Detoxification of Lignocellulose-Derived Microbial Inhibitory Compounds by Clostridium Beijerinckii NCIMB 8052 During Acetone-Butanol-Ethanol Fermentation

Home , Butyrate kinase, Clostridium aerotolerans, Clostridium beijerinckii, Toxic alcohol

Detoxification of Lignocellulose-derived Microbial Inhibitory Compounds by Clostridium beijerinckii NCIMB 8052 during Acetone-Butanol-Ethanol Fermentation

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Yan Zhang

Graduate Program in Animal Sciences

The Ohio State University

2013

Dissertation Committee:

Thaddeus C. Ezeji, Advisor

Steven C. Loerch

Sandra G. Velleman

Zhongtang Yu

Venkat Gopalan

Copyrighted by

Yan Zhang

2013

Abstract

Pretreatment and hydrolysis of lignocellulosic biomass to fermentable sugars generate a complex mixture of microbial inhibitors such as furan aldehydes (e.g., furfural), which at sublethal concentration in the fermentation medium can be tolerated or detoxified by acetone butanol ethanol (ABE)-producing Clostridium beijerinckii NCIMB

8052. The response of C. beijerinckii to furfural at the molecular level, however, has not been directly studied. Therefore, this study was to elucidate mechanism employed by C. beijerinckii to detoxify lignocellulose-derived microbial inhibitors and use this information to develop inhibitor-tolerant C. beijerinckii.

Towards the long-term goal of developing inhibitor-tolerant Clostridium strains, the first objective was to evaluate ABE fermentation by C. beijerinckii using different proportions of Miscanthus giganteus hydrolysates as carbon source. Compared to the growth of C. beijerinckii in control medium, C. beijerinckii experienced different degrees of inhibition. The degree of inhibition was dose-dependent, and C. beijerinckii did not grow in P2 medium with greater than 25% (v/v) Miscanthus giganteus hydrolysates. To improve tolerance of C. beijerinckii to inhibitors, supplementation of P2 medium with undiluted (100%) Miscanthus giganteus hydrolysates with 4 g/L CaCO3 resulted in successful growth of and ABE production by C. beijerinckii. Spectrophotometric and

HPLC analyses revealed that C. beijerinckii transformed lignocellulose-derived furan

aldehydes such as furfural and hydroxymethylfurfural to furfuryl alcohol and 2, 5-bis- hydroxymethylfuran, respectively and at a rate of 0.15 and 0.08 g/L/h, respectively.

The next study aimed to compare differential gene expressions between C. beijerinckii cultures grown in P2 medium supplemented with and without furfural during acidogenic and solventogenic growth phases. The genomic microarray was used to comprehensively evaluate the inhibitory effects of furfural on C. beijerinckii, and potential adaptation mechanisms to furfural stress. Functional gene group analysis showed that increased expression of genes related to redox balancing may be responsible for the reduction of toxic effects of furfural and alleviation of furfural induced oxidative stress in C. beijerinckii during acidogenic growth phase. However, ABE accumulation, redox balance perturbations, and repression of phosphotransferase system may have caused the termination of the growth of C. beijerinckii following furfural challenge at the solventogenic growth phase.

The last objective was to accelerate biotransformation of furfural to furfuryl alcohol by overexpression of furfural-reducting enzymes in C. beijerinckii. Based on results obtained from the transcriptomic analysis of C. beijerinckii, two candidate genes, aldo/keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR) encoded by

Cbei_3974 and Cbei_3904 respectively, were selected, cloned and expressed in

Escherichia coli to generate polyhistidine-tagged proteins and confirm the role of these enzymes in furfural reduction. Those (His)6-tagged proteins were purified by immobilized metal affinity chromatography. AKR and SDR reduced furfural to furfuryl alcohol using NADPH as cofactor, and they showed catalytic activities over a broad range of temperature, pH, and substrate specificity. Subsequently, AKR and SDR were

iii

overexpressed in C. beijerinckii, which resulted in the development of inhibitor-tolerant strains, C. beijerinckii AKR+ and C. beijerinckii SDR+. This integrated study enhanced our understanding of inhibitory effects of lignocellulose-derived aldehydes, and discussed poteintial strategies to engineer clostridia with high tolerance to lignocellulose hydrolysates.

Dedication

This document is dedicated to my Dearest Mother.

Acknowledgments

I would like to take this opportunity to thank all the people who generously supported my research and helped me accomplish this dissertation.

First of all, I would like to express my deepest gratitude to my advisor, Dr.

Thaddeus Ezeji, for providing me the opportunity to one of my favorite fields, for his careful and thoughtful guidance on each of my experiment, and for all the scientific training he gave me to help me become a scientist.

I would also like to thank all my committee members for discussing on my progress updates. Dr. Sandra Velleman always inspires me with her questions and comments on my presentations and progress reports. Dr. Zhongtang Yu offered me a great chance to work and learn techniques in his laboratory when I took classes in the

Columbus campus, and continued to encourage me after I moved to Wooster. Dr. Steve

Loerch has taught me and warmed me more than he will ever know. I really appreciate every word he talked to me, especially when he pointed out my weaknesses and helped me to improve. Dr. Venkat Gopalan is one of the best teachers I have ever met. He is so knowledgeable and his ideas are always logical. I benefited a lot from his questions and suggestions.

I would like to thank Dr. Bei Han for being a great colleague during my first two years of graduate study and a great friend to date. I also appreciate Dr. Victor Ujor for his help on my molecular cloning experiment. Many thanks to Christopher Abraham,

Angufor Numfor, Catherine Richmond, and Jenny Orozco for the conversations and laughter in the lab which made life not that stressful. I would also like to thank Lingling

Wang, Danni Ye, Shan Wei, Min Seok Kim, Jill Stiverson, Thavamathi Annamalai, and

Revathi Shanmugasundaram for their friendship.

Finally, I would like to thank my family. No words can express how happy and thankful I am as the daughter of my Mom and Dad. I would never have made it this far without your love and support. I am very grateful to have my sister with me to share my happiness and sadness. The most sincere thanks go to my husband, Bin You, the person who knows me best and supports me through good times and bad.

vii

Vita

2007...... B.S. Biotechnology, Shandong Normal

University

2008 to present ...... Graduate Research Associate, Department

of Animal Sciences, The Ohio State

University

Publications

Zhang, Y., Han, B., Ezeji, T.C., 2012. Biotransformation of furfural and 5-

hydroxymethyl furfural (HMF) by Clostridium acetobutylicum ATCC 824 during

butanol fermentation. N. Biotechnol. 29, 345-351.

Zhang, Y., Ezeji, T.C., 2013. Transcriptional analysis of Clostridium beijerinckii NCIMB

8052 to elucidate role of furfural stress during acetone butanol ethanol fermentation.

Biotechnol. Biofuels. In revision.

viii

Fields of Study

Major Field: Animal Sciences

Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments...... vi

Vita ...... viii

Publications ...... viii

Fields of Study ...... ix

List of Tables ...... xviii

List of Figures ...... xxi

Chapter 1: Introduction ...... 1

References ...... 5

Chapter 2: Literature Review ...... 8

2.1 Acetone-Butanol-Ethanol (ABE) fermentation ...... 8

2.1.1 History of ABE fermentation ...... 8

2.1.2 Solventogenic Clostridium sp. and mechanisms of ABE fermentation ..... 10

2.2 Lignocellulosic biomass ...... 13

2.2.1 Sources of Lignocellulosic biomass...... 13

2.2.2 Pretreatment of lignocellulosic biomass ...... 14

2.2.3 Enzymatic hydrolysis of lignocellulosic biomass ...... 16

2.2.4 Lignocellulose-derived microbial inhibitory compounds ...... 17

2.3 Molecular mechanisms of in situ detoxification ...... 22

2.4 Transcriptome analysis using microarray ...... 25

2.5 Conclusions ...... 27

References ...... 29

Chapter 3: Acetone Butanol Ethanol Production from Miscanthus giganteus by

Clostridium beijerinckii NCIMB 8052 ...... 45

3.1 Abstract ...... 45

3.2 Introduction ...... 46

3.3 Materials and Methods ...... 48

3.3.1 Microorganism and culture conditions ...... 48

3.3.2 Pretreatment of Miscanthus giganteus ...... 48

3.3.3 Acetone butanol ethanol production from Miscanthus giganteus

hydrolysates by C. beijerinckii 8052 ...... 49

3.3.4 Analytical methods ...... 49

3.4 Results and Discussion ...... 50

3.4.1 Pretreatment and hydrolysis of Miscanthus giganteus ...... 50

3.4.2 Inhibition on the growth of and ABE production by C. beijerinckii 8052

using Miscanthus giganteus hydrolysates ...... 51

3.4.3 Mitigation of lignocellulose-derived microbial inhibitors by CaCO3...... 53

3.5 Conclusion ...... 55

References ...... 57

Chapter 4: Biotransformation of Lignocellulose-derived Furan Aldehydes and Phenolic

Compounds by Clostridium beijerinckii NCIMB 8052 during Butanol Fermentation .... 65

4.1 Abstract ...... 65

4.2 Introduction ...... 66

4.3 Materials and Methods ...... 68

4.3.1 Microorganism and culture conditions ...... 68

4.3.2 Fermentation in P2 medium containing lignocellulose-derived inhibitors . 68

4.3.3 Analytical methods ...... 69

4.4 Results and Discussion ...... 69

4.4.1 Decrease of furfural, HMF, 4-hydroxybenzaldehyde and p-coumaric acid

by C. beijerinckii 8052 ...... 69

4.4.2 Impact of furfural, HMF, and furfuryl alcohol on C. beijerinckii 8052

growth and ABE fermentation ...... 71

xii

4.4.3 Impact of 4-hydroxybenzaldehyde and p-coumaric acid on C. beijerinckii

8052 growth and ABE fermentation ...... 74

4.5 Conclusion ...... 76

References ...... 77

Chapter 5: Transcriptional Analysis of Clostridium beijerinckii NCIMB 8052 to

Elucidate Role of Furfural Stress During Acetone Butanol Ethanol Fermentation ...... 88

5.1 Abstract ...... 88

5.2 Introduction ...... 89

5.3 Materials and Methods ...... 91

5.3.1 Bacterial strains, culture conditions ...... 91

5.3.2 Furfural-challenged experiments and ABE Fermentation ...... 92

5.3.3 Total RNA purification ...... 94

5.3.4 Comparative microarray hybridization ...... 95

5.3.5 Microarray data analysis ...... 96

5.3.6 Microarray data accession number ...... 98

5.3.7 Real-time quantitative reverse transcription PCR (Q-RT-PCR) ...... 98

5.4 Results ...... 99

5.4.1 Microarray analysis of C. beijerinckii 8052 transcriptome under furfural

stress ...... 99

xiii

5.4.2 Expression of C. beijerinckii 8052 redox and cofactor genes in the presence

of furfural ...... 100

5.4.3 Expression of membrane transporter genes in C. beijerinckii 8052 ...... 102

5.4.4 Expression of a two-component signal transduction system, chemotaxis,

and cell motility genes in C. beijerinckii 8052 ...... 105

5.4.5 Validation of gene expression data from microarray analysis by Q-RT-PCR

...... 107

5.4.6 Interactive effect of furfural reduction and ABE production ...... 107

5.5 Discussion ...... 110

5.5.1 Redox and cofactor genes are crucial for detoxification of furfural by C.

beijerinckii 8052 ...... 111

5.5.2 Membrane transporter genes play active role in furfural tolerance and

detoxification by C. beijerinckii 8052...... 114

5.5.3 Furfural influences the adaptation machinery of C. beijerinckii 8052 ..... 116

5.6 Conclusions ...... 120

References ...... 122

Chapter 6: Purification and Characterization of Aldo/keto Reductase and Short-chain

Dehydrogenase/reductase Catalyzing NADPH-coupled Furfural Reduction in Clostridium beijerinckii NCIMB 8052 ...... 145

6.1 Abstract ...... 145

xiv

6.2 Introduction ...... 146

6.3 Materials and Methods ...... 148

6.3.1 Bacterial strains, plasmids and culture conditions ...... 148

6.3.2 Cloning of AKR and SDR genes ...... 149

6.3.3 Expression of AKR and SDR in Rosetta-gami™ B(DE3)pLysS ...... 151

6.3.4 Purification of AKR and SDR ...... 151

6.3.5 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) ...

...... 152

6.3.6 AKR and SDR assay ...... 153

6.3.7 Characterization of AKR and SDR enzymes ...... 154

6.4 Results and Discussion ...... 154

6.4.1 Construction of recombinant plasmids ...... 154

6.4.2 Expression and purification of AKR and SDR ...... 155

6.4.3 Effect of temperature and pH on the activity of AKR and SDR ...... 158

6.4.4 Kinetic properties of AKR and SDR ...... 159

6.4.5 Substrate specificity of AKR and SDR ...... 160

6.5 Conclusion ...... 161

References ...... 163

Chapter 7: Increased Furfural Tolerance due to Overexpression of NADPH-dependent

Aldo/keto Reductase and Short-chain Dehydrogenase/reductase in Clostridium beijerinckii NCIMB 8052 ...... 183

7.1 Abstract ...... 183

7.2 Introduction ...... 184

7.3 Materials and Methods ...... 186

7.3.1 Bacterial strains, plasmids and culture conditions ...... 186

7.3.2 Construction of recombinant plasmids pWUR460_3974 and

pWUR460_3904 ...... 187

7.3.3 Electrotransformation of recombinant plasmids to C. beijerinckii 8052 .. 189

7.3.4 Purification of C. beijerinckii AKR and SDR and enzyme assay...... 190

7.3.5 Effect of furfural on C. beijerinckii and ABE fermentation ...... 191

7.3.6 Analytical methods ...... 192

7.4 Results and Discussion ...... 192

7.4.1 Gene cloning of AKR and SDR ...... 192

7.4.2 Purification of AKR and SDR and enzyme assay ...... 193

7.4.3 Effect of increasing AKR and SDR expression in C. beijerinckii on furfural

tolerance ...... 195

7.5 Conclusion ...... 198

References ...... 199

xvi

Chapter 8: Conclusion...... 210

Bibliography ...... 215

Appendix A: Additional Analysis of Differental Gene Expression due to Furfural

Challenge during Acidogenesis and Solventogenesis by C. beijerinckii 8052 ...... 242

xvii

List of Tables

Table 3.1 Analysis of sugars and inhibitors production after pretreatment and

hydrolysis of Miscanthus giganteus...... 60

Table 3.2 Estimation of glucose and xylose yield of Miscanthus giganteus

hydrolysates produced from liquid hot water pretreatment and enzymatic

hydrolysis...... 61

Table 3.3 Analysis of lignocellulose-derived microbial inhibitors at 0 h and 12 h of

fermentation by C. beijerinckii 8052 using P2 medium supplemented with

subleathal level of Miscanthus giganteus hydrolysates...... 62

Table 4.1 Effects of furfural, hydroxymethyl furfural (HMF), hydroxybenzaldehyde

and coumaric acid on microorganisms...... 86

Table 5.1 List of 30 genes and sequences of primers used in validation of microarray

analysis by Q-RT-PCR ...... 131

Table 6.1 PCR primers used to amplify AKR and SDR using C. beijerinckii 8052

genomic DNA as template...... 168

xviii

Table 6.2 Purification of Cbei_3974 (AKR) and Cbei_3904 (SDR) from Rosetta-

gami™ B(DE3)pLysS...... 169

Table 6.3 Kinetic parameters of purified AKR and SDR...... 170

Table 6.4 Substrate specificity of AKR and SDR...... 171

Table 7.1 PCR primers used to amplify aldo/keto reductase (AKR) and short-chain

dehydrogenase/reductase (SDR) using pET 15b_3974 and pET 15b_3904,

respectively, as the template...... 203

Table 7.2 Purification of AKR and SDR from C. beijerinckii 8052...... 204

Table A1a Enriched Up-regulated Gene Ontology Groups in the experiment of furfural

challenge during acidogenesis...... 243

Table A1b Enriched Down-regulated Gene Ontology Groups in the experiment of

furfural challenge during acidogenesis ...... 244

Table A1c Enriched Up-regulated Gene Ontology Groups in the experiment of furfural

challenge during solventogenesis ...... 245

Table A1d Enriched Down-regulated Gene Ontology Groups in the experiment of

furfural challenge during solventogenesis ...... 247

Table A2 Significantly regulated KEGG classifications during furfural challenge

experiment ...... 249

Table A3a Genes up-regulated by more than 3 folds during acidogenic furfural-

challenge...... 250

Table A3b Genes down-regulated by more than 3 folds during acidogenic furfural-

challenge ...... 253

xix

Table A3c Genes up-regulated by more than 3 folds during solventogenic furfural-

challenge ...... 255

Table A3d Genes down-regulated by more than 3 folds during solventogenic furfural-

challenge ...... 262

Table A4 Fold change of solvent production genes according to microarray analysis

...... 276

List of Figures

Figure 2.1 Schematic representation on pretreatment of lignocellulosic...... 41

Figure 2.2 Production of sugars and microbial inhibitors during lignocellulosic

biomass pretreatment and hydrolysis...... 42

Figure 2.3 The formation of furfural (a) and HMF (b) during the pretreatment of

lignocellulosic biomass...... 43

Figure 2.4 Simplified ABE production pathway in solventogenic clostridia...... 44

Figure 3.1 Cell growth and ABE production by C. beijerinckii 8052 using Miscanthus

giganteus hydrolysate as substrate...... 63

Figure 3.2 Cell growth and ABE production by C. beijerinckii 8052 using Miscanthus

giganteus hydrolysate as substrate supplemented with 4 g/L CaCO3...... 64

Figure 4.1 Biotransformation of furfural, HMF, 4-hydroxybenzaldehyde and p-

coumaric acid during ABE fermentation by C. beijerinckii 8052...... 82

Figure 4.2 Growth and ABE production profiles by C. beijerinckii 8052 using P2

medium with 2 g/L furfural or HMF...... 83

Figure 4.3 Growth and ABE production profiles by C. beijerinckii 8052 using P2

medium with 2 to 10 g/L furfuryl alcohol...... 84

xxi

Figure 4.4 Growth and ABE production profiles by C. beijerinckii 8052 using P2

medium with 0.5 g/L of 4-hydroxybenzaldehyde and 0.3 g/L p-coumaric

acid...... 85

Figure 5.1 Comparison of gene expression after furfural challenge at acidogenic and

solventogenic phases...... 134

Figure 5.2 Microarray validation of furfural-challenged gene expression during

acidogenic phase (A) and solventogenic phase (B)...... 138

Figure 5.3 Cell growth, solvent production and furfural reduction after furfural

challenge at the acidogenic phase...... 139

Figure 5.4 ABE production after furfural challenge at the acidogenic phase...... 140

Figure 5.5 Cell growth, solvent production and furfural reduction after furfural

challenge at the solventogenic phase...... 141

Figure 5.6 ABE production after furfural challenge at the solventogenic phase phase.

...... 142

Figure 5.7 ABE production after furfural challenge at the solventogenic phase phase.

...... 143

Figure 5.8 Reduction of furfural by acidogenic C. beijerinckii 8052 culture challenged

with furfural and supplemented ABE...... 144

Figure 6.1 Schematic representation of recombinant plasmids pET-15b_3974 (a) and

pET-15b_3904 (b)...... 172

Figure 6.2 SDS-PAGE analysis of AKR and SDR expressions induced by different

concentrations of IPTG...... 173

xxii

Figure 6.3 SDS-PAGE analysis of induction and purification of AKR (a) and SDR (b)

from Rosetta-gami™ B(DE3)pLysS...... 174

Figure 6.4 Codon usage frequency of Cbei_3974 (a) and Cbei_3904 (b) analyzed by

E. coli Codon Usage Analysis 2.0 ...... 175

Figure 6.5 SDS-PAGE analysis of Cbei_3974 (AKR) (a) and Cbei_3904 (SDR) (b)

from Rosetta-gami™ B(DE3)pLysS during puriﬁcation process...... 180

Figure 6.6 Effect of temperature on activity of furfural reduction by AKR (a) and SDR

(b)...... 181

Figure 6.7 Effect of pH on activity of furfural reduction by AKR (a) and SDR (b). . 182

Figure 7.1 Schematic representation of recombinant plasmids pWUR460_3974 (a) and

pWUR460_3904 (b)...... 205

Figure 7.2 SDS-PAGE analysis of purification of AKR (a) and SDR (b) from C.

beijerinckii 8052...... 206

Figure 7.3 Batch fermentations in P2 medium with 4 g/L furfural by C. beijerinckii

AKR+ (C. beijerinckii 8052 overexpressing AKR) and C. beijerinckii SDR+

(C. beijerinckii 8052 overexpressing SDR)...... 207

xxiii

Chapter 1: Introduction

Biofuel production using renewable substrates is increasingly recognized as a strategy to decrease our reliance on non-renewable fossil fuels. Biomass-based butanol is an attractive alternative to conventional ethanol biofuel because butanol has a number of advantages over ethanol, such as greater energy content, lower vapor pressure, and lower hygroscopicity (Dürre, 2008). Butanol can be produced by traditional acetone-butanol- ethanol (ABE) fermentation from biomass-derived sugars by solventogenic Clostridium species such as Clostridium beijerinckii NCIMB 8052 and Clostridium acetobutylicum

ATCC 824 (Qureshi and Ezeji, 2008). One of the most studied butanol-producing bacterium is Clostridium beijerinckii NCIMB 8052. It is a Gram-positive, obligately anaerobic, and endospore-forming bacterium (Mitchell, 1997), capable of utilizing cellobiose, glucose, xylose, arabinose and mannose to produce acetic acid, butyric acid, acetone, butanol, and ethanol. These solvents are produced via a metabolically biphasic fermentative stages commonly known as acidogenic (exponential) and solventogenic

(stationary) phases (Jones and Woods, 1986).

In the US, the most common biomass feedstock presently used for biofuel production is corn sugar, which has resulted in a heated debate over the impact on human

and animal food supply, and consequently the cost of corn (Yacobucci et al., 2007).

Hence, with regards to cost and renewability of substrates, lignocellulosic biomass is considered the most promising alternative for biofuel production (Hamelinck and Faaij,

2006). Lignocellulosic biomass refers to plant biomass primarily composed of cellulose, hemicelluloses and lignin. The carbohydrate polymers (cellulose and hemicelluloses) are tightly bound to the rigid lignin polymers (Mosier et al., 2005). To make the sugars from cellulose and hemicelluloses accessible to microbes, pretreatment and hydrolysis of lignocellulosic biomass is required.

ABE fermentation by solventogenic Clostridium species using lignocellulosic biomass remains a challenge as the release of fermentable sugars by acid/heat-aided hydrolysis is inevitably accompanied by the production of microbial inhibitors (Kumar et al., 2009).

Removal of these inhibitory compounds prior to fermentation has been reported by the use of chemical, physical, and biological methods (Larsson et al., 1999; Martinez et al.,

2001). However, these processes do not efficiently remove inhibitors from hydrolysates.

Furthermore, the increase in cost associated with these processes in addition to loss of fermentable sugars along with inhibitory compounds makes them less attractive for industrial scale application (Almeida et al., 2009). Alternatively, development of more robust microbial strains capable of in situ detoxification of lignocellulose-derived inhibitors is considered a more cost-effective approach for the advancement of viable biomass-to-biofuel industry. To achieve this goal, understanding the toxicity of lignocellulose-derived inhibitors and the tolerance mechanisms of solventogenic

Clostridium species is crucial.

Microbial inhibitors generated from lignocellulosic biomass are generally classified into three groups; aldehydes, phenols and organic acids (Liu and Blaschek,

2010). The most studied lignocellulose-derived inhibitors are furan aldehydes, including furfural and hydroxymethyl furfural (HMF). Furan aldehydes damage microbial cell walls and cell membranes, inhibit activities of glycolytic enzymes and RNA synthesis

(Zaldivar et al., 1999; Liu and Blaschek, 2010). Previous studies have demonstrated dose- dependent inhibitory effects of furan aldehydes on ethanol fermentation (Liu et al., 2004), and the conversion of furfural and HMF to their corresponding alcohols by ethanologenic yeast strains has also been reported (Liu et al., 2008). Phenolic compounds such as 4- hydroxybenzaldehyde, syringaldehyde and and p-coumaric acid appeared to exert much greater inhibitory effects than furans on most yeast (Delgenes et al., 1996) and bacterial strains (Zaldivar et al., 1999; Richmond et al., 2012) even at very low concentrations.

With regard to butanol fermentation by solventogenic Clostridium species, the impact of biomass degradation products such as aldehydes, organic acids, and phenolic compounds have been studied (Ezeji et al., 2007). The biotransformation of these inhibitors by solventogenic Clostridium species, however, has yet to be investigated, and more importantly, the underlying mechanisms of inhibitor detoxification and cell tolerance remain unclear. These gaps in knowledge greatly hinder the engineering of inhibitor- tolerant Clostridium strains.

The objective of this study, therefore, was to gain a better understanding of the interplay between the toxicity of lignocellulose-derived inhibitors and the mechanisms of inhibitor tolerance by butanol-producing C. beijerinckii 8052. To investigate the degree of inhibition of lignocellulosic hydrolysates and biotransformation of the attendant

inhibitors, ABE fermentation by C. beijerinckii 8052 was conducted using hydrolysates of Miscanthus giganteus, during which cell growth, ABE production, and inhibitor conversion were analyzed. In light of the finding that supplementation of fermentation broth by CaCO3 enhanced butanol tolerance by C. beijerinckii 8052 (Han et al., 2013), it was hypothesized that addition of CaCO3 to the fermentation broth might also mitigate inhibitor-mediated toxicity of Miscanthus giganteus hydrolysates. Hence, supplementation of Miscanthus giganteus hydrolysates with CaCO3 during ABE fermentation was conducted. Following our initial observation of the inherent ability of

C. beijerinckii 8052 to transform furfural (2 g/L), HMF (2 g/L), 4-hydroxybenzaldehyde

(0.5 g/L) and p-coumaric acid (0.3 g/L) (Chapter 4), transcriptomic analysis of C. beijerinckii 8052 challenged with 2 and 3 g/L furfural during acidogenic and solventogenic phase, respectively, was performed to uncover the underlying mechanisms at the mRNA level. Based on the results from the transcriptomic study, two candidate proteins, an aldo/keto reductase (AKR, encoded by Cbei_3974) and a short-chain dehydrogenase/reductase (SDR, encoded by Cbei_3904) were hypothesized to play key roles in the biotransformation of furfural to furfuryl alcohol. This hypothesis was confirmed by in vitro enzyme characterization using recombinant AKR and SDR obtained after overexpression in E. coli. The final objective of this study, therefore, was to (i) independently clone and overexpress AKR and SDR in C. beijerinckii 8052 and (ii) evaluate the furfural detoxification capacity of the resulting strains during growth and

ABE fermentation in medium supplemented with furfural.

References

Almeida, J.R.M., Bertilsson, M., Gorwa-Grauslund, M.F., Gorsich, S., Lidén, G., 2009.

Metabolic effects of furaldehydes and impacts on biotechnological processes. Appl.

Microbiol. Biotechnol. 82, 625-638.

Delgenes, J.P., Moletta, R., Navarro, J.M., 1996. Effects of lignocellulose degradation

products on ethanol fermentations of glucose and xylose by Saccharomyces

cerevisiae, Zymomonas mobilis, Pichia stipitis, and Candida shehatae. Enzyme

Microb.Technol. 19, 220-225.

Dürre, P., 2008. Fermentative Butanol Production, Ann. N. Y. Acad. Sci. 1125, 353-362.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Hamelinck, C.N., Faaij, A.P.C., 2006. Outlook for advanced biofuels, Energy Policy. 34,

3268-3283.

Han, B., Ujor, V., Lai, L.B., Gopalan, V., Ezeji, T.C., 2013. Use of Proteomic Analysis

To Elucidate the Role of Calcium in Acetone-Butanol-Ethanol Fermentation by

Clostridium beijerinckii NCIMB 8052. Appl. Environ. Microbiol. 79, 282-293.

Jones, D.T., Woods, D.R., 1986. Acetone-butanol fermentation revisited. Microbiol.Rev.

50, 484-524.

Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of

lignocellulosic biomass for efficient hydrolysis and biofuel production, Ind. Eng.

Chem. Res. 48, 3713-3729.

Larsson, S., Reimann, A., Nilvebrant, N., Jönsson, L.J., 1999. Comparison of different

methods for the detoxification of lignocellulose hydrolyzates of spruce, Appl.

Biochem. Biotechnol. 77, 91-103.

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxification,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Liu, Z.L., Moon, J., Andersh, B.J., Slininger, P.J., Weber, S., 2008. Multiple gene-

mediated NAD (P) H-dependent aldehyde reduction is a mechanism of in situ

detoxification of furfural and 5-hydroxymethylfurfural by Saccharomyces cerevisiae.

Appl. Microbiol. Biotechnol. 81, 743-753.

Liu, Z.L., Slininger, P.J., Dien, B.S., Berhow, M.A., Kurtzman, C.P., Gorsich, S.W.,

2004. Adaptive response of yeasts to furfural and 5-hydroxymethylfurfural and new

chemical evidence for HMF conversion to 2,5-bis-hydroxymethylfuran. J. Ind.

Microbiol. Biotechnol. 31, 345-352.

Martinez, A., Rodriguez, M.E., Wells, M.L., York, S.W., Preston, J.F., Ingram, L.O.,

2001. Detoxification of dilute acid hydrolysates of lignocellulose with lime,

Biotechnol. Prog. 17, 287-293.

Mitchell, W.J., 1997. Physiology of carbohydrate to solvent conversion by clostridia.

Adv. Microb. Physiol. 39, 31-130.

Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y., Holtzapple, M., Ladisch, M.,

2005. Features of promising technologies for pretreatment of lignocellulosic biomass.

Bioresour. Technol. 96, 673-686.

Qureshi, N., Ezeji, T.C., 2008. Butanol,‘a superior biofuel’production from agricultural

residues (renewable biomass): recent progress in technology, Biofuels Bioprod.

Bioreﬁn. 2, 319-330.

Richmond, C., Ujor, V., Ezeji, T.C., 2012. Impact of syringaldehyde on the growth of

Clostridium beijerinckii NCIMB 8052 and butanol production. 3 Biotech. 2, 159-167.

Yacobucci, B.D., Schnepf, R.D., Library of Congress. Congressional Research Service,

2007. Ethanol and biofuels: agriculture, infrastructure, and market constraints related

to expanded production.

Zaldivar, J., Martinez, A., Ingram, L.O., 1999. Effect of selected aldehydes on the growth

and fermentation of ethanologenic Escherichia coli. Biotechnol. Bioeng. 65, 24-33.

Chapter 2: Literature Review

2.1 Acetone-Butanol-Ethanol (ABE) fermentation

2.1.1 History of ABE fermentation

The production of butanol by microbial fermentation was first described by the famous French microbiologist Louis Pasteur in 1862, and during the latter part of 19th century microbiologists isolated pure bacterial strains producing butanol (Dürre, 2008).

But the production of acetone by fermentation process was first reported in early 1900s

(Jones and Woods, 1986). And these findings contributed to the launch of solvent production for use in synthetic rubber and other industrial feedstock chemicals, such as acetone, butanol and amyl alcohol. During this stage, Chaim Weizmann, a chemist who believed that alcohol production by fermentation was essential for rubber synthesis, trained himself to be a microbiologist and successfully isolated an organism from the soil that produced much more acetone and butanol than the previous strains, which was granted a patent in 1915 (Dürre, 2008). This bacterium, known as the Weizmann organism that was later named as Clostridium acetobutylicum, became very famous in not only biotechnology but also politics since the outbreak of World War I (WWI).

During WWI, the urgent requirement for acetone, as an additive for the production of explosives, made a fast development of acetone production by fermentation using

Weizmann organism (Dürre, 2008). At the same time, the by-product of acetone fermentation, butanol, was utilized in a relative small amount by Japan as a fuel extender for airplanes (Ezeji et al., 2010), while most of the butanol was simply stored for further use (Jones and Woods, 1986). After WWI, because of the dramatically increased production of automobile and the Prohibition in the United States, the shortage of the original solvent for car manufacture led to a demand for alternative solvent production

(Dürre, 2008), and therefore more butanol fermentation plants had been built (Mitchell,

1997). The onset of the Second World War again brought about the demand of acetone using butanol fermentation plants (Jones and Woods, 1986). However, during the middle of 20th century ABE production via fermentation process underwent a decline and most production plants in North America, Europe, and Russia had been shut down or converted for other purposes by 1980 (Ni and Sun, 2009). Instead of fermentation process, butanol was produced chemically from petroleum feedstock, since butanol biosynthesis could not beat the petrochemical synthesis method due to the low prices of crude oil at that time, the increasing price of fermentation substrate and the low yield.

There was a short stage of stimulation on the production of fuels from renewable resources during the oil crisis in 1973 to 1979 due to the geopolitical conflict (Dürre,

2008). And the constant shortage of fuel produced chaos in the West attracted great attention from scientists and politicians to the issues of energy crisis and the development of alternative fuels (Mitchell, 1997). However, the interest on biofuel production from renewable feedstock decreased since the subsequently decrease of crude oil price made the biofuel production not cost-competitive (Busche and Allen, 1989). Recently, academia and government have once again shifted their interest again to biofuels mainly

resulting from the increasing price of crude oil, the depletion of fossil fuel, the political instability of gasoline producing countries, and the environmental concerns (Antoni et al.,

2007). Since 2003, the petrochemical giant British Petroliem (BP) and biotechnology firm DuPont have been working together to develop advanced biofuel, and they announced the creation of partnership on 2006 to develop and produce the next generation of biofuel to meet the increasing global demand for biofuels and that their first product to market would be biobutanol (Hess, 2006). Today, researches on ABE fermentation mainly focus on the relieve of end products inhibition (Qureshi et al., 1992), improvement of organism tolerance to toxic metabolites (Ezeji et al., 2010), selection of high production strains (Qureshi and Blaschek, 2001), application of new technologies to fermentation process (Qureshi and Maddox, 1992), etc.

2.1.2 Solventogenic Clostridium sp. and mechanisms of ABE fermentation

2.1.2.1 Solventogenic Clostridium sp.

The production of acetone, ethanol and butanol by fermentation process is usually conducted by solventogenic Clostridium sp. No other genus in the three domains

(Bacteria, Archaea, and Eucarya) has better capability to produce butanol than solventogenic Clostridium (Qureshi and Ezeji, 2008). Clostridium are rod-shaped Gram- positive bacteria, and they are obligate anaerobes that can form endospores (Mitchell,

1997). Different species of solventogenic Clostridium have been recognized mainly based on their difference in type and ratio of the solvent they produced (Jones and

Woods, 1986). The strains most commonly used for ABE production are C. acetobutylicum ATCC 824 and C. beijerinckii 8052. These two strains have especially high solvent production capabilities and fully sequenced genomes. Many other solvent-

producing strains have also been reported, such as C. acetobutylicum P262, C. acetobutylicum NRRL B643, C. acetobutylicum B18, C. beijerinckii P260, C. beijerinckii

BA101, C. beijerinckii LMD 27.6, C. aurantibutyricum, and C. tetanomorphum (Qureshi and Ezeji, 2008). C. beijerinckii produces solvents in approximately the same ratio as C. acetobutylicum (a 6:3:1 ratio for butanol:acetone:ethanol), but isopropanol may also be produced, and C. aurantibutyricum produces both acetone and isopropanol in addition to butanol. However, C. tetanomorphum produces almost equimolar amounts of butanol and ethanol but no other solvents (Jones and Woods, 1986).

2.1.2.2 Mechanisms of ABE fermentation

The biochemical pathways utilized by solventogenic Clostridium sp. (Figure 2.4) to convert carbohydrates to fermentation products start with transport of monosaccharide across the cytoplasmic membrane to the cytoplasm via specific membrane-bound transport systems. For example, glucose, fructose, glucitol and mannitol were considered to be transported by the phosphoenolpyruvate (PEP)-dependent phosphotransferase system (PTS) (Mitchell, 1996), while in a hyper-butanol-producing mutant C. beijerinckii, glucose uptake and accumulation might rely on both PEP-dependent phosphotransferase system and adenosine triphosphate (ATP)-dependent glucose phosphorylation, with each predominanting in different fermentation stages (Lee and

Blaschek, 2001). Recently, transporters of xylose and arabinose were reported and overexpressed, resulting in an elevated substrate consumption and ABE production (Xiao et al., 2011). After transported into the cell, different sugars are then metabolized via different series of enzymatic reactions. One mole of glucose is broken down via the

Embden-Meyerhof-Parnas (EMP) pathway with the production of 2 moles of pyruvate as

well as the net production of 2 moles of adenosine triphosphate (ATP) and 2 moles of reduced nicotinamide adenine dinucleotide (NADH). While pentoses are fermented by pentose phosphate pathway with the conversion of 3 moles of pentose to 5 moles of ATP and 5 moles of NADH and the production of fructose 6-phosphate and glyceraldehyde 3- phosphate, the latter two of which are subsequently converted into pyruvate through EMP pathway (Jones and Woods, 1986).

In typical batch fermentation, production of ABE by solvent-producing

Clostridium species is a biphasic process including an acidogenic phase during the initial growth stage and a solventogenic phase after the culture enters the stationary growth stage. During the acidogenic phase, pyruvate is oxidized to acetyl-CoA, followed by the production of acetic acid from acetyl-CoA and butyric acid from butyryl-CoA with the generation of ATP. During this stage, the pH of culture medium decreases due to the acids formation. The shift to solvent production during the second phase of fermentation involves the reassimilation of previously produced acetic acid and butyric acid to generate acetone, butanol and ethanol with a concomitant increase in the culture pH. The production of acetone is directly coupled to the re-uptake of acetic acid and butyric acid by means of an acetoacetyl-CoA:acetate/butyrate:CoA transferase and subsequently acetoacetate decarboxylase or non-enzymatic decarboxylation (Han et al., 2011). Acetone production drives the unfavorable reaction from acetic and butyric acids to acetyl-and butyryl-CoA which is further converted to ethanol and butanol by aldehyde and alcohol dehydrogenases (Jones and Woods, 1986). It is suggested that acid uptake functions as a detoxification process to maintain metabolism in an unfavorable condition, since solvents are less toxic than the acids produced (Hartmanis et al., 1984). The regulations which

underlie the shift from the acidogenic to solventogenic phase of bacterial growth have been investigated previously. Researches indicate that the low internal pH due to the accumulation of acids plays an important role in triggering the onset of solventogenisis

(Gottwald and Gottschalk, 1985). Besides, the intracellular level of butyryl phosphate has been suggested to affect solvent formation since butyryl phosphate may act as a phosphodonor of transcriptional factor, and high level of butyryl phosphate upregulates the solvent production and stress genes leading to the initiation of solvent production much earlier and stronger than that of low level butyryl phosphate strains (Zhao et al.,

2005). Another factor proposed to induce the shift from acid to solvent production is the high concentration of NADH which is accumulated during the initial growth phase. A high level of NADH results in a high production of solvents with increased activities of

NADH-dependent alcohol and aldehyde dehydrogenase (Girbal et al., 1995). In addition, since the regeneration of NAD+ is very important for the continued sugar break down via glycolysis pathway (Ezeji et al., 2010), the consumption of NADH may also be seen as an adaptive mechanism towards cell survival.

2.2 Lignocellulosic biomass

2.2.1 Sources of Lignocellulosic biomass

Lignocellulosic biomass refers to plant biomass composed of cellulose, hemicelluloses and lignin. The carbohydrate polymers including cellulose and hemicelluloses are tightly encased by the rigid lignin polymers (Mosier et al., 2005).

Using lignocellulosic biomass for biofuel production offers several benefits, for example, abundant and diverse raw material compared to sources such as corn and sugarcane, reduction of market stress on food crops and increase of biofuel net energy yield

(Hamelinck and Faaij, 2006). Lignocellulosic biomass includes several categories like agricultural residuals (e.g. corn stover and wheat straw), forest residuals (e.g. branches and foliage), organic fractions of municipal solid waste (e.g. kitchen and yard waste, and packing waste), and dedicated energy crops (e.g. giant miscanthus and switch grass). It is estimated that the cropland covers an area of only 20% of the acreage of the United

States, but 50% of the total acres in the country has the capability of growing biomass

(Bohlmann, 2006). And every year about 400 million tons of lignocellulosic biomass is produced in the United States and 1.3 billion tons per year is expected to be produced by

2030 (Yacobucci et al., 2007). Lignocellulosic biomass is the most abundant renewable material on the planet with many other advantaged such as inexpensive and environmental friendly, and only 3% of it is used by human, which is therefore a great potential starting material for many industrial processes (Lucia, 2008).

2.2.2 Pretreatment of lignocellulosic biomass

Unlike the easily accessible sugars such as starch from corn and sucrose from sugarcane, simple sugars from lignocellulosic biomass have to be released for microbial fermentation through pretreatment processes (Figure 2.1) (Mosier et al., 2005). To overcome the recalcitrance of lignocellulosic biomass, appropriate pretreatment process is required to remove lignin, depolymerize hemicelluloses, reduce the crystallinity of cellulose, and increase the surface area prior to further hydrolysis (Kumar et al., 2009).

Different types of pretreatment techniques have been applied to process different biomass for the production of biofuel, and pretreatment techniques can be generally classified into three categories based on their applications and types of catalyst, that is, physical, chemical, and biological pretreatment (Zheng et al., 2009). Physical pretreatment,

without using any chemical agents, includes uncatalyzed steam explosion, liquid hot water pretreatment (LHW), mechanical comminution, and high energy radiation.

Chemical pretreatment, the most studied pretreatment technique, uses catalysts such as acid, alkaline, organosolv, etc. Different from physical and chemical methods, the biological pretreatment are conducted under a mild condition using biomass-degrading microorganisms especially white-, brown-, soft-rot fungi to modify the composition and structure of lignocellulosic biomass and thus to facilitate further enzymatic digestion

(Zheng et al., 2009). Besides those individual pretreatments, combinations of two or more pretreatment techniques are also very common, for example, physicochemical pretreatment using steam explosion with addition of H2SO4 or CO2, and ammonia fiber explosion (AFEX) (Kumar et al., 2009). An effective pretreatment process must meet the following criteria: (1) increasing the release of sugars or the potential to generate simple sugars, (2) preserving the sugars from degradation during physical and chemical pretreatments or utilization by other organisms during biological pretreatment, (3) limiting the release and formation of microbial inhibitors, (4) reducing the cost by minimizing energy input, using inexpensive catalysts, recovering high-value by products, etc. Given these requirements, none of the currently available pretreatment technologies is universally ideal, and each has its own advantages and disadvantages. For example, in liquid hot water (LHW) pretreatment, biomass undergoes high pressure cooking in hot water (160°C to 240°C) without any addition of acids, which reduces cost on catalyst as well as chemicals used for neutralization and also produces less amount of inhibitors compared to pretreatment with acidic catalysts. However, the need of high pressure, high temperature and a large amount of water supplied to the system increase the energy input

and cost (Brodeur et al., 2011). While another example, dilute acid pretreatment, uses catalysts such as dilute sulfuric acid, dilute hydrochloric acid, and dilute phosphoric acid and performs well on a wide range of biomass materials (Zheng et al., 2009). The acid- based pretreatment has an advantage of solublizing hemicelluloses and releasing xylose up to 90% of the theoretical value, and it is also relatively inexpensive and effective especially when using dilute sulfuric acid as the catalyst. However, acid treatment results in production of microbial inhibitors, requirement for neutralization, generation of salts which are potential microbial inhibitor, and corrosion problems (Zheng et al., 2009).

2.2.3 Enzymatic hydrolysis of lignocellulosic biomass

Although solventogenic clostridia secrete numerous enzymes such as amylase, glucosidase, glucoamylase, pullulanase, and amylopullulanase enzymes to break down the polymeric carbohydrate to monomeric sugars, there is no evidence to show that solventogenic clostridia are capable of directly using crystalline cellulose or hemicelluloses as a carbon source (Ezeji et al., 2007a). Since the pretreatment process of lignocellulosic biomass is designed to degrade lignin and may partially destroy the cellulose and hemecellulose structures, further hydrolysis of these polymers to monomeric sugars is necessary. Compared with acid or alkaline hydrolysis, enzymatic hydrolysis of biomass has many advantages especially the lower utility cost because of the mild conditions used during the process as well as the absence of corrosion problems

(Sun and Cheng, 2002).

The major sugars fermented by microbes are glucose generated from cellulose and many other hexoses and pentoses from hemicelluloses such as xylose, mannose, galactose and arabinose. The hydrolysis of cellulose to glucose requires cellulases. They

are a mixture of enzymes including (1) endoglucanase (EC 3.2.1.4) that randomly cleaves internal bonds and create free-chain ends, (2) exoglucanase (EC 3.2.1.91) that usually cleaves two units from the free-chain ends produced by endoglucanase, releasing cellobiose, and (3) β-glucosidase or cellobiase (EC 3.2.1.21) which hydrolyzes cellobiose, the product of exoglucanase, to produce glucose (Sun and Cheng, 2002). To completely hydrolyze hemicelluloses, a number of enzymes are required including mainly xylanase and β-xylosidase, as well as acetylxylan esterase, α-arabinofuranosidase,

α-glucuronidase, α-galactosidase, ferulic and/or p-coumaric acid esterase (Ezeji et al.,

2007a).

2.2.4 Lignocellulose-derived microbial inhibitory compounds

A major limitation of using lignocellulosic biomass as the chief substrate for biofuel production is the generation of numerous inhibitory compounds during the pretreatment and hydrolysis of lignocellulosic biomass (Figure 2.2), because these compounds in biomass hydrolysates generally inhibit the growth of fermenting microorganisms, including yeast and bacteria strains (Klinke et al., 2004), and production of biofuel (Ezeji et al., 2007b). Since these inhibitory compounds are often part of the lignin or hemicellulose structure, it is difficult to avoid releasing them during complete hydrolysis of lignocellulosic biomass to monomeric sugars. To facilitate fermentation process, removing these inhibitory compounds from biomass hydrolysates prior to fermentation is necessary but costly, and adds extra steps to the process (Almeida et al.,

2009). Therefore there is increasing interest in studies of inhibitor detoxification by microorganisms have been increasing during the past few decades in the aim of developing inhibitor tolerant strains during biomass to biofuel processes (Liu, 2011).

2.2.4.1 Classification of inhibitors

The variety and concentration of degradation products resulting from the pretreatment of lignocellulosic biomass may vary significantly due to the natural composition of biomass resources as well as the pretreatment conditions. However, these inhibitors are generally grouped into several categories: aldehydes, phenols and organic acids (Liu and Blaschek, 2010). Aldehyde inhibitors are organic compounds containing one or more functional aldehyde groups and the base structure of aldehydes could be a furan ring, e.g. furfural and 5-hydroxymethylfurfural (HMF), or a phenol-related structure, e.g. 4-hydroxybenzaldehyde, syringaldehyde and cinnamaldehyde. Aldehyde inhibitors are generated either from degradation of sugars or from lignin depolymerization. For example, furfural and HMF, each containing a furan ring and an aldehyde group, are produced by dehydration of pentose and hexose, respectively, and those sugars are released from hemicelluloses and cellulose during lignocellulose decomposition (Figure 2.3). Some other aldehyde inhibitors are released from the breakdown of lignin structure for example vanillin and syringaldehyde (Mialon et al.,

2010). Besides the phenolic aldehydes, other inhibitory phenolic compounds are also generally found from degradation of lignin, including phenol, 2-methlphenol, coniferyl alcohol, vanillyl alcohol, etc (Liu and Blaschek, 2010). Another important inhibitor category is organic acids, and carboxylic acid is the most commonly type found in lignocellulosic hydrolysates. Simple acids may produced by depolymerization of hemicelluloses, such as acetic acid and furoic acid (Liu and Blaschek, 2010), or by further degradation of monomer sugars, for example, levulinic acid can be produced under high temperature and acidic conditions from degradation of HMF, the dehydration

product of hexose (Girisuta et al., 2006), and similarly formic acid may be produced due to further breakdown of furfural which is derived from pentose (Lamminp et al., 2012).

In addition, many phenol-based acids are classified as organic acid inhibitors because their inhibition actions are more likely from the carboxyl functional group, such as 4- hydroxybenzoic acid, vanillic acid, syringic acid and ferulic acid (Liu and Blaschek,

2010). Similar to the source of other phenolic compounds, these phenolic acids are generally from lignin degradation (Hedges and Ertel, 1982).

2.2.4.2 Effects of inhibitors

The effects of different inhibitors usually vary when different organisms or strains are used during the fermentation process. The degree of inhibition of the aldehyde, phenol and organic acid also depends on many other factors such as the concentrations and types of inhibitors, temperature and pH of the culture medium, ratio of cells to inhibitors, and physiological conditions of cells (Leonard and Hajny, 1945).

The most well-studied aldehyde inhibitors are furan aldehydes, including furfural and HMF. Since furfural and HMF are chemically related compounds in that each has one furan ring and an aldehyde group, it is not surprising that they exhibt similar effects.

Both furans have been reported to cause a dose-dependent inhibition on cell growth and metabolic activities during fermentation by yeast (Liu et al., 2004). In addition, conversions of furfural and HMF to their corresponding alcohols have been reported previously (Liu et al., 2008). Despite their similarities, furfural and HMF also have important differences on their inhibitory effects. In a study on ethanol fermentation by

Escherichia coli, supplementation of furfural resulted in a stronger and immediate inhibition than HMF and other aldehydes, suggesting a direct inhibitory effect of furfural

on fermentation enzymes, for example, alcohol dehydrogenase, as well as glycolytic enzymes (Zaldivar et al., 1999). Also, a study on Saccharomyces cerevisiae shows that the conversion of furfural is approximately four-fold faster than that of HMF, suggesting possible differences in membrane transport rates or affinities to their target enzymes

(Taherzadeh et al., 2000). While these furan compounds resulted obvious inhibition to yeast and E. coli strains even at a lower concentration, they were found to be stimulatory rather than inhibitory to solventogenic clostridia at a concentration up to 3 g/L of each; but the mixture of the two still negatively affected the culture (Ezeji et al., 2007b). These furans have also been reported to damage cell walls and cell membranes, inhibit activities of enzymes, and inhibit RNA synthesis in microorganisms (Liu and Blaschek, 2010).

Other aldehyde inhibitors such as 4-hydroxybenzaldehyde and syringaldehyde appeared to exert a much greater inhibition to most yeast (Delgenes et al., 1996) and bacteria strains (Zaldivar et al., 1999; Richmond et al., 2012) even at concentrations less than 5 mM. Compared to the aldehyde inhibitors derived from sugar degradation, i.e. furfural and HMF, the aldehydes from lignin degradation (e.g. 4-hydroxybenzaldehyde and syringaldehyde) are more toxic (Parajó et al., 1998).

Regarding another inhibitor group, weak organic acids inhibit the growth of microorganisms and are therefore used as food preservatives. In general, weak acids are more toxic to bacteria such as E. coli (Klinke et al., 2004) and Clostridium sp. (Ezeji et al., 2007b; Chamkha et al., 2001) than to yeast strains. The inhibition of weak acids on cell growth may result from the inflow of undissociated acid from the extracellular environment into cytoplasm, leading to dissociation of weak acid in the neutral intracellular condition and therefore a decrease of intracellular pH, which has a negative

effect on cell proliferation and viability as well as pH-dependent fermentation process

(Palmqvist and Hahn-Hägerdal, 2000). It has also been proposed that the drop of intracellular pH induces the adaptive action of proton pumping out of the cell with the comsumption of ATP to neutralize the cytoplasmic pH. However, high concentration of acid may exhaust the proton pumping capacity by depletion of ATP content (Palmqvist and Hahn-Hägerdal, 2000). The inhibitory effects of phenolic compounds are not as clear as the other categories, mainly because of a lack of accurate qualitative and quantitative analyses. The toxicity of phenols may arise from the hydrophobic effect and thus phenols can cause irreversible permeabilization of the cell membrane, thereby affecting their functions as selective barriers and enzyme matrices (Palmqvist and Hahn-Hägerdal,

2000).

2.2.4.3 Removal of inhibitors

Given the production of toxic compounds during the pretreatment and hydrolysis processes of lignocellulosic biomass, removal of these inhibitors from biomass hydrolysates is necessary to facilitate cell growth and fermentation. Depending on biomass material, the pretreatment technology and the microorganism used for fermentation, different inhibitor removal approaches may be applied. The most commonly used methods for inhibitor removal can be grouped into three categories: physical, chemical, or biological method (Liu and Blaschek, 2010). Vacuum evaporation is a physical detoxification method with the advantage of reducing the concentration of volatile compounds such as furfural, acetic acid, and vanillin in the biomass hydrolysate

(Mussatto and Roberto, 2004). However, fermentations using hydrolysate that were treated with this method was drastically inhibited due to the concentrations of non-

volatile lignin derivatives or non-volatile dissociated form of acetic acid (Mussatto and

Roberto, 2004). As part of the chemical methods, some popular extraction chemicals are used to remove inhibitory compounds including methyl tert-butyl ether (MTBE) following pre-treatment and hydrolysis of sawdust (Ranatunga et al., 1997), and ethyl acetate to remove inhibitors formed from corn stover, switchgrass and poplar hydrolyates

(Fenske et al., 1998). Activated charcoal has veen used to adsorb the inhibitory compounds; Ca(OH)2 and NaOH have been applied to adjust the pH, thus precipitating inhibitory compounds out of solution (Martinez et al., 2001), which may reduce phenol and aldehyde compounds and improve fermentation production. However, the obvious drawback of this method is that additional chemicals (e.g. CaSO4) are generated, whose removal before fermentation increases the production cost (Liu and Blaschek, 2010).

Moreover, acids are required to neutralize the hydrolysate after treatment with the alkali, which not only increases cost but also generate salts that may be fermentation inhibitors as well (Nakas et al., 1983). Biological method of detoxification involves the use of either enzymatic treatment or microorganisms especially fungi. Glucose utilization and solvent production were enhanced when laccase and lignin peroxidase were added to the willow hydrolysate (Jönsson et al., 1998). The fungus Coniochaeta ligniaria has been reported to remove furfural and hydroxymethylfurfural from corn stover hydrolysate

(Lopez et al., 2004).

2.3 Molecular mechanisms of in situ detoxification

The biotransformation of furfural and HMF has been reported using aerobic or facultative microorganisms such as S. cerevisiae (Heer et al., 2009; Taherzadeh et al.,

2000) and E. coli (Gutiérrez et al., 2002; Boopathy et al., 1993). The mechanisms of

detoxification of these furan inhibitors by yeast or E. coli are not involved in utilization or degradation of these compounds, but microorganism catalyzes the reduction of furfural and HMF to the furfuryl alcohol and HMF alcohol, respectively, because the alcohol forms are less toxic than their aldehyde forms on a molar or weight basis (Zaldivar et al.,

2000). Catalyzed by oxidoreductases, furfural is converted to furfuryl alcohol by accepting 2e¯ carried with NAD(P)H (Boopathy, 2009), and similarly, HMF is reduced to its corresponding alcohol, identified as being 2,5-bis-hydroxymethylfuran (Liu et al.,

2004). Furfural and HMF show maximum absorbance at 276 and 282 nm, respectively, while furfuryl alcohol and HMF alcohol have maximum absorbance at 220 and 222 nm, respectively. Thus, concentrations of furans in the samples are easily determined with a

UV-Vis spectrophotometer (Boopathy et al., 1993) or using high-performance liquid chromatography (HPLC) equipped with photodiode array detector (Gutiérrez et al.,

2002). In yeast cell, a shortage of NADH has been observed when furfural is added to the culture, indicating a competition of NADH between furfural reduction and glycolysis, and this leads to an accumulation of acetaldehyde and thus a delay of ethanol production

(Liu and Blaschek, 2010).

With respect to the enzymes catalyzing the conversion of furfural, resent studies investigated different oxidoreductases such as aldehyde dehydrogenase IV (ALD4), alcohol dehydrogenase VI (ADH6), alcohol dehydrogenase VII (ADH7), aldose reductase III (GRE3) (Liu et al., 2008), aldehyde reductase Ari1P (Bowman et al., 2010), aldehyde reductase YqhD and DkgA (Jarboe, 2011). As for the enzyme cofactors, preferences on both NADH or NADPH have been observed. For example, ADH7, ALD4 and GRE3 have clear NADH preference on both furfural and HMF reduction, while

ADH6, AriP, YqhD and DkgA are NADPH-dependent (Liu et al., 2008; Bowman et al.,

2010; Jarboe, 2011). Regarding the substrate specificity, most of the identified enzymes showed a broad-substrate range. For example, ADH6 and ADH7 showed similar substrate activities toward a wide range of aldehyde (Larroy et al., 2002a; Larroy et al.,

2002b); AriP is shown to have the ability to reduce at least 14 aldehyde substrates

(Bowman et al., 2010); and YqhD has the reductase activity for more than 10 aldehyde substrate (Jarboe, 2011). Despite the enzymes having been known to play important roles in aldehyde detoxification, no report has been shown that a single gene deletion in these enzymes exhibited obvious defects on cell growth or the biotransformation of aldehyde

(Liu and Blaschek, 2010). Therefore, it is very clear that the mechanisms of detoxification are performed by a very complex metabolic network with numerous genes and regulatory cascades, rather than a single gene. For example, a study on S. cerevisiae has shown that overexpression of genes associated with the pentose phosphate pathway benefits the tolerance of cells to furfural stress probably due to the increased abundance of reducing power (Gorsich et al., 2006). Proteomic analyses have shown that furfural stress induces cellular adaptations including a central carbon metabolism rearrangement as well as responses to stresses like unfolded protein, oxidative stress, osmotic and salt stress, DNA damage and nutrient starvation (Lin et al., 2009). Although not studied as much as aldehyde inhibitors, biotransformations of phenolic compounds have also been investigated using yeast and bacteria. A yeast species Brettanomyces anomalus is capable of converting p-coumaric, caffeic and ferulic acid to 4-vinyl and 4-ethyl derivatives, but the reaction enzymes and pathways were not characterized (Edlin et al., 1995). Under anaerobic conditions, aromatic compounds such as vanilline, phenol, hydroxybenzoic

acid and benzaldehyde may be metabolized through benzoyl-CoA pathway, and the resulting benzoyl-CoA may be further coverted to acetyl-CoA by bacteria which is an important metabolic intermediate (Harwood et al., 1998). Since the mechanism of detoxification is regulated by numerous genes rather than a single gene or a couple of genes, genomic-based technologies have been showing great advantages in selecting the genes responding to inhibition effects from thousands or tens of thousands genes, and uncovering the gene interaction and regulatory network, which greatly benefits the development of tolerant strain with the aim of efficient detoxification of inhibitors derived from lignocellulosic biomass.

2.4 Transcriptome analysis using microarray

Transcriptome analysis is a method for examination of total set or subset of transcripts produced in a particular cell type. Since mRNA transcripts may vary with the external environmental conditions, the study of transcriptome is expected to detect and compare the expression of differential mRNA levels in different cell samples. It is increasingly important in revealing the alteration of gene expression due to an environmental challenge or during disease progression. Analysis of global pattern of gene expression provides researchers with comprehensive information to understand the molecular mechanisms and metabolic pathways to regulate the cellular defense, mutation, programmed death, etc.

Microarray technology is currently the most widely used method to study genome-wide differential RNA expression. DNA microarray is a small chip with a collection of tens of thousands or more of oligonucleotide probes immobilized on its solid surface such as glass, plastic and silicon (Heller, 2002), and complementary nucleic

acid sequences can be detected and quantified when bound to their specific probes at a given condition. During the past couple of decades, an enormous number of publications have appeared using microarray analysis on the area of molecular biological and genomic research, pharmacogenomic research, infectious and genetic disease and cancer diagnostics for the purpose of gene annotation, functional pathway prediction, new drug discovery, as well as disease illumination and prediction (Stoughton, 2005).

Microarray assays, especially high-density microarrays, generate large dataset because each study comprises multiple arrays and each array includes tens of thousands of oligonucleotide probes. To precisely and accurately extract useful information from the complex data, statistical and bioinformatics tools have been rapidly developed and exploited. Statistical analysis includes reducing background noise when acquiring data from the scanned image, assessing the reproducibility of each spot value, adjusting bias from non-biological variations, and selecting differentially expressed genes (Smyth et al.,

2003). Statistical analysis usually generates a large set of genes that are differentially expressed among the samples. To uncover the biological meaning of statistical results, multiple bioinformatic tools can be used. The task of analyzing the resulting gene set is to systematically map the genes with similar behavior in various of conditions into an associated biological annotation, for example, gene ontology terms and metabolic pathways. The analysis of large gene lists is more of an exploratory computational procedure benefiting from the development of clustering and visualizing software tools such as Cluster and TreeView (Eisen et al., 1998), GenMAPP (Salomonis et al., 2007),

TIGR MultiExperiment Viewer (Sutton et al., 1995), JAVA Treeview(Saldanha, 2004) and DAVID bioinformatics resources (Huang et al., 2009). This analysis assists biologists

to the discovery of novel genes associated with biological processes and molecular functions (Liu, 2002). The reason is that multiple genes might be involved in the same cell activity and these genes often have the similar expression pattern under the same environmental condition. When an unknown gene is classified into a group of genes with a known biological function, it is capable of giving an indication of the role of the unknown gene, and moreover, the coexpression of the genes also suggests the possibility of coregulation under a same regulon (Moreau et al., 2002).

Microarray analysis has an obvious advantage to measure the expression of a large number of genes simultaneously. However, the quality of data generated from microarray may vary greatly with regard to the platforms and procedures used for experimental design, sample preparation, hybridization process, as well as data acquisition and normalization. Therefore, an independent method of gene expression analysis is required to validate the microarray result, and quantitative real-time PCR

(qPCR) is a commonly used tool for this purpose. The correlation of gene expression results from microarray and qPCR is analyzed to indicate the agreement between the two methods. However, the data from qPCR and microarray often result in disagreement, and the factors influences the analysis include up- or down-regulation of genes and qPCR cycle threshold (Ct) value (Morey et al., 2006).

2.5 Conclusions

Clostridium species are bacteria that have especially high butanol production capabilities, by virtue of their ability to utilize sugars present in lignocellulosic hydrolysates. In addition to fermentable sugars, pretreatment of lignocellulose-based raw materials generates complex mixtures of microbial inhibitory compounds. The inhibitory

properties of biomass degradation products are widely recognized as one of the major limitations to the conversion of biomass to biofuel, thereby impeding the commercialization of biofuels. While some studies have been conducted on effects of microbial inhibitors, very little is known on the furfural mechanism during biobutanol fermentation process using Clostridium species. This gap in knowledge continues to hamper attempts at engineering inhibitor-tolerant strains, evident in the persisting low butanol productivity from biomass feedstock, hence, the high cost of butanol production at industrial scale. The overall objective of this dissertation was to investigate the impacts of lignocellulose-derived inhibitory compounds (using furfural as a model) to butanol fermentation by solventogenic Clostridium species on inhibitor tolerance and butanol production, and to identify and characterize furfural transformation genes and enzymes to enhance understanding of furfural toxicity mechanisms. The hypothesis was that challenging C. beijerinckii 8052 culture with inhibitory compound, furfural, would alter patterns of gene expression in this microorganism, and consequently, result in the induction of adaptation machineries against furfural. Identification and overespression of genes involved in the biotransformation of aldehydes (e.g. furfural and HMF) to the corresponding alcohols inC. beijerinckii 8052 would enhance its furfural biotransformation rate and tolerance.

References

Almeida, J.R.M., Bertilsson, M., Gorwa-Grauslund, M.F., Gorsich, S., Lidén, G., 2009.

Metabolic effects of furaldehydes and impacts on biotechnological processes. Appl.

Microbiol. Biotechnol. 82, 625-638.

Antoni, D., Zverlov, V.V., Schwarz, W.H., 2007. Biofuels from microbes. Appl.

Microbiol. Biotechnol. 77, 23-35.

Augenlicht, L.H., Kobrin, D., 1982. Cloning and screening of sequences expressed in a

mouse colon tumor. Cancer Res. 42, 1088-1093.

Augenlicht, L.H., Wahrman, M.Z., Halsey, H., Anderson, L., Taylor, J., Lipkin, M.,

1987. Expression of cloned sequences in biopsies of human colonic tissue and in

colonic carcinoma cells induced to differentiate in vitro. Cancer Res. 47, 6017-6021.

Bohlmann, G.M., 2006. Process economic considerations for production of ethanol from

biomass feedstocks. Ind. Biotechnol. 2, 14-20.

Boopathy, R., 2009. Anaerobic biotransformation of furfural to furfuryl alcohol by a

methanogenic archaebacterium. Int. Biodeterior. Biodegrad. 63, 1070-1072.

Boopathy, R., Bokang, H., Daniels, L., 1993. Biotransformation of furfural and 5-

hydroxymethyl furfural by enteric bacteria. J. Ind. Microbiol. Biotechnol. 11, 147-

150.

Bowman, M.J., Jordan, D.B., Vermillion, K.E., Braker, J.D., Moon, J., Liu, Z.L., 2010.

Stereochemistry of furfural reduction by a Saccharomyces cerevisiae aldehyde

reductase that contributes to in situ furfural detoxification. Appl. Environ. Microbiol.

76, 4926-4932.

Brodeur, G., Yau, E., Badal, K., Collier, J., Ramachandran, K., Ramakrishnan, S., 2011.

Chemical and Physicochemical Pretreatment of Lignocellulosic Biomass: A Review.

Enz. Research. 2011, 1-17.

Busche, R.M., Allen, B., 1989. Technoeconomics of butanol extractive fermentation in a

multiphase fluidized bed bioreactor. Appl. Biochem. Biotechnol. 20, 357-374.

Chamkha, M., Garcia, J.L., Labat, M., 2001. Metabolism of cinnamic acids by some

Clostridiales and emendation of the descriptions of Clostridium aerotolerans,

Clostridium celerecrescens and Clostridium xylanolyticum. Int. J. Syst. Evol.

Microbiol. 51, 2105-2111.

Delgenes, J.P., Moletta, R., Navarro, J.M., 1996. Effects of lignocellulose degradation

products on ethanol fermentations of glucose and xylose by Saccharomyces

cerevisiae, Zymomonas mobilis, Pichia stipitis, and Candida shehatae. Enzyme

Microb. Technol. 19, 220-225.

Dürre, P., 2008. Fermentative Butanol Production. Ann. N. Y. Acad. Sci. 1125, 353-362.

Edlin, D.A.N., Narbad, A., Dickinson, J.R., Lloyd, D., 1995. The biotransformation of

simple phenolic compounds by Brettanomyces anomalus. FEMS Microbiol. Lett. 125,

311-315.

Eisen, M.B., Spellman, P.T., Brown, P.O., Botstein, D., 1998. Cluster analysis and

display of genome-wide expression patterns. Proc. Natl. Acad. Sci. 95, 14863- 14868.

Ezeji, T., Milne, C., Price, N.D., Blaschek, H.P., 2010. Achievements and perspectives to

overcome the poor solvent resistance in acetone and butanol-producing

microorganisms. Appl. Microbiol. Biotechnol. 85, 1697-1712.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007a. Bioproduction of butanol from biomass:

from genes to bioreactors. Curr. Opin. Biotechnol. 18, 220-227.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007b. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Fenske, J., Griffin, D., Penner, M., 1998. Comparison of aromatic monomers in

lignocellulosic biomass prehydrolysates. J. Ind. Microbiol. Biotechnol. 20, 364-368.

Girbal, L., Croux, C., Vasconcelos, I., Soucaille, P., 1995. Regulation of metabolic shifts

in Clostridium acetobutylicum ATCC 824. FEMS Microbiol. Rev. 17, 287-297.

Girisuta, B., Janssen, L., Heeres, H., 2006. A kinetic study on the decomposition of 5-

hydroxymethylfurfural into levulinic acid. Green Chem. 8, 701-709.

Gorsich, S., Dien, B., Nichols, N., Slininger, P., Liu, Z., Skory, C., 2006. Tolerance to

furfural-induced stress is associated with pentose phosphate pathway genes ZWF1,

GND1, RPE1, and TKL1 in Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol.

71, 339-349.

Gottwald, M., Gottschalk, G., 1985. The internal pH of Clostridium acetobutylicum and

its effect on the shift from acid to solvent formation. Arch. Microbiol. 143, 42-46.

Gutiérrez, T., Buszko, M.L., Ingram, L.O., Preston, J.F., 2002. Reduction of furfural to

furfuryl alcohol by ethanologenic strains of bacteria and its effect on ethanol

production from xylose. Appl. Biochem. Biotechnol. 98-100, 327-340.

Hamelinck, C.N., Faaij, A.P.C., 2006. Outlook for advanced biofuels. Energy Policy. 34,

3268-3283.

Han, B., Gopalan, V., Ezeji, T.C., 2011. Acetone production in solventogenic

Clostridium species: new insights from non-enzymatic decarboxylation of

acetoacetate. Appl. Microbiol. Biotechnol. 91, 565-576.

Hartmanis, M.G.N., Klason, T., Gatenbeck, S., 1984. Uptake and activation of acetate

and butyrate in Clostridium acetobutylicum. Appl. Microbiol. Biotechnol. 20, 66-71.

Harwood, C.S., Burchhardt, G., Herrmann, H., Fuchs, G., 1998. Anaerobic metabolism of

aromatic compounds via the benzoyl-CoA pathway. FEMS Microbiol. Rev. 22, 439-

458.

Hedges, J.I., Ertel, J.R., 1982. Characterization of lignin by gas capillary chromatography

of cupric oxide oxidation products. Anal. Chem. 54, 174-178.

Heer, D., Heine, D., Sauer, U., 2009. Resistance of Saccharomyces cerevisiae to high

concentrations of furfural is based on NADPH-dependent reduction by at least two

oxireductases. Appl. Environ. Microbiol. 75, 7631-7638.

Heller, M.J., 2002. DNA microarray technology: devices, systems, and applications.

Annu. Rev. Biomed.Eng. 4, 129-153.

Hess, G., 2006. BP and DuPont to Make ‘Biobutanol,’. Chem. Eng. News. 84, 9-10.

Huang, D.W., Sherman, B.T., Lempicki, R.A., 2009. Systematic and integrative analysis

of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44-57.

Jarboe, L.R., 2011. YqhD: a broad-substrate range aldehyde reductase with various

applications in production of biorenewable fuels and chemicals. Appl. Microbiol.

Biotechnol. 89, 249-257.

Jones, D.T., Woods, D.R., 1986. Acetone-butanol fermentation revisited. Microbiol. Rev.

50, 484-524.

Jönsson, L., Palmqvist, E., Nilvebrant, N.O., Hahn-Hägerdal, B., 1998. Detoxification of

wood hydrolysates with laccase and peroxidase from the white-rot fungus Trametes

versicolor. Appl. Microbiol. Biotechnol. 49, 691-697.

Klinke, H.B., Thomsen, A.B., Ahring, B.K., 2004. Inhibition of ethanol-producing yeast

and bacteria by degradation products produced during pre-treatment of biomass. Appl.

Microbiol. Biotechnol. 66, 10-26.

Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of

lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng.

Chem. Res. 48, 3713-3729.

Lamminp , K., Ahola, J., Tanskanen, J., 2012. Kinetics of xylose dehydration into

furfural in formic acid. Ind. Eng. Chem. Res. 51, 6297-6303.

Larroy, C., Fernández, M.R., González, E., Parés, X., Biosca, J.A., 2002a.

Characterization of the Saccharomyces cerevisiae YMR318C (ADH6) gene product

as a broad specificity NADPH-dependent alcohol dehydrogenase: relevance in

aldehyde reduction. Biochem. J. 361, 163.

Larroy, C., Parés, X., Biosca, J.A., 2002b. Characterization of a Saccharomyces

cerevisiae NADP (H)‐dependent alcohol dehydrogenase (ADHVII), a member of the

cinnamyl alcohol dehydrogenase family. Eur. J. Biochem. 269, 5738-5745.

Lee, J., Blaschek, H.P., 2001. Glucose uptake in Clostridium beijerinckii NCIMB 8052

and the solvent-hyperproducing mutant BA101. Appl. Environ. Microbiol. 67, 5025-

5031.

Leonard, R.H., Hajny, G.J., 1945. Fermentation of Wood Sugars to Ethyl Alcohol. Ind.

Eng. Chem. 37, 390-395.

Lin, F.M., Qiao, B., Yuan, Y.J., 2009. Comparative proteomic analysis of tolerance and

adaptation of ethanologenic Saccharomyces cerevisiae to furfural, a lignocellulosic

inhibitory compound. Appl. Environ. Microbiol. 75, 3765-3776.

Liu, J.S., 2002. Bioinformatics: microarrays analyses and beyond. Amstat News. 59-67.

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxification,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Liu, Z.L., Moon, J., Andersh, B.J., Slininger, P.J., Weber, S., 2008. Multiple gene-

mediated NAD (P) H-dependent aldehyde reduction is a mechanism of in situ

detoxification of furfural and 5-hydroxymethylfurfural by Saccharomyces cerevisiae.

Appl. Microbiol. Biotechnol. 81, 743-753.

Liu, Z.L., 2011. Molecular mechanisms of yeast tolerance and in situ detoxification of

lignocellulose hydrolysates. Appl. Microbiol. Biotechnol. 90, 809-825.

Liu, Z.L., Slininger, P.J., Dien, B.S., Berhow, M.A., Kurtzman, C.P., Gorsich, S.W.,

2004. Adaptive response of yeasts to furfural and 5-hydroxymethylfurfural and new

chemical evidence for HMF conversion to 2,5-bis-hydroxymethylfuran. J. Ind.

Microbiol. Biotechnol. 31, 345-352.

Lopez, M., Nichols, N., Dien, B., Moreno, J., Bothast, R., 2004. Isolation of

microorganisms for biological detoxification of lignocellulosic hydrolysates. Appl.

Microbiol. Biotechnol. 64, 125-131.

Lucia, L.A., 2008. Lignocellulosic biomass: A potential feedstock to replace petroleum.

BioResources. 3, 981-982.

Martinez, A., Rodriguez, M.E., Wells, M.L., York, S.W., Preston, J.F., Ingram, L.O.,

2001. Detoxification of dilute acid hydrolysates of lignocellulose with lime.

Biotechnol. Prog. 17, 287-293.

Mialon, L., Pemba, A.G., Miller, S.A., 2010. Biorenewable polyethylene terephthalate

mimics derived from lignin and acetic acid. Green Chem. 12, 1704-1706.

Mitchell, W.J., 1997. Physiology of carbohydrate to solvent conversion by clostridia.

Adv. Microb. Physiol. 39, 31-130.

Mitchell, W., 1996. Carbohydrate Uptake and Utilization by Clostridium beijerinckii

NCIMB 8052. Anaerobe. 2, 379-384.

Moreau, Y., De Smet, F., Thijs, G., Marchal, K., De Moor, B., 2002. Functional

bioinformatics of microarray data: from expression to regulation. Proc. IEEE. 90,

1722-1743.

Morey, J.S., Ryan, J.C., Van Dolah, F.M., 2006. Microarray validation: factors

influencing correlation between oligonucleotide microarrays and real-time PCR. Biol.

Proc. Online. 8, 175-193.

Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y., Holtzapple, M., Ladisch, M.,

2005. Features of promising technologies for pretreatment of lignocellulosic biomass.

Bioresour. Technol. 96, 673-686.

Mussatto, S.I., Roberto, I.C., 2004. Alternatives for detoxification of diluted-acid

lignocellulosic hydrolyzates for use in fermentative processes: a review. Bioresour.

Technol. 93, 1-10.

Nakas, J., Schaedle, M., Parkinson, C., Coonley, C., Tanenbaum, S., 1983. System

development for linked-fermentation production of solvents from algal biomass. Appl.

Environ. Microbiol. 46, 1017-1023.

Ni, Y., Sun, Z., 2009. Recent progress on industrial fermentative production of acetone–

butanol–ethanol by Clostridium acetobutylicum in China. Appl. Microbiol.

Biotechnol. 83, 415-423.

Palmqvist, E., Hahn-Hägerdal, B., 2000. Fermentation of lignocellulosic hydrolysates. II:

inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 25-33.

Parajó, J.C., Dominguez, H., Domínguez, J.M., 1998. Biotechnological production of

xylitol. Part 3: Operation in culture media made from lignocellulose hydrolysates.

Bioresour. Technol. 66, 25-40.

Qureshi, N., Blaschek, H., 2001. Recent advances in ABE fermentation: hyper-butanol

producing Clostridium beijerinckii BA101. J. Ind. Microbiol. Biotechnol. 27, 287-

291.

Qureshi, N., Ezeji, T.C., 2008. Butanol,‘a superior biofuel’production from agricultural

residues (renewable biomass): recent progress in technology. Biofuels Bioprod.

Bioref. 2, 319-330.

Qureshi, N., Maddox, I.S., Friedl, A., 1992. Application of continuous substrate feeding

to the ABE fermentation: relief of product inhibition using extraction, perstraction,

stripping, and pervaporation. Biotechnol. Prog. 8, 382-390.

Qureshi, N., Maddox, I., 1992. Application of novel technology to the ABE fermentation

process. Appl. Biochem. Biotechnol. 34, 441-448.

Ranatunga, T.D., Jervis, J., Helm, R.F., McMillan, J.D., Hatzis, C., 1997. Identification

of inhibitory components toxic toward Zymomonas mobilis CP4 (pZB5) xylose

fermentation. Appl. Biochem. Biotechnol. 67, 185-198.

Richmond, C., Ujor, V., Ezeji, T.C., 2012. Impact of syringaldehyde on the growth of

Clostridium beijerinckii NCIMB 8052 and butanol production. 3 Biotech. 2, 159-167.

Saldanha, A.J., 2004. Java Treeview—extensible visualization of microarray data.

Bioinformatics. 20, 3246-3248.

Salomonis, N., Hanspers, K., Zambon, A.C., Vranizan, K., Lawlor, S.C., Dahlquist, K.D.,

Doniger, S.W., Stuart, J., Conklin, B.R., Pico, A.R., 2007. GenMAPP 2: new features

and resources for pathway analysis. BMC Bioinformatics. 8, 217.

Smyth, G.K., Yang, Y.H., Speed, T., 2003. Statistical issues in cDNA microarray data

analysis. Methods Mol. Biol. 224, 111-136.

Stoughton, R.B., 2005. Applications of DNA microarrays in biology. Annu. Rev.

Biochem. 74, 53-82.

Sun, Y., Cheng, J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a

review. Bioresour. Technol. 83, 1-11.

Sutton, G.G., White, O., Adams, M.D., Kerlavage, A.R., 1995. TIGR Assembler: A new

tool for assembling large shotgun sequencing projects. Genome Sci. Technol. 1, 9-19.

Taherzadeh, M.J., Gustafsson, L., Niklasson, C., Liden, G., 2000. Physiological effects of

5-hydroxymethylfurfural on Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol.

53, 701-708.

Xiao, H., Gu, Y., Ning, Y., Yang, Y., Mitchell, W.J., Jiang, W., Yang, S., 2011.

Confirmation and elimination of xylose-metabolic bottlenecks in glucose-PTS-

deficient Clostridium acetobutylicum to realize simultaneous utilization of glucose,

xylose and arabinose. Appl. Environ. Microbiol. 77 (22), 7886-95.

Yacobucci, B.D., Schnepf, R.D., Library of Congress. Congressional Research Service,

2007. Ethanol and biofuels: agriculture, infrastructure, and market constraints related

to expanded production.

Zaldivar, J., Martinez, A., Ingram, L.O., 2000. Effect of alcohol compounds found in

hemicellulose hydrolysate on the growth and fermentation of ethanologenic

Escherichia coli. Biotechnol. Bioeng. 68, 524-530.

Zaldivar, J., Martinez, A., Ingram, L.O., 1999. Effect of selected aldehydes on the growth

and fermentation of ethanologenic Escherichia coli. Biotechnol. Bioeng. 65, 24-33.

Zhao, Y., Tomas, C.A., Rudolph, F.B., Papoutsakis, E.T., Bennett, G.N., 2005.

Intracellular butyryl phosphate and acetyl phosphate concentrations in Clostridium

acetobutylicum and their implications for solvent formation. Appl. Environ.

Microbiol. 71, 530-537.

Zheng, Y., Pan, Z., Zhang, R., 2009. Overview of biomass pretreatment for cellulosic

ethanol production. Int. J. Agric. Biol. Eng. 2, 51-68.

Figure 2.1 Schematic representation on pretreatment of lignocellulosic.

Figure 2.2 Production of sugars and microbial inhibitors during lignocellulosic biomass pretreatment and hydrolysis.

(a)

(b)

Figure 2.3 The formation of furfural (a) and HMF (b) during the pretreatment of lignocellulosic biomass.

Figure 2.4 Simplified ABE production pathway in solventogenic clostridia. 1, glucose uptake by the phosphotransferase system (PTS) and conversion to pyruvate by the Emden–Meyerhof–Parnas pathway; 2, pyruvate-ferrodoxin oxidoreductase; 3, thiolase or acetyl-CoA acetyltransferase; 4, 3-hydroxybutyryl-CoA dehydrogenase; 5, crotonase; 6, butyryl-CoA dehydrogenase; 7, phosphate butyltransferase (phosphotrans- butyrylase);8, butyrate kinase; 9, phosphate acetyltransferase (phosphostransacetylase); 10, acetate kinase; 11, butyraldehyde dehydrogenase and alcohol/aldehyde dehydrogenase; 12, butanol dehydrogenase; 13, acetoacetyl-CoA:acetate/butyrate:CoA transferase;14, acetoacetate decarboxylase; 15, acetaldehyde dehydrogenase; 16, ethanol dehydrogenase.

Chapter 3: Acetone Butanol Ethanol Production from Miscanthus giganteus by Clostridium beijerinckii NCIMB 8052

3.1 Abstract

Inhibition of fermenting microorganisms by degradation products of lignocellulosic biomass are increasingly recognized as one of the major limitations of the biomass conversion to biofuel. Towards the understanding of toxicity of lignocellulose- derived inhibitors and tolerance of solventogenic clostridia, physiological alterations of fermentation by Clostridium beijerinckii NCIMB 8052 using hydrolysate of Miscanthus giganteus was investigated. Fermentation P2 medium containing Miscanthus giganteus hydrolysates, even as low as 10% (v/v), remarkably inhibited cell growth of C. beijerinckii 8052, and delayed fermentation by 12 h. Although levels of acetone butanol ethanol (ABE) produced by C. beijerinckii 8052 in the control medium (P2 medium without lignocellulosic biomass hydrolysates) and medium with hydrolysates (10%, v/v) are similar, C. beijerinckii 8052 was unable to achieve solventogenesis when grown in P2 medium containing ≥ 25% (v/v) Miscanthus hydrolysates. Additionally, C. beijerinckii

8052 did not grow in P2 medium with ≥ 40% (v/v) Miscanthus hydrolysates.

Nonetheless, C. beijerinckii 8052 detoxified lignocellulose-derived microbial inhibitory compounds such as furfural, hydroxymethylfurfural (HMF), 4-hydroxybenzaldehyde, and p-coumaric acid in Miscanthus hydrolysates during ABE fermentation. Additionally, it was determined that supplementation of Miscanthus hydrolysates (100%, v/v) based P2

medium with CaCO3 enabled C. beijerinckii to utilize undetoxified and undiluted

Miscanthus hydrolysates for growth and ABE production. Collectively, detoxification of lignocellulose-derived microbial inhibitory compounds by C. beijerinckii, buffering role of CaCO3, and its potential stabilization of C. beijerinckii metabolism during fermentation using Miscanthus hydrolysates are discussed.

3.2 Introduction

Researches on biofuel production may help reduce the use of fossil fuels and therefore alleviate emission of greenhouse gases and global energy crisis. Conventional biofuel production has limitations including low energy content and high cost of feedstock. To overcome these challenges, butanol production from lignocellulosic biomass has shown promises since the feedstock is the most abundant and relatively cheap renewable material (Ezeji et al., 2007; Green, 2011). Lignocellulosic biomass refers to plant biomass composed of cellulose, hemicelluloses and lignin, for example, corn stover, wheat straw, branches, foliage, as well as dedicated energy crops such as

Miscanthus and switchgrass (Singh et al., 2010). Using lignocellulosic biomass for biofuel production offers many benefits such as abundance and diverse raw material compared to sources such as corn and sugarcane, and potential reduction of market stress due to use of food crops for biofuels production, and increase in net energy yield

(Hamelinck and Faaij, 2006). To overcome the recalcitrance of lignocellulosic biomass and release fermentable sugars, appropriate pretreatment process is required to remove lignin, depolymerize hemicelluloses, and reduce the crystallinity of cellulose, however, numerous microbial inhibitors are generated during the pretreatment and hydrolysis of lignocellulosic biomass (Kumar et al., 2009). These toxic compounds may have large

negative effects to butanol production mainly due to inhibition of cell growth or metabolites production by cell membrane disruption, nucleic acid damage, and enzyme inhibition (Mills et al., 2009). Although several inhibitor removal methods, including physical, chemical, or biological, may facilitate cell growth and fermentation, the removal of inhibitors from hydrolysates prior to fermentation may not be an economically feasible approach, considering the cost of additional processing steps and potential loss of fermentable sugars during inhibitors removal (Liu and Blaschek, 2010).

Towards the understanding of toxicity of lignocellulose hydrolysate to Clostridium species, the objective of this study was to examine how Miscanthus giganteus hydrolysates influence cell growth and ABE production by C. beijerinckii 8052, and how to mitigate their toxicity to cell metabolism.

To study inhibitory characteristics of lignocellulose hydrolysates, ABE fermentation by C. beijerinckii 8052 was conducted using hydrolysates of Miscanthus giganteus generated by liquid hot water (LHW) pretreatment. By analyzing inhibitory compounds during fermentation, the inherent capacity of C. beijerinckii 8052 to detoxify furfural, hydroxymethylfurfural (HMF), 4-hydroxybenzaldehyde, and p-coumaric acid was discovered. Toward alleviating toxicity of Miscanthus hydrolysates, the benefit of fermentation-broth additive, CaCO3, was evaluated. Notably, CaCO3 increases stabilization of biosynthesis and growth machinery of C. beijerinckii 8052 leading to a robust growth, and therefore improves cell tolerance to environmental stress (Han et al.,

2013). These findings provide new insights and approaches for biofuel productions from undetoxified lignocellulosic biomass which is crucial for making large scale bio-butanol production from biomass economically feasible.

3.3 Materials and Methods

3.3.1 Microorganism and culture conditions

Clostridium beijerinckii NCIMB 8052 was obtained from the American Type

Culture Collection, Manassas, VA. C. beijerinckii 8052 stocks were maintained as spores in sterile double distilled water at 4°C. Spores of C. beijerinckii 8052 were heat shocked at 75°C for 10 min, cooled on ice, and then inoculated into 10-ml anoxic tryptone– glucose–yeast extract (TGY) medium (Zhang et al., 2012). The culture was incubated anaerobically at 35±1°C for 12 to 14 h until OD600 attained 0.9–1.1. Then 8% (v/v) of the actively growing culture was transferred into TGY medium and grown for 4 to 5 h until

OD600 0.9–1.1.

3.3.2 Pretreatment of Miscanthus giganteus

Prior to pretreatment, Miscanthus giganteus was ground using a Thomas-Wiley mill and 1 mm sieve (Thomas Scientific, Swedesboro, NJ). The moisture content was measured by an infrared moisture analyzer (Mettler Toledo, Columbus, OH) according to

NREL Laboratory Analytical Procedure (LAP) (Hames et al., 2008), which is necessary to calculate the amount of water to be mixed to generate different percentages of dry solid loading. The 119.8 g of biomass material, with a moisture content of 9.87%, was mixed with 600 ml distilled or de-ionized water to obtain 15% (w/v) solids loading, calculated on weight/weight basis. The liquid hot water (LHW) pretreatment was carried out in Parr Reactor (Parr, Moline, IL) at 190°C for 20 min. After cooling down to 25 °C, hydrolysates slurry was neutralized to pH 6.0 with ammonium hydroxide. Enzymatic hydrolysis of the pretreated Miscanthus was carried out by adding cellulase (15 FPU/g cellulose), β-glucosidase (Novozyme 188) (40 U/g cellulose), and xylanase (0.25

g/100ml) as described previously (Ezeji et al., 2007). Enzymatic hydrolysis was carried out for 72 h at 50 °C. Pretreated material was stored at – 20 °C until analysis and fermentation experiments.

3.3.3 Acetone butanol ethanol production from Miscanthus giganteus hydrolysates by C. beijerinckii 8052

Batch fermentations were performed in 150-mL Pyrex screw-capped media bottles with C. beijerinckii 8052 in triplicate. The P2 medium containing 46 g/L glucose and 14 g/L xylose, instead of 60g/L glucose, was used as the control medium. For

Miscanthus hydrolysates treatment medium, 10%, 25% or 40% (v/v) Miscanthus hydrolysates was used for the ABE fermentation. The treatment medium was supplemented with glucose and xylose to ensure the same amount of substrate concentration as the control. All fermentation media were supplemented with 1% (v/v) of sterile P2 medium stock solutions (buffer, mineral and vitamin) (Zhang et al., 2012) before inoculation with 6% (v/v) of C. beijerinckii 8052 preculture. To evaluate the effect of CaCO3 on ABE fermentation using Miscanthus hydrolysates as substrate, 4 g/L CaCO3 was added to control medium, 10%, 25%, 50% and 100% Miscanthus hydrolysates P2 based medium. Pyrex screw capped media bottles containing 100 mL fermentation medium were autoclaved at 121°C for 15 min. During the course of fermentation, 5-mL samples were collected every 12 h for optical density, pH, ABE, and acid analysis.

3.3.4 Analytical methods

Concentrations of inhibitors in the samples were determined with a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA). The maximum absorbance spectra of inhibitors were at 276, 220, 282, 222, 277, and 284 nm for furfural, furfuryl alcohol,

HMF, HMF alcohol, 4-hydroxybenzaldehyde and p-coumaric acid, respectively. Residual concentrations of inhibitors in the samples were calculated using standards of known concentrations. Residual concentrations of furfural, and HMF in the sample were confirmed by HPLC equipped with a photodiode array (PDA) detector (Waters, Milford,

MA) and a 3.5 μm Xbridge C18, 150 mm × 4.6 mm column (Waters, Milford, MA) as previous work (Zhang et al., 2012). Sugars were analyzed by HPLC equipped with evaporative light scattering (ELSD) detector (Waters, Milford, MA) and a 9 μm Aminex

HPX-87P, 300 mm × 7.8 mm column with a 4.6 mm ID × 3 cm long Aminex deashing guard column (Bio-Rad, Hercules, CA) maintained at 65°C. The mobile phase was

HPLC-grade water operated at a flow rate of 0.6 mL/min (Zhang et al., 2012). Samples were filtered through 0.45 μm syringe filters before injections. The concentrations of

ABE and acids were quantified using a gas chromatography system (Agilent

Technologies 7890A, Agilent Technologies Inc., Wilmington, DE), equipped with a flame ionization detector (FID) and 30 m (length) × 320 μm (internal diameter) × 0.50 μm (HP-Innowax film) J × W 19091N-213 capillary column (Han et al.,

2013).

3.4 Results and Discussion

3.4.1 Pretreatment and hydrolysis of Miscanthus giganteus

To release simple sugars from Miscanthus for microbial fermentation, the liquid hot water (LHW) pretreatment was used, and composition of Miscanthus giganteus hydrolysates was analyzed for monomeric sugars and biomass degradation products using the HPLC as described in the materials and methods section above. LHW pretreatment allows biomass undergo high pressure cooking at elevated temperatures (190 °C in this

study) without any addition of acids, which reduces cost of catalyst as well as chemicals used for neutralization, and may produce less amounts of lignocellulose-derived microbial inhibitors compared to pretreatment with acidic catalysts (Brodeur et al., 2011).

The main sugars generated after pretreatment and hydrolysis of Miscanthus giganteus were glucose, xylose and arabinose from depolymerization of cellulose and hemicelluloses (Table 3.1). Glucose and xylose yield was 0.22 g and 0.08 g, respectively, per gram of Miscanthus giganteus biomass (Table 3.2), which is lower than the reported value (0.44 g glucose and 0.21 g xylose per gram of biomass) by using dilute acid presoaking combined with wet explosion and enzymatic treatment (Sørensen et al.,

2008). These results indicate that LHW pretreatment at 190 °C followed by enzymatic hydrolysis is not sufficient to recover more than 50% of the sugars, and thus, harsher conditions such as temperatures and process times greater than 190 °C and 20 min or addition of acids or alkaline in order to increase sugar yield may be required. With the aim of studying inhibitory effects of biomass hydrolysate, however, this pretreatment condition generated a good amount of biomass degradation products. Potential microbial inhibitors produced from lignocellulosic biomass include sugar degradation products (e.g. furfural and HMF from xylose and glucose, respectively) (Palmqvist and Hahn-Hägerdal,

2000) and lignin-hemicellulose decomposition (e.g. 4-hydroxybenzaldehyde, syringic acid, syringaldehyde, p-coumaric acid and ferulic acid) (Hedges and Ertel, 1982), and concentrations of these inhibitory compounds are listed in Table 3.1.

3.4.2 Inhibition on the growth of and ABE production by C. beijerinckii 8052 using Miscanthus giganteus hydrolysates

To evaluate the cell growth and ABE production profile by C. beijerinckii 8052 using Miscanthus as the substrate, batch fermentations were performed using medium containing 10%, 25% and 40% of Miscanthus hydrolysate. Severe inhibition of C. beijerinckii 8052 growth was observed when dilute Miscanthus hydrolysates was used as carbon source (Figure 3.1A). Medium supplemented with 10% and 25% Miscanthus giganteus hydrolysates resulted in approximately 40% and 30% cell growth, respectively, compared with control medium without Miscanthus giganteus hydrolysates. When

Miscanthus giganteus hydrolysates content of the fermentation medium was increased to

40% (v/v), further inhibition of C. beijerinckii 8052 growth was observed (Figure 3.1A).

Although the growth of C. beijerinckii 8052 was inhibited when grown in medium with

10% Miscanthus giganteus hydrolysates, the amount of ABE (11.8 g/L) produced was similar to that produced in the control medium (11.4 g/L) (Figure 3.1B and C). However,

ABE production by C. beijerinckii 8052 in medium containing 10% Miscanthus giganteus hydrolysates was delayed by 12 h (Figure 3.1B and C). It is likely that during this time C. beijerinckii 8052 transformed some of the inhibitors such as furfural, HMF,

4-hydroxybenzaldehyde, syringic acid, syringaldehyde, p-coumaric acid (Table 3.3) into less toxic compounds. Differently, 25% of Miscanthus hydrolysate repressed the growth of C. beijerinckii 8052 and completely inhibited ABE production by C. beijerinckii 8052

(Figure 3.1D). Under this treatment, no acetic acid uptake was observed, and there was a built-up of butyric acid in the fermentation medium (Figure 3.1D). Also, no ethanol was produced, and final ABE concentration (0.5 g/L) was produced by C. beijerinckii 8052

(Figure 3.1D), unlike ABE production in the control medium (Figure 3.1B).

Notably, inhibition of C. beijerinckii by Miscanthus giganteus hydrolysates is due to degradation products generated during pretreatment and hydrolysis of Miscanthus giganteus. The variety and concentration of biomass degradation products may vary significantly due to the natural composition of biomass resources as well as the pretreatment conditions. Until now, relatively smaller number of lignocellulosic derived inhibitory compounds has been identified although hundreds or more may generated during acid pretreatment of biomass (Liu and Blaschek, 2010). This is one of the reasons why the degree of inhibition was so high in this study even though diluted hydrolysate was used during fermentation process. Besides, although composition analysis of hydrolysate shows low levels of identified inhibitory compounds, combination of these compounds may present additive and synergistic toxicity to C. beijerinckii 8052.

Since ABE productions by C. beijerinckii 8052 were similar between the control and the medium containing 10% Miscanthus giganteus hydrolysates, it was hypothesized that at a sublethal concentration, C. beijerinckii has an inherent capacity to tolerate and detoxify some lignocellulosic derived microbial inhibitory compounds during ABE fermentation. To test this hypothesis, concentrations of biomass degradation products in

Miscanthus giganteus hydrolysates were measured before and 12 h after ABE fermentation by C. beijerinckii 8052. Interestingly, out of seven compounds analyzed, about four (furfural, HMF, 4-hydroxybenzaldehyde, and p-coumaric acid) were depleted within 12 h of fermentation (Table 3.3).

3.4.3 Mitigation of lignocellulose-derived microbial inhibitors by CaCO3

Calcium carbonate has been widely used as additives to improve fermentation by yeast (Wilkins et al., 2007), fungus (Ohta et al., 1993), and bacteria (Jiang et al., 2009),

mainly owing to its regulation of cultural pH. Moreover, previous studies have demonstrated some benefits of supplementation of fermentation medium with CaCO3 such as increase in buffering capacity of the medium, stabilization on cellular activity, and enhanced tolerance to microbial inhibitory products such as butanol (Kanouni et al.,

1998; Han et al., 2013). In light of these findings, it was hypothesized that addition of

CaCO3 to the fermentation broth would mitigate growth and fermentation inhibition caused by lignocellulose-derived microbial inhibitory compounds and consequently, facilitate C. beijerinckii 8052 growth and ABE production. To test this hypothesis, batch

ABE fermentation by C. beijerinckii 8052 using the P2 medium supplemented with 4 g/L

CaCO3, and 10%, 25%, 50% or 100% Miscanthus hydrolysates was conducted. Figure

3.2A shows that in the presence of CaCO3, P2 medium supplemented with Miscanthus hydrolysates supported the growth of C. beijerinckii 8052 and maximum OD600 ranging from 3.7 to 5.9 was attained at fermentation time of 24 h (Figure 3.2A). Moreover, growth of C. beijerinckii 8052 in P2 medium supplemented with 10-50% Miscanthus hydrolysates achieved a higher maximum cell density than that of control medium without CaCO3 (Figure 3.1A). It should be noted that without CaCO3 C. beijerinckii 8052 did not grow in P2 medium supplemented with 40% Miscanthus hydrolysates (Figure

3.1A). ABE production using P2 medium supplemented with Miscanthus hydrolysates

(10 – 100%) and 4 g/L CaCO3 ranged from 8.3 to 17.3 g/L (Figure 3.2B-F), while without CaCO3, even 25% Miscanthus hydrolysates supplemented to P2 led to no ABE production (Figure 3.1D). In addition, the concentration of acetic and butyric acids was built up in P2 medium supplemented with increasing amount of Miscanthus hydrolysates

(Figure 3.2B-F). Notably, fermentation broth with undiluted Miscanthus hydrolysates

accumulated 8 g/L acids (acetic and butyric acids) (Figure 3.2F), while only less than 1 g/L acids remained in the control medium (Figure 3.2B). Improvement of cell growth and solvent production indicates that CaCO3 plays an important role in overcoming the synergistic toxicity of biomass degradation products. Buffering capacity is one of the reasons that CaCO3 could facilitate ABE fermentation using lignocellulose hydrolysate as substrate, because it can neutralize organic acids produced after biomass hydrolysis.

Furthermore, Ca2+ has been demonstrated to increase the expression of heat-shock proteins, enhance sugar utilization, induce oxidative responses, stabilize cell membrane and nucleic acids, and increase activity of solvent production enzymes in C. beijerinckii

8052 (Han et al., 2013). It is plausible that C. beijerinckii 8052 was able to grow and produce ABE in P2 medium supplemented with 100% Miscanthus hydrolysates due to amelioration of toxic effects of lignocellulose-derived microbial inhibitory compounds on the microorganism and induction of cell stabilization proteins in C. beijerinckii 8052 by

2+ CaCO3 and Ca .

3.5 Conclusion

Pretreatment and hydrolysis of Miscanthus giganteus generate of fermentable sugars and many other microbial inhibitors such as organic acids, aldehydes, and phenolics. Growth of and ABE production by C. beijerinckii 8052 was completely inhibited by medium supplemented with more than 40% (v/v) Miscanthus giganteus hydrolysates, while medium supplemented with 25% hydrolysates resulted in a suppressed cell growth and no ABE production. C. beijerinckii 8052 was determined to be able to detoxify some lignocellulose-derived microbial inhibitory compounds such as furfural, HMF, 4-hydroxybenzaldehyde, and p-coumaric acid during ABE fermentation

using the Miscanthus giganteus hydrolysates as substrate. In addition, CaCO3 was determined to ameliorate toxic effects of lignocellulose-derived microbial inhibitory compounds on C. beijerinckii 8052 and consequently, improved bioconversion of

Miscanthus giganteus hydrolysates to ABE. To develop inhibitor-tolerant strains of C. beijerinckii 8052, understanding mechanisms with which C. beijerinckii 8052 uses to transform or tolerate these lignocellulose-derived inhibitory compounds is necessary.

References

Brodeur, G., Yau, E., Badal, K., Collier, J., Ramachandran, K., Ramakrishnan, S., 2011.

Chemical and Physicochemical Pretreatment of Lignocellulosic Biomass: A Review.

Enz. Research. 2011, 1-17.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Green, E.M., 2011. Fermentative production of butanol—the industrial perspective. Curr.

Opin. Biotechnol. 22, 337-343.

Hamelinck, C.N., Faaij, A.P.C., 2006. Outlook for advanced biofuels. Energy Policy. 34,

3268-3283.

Hames, B., Hyman, D., Payne, C., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Nrel,

J.W., 2008. Determination of total solids in biomass and total dissolved solids in

liquid process samples. Laboratory Analytical Procedure (LAP). NREL/TP-510–

42621. National Renewable Energy Laboratory.

Han, B., Ujor, V., Lai, L.B., Gopalan, V., Ezeji, T.C., 2013. Use of Proteomic Analysis

To Elucidate the Role of Calcium in Acetone-Butanol-Ethanol Fermentation by

Clostridium beijerinckii NCIMB 8052. Appl. Environ. Microbiol. 79, 282-293.

Hedges, J.I., Ertel, J.R., 1982. Characterization of lignin by gas capillary chromatography

of cupric oxide oxidation products. Anal. Chem. 54, 174-178.

Jiang, Y., Xu, C., Dong, F., Yang, Y., Jiang, W., Yang, S., 2009. Disruption of the

acetoacetate decarboxylase gene in solvent-producing Clostridium acetobutylicum

increases the butanol ratio. Metab. Eng. 11, 284-291.

Kanouni, A.E., Zerdani, I., Zaafa, S., Znassni, M., Loutfi, M., Boudouma, M., 1998. The

improvement of glucose/xylose fermentation by Clostridium acetobutylicum using

calcium carbonate. World J. Microbiol. Biotechnol. 14, 431-435.

Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of

lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng.

Chem. Res. 48, 3713-3729.

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxification,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and

tolerance mechanisms in Escherichia coli. Biotechnol. Biofuels. 2, 26.

Ohta, K., Hamada, S., Nakamura, T., 1993. Production of high concentrations of ethanol

from inulin by simultaneous saccharification and fermentation using Aspergillus

niger and Saccharomyces cerevisiae. Appl. Environ. Microbiol. 59, 729-733.

Palmqvist, E., Hahn-Hägerdal, B., 2000. Fermentation of lignocellulosic hydrolysates. II:

inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 25-33.

Singh, A., Pant, D., Korres, N.E., Nizami, A.S., Prasad, S., Murphy, J.D., 2010. Key

issues in life cycle assessment of ethanol production from lignocellulosic biomass:

challenges and perspectives. Bioresour. Technol. 101, 5003-5012.

Sørensen, A., Teller, P.J., Hilstrøm, T., Ahring, B.K., 2008. Hydrolysis of Miscanthus for

bioethanol production using dilute acid presoaking combined with wet explosion pre-

treatment and enzymatic treatment. Bioresour. Technol. 99, 6602-6607.

Wilkins, M.R., Widmer, W.W., Grohmann, K., 2007. Simultaneous saccharification and

fermentation of citrus peel waste by Saccharomyces cerevisiae to produce ethanol.

Proc. Biochem. 42, 1614-1619.

Zhang, Y., Han, B., Ezeji, T.C., 2012. Biotransformation of furfural and 5-

hydroxymethyl furfural (HMF) by Clostridium acetobutylicum ATCC 824 during

butanol fermentation. N. Biotechnol. 29, 345-351.

Table 3.1 Analysis of sugars and inhibitors production after pretreatment and hydrolysis of Miscanthus giganteus. Liquid hot water (LHW) pretreatment of Miscanthus giganteus (15% dry mass loading) was conducted at190°C for 20 min. After pH adjustment to 5.5 by ammonia addition, enzymatic hydrolysis were performed by adding cellulase (15 FPU/g cellulose), beta- glucosidase (Novozyme 188) (40 U/g cellulose), and xylanase (0.25g/100ml), followed by incubating for 72 h at 50 °C with agitation at 200 rpm. Sugars and inhibitors were analyzed by HPLC equipped with Aminex HPX-87P column (Bio-Rad, Hercules, CA) and Xbridge C18 column (Waters, Milford, MA), respectively.

Concentration Concentration Sugars Inhibitors (g/L) (mg/L)

HMF 176±0.58 Glucose 38.84±0.30 Furfural 1459±1.00

4-Hydroxybenzaldehyde 63±0.00 Xylose 14.54±0.10 Syringic acid 29±0.00

Syringaldehyde 42±0.58 Arabinose 4.33±0.00 p-Coumaric acid 219±1.00

Total 57.71±0.21 Ferulic acid 247±2.52

Table 3.2 Estimation of glucose and xylose yield of Miscanthus giganteus hydrolysates produced from liquid hot water pretreatment and enzymatic hydrolysis. Sugar concentration was measured by HPLC equipped with Aminex HPX-87P column (Bio-Rad, Hercules, CA). The amount of glucose and xylose obtained in hydrolysate was calculated as the concentration of sugar concentration multiplied by volume of hydrolysate. Sugar yield was estimated by the amount of glucose and xylose obtained in hydrolysate per gram of loaded biomass.

Glucose Xylose

Volume of hydrolysate (L) 0.68 Sugar Conc. (g/L) 38.84 14.54 Sugars in hydrolysate (g) 26.41 9.89 Loaded biomass (g) 119.80 Sugar yield (g/g biomass) 0.22 0.08

Table 3.3 Analysis of lignocellulose-derived microbial inhibitors at 0 h and 12 h of fermentation by C. beijerinckii 8052 using P2 medium supplemented with subleathal level of Miscanthus giganteus hydrolysates.

Concentration Concentration Inhibitors (mg/L) at 0 h (mg/L) at 12 h HMF 21 ND Furfural 236 ND 4-Hydroxybenzaldehyde 15 ND Syringic acid 9 7 Syringaldehyde 13 12 Coumaric acid 40 ND Ferulic acid 46 42 ND: Not detected.

A B

C D

Figure 3.1 Cell growth and ABE production by C. beijerinckii 8052 using Miscanthus giganteus hydrolysate as substrate. Batch ABE fermentation by C. beijerinckii 8052 was conducted using control mediumP2 medium (B), supplemented with 10% (C) and 25% (D) of Miscanthus giganteus hydrolysate as substrate. Cell growth (A) was estimated by optical density at 600 nm; solvent production was measured by gas chromatography, and are represented (B-D) as follows: Acetone, solid circles; Ethanol, empty circles; Butanol, solid triangles; Acetic acid, empty triangles; Butyric acid, solid squares.

A B

C D

E F

Figure 3.2 Cell growth and ABE production by C. beijerinckii 8052 using Miscanthus giganteus hydrolysate as substrate supplemented with 4 g/L CaCO3. Cell growth (A) and solvent production profiles during batch ABE fermentation by C. beijerinckii 8052 were obtained using 4 g/L CaCO3 supplemented in P2 medium (B), 10% (C), 25% (D), 50% (D) and 100% (F) of Miscanthus hydrolysate. Products concentrations (B-F) are represented as follows: Acetone, solid circles; Ethanol, empty circles; Butanol, solid triangles; Acetic acid, empty triangles; Butyric acid, solid squares.

Chapter 4: Biotransformation of Lignocellulose-derived Furan Aldehydes and Phenolic Compounds by Clostridium beijerinckii NCIMB 8052 during Butanol Fermentation

4.1 Abstract

Tolerance of lignocellulose-derived microbial inhibitory compounds will enhance efficient bioconversion of lignocellulosic biomass hydrolysates to fuels and chemicals.

The effect of furan derivatives (furfural, furfuryl alcohol, and hydroxymethyl furfural

[HMF]) and phenolic compounds (4-hydroxybenzaldehyde and p-coumaric acid) on butanol production by Clostridium beijerinckii NCIMB 8052 was investigated. Batch acetone butanol ethanol (ABE) fermentations demonstrated that furfural and HMF were transformed to their corresponding alcohols, furfuryl alcohol, and 2,5-bis- hydroxymethylfuran, at a rate of 0.15 and 0.08 g/L/h, respectively, whereas 4- hydroxybenzaldehyde and p-coumaric acid were transformed at a slower rate, 0.03 and

0.05 g/L/h, respectively, by C. beijerinckii 8052. Although C. beijerinckii 8052 challenged with 2 g/L furfural or HMF experienced an extended lag phase, maximum cell density attained and ABE produced were not adversely affected. When fermentation medium was supplemented with ≤ 3 g/L furfuryl alcohol, there was no decrease in the growth of and ABE fermentation by C. beijerinckii 8052. Compared to furan derivatives, phenolic compounds, 4-hydroxybenzaldehyde and p-coumaric acid, significantly inhibited cell growth, acids uptake, and ABE production, even at a very low

concentration of less than 0.5 g/L, indicating higher toxicity to C. beijerinckii 8052. This study improved our understanding on the effect of lignocellulose-derived furan aldehydes and phenolic compounds on butanol fermentation and inherent ability of detoxification by solventogenic clostridia

4.2 Introduction

Solventogenic Clostridium species are able to convert a wide range of substrates to acetone, ethanol and butanol during fermentation, a process commonly called acetone– butanol–ethanol (ABE) fermentation. The cost of butanol production is heavily influenced by the type of substrates used (Qureshi and Blaschek, 2001; Green, 2011).

Renewable sources of sugars such as grasses, wood, agriculture residuals and municipal waste (lignocellulosic biomass) have great potential as fermentation substrates. Efficient utilization of lignocellulosic biomass for the production of fuels and chemicals is limited by the action of biomass-derived microbial inhibitors (Liu and Blaschek, 2010; Ezeji et al., 2007; Ezeji and Blaschek, 2008; Sakai et al., 2007). Pre-fermentation detoxification steps are intended to remove lignocellulosic derived microbial inhibitory compounds and consequently, enhance the growth of fermenting microorganisms, but they significantly increase overall fuel and chemical production costs due to costs associated with the process and potential loss of fermentable sugars during detoxification (Ezeji et al., 2007;

Almeida et al., 2009). Significant improvements in the fermentation of lignocellulosic biomass hydrolysates can be achieved if sugars and biomass-derived microbial inhibitors could be simultaneously metabolized by fermenting microorganisms during fermentation.

Furfural and hydroxymethyl furfural (HMF) are common furan aldehydes derived from dehydration of pentoses and hexoses during pretreatment and hydrolysis of

lignocellulosic biomass (Liu and Blaschek, 2010). Depending on the source of biomass and the type of pretreatment employed, the concentration of furan aldehydes in lignocellulosic hydrolysates can reach 3.5 g/L and 5.9 g/L for furfural and HMF, respectively (Klinke et al., 2004). Their effects on biofuel fermentation by

Saccharomyces cerevisiae (Taherzadeh et al., 2000), Escherichia coli (Liu and Blaschek,

2010) or solventogenic Clostridium species (Ezeji et al., 2007; Zhang et al., 2012) have been investigated previously. These furans are thought to damage cell membranes, inhibit activities of glycolytic enzymes and proteins, and inhibit RNA synthesis in microorganisms (Zaldivar et al., 1999; Liu and Blaschek, 2010). Besides furan aldehydes, phenolic compounds, mainly produced from lignin decomposition, also show inhibitory effects on cell growth and alcohol fermentation, especially low molecular weight phenolics such as 4-hydroxybenzaldehyde and p-coumaric acid (Hedges and Ertel, 1982;

Klinke et al., 2004). The inhibitory effects of phenolic compounds are not as clear as furan aldehydes, mainly because of a lack of accurate qualitative and quantitative analyses (Palmqvist and Hahn-Hägerdal, 2000). Since the biotransformation of furfural,

HMF, 4-hydroxybenzaldehyde and p-coumaric acid by solventogenic Clostridium species has never been investigated, this gap in knowledge hinders the development of inhibitor- tolerant Clostridium strains for ABE production.

In this study, C. beijerinckii 8052 was challenged with different levels of furfural,

HMF, 4-hydroxybenzaldehyde, or p-coumaric acid, and cell growth, microbial inhibitors transformations, and ABE production were monitored and compared with control groups.

Additionally, potential mechanisms of inhibitions by furan aldehydes, phenolic

aldehydes, and phenolic acid on C. beijerinckii 8052 in conjuction with tolerance to these inhibitors are discussed.

4.3 Materials and Methods

4.3.1 Microorganism and culture conditions

Clostridium beijerinckii NCIMB 8052 was obtained from the American Type

Culture Collection, Manassas, VA. C. beijerinckii 8052 stocks were maintained as spores in sterile double distilled water at 4°C. Spores of C. beijerinckii 8052 were heat shocked at 75°C for 10 min, cooled on ice, and then inoculated into 10-ml anoxic tryptone– glucose–yeast extract (TGY) medium (Zhang et al., 2012). The culture was incubated anaerobically at 35±1°C for 12 to 14 h until OD600 attained 0.9–1.1. Then 8% (v/v) of the actively growing C. beijerinckii 8052 culture was transferred into TGY medium and grown for 4 to 5 h until OD600 0.9–1.1 was attained.

4.3.2 Fermentation in P2 medium containing lignocellulose-derived inhibitors

Batch fermentations were performed in 150-mL Pyrex screw-capped media bottles with C. beijerinckii 8052 in triplicate. Pyrex screw capped media bottles containing 100 mL fermentation P2 medium (glucose 60 g/L and yeast extract 1 g/L) were autoclaved at 121°C for 15 min. Media for treatment samples were supplemented with 1–2 g/L furfural, 1–2 g/L HMF, 2–10 g/L furfuryl alcohol, and 0.2 – 0.5 g/L 4- hydroxybenzaldehyde or p-coumaric acid. After cooling to 40°C, loosely capped media bottles were transferred into the anaerobic chamber (Coy, Ann Arbor, MI, USA) at 35°C for 24 h for anaerobiosis. Sterile P2 medium stock solutions including buffer, mineral and vitamin were added prior to inoculation of 6% (v/v) of C. beijerinckii 8052 preculture

(Zhang et al., 2012). During the course of fermentation, 5 mL samples were collected

every 12 h for optical density, pH, residual inhibitory compounds, ABE, and acid analysis.

4.3.3 Analytical methods

Growth of C. beijerinckii 8052 was estimated by measuring its optical density

(OD600) using a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA).

Concentrations of ABE and acids were quantified by a gas chromatography system

(Agilent Technologies 7890A, Agilent Technologies Inc., Wilmington, DE), equipped with a flame ionization detector and 30 m (length) × 320 μm (internal diameter) × 0.50 μm (HP-Innowax film) J × W 19091N-213 capillary column (Han et al.,

2013). Concentrations of microbial inhibitors in samples were determined with a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA). The maximum absorbance spectra of inhibitors were 276, 220, 282, 222, 277, and 284 nm for furfural, furfuryl alcohol,

HMF, HMF alcohol, 4-hydroxybenzaldehyde and p-coumaric acid, respectively. Residual concentrations of microbial inhibitors in samples were calculated using standards of known concentrations. Sample treatments were analyzed statistically to determine level of significance using SAS Version 9.1.3 (SAS Institute Inc., Cary, NC). Independent two- sample t-test between the control and each treatment was conducted to evaluate the effect of inhibitors on cell growth and ABE production by C. beijerinckii 8052.

4.4 Results and Discussion

4.4.1 Decrease of furfural, HMF, 4-hydroxybenzaldehyde and p-coumaric acid by C. beijerinckii 8052

Ability of C. beijerinckii 8052 to transform lignocellulose-derived microbial inhibitory compounds to less inhibitory forms was investigated. Studies focused on two

furan aldehydes, furfural and HMF, and two phenolic compounds, 4- hydroxybenzaldehyde and p-coumaric acid, because they are compounds commonly present in Miscanthus giganteus hydrolysates (See Chapter 3). During growth of and

ABE fermentation by C. beijerinckii 8052, reduction of furfural and HMF was detected and quantified by spectrophotometry at the absorbance of 276 nm and 282 nm, respectively (Figure 4.1A and B). Concurrently, furfuryl alcohol and 2,5-bis- hydroxymethylfuran (HMF alcohol) were produced by the reduction of furfural and

HMF, respectively and generation of these products were detected by spectrophotometry at the absorbance of 220 nm for furfuryl alcohol, and 222 nm for 2,5-bis- hydroxymethylfuran (Figure 4.1A and B). The specific conversion rate of furfural to furfuryl alcohol by C. beijerinckii 8052 (0.15 g/L/h) was higher than that of HMF to 2,5- bis-hydroxymethylfuran (0.08 g/L/h), which agrees with my previous study using C. acetobutylicum ATCC 824 (Zhang et al., 2012; Table 4.1). Biotransformation of furan aldehyde to its corresponding less toxic alcohol during ethanol fermentation by bacteria and yeast strains, methanol production by archaea, as well as butanol fermentation by C. acetobutylicum, has been reported previously (Table 4.1). It has been recognized that hydroxymethyl group on furfuryl alcohol or 2,5-bis-hydroxymethylfuran is generated when the aldehyde group on the furan ring accepts two electrons from electron carrier molecules such as NADH or NADPH (Gutiérrez et al., 2006). In contrast, lower concentrations of 4-hydroxybenzaldehyde (0.5 g/L) and p-coumaric acid (0.3 g/L), were used in this study due to their greater inhibitory effects on microorganisms as discussed in the following sections.

C. beijerinckii 8052 depleted 0.5 g/L of 4-hydroxybenzaldehyde within 18 h during ABE fermentation with a conversion rate of 0.03 g/L/h, while 0.3 g/L p-coumaric acid was completely transformed into an unknown compound at fermentation of 6 h with a conversion rate of 0.05 g/L/h (Figure 4.1C and Table 4.1). The reaction products of 4- hydroxybenzaldehyde and p-coumaric acid could not be detected in this study, however, previous reports have shown biotransformation products of other microorganisms:

Clostridium formicoaceticum has the ability to oxidize 4-hydroxybenzaldehyde to 4-

Hydroxybenzoate by constitutive aldehyde oxidoreductase (Frank et al., 1998). Studies on denitrifying, phototrophic and fermenting bacteria have demonstrated transformation of 4-hydroxybenzaldehyde to the central intermediate benzoyl-CoA through a series of reactions catalyzed by aldehyde reductase, CoA ligase and 4-hydroxybenzoyl-CoA reductase, and benzoyl-CoA are finally metabolized to acetyl-CoA and CO2 (Harwood et al., 1998). Additionally, Clostridium species are able to transform p-coumaric acid to p- hydroxyhydrocinnamic acid by reduction or to 4-vinylphenol and then 4 ethylphenol by decarboxylation and reduction, however, no cleavage of aromatic ring was observed

(Chamkha et al., 2001). It is likely that aromatic rings in furfural, HMF and p-coumaric acid are not easily attacked and only side chain substitution on the aromatic ring occurs, which prevents complete degradation of aromatic structure. However, when a high energy bond (thioester link) is introduced by Coenzyme A to 4-hydroxybenzaldehyde, the aromatic ring can open up and become completely metabolized without accumulation of intermediate products (Harwood et al., 1998).

4.4.2 Impact of furfural, HMF, and furfuryl alcohol on C. beijerinckii 8052 growth and ABE fermentation

To study the effect of furfural and HMF on cell growth and ABE production, batch fermentations were performed by C. beijerinckii 8052 challenged with furfural or

HMF. Generally, the growth of C. beijerinckii 8052 was not affected by the addition of 2 g/L furfural because there was no significant difference between the maximum cell density attained by C. beijerinckii 8052 grown in control and media with 2 g/L furfural

(P>0.05) (Figure 4.2A). However, C. beijerinckii 8052 challenged with 2 g/L HMF underwent a growth inhibition (P≤0.05), and both furfural or HMF (2 g/L) extented the lag phase to 6 h while the growth of the C. beijerinckii 8052 in control medium (without furfural or HMF) started in less than 15 min (Figure 4.2A). The growth of furan challenged C. beijerinckii 8052 cultures was delayed when furfural or HMF was still present in the culture medium, but quickly recovered after 12 and 24 h following depletion of furfural and HMF, respectively (Figure 4.1A and B). Further, a typical biphasic fermentation process was indicated by an initial decrease in pH during exponential growth phase due to acids production, followed by increase and fluctuations in culture pH during stationary growth phase due to acids re-assimilation and ABE production. Clearly, furfural and HMF did not alter this profile (Figure 4.2B). Previous studies on effects of furfural and HMF on the growth of microorganisms are summarized in Table 4.1. Supplementation of fermentation media with furfural ranging from 1.0 g/L to 3.7 g/L during fermentations by microorganisms including yeast, bacteria and archaea resulted in cell growth of about 1% to 50%, except for two Methanococcus strains and two solventogenic Clostridium strains that did not experience any growth inhibition by addition of 1 – 2 g/L furfural (Table 4.1). Similarly, supplementation of fermentation medium with 0.9 g/L to 5.0 g/L HMF often leads to inhibition of cell growth (1.4% -

78%), unlike C. beijerinckii 8052 and C. acetobutylicum 824 that was not inhibited by the presence of 2 g/L HMF in the fermentation medium (Table 4.1). Figure 4.2 C, D and E depict comparison of ABE production by C. beijerinckii 8052 challenged with 2 g/L furfural or HMF. Concentrations of acetone, butanol and ethanol in the fermentation medium increased steadily during ABE fermentation C. beijerinckii 8052 grown in the control and treatment media, and no significant difference of total ABE production was detected between the control and each treatment (P>0.05). This result is consistent with previous studies on C. beijerinckii BA101 (Ezeji et al., 2007) and C. acetobutylicum

ATCC 824 (Zhang et al., 2012), showing that furfural and HMF at tested concentration are not inhibitory to cell growth and ABE production.

As the reduction product of furfural, furfuryl alcohol has been recognized as a lesser toxic compound than furfural (Zaldivar et al., 2000). The effect of furfuryl alcohol on ABE fermentation by solventogenic Clostridium species, however, has not been studied. Figure 4.3 illustrates the impact of furfuryl alcohol (0 g/L, 2 g/L, 3 g/L, 5 g/L and 10 g/L) on cell growth and fermentation by C. beijerinckii 8052. Low concentration

(2 or 3 g/L) of furfuryl alcohol did not significantly alter maximum cell density (P>0.05).

However, challenging with 5 or 10 g/L furfural alcohol, C. beijerinckii 8052 cultures got much lower cell density, suggesting an inhibitory effect on cell growth (Figure 4.3A)

(P≤0.05). The pH profile (Figure 4.3B) indicates a typical biphasic (acidogenic and solventogenic) phase that was not altered by furfuryl alcohol (Figure 4.3F and G).

Whereas supplementation of fermentation medium with 2 or 3 g/L furfuryl alcohol did not decrease total ABE production by C. beijerinckii 8052 (P>0.05), rather, it slightly increased total ABE production by 4% and 9% for growth in medium containing 2 g/L

and 3 g/L furfuryl alcohol, respectively; growth of C. beijerinckii 8052 in fermentation medium containing 5 g/L or 10 g/L furfuryl alcohol resulted in the production of ABE approximately 70% and 40%, respectively of that of the control(P≤0.05) (Figure 4.3C, D and E) The presence of furfuryl alcohol in the fermentation medium may have altered ethanol-butanol ratio in the ABE produced by C. beijerinckii 8052 (Figure 4.3C, D and

E). Interestingly, similar effect of furfuryl alcohol (< 3 g/L) on ethanol fermentation by yeast has been observed to increase ethanol production while a higher concentration decreased its production (Thygesen et al., 2012).This result is consistent with previous study in which furfuryl alcohol was demonstrated to have less inhibitory effect on fermenting microorganisms than furfural on a molar basis (Zaldivar et al., 2000). The toxicity of furfuryl alcohol may be related to hydrophobicity, and it is more likely to disrupt microbial plasma membrane and cause leakage of cellular content (Zaldivar et al., 2000), which is different from the mode of toxicity for furfural or HMF whose primary inhibition targets may be intracellular sites instead of cell membrane (Mills et al.,

2009).

4.4.3 Impact of 4-hydroxybenzaldehyde and p-coumaric acid on C. beijerinckii

8052 growth and ABE fermentation

Although C. beijerinckii 8052 can tolerate furan aldehydes as much as 2 g/L, it is more vulnerable to phenolic compounds (4-hydroxybenzaldehyde and p-coumaric acid) even at a concentrations of less than 0.5 g/L (Figure 4.4). Treatment of C. beijerinckii

8052 with 0.5 g/L of 4-hydroxybenzaldehyde or 0.3 g/L of p-coumaric acid resulted in inhibition on cell growth (P≤0.05) (Figure 4.4A). Specifically, p-coumaric acid delayed cell growth by 44% and extended fermentation time to 84 h. While 4-

hydroxybenzaldehyde did not slow down C. beijerinckii 8052 growth, the maximum cell density attained by C. beijerinckii 8052 challenged with 0.5 g/L 4-hydroxybenzaldehyde was 87% of that obtained in control culture, and the ABE fermentation terminated at fermentation time of 60 h instead of 72 h (Figure 4.4A). Previous studies (Table 4.1) also demonstrate that low levels of 4-hydroxybenzaldehyde or p-coumaric acid (≤1.1 g/L) may lead to inhibition of cell growth by 0.7% to 75%. Figure 4.4B shows all control and treatments had an initial decrease in pH due to acidogenesis during exponential phase, followed by increase and fluctuations in culture pH due to acids uptake and ABE production. The ABE production was significantly impeded when C. beijerinckii 8052 was challenged with either 4-hydroxybenzaldehyde or p-coumaric acid (P≤0.05) (Figure

4.4C, D and E). Maximum butanol and ethanol concentrations obtained from C. beijerinckii 8052 grown in fermentation medium supplemented 0.5 g/L 4- hydroxybenzaldehyde was approximately 50% and 40% of that obtained from control culture, respectively. Although cultures challenged with 0.3 g/L p-coumaric acid produced about 80% and 70% of butanol and ethanol, respectively, compared with the control, the fermentation time was longer, which resulted in lower ABE productivity

(Figure 4.4C and 4D). Since ABE are mainly produced from the uptake of acetic acid and butyric acid, alteration of acids production profile also refected inhibitor toxicity during fermentation process. Final concentrations of both acetic acid and butyric acid were higher in cultures treated with 4-hydroxybenzaldehyde and p-coumaric acid than in the control (Figure 4.4F and G). Different from p-coumaric acid that only decreased acids uptake, 4-hydroxybenzaldehyde resulted in a continuous accumulation of butyric acid

after 24 h (Figure 4.4G) indicating impaired re-assimilation of butyric acid, which is consistent with decreased production of butanol (Figure 4.4C and D).

Consistent with previous study which demonstrated that degradation products of lignin such as phenolic compounds (e.g. 4-hydroxybenzaldehyde and p-coumaric acid) are more toxic to bacteria than degradation products of hemicelluloses or cellulose (e.g. furfural and HMF) (Lee et al., 1999). Notably, the degree of microbial inhibition by a compound is relative to the compound’s hydrophobicity. Since phenolic compounds are often more hydrophobic than furan derivatives, phenolics with the same functional groups as furans are more toxic to microbial cells probably due to increased membrane fluidity and permeability caused by phenolics (Liu and Blaschek, 2010). Although phenolic acids and aldehydes caused only partial or even no detectable damage to the cell membrane, their targets may be located in intracellular hydrophobic components which may influence crucial metabolism, for example, macromolecular synthesis (Mills et al.,

2009).

4.5 Conclusion

C. beijerinckii 8052 transformed and potentially detoxified furfural, HMF, 4- hydroxybenzaldehyde, and p-coumaric acid into less inhibitory compounds during ABE fermentation. Furfural and HMF were transformed to less inhibitory compounds, furfuryl alcohol and 2,5-bis-hydroxymethylfuran, respectively, by C. beijerinckii 8052, whereas

4-hydroxybenzaldehyde and p-coumaric acid were transformed to unknown compounds.

Although C. beijerinckii 8052 grown in fermentation medium supplemented with furfural or HMF increased the lag phase, these compounds (and furfuryl alcohol) did not have inhibitory effect on maximum cell growth and ABE production at concentration less than

3 g/L. However, 4-hydroxybenzaldehyde and p-coumaric acid (at concentrations below

0.5 g/L) suppressed C. beijerinckii 8052 growth, acids uptake and consequently ABE production. This study is significant because obtained results will improve our understanding of impact of lignocellulose-derived inhibitory compounds on C. beijerinckii 8052 during ABE fermentation and conversion of lignocellulosic biomass to fuels and chemicals.

References

Almeida, J.R.M., Bertilsson, M., Gorwa-Grauslund, M.F., Gorsich, S., Lidén, G., 2009.

Metabolic effects of furaldehydes and impacts on biotechnological processes, Appl.

Microbiol. Biotechnol. 82, 625-638.

Belay, N., Boopathy, R., Voskuilen, G., 1997. Anaerobic Transformation of Furfural by

Methanococcus deltae (Delta) LH. Appl. Environ. Microbiol. 63, 2092-2094.

Boopathy, R., 2009. Anaerobic biotransformation of furfural to furfuryl alcohol by a

methanogenic archaebacterium. Int. Biodeterior. Biodegrad. 63, 1070-1072.

Boopathy, R., Bokang, H., Daniels, L., 1993. Biotransformation of furfural and 5-

hydroxymethyl furfural by enteric bacteria. J. Ind. Microbiol. Biotechnol. 11, 147-

150.

Chamkha, M., Garcia, J.L., Labat, M., 2001. Metabolism of cinnamic acids by some

Clostridiales and emendation of the descriptions of Clostridium aerotolerans,

Clostridium celerecrescens and Clostridium xylanolyticum. Int. J. Syst. Evol.

Microbiol. 51, 2105-2111.

Delgenes, J.P., Moletta, R., Navarro, J.M., 1996. Effects of lignocellulose degradation

products on ethanol fermentations of glucose and xylose by Saccharomyces

cerevisiae, Zymomonas mobilis, Pichia stipitis, and Candida shehatae. Enzyme

Microb.Technol. 19, 220-225.

Ezeji, T.C., Blaschek, H.P., 2008. Fermentation of dried distillers' grains and solubles

(DDGS) hydrolysates to solvents and value-added products by solventogenic

clostridia, Bioresour. Technol. 99, 5232-5242.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Frank, C., Schwarz, U., Matthies, C., Drake, H.L., 1998. Metabolism of aromatic

aldehydes as cosubstrates by the acetogen Clostridium formicoaceticum. Arch.

Microbiol. 170, 427-434.

Garcia, J.L., Patel, B.K.C., Ollivier, B., 2000. Taxonomic, phylogenetic, and ecological

diversity of methanogenic Archaea. Anaerobe. 6, 205-226.

Green, E.M., 2011. Fermentative production of butanol—the industrial perspective. Curr.

Opin. Biotechnol. 22, 337-343.

Gutiérrez, T., Buszko, M.L., Ingram, L.O., Preston, J.F., 2002. Reduction of furfural to

furfuryl alcohol by ethanologenic strains of bacteria and its effect on ethanol

production from xylose. Appl. Biochem. Biotechnol. 98-100, 327-340.

Gutiérrez, T., Ingram, L.O., Preston, J.F., 2006. Purification and characterization of a

furfural reductase (FFR) from Escherichia coli strain LYO1—An enzyme important

in the detoxification of furfural during ethanol production. J. Biotechnol. 121, 154-

164.

Han, B., Ujor, V., Lai, L.B., Gopalan, V., Ezeji, T.C., 2013. Use of Proteomic Analysis

To Elucidate the Role of Calcium in Acetone-Butanol-Ethanol Fermentation by

Clostridium beijerinckii NCIMB 8052. Appl. Environ. Microbiol. 79, 282-293.

Harwood, C.S., Burchhardt, G., Herrmann, H., Fuchs, G., 1998. Anaerobic metabolism of

aromatic compounds via the benzoyl-CoA pathway. FEMS Microbiol. Rev. 22, 439-

458.

Hedges, J.I., Ertel, J.R., 1982. Characterization of lignin by gas capillary chromatography

of cupric oxide oxidation products. Anal. Chem. 54, 174-178.

Klinke, H.B., Thomsen, A.B., Ahring, B.K., 2004. Inhibition of ethanol-producing yeast

and bacteria by degradation products produced during pre-treatment of biomass.

Appl. Microbiol. Biotechnol. 66, 10-26.

Lee, W.G., Lee, J.S., Shin, C.S., Park, S.C., Chang, H.N., Chang, Y.K., 1999. Ethanol

production using concentrated oak wood hydrolysates and methods to detoxify. Appl.

Biochem. Biotechnol. 78, 547-559.

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxification,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and

tolerance mechanisms in Escherichia coli. Biotechnol. Biofuels. 2, 26.

Palmqvist, E., Almeida, J.S., Hahn‐Hägerdal, B., 1999. Influence of furfural on anaerobic

glycolytic kinetics of Saccharomyces cerevisiae in batch culture. Biotechnol. Bioeng.

62, 447-454.

Palmqvist, E., Hahn-Hägerdal, B., 2000. Fermentation of lignocellulosic hydrolysates. II:

inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 25-33.

Qureshi, N., Blaschek, H.P., 2001. ABE production from corn: a recent economic

evaluation. J. Ind. Microbiol. Biotechnol. 27, 292-297.

Sakai, S., Tsuchida, Y., Okino, S., Ichihashi, O., Kawaguchi, H., Watanabe, T., Inui, M.,

Yukawa, H., 2007. Effect of lignocellulose-derived inhibitors on growth of and

ethanol production by growth-arrested Corynebacterium glutamicum R. Appl.

Environ. Microbiol. 73, 2349-2353.

Taherzadeh, M.J., Gustafsson, L., Niklasson, C., Liden, G., 2000. Physiological effects of

5-hydroxymethylfurfural on Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol.

53, 701-708.

Thygesen, A., Vahlgren, L., Frederiksen, J.H., Linnane, W., Thomsen, M.H., 2012. SSF

Fermentation of Rape Straw and the Effects of Inhibitory Stress on Yeast, in:

Bioethanol. InTech, pp. 209-221.

Wierckx, N., Koopman, F., Bandounas, L., De Winde, J.H., Ruijssenaars, H.J., 2010.

Isolation and characterization of Cupriavidus basilensis HMF14 for biological

removal of inhibitors from lignocellulosic hydrolysate. Microb. Biotechnol. 3, 336-

343.

Zaldivar, J., Martinez, A., Ingram, L.O., 2000. Effect of alcohol compounds found in

hemicellulose hydrolysate on the growth and fermentation of ethanologenic

Escherichia coli. Biotechnol. Bioeng. 68, 524-530.

Zaldivar, J., Martinez, A., Ingram, L.O., 1999. Effect of selected aldehydes on the growth

and fermentation of ethanologenic Escherichia coli. Biotechnol. Bioeng. 65, 24-33.

Zhang, Y., Han, B., Ezeji, T.C., 2012. Biotransformation of furfural and 5-

hydroxymethyl furfural (HMF) by Clostridium acetobutylicum ATCC 824 during

butanol fermentation. N. Biotechnol. 29, 345-351.

A B

Figure 4.1 Biotransformation of furfural, HMF, 4-hydroxybenzaldehyde and p- coumaric acid during ABE fermentation by C. beijerinckii 8052. Biotransformation of furfural (A) and HMF (B) to furfuryl alcohol and 2,5-bis- hydroxymethylfuran (HMF alcohol) respectively, and decrease of 4- hydroxybenzaldehyde and p-coumaric acid (C) by C. beijerinckii 8052 were observed using P2 medium supplemented with individual compound. Concentrations of Furfural, furfuryl alcohol, HMF, HMF alcohol, 4-hydroxybenzaldehyde and p-coumaric acid were analyzed by absorbance at 276, 220, 282, 222, 277, and 284 nm, respectively. Except for HMF alcohol, concentrations of other compounds were calculated using standards of known concentrations.

A B

C D

Figure 4.2 Growth and ABE production profiles by C. beijerinckii 8052 using P2 medium with 2 g/L furfural or HMF. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium (solid circles), P2 medium supplemented with 2 g/L of furfural (empty circles), and P2 medium supplemented with 2 g/L HMF (solid triangles). Cell growth (A), pH (B), as well as production of acetone (C), ethanol (D) and butanol (E) were illustrated.

Figure 4.3 Growth and ABE production profiles by C. beijerinckii 8052 using P2 medium with 2 to 10 g/L furfuryl alcohol. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium (solid circles), P2 supplemented with 2 g/L (empty circles), 3 g/L (solid triangles), 5 g/L (empty triangles), and 10 g/L (solid square) furfuryl alcohol. Cell growth (A), pH (B), as well as production of acetone (C), ethanol (D), butanol (E), acetic acid (F), and butyric acid (G) were illustrated.

Figure 4.4 Growth and ABE production profiles by C. beijerinckii 8052 using P2 medium with 0.5 g/L of 4-hydroxybenzaldehyde and 0.3 g/L p-coumaric acid. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium (solid circles), P2 medium supplemented with 0.5 g/L of 4-hydroxybenzaldehyde (empty circles), and P2 medium with 0.3 g/L p-coumaric acid (solid triangles). Cell growth (A), pH (B), as well as production of acetone (C), ethanol (D), butanol (E), acetic acid (F), and butyric acid (G) were illustrated.

Table 4.1 Effects of furfural, hydroxymethyl furfural (HMF), hydroxybenzaldehyde and coumaric acid on microorganisms.

Inhibitor/ Growth Specific conversion Microorganism Ref Concentration (g/L) (%) rate (g/L/h) Furfural 2.0 Candida shehatae ATCC 22984 9.7 0.07 (Delgenes et al., 1996) 2.0 Pichia stipitis NRRL Y7124 1 0.07 (Delgenes et al., 1996) 2.0 Saccaromyces cerevisiae CBS 1200 10 0.08 (Delgenes et al., 1996) 2.0 Zymomonas mobilis ATCC 10988 44 0.07 (Delgenes et al., 1996) 4.0 Saccaromyces cerevisiae 27 0.40 (Palmqvist et al., 1999) Corynebacterium glutamicum strain R- 1.0 40 * (Sakai et al., 2007) ldhA-pCRA723 2.0 Escherichia coli LY01 75 * (Zaldivar et al., 1999) 0.4 Escherichia coli KO11 * 0.09 (Gutiérrez et al., 2002) 0.4 Escherichia coli LY01 * 0.09 (Gutiérrez et al., 2002) 0.4 Klebsiella oxytoca P2 * 0.08 (Gutiérrez et al., 2002) 3.4 Escherchia coli ATCC 1175 50 * (Boopathy et al., 1993) 86 3.7 Enterobacter aerogenes 50 * (Boopathy et al., 1993) 3.1 Citrobacter freundii 50 * (Boopathy et al., 1993) 3.5 KIebsiella pneumoniae H strain 50 * (Boopathy et al., 1993) 2.4 Edwardsie,lla sp. 50 * (Boopathy et al., 1993) 1.6 Proteus vulgaris 50 * (Boopathy et al., 1993) 1.4 Proteus mirabilis 50 * (Boopathy et al., 1993) 3.5 Cupriavidus basilensis HMF14 * 0.5 (Wierckx et al., 2010) 1.0 Methanococcus deltae DLH 110 0.03 (Belay et al., 1997) 1.4 Methanococcus sp., strain B 125 0.01 (Boopathy, 2009) 1.7 Clostridium acetobutylicum ATCC 824 115 0.14 (Zhang et al., 2012) 1.9 Clostridium beijerinckii NCIMB 8052 116 0.15 This work

(Continued)

Table 4.1: Continued.

Hydroxymethyl 5.0 Candida shehatae ATCC 22984 8 0.06 (Delgenes et al., 1996) furfural 5.0 Pichia stipitis NRRL Y7124 1.4 0.07 (Delgenes et al., 1996) 5.0 Saccaromyces cerevisiae CBS 1200 11 0.16 (Delgenes et al., 1996) 5.0 Zymomonas mobilis ATCC 10988 33 0.09 (Delgenes et al., 1996) Corynebacterium glutamicum strain R- 0.9 78 * (Sakai et al., 2007) ldhA-pCRA723 2.3 Escherichia coli LY01 75 * (Zaldivar et al., 1999) 2.6 Escherchia coli ATCC 1175 50 * (Boopathy et al., 1993) 2.5 Enterobacter aerogenes 50 * (Boopathy et al., 1993) 2.6 Citrobacter freundii 50 * (Boopathy et al., 1993) 2.3 KIebsiella pneumoniae H strain 50 * (Boopathy et al., 1993) 3.3 Edwardsie,lla sp. 50 * (Boopathy et al., 1993) 1.9 Proteus vulgaris 50 * (Boopathy et al., 1993) 1.3 Proteus mirabilis 50 * (Boopathy et al., 1993) 2.0 Clostridium acetobutylicum ATCC 824 115 0.08 (Zhang et al., 2012) 1.9 Clostridium beijerinckii NCIMB 8052 95 0.08 This work

87 4- 0.5 Candida shehatae ATCC 22984 60 * (Delgenes et al., 1996)

Hydroxybenzald 0.5 Pichia stipitis NRRL Y7124 57 * (Delgenes et al., 1996) ehyde 0.5 Saccaromyces cerevisiae CBS 1200 75 * (Delgenes et al., 1996) 0.5 Zymomonas mobilis ATCC 10988 16 * (Delgenes et al., 1996) Corynebacterium glutamicum strain R- 1.1 60 * (Sakai et al., 2007) ldhA-pCRA723 0.3 Escherichia coli LY01 50 * (Zaldivar et al., 1999) Clostridium formicoaceticum ATCC 0.6 0.7 0.02 (Frank et al., 1998) 27076 0.5 Clostridium beijerinckii NCIMB 8052 87 0.03 This work p-Coumaric acid * Clostridium aerotolerans DSM 5434 * * (Chamkha et al., 2001) * Clostridium xylanolyticum DSM 6555 * * (Chamkha et al., 2001) 0.3 Clostridium beijerinckii NCIMB 8052 66 0.05 This work * Not determined

Chapter 5: Transcriptional Analysis of Clostridium beijerinckii NCIMB 8052 to Elucidate Role of Furfural Stress During Acetone Butanol Ethanol Fermentation

5.1 Abstract

Furfural is the prevalent microbial inhibitor generated during pretreatment and hydrolysis of lignocellulosic biomass to monomeric sugars, but the response of acetone butanol ethanol (ABE) producing Clostridium beijerinckii NCIMB 8052 to this compound at the molecular level is unknown. To discern the effect of furfural on C. beijerinckii and to gain insight into molecular mechanisms of action and detoxification, physiological changes of furfural-stressed cultures during acetone butanol ethanol (ABE) fermentation were studied, and differentially expressed genes were profiled by genome- wide transcriptional analysis.

A total of 5,003 C. beijerinckii 8052 genes capturing about 99.7% of the genome were examined. About 111 genes were differentially expressed (up- or down-regulated) by C. beijerinckii when it was challenged with furfural at acidogenic growth phase compared with 721 genes that were differentially expressed (up- or down-regulated) when C. beijerinckii was challenged with furfural at solventogenic growth phase. The differentially expressed genes include genes related to redox and cofactors, membrane transporters, carbohydrate, amino sugar and nucleotide sugar metabolisms, heat shock

proteins, DNA repair, and two-component signal transduction system. While C. beijerinckii exposed to furfural stress during the acidogenic growth phase produced 13% more ABE than the unstressed control, ABE production by C. beijerinckii ceased following exposure to furfural stress during the solventogenic growth phase.

Genome-wide transcriptional response of C. beijerinckii to furfural stress was investigated for the first time using microarray analysis. Stresses emanating from ABE accumulation in the fermentation medium; redox balance perturbations; and repression of genes that code for the phosphotransferase system, cell motility and flagellar proteins

(and combinations thereof) may have caused the premature termination of C. beijerinckii

8052 growth and ABE production following furfural challenge at the solventogenic phase.This study provides basis for metabolic engineering of C. beijerinckii 8052 for enhanced tolerance of lignocellulose-derived microbial inhibitory compounds, thereby improving bioconversion of lignocellulosic biomass hydrolysates to biofuels and chemicals.

5.2 Introduction

Clostridium species are gram positive, anaerobic, spore-forming bacteria

(Mitchell, 1997). Production of ABE (acetone-butanol-ethanol) by solventogenic

Clostridium species is attractive by virtue of their ability to utilize sugars such as cellobiose, glucose, xylose, arabinose and mannose that are present in lignocellulosic hydrolysate. Butanol production from lignocellulosic biomass has shown promise, and the feedstock is abundant, renewable and relatively cheap (Ezeji et al., 2007). However, microbial inhibitors, such as furfural, hydroxymethyl furfural (HMF),

hydroxybenzaldehyde, and coumaric acid, are produced during lignocellulosic biomass pretreatment and impede biofuel production from generated hydrolysates (Ezeji et al.,

2007).

These lignocellulose-derived inhibitory compounds inhibit cell growth and biofuel production by disrupting cell membranes, damaging polynucleotides, repressing central metabolic enzymes, decreasing intracellular pH, increasing cell turgor pressure, and inducing oxidative stress (Mills et al., 2009; Allen et al., 2010). Although physical, chemical and biological inhibitor removal methods may facilitate substrate utilization and butanol fermentation, removal of inhibitors from hydrolysates prior to fermentation may not be economically feasible due to the cost associated with additional processing steps and the potential loss of fermentable sugars (Liu and Blaschek, 2010). To make bioconversion of lignocellulosic biomass to butanol economically feasible, development of an inhibitor-tolerant strain of bacteria is crucial. Previous studies have investigated the impact of lignocellulose-derived inhibitors on butanol fermentation. Ezeji et al. (2007) have demonstrated the inhibitory effect of corn-fiber-derived aldehyde, organic acid and phenolic compounds on cell growth and ABE production by C. beijerinckii BA101, and they have shown the synergistic effect of mixtures of inhibitors over the sum of individual toxic effects. Chemicals present in wheat straw hydrolysates, mainly furfural and HMF, have also been reported to enhance butanol productivity when using C. beijerinckii P260 (Qureshi et al., 2011). Furthermore, the detoxification of furfural and

HMF by C. acetobutylicum ATCC 824 has been shown during butanol fermentation

(Zhang et al., 2012). However, the underlying mechanisms for inhibitor detoxification

and tolerance by fermenting solventogenic Clostridium species remain unclear. This gap in knowledge continues to hamper attempts at engineering inhibitor-tolerant bacterial strains, as evidenced by the persistent low butanol productivity from biomass feedstock and, hence, the high cost of butanol production.

The objective of this study was to examine the response of C. beijerinckii 8052 at the mRNA level to the challenge of furfural to better understand the interplay of furfural toxicity and corresponding bacterial tolerance mechanisms. The impact of furfural, the most representative lignocellulose-derived inhibitor, on C. beijerinckii 8052 was studied to understand its potential mechanism and that of other lignocellulose-derived aldehydes.

To gain insight into mechanisms of furfural toxicity and tolerance, the interactive effect of furfural conversion and ABE production was studied by challenging fermentation cultures with different doses of furfural at different growth stages. To examine physiological alterations in C. beijerinckii 8052 cell growth and ABE production as a consequence of furfural stress, gene expression patterns between control and furfural- stressed treatment cultures were compared by genome-wide transcriptional microarray analysis. The comparison of gene expression patterns in relation to ABE production is expected to provide insights toward metabolically engineering C. beijerinckii 8052 with enhanced tolerance for lignocellulose-derived microbial inhibitory compounds and better utilization of lignocellulosic biomass hydrolysates.

5.3 Materials and Methods

5.3.1 Bacterial strains, culture conditions

Clostridium beijerinckii NCIMB 8052 (ATCC 51743) was obtained from the

American Type Culture Collection (Manassas, VA) and was used in all experiments unless noted otherwise. Stocks of C. beijerinckii 8052 spores were stored in sterile, double-distilled water at 4 °C. To revive C. beijerinckii 8052 spores, a 200 μL aliquot was heat-shocked for 10 min at 75 °C followed by cooling on ice. The heat-shocked spores were inoculated into 10 mL anoxic pre-sterilized tryptone–glucose–yeast extract

(TGY) medium and incubated in an anaerobic chamber (Coy Laboratory Products Inc.,

Ann Arbor, Michigan) with a modified atmosphere of 82% N2, 15% CO2, and 3% H2 for

12 h to 14 h at 35 °C ± 1 °C until active growth (OD600 0.9–1.1) was attained (Han et al.,

2011).Eight ml of actively growing culture was subsequently transferred into 92 mL of anoxic TGY medium. The culture was grown anaerobically at 35 °C ± 1 °C for 4 h to 5 h; during this time it reached an optical density (OD600) of 0.9 - 1.1, when it was used as the pre-culture.

5.3.2 Furfural-challenged experiments and ABE Fermentation

Batch ABE fermentation by C. beijerinckii 8052 was performed in 250-mL Pyrex screw-capped media bottles containing 200 mL anoxic P2 medium (glucose 60 g/L and yeast extract 1 g/L) and P2 stock solutions as described previously (Zhang et al., 2012).

To evaluate the response of C. beijerinckii 8052 to furfural during acidogenic and solventogenic phases, a 1-L flask containing 600 mL anoxic P2 medium plus P2 stock solutions was inoculated (6% v/v) with C. beijerinckii 8052 pre-culture and incubated anaerobically for 8 h; during this time the OD600 attained 1.4 ± 0.05 (acidogenic phase).

The culture was then subdivided into aliquots of 100 mL in six 150-mL Pyrex screw-

capped media bottles. Three bottles in the treatment group were challenged with furfural

(2 g/L) and the other three bottles were left unchallenged as the control. After 2 to 3 h growth at 35 C, the original concentration of furfural in the growth medium was reduced by more than half, and C. beijerinckii 8052 samples were collected from each bottle and triplicate control and furfural-challenged samples were pooled separately. Notably, cell density of C. beijerinckii 8052 during solventogenic phase was markedly higher than that at acidogenic growth phase. Given the greater number of C. beijerinckii 8052 cells in the fermentation medium at solventogenic phase; we decided to increase the concentration of furfural to 3 g/L. For furfural challenge at solventogenic phase, 600 ml of C. beijerinckii

8052 culture was incubated anaerobically for 24 h; during this time the OD600 attained 4.0

± 0.2, and then subdivided into aliquots of 100 mL in six 150-mL bottles. Three bottles were challenged with 3 g/L of furfural which were the treatment group and the other three bottles were unchallenged (control). Following 2 to 3 h post-furfural challenge, during which time the original concentration of furfural in the growth medium was reduced by more than half, C. beijerinckii 8052 samples were collected from each bottle and triplicate control and furfural-challenged samples were pooled separately for analysis. Aliquots of C. beijerinckii 8052 samples were placed on ice immediately after removal from the bottles. C. beijerinckii 8052 pellets were obtained by centrifugation at

5000 g for 10 min at 4 °C prior to suspension in a solution containing a 2:1 ratio of

RNAprotect cell reagent to phosphate-buffered saline (PBS) (Qiagen Inc., Valencia, CA) to stabilize the RNA. The suspension was incubated at room temperature for 5 min,

centrifuged to obtain cell pellets, and stored at -80 °C overnight as described previously

(Servinsky et al., 2010).

The pH profile of C. beijerinckii 8052 fermentation was monitored with a

Beckman Ф500 pH meter (Beckman Coulter Inc., Brea, CA). Growth of C. beijerinckii

8052 was estimated using a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA) to measure the OD600. Concentrations of fermentation products–acetate, butyrate, acetone, butanol, and ethanol were measured using a 7890A Agilent Technologies gas chromatograph (Agilent Technologies Inc., Wilmington, DE) equipped with a flame ionization detector (FID) and 30 m (length) x 320 m (internal diameter) x 0.50 m (HP-

Innowax film) J x W 19091N-213 capillary column as described previously (Han et al.,

2011). Initial and residual furfural concentrations in the growth medium were determined as described previously (Zhang et al., 2012). Sample treatments were analyzed statistically to determine level of significance using SAS Version 9.1.3 (SAS Institute

Inc., Cary, NC). Independent two-sample t-test between the control and each treatment was conducted to evaluate the effect of inhibitors on ABE production by C. beijerinckii

8052.

5.3.3 Total RNA purification

The total cellular RNA was purified from 2-mL C. beijerinckii 8052 culture.

Briefly, C. beijerinckii 8052 cell pellets from 2-mL culture aliquots were thawed on ice and re-suspended in 750 μL RLT buffer™ (Qiagen Inc., Valencia, CA) supplemented with 1% (v/v) β-mercaptoethanol (β-ME). For complete lysis of C. beijerinckii 8052, a 2- mL microcentrifuge tube was half filled with 0.1 mm diameter zirconia/silica beads

(Biospec, Bartlesville, OK) followed by 750 μL RLT buffer™ supplemented with 1%

(v/v) β-ME and then chilled on ice. C. beijerinckii 8052 suspension (750 μL) was added to the pre-chilled tube with beads and agitated in a TissueLyser LT (Qiagen Inc.,

Valencia, CA) for 8 min at a setting of 50 Hz to break the C. beijerinckii 8052 cells. Total

RNA was purified from homogenized cells using an RNeasy mini kit (Qiagen Inc.,

Valencia, CA) according to the manufacturer's instructions. RNA quality was analyzed using a Nanochip 2100 bioanalyzer (Agilent Technologies Inc., Wilmington, DE), and

RNA concentration was measured by NanoDrop 3300 (Thermo scientific, Wilmington,

DE) according to the manufacturer's instructions.

5.3.4 Comparative microarray hybridization

The microarray was constructed by MYcroarray Inc. (Ann Arbor, MI). A total of

5,003 C. beijerinckii 8052 genes capturing about 99.7% of the genome were examined.

To minimize error, five identical replicates of each C. beijerinckii 8052 (Cb) probe sequence (45-47mer) were designed and fabricated onto the microarray chip. To enhance hybridization, 10 µg of total RNA was converted to enriched mRNA using a

MICROBExpress™ Bacterial mRNA Enrichment Kit (Life Technologies, Grand Island,

NY) and following manufacturer’s protocol. About 200 ng of enriched mRNA was converted to complementary RNA (cRNA) using a MessageAmp™ II-Bacteria RNA

Amplification Kit (Life Technologies, Grand Island, NY) and following manufacturer’s protocol. Alexa Fluor 555 (Life Technologies, Grand Island, NY) was coupled to cRNA following the manufacturer’s instructions. Removal of unincorporated dye was conducted using RNeasy Mini Columns (Qiagen Inc., Valencia, CA) and following manufacturer’s

protocol. About 30 – 50 μL eluate containing ~40 μg labeled cRNA was generated. The resulting dye-coupled cRNA was made up to 60 μL with 2 μL 150 mM ZnSO4 and elution buffer to bring the final concentration of ZnSO4 to 5 mM; this dilution was followed by incubation (fragmentation) at 75 °C for 10 min. Ten micrograms of each fluor-labeled cRNA was hybridized separately to one array, and hybridization was performed for 20 h at 45 °C in a hybridization buffer containing 6x SSPE (20X SSPE stock: 3 M NaCl, 20 mM EDTA, 118.2 mM NaH2PO4 and 81.8 mM Na2HPO4), 10% de- ionized formamide, 0.01 mg/mL acetylated BSA, 0.01% Tween-20, and 1% control oligos (provided by Mycroarray Inc). After hybridization, the slide was washed with fresh 6x SSPE buffer (24 °C) and transferred to fresh 0.5x SSPE buffer prior to drying.

The slide was dried immediately by centrifugation at 2000 g for 3 min prior to scanning by an Axon 4000B Scanner (Molecular Devices, Sunnyvale, CA).

5.3.5 Microarray data analysis

The slide was scanned using an Axon 4000B Scanner set at 5 μm per pixel resolution and 100% laser power. The scanned images were extracted and analyzed using version 6.1.0.4 GenePix Pro Software (Molecular Devices, Sunnyvale, CA). For signal extraction, circular feature indicators (35 μm diameter) were centered over each spot, and the median feature pixel intensity was extracted. Data images were extracted from the center of the spot area (35 µm in diameter), where the sequence fidelity is exceptional, instead of from the larger indicator feature spot diameter area (~60 µm). To minimize error due to differences in sample behavior from array to array, a scale factor was created

to normalize the signal across all arrays. A scale factor (SF) for each array was calculated as follows:

SF = μ (median pixel signal)control/μ (median pixel signal)treatment. Trimmed mean was used to generate the final signal value for the five identical probe replicates, which identified the differential expression pattern of each gene on the array. The trimmed mean was calculated by discarding the maximum and minimum adjusted signal within each set of five probe replicates, followed by averaging the adjusted signal values of the remaining three probes. Relative expression levels (gene expression ratio) for each gene were calculated by dividing the signal intensity of the array from the furfural-challenged

C. beijerinckii 8052 culture by the intensity of the unchallenged control culture. To facilitate a fair comparison of up- and down-regulated genes, fold change was calculated as follows: for genes with an expression ratio ≥1, the fold change is the same as the expression ratio, whereas the fold change of genes whose expression ratio <1 equals the reciprocal of the expression ratio multiplied by -1 (Babu, 2004). To make data distribution symmetrical, the gene expression ratio was used to calculate the log2 transformation ratio as described previously (Quackenbush, 2002). Expression patterns were visualized colorimetrically using TreeView (version 1.60). Enrichment analysis of

Gene Ontology terms, including biological process, cellular component and molecular function, and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment pathway analysis were performed using a DAVID Functional Annotation Bioinformatics

Microarray analysis to identify statistically over-represented biological terms (Huang et al., 2009).

5.3.6 Microarray data accession number

All microarray data have been submitted to NCBI’s Gene Expression Omnibus database at http://www.ncbi.nlm.nih.gov/geo/ with accession number GSE42597.

5.3.7 Real-time quantitative reverse transcription PCR (Q-RT-PCR)

Following microarray analysis, several genes were differentially induced or repressed in response to furfural stress, and several genes were selected for further analysis using Q-RT-PCR to validate microarray results. Briefly, C. beijerinckii 8052 cultures were grown anaerobically and challenged with 2-3 g/L furfural during acidogenic and solventogenic growth phases followed by centrifugation to collect cell pellets as described above. Total RNA was purified from cell pellets of furfural- challenged and unchallenged C. beijerinckii 8052 cultures and lysed with TissueLyser LT as described above. Genomic DNA was removed from the total RNA isolate using

RNase-free DNase (New England Biolabs Inc, Ipswich, MA). Total RNA (2 µg) was reverse transcribed to form first strand cDNA by random hexamer-primed reverse transcription reactions using SuperScript III reverse transcriptase (Life Technologies,

Grand Island, NY) following the manufacturer’s instructions. For quantitative reverse transcription polymerase chain reaction (Q-RT-PCR), cDNA, specific primers and

GoTaq® qPCR Master Mix containing Bryt™ Green dye (Promega, Madison, WI) were proportionately mixed following manufacturer’s protocol. The forward and reverse gene- specific primers used for amplification of specific genes were synthesized by Eurofins

MWG Operon and are listed in Table 5.1. The 16S rRNA of C. beijerinckii 8052, which was amplified with gene-specific forward (5’- GAA GAA TAC CAG TGG CGA AGG

C-3’) and reverse (5’- ATT CAT CGT TTA CGG CGT GGA C-3’) primers, was used as the internal standard. Prior to selection of 16S rRNA as an internal standard, the expression of 16s rRNA of furfural-challenged and unchallenged C. beijerinckii 8052 cultures was analyzed and confirmed for constant expression under the reaction condition of the study. The mRNA levels of genes of interest (Table 5.1) were quantified by subjecting cDNA to Q-RT-PCR analysis in triplicate using a Bio-Rad iCycler continuous fluorescence detection system (Bio-Rad, Hercules, CA); Q-RT-PCR reaction conditions were as follows: step 1, 95 °C for 2 min (hot-start activation), step 2, 95 °C for 15 sec

(denaturation), step 3, 55 °C for 30 sec (annealing and extension), 40 cycles of step 2 and

3, step 4, 95 °C for 1 min (denature of PCR product), step 5, 55 °C for 1 min (annealing of PCR product), and step 6, heat from 65°C to 95°C with a ramp speed of 1°C per 10 sec, resulting in melting curves.. Expression levels of C. beijerinckii 8052 genes were quantified by the comparative CT method as previously described (Schmittgen and Livak,

2008).

5.4 Results

5.4.1 Microarray analysis of C. beijerinckii 8052 transcriptome under furfural stress

Furfural was chosen as the model lignocellulose-derived inhibitory compound for this investigation because it is the most prevalent microbial inhibitor generated during pretreatment and hydrolysis of lignocellulosic biomass to monomeric sugars. However, it has been shown previously to enhance solventogenic Clostridium species growth and

ABE production when the fermentation medium is supplemented with <3 g/L furfural

prior to fermentation (Ezeji et al., 2007, Ezeji and Blaschek, 2008). To better understand mechanisms with which furfural affect C. beijerinckii 8052 physiology, the global response of C. beijerinckii 8052 to the challenge of furfural during both acidogenesis and solventogenesis at the mRNA level was profiled using whole genome microarray analysis. The findings from the present study are grouped into different attributes.

5.4.2 Expression of C. beijerinckii 8052 redox and cofactor genes in the presence of furfural

After the furfural stress during the acidogenic phase, the expression of some redox proteins in C. beijerinckii 8052 increased by up to 16-fold compared with that in the control group (Figure 5.1A and Appendix A: Table A3a). Gene ontology (GO) analysis (Appendix A: Table A1a) shows that three of these redox proteins are involved in antioxidant activity (GO:0016209): thioredoxin reductase (Cbei_2681), redoxin domain-containing protein (Cbei_2680), and glutathione peroxidase (Cbei_0389); the latter two possess oxidoreductase activity acting on superoxide as an acceptor

(GO:0016684) and in response to oxidative stress (GO:0006979). Another group of genes that was up-regulated by more than threefold encodes oxidoreductases acting on CH or

CH2 groups, with disulfide as an acceptor (GO:0016728) (Appendix A: Table A1a).

According to KEGG enrichment pathway analysis, this group of genes is associated with purine (cbe00230) and pyrimidine (cbe00240) metabolisms (Appendix A: Table A2), and includes anaerobic ribonucleoside triphosphate reductase (Cbei_0068), adenosylcobalamin-dependent ribonucleoside-triphosphate reductase (Cbei_2522), and ribonucleotide-diphosphate reductase subunits (Cbei_0194 and Cbei_0195). The

100

remaining oxidoreductases (GO:0016491) (Appendix A: Table A1a) that have higher expression in the furfural treatment culture than in the control culture are aldo/keto reductase (Cbei_3974), short-chain dehydrogenase/reductase (SDR) (Cbei_3904), disulfide bond formation A (DSBA) oxidoreductase (Cbei_2058), FAD linked oxidase domain-containing protein (Cbei_0312), and alcohol dehydrogenase (Cbei_1464) (Figure

5.1A and Appendix A: Table S3A). The transcriptome of C. beijerinckii 8052 after furfural challenge at the solventogenic phase shows some similarities in terms of redox enzymes. All the above genes, except FAD linked oxidase domain-containing protein

(Cbei_0312), and alcohol dehydrogenase (Cbei_1464), were also induced by furfural challenge at the solventogenic phase (Figure 5.1A and Appendix A: Table A3c).

Besides redox enzymes, components associated with redox reactions were also highly expressed in cultures challenged with furfural at the acidogenic phase. One of the related components affected by furfural treatment is the iron-sulfur cluster. The expression of genes encoding iron-sulfur cluster assembly proteins (Cbei_1848,

Cbei_1849, Cbei_1850, Cbei_1851 and Cbei_1852) increased by up to fivefold (Figure

5.1A and Appendix A: Table A3a); these genes are classified into the cofactor biosynthetic process (GO:0051188) (Appendix A: Table A1a). Another group of genes classified into the same group (GO:0051188), as well as into the vitamin biosynthetic process (GO:0009110) (Appendix A: Table A1a), includes those encoding cobalt ABC transporter ATPase (Cbei_3693), cobalt ABC transporter permease (Cbei_3694) and cobalt transport protein CbiM (Cbei_3695) (Figure 5.1A and Appendix A: Table A3a). In addition, differential expression was also observed in furfural-challenged cultures for

101

several members of riboflavin biosynthesis genes (Cbei_1224, Cbei_1225, Cbei_1226,

Cbei_1227) (Figure 5.1A and Appendix A: Table A3a). This group of genes belongs to the Gene Ontology term riboflavin metabolic process (GO:0006771) (Appendix A: Table

A1a), and if classified by KEGG pathway analysis, these genes are involved in riboflavin metabolism (cbe00740) (Appendix A: Table A2). However, furfural challenge during solventogenesis affected gene expression differently from that at acidogenesis in terms of redox enzyme cofactors. First, expression of genes that code for iron-sulfur cluster assembly proteins was even higher during solventogenesis (Figure 5.1A and Appendix A:

Table A3c), and those genes (Cbei_1848, Cbei_1849, Cbei_1850, Cbei_1851 and

Cbei_1852) were up-regulated in furfural-challenged cultures by up to 54-fold compared to no more than fivefold during acidogenesis (Figure 5.1A and Appendix A: Table A3c).

On the other hand, the expression of genes involved in synthesis of other cofactors, including riboflavin and cobalamin, did not show obvious alterations during furfural challenge at solventogenesis (Figure 5.1A), although these genes were highly induced during furfural challenge at acidogenesis (Figure 5.1A and Appendix A: Table A3a).

5.4.3 Expression of membrane transporter genes in C. beijerinckii 8052

Gene expression analysis of C. beijerinckii 8052 responding to furfural stress during butanol fermentation was performed to determine not only the effect of furfural on

C. beijerinckii 8052 growth and ABE production but also on molecular physiological changes. Furfural in the ABE fermentation medium altered expressions of the membrane transport system, including ATP-binding cassette transporters (ABC-transporter) and phosphotransferase system (PTS), in C. beijerinckii 8052 during both acidogenic and

102

solventogenic phases (Figure 5.1B). While some ABC-transporter genes such as galactoside ABC transporter (Cbei_3298), multidrug ABC transporter ATPase

(Cbei_3299 and Cbei_3300), and cobalt ABC transporter ATPase (Cbei_3693) were expressed up to seven-fold in furfural-challenged C. beijerinckii 8052 at the acidogenic phase (Figure 5.1B and Appendix A: Table A3a), expression of these transporter genes was increased by a greater fold during the solventogenic phase (Figure 5.1B and

Appendix A: Table A3c). According to KEGG pathway analysis, some ABC-transporter- related genes may be classified into what is known as KEGG pathway ABC transporters

(cbe02010) (Appendix A: Table A2), which include transport proteins that catalyze transmembrane movement of different substrates including sulfate, phosphate and branched-chain amino acid. Expression of these genes increased up to twelvefold in furfural-challenged C. beijerinckii 8052 during solventogenesis (Figure 5.1B and

Appendix A: Table A3c). Specifically, genes involved in sulfate transportation include sulfate ABC transporter ATPase (Cbei_4190), sulfate ABC transporter inner membrane protein (Cbei_4191 and Cbei_4192), and sulfate ABC transporter substrate-binding protein (Cbei_4193); phosphate transporters include phosphate binding protein

(Cbei_1127), phosphate ABC transporter permease (Cbei_1128 and Cbei_1129), and phosphate ABC transporter ATPase (Cbei_1130); and genes that code for branched-chain amino acids include extracellular ligand-binding receptor (Cbei_1762, Cbei_1767 and

Cbei_5042), inner-membrane translocator (Cbei_1763, Cbei_1764, Cbei_5043, and

Cbei_5044), and ABC transporter (Cbei_1765, Cbei_1766, Cbei_5045, Cbei_5046, and

Cbei_2145). In addition, genes for cyanate or nitrite transport that belong to ABC

103

transporters (Cbei_2089 and Cbei_3331) are equally induced in furfural-challenged C. beijerinckii 8052 during solventogenesis (Figure 5.1B and Appendix A: Table A3c).

Unlike ABC-transporter genes whose expression was increased by furfural- challenged C. beijerinckii 8052, another member of the membrane transporter system, phosphotransferase system (PTS), was repressed in furfural-challenged cultures of C. beijerinckii 8052 at both acidogenic and solventogenic phases (Figure 5.1B). Prominent among the PTS are the PTS system mannose/fructose/sorbose family transporters involved in fructose and mannose metabolism (cbe00051) and amino sugar and nucleotide sugar metabolism (cbe00520) (Appendix A: Table A2). While mostly genes encoding PTS system mannose/fructose/sorbose family transporter subunits IIA

(Cbei_4914), IIB (Cbei_4913), IIC (Cbei_4912 and Cbei_3872), and IID (Cbei_4911 and

Cbei_3871) were repressed by up to four-fold when cultures of C. beijerinckii 8052 were challenged with furfural during the acidogenic growth phase (Figure 5.1B and Appendix

A: Table A3b), a wider spectrum of genes, including the PTS system mannose/fructose/sorbose family transporters, was repressed when cultures of C. beijerinckii 8052 were challenged with furfural at the solventogenic growth phase (Figure

5.1B and Appendix A: Table A3d). The repressed genes associated with sugar metabolism during the solventogenic growth phase include mannose/fructose/sorbose family transporter subunit IID (Cbei_0958, Cbei_2196, Cbei_3871, Cbei_4557, and

Cbei_4911), mannose/fructose/sorbose family IIC subunit (Cbei_3872), sorbose-specific transporter subunit IIC (Cbei_4558), sorbose subfamily transporter subunit IIB

(Cbei_4559), mannose-6-phosphate isomerase (Cbei_0996), and glucitol/sorbitol-specific

104

transporter subunit IIC (Cbei_0336) (Figure 5.1B). Besides the listed genes, N- acetylglucosamine-specific IIBC subunit (Cbei_4532) and glucose subfamily transporter subunit IIA (Cbei_4533) are also involved in amino sugar and nucleotide sugar metabolism (cbe00520) (Appendix A: Table A2). The repression of these genes may affect the transportation and metabolism of sugars such as those in the glucose family (N- acetyl-D-glucosamine, D-glucosamine and glucosides), the lactose and cellobiose families, the mannose family (mannose and galactosamine), and others such as sorbose, sorbitol, glucitol, and L-ascorbate, many of which are monomeric sugars of lignocellulosic biomass. Furfural challenge of C. beijerinckii 8052 during the solventogenic phase, in addition, inhibited other specific PTS systems, including lactose/cellobiose-specific subunits (Cbei_2663, Cbei_2740, Cbei_4634, Cbei_4639,

Cbei_4640, and Cbei_4683), sorbose-specific subunits (Cbei_2907) and subunit IIA-like nitrogen-regulatory protein PtsN (Cbei_2741) (Figure 5.1B and Appendix A: Table A3d).

5.4.4 Expression of a two-component signal transduction system, chemotaxis, and cell motility genes in C. beijerinckii 8052

As with membrane transporter genes, the expression of genes associated with the two-component signal transduction system (cbe02020) was altered in furfural-challenged

C. beijerinckii 8052 at both acidogenic and solventogenic phases (Appendix A: Table

A2). Following furfural challenge of C. beijerinckii 8052 during the acidogenic growth phase, only two genes (Cbei_4019, chemotaxis protein CheA and Cbei_4273,

MotA/TolQ/ExbB proton channel) involved in the two-component signal transduction system were repressed by about fourfold (Figure 5.1C and Appendix A: Table A3b).

105

When the C. beijerinckii 8052 culture was challenged with furfural at the solventogenic growth phase, more than 40 genes were repressed by up to 18-fold (Figure 5.1C and

Appendix A: Table A3d). Notably, the two major functional categories of genes belonging to the two-component signal transduction system are bacterial chemotaxis

(cbe02030) and flagellar assembly (cbe02040) (Appendix A: Table A2).

Although chemotaxis is the most widely studied two-component sensory system in bacteria, not much has been reported about the system in relation to furfural stress in solventogenic Clostridium species. When the C. beijerinckii 8052 culture was challenged with furfural at the solventogenic growth phase, many genes associated with the chemotaxis sensory system in C. beijerinckii 8052, such as methyl-accepting chemotaxis sensory transducer (Cbei_0287, Cbei_0804, Cbei_2787, Cbei_3356, Cbei_3671,

Cbei_3961, Cbei_4161, Cbei_4821, and Cbei_4828) and genes that code for chemotaxis proteins (CheA Cbei_4307, Cbei_4829, and Cbei_4183; CheB Cbei_4309 and

Cbei_4826; CheR Cbei_4827; CheW Cbei_4184 and Cbei_4822; CheY Cbei_4819 and

Cbei_4015; and MotA Cbei_4273), were differentially repressed (Figure 5.1C and

Appendix A: Table A3d). Since chemotaxis directs flagellar motion and controls the swimming pattern of the cell (Armitage and Schmitt, 1997), genes encoding flagellar assembly proteins were also differentially repressed by furfural. These flagellar proteins include FliS (Cbei_4292), FliR and FlhB (Cbei_4254), FliH (Cbei_4266), MotA

(Cbei_4273), FlgL (Cbei_4297), FliC (Cbei_4274 and Cbei_4289), and FliD

(Cbei_4291), as shown in Figure 5.1C and Appendix A: Table A3d. Additionally, there are two-component signal transduction systems related to genes encoding proteins that

106

partake in many cellular functions such as quorum sensing and flagella assembly

(flagellin domain-containing protein Cbei_4274 and Cbei_4289, and MotA/TolQ/ExbB proton channel Cbei_4273), carbon storage regulation (carbon storage regulator CsrA

Cbei_4295), nitrogen assimilation (glutamine synthetase Cbei_0444), and cell cycle progression and development (signal transduction histidine kinase regulating citrate/malate metabolism Cbei_4175, multi-sensor signal transduction histidine kinase

Cbei_4430, and histidine kinase internal region Cbei_4458) that were differentially repressed when the C. beijerinckii 8052 culture was challenged with furfural at the solventogenic growth phase (Figure 5.1C and Appendix A: Table A3d).

5.4.5 Validation of gene expression data from microarray analysis by Q-RT-

PCR

To validate differential gene expressions obtained using microarray analysis, Q-

RT-PCR was applied to quantify gene expression levels in biological replicate cultures of

C. beijerinckii 8052 using treatment conditions that mimicked microarray treatment but were independent of the cultures used for microarray analysis. Briefly, the C. beijerinckii

8052 culture was challenged with furfural at acidogenic and solventogenic growth phases during which 19 and 23 genes, respectively, were evaluated. The genes were selected randomly within each range of fold change. Differential gene expressions in furfural- challenged C. beijerinckii 8052 determined via microarray analysis and Q-RT-PCR were found to have a high degree of correlation between them at both acidogenic (R = 0.87) and solventogenic phases (R = 0.84) (Figure 5.2, Table 5.1).

5.4.6 Interactive effect of furfural reduction and ABE production

107

To determine effects of furfural on C. beijerinckii 8052 growth and ABE production at acidogenic and solventogenic phases, a C. beijerinckii 8052 culture grown in P2 medium was challenged with furfural, and changes in cell density, acid production and ABE production were measured relative to cultures grown in P2 medium without furfural. Challenge of C. beijerinckii 8052 with 2 g/L furfural during the acidogenic phase (fermentation time 8 h) when OD600 was between 1.5 and 2.0 resulted in complete depletion of furfural within 4 h, and cell densities of the furfural-challenged C. beijerinckii 8052 and control cultures were nearly indistinguishable (Figure 5.3A).

However, in the presentce of furfural, ABE production was significantly inhibited compared with that in the control (P≤0.05) (Figure 5.3A-E). The acetic and butyric acid levels measured in both the furfural-challenged C. beijerinckii 8052 and the unchallenged control cultures were reflective of the respective acetone and butanol production profiles

(Figure 5.3). Notably, although ABE production and acid re-assimilation by furfural- challenged C. beijerinckii 8052 were inferior to that of the unchallenged control, the fermentation proceeded rapidly following depletion of furfural, and the maximum concentrations of total acetone, butanol, and ethanol produced by the furfural-challenged

C. beijerinckii 8052 were not significantly different between the control and the challenged culture (P>0.05) (Figure 5.4A-C). Interestingly, acid assimilation was stimulated in the furfural-challenged C. beijerinckii 8052 following furfural depletion in the fermentation medium, and the final concentrations of acetic and butyric acid were lower than that of the control by 22% and 19%, respectively (Figure 5.4D and E).

108

While C. beijerinckii 8052 cultures challenged with furfural at the acidogenic phase could tolerate furfural and produce same amount of ABE as the control following the depletion of furfural (P>0.05), challenging C. beijerinckii 8052 culture with furfural during the solventogenic phase (fermentation time 25 h; OD600 5.0-5.5) resulted in shut down of ABE production and rapid accumulation of acetic and butyric acid in the fermentation medium (Figure 5.5). Unlike the control, C. beijerinckii 8052 grown in P2 medium without furfural underwent a normal fermentation process (Figures 5 and 6). At solventogenesis, furfural reduction was impeded (Figure 5G) when concentrations of acetone, ethanol and butanol were high (3.40 g/L ± 0.27 g/L, 0.22 g/L ± 0.02 g/L, and

5.93 g/L ± 0.12 g/L, respectively) (Figure 5.5A-C). Although the cell density of C. beijerinckii 8052 in the solventogenic phase culture was four times higher than in the acidogenic phase culture, 3 g/L furfural was reduced by only 80% in 4 h (Figure 5.5G).

Similarly, when 2 g/L furfural was used to challenge C. beijerinckii 8052 at the solventogenic phase, the 2 g/L furfural was depleted before 2 h. To collect cells that were treated by furfural for more than one generation time, 3 g/L furfural was used for microarray assay. However, ABE production was shut down (Figure 5.7A-C) followed by accumulation of acetic and butyric acids (Figure 5.7D-E) in the fermentation medium, and the culture did not recover following the depletion of furfural.

To independently verify whether the presence of ABE in the fermentation medium was contributing to the toxicity of furfural to C. beijerinckii 8052 and decreasing furfural reduction during the solventogenic phase, the acidogenic phase culture of C. beijerinckii 8052 was supplemented with 2 g/L furfural together with acetone, ethanol

109

and butanol at concentrations that mimic their concentrations at the solventogenic phase.

Interestingly, this situation reduced the concentration of furfural in the fermentation medium by only 75% after 4 h post-furfural challenge, unlike the control without ABE supplementation, which depleted the furfural in 3 h (Figure 5.8). However, the presence of furfural in the fermentation medium during the solventogenic growth phase did not have remarkable impact on the expression of ABE production genes in C. beijerinckii

8052 (Appendix A: Table A4).

5.5 Discussion

While inhibitory properties of degradation products of lignocellulosic biomass are widely recognized as a major limitation to bioconversion of biomass to biofuels and chemicals (Liu and Blaschek, 2010; Singh et al., 2010; Sakai et al., 2007), the gap in knowledge with respect to detoxification of these lignocellulose-derived inhibitory compounds by fermenting microorganisms continues to impede the development of inhibitor-tolerant strains vis–à–vis commercialization of biofuels. This study presents the physiological changes and transcriptional responses of C. beijerinckii 8052 to furfural challenge at different growth and fermentation stages, highlighting a systematic pattern of gene regulations and revealing potential target genes for strain improvements by genetic engineering.

Genome-wide microarray analysis demonstrated a clear perspective on the effect of the lignocellulosic biomass-derived inhibitor furfural on the transcriptional profile of

C. beijerinckii 8052. This study revealed for the first time that changes in physiological activities of furfural-challenged cultures of C. beijerinckii are coordinated with

110

transcriptional variations during ABE fermentation. Validation of microarray data by Q-

RT-PCR using samples from independent biological treatment showed high degrees of correlation coefficient (R) at both acidogenic and solventogenic phases (0.87 and 0.84, respectively), which fall into the upper values of the reported range (-0.48 to +0.93)

(Etienne et al., 2004), thus, confirming strong reliability of data obtained by microarray analysis. This result is significant because the correlation coefficient is the generally accepted criterion for assessing reliability of microarray data (Morey et al., 2006).

Ramifications of obtained results are discussed below under different attributes.

5.5.1 Redox and cofactor genes are crucial for detoxification of furfural by C. beijerinckii 8052

Furfural challenge increases the expression of genes encoding redox proteins in C. beijerinckii 8052 (Figure 5.1A and Appendix A: Table A3a and A3c). Differential expression of redox genes involved in antioxidant activity suggests that furfural causes oxidative stress in C. beijerinckii 8052. In yeast, furfural induces the accumulation of reactive oxygen species (ROS) that are known to damage DNA, lipids and proteins and that subsequently induce programmed cell death (Allen et al., 2010). Glutathione peroxidases and thioredoxin peroxidases, which were differentially induced in furfural- challenged C. beijerinckii 8052 (Figure 5.1A and Appendix A: Table A3a and A3c), can reduce H2O2 to H2O via oxidation of thiol groups. The reduction of oxidized glutathione and thioredoxin is catalyzed by glutathione reductase or thioredoxin reductase, respectively, using NADPH as the electron donor (Veal et al., 2007). Thioredoxin and glutathione can also function as oxygen quenchers and hydroxyl radical scavengers

111

(Zeller and Klug, 2006; Bisby et al., 1999; Kullisaar et al., 2002). Given that this study was conducted in an anaerobic chamber with less than 1 ppm of molecular oxygen

(monitored by an oxygen detector), the origin of ROS is not clear. However, production of ROS or other radical species by anaerobes has been hypothesized previously (Rocha and Smith, 1999). Organic peroxidize generated during anaerobic condition also required the antioxidative function of thiol peroxidase (Cha et al., 2004). In addtion, a gene encoding thioredoxin family protein dsbA oxidoreductase, a periplasmic oxidoreductase that facilitates disulfide bond formation in proteins, was found in this study to be induced by furfural. Overexpression of dsbA in E. coli increased soluble protein level in the periplasm and improved enzyme secretion and activity (Zheng et al., 2012). Elevated levels of antioxidant activity due to furfural challenge indicate increased oxidative stress, thus, accentuating innate detoxification capabilities of C. beijerinckii under the influence of furfural stress. Additionally, thioredoxin and thioredoxin reductase work in tandem with ribonucleotide reductase during reduction of ribonucleoside diphosphates to deoxyribonucleoside diphosphates (Arnér and Holmgren, 2000). The induced expression of these redox enzymes (Figure 5.1A and Appendix A: Table A3a) involved in purine and pyrimidine metabolism in C. beijerinckii 8052 (Appendix A: Table A2) suggests greater demand of nucleotides due to furfural stress. This premise is supported by the fact that DNA molecules are prone to damage in the presence of furfural (Mills et al., 2009), and in this case, DNA repair or biosynthesis is activated leading to induced expression of redox enzymes.

112

The differentially induced genes encoding oxidoreductases such as aldo/keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR) in C. beijerinckii, which are involved in the reduction of furfural to furfuryl alcohol, have been reported elsewhere as scavengers of furfural in Escherichia coli (Miller et al., 2009b; Jarboe,

2011), Saccharomyces cerevisiae (Liu et al., 2008), and Zymomonas mobilis (Agrawal and Chen, 2011) fermentations. Direct reduction of furfural to the less toxic furfuryl alcohol (Zhang et al., 2012) is another strategy C. beijerinckii 8052 uses to mitigate toxic effects of furfural.

Moreover, differential expression of genes encoding the iron-sulfur cluster and cobalamin- and riboflavin-associated proteins, was observed in furfural-challenged C. beijerinckii 8052 (Figure 5.1A and Appendix A: Table A3a), thus, accentuating cellular responses to furfural stress by redox balancing because these proteins require cofactors such as NADH and NADPH to facilitate catalysis. Notably, the iron-sulfur cluster plays important roles in electron transfer by redox enzymes, disulfide reduction by ferredoxin:thioredoxin reductase, regulation of gene expressions associated with

Ferredoxin-NADP+ reductase, and iron and sulfur storage in ferredoxins (Johnson et al.,

2005). Similar to iron-sulfur clusters, cobalamin, known as vitamin B12, can also function as redox enzyme cofactors (Schumacher et al., 1997); a typical example is the cobalamin-mediated biodegradation of chloroform by the methanogenic consortium obtained from an anaerobic distillery waste water treatment plant (Guerrero‐Barajas and

Field, 2005). Riboflavin, the redox active moiety of flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), has been shown to be differentially induced in

113

response to furfural challenge during ethanol production by S. cerevisiae (Li and Yuan,

2010). Broadly, these results support the idea that induction of genes encoding redox proteins and cofactors and transformation of furfural to less toxic furfuryl alcohol are important mechanisms that C. beijerinckii 8052 uses to restore redox balance under furfural stress and mitigation of toxic effects of furfural on the cell.

5.5.2 Membrane transporter genes play active role in furfural tolerance and detoxification by C. beijerinckii 8052.

Fluctuations in differential expressions of the membrane transport system, including ATP-binding cassette transporters (ABC-transporter) and phosphotransferase system (PTS), signify possible physiological adaptations in C. beijerinckii 8052 in response to furfural stress (Figure 5.1B). The increased expression of sulfate ABC transporter genes (Figure 5.1B) in C. beijerinckii 8052 may be interpreted following a previously proposed model in E. coli (Miller et al., 2009a) in which furfural depresses sulfur assimilation with concomitant inhibition of cell growth, but supplementing the fermentation medium with sulfur-containing amino acids (cysteine and methionine) reversed cell growth and increased cell tolerance to furfural. It is plausible that under furfural stress, C. beijerinckii 8052 may sense sulfur limitation and consequently elevate expression of sulfate ABC transporter genes in preparation for potential increased absorption of sulfur from the fermentation medium.

The expression of the phosphate-specific transport (Pst) system in C. beijerinckii

8052 was significantly enhanced under furfural stress (Figure 5.1B and Appendix A:

Table A3a). Phosphate is an essential component of nucleotides; hence, it plays a central

114

role in chemical energy and DNA/RNA synthesis. The elevated expression of the Pst system may indicate shortage of intracellular phosphates, thus, the need for increased absorption of phosphorus from the environment. Elsewhere, while Pst in E. coli was demonstrated to have decreased expression in the presence of excessive inorganic phosphate, phosphate limitation induces the expression of Pst (van Veen, 1997).

Similarly, elevated expression of multiple operons encoding ABC transporters for branched-chain amino acid transportation was observed (Figure 5.1B and Appendix A:

Table A3a and A3c). It is conceivable that the biosynthesis of branched-chain amino acids (leucine, isoleucine and valine) in C. beijerinckii 8052 is perturbed when furfural is present in the medium, hence, the induction of genes encoding a related membrane transportation system to mitigate the perturbation. This line of reasoning agrees with the fact that furfural induces the accumulation of reactive oxygen species and superoxide anions, which may damage the synthesis of amino acids, especially the branched chain amino acids (Storz et al., 1990). Therefore, it is reasonable that C. beijerinckii 8052 increases the expression of these ABC transporters to facilitate enhanced absorption of exogenous amino acids.

In contrast to ABC transporters, the phosphotransferase system (PTS) reveals decreased expression in furfural-challenged C. beijerinckii 8052 (Figure 5.1B and

Appendix A: Table A3b and A3d). Since the bacterial PTS plays crucial roles in sugar reception, transport and phosphorylation in addition to regulation of catabolic pathways

(Saier Jr, 2001), the expression level of PTS may reflect the physiological state of cell metabolism and, consequently, could rationalize the low ABE production and premature

115

termination of the C. beijerinckii 8052 fermentation process after furfural challenge at the solventogenic growth phase (Figures 5 - 7).

5.5.3 Furfural influences the adaptation machinery of C. beijerinckii 8052

The two-component signal transduction system (TCS) is a stimulus-response coupling signal transduction machinery that allows bacteria to respond and adapt to changes in a wide range of environmental conditions, such as nutrient assimilation

(Görke and Stülke, 2008), cellular redox state (Fernández-Piñar et al., 2008) and bacterial virulence regulation (Beier and Gross, 2006). Chemotaxis, controlled by TCS, is the cells’ response to stressful environments. In chemotaxis, signals are first sensed by transmembrane receptors known as methyl-accepting chemotaxis proteins (MCPs), which control the autophosphorylation of a kinase protein and then a regulator protein. The regulator protein interacts directly with flagellar proteins that act as motor switches and, thus, controls the swimming pattern of the bacterial cell (Faguy and Jarrell, 1999).

Exposure of C. beijerinckii 8052 to furfural stress elicits repression of genes that code for

MCPs, CheA, CheY, and flagellar proteins (Figure 5.1C), plausibly causing temporary (at the acidogenic phase) and permanent (at the solventogenic phase) defects in the adaptation machinery of C. beijerinckii 8052. However, a previous study (Nichols et al.,

2012) showed that chemotaxis machinery was induced in Pseudomonas strains when using furfural as substrate. It is likely that furfural is an attractant to Pseudomonas strains since these bacteria are able to use furfural as a carbon source and therefore respond chemotactically to furfural. However, furfural acts as a toxin to C. beijerinckii 8052,

116

which can only be transformed but not be metabolized, and therefore many cell functions including chamotaxis may be down-regulated.

In bacteria, the carbon storage regulator (CsrA) is recognized as an activator of glycolysis, acetate metabolism, and flagellum biosynthesis (Dubey et al., 2003) and as a global regulator of bacterial virulence and stress response (Barnard et al., 2004). E. coli csrA (csrA deficient) strains are known to have severe growth problems due to central carbon stress (Wei et al., 2000), and the csrA strain of Helicobacter pylori significantly attenuates its virulence (Barnard et al., 2004). In the presence of furfural, the global regulator CsrA in C. beijerinckii 8052 was significantly repressed (Figure 5.1C and

Appendix A: Table A3b and A3d), which may result in the repression of glycolysis and consequently, may trigger repertoires of transformations in stationary-phase physiology

(Wei et al., 2000). This could rationalize the low ABE production and premature termination of the C. beijerinckii 8052 fermentation process following furfural challenge at the solventogenic growth phase (Figures 5 - 7).

Furthermore, repression of glutamine synthetase (GS) in C. beijerinckii 8052 during ABE fermentation in the presence of furfural may decrease the production of glutamine, which may have undesirable effects with respect to nutrient assimilation and cellular redox balance. The GS strain (glutamine-requiring strain) of Bacillus subtilis was found to cause pleiotropic effects on glucose catabolite repression (Fisher and

Sonenshein, 1984). Moreover, glutamine is a precursor of glutamate, which may be used to synthesize glutathione, an important cellular antioxidant (albeit in reduced form) that mitigates stresses (Matés et al., 2002). Since the product of GS plays an important role in

117

cellular redox balance, the repression of GS may impair the tolerance of C. beijerinckii

8052 to furfural.

5.5.4 Basis for both stimulatory and inhibitory effects of furfural on C. beijerinckii 8052

Addition of furfural (<3 g/L) to the fermentation medium inhibits ABE production by C. beijerinckii 8052 to various degrees regardless of the growth stage

(acidogenic or solventogenic) of the culture (Figure 5.3 and 5). While C. beijerinckii

8052 challenged with furfural during the acidogenic phase experienced short-term ABE production inhibition (Figure 5.3), rapid depletion of furfural in the fermentation medium

(Figure 5.3G), full recovery following exhaustion of furfural in the growth medium

(Figure 5.4), C. beijerinckii 8052 challenged with furfural at the solventogenic growth phase resulted in immediate termination of ABE fermentation (Figures 5 and 6). This finding partly agrees with previous investigations (Ezeji et al., 2007; Zhang et al., 2012), which reported that furfural could stimulate growth and ABE production when added at the beginning of fermentation, but it also expands knowledge in the field by uncovering the fact that furfural is most toxic to C. beijerinckii 8052 during the solventogenic growth phase.

C. beijerinckii 8052 did not recover from the toxic effect of furfural when it was challenged with it at the solventogenic growth phase, yet why didn’t furfural inhibit growth of C. beijerinckii 8052 when it was added either at the beginning of fermentation or during the acidogenic growth phase? The answer may be found in the genes. Genes

GrpE, DnaK and DnaJ in DnaK operon encoding GrpE, DnaK and DnaJ proteins are

118

induced under furfural stress (Appendix A: Table A3a), and they play an important role in mitigating harmful effects of environmental stresses such as UV irradiation (Krueger and Walker, 1984), ethanol (Weng et al., 2001), and butanol (Narberhaus et al., 1992) on microorganisms, and stresses on the cellular chaperone machinery (Schröder et al., 1993).

While GroES and GroEL in the groE operon [which are also highly conserved molecular chaperons and are known to be induced by the presence of butanol (Tomas et al., 2003) were differentially induced by more than three-fold when C. beijerinckii 8052 was challenged with furfural during the solventogenic growth phase (Appendix A: Table

A3c), the operon was differentially induced by less than three-fold when C. beijerinckii

8052 was challenged with furfural during the acidogenic growth phase (Appendix A:

Table A3a). Notably, overexpression of groES and groEL increases production of ABE, tolerance to toxic products, and metabolism in solventogenic Clostridium species (Tomas et al., 2003), but their increased expression during the solventogenic phase was perhaps to overcome (albeit increased expression was not enough to overcome furfural toxicity) the high toxicity of furfural to C. beijerinckii 8052 at this physiological growth phase.

Then, why is furfural more toxic to C. beijerinckii 8052 during the solventogenic phase than during the acidogenic phase? Three hypotheses are proposed. First, the presence of ABE enhances cell membrane fluidity and inhibits cell metabolism (Ezeji et al., 2010), which leads to significant loss of cell functions and weakening of the cellular defense system to furfural. This was underscored by the fact that C. beijerinckii 8052 was unable to reduce 2 g/L furfural in the presence of ABE, unlike the control without ABE, which reduced the entire amount of furfural (Figure 5.8). Second, the biotransformation

119

of furfural, which is catalyzed by NAD(P)H-dependent oxidoreductase (Zhang et al.,

2012), competes with NAD(P)H-dependent dehydrogenase (that catalyzes alcohol production) for NAD(P)H coenzymes (Tummala et al., 2003). The need for NAD(P)H by

NAD(P)H-dependent oxidoreductase to boost cellular defense against furfural is a high priority, which leads to a decrease in the NAD(P)H pool and subsequently impedes alcohol production by NAD(P)H-dependent dehydrogenase. This hypothesis is supported by a previous finding in ethanologenic E. coli (Miller et al., 2009b; Turner et al., 2011), wherein silence of an oxidoreductase involved in furfural conversion relieves the diversion of NAD(P)H away from other important biosynthetic processes, thus, increasing cell growth and furfural tolerance. Competition between oxidoreductase and alcohol dehydrogenase for NAD(P)H is severe at the solventogenic growth phase, during which NAD(P)H is needed for the conversion of butyryl-CoA to butyrylaldehyde and subsequently to butanol, unlike the acidogenic growth phase, during which acids are produced in tandem with NAD(P)H production (Ezeji et al., 2010). Third, while furfural repressed the expression of only two genes involved in cell motility by more than threefold when C. beijerinckii 8052 was challenged with furfural during the acidogenic phase, more than forty genes were differentially repressed by up to 18-fold when C. beijerinckii 8052 was challenged with furfural at the solventogenic phase (Figure 5.1C and Appendix A: Table A3b and A3d). Notably, the non-motile strain of C. acetobutylicum has been shown to produce lower ABE than the motile parent strain

(Gutierrez and Maddox, 1990).

5.6 Conclusions

120

While elevated expression of redox and cofactor genes, heat shock genes, and redox balancing may contribute to enhanced ABE fermentation when C. beijerinckii 8052 was challenged with furfural at the acidogenic phase, stresses emanating from ABE production; redox balance perturbations; and repression of genes that code for the phosphotransferase system, cell motility and flagellar proteins (and combinations thereof) may have caused the premature termination of C. beijerinckii 8052 growth and ABE production following furfural challenge at the solventogenic phase. Transcriptomic and fermentation studies carried out in this work provided a new multifarious basis for both stimulatory and inhibitory effects of furfural on C. beijerinckii 8052 during ABE fermentation. Collectively, this study provided insights that could form the basis for metabolic engineering of C. beijerinckii 8052 for enhanced tolerance of lignocellulose- derived microbial inhibitory compounds, thereby improving bioconversion of lignocellulosic biomass hydrolysates to biofuels and chemicals.

121

References

Agrawal, M., Chen, R.R., 2011. Discovery and characterization of a xylose reductase

from Zymomonas mobilis ZM4. Biotechnol. Lett. 33, 2127-2133.

Allen, S.A., Clark, W., McCaffery, J.M., Cai, Z., Lanctot, A., Slininger, P.J., Liu, Z.L.,

Gorsich, S.W., 2010. Furfural induces reactive oxygen species accumulation and

cellular damage in Saccharomyces cerevisiae. Biotechnol. Biofuels. 3, 2.

Armitage, J.P., Schmitt, R., 1997. Bacterial chemotaxis: Rhodobacter sphaeroides and

Sinorhizobium meliloti - variations on a theme? Microbiology. 143, 3671-3682.

Arnér, E.S.J., Holmgren, A., 2000. Physiological functions of thioredoxin and

thioredoxin reductase. Eur. J. Biochem. 267, 6102-6109.

Babu, M.M., 2004. Introduction to microarray data analysis, in: Computational

Genomics: Theory and Application. Horizon Press, Norwich, pp. 225-249.

Barnard, F.M., Loughlin, M.F., Fainberg, H.P., Messenger, M.P., Ussery, D.W.,

Williams, P., Jenks, P.J., 2004. Global regulation of virulence and the stress response

by CsrA in the highly adapted human gastric pathogen Helicobacter pylori. Mol.

Microbiol. 51, 15-32.

Beier, D., Gross, R., 2006. Regulation of bacterial virulence by two-component systems.

Curr. Opin. Microbiol. 9, 143-152.

122

Bisby, R.H., Morgan, C.G., Hamblett, I., Gorman, A.A., 1999. Quenching of singlet

oxygen by trolox C, ascorbate, and amino acids: effects of pH and temperature. J.

Phys. Chem. A. 103, 7454-7459.

Cha, M., Kim, W., Lim, C., Kim, K., Kim, I., 2004. Escherichia coli periplasmic thiol

peroxidase acts as lipid hydroperoxide peroxidase and the principal antioxidative

function during anaerobic growth, J. Biol. Chem. 279, 8769-8778.

Dubey, A.K., Baker, C.S., Suzuki, K., Jones, A.D., Pandit, P., Romeo, T., Babitzke, P.,

2003. CsrA regulates translation of the Escherichia coli carbon starvation gene, cstA,

by blocking ribosome access to the cstA transcript. J. Bacteriol. 185, 4450-4460.

Etienne, W., Meyer, M.H., Peppers, J., Meyer, R.A., 2004. Comparison of mRNA gene

expression by RT-PCR and DNA microarray, BioTechniques. 36, 618-627.

Ezeji, T., Milne, C., Price, N.D., Blaschek, H.P., 2010. Achievements and perspectives to

overcome the poor solvent resistance in acetone and butanol-producing

microorganisms. Appl. Microbiol. Biotechnol. 85, 1697-1712.

Ezeji, T.C., Blaschek, H.P., 2008. Fermentation of dried distillers' grains and solubles

(DDGS) hydrolysates to solvents and value-added products by solventogenic

clostridia, Bioresour.Technol. 99, 5232-5242.

123

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Faguy, D., Jarrell, K., 1999. A twisted tale: the origin and evolution of motility and

chemotaxis in prokaryotes, Microbiology. 145, 279-281.

Fernández-Piñar, R., Ramos, J.L., Rodríguez-Herva, J.J., Espinosa-Urgel, M., 2008. A

two-component regulatory system integrates redox state and population density

sensing in Pseudomonas putida. J. Bacteriol. 190, 7666-7674.

Fisher, S.H., Sonenshein, A.L., 1984. Bacillus subtilis glutamine synthetase mutants

pleiotropically altered in glucose catabolite repression. J. Bacteriol. 157, 612-621.

Görke, B., Stülke, J., 2008. Carbon catabolite repression in bacteria: many ways to make

the most out of nutrients. Nat. Rev. Microbiol. 6, 613-624.

Guerrero‐Barajas, C., Field, J.A., 2005. Riboflavin‐and cobalamin‐mediated

biodegradation of chloroform in a methanogenic consortium. Biotechnol. Bioeng. 89,

539-550.

Gutierrez, N., Maddox, I., 1990. Isolation and partial characterization of a non-motile

mutant of Clostridium acetobutylicum. Biotechnol. Lett. 12, 853-856.

124

Han, B., Gopalan, V., Ezeji, T.C., 2011. Acetone production in solventogenic

Clostridium species: new insights from non-enzymatic decarboxylation of

acetoacetate. Appl. Microbiol. Biotechnol. 91, 565-576.

Huang, D.W., Sherman, B.T., Lempicki, R.A., 2009. Systematic and integrative analysis

of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44-57.

Jarboe, L.R., 2011. YqhD: a broad-substrate range aldehyde reductase with various

applications in production of biorenewable fuels and chemicals. Appl. Microbiol.

Biotechnol. 89, 249-257.

Johnson, D.C., Dean, D.R., Smith, A.D., Johnson, M.K., 2005. Structure, function, and

formation of biological iron-sulfur clusters. Annu. Rev. Biochem. 74, 247-281.

Krueger, J.H., Walker, G.C., 1984. groEL and dnaK genes of Escherichia coli are

induced by UV irradiation and nalidixic acid in an htpR+-dependent fashion. Proc.

Natl. Acad. Sci. 81, 1499-1503.

Kullisaar, T., Zilmer, M., Mikelsaar, M., Vihalemm, T., Annuk, H., Kairane, C., Kilk, A.,

2002. Two antioxidative lactobacilli strains as promising probiotics. Int. J. Food

Microbiol. 72, 215-224.

Li, B.Z., Yuan, Y.J., 2010. Transcriptome shifts in response to furfural and acetic acid in

Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 86, 1915-1924.

125

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxfication,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Liu, Z.L., Moon, J., Andersh, B.J., Slininger, P.J., Weber, S., 2008. Multiple gene-

mediated NAD (P) H-dependent aldehyde reduction is a mechanism of in situ

detoxification of furfural and 5-hydroxymethylfurfural by Saccharomyces cerevisiae.

Appl. Microbiol. Biotechnol. 81, 743-753.

Matés, J.M., Pérez-Gómez, C., de Castro, I.N., Asenjo, M., Márquez, J., 2002. Glutamine

and its relationship with intracellular redox status, oxidative stress and cell

proliferation/death. Int. J. Biochem. Cell Biol. 34, 439-458.

Miller, E.N., Jarboe, L.R., Turner, P.C., Pharkya, P., Yomano, L.P., York, S.W., Nunn,

D., Shanmugam, K., Ingram, L.O., 2009a. Furfural inhibits growth by limiting sulfur

assimilation in ethanologenic Escherichia coli strain LY180. Appl. Environ.

Microbiol. 75, 6132-6141.

Miller, E.N., Jarboe, L., Yomano, L., York, S., Shanmugam, K., Ingram, L., 2009b.

Silencing of NADPH-dependent oxidoreductase genes (yqhD and dkgA) in furfural-

resistant ethanologenic Escherichia coli. Appl. Environ. Microbiol. 75, 4315-4323.

Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and

tolerance mechanisms in Escherichia coli. Biotechnol. Biofuels. 2, 26.

126

Mitchell, W.J., 1997. Physiology of carbohydrate to solvent conversion by clostridia.

Adv. Microb. Physiol. 39, 31-130.

Morey, J.S., Ryan, J.C., Van Dolah, F.M., 2006. Microarray validation: factors

influencing correlation between oligonucleotide microarrays and real-time PCR. Biol.

Proc. Online. 8, 175-193.

Narberhaus, F., Giebeler, K., Bahl, H., 1992. Molecular characterization of the dnaK

gene region of Clostridium acetobutylicum, including grpE, dnaJ, and a new heat

shock gene. J. Bacteriol. 174, 3290-3299.

Nichols, N.N., Lunde, T.A., Graden, K.C., Hallock, K.A., Kowalchyk, C.K., Southern,

R.M., Soskin, E.J., Ditty, J.L., 2012. Chemotaxis to furan compounds by furan-

degrading Pseudomonas strains, Appl. Environ. Microbiol. 78, 6365-6368.

Quackenbush, J., 2002. Microarray data normalization and transformation. Nat. Genet.

32, 496-501.

Qureshi, N., Bowman, M., Saha, B., Hector, R., Berhow, M., Cotta, M., 2011. Effect of

cellulosic sugar degradation products (furfural and hydroxymethyl furfural) on

acetone-butanol-ethanol (ABE) fermentation using Clostridium beijerinckii P260.

Food Bioprod. Process. 90, 533-540.

127

Rocha, E.R., Smith, C.J., 1999. Role of the alkyl hydroperoxide reductase (ahpCF) gene

in oxidative stress defense of the obligate anaerobe Bacteroides fragilis, J. Bacteriol.

181, 5701-5710.

Saier Jr, M.H., 2001. The bacterial phosphotransferase system: structure, function,

regulation and evolution. J. Mol. Microbiol. Biotechnol. 3, 325-328.

Schmittgen, T.D., Livak, K.J., 2008. Analyzing real-time PCR data by the comparative

CT method, Nat. Protoc. 3, 1101-1108.

Schröder, H., Langer, T., Hartl, F., Bukau, B., 1993. DnaK, DnaJ and GrpE form a

cellular chaperone machinery capable of repairing heat-induced protein damage.

EMBO J. 12, 4137-4144.

Schumacher, W., Holliger, C., Zehnder, A.J.B., Hagen, W.R., 1997. Redox chemistry of

cobalamin and iron-sulfur cofactors in the tetrachloroethene reductase of

Dehalobacter restrictus. FEBS Lett. 409, 421-425.

Servinsky, M.D., Kiel, J.T., Dupuy, N.F., Sund, C.J., 2010. Transcriptional analysis of

differential carbohydrate utilization by Clostridium acetobutylicum. Microbiology.

156, 3478-3491.

Singh, A., Pant, D., Korres, N.E., Nizami, A.S., Prasad, S., Murphy, J.D., 2010. Key

issues in life cycle assessment of ethanol production from lignocellulosic biomass:

challenges and perspectives. Bioresour. Technol. 101, 5003-5012.

128

Storz, G., Tartaglia, L.A., Farr, S.B., Ames, B.N., 1990. Bacterial defenses against

oxidative stress. Trends Genet. 6, 363-368.

Tomas, C.A., Welker, N.E., Papoutsakis, E.T., 2003. Overexpression of groESL in

Clostridium acetobutylicum results in increased solvent production and tolerance,

prolonged metabolism, and changes in the cell's transcriptional program. Appl.

Environ. Microbiol. 69, 4951-4965.

Tummala, S.B., Junne, S.G., Papoutsakis, E.T., 2003. Antisense RNA downregulation of

coenzyme A transferase combined with alcohol-aldehyde dehydrogenase

overexpression leads to predominantly alcohologenic Clostridium acetobutylicum

fermentations. J. Bacteriol. 185, 3644-3653.

Turner, P.C., Miller, E.N., Jarboe, L.R., Baggett, C.L., Shanmugam, K., Ingram, L.O.,

2011. YqhC regulates transcription of the adjacent Escherichia coli genes yqhD and

dkgA that are involved in furfural tolerance. J. Ind. Microbiol. Biotechnol. 38, 431-

439. van Veen, H.W., 1997. Phosphate transport in prokaryotes: molecules, mediators and

mechanisms. Antonie Van Leeuwenhoek. 72, 299-315.

Veal, E.A., Day, A.M., Morgan, B.A., 2007. Hydrogen peroxide sensing and signaling.

Mol. Cell. 26, 1-14.

129

Wei, B., Shin, S., LaPorte, D., Wolfe, A.J., Romeo, T., 2000. Global regulatory mutations

in csrA and rpoS cause severe central carbon stress in Escherichia coli in the presence

of acetate. J. Bacteriol. 182, 1632-1640.

Weng, S.F., Tai, P.M., Yang, C.H., Wu, C.D., Tsai, W.J., Lin, J.W., Tseng, Y.H., 2001.

Characterization of stress-responsive genes, hrcA-grpE-dnaK-dnaJ, from

phytopathogenic Xanthomonas campestris. Arch. Microbiol. 176, 121-128.

Zeller, T., Klug, G., 2006. Thioredoxins in bacteria: functions in oxidative stress response

and regulation of thioredoxin genes. Naturwissenschaften. 93, 259-266.

Zhang, Y., Han, B., Ezeji, T.C., 2012. Biotransformation of furfural and 5-

hydroxymethyl furfural (HMF) by Clostridium acetobutylicum ATCC 824 during

butanol fermentation. N. Biotechnol. 29, 345-351.

Zheng, Z., Chen, T., Zhao, M., Wang, Z., Zhao, X., 2012. Engineering Escherichia coli

for succinate production from hemicellulose via consolidated bioprocessing. Microb.

Cell Fact. 11, 37.

130

Table 5.1 List of 30 genes and sequences of primers used in validation of microarray analysis by Q-RT-PCR

Gene Fold change Forward primer Reverse primer Gene name symbol Microarray Q-RT- sequence (5’-3’) sequence (5’-3’) analysis PCR MerR family a a AGAACACGAA Cbei_ 5.43 2.86 TGAACATTTTC transcriptional TATGCCTATTG 3973 CCTGTGCTTTA regulator AG cobalt ABC a a GCCATTGCTTT CTTTGAGAAC Cbei_ 7.26 2.39 transporter AATAGGTGTA TCTTCATTTTG 3693 ATPase A C Cbei_ acetylglutamate -5.17a -1.69a GTTGTTGGTA TGGAACGTCA 4519 kinase GCGTAGCC GTAAGCAGT a a AATGGGTATC Cbei_ assembly protein 4.06 3.38 AAGTGTTTGG AGGTTCTTTTG 1849 SufB TGCAGCGTGT G a a GTTGTAAATC Cbei_ hypothetical 4.47 4.70 AACAGGAGCC CCCTACCTTGA 1138 protein ATTTTAGCAA C glyceraldehyde- -1.05a, 1.35a, Cbei_ 3-phosphate -1.48 1.25 GGTGCTCAAA GCTTTCATAGC 0597 dehydrogenase, GAGTTCCA AGCGTTA type I Cbei_ alcohol 3.78a 4.16a CTAAAAGAGC AAACGCCACG 1464 dehydrogenase AGGGGCAGAT TCAACTCC Cbei_ type II secretion 4.41a 1.95a GCAGCTATAA CGCAATAACT 4218 system protein E CAGGACATTT CCCACAAC Cbei_ hypothetical 80.04 873.10 TGCAGTAGCG TAATCCTGCG 2057 protein ATTGAACA GCTAAGAA groEL TAGAAGAGCC Cbei_ TCCACTATTCC chaperonin 3.84 1.87 AGTAAGACAA 0329 ACCTTTTATC GroEL A Cbei_ pyruvate 6.11 55.72 CATATTGGGA GCCTAAAGCA 0315 formate-lyase TGGACACTCA AGCCACTC Cbei_ acetyl-CoA 3.04 1.77 GGTGATGCTG TTCTCCATGTC 3630 acetyltransferase ATGCTATCG CCATTCTTT (Continued)

131

Table 5.1: Continued.

heavy metal ACAGCAACTC Cbei_ CCTAAAGACG transport/detoxifi 4.26 1.32 ATACCACTAC 1435 GGGACATA cation protein TA PTS system, N- Cbei_ acetylglucosami TAAAGGGTGT CCAGATACAC 4532 ne-specific IIBC -17.89 0.34 TGGAGGAA CTGCTGATTT subunit Cbei_ alcohol -5.74 -16.68 CTTTATGTTGG CCAATAATCC 0685 dehydrogenase GGTTTGC TACCCTCTTC Cbei_ carbon starvation -12.64 -4.89 CTATCTGGGTT GCCATTAGGG 0554 protein CstA CCATTCAC AGACAAAA 2-oxoglutarate Cbei_ ferredoxin GAGAAGGGGA TCCGATGTCA 4041 oxidoreductase -3.69 -1.20 TTCTTATGG GTTGTAGGT subunit beta aldehyde oxidase and xanthine Cbei_ CAGGACTGGG GTCGGACAAG dehydrogenase 1982 ATGTAATGAT TAATGGGT molybdopterin- -3.23 -1.26 binding subunit AATGGGTATC Cbei_ FeS assembly AAGTGTTTGG 24.45 114.56 AGGTTCTTTTG 1849 protein SufB TGCAGCGTGT G 14.03a , 5.39a, Cbei_ hypothetical ACAGGCGTTA CCTTGATGTTG 2445 protein 18.03 19.56 TTTTAGCGAG AACTGCTGAC

2.37a, Cbei_ thioredoxin GGTAGAGCCG AACCTGTTAT a 2681 reductase 3.18 , 9.82 53.82 GATTAGATGC GCTGCGAAAA

20.39a , 3.67a, XRE family AGACCAAATT Cbei_ CATCTTGCCAC transcriptional CGCATGATTT 3616 AAACCTTCTTC regulator 6.69 3.43 AGA

(Continued)

132

Table 5.1: Continued.

2.33a, CCATTAGCCC Cbei_ aldo/keto CGCCATTTGA a AAGGGACATT 3974 reductase 9.06 , 3.87 GCAAGAGTTT 2.45 A response -4.66a , - regulator a Cbei_ receiver sensor 1.12 , AACCTGGCGA GGCTGAACCT 2725 signal 17.21 AGGAACTG TCCCCATA transduction 0.81 histidine kinase -3.41a , - Cbei_ coenzyme A TGGAGCATCA TTCCCTGCTTG 3278 transferase 4.41 1.68 ATAAACCC TCTACTTCT

-3.81 a , - -3.76a, Cbei_ MotA/TolQ/Exb GAGGGTTACA CTATAATCTTC 4273 B proton channel 4.76 -1.16 AATGGTGG ATGCCTTGC

PTS system -3.71a , a mannose/fructos -1.14 , Cbei_ CTGGGGAACA CCATAGAACA e/sorbose family 4911 -3.89 CTAAGACCT TTCCATACCA transporter 2.28 subunit IID 1.54a, iron-containing Cbei_ AGTTATTGCG GAACCAGTCG alcohol 2421 a 2.16 GCAGGAGT CTGATAGTGT dehydrogenase 2.73 , 1.26

-1.44a , - -1.04a, Cbei_ aldo/keto AGAGGATTAA GCCCATTCTGC 2676 reductase 1.62 2.10 AGGCTGCTAA TGGAGTA

2.38a, Short-chain GCAGCCACAA TTCGGTATTTA Cbei_ dehydrogenase/ a AAGGAGCAGT TTGGACCAGG 3904 4.14 , 1.92 reductase 1.68 T AG aFold change when C. beijerinckii 8052 was challenged with furfural at acidogenic phase; values without superscript represent fold change when C. beijerinckii 8052 was challenged with furfural at solventogenic phase. 133

Figure 5.1 Comparison of gene expression after furfural challenge at acidogenic and solventogenic phases. The findings are grouped into different attributes: redox and cofactor genes (A), membrane transporter genes (B), and two-component signal transduction system and chemotaxis genes (C).

134

(Continued)

135

Figure 5.1: Continued.

(Continued)

136

Figure 5.1: Continued.

137

2 A

Ratio

2 R  0.87 0

-1

-2

Q-RT-PCR Log -3

-4 -3 -2 -1 0 1 2 3 4 5

Microarray Log2 Ratio

12 10 B 8

Ratio 6

2 4 R  0.84 2

-2

Q-RT-PCR Log -4

-6 -6 -4 -2 0 2 4 6 8

Microarray Log2 Ratio

Figure 5.2 Microarray validation of furfural-challenged gene expression during acidogenic phase (A) and solventogenic phase (B).

138

4.0 1.6 2.0

1.4 B 1.6 C 3.0 A 1.2 1.2 2.0 1.0 600nm 0.8 0.8 OD 1.0

Butanol (g/L) Acetone (g/L) 0.6 0.4

0.0 0.4 0.0 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h) 0.4 5.5 1.4

5.0 E 1.3 F 0.3 D 4.5 1.2 0.2 4.0 1.1

Ethanol (g/L) 0.1 3.5 1.0

Acetic acid (g/L)

Butyric acid (g/L) Butyric

0.0 3.0 0.9 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

2.0 G 1.5

1.0 P2 P2 + Furfural 2 g/L 0.5

Furfural(g/L)

0.0 0 1 2 3 4 Time post-challenge (h)

Figure 5.3 Cell growth, solvent production and furfural reduction after furfural challenge at the acidogenic phase. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until acidogenic phase. The fermentation broth was then divided to two aliquots, one of which was added with 2 g/L furfural and the other left unchallenged as control. Control and treatment samples were taken every hour until furfural depletion to analyze cell growth (A), production of acetone (B), butanol (C), ethanol (D), acetic acid (E) and butyric acid (F), and reduction of furfural in the treatment medium (F).

139

8 14 0.7 7 12 0.6 C 6 A B 10 0.5 5 8 0.4 4 0.3 3 6 4 0.2 2 Ethanol (g/L)

Acetone (g/L)

Butanol (g/L) 1 2 0.1 0 0 0.0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

5.5 1.4 5.0 D 1.2 E 4.5 4.0 1.0 P2 3.5 0.8 A P2 + Furfural 2 g/L 3.0 0.6 2.5

Acetic acid (g/L) 0.4 2.0 acid (g/L) Butyric 1.5 0.2 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Time post-challenge (h) Time post-challenge (h)

Figure 5.4 ABE production after furfural challenge at the acidogenic phase. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until acidogenic phase. The fermentation broth was then divided to two aliquots, one of which was added with 2 g/L furfural and the other left unchallenged as control. Control and treatment samples were taken every hour until furfural depletion and then every 12 h until end of fermentation. Production of acetone (A), butanol (B), ethanol (C), acetic acid (D) and butyric acid (E) were analyzed.

140

5.0 8.5 6.0 A 4.5 B 8.0 C 4.5 7.5 4.0 7.0

600nm 3.0 3.5 6.5

Butanol (g/L) Butanol

1.5 Acetone (g/L) 3.0 6.0

0.0 2.5 5.5 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

0.5 2.8 1.5 D E F 0.4 2.7 1.2

2.6 0.3 0.9 2.5

Ethanol (g/L) Ethanol 0.2 0.6 2.4

Acetic(g/L) acid

Butyric acid (g/L) acid Butyric

0.1 2.3 0.3 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

3.0

2.5 G

2.0

1.5 P2 1.0 P2 + Furfural 3 g/L

Furfural (g/L) Furfural 0.5

0.0 0 1 2 3 4 Time post-challenge (h)

Figure 5.5 Cell growth, solvent production and furfural reduction after furfural challenge at the solventogenic phase. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until solventogenic phase. The fermentation broth was then divided to two aliquots, one of which was added with 3 g/L furfural and the other left unchallenged as control. Control and treatment samples were taken every hour to analyze cell growth (A), production of acetone (B), butanol (C), ethanol (D), acetic acid (E) and butyric acid (F), and reduction of furfural in the treatment medium (F).

141

0.8 6.0 A 15.0 B C 0.6 12.5 4.5 10.0 0.4

3.0 7.5 Etanol (g/L) Etanol 0.2

Butanol (g/L) Butanol

Acetone (g/L) Acetone 5.0 1.5 0.0 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

3.5 2.0 D E 3.0 1.6

2.5 1.2 A 2.0 0.8 P2 1.5 0.4 P2 + Furfural 3 g/L

Acetic acid (g/L) acid Acetic

Butyric acid (g/L) acid Butyric 1.0 0.0 0 10 20 30 40 0 10 20 30 40 Time post-challenge (h) Time post-challenge (h)

Figure 5.6 ABE production after furfural challenge at the solventogenic phase phase. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until solventogenic phase. The fermentation broth was then divided to two aliquots, one of which was added with 3 g/L furfural and the other left unchallenged as control. Control and treatment samples were taken every hour during the first 4 h post-challenge and then every 12 h until end of fermentation. Production of acetone (A), butanol (B), ethanol (C), acetic acid (D) and butyric acid (E) were analyzed.

142

5.5 0.6

5.0 12.0 A B 0.5 C 4.5 10.5 4.0 0.4 9.0 3.5 0.3 3.0 7.5

Ethanol (g/L) Ethanol

Butanol (g/L) Butanol Acetone (g/L) Acetone 0.2 2.5 6.0 2.0 0.1 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 Time post-challenge (h) Time post-challenge (h) Time post-challenge (h)

3.5 1.8 D 1.6 E 3.0 1.4 1.2 A P2 2.5 1.0 P2 + Furfural 2 g/L 0.8

Acetic acid (g/L) acid Acetic 2.0

Butyric acid (g/L) acid Butyric 0.6 0.4 0 10 20 30 40 0 10 20 30 40 Time post-challenge (h) Time post-challenge (h)

Figure 5.7 ABE production after furfural challenge at the solventogenic phase phase. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until solventogenic phase. The fermentation broth was then divided to two aliquots, one of which was added with 2 g/L furfural and the other left unchallenged as control. Control and treatment samples were taken every hour during the first 4 h post-challenge and then every 12 h until end of fermentation. Production of acetone (A), butanol (B), ethanol (C), acetic acid (D) and butyric acid (E) were analyzed.

143

2.5

2.0

1.5

1.0

Furfural (g/L) Furfural

0.5

0.0 0 1 2 3 4 Time post-challenge (h)

P2 + Furfural 2 g/L P2 + ABE + Furfural 2 g/L

Figure 5.8 Reduction of furfural by acidogenic C. beijerinckii 8052 culture challenged with furfural and supplemented ABE. Batch ABE fermentation was conducted by C. beijerinckii 8052 using P2 medium until acidogenic phase. The fermentation broth was then divided to two aliquots: one was added with 2 g/L furfural, and the other was added with 2 g/L furfural plus acetone (2.80 g/L), ethanol (0.15 g/L) and butanol (5.53 g/L) to bring the final concentrations of acetone, ethanol and butanol to 3.40 g/L, 0.22 g/L and 5.93 g/L, respectively. Decrease of furfural was detected every hour post challenge until depletion.

144

Chapter 6: Purification and Characterization of Aldo/keto Reductase and Short-chain Dehydrogenase/reductase Catalyzing NADPH-coupled Furfural Reduction in Clostridium beijerinckii NCIMB 8052

6.1 Abstract

Butanol fermentation by Clostridium beijerinckii NCIMB 8052 using lignocellulosic biomass hydrolysates as substrate is challenging due to microbial inhibitory compounds generated during biomass pretreatment. Furfural is one of the representative inhibitors that interfere with C. beijerinckii 8052 growth and acetone butanol ethanol (ABE) fermentation although this microorganism can tolerate and transform low levels of furfural in the fermentation medium to lesser inhibitory compound, furfuryl alcohol. To understand the mechanism of furfural biotransformation by C. beijerinckii 8052, two oxidoreductases, aldo/keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR), were cloned and expressed as histidine tagged proteins in an E. coli strain Rosetta-gami™ B(DE3)pLysS, and purified by immobilized metal affinity chromatography. Protein gel analysis showed that the molecular mass of the purified AKR and SDR were close to the expected values of 37 kDa and 27 kDa, respectiviely. While AKR has apparent Km and Vmax values for furfural of 32.4 mM and

254.2 mM/s, respectively, SDR has lower Km (26.4 mM) and Vmax (22.6 mM/s) for furfural than AKR. AKR and SDR showed broad substrate specificity and catalyzed the 145

reduction of different aldehydes in the presence of NADPH. Based on these results, it is proposed that AKR and SDR may be involved in the biotransformation of furfural to furfuryl alcohol by C. beijerinckii 8052.

6.2 Introduction

Lignocellulosic biomass such as agricultural residuals and dedicated energy crops can serve as feedstock for the production of liquid biofuels such as ethanol (Sheehan et al., 2003; Saha et al., 2005; Sørensen et al., 2008) and butanol (Qureshi et al., 2010).

Lignocellulosic biomass is a more attractive substrate than the conventional biofuel substrate in the US such as corn because of its abundance, low production cost and the alleviation of food-fuel competition (Stöcker, 2008). However, one of the major barriers in its conversion to biofuel is the presence of degradation products generated during biomass pretreatment and hydrolysis to fermentable sugars, which subsequently may inhibit the growth of fermenting microorganisms during fermentation (Parajó et al.,

1998). These microbial inhibitors include aldehydes, phenols and organic acids (Liu and

Blaschek, 2010). Furfural and 5-hyroxymethylfurfural, which are generated from dehydration of xylose and glucose, respectively, may disrupt cell membrane, damage nucleic acid, repress protein synthesis, inhibit enzyme activities, induce oxidative stress, and repress substrate utilization (Ezeji et al., 2007; Mills et al., 2009; Allen et al., 2010), and these undesirable by-products of sugars are commonly present in low cost dilute acid pretreatedbiomass (Liu and Blaschek, 2010).

Previous studies have shown that solventogenic Clostridium species depleted relatively low levels of furfural or HMF (≤ 2 g/L) in the fermentation medium when

146

Miscanthus hydrolysates or P2 medium supplemented with furfural or HMF was used as substrate for butanol fermentation (Zhang et al., 2012; Chapter 3; Chapter 4). Although furfural has been reported to be oxidized to furoic acid by a Pseudomonas strain (Koenig and Andreesen, 1990), solventogenic Clostridium species such as C. acetobutylicum 824 and C. beijerinckii 8052 convert furfural to furfuryl alcohol through a one-step reduction process, which is consistent with previous reports on other microorganisms including yeast, bacteria, and archaea (Zhang et al., 2012). A transcriptome analysis study, which was conducted to examine the response of C. beijerinckii 8052 to furfural stress at the mRNA level, showed that the presence of furfural in the fermentation medium increased the expression of genes encoding redox proteins in C. beijerinckii 8052 (Chapter 5).

Notably, oxidoreductases are known to be involved at mitigating the secondary oxidative effect of furfural (e.g. reactive oxygen species and inappropriate disulfide bond formation in proteins) on microbial cells (Allen et al., 2010). Moreover, oxidoreductases such as aldo/keto reductase (AKR), encoded by gene Cbei_3974, and short-chain dehydrogenase/reductase (SDR), encoded by gene Cbei_3904, have been proposed to be used by C. beijerinckii to catalyze the transformation of furfural to furfuryl alcohol

(Chapter 5). Several enzymes in AKR family have been known to reduce furfural and

HMF, such as ZMO0976 from Zymomonas mobilis ZM4 (Agrawal and Chen, 2011),

YqhD and DkgA from Escherichia coli (Miller et al., 2009), and GRE3 from

Saccharomyces cerevisiae (Liu et al., 2008). Enzymes belonging to SDR family have also shown furfural reduction activity, for example, Ari1 from S. cerevisiae (Liu, 2011).

Moreover, deduced Cbei_3974 (AKR) protein sequence shows good sequence similarity

147

with GRE3 protein of S. cerevisiae (query coverage 92% and E value 1e−11), indicating potential similarity on furfural conversion. However, no enzyme from Clostridium species has been reported to catalyze the reduction of furfural or other lignocellulose- derived aldehyde to the corresponding less microbial inhibitory alcohols. The objective of this study was to purify Cbei_3974 (AKR) and Cbei_3904 (SDR) proteins by molecular cloning, heterologous protein expression, purification, and characterization in terms of reaction temperature, pH, kinetic parameters, and substrate specificity, in an effort to understand the mechanism with which C. beijerinckii tolerates and transforms aldehydes into alcohols.

6.3 Materials and Methods

6.3.1 Bacterial strains, plasmids and culture conditions

Clostridium beijerinckii NCIMB 8052 (ATCC 51743) was obtained from the

American Type Culture Collection (Manassas, VA). Escherichia coli DH5α from New

England Biolabs (Ipswich, MA) was used in constructing and maintaining of recombinant plasmids. E. coli strain Rosetta-gami™ B(DE3)pLysS competent cells was purchased from EMD Biosciences (San Diego, CA), and used as host for recombinant protein expression. Plasmid vector pET-15b was purchased from EMD Millipore (Billerica,

MA). Recombinant plasmids pET-15b_3974 and pET-15b_3904 contain coding regions of Cbei_3974 aldo/keto reductase (AKR) and Cbei_3904 short-chain dehydrogenase/reductase (SDR), respectively.

Laboratory stocks of C. beijerinckii 8052 were routinely maintained as spore suspensions in sterile double distilled water at 4 °C. To revive C. beijerinckii 8052

148

spores, 200 μL stock was heat-shocked for 10 min at 75 °C followed by cooling on ice.

The heat-shocked spores were inoculated into 10 mL anoxic pre-sterilized tryptone– glucose–yeast extract (TGY) medium and incubated in an anaerobic chamber (Coy

Laboratory Products Inc., Ann Arbor, Michigan) as previously described (Ezeji and

Blaschek, 2008). E. coli DH5α was grown in Lysogeny broth (LB) medium, and Rosetta- gami™ B(DE3)pLysS was grown in LB medium supplemented with chloramphenicol

(34 μg/ml) to maintain its plasmid pRARE. E. coli strains (DH5α and Rosetta-gami™

B(DE3)pLysS) with recombinant plasmids pET-15b_3974 or pET-15b_3904 were grown in LB medium with additional ampicillin (50 μg/ml).

6.3.2 Cloning of AKR and SDR genes

The genomic DNA of C. beijerinckii 8052 was prepared according to the method described previously (Cornillot et al., 1997). The AKR and SDR gene fragments were amplified by polymerase chain reaction (PCR) using C. beijerinckii 8052 genomic DNA as template and Phusion® High-Fidelity DNA Polymerase (New England Biolabs,

Ipswich, MA). The PCR primers (synthesized by Eurofins MWG Operon, Huntsville,

AL) were designed by Primer Premier 5 (PREMIER Biosoft Int., Palo Alto, CA) according to the open reading frame (ORF) of each gene (Table 6.1).

The PCR reaction conditions consist of the following steps: step 1, 98 °C for 30 sec (initial denaturation), step 2, 98 °C for 10 sec (denaturation), step 3, 59 °C for 30 sec

(annealing), step 4, 72 °C for 30 sec (extension), 35 cycles of step 2, 3, and 4, step 5, 72

°C for 5 min (final extension)The PCR product obtained was digested with NdeI and ClaI for AKR, and BamHI and ClaI for SDR, whose recognition sites were included in the

149

PCR primers indicated in the underlined sequences (Table 6.1). Restriction enzymes used in this study were purchased from New England Biolabs (Ipswich, MA). Vector pET-15b was isolated from E. coli using the GenCatch plus plasmid DNA miniprep kit

(Epoch Life Science, Sugar Land, TX).The vector was digested by appropriate restriction enzymes, NdeI and ClaI for cloning of Cbei_3974, and BamHI and ClaI for cloning

Cbei_3904. The digested vectors and PCR products were purified by agarose gel electrophoresis using the GenCatch advanced PCR extraction kit (Epoch Life Science,

Sugar Land, TX). The ligation of PCR fragment and vector was performed at 4 °C overnight by T4 DNA ligase (New England Biolabs, Ipswich, MA) in an insert:vector proportion 10:1. The resulting ligation mixture was purified using the GenCatch advanced PCR extraction kit (Epoch Life Science, Sugar Land, TX). Electroporation of recombinant DNA into E. coli DH5α was performed in 1mm cuvette using a Bio-Rad

Gene Pulser Xcell™ system set at 1.8 KV, 25 μF capacitance, and 200Ω resistance.

Electroporation duration time was automatically set by the Bio-Rad Gene Pulser Xcell™ machine (Richmond, CA) which was usually between 4.3 to 4.4 milliseconds. Following electroporation, cells were diluted in 1ml SOC medium (2% w/v bacto-tryptone, 0.5% w/v Yeast extract, 10mM NaCl, 2.5mM KCl, 10mM MgCl2, and 20mM glucose) and incubated for 1 h at 37°C with shaking at 250 rpm. The recovered cells were plated on

LB agar with ampicillin (50 μg/ml), and colonies containing recombinant plasmids were verified by PCR and agarose gel electrophoresis. Recombinant plasmids, pET-15b_3974 or pET-15b_3904, were amplified by the growth of E. coli transformants in LB medium supplemented with ampicillin, purified using the GenCatch plus plasmid DNA miniprep

150

kit (Epoch Life Science, Sugar Land, TX), and stored at - 80°C. The sequences of cloned genes were verified via DNA sequencing performed by the Plant–Microbe Genome

Facility at The Ohio State University (Columbus, OH).

6.3.3 Expression of AKR and SDR in Rosetta-gami™ B(DE3)pLysS

The expression plasmids, pET-15b_3974 or pET-15b_3904, were transformed into Rosetta-gami™ B(DE3)pLysS competent cells according to the manufacturer's instructions. The transformed strains were selected on LB plates supplemented with 34

μg/ml chloramphenicol and 50 μg/ml ampicillin, and grown at 37 °C overnight. A single colony was picked from the plate and cultured in 5 mL LB medium with the same antibiotics overnight at 37 °C. The culture was then inoculated into 250 mL of the same medium followed by incubation in a rotary shaker at 250 rpm and 37 °C. When the culture reached optical density (OD600) of 0.6 as measured by a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA), isopropyl-β-D- thiogalactopyranoside (IPTG) was added to a final concentration of 400 μM followed by incubation at 20 °C for 6 h to induce the expression of the recombinant proteins AKR and

SDR. The control cultures grown in parallel included host strain (Rosetta-gami™

B(DE3)pLysS) with vector pET-15b without and with IPTG induction, plus host strain with plasmid pET-15b_3974 and pET-15b_3904 without IPTG induction. Induced and uninduced cells were harvested by centrifugation at 1,500 × g for 15 min at 4 °C

(Beckman Model J2 21 M induction drive centrifuge and rotor JA-17, Beckman, Palo

Alto, CA) followed by storage at – 80 °C.

6.3.4 Purification of AKR and SDR

151

E. coli cell pellet from 50 ml culture was disrupted by suspension in 3ml

BugBuster protein extraction reagent (Novagen, EMD Chemicals Inc., San Diego, CA) supplemented with 10 μl DNase I (New England Biolabs, Ipswich, MA) and 3 mM of

MgCl2, followed by incubation at room temperature for 20 min. The insoluble cell debris was removed by centrifugation at 12,000 × g for 20 min at 4 °C. The supernatant was transferred into a fresh tube and an aliquot was saved for protein analysis by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The supernatant (crude protein) was loaded onto metal ion affinity chromatography His GraviTrap column (GE

Healthcare, Uppsala, Sweden) pre-equilibrated with 10 ml of binding buffer (20 mM sodium phosphate, 500 mM NaCl, 20 mM imidazole, and 20% glycerol, pH 7.4). The column with loaded crude protein was washed by 10 ml binding buffer followed by the addition of 1 ml elution buffer (20 mM sodium phosphate, 500 mM NaCl, 500 mM imidazole, and 20% glycerol, pH 7.4). The elution was repeated three more times to collect 4 fractions of 1 ml each. Eluted fractions were loaded onto Zeba™ Spin Desalting

Columns (Thermo Scientific, Rockford, IL) to exchange elution buffer to 25 mM potassium phosphate buffer (pH 7.0) with 20% glycerol following manufacturer’s protocols. Purified AKR and SDR were stored at – 20 °C until further use.

6.3.5 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)

The purity and molecular weights of AKR and SDR were assessed by SDS-

PAGE (resolving gel, 10% acrylamide/bis-acrylamide, pH 8.8; and stacking gel, 5% acrylamide/bis-acrylamide, pH 6.8). The electrophoresis was carried out with a constant voltage of 130 V. Standard protein markers used for the estimation of AKR and SDR

152

molecular weights was Pierce prestained protein molecular weight marker

(Pierce/Thermo Scientific, Rockford, IL). After electrophoresis, gels were transferred into Coomassie Blue R250 solution and swirled gently at room temperature for 60 min to stain. Removal of excess stain was accomplished by transferring the stained gel into a destaining solution containing 45% (v/v) of methanol and 10% (v/v) of acetic acid followed by gentle swirling for about 30 min at room temperature.

6.3.6 AKR and SDR assay

To quantify activities of AKR and SDR, a reaction mixture consisting of 25 mM potassium phosphate buffer (pH 7.0), 40 mM furfural, 0.4 mM NADPH, and 20 μl AKR or SDR (total reaction volume, 0.5 ml) was prepared as described previously (Gutiérrez et al., 2006). The assay was conducted at 40 °C and activity was measured by the rate of

NADPH oxidation to NADP+ at 340 nm using a DU800 spectrophotometer (Beckman

Coulter Inc., Brea, CA). A circulating water bath was connected directly to the cuvette compartment of the spectrophotometer to control reaction temperature. Reaction reading was blanked after the addition of furfural to the buffer. A 5-min equilibration time was allowed after the addition of NADPH to the reaction mixture. The reaction was initiated by the addition of AKR or SDR to the mixture followed by reading at 10 sec intervals for

3 min. The initial velocity of the reaction is the slope of the initial linear portion of the resulting curve. One unit of enzyme activity was defined as the number of micromoles of

NADPH oxidized per minute, and the specific activity was defined as the number of micromoles of NADPH oxidized per minute per milligram protein. Protein

153

concentrations were determined using the Bradford assay (Amresco®, Bradford Reagent,

Solon, OH) using bovine serum albumin as the standard.

6.3.7 Characterization of AKR and SDR enzymes

To determine optimal temperature, the activities of AKR and SDR were measured using 25 mM potassium phosphate buffer (pH 7.0) at different temperatures ranging from

20 °C to 65 °C. The optimal pH for the activity of AKR or SDR was determined using 25 mM potassium phosphate buffer (pH 4.5 – 8.5) and 25 mM sodium acetate buffer (pH

4.0) at 40 °C. To study kinetic properties of AKR and SDR, initial reaction velocity was measured at furfural concentrations ranging from 2 mM to 50 mM. Kinetic constants Km and Vmax values were determined according to Michaelis-Menten equation using KaleidaGraph (Synergy Software, Reading PA). Kcat value was calculated as the ratio of Vmax to total amount of enzyme. Substrate specificity of AKR and SDR were analyzed by evaluating activities of AKR and SDR using hydroxymethylfurfural (HMF), benzaldehyde, butyraldehyde, and furfural as substrates.

6.4 Results and Discussion

6.4.1 Construction of recombinant plasmids

The open reading frame (ORF) of Cbei_3974 has 996 bp with an ATG start codon and a TAG stop codon. It encodes a protein of 331 amino acids residues with a calculated molecular mass of 37,239 Da. This protein was predicted to be an oxidoreductase that may belong to the superfamily of aldo/keto reductase (AKR). Additionally, Cbei_3904 has a length of 741 bp with an ATG start codon and a TAA stop codon, and it was predicted to code for another oxidoreductase that may belong to short-chain

154

dehydrogenase/reductase (SDR) family, formerly known as short-chain alcohol dehydrogenase. This SDR has 246 amino acids residues with a calculated molecular mass of 26,814 Da. Transcriptional analysis of C. beijerinckii 8052 grown in the presence of furfural showed Cbei_3974 and Cbei_3904 genes were significantly up-regulated

(Chapter 6), and AKR and SDR have been reported to be involved in furfural reduction to furfuryl alcohol in many other species (Zhang et al., 2012). Therefore, it was hypothesized that Cbei_3974 and Cbei_3904 may express proteins with capacity to mitigate toxic effect of furfural on C. beijerinckii 8052 by catalyzing the conversion of furfural to furfuryl alcohol. To test this hypothesis, Cbei_3974 and Cbei_3904 were cloned and overexpressed in E. coli followed by in vitro characterization of purified enzymes. Briefly, the full length of Cbei_3974 and Cbei_3904 were successfully amplified using genomic DNA of C. beijerinckii 8052 as template, and then cloned in expression vector (pET-15b) under the control of a strong T7 promoter (Figure 6.1). The cloned AKR and SDR sequences were tagged with a sequence on 5’ end encoding six successive histidine residues that facilitate target protein purification (Figure 6.1). To amplify and purify the resulting recombinant plasmids, pET-15b_3974 and pET-

15b_3904 were transformed into E. coli DH5α followed by selection on LB plates containing ampicillin. PCR amplicons generated from gene-specific primers confimed the successful cloning of the target gene, and DNA sequencing results confirmed insertion site, sequence orientation, and the absence of mutation.

6.4.2 Expression and purification of AKR and SDR

155

E. coli Lemo21(DE3) (New England Biolabs, Ipswich, MA) was first used as the host strain to express AKR and SDR. While this E. coli strain successfully expressed the

AKR encoded by Cbei_3974, expression of SDR from Cbei_3904 by this strain was not successful (data not shown). To troubleshoot this occurrence, codon usage frequency of

Cbei_3974 and Cbei_3904 in E. coli was analyzed by E. coli Codon Usage Analyzer 2.1

(http://faculty.ucr.edu/~mmaduro/codonusage/usage.htm). Figure 6.4 shows that 29% of the Cbei_3974 codons have a usage frequency less than 10% in E. coli, and 24% of the

Cbei_3904 codons are rarely used by E. coli. It has been suggested that the presence of rare codons near the 5’ of the transcripts likely affects transcriptional efficiency (Hannig and Makrides, 1998). A closer examination of codon usage shows that as high as 50% of the codons that encode the first 30 amino acid residues in the 5’terminus of Cbei_3904 are rare in E. coli, compared to only 27% of that in Cbei_3974. Also, the second codon of

Cbei_3904 is of extremely rare use by E. coli (frequency ≤ 1%), and there are 3 and 5 successive very rare codons. As suggested previously (Gustafsson et al., 2004), these rare codons may explain why E. coli Lemo21(DE3) failed to express SDR.

Given the inability of E. coli Lemo21(DE3) to express SDR, E. coli strains that can correct condon bias were sought. Rosetta-gami™ B(DE3)pLysS was chosen as the host strain because it supplies six rare tRNAs to reduce codon usage bias in E. coli, hence improving translation efficiency and accuracy of heterologous proteins (Terpe, 2006).

Since the induction of target protein in the E. coli host is dependent on IPTG concentration, different doses of IPTG (100 µM, 400 µM, and 1000 µM) were tested and the induced protein expressions were compared using the SDS-PAGE (Figure 6.2). The

156

result shows that IPTG induced the expression of both AKR and SDR at all concentrations tested, but 400 µM IPTG gave the maximum protein expression level.

Controls for each protein expression include cell lysate from host strain (Rosetta-gami™

B(DE3)pLysS) with vector pET-15b without and with IPTG induction (Lane 2 and Lane

3, respectively, Figure 6.3), plus host strain with plasmid pET-15b_3974 (Lane 4, Figure

6.3a) and pET-3904 (Lane 4, Figure 6.4b) without IPTG induction. None of the controls expressed a high level of protein whose molecular weight is equal or close to that of AKR or SDR. However, large bands with approximate molecular weights of 37 kDa for AKR and 27 kDa for SDR were obtained from the IPTG induced strain harboring plasmids with target genes (Lane 5, Figure 6.3a for AKR and 3b for SDR). The obtained molecular weights of AKR and SDR are consistent with the predicted size (AKR, 37,239 Da and

SDR, 26,814 Da).

Purifications of histidine-tagged AKR and SDR were carried out using immobilized metal affinity chromatography (IMAC) in gravity-flow columns. To load onto purification column, soluble fraction of crude cell extract was obtained by gentle disruption of induced cultures using BugBuster protein extraction reagent. Cell lysate, flow-through, washes, and elution fractions obtained during the purification process were analyzed by SDS-PAGE (Figure 6.5). For both proteins, the first 1-ml of elution buffer did not efficiently elute the target protein from the columns. Second fraction yielded target proteins of approximately 95% purity for AKR (Lane E2, Figure 6.5a) and 98% for

SDR (Lane E2, Figure 6.5b). The subsequent elution fractions gave higher purities but lower protein concentrations (Lane E3 – E5, Figure 6.5). Enzyme activities were

157

measured before and after buffer exchange from elution buffer to storage buffer but no remarkable difference was observed, indicating that AKR and SDR are not sensitive to up to 0.5 M of NaCl and imidazole. A summary of the AKR and SDR purification is listed in Table 6.2. The purification method generated approximately 11-fold purification for

AKR with a 38% yield, and 9-fold purification for SDR with a 31% yield. The method of heterologous protein cloning and purification of polyhistidine tagged protein has been used previously to characterize furfural reductase enzymes such as FucO from E. coli

(Wang et al., 2011) and Ari1 from S. cerevisiae (Liu and Moon, 2009). Other purification steps and methods such as precipitation, ultracentrifugation, size exclusion, ion exchange, and more often combination of several different purification steps have also been used to purify reductase enzymes (Gutiérrez et al., 2006; Li et al., 2011).

6.4.3 Effect of temperature and pH on the activity of AKR and SDR

Using furfural as the substrate and NADPH as the cofactor, AKR and SDR showed catalytic activities at broad ranges of temperature and pH. The effect of temperature on the activity of AKR was determined at temperatures ranging from 20 °C to 65 °C at pH 7.0. AKR showed maximum activity between 40 °C and 50 °C, and the furfural reduction activity was almost doubled when reaction temperature was increased from 20 °C to 40 °C. The activity of AKR started to decrease beyond 50 °C but retained about 30% of the maximum activity at 65 °C (Figure 6.6a). Similarly, SDR showed maximum activity between 35 °C and 50 °C. About 25% of the maximum activity was measured at 20 °C and less than 10% was measured when the assay temperature exceeded 60 °C (Figure 6.6b). The effect of pH on the activity of AKR and SDR were

158

determined between pH 4.0 and pH 8.5. Figure 6.7 shows that both AKR and SDR had their optimum activity at pH 7.0. While AKR retained more than 50% of its activity at pH

5.5 and pH 8.5, SDR had more than 60% of its activity between pH 5.0 and pH 8.5.

Substantial activities of AKR and SDR at broad ranges of temperature and pH is desirable given the biphasic physiology of C. beijerinckii 8052, in which the microorganism produces acid (thus lowring pH to below 5.5) during exponential growth phase and consequently, but then assimilates the acid (thus increasing the pH) during the stationary/solventogenic growth phase to produce acetone butanol ethanol (ABE) (Zhang et al., 2012).

6.4.4 Kinetic properties of AKR and SDR

To determine the kinetic constants of AKR and SDR, the initial furfural reduction rate was measured at furfural concentrations ranging from 2 mM to 50 mM at 40 °C and pH 7.0, in the presence of NADPH since both enzymes showed strong furfural reduction activities coupled with cofactors NADPH but not significant activities with NADH (data not shown). To calculate Km and Vmax values, the curve of reaction velocity vs. furfural concentration was plotted and fit to the Michaelis-Menten equation with non-linear regression in KaleidaGraph. The Michaelis-Menten equation of AKR was V =

254.2*[furfural]/(32.4+[furfural]) (R = 0.995); and that of SDR was V =

22.6*[furfural]/(26.4+[furfural]) (R = 0.986). Based on these equations, kinetic parameters of purified AKR and SDR Km, Vmax, Kcat (ratio of Vmax to the total amount of enzyme) and the ratio of Kcat to Km were calculated and presented in Table 6.3. Notably,

SDR has lower Km (26.4 mM) value than AKR (32.4 mM), indicating a better affinity of

159

SDR to furfural than AKR. Km values of SDR and AKR obtained from this study are higher than the previously reported NADPH-dependent aldehyde reductase (Ari1p) from

S. cerevisiae (12.8 mM) (Jordan et al., 2011), NADPH-dependent aldehyde reductase

YqhD of E. coli (9.0 mM) (Miller et al., 2009), and NADH-dependent dependent propanediol oxidoreductase (FucO) of E. coli (0.4 mM) (Wang et al., 2011). When the reaction velocity of AKR and SDR were compared at the maximum reaction conditions

(Table 6.3), both AKR and SDR have higher Vmax than that of FucO using furfural as the substrate (1.9 µmol/min/mg protein) (Wang et al., 2011), and the turnover number (Kcat) for AKR and SDR showed higher speed of catalysis than that of Ari1P (Jordan et al.,

2011) and YqhD (Jarboe, 2011). The ratio of Kcat to Km for AKR is greater than that of

SDR, which implies that AKR is catalytically more efficient than SDR with regard to furfural reduction to furfuryl alcohol (Table 6.3).

6.4.5 Substrate specificity of AKR and SDR

As shown in Table 6.4, AKR and SDR showed activities with many aldehydes including furfural, HMF, benzaldehyde, and butyraldehyde. To compare activities of

AKR and SDR on different substrate, enzyme activities on other compounds were presented related to the activity towards furfural. Compared with furfural, HMF had about 80% and 50% affinity for AKR and SDR, respectively. This result agrees with a previous report (Liu et al., 2008) that oxidoreductases having reduction activity on furfural coupled with either NADH or NADPH usually demonstrate activity on HMF, probably due to their structural similarities as furan aldehydes. While AKR showed a relative activity of 35% when benzaldehyde was used as substrate, SDR showed more

160

than 16-fold relative activity with benzaldehyde as the substrate (Table 6.4). In a previous report, it was shown that purified furfural reductase from an E. coli strain had 1.2 fold increase in activity when benzaldehyde was used as the substrate (Gutiérrez et al., 2006).

Since benzaldehyde derivatives have been reported to be present in the lignocellulosic biomass hydrolysate (Larsson et al., 2000), the activity of SDR on benzaldehyde indicates a potential application in the detoxification of phenolic compounds or other aldehydes during fermentation of lignocellulosic biomass hydrolysates to ABE by C. beijerinckii. Given that butyraldehyde is the precursor of butanol and the last step

(butyraldehyde reduced to butanol by butanol dehydrogenase) in the butanol production pathway in solventogenic Clostridium species, (Han et al., 2011), activities of AKR and

SDR were evaluated using butyraldehyde as substrate. Notably, AKR and SDR showed relative activities of approximately 25% and 32%, respectively with butyraldehyde substrate (Table 6.4). Therefore the overexpression of AKR and SDR may not only improve cellular ability on detoxification of aldehyde inhibitors, but also facilitate the production of butanol from butyraldehyde, leading to an efficient conversion of biomass to biofuel.

6.5 Conclusion

Enzymes encoded by Cbei_3974 and Cbei_3904 belonging to AKR and SDR families, respectively, were successfully cloned and expressed in E. coli. Both purified enzymes demonstrated furfural reduction activities using NADPH as the cofactor, suggesting that multiple NADPH-dependent oxidoreductases may be responsible for the biotransformation of furfural to furfuryl alcohol by C. beijerinckii 8052. While both

161

AKR and SDR have a broad substrate specificity, AKR has more catalytic efficiency for furfural than SDR, but SDR has a higher activity with benzaldehyde as substrate, implying various applications of these enzymes on detoxification of lignocellulose- derived inhibitors during ABE fermentation by C. beijerinckii 8052.

162

References

Agrawal, M., Chen, R.R., 2011. Discovery and characterization of a xylose reductase

from Zymomonas mobilis ZM4, Biotechnol. Lett. 33, 2127-2133.

Allen, S.A., Clark, W., McCaffery, J.M., Cai, Z., Lanctot, A., Slininger, P.J., Liu, Z.L.,

Gorsich, S.W., 2010. Furfural induces reactive oxygen species accumulation and

cellular damage in Saccharomyces cerevisiae. Biotechnol. Biofuels. 3, 2.

Cornillot, E., Nair, R.V., Papoutsakis, E.T., Soucaille, P., 1997. The genes for butanol

and acetone formation in Clostridium acetobutylicum ATCC 824 reside on a large

plasmid whose loss leads to degeneration of the strain. J. Bacteriol. 179, 5442-5447.

Ezeji, T.C., Blaschek, H.P., 2008. Fermentation of dried distillers' grains and solubles

(DDGS) hydrolysates to solvents and value-added products by solventogenic

clostridia. Bioresour. Technol. 99, 5232-5242.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Gustafsson, C., Govindarajan, S., Minshull, J., 2004. Codon bias and heterologous

protein expression. Trends Biotechnol. 22, 346-353.

Gutiérrez, T., Ingram, L.O., Preston, J.F., 2006. Purification and characterization of a

furfural reductase (FFR) from Escherichia coli strain LYO1—An enzyme important 163

in the detoxification of furfural during ethanol production. J. Biotechnol. 121, 154-

164.

Han, B., Gopalan, V., Ezeji, T.C., 2011. Acetone production in solventogenic

Clostridium species: new insights from non-enzymatic decarboxylation of

acetoacetate. Appl. Microbiol. Biotechnol. 91, 565-576.

Hannig, G., Makrides, S.C., 1998. Strategies for optimizing heterologous protein

expression in Escherichia coli. Trends Biotechnol. 16, 54-60.

Jarboe, L.R., 2011. YqhD: a broad-substrate range aldehyde reductase with various

applications in production of biorenewable fuels and chemicals. Appl. Microbiol.

Biotechnol. 89, 249-257.

Jordan, D.B., Braker, J.D., Bowman, M.J., Vermillion, K.E., Moon, J., Liu, Z.L., 2011.

Kinetic mechanism of an aldehyde reductase of Saccharomyces cerevisiae that

relieves toxicity of furfural and 5-hydroxymethylfurfural. BBA Proteins and

Proteomics. 1814, 1686-1694.

Koenig, K., Andreesen, J., 1990. Xanthine dehydrogenase and 2-furoyl-coenzyme A

dehydrogenase from Pseudomonas putida Fu1: two molybdenum-containing

dehydrogenases of novel structural composition. J. Bacteriol. 172, 5999-6009.

Larsson, S., Quintana-Sainz, A., Reimann, A., Nilvebrant, N., Jönsson, L.J., 2000.

Influence of lignocellulose-derived aromatic compounds on oxygen-limited growth

164

and ethanolic fermentation by Saccharomyces cerevisiae. Appl. Biochem. Biotechnol.

84, 617-632.

Li, Q., Metthew Lam, L., Xun, L., 2011. Cupriavidus necator JMP134 rapidly reduces

furfural with a Zn-dependent alcohol dehydrogenase. Biodegradation. 22, 1215-1225.

Liu, Z.L., Blaschek, H.P., 2010. Biomass conversion inhibitors and in situ detoxfication,

in: Biomass to Biofuels: Strategies for Global Industries. Wiley, pp. 233-259.

Liu, Z.L., Moon, J., 2009. A novel NADPH-dependent aldehyde reductase gene from

Saccharomyces cerevisiae NRRL Y-12632 involved in the detoxification of aldehyde

inhibitors derived from lignocellulosic biomass conversion. Gene. 446, 1-10.

Liu, Z.L., Moon, J., Andersh, B.J., Slininger, P.J., Weber, S., 2008. Multiple gene-

mediated NAD (P) H-dependent aldehyde reduction is a mechanism of in situ

detoxification of furfural and 5-hydroxymethylfurfural by Saccharomyces cerevisiae.

Appl. Microbiol. Biotechnol. 81, 743-753.

Liu, Z.L., 2011. Molecular mechanisms of yeast tolerance and in situ detoxification of

lignocellulose hydrolysates. Appl. Microbiol. Biotechnol. 90, 809-825.

Miller, E.N., Jarboe, L., Yomano, L., York, S., Shanmugam, K., Ingram, L., 2009.

Silencing of NADPH-dependent oxidoreductase genes (yqhD and dkgA) in furfural-

resistant ethanologenic Escherichia coli. Appl. Environ. Microbiol. 75, 4315-4323.

165

Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and

tolerance mechanisms in Escherichia coli. Biotechnol. Biofuels. 2, 26.

Parajó, J.C., Dominguez, H., Domínguez, J.M., 1998. Biotechnological production of

xylitol. Part 3: Operation in culture media made from lignocellulose hydrolysates.

Bioresour. Technol. 66, 25-40.

Qureshi, N., Saha, B.C., Hector, R.E., Dien, B., Hughes, S., Liu, S., Iten, L., Bowman,

M.J., Sarath, G., Cotta, M.A., 2010. Production of butanol (a biofuel) from

agricultural residues: Part II–Use of corn stover and switchgrass hydrolysates.

Biomass Bioenergy. 34, 566-571.

Saha, B.C., Iten, L.B., Cotta, M.A., Wu, Y.V., 2005. Dilute acid pretreatment, enzymatic

saccharification and fermentation of wheat straw to ethanol. Process Biochem. 40,

3693-3700.

Sheehan, J., Aden, A., Paustian, K., Killian, K., Brenner, J., Walsh, M., Nelson, R., 2003.

Energy and environmental aspects of using corn stover for fuel ethanol. J. Ind. Ecol.

7, 117-146.

Sørensen, A., Teller, P.J., Hilstrøm, T., Ahring, B.K., 2008. Hydrolysis of Miscanthus for

bioethanol production using dilute acid presoaking combined with wet explosion pre-

treatment and enzymatic treatment. Bioresour. Technol. 99, 6602-6607.

166

Stöcker, M., 2008. Biofuels and biomass‐to‐liquid fuels in the biorefinery: Catalytic

conversion of lignocellulosic biomass using porous materials. Angew. Chem. Int. Ed.

47, 9200-9211.

Terpe, K., 2006. Overview of bacterial expression systems for heterologous protein

production: from molecular and biochemical fundamentals to commercial systems,

Appl. Microbiol. Biotechnol. 72, 211-222.

Wang, X., Miller, E., Yomano, L., Zhang, X., Shanmugam, K., Ingram, L., 2011.

Overexpression of NADH-dependent oxidoreductase fucO increases furfural

tolerance in Escherichia coli strains engineered for the production of ethanol and

lactate. Appl. Environ. Microbiol. 77, 5132-5140.

Zhang, Y., Han, B., Ezeji, T.C., 2012. Biotransformation of furfural and 5-

hydroxymethyl furfural (HMF) by Clostridium acetobutylicum ATCC 824 during

butanol fermentation. N. Biotechnol. 29, 345-351.

167

Table 6.1 PCR primers used to amplify AKR and SDR using C. beijerinckii 8052 genomic DNA as template. Restriction sites are underlined.

Gene Primer Sequence

5′- GCG CAT ATG ATG AAA TAT CAA GCA TCA AAA 3974_For Cbei_3974 AAT AGA -3′

AKR 5′- CGCT ATC GAT CTA GTT AGC ACT TAT TTC ATC 3974_Rev AAT CAT C -3′

5′- CCA GGA TCC GAT GAG AAA CTT AGA AGG TAA 3904_For Cbei_3904 AGT TGC -3′

SDR 5′- TCG CAT CGA TTT AAA CAT ATC CAC CAT TAA TTC 3904_Rev GA -3′

168

Table 6.2 Purification of Cbei_3974 (AKR) and Cbei_3904 (SDR) from Rosetta- gami™ B(DE3)pLysS. AKR and SDR expressed in Rosetta-gami™ B(DE3)pLysS were purified by immobilized metal affinity chromatography (IMAC) in gravity-flow columns His GraviTrap column (GE Healthcare, Uppsala, Sweden). Crude proteins were the soluble fractions after cell disruption, and purified proteins were obtained by affinity chromatography. Specific activity was determined as micromoles of NADPH oxidized per minute per milligram of enzyme. Total activity was calculated as specific activity multiplied by total amount of protein. Protein yield was estimated as the proportion of total activity of the purified protein to that of the crude protein. And purification fold was the ratio of specific activity of the purified protein to that of the crude one.

Specific Protein Protein Total Total activity Yield Purification Protein conc. volume protein activity (µmol/min (%) fold (mg/ml) (ml) (mg) (μmol/min) /mg ) Crude 8.76 762.0 6.2 54.3 41376.5 100.0 1.0 AKR Purified 1.32 8028.5 1.5 2.0 15851.3 38.3 10.5 AKR Crude 8.61 46.9 6.3 54.3 2545.3 100.0 1.0 SDR Purified 1.33 401.4 1.5 2.0 798.5 31.4 8.6 SDR

169

Table 6.3 Kinetic parameters of purified AKR and SDR. AKR and SDR were expressed and purified from Rosetta-gami™ B(DE3)pLysS. Reactions were conducted at 40 °C in a mixture consisting of 25 mM potassium phosphate buffer (pH 7.0), 40 mM furfural, 0.4 mM NADPH, and 20 μl pure AKR or SDR in total volume of 0.5 ml. Kinetic constants Km and Vmax values were determined according to Michaelis-Menten equation using KaleidaGraph (Synergy Software, Reading PA). Kcat value was calculated as the ratio of Vmax to total amount of enzyme.

K V K K /K Enzyme m max cat cat m (mM) (mM/s) (s-1) (s-1 * mM-1)

AKR 32.4 254.2 1.4x105 4.2x103

SDR 26.4 22.6 9.3x103 3.5x102

170

Table 6.4 Substrate specificity of AKR and SDR. AKR and SDR were expressed and purified from Rosetta-gami™ B(DE3)pLysS. Reactions were conducted at 40 °C in a mixture consisting of 25 mM potassium phosphate buffer (pH 7.0), 10 mM of the substrate, 0.4 mM NADPH, and 20 μl pure AKR or SDR in total volume of 0.5 ml. Specific activity was determined as micromoles of NADPH oxidized per minute per milligram of protein. Relative activities on different substrate are presented as the proportions of the activity on furfural.

AKR specific Relative SDR specific Relative SDR Substrate activity AKR activity (µmol activity % (µmol/min/mg) activity % /min/mg) Furfural 795.7 ± 31.2 100.00 112.8 ± 3.8 100.0

HMF 635.2 ± 24.6 79.83 52.4 ± 19.2 46.4

Benzaldehyde 277.3 ± 121.6 34.85 1820.9 ± 256.4 1613.7

Butyraldehyde 198.9 ± 34.2 25.00 36.1 ± 4.9 32.0

171

(a)

(b)

Figure 6.1 Schematic representation of recombinant plasmids pET-15b_3974 (a) and pET-15b_3904 (b). The recombinant plasmid contains origins (ori), ampicillin resistance (Ap), histidine tag (His tag), protease thrombin, restriction sites (NdeI and ClaI in (a) and BamHI and ClaI in (b)), the lac repressor (lacI), as well as aldo/keto reductase and short-chain dehydrogenase/reductase for (a) and (b), respectively.

172

Figure 6.2 SDS-PAGE analysis of AKR and SDR expressions induced by different concentrations of IPTG. Samples in lanes: 1. Pierce prestained protein molecular weight marker 2. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 without IPTG induction 3. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 induced by 100 µM of IPTG 4. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 induced by 400 µM of IPTG 5. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 induced by 1000 µM of IPTG 6. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3904 without IPTG induction 7. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3904 induced by 100 µM of IPTG 8. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3904 induced by 400 µM of IPTG 9. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3904 induced by 1000 µM of IPTG 10. Cell lysate of Rosetta-gami™ B(DE3)pLysS

173

(a)

(b)

Figure 6.3 SDS-PAGE analysis of induction and purification of AKR (a) and SDR (b) from Rosetta-gami™ B(DE3)pLysS. Samples in lanes: 1. Pierce prestained protein molecular weight marker 2. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b without IPTG induction 3. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b with IPTG induction 4. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 (a) and pET-15b_3904 (b) without IPTG induction 5. Cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET-15b_3974 (a) and pET-15b_3904 (b) with IPTG induction 6. Induced cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET- 15b_3974 (a) and pET-15b_3904 (b) before loaded into column purification 7. Induced cell lysate of Rosetta-gami™ B(DE3)pLysS with plasmids pET- 15b_3974 (a) and pET-15b_3904 (b) flow through 8. Purified AKR(a) encoded by Cbei_3974 and SDR (b) encoded by Cbei_3904

174

Figure 6.4 Codon usage frequency of Cbei_3974 (a) and Cbei_3904 (b) analyzed by E. coli Codon Usage Analysis 2.0

175

(a) 100 50 Codon Freq.

A A T C G T A A A T A G A A T A A T G G A G T A C C A A T T G C T C A T G A A G G T A A A A G T T T Codon T A A A C C A A G A A A T A A G A G G A G G T A T C T T C T G T G A A T G G A C A A A A T A A T G T G A T A A A A T A T T G G G T T A T G A T T A G T T G A C T T T G T T T A C T A T T T C G A A A T T

Amino Acid M K Y Q A S K N R Y N E M K Y S K C G E S G L K L P M I S F G L W H N F G S N A D Y N N M K E L C F Position 10 20 30 40 50

100 50 Codon Freq. 0

A G T G A G A A C T G T G A A T G C G C G A G G G A T G C A T A G G T G A T C G G T T A A A A G G T Codon C C T A A G T C A T A T C A A A G C T C G G C A A A T G G T T G A A T C C A G A A T T T G C A C G A T T T T T A A C T T T G A T C T G A T G A T A A A T T A T T A A C T A G A T T T A A A T T A A A T T

Amino Acid T A F D N G I T H F D L A N N Y G P V P G S A E E N F G R I L R D D L A T Y R D E L L I S T K A G Y Position 60 70 80 90 100

100 50 Codon Freq.

A A T G G C T G G T G A C A T A T G A T G C A T A C A G T G T G G A T T C C C A G C G A C T G G T A Codon A T G A G C A G A T G G G A A T T C G T A A G T A G T G T A A T A T T A A A G T A C A C C T A A C T G G G A A T T A C T A C T A T A A G T G C G C A G T G A G A T A T A C T C T A G T C T T A A A A T G

Amino Acid K M W E G P Y G D F G S R K Y I L A S L D Q S L K R M G L E Y V D I F Y H H R M D P D T P L E E S M Position 110 120 130 140 150

(Continued)

176

Figure 6.4: Continued.

100 50 Codon Freq.

A G C G A G G A A G A G C T G G A T A T A G G A A G A G G G A T A G T A T C T G A A C A A T T A T G Codon T C T A C C T A G G A C T A C G T C A A A G A C T A A C C C T T A A T A G C T T T A A A G A C T T A G T T T A T A A T A A A T C A A T A T T T A G A G A G A T G T A T A A A T A T A T T A T A T T A T T

Amino Acid M A L D T A V K S G K A L Y A G I S N Y N G E T M E K A A A I L N E L K C P F V I N Q N R Y S I F D Position 160 170 180 190 200

100 50

A A A G A A G C A A G G A G A G A G A A G T A C T G C G A T A G A T T A G A C G G A C A A G G G A T Codon G C T A A A G T A G C C A A A G A G T T C T G C T C A G C T C A A A T G G T C A A G G T A T A G G T A T T A T T A T A A A A A A C A A A T A G T T A A C A G A A A T G T A T A T A C T C G A G G T A A T

Amino Acid R T I E N N G L K R A A K E N G K G I I A F S P L A Q G T L T D K Y L S G I P D D S R I K V D G R F Position 210 220 230 240 250

100 50 Codon Freq.

T A C G A T A G A A T G C A A A T A A A G T A A G C A C G C A G C A T G C A G T G G A A G C A G G T Codon T A A A T T C A A A T A A T G G T A A T C T A G G A C T C A T C T G G T T A A C A T C G T T T G C C A G A T A G A G G G A G A T G A A C T T T A T A G A T T T A G G T C G A A A T A A A A T A T T A A A

Amino Acid L K Q D I L T E K K L E Q I R R L N N I A L N R G Q T L A Q M A L S W V L K D S E V T S V L I G A S Position 260 270 280 290 300

(Continued)

177

Figure 6.4: Continued.

100 50 Codon Freq.

A C T C A A G A G G A G C A A G T A G G G T A A A G G A A G A T Codon A C C A T T A A T G T T A A T G T C A A A T T T T A A T G C A A A G T A T T A T G A A T T G T A C G T A G A G G T T A A T T C G

Amino Acid K P S Q I I E N V G I V H K I G F T D E E L M M I D E I S A N Z Position 310 320 330

Colors: = less than 10% of codons for same amino acid; = at least 10%

Fraction of sense codons below threshold (=10.00): 99/331 (29%)

(b)

100 50 Codon Freq.

A A A T G G A G G A A A G G T A G A G A G A G A C T T G C G G A G G G A T T A A G G A G G G G G G G Codon T G A T A G A T C T T C G C C G G T G G C T C G A T C C T G C A T T T A A C A A C T A C A A T T A A G A C A A T A T A A A A T A T A A A A T T A T A A A T A T A T A A T C T T T T T A A A A A A T A A A

Amino Acid M R N L E G K V A I I T G A S R G I G S A I A R Q L S A L G A K V V V N Y S N N A V K A E E V V E E Position 10 20 30 40 50

100 50 Codon Freq. 0

A A A T G G C G G G A A G G G A A A A G G G A T T T G A A A A T G A G G A T A A A G G G A C T A T C Codon T C A C G A A C T C T A C A T G A T A A T A A T T C A C T C A T G G T A T T T A A C G T T T A A T T A T A A T G A C T A T A T T T T T A A C T A A A T A A A A T A T A G A T A A A T T A C A A T T A A T

Amino Acid I T K S G E Q A V A I K A D V S N I K D V E K L F S E T I T K F G R V D I L I N N A G V I L Y K L L Position 60 70 80 90 100

(Continued) 178

Figure 6.4: Continued.

100 50 Codon Freq.

T G G A G G G T G A C T A A A G A G A T T G T C C G A A C A G A A G A A A A T T A T G G G A A T C A Codon C A T C A A A T A A T T A T A T A G C A T C G A A C T A A T A A A G G T T A T C C C T T G G T T C C T C A A A A A T T G T T T A T A G T T T T A C G A T G A T G G T C A A C T T T A C A G T A T G T T T

Amino Acid S D V T E E E F D K L F N I N V K G T Y F A C Q Q A M K H M E N N G R I I N F S T S V V G S M F P T Position 110 120 130 140 150

100 50 Codon Freq. 0

T A G T G G A A G G G G C A A C C T G A G T G C A A A A A A G G G C G C A A A G C T A G G A A G G C Codon A G T A C C C A G C T A A T C G A T C A A T G C A A T C T A C T C C G C T A C A T T A T G A C A A A T C T T A C A A A A T A A C A C A A A G A T T T A A T A T T T A T T T A A T C A T C C T A G A T G G

Amino Acid Y S V Y A A T K G A V E Q I T R Q L A K E F G P K K I T I N A V A P G P I N T E L F N V G K T D E Q Position 160 170 180 190 200

100 50 Codon Freq.

A G G A A C A A T T G C A G G C G G A G A A A G T T G A G A G C T A A G C A C C A A G G T G T Codon T A C T G A T A C T G G T G A C A A T C A C T A T T T G A A C A G T C G A C T G T A G G A T A A A A T A A G T T T A T T T A A T T A A T C T A T G T T T G T A G A A T A T T A T T T A T T A

Amino Acid I E A I R Q M N S F G R I G E P D D I A N T I E F L V S D K A Q W I T G Q T L R I N G G Y V Z Position 210 220 230 240

Colors: = less than 10% of codons for same amino acid; = at least 10%

Fraction of sense codons below threshold (=10.00): 61/246 (24%)

179

(a)

(b)

Figure 6.5 SDS-PAGE analysis of Cbei_3974 (AKR) (a) and Cbei_3904 (SDR) (b) from Rosetta-gami™ B(DE3)pLysS during puriﬁcation process. Samples in lanes: M: Pierce prestained protein molecular weight marker L: Cell lysate supernatant F: Flow through W1: 1st Wash by binding buffer containing 20 mM IMD W2: 2nd Wash by binding buffer containing 50 mM IMD E1-5: Elution by 1 – 5 ml of elution buffer

180

(a)

Specific activity (mmol/min/mg protein) Specific activity (mmol/min/mg 0 20 30 40 50 60 70 Temperature (°C)

(b)

1.0

0.8

0.6

0.4

0.2

Specific activity (mmol/min/mg protein) (mmol/min/mg activity Specific 0.0 20 30 40 50 60 Temperature (°C)

Figure 6.6 Effect of temperature on activity of furfural reduction by AKR (a) and SDR (b). Reactions were conducted in a mixture consisting of 25 mM potassium phosphate buffer (pH 7.0), 40 mM furfural, 0.4 mM NADPH, and 20 μl pure AKR or SDR in total volume of 0.5 ml.

181

(a)

Specific activity (mmol/min/mg protein) (mmol/min/mg activity Specific 0 4 5 6 7 8 9 pH

(b)

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

Specific activity (mmol/min/mg protein) (mmol/min/mg activity Specific 0.2 4 5 6 7 8 9 pH

Figure 6.7 Effect of pH on activity of furfural reduction by AKR (a) and SDR (b). Reactions were conducted at 40 °C in a mixture consisting of 25 mM potassium phosphate buffer (pH 5.0 – 8.5) or 25 mM of soldium acetate (pH 4.0), 40 mM furfural, 0.4 mM NADPH, and 20 μl pure AKR or SDR in total volume of 0.5 ml.

182

Chapter 7: Increased Furfural Tolerance due to Overexpression of NADPH-dependent Aldo/keto Reductase and Short-chain Dehydrogenase/reductase in Clostridium beijerinckii NCIMB 8052

7.1 Abstract

Clostridium beijerinckii NCIMB 8052, one of the best known butanol-producing bacteria, is capable of using lignocellulosic biomass hydrolysates as carbon source for growth and acetone butanol ethanol (ABE) production, albeit poorly due to the presence of microbial inhibitory compounds. To improve tolerance of C. beijerinckii 8052 to microbial inhibitory compounds generated during pretreatment of lignocellulosic biomass, aldo/keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR) were chosen based on transcriptional response of the microorganism to furfural stress, and overexpressed in C. beijerinckii 8052 to generate C. beijerinckii AKR+ and C. beijerinckii SDR+. While C. beijerinckii AKR+ and C. beijerinckii SDR+ grew in P2 medium containing 4 g/L furfural, which was depleted prior to fermentation time of 24 h, control strains with no additional copy of AKR or SDR did not grow in the presence of

4 g/L furfural. Additionally, whereas C. beijerinckii AKR+ produced 6.8 g/L butanol and

3.6 g/L acetone during fermentation in the presence of 4 g/L furfural, C. beijerinckii

SDR+ produced 8.3 g/L butanol and 5.5 g/L acetone under same growth conditions.

183

Moreover, AKR and SDR were purified from C. beijerinckii AKR+ and C. beijerinckii

SDR+ strains, respectively, and NADPH-dependent furfural reductions using pure AKR and SDR enzymes were quantified to validate their roles in the developed strains. This study demonstrated for the first time that overexpression of AKR and SDR in C. beijerinckii is an effective strategy for enhancing furfural tolerance and detoxification, and possible implications of this finding on utilization of lignocellulosic biomass hydrolysates for ABE fermentation are discussed.

7.2 Introduction

Solventogenic Clostridium species are gram-positive, endospore-forming, and obligately anerobic bacteria that have robust high butanol production capabilities by virtue of their ability to utilize sugars (e.g. cellobiose, glucose, xylose, arabinose and mannose) present in lignocellulosic biomass hydrolysates (Ezeji et al., 2007).

Fermentable sugars are naturally trapped inside of lignin structure of lignocellulose, which necessitates pretreatment of lignocellulose-based feedstock to disrupt lignin, depolymerize hemicelluloses, reduce the crystallinity of cellulose, and consequently, make sugar components of lignocellulose accessible to bacteria (Kumar et al., 2009).

However, pretreatment process generate complex mixtures of microbial inhibitory compounds, which has been widely recognized as one of the major limitations of lignocellulosic biomass utilization for biofuel production, and furfural is a common microbial inhibitor produced especially during low-cost dilute acid pretreatment of lignocellulosic biomass (Liu and Blaschek, 2010).

184

Clostridium species have been shown to detoxify microbial inhibitors that belong to aldehyde group, although the underlying mechanism was not clear (Ezeji et al., 2007).

In the current study, effect of lignocellulose-derived inhibitory compounds on growth of and ABE fermentation by C. beijerinckii 8052 was examined (See Chapter 4) followed by transcriptome analysis of C. beijerinckii 8052 grown in the presence of furfural (See chapter 5). The study revealed that genes encoding redox proteins, membrane transporters, and signal transduction system were differentially expressed when C. beijerinckii was grown in the presence of furfural (See Chapter 5). At that juncture, the study hypothesized that two oxidoreductases, aldo/keto reductase (AKR) and short-chain dehydrogenase/reductase (SDR) encoded by Cbei_3974 and Cbei_3904, respectively, were involved in the biotransformation of furfural to furfuryl alcohol, a potential mechanism with which C. beijerinckii 8052 uses to mitigate toxic effects of furfural.

To test this hypothesis, AKR and SDR were overexpressed in C. beijerinckii

8052, and the resulting strains, C. beijerinckii AKR+ and C. beijerinckii SDR+, were grown in P2medium supplemented with high dose of furfural (4 g/L) to explore potential advantages in terms of detoxification and fermentation over C. beijerinckii 8052 wild type. It should be noted that C. beijerinckii 8052 can tolerate and grow in medium containing ≤ 3 g/L furfural (Chapter 4). Since C. beijerinckii 8052 may have other reductases capable of furfural reduction, it is important to purify AKR and SDR from C. beijerinckii AKR+ and C. beijerinckii SDR+, and analyze their activities using furfural as the substrate, to convincingly establish their roles in furfural tolerance and detoxification.

Results obtained from this study confirmed the role of AKR and SDR in furfural

185

reduction to furfuryl alcohol as demonstrated via enzyme activity assays and enhanced tolerance of furfural by C. beijerinckii AKR+ and C. beijerinckii SDR+.

7.3 Materials and Methods

7.3.1 Bacterial strains, plasmids and culture conditions

Clostridium beijerinckii NCIMB 8052 (ATCC 51743) was obtained from the

American Type Culture Collection (Manassas, VA). Escherichia coli DH5α was purchased from New England Biolabs (Ipswich, MA) and was used for the construction and maintenance of recombinant plasmids. Plasmid vector pWUR460 was obtained from

Dr. Wouter Kuit, Wageningen University and Research Centre, Wageningen,

Netherlands (Siemerink et al., 2011). Recombinant plasmids pWUR460_3974 and pWUR460_3904 contains coding regions of Cbei_3974 aldo/keto reductase (AKR) and

Cbei_3904 short-chain dehydrogenase/reductase (SDR), respectively, and were tagged with sequences encoding six successive histidine residues on the 5’ end.

Laboratory stocks of C. beijerinckii 8052 were routinely maintained as spore suspensions in sterile double distilled water at 4 °C. C. beijerinckii 8052 spores (200 μL) were heat-shocked at 75 °C for 10 min followed by cooling on ice. The heat-shocked spores were inoculated into 10 mL anoxic pre-sterilized tryptone–glucose–yeast extract

(TGY) medium (Ezeji and Blaschek, 2008). The culture was incubated in an anaerobic chamber (Coy Laboratory Products Inc., Ann Arbor, Michigan) at 35 °C overnight until optical density (OD) at 600 nm reached 1.0±0.1. About 8% of actively growing culture was transferred into TGY medium and grown for another 4 to 5h until OD600 of 1.0±0.1 was attained, and this preculture was used as the inoculum for ABE fermentation (Ezeji

186

and Blaschek, 2008). E. coli DH5α was grown in Lysogeny broth (LB) medium, and E. coli strains with recombinant plasmids pWUR460_3974 and pWUR460_3904 were grown in LB medium supplemented with ampicillin (50 μg/ml).

7.3.2 Construction of recombinant plasmids pWUR460_3974 and pWUR460_3904

The histidine-tagged AKR and SDR gene fragments were amplified by polymerase chain reaction (PCR) using the purified plasmids pET-15b_3974 or pET-

15b_3904, respectively, as the template (See Chapter 6) , and cloned into Escherichia coli-solventogenic Clostridium shuttle vector pWUR460 (Siemerink et al., 2011). The

PCR primers (synthesized by Eurofins MWG Operon, Huntsville, AL) were designed by

Primer Premier 5 (PREMIER Biosoft Int., Palo Alto, CA) and are listed in Table 7.1. The enzyme used for PCR amplification was Phusion High-Fidelity DNA Polymerase (New

England Biolabs, Ipswich, MA). The PCR reaction conditions consist of the following steps: step 1, 98 °C for 30 sec (initial denaturation), step 2, 98 °C for 10 sec

(denaturation), step 3, 59 °C for 30 sec (annealing), step 4, 72 °C for 30 sec (extension),

35 cycles of step 2, 3, and 4, step 5, 72 °C for 5 min (final extension). For cloning of

AKR, the PCR product obtained was digested with ApaI and XhoI, and recognition sites which were included in the PCR primers are indicated by underlined sequences (Table

7.1). Restriction enzymes used in this study were purchased from New England Biolabs

(Ipswich, MA). Vector pWUR460 was maintained in E. coli DH5α and was isolated by using the GenCatch plus plasmid DNA miniprep kit (Epoch Life Science, Sugar Land,

TX) followed by digestion with ApaI and XhoI (Siemerink et al., 2011). For cloning of

187

SDR, vector pWUR460 was first linearized by XhoI, and the fragment was blunt-ended by treatment with DNA polymerase I large (Klenow) fragment (New England Biolabs,

Ipswich, MA), which was followed by a secondary digestion with ApaI. The digested vectors and PCR products were purified by agarose gel electrophoresis, using the

GenCatch advanced PCR extraction kit (Epoch Life Science, Sugar Land, TX). The ligation of PCR fragment (histidine tagged Cbei_3974 or Cbei_3904) to the vector

(pWUR460) was performed at 4 °C overnight by T4 DNA ligase (New England Biolabs,

Ipswich, MA) in an insert:vector proportion 10:1. The recombinant plasmid was purified using the GenCatch advanced PCR extraction kit (Epoch Life Science, Sugar Land, TX).

Electroporation of the recombinant plasmid into E. coli DH5α was performed in 1mm cuvette using a Bio-Rad Gene Pulser Xcell™ system set at 1.8 KV, 25 μF capacitance, and 200Ω resistance. The duration time was automatically set by the Bio-Rad Gene

Pulser Xcell™ machine (Richmond, CA) which was usually between 4.3 to 4.4 milliseconds. Following electroporation, E. coli DH5α cells were diluted in 1ml SOC medium (2% w/v bacto-tryptone, 0.5% w/v Yeast extract, 10mM NaCl, 2.5mM KCl,

10mM MgCl2, and 20mM glucose) and incubated for at 37°C for 1 h with shaking at 250 rpm. The recovered cells were plated on LB agar supplemented with ampicillin (50

μg/ml). Colonies containing recombinant plasmids were verified by PCR and gel electrophoresis in agarose gel. Recombinant plasmids, pWUR460_3974 or pWUR460_3904, were amplified by growing E. coli transformants in LB medium supplemented with ampicillin followed by purification using the GenCatch plus plasmid

DNA miniprep kit (Epoch Life Science, Sugar Land, TX) and storage at – 80 °C.

188

Sequences of cloned genes were verified via DNA sequencing performed by the Plant–

Microbe Genome Facility at The Ohio State University (Columbus, OH).

7.3.3 Electrotransformation of recombinant plasmids to C. beijerinckii 8052

C. beijerinckii 8052 preculture (6%) was inoculated into P2 medium (glucose 60 g/L and yeast extract 1 g/L) plus P2 buffer, mineral and vitamin stock solutions (Ezeji and Blaschek, 2008), and grown at 35 °C until cell optical density of 0.9–1.1 at 600 nm.

To prepare competent cells, C. beijerinckii 8052 was harvested by centrifugation at 1,200

× g for 10 min at 4 °C (Beckman Model J2 21 M induction drive centrifuge and rotor

JA-17, Beckman, Palo Alto, CA), and washed twice with ice-cold 10% (w/v)

Polyethylene glycol (PEG) 8000. The cell pellet was resuspended with 1/20 volume of

10% PEG 8000 (Zhou and Johnson, 1993). To transform pWUR460_3974 or pWUR460_3904 into C. beijerinckii 8052, 5 μg of each plasmid was gently mixed with freshly prepared C. beijerinckii 8052 competent cells (400 μl) in 0.2 cm electroporation cuvette (Bio-Rad Laboratories, Hercules, CA) followed by incubation of the mixture on ice for 2 min. Electroporation was conducted by using the Bio-Rad Gene Pulser Xcell™ system (Bio-Rad Laboratories, Hercules, CA) set at 2.5 kV, 25 μF capacitance, and infinite resistance, and the pulse delivery time was automatically set by the Bio-Rad Gene

Pulser Xcell™ machine (Zhou and Johnson, 1993). After electroporation, cells were diluted in 9 volumes of TGY medium and incubated anaerobically at 35 °C for 6 h to allow cell recovery and expression of the antibiotic resistant gene. The cell suspension was then mixed with semi-solid TGY agar (0.45% w/v) containing 25 μg/ml erythromycin followed by incubation in an anaerobic chamber for 48–72 h. Colonies

189

were picked and inoculated in TGY medium with 35 μg/ml erythromycin and grown at

35 °C overnight. About 10% of the overnight culture was transferred into TGY medium with 35 μg/ml erythromycin followed by incubation in the anaerobic chamber at 35 °C until active growth (OD600 0.9 – 1.1) was attained. To maintain laboratory stocks of transformants (C. beijerinckii AKR+ and C. beijerinckii SDR+), actively-growing C. beijerinckii AKR+ or C. beijerinckii SDR+ was inoculated (6%, v/v) into P2 medium supplemented with 35 μg/ml erythromycin and grown anaerobically as described above.

At the end of the fermentation (about 5-7 days), C. beijerinckii AKR+ or C. beijerinckii

SDR+ spores was harvested by centrifugation at 6,000 × g for 15 min at 4 °C followed by the decantation of the supernatant. The pellets were washed twice with sterile distilled water followed by resuspension in sterile distilled water and storage at 4 °C.

7.3.4 Purification of C. beijerinckii AKR and SDR and enzyme assay

C. beijerinckii AKR+ and C. beijerinckii SDR+ harboring plasmids pWUR460_3974 and pWUR460_3904, respectively, were grown anaerobically in P2 medium supplemented with 35 μg/ml erythromycin for 24 h. C. beijerinckii AKR+ and C. beijerinckii SDR+ cells were harvested from 50 ml culture by centrifugation at 1,200 × g for 15 min at 4 °C and stored at – 80 °C prior to use. Cells were disrupted by suspension in 3 ml BugBuster protein extraction reagent (Novagen, EMD Chemicals Inc., San Diego,

CA) supplemented with 10 μl of DNase I (New England Biolabs, Ipswich, MA) and 3 mM MgCl2. The cell debris was removed by centrifugation at 12,000 × g for 20 min at 4

°C, and the supernatant was loaded onto the His GraviTrap column (GE Healthcare,

Uppsala, Sweden) pre-equilibrated with 10 ml of binding buffer (20 mM sodium

190

phosphate, 500 mM NaCl, 20 mM imidazole, and 20% glycerol, pH 7.4). The column was then washed by 10 ml of the binding buffer followed by the addition of 1 ml elution buffer (20 mM sodium phosphate, 500 mM NaCl, 500 mM imidazole, and 20% glycerol, pH 7.4). The elution was repeated three more times to collect 4 fractions of 1 ml each.

The purity and molecular weight of eluted fractions were assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentrations of eluates were determined by Bradford reagent (Amresco, Bradford Reagent, Solon, OH) using bovine serum albumin as the standard.

NADPH oxidation to NADP+ at 340 nm using a DU800 spectrophotometer (Beckman

Coulter Inc., Brea, CA). Reaction was blanked when adding furfural to the buffer prior to addition of NADPH and enzyme. Following a 5-min equilibration time after addition of

NADPH to the reaction mixture, the reaction was initiated by the addition of pure 20 μl

AKR or SDR. Each reaction was performed in duplicate.

7.3.5 Effect of furfural on C. beijerinckii and ABE fermentation

Batch fermentations were performed in 150 mL Pyrex screw capped media bottles containing 100 mL of anoxic P2 medium (glucose 60 g/L and yeast extract 1 g/L) plus P2 stock solutions (Zhang et al., 2012). To evaluate the response of C. beijerinckii AKR+ and

C. beijerinckii SDR+ furfural, 6% (v/v) of actively growing preculture of C. beijerinckii

191

AKR+ or C. beijerinckii SDR+ was inoculated into the P2 medium supplemented with 4 g/L furfural. During the course of fermentation, aliquots of 3 mL were taken every 12 h for the analysis of cell growth, ABE production, and the concentration of furfural.

Control fermentations were conducted by inoculating P2 medium with C. beijerinckii

8052 and C. beijerinckii 8052 with plasmid pMTL500E (the vector used to construct pWUR460) (Siemerink et al., 2011) and grown as described above.

7.3.6 Analytical methods

Growth of C. beijerinckii 8052 was estimated by measuring optical density

(OD600) using a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA). The concentrations of ABE and acids were quantified by a gas chromatography system

(Agilent Technologies 7890A, Agilent Technologies Inc., Wilmington, DE), equipped with a flame ionization detector (FID) and 30 m (length) × 320 μm (internal diameter) × 0.50 μm (HP-Innowax film) J × W 19091N-213 capillary column (Han et al.,

2013). Furfural concentrations in the samples were determined by the UV absorbance at

276 nm with a DU800 spectrophotometer (Beckman Coulter Inc., Brea, CA) (Zhang et al., 2012).

7.4 Results and Discussion

7.4.1 Gene cloning of AKR and SDR

Histidine-tagged AKR and SDR were amplified from previously cloned plasmid pET-15b_3974 or pET-15b_3904, respectively (See Chapter 6), and cloned into

Escherichia coli-solventogenic Clostridium shuttle vector pWUR460 because vector pET-15b can propagate only in E. coli but not in solventogenic Clostridium species,

192

while pWUR460 works for both (Siemerink et al., 2011). A restriction site (ApaI) was included in the forward primer followed by a ribosomal binding site (RBS) and a 9 nt space between RBS and the start codon (Figure 7.1). For cloning of AKR, restriction sites of ApaI and XhoI were included in forward and reverse primers, respectively; for SDR, sequence of ApaI was included in the forward primers, but no restriction site was in the reverse primer since sequence of XhoI was found in the coding region of SDR. Therefore the cloning of SDR required Klenow fragment to fill up the sticky end of the vector generated by XhoI and blunt ligation was necessary. Both AKR and SDR were cloned under the control of thl which is a strong constitutive promoter of thiolase (Siemerink et al., 2011), and thus both enzymes were expressed during the entire growth cycle or fermentation. Additionally, the six successive histidine residues tagged on the 5’ end of the cloned sequences (Figure 7.1) facilitated the purification of the target protein by metal ion affinity chromatography. The recombinant plasmids pWUR460_3974 and pWUR460_3904 were transformed into E. coli DH5α for further plasmid amplification and storage at •80 C° prior to use. The recombinant plasmids pWUR460_3974 and pWUR460_3904 were electroporated into C. beijerinckii 8052 for expression and enzyme purification. By using gene-specific primers, amplicons of PCR verified the insertion of the target gene into the vector, and DNA sequencing results confirmed the correct insertion site and orientation.

7.4.2 Purification of AKR and SDR and enzyme assay

Since AKR and SDR were cloned under the control of the thl promoter, both enzymes were expressed constitutively in C. beijerinckii 8052. Purification of AKR and

193

SDR was performed as mentioned in materials and methods. SDS-PAGE analysis shows that histidine-tagged AKR and SDR were successfully purified using the metal ion affinity chromatography (Figure 7.2). Using the cell free extracts and purified AKR and

SDR, strong furfural reduction activities were shown coupled with cofactors NADPH but not significant activities with NADH (data not shown). Specific activity was calculated as number of µM NADPH oxidized per minute per milligram AKR or SDR. Protein yield was estimated as the proportion of total activity of the purified protein to that of the crude protein, and purification fold was the ratio of specific activity of the purified protein to that of the crude one. While the purification of AKR gave a high yield of 81% and a 46- fold increase in specific activity, only 31% yield with a 20-fold purification was obtained when SDR was purified using the same purification method (Table 7.2). The actural purification folds are expected to be much higher the the obtained values. The reason is that C. beijerinckii 8052 crude protein contains more than one enzyme catalyzing furfural reduction which makes the ratio of total activity of the purified protein to that of the crude protein (purification fold) smaller than the actural values. Although many oxidoreductase purifications from Clostridium species were carried out under anaerobic conditions (Meinecke et al., 1989; Ismaiel et al., 1993), purification processes and enzyme activity assays of AKR and SDR were conducted aerobically, indicating that neither of them was markedly inactivated by oxygen. This result is consistent with previously reported reductases that catalyze furfural reduction reactions (Liu et al., 2008;

Liu and Moon, 2009; Miller et al., 2009; Wang et al., 2011).

194

7.4.3 Effect of increasing AKR and SDR expression in C. beijerinckii on furfural tolerance

While no growth was observed when control strains (C. beijerinckii 8052 and C. beijerinckii 8052 harboring vector without AKR or SDR gene insert) were cultivated in

P2 medium supplemented with 4 g/L furfural, C. beijerinckii AKR+ and C. beijerinckii

SDR+ with overexpressed AKR and SDR, respectively, grew without a elongated lag phase, and 4 g/L furfural was depleted before 24 h of growth (Figure 7.3). Following the depletion of furfural in the P2 medium, growth of C. beijerinckii AKR+ and C. beijerinckii SDR+ increased resulting in maximum optical densities at 600 nm of 3 and 5, respectively (Figure 7.3a).

Figure 7.3b and c depict the production of ABE (acetone, ethanol, and butanol) and acids (acetic acid and butyric acid) by C. beijerinckii AKR+ and C. beijerinckii SDR+ strains challenged with 4 g/L of furfural. Concentrations of acetone, butanol and ethanol increased continuously during fermentation, however, ABE production was barely detected before 12 h of growth. Following fermentation time of 12 h, ABE concentration in the fermentation medium started to increase suggesting that furfural had inhibitory effects on ABE fermentation by C. beijerinckii, and fermentation process resumed after the depletion of furfural in the medium (Figure 7.3). This observation is consistent with previous study that showed furfural inhibited solventogenic enzymes, aldehyde dehydrogenase and alcohol dehydrogenase by a competitive model (Modig et al., 2002).

Before 24 h of fermentation, no uptake of acetate was observed by C. beijerinckii AKR+ and C. beijerinckii SDR+ in P2 medium with 4 g/L furfural (Figure 7.3b and c), which

195

may be due to the interruption of the ongoing furfural conversion. Since reduction of furfural by AKR and SDR are NADPH dependent, and butanol production from butyraldehyde also requires NADPH as the electron donor (Dürre et al., 1987), there may be a competition for use of cofactor NADPH between reduction of furfural and production of butanol; and consequently, the cell may be drained of reducing power

NADPH in the presence of furfural. It is plausible that C. beijerinckii AKR+ and C. beijerinckii SDR+ are more proficient than the wild type at tolerating furfural by directing most of the cellular NADPH towards furfural detoxification while delaying growth and butanol production (Figure 7.3b and c). Given that butanol production was temporarily repressed, the upstream intermediates of this pathway (e.g. buryraldehyde and butyryl-

CoA) (Jones and Woods, 1986) may likely accumulate leading to an inhibition of thiolase that catalyzes condensation of two molecules of acetyl-CoA to one molecule of acetoacetyl-CoA (Wiesenborn et al., 1988). Since acetyl-CoA is the precursor of two- carbon acetate and acetoacetyl-CoA is the precursor of three- (acetone) and four-

(butanol) carbon solvent, inhibition of acetyl-CoA condensation may re-direct the carbon flow towards acid production, and potentially, leading to accumulation of acids in the fermentation medium (Figure 7.3b and c) (Girbal and Soucaille, 1998).

A close examination of C. beijerinckii SDR+ shows that it produced comparable level of butanol in the presence of furfural (Figure 7.3c) to that produced by C. beijerinckii 8052 in P2 medium (See Chapter 4). Comparatively, butanol production of the strain C. beijerinckii AKR+ was even lower (Figure 7.3b). Although furfural was reduced to less toxic furfuryl alcohol, 4 g/L furfuryl alcohol in the fermentation medium,

196

together with other alcohol products, may still have an inhibitory effect (See Chapter 4), resulting in the lower butanol production by C. beijerinckii SDR+ and C. beijerinckii

AKR+. In addition, constitutive expression of AKR and SDR may result in a lower ratio of NADPH/NADP+ in C. beijerinckii AKR+ and C. beijerinckii SDR+ than that in wild type even after complete reduction of furfural, since AKR and SDR may use other aldehydes as substrate (Chapter 6). This shortage of reducing power or modification of electron flow has been suggested to alter the solvent ratio (Nakayama et al., 2008), which explains the elevated ratio of acetone to butanol, 0.54 and 0.66 for C. beijerinckii AKR+ and C. beijerinckii SDR+ strains, respectively (Figure 7.3b and c), compared to 0.32 obtained from ABE fermentation by C. beijerinckii 8052 (See Chapter 4). Since the reaction efficiency of AKR was higher than that of SDR (Chapter 6), the imbalance of

NADPH/NADP+ may be worse in C. beijerinckii AKR+ than that in C. beijerinckii SDR+.

Previous study has suggested that decreased ratio of NADPH/NADP+ may be detrimental to cell metabolism (Tian et al., 1998), which can explain the lower cell density and butanol production by C. beijerinckii AKR+ than those by C. beijerinckii SDR+ (Figure

7.3).

Collectively, although in vitro enzyme assay (See Chapter 6) showed that AKR has a high furfural reduction rate, AKR may in fact have a double-edged effect on C. beijerinckii AKR+ because excess AKR in the cell may improve tolerance to furfural stress but may perturb redox balance (NADPH/NADP+ ratio) and metabolic pathways associated with cell proliferation and ABE fermentation. Accordingly, any strategy for improving in situ detoxification of lignocellulose-derived microbial inhibitors should

197

consider striking a balance between inhibitor detoxification and fermentation performance. From this perspective, SDR appears to be a better candidate than AKR for furfural detoxification during ABE fermentation by C. beijerinckii 8052.

7.5 Conclusion

Overexpression of furfural reduction enzymes (AKR and SDR) conferred C. beijerinckii 8052 with enhanced capacity to detoxify furfural during growth and ABE fermentation. In vitro enzyme activity using AKR and SDR purified from C. beijerinckii

8052 verified the NADPH-dependent furfural reduction more convincingly. And alleviation of toxic effects by accelerated furfural reduction allowed understanding of one important mechanisms of inhibitor detoxification. To further enhance butanol production using lignocellulosic biomass as feedstock, investigations on balance of NADPH/NADP+ and re-direction of carbon flux are suggested to minimize the side-effect of inhibitor reduction on butanol production.

198

References

Dürre, P., Kuhn, A., Gottwald, M., Gottschalk, G., 1987. Enzymatic investigations on

butanol dehydrogenase and butyraldehyde dehydrogenase in extracts of Clostridium

acetobutylicum. Appl. Microbiol. Biotechnol. 26, 268-272.

Ezeji, T.C., Blaschek, H.P., 2008. Fermentation of dried distillers' grains and solubles

(DDGS) hydrolysates to solvents and value-added products by solventogenic

clostridia, Bioresour. Technol. 99, 5232-5242.

Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Butanol production from agricultural

residues: Impact of degradation products on Clostridium beijerinckii growth and

butanol fermentation. Biotechnol. Bioeng. 97, 1460-1469.

Girbal, L., Soucaille, P., 1998. Regulation of solvent production in Clostridium

acetobutylicum. Trends Biotechnol. 16, 11-16.

Gutiérrez, T., Ingram, L.O., Preston, J.F., 2006. Purification and characterization of a

furfural reductase (FFR) from Escherichia coli strain LYO1—An enzyme important

in the detoxification of furfural during ethanol production, J. Biotechnol. 121, 154-

164.

Han, B., Ujor, V., Lai, L.B., Gopalan, V., Ezeji, T.C., 2013. Use of Proteomic Analysis

To Elucidate the Role of Calcium in Acetone-Butanol-Ethanol Fermentation by

Clostridium beijerinckii NCIMB 8052. Appl. Environ. Microbiol. 79, 282-293.

199

Ismaiel, A.A., Zhu, C., Colby, G., Chen, J., 1993. Purification and characterization of a

primary-secondary alcohol dehydrogenase from two strains of Clostridium

beijerinckii. J. Bacteriol. 175, 5097-5105.