr. Ju’s research group for their help and suggestions during the course of this work. In addition I would like to thank my friends Radhika Gummadi,

Narayanan, Rajesh, Keerthi, Nana, Radhika, Bindu and Uma for their support throughout these years.

I would like to acknowledge the financial support from United Soybean Board and the late Dr. Tsung M. Kua from USDA for their collaboration with this project and also chemical and biomoleculer engineering department for the financial support and Dr.

Don Ott for letting me use his optical microscope.

TABLE OF CONTENTS Page

LIST OF TABLES………………………………………………………………………xii

LIST OF FIGURES………………………………………………………………...... xiii

CHAPTER

I. INTRODUCTION ...... 1

1.1 Background and significance ...... 1

1.2 Scope of research ...... 5

1.3 Research objectives ...... 7

1.4 Structure of dissertation ...... 8

II. LITERATURE SURVEY ...... 10

2.1 Biodiesel byproduct glycerol ...... 10

2.2 Polyols...... 12

2.3 Arabitol ...... 14

2.3.1 Properties and applications………………………………………………………15

2.3.2 Production of arabitol and its biosynthetic pathway……………………………..18

2.4 Biosurfactants ...... 22

2.5 Sophorolipids ...... 23

2.5.1 Properties and applications………………………………………………………24

2.5.2 Production………………………………………………………………………..27

III. MATERIALS AND METHODS ...... 32

3.1 Microorganisms and medium ...... 32

3.2 Experimental setup and equipment ...... 33

3.2.1 Arabitol production ...... 33

3.2.2 Sophorolipids production ...... 35

3.3 Analytical methods ...... 36

3.3.1 Glucose analysis...... 36

3.3.2 Sugar and polyols analysis ...... 36

3.3.3 Cell concentration measurement ...... 37

3.3.4 Ammonia nitrogen analysis ...... 39

3.3.5 Sophorolipids analysis ...... 39

IV. PRODUCTION OF ARABITOL FROM GYCEROL: STRAIN SCREENING AND STUDY FACTORS AFFECTING PRODUCTION YIELD ...... 40

4.1 Introduction ...... 40

4.2 Materials and Methods ...... 42

4.2.1 Yeast strain screening ...... 42

4.2.2 Media ...... 42

4.2.3 Culture methods ...... 43

4.2.4 Analytical methods ...... 45

4.3 Results and discussion ...... 48

4.3.1 Screening for arabitol production from glycerol ...... 48

4.3.2 Effect of culture volume in shake flasks ...... 51

4.3.3 Effect of temperature ...... 53

vii

4.3.4 Effect of initial glycerol concentration ...... 54

4.3.5 Effect of salt concentration ...... 55

4.3.6 Effects of addition of other carbon substrates ...... 59

4.4 Conclusions ...... 64

V. ARABITOL PRODUCTION FROM GLYCEROL USING DEBARYOMYCES HANSENII: EFFECTS OF MEDIUM N/P RATIO, pH AND DISSOLVED OXYGEN CONCENTRATION ...... 66

5.1. Introduction ...... 66

5.2. Materials and Methods ...... 68

5.2.1 Microorganism and medium ...... 68

5.2.2 Inoculum preparation ...... 69

5.2.3 Fermentation conditions...... 70

5.2.4 Effect of medium N-to-P ratio ...... 70

5.2.5 Effects of pH and DO ...... 71

5.2.6 Analytical methods ...... 71

5.3. Results and Discussion ...... 72

5.3.1 Effects of medium nitrogen-to-phosphorous ratio ...... 72

5.3.2 Effect of DO ...... 75

5.3.3 Effect of pH...... 78

5.4. Conclusion ...... 84

VI. MODIFICATION OF MEDIUM COMPOSITION IN IMPROVING SPECIFIC PRODUCTIVITY AND VOLUMETRIC PRODUCTIVITY OF ARABITOL ...... 85

6.1. Introduction ...... 85

6.2. Materials and Methods ...... 87

viii

6.2.1 Microorganism and medium ...... 87

6.2.2 Inoculum preparation ...... 88

6.2.3 Experimental studies ...... 88

6.2.4 Analytical methods ...... 93

6.3. Results and Discussion ...... 93

6.3.1 Effect of different N source combinations ...... 94

6.3.2 Effect of magnesium concentration ...... 97

6.3.3 Effect of phosphorous concentration ...... 98

6.3.4 Arabitol production in fermentor with organic N source and acid addition to lower pH at late growth phase ...... 99

6.3.5 pH and arabitol production affected by addition of ammonium sulfate at different ratios of NH3-N to organic N ...... 101

6.4. Conclusion ...... 106

VII. METHODS OF IMPROVING GLYCERROL CONSUMPTION AND ARABITOL PRODUCTION IN FERMENTATION WITH DEBARYOMYCES HANSENII .... 108

7.1. Introduction ...... 108

7.2. Materials and Methods ...... 110

7.2.1 Microorganism and media ...... 111

7.2.2 Culture conditions and procedures...... 112

(a) Slow addition of glucose ...... 113

(b) Improving glycerol uptake and arabitol production ...... 113

7.2.3 Analytical methods ...... 116

Cell concentration ...... 116

Glycerol, glucose, xylose and arabitol analyses ...... 116

7.3. Results and Discussion ...... 117

7.3.1 Glucose and xylose as carbon source...... 117

7.3.2 Comparison of glucose and glycerol for arabitol production in pH and DO controlled fermentors ...... 120

7.3.3 Slow glucose addition for arabitol production ...... 122

7.3.4 Arabitol production by glucose addition to stationary-phase glycerol-grown culture… ...... 126

7.3.5 Effects of pH adjustment, N-source addition and xylose addition on improving glycerol depletion and arabitol production ...... 128

7.4. Conclusion ...... 135

VIII. PRODUCTION OF SOPHOROLIPIDS USING BIODIESEL BYPRODUCT GLYCEROL AND VEGETABLE OIL AS CARBON SOURCE ...... 136

8.1. Introduction ...... 136

8.2. Materials and methods ...... 139

8.2.1 Culture maintenance and inoculum preparation ...... 139

8.2.2 Fermentation conditions...... 139

8.2.3 Analytical methods ...... 142

8.3 Results and discussion ...... 143

8.4 Conclusions ...... 149

IX. CONCLUSION...... 151

9.1 Summary and conclusions ...... 151

9.2 Future recommendations ...... 156

REFRENCES ...... 158

APPENDIX ...... 179

LIST OF TABLES

Table Page

2.1 MIC of different sophorolipids resulting in reducing growth up to 50% of normal growth…………………………………………………………………...13

2.2 Comparison of properties of biosurfactants and synthetic surfactants………………………………………………………………………………..27

2.3 Summary of properties of some of the polyols………………………………………27

3.1 Genera and number of strains screened for arabitol production from glycerol as substrate……………………………………………………………………...33

4.1 Genera and number of strains screened……………………………………………...49

4.2 Percentages of different polyols produced by some osmotolerant yeast strains………………………………………………………………………………51

5.1 Compositions of media used for N/P ratio study…………………………………….69

6.1 Media used in shake flask study for effects of different N source combinations……………………………………………………………………………..91

6.2 Sources of N and P for media used in the study of effects of ammonium sulfate addition on pH change and arabitol production……………………..92

7.1 Nine systems compared for improving glycerol consumption and arabitol production of a stationary-phase culture pre-grown in the glycerol-based medium – Systems, with or without pH adjustment to 5, were supplemented with xylose (Xyl), yeast extract (YE), ammonium sulfate (AS), or a certain combination of them……………………………………………………………………116

8.1 Comparison of volumetric productivity and total sophorolipids produced by various cost effective methods using different low cost raw materials about 70 g/L in 5 days………………………………………………………..151

A.1 Substrate concentrations that were considered for determining the volumetric oxygen transfer rate and oxygen uptake rate……………………………….184

A.2 kLa and OUR values for Run 1 and Run 2 with No cells and in the presence of different concentrations of glycerol and xylitol…………………………...186

xii

LIST OF FIGURES Figure Page

2.1 Biodiesel production in millions of gallons over the years in United States of America…………………………………………………………………11

2.2 Market price of refined and crude glycerol produced from biodiesel industry over the years…………………………………………………………….11

2.3 Molecular structure of Sophorolipids……………………………………….…….14

2.4 Possible biosynthetic pathways for sophorolipids production using carbohydrate as substrate…………………………………………………………18

2.5 Possible biosynthetic pathways for the production of sophorolipids using triglycerides as substrate……………………………………………...…….21

2.6 Structures of xylitol and arabitol……………………………………….………….24

2.7 Xylitol polymerization steps with (A) citric acid or (B) sebacic acid to form hydrogel and elastomer……………………………………………...28

2.8 Possible pathways for the production of arabitol with substrates such as glucose, glycerol, sorbitol and xylose………………………...…………..29

3.1 Experimental set-up for batch fermentation for arabitoL production………...... 34 4.1 Concentrations of arabitol produced, glycerol consumed and cells grown in the systems with different initial culture volumes. Arabitol and glycerol concentrations were measured at 120 h; cell concentrations were measured at 80 h………………………..…………………………………………52

4.2 Arabitol produced by selected strains of Debaryomyces, Geotrichum and Metschnikowia at different temperatures. Samples were taken after 3 days of cultivation……………..………………………………………………………..53

4.3 Effects of different initial glycerol concentrations on D. hansenii SBP-1 fermentation: (a) cell growth profiles, (b) arabitol production profiles, (c) concentrations of glycerol consumed and arabitol produced at 120 h, and (d) arabitol yield (from consumed glycerol) at 120 h…………………………..…55

xiii

4.4 Effects of different salt concentrations on D. hansenii SBP-1 fermentation with 100 g/L of initial glycerol concentration: Cell growth and arabitol production profiles…………………………………………………..56

4.5 Effects of salt addition on arabitol production by D. hansenii SBP-1 in media with 100 g/L of initial glycerol concentration. The salt was added after 2 days of growth………………………………………………....58

4.6 Effects of addition of 30 g/L glucose, xylose or sorbitol as a potential second carbon source, along with 50 g/L glycerol, on arabitol production by stationary-phase D. hansenii SBP-1. Cells were grown for 74 h in media containing 30 g/L glycerol before the second C-source and glycerol were added. The control system was added with 80 g/L glycerol Arabitol production profiles are compared in (a). Concentrations of the second C-source and glycerol consumed by the stationary-phase cultures (during 74-145 h) are shown in (b)………………………………………………..61

4.7 Possible pathways for the conversion of various substrates to arabitol…………………………………………………………………...………..65

5.1 Profiles of cell growth and arabitol production in systems with different N-to-P ratios in the media………………………...……………………..73

5.2 pH and DO profiles for the System 5 fermentation………………………….……75

5.3 Time profiles of (a) cell growth and (b) glycerol consumption and arabitol production, and (c) the overall and stationary-phase arabitol yields observed in systems with 5, 10 and 20% DO control values……………………………………………………………………….……..76

5.4 (a) Time profiles of arabitol concentration (g/L) produced at different pH (3, 3.5 and 4) in systems with frequent glycerol addition to maintain about 100 g/L glycerol, and (b) their corresponding plots of amounts (g) of arabitol produced versus amounts (g) of glycerol consumed……………………………….………………………………………....79

5.5 Profiles of (a) cell growth and (b) arabitol production in fermentations with same (5%) control DO value but different pH control values, i.e., 3, 3.5, 4, 5 and 6…...…………..……………………………………………………….....80

6.1 Profiles of (a) cell growth and (b) pH for systems with different N sources………..…………………………………………………………………95

6.2 Cell and arabitol concentrations of samples taken at 50 h from systems with different N sources, plotted against the total P concentration in the initial medium……………..…………………………………………………..97

xiv

6.3 Cell growth represented as optical density measured at 610 nm for systems with 1, 1.5 and 2 g/L MgSO4∙7H2O in both the (YE + AS) and (YE + Pep) medium systems……………………..……………………………….99

6.4 Cell growth profiles in media with varying phosphorous concentrations……………………………………………………………………100

6.5 (a) pH (b) Cell, glycerol and arabitol concentration profiles for fermentation run with YE and Pep as nitrogen source……………..……………102

6.6 pH change profile for systems with different ratios of organic N to NH3-N…………………………………………………………………………….103

6.7 Profiles of (a) cell growth and (b) arabitol production for systems with different ratios of organic N to NH3-N……………………….……………105

6.8 (a) Doubling time of cell growth which was calculated from 0-24 hrs, time period when the cell growth was linear and (b) arabitol yield (measured at 75 h) for systems with different ratios of organic N to NH3-N……………………………………………………………………………106

7.1 (a) pH profiles comparison (b) Cell growth profiles comparison followed by measurements of optical density at 610 nm in glucose, glycerol, xylose and Glucose+Xylose systems…………………………………………....118

7.2 (a) Arabitol production comparison in glucose, glycerol, xylose and Glucose+Xylose systems (b) Glucose, glycerol and xylose consumption and (c) volumetric productivity for systems containing different carbon sources such as glucose, xylose and glycerol………………..…………………………...120

7.3 (a) Cell growth comparison (b) Comparison of volumetric productivity in glucose and glycerol fermentors and glycerol production in glucose fermentor……….....………………………………………………….….122

7.4 (a) substrate consumption in glucose and glycerol fermenters and (b) Arabitol production profiles in glucose and glycerol fermentors and glycerol production in glucose fermentor………...……………...………………123

7.5 (a) Glucose concentration in g/L and (b)Glucose consumed and arabitol produced in gms in systems with 0.5,1 and 1.5 g/L.hr continuous glucose addition rates done 250 ml erlenmeyer flasks…………….....…………125

7.6 Glycerol, glucose consumption and arabitol production profile in the fermentation studies done to see the effect of addition of 30 g/L glucose to improve the glycerol uptake rate………………………...……………………….128

7.7 (a) Cell growth profile for systems with different xylose and nitrogen source concentration in comparison with control with no external xylose or nitrogen source additions after 160 h of fermentation broth……………………………………………………………...………………130

7.8 (a) pH profile and (b) Arabitol production profile for systems with different xylose and nitrogen source concentration in comparison with control with no external xylose or nitrogen source additions after 160 h of fermentation broth………………...………………………………………...……132

7.9 (a) Concentration of total glycerol & xylose consumed and arabitol produced (b) Total carbon yield from total carbon consumed contributed from glycerol and xylose present in the medium and (c) Concentration of total carbon consumed and produced in all the systems with different xylose and nitrogen source in the medium measured from 160-225 h………………………..………………………………134

8.1 Rate of vegetable oil addition in ml/h for fermentor with glucose as primary carbon source…………….…………………………………142

8.2 Rate of vegetable oil addition in ml/h for fermentor with glycerol as primary carbon source…………………………………………………………..143

8.3 Cell growth and ammonia nitrogen consumption profile in glucose and glycerol as hydrophilic substrates and vegetable oil as hydrophobic source for sophorolipids production at different times of fermentation………………………………………………………………….….145

8.4 Comparison of sophorolipids production profile for fermentors with glucose and glycerol as hydrophilic carbon source and vegetable oil as hydrophobic carbon source………………….…………………………………..146

8.5 Comparison of volumetric productivity of sophorolipids in fermentors with glucose and glycerol as hydrophilic carbon source and vegetable oil as hydrophobic carbon source……………………………………148

8.6 Residual glucose and glycerol concentrations in g/L measured at different times of fermentation. The increase in glucose concentration indicates the batch addition……………………………………………………………….148

A.1 Typical dissolved oxygen profile using dynamic method for determining volumetric oxygen transfer rate…………………………………...183

A.2 kLa and OUR values with No cells and in the presence of different concentrations of glycerol and xylitol…………………………………………………………..187

xvi

CHAPTER I

INTRODUCTION

1.1 Background and significance

Biodiesel is an attractive alternative fuel to petroleum. Biodiesel can be produced from renewable resources like soybean oil, sunflower oil, coconut oil, rapeseed oil and palm oil. Over the recent years, importance has been given for biodiesel production because of the increased prices of crude oil and also due to the environmental concerns.

Thus, there is a great possibility for the biodiesel market to increase sharply [1].

Biodiesel produced from renewable sources like vegetable oil and animal fat is an attractive alternative fuel [2]. Several advantages of biodiesel over petroleum are that it is

(1) produced from renewable resources, (2) environmental friendly, and (3) biodegradable. In biodiesel production using transesterification of triglycerides, glycerol is the major inevitable byproduct produced: about 1 kg of glycerol is formed for every 9 kg of biodiesel produced [3]. Biodiesel consumption in the United States has increased dramatically from 75 million gallons in 2005 to 700 million gallons in 2008. The latter resulted in the production of around 45 million gallons of glycerol [4]. Refined glycerol has numerous applications in food, drug, textile and cosmetic industries whereas crude glycerol produced from biodiesel industry is of low value because of its impurities like spent catalyst, salts after neutralization, residual methanol, methyl esters and free fatty

acids [5, 6]. The economics of biodiesel industry is strongly influenced by the value of its byproducts. Developing new uses of biodiesel glycerol is imperative to economics and sustainability of the biodiesel industry [7, 8]. The main objective of this project is to utilize glycerol as the carbon source for the production of high-value biosurfactants and polyol.

Biosurfactants in recent years have got attention due to their advantages over their synthetic counterparts for their applications in various industries like petroleum, cosmetics, food processing and pharmaceuticals. Biosurfactants are biomolecules that are amphiphilic, have high surface activity, and partition at interface [20]. Biosurfactants have a wide variety of chemical structures such as glycolipids, lipopeptides, fatty acids, polysaccharide-protein complexes and neutral lipids [21]. During the cultivation on various carbon sources and hydrophobic substances such as hydrocarbons, yeast, bacteria and filamentous fungi are found to synthesize biosurfactants [22]. The advantages of biosurfactants over their synthetic counterparts include (1) high specificity (2) biocompatibility (3) low toxicity and (4) biodegradability [4]. Some biosurfactants exhibit good thermal and chemical stability [20]. Biosurfactants like sophorolipids and rhamnolipids have potential applications in remediation of contaminated water and soil

[4]. Since synthetic surfactants are having health and environmental impact, biosurfactants are more attractive as they can be produced from renewable resources like fats and oils. In addition, biosurfactants have a novel structural characteristics and physical properties which make them suitable for various applications such as phase separation, wetting, foaming, emulsification, de-emulsification, solubulization, corrosion inhibition and viscosity reduction [23].

Among the most important biosurfactants, sophorolipids can be produced by

Candida bombicola. Sophorolipids consists of a dimeric carbohydrate sophorose linked to a long chain of fatty acid [16]. Sophorolipids have good market value because cheaper production cost compared to other biosurfactants [8]. Glucose is mainly used as a carbon source and fatty acids are used as secondary carbon source for the production of sophorolipids [9]. In this study biodiesel glycerol is used as a primary carbon source and vegetable oil is used as the secondary carbon source. Important factors like concentration of fatty acids and dissolved oxygen content were studied for the effective production of

SL.

Sophorolipids are the cheapest biosurfactant and are highly industrially significant in commercial market [19]. The acetylated lactonic sophorolipids have applications in cosmetics as antidandruff, deodorant, bacteriostatic and moisturizers [24] [25].

Sophorolipids can also be used as an anti-fungal agent against plant and seed pathogenic fungi [26]. The production of sophorolipids from glycerol from biodiesel industry has been studied as a part of this project.

Arabitol is a polyhydric alcohol that can be used as a low calorie sweetener [9]. In addition, a study by the Department of Energy identified arabitol, and its enantiomer xylitol, as one of the top twelve biomass-derivable building block chemicals. Arabitol and xylitol can be transformed into several groups of chemicals like xylaric/xylonic acid, arabonic/arabinoic acid, propylene glycol and ethylene glycol [10]. Arabitol and xylitol have melting points of 103 °C and 93 °C, respectively. Both are highly soluble in water and both form white crystals when purified [11, 12]. The catabolism of arabitol by

Escherichia coli involves the formation of arabitol phosphate which induces the synthesis

3 of compounds that inhibit the bacterial metabolism [13]. While more studies are required, the above property makes it possible to use arabitol as sweetener for reducing dental caries. Also, the caloric value of arabitol is 0.2 kcal/g whereas it is 2.4 kcal/g for xylitol

[9, 14-16]. It is highly possible that arabitol can be used in many of the known applications of xylitol, as a natural sweetener, a dental caries reducer and a sugar substitute for diabetic patients [17]. If desirable, arabitol can also be converted to xylitol, for example, by using Glucanobacter oxydans [18]. This bacterium was capable of oxidizing D-arabitol to D-xylulose using the membrane-bound D-arabitol dehydrogenase and then converting D-xylulose to D-xylitol using the also membrane-bound D-xylitol dehydrogenase. Xylitol yield of around 25% has been reported [19].

Xylitol is currently produced by chemical reduction of xylose derived from wood hydrolysate under alkaline conditions [20]. This process requires high pressure (50 atm) and temperature (80-140 °C), use of expensive catalyst, and extensive separation steps.

Xylitol production from xylose by biological processes has also been explored [21-25].

Yeast can covert xylose to xylitol using NAD(P)H-coupled xylose reductase.

Unfortunately, the xylitol produced tends to be oxidized to xylulose by NAD+-coupled xylitol dehydrogenase. Good xylitol yields from such a process require tightly controlled, high intracellular NAD(P)H/NAD+ ratios. This control is not an easy task in large-scale industrial operations where the environment (particularly the dissolved oxygen concentrations) inside the large bioreactors is non-homogeneous. The above chemical and biological processes require expensive separation of xylose from the complex sugar mixtures in the biomass hydrolysate. The alternative approach of producing arabitol from

4 biodiesel glycerol and then, if desirable, converting arabitol to xylitol may prove economically attractive.

Arabitol is known to be produced by osmophilic yeast species such as

Debaryomyces [26], Candida [27], Pichia [28], Hansenula [29] and Endomycopsis [30].

When exposed to osmotic stress, the yeast accumulates compatible solutes such as arabitol, glycerol, xylitol, erythritol and mannitol to balance the osmotic pressure across the cell membrane. As mentioned, glycerol was used as carbon source for the production of arabitol. An objective of this study was to select the yeast strains that produce large amounts of arabitol with high yields and minimal other polyols (for easier downstream separation) using glycerol as substrate. The work was supported by grants from the

United Soybean Board (Projects 9435 and 1475). The productive D. hansenii strain was identified by the late Dr. Tsung M. Kuo (NCAUR, USDA, Peoria, IL) our collaborator of this project. Experiments were then conducted to characterize the most promising strains for growth and arabitol production under different culture conditions. Another objective was to improve the production and yield of arabitol by optimizing the operating conditions like temperature, pH and dissolved oxygen concentration and improving the medium composition.

1.2 Scope of research

This research focused on two main subjects. One was on the production of sophorolipids with Candida bombicola using biodiesel byproduct glycerol and other was the production of arabitol with Debaryomyces hansenii using biodiesel byproduct glycerol. Much of the research of this project was concentrated on the optimization of arabitol production. Arabitol is an expensive bioproduct since there the methods of its

5 production are not cost effective enough. Hence, we focused on the optimization of arabitol production and the details of this are discussed as follows.

One of the subjects was relating to the production of arabitol. The first step was to identify a strain that could utilize glycerol and produce arabitol. The strain screening process for identifying such strain was done at United States Department of Agriculture were 214 strains belonging to different genera such as Debaryomyces, Geotrichum,

Candida, Metschnikowia, Pichia and others which are reported in the later chapters. This part of the project was supported by USDA and the strain was identified by late Dr.

Tsung M. Kuo (NCAUR, USDA, Peoria, IL). After identifying the strain, factors that are critical in affecting the arabitol production such as operating conditions, substrate concentration and medium composition were studied.

Arabitol production initiated during exponential growth phase and continued in stationary phase of the yeast cells. However, arabitol production rate was seen to reduce in stationary phase. This problem was addressed by the addition of nitrogen source and xylose to the fermentation broth and other methods to improve the production rate during the stationary phase.

In general, for polyol like arabitol production by other species glucose is the most widely used carbon source. When glycerol is used as carbon source, it should go into gluconeogenisis pathway for entering into pentose phosphate pathway for arabitol production. When glucose was used as substrate arabitol production was faster than that with glycerol. An effort was done to improve this process of conversion of arabitol to glycerol. With the aim of using cheaper substrate for arabitol production plant bagasse was used as substrate. Major carbon source in bagasse after pretreatment is glucose

6 followed by xylose. These carbon sources and their most prominent ratio was used to mimic bagasse medium for arabitol production was also studied.

Initial part of this study deals with the usage of glycerol as a carbon source for the production of sophorolipids.

Another part of this study dealt with the usage of glycerol as a carbon source for the production of sophorolipids. Glucose is the most used carbon source for the production of sophorolipids. In this study, glycerol was compared with glucose as a carbon source for the production of sophorolipids using Candida bombicola also comparing the sophorolipids production from other processes using biodiesel waste stream.

1.3 Research objectives

One of the objectives of this project was to identify strain that uses biodiesel glycerol and produce arabitol. Arabitol is stereoisomer to xylitol and has many properties similar to xylitol. However, not much research has been done with arabitol compared to xylitol.

Also, not much research was done on the production of arabitol and the market value of arabitol is higher than xylitol. Hence, there is a need to find an effective and economical method for the production of arabitol for further research on its applications as a polyol.

Another objective of this project was to study the production of sophorolipids using biodiesel glycerol. There were no reports on the production of sophorolipids using biodiesel glycerol in comparison to other substrates such as glucose.

Accordingly, the following are the specific objectives that were addressed in this project:

1. Select a strain for its ability to produce arabitol as the only polyol using biodiesel

glycerol as the substrate.

2. Develop a process that produces arabitol with high yield using glycerol as a substrate

by improving the medium composition and optimizing the pH, temperature and

dissolved oxygen concentration.

3. Improve the productivity of arabitol and reducing the amount of residual glycerol

concentration at the end of the fermentation process.

4. Check the ability of Candida bombicola to produce sophorolipids from glycerol as

carbon source in comparison with glucose as carbon source.

1.4 Structure of dissertation

The dissertation was divided into two parts: (1) Studies on improving arabitol production from glycerol with total 9 chapters. (2) Production of sophorolipids using biodiesel glycerol.

Chapter I contained Background, scope and objectives of this research work.

Chapter II contained the summary of the biodiesel production and glycerol market value from this industry. Also contained in this chapter was literature survey on sophorolipids, their structure, production and applications. Finally, the chapter contained literature survey about the polyols especially arabitol, its properties, production and potential applications compared to xylitol.

Chapter III contained the materials, microorganisms and analytical methods that were used in this research study. Chapter IV, chapter V, chapter VI and chapter VII dealt with another part of the project which was the production of arabitol using glycerol as substrate.

Chapter IV and chapter V contained studies done to improve the yield of arabitol production. Chapter IV dealt with the strain screening process for arabitol production

8 using glycerol as substrate and studying factors that affected the production and yield of arabitol. Chapter V contained the results of the optimization studies to improve arabitol production. Chapter VI and chapter VII contained the results from the studies conducted to improve the arabitol productivity and cell growth rate. Chapter VII dealt with improving the productivity of arabitol by studying the effects of magnesium concentration, different nitrogen source, inoculum size and use of combination of organic and inorganic nitrogen source. Studies in chapter VII contained attempts to improve arabitol productivity and reduce the residual glycerol concentration at the end of the fermentation by the addition of xylose. This chapter also contained the studies that evaluate the ability of Debaryomyces hansenii to produce arabitol using bagasse hydrolysate mainly containing glucose and xylose as carbon sources.

Chapter VIII was another part of the project that dealt with the production of sophorolipids with Candida bombicola using biodiesel glycerol as substrate and its comparison with glucose as substrate.

Finally, chapter IX summarized all the conclusions obtained from this project and future recommendations were also included.

CHAPTER II

LITERATURE SURVEY

2.1 Biodiesel byproduct glycerol

Biodiesel is an attractive alternative fuel to petroleum that can be produced from renewable resources such as soy bean oil, sunflower oil, coconut oil, palm oil and rapeseed oil [2]. Petroleum usage raises concerns about its increasing price, availability and its effect on global warming. Thus, biodiesel having advantages of being produced from renewable sources, biodegradability and non-toxic nature makes it preferable over petroleum. However, price of biodiesel production is high than petroleum and hence there is a need to improve the economics of biodiesel production. In the biodiesel production process, for every 9 kg of biodiesel produced there is 1 kg of glycerol produced as a byproduct with impurities like spent catalyst, residual methanol, methyl esters and free fatty acids. Methylesters/biodiesel is produced from triglycerides and fatty acids. From triglycerides, methyesters are produced by transesterification with methanol in the presence of base catalysts [31].

Hence, glycerol is an inevitable byproduct in any path of biodiesel production.

Biodiesel production was increasing until 2008 and the production is as high as 500 millions of gallons/year in 2009 as shown in Figure 2.1, resulting in the production of about 45 million gallons of impure glycerol. Also, the market value of both crude and

refined glycerol decreased over the years [32] (Figure 2.2). Crude glycerol does not find many applications and hence will become a waste unless it finds applications making it a valuable byproduct thereby improving the economics of the biodiesel industry.

Figure 2.1 Biodiesel production in millions of gallons over the years in United States of

America.

Figure 2.2 Market price of refined and crude glycerol produced from biodiesel industry over the years.

2.2 Polyols

Polyols refer to alcohols containing multiple hydroxyl groups. Polyols are classified based on the number of carbon atoms. Sugar alcohols are found in fruits and vegetables. Sugar alcohols are a group of polyols that are most commonly used in food industry or known as compounds containing multiple hydroxyl groups available for organic reactions for polymer synthesis. Properties of polyols include anticariogenicity, cariostatic effect which makes them advantageous to use as a sweetener [12] [13].

Polyols such as xylitol, maltitol and lactitol are mostly used as sucrose substitutes.

Polyols that are used in food industry are mostly used as sweeteners due to their low caloric content. Polyols are used as starting materials for the production of rare sugars such as L-xylulose, L-sorbose and D-tagatose [33-35]. Properties of some of these polyols are summarized in Table 2.1 [36-38].

Chemical reduction of sugars produces their corresponding polyols. For example, polyols such as xylitol is currently produced by chemical reduction of xylose derived from wood hydrolysate under alkaline conditions [20]. This process requires high pressure (50 atm) and temperature (80-140 °C), use of expensive catalyst and extensive separation steps. Such process requires high operating costs and poses difficulties during purification. Xylitol production from xylose by biological processes has also been explored and such microbial production path could reduce the operating costs [21-25].

Yeast species such as Candida tropicalis, Candida parapsilosis and Candida guilliermondi produces xylitol when xylose is used as substrate. This conversion of xylitol from xylose is catalyzed by NAD(P)H coupled xylose reductase enzyme [22].

0.7 0.2 Non None Erythritol cariogenic Very Strong

1 4 high Weak Sucrose Cariogenic Moderately

0.5

- 2.6 Low Strong 0.4 Slightly Sorbitol cariogenic and laxative

0.5

- 1.6 and Low Strong 0.4 Slightly laxative Mannitol cariogenic

1 2.4 Low Very Anti strong Xylitol laxative cariogenic and mildly

25; Weak<5 - -

0.6 0.2 Non None Strong Arabitol cariogenic and laxative

35; Moderate35; 5

0.5 None Weak Strong laxative Isomaltitol polyols

0.9 2.1 Non Weak Slightly Maltitol cariogenic and laxative

2 0.3 Non None Lactitol laxative Moderate cariogenic and strong H) kcal/g:Very strong >35; Strong 25 ∆

Other ling effect Coloric lood sugar secretion Effect on Sweetener properties Sweeteness B Sweetness was related to sucrose sweetness=1 Cooling effect(

level & insulin content, kcal/g Coo - - Table 2.1 ofSummary properties of some of the 1 2

Xylitol in particular is the best sugar among other polyols due to its ability to replace sucrose as a sweetener because of its high sweetening and anitcariogenic property. Also it can be used as substitute for diabetic patients since it does not require insulin and glucose-6-phosphate dehydrogenase for regulation of metabolism [39, 40].

2.3 Arabitol

D-arabitol is also known as D-arabitol, D-arabitol, and D-lyxitol. It is a polyol found in lichens [41]and mushrooms [42]. Chemically it is pentane-1,2,3,4,5-pentol. Based on the structure of pentane-1,2,3,4,5-pentol, it exhibits optical isomerism with L and D forms of arabitol. Also, D-arabitol is a stereoisomer to xylitol and the structures are shown in

Figure 2.3. Arabitol is known to be produced by osmophilic yeast species such as

Saccharomyces rouxii [43], Debaryomyces hansenii, Pichia guilliermondii [44],

Hansenula polymorpha [45], Endomycopsis chodati, Candida tropicalis [46], yeast like fungus Moniliella tomentosa [47] and fungus Geotrichum candidum [48].

Xylitol Arabitol

Figure 2.3 Structures of xylitol and arabitol.

2.3.1 Properties and applications

One of the important and possible applications of arabitol is as an alternative sweetener.

The use of alternative sugars is increasing these days and hence the demand. These alternative sugars should possess some of the following properties to make them replaceable for sucrose: (1) low caloric content, (2) diabetic reducer, (3) control of dental caries, (4) at least as sweet as sucrose, (5) colorless, odorless, non-cariogenic and metabolized normally, and (6) cost effective production. There has not been much research done on arabitol as a sweetener but xylitol is a well-known sugar alternative to sucrose with several advantages such as anticariogenic, high cooling effect upon consumption (because xylitol dissolution is an endothermic reaction resulting in reducing temperature), and nearly as sweet as sucrose. Xylitol finds many applications in food industry like chocolate, chewing gum, hard coating applications, baked goods and dairy products [37, 49-51].

Arabitol being a stereoisomer to xylitol possesses some of the similar properties such as the non-cariogenic property and cooling effect. Properties of arabitol, xylitol and some other polyols are compared in Table 2.1. It is showed that arabitol has lower caloric content than xylitol and it can be used as an alternative sweetener to sucrose for diabetic patients. Wallace [52] conducted an experiment on metabolism of xylitol and arabitol by rumen bacteria of sheep, the bacteria mostly belonging to the following genera: Flavobacterium, Pseudomonas, Proteus, Micrococcus, Bacilli, Propionibacteria,

Lactobacilli, Streptococci and Cocii . It was found that only 0.24 g/L arabitol and no xylitol could be consumed from the feeding of 1 g/L arabitol 1 g/L and xylitol.

Streptococcus mutans is also among the most common bacteria responsible for dental

15 caries [53]. The above finding suggested the potential application of arabitol in reducing dental caries. This type of bacterial inhibition was supported by another study with E. coli where it was hypothesized that the catabolism of arabitol by E. coli involved the production of arabitol phosphate, which induced the synthesis of compounds that inhibit the metabolism of bacteria [13]. Arabitol could also be used as a diet control sweetener since it has low caloric content and is not metabolized by humans [9, 14, 54].

It is highly possible that arabitol can be used in many of the known applications of xylitol, as a natural sweetener, a dental caries reducer, and a sugar substitute for diabetic patients [17]. If desirable, arabitol can also be converted to xylitol, for example, by using Glucanobacter oxydans [18]. This bacterium was capable of oxidizing D- arabitol to D-xylulose using the membrane-bound D-arabitol dehydrogenase and then converting D-xylulose to D-xylitol using the membranebound D-xylitol dehydrogenase.

Xylitol yield of around 25% has been reported [19].

Another possible application of arabitol is for synthesis of polyol-based polymers.

In a study conducted by Langer et al., xylitol based polymers were made using citric acid and sebacic acid. Figure 2.4 represented the scheme of reactions in producing xylitol based polymers. Xylitol upon polycondensation with water soluble citric acid yielded biodegradable water soluble polymer. Hydroxyl groups of this pre polymer were functionalized using methacrylanhydride. The water soluble prepolymer was then photopolymerized using a photoinitiator which resulted in a poly(xylitol-co-citrate) hydrogel. Prepolymer formed by the polycondensation of xylitol and sebacic acid resulted in an elastomer up on further polycondensation. Elastomer thus formed was reported to be a more biocompatible material than poly(L-lactic-co-glycolic acid) and

16 have potential applications as a biomaterial which have good biocompatibility both in vitro and in vivo [55].

Figure 2.4 Xylitol polymerization steps with (A) citric acid or (B) sebacic acid to form hydrogel and elastomer.

A study by the Department of Energy identified arabitol, and its enantiomer xylitol, as one of the top 12 biomass-derivable building block chemicals. Arabitol and xylitol can be transformed into several groups of chemicals like arabonic/ arabinoic acid, xylaric/xylonic acid, propylene glycol, and ethylene glycol [56].

2.3.2 Production of arabitol and its biosynthetic pathway

D-arabitol could be produced by catalytic reduction of D-arabinose or lyxose or by chemical reduction of lactones of arabinonic and lyxonic acids [57-59]. This process

18 uses expensive catalyst and high temperatures of 100 °C. Instead, arabitol could be produced by osmophilic yeast species. Osmophilic yeasts accumulate polyols under osmotic stress and the details of this mechanism were described as follows: Increasing the external osmolarity causes hyperosmotic stress on yeast cells resulting in outflow of water [60]. Under such circumstances yeast adapts to this environment by accumulation of polyols such as glycerol, erythritol, arabitol, xylitol, and mannitol in the cytoplasm of the cell to prevent dehydration [61, 62]. These compounds that are accumulated by yeast are known as osmolytes or compatible solutes and their role is to increase the osmolarity in the cell. The concentration of these osmolytes or compatible solutes was regulated either by metabolic pathway or by membrane transport control [63, 64]. Some yeast species belonging to Saccharomyces genus accumulated only one polyol whereas other yeasts accumulate mixture of polyols. Also, the type of polyol accumulated was independent of the solute in the medium responsible for osmotic stress [65]. However, there was a conflict in such studies since Onishi reported that the yeast Saccharomyces rouxii accumulates glycerol under salt stressed conditions and arabitol under glucose stressed conditions [66].

There is a wide range of carbon sources that could be used for arabitol production including glucose, sucrose, xylose and arabinose [43, 45, 67].

Arabitol is synthesized via the pentose phosphate pathways as shown in Figure

2.5. Ribulose-5-phosphate is considered as an important precursor for production of polyols like arabitol, xylitol and erythritol [68]. When glucose is used as the substrate, two routes have been reported for Zygosaccharomyces rouxii, that glucose is converted to ribulose 5-phosphate, which is then converted either to ribulose by ribulokinasae or to

19 xylulose 5-phosphate by ribulose 5-phosphate epimerase. Ribulose is reduced to arabitol by an NADPH-dependent arabitol dehydrogenase. Xyluolse 5-phosphate is dephosphorylated to xylulose by xylulokinase and then reduced to arabitol by an NADH- dependent arabitol dehydrogenase [69-71].

Arabitol synthesis from xylose also may follow two possible ways (Figure 2.5), as reported in the studies with Z. rouxii and Aerobacter aerogenes [72]. In the first route xylose is reduced to xylitol and then to xylulose. In the second route xylose is directly converted to xylulose by xylose isomerase. The xylulose formed from either route is then reduced to arabitol by arabitol dehydrogenase.

Arabitol syntheses from sorbitol and glycerol, if occurring, are expected to follow similar routes as the synthesis from glucose after they are converted to glucose-6- phosphate (Figure 2.5). Sorbitol is first converted to fructose-6-phosphate (via fructose) or to sorbitol-6-phosphate. Fructose-6-phosphate is then converted to glucose-6- phosphate. As for glycerol, the metabolic pathway in yeasts like Candida utilis and

Saccharomyces cerevisiae is initiated by glycerol kinase and a mitochondrial sn-glycerol

3-phosphate dehydrogenase [73]. An alternative pathway in yeasts lacking glycerol kinase is indicated by the presence of NAD-dependent glycerol dehydrogenase and dihydroxyacetone kinase [74]. Dihydroxyacetone phosphate, formed in the above routes, is converted to glyceraldehyde-3-phosphate and, subsequently via gluconeogenesis pathway, to glucose-6-phosphate. C. utilis has been reported to utilize glycerol faster than

S. cerevisiae [73]. There seem to be no reports on the uptake transport system of glycerol in C. utilis, although glycerol transport by simple diffusion has been described for S. cerevisiae [75].

Figure 2.5 Possible pathways for the production of arabitol with substrates such as glucose, glycerol, the production withglucose, substrates for sorbitolxylose such as 2.5 Possiblearabitol and Figure of pathways

2.4 Biosurfactants

Surfactants are amphiphilic molecules that contain both hydrophobic and hydrophilic moieties and are able to partition themselves at interface between fluids of different polarity and hydrogen bonding. This partitioning effect results in reducing interfacial tension and also can dominate the interfacial rheological behavior and mass transfer, reduces surface tension and critical micelle concentration. Critical micelle concentration or CMC is the concentration of the surfactant above which addition of any more surfactant results in the formation of micelles.

Surfactants find many applications in emulsion, detergent, textile, food processing and cosmetic industries due to these properties. Biosurfactants are biological molecules that are amphiphilic in nature and partition at interface especially with high surface activity. Synthetic and biological surfactants have comparable physiochemical properties such as chemical diversity, decreasing interfacial tension, pH stability and emulsifying properties. Apart from these properties, biosurfactants have the advantage of being easily biodegradable and environmentally friendly. Thus, biosurfactants find more applications in bioremediation, bio pesticides, food industry and dispersion of oil spills than their synthetic counterparts.

Biosurfactants are classified according to their chemical composition and their microbial origin. Most biosurfactants have a hydrophilic moiety consisting of either ionic or nonionic and mono-, di-, or polysaccharides, amino acids, carboxylic acids or peptides and a hydrophobic moiety consisting of saturated, unsaturated or hydroxylated fatty acids

[76]. Biosurfactants are predominantly known to be produced by hydrocarbon degrading microorganisms since the function of biosurfactants was related to its hydrocarbon

22 uptake. However, biosurfactants were also produced from glucose, sucrose, glycerol and ethanol [77-81]. In case of a strain of bacillus subtilis, the structure of the surfactant produced depended on the amino acid concentration in the medium. For example, the ratio of homologous lipopeptides Val7-surfactin and leu7-surfactin varied with the addition of different α-amino acids [82]. Biosurfactants have a wide variety of chemical structures such as glycolipids, lipopeptides, fatty acids, polysaccharide-protein complexes and neutral lipids [21]. Glycolipids and phospholipids are the most common and well- studied groups of biosurfactants. Among these most studied biosurfactants are sophorolipids and rhamnolipids, which belong to the glycolipid group that are carbohydrates in combination with long-chain aliphatic acids or hydroxylaliphatic acids.

Despite the aforementioned advantages of biosurfactants compared to synthetic surfactants, biosurfactants are not widely used due mainly to economic reasons.

Developing a process that uses inexpensive substrates and maximizes yield from the production fermentation processes is important for the expanded use of biosurfactants.

2.5 Sophorolipids

Among the most important biosurfactants, sophorolipids are important biosurfactants which are produced mostly by Candida bombicola [83], Candida apicola

[84], Wickerhaamiella domericqiae [85]and Torulopsis gropengiesseri [86].

Sophorolipids are extracellular biosurfactant molecules consisting of dimeric sugar linked with a glycosidic bond to a hydroxyl fatty acid [16]. The nature of the fatty acid in the sophorolipids varies with the type of hydrophobic carbon source used and mixtures of sophorolipids are formed. Acidic and lactonic sophorolipids are two different types which differ in their lactonization, acetylation of the sugar moiety and fatty acid saturation [87].

Opened (acidic) and closed (lactonic) structures as shown in Figure 2.6. When the acid is attached to the carbohydrate by the hydroxyl group, leaving a free carboxyl function then it forms an opened (acidic) structure. On the other hand, when the free carboxyl group is linked to the sophorose molecule it forms a closed (lactonic) structure [17]. Due to the existence of lactonic and acidic forms, sophorolipids have diverse physiochemical properties. Properties such as lowering surface tension, foam formation capacity, solubility in organic solvents and water and antimicrobial activity are affected by the degree of lactone formation.

[17]

Opened structure (Acidic SL) Closed structure (Lactonic SL)

R, R1, R2 = Acetate or Hydrogen atom

Figure 2.6 Molecular structure of Sophorolipids

2.5.1 Properties and applications

Sophorolipids have diverse physiochemical properties such as solubility in organic solvents and water and surface activity depending on their structure. In general, lactonic sophorolipids have stronger properties of lowering surface tension and antimicrobial activity [88]. And the acidic sophorolipids at pH > pKa are more soluble in aqueous solutions and more foaming [18].

The presence of acetyl groups on sophorolipids lowers the hydrophilicity and enhances the antiviral and cytokine stimulating effects of glycolipids [88]. The non- acetylated sophorolipids are readily soluble in water whereas diacetylated sophorolipids are not due to their oily nature with maximum of 70 mg/L solubility [89]. Acidic sophorolipids solubility depends on the pH. For pH greater than 6 the mixture is 30 times more soluble than at pH 4. For example, solubility of acidic sophorolipids was 0.4 g/L in water with no pH control and at pH adjusted to 6-8 has 12-15 g/L solubility [90, 91]. Due to the differences in the composition and purity of samples and measurement methods used, there are different values of interfacial and surface tension and CMC reported. It was reported that sophorolipids lowered surface tension of water from 72.8 mN/m to 40-

30 mN/m, with CMC of 40-100 mg/L [92]. 10 mg/L of purified lactonic sophorolipids reduced the interfacial tension between n-hexadecane and water from 40 mN/m to 5 mN/m at 20-90 °C and pH 6-9. Acidic sophorolipids, on the other hand, reduced interfacial tension of hexadecane to 1-2 mN/m [93]. It was reported that a mixture of purified lactonic sophorolipids lowered surface tension of water to 36 mN/m, with 10 mg/L CMC [89]. Recently, Ashby et al. reported that a mixture of sophorolipids, mostly

(≥ 92%) lactonic sophorolipids, showed a minimum surface tension of 35-36 mN/m,

CMC of 35 to > 200 mg/L(depending on the lipid precursor used such as steric acid, oleic acid or linoleic acid), and a rapeseed oil-water interfacial tension of 3-7 mN/m [94].

Hydrophilic/lipophilic balance (HLB) values vary in a wide range for sophorolipids with higher values for acidic sophorolipids and lower values for acetylated lactonic sophorolipids. HLB of acetylated acidic sophorolipids was 8-15 in acid form or as sodium or potassium salts and 4-7 in the form of calcium salts. Non-acetylated acidic

25 sophorolipids have a HLB value of 30-40 with good cleansing ability; diacetylated acidic sophorolipids act as excellent emulsifiers; and monoacetylated acidic sophorolipids have foaming and cleansing ability [95, 96].

Sophorolipids are the cheapest and highly industrially significant in commercial market [19]. The acetylated lactonic sophorolipids have applications in cosmetics as antidandruff, deodorant, bacteriostatic and moisturizers [97] [25]. Sophorolipids can also be used as an anti-fungal agent against plant and seed pathogenic fungi [26].

Sophorolipids are used in washing the drill cutting polluted by hydrocarbons and in generating hydrocarbon from mud [98]. Sophorolipids have properties such as inhibiting free radical and elastase activity, and stimulating dermal fibrinolytic property [96, 99,

100]. Also, it can be used as a cleaning agent in dish washer [100]. Antibacterial properties of sophorolipids make it applicable in fighting against infectious diseases and

Table 2.2 summarizes the Minimum Inhibition Concentration (MIC) for different bacterial species [101, 102]. Also, sophorolipids showed cytotoxic effect on cancer cells leading to its application in liver and pancreatic cancer treatment [103, 104]. A comparison of the properties such as surface tension and CMC of sophorolipids, rhamnolipids, alkyl dodecyl benzene and sodium dodecyl sulfate is shown in the following Table 2.3.

MIC (µg/mL) of Microorganisms SL-1 SL-2 SL-3

Bacillus subtilis 6 25 500

Staphylococcus epidermidis 6 25 NI

Streptococcus faecium 15 29 NI

Glomerella cingulate NI 50 NI

Table 2.2 MIC of different sophorolipids resulting in reducing growth up to 50% of normal growth.

NI-No Inhibition

Surfactant Minimum surface tension, CMC, mg/L mN/m

Sophorolipids 33-37 10-200

Rhamnolipids 26-29 5-200

Detergent alkyl dodecyl 47 590 benzene Sodium dodecyl sulfate 37 2000-3000

Table 2.3 Comparison of properties of biosurfactants and synthetic surfactants.

2.5.2 Production

Sophorolipids are glycolipid surfactants that were first identified to be produced by yeast Torulopsis magnoliae isolated from sow thistle petals [105]. This species was later renamed as Torulopsis apicola and is currently known as Candida bombicola.

Recently, sophorolipids were also found to be produced by Wickerhamiella domericqiae, which was isolated from oil-containing waste water samples [85].

The hydrophilic moiety of sophorolipid is a dimeric sophorose, which is linked with a glycosidic bond to a hydroxyl group of a fatty acid, the hydrophobic moiety.

Sophorolipids were produced industrially by Candida (Torulopsis) bombicola fermentation [9].Using C. bombicola sophorolipids could be produced by using only hydrophilic substrates such as glucose, sucrose, cheese whey, molasses and glycerol or only hydrophobic substrates such as fatty acids, hydrocarbons, fatty acid esters and glycerides or both hydrophilic and hydrophobic substrates. [106-109]. The possible pathway for the production of sophorolipids from carbohydrates is shown in Figure 2.7.

Sophorose is obtained by modification and transformation of carbohydrates and fatty acids are formed via acetyl-CoA from glycolysis pathway. Sophorose and fatty acids are connected to form sophorolipids [110].

Figure 2.7. Possible biosynthetic pathway for sophorolipids production using carbohydrate as substrate.

Another possible pathway from triglycerides as substrate is shown in Figure 2.8 where triglycerides are converted to glycerol and fatty acids by hydrolysis and glycerol is then converted to sugar (sophorose) via gluconeogenesis. Sophorose and fatty acids combine to form sophorolipids [110].

Figure 2.8. Possible biosynthetic pathway for the production of sophorolipids using triglycerides as substrate.

Sophorolipids are secondary metabolites of C. bombicola, that is, sophorolipids production is not associated with cell growth and the production starts only when the limiting nutrient (in most cases, the nitrogen source) is exhausted. Sophorolipids produced are excreted into the culture broth during fermentation. Factors affecting sophorolipids production are the type of carbon source, energy substrates and/or precursors, pH, dissolved oxygen concentration, temperature and medium composition.

Sophorolipids could be produced by primary carbon sources alone such as glucose, sucrose, galactose, molasses and whey or by secondary carbon source / lipid precursors alone such as oils and fatty acids. Primary carbon sources mostly act as the energy sources and also as precursor for carbohydrate backbone. Secondary carbon source/lipid precursor is mostly used as precursor for lipid moiety. However, using a

29 combination of both resulted in increased sophorolipids production and yield [111]. The type of pre-cursor/lipidic substrate used has an effect on sophorolipids production. In general, fatty acid chain of sophorolipids has 16-18 carbon atoms. Using pre-cursors of chain lengths C16- C19 resulted in direct incorporation of this into the hydrophobic moiety of sophorolipids without altering the length or structure of hydrocarbon. For pre-cursors of chain lengths greater than C19 and less than C16 the cells had to shorten or lengthen the chain in order to incorporate into the hydrophobic moiety of sophorolipids. In comparison, yield of sophorolipids production was higher for pre-cursors with chain lengths C16-C24 than those with chain lengths C12-C15 [112, 113].

For higher productivity sophorolipids could be produced by both fed-batch and continuous fermentation; however, it should be noted that the lipid pre-cursors are to be added step-wise in continuous fermentation due to the inhibitory effect on yeast cells upon accumulation of fatty acids [107, 109]. Accumulation of fatty acids represses

NADPH production by inhibiting glucose-6-phosphate dehydrogenase, the first enzyme of pentose phosphate pathway [107]. NADH production is vital as it is the cofactor of hydroxylase which catalyzes the hydroxylation of fatty acids to form the hydroxyl fatty acids required for sophorolipids production. Hence, the added fatty acids should not get accumulated in the fermentation broth.

Dissolved oxygen concentrations of ≥ 5-15% were required to achieve high sophorolipids production yields by C. bombicola [114-116]. Garcia-Ochoa and Casas investigated the effect of dissolved oxygen concentration on sophorolipids production using an unstructured kinetic model. This study showed that high dissolved oxygen concentrations improved the sophorolipids production and yield by at least 60% [117].

Temperature also plays an important role in sophorolipids production. It was observed that 21 °C was optimal temperature for sophorolipids production by C. bombicola [118]. C. bombicola was able to grow at the temperature range of 21-37 °C but sophorolipids production significantly dropped at temperatures higher than 30 °C

[114].

Different pH values in the fermentor results in the production of different types of sophorolipids. For example, water soluble acidic sophorolipids were formed at pH less than 2 and crystalline lactonic sophorolipids were formed at pH 3-5 [114, 118-120]. At any pH higher than 5, sophorolipids production deteriorated [114, 119]. Resting-cell fermentation for sophorolipids production had optimal pH of 3.5 while that of batch fermentation was 3.5-4.5 because the higher pH was more favorable for cell growth.

There were some reports regarding the change in the nature of sophorolipids at different pH. For example, solubility of sophorolipids produced in the aqueous medium increased when pH was raised from 3.5 to 5 [120]. It was further observed that water soluble sophorolipids was converted into crystalline sophorolipids by pH adjustment but not vice versa [119]. This change in the nature of sophorolipids was not observed in two individual studies by Spencer et al and Goebbert [114, 118]. However, the effect of pH on sophorolipids production was consistent in most of the reported studies.

CHAPTER III

MATERIALS AND METHODS

3.1 Microorganisms and medium

Candida bombicola and Debaryomyces hansenii are the strains use in this study. C. bombicola (ATCC 22214) was obtained from United States Department of agriculture,

Agricultural Research Service Culture Collection, Peoria, IL. D. hansenii was also obtained from Agricultural Research Service culture collection –Northern Regional

Research Laboratories (ARS-NRRL), USDA. This strain was selected from screening about 214 strains belonging to different genera as shown in Table 3.1 for their ability to produce arabitol from glycerol as a substrate. Among these several strains that were screened, strains belonging to genera Dearyomyces and Geotrichum resulted in noticeable amounts of polyols (> 5g/L) after 3 days of cultivation in shake flasks. But

Geotrichum strains shoed the production of mannitol along with arabitol with arabitol / mannitol ration about 2. On the other hand, some of the strains belonging to

Debaryomyces produced arabitol as the only polyol. Hence, Debaryomyces strain was selected for further optimization studies. These cultures were stored on agar plates in refrigerator (at -4 °C) in 5 g/L of peptone, 20 g/L glucose and 3 g/L of yeast extract and 3 g/L of malt extract medium solution.

# of # of Genera strains Genera strains Debaryomyces 67 Lachancea 1 Geotrichum 41 Torulaspora 1 Metschnikowia 37 Naumovozyma 1 Candida 24 Kodamaea 1 Dipodascus 14 Sugiyamaella 1 Pichia 5 Hanseniaspora 1 Trigonopsis 4 Cephaloascus 1 Galactomyces 4 Botryozyma 1 Zygosaccharomyces 2 Trichomonascus 1 Citeromyces 1 Sporopachydermia 1 Saccharomycopsis 1 Endomyces 1 Hyphopichia 1 Schizoblastosporion 1 Wickerhamia 1

Table 3.1 Genera and number of strains screened for arabitol production from glycerol as substrate.

3.2 Experimental setup and equipment

The experimental set up and medium compositions used in different chapters of the dissertation are described in the following sections.

3.2.1 Arabitol production

Some of the studies were done in Queue orbital shaker (Queue systems, Parkersburg,

West Virginia) in 250 ml Erlenmyer flasks at 250 rpm and 30 °C. Most of the optimization studies such as pH, dissolved oxygen concentration and medium composition were done in 2 L Bioflo fermentors (model number 110). The experimental

33 set-up for this study was shown in Figure 3.1. The dissolved oxygen concentration in the fermentation broth was measured using Ingold polarographic DO probe (InPro 6830

Model # 52201017). The dissolved oxygen concentration was controlled at required value by automatic variation in the agitation speed by the Bioflo software. pH in the broth was measured by a pH probe and was automatically controlled by the addition of 1 N sulfuric acid and 0.05 N sodium hydroxide.

The inoculum preparation and starting of the fermentation was described as follows: One loopful of cells was tr ansferred from the agar plates that were stored in the refrigerator into 250 ml Erlenmeyer flask containing 50 ml of growth medium. The cells were then allowed to grow by stirring on the stir plate at room temperature for 24 hrs.

Then cells were transferred into 100 ml of medium with 5% inoculum size and were allowed to grow for 36 h by vigorous stirring. 5% inoculum size cells were transferred into 2L culture with 1 L culture volume under sterile conditions.

Figure 3.1 Experimental set-up for batch fermentation for arabitol production

3.2.2 Sophorolipids production

Medium composition that was used for sophorolipids production was as follows in g/L:

Glucose/glycerol, 100; NH4Cl, 4; MgSO4.H2O, 0.5; yeast extract, 5; and KH2PO4, 1. This study was done in 600 ml glass fermentor with 400 ml culture volume. This culture was maintained at room temperature, pH of 4.0 and dissolved oxygen concentration (DO) above 5% air saturation (maintained only during exponential growth phase). DO was manually controlled by varying agitation while keeping aeration constant at 1 VVM (i.e.,

400 ml/min of air flow rate). Dissolved oxygen concentration was manually controlled only during exponential growth phase that was between 24-65 h. Dissolved oxygen concentration was measured by an optical micro-sensor PSt3 (PreSens Precision Sensing

GmbH) that was attached on the inner side of the fermentor wall. The micro-sensor has an oxygen-sensitive dye immobilized in a silicone matrix attached on a polyester foil.

The dissolved oxygen concentration was then acquired by the tip of an optical fiber that was held on the sensor on the outside wall of the fermentor. The other end of the optical fiber was connected to the oxygen meter Fibox 3-trace v3 (PreSens Precision sensing

GmbH) which in turn connected to a computer for data acquisition and recording. pH in the medium was maintained at 4.0 by the addition of 1 N hydrochloric acid and 0.5 N sodium hydroxide. Vegetable oil (Soybean) was also added to the medium as an additional carbon source (lipid precursor) using a syringe pump. Vegetable oil was added at a rate that would result in glycerol/ vegetable oil weight ratio of about 5 during the fermentation. The vegetable oil addition was adjusted according to this ratio since it was reported earlier that sophorolipids production increased with controlled addition of hydrophobic carbon source (lipid precursor).

3.3 Analytical methods

The sample collected from the fermentation broth at any time of fermentation was centrifuged for both sophorolipids and arabitol production. Cell pellet thus obtained was used for cell concentration measurements and the supernatant was saved in freezer for sugar, polyols and ammonium analysis.

3.3.1 Glucose analysis

For sophorolipids system when glucose was used as substrate, the following enzymatic glucose assay kit (Sigma P7119, procedure No.510) was used. This method was based on the following enzymatic reactions:

Glucose oxidase Glucose + 2H2O +O2 Gluconic acid + H2O2 ------(3.1) Peroxide H2O2 + o-Dianisidine Oxidized o-Dianisidine (Brown) ------(3.2)

Glucose was first oxidized to gluconic acid and hydrogen peroxide by glucose oxidase as shown in equation (1). Hydrogen peroxide thus formed oxidizes colorless o-dianisidine in the presence of peroxide to form brown o-Dianisidine. The absorbance of this sample was measured at 450 nm with UV/Vis spectrophotometer (Model UV-1601, Shimadzu

Corporation, Columbia, MD). This absorbance measured was converted to glucose concentration by developing calibration from their standards using this method.

3.3.2 Sugar and polyols analysis

For samples taken in the later studies with D. hansenii, the glycerol, glucose, xylose, and arabitol concentrations were similarly measured by HPLC (Shimadzu) with a

36 refractive index detector, using a different carbohydrate column (Supelco column H, 250 x 4.6 mm, with a guard column, 50 x 4.6 mm) maintained at ambient temperature. The mobile phase used was 0.1% H3PO4 at a flow rate of 0.17 ml/min. The representative retention times for glucose, xylose, arabitol and glycerol were 14.2, 15.3, 17.2 and 20.7 minutes, respectively. Sorbitol and glycerol concentration was analyzed using a

Supelcosil LC-NH2 column (250 x 4.6 mm) with 75:25 acetonitrile: water as the mobile phase at a flow rate of 1.0 ml/min. The sorbitol peak appeared at a retention time of about

16.3 minutes. A linear equation was obtained from the calibration line of respective standards of 1 to 10 g/L. Samples were diluted accordingly with their respective mobile phase to be within their calibration range.

3.3.3 Cell concentration measurement

Two types of yeast strains were used for the production of arabitol and sophorolipids.

Candida bombicola was used for sophorolipids production and Debaryomyces hansenii was used for arabitol production.

Candida bombicola:

Cell concentration was determined using Bradford protein assay kit II (Bio-rad

Laboratories, Hercules, CA) by measuring the intracellular protein concentration. This method was based on Bradford dye-binding method which involves the binding of proteins with aromatic and basic amino acids by comassie blue G-250 dye [121]. Bovine serum albumin was used as a standard for protein calibration. A 5-ml broth sample was centrifuged at 8,000 rpm for 10 min (Sorvall RC 5c, DuPont, Wilmington, DE). The supernatant was collected and frozen for future analyses of substrate and product

(arabitol) concentrations. The cell pellet was washed twice with de-ionized water and then lysed by addition of 5 ml of 0.2 N NaOH and heating at 100 °C for 20 min. The protein concentration of the lysate was measured according to the Bradford assay, with the absorbance at 595 nm measured using a UV/VIS spectrophotometer (Model UV-

1601, Shimadzu Corporation, Columbia, MD).

Debaryomyces hansenii

Cell growth was measured either by dry weight or from intracellular protein measurement analysis. In case of dry weight analysis 10 ml of the sample taken from the fermentation broth was centrifuged and separated from supernatant. The Cell pellet thus obtained was re-suspended in de-ionized water and centrifuged again. Supernatant thus obtained was discarded and re-suspended in 10 ml of de-ionized water. This solution was transferred onto pre-weighted aluminum pan and was allowed to dry in an oven at 100 °C for 24 hrs. Other method of measuring the cell concentration was described as follows.

Cell growth was observed by measuring optical density at 610nm using UV/VIS spectrophotometer (Model UV-1601, Shimadzu Corporation, Columbia, MD) for all the shaker studies. Elsewhere cell concentrations were determined from the intracellular protein concentrations measured using the Bradford protein assay kit II (Bio-rad

Laboratories, Hercules, CA) as described in section 3.2.2 for Candida bombicola. The relationship between the intracellular protein concentration and the cell dry-weight concentration was established with the samples taken during the exponential growth phase of 2 repeated batch fermentation experiments. The following relationship was obtained (R2 = 0.92):

Cell dry-weight concentration (g/L) = Intracellular protein concentration (g/L) × 12.42.

3.3.4 Ammonia nitrogen analysis

Additionally, a more exact measurement of ammonia, and nitrate concentrations was carried out when necessary using an ammonia electrode. Using this method, the ammonium concentrations could be accurately measured in a wide range from 1–1000 mg/L of NH4+-N. The nitrate (and nitrite) in the sample were reduced to ammonia using titanous chloride and measured using the ammonia electrode as well. For this case, the sample was properly diluted to match a narrower concentration range of 1 – 20 mg/L.

3.3.5 Sophorolipids analysis

Sophorolipids analysis was done using Anthrone analysis method [122]. Sample taken from fermentation broth was not separated from cells and the pH of that sample was adjusted to 2 to 2.5 by the addition of 1N HCL. This sample was then extracted with four-fold volume of ethyl acetate for 12-15 hrs. Ethyl acetate phase was then separated from the aqueous phase. Ethyl acetate was then vaporized by air drying. The lipids thus remained were dissolved in 0.05 M NaHCO3. After dissolving and anthrone reagent (2 g/L anthrone in sulfuric acid) was added and the reaction mixture was kept in a heating block maintained at 95 °C for 16 minutes. Then the absorbance was measured at 625 nm with the UV/Vis spectrophotometer. A calibration curve using sophorose (Sigma-

Aldrich) as the standard was developed which was used for the conversion absorbance to released sophorose. Sophorolipids concentration was estimated by multiplying the measured sophorose concentration by 2.04, which was the ratio of molecular weight of two main sophorolipids (687.9 & 705.8) that were produced in the fermentation and molecular weight of sophorose (342).

CHAPTER IV

PRODUCTION OF ARABITOL FROM GYCEROL: STRAIN SCREENING AND

STUDY FACTORS AFFECTING PRODUCTION YIELD

4.1 Introduction

Biodiesel motor fuel produced from renewable sources such as vegetable oil and animal fat is an attractive alternative to petroleum-derived fuel [123]. In biodiesel production using trans esterification of triglycerides, glycerol is the major byproduct produced: about 1 kg of glycerol is formed for every 9 kg of biodiesel produced [3].

Biodiesel consumption in the United States has increased dramatically from 75 million gallons in 2005 to 450 million gallons in 2007. Accompanying this increase was the production of around 45 million gallons of glycerol in 2007 [4]. Refined glycerol has numerous applications in the food, drug, textile and cosmetic industries whereas crude glycerol produced from the biodiesel industry is of low value because of impurities such as spent catalyst, salts after neutralization, residual methanol, methyl esters and free fatty acids [5, 6]. The economics of the biodiesel industry is strongly influenced by the value

of its byproducts, and developing new uses for biodiesel glycerol is imperative to the sustainability of the biodiesel industry [7, 8].

In this study, the biodiesel byproduct glycerol was used as the substrate for production of arabitol, a polyhydric alcohol. A study by the Department of Energy identified arabitol, and its enantiomer xylitol, as one of the top twelve biomass-derivable building block chemicals. Arabitol and xylitol can be transformed into several groups of chemicals like xylaric/xylonic acid, arabonic/arabinoic acid, propylene glycol and ethylene glycol [10]. Arabitol and xylitol have melting points of 103 °C and 93 °C, respectively. Both are highly soluble in water and both form white crystals when purified

[11, 12]. The catabolism of arabitol by Escherichia coli involves the formation of arabitol phosphate which induces the synthesis of compounds that inhibit bacterial metabolism

[13]. This property makes it possible to use arabitol as a sweetener that reduces dental caries. Also, the caloric value of arabitol is 0.2 kcal/g, whereas it is 2.4 kcal/g for xylitol

[9, 14-16]. It is highly possible that arabitol can be used in many of the known applications of xylitol, as a natural sweetener, a dental caries reducer and a sugar substitute for diabetic patients [17]. If desirable, arabitol can also be converted to xylitol, for example, by using Glucanobacter oxydans [18]. This bacterium was capable of oxidizing D-arabitol to D-xylulose using the membrane-bound D-arabitol dehydrogenase and then converting D-xylulose to D-xylitol using the membrane-bound D-xylitol dehydrogenase. Xylitol yield of around 25% has been reported [19].

Arabitol is known to be produced by osmophilic yeast species such as

Debaryomyces [26], Candida [27], Pichia [28], Wickerhamomyces (Hansenula) [29] and

Saccharomycopsis (Endomycopsis) [30]. When exposed to osmotic stress, the yeast

41 accumulates compatible solutes such as arabitol, glycerol, xylitol, erythritol and mannitol to balance the osmotic pressure across the cell membrane. An objective of this study was to select the yeast strains that produce large amounts of arabitol with high yields and minimal other polyols (for easier downstream separation) using glycerol as substrate.

Experiments characterized the most promising strains for growth and arabitol production under different culture conditions.

4.2 Materials and Methods

This section presents the different strains that were screened for their ability to produce arabitol from glycerol, culture conditions and the medium composition that were used for all the studies conducted in this chapter.

4.2.1 Yeast strain screening

Extensive culture screening of 214 strains from 25 genera was conducted for arabitol production from glycerol. The following 5 genera contained the largest numbers of strains screened: Debaryomyces, Geotrichum, Metschnikowia, Candida and

Dipodascus. A complete list of the genera and the numbers of screened strains from each genus is given in Table 6.1. All of these strains were obtained from the Agricultural

Research Service (ARS) Culture Collection (NRRL) at National Center for Agricultural

Utilization Research (NCAUR), United States Department of Agriculture, Peoria, IL.

4.2.2 Media

The yeasts were maintained on YM agar slants (5.0 g peptone, 10.0 g glucose, 3.0 g yeast extract, 3.0 g of malt extract, and 20.0 g of agar in 1L deionized water). Cultures were grown for inoculation in YM broth (without agar). The medium used in the 42 screening study and in later studies on cell growth and arabitol production by D. hansenii

SBP-1 (NRRL Y-7483) had the following composition (per liter of solution): yeast extract, 3 g; (NH4)2SO4, 2 g; K2HPO4, 2.4 g; KH2PO4, 1.6 g; MgSO47H2O, 1 g; and glycerol 100 g (unless specified otherwise). The glycerol used in this study was a crude glycerol from a biodiesel plant without a glycerol refinery (provided by Biodiesel

Systems, Madison, WI). The crude glycerol had 88% glycerol. To make 100 g/L glycerol in the culture medium, 113 g/L of the crude glycerol was added. The medium had an initial pH of 6.6-6.8. Glycerol (and other carbon sources used in some studies, i.e., glucose, xylose and sorbitol) was autoclaved separately from other medium components.

4.2.3 Culture methods

(a) Screening

Three ml of YM broth in a sterile screw top test tube was inoculated with yeast cells from an YM slant. The tube was then placed in an Innova 4430 incubator/shaker

(New Brunswick Scientific) set at 25 °C and 200 rpm for 24 hours. Next, 100 µl of

YM/yeast culture was pipetted into 10 ml of the screening medium in a 50 ml flask. The flask was then placed in an Innova 4430 or an Innova 4335 incubator shaker set at 30 °C and 200 rpm for 72 hours. Initial screening was completed with a single replicate of each yeast strain tested. The results from these experiments were later confirmed by re- screening in duplicate. After 72 hours of reaction time, the flasks were removed from the incubator/shaker and processed for analysis of glycerol and polyol product concentrations.

(b) Culture conditions for D. hansenii SBP-1 (NRRL Y-7483)

Inoculum was prepared by transferring a loop of cells from an agar plate to 50 ml

YM broth in a 250 ml flask covered with cheese cloth. The culture was grown at room temperature (22 ± 1 C) for 24 h under vigorous magnetic stirring. 2.5 ml of the inoculum thus prepared was added to each flask in the subsequent study of culture conditions, which were made with a 50 ml medium volume in 250 ml flasks shaken at 200 rpm. The temperature used in these studies was 30 °C except in the study of temperature effects.

Multiple samples were taken during cultivation to establish the profiles of cell growth, substrate consumption and product formation.

(c)Effect of medium volume in shaker flasks

Shake flasks are not very suitable for studying the effects of dissolved oxygen concentrations (DO) on cell growth and product formation. Nonetheless, to obtain an indication of culture sensitivity to low DO or anaerobic conditions, a study was done with

D. hansenii SBP-1 in 250 ml shake flasks containing the following different medium volumes: 30, 50, 75, 100 and 150 ml. Under the same shake speed (200 rpm), the flasks with smaller volumes were expected to have better oxygen transfer efficiency via surface aeration, resulting in higher broth DO for the cultures of similar cell concentrations reached in the N-limited culture medium.

4.2.4 Analytical methods

Cell concentration and sugar concentration profiles were obtained using the following analysis methods. Among the different sugar and polyols that were analyzed were glucose, arabitol, xylitol, glycerol, sorbitol and xylose.

Cell concentration

Cell concentrations were mostly determined from the intracellular protein concentrations measured using the Bradford protein assay kit II (Bio-rad Laboratories,

Hercules, CA). A 5-ml broth sample was centrifuged at 8,000 rpm for 10 min (Sorvall

RC 5c, DuPont, Wilmington, DE). The supernatant was collected and frozen for future analyses of substrate and product (arabitol) concentrations. The cell pellet was washed twice with de-ionized water and then lysed by addition of 5 ml of 0.2 N NaOH and heating at 100 °C for 20 min. The protein concentration of the lysate was measured according to the Bradford assay, with the absorbance at 595 nm measured using a

UV/VIS spectrophotometer (Model UV-1601, Shimadzu Corporation, Columbia, MD).

The relationship between the intracellular protein concentration and the cell dry-weight concentration was established with the samples taken during the exponential growth phase of 2 repeated batch fermentation experiments. The following relationship was obtained (R2 = 0.92):

Cell dry-weight concentration (g/L) = Intracellular protein concentration (g/L) × 12.42.

Substrate and product concentrations

For culture samples taken during the screening study, 10 g/L ribitol or rhamnose was added as internal standard. The samples were then centrifuged at 12,000 rpm, 10 °C

45 for 15 minutes using a Beckman Coulter model J2-21 centrifuge to collect the supernatant. The supernatant was concentrated about twofold using a speed vacuum

(Thermo Electron Corporation, Waltham, MA) set at 65 °C. After being passed through a pre-swollen DE52 (diethylaminoethyl cellulose, Whatman) anionic exchange resin in

Bio-Rad mini-columns, the sample was filtered and analyzed by high-performance liquid chromatography (HPLC, Shimadzu) with a Waters 410 differential refractive index detector. Used at a flow rate of 0.6 ml/min, the mobile phase was 100% water generated by filtering deionized water through an ELGA PURELAB Ultra system. The main column used was a Rezex RCM-Monosaccharide column (300 × 7.8 mm, 8 µm,

Phenomenex, Torrance, CA) maintained at 85-90 °C by a column heater. Three guard columns were used in series: a Phenomenex Security Guard with Carbo-Ca (4 × 3 mm) cartridge and 2 Phenomenex Rezex RCM (50 × 7.8 mm, 8 µm) guard columns. The resulting chromatograms were compared to the chromatograms of known standards and calibration curves for identification and quantification of the polyols present. The following were representative HPLC retention times (in minutes) observed for the internal standards and relevant polyols: rhamnose - 16.8, ribitol – 20, glycerol - 21.6, mannitol - 23.3, arabitol - 23.5 and xylitol - 27.2. The typical standard errors in the analyzed polyol concentrations were 1-7%.

For samples taken in the later studies with D. hansenii SBP-1, the glycerol, glucose, xylose, and arabitol concentrations were similarly measured by HPLC

(Shimadzu) with a refractive index detector, using a different carbohydrate column

(Supelco column H, 250 x 4.6 mm, with a guard column, 50 x 4.6 mm) maintained at ambient temperature. The mobile phase used was 0.1% H3PO4 at a flow rate of 0.17

46 ml/min. The representative retention times for glucose, xylose, arabitol and glycerol were

14.2, 15.3, 17.2 and 20.7 minutes, respectively. Sorbitol concentration was analyzed using a Supelcosil LC-NH2 column (250 x 4.6 mm) with 75:25 acetonitrile:water as the mobile phase at a flow rate of 1.0 ml/min. The sorbitol peak appeared at a retention time of about 16.3 minutes.

With retention times of 23.3 and 23.5 minutes, mannitol and arabitol could not be reliably could not be separated by the HPLC method used used in the screening study.

When a broader (or double/shouldering) mannitol/arabitol HPLC peak was detected, a

GC analysis was done to determine the ratio of mannitol and arabitol. This ratio was then used to quantify the two polyols from the peak in the HPLC analysis. For GC analysis the samples (50 µl) contained in test tubes were first dried in a heat block (Thermolyne type

17600 Dri-Bath) at 60 °C under air using a Pierce model 18780 Reacti-Vap evaporating unit. Then, 1 ml of 45 g/L hydroxylamine hydrochloride in pyridine was added to each sample, vortexed, and then placed in the heat block for 10 minutes. The samples were vortexed again and placed back into the heat block for another 10 minutes. Next, the samples were removed from the heat block and cooled under cold running water to near room temperature. Then, 1 ml of acetic anhydride was added to each test tube and the above vortexing and 10-minute heating in the heat block were repeated twice. After being cooled again under cold running water to near room temperature, the samples were completely dried under air. Following this treatment, each test tube received 3 ml of water and 2 ml of ethyl acetate, was vortexed thoroughly, and then left standing for phase separation. The top layer (ethyl acetate phase) was collected into a 1 dram vial. Another

2 ml of ethyl acetate was added to each test tube and the above extraction procedure was

47 repeated. The ethyl acetate extract was again collected and combined with the previous extract in the 1 dram vial. The extract in the vial was completely dried under nitrogen using an Organomation N-EVAP analytical evaporator, and then redissolved in 300 µl of ethyl acetate. After being transferred into a GC vial, the sample was analyzed on a

Hewlett Packard HP 6890 series GC with a Phenomenex Zebron ZB-5 column. The resulting chromatograms were compared to chromatograms of known polyol standards mainly for determination of the ratios of mannitol and arabitol which could not be separated by the HPLC method used. The following were representative retention times

(in minutes) observed in the GC analysis: glycerol - 6.8, ribitol (internal standard) - 17.1, arabitol - 17.4, xylitol - 17.8 and mannitol - 22.1.

4.3 Results and discussion

The initial part of this section has the results from the screening process and strain selection. Later parts of this section dealt with the factors such as temperature, initial glycerol concentration and substrates other than glycerol as carbon source that affect the production of arabitol with the selected strain.

4.3.1 Screening for arabitol production from glycerol

Among the cultures screened (Table 4.1), the genera Debaryomyces and

Geotrichum had the largest numbers of strains that produced noticeable amounts ( 5 g/L) of polyols from glycerol, after 3 days of cultivation in the shake flasks.

# of # of

Genera strains Genera strains

Debaryomyces 67 Lachancea 1

Geotrichum 41 Torulaspora 1

Metschnikowia 37 Naumovozyma 1

Candida 24 Kodamaea 1

Dipodascus 14 Sugiyamaella 1

Pichia 5 Hanseniaspora 1

Trigonopsis 4 Cephaloascus 1

Galactomyces 4 Botryozyma 1

Zygosaccharomyces 2 Trichomonascus 1

Citeromyces 1 Sporopachydermia 1

Saccharomycopsis 1 Endomyces 1

Hyphopichia 1 Schizoblastosporion 1

Wickerhamia 1

Table 4.1 Genera and number of strains screened

Debaryomyces and Metschnikowia strains tended to produce predominantly arabitol whereas Geotrichum strains produced significant amounts of mannitol, in addition to arabitol. Examples for the distribution of different polyols produced are compared in Table 4.2 for several strains. Selected strains from these genera, specifically

D. hansenii (SBP-1, NRRL Y-7483), G. candidum (SBP-12, NRRL Y-552), G. cucujoidarum (SBP-219, Y-27732), and M. zobellii (SBP-14, NRRL Y-5387), were examined further for the effects of some cultivation conditions. More thorough studies were done with D. hansenii (SBP-1) because the minimal amount of non-arabitol polyols

49 produced by this strain was expected to significantly simplify the downstream arabitol purification process.

The screening study conducted in this work showed that the species from different genera produced different polyols or polyol mixtures from glycerol. Using glucose as the substrate, others have reported the production of various polyols such as erythritol, xylitol and mannitol by different species [71, 124-126]. D. hansenii SBP-1 was chosen for more detailed evaluations in this work because it produced arabitol as the predominant polyol and with high concentrations (> 10 g/L in shake flasks).

Total Polyol distribution (%)

Species SBP # NRRL # polyol Arabitol Xylitol Mannitol Ribitol (g/L)

Debaryomyces 1 Y-7483 10 97.8 1.6 ND 0.6 hansenii

D. hansenii 2 Y-1015 11 97.4 2.6 ND ND

Geotrichum candidum 12 Y-552 14 65.3 1.0 33.7 ND

G. cucujoidarum 194 Y-27731 19 59.0 0.8 39.4 0.8

G. cucujoidarum 219 Y-27732 13 71.7 0.8 25.9 1.6

Metschnikowia zobellii 14 Y-5387 5 94.9 ND ND 5.1

Table 4.2 Percentages of different polyols produced by some osmotolerant yeast strains

ND-Not detectable.

4.3.2 Effect of culture volume in shake flasks

The different medium volumes (30, 50, 75, 100, and 150 ml) used in the systems studied were expected to cause different profiles (varying with time) of dissolved oxygen concentrations (DO) in the broth. DO profiles were, however, difficult to follow in shake- flask cultures. Instead, the concentrations of D. hansenii SBP-1 cells, arabitol produced and glycerol consumed were compared in Figure 4.1 to show the possible effects of DO.

The cell concentrations were measured at 80 h because the preliminary study had shown that the cultures would typically have reached the stationary phase by 80 h. Arabitol and glycerol concentrations were measured at 120 h, to allow ample time for arabitol production. The systems with 30, 50 and 75 ml medium were found to have comparable results for all 3 concentrations (cells, arabitol and glycerol), with p > 0.05 from one-way

ANOVA (analysis of variance) using Minitab (Minitab Inc, State College, PA). The systems with 100 and 150 ml medium reached lower cell and arabitol concentrations and consumed less glycerol, presumably due to the insufficient oxygen transfer in these larger-volume systems. More importantly, the yields of arabitol from consumed glycerol remained about 20% (19%-22%) in the 3 systems with lower volumes but decreased to

10% and 5% as the volume increased to 100 ml and 150 ml, respectively. The results indicated that the 50 ml volume used in the initial screening study was suitable. The same volume was used in all the subsequent shake-flask studies. The results also suggested that very low or zero DO, corresponding to the systems of larger medium volumes, was not good for arabitol production.

40 Arabitol produced Glycerol consumed 30 Cells

10 Concentration (g/L) Concentration 0 0 50 100 150 Initial medium volume (ml)

Figure 4.1 Cell concentration measured in g/L at 80 h and concentration of arabitol and glycerol at 120 h in g/L in systems with different culture volumes.

4.3.3 Effect of temperature

The concentrations of arabitol produced at different temperatures by D. hansenii

(SBP-1), G. candidum (SBP-12) and M. zobellii (SBP-14), after 3 days of cultivation, were compared in Figure 4.2. All of these strains showed maximal arabitol production at

30 °C. D. hansenii (SBP-1) was found particularly sensitive to higher temperature, giving negligible arabitol production at 35 °C. Arabitol production by M. zobellii (SBP-14) was, on the other hand, similar at 30 °C and 35 °C.

Figure 4.2 Arabitol produced by selected strains of Debaryomyces, Geotrichum and

Metschnikowia at different temperatures. Samples were taken after 3 days of cultivation.

4.3.4 Effect of initial glycerol concentration

Arabitol production is associated with osmophilic yeasts [127]. The effects of glycerol and salt concentrations, both of which can provide osmotic pressure to the cells, are described in this and the next sections, respectively.

Shown in Figure 4.3(a) are the cell concentrations of D. hansenii SBP-1 at 0, 72, and 120 h in the systems with 50, 90, 120 and 150 g/L of glycerol in the initial media.

The cell concentrations were comparable, reaching 17-20 g/L, presumably because all were limited by the same N-source concentration in the media. Glycerol was not completely exhausted in any systems at 120 h (glycerol concentration data not shown).

The profiles of arabitol production in these systems are shown in Figure 4.3(b). In the system with 50 g/L glycerol initially, the arabitol production essentially stopped after 72 h (when the remaining glycerol concentration dropped below 20 g/L). Arabitol production continued after 72 h in the systems with higher initial glycerol concentrations.

The concentrations of arabitol produced and glycerol consumed at 120 h are summarized in Figure 4.3(c). The arabitol production in the 3 systems with high initial glycerol concentrations ( 90 g/L) appeared to be comparable whereas the glycerol consumption decreased with increasing initial glycerol concentrations. The resultant arabitol yields from the consumed glycerol at 120 h were shown in Figure 4.3(d). The arabitol yield increased with the increase in initial glycerol concentration, particularly from 50 g/L to

90 g/L. The arabitol yield reached about 50% in the system with 150 g/L of initial glycerol. The findings suggested that certain glycerol concentration (and/or its associated osmotic pressure) was required for arabitol synthesis by the osmophilic yeast.

Figure 4.3 Effects of different initial glycerol concentrations on D. hansenii SBP-1 fermentation: (a) cell growth profiles, (b) arabitol production profiles, (c) concentrations of glycerol consumed and arabitol produced at 120 h, and (d) arabitol yield (from consumed glycerol) at 120 h

4.3.5 Effect of salt concentration

The above results also indicated that certain concentrations of glycerol would remain unconsumed when the arabitol production by D. hansenii SBP-1 became very

55 slow or stopped. The remaining glycerol would complicate the downstream collection and purification of arabitol. It was thought that salt (NaCl) might offer the necessary osmotic stress for complete conversion of glycerol to arabitol. D. hansenii was reported to tolerate high salt concentrations, up to 4M NaCl [128].

The study was made in media containing 100 g/L of glycerol and 0, 50, 100 and

150 g/L of NaCl, respectively. The cell growth was not affected by addition of 50 and

100 g/L NaCl but was slowed significantly in the system with 150 g/L NaCl (Figure 6.4).

Arabitol production was more sensitive to the salt addition (Figure 4.4). Presence of even

50 g/L NaCl caused significantly poorer arabitol production. The system with 150 g/L

NaCl produced less than 1 g/L of arabitol.

Figure 4.4 Effects of different salt concentrations on D. hansenii SBP-1 fermentation with 100 g/L of initial glycerol concentration: Cell growth and arabitol production profiles To separate the effect of NaCl addition on arabitol production from that on cell growth, a subsequent study was made with the salt being added after 2 days of cell growth in the medium with 100 g/L glycerol. Three systems, with 0 (control), 100 and

150 g/L NaCl, respectively, were included for comparison. Delaying the salt addition successfully minimized the negative effect on cell growth (data not shown). Arabitol production was, however, completely stopped after the salt addition (Figure 4.5). It is therefore concluded that high salt concentrations have negative effects on arabitol production by D. hansenii SBP-1. It is infeasible to use salt addition to apply osmotic pressure for complete conversion of glycerol to arabitol.

Effects of NaCl addition (25, 50 and 100 g/L) on arabitol production were also evaluated with other strains, including D. hansenii SBP-2 and SBP-5, G. candidum SBP-

12, and G. cucujoidarum SBP-194 and SBP-219. Salt addition was found to have similar negative effects on arabitol production by these strains (data not shown).

Figure 4.5 Effects of salt addition on arabitol production by D. hansenii SBP-1 in media with 100 g/L of initial glycerol concentration. The salt was added after 2 days of growth.

D. hansenii is known to tolerate high concentrations of sugar and salt [129]. It was reported to tolerate salt stress up to 240 g/L although the cell growth rate decreased at NaCl concentrations above 160 g/L [130]. In the current study, the growth of D. hansenii SBP-1 was observed not to be affected by the presence of 100 g/L NaCl, but the growth rate decreased significantly at 150 g/L NaCl.

Zygosaccharomyces rouxii was found to produce glycerol and arabitol in glucose- based media [131]. The glycerol and arabitol distribution in the polyol mixture produced depended on the salt and glucose concentrations used. Less arabitol (10 g/L) than glycerol (29 g/L) was produced when the initial medium contained 180 g/L NaCl and 100 g/L glucose. On the other hand, the concentrations of arabitol and glycerol produced became similar (about 32 g/L) when the medium had a higher glucose concentration (300 g/L) and a low NaCl concentration (0.1 g/L). According to the study results, high osmotic

58 pressures (generated by high salt and/or sugar concentrations) induced this yeast to produce glycerol while high sugar and low salt concentrations were more favorable for arabitol production. These findings were largely consistent with the results of the current study. In the current study, high initial glycerol concentrations ( 90 g/L) and no or low

NaCl concentrations (<< 50 g/L) were found to be favorable for arabitol production and yield. High salt concentrations, on the other hand, tended to inhibit yeast growth and, particularly, arabitol production. Cell growth was affected at salt concentrations higher than 100 g/L; arabitol production was inhibited even at 50 g/L NaCl.

In another study with a different strain of D. hansenii, the intra- and extra-cellular contents of polyols, i.e., glycerol and arabitol, were measured in response to increasing salinity in a medium containing only 5 g/L glucose [132]. It was found that in the medium containing 181 g/L NaCl, the yeast accumulated intracellular glycerol and released extracellular glycerol (without production of either intra- or extracellular arabitol) while glucose was available; after glucose was depleted, both intra- and extracellular glycerol were consumed and the intracellular arabitol content increased (still without production of extracellular arabitol). While high salt concentrations were again found to cause glycerol production from sugar, the sugar concentration was too low (5 g/L) in this study to allow production of extracellular arabitol.

4.3.6 Effects of addition of other carbon substrates

Since salt addition could not be used to promote more complete conversion of glycerol to arabitol, the effects of addition of a second carbon source (glucose, xylose, or sorbitol), along with glycerol, on arabitol production by D. hansenii SBP-1 were investigated in four systems. The study was made in four systems. The culture was first

59 grown in the medium with an initial glycerol concentration of 30 g/L. After 74 h (when the cultures were in the early stationary phase), 30 g/L glucose, xylose, or sorbitol plus 50 g/L glycerol were added to 3 of the systems, and 80 g/L glycerol was added to the 4th

(control) system. All of the systems reached similar maximum cell concentrations (about

16 g/L, data not shown). The resultant arabitol concentration profiles are shown in Figure

6.6(a). Before the addition of more carbon substrates at 74 h, all of the systems produced about 2 g/L arabitol. The subsequent addition of 80 g/L glycerol (in the control system) did not lead to much more arabitol production (Figure 4.6(a)). Such a two-step addition of glycerol (30 g/L and then 80 g/L) appeared to be less favorable for arabitol production, when compared to the addition of all the glycerol in the initial medium (see the arabitol profiles for the systems with 90 g/L and 120 g/L of initial glycerol concentrations in

Figure 4.3(b) and the profiles for the control systems in Figures 4.4 and 4.5). The addition of sorbitol along with glycerol also did not give good arabitol production (Figure

4.6(a)). On the other hand, additions of glucose and xylose significantly improved the arabitol production. Concentrations of the potential second C-source (glucose, xylose or sorbitol) and glycerol consumed after the addition in the stationary phase (during 74-145 h) are summarized in Figure 4.6(b). More glycerol was consumed in the control system

(added with only glycerol) than in the other 3 systems. Little sorbitol was consumed. The lower glycerol consumption in this system might be caused by the lower added glycerol concentration (50 g/L, as compared to 80 g/L in the control) or by the inhibition of sorbitol. On the other hand, glucose and xylose were simultaneously or preferentially consumed by the yeast. Arabitol was the only major metabolite even in the presence of other carbon source such as xylose and glucose.

50.0

(b) 40.0

30.0

20.0 Suorce Suorce Consumed (g/L) -

C 10.0

0.0 Glu + Gly Xyl + Gly Sorb + Gly Gly

2nd C-source Gly

Figure 4.6 Effects of addition of 30 g/L glucose, xylose or sorbitol as a potential second carbon source, along with 50 g/L glycerol, on arabitol production by stationary-phase D. hansenii SBP-1. Cells were grown for 74 h in media containing 30 g/L glycerol before the second C-source and glycerol were added. The control system was added with 80 g/L glycerol. Arabitol production profiles are compared in (a). Concentrations of the second

C-source and glycerol consumed by the stationary-phase cultures (during 74-145 h) are shown in (b). Although not shown, the standard deviations of the consumed concentrations in (b) were in the range of 13% to 22%.

Arabitol production was found to be improved by addition of glucose and xylose

(but not sorbitol). Arabitol is synthesized via the pentose phosphate pathways [71]. The possible routes are summarized in Figure 4.7.

Ribulose-5-phosphate is considered as an important precursor for production of polyols like arabitol, xylitol and erythritol [68] With glucose as the substrate, two routes have been reported for Zygosaccharomyces rouxii [71] [127, 133]: Glucose is converted to ribulose 5-phosphate, which is then converted either to ribulose by ribulokinasae or to xylulose 5-phosphate by ribulose 5-phosphate epimerase. Ribulose is reduced to arabitol by an NADPH-dependent arabitol dehydrogenase. Xyluolse 5-phosphate is dephosphorylated to xylulose by xylulokinase and then reduced to arabitol by an NADH- dependent arabitol dehydrogenase.

Arabitol synthesis from xylose also may follow two possible ways (Figure 4.7), as reported in the studies with Z. rouxii and Aerobacter aerogenes [71, 72]. In the first route xylose is reduced to xylitol and then to xylulose. In the second route xylose is directly converted to xylulose by xylose isomerase. The xylulose formed from either route is then reduced to arabitol by arabitol dehydrogenase.

Arabitol syntheses from sorbitol and glycerol, if occurring, are expected to follow similar routes as the synthesis from glucose after they are converted to glucose-6- phosphate (Figure 4.7). Sorbitol is first converted to fructose-6-phosphate via fructose or sorbitol-6-phosphate. Fructose-6-phosphate is then converted to glucose-6-phosphate. As for glycerol, the metabolic pathway in yeasts like Candida utilis and Saccharomyces cerevisiae is initiated by glycerol kinase and a mitochondrial sn-glycerol 3-phosphate dehydrogenase [134]. An alternative pathway in yeasts lacking glycerol kinase is

62 indicated by the presence of NAD-dependent glycerol dehydrogenase and dihydroxyacetone kinase [74]. Dihydroxyacetone phosphate, formed in the above routes, is converted to glyceraldehyde-3-phosphate and, subsequently via gluconeogenesis pathway, to glucose-6-phosphate. C. utilis has been reported to utilize glycerol faster than

S. cerevisiae [134]. There seem to be no reports on the uptake transport system of glycerol in C. utilis, although glycerol transport by simple diffusion has been described for S. cerevisiae [135].

The nucleotide sequences for the relevant enzymes reported have been searched and compared with the D. hansenii genome (NC_006048) using the NCBI-BLAST

(National Center for Biotechnology Information-Basic Local Alignment Search Tool).

The matching percentages are indicated in Figure 4.7 by different arrow styles. This yeast’s ability to convert sorbitol to fructose-6-phosphate is noticeably less certain, potentially responsible for the insignificant sorbitol utilization observed in this study

(Figure 4.6(b)).

As mentioned in the Introduction, xylitol (but not arabitol) has many known or existent commercial applications. It may therefore be desirable to produce xylitol from arabitol. Xylitol is currently produced by chemical reduction of xylose derived from wood hydrolysate [20]. The chemical process uses an expensive catalyst at high pressure

(50 atm) and temperature (80-140 °C). Xylitol production from xylose by biological processes has also been explored [21-25]. Yeasts can convert xylose to xylitol using

NAD(P)H-coupled xylose reductase. Unfortunately, the xylitol produced tends to be oxidized to xylulose by NAD+-coupled xylitol dehydrogenase. Good xylitol yields from such a process require tightly controlled, high intracellular NAD(P)H/NAD+ ratios. This

63 control is not an easy task in large-scale industrial operations where the environment

(particularly the dissolved oxygen concentrations) inside the large bioreactors is non- homogeneous. Furthermore, both of the above chemical and biological processes require expensive separation of xylose from the complex sugar mixtures in the biomass hydrolysate. The alternative approach of producing arabitol and then converting the arabitol to xylitol has the potential of being more economically attractive.

4.4 Conclusions

In summary, strains belonging to the genera Debaryomyces and Geotrichum produced total polyol concentrations  5 g/L after 3 days of cultivation among the cultures screened. However, while Debaryomyces strains were able to produce arabitol as the main metabolite, Geotrichum strains produced mannitol along with arabitol. The optimal temperature for arabitol production by D. hansenii SBP-1 was 30°C. Very low

DO or anaerobic conditions inhibited arabitol production. High initial glycerol concentrations improved the yield of arabitol, reaching approximately 50% (w/w) at about 150 g/L initial glycerol. Though the growth of D. hansenii SBP-1 was not noticeably affected in media with up to 100 g/L NaCl, arabitol production by the yeast was negatively affected by high salt concentrations. Presence of glucose and xylose, in addition to glycerol, in the medium improved the arabitol production. Sorbitol, on the other hand, was not a good substrate for the yeast and replacing a portion of glycerol with sorbitol led to poorer arabitol production. A study to further improve the arabitol production is being conducted, by optimizing the operating parameters such as pH, dissolved oxygen concentration, choice of limiting nutrient, and medium composition.

for the conversion of various substrates to arabitol the substrates various for conversion of .7 Possible pathways .7 Possible pathways Figure 4 Figure

CHAPTER V

ARABITOL PRODUCTION FROM GLYCEROL USING DEBARYOMYCES

HANSENII: EFFECTS OF MEDIUM N/P RATIO, pH AND DISSOLVED OXYGEN

CONCENTRATION

5.1. Introduction

Arabitol, a polyhydric alcohol, is a stereoisomer to xylitol. Xylitol has many known applications, particularly as a low calorie sweetener and an anticariogenic agent

[37, 38]. Arabitol and xylitol were also considered among the top twelve biomass derivable building block chemicals that can be transformed into several other groups of useful chemicals and unique polymers [10]. As enantiomers, arabitol and xylitol are likely to have some similar properties. Xylitol has been shown in many studies to inhibit growth of bacteria [136]. Although much less studied, arabitol has also been reported to inhibit growth of E. coli, via production of arabitol phosphate that induces synthesis of other toxic metabolites [13]. Similar to xylitol, arabitol may be applied as an anticariogenic agent by inhibiting the growth of oral bacteria. Arabitol has an extremely low calorie content of 0.2 kcal/g, as compared to 2.4 kcal/g for xylitol and 4 kcal/g for sucrose [14, 15, 137]. Hence, arabitol may find unique applications as a natural, very

low-calorie sweetener and sugar substitute for diabetic patients [17]. Arabitol can also be effectively converted to xylitol, for example, by Glucanobacter oxydans [18].

This arabitol-to-xylitol route is an alternative to the rather expensive xylitol production processes based on chemical [20] or biological [138] reduction of xylose, which needs to be first purified (or at least significantly enriched) from the hydrolysate of lignocellulosic materials.

Osmophilic yeast like Candida, Pichia, Debaryomyces, Hansenula and

Endomycopsis are known to produce arabitol. Glycerol, arabitol, xylitol, erythritol and mannitol are the possible polyols accumulated by these species when exposed to osmotic stress [26, 28, 68, 139, 140]. Different carbon sources such as sucrose, glucose, arabinose, xylose and xylulose can be used for arabitol production but the arabitol produced is typically mixed with other polyols (including glycerol) and sometimes ethanol [43, 71, 141]. In an earlier study we reported the production of arabitol as the only polyol when a selected Debaryomyces hansenii strain was cultured on glycerol

[142]. Glycerol is a major byproduct from biodiesel production processes and represents a potentially inexpensive substrate for lowering the arabitol production cost. In the earlier study the optimal temperature was found to be 30C. From experiments made in shake flasks, there was also qualitative indication that the dissolved oxygen concentration

(DO) should not be too low. In the literature reports, with glucose as the substrate, different yeast species such as Moniliella tomentosa, Metschnikowia reukaufii,

Kodamaea ohmeri and Zygosaccharomyces rouxii showed different optimal pH for arabitol production [47, 71, 124, 143]. These studies used 100-200 g/L glucose in batch fermentation and the optimal pH ranged from 3.4 to 6. For example, pH 5 was reported as

67 the optimal for arabitol production, at higher than 60% yield, by M. reukaufii using 200 g/L glucose [124].

In this current work D. hansenii fermentation was studied in pH- and DO- controlled fermentors. The objective was to determine the effects of pH, DO and medium nitrogen-to-phosphorus (N/P) ratio on the arabitol yield and productivity.

5.2. Materials and Methods

This section gave the details of the microorganism used, culture conditions, medium compositions, analytical methods and type of studies conducted to improve the arabitol production by studying the optimal conditions required.

5.2.1 Microorganism and medium

The yeast culture Debaryomyces hansenii (NRRL Y-7483) was obtained from

USDA ARS culture collection (Peoria, IL). Culture was maintained on agar plates containing 10 g/L glucose, 3 g/L yeast extract, 3 g/L malt extract, 5 g/L peptone, and 20 g/L agar. The medium used to prepare inocula for the initial study of medium N-to-P ratio had the following composition (per liter): glycerol, 100 g; yeast extract, 3 g;

(NH4)2SO4, 2 g; K2HPO4, 2.4 g; KH2PO4, 1.6 g; and MgSO47H2O, 1 g. Compositions of the media evaluated in the N-to-P ratio study are given in Table 5.1. All the later studies and inoculum preparation were made with the improved medium composition (per liter): glycerol, 100 g; yeast extract, 6 g; (NH4)2SO4, 4 g; K2HPO4, 0.316 g; KH2PO4, 0.211 g; and MgSO47H2O, 1 g. The medium pH was 6.7 ± 0.2 after autoclaving. Peptone, yeast extract, malt extract and MgSO47H2O were purchased from Sigma-Aldrich; agar,

68 glucose, ammonium sulfate and glycerol from Fisher Scientific; K2HPO4 from Merck; and KH2PO4 from EMD Chemicals.

Chemical, g/L System 1 System 2 System 3 System 4 System 5

Glycerol 100 100 100 100 100

Yeast extract 2.30 3 3 2.30 6

(NH4)2SO4 2.36 2 2 2.36 4

K2HPO4 3.85 2.4 0.16 0 0.32

KH2PO4 2.57 1.6 0.11 0 0.21

MgSO4∙7H2O 1 1 1 1 1

KCl 0 0 0 1.57 0

N, g/L 0.75 0.75 0.75 0.75 1.50

P, g/L 1.25 0.83 0.08 0.03 0.17

N/P 0.6 0.9 9 30 9

Table 5.1 Compositions of media used for N/P ratio study

5.2.2 Inoculum preparation

A loop of cells was transferred from an agar plate to 50-ml medium in a 250-ml

Erlenmeyer flask covered with cheese cloth sandwiched cotton. The culture was grown at room temperature (22 ± 1 C) for 24 h under vigorous magnetic stirring. The inoculum

69 thus prepared was added at 5% of the final broth volume in the subsequent culture condition study. A preliminary study had been done to compare the cell growth at inoculum sizes of 5%, 10%, 15%, 20% and 25%. The results confirmed that 5% inoculum is adequate (data not shown).

5.2.3 Fermentation conditions

Each batch of studies was done in 2 parallel 2.5-L New Brunswick Bioflo 110 fermentors with 1-L working volume. All studies were done at 30°C, controlled by a heating jacket. Aeration was fixed at 1 L/min of 0.2-µm filter-sterilized air. DO was allowed to drop from about 100% air saturation to the studied value (5%, 10% or 20%) and then controlled at the value by automatic adjustment of the agitation speed (initially at 400 rpm). For the initial study of medium N-to-P ratio, DO was controlled at 10%.

After the DO effect study, DO was controlled at 5% for all the later studies. pH was also allowed to drop from about 6.7 to 3.5 (except in the pH effect study or specified otherwise) and then controlled by automatic acid (1 N HCl) or base (0.5 N NaOH) addition. Vegetable oil, less than 10 ml for each fermentation batch, was added as antifoam when needed.

5.2.4 Effect of medium N-to-P ratio

These earlier experiments were done with pH and DO controlled (after dropping naturally from their initial values) at 4.0 and 10%, respectively (instead of 3.5 and 5% as in the later studies). These fermentation studies were also done with continuous glycerol addition to maintain the concentration around 100 g/L, by adjusting the addition rate according to the measured glycerol consumption rate. The glycerol concentration was

70 maintained high because the previous study had suggested the benefit of high glycerol concentrations on arabitol production [142]

5.2.5 Effects of pH and DO

Initial pH studies were done at 10% DO and with continuous glycerol addition to maintain an approximately 100 g/L concentration. The pH was allowed to drop (from about 6.7) gradually with cell growth before being controlled at 3, 3.5 or 4. pH 3.5 was chosen from the above studies to be used for studying the effect of DO. DO of 5, 10 and

20% were compared in fermentations made with glycerol only from the initial medium

(without continuous addition). Three to five fermentor batches were repeated for each

DO condition. A second round of studies for pH effect were then repeated with the best

DO (5%) found from the DO effect studies. pH 3, 3.5, 4, 5 and 6 were evaluated, also without glycerol addition during the batch fermentation.

5.2.6 Analytical methods

Cell growth was monitored by measuring the intracellular protein concentrations in periodic samples using the Bradford protein assay kit II (Bio-rad Laboratories,

Hercules, CA). The intracellular protein concentrations were then converted to cell dry- weight concentrations using a pre-established calibration curve [142]. For the intracellular protein concentration measurement, a 5-ml broth sample was centrifuged at

9,000 g for 10 min (Sorvall RC 5c, DuPont, Wilmington, DE). The supernatant was collected and frozen for future analyses of glycerol and arabitol concentrations.

Arabitol and glycerol concentrations were measured using HPLC (Shimadzu LC

10A) with a refractive index detector (RID-10A). A carbohydrate column (Supelco

71 column H, 250 x 4.6 mm) with a guard column (No. 59319, 50 x 4.6 mm) was used at ambient temperature. The mobile phase was 0.1% (v/v) H3PO4 at a flow rate of 0.14 ml/min. The retention times for glycerol and arabitol were 20 and 25 min, respectively.

Calibration curves for converting the peak areas to concentrations were established with standard solutions of glycerol and arabitol.

5.3. Results and Discussion

This section presents the results of the optimal conditions required for the maximum arabitol production using glycerol as substrate and the reasons why these optimal conditions were found to be better in resulting better arabitol yield.

5.3.1 Effects of medium nitrogen-to-phosphorous ratio

The cell growth and arabitol production profiles observed in five different medium compositions are shown in Figure 5.1. The media for Systems 1 – 4 had the same N concentration of 0.75 g/L but varying P concentrations (1.25, 0.83, 0.08, and 0.03 g/L), giving the corresponding N/P ratios (w/w) of 0.6, 0.9, 9, and 30. Also, the media for Systems 1 and 4 (with a yeast extract-to-ammonium sulfate ratio of nearly 1) had a lower ratio of organic-N to inorganic-N than the media for Systems 2 and 3 did (with a

YE-to-AS ratio of 1.5). The medium for System 5 was essentially a 2X medium of

System 3, except that the glycerol and MgSO4 concentrations were kept the same.

Cell growth in Systems 1 and 4, with less organic N, appeared to be slower initially or have longer lag phases than the growth in the other systems. This effect was seen reproducibly in other shake flask experiments (data not shown). The maximum cell concentrations reached were however comparable (12-15 g/L) in all 4 systems (1 – 4).

The effect of medium N/P ratio was more significant on arabitol production. The system with too little P (System 4, N/P = 30) or too much P (System 1, N/P = 0.6) was found to be unsuitable for arabitol production. System 3, with an N/P ratio of 9, gave the highest arabitol production.

With the same N/P ratio but 2X concentrations, the medium for System 5 allowed the yeast cells to grow to 23-25 g/L, almost doubling the maximum cell concentration in

System 3, i.e., 12-15 g/L. Arabitol production in System 5 was also significantly higher.

Accordingly, this medium composition was used as the basis for the later studies. Note that this medium contained 1 g/L MgSO4∙7H2O, which was the only medium component whose concentration was not increased from that in the other media. A separate study was done with 6 shake flasks to compare the cell growth in media having 1, 1.5 and 2 g/L of MgSO4∙7H2O. No differences in growth rate or maximum cell concentration were observed (data not shown). It was thus confirmed that neither Mg nor S was limiting in the medium for System 5.

25 N/P=0.6 N/P=0.9 (a)

) N/P=9 L / 20 N/P=30 (g

N/P=9 (Double N) n o

ti 15 a tr n e c

n 10 o C

l l e

C 5

0 Time (h) 0 50 100 150

30 N/P=0.6 N/P=0.9 (b) 25 N/P=9 N/P=30

) N/P=9 (Double N)

L 20 / (g

to 15 i b a r

A 10

0 0 50 Time (h) 100 150

Figure 5.1 Profiles of cell growth and arabitol production in systems with different N-to-

P ratios in the media

As an example, the DO and pH profiles observed in System 5 are shown in Figure

5.2. DO and pH both dropped naturally to the control values of 10% DO and pH 4, in 15-

20 h and 25-30 h, respectively. Note that almost all the arabitol production occurred after the low control pH was reached.

There is still room for a closer study of the effect of medium N/P ratio. In general, yeast has about 7.5% N and 1.7% P of its dry weight [144]. Accordingly, yeast biomass has an N/P ratio of about 4.4. Phosphorous is important for the cell membrane formation and function. How the P-deficiency at slightly higher N/P ratios would affect the arabitol production is unknown. Here we examined a very wide range of N/P ratio with only 4 medium compositions (N/P = 0.6, 0.9, 9, and 30). The P-deficient medium with N/P of 9 showed the highest arabitol production among the studied systems. More studies in a narrower N/P range, say, 2 to 15, may identify an even more optimal N/P ratio for arabitol production.

120 8

%DO 7 100 pH 6 80 5

60 4 pH %DO 3 40 2 20 1

0 0 0 10 20 30 40 50 Time (h)

Figure 5.2 pH and DO profiles for the System 5 fermentation shown in Figure 5.1

5.3.2 Effect of DO

Profiles of cell growth, arabitol production and glycerol consumption are compared in Figure 5.3 for the fermentations made with different control DO values: 5%,

10% and 20%. Three to five repeated fermentations were made at each DO. Shown in

Figure 5.3 are the average results and the associated standard deviations.

Overall, the cell growth profiles at these DO values did not differ significantly, although a higher DO appeared to allow the system to reach the maximum cell concentration (about 22 g/L) faster. The system DO would not drop to the different control DO values before, on average, 15 (± 3) h (data not shown). So the cell growth rates in the first 24 h in these systems were, as expected, similar, with an average cell doubling time of 3.7 (± 0.1) h. The DO effect on growth would become more apparent in the second day of cultivation. Cell growth during the second day was significantly slower than that during the first 24 h, because the pH would also drop to the low control value (3.5) after about 27 (± 4) h. The cell doubling time after the systems reached both 75 the control pH and DO was estimated to be approximately 30 h for the 5% DO systems and 20 h for the 10% and 20% DO systems.

Overall the arabitol production and glycerol consumption profiles were also not markedly different in the systems of different control DO values. The most apparent difference in Figure 5.3(b) is that the systems with 20% DO had higher rates of arabitol production and glycerol consumption during the period of 30-72 h, compared to the rates for the systems with 5% and 10% DO. However, afterwards the arabitol production in the 20% DO systems slowed down more noticeably than the production in the lower DO systems, although plenty of glycerol (at least 40 g/L) was still available. By 120-144 h all the systems produced, on average, about 38-40 g/L arabitol from about 75-90 g/L glycerol consumed. As shown in Figure 5.3(c), the overall yield of arabitol (from consumed glycerol) was estimated at 55% (± 3%) for the 5% DO systems, 40% (± 9%) for the 10% DO systems, and 45% (± 9%) for the 20% DO systems. Also shown in

Figure 5.3(c) are the yields of arabitol produced during the stationary phase. Presumably because no glycerol was consumed for cell growth, the stationary-phase yields were higher than the overall yields.

30 5%DO (a) 25 10% DO 20% DO 20

Cell concentration (g/L) 5

0 0 20 40 60 80 100 120 140 Time (h)

160 50 Glycerol-5%DO Glycerol-10%DO Glycerol-20%DO 140 Arabitol-5%DO Arabitol-10%DO Arabitol-20%DO 40 120

100 (b) 30 80

20 60 Arabitol (g/L) Glycerol (g/L)

40 10 20

0 0 0 20 40 60 80 100 120 140 Time (h)

Stationary-phase yield 70 (c) Max Overall yield 60

50 d l e 40 Yi

% 30

0 5% 10% 20% %Dissolved Oxygen

Figure 5.3 Time profiles of (a) cell growth and (b) glycerol consumption and arabitol production, and (c) the overall and stationary-phase arabitol yields observed in systems with 5, 10 and 20% DO control values.

Despite the large standard deviations in the 10% and 20% DO systems, the arabitol yield was significantly higher in the 5% DO systems. By using Minitab (Minitab

Inc, State college, PA) the one-way ANOVA (analysis of variance) on the yields at different DO values gives p = 0.042 and 0.001 for the overall and stationary-phase yields,

77 respectively; both are < 0.05. The higher arabitol yields at 5% DO might be related to the coupled NADH oxidation (to NAD+) with arabitol formation [142, 145, 146]; the low DO condition might be more favorable for maintaining a higher intracellular ratio of

NADH/NAD+, which in turn had a positive effect on arabitol formation. Overall, it appears that a higher DO (20%) was more favorable for faster arabitol production while a lower DO (5%) was more favorable for higher product yield. It should be noted that the rather mild effect of DO in the range of 5% to 20% is good for large scale bioprocesses.

In large fermentors DO variation in different parts of the bioreactors is inevitable: higher in the high-shear zones with finer bubbles near the impellers and lower in the low-shear zones with fewer and larger bubbles. The wide range of suitable DO makes arabitol production easier in large fermentors.

5.3.3 Effect of pH

The time profiles of arabitol production in fermentation experiments made with frequent glycerol addition (to maintain glycerol concentration at approximately 100 g/L) at the same 10% DO but different control pH (3, 3.5, and 4) are shown in Figure 5.4(a).

The amounts (g) of arabitol produced are also plotted against the amounts of glycerol consumed in Figure 5.4(b), for estimation of arabitol yield from the consumed glycerol.

Arabitol production at pH 3 and 3.5 were similar up to about 100 h but the production at the pH 3.5 system slowed down significantly after 100 h, as shown in

Figure 5.4(a). Nonetheless, the arabitol yield (from consumed glycerol) was essentially the same for these two systems (Figure 5.4(b)), i.e., 17.5% for up to about 25 g of arabitol produced; lower at higher arabitol amounts. Although each of these systems with frequent glycerol addition was examined in only one fermentation experiment (without

78 repetitions), it indicated that the system of a higher pH (4) performed much worse, in both production rate and product yield, than the systems with lower control pH values (3 and 3.5).

60 pH 3 (a) 50 pH 3.5 pH 4 40

20 Arabitol conc(g/L)

0 0 50 100 150 200 Time (h)

(b)

Figure 5.4. (a) Time profiles of arabitol concentration (g/L) produced at different pH (3,

3.5 and 4) in systems with frequent glycerol addition to maintain about 100 g/L glycerol, and (b) their corresponding plots of amounts (g) of arabitol produced versus amounts (g) of glycerol consumed.

The pH effect was then studied more closely in fermentations without further glycerol addition because it had become evident by then that the frequent glycerol addition had a negative effect on the arabitol yield (< 17.5%, even lower than the yields obtained in batch fermentations made in shake flasks). The cell growth and arabitol production profiles for the fermentations made at the same 5% DO but different control pH values of 3, 3.5, 4, 5 and 6 are shown in Figure 7.5. The pH 5 fermentation was made only once while the fermentation was repeated 3 times for pH 3, 5 times for pH 3.5, and twice for pH 4 and 6. Shown in Figure 5.5 are the average results and the associated standard deviations.

40 pH 3 35 (a) pH 3.5 30 pH 4.0

25 pH 5.0

20 pH 6.0

10 Cell concentration (g/L)

0 0 20 40 60 80 100 120 Time (h)

40 pH 3 (b) 35 pH 3.5 30 pH 4.0

25 pH 5.0

20 pH 6.0 15 Arabitol (g/L) 10

5 0 0 20 40 60 80 100 120 Time (h)

Figure 5.5. Profiles of (a) cell growth and (b) arabitol production in fermentations with same (5%) control DO value but different pH control values, i.e., 3, 3.5, 4, 5 and 6

Similar to the discussion for the DO effect, because the systems would reach their different pH control values only after 24 (± 2) h, the cell growth rates in these systems during the first 24 h were about the same, with an average cell doubling time of 3.9 (±

0.2) h. But afterwards the cell growth profiles differed much for systems with different pH control values. For the two low pH systems (3 and 3.5), the cell doubling time during the second day (24 – 48 h) was estimated at 27.6 (± 0.3) h; for the three higher pH systems (4, 5 and 6), the doubling time was shorter at approximately 18.1 (± 1.8) h. The maximum cell concentrations reached were also different: roughly 17 g/L for pH 3, 21 g/L for pH 3.5, 23 g/L for pH 4, 25 g/L for pH 5, and 30 g/L for pH 6. It was not surprising to find that a higher pH (closer to neutral) was more favorable to the growth of yeast cells.

As for the pH effect on arabitol production, high pH values (4 and above) were clearly less ideal, consistent with the finding from the systems with frequent glycerol addition (Figure 5.4). But in these fermentations pH 3.5 gave much better arabitol 81 production (averaged at 38 g/L by 120 h) than pH 3 did (18 g/L) whereas in the fermentations with frequent glycerol addition, the arabitol production at these two pH values was more comparable. The reason(s) for this discrepancy remains unknown. One possibility is that pH 3 is critically low for the cells and because of the log-scale nature of pH, a small pH difference at this low level corresponds to a large difference in the proton

(H+) concentration. Therefore, for the pH 3 systems, the cell metabolism could be much more sensitive to a small deviation from the control pH value. In any case, it is clear that pH 3.5 is a much more robust control value for optimal arabitol production with the current medium and process design. Overall, it was found that for D. hansenii a high pH was more favorable for cell growth but a low pH at 3.5 was optimal for arabitol production.

The DO and pH effects found in the present study suggested that the current process and medium designs are suitable, i.e., starting with high DO (~ 100%) and neutral pH to encourage faster cell growth and allowing these parameters to drop naturally to the lower control values for higher arabitol production in the late exponential-growth phase and stationary phase. It however should be noted that lowering the pH forcefully did not work well. In another experiment (data not shown) the medium contained only organic N sources (yeast extract and peptone), to encourage faster cell growth initially. Without the consumption of ammonium to help lower the system pH, the system with only organic-N had a relatively high pH during the first 24 h. Acid was then gradually added over a period of 4 h to drop the pH to 3.5. Arabitol production was found to be completely curbed by this slow acid addition for pH adjustment. In a subsequent experiment (data not shown), instead of acid, (NH4)2SO4 was added at 27 h to

82 help lower the pH. Again the arabitol production stopped prematurely. The reasons for these negative outcomes remain unknown but the arabitol production appeared to be rather sensitive to perturbation of culture conditions.

An important issue to address is the incomplete consumption of glycerol.

Typically about 30 g/L glycerol would still be present when the arabitol production stopped at about 40-50 g/L. The observation is not too surprising, considering that arabitol production is believed to be associated with environments of high osmotic pressures [45, 147] (although high salt concentrations had been shown to have negative effect on arabitol production [147]). When the molar glycerol concentration dropped below the molar arabitol concentration, there was no more benefit for the cells to continue converting glycerol to arabitol, instead of switching to consume arabitol as the substrate. In some earlier fermentation experiments where a lower initial glycerol concentration (about 50 g/L) was used, the arabitol concentration was indeed observed to decrease when the glycerol concentration dropped low (data not shown). Nevertheless, the presence of relatively high concentrations of glycerol, which is a polyol similar to arabitol, would complicate the downstream arabitol purification. A better process design to allow more complete depletion of glycerol is desirable.

The highest volumetric productivity of arabitol obtained was about 0.35 g/L-h.

This productivity could be increased as follows. The fast cell growth period may be extended to reach high cell concentrations sooner. As reported earlier, the doubling time in the first day was about 4 h but increased to 20-30 h in the second day. High DO, closer to neutral pH and more organic-N can help to achieve this. But, it is important to

83 do so without the need to forcefully drop the pH to the low value preferred for arabitol production.

The glycerol-based medium may be supplemented by other precursor substrates for arabitol production. Xylitol has been shown to be produced more effectively by

Candida tropicalis using a glucose-xylose mixture as the carbon substrate, with glucose being the substrate supporting cell growth and xylose the precursor for xylitol production

[23, 148]. A similar approach may be applicable for arabitol production. While the xylitol production by C. tropicalis requires the use of rather pure and expensive xylose as precursor, D. hansenii can convert a wider range of substrates to arabitol, proven with glucose and xylose [147] and potentially with arabinose and other sugars in the lignocellulosic hydrolysate. Production of arabitol by this approach will be more economical than the production of xylitol.

5.4. Conclusion

The preferred operation for arabitol production was to start the fermentation at high DO and neutral pH and allow them to drop naturally, along cell growth, to the optimal control values of 5% DO and pH 3.5. With this operation, 40 g/L arabitol was produced in 5 days with 55% yield and 0.33 g/L-h productivity in the medium with an

N/P ratio of 9. Higher DO (10% and 20%) were slightly less favorable (up to 45% arabitol yield). pH effect was more significant: the systems with higher (≥ 4) or lower (3) pH produced no more than 20 g/L arabitol.

CHAPTER VI

MODIFICATION OF MEDIUM COMPOSITION IN IMPROVING SPECIFIC

PRODUCTIVITY AND VOLUMETRIC PRODUCTIVITY OF ARABITOL

6.1. Introduction

Arabitol, an enantiomer of xylitol, is a polyhydric alcohol known to be produced by species such as Saccharomyces cerevisiae, Debaryomyces hansenii, Pichia farinosa,

Hansenula anomaloa and Endomycopsis chodati [26, 28, 68, 139, 140] using substrates such as glucose, glycerol, xylose, xylulose and sucrose [43, 71, 141]. Glucose was the most commonly used substrate for arabitol production so far. When glycerol was used as a substrate using a selected D. hansenii strain, a total arabitol yield of 55% with 0.3 g/L-h volumetric productivity was achieved [147]. And the optimal conditions for this yeast were at a temperature of 30 °C, pH less than 4 and 5% dissolved oxygen concentration

(DO). Under these optimal conditions, yeast cells doubling time was about 7 h for initial

0-50 h with a specific productivity of 0.1 h-1 (or g arabitol per g cell dry weight per h).

Doubling times were lower of around 6 h at higher pH (5-6) and slightly more of around

6.7 h at 10-20% DO. However, arabitol production at these conditions of high pH and DO was lower than at pH 3 and 5% DO.

In this study attempts were made to modify the medium composition to improve the specific productivity and the volumetric productivity. Yeast extract and peptone are major nitrogen sources used for arabitol production with various yeast species [43, 71,

149, 150]. The choice of nitrogen source could have a greater effect on yield of cell growth and hence the arabitol production. In this study, the type of N source affected how pH changed along with cell growth (before the control pH value was reached and the pH control kicked in), resulting in different maximum cell and arabitol concentrations attained. Earlier studies with D. hansenii using glycerol as carbon (C) source showed that pH was an important factor and that low pH (< 4) was favorable for arabitol production

(from the results shown in chapter 5). A combination of organic and inorganic nitrogen source was studied to determine the optimal ratio for better cell growth and arabitol production. Another important parameter in the optimization process is inoculum size.

Low inoculum volume usually results in reduction of cell growth rate leading to long lag phase and high inoculum volume are also not necessary. Arabitol production could be influenced with the inoculum size and was found to be increasing with increase in inoculum size in case of Saccharomyces rouxii with glucose as carbon source [71]. In general, optimal inoculum size used for arabitol production in many studies was about

5%. Optimal inoculum size sufficient to reduce the lag phase was determined in this study [150].

Magnesium and phosphate are among the potential nutrients required in the medium as they are essential for many biological functions and enzyme reactions

(Dehydrogenases, kinases, phoshorylases, etc.,) [151]. Arabitol production was stimulated by low phosphate ion concentration for Saccharomyces rouxii using glucose

[152]. High inorganic phosphate concentration results in reduction in the activity of acid phosphatase necessary for the conversion of ribulose-5 phosphate to ribulose which are the precursors required for the production of arabitol [152]. The minimum concentration of magnesium and phosphate required was determined above which there was no effect on the cell growth or arabitol production.

In this report, the effects of inoculum size, magnesium concentration, and type of nitrogen (N) source on D.hansenii growth rate and arabitol production are presented. In addition, the effect of reducing pH by acid addition to the culture medium when yeast extract and peptone were used as nitrogen source was studied as the pH of the medium was >5 with these nitrogen sources.

6.2. Materials and Methods

This section presents the microorganism used for arabitol production, medium composition used for different studies that were mentioned in this chapter. The analytical methods that were used for measuring cell and sugar concentrations were also presented.

6.2.1 Microorganism and medium

The yeast culture Debaryomyces hansenii (NRRL Y-7483) was obtained from

USDA ARS culture collection (Peoria, IL). Culture was maintained on agar plates containing 10 g/L glucose, 3 g/L yeast extract, 3 g/L malt extract, 5 g/L peptone, and 20 g/L agar. The medium used for inoculum preparation had the following composition (per liter): yeast extract, 6 g; (NH4)2SO4, 4 g; K2HPO4, 0.316 g; KH2PO4, 0.211 g;

MgSO47H2O, 1 g; and glycerol, 100 g. The medium had a nitrogen-to-phosphorus (N/P) ratio of 9 and had been found better for arabitol production (than other tested media with

N/P ratios of 0.6, 0.9 and 30) in an earlier study using fermentors with pH and DO controls (results from Chapter 5). This medium was used as the control for comparison in this study. Various media with compositions modified to have different types and/or concentrations of N source, P source and MgSO47H2O were evaluated in this study for cell growth and arabitol production profiles. More details are given in the following subsections. All the media had initial pH around 6.7±0.2 after autoclaving. Peptone, yeast extract, malt extract and MgSO4∙7H2O were purchased from Sigma-Aldrich. Agar, glucose, ammonium sulfate and glycerol were purchased from Fisher Scientific. K2HPO4 and KH2PO4 were purchased from Merck and EMD Chemicals, respectively.

6.2.2 Inoculum preparation

Inoculum for culture study was prepared by transferring a loop of cells from an agar plate to 50 ml culture medium in a 250-ml Erlenmeyer flask covered with cotton wrapped in cheese cloth. The culture was grown at room temperature (22 ± 1 C) for 24 h under vigorous magnetic stirring. The grown culture was inoculated at 5% of the final broth volume in the subsequent studies. A study had been first done to compare the cell growth at inoculum sizes of 5%, 10%, 15%, 20% and 25%. 5% inoculum was found adequate without causing a long lag phase (data not shown).

6.2.3 Experimental studies

All studies were done in multiple 250-ml shake flasks with 50-ml working volume. Cultures were grown at 30°C under 250 rpm shaking in a Queue Orbital Shaker

(Queue systems, Parkersburg, West Virginia). DO and pH were not controlled. Samples taken periodically were measured for pH to follow the trends of pH change.

6.2.3.1 Effect of N source

Two groups of studies were done. In the first group, seven combinations of N sources were evaluated. One was the control medium given above. The N source combinations used in the other six media were (1) yeast extract only (YE), (2) peptone only (Pep), (3) ammonium sulfate only (AS), (4) yeast extract and peptone (YE + Pep),

(5) yeast extract and ammonium sulfate (YE + AS), and (6) ammonium sulfate and peptone (AS + Pep). Total N concentrations in the seven media were made approximately the same at 1.5 g/L by adjusting the concentration of each N-source component, based on the following N contents: 10.9% in YE, 12.9% in Pep and 21.2% in AS according to the bionutrients manual [152] for YE and Pep. The total P concentrations in these media could similarly be estimated with the following P contents: 1.1%, 0.45%, 17.8% and

22.8% in YE, Pep, K2HPO4 and KH2PO4, respectively. The total P concentrations in these media varied from 0.09 g/L to 0.24 g/L. The actual amount of each N-source component used in these seven media are summarized in Table 6.1, together with the resultant total N concentrations, total P concentrations, ratios of organic-N to NH3-N, and the weight ratios of total N to total P.

The second group of studies was done to further investigate the effects of the ratio between organic-N and NH3-N on the culture behaviors. The different systems used and their respective medium compositions are given in Table 6.2. In this group of studies, AS was added after 24 h for two purposes: first, to allow faster initial cell growth with organic N source alone; second, to observe how the addition of AS to media at different

89 organic-N to NH3-N ratios would affect the culture pH change, cell growth, and arabitol production.

Chemical YE + YE + AS +

(g/L) System YE AS Pep AS Pep Pep Control

Glycerol 120 120 120 120 120 120 120

Yeast extract 13 – – 6.9 6.9 – 6

(NH4)2SO4 – 7.1 – 3.5 – 3.5 4

Peptone – – 12 – 6 6 –

K2HPO4 0.32 0.32 0.32 0.32 0.32 0.32 0.32

KH2PO4 0.21 0.21 0.21 0.21 0.21 0.21 0.21

MgSO4∙7H2O 1 1 1 1 1 1 1

Total N 1.42 1.5 1.55 1.5 1.53 1.52 1.5

Total P 0.24 0.09 0.15 0.17 0.19 0.12 0.16

N/P 5.8 16 10.5 9 7.9 12.6 9.5

Table 6.1 Media used in shake flask study for effects of different N source combinations

System YE Pep AS K2HPO4 KH2PO4 MgSO4.7H2O

(YE+Pep)/AS=0.4 1.94 1.67 5.05 0.428 0.285 1

(YE+Pep)/AS=0.774 2.95 2.57 3.99 0.382 0.254 1

(YE+Pep)/AS=1.572 4.14 3.6 2.75 0.328 0.219 1

(YE+Pep)/AS=3 5.08 4.42 1.77 0.285 0.19 1

(YE+Pep)/AS=6 5.81 5.06 1.01 0.253 0.168 1

(YE+Pep)/AS=10 6.46 5.36 0.64 0.227 0.151 1

(YE+Pep)/AS=15 6.16 5.53 0.44 0.234 0.156 1

(YE+Pep)/AS=20 6.36 5.63 0.33 0.226 0.151 1

(YE+Pep)/AS=40 6.61 5.76 0.17 0.216 0.144 1

(YE+Pep) 6.89 6 0 0.316 0.211 1

Table 6.2 Sources of N and P for media used in the study of effects of ammonium sulfate addition on pH change and arabitol production.

Note: All media contained 120 g/L glycerol and 1 g/L MgSO4∙7H2O, in addition to the N and P sources given in the table and (YE+Pep)/AS ratio was the ratio of N content in (YE+P) to

AS. Total N and P in all the systems were 1.5 and 0.2 g/L respectively.

Periodical samples were taken for measurements of pH, optical density at 610 nm (for cell concentration), and concentrations of glycerol and arabitol.

6.2.3.2 Effect of magnesium concentration

To make sure that the media used in this study were not limited by the 1 g/L

MgSO4∙7H2O concentration included, a study was done to compare the cell growth in 91 both the YE + AS and YE + Pep media (described above and in Table 9.1) containing different (1.0, 1.5 and 2 g/L) concentrations of MgSO4∙7H2O but the same concentrations for all other components.

6.2.3.3 Effects of P concentration

The effect of different phosphorus concentration on cell growth and arabitol production was conducted by varying the amount of K2HPO4 and KH2PO4. Total nitrogen in the medium was kept constant (Nitrogen source was from yeast extract and peptone) in all the systems that were considered for this study. Systems with phosphorus concentration of 0.1-0.25 g/L thereby have N/P of 15.3-6.1 were considered for this study. These systems were compared to a control system with yeast extract and peptone as nitrogen source and 0.2 g/L phosphorus concentration.

6.2.3.4 Effect of decreasing the fermentation broth pH after 24 h of fermentation in improving arabitol production

From the studies of effect of nitrogen source it was concluded that YE+Pep medium had better cell growth compared to other systems, but poor arabitol production.

Such poor arabitol production could be attributed to pH being more than 4 (Optimal pH was 3.5 from earlier studies). This study was done to observe the effect of reducing the pH from 5 to 3.5 around 24 h of fermentation with YE+Pep medium.

YE+Pep medium composition as listed in Table 6.1. was used in 2L fermentor with 1L working volume. Pre-culture for this fermentation run had same medium composition as YE+Pep and was cultured in 250 ml Erlenmeyer flask with 100 ml culture volume. Pre-culture was grown at 30°C under 250 rpm shaking in a Queue

Orbital Shaker (Queue systems, Parkersburg, West Virginia). After 48 h, 5% of inoculum was transferred into 2L fermentor with 1L culture volume maintained at 30 °C. Dissolved oxygen concentration (DO) was allowed to drop from ~ 100% (initial) to 10% and was then maintained at 10% till 60 h of fermentation 5% thereafter by varying agitation speed. Around 24 h, pH was reduced from 5 to 3.5 by gradual addition of 1N HCL over a time period of 4 h and was maintained at 3.5 thereafter.

6.2.4 Analytical methods

Cell growth in all the shake flask studies was observed by measuring the optical densities at 610 nm of periodical samples, using a UV/VIS spectrophotometer (Model

UV-1601, Shimadzu Corporation, Columbia, MD). For glycerol and arabitol concentrations, a 5-ml broth sample was centrifuged at 9268 g for 10 min (Sorvall RC 5c,

DuPont, Wilmington, DE) and the supernatant was collected and frozen for future HPLC analyses using a Shimadzu LC 10A system with a refractive index detector (RID-10A). A carbohydrate column (Supelco column H, 250 x 4.6 mm) with a guard column (59319-50 x 4.6 mm) were used at ambient temperature. The mobile phase was 0.1% (v/v) H3PO4 at a flow rate of 0.14 ml/min. The retention times for glycerol and arabitol were 20 and 25 min, respectively. Calibration curves for converting the peak areas to concentrations were established with standard solutions of pure glycerol and arabitol.

6.3. Results and Discussion

This section presents the results of the effect of different nitrogen sources, magnesium and phosphorous concentration and methods of specific growth rate were discussed.

6.3.1 Effect of different N source combinations

Yeast has the ability to grow on both organic and inorganic N sources. The effect of such nitrogenous compounds on the cell growth rate was studied. Cell growth profiles in terms of optical density for systems with different combinations of N sources (shown in Table 6.1) were shown in Figure 6.1(a). The maximum cell concentrations reached were clearly different for the systems with different N sources. Also shown in an inset were the cell growth profiles in the first 24 h in semi-logarithmic plots, for clearer viewing of the associated lag phases. The all organic-N systems, YE and (YE + Pep), had the highest maximum cell concentrations and the shortest lag phases. The all inorganic-N

AS system had very poor growth. The poor growth in the AS system was associated with the very early decrease in pH to values below 3.5 (Figure 6.1(b)) due to the consumption of NH3 as the N source for growth. However, cell growth was also poor for the (AS +

Pep) system where pH did not drop to acidic conditions. In addition the cell growth in the

Pep system was clearly poorer than the other two all organic-N systems, YE and (YE +

Pep). It appears that as an organic-N source for the growth of D. hansenii, peptone was not as effective as yeast extract.

70 YE AS

(a) Pep YE+AS 60 YE+Pep AS+Pep Control 50

40 610

OD 30 20 10 0 0 20 Time94 (h)40 60

8 (b) 7

pH YE 3 AS Pep 2 YE+AS YE+Pep 1 AS+Pep Control 0 0 20 40 60 Time (h)

Figure 6.1. Profiles of (a) cell growth and (b) pH for systems with different N sources

For this set of experiments, the produced arabitol concentrations were measured only at the end, i.e., 50 h. As these media had about the same total N concentration but different total P concentrations, the maximum concentrations of cells and arabitol achieved (at 50 h) in these systems were plotted against their total P concentrations

(Figure 6.2). It was interesting to note that the maximum cell growth appeared to increase with increasing total P concentration. Arabitol production, however, did not increase with the total P concentration. Instead, the two highest arabitol-producing systems, i.e., the control and (YE + AS), were the two showing clear pH decrease to 3.0 – 3.5 values after about 20 h (Figure 6.1(b)). This finding is consistent with the earlier study as mentioned in Chapter 5, showed that the low pH < 4 was more favorable for arabitol production by this yeast. Note that pH of the AS system also decreased to < 3.5 but this occurred too

95 early, before cells could grow to a high concentration to effect an acceptable volumetric productivity of arabitol.

610 40

OD 30

0 9 8

7 6

5 4 3 Arabitol (g/L) Arabitol 2 1 0

Phosphorous (System) g/L

Figure 6.2. Cell and arabitol concentrations of samples taken at 50 h from systems with different N sources, plotted against the total P concentration in the initial medium

Dalhiya and Palnitkar studied the effect of N source on xylitol production by

Petromyces albertensis and Candida shehatae using xylose as a substrate, and found that the organic N source, mainly yeast extract, gave higher xylitol production with 40% yield

[138, 153]. In this study with D. hansenii, better cell growth was obtained with organic N sources, i.e., yeast extract and peptone, and yeast extract was found a better N source than peptone. But these systems with organic N sources did not do well for arabitol production because this yeast strain strongly favored low pH (~ 3-3.5) for arabitol production

6.3.2 Effect of magnesium concentration

The cell growth profiles for two medium systems, i.e., the (YE + Pep) and (YE +

AS) media, containing 1, 1.5 and 2 g/L MgSO4∙7H2O are shown in Figure 6.3. Clearly, at these concentrations, magnesium sulfate was not the limiting nutrient in the media and had no effects on cell growth. Further, it was confirmed that the media with all organic N sources supported better growth than the media with mixed organic and inorganic N sources.

45 1 g/L-(YE+Pep) 40 1.5 g/L-(YE+Pep) 2 g/L-(YE+Pep) 35 1 g/L-(Control) 30 1.5 g/L-(Control) 2 g/L-(Control) 25 610 20 OD 15 10 5 0 0 10 20 30 40 50 Time (h)

Figure 6.3. Cell growth represented as optical density measured at 610 nm for systems with 1, 1.5 and 2 g/L MgSO4∙7H2O in both the (YE + AS) and (YE + Pep) medium systems.

6.3.3 Effect of phosphorous concentration

As shown in Figure 6.2, for the earlier shake flask study with different N source combinations, the maximum cell concentrations attained appeared to increase with the total P concentrations in the media. Also, the dry biomass of yeast has been reported to have an N/P ratio of around 4.4 [154]. To test if the systems were indeed P-limited, another set of shake flask study was made. The media were modified from the (YE +

Pep) medium to have different total P concentrations (from 0.1 to 0.25, corresponding to

N/P ratios of 15.3 to 6.1) by varying the concentrations of K2HPO4 and KH2PO4 used, while all other components were kept the same.

The cell growth profiles represented by the OD610 values are shown in Figure 6.4.

All the systems had the same growth profiles till the end of the study (68 h), with the lowest total P system (0.1 g/L, corresponding N/P = 15.3) being the only possible exception at the last data point. The results indicated that P limitation was not responsible for the different maximum cell concentrations observed in the earlier study with different

N-source combinations. Note also that the resulting pH profiles in all the systems were very similar (data not shown), dropping from 6.7 (initial) to only 5.0 at the end. The pH was not low enough for good arabitol production, reaching only 3-5 g/L in all systems

(data not shown).

60 P - 0.1 g/L P - 0.125 g/L 50 P - 0.15 g/L P - 0.175 g/L P - 0.2 g/L 40 P - 0.225 g/L P - 0.25 g/L 610 30 P - 0.24 (Control) D O 20

0 0 20 Time (h) 40 60

Figure 6.4. Cell growth profiles in media with varying phosphorous concentrations

6.3.4 Arabitol production in fermentor with organic N source and acid addition to lower pH at late growth phase

The results from the study with media of different N source combinations indicated the possibility of improving arabitol productivity by using the medium with

(YE + Pep) as N source (for faster initial cell growth, since this system had doubling time of 4.3 h compared to the control i.e., (YE+AS) with 3.9 h which was calculated at 30 h which falls under the linear portion of cell growth profile as shown in Figure 6.5(b)) and then lower the pH at late growth phase by acid addition (for more favorable arabitol production).

This experiment was done in the fermentor with DO maintained at 10% during the first 60 h (including the growth phase) and 5% thereafter. Since pH was above 5 till

24 h and optimal condition for better arabitol production was at pH 3.5, pH was reduced to 3.5 by gradual addition of 1N HCl over a 4 h period. Surprisingly, arabitol production

(and cell growth) stopped after 30 h as a result of this slow pH adjustment. The profiles of cell growth, arabitol production and pH change for this fermentation are shown in

Figure 6.5. Arabitol production by the yeast appeared to be very sensitive to the culture conditions. Another gentler method of lowering culture pH was needed.

8 (a) 7 6 5

pH 4 3 2 1 0 0 10 20 30 40 50 60 Time (h)

120 25 (b) Glycerol Cell concentration 100 20 Arabitol 80 15 60 10 40 ` Glycerol (g/L) Arabitol / Cell

20 5 concentration (g/L)

0 0 0 20 40 60 80 100 120 140 Time (h)

Figure 6.5(a) pH (b) Cell, glycerol and arabitol concentration profiles for fermentation run with YE and Pep as nitrogen source

100

6.3.5 pH and arabitol production affected by addition of ammonium sulfate at different ratios of NH3-N to organic N

Since lowering pH by slow acid addition interrupted arabitol production in the above fermentor experiment, another set of shake flask studies was made to see if ammonium sulfate could be added at proper ratios to organic N, to trigger natural pH decrease and high arabitol production. Cultures for these shake flask studies were grown originally in the (YE + Pep) medium and then added with ammonium sulfate at 24 h. In this study, a set of systems were designed in a way that the nitrogen was coming from yeast extract, peptone and ammonium sulfate that would result in different pH profiles affecting both cell growth and arabitol production. The medium compositions for these systems are summarized in Table 6.2.

Figure 6.6 pH change profile for systems with different ratios of organic N to NH3-N

The profiles of pH, cell growth and arabitol production for these systems with different organic-N/NH3-N ratios are shown in Figure 6.6. The pH drop caused by the

101 ammonium sulfate addition accorded well with the organic-N/NH3-N ratio: the drops were faster and deeper at lower ratios (Figure 6.6). In addition, the systems high in organic-N/NH3-N ≥3had higher cell growth (Figure 6.7(a)). But systems high in organic-

N/NH3-N ≥6 had shorter doubling times of around 5.5 h which was calculated at initial

24 h period (Figure 6.8(a)). But the arabitol production till the end of the study (~ 75 h) was not high in these systems except for system with organic-N/NH3-N >6 (Figure

6.7(b)), presumably because the pH of these systems did not drop to the low values (< 4) favorable for arabitol production (Figure 6.6). The systems with organic-N/NH3-N = 1.57 and 3 showed higher arabitol production and yield compared to the other systems (Figure

6.8(b)), with the pH dropped to 3.5 at the suitable rates without external pH adjustments

(Figure 6.6). The pH of the system with organic-N/NH3-N = 0.77 dropped too fast

(Figure 6.6); the arabitol production started earlier (the highest among all systems at 33 h) but did not continue afterwards (Figure 6.7(b)) presumably because of the prematurely diminished cell activity as indicated by the halted cell growth (Figure 6.7(a)). The system with the even lower ratio (0.4) of organic N to NH3-N fared even worse.

102

(a)

(b)

Figure 6.7. Profiles of (a) cell growth and (b) arabitol production for systems with different ratios of organic N to NH3-N

103

The results clearly confirmed the importance of pH to arabitol production by the yeast strain. It was also shown that good arabitol production can be achieved by this method of adding optimal concentrations of ammonium sulfate at a late growth phase to media containing yeast extract and peptone as N source, for reducing pH without acid addition.

(a)

6 Doubling time (h) Doubling

5 1

(YE+Pep)

(YE+Pep)/AS=3(YE+Pep)/AS=6 (YE+Pep)/AS=0.4 (YE+Pep)/AS=10(YE+Pep)/AS=15(YE+Pep)/AS=20(YE+Pep)/AS=40 (YE+Pep)/AS+0.774(YE+Pep)/AS=1.572 System

104

as (b) 40

% Total arabitol yield 0

(YE+Pep)

(YE+Pep)/AS=3(YE+Pep)/AS=6 (YE+Pep)/AS=0.4 (YE+Pep)/AS=10(YE+Pep)/AS=15(YE+Pep)/AS=20(YE+Pep)/AS=40 (YE+Pep)/AS+0.774(YE+Pep)/AS=1.572

Figure 6.8. (a) Doubling time of cell growth which was calculated from 0-24 hrs, time period when the cell growth was linear and (b) arabitol yield (measured at 75 h) for systems with different ratios of organic N to NH3-N.

The possible mechanisms for improved arabitol production by the addition of inorganic nitrogen source could be explained as follows. Yeast can assimilate ammonia into glutamate and glutamine by the action of NADPH dependent glutamate dehydrogenase or nitrate reductase. Glutamate and glutamine are important precursors for amino acid synthesis [155]. Hence, NADPH is important in the formation of amino acids when the medium is based on ammonium or nitrate as N source. Reduction of glucose 6- phosphate (enzyme required in the first step of the oxidative pentose phosphate pathway) by glucose 6-phosphate dehydrogenase, a NADP linked enzyme releases NADPH.

Nitrate or ammonium based medium requires NADPH for amino acid synthesis and hence PPP enzymes based on NADPH was more active which helped in xylitol production from xylose [156, 157]. Osmond and Rees [158] studied the activity of

105 several glycolytic and pentose phosphate pathway (PPP) enzymes when Candida utilis was grown on a nitrate based medium (10 g/L KNO3) compared to an amino acid based medium (10 g/L yeast extract). It was observed that the activities of glucose 6-phosphate dehydrogenase and transketolase (PPP enzymes) were 2.5 fold higher in the nitrate based medium than in the amino acid based medium. This suggest that yeasts feeding inorganic nitrogen source such as nitrate and ammonium salts improve arabitol production by increasing the activity of PPP enzymes. Organic nitrogen source such as YE and Pep were found to be good for cell growth in this study. Among the different nitrogen sources tested for better xylitol production from xylose using Petromyces albertensis showed yeast extract best choice as organic nitrogen source with 30.6 g/L xylitol and ammonium acetate as best inorganic nitrogen source with 16.7 g/L in 12 days [138].

6.4. Conclusion

Inoculum size from 5 to 25% had no effect on the cell growth rate probably due to low inoculum concentration. Combination of yeast extract and peptone as nitrogen a source was best in terms of cell growth rate and arabitol production followed by Y.E+A.S combination. These two systems were the same in case of arabitol production. Y.E+A.S combination was further used due to the low price of A.S compared to peptone.

Cell growth rate did not depend on phosphorous concentration when Y.E+P was used as nitrogen source. The corresponding arabitol produced showed that it depends on phosphorous concentration and N/P in the medium. It was observed that N/P ratio of 9 enhances arabitol production and any value below or above these resulted in poor or less than 10 g/L arabitol in 70 hrs. Magnesium concentration up to 2 g/L in the medium did not cause any inhibitory effect on the growth of Debaryomyces hansenii. Yeast extract

106 and peptone are best organic nitrogen sources with ammonium sulfate as inorganic nitrogen source and resulted in good arabitol production and yield of about 40%.

107

CHAPTER VII

METHODS OF IMPROVING GLYCERROL CONSUMPTION AND ARABITOL

PRODUCTION IN FERMENTATION WITH DEBARYOMYCES HANSENII

7.1. Introduction

Arabitol is a polyhydric alcohol known to be produced by osmophilic yeast species such as Candida, Debaryomyces, Pichia, Saccharomyces, Hansenula and

Endomycopsis [29, 68, 140, 159, 160]. Arabitol production is of interest because of its potential use as reducer of dental caries and as a sweetener with low calorie content [13,

38, 161]. Chemically, arabitol could be produced by catalytic hydrogenation of arabinose and reduction of lactones of arabinoic and lyxonic acids [162]. Arabitol could also be produced by microbial fermentation, using glucose, xylose, arabinose, glycerol, sucrose and xylulose as carbon sources [43, 71, 141]. The metabolic routes for arabitol synthesis in the pentose phosphate pathway are also known, typically involving ribulose-5- phosphate being converted to either (1) ribulose and then arabitol, or (2) xylulose-5- phosphate, then xylulose, and then arabitol [70, 71, 163].

108

In earlier studies (refs) more than 200 species and strains were screened and a

Debaryomyces hansenii strain was identified as the best arabitol producer from glycerol that formed arabitol as the only polyol metabolite. The optimal condition for arabitol production was found to be 30°C, 3.5 pH and 5% dissolved oxygen (DO) concentration, in a nitrogen-limited and phosphorus-deficient medium with an N/P ratio of 9. The preferred operation was to start the fermentation at high DO and neutral pH for faster cell growth, allow DO and pH to drop naturally to 5% DO and 3.5 pH by late exponential- growth phase, and then control them at those values for good arabitol production. With

100 g/L of initial glycerol, up to 40 g/L arabitol could be produced in 5 days with a maximum overall yield of 55%. About 30-40 g/L glycerol would however remain unconsumed. A downstream process to purify arabitol from the fermentation broth was also developed but, because glycerol and arabitol are both polyols, the selective extraction to remove glycerol would still remove some arabitol. The purification costs would be significantly lower if glycerol could be more completely converted to arabitol.

One main objective of this work is therefore to investigate methods for achieving more complete consumption of glycerol and higher arabitol production.

The second objective of this work is to investigate the possible way of producing arabitol from sugars in lignocellulose hydrolysate, as a higher-value biorefinery byproduct. Xylitol, a stereoisomer of arabitol, had been shown producible by fermentation using agricultural lignocellulosic residues such as wheat straw, sugarcane bagasse, rice straw, hardwood waste and corn stalks [164-167]. The lignocellulosic materials contain polysaccharides that, by pretreatment and hydrolysis, can be broken down to monosaccharides such as glucose, xylose, arabinose, mannose, and galactose.

109

Most of the lignocellulosic material is cellulose followed by hemicellulose and lignin.

Glucose is the major constituent of cellulose and xylose is the major constituent of hemicellulose [168, 169]. As mentioned glucose was the most used carbon source and xylose is also used for arabitol production. However, xylitol was also produced along with arabitol [170]. The two sugars were examined in this work for different ways to be used for arabitol production by the D. hansenii strain, either as the only substrate or as a cosubstrate with glycerol.

The specific studies made for the two aforementioned objectives are described in

Materials and Methods. Results of the studies are then presented and discussed, including recommendations for further improvement.

7.2. Materials and Methods

The following studies were made: (1) investigation of cell growth and arabitol production in shake flasks with media having different carbon substrates (glucose, xylose, glycerol, and glucose-xylose mixture); (2) comparison of arabitol production from glucose and glycerol, respectively, as the sole substrate, in pH and DO controlled fermentors; (3) determination of the necessary continuous addition rate of glucose (as sole substrate) to support active arabitol production, as an important information for future design of processes coupling lignocellulose hydrolysis and arabitol production; (4) investigation of glucose addition at the stationary phase of fermentation made with glycerol as the starting substrate, in an attempt to achieve more complete glycerol consumption and higher arabitol production; and (5) systematic comparison of the individual or combined effects of pH adjustment, N-source addition, and xylose addition to stationary-phase culture

110 broths pre-grown with glycerol as the starting substrate, for improved glycerol depletion and arabitol production.

7.2.1 Microorganism and media

The yeast culture Debaryomyces hansenii (NRRL Y-7483) was obtained from the

USDA ARS culture collection (Peoria, IL). The basic medium composition used was

(per L): glycerol, 120 g; yeast extract, 6 g; (NH4)2SO4, 4 g; K2HPO4, 0.32 g; KH2PO4,

0.21 g; and MgSO47H2O, 1 g. This medium was selected from an earlier optimization study mentioned in Chapter V. To prepare inoculum for each culture study, a loop of cells was transferred from an agar plate to a 250-ml Erlenmeyer flask covered with cheese-clothed cotton. The flask contained 50 ml of the medium described above. The culture was grown at 30C for 24 h in an orbital shaker (Queue Systems, Parkersburg,

West Virginia). The inoculum thus prepared was added at 5% of the final broth volume in the subsequent culture study.

The media were modified on purpose in some of the studies. For the study of comparing different carbon substrates, glycerol was replaced by glucose (130 g/L), xylose (110 g/L), or a mixture of glucose (110 g/L) and xylose (75 g/L). For the study of slow glucose addition, the initial medium contained (per L) 30 g glucose, 2.1 g yeast extract, 1.4 g (NH4)2SO4, 0.11 g K2HPO4, 0.07 g KH2PO4, and 1 g MgSO47H2O. This medium had about 1/3 of the N and P concentrations in the basic medium, to keep the cell concentration lower to avoid DO limitation in the magnetically stirred flasks. Similarly, to avoid DO limitation in shake flasks, a medium with a lower N concentration was used to prepare a culture in the fermentor, before the broth was divided into multiple shake flasks for studying the potential ways (pH adjustment, N-source addition and xylose

111 addition) of improving glycerol depletion and arabitol production. The lower-N medium used was (per L): yeast extract, 3g; (NH4)2SO4, 2 g; K2HPO4, 0.32 g; KH2PO4, 0.21 g;

MgSO47H2O, 1 g; and glycerol, 120 g. Lastly, the media used for comparing arabitol production in fermentors with glucose and glycerol, respectively, as the carbon substrate had (per L) yeast extract, 3 g; (NH4)2SO4, 2 g; K2HPO4, 2.4 g; KH2PO4, 1.6 g;

MgSO47H2O, 1 g; and glucose or glycerol, 120 g.

7.2.2 Culture conditions and procedures

All the studies, except that with continuous, slow glucose addition, were made at

30°C, the optimal temperature determined for maximum arabitol production [142]. The studies made in fermentors (2.5 L New Brunswick Bioflo 110 with 1 L working volume) were operated by allowing DO and pH to drop naturally with cell growth till 5% DO and pH 3.5 and, thereafter, controlling at these values, by automatic adjustment of agitation speed and base (1 N NaOH) addition (ref). The continuous glucose addition study was made at room temperature, in 250 ml Erlenmeyer flasks with 50 ml culture volume under vigorous magnetic stirring. The studies made in shake flasks were not DO or pH controlled but the periodic samples taken were measured for pH. The shake flask studies were all done in 250 ml Erlenmeyer flasks with 50 ml culture volume, mixed by 250 rpm shaking in the Queue orbital shaker.

The experimental designs and procedures for two of the studies were more complicated and, therefore, described in more details in the following subsections. The two studies are (1) the evaluation of arabitol production from glucose added slowly and

(2) the comparison of pH adjustment, N-source addition, and xylose addition to

112 stationary-phase, glycerol-grown culture as means for stimulating glycerol depletion and further arabitol production.

(a) Slow addition of glucose

This study was intended to extract information on how D. hansenii responds to different rates of glucose addition for arabitol production. The information would be useful in considering the coupling of a stage of lignocellulose hydrolysis to release sugars to support continuous arabitol production. The study was made with 3 simultaneous flasks (250 ml, with 50 ml culture volume) under vigorous magnetic stirring. Cells were first grown in batch mode in the medium with 30 g/L glucose and lower (1/3) N and P concentrations (composition given above), aiming to reach about 7-8 g/L maximum cell concentration. When the glucose concentration dropped to about 5 g/L (at 40-50 h), a 100 g/L concentrated glucose solution was added continuously to the 3 flasks using 3 syringe pumps (KDS 100, KD scientific) to provide glucose at the rates of 0.5, 1.0 and 1.5 g/L-h, respectively, throughout the rest of the study. Periodic samples were taken and analyzed to follow the cell, glucose and arabitol concentrations.

(b) Improving glycerol uptake and arabitol production

The study was to address the problem that after certain period of time, the stationary-phase culture grown on glycerol would stop producing arabitol with 30-40 g/L glycerol remaining. The problem compromises arabitol productivity and complicates the downstream purification of arabitol from the fermentation broth.

Glycerol would be consumed if to support cell growth and/or conversion to arabitol. A preliminary shake flask study showed that the yeast did not grow (or barely) at

113 the low pH 3.5 condition (data not shown). The yeast stopped converting glycerol to arabitol for two possible reasons. (1) The culture had been kept too long at the low-pH stationary phase and had become inactive metabolically. (2) The glycerol consumption and arabitol production in the earlier culture stages had led to low glycerol-to-arabitol ratios, which were no longer favorable for further conversion of glycerol to arabitol.

Conceptually, further glycerol depletion (and conversion to arabitol) might require reinvigoration of the culture by raising the pH and/or adding fresh N-sources to stimulate cell growth again. Also, a cosubstrate such as xylose or arabinose, which, when converted to arabitol, drains the reducing coenzyme NADH [171], might be added to trigger further glycerol consumption, which reduces NAD+ to NADH, while significantly improving arabitol productivity.

Accordingly, the design of this study was to first prepare the stationary culture in a fermentor and then distribute the broth to multiple shake flasks for evaluation the individual or combined effects of pH adjustment, N-source addition, and xylose addition.

To prepare the fermentor culture, cells were grown at 30°C and pH and DO were controlled at 3.5 and 5%, respectively, according to the preferred operation described before. At 160 h the broth was collected and transferred, inside a laminar flow hood, into nine 250 ml Erlenmeyer flasks, each with 50 ml of the broth. The nine conditions evaluated are summarized in Table 7.1. The broths of Systems 1, 2 and 3 were adjusted to pH 5 at the beginning of the study, by addition of a sterilized 2 N NaOH solution.

Systems 4 to 9 were not adjusted for pH and the initial pH of these systems was 3.5, the same as the pH maintained in the fermentor. 30 g/L xylose, 3 g/L yeast extract, and 2 g/L ammonium sulfate were added to Systems 1 and 5. 3 g/L yeast extract and 2 g/L

114 ammonium sulfate were added to Systems 2 and 6. 30 g/L xylose was added to System 3 and 8. 30 g/L xylose and 6.9 g/L yeast extract (with the same N concentration as that provided by 3 g/L yeast extract + 2 g/L ammonium sulfate) were added to System 4. 6.9 g/L yeast extract was added to System 7. System 9 was the control without any medium supplementation or pH adjustment. Note the inclusion of two different types of N-source supplementation (3 g/L yeast extract + 2 g/L ammonium sulfate and 6.9 g/L yeast extract) in the study. The higher organic-N supplementation was expected to cause a longer period of gradual pH increase, associated with NH3 release, which might be additionally effective in reinvigorating the stationary-phase culture.

System Supplementation pH

1 30 g/L Xyl + 3 g/L YE+ 2 g/L AS 5

2 3 g/L YE + 2 g/L AS 5

3 30 g/L Xyl 5

4 30 g/L Xyl + 6.9 g/L YE NA

5 30 g/L Xyl + 3 g/L YE+ 2 g/L AS NA

6 3 g/L YE + 2 g/L AS NA

7 6.9 g/L YE NA

8 30 g/L Xyl NA

9 Control (no supplementation) NA

Note: NA – No adjustment of fermentor broth pH at the beginning of the shake flask study

Table 7.1 Nine systems compared for improving glycerol consumption and arabitol production of a stationary-phase culture pre-grown in the glycerol-based medium –

Systems, with or without pH adjustment to 5, were supplemented with xylose (Xyl), yeast extract (YE), ammonium sulfate (AS), or a certain combination of them.

115

7.2.3 Analytical methods

Cell concentration

Cell growth in the shake flask studies with glucose and xylose as carbon source was recorded by measuring the optical densities at 610 nm of periodical samples, using a

UV/VIS spectrophotometer (Model UV-1601, Shimadzu Corporation, Columbia, MD).

Samples taken from all other studies were measured for intracellular protein concentrations using the Bradford protein assay kit II (Bio-rad Laboratories, Hercules,

CA). The intracellular protein concentrations were then converted to cell dry-weight concentrations using a pre-established calibration curve[142].

Glycerol, glucose, xylose and arabitol analyses

A 5-ml broth sample was centrifuged at 9000 g for 10 min (Sorvall RC 5C,

DuPont, Wilmington, DE). The supernatant was collected and frozen for future analyses of substrate and product (arabitol) co ncentrations by high-performance liquid chromatography (HPLC). A Shimadzu LC 10A system was used with a refractive index detector (RID-10A). Arabitol and glycerol concentrations were measured using a carbohydrate column (Supelco column H, 250 x 4.6 mm) with a guard column (59319-50 x 4.6 mm). The mobile phase was 0.1% (v/v) H3PO4 at a flow rate of 0.14 ml/min. The retention times for glycerol and arabitol were 20 and 25 min, respectively. Glucose and xylose concentrations were measured using a Supelcosil LC-NH2 column (250 x 4.6 mm) with 75:25 acetonitrile:water as the mobile phase at a flow rate of 1.0 ml/min. Glucose and xylose had retention times of 15.2 and 16.5 min, respectively. Calibration curves were established with standard solutions of these compounds.

116

7.3. Results and Discussion

The results from the ability of D. hansenii to use glucose and xylose both for growth and arabitol production, potential use of lignocellulosic biomass for arabitol production and improvement of arabitol production and glycerol consumption were presented and discussed.

7.3.1 Glucose and xylose as carbon source

Four shake flask systems, with glucose, xylose, glucose-xylose mixture, and glycerol as C source, were considered to evaluate the possibility of arabitol production by the D. hansenii strain from hydrolysate of plant biomass. Cell growth was initially faster in glucose system but decreased later due to the decrease in pH to 4 (Figure 7.1(a)) in 20 h whereas in glycerol system, pH dropped to low value only around 35 h. D. hansenii was able to grow on all the systems except the one with xylose as the only carbon source

(Figure 7.1b).

117

Figure 7.1 (a) pH profiles comparison (b) Cell growth profiles comparison followed by measurements of optical density at 610 nm in glucose, glycerol, xylose and

Glucose+Xylose systems

There was also no arabitol production in the system with only xylose as C source.

The concentrations of arabitol produced in 140 h of fermentation were highest and comparable, at 35 g/L, in the two glucose-containing systems (Figure 8.2(a)). The glycerol system produced 24 g/L arabitol. The volumetric productivity in the glucose systems increased from about 0.3 g/L-h to 0.45 g/L-h before dropping back to 0.25-0.30 g/L-h as shown in Figure 7.2(c). (The increase-then-decrease profile was even clearer in the glucose-based fermentation made in the fermentor, as shown later in Figure 7.3(b)).

The productivity in the system with glucose-xylose mixture varied less, in the range of

0.25-0.35 g/L-h. Note also that the productivity was high from very early stages in these systems. On the other hand, arabitol productivity in the glycerol system increased slowly to 0.2 g/L-h after about 50 h. It could be concluded that glucose and xylose, which are

118 usually the major monosaccharide constituents present in plant biomass, could be used for the arabitol production by the D. hansenii strain.

(a)

(b)

119

(c)

Figure 7.2 (a) Arabitol production comparison in glucose, glycerol, xylose and

Glucose+Xylose systems (b) Glucose, glycerol and xylose consumption and (c) volumetric productivity for systems containing different carbon sources such as glucose, xylose and glycerol.

7.3.2 Comparison of glucose and glycerol for arabitol production in pH and DO controlled fermentors

As shown in Figure 7.3(a), the cell growth was similar in the glucose and glycerol systems, with a specific growth rate of 0.17 h-1 during initial 30 h. Volumetric productivity was high in the range of 0.30-0.66 g/L-h, peaking at about 50 h, in the glucose system. Whereas, it was 0.2 g/L.hr during stationary phase in glycerol system.

Volumetric productivity in case of glucose system reduced during stationary phase but glycerol system maintained a constant volumetric productivity throughout the run at 0.2 g/L.hr (Figure 7.3(b)). Though the cell concentration achieved by both the systems was about the same (20-25 g/L), arabitol production (Figure 7.4(b))was more in glucose

120 system (53 g/L at 120 h) than in glycerol system (23 g/L at 120 h) and the total arabitol yield was 54 and 33% in glucose and glycerol system respectively. In glucose as carbon source system, small amounts (3 g/L) glycerol was produced after 24 to 50 h but was later consumed completely by 120 h. In glycerol as carbon source, there was no other product formation apart from arabitol.

a b

Figure 7.3(a) Cell growth comparison (b) Comparison of volumetric productivity in glucose and glycerol fermentors and glycerol production in glucose fermentor.

(a)

121

(b)

Figure 7.4 (a) substrate consumption in glucose and glycerol fermenters and (b) Arabitol production profiles in glucose and glycerol fermentors and glycerol production in glucose fermentor.

Arabitol productivity and yield were 0.4 g/L.hr and 54% respectively when glucose was used as carbon source compared to 0.2 g/L.hr volumetric productivity and

33% arabitol yield using glycerol as carbon source. The possible pathway for the production of arabitol from glucose and glycerol were discussed in earlier studies [172] and it was mentioned that when glycerol was used as carbon source it would be converted to glyceraldehyde-3-phosphate and it should then go up into the gluconeogenesis route and then to pentose phosphate pathway for arabitol production [173].

7.3.3 Slow glucose addition for arabitol production

As mentioned in materials and methods section three systems with 0.5, 1 and 1.5 g/L.hr glucose addition rates were considered once the residual glucose concentration in each system was about 5 g/L. Cell concentrations reached 7 g/L at 24 h and glucose concentrations dropped to 5 g/L at 40-50 h in all the systems. Figure 7.5(a) shows the

122 residual glucose concentration in all these 3 systems at different times of the run and it was observed that in system 2 and 3 which had the continuous glucose additions at the rate of 1 and 1.5 g/L.hr, glucose accumulated over the time as the consumption rate was less than the addition rate and hence the residual glucose concentration was always high.

Whereas, system 1 with continuous glucose addition at the rate of 0.5 g/L.hr had residual glucose concentration always lower than 10 g/L, implying that this rate of glucose addition was low for the cells to consume. The arabitol produced and glucose consumed in grams from these systems over the time was shown in Figure 7.5(b). It was observed from earlier studies that arabitol production was better at higher substrate concentrations especially earlier studies on this species showed better arabitol yield at initial glycerol concentrations more than 50 g/L. Similarly, system 2 and 3 with residual glucose concentrations more than 20 g/L had more arabitol produced with 1.2 and 2.7 g in system

2 and 3 respectively compared to system 1 with only 0.5 g of arabitol. Also, the total arabitol yield was 68%, 28% and 8% for Systems 3, 2 and 1 respectively. Hence, higher bagasse concentration would be needed to be used for releasing high concentrations of sugars (glucose and xylose in major amount) which results in at least 1-1.5 g/L sugar releasing every hour. 6 g/L sugar was released in 24 h with enzyme hydrolysis using 100 g/L bagasse in buffer solution using commercial cellulase enzyme. Rate of sugar release was 0.25 g/L-h and hence the amount of bagasse has to be increased at least 4 times for release of sugars sufficient for continuous arabitol production for cell concentration of about 7 g/L. For higher cell concentration much higher release rate would be required.

123

(a)

(b)

Figure 7.5. (a) Glucose concentration in g/L and (b)Glucose consumed and arabitol produced in gms in systems with 0.5,1 and 1.5 g/L.hr continuous glucose addition rates done 250 ml erlenmeyer flasks.

A cost effective method for producing arabitol could be by using lignocellulosic materials such as plant biomass containing hexoses and pentoses in the form of polymers.

In this study it was clear that glucose was a better substrate for arabitol production among most of the lignocellulosic materials. Hence, pretreated plant biomass could be used for arabitol production using cellulase enzyme for breaking down cellulose and

124 hemicellulose into glucose, xylose and arabinose and the ability of Debaryomyces hansenii to use glucose and xylose as a carbon source showed that arabitol production was good (about 35 g/L produced) than glycerol (25 g/L arabitol produced). There are studies reported on the production of xylitol using lignocellulosic materials such as wheat straw, sugarcane bagasse, hardwood waste, corn stalks and rice straw [164-167] but there were none reported about the arabitol production using lignocellulosic materials.

However, small amounts of arabitol production (~1 g/L) were reported from bagasse used for xylitol production. It was interesting to find in this study that xylose was not metabolized by Debaryomyces hansenii cells but was used for arabitol production. Also, unlike other yeast cells, D.hansenii cells in this study did not produce xylitol when xylose was used along with glucose as opposed to some studies by Zygosaccharomyces rouxii

[71]. Since there was no xylitol detected at all in this study when xylose was used the conversion of xylose to arabitol could be explained by following possible pathway.

Xylose was converted to xylulose using xylose isomerase followed by the conversion of xylulose to arabitol using arabitol dehydrogenase instead of converting xylose to xylitol and then to xylulose followed by conversion to arabitol [71, 174]. . From earlier NCBI

(National Center for Biotechnology Information’s) search studies on the presence relevant nucleotide sequences of different enzymes in this yeast [172] showed that the possibility of the presence of nucleotide sequence of xylose isomerase is 76% only. Also, there was no xylitol production detected in the fermentation broth hence, there was no arabitol production in xylose system. In this study we used 112 g/L glucose and 80 g/L xylose but the total sugar content in most of the lignocellulosic materials is less and as a reference the total sugars in 100 g/L guayule plant biomass is 10 g/L. So, a large amount

125 of pretreated biomass has to be used in order to get high arabitol yield. In this study we evaluated the minimum amount of glucose release rate from enzyme hydrolysis of pretreated biomass for continuous arabitol production and it was determined that at least

1 g/L sugar concentration need to be released into the medium from enzyme hydrolysis of bagasse for continuous and better arabitol production. Any release rate lower than this resulted in almost no arabitol production.

7.3.4 Arabitol production by glucose addition to stationary-phase glycerol-grown culture

From earlier studies with D.hansenii for the production of arabitol using glycerol as carbon source showed that glycerol consumption rate during stationary phase and arabitol production rate also decreases when compared to the late exponential growth phase [142]. To improve the arabitol production and glycerol consumption, glucose was added during late exponential phase so that it could be used for cell growth and maintenance and glycerol for arabitol production. Fermentation was carried with glycerol as the carbon source and with the medium composition as described in materials and methods. Yeast cells grew to about 20 g/L cell concentration in 60 h. Sterilized glucose solution was added to the fermentation broth to make it up to 30 g/L in the medium. From the earlier runs and in this study shows that rate of glycerol consumption reduces after about 55 h. Hence, glucose was added hoping to improve arabitol production from glycerol consumption and glucose for the cell growth and maintenance. As shown in

Figure 7.6, glycerol consumption decreased after the addition of 30 g/L glucose. Glucose addition repressed the consumption of glycerol completely. Glucose was completely consumed in 40 h after its addition to the fermentation broth and this resulted in the production of about 30±5g/L arabitol. Perhaps, 30 g/L glucose concentration might be too

126 high making it the preferable carbon source for the arabitol production and some of the glycerol was diverted to cell growth and maintenance instead.

120 60

Glycerol Glucose Arabitol 100 50

80 40

60 30 Arabitol (g/L) Arabitol Glycerol (g/L) Glycerol 40 20

20 10

0 0 0 20 40 60 80 100 120 Time (h)

Figure 7.6 Glycerol, glucose consumption and arabitol production profile in the fermentation studies done to see the effect of addition of 30 g/L glucose to improve the glycerol uptake rate.

Addition of glucose resulted in almost no consumption of glycerol and only glucose consumption; however arabitol production was improved from this glucose consumption. D.hansenii cells seem to have higher affinity for glucose than glycerol when both the substrates were present in the medium. Similar studies for xylitol production were done using Saccharomyces cerevisiae where xylose was used as a primary substrate and glucose was used as cosubstrate as energy supplier for endogenous metabolism and for cofactor regeneration. This study reported very less xylose consumption (5 g/L) when 50 g/L glucose and 45 g/L xylose were present initially in the medium and there was also ethanol accumulation with this medium [175]. Glucose and

127 xylose competitively inhibit each other for the membrane transportation as they seem to use the same membrane transporter and this membrane transporter has higher affinity to glucose than xylose and this resulted in ethanol formation [176, 177]. This ethanol formation was reduced by adding low amounts of glucose continuously to maintain a concentration of 0.35 g/L glucose all the time during the fermentation run which supports just for the energy purposes and cofactor regeneration. Hence, the glycerol uptake and arabitol production in this study could be improved by slow addition of glucose instead of high amounts of glucose addition.

7.3.5 Effects of pH adjustment, N-source addition and xylose addition on improving glycerol depletion and arabitol production

It was reported from earlier studies with this yeast that when xylose was used as a substrate along with glycerol, arabitol production was better than that compared with only glycerol [172]. Also, in this study we observed that D.hansenii cells were not able to grow on xylose alone, this knowledge was used to add xylose during stationary phase when glycerol consumption rate becomes low to improve the consumption of glycerol for energy purposes required for the conversion of xylose to arabitol. Shown in Figure 7.7 are the profiles of cell growth, glycerol consumption and arabitol production in the fermentation made for preparation of the broth to be used in the subsequent shake flask experiments. Harvested at 160 h the broth had about 15 g/L cells, 90 g/L glycerol, and 14 g/L arabitol, and the arabitol production had slowed down significantly for a while.

After transferring the culture broth into the 9 shake flasks (conditions summarized in Table 7.1), the cell concentration profiles were followed and shown in Figure 7.7.

Cells did not grow in Systems 3, 8 and 9 because of the absence of N source. pH

128 adjustment (to 5) did not have a significant effect on cell growth, as seen from the similar growth profiles between Systems 1 and 5 and between Systems 2 and 6. Presumably this was because, even without initial pH adjustment, the pH in Systems 5 and 6 increased as a result of the NH3 release from the metabolism of added yeast extract (Figure 7.8(a)).

More significant effect was seen with the use of a higher amount of organic N source: cell concentration increased from 13.5 to 22-24 g/L in Systems 4 and 7 with 6.9 g/L yeast extract while it increased only to 17-19 g/L in Systems 1, 2, 5 and 6 with 3 g/L yeast extract and 2 g/L ammonium sulfate. This slightly low growth than other systems could be related to their low pH. Note that because the broth had 90 g/L glycerol at the beginning of the shake flask experiments, the media in all N-supplemented systems were

N-limited.

27.5 25 1 22.5 2 20 3 17.5 15 4 12.5 5 10 6 7.5

Cell concentration (g/L) 7 5 2.5 8 0 9 150 170 190 210 230 Time (h)

Figure 7.7 Cell growth profile for systems with different xylose and nitrogen source concentration in comparison with control with no external xylose or nitrogen source additions after 160 h of fermentation broth.

129

As shown in Figure 7.8(a), pH in all the systems dropped below 3.5 around 190 h except that in system 7, which were still around 3.8. Hence, though cell growth was high in system 5, arabitol production is low in system 7.

Arabitol production profiles are shown in Figure 7.8(b). The top 3 producing

Systems 4, 1 and 5, in decreasing order, all were added with xylose and N source.

Lacking either one resulted in much poorer production in all other systems. System 4 gave a high productivity of 0.5g/L-h, followed by System 1 with 0.4 g/L-h and control had 0.1 g/L h. This volumetric productivity was calculated after the addition of nitrogen and xylose to the separated fermentation broth i.e., from 160 to 225 h. Initial pH adjustment and the use of higher organic N source appeared to have secondary beneficial effects. The pH adjustment effect can be seen from the higher final arabitol concentration

(38 g/L) in the pH adjusted Systems 1 than that in the System 5 (30 g/L) without pH adjustment. The positive effect of a higher organic N source is indicated by the higher production in System 4 (with 6.9 g/L yeast extract) than in System 1 (with 3 g/L yeast extract and 2 g/L ammonium sulfate). It is unclear if the effect of higher organic N source was also only because of the higher and longer pH increase that it caused (Figure 8.8(a)).

130

(a)

50 45 (b) 1 40 2 35 3 30 4 25 5 20

Arabitol (g/L) 6 15 7 10 8 5 9 0 150 170 190 210 230 Time (h)

Figure 7.8 (a) pH profile and (b) Arabitol production profile for systems with different xylose and nitrogen source concentration in comparison with control with no external xylose or nitrogen source additions after 160 h of fermentation broth.

The profiles of glycerol and xylose consumption were not shown, as the figures will be too messy. Instead, the total concentrations of glycerol and xylose consumed at

225 h (the end of the study) are shown in Figure 7.9(a) together with the concentrations of arabitol produced. In addition, the overall yields of arabitol carbon (C) from the total C

131 consumed from xylose and/or glycerol are shown in Figure 7.9(b), and the concentrations of total C consumed and produced are shown in Figure 7.9(c).

For the purpose of stimulating glycerol depletion, the use of a high organic N source without xylose addition was clearly most effective, as shown by the highest (50 g/L) glycerol consumption in System 7 (Figure 7.9(a)). System 7 also had the highest new cell growth (about 10 g/L, Figure 7.9(a)) and produced about 10 g/L additional arabitol. Glycerol was consumed to support both cell growth and arabitol production. The use of the mixed N sources of yeast extract and ammonium sulfate, without xylose addition, in System 6 gave substantially poorer results in both glycerol consumption and arabitol production.

Conversion of xylose to arabitol requires NADH and releases NAD+. Under low aerobic conditions NAD+ levels are low and this pathway of arabitol conversion from xylose releases NAD+ necessary for energy purposes [174]. This hypothesis seemed to work in this study when xylose was added to glycerol containing culture broth after 190 h of fermentation. In this study, 8 g/L arabitol was produced in 60 h with xylose addition after 160 h fermentation whereas the system with no xylose addition produced 3 g/L arabitol. Also, the uptake rate of glycerol was better in systems with xylose addition and nitrogen source addition, as seen in system with 6.9 g/L yeast extract and 30 g/L xylose addition which resulted in 30 g/L of arabitol production in 60 h after 160 h of fermentation. About 30 g/L of glycerol and 30 g/L xylose was consumed leaving only trace amounts of xylose making it the better cosubstrate for improving arabitol production as well as for glycerol consumption. For the system with no xylose or nitrogen source addition the total glycerol consumed was only 10 g/L. This study

132 developed an efficient method to reduce the residual glycerol concentration and improve the arabitol production during the end of the fermentation process making it beneficial for the downstream processing for arabitol recovery and purification.

60 60 Xylose consumed C yield 50 (b) 50 Glycerol consumed (a) Arabitol produced 40 40

30 30 Yield (%)

20 20 Concentration (g/L)

10 10

0 0 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 System System

25 (c) C produced g/L 20 15 C consumed 10 5 0 1 2 3 4 5 6 C produced 7 8 9 System

Figure 7.9 (a) Concentration of total glycerol & xylose consumed and arabitol produced

(b) Total carbon yield from total carbon consumed contributed from glycerol and xylose present in the medium and (c) Concentration of total carbon consumed and produced in all the systems with different xylose and nitrogen source in the medium measured from

160-225 h.

Xylose addition was clearly very beneficial for arabitol production. It gave not only the highest additional arabitol concentrations (30 g/L for System 4 and 23 g/L for

System 1) but also the highest arabitol-C yields (52-53%, Figure 7.9(b)). Also, while the total C consumption in System 7 was similar to those in Systems 1, 4 and 5 133

(Figure7.9(c)), System 7, without xylose, gave much lower arabitol concentration and yield (Figure 7.9(a) and (b)).

Systems 1&2 and systems 5&6 could be compared to see the effect of addition of xylose. As shown in Figure 7.8(b), arabitol production was more in systems 1&5 with 30 g/L xylose in the medium than in systems 2&6 with no xylose in the medium. It was mentioned in the earlier section that xylose was not metabolized by Debaryomyces hansenii cells, so all the xylose that was consumed was used only for arabitol production.

Arabitol was the only polyol that was produced in these systems with the presence of xylose and there was no xylitol production. Also, systems 4&7 could be compared for the effect of xylose where, system 4 resulted in more arabitol (30 g/L) produced compared to system 7 (11 g/L). Total arabitol produced in system 4 was highest compared to rest of the systems and also highest in cell concentration due to the presence of more nitrogen source (6.9 g/L yeast extract).

It is inconclusive whether xylose addition had a negative effect on glycerol consumption. On one hand, glycerol consumption was much lower in System 4 (with xylose and high yeast extract) than in System 7 (same yeast extract but no xylose), suggesting a negative effect of xylose on glycerol consumption. On the other hand, the concentrations of glycerol consumed in Systems 1 and 5 (with xylose and yeast extract- ammonium sulfate mixture) were comparable to that in System 6 (same N-source mixture but no xylose) (Figure 7.9(a)). This uncertainty will have to be clarified in future experiments made in pH and DO controlled fermentors.

134

7.4. Conclusion

Glucose, xylose, glycerol, sucrose and xylose were used as carbon sources for arabitol production [43, 71, 141]. Among these glucose is the most used carbon source by species such as Saccharomyces, Debaryomyces and Candida [43, 141, 178]. It is always economical to use a low price substrate for better process economics. Hence glycerol

(biodiesel byproduct) as a substrate makes a better substrate for arabitol production and could improve the process economics. In this study we compared the kinetic data and arabitol production using glucose and glycerol as carbon source using Debaryomyces hansenii. Volumetric productivity of arabitol with glucose was higher (0.6 g/L-h) than glycerol (0.2 g/L-h) but decreased during stationary phase. Debaryomyces hansenii showed better arabitol production when xylose was used along with glucose in comparison with glycerol as carbon source suggesting the possible use of lignocellulose hydrolysate for arabitol production. Glycerol consumption and arabitol production in the stationary phase was improved by the addition of nitrogen source, xylose along with the adjustment of the pH (low pH favors arabitol production).

135

CHAPTER VIII

PRODUCTION OF SOPHOROLIPIDS USING BIODIESEL BYPRODUCT

GLYCEROL AND VEGETABLE OIL AS CARBON SOURCE

8.1. Introduction

Sophorolipids are extracellular biosurfactants. Sophorolipids molecules consist of dimeric sugar linked with a glycosidic bond to a hydroxyl fatty acid. Sophorolipids can be produced by Candida bombicola [109], Candida apicola [119], Candida bogoriensis

[179], Candida batistae [180] and Wickerhamiella domericqiae [103]. Sophorolipids are of significant interest because of their properties such as antimicrobial activity, low molecular weight, and mostly for their ability to act as surfactant. Sophorolipids are considered as low molecular weight surfactants compared to high molecular weight surfactants such as proteins and lipopolysaccharides and these low molecular weight surfactants have the ability to lower the surface and interfacial tensions [93]. These properties enable sophorolipids to have many potential applications in food [181], cosmetic [97], detergent [182], healthcare [103], textile and agricultural industry [183].

By fermentation processes, sophorolipids can be produced using media containing only a hydrophilic (water soluble) carbon source, only a hydrophobic (water insoluble)

136

carbon source, or a mixture of hydrophobic and hydrophilic carbon sources. The mixed carbon sources generally afford more effective sophorolipids production (refs).

Glucose is the most widely used hydrophilic substrate, followed by sucrose, galactose, lactose, cheese whey, and molasses [184-187]. The latter two were used for reducing the raw material cost. When lactose was used as the only substrate in Torulopsis bombicola fermentation, it could not support cell growth and sophorolipids production unless a suitable hydrophobic substrate (such as sunflower, canola or olive oil) was supplemented

[185, 186]. The hydrophobic substrates or lipid precursors used for sophorolipids production have included oils, fatty acids, their corresponding esters, and alkanes. Oils are widely used as lipid precursors for sophorolipids production and some of these that are tested include soybean oil, linseed oil, grape seed oil, palm oil, coconut oil, olive oil, canola oil, sunflower oil, corn oil, and rapeseed oil [106, 107, 113, 186, 188-192].

The production cost of sophorolipids is currently in the range of $2-3/kg from various sources using different substrates [193, 194]. The relatively high cost is acceptable for low-volume specialty products in cosmetics and health care industry but not competitive enough with synthetic, petroleum-based surfactants for high-volume commodity applications such as oil recovery and industrial cleaning [195]. Cost of the raw materials used in the production is an important factor since it contributes to about

75% of the total production cost. This raw material cost can be reduced by using agro- based wastes such as protein-removed whey and soy molasses [89, 187, 196, 197].

Use of renewable resources is increasing over the years due to the concerns of pollution and importance of sustainable energy sources. Most of the renewable sources are cheaper and more environmentally friendly. Biodiesel is a major liquid biofuel for

137 transportation usage. Biodiesel is generally produced from soybean, sunflower, coconut, palm and rapeseed oil by trans esterification with methanol or ethanol [2, 198]. Over the years biodiesel consumption has been increasing and such increase results in the generation of large amounts of co-products, byproducts or waste. Biodiesel production would be economically and environmentally feasible only if when the disposal of these byproducts and waste could find beneficial uses or be cheaply and safely disposed.

For every 9kg of biodiesel produced from triglycerides, about 1 kg of glycerol is produced as byproduct. By 2009 there was about 450 million gallons of biodiesel produced, resulting in about 50 million gallons of glycerol as byproduct [32]. This glycerol from the biodiesel industry has several impurities such as spent catalyst, salts after neutralization, residual methanol, free fatty acids and methyl esters. Such glycerol does not find many applications compared to pure glycerol and hence has low market value [6]. In our research, this glycerol from the biodiesel industry was used as carbon source for sophorolipids production in this study using the yeast C. bombicola. Using such low priced carbon source would be an alternative to glucose in reducing the total production costs for sophorolipids production and also could improve the economics of biodiesel industry. It was reported that biodiesel co-product stream containing glycerol, free fatty acids (FFA) , free acid methyl esters (FAME) was used for sophorolipids production using Candida bombicola which resulted in production of 60 g/L of sophorolipids in 5 days [197]. A comparison of glucose and biodiesel glycerol was done in this study showing the possible application of biodiesel glycerol as a potential carbon source for sophorolipids production.

138

8.2. Materials and methods

Medium composition, culture conditions and experimental set up details that were used in this chapter were mentioned in this section.

8.2.1 Culture maintenance and inoculum preparation

Yeast Candida bombicola was maintained on agar plates at 4 °C with the following medium composition (g/L): yeast extract, 3; malt extract, 3; peptone, 5; and glucose, 10.

This agar plate was kept in incubator for 24 h at 30 °C before the culture was inoculated for growing pre-culture required for the fermentation.

The medium used for inoculum preparation had the following composition (g/L): glucose, 100; ammonium chloride, 4; yeast extract, 5; MgSO4.H2O, 0.5; and KH2PO4, 1.

10 ml of this medium was sterilized in a 30 ml vial by autoclaving. After the medium was cooled down, one loopful of cells were transferred from the agar plate to the vial. Cells were allowed to grow under vigorous stirring for 24 h on a stirring plate at room temperature. 5 ml of this cell suspension was transferred into a250 ml Erlenmeyer flask containing 50 ml fresh medium. The cells were again grown for 24 h on a magnetic stir plate. The grown culture was inoculated (at about 10% v/v) into a700 ml fermentor with

500 ml fresh medium. The medium composition used in the fermentor was same as the medium used for growing the pre-culture as mentioned above. In addition vegetable oil

(Soybean oil) was added to the fermentation broth at a rate described in later sections.

8.2.2 Fermentation conditions

Initial pH of the medium was around 5.0. The pH was allowed to drop (upon ammonium sulfate consumption as the nitrogen source) and then maintained at 3.6. Two fermentation

139 runs were done with different hydrophilic substrate. One was with glucose and another with biodiesel glycerol (88% pure). Biodiesel glycerol was obtained from Biodiesel systems and the elemental composition of this glycerol was as follows: 800 ppm chloride,

5400 ppm sulfate, <380 ppm nitrite nitrogen, 98 ppm phosphorous, 20 ppm magnesium,

<50 ppm nitrate nitrogen, 35 ppm bromide, 25 ppm fluoride and trace amounts of Cd, Ca,

Ni, Zn, Mn, Cu, Fe and Pb. Glucose or glycerol was added almost continuously to the fermentor as described in the following. 25 g of glucose were added to the fermentor every 24 h. This glucose was added as a powder and was sterilized in dry heat at 105 °C for 2 h. Loss of glucose (~0.35%) during this sterilization part was also taken into account for concentration calculations. For the glycerol-based fermentation, 50 g/L glycerol was included in the initial medium and more glycerol was added continuously using a syringe pump at a rate of 1.2 ml/h for first 24 h and then at 0.3 ml/h for the rest of the run.

Vegetable oil was used as the hydrophobic carbon source or precursor for sophorolipids production. In a study by Rau et al. [199] rapeseed oil was used as the hydrophobic carbon source and was continuously added at a low rate (0.75 g/L-h) during the growth phase and at a high rate (1.75 g/L.hr) during the stationary phase. This addition rate was further adjusted according to the residual rapeseed oil concentration in the medium since accumulation of more than 10 g/L of rapeseed oil showed a decrease in the sophorolipids production. In their study a maximum of 300 g/L sophorolipids was produced with glucose as hydrophilic carbon source and with controlled hydrophobic carbon source addition making sure that there was always about 1 g/L rapeseed oil present in the medium throughout the run. Similar conditions were adopted in our study

140 with controlled vegetable oil addition and glucose and glycerol as the carbon source

[199]. Sterilized vegetable oil was continuously added in both glucose- and glycerol- based fermentations at different rates as shown in Figures 8.1 and 8.2, respectively. The addition rate was altered according to the oil remained in the fermentor system, which was measured at each sampling time by measuring roughly the amount of oil collected in a centrifuge tube present on top of the known amount of sample.

1.2

1 ) r h

/ 0.8 l m (

n o

i 0.6 t i d d a

i 0.4 O

0.2

0 0 20 40 60 80 100 120 140 Time (h)

Figure 8.1 Rate of vegetable oil addition in ml/h for fermentor with glucose as primary carbon source.

141

1.4

1.2 ) r

h 1 / l (m 0.8 n o ti i

d 0.6 d a

l i 0.4 O

0.2

0 0 20 40 60 80 100 120 140 160 Time (h)

Figure 8.2 Rate of vegetable oil addition in ml/h for fermentor with glycerol as primary carbon source.

8.2.3 Analytical methods

Cell concentration

10 ml of culture broth was centrifuged at 8000 rpm and supernatant was stored in freezer at -4°C for further analysis of glycerol concentration. Cell concentration was measured by measuring the protein concentration using intracellular protein analysis, which procedure was discussed in section 3.3.3 for Candida bombicola.

Glycerol analysis

The supernatant collected at different times of fermentation run was used for analyzing glycerol concentration with HPLC using column LC-NH2 by diluting the sample with mobile phase to have the concentration within the range of the standards used. Detailed description of the analysis was described in section 3.3.2.

142

Sophorolipids analysis

5 ml of sample was collected from the fermentation broth and was acidified to pH 2 and extracted with 25 ml of ethyl acetate for 10-12 h. The sample was air dried to remove ethyl acetate and was then dissolved in 0.05 N NaHCO3. Detailed method described in section 3.3.5 was used for sophorolipids concentration measurement.

8.3 Results and discussion

This chapter compares the cell growth and sophorolipids production ability of C. bombicola using glucose and biodiesel glycerol as hydrophilic carbon source and vegetable oil as hydrophobic carbon source. Both systems had the same nitrogen concentration and the medium was nitrogen limited, meaning that the cell concentration reached to a maximum concentration when the nitrogen source in the medium became depleted. As shown in Figure 8.3, the cell growth profiles were similar in both glucose and glycerol systems with an about 25 g/L maximum cell concentration. The total concentrations of remaining ammonia nitrogen in the medium at different times were also shown in Figure 8.3. It was observed that rates of consumption of ammonia nitrogen were similar in both systems. Also, the yeast cells reached their maximum cell concentration as soon as the ammonia nitrogen concentration was consumed completely, confirming that the medium was nitrogen limited. Glycerol was shown to be effective in supporting the growth of C. bombicola without any apparently negative effects.

The ability of C. bombicola to produce sophorolipids using biodiesel glycerol was also evaluated. The concentrations of produced sophorolipids in the glucose and glycerol- based systems are compared in Figure 8.4 for samples collected at different times of

143 fermentation. As described in materials and methods section, vegetable oil was added to the fermentor broth at a rate that would allow the ratio of glycerol to vegetable oil about 3

(w/w) to improve the production of sophorolipids during the stationary phase.

Figure 8.3 Cell growth and ammonia nitrogen consumption profile in glucose and glycerol as hydrophilic substrates and vegetable oil as hydrophobic source for sophorolipids production at different times of fermentation.

Sophorolipids are non-growth associated products hence their production was less during the growth phase and more during the stationary phase [200]. As shown in Figure

8.4, the rate of sophorolipids production was less until stationary phase, which was about

24 h, and was more when cells reached their maximum concentration and were no longer growing. Both glucose and glycerol systems had vegetable oil (contains mainly soybean oil) as the additional carbon (hydrophobic) source. Sophorolipids production was higher in the glucose system than in the glycerol system (Figure 8.4). The maximum

144 sophorolipids concentrations achieved at 120 h were 90 g/L in the glucose system and 70 g/L in the glycerol system.

Figure 8.4 Comparison of sophorolipids production profile for fermentors with glucose and glycerol as hydrophilic carbon source and vegetable oil as hydrophobic carbon source.

The volumetric productivity profiles of sophorolipids in glucose and glycerol fermentors are shown in Figure 8.5. The productivity was also higher with glucose than with glycerol as the hydrophilic carbon source. The volumetric productivity in the glucose-based fermentation peaked at 1.1 g/L-h at 60 h whereas in the glycerol-based fermentation the productivity remained relatively constant once reached about 0.5 g/L-h after 40 h. As shown in Figure 8.6, both glucose and glycerol concentrations were kept relatively high in the two fermentation systems. More specifically, the glycerol concentration accumulated to as high as 150 g/L. It is unknown if the resultant high osmotic stress on the cells had any effects on the lower sophorolipids production in the

145 glycerol-based system as compared to that in the glucose-based fermentation though it did not affect the cell growth.

In a previous study by Ashby et al., a biodiesel co-product stream containing 40% glycerol, 34% hexane-soluble and 26% water was used for sophorolipids production. 60 g/L sophorolipids were produced in 5 days with biodiesel co-product stream. Our study also produced about 60 g/L sophorolipids in 5 days, but the cell concentration in our study was about half of that obtained in Ashby et al., study. This low cell concentration was due to low total nitrogen concentration in our medium. Thus, if we could use higher nitrogen concentration and achieve higher cell concentration, sophorolipids concentration would be much higher. It was reported that most of the sophorolipids that were produced using biodiesel co-product stream were of lactonic forms [197]. Their study compared sophorolipids production with pure glycerol which resulted in sophorolipids which are

99% lactonic form. Thus, the sophorolipids that were produced in our study could be assumed to be of the lactonic form. Lactonic sophorolipids generally find applications in cosmetic industries as moisturizer, antidandruff agent and deodorant and also be used for treating skin diseases [201].

146

1.2 Glucose 1 Glycerol

0.8

0.6

0.4

0.2 Volumetric productivity (g/L.hr) 0 0 20 40 60 80 100 120 140 Time (h)

Figure 8.5 Comparison of volumetric productivity of sophorolipids in fermentors with glucose and glycerol as hydrophilic carbon source and vegetable oil as hydrophobic carbon source.

Figure 8.6 Residual glucose and glycerol concentrations in g/L measured at different times of fermentation. The increase in glucose concentration indicates the batch addition of glucose at that time and increase in the glycerol concentration indicates the continuous addition of glycerol throughout the run.

147

Most of the processes have raw material costs coming to about 75% of the total production cost and the process economics could be improved by choosing an inexpensive and effective raw material for sophorolipids production and hence we analyzed the ability of biodiesel glycerol as the carbon source. Similar attempts for reducing the cost of raw materials was done by using restaurant waste cooking oil, biodiesel co-product stream (BCS-40% glycerol, 34% hexane-solubles and 26% water), corn syrup, whey and sugarcane molasses (62% sugar, 10% non-sugars and 8% inorganic salts) [89, 197, 202]. Sophorolipids production from these various studies was compared in Table 8.1.

The study by Ashby et al. [203] where BCS containing 40% glycerol was used for sophorolipids production reported that 60 g/L sophorolipids was produced in 5 days which was comparable with this study using biodiesel glycerol (88% pure). Their study also showed that using pure glycerol (without a second hydrophobic substrate), only 9 g/L sophorolipids were produced and the cell concentration was low (~9 g/L) compared to BCS (~43 g/L) with same nitrogen concentration. This low cell growth in pure glycerol system was due to the presence of high glycerol concentrations than in BCS system with 40 g/L glycerol i.e., high glycerol concentration deteriorates the cell growth.

In the current study, the 88% pure biodiesel glycerol did not inhibit cell growth and had the same cell concentration as in the system with glucose as the carbon source. Among the other low cost raw materials used for sophorolipids production listed in Table 8.1, biodiesel glycerol stands second in terms of volumetric productivity of sophorolipids. But this volumetric productivity was low compared to other processes that use glucose greater than at least 1 g/L-h.

148

8.4 Conclusions

In summary, this study confirms that biodiesel glycerol does not show any inhibitory effect on the cell growth and it similar to that produced by glucose as substrate.

However, sophorolipids production was less than glucose but better than other processes using low-cost raw materials used for sophorolipids production. Additionally, sophorolipids produced by biodiesel glycerol was comparable with another similar process using biodiesel co-product stream with 40% glycerol. It was possible that the sophorolipids produced by biodiesel glycerol could be lactonic form which has applications in cosmetic industries. Residual glycerol concentrations need to be reduced to improve the sophorolipids recovery process.

149

Hydrophobic Species Hydrophilic lipid pre- Time SL Volumetric Ref Substrate cursor days g/L productivity g/L.hr

Candida Biodiesel Soybean oil 5-6 70 0.5 This bombicola glycerol study (88% pure) Sugarcane Candida molasses Soybean oil 7.5 60 0.33 [202] bombicola Biodiesel co- product stream (40% Candida glycerol, 34% — 5 60 0.5 [203] bombicola hexane- solubles & 26% water) Glucose Restaurant Candida waste oil 10 34 0.14 [204] bombicola

Cryptococcus curvatus and Whey Rapeseed oil NS 422 [89] Candida bombicola * Dairy waste water + Candida Glucose Soy bean oil 8-10 62 0.3 [202] bombicola

Candida Glycerol FAE 7 46 0.27 [205] bombicola

NS- Not specified

FAE-Fatty acid esters prepared by transesterification of soy oil with methanol. *- Cryptococcus curvatus was used for the conversion of whey into biomass with intracellular triglycerides (single-cell oil) and Candida bombicola uses this SCO for growth and SL production.

Table 8.1 Comparison of volumetric productivity and total sophorolipids produced by various cost effective methods using different low cost raw materials about 70 g/L in 5 days.

150

CHAPTER IX

CONCLUSION

9.1 Summary and conclusions

In this research, the following studies were conducted and the conclusions from these studies were presented in this chapter.

1. Screening and selection of strain for the production of arabitol from biodiesel

glycerol.

2. Optimization of process parameters of fermentation and factors affecting the

arabitol production.

3. Investigation of potential use of lignocellulose hydrolysate for arabitol

production.

4. Modification of medium composition to improve the specific productivity and

volumetric productivity of arabitol using Debaryomyces hansenii.

5. Investigation of methods to improve the complete consumption of glycerol and

higher arabitol productivity.

6. Comparison of production of sophorolipids using glycerol with that using

glucose.

The conclusions from these studies were as follows:

151

1. There is much research contributed on the production of xylitol from different

carbon sources till now but there was little research done on the production of

arabitol. Hence, there is no inexpensive method for its production. In this research

we addressed the production of arabitol using low market valued biodiesel

glycerol. USDA performed the strain screening for this part of the research and

among 214 strains screened Debaryomyces hansenii (NRRL# Y-7483) produced

arabitol as the only polyol. This species was unique itself for producing arabitol as

the only polyol compared to most of the species used for arabitol production

result in the production of polyols such as mannitol, xylitol and erythritol which

makes the downstream processing difficult for the separation of arabitol from rest

of the polyols. Hence, the downstream processing with this species would be

better compared to other species.

2. For this species, we developed the medium composition and required process

parameters for optimal production of arabitol by studying the factors that affect

the cell growth and product yield. Optimal operating conditions for better arabitol

yield of approximately 55% were with temperature of 30 °C, 3.5 pH and 5% DO

in 5 days. DO higher than 5% favored faster cell growth but arabitol yield was

lower. Similarly, higher pH above 3.5-6 favored in reaching higher cell

concentration but arabitol production decreased dramatically especially at pH 6.

Apart from these operating conditions, other factors that influence the yield of

arabitol were the initial concentration of glycerol and N-to-P ratio in the medium

composition. High initial glycerol concentration (100-150 g/L) resulted in 40-55%

arabitol yield. Low glycerol concentration of 50 g/L yielded only 15% of arabitol.

152

A minimum of 100 g/L glycerol should be used in the medium for achieving

arabitol yields of 50%. High glycerol concentration offers high osmotic stress on

the medium which could favor the accumulation of arabitol in yeast. But, NaCl

could not be used to provide the osmotic stress in the medium since it hindered

the arabitol production. 50 g/L NaCl was tolerable for cell growth above which

cell growth was reduced drastically. Hence, NaCl could not be used for the

production of arabitol using Debaryomyces hansenii.

Different N-to-P ratios studied were 0.6, 0.9, 9 and 30. Cell growth was

not affected much with these different N-to-P ratios in the medium, but arabitol

production varied. Arabitol production was more in N/P =9 followed by N/P=0.9,

N/P=0.6 and N/P=30. The P-deficient medium N/P=9 with lower P-concentration

was more suitable but N/P=30 with even lower P-concentration was not suitable.

Hence, there was a limit of how low the P-concentration could be used in the

medium without reducing the yield.

3. Lignocellulose hydrolysate generally consists of glucose, xylose and trace

amounts of arabinose. Arabitol production was not studied earlier from such

source. This hydrolysate usage for the production of arabitol was studied. The

ability of D.hansenii in consuming glucose alone, xylose alone and a combination

of glucose and xylose was studied. D.hansenii was able to metabolize glucose but

was not able to metabolize xylose. However, xylose consumption was observed

when a combination of glucose and xylose was used. This suggested that xylose

could not be used for cell growth but could be used for arabitol production via

pentose phosphate pathway. Also, when compared to glycerol, glucose and xylose

153

together resulted in higher arabitol production suggesting the possible application

of lignocellulose hydrolysate for arabitol production.

One of the popular methods of producing lignocellulose hydrolysate is by

enzyme hydrolysis using expensive commercial enzymes. This usage of

expensive enzymes could be eliminated by coupling the arabitol production using

D.hansenii with cellulase producing T.reesei for conversion of lignocellulose into

sugars. However, the rate of release of sugars is an important factor as the arabitol

production was affected by the substrate concentration. So we performed the

study of effect of rate of sugar release on arabitol production. Three rate of

glucose additions that were considered are 0.5, 1 and 1.5 g/L-h. A minimum of 1-

1.5 g/L-h sugar has to be released for continuous production of arabitol. Any rate

of sugar release below this resulted in the consumption of arabitol.

4. After optimizing the medium composition with optimal N/P value for improving

arabitol production, other medium components were further optimized such as

magnesium concentration and nitrogen source. Cell growth at different

MgSO4.7H2O concentrations of 1, 1.5 and 2 g/L were studied to see any

inhibitory effect of high magnesium concentration on the growth of yeast cells if

any. Cell growth profile and maximum cell concentration reached by these

systems was all similar. Hence, there was no inhibition by 2 g/L MgSO4.7H2O at

which the concentration of magnesium was more than required by yeast cells to

reach about 24 g/L cell concentration. Nitrogen source played a vital role in the

cell growth rate. It was understood from our study that cell growth rate was high

in case of only organic nitrogen source such as yeast extract or peptone or a

154

combination of both rather than using inorganic nitrogen source indicated by the

doubling time of the cells. Following are the decreasing order of the systems with

different nitrogen source based on their doubling time in h-1: Yeast

extract+Peptone (3.9) > Peptone=Yeast extract (4.26) > Yeast

extract+Ammonium sulfate (4.31) > Ammonium sulfate+Peptone (4.55) >

Ammonium sulfate (9.53). Using only inorganic nitrogen source poorly supported

cell growth whereas, organic nitrogen source proved to be a best growth nutrient.

5. As mentioned in earlier chapters, high initial glycerol concentrations were

required to improve the arabitol yield, but glycerol rate of consumption reduced

during stationary phase and it would take longer run times for complete

consumption of glycerol. This issue was investigated in two ways.

In the first way, 30 g/L glucose was added during the stationary phase

hoping to see glucose consumption for cell maintenance purposes and glycerol for

arabitol production purpose. On the contrary, up on the addition of glucose, yeast

cells consumed almost only glucose and very small amounts of glycerol. Glucose

addition does not help in improving glycerol consumption.

Second way was to add only nitrogen source (6.9 g/L YE) or nitrogen

source and xylose addition with pH adjusted to 5 (3 g/L YE, 2 g/ AS and 30 g/L

Xylose). Both the systems resulted in maximum depletion of glycerol where, it

was used for both cell growth and arabitol production. However, ‘C’ yield from

total ‘C’ consumed was 53% for the latter system compared to only 21% for

former system. This suggests the importance of adding xylose along with the

nitrogen source. Also, adjustment of the pH after adding the nitrogen source was

155

important because there the yield was only around 30% for nitrogen source and

xylose (3 g/L YE, 2 g/L AS and 30 g/L Xylose) addition without pH adjustment.

This was because optimal pH for arabitol production was at lower pH. Such way

of adding nitrogen source and xylose during the stationary phase with the pH

adjustment resulted in improved glycerol consumption and arabitol production

making the downstream processing easy for arabitol recovery.

9.2 Future recommendations

The volumetric productivity of arabitol after optimizing all the conditions was about 0.5 g/L-h. It would take longer run times to reach high arabitol production. Such volumetric productivity was low compared to xylitol with a volumetric productivity of

3.5 g/L-h. Volumetric oxygen transfer coefficient effect on the production of arabitol would be a good point to study about increasing the volumetric productivity, since the arabitol production seemed to be affected by the available oxygen and its role in the activity of arabitol dehydrogenase enzyme (required for the conversion of xylulose to arabitol) through the NADPH levels in the cell. Finding the best kLa required for better cell growth and arabitol production could improve the arabitol productivity and for future application of this process at an industrial scale.

Additionally, the medium composition can be adjusted further to grow the cells to higher maximum cell concentrations. If the specific productivity can be maintained, the higher cell concentrations would translate into higher volumetric productivity. The upper limit of cell concentration employable is likely set by the oxygen supply rate of the fermentor used. The requirement of only a low DO (5%) in the process is encouraging for this approach.

156

Another aspect of interesting future study would be the optimization of sophorolipids production using glycerol. In this research it was found out that sophorolipids production was low compared to glucose as carbon source and no more studies were done regarding the improvements of this process. Also there is no literature reported on the effective application of biodiesel glycerol for sophorolipids production.

Some of the factors that could be optimized are the concentration of glycerol, choice of lipid pre-cursor and the source of nitrogen and its concentration.

157

REFRENCES

1. Ma, F. and M.A. Hanna, Biodiesel production: a review1. Bioresource technology, 1999. 70(1): p. 1-15.

2. Krawczyk, T., Biodiesel-alternative fuel makes inroads but hurdles remain. Inf, 1996. 7(8): p. 801-808.

3. Dasari, M.A., et al., Low-pressure hydrogenolysis of glycerol to propylene glycol. Applied Catalysis A, General, 2005. 281(1-2): p. 225-231.

4. National, Fuel Facts sheets. . Biodiesel Board, 2007.

5. Liu, Y., et al., Critical influence of osmotic pressure on continuous production of glycerol by an osmophilic strain of Candida krusei in a multistage cascade bioreactor. Process biochemistry, 2002. 38(3): p. 427-432.

6. Bournay, L., et al., New heterogeneous process for biodiesel production: A way to improve the quality and the value of the crude glycerin produced by biodiesel plants. Catalysis Today, 2005. 106(1-4): p. 190-192.

7. Demirbas, A., Biodiesel fuels from vegetable oils via catalytic and non-catalytic supercritical alcohol transesterifications and other methods: a survey. Energy Conversion and Management, 2003. 44(13): p. 2093-2109.

8. Haas, M.J., et al., A process model to estimate biodiesel production costs. Bioresource technology, 2006. 97(4): p. 671-678.

9. Huck, J.H.J., et al., Evaluation of pentitol metabolism in mammalian tissues provides new insight into disorders of human sugar metabolism. Molecular genetics and metabolism, 2004. 82(3): p. 231-237.

158

10. Werpy, T. and G. Petersen, Top Value Added Chemicals from Biomass: Volume I- -Results of Screening for Potential Candidates from Sugars and Synthesis Gas. 2004, DOE/GO-102004-1992, National Renewable Energy Lab., Golden, CO (US).

11. Le Tourneau, D., Trehalose and acyclic polyols in sclerotia of Sclerotinia sclerotiorum. Mycologia, 1966. 58(6): p. 934-942.

12. Talja, R.A. and Y.H. Roos, Phase and state transition effects on dielectric, mechanical, and thermal properties of polyols. Thermochimica Acta, 2001. 380(2): p. 109-121.

13. Scangos, G.A. and A.M. Reiner, A unique pattern of toxic synthesis in pentitol catabolism: implications for evolution. Journal of Molecular Evolution, 1979. 12(3): p. 189-195.

14. McCormick, D.B. and O. Touster, Conversion of d-[1-14C] arabitol, l-[1-14C] arabitol, and d-[1-14C] ribitol to liver glycogen in the rat and guinea-pig. Biochimica et Biophysica Acta, 1961. 54(3): p. 598-600.

15. Mitchell, H.L., Sweeteners and sugar alternatives in food technology. 2006: Wiley-Blackwell.

16. Crick, R., Improvements in or relating to sweetening agents for food. 1961, US patent 884,961.

17. Gare, F., The Sweet Miracle of Xylitol: The All Natural Sugar Substitute Approved by the FDA as a Food Additive. 2003: Basic Health Pubns.

18. Suzuki, S., et al., Novel enzymatic method for the production of xylitol from D- arabitol by Gluconobacter oxydans. Bioscience, biotechnology, and biochemistry, 2002. 66(12): p. 2614-2620.

19. Sugiyama, M., et al., Cloning of the xylitol dehydrogenase gene from Gluconobacter oxydans and improved production of xylitol from D-arabitol. Bioscience, biotechnology, and biochemistry, 2003. 67(3): p. 584-591.

159

20. Melaja, A.J. and L. Hamalainen, Process for making xylitol. 1977, Google Patents.

21. Buhner, J. and F.A. Agblevor, Effect of detoxification of dilute-acid corn fiber hydrolysate on xylitol production. Applied biochemistry and biotechnology, 2004. 119(1): p. 13-30.

22. Kastner, J.R., M.A. Eiteman, and S.A. Lee, Effect of redox potential on stationary-phase xylitol fermentations using Candida tropicalis. Applied microbiology and biotechnology, 2003. 63(1): p. 96-100.

23. Kim, J.H., et al., Optimization of fed-batch fermentation for xylitol production by Candida tropicalis. Journal of Industrial Microbiology and Biotechnology, 2002. 29(1): p. 16-19.

24. Leathers, T.D., B.S. Dien, and A.R.S. Usda, Xylitol production from corn fibre hydrolysates by a two-stage fermentation process. 2000.

25. Leathers, T.D. and B.S. Dien, Xylitol production from corn fibre hydrolysates by a two-stage fermentation process. Process Biochemistry, 2000. 35(8): p. 765-769.

26. Nobre, M.F. and M.S. Costa, The accumulation of polyols by the yeast Debaryomyces hansenii in response to water stress. Canadian Journal of Microbiology, 1985. 31(11): p. 1061-1064.

27. Bernard, E.M., et al., Rate of arabinitol production by pathogenic yeast species. Journal of Clinical Microbiology, 1981. 14(2): p. 189-194.

28. Bisping, B., U. Baumann, and R. Simmering, Effect of immobilization on polyol production by Pichia farinosa. Prog. Biotechnol, 1996. 11: p. 395–401.

29. Van Eck, J.H., B.A. Prior, and E.V. Brandt, Accumulation of polyhydroxy alcohols by Hansenula anomala in response to water stress. Journal of General Microbiology, 1989. 135(12): p. 3505.

160

30. Hajny, G.J., D-Arabitol production by Endomycopsis chodati. Applied and Environmental Microbiology, 1964. 12(1): p. 87-92.

31. Tyson, K.S., et al., Biomass oil analysis: research needs and recommendations. 2004: Citeseer.

32. Board, U.S., The Potential Impact of Rising Petrochemical Prices on Soy Use for Industrial Applications. 2010.

33. Izumori, K., et al., Production of D-tagatose from dulcitol by Arthrobacter globiformis. Applied and environmental microbiology, 1984. 48(5): p. 1055.

34. Khan, A.R., et al., Conversion of xylitol to L-xylulose by Alcaligenes sp. 701B- cells. Journal of Fermentation and Bioengineering, 1991. 72(6): p. 488-490.

35. Gupta, A., et al., Gluconobacter oxydans: its biotechnological applications. Journal of molecular microbiology and biotechnology, 2001. 3(3): p. 445-456.

36. Belitz, H.-D., Grosch, Werner, Schieberle, Peter, Food chemistry. 4 ed. 2009, Berlin Heidelberg: Springer-verlag. 1070.

37. Lee, C.K., Structural functions of taste in the sugar series: Taste properties of sugar alcohols and related compounds. Food Chemistry, 1977. 2(2): p. 95-105.

38. Mitchell, H., Sweeteners and Sugar Alternatives in Food Technology. 2006, Oxford, UK: Blackwell Publishing.

39. Makinen, K.K., Xylitol and oral health. Adv. Food Res, 1979. 25: p. 137-158.

40. Ylikahri, R., Metabolic and nutritional aspects of xylitol. Adv. Food Res, 1979. 25: p. 159-180.

41. Lindberg, B., A. Misiorny, and C.A. Wachtmeister, Studies on the chemistry of lichens. IV. Investigation of the low-molecular carbohydrate constituents of different lichens. Acta chem. scand, 1953. 7: p. 591.

161

42. Frerejacque, M., Presence of D-arabitol in Fistulina hepatica. Comptes Rendus Hebdomadaire des Seances de l’Academie des Sciences, 1939. 208: p. 1123– 1124.

43. Moran, J.W. and L.D. Witter, Effect of sugars on D-arabitol production and glucose metabolism in Saccharomyces rouxii. Journal of bacteriology, 1979. 138(3): p. 823.

44. Saha, B.C. and R.J. Bothast, Production of L-arabitol from L-arabinose by Candida entomaea and Pichia guilliermondii. Applied Microbiology and Biotechnology, 1996. 45(3): p. 299-306.

45. Escalante, J., et al., Production of arabitol from glucose by Hansenula polymorpha. Journal of Fermentation and Bioengineering, 1990. 70(4): p. 228- 231.

46. Hattori, K. and T. Suzuki, Microbial Production of D-Arabitol by n-Alkane-grown Candida tropicalis. Agricultural and Biological Chemistry, 1974. 38(10): p. 1875- 1881.

47. Hanssens, L., A. Van Regenmortel, and H. Verachtert, pH-dependent polyol production in Moniliella tomentosa. Applied and Environmental Microbiology, 1972. 24(5): p. 831.

48. Costa, M.S. and D.J. Niederpruem, Arabitol accumulation in Geotrichum candidum. Archives of Microbiology, 1982. 131(3): p. 283-286.

49. Moskowitz, H.R., The sweetness and pleasantness of sugars. The American journal of psychology, 1971. 84(3): p. 387-405.

50. Lindley, M.G., G.G. Birch, and R. Khan, Sweetness of sucrose and xylitol. Structural considerations. Journal of the Science of Food and Agriculture, 1976. 27(2): p. 140-144.

51. Yamaguchi, S., et al., Studies on the taste of some sweet substances. Agricultural and Biological Chemistry, 1970. 34(2): p. 187-197.

162

52. Wallace, R.J., Identification of slowly metabolized sugars and sugar derivatives that could be used to establish new or modified microbial species in the rumen. Current Microbiology, 1989. 19(4): p. 271-274.

53. Alaluusua, S. and O.V. Renkonen, Streptococcus mutans establishment and dental caries experience in children from 2 to 4 years old. European Journal of Oral Sciences, 1983. 91(6): p. 453-457.

54. Crick, R., Improvements in or relating to sweetening agents for food. 1961.

55. Bruggeman, J.P., et al., Biodegradable Xylitol Based Polymers. Advanced Materials, 2008. 20(10): p. 1922-1927.

56. Werpy, T., et al., Top Value Added Chemicals From Biomass. Volume 1-Results of Screening for Potential Candidates From Sugars and Synthesis Gas. 2004.

57. Kirk-Othmer, Encyclopedia of Chemical Technology. Vol. 3rd edn. 1978, New York: John Wiley & Sons.

58. Elseviers, M., et al., Process for the production of arabinitol. 1998, Google Patents.

59. Fabre, L., P. Gallezot, and A. Perrard, Catalytic hydrogenation of arabinonic acid and lactones to arabitol. Journal of Catalysis, 2002. 208(1): p. 247-254.

60. Tamás, M.J. and S. Hohmann, 4 The osmotic stress response of Saccharomyces cerevisiae. [Without Title], 2003: p. 121-200.

61. Van Eck JH, P.B.B.E., The water relations of growth and polyhydroxy alcohol production by ascomycetous yeasts. Journal of general microbiology, 1993. 139: p. 1047–1054.

62. Albertyn J, H.S., Thevelein JM & Prior BA, GDP1, which encodes glycerol-3- phosphate dehydrogenase, is essential for growth under osmotic stress in

163

Saccharomyces cerevisiae, and its expression is regulaed by high osmolarity glycerol response pathway. Molecular cell biology, 1994. 14: p. 4135–4144.

63. Kayingo, G., S.G. Kilian, and B.A. Prior, Conservation and release of osmolytes by yeasts during hypo-osmotic stress. Archives of microbiology, 2001. 177(1): p. 29-35.

64. Hohmann, M.J.T.S., The osmotic stress response of Saccharomyces cerevisiae. Yeast Stress Responses (Hohmann S & Mager WH, eds), 2003: p. 122–199.

65. Adler, A.B.L., Physiology of osmotolerance in fungi. Advanced microbial physiology, 1992. 33: p. 145–212.

66. Onishi, H., Studies on osmophilic yeasts VII: production of polyalcohols by Saccharomyces rouxii in the concentrated media of sodium chloride and sugars, and identification of poly-alcohols produced. Bulletin of the agricultural chemical society of japan, 1960. 24: p. 126-130.

67. McMillan, J.D. and B.L. Boynton, Arabinose utilization by xylose-fermenting yeasts and fungi. Applied biochemistry and biotechnology, 1994. 45(1): p. 569- 584.

68. Bernard, E.M., et al., Rate of arabinitol production by pathogenic yeast species. Journal of Clinical Microbiology, 1981. 14(2): p. 189.

69. ER Blakley, J.F.S., Studies on the formation of Darabitol by osmophilic yeasts. Canadian journal of biochemistry and physiology, 1962. 40: p. 1737– 1748.

70. Ingram, J.M. and W.A. Wood, Enzymatic basis for D-arabitol production by Saccharomyces rouxii. Journal of Bacteriology, 1965. 89(5): p. 1186.

71. Saha, B.C., Y. Sakakibara, and M.A. Cotta, Production of d-arabitol by a newly isolated Zygosaccharomyces rouxii. Journal of Industrial Microbiology and Biotechnology, 2007. 34(7): p. 519-523.

164

72. Wilson, B.L. and R.P. Mortlock, Regulation of D-xylose and D-arabitol catabolism by Aerobacter aerogenes. Journal of Bacteriology, 1973. 113(3): p. 1404.

73. Gancendo, C.G.a.J., Pathways of utilization and production. European journal of biochemistry, 1968. 5(2): p. 165–172.

74. Babel, W. and K.H. Hofmann, The relation between the assimilation of methanol and glycerol in yeasts. Archives of Microbiology, 1982. 132(2): p. 179-184.

75. Lages, F. and C. Lucas, Contribution to the physiological characterization of glycerol active uptake in Saccharomyces cerevisiae. Biochimica et Biophysica Acta (BBA)-Bioenergetics, 1997. 1322(1): p. 8-18.

76. Georgiou, G., S.C. Lin, and M.M. Sharma, Surface–Active Compounds from Microorganisms. Bio/technology, 1992. 10(1): p. 60-65.

77. Guerra-Santos, L.H., O. Käppeli, and A. Fiechter, Dependence of Pseudomonas aeruginosa continous culture biosurfactant production on nutritional and environmental factors. Applied Microbiology and Biotechnology, 1986. 24(6): p. 443-448.

78. Cooper, D.G. and B.G. Goldenberg, Surface-active agents from two Bacillus species. Applied and environmental microbiology, 1987. 53(2): p. 224.

79. Palejwala, S. and J.D. Desai, Production of an extracellular emulsifier by a gram negative bacterium. Biotechnology Letters, 1989. 11(2): p. 115-118.

80. Passeri, A., et al., Marine biosurfactants. IV. Production, characterization and biosynthesis of an anionic glucose lipid from the marine bacterial strain MM1. Applied Microbiology and Biotechnology, 1992. 37(3): p. 281-286.

81. Hommel, R.K. and K. Huse, Regulation of sophorose lipid production by Candida (Torulopsis) apicola. Biotechnology Letters, 1993. 15(8): p. 853-858.

82. Peypoux, F. and G. Michel, Controlled biosynthesis of Val7-and Leu7-surfactins. Applied Microbiology and Biotechnology, 1992. 36(4): p. 515-517.

165

83. Spencer, J.F.T., P.A.J. Gorin, and A.P. Tulloch, Torulopsis bombicola sp. n. Antonie van Leeuwenhoek, 1970. 36(1): p. 129-133.

84. Tulloch, A.P., A. Hill, and J.F.T. Spencer, A new type of macrocyclic lactone from Torulopsis apicola. Chem. Commun.(London), 1967(12): p. 584-586.

85. Chen, J., et al., Production, structure elucidation and anticancer properties of sophorolipid from Wickerhamiella domercqiae. Enzyme and microbial technology, 2006. 39(3): p. 501-506.

86. Jones, D.F., Novel macrocyclic glycolipids from Torulopsis gropengiesseri. J. Chem. Soc. C, 1967(0): p. 479-484.

87. Kosaric, N., Biosurfactants: production, properties, applications. Vol. 48. 1993: CRC.

88. Shah, V., et al., Sophorolipids, microbial glycolipids with anti-human immunodeficiency virus and sperm-immobilizing activities. Antimicrobial agents and chemotherapy, 2005. 49(10): p. 4093.

89. Otto, R.T., et al., Production of sophorolipids from whey. Applied microbiology and biotechnology, 1999. 52(4): p. 495-501.

90. Rau, U., et al., Enzymatic conversion of a sophorolipid into a glucose lipid. Biotechnology Letters, 1999. 21(11): p. 973-977.

91. Hu, Y. and L.K. Ju, Purification of lactonic sophorolipids by crystallization. Journal of biotechnology, 2001. 87(3): p. 263-272.

92. Davila, A.M., et al., Identification and determination of individual sophorolipids in fermentation products by gradient elution high-performance liquid chromatography with evaporative light-scattering detection. Journal of Chromatography A, 1993. 648(1): p. 139-149.

166

93. Rosenberg, E. and E.Z. Ron, High-and low-molecular-mass microbial surfactants. Applied microbiology and biotechnology, 1999. 52(2): p. 154-162.

94. Ashby, R.D., D.K.Y. Solaiman, and T.A. Foglia, Property control of sophorolipids: influence of fatty acid substrate and blending. Biotechnology letters, 2008. 30(6): p. 1093-1100.

95. Inoue, S., Y. Kimura, and M. Kinta, Dehydrating purification process for a fermentation product. 1980, Google Patents.

96. Hillion, G., et al., Use of a sophorolipid to provide free radical formation inhibiting activity or elastase inhibiting activity. 1998, Google Patents.

97. Abe, Y., S. Inoue, and A. Ishida, Cosmetic composition for skin and hair treatment. 1981, Google Patents.

98. Baviere, M., D. Degouy, and J. Lecourtier, Process for washing solid particles comprising a sophoroside solution. 1994, Google Patents.

99. Maingault, M., Utilization of sophorolipids as therapeutically active substances or cosmetic products, in particular for the treatment of the skin. 1999, Google Patents.

100. Furuta, T., K. Igarashi and Y. Hirata., Biodegradable low-foaming detergent compositions containing sophorolipid, in Saraya Co., Ltd., I.P.-J. pp., Editor. 2003: Japan.

101. Napolitano, L.M., Sophorolipids in sepsis: Antiinflammatory or antibacterial?*. Critical care medicine, 2006. 34(1): p. 258.

102. S Lang, E.K., F Wagner, Antimicrobial Effects of Biosurfactants. European journal of lipid science and technology, 1989. 91(9).

103. Chen, J., et al., Sophorolipid produced from the new yeast strain Wickerhamiella domercqiae induces apoptosis in H7402 human liver cancer cells. Applied microbiology and biotechnology, 2006. 72(1): p. 52-59.

167

104. Fu, S.L., et al., Sophorolipids and their derivatives are lethal against human pancreatic cancer cells. Journal of Surgical Research, 2008. 148(1): p. 77-82.

105. Gorin, P.A.J., J.F.T. Spencer, and A.P. Tulloch, Hydroxy fatty acid glycosides of sophorose from Torulopsis magnoliae. Canadian Journal of Chemistry, 1961. 39(4): p. 846-855.

106. Rau, U., C. Manzke, and F. Wagner, Influence of substrate supply on the production of sophorose lipids by Candida bombicola ATCC 22214. Biotechnology Letters, 1996. 18(2): p. 149-154.

107. Kim, S.Y., et al., Effect of soybean oil and glucose on sophorose lipid fermentation by Torulopsis bombicola in continuous culture. Applied microbiology and biotechnology, 1997. 48(1): p. 23-26.

108. Hu, Y. and L.K. Ju, Sophorolipid production from different lipid precursors observed with LC-MS. Enzyme and Microbial Technology, 2001. 29 (10): p. 593- 601.

109. Fiehler, K., A. Albrecht, D. Rasch and U. Rau., Continuous production of sophorolipids by Candida bombicola ATCC 22214. Fett/Lipid, 1997. 99(9): p. 19- 24.

110. Siegmund, L. and F. Wagner, Bioconversion of Oils and Sugars to Glycolipids. Biosurfactants: production, properties, applications, 1993: p. 205.

111. Hommel, R.K., et al., Production of sophorose lipid by Candida (Torulopsis) apicola grown on glucose. Journal of biotechnology, 1994. 33(2): p. 147-155.

112. Tulloch, A.P.S., J. F. T. Gorin, P. A. J., The fermentation of long-chain Compounds by Torulopsis magnoliae: I. Structures of the hydroxy fatty acids obtained by the fermenatation of fatty acids and hydrocarbons. Canadian Journal of Chemistry, 1962. 40(7): p. 1326-1338.

113. Davila, A.M., R. Marchal, and J.P. Vandecasteele, Sophorose lipid production from lipidic precursors: Predictive evaluation of industrial substrates. Journal of Industrial Microbiology & Biotechnology, 1994. 13(4): p. 249-257.

168

114. Spencer, J.F.T., D.M. Spencer, and A.P. Tulloch, Extracellular glycolipids of yeasts. Economic microbiology, 1979. 3: p. 523-540.

115. Guilmanov, V., et al., Oxygen transfer rate and sophorose lipid production by Candida bombicola. Biotechnology and bioengineering, 2002. 77(5): p. 489-494.

116. Lang, S., et al., Sophoroselipids, method for their production and use. 2002, Google Patents.

117. Garcia-Ochoa, F. and J.A. Casas, Unstructured kinetic model for sophorolipid production by Candida bombicola. Enzyme and Microbial Technology, 1999. 25(7): p. 613-621.

118. Göbbert, U., S. Lang, and F. Wagner, Sophorose lipid formation by resting cells of Torulopsis bombicola. Biotechnology Letters, 1984. 6(4): p. 225-230.

119. Stüwer, O., et al., Production of crystalline surface-active glycolipids by a strain of Torulopsis apicola. Journal of Biotechnology, 1987. 6(4): p. 259-269.

120. Davila, A.M., R. Marchal, and J.P. Vandecasteele, Sophorose lipid fermentation with differentiated substrate supply for growth and production phases. Applied Microbiology and Biotechnology, 1997. 47(5): p. 496-501.

121. Bradford, M.M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical biochemistry, 1976. 72(1-2): p. 248-254.

122. Scott Jr, T.A. and E.H. Melvin, Determination of dextran with anthrone. Analytical Chemistry, 1953. 25(11): p. 1656-1661.

123. Krawczyk, T., Biodiesel-alternative fuel makes inroads but hurdles remain. Inform, 1996. 7(8): p. 801-808.

169

124. Nozaki, H., et al., Production of D-Arabitol by Metschnikowia reukaufii AJ14787. Bioscience, biotechnology, and biochemistry, 2003. 67(9): p. 1923-1929.

125. Groleau, D., P. Chevalier, and T. Yuen, Production of polyols and ethanol by the osmophilic yeastZygosaccharomyces rouxii. Biotechnology Letters, 1995. 17(3): p. 315-320.

126. Spencer, J.F., Production of polyhydric alcohols by yeasts. Progress in industrial microbiology, 1968. 7: p. 1.

127. Blakley, E.R. and J.F. Spencer, Studies on the formation of D-arabitol by osmophilic yeasts. Canadian journal of biochemistry and physiology, 1962. 40: p. 1737.

128. Larsson, C., et al., Osmoregulation of the salt-tolerant yeast Debaryomyces hansenii grown in a chemostat at different salinities. Journal of bacteriology, 1990. 172(4): p. 1769-1774.

129. Tokuoka, K., Sugar- and salt-tolerant yeasts. Journal of applied bacteriology, 1993. 74(2): p. 101-110.

130. Norkrans, B., Studies on marine occurring yeasts: growth related to pH, NaCl concentration and temperature. Archives of Microbiology, 1966. 54(4): p. 374- 392.

131. Onishi, H. and Y. Shiromaru, Physiological changes induced by salt stress in a salt-tolerant soy-yeast, Saccharomyces rouxii. FEMS Microbiology Letters, 1984. 25(2-3): p. 175-178.

132. Adler, L. and L. Gustafsson, Polyhydric alcohol production and intracellular amino acid pool in relation to halotolerance of the yeast Debaryomyces hansenii. Archives of Microbiology, 1980. 124(2): p. 123-130.

133. Ingram, J.M. and W.A. Wood, Enzymatic Basis for d-Arabitol Production by Saccharomyces rouxii 1. Journal of Bacteriology, 1965. 89(5): p. 1186-1194.

170

134. Gancedo, C.J.M.G.a.A.S., Pathways of utilization and production. European journal of biochemistry, 1968. 5(2): p. 165-172.

135. Lages, F. and C. Lucas, Physiological characterization of mediated glycerol uptake in Saccharomyces cerevisiae. Biochim. Biophys. Acta, 1997. 1322: p. 8- 18.

136. Assev, S. and G. RÖLla, Further studies on the growth inhibition of Streptococcus mutans OMZ 176 by xylitol. Acta Pathologica Microbiologica Scandinavica Series B: Microbiology, 1986. 94(1 6): p. 97-102.

137. Crick, R., Improvements in or relating to sweetening agents for food, U. patent, Editor. 1961.

138. Dahiya, J.S., Xylitol production by Petromyces albertensis grown on medium containing D-xylose. Canadian journal of microbiology, 1991. 37(1): p. 14-18.

139. Van Eck, J.H., B.A. Prior, and E.V. Brandt, Accumulation of polyhydroxy alcohols by Hansenula anomala in response to water stress. Journal of general microbiology, 1989. 135(12): p. 3505-3513.

140. Hajny, G.J., D-Arabitol production by Endomycopsis chodati. Applied and Environmental Microbiology, 1964. 12(1): p. 87.

141. Gírio, F.M., et al., Polyols production during single and mixed substrate fermentations in Debaryomyces hansenii. Bioresource technology, 2000. 71(3): p. 245-251.

142. Koganti, S., et al., Production of arabitol from glycerol: strain screening and study of factors affecting production yield. Applied Microbiology and Biotechnology, 2011: p. 1-11.

143. Zhu, H.Y., et al., Production of d-arabitol by a newly isolated Kodamaea ohmeri. Bioprocess and biosystems engineering. 33(5): p. 565-571.

171

144. Malcolm Rhodes P, P.F.S., Applied microbial physiology. 1997, Oxford: IRL Press at Oxford University Press. 270.

145. Sarote, S., Screening of Yeasts for Production of Xylitol from D-Xylose. Journal of Fermentation and Bioengineering, 1995. 80(6): p. 565-570.

146. Hallborn, J., et al., The influence of cosubstrate and aeration on xylitol formation by recombinantSaccharomyces cerevisiae expressing theXYL1 gene. Applied Microbiology and Biotechnology, 1994. 42(2): p. 326-333.

147. Koganti, S., et al., Production of arabitol from glycerol: strain screening and study of factors affecting production yield. Applied microbiology and biotechnology: p. 1-11.

148. Kim, J.H., Y.W. Ryu, and J.H. Seo, Analysis and optimization of a two-substrate fermentation for xylitol production using Candida tropicalis. Journal of Industrial Microbiology and Biotechnology, 1999. 22(3): p. 181-186.

149. Nobre, M.F. and M.S. da Costa, Factors favouring the accumulation of arabinitol in the yeast Debaryomyces hansenii. Canadian journal of microbiology, 1985. 31(5): p. 467-471.

150. Zhu, H.Y., et al., Production of d-arabitol by a newly isolated Kodamaea ohmeri. Bioprocess and biosystems engineering, 2010. 33(5): p. 565-571.

151. Birch, N.J., Magnesium and the cell. 1993: Academic press. 289.

152. Spencer, J.F.T. and P. Shu, Polyhydric alcohol production by osmophilic yeasts: effect of oxygen tension and inorganic phosphate concentration. Canadian Journal of Microbiology, 1957. 3(4): p. 559-567.

153. Palnitkar, S. and A. Lachke, Effect of nitrogen sources on oxidoreductive enzymes and ethanol production during D-xylose fermentation by Candida shehatae. Canadian Journal of Microbiology, 1992. 38: p. 158-162.

172

154. Rhodes, P.M. and P.F. Stanbury, Applied microbial physiology: A practical approach. 1997: IRL Press at Oxford University Press. 270.

155. Magasanik, B., Ammonia assimilation by Saccharomyces cerevisiae. Eukaryotic Cell, 2003. 2(5): p. 827-829.

156. Jeffries, T.W., Effects of Nitrate on Fermentation of Xylose and Glucose by Pachysolen Tannophilus. Nature Biotechnology, 1983. 1(6): p. 503-506.

157. Irene Witt, L.H., Regulation of pyruvate decarboxylase (E.C. 4.1.1.1) synthesis by coenzyme induction in Saccharomyces cerevisiae. Biochemical and Biophysical research communications, 1966. 25(3): p. 340-345.

158. Osmond, C.B. and T. Ap Rees, Control of the pentose-phosphate pathway in yeast. Biochimica et Biophysica Acta (BBA)-General Subjects, 1969. 184(1): p. 35-42.

159. Nobre, M.F. and M.S. da Costa, The accumulation of polyols by the yeast Debaryomyces hansenii in response to water stress. Canadian journal of microbiology, 1985. 31(11): p. 1061-1064.

160. Bisping, B., U. Baumann, and R. Simmering, Effects of immobilization on polyol production by Pichia farinosa. Progress in Biotechnology, 1996. 11: p. 395-401.

161. Crick, R., Improvements in or relating to sweetening agents for food. 1961: US.

162. Kirk, O., Encyclopedia of chemical technology. 3rd ed. Vol. 1. 1978, New York: John Wiley & Sons. 754-778.

163. Blakley ER, S.J., Studies on the formation of Darabitol by osmophilic yeasts. Can J Biochem Physiol, 1962. 40: p. 1737–1748. 164. Ko, C.H., et al., Integrated xylitol production by fermentation of hardwood wastes. Journal of Chemical Technology & Biotechnology, 2008. 83(4): p. 534- 540.

173

165. Canilha, L., W. Carvalho, and J.B. Almeida e Silva, Influence of medium composition on xylitol bioproduction from wheat straw hemicellulosic hydrolysate. World Journal of Microbiology and Biotechnology, 2005. 21(6): p. 1087-1093.

166. Rao, R.S. and C. Jyothi, Xylitol production from corn fiber and sugarcane bagasse hydrolysates by Candida tropicalis. Bioresource technology, 2006. 97(15): p. 1974-1978.

167. Roberto, I.C., et al., Evaluation of rice straw hemicellulose hydrolysate in the production of xylitol byCandida guilliermondii. Biotechnology letters, 1994. 16(11): p. 1211-1216.

168. Sun, Y. and J. Cheng, Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresource technology, 2002. 83(1): p. 1-11.

169. Saha, B.C., Hemicellulose bioconversion. Journal of industrial microbiology & biotechnology, 2003. 30(5): p. 279-291.

170. Saha, B.C., Y. Sakakibara, and M.A. Cotta, Production of D-arabitol by a newly isolated Zygosaccharomyces rouxii. Journal of industrial microbiology & biotechnology, 2007. 34(7): p. 519-523.

171. Wong, B., et al., D-arabitol metabolism in Candida albicans: studies of the biosynthetic pathway and the gene that encodes NAD-dependent D-arabitol dehydrogenase. Journal of bacteriology, 1993. 175(19): p. 6314.

172. Koganti, S., et al., Production of arabitol from glycerol: strain screening and study of factors affecting production yield. Applied microbiology and biotechnology, 2011. 90(1): p. 257-267.

173. Gancedo, C., J.M. Gancedo, and A. Sols, Glycerol metabolism in yeasts. European Journal of Biochemistry, 1968. 5(2): p. 165-172.

174. Wilson BL, M.R., Regulation of D-xylose and Darabitol catabolism by Aerobacter aerogenes. Journal of bacteriology, 1973. 113(3): p. 1404–1411.

174

175. Lee, W.J., Y.W. Ryu, and J.H. Seo, Characterization of two-substrate fermentation processes for xylitol production using recombinant Saccharomyces cerevisiae containing xylose reductase gene. Process Biochemistry, 2000. 35(10): p. 1199-1203.

176. Cirillo, V.P., Relationship between sugar structure and competition for the sugar transport system in Bakers' yeast. Journal of Bacteriology, 1968. 95(2): p. 603.

177. Kotyk, A., Properties of the sugar carrier in baker's yeast. Folia microbiologica, 1967. 12(2): p. 121-131.

178. Wong, B., et al., D-Arabitol metabolism in Candida albicans: Construction and analysis of mutants lacking D-arabitol dehydrogenase. Journal of bacteriology, 1995. 177(11): p. 2971.

179. Cutler, A.J. and R.J. Light, Regulation of hydroxydocosanoic acid sophoroside production in Candida bogoriensis by the levels of glucose and yeast extract in the growth medium. Journal of Biological Chemistry, 1979. 254(6): p. 1944.

180. Konishi, M., et al., Production of new types of sophorolipids by Candida batistae. Journal of Oleo Science, 2008. 57(6): p. 359-369.

181. Akari, S. and Y. Akari, Method of modifying quality of wheat flour product. Japanese patent, 1987. 61205449.

182. Lew, H.Y., Low foaming detergent compositions. 1966, Google Patents.

183. Ducreux, J., et al., Composition containing a surface active compound and glycolipids and decontamination process for a porous medium polluted by hydrocarbons. 1997, Google Patents.

184. Klekner, V., N. Kosaric, and Q.H. Zhou, Sophorose lipids produced from sucrose. Biotechnology letters, 1991. 13(5): p. 345-348.

175

185. Zhou, Q.H. and N. Kosaric, Effect of lactose and olive oil on intra-and extracellular lipids of Torulopsis bombicola. Biotechnology Letters, 1993. 15(5): p. 477-482.

186. Zhou, Q.H. and N. Kosaric, Utilization of canola oil and lactose to produce biosurfactant with Candida bombicola. Journal of the American Oil Chemists' Society, 1995. 72(1): p. 67-71.

187. Solaiman, D.K.Y., et al., Production of sophorolipids by Candida bombicola grown on soy molasses as substrate*,**. Biotechnology letters, 2004. 26(15): p. 1241-1245.

188. Cooper, D.G. and D.A. Paddock, Production of a biosurfactant from Torulopsis bombicola. Applied and environmental microbiology, 1984. 47(1): p. 173.

189. Lee, K.H. and J.H. Kim, Distribution of substrates carbon in sophorose lipid production by Torulopsis bombicola. Biotechnology Letters, 1993. 15(3): p. 263- 266.

190. Casas, J.A. and F. García-Ochoa, Sophorolipid production by Candida bombicola: medium composition and culture methods. Journal of bioscience and bioengineering, 1999. 88(5): p. 488-494.

191. Cavalero, D.A. and D.G. Cooper, The effect of medium composition on the structure and physical state of sophorolipids produced by Candida bombicola ATCC 22214. Journal of biotechnology, 2003. 103(1): p. 31-41.

192. Pekin, G., F. Vardar Sukan, and N. Kosaric, Production of sophorolipids from Candida bombicola ATCC 22214 using Turkish corn oil and honey. Engineering in Life Sciences, 2005. 5(4): p. 357-362.

193. Deshpande, M. and L. Daniels, Evaluation of sophorolipid biosurfactant production by Candida bombicola using animal fat. Bioresource technology, 1995. 54(2): p. 143-150.

176

194. Rau, U., et al., Produktion und Charakterisierung von eukaryotischen Glykolipiden. Fachagentur Nachwachsende Rohstoffe Landwirtschaftsverlag (Ed.). Chemische Nutzung heimischer PflanzenÃle, 1998.

195. Kosaric, N., et al., The role of nitrogen in multiorganism strategies for biosurfactant production. Journal of the American Oil Chemists' Society, 1984. 61(11): p. 1735-1743.

196. Felse, P.A., et al., Sophorolipid biosynthesis by Candida bombicola from industrial fatty acid residues. Enzyme and Microbial Technology, 2007. 40(2): p. 316-323.

197. Ashby, R.D., et al., Sophorolipid biosynthesis from a biodiesel co-product stream. Journal of the American Oil Chemists' Society, 2005. 82(9): p. 625-630.

198. Knothe, G., R.O. Dunn, and M.O. Bagby. Biodiesel: The use of vegetable oils and their derivatives as alternative diesel fuels. 1997: ACS Publications.

199. Rau, U., et al., Sophorolipids: a source for novel compounds. Industrial Crops and Products, 2001. 13(2): p. 85-92.

200. Casas, J.A., S. Garcia de Lara, and F. Garcia-Ochoa, Optimization of a synthetic medium for Candida bombicola growth using factorial design of experiments. Enzyme and microbial technology, 1997. 21(3): p. 221-229.

201. Mager, H., R. Roethhsbeger, and F. Wagner, Preparation of sophorose lipid lactones for use in cosmetics, especially as antidandruff and bacteriostatic agents and deodorant. Germany Patent, De, 1987. 3: p. 417.

202. Daverey, A. and K. Pakshirajan, Kinetics of growth and enhanced sophorolipids production by Candida bombicola using a low-cost fermentative medium. Applied biochemistry and biotechnology, 2010. 160(7): p. 2090-2101.

203. Ashby, R.D., et al., Sophorolipid biosynthesis from a biodiesel co-product stream. Journal of the American Oil Chemists' Society, 2005. 82(9): p. 625-630.

177

204. Shah, V., M. Jurjevic, and D. Badia, Utilization of restaurant waste oil as a precursor for sophorolipid production. Biotechnology progress, 2007. 23(2): p. 512-515.

205. Ashby, R.D., D.K.Y. Solaiman, and T.A. Foglia, The use of fatty acid esters to enhance free acid sophorolipid synthesis. Biotechnology letters, 2006. 28(4): p. 253-260.

206. Robinson, C.W. and C.R. Wilke, Simultaneous measurement of interfacial area and mass transfer coefficients for a well—mixed gas dispersion in aqueous electrolyte solutions. AIChE Journal, 1974. 20(2): p. 285-294.

207. De Figueiredo, P.H. and M.M. Lopes, The scale-up of aerated mixing vessels for specified oxygen dissolution rates. Chemical Engineering Science, 1979. 34(11): p. 1333-1338.

208. Tsao, G.T., Interfacial oxygen transfer in fermentation. Biotechnology and Bioengineering, 1977. 19(4): p. 557-559.

209. Bandyopadhyay, B., A.E. Humphrey, and H. Taguchi, Dynamic measurement of the volumetric oxygen transfer coefficient in fermentation systems. Biotechnology and Bioengineering, 1967. 9(4): p. 533-544.

210. Gimenes, M.A.P., et al., Oxygen uptake rate in production of xylitol by Candida guilliermondii with different aeration rates and initial xylose concentrations. Applied biochemistry and biotechnology, 2002. 98(1): p. 1049-1059.

211. Roseiro, J.C., et al., The effects of the oxygen transfer coefficient and substrate concentration on the xylose fermentation by Debaryomyces hansenii. Archives of microbiology, 1991. 156(6): p. 484-490.

178

APPENDIX

MEASUREMENT OF kLa IN THE PRESENCE OF DIFFERENT SUBSTRATES FOR

THE PRODUCTION OF ARABITOL

Introduction

Arabitol synthesis takes place in pentose phosphate pathway via ribulose-5- phosphate a precursor to pentose phosphate pathway. Ribulose-5-phosphate is either converted to ribulose or xylulose-5-phosphate. Ribulose is reduced to arabitol by

NAD(P)H dependent arabitol dehydrogenase and xylulose-5-phosphate is dephosphorylated to xylulose which is further reduced to arabitol by NAD(P)H dependent arabitol dehydrogenase [70, 71]. The catalytic activity of arabitol dehydrogenase depends on the intracellular cofactor (NAD(P)H) concentrations. Arabitol could be converted back to xylulose or ribulose depending on the intracellular NAD+ concentrations as NAD+ is a cofactor for arabitol dehydrogenase for these conversions.

This NAD+ regeneration is influenced by yeast respiratory activity which depends on the dissolved oxygen concentrations in the fermentation medium. Hence, the aeration level is a critical factor in the formation of arabitol.

179

Dissolved oxygen is an important substrate in aerobic fermentations and in most well designed and optimized fermentations. In aerobic processes, oxygen transfer is one of the limiting factor for achieving higher cell concentration and product formation. In such aerobic processes, low solubility of oxygen in aqueous media is a major problem often encountered [206, 207]. The solubility of oxygen in water is 0.244 mmol/L at 25 °C and is usually lower in the culture medium due to the presence of salts and nutrients.

Oxygen transfer is usually limited by the liquid film surrounding the gas bubbles and the rate of oxygen transfer in a fermentation medium is usually represented as below. dCL/dt = kLa(C*-CL) – OUR ------(A.1)

Where dCL/dt is the change in oxygen transfer with time, C* is the saturated dissolved oxygen concentration, CL is the dissolved oxygen concentration in bulk liquid phase and

OUR is the oxygen uptake rate. The oxygen transfer efficiency is evaluated by the volumetric oxygen transfer coefficient kLa which depends on the operating conditions such as aeration rate and agitation speed and on other factors such as the medium composition and aeration type (bubbled aeration or surface aeration) particularly the components that affect the medium viscosity and surface tension [208].

Many studies were reported on the effect of kLa on the production of xylitol. This first part of the chapter deals with the measurement of kLa and OUR when different concentrations of glycerol were used as the substrate for the production of arabitol using

Debaryomyces hansenii. High glycerol concentrations were expected to raise the medium viscosity. These values were of kLa and OUR were obtained using the dynamic or

“gassing-in” measurement method [208, 209].

180

Fermentation conditions

kLa studies kLa and OUR measurement studies were carried out in 600 ml with 400 ml culture volume with the medium composition as mentioned in section 4.2.2 for arabitol production. Culture was maintained at room temperature and pH at 4.0 by the addition of

1N NaOH. Culture had aeration of 1VVM and agitation of 550 rpm. Dissolved oxygen was continuously monitored throughout the run.

kLa studies

Volumetric oxygen transfer rate studies were done at 70 h making sure yeast cells reached stationary phase. Two separate fermentation runs were carried out for measuring kLa at different glycerol and xylitol concentrations. kLa was measured using dynamic or

“gassing-in” method [209] . This method involves a sudden interruption by turning off the air supply to the fermentor when the cells were actively respiring followed by a subsequent resumption of air supply. DO was continuously monitored throughout this process and DO drops during the sudden interruption of air supply as a result of the respiration of cells and increases to the initial value once the air supply was resumed, such DO profile was shown in Figure A.1

The rate of oxygen transfer in the fermentation medium was given as below dCL/dt = kLa(C*-CL)-OUR.

Where dCL/dt is the volumetric oxygen transfer rate, C* is the oxygen concentration in bulk gas, CL is the oxygen concentration in bulk liquid phase and OUR is the oxygen uptake rate. Glycerol and xylitol concentrations that were considered for determining kLa

181 are listed in Table A.1. Two separate fermentation runs were considered for these studies.

As shown in Table A.1, known glycerol amount was added to the fermentor at stationary phase to make 20 g/L residual glycerol concentration in the fermentor. After this glycerol addition kLa experiment was carried out. Once the kLa experiment was finished, known amount of xylitol was added to the same fermentor resulting in 20 g/L glycerol and 20 g/L xylitol. Then, kLa experiment was carried out under these substrate concentrations.

Similarly, after finishing this experiment, known amount of glycerol was added to make the residual substrate concentration as 50 g/L glycerol and 20 g/L xylitol. Similarly, different substrate additions were considered in run 2 as shown in Table A.1.

on i t a r nt e

c Air turned off on c ) n O D ( ge y ox

d Slope = -OUR e v ol s s i D Air turned on Time

Figure A.1 Typical dissolved oxygen profile using dynamic method for determining volumetric oxygen transfer rate.

182

Run 1

Substrate added, k La studies at substrate g/L concentration

+ 20 g/L glycerol 20 g/L glycerol

+ 20 g/L Xylitol 20 g/L glycerol + 20 g/L Xylitol

+ 30 g/L glycerol 50 g/L glycerol + 20 g/L xylitol

+ 50 g/L Xylitol 50 g/L glycerol + 70 g/L Xylitol

+50 g/L glycerol 100 g/L glycerol + 70 g/L Xylitol

Run 2

Substrate added, k La studies at substrate g/L concentration

+ 50 g/L Glycerol 50 g/L Glycerol

+ 50 g/L Xylitol 50 g/L Glycerol + 50 g/L xylitol

Table A.1 Substrate concentrations that were considered for determining the volumetric oxygen transfer rate and oxygen uptake rate. kLa values from these experiments were determined by integration method. Integrating equation A.1 yields the following equation.

 L )( taK  L )( taK  oL max1 eCeCC 

Simplifying the above equation gives,

  CC   L max   ln  L taK  o  CC max 

183

Where C0 is the DO at (t=0) start of aeration and Cmax is the maximum DO during the aeration cycle. Both C0 and Cmax were experimentally obtained from the aeration curve shown in Figure A.1.

Results and discussion

As mentioned in Table A.1, respective substrates were added and dynamic method was conducted. The kLa and OUR values obtained at these conditions were reported in Table A.2. The kLa and OUR values were recorded before the inoculation of the cells into the fermentor with only fresh medium containing 50 g/L glycerol. After the cells reached stationary phase, i.e. once the cells reached their maximum cell concentration after the exhaustion of the limiting nutrient which is nitrogen, glycerol or xylitol were added and the dynamic method was then performed. As shown in Table A.2, kLa value was low with cells present in the cells than in their absence which was due to the respiration of the cells. Addition of glycerol always resulted in the reduction of kLa suggesting the decreased oxygen transfer rate which could be due to the increased viscosity of the broth. In run 2 however addition of xylitol did not change kLa value. So the ratio of kLa of fresh medium with no cells and with cells and different concentrations of glycerol and xylitol were shown in Figure A.2.

184

Run 1

k La, OUR,

-1 Medium sec mg O2/L-sec

No cells 0.041 0.099

20 g/L glycerol 0.031 0.089

20 g/L glycerol + 20 g/L Xylitol 0.038 0.069

50 g/L glycerol + 20 g/L xylitol 0.029 0.075

50 g/L glycerol + 70 g/L Xylitol 0.033 0.064

100 g/L glycerol + 70 g/L Xylitol 0.036 0.065

Run 2

No cells 0.0179 0.0382

50 g/L glycerol 0.0242 0.294

50 g/L Glycerol + 50 g/L Xylitol 0.024 0.295

Table A.2 kLa and OUR values for Run 1 and Run 2 with No cells and in the presence of different concentrations of glycerol and xylitol

Addition of glycerol did not affect OUR much probably due to the acclimation of yeast cells to glycerol during their initial growth period. But, addition of xylitol decreased the OUR significantly. High substrate concentrations were found to be important in the production of arabitol since arabitol production was reported to be increasing with increase in glucose concentration in the medium due to the increased activity of the arabitol dehydrogenase enzyme [43]. It is required that glycerol concentrations have to be high for high arabitol production and hence kLa value should be improved at these high

185 glycerol conditions by increasing agitation or aeration. The arabitol production under these conditions was not measured but it could be compared with the literature studies about kLa values affecting xylitol production. From Figure A.2, addition of glycerol, up to 100 g/L, decreased kLa by up to 40%, presumably due to the increased broth viscosity.

Surprisingly, addition of xylitol, up to 50 g/L, caused increase in kLa by up to 35%. The responsible mechanism is unknown. Addition of xylitol, up to 70 g/L, caused up to 35% decrease in OUR. Addition of glycerol, up to 100 g/L, did not have appreciable effect to

OUR, presumably because the cells were cultivated in a medium with 50 g/L glycerol and were acclimated to the presence of relatively high glycerol concentrations. The observations seemed to suggest that within the studied glycerol concentration range, cell metabolism might not change significantly. On the other hand, cell metabolism changed dramatically with the addition of xylitol. It would be important to know if and how xylitol addition affects the arabitol production.

186

Figure A.2 kLa and OUR values with No cells and in the presence of different concentrations of glycerol and xylitol

-1 Cell growth studies were done to see the effect of different kLa (0.03 sec , 0.012 sec-1, and 0.0042 sec-1) with Candida guilliermondii for xylitol production using xylose.

These kLa values were obtained by varying the aeration rate. Cell growth increased from

-1 -1 -1 4 g/L to 12 g/L in 50 h for 0.03 sec system, whereas for 0.01 sec and 0.042 sec kLa systems there was not much growth observed [210]. High cell growth generally leads to better product production unless the product is based on anaerobic process. Compared to these kLa values, the kLa values obtained with high glycerol concentration in our study with Debaryomyces hansenii were similar of about 0.036 at 100 g/L glycerol and 70 g/L xylitol. So, there is a possibility of arabitol production to be high.

Xylitol production from xylose was increased from 15 g/L to 60 g/L with increase

-1 -1 in kLa value from 0.007 sec to 0.013 sec using Candiada boidinii. But when higher kLa values were adopted such as 0.022 sec-1xylitol yield decreased to about 22% from 44% at

187

0.018 sec-1. There seems to be a relation between the production of xylitol versus aeration rate. Maximum yield was obtained at low kLa but beyond certain value this relationship is inversely related. At higher kLa values which provides increased avialbility of oxygen to yeast cell, might lead to the regeneration of NAD+. Increased regeneration at high oxygen availability leads to the accumulation of NAD+ which oxidises produced xylitol to xylulose and further phosphorylation of xylulose to xylulose-5-phosphate. This could be further reduced using transketolase and produce fructose-6-phosphate and glyceraldehyde-3-phosphate leading into glycolysis. Hence, there will not be any accumulation of xylitol [211]. Similar pathway could be observed in case of arabitol where at high dissolved oxygen and high kLa values arabitol also could be oxidized to xylulose and its further conversion to be used in glycolysis.

Further studies focusing on the analysis of arabitol production at various kLa values should be carried out to improve the production of arabitol and prevent its oxidation and further use as a substrate instead of glycerol.

188