Engineering Zymomonas Mobilis for the Production of Biofuels and Other Value-Added Products

ENGINEERING ZYMOMONAS MOBILIS FOR THE PRODUCTION OF BIOFUELS AND OTHER VALUE-ADDED PRODUCTS

KORI LYNN DUNN

DISSERTATION

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Chemical Engineering in the Graduate College of the University of Illinois at Urbana-Champaign, 2015

Urbana, Illinois

Doctoral Committee:

Associate Professor Christopher V. Rao, Chair Associate Professor Yong-Su Jin Associate Professor Hyunjoon Kong Associate Professor Mary Kraft

Abstract

Zymomonas mobilis is a promising organism for lignocellulosic biofuel production as it can efficiently produce ethanol from simple sugars using unique metabolic pathways. Z. mobilis displays what is known as the “uncoupled growth” phenomenon, meaning cells will rapidly convert sugars to ethanol regardless of their energy requirements for growth. This makes Z. mobilis attractive not only for ethanol production, but for alternative product synthesis as well.

One limitation of Z. mobilis for cellulosic ethanol production is that this organism cannot natively ferment the pentose sugars, like xylose and arabinose, which are present in lignocellulosic hydrolysates. While it has been engineered to do so, the fermentation rates of these sugars are still extremely low. In this work, we have investigated sugar transport as a possible bottleneck in the fermentation of xylose by Z. mobilis. We showed that transport limits xylose fermentations in this organism, but only when the starting sugar concentration is high. To discern additional bottlenecks in pentose fermentations by

Z. mobilis, we then used adaptation and high-throughput sequencing to pinpoint genetic mutations responsible for improved growth phenotypes on these sugars. We found that the transport of both xylose and arabinose through the native sugar transporter, Glf, limits pentose fermentations in Z. mobilis, thereby confirming our previous results. We also found that mutations in the AddB protein increase plasmid stability and can reduce cellular aggregation in these strains. Consistent with previous research, we found that reduced xylitol production improves xylose fermentations in Z. mobilis. We also found that increased transketolase activity and reduced glyceraldehyde-3-phosphate dehydrogenase activity improve arabinose fermentations in Z. mobilis.

In order for Z. mobilis to prosper as an industrial host for alternative product synthesis, the genetic techniques utilized in this organism must be improved. Toward this goal, we have adapted the λ Red recombinase system for use in Z. mobilis. We have shown that this system increases the frequency of double crossover events for the purpose of constructing gene knockouts in the strain. We have also constructed an expression system for the type II CRISPR/CRISPR-associated (Cas) bacterial adaptive immunity system in Z. mobilis. We have shown that the Cas9 nuclease can be directed by small

RNAs to target the Z. mobilis genome for the purpose of genome editing, and that this system can also likely be used to facilitate gene knockouts in this organism. Finally, toward improving heterologous protein production in Z. mobilis, we have constructed a constitutive promoter library that leads to a range of gene expression levels in the strain.

Collectively, our results provide the framework for the development of an industrial production process utilizing Z. mobilis as the microbial host.

iii

Acknowledgements

I am forever indebted to the many people who have helped me throughout my years in graduate school. First and foremost, I am especially grateful for my advisor, Dr.

Christopher Rao. Thank you for being patient with me as I figured out Zymomonas, for affording me the independence that I prefer, and for always trusting me to solve problems in my own way. I am leaving your lab today with a wealth of knowledge that I couldn’t have gotten under the direction of anyone else. It has been a privilege working with you.

Thank you also to Dr. Yong-Su Jin, Dr. Mary Kraft, and Dr. Hyunjoon Kong for being on my committee and for providing me with valuable advice on my projects, and the Energy

Biosciences Institute for funding my research.

My days in lab were enjoyable thanks to the members of the Rao lab, past and present. Ahmet Badur, Santosh Koirala, and Shuyan Zhang, I couldn’t have asked for better people with whom to share all 6 years of my PhD. Thank you for always being on my side and for sharing my sense of humor. Thank you for appreciating the real me, and for making me feel like that person was someone special. To the many other members of the Rao lab, thank you for the frequent fun times and the acceptance. To my dearest friends outside of lab, Ritika Mohan and Dawn Eriksen, thank you for making the days less daunting. My years in graduate school will be remembered as some of the best of my life, thanks to all of you.

To my sister, Briana, thank you for always being my biggest fan. Thank you for being by my side through it all. We made it this far together, and we will make it even further together, too. Nothing in my life means more to me than you. I don’t know what I would do without you.

To my counterpart, Justin Bradley, thank you for moving to Illinois for me. Thank you for giving up your life to join mine. Thank you for making the most of the situation, for taking the opportunity to better yourself, and of course for never complaining when I had to work late.

Thank you also to my family for making the drive to Illinois and for understanding that education is important. Thank you to my grandparents for the encouragement and for making me feel loved. Finally, thank you to my brothers, Evan and Nathan, for bringing me so much happiness in my life. I hope you will both read this document and aim to follow in my footsteps.

Table of Contents

Chapter 1: Introduction ...... 1 1.1 Motivation ...... 1 1.2 The physiology of Zymomonas mobilis ...... 2 1.3 Genetically modifying Z. mobilis ...... 9 1.4 Z. mobilis for ethanol production from lignocellulosic biomass ...... 12 1.5 Z. mobilis for alternative product synthesis ...... 17 1.6 Conclusions ...... 21 1.7 Figures ...... 23 Chapter 2: Expression of a xylose-specific transporter improves ethanol production in metabolically engineered Zymomonas mobilis ...... 27 2.1 Introduction ...... 27 2.2 Materials and methods ...... 28 2.3 Results ...... 33 2.4 Discussion ...... 36 2.5 Tables and figures ...... 39 Chapter 3: High-throughput sequencing reveals adaptation-induced mutations in pentose-fermenting strains of Zymomonas mobilis ...... 44 3.1 Introduction ...... 44 3.2 Materials and methods ...... 45 3.3 Results ...... 56 3.4 Discussion ...... 63 3.5 Tables and figures ...... 68 Chapter 4: Genetically modifying Z. mobilis using the λ Red recombinase system and a CRISPR-associated nuclease ...... 76 4.1 Introduction ...... 76 4.2 Materials and methods ...... 80 4.3 Results ...... 88 4.4 Discussion ...... 94 4.5 Tables and figures ...... 100 Chapter 5: Development of a promoter library for use in Z. mobilis...... 111

5.1 Introduction ...... 111 5.2 Materials and methods ...... 112 5.3 Results ...... 116 5.4 Discussion ...... 118 5.5 Tables and figures ...... 120 Chapter 6: Conclusions and recommendations ...... 128 6.1 Contributions ...... 128 6.2 Future directions ...... 128 References ...... 135

vii

Chapter 1: Introduction

1.1 Motivation

Microorganisms have historically been used for the production of a variety of value-added products, including pharmaceuticals, foods, and biofuels (da Silva et al.

2013). The yeast Saccharomyces cerevisiae and the bacterium Escherichia coli are popular hosts for product synthesis because they are well-characterized and are highly amenable to genetic manipulation (Leonelli and Ankeny 2013). However, researchers in recent decades have realized the potential of less common, non-model organisms for answering key biological questions and, consequently, for producing chemicals and fuels

(Armengaud et al. 2014). The study of these organisms has allowed for the analysis of a diversity of mechanistic phenomena that would be lost if focus was placed strictly on the commonly used hosts.

One of the most intriguing non-model organisms is the bacterium Zymomonas mobilis. Z. mobilis is the only bacterium that uses the Entner-Doudoroff pathway anaerobically in combination with the pyruvate decarboxylase and alcohol dehydrogenase enzymes (Swings and De Ley 1977). Combining this with the fact that Z. mobilis displays the “uncoupled growth” phenomenon, which enables it to metabolize sugars rapidly regardless of its requirements for growth, makes Z. mobilis an outstanding ethanol producer (Kalnenieks 2006). This capacity for ethanol production has prompted interest in Z. mobilis for industrial biofuel production since the early 1980s (Baratti and

Bu'lock 1986). Around that time, oil embargoes were threatening the United States’ fuel supplies. In response, the development of alternative energy sources became a priority, and the Department of Energy created the Office of Alcohol Fuels (OAF). The OAF was

1 charged with accelerating the scale-up of ethanol production by funding research projects aimed at producing the fuel cheaply in an environmentally-conscious manner (Wyman

2001). This research brought Z. mobilis into the spotlight as an organism with the potential to bring the United States out of its energy crisis, commencing years of investigation into this organism’s interesting physiology.

1.2 The physiology of Zymomonas mobilis

1.2.1 Central metabolic pathways

Z. mobilis is a Gram-negative, obligately fermentative, motile, rod-shaped bacterium frequently found in tropical plant saps and spoiled cider (Swings and De Ley

1977). It produces ethanol and carbon dioxide in equimolar, near-theoretical amounts from the sugars glucose, fructose, and sucrose using the Entner-Doudoroff (ED) pathway, also known as the 2-keto-3-deoxy-6-phosphogluconate (KDPG) pathway, in combination with the pyruvate decarboxylase (Pdc) and alcohol dehydrogenase (Adh) enzymes

(Dawes et al. 1966; Gibbs and Demoss 1954). In contrast to other bacteria, Z. mobilis uses the ED pathway anaerobically to produce only one mole of ATP per mole of glucose, in place of the Embden-Meyerhof-Parnas (EMP) glycolysis pathway which produces 2 (Conway 1992; Kalnenieks 2006). Z. mobilis is restricted to the ED pathway for the metabolism of sugars because its genome lacks sequences for the 6- phosphofructokinase enzyme of the EMP pathway and the 2-oxoglutarate dehydrogenase complex and malate dehydrogenase enzymes of the tricarboxylic acid (TCA) cycle (Raps and Demoss 1962; Seo et al. 2005). Genes for most pentose phosphate pathway (PPP) enzymes are also not present on the Z. mobilis chromosome (Seo et al. 2005). The TCA cycle and PPP enzymes that are present, along with phosphoenolpyruvate carboxylase

2 and malic enzyme are used for anaplerotic reactions (Bringermeyer and Sahm 1989;

Sprenger 1996). Similar to glucose and fructose, sucrose metabolism also proceeds through the ED pathway in Z. mobilis, with the sugar first being degraded into glucose and fructose monomers by extracellular levansucrases, frequently with the formation of a levan byproduct (Goldman et al. 2008; Gunasekaran et al. 1990; Kannan et al. 1995a;

Kannan et al. 1995b; Kyono et al. 1995; Senthilkumar et al. 2009; Silbir et al. 2014; Song et al. 1993; Song et al. 1994). Typically little more than 2% of carbon source utilized by

Z. mobilis is converted into biomass (Belauich and Senez 1965; McGill and Dawes

1971). This low growth yield is a characteristic of Z. mobilis that makes it ideal for product synthesis.

1.2.2 Uncoupled growth

A distinguishing characteristic of Z. mobilis is that it displays what is known as the “uncoupled growth” phenomenon, meaning cells will consume sugars rapidly regardless of their requirements for growth (Belauich and Senez 1965). The catabolic rates in this organism are up to 5 times faster than those seen in yeast, reaching up to 1.0

µmol glucose mg dry wt.-1 min-1 during anaerobic exponential growth (Kalnenieks 2006).

All ED enzymes are highly expressed, most of them constitutively, and ED proteins comprise up to half of the cells’ total protein (Algar and Scopes 1985; An et al. 1991).

Although the ATP yield of the ED pathway is the lowest of the fermentative pathways, the high catabolic rate in Z. mobilis prevents the organism from suffering from a lack of

ATP (Kalnenieks 2006). In fact, due to its low biomass yield, Z. mobilis has a surfeit of

ATP that must somehow be exhausted. The energy consumption reactions for exhausting this ATP balance well with energy production reactions during log phase growth,

3 regardless of the media used (Lazdunsk.A and Belaich 1972). Lazdunski and coworkers provided evidence for a membrane-associated ATPase that eliminates the excess ATP

(Lazdunsk.A and Belaich 1972). Reyes and Scopes later purified what they believed to be the same ATPase, and found that it was of the proton-pumping F0F1-type and that it likely accounted for 20% of the total ATP turnover in the cells (Reyes and Scopes 1991). The rest of the ATP turnover is likely due to the action of acid and alkaline phosphatases and a periplasmic 5’-nucleotidase (Reyes and Scopes 1991). The bacterium Streptococcus bovis also displays the uncoupled growth phenomenon (Forrest 1967). In this bacterium,

ATP spilling by the ATPase is coupled to a cycling of protons across the membrane

(Cook and Russell 1994). Presumably Z. mobilis has a similar conformation, because its

ATPase does not actually dissipate energy but instead converts it into proton-motive force (PMF) (Kalnenieks 2006). While a large pH gradient has been observed across the

Z. mobilis membrane (Barrow et al. 1984; Kalnenieks et al. 1987; Osman et al. 1987), no accurate determination of the transmembrane electric potential has been made to verify this hypothesis (Kalnenieks et al. 1987; Ruhrmann and Kramer 1992). In addition to the cycling of protons across the membrane, Kalnenieks and coworkers proposed that the

PMF could be dissipated through the export of bicarbonate anions from the cell

(Kalnenieks 2006). This hypothesis is supported by the presence of the gene for carbonic anhydrase, which converts carbon dioxide into bicarbonate anions, on the Z. mobilis chromosome (Seo et al. 2005).

1.2.3 Respiration

Although obligately fermentative, Z. mobilis does possess a constitutive respiratory chain (Belauich and Senez 1965). The respiration rate is higher in Z. mobilis

4 than in several well-studied organisms, including E. coli and S. cerevisiae, but the physiological role of respiration in this organism is still uncertain (Kalnenieks 2006).

Experiments using inhibitors like cyanide, chlorpromazine, and myxothiazol to quantify and differentiate terminal oxidases support a respiratory chain with a branched structure

(Kalnenieks et al. 1998; Kita et al. 1984a; Kita et al. 1984b; Poole 1994). The primary electron donors are NADH and NADPH (Bringer et al. 1984). Glucose and D-lactate are also able to donate electrons to the respiratory chain, but to a much lesser extent—the activity of their dehydrogenases is only about 5% or 10-20% of that of the NADH dehydrogenases, respectively (Strohdeicher et al. 1989; Strohdeicher et al. 1990;

Strohdeicher et al. 1988). NADH donates electrons to either a NADH:ubiquinone oxidoreductase complex or a type II NADH dehydrogenase (ndh) (Hayashi et al. 2012;

Seo et al. 2005). The NADH:ubiquinone oxidoreducatse is closely homologous to the genes of an electron transport complex responsible for nitrogen fixation in the bacterium

Rhodobacter capsulatus (Saez et al. 2001; Schmehl et al. 1993). Evidence of nitrogen fixation in Z. mobilis was provided recently, after years of failed attempts (Kremer et al.

2015; Seo et al. 2005). The ability to oxidize NADPH in the respiratory chain of bacteria is rare, but Z. mobilis appears to do so using its type II NADH dehydrogenase

(Kalnenieks et al. 2008). Glucose and D-lactate donate electrons to membrane-bound glucose dehydrogenase and D-lactate dehydrogenase, respectively (Kalnenieks et al.

1998; Strohdeicher et al. 1989; Strohdeicher et al. 1988). All respiratory dehydrogenases then pass their electrons onto the vitamin-like substance ubiquinone, or coenzyme Q10.

From Q10 the electrons are passed to terminal oxidases that are able to reduce oxygen to water. Researchers are confident that a cytochrome bd terminal oxidase is present in Z.

5 mobilis (Kalnenieks et al. 1998; Seo et al. 2005), but the identity of the oxidase that terminates the other branch of the Z. mobilis respiratory chain is still unclear.

1.2.4 Oxidative phosphorylation

The presence of the confirmed components of the respiratory chain combined with the confirmed presence of the F0F1-type ATPase mentioned above would indicate that Z. mobilis is able to perform oxidative phosphorylation for energy production, such that its aerobic growth yield would be higher than its anaerobic yield (Kalnenieks 2006).

However, this is not the case. Growth yields are typically the same or lower when Z. mobilis is grown aerobically versus when it is grown anaerobically (Belauich and Senez

1965; Toh and Doelle 1997b). While this implies a non-functional respiratory pathway,

Dawes and Large did show that ATP could be produced by starved aerated Z. mobilis from ethanol, presumably through oxidative phosphorylation (Dawes and Large 1970).

Kalnenieks and coworkers went on to provide support for this hypothesis by showing that a proton gradient was present in these cells and that the production of ATP was coupled to NADH oxidation. They also showed that the Z. mobilis ATPase could directly produce

ATP by providing it with an artificial proton gradient (Kalnenieks et al. 1993). In 1997,

Zikmanis and coworkers showed that they could in fact achieve a higher growth yield during exponential growth in the presence of oxygen than they could without oxygen present, but the effect was only seen under precise culture conditions. In addition to the higher specific growth rate, a lower specific glucose consumption rate and a lower specific ethanol production rate were seen, providing evidence for the importance of aerobic ATP production under these conditions (Zikmanis et al. 1997). Toh and Doelle were also able to achieve higher growth yields in the presence of oxygen, although they

6 were not completely convinced that oxidative phosphorylation was occurring (Toh and

Doelle 1997a). Regardless, it is clear that if oxidative phosphorylation is functional in Z. mobilis, the efficiency of the process is extremely low. The original consensus among researchers was that this is due to the accumulation of acetaldehyde in the culture medium, which is more pronounced in aerobic cultures due to cofactor competition

(Tanaka et al. 1990). Acetaldehyde was believed to inhibit Z. mobilis growth when present in extremely low amounts (Wecker and Zall 1987). However, it was later determined that under conditions of environmental shock acetaldehyde can actually be beneficial (Stanley et al. 1997), and that reports of its inhibitory effects were inconsistent

(Kalnenieks et al. 2000; Zikmanis et al. 1999). It is also known that the inhibition of the oxidative phosphorylation process itself by acetaldehyde is negligible (Kalnenieks et al.

1993). Therefore, the cause of inefficient respiration in Z. mobilis is still unclear.

1.2.5 Nutrition and environmental effects

While the tolerance of Z. mobilis to the ethanol precursor acetaldehyde may be low, its tolerance to ethanol is high. Strains have been reported to tolerate more than 100 g/L, higher than the tolerance of the yeast S. cerevisiae (Sprenger 1996; Swings and De

Ley 1977). At ethanol concentrations above this amount, leakage of essential cofactors and coenzymes through the plasma membrane occurs (Osman and Ingram 1985). Z. mobilis is also able to tolerate high sugar concentrations of up to 400 g/L (Swings and De

Ley 1977). When growing on sucrose, sorbitol is produced by the Z. mobilis glucose- fructose oxidoreductase (GFOR) which helps to overcome the osmotic stress of growth in high sugar concentrations. If sucrose is not the carbon source, the addition of sorbitol to the culture medium improves growth at high sugar concentrations and also at high

7 ethanol concentrations (Loos et al. 1994; Sootsuwan et al. 2013). Similarly, high salt concentrations also lead to the formation of sorbitol during growth on sucrose to help overcome osmotic stress (Bekers et al. 2000). However, Z. mobilis is not able to grow in media containing more than 20 g/L sodium chloride (Swings and De Ley 1977). The optimum temperature for the growth of Z. mobilis is 30°C, with most strains also able to tolerate temperatures up to 36°C fairly well (Fieschko and Humphrey 1983; Swings and

De Ley 1977). Temperatures above this range cause the lipid-to-protein ratio in the membrane to decline, leading to the leakage of soluble proteins into the surroundings

(Benschoter and Ingram 1986). Growth occurs in a large range of pH values, from about

3.5 to 7.5 (Swings and De Ley 1977). Baumler and coworkers found that the expression of a proton-buffering peptide in Z. mobilis could improve its survival rate in pH 3 to pH

3.5, but its growth rate at these pH levels was unaffected (Baumler et al. 2006). The need for pantothenate, a precursor of coenzyme A, in Z. mobilis minimal medium has been confirmed (Belauich and Senez 1965; Swings and De Ley 1977). In rich medium, yeast extract is not a suitable carbon source and a fermentable sugar must be added (Swings and De Ley 1977). All amino acids, ammonium salts, and peptides can act as nitrogen sources for Z. mobilis (Belauich and Senez 1965). Most assimilated sulfur comes from sulfate and methionine (Rogers, Lee et al. 1982). The addition of inorganic phosphate and certain metallic ions can also improve the growth of Z. mobilis (Belauich and Senez

1965; Dawes and Large 1970).

1.2.6 Strain classification

The classification of Z. mobilis into two subspecies, Z. mobilis subsp. mobilis and

Z. mobilis subsp. pomaceae, was proposed by De Ley and Swings in 1976 (Deley and

Swings 1976). Nearly all of the strains analyzed by the authors fell into the subspecies mobilis, while the strains that caused cider sickness belonged to subspecies pomaceae.

The type strain for subspecies mobilis was designated Zymomonas mobilis (Lindner)

Kluyver and van Niel, or ATCC 10988 (Swings and De Ley 1977). While strain ATCC

10988 produced ethanol efficiently, it was shown in 1981 that strain ZM4 (CP4) produced ethanol from glucose more quickly than ATCC 10988 and also had a higher tolerance to this ethanol (Skotnicki et al. 1981). Therefore, strain ZM4 was chosen for further evaluation and engineering and has remained the Z. mobilis strain of choice for genetic modification and ethanol production since then.

1.3 Genetically modifying Z. mobilis

1.3.1 Early progress

After the ethanol-producing potential of Z. mobilis was brought to light by the laboratory of P.L. Rogers (Rogers et al. 1979), attempts to modify the strain for industrial use intensified. In 1980, Skotnicki and coworkers showed that plasmids of the IncP and

IncFII incompatibility groups could be transferred into the strain by conjugation from E. coli and Pseudomonas aeruginosa (Skotnicki et al. 1980). Realizing that the substrate range of Z. mobilis would need to be expanded in order for it to prosper in the ethanol industry, Carey and coworkers used a conjugable vector containing the tetracycline resistance gene to express a lactose operon in the strain (Carey et al. 1983). In 1984, it was shown that Z. mobilis could be chemically transformed with a hybrid vector created by fusing an E. coli vector and a Z. mobilis native vector (Browne et al. 1984). Additional

Z. mobilis native vectors have since been characterized and used in shuttle vectors

(Afendra et al. 1999; Arvanitis et al. 2000; Dally et al. 1982; Drainas et al. 1983; Misawa

9 and Nakamura 1989; Reynen et al. 1990; Scordaki and Drainas 1987; Scordaki and

Drainas 1990; Strzelecki et al. 1987). Conway and coworkers were the first to construct a vector designed specifically for protein expression in Z. mobilis (Conway et al. 1987a).

The vector contained a broad-host-range origin of replication, a chloramphenicol acyltransferase gene, and a sequenced Z. mobilis promoter directly upstream of a restriction site that could be used for gene cloning (Conway et al. 1987a). The team also sequenced the promoter of pyruvate decarboxylase, one of the most abundant proteins in

Z. mobilis (Conway et al. 1987c). This strong promoter has been used frequently for heterologous protein expression in the strain since then. Additionally, Conway and coworkers characterized promoter structure in general in Z. mobilis using lacZ as a reporter (Conway et al. 1987b). They found that Z. mobilis promoter sequences share several features with E. coli promoter sequences and contain similar consensus sequences in the -10 region and partial sequence homology in the -35 region. The sequence AGGA was also identified as a possible ribosome binding site in Z. mobilis, appearing 8 to 12 base pairs upstream of the start codon (Conway et al. 1987b). In 1997 Pappas and coworkers were the first to show that transposable elements present on suicide vectors could be used to mutagenize the Z. mobilis chromosome (Pappas et al. 1997). The authors successfully used the mini Mu transposon to construct stable Z. mobilis auxotrophs, thereby demonstrating the feasibility of the method for chromosomal integrations.

1.3.2 Later difficulties with genetic modification

While initial attempts to genetically modify Z. mobilis showed promise, several difficulties were later realized. Although gene expression from the lactose operon of

Carey and coworkers was confirmed in Z. mobilis, the strain was unable to grow on

10 lactose as its sole carbon source (Carey et al. 1983). Later attempts to expand the substrate range of Z. mobilis were also met with difficulty (Byun et al. 1986; Feldmann et al. 1992; Liu et al. 1988; Yanase et al. 1988). While conjugation was a reliable method for introducing plasmid DNA into Z. mobilis, chemical transformation and electroporation were not (Buchholz and Eveleigh 1986; Skotnicki et al. 1982; Sprenger

1993). If plasmid DNA was successfully transformed into the cells, it was often unstable and readily lost, especially if it did not contain an origin of replication from a Z. mobilis native vector (Afendra and Drainas 1987b; Bresticgoachet et al. 1990; Dong et al. 2011;

Pappas et al. 1997; Scordaki and Drainas 1987; Strzelecki et al. 1990). Low plasmid stability in the strain is thought to be due to the presence of both type I and type IV restriction-modification systems, which are able to degrade foreign DNA (Kerr et al.

2011; Typas and Galani 1992). Protocols for chromosomal integrations and gene knockouts have not improved since the suicide vector technique of Pappas and coworkers

(Pappas et al. 1997), and targeted methods that utilize suicide vectors require hundreds or thousands of base pairs of homology to the target region but still yield only a few transformants (Agrawal et al. 2012b; Delgado et al. 2002; Kerr et al. 2011; Senthilkumar et al. 2004). Further, an accepted set of promoters and ribosome binding sites (RBSs) for the precise control of gene expression in Z. mobilis has not been developed. Studies to date have primarily relied on the use of strong and constitutive promoters for the uncontrolled expression of heterologous genes, further complicating gene expression optimization in the strain (Deanda et al. 1996b; Jeon et al. 2005b; Zhang et al. 1995).

Collectively, the resistance of Z. mobilis to genetic modification has led to the emergence of only a few recombinant strains (Kerr et al. 2011). To date, Z. mobilis has not been

11 accepted as an industrial organism, regardless of its unique physiology and rapid ethanol production. The acceptance of Z. mobilis for the industrial production of ethanol and other value-added products will require the improvement and standardization of genetic techniques.

1.4 Z. mobilis for ethanol production from lignocellulosic biomass

1.4.1 The cellulosic ethanol production process

Traditionally, fuel ethanol was produced by the yeast S. cerevisiae from the six- carbon sugars found in food crops such as corn and sugarcane (Wheals et al. 1999). More recently, focus has shifted to producing biofuels from lignocellulosic feedstocks, such as agricultural and municipal wastes (Schubert 2006). Cellulosic feedstocks are cheap, abundant, and do not compete with the nation’s food supply (Lynd et al. 1991). They are made up of three primary components: cellulose, hemicellulose, and lignin (Demirbas

2009). After mechanical degradation and an acid pretreatment, cellulose and hemicellulose can be broken down enzymatically into a mixture primarily composed of the six-carbon sugar glucose and the five-carbon sugars xylose and arabinose, which can then be fermented into ethanol or other biofuels by microorganisms (Carroll and

Somerville 2009; Lawford and Rousseau 2002). The sugar mixture may also contain a variety of growth inhibitors, including acetic acid, formic acid, furfural, 5- hydroxymethylfurfural and levulinic acid (Shuai et al. 2010). In order for the cellulosic ethanol production process to be economical, it is important that the microorganism used have several important characteristics. It must be able to ferment both five and six-carbon sugars rapidly to ethanol, in the process tolerating high concentrations of both the sugars and the ethanol, and it must also have a high tolerance to the inhibitors in the hydrolysate

(Bothast et al. 1999; Dien et al. 2003). Further, the organism should be safe, resistant to contamination, and amenable to genetic manipulation (Nevoigt 2008).

1.4.2 Advantages and disadvantages of Z. mobilis for cellulosic ethanol production

Due to its naturally high ethanol productivity, Z. mobilis is an obvious contender in the search for an industrial cellulosic ethanologen. This productivity is up to 5 times higher than that observed in S. cerevisiae, the traditional host for ethanol production

(Rogers et al. 1979; Zhang et al. 1995). Ethanol yields of up to 97% of theoretical are possible with Z. mobilis, primarily due to its unique homoethanologenic metabolism and the low biomass yields obtained with this strain (Belauich and Senez 1965; McGill and

Dawes 1971). Further, Z. mobilis has an extremely high tolerance to both ethanol and to sugars present in the hydrolysate (Rogers et al. 1979). As an organism frequently used for beverage making, Z. mobilis is generally regarded as safe for industrial use (Swings and

De Ley 1977). It is resistant to contamination by other microorganisms, and there are no known viral invaders (Sahm et al. 2006). Its ability to grow at an acidic pH further renders it robust to contamination (Belauich and Senez 1965; McGill and Dawes 1971).

Finally, Z. mobilis is aerotolerant but does not require oxygen for growth, simplifying process design requirements and reducing production costs (Swings and De Ley 1977).

In spite of these attractive advantages, several characteristics of Z. mobilis have prevented it from gaining acceptance for cellulosic ethanol production. Primarily, Z. mobilis has a very narrow substrate range and is unable to natively digest the five carbon sugars xylose and arabinose that are present in the lignocellulosic hydrolysate (Swings and De Ley 1977). Additionally, Z. mobilis has a low tolerance to some of the inhibitors

13 present in this hydrolysate, and is especially sensitive to acetic acid (Jeon et al. 2002;

Kim et al. 2000c; Ranatunga et al. 1997). Finally, as discussed above, researchers have faced several difficulties genetically modifying Z. mobilis (Kerr et al. 2011). In order for

Z. mobilis to be competitive in the cellulosic ethanol industry, these difficulties must be overcome.

1.4.3 Engineering Z. mobilis for improved lignocellulosic ethanol production

Early attempts to engineer Z. mobilis specifically for lignocellulosic ethanol production focused on expanding the organism’s substrate range to include xylose and arabinose (Rogers et al. 2007a). In 1987, Liu and coworkers expressed xylose isomerase, xylose permease, and xylulokinase from Xanthomonas XA1-1 in Z. mobilis. Although enzyme expression was confirmed, the strain was unable to grow on xylose as its sole carbon source (Liu et al. 1988). Similarly, Feldmann and coworkers expressed xylose isomerase and xylulokinase from E. coli in Z. mobilis, and used this strain to isolate a mutant that produced less xylitol phosphate, a potent growth inhibitor produced when xylose is consumed (Feldmann et al. 1992). They then expressed the E. coli transketolase gene in the mutant strain, and again, although enzyme activities were confirmed, the strain was unable to grow on xylose as its sole carbon source. The authors determined that this was likely due to the absence of detectable transaldolase activity in Z. mobilis, but since no transaldolase gene had yet been cloned the authors did not attempt to express this protein in the recombinant strain (Feldmann et al. 1992). It wasn’t until 1995, when researchers at the National Renewable Energy Laboratory were able to clone and express the putative E. coli transaldolase gene together with the E. coli xylose isomerase, xylulokinase, and transketolase genes in Z. mobilis that the strain was able to grow on the

14 sugar as its sole carbon source (Zhang et al. 1995). Shortly thereafter, the same laboratory was able to engineer Z. mobilis to consume arabinose as its sole carbon source by expressing the arabinose isomerase, ribulokinase, ribulose-5-phosphate-4-epimerase, transaldolase, and transketolase genes from E. coli in the strain (Deanda et al. 1996b).

Unfortunately, however, the rates of xylose and arabinose consumption in these strains were far slower than the rate of glucose consumption, even though the ATP yields are the same for the two sugars on a molar basis (Agrawal et al. 2011b; De Graaf et al.

1999a; Deanda et al. 1996b; Joachimsthal and Rogers 2000b; Lawford and Rousseau

2000a; Zhang et al. 1995). Consequently, later efforts focused on investigating the bottlenecks present in xylose and arabinose metabolism in Z. mobilis. The production of xylitol and xylitol phosphate, which was uncovered by Feldmann and coworkers in 1992, was determined to play an important role in limiting xylose fermentations by the strain

(Akinterinwa and Cirino 2009; Feldmann et al. 1992; Kim et al. 2000a). At least two pathways contribute to xylitol formation in Z. mobilis—one involving the glucose- fructose oxidoreductase of central metabolism, and the other involving an NADPH- dependent xylose reductase (Feldmann et al. 1992; Zhang et al. 2009). Agrawal and Chen later found that a strain of Z. mobilis adapted to growth on xylose as its sole carbon source had a mutation in the NADPH-dependent xylose reductase that greatly decreased its activity (Agrawal et al. 2011b). They subsequently deleted this reductase from the chromosome of their xylose-fermenting strain, and found that xylitol production was decreased with a concomitant increase in the rate of xylose metabolism (Agrawal and

Chen 2011). Aside from xylitol formation, Kim and coworkers found that Z. mobilis grown on xylose was in a less-energized state than Z. mobilis grown on glucose, possibly

15 due to the weak activities of the heterologously-expressed xylose isomerase and xylulokinase enzymes (Kim et al. 2000b). Although the studies are unrelated, the adapted strain of Agrawal and coworkers did have a mutation in the promoter region of xylose isomerase that increased its activity and presumably expression by roughly fivefold, revealing the importance of the proper expression level of this enzyme. However, in another separate study, increased expression of xylulokinase was found to not increase the rate of xylose metabolism (Jeon et al. 2005a). If anything, it decreased it, possibly due to the associated increase in xylitol production, although no increase in intracellular xylitol-5-phosphate was observed in the strain. Less work has been done to determine bottlenecks present in arabinose metabolism in Z. mobilis. Deanda and coworkers were not able to discern why their strain did not perform as well on arabinose as it did on glucose, but the authors did propose that the inefficiency of arabinose transport by the Z. mobilis native transporter, the glucose facilitated diffusion protein or Glf, may be limiting arabinose fermentations in the strain (Deanda et al. 1996b).

Work has also been done to increase the tolerance of Z. mobilis to the inhibitors that are present in lignocellulosic hydrolysates. Ranatunga and coworkers determined that acetic acid had by far the largest detrimental effect on xylose-fermenting Z. mobilis, with cells performing at only 9% of maximal in the presence of acetic acid concentrations typically present in cellulosic hydrolysates (Ranatunga et al. 1997). Far behind were caproic acid and furfural, which caused cells to perform at only 57 and 58% of maximal, respectively. Kim and coworkers used nuclear magnetic resonance studies to determine that acetic acid inhibits Z. mobilis by acidifying the cytoplasm and decreasing the amounts of nucleotide triphosphates and sugar phosphates present in the cell (Kim et al.

2000c). Yang and coworkers later isolated a strain of Z. mobilis with increased acetic acid tolerance using chemical mutagenesis and by gradually increasing the concentration of sodium acetate present in the culture medium. They discovered that their strain had increased expression of the sodium-proton antiporter gene nhaA that likely improved the sodium tolerance of their strain (Yang et al. 2010a). Chen and colleagues also isolated a strain of Z. mobilis with improved acetic acid tolerance using adaptation, and separately isolated strains with improved tolerance to other inhibitors present in lignocellulosic hydrolysates, also using adaptation (Chen et al. 2009). Skerker and coworkers took a different approach and constructed a library of Z. mobilis mutants using transposons.

They then analyzed the library for inhibitor tolerance using chemogenomic profiling and found 44 genes important for hydrolysate tolerance in Z. mobilis. The overexpression of one of these genes, ZMO1875, improved the specific ethanol productivity of the strain

2.4-fold in the presence of miscanthus hydrolysate (Skerker et al. 2013).

While some bottlenecks in xylose and arabinose fermentations by Z. mobilis have been uncovered and its tolerance to important inhibitors has been increased, ethanol production by the strain from lignocellulosic pentose sugars still lags behind ethanol production from hexose sugars. In order for the industrial ethanol production process to be economical, pentose sugar fermentations by Z. mobilis must be improved further.

1.5 Z. mobilis for alternative product synthesis

1.5.1 Advantages and disadvantages of Z. mobilis for alternative product synthesis

Z. mobilis has historically received considerable attention for its high ethanol productivity from glucose. However, researchers often overlook the fact that Z. mobilis

17 can also be used to produce other valuable metabolites under the correct culture conditions (Johns et al. 1991). When grown on sucrose, Z. mobilis produces considerably less ethanol and considerably more byproducts than when grown on glucose, suggesting that glucose availability and concentration affect whether cells channel their carbon source into the ED pathway to produce ethanol or elsewhere to produce byproducts

(Johns et al. 1991). Z. mobilis is a promising biocatalyst for alternative product synthesis because of its low biomass yield and high sugar uptake rate, which allows carbon to quickly become fuels or chemicals instead of biomass (Belauich and Senez 1965; McGill and Dawes 1971). It also produces a wide variety of growth-linked byproducts that hold commercial value, allowing for these products to all be isolated in a single process (Johns et al. 1991). Further, it is a rich source of many enzymes currently used in diagnostic analysis and research (Rogers et al. 2007a). As with ethanol production, Z. mobilis is also a desirable host for alternative product synthesis because it is generally regarded as safe, it is resistant to contamination, and it has relatively simple growth requirements

(Belauich and Senez 1965; McGill and Dawes 1971; Swings and De Ley 1977).

Regardless of the potential of Z. mobilis for the production of value-added products, the organism will need to be further engineered to make any of the processes economical. Cellular metabolism should be shifted away from ethanol production and toward the production of the desired product—this will require the construction of knockout and overexpression mutants and the precise control of heterologous gene expression in Z. mobilis (Rogers et al. 2007a). Therefore, genetic techniques for constructing these recombinant strains clearly need to be improved in order for the potential of Z. mobilis for the production of value-added products to be realized.

1.5.2 Engineering Z. mobilis for alternative product synthesis

Z. mobilis is natively able to produce the byproducts fructose, sorbitol, gluconate, fructooligosaccharides, and levan at concentrations exceeding 1 g/L (Johns et al. 1991).

The strain also produces a variety of byproducts in small amounts, including mannitol, acetaldehyde, glycerol, dihydroxyacetone, and lactic, gluconic, succinic, and acetic acids

(Jobses et al. 1985; Schmidt and Schugerl 1987; Viikari and Gisler 1986; Wecker and

Zall 1987). Z. mobilis has received some attention for the commercial production of fructose from sucrose with the concomitant production of ethanol, both of which are value-added products (Doelle and Greenfield 1985). In order for this to be possible at industrially desirable levels, however, a fructokinase-negative mutant must be used so that the cells cannot consume the fructose that is being produced by sucrose splitting, and instead only convert the glucose into ethanol (Pankova et al. 1988). This process is desirable over the process currently used to produce the food sweetener high-fructose corn syrup because fructose and ethanol are relatively easy to separate from one another

(Johns et al. 1991). Sorbitol, levan, and fructooligosaccharides are also value-added products that can be produced by Z. mobilis during growth on sucrose under very specific culture conditions. These products are commonly used in the food and pharmaceutical industries. A high rate of sucrose hydrolysis favors sorbitol production, while a low rate favors levan production (Doelle and Greenfield 1985; Viikari and Gisler 1986).

Fructooligosaccharide production occurs as an alternative to levan production when salt concentrations are high (Doelle and Doelle 1990; Doelle et al. 1990). Sorbitol production in Z. mobilis has been improved by overexpressing the glucose-fructose oxidoreductase enzyme, responsible for producing sorbitol in a single step from glucose and fructose

(Liu et al. 2010). Levan production in Z. mobilis has been improved by lowering the rate of sucrose hydrolysis using a strain with the sacC gene inactivated (Senthilkumar et al.

2004). Z. mobilis can also be used for the isolation of a variety of industrially-relevant enzymes, including fructokinase, glucose-6-phosphate dehydrogenase, pyruvate kinase, pyruvate decarboxylase and glucose-fructose oxidoreductase (Johns et al. 1991). A mutant form of pyruvate decarboxylase was engineered and used in the production process of an important pharmaceutical (Goetz et al. 2001; Iwan et al. 2001; Pohl 1997;

Rogers et al. 1997), but little else has been done toward improving enzyme yields for commercial use in Z. mobilis.

While Z. mobilis is promising for industrial fructose production, the process is tainted by the concomitant production of large amounts of sorbitol. This is detrimental because the amount of sorbitol allowed in food products is often limited, and fructose and sorbitol are difficult to separate. While growth condition optimization can be used to decrease the amount of sorbitol that is produced, careful genetic engineering of the fructokinase-negative mutants can possibly decrease sorbitol production further, making the process more industrially feasible. Similarly, it is possible that overexpressing the Z. mobilis sucrases could improve sorbitol production in the strain and improve its industrial outlook. In general, all routes of alternative product synthesis in Z. mobilis would benefit from additional engineering of the strains to decrease ethanol production and increase production of the desired products. In working toward this goal, the current consensus in the field is that genetic engineering techniques need to be improved and modernized in order for Z. mobilis to be considered as a viable platform for future biorefineries (He et al. 2014).

1.6 Conclusions

Z. mobilis clearly holds potential for the commercial production of the biofuel ethanol in the present and for the commercial production of other value-added products in the future. In order for these processes to be economical, however, more work needs to be done. The digestion of lignocellulosic pentose sugars by the strain is still slow and lags far behind the digestion of hexose sugars; the full reason for this is unknown. Ideally, all pentose and hexose sugars would be consumed simultaneously. Further, although Z. mobilis is a promising platform for the isolation of other biofuels and chemicals, little work can be done to improve these processes with the outdated and inefficient genetic techniques available to researchers today. Ideally, modern genetic engineering techniques would be applied for use in Z. mobilis.

In this work, we have attempted to overcome these issues in Z. mobilis. In chapter

2, we describe how we used sugar transport engineering to improve the consumption rate of the pentose sugar xylose in engineered Z. mobilis. In chapter 3, we explain our use of adaptive evolution to isolate strains of Z. mobilis capable of more rapid growth on the pentose sugars xylose and arabinose, and explain how we used high-throughput sequencing to determine the chromosomal mutations responsible for the altered phenotypes in order to elucidate additional bottlenecks in pentose sugar metabolism by the strain. Toward genetic technique improvement, in chapter 4 we describe how we have applied the lambda Red system combined with the yeast flippase enzyme for constructing clean chromosomal deletions in Z. mobilis. We also provide proof-of-concept results for chromosomal editing using the CRISPR/Cas9 system in the strain. In chapter 5, we

21 describe the construction of a promoter library for the precise control of heterologous gene expression in Z. mobilis. Finally, in chapter 6, we conclude and discuss future work.

1.7 Figures

Figure 1.1 Image of Z. mobilis cells taken using differential interference contrast (DIC) microscopy (bar = 10 µm).

Figure 1.2 Pathways for glucose, xylose, and arabinose metabolism in engineered Z. mobilis. Xylose and arabinose enter central metabolism through the pentose phosphate pathway (left), while glucose enters the Entner-Doudoroff pathway directly (right). Heterologous enzymes are shown in bold.

Figure 1.3 The lignocellulosic ethanol production process. Cellulosic feedstocks, such as switchgrass and miscanthus, are first broken down into smaller particles. Dilute acid or steam is then used to break these particles into polysaccharides that become primarily the sugars glucose, xylose, and arabinose during the hydrolysis step. These sugars can then be fermented by engineered Z. mobilis into ethanol that is purified using distillation.

Figure 1.4 Pathways for byproduct formation in Z. mobilis.

Chapter 2: Expression of a xylose-specific transporter improves ethanol production in metabolically engineered Zymomonas mobilis1

2.1 Introduction

Z. mobilis produces ethanol from glucose at rates and yields that surpass those of other microorganisms (Rogers et al. 2007b; Yang et al. 2010b). It has also been engineered to consume the cellulosic pentose sugar xylose (Zhang et al. 1995), but it does so very inefficiently (Agrawal et al. 2011a; De Graaf et al. 1999b; Joachimsthal and

Rogers 2000a). The fermentation rates are almost an order of magnitude slower than they are for glucose even though the molar yields of ATP are the same for the two sugars

(Lawford and Rousseau 2000b). These strains are also unable to simultaneously ferment glucose and xylose. Rather, in a manner reminiscent of catabolite repression, they sequentially ferment these two sugars (Gao et al. 2002; Lawford and Rousseau 1999).

This selective utilization is not thought to be due to any known intrinsic regulatory system but rather to inefficiencies in the metabolism of xylose.

As discussed in chapter 1, a number of studies have investigated bottlenecks in xylose metabolism by Z. mobilis. One aspect of xylose metabolism that has not been investigated in detail is transport. Xylose enters the cell through the promiscuous glucose facilitated diffusion protein, Glf (Dimarco and Romano 1985; Parker et al. 1995). While the affinity of Glf for xylose is lower than for glucose (KM=40 mM versus KM=4.1 mM), the Vmax is twice as great (Weisser et al. 1996a). However, these measurements were made in a strain of E. coli where the endogenous xylose transporters were still present,

1Portions of this chapter were reprinted from “Expression of a xylose-specific transporter improves ethanol production by metabolically engineered Zymomonas mobilis” in the journal Applied Microbiology and Biotechnology, August 2014, Volume 98, Issue 15, pp 6897-6905, by Dunn and Rao with kind permission from Springer Science and Business Media. 27 possibly leading to a biased estimate of the Vmax. In addtition, glucose may block the uptake of xylose by the Glf transporter through a competitive inhibitory mechanism.

Such a mechanism would explain why glucose is preferentially utilized by xylose- fermenting strains of Z. mobilis (Deanda et al. 1996a). Therefore, one potential solution to improve xylose fermentation, both as a single carbon source and in mixtures with glucose, would be to introduce pentose-specific transporters into the cell.

In this chapter, we tested how the expression of a xylose-specific transporter in a xylose-fermenting strain of Z. mobilis affects its ability to utilize xylose, both as a single carbon source and in mixtures with glucose. Our results show that the introduction of a xylose-specific transporter increases the rate of xylose utilization at high sugar concentrations. The increase is most pronounced when xylose is the sole carbon source.

We also found that over-expression of the native Glf transporter improved the efficiency of xylose utilization at high concentrations when it was the sole carbon source. These results demonstrate that transport is a bottleneck for xylose utilization by Z. mobilis and that the expression of xylose-specific transporters can help relieve this bottleneck.

2.2 Materials and methods

2.2.1 Bacterial strains, media, and growth conditions

All experiments were performed in Zymomonas mobilis subsp. mobilis ZM4

(ATCC 31821). RM liquid and solid medium (10 g/L yeast extract and 2 g/L potassium phosphate, dibasic) supplemented with glucose or xylose at the designated concentrations were used for all Z. mobilis experiments. All Z. mobilis cultures were grown at 30°C.

Antibiotics were used at the following concentrations: spectinomycin, 100 µg/mL; and chloramphenicol, 100 µg/mL.

2.2.2 Plasmid construction

All plasmids used in this study are listed in Table 2.1. All primers used in this study are listed in Table 2.2. To construct a plasmid containing the xylose metabolism and pentose phosphate pathway genes, the xylose isomerase (xylA), xylulokinase (xylB), transaldolase (talB), and transketolase (tktA) genes were isolated from E. coli MG1655 genomic DNA by PCR synthesis using primers KD002F, KD002R, KD004F, KD004R,

KD005F, and KD005R, respectively. To express the xylose isomerase and xylulokinase genes (xylAB) in Z. mobilis, they were fused from the xylA start codon to the xylB terminator sequence to the strong and constitutive 310 bp Z. mobilis glyceraldehyde-3- phosphate dehydrogenase (gap) promoter, amplified from Z. mobilis ZM4 genomic DNA using primers KD001F and KD001R, by overlap extension PCR (Horton et al. 1989).

Specifically, overlap extension PCR reactions were performed in a 50 µL volume containing 500 ng of the larger gene fragment and an equimolar amount of the second.

No single-stranded DNA primers were added to the splicing reaction. Double-stranded

DNA fragments were annealed and extended using Phusion high-fidelity DNA polymerase (New England Biolabs) during 25 PCR cycles with an annealing temperature of 50°C. The resulting 3264 bp Pgap-xylAB synthetic operon was then inserted into plasmid pJS71 to form plasmid pKLD1. To express the transaldolase and transketolase genes in Z. mobilis, the transaldolase gene from its start to stop codon was first fused to the strong and constitutive 308 bp Z. mobilis enolase (eno) promoter, amplified from Z. mobilis ZM4 genomic DNA using primers KD003F and KD003R, using overlap extension PCR. The constructed 1264 bp Peno-talB fragment was then inserted into plasmid pKLD1 to yield plasmid pKLD2. Finally, the transketolase gene including its

29 native ribosome binding site and native terminator sequences was inserted directly downstream of the talB stop codon in plasmid pKLD2 to create the 3348 bp Peno-talB- tktA synthetic operon and yielding plasmid pKLD3.

To construct a pentose-specific transporter plasmid, the xylE gene was amplified from E. coli MG1655 genomic DNA by PCR synthesis using primers KD007F and

KD007R. The 500 bp strong and constitutive pyruvate decarboxylase (pdc) promoter was also amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers

KD006F and KD006R. For expression in Z. mobilis, the xylE gene was then fused from its start codon and including its native terminator sequence to the pdc promoter using overlap extension PCR. The Ppdc-xylE construct was then inserted into plasmid pEZCm to yield plasmid pKLD4. For overexpression of Glf in Z. mobilis, the pdc promoter and the glf gene were amplified from Z. mobilis ZM4 genomic DNA using primers KD008F,

KD008R, KD009F, and KD009R, respectively. The glf gene was then fused from its start codon and including its native terminator sequence to the pdc promoter using overlap extension PCR. The Ppdc-glf construct was then inserted into plasmid pEZCm to yield plasmid pKLD5.

To construct plasmids for analyzing promoter strengths, the Venus fluorescent protein gene was amplified from its start codon and including its native terminator sequence from plasmid pVenus by PCR synthesis using primers KD011F, KD011R,

KD013F, and KD013R. The 310 bp glf promoter and the 500 bp pdc promoter were also amplified separately from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers

KD010F, KD010R, KD012F, and KD012R, respectively. The Venus gene was then fused separately to the glf and pdc promoters using overlap extension PCR. The resulting Pglf-

Venus and Ppdc-Venus constructs were then inserted into plasmid pEZCm to yield plasmids pKLD6 and pKLD7, respectively.

To construct a plasmid for analyzing transporter localization, the 310 bp glf promoter was amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers KD014F and KD014R. The xylE gene was amplified without its stop codon and including a polypeptide linker sequence from E. coli MG1655 genomic DNA by PCR synthesis using primers KD015F and KD015R. The Venus fluorescent protein gene was amplified from its start codon and including its native terminator sequence from plasmid pVenus by PCR synthesis using primers KD016F and KD016R. The glf promoter, the xylE gene, and the Venus gene were then fused together using overlap extension PCR, yielding the Pglf-xylE-venus translational fusion construct. The fusion was then inserted into plasmid pEZCm to yield plasmid pKLD8. Z. mobilis expressing plasmid pKLD6 was used as a control in localization experiments.

2.2.3 Plasmid transformation into Z. mobilis using electroporation

Z. mobilis cells grown in RM containing 20 g/L glucose to an OD600nm of at least

0.8 were harvested and washed once with water and once with 10% (v/v) glycerol. Cells were then resuspended in 10% glycerol to 1/100 of the original volume. For each transformation, 2.5 µg of TypeOne Restriction Inhibitor (Epicentre) and at least 1 µg of unmethylated plasmid DNA isolated from E. coli strain GM119 or GM2929 were added to 40 µL of competent cells. Cuvettes with 0.1 cm gaps were used in a BioRad XCell gene pulser set at 1.6 kV, 200 Ω, and 25 µF. Following transformation, cells were resuspended in RM and allowed to recover at 30°C for 1 to 3 h. Transformants were isolated after growth for 3 to 7 days on RM plates containing the appropriate antibiotic.

Plasmid presence was confirmed by plasmid isolation and subsequent transformation into

E. coli DH5α. For strains harboring two plasmids, transformations were performed sequentially.

2.2.4 Fermentations

Pre-seed cultures were prepared by inoculating a single colony from an RM agar plate into liquid RM in a 5 mL glass tube. Pre-seed cultures were grown with shaking until stationary phase, when they were used to inoculate seed cultures. Seed cultures were grown with shaking in 15 mL sealed conical tubes filled with RM to 60% volume until late exponential phase. All pre-seed and seed cultures were grown in RM medium containing 20 g/L glucose as carbon source. Fermentation cultures used for analyses were inoculated with seed culture to an OD600nm of 0.02. They were then grown with shaking at

140 rpm in 50 mL sealed conical tubes filled with RM to 90% volume and containing the appropriate antibiotics and the appropriate amounts of glucose and/or xylose as carbon source(s).

2.2.5 Analytical methods

Cell growth was determined by the optical density (OD) at 600 nm. Fermentation samples were passed through 0.22 µm polyethersulfone syringe filters and the resulting supernatant was analyzed for glucose, xylose, xylitol, and ethanol concentrations using high-performance liquid chromatography (HPLC). A Shimadzu HPLC system with a refractive index detector (Shimadzu RID-10A) was used together with an Aminex HPX-

87H carbohydrate analysis column and a cation H micro-guard cartridge (Bio-Rad) kept at a temperature of 65°C. The mobile phase was 5 mM H2SO4, pumped at a flow rate of

0.6 mL/min. Peaks were identified and quantified by retention time comparison to

32 authentic standards (glucose, xylose, xylitol, and ethanol). All data reported are the average and standard deviation of three replicates.

For promoter strength analyses, pre-seed, seed, and fermentation cultures were prepared as described above, except fermentation cultures were filled to 40% volume to allow for the aeration necessary for proper expression of the Venus protein. At late exponential phase, 150 µL was withdrawn from the fermentation culture and transferred to a single well of a black-walled, clear-bottomed Costar 96-well microtiter plate. The fluorescence (Ex/Em: 515/528 nm) and optical density (600 nm) were measured using a

Tecan Safire2 microplate reader. Fluorescence measurements are reported as the relative fluorescence normalized to the optical density of the sample to correct for variation in cell density and given in terms of relative fluorescence units (RFU).

For transporter localization analyses, cells were mounted on an agar pad on a #1.5 coverslip. Total internal reflection fluorescence (TIRF) microscopy images were captured on a Leica AF6000 microscope fitted with a Hamamatsu EM-1K EM-CCD camera, automated stage, adaptive focus control (AFC), and a 488 nm laser. A 3x magnification,

100x/1.47 oil-immersion objective was used, and the evanescent field penetration depth was set to 250 nm.

2.3 Results

2.3.1 Construction of a Z. mobilis strain expressing a xylose-specific transporter

We hypothesized that transport limits the rate of xylose utilization in Z. mobilis.

To test this hypothesis, we expressed XylE, the low-affinity xylose transporter from E. coli (Davis and Henderson 1987), in a xylose-utilizing strain of Z. mobilis (Figure 2.1).

We tested both the native pdc and glf promoters for expressing XylE and found that the pdc promoter yielded significantly better results. As a consequence, we only report the results for this promoter. We separately tested the expression of these two promoters using transcriptional fusions to the fluorescent protein Venus (Nagai et al. 2002) and found that the pdc promoter is significantly stronger than the glf promoter (7582 ± 661

RFU versus 3032 ± 255 RFU). In addition, we confirmed that the XylE protein is able to localize to the Z. mobilis membrane using a translational fusion to Venus (Figure 2.2).

We also tested AraE, the low-affinity arabinose transporter from E. coli, which is known to also transport xylose (Desai and Rao 2010; Novotny and Englesberg 1966;

Shamanna and Sanderson 1979). However, we found that the expression of AraE inhibited growth in Z. mobilis. Specifically, we observed a large lag in the growth (~72 hours) of strains expressing AraE. We also tested a xylose-utilizing strain of Z. mobilis expressing the native Glf transporter using the pdc promoter from a plasmid. These results are included in our analysis. Our control was a xylose-utilizing strain of Z. mobilis containing the empty plasmid used to express the transporter.

The xylose-utilizing strain was made by expressing the xylose isomerase (xylA), xylulokinase (xylB), transaldolase (talB) and transketolase (tktA) genes from E. coli in Z. mobilis. This design is based on the one previously developed by Zhang and coworkers

(Zhang et al. 1995). In order for this strain to consume xylose as its sole carbon source, we needed to first grow the strain in glucose and then slowly adapt it to xylose as reported previously (Agrawal et al. 2011a).

2.3.2 Expression of XylE and Glf improves xylose fermentation in xylose-only media

We first tested our different strains using xylose as the sole carbon source. We examined three different xylose concentrations: 10 g/L, 50 g/L, and 100 g/L (Table 2.3).

At a xylose concentration of 10 g/L, we did not observe any improvement in xylose utilization or ethanol production when XylE or Glf were expressed. At a xylose concentration of 50 g/L, we observed a small increase in the rate of xylose utilization

(~30%) and production of ethanol (~20%). The increases were similar for both XylE and

Glf. At a concentration of 100 g/L (Figure 2.3), we observed significant increases in both the rate of xylose utilization (~50-140%) and production of ethanol (~70-100%) when

XylE or Glf were expressed in the cells. The increases were greater for XylE than Glf.

Collectively, these results demonstrate that transport limits xylose metabolism at high sugar concentrations. These results also demonstrate that the expression of xylose transporters, including the promiscuous Glf transporter, can improve the rate of xylose utilization and ethanol production.

We also measured xylitol as it has previously been shown to inhibit xylose utilization in Z. mobilis. Consistent with increased rates of xylose utilization, we found that cells expressing XylE or Glf produced more xylitol than those that did not. This increase, however, was seen only at the higher xylose concentrations (Table 2.3).

2.3.3 Expression of XylE improves xylose fermentation in glucose-xylose mixed media

We next tested whether the expression of XylE or Glf would improve xylose utilization when glucose is also present. Once again, we tested the performance of these

35 strains at three sugar concentrations (Table 2.4). At a concentration of 10 g/L xylose and

10 g/L glucose, we did not observe any increase in the rate of xylose utilization or ethanol production. In fact, the utilization and productivity rates were less than the wild- type control though the final ethanol titers were the same. At a concentration of 25 g/L and 25 g/L glucose, we again observed no increase in the rate of xylose utilization or ethanol production: rates and final titers were roughly the same for the different strains.

Once again, only at the highest concentrations of sugars (50 g/L xylose and 50 g/L glucose) did we see an effect (Figure 2.4). In this case, expression of XylE increased the xylose utilization rate by approximately 50% and final ethanol titer by 15%. Interestingly, expression of XylE reduced the glucose consumption rate. Even then, however, sugar utilization was not simultaneous as xylose metabolism is still far slower than glucose metabolism. We also note that less xylitol is produced in strains expressing XylE for reasons unknown. Finally, we did not observe any difference in strains expressing Glf relative to the control.

2.4 Discussion

We investigated whether the expression of xylose-specific transporters improves the rate of xylose metabolism in Z. mobilis. We found that XylE expression improves ethanol production in both xylose only and glucose-xylose mixed media, but only when the starting sugar concentrations are high (>50 g/L). Despite these increases, no significant improvements in glucose-xylose co-consumption were observed. In addition, we found that overexpression of the native Glf transporter improves ethanol production in the pentose-fermenting strain, but only in xylose-only media and again only when the

36 starting sugar concentrations are high. Collectively, these results suggest that transport is a bottleneck in xylose metabolism but only at high sugar concentrations.

To potentially explain this result, we note that Glf is expected to saturate at a xylose concentration of approximately 45 g/L (assuming KM=30 mM and the transport rate is 90% maximal). This value is roughly the concentration where transporter expression increases xylose utilization. This would suggest that the expression of XylE simply increases the overall capacity for transport rather than increasing the specific affinity for xylose. Consistent with this conclusion, we also found that over-expression of

Glf yielded similar results as those involving the xylose-specific transporter XylE.

Moreover, with regards to glucose-xylose co-utilization, our data would suggest that glucose does not inhibit xylose uptake – rather, the Glf transporter simply saturates at higher xylose concentrations. The apparent diauxie is likely due to inefficiencies in xylose metabolism and not transport. In other words, the sugars are consumed sequentially because xylose is consumed much more slowly than glucose.

Interestingly, the KM of XylE for xylose is approximately 100 μM (Sumiya et al.

1995), which suggests that the transporter would be saturated at concentrations less than

1 g/L. This would suggest that expression of XylE should increase xylose utilization at low concentrations, something we did not observe. While we cannot fully explain this result, it may be worth noting that XylE is a proton-linked sugar transporter. It utilizes the proton-motive force as the energy source for sugar transport. Glf, on the other hand, is not energy dependent and transport is due solely to concentration differences (Parker et al. 1995). One possibility is that the membrane is not equivalently energized in Z. mobilis, leading to a higher effective KM.

We conclude by first noting that similar strategies have also been employed in xylose-fermenting strains of yeast (Du et al. 2010b; Leandro et al. 2006a; Runquist et al.

2009b; Saloheimo et al. 2007; Subtil and Boles 2011; Young et al. 2011). Of note is the work of Young and coworkers, where they applied directed evolution to xylose-specific transporters for expression in yeast (Young et al. 2012a; Young et al. 2014a). In principle, analogous approaches can be employed to further improve xylose transport in

Z. mobilis. Second, we note that transporter expression complements other approaches employed to improve xylose metabolism in Z. mobilis including serial adaptation and targeted deletions (Agrawal and Chen 2011; Agrawal et al. 2011a; Agrawal et al. 2012a).

These strategies can potentially be combined to further improve xylose metabolism in Z. mobilis.

2.5 Tables and figures

Table 2.1 Plasmids used in this study.

Plasmid Genotype or relevant characteristics Source or reference pEZCm CmR, pACYC184-pZMO1 Jeff Skerker pJS71 SpR, pBBR1MCS-lacZα (Skerker and Shapiro 2000a) pVenus KanR, MCS yfp(venus) t0 attλ, oriR6K (Saini et al. 2009) pKLD1 pJS71-PgapxylAB This study pKLD2 pJS71-PgapxylAB-PenotalB This study pKLD3 pJS71-PgapxylAB-PenotalB-tktA This study pKLD4 pEZCm-PpdcxylE This study pKLD5 pEZCm-Ppdcglf This study pKLD6 pEZCm-Pglfvenus This study pKLD7 pEZCm-Ppdcvenus This study pKLD8 pEZCm-PglfxylE-venus This study

Table 2.2 Oligonucleotide primers used in this study.

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD001F Pgap, for ATGCCTCGAGACTTTGTTCGATCAACAACCCG KD001R xylAB GTTTATTCTCCTAACTTATTAAGTAGCTACTATATTCC

KD002F xylAB, for GTTAGGAGAATAAACATGCAAGCCTATTTTGACCAGCTCGATCGCG KD002R Pgap ATGCTCTAGAGCCATTAACAAATGATTTTCAGAATA

KD003F Peno, for ATGCTCTAGACCTTCATGTTTTGCTTCATG KD003R talB ATCGAAACCTTTCTTAAAATCTTTTAGAC

KD004F talB, for AAGAAAGGTTTCGATATGACGGACAAATTGACCTC KD004R Peno ATGCGGGCCCTTACAGCAGATCGCCGATCA

KD005F tktA ATGCGGGCCCTCATCCGATCTGGAGTCAAAATGTCCTCACG KD005R ATGCGAGCTCCCGCAAACGGACATATCAAGG

KD006F PPDC, for ATGCTCTAGATGCGGTATAATATCTGTAACAGCTCATTGAT KD006R xylE TGCTTACTCCATATATTCAAAAC

KD007F xylE, for TATATGGAGTAAGCAATGAATACCCAGTATAATTCCAG KD007R PPDC ATGCTCTAGAGAAACGGCTGAAGCGGG

KD008F PPDC, for ATGCTCTAGATGCGGTATAATATCTGTAACAGCTCATTGAT KD008R glf CAGAACTCATTGCTTACTCCATATATTCAAAACACTATGTCTGAA

KD009F glf, for GTGTTTTGAATATATGGAGTAAGCAATGAGTTCTGAAAGTAGTCAGGGTCT AG KD009R PPDC ATGCTCTAGAAGTATAATTGGTTAAGACTTATCTAAAAAAGACAAAAGGA TTC

KD010F Pglf, for ATGCTCTAGATTTTTTAAAAAGAAAACTGTTTTTTTAAACACTTATGTTGC KD010R Venus CTTCTCCTTTACTCATGGCGATTCCTCTCCCGCC

KD011F Venus, for AGGAATCGCCATGAGTAAAGGAGAAGAACTTTTCACTGGAG KD011R Pglf ATGCCTCGAGACCGAGCGTTCTGAACAAATCCAG

KD012F Ppdc, for ATGCTCTAGATGCGGTATAATATCTGTAACAGCTCATTGAT KD012R Venus TTCTTCTCCTTTACTCATTGCTTACTCCATATATTCAAAACACTATGTCTGA

KD013F Venus, for TGGAGTAAGCAATGAGTAAAGGAGAAGAACTTTTCACTGGAG KD013R Ppdc ATGCCTCGAGACCGAGCGTTCTGAACAAATCCAG

KD014F Pglf, for ATGCTCTAGATTTTTTAAAAAGAAAACTGTTTTTTTAAACACTTATGTTGCC KD014R xylE GGCGATTCCTCTCCCGCC

KD015F xylE, for Pglf GGCGGGAGAGGAATCGCCATGAATACCCAGTATAATTCCAGTTATATATT TTC KD015R and Venus CTCCTTTACTCATGCTACCGCTGCCGCTACCCAGCGTAGCAGTTTGTTGTGT TTTCT

KD016F Venus, for GCAGCGGTAGCATGAGTAAAGGAGAAGAACTTTTCACTGGA KD016R xylE ATGCCTCGAGACCGAGCGTTCTGAACAAATCCA

Table 2.3 Performance summary of the transporter constructs in RM containing xylose as carbon source.

Max. Ethanol Max. ethanol Max. xylose Trans- Max. xylitol Xylitol yield ethanol yield (g/g prod. rate cons. rate porter conc. (g/L) (g/g sugar) conc. (g/L) sugar) (g/L h) (g/L h) 10 g/L xylose None 3.4 ± 0.1 0.33 ± 0.0 0.17 ± 0.02 0.45 ± 0.05 0.40 ± 0.03 0.04 ± 0.002 XylE 3.4 ± 0.09 0.32 ± 0.02 0.13 ± 0.007 0.37 ± 0.02 0.37 ± 0.05 0.04 ± 0.005 Glf 3.4 ± 0.07 0.34 ± 0.02 0.15 ± 0.01 0.44 ± 0.03 0.36 ± 0.03 0.04 ± 0.002 50 g/L xylose None 15.7 ± 0.1 0.36 ± 0.01 0.14 ± 0.005 0.37 ± 0.004 4.2 ± 0.02 0.10 ± 0.001 XylE 16.9 ± 0.1 0.37 ± 0.007 0.17 ± 0.004 0.48 ± 0.01 4.8 ± 0.09 0.11 ± 0.003 Glf 16.6 ± 0.2 0.37 ± 0.01 0.18 ± 0.003 0.47 ± 0.005 4.0 ± 0.4 0.09 ± 0.008 100 g/L xylose None 11.9 ± 1.1 0.35 ± 0.01 0.09 ± 0.004 0.29 ± 0.02 3.7 ± 0.04 0.12 ± 0.01 XylE 21.8 ± 0.8 0.40 ± 0.004 0.18 ± 0.008 0.70 ± 0.07 5.5 ± 0.09 0.07 ± 0.002 Glf 18.9 ± 0.5 0.39 ± 0.008 0.16 ± 0.006 0.50 ± 0.07 4.7 ± 0.04 0.08 ± 0.003

Table 2.4 Performance summary of the transporter constructs in RM containing glucose and xylose as carbon sources.

Max. Max. Max. Max. Max. Ethanol Trans- ethanol ethanol glucose xylose xylitol Xylitol yield yield (g/g porter conc. prod. rate cons. rate cons. rate conc. (g/g sugar) sugar) (g/L) (g/L h) (g/L h) (g/L h) (g/L) 10 g/L glucose, 10 g/L xylose None 8.3±0.10 0.42±0.01 0.51±0.02 0.49±0.02 0.78±0.01 0.63±0.01 0.03±0.001 XylE 8.3±0.09 0.42±0.02 0.37±0.02 0.63±0.01 0.58±0.04 0.65±0.01 0.03±0.001 Glf 8.3±0.08 0.41±0.02 0.51±0.02 0.46±0.01 0.41±0.05 0.61±0.01 0.03±0.001 25 g/L glucose, 25 g/L xylose None 20.7±0.2 0.42±0.01 1.00±0.02 1.42±0.05 0.94±0.03 2.45±0.06 0.05±0.001 XylE 20.7±0.3 0.45±0.01 0.91±0.02 1.68±0.05 0.92±0.09 1.91±0.02 0.04±0.001 Glf 20.7±0.2 0.43±0.01 0.99±0.02 1.42±0.03 0.96±0.05 2.20±0.07 0.05±0.002 50 g/L glucose, 50 g/L xylose None 34.4±0.3 0.44±0.01 1.92±0.01 3.26±0.08 0.53±0.08 6.67±0.05 0.09±0.0 XylE 39.6±0.2 0.46±0.01 1.06±0.04 1.62±0.05 0.82±0.03 4.31±0.08 0.05±0.001 Glf 33.8±1.3 0.44±0.02 1.92±0.02 3.29±0.03 0.57±0.05 5.99±0.13 0.08±0.001

Figure 2.1 Schematic representations of plasmids. a Plasmid used to express the xylose metabolism genes. b Plasmid used to express the transporters.

Figure 2.2 Transporter localization analysis. a Z. mobilis expressing fluorescently- labeled XylE. b Z. mobilis expressing free fluorescent protein. Note that fluorescence is concentrated at the cell’s edges in a, showing that XylE is correctly expressed in the membrane (bar = 5 µm).

Figure 2.3 Analysis of the Ppdc-xylE transporter construct in sugar-rich xylose-only media. a,b Xylose and ethanol concentrations for Z. mobilis with pKLD3 and pEZCm (Control) and Z. mobilis with pKLD3 and pKLD4 (+XylE). c Growth of the strains as determined by OD600 nm.

Figure 2.4 Analysis of the Ppdc-xylE transporter construct in sugar-rich glucose-xylose mixed media. a-c Glucose, xylose, and ethanol concentrations for Z. mobilis with pKLD3 and pEZCm (Control), and Z. mobilis with pKLD3 and pKLD4 (+XylE). d Growth of the strains as determined by OD600 nm.

Chapter 3: High-throughput sequencing reveals adaptation- induced mutations in pentose-fermenting strains of Zymomonas mobilis2

3.1 Introduction

In chapter 2, we found that pentose-specific transporter expression can improve xylose fermentations in metabolically-engineered Z. mobilis (Dunn and Rao 2014).

However, this result was only true at high sugar concentrations, primarily in media containing xylose as the sole carbon source. Therefore, we wanted to elucidate additional bottlenecks in pentose metabolism. To do so, we used adaptive evolution and high- throughput sequencing.

Adaptation is a powerful technique for strain development, because it does not require knowledge of the specific metabolic bottlenecks in need of targeting (Lawford et al. 1998; Rosenberg 2001). In a notable study, Agrawal and coworkers used adaptive evolution to isolate a strain of Z. mobilis capable of consuming xylose up to three times faster than its parent (Agrawal et al. 2011b). Previous studies had identified xylitol production as an important bottleneck in xylose metabolism due to the loss of carbon and energy (Feldmann et al. 1992; Kim et al. 2000a). Aware of this finding, Agrawal and coworkers determined that reduced xylitol production was partially responsible for their strain’s improved performance. Additionally, they found that the activity of the heterologously-expressed xylose isomerase enzyme was increased by roughly fivefold in their adapted strain. They did not, however, determine all of the genetic mutations collectively responsible for the observed phenotype (Agrawal et al. 2011b; Agrawal et al.

2012b). Similarly, Deanda and coworkers noted that their arabinose-fermenting strain of 2Portions of this chapter were reprinted with permission from “High-throughput sequencing reveals adaptation-induced mutations in pentose-fermenting strains of Zymomonas mobilis” in the journal Biotechnology and Bioengineering, 2015, by Dunn and Rao. This article is protected by copyright. All rights reserved. 44

Z. mobilis performed better after adaptation in arabinose-only media, but the authors were unable to attribute the improvement to any obvious differences in arabinose-related enzyme activities (Deanda et al. 1996b).

Intrigued by these studies, we also adapted xylose and arabinose-fermenting strains of Z. mobilis to growth on these pentose sugars as their sole carbon sources (Dunn and Rao 2015). We isolated strains with rapid growth, sugar consumption, and ethanol production rates. In order to identify additional bottlenecks in pentose metabolism, we then used high-throughput sequencing to pinpoint the genetic changes responsible for the improved phenotypes. In both the xylose and the arabinose-adapted strains, we found mutations in the coding region of the primary sugar transporter, Glf, and in the DNA repair protein AddB. Consistent with the results of Agrawal and coworkers, we found a mutation in the xylose reductase gene in the xylose-fermenting strain. In the arabinose- fermenting strain, we found mutations in the native glyceraldehyde-3-phosphate dehydrogenase and transketolase coding sequences. Collectively, our results reveal potential targets for the future engineering of pentose metabolism in Z. mobilis.

3.2 Materials and methods

3.2.1 Bacterial strains, media, and growth conditions

All bacterial strains used in this study are listed in Table 3.1. Liquid and solid rich medium (RM) with 10 g/L yeast extract, 2 g/L potassium phosphate, dibasic supplemented with glucose, xylose, or arabinose at the designated concentrations were used for all Z. mobilis experiments. All Z. mobilis cultures were grown at 30°C and antibiotics were used at the following concentrations: spectinomycin, 100 µg/mL; and chloramphenicol, 100 µg/mL.

3.2.2 Plasmid construction

All plasmids used in this study are listed in Table 3.1. All primers used in this study are listed in Table 3.2. The construction of a plasmid containing the xylose fermentation genes (Figure 3.1a) was described previously (Dunn and Rao 2014). To construct a plasmid containing the arabinose metabolism and pentose phosphate pathway genes, the L-arabinose isomerase (araA), L-ribulokinase (araB), L-ribulose-5-phosphate

4-epimerase (araD), transaldolase (talB), and transketolase (tktA) genes were isolated from E. coli MG1655 genomic DNA by PCR synthesis using primers KD024F, KD024R,

KD025F, KD025R, KD021F, KD021R, KD022F, and KD022R, respectively. To express the transaldolase and transketolase genes in Z. mobilis, the transaldolase gene from its start to stop codon was first fused to the strong and constitutive 308 bp Z. mobilis enolase

(eno) promoter, amplified from Z. mobilis ZM4 genomic DNA using primers KD020F and KD020R, using overlap extension PCR (Horton et al. 1989). The constructed 1264 bp Peno-talB fragment was then inserted into plasmid pJS71 to yield plasmid pKLD9.

Next, the transketolase gene including its native ribosome binding site and native terminator sequences was inserted directly downstream of the talB stop codon in the plasmid pKLD9 to create the 3348 bp Peno-talB-tktA synthetic operon, yielding plasmid pKLD10. To express the araBAD genes in Z. mobilis, repetitive palindromic sequences were first removed from between araA and araD, as was done by Deanda and coworkers

(Deanda et al. 1996b). The araBA and araD gene fragments were then fused from the araB start codon to the araD terminator sequence to the strong and constitutive 310 bp Z. mobilis glyceraldehyde-3-phosphate dehydrogenase (gap) promoter, amplified from Z. mobilis ZM4 genomic DNA using primers KD023F and KD023R, using overlap

46 extension PCR. The resulting 4397 bp Pgap-araBAD synthetic operon was then inserted into plasmid pKLD10 to form plasmid pKLD11 (Figure 3.1b).

To construct plasmids for mutant transporter analyses, glf including its native promoter and terminator sequences was amplified from Z. mobilis ZM4 genomic DNA, glfA18T including its native promoter and terminator sequences was amplified from the genomic DNA of strain KLD1, and glfV275F including its native promoter and terminator sequences was amplified from the genomic DNA of strain KLD2, by PCR synthesis using primers KD026F and KD026R. The glf genes were then inserted separately into plasmid pEZCm to yield plasmids pKLD12, pKLD13, and pKLD14, respectively. To construct a vector containing the glfA18TV275F allele, the first half of the glf gene containing the A18T mutation was amplified from the genomic DNA of strain KLD1 by PCR synthesis using primers KD027F and KD027R. The second half of the glf gene containing the V275F mutation was amplified from the genomic DNA of strain KLD2 by PCR synthesis using primers KD028F and KD028R. The two glf fragments were then fused together using overlap extension PCR. The constructed glfA18TV275F fragment was then inserted into plasmid pEZCm to yield plasmid pKLD15.

To construct a plasmid for the purification of xylulokinase from E. coli, the XKS1 coding region was amplified from Saccharomyces cerevisiae BY4742 genomic DNA by

PCR synthesis using primers KD029F and KD029R. The XKS1 gene was then inserted into plasmid pPROTet.E in-frame with the vector’s 6xHN tag to yield plasmid pKLD16.

To construct a vector for the inactivation of addB, the region 1 kb upstream of the addB gene was amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using

47 primers KD030F and KD030R. The region 1 kb downstream of the addB gene was also amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers KD031F and KD031R. The spectinomycin resistance gene was amplified from plasmid pJS71 by

PCR synthesis using primers KD032F and KD032R to include flippase recognition target

(FRT) sequences for the future removal of the resistance from the chromosome, if necessary (Datsenko and Wanner 2000; Martinez-Morales et al. 1999). The pPROTet.E vector backbone was amplified from the pPROTet.E vector by PCR synthesis using primers KD033F and KD033R. The four DNA fragments were then phosphorylated using

T4 polynucleotide kinase (New England Biolabs), and combined with Ampligase thermostable DNA ligase and buffer (epicentre) and bridging oligonucleotides KD034F,

KD034R, KD035F, and KD035R. The fragments were then joined together such that the addB upstream and downstream regions were flanking the spectinomycin gene in the pPROTet.E vector using the optimized ligase cycling reaction parameters reported by de

Kok and coworkers (de Kok et al. 2014) to yield plasmid pKLD17.

To construct a pBBR1-based vector for plasmid stability analyses, the 500 bp pyruvate decarboxylase (pdc) promoter was amplified from Z. mobilis ZM4 genomic

DNA by PCR synthesis using primers KD012F and KD012R (Dunn and Rao 2014). The

Venus fluorescent protein gene was amplified from its start codon and including its native terminator sequence from plasmid pVenus by PCR synthesis using primers KD013F and

KD013R (Dunn and Rao 2014). The pdc promoter was then fused to the Venus gene using overlap extension PCR. The resulting Ppdc-Venus construct was then inserted into plasmid pJS14 to yield plasmid pKLD18.

3.2.3 Plasmid transformation into Z. mobilis using electroporation

Z. mobilis cells grown to an OD600nm of at least 0.8 in RM containing 20 g/L glucose were harvested and washed once with water and once with 10% (v/v) glycerol.

Cells were then resuspended in 10% glycerol to 1% of the original culture volume. For each transformation, at least 1 µg of unmethylated plasmid DNA isolated from E. coli strain GM119 and 2.5 µg of TypeOne Restriction Inhibitor (epicentre) were added to 40

µL of competent cells. Cuvettes with 0.1 cm gaps were used in a BioRad XCell gene pulser set at 1.6 kV, 200 Ω, and 25 µF. Following transformation, cells were allowed to recover in RM at 30°C for 1 to 3 h. Transformants were isolated after growth for 3 to 7 days on RM plates containing the appropriate antibiotic. Plasmid presence was confirmed by plasmid isolation and subsequent transformation into E. coli DH5α. When two plasmids were used, transformations were performed sequentially.

3.2.4 Strain adaptation

Z. mobilis containing plasmid pKLD3 was grown with shaking in liquid RM containing 50 g/L xylose as sole carbon source in a 5 mL glass tube until stationary phase to select for cells able to grow in conditions of low xylose, high ethanol, and high xylitol.

Cells were then subcultured in fresh RM containing 50 g/L xylose for a total of 30 transfers. After this initial adaptation, cells were subcultured into fresh RM containing

100 g/L xylose for an additional 60 transfers (total: roughly 300 generations). The resulting strain was denoted KLD1. Z. mobilis containing plasmid pKLD11 was grown with shaking in liquid RM containing 25 g/L arabinose as sole carbon source in a 5 mL glass tube until stationary phase to select for cells able to grow in conditions of low arabinose and high ethanol. Cells were then subcultured into fresh RM containing 25 g/L

49 arabinose for a total of 60 transfers (roughly 200 generations). The resulting strain was denoted KLD2.

3.2.5 Fermentations

Cultures used for analyses were grown as was described previously (Dunn and

Rao 2014). Briefly, to prepare pre-seed cultures, a single colony from an RM agar plate was inoculated into liquid RM in a 5 mL glass tube and grown until stationary phase. Pre- seed cultures were then used to inoculate seed cultures. Seed cultures were grown with shaking until late exponential phase in 15 mL sealed conical tubes filled with RM to 60% volume. All pre-seed and seed cultures were grown in RM containing 20 g/L glucose as carbon source. Cultures used for analyses were inoculated with seed culture to an

OD600nm of 0.02. They were then grown with shaking at 140 rpm in 50 mL sealed conical tubes filled with RM to 90% volume and containing the appropriate amounts of xylose or arabinose as carbon source and the appropriate amount of antibiotic(s).

3.2.6 Analytical methods

Fermentation cultures were analyzed as was described previously (Dunn and Rao

2014). Briefly, the optical density (OD) at 600 nm was used to monitor cell growth.

Culture samples were purified using 0.22 µm polyethersulfone syringe filters and the resulting supernatant was analyzed for xylose, arabinose, xylitol, and ethanol concentrations using high-performance liquid chromatography (HPLC). A Shimadzu

HPLC system with a Shimadzu RID-10A refractive index detector was used together with an Aminex HPX-87H carbohydrate analysis column and a cation H micro-guard cartridge (Bio-Rad) kept at a temperature of 65°C. The 5 mM H2SO4 mobile phase was pumped at a flow rate of 0.6 mL/min. Peaks were identified and quantified by retention

50 time comparison to authentic xylose, arabinose, xylitol, and ethanol standards. For cell aggregation analyses, cells grown to late exponential phase were mounted on an agar pad.

Phase contrast images were captured on a Leica AF6000 microscope fitted with a

Hamamatsu EM-1K EM-CCD camera, automated stage, and adaptive focus control. A x100/1.47 oil immersion objective was used.

3.2.7 Sequencing reactions

Chromosomal DNA was isolated from Z. mobilis strains using the Qiagen

QIAamp DNA mini kit according to the manufacturer’s instructions. Shotgun genomic libraries were prepared at the Roy J. Carver Biotechnology Center (CBC) at the

University of Illinois at Urbana-Champaign using the TruSeq DNA sample preparation kit (Illumina). Libraries were sequenced by the CBC on an Illumina HiSeq 2500 instrument. Paired-end reads were generated for all samples. Both the forward and reverse reactions for each sample generated over 23 million reads. Fastq files were generated using Casava 1.8.2 (Illumina). Fastq files were analyzed using the CLC

Genomics Workbench 7.0.3 (http://www.clcbio.com).

3.2.8 Quantitative PCR (qPCR) reactions

For qPCR reactions carried out using DNA, total DNA was isolated from Z. mobilis strains using the Qiagen QIAamp DNA mini kit according to the manufacturer’s instructions. PCR amplification of this DNA using primers KD037F- KD043R, designed using the Primer3Plus software (Untergasser et al. 2007), was visualized using the

HotStart-IT SYBR Green qPCR Master Mix with UDG (Affymetrix) and a Bio-Rad

MiniOpticon Real-Time PCR System. For qPCR reactions carried out using RNA, total

RNA was isolated from Z. mobilis strains grown to mid-exponential phase using the

Qiagen RNeasy Mini Kit according to the manufacturer’s instructions. cDNA was generated from RNA using the Qiagen QuantiTect Reverse Transcription Kit according to the manufacturer’s instructions. PCR amplification of cDNA using primers KD037F-

KD049R, designed using the Primer3Plus software (Untergasser et al. 2007), was visualized using the HotStart-IT SYBR Green qPCR Master Mix with UDG (Affymetrix) and a Bio-Rad MiniOpticon Real-Time PCR System. All results were quantified by PCR cycle-number comparison to standard curves prepared using E. coli MG1655 or Z. mobilis ZM4 genomic DNA. The Z. mobilis rrsA gene was used as a control for normalizing differences in total DNA or RNA quantities (He et al. 2012).

3.2.9 Enzymatic assays

Enzymatic assay methods were adapted from Agrawal and coworkers (Agrawal et al. 2011b). Briefly, Z. mobilis cells grown in RM containing 50 g/L xylose and 50 g/L glucose for 48 h or 25 g/L arabinose and 25 g/L glucose for 24 h were harvested by centrifugation and washed once with freshly prepared extraction buffer (10 mM Tris-HCl pH 7.5, 5 mM MgCl2, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride).

Cell pellets were resuspended in extraction buffer to an OD600nm of 100 and sonicated for

8 10 s cycles. Sonicated cells were spun at 52,000 x g for 20 min at 4°C and the resulting supernatant was used for enzymatic assays. Reactions were carried out in quartz cuvettes in a final volume of 400 µL. The absorbance of NADH was measured at 340 nm using a

Shimadzu BioSpec-1601 DNA/ Protein/ Enzyme Analyzer. Glyceraldehyde-3-phosphate dehydrogenase activity was assayed in 42 mM Tris-HCl buffer, pH 8.5, containing 10 mM trisodium phosphate, 1 mM NAD, 3mM β-mercaptoethanol, 1 mM DL- glyceraldehyde-3-phosphate, and 20 µL cell-free extract. Transketolase activity was

52 measured using the coupling assay reported by Lee and coworkers (Lee et al. 2008).

Briefly, xylulokinase from S. cerevisiae was overexpressed in E. coli using plasmid pKLD16. Cells were harvested and washed once with phosphate buffered saline (PBS).

The cell pellet was resuspended in 3 mL lysis buffer (50 mM NaH2PO4, 300 mM NaCl, and 10 mM imidazole, pH 8.0) for every gram of cells and sonicated for 6 15 s cycles.

Sonicated cells were spun at 2,934 x g for 20 min at 4°C. The resulting supernatant was then passed through a gravity chromatography column containing Ni-NTA resin to yield purified 6xHN-Xks1. The purified 6xHN-Xks1 was used to create xylulose 5-phosphate for the transketolase assay in 100 mM Tris-HCl buffer, pH 7.8, containing 2.5 mM xylulose, 2.5 mM ATP, and 20 μg/mL purified 6xHN-Xks1. Transketolase activity was assayed in 42 mM Tris-HCl buffer, pH 8.5, containing 5mM MgCl2, 1 mM thiamine pyrophosphate, 0.5 mM ribose-5-phosphate, 0.2 mM NADH, 2.5 U/mL glycerophosphate dehydrogenase, 25 U/mL triose phosphate isomerase, 80 μL of the 6xHN-Xks1 reaction mixture, and 40 μL cell-free extract. All chemicals and commercial enzymes used for assays were purchased from Sigma-Aldrich (St. Louis, MO). Results were normalized to the total protein content of the lysates, determined using the PierceTM BCA Protein Assay

Kit (Thermo Scientific) according to the manufacturer’s instructions. Relative activities were then determined by setting the activity of the unadapted strain lysate to 1. Statistical significance of the enzyme assay data was determined using the Microsoft Excel t-test, two-sample with unequal variance. Assay results were significantly different (P < 0.05).

3.2.10 Construction of an addB knockout strain

Plasmid pKLD17 was linearized by digestion with EcoRI and KpnI to reduce the likelihood of single crossover events. At least 1 µg of the digested vector was transformed into Z. mobilis as described above. Cells were allowed to recover for 3 h then plated on spectinomycin to select for transformants. Insertional inactivation of addB was confirmed by PCR using primers KD036F and KD036R and genomic DNA chemically extracted using standard protocols. Successful inactivation yielded a PCR product 1.5 kb in length. Unsuccessful inactivation yielded a PCR product 3.0 kb in length. Single crossover events yielded two PCR products, one 1.5 kb in length and another 3.0 kb in length.

3.2.11 Plasmid stability experiments

Plasmid stability was determined using a reporter gene method similar to that of

Meima and coworkers (Meima et al. 1995). Briefly, unadapted Z. mobilis containing plasmid pKLD3 and pKLD7 (Dunn and Rao 2014), unadapted Z. mobilis containing plasmid pKLD11 and pKLD7, and strains KLD1, KLD2, and KLD3 containing plasmid pKLD7 were grown in RM containing spectinomycin and chloramphenicol for a total of

20 media transfers. Unadapted strains of Z. mobilis containing plasmids pKLD3 and pKLD11 and strains KLD1 and KLD2 were also transformed with plasmid pKLD18 and grown in RM containing chloramphenicol only for a total of 20 media transfers (because pKLD3, pKLD11, and pKLD18 have the same origin of replication and therefore cannot be stably co-maintained). Strain KLD3 containing pKLD18 was grown in RM containing spectinomycin and chloramphenicol for a total of 20 media transfers. After this brief adaptation, cells were analyzed using flow cytometry. For a typical flow cytometry

54 experiment, cells were grown overnight in RM containing antibiotic(s), then subcultured to an OD600nm of 0.05 and again grown overnight. At late-exponential phase cells were harvested by centrifugation and washed once with PBS. Cells were then resuspended in

PBS and analyzed using a BD LSR II flow cytometer. Fluorescence values for approximately 500,000 cells were recorded using the fluorescein isothiocyanate (FITC) channel (excitation, 488 nm; emission, 530/30 nm). Separately, cells were stained with

4’,6-diamidino-2-phenylindole (DAPI) to distinguish them from other debris. For a typical flow cytometry experiment with DAPI stain, cells were grown as reported above, harvested and washed once with PBS, then resuspended in 14.3 µM DAPI buffer containing 1 mg/mL chloramphenicol. Cells were incubated at room temperature for 30 minutes then analyzed as reported above, except the Pacific Blue channel (excitation, 405 nm; emission, 450/50 nm) was used. The value of fluorescence at which cells were considered fluorescent versus non-fluorescent was determined by comparison to a Z. mobilis wild-type control. FCS Express version 4 (De Novo Software) was used to extract and analyze the data, which was then exported to Microsoft Excel for the final determination of the percentage of fluorescent and non-fluorescent cells in each sample.

Statistical significance of the differences in plasmid stability in the unadapted versus the adapted strains for xylose and arabinose was determined using the Microsoft Excel t-test, two-sample with unequal variance. Plasmid stabilities were significantly different (P <

0.05).

3.3 Results

3.3.1 Construction, adaptation, and sequencing of pentose-fermenting strains of Z. mobilis

Adaptation can be used to quickly isolate bacterial strains that have improved performance under controlled environmental conditions. We used adaptation to isolate strains of Z. mobilis that perform better when consuming either xylose or arabinose as sole carbon source. The development of a xylose-fermenting strain of Z. mobilis was described previously (Dunn and Rao 2014). An arabinose-fermenting strain was made by expressing the L-arabinose isomerase (araA), L-ribulokinase (araB), L-ribulose-5- phosphate 4-epimerase (araD), transaldolase (talB), and transketolase (tktA) genes from

E. coli in Z. mobilis. Our design is based on the one previously developed by Deanda and coworkers (Deanda et al. 1996b). In order for this strain to consume arabinose as its sole carbon source, we needed to first grow the strain on glucose and then slowly adapt it to arabinose, as reported previously for xylose-fermenting strains (Agrawal et al. 2011b) except cells from the final glucose-containing culture were plated on an RM agar plate containing arabinose as sole carbon source. Attempts to isolate an arabinose-fermenting strain from liquid RM containing arabinose as sole carbon source were unsuccessful.

Following the initial adaptation periods necessary to isolate the pentose- fermenting strains, subculturing was continued for 30 transfers in media containing 5% xylose followed by 60 transfers in media containing 10% xylose, or for 60 transfers in media containing 2.5% arabinose. Sugar concentrations used during adaptation periods were chosen to emulate the conditions the bacterium could experience in industrial fermentations with concentrated lignocellulosic hydrolysates (Dominguez et al. 1996;

Huang et al. 2009). Following adaptation, the pentose-fermenting strains showed improved fermentation characteristics (Figure 3.2). Specifically, we observed significant increases in the maximum rates of pentose sugar utilization (~50-250%) and in the final ethanol titers (~60-80%). To determine additional bottlenecks in pentose fermentations by Z. mobilis, we then used high-throughput sequencing to determine the genetic mutations responsible for the altered phenotypes. The mutations we found compared to the sequence of our wild-type Z. mobilis are listed in Table 3.3. From these, several mutations were selected from each strain for further analysis. The plasmids from the adapted strains were also sequenced but no mutations were found.

3.3.2 Improved pentose sugar transport improves xylose and arabinose fermentation in Z. mobilis

We found mutations in the coding sequence of the native sugar transporter, the glucose facilitated diffusion protein (Glf), in both strain KLD1 and strain KLD2 (Table

3.3). Quantitative PCR results revealed that the mutations did not affect the expression levels of this protein in the adapted strains (Figure 3.4b). Therefore, to predict the effect the mutations had on the activity of the protein, we used the HMMTOP software to perform a sequence-based analysis (Tusnady and Simon 1998; Tusnady and Simon

2001). Consistent with previous reports, we found that the Glf protein likely consists of

12 transmembrane helices and one major cytoplasmic loop (Ren et al. 2009). The A18T mutation likely appears in the first transmembrane helix, while the V275F mutation likely appears in the seventh transmembrane helix. Previous studies have shown that these helices in particular are important to the sugar specificity of Glf and other transporters that also contain 12 transmembrane helices (Arbuckle et al. 1996; Eriksen et al. 2013;

Hashiramoto et al. 1992; Ren et al. 2009; Seatter et al. 1998; Young et al. 2012b). To then confirm that the mutant transporters improve pentose fermentations in Z. mobilis, we expressed the glfA18T and glfV275F alleles from vectors in the xylose and arabinose unadapted strains. We also expressed glfA18TV275F, an allele containing both mutations, in the unadapted strains. We then compared these strains to a strain expressing the wild-type glf from a vector. Strains expressing glfA18TV275F performed poorly; therefore, this construct was removed from further analysis. However, cells expressing glfA18T and glfV275F outperformed the strain expressing the wild-type transporter.

Specifically, glfA18T expression led to a 44% higher ethanol titer in the unadapted xylose-fermenting strain than did expression of the wild-type glf (Figure 3.3). In the unadapted arabinose-fermenting strain, expression of glf from a vector led to an extremely long lag phase, while the expression of glfV275F did not (Figure 3.3). The

GlfA18T xylose-adapted transporter also improved ethanol production in the unadapted arabinose-fermenting strain, and the GlfV275F arabinose-adapted transporter improved ethanol production in the unadapted xylose-fermenting strain. Interestingly, the improvements seen were greater when the expressed transporter was adapted on the opposite sugar, for reasons unknown (Figure 3.3). Consistent with our previous findings

(Dunn and Rao 2014), these results demonstrate that sugar transport is a bottleneck in pentose fermentations by Z. mobilis. They also suggest that the sugar transport limitation is more pronounced during growth on arabinose than during growth on xylose, because the mutant transporters lead to substantially better performance than does the wild-type

Glf during growth on this sugar.

3.3.3 Improved plasmid stability improves xylose and arabinose fermentation in Z. mobilis

We found mutations in the coding sequence of the double-strand break repair protein, AddB, in both strain KLD1 and strain KLD2 (Table 3.3). In other bacteria,

AddB has been shown to affect concatemer formation, typically from plasmids that replicate via a rolling-circle mechanism. Specifically, long repeats of plasmid DNA have been found in cells expressing a mutant form of addB (Viret and Alonso 1987).

Concurrently, the copy number of certain vectors was also affected (Viret and Alonso

1987). Because plasmids pKLD3 and pKLD11 do not replicate via rolling-circle replication (Antoine and Locht 1992), it was unlikely that concatemer formation and plasmid copy number were affected in adapted Z. mobilis due to mutations in addB.

Quantitative PCR reactions confirmed this hypothesis—no significant differences were seen in the DNA copy numbers of plasmid-bound genes in the unadapted versus the adapted strains (Figure 3.4a). Additionally, no significant differences in the transcript levels of plasmid-bound genes were seen (Figure 3.4b).

AddB has also been shown to influence structural plasmid stability, specifically affecting the frequency of illegitimate recombination of plasmid DNA (Meima et al.

1997; Meima et al. 1995). To determine if the addB mutations improve plasmid stability in the adapted strains of Z. mobilis, mutation frequency was monitored by the expression of the Venus fluorescent protein on two vectors: one that replicates via rolling-circle replication (pKLD7), and one that replicates by the same mechanism as pKLD3 and pKLD11 (pKLD18). Antibiotic-resistant progeny that no longer fluoresce are assumed to

59 have lost fluorescence due to deletion or inversion mutations on the vector (Meima et al.

1995).

For the vector that replicates via a rolling-circle mechanism, pKLD7, we found that strains KLD1 and KLD2 had the lowest frequency of fluorescence loss after 20 transfers in media containing antibiotics (~2%), and therefore had the highest plasmid stability. The unadapted strains had slightly higher frequencies of fluorescence loss (~4-

8%) while an addB knockout strain of Z. mobilis had a significantly higher frequency of fluorescence loss (~83%) and therefore had the lowest plasmid stability of all of the strains tested (Figure 3.5). These results indicate that the AddB protein plays an important role in the stability of vectors that replicate using a rolling-circle mechanism in

Z. mobilis.

In the case of the vector that does not replicate via rolling-circle replication, pKLD18, we found that strains KLD1 and KLD2 had lower frequencies of fluorescence loss (~3-5%) after 20 transfers in media containing antibiotics compared to the unadapted strains (~11%), suggesting again that structural plasmid stability was increased in these strains (Figure 3.5). Interestingly, however, the addB knockout strain of Z. mobilis performed differently with pKLD18—the fluorescence of nearly all of the cells increased after 20 transfers in media containing antibiotics (Figure 3.5), a result not seen with plasmid pKLD7. An increase in fluorescence is consistent with the formation of cell aggregates. Flow cytometry results also suggested aggregation in the unadapted strains, while suggesting no significant aggregation in the adapted strains, KLD1 and KLD2

(Figure 3.5). Phase-contrast microscopy confirmed these aggregation patterns (Figure

3.6). To potentially explain why aggregation is occurring, we note that in addition to not

60 replicating via rolling circle replication, the vector pKLD18 has another feature that distinguishes it from pKLD7—it contains a functional origin of transfer and is therefore mobilizable. We also note that the RecBCD protein complex commonly replaces the

AddAB complex in other bacteria (Chedin et al. 1998b). RecBCD deletion mutants have been shown to acquire foreign DNA through horizontal gene transfer at a higher frequency than wild-type cells (Harms and Wackernagel 2008). Further, a phylogenetic tree analysis suggests that horizontal gene transfer is promoted when RecBCD or AddAB are lost from the genome (Cromie 2009). Additionally, it has been shown that the Z. mobilis strain used in the present study, ZM4, contains at least two genes necessary for the conjugal transfer of mobilizable plasmids (Seo et al. 2005). Therefore, we hypothesize that in AddB deletion mutants and to a lesser extent in wild-type Z. mobilis, the cells aggregate in an attempt to carry out conjugation for the purpose of horizontal gene transfer. This is possibly in an attempt to obtain a better-performing analogue of the

AddB protein. The mutated forms of AddB present in strains KLD1 and KLD2 may therefore prevent the cells from attempting a futile conjugation process, thereby relieving this metabolic burden.

3.3.4 Reduced xylitol production improves xylose fermentation in Z. mobilis

Agrawal and coworkers found that reduced xylitol production due to a mutation in the Z. mobilis xylose reductase (xr) gene, ZMO0976, was partially responsible for their adapted strain’s improved performance on xylose (Agrawal et al. 2012b). We, too, found a mutation at this locus in strain KLD1 (Table 3.3). We measured the amount of xylitol produced by the unadapted and the adapted xylose-fermenting strains. While strain KLD1 consumed more xylose and therefore produced more total xylitol, ~18% less carbon

61 became xylitol on a gram per gram basis compared to the unadapted strain (Table 3.4).

While the functionality of the enzyme was altered in the adapted strain, quantitative PCR showed that the xr expression level was not affected (Figure 3.4b). These results confirm the importance of xylitol production in xylose fermentation by Z. mobilis.

3.3.5 Reduced glyceraldehyde-3-phosphate dehydrogenase activity and increased transketolase activity improve arabinose fermentation in Z. mobilis

In strain KLD2, we found a mutation in the coding region of the glyceraldehyde-

3-phosphate dehydrogenase (GAPDH) enzyme (Table 3.3). GAPDH is a glycolytic enzyme that reversibly catalyzes the phosphorylation of glyceraldehyde-3-phosphate to yield 1,3-diphosphoglycerate (Conway et al. 1987d). Therefore, this enzyme has the same substrate as the heterologously-expressed transaldolase enzyme, which converts glyceraldehyde-3-phosphate and sedoheptulose-7-phosphate into fructose-6-phosphate and erythrose-4-phosphate (Sprenger et al. 1995). We found that the glyceraldehyde-3- phosphate dehydrogenase activity in the unadapted arabinose-fermenting strain grown in glucose-arabinose mixed media until the arabinose phase of growth was ~67% higher than in the adapted strain, KLD2 (Table 3.5). Quantitative PCR showed that this was not due to differences in expression levels (Figure 3.4b). The reduced GAPDH activity in the adapted strain may increase the access of transaldolase to its substrate, thereby increasing the amount of carbon source that is able to enter central metabolism to become ethanol and ATP.

In strain KLD2, we also found a mutation in the coding region of the Z. mobilis native transketolase (Tkt) enzyme (Table 3.3). Tkt is an enzyme of the pentose phosphate pathway that, together with transaldolase, converts xylulose-5-phosphate into

62 intermediates of central metabolism (Zhang et al. 1995). The activity of this enzyme in wild-type Z. mobilis is low (Feldmann et al. 1992). We found that the Tkt activity in strain KLD2 grown in glucose-arabinose mixed media until the arabinose phase of growth was ~70% higher than in the unadapted arabinose-fermenting strain (Table 3.5).

Quantitative PCR showed that this was not due to differences in expression levels

(Figure 3.4b). These results suggest that low Tkt activity limits arabinose fermentations in Z. mobilis.

3.4 Discussion

We used adaptive evolution and high-throughput sequencing to determine bottlenecks in xylose and arabinose fermentations by Z. mobilis. We found that mutations in the native sugar transporter, Glf, improve both xylose and arabinose fermentations by this organism, suggesting that sugar transport is a bottleneck in pentose metabolism. We also found that mutations in the AddB protein improve structural plasmid stability and reduce the aggregation of cells expressing mobilizable vectors, revealing that the wild- type version of this protein limits pentose fermentations carried out using vector DNA.

Additionally, we found that a mutation in the xylose reductase gene led to reduced xylitol production and improved xylose fermentations by Z. mobilis. Finally, we found mutations in the glyceraldehyde-3-phosphate dehydrogenase and transketolase coding regions that decreased and increased the activity of these enzymes, respectively, and led to improved arabinose fermentations by Z. mobilis.

We previously investigated sugar transport as a bottleneck in xylose fermentations by Z. mobilis. We found that sugar transport was a process bottleneck, but only at high xylose concentrations (Dunn and Rao 2014). In this study, we adapted Z. mobilis to

63 growth at high xylose concentrations. Again, we found that improved xylose transport led to improved xylose fermentation performance, confirming our previous results. Previous reports have also predicted that sugar transport is a limiting factor in arabinose fermentations by Z. mobilis (Deanda et al. 1996b). The Glf protein likely transports this sugar into the cell, but the affinity of Glf for arabinose is much lower than its affinity for other sugars, including glucose and xylose (Weisser et al. 1996b). Consistent with this finding, our results suggest that sugar transport is a more important bottleneck in arabinose fermentations than in xylose fermentations, as mutant forms of Glf lead to more pronounced improvements in fermentation performance by arabinose-fermenting strains. The overexpression of the wild-type Glf nearly abolishes growth by these strains, likely due to the combined modest transport rate of arabinose through this protein and the additional metabolic burden presented by its overexpression. In general, however, the presence of mutations in glf in both the xylose and the arabinose-adapted strains of Z. mobilis suggests that sugar transport engineering is an important strategy toward expanding the substrate range of this organism. In support of this conclusion, sugar transport engineering has been an effective approach toward the improvement of substrate utilization in many microorganisms (Du et al. 2010a; Kuyper et al. 2005;

Leandro et al. 2006b; Runquist et al. 2009a; Young et al. 2012b; Young et al. 2014b).

It is well-known that plasmid stability is an issue in Z. mobilis (Afendra and

Drainas 1987a; Byun et al. 1986; Dong et al. 2011; Kerr et al. 2011). Prior studies have shown that members of the DNA repair system and the DNA restriction-modification systems play important roles in plasmid stability and maintenance in Z. mobilis (Kerr et al. 2011; Vartholomatos et al. 1993). In this study, we found another member of the DNA

64 repair system that is important to plasmid stability: the double-strand break repair protein,

AddB. AddB is the nuclease portion of the AddAB helicase-nuclease protein complex

(Chedin et al. 1998a). Using a vector that replicates using a rolling-circle mechanism and one that does not, we found that strains expressing mutant forms of the AddB protein show increased structural plasmid stability. Consistent with this result, several studies have shown that this protein affects plasmid stability in other bacteria (Kupsch et al.

1989; Meima et al. 1997; Meima et al. 1995; Meima et al. 1996; Peijnenburg et al. 1987).

The precise mechanism by which the AddB protein affects structural plasmid stability in

Z. mobilis is currently unknown. However, in general, the presence of mutations in addB that improve plasmid stability in both the xylose and the arabinose-fermenting strains suggests that the chromosomal integration of heterologous genes is an important step in the development of industrial strains of Z. mobilis. In cases where vector DNA is desired, however, strains expressing the mutant forms of AddB reported here may be viable options. Further, using a mobilizable vector that does not replicate using a rolling-circle mechanism, we found that the loss of the AddB protein leads to the aggregation of cells, possibly to encourage horizontal gene transfer through conjugation. Cells expressing mutant forms of AddB reported here, however, did not aggregate. This result suggests that removing the origin of transfer or mobilization element from pBBR1-based vectors before electroporation into Z. mobilis may improve the performance of strains utilizing these plasmids. If conjugation must be used for plasmid insertion into the cells, however, strains expressing the mutant forms of AddB reported here can serve as viable hosts.

Several previous reports have found that xylitol production and the concurrent production of the toxic by-product xylitol-5-phosphate limit xylose fermentations in Z.

65 mobilis (Agrawal et al. 2011b; Feldmann et al. 1992; Jeon et al. 2005a; Kim et al. 2000a).

We found that reduced xylitol production by Z. mobilis due to a mutation in the xylose reductase gene, ZMO0976, improves xylose fermentations in this organism. Agrawal and

Chen’s strain of Z. mobilis adapted to growth on xylose also contained a mutation at this locus that led to a significant reduction in xylitol production (Agrawal and Chen 2011;

Agrawal et al. 2011b). Consequently, the authors deleted this gene from the chromosome, and found that an even greater reduction in xylitol production could be achieved compared to their adapted strain when xylose isomerase was overexpressed in the knockout cells (Agrawal et al. 2012b). Potentially, a xylose reductase knockout can be combined with the additional mutations reported here to further improve xylose metabolism in Z. mobilis.

Recent studies have suggested that the heterologously-expressed metabolic enzymes may be rate-limiting in xylose fermentations by Z. mobilis (De Graaf et al.

1999a; Kim et al. 2000a). In this study, we found that the same may be true for arabinose fermentations—the native Z. mobilis transketolase gene contained a mutation that increased its activity by ~70%, suggesting that the activity of the heterologous transketolase is insufficient for proper arabinose assimilation. Also in support of this hypothesis, the Z. mobilis glyceraldehyde-3-phosphate dehydrogenase (GAPDH) enzyme contained a mutation that reduced its activity in the adapted cells by ~40%. Because the substrate of GAPDH and the heterologous transaldolase are the same, it is likely that the reduced GAPDH activity increases the access of transaldolase to its substrate. This suggests that the activity of the heterologous transaldolase is also insufficient for proper

66 arabinose assimilation. Interestingly, the transketolase and transaldolase enzymes were not found to be rate-limiting in the xylose-fermenting strain of Z. mobilis.

Collectively, our results reveal additional bottlenecks in pentose fermentations by

Z. mobilis. These bottlenecks are potential targets for the future engineering of pentose metabolism in the strain. Additionally, several of the mutations reported here are not specific to xylose and arabinose metabolism—they can be applied toward the metabolic engineering of Z. mobilis to consume additional substrates and to produce additional products.

3.5 Tables and figures

Table 3.1 Bacterial strains and plasmids used in this study.

Genotype or relevant characteristics Source or reference Strains KLD1 ATCC 31821 [pKLD3], adapted to growth on xylose This study KLD2 ATCC 31821 [pKLD11], adapted to growth on arabinose This study KLD3 ATCC 31821 ΔaddB::SpR This study Plasmids pEZCm CmR, pACYC184-pZMO1 (Skerker and Shapiro 2000b) pJS71 SpR, pBBR1MCS-lacZα (Skerker and Shapiro 2000b) pJS14 CmR, pBBR1MCS-lacZα (Skerker and Shapiro 2000b) R pPROTet.E Cm , PLTetO-1-RBS-6xHN-MCS-T1 Clontech pVenus KanR, MCS yfp (Venus) t0 attλ, oriR6K (Saini et al. 2009) pKLD3 pJS71-PgapxylAB-PenotalB-tktA (Dunn and Rao 2014) pKLD7 pEZCm-PpdcVenus (Dunn and Rao 2014) pKLD9 pJS71-PenotalB This study pKLD10 pJS71-PenotalB-tktA This study pKLD11 pJS71-PenotalB-tktA-PgaparaBAD This study pKLD12 pEZCm-glf This study pKLD13 pEZCm-glf A18T This study pKLD14 pEZCm-glf V275F This study pKLD15 pEZCm-glf A18TV275F This study pKLD16 pPROTet.E-XKS1 This study pKLD17 pPROTet.E-addBUP-SpR-addBDOWN This study pKLD18 pJS14-PpdcVenus This study

Table 3.2 Oligonucleotide primers used in this study.

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD020F Peno, for ATGCTCTAGACCTTCATGTTTTGCTTCATG KD020R talB ATCGAAACCTTTCTTAAAATCTTTTAGAC

KD021F talB, for AAGAAAGGTTTCGATATGACGGACAAATTGACCTC KD021R Peno ATGCGGGCCCTTACAGCAGATCGCCGATCA

KD022F tktA ATGCGGGCCCTCATCCGATCTGGAGTCAAAATGTCCTCACG KD022R ATGCGAGCTCCCGCAAACGGACATATCAAGG

KD023F Pgap, for ATGCCTCGAGTTTGTTCGATCAACAACCCGA KD023R araBA GTTTATTCTCCTAACTTATTAAGT

KD024F araBA, for ACTTAATAAGTTAGGAGAATAAACATGGCGATTGCAATTGGCCT KD024R Pgap and araD TTAGCGACGAAACCCGTAATA

KD025F araD, for TATTACGGGTTTCGTCGCTAAGGAAGGAGTCAACATGTTAGA KD025R araBA ATGCTCTAGAGGCCCGTTGTCCGTCGC

KD026F glf ATGCCTCGAGTTGTTGTCGCGGGAGAGGC KD026R ATGCGGTACCAAGCAAACCATCGGCATCAAGAC

KD027F glfA18T, for ATGCCTCGAGTTGTTGTCGCGGGAGAGGC KD027R glfV275F CCTTCTGAAGCCGGAGACCAG

KD028F glfV275F, for CTGGTCTCCGGCTTCAGAAGG KD028R glfA18T ATGCGGTACCAAGCAAACCATCGGCATCAAGAC

KD029F XKS1 ATGCGTCGACTTGTGTTCAGTAATTCAGAGACAGACAAGAG KD029R ATGCTCTAGATTAGATGAGAGTCTTTTCCAGTTCGCTTA

KD030F addB, 1 kb TCGTAGAATGGCCAGAAAGATTGGG KD030R upstream ATTCAAAATTAACCGCGTCTAAGAATGGTATC

KD031F addB, 1 kb GGGAATCCTATGGCTCATCCACAG KD031R downstream TTAGGCTGATAATCTGGTGAATCGCC

KD032F SpR GAAGTTCCTATACTTTCTAGAGAATAGGAACTTCGGAATAGGAACTT CCGAACCACTTCATCCGGGGT KD032R GAAGTTCCTATTCCGAAGTTCCTATTCTCTAGAAAGTATAGGAACTTC CAATTATGTGCTTAGTGCATCTAACGCTT

KD033F pPROTet.E CCTAGGGCGTTCGGCTGC KD033R TCTAGGGCGGCGGATTTGT

KD034F pPROTet.E-addB ACAAATCCGCCGCCCTAGATCGTAGAATGGCCAGAAAGATTGGG up bridge KD034R addB up- SpR ACCATTCTTAGACGCGGTTAATTTTGAATGAAGTTCCTATACTTTCTA bridge GAGAATAGGAAC

KD035F SpR- addB down AGAGAATAGGAACTTCGGAATAGGAACTTCGGGAATCCTATGGCTCA bridge TCCACA KD035R addB down GGCGATTCACCAGATTATCAGCCTAACCTAGGGCGTTCGGCTGC bridge- pPROTet.E

KD036F ΔaddB::SpR GATGTCAGCACGCCCAAAGC KD036R check GTTCTGCCTCCAAAGGCATTAATCG

Table 3.2 (cont.)

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD037F xylA qPCR TGCTACTGGCACACCTTCTG KD037R TCCACATCGTGGAAGCAATA

KD038F xylB qPCR AGCATTCTCTGCAGGACGTT KD038R CGCTGAACCCATAGCAATTT

KD039F talB qPCR CGTTTGTTGGCCGTATTCTT KD039R AGAATTTCGCCGATGTTACG

KD040F tktA qPCR GAGCATGGACGCAGTACAGA KD040R GCAGCTGACGGAAGTTTTTC

KD041F araB qPCR CCTCGATTTTGGCAGTGATT KD041R GCTGTTCGACGCTAAGCTCT

KD042F araA qPCR GTCACCGATGGCGATAAAGT KD042R TCATGGTGTAGCAGCTTTCG

KD043F araD qPCR AACCGGTGAAGTGGTTGAAG KD043R GATTTCTGCGTCGGTCATTT

KD044F addB qPCR ACAGCCACCCAGTGAAAAAC KD044R GAGCGAAGGCTGTTTAATGC

KD045F rrsA qPCR GTTCGGAATTACTGGGCGTA KD045R CCACTGGTGTTCTTCCGAAT

KD046F gapdh, qPCR TCCGGATTCTGGACTTGAAC KD046R GGTGAAGATACCCGTGCATT

KD047F tklB qPCR AAGGCTCAGTTCGGTGAAGA KD047R ATGACCATCACAAGCCACAA

KD048F xr qPCR CGTTAGGCACATGGGCTATT KD048R CCGACTTTGGTCGCAATAAT

KD049F glf qPCR AGTCACGCGACTAGCCCTAA KD049R ACAACGACCATCCCAGAAAG

Table 3.3 Mutations present in the adapted strains.

KLD1, xylose-adapted strain Position Reference Allele Type Gene Description 4912 C A SNP ZMO0004 cysN- Sulfate adenylyltransferase subunit 285550 G A SNP ZMO0282 Efflux transporter, unknown substrate 368732 G A SNP ZMO0366 glf- Sugar transporter 688071 T G SNP ZMO0692 gyrA- DNA gyrase 815545 C A SNP - Non-coding region 822843 G A SNP ZMO0821 cysZ- Cysteine synthase 992357 G A SNP ZMO0976 Aldo/keto reductase 1114760 G A SNP ZMO1099 addB- Double-strand break repair protein 1120509 T G SNP ZMO1105 Histidine kinase 1387762 G A SNP ZMO1375 rnc- Ribonuclease III 1633801 C T SNP - Non-coding region 1759483 TCT - Deletion ZMO1711 Predicted permease 2046111 G T SNP ZMO1986 TonB-dependent receptor plug KLD2, arabinose-adapted strain Position Reference Allele Type Gene Description 161168 C G SNP ZMO0176 tklB- Transketolase 161483 G A SNP ZMO0177 gap- Glyceraldehyde-3-P dehydrogenase 369503 G T SNP ZMO0366 glf- Sugar transporter 769036 C T SNP - Non-coding region 1115747 A G SNP ZMO1099 addB- Double-strand break repair protein 1157061 C T SNP ZMO1139 ilvl- Acetolactate synthase

Table 3.4 Xylitol yield from the xylose-fermenting unadapted (ATCC 31821 [pKLD3]) and adapted (KLD1) strains grown in RM containing 100 g/L xylose after 5 days.

Strain Xylitol yield (g/g xylose) ATCC 31821 [pKLD03] 0.131 ± 0.004 KLD1 0.107 ± 0.004

Table 3.5 Relative enzymatic activities in cell-free extracts of the arabinose-fermenting unadapted (ATCC 31821 [pKLD11]) and adapted (KLD2) strains, normalized to the activity of the unadapted strain. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Tkt, transketolase.

Strain GAPDH Tkt ATCC 31821 [pKLD11] 1.0 ± 0.06 1.0 ± 0.1 KLD2 0.6 ± 0.03 1.7 ± 0.2

Figure 3.1 Schematic representations of plasmids. a Plasmid used to express the xylose metabolism genes. b Plasmid used to express the arabinose metabolism genes.

Figure 3.2 Adapted strain analyses. a-b Xylose and ethanol concentrations for strain ATCC 31821 [pKLD3] (unadapted) and strain KLD1 (adapted). c Growth of the strains as determined by OD600nm. d-e Arabinose and ethanol concentrations for strain ATCC 31821 [pKLD11] (unadapted) and strain KLD2 (adapted). f Growth of the strains as determined by OD600nm.

Figure 3.3 Quantitative PCR results. a Quantitative PCR results obtained from total cell DNA from strains ATCC 31821 [pKLD3] (Xylose unadapted), KLD1 (Xylose adapted), ATCC 31821 [pKLD11] (Arabinose unadapted) and KLD2 (Arabinose adapted). b Quantitative PCR results obtained from total cell RNA. tklB and gapdh reactions were carried out for the arabinose-fermenting strains only. xr reaction was carried out for the xylose-fermenting strains only. tktA, E. coli transketolase; talB, transaldolase; addB, double-strand break repair protein; xylA, xylose isomerase; xylB, xylulokinase; araA, L- arabinose isomerase; araB, L-ribulokinase; araD, L-ribulose-5-phosphate 4-epimerase; tklB, Z. mobilis transketolase; gapdh, glyceraldehyde-3-phosphate dehydrogenase; xr, xylose reductase; glf, glucose facilitated diffusion protein.

Figure 3.4 Mutant transporter analyses. a-b Xylose and ethanol concentrations for strain ATCC 31821 [pKLD3] [pKLD12] (+Glf), strain ATCC 31821 [pKLD3] [pKLD13] (+Xyl adapted Glf), and strain ATCC 31821[pKLD3] [pKLD14] (+Ara adapted Glf). c Growth of the strains as determined by OD600nm. d-e Arabinose and ethanol concentrations for strain ATCC 31821 [pKLD11] [pKLD12] (+Glf), strain ATCC 31821 [pKLD11] [pKLD13] (+Xyl adapted Glf), and strain ATCC 31821 [pKLD11] [pKLD14] (+Ara adapted Glf). f Growth of the strains as determined by OD600nm.

Figure 3.5 AddB analysis. Flow cytometry fluorescence profiles of strains KLD3 (ΔaddB::SpR, +Venus), ATCC 31821 [pKLD3] (Xyl unadapted, +Venus), KLD1 (Xyl adapted, +Venus), ATCC 31821 [pKLD11] (Ara unadapted, +Venus), and KLD2 (Ara adapted, +Venus) containing either pKLD7 (left) or pKLD18 (right) after 20 media transfers in RM with antibiotic selection. Non-fluorescent ATCC 31821 (WT) cells were used for comparison in determining the percentages of fluorescent and non-fluorescent cells present in each culture.

Figure 3.6 Aggregation analysis. a The arabinose-adapted strain background containing pKLD18 shows little to no aggregation. The xylose-adapted strain background containing pKLD18 behaves similarly. b The arabinose-unadapted strain background containing pKLD18 shows some aggregation. The xylose-unadapted strain background behaves similarly. c The AddB knockout strain, KLD3, containing pKLD18 shows significant aggregation (bar = 10 µm).

Chapter 4: Genetically modifying Z. mobilis using the λ Red recombinase system and a CRISPR-associated nuclease

4.1 Introduction

Although Z. mobilis is primarily known for its ability to produce the biofuel ethanol, this bacterium also holds promise for the production of other value-added products (Belauich and Senez 1965; Johns et al. 1991; McGill and Dawes 1971). In order for alternative product yields and titers to be economical, however, further engineering of the strain to shift metabolism away from ethanol production must be done. Making the necessary chromosomal modifications to Z. mobilis for alternative product synthesis using the outdated and inefficient techniques available to researchers today would be difficult. Specifically, techniques for the construction of gene knockouts and chromosomal integrations, as well as techniques for genome regulation need to be improved if Z. mobilis is to be accepted as an industrial host for products other than biofuels.

Gene knockouts and chromosomal integrations have historically been constructed in Z. mobilis using suicide vectors that require hundreds or thousands of base pairs of homology to the target region but still yield only a few transformants (Agrawal et al.

2012b; Delgado et al. 2002; Kerr et al. 2011; Senthilkumar et al. 2004). Alternatively, bacteriophage recombination systems have been shown to effectively induce homologous recombination between linear DNA fragments and bacterial chromosomes (Datsenko and

Wanner 2000). These systems can therefore be used to introduce a variety of genetic alterations to a bacterial host, including gene knockouts, insertions, and point mutations

(Baba et al. 2006; Datsenko and Wanner 2000; Lemuth et al. 2011; Posfai et al. 2006;

Wang and Pfeifer 2008).

The recombination system that has received the most attention for gene knockouts and integrations is the bacteriophage λ Red system, which contains the proteins Gamma,

Beta, and Exo (Murphy 1998). The Gamma protein is able to inhibit RecBCD-type host nucleases from degrading linear DNA, protecting the cassette (Karu et al. 1975). Exo acts as an exonuclease on double-stranded cassettes to create single-stranded sticky ends that are then bound by the Beta protein and directed to anneal and recombine with complementary DNA on the chromosome (Sharan et al. 2009). When expressed in E. coli, these three proteins can easily induce a linear selectable marker DNA cassette flanked by only 35 bp of homology to crossover into the chromosome (Datsenko and

Wanner 2000). When the selectable marker is flanked by flippase recognition target

(FRT) sites, it can then be easily removed from the chromosome by yeast flippase, leaving behind only a FRT scar (Cherepanov and Wackernagel 1995). This system of using the λ Red proteins to introduce a selectable marker in place of a gene of interest followed by removing the selectable marker gene with flippase has allowed for the rapid construction of gene knockouts and integrations in several species of bacteria (Datsenko and Wanner 2000; Katashkina et al. 2009; Lesic and Rahme 2008; Yamamoto et al.

2009). The system may therefore be effective in Z. mobilis.

The bacteriophage λ protein Beta is also capable of mediating the modification of chromosomal genes using single stranded DNA (ssDNA) oligonucleotides. These oligonucleotides are incorporated into the host genome at the replication fork during

DNA synthesis, leading to allelic replacement (Ellis et al. 2001; Wang et al. 2009; Zhang et al. 1998). The process works most efficiently when the mutS gene, responsible for mismatch repair, has been removed from the chromosome (Costantino and Court 2003).

The oligonucleotides can be designed to target any chromosomal locus or loci. When several cycles of ssDNA incorporation are performed to accumulate mutations, the process is called multiplex automated genome engineering, or MAGE (Wang et al. 2009).

MAGE enables rapid, large-scale evolution of cells, allowing for the isolation of mutagenic strains with improved phenotypes rapidly, efficiently, and inexpensively

(Wang et al. 2009). Theoretically, if the λ Red proteins successfully mediate gene knockouts and integrations in Z. mobilis, MAGE should also be possible in this strain.

While the λ Red recombinase system is a very powerful tool for constructing gene knockouts and integrations with the use of selectable markers, researchers have also recently discovered the power of another system for modifying genomes without selection. The Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) adaptive immunity system utilizes small RNAs for sequence-specific silencing of foreign, invading nucleic acids (Jinek et al. 2012; Wiedenheft et al. 2012). The type II

CRISPR/Cas9 system from Streptococcus pyogenes has been widely utilized for marker- free genome engineering (Cong et al. 2013; Hwang et al. 2013; Jiang et al. 2013; Mali et al. 2013). This system has three primary components: the CRISPR RNA (crRNA), the trans-activating CRISPR RNA (tracrRNA), and the CRISPR associated (Cas) nuclease,

Cas9 (Jinek et al. 2012). The tracrRNA acts as a scaffold for the crRNA, which contains about 20 to 30 base pairs of sequence homology to the invading or target DNA (Mali et al. 2013). This homology is known as the protospacer sequence, and it is flanked by direct repeats that are recognized by the tracrRNA (Ishino et al. 1987; Jinek et al. 2012;

Makarova et al. 2011). The tracrRNA and crRNA can also be fused into a single transcript called guide RNA (gRNA) (Jinek et al. 2012). The tracrRNA:crRNA complex

78 or gRNA transcript recruit the Cas9 protein to the homologous base pairs in the target

DNA (Jinek et al. 2012). In order for Cas9 to successfully bind, however, the target sequence must also contain a Protospacer Adjacent Motif (PAM) sequence at its 3’ end

(Deveau et al. 2008). For Cas9, the optimal PAM sequence is NGG (Mojica et al. 2009).

After successful binding to DNA containing an adjacent PAM, Cas9 is able to cleave, leading to a double-stranded break (Jinek et al. 2012).

For the generation of chromosomal knockouts or integrations utilizing the

CRISPR-Cas system in bacteria, double-stranded DNA flanked by regions of homology to the target gene are introduced into the cells (Jiang et al. 2013). The dsDNA cassette does not need to be a marker gene, thereby allowing for the replacement of chromosomal genes with any gene of interest. DNA coding for the Cas9 protein and gRNA or tracrRNA:crRNA complementary to the unmutated target gene is simultaneously transformed into the cells. After crossover into the chromosome, the cassette will prevent recognition of the target by the Cas9-tracrRNA:crRNA complex, allowing mutated cells to evade cleavage by Cas9. The DNA of unmutated cells, however, will be targeted by the endonuclease and the resulting double-stranded break will likely lead to cell death

(Gudbergsdottir et al. 2011; Jiang et al. 2013; Marraffini and Sontheimer 2010). Due to the broad applicability of the CRISPR system and its proven functionality in a wide variety of cell types and historically recalcitrant organisms (Sander and Joung 2014), it is likely that this system can also be applied for genome engineering in Z. mobilis.

In this chapter, we constructed a λ Red expression system for use in Z. mobilis.

We tested the effectiveness of this system for the development of gene knockouts compared to the effectiveness of natural homologous recombination. We found that

79 although the number of colony forming units was relatively unchanged when the λ Red proteins were expressed, a larger fraction of these colony forming units displayed the desired mutant genotype. We also constructed a vector for the expression of yeast flippase in Z. mobilis to enable marker removal from the chromosome. We found that this enzyme was highly effective in the strain. Because our knockout results suggested activity of the λ Beta protein in Z. mobilis, we constructed a host strain for MAGE.

Unfortunately, however, we were unable to provide evidence for ssDNA incorporation in the strain under the conditions tested here. Finally, we also constructed an expression system for the CRISPR small RNAs and the Cas9 protein in Z. mobilis. We tested the system in Z. mobilis and found that the Cas9 protein is able to induce cell death when directed to a chromosomal gene by the tracrRNA:crRNA complex. We also found that the Cas9 system is likely able to induce dsDNA crossover into the chromosome.

Collectively, our work provides proof-of-concept for the use of modern genome engineering techniques in Z. mobilis and sets the stage for the acceptance of this strain as an alternative industrial host.

4.2 Materials and methods

4.2.1 Bacterial strains, media, and growth conditions

All experiments were performed in Zymomonas mobilis subsp. mobilis ZM4

(ATCC 31821). RM liquid and solid medium (10 g/L yeast extract and 2 g/L potassium phosphate, dibasic) supplemented with 20 g/L glucose was used for all Z. mobilis experiments. All Z. mobilis cultures were grown at 30°C. Antibiotics were used at the following concentrations: spectinomycin, 100 µg/mL; chloramphenicol, 100 µg/mL; and tetracycline, 25 µg/mL.

4.2.2 Plasmid construction

All plasmids used in this study are listed in Table 4.1. All primers used in this study are listed in Table 4.2. To construct a plasmid to express the λ Red genes in Z. mobilis, the gam-beta-exo operon was amplified to include its native ribosome binding site and terminator sequences from plasmid pKD46 by PCR synthesis using primers

KD050F and KD050R. The fragment was then inserted into plasmid pSRKTc directly downstream of the vector’s engineered lac promoter to form plasmid pKLD19.

To construct a plasmid for the expression of yeast flippase in Z. mobilis, the flippase (flp) gene was amplified from its start codon and including its native terminator sequence from S. cerevisiae BY4742 DNA by PCR synthesis using primers KD051F and

KD051R or KD053F and KD053R. The 500 bp pdc promoter was amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers KD054F and KD054R. The two fragments were then joined together using overlap extension PCR (Horton et al.

1989). The Ppdc-flp gene fragment was then inserted into plasmid pJS14 to form plasmid

q pKLD20. Separately, the lacI -Plac fragment was amplified from plasmid pSRKTc by

q PCR synthesis using primers KD052F and KD052R. The lacI -Plac and flp fragments

q were then joined together using overlap extension PCR, and the resulting lacI -Placflp fragment was inserted into plasmid pJS14 to yield plasmid pKLD21.

To construct vectors for the generation of linear dsDNA cassettes, the spectinomycin resistance gene including its native promoter and terminator sequences was amplified by PCR synthesis from plasmid pJS71 using primers KD055F and

KD055R, such that FRT sites were added at both ends of the gene. When FRT sites were not added, primers KD056F and KD056R were used. The pPROTet.E vector was also

81 amplified by PCR synthesis from plasmid pPROTet.E using primers KD057F and

KD057R. One kilobase fragments both upstream and downstream of five of the six target genes (Table 4.3) were amplified using primers KD058F-KD062R and KD063F-

KD067R, respectively. The appropriate DNA fragments were then phosphorylated using

T4 polynucleotide kinase (New England Biolabs), and combined with Ampligase thermostable DNA ligase and buffer (epicentre) and bridging oligonucleotides KD068F-

KD077R, as necessary. The fragments were then joined together such that the target gene upstream and downstream regions were flanking the spectinomycin gene in the pPROTet.E vector using the optimized ligase cycling reaction parameters reported by de

Kok and coworkers (de Kok et al. 2014) to yield plasmids pKLD22-pKLD26, respectively. For the construction of vector pKLD29, the sacB upstream and sacC downstream regions were amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers KD078F, KD078R, KD079F, and KD079R, respectively, and inserted sequentially into pPROTet.E to form plasmids pKLD27 and pKLD28. The spectinomycin gene was amplified from plasmid pJS71 by PCR synthesis using primers

KD080F and KD080R and inserted between the sacB upstream and sacC downstream regions in plasmid pKLD28 to form plasmid pKLD29.

To construct a vector for the development of an ssDNA recombineering host strain, the regions 1 kb upstream and 1 kb downstream of the zmrr gene were amplified using primers KD062F, KD062R, KD067F, and KD067R, respectively. The pUC19 plasmid was amplified from pUC19 using primers KD081F and KD081R. The spectinomycin resistance gene was amplified by PCR synthesis to contain upstream KpnI and BamHI restriction sites from plasmid pJS71 using primers KD082F and KD082R.

The four fragments were then joined together as was described above using primers

KD083F-KD084R and the optimized ligase cycling reaction parameters reported by de

Kok and coworkers to yield plasmid pKLD30. Separately, the plasmid pKD3 was amplified using primers KD085F and KD085R such that a premature stop codon was introduced into the chloramphenicol resistance gene. This PCR product was then phosphorylated using T4 polynucleotide kinase (New England Biolabs) and self-ligated to yield plasmid pKLD31. The chloramphenicol resistance gene containing a premature stop codon was then amplified from plasmid pKLD31 by PCR synthesis using primers

KD086F and KD086R. This allele was then inserted at the KpnI and BamHI sites of pKLD30 to yield pKLD32.

To construct a vector for constitutive expression of the λ Red genes in Z. mobilis, the gam beta exo operon was amplified from its start codon and including its native terminator sequence from plasmid pKD46 by PCR synthesis using primers KD087F and

KD087R. Separately, the strong and constitutive 500 bp pdc promoter was amplified from Z. mobilis ZM4 genomic DNA by PCR synthesis using primers KD088F and

KD088R. The two fragments were then joined together using overlap extension PCR.

The resulting Ppdc-gam-beta-exo DNA was then inserted into plasmid pSRKTc to form plasmid pKLD33.

To construct all-inclusive CRISPR/Cas expression vectors for use in Z. mobilis, the 1.6 kb plasmid pZMO1 maintenance region from pEZCm was amplified by PCR synthesis using primers KD089F and KD089R. The resulting fragment was then inserted into plasmid pCas9 to form plasmid pKLD34. Plasmid pCas9 was a gift from Luciano

Marraffini (Addgene plasmid # 42876). Separately, the 5.3 kb region of pCas9

83 comprising the tracrRNA, cas9, and direct repeat coding regions was amplified by PCR synthesis using primers KD090F and KD090R and inserted into plasmid pSRKTc to form plasmid pKLD35.

To construct co-expression vectors for the CRISPR/Cas system in Z. mobilis, the

CRISPR direct repeats including the 100 bp upstream and downstream regions were amplified by PCR synthesis using primers KD091F and KD091R from plasmid pCas9. A

635 bp fragment containing the tracrRNA coding sequence was also amplified by PCR synthesis from plasmid pCas9 using primers KD092F and KD092R. The two fragments were then joined together using overlap extension PCR and inserted into plasmid pEZCm to yield plasmid pKLD36. Separately, the cas9 gene including its native promoter and terminator sequences was amplified by PCR synthesis using primers KD093F and

KD093R from plasmid pCas9 and inserted into plasmid pSRKTc to form plasmid pKLD37, into plasmid pEZCm to form plasmid pKLD38, and into plasmid pJS71 to form plasmid pKLD39. To insert a zmrr targeting spacer into plasmid pKLD36, a method similar to that of Jiang and coworkers was used (Jiang et al. 2013). Briefly, 2 μM of each primer KD094F and KD094R were first phosphorylated using T4 polynucleotide kinase

(New England Biolabs) in a total reaction volume of 50 μL. The primers were then annealed by adding 50 mM NaCl, incubating for 5 minutes at 95°C, and then slowly cooling to room temperature. The annealed primers were then diluted and ligated into

BsaI-digested pKLD36 to form plasmid pKLD40.

4.2.3 Plasmid transformation into Z. mobilis using electroporation

For standard plasmid transformations, Z. mobilis cells grown to an OD600nm of at least 0.8 in RM containing 20 g/L glucose were harvested and washed once with water

84 and once with 10% (v/v) glycerol. Cells were then resuspended in 10% glycerol to 1% of the original culture volume. For each transformation, at least 1 µg of unmethylated plasmid DNA isolated from E. coli strain GM119 and 2.5 µg of TypeOne Restriction

Inhibitor (epicentre) were added to 40 µL of competent cells. Cuvettes with 0.1 cm gaps were used in a BioRad XCell gene pulser set at 1.6 kV, 200 Ω, and 25 µF. Following transformation, cells were allowed to recover in RM at 30°C for 1 to 3 h. Transformants were isolated after growth for 3 to 7 days on RM plates containing the appropriate antibiotic. Plasmid presence was confirmed by plasmid isolation and subsequent transformation into E. coli DH5α. When two plasmids were used, transformations were performed sequentially unless otherwise indicated.

4.2.4 Z. mobilis λ Red-mediated knockout experiments

Z. mobilis containing plasmid pKLD19 and Z. mobilis without pKLD19 were inoculated from single colonies on agar plates in liquid RM (containing antibiotic as necessary) in 5 mL glass tubes and grown with shaking overnight. Cells were then subcultured to an OD600nm of 0.05 in 50 mL RM containing 500 μM IPTG in 250 mL flasks and grown for 16 hours. Cells were then induced a second time with 500 μM

IPTG. After 3 h, cells were prepared as described above and electroporated with at least 1

μg of cassette plasmid (pKLD22-pKLD26 or pKLD29) linearized by a double restriction digest with enzymes that cut in the vector backbone. Cells were allowed to recover for 3 h, at which time they were plated on RM plates containing spectinomycin and allowed to grow for 3 to 7 days.

Insertional inactivation of target genes was confirmed by PCR using primers

KD095F-KD100R and genomic DNA extracted using standard protocols. For all target

85 genes, successful inactivation through a double crossover event yielded a single PCR product 1.5 kb in length. Unsuccessful inactivation through a single crossover event yielded two PCR products, one 1.5 kb in length and the other the length of the target gene. Transformants obtained through single crossover events were also resistant to chloramphenicol, due to the insertion of the entire cassette vector backbone into the chromosome. False positive spectinomycin-resistant cells yielded a single PCR product the length of the target gene.

4.2.5 Motility plate assays

The FliA protein is a transcriptional regulator of motility (Seo et al. 2005). When it is absent, cells are no longer motile (Chen and Helmann 1992). Motility loss can be detected using a simple plate assay. Specifically, cells were inoculated from single colonies on agar plates in liquid RM in 5 mL glass tubes and grown with shaking until late exponential phase. Roughly 5 x 105 cells were then spotted in the center of an RM agar plate containing 3 g/L agar to allow the cells to swim. Plated cells were allowed to grow overnight, at which time the size of the halo was used as an indicator of motility.

4.2.6 Marker removal from Z. mobilis

Cells with confirmed insertional inactivation of a target gene due to replacement with spectinomycin flanked by FRT sites were transformed with plasmid pKLD20 as was described above. Successful transformants isolated on chloramphenicol agar plates were streaked for individual colonies again on an RM agar plate containing chloramphenicol and grown for 2 days at 30°C. Colonies were then tested for the loss of spectinomycin resistance by streaking on an RM agar plate and an RM agar plate containing spectinomycin. Plasmid pKLD21 was tested in a similar manner, except cells were

86 induced with 500 μM IPTG 3 h before competent cells were prepared. RM agar plates containing chloramphenicol also contained 500 μM IPTG when pKLD21 was used.

4.2.7 ssDNA recombineering experiments

To construct an ssDNA recombineering host strain, vector pKLD32 was linearized with the NdeI and AatII restriction enzymes then used to construct a zmrr knockout in strain KLD10. This led to the chromosomal integration of a non-functional chloramphenicol resistance gene containing a premature stop codon. This strain was designated KLD11. Plasmid pKLD19 and plasmid pKLD33 were then separately electroporated into this strain. These cells were then grown in RM containing spectinomycin and tetracycline, prepared as was described above, and electroporated with

6 μM oligonucleotide KD101F. Cells were allowed to grow for 3 hours before plating on

RM agar plates containing chloramphenicol to select for transformants. When IPTG was necessary, cells were induced as was described above for λ Red-mediated knockout experiments. Experiments with oligonucleotides KD101R and KD102F were similar except Z. mobilis containing plasmid pKLD33 was used as the host strain. All oligonucleotides were designed to target the lagging strand during DNA replication.

4.2.8 Z. mobilis CRISPR/Cas experiments

Z. mobilis cells containing plasmid pKLD37 and Z. mobilis without pKLD37 were inoculated from single colonies on agar plates in liquid RM (containing antibiotic as necessary) in 5 mL glass tubes and grown with shaking overnight. Cells were then prepared and electroporated as was described above, except both plasmid pKLD40 and linearized plasmid pKLD26 were transformed at once. After recovery for 3 h, cells were plated on RM agar plates containing spectinomycin and chloramphenicol and allowed to

87 grow for 3 to 7 days to select for transformants. Insertional inactivation of the target gene was confirmed by PCR using primers KD099F and KD099R as was described above. In a separate experiment, plasmid pKLD40 was transformed individually into Z. mobilis with and without pKLD39. The entire transformations were plated on RM agar plates containing chloramphenicol and the resulting number of colony forming units was quantified.

4.3 Results

4.3.1 Construction of Z. mobilis λ Red recombinase and CRISPR system host strains

Genetic techniques available for use in Z. mobilis are inefficient and unreliable. In order to ensure that the strain reaches its full industrial potential, we wanted to develop protocols for modern genome engineering strategies in Z. mobilis. Specifically, to improve the construction efficiency of gene knockouts and integrations, we developed a strain of Z. mobilis that expresses the phage λ Red proteins Beta, Exo, and Gamma under the control of an inducible lac promoter with a low basal level of expression from a vector (Figure 4.1a). These proteins are able to mediate homologous recombination between linear dsDNA and the chromosome (Datsenko and Wanner 2000). Similarly, we also developed a strain that expresses the three proteins constitutively under the control of the strong Z. mobilis pdc promoter. Exclusively, the Beta protein is also able to mediate the incorporation of ssDNA oligonucleotides into bacterial chromosomes (Ellis et al. 2001). To test if it could do so in Z. mobilis, we constructed a Beta-expressing strain deficient in the MutS protein. MutS is responsible for DNA mismatch repair—when it is

88 not present, the efficiency of ssDNA recombineering is greatly improved (Costantino and

Court 2003).

Another system for genome engineering that has received considerable attention in recent years is the CRISPR/Cas system for marker-free gene knockouts, insertions, and point mutations. To test this system in Z. mobilis, we constructed a strain that expresses the Cas9 protein under the control of its native promoter from a broad-host range vector

(Figure 4.1c). We also constructed a vector for the expression of the small crRNA and tracrRNA transcripts (Figure 4.1b).

4.3.2 The λ Red recombinase system increases double crossover events in Z. mobilis

To test the λ Red recombinase system in our strain of Z. mobilis expressing

Gamma, Beta, and Exo, we first needed to generate linear dsDNA cassettes with regions of homology to the chromosome. Recombineering in alternative hosts frequently requires homology regions of up to 1,000 bp for the best results (Lesic and Rahme 2008; Pontes and Dale 2011; van Kessel and Hatfull 2007). Additionally, Z. mobilis requires at least 1

μg of DNA for a typical transformation to be successful. We attempted to use a 2-step

PCR process to construct cassettes with the required 1,000 bp of homology, but we found that cassette yields were not high enough to ensure transformation success. Therefore, we utilized the optimized ligase cycling reaction parameters reported by de Kok and coworkers (de Kok et al. 2014) to quickly construct high-quality cassettes in a high-copy vector, which we then linearized by restriction double digest. These linear cassettes were then electroporated into Z. mobilis containing plasmid pKLD19 that had been induced with IPTG, and Z. mobilis containing plasmid pKLD19 that had not been induced with

IPTG. After selection, we found roughly the same number of colony forming units on the induced and uninduced plates. However, we had previously observed that single crossover events that led to the chromosomal integration of the entire cassette vector

(including its resistance gene) in Z. mobilis were relatively common (Figure 4.2a).

Therefore, we wanted to determine if the expression of the λ Red proteins increased the likelihood of desirable double crossover events (Figure 4.2b). Using PCR verification

(Figure 4.3) and antibiotic resistance checks, we found that a higher percentage of colonies tested were in fact knockout mutants when the λ Red proteins were induced.

Specifically, in the best case, ~90% of colonies tested displayed the desired ΔsacBC::SpR genotype when the λ Red proteins were induced, while only ~19% of colonies displayed this genotype when they were not (Table 4.4). The constitutive λ Red expression vector pKLD33 yielded similar results as pKLD19 compared to a wild-type Z. mobilis control.

When possible, we also utilized phenotypic characterization to confirm our knockouts

(Figure 4.4).

In an attempt to increase the overall efficiency of the process, λ Red-mediated knockouts were also attempted using KLD5 as the host strain. Strain KLD5 is deficient in

AddA, the Z. mobilis analogue of RecB. However, the addA knockout strain carrying the recombinase vector pKLD19 or pKLD33 grew slowly, did not revive well from freezer stock, and led to only one or two colony forming units after cassette transformation. This made it difficult to obtain meaningful results with this strain.

4.3.3 Yeast flippase removes marker genes from the Z. mobilis chromosome

When FRT sites are placed at either end of an integrated resistance gene, the yeast flippase enzyme can be used to remove the gene from the chromosome. We wanted to

90 determine if flippase was functional in Z. mobilis. We constructed a Z. mobilis flippase expression system by placing this enzyme under the control of an inducible lac promoter with a low basal level of expression from a broad-host range vector. However, very few colonies actually lost resistance when flp expression was induced. To attempt to improve this system, we then constructed a constitutive expression vector with flp placed under the control of the strong and constitutive pdc promoter. With this vector, we found consistently that nearly all candidates tested lost the FRT-flanked spectinomycin resistance cassette, proving that flippase is an effective tool for genome engineering in Z. mobilis.

4.3.4 The λ Red Beta protein does not mediate ssDNA recombineering in a mutS-deficient strain of Z. mobilis

While we did not see λ Red knockout efficiencies in Z. mobilis that were close to those that have been reported in E. coli, our results did suggest that the proteins were active in the strain. Therefore, we wanted to attempt ssDNA recombineering in Z. mobilis. ssDNA recombineering only requires the action of the λ Red Beta protein, but it is common practice to utilize strains expressing all three λ Red proteins during ssDNA oligonucleotide incorporation (Copeland et al. 2001; Wang et al. 2009). The process is also much more efficient in strains deficient in the MutS protein, which is responsible for mismatch repair (Costantino and Court 2003). Therefore, we first constructed a mutS knockout strain of Z. mobilis. To this strain we added a non-functional chloramphenicol resistance cassette containing a premature stop codon at the zmrr locus. This strain, designated KLD11, was confirmed to be sensitive to chloramphenicol. Vector pKLD19 or pKLD33 was then electroporated into strain KLD11 to yield Z. mobilis ssDNA

91 recombineering host strains. The 60 bp ssDNA oligonucleotide KD101F was then electroporated into these strains to attempt to correct the premature stop codon by oligonucleotide incorporation into the chromosome. However, after several attempts, no transformants were recovered on chloramphenicol agar plates. Separately, 60 bp oligonucleotides KD101R and KD102F were transformed into Z. mobilis containing plasmid pKLD33. If incorporated, these oligonucleotides would lead to a loss of motility or a loss of the ability to grow on fructose, respectively, but no fliA or frk null mutants could be found using motility plate assays and fructose growth tests. While ssDNA recombineering may still be possible in Z. mobilis, we found that it is likely not possible under the precise conditions tested here.

4.3.5 The Cas9 protein induces cell death when directed to a Z. mobilis chromosomal gene

As a simple initial test of the CRISPR/Cas system in Z. mobilis, we wanted to see if the Cas9 protein could be directed to create a double-stranded break in the chromosome by a tracrRNA:crRNA complex containing a spacer with genomic homology and the required adjacent PAM sequence (Figure 4.5). We attempted to transform Z. mobilis with two different all-inclusive Cas expression vectors, pKLD34 and pKLD35. pKLD34 contains the tracrRNA-cas9-dr coding region and the pZMO1 native vector replication origin. pKLD35 contains the tracrRNA-cas9-dr coding region and the pBBR1 replication origin. When we attempted to transform these vectors into Z. mobilis, we found no colony forming units on chloramphenicol or tetracycline RM agar plates, respectively, despite several attempts. Because it was possible that simply the expression level of the

Cas9 protein was lethal to the cells, we also constructed three vectors containing the cas9

92 coding region only—pKLD37, pKLD38, and pKLD39. Without the tracrRNA:crRNA complex, the Cas9 protein would no longer be directed to the chromosome, so if these vectors were also unable to be transformed into the cells then we knew Cas9 expression levels and not chromosomal cleavage were responsible for cell death. The three vectors also contained the pBBR1 or pZMO1 vector origins to account for copy number effects.

When we transformed these vectors into Z. mobilis, we found many colony forming units on tetracycline, chloramphenicol, and spectinomycin plates, respectively.

A similar test was carried out using two separate vectors, one carrying the Cas9 protein and the other carrying the small RNAs and a spacer to target the Z. mobilis chromosome at a site with the required adjacent PAM sequence. Cells containing pKLD39 and cells without pKLD39 were transformed with plasmid pKLD40. With Cas9 present on pKLD39, we most frequently saw no colony forming units on an RM agar plate containing chloramphenicol, because the tracrRNA:crRNA complex directed Cas9 to cleave the chromosome (Table 4.5). When Cas9 was not present, we saw many colony forming units on an RM agar plate containing chloramphenicol. Consistent with previous results, we did find a single colony expressing Cas9 that was able to evade cleavage by the nuclease, likely due to a mutation near the PAM site (Jiang et al. 2013). Collectively, these results provide preliminary proof that the tracrRNA:crRNA complex is able to direct Cas9 cleavage of the Z. mobilis chromosome.

4.3.6 The Cas system mediates dsDNA crossover into the Z. mobilis chromosome

To prove the utility of the CRISPR/Cas system in Z. mobilis, we wanted to use it to mediate a gene knockout in the strain. Therefore, we simultaneously transformed cells

93 containing pKLD37 and cells without pKLD37 with pKLD40 and pKLD26 linearized by double restriction digest. In the cells expressing Cas9, the Cas9 protein would be directed to cut the chromosome at the wild-type zmrr allele if the cassette was incorporated through a single crossover event or not incorporated at all, while cells with a successful knockout genotype would avoid cleavage by Cas9 and represent the majority of transformants. In cells not expressing Cas9, undesirable single crossover events would likely represent the majority of the transformants. As expected, PCR verification showed the presence of the wild-type zmrr gene on the chromosome of cells not expressing Cas9, while cells that were expressing Cas9 no longer contained the wild-type zmrr allele

(Figure 4.6). However, when Cas9 was expressed, the chloramphenicol resistance gene was incorporated at the zmrr allele instead of the spectinomycin resistance gene, as expected, for reasons unknown. Collectively, because cells expressing Cas9 lost the wild- type allele and cells without Cas9 did not, our results suggest that the CRISPR/Cas system is likely useful for genome editing in Z. mobilis.

4.4 Discussion

We attempted to apply modern genetic engineering techniques for use in Z. mobilis. We found that the proteins of the λ Red recombinase system are able to increase the frequency of desirable double crossover events when used to mediate gene knockouts in the strain. We also found that the yeast flippase enzyme is very efficient at removing genes flanked by FRT sites from the Z. mobilis chromosome. Unfortunately, we were unable to provide evidence for ssDNA oligonucleotide incorporation into the chromosome of the strain. We did, however, provide evidence of tracrRNA:crRNA- directed Cas9 cleavage of the chromosome.

The proteins of the λ Red system have been used to mediate allelic exchange in a variety of organisms (Datsenko and Wanner 2000; Katashkina et al. 2009; Lesic and

Rahme 2008; Yamamoto et al. 2009). While our results suggest that the proteins are active in Z. mobilis, they also suggest that the system is not highly efficient in the strain.

To potentially explain this result, we first note that Z. mobilis is known to contain highly active type I and type IV DNA restriction-modification (R-M) systems (Kerr et al. 2011).

These systems efficiently degrade any foreign DNA that enters the cell, including our dsDNA cassettes, and are the primary reason why the general transformation efficiency of Z. mobilis is so low (Kerr et al. 2011). In this study, we constructed a strain of Z. mobilis (KLD8) deficient in a putative methylated adenine recognition and restriction

(mrr) gene, zmrr. The zmrr gene is a member of the type IV R-M system. It is possible that using KLD9 as a host strain for λ Red-mediated knockouts would improve the efficiency of the system in Z. mobilis, by decreasing the likelihood that the dsDNA cassette is degraded immediately upon entry into the cell. Further, strains deficient in the type I R-M genes identified by Kerr and coworkers (Kerr et al. 2011) may also serve as better Z. mobilis λ Red recombineering hosts.

More importantly, we also note that the λ Red Gamma protein is responsible for inhibiting the host RecBCD helicase-nuclease complex from degrading linear DNA

(Karu et al. 1975). Z. mobilis does not contain RecBCD and instead contains the functionally similar AddAB complex (Cromie 2009). The Gamma protein is specifically able to inhibit the RecB portion of RecBCD (Marsic et al. 1993). The AddA protein of the AddAB complex shares homology with RecB (Haijema et al. 1996). Additionally, chemical inhibitors of the RecBCD complex are also able to inhibit the AddAB complex

(Amundsen et al. 2012), and recombineering in other alternative hosts was shown to occur without a Gamma-equivalent function (Datta et al. 2008). Therefore, it was reasonable to believe that the λ Red Gamma protein might also display activity toward

AddA, or that the system would still function efficiently if it did not. Our results, however, suggest that this is likely not the case. Attempts to use the λ Red recombinase system in Bacillus subtilis, another bacterium that contains AddAB instead of RecBCD, were also met with minimal success (Wang et al. 2012). Therefore, an obvious step in using the λ Red system in Z. mobilis was to attempt the process in our addA knockout strain, KLD5. Unfortunately, however, attempts to do so were met with limited success due to the low viability of cells deficient in AddA. Therefore, adding small-molecule inhibitors of AddAB, such as those identified by Amundsen and coworkers (Amundsen et al. 2012), to the λ Red recombinase transformation mixture might be a viable alternative to removing addA from the Z. mobilis chromosome.

Although the λ Red expression system constructed here did not lead to knockout efficiencies in Z. mobilis that rival those seen in E. coli, it is quite possible that the system can be used to improve knockout and integration efficiencies in other bacterial hosts. In our design, the three λ Red proteins are expressed on a broad-host range vector, whose origin has been shown to be stably maintained in E. coli, Bordetella pertussis, Vibrio cholera, Rhizobium meliloti, and Pseudomonas putida, among others (Antoine and Locht

1992). On pKLD19, the proteins are expressed from an engineered lac promoter that has also been shown to function in a great diversity of bacteria (Khan et al. 2008).

Frequently, the λ Red proteins are lethal when expressed too highly (Katashkina et al.

2009; Sergueev et al. 2001). Because the lac promoter is inducible, the protein expression level can also be optimized for the host of interest using our system.

Efficiently removing antibiotic resistance cassettes from the Z. mobilis chromosome using yeast flippase is a very useful technique that came from this work. Z. mobilis is naturally resistant to many commonly used antibiotics, including ampicillin, gentamycin, and kanamycin, among others (Swings and De Ley 1977). Our experience shows that only three antibiotics can be reliably used in the strain: chloramphenicol, spectinomycin, and tetracycline. Another lab reports that even fewer antibiotics can be used for their strain of Z. mobilis (Kerr et al. 2011). This is an issue when multiple genetic modifications are to be made to a single chromosome; the number of modifications is limited to the number of available antibiotics, in general. However, when antibiotic resistance cassettes are removed from the chromosome with flippase, they can be recycled and used again for the construction of additional chromosomal modifications.

Further, the removal of antibiotic resistances is also very important if a strain is going to be used industrially (de Lorenzo 1992). Separately, Zou and coworkers also showed that flippase is highly functional in Z. mobilis, thereby confirming our positive results (Zou et al. 2012).

While the efficiency of λ Red-mediated chromosomal modifications using dsDNA is low in certain species of bacteria, attempts to modify genomes using ssDNA oligonucleotides have been met with greater rates of success. For example, the genome of

B. subtilis has been successfully modified with ssDNA and the λ Red Beta protein but not with dsDNA and the three λ Red proteins (Wang et al. 2012). In general, the efficiency of ssDNA incorporation into the chromosome is 10-100 times greater than the efficiency of

97 dsDNA incorporation (Ellis et al. 2001). Therefore, it was reasonable to believe that ssDNA recombineering would work in Z. mobilis. However, we were unable to provide evidence that this was the case. To attempt to explain this result, we note that the efficiency of ssDNA incorporation in E. coli was improved after endogenous nucleases were removed from the chromosome (Mosberg et al. 2012). This implies that expression of the λ Red Beta protein alone is not enough for total ssDNA evasion of host nucleases, even in E. coli. The low transformation efficiency as well as the to-date phage resistance of Z. mobilis suggests that host nucleases in this strain are highly active, making it even less likely that the λ Red Beta protein will be sufficient at protecting ssDNA until it is incorporated into the chromosome. To overcome this issue and improve host nuclease evasion, phosphorothioation of oligonucleotides has proven useful (Wang et al. 2011), and alternatives to the Beta protein have been used that are more efficient at protecting ssDNA in alternative hosts (Datta et al. 2008). Secondly, we also note that it is possible that 60 bp ssDNA oligonucleotides are too short for incorporation into the Z. mobilis chromosome. Previous research has shown that longer oligonucleotides are incorporated with higher efficiencies (DiCarlo et al. 2013a; Wang et al. 2009), and incorporation into the B. subtilis chromosome required single-stranded PCR products with two regions of homology (Wang et al. 2012). Attempting ssDNA recombineering in Z. mobilis with longer oligonucleotides may lead to positive results.

A system with even broader applicability compared to the λ Red recombinase system is the CRISPR/Cas adaptive immunity system. This system does not require the use of selective markers, and it has been shown to function in an even greater range of species than λ Red—Cas systems have been used successfully in bacteria, fungi, plants,

98 mice, and even human cells (DiCarlo et al. 2013b; Jiang et al. 2013; Mali et al. 2013;

Shan et al. 2013; Wang et al. 2013). The broad functionality of the CRISPR-associated system arises because the Cas proteins and small RNAs do not interact with specific host proteins, and instead only interact with DNA (Hwang et al. 2013). Here, we showed evidence that two small RNAs are able to direct Cas9 cleavage of the Z. mobilis chromosome. We also showed that the Cas system is likely able to mediate dsDNA crossover into the Z. mobilis chromosome. These results are promising for the future genome editing of Z. mobilis—nearly all genetic techniques can be improved through the action of Cas9. Specifically, gene knockouts, integrations, point mutations, inversions, and internal deletions and insertions can all be mediated with CRISPR/Cas genes. An important next step for the system in Z. mobilis is to show that the efficiency is high enough to allow for marker-free genome engineering. Combining the action of the λ Red proteins with the CRISPR/Cas system may prove useful for constructing these marker- free mutants (Jiang et al. 2013).

Collectively, our results provide a strong groundwork for the future engineering of Z. mobilis for the production of biofuels and other value-added products. We have attempted dsDNA λ Red recombineering, ssDNA λ Red recombineering, and

CRISPR/Cas-mediated genome engineering in the strain and have provided positive results for two of the three. It is imperative that Z. mobilis be made more amenable to genetic manipulation if it is ever to be accepted as an industrial host. The systems and results reported here can be used toward achieving this goal.

4.5 Tables and figures

Table 4.1 Bacterial strains and plasmids used in this study.

Genotype or relevant characteristics Source or reference Strain KLD4 ATCC 31821 ΔaddA::FRTSpRFRT This study KLD5 ATCC 31821 ΔaddA::FRT This study KLD6 ATCC 31821 ΔfliA::FRTSpRFRT This study KLD7 ATCC 31821 ΔmutS::FRTSpRFRT This study KLD8 ATCC 31821 Δxr::FRTSpRFRT This study KLD9 ATCC 31821 Δzmrr::SpR This study KLD10 ATCC 31821 ΔmutS::FRT This study KLD11 ATCC 31821 ΔmutS::FRT Δzmrr::CmR Y33Stop-SpR This study Plasmid pCas9 CmR, tracrRNA-cas9-dr, p15a ori (Jiang et al. 2013) pEZCm CmR, pACYC184-pZMO1 (Skerker and Shapiro 2000b) pJS14 CmR, pBBR1MCS-lacZα (Skerker and Shapiro 2000a) pJS71 SpR, pBBR1MCS-lacZα (Skerker and Shapiro 2000a) pKD3 AmpR, FRTCmRFRT, oriR6K (Datsenko and Wanner 2000) R pKD46 Amp , PBADgam-beta-exo, pSC101 ori(Ts) (Datsenko and Wanner 2000) pPROTet.E CmR, colE1 ori Clontech R q pSRKTc Tet , pBBR1MCS-3-lacI -Plac (Khan et al. 2008) pUC19 AmpR, lacZα, pMB1 ori New England Biolabs pKLD19 pSRKTc-gam-beta-exo This study pKLD20 pJS14-Ppdcflp This study q pKLD21 pJS14-lacI -Placflp This study pKLD22 pPROTet.E-addAUP-FRTSpRFRT-addADOWN This study pKLD23 pPROTet.E-fliAUP-FRTSpRFRT-fliADOWN This study pKLD24 pPROTet.E-mutSUP-FRTSpRFRT-mutSDOWN This study pKLD25 pPROTet.E-xrUP-FRTSpRFRT-xrDOWN This study pKLD26 pPROTet.E-zmrrUP-SpR-zmrrDOWN This study pKLD27 pPROTet.E-sacBUP This study pKLD28 pPROTet.E-sacBUP-sacCDOWN This study pKLD29 pPROTet.E-sacBUP-SpR-sacCDOWN This study pKLD30 pUC19- zmrrUP-SpR-zmrrDOWN This study pKLD31 pKD3, CmR Y33Stop This study pKLD32 pUC19- zmrrUP-CmR Y33Stop-SpR-zmrrDOWN This study pKLD33 pSRKTc-Ppdcgam-beta-exo This study pKLD34 pCas9-pZMO1 This study pKLD35 pSRKTc-tracrRNA-cas9-dr This study pKLD36 pEZCm-tracrRNA-dr This study pKLD37 pSRKTc-cas9 This study pKLD38 pEZCm-cas9 This study pKLD39 pJS71-cas9 This study pKLD40 pEZCm-tracrRNA-dr-zmrr spacer This study

100

Table 4.2 Oligonucleotide primers used in this study.

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD050F gam-beta-exo ATGCGAGCTCAAGGAGGTTATAAAAAATGGATATTAATACTGAAACT GAGA KD050R ATGCTCTAGATTCTTCGTCTGTTTCTACTGGTATTGGC

KD051F flp, for ATGTTAAAGGAGGTAAAGATAAAATGCCACAATTTGGTATATTATGT AAAACACCA q KD051R lacI -Plac ATGCCTCGAGTTATATGCGTCTATTTATGTAGGATGAAAGGTAGTCT

q KD052F lacI -Plac, ATGCGGTACCATTGAATTCTGATTGACACCATCGAATGG KD052R for flp ATTGTGGCATTTTATCTTTACCTCCTTTAACATATGATGCTGTTTCCTG TGTGAAATTGT

KD053F flp, for Ppdc TTTTGAATATATGGAGTAAGCAATGCCACAATTTGGTATATTATGTA AAACACCA KD053R ATGCCTCGAGTATATACATGAGAAGAACGGCATAGTGCG

KD054F Ppdc, for ATGCGGTACCTGCGGTATAATATCTGTAACAGCTCATTGA KD054R flp ATTGTGGCATTGCTTACTCCATATATTCAAAACACTATGTCTGA

KD055F FRTSpRFRT GAAGTTCCTATACTTTCTAGAGAATAGGAACTTCGGAATAGGAACTT CCGAACCACTTCATCCGGGGT KD055R GAAGTTCCTATTCCGAAGTTCCTATTCTCTAGAAAGTATAGGAACTTC CAATTATGTGCTTAGTGCATCTAACGCTT

KD056F SpR CGAACCACTTCATCCGGGGT KD056R CAATTATGTGCTTAGTGCATCTAACGCTT

KD057F pPROTet.E CCTAGGGCGTTCGGCTGC KD057R TCTAGGGCGGCGGATTTGT

KD058F addAUP AGCGCCAACCATGCCATC KD058R AGGATTCCCTTAATTCTGTGCCGA

KD059F fliAUP TTTGCAATTGTGACCTCGCCG KD059R TCACCATCTCCAAGCTGGAAAAGA

KD060F mutSUP GAAGTCAATATTCGCTTGGAGTGAATCTATAGC KD060R CGGTGTTATATTCCGCGATTTTGAAGC

KD061F xrUP CTATCAGTCGTCCCTTATATGGTCTGACG KD061R CAGAAACTCTCCACATTTTTTAAAAAAGACCCC

KD062F zmrrUP CGGATGCTGCTTTGGATTTGATTAGT KD062R TATCGAGCCTTTTACTTAAAATAATCTATCGAAAGAAAAAC

KD063F addADOWN TGATCGGATTTTCGTCATTGCATGAAA KD063R CAGAGTTTGTCGCCTTCAAGGC

KD064F fliADOWN AAAGATAATAAAAATATAAATTCAATAAAAATAAGTAGAATTGATC CAATATTAAAAAAG KD064R TTCAACATCAGGGGTATGTTGCATCA

KD065F mutSDOWN AAGAAAAATATATTCTCTGGCTTTTTCTCAATAATTTTTTGG KD065R TCCTTACCTGCTCGTCGGAATT

KD066F xrDOWN GAGCGGATATAAAACAAGCCGTCAC KD066R CTCCTTTTGGGTGCGGCG

101

Table 4.2 (cont.)

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD067F zmrrDOWN GCTTTTTGAAAAAGCCGTTTCCTGTATTG KD067R TTACCAGCATGTTTTATCAGGAATCGGA

KD068F pPROTet.E- ACAAATCCGCCGCCCTAGAAGCGCCAACCATGCCATC addAUP bridge KD068R addAUP- ATCGGCACAGAATTAAGGGAATCCTGAAGTTCCTATACTTTCTAGAG FRTSpRFRT AATAGGAACTTCG bridge

KD069F FRTSpRFRT- AGAGAATAGGAACTTCGGAATAGGAACTTCTGATCGGATTTTCGTCA addADOWN TTGCATGAAATG bridge KD069R addADOWN- GCCTTGAAGGCGACAAACTCTGCCTAGGGCGTTCGGCTGC pPROTet.E bridge

KD070F pPROTet.E- ACAAATCCGCCGCCCTAGATTTGCAATTGTGACCTCGCCG fliAUP bridge KD070R fliAUP- TCTTTTCCAGCTTGGAGATGGTGAGAAGTTCCTATACTTTCTAGAGAA FRTSpRFRT TAGGAACTTCG bridge

KD071F FRTSpRFRT- AGAGAATAGGAACTTCGGAATAGGAACTTCAAAGATAATAAAAATA fliADOWN bridge TAAATTCAATAAAA KD071R fliADOWN- TGATGCAACATACCCCTGATGTTGAACCTAGGGCGTTCGGCTGC pPROTet.E bridge

KD072F pPROTet.E- ACAAATCCGCCGCCCTAGAGAAGTCAATATTCGCTTGGAGTGAATCT mutSUP bridge ATAGC KD072R mutSUP- CTTCAAAATCGCGGAATATAACACCGGAAGTTCCTATACTTTCTAGA FRTSpRFRT GAATAGGAACTT bridge

KD073F FRTSpRFRT- AGAATAGGAACTTCGGAATAGGAACTTCAAGAAAAATATATTCTCTG mutSDOWN GCTTTTTCTCAAT bridge KD073R mutSDOWN- AATTCCGACGAGCAGGTAAGGACCTAGGGCGTTCGGCTGC pPROTet.E bridge

KD074F pPROTet.E-xrUP ACAAATCCGCCGCCCTAGACTATCAGTCGTCCCTTATATGGTCTGAC bridge G KD074R xrUP-SpR bridge GGGGTCTTTTTTAAAAAATGTGGAGAGTTTCTGCGAACCACTTCATCC GGGGT

KD075F SpR-xrDOWN AAGCGTTAGATGCACTAAGCACATAATTGGAGCGGATATAAAACAA bridge GCCGTCAC KD075R xrDOWN- CGCCGCACCCAAAAGGAGCCTAGGGCGTTCGGCTGC pPROTet.E bridge

KD076F pPROTet.E- ACAAATCCGCCGCCCTAGACGGATGCTGCTTTGGATTTGATTAGT zmrrUP bridge KD076R zmrrUP-SpR TTTTTCTTTCGATAGATTATTTTAAGTAAAAGGCTCGATACGAACCAC bridge TTCATCCGGGGT

102

Table 4.2 (cont.)

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD077F SpR-zmrrDOWN AAGCGTTAGATGCACTAAGCACATAATTGGCTTTTTGAAAAAGCCGT bridge TTCCTGTATTG KD077R zmrrDOWN- TCCGATTCCTGATAAAACATGCTGGTAACCTAGGGCGTTCGGCTGC pPROTet.E bridge

KD078F sacBUP ATGCGAATTCCAAGATTTTTGACAGACATCACCGAATGAT KD078R ATGCGGTACCAAGAACTATCCTAATAAATGATGAAATAAAGCCGAA AA

KD079F sacCDOWN ATGCGTCGACTTTTTGACGACAAAAAATTGCGCTGAAA KD079R ATGCGGATCCCTAAAAATCTGAGCGATTTTTCGGGTGT

KD080F SpR ATGCGGTACCCGAACCACTTCATCCGGGGT KD080R ATGCGTCGACCAATTATGTGCTTAGTGCATCTAACGCTT

KD081F pUC19 GGCGTAATCATGGTCATAGCTGTTTC KD081R ACTGGCCGTCGTTTTACAACGT

KD082F KpnI BamHI GGTACCGGATCCCGAACCACTTCATCCGGGGT KD082R SpR CAATTATGTGCTTAGTGCATCTAACGCTT

KD083F pUC19-zmrrUP ACGTTGTAAAACGACGGCCAGTCGGATGCTGCTTTGGATTTGATTAG bridge TG KD083R zmrrUP-KpnI TTTTTCTTTCGATAGATTATTTTAAGTAAAAGGCTCGATAGGTACCGG BamHI SpR ATCCCGAACCAC bridge

KD084F KpnI BamHI SpR- AAGCGTTAGATGCACTAAGCACATAATTGGCTTTTTGAAAAAGCCGT zmrrDOWN TTCCTGTATTG bridge KD084R zmrrDOWN- ATCCGATTCCTGATAAAACATGCTGGTAAGGCGTAATCATGGTCATA pUC19 bridge GCTGTTTC

KD085F pKD3 CmR TGTACCTAAAACCAGACCGTTCAGC KD085R Y33Stop TTGAGCAACTGACTGAAATGCCTCA

KD086F CmRY33Stop ATGCGGTACCTGAGACGTTGATCGGCACGTA KD086R ATGCGGATCCATTTAAATGGCGCGCCTTACGC

KD087F gam-beta-exo, for TGTTTTGAATATATGGAGTAAGCAATGGATATTAATACTGAAACTGA Ppdc GATCAAGCAAAAG KD087R ATGCCTCGAGTTCTTCGTCTGTTTCTACTGGTATTGGC

KD088F Ppdc, for ATGCGGTACCTGCGGTATAATATCTGTAACAGCTCATTGA KD088R gam-beta-exo TAATATCCATTGCTTACTCCATATATTCAAAACACTATGTCTGA

KD089F pZMO1 ATGCTCTAGACGCGAACGGACAAGCCC KD089R ATGCATCGATGGCAGGAGTGCAGGAAATAGCT

KD090F tracrRNA- ATGCTCTAGATATCTCTTCAAATGTAGCACCTGAAGTCAG KD090R cas9-dr ATGCCTCGAGGCGATGCTGTCGGAATGGAC

KD091F dr, for ATGCTCTAGACTCTTAATAAATGCAGTAATACAGGGGCTTTTC KD091R tracrRNA TAACTGTGATTAACCCTCTTTCTCAAGTTATCATCGGC

103

Table 4.2 (cont.)

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD092F tracrRNA, for dr TGATAACTTGAGAAAGAGGGTTAATCACAGTTAAATTGCTAACGCA GTCAG KD092R ATGCTCTAGATGATGATTTCAATAATAGTTTTAATGACCTCCGAAAT T

KD093F cas9 ATGCTCTAGAAAAAAGAAAAATTAAAGCATATTAAACTAATTTCGG AGGTC KD093R ATGCCTCGAGATACCATATTTTTAGTTATTAAGAAATAATCTTCATC TAAAATATACTTC

KD094F zmrr spacer AAACTGGTGGCTATGGGCTATGGCGGCAATCATGG KD094R AAAACCATGATTGCCGCCATAGCCCATAGCCACCA

KD095F ΔaddA::FRT TTTTGATGAATGGCGCAGTAATCCC SpRFRT KD095R check AGTATGATCTAATAGAAATACCATTTCATGCAATGACG

KD096F ΔfliA::FRT GCTGTCTGGATTTATTCCCCAGATCA SpRFRT KD096R check AAATAAATCTTTTGATAGATAATAGCTTTTTTAATATTGGATCAATT CTACT

KD097F ΔmutS::FRT ACTGTCAAAAAGACTAAAGCTGCAAAGAA SpRFRT KD097R check TCACTTATTCCAAAAAATTATTGAGAAAAAGCCAGA

KD098F Δxr::SpR TTAAAGGCGGATCAACAGGCCA KD098R check CGACAGTGACGGCTTGTTTTATATCC

KD099F Δzmrr::SpR ATTCTGTCGCCGGATAGGCG KD099R check GCCAGAATCCTAACAATACAGGAAACG

KD100F ΔsacBC::SpR CTTTTTTCGGCTTTATTTCATCATTTATTAGGATAGTTCTT KD100R check TTCAGCGCAATTTTTTGTCGTCAAAAA

KD101F CmRY33Stop GGCATTTCAGTCAGTTGCTCAATGTACCTATAACCAGACCGTTCAGC repair TGGATATTACGGC KD101R fliAL33Stop AAGAATGTCGAAGCGTTAATCAAAACCCATTAACAACTGGTTCGGA AAATTGCATGGCAT

KD102F frkE33Stop ATTGATTCTGACCGGAAGATGCTGGCTGTGTAACGTGTTCCGACCAC AACCCCTGAAGAA

104

Table 4.3 Z. mobilisλ Red-mediated knockout target genes.

Gene Description Protein significance ZMO1098 addA- double-strand break repair helicase RecB analogue ZMO0626 fliA- flagellar-specific sigma factor Regulates motility ZMO1907 mutS- DNA mismatch repair protein Reduces ssDNA recombineering efficiency ZMO0374-5 sacBC- levansucrases Required for sucrose utilization ZMO0976 xr- aldo/keto reductase Contributes to xylitol production ZMO0028 zmrr- restriction endonuclease Decreases transformation efficiency

Table 4.4 Frequency of ΔsacBC::SpR double crossover events in cells expressing pKLD19 with IPTG added (Induced) or without IPTG (Uninduced).

Percent of colonies tested displaying the correct genotype Induced 90 ± 2 Uninduced 19 ± 5

Table 4.5 Colony forming units obtained in the presence or absence of Cas9. Cells with pKLD39 (+Cas9) and without pKLD39 (-Cas9) were transformed with pKLD40. When Cas9 is present, the tracrRNA:crRNA complex directs it to cleave the chromosome, leading to cell death.

Colony forming units +Cas9 0 or 1 -Cas9 48 ± 11

105

Figure 4.1 Schematic representations of plasmids. a Plasmid used to express the λ Red recombinase system. b Plasmid used to express the tracrRNA:crRNA complex. c One of three plasmids used to express the Cas9 protein.

106

Figure 4.2 Single versus double crossover events. a When DNA is incorporated into the chromosome through a single crossover event, the entire circular vector is integrated into the chromosome. b When DNA is incorporated into the chromosome through a double crossover event, the gene of interest (zmrr) is removed from the chromosome, leading to a gene knockout.

107

Figure 4.3 PCR verification of knockouts. Control DNA yielded a PCR product the length of the wild-type allele. Desirable double knockout mutants yielded a PCR product the length of the spectinomycin resistance gene, 1.5 kb. Single crossover mutants yielded two PCR products, one the length of the wild-type allele and one the length of the spectinomycin resistance gene, 1.5 kb. Results shown are for the zmrr gene, whose wild- type PCR product is 1 kb in length.

Figure 4.4 Phenotypic verification of the Z. mobilis ΔfliA::FRTSpRFRT strain. The FliA protein is a transcriptional regulator of motility. When it is absent, cells are no longer motile. Motility loss can be detected using a simple plate assay. a Wild-type Z. mobilis control cells swim toward the edge of the plate. b Z. mobilis ΔfliA::FRTSpRFRT cells cannot swim.

108

Figure 4.5 Mechanism of the CRISPR/Cas9 system. The tracrRNA acts as a scaffold for the crRNA, which contains homology to the target protospacer sequence. Directly adjacent to the protospacer sequence on the chromosome is the NGG PAM sequence. The tracrRNA:crRNA complex directs Cas9 to create a double-stranded break in the DNA homologous to the protospacer sequence.

109

Figure 4.6 PCR verification of CRISPR/Cas9-mediated knockouts. Cells expressing pKLD37 (+Cas9) no longer contain the wild-type zmrr allele, confirmed by the lack of a PCR product at 1 kb. Cells lacking pKLD37 (-Cas9) are unsuccessful knockouts and still yield a 1 kb zmrr PCR product.

110

Chapter 5: Development of a promoter library for use in Z. mobilis

5.1 Introduction

In chapter 4, we attempted to improve upon the genetic techniques used for modifying the chromosome of Z. mobilis. In addition to chromosomal modifications, the efficient production of heterologous proteins will also be important if Z. mobilis is ever to be accepted as a host for alternative product synthesis. Efficient protein production requires the precise control of heterologous gene expression at the transcriptional level.

Currently, no set of Z. mobilis gene promoters of varying strengths exists for this purpose.

Typically, Z. mobilis native promoters, such as the gap, eno, and pdc promoters are used to express heterologous enzymes in this host (Deanda et al. 1996b; Zhang et al.

1995). These promoters are strong and constitutive, so expression from them is uncontrolled and not optimized. The inducible tac promoter, which is a hybrid of the trp and lac UV5 promoters, has been shown to function in Z. mobilis for the controlled expression of heterologous genes (de Boer et al. 1983). However, the system showed only partial control and led to a high level of basal expression (Arfman et al. 1992). In previous chapters, we successfully utilized an engineered lac promoter for protein expression in Z. mobilis. We saw tightly regulatable expression from this promoter, but the maximal expression level was considerably lower than what can be achieved with Z. mobilis native promoters. Therefore, a set of gene promoters of varying strengths designed for use in Z. mobilis would be a very useful tool set for the future engineering of the strain.

111

In this chapter, we used error-prone PCR to construct a large promoter library based on the strong and constitutive Z. mobilis pdc promoter. We then used flow cytometry to select the candidates displaying the highest level of single-cell homogeneity.

These stable candidates were then characterized using transcriptional fusions to the

Venus and mCherry fluorescent proteins, and transcriptional-level control was confirmed using qPCR. The final library candidates identified here are useful for pathway optimization and heterologous protein expression in Z. mobilis.

5.2 Materials and methods

5.2.1 Bacterial strains, media, and growth conditions

All experiments were performed in Zymomonas mobilis subsp. mobilis ZM4

(ATCC 31821). Liquid and solid rich medium (RM) with 10 g/L yeast extract, 2 g/L potassium phosphate, dibasic supplemented with 20 g/L glucose was used for all Z. mobilis experiments. All Z. mobilis cultures were grown at 30°C and antibiotics were used at the following concentrations: spectinomycin, 100 µg/mL; chloramphenicol, 100

µg/mL; and tetracycline, 25 µg/mL.

5.2.2 Plasmid construction

All strains and plasmids used in this study are listed in Table 5.1. All primers used in this study are listed in Table 5.2. To construct a plasmid for library development, the pEZCm vector was first amplified by PCR synthesis using primers KD103F and

KD103R such that it contained XbaI and BamHI restriction sites, the Z. mobilis gap ribosome binding site sequence, and the KpnI and XhoI restriction sites, respectively.

This PCR product was digested with BamHI and self-ligated to form plasmid pKLD40.

The venus fluorescent protein gene was amplified from plasmid pVenus by PCR

112 synthesis using primers KD104F and KD104R and inserted at the KpnI and XhoI sites of pKLD40 to form plasmid pKLD41. Separately, to construct a vector for additional library analyses, the mCherry fluorescent protein gene was amplified from plasmid pSZ by PCR synthesis using primers KD105F and KD105R and also inserted at the KpnI and XhoI sites of pKLD40 to form plasmid pKLD42.

To construct a vector containing the pdc promoter for use as an error-prone PCR template, the pdc promoter was amplified by PCR synthesis using primers KD106F and

KD106R from Z. mobilis ZM4 genomic DNA. This promoter was inserted into the plasmid pJS71 to yield plasmid pKLD43.

To construct vectors for pdc promoter analysis upstream of mCherry, promoters pdc 1-8 were amplified by PCR synthesis from vectors pKLD44-pKLD51, respectively, using primers KD107F and KD107R. The promoters were then inserted separately into plasmid pKLD42 to form plasmids pKLD52-pKLD59, respectively.

To construct vectors for testing inducible promoters in Z. mobilis, the venus fluorescent protein gene was amplified by PCR synthesis to include its ribosome binding site using primers KD108F and KD108R. The venus gene fragment was then inserted into pSRKTc downstream of the engineered lac promoter to form plasmid pKLD60.

Separately, the araC-ParaBAD region was amplified from E. coli MG1655 genomic DNA by PCR synthesis using primers KD109F and KD109R. The venus gene was amplified from pVenus by PCR synthesis using primers KD110F and KD110R. The araC-ParaBAD and venus gene fragements were then joined together by overlap extension PCR to yield the araC-ParaBADvenus gene fragment. This fragment was then inserted into plasmid pJS14 to yield plasmid pKLD61.

113

5.2.3 Library construction

A library of pdc promoter mutants was constructed using error-prone PCR.

Specifically, plasmid pKLD43 was used as template for nucleotide analogue mutagenesis in the presence of 20 μM 8-oxo-2’deoxyguanosine (8-oxo-dGTP) and 6-(2-deoxy-β-D- ribofuranosyl)-3,4-dihydro-8H-pyrimido-[4,5-c][1,2]oxazin-7-one (dPTP) (TriLink

Biotechnologies) (Zaccolo and Gherardi 1999) using primers KD111F and KD111R.

These primers bound to pKLD43 such that the entire 150 bp pdc promoter length could be mutagenized. Reactions were carried out in a 50 μL volume using GoTaq® master mix (Promega). An annealing temperature of 55ºC was used and 30 amplification cycles were performed. The PCR products were then cleaned and digested for insertion between the XbaI and BamHI sites of plasmid pKLD41. After ligation, the product was transformed into electrocompetent DH5α. The entire library of colonies was transferred to 50 mL liquid LB medium using a sterile loop and grown to mid-exponential phase, at which time 10 mL was withdrawn from the culture for plasmid isolation. The isolated library plasmid miniprep was then transformed into electrocompetent E. coli strain

GM119. The entire library was again transferred to 50 mL liquid LB medium using a sterile loop and grown to mid-exponential phase, at which time 10 mL was withdrawn from the culture for plasmid isolation and subsequent transformation into electrocompetent Z. mobilis.

5.2.4 Fluorescence measurements

Colonies containing library candidates were transferred to 1 mL RM or LB media in a single well of a polypropylene 2.2 mL 96-well deep microtiter plate (VWR) and grown with shaking at 600 rpm overnight. At late exponential phase, cells were

114 subcultured 1:100 in 1 mL RM or LB media in a single well of a polypropylene 2.2 mL

96-well deep microtiter plate and grown with shaking at 600 rpm for 18 h (Z. mobilis) or

8 h (E. coli), at which time 150 μL of culture was transferred to a single well of a black- walled, clear-bottomed Costar 96-well microtiter plate. For final analysis of the refined 8- member library, cells were grown in 2 mL RM in 5 mL glass tubes with shaking for 18 h instead of in a 96-well deep microtiter plate. The fluorescence (Venus Ex/Em: 515/528 nm, mCherry Ex/Em: 587/610 nm) and optical density (600 nm) were measured using a

Tecan Safire2 microplate reader. Fluorescence measurements are reported as the relative fluorescence normalized to the optical density of the sample to correct for variation in cell density. Fluorescences normalized to the optical density were then normalized to the fluorescence value obtained from the wild-type pdc promoter.

5.2.5 Flow cytometry experiments

A subset of 25 Z. mobilis library members were selected with a wide range of

Venus fluorescence levels for flow cytometry analysis, to ensure that bulk culture behavior accurately represented the fluorescence seen at the single-cell level. For a typical flow cytometry experiment, cells were grown overnight in RM containing chloramphenicol, then subcultured to an OD600nm of 0.05 and again grown overnight. At late-exponential phase, cells were harvested by centrifugation and washed once with

PBS. Cells were then resuspended in PBS and analyzed using a BD LSR II flow cytometer. Fluorescence values for approximately 500,000 cells were recorded using the fluorescein isothiocyanate (FITC) channel (excitation, 488 nm; emission, 530/30 nm).

The eight candidates with clean monovariate distributions were selected from the original

115

25 for further analysis. These candidates were designated promoters pdc 1-8, in vectors pKLD44-pKLD51, respectively.

5.2.6 Quantitative PCR (qPCR) reactions

Total RNA was isolated from Z. mobilis strains grown to mid-exponential phase using the Qiagen RNeasy Mini Kit according to the manufacturer’s instructions. cDNA was generated from RNA using the Qiagen QuantiTect Reverse Transcription Kit according to the manufacturer’s instructions. PCR amplification of cDNA using primers

KD112F, KD112R, KD045F, and KD045R (Dunn and Rao 2015), designed using the

Primer3Plus software (Untergasser et al. 2007), was visualized using the HotStart-IT

SYBR Green qPCR Master Mix with UDG (Affymetrix) and a Bio-Rad MiniOpticon

Real-Time PCR System. All results were quantified by PCR cycle-number comparison to standard curves prepared using E. coli strain AB273 or Z. mobilis ZM4 genomic DNA.

The Z. mobilis rrsA gene was used as a control for normalizing differences in total RNA quantities (He et al. 2012).

5.3 Results

5.3.1 Error-prone PCR leads to a library of promoters with varying strengths in Z. mobilis

It is important that gene expression levels be precisely controlled when producing heterologous proteins in a bacterial host. Inducible promoters are often useful for this purpose. We tested two inducible promoter systems in Z. mobilis—the lac promoter, inducible with IPTG, and the araBAD promoter, inducible with arabinose. Using transcriptional fusions to the venus fluorescent protein gene, we found that even at full induction these promoters led to only ~30% and ~5% of the activity of the wild-type pdc

116 promoter, respectively. Therefore, to cover the whole range of promoter strengths, we used nucleotide analogue mutagenesis to construct a library of promoters based on the strong and constitutive Z. mobilis pdc promoter. This approach is similar to the one used by Alper and coworkers (Alper et al. 2005). The promoters were cloned upstream of the venus fluorescent protein gene in a vector designed to facilitate library construction and analysis (Figure 5.1). The resulting promoter library led to a range of Venus fluorescence levels in both E. coli and Z. mobilis (Figure 5.2). While the library initially contained hundreds of members, we found that very few promoter variants led to consistent results in Z. mobilis. Cells initially displaying a very high level of fluorescence in Z. mobilis were particularly unstable. Through repeated fluorescence tests and by using flow cytometry to identify those candidates with homogeneous single-cell fluorescence distributions, the library was refined to contain 8 variants that led to reproducible results in Z. mobilis. The DNA sequences of these candidates can be found in Table 5.3. The promoter strengths of the 8 library members were found to span a 160-fold range when

Venus was used as the reporter (Figure 5.3).

5.3.2 Expression from the library of promoters is constitutive

To verify the constitutive nature of the 8 pdc promoter variants, each was cloned upstream of the mCherry fluorescent protein gene in the same library vector. Again, we found that the promoters led to reproducible fluorescence results and behavior consistent with that seen when venus was used as the reporter gene. It is important to note, however, that not all Z. mobilis cells will produce mCherry when the vector is present in the cells, for reasons unknown. Therefore, results shown are for colonies with confirmed

117 fluorescence. With mCherry, the promoter strengths of the 8 library members were found to span a 140-fold range (Figure 5.4).

5.3.3 Expression-level variation is due to transcriptional differences

To characterize the promoter library at the transcriptional level, we used quantitative PCR to measure the relative venus transcript levels in Z. mobilis cells expressing the 8 different library candidates. The mRNA levels (ng) spanned a 256-fold range, and normalized results showed a high level of correlation with fluorescence results

(Figure 5.5). This confirms that expression-level variation between the 8 library members is due to transcriptional differences.

5.4 Discussion

Expressing heterologous genes in Z. mobilis under the control of strong and constitutive native promoters is not always the best option. Therefore, we used error- prone PCR to construct a library of promoter variants based on the strong and constitutive

Z. mobilis pdc promoter. We found that the library led to a range of venus fluorescence levels in both E. coli and Z. mobilis. Transcriptional fusions of library members to the mCherry fluorescent protein led to similar results, confirming the constitutive nature of the promoters. qPCR results confirmed that fluorescence differences were due to control at the transcriptional level.

The promoter library constructed here is a very useful toolkit for heterologous protein production in Z. mobilis. While inducible promoters can also be used to control gene expression levels, the level of inducer is not homogeneous at the single-cell level, and adding inducer to large cultures can be expensive (Siegele and Hu 1997). Further, we have found that typical inducible promoters do not lead to high expression levels in Z.

118 mobilis, even at full induction. Although we only isolated one stable candidate with an expression level that is higher than that of the wild-type pdc promoter, this result has also been observed in other alternative hosts (Markley et al. 2015), and the wild-type pdc promoter expression level is already considerably high. Regardless, the promoter library is useful for determining the precise impact of gene dosage on a specific phenotype. For example, by cloning the E. coli transketolase and transaldolase genes under the control of the 8 different library members and then analyzing the effect each expression level has on arabinose fermentation by Z. mobilis, we can discern exactly how limiting each enzyme was in our unadapted arabinose-fermenting strain constructed in chapter 3.

As Z. mobilis gains popularity for alternative product synthesis, the ability to fine- tune protein expression levels will become increasingly important. The promoter library constructed here will be useful for this purpose. Should more promoter options be necessary, the initial library containing hundreds of candidates can be scanned again to isolate additional stable candidates. Further, error-prone PCR can be performed on other

Z. mobilis native promoters if needed. The entire library can also be scanned for variants that respond to distinct environmental conditions, if needed for a specific application. As a whole, our promoter library facilitates the process of constructing synthetic gene networks in Z. mobilis.

119

5.5 Tables and figures

Table 5.1 Strains and plasmids used in this study.

Genotype or relevant characteristics Source or reference Strain R AB273 E. coli MG1655 attλ::[kan PptsGvenus] Ahmet Badur Plasmid pEZCm CmR, pACYC184-pZMO1 Jeff Skerker pJS71 SpR, pBBR1MCS-lacZα (Skerker and Shapiro 2000a) R pSZ Cm , pPROTet.E-PLTet.0-mCherry Shuyan Zhang R q pSRKTc Tet , pBBR1MCS-3-lacI -Plac (Khan et al. 2008) pVenus KanR, MCS yfp(venus) t0 attλ, oriR6K (Saini et al. 2009) pKLD40 pEZCm-gapRBS This study pKLD41 pEZCm-gapRBSvenus This study pKLD42 pEZCm-gapRBSmCherry This study pKLD43 pJS71-Ppdc This study pKLD44 pEZCm-Ppdc1gapRBSvenus This study pKLD45 pEZCm-Ppdc2gapRBSvenus This study pKLD46 pEZCm-Ppdc3gapRBSvenus This study pKLD47 pEZCm-Ppdc4gapRBSvenus This study pKLD48 pEZCm-Ppdc5gapRBSvenus This study pKLD49 pEZCm-Ppdc6gapRBSvenus This study pKLD50 pEZCm-Ppdc7gapRBSvenus This study pKLD51 pEZCm-Ppdc8gapRBSvenus This study pKLD52 pEZCm-Ppdc1gapRBSmCherry This study pKLD53 pEZCm-Ppdc2gapRBSmCherry This study pKLD54 pEZCm-Ppdc3gapRBSmCherry This study pKLD55 pEZCm-Ppdc4gapRBSmCherry This study pKLD56 pEZCm-Ppdc5gapRBSmCherry This study pKLD57 pEZCm-Ppdc6gapRBSmCherry This study pKLD58 pEZCm-Ppdc7gapRBSmCherry This study pKLD59 pEZCm-Ppdc8gapRBSmCherry This study pKLD60 pSRKTc-venus This study pKLD61 pJS14-araC-ParaBADvenus This study

120

Table 5.2 Oligonucleotide primers used in this study.

Primer Primer use Forward and reverse sequences, respectively (5’-3’) number KD103F pEZCm, gap ATGCGGATCCATAAGTTAGGAGAATGGTACCCCTCGAGTATCGCGAA RBS CGGACAAGCCC KD103R ATGCGGATCCTCTAGAAATATTTTATCTGATTAATAAGATGATCTTCTT GAGATCG

KD104F Venus ATGCGGTACCATGAGTAAAGGAGAAGAACTTTTCACTGGA KD104R ATGCCTCGAGACCGAGCGTTCTGAACAAATCCA

KD105F mCherry ATGCGGTACCATGGTGAGCAAGGGCGAAGA KD105R ATGCCTCGAGTTACTTGTACAGCTCGTCCATGCC

KD106F Ppdc ATGCGGTACCTCTAGATAAAAAAGATAAAAAGTCTTTTCGCTTCGGCA KD106R ATGCCTCGAGGGATCCATATATTCAAAACACTATGTCTGAATCAGGAT GACC

KD107F Ppdc1-8 into CAAAACGATCTCAAGAAGATCATCTTATTAATCAGATAAAATATT KD107R pKLD42 CCTTTACTCATGGTACCATTCTCCTAACTTAT

KD108F Venus ATGCCTCGAGCTAACTAACTAAAGATTAAC KD108R ATGCGGTACCACCGAGCGTTCTGAACAAATCC

KD109F araC-ParaBAD, ATGCGTCGACGGGATTCTGCAAACCCTATGCTACT KD109R for Venus GCCAAAATCGAGGCCAATTGC

KD110F Venus, for GGCCTCGATTTTGGCATGAGTAAAGGAGAAGAACTTTTCACTGG KD110R araC-ParaBAD ATGCGAATTCACCGAGCGTTCTGAACAAATCC

KD111F Ppdc error GGAACAAAAGCTGGGTACCTCTAGA KD111R CGTCGACCTCGAGGGATCC

KD112F Venus qPCR GTCAGTGGAGAGGGTGAAGG KD112R TACATAACCTTCGGGCATGG

121

Table 5.3 DNA sequences of the wild-type pdc promoter and the 8 library candidates. Mutated oligonucleotides are shown in red.

Promoter Sequence (5’-3’) Ppdc TAAAAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCATCATGAACAAAAA TTCGGCATTTTTAAAAATGCCTATAGCTAAATCCGGAACGACACTTTAGAGGTTTCT GGGTCATCCTGATTCAGACATAGTGTTTTGAATATAT

Ppdc1 TAAAAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCCTCATGAACGAAAA TTCGGCATCTTTGGAAGTGCCTACAGCTGAATTCGGAACGACACTTCAGAGGTTCCT GGGTCATCCTGATTCAGACATAGTGTTTTGAGTATAT

Ppdc2 TAAAAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTCCATCATGAACGAAAA TTCGGCATCTTTAAAAGTGCCTATAGCTAGATCCGGAACGACACTTTAGAGGTTTCT GGGTCATCCTGATTCAGACATAGTGTTTTGAATATAT

Ppdc3 AAAAAAAGATAAAAAGTCTTTTCGCTCCGGCAGAAGAGACTCATCATGAACAAAAG TTCGGCATCTTTAAAAATGCCTATAGCTAAATCCGGAACGACACTTTAGAGGTTTCT GGGTCACCCTGATTCAAACATAGTGTTTTGAATATAT

Ppdc4 TAAGAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCATCACGAACAAAAA TTCGGCATCTTTAAAAATGCCTATAGCTAAATCCGGAACGACACTTTAGAGGTTTCT GGGCCGCCCTGATTCAGACATAGTGTTTTGAGTATAT

Ppdc5 TAAAAAAGATAAGAAGTCTTTTCGCTTCGGCAGAAGAGGTTCATCATGAGCAAAAA TTCGGCATTTTTAAAAATGCCTATAGCTAAATCCGGGACGACACTTTAGAGGTTTCT GGGTCATCCTGATTCAGACATAGTGTTTTGAATATAT

Ppdc6 TAAAAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCATCATGAACAAAAA TTCGGCATTCTTAAAAATGCCTATAGCTAAATCCGGAACGACACTTTAGAGGTTTCT GGGTCATCCTGATTCAGACATAGTGTTTTGAATATAT

Ppdc7 TAAAAAAGATAAAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCGTCATGAACAAAAA TTCGACATTTTTAAAAATGCCTATAGCTAAATCCGGAGCGACACTTTAGAGGTTTCT GGGTCATCCTGATTCAGACATAGTGTCTTGAATATAT

Ppdc8 TAAAAAGGATAGAAAGTCTTTTCGCTTCGGCAGAAGAGGTTCATCATGAACAAAAA TTCGGCATTTTTAAAAATGCCTATAGCTAAATCCGGAACGACACTTTAGAGGTCTCT GGGTCATCCTGATTCAGACATAGTGTTTTGAATATGT

122

Figure 5.1 Schematic representation of plasmid constructed for library characterization.

123

Figure 5.2 Venus fluorescence of library candidates before refinement. The library vectors led to a range of fluorescences in both E. coli (a) and Z. mobilis (b).

124

Figure 5.3 Venus fluorescence of the wild-type pdc promoter and the 8 stable library candidates in Z. mobilis.

125

Figure 5.4 mCherry fluorescence of the wild-type pdc promoter and the 8 stable library candidates in Z. mobilis.

126

Figure 5.5 Quantitative PCR results showing relative venus transcript levels in Z. mobilis.

127

Chapter 6: Conclusions and recommendations

6.1 Contributions

Z. mobilis has received considerable attention for its ability to produce the biofuel ethanol at near theoretical yields using unique metabolic pathways. In this work, we made improvements to Z. mobilis and Z. mobilis genetic techniques to render the strain a feasible industrial host, not only for the production of ethanol but for the production of other value-added products as well. In chapter 2, we used sugar transport engineering to improve the rate at which Z. mobilis produces ethanol from the five-carbon sugar xylose.

In chapter 3, we found several chromosomal mutations in xylose and arabinose- fermenting strains of Z. mobilis that improved the performance of these strains during growth on these sugars as sole carbon sources; we then thoroughly characterized several of these mutations to elucidate their individual contributions to the global phenotypes. In chapter 4, we constructed expression systems for the λ Red recombinase proteins and the bacterial CRISPR/Cas9 system for use in Z. mobilis. We proceeded to use these systems to facilitate genome editing in the strain. In chapter 5, we constructed a promoter library to be used for the precise control of gene expression levels in Z. mobilis, which is useful for heterologous protein production in the strain. Overall, we have contributed significantly to the advancement of Z. mobilis as an industrial workhorse. However, there is more work to be done.

6.2 Future directions

6.2.1 Cellulosic ethanol production

One of the primary limitations associated with Z. mobilis for bioethanol production is that the substrate range of this organism is limited to the six carbon sugars

128 glucose, fructose, and sometimes sucrose (Dawes et al. 1966). For ethanol production from cellulosic feedstocks, pentose sugar digestion is also important. While previous studies have developed engineered strains of Z. mobilis capable of fermenting the five carbon sugars xylose and arabinose (Deanda et al. 1996b; Zhang et al. 1995), and additional studies have improved upon these strains (Agrawal et al. 2011b; Agrawal et al.

2012b; Feldmann et al. 1992), pentose sugar digestion still lags far behind glucose consumption in Z. mobilis. Our work contributed greatly to determining why this is the case, but our engineered strains still do not consume xylose and arabinose as rapidly as they do glucose. Therefore, more work is needed to construct a strain of Z. mobilis that truly ferments pentoses and hexoses simultaneously. Toward this goal, further analysis of the mutations we found in adapted strains of Z. mobilis in chapter 3 may uncover additional metabolic pathways that inhibit pentose fermentations. Exploiting these pathways further may yield even greater improvements in the performance of Z. mobilis on these sugars. It is also possible that mutations found in efficient xylose-fermenting strains could improve arabinose fermentations, and vice versa. Moreover, we have determined in both chapters 2 and 3 that sugar transport is very important to both xylose and arabinose fermentations in Z. mobilis. Therefore, additional experiments focused on improving sugar transport should be done. Specifically, the mutant forms of the native Z. mobilis sugar transporter, Glf, that we discovered in chapter 3 can be combined with the wild-type Glf or the E. coli XylE in a single adapted strain to ensure that all types of sugars are being transported efficiently. Because xylose transport by XylE has been shown to be subject to a certain degree of “dead-end” glucose repression (Lam et al.

1980), mutant forms of XylE unable to bind glucose may be useful for expression in Z.

129 mobilis. Finally, because our xylose-fermenting strain still produced large amounts of the inhibiting by-product xylitol, and because removal of the associated aldo/keto reductase gene from the chromosome has been shown to improve xylose fermentation by Z. mobilis

(Agrawal et al. 2012b), a strain deficient in this gene but still containing the other adaptation-induced mutations should be constructed.

In addition to modifications that stem directly from our results, there are other modifications that can be made to Z. mobilis to improve it further for cellulosic ethanol production. The sugars mannose, galactose, and rhamnose are also common in cellulosic hydrolysates (Dwivedi et al. 2009). Therefore, to make Z. mobilis completely ideal for industrial fuel production, it should be engineered to ferment these sugars in addition to glucose, xylose, and arabinose. Adaptation can then be used to improve the growth of Z. mobilis on these sugars, and high-throughput sequencing can be used to reveal the genes implicated in the improved growth phenotypes. Additionally, Z. mobilis is known to be highly sensitive to the acetic acid that is present in lignocellulosic hydrolysates.

Adaptation on acetic acid has been performed to increase the strain’s tolerance to this inhibitor (Agrawal et al. 2012b), but a superior approach may be to engineer Z. mobilis to convert acetic acid into ethanol. Acetic acid consumption has proven successful in yeast, but it required coupling with an NADH-producing xylose utilization pathway to maintain redox balance (Wei et al. 2013). Because Z. mobilis displays the uncoupled growth phenomenon, it is possible that NADH and ATP are readily available for acetic acid consumption in wild-type cells. We constructed a vector to express the E. coli acetic acid consumption genes, acetyl-coenzyme A synthetase (acs) and acetaldehyde dehydrogenase (mhpF), in Z. mobilis, but the plasmid would not transform into the cells.

130

This suggests that redox balance was in fact an issue. However, we did utilize strong and constitutive Z. mobilis promoters for gene expression, so the expression level may have simply been too high for the proteins to only consume extra ATP and reducing equivalents that were not needed for growth. Further optimization of the protein levels using the promoter library we constructed in chapter 5 may lead to the successful consumption of acetic acid by Z. mobilis.

6.2.2 Genetic techniques

The highly active restriction-modification system in Z. mobilis has limited the development of new, efficient genetic techniques for use in this organism for many years.

We have attempted to overcome this issue by applying the λ Red recombinase system and the CRISPR/Cas bacterial adaptive immunity system for genome editing in Z. mobilis.

While our attempts were met with some success, there is still room for improvement. The

λ Red proteins increased the likelihood that DNA would be incorporated into the Z. mobilis chromosome through double crossover events, but the number of colony forming units we obtained upon completion of our knockout trials was still low. Optimizing protocol parameters only increased the likelihood of double crossover events and did not increase the number of colony forming units to the levels seen when these proteins are used in E. coli. We also were not able to utilize these proteins for the incorporation of ssDNA oligonucleotides into the Z. mobilis chromosome. On the other hand, our results using the CRISPR/Cas9 system are promising and may lead to the development of a high-efficiency knockout/integration protocol for Z. mobilis. Regardless, to ensure that all protocols developed in the future for genome editing in the strain are functional and efficient, the Z. mobilis restriction-modification system must be very thoroughly

131 investigated. Three genes of the type I and type IV R-M systems were identified and characterized by Kerr and coworkers, but a third plasmid-bound type I system has not been characterized at all (Kerr et al. 2011). The fact that no bacteriophage capable of attacking Z. mobilis has surfaced shows just how robust the organism’s R-M system really is. While this is an excellent reason to pursue Z. mobilis for use in future biorefineries, phage pose a huge threat to industrial processes currently carried out using other bacteria. Therefore, further investigation into the R-M proteins present in Z. mobilis may be useful for engineering phage-resistance into other industrial hosts.

Because the results we obtained using the CRISPR/Cas9 system were the most promising for Z. mobilis genome editing, future work should focus on improving this system. We expressed the Cas9 protein under the control of its native S. pyogenes promoter. We also constructed a vector to express Cas9 under the control of the strong and constitutive Z. mobilis pdc promoter, but have not yet analyzed this construct. It is well-known that the expression level of the Cas9 protein is very important to the functionality of the process in other hosts (Duda et al. 2014), so optimization of the Cas9 expression level in Z. mobilis is an important next step in improving the process.

Additionally, using the system for a marker-free knockout or integration will shed a lot of light on the true potential of this tool in Z. mobilis. Further, software should be developed to assist with the design of the Z. mobilis spacer sequence to reduce off-target effects.

This is especially important in bacteria, because Cas9 cleavage of the chromosome typically leads to cell death in these hosts. Another way to diminish the effects of off- target binding by Cas9 is to use a catalytically-inactive form of Cas9, known as dCas9, to carry out CRISPR interference (Qi et al. 2013). CRISPR interference occurs when dCas9

132 binds to a target gene without actually cleaving the chromosome. In this way, it blocks transcription and leads to gene silencing. A benefit of dCas9 over Cas9 is that the effects of dCas9 are reversible. We constructed an expression system for dCas9 in Z. mobilis, including a vector expressing dCas9 under the control of an inducible promoter, a strain of Z. mobilis with the venus fluorescent protein reporter integrated into the chromosome, and a vector containing the DNA coding for the gRNA complex and a spacer to target the promoter of the integrated venus gene. However, at the time of publication, we had not yet thoroughly tested the system in Z. mobilis. Future experiments will therefore reveal if dCas9 is a useful tool for genome regulation in the strain.

6.2.3 Alternative product synthesis

The interesting physiology displayed by Z. mobilis is worth manipulating for the production of alternative products. We worked to improve genetic techniques in the strain for this purpose, and our results on improving pentose sugar consumption by Z. mobilis can also be applied to alternative production processes. Central to the development of Z. mobilis strains capable of producing economical amounts of alternative products is shifting metabolic flux away from ethanol production. Specifically, the pyruvate decarboxylase gene must be silenced or removed from the chromosome. We attempted to remove pdc from the chromosome of a strain of Z. mobilis that we engineered to overproduce lactate for redox balance. However, our attempts were unsuccessful, even in pH-controlled media. We also attempted to use pdc antisense RNA to silence the gene in a lactate-overproducing strain, but no significant decrease in ethanol production was seen. It is clear that the successful development of a strain of Z. mobilis deficient in pdc would be extremely useful for alternative product synthesis—the high metabolic flux

133 could be directed toward a new, non-ethanol product. With current industry trends suggesting that cellulosic ethanol may not be the fuel of the future, it is likely that researchers will start focusing on the potential of Z. mobilis for the production of fatty acids, sorbitol, gluconate, fructose, butanol, and other value-added products. We hope that our work will be helpful in making an industrial production process utilizing Z. mobilis a reality.

134

References

Afendra AS, Drainas C (1987a) Expression and Stability of a Recombinant Plasmid in Zymomonas mobilis and Escherichia-Coli. Journal of General Microbiology 133:127-134 Afendra AS, Drainas C (1987b) Expression and stability of a recombinant plasmid in Zymomonas mobilis and Escherichia coli. J Gen Microbiol 133(1):127-34 Afendra AS, Vartholomatos G, Arvanitis N, Drainas C (1999) Characterization of the mobilization region of the Zymomonas mobilis ATCC10988 plasmid pZMO3. Plasmid 41(1):73-7 doi:S0147-619X(98)91374-9 [pii] 10.1006/plas.1998.1374 Agrawal M, Chen RR (2011) Discovery and characterization of a xylose reductase from Zymomonas mobilis ZM4. Biotechnology Letters 33(11):2127-33 doi:10.1007/s10529-011-0677-6 Agrawal M, Mao Z, Chen RR (2011a) Adaptation yields a highly efficient xylose- fermenting Zymomonas mobilis strain. Biotechnol Bioeng 108(4):777-85 doi:10.1002/bit.23021 Agrawal M, Wang Y, Chen RR (2012a) Engineering efficient xylose metabolism into an acetic acid-tolerant Zymomonas mobilis strain by introducing adaptation-induced mutations. Biotechnol Lett 34(10):1825-32 doi:10.1007/s10529-012-0970-z Akinterinwa O, Cirino PC (2009) Heterologous expression of D-xylulokinase from Pichia stipitis enables high levels of xylitol production by engineered Escherichia coli growing on xylose. Metab Eng 11(1):48-55 doi:10.1016/j.ymben.2008.07.006 S1096-7176(08)00046-3 [pii] Algar EM, Scopes RK (1985) Studies on Cell-Free Metabolism - Ethanol-Production by Extracts of Zymomonas mobilis. Journal of Biotechnology 2(5):275-287 doi:Doi 10.1016/0168-1656(85)90030-6 Alper H, Fischer C, Nevoigt E, Stephanopoulos G (2005) Tuning genetic control through promoter engineering. Proc Natl Acad Sci U S A 102(36):12678-83 doi:0504604102 [pii] 10.1073/pnas.0504604102 Amundsen SK, Spicer T, Karabulut AC, Londono LM, Eberhart C, Fernandez Vega V, Bannister TD, Hodder P, Smith GR (2012) Small-molecule inhibitors of bacterial AddAB and RecBCD helicase-nuclease DNA repair enzymes. ACS Chem Biol 7(5):879-91 doi:10.1021/cb300018x An H, Scopes RK, Rodriguez M, Keshav KF, Ingram LO (1991) Gel electrophoretic analysis of Zymomonas mobilis glycolytic and fermentative enzymes: identification of alcohol dehydrogenase II as a stress protein. J Bacteriol 173(19):5975-82 Antoine R, Locht C (1992) Isolation and molecular characterization of a novel broad- host-range plasmid from Bordetella bronchiseptica with sequence similarities to plasmids from gram-positive organisms. Mol Microbiol 6(13):1785-99 Arbuckle MI, Kane S, Porter LM, Seatter MJ, Gould GW (1996) Structure-function analysis of liver-type (GLUT2) and brain-type (GLUT3) glucose transporters: expression of chimeric transporters in Xenopus oocytes suggests an important role for putative transmembrane helix 7 in determining substrate selectivity. Biochemistry 35(51):16519-27 doi:10.1021/bi962210nbi962210n [pii]

135

Arfman N, Worrell V, Ingram LO (1992) Use of the tac promoter and lacIq for the controlled expression of Zymomonas mobilis fermentative genes in Escherichia coli and Zymomonas mobilis. J Bacteriol 174(22):7370-8 Armengaud J, Trapp J, Pible O, Geffard O, Chaumot A, Hartmann EM (2014) Non- model organisms, a species endangered by proteogenomics. J Proteomics doi:S1874-3919(14)00019-0 [pii]10.1016/j.jprot.2014.01.007 Arvanitis N, Pappas KM, Kolios G, Afendra AS, Typas MA, Drainas C (2000) Characterization and replication properties of the Zymomonas mobilis ATCC 10988 plasmids pZMO1 and pZMO2. Plasmid 44(2):127-37 doi:10.1006/plas.2000.1480S0147-619X(00)91480-X [pii] Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H (2006) Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol 2:2006 0008 doi:msb4100050 [pii]10.1038/msb4100050 Baratti JC, Bu'lock JD (1986) Zymomonas mobilis: a bacterium for ethanol production. Biotechnol Adv 4(1):95-115 doi:0734975086900066 [pii] Barrow KD, Collins JG, Rogers PL, Smith GM (1984) The structure of a novel polysaccharide isolated from Zymomonas mobilis determined by nuclear magnetic resonance spectroscopy. Eur J Biochem 145(1):173-9 Baumler DJ, Hung KF, Bose JL, Vykhodets BM, Cheng CM, Jeong KC, Kaspar CW (2006) Enhancement of acid tolerance in Zymomonas mobilis by a proton- buffering peptide. Appl Biochem Biotechnol 134(1):15-26 doi:ABAB:134:1:15 [pii] Bekers M, Vigants A, Laukevics J, Toma M, Rapoports A, Zikmanis P (2000) The effect of osmo-induced stress on product formation by Zymomonas mobilis on sucrose. Int J Food Microbiol 55(1-3):147-50 Belauich JP, Senez JC (1965) Influence of Aeration and of Pantothenate on Growth Yields of Zymomonas mobilis. J Bacteriol 89:1195-200 Benschoter AS, Ingram LO (1986) Thermal Tolerance of Zymomonas mobilis: Temperature-Induced Changes in Membrane Composition. Appl Environ Microbiol 51(6):1278-84 Bothast RJ, Nichols NN, Dien BS (1999) Fermentations with new recombinant organisms. Biotechnol Prog 15(5):867-75 doi:bp990087w [pii] 10.1021/bp990087w Bresticgoachet N, Gunasekaran P, Cami B, Baratti J (1990) Transfer and Expression of a Bacillus-Licheniformis Alpha-Amylase Gene in Zymomonas mobilis. Archives of Microbiology 153(3):219-225 doi:Doi 10.1007/Bf00249071 Bringer S, Sahm H, Swyzen W (1984) Ethanol-Production by Zymomonas mobilis and Its Application on an Industrial-Scale. Biotechnology and Bioengineering:311-319 Bringermeyer S, Sahm H (1989) Junctions of Catabolic and Anabolic Pathways in Zymomonas mobilis - Phosphoenolpyruvate Carboxylase and Malic Enzyme. Appl Microbiol Biot 31(5-6):529-536 Browne GM, Skotnicki ML, Goodman AE, Rogers PL (1984) Transformation of Zymomonas mobilis by a hybrid plasmid. Plasmid 12(3):211-4 Buchholz SE, Eveleigh DE (1986) Transfer of Plasmids to an Antibiotic-Sensitive Mutant of Zymomonas mobilis. Appl Environ Microb 52(2):366-370

136

Byun MOK, Kaper JB, Ingram LO (1986) Construction of a New Vector for the Expression of Foreign Genes in Zymomonas mobilis. Journal of Industrial Microbiology 1(1):9-15 doi:Doi 10.1007/Bf01569411 Carey VC, Walia SK, Ingram LO (1983) Expression of a Lactose Transposon (Tn951) in Zymomonas mobilis. Appl Environ Microbiol 46(5):1163-8 Carroll A, Somerville C (2009) Cellulosic biofuels. Annu Rev Plant Biol 60:165-82 doi:10.1146/annurev.arplant.043008.092125 Chedin F, Noirot P, Biaudet V, Ehrlich SD (1998a) A five-nucleotide sequence protects DNA from exonucleolytic degradation by AddAB, the RecBCD analogue of Bacillus subtilis. Mol Microbiol 29(6):1369-77 Chedin F, Noirot P, Biaudet V, Ehrlich SD (1998b) A five-nucleotide sequence protects DNA from exonucleolytic degradation by AddAB, the RecBCD analogue of Bacillus subtilis. Molecular Microbiology 29(6):1369-77 Chen RR, Wang Y, Shin HD, Agrawal M, Mao Z (2009) Improved strains of Zymomonas mobilis for fermentation of biomass. Google Patents Chen YF, Helmann JD (1992) Restoration of motility to an Escherichia coli fliA flagellar mutant by a Bacillus subtilis sigma factor. Proc Natl Acad Sci U S A 89(11):5123-7 Cherepanov PP, Wackernagel W (1995) Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic- resistance determinant. Gene 158(1):9-14 doi:037811199500193A [pii] Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, Hsu PD, Wu X, Jiang W, Marraffini LA, Zhang F (2013) Multiplex genome engineering using CRISPR/Cas systems. Science 339(6121):819-23 doi:10.1126/science.1231143 science.1231143 [pii] Conway T (1992) The Entner-Doudoroff pathway: history, physiology and molecular biology. FEMS Microbiol Rev 9(1):1-27 Conway T, Byun MO, Ingram LO (1987a) Expression Vector for Zymomonas mobilis. Appl Environ Microbiol 53(2):235-41 Conway T, Osman YA, Ingram LO (1987b) Gene expression in Zymomonas mobilis: promoter structure and identification of membrane anchor sequences forming functional lacZ' fusion proteins. J Bacteriol 169(6):2327-35 Conway T, Osman YA, Konnan JI, Hoffmann EM, Ingram LO (1987c) Promoter and nucleotide sequences of the Zymomonas mobilis pyruvate decarboxylase. J Bacteriol 169(3):949-54 Conway T, Sewell GW, Ingram LO (1987d) Glyceraldehyde-3-Phosphate Dehydrogenase Gene from Zymomonas mobilis - Cloning, Sequencing, and Identification of Promoter Region. Journal of Bacteriology 169(12):5653-5662 Cook GM, Russell JB (1994) Energy-spilling reactions of Streptococcus bovis and resistance of its membrane to proton conductance. Appl Environ Microbiol 60(6):1942-8 Copeland NG, Jenkins NA, Court DL (2001) Recombineering: a powerful new tool for mouse functional genomics. Nat Rev Genet 2(10):769-79 doi:10.1038/35093556 35093556 [pii]

137

Costantino N, Court DL (2003) Enhanced levels of lambda Red-mediated recombinants in mismatch repair mutants. Proc Natl Acad Sci U S A 100(26):15748-53 doi:10.1073/pnas.2434959100 2434959100 [pii] Cromie GA (2009) Phylogenetic ubiquity and shuffling of the bacterial RecBCD and AddAB recombination complexes. Journal of Bacteriology 191(16):5076-84 doi:10.1128/JB.00254-09 JB.00254-09 [pii] da Silva TL, Gouveia L, Reis A (2013) Integrated microbial processes for biofuels and high value-added products: the way to improve the cost effectiveness of biofuel production. Appl Microbiol Biotechnol doi:10.1007/s00253-013-5389-5 Dally EL, Stokes HW, Eveleigh DE (1982) A Genetic Comparison of Strains of Zymomonas mobilis by Analysis of Plasmid DNA. Biotechnology Letters 4(2):91- 96 doi:Doi 10.1007/Bf01091343 Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97(12):6640-5 doi:10.1073/pnas.120163297 120163297 [pii] Datta S, Costantino N, Zhou X, Court DL (2008) Identification and analysis of recombineering functions from Gram-negative and Gram-positive bacteria and their phages. Proceedings of the National Academy of Sciences 105(5):1626-1631 doi:10.1073/pnas.0709089105 Davis EO, Henderson PJ (1987) The cloning and DNA sequence of the gene xylE for xylose-proton symport in Escherichia coli K12. J Biol Chem 262(29):13928-32 Dawes EA, Large PJ (1970) Effect of starvation on the viability and cellular constituents of Zymomonas anaerobia and Zymomonas mobilis. J Gen Microbiol 60(1):31-42 Dawes EA, Ribbons DW, Large PJ (1966) The route of ethanol formation in Zymomonas mobilis. Biochem J 98(3):795-803 de Boer HA, Comstock LJ, Vasser M (1983) The tac promoter: a functional hybrid derived from the trp and lac promoters. Proc Natl Acad Sci U S A 80(1):21-5 De Graaf AA, Striegel K, Wittig RM, Laufer B, Schmitz G, Wiechert W, Sprenger GA, Sahm H (1999a) Metabolic state of Zymomonas mobilis in glucose-, fructose-, and xylose-fed continuous cultures as analysed by 13C- and 31P-NMR spectroscopy. Arch Microbiol 171(6):371-85 doi:91710371.203 [pii] de Kok S, Stanton LH, Slaby T, Durot M, Holmes VF, Patel KG, Platt D, Shapland EB, Serber Z, Dean J, Newman JD, Chandran SS (2014) Rapid and Reliable DNA Assembly via Ligase Cycling Reaction. Acs Synth Biol 3(2):97-106 doi:Doi 10.1021/Sb4001992 de Lorenzo V (1992) Genetic engineering strategies for environmental applications. Curr Opin Biotechnol 3(3):227-31 Deanda K, Zhang M, Eddy C, Picataggio S (1996a) Development of an arabinose- fermenting Zymomonas mobilis strain by metabolic pathway engineering. Appl Environ Microbiol 62(12):4465-70 Deley J, Swings J (1976) Phenotypic Description, Numerical-Analysis, and Proposal for an Improved Taxonomy and Nomenclature of Genus Zymomonas Kluyver and Van Niel 1936. Int J Syst Bacteriol 26(2):146-157 Delgado OD, Martinez MA, ABate CM, Sineriz F (2002) Chromosomal integration and expression of green fluorescent protein in Zymomonas mobilis. Biotechnology Letters 24(15):1285-1290 doi:Doi 10.1023/A:1016266110391

138

Demirbas MF (2009) Biorefineries for biofuel upgrading: A critical review. Appl Energ 86:S151-S161 doi:DOI 10.1016/j.apenergy.2009.04.043 Desai TA, Rao CV (2010) Regulation of arabinose and xylose metabolism in Escherichia coli. Appl Environ Microbiol 76(5):1524-32 doi:10.1128/AEM.01970-09 Deveau H, Barrangou R, Garneau JE, Labonte J, Fremaux C, Boyaval P, Romero DA, Horvath P, Moineau S (2008) Phage response to CRISPR-encoded resistance in Streptococcus thermophilus. J Bacteriol 190(4):1390-400 doi:JB.01412-07 [pii] 10.1128/JB.01412-07 DiCarlo JE, Conley AJ, Penttila M, Jantti J, Wang HH, Church GM (2013a) Yeast oligo- mediated genome engineering (YOGE). Acs Synth Biol 2(12):741-9 doi:10.1021/sb400117c DiCarlo JE, Norville JE, Mali P, Rios X, Aach J, Church GM (2013b) Genome engineering in Saccharomyces cerevisiae using CRISPR-Cas systems. Nucleic Acids Res 41(7):4336-43 doi:10.1093/nar/gkt135 gkt135 [pii] Dien BS, Cotta MA, Jeffries TW (2003) Bacteria engineered for fuel ethanol production: current status. Appl Microbiol Biotechnol 63(3):258-66 doi:10.1007/s00253-003- 1444-y Dimarco AA, Romano AH (1985) d-Glucose Transport System of Zymomonas mobilis. Appl Environ Microbiol 49(1):151-7 Doelle HW, Greenfield PF (1985) The Production of Ethanol from Sucrose Using Zymomonas mobilis. Appl Microbiol Biot 22(6):405-410 Doelle MB, Doelle HW (1990) Sugarcane Molasses Fermentation by Zymomonas mobilis. Appl Microbiol Biot 33(1):31-35 Doelle MB, Greenfield PF, Doelle HW (1990) The Relationship between Sucrose Hydrolysis, Sorbitol Formation and Mineral Ion Concentration during Bioethanol Formation Using Zymomonas mobilis 2716. Appl Microbiol Biot 34(2):160-167 Dominguez JM, Gong CS, Tsao GT (1996) Pretreatment of sugar cane bagasse hemicellulose hydrolysate for xylitol production by yeast. Appl Biochem Biotechnol 57-58:49-56 Dong HW, Bao J, Ryu DD, Zhong JJ (2011) Design and construction of improved new vectors for Zymomonas mobilis recombinants. Biotechnol Bioeng 108(7):1616-27 doi:10.1002/bit.23106 Drainas C, Slater AA, Coggins L, Montague P, Costa RG, Ledingham WM, Kinghorn JR (1983) Electron-Microscopic Analysis of Zymomonas mobilis, Strain Atcc-10988 Plasmid DNA. Biotechnology Letters 5(6):405-408 doi:Doi 10.1007/Bf00131281 Du J, Li S, Zhao H (2010a) Discovery and characterization of novel d-xylose-specific transporters from Neurospora crassa and Pichia stipitis. Mol Biosyst 6(11):2150- 6 doi:10.1039/c0mb00007h Duda K, Lonowski LA, Kofoed-Nielsen M, Ibarra A, Delay CM, Kang Q, Yang Z, Pruett-Miller SM, Bennett EP, Wandall HH, Davis GD, Hansen SH, Frodin M (2014) High-efficiency genome editing via 2A-coupled co-expression of fluorescent proteins and zinc finger nucleases or CRISPR/Cas9 nickase pairs. Nucleic Acids Res 42(10):e84 doi:10.1093/nar/gku251 gku251 [pii] Dunn KL, Rao CV (2014) Expression of a xylose-specific transporter improves ethanol production by metabolically engineered Zymomonas mobilis. Appl Microbiol Biotechnol doi:10.1007/s00253-014-5812-6

139

Dunn KL, Rao CV (2015) High-throughput sequencing reveals adaptation-induced mutations in pentose-fermenting strains of Zymomonas mobilis. Biotechnol Bioeng doi:10.1002/bit.25631 Dwivedi P, Alavalapati JRR, Lal P (2009) Cellulosic ethanol production in the United States: Conversion technologies, current production status, economics, and emerging developments. Energy Sustain Dev 13(3):174-182 doi:DOI 10.1016/j.esd.2009.06.003 Ellis HM, Yu D, DiTizio T, Court DL (2001) High efficiency mutagenesis, repair, and engineering of chromosomal DNA using single-stranded oligonucleotides. Proc Natl Acad Sci U S A 98(12):6742-6 doi:10.1073/pnas.121164898 121164898 [pii] Eriksen DT, Hsieh PC, Lynn P, Zhao H (2013) Directed evolution of a cellobiose utilization pathway in Saccharomyces cerevisiae by simultaneously engineering multiple proteins. Microb Cell Fact 12:61 doi:10.1186/1475-2859-12-61 1475- 2859-12-61 [pii] Feldmann SD, Sahm H, Sprenger GA (1992) Pentose Metabolism in Zymomonas mobilis Wild-Type and Recombinant Strains. Appl Microbiol Biot 38(3):354-361 Fieschko J, Humphrey RE (1983) Effects of temperature and ethanol concentration on the maintenance and yield coefficient of Zymomonas mobilis. Biotechnol Bioeng 25(6):1655-60 doi:10.1002/bit.260250618 Forrest WW (1967) Energies of activation and uncoupled growth in Streptococcus faecalis and Zymomonas mobilis. J Bacteriol 94(5):1459-63 Gao Q, Zhang M, McMillan JD, Kompala DS (2002) Characterization of heterologous and native enzyme activity profiles in metabolically engineered Zymomonas mobilis strains during batch fermentation of glucose and xylose mixtures. Appl Biochem Biotechnol 98-100:341-55 Gibbs M, Demoss RD (1954) Anaerobic dissimilation of C14-labeled glucose and fructose by Pseudomonas lindneri. J Biol Chem 207(2):689-94 Goetz G, Iwan P, Hauer B, Breuer M, Pohl M (2001) Continuous production of (R)- phenylacetylcarbinol in an enzyme-membrane reactor using a potent mutant of pyruvate decarboxylase from Zymomonas mobilis. Biotechnology and Bioengineering 74(4):317-325 doi:Doi 10.1002/Bit.1122 Goldman D, Lavid N, Schwartz A, Shoham G, Danino D, Shoham Y (2008) Two active forms of Zymomonas mobilis levansucrase. An ordered microfibril structure of the enzyme promotes levan polymerization. J Biol Chem 283(47):32209-17 doi:10.1074/jbc.M805985200 M805985200 [pii] Gudbergsdottir S, Deng L, Chen Z, Jensen JV, Jensen LR, She Q, Garrett RA (2011) Dynamic properties of the Sulfolobus CRISPR/Cas and CRISPR/Cmr systems when challenged with vector-borne viral and plasmid genes and protospacers. Mol Microbiol 79(1):35-49 doi:10.1111/j.1365-2958.2010.07452.x Gunasekaran P, Karunakaran T, Cami B, Mukundan AG, Preziosi L, Baratti J (1990) Cloning and sequencing of the sacA gene: characterization of a sucrase from Zymomonas mobilis. J Bacteriol 172(12):6727-35 Haijema BJ, Venema G, Kooistra J (1996) The C terminus of the AddA subunit of the Bacillus subtilis ATP-dependent DNase is required for the ATP-dependent exonuclease activity but not for the helicase activity. J Bacteriol 178(17):5086-91

140

Harms K, Wackernagel W (2008) The RecBCD and SbcCD DNases suppress homology- facilitated illegitimate recombination during natural transformation of Acinetobacter baylyi. Microbiol-Sgm 154:2437-2445 doi:DOI 10.1099/mic.0.2008/018382-0 Hashiramoto M, Kadowaki T, Clark AE, Muraoka A, Momomura K, Sakura H, Tobe K, Akanuma Y, Yazaki Y, Holman GD, et al. (1992) Site-directed mutagenesis of GLUT1 in helix 7 residue 282 results in perturbation of exofacial ligand binding. J Biol Chem 267(25):17502-7 Hayashi T, Kato T, Furukawa K (2012) Respiratory chain analysis of Zymomonas mobilis mutants producing high levels of ethanol. Appl Environ Microbiol 78(16):5622-9 doi:10.1128/AEM.00733-12 AEM.00733-12 [pii] He MX, Wu B, Qin H, Ruan ZY, Tan FR, Wang JL, Shui ZX, Dai LC, Zhu QL, Pan K, Tang XY, Wang WG, Hu QC (2014) Zymomonas mobilis: a novel platform for future biorefineries. Biotechnology for Biofuels 7 doi:Artn 101Doi 10.1186/1754- 6834-7-101 He MX, Wu B, Shui ZX, Hu QC, Wang WG, Tan FR, Tang XY, Zhu QL, Pan K, Li Q, Su XH (2012) Transcriptome profiling of Zymomonas mobilis under ethanol stress. Biotechnol Biofuels 5(1):75 doi:10.1186/1754-6834-5-75 1754-6834-5-75 [pii] Horton RM, Hunt HD, Ho SN, Pullen JK, Pease LR (1989) Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77(1):61-8 doi:0378-1119(89)90359-4 [pii] Huang C, Zong MH, Wu H, Liu QP (2009) Microbial oil production from rice straw hydrolysate by Trichosporon fermentans. Bioresour Technol 100(19):4535-8 doi:10.1016/j.biortech.2009.04.022 S0960-8524(09)00387-3 [pii] Hwang WY, Fu Y, Reyon D, Maeder ML, Tsai SQ, Sander JD, Peterson RT, Yeh JR, Joung JK (2013) Efficient genome editing in zebrafish using a CRISPR-Cas system. Nat Biotechnol 31(3):227-9 doi:10.1038/nbt.2501 nbt.2501 [pii] Ishino Y, Shinagawa H, Makino K, Amemura M, Nakata A (1987) Nucleotide sequence of the iap gene, responsible for alkaline phosphatase isozyme conversion in Escherichia coli, and identification of the gene product. J Bacteriol 169(12):5429- 33 Iwan P, Goetz G, Schmitz S, Hauer B, Breuer M, Pohl M (2001) Studies on the continuous production of (R)-(-)-phenylacetylcarbinol in an enzyme-membrane reactor. J Mol Catal B-Enzym 11(4-6):387-396 doi:Doi 10.1016/S1381- 1177(00)00029-1 Jeon YJ, Svenson CJ, Joachimsthal EL, Rogers PL (2002) Kinetic analysis of ethanol production by an acetate-resistant strain of recombinant Zymomonas mobilis. Biotechnology Letters 24(10):819-824 doi:Doi 10.1023/A:1015546521000 Jeon YJ, Svenson CJ, Rogers PL (2005a) Over-expression of xylulokinase in a xylose- metabolising recombinant strain of Zymomonas mobilis. Fems Microbiology Letters 244(1):85-92 doi:DOI 10.1016/j.femsle.2005.01.025 Jeon YJ, Svenson CJ, Rogers PL (2005b) Over-expression of xylulokinase in a xylose- metabolising recombinant strain of Zymomonas mobilis. FEMS Microbiol Lett 244(1):85-92 doi:S0378-1097(05)00039-X [pii] 10.1016/j.femsle.2005.01.025

141

Jiang W, Bikard D, Cox D, Zhang F, Marraffini LA (2013) RNA-guided editing of bacterial genomes using CRISPR-Cas systems. Nat Biotechnol 31(3):233-9 doi:10.1038/nbt.2508 nbt.2508 [pii] Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E (2012) A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337(6096):816-21 doi:10.1126/science.1225829 science.1225829 [pii] Joachimsthal EL, Rogers PL (2000a) Characterization of a high-productivity recombinant strain of Zymomonas mobilis for ethanol production from glucose/xylose mixtures. Appl Biochem Biotechnol 84-86:343-56 Joachimsthal EL, Rogers PL (2000b) Characterization of a high-productivity recombinant strain of Zymomonas mobilis for ethanol production from glucose/xylose mixtures. Appl Biochem Biotechnol 84-86:343-56 Jobses IM, Egberts GT, van Baalen A, Roels JA (1985) Mathematical modelling of growth and substrate conversion of Zymomonas mobilis at 30 and 35 degrees C. Biotechnol Bioeng 27(7):984-95 doi:10.1002/bit.260270709 Johns MR, Greenfield PF, Doelle HW (1991) Byproducts from Zymomonas mobilis Bioreactor Systems and Effects. Advances in Biochemical Engineering/Biotechnology, vol 44. Springer Berlin Heidelberg, pp 97-121 Kalnenieks U (2006) Physiology of Zymomonas mobilis: some unanswered questions. Adv Microb Physiol 51:73-117 doi:S0065-2911(06)51002-1 [pii] 10.1016/S0065- 2911(06)51002-1 Kalnenieks U, Degraaf AA, Bringermeyer S, Sahm H (1993) Oxidative-Phosphorylation in Zymomonas mobilis. Archives of Microbiology 160(1):74-79 Kalnenieks U, Galinina N, Bringer-Meyer S, Poole RK (1998) Membrane D-lactate oxidase in Zymomonas mobilis: evidence for a branched respiratory chain. FEMS Microbiol Lett 168(1):91-7 doi:S0378-1097(98)00424-8 [pii] Kalnenieks U, Galinina N, Strazdina I, Kravale Z, Pickford JL, Rutkis R, Poole RK (2008) NADH dehydrogenase deficiency results in low respiration rate and improved aerobic growth of Zymomonas mobilis. Microbiology 154(Pt 3):989-94 doi:10.1099/mic.0.2007/012682-0 154/3/989 [pii] Kalnenieks U, Galinina N, Toma MM, Poole RK (2000) Cyanide inhibits respiration yet stimulates aerobic growth of Zymomonas mobilis. Microbiology 146 ( Pt 6):1259- 66 Kalnenieks UZ, Pankova LM, Shvinka YE (1987) Proton Motive Force in the Bacterium Zymomonas mobilis. Biochemistry-Moscow+ 52(5):617-620 Kannan R, Mukundan G, Ait-Abdelkader N, Augier-Magro V, Baratti J, Gunasekaran P (1995a) Molecular cloning and characterization of the extracellular sucrase gene (sacC) of Zymomonas mobilis. Arch Microbiol 163(3):195-204 Kannan R, Pitchaimani K, Gunasekaran P, Ait-Abdelkader N, Baratti J (1995b) Overexpression of extracellular sucrase (SacC) of Zymomonas mobilis in Escherichia coli. FEMS Microbiol Lett 133(1-2):29-33 Karu AE, Sakaki Y, Echols H, Linn S (1975) The gamma protein specified by bacteriophage gamma. Structure and inhibitory activity for the recBC enzyme of Escherichia coli. J Biol Chem 250(18):7377-87

142

Katashkina JI, Hara Y, Golubeva LI, Andreeva IG, Kuvaeva TM, Mashko SV (2009) Use of the lambda Red-recombineering method for genetic engineering of Pantoea ananatis. BMC Mol Biol 10:34 doi:10.1186/1471-2199-10-34 1471-2199-10-34 [pii] Kerr AL, Jeon YJ, Svenson CJ, Rogers PL, Neilan BA (2011) DNA restriction- modification systems in the ethanologen, Zymomonas mobilis ZM4. Appl Microbiol Biotechnol 89(3):761-9 doi:10.1007/s00253-010-2936-1 Khan SR, Gaines J, Roop RM, 2nd, Farrand SK (2008) Broad-host-range expression vectors with tightly regulated promoters and their use to examine the influence of TraR and TraM expression on Ti plasmid quorum sensing. Appl Environ Microbiol 74(16):5053-62 doi:10.1128/AEM.01098-08 AEM.01098-08 [pii] Kim IS, Barrow KD, Rogers PL (2000a) Kinetic and nuclear magnetic resonance studies of xylose metabolism by recombinant Zymomonas mobilis ZM4(pZB5). Appl Environ Microbiol 66(1):186-93 Kim IS, Barrow KD, Rogers PL (2000c) Nuclear magnetic resonance studies of acetic acid inhibition of rec Zymomonas mobilis ZM4(pZB5). Appl Biochem Biotechnol 84-86:357-70 Kita K, Konishi K, Anraku Y (1984a) Terminal oxidases of Escherichia coli aerobic respiratory chain. I. Purification and properties of cytochrome b562-o complex from cells in the early exponential phase of aerobic growth. J Biol Chem 259(5):3368-74 Kita K, Konishi K, Anraku Y (1984b) Terminal oxidases of Escherichia coli aerobic respiratory chain. II. Purification and properties of cytochrome b558-d complex from cells grown with limited oxygen and evidence of branched electron-carrying systems. J Biol Chem 259(5):3375-81 Kremer TA, LaSarre B, Posto AL, McKinlay JB (2015) N-2 gas is an effective fertilizer for bioethanol production by Zymomonas mobilis. P Natl Acad Sci USA 112(7):2222-2226 doi:DOI 10.1073/pnas.1420663112 Kupsch J, Alonso JC, Trautner TA (1989) Analysis of Structural and Biological Parameters Affecting Plasmid Deletion Formation in Bacillus subtilis. Molecular & General Genetics 218(3):402-408 doi:Doi 10.1007/Bf00332402 Kuyper M, Toirkens MJ, Diderich JA, Winkler AA, van Dijken JP, Pronk JT (2005) Evolutionary engineering of mixed-sugar utilization by a xylose-fermenting Saccharomyces cerevisiae strain. Fems Yeast Res 5(10):925-934 doi:DOI 10.1016/j.femsyr.2005.04.004 Kyono K, Yanase H, Tonomura K, Kawasaki H, Sakai T (1995) Cloning and characterization of Zymomonas mobilis genes encoding extracellular levansucrase and invertase. Biosci Biotechnol Biochem 59(2):289-93 Lam VMS, Daruwalla KR, Henderson PJF, Jonesmortimer MC (1980) Proton-Linked D- Xylose Transport in Escherichia coli. Journal of Bacteriology 143(1):396-402 Lawford HG, Rousseau JD (1999) The effect of glucose on high-level xylose fermentations by recombinant Zymomonas in batch and fed-batch fermentations. Appl Biochem Biotech 77-9:235-249 doi:Doi 10.1385/Abab:77:1-3:235 Lawford HG, Rousseau JD (2000a) Comparative energetics of glucose and xylose metabolism in recombinant Zymomonas mobilis. Appl Biochem Biotechnol 84- 86:277-93

143

Lawford HG, Rousseau JD (2000b) Comparative energetics of glucose and xylose metabolism in recombinant Zymomonas mobilis. Appl Biochem Biotech 84- 6:277-293 doi:Doi 10.1385/Abab:84-86:1-9:277 Lawford HG, Rousseau JD (2002) Performance testing of Zymomonas mobilis metabolically engineered for cofermentation of glucose, xylose, and arabinose. Appl Biochem Biotechnol 98-100:429-48 Lawford HG, Rousseau JD, Mohagheghi A, McMillan JD (1998) Continuous culture studies of xylose-fermenting Zymomonas mobilis. Appl Biochem Biotechnol 70- 72:353-67 doi:10.1007/BF02920151 Lazdunsk.A, Belaich JP (1972) Uncoupling in Bacterial Growth - Atp Pool Variation in Zymomonas mobilis Cells in Relation to Different Uncoupling Conditions of Growth. Journal of General Microbiology 70(Apr):187-& Leandro MJ, Goncalves P, Spencer-Martins I (2006a) Two glucose/xylose transporter genes from the yeast Candida intermedia: first molecular characterization of a yeast xylose-H+ symporter. Biochem J 395(3):543-9 doi:10.1042/BJ20051465 Lee JY, Cheong DE, Kim GJ (2008) A novel assay system for the measurement of transketolase activity using xylulokinase from Saccharomyces cerevisiae. Biotechnol Lett 30(5):899-904 doi:10.1007/s10529-007-9616-y Lemuth K, Steuer K, Albermann C (2011) Engineering of a plasmid-free Escherichia coli strain for improved in vivo biosynthesis of astaxanthin. Microb Cell Fact 10:29 doi:10.1186/1475-2859-10-29 1475-2859-10-29 [pii] Leonelli S, Ankeny RA (2013) What makes a model organism? Endeavour 37(4):209-12 doi:10.1016/j.endeavour.2013.06.001 S0160-9327(13)00037-9 [pii] Lesic B, Rahme LG (2008) Use of the lambda Red recombinase system to rapidly generate mutants in Pseudomonas aeruginosa. BMC Mol Biol 9:20 doi:10.1186/1471-2199-9-20 1471-2199-9-20 [pii] Liu C, Dong H, Zhong J, Ryu DD, Bao J (2010) Sorbitol production using recombinant Zymomonas mobilis strain. J Biotechnol 148(2-3):105-12 doi:10.1016/j.jbiotec.2010.04.008 S0168-1656(10)00193-8 [pii] Liu CQ, Goodman AE, Dunn NW (1988) Expression of Cloned Xanthomonas D-Xylose Catabolic Genes in Zymomonas mobilis. Journal of Biotechnology 7(1):61-70 doi:Doi 10.1016/0168-1656(88)90035-1 Loos H, Kramer R, Sahm H, Sprenger GA (1994) Sorbitol promotes growth of Zymomonas mobilis in environments with high concentrations of sugar: evidence for a physiological function of glucose-fructose oxidoreductase in osmoprotection. J Bacteriol 176(24):7688-93 Lynd LR, Cushman JH, Nichols RJ, Wyman CE (1991) Fuel ethanol from cellulosic biomass. Science 251(4999):1318-23 doi:251/4999/1318 [pii] 10.1126/science.251.4999.1318 Makarova KS, Haft DH, Barrangou R, Brouns SJ, Charpentier E, Horvath P, Moineau S, Mojica FJ, Wolf YI, Yakunin AF, van der Oost J, Koonin EV (2011) Evolution and classification of the CRISPR-Cas systems. Nat Rev Microbiol 9(6):467-77 doi:10.1038/nrmicro2577 nrmicro2577 [pii] Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, Norville JE, Church GM (2013) RNA-guided human genome engineering via Cas9. Science 339(6121):823-6 doi:10.1126/science.1232033 science.1232033 [pii]

144

Markley AL, Begemann MB, Clarke RE, Gordon GC, Pfleger BF (2015) Synthetic Biology Toolbox for Controlling Gene Expression in the Cyanobacterium Synechococcus sp. strain PCC 7002. Acs Synth Biol 4(5):595-603 doi:10.1021/sb500260k Marraffini LA, Sontheimer EJ (2010) Self versus non-self discrimination during CRISPR RNA-directed immunity. Nature 463(7280):568-71 doi:10.1038/nature08703 nature08703 [pii] Marsic N, Roje S, Stojiljkovic I, Salaj-Smic E, Trgovcevic Z (1993) In vivo studies on the interaction of RecBCD enzyme and lambda Gam protein. J Bacteriol 175(15):4738-43 Martinez-Morales F, Borges AC, Martinez A, Shanmugam KT, Ingram LO (1999) Chromosomal integration of heterologous DNA in Escherichia coli with precise removal of markers and replicons used during construction. Journal of Bacteriology 181(22):7143-8 McGill DJ, Dawes EA (1971) Glucose and fructose metabolism in Zymomonas anaerobia. Biochem J 125(4):1059-68 Meima R, Haijema BJ, Dijkstra H, Haan GJ, Venema G, Bron S (1997) Role of enzymes of homologous recombination in illegitimate plasmid recombination in Bacillus subtilis. J Bacteriol 179(4):1219-29 Meima R, Haijema BJ, Venema G, Bron S (1995) Overproduction of the ATP-dependent nuclease AddAB improves the structural stability of a model plasmid system in Bacillus subtilis. Mol Gen Genet 248(4):391-8 Meima R, Venema G, Bron S (1996) A positive selection vector for the analysis of structural plasmid instability in Bacillus subtilis. Plasmid 35(1):14-30 doi:S0147- 619X(96)90002-5 [pii] 10.1006/plas.1996.0002 Misawa N, Nakamura K (1989) Nucleotide-Sequence of the 2.7 Kb Plasmid of Zymomonas mobilis Atcc10988. Journal of Biotechnology 12(1):63-70 doi:Doi 10.1016/0168-1656(89)90129-6 Mojica FJM, Díez-Villaseñor C, García-Martínez J, Almendros C (2009) Short motif sequences determine the targets of the prokaryotic CRISPR defence system. Microbiology 155(3):733-740 doi:10.1099/mic.0.023960-0 Mosberg JA, Gregg CJ, Lajoie MJ, Wang HH, Church GM (2012) Improving lambda red genome engineering in Escherichia coli via rational removal of endogenous nucleases. PLoS One 7(9):e44638 doi:10.1371/journal.pone.0044638 PONE-D- 12-12936 [pii] Murphy KC (1998) Use of bacteriophage lambda recombination functions to promote gene replacement in Escherichia coli. J Bacteriol 180(8):2063-71 Nagai T, Ibata K, Park ES, Kubota M, Mikoshiba K, Miyawaki A (2002) A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications. Nat Biotechnol 20(1):87-90 doi:10.1038/nbt0102-87 Nevoigt E (2008) Progress in metabolic engineering of Saccharomyces cerevisiae. Microbiol Mol Biol Rev 72(3):379-412 doi:10.1128/MMBR.00025-07 72/3/379 [pii] Novotny CP, Englesberg E (1966) The L-arabinose permease system in Escherichia coli B/r. Biochim Biophys Acta 117(1):217-30

145

Osman YA, Conway T, Bonetti SJ, Ingram LO (1987) Glycolytic flux in Zymomonas mobilis: enzyme and metabolite levels during batch fermentation. J Bacteriol 169(8):3726-36 Osman YA, Ingram LO (1985) Mechanism of ethanol inhibition of fermentation in Zymomonas mobilis CP4. J Bacteriol 164(1):173-80 Pankova LM, Shvinka JE, Beker MJ (1988) Regulation of Intracellular H+ Balance in Zymomonas mobilis 113 during the Shift from Anaerobic to Aerobic Conditions. Appl Microbiol Biot 28(6):583-588 Pappas KM, Galani I, Typas MA (1997) Transposon mutagenesis and strain construction in Zymomonas mobilis. J Appl Microbiol 82(3):379-88 Parker C, Barnell WO, Snoep JL, Ingram LO, Conway T (1995) Characterization of the Zymomonas mobilis glucose facilitator gene product (glf) in recombinant Escherichia coli: examination of transport mechanism, kinetics and the role of glucokinase in glucose transport. Mol Microbiol 15(5):795-802 Peijnenburg AA, Bron S, Venema G (1987) Structural plasmid instability in recombination- and repair-deficient strains of Bacillus subtilis. Plasmid 17(2):167-70 doi:0147-619X(87)90023-0 [pii] Pohl M (1997) Protein design on pyruvate decarboxylase (PDC) by site-directed mutagenesis. Application to mechanistical investigations, and tailoring PDC for the use in organic synthesis. Advances in biochemical engineering/biotechnology 58:15-43 Pontes MH, Dale C (2011) Lambda red-mediated genetic modification of the insect endosymbiont Sodalis glossinidius. Appl Environ Microbiol 77(5):1918-20 doi:10.1128/AEM.02166-10 AEM.02166-10 [pii] Poole RK (1994) Oxygen reactions with bacterial oxidases and globins: binding, reduction and regulation. Antonie Van Leeuwenhoek 65(4):289-310 Posfai G, Plunkett G, 3rd, Feher T, Frisch D, Keil GM, Umenhoffer K, Kolisnychenko V, Stahl B, Sharma SS, de Arruda M, Burland V, Harcum SW, Blattner FR (2006) Emergent properties of reduced-genome Escherichia coli. Science 312(5776):1044-6 doi:1126439 [pii] 10.1126/science.1126439 Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, Lim WA (2013) Repurposing CRISPR as an RNA-Guided Platform for Sequence-Specific Control of Gene Expression. Cell 152(5):1173-1183 doi:DOI 10.1016/j.cell.2013.02.022 Ranatunga TD, Jervis J, Helm RF, McMillan JD, Hatzis C (1997) Identification of inhibitory components toxic toward Zymomonas mobilis CP4(pZB5) xylose fermentation. Appl Biochem Biotech 67(3):185-198 doi:Doi 10.1007/Bf02788797 Raps S, Demoss RD (1962) Glycolytic enzymes in Zymomonas mobilis. J Bacteriol 84:115-8 Ren C, Chen T, Zhang J, Liang L, Lin Z (2009) An evolved xylose transporter from Zymomonas mobilis enhances sugar transport in Escherichia coli. Microb Cell Fact 8:66 doi:10.1186/1475-2859-8-66 1475-2859-8-66 [pii] Reyes L, Scopes RK (1991) Membrane-associated ATPase from Zymomonas mobilis; purification and characterization. Biochim Biophys Acta 1068(2):174-8 doi:0005- 2736(91)90207-O [pii]

146

Reynen M, Reipen I, Sahm H, Sprenger GA (1990) Construction of expression vectors for the gram-negative bacterium Zymomonas mobilis. Mol Gen Genet 223(2):335- 41 Rogers PL, Jeon YJ, Lee KJ, Lawford HG (2007a) Zymomonas mobilis for Fuel Ethanol and Higher Value Products. In: Olsson L (ed) Biofuels. Advances in Biochemical Engineering/Biotechnology, vol 108. Springer Berlin Heidelberg, pp 263-288 Rogers PL, Jeon YJ, Lee KJ, Lawford HG (2007b) Zymomonas mobilis for fuel ethanol and higher value products. Adv Biochem Eng Biotechnol 108:263-88 doi:10.1007/10_2007_060 Rogers PL, Lee KJ, Tribe DE (1979) Kinetics of Alcohol Production by Zymomonas mobilis at High Sugar Concentrations. Biotechnology Letters 1(4):165-170 doi:Doi 10.1007/Bf01388142 Rogers PL, Shin HS, Wang B (1997) Biotransformation for L-ephedrine production. Adv Biochem Eng Biotechnol 56:33-59 Rosenberg SM (2001) Evolving responsively: adaptive mutation. Nat Rev Genet 2(7):504-15 doi:10.1038/35080556 35080556 [pii] Ruhrmann J, Kramer R (1992) Mechanism of glutamate uptake in Zymomonas mobilis. J Bacteriol 174(23):7579-84 Runquist D, Fonseca C, Radstrom P, Spencer-Martins I, Hahn-Hagerdal B (2009a) Expression of the Gxf1 transporter from Candida intermedia improves fermentation performance in recombinant xylose-utilizing Saccharomyces cerevisiae. Appl Microbiol Biotechnol 82(1):123-30 doi:10.1007/s00253-008- 1773-y Saez LP, Garcia P, Martinez-Luque M, Klipp W, Blasco R, Castillo F (2001) Role for draTG and rnf genes in reduction of 2,4-dinitrophenol by Rhodobacter capsulatus. J Bacteriol 183(5):1780-3 doi:10.1128/JB.183.5.1780-1783.2001 Sahm H, Bringer-Meyer S, Sprenger GA (2006) The Genus Zymomonas. Prokaryotes: A Handbook on the Biology of Bacteria, Vol 5, Third Edition:201-221 doi:Doi 10.1007/0-387-30745-1_10 Saini S, Pearl JA, Rao CV (2009) Role of FimW, FimY, and FimZ in regulating the expression of type i fimbriae in Salmonella enterica serovar Typhimurium. Journal of Bacteriology 191(9):3003-10 doi:10.1128/JB.01694-08 JB.01694-08 [pii] Saloheimo A, Rauta J, Stasyk OV, Sibirny AA, Penttila M, Ruohonen L (2007) Xylose transport studies with xylose-utilizing Saccharomyces cerevisiae strains expressing heterologous and homologous permeases. Appl Microbiol Biot 74(5):1041-52 doi:10.1007/s00253-006-0747-1 Sander JD, Joung JK (2014) CRISPR-Cas systems for editing, regulating and targeting genomes. Nat Biotechnol 32(4):347-55 doi:10.1038/nbt.2842 nbt.2842 [pii] Schmehl M, Jahn A, Meyer zu Vilsendorf A, Hennecke S, Masepohl B, Schuppler M, Marxer M, Oelze J, Klipp W (1993) Identification of a new class of nitrogen fixation genes in Rhodobacter capsulatus: a putative membrane complex involved in electron transport to nitrogenase. Mol Gen Genet 241(5-6):602-15 Schmidt W, Schugerl K (1987) Continuous Ethanol-Production by Zymomonas mobilis on a Synthetic Medium. Chem Eng J Bioch Eng 36(3):B39-B48 doi:Doi 10.1016/0300-9467(87)80030-8

147

Schubert C (2006) Can biofuels finally take center stage? Nature Biotechnology 24(7):777-784 doi:Doi 10.1038/Nbt0706-777 Scordaki A, Drainas C (1987) Analysis of Natural Plasmids of Zymomonas mobilis Atcc- 10988. Journal of General Microbiology 133:2547-2556 Scordaki A, Drainas C (1990) Analysis and stability of Zymomonas mobilis ATCC 10988 plasmid pZMO3. Plasmid 23(1):59-66 doi:0147-619X(90)90044-D [pii] Seatter MJ, De la Rue SA, Porter LM, Gould GW (1998) QLS motif in transmembrane helix VII of the glucose transporter family interacts with the C-1 position of D- glucose and is involved in substrate selection at the exofacial binding site. Biochemistry 37(5):1322-6 doi:10.1021/bi972322u bi972322u [pii] Senthilkumar V, Rajendhran J, Busby SJ, Gunasekaran P (2009) Characterization of multiple promoters and transcript stability in the sacB-sacC gene cluster in Zymomonas mobilis. Arch Microbiol 191(6):529-41 doi:10.1007/s00203-009- 0479-6 Senthilkumar V, Rameshkumar N, Busby SJ, Gunasekaran P (2004) Disruption of the Zymomonas mobilis extracellular sucrase gene (sacC) improves levan production. J Appl Microbiol 96(4):671-6 doi:2169 [pii] Seo JS, Chong H, Park HS, Yoon KO, Jung C, Kim JJ, Hong JH, Kim H, Kim JH, Kil JI, Park CJ, Oh HM, Lee JS, Jin SJ, Um HW, Lee HJ, Oh SJ, Kim JY, Kang HL, Lee SY, Lee KJ, Kang HS (2005) The genome sequence of the ethanologenic bacterium Zymomonas mobilis ZM4. Nat Biotechnol 23(1):63-8 doi:nbt1045 [pii] 10.1038/nbt1045 Sergueev K, Yu D, Austin S, Court D (2001) Cell toxicity caused by products of the p(L) operon of bacteriophage lambda. Gene 272(1-2):227-35 doi:S0378111901005352 [pii] Shamanna DK, Sanderson KE (1979) Uptake and catabolism of D-xylose in Salmonella typhimurium LT2. J Bacteriol 139(1):64-70 Shan Q, Wang Y, Li J, Zhang Y, Chen K, Liang Z, Zhang K, Liu J, Xi JJ, Qiu JL, Gao C (2013) Targeted genome modification of crop plants using a CRISPR-Cas system. Nat Biotechnol 31(8):686-8 doi:10.1038/nbt.2650 nbt.2650 [pii] Sharan SK, Thomason LC, Kuznetsov SG, Court DL (2009) Recombineering: a homologous recombination-based method of genetic engineering. Nat Protoc 4(2):206-23 doi:10.1038/nprot.2008.227 nprot.2008.227 [pii] Shuai L, Yang Q, Zhu JY, Lu FC, Weimer PJ, Ralph J, Pan XJ (2010) Comparative study of SPORL and dilute-acid pretreatments of spruce for cellulosic ethanol production. Bioresour Technol 101(9):3106-14 doi:10.1016/j.biortech.2009.12.044 S0960-8524(09)01709-X [pii] Siegele DA, Hu JC (1997) Gene expression from plasmids containing the araBAD promoter at subsaturating inducer concentrations represents mixed populations. Proc Natl Acad Sci U S A 94(15):8168-72 Silbir S, Dagbagli S, Yegin S, Baysal T, Goksungur Y (2014) Levan production by Zymomonas mobilis in batch and continuous fermentation systems. Carbohydr Polym 99:454-61 doi:10.1016/j.carbpol.2013.08.031 S0144-8617(13)00815-1 [pii] Skerker JM, Leon D, Price MN, Mar JS, Tarjan DR, Wetmore KM, Deutschbauer AM, Baumohl JK, Bauer S, Ibanez AB, Mitchell VD, Wu CH, Hu P, Hazen T, Arkin

148

AP (2013) Dissecting a complex chemical stress: chemogenomic profiling of plant hydrolysates. Mol Syst Biol 9:674 doi:10.1038/msb.2013.30 msb201330 [pii] Skerker JM, Shapiro L (2000a) Identification and cell cycle control of a novel pilus system in Caulobacter crescentus. Embo J 19(13):3223-34 doi:10.1093/emboj/19.13.3223 Skotnicki ML, Lee KJ, Tribe DE, Rogers PL (1981) Comparison of ethanol production by different zymomonas strains. Appl Environ Microbiol 41(4):889-93 Skotnicki ML, Lee KJ, Tribe DE, Rogers PL (1982) Genetic alteration of Zymomonas mobilis for ethanol production. Basic Life Sci 19:271-90 Skotnicki ML, Tribe DE, Rogers PL (1980) R-Plasmid Transfer in Zymomonas mobilis. Appl Environ Microbiol 40(1):7-12 Song KB, Joo HK, Rhee SK (1993) Nucleotide sequence of levansucrase gene (levU) of Zymomonas mobilis ZM1 (ATCC10988). Biochim Biophys Acta 1173(3):320-4 Song KB, Lee SK, Joo HK, Rhee SK (1994) Nucleotide and derived amino acid sequences of an extracellular sucrase gene (invB) of Zymomonas mobilis ZM1 (ATCC10988). Biochim Biophys Acta 1219(1):163-6 doi:0167-4781(94)90262-3 [pii] Sootsuwan K, Thanonkeo P, Keeratirakha N, Thanonkeo S, Jaisil P, Yamada M (2013) Sorbitol required for cell growth and ethanol production by Zymomonas mobilis under heat, ethanol, and osmotic stresses. Biotechnol Biofuels 6(1):180 doi:10.1186/1754-6834-6-180 1754-6834-6-180 [pii] Sprenger GA (1993) Approaches to Broaden the Substrate and Product Range of the Ethanologenic Bacterium Zymomonas mobilis by Genetic-Engineering. Journal of Biotechnology 27(3):225-237 doi:Doi 10.1016/0168-1656(93)90087-4 Sprenger GA (1996) Carbohydrate metabolism in Zymomonas mobilis: A catabolic highway with some scenic routes. Fems Microbiology Letters 145(3):301-307 Sprenger GA, Schorken U, Sprenger G, Sahm H (1995) Transaldolase B of Escherichia coli K-12: cloning of its gene, talB, and characterization of the enzyme from recombinant strains. Journal of Bacteriology 177(20):5930-6 Stanley GA, Hobley TJ, Pamment NB (1997) Effect of acetaldehyde on Saccharomyces cerevisiae and Zymomonas mobilis subjected to environmental shocks. Biotechnol Bioeng 53(1):71-8 doi:10.1002/(SICI)1097-0290(19970105)53:1<71::AID- BIT10>3.0.CO;2-C Strohdeicher M, Bringermeyer S, Neuss B, Vandermeer R, Duine JA, Sahm H (1989) Glucose-Dehydrogenase from Zymomonas mobilis - Evidence for a Quinoprotein. Pqq and Quinoproteins:103-105 Strohdeicher M, Neuss B, Bringermeyer S, Sahm H (1990) Electron-Transport Chain of Zymomonas mobilis - Interaction with the Membrane-Bound Glucose- Dehydrogenase and Identification of Ubiquinone-10. Archives of Microbiology 154(6):536-543 Strohdeicher M, Schmitz B, Bringermeyer S, Sahm H (1988) Formation and Degradation of Gluconate by Zymomonas mobilis. Appl Microbiol Biot 27(4):378-382 Strzelecki AT, Goodman AE, Cail RG, Rogers PL (1990) Behavior of the hybrid plasmid pNSW301 in Zymomonas mobilis grown in continuous culture. Plasmid 23(3):194-200

149

Strzelecki AT, Goodman AE, Rogers PL (1987) Behavior of the IncW plasmid Sa in Zymomonas mobilis. Plasmid 18(1):46-53 doi:0147-619X(87)90077-1 [pii] Subtil T, Boles E (2011) Improving L-arabinose utilization of pentose fermenting Saccharomyces cerevisiae cells by heterologous expression of L-arabinose transporting sugar transporters. Biotechnol Biofuels 4:38 doi:10.1186/1754-6834- 4-38 Sumiya M, Davis EO, Packman LC, McDonald TP, Henderson PJ (1995) Molecular genetics of a receptor protein for D-xylose, encoded by the gene xylF, in Escherichia coli. Receptors Channels 3(2):117-28 Swings J, De Ley J (1977) The biology of Zymomonas. Bacteriol Rev 41(1):1-46 Tanaka H, Ishikawa H, Osuga K, Takagi Y (1990) Fermentation Performance of Zymomonas mobilis against Oxygen-Supply .1. Fermentative Ability of Zymomonas mobilis under Various Oxygen-Supply Conditions in Batch Culture. J Ferment Bioeng 69(4):234-239 doi:Doi 10.1016/0922-338x(90)90219-M Toh H, Doelle H (1997a) Changes in the growth and enzyme level of Zymomonas mobilis under oxygen-limited conditions at low glucose concentration. Arch Microbiol 168(1):46-52 Toh H, Doelle H (1997b) Changes in the growth and enzyme level of Zymomonas mobilis under oxygen-limited conditions at low glucose concentration. Archives of Microbiology 168(1):46-52 doi:DOI 10.1007/s002030050468 Tusnady GE, Simon I (1998) Principles governing amino acid composition of integral membrane proteins: application to topology prediction. J Mol Biol 283(2):489- 506 doi:S0022-2836(98)92107-6 [pii] 10.1006/jmbi.1998.2107 Tusnady GE, Simon I (2001) The HMMTOP transmembrane topology prediction server. Bioinformatics 17(9):849-50 Typas MA, Galani I (1992) Chemical and Uv Mutagenesis in Zymomonas mobilis. Genetica 87(1):37-45 doi:Doi 10.1007/Bf00128771 Untergasser A, Nijveen H, Rao X, Bisseling T, Geurts R, Leunissen JA (2007) Primer3Plus, an enhanced web interface to Primer3. Nucleic Acids Res 35(Web Server issue):W71-4 doi:gkm306 [pii] 10.1093/nar/gkm306 van Kessel JC, Hatfull GF (2007) Recombineering in Mycobacterium tuberculosis. Nat Methods 4(2):147-52 doi:nmeth996 [pii] 10.1038/nmeth996 Vartholomatos G, Typas MA, Drainas C (1993) An ultraviolet-sensitive mutant of Zymomonas mobilis affecting the stability of its natural plasmid pZMO2. Plasmid 29(1):10-8 doi:S0147-619X(83)71002-4 [pii] 10.1006/plas.1993.1002 Viikari L, Gisler R (1986) By-Products in the Fermentation of Sucrose by Different Zymomonas-Strains. Appl Microbiol Biot 23(3-4):240-244 Viret JF, Alonso JC (1987) Generation of linear multigenome-length plasmid molecules in Bacillus subtilis. Nucleic Acids Res 15(16):6349-67 Wang H, Yang H, Shivalila CS, Dawlaty MM, Cheng AW, Zhang F, Jaenisch R (2013) One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-mediated genome engineering. Cell 153(4):910-8 doi:10.1016/j.cell.2013.04.025 S0092-8674(13)00467-4 [pii] Wang HH, Isaacs FJ, Carr PA, Sun ZZ, Xu G, Forest CR, Church GM (2009) Programming cells by multiplex genome engineering and accelerated evolution. Nature 460(7257):894-8 doi:10.1038/nature08187 nature08187 [pii]

150

Wang L, Chen S, Vergin KL, Giovannoni SJ, Chan SW, DeMott MS, Taghizadeh K, Cordero OX, Cutler M, Timberlake S, Alm EJ, Polz MF, Pinhassi J, Deng Z, Dedon PC (2011) DNA phosphorothioation is widespread and quantized in bacterial genomes. Proc Natl Acad Sci U S A 108(7):2963-8 doi:10.1073/pnas.1017261108 1017261108 [pii] Wang Y, Pfeifer BA (2008) 6-deoxyerythronolide B production through chromosomal localization of the deoxyerythronolide B synthase genes in E. coli. Metab Eng 10(1):33-8 doi:S1096-7176(07)00050-X [pii] 10.1016/j.ymben.2007.09.002 Wang Y, Weng J, Waseem R, Yin X, Zhang R, Shen Q (2012) Bacillus subtilis genome editing using ssDNA with short homology regions. Nucleic Acids Res 40(12):e91 doi:10.1093/nar/gks248 gks248 [pii] Wecker MS, Zall RR (1987) Production of Acetaldehyde by Zymomonas mobilis. Appl Environ Microbiol 53(12):2815-20 Wei N, Quarterman J, Kim SR, Cate JHD, Jin YS (2013) Enhanced biofuel production through coupled acetic acid and xylose consumption by engineered yeast. Nat Commun 4 doi:Artn 2580 Doi 10.1038/Ncomms3580 Weisser P, Kramer R, Sprenger GA (1996a) Expression of the Escherichia coli pmi gene, encoding phosphomannose-isomerase in Zymomonas mobilis, leads to utilization of mannose as a novel growth substrate, which can be used as a selective marker. Appl Environ Microbiol 62(11):4155-61 Wheals AE, Basso LC, Alves DMG, Amorim HV (1999) Fuel ethanol after 25 years. Trends Biotechnol 17(12):482-487 doi:Doi 10.1016/S0167-7799(99)01384-0 Wiedenheft B, Sternberg SH, Doudna JA (2012) RNA-guided genetic silencing systems in bacteria and archaea. Nature 482(7385):331-8 doi:10.1038/nature10886 nature10886 [pii] Wyman CE (2001) Twenty years of trials, tribulations, and research progress in bioethanol technology: selected key events along the way. Appl Biochem Biotechnol 91-93:5-21 Yamamoto S, Izumiya H, Morita M, Arakawa E, Watanabe H (2009) Application of lambda Red recombination system to Vibrio cholerae genetics: simple methods for inactivation and modification of chromosomal genes. Gene 438(1-2):57-64 doi:10.1016/j.gene.2009.02.015 S0378-1119(09)00107-3 [pii] Yanase H, Kurii J, Tonomura K (1988) Fermentation of Lactose by Zymomonas mobilis Carrying a Lac+ Recombinant Plasmid. Journal of Fermentation Technology 66(4):409-415 doi:Doi 10.1016/0385-6380(88)90007-6 Yang S, Land ML, Klingeman DM, Pelletier DA, Lu TY, Martin SL, Guo HB, Smith JC, Brown SD (2010a) Paradigm for industrial strain improvement identifies sodium acetate tolerance loci in Zymomonas mobilis and Saccharomyces cerevisiae. Proc Natl Acad Sci U S A 107(23):10395-400 doi:10.1073/pnas.0914506107 0914506107 [pii] Young E, Poucher A, Comer A, Bailey A, Alper H (2011) Functional survey for heterologous sugar transport proteins, using Saccharomyces cerevisiae as a host. Appl Environ Microbiol 77(10):3311-9 doi:10.1128/AEM.02651-10 Young EM, Comer AD, Huang H, Alper HS (2012a) A molecular transporter engineering approach to improving xylose catabolism in Saccharomyces cerevisiae. Metab Eng 14(4):401-11 doi:10.1016/j.ymben.2012.03.004

151

Young EM, Tong A, Bui H, Spofford C, Alper HS (2014a) Rewiring yeast sugar transporter preference through modifying a conserved protein motif. Proc Natl Acad Sci U S A 111(1):131-6 doi:10.1073/pnas.1311970111 Zaccolo M, Gherardi E (1999) The effect of high-frequency random mutagenesis on in vitro protein evolution: a study on TEM-1 beta-lactamase. J Mol Biol 285(2):775- 83 doi:S0022-2836(98)92262-8 [pii] 10.1006/jmbi.1998.2262 Zhang M, Eddy C, Deanda K, Finkelstein M, Picataggio S (1995) Metabolic Engineering of a Pentose Metabolism Pathway in Ethanologenic Zymomonas mobilis. Science 267(5195):240-3 doi:267/5195/240 [pii] 10.1126/science.267.5195.240 Zhang X, Chen G, Liu W (2009) Reduction of xylose to xylitol catalyzed by glucose- fructose oxidoreductase from Zymomonas mobilis. FEMS Microbiol Lett 293(2):214-9 doi:10.1111/j.1574-6968.2009.01529.x FML1529 [pii] Zhang Y, Buchholz F, Muyrers JP, Stewart AF (1998) A new logic for DNA engineering using recombination in Escherichia coli. Nat Genet 20(2):123-8 doi:10.1038/2417 Zikmanis P, Kruce R, Auzina L (1997) An elevation of the molar growth yield of Zymomonas mobilis during aerobic exponential growth. Arch Microbiol 167(2/3):167-71 doi:71670167.203 [pii] Zikmanis P, Kruce R, Auzina L (1999) Molar growth yields of Zymomonas mobilis on glucose after the transition from anaerobic to aerobic continuous growth. Acta Biotechnol 19(1):69-75 doi:DOI 10.1002/abio.370190111 Zou SL, Hong LF, Wang C, Jing X, Zhang MH (2012) Construction of an Unmarked Zymomonas mobilis Mutant Using a Site-Specific FLP Recombinase. Food Technol Biotech 50(4):406-411

152