PSEUDOMONAS AERUGINOSA METABOLISM IN THE CYSTIC FIBROSIS LUNG By Michael Jay Opperman A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Cell and Molecular Biology - Doctor of Philosophy 2016ABSTRACT PSEUDOMONAS AERUGINOSA METABOLISM IN THE CYSTIC FIBROSIS LUNG By Michael Jay Opperman Cystic fibrosis of the lung is an autosomal recessive disease caused by defects in the Cystic Fibrosis Transmembrane conductance Regulator. While current treatments and interventions have greatly reduced morbidity and increased life expectancy in this population, cystic fibrosis patients face chronic infection of the lung. These chronic infections cause a chronic inflammatory state of the lung to develop, leading to slow degradation of the lung parenchyma and expansion of the conductive pathways, a disease known as bronchiectasis. Treatment at the late stages of this pulmonary disease is by heart and lung transplantation, a surgical intervention that carries significant risk as well as limited by the total availability of such organs. A potential intervention in this disease process that may significantly reduce the progression of bronchiectasis is in the effective removal of pathogenic organisms from the cystic fibrosis lung. Of the various organisms that cause chronic infection of the cystic fibrosis lung, Pseudomonas aeruginosa demonstrates the highest prevalence of infection and greatest correlated risk of morbidity and mortality in this population. Pseudomonas aeruginosa is a Gram negative bacterial species know for a wide range of opportunistic infections (e.g. burn wounds, urinary tract infections secondary to catheterization, and ventilator-associated pneumonia) and a diverse range of environments that it inhabits. Effective treatment of P. aeruginosa infection is greatly complicated due to its intrinsic resistance to many different classes of antibiotics as well as the rapid development of resistance to treatment during chronic cystic fibrosis lung infection. A greater understanding of the metabolic phenotypes of P. aeruginosa may lead to the development of both better clinical tests monitoring the state of infection of the lung as well as the intelligent design of pharmaceutical interventions that target metabolic pathways important to growth in this organism during infection. In this dissertation, I present the quantified carbon metabolism of six cystic fibrosis P. aeruginosa isolates and the laboratory reference strain PA01. In these studies, I demonstrate the difference of metabolism under glucose only minimal media growth (M9) and in a media formulated to mimic the cystic fibrosis lung under aerobic and anoxic growth (Simplified Synthetic Cystic Fibrosis sputum Media). These studies reveal shared and divergent metabolic phenotypes across several isolates from different geographical as well as genetic backgrounds. These studies also identify the importance of the glyoxylate cycle and polyamine metabolism during the growth of P. aeruginosa and to demonstrate the variation in utilization of other metabolic sub-systems of central carbon metabolism. Finally, these studies demonstrate the suboptimal biomass production of P. aeruginosa and identify overproduction of reductant as a key product of the central metabolism of P. aeruginosa. Overall, this research makes substantial progress in understanding not only the metabolism of P. aeruginosa in the cystic fibrosis lung but identifying key metabolic pathways for future investigation that may have a significant impact in the understanding of biofilm formation and pathogenic infection of other bacteria.Copyright by MICHAEL JAY OPPERMAN 2016v ACKNOWLEDGMENTS I wish to thank the current and former members of the Shachar-Hill lab during my time as a student for both their technical as well as their emotional support during my PhD years at MSU. I wish to specifically thank Lisa Carey, Dr. Wilson Chen, and Dr. Rahul Deshpande for their technical support as well as their mentorship and guidance during the early years of my research career. I greatly thank Teresa Clark for her technical support and guidance, especially with the MFA process on the HPCC. I also wish to thank Nathan Praschan and Raven Batshon both for their culturing work during the start of this research as well as the experience of mentorship. Finally, I wish to thank Matthew Jeurgens for their technical expertise and friendship during these last, stressful years, of my research. I wish to warmly thank Dr. Cindy Arvidson and Dr. Chris Waters for their allowance of laboratory space for my experiments. I wish to also thank Dr. Martha Mulks for her gift of the strain PA01 as well as guidance during my early experiments. I also greatly thank my guidance committee, Dr. Cindy Arvidson, Dr. Claire Vieille, Dr. Robert Worden, Dr. Sheng-Yang He, and Dr. Yair Shachar-Hill both for their time as well as for their appreciated role as mentors. Finally, I wish to thank my ever-loving wife and companion of many years, Amanda Opperman, whom I could write many words but have so, so little space left. I wish to thank my parents, Roderick fiDabbafl Opperman and Nancy Opperman, for their constant love and support. I cannot forget to thank my bother Nathaniel Opperman, my lifelong friend, and his wife Dr. Lindsey VanSchoiack. I finally wish to thank my Grandparents: Earl and Lodema Harmon; and Herb and Marjorie Opperman for their lifelong love and support as well as the rest of my extended family, too numerous for this page, who have helped me to become the man I am today.vi TABLE OF CONTENTS LIST OF TABLES .............................................................................................. viii LIST OF FIGURES .............................................................................................. ix KEY TO ABBREVIATIONS ............................................................................... xi CHAPTER 1 Œ Pathogenesis of Cystic Fibrosis and Pseudomonas aeruginosa Infection: A Literature Review .............................................................................. 1 The Medical History of Cystic Fibrosis ....................................................... 3 The Molecular Basis for Cystic Fibrosis ...................................................... 6 The Pathophysiology of Cystic Fibrosis ....................................................... 8 The Taxonomy and Metabolism of Pseudomonas aeruginosa ................... 12 Members of Note of the Pseudomonas genus............................................. 14 The Human Pathogen Pseudomonas aeruginosa ........................................ 15 Pseudomonas aeruginosa and the Cystic Fibrosis Lung............................. 18 Metabolic Network Analysis ...................................................................... 19 Metabolic Flux Analysis ............................................................................ 20 Flux Balance Analysis ............................................................................... 21 REFERENCES .......................................................................................... 24 CHAPTER 2 Œ Metabolic Flux Analyses of Pseudomonas aeruginosa Cystic Fibrosis Isolates ................................................................................................... 47 ABSTRACT .............................................................................................. 48 INTRODUCTION ..................................................................................... 49 MATERIALS AND METHODS ............................................................... 52 Strain selection and culture .............................................................. 52 Cell and media sample harvest ......................................................... 53 Determination of substrate uptake rates and product secretion ......... 53 Determination of CO2 efflux rates .................................................... 55 Analysis of Amino Acid labeling ..................................................... 55 Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA) ............................................................................... 56 13C Metabolic Flux Analysis ............................................................ 56 Flux Balance Analysis ..................................................................... 58 RESULTS .................................................................................................. 58 DISCUSSION ............................................................................................ 72 vii ACKNOWLEDGMENTS ......................................................................... 77 APPENDICES ........................................................................................... 78 APPENDIX A - Electronic Supplemental Descriptions ................... 79 APPENDIX B - Additional Figure ................................................... 80 REFERENCES .......................................................................................... 81 CHAPTER 3 Œ Quantification of Pseudomonas Aeruginosa Metabolism in Cystic Fibrosis Lung Like Conditions ............................................................................. 95 ABSTRACT .............................................................................................. 96 INTRODUCTION ..................................................................................... 97 METHODS AND MATERIAL ................................................................ 102 Bacterial strains and culture ............................................................ 102 Cell and media sample harvest ........................................................ 105 Determination of substrate uptake rates and product secretion ........ 105 Determination of CO2 efflux rates ................................................... 106 Determination of NO3 uptake rates ................................................. 106 Analysis of Amino Acid labeling .................................................... 107 Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA) .............................................................................. 108 13C Metabolic Flux Analysis ........................................................... 108 Flux Balance Analysis .................................................................... 110 RESULTS ................................................................................................. 110 DISCUSSION ........................................................................................... 135 ACKNOWLEDGEMENTS ...................................................................... 141 APPENDICES .......................................................................................... 142 APPENDIX A - Electronic Supplemental Descriptions .................. 143 APPENDIX B - Additional Tables and Figures .............................. 144 REFERENCES ......................................................................................... 148 CHAPTER 4 Œ Conclusions and Future Directions ............................................. 163 Comparison of Metabolism ....................................................................... 164 Metabolic Fingerprinting and Metabolic Flux Analysis ............................ 166 Virulence and Metabolism ........................................................................ 167 Future Directions ...................................................................................... 170 REFERENCES ......................................................................................... 173 viii LIST OF TABLES Table 2.1. Pseudomonas aeruginosa strains analyzed .............................................................. 54 Table 2.2. Net and exchange (Xch) fluxes determined by 13C MFA with 90% confidence intervals. .................................................................................................................................... 68 Table 2.3. Flux ratios and cofactor production rates for P. aeruginosa strains. ....................... 71 Table 3.1. List of Pseudomonas aeruginosa strains used in this experiment. ........................ 103 Table 3.2. Uptake and CO2 efflux rate of Pseudomonas aeruginosa strains under aerobic growth conditions in Simplified-SCFM. ................................................................................. 117 Table 3.3. Amino acid and nitrate uptake rates and acetate efflux rates of Pseudomonas aeruginosa strains under anoxic growth conditions in Simplified-SCFM. ............................. 117 Table 3.4. Net and exchange (Xch) fluxes determined by 13C MFA of aerobic SCFM with 90% confidence intervals................................................................................................................. 123 Table 3.5. Net and exchange (Xch) fluxes determined by 13C MFA of anoxic SCFM with 90% confidence intervals................................................................................................................. 128 Table 3.6. Flux ratio tables of experimental strains in major points of carbon diversion in central metabolism and estimated energy production of MFA determined strains flux rates. 133 Table 3.7. MFA constrained FBA to FBA maximized growth results. .................................. 134 Table S3.1. Average and standard deviation of remaining carbon compounds during growth in Synthetic Cystic Fibrosis Media (SCFM) under aerobic conditions....................................... 147 Table S3.2. Average and standard deviation of remaining carbon compounds during growth in Synthetic Cystic Fibrosis Media (SCFM) under Anoxic conditions. ..................................... 147 ix LIST OF FIGURES Figure 2.1. Growth physiology. ............................................................................................. 61 Figure 2.2. Carbon conversion efficiency during growth. ..................................................... 62 Figure 2.3 Discrimination among strains using 13C labeling fingerprinting ......................... 63 Figure 2.4. 13C MFA flux map of strain PA01. ..................................................................... 65 Figure 2.5. Flux map from FBA analysis of strain PA01. ..................................................... 66 Figure 2.6. 13C MFA flux maps of P. aeruginosa cystic fibrosis clinical isolates. ............... 69 Figure 2.7. Comparison of select metabolic fluxes determined by 13C MFA with growth rates of P. aeruginosa strains. ............................................................................................... 70 Figure 2.8. Discrimination of metabolic phenotypes among P. aeruginosa strains. ............. 70 Figure S2.1. Averaged carbon-uptake normalized flux map from 13C MFA analysis of P. aeruginosa cystic fibrosis isolates grown in M9. .................................................................. 80 Figure 3.1. Analysis of amino acids, glucose, and lactate consumption in aerobic Synthetic Cystic Fibrosis sputum Media (SCFM) growth of seven strains of Pseudomonas aeruginosa. ............................................................................................................................................. 112 Figure 3.2. Analysis of amino acids, glucose, and lactate consumption in anoxic Synthetic Cystic Fibrosis sputum Media (SCFM) growth of seven strains of Pseudomonas aeruginosa. ............................................................................................................................................. 113 Figure 3.3. Specific growth rates of Pseudomonas aeruginosa strains under the growth conditions............................................................................................................................. 114 Figure 3.4. Percent of total carbon uptake from Simplified Synthetic Cystic Fibrosis Media (S-SCFM) during aerobic growth by Pseudomonas aeruginosa strains by compound. ..... 118 Figure 3.5. Percent of total carbon uptake from Simplified Synthetic Cystic Fibrosis Media (S-SCFM) during anoxic growth by Pseudomonas aeruginosa strains by compound. ..... 119 Figure 3.6. Determination of Strain Diversity by 13C-Labeling fingerprinting. .................. 121 Figure 3.7. Flux map from 13C MFA analysis for strain PA01 grown Simplified-SCFM under aerobic conditions. ..................................................................................................... 122 x Figure 3.8. Flux map from FBA analysis for strain PA01 grown in Simplified-SCFM under aerobic conditions. ............................................................................................................... 124 Figure 3.9. Averaged carbon-uptake normalized flux map from 13C MFA analysis of P. aeruginosa cystic fibrosis isolates grown in aerobic Simplified-SCFM conditions. .......... 126 Figure 3.10. Flux map from 13C MFA analysis for strain PA01 grown in Simplified-SCFM under anoxic conditions. ...................................................................................................... 127 Figure 3.11. Flux map from FBA analysis for strain PA01 grown in Simplified-SCFM under anoxic conditions. ................................................................................................................ 130 Figure 3.12. Flux map from 13C MFA analysis of P. aeruginosa cystic fibrosis isolates grown in anoxic Simplified-SCFM conditions. ................................................................... 131 Figure 3.13. Hierarchical Cluster Analysis of MFA derived flux values. ........................... 132 Figure S3.1. Correlation of Pseudomonas aeruginosa strain intrinsic growth rates under media and oxygen conditions.. ............................................................................................ 144 Figure S3.2. Carbon conversation efficiency (CCE) calculation of Pseudomonas aeruginosa strains under aerobic growth conditions in Simplified-SCFM with 90% CI. ..................... 145 Figure S3.3. Nitrogen conversation (NC) percentage calculation of Pseudomonas aeruginosa strains under aerobic and anoxic growth conditions .. ..................................... 146 xi KEY TO ABBREVIATIONS KG Œ -ketoglutarate 13C Œ Carbon-13 6PGC Œ 6-phospho-gluconate ABC Œ ATP-binding cassette protein Ac-CoA Œ acetyl-CoA ARG Œ arginine ASP Œ aspartate CCE Œ Carbon Conversion Efficiency CIT Œ citrate CITR Œ citrulline CI Œ confidence interval CF Œ cystic fibrosis CFTR Œ Cystic Fibrosis Transmembrane conductance Regulator EMP Œ EmbdenŒMeyerhofŒParnas pathway EDP Œ Entner-Doudoroff pathway E4P Œ erythrose 4-phosphate FBA Œ Constraints-based Flux Balance Analysis FVA Œ Flux Variability Analysis FUM Œ fumarate GLU Œ glutamate G5SH Œ glutamate-5-semialdehyde G3P Œ glyceraldehyde 3-phosphate xii GLXŒ glyoxylate or glyoxylate cycle H6P Œ hexose-6-phosphate HCA Œ Hierarchical Clustering Analysis ICIT Œ isocitrate LPS Œ lipopolysaccharide MAL Œ malate MFA Œ Metabolic Flux Analysis ORN Œ ornithine OAA Œ oxaloacetate OPPP Œ Oxidative Pentose Phosphate Pathway P5P Œ pentose-5-phosphate PEP Œ phosphoenolpyruvate PGN Œ polygalacturan PCA Œ Principal Component Analysis PKA Œ Protein Kinase A PYR Œ pyruvate S7P Œ sedoheptulose 7-phosphate SUCC Œ succinate SSres Œ sum of squared residuals TCA Œ tricarboxylic acid cycle T2SS Œ Type II Secretion System T3SS ŒType III secretion system 1 CHAPTER 1 Pathogenesis of Cystic Fibrosis and Pseudomonas aeruginosa Infection: A Literature Review2 The work described in this thesis concerns metabolic changes during adaptation and pathogenesis of the bacteria Pseudomonas aeruginosa in chronic infection of the cystic fibrosis lung. Pseudomonas aeruginosa is a significant factor in the morbidity and mortality of cystic fibrosis (CF). Cystic fibrosis is an autosomal recessive disease caused by a defect in the Cystic Fibrosis Transmembrane Conductance Regulator, which causes pathological changes in bicarbonate secretion and mucus function. The research described in this thesis uses Metabolic Flux Analysis and Flux Balance Analysis to quantify central carbon metabolism in cystic fibrosis isolates of P. aeruginosa. These techniques yield characteristic flux maps that define the metabolic phenotypes of the strains under environmental and lung-like conditions. Such maps yield a better systemic understanding of the metabolic evolution of the organism in the cystic fibrosis lung. The metabolic phenotypes can also aid in identifying potential metabolic drug targets. This introductory chapter describes the medical history, pathogenesis, molecular basis, and current treatment of cystic fibrosis. Also included is a description of the taxonomy and the metabolic characteristics of the genus Pseudomonas. The virulence characteristics of P. aeruginosa, its pathogenesis during pulmonary infection in cystic fibrosis, and the current complications in the treatment of lung infections of cystic fibrosis are presented next. Lastly, an introduction is given to the technical aspects of Metabolic Flux Analysis and Flux Balance Analysis, the methods used in the studies presented in subsequent chapters on the metabolism of P. aeruginosa. 3 The Medical History of Cystic Fibrosis Cystic fibrosis of the pancreas, also known as mucoviscidosis but more commonly as cystic fibrosis, was named for the observation of pancreatic cystic and fibrotic lesions observed during autopsy [1]. Modern investigation of the disease can be traced to the published case study of 49 patients with steatorrhea (excessive fat in feces) caused by pancreatic insufficiency who were subsequently shown to have cyst formation and fibrotic lesions in the pancreas at autopsy by Dr. Dorothy Andersen [1]. While early descriptions of the disease focused on the Celiac-like symptoms of the disease, steatorrhea and failure to thrive, a frequent note was the associated chronic respiratory infections in patients that survived the neonatal period [1,2]. Patients with the disease seemed to fall into two groups, one that would die due to obstruction of the digestive system, meconium ileus, within the neonatal period and those that would live past infancy but succumb to chronic respiratory infection [1,2]. Early work to understand the pulmonary complications of cystic fibrosis showed infection principally by Staphylococcus aureus and treatment was diet supplementation with vitamins, limiting fat intake, and ingestion of pancreatic extracts with sulfadiazine or other sulfa drugs used as antibacterial therapy [3]. During this time, 50 percent of those with meconium ileus would die within the first two weeks of life, and 50 percent of those without meconium ileus would die within the first 6 months of life [4]. While early observations included the discussion of many bodily glands possibly being affected, it was not until the 1940s that abnormal mucus gland function was investigated as the basis for the clinical presentation of this disease [5,6]. The study conducted by Farber et al. demonstrated that unlike Celiac disease, cystic fibrosis patient secretions were lower in viscosity and were greatly reduced in the secretion of the pancreatic enzymes trypsin and lipase [6]. A second paper by the Farber gave the disease the name mucoviscidosis to describe the viscous mucus secretions that provided the mechanical basis for the decreased enzyme secretion of the 4 pancreas as well as the increased viscosity of the fluids seen [7]. This conclusion of thickened mucus being a central symptom of the disease would soon be used to explain the strong association between meconium ileus and cystic fibrosis [8,9]. Realization that the disorder involved all exocrine glands would come in the late 1940s. A heat wave in New York City in August 1948 led to the discovery of sweat gland involvement, with an obvious bias toward cystic fibrosis children being hospitalized for heat exhaustion and further investigation showing examples of patients moved to warmer climates to protect against respiratory infections dying quickly as a result of heat exhaustion complications [10]. This observation led to a case-control study and subsequent direct experimental measurement of sweat and blood salt concentration in patients with suspected cystic fibrosis in 1953 [11]. This study provided evidence of abnormal salt loss in cystic fibrosis with increased secretion of sodium, chloride, and potassium by sweat and salivary glands, even under salt-limited intake [11].These results would lead to the development of sweat testing to quickly and reliably determine cystic fibrosis in patient populations [12]. The next decades of cystic fibrosis research, while providing more effective treatments to prolong life, did not yield the true cause of the disease. Treatment of the severest of disease symptoms, meconium ileus, did not begin until 1948, leading to increased survival in some patients [13,14]. Antibiotic treatment, new to medical science, also increased survival of cystic fibrosis patients, but soon led to cases of infection by gram negative organisms, chiefly Pseudomonas aeruginosa, in the 1950s [14Œ17]. In the 1960s and 1970s, research into proposals that cystic fibrosis is an inborn error of metabolism or is caused by abnormal production of mucus proteins would prove these ideas to be wrong, while the discovery of abnormal ciliary motility and subsequent proposal of a cilliary-toxic factor defining cystic fibrosis also failed to 5 yield a clear understanding of the disease [18Œ21]. Ultimately, the lack of another organism with similar presentation and difficulty in culturing cells from cystic fibrosis patients hindered understanding the molecular basis of the disease well into the 1980s [19,22]. Abnormalities in salt content of pulmonary secretions would finally be demonstrated in lung epithelial tissue in 1981,directly showing excessive sodium uptake and demonstrating a likely defect in chloride ion regulation on the apical membrane in cystic fibrosis patients [23]. This theory was further supported by work in cultured cells from the tracheal mucosa of a cystic fibrosis patient [24]. While evidence of chloride ion dysfunction in epithelial tissue was beginning to be established, the genetics of the disease were also beginning to be untangled. Early work by Andersen et al. demonstrated the likelihood that the disease was an autosomal recessive genetic disease [5]. By the 1970s, it had become clear this was a genetic disease predominately affecting Caucasian populations [25Œ27] and was rarely seen in non-Caucasian populations [26,28Œ30]. In the US, rates of gene prevalence were shown to be approximately 2%-5%, but multiple studies failed to show if this was a multi-loci or multi-allelic disease [31Œ33]. The molecular basis for the disease and its genetic quirks would finally be revealed with the development and application of modern DNA technology. Development of genetic linkage analysis using restriction fragment length polymorphisms [34] would lead to the localization of the cystic fibrosis gene to band q31 of chromosome 7 [35,36]. In 1989, Collins, Riordan and Tsui cloned and named the Cystic Fibrosis Transmembrane conductance Regulator (CFTR) [37Œ39]. This work would establish the CFTR as being an ATP-binding cassette (ABC) protein, as well as describe what would be known as the most common mutation, F508del [37,40]. Confirmation of the chloride channel activity of the CFTR would come from work on CFTR point mutations and direct purification and testing of bound CFTR in the early 1990s [41Œ43]. These techniques have 6 also led to a much clearer understanding of the world-wide prevalence of the disease, with those of Northern European descent having the highest risk of cystic fibrosis in the United States with a birth rate of 1 in 3000 live births and worldwide prevalence of the disease varying from 1 in 900 live births in some populations in Canada to 1 in 10,500 live births in Native American populations [44]. With a relatively high allelic rate in northern European populations, it has been proposed that there may be a protective effect for heterozygotes against certain pulmonary infections or conditions, but the true cause for the selective pressure retaining the genetic defect remains an open scientific question [45,46]. The Molecular Basis for Cystic Fibrosis Much is now know about the protein responsible for cystic fibrosis, the Cystic Fibrosis Transmembrane conductance Regulator (CFTR). The CFTR is a cAMP-regulated ABC transporter with the unique property among such proteins of transporting several ions, preferentially chloride ions [47Œ49]. These properties help the protein regulate epithelial sodium channels, regulate ATP-sensitive outwardly rectifying chloride channels, directly transport ATP, regulate potassium channels, and aid in functioning of the bicarbonate/chloride exchange in the pancreas [50Œ53]. CFTR resides in the apical membrane surface with two membrane spanning domains that make up the pore that gates chloride ion transfer [54]. It also contains two cytoplasmic domains that define the protein as a member of the ATP-binding cassette transporter family, an ATP binding domain and an ATP hydrolyzing domain [55]. Regulation of CFTR activity is primarily through phosphorylation by Protein Kinase A (PKA), which is activated by cAMP, but it can also be activated by other proteins (e.g. PKGII) [56Œ58]. Presently under investigation is the role of the CFTR in bicarbonate transport outside of pancreatic exocrine 7 secretion, as acidic secretions are a prominent feature of cystic fibrosis that best explain the broad pathological features of the disease [59Œ61]. With these advances in the molecular understanding of CFTR, the more than 2000 known genetic defects in the CFTR have been classified as pathologies of protein quantity or quality[44]. Defects in protein quantity can involve correct mRNA production, which can be caused by abnormal gene transcription as a result of missense (39.9% of reported CFTR variants), nonsense (8.3%), frameshift (15.7%), or splicing (11.5%) mutations, which lead to premature stop codons or accelerated mRNA degradation and subsequent lack of sufficient CFTR in the cell membrane [62,63]. Mutations causing these transcription defects are more commonly seen in Mediterranean countries (G542X) and in Ashkenazi Jewish populations (W1282X) [55,64]. Additional mutations in the gene promoter region or alternative gene splicing of the CFTR, seen in A455E and P574H defects, can further lead to decreased transcript production and thus decreased amounts of CFTR in the cell membrane [64]. Protein manufacturing defects also reduce protein quantity, such as the endoplasmic reticulum quality control degradation of the CFTR in F508del, which leads to few successful insertions of the CFTR into the cell membrane [65]. Finally, the CFTR itself can be intrinsically unstable, with the Q1412X leading to a CFTR protein with a reduced half-life, caused by the loss of the last 70 amino acids of the protein [64]. Cystic fibrosis pathologies are also caused by a second class of CFTR defects, which are due to intrinsic issues with protein function [44,56,64]. CFTR activity can be affected by genetic mutations that allow for proper protein insertion and levels to be present in the cell, but cause incorrect opening of the channel or limited ion conductance when open [44,56,66]. The 3rd most common cystic fibrosis defect, G551D, is due to dysfunction of the ATP-binding site, which 8 causes obstruction of the open conformation of the protein [55,64]. The conductance of the CFTR can also be affected, as seen in the R117H genetic defect, changing the current-voltage relationship from linear to strongly inwardly rectifying [56,62,64]. The Pathophysiology of Cystic Fibrosis Ultimately, the defective bicarbonate ion localization due to a CFTR defect leads to the central pathology of thickened secretions as a result of improper mucin expansion on epithelial surfaces. The pathophysiology of cystic fibrosis is that chronically dehydrated mucus leads to thickened and sticky secretions, causing the disease pathology [60,61]. While previous theories have explained the lack of proper mucus hydration on the lack of correct chloride transport, a unified theory on cystic fibrosis based on chloride transport does not fully explain the dehydrated mucus seen in the alimentary tract [61,67]. Instead, recent work has established the importance of calcium ions and bicarbonate; calcium ions and bicarbonate regulate charge shielding and alkaline pH, which is critical for the proper unpacking and unfolding of mucin polymers to form a high viscosity hydrated gel [67Œ70]. This explanation of how defective CFTR function affects bicarbonate secretion and consequently mucus function connects the molecular changes to the organ-level dysfunction and symptom presentation in cystic fibrosis. The accumulation of mucus in the digestive system, in particular the duodenum, gives rise to the earliest clinical presentation of the disease. The duodenum, the most proximal section of the small intestine contains the anatomical channels for both the pancreas and the liver, and is where bicarbonate and mucus secretions from the pancreas and duodenum neutralize, and protect the organ from, the acidic stomach contents [71]. This creates the proper environment for micelle formation of lipids and activation of digestive enzymes from the pancreas [72,73]. In about 15% of infants with cystic fibrosis, the first indication of a mucus-plugged small intestine appears 9 with the failure to pass the first stool within the first few days of life, a condition called meconium ileus [44,74]. The CFTR in the pancreatic ducts is central to the high rate of bicarbonate secretion, and the degree of loss of bicarbonate transport in the expressed CFTR allele determines whether pancreatic insufficiency is seen in a patient, which manifests as the pancreatic duct clogged with thickened mucus and accumulating proenzymes beginning in the neonatal period [75Œ78]. The closed pancreatic duct allows the digestive enzymes to activate within the pancreas itself, causing enzymatic digestion, pancreatitis and fibrosis [79,80]. In time, the lack of secretion of digestive enzymes by the pancreas begins to manifest in fatty stools (steatorrhea) and failure to thrive in infants and toddlers due to fat soluble vitamin and lipid absorption failure [80]. Current treatment for malnutrition caused by cystic fibrosis is to increase caloric intake and supplement pancreatic enzymes and fat soluble vitamins [81]. Additional symptoms of cystic fibrosis present as patients enter adulthood. Without sufficient bicarbonate secretion in the duodenum, the small intestine remains acidic, producing an environment hospitable to bacterial overgrowth, further complicating the malnutrition symptoms of the disease [82]. Continuous failure to neutralize gastric contents leads to increased gastric emptying times and decreased lower esophageal tone, causing gastroesophageal reflux disease and peptic ulcers [83Œ85]. In time, chronic production of thickened mucus in the digestive tract can obstruct the small bowel, presenting in Distal Intestinal Obstruction Syndrome, Meconium Plug Syndrome, constipation, intussusception, and, in extreme cases, Pneumatosis Intestinalis, where an intussuscepted bowel ruptures and leads to bowel contents emptying into the abdominal cavity [84]. 97% of males with cystic fibrosis are sterile due to congenital bilateral absence of vas deferens, an obstructive azoospermia likely related to the lack of necessary CFTR expression for proper male reproductive tract development [86,87]. (It should 10 be noted that there is no relationship between sterility in women and cystic fibrosis, and the pregnancies of those with a normal range of lung function and BMI are regarded as low-risk [44,81].) Pancreatic enzyme secretion obstruction due to mucus plugs in the secreting lobes causes cellular metaplasia in mucus secreting cells and, finally, fibrosis of the organ, leading to episodes of pancreatitis[80,84,88]. The endocrine function of the pancreas is affected in 50% of cystic fibrosis patients who are 30 or older, although it is unclear if this is due to direct destruction of the beta cells of the pancreas or dysfunction related to decreased CFTR expression [89,90]. However, pathology in the pulmonary system is the greatest cause of mortality in cystic fibrosis patients. Current understanding is that the pseudostratified ciliated columnar epithelium provides innate bacterial defense by using cilia to move a thin mucus layer on the cell surface towards the larynx, pulling inhaled pathogens away from the lower airways of the lung towards the destructive environment of the stomach [91]. To have sufficient water to allow for proper fluidity of this mucus layer, the proper bicarbonate and chloride transport by the CFTR and sodium by the epithelial sodium channel (ENaC) must be in place to form a pericillary layer below the normal mucus layer [92,93]. In cystic fibrosis, this mucus layer is disturbed, although it is currently unknown if the liquid layer height or viscosity of the mucus is the primary basis of the dysfunction [94,95]. Ultimately, the thickened mucus cannot be moved by the ciliated cells, causing pooled mucus to form and providing a non-hostile environment for bacteria and non-bacterial species to grow [93]. Commonly seen infectious species in this population are Pseudomonas aeruginosa, Staphylococcus aureus, and Hemophilus influenza, although Streptoccocus sp., Burkholderia sp., Streptococcus pneumoniae, and Aspergillus sp. are also of concern [96]. With this increase in susceptibility to infection, cystic fibrosis patients face a 11 pathological cycle of lung destruction due to infection, inflammation, lung tissue destruction due to the immune response, infection clearance, and, followed by another infection [97Œ99]. This prolonged multiple infection leads to a chronic inflammatory state of the lung causes repeated cycles of neutrophil elastase digestion of the lung parenchyma, which causes the obstructive lung disease Bronchiectasis, with a majority of cystic fibrosis patients demonstrating the disease by three years of age [100Œ102]. While treatment of the pulmonary conditions of cystic fibrosis is possible, many issues do arise. Chronic inflammation caused by infection is treated with high-dose ibuprofen, but such treatment can cause Analgesic nephropathy [103]. While alternative anti-inflammatory treatments are under investigation, including Nitric Oxide modulators such as Riociguat or GSNOR inhibitors, ibuprofen remains the most common current treatment [104]. Acute and chronic infection is treated with antibiotics, most commonly aminoglycosides such as tobramycin, vaccination, or monoclonal antibodies against specific bacteria or viruses, but the ever-present issue of antibacterial resistance greatly complicates chronic infection treatment[96,105,106]. DNAses are also used to help reduce thickened mucus secretions [106]. Perhaps the most hopeful treatments involve direct treatment of the defective CFTR, such as the CFTR potentiator, ivacaftor, and CFTR modulators such as lumacaftor, but these cannot treat all of the known defects in the CFTR [104,107]. Epidemiological studies of cystic fibrosis patients with mild and severe disease both show a strong decline in lung function starting at 40 years of age, with no difference seen in pulmonary disease severity [108]. Final treatment for decreased lung function caused by bronchiectasis due to chronic infection is lung transplantation, though the lack of sufficient organs for transplant makes this option unavailable to most [109,110]. Because the leading cause of morbidity and mortality in cystic fibrosis patients is chronic lung 12 infection, understanding Pseudomonas aeruginosa infection in cystic fibrosis patients is of crucial importance [96,105]. Current prevalence rates of P. aeruginosa infection of has been shown to be 74% of all cystic fibrosis patients 26 years and older, with over 50% of patients over the age of 18 demonstrating infection [111]. The current theory of infection of the lung by P. aeruginosa begins with a clinically independent exposure to the bacteria, due to the bacterium's environmental ubiquity [112]. Inhaled P. aeruginosa begins to establish persistent biofilm communities within the lower respiratory tract, beginning a cycle of innate immune system response, inflammation, decreased P. aeruginosa concentration and then re-infection due to a failure of clearance of the biofilm [113,114]. This chronic cycle of inflammation causes the scarring and widening of the airways of the lower respiratory tract known as Bronchiectasis, a condition that can only be treated with lung transplantation in the late stages of the disease [102]. Pseudomonas aeruginosa treatment using current antibacterial therapy is insufficient, as biofilm increases microbial survival and recolonization after treatment, eventually leading to a state of nearly complete antibiotic resistance in chronic infection [115]. The long range goal of the research described in this thesis is to identify metabolic features of P. aeruginosa during cystic fibrosis infections which can be used to aid in treatment. The Taxonomy and Metabolism of Pseudomonas aeruginosa The Pseudomonadaceae are a Family of organisms that have been observed since the beginning of the study of bacteria, with the first description and use of the name Pseudomonas occurring in 1885 [116,117]. With the establishment of ribosomal RNA-based taxonomy the present genera of bacteria in this family consists primarily of Pseudomonas sp., Azotobacter sp., Azomonas sp., Azorhizophilus sp., and Cellvibrio sp.[118]. Pseudomonas aeruginosa belongs to 13 the genus Pseudomonas, the Type Genus of the family[118,119]. While initially classified as most genetically similar to P. fluorescens and P. putida, inclusion of several house-keeping gene sequences to establish higher resolution inter-species genetic lineages has demonstrated P. oleovorans and P. stutzeri as belonging to the P. aeruginosa subgroup and P. fluorescens and P. putida belonging to separate subgroups of Pseudomonas [119,120]. Members of Pseudomonas share several physically and chemically definable characteristics. Pseudomonas sp. are Gram negative bacilli with polar flagella found ubiquitously in water and land ecosystems between 4°C-42°C and a pH range of 4-8, making many environments potential reservoirs [118,121]. They are common opportunistic pathogens of animals and plants [118,119]. A defining characteristic of some members of the genus, though not expressed by all members, is the production of fluorescent colored compounds, such as phenazine pigments, toxins that can act as siderophores [122,123]. The metabolic characteristics of the genus Pseudomonas help further define the group. Pseudomonas are chemoorganotrophic, nonphotosynthetic bacteria that require only simple carbon compounds to survive, with only three known species of the genus requiring more complex carbon compounds to grow [118]. While all genes for tricarboxylic acid cycle (TCA) and oxidative pentose phosphate pathway (OPPP) enzymes have been found through genome sequencing of several Pseudomonas species [124Œ127], the general experimental finding in the genus is the unique lack of an active 6-phosphofructokinase[128Œ131], causing a loss of a fully functional EmbdenŒMeyerhofŒParnas (EMP) pathway and the reliance on the Entner-Doudoroff (ED) pathway[118]. The Entner-Doudoroff pathway is an enzymatic pathway that allows for the cleavage of 6-phospho-gluconate to one pyruvate and one glyceraldehyde-3-phosphate using the enzymes 6-phosphogluconate dehydratase and 2keto-3-deoxyphosphogluconate aldolase [132]. This 14 pathway is seen in many bacterial species, and may be a more primitive form of carbohydrate catabolism than the EMP pathway [133]. This pathway only produces one ATP, one NADPH, and one NADH compared to the EMP pathway's net production of 2 ATP and 2 NADH, but the ED pathway has been proposed to decrease total protein need for glycolytic metabolism, allowing for a total net increase in energy with respect to metabolic need for enzyme production [134]. Anaerobic metabolism and fermentation pathways do exist in Pseudomonas, which are often generally mis-described as a genus of obligate aerobes. Arginine can be deaminated to ornithine to directly produce ATP in several Pseudomonas species, with growth only observed under rich media conditions [135Œ137]. Pyruvate fermentation is known to occur under anaerobic growth in only two members of the genus, Pseudomonas extremaustralis and Pseudomonas aeruginosa, with P. aeruginosa producing lactate, acetate, acetaldehyde, or ethanol [138,139]. A striking metabolic characteristic of the genus Pseudomonas is the ability to use nitrogen compounds as alternative electron acceptors[140Œ143]. Using the enzymes nitrate reductase, nitrite reductase, nitric oxide reductase, and nitrous oxide reductase, Pseudomonas are able to reduce nitrate to nitrite, nitric oxide, nitrous oxide, and finally to nitrogen gas [144]. These enzymes receive electrons from the same electron donors to oxygen oxidases, ubiquinols and the cytochrome bc1 complex that are employed in aerobic respiration. This allows for the electron transfer chain of oxidative phosphorylation to continue functioning in anoxia with only a slight decrease in energy transfer efficiency [144,145]. Members of Note of the Pseudomonas genus Members of the Pseudomonas genus have been studied for various scientific and industrial properties. Pseudomonas putida has been studied in the potential "green" production of 15 food flavoring additives, (e.g. cinnamic acid, cis-cis-muconate, p-hydroxy- benzoate), important aromatic chemical precursors (e.g. p-cuomarate), and pharmaceutical chemical precursors (e.g.myxochromide) [146Œ148]. Pseudomonas putida strain KT2440 is noteworthy, as it is a well-studied, sequenced organism that is a Generally Recognized as Safe certified strain [146,149]. Pseudomonas fluorescensis a well-known rhizosphere bacteria, known for its commensal nature with plant roots in helping obtain nutrients while producing anti-fungal compounds [150]. The Pseudomonas genus members most similar to P. aeruginosa, P. oleovorans and P. stutzeri are also of scientific and practical interest. P. oleovorans, a bacteria that can directly oxidize petroleum derived hydrocarbons, is under continued study for use in the "green" production of biodegradable thermoplastics[151Œ153]. P. stutzeri has been of interest for its bioremediation potential because of its ability to digest aromatic hydrocarbons and its resistance to various heavy metals [154]. Pseudomonas syringae is a well-known plant pathogen of model and agricultural plants such as kiwifruit producing Actinidia deliciosa and hazelnut producing Corylus avellana; the bacterium is used as a model system for plant-pathogen interactions [155,156]. The pathogenesis of P. syringae in plants involves entering a plant at sites of injury or openings such as the stomata[157,158]. Once inside, the bacterium uses a Type III secretion system (T3SS) to directly inject the plant cell with virulence factors to suppress plant cell recognition as well as developing a biofilm to sustain infection[157,159]. The Human Pathogen Pseudomonas aeruginosa Of the Pseudomonads, P. aeruginosa is the most significant for human health. The name aeruginosa, literally Latin for "copper rust", is a reference to the similarity of the pigmented 16 organism's colonies to that of the color of oxidized copper (verdigris) [119].The first recorded isolation and description of P. aeruginosa was in 1872 [119]. It is an opportunistic pathogen and, in addition to humans, can infect amoebas, fungi, plants, nematodes and other mammals [160Œ163]. Opportunistic infections of P. aeruginosa occur as a result of the breakdown or weakening of anatomical barriers of the host; this makes the bacterium a common cause of nosocomial infection, principally from contaminated medical equipment that bypasses physical barriers: catheterization-related Urinary Tract Infections, ventilator-associated pneumonia, and a general infection risk of wounds and burns in a hospital setting [124,164Œ166]. P. aeruginosa, in a non-medical setting, is more commonly seen in "hot tub folliculitis", an infection of hair follicles seen post exposure to a water source that has not been properly treated against bacterial growth [166]. The study of the organism as a causative agent of human disease has led to a close examination of the many factors responsible for its pathogenicity. P. aeruginosa use Type IV pili and fimbriae for motility and attachment to epithelial cell membranes and other materials [167]. P. aeruginosa cells that have entered the respiratory tract have been shown to bind to the sialic acid structures present on epithelial cells and mucin proteins [168,169]. Sialic acid structures on cells are believed to be sources of self-recognition by the immune system and have been shown to increase during inflammation [170,171]. In the consistently pro-inflammatory state with thickened and static pools of mucin in the pulmonary system of cystic fibrosis, P. aeruginosa quickly establishes residency [172]. After attaching, P. aeruginosa begins to secrete various compounds to begin digestion of the host environment. P. aeruginosa has a well described Type III secretion system (T3SS) allowing for direct injection of host cells with virulence factors [114]. Expression of this 17 virulence factor alone is associated with a six-fold increase in the relative risk of mortality in P. aeruginosa associated pneumonia [173]. P. aeruginosa secretes several proteins through a Type II Secretion System (T2SS), such as Exotoxin A, a diptheria-toxin like toxin that causes ADP-ribosylation of EF-2 that may help in breakdown of tissue barriers during infection [174,175]. Various proteases are also secreted via the T2SS, importantly: the elastase LasB, which cleaves elastin to cause the breakdown of the mechanical barriers that hold tissues together; the lysine-specific zinc metallo-endopeptidase LasA that enhances elastase activity and cleaves Staphylococcus aeures peptidoglycan; and phospholipases to cleave the phospholipid cell membrane of the host cells [115,175Œ179]. With sufficient concentration of P. aeruginosa, formation of the virulence factor with the greatest scientific and medical interest, biofilm, begins. Biofilm production by bacteria is widely reported, and is seen as a protective mechanism of bacteria against their environment and especially against antibacterial agents [164,180Œ182]. Biofilm production is a complex and coordinated communal behavior that relies on group signaling, known as quorum signaling, to produce coordinated production of biofilm precursor products at the correct bacteria density to form a mature biofilm [183Œ185]. In P. aeruginosa production of sufficient quantities of the quorum signaling molecules belonging to the family of N-acyl homoserine lactone, 3-oxo-C12-HSL and C4ŒHSL, leads to activation of the lasRI and the rhlRI gene regulation systems [184]. This regulation system causes metabolic and cellular changes to shift the bacteria into production of biofilm components, L-guluronic acid and D-mannuronic acid, and physiological changes for biofilm niche living [186]. Important changes to respiration are likely to occur as well, as low oxygen tension within cystic fibrosis mucus is further reduced within biofilm [187]. The biofilm community also fosters the growth of a small population of highly resistant cells known as 18 persister cells, capable of replenishing the biofilm with bacteria after exposure to an antibiotic [188,189]. P. aeruginosa expresses a diverse array of antibiotic resistance beside biofilm production. P. aeruginosa has an inducible protein AmpC, a cephalosporinase that can hydrolyze the -lactam ring found in the penicillin drug family of antibiotics to cause inactivation [190]. The bacteria also contains several efflux pump proteins, capable of removing the antibacterial drug families of fluoroquinolones and aminoglycosides from the cell's cytosol [191]. Disturbingly, strains of P. aeruginosa have been reported, that are resistant to colistin, a last-resort antibiotic [192]. Pseudomonas aeruginosa and the Cystic Fibrosis Lung Many studies of P. aeruginosa as a cystic fibrosis pathogen have focused on identifying genetic changes during chronic infection and the production of biofilms and virulence factors, but much less is known about the system-wide metabolic phenotypes of this and other pathogenic microbes or their physiological adaptations during chronic infections [112,124,164,193Œ202]. Omic studies of pathogenic P. aeruginosa strains have described changes at the genomic [203Œ212] transcriptomic [206,213Œ217] and proteomic [181,213,218] level that point to numerous changes in adaptation to the cystic fibrosis lung environment. Evidence in these and other studies indicates that this organism grows in a microaerobic to anaerobic environment within the lung; nitrate is available for respiration and available primary carbon sources include amino acids, glucose, and lactic acid [219,220]. Direct measures of P. aeruginosa cystic fibrosis isolates found that the overproduction of the biofilm component alginate is commonly associated with increased morbidity and mortality in the patient population. This produces a mucoid colony phenotype in culture [176,221Œ223]. Chronic 19 infection isolates of P. aeruginosa also demonstrate decreased production of other known virulence factors [197]. However, these identified changes, have not led to a clear understanding of the metabolic adaptations that the organism undergoes for two distinct reasons. First, chronic infection strains of P. aeruginosa often lack mismatch repair enzymes, creating the hypermutable phenotype, allowing for increased niche specialization in the pulmonary system [224]. This causes most chronic infection strain isolates P. aeruginosa to have significant genomic instability, introducing significant uncertainty in the direct genomic measurements due to the very high genomic diversity observed [217,224Œ226]. Second, different Omic datasets do not correlate well (Pearson™s correlation coefficients (r2) values are generally around 0.4-0.5) within data sets or with comparisons to other Omic measures of the same samples, giving uncertainty to the true meaning of the differences recorded with these techniques [227Œ229]. To identify which changes are actually significant for pathogenesis and to measure and predict functional changes in metabolism, a toolbox of network based computational and experimental methods is available. Metabolic Network Analysis Metabolic Network Analysis is the application of topological analysis and computational theory to biological measurements to better understand the interplay between genetics and biochemistry to produce metabolism [230]. The methods for describing metabolism mathematically and modeling it computationally were developed for the purpose of directed genetic engineering of microbes to produce desirable biomolecules [231,232]. The major approaches to analyzing and modeling metabolic fluxes can be divided into enzyme kinetics based representation of chemical reactions or steady-state assumption based approaches 20 [233].The work in this thesis focuses on two steady-state based modeling techniques, Metabolic Flux Analysis and Flux Balance Analysis. Metabolic Flux Analysis Metabolic Flux Analysis (MFA) centers on the use of isotopically labeled substrates (usually 13C) in experiments allowed to reach a stable pattern of isotopic distribution (end point labeling) [234]. The labeling patterns in metabolic products or intermediates are used to deduce internal metabolic state and pattern of carbon fluxes through the cellular metabolic system: a flux map[235]. To produce such a map measurements are made of extracellular input and output fluxes such as biomass production, gas exchange, and carbon inputs[235]. Second, metabolite pools must also accurately represent subcompartmentalization of a cell such that measured mass isomers (isotopomers) accurately represent labeling in different parts of the cells [236]. Third, the major biochemical pathways need to be known and the stoichiometry and carbon positional transformations for each reaction are needed for each reaction step [237]. Fourth it is assumed (based on past metabolic studies) that there are no changes in enzyme kinetics due to increases in mass due to isotopic labeling [234]. With these assumptions in mind, the computational modeling and fitting of 13C labeling data can be discussed. Metabolism of carbon-containing compounds by enzymatic processes occurs at defined carbon positions within the molecule due to the biochemical specificity of enzymatic catalysis; this fundamental property of enzymatic reactions allows for characteristic labeling patterns to develop for the carbon in the molecule [238]. To describe the re-shuffling of label for a reaction, an atom mapping matrix is constructed for each reaction in the network [239]. Since each carbon atom in a metabolite is either labeled or unlabeled, there are a total of 2nisotopomers [239]. Using mapping matrices and isotopomer matrix representations matrix 21 algebra can be used to predict the labeling pattern at steady state in all metabolites by assuming reaction rates (the values of the metabolic fluxes) [235]. The computed results are then compared to the measured 13C labeled measurements and the flux values are iteratively changed to optimize the fit of the modeled to the measured labeling patterns [235]. Computational software such as 13C-Flux is available for this stage of MFA [240]. This results in MFA quantifying carbon fluxes through metabolism. To date, most MFA studies of microbial systems have focused on questions related to biotechnology and metabolic engineering, microbial physiology, and gene function [241Œ245]. Several recent studies have analyzed the metabolic interactions of pathologically relevant bacteria within their host environment [246,247] and metabolic differences between mutant strains [248]. A recent study of pathogenesis related metabolism in P. aeruginosa employed MFA to compare 17 uropathogenic strains [249]. A more detailed discussion with a methodologically improved study directed specifically at cystic fibrosis isolates can be found in the next chapter of this dissertation. Flux Balance Analysis In addition to the use of MFA to quantify metabolism, Constraints-based Flux Balance Analysis (FBA) is a biochemically-based mathematical modeling approach to the steady state analysis of metabolism that incorporates carbon and non-carbon elements in a large-scale mass balance analysis; energy and redox co-factors are also balanced and no labeling experimental data are needed [250Œ252]. FBA can extend the metabolic scope beyond carbon and beyond central metabolism to a complete genomic scale coverage of metabolism [251,253]. The biochemical equations that make up the individual steps of metabolism are converted into a matrix equation in which the stoichiometry of each reaction (how many of which metabolites are 22 consumed and produced) is described in a Stoichiometric or S matrix and this acts on a vector containing the carbon flux values for each reaction to yield a vector containing the change of metabolite concentrations over time (Equation 1). (1) Where: is change of concentration over time S is the stoichometrixmatrix v is the flux vector Equation 1 can be constrained to reduce the possible mathematical solution space using assumptions, such as the steady state assumption of no net accumulation of internal metabolite pools, and through measurements like proteomic or transcriptomic data to restrict the reactions that are potentially active in the model [251]. Additionally, optimization algorithms can be used to maximize or minimize selected individual fluxes or combinations of fluxes (called objective functions)in order to produce the most likely carbon flux of the metabolic network being tested [251]. The COBRA (Constraints Based Reconstruction Analysis) toolbox in MATLAB is an FBA computational system with access to these functions as described above were used for FBA in this research project [254]. FBA investigations of pathogenic organisms have been used to search for novel drug targets and have identified potential metabolic targets not affected by current therapeutics, such as amino acid production or fatty acid metabolism [255Œ262]. Recent FBA research on P. aeruginosa includes a genome-based metabolic and transport model by Oberhardt and colleagues [263], and used Flux Balance Analysis (FBA) with transcript data from two cystic fibrosis clinical strains isolated 44 months apart to investigate P. aeruginosa™s metabolic capabilities and 23 potential metabolic changes during prolonged infection [264]. FBA has also been used to identify potential metabolic drug targets during biofilm growth of P. aeruginosa[265]. The combination of both FBA and MFA allows for investigation of suboptimal growth and testing of alternative objective functions that may be optimized besides maximal biomass production. However, there have been surprisingly few studies in which these complementary network flux analysis approaches have been combined [146,227,241,266Œ271], let alone in direct application to understand bacterial pathogenicity. In the following chapters, the investigation of Pseudomonas aeruginosa metabolism and its changes due to exposure to and growth in the cystic fibrosis lung environment will be described. 24 REFERENCES25 REFERENCES 1. ANDERSEN DH. CYSTIC FIBROSIS OF THE PANCREAS AND ITS RELATION TO CELIAC DISEASE. Am J Dis Child. 1938;56: 344. doi:10.1001/archpedi.1938.01980140114013 2. Jeffery FW. Cystic Fibrosis of the Pancreas. Can Med Assoc J. 1941;45: 224Œ229. Available: http://www.ncbi.nlm.nih.gov/pubmed/1827103 3. di SANT™AGNESE PEA, Anderson DH. CELIAC SYNDROME IV. Chemotherapy in Infections of the Respiratory Tract Associated with Cystic Fibrosis of the Pancreas. Am J Dis Child. 1946;72: 17. doi:10.1001/archpedi.1946.02020300024003 4. Mantle ADJ, Norman AP. Life-Table For Cystic Fibrosis. Br Med J. 1966;2: 1238Œ1241. 5. ANDERSEN DH. CELIAC SYNDROME V. Genetics of Cystic Fibrosis of the Pancreas With a Consideration of Etiology. Am J Dis Child. 1946;72: 62. doi:10.1001/archpedi.1946.02020300069004 6. Farber S, Shwachman H, Maddock CL. PANCREATIC FUNCTION AND DISEASE IN EARLY LIFE. I. PANCREATIC ENZYME ACTIVITY AND THE CELIAC SYNDROME. J Clin Invest. 1943;22: 827Œ38. doi:10.1172/JCI101456 7. Farber S. Some organic digestive disturbances in early life. J Mich Med Soc. 1945;44: 587. 8. Bodian M (Editor). Fibrocystic disease of the pancreas. A congenital disorder of mucus productionŠmucosis. Eugen Rev. 1953;45: 110Œ111. Available: http://www.ncbi.nlm.nih.gov/pubmed/2973441 9. Glanzmann E. Einführung in die Kinderheilkunde: In 195 Vorlesungen für Studierende undÄrzte. Vienna: Springer Vienna; 1949. pp. 304Œ320. doi:10.1007/978-3-7091-3548-8_70 10. KESSLER WR, ANDERSEN DH. Heat prostration in fibrocystic disease of the pancreas and other conditions. Pediatrics. 1951;8: 648Œ56. Available: http://www.ncbi.nlm.nih.gov/pubmed/14891321 11. DI SANT™AGNESE PA, DARLING RC, PERERA GA, SHEA E. Abnormal electrolyte composition of sweat in cystic fibrosis of the pancreas; clinical significance and relationship to the disease. Pediatrics. 1953;12: 549Œ63. Available: http://www.ncbi.nlm.nih.gov/pubmed/13111855 12. GIBSON LE, COOKE RE. A test for concentration of electrolytes in sweat in cystic fibrosis of the pancreas utilizing pilocarpine by iontophoresis. Pediatrics. 1959;23: 545Œ9. Available: http://www.ncbi.nlm.nih.gov/pubmed/13633369 13. HIATT RB, WILSON PE. CELIAC SYNDROME .7. THERAPY OF MECONIUM 26 ILEUS - REPORT OF 8 CASES WITH A REVIEW OF THE LITERATURE. Surg Gynecol Obstet. 55 E ERIE ST, CHICAGO, IL 60611: FRANKLIN H MARTIN FOUNDATION; 1948;87: 317Œ327. 14. SHWACHMAN H, PRYLES C V, GROSS RE. Meconium ileus; a clinical study of twenty surviving patients. AMA J Dis Child. 1956;91: 223Œ44. Available: http://www.ncbi.nlm.nih.gov/pubmed/13301137 15. GARRARD SD, RICHMOND JB, HIRSCH MM. Pseudomonas aeruginosa infection as a complication of therapy in pancreatic fibrosis (muscoviscidosis). Pediatrics. 1951;8: 482Œ8. Available: http://www.ncbi.nlm.nih.gov/pubmed/14882898 16. Donnison AB, Shwachman H, Gross RE. A review of 164 children with meconium ileus seen at the Children™s Hospital Medical Center, Boston. Pediatrics. 1966;37: 833Œ850. Available: http://sfx.scholarsportal.info/mcmaster?sid=OVID:medline&id=pmid:5932631&id=doi:&issn=0031-4005&isbn=&volume=37&issue=5&spage=833&pages=833-50&date=1966&title=Pediatrics&atitle=A+review+of+164+children+with+meconium+ileus+seen+at+the+Children™s+Hospital 17. Huang NN, Van Loon EL, Sheng KT. The flora of the respiratory tract of patients with cystic fibrosis of the pancreas. J Pediatr. 1961;59: 512Œ521. doi:10.1016/S0022-3476(61)80234-5 18. di Sant™Agnese PA, Davis PB. Research in Cystic Fibrosis. N Engl J Med. 1976;295: 481Œ485. doi:10.1056/NEJM197608262950905 19. Di Sant™Agnese PA, Davis PB. Research in Cystic Fibrosis. N Engl J Med. 1976;295: 534Œ541. doi:10.1056/NEJM197609022951005 20. Di Sant™Agnese PA, Davis PB. Research in Cystic Fibrosis. N Engl J Med. 1976;295: 597Œ602. doi:10.1056/NEJM197609092951105 21. di Sant™Agnese PA, Talamo RC. Pathogenesis and Physiopathology of Cystic Fibrosis of the Pancreas. N Engl J Med. 1967;277: 1399Œ1408. doi:10.1056/NEJM196712282772605 22. Davis PB, Di PA, Agnese S, Agnese DIS. A Review. Cystic Fibrosis at Forty-Quo Vadis? 1980;87: 83Œ87. 23. Knowles M, Gatzy J, Boucher R. Increased bioelectric potential difference across respiratory epithelia in cystic fibrosis. N Engl J Med. 1981;305: 1489Œ95. doi:10.1056/NEJM198112173052502 24. Widdicombe JH, Welsh MJ, Finkbeiner WE. Cystic fibrosis decreases the apical membrane chloride permeability of monolayers cultured from cells of tracheal epithelium. Proc Natl Acad Sci U S A. 1985;82: 6167Œ71. doi:10.1073/pnas.82.18.6167 25. KRAMM ER, CRANE MM, SIRKEN MG, BROWN ML. A cystic fibrosis pilot survey in three New England states. Am J Public Health Nations Health. 1962;52: 2041Œ57. Available: http://www.ncbi.nlm.nih.gov/pubmed/14035460 27 26. Steinberg AG, Brown DC. On the Incidence of Cystic Fibrosis of the Pancreas. Am J Hum Genet. 1960;12: 416Œ24. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1932169&tool=pmcentrez&rendertype=abstract 27. Walbaum R. Is mucoviscidosis transmited by a recessive autosomal gene? Lille médical J la Fac médecine Pharm l™Université Lille. 1972;17: 980. 28. GOODMAN HO, REED SC. Heredity of fibrosis of the pancreas; possible mutation rate of the gene. Am J Hum Genet. 1952;4: 59Œ71. Available: http://www.ncbi.nlm.nih.gov/pubmed/14943708 29. Wright SW, Morton NE. Genetic studies on cystic fibrosis in Hawaii. Am J Hum Genet. 1968;20: 157Œ69. 30. Kulczycki LL. Cystic Fibrosis in Blacks in Washington, DC. Am J Dis Child. 1974;127: 64. doi:10.1001/archpedi.1974.02110200066009 31. CONNEALL.PM, MERRITT AD, YU PL. CYSTIC-FIBROSIS - POPULATION GENETICS. Tex Rep Biol Med. MEDICAL BRANCH, GALVESTON, TX 77550: UNIV TEXAS; 1973;31: 639Œ650. 32. MERRITT AD, TODD CW, MYERS TL, HANNA BL. INCIDENCE AND MODE OF INHERITANCE OF CYSTIC FIBROSIS. J Lab Clin Med. 11830 WESTLINE INDUSTRIAL DR, ST LOUIS, MO 63146-3318: MOSBY-YEAR BOOK INC; 1962;60: 998Œ&. 33. Grand RJ. Pregnancy in Cystic Fibrosis of the Pancreas. JAMA J Am Med Assoc. 1966;195: 993. doi:10.1001/jama.1966.03100120061014 34. Jeffreys a J, Wilson V, Thein SL. Individual-specific fifingerprintsfl of human DNA. Nature. 1985;316: 76Œ79. doi:10.1038/316076a0 35. Wainwright BJ, Scambler PJ, Schmidtke J, Watson EA, Law HY, Farrall M, et al. Localization of cystic fibrosis locus to human chromosome 7cen-q22. Nature. 1985;318: 384Œ5. doi:10.1038/318384a0 36. Tsui L, Buchwald M, Barker D, Braman J, Knowlton R, Schumm J, et al. Cystic fibrosis locus defined by a genetically linked polymorphic DNA marker. Science (80- ). 1985;230: 1054Œ1057. doi:10.1126/science.2997931 37. Kerem B, Rommens JM, Buchanan J a, Markiewicz D, Cox TK, Chakravarti A, et al. Identification of the cystic fibrosis gene: genetic analysis. Science. 1989;245: 1073Œ80. Available: http://www.ncbi.nlm.nih.gov/pubmed/2570460 38. Rommens J, Iannuzzi M, Kerem B, Drumm M, Melmer G, Dean M, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science (80- ). 1989;245: 1059 Œ1065. doi:10.1126/science.2772657 39. Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, et al. 28 Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science. 1989;245: 1066Œ1073. doi:10.1126/science.2475911 40. Morral N, Bertranpetit J, Estivill X, Nunes V, Casals T, Giménez J, et al. The origin of the major cystic fibrosis mutation (delta F508) in European populations. Nat Genet. 1994;7: 169Œ175. doi:10.1038/ng0694-169 41. Anderson MP, Gregory RJ, Thompson S, Souza DW, Paul S, Mulligan RC, et al. Demonstration that CFTR is a chloride channel by alteration of its anion selectivity. Science. 1991;253: 202Œ5. doi:10.1126/science.1712984 42. Bear CE, Li C, Kartner N, Bridges RJ, Jensen TJ, Ramjeesingh M, et al. Purification and functional reconstitution of the cystic fibrosis transmembrane conductance regulator (CFTR). Cell. 1992;68: 809Œ818. doi:10.1016/0092-8674(92)90155-6 43. Tabcharani JA, Chang XB, Riordan JR, Hanrahan JW. Phosphorylation-regulated Cl- channel in CHO cells stably expressing the cystic fibrosis gene. Nature. 1991;352: 628Œ631. doi:10.1038/352628a0 44. O™Sullivan BP, Freedman SD. Cystic fibrosis. Lancet. Elsevier Ltd; 2009;373: 1891Œ1904. doi:10.1016/S0140-6736(09)60327-5 45. Meindl RS. Hypothesis: A Selective Advantage for Cystic Fibrosis Heterozygotes. Am J Phys Anthropol. 1987;74: 39Œ45. 46. Poolman EM, Galvani AP. Evaluating candidate agents of selective pressure for cystic fibrosis. J R Soc Interface. 2007;4: 91Œ98. doi:10.1098/rsif.2006.0154 47. Gadsby DC, Vergani P, Csanády L. The ABC protein turned chloride channel whose failure causes cystic fibrosis. Nature. 2006;440: 477Œ483. doi:10.1038/nature04712 48. Riordan, Rommens J, Kerem B, Alon N, Rozmahel R, Grzelczak Z, et al. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science (80- ). 1989;245: 1066Œ1073. doi:10.1126/science.2475911 49. Dean M, Allikmets R. Complete Characterization of the Human ABC Gene Family. J Bioenerg Biomembr. 2001;33: 475Œ479. doi:10.1023/A:1012823120935 50. Stutts MJ, Canessa CM, Olsen JC, Hamrick M, Cohn JA, Rossier BC, et al. CFTR as a cAMP-dependent regulator of sodium channels. Science. 1995;269: 847Œ850. doi:10.1126/science.7543698 51. Schwiebert EM, Egan ME, Hwang TH, Fulmer SB, Allen SS, Cutting GR, et al. CFTR regulates outwardly rectifying chloride channels through an autocrine mechanism involving ATP. Cell. 1995;81: 1063Œ1073. doi:10.1016/S0092-8674(05)80011-X 52. Reisin IL, Prat a G, Abraham EH, Amara JF, Gregory RJ, Ausiello D a, et al. The cystic fibrosis transmembrane conductance regulator is a dual ATP and chloride channel. J Biol Chem. 1994;269: 20584Œ91. Available: http://www.ncbi.nlm.nih.gov/pubmed/7519611 29 53. Mehta A. CFTR: More than just a chloride channel. Pediatr Pulmonol. 2005;39: 292Œ298. doi:10.1002/ppul.20147 54. Anderson MP, Sheppard DN, Berger H a, Welsh MJ. Chloride channels in the apical membrane of normal and cystic fibrosis airway and intestinal epithelia. Am J Physiol. 1992;263: L1Œ14. Available: http://www.ncbi.nlm.nih.gov/pubmed/1322048 55. Cant N, Pollock N, Ford RC. CFTR structure and cystic fibrosis. Int J Biochem Cell Biol. Elsevier Ltd; 2014;52: 15Œ25. doi:10.1016/j.biocel.2014.02.004 56. Vankeerberghen A, Cuppens H, Cassiman JJ. The cystic fibrosis transmembrane conductance regulator: An intriguing protein with pleiotropic functions. J Cyst Fibros. 2002;1: 13Œ29. doi:10.1016/S1569-1993(01)00003-0 57. Berger HA, Anderson MP, Gregory RJ, Thompson S, Howard PW, Maurer RA, et al. Identification and regulation of the cystic fibrosis transmembrane conductance regulator-generated chloride channel. J Clin Invest. 1991;88: 1422Œ1431. doi:10.1172/JCI115450 58. Sheppard DN, Welsh MJ. Structure and function of the CFTR chloride channel. Physiol Rev. 1999;79: S23Œ45. Available: http://www.ncbi.nlm.nih.gov/pubmed/9922375 59. Choi JY, Lee MG, Ko S, Muallem S. Cl(-)-dependent HCO3- transport by cystic fibrosis transmembrane conductance regulator. Jop. 2001;2: 243Œ246. 60. Borowitz D. CFTR, bicarbonate, and the pathophysiology of cystic fibrosis. Pediatr Pulmonol. 2015;50: S24ŒS30. doi:10.1002/ppul.23247 61. Quinton PM. Cystic fibrosis: impaired bicarbonate secretion and mucoviscidosis. Lancet. 2008;372: 415Œ417. doi:10.1016/S0140-6736(08)61162-9 62. Ferec C, Cutting GR. Assessing the disease-liability of mutations in CFTR. Cold Spring Harb Perspect Med. 2012;2. doi:10.1101/cshperspect.a009480 63. Brennan M-L, Schrijver I. Cystic Fibrosis. J Mol Diagnostics. American Society for Investigative Pathology and the Association for Molecular Pathology; 2016;18: 3Œ14. doi:10.1016/j.jmoldx.2015.06.010 64. Wang Y, Wrennall JA, Cai Z, Li H, Sheppard DN. Understanding how cystic fibrosis mutations disrupt CFTR function: From single molecules to animal models. Int J Biochem Cell Biol. Elsevier Ltd; 2014;52: 47Œ57. doi:10.1016/j.biocel.2014.04.001 65. Cutting GR. Cystic fibrosis genetics: from molecular understanding to clinical application. Nat Rev Genet. Nature Publishing Group; 2015;16: 45Œ56. doi:10.1038/nrg3849 66. Wang Y, Li C, Gao C, Ma C, Xu P. Genome Sequence of the Nonpathogenic Pseudomonas aeruginosa Strain ATCC 15442. Genome Announc. 2014;2: 2Œ3. doi:10.1128/genomeA.00421-14 67. Garcia MAS, Yang N, Quinton PM. Normal mouse intestinal mucus release requires cystic fibrosis transmembrane regulator-dependent bicarbonate secretion. J Clin Invest. 30 2009;119: 2613Œ2622. doi:10.1172/JCI38662 68. Ambort D, Johansson ME V., Gustafsson JK, Nilsson HE, Ermund A, Johansson BR, et al. Calcium and pH-dependent packing and release of the gel-forming MUC2 mucin. Proc Natl Acad Sci. 2012;109: 5645Œ5650. doi:10.1073/pnas.1120269109 69. Verdugo P, Deyrup-Olsen I, Aitken M, Villalon M, Johnson D. Molecular Mechanism of Mucin Secretion: I. The Role of Intragranular Charge Shielding. J Dent Res. 1619 DUKE ST, ALEXANDRIA, VA 22314: AMER ASSOC DENTAL RESEARCH; 1987;66: 506Œ508. doi:10.1177/00220345870660022001 70. Nordman H, Davies JR, Herrmann a, Karlsson NG, Hansson GC, Carlstedt I. Mucus glycoproteins from pig gastric mucosa: identification ofdifferent mucin populations from the surface epithelium. Biochem J. 1997;326 ( Pt 3: 903Œ910. 71. Elmunzer BJ, Elta GH. Biliary Tract Motor Function and Dysfunction. Sleisenger and Fordtran™s Gastrointestinal and Liver Disease Tenth Edition. 2016. pp. 1078Œ1084. 72. Roy CC, Weber AM, Lepage G, Smith L, Levy E. Digestive and Absorptive Phase Anomalies Associated with the Exocrine Pancreatic Insufficiency of Cystic Fibrosis. J Pediatr Gastroenterol Nutr. 227 EAST WASHINGTON SQ, PHILADELPHIA, PA 19106: LIPPINCOTT-RAVEN PUBL; 1988;7: S1ŒS7. doi:10.1097/00005176-198811001-00002 73. HEIZER WD, CLEAVELAND CR, IBER FL. GASTRIC INACTIVATION OF PANCREATIC SUPPLEMENTS. Bull Johns Hopkins Hosp. JOURNALS PUBLISHING DIVISION, 2715 NORTH CHARLES ST, BALTIMORE, MD 21218-4319: JOHNS HOPKINS UNIV PRESS; 1965;116: 261Œ70. Available: http://www.ncbi.nlm.nih.gov/pubmed/14272432 74. De Lisle RC, Borowitz D. The Cystic Fibrosis Intestine. Cold Spring Harb Perspect Med. 2013;3: a009753Œa009753. doi:10.1101/cshperspect.a009753 75. Gustafsson JK, Ermund A, Ambort D, Johansson ME V, Nilsson HE, Thorell K, et al. Bicarbonate and functional CFTR channel are required for proper mucin secretion and link cystic fibrosis with its mucus phenotype. J Exp Med. 2012;209: 1263Œ72. doi:10.1084/jem.20120562 76. Steward MC, Ishiguro H, Case RM. Mechanisms of Bicarbonate Secretion in the Pancreatic Duct. Annu Rev Physiol. 2005;67: 377Œ409. doi:10.1146/annurev.physiol.67.031103.153247 77. Ishiguro H, Yamamoto A, Nakakuki M, Yi L, Ishiguro M, Yamaguchi M, et al. Physiology and pathophysiology of bicarbonate secretion by pancreatic duct epithelium. Nagoya J Med Sci. 2012;74: 1Œ18. Available: http://www.ncbi.nlm.nih.gov/pubmed/22515107 78. Choi JY, Muallem D, Kiselyov K, Lee MG, Thomas PJ, Muallem S. Aberrant CFTR-dependent HCO3- transport in mutations associated with cystic fibrosis. Nature. 2001;410: 31 94Œ97. doi:10.1038/35065099 79. Meyerholz DK, Stoltz DA, Pezzulo AA, Welsh MJ. Pathology of gastrointestinal organs in a porcine model of cystic fibrosis. Am J Pathol. American Society for Investigative Pathology; 2010;176: 1377Œ89. doi:10.2353/ajpath.2010.090849 80. Gibson-Corley KN, Meyerholz DK, Engelhardt JF. Pancreatic pathophysiology in cystic fibrosis. J Pathol. 2016;238: 311Œ320. doi:10.1002/path.4634 81. Schindler T, Michel S, Wilson AWM. Nutrition Management of Cystic Fibrosis in the 21st Century. Nutr Clin Pract. 2015;30: 488Œ500. doi:10.1177/0884533615591604 82. Lisowska A, Wójtowicz J, Walkowiak J. Small intestine bacterial overgrowth is frequent in cystic fibrosis: Combined hydrogen and methane measurements are required for its detection. Acta Biochim Pol. 2009;56: 631Œ634. doi:20091820 [pii] 83. Mousa HM, Woodley FW. Gastroesophageal reflux in cystic fibrosis: Current understandings of mechanisms and management. Curr Gastroenterol Rep. 2012;14: 226Œ235. doi:10.1007/s11894-012-0261-9 84. Kelly T, Buxbaum J. Gastrointestinal Manifestations of Cystic Fibrosis. Dig Dis Sci. Springer US; 2015;60: 1903Œ1913. doi:10.1007/s10620-015-3546-7 85. Stallings VA, Stark LJ, Robinson KA, Feranchak AP, Quinton H. Evidence-Based Practice Recommendations for Nutrition-Related Management of Children and Adults with Cystic Fibrosis and Pancreatic Insufficiency: Results of a Systematic Review. J Am Diet Assoc. 2008;108: 832Œ839. doi:10.1016/j.jada.2008.02.020 86. Chen H, Ruan YC, Xu WM, Chen J, Chan HC. Regulation of male fertility by CFTR and implications in male infertility. Hum Reprod Update. 2012;18: 703Œ713. doi:10.1093/humupd/dms027 87. Tizzano EF, Chitayat D, Buchwald M. Cell-specific localization of CFTR mRNA shows developmentally regulated expression in human fetal tissues. Hum Mol Genet. 1993;2: 219Œ224. doi:10.1093/hmg/2.3.219 88. Allan Tucker J, Spock A, Spicer SS, Shelburne JD, Bradford W. Inspissation of Pancreatic Zymogen Material in Cystic Fibrosis. Ultrastruct Pathol. 2003;27: 323Œ335. doi:10.1080/01913120390231717 89. Ntimbane T, Mailhot G, Spahis S, Rabasa-Lhoret R, Kleme ML, Melloul D, et al. CFTR silencing in Pancreatic Beta-Cells Reveals a Functional Impact on Glucose- Stimulated Insulin Secretion and Oxidative Stress Response. Am J Physiol - Endocrinol Metab. 2015;1: ajpendo.00333.2015. doi:10.1152/ajpendo.00333.2015 90. Hameed S, Jaffe A, Verge CF. Cystic Fibrosis Related Diabetes (CFRD)-The End Stage of Progressive Insulin Deficiency. Pediatr Pulmonol. 2011;46: 747Œ760. doi:10.1002/ppul.21495 91. Knowles MR, Boucher RC. Mucus clearance as a primary innate defense mechanism for 32 mammalian airways. J Clin Invest. 2002;109: 571Œ577. doi:10.1172/JCI200215217 92. Collawn JF, Matalon S. CFTR and Lung Homeostasis. Am J Physiol Lung Cell Mol Physiol. 2014;35294: ajplung.00326.2014. doi:10.1152/ajplung.00326.2014 93. Livraghi A, Randell SH. Cystic fibrosis and other respiratory diseases of impaired mucus clearance. Toxicol Pathol. 2007;35: 116Œ29. doi:10.1080/01926230601060025 94. Derichs N, Jin B-J, Song Y, Finkbeiner WE, Verkman a S. Hyperviscous airway periciliary and mucous liquid layers in cystic fibrosis measured by confocal fluorescence photobleaching. FASEB J. 2011;25: 2325Œ2332. doi:10.1096/fj.10-179549 95. Chen J-H, Stoltz DA, Karp PH, Ernst SE, Pezzulo AA, Moninger TO, et al. Loss of Anion Transport without Increased Sodium Absorption Characterizes Newborn Porcine Cystic Fibrosis Airway Epithelia. Cell. 2010;143: 911Œ923. doi:10.1016/j.cell.2010.11.029 96. Lipuma JJ. The changing microbial epidemiology in cystic fibrosis. Clin Microbiol Rev. 2010;23: 299Œ323. doi:10.1128/CMR.00068-09 97. Bonfield TL, Panuska JR, Konstan MW, Hilliard KA, Hilliard JB, Ghnaim H, et al. Inflammatory cytokines in cystic fibrosis lungs. Am J Respir Crit Care Med. 1995;152: 2111Œ2118. doi:10.1164/ajrccm.152.6.8520783 98. Muhlebach MS, Stewart PW, Leigh MW, Noah TL. Quantitation of inflammatory responses to bacteria in young cystic fibrosis and control patients. Am J Respir Crit Care Med. 1999;160: 186Œ191. doi:10.1164/ajrccm.160.1.9808096 99. Muhlebach MS, Reed W, Noah TL. Quantitative Cytokine Gene Expression in CF Airway. Pediatr Pulmonol. 2004;37: 393Œ399. doi:10.1002/ppul.20010 100. Mott LS, Park J, Murray CP, Gangell CL, de Klerk NH, Robinson PJ, et al. Progression of early structural lung disease in young children with cystic fibrosis assessed using CT. Thorax. 2011;67: 509Œ516. doi:10.1136/thoraxjnl-2011-200912 101. Sly P, Gangell C, Chen L, Ware R, Ranganathan S, Mott LS, et al. Risk factors for bronchiectasis in children with cystic fibrosis. N Engl J Med. 2013;368: 1963Œ70. doi:10.1056/NEJMoa1301725 102. O™Donnell AE. Bronchiectasis. Chest. 2008;134: 815Œ823. doi:10.1378/chest.08-0776 103. Savant AP, McColley SA. 2014 year in review: Cystic fibrosis. Pediatr Pulmonol. 2015;50: 1147Œ1156. doi:10.1002/ppul.23309 104. Solomon GM, Marshall SG, Ramsey BW, Rowe SM. Breakthrough therapies: Cystic fibrosis (CF) potentiators and correctors. Pediatr Pulmonol. 2015;50: S3ŒS13. doi:10.1002/ppul.23240 105. Lyczak JB, Cannon CL, Pier GB. Lung infections associated with cystic fibrosis. Clin Microbiol Rev. 2002;15: 194Œ222. doi:10.1128/CMR.15.2.194 33 106. Mogayzel PJ, Naureckas ET, Robinson KA, Mueller G, Hadjiliadis D, Hoag JB, et al. Cystic fibrosis pulmonary guidelines: Chronic medications for maintenance of lung health. Am J Respir Crit Care Med. 2013;187: 680Œ689. doi:10.1164/rccm.201207-1160OE 107. Bell SC, De Boeck K, Amaral MD. New pharmacological approaches for cystic fibrosis: Promises, progress, pitfalls. Pharmacol Ther. Elsevier Inc.; 2014;145: 19Œ34. doi:10.1016/j.pharmthera.2014.06.005 108. Nick JA, Chacon CS, Brayshaw SJ, Jones MC, Barboa CM, St. Clair CG, et al. Effects of gender and age at diagnosis on disease progression in long-term survivors of cystic fibrosis. Am J Respir Crit Care Med. 2010;182: 614Œ626. doi:10.1164/rccm.201001-0092OC 109. Blatter J, Sweet S. Lung Transplantation in Cystic Fibrosis: Trends and Controversies. Pediatr Allergy Immunol Pulmonol. 2015;28: 237Œ243. doi:10.1089/ped.2015.0564 110. Flores JS, Rovedder PME, Ziegler B, Pinotti AFF, Barreto SSM, Dalcin P d. TR. Clinical Outcomes and Prognostic Factors in a Cohort of Adults With Cystic Fibrosis: A 7-Year Follow-Up Study. Respir Care. 2016;61: 192Œ199. doi:10.4187/respcare.04097 111. Salsgiver EL, Fink AK, Knapp EA, LiPuma JJ, Olivier KN, Marshall BC, et al. Changing Epidemiology of the Respiratory Bacteriology of Patients with Cystic Fibrosis. Chest. Elsevier Inc; 2015;149: 390Œ400. doi:10.1378/chest.15-0676 112. Hogardt M, Heesemann J. Adaptation of Pseudomonas aeruginosa during persistence in the cystic fibrosis lung. Int J Med Microbiol. Elsevier GmbH.; 2010;300: 557Œ562. doi:10.1016/j.ijmm.2010.08.008 113. Zanin M, Baviskar P, Webster R, Webby R. The Interaction between Respiratory Pathogens and Mucus. Cell Host Microbe. Elsevier Inc.; 2016;19: 159Œ168. doi:10.1016/j.chom.2016.01.001 114. Gonçalves-de-Albuquerque CF, Silva AR, Burth P, Rocco PRM, Castro-Faria MV, Castro-Faria-Neto HC. Possible mechanisms of Pseudomonas aeruginosa-associated lung disease. Int J Med Microbiol. 2016;306: 20Œ28. doi:10.1016/j.ijmm.2015.11.001 115. Chatterjee M, Anju CP, Biswas L, Anil Kumar V, Gopi Mohan C, Biswas R. Antibiotic resistance in Pseudomonas aeruginosa and alternative therapeutic options. Int J Med Microbiol. Elsevier GmbH.; 2016;306: 48Œ58. doi:10.1016/j.ijmm.2015.11.004 116. Winslow CE, Broadhurst J, Buchanan RE, Krumwiede C, Rogers LA, Smith GH. The Families and Genera of the Bacteria: Preliminary Report of the Committee of the Society of American Bacteriologists on Characterization and Classification of Bacterial Types. J Bacteriol. 1917;2: 505Œ66. Available: http://www.ncbi.nlm.nih.gov/pubmed/16558764 34 117. Gunsalus IC. Pseudomonas: A century of biodiversity. In: Nakazawa, T and Furukawa, K and Haas, D and Silver, S, editor. MOLECULAR BIOLOGY OF PSEUDOMONADS. 1325 MASSACHUSETTS AVE NW, WASHINGTON, DC 20005-4171: AMER SOC MICROBIOLOGY; 1996. pp. 8Œ21. 118. Moore ERB, Tindall BJ, Martins Dos Santos VAP, Pieper DH, Ramos J-L, Palleroni NJ. Nonmedical: Pseudomonas. In: Intergovernmental Panel on Climate Change, editor. The Prokaryotes. New York, NY: Springer New York; 2006. pp. 646Œ703. doi:10.1007/0-387-30746-X_21 119. Yahr TL, Parsek MR. Pseudomonas aeruginosa. The Prokaryotes. Springer New York; 2006. pp. 704Œ713. doi:10.1007/0-387-30746-X_22 120. Mulet M, Lalucat J, García-Valdés E. DNA sequence-based analysis of the Pseudomonas species. Environ Microbiol. 2010;12: 1513Œ1530. doi:10.1111/j.1462-2920.2010.02181.x 121. Migula W. System der bakterien. Fischer; 1897. 122. Meyer J, Meyer J, Baida N, Baida N, Gardan L, Gardan L, et al. Siderophore Typing, a Powerful Tool for the Identi cation of Fluorescent and Non uorescent Pseudomonads. Society. 2002;68: 2745Œ2753. doi:10.1128/AEM.68.6.2745 123. Meyer J, Stintzi A. Iron metabolism and siderophores in Pseudomonas and related species. In: Montie T, editor. PSEUDOMONAS. PLENUM PRESS DIV PLENUM PUBLISHING CORP, 233 SPRING ST, NEW YORK, NY 10013 USA; 1998. pp. 201Œ243. 124. Stover C, Pham X, Erwin a, Mizoguchi S. Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen. Nature. 2000;406: 959Œ964. Available: http://www.nature.com/nature/journal/v406/n6799/abs/406959a0.html 125. Nelson KE, Weinel C, Paulsen IT, Dodson RJ, Hilbert H, Martins dos Santos VAP, et al. Complete genome sequence and comparative analysis of the metabolically versatile Pseudomonas putida KT2440. Environ Microbiol. 2002;4: 799Œ808. doi:10.1046/j.1462-2920.2002.00366.x 126. Buell CR, Joardar V, Lindeberg M, Selengut J, Paulsen IT, Gwinn ML, et al. The complete genome sequence of the Arabidopsis and tomato pathogen Pseudomonas syringae pv. tomato DC3000. Proc Natl Acad Sci U S A. 2003;100: 10181Œ10186. doi:10.1073/pnas.1731982100 127. Paulsen IT, Press CM, Ravel J, Kobayashi DY, Myers GS a, Mavrodi D V, et al. Complete genome sequence of the plant commensal Pseudomonas fluorescens Pf-5. Nat Biotechnol. 2005;23: 873Œ878. doi:10.1038/nbt1110 128. Vicente M, Canovas JL. Glucolysis in Pseudomonas putida: physiological role of alternative routes from the analysis of defective mutants. J Bacteriol. 1973;116: 908Œ914. 35 129. Heath HE, Gaudy ET. Relationship between catabolism of glycerol and metabolism of hexosephosphate derivatives by Pseudomonas aeruginosa. J Bacteriol. 1978;136: 638Œ646. 130. O™Brien RW. Enzymatic analysis of the pathways of glucose catabolism and gluconeogenesis in Pseudomonas citronellolis. Arch Microbiol. 1975;103: 71Œ6. Available: http://www.ncbi.nlm.nih.gov/pubmed/239656 131. TIWARI N, CAMPBELL J. Enzymatic control of the metabolic activity of Pseudomonas aeruginosa grown in glucose or succinate media. Biochim Biophys Acta - Gen Subj. 1969;192: 395Œ401. doi:10.1016/0304-4165(69)90388-2 132. Conway T. The Entner-Doudoroff pathway: history, physiology and molecular biology. FEMS Microbiol Rev. 1992;9: 1Œ27. doi:10.1016/0378-1097(92)90334-K 133. Romano a. H, Conway T. Evolution of carbohydrate metabolic pathways. Res Microbiol. 1996;147: 448Œ455. doi:10.1016/0923-2508(96)83998-2 134. Flamholz A, Noor E, Bar-Even A, Liebermeister W, Milo R. Glycolytic strategy as a tradeoff between energy yield and protein cost. Proc Natl Acad Sci U S A. 2013;110: 10039Œ44. doi:10.1073/pnas.1215283110 135. Liu YM, Sun ZH, Ni Y, Zheng P, Liu YP, Meng FJ. Isolation and identification of an arginine deiminase producing strain Pseudomonas plecoglossicida CGMCC2039. World J Microbiol Biotechnol. 2008;24: 2213Œ2219. doi:10.1007/s11274-008-9732-4 136. Shibatani, T Kakimoto, T Chibata I. Crystallization and properties of L-arginine deiminase of Pseudomonas putida. 1975; 4580Œ4583. 137. Mercenier a, Simon JP, Haas D, Stalon V. Catabolism of L-arginine by Pseudomonas aeruginosa. J Gen Microbiol. 1980;116: 381Œ389. doi:10.1099/00221287-116-2-381 138. Raiger Iustman LJ, Tribelli PM, Ibarra JG, Catone M V., Solar Venero EC, L??pez NI. Genome sequence analysis of Pseudomonas extremaustralis provides new insights into environmental adaptability and extreme conditions resistance. Extremophiles. 2015;19: 207Œ220. doi:10.1007/s00792-014-0700-7 139. Williams HD, Zlosnik JE a, Ryall B. Oxygen, Cyanide and Energy Generation in the Cystic Fibrosis Pathogen Pseudomonas aeruginosa. Advances in Microbial Physiology. 2006. doi:10.1016/S0065-2911(06)52001-6 140. Wilson LP, Bouwer EJ. Biodegradation of aromatic compounds under mixed oxygen/denitrifying conditions: a review. J Ind Microbiol Biotechnol. 1997;18: 116Œ130. doi:10.1038/sj.jim.2900288 141. Su W, Zhang L, Li D, Zhan G, Qian J, Tao Y. Dissimilatory nitrate reduction by Pseudomonas alcaliphila with an electrode as the sole electron donor. Biotechnol Bioeng. 2012;109: 2904Œ2910. doi:10.1002/bit.24554 36 142. Gamble TN, Betlach MR, Tiedje JM. Numerically Dominant Denitrifying Bacteria from World Soils. Appl Environ Microbiol. 1325 MASSACHUSETTS AVENUE, NW, WASHINGTON, DC 20005-4171: AMER SOC MICROBIOLOGY; 1977;33: 926Œ939. 143. Delorme S, Philippot L, Edel-Hermann V, Deulvot C, Mougel C, Lemanceau P. Comparative Genetic Diversity of the narG, nosZ, and 16S rRNA Genes in Fluorescent Pseudomonads. Appl Environ Microbiol. 2003;69: 1004Œ1012. doi:10.1128/AEM.69.2.1004-1012.2003 144. Zumft W. Cell biology and molecular basis of denitrification. Microbiol Mol Biol Rev. 1997;61: 533Œ616. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=232623&tool=pmcentrez&rendertype=abstract 145. Arai H. Regulation and function of versatile aerobic and anaerobic respiratory metabolism in Pseudomonas aeruginosa. Front Microbiol. 2011;2: 1Œ13. doi:10.3389/fmicb.2011.00103 146. Puchalka J, Oberhardt MA, Godinho M, Bielecka A, Regenhardt D, Timmis KN, et al. Genome-Scale Reconstruction and Analysis of the Pseudomonas putida KT2440 Metabolic Network Facilitates Applications in Biotechnology. Plos Comput Biol. 2008;4: e1000210. doi:10.1371/journal.pcbi.1000210 147. Nijkamp K, Westerhof RGM, Ballerstedt H, De Bont JAM, Wery J. Optimization of the solvent-tolerant Pseudomonas putida S12 as host for the production of p-coumarate from glucose. Appl Microbiol Biotechnol. 2007;74: 617Œ624. doi:10.1007/s00253-006-0703-0 148. Stephan S, Heinzle E, Wenzel SC, Krug D, M??ller R, Wittmann C. Metabolic physiology of Pseudomonas putida for heterologous production of myxochromide. Process Biochem. 2006;41: 2146Œ2152. doi:10.1016/j.procbio.2006.06.022 149. Loeschcke A, Thies S. Pseudomonas putidaŠa versatile host for the production of natural products. Appl Microbiol Biotechnol. 2015;99: 6197Œ6214. doi:10.1007/s00253-015-6745-4 150. Berg G, Smalla K. Plant species and soil type cooperatively shape the structure and function of microbial communities in the rhizosphere. FEMS Microbiol Ecol. 2009;68: 1Œ13. doi:10.1111/j.1574-6941.2009.00654.x 151. Pivnick H, Sabina LR. Oxidation of soluble oil emulsions and emulsifiers by Pseudomonas oleovorans and Pseudomonas formicans. Appl Microbiol. 1956;4: 171Œ5. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1057195&tool=pmcentrez&rendertype=abstract 152. Akaraonye E, Keshavarz T, Roy I. Production of polyhydroxyalkanoates: The future green materials of choice. J Chem Technol Biotechnol. 2010;85: 732Œ743. 37 doi:10.1002/jctb.2392 153. de Smet MJ, Eggink G, Witholt B, Kingma J, Wynberg H. Characterization of intracellular inclusions formed by Pseudomonas oleovorans during growth on octane. J Bacteriol. 1983;154: 870Œ878. 154. Lalucat J, Bennasar A, Bosch R, García-Valdés E, Palleroni NJ. Biology of Pseudomonas stutzeri. Microbiol Mol Biol Rev. 2006;70: 510Œ547. doi:10.1128/MMBR.00047-05 155. Abramovitch RB, Martin GB. Strategies used by bacterial pathogens to suppress plant defenses. Curr Opin Plant Biol. 2004;7: 356Œ364. doi:10.1016/j.pbi.2004.05.002 156. Lamichhane JR, Varvaro L, Parisi L, Audergon J-M, Morris CE. Chapter Four Œ Disease and Frost Damage of Woody Plants Caused by Pseudomonas syringae: Seeing the Forest for the Trees [Internet]. Advances in Agronomy. Elsevier; 2014. doi:10.1016/B978-0-12-800132-5.00004-3 157. Ichinose Y, Taguchi F, Mukaihara T. Pathogenicity and virulence factors of Pseudomonas syringae. J Gen Plant Pathol. 2013;79: 285Œ296. doi:10.1007/s10327-013-0452-8 158. Melotto M, Underwood W, He SY. Role of stomata in plant innate immunity and foliar bacterial diseases. Annu Rev Phytopathol. 2008;46: 101Œ22. doi:10.1146/annurev.phyto.121107.104959 159. Boller T, He SY. Innate Immunity in Plants: An Arms Race Between Pattern Recognition Receptors in Plants and Effectors in Microbial Pathogens. Science (80- ). 2009;324: 742Œ744. doi:10.1126/science.1171647 160. Elrod RP, Braun AC. Pseudomonas aeruginosa; its role as a plant pathogen. J Bacteriol. 1942;44: 633Œ645. 161. Yorgey P, Rahme LG, Tan MW, Ausubel FM. The roles of mucD and alginate in the virulence of Pseudomonas aeruginosa in plants, nematodes and mice. Mol Microbiol. 2001;41: 1063Œ1076. doi:10.1046/j.1365-2958.2001.02580.x 162. Qureshi MN, Perez AA, Madayag RM, Bottone EJ. Inhibition of Acanthamoeba species by Pseudomonas aeruginosa: rationale for their selective exclusion in corneal ulcers and contact lens care systems. J Clin Microbiol. 1993;31: 1908Œ10. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=265656&tool=pmcentrez&rendertype=abstract 163. Kerr JR. Suppression of fungal growth exhibited by Pseudomonas aeruginosa. J Clin Microbiol. 1994;32: 525Œ7. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=263067&tool=pmcentrez&rendertype=abstract 164. Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science. 1999;284: 1318Œ1322. doi:10.1126/science.284.5418.1318 165. Cross A, Allen JR, Burke J, Ducel G, Harris A, John J, et al. Nosocomial infections due to 38 Pseudomonas aeruginosa: review of recent trends. Rev Infect Dis. 1983;5 Suppl 5: S837Œ45. Available: http://www.ncbi.nlm.nih.gov/pubmed/6361960 166. Mena KD, Gerba CP. Risk Assessment of Pseudomonas aeruginosa in Water. Toxicology. 2009. pp. 71Œ115. doi:10.1007/978-1-4419-0032-6_3 167. Wall D, Kaiser D. Type IV pili and cell motility. Mol Microbiol. 1999;32: 1Œ10. doi:10.1111/j.1365-2958.2003.03977.x 168. Lillehoj EP, Kim BT, Kim KC. Identification of Pseudomonas aeruginosa flagellin as an adhesin for Muc1 mucin. Am J Physiol Lung Cell Mol Physiol. 2002;282: L751Œ6. doi:10.1152/ajplung.00383.2001 169. Jeffries JL, Jia J, Choi W, Choe S, Miao J, Xu Y, et al. Pseudomonas aeruginosa pyocyanin modulates mucin glycosylation with sialyl-Lewisx to increase binding to airway epithelial cells. Mucosal Immunol. Society for Mucosal Immunology; 2015; Available: http://dx.doi.org/10.1038/mi.2015.119 170. Mahajan VS, Pillai S. Sialic acids and autoimmune disease. Immunol Rev. 2016;269: 145Œ161. doi:10.1111/imr.12344 171. Lillehoj EP, Hyun SW, Feng C, Zhang L, Liu A, Guang W, et al. NEU1 sialidase expressed in human airway epithelia regulates epidermal growth factor receptor (EGFR) and MUC1 protein signaling. J Biol Chem. 2012;287: 8214Œ31. doi:10.1074/jbc.M111.292888 172. Berger M. Lung inflammation early in cystic fibrosis: Bugs are indicted, but the defense is guilty. Am J Respir Crit Care Med. 2002;165: 857Œ858. doi:10.1164/rccm.2202030 173. Driscoll J a, Brody SL, Kollef MH. The Epidemiology , Pathogenesis and Treatment of Pseudomonas aeruginosa Infections. 2007;67: 351Œ368. 174. Stuart RK, Pollack M. Pseudomonas aeruginosa Exotoxin A Inhibits Proliferation of Human Bone Marrow Progenitor Cells In Vitro. 1982;38: 206Œ211. 175. Azghani AO. Pseudomonas aeruginosa and Epithelial Permeability: Role of Virulence Factors Elastase and Exotoxin A. Am J Respir Cell Mol Biol. 1996;15: 132Œ140. doi:10.1165/ajrcmb.15.1.8679217 176. Govan JR, Deretic V. Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol Rev. 1996;60: 539Œ574. doi:10.1111/j.1365-2672.2007.03706.x 177. Toder DS, Ferrell SJ, Nezezon JL, Rust L, Iglewski BH. aeruginosa: analysis of transcription and gene lasA and lasB Genes of Pseudomonas aeruginosa: Analysis of Transcription and Gene Product Activity. 1994;62: 1320Œ1327. 178. Kessler E, Safrin M, Abrams WR, Rosenbloom J, Ohman DE. Inhibitors and Specificity of Pseudomonas-Aeruginosa Lasa. J Biol Chem. 1997;Vol 272: 9884Œ9889. doi:10.1074/jbc.272.15.9884 39 179. Grande KK, Gustin JK, Kessler E, Ohman DE. Identification of critical residues in the propeptide of LasA protease of Pseudomonas aeruginosa involved in the formation of a stable mature protease. J Bacteriol. 2007;189: 3960Œ3968. doi:10.1128/JB.01828-06 180. Mann EE, Wozniak DJ. Pseudomonas biofilm matrix composition and niche biology. FEMS Microbiol Rev. 2012;36: 893Œ916. doi:10.1111/j.1574-6976.2011.00322.x 181. Park AJ, Murphy K, Krieger JR, Brewer D, Taylor P, Habash M, et al. A temporal examination of the planktonic and biofilm proteome of whole cell Pseudomonas aeruginosa PA01 using quantitative mass spectrometry. Mol Cell Proteomics. 2014;13: 1095Œ105. doi:10.1074/mcp.M113.033985 182. Høiby N, Krogh Johansen H, Moser C, Song Z, Ciofu O, Kharazmi A. Pseudomonas aeruginosa and the in vitro and in vivo biofilm mode of growth. Microbes Infect. 2001;3: 23Œ35. doi:10.1016/S1286-4579(00)01349-6 183. Darch SE, West S a., Winzer K, Diggle SP. Density-dependent fitness benefits in quorum-sensing bacterial populations. Proc Natl Acad Sci. 2012;109: 8259Œ8263. doi:10.1073/pnas.1118131109 184. Nadal Jimenez P, Koch G, Thompson J a., Xavier KB, Cool RH, Quax WJ. The Multiple Signaling Systems Regulating Virulence in Pseudomonas aeruginosa. Microbiol Mol Biol Rev. 2012;76: 46Œ65. doi:10.1128/MMBR.05007-11 185. Balasubramanian D, Schneper L, Kumari H, Mathee K. A dynamic and intricate regulatory network determines Pseudomonas aeruginosa virulence. Nucleic Acids Res. 2013;41: 1Œ20. doi:10.1093/nar/gks1039 186. Hay ID, Wang Y, Moradali MF, Rehman ZU, Rehm BH a. Genetics and regulation of bacterial alginate production. Environ Microbiol. 2014;16: 2997Œ3011. doi:10.1111/1462-2920.12389 187. Worlitzsch D, Tarran R, Ulrich M, Schwab U, Cekici A, Meyer KC, et al. Effects of reduced mucus oxygen concentration in airway Pseudomonas infections of cystic fibrosis patients. J Clin Invest. 2002;109: 317Œ325. doi:10.1172/JCI200213870 188. Fauvart M, de Groote VN, Michiels J. Role of persister cells in chronic infections: Clinical relevance and perspectives on anti-persister therapies. J Med Microbiol. 2011;60: 699Œ709. doi:10.1099/jmm.0.030932-0 189. Lewis K. Persister cells. Annu Rev Microbiol. 2010;64: 357Œ372. doi:10.1146/annurev.micro.112408.134306 190. Lister PD, Wolter DJ, Hanson ND. Antibacterial-resistant Pseudomonas aeruginosa: Clinical impact and complex regulation of chromosomally encoded resistance mechanisms. Clin Microbiol Rev. 2009;22: 582Œ610. doi:10.1128/CMR.00040-09 191. Livermore DM. Multiple Mechanisms of Antimicrobial Resistance in Pseudomonas 40 aeruginosa: Our Worst Nightmare? Clin Infect Dis. 2002;34: 634Œ640. doi:10.1086/338782 192. Denton M, Kerr K, Mooney L, Keer V, Rajgopal A, Brownlee K, et al. Transmission of colistin-resistant Pseudomonas aeruginosa between patients attending a pediatric cystic fibrosis center. Pediatr Pulmonol. 2002;34: 257Œ261. doi:10.1002/ppul.10166 193. Cheng K, Smyth RL, Govan JR, Doherty C, Winstanley C, Denning N, et al. Spread of beta-lactam-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. Lancet. 1996;348: 639Œ642. doi:10.1016/S0140-6736(96)05169-0 194. Carter MEK, Fothergill JL, Walshaw MJ, Rajakumar K, Kadioglu A, Winstanley C. A subtype of a Pseudomonas aeruginosa cystic fibrosis epidemic strain exhibits enhanced virulence in a murine model of acute respiratory infection. J Infect Dis. 2010;202: 935Œ942. doi:10.1086/655781 195. Fothergill JL, Panagea S, Hart C a, Walshaw MJ, Pitt TL, Winstanley C. Widespread pyocyanin over-production among isolates of a cystic fibrosis epidemic strain. BMC Microbiol. 2007;7: 45. doi:10.1186/1471-2180-7-45 196. Bragonzi A, Paroni M, Nonis A, Cramer N, Montanari S, Rejman J, et al. Pseudomonas aeruginosa microevolution during cystic fibrosis lung infection establishes clones with adapted virulence. Am J Respir Crit Care Med. 2009;180: 138Œ145. doi:10.1164/rccm.200812-1943OC 197. Lorè NI, Cigana C, De Fino I, Riva C, Juhas M, Schwager S, et al. Cystic Fibrosis-Niche Adaptation of Pseudomonas aeruginosa Reduces Virulence in Multiple Infection Hosts. PLoS One. 2012;7: e35648. doi:10.1371/journal.pone.0035648 198. Mulcahy LR, Burns JL, Lory S, Lewis K. Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol. 2010;192: 6191Œ6199. doi:10.1128/JB.01651-09 199. Sauer K, Cullen MC, Rickard a H, Zeef L a H, Gilbert P, Davies DG. Characterization of Nutrient-Induced Dispersion in Pseudomonas aeruginosa PA01 Biofilm. J Bacteriol. 2004;186: 7312Œ7326. doi:10.1128/JB.186.21.7312 200. Mavrodi D V, Bonsall RF, Delaney SM, Soule MJ, Phillips G, Thomashow LS. Functional Analysis of Genes for Biosynthesis of Pyocyanin and Phenazine-1-Carboxamide from Pseudomonas aeruginosa PA01. J Bacteriol. 2001;183: 6454Œ6465. doi:10.1128/JB.183.21.6454 201. Clark ST, Diaz Caballero J, Cheang M, Coburn B, Wang PW, Donaldson SL, et al. Phenotypic diversity within a Pseudomonas aeruginosa population infecting an adult with cystic fibrosis. Sci Rep. Nature Publishing Group; 2015;5: 10932. doi:10.1038/srep10932 202. Hogardt M, Heesemann J. Microevolution of pseudomonas aeruginosa to a chronic pathogen of the cystic fibrosis lung. Curr Top Microbiol Immunol. 2013;358: 91Œ118. doi:10.1007/82_2011_199 41 203. Chung JCS, Becq J, Fraser L, Schulz-Trieglaff O, Bond NJ, Foweraker J, et al. Genomic variation among contemporary Pseudomonas aeruginosa isolates from chronically infected cystic fibrosis patients. J Bacteriol. 2012;194: 4857Œ4866. doi:10.1128/JB.01050-12 204. Rau MH, Marvig RL, Ehrlich GD, Molin S, Jelsbak L. Deletion and acquisition of genomic content during early stage adaptation of Pseudomonas aeruginosa to a human host environment. Environ Microbiol. 2012;14: 2200Œ2211. doi:10.1111/j.1462-2920.2012.02795.x 205. Cramer N, Klockgether J, Wrasman K, Schmidt M, Davenport CF, Tümmler B. Microevolution of the major common Pseudomonas aeruginosa clones C and PA14 in cystic fibrosis lungs. Environ Microbiol. 2011;13: 1690Œ1704. doi:10.1111/j.1462-2920.2011.02483.x 206. Salunkhe P, Smart CHM, Morgan AW, Panagea S, Walshaw MJ, Hart A, et al. A Cystic Fibrosis Epidemic Strain of Pseudomonas aeruginosa Displays Enhanced Virulence and Antimicrobial Resistance. J Bacteriol. 2005;187: 4908Œ4920. doi:10.1128/JB.187.14.4908 207. Bragonzi A, Wiehlmann L, Klockgether J, Cramer N, Worlitzsch D, Döning G, et al. Sequence diversity of the mucABD locus in Pseudomonas aeruginosa isolates from patients with cystic fibrosis. Microbiology. 2006;152: 3261Œ3269. doi:10.1099/mic.0.29175-0 208. Jeukens J, Boyle B, Kukavica-Ibrulj I, Ouellet MM, Aaron SD, Charette SJ, et al. Comparative genomics of isolates of a Pseudomonas aeruginosa epidemic strain associated with chronic lung infections of cystic fibrosis patients. PLoS One. 2014;9: 1Œ15. doi:10.1371/journal.pone.0087611 209. Mena A, Smith EE, Burns JL, Speert DP, Moskowitz SM, Perez JL, et al. Genetic adaptation of Pseudomonas aeruginosa to the airways of cystic fibrosis patients is catalyzed by hypermutation. J Bacteriol. 2008;190: 7910Œ7917. doi:10.1128/JB.01147-08 210. Caballero JD, Clark ST, Coburn B, Zhang Y, Wang PW, Donaldson SL, et al. Selective sweeps and parallel pathoadaptation drive pseudomonas aeruginosa evolution in the cystic fibrosis lung. MBio. 2015;6: 1Œ14. doi:10.1128/mBio.00981-15 211. Darch SE, McNally A, Harrison F, Corander J, Barr HL, Paszkiewicz K, et al. Recombination is a key driver of genomic and phenotypic diversity in a Pseudomonas aeruginosa population during cystic fibrosis infection. Sci Rep. 2015;5: 7649. doi:10.1038/srep07649 212. Markussen T, Marvig RL, Gomez-Lozano M, Aanaes K, Burleigh AE, Hoiby N, et al. Environmental Heterogeneity Drives Within-Host Diversification and Evolution of Pseudomonas aeruginosa. MBio. 2014;5: e01592Œ14Œe01592Œ14. doi:10.1128/mBio.01592-14 42 213. Hoboth C, Hoffmann R, Eichner A, Henke C, Schmoldt S, Imhof A, et al. Dynamics of adaptive microevolution of hypermutable Pseudomonas aeruginosa during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2009;200: 118Œ130. doi:10.1086/599360 214. Son MS, Matthews WJ, Kang Y, Nguyen DT, Hoang TT. In vivo evidence of Pseudomonas aeruginosa nutrient acquisition and pathogenesis in the lungs of cystic fibrosis patients. Infect Immun. 2007;75: 5313Œ5324. doi:10.1128/IAI.01807-06 215. Huse HK, Kwon T, Zlosnik JEA, Speert DP, Marcotte EM, Whiteley M. Parallel evolution in Pseudomonas aeruginosa over 39,000 generations in vivo. MBio. 2010;1: e00199Œ10. doi:10.1128/mBio.00199-10.Editor 216. Harmer C, Alnassafi K, Hu H, Elkins M, Bye P, Rose B, et al. Modulation of gene expression by Pseudomonas aeruginosa during chronic infection in the adult cystic fibrosis lung. Microbiology. 2013;159: 2354Œ63. doi:10.1099/mic.0.066985-0 217. Varga JJ, Barbier M, Mulet X, Bielecki P, Bartell JA, Owings JP, et al. Genotypic and phenotypic analyses of a Pseudomonas aeruginosa chronic bronchiectasis isolate reveal differences from cystic fibrosis and laboratory strains. BMC Genomics. 2015;16: 883. doi:10.1186/s12864-015-2069-0 218. Wu X, Held K, Zheng C, Staudinger BJ, Chavez JD, Weisbrod CR, et al. Dynamic Proteome Response of Pseudomonas aeruginosa to Tobramycin Antibiotic Treatment. Mol Cell Proteomics. 2015;14: 2126Œ2137. doi:10.1074/mcp.M115.050161 219. Palmer KL, Aye LM, Whiteley M. Nutritional cues control Pseudomonas aeruginosa multicellular behavior in cystic fibrosis sputum. J Bacteriol. 2007;189: 8079Œ8087. doi:10.1128/JB.01138-07 220. Schobert M, Jahn D. Anaerobic physiology of Pseudomonas aeruginosa in the cystic fibrosis lung. Int J Med Microbiol. Elsevier GmbH.; 2010;300: 549Œ556. doi:10.1016/j.ijmm.2010.08.007 221. Rau MH, Hansen SK, Johansen HK, Thomsen LE, Workman CT, Nielsen KF, et al. Early adaptive developments of Pseudomonas aeruginosa after the transition from life in the environment to persistent colonization in the airways of human cystic fibrosis hosts. Environ Microbiol. 2010;12: 1643Œ1658. doi:10.1111/j.1462-2920.2010.02211.x 222. Terry JM, Piña SE, Mattingly SJ. Environmental conditions which influence mucoid conversion Pseudomonas aeruginosa PA01. Infect Immun. 1991;59: 471Œ477. 223. Ryall B, Carrara M, Zlosnik JE a, Behrends V, Lee X, Wong Z, et al. The mucoid switch in pseudomonas aeruginosa represses quorum sensing systems and leads to complex changes to stationary phase virulence factor regulation. PLoS One. 2014;9. doi:10.1371/journal.pone.0096166 43 224. Didelot X, Walker AS, Peto TE, Crook DW, Wilson DJ. Within-host evolution of bacterial pathogens. Nat Rev Microbiol. Nature Publishing Group; 2016;14: 150Œ162. doi:10.1038/nrmicro.2015.13 225. Hogardt M, Hoboth C, Schmoldt S, Henke C, Bader L, Heesemann J. Stage-specific adaptation of hypermutable Pseudomonas aeruginosa isolates during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2007;195: 70Œ80. doi:10.1086/509821 226. Rodríguez-Rojas A, Oliver A, Blázquez J. Intrinsic and environmental mutagenesis drive diversification and persistence of pseudomonas aeruginosa in chronic lung infections. J Infect Dis. 2012;205: 121Œ127. doi:10.1093/infdis/jir690 227. Maier T, Güell M, Serrano L. Correlation of mRNA and protein in complex biological samples. FEBS Lett. Federation of European Biochemical Societies; 2009;583: 3966Œ3973. doi:10.1016/j.febslet.2009.10.036 228. Haider S, Pal R. Integrated analysis of transcriptomic and proteomic data. Curr Genomics. 2013;14: 91Œ110. doi:10.2174/1389202911314020003 229. Maier T, Schmidt A, Güell M, Kühner S, Gavin A-C, Aebersold R, et al. Quantification of mRNA and protein and integration with protein turnover in a bacterium. Mol Syst Biol. 2011;7: 511. doi:10.1038/msb.2011.38 230. Lacroix V, Cottret L, Thebault P, Sagot M-F. An Introduction to Metabolic Networks and Their Structural Analysis. IEEE/ACM Trans Comput Biol Bioinforma. 2008;5: 594Œ617. doi:10.1109/TCBB.2008.79 231. Bailey JE. Toward a science of metabolic engineering. Science (80- ). 1991;252: 1668Œ1675. doi:10.1126/science.2047876 232. Stephanopoulos G. Metabolic fluxes and metabolic engineering. Metab Eng. 1999;1: 1Œ11. doi:10.1006/mben.1998.0101 233. Ruppin E, Papin JA, de Figueiredo LF, Schuster S. Metabolic reconstruction, constraint-based analysis and game theory to probe genome-scale metabolic networks. Curr Opin Biotechnol. Elsevier Ltd; 2010;21: 502Œ510. doi:10.1016/j.copbio.2010.07.002 234. Wiechert W, de Graaf a a. Bidirectional reaction steps in metabolic networks: I. Modeling and simulation of carbon isotope labeling experiments. Biotechnol Bioeng. 1997;55: 101Œ17. doi:10.1002/(SICI)1097-0290(19970705)55:1<101::AID-BIT12>3.0.CO;2-P 235. Wiechert W, Siefke C, de Graaf a a, Marx a. Bidirectional reaction steps in metabolic networks: II. Flux estimation and statistical analysis. Biotechnol Bioeng. 1997;55: 118Œ135. doi:10.1002/(SICI)1097-0290(19970705)55:1<118::AID-BIT13>3.0.CO;2-I 236. HOLZHUTTER HG. COMPARTMENTAL ANALYSIS - THEORETICAL ASPECTS AND APPLICATION. Biomed Biochim Acta. MUHLENSTRASSE 33-34, D-13187 BERLIN, GERMANY: AKADEMIE VERLAG GMBH; 1985;44: 863Œ873. 44 237. Wiechert W, De Graaf A a. In vivo stationary flux analysis by 13C labeling experiments. Adv Biochem Eng Biotechnol. 1996;54: 109Œ54. 238. Wiechert W, Möllney M, Isermann N, Wurzel M, De Graaf A a. Bidirectional reaction steps in metabolic networks: III. Explicit solution and analysis of isotopomer labeling systems. Biotechnol Bioeng. 1999;66: 69Œ85. doi:10.1002/(SICI)1097-0290(1999)66:2<69::AID-BIT1>3.0.CO;2-6 239. Zupke C, Stephanopoulos G. Modeling of Isotope Distributions and Intracellular Fluxes in Metabolic Networks Using Atom Mapping Matrixes. Biotechnol Prog. 1994;10: 489Œ498. doi:10.1021/bp00029a006 240. Zamboni N, Fendt S-M, Rühl M, Sauer U. (13)C-based metabolic flux analysis. Nat Protoc. 2009;4: 878Œ892. doi:10.1038/nprot.2009.58 241. García Martín H, Kumar VS, Weaver D, Ghosh A, Chubukov V, Mukhopadhyay A, et al. A Method to Constrain Genome-Scale Models with 13C Labeling Data. Maranas CD, editor. PLOS Comput Biol. 2015;11: e1004363. doi:10.1371/journal.pcbi.1004363 242. Ghosh A, Nilmeier J, Weaver D, Adams PD, Keasling JD, Mukhopadhyay A, et al. A peptide-based method for 13C Metabolic Flux Analysis in microbial communities. PLoS Comput Biol. 2014;10: e1003827. doi:10.1371/journal.pcbi.1003827 243. Heux S, Poinot J, Massou S, Sokol S, Portais J-C. A novel platform for automated high-throughput fluxome profiling of metabolic variants. Metab Eng. 2014;25: 8Œ19. doi:10.1016/j.ymben.2014.06.001 244. McAtee AG, Jazmin LJ, Young JD. Application of isotope labeling experiments and 13C flux analysis to enable rational pathway engineering. Curr Opin Biotechnol. Elsevier Ltd; 2015;36: 50Œ56. doi:10.1016/j.copbio.2015.08.004 245. Adebiyi AO, Jazmin LJ, Young JD. 13C flux analysis of cyanobacterial metabolism. Photosynth Res. Springer Netherlands; 2015;126: 19Œ32. doi:10.1007/s11120-014-0045-1 246. Götz A, Eylert E, Eisenreich W, Goebel W. Carbon metabolism of enterobacterial human pathogens growing in epithelial colorectal adenocarcinoma (Caco-2) cells. PLoS One. 2010;5: 8Œ9. doi:10.1371/journal.pone 247. Beste DJ V, Nöh K, Niedenführ S, Mendum T a., Hawkins ND, Ward JL, et al. 13C-flux spectral analysis of host-pathogen metabolism reveals a mixed diet for intracellular mycobacterium tuberculosis. Chem Biol. 2013;20: 1012Œ1021. doi:10.1016/j.chembiol.2013.06.012 248. Bücker R, Heroven AK, Becker J, Dersch P, Wittmann C. The Pyruvate-Tricarboxylic Acid Cycle Node. J Biol Chem. 2014;289: 30114Œ30132. doi:10.1074/jbc.M114.581348 249. Berger A, Dohnt K, Tielen P, Jahn D, Becker J, Wittmann C. Robustness and plasticity of metabolic pathway flux among uropathogenic isolates of Pseudomonas aeruginosa. PLoS One. 2014;9: e88368. doi:10.1371/journal.pone.0088368 45 250. Orth JD, Thiele I, Palsson BØ. What is flux balance analysis? Nat Biotechnol. 2010;28: 245Œ248. doi:10.1038/nbt.1614.What 251. Lewis NE, Nagarajan H, Palsson BO. Constraining the metabolic genotype-phenotype relationship using a phylogeny of in silico methods. Nat Rev Microbiol. 2012;10: 291Œ305. doi:doi:10.1038/nrmicro2737 252. Karr JR, Takahashi K, Funahashi A. The principles of whole-cell modeling. Curr Opin Microbiol. Elsevier Ltd; 2015;27: 18Œ24. doi:10.1016/j.mib.2015.06.004 253. Hyduke D, Hyduke D, Schellenberger J, Que R, Fleming R, Thiele I, et al. COBRA Toolbox 2.0. Protoc Exch. 2011; doi:10.1038/protex.2011.234 254. Mahadevan R, Schilling CH. The effects of alternate optimal solutions in constraint-based genome-scale metabolic models. Metab Eng. 2003;5: 264Œ276. doi:10.1016/j.ymben.2003.09.002 255. Kim HU, Kim SY, Jeong H, Kim TY, Kim JJ, Choy HE, et al. Integrative genome-scale metabolic analysis of Vibrio vulnificus for drug targeting and discovery. Mol Syst Biol. Nature Publishing Group; 2014;7: 460Œ460. doi:10.1038/msb.2010.115 256. Song C, Chiasson M a, Nursimulu N, Hung SS, Wasmuth J, Grigg ME, et al. Metabolic reconstruction identifies strain-specific regulation of virulence in Toxoplasma gondii. Mol Syst Biol. Nature Publishing Group; 2013;9: 708. doi:10.1038/msb.2013.62 257. Lee D-Y, Chung BKS, Yusufi FNK, Selvarasu S. In silico genome-scale modeling and analysis for identifying anti-tubercular drug targets. Drug Dev Res. 2011;72: 121Œ129. doi:10.1002/ddr.20408 258. Wang C, Deng Z-L, Xie Z-M, Chu X-Y, Chang J-W, Kong D-X, et al. Construction of a genome-scale metabolic network of the plant pathogen Pectobacterium carotovorum provides new strategies for bactericide discovery. FEBS Lett. Federation of European Biochemical Societies; 2014;589: 285Œ94. doi:10.1016/j.febslet.2014.12.010 259. Lee D-S, Burd H, Liu J, Almaas E, Wiest O, Barabasi A-L, et al. Comparative Genome-Scale Metabolic Reconstruction and Flux Balance Analysis of Multiple Staphylococcus aureus Genomes Identify Novel Antimicrobial Drug Targets. J Bacteriol. 2009;191: 4015Œ4024. doi:10.1128/JB.01743-08 260. Veith N, Solheim M, van Grinsven KWA, Olivier BG, Levering J, Grosseholz R, et al. Using a Genome-Scale Metabolic Model of Enterococcus faecalis V583 To Assess Amino Acid Uptake and Its Impact on Central Metabolism. Appl Environ Microbiol. 2015;81: 1622Œ33. doi:10.1128/AEM.03279-14 261. Kim HU, Kim TY, Lee SY. Genome-scale metabolic network analysis and drug targeting of multi-drug resistant pathogen Acinetobacter baumannii AYE. Mol Biosyst. 2010;6: 339Œ348. doi:10.1039/b916446d 46 262. Larocque M, Chénard T, Najmanovich R. A curated C. difficile strain 630 metabolic network: prediction of essential targets and inhibitors. BMC Syst Biol. 2014;8: 117. doi:10.1186/s12918-014-0117-z 263. Oberhardt M a, Puchaka J, Fryer KE, Martins dos Santos V a P, Papin J a. Genome-scale metabolic network analysis of the opportunistic pathogen Pseudomonas aeruginosa PA01. J Bacteriol. 2008;190: 2790Œ2803. doi:10.1128/JB.01583-07 264. Oberhardt M a., Goldberg JB, Hogardt M, Papin J a. Metabolic network analysis of Pseudomonas aeruginosa during chronic cystic fibrosis lung infection. J Bacteriol. 2010;192: 5534Œ5548. doi:10.1128/JB.00900-10 265. Sigurdsson G, Fleming RMT, Heinken A, Thiele I. A systems biology approach to drug targets in pseudomonas aeruginosa biofilm. PLoS One. 2012;7: 1Œ9. doi:10.1371/journal.pone.0034337 266. Schuetz R, Kuepfer L, Sauer U. Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. Mol Syst Biol. 2007;3. doi:10.1038/msb4100162 267. Montezano D, Meek L, Gupta R, Bermudez LE, Bermudez JCM. Flux Balance Analysis with Objective Function Defined by Proteomics Data-Metabolism of Mycobacterium tuberculosis Exposed to Mefloquine. PLoS One. 2015;10: e0134014. doi:10.1371/journal.pone.0134014 268. Toya Y, Ishii N, Nakahigashi K, Hirasawa T, Soga T, Tomita M, et al. 13C-Metabolic flux analysis for batch culture of Escherichia coli and its pyk and pgi gene knockout mutants based on mass isotopomer distribution of intracellular metabolites. Biotechnol Prog. 2010;26: 975Œ992. doi:10.1002/btpr.420 269. Wu C, Xiong W, Dai J, Wu Q. Genome-Based Metabolic Mapping and 13 C Flux Analysis Reveal Systematic Properties of an Oleaginous Microalga Chlorella protothecoides. Plant Physiol. 2015;167: 586Œ599. doi:10.1104/pp.114.250688 270. Williams TCR, Poolman MG, Howden AJM, Schwarzlander M, Fell DA, Ratcliffe RG, et al. A Genome-Scale Metabolic Model Accurately Predicts Fluxes in Central Carbon Metabolism under Stress Conditions. Plant Physiol. 2010;154: 311Œ323. doi:10.1104/pp.110.158535 271. Chen X, Alonso AP, Allen DK, Reed JL, Shachar-Hill Y. Synergy between 13C-metabolic flux analysis and flux balance analysis for understanding metabolic adaption to anaerobiosis in E. coli. Metab Eng. Elsevier; 2011;13: 38Œ48. doi:10.1016/j.ymben.2010.11.004 47 CHAPTER 2 Œ Metabolic Flux Analyses of Pseudomonas aeruginosa Cystic Fibrosis Isolates** This manuscript was submitted to Metabolic Engineering Journal in March 2016 (Opperman and Shachar-Hill); it is currently under review. 48 ABSTRACT Pseudomonas aeruginosa is a metabolically versatile wide-ranging opportunistic pathogen. In humans P. aeruginosa causes infections of the skin, urinary tract, blood, and the lungs of Cystic Fibrosis patients. In addition, P. aeruginosa™s broad environmental distribution, relatedness to biotechnologically useful species, and ability to form biofilms have made it the focus of considerable interest. We used 13C Metabolic Flux Analysis and Flux Balance Analysis to understand energy and redox production and consumption and to explore the metabolic phenotypes of one reference strain and five strains isolated from the lungs of Cystic Fibrosis patients. Our results highlight the importance of the Oxidative Pentose Phosphate and Entner-Doudoroff pathways in P. aeruginosa growth. Among clinical strains, we report two divergent metabolic strategies and identify changes between genetically related strains that have emerged during a chronic infection of the same patient. MFA revealed that the magnitude of fluxes through the glyoxylate cycle correlates with growth rates. 49 INTRODUCTION Pseudomonas aeruginosa is a gram negative, environmentally widespread, opportunistically pathogenic bacterium; it causes infections, growth inhibition and death in organisms as diverse as amoebas, fungi, plants, nematodes and mammals [1Œ4]. In humans this organism causes infections of the skin, urinary tract, blood, and lung; it is among the most frequent and harmful causes of hospital-acquired infections [5Œ7]. P. aeruginosa, which forms biofilms in many environmental and pathogenic situations, is metabolically versatile [8] and intrinsically resistant to many antibiotics; it develops further resistance during chronic infections resulting in treatment failure [9Œ11]. P. aeruginosa infections of the lungs of Cystic Fibrosis (CF) patients are of particular concern. CF is a genetic disease involving defects in the Cystic Fibrosis Transmembrane conductance Regulator (CFTR) protein that affects an estimated 70,000 individuals worldwide [12Œ17]. Defective membrane transport leads to pancreatic insufficiency, diabetes mellitus, bronchiectasis, and chronic bacterial infection of the pulmonary system in adulthood [18,19]. While new treatments have begun to change how some CF sub-types are treated (e.g. ivacaftor; [20]), most adults with CF still face chronic bacterial infection, especially by P. aeruginosa, leading to the development of highly resistant strains and to the clinical failure of pulmonary treatment, chronic inflammation and progressive damage to the lung, pulmonary failure, and eventual lung transplantation or death in mid-adulthood [5,9,11,12,21,22]. While many studies of P. aeruginosa as a pathogen have focused on identifying genetic changes during chronic infection, and on the production of biofilms and virulence factors, much less is known about the system-wide metabolic phenotypes of this and other pathogenic microbes or their physiological adaptations during chronic infections [5,7,23Œ32]. Omic studies of 50 pathogenic P. aeruginosa strains have described changes at the genomic [33Œ41], transcriptomic [41Œ46], and proteomic [47,43,48] levels. Such studies have demonstrated numerous changes in the cellular inventory during evolution in the lung environment. To identify which of these many changes are significant for pathogenesis and to measure and predict functional changes in metabolism, a toolbox of network based computational and experimental methods is available. In recent years, these tools have begun to be applied to pathogenic microbes, including P. aeruginosa. Constraints-based Flux Balance Analysis (FBA) uses the structure of the metabolic network and the stoichiometries of the reactions of which they are composed to build computational models of metabolism (see [49Œ51] for recent reviews). Such models are used to investigate the potential flows of carbon and other elements as well as cofactor balances and can extend to genome-wide coverage [50,52]. In addition to identifying reactions and conditions essential for growth and improving gene annotations, FBA and related tools allow the prediction of maximal growth rates and the exploration of predicted metabolic flux distributions under the assumption of different optimization strategies (fiobjective functionsfl, most commonly maximal growth efficiency). FBA investigations of pathogenic organisms have been used to search for novel drug targets and have pointed to potential metabolic targets not affected by current therapeutics, such as amino acid production or fatty acid metabolism [53Œ60]. Oberhardt and colleagues have constructed a genome-based model of metabolism and transport in P. aeruginosa [61], and used Flux Balance Analysis (FBA) with transcript data from two CF clinical strains to investigate P. aeruginosa™s metabolic capabilities and potential metabolic changes during prolonged infection [62]. FBA has also been used to identify potential metabolic drug targets during biofilm growth of P. aeruginosa [63]. To map network-wide metabolic 51 fluxes without assumptions about the strategies or objectives being pursued during cellular metabolism, isotopic labeling experiments are used together with metabolic network models [64]. Metabolic Flux Analysis (MFA) combines 13C labeling results with the growth and uptake measurements used for FBA to yield estimates of carbon fluxes through central metabolism [65Œ68]. Most MFA studies of microbial systems to date have focused on questions related to biotechnology and metabolic engineering, microbial physiology, and gene function [69Œ73]. Several recent studies have analyzed the metabolic interactions of pathologically relevant bacteria within their host environment [74,75] and metabolic differences between mutant strains [76]. A recent study of pathogenesis related metabolism in P. aeruginosa employed MFA to describe 17 uropathogenic strains [77]. The combination of FBA and MFA allows the testing of alternative objective functions, and can identify the origins of sub-maximal growth rates. However there have been surprisingly few studies in which these complementary network flux analysis approaches have been combined [70,78Œ85]. Here we performed FBA analysis together with 13C MFA of one reference and five selected CF pathogenic strains of P. aeruginosa to understand energy and redox production and consumption processes and to explore metabolic (in)efficiencies and metabolic phenotypes. Our results highlight the importance of the Oxidative Pentose Phosphate Pathway (OPPP) and Entner-Doudoroff Pathway (EDP) in P. aeruginosa growth and point to a substantially lower flux around the tricarboxylic acid pathway than has been previously reported for the same reference strain under similar conditions. The differences were explained by the smaller experimental dataset previously used. Among the clinical strains we report two metabolic strategies and identify changes between genetically related strains that have emerged 52 during the course of a chronic infection of the same patient. MFA identified fluxes through the glyoxylate cycle, whose magnitude correlated with growth rates across strains, although this pathway is not predicted by FBA to increase growth efficiency. Finally, we compared the degree of relatedness of metabolism among strains obtained in two ways: in the first approach 13C labeling data alone was used, and in the second the fluxes obtained by MFA were compared. While both approaches clearly discriminated among the strains, the patterns of relatedness revealed by labeling data did not correspond to those apparent from the flux analyses. MATERIALS AND METHODS Strain selection and culture Pseudomonas aeruginosa strains are described in Table 2.1. Strains were chosen that have sequenced genomes, clinical and epidemiological importance, and to include an example of an ancestral and descendant strains [86]. Strains AA2, AA43, AA44, and LES 400 were obtained from the Belgian Co-ordinated Collections of Micro-organisms (Ghent, Belgium; LMG numbers: 27630, 27631, 27632, and 27623 respectively). Pseudomonas aeruginosa strain AMT 0023-30 was obtained from Cystic Fibrosis Isolate Core (Seattle, Washington). Pseudomonas aeruginosa strain PA01 was obtained from Dr. M. Mulks at Michigan State University. M9 defined minimal media with 22 mmol glucose as the sole carbon source was used as the culture media for all experiments [87]. All cultures were pre-cultured from -80°C frozen stock samples onto M9 media solidified with 1.5% agarose for 24 hours at 37°C. Single colonies were transferred into 30ml of M9 liquid media in 250 ml Bellco triple baffled shake flasks, incubated at 37°C and shaken at 120 RPM until stationary phase (12 to 14 hours) before spectrophotometric measurement of cell density as absorbance at 600nm (OD600) and inoculation 53 of experimental cultures to initial densities of 0.01 OD600. For 13C-labeling experiments cells were cultured using either 99.9% (mol/mol) [1-13C] glucose or 20% (mol/mol) [U-13C] glucose. Cell and media sample harvest Cell pellets were collected by centrifugation of cultures at mid-log phase (OD6000.5). 1 ml of culture was centrifuged at 15000 g for 5 minutes. The supernatant was removed and 100 l of 6N HCl was added to the pellet, which was stored at -20°C. Cell pellet sample preparation for amino acid analysis was based on previously reported methods [78,88]. Determination of substrate uptake rates and product secretion Glucose uptake rates for each strain were determined from 1mL culture supernatants using at least three biological replicate samples for each of three log phase time points taken at 15 minute intervals. After lyophilization, samples were resuspended in 600 L of 99% D2O, and then lyophilized and resuspended in 600 L of 99.9% D2O. 1H-NMR was performed on an Agilent DirectDrive2 500MHz instrument using an Agilent OneNMR Probe with Protune for auto-tuning. Spectra were obtained at 500 MHz with a pulse angle of <45o, acquisition time of2.05 s, and recycle delay of 2 s. Glucose concentrations relative to the initial 22mM were determined with reference to 10 mM methylphosphamide added as internal standard. No secreted products were detected at significant levels in the media after culture growth for any of the strains. Calculation of glucose uptake rates during exponential growth used the following equation: Where is intrinsic growth rate (hr-1); C is the concentration of glucose (mM) at time t; Xi is the initial density (g·DW/L) of cells; t is time (hours); and v is uptake flux (mmol/g·DW/hr). 54 Table 2.1. Pseudomonas aeruginosa strains analyzed. LPS, Lipopolysaccharide; PGN, polygalacturan. N/A, not applicable. Strain Origin Related Isolates Relevant Characteristics References PA01 Australia Wound Isolate (1955) N/A Standard Reference Strain. [7,30,32] AMT 0023-30 U.S.A. CF - Pediatric Clinical Isolate (~1998) AMT 0023-34 Persister Cells Present [31,39] LES 400 U.K. CF - Adult Clinical Isolate (2003) Liverpool Epidemic Strains (LES) Increased Acute Morbidity Risk Decreased Life Expectancy with Infection Pyrocyanin Overproduction Alginate Overproduction CF Transmissible [24,25,27,37,89] AA2 Germany CF Œ Adult Clinical Sequential Isolates (~1998 & ~2003) AA43, AA44 Parental strain of AA43, AA44 [23,29,33] AA43 AA2, AA44 7.5 years total lung colonization time Increased Acute Morbidity risk Motility defect Protease reduction LPS and PGN changes AA44 AA2, AA43 55 Determination of CO2 efflux rates Cultures were grown, as described above, to early log phase that were sealed with gas-tight caps and incubated for two hours. Due to the high sensitivity and low maximum range of CO2 detection of the LICOR CO2 measurement, P. aeruginosa growth was limited to a low range of optical density for this measurement. The cell density of the culture was measured and growth was stopped by adding 1 mL 6N HCl to the culture, which also converted dissolved bicarbonate to CO2. Total CO2 was measured using a LICOR LI-6400with dryCO2 free air as input into the culture flask at a rate of 500l/s. Readings of total CO2 were recorded over 5 minutes (1s time resolution) to allow for removal of CO2 from the flask. Total CO2 evolved was measured in at least three replicate cultures with CO2 levels integrated for total CO2 efflux. Due to the very small change in OD during the CO2 measurement period, CO2 efflux was calculated from total CO2 evolved per change in time per total change in dry-weight in grams. Analysis of Amino Acid labeling Cell pellets were suspended in 1 mL of 6N HCl and incubated at 100°C for 24 hours, and dried at 60°C under a stream of N2. Amino acids were derivatized using N-Methyl-N-[tert-butyldimethyl-silyl] trifluoroacetimide (MTBSTFA, SIGMA-ALDRICH) in a 25l pyrimidine and 25l 1% MTBSTFA solution incubated at 40°C for 1hour as previously described [78,88,90]. GC-MS analyses of derivatized amino acids were performed on an Agilent 5973GC/quadrupole MS. GC-MS signals were corrected for natural isotope abundance [88]. Amino acid fragments used in 13C MFA were based on reported reliability [90]. Unlabeled amino acid samples collected were used to confirm the accuracy of natural abundance correction. GC-MS amino acid data was corrected for natural abundance isotopic contents with average values and standard deviation calculated for biological replicates used as model inputs. 56 Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA) GC-MS measured mass fragment mass abundance data for amino acids of strains harvested after growth in 100% 1-13C glucose or 20% uniformly labeled (1,2,3,4,5,6-13C) glucose was collected. Amino acid fragments that were measureable in three biological replicates were then processed in MATLAB (2010a, Mathworks, Natick MA, U.S.A.) using the PCA function to produce a principal component analysis [91]. HCA was performed on this data set as well as on fluxes determined by MFA that were normalized to each strain's glucose uptake rate. HCA was implemented in the programming language R using complete linkage criteria using Euclidian distance [92]. 13C Metabolic Flux Analysis A carbon isotopic network of P. aeruginosa central metabolism was constructed by an approach similar to that previously used for E. coli [78]. The iM01056 FBA model [61] reaction network was simplified by only including carbon atoms of reactants and products, by introducing both net and exchange fluxes for reversible fluxes, by combining linear reaction sequences that do not alter carbon positions, and by condensing secondary metabolic pathways into a growth equation that consumes central metabolic intermediates as precursors to produce cellular biomass as previously defined [61,93]. By preserving the architecture of the FBA model, results of FBA and 13C MFA could be directly compared, and 13C MFA results examined by FBA [78]. Measured rates of glucose uptake, CO2 efflux, and 13C labeling patterns in amino acids together with growth rates were used to estimate internal fluxes of central carbon metabolism by fitting flux values to the experimental data using the 13C-FLUX software as previously described [67,68,78]. For most labeling data, variation among biological replicates was very small (see Supplemental 2.1); since experimental standard deviations (SDs) are not a reliable estimate of 57 true population SDs when 5 or fewer replicates are analyzed, experimentally observed SD™s with low SD™s were increased to 2% of mean values for MFA modeling as previously described [78]. This also avoids excessively constraining modeling results to labeling data at the expense of direct flux measurements and reduces the potential for distortions due to precise but inaccurate mass isomer quantification [78]. To allow for uncertainties in P. aeruginosa biomass compositions, production rates were constrained to be within 50% of the E. coli biomass values [78]. To minimize the risk that solutions represented local rather than global optimization minima, multiple randomly generated initial fluxes constrained by sampling the feasible solution space were used. First, at least 100 randomly generated feasible starting points that produced optimized fits by the 13C-FLUX program were found. Next, the 10 starting points that yielded the lowest final residuum values were used in the second stage to generate 1000 more starting points by randomly perturbing these starting points to yield 100 new points each. The final, lowest residuum optimized flux values that resulted in the best fit to the data were used. Confidence intervals for flux values were estimated using a Monte Carlo approach to randomly generate values of: biomass, glucose uptake rate, CO2 efflux rate, and amino acid labeling data for each strain [94] based on the experimentally determined standard deviations. At least 20 such datasets were then fitted, as described above to yield best-fit flux maps for each strain. These 20 flux value sets were then used to calculate 90% confidence intervals for each flux modeled for each strain. All MFA computations were performed using the High Performance Computer Center, Michigan State University, using a parallel 1536 core cluster of 192 nodes (two four-core Intel 58 Xeon E5620s at 2.4 GHz with 24 GB of RAM and 250 GB local disk space per node). Global sum of squared residuals (SSres) for each strain is listed in Supplemental 2.2. Flux Balance Analysis The genome-derived stoichiometric P. aeruginosa model iM01056 developed by Oberhardt et al. was used for FBA [61]. The results of PA01are representative of predicted central metabolism for all the strains studied here since their genomes do not appear to lack any functional central metabolic fluxes [23,39,41]. The model was modified to better account for lipid production (see Supplemental 2.3) resulting in a total of 1013 reactions and 875 metabolites. The COBRA Toolbox 2.0 in MATLAB (Mathworks, Natick MA, U.S.A.) implementing the Gurobi 6 optimizer was used for FBA using the objective function of maximal biomass production [52]. FBA simulations used measured glucose uptake rates by strain PA01. To determine the range of fluxes that allow 99% or 90% of the maximal growth rates, the Flux Variability Analysis (FVA) function of the COBRA Toolbox was used [95]. RESULTS The physiology, growth and yield of the reference and clinical strains were compared during growth in defined medium to characterize any intrinsic or evolved differences among them. Specific growth rates, final culture density (as a measure of carbon conversion efficiency over the full growth cycle), glucose uptake rates, and CO2 efflux rates were measured as described in Materials and Methods and are shown in Figure 2.1. The growth rates show a 40% range of values, and both glucose uptake and CO2 efflux rates show a twofold range across strains. Compared to those reported for uropathogenic P. aeruginosa strains grown in an artificial urine medium [77], strain LES 400 showed a growth rate as high as or higher than the highest previously reported (0.96 hr-1 vs 0.91 hr-1) and strain 59 AA44 showed a glucose uptake rate as low as or lower than the lowest rate reported in uropathogenic strains (4.92 mmol gDW-1 hr-1 vs 5.37 mmol gDW-1 hr-1). P. aeruginosa pathogenic strains display a large range of growth parameters under defined growth conditions pointing to divergent physiological phenotypes. By contrast, the maximum optical densities attained at stationary phase showed no statistically significant differences among strains, indicating that the growth yields over the culture period are similar. To explore whether final yields are reflected in substrate use efficiencies during growth, Carbon Conversion Efficiency (CCE) was calculated during log phase growth using biomass production rates compared to: (a) glucose uptake rates; and (b) CO2 production rates (Figure 2.2). CCE values for both methods are not significantly different for any strain, as seen from the 95% confidence intervals and confirmed by a heteroscedastic two sided t-test (p>0.05). These results are consistent with the absence of detected secreted products in culture media (closed carbon balance). Previous studies of P. aeruginosa and P. fluorescens also found no evidence for significant metabolite export during growth in defined simple media [77,96]. The values of CCE% found for these strains (59%-72%) are higher than the 52% reported for E. coli growing under the same conditions [97]. In that case the CCE is lowered by acetate secretion, but the range here is lower than the estimated 86% for P. fluorescens (as calculated from reported MFA results [96]). To assess whether the diverse pathogenic strains have evolved divergent metabolic phenotypes, 13C labeling data for amino acids from cultures grown to steady state with labeled glucose, were analyzed by Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA). Both analyses (Figure 2.3) demonstrate clear differences between strains, with 70% of the variation of the amino acid label data contained in the first three principal 60 components of the PCA. Sample labeled amino acid data used in Figure 2.3 can be found in Supplemental 2.1. Both PCA and HCA show two groups of strains, one containing AMT 0023-30 and PA01 and the other consisting of AA43, AA2, AA44, and LES 400. The 13C MFA flux map of P. aeruginosa reference strain (PA01) is shown in Figure 2.4. The MFA model, experimental data, and flux tables are given in Supplemental 2.2. The map shows that the flux through the Entner-Doudoroff Pathway (EDP) is equivalent to 60% of the glucose uptake rate. Flux through the decarboxylation step of the Oxidative Pentose Phosphate Pathway is equivalent to 85% of the glucose uptake. Published values for aerobic E. coli grown under similar conditions show a flux of 25%-27% through the OPPP compared to glucose uptake with glycolysis carrying close to 80% of the glucose uptake flux [78,98]. A substantial flux recycles carbon from the OPPP to hexose-monophosphates so that the sum of these fluxes (the total flux from hexose-6-phophate to 6-phosphogluconate) is higher than the glucose uptake rate. Escherichia coli studies have also reported cyclic OPPP fluxes under these conditions [78,98]. Little or no net flux was estimated to occur from triose-phosphate to hexose-6-phosphate via the reversible steps of glycolysis. Also noteworthy is the low tri-carboxylic acid (TCA) cycle flux under these aerobic conditions, with significant fluxes estimated though the glyoxylate cycle and anapleurosis from phosphoenolpyruvate carboxylase. Confidence interval (CI) calculations of the MFA model indicate that the fluxes are well estimated by the data, with 90% CIs for net fluxes limited to +/-10% of flux values in the EDP and OPPP and close to +/-20% of fluxes in the TCA cycle. 61 Figure 2.1. Growth physiology. (A) Specific growth rates (n=5, AA43 n=4); (B) Final cell densities (OD600, n=3); (C) Glucose uptake rates (AA44 n=7, LES400 n=3, AMT0023-3 n=6, others n=5), and (D) CO2 efflux rates (AMT0023-30 n=4, AA2 n=5, others n=3). Error bars represent 95% confidence intervals. Horizontal bars in (A), (C), and (D) connect strains that do not show significant difference (0.05) under two-sided t-tests, [no significant differences among strains in (B)]. 62 Figure 2.2. Carbon conversion efficiency during growth. Filled bars show values calculated from glucose uptake and biomass production; Empty bars show values calculated from CO2 efflux and biomass production. Error bars represent 95% confidence intervals. (CCE% Glucose for LES400 n=4; CCE % CO2 AA2 n=5, AMT0023-30 n=4; others n=3) Figure 2.3. Discrimination among strains using the isotopomers of proteinogenic amino acids and their fragments. Steady state labeling was measured in three biological replicates from 100% 1-13C and 20% U6-13C Glucose experiments (see materials and methods). Strains: AMT0023AA43 (squares), AA44 (diamonds), LES 400 (pentagrams), and PA01 (hexagrams).(Analysis (HCA, see materials and methods) of the same data used in Figure 2.3A. 63 Discrimination among strains using 13C labeling fingerprinting. (Figure 2.3A - Left) Principal Component Analysis of no acids and their fragments. Steady state labeling was measured in three biological replicates C Glucose experiments (see materials and methods). Strains: AMT0023-30 (circles), AA2 (triangles), LES 400 (pentagrams), and PA01 (hexagrams).(Figure 2.3B - Right) Hierarchical Clustering Analysis (HCA, see materials and methods) of the same data used in Figure 2.3A. Principal Component Analysis of no acids and their fragments. Steady state labeling was measured in three biological replicates 30 (circles), AA2 (triangles), Right) Hierarchical Clustering 64 Flux Balance Analysis of Pseudomonas aeruginosa strain PA01 was performed using a modified genome-derived model (see methods) together with the glucose uptake rate measured for PA01. A predicted flux map based on maximum growth efficiency (maximal biomass production for the glucose uptake rate) as the objective function is shown in Figure 2.5. A listing of the predicted net fluxes for optimal growth is in the supplemental materials (Supplemental 2.3). The FBA flux map shows several differences from the MFA experimentally based map for PA01. Cyclic flux in the oxidative pentose phosphate pathway is not seen in the FBA map; instead the majority of the carbon flows through the Entner-Doudoroff pathway and the OPPP decarboxylation flux is predicted to be lower than estimated by MFA. FBA predicts a robust tri-carboxylic acid cycle flux that is not consistent with the fluxes estimated by MFA. FBA predicts maximal growth efficiency (CCE) to be 70%, compared to the 60-65% observed. To assess whether particular fluxes estimated by MFA are likely to be responsible for the submaximal CCE observed for PA01, flux variability analysis (FVA) was used to determine the ranges of fluxes consistent with 99-100% of maximal predicted. The results are shown in Figure 2.5 and Table 2.2 and predict that near-maximal growth could be sustained with a large range of individual fluxes. For example the first committed step of the OPPP (decarboxylation of6-phosphogluconate) can have 0-130% of the predicted optimal value. The MFA map for PA01 shows several net fluxes outside the range defined by FVA for near-optimal growth: (A) G3P to PEP flux is estimated by MFA at 20% of glucose uptake with a range of 49%-81%calculated by FVA;(B) pyruvate to Ac-CoA and CO2 (19% of glucose uptake in MFA vs 51-133% from VFA); and (C) PEP to pyruvate (6% vs. 49-81%). Figure 2.4. 13C MFA flux mapmedia with either 100% 1-13C or 20% Usteady state labeling during exponential growthmodeled as described in materials and methods. Arrow thicknesses are proportional to net fluxes. Numbers represent net carbon fluxes in units of c90% confidence intervals (CI). The rage of fluxes catalyzed by Aldolasedehydrogenase show the optimized model results paired with the upper range of the calculated CI90% as these fluxes are limited to 0 (irreversible). Nonabbreviations: 6PGC, 6-phosphophosphate; P5P, Pentose-5-Phosphate. 65 flux map of strain PA01. Strain PA01 was grown in defined minimal C or 20% U6-13C Glucose (n=3 replicates for each substrate) to steady state labeling during exponential growth. Labeling and external fluxes were measured and as described in materials and methods. Arrow thicknesses are proportional to net fluxes. Numbers represent net carbon fluxes in units of c-mmol flux/(gDWŁhr); values are given 90% confidence intervals (CI). The rage of fluxes catalyzed by Aldolase and alphadehydrogenase show the optimized model results paired with the upper range of the calculated CI90% as these fluxes are limited to 0 (irreversible). Non-standard or potentially ambiguous phospho-gluconate; G3P, Glyceraldehyde 3-phosphate; H6P, hexosePhosphate. Strain PA01 was grown in defined minimal C Glucose (n=3 replicates for each substrate) to . Labeling and external fluxes were measured and as described in materials and methods. Arrow thicknesses are proportional to net fluxes. flux/(gDWŁhr); values are given +/- and alpha-ketogluarate dehydrogenase show the optimized model results paired with the upper range of the calculated standard or potentially ambiguous phosphate; H6P, hexose-6-66 Figure 2.5. Flux map from FBA analysis of strain PA01. Uptake of glucose, production of CO2 and growth rate were used to constrain a stoichiometric genome-derived metabolic model (see materials and methods) with the objective function being maximal growth (corresponding to maximal growth yield on glucose). Numbers are net fluxes as in Figure 2.4. Ranges are from flux variability analysis (range of each flux consistent with 99-100% of maximal growth). 67 To quantify the metabolic differences among strains isolated from the cystic fibrosis lung environment, Metabolic Flux Analysis was conducted on the clinical strains (Table 2.1). Flux values estimated by MFA are listed in Table 2.2 (full flux lists can be found in Supplemental 2.2) and net fluxes are displayed relative to the PA01 flux values in Figure 2.6A while Figure 2.6B shows flux values relative to the uptake rate of glucose for each strain. This figure demonstrates that CO2 efflux and phosphoenolpyruvate carboxylase flux rates across all strains are within 30% of the reference strain flux even though glucose uptake rates show a variation of 49-106%. The estimated EDP and TCA cycle fluxes vary widely across strains with 2-fold and 3.5-fold ranges, respectively. The -ketoglutarate dehydrogenase flux is very low in reference strain PA01, and is highly variable across the other strains (Figure 2.6A). Finally, it was observed that the flux rate of isocitrate lyase plotted against the measured intrinsic growth rate of all strains showed the greatest linear correlation (0.599 R2) of flux to growth rate outside of the biomass related synthesis fluxes. This value increased to a R2 value of 0.982 with the removal of strain AA2™s MFA derived flux from this data set (Figure 2.7). Hierarchical Clustering Analysis (HCA) of MFA flux values normalized to measured glucose uptake rates (Figure 2.8) separated strains into two groups with distinct metabolic strategies. LES 400, AA44, and PA01 were in one group and AA2, AA43, and AMT 0023-30 were in the other. The group containing LES 400, PA01, and AA44 shows little or no decarboxylation of 2-oxoglutarate (-ketoglutarate, in the TCA cycle) and much higher OPPP fluxes in comparison to the AA2, AMT0023-30, and AA43 group. To investigate the differences that underlie these groupings, the ratios of fluxes through key pathways are shown in Table 2.3. 68 Strain PA01 LES 400 AMT 0023-30 AA2 AA43 AA44 Flux Net Flux Xch Flux Net Flux Xch Flux Net Flux Xch Flux Net Flux Xch Flux Net Flux Xch Flux Net Flux Xch Flux Glucose Uptake 9.45±0.00 N/A 10.05±0.22 N/A 9.25±0.24 N/A 6.96±0.23 N/A 5.79±0.23 N/A 4.80±0.11 N/A H6P6PGC 13.71±0.47 N/A 14.42±0.52 N/A 12.69±0.42 N/A 6.34±0.28 N/A 5.34±0.43 N/A 5.75±0.20 N/A 6PGCG3P +PYR 5.67±0.57 N/A 10.13±0.65 N/A 11.35±0.55 N/A 5.96±0.21 N/A 4.97±0.25 N/A 3.26±0.11 N/A G3PPEP 3.69±0.32 0.84±0.02 2.87±0.37 0.86±0.03 2.19±0.43 0.86±0.044.36±0.20 0.81±0.03 3.62±0.22 0.95±0.04 2.31±0.07 0.78±0.04 PEPPYR 1.14±0.31 0.56±0.03 0.75±0.29 0.57±0.03 -0.24±0.44 0.58±0.022.19±0.19 0.53±0.02 1.64±0.17 0.44±0.03 0.97±0.10 0.45±0.02 2 G3PH6P 0.00±0.45 N/A 2.85±0.54 N/A 3.44±0.51 N/A 0.00±0.00 N/A 0.00±0.00 N/A 0.00±0.04 N/A 6PGC2 R5P 8.04±0.63 N/A 4.29±0.72 N/A 1.34±0.46 N/A 0.38±0.23 N/A 0.36±0.37 N/A 2.48±0.18 N/A 2 R5PS7P + G3P 2.50±0.21 0.51±0.09 1.20±0.23 0.47±0.03 0.31±0.15 0.29±0.05-0.01±0.08 0.35±0.02 0.02±0.12 0.28±0.04 0.72±0.06 0.20±0.05 S7P + G3P R5P + E4P 2.50±0.21 0.95±0.05 1.20±0.23 0.95±0.00 0.31±0.15 0.95±0.00-0.01±0.08 0.95±0.00 0.02±0.12 0.95±0.00 0.72±0.06 0.95±0.00 R5P + E4P H6P + G3P2.13±0.21 0.17±0.07 0.77±0.22 0.00±0.01 0.00±0.14 0.00±0.04-0.28±0.09 0.28±0.06 -0.22±0.12 0.14±0.06 0.50±0.05 0.14±0.04 PYRCO2+ AcCoA 3.54±0.52 N/A 6.84±0.45 N/A 8.14±0.38 N/A 5.41±0.28 N/A 4.31±0.31 N/A 2.17±0.16 N/A OAA + AcCoA CIT 1.93±0.37 N/A 3.76±0.28 N/A 6.86±0.56 N/A 4.14±0.30 N/A 3.49±0.31 N/A 1.13±0.13 N/A CITICIT 1.93±0.37 N/A 3.76±0.28 N/A 6.86±0.56 N/A 4.14±0.30 N/A 3.49±0.31 N/A 1.13±0.13 N/A ICIT KG + CO2 1.07±0.25 0.00±0.11 140±0.16 0.21±0.09 6.20±0.72 0.00±0.063.62±0.35 0.05±0.09 3.30±0.38 0.00±0.10 0.48±0.14 0.00±0.06 KG SUCC + CO2 0.09±0.27 N/A 0.19±0.15 N/A 5.32±0.71 N/A 2.85±0.41 N/A 2.68±0.37 N/A 0.01±0.12 N/A SUCC FUM 0.96±0.38 0.01±0.00 2.55±0.23 0.01±0.00 5.98±0.55 0.01 3.37±0.32 0.01±0.00 2.88±0.30 0.01±0.00 0.66±0.09 0.01±0.00 FUM MAL 0.97±0.20 0.95±0.15 1.89±0.11 0.00±0.15 3.39±0.28 0.16±0.132.03±0.16 0.69±0.15 1.72±0.16 0.67±0.11 0.60±0.05 0.01±0.13 FUM MAL 0.97±0.20 0.95±0.15 1.89±0.11 0.00±0.15 3.39±0.28 0.16±0.132.03±0.16 0.69±0.15 1.72±0.16 0.67±0.11 0.60±0.05 0.01±0.13 MAL OAA 2.81±0.54 0.95±0.15 6.13±0.39 0.95±0.00 7.43±0.43 0.95±0.064.58±0.29 0.95±0.06 3.63±0.31 0.95±0.08 1.86±0.15 0.95±0.06 ICIT GLX + SUCC 0.86±0.16 N/A 2.36±0.20 N/A 0.66±0.24 N/A 0.52±0.18 N/A 0.20±0.16 N/A 0.65±0.08 N/A GLX + AcCoA MAL 0.86±0.16 N/A 2.36±0.20 N/A 0.65±0.25 N/A 0.52±0.19 N/A 0.20±0.17 N/A 0.65±0.09 N/A MAL PYR + CO2 0.00±0.00 0.00±0.10 0.00±0.00 0.76±0.08 0.00±0.00 0.92±0.060.00±0.01 0.79±0.04 0.00±0.00 0.82±0.08 0.00±0.00 0.40±0.06 PEP + CO2 OAA 1.81±0.19 0.00±0.05 1.25±0.16 0.00±0.02 1.82±0.28 0.00±0.041.62±0.17 0.17±0.07 1.50±0.16 0.00±0.08 0.89±0.06 0.00±0.04 2 GLX PEP 0.00±0.00 0.00±0.03 0.00±0.00 0.00±0.02 0.00±0.05 0.00±0.010.00±0.06 0.00±0.00 0.00±0.01 0.00±0.00 0.00±0.01 0.00±0.01 CO2 Efflux 13.62±0.95 N/A 14.13±0.77 N/A 21.67±1.76 N/A 13.29±0.61 N/A 11.58±0.78 N/A 7.82±0.55 N/A Final Residuum 182.27 142.11 99.88 73.72 72.75 86.43 Table 2.2. Net and exchange (Xch) fluxes determined by 13C MFA with 90% confidence intervals. (See materials and methods) Residuum is the minimal sum-of-squares optimized value found using 13C-Flux. Abbreviations as for Figure 2.4. 69 Figure 2.6. 13C MFA flux maps of P. aeruginosa cystic fibrosis clinical isolates. (A) Net fluxes relative to those of the reference strain, PA01. (B) MFA determined fluxes (moles of C) in experimental strains relative to the carbon molar flux of glucose uptake. Each flux is expressed with its +/- 90% confidence intervals. Several arrows were removed in comparison to Figure 2.3 due to either low flux in the reference strain (ED7 in A) or zero fluxes seen in reaction for all strains. From top to bottom: AA2, AA44, AA43, LES400, AMT0023-30, and PA01 (B only). Arrow weights of figure do not quantitatively correspond to flux sizes. The reactions catalyzed by Aldolase and alpha-ketogluarate dehydrogenase (strain PA01) show the optimized model results paired with the upper range of the calculated CI 90% as modeled confidence interval results are limited to a flux of zero and notation used for other reactions would imply a reversible flux. 70 Figure 2.7 Comparison of select metabolic fluxes determined by 13C MFA with growth rates of P. aeruginosa strains. MFA derived fluxes are plotted against measured intrinsic growth rate with coefficient of determination of linear regression of fluxes for glyceraldehyde phosphate dehydrogenase (diamond), isocitrate lyase (filled squares; the value for strain AA2 is represented with and empty square), and aspartate synthesis flux (filled circles). The R2 value for isocitrate lyase with the inclusion of strain AA2 is 0.599. Net flux values are as in Table 2.2. Figure 2.8. Discrimination of metabolic phenotypes among P. aeruginosa strains. Hierarchical Clustering Analysis using MFA derived flux values was employed to identify two main groupings as discussed in the text.71 Growth Rate () GLX/TCA ED/TCA OPPP/ED OPPP/TCA NAD(P)H ATP FADH2 Respiratory ATP Synthesis PA01 FBA 0.85 0.00 1.87 0.07 0.14 38.71 -2.82 3.42 99.11 LES 400 0.96±0.12 0.63±0.11 2.69±0.10 0.20±0.20 0.53±0.21 39.35±2.54 -9.12±0.67 2.55±0.23 93.09±0.97 PA01 0.78±0.05 0.45±0.27 2.93±0.22 0.74±0.13 2.16±0.21 38.86±2.48 -8.21±0.58 0.96±0.38 90.36±1.34 AA44 0.73±0.03 0.57±0.17 2.89±0.12 0.37±0.09 1.06±0.14 17.69±0.98 -3.81±0.33 0.66±0.09 41.40±0.66 AA2 0.89±0.08 0.13±0.36 1.44±0.08 0.00±9.79 0.00±9.79 29.10±1.99 -1.91±0.83 3.37±0.32 75.88±1.15 AMT 0023-300.71±0.07 0.10±0.38 1.65±0.09 0.05±0.50 0.07±0.51 44.41±3.41 -4.17±1.39 5.98±0.55 115.82±1.72 AA43 0.67±0.03 0.06±0.82 1.42±0.10 0.01±5.81 0.01±5.81 24.71±2.38 -1.47±0.83 2.88±0.30 64.62±1.23 P. fluorescens SBW25 0.04 0 0.59±0.00 0.35±40.61 0.21±40.61 15.03±1.56 0.20±2.04 3.06±100.00 42.36±102 P. fluorescens mucA- algC 0.04 0.11±0.0 0.44±0.00 1.34±0.00 0.59±0.00 9.19±1.86 2.45±3.80 1.47±100.20 27.63±212 E. coli aerobic 0.58±0.01 0.14±0.79 5.14±4.05# 0.33±0.03# 1.68±0.29 31.89±2.32 10.77±1.93 0.27±0.33 90.90±3.98 Table 2.3. Flux ratios and cofactor production rates for P. aeruginosa strains. The ratios of fluxes at major central metabolic branch points, as discussed in the text, separate the strains into the two groupings shown in Figure 2.8. Estimated cofactor production rates are from MFA determined flux maps normalized to the glucose uptake rate for each strain. ATP, substrate level phosphorylation; NADP(H) total cellular NADH + NADPH production. Strains LES 400, PA01, and AA44 show higher flux ratios but lower relative ATP generation rates than strains AA2, AMT 0023-30, and AA43. Flux ratios calculated using: GLX/TCA = V(ICITGLX+SUCC)/V(OAA+AC-COACIT); ED/TCA = V(6PGG3P+PYR)/V(OAA+AC-COACIT); ED/OPPP = V(6PGG3P+PYR)/V(P5P+P5PS7P+G3P) ; OPPP/TCA = V(P5P+P5PS7P+G3P)/ V(OAA+AC-COACIT). These results show the differences in energy turnover, with strain AMT0023-30 being the highest in energy produced per glucose uptake and ATP production per unit of growth. NADPH consumption related to biosynthesis estimated from FBA values were removed from the total NAD(P)H pool before ATP equivalent calculation. 72 These ratios were based on major branch points such as between the TCA and Glyoxylate cycles and EDP:OPPP as well as on alternate sources of co-factors. Total NAD(P)H and ATP production rates are also shown. This analysis points to the ratios GLX/TCA, ED/TCA, OPPP/ED and OPPP/TCA being markedly different between the two flux phenotypes discriminated by HCA (Figure 2.8). A calculation of flux ratios derived from FBA using a maximized biomass production objective function is included, showing that with the exception of the GLX/TCA flux, maximal growth fluxes fall within the variation across strains while none of the strains shows flux partitioning predicted by maximal growth. Flux ratios derived from MFA aerobic E. coli measures [78] also are included, with EMPP used in place of EDP, for comparison to another gram-negative bacteria grown under similar conditions. While the E. coli is closer in its flux ratios to strains LES 400, PA01, and AA44, Table 2.3 clearly demonstrates very different energy utilization strategies for glucose consumption, with a net negative total direct ATP production in P. aeruginosa compared to a positive ATP production in E. coli. P. aeruginosa growth is also significantly higher than E. coli under similar conditions. A related species, P. fluorescens, MFA[96] under similar conditions is also included in this table, demonstrating metabolic differences between alginate non-producing wild-type (SBW25) and an alginate operon induced but non-alginate producing mutant (mucA-algC; [99]). DISCUSSION Many studies of Pseudomonas aeruginosa have focused on identifying genetic changes [33,37,39], production of biofilm [30,100Œ102], variation in biofilm components [28,31,103], biofilm dispersal [32,104], virulence factor production [23,24,27,29,41], understanding antibiotic resistance [22], and identifying pathogenic and phenotypic differences between environmental isolates and pathogenic isolates [105,106]. Quantifying metabolic fluxes across strains can show 73 whether the genetic and physiological differences correspond to metabolic flux patterns. This study provides the first direct evidence of distinct metabolic phenotypes among clinical isolates from cystic fibrosis infections by P. aeruginosa. An earlier MFA study of uropathogenic P. aeruginosa strains reported a significantly different map of central metabolic fluxes for the PA01 reference strain to the one we observed (Table 2.2). These include a much smaller carbon flux through the OPPP (1.07 mmol gDW-1 hr-1 vs our finding of 8.03 mmol gDW-1hr-1), a larger portion of carbon passing through the EDP (8.24 mmol gDW-1hr-1 vs 5.67 mmol gDW-1hr-1), a substantial TCA cyclic flux (6.42 mmol gDW-1hr-1 vs1.93 mmol gDW-1hr-1), and import of carbon into the TCA cycle from pyruvate and its metabolites with an output flux from the TCA cycle from oxaloacetate to phosphoenolpyruvate [77]. If this divergence were due to the effects of the media (moderately higher salt content in the earlier study), it would highlight metabolic plasticity in P. aeruginosa. The alternative is that experimental/analytical differences account for the flux map discrepancies. Since Berger et al., in keeping with common practice, used a single labeling scheme (100% 1-13C glucose), we repeated our MFA analysis using only the 100% 1-13C glucose labeled amino acid data obtained in this study. This yielded fluxes closer to that reported by Berger et al., with a decreased OPPP decarboxylation flux, and increased fluxes in the TCA cycle and EDP (see supplemental 2.4). We also tested the effect of removing the measurements of CO2 efflux rate (which limited this study's MFA CO2 efflux rate to be between 23-39% of glucose uptake rates and which is not measured in most MFA studies). This additional reduction in constraining data resulted in estimated fluxes that are very similar to the one reported by Berger et al. (see supplemental 2.4). Crown et al. [98] recently reported that multiple glucose labeling experiments yield better resolution and confidence in 13C MFA of E. coli than one labeled isotopomer or one 74 combination of isotopomers. This is consistent with previous label design studies (reviewed in [107]), and our observations here and previously [78,97,108,109], which also support the value of gas exchange measurements in 13C MFA. Due to the generally low throughput of MFA, there have been efforts both to speed up flux analyses [72,110], to use 13C labeling patterns in metabolite profile datasets without flux mapping to identify pathway activities [111] and to correlate differences in steady state labeling of biomass (protein amino acids) with alterations in particular fluxes among knockout mutants or different substrate use [112]. Here, we examined whether relatedness of strains in their overall labeling patterns was linked to relatedness in flux maps, which would be particularly valuable in flux phenotype screening of multiple strains. Our results show that while strains can be reliably separated using amino acid labeling patterns, the nature and degree of metabolic differences cannot be straightforwardly inferred. Previous sequencing work creates the potential to assess genotype-phenotype relationships. Strain LES 400 was genomically compared to strain PA01, identifying multiple potentially significant genetic differences between these strains [37]. Under these conditions, the genomic differences did not cause large changes in the flux patterns. Previous study of the closely related strains AA2, AA43, and AA44, identified AA44 as being different in its non-mucoid character [23]. Here we observed substantial flux pattern differences between AA2 and AA43 on the one hand and AA44 on the other under conditions where no significant exopolysaccharide production was detected. Strain AMT0023-30, which was characterized as having persister cells [31] segregated with these mucoid strains. Persister cells are often associated with biofilm formation [113] so the flux differences observed here may be functionally related to differences between strains that for biofilms versus non-biofilm formers. 75 A MFA study of a wild-type strain of P. fluorescens and a mutant that has induced alginate production signaling with corresponding deletion of a necessary enzyme for alginate production strain reported flux map differences that resemble the differences seen in this study between the two phenotypic groups: high TCA, low OPPP flux in one; and low TCA and high OPPP flux in the other [96]. An FBA study of Neisseria meningitidis, another species lacking a complete EMPP, also predicted that these two extremes in TCA and OPPP flux can exist [114]. We note that the flux phenotypes of strains AA43 and AA44 show divergent metabolic adaptations compared to their shared ancestor: strain AA2. Thus different basal metabolic strategies can be observed in related strains after prolonged survival in the cystic fibrosis lung [29]. Several studies have investigated the importance and interplay between metabolism and pathogenicity. The Entner-Doudoroff Pathway (EDP) is a well conserved pathway in the Pseudomonas genus[115,116], whose members lack a complete EMPP, and has been shown to be preferentially utilized by glucose consuming marine bacteria and to correlate with oxidative stress tolerance [117]. The EDP was postulated to have selective advantages based on the observation of lowered fitness when Fructose-bisphosphate aldolase was introduced in a related species, Pseudomonas putida [118]. The EDP has also been argued to have advantages due to lower protein expression requirements and increased NADPH production in comparison to the EMPP [118,119]. A previous transcriptomics study on sequential CF isolates of P. aeruginosa from three separate patients demonstrated no change in expression levels of EDP enzymes, which contrasts with the variability in the carbon flux among the strains investigated here [43]. This study also demonstrates through MFA findings that efficient growth relies on utilization of EDP over the OPPP. Glyceraldehyde 3 phosphate cycling from the EDP has also been discussed 76 in the literature of Pseudomonas, with evidence of this occurring in mutants [115] and alginate label studies in Pseudomonas mendocina [120]. Decreases in TCA utilization in pathogenic organisms has also been shown to correspond to changes in virulence. These metabolic changes are associated with increases in survivability and growth within the oxidative environment of activated macrophages in Salmonella typhimurium [121,122], attenuation of virulence in Salmonella enterica and Yersinia pseudotuberculosis [76,123], and decrease in type III secretion system expression in Y. pseudotuberculosis and P. aeruginosa [76,124,125]. Finally, the glyoxylate shunt has been implicated in pathogenesis of organisms in human disease [126]. In addition to upregulation of isocitrate lyase in CF infections [45], direct evidence of the importance of the glyoxylate shunt in P. aeruginosa pathogenicity was shown in a mutant screen, where the knockout of isocitrate lyase prevented infection of alfalfa seedlings and substantially reduced lung infection [127]. Here we observed small-to-moderate fluxes through this shunt in all strains, and a significant correlation between this flux and intrinsic growth rate across strains. This pathway activity under conditions when neither fatty acids nor acetate were provided, points to a possible anapleurotic role. Oxidative stress during chronic infection may explain some of the flux patterns observed. Oxidative bursts are seen in many different organisms™ response to bacterial infection [128Œ131], and high oxidative stress in cystic fibrosis patients has been well documented [132,133]. A correlation has been established between oxidative stress and strain diversity in P. aeruginosa [134]. Glucose catabolism via the EDP and OPPP pathways produces NADPH, which is needed in antioxidant production/regeneration. The ratio of OPPP to EDP fluxes would regulate the NADPH production rate, since the first produces two, and the second one NADPH per glucose 77 equivalent. Immune cell oxidative attack has been shown to result in inhibition of glycolytic metabolism beyond glyceraldehyde 3 phosphate [135], perhaps favoring utilization of the EDP to bypass the lower steps of glycolysis. The gene producing glucose-6-phosphate dehydrogenase, zwf, normally under repression when non-glucose carbon sources are available, becomes dysregulated in some cystic fibrosis isolates [136,137]. It will be necessary to extend the present approach to characterizing clinical strains to determine whether the flux phenotypes and their correlates observed here also hold under conditions more closely resembling those in the CF lung. ACKNOWLEDGMENTS We wish to thank Matthew Juergens for performing the Hierarchical Clustering Analysis and for the preliminary work and culturing for this project by Nathan Praschan and Raven Batshon. We also wish to thank Seattle Children's Hospital Cystic Fibrosis Isolate Core (NIH P30 DK089507) for their donation of strain AMT0023-30 and the RTSF Mass Spectrometry and Metabolomics Core for at Michigan State University for their analytical support. Michael Opperman was supported in part by funds from the Cellular and Molecular Biology Graduate Program and College of Human Medicine at Michigan State University and by a MD-PhD Fellowship grant by Spectrum Health. 78 APPENDICES79 APPENDIX A - Electronic Supplemental Descriptions Descriptions of the supplemental material uploaded as an electronic supplement of this dissertation. Supplemental 2.1: 13C labeling data of amino acids used in Principal Component Analysis and Metabolic Flux Analysis models of this study. Supplemental 2.2: Complete list of reactions in Metabolic Flux Analysis model with complete flux table results and list of biomass values set for each strain. Supplemental 2.3: Complete flux results for Flux Balance Analysis of biomass optimized growth of strain PA01and Flux Variability Analysis results for 99% and 90% maximally optimized biomass flux rate. Alterations to the biomass equation and additional reactions added to Oberhardt et al. FBA model are also described. Supplemental 2.4: Metabolic Flux Analysis flux results for strain PA01 model with removal of 20% 1,2,3,4,5,6-13C Glucose AA label data and CO2 efflux data. Supplemental 2.5: The version of the formatted Flux Balance Analysis model used in this study (.xls format). Figure S2.1. Averaged carbonaeruginosa cystic fibrosis isolates grown in M9.experimental strains normalized to total carbon uptake averaged with standard deviation determined. Arrow thicknesses are proportional are colorized to demonstrate variability of total uptake carbon through experimental strains.80 APPENDIX B - Additional Figure Figure S2.1. Averaged carbon-uptake normalized flux map from 13C MFA analysis of cystic fibrosis isolates grown in M9. MFA determined c-mol fluxes of individual experimental strains normalized to total carbon uptake averaged with standard deviation determined. Arrow thicknesses are proportional to averaged % of total carbon uptake. Arrows are colorized to demonstrate variability of total uptake carbon through experimental strains. MFA analysis of P. mol fluxes of individual experimental strains normalized to total carbon uptake averaged with standard deviation to averaged % of total carbon uptake. Arrows are colorized to demonstrate variability of total uptake carbon through experimental strains.81 REFERENCES82 REFERENCES 1. Elrod RP, Braun AC. Pseudomonas aeruginosa; its role as a plant pathogen. J Bacteriol. 1942;44: 633Œ645. 2. Kerr JR. Suppression of fungal growth exhibited by Pseudomonas aeruginosa. J Clin Microbiol. 1994;32: 525Œ7. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=263067&tool=pmcentrez&rendertype=abstract 3. Qureshi MN, Perez AA, Madayag RM, Bottone EJ. Inhibition of Acanthamoeba species by Pseudomonas aeruginosa: rationale for their selective exclusion in corneal ulcers and contact lens care systems. J Clin Microbiol. 1993;31: 1908Œ10. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=265656&tool=pmcentrez&rendertype=abstract 4. Yorgey P, Rahme LG, Tan MW, Ausubel FM. The roles of mucD and alginate in the virulence of Pseudomonas aeruginosa in plants, nematodes and mice. Mol Microbiol. 2001;41: 1063Œ1076. doi:10.1046/j.1365-2958.2001.02580.x 5. Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science. 1999;284: 1318Œ1322. doi:10.1126/science.284.5418.1318 6. Cross A, Allen JR, Burke J, Ducel G, Harris A, John J, et al. Nosocomial infections due to Pseudomonas aeruginosa: review of recent trends. Rev Infect Dis. 1983;5 Suppl 5: S837Œ45. Available: http://www.ncbi.nlm.nih.gov/pubmed/6361960 7. Stover C, Pham X, Erwin a, Mizoguchi S. Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen. Nature. 2000;406: 959Œ964. Available: http://www.nature.com/nature/journal/v406/n6799/abs/406959a0.html 8. Alvarez-Ortega C, Harwood CS. Responses of Pseudomonas aeruginosa to low oxygen indicate that growth in the cystic fibrosis lung is by aerobic respiration. Mol Microbiol. 2007;65: 153Œ165. doi:10.1111/j.1365-2958.2007.05772.x 9. Hauser AR, Jain M, Bar-Meir M, McColley S a. Clinical significance of microbial infection and adaptation in cystic fibrosis. Clin Microbiol Rev. 2011;24: 29Œ70. doi:10.1128/CMR.00036-10 10. Livermore DM. Multiple Mechanisms of Antimicrobial Resistance in Pseudomonas aeruginosa: Our Worst Nightmare? Clin Infect Dis. 2002;34: 634Œ640. doi:10.1086/338782 83 11. Lyczak JB, Cannon CL, Pier GB. Lung Infections Associated with Cystic Fibrosis. Clin Microbiol Rev. 2002;15: 194Œ222. doi:10.1128/CMR.15.2.194 12. Cohen TS, Prince A. Cystic fibrosis: a mucosal immunodeficiency syndrome. Nat Med. Nature Publishing Group; 2012;18: 509Œ519. doi:10.1038/nm.2715 13. Gadsby DC, Vergani P, Csanády L. The ABC protein turned chloride channel whose failure causes cystic fibrosis. Nature. 2006;440: 477Œ483. doi:10.1038/nature04712 14. Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, et al. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science. 1989;245: 1066Œ1073. doi:10.1126/science.2475911 15. Kerem B, Rommens JM, Buchanan J a, Markiewicz D, Cox TK, Chakravarti A, et al. Identification of the cystic fibrosis gene: genetic analysis. Science. 1989;245: 1073Œ80. Available: http://www.ncbi.nlm.nih.gov/pubmed/2570460 16. Rommens J, Iannuzzi M, Kerem B, Drumm M, Melmer G, Dean M, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science (80- ). 1989;245: 1059 Œ1065. doi:10.1126/science.2772657 17. Sosnay PR, Siklosi KR, Van Goor F, Kaniecki K, Yu H, Sharma N, et al. Defining the disease liability of variants in the cystic fibrosis transmembrane conductance regulator gene. Nat Genet. 2013;45: 1160Œ7. doi:10.1038/ng.2745 18. Manaker S, Tino G. Natural history and prognosis of advanced lung disease. Clin Chest Med. 1997;18: 435Œ55. doi:10.1016 19. Mitchell RMS, Byrne MF, Baillie J. Pancreatitis. Lancet (London, England). 2003;361: 1447Œ55. doi:3001 20. Bell SC, De Boeck K, Amaral MD. New pharmacological approaches for cystic fibrosis: Promises, progress, pitfalls. Pharmacol Ther. Elsevier Inc.; 2014;145: 19Œ34. doi:10.1016/j.pharmthera.2014.06.005 21. Bjarnsholt T, Jensen PØ, Fiandaca MJ, Pedersen J, Hansen CR, Andersen CB, et al. Pseudomonas aeruginosa biofilms in the respiratory tract of cystic fibrosis patients. Pediatr Pulmonol. 2009;44: 547Œ558. doi:10.1002/ppul.21011 22. Breidenstein EBM, de la Fuente-Núñez C, Hancock REW. Pseudomonas aeruginosa: All roads lead to resistance. Trends Microbiol. 2011;19: 419Œ426. doi:10.1016/j.tim.2011.04.005 84 23. Bragonzi A, Paroni M, Nonis A, Cramer N, Montanari S, Rejman J, et al. Pseudomonas aeruginosa microevolution during cystic fibrosis lung infection establishes clones with adapted virulence. Am J Respir Crit Care Med. 2009;180: 138Œ145. doi:10.1164/rccm.200812-1943OC 24. Carter MEK, Fothergill JL, Walshaw MJ, Rajakumar K, Kadioglu A, Winstanley C. A subtype of a Pseudomonas aeruginosa cystic fibrosis epidemic strain exhibits enhanced virulence in a murine model of acute respiratory infection. J Infect Dis. 2010;202: 935Œ942. doi:10.1086/655781 25. Cheng K, Smyth RL, Govan JR, Doherty C, Winstanley C, Denning N, et al. Spread of beta-lactam-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. Lancet. 1996;348: 639Œ642. doi:10.1016/S0140-6736(96)05169-0 26. Clark ST, Diaz Caballero J, Cheang M, Coburn B, Wang PW, Donaldson SL, et al. Phenotypic diversity within a Pseudomonas aeruginosa population infecting an adult with cystic fibrosis. Sci Rep. Nature Publishing Group; 2015;5: 10932. doi:10.1038/srep10932 27. Fothergill JL, Panagea S, Hart C a, Walshaw MJ, Pitt TL, Winstanley C. Widespread pyocyanin over-production among isolates of a cystic fibrosis epidemic strain. BMC Microbiol. 2007;7: 45. doi:10.1186/1471-2180-7-45 28. Hogardt M, Heesemann J. Microevolution of pseudomonas aeruginosa to a chronic pathogen of the cystic fibrosis lung. Curr Top Microbiol Immunol. 2013;358: 91Œ118. doi:10.1007/82_2011_199 29. Lorè NI, Cigana C, De Fino I, Riva C, Juhas M, Schwager S, et al. Cystic Fibrosis-Niche Adaptation of Pseudomonas aeruginosa Reduces Virulence in Multiple Infection Hosts. PLoS One. 2012;7: e35648. doi:10.1371/journal.pone.0035648 30. Mavrodi D V, Bonsall RF, Delaney SM, Soule MJ, Phillips G, Thomashow LS. Functional Analysis of Genes for Biosynthesis of Pyocyanin and Phenazine-1-Carboxamide from Pseudomonas aeruginosa PA01. J Bacteriol. 2001;183: 6454Œ6465. doi:10.1128/JB.183.21.6454 31. Mulcahy LR, Burns JL, Lory S, Lewis K. Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol. 2010;192: 6191Œ6199. doi:10.1128/JB.01651-09 32. Sauer K, Cullen MC, Rickard a H, Zeef L a H, Gilbert P, Davies DG. Characterization of Nutrient-Induced Dispersion in Pseudomonas aeruginosa PA01 Biofilm. J Bacteriol. 2004;186: 7312Œ7326. doi:10.1128/JB.186.21.7312 85 33. Bragonzi A, Wiehlmann L, Klockgether J, Cramer N, Worlitzsch D, Döning G, et al. Sequence diversity of the mucABD locus in Pseudomonas aeruginosa isolates from patients with cystic fibrosis. Microbiology. 2006;152: 3261Œ3269. doi:10.1099/mic.0.29175-0 34. Caballero JD, Clark ST, Coburn B, Zhang Y, Wang PW, Donaldson SL, et al. Selective sweeps and parallel pathoadaptation drive pseudomonas aeruginosa evolution in the cystic fibrosis lung. MBio. 2015;6: 1Œ14. doi:10.1128/mBio.00981-15 35. Chung JCS, Becq J, Fraser L, Schulz-Trieglaff O, Bond NJ, Foweraker J, et al. Genomic variation among contemporary Pseudomonas aeruginosa isolates from chronically infected cystic fibrosis patients. J Bacteriol. 2012;194: 4857Œ4866. doi:10.1128/JB.01050-12 36. Darch SE, McNally A, Harrison F, Corander J, Barr HL, Paszkiewicz K, et al. Recombination is a key driver of genomic and phenotypic diversity in a Pseudomonas aeruginosa population during cystic fibrosis infection. Sci Rep. 2015;5: 7649. doi:10.1038/srep07649 37. Jeukens J, Boyle B, Kukavica-Ibrulj I, Ouellet MM, Aaron SD, Charette SJ, et al. Comparative genomics of isolates of a Pseudomonas aeruginosa epidemic strain associated with chronic lung infections of cystic fibrosis patients. PLoS One. 2014;9: 1Œ15. doi:10.1371/journal.pone.0087611 38. Markussen T, Marvig RL, Gomez-Lozano M, Aanaes K, Burleigh AE, Hoiby N, et al. Environmental Heterogeneity Drives Within-Host Diversification and Evolution of Pseudomonas aeruginosa. MBio. 2014;5: e01592Œ14Œe01592Œ14. doi:10.1128/mBio.01592-14 39. Mena A, Smith EE, Burns JL, Speert DP, Moskowitz SM, Perez JL, et al. Genetic adaptation of Pseudomonas aeruginosa to the airways of cystic fibrosis patients is catalyzed by hypermutation. J Bacteriol. 2008;190: 7910Œ7917. doi:10.1128/JB.01147-08 40. Rau MH, Marvig RL, Ehrlich GD, Molin S, Jelsbak L. Deletion and acquisition of genomic content during early stage adaptation of Pseudomonas aeruginosa to a human host environment. Environ Microbiol. 2012;14: 2200Œ2211. doi:10.1111/j.1462-2920.2012.02795.x 41. Salunkhe P, Smart CHM, Morgan JAW, Panagea S, Walshaw MJ, Hart CA, et al. A Cystic Fibrosis Epidemic Strain of Pseudomonas aeruginosa Displays Enhanced Virulence and Antimicrobial Resistance. J Bacteriol. 2005;187: 4908Œ4920. doi:10.1128/JB.187.14.4908-4920.2005 86 42. Harmer C, Alnassafi K, Hu H, Elkins M, Bye P, Rose B, et al. Modulation of gene expression by Pseudomonas aeruginosa during chronic infection in the adult cystic fibrosis lung. Microbiology. 2013;159: 2354Œ63. doi:10.1099/mic.0.066985-0 43. Hoboth C, Hoffmann R, Eichner A, Henke C, Schmoldt S, Imhof A, et al. Dynamics of adaptive microevolution of hypermutable Pseudomonas aeruginosa during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2009;200: 118Œ130. doi:10.1086/599360 44. Huse HK, Kwon T, Zlosnik JEA, Speert DP, Marcotte EM, Whiteley M. Parallel evolution in Pseudomonas aeruginosa over 39,000 generations in vivo. MBio. 2010;1: e00199Œ10. doi:10.1128/mBio.00199-10.Editor 45. Son MS, Matthews WJ, Kang Y, Nguyen DT, Hoang TT. In vivo evidence of Pseudomonas aeruginosa nutrient acquisition and pathogenesis in the lungs of cystic fibrosis patients. Infect Immun. 2007;75: 5313Œ5324. doi:10.1128/IAI.01807-06 46. Varga JJ, Barbier M, Mulet X, Bielecki P, Bartell JA, Owings JP, et al. Genotypic and phenotypic analyses of a Pseudomonas aeruginosa chronic bronchiectasis isolate reveal differences from cystic fibrosis and laboratory strains. BMC Genomics. 2015;16: 883. doi:10.1186/s12864-015-2069-0 47. Park AJ, Murphy K, Krieger JR, Brewer D, Taylor P, Habash M, et al. A temporal examination of the planktonic and biofilm proteome of whole cell Pseudomonas aeruginosa PA01 using quantitative mass spectrometry. Mol Cell Proteomics. 2014;13: 1095Œ105. doi:10.1074/mcp.M113.033985 48. Wu X, Held K, Zheng C, Staudinger BJ, Chavez JD, Weisbrod CR, et al. Dynamic Proteome Response of Pseudomonas aeruginosa to Tobramycin Antibiotic Treatment. Mol Cell Proteomics. 2015;14: 2126Œ2137. doi:10.1074/mcp.M115.050161 49. Orth JD, Thiele I, Palsson BØ. What is flux balance analysis? Nat Biotechnol. 2010;28: 245Œ248. doi:10.1038/nbt.1614.What 50. Lewis NE, Nagarajan H, Palsson BO. Constraining the metabolic genotype-phenotype relationship using a phylogeny of in silico methods. Nat Rev Microbiol. 2012;10: 291Œ305. doi:doi:10.1038/nrmicro2737 51. Karr JR, Takahashi K, Funahashi A. The principles of whole-cell modeling. Curr Opin Microbiol. Elsevier Ltd; 2015;27: 18Œ24. doi:10.1016/j.mib.2015.06.004 52. Hyduke D, Hyduke D, Schellenberger J, Que R, Fleming R, Thiele I, et al. COBRA Toolbox 2.0. Protoc Exch. 2011; doi:10.1038/protex.2011.234 53. Kim HU, Kim SY, Jeong H, Kim TY, Kim JJ, Choy HE, et al. Integrative genome-scale 87 metabolic analysis of Vibrio vulnificus for drug targeting and discovery. Mol Syst Biol. Nature Publishing Group; 2014;7: 460Œ460. doi:10.1038/msb.2010.115 54. Larocque M, Chénard T, Najmanovich R. A curated C. difficile strain 630 metabolic network: prediction of essential targets and inhibitors. BMC Syst Biol. 2014;8: 117. doi:10.1186/s12918-014-0117-z 55. Lee D-S, Burd H, Liu J, Almaas E, Wiest O, Barabasi A-L, et al. Comparative Genome-Scale Metabolic Reconstruction and Flux Balance Analysis of Multiple Staphylococcus aureus Genomes Identify Novel Antimicrobial Drug Targets. J Bacteriol. 2009;191: 4015Œ4024. doi:10.1128/JB.01743-08 56. Lee D-Y, Chung BKS, Yusufi FNK, Selvarasu S. In silico genome-scale modeling and analysis for identifying anti-tubercular drug targets. Drug Dev Res. 2011;72: 121Œ129. doi:10.1002/ddr.20408 57. Song C, Chiasson M a, Nursimulu N, Hung SS, Wasmuth J, Grigg ME, et al. Metabolic reconstruction identifies strain-specific regulation of virulence in Toxoplasma gondii. Mol Syst Biol. Nature Publishing Group; 2013;9: 708. doi:10.1038/msb.2013.62 58. Veith N, Solheim M, van Grinsven KWA, Olivier BG, Levering J, Grosseholz R, et al. Using a Genome-Scale Metabolic Model of Enterococcus faecalis V583 To Assess Amino Acid Uptake and Its Impact on Central Metabolism. Appl Environ Microbiol. 2015;81: 1622Œ33. doi:10.1128/AEM.03279-14 59. Wang C, Deng Z-L, Xie Z-M, Chu X-Y, Chang J-W, Kong D-X, et al. Construction of a genome-scale metabolic network of the plant pathogen Pectobacterium carotovorum provides new strategies for bactericide discovery. FEBS Lett. Federation of European Biochemical Societies; 2014;589: 285Œ94. doi:10.1016/j.febslet.2014.12.010 60. Kim HU, Kim TY, Lee SY. Genome-scale metabolic network analysis and drug targeting of multi-drug resistant pathogen Acinetobacter baumannii AYE. Mol Biosyst. 2010;6: 339Œ348. doi:10.1039/b916446d 61. Oberhardt MA, Puchaka J, Fryer KE, Martins dos Santos V a P, Papin JA. Genome-scale metabolic network analysis of the opportunistic pathogen Pseudomonas aeruginosa PA01. J Bacteriol. 2008;190: 2790Œ2803. doi:10.1128/JB.01583-07 62. Oberhardt M a., Goldberg JB, Hogardt M, Papin J a. Metabolic network analysis of Pseudomonas aeruginosa during chronic cystic fibrosis lung infection. J Bacteriol. 2010;192: 5534Œ5548. doi:10.1128/JB.00900-10 63. Sigurdsson G, Fleming RMT, Heinken A, Thiele I. A systems biology approach to drug targets in pseudomonas aeruginosa biofilm. PLoS One. 2012;7: 1Œ9. doi:10.1371/journal.pone.0034337 88 64. Niedenführ S, Wiechert W, Nöh K. How to measure metabolic fluxes: a taxonomic guide for (13)C fluxomics. Curr Opin Biotechnol. 2015;34: 82Œ90. doi:10.1016/j.copbio.2014.12.003 65. Kelleher JK. Flux estimation using isotopic tracers: common ground for metabolic physiology and metabolic engineering. Metab Eng. 2001;3: 100Œ10. doi:10.1006/mben.2001.0185 66. Kohlstedt M, Becker J, Wittmann C. Metabolic fluxes and beyondŠsystems biology understanding and engineering of microbial metabolism. Appl Microbiol Biotechnol. 2010;88: 1065Œ1075. doi:10.1007/s00253-010-2854-2 67. Wiechert W, Möllney M, Petersen S, De Graaf AA. A universal framework for 13C metabolic flux analysis. Metab Eng. 2001;3: 265Œ283. doi:10.1006/mben.2001.0188 68. Wiechert W, Möllney M, Isermann N, Wurzel M, De Graaf A a. Bidirectional reaction steps in metabolic networks: III. Explicit solution and analysis of isotopomer labeling systems. Biotechnol Bioeng. 1999;66: 69Œ85. doi:10.1002/(SICI)1097-0290(1999)66:2<69::AID-BIT1>3.0.CO;2-6 69. Adebiyi AO, Jazmin LJ, Young JD. 13C flux analysis of cyanobacterial metabolism. Photosynth Res. Springer Netherlands; 2015;126: 19Œ32. doi:10.1007/s11120-014-0045-1 70. García Martín H, Kumar VS, Weaver D, Ghosh A, Chubukov V, Mukhopadhyay A, et al. A Method to Constrain Genome-Scale Models with 13C Labeling Data. Maranas CD, editor. PLOS Comput Biol. 2015;11: e1004363. doi:10.1371/journal.pcbi.1004363 71. Ghosh A, Nilmeier J, Weaver D, Adams PD, Keasling JD, Mukhopadhyay A, et al. A peptide-based method for 13C Metabolic Flux Analysis in microbial communities. PLoS Comput Biol. 2014;10: e1003827. doi:10.1371/journal.pcbi.1003827 72. Heux S, Poinot J, Massou S, Sokol S, Portais J-C. A novel platform for automated high-throughput fluxome profiling of metabolic variants. Metab Eng. 2014;25: 8Œ19. doi:10.1016/j.ymben.2014.06.001 73. McAtee AG, Jazmin LJ, Young JD. Application of isotope labeling experiments and 13C flux analysis to enable rational pathway engineering. Curr Opin Biotechnol. Elsevier Ltd; 2015;36: 50Œ56. doi:10.1016/j.copbio.2015.08.004 74. Beste DJ V, Nöh K, Niedenführ S, Mendum T a., Hawkins ND, Ward JL, et al. 13C-flux spectral analysis of host-pathogen metabolism reveals a mixed diet for intracellular mycobacterium tuberculosis. Chem Biol. 2013;20: 1012Œ1021. doi:10.1016/j.chembiol.2013.06.012 89 75. Götz A, Eylert E, Eisenreich W, Goebel W. Carbon metabolism of enterobacterial human pathogens growing in epithelial colorectal adenocarcinoma (Caco-2) cells. PLoS One. 2010;5: 8Œ9. doi:10.1371/journal.pone 76. Bücker R, Heroven AK, Becker J, Dersch P, Wittmann C. The Pyruvate-Tricarboxylic Acid Cycle Node. J Biol Chem. 2014;289: 30114Œ30132. doi:10.1074/jbc.M114.581348 77. Berger A, Dohnt K, Tielen P, Jahn D, Becker J, Wittmann C. Robustness and plasticity of metabolic pathway flux among uropathogenic isolates of Pseudomonas aeruginosa. PLoS One. 2014;9: e88368. doi:10.1371/journal.pone.0088368 78. Chen X, Alonso AP, Allen DK, Reed JL, Shachar-Hill Y. Synergy between 13C-metabolic flux analysis and flux balance analysis for understanding metabolic adaption to anaerobiosis in E. coli. Metab Eng. Elsevier; 2011;13: 38Œ48. doi:10.1016/j.ymben.2010.11.004 79. Maier T, Güell M, Serrano L. Correlation of mRNA and protein in complex biological samples. FEBS Lett. Federation of European Biochemical Societies; 2009;583: 3966Œ3973. doi:10.1016/j.febslet.2009.10.036 80. Montezano D, Meek L, Gupta R, Bermudez LE, Bermudez JCM. Flux Balance Analysis with Objective Function Defined by Proteomics Data-Metabolism of Mycobacterium tuberculosis Exposed to Mefloquine. PLoS One. 2015;10: e0134014. doi:10.1371/journal.pone.0134014 81. Puchalka J, Oberhardt MA, Godinho M, Bielecka A, Regenhardt D, Timmis KN, et al. Genome-Scale Reconstruction and Analysis of the Pseudomonas putida KT2440 Metabolic Network Facilitates Applications in Biotechnology. Plos Comput Biol. 2008;4: e1000210. doi:10.1371/journal.pcbi.1000210 82. Schuetz R, Kuepfer L, Sauer U. Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. Mol Syst Biol. 2007;3. doi:10.1038/msb4100162 83. Toya Y, Ishii N, Nakahigashi K, Hirasawa T, Soga T, Tomita M, et al. 13C-Metabolic flux analysis for batch culture of Escherichia coli and its pyk and pgi gene knockout mutants based on mass isotopomer distribution of intracellular metabolites. Biotechnol Prog. 2010;26: 975Œ992. doi:10.1002/btpr.420 84. Williams TCR, Poolman MG, Howden AJM, Schwarzlander M, Fell DA, Ratcliffe RG, et al. A Genome-Scale Metabolic Model Accurately Predicts Fluxes in Central Carbon Metabolism under Stress Conditions. Plant Physiol. 2010;154: 311Œ323. doi:10.1104/pp.110.158535 90 85. Wu C, Xiong W, Dai J, Wu Q. Genome-Based Metabolic Mapping and 13 C Flux Analysis Reveal Systematic Properties of an Oleaginous Microalga Chlorella protothecoides. Plant Physiol. 2015;167: 586Œ599. doi:10.1104/pp.114.250688 86. De Soyza A, Hall AJ, Mahenthiralingam E, Drevinek P, Kaca W, Drulis-Kawa Z, et al. Developing an international Pseudomonas aeruginosa reference panel. Microbiologyopen. 2013;2: 1010Œ1023. doi:10.1002/mbo3.141 87. Sambrook J, Russel DW. Molecular Cloning, A Laboratory Manual. Spring Harbor Laboratory Press, Cold Spring Habor, NY.; 2001. 88. Schwender J, Ohlrogge JB, Shachar-Hill Y. A flux model of glycolysis and the oxidative pentosephosphate pathway in developing Brassica napus embryos. J Biol Chem. 2003;278: 29442Œ29453. doi:10.1074/jbc.M303432200 89. Salunkhe P, Smart CHM, Morgan AW, Panagea S, Walshaw MJ, Hart A, et al. A Cystic Fibrosis Epidemic Strain of Pseudomonas aeruginosa Displays Enhanced Virulence and Antimicrobial Resistance. J Bacteriol. 2005;187: 4908Œ4920. doi:10.1128/JB.187.14.4908 90. Allen DK, Shachar-Hill Y, Ohlrogge JB. Compartment-specific labeling information in 13C metabolic flux analysis of plants. Phytochemistry. 2007;68: 2197Œ2210. doi:10.1016/j.phytochem.2007.04.010 91. Pearson K. LIII. On lines and planes of closest fit to systems of points in space. Philos Mag Ser 6. 1901;2: 559Œ572. doi:10.1080/14786440109462720 92. Dash M, Liu H, Scheuermann P, Tan KL. Fast hierarchical clustering and its validation. Data Knowl Eng. 2003;44: 109Œ138. doi:10.1016/S0169-023X(02)00138-6 93. Neidhardt FC, Edwin Umbarger H. Chemical Composition of Escherichia coli. Chem Compos Escherichia coli Neidhardt, FC (Ed), Escherichia coli Salmonella Cell Mol Biol. 1996;Escherichi. 94. Schmidt K, Nørregaard LC, Pedersen B, Meissner A, Duus JØ, Nielsen JØ, et al. Quantification of Intracellular Metabolic Fluxes from Fractional Enrichment and13CŒ13C Coupling Constraints on the Isotopomer Distribution in Labeled Biomass Components. Metab Eng. 1999;1: 166Œ179. doi:10.1006/mben.1999.0114 95. Mahadevan R, Schilling CH. The effects of alternate optimal solutions in constraint-based genome-scale metabolic models. Metab Eng. 2003;5: 264Œ276. doi:10.1016/j.ymben.2003.09.002 91 96. Lien SK, Niedenführ S, Sletta H, Nöh K, Bruheim P. Fluxome study of Pseudomonas fluorescens reveals major reorganisation of carbon flux through central metabolic pathways in response to inactivation of the anti-sigma factor MucA. BMC Syst Biol. 2015;9: 6. doi:10.1186/s12918-015-0148-0 97. Chen X, Shachar-Hill Y. Insights into metabolic efficiency from flux analysis. J Exp Bot. 2012;63: 2343Œ2351. doi:10.1093/jxb/ers057 98. Crown SB, Long CP, Maciek R. Integrated 13C-metabolic flux analysis of 14 parallel labeling experiments in Escherichia coli. Metab Eng. Elsevier; 2015;28: 151Œ158. doi:10.1016/j.ymben.2015.01.001 99. Borgos SEF, Bordel S, Sletta H, Ertesvåg H, Jakobsen Ø, Bruheim P, et al. Mapping global effects of the anti-sigma factor MucA in Pseudomonas fluorescens SBW25 through genome-scale metabolic modeling. BMC Syst Biol. 2013;7: 19. doi:10.1186/1752-0509-7-19 100. Ma L, Conover M, Lu H, Parsek MR, Bayles K, Wozniak DJ. Assembly and development of the Pseudomonas aeruginosa biofilm matrix. PLoS Pathog. 2009;5. doi:10.1371/journal.ppat.1000354 101. Van Alst NE, Picardo KF, Iglewski BH, Haidaris CG. Nitrate sensing and metabolism modulate motility, biofilm formation, and virulence in Pseudomonas aeruginosa. Infect Immun. 2007;75: 3780Œ3790. doi:10.1128/IAI.00201-07 102. Wagner VE, Iglewski BH. P. aeruginosa Biofilms in CF Infection. Clin Rev Allergy Immunol. 2008; 1Œ11. doi:10.1007/s12016-008-8079-9 103. Stewart PS, Franklin MJ. Physiological heterogeneity in biofilms. Nat Rev Microbiol. 2008;6: 199Œ210. doi:10.1038/nrmicro1838 104. McDougald D, Rice S a., Barraud N, Steinberg PD, Kjelleberg S. Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nat Rev Microbiol. Nature Publishing Group; 2011;10: 39Œ50. doi:10.1038/nrmicro2695 105. Alonso a, Rojo F, Martínez JL. Environmental and clinical isolates of Pseudomonas aeruginosa show pathogenic and biodegradative properties irrespective of their origin. Environ Microbiol. 1999;1: 421Œ430. doi:emi52 [pii] 106. Head NE, Yu H. Cross-Sectional Analysis of Clinical and Environmental Isolates of Pseudomonas aeruginosa: Biofilm Formation, Virulence, and Genome Diversity. Infect Immun. 2004;72: 133Œ144. doi:10.1128/IAI.72.1.133-144.2004 92 107. Antoniewicz MR. 13C metabolic flux analysis: optimal design of isotopic labeling experiments. Curr Opin Biotechnol. Elsevier Ltd; 2013;24: 1116Œ1121. doi:10.1016/j.copbio.2013.02.003 108. Alonso AP, Goffman FD, Ohlrogge JB, Shachar-Hill Y. Carbon conversion efficiency and central metabolic fluxes in developing sunflower (Helianthus annuus L.) embryos. Plant J. 2007;52: 296Œ308. doi:10.1111/j.1365-313X.2007.03235.x 109. Paula Alonso A, Dale VL, Shachar-Hill Y. Understanding fatty acid synthesis in developing maize embryos using metabolic flux analysis. Metab Eng. Elsevier; 2010;12: 488Œ497. doi:10.1016/j.ymben.2010.04.002 110. Junker BH. Flux analysis in plant metabolic networks: Increasing throughput and coverage. Curr Opin Biotechnol. 2014;26: 183Œ188. doi:10.1016/j.copbio.2014.01.016 111. Chokkathukalam A, Kim D-H, Barrett MP, Breitling R, Creek DJ. Stable isotope-labeling studies in metabolomics: new insights into structure and dynamics of metabolic networks. Bioanalysis. 2014;6: 511Œ24. doi:10.4155/bio.13.348 112. Zamboni N, Sauer U. Model-independent fluxome profiling from 2H and 13C experiments for metabolic variant discrimination. Genome Biol. 2004;5: R99. doi:10.1186/gb-2004-5-12-r99 113. Lewis K. Persister cells. Annu Rev Microbiol. 2010;64: 357Œ372. doi:10.1146/annurev.micro.112408.134306 114. Baart GJE, Willemsen M, Khatami E, de Haan A, Zomer B, Beuvery EC, et al. Modeling Neisseria meningitidis B metabolism at different specific growth rates. Biotechnol Bioeng. 2008;101: 1022Œ35. doi:10.1002/bit.22016 115. Kessie TG. Alterative Pathways of Carbohydrate Utilization in Pseudomonads. Annu Rev Microbiol. 1984;38: 359Œ87. 116. Romano a. H, Conway T. Evolution of carbohydrate metabolic pathways. Res Microbiol. 1996;147: 448Œ455. doi:10.1016/0923-2508(96)83998-2 117. Klingner A, Bartsch A, Dogs M, Wagner-Döbler I, Jahn D, Simon M, et al. Large-scale 13 C-flux profiling reveals conservation of the Entner-Doudoroff pathway as glycolytic strategy among glucose-using marine bacteria. Appl Environ Microbiol. 2015;81: AEM.03157Œ14. doi:10.1128/AEM.03157-14 118. Chavarría M, Nikel PI, Pérez-Pantoja D, De Lorenzo V. The Entner-Doudoroff pathway empowers Pseudomonas putidaKT2440 with a high tolerance to oxidative stress. Environ Microbiol. 2013;15: 1772Œ1785. doi:10.1111/1462-2920.12069 93 119. Flamholz A, Noor E, Bar-Even A, Liebermeister W, Milo R. Glycolytic strategy as a tradeoff between energy yield and protein cost. Proc Natl Acad Sci U S A. 2013;110: 10039Œ44. doi:10.1073/pnas.1215283110 120. Anderson AJ, Hacking AJ, Dawes EA. Alternative Pathways for the Biosynthesis of Alginate from Fructose and Glucose in Pseudomonas Mendocina and Azotobacter Vinelandii. Microbiology. 1987;133: 1045Œ1052. doi:10.1099/00221287-133-4-1045 121. Abernathy J, Corkill C, Hinojosa C, Li X, Zhou H. Deletions in the pyruvate pathway of Salmonella Typhimurium alter SPI1-mediated gene expression and infectivity. J Anim Sci Biotechnol. 2013;4: 5. doi:10.1186/2049-1891-4-5 122. Bowden SD, Ramachandran VK, Knudsen GM, Hinton JCD, Thompson A. An Incomplete TCA Cycle Increases Survival of Salmonella Typhimurium during Infection of Resting and Activated Murine Macrophages. May RC, editor. PLoS One. 2010;5: e13871. doi:10.1371/journal.pone.0013871 123. Mercado-Lubo R, Leatham MP, Conway T, Cohen PS. Salmonella enterica Serovar Typhimurium Mutants Unable To Convert Malate to Pyruvate and Oxaloacetate Are Avirulent and Immunogenic in BALB/c Mice. Infect Immun. 2009;77: 1397Œ1405. doi:10.1128/IAI.01335-08 124. Dacheux D, Epaulard O, de Groot A, Guery B, Leberre R, Attree I, et al. Activation of the Pseudomonas aeruginosa Type III Secretion System Requires an Intact Pyruvate Dehydrogenase aceAB Operon. Infect Immun. 2002;70: 3973Œ3977. doi:10.1128/IAI.70.7.3973-3977.2002 125. Wilharm G, Heider C. Interrelationship between type three secretion system and metabolism in pathogenic bacteria. Front Cell Infect Microbiol. 2014;4: 1Œ10. doi:10.3389/fcimb.2014.00150 126. Dunn MF, Ramírez-Trujillo JA, Hernández-Lucas I. Major roles of isocitrate lyase and malate synthase in bacterial and fungal pathogenesis. Microbiology. 2009;155: 3166Œ3175. doi:10.1099/mic.0.030858-0 127. Lindsey TL, Hagins JM, Sokol PA, Silo-Suh LA. Virulence determinants from a cystic fibrosis isolate of Pseudomonas aeruginosa include isocitrate lyase. Microbiology. 2008;154: 1616Œ1627. doi:10.1099/mic.0.2007/014506-0 128. Jain C, Yun M, Politz SM, Rao RP. A pathogenesis assay using Saccharomyces cerevisiae and Caenorhabditis elegans reveals novel roles for yeast AP-1, Yap1, and host dual oxidase BLI-3 in fungal pathogenesis. Eukaryot Cell. 2009;8: 1218Œ27. doi:10.1128/EC.00367-08 94 129. Kavanagh K, Reeves EP. Exploiting the potential of insects for in vivo pathogenicity testing of microbial pathogens. FEMS Microbiol Rev. 2004;28: 101Œ112. doi:10.1016/j.femsre.2003.09.002 130. Lamb C, Dixon RA. The Oxidative Burst in Plant Disease Resistance. Annu MolBiol. 1997;48: 251Œ275. doi:10.1146/annurev.arplant.48.1.251 131. Lambeth JD. NOX enzymes and the biology of reactive oxygen. Nat Rev Immunol. 2004;4: 181Œ189. doi:10.1038/nri1312 132. Lagrange-Puget M, Durieu I, Ecochard R, Abbas-Chorfa F, Drai J, Steghens JP, et al. Longitudinal study of oxidative status in 312 cystic fibrosis patients in stable state and during bronchial exacerbation. Pediatr Pulmonol. 2004;38: 43Œ49. doi:10.1002/ppul.20041 133. Ziady AG, Hansen J. Redox balance in cystic fibrosis. Int J Biochem Cell Biol. Elsevier Ltd; 2014;52: 113Œ123. doi:10.1016/j.biocel.2014.03.006 134. Ciofu O, Riis B, Pressler T, Enghusen H, Høiby N, Poulsen HE. Occurrence of hypermutable Pseudomonas aeruginosa in cystic fibrosis patients is associated with the oxidative stress caused by chronic lung inflammation. Antimicrob Agents Chemother. 2005;49: 2276Œ2282. doi:10.1128/AAC.49.6.2276 135. Deng X, Liang H, Ulanovskaya OA, Ji Q, Zhou T, Sun F, et al. Steady-state hydrogen peroxide induces glycolysis in staphylococcus aureus and pseudomonas aeruginosa. J Bacteriol. 2014;196: 2499Œ2513. doi:10.1128/JB.01538-14 136. Ma J, Hager PW, Howell ML, Phibbs P V, Hassett DJ. Cloning and Characterization of the Pseudomonas aeruginosa zwf Gene Encoding Enzyme Important in Resistance to Methyl Cloning and Characterization of the Pseudomonas aeruginosa zwf Gene Encoding Glucose-6-Phosphate Dehydrogenase , an Enzyme Important in R. 1998;180: 1741Œ1749. 137. Silo-suh L, Suh S, Phibbs P V, Ohman DE. Adaptations of Pseudomonas aeruginosa to the Cystic Fibrosis Lung Environment Can Include Deregulation of zwf, Encoding Glucose-6-Phosphate Dehydrogenase. J Bacteriol. 2005;187: 7561Œ7568. doi:10.1128/JB.187.22.7561 95 CHAPTER 3 Œ Quantification of Pseudomonas aeruginosa Metabolism in Cystic Fibrosis Lung Like Conditions96 ABSTRACT Pseudomonas aeruginosa is an opportunistic pathogen that infects a wide range of species and of significant importance to human health. Known for causing nosocomial infections of burns and of the urinary tract, as well as those related to ventilator-associated pneumonia, P. aeruginosa is a major factor in the morbidity and mortality of cystic fibrosis (CF) patients. With intrinsic abilities to resist antibacterial agents and quick adaptation against new antibiotics, P. aeruginosa is an important organism to study in order to develop treatments that are more effective. Many P. aeruginosa isolates have been investigated using many different Omic and direct virulence factor measures in an attempt to describe shared virulence and metabolic changes during chronic infection of the CF lung. In this study, we used 13C Metabolic Flux Analysis and Flux Balance Analysis on one reference strain and six strain isolates from the lungs of CF patients to quantify their metabolic phenotypes. The strains were grown under aerobic and anoxic conditions in a media formulated to mimic the CF lung. Our results highlight the importance of putrescine (a polyamine) metabolism and the use of the glyoxylate cycle of P. aeruginosa during growth in the CF lung while demonstrating low utilization of the Entner-Doudoroff and Oxidative Pentose Phosphate Pathways. We also report four metabolic phenotypes under growth in aerobic and anoxic conditions and the divergence of metabolism of genetically related isolates that have emerged during chronic CF infection. 97 INTRODUCTION Cystic fibrosis of the pancreas (CF), also known as mucoviscidosis or more commonly simply as cystic fibrosis, is an autosomal recessive disease caused by genetic defects in the Cystic Fibrosis Transmembrane conductance Regulator (CFTR) [1Œ5]. The CFTR, expressed in many epithelial tissues of the body, is especially important for the correct secretion of chloride and bicarbonate ions in the pancreas, lungs, and duodenum [6Œ14]. Defective CFTR expression leads to thickened mucus secretions in the body of CF patients, causing symptoms including: pancreatic duct obstruction; pancreatitis and fibrosis of the pancreas; meconium ileus; gastroesophageal reflux disease and peptic ulcers; and Bronchiectasis, which poses the greatest risk of morbidity and mortality in CF [6Œ9,15Œ24]. Bronchiectasis, an obstructive lung disease, is caused by widening of the lower bronchioles due to chronic inflammation caused by repeated cycles of infection and immune clearance, damaging lung parenchyma [25Œ27]. In CF, secreted mucus has abnormally low viscosity, causing pooling of mucus that allows growth of inhaled bacteria in the lung [28Œ30]. Final stage treatment for decreased lung function caused by bronchiectasis is lung transplantation, though the lack of sufficient organs for transplant makes this option unavailable to most [31,32]. Common infectious species seen in CF include Gram positive bacteria (such as Staphylococcus aureus, Streptoccocus sp.), Gram negative bacteria (Pseudomonas aeruginosa, Haemophilus influenzae) and non-bacterial species (Aspergillus sp.) [33]. Of greatest concern related to the possible lung pathogens in CF is P. aeruginosa. Pseudomonas aeruginosa is a gram negative opportunistic pathogen with a wide range of suitable environments and hosts [34Œ39]. Current prevalence rates of P. aeruginosa infection among CF patients have been reported to be as high as 74% of all CF patients 26 years and older 98 [40,41]. Due to its potential ubiquity in suitable human habitats, the high rate of infection of CF patients is currently thought to be due to random environmental exposure and, rarely, clinical exposure [42]. Additionally, P. aeruginosa is resistant to many current therapies, due to its many intrinsic defenses against current antibacterials (e.g. biofilm, cephalosporinase, and efflux pump expression), as well as its rapid adaption due to the commonly seen hypermutable phenotype (caused by acquired defects in mismatch repair enzymes) [43Œ47]. These intrinsic defenses and rapid adaptations against antibacterial treatment, along with the pathogen™s environmental ubiquity, lead to repeated infection cycles and clinical failure of treatment, thus making P. aeruginosa of major clinical interest in the treatment of CF. Many studies of P. aeruginosa CF isolates have been focused on identifying the changes in the organism during chronic infection. Studies have used various Omic tools [48Œ64], measurements of known virulence factor expression, and measures of physiological changes [42,65Œ76] to establish a wide range of known adaptations. Evidence in these and other studies generally demonstrate an organism growing in a microaerobic to anaerobic environment within the lung (with nitrate available for respiration), with its primary carbon sources being amino acids, glucose, and lactic acid present in the lung sputum [77,78]. Additionally, studies of P. aeruginosa indicated that overproduction of the biofilm component alginate is associated with increased morbidity and mortality in the CF patient population [79Œ82]. Chronic infection isolates of P. aeruginosa also demonstrate decreased production of other known virulence factors [71]. However, these identified changes have not led to a clear understanding of the metabolic adaptations that the organism undergoes for two distinct reasons. First, chronic infection strains of P. aeruginosa are often of the hypermutable phenotype, causing significant genomic 99 instability and diversity, introducing significant uncertainty in direct genomic measurements [62,43,83,84]. Second, different Omic datasets do not correlate well within data sets or in comparisons to other Omic measures of the same samples, rendering uncertain the true meaning of the differences recorded with these techniques [85Œ87]. To identify which metabolic changes are actually significant for pathogenesis and to measure and predict functional changes in metabolism, a toolbox of network based computational and experimental methods is available. The work in this paper focuses on two steady-state based metabolic modeling techniques, Metabolic Flux Analysis and Flux Balance Analysis. Metabolic Flux Analysis (MFA) centers on the use of isotopically labeled substrates (usually 13C) in experiments allowed to reach a stable pattern of isotopic distribution (end point labeling) [88]. The labeling patterns in metabolic products or intermediates are used to deduce internal metabolic state and a pattern of carbon fluxes through the cellular metabolic system: a flux map [89]. To produce such a map, a standard list of reactions with carbon positional transformations in the organism and stoichiometries must be known, and measurements must be taken of extracellular input and output fluxes such as biomass production, gas exchange, and carbon inputs [89Œ91]. Metabolism of carbon-containing compounds by enzymatic processes occurs at defined carbon positions within the molecule due to the biochemical specificity of enzymatic catalysis; this fundamental property of enzymatic reactions allows for characteristic labeling patterns to develop for the carbon in the molecule [92]. To describe the re-shuffling of the label for a reaction, an atom mapping matrix is constructed for each reaction in the network [93]. With mapping matrices and isotopomer matrix representations, matrix algebra can be used to predict the labeling pattern at steady state in all metabolites by assuming reaction rates (the values of the metabolic fluxes) [89]. The computed results are then compared to the measured 100 13C labeled measurements and the assumed flux values are iteratively changed to optimize the fit of the model to the measured labeling patterns [89]. This results in MFA quantifying carbon fluxes through metabolism. The second modeling technique used, Constraints-based Flux Balance Analysis (FBA), is a biochemically-based mathematical modeling approach to the steady state analysis of metabolism that incorporates carbon and non-carbon elements in a large-scale mass balance analysis; energy and redox co-factors are also balanced and no labeling experimental data are needed[94Œ96]. FBA can extend the metabolic scope beyond carbon and beyond central metabolism for complete genomic scale coverage of metabolism [95,97]. The biochemical equations that make up the individual steps of metabolism are converted into a matrix equation in which the stoichiometry of each reaction is analyzed using software packages such as The COBRA (Constraints Based Reconstruction Analysis) toolbox in MATLAB [98] in order to optimize the model reactions. Additionally, optimization algorithms can be used to maximize or minimize selected individual fluxes or combinations of fluxes (called objective functions) in order to produce the most likely carbon flux of the metabolic network being tested [95]. To date, most MFA studies of microbial systems have focused on questions related to biotechnology and metabolic engineering, microbial physiology, and gene function [99Œ103]. Several recent studies have analyzed the metabolic interactions of pathologically relevant bacteria within their host environment [104,105], or the metabolic differences between mutant strains [106]. A recent study of pathogenesis-related metabolism in P. aeruginosa employed MFA to compare 17 uropathogenic strains [107]. FBA investigations of pathogenic organisms have been used to search for novel drug targets and have identified potential metabolic targets not affected by current therapeutics, such as amino acid production or fatty acid metabolism 101 [108Œ115]. Recent FBA research on P. aeruginosa includes a genome-based metabolic and transport model by Oberhardt and colleagues [116], and uses FBA with transcript data from two CF clinical strains isolated 44 months apart in order to investigate P. aeruginosa™s metabolic capabilities and potential metabolic changes during prolonged infection [117]. FBA has also been used to identify potential metabolic drug targets during biofilm growth of P. aeruginosa [118]. The combination of both FBA and MFA allows for investigation of suboptimal growth and testing of alternative objective functions that may be optimized besides maximal biomass production. However, there have been surprisingly few studies in which these complementary network flux analysis approaches have been combined [119,85,99,120Œ125], let alone in direct application to understand bacterial pathogenicity. Here, we performed FBA and 13C MFA of reference strain PA01 and 6 selected CF pathogenic isolates to quantify redox production and consumption, metabolic inefficiencies, and the metabolic phenotypes in media that mimics the CF lung. In this study, we highlight the importance of amino acid catabolism by putrescine (polyamine) metabolism, with significant carbon fluxes not described in the scientific literature of any MFA study. We also demonstrate the differences in utilization of the glyoxylate cycle under aerobic and anoxic growth. Among the clinical strains we report four general metabolic phenotypes, two in aerobic growth and two in anoxic growth, which appear during chronic infection of the CF lung. We also demonstrate divergence in metabolic phenotype during chronic infection of two genetically related strains. Finally, we report strain similarities and differences using 13C labeling data of proteogenic amino acids and with obtained 13C MFA results. While discrimination of strains was obtained in both approaches in aerobic growth results, failure of 102 discrimination during anoxic growth using 13C labeling data alone was inconsistent with 13C MFA results. METHODS AND MATERIAL Bacterial strains and culture Pseudomonas aeruginosa strains are described in Table 3.1. Strains AA2, AA43, AA44, and LES 400 were obtained from the Belgian Co-ordinated Collections of Micro-organisms (Ghent, Belgium; LMG numbers:27630, 27631, 27632, and 27623 respectively). Pseudomonas aeruginosa strain AMT 0023-30 and AMT 0023-34 was obtained from Cystic Fibrosis Isolate Core (Seattle, Washington). Pseudomonas aeruginosa strain PAO1 was obtained from Dr. M. Mulks at Michigan State University. Synthetic Cystic Fibrosis sputum Media (SCFM), a defined minimal media designed to simulate both the carbon sources used by P. aeruginosa and the environment of the cystic fibrosis lung, was one of the culture medias used in this experiment [77].103 Strain Origin Related Isolates Relevant Characteristics References PAO1 Australia Wound Isolate (1955) N/A Commonly Used Reference Strain. [65,73,74] AMT 0023-30 USA CF - Pediatric Clinical Isolate (~1998) AMT 0023-34 Persister Cells Present [54,72] AMT 0023-34 USA CF - Pediatric Clinical Isolate (~2006) AMT 0023-30 Persister Cells Present LES 400 U.K. CF - Adult Clinical Isolate (2003) Liverpool Epidemic Strains (LES) Increased Acute Morbidity Risk Decreased Life Expectancy with Infection Pyrocyanin Overproduction Alginate Overproduction CF Transmissible [53,67Œ69,126] AA2 German CF Œ Adult Clinical Sequential Isolates (~1998 & ~2003) AA43, AA44 Parental strain of AA43, AA44 [52,70,71] AA43 AA2, AA44 7.5 years total lung colonization time Increased Acute Morbidity risk Motility defect Protease reduction LPS and PGN changes AA44 AA2, AA43 Table 3.1. List of Pseudomonas aeruginosa strains used in this experiment. LPS Œ Lipopolysaccharide; PGN Œ polygalacturan 104 For the purpose of 13C MFA labeling experiments, a reduced carbon source media based on SCFM was prepared (Simplified-SCFM), with the following changes to the carbon sources: L-glutamate 9.2 mM, L-alanine 11 mM, L-ornithine 4.3 mM, L-arginine 1.8 mM, and L-aspartate 4.9 mM and retaining all other described components (e.g. salts) [77]. These concentrations and substrates were determined by the major carbon compounds utilized by all experimental strains of P. aeruginosa during logarithmic growth of experimental cultures, the availability and relatively low cost of 13C used in this study, and the need to maintain the osmolaric concentration of SCFM. All cultures were pre-cultured from -80°C frozen stock samples onto the media of subsequent experimental culturing (SCFM or Simplified-SCFM) that was solidified with 1.5% agarose. Cultures were grown for 24 hours at 37° C to produce isolated colonies. For aerobic cultures, single colonies were transferred into 30 mL of SCFM or simplified-SCFM liquid media in 250 mL Bellco triple baffled shake flasks, incubated at 37°C and shaken at 120 RPM until stationary phase (12 to 14 hours) before spectrophotometric measurement of cell density as absorbance at 600nm (OD600) and inoculation of experimental cultures to initial densities of 0.01 OD600. For all anoxic cultures, glass pressure tubes were used and media concentrations of KNO3 were increased to 50mM. All pressure tubes were purged of oxygen content through degassing by bubbling the media with pure N2 gas for 15 minutes uncovered and then for 5 additional minutes covered by rubber stopper. All tubes were then fully closed by rubber stoppers, sealed in place by a metal crimp, and autoclaved. All culturing techniques involving puncturing the rubber stopper were executed by purging hypodermic needles with N2 gas before the removal or addition of liquid culture media. For anoxic culture inoculation, single colonies were transferred onto the side of pressure tubes filled with 10 mL of SCFM or simplified-SCFM liquid media in 20 mL pressure tubes, vigorously shaken for 10 105 minutes, and then incubated at 37° C until stationary phase (24 to 48 hours) before spectrophotometric measurement of cell density as absorbance at 600nm (OD600). Inoculation of experimental cultures occurred at initial densities of 0.01 OD600. For 13C labeling experiments, cells were cultured using either 75% (mol/mol) [3-13C] L-alanine or 99% (mol/mol) [U-13C] L-glutamate. Cell and media sample harvest Cell pellets were collected by centrifugation of cultures at mid-log phase (OD6000.5 for aerobic, OD6000.2 for anoxic). 1 mL of culture was centrifuged at 15000 g for 5 minutes. The supernatant was removed and 100 L of 6N HCl was added to the pellet which was stored at -20° C. Cell pellet sample preparation for amino acid analysis was based on previously reported methods [125,127]. Determination of substrate uptake rates and product secretion Amino acid uptake rates were determined using at least three biological replicate samples collected during log phase time points. Samples were filtered using Millipore 20 m pore size filters with filtered media collected and placed in -20° C for storage. Samples were quantified using a Waters Quattro micro API LC/MS/MS as described by Gu et al., 2012 [128]. Sample preparation used 100 L of samples diluted in 900 L of milipore filtered H2O with 100 M of phenylalanine-d8 as internal standard. Glucose and lactate uptake rates and acetate secretion rates for each strain were determined from 1 mL culture supernatants using at least three biological replicate samples for each of three log phase time points taken with 10 mM methylphosphamide added as internal standard. After lyophilization, samples were resuspended in 600 L of 99% D2O, and then lyophilized and resuspended in 600 L of 99% D2O. 1H-NMR was performed on an Agilent 106 DirectDrive2 500MHz instrument using an Alginate OneProbe with Protune for auto-tuning. Spectra were obtained at 500 MHz with a pulse angle of <45o, acquisition time of 2.05 s, and recycle delay of 2 s. No other secreted products were detected at significant levels in the media after culture growth for any of the strains. Calculation of uptake and secretion rates during exponential growth used the following equation: Where is intrinsic growth rate (hr-1); C is the concentration of glucose (mM) at time t; Xi is the initial density (gDW/L) of cells; t is time (hours); and v is uptake flux (mmol/g·DW-hr). Determination of CO2 efflux rates As described above, aerobic cultures were sealed with gas-tight caps and incubated for two hours until reaching early log phase. Due to the high sensitivity and low maximum range of CO2 detection of the LICOR CO2 measurement, P. aeruginosa growth was limited to a low range of optical density for this measurement. The cell density of the culture was measured and growth was stopped by adding 1 mL 6N HCl to the culture, which also converted dissolved bicarbonate to CO2. Total CO2 was measured using a LICOR LI-6400 with dry CO2 free air as input into the culture flask at a rate of 500 L/s. Readings of total CO2 were recorded over 5 minutes (1 s time resolution) to allow for removal of CO2 from the flask. Total CO2 evolved was measured in at least three replicate cultures with CO2 levels integrated for total CO2 efflux. CO2 efflux was calculated as described in determination of substrate uptake rates and product secretion. Determination of NO3 uptake rates Nitrate uptake rates were determined using at least three biological replicate samples collected during log phase time points of anoxic cultures. Samples were filtered using Millipore 20 m pore size filters with filtered media collected and placed in -20° C for storage. Sample 107 preparation used 100 L of samples diluted in 900 L of milipore filtered H20 with 10 M of 14NO3 as internal standard. A 3200 QTRAP® LC/MS/MS coupled to a Shimadzu (Columbia, MD) LC-20ADvp HPLC system and SIL-HTC autosampler was used for LCMS analysis. A ZIC®-pHILIC column (50 × 2.1 mm, 5 µm particle size) was used with column oven temperature at 40° C. The injection volume was 1 L, and the flow rate was 0.2 mL/min using water/acetonitrile (A/B) gradient at ambient temperature. The initial gradient (A/B)=10/90, was held until 2 minutes, followed by a linear gradient to 30/70 at 3 minutes, held at 30/70 until 4 minutes, followed by a ramp to 10/90 at 4.01. Mass spectra were acquired using turbo VTM ion source and multiple ion monitoring scan type in negative ion mode. Masses of 62 (unlabeled nitrate) and 63 (15N labeled nitrate) were measured at retention time of 1.3 min. To determine the concentration of 15N labeled and unlabeled nitrate, a standard curve was run for authentic 15N labeled nitrate standard with concentration range from 0 to 100 M. Data were processed for calibration and for quantification of the analytes with QTRAP analyst software. Analysis of Amino Acid labeling Cell pellets were suspended in 1mL of 6N HCl, incubated at 100° C for 24 hours, and dried at 60° C under a stream of N2. Amino acids were derivatized using N-Methyl-N-[tert-butyldimethyl-silyl] trifluoroacetimide (MTBSTFA,SIGMA-ALDRICH) in a 25 L pyrimidine and 25L 1% MTBSTFA solution incubated at 40° C for 1 hour as previously described [125,127,129]. GC-MS analyses of derivatized amino acids were performed on an Agilent 5973GC/quadrupole MS. GC-MS signals were corrected for natural isotope abundance [127]. Amino acid fragments used in 13C MFA were based on reported reliability [129]. Unlabeled amino acid samples collected were used to confirm the accuracy of natural abundance correction. 108 GC-MS amino acid data was corrected for natural abundance isotopic contents with average values and standard deviation calculated for biological replicates used as model inputs. Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA) GC-MS measured mass fragment mass abundance data for amino acids of strains harvested after growth in 75% 3-C13 L-alanine or 100% uniformly labeled (1,2,3,4,5-C13) L-glutamate was collected. Amino acid fragments that were measureable in three biological replicates were then processed in MATLAB (2010a) using the PCA function to produce a principal component analysis [130]. HCA was performed on this data set as well as on fluxes determined by MFA that were normalized to each strain's glucose uptake rate. HCA was implemented in the programming language R using complete linkage criteria which use Euclidian distance [131]. 13C Metabolic Flux Analysis A carbon isotopic network of Pseudomonas aeruginosa central metabolism was constructed by an approach similar to that previously used for E coli and previous P. aeruginosa work [125,132].The iM01056 FBA model [116] reaction network was simplified by only including carbon atoms of reactants and products, by introducing both net and exchange fluxes for reversible fluxes, by combining linear reaction sequences that do not alter carbon positions, and by condensing secondary metabolic pathways into a growth equation that consumes central metabolic intermediates as precursors to produce cellular biomass as previously defined [116,133]. Additionally, the preservation of the architecture of the FBA model allowed the results of FBA and 13C MFA to be directly compared and the 13C MFA results to be used to constrain the FBA [125]. 109 Measured rates of glucose uptake, CO2 efflux, and 13C labeling patterns in amino acids, together with growth rates, were used to estimate internal fluxes of central carbon metabolism [133] by fitting flux values to the experimental data using the 13C-FLUX software as previously described [92,125,134]. For labeling data, variation among biological replicates was very small (see Supplemental 3.1); since experimental standard deviations (SDs) are not a reliable estimate of true population SDs when 5 or fewer replicates are analyzed, the experimentally observed SD™s were increased to 2% of mean values for MFA modeling as previously described [125,132]. Measurements with a replicate standard deviation greater than 2% were adjusted to this value to account for the measurement uncertainty. Measurements with only one replicate had their deviation adjusted to 5% of the mean value. This also avoids excessively constraining modeling results to labeling data at the expense of direct flux measurements and reduces the potential for distortions due to precise but inaccurate mass isomer quantification [125,132]. To allow for uncertainties in P. aeruginosa biomass compositions, production rates were constrained to be within 50% of the E. coli biomass values [125]. To minimize the risk that solutions represented local rather than global optimization minima, multiple randomly generated initial fluxes constrained by sampling the feasible solution space were used. First, at least 100 randomly generated feasible starting points that produced optimized fits by the 13C-FLUX program were found. Next, the 10 starting points that yielded the lowest final residuum values were used in the second stage to generate 1000 more starting points by randomly perturbing the 10 starting points to yield 100 new points each. The final, lowest residuum optimized flux values that resulted in the best fit to the data were used. Confidence intervals for flux values were estimated using a Monte Carlo approach to randomly generate values of: biomass, glucose uptake rate, CO2 efflux rate, and amino acid labeling data for each 110 strain [135] based on the experimentally determined standard deviations. At least 20 such datasets were then fitted, using the 10 best fit (lowest SSres) starting and ending flux values determined for each strain as described above to yield best-fit flux maps for each strain. These flux value sets were then used to calculate 90% confidence intervals for each flux modeled for each strain. All MFA computations were performed at the High Performance Computer Center, Michigan State University, using a parallel 1536 core cluster of 192 nodes (two four-core Intel Xeon E5620s at 2.4 GHz with 24 GB of RAM and 250 GB local disk space per node). Global SSres for each strain is listed in Supplemental 3.2. Flux Balance Analysis The genome-derived stoichiometric Pseudomonas aeruginosa model iM01056 developed by Oberhardt et al. with previously described updates were used for FBA [116,132]. The model was modified to more fully account for alanine catabolism (see Supplemental 3.3), resulting in a total of 1016 reactions and 876 metabolites. The COBRA Toolbox 2.0 in MATLAB (2014a) with the implementation of the Gurobi 6 optimizer as optimization solver was used for FBA, using the objective function of maximal biomass production [97]. FBA simulations used measured amino acid uptake rates, NO3 uptake rates, and acetic acid and CO2 efflux rates by strain. To determine the range of fluxes that allow 99% of the maximal growth rates, the Flux Variability Analysis (FVA) function of the COBRA Toolbox was used [98]. RESULTS To determine the preferred carbon sources in the P. aeruginosa isolates in the cystic fibrosis lung and to reduce the complexity and cost in 13C substrate labeling experiments, strains were grown under aerobic and anoxic conditions in the defined complex Synthetic Cystic 111 Fibrosis sputum Media (SCFM) [77]. Filtered media samples of P. aeruginosa strains grown in SCFM under aerobic conditions were quantified using HPLC with the results shown in Figure 3.1. The average concentrations and uptake rates of P. aeruginosa grown in SCFM under aerobic conditions measured at early-log phase growth demonstrate average strains preference for L-lactate, ornithine, proline, alanine, glycine, leucine, lysine and aspartate under uptake rates, carbon uptake rates, and lowest % remaining compared to original media concentrations. Individual strain results (see Supplemental Table S3.1) demonstrate strain AA43 having little to no uptake of leucine and lysine and strain AMT0023-30 having little uptake of glycine under aerobic SCFM growth. Consideration of similar catabolic pathway utilization allows for the selection of L-lactate or alanine as well as glutamate or proline, with little reduction in determination of metabolic pathways. Anoxic growth in SCFM results also demonstrate similarities in uptake preference in aerobic SCFM, with a notable increase in leucine consumption and decrease in L-Lactate (see Supplemental Table S3.2). These combined results led to a proposed Simplified-SCFM (S-SCFM) for 13C labeling experiments containing the amino acids of glutamate, alanine, ornithine, arginine (included for proposed anoxic growth utilization in P. aeruginosa [136]), and aspartate in P. aeruginosa strains. Similar amino acid uptake preferences were also demonstrated by Palmer et al., with the exception of ornithine [77]. Comparisons of P. aeruginosa intrinsic growth rates in Simplified-SCFM and SCFM under aerobic and anoxic conditions for each strain are shown in Figures 3.2 and 3.3. Figure 3.2 demonstrates similar growth rates (within 20%) between S-SCFM and SCFM of all strains under aerobic and anoxic growth. Figure 3.3 demonstrates the similarity of growth rate order between strains under aerobic and anoxic conditions, with high R² values demonstrating similarity of growth to SCFM (see Supplemental Figure S3.1). 112 Figure 3.1. Analysis of amino acids, glucose, and lactate consumption in aerobic Synthetic Cystic Fibrosis sputum Media (SCFM) growth of seven strains of Pseudomonas aeruginosa. Media harvest occurred at early logarithmic growth (~3-5 hrs post inoculation, with strain AMT 0023-34 at ~8 hrs post inoculation). A - Averaged values of compound uptake rates, units are in mMol/(g·DW-hr). B - Averaged values of the carbon uptake, Units are in C-mMol/(g·DW-hr). C - Averaged values of the % of the compound remaining in SCFM media. D - % of total carbon uptake from the media by compound. All measures are n=21 (3 measures per strain); standard deviation ranges are represented as error bars. 113 Figure 3.2. Analysis of amino acids, glucose, and lactate consumption in anoxic Synthetic Cystic Fibrosis sputum Media (SCFM) growth of seven strains of Pseudomonas aeruginosa. Media harvest occurred at early logarithmic growth (~ 3-5 hrs post inoculation, with strain AMT 0023-34 at ~8 hrs post inoculation). A - Averaged values of compound uptake rates, units are in mMol/(g·DW-hr). B - Averaged values of the carbon uptake, Units are in C-mMol/(g·DW-hr). C - Averaged values of the % of the compound remaining in SCFM media. D - % of total carbon uptake from the media by compound. All measures are n=21 (3 measures per strain); standard deviation ranges are represented as error bars. 114 Figure 3.3. Specific growth rates of Pseudomonas aeruginosa strains under the growth conditions. Aerobic SCFM (black; AA2 n=4, LES400 n=4, others n=5), Aerobic Simplified-SCFM (dark grey; n=3), Anoxic SCFM (light grey; PA01 n=3, all others n=5), and Anoxic Simplified-SCFM (white; n=3). Error bars represent standard deviation of measurements. 115 Uptake and efflux rates determined under aerobic and anoxic growth in S-SCFM are shown in Tables 3.2 and 3.3 and in Figures 3.4 and 3.5. Strains growing under aerobic S-SCFM uptake rates in Table 3.2 demonstrate a nearly threefold difference in alanine (12.15 mmol gDW-1 hr-1 vs 5.85 mmol gDW-1 hr-1), arginine (0.94 mmol gDW-1 hr-1 vs 3.00 mmol gDW-1 hr-1), and ornithine (0.52 mmol gDW-1 hr-1 vs 1.54 mmol gDW-1 hr-1) uptake rates, a nearly sixfold difference in the uptake rate of aspartate (0.59 mmol gDW-1 hr-1 vs 3.52 mmol gDW-1 hr-1), and a twofold difference in glutamate uptake rate (2.51 mmol gDW-1 hr-1 vs 5.69 mmol gDW-1 hr-1). Total percentage uptake of carbon by compound demonstrates a range of amino acid by uptake as well, shown in Figure 3.4. The two compounds that make up a majority of the carbon uptake in each strain are alanine and glutamate. The cross-strain ranges of the percentage of total carbon uptake made up by these two compounds demonstrates large variation in the percentage of alanine in total carbon uptake (23% in PA01 to 50% in AA44) and narrower variation in the percentage of glutamate in total carbon uptake (22% in AA44 to 37% in PA01). Measured uptake rates of amino acids by P. aeruginosa during growth in anoxic S-SCFM also demonstrate large ranges over substrate uptakes. As shown in Table 3.3, alanine uptake rates demonstrate a fourfold difference (15.09 mmol gDW-1 hr-1 vs 1.43 mmol gDW-1 hr-1); arginine, a sixfold difference (1.43 mmol gDW-1 hr-1 vs 0.22 mmol gDW-1 hr-1); aspartate, a fivefold difference (6.88 mmol gDW-1 hr-1 vs 1.24 mmol gDW-1 hr-1); and glutamate, a fivefold difference (6.88 mmol gDW-1 hr-1 vs 1.24 mmol gDW-1 hr-1). Additionally, ornithine uptake rates demonstrate a range of 19% excretion of total carbon uptake to 16% of total carbon uptake (3.79 mmol gDW-1 hr-1 vs -0.97 mmol gDW-1 hr-1). Figure 3.5 also demonstrate the ranges of total carbon uptake by amino acid. These rates also demonstrate ranges of 35% (PA01) to 69% (AA2) for alanine and 12% (AA2 and AMT 0023-34) to 29% (PA01 and AA43) for glutamate, the two compounds 116 which collectively form the majority of carbon uptake. Based on the consistency of carbon flow in both anoxic and aerobic growth, a parallel 13C labeling experimental scheme of 75% 13C-3 alanine and 100% U-13C glutamate was used for 13C labeling for the MFA studies. Proteogenic 13C labeled amino acids obtained from parallel 13C labeling experiments were used to construct a Principal Component Analysis plot of the aerobic and anoxic strain in this experiment to determine strain relationships under aerobic and anoxic growth. Figure 3.6A shows clear differences under aerobic growth of strains, while Figure 3.6B indicates that strains AA2, AA44, and LES400 demonstrate labeling overlap in anoxic growth. Full proteogenic amino acid labeling data can be found in electronic Supplemental 3.1. The 13C MFA flux map for reference strain PA01 during aerobic growth of S-SCFM is shown in Figure 3.7, with the list of fluxes shown in the flux map provided in Table 3.4 (the complete flux listing can be found in electronic Supplemental 3.2). The MFA model shows the funneling of carbon through ornithine from glutamate towards succinate with only a third of the incoming glutamate flux towards -ketoglutarate directly. The flux model also demonstrates cycling of carbon from the TCA (tricarboxcylic acid) cycle in the form of malate to pyruvate and then back as citrate. The model also demonstrates flux through the glyoxylate cycle with all glyoxylate being converted to phosphoenolpyruvate. The Entner-Doudoroff and Oxidative Pentose Phosphate Pathways have little relative carbon flux through them, and demonstrate minimal flux allowing for necessary biosynthesis of bases and amino acids for biomass production. Most fluxes in this model demonstrate 90% CI values of 20% or less of the total flux, with the exception of the urea cycle components in this model demonstrating potential reversibility and utilization of this pathway.117 Strain Alanine Arginine Aspartate Glutamate Ornithine Total Carbon Uptake CO2 Efflux PA01 5.97±1.38 1.49±0.26 3.52±0.89 5.69±1.51 1.54±0.47 77.1±19.1 31.25±6.41 AA2 6.60±1.22 0.94±0.29 1.79±0.11 3.80±0.63 0.92±0.29 56.2±10.4 31.4±1.68 AA43 5.85±1.26 1.29±0.29 1.32±0.08 2.51±0.63 0.52±0.26 45.8±10.3 43.1±5.78 AA44 10.17±2.09 1.32±0.36 1.52±0.44 2.75±1.44 0.59±0.32 61.2±18.96 31.2±1.48 LES400 10.88±1.85 1.12±0.21 2.10±0.97 4.62±1.49 1.18±0.78 76.7±22.0 27.9±3.28 AMT 0023-30 12.15±2.53 3.00±0.51 2.35±1.51 4.97±2.38 1.24±0.35 94.9±30.3 73.0±3.08 AMT 0023-34 7.29±1.79 1.22±0.40 0.59±0.31 3.25±0.24 0.66±0.19 51.1±11.2 44.7±3.47 Table 3.2. Uptake and CO2 efflux rate of Pseudomonas aeruginosa strains under aerobic growth conditions in Simplified-SCFM. All units are in mMol/(g-DW-hr) except total carbon uptake, which is in C-mMol/(g·DW-hr) For amino acid uptake rates: PA01 n=6; AA2 n=5; AA44, LES400, AMT30 n=4; AA43 and AMT34 n=3. For CO2 efflux rates n=6 for PA01, AA43, AA44, and AMT 0023-34; n=3 for AA2, LES400, AMT 0023-30. Strain Alanine Arginine Aspartate Glutamate Ornithine Total Carbon Uptake Nitrate Uptake Acetate Secretion Pa01 6.82±0.37 0.29±0.09 2.59±0.35 3.33±0.36 1.80±0.19 58.2±5.82 17.3±3.68 0.68±0.16 AA2 11.116.6 1.051.17 2.914.41 0.162.28 -0.25-0.45 49.883.6 103±37.3 3.00±0.65 AA43 3.57±0.08 0.26±0.05 1.24±0.01 1.68±0.18 0.63±0.33 28.8±3.13 13.3±1.78 0.28±0.09 AA44 6.21±0.72 0.22±0.12 1.44±0.94 1.95±0.68 0.37±0.77 37.3±13.9 68.8±0.96 0.74±0.14 LES400 15.1±2.81 1.43±0.16 6.88±1.24 3.243.35 3.79±2.49 107±27.6 28.5±8.09 1.36±0.49 AMT 0023-30 10.915.1 0.911.34 1.411.60 0.861.83 -0.690.39 40.470.8 38.2±9.91 ND AMT 0023-34 4.76±1.07 0.51±0.23 1.55±1.53 0.65±0.80 -0.97±1.42 21.9±21.8 27.6±3.91 ND Table 3.3. Amino acid and nitrate uptake rates and acetate efflux rates of Pseudomonas aeruginosa strains under anoxic growth conditions in Simplified-SCFM. All units are in mMol/(g·DW-hr) except Total carbon uptake, which is in C-mMol/(g·DW-hr). Measurements with only two measurements are given as the range of measurement. Amino acid uptake rates for strains AA2 and AMT 0023-30, n=2; LES400 Glutamate measurement n=2; all others n=3. Nitrate uptake rates n=3. Acetate secretion rates for strains for PA01 and AA44 n=4; all others n=3; ND= not detected. 118 Figure 3.4. Percent of total carbon uptake from Simplified Synthetic Cystic Fibrosis Media (S-SCFM) during aerobic growth by Pseudomonas aeruginosa strains by compound. Alanine (black) Arginine (white), Aspartate (dark grey), Glutamate (medium grey), and Ornithine (light grey). PA01 n=6; AA2 n=5; AA44, LES400, AMT30 n=4; AA43 and AMT34 n=3. Error bars represent standard deviation of measurements. 119 Figure 3.5. Percent of total carbon uptake from Simplified Synthetic Cystic Fibrosis Media (S-SCFM) during anoxic growth by Pseudomonas aeruginosa strains by compound. Alanine (black) Arginine (white), Aspartate (dark grey), Glutamate (medium grey), and Ornithine (light grey). Uptake rates for strains AA2 and AMT 0023-30, n=2; LES400 Glutamate measurement n=2; all others n=3. Error bars represent standard deviation of measurements or as ranges where measurements are n=2. 120 Flux Balance Analysis of P. aeruginosa reference strain PA01 was performed using a modified genome-derived model together with the amino acid uptake rates measured for PA01 under aerobic conditions. A predicted flux map based on maximum growth efficiency (maximal biomass production for amino acid uptake rates) as the objective function is shown in Figure 3.8, with the listing of flux ranges for 99% of maximal biomass determined by Flux Variability Analysis shown. The FBA results demonstrate striking differences from 13C MFA flux, showing that optimal growth in P. aeruginosa strain PA01 would involve carbon flux into L-glutamate and then into the TCA cycle via -ketoglutarate. Most carbon flux from the TCA cycle would be through the conversion of oxaloacetate to phosphoenolpyruvate under optimized growth. The normalized to total carbon uptake 13C MFA flux results for all strains grown aerobically are shown in Figure 3.9 with values listed as a percentage of total carbon uptake through each reaction pathway for each strain. A complete listing of absolute fluxes in the figure can be found in Table 3.4 and a full result listing in electronic Supplemental 3.2. These results demonstrate a carbon flux similar to that for PA01, with movement of glutamate to ornithine and then to the TCA cycle making up a range of 30%-77% of the total uptake carbon for all strains. Strains AA44 and AMT0023-30 utilize glutamate dehydrogenase to move carbon from the TCA cycle to glutamate. The -ketoglutarate to succinate flux shows great variability in the strains, with a range of no flux to more than half of the carbon flux through this pathway across all of the strains. The MFA demonstrates cycling of carbon from the TCA cycle in the form of malate to pyruvate and then back as citrate in all of the strain models. The fluxes (Table 3.4) also demonstrate little flux through the EDP and OPPP pathways for each strain. Figure 3.6. Determination of Strain Diversity by isotopomers of proteinogenic amino acids and their fragments during the isotopomers of proteinogenic amino acids and their fragments during biological replicates from 100% 3-C13 Alanine and 100% UAMT0023-30 (filled circle), AMT0023-34 (empty circle), AA2 (triangle), AA43 (square), AA44 (diamond), LES 400 (pentagram), and PA01 (hexagram). 121 . Determination of Strain Diversity by 13C-Labeling fingerprinting. 3.6A: Principal Component Analysis of the isotopomers of proteinogenic amino acids and their fragments during aerobic growth. Figure 3.6B: Principal Component Analysis of the isotopomers of proteinogenic amino acids and their fragments during anoxic growth. Steady state labeling was measured in three Alanine and 100% U5-C13 Glutamate Aerobic experiments (see materials and methods). Strains: 34 (empty circle), AA2 (triangle), AA43 (square), AA44 (diamond), LES 400 (pentagram), Principal Component Analysis of the : Principal Component Analysis of growth. Steady state labeling was measured in three nts (see materials and methods). Strains: 34 (empty circle), AA2 (triangle), AA43 (square), AA44 (diamond), LES 400 (pentagram), Figure 3.7. Flux map from 13C MFA analysis for strain PA01 grown Simwas grown in defined S-SCFM with either 100% 1labeling during exponential growth. Labeling and external fluxes were measured aArrow thicknesses are proportional to net fluxes. Numbers represent net carbon fluxes in units of cgiven +/- 90% confidence intervals (CI). Non-standard or potentially ambiguG3P,Glyceraldehyde 3-phosphate; H6P, hexose-citrulline. 122 MFA analysis for strain PA01 grown Simplified-SCFM under aerobic conditions.SCFM with either 100% 1-C13 or 20% U6-C13 Glucose (n=3 replicates for each substrate) to steady state labeling during exponential growth. Labeling and external fluxes were measured and modeled as described in materials and methods. Arrow thicknesses are proportional to net fluxes. Numbers represent net carbon fluxes in units of c-mmol flux/(gDWŁhr); values are standard or potentially ambiguous abbreviations: 6PGC, 6-phospho-6-phosphate; P5P, Pentose-5-Phosphate; G5SH, glutamate-5-semialdehyde; CITR, SCFM under aerobic conditions. Strain PA01 Glucose (n=3 replicates for each substrate) to steady state nd modeled as described in materials and methods. mmol flux/(gDWŁhr); values are phospho-gluconate; semialdehyde; CITR, 123 Table 3.4. Net and exchange (Xch) fluxes determined by 13C MFA of aerobic SCFM with 90% confidence intervals. (See materials and methods for description of CI calculation) Residuum is the minimal sum-of-squares optimized value found using C13-Flux. Abbreviations as for Figure 3.7, NA - flux description is not applicable. 124 Figure 3.8. Flux map from FBA analysis for strain PA01 grown in Simplified-SCFM under aerobic conditions. Uptake of amino acids, production of CO2, and growth rate were used to constrain a stoichiometric genome-derived metabolic model (see materials and methods) with the objective function being maximal growth (corresponding to maximal growth yield on S-SCFM defined amino acids). Numbers are net fluxes as in Figure 3.4. Ranges are from flux variability analysis (range of each flux consistent with 99-100% of maximal growth). 125 The 13C MFA flux map for reference strain PA01 during anoxic growth in S-SCFM is shown in Figure 3.10, with the complete list of fluxes listed in Table 3.5. As seen in the aerobic 13C MFA results, carbon flux from glutamate is moved through ornithine from glutamate towards succinate with additional carbon converted from -ketoglutarate to glutamate. The glyoxylate cycle is utilized heavily in anoxic conditions, as is an incomplete TCA cycle. Most fluxes in this model demonstrate 90% CI values of 20% or less of the total flux, with the exception of the urea cycle components in this model demonstrating potential reversibility and utilization of this pathway. The FBA results for optimized biomass growth under anoxic conditions with the FVA determined ranges of 99% optimized biomass are shown in Figure 3.11. The FBA results demonstrate movement of amino acid carbon from ornithine to glutamate to be taken into the TCA cycle as -ketoglutarate. FBA also demonstrates full utilization of the TCA cycle and no flux through the glyoxylate cycle. These results stand in stark contrast with the 13C MFA results for PA01. The FBA results are in agreement with the MFA results of little carbon flux through the EDP and OPPP.126 Figure 3.9. Averaged carbon-uptake normalized flux map from 13C MFA analysis of P. aeruginosa cystic fibrosis isolates grown in aerobic Simplified-SCFM conditions. MFA determined c-mol fluxes of individual experimental strains normalized to total carbon uptake averaged with standard deviation determined. Arrow thicknesses are proportional to averaged % of total carbon uptake. Arrows are colorized to demonstrate variability of total uptake carbon through experimental strains. 127 Figure 3.10. Flux map from 13C MFA analysis for strain PA01 grown in Simplified-SCFM under anoxic conditions. Strain PA01 was grown in defined S-SCFM with either 100% 1-C13 or 20% U6-C13 Glucose (n=3 replicates for each substrate) to steady state labeling during exponential growth. Labeling and external fluxes were measured and modeled as described in materials and methods. Arrow thicknesses are proportional to net fluxes. Numbers represent net carbon fluxes in units of c-mmol flux/(gDWŁhr); values are given +/- 90% confidence intervals (CI). 128 Table 3.5. Net and exchange (Xch) fluxes determined by 13C MFA of anoxic SCFM with 90% confidence intervals. (See materials and methods for description of CI calculation) Residuum is the minimal sum-of-squares optimized value found using C13-Flux. Abbreviations as for Figure 3.7, NA - flux description is not applicable.129 Net carbon fluxes normalized to total carbon uptake rates of the 13C MFA results for all strains are shown in Figure 3.12, with values listed as a percentage of total carbon uptake through each reaction pathway for each strain. A complete listing of fluxes in the figure can be found in Table 3.5 and a full result listing in Supplemental 3.2. These results demonstrate that all strains but AMT0023-34 have an incomplete TCA cycle, with most strains (besides PA01 and LES400) transferring carbon from -ketoglutarate to isocitrate into the glyoxylate cycle. The glyoxylate cycle also has a large carbon flux for all strains, having a range of 57%-87% of total input carbon. These results also demonstrate the cycling of carbon from the TCA cycle in the form of malate to pyruvate and then back as citrate. The movement of carbon towards the OPPP and ED pathways represents 11%-29% of total carbon uptake. 130 Figure 3.11. Flux map from FBA analysis for strain PA01 grown in Simplified-SCFM under anoxic conditions. Uptake of amino acids, production of CO2, and growth rate were used to constrain a stoichiometric genome-derived metabolic model (see materials and methods) with the objective function being maximal growth (corresponding to maximal growth yield on S-SCFM defined amino acids). Numbers are net fluxes as in Figure 3.4. Ranges are from flux variability analysis (range of each flux consistent with 99-100% of maximal growth). 131 Figure 3.12. Flux map from 13C MFA analysis of P. aeruginosa cystic fibrosis isolates grown in anoxic Simplified-SCFM conditions. MFA determined c-mol fluxes of individual experimental strains normalized to total carbon uptake averaged with standard deviation determined. Arrow thicknesses are proportional to averaged % of total carbon uptake. Arrows are colorized to demonstrate variability of total uptake carbon through experimental strains. 132 Figure 3.13. Hierarchical Cluster Analysis of MFA derived flux values. Hierarchical Clustering Analysis using MFA derived flux values was employed to identify the main groupings as discussed in the text. A- Results from the Aerobic S-SCFM 13C MFA determined fluxes. B - Results from the Anoxic S-SCFM 13C MFA determined fluxes. 133 Table 3.6. Flux ratio tables of experimental strains in major points of carbon diversion in central metabolism and estimated energy production of MFA determined strains flux rates. Flux ratios and cofactor production rates for P. aeruginosa strains. The ratios of fluxes at major central metabolic branch points, as discussed in the text, with significant flux and strain variability shown in Figure 3.9 and 3.13. Estimated cofactor production rates are from MFA determined flux maps normalized to the glucose uptake rate for each strain (direct summation of FBA results listed in brackets). ATP, substrate level phosphorylation; NADP(H) total cellular NADH + NADPH production. Flux ratios calculated using: GLX/TCA = V(OAA+AC-COACIT)/V(ICITGLX+SUCC); ORN/GLU =V(ORNSUCC+CO2)/V(GLUKG) ; PEP/PYR = V(OAAPEP+CO2)/V(MALPYR+CO2) ; G6P/R5P =V(G3P+G3PG6P)/V(G3P+S7PE4P+R5P). 134 Table 3.7. MFA constrained FBA to FBA maximized growth results. Results higher than 100% correspond to increased production compared to growth optimized results. 135 Hierarchical Cluster Analysis of the carbon uptake normalized 13C MFA flux results are shown in Figure 3.13.. These results demonstrate two flux patterns in the aerobic and anoxic results. Under aerobic growth (Figure 3.13A), strains LES 400 and AMT 0023-34 show the most similarity in comparison to the other strains. Under anoxic growth (Figure 3.13B), AMT0023-34, AA2 and AA43 show the most similarity in comparison to the other strains. To understand the flux differences that HCA has shown, the ATP, redox cofactors, and flux ratios of major metabolic flux divisions were calculated. As listed in Table 3.6, ratios of TCA cycle flux to glyoxylate cycle flux are higher under aerobic growth compared to similar to higher flux ratios through the glyoxylate cycle flux under anoxic conditions. Movement of carbon from glutamate or ornithine shows the utilization of ornithine to succinate conversion by all strains under aerobic growth, but via glutamate by strains AA2, AA43, and AMT 0023-34 under anoxic conditions. Most strains also demonstrate net consumption of NADPH and ATP during growth (except aerobic growth of AA43 and AMT0023-30). To determine differences in the ATP and redox cofactor production in relation to optimized biomass growth, FBA constrained by 13C MFA results was compared to FBA optimized biomass growth. As shown in Table 3.7, MFA constrained FBA shows increased NADPH and NADH production over biomass synthesis reduction. DISCUSSION Many studies of Pseudomonas aeruginosa have been conducted to better understand the various adaptations that happen during chronic infection of CF patients. A frequent concern related to P. aeruginosa metabolism in the CF lung has been understanding its metabolic flexibility under micro-aerobic to anoxic conditions, especially as the organism™s growth is likely to be under these conditions in the CF lung [77,78,137]. Additionally, various Omic measures of 136 serial CF isolates have been conducted to measure the genomic [48,49,51Œ54,56,57,75,138], transcriptomic [51,58Œ62], and proteomic [58,63,123] changes in chronic infection adaptations to the CF lung environment. This study provides the first direct quantization and comparison of the metabolic phenotypes of P. aeruginosa under CF lung-like conditions and the metabolic changes seen under oxygen and nitrate respiration. This study was based on the work of Palmer et al. on defining the likely carbon and salt environment seen in the CF lung [77]. While this media lacks a source of lipids, especially those seen in the CF lung, the findings in this study would be compatible with possible uptake of a lipid for catabolism as the TCA cycle is the principal catabolic pathway utilized by these strains [59]. This study also demonstrates an uptake preference of amino acids and lactate similar to that reported by Palmer et al., with the notable exception of ornithine uptake preference in the strains of this study [77]. Due to cost and availability concerns, 13C L-lactate was not used in this study. Simulated FBA results of L-lactate uptake under aerobic growth, using MFA flux rates as internal flux constraints with SCFM L-lactate uptake rates, did show ready uptake and increase in biomass synthesis rates in all strains (see Supplemental 3.3). These results showed a significant utilization of the enzyme L-Lactate dehydrogenase ubiquinone, which converted L-lactate to pyruvate with reduced ubiquinone available for cellular respiration. Additionally, leucine was a major carbon source in anoxic growth but was not used in these experiments to allow direct comparison to aerobic and anoxic growth. Simulated FBA results of SCFM measured leucine uptake rates with constrained MFA did demonstrate uptake in some strains and allowed for additional reductant production in L-leucine catabolism (see Supplemental 3.3). Finally, aromatic amino acids, which have been shown to increase virulence factor production in CF isolates of P. aeruginosa, were not included in this study [139]. While this may affect central 137 carbon metabolism due to increased production of virulence factors, the current MFA shows that production fluxes of these aromatic amino acids are minimal and thus likely to have a minimal direct effect on aromatic amino acid anabolism. The 13C MFA and FBA results demonstrate notable phenotypic properties of P. aeruginosa under CF lung-like growth conditions. With oxidative bursts a common anti-bacterial strategy of many species [140Œ143], the high rates of reductant production seen in these P. aeruginosa strains are notable. Comparisons of the 13C MFA constrained FBA to optimized growth also demonstrates both significant increases in NAD(P)H production over that in biomass optimized growth and the reduction in biomass production rates. FBA constrained by 13C MFA results and nitrate uptake rates intriguingly showed incomplete reduction of nitrate in some strains, being excreted as nitrite (see supplemental 3.3), pointing to the possibility of incomplete respiration of nitrate under anoxic conditions. The inability to demonstrate similarity of anoxic 13C labeled proteogenic amino acids PCA results to 13C MFA HCA results is again noted. The efforts to build a high throughput system for analysis of metabolic similarities and differences using 13C label patterns of metabolites while avoiding the time intensive work of 13C MFA is desirable [101,144]. While previous results showed that PCA of 13C labeled proteogenic amino acids may differentiate strain similarity but not true metabolic phenotypes [132], the results shown in this study demonstrate that 13C MFA determined fluxes can demonstrate strain phenotypes even in PCA determined strains with amino acid label similarities. Previous work on CF isolates™ metabolism in minimal media demonstrates some similarities to the findings in this work. The HCA of M9 aerobic growth and S-SCFM aerobic 138 growth reproduces the close flux relationships between strains AA44 and PA01 and between strains AA2 and AA43 [132]. During aerobic growth in S-SCFM, strains PA01 and LES 400 and strains AA44 and AA2 demonstrate greater differences than were shown in aerobic minimal media growth [132]. While not surprising results, this does demonstrate that different metabolic phenotypes can be expressed by each strain than can be predicted under more simplified media conditions. These results do represent important factors in the consideration of the role of metabolism in pathogenicity, as under both media conditions the utilization of the glyoxylate cycle is higher than would be predicted under maximized biomass growth demonstrated in FBA [132]. These studies™ results also indicate that metabolic phenotypes of multiple strains may fall in a few generalized categories, such as two phenotypes demonstrated in the previous study and those demonstrated in this study [132]. These fewer phenotypes resolved by 13C MFA stand in contrast to the range and number of genotypes seen in genomic based methods, but analysis of more strains would be needed to fully understand the metabolic flexibility of P. aeruginosa in chronic CF infection [48,52,53,56,60]. The glyoxylate cycle is important to consider both as a metabolic target for pharmaceutical development and as a potential pathogenic factor in P. aeruginosa infection. A study conducted on the reduction of pathogenicity of P. aeruginosa infection of alfalfa seeds seen with the removal of the glyoxylate cycle enzyme isocitrate lyase also demonstrated reduction in rates of lung infection [145]. The work in this study as well as that in previous studies that utilized M9 minimal media demonstrates the high utilization of this pathway even in growth without lipids, as the catabolism of lipids found in the lung has been proposed as a possible explanation of glyoxylate cycle utilization in chronic infection isolates of P. aeruginosa as well as other pathogenic organisms of the lung [58,132,146]. The 13C MFA results also 139 demonstrate the utilization of glyoxylate to form phosphoglycerate under anoxic conditions by using glyoxylate carboligase, 2-hydroxy-3-oxopropionate reductase NAD(P), and glycerate kinase, but not utilizing the common second enzyme of the glyoxylate cycle, malate synthase. This study also helps to shed light on published results of the various Omic techniques used to understand chronic infection adaptation of P. aeruginosa. The 13C MFA results contextualize the transcriptomic work on chronic infection isolates from CF lungs by Hoboth et al., with both showing the utilization of glutamate and ornithine metabolism [58]. Importantly, 13C MFA demonstrated that the carbon flow was the reverse of that theorized by Hoboth et al. [58]. Other published transcriptomic studies show results similar to this study's 13C MFA, with upregulation of ornithine and arginine catabolic pathways and downregulation of glutamate dehyrdogenase and EDP related enzymes [147]. The Entner-Doudoroff and Oxidative Pentose Phosphate Pathways may not be particularly important in planktonic growth in the CF lung, as MFA results demonstrate little carbon flux through these pathways, especially after biosynthesis of biomass components are taken into account. These results may be different in the CF lung, due to the presence of carbohydrates and direct oxidative stress, as well as measured transcriptomic changes of P. aeruginosa CF isolate strains demonstrating increased EDP and OPPP gene expression [51,148]. Putrescine and other polyamines are common molecules found in many organisms with multiple described functions (e.g. antioxidant, nucleic acid stabilization, acid resistance, sidephores, etc.) [149]. Pathogenesis in several bacterial species (e.g., Staphylococcus sp., Yersinia pestis, Bacillus anthracis, and Vibrio cholerae) has been linked to the expression of various polyamines [150]. The importance of polyamine production in biofilm formation of several bacteria species (e.g., Vibrio cholerae, Yersinia pestis, Bacillus subtilis, and 140 Staphylococcus aureus) has been more recently demonstrated, with defects in enzymes necessary for polyamine metabolism, or exposure to certain polyamines, reducing biofilm formation but not affecting growth [151Œ157]. The metabolism of P. aeruginosa heavily utilizing putrescine metabolism during catabolism of amino acids, as well as multiple biofilm forming species failing to form biofilms without intact polyamine metabolic pathways, may demonstrate a shared metabolism in biofilm forming pathogenic bacteria and a possible shared pharmaceutical target. Direct production and utilization of polyamines by P. aeruginosa has also been described. Polyamine localization to the outer membrane of P. aeruginosa have been described and shown to be protective against extracellular DNA, oxidative damage, and antibiotic action [158]. A high amount of putrescine production has been reported in P. aeruginosa strain PA01 grown under rich media conditions and showing decreased production when supplemented with addition of glucose or fructose to the media [159]. Reports of samples of CF sputum during exacerbation have shown an increase in putrescine concentration and may be explained by the metabolism of P. aeruginosa demonstrated in this work [160]. These possible utilizations of putrescine rely on the accumulation of the metabolite, and kinetic data for these enzymes would be beneficial in determining the possibility of this accumulation. Putrescine or agmatine being utilized as a form of signaling in P. aeruginosa may be another possible explanation of the 13C MFA results described in this work. Genetic and transcriptomic work by Chou et al. during growth of P. aeruginosa mutants to characterize metabolism of agmatine and putrescine demonstrated induction of enzymes responsible for catabolism when exposed to agmatine and acetyl-putrescine [161]. Chou et al. also report the induction of the oprH-phoP-phoQ operon in P. aeruginosa exposed to exogenous agmatine and putrescine, allowing for the expression of known antibacterial resistance genes [161]. A gene 141 that functions to enhance biofilm growth and whose promoter binds to agmatine, agu2ABCA, has been described and found in some P. aeruginosa CF isolates [162]. Other findings of various polyamine exposures at millimolar concentrations causing an increased susceptibility to penicillin-class antibiotics and polymyxin B but decreasing susceptibility towards other classes (aminoglycosides, quinolones, etc.) in P. aeruginosa add to evidence of the role of polyamines as a possible regulatory agent in this species, especially in the reactivation of persister cell growth [163Œ165]. Other organisms have also demonstrated some of these adaptations, with work in Escherichia coli showing the role of putrescine metabolism in persister cell development [166]. A study of modulation of biofilm formation though exogenous exposure of various polyamines in pathogenic bacteria also demonstrate a possible role of polyamines in biofilm formation signaling or as a metabolic factor, but this relationship remains to be fully elucidated [156,157]. ACKNOWLEDGEMENTS We wish to thank Matthew Juergens for performing the Hierarchical Clustering Analysis. We also wish to thank Seattle Children's Hospital Cystic Fibrosis Isolate Core (NIH P30 DK089507) for their donation of strain AMT0023-30 and the RTSF Mass Spectrometry and Metabolomics Core for at Michigan State University for their analytical support. Dr. Gemma Reguera graciously provided anoxic culturing technique training and equipment. Michael Opperman was supported in part by funds from the Cellular and Molecular Biology Graduate Program and College of Human Medicine at Michigan State University and by a MD-PhD Fellowship grant by Spectrum Health. 142 APPENDICES 143 APPENDIX A - Electronic Supplemental Descriptions Descriptions of the supplemental material uploaded as an electronic supplement of this dissertation. Supplemental 3.1: 13C labeling data of amino acids used in Principal Component Analysis and Metabolic Flux Analysis models of this study. Supplemental 3.2: Complete list of reactions in Metabolic Flux Analysis model with complete flux table results and list of biomass values set for each strain. Supplemental 3.3: Complete flux results for Flux Balance Analysis of biomass optimized growth of strains PA01, AA2, AA43, AA44, LES 400, AMT 0023-30, and AMT 0023-34 in aerobic and anoxic growth conditions. Flux Variability Analysis results for 99% and maximally optimized biomass flux rate are also shown. Additional modeling (e.g. lactate and leucine uptake) can also be found here. Alterations to the biomass equation and additional reactions added to Opperman et al. FBA model are also described. Supplemental 3.4: Current version of formatted Flux Balance Analysis model used in this study (.xls format).APPENDIX B Figure S3.1. Correlation of Pseudomonas aeruginosamedia and oxygen conditions.(squares) and anoxic (triangles) conditions plotted SCFM intrinsic growth rates against Simplified-SCFM intrinsic growth rates. Cocorrelations, demonstrating the similar biomass growth rates and thus metabolic conditions in both media conditions. 144 APPENDIX B - Additional Tables and Figures Pseudomonas aeruginosa strain intrinsic growth rates under media and oxygen conditions. Intrinsic growth rates of P. aeruginosa strains under aerobic (squares) and anoxic (triangles) conditions plotted SCFM intrinsic growth rates against SCFM intrinsic growth rates. Correlation coefficients for both conditions show strong correlations, demonstrating the similar biomass growth rates and thus metabolic conditions in train intrinsic growth rates under strains under aerobic (squares) and anoxic (triangles) conditions plotted SCFM intrinsic growth rates against rrelation coefficients for both conditions show strong correlations, demonstrating the similar biomass growth rates and thus metabolic conditions in 145 Figure S3.2. Carbon conversation efficiency (CCE) calculation of Pseudomonas aeruginosa strains under aerobic growth conditions in Simplified-SCFM with 90% CI. Empty bars show values calculated from amino acid uptake and biomass production; Filled bars show values calculated from CO2 efflux and biomass production. 2-sided student™s t-test of CCE demonstrate no significant difference (P >0.05) for each strain. CCE% calculation based on amino acid uptake for PA01 and AA43 n=6; AA2 n=5; AA44, LES400, and AMT 0023-30 n=3; AMT 0023-34 n=3. CCE% calculation based on CO2 efflux for PA01, AA44, AA43, and AMT 0023-34 n=6; all rest are n=3. 146 Figure S3.3. Nitrogen conversation (NC) percentage calculation of Pseudomonas aeruginosa strains under aerobic and anoxic growth conditions. Aerobic (filled) and anoxic growth (empty) conditions in Simplified-SCFM with standard deviation as error bars. NC% calculation based on aerobic growth for PA01 and AA43 n=6; AA2 n=5; AA44, LES400, and AMT 0023-30 n=3; AMT 0023-34 n=3. N=3 for all NCE% anoxic calculations. NCE% greater than 100% imply uptake of nitrogen from sources other than amino acids. NCE% less than 100% imply secretion of nitrogenous waste due to amino acid catabolism. 147 Table S3.1. Average and standard deviation of remaining carbon compounds during growth in Synthetic Cystic Fibrosis Media (SCFM) under aerobic conditions. Amino Acids measured by LCMS and remaining Glucose and Lactate measured by NMR. Table S3.2. Average and standard deviation of remaining carbon compounds during growth in Synthetic Cystic Fibrosis Media (SCFM) under Anoxic conditions. Amino Acids measured by LCMS and remaining Glucose and Lactate measured by NMR. 148 REFERENCES 149 REFERENCES 1. ANDERSEN DH. CYSTIC FIBROSIS OF THE PANCREAS AND ITS RELATION TO CELIAC DISEASE. Am J Dis Child. 1938;56: 344. doi:10.1001/archpedi.1938.01980140114013 2. Farber S. Some organic digestive disturbances in early life. J Mich Med Soc. 1945;44: 587. 3. Kerem B, Rommens JM, Buchanan J a, Markiewicz D, Cox TK, Chakravarti A, et al. Identification of the cystic fibrosis gene: genetic analysis. Science. 1989;245: 1073Œ80. Available: http://www.ncbi.nlm.nih.gov/pubmed/2570460 4. Rommens J, Iannuzzi M, Kerem B, Drumm M, Melmer G, Dean M, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science (80- ). 1989;245: 1059 Œ1065. doi:10.1126/science.2772657 5. Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, et al. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science. 1989;245: 1066Œ1073. doi:10.1126/science.2475911 6. Gustafsson JK, Ermund A, Ambort D, Johansson ME V, Nilsson HE, Thorell K, et al. Bicarbonate and functional CFTR channel are required for proper mucin secretion and link cystic fibrosis with its mucus phenotype. J Exp Med. 2012;209: 1263Œ72. doi:10.1084/jem.20120562 7. Steward MC, Ishiguro H, Case RM. Mechanisms of Bicarbonate Secretion in the Pancreatic Duct. Annu Rev Physiol. 2005;67: 377Œ409. doi:10.1146/annurev.physiol.67.031103.153247 8. Ishiguro H, Yamamoto A, Nakakuki M, Yi L, Ishiguro M, Yamaguchi M, et al. Physiology and pathophysiology of bicarbonate secretion by pancreatic duct epithelium. Nagoya J Med Sci. 2012;74: 1Œ18. Available: http://www.ncbi.nlm.nih.gov/pubmed/22515107 9. Choi JY, Muallem D, Kiselyov K, Lee MG, Thomas PJ, Muallem S. Aberrant CFTR-dependent HCO3- transport in mutations associated with cystic fibrosis. Nature. 2001;410: 94Œ97. doi:10.1038/35065099 10. Ambort D, Johansson ME V., Gustafsson JK, Nilsson HE, Ermund A, Johansson BR, et al. Calcium and pH-dependent packing and release of the gel-forming MUC2 mucin. Proc Natl Acad Sci. 2012;109: 5645Œ5650. doi:10.1073/pnas.1120269109 11. Verdugo P, Deyrup-Olsen I, Aitken M, Villalon M, Johnson D. Molecular Mechanism of Mucin Secretion: I. The Role of Intragranular Charge Shielding. J Dent Res. 1619 DUKE ST, ALEXANDRIA, VA 22314: AMER ASSOC DENTAL RESEARCH; 1987;66: 506Œ508. doi:10.1177/00220345870660022001 150 12. Nordman H, Davies JR, Herrmann a, Karlsson NG, Hansson GC, Carlstedt I. Mucus glycoproteins from pig gastric mucosa: identification ofdifferent mucin populations from the surface epithelium. Biochem J. 1997;326 ( Pt 3: 903Œ910. 13. Garcia MAS, Yang N, Quinton PM. Normal mouse intestinal mucus release requires cystic fibrosis transmembrane regulator-dependent bicarbonate secretion. J Clin Invest. 2009;119: 2613Œ2622. doi:10.1172/JCI38662 14. Quinton PM. Cystic fibrosis: impaired bicarbonate secretion and mucoviscidosis. Lancet. 2008;372: 415Œ417. doi:10.1016/S0140-6736(08)61162-9 15. Mousa HM, Woodley FW. Gastroesophageal reflux in cystic fibrosis: Current understandings of mechanisms and management. Curr Gastroenterol Rep. 2012;14: 226Œ235. doi:10.1007/s11894-012-0261-9 16. Kelly T, Buxbaum J. Gastrointestinal Manifestations of Cystic Fibrosis. Dig Dis Sci. Springer US; 2015;60: 1903Œ1913. doi:10.1007/s10620-015-3546-7 17. Stallings VA, Stark LJ, Robinson KA, Feranchak AP, Quinton H. Evidence-Based Practice Recommendations for Nutrition-Related Management of Children and Adults with Cystic Fibrosis and Pancreatic Insufficiency: Results of a Systematic Review. J Am Diet Assoc. 2008;108: 832Œ839. doi:10.1016/j.jada.2008.02.020 18. De Lisle RC, Borowitz D. The Cystic Fibrosis Intestine. Cold Spring Harb Perspect Med. 2013;3: a009753Œa009753. doi:10.1101/cshperspect.a009753 19. O™Sullivan BP, Freedman SD. Cystic fibrosis. Lancet. Elsevier Ltd; 2009;373: 1891Œ1904. doi:10.1016/S0140-6736(09)60327-5 20. Meyerholz DK, Stoltz DA, Pezzulo AA, Welsh MJ. Pathology of gastrointestinal organs in a porcine model of cystic fibrosis. Am J Pathol. American Society for Investigative Pathology; 2010;176: 1377Œ89. doi:10.2353/ajpath.2010.090849 21. Gibson-Corley KN, Meyerholz DK, Engelhardt JF. Pancreatic pathophysiology in cystic fibrosis. J Pathol. 2016;238: 311Œ320. doi:10.1002/path.4634 22. Mott LS, Park J, Murray CP, Gangell CL, de Klerk NH, Robinson PJ, et al. Progression of early structural lung disease in young children with cystic fibrosis assessed using CT. Thorax. 2011;67: 509Œ516. doi:10.1136/thoraxjnl-2011-200912 23. Sly P, Gangell C, Chen L, Ware R, Ranganathan S, Mott LS, et al. Risk factors for bronchiectasis in children with cystic fibrosis. N Engl J Med. 2013;368: 1963Œ70. doi:10.1056/NEJMoa1301725 24. O™Donnell AE. Bronchiectasis. Chest. 2008;134: 815Œ823. doi:10.1378/chest.08-0776 25. Bonfield TL, Panuska JR, Konstan MW, Hilliard KA, Hilliard JB, Ghnaim H, et al. Inflammatory cytokines in cystic fibrosis lungs. Am J Respir Crit Care Med. 1995;152: 2111Œ2118. doi:10.1164/ajrccm.152.6.8520783 151 26. Muhlebach MS, Stewart PW, Leigh MW, Noah TL. Quantitation of inflammatory responses to bacteria in young cystic fibrosis and control patients. Am J Respir Crit Care Med. 1999;160: 186Œ191. doi:10.1164/ajrccm.160.1.9808096 27. Muhlebach MS, Reed W, Noah TL. Quantitative Cytokine Gene Expression in CF Airway. Pediatr Pulmonol. 2004;37: 393Œ399. doi:10.1002/ppul.20010 28. Derichs N, Jin B-J, Song Y, Finkbeiner WE, Verkman a S. Hyperviscous airway periciliary and mucous liquid layers in cystic fibrosis measured by confocal fluorescence photobleaching. FASEB J. 2011;25: 2325Œ2332. doi:10.1096/fj.10-179549 29. Chen J-H, Stoltz DA, Karp PH, Ernst SE, Pezzulo AA, Moninger TO, et al. Loss of Anion Transport without Increased Sodium Absorption Characterizes Newborn Porcine Cystic Fibrosis Airway Epithelia. Cell. 2010;143: 911Œ923. doi:10.1016/j.cell.2010.11.029 30. Livraghi A, Randell SH. Cystic fibrosis and other respiratory diseases of impaired mucus clearance. Toxicol Pathol. 2007;35: 116Œ29. doi:10.1080/01926230601060025 31. Blatter J, Sweet S. Lung Transplantation in Cystic Fibrosis: Trends and Controversies. Pediatr Allergy Immunol Pulmonol. 2015;28: 237Œ243. doi:10.1089/ped.2015.0564 32. Flores JS, Rovedder PME, Ziegler B, Pinotti AFF, Barreto SSM, Dalcin P d. TR. Clinical Outcomes and Prognostic Factors in a Cohort of Adults With Cystic Fibrosis: A 7-Year Follow-Up Study. Respir Care. 2016;61: 192Œ199. doi:10.4187/respcare.04097 33. Lipuma JJ. The changing microbial epidemiology in cystic fibrosis. Clin Microbiol Rev. 2010;23: 299Œ323. doi:10.1128/CMR.00068-09 34. Elrod RP, Braun AC. Pseudomonas aeruginosa; its role as a plant pathogen. J Bacteriol. 1942;44: 633Œ645. 35. Yorgey P, Rahme LG, Tan MW, Ausubel FM. The roles of mucD and alginate in the virulence of Pseudomonas aeruginosa in plants, nematodes and mice. Mol Microbiol. 2001;41: 1063Œ1076. doi:10.1046/j.1365-2958.2001.02580.x 36. Qureshi MN, Perez AA, Madayag RM, Bottone EJ. Inhibition of Acanthamoeba species by Pseudomonas aeruginosa: rationale for their selective exclusion in corneal ulcers and contact lens care systems. J Clin Microbiol. 1993;31: 1908Œ10. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=265656&tool=pmcentrez&rendertype=abstract 37. Kerr JR. Suppression of fungal growth exhibited by Pseudomonas aeruginosa. J Clin Microbiol. 1994;32: 525Œ7. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=263067&tool=pmcentrez&rendertype=abstract 38. Migula W. System der bakterien. Fischer; 1897. 152 39. Moore ERB, Tindall BJ, Martins Dos Santos VAP, Pieper DH, Ramos J-L, Palleroni NJ. Nonmedical: Pseudomonas. In: Intergovernmental Panel on Climate Change, editor. The Prokaryotes. New York, NY: Springer New York; 2006. pp. 646Œ703. doi:10.1007/0-387-30746-X_21 40. Berger M. Lung inflammation early in cystic fibrosis: Bugs are indicted, but the defense is guilty. Am J Respir Crit Care Med. 2002;165: 857Œ858. doi:10.1164/rccm.2202030 41. Salsgiver EL, Fink AK, Knapp EA, LiPuma JJ, Olivier KN, Marshall BC, et al. Changing Epidemiology of the Respiratory Bacteriology of Patients with Cystic Fibrosis. Chest. Elsevier Inc; 2015;149: 390Œ400. doi:10.1378/chest.15-0676 42. Hogardt M, Heesemann J. Adaptation of Pseudomonas aeruginosa during persistence in the cystic fibrosis lung. Int J Med Microbiol. Elsevier GmbH.; 2010;300: 557Œ562. doi:10.1016/j.ijmm.2010.08.008 43. Didelot X, Walker AS, Peto TE, Crook DW, Wilson DJ. Within-host evolution of bacterial pathogens. Nat Rev Microbiol. Nature Publishing Group; 2016;14: 150Œ162. doi:10.1038/nrmicro.2015.13 44. Fauvart M, de Groote VN, Michiels J. Role of persister cells in chronic infections: Clinical relevance and perspectives on anti-persister therapies. J Med Microbiol. 2011;60: 699Œ709. doi:10.1099/jmm.0.030932-0 45. Lewis K. Persister cells. Annu Rev Microbiol. 2010;64: 357Œ372. doi:10.1146/annurev.micro.112408.134306 46. Lister PD, Wolter DJ, Hanson ND. Antibacterial-resistant Pseudomonas aeruginosa: Clinical impact and complex regulation of chromosomally encoded resistance mechanisms. Clin Microbiol Rev. 2009;22: 582Œ610. doi:10.1128/CMR.00040-09 47. Livermore DM. Multiple Mechanisms of Antimicrobial Resistance in Pseudomonas aeruginosa: Our Worst Nightmare? Clin Infect Dis. 2002;34: 634Œ640. doi:10.1086/338782 48. Chung JCS, Becq J, Fraser L, Schulz-Trieglaff O, Bond NJ, Foweraker J, et al. Genomic variation among contemporary Pseudomonas aeruginosa isolates from chronically infected cystic fibrosis patients. J Bacteriol. 2012;194: 4857Œ4866. doi:10.1128/JB.01050-12 49. Rau MH, Marvig RL, Ehrlich GD, Molin S, Jelsbak L. Deletion and acquisition of genomic content during early stage adaptation of Pseudomonas aeruginosa to a human host environment. Environ Microbiol. 2012;14: 2200Œ2211. doi:10.1111/j.1462-2920.2012.02795.x 50. Cramer N, Klockgether J, Wrasman K, Schmidt M, Davenport CF, Tümmler B. Microevolution of the major common Pseudomonas aeruginosa clones C and PA14 in cystic fibrosis lungs. Environ Microbiol. 2011;13: 1690Œ1704. doi:10.1111/j.1462-2920.2011.02483.x 153 51. Salunkhe P, Smart CHM, Morgan AW, Panagea S, Walshaw MJ, Hart A, et al. A Cystic Fibrosis Epidemic Strain of Pseudomonas aeruginosa Displays Enhanced Virulence and Antimicrobial Resistance. J Bacteriol. 2005;187: 4908Œ4920. doi:10.1128/JB.187.14.4908 52. Bragonzi A, Wiehlmann L, Klockgether J, Cramer N, Worlitzsch D, Döning G, et al. Sequence diversity of the mucABD locus in Pseudomonas aeruginosa isolates from patients with cystic fibrosis. Microbiology. 2006;152: 3261Œ3269. doi:10.1099/mic.0.29175-0 53. Jeukens J, Boyle B, Kukavica-Ibrulj I, Ouellet MM, Aaron SD, Charette SJ, et al. Comparative genomics of isolates of a Pseudomonas aeruginosa epidemic strain associated with chronic lung infections of cystic fibrosis patients. PLoS One. 2014;9: 1Œ15. doi:10.1371/journal.pone.0087611 54. Mena A, Smith EE, Burns JL, Speert DP, Moskowitz SM, Perez JL, et al. Genetic adaptation of Pseudomonas aeruginosa to the airways of cystic fibrosis patients is catalyzed by hypermutation. J Bacteriol. 2008;190: 7910Œ7917. doi:10.1128/JB.01147-08 55. Caballero JD, Clark ST, Coburn B, Zhang Y, Wang PW, Donaldson SL, et al. Selective sweeps and parallel pathoadaptation drive pseudomonas aeruginosa evolution in the cystic fibrosis lung. MBio. 2015;6: 1Œ14. doi:10.1128/mBio.00981-15 56. Darch SE, McNally A, Harrison F, Corander J, Barr HL, Paszkiewicz K, et al. Recombination is a key driver of genomic and phenotypic diversity in a Pseudomonas aeruginosa population during cystic fibrosis infection. Sci Rep. 2015;5: 7649. doi:10.1038/srep07649 57. Markussen T, Marvig RL, Gomez-Lozano M, Aanaes K, Burleigh AE, Hoiby N, et al. Environmental Heterogeneity Drives Within-Host Diversification and Evolution of Pseudomonas aeruginosa. MBio. 2014;5: e01592Œ14Œe01592Œ14. doi:10.1128/mBio.01592-14 58. Hoboth C, Hoffmann R, Eichner A, Henke C, Schmoldt S, Imhof A, et al. Dynamics of adaptive microevolution of hypermutable Pseudomonas aeruginosa during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2009;200: 118Œ130. doi:10.1086/599360 59. Son MS, Matthews WJ, Kang Y, Nguyen DT, Hoang TT. In vivo evidence of Pseudomonas aeruginosa nutrient acquisition and pathogenesis in the lungs of cystic fibrosis patients. Infect Immun. 2007;75: 5313Œ5324. doi:10.1128/IAI.01807-06 60. Huse HK, Kwon T, Zlosnik JEA, Speert DP, Marcotte EM, Whiteley M. Parallel evolution in Pseudomonas aeruginosa over 39,000 generations in vivo. MBio. 2010;1: e00199Œ10. doi:10.1128/mBio.00199-10.Editor 61. Harmer C, Alnassafi K, Hu H, Elkins M, Bye P, Rose B, et al. Modulation of gene expression by Pseudomonas aeruginosa during chronic infection in the adult cystic fibrosis lung. Microbiology. 2013;159: 2354Œ63. doi:10.1099/mic.0.066985-0 154 62. Varga JJ, Barbier M, Mulet X, Bielecki P, Bartell JA, Owings JP, et al. Genotypic and phenotypic analyses of a Pseudomonas aeruginosa chronic bronchiectasis isolate reveal differences from cystic fibrosis and laboratory strains. BMC Genomics. 2015;16: 883. doi:10.1186/s12864-015-2069-0 63. Park AJ, Murphy K, Krieger JR, Brewer D, Taylor P, Habash M, et al. A temporal examination of the planktonic and biofilm proteome of whole cell Pseudomonas aeruginosa PAO1 using quantitative mass spectrometry. Mol Cell Proteomics. 2014;13: 1095Œ105. doi:10.1074/mcp.M113.033985 64. Wu X, Held K, Zheng C, Staudinger BJ, Chavez JD, Weisbrod CR, et al. Dynamic Proteome Response of Pseudomonas aeruginosa to Tobramycin Antibiotic Treatment. Mol Cell Proteomics. 2015;14: 2126Œ2137. doi:10.1074/mcp.M115.050161 65. Stover C, Pham X, Erwin a, Mizoguchi S. Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature. 2000;406: 959Œ964. Available: http://www.nature.com/nature/journal/v406/n6799/abs/406959a0.html 66. Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science. 1999;284: 1318Œ1322. doi:10.1126/science.284.5418.1318 67. Cheng K, Smyth RL, Govan JR, Doherty C, Winstanley C, Denning N, et al. Spread of beta-lactam-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. Lancet. 1996;348: 639Œ642. doi:10.1016/S0140-6736(96)05169-0 68. Carter MEK, Fothergill JL, Walshaw MJ, Rajakumar K, Kadioglu A, Winstanley C. A subtype of a Pseudomonas aeruginosa cystic fibrosis epidemic strain exhibits enhanced virulence in a murine model of acute respiratory infection. J Infect Dis. 2010;202: 935Œ942. doi:10.1086/655781 69. Fothergill JL, Panagea S, Hart C a, Walshaw MJ, Pitt TL, Winstanley C. Widespread pyocyanin over-production among isolates of a cystic fibrosis epidemic strain. BMC Microbiol. 2007;7: 45. doi:10.1186/1471-2180-7-45 70. Bragonzi A, Paroni M, Nonis A, Cramer N, Montanari S, Rejman J, et al. Pseudomonas aeruginosa microevolution during cystic fibrosis lung infection establishes clones with adapted virulence. Am J Respir Crit Care Med. 2009;180: 138Œ145. doi:10.1164/rccm.200812-1943OC 71. Lorè NI, Cigana C, De Fino I, Riva C, Juhas M, Schwager S, et al. Cystic Fibrosis-Niche Adaptation of Pseudomonas aeruginosa Reduces Virulence in Multiple Infection Hosts. PLoS One. 2012;7: e35648. doi:10.1371/journal.pone.0035648 72. Mulcahy LR, Burns JL, Lory S, Lewis K. Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol. 2010;192: 6191Œ6199. doi:10.1128/JB.01651-09 155 73. Sauer K, Cullen MC, Rickard a H, Zeef L a H, Gilbert P, Davies DG. Characterization of Nutrient-Induced Dispersion in Pseudomonas aeruginosa PAO1 Biofilm. J Bacteriol. 2004;186: 7312Œ7326. doi:10.1128/JB.186.21.7312 74. Mavrodi D V, Bonsall RF, Delaney SM, Soule MJ, Phillips G, Thomashow LS. Functional Analysis of Genes for Biosynthesis of Pyocyanin and Phenazine-1-Carboxamide from Pseudomonas aeruginosa PAO1. J Bacteriol. 2001;183: 6454Œ6465. doi:10.1128/JB.183.21.6454 75. Clark ST, Diaz Caballero J, Cheang M, Coburn B, Wang PW, Donaldson SL, et al. Phenotypic diversity within a Pseudomonas aeruginosa population infecting an adult with cystic fibrosis. Sci Rep. Nature Publishing Group; 2015;5: 10932. doi:10.1038/srep10932 76. Hogardt M, Heesemann J. Microevolution of pseudomonas aeruginosa to a chronic pathogen of the cystic fibrosis lung. Curr Top Microbiol Immunol. 2013;358: 91Œ118. doi:10.1007/82_2011_199 77. Palmer KL, Aye LM, Whiteley M. Nutritional cues control Pseudomonas aeruginosa multicellular behavior in cystic fibrosis sputum. J Bacteriol. 2007;189: 8079Œ8087. doi:10.1128/JB.01138-07 78. Schobert M, Jahn D. Anaerobic physiology of Pseudomonas aeruginosa in the cystic fibrosis lung. Int J Med Microbiol. Elsevier GmbH.; 2010;300: 549Œ556. doi:10.1016/j.ijmm.2010.08.007 79. Govan JR, Deretic V. Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol Rev. 1996;60: 539Œ574. doi:10.1111/j.1365-2672.2007.03706.x 80. Rau MH, Hansen SK, Johansen HK, Thomsen LE, Workman CT, Nielsen KF, et al. Early adaptive developments of Pseudomonas aeruginosa after the transition from life in the environment to persistent colonization in the airways of human cystic fibrosis hosts. Environ Microbiol. 2010;12: 1643Œ1658. doi:10.1111/j.1462-2920.2010.02211.x 81. Terry JM, Piña SE, Mattingly SJ. Environmental conditions which influence mucoid conversion Pseudomonas aeruginosa PAO1. Infect Immun. 1991;59: 471Œ477. 82. Ryall B, Carrara M, Zlosnik JE a, Behrends V, Lee X, Wong Z, et al. The mucoid switch in pseudomonas aeruginosa represses quorum sensing systems and leads to complex changes to stationary phase virulence factor regulation. PLoS One. 2014;9. doi:10.1371/journal.pone.0096166 83. Hogardt M, Hoboth C, Schmoldt S, Henke C, Bader L, Heesemann J. Stage-specific adaptation of hypermutable Pseudomonas aeruginosa isolates during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2007;195: 70Œ80. doi:10.1086/509821 156 84. Rodríguez-Rojas A, Oliver A, Blázquez J. Intrinsic and environmental mutagenesis drive diversification and persistence of pseudomonas aeruginosa in chronic lung infections. J Infect Dis. 2012;205: 121Œ127. doi:10.1093/infdis/jir690 85. Maier T, Güell M, Serrano L. Correlation of mRNA and protein in complex biological samples. FEBS Lett. Federation of European Biochemical Societies; 2009;583: 3966Œ3973. doi:10.1016/j.febslet.2009.10.036 86. Haider S, Pal R. Integrated analysis of transcriptomic and proteomic data. Curr Genomics. 2013;14: 91Œ110. doi:10.2174/1389202911314020003 87. Maier T, Schmidt A, Güell M, Kühner S, Gavin A-C, Aebersold R, et al. Quantification of mRNA and protein and integration with protein turnover in a bacterium. Mol Syst Biol. 2011;7: 511. doi:10.1038/msb.2011.38 88. Wiechert W, de Graaf a a. Bidirectional reaction steps in metabolic networks: I. Modeling and simulation of carbon isotope labeling experiments. Biotechnol Bioeng. 1997;55: 101Œ17. doi:10.1002/(SICI)1097-0290(19970705)55:1<101::AID-BIT12>3.0.CO;2-P 89. Wiechert W, Siefke C, de Graaf a a, Marx a. Bidirectional reaction steps in metabolic networks: II. Flux estimation and statistical analysis. Biotechnol Bioeng. 1997;55: 118Œ135. doi:10.1002/(SICI)1097-0290(19970705)55:1<118::AID-BIT13>3.0.CO;2-I 90. HOLZHUTTER HG. COMPARTMENTAL ANALYSIS - THEORETICAL ASPECTS AND APPLICATION. Biomed Biochim Acta. MUHLENSTRASSE 33-34, D-13187 BERLIN, GERMANY: AKADEMIE VERLAG GMBH; 1985;44: 863Œ873. 91. Wiechert W, De Graaf A a. In vivo stationary flux analysis by 13C labeling experiments. Adv Biochem Eng Biotechnol. 1996;54: 109Œ54. 92. Wiechert W, Möllney M, Isermann N, Wurzel M, De Graaf A a. Bidirectional reaction steps in metabolic networks: III. Explicit solution and analysis of isotopomer labeling systems. Biotechnol Bioeng. 1999;66: 69Œ85. doi:10.1002/(SICI)1097-0290(1999)66:2<69::AID-BIT1>3.0.CO;2-6 93. Zupke C, Stephanopoulos G. Modeling of Isotope Distributions and Intracellular Fluxes in Metabolic Networks Using Atom Mapping Matrixes. Biotechnol Prog. 1994;10: 489Œ498. doi:10.1021/bp00029a006 94. Orth JD, Thiele I, Palsson BØ. What is flux balance analysis? Nat Biotechnol. 2010;28: 245Œ248. doi:10.1038/nbt.1614.What 95. Lewis NE, Nagarajan H, Palsson BO. Constraining the metabolic genotype-phenotype relationship using a phylogeny of in silico methods. Nat Rev Microbiol. 2012;10: 291Œ305. doi:doi:10.1038/nrmicro2737 96. Karr JR, Takahashi K, Funahashi A. The principles of whole-cell modeling. Curr Opin Microbiol. Elsevier Ltd; 2015;27: 18Œ24. doi:10.1016/j.mib.2015.06.004 157 97. Hyduke D, Hyduke D, Schellenberger J, Que R, Fleming R, Thiele I, et al. COBRA Toolbox 2.0. Protoc Exch. 2011; doi:10.1038/protex.2011.234 98. Mahadevan R, Schilling CH. The effects of alternate optimal solutions in constraint-based genome-scale metabolic models. Metab Eng. 2003;5: 264Œ276. doi:10.1016/j.ymben.2003.09.002 99. García Martín H, Kumar VS, Weaver D, Ghosh A, Chubukov V, Mukhopadhyay A, et al. A Method to Constrain Genome-Scale Models with 13C Labeling Data. Maranas CD, editor. PLOS Comput Biol. 2015;11: e1004363. doi:10.1371/journal.pcbi.1004363 100. Ghosh A, Nilmeier J, Weaver D, Adams PD, Keasling JD, Mukhopadhyay A, et al. A peptide-based method for 13C Metabolic Flux Analysis in microbial communities. PLoS Comput Biol. 2014;10: e1003827. doi:10.1371/journal.pcbi.1003827 101. Heux S, Poinot J, Massou S, Sokol S, Portais J-C. A novel platform for automated high-throughput fluxome profiling of metabolic variants. Metab Eng. 2014;25: 8Œ19. doi:10.1016/j.ymben.2014.06.001 102. McAtee AG, Jazmin LJ, Young JD. Application of isotope labeling experiments and 13C flux analysis to enable rational pathway engineering. Curr Opin Biotechnol. Elsevier Ltd; 2015;36: 50Œ56. doi:10.1016/j.copbio.2015.08.004 103. Adebiyi AO, Jazmin LJ, Young JD. 13C flux analysis of cyanobacterial metabolism. Photosynth Res. Springer Netherlands; 2015;126: 19Œ32. doi:10.1007/s11120-014-0045-1 104. Götz A, Eylert E, Eisenreich W, Goebel W. Carbon metabolism of enterobacterial human pathogens growing in epithelial colorectal adenocarcinoma (Caco-2) cells. PLoS One. 2010;5: 8Œ9. doi:10.1371/journal.pone 105. Beste DJ V, Nöh K, Niedenführ S, Mendum T a., Hawkins ND, Ward JL, et al. 13C-flux spectral analysis of host-pathogen metabolism reveals a mixed diet for intracellular mycobacterium tuberculosis. Chem Biol. 2013;20: 1012Œ1021. doi:10.1016/j.chembiol.2013.06.012 106. Bücker R, Heroven AK, Becker J, Dersch P, Wittmann C. The Pyruvate-Tricarboxylic Acid Cycle Node. J Biol Chem. 2014;289: 30114Œ30132. doi:10.1074/jbc.M114.581348 107. Berger A, Dohnt K, Tielen P, Jahn D, Becker J, Wittmann C. Robustness and plasticity of metabolic pathway flux among uropathogenic isolates of Pseudomonas aeruginosa. PLoS One. 2014;9: e88368. doi:10.1371/journal.pone.0088368 108. Kim HU, Kim SY, Jeong H, Kim TY, Kim JJ, Choy HE, et al. Integrative genome-scale metabolic analysis of Vibrio vulnificus for drug targeting and discovery. Mol Syst Biol. Nature Publishing Group; 2014;7: 460Œ460. doi:10.1038/msb.2010.115 109. Song C, Chiasson M a, Nursimulu N, Hung SS, Wasmuth J, Grigg ME, et al. Metabolic reconstruction identifies strain-specific regulation of virulence in Toxoplasma gondii. Mol Syst Biol. Nature Publishing Group; 2013;9: 708. doi:10.1038/msb.2013.62 158 110. Lee D-Y, Chung BKS, Yusufi FNK, Selvarasu S. In silico genome-scale modeling and analysis for identifying anti-tubercular drug targets. Drug Dev Res. 2011;72: 121Œ129. doi:10.1002/ddr.20408 111. Wang C, Deng Z-L, Xie Z-M, Chu X-Y, Chang J-W, Kong D-X, et al. Construction of a genome-scale metabolic network of the plant pathogen Pectobacterium carotovorum provides new strategies for bactericide discovery. FEBS Lett. Federation of European Biochemical Societies; 2014;589: 285Œ94. doi:10.1016/j.febslet.2014.12.010 112. Lee D-S, Burd H, Liu J, Almaas E, Wiest O, Barabasi A-L, et al. Comparative Genome-Scale Metabolic Reconstruction and Flux Balance Analysis of Multiple Staphylococcus aureus Genomes Identify Novel Antimicrobial Drug Targets. J Bacteriol. 2009;191: 4015Œ4024. doi:10.1128/JB.01743-08 113. Veith N, Solheim M, van Grinsven KWA, Olivier BG, Levering J, Grosseholz R, et al. Using a Genome-Scale Metabolic Model of Enterococcus faecalis V583 To Assess Amino Acid Uptake and Its Impact on Central Metabolism. Appl Environ Microbiol. 2015;81: 1622Œ33. doi:10.1128/AEM.03279-14 114. Kim HU, Kim TY, Lee SY. Genome-scale metabolic network analysis and drug targeting of multi-drug resistant pathogen Acinetobacter baumannii AYE. Mol Biosyst. 2010;6: 339Œ348. doi:10.1039/b916446d 115. Larocque M, Chénard T, Najmanovich R. A curated C. difficile strain 630 metabolic network: prediction of essential targets and inhibitors. BMC Syst Biol. 2014;8: 117. doi:10.1186/s12918-014-0117-z 116. Oberhardt MA, Puchaka J, Fryer KE, Martins dos Santos V a P, Papin JA. Genome-scale metabolic network analysis of the opportunistic pathogen Pseudomonas aeruginosa PAO1. J Bacteriol. 2008;190: 2790Œ2803. doi:10.1128/JB.01583-07 117. Oberhardt M a., Goldberg JB, Hogardt M, Papin J a. Metabolic network analysis of Pseudomonas aeruginosa during chronic cystic fibrosis lung infection. J Bacteriol. 2010;192: 5534Œ5548. doi:10.1128/JB.00900-10 118. Sigurdsson G, Fleming RMT, Heinken A, Thiele I. A systems biology approach to drug targets in pseudomonas aeruginosa biofilm. PLoS One. 2012;7: 1Œ9. doi:10.1371/journal.pone.0034337 119. Puchalka J, Oberhardt MA, Godinho M, Bielecka A, Regenhardt D, Timmis KN, et al. Genome-Scale Reconstruction and Analysis of the Pseudomonas putida KT2440 Metabolic Network Facilitates Applications in Biotechnology. Plos Comput Biol. 2008;4: e1000210. doi:10.1371/journal.pcbi.1000210 120. Schuetz R, Kuepfer L, Sauer U. Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. Mol Syst Biol. 2007;3. doi:10.1038/msb4100162 159 121. Montezano D, Meek L, Gupta R, Bermudez LE, Bermudez JCM. Flux Balance Analysis with Objective Function Defined by Proteomics Data-Metabolism of Mycobacterium tuberculosis Exposed to Mefloquine. PLoS One. 2015;10: e0134014. doi:10.1371/journal.pone.0134014 122. Toya Y, Ishii N, Nakahigashi K, Hirasawa T, Soga T, Tomita M, et al. 13C-Metabolic flux analysis for batch culture of Escherichia coli and its pyk and pgi gene knockout mutants based on mass isotopomer distribution of intracellular metabolites. Biotechnol Prog. 2010;26: 975Œ992. doi:10.1002/btpr.420 123. Wu C, Xiong W, Dai J, Wu Q. Genome-Based Metabolic Mapping and 13 C Flux Analysis Reveal Systematic Properties of an Oleaginous Microalga Chlorella protothecoides. Plant Physiol. 2015;167: 586Œ599. doi:10.1104/pp.114.250688 124. Williams TCR, Poolman MG, Howden AJM, Schwarzlander M, Fell DA, Ratcliffe RG, et al. A Genome-Scale Metabolic Model Accurately Predicts Fluxes in Central Carbon Metabolism under Stress Conditions. Plant Physiol. 2010;154: 311Œ323. doi:10.1104/pp.110.158535 125. Chen X, Alonso AP, Allen DK, Reed JL, Shachar-Hill Y. Synergy between 13C-metabolic flux analysis and flux balance analysis for understanding metabolic adaption to anaerobiosis in E. coli. Metab Eng. Elsevier; 2011;13: 38Œ48. doi:10.1016/j.ymben.2010.11.004 126. Salunkhe P, Smart CHM, Morgan JAW, Panagea S, Walshaw MJ, Hart CA, et al. A Cystic Fibrosis Epidemic Strain of Pseudomonas aeruginosa Displays Enhanced Virulence and Antimicrobial Resistance. J Bacteriol. 2005;187: 4908Œ4920. doi:10.1128/JB.187.14.4908-4920.2005 127. Schwender J, Ohlrogge JB, Shachar-Hill Y. A flux model of glycolysis and the oxidative pentosephosphate pathway in developing Brassica napus embryos. J Biol Chem. 2003;278: 29442Œ29453. doi:10.1074/jbc.M303432200 128. Gu L, Jones AD, Last RL. Rapid LCŒMS/MS Profiling of Protein Amino Acids and Metabolically Related Compounds for Large-Scale Assessment of Metabolic Phenotypes. In: Alterman MA, Hunziker P, editors. Amino Acids. Totowa, NJ: Humana Press; 2012. pp. 1Œ11. doi:10.1007/978-1-61779-445-2_1 129. Allen DK, Shachar-Hill Y, Ohlrogge JB. Compartment-specific labeling information in 13C metabolic flux analysis of plants. Phytochemistry. 2007;68: 2197Œ2210. doi:10.1016/j.phytochem.2007.04.010 130. Pearson K. LIII. On lines and planes of closest fit to systems of points in space. Philos Mag Ser 6. 1901;2: 559Œ572. doi:10.1080/14786440109462720 131. Dash M, Liu H, Scheuermann P, Tan KL. Fast hierarchical clustering and its validation. Data Knowl Eng. 2003;44: 109Œ138. doi:10.1016/S0169-023X(02)00138-6 160 132. Opperman MJ, Shachar-Hill Y. Metabolic Flux Analyses of Pseudomonas aeruginosa Cystic Fibrosis Isolates. (In Rev.) 2016; 133. Neidhardt FC, Edwin Umbarger H. Chemical Composition of Escherichia coli. Chem Compos Escherichia coli Neidhardt, FC (Ed), Escherichia coli Salmonella Cell Mol Biol. 1996;Escherichi. 134. Wiechert W. 13C metabolic flux analysis. Metab Eng. 2001;3: 195Œ206. doi:10.1006/mben.2001.0187 135. Schmidt K, Nørregaard LC, Pedersen B, Meissner A, Duus JØ, Nielsen JØ, et al. Quantification of Intracellular Metabolic Fluxes from Fractional Enrichment and13CŒ13C Coupling Constraints on the Isotopomer Distribution in Labeled Biomass Components. Metab Eng. 1999;1: 166Œ179. doi:10.1006/mben.1999.0114 136. Zumft W. Cell biology and molecular basis of denitrification. Microbiol Mol Biol Rev. 1997;61: 533Œ616. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=232623&tool=pmcentrez&rendertype=abstract 137. Hassett DJ, Sutton MD, Schurr MJ, Herr AB, Caldwell CC, Matu JO. Pseudomonas aeruginosa hypoxic or anaerobic biofilm infections within cystic fibrosis airways. Trends Microbiol. 2009;17: 130Œ138. doi:10.1016/j.tim.2008.12.003 138. Cramer N, Klockgether J, Wrasman K, Schmidt M, Davenport CF, T??mmler B. Microevolution of the major common Pseudomonas aeruginosa clones C and PA14 in cystic fibrosis lungs. Environ Microbiol. 2011;13: 1690Œ1704. doi:10.1111/j.1462-2920.2011.02483.x 139. Palmer GC, Palmer KL, Jorth P a., Whiteley M. Characterization of the Pseudomonas aeruginosa transcriptional response to phenylalanine and tyrosine. J Bacteriol. 2010;192: 2722Œ2728. doi:10.1128/JB.00112-10 140. Jain C, Yun M, Politz SM, Rao RP. A pathogenesis assay using Saccharomyces cerevisiae and Caenorhabditis elegans reveals novel roles for yeast AP-1, Yap1, and host dual oxidase BLI-3 in fungal pathogenesis. Eukaryot Cell. 2009;8: 1218Œ27. doi:10.1128/EC.00367-08 141. Kavanagh K, Reeves EP. Exploiting the potential of insects for in vivo pathogenicity testing of microbial pathogens. FEMS Microbiol Rev. 2004;28: 101Œ112. doi:10.1016/j.femsre.2003.09.002 142. Lamb C, Dixon RA. The Oxidative Burst in Plant Disease Resistance. Annu MolBiol. 1997;48: 251Œ275. doi:10.1146/annurev.arplant.48.1.251 143. Lambeth JD. NOX enzymes and the biology of reactive oxygen. Nat Rev Immunol. 2004;4: 181Œ189. doi:10.1038/nri1312 144. Junker BH. Flux analysis in plant metabolic networks: Increasing throughput and coverage. Curr Opin Biotechnol. 2014;26: 183Œ188. doi:10.1016/j.copbio.2014.01.016 161 145. Lindsey TL, Hagins JM, Sokol PA, Silo-Suh LA. Virulence determinants from a cystic fibrosis isolate of Pseudomonas aeruginosa include isocitrate lyase. Microbiology. 2008;154: 1616Œ1627. doi:10.1099/mic.0.2007/014506-0 146. Dunn MF, Ramírez-Trujillo JA, Hernández-Lucas I. Major roles of isocitrate lyase and malate synthase in bacterial and fungal pathogenesis. Microbiology. 2009;155: 3166Œ3175. doi:10.1099/mic.0.030858-0 147. Palmer KL, Mashburn LM, Singh PK, Whiteley M. Cystic Fibrosis Sputum Supports Growth and Cues Key Aspects of Pseudomonas aeruginosa Physiology Cystic Fibrosis Sputum Supports Growth and Cues Key Aspects of Pseudomonas aeruginosa Physiology. J Bacteriol. 2005;187: 5267Œ5277. doi:10.1128/JB.187.15.5267 148. Deng X, Liang H, Ulanovskaya OA, Ji Q, Zhou T, Sun F, et al. Steady-state hydrogen peroxide induces glycolysis in staphylococcus aureus and pseudomonas aeruginosa. J Bacteriol. 2014;196: 2499Œ2513. doi:10.1128/JB.01538-14 149. Wortham BW, Patel CN, Oliveira M a. Polyamines in bacteria: pleiotropic effects yet specific mechanisms. Adv Exp Med Biol. 2007;603: 106Œ15. doi:10.1007/978-0-387-72124-8_9 150. Shah P, Swiatlo E. A multifaceted role for polyamines in bacterial pathogens. Mol Microbiol. 2008;68: 4Œ16. doi:10.1111/j.1365-2958.2008.06126.x 151. Karatan E, Michael AJ. A wider role for polyamines in biofilm formation. Biotechnol Lett. 2013;35: 1715Œ1717. doi:10.1007/s10529-013-1286-3 152. Burrell M, Hanfrey CC, Murray EJ, Stanley-Wall NR, Michael AJ. Evolution and multiplicity of arginine decarboxylases in polyamine biosynthesis and essential role in Bacillus subtilis biofilm formation. J Biol Chem. 2010;285: 39224Œ39238. doi:10.1074/jbc.M110.163154 153. Patel CN, Wortham BW, Lines JL, Fetherston JD, Perry RD, Oliveira MA. Polyamines are essential for the formation of plague biofilm. J Bacteriol. 2006;188: 2355Œ2363. doi:10.1128/JB.188.7.2355-2363.2006 154. Zhang X, Zhang Y, Liu J, Liu H. PotD protein stimulates biofilm formation by Escherichia coli. Biotechnol Lett. 2013;35: 1099Œ1106. doi:10.1007/s10529-013-1184-8 155. Lee J, Sperandio V, Frantz DE, Longgood J, Camilli A, Phillips MA, et al. An alternative polyamine biosynthetic pathway is widespread in bacteria and essential for biofilm formation in Vibrio cholerae. J Biol Chem. 2009;284: 9899Œ9907. doi:10.1074/jbc.M900110200 156. McGinnis MW, Parker ZM, Walter NE, Rutkovsky AC, Cartaya-Marin C, Karatan E. Spermidine regulates Vibrio cholerae biofilm formation via transport and signaling pathways. FEMS Microbiol Lett. 2009;299: 166Œ174. doi:10.1111/j.1574-6968.2009.01744.x 162 157. Goytia M, Dhulipala VL, Shafer WM. Spermine impairs biofilm formation by Neisseria gonorrhoeae. FEMS Microbiol Lett. 2013;343: 64Œ69. doi:10.1111/1574-6968.12130 158. Johnson L, Mulcahy H, Kanevets U, Shi Y, Lewenza S. Surface-localized spermidine protects the Pseudomonas aeruginosa: Outer membrane from antibiotic treatment and oxidative stress. J Bacteriol. 2012;194: 813Œ826. doi:10.1128/JB.05230-11 159. Landete JM, Arena ME, Pardo I, Manca de Nadra MC, Ferrer S. Comparative survey of putrescine production from agmatine deamination in different bacteria. Food Microbiol. 2008;25: 882Œ887. doi:10.1016/j.fm.2008.06.001 160. Twomey KB, Alston M, An SQ, O™Connell OJ, McCarthy Y, Swarbreck D, et al. Microbiota and metabolite profiling reveal specific alterations in bacterial community structure and environment in the cystic fibrosis airway during exacerbation. PLoS One. 2013;8. doi:10.1371/journal.pone.0082432 161. Han TC, Kwon DH, Hegazy M, Lu CD. Transcriptome analysis of agmatine and putrescine catabolism in Pseudomonas aeruginosa PAO1. J Bacteriol. 2008;190: 1966Œ1975. doi:10.1128/JB.01804-07 162. Williams BJ, Du RH, Calcutt MW, Abdolrasulnia R, Christman BW, Blackwell TS. Discovery of an operon that participates in agmatine metabolism and regulates biofilm formation in Pseudomonas aeruginosa. Mol Microbiol. 2010;76: 104Œ119. doi:10.1111/j.1365-2958.2010.07083.x 163. Manuel J, Zhanel GG, De Kievit T. Cadaverine suppresses persistence to carboxypenicillins in Pseudomonas aeruginosa PAO1. Antimicrob Agents Chemother. 2010;54: 5173Œ5179. doi:10.1128/AAC.01751-09 164. Kwon DH, Lu C. Polyamines Increase Antibiotic Susceptibility in Pseudomonas aeruginosa Polyamines Increase Antibiotic Susceptibility in Pseudomonas aeruginosa. 2006;50: 1623Œ1627. doi:10.1128/AAC.50.5.1623 165. El-Halfawy OM, Valvano MA. Chemical Communication of Antibiotic Resistance by a Highly Resistant Subpopulation of Bacterial Cells. PLoS One. 2013;8. doi:10.1371/journal.pone.0068874 166. Tkachenko AG, Kashevarova NM, Karavaeva EA, Shumkov MS. Putrescine controls the formation of Escherichia coli persister cells tolerant to aminoglycoside netilmicin. FEMS Microbiol Lett. 2014;361: 25Œ33. doi:10.1111/1574-6968.12613 163 CHAPTER 4 Œ Conclusions and Future Directions 164 The work described in this thesis was concerned with the metabolic changes during adaptation to the cystic fibrosis lung environment and pathogenesis by the bacterium Pseudomonas aeruginosa in chronic infection. The research used Metabolic Flux Analysis and Flux Balance Analysis to quantify central carbon metabolism in CF isolates of P. aeruginosa to add to the understanding of the bacterium's metabolism and the changes induced by exposure to and growth in the cystic fibrosis lung environment. Due to the significance of P. aeruginosa as a cause of morbidity and mortality in cystic fibrosis (CF) and its intrinsic and acquired resistance against antibiotic therapy, the additional goal of this study was to propose possible metabolic targets for potential future pharmaceutical development. Comparison of Metabolism A central question of this thesis is whether metabolic differences can be determined in the role of adaptation towards chronic infection of lung, with the comparison of P. aeruginosa CF lung isolates to a wound isolate as well as to other published studies. The results show the extensive metabolic range of P. aeruginosa, both in the range of flux patterns though carbon metabolism under different conditions as well as in the carbon metabolic phenotypes present in genetically similar and diverse strains. Both of the studies in this dissertation also help to contextualize the metabolism of P. aeruginosa in comparison to other 13C MFA organisms under aerobic conditions with glucose as the carbon source and under conditions mimicking the CF lung. The minimal media 13C MFA study describes a standardized growth condition that can be used to identify and quantify differences in the growth physiology and metabolism between bacterial species. The results in this study also contribute to a comparison of metabolism based on metabolism of glucose through the Embden-Meyerhof-Parnas Pathway (EMP) versus the 165 Entner-Doudoroff Pathway (EDP). While theories of reduced protein production cost and energy yield as explanations of the loss of EMP with conservation and utilization of EDP in some organisms have been demonstrated mathematically, direct demonstration in organisms have not been shown [1]. The measured utilization rates of the EDP in several of the P. aeruginosa strains (especially AA44) and E. coli utilization of EMP during aerobic growth in minimal media demonstrate similar rates of flux ratios through these pathways with increased growth rates in P. aeruginosa [2]. The study of aerobic metabolism during growth on glucose also allowed the direct comparison of a related species P. fluorescens to P. aeruginosa. The findings demonstrated not only the shared metabolic phenotypes in these species, but also provide evidence that one of the minimal media growth phenotypes in this study may be demonstrating a proto-biofilm synthesis metabolism [3]. The S-SCFM 13C MFA study allowed a detailed elucidation of the carbon fluxes through central metabolism in P. aeruginosa under more CF lung-like conditions. The metabolic findings of the study can be compared and contrasted to the proteomic work by Hoboth et al. (2009). That group's findings of increased levels of Tricarboxylic acid (TCA) cycle enzymes and of proteins responsible for the interconversion of ornithine, arginine, and glutamate and TCA cycle intermediates in late CF isolates in comparison to early CF isolates are consistent with the large carbon fluxes in the S-SCFM 13C MFA flux maps obtained here [4]. But, 13C MFA quantification of carbon fluxes here demonstrated that the utilization of amino acids by these strains under both aerobic and anoxic conditions drives carbon flux from glutamate to ornithine to succinate, which is opposite to the conclusions of the prior Omic study [4]. The additional analysis of the anoxic growth condition to better understand differences in aerobic and anoxic growth in the presence of substrates present in the CF lung also provided insight into the 166 utilization of amino acids and the role of the glyoxylate cycle enzyme isocitrate lyase under anoxic growth. The analyses in lung-like media under both aerobic and anoxic conditions highlight the value of studying metabolic phenotypes under experimental conditions that mimic relevant environmental circumstances. Indeed there is evidence that the CF lung includes a range of microenvironments from aerobic to anoxic [5Œ7]. Metabolic Fingerprinting and Metabolic Flux Analysis Metabolic Fingerprinting using stable isotopic labeling is a method that has been proposed to increase the speed and throughput of metabolic phenotyping by identifying metabolic fluxes without full flux analysis [8,9]. Metabolic flux fingerprinting involves using LC-MS, GC-MS, and/or NMR to produce a set of labeling data on the system of interest and applying statistical methods (Analysis of Variance, ANOVA; Principal Component Analysis, PCA; or Independent Component Analysis, ICA) to identify differences and/or degrees of difference between treatments or genotypes. Features revealed may either describe the system's metabolic features or identify specific measures that could be used for identification of a specific organism [10]. A proposed refinement of this method is the use of stable isotopic labeling to produce labeled metabolites that capture the underlying metabolic network that produced the measured metabolites [11]. This process is meant to identify metabolic similarities and differences more efficiently (days of analysis) than using 13C MFA (months of modeling). We investigated the power of this approach to discriminate between metabolic flux phenotypes exhibited by different strains and whether degrees of similarity among strains identified by fingerprinting agreed with the results of full flux analysis. The results of these studies cast doubt on the ability of 13C metabolic fingerprinting to differentiate reliably between strains of the same species. Thus we found that HCA using fluxes obtained by 13C MFA was able 167 to clearly distinguish among P. aeruginosa strains grown anoxically in S-SCFM; whereas the 13C labeling data used in the MFA was insufficient to distinguish the strains using PCA. Under aerobic growth (either in minimal or S-SCFM media) 13C metabolic fingerprinting was able to identify that each strain was different from the others, but the overall degree of difference between one strain and another, did not accord with the degrees of difference/relatedness shown by HCA of 13C MFA flux analyses. Because of the time-consuming nature of rigorous 13C MFA and the value of increasing throughput, further investigation should be conducted to determine which if any statistical tools can be used to determine metabolic differences among strains of the same species. It may be that specific labeling information (e.g. increased measurements of labeling in particular metabolites), and/or different statistical methods (e.g. Independent Component Analysis) may increase the predictability of these analyses. It seems likely that progress in this direction will be easier on a case-by-case basis for specific metabolic processes Œ that is using a particular labeled substrate to discriminate particular fluxes, and that developing profiling methods for system-wide untargeted differences will be harder, especially where multiple substrates are utilized simultaneously. Virulence and Metabolism A central question motivating the work of this thesis was fiCan metabolism and its changes be meaningfully related to an organism™s adaptation towards greater virulence?fl By quantifying metabolic changes in strains that have adapted to a pathogenic environment, these changes could be general to organism adaptation in certain body environments. Such differences could provide a basis for the production of tests to determine the course of chronic infections, and investigated as potential targets to reduce or eliminate infections. Presented below is a 168 summary of the metabolic findings of this thesis, which could be potential targets for pharmaceutical intervention. The glyoxylate cycle has been shown to be an important flux in both minimal media and CF lung-like growth of P. aeruginosa. With the use of FBA in these studies, it has been shown that optimized biomass production (growth) in P. aeruginosa does not explain these fluxes. The observation of a correlation between growth rate and glyoxylate cycle flux observed here, is consistent with the results of a previous study that demonstrated reduced virulence in both alfalfa seeds and mouse lung by P. aeruginosa in a strain with defective isocitrate lyase [12]. The finding of a functional glyoxylate cycle as being an important metabolic pathway for persistence in the body in pathogenic organisms (especially Mycobacterium tuberculosis) has been noted in the scientific literature and in the context of gluconeogenesis during lipid metabolism [13]. This utilization of the glyoxylate cycle by P. aeruginosa during non-lipid metabolism and especially anoxic amino acid metabolism is intriguing, as the generally accepted role of this cycle is for anabolism from two carbon acetyl-CoA units derived from lipid catabolism. The utilization of this cycle by P. aeruginosa strains under the conditions used here could potentially be an inherent signaling response that may be expressed at higher levels under anoxic conditions, but more direct measures (such as transcriptomics and metabolic analyses of mutant and transgenic strains) would need to be conducted to further elucidate the interplay between environment and glyoxylate cycle expression. The results of this study also demonstrate that isocitrate lyase and 2-hydroxy-3-oxopropionate reductase, operating without a complete glyoxylate cycle, is heavily utilized during anoxic S-SCFM growth in P. aeruginosa and could possibly be pharmaceutical targets to reduce or eliminate this organism from the CF lung environment. 169 The catabolism of the amino acids glutamate, ornithine, and arginine via putrescine into the TCA cycle and not via -ketoglutarate is a novel finding. As shown by FBA growth optimization, putrescine production is less efficient for growth, likely due to the increased ATP consumption by this pathway. A possible explanation of this pathway, though kinetic data for these enzymes would be beneficial in determining the possibility of accumulation, in is the utilization of polyamines, such as putrescine, in P. aeruginosa as a general cellular and plasma membrane anti-oxidant [14]. The expression of polyamines, especially putrescine, in the cell membrane surface of P. aeruginosa has been demonstrated in the context of antimicrobial stress (e.g. extracellular DNA) [15]. A polyamine metabolite derived from arginine decarboxylation, agmatine, has also been demonstrated as a signaling molecule in biofilm formation that develops in some P. aeruginosa isolates, providing additional evidence of an adaptation of metabolism towards signaling of immune stress in this organism [16]. The discovery in this study of substantial fluxes via putrescine also provide a possible explanation for the observed correlation between CF exacerbations and increases in putrescine concentration from sputum samples [17]. These studies also provide evidence of metabolism of P. aeruginosa towards production of reductant at the expense of biomass synthesis. 13C MFA of strains growing in Minimal media demonstrated a metabolic phenotype that differed from the one predicted by FBA for optimized biomass production. A key difference is towards the overproduction of reductant that is seen in the increased OPPP flux through relative to fluxes through the ED pathway and TCA cycle. This overproduction of reductant in comparison to optimized biomass production was further accentuated in S-SCFM growth as seen in the calculation of reductant produced in 13C MFA constrained fluxes of FBA optimized growth compared to FBA optimized growth. A possible 170 explanation for this finding is that P. aeruginosa uses high rates of NADPH production to survive in the oxidative environment of the CF lung or in other hostile environments. Future Directions To better understand the metabolic adaptation of P. aeruginosa to the CF lung environment, a more life-like replication of the CF lung environment would be useful. 13C-label experiments would ideally be conducted in an animal model infected with P. aeruginosa, providing the most accurate in vivo study of metabolism and infection. Unfortunately, such an experiment would be both expensive (in 13C substrates), and difficult to interpret, since measuring uptake rates and accounting for metabolic transformations outside the P. aeruginosa cells would be invasive and extremely challenging. Additionally, while the genetic cause of CF has been well established, the underlying pathology has not been well replicated in organ systems of any model animal used [18]. Under these circumstances, it is more reasonable to propose the use of biofilm cultivation techniques under lung-like conditions as the next step in experimental techniques [19]. Using such a system to produce biofilm under more lung-like conditions would complicate the ability to perform steady state MFA due to the timescales necessary for complete isotopic labeling of analytic substrate as well as the heterogeneous environment of the biofilm [20]. While isotopically non-stationary MFA techniques have been established that would allow for such a biofilm study to be feasible [21], technical obstacles remain. These include designing and verifying a suitable biofilm growth chamber; methodological development of consistent growth of P. aeruginosa biofilm cultures; identification, harvesting and isolation of biofilm layer(s) of phenotypic difference; and measurement of uptake and secretion rates of metabolites. However, such an investigation should be feasible and represents a logical next step in this research. 171 Besides discussion of the natural progression of experiments conducted in this thesis, other experiments could be conducted to verify and explore the conclusions reported. Growth of the strains in these experiments to measure and quantify biofilm differences in the media used in these experiments could be completed. These experiments could be used to confirm whether pro-biofilm growth phenotypes are expressed under planktonic growth. Measurement of biofilm production rates and type of biofilm for each strain could then be compared to the identified biofilm and non-biofilm metabolic phenotypes demonstrated under M9 growth to identify if MFA determined phenotypes under planktonic growth do correlate with alginate biofilm producers. Biofilm production difference by strains grown in S-SCFM may provide additional insight into the metabolic differences seen as well. Additionally, all the metabolic results found could be further augmented with measurement of known virulence factors, quorum signaling molecules, and transcriptomic measurements. These measurements may provide finer quantification of pathogenicity or virulence under these growth conditions that may provide a contrastable difference of virulence that may explain the underlying metabolic phenotypes reveled in these studies. Additional 13C MFA experiments could also be conducted using a lipid source instead of amino acid sources of carbon. These result could help to shed light on the differences in lipid and amino acid metabolism of these CF isolates and may help in determining possible metabolic utilization differences that may occur is some strains. Measurement of secondary metabolites (known virulence factors, quorum signaling molecules, etc.) could also be used to further constrain metabolic measurements in FBA constrained MFA that may lead to additional discoveries about P. aeruginosa metabolism. 172 Additional strains of P. aeruginosa could be analyzed using 13C MFA to further validate and identify additional metabolic phenotypes. As demonstrated in this work, the reference strain PA01 did not show significant difference from the CF isolates used in this study. The use of environmental strains, as opposed to pathogenic strains like PA01, as the metabolic reference may help in the understanding of pathogenic adaptation seen in this organism. Other CF isolates could also be included to further encompass identified phenotypes beyond the genetically sequential isolates chosen for this study. Finally, identification and utilization of inhibitors of key enzymatic pathways, or of P. aeruginosa mutants defective in these pathways, could be used to validate the results and proposed mechanisms of this thesis. Inhibition of the enzyme of the Entner-Doudoroff pathway would validate the ability of P. aeruginosa to continue to grow under minimal and complex media. Use of an inhibitor or known mutants in polyamine metabolism would be of additional interest, and would help in the illumination of polyamine metabolism and biofilm formation [22]. These experiments would also need to be paired with biofilm growth experiments to fully validate the relationship between polyamines and biofilm formation. 173 REFERENCES 174 REFERENCES 1. Flamholz A, Noor E, Bar-Even A, Liebermeister W, Milo R. Glycolytic strategy as a tradeoff between energy yield and protein cost. Proc Natl Acad Sci U S A. 2013;110: 10039Œ44. doi:10.1073/pnas.1215283110 2. Chen X, Alonso AP, Allen DK, Reed JL, Shachar-Hill Y. Synergy between 13C-metabolic flux analysis and flux balance analysis for understanding metabolic adaption to anaerobiosis in E. coli. Metab Eng. Elsevier; 2011;13: 38Œ48. doi:10.1016/j.ymben.2010.11.004 3. Lien SK, Niedenführ S, Sletta H, Nöh K, Bruheim P. Fluxome study of Pseudomonas fluorescens reveals major reorganisation of carbon flux through central metabolic pathways in response to inactivation of the anti-sigma factor MucA. BMC Syst Biol. 2015;9: 6. doi:10.1186/s12918-015-0148-0 4. Hoboth C, Hoffmann R, Eichner A, Henke C, Schmoldt S, Imhof A, et al. Dynamics of adaptive microevolution of hypermutable Pseudomonas aeruginosa during chronic pulmonary infection in patients with cystic fibrosis. J Infect Dis. 2009;200: 118Œ130. doi:10.1086/599360 5. Palmer KL, Aye LM, Whiteley M. Nutritional cues control Pseudomonas aeruginosa multicellular behavior in cystic fibrosis sputum. J Bacteriol. 2007;189: 8079Œ8087. doi:10.1128/JB.01138-07 6. Schobert M, Jahn D. Anaerobic physiology of Pseudomonas aeruginosa in the cystic fibrosis lung. Int J Med Microbiol. Elsevier GmbH.; 2010;300: 549Œ556. doi:10.1016/j.ijmm.2010.08.007 7. Hassett DJ, Sutton MD, Schurr MJ, Herr AB, Caldwell CC, Matu JO. Pseudomonas aeruginosa hypoxic or anaerobic biofilm infections within cystic fibrosis airways. Trends Microbiol. 2009;17: 130Œ138. doi:10.1016/j.tim.2008.12.003 8. Zamboni N, Sauer U. Model-independent fluxome profiling from 2H and 13C experiments for metabolic variant discrimination. Genome Biol. 2004;5: R99. doi:10.1186/gb-2004-5-12-r99 9. Hollinshead WD, Henson WR, Abernathy M, Moon TS, Tang YJ. Rapid metabolic analysis of Rhodococcus opacus PD630 via parallel 13C-metabolite fingerprinting. Biotechnol Bioeng. 2016;113: 91Œ100. doi:10.1002/bit.25702 10. Gao P, Xu G. Mass-spectrometry-based microbial metabolomics: Recent developments and applications. Anal Bioanal Chem. 2015;407: 669Œ680. doi:10.1007/s00216-014-8127-7 175 11. Tang JK-H, You L, Blankenship RE, Tang YJ. Recent advances in mapping environmental microbial metabolisms through 13C isotopic fingerprints. J R Soc Interface. 2012;9: 2767Œ80. doi:10.1098/rsif.2012.0396 12. Lindsey TL, Hagins JM, Sokol P a., Silo-Suh L a. Virulence determinants from a cystic fibrosis isolate of Pseudomonas aeruginosa include isocitrate lyase. Microbiology. 2008;154: 1616Œ1627. doi:10.1099/mic.0.2007/014506-0 13. Lorenz MC, Fink GR. Life and death in a macrophage: Role of the glyoxylate cycle in virulence. Eukaryot Cell. 2002;1: 657Œ662. doi:10.1128/EC.1.5.657-662.2002 14. Shah P, Swiatlo E. A multifaceted role for polyamines in bacterial pathogens. Mol Microbiol. 2008;68: 4Œ16. doi:10.1111/j.1365-2958.2008.06126.x 15. Johnson L, Mulcahy H, Kanevets U, Shi Y, Lewenza S. Surface-localized spermidine protects the Pseudomonas aeruginosa: Outer membrane from antibiotic treatment and oxidative stress. J Bacteriol. 2012;194: 813Œ826. doi:10.1128/JB.05230-11 16. Williams BJ, Du RH, Calcutt MW, Abdolrasulnia R, Christman BW, Blackwell TS. Discovery of an operon that participates in agmatine metabolism and regulates biofilm formation in Pseudomonas aeruginosa. Mol Microbiol. 2010;76: 104Œ119. doi:10.1111/j.1365-2958.2010.07083.x 17. Twomey KB, Alston M, An SQ, O™Connell OJ, McCarthy Y, Swarbreck D, et al. Microbiota and metabolite profiling reveal specific alterations in bacterial community structure and environment in the cystic fibrosis airway during exacerbation. PLoS One. 2013;8. doi:10.1371/journal.pone.0082432 18. Wang Y, Wrennall JA, Cai Z, Li H, Sheppard DN. Understanding how cystic fibrosis mutations disrupt CFTR function: From single molecules to animal models. Int J Biochem Cell Biol. Elsevier Ltd; 2014;52: 47Œ57. doi:10.1016/j.biocel.2014.04.001 19. Merritt JH, Kadouri DE, O™Toole G a. Growing and analyzing static biofilms. Curr Protoc Microbiol. 2011; 1Œ18. doi:10.1002/9780471729259.mc01b01s22 20. Stewart PS, Franklin MJ. Physiological heterogeneity in biofilms. Nat Rev Microbiol. 2008;6: 199Œ210. doi:10.1038/nrmicro1838 21. Wiechert W, Nöh K. Isotopically non-stationary metabolic flux analysis: Complex yet highly informative. Curr Opin Biotechnol. 2013;24: 979Œ986. doi:10.1016/j.copbio.2013.03.024 22. Han TC, Kwon DH, Hegazy M, Lu CD. Transcriptome analysis of agmatine and putrescine catabolism in Pseudomonas aeruginosa PA01. J Bacteriol. 2008;190: 1966Œ1975. doi:10.1128/JB.01804-07