. v..u irrefi. . : ... 'h z .ch .. .1 t . .{4:53.- .3. ”up: i. . *5, E: .03.... 51 . an (a. . .C. r I . Nixuih: up. kw P a! P 2.! Z 1... “Mn...” 8.. . . . _ .lfli‘vlh‘lnll I . i! 94 . «1.9%! ;: a :‘z This is to certify that the dissertation entitled SYNTHESIS OF AROMATICS AND HYDROAROMATICS FROM D-GLUCOSE VIA A NATIVE AND A VARIANT OF THE SHIKIMATE PATHWAY presented by I l o H a ”865? NINGQING RAN E a 5 a .99.: .4 as has been accepted towards fulfillment 0' . 2 of the requrrements for the I <0 l Ph.D. degree in CHEMISTRY the 09944171 / Major Professor’s Signature 8/2 01/04] Date MSU is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/07 p:/ClRC/DaleDue.indd-p.1 SYNTHESIS OF AROMATICS AND HYDROAROMATICS FROM D-GLUCOSE VIA A NATIVE AND A VARIANT OF THE SHIKIMATE PATHWAY By Ningqing Ran A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 2004 ABSTRACT SYNTHESIS OF AROMATICS AND HYDROAROMATICS FROM D-GLUCOSE VIA A NATIVE AND A VARIANT OF THE SHIKIMATE PATHWAY By Ningqing Ran Microbial synthesis of aromatic chemicals from non-toxic D-glucose constitutes an intriguing alternative to current use of toxic benzene as a starting material. In addition, D-glucose is derived from renewable feedstocks such as starch and cellulose while benzene is derived from nonrenewable petroleum. As a case study of the challenges presented by the microbial toxicity of many aromatic chemicals, the synthesis of hydroquinone from D-glucose is examined. Construction and fennentor-controlled cultivation of an Escherichia coli catalyst is detailed for the conversion of D-glucose into the nontoxic hydroaromatic quinic acid via reduction of the shikimate pathway intermediate 3-dehydroquinic acid. Chemical oxidation of quinic acid in clarified, decolorized, ammonia-free culture supernatant then affords hydroquinone in high yield. By interfacing microbial synthesis with chemical synthesis, the toxicity of hydroquinone towards E. coli is circumvented. Another value-added aromatic chemical that can be synthesized by way of the shikimate pathway is gallic acid. However, key aspects of the biosynthesis of gallic acid have not previously been elaborated. Gallic acid has been hypothesized to be derived via oxidation of 3-dehydroshikimic acid or alternatively, by hydroxylation of protocatechuic acid. 3-Dehydroshikimic acid is an intermediate in the shikimate pathway. Dehydration of 3—dehydroshikimic acid leads to gallic acid. In this thesis, biosynthesis of gallic acid in E. coli is examined with chemically synthesized [5-‘80]-3-dehydroshikimic acid. Analysis by mass spectrometry of the m0 labeling in gallic acid synthesized by E. coli from [5-1801-3-dehydroshikimic acid revealed full retention of the l8O-labeling. These results suggest that oxidation of 3-dehydroshikimic acid is the primary biosynthetic route to gallic acid. In addition to target-oriented biosynthesis of aromatics, a new strategy is examined to increase the concentration and yield of natural products biosynthesized via the shikimate pathway in E. coli by increasing phosphoenolpyruvate availability. A pyruvate-based shikimate pathway is created by the directed evolution of 2-keto-3-deoxy- 6-phosphogalactonate (KDPGal) aldolase. The result is an enzyme that catalyzes the condensation of pyruvic acid with D-erythrose 4-phosphate to produce 3-deoxy-D- arabino-heptluosonic acid 7-phosphate (DAHP), which is the first committed intermediate in the shikimate pathway. By contrast, DAHP in the native shikimate pathway is formed by the condensation of phosphoenolpyruvate with D-erythrose 4- phosphate catalyzed by DAHP synthase. The ability of the created pyruvate-based shikimate pathway to support microbial growth and synthesize 3-dehydroshikimic acid is examined in E. coli constructs that lack phosphoenolpyruvate-based DAHP synthase activity. Copyright by Ningqing Ran 2004 To my parents, for their love and support. ACKNOWLEDGMENTS It is with great pleasure that I take the opportunity to thank all the people that I am indebted to for their help. First and foremost, I want to thank my research advisor Prof. John Frost for his guidance and encouragement throughout the the course of my graduate career. His enthusiasm, intergral View on research and excellent writing skill have made a deep impression on me. I owe him lots of gratitute for showing me the way of research. In addition, I would like to thank the members of my graduate committee, Prof. Babak Borhan, Prof. John McCracken and Prof. Merlin Bruening for their intellectual input during the preparation of this dissertation. I am grateful to Dr. Karen Draths for her inspiration, technique guidance and support to my research, particularly for the many stimulating and instructive discussions we had and for the arduous task of critic and proof reading of my dissertation. I also wish to thank all of the Frost group members, both past and present, Dr. Chad Hansen, Dr. Dave Knop, Dr. Spiros Kambourakis, Dr. Kai Li, Dr. Feng Tian, Dr. Jessica Barker, Dr. Sunil Chandran, Dr. Padmesh Venkitasubramanian, Dr. Jian Yi, Dr. Jiantao Guo, Dr. Wei Niu, Dr. Dongming Xie, Dr. Jihane Achkar, Dr. Mo Xian, Ms. Mapitso Molefe, Ms. Heather Stueben, Mr. Xiaofei J ia, Mr. Wensheng Li, Mr. Justus J ancauskas, Mr. Man-Kit Lau, Mr. Jingsong Yang and Mr. Kin Sing Lee for their assistant and friendship during the progression of this research. This thesis is dedicated to my parents Quanyin Ran and Su Ye and my brother N ingyu Ran for their love and support. vi TABLE OF CONTENTS LIST OF TABLES ....................................................................................................... viii LIST OF FIGURES ......................................................................................................... x LIST OF ABBREVIATIONS ........................................................................................ xv CHAPTER ONE - Introduction ....................................................................................... 1 Biosyntheses of value-added chemicals from D-glucose ........................................... 2 Directed evolution ................................................................................................. 25 CHAPTER TWO - Benzene-Free Synthesis of Hydroquinone ....................................... 42 Introduction ............................................................................................................... 42 Microbial synthesis of quinic acid from D-glucose. .................................................... 48 Chemical synthesis of hydroquinone from quinic acid. .............................................. 79 Discussion ................................................................................................................. 86 CHAPTER THREE - Analysis of Gallic Acid Biosynthetic Pathway in Escherichia coli with [5-‘8O]-3-Dehydroshikimic Acid ......................................................................... 102 Introduction ............................................................................................................. 102 Synthesis of [5-‘801-3-dehydroshikimic acid ............................................................ 110 Biosynthesis of gallic acid in E. coli. ....................................................................... 111 Discussion ............................................................................................................... 123 CHAPTER FOUR - Creation of a Pyruvate-Based Shikimate Pathway in Escherichia coli .................................................................................................................................... 131 Introduction ............................................................................................................. 1 3 1 Directed evolution of KDPGal aldolase ................................................................... 148 Discussion ............................................................................................................... 199 CHAPTER FIVE - Experimental ................................................................................. 209 General chemistry .................................................................................................... 209 Reagents and solvents .............................................................................................. 209 Chromatography ...................................................................................................... 210 Spectroscopic and analytical measurements ............................................................. 210 Microbial strains and plasmids ................................................................................. 211 Storage of microbial strains and plasmids ................................................................ 213 Culture medium ....................................................................................................... 213 Fed-batch fermentation (general) ............................................................................. 215 Genetic manipulations ............................................................................................. 219 Enzyme assays ........................................................................................................ 232 CHAPTER TWO ..................................................................................................... 235 CHAPTER THREE ................................................................................................. 244 CHAPTER FOUR ................................................................................................... 248 vii LIST OF TABLES Table 1. Concentrations and yields of quinic acid and 3-dehydroquinic acid synthesized by various E. coli strains under different culture conditions. .................................. 58 Table 2. Shikimate dehydrogenase specific activities for various strains cultured under different conditions. ............................................................................................... 59 Table 3. DAHP synthase specific activities for various strains cultured under different conditions. ............................................................................................................. 59 Table 4. Phosphoenolpyruvate synthase specific activity during fermentation runs. ....... 65 Table 5. Reaction conditions and yields for chlorine-free oxidation of quinic acid. ........ 86 Table 6. l8O-Enrichment determined by FAB(-) mass spectrometry. ............................ 114 Table 7. Purification of E. coli KDPGal aldolase. ........................................................ 150 Table 8. Effect of expression of wild—type E. coli KDPGal aldolase on growth characteristics of E. coli strains. ........................................................................... 157 Table 9. Directed evolution of E. coli KDPGal aldolase ............................................... 165 Table 10. Mutations and specific activities of E. coli KDPGal aldolase variants. ......... 165 Table 11. KDPGal aldolases from various microorganisms. ........................................ 167 Table 12. Directed evolution of K. pneumoniae KDPGal aldolase. .............................. 169 Table 13. Mutations and specific activities of K. pneumoniae KDPGal aldolase variants. ............................................................................................................................ 169 Table 14. Directed evolution of S. typhimurium KDPGal aldolase. .............................. 173 Table 15. Mutations and specific activities of S. typhimuriume KDPGal aldolase variants. ............................................................................................................................ 173 Table 16. Specific activities of wild-type and evolved KDPGal aldolase isozymes. ..... 174 Table 17. Synthesis of 3-dehydroshikimic acid under fermentor-controlled conditions. 180 Table 18. Evolved KDPGal aloldase activities towards DAHP formation. ................... 180 viii Table 19. Chimeric dgoA genes evolved by cross-species DNA family shuffling ......... 188 Table 20. Kinetic parameters of the wild-type KDPGal aldolases and the evolved variants from cross-species DNA family shuffling. .............................................. 189 Table 21. Microbial strains and plasmids. ................................................................... 211 ix LIST OF FIGURES Figure 1. Selected fermentation products derived from D-glucose .................................... 3 Figure 2. Commercially available and potential derivatives of L-lactic acid. .................... 5 Figure 3. Synthesis of polytrimethylene terephthalate. ..................................................... 6 Figure 4. Potential derivatives of succinic acid. ............................................................... 7 Figure 5. Chemicals synthesized from L-phenylalanine, L-tyrosine and L-tryptophan ....... 8 Figure 6. The shikimate pathway. .................................................................................. 10 Figure 7. Value-added aromatic compounds synthesized from D-glucose via 3-dehydroshikimic acid intermediacy ..................................................................... 13 Figure 8. Conventional synthesis and biosynthesis of catechol and adipic acid. ............. 16 Figure 9. Synthesis of vanillin via benzene or D-glucose ............................................... 18 Figure 10. Synthesis of p-hydroxybenzoic acid from glucose and potassium phenoxide. 19 Figure 11. Synthesis of gallic acid and pyrogallol from D-glucose. ................................ 21 Figure 12. Synthesis of phenol and p-hydroxybenzoic acid via intermediacy of shikimic acid .......................................................................................................... 22 Figure 13. Syntheses of polyhydroxybenzenes from D-glucose. ..................................... 24 Figure 14. Conventional approach and DNA shuffling for protein evolution .................. 27 Figure 15. Chemical synthesis of hydroquinone from benzene ....................................... 43 Figure 16. Microbial routes to hydroquinone from phenol and benzene. ........................ 44 Figure 17. Proposed biocatalytic synthesis of hydroquinone from D-glucose. ................ 45 Figure 18. Previously examined synthesis of hydroquinone from D-glucose. ................ 46 Figure 19. Selected molecules synthesized from quinic acid. ........................................ 49 Figure 20. Chemical synthesis of quinic acid from D-arabinose. .................................... 50 Figure 21. Chemical synthesis of quinic acid from ketol silyl ether 1 ............................. 51 Figure 22. The truncated aromatic amino acid biosynthetic pathway with quinic acid biosynthesis. .......................................................................................................... 53 Figure 23. Construction of plasmid pKDl2.l38 ............................................................. 55 Figure 24. E. coli QPl.1/pKD12.138 cultured under glucose-limited, fermentor- controlled conditions .............................................................................................. 57 Figure 25. Construction of plasmid pNR4.230. .............................................................. 60 Figure 26. E. coli QPl.1/pNR4.230 cultured under glucose-limited, fermentor- controlled conditions .............................................................................................. 61 Figure 27. Schematic of glucose transport and phosphorylation using PIS system. ....... 62 Figure 28. E. coli QPl.1/pKD15.071 cultured under glucose-limited, fermentor- controlled conditions .............................................................................................. 64 Figure 29. E. coli QPl.1/pNR4.272 cultured under glucose-limited, fermentor- controlled conditions .............................................................................................. 65 Figure 30. Construction of plasmid pKD15.071 ............................................................. 66 Figure 31. Construction of plasmid pNR4.272. .............................................................. 67 Figure 32. Schematic of non-PTS glucose transport and phosphorylation. ..................... 68 Figure 33. Construction of plasmid pNR4.276. .............................................................. 70 Figure 34. Synthesis of 3-dehydroquinic acid from quinic acid. ..................................... 71 Figure 35. Biosynthesis of quinic acid from the added 3-dehydroquinic acid by E. coli QPl.1/pNR4.276 under glucose-limited, fermentor—controlled conditions .............. 72 Figure 36. E. coli QPl.1/pKD12.138 cultured under glucose-rich, fermentor- controlled conditions ............................................................................................. 73 Figure 37. E. coli QPl.1/pNR4.230 cultured under glucose-rich, fermentor- controlled conditions ............................................................................................. 74 Figure 38. E. coli QPl.1ptsG/pKD12.138 cultured under glucose-rich, fermentor- controlled conditions .............................................................................................. 76 Figure 39. 3-Dehydroquinate-synthesizing E. coli QPl.1/pKL4.33. ............................... 78 xi Figure 40. Conversion of quinic acid to hydroquinone. .................................................. 81 Figure 41. Dehydration and aromatization of 3(R),5(R)-trihydroxycyclohexanone (10) to hydroquinone. ........................................................................................... 82 Figure 42. Syntheses of enone intermediates 11 and 12 from quinic acid. ...................... 83 Figure 43. Diagram of electrochemical oxidation apparatus. .......................................... 84 Figure 44. Industrial chemicals derived from gallic acid. ............................................. 102 Figure 45. Biosynthesis of 1,2,3,4,6-0-pentagalloylglucose ......................................... 104 Figure 46. Three proposed pathways for biosynthesis of gallic acid. ............................ 105 Figure 47. l3C-Abundance (in percentage) of L—phenylalanine, L-tyrosine and L-tryptophan from P. blakesleeanus cultured with [1-‘3C]-glucose. ...................... 108 Figure 48. Predicted and observed labeling patterns of gallic acid synthesized by P. blakesleeanus cultured with [l-‘3C]-glucose. ................................................... 109 Figure 49. Strategy for distinguishing proposed gallic acid biosynthetic pathways with [5-‘80]-3-dehydroshikimic acid. .................................................................. 110 Figure 50. Synthesis of [5-'SO]-3-dehydroshikimic acid from shikimic acid ................. 111 Figure 51. Gallic acid and 3-dehydroquinic acid production upon addition of [5-‘80]-enriched 3-dehydroshikimic acid by E. coli KL3/pRC1.55B. ................... 116 Figure 52. Microbial synthesis of gallic acid by exploiting pobA* activity ................... 118 Figure 53. Gallic acid and 3-dehydroquinic acid production upon addition of [5-'SO]-enriched 3-dehydroshikimic acid by E. coli KL7/pSK6.76. ...................... 119 Figure 54. Mechanism for base-catalyzed air oxidation of 3-dehydroshikimic acid ...... 121 Figure 55. Proposed mechanism for 'sO-exchange of 3-dehydroshikimic acid in inorganic phosphate buffered water ................................................................. 122 Figure 56. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains reliant on PT S for glucose transport. .............................................. 133 Figure 57. Previously examined strategies to increase phosphoenolpyruvate availability. ............................................................................................................................ 135 xii Figure 58. Proposed pyruvate-based shikimate pathway and the native phosphoenolpyruvate-based shikimate pathway. .................................................. 136 Figure 59. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains employing pyruvate-based shikimate pathway .......................... 137 Figure 60. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains with fully recycling pyruvate to phosphoenolpyruvate. ............. 140 Figure 61. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains employing non-PTS glucose transport. ..................................... 142 Figure 62. Pathway for D-galactonate catabolism in E. coli .......................................... 146 Figure 63. KDPGal aldolase (dgoA) and KDPG aldolase (eda) catalyzed reactions ...... 147 Figure 64. Synthesis of 2-keto-3-deoxy-6-phosphogalactonate. .................................. 152 Figure 65. Byproduct formation mechanism in the fi-elimination reaction. .................. 152 Figure 66. Construction of plasmid pNR5.223. ............................................................ 154 Figure 67. Construction of plasmid pNR6. 106. ............................................................ 155 Figure 68. Construction of plasmid pNR7.088. ............................................................ 162 Figure 69. Construction of plasmid libraries of pEC01, pEC02 and pECO3 .................. 164 Figure 70. Construction of plasmid libraries of pKPOl, pKPOZ and pKPO3 .................. 170 Figure 71. Construction of plasmid libraries of pSTOl , pST02, pSTO3 and pST04 ....... 172 Figure 72. Growth in the absence of aromatic amino acid and aromatic vitamin supplementations in glucose-containing minimal salts medium under shake-flask conditions .......................................................................................... 175 Figure 73. Construction of plasmid pNR7. 126. ............................................................ 176 Figure 74. Construction of plasmid pNR8.172. ............................................................ 182 Figure 75. Construction of plasmid pKP03-3serA ........................................................ 183 Figure 76. Construction of plasmid pST04-58erA. ....................................................... 184 Figure 77. E. coli NR7/pKP03-3serA cultured under glucose—rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 185 xiii Figure 78. E. coli NR7/pNR8.074 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 185 Figure 79. E. coli NR7/pNR8.172 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 186 Figure 80. E. coli NR7/pNR8.170 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 186 Figure 81. E. coli NR7/pST04-58erA cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 187 Figure 82. E. coli NR7/pNR8.121 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 187 Figure 83. E. coli NR7/pNR8.165-4serA cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 191 Figure 84. E. coli NR7/pNR8.180 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 191 Figure 85. E. coli NR7/pNR8.182 cultured under glucose-rich, fermentor- controlled conditions (36 °C, 20% D. O.) ............................................................. 192 Figure 86. Construction of plasmids pNR8.165-25erA and pNR8.165-4serA. .............. 193 Figure 87. Construction of plasmid pNR8.180. ............................................................ 194 Figure 88. Construction of plasmid pNR8.182. ............................................................ 195 Figure 89. Construction of plasmid pNR8.187. ............................................................ 196 Figure 90. Construction of plasmid pNR8.189. ............................................................ 197 Figure 91. Construction of plasmid pNR8. 190. ............................................................ 198 xiv Ac ADP ATP AP ApR CA CIAP Cm C m” DAHP DCU DEAD DEAE DHQ DHS D.O. D'I'I‘ E4P EPSP FBR GA HPLC LIST OF ABBREVIATIONS acetyl adenosine diphosphate adenosine triphosphate ampicillin ampicillin resistance gene base pair chorismic acid calf intestinal alkaline phosphatase chloramphenicol chloramphenicol resistance gene 3-deoxy-D-arabino-heptulosonic acid 7-phosphate digital control unit diethyl azodicarboxylate diethylaminoethyl 3-dehydroquinic acid 3-dehydroshikimic acid dissolved oxygen dithiothreitol D-erythrose 4-phosphate 5-enolpyruvoylshikimate 3-phosphate feedback resistant gallic acid hour high pressure liquid chromatography XV Kan Kan“ kb KDPG KDPGal kg m LB M9 min NAD N ADH NADP NADPH NMR OD ORF PCA PEP PHB isopropyl fi-D-thiogalactopyranoside kanamycin kanamycin resistance gene kilobase 2-keto-3-deoxy-6-phosphogluconate 2-keto-3-deoxy-6-phosphoga1actonate kilogram Michaelis constant Luria-Bertani molar minimal salts minute milliliter microliter rrrillimolar micromolar nicotinamide adenine dinucleotide, oxidized form nicotinamide adenine dinucleotide, reduced form nicotinamide adenine dinucleotide phosphate, oxidized form nicotinamide adenine dinucleotide phosphate, reduced form nuclear magnetic resonance spectroscopy optical density open reading frame protocatechuic acid phosphoenolpyruvic acid p-hydroxybenzoic acid xvi PID PCR Phe Pyr Tyr Trp psi PTS QA SA SDS S3P Tc TCA TSP TsOH UDP UV proportional-integral-derivative polymerase chain reaction L-phenylalanine pyruvate L-tyrosine L-tryptophan pounds per square inch phosphotransferase system quinic acid rotations per minute shikirrric acid sodium dodecyl sulfate shikimate 3-phosphate tetracycline tricarboxylic acid sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4 p-toluenesulfonic acid Uridine 5’-diphosphate ultraviolet xvii CHAPTER ONE Introduction Today the chemical industry is dominated by technologies that rely on starting materials derived from petroleum and natural gas. The increasing scarcity and cost of petroleum and natural gas1 are contributing to a growing interest in the use of renewable feedstocks to improve long-term sustainability. Carbohydrates such as D-glucose derived from plant-derived starch and cellulose represent a promising renewable starting material that could serve as an alternative to starting materials derived from fossil fuels. Consequently, biocatalysis has now emerged as a promising and fast-developing tool to harvest the energy and chemical building blocks locked in the carbon and oxygen atoms of carbohydrates. Starch from corn serves as the primary feedstock for today’s bioproducts industry.2 The starch generated by com wet milling is hydrolyzed to yield D-glucose that is then converted via microbial synthesis into products ranging from fuel ethanol to amino acids such as L-lysine as well as an emerging new product polylactide polymer. Cellulose is likely to serve as a valuable source of D-glucose in the future when cost- efficient depolymerization technology is available.3 With the vast amount of carbon available from the glucose in cellulose, microbial synthesis holds great potential for the manufacturing of many new industrial chemicals. In Chapter 2 of this dissertation, the synthesis of hydroquinone from D-glucose via a combination of microbial and chemical synthesis is described. Previous work in the Frost group explored microbial synthesis of quinic acid via the shikimate pathway in a genetically modified Escherichia coli strain followed by chemical conversion of quinic acid to hydroquinone.4 The overall yield of hydroquinone synthesized from D-glucose was 3%. Chapter 2 details the further improvement of microbial synthesis of quinic acid and the elaboration of a high-yielding chemical conversion of quinic acid into hydroquinone. Chapter 3 investigates the biosynthetic pathway responsible for the formation of gallic acid in Escherichia coli. This study begins with the chemical synthesis of [5-130]-3-dehydroshikimic acid. The 18O labeling in gallic acid synthesized by E. coli from [5-180]-3—dehydroshikimic acid is then determined. Examination of the extent of 18O labeling in product gallic acid excludes the involvement of protocatechuic acid as an intermediate and favors the direct oxidation of 3-dehydroshikimic acid as the primary biosynthetic route to gallic acid in E. coli. Chapter 4 presents the creation of a pyruvate-based shikimate pathway variant. In this novel shikimate pathway, the first catalytic step entails the condensation of pyruvate with D-erythrose 4-phosphate catalyzed by 2-keto—3-deoxy-6-phosphogalactonate (KDPGal) aldolase mutants created by directed evolution. By contrast, the first committed step in the native shikimate pathway is the condensation of phosphoenolpyruvate with D-erythrose 4-phosphate. The pyruvate-based shikimate pathway establishes a completely new strategy for increasing phosphoenolpyruvate availability in microbes. Biosyntheses of value-added chemicals from D-glucose Interestingly, only one hundred years ago biomass-derived carbohydrates played a significant role in the US economy. In the late 1800’s, the largest selling chemicals were alcohols made from wood and grain, and the first man-made plastics, parkesine and celluloid (cellulose nitrate), were derived from cellulose.2'S By the 1970’s, hydrocarbon resources were established as the world’s primary feedstock for chemical manufacture. In many cases, it is still more economical to produce chemicals from petroleum or natural gas than from plant-derived carbohydrates. Today the advances in biocatalysis and bioprocessing are beginning to make an impact on reducing the cost of producing industrial chemicals from renewable carbohydrates and making them more competitive with those products derived from petroleum (Figure 1). /\OH ethanol HO HO O _ Reichstein S. cerevrsrae OH synthesis 0 HO OH succinic acid L—ascorbic acid \\ / A. subxoydans OH 5 00” AOH L b I ' i OH . u gancus HO OH O. glutaminicum 9H // + /'\n’OH D-glucose NH3 0 H3KIWC02- L-lactic acid I E. coli L-Iysine 1 ,3-propanediol Figure 1. Selected fermentation products derived from D-glucose. One example is the microbial synthesis of L-lactic acid from corn starch for use in production of polylactides. L—Lactic acid is primarily used as an acidulant in food and beverages, an electroplating bath additive, and a textile and leather additive.6 It is estimated that 33 x 106 kg/year of L-lactic acid are used in the food industry alone. Homofermentative production of L-lactic acid by microbes like Lactobacillus bulgaricus is accomplished at titers of 100 g/L and in yields of 95% from starch by an enzymatic saccharification/fermentation process.7 Polylactide is a biodegradable thermoplastic polymer produced from L-lactic acid through condensation and ring-opening polymerization performed using a solvent—free melt process.8 Polylactides exhibit several desirable properties including high-stiffness, clarity and gloss, grease resistance and flavor barrier properties suitable for use in the packaging industry.9 A joint venture between Cargill and Dow Chemical started up their first large-scale polylactide plant in Blair, Nebraska, with a 140 x 106 kg/year capacity.l0 Cargill Dow projects a possible market of 3.6 x 109 kg/year by 2020.ll Ethyl lactate is another L-lactic acid derivative that has recently been commercialized. It’s an environmentally benign solvent with properties superior to many conventional petroleum-based solvents. The use of ethyl lactate has been limited due to high production costs leading to high selling prices ($3.30~4.40/kg). Ethyl lactate is currently being used by Vertec Biosolvents Inc. in soy oil-solventblends.12 Acrylic acid, pyruvic acid, L-propylene glycol and 2,5-dimethyl-1,4- dioxane are other potential L-lactic acid derivatives which are currently produced from petroleum (Figure 2).2 O /\n,O\/ O n O / polylactide ethyl lactate 9” 0H /\n/OH ______ * ArOH /-\/OH 0 ~. 0 L-lactic acid acrylic acid L-propylene glycol ‘ \ O O Jere. /[ 3’ O O pyruvic acid 2,5—dimethyl-1,4—dioxane Figure 2. Commercially available and potential derivatives of L-lactic acid. Recent advances in biotechnology also made an inroad in manufacturing of 1,3- propanediol. 1,3-Propanediol, together with terephthalic acid, is used to produce polytrimethylene terephthalate (PTT, Figure 3). PTT is a polymer with properties such as good resilience, stain resistance and low static generation relative to its competitors PET (polyethylene terephthalate) and nylon in fiber and textile applications. PTT is currently manufactured by Shell Chemical and Dupont. Shell developed a petrochemical route to 1,3-propanediol by the reaction of ethylene oxide with carbon monoxide and hydrogen.13 Dupont currently makes 1,3-propanediol from propylene via acrolein. Genecor International and Dupont have been collaborating to develop the metabolic pathway in E. coli to produce 1,3-propanediol from D-glucose at a lower cost. In 2002, a pilot-scale 1,3-propanediol operation was built in Decatur, Illinois.” It is estimated that 1,3- propanediol has a potential 2020 market of 220 x 10‘5 kg. COZH 0M0, HO OH HOZC i Figure 3. Synthesis of polytrimethylene terephthalate. Succinic acid, with an annual production at 15 x 106 kg, is used as a surfactant and ion chelator in electropolating, food and pharmaceuticals. Industrial succinic acid is currently produced from butane via intermediacy of maleic anhydride.15 Food-grade succinic acid could be produced through fermentation of glucose. Over the past 5-10 years, advances in fermentation and especially separation technology have reduced the potential production cost of microbial synthesized succinic acid from $3.30 - $4.40/kg in 1992 to about $1.10/kg today. Further advances could significant reduce the cost of the bio-based succinic acid that would expend the current market and open up new applications. Succinic acid forms a platform from which many chemicals can be produced. Routes from succinic acid to chemicals including tetrahydrofuran (THF), 1,4- butanediol, y-butyrolactone and succinate salts are nearly cost-competitive with their fossil fuel-based counterparts. Current production of 1,4-butanediol is based on the Reppe process in which acetylene is reacted with formaldehyde to butynediol followed by catalytic hydrogenation to 1,4-butanediol.l6 It can also be produced via acetoxylation of butadiene or hydroformation of allyl alcohol.16 THF is an important solvent and is also used in the manufacture of polytetramethylene glycol. It is currently manufactured by the cyclodehydration of 1,4-butanediol and accounts for 48% of total 1,4-butanediol consumption. HONOH U o O HO ______ § 1 ,4-butanediol THF NOH ....... - O succinic acid 0 9H3 do Cr” y—butyrolactone N-methyl pyrrolidone Figure 4. Potential derivatives of succinic acid. Amino acids. Microbial production of amino acids has long been exploited in industry. The largest volume amino acids produced are L-glutamate, L-lysine and D,L- methionine. For instance, about 1 x 109 kg of L-glutamate are produced annually as a flavoring agent, and more than 4.4 x 108 kg/year of L-lysine are produced worldwide which is mainly used as an animal feed additive.l7 Both L-glutamate and L-lysine are produced industrially using Corynebacterium glutamicum, which is a L—glutamate accumulating soil bacterium discovered in I957, from molasses (cane or sugar-beet) or starch hydrosylates. Microbe-synthesized L-lysine can reach a titer of 170 g/L with 54% yield based on carbohydrate starting material.'7 I Compared to the bulk production of L-glutamate and L-lysine, the aromatic amino acids L-phenylalanine, L-tyrosine and L-tryptophan are produced at significantly smaller volumes. L-Phenylalanine is produced predominantly for the production of the low- calorie sweeter aspartame using the NutraSweet process.18 The worldwide production for L-phenylalanine is 1.2 x 107 kg/year. Fermentation processes for L-phenylalanine synthesis have been developed for genetically engineered strains of C. glutamicum, Brevibacterium lactofermentum, Bacillus sublitis and E. coli."”20 L-Tyrosine is produced in an even smaller volume (1.5 x 105 kg/year) and is used for the production of the anti- Parkinson’s drug L-DOPA and as a dietary supplement.21 Microbial synthesis of L- tyrosine mainly employs strains of E. coli, Bacillus substilis or various forms of coryneform bacteria.22 L-Tryptophan is produced at about 6 x 105 kg/year as a feed additive. Introduction of naphthalene dioxygenase into a tryptophan—synthesizing microbe that also expresses tryptophanase results in biocatalytic synthesis of indigo,23 the vat dye that gives blue jeans the faded-blue coloration. Microbial engineering for production of L-tryptophan, has primarily been carried out in E. coli,24 Corynebacterium glutamicum25 and Bacillus subtilis.26 co H 0 2 ———> PERI-J N OCH3 NH2 ' H L-phenylalanine aspartame COZH HO COZH HO 2 HO 2 L-tyrosine L'DOPA C02H 0 Ki Q50... ——»__.. g _ 0 N N H H O L-tryptophan indigo Figure 5. Chemicals synthesized from L-phenylalanine, L-tyrosine and L- tryptophan. The shikimate pathway. The shikimate pathway, also referred to as the common pathway of aromatic amino acid biosynthesis, is essential for the transformation of simple carbohydrate precursors into the aromatic amino acids L-phenylalanine, L-tyrosine and L- tryptophan in plants, bacteria and fungi.27 Mammals are incapable of de novo biosynthesis of aromatic amino acids. Therefore L-phenylalanine and L-tryptophan are essential components of animal diet, while animals can synthesize L-tyrosine in a single step from L-phenylalanine. The substrates and products of the seven enzymatic reactions that convert phosphoenolpyruvate (PEP) and D-erythrose 4-phosphate (E4P) into chorismic acid (Figure 6) were identified by the early 1960’s from studies of auxotrophs of Escherichia coli and Klebsiella aerogenes. The first committed step catalyzed by DAHP synthase involves condensation of PEP and E4P to form 3-deoxy-D-arabino-heptulosonic acid 7- phosphate (DAHP). Three isozymes of DAHP synthase exist in E. coli, each of which is sensitive to feedback inhibition by one of the three aromatic amino acids. The genes aroF, aroG and wall encode for L-tyrosine-sensitive, L-phenylalanine-sensitive, and L- tryptophan-sensitive isozymes of DAHP synthase, respectively. DAHP is converted into 3-dehydroquinic acid (DHQ) by DHQ synthase which is encoded by aroB28 in a complex reaction involving an intramolecular oxidation-reduction at C-5 of DAHP with a very tightly bound NAD+ cofactor, a syn elimination of phosphate, and an intramolecular cyclization.29 A syn elimination of water from DHQ affords 3-dehydroshikimic acid (DHS).30 This reaction is catalyzed by 3-dehydroquinate dehydratase, which is a type I enzyme encoded by aroD. H203PO phosphoenolpcrozgt H9304 HO CC32H H3PO4 HO, COZH H20 y uv e :. H aI‘OH H203PO 5H OH H203PO oH DAHP 3-dehydroquinic acid D-erythrose 4-phosphate 002 H OHZ COZH NAD(P) )H NAD(P) ATP ADP PEP H3PO4 0/131 OH aroE 0,.[6 01M aroK - aroA m3 H OH and H203Po“ i OH OH OH 3—dehydroshikimic acid shikimic acid shikimate 3-phosphate COzH H3P04 COzH C02H 4 —'—>* Ubiquinones Hzoapo‘“ i OJLCOZH 3’06 .zo’cho H :H OH EPSP chorismic acid hydrooxybenzoic acid CO2H HO COzH 002H " COZH (j €3.01 0 COQH N...2 i 2-aminobenzoic acid isochorismic acid 4—aminobenzoic acid prephenic acid _ menaquinones - - L-tyrosine L tryptophan enterobactins mm and L-phenylalanine Figure 6. The common pathway of aromatic amino acid biosynthesis and pathways beyond chorismate. Intermediates (abbreviations): 3-deoxy-D-arabino-heptulosonic acid 7-phosphate (DAHP), 5-enolpyruvylshikimate-3-phosphate (EPSP). Genes, enzymes: aroF, DAHP synthase (tyrosine); aroG, DAHP synthase (phenylalanine); aroH, DAHP synthase (tryptophan); aroB, DHQ synthase; aroD, DHQ dehydratase; aroE, shikimate dehydrogenase; ydiB, shikimate dehydrogenase; aroK, shikimate kinase 1; aroL, shikimate kinase II; aroA, EPSP synthase; aroC, chorismate synthase. 10 Reduction of 3-dehydroshikimic acid to shikimic acid in the presence of NADPH is catalyzed by aroE-encoded shikimate dehydrogenase.31 Recently, a NADH-dependent shikimate dehydrogenase isozyme YdiB32 was identified in E. coli. Shikimic acid is further converted to shikimate 3-phosphate by phosphoryl group transfer from ATP catalyzed by shikimate kinase. E. coli K12 carries two isozymes of shikimate kinase encoded by the loci aroL33 and aroK.34 5-Enolpyruvoylshikimate 3-phosphate (EPSP) synthase, encoded by aroA,35 catalyzes the reversible formation of 5- enolpyruvoylshikimate 3-phosphate and inorganic phosphate from shikimate 3-phosphate and phosphoenolpyruvate. The last enzyme of the common pathway is chorismate synthase,36 encoded by aroC. It catalyzes the concerted 1,4-trans elimination of phosphate from 5-enolpyruvylshikimate-3-phosphate to afford chorismic acid, the precursor for biosynthesis of three aromatic amino acids L-tryptophan, L-tyrosine, and L- phenylalanine (Figure 6). In addition, biosynthetic pathways leading to ubiquinone, folic acid and enterochelin also branch away from the common pathway at chorismic acid (Figure 6). Folic acid-derived coenzymes are frequently involved in the biosynthetic transfer of one carbon fragments, ubiquinones are involved in electron transport, and enterochelin is an iron chelator responsible for iron uptake in numerous microorganisms. Regulation of the shikimate pathway in prokaryotes such as E. coli involves both feedback inhibition and transcriptional control. DAHP synthase isozymes are sensitive to feedback inhibition by the respective aromatic amino acids. Transcription of genes encoding the L-tyrosine-and L-phenylalanine-sensitive isozymes are regulated by the tyrR repressor and transcription of the L-tryptophan-sensitive isozyme is regulated by the trpR repressor.37 Shikimate kinase may represent a secondary control in the shikimate 11 pathway. E. coli shikimate kinase encoded by aroK is synthesized constitutively, but the synthesis of the other shikimate kinase encode by aroL, which provides the major activity in vivo, is under control of the tyrR and trpR proteins.37 The essential role of the shikimate pathway in bacteria, plants, fungi and parasites but its absence in mammals makes the enzymes involved in this pathway appealing targets for the development of broad-spectrum antibiotic drugs and herbicides. The widely used herbicide glyphosate, marketed under the trade name Roundup”,38 inhibits the sixth enzyme in the pathway, 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase. In addition to herbicidal activity, glyphosate also displayed antiparasite activity towards apicomplexan parasites including Toxoplasma gondii, which causes toxoplasmosis.39 Biocatalytic syntheses of aromatics by manipulating the shikimate pathway. The shikimate pathway has provided convenient access to a variety of aromatic chemicals including catechol,40 vanillin,41 benzoquinone and hydroquinone,42 gallic acid,” 1’43 145 pyrogallo p-hydroxybenzoic acid,44 and pheno in addition to various non-aromatic compounds such as shikimic acid,4 3-dehydroshikimic acid,46 quinic acid4 and adipic acid.47 The benzene-free synthesis of aromatic molecules from carbohydrates constitutes a different path from conventional organic synthesis used by the chemical industry. Use of non-toxic and non-volatile carbohydrate feedstocks avoids the health problems and expenses associated with the use of carcinogenic and volatile benzene and its derivatives. Biocatalytic conversion of a shikimate pathway metabolite into an aromatic molecule represents a straightforward approach to harness the potential of this pathway. Examples include the microbial conversion of D-glucose into catechol, gallic acid, 12 pyrogallol, vanillic acid and p-hydroxybenzoic acid by a genetically-modified microbial host. 3-Dehydroshikimic acid serves as a branch point for the biosyntheses of these aromatics via the common pathway (Figure 7). COzH COZH £5. ->——» O. I i O 5 OH e O COZH HO OH OH OH D—glucose 3-dehydroshikimic chorismic acid acid i/ la \2 COzH COZH o’ H:Hd o’ H:20H ‘ Ho“ : H'OH protocatechuic acid gallic acid shikimic acid 3/ \‘i "/ \1 1f COzH COZH [:[OH CH30‘ :H HO' ;H OH 5 OH catechol vanillic acid PYrogallol phenol p-hydroxybezoic acid Figure 7. Value-added aromatic compounds synthesized from D-glucose via 3- dehydroshikimic acid intermediacy. Keys: (a) aroZ, 3-dehydroshikimate dehydratase; (b) aroY, protocatechuate decarboxylase; (c) (i) COMT, catechol-0-methyltransferase, (ii) aryl aldehyde dehydrogenase; (d) p0bA*, p-hydroxybenzoate hydroxylase; (e) 02, Cu“, Zn“, AcOH; (f) aro Y, PCA decarboxylase; (g) aroE, shikimate dehydrogenase; (h) H20, 350 °C; (i) 1 M H2804 in AcOH; (j) ubiC, chorismate-pyruvate lyase. 13 Li et al. developed E. coli KL3/pKL4.13OB to convert D-glucose into 3- dehydroshikimic acid.“S E. coli KL3 contains a mutation in the aroE gene encoding shikimate dehydrogenase, which renders E. coli KL3 incapable of converting 3- dehydroshikimic acid into shikimic acid. Enhancing the production of 3- dehydroshikimic acid requires transforming E. coli KL3 with a multicopy plasmid pKL4.130B containing PamparoFFBR, which encodes a feedback-insensitive isozyme of DAHP synthase, and a copy of tktA48 to increase the activity of transketolase. Overexpression of AroFFBR and transketolase increase carbon flow into the shikimate pathway. E. coli KL3/pKL4.130B synthesized 69 g/L of 3-dehydroshikimic acid in 30% (mol/mol) yield from D-glucose under fed-batch fermentor conditions. The yield was further improved to 35% (mol/mol) by overexpression of a ppsA-encoded phosphoenolpyruvate synthase in a similar construct developed by Yi et al.49 Catechol and adipic acid. 3-Dehydroshikimic acid can be converted into catechol by introducing DHS dehydratase and PCA decarboxylase activities into a DHS- synthesizing strain. Catechol is an important chemical building block used in synthesis of a variety of flavors (vanillin, eugenol, isoeugenol), pharmaceuticals (L-DOPA, adrenaline, papaverine), agrochemicals (carbofuran, propoxur) and polymerization inhibitors and antioxidants (4-tert-butylcatechol, veratrol). Global production of catechol is estimated at 2.5 x 107 kg/year.50 Most catechol production begins with Friedel-Crafts alkylation of benzene to afford cumene followed by Hock-type air oxidation of cumene to acetone and phenol (Figure 8). The phenol is then oxidized to a mixture of catechol and hydroquinone using 70% hydrogen peroxide either in the presence of transition metal catalysts or in formic acid solution where performic acid is the actual oxidant. Catechol l4 and hydroquinone are separated by distillation.51 As an alternative to benzene as a starting material, D-glucose can be converted in an aroE auxotroph to 3-dehydroshikimic acid by plasmid-based expression of AroFFBR, AroB, and TktA. A second plasmid incorporates a copy of aroZ and aroY isolated from Klebsiella pneumoniae. 3- Dehydroshikimic acid is aromatized to protocatechuic acid by aroZ-encoded 3- dehydroshikimic acid dehydratase, which then undergoes an enzymatic decarboxylation to catechol by the aroY—encoded PCA decarboxylase. Biocatalytic production with the pre-grown cells of AB2834/pKDl36/pKD9.069A afforded catechol at concentrations of 2 g/L and 33% yield in minimal salts medium containing D-glucose.40 Although the toxicity of catechol towards E. coli hindered significant accumulation of catechol, inclusion of a catA encoding catechol 1,2-dioxygenase in the catechol producing strain converted catechol into cis,cis-muconic acid. The catA- encoded catechol 1,2-dioxygenase was isolated from Acinetobacter calcoaceticus.52 Catalytic hydrogenation of cis,cis—muconic acid under mild conditions afforded adipic acid (Figure 8),47a a monomer for synthesis of nylon-6,6.53 The global demand for adipic acid exceeds 1.9 x 109 kg/yr. Manufacture of adipic acid starts with hydrogenation of benzene to cyclohexane followed by air oxidation to cyclohexanol and cyclohexanone (Figure 8). Exhaustive oxidation of the cyclohexyl intermediates with 60 per cent nitric acid in the presence of copper and vanadium catalysts yields adipic acid.53 Nitrous oxide is produced as a major byproduct of this process. Adipic acid production was estimated to account for 10% of the atmospheric nitrous oxide level which contributes to ozone depletion and global warming.54 The process also requires use of elevated temperatures and pressures. 15 acetone o9» o*—.» W hydroquinone benzene cumene phenol >3 OH COZH COZH _. _, _, Q i OH O i OH HO OH HO OH OH OH D-glucose 3-dehydroshikimic protocatechuic catechol acid acid OH 1d (j © fig cyclohexanol h ——> —-> O benzene H020 adipic acid cyclohexanone e / COZH H02C / cis,cis-muconic acid Figure 8. Conventional synthesis and biosynthesis of catechol and adipic acid. Keys: (a) propylene, 400-600 psi., solid H3PO4 catalyst, ZOO-260°C; (b) 02, 80-130 °C, 802, 60-100 °C; (0) 70% H202, EDTA, Fe2+ or C0“, 70-80 °C; ((1) E. coli WNl/pWN2.248; (e) H2, 50 psi., 10% Pt/C; (f) Ni-A1203, H2, 370-800 psi., 150-250 °C; (g) Co, 02, 120-140 psi., 150-160 °C; (h) Cu, NH4VO3, 60% HNO3, 60-80 °C. The microbial-based route to adipic acid via catechol intermediacy has been improved from the previously reported synthesis under shake-flask conditions. The microbial synthesis utilized an aroE auxotroph E. coli WN 1 expressing aroZ-encoded DHS dehydratase for the conversion of DHS to protocatechuic acid, aroY-encoded PCA decarboxylase for the conversion of PCA to catechol, and catA-encoded catechol 1,2- dioxygenase for the conversion of catechol to cis,cis-muconic acid. The construct E. coli 16 WNl/pWN2.248 synthesized 37 g/L cis,cis-muconic acid from D-glucose in 23% yield (mol/mol) (Figure 8). Hydrogenation of clarified cell broth with 10% WC (5% mol/mol) and 50 psi H2 afforded adipic acid in 97% yield from cis,cis-muconic acid. The interface of biocatalysis and catalytic hydrogenation allows adipic acid to be produced without formation of large amounts of nitrous oxide. Vanillin. Vanillin is one of the most important aromatic flavor compounds used in the food and beverage industry and also finds use in formulations for perfumes.55 Natural vanillin obtained from the vanilla beans of the orchid Vanilla planifolia accounts for only 2 x 104 kg/yr of the world’s 1.2 x 107 kg/yr demand for vanillin.56 The difference is made up by synthetic vanillin starting from petroleum-derived guaiacol. Condensation of guaiacol with glyoxylic acid affords mandelic acid (Figure 9). Oxidation of mandelic acid and subsequent decarboxylation results in vanillin. Synthetic vanillin sells for $12/kg while natural vanilla flavoring extracted from vanilla bean containing 2% vanillin sells for $30-120/kg.57 The high price for natural vanilla flavoring reflects the labor-intensive cultivation, pollination, harvesting and curing of vanilla beans. The demand for natural flavorings has, in turn, prompted the development of biocatalytic routes to vanillin. Biocatalytic conversion of D-glucose to vanillin passes through the intermediacy of 3-dehydroshikimic acid. Heterologous expression of the aroZ locus in E. coli aroE auxotroph KL7 leads to protocatechuic acid as previously discussed. Expression of rat- liver COMT-encoded catechol-0-methyltransferase in KL7 afforded 4.9 g/L of vanillic acid by fed-batch fermentation from D-glucose when the construct was supplemented with L-methionine (Figure 9). COMT catalyzes the methylation of protocatechuic acid to 17 a mixture of vanillic acid and isovanillic acid. The in vitro reduction of vanillic acid to vanillin was carried out by aryl aldehyde dehydrogenase purified from the fungus Neurospora crassa in 66% yield.58 This two-step biocatalytic synthesis of vanillin is the only biocatalytic synthesis of vanillin using a carbohydrate as a starting material.57 HO COZH +—a «r [2 H300 H300 OH OH benzene mandelic guaiacol acid lb COCOZH CHO I COZH <03 e -—C-> «— H3CO H300 H300 OH OH phenylglyoxilic vanillin vanillic acid acid Id COZH COZH d dHOE> a on . HO OH OH D-glucose 3-dehydroshikimic protocatHechuic acid acid Figure 9. Synthesis of vanillin via benzene or D-glucose. Keys: (a) HCOCOZH; (b) 02; (c) H+; (d) KL7/pKL5.26A; (e) N. crassa aryl aldehyde dehydrogenase. p-Hydroxybenzoic acid. p—Hydroxybenzoic acid is a component of liquid crystal pOIymers such as Xydar,59 which have attracted considerable attention because of their 18 use in high-performance applications. Esters of p-hydroxybenzoic acid are also widely used as food preservatives.60 p -Hydroxybenzoic acid is currently manufactured by Kolbe-Schmitt reaction of dried potassium phenoxide with 20 atm dry CO2 at 180-250 °C. Product potassium p-hydroxybenzoate is converted to its free acid upon addition of mineral acid (Figure 11). Besides the required temperatures and pressures, p- hydroxybenzoic acid manufacture has to contend with handling of phenol which is a highly toxic, corrosive chemical.“ Chorismic acid can also be converted directly into p-hydroxybenzoic acid in a reaction catalyzed by ubiC-encoded62 chorismate-pyruvate lyase. A genetically-modified E. coli biocatalyst that synthesizes elevated concentrations of chorismic acid and overexpresses the ubiC-encoded chorismate-pyruvate lyase afforded 12 g/L of p- hydroxybenzoic acid in 13% yield from glucose under fermentor-controlled conditions.44 OH COZH COZH ,\OH —> . —-> JL é OH Ho“ i OH 5 O CO2H HO OH OH OH D-glucose shikimic acid chorismic acid lo COZK ' cozH O . b —-> -—> OK OH OH potassium potassium p-hydroxybenzoic phenoxide p-hydroxybenzoate acid Figure 11. Synthesis of p-hydroxybenzoic acid from glucose and potassium phenoxide. Keys: (a) 20 atm C02, 180-250 °C; (b) H*; (c) ubiC, chorismate-pyruvate lyase. 19 Gallic acid and pyrogallol. The biosynthesis of another molecule that stems from 3-dehydroshikimic acid is 3,4,5-tn'hydroxybenzoic acid, which is commonly known as gallic acid. This polyhydroxylated aromatic is currently isolated from gall nuts or from seed pods of Coulteria tinctoria trees found in Peru.63 Thermal decarboxylation of gallic acid in copper autoclaves affords pyrogallol.63 Two biocatalytic routes were developed with the aim to supplant isolation of gallic acid and pyrogallol from scarce natural resources. In one way, 3-dehydroshikimic acid in acetic acid solution can be oxidized by 02 in the presence of catalytic amounts of Cu2+ and Zn2+ to afford gallic acid in 67% yield.64 Alternatively, gallic acid can be obtained directly from D-glucose via protocatechuic acid intermediary (Figure 10).43 E. coli KL7/pSK6.161 expresses a plasmids-localized mutant isozyme of p-hydroxybenzoate hydroxylase encoded by plasmid-localized p0bA*,“ DHS dehydratase encoded by a genomic copy of aroZ and feedback-insensitive DAHP synthase encoded by plasmid-localized aroFtBR. E. coli KL7/pSK6.l61 afforded 20 g/L of gallic acid in 12% (mol/mol) yield from D-glucose under fermentor-controlled conditions.43 Decarboxylation of gallic acid to pyrogallol was conducted by E. coli RB791serAzzaroB/pSK6.234 expressing aro Y—encoded PCA decarboxylase. Addition of gallic acid to a batch culture of E. coli RB79lserAzzaroB/pSK6.234 during its stationary phase of grth afforded pyrogallol in a concentration of 14 g/L in 97% (mol/mol) yield. The high-yielding biocatalytic decarboxylation of gallic acid to pyrogallol provides an attractive alternative to currently employed chemical decarboxylation process. The toxicity of pyrogallol towards growing E. coli cells precluded the direct synthesis of pyrogallol from D-glucose using a single microbial construct. 20 OH COzH COZH a a -—> —> 5 OH O i OH HO HO OH OH OH D-glucose 3-dehydroshikimic protocatechuic acid acid b\ / a COZH E P c ‘— HO OH HO OH OH OH pyrogallol gallic acid Figure 10. Synthesis of gallic acid and pyrogallol from D-glucose. Keys: (a) E. coli KL7/pSK6.161, 12%; (b) 02, Cu2+, Zn2+, AcOH, 67%; (c) E. coli RB791serA::aroB/pSK6.234, 97%. Syntheses of aromatics via a combination of biocatalytic and chemical conversions. The direct conversion approaches for biosynthesis of aromatics must contend with toxicity of the aromatic products towards the microbial biocatalysts. For example, it only takes 2.5 mM catechol to completely inhibit growth of E. coli.40 Cells already in stationary phase can tolerate higher concentrations of catechol. As a result, the direct conversion of D-glucose to catechol was carried out by initial growth of the microbial biocatalyst to stationary phase followed by resuspension and culturing in minimal salts medium where synthesis of catechol occurred. Synthesis of pyrogallol also entailed addition of microbially synthesized gallic acid to pre-grown cells.43 Although it is possible to continuously remove the toxic aromatics during the biocatalytic conversions, an alternative strategy has been developed to avoid the toxicity of aromatics to biocatalysts. This strategy involves employing microbial catalysis to synthesize an 21 intermediate, which is not toxic to the microbe, followed by chemical conversion of the intermediate to the desired aromatic product. Examples include the syntheses of phenol, p-hydroxybenzoic acid, 1,2,3,4-tetrahydroxybenzene and hydroxyhydroquinone from D- glucose. Phenol and p-hydroxybenzoic acid. Phenol is used to make synthetic resins, dyes, pharmaceuticals, pesticides, perfumes, lubricating oils and solvents.66 The Hock oxidation of benzene-derived cumene is currently the predominant method used in the production of phenol with an annual production of 5 x 109 kg (Figure 8). It is estimated that 20% of the global benzene production is directed to the manufacture of phenol.67 OH COZH .\OH a _—_> . 3 OH Ho“ _._ OH HO OH OH D-glucose shikimic acid COzEt / i b X OOH gov. 'INH2°H3PO4 NHAC I OH (as-4104 phenol p-hydroxybenzoic acid COZH Figure 12. Synthesis of phenol and p-hydroxybenzoic acid via intermediacy of shikimic acid. Keys: (a) E. coli SP1.1/pKD12.138 or SP1.1pts/pSC6.090; (b) H20, 350 °C; (C) 1 M H2504 in ACOH. Commercially available shikimic acid is isolated from fruits of Illicium plant at a cost of approximately $10,000/kg.27a E. coli SP1.1/pKDl2.1384 was constructed for 22 microbial production of shikimic acid from glucose, and a further improved construct SP1.1pts/pSC6.090 has achieved a titer to 87 g/L and 36% (mol/mol) yield."3 The shikimate-producing biocatalyst has been licensed by Roche for the synthesis of neuraminidase inhibitor GS-4104, an anti-influenza drug sold under the name Tamiflumfr’9 E. coli synthesis of shikimic acid results from disruption of the genomic aroL and aroK loci, overexpression of feedback-insensitive, aroFFBR-encoded DAHP synthase, aroE-encoded shikimate dehydrogenase and tktA-encoded transketolase to channel more carbon into the shikimate pathway.4 Shikimic acid can also be used as an intermediate in the synthesis of phenol. A benzene-free route to phenol was achieved by aromatization and decarboxylation of microbes synthesized shikimic acid in near-critical water to afford phenol in an overall 14% yield from D-glucose (Figure 12).“ Conversion of shikimic acid to phenol in near-critical water circumvents the potent antimicrobial activity of phenol and avoids the use of strong acids. p-Hydroxybenzoic acid can also be synthesized from shikimic acid by refluxing shikimic acid in acetic acid containingl M sulfuric acid in an overall 15% yield from D-glucose.4S Polyhydroxybenzenes. The strategy of combining biocatalysis and chemical catalysis has also been used in the synthesis of 1,2,3,4-tetrahydroxybenzene from D- glucose (Figure 13). D-Glucose was converted into myo-inositol in 11% yield by an E. coli strain expressing myo-inositol l-phosphate synthase from Saccharomyces cerevisiae. Oxidation of the biosynthesized myo-inositol to myo-inosose was carried out by Gluconobacter oxidans ATCC 621 in 95% yield. myo-Inosose underwent aromatization under acidic conditions to afford l,2,3,4-tetrahydroxybenzene in 66% yield. In a similar 1.70 approach, Katinuma et a synthesized 2-deoxy-scyllo-inosose from D-glucose with an 23 enzymatic reaction catalyzed by 2-deoxy-scyllo-inosose synthase from Bacillus circulans in 38% yield. Treatment of the 2-deoxy-scyllo-inosose with H1 in acetic acid afforded catechol in 59% yield. Hansen et al. converted the 2-deoxy-scyllo-inosose into hydroxyhydroquinone in 39% yield by acid—catalyzed dehydration.7| OH OH OH Ho ' OH b Ho ' OH O HO 0 —-» —» HO:’:OH Ho‘o Ho OH OH OH myo-inositol mon-inosose 1,2,3,4-tetrahydroxy- benzene .\OI"I OH OH \d‘ i 0 6 HO OH H Oh. / HO D-glucose . OH HO OH \I‘ hydroxyhydroquinone OH 2-deoxy-scyIIo-inosose CEOH OH catechol Figure 13. Syntheses of polyhydroxybenzenes from D-glucose. Keys: (a) WP 1/pAD1.88A (b) Gluconobacter oxidans; (c) 0.5 M H2804, reflux; (d) (i) hexokinase, ATP, Mg“, (ii) 2-deoxy-scyllo-inosose synthase, NAD*; (e) 0.5 M H3PO4, reflux; (f) HI, HOAQ In Chapter 2 of this thesis, a synthetic route to hydroquinone from glucose that utilized this hybrid synthesis strategy will be presented. A genetically engineered E. coli construct produces quinic acid as a non-toxic intermediate which is then chemically converted into hydroquinone. This synthesis was accomplished by manipulation of the 24 shikimate pathway for the synthesis of quinic acid and the development of high-yielding methodologies for conversion of quinic acid into hydroquinone in purified fermentation broth. As seen from the syntheses of phenol and polyhydroxybenzenes, utilization of this strategy enables glucose to replace benzene as the starting material for synthesis of hydroquinone and also avoids the microbial toxicity of hydroquinone. Directed evolution Continuously expanding applications of enzymes and whole cells as biocatalysts for the chemical, pharmaceutical and food industries create a growing demand for enzymes that exhibit superior properties including higher operational stability, higher activities with unnatural substrates and higher enantioselectivity. Traditionally, enzymes used in industry were isolated for these purposes by the screening of microorganisms from culture collections or from extreme environments.” Over the past few years, rapidly increasing availability of entire genome-sequences together with development of bioinformatic tools have significantly broadened the possible range of enzymes that could be used in industry. However, the performance of wild-type enzymes that evolved over billions of years for highly specific and controlled reactions in the context of living organisms are not always suited for specific industrial applications. In this respect, directed evolution has emerged as a powerful tool for improving the characteristics of enzymes in a targeted manner. Previously, enzymes catalysis was modified using mutation and selection with approaches such as UV and chemical mutagenesis, mutator 73,74 strains and modeling-based point mutagenesis. Directed evolution combines random mutagenesis and/or in vitro recombination together with screening or selection for the 25 rapid evolution of individual proteins, biosynthetic pathways, and bacterial genomes.75 Contrary to traditional random and directed mutagenesis, which are slow processes in general, directed evolution can accelerate the improvement of desired properties. The first step in directed evolution is to generate genetic diversity starting from a single gene or a family of related genes by random mutagenesis or gene recombination, the resulting gene library is cloned back into an expression vector and transformed into a suitable microorganism for protein expression. Clones expressing improved proteins are identified by selection or screening, and those genes encoding improved proteins are isolated and used as parents for the next round of directed evolution. This process of improvement is repeated until the goal is achieved (Figure 14).76 Error-prone polymerase chain reaction (EP-PCR) is a commonly used method to generate random point mutations within a gene sequence.77 Error-prone PCR utilizes a low-fidelity polymerase such as Thermus aquaticus (T aq) polymerase in combination with reaction conditions such as adding Mn2+ and lowering annealing temperature which further decrease the fidelity of the polymerase. In contrast to the point mutations generated by error-prone PCR, gene recombination involves the recombination of multiple DNA sequences to create a library of chimeras. DNA shuffling, the first developed and still the most widely used technique for in vitro gene recombination, was reported by Stemmer in 1994.73 In this method, parental genes are first randomly fragmented with DNase I, and the purified fragments are then reassembled into full— length gene products by repeated cycles of overlap extension reactions. Recombinogenic events occur when fragments derived from different parental genes prime on each other. 26 Wild-type gene Homologous genes Random mutagenesis Recombination O O +— | '- Selection or 1 screening 0 . , Selection or - screening Random .. __ mutagenesis _—._ O O ‘ . ° 1 Recombination Selection or screening Selection or screening «+- -4—-4— Figure 14. Conventional approach and DNA shuffling for protein evolution. The end products of gene recombination are chimeric genes containing pieces of DNA from different parental genes.78 Following the pioneer work of Stemmer, new tools were developed that enhance existing methods of mutation and screening, such as single-strand DNA shuffling,79 random chimeragenesis on transient template (RACHITT),80 incremental truncation for the creation of hybrid enzymes (ITCHY),8' random priming 27 recombination, (RPR),82 staggered extension process (StEP),83 sequence homology- independent protein recombinantion (SHIPREC)84 and synthetic ligation assembly.85 The crucial point and also the most challenging step in the directed evolution is development of an efficient and sensitive method for the identification and isolation of desired mutants out of a library of 104-106 variants.76 This is performed either by selection or by screening using high-throughput technologies. Selection might be applicable86 in cases when the desired protein function is essential for the survival or growth of the host organism. This is often achieved by genetic complementation of host organisms that are deficient in a certain pathway or activity.” Using high-throughput selection, large libraries of enzyme variants can be assayed and the size of the library is only limited by the cell transformation efficiency. Unfortunately, devising a selection method for a given enzyme can be quite difficult. Screening is the process in which every library member is assayed individually using biochemical or biophysical analysis such as colorimetry, high-pressure liquid chromatography (HPLC) or gas chromatography-mass spectrometry (GC-MS). This is commonly done in either a 96- well or 384-well microtiter plate using a detectable signal developed on the basis of the targeted catalytic activity. These high-throughput screens can be done with whole cells, cell lysates, or partially purified enzymes. Screening often limits the size of library that can be assayed to 104 - 10“, even when robotic hardware is employed. On the other hand, screening assays are flexible since experimental conditions can be tailored to meet a specific requirement.78 Other screening methods were also developed for in vitro or in vivo screening or selection. These include compartmentalization and selection based on an in vitro selection system confined with a water-in-oil emulsion,88 phage display 28 coupled with substrate co-immobilization onto the phage surface, leading to modified indentifiable phages carrying the desired activity,89 and a three-hybrid system to detect desired enzymatic activity in vivo.90 Improving activity. Directed evolution is frequently used to improve enzyme fitness for a given industrial and biotechnology application. For example, Arnold and coworkers reported the use of random mutagenesis and saturation mutagenesis to improve the thermal stability of protease subtilisin E.91 Gray et al. improved the thermal stability of haloalkane dehalogenase from Rhodococcus rhodochrous for the hydrolysis of 1,2,3-trichloropropane to the corresponding alcohol 2,3-dichloropropanol, which can be further converted to epoxides, by saturation mutagenesis.92 A recent example is the generation of red fluorescent protein suitable for living cell imaging.93 The red fluorescent protein (DsRed) from Discosoma coral holds great potential for a spectrally distinct companion of the widely used green fluorescent protein (GFP) from Aequorea jellyfish as a reporter for gene expression and regulation. However, the potential of the red fluorescent protein to be a generally accepted tool is hampered by its slow chromophore maturation and obligate tetramerization. Although directed evolution has been applied successfully to green fluorescent protein to improve the fluorescent signal to 45-fold greater than commercially available GFP,94 engineering of the red fluorescent protein has proven to be difficult. Bevis et al.” used several rounds of directed and random mutagenesis with library sizes ranging from 103 to 105 variants to create red fluorescent protein variants that mature 15 times faster than the wild-type protein. Tsien et a1. further evolved this DsRed variant to be monomeric using a combination of structure-based design and directed evolution.93b 29 Altering stereoselectivity. DNA shuffling has also been applied successfully to changing the stereoselectivity of an enzyme. For instance, Arnold and coworkers improved the hydantoinase process for production of L-methionine in E. coli by inverting the enantioselectivity of the D-selective hydantoinase (40% enantiomeric excess) from Arthrobacter sp. DSM 9771 into an L-selective enzyme (20% enantiomeric excess) and increasing its total activity by five-fold.95 These engineered mutants were licensed to Degussa Fine Chemicals, which is currently developing a biocatalytic process for the synthesis of various L-amino acids.78 Reetz and coworkers96 reported the inversion of the stereoselectivity of a lipase from Pseudomonas aeruginosa by carrying out successive rounds of error-prone PCR combined with DNA shuffling. One mutant showed pronounced R-selectivity (E=30) compared with the wild-type enzyme that only showed modest S-selectivity. Modifying substrate specificity. Modifying an enzyme for new substrate recognition and regio- and stereospecificity has been a long-standing goal. In general, modifying the substrate selectivity of an enzyme is more difficult than improving the activity of an enzyme in a different environment due to the complex relationship between the enzyme structure and its substrate.97 There have been some impressive advances in this arena. For example, Yano et al.98 reported successful reengineering of an aspartate aminotransferase to a B—branched aminotransferase using random mutagenesis and selection for complementation of an engineered deficiency in the endogenous B—branched aminotransferase. After five cycles of mutation and selection with 10‘5-107 variants in each cycle, an aspartate aminotransferase variant was generated that utilized 2-oxo acids 105 times more 30 efficiently than the wild-type enzyme. Recent efforts on modifying the substrate selectivity by directed evolution include 2-hydroxybiphenyl 3-monooxygenase and biphenol dioxygenase.99 A particularly successful example is the directed evolution of an orthogal aminoacyl-tRNA synthase suitable for incorporating synthetic amino acids into proteins via an amber suppressor tRNA in vivo. Tyrosyl-tRNA synthetase from Methanococcus jannaschii was evolved to an efficient O-methyl-L-tyrosine tRNA synthetase.100 Based on those literature examples, it seems that a large library size and a powerful in vivo selection method were the key to the successful enzyme evolution. Chapter 3 of this thesis details the creation of a pyruvate-based shikimate pathway. The centerpiece of this successful creation lies on modifying substrate specificity of the dgoA-encoded 2-keto-3-deoxy-6-phosphogalactonate aldolase. Wild- type 2—keto-3-deoxy-6-phosphogalactonate aldolase catalyzes the reversible condensation of pyruvate and D-glyceradehyde 3-phosphate. After 4—5 rounds of PCR mutagenesis and DNA shuffling, the evolved 2-keto-3-deoxy-6-phosphogalactonate aldolase exhibits a 25- fold increase in Km/kcat towards the catalyzed condensation of pyruvate and D-erythrose 4- phosphate to form 3-deoxy-D-arabino-heptulosonic acid 7-phosphate. 31 Reference |http://www.eia.doe.gov/emeu/steo/pub/contents.html Short-term energy outlook- February 2004. 2 Paster, M.; Pellegrino, J. L.; Carole, T. M. Industrial Biopraducts: Today and Tomorrow; U. S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Office of the Biomass Program, 2003. From http://www.bioproducts- bioenergygov. 3 (a) Bozell, J. J .; Landucci, R. Alternate F eedstocks Program Technical and Economic Assessment; U. S. Department of Energy, Office of Industrial Technologies, 1993. (b) Lynd, L. R.; Cushman, J. H.; Nichols, R. J .; Wyman, C. E. Fuel ethanol from cellulose biomass. Science 1991, 251, 1318-1323. (c) Zaldivar, J.; Nielson, J .; Olsson, L. Fuel ethanol production from lignocellulose: a challenge for metabolic engineering and process integration. Appl. Microbial. Biotechnol. 2001, 56, 17-34. 4 Draths, K. M.; Knop, D. R.; Frost, K. M. Shikimic acid and quinic acid: replacing isolation from plant sources with recombinant biocatalysis. J. Am Chem. Soc. 1999, 121, 1603-1604. 5 The History of Plastics, http://inventors.about.com/libra_ry/inventors/blplastic.htm ‘5 Datta, R. In Kirk-0thmer Encyclopedia of Chemical Technology, 4‘11 ed.; Kroschwitz, J. 1.; Howe-Grant, M. Eds.; Wiley: New York, 1997; Vol. 13, p. 1042-1062. 7 Datta, R.; Tsai, S-H.; Bonsignore, P.; Moon, S-H.; Frank, J. R. Technological and economic potential of poly(lactic acid) and lactic acid derivatives. FEMS Microbial. Rev. 1995, 16, 221-231. 8 Garlotta, D. A literature review of poly(lactic acid). J. Polym. Environ. 2001, 9, 63-84. 9 New food service alternative from annually renewable resources unveiled for use in major avenues. Cargill Dow LLC press release, March 5, 2001. http://www.cargilldow.com. 10 Cargill Dow’s world-scale manufacturing facility comes on line. Cargill Dow LLC press release, April 4, 2002. http://www.cargilldow.com. ” Fahey, J. Forbes Magazine, November 26, 2001, 206. ‘2 Chang, J. Vertec biosolvents on verge of breaking out with new replacement applications, Chemical Market Reporter, September 2, 2002. 32 '3 Stinson, S. C. Fine and intermediate chemical makers emphasize new projects and processes. Chem. Eng. News. 1995, 17, 10-14. '4 http://www.dupont.com/biotech/resources ‘5 Fumagalli, C. In Kirk-0thmer Encyclopedia of Chemical Technology, 4th ed.; Kroschwitz, J. I.; Howe-Grant, M. Eds; Wiley: New York, 1997; Vol. 22, p. 1074-1102. ’6 Hort, E. V.; Taylor, P. In Kirk-0thmer Encyclopedia of Chemical Technology, 4th ed.; Kroschwitz, J. I.; Howe-Grant, M., Eds; Wiley: New York, 1997; Vol. 1, p. 210-211. ‘7 (a) Salm, H.; Eggeling, L.; de Graaf, A. A. Pathway analysis and metabolic engineering in Corynebacterium glutamicum. Biol. Chem. 2000, 381, 899-910. (b) Eggling, L.; Sahm, H. L-Glutamate and L-lysine: traditional products with impetuous developments. Appl. Microbial. Biotechnol. 1999, 52, 146-153. '8 McCann, J. E. Sweet Success: Haw NutraSweet Created a Billion Dollar Business; Business One Irwin: Homewood, IL, 1990. ‘9 De Boer, L.; Dijkhuizen, L. Microbial and enzymatic processes for L-phenylalanine production. Adv. Biachem. Eng./Biatechnal. 1990, 41, 1-27. 2° Grinter, N. J. Developing an L-phenylalanine process. Chemtech 1998, 7, 33-37. 2‘ Bongaerts, J .; Kramer, M.; Muller, U.; Raeven, L.; Wubbolts, M. Metabolic engineering for microbial production of aromatic amino acids and derived compounds. Metab. Eng. 2001, 3, 289-300. 22 Maiti, T. K.; Roy, A.; Mukherjee, S. K. Chatterjee, S. P. Microbial production of L- tyrosine: a review. Hindustan Antibiat. Bull. 1995, 37, 51-65. 23 (a) Murdock, D.; Ensley, B. D.; Serdar, C.; Thalen, M. Construction of metabolic operons catalyzing the de nova biosynthesis of indigo in Escherichia coli. Biotechnology 1993, 11, 381-386. (b) Berry, A.; Battist, S.; Chotani, G.; Dodge, T. C.; Peck, 8.; Power, 8.; Weyler, W. Biosynthesis of indigo using recombinant E. coli: development of a biological system for the cost-effective production of a large volume chemical. Proceedings-Biomass Conference Americas: Energy, Environment, Agriculture and Industry, 2““, Portland, Oreg. Aut. 21-24, 1995, p. 1121-1129. (c) Berry, A.; Dodge, T. C.; Pepsin, M.; Weyler, W. Application of metabolic engineering to improve both the production and use of biotech indigo. Indust. Microbial. Biotech. 2002, 28, 127-133. 24 Berry, A. Improving production of aromatic compounds in Escherichia coli by metabolic engineering. Trends Biotechnol. 1996, 14, 250-256. 33 7‘5 Katsumata, R.; Ikeda, M. Hyperproduction of tryptophan in Corynebacterium glutamicum by pathway engineering. Biotechnology 1993, I], 921-925. 26 Kurahashi, 0.; Tsuchida, T.; Kawashima, N.; Ei, H.; Yamane, K. L-Tryptophan production by transformed Bacillus subtilis. JP 61096990, 1984. 27 (a) Haslam, E. In Shikimic Acid: Metabolism and Metabolites; Wiley: New York, 1993. (b) Bentley, R. The shikimate pathway - a metabolic tree with many branches. Crit. Rev. Biochem. Mol. Biol. 1990, 25, 307-384. (c) Herrmann, K. M. In Amino Acids: Biosynthesis and Genetic Regulation; Herrmann, K. M., Somerville, R. L., Ed.; Addison- Wesley: Reading, 1983: p. 301. 2” Frost, J. W.; Bender, J. L.; Kadonaga, J. T.; Knowles, J. R. Dehydroquinate synthase from Escherichia coli: purification, cloning, and construction of overproducers of the enzyme. Biochemistry 1984, 23, 4470-4475. 29 Carpenter, E. P.; Hawkins, A. R.; Frost, J. W.; Brown, K. A. Structure of dehydroquinate synthase reveals an active site capable of multistep catalysis. Nature 1998, 394, 299-302. 3° (a) Smith, B. W.; Turner, M. J.; Haslam, E. Shikimate pathway. 4. Stereochemistry of 3-dehydroquinate dehydratase reaction and observations on 3-dehydroquinate synthetase. J. Chem. Sac., Perkins Trans. 1 1975, I, 52-55. (b) Duncan, K.; Chaudhuri, S.; Campbell, M. S.; Coggins, J. R. The overexpression and complete amino-acid-sequence of Escherichia coli 3-dehydroquinase. Biochem. J. 1986, 238, 475-483. 3‘ (a) Anton, I. A.; Coggins, J. R. Sequencing and overexpression of the Escherichia coli aroE gene encoding shikimate dehydrogenase. Biochem. J. 1988, 249, 319-326. (b) Chaudhuri, S.; Coggins, J. R. The purification of shikimate dehydrogenase from Escherichia coli. Biochem. J. 1985, 226, 217-223. ’2 Benach, J.; Lee, 1.; Edstrom, W.; Kuzin, A. P.; Chiang, Y. W.; Acton, T. B.; Montelione, G. T.; Hunt, J. F. The 2.3-angstrom crystal structure of the shikimate 5- dehydrogenase orthologue YdiB from Escherichia coli suggests a novel catalytic environment for an NAD-dependent dehydrogenase. J. Biol. Chem. 2003, 278, 19176- 19182. 33 (a) DeFeyter, R. C.; Pittard, J. Genetic and molecular analysis of AroL, the gene for Shikimate kinase-II in Escherichia coli K-12. J. Bacteriol. 1986, 165, 226-232. (b) DeFeyter, R. C.; Pittard, J. Purification and properties of shikimate kinase-II from Escherichia coli K-l2. J. Bacterial. 1986, 165, 331-333. 3" Laner-Olesen, A.; Marinus, M. G. Identification of the gene (Arok) encoding sh ikimic acid kinase-I of Escherichia coli. J. Bacterial. 1992, I 74, 525-529. 34 35 Duncan, K.; Coggins, J. R. The Serc-AroA operon of Escherichia coli: a mixed function operon encoding enzymes from 2 different amino-acid biosynthetic pathways. Biochem. J. 1986, 234, 49-57. (b) Duncan, K.; Lewendon, A.; Coggins. J. R. The purification of 5-enolpyruvylshikimate 3-phosphate synthase from an overproducing strain of Escherichia coli. FEBS Lett. 1984, 165, 121-127. 36 White, P. J .; Millar, G.; Coggins, J. The overexpression, purification and complete amino-acid sequence of chorismate synthase from Escherichia coli K12 and its comparison with the enzyme from Neurospora crassa. Biochem. J. 1988, 251, 313-322. ’7 Pittard, A. J. Biosynthesis of the aromatic amino acids. In Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. Neidhardt, F. C. ed. ASM Press (Washington, DC: American Society for Microbiology) 1996, p. 458-484. 3” Kishore, G. M.; Shah, D. M. Amino acid biosynthesis inhibitors as herbicides. Annu. Rev. Biochem. 1988, 57, 627-663. 39 Roberts, R; Roberts, C. W.; Johnson, J. J.; Kyle, D. E.; Krell, T.; Coggins, J. R.; Coombs, G. H.; Milhous, W. K.; Tzipori, S.; Ferguson, D. J. P.; Chakrabarti, D.; McLeod, R. Evidence for the shikimate pathway in apicomplexan parasites. Nature 1998, 393, 801-805. 4° Draths, K. M.; Frost, J. W. Environmentally compatible synthesis of catechol from D- glucose. J. Am. Chem. Soc. 1995, 117, 2395-2400. 4‘ Li, K.; Frost, J. W. Synthesis of vanillin from glucose. J. Am. Chem. Soc. 1998, 120, 10545-10546. ‘2 Draths, K. M.; Ward, T. L.; Frost, J. W. Biocatalysis and nineteenth century organic chemistry: conversion of D-glucose into quinoid organics. J. Am. Chem. Soc. 1992, 114, 9725-9726. ‘3 Kambourakis, S.; Draths, K. M.; Frost, J. W. Synthesis of gallic acid and pyrogallol from glucose: replacing natural product isolation with microbial catalysis. J. Am. Chem. Soc. 2000, 122, 9042-9043. 4" Barker, J. L.; Frost, J. W. Microbial synthesis of p-hydroxybenzoic acid from glucose. Biotechnol. Bioeng. 2001, 76, 376-390. 45 Gibson, J. M.; Thomas, P. S.; Thomas, J. D.; Barker, J. L.; Chandran, S. S.; Harrup, M. K. ; Draths, K. M.; Frost, J. W. Benzene-free synthesis of phenol. Angew. Chem, Intl. Ed. 2001, 40, 1945-1948. 35 46 Li, K.; Mikola, M. R.; Draths, K. M.; Worden, R. M.; Frost, J. W. Fed-batch fermentor synthesis of 3-dehydroshikimic acid using recombinant Escherichia coli. Biotechnol. Bioeng. 1999, 64, 61-73. ’7 (a) Draths, K. M.; Frost, J. W. Environmentally compatible synthesis of adipic acid from D-glucose. J. Am. Chem. Soc. 1994, 116, 399-400. (b) Niu, W.; Draths, K. M.; Frost, J. W. Benzene-free synthesis of adipic acid. Biatechnol. Prag. 2002, 18, 201-211. 4" Draths, K. M.; Pompliano, D. L.; Conley, D. L.; Frost, J. W.; Berry, A.; Disbrow, G. L.; Staversky, R. J .; Lievense, J. C. Biocatalytic synthesis of aromatics from D-glucose: the role of transketolase. J. Am. Chem. Soc. 1992, 114, 3956-3962. ‘9 Yi, J .; Li, K.; Draths, K. M.; Frost, J. W. Modulation of phosphoenolpyruvate synthase expression increases shikimate pathway product yields in E. coli. Biatechnal. Prag. 2002, 18,1141-1148. 5° Krumenacker, L.; Costantini, M.; Pontal, P.; Sentenac, J. In Kirt-Othmer Encyclopedia of Chemical Technology, 4‘h ed.; Kroschwitz, J. I., Howe-Grant, M., Eds; Wiley, New York, 1995, Vol 13, p. 996. 5' (a) Franck, H.-G.; Stadelhofer, J. W. Industrial Aromatic Chemistry; Springer-Verlag: New York, 1988; p. 183-190. (b) Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York: Wiley, 1989, p. 512-519. 52 Neidle, E. L.; Ornston, L. N. Cloning and expression of Acinetobacter calcaaceticus catechol 1,2-dioxygenase structural gene catA in Escherichia coli. J. Bacterial. 1986, 168, 815-820. 53 Davis, D. D.; Kemp, D. R. In Kirk-0thmer Encyclopedia of Chemical Technology, 4m ed.; Kroschwitz, J. I.; Howe-Grant, M., Eds; Wiley: New York, 1997; Vol. 1, p. 466- 493. 54 (a) Thiemens, M. H.; Trogler, W. C. Nylon production — an unknown source of atmospheric nitrous oxide. Science 1991, 251, 932. (b) Dickinson, R. E.; Cicerone, R. J. Future global warming from atmosphere trace gases. Nature 1986, 319, 109-115. 55 Esposito, L.; Forrnanek, K.; Kientz, G.; Mauger, F.; Maureux, V.; Robert, G.; T ruchet, F. In Kirk-0thmer Encyclopedia of Chemical Technology, 4‘h ed.; Kroschwitz, J. I., Howe-Grant, M., Eds; Wiley: New York, 1997; Vol. 24, p. 812. 56 Clark, G. s. Vanillin. Perf. Flavor. 1990, 15, 45-54. 57 Priefert, H.; Rabenhorst, J .; Steinbuchel, A. Biotechnological production of vanillin. Appl. Microbial. Biotechnol. 2001, 56, 296-314. 36 58 (a) Gross, G. G.; Bolkart, K. H.; Zenk, M. H. Reduction of cinnamic acid to cinnamaldehyde and alcohol. Biochem. Biophys. Res. Commun. 1968, 32, 173-178. (b) Gross, G. G.; Zenk, M. H. Reduction of aromatic acids to aldehydes and alcohols in a cell-free system. 1. Purification and properties of aryl aldehyde: NADP oxidoreductase from Neurospora crassa. Eur. J. Biochem. 1969, 8, 413-419. (c) Gross, G. G. Formation and reduction of intermediate acyladenylate by aryl-aldehyde. NADP oxidoreductase from Neurospora crassa. Eur. J. Biochem. 1972, 11, 585-592. (d) Zenk, M. H.; Gross, G. G. Enzymatic reduction of cinnamic acids. Recent Adv. Phytachem. 1972, 4, 87-106. “’9 Kirsch, M. A.; Williams, D. J. Understanding the thermoplastic polyester business. Chemtech 1994, 24 , 40-49. 6° Szmant, H. H. In Organic Building Blocks of the Chemical Industry, New York: Wiley, l989,p.467. 6‘ Lenga, R. E.; Votoupal, K. L. The Sigma-Aldrich Library of Regulatory and Safety Data; Sigma-Aldrich: Milwaukee, WI, 1993. 62 Siebert, M.; Berchthold, A.; Melzer, M.; May, U. Berger, U. Ubiquinone biosynthesis - cloning of the genes coding for chorismate pyruvate-lyase and 4-hydroxybenzoate octaprenyl transferase from Escherichia coli. F EBS Lett. 1992, 307, 347-350. 63 Leston, G. In Kirk-Othmer Encyclopedia of Chemical Technology; 4‘h ed.; Kroschwitz, J. I., Howe-Grand, M., Eds; New York, 1997; Vol. 19, p. 778. 6‘ Kambourakis, S.; Frost, J. W. Synthesis of gallic acid: Cu2’-mediated oxidation of 3- dehydroshikirrric acid. J. Org. Chem. 2000, 6904-6909. 65 (a) Entsch, B.; Palfey, B. A.; Ballou, D. P.; Massey, V. Catalytic function of tyrosine residues in para-hydroxybenzoate hydroxylase as determined by the study of site-directed mutants. J. Biol. Chem. 1991, 266, 17341-17349. (b) Eschrich, K.; van der Bolt, F. J. T.; de Kok, A.; van Berkel, W. J. H. Role of Tyr201 and Tyr385 in substrate activation by p- hydroxybenzoate hydroxylase from Pseudomonas fluorescens. Eur. J. Biochem. 1993, 216, 137-146. 66 Wallace, J. In Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed., Vol. 18; Kroschwitz, J. I.; Howe-Grant, M., eds; Wiley: New York, 1998, p. 592-602. 67 Szmant, H. H. In Organic Building Blocks of the Chemical Industry; Wiley: New York, 1989, p. 434. 68 Chandran, S. C.; Yi, J.;, Draths, K. M.; von Daeniken, R.; Weber, W.; Frost, J. W. Phosphoenolpyruvate availability and the biosynthesis of shikimic acid. Biatechnal. Prog. 2003, 19, 808-814. 37 69 (a) Federspiel, M.; Fischer, R.; Henning, M.; Mair, H.-J.; Oberhauser, T.; Rimmler, G.; Albiez, T.; Bruhin, J .; Estermann, H.; Gandert, C.; Gockel, V.; Gotzo, S.; Hoffmann, U.; Huber, G.; Janatsch, G.; Lauper, S.; Odette, T.-S.; Trussardi, R.; Zwahlen, A. G. Industrial synthesis of the key precursor in the synthesis of the anti-influenza drug Oseltamivir phosphate (Ro 64-0796/002, GS-4104-02): ethyl (3R,4S,55)-4,5-epoxy-3-(1- ethyl-propoxy)-cyclohex-l-ene-l-carboxylate. Org. Process Res. Dev. 1999, 3, 266-274. (b) Kim, C. U.; Lew, W.; Williams, M. A.; Liu, H.; Zhang, L.; Swaminathan, S.; Bischofberger, N.; Chen, M. 8.; Mendel, D. B.; Tai, C. Y.; Laver, W. G.; Stevens, R. C. Influenza neuraminidase inhibitors possessing a novel hydrophobic interaction in the enzyme active site: design, synthesis, and structural analysis of carbocyclic sialic acid analogues with potent anti-influenza activity. J. Am. Chem. Soc. 1997, 119, 681-690. (c) Rohloff, J. C.; Kent, K. M.; Postich, M. J.; Becker, M. W.; Chapman, H. H.; Kelly, D. E.; Lew, W.; Louie, M. S.; McGee, L. R.; Prisbe, E. J .; Schultze, L. M.; Yu, R. H.; Zhang, L. Practical total synthesis of the anti-influenza drug GS-4104. J. Org. Chem. 1998, 63, 4545-4550. 7° Kakinuma, K.; Nango, E.; Kudo, F.; Matsushima, Y.; Eguchi, T. An expeditious chemo-enzymatic route from glucose to catechol by the use of 2-deoxy-scylla-inosose synthase. Tetrahedron Lett. 2000, 41, 1935-1938. 71 Hansen, C.; Frost, J. W. Deoxygenation of polyhydroxybenzene: an alternative strategy for the benzene-free synthesis of aromatic chemicals. J. Am. Chem. Soc. 2001, 124, 5926- 5927. 72 (a) Cheetham, P. S. J. Screening for novel biocatlysts. Enzyme Micrab. Technol. 1987, 9, 194-213. (b) Ogawa, J .; Shimizu, S. Microbial enzymes: new industrial applications from traditional screening methods. Trends Biotechnol. 1999, 17, 13-20. 73 (a) Stemmer, W. P. C. DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution. Proc. Natl. Acad. Sci. USA 1994, 91, 10747- 10751. (b) Stemmer, W. P. C. Rapid evolution of a protein in vitro by DNA shuffling. Nature 1994, 370, 389-391. 7" Crameri, A.; Raillard, S-A.; Bermudez, E.; Stemmer, W. P. C. DNA shuffling of a family of genes from diverse species accelerates directed evolution. Nature 1998, 391, 288-291. 75 Zhang, Y-X.; Perry, K.; Vinci, V. A.; Powell, K.; Stemmer, W. P. C.; del Cardayre, S. B. Genome shuffling leads to rapid phenotypic improvement in bacteria. Nature 2002, 415, 644-646. 76 Schmidt-Dannert, C. Directed evolution of single proteins, metabolic pathways and Viruses. Biochemistry 2001, 44, 13125-13136. 38 77 Cadwell, R. C.; Joyce, G. F. Randomization of genes by PCR mutagenesis. PCR Methods Appl. 1992, 2, 28-33. 73 Woodyer, R.; Chen, W.; Zhao, H. Outrunning nature: directed evolution of superior biocatalysts. J. Chem. Edu. 2004, 81, 126-133. 79 Zha, W.; Zhu, T.; Zhao, H. Family shuffling with single-stranded DNA. In Methods in Molecular Biology (Totowa, NJ, United States) 2003, 231(Directed Evolution Library Creation), 91-97. 8° Coco, W. M.; Levinson, W. E.; Crist, M. J .; Hektor, H. J .; Darzins, A.; Pienkos, P. T.; Squires, C. H.; Monticello, D. J. DNA shuffling method for generating highly recombined genes and evolved enzymes. Nat. Biotechnol. 2001, 19, 354-359. 8' Ostermeier, M.; Shim, J. H.; Benkovic, S. J. A combinatorial approach to hybrid enzymes independent of DNA homology. Nat. Biotechnol. 1999, 17, 1205-1209. 32 Shao, Z.; Zhao, H.; Giver, L.; Arnold, F. H. Random-priming in vitro recombination: an effective tool for directed evolution. Nucleic Acids Res. 1998, 26, 681-683. 8’ Zhao, H.; Giver, L.; Shao, Z.; Affholter, J. A.; Arnold, F. H. Molecular evolution by staggered extension process (StEP) in vitro recombination. Nat. Biotechnol. 1998, 16, 258-261. 8" Sieber, V.; Martinez C. A.; Arnold F. H. Libraries of hybrid proteins from distantly related sequences. Nat. Biotechnol. 2001, 19, 456-60. 85 Short, J. M. Diversa Corporation, Synthetic ligation assembly in directed evolution. US Patent 6,605,449 B l, 2003. 8" Fastrez, J. In viva versus in vitro screening or selection for catalytic activity in enzymes and abzymes. Mal. Biotechnol. 1997, 7, 37-55. ’7 Naki, D.; Paech, C., Ganshaw, G., Schellenberger, V. Selection of a subtilisin- hyperproducing Bacillus in a highly structured environment. Appl. Environ. Microbial. 1998, 49, 290-294. 88 Tawfik, D. S.; Griffins, A. D. Man-made cell-like compartments for molecular evolution. Nat. Biotechnol. 1998, 16, 652-656. 89 (a) Pedersen, H.; Holder, S.; Sutherlin, D. P.; Schwitter, U.; King, D. S.; Schultz, P. G. A method for directed evolution and functional cloning of enzymes. Proc. Natl. Acad. Sci. USA 1998, 95, 10523-10528. (b) Demartis, S.; Huber, A.; Viti, F.; Lozzi, L.; Giovannoni, L.; Neri, P.; Winter, G.; Neri, D. A strategy for the isolation of catalytic activities from repertoires of enzymes displayed on phage. J. Mol. Biol. 1999, 286, 617- 39 633. 9° Firestine, S. M.; Salinas, F.; Nixon, A. E.; Baker, S. J .; Benkovic, S. J. Using an AraC- based three-hybrid system to detect biocatalysts in viva. Nat. Biotechnol. 2000, 18, 544- 547. 9' Miyazaki, K.; Arnold, F. H. Exploring nonnatural evolutionary pathway by saturation mutagenesis: rapid improvement of protein function. J. Mol. Eval. 1999, 49, 716-720. 92 Gray, K. A.; Richardson, T. H.; Kretz, K.; Short, J. M.; Bartnek, F.; Knowles, R.; Kan, L.; Swanson, P. E.; Robertson, D. E. Rapid evolution of reversible denaturation and elevated melting temperature in a microbial haloalkane dehalogenase. Adv. Synth. Catal. 2001, 343, 607-617. 93 (a) Bevis, B. J.; Glick, B. S. Rapidly maturing variants of the Discosoma red fluorescent protein (DsRed). Nat. Biotechnol. 2002, 20, 83-87. (b) Campbell, R. E.; Tour, 0.; Palmer, A. E.; Steinbach, P. A.; Baird, G. S.; Zacharias, D. A.; Tsien, R. Y. A monomeric red fluorescent protein. Proc. Natl. Acad. Sci. USA 2002, 99, 7877-7882. 9" Crameri, A.; Whitehom, E. A.; Tate, E.; Stemmer, W. P. C. Improved green fluorescent protein by molecular evolution using DNA shuffling. Nat. Biotechnol. 1996, 14, 315-319. 95 May, 0.; Nguyen, P. T.; Arnold, F. H. Inverting enantioselectivity by directed evolution of hydantoinase for improved production of L-methionine. Nat. Biotechnol. 2000, 18, 317-320. 9" Zha, D.; Wilensek, S.; Hermes, M.; Jaeger, K. A.; Reetz, M. T. Complete reversal of enantioselectivity of an enzyme-catalyzed reaction by directed devolution. Chem. Commun. 2001, 2664-2665. 97 Tao, H.; Cornish, V. W. Milestones in directed enzyme evolution. Curr. Opin. Chem. Biol. 2002, 6, 858-864. 98 Yano, T.; Oue, S.; Kagamiyama, H. Directed evolution of an aspartate aminotransferase with new substrate specificities. Proc. Natl. Acad. Sci. USA 1998, 95, 5511-5515. 99 (a) Meyer, A.; Schmid, A.; Held, M.; Westphal, A. H.; Rothlisberger, M.; Kohler, H. P.; van Berkel, W. J.; Witholt, B. Changing the substrate reactivity of 2-hydroxybiphenyl 3-monooxygenase from Pseudomonas azeleica HBPl by directed evolution. J. Biol. Chem. 2002, 277, 5575-5582. (b) Barriault, D.; Plante, M. M.; Sylvestre, M. Family shuffling of a targeted bphA region to engineer biphenyl dioxygenase. J. Bacterial. 2002, 184, 3794-3800. 40 ‘00 Wang, L.; Brock, A.; Herberich, B.; Schultz, P. G. Expanding the genetic code of Escherichia coli. Science 2001, 292, 498-500. 41 SHAPE—MED Benzene-Free Synthesis of Hydroquinone Introduction Hydroquinone is a pseudocommodity chemical used in photographic developer and in the synthesis of polymerization inhibitors and rubber antioxidants.l Global demand for hydroquinone was estimated at 42 x 106 kg/year in 1997, with a projected annual growth rate of 3-5%.2 Previously synthesized by a route employing stoichiometric amounts of MnO2 to oxidize aniline, manufacture of hydroquinone is now dominated by Hock oxidation of 1,4-diisopropylbenzene and peroxide hydroxylation of phenol by homogeneous or heterogeneous catalysis (Figure 15). In addition to chemical synthesis, microbial hydroxylation of benzene and phenol to hydroquinone has also been achieved in low yield. All these syntheses have the common feature that volatile and carcinogenic benzene and its derivatives are the starting material. Synthesis of hydroquinone via intermediacy of aniline and nitrobenzene generates large quantities of salt stream including MnSO,, (NH4)2SO4, and iron oxide salts.l Both Hock-type and peroxide oxidations constitute improvements in the synthesis of hydroquinone by virtue of reducing the number of required synthetic steps and eliminating byproduct salt streams. Manufacture of hydroquinone by Hock oxidation of 1,4-diisopropylbenzene and peroxide hydroxylation of phenol (Figure 15) accounts for about 60% and 30% of global hydroquinone production, respectively.‘ The 1,4- diisopropylbenzene is synthesized by Freidel-Crafts reaction of benzene or cumene with prOpylene or 2-propanol. Catalyzed air oxidation of 1,4-diisopropylbenzene produces 42 dihydrOperoxide, hydroxyhydroperoxide, and dicarbinol products. The hydroxyhydroperoxide and dicarbinol are converted into the dihydroperoxide upon treatment with H202. Acid-catalyzed cleavage of the dihydroperoxide produces acetone and hydroquinone. One drawback in this route is the formation of explosive acetone hydroperoxides during the acid-catalyzed cleavage reactions (Figure 15).l Reaction of phenol with H202 in the presence of acid catalysts leads to a mixture of hydroquinone and catechol (Figure 8, Chapter 1). The acid catalyst employed such as formic, sulfuric, trifluoromethanesulfonic acids and synthetic zeolites during reaction of phenol with H202 significantly controls the ratio of hydroquinone to catechol.l OH ’ I: ———> a / cumene phenol O lg a OH benzene \ b ——-> OH 0 . 1 4_ -diisopropyl hydroqumone benzene 1f N02 NH2 8 —- e ——-» o nitrobenzene aniline benzoquinone Figure 15. Chemical synthesis of hydroquinone from benzene. Keys: (a) 2-propene, HZSM-12; (b) i) 02, NaOH, ii) HZSO4; (c) HNO3, H2804; (d) Cu/SiOz, H2; (e) MnOZ, H2S04; (f) Fe°; (g) HCOZH, HCO3H. 43 Similar to the chemical synthesis of hydroquinone from benzene and phenol, microbe-catalyzed routes to hydroquinone have included the synthesis from phenol by butane-catabolizing bacterium Mycobacterium sp. HB50, and the conversion from benzene by the methane-catabolizing Methylasinus trichasparium OB3b (Figure 16).3 Phenol, hydroquinone, and benzoquinone are all toxic toward rrricrobes.3a To reduce the toxicity of hydroquinone toward the microbial catalyst, a continuous reaction system was developed for Mycobacterium sp. HBSO allowing phenol to be converted into hydroquinone with a specific volumetric productivity of 2-3 g/L/h. Methylasinus trichasparium OB3b synthesized approximately 1 g/L of hydroquinone from benzene with use of a chemostat to control reaction conditions. The major advantage of using Mycobacterium sp. HB50 was its greater resilience toward hydroquinone and benzoquinone toxicity relative to Methylasinus trichasparium.3n OH OH © Mycobacterium sp. H850 phenol OH hydroquinone OH (2 Methylasinus trichasparium OB3b_ benzene OH hydroquinone Figure 16. Microbial routes to hydroquinone from phenol and benzene. Conversion of phenol to hydroquinone catalyzed by Mycobacterium sp. HB50 and conversion of benzene to hydroquinone catalyzed by Methylasinus trichasparium 0B3b share with the aforementioned (Figure 15) chemical routes the feature of directly or indirectly using fossil fuel-derived benzene as the starting material. A biocatalytic synthesis of hydroquinone from glucose would provide a fundamentally different access to the aromatic molecule. The use of carcinogenic benzene and its derivatives would be completely avoided in this route. In theory, glucose can be converted into hydroquinone via a biocatalytic route involving intermediacy of p-hydroxybenzoic acid (Figure 17). Glucose can be converted into p-hydroxybenzoic acid at a concentration of 12 g/L in 13% (mol/mol) yield using recombinant Escherichia coli constructs" Conversion of p- hydroxybenzoic acid into hydroquinone would be catalyzed by p-hydroxybenzoate 1- hydroxylase, an enzyme found in Candida parapsilasis.5 To construct a single microbe capable of catalyzing the conversion of glucose into hydroquinone, the C. parapsilasis gene encoding p-hydroxybenzoate l-hydroxylase would likely have to be isolated and then expressed in a p-hydroxybenzoate-synthesizing microbial host. OH . COZH OH E. 00/! p-hydroxybenzoate ‘ JB1 61 /pJ 82.274 1 -hydroxy|ase i OH 7 / \ 7 HO OH OH NADPH NADPI OH D-glucose p-hydroxybenzoic 02 hydroquinone acid Figure 17. Proposed biocatalytic synthesis of hydroquinone from D-glucose. 45 The proposed microbial synthesis of hydroquinone from glucose via intermediacy of p-hydroxybenzoic acid, although avoiding use of benzene as starting material, still must contend with the microbial toxicity of product hydroquinone. To circumvent the toxicity of the aromatic product, a two-step synthesis that involves biocatalytic conversion of glucose to quinic acid followed by chemical conversion of quinic acid to 6 constructed E. coli hydroquinone was examined. Draths et a1. A32848/pKD136/pTW809OA, which synthesized 4.8 g/L of quinic acid in 31% yield from glucose by expression of the qad—encoded quinate dehydrogenase from Klebsiella pneumoniae. Oxidation of the biosynthesized quinic acid with 100 equivalents of MnO2 afforded 10% yield of hydroquinone upon heating at 100°C for 18 h.6 Microbial synthesis of quinic acid by the heterologous E. coli strain provided a relatively inexpensive source of the hydroaromatic for conversion to the pseudocommodity chemical hydroquinone. The key distinction between the chemical oxidation of microbe- synthesized quinic acid as a route to hydroquinone relative to all reported or theoretical biocatalytic syntheses of hydroquinone is that the microbial catalyst never comes into contact with a toxic aromatic starting material, aromatic intermediate or aromatic product. Quinic acid at neutral pH does not adversely affect either growth or metabolism of E. coli. Ho. cozH 0” a b ———> ————-—> Ho“ é OH OH OH D-glucose quinic acid hydroquinone Figure 18. Previously examined synthesis of hydroquinone from D-glucose. Keys: (a) E. coli AB3248/pKD136/pTW 8090A, 31%; (b) MnO2 (100 equiv.), 100 °C, 10%. 46 In order to be cost-competitive with the commercial chemical process for hydroquinone, both the microbial synthesis of quinic acid and the chemical conversion of quinic acid to hydroquinone need to be high yielding and low cost. In the reported biocatalytic synthesis of hydroquinone, the E. coli AB3248/pKD136/pTW8090A was cultured under shake-flask conditions. It is generally difficult to control oxygenation and pH during growth of microbes in shake flasks. Microbes cultured under shake-flask conditions are also under physiological stress due to the glucose-rich environment present immediately after inoculation that changes to a glucose-limited environment as the cultures grow. The two-plasmid system and the heterologous expression of quinate dehydrogenase from K. pneumoniae also hindered cultivation of the biocatalyst under fed-batch fermentation conditions. The finding that E. coli aroE-encoded shikimate dehydrogenase also catalyzes the reduction of 3-dehydroquinate to quinic acid opens a new avenue to biosynthesis of quinic acid with a homogeneous E. coli construct.7 The first part of this chapter will discuss the construction and strategies to improve microbial synthesis of quinic acid. The reported chemical conversion of quinic acid to hydroquinone utilized manganese dioxide, which has the same environmental problem associated with the commercial route to hydroquinone where stoichiometric amounts of MnO2 are used to oxidize aniline. To reduce the environmental pollution and improve the yield of this conversion, chemical conversion of quinic acid into hydroquinone requires the development of high-yielding reactions using inexpensive reagents. Due to the potentially high cost of isolating quinic acid from fermentation broth. the developed methodology should be appropriate for use in aqueous medium. The second part of the 47 chapter will deliberate the high-yielding chemical conversions developed for synthesis of hydroquinone from quinic acid. Microbial synthesis of quinic acid from D-glucose. (-)-Quinic acid is found in a variety of plant materials ranging from Cinchona bark to tobacco leaves to cranberries.8 Its isolation from crude quinine (an anti-malaria drug isolated from Cinchona bark) dated back to 1790,9 although its structure and stereochemistry were assigned more than a hundred years later by Fisher and Dangschzt.10 As a chiral synthon with its highly-functionalized, six-membered carbocyclic ring and multiple asymmetric centers, quinic acid has garnered attention for combinatorial synthesis'1 and natural products syntheses. Numerous biologically active molecules have been constructed in whole, or in part, from quinic acid. The diversity of these molecules encompasses the anti-influenza drug GS4104,12 (-)-sugiresinol dimethyl ether, a derivative of (-)-sugiresinol isolated from Cryptomeria japonica,13 the epoxycyclohexenone core of scyphostatin, a powerful inhibitor of neutral sphongomyelinase,” (+)-eutypoxide B, a secondary metabolite of fungus Eutypa lata responsible for pathogenic vineyard die-back disease,15 the A ring of 1a ,25 - dihydroxyvitamin D3 derivatives as potential drugs for treatment of osteoporosis and psoriasis,16 the 2-iodocyclohexenone acetal portion of anticancer drug taxol,l7 and the bicyclic core structure of the potent enediyne antitumoral agent esperimicin-Al18 (Figure 19). Quinic acid has also been used to prepare a chiral ketone for generation of chiral dioxirane used in asymmetric epoxidation of prochiral olefins (Figure 19).19 48 002Et fij O O 0“ "NH -H PO 2 3 4 O NHAc "0H GS-41 O4 (-)-sugiresinol dimethyl ether (+)-eutypoxide B o EOQ‘OTBS 1 a,25-dihydroxy-19-norvitamin D3 2-lodocyclohexenone acetal 01— TBSO H O H300zC:. O “‘2“; (I t ‘ O 0‘ .- \~“ I l O H3CO bicyclic core of esperamicin-A1 chiral ketone for asymmetric epoxidation Figure 19. Selected molecules synthesized from quinic acid. The portions from quinic acid are indicated in bold. 49 Commercially available quinic acid is isolated from Cinchona bark.20 Two stereospecific syntheses of quinic acid from D-arabinose‘2| and ketol silyl ether 1,22 respectively have been reported in the literature. D-Arabinose was reduced by catalytic hydrogenation to D-arabitol followed by several protection steps to 1,5-0-ditrityl-2,3,4- O-benzyl-D-arabitol, which was then converted S-methylene-l,2,3-cyclohexanetriol. Oxidative cleavage of the methylene compound by OsO.,/NaIO4 afforded a cyclohexanone compound, which was further converted to a cyanohydrin followed by hydrolysis and de-protection of the acetyl groups to yield quinic acid (Figure 20).2| 0 OH 9H OBz QBz “Ono” 3, HOMOH __b_,'rso\/'\_/'\,0Ts Ho“ OH 6H 632 D-arabinose D-arabitol 1,5-0-ditrityl-2,3,4-0- benzyl-D-arabitol CH2 CH2 0 ° (1 d e b. ——> ——> .__.> 320“. OBz HO“. ; OH AcO‘V OAc OBz 6H 6Ac 5-methylene-1 ,2,3- 1 ,2,3-O-triacetyl- cyclohexanetriol cyclohexanone HO COZH H04 CN ,. L» CL —9» (it Aco“' OAc Ho“' . OH OAc OH triacetquuinic acid nitrile quinic acid Figure 20. Chemical synthesis of quinic acid from D-arabinose. Keys: (a) Ni, H2, 96%; (b) i) triphenylmethyl chloride, pyridine, 75%; ii) benzyl chloride, KOH, 80%; iii) 70% AcOH, 69%; iv) p-TsCl, pyridine, 76%; (c) i) methylenetriphenylphosphorane, ii) CH20, 82%; (d) Na, liquid NH3, 80%; (e) i) ACZO, pyridine, 82%, ii) cat. 0504, NaIO4, 92%; f) HCN, 86%; (g) i) A020, pyridine, 65%; ii) HBr/ACOH; iii) H20, 78%; iv) NaOH, H20, 95%. 50 Synthesis of quinic acid from ketol silyl ether I proceeded through a hydroxylation reaction followed by a retro-Diels-Alder reaction to give the cyclohexenone 3. Dihydroxylation of 3 by OsO4lNMO afforded a cis-diol as major product. Reduction of the ketone 4 by N aBH4 followed by deoxygenation reaction to give fully protected quinic acid. Removal of the three alcohol protecting groups by refluxing with CBr4 and hydrolysis of the methyl ester with NaOH afforded quinic acid (Figure 21). , OTDBMSb OQAC —-> . COch3—> - OH O O COzCHa TBDMSO 1 2 3 AcO, COZCH3 HO, COzH Im ' ><". O'oocha _ L-phenylalanine H203Po“ OH L-tryptophan shikimate 3- -phosphate Figure 22. The truncated aromatic amino acid biosynthetic pathway with quinic acid biosynthesis. Intermediates (abbreviations): phosphoenolpyruvate (PEP), D- erythrose 4-phosphate (E4P), 3-deoxy-D-arabin0-heptulosonic acid 7-phosphate (DAHP). Genes (enzymes): aroF, aroG, aroH, DAHP synthase; aroB, 3-dehydroquinate synthase; aroD, 3-dehydroquinate dehydratase; aroE, shikimate dehydrogenase; aroK, aroL, shikimate kinase. 53 Along with serA and the aforementioned aroFFBR, aroE and tktA were included on plasmid pKD12.138 (Figure 23).26 Overexpression of aroE-encoded shikimate dehydrogenase reduces 3-dehydroquinic acid into quinic acid. Utilization of E. coli aroE-encoded shikimate dehydrogenase instead of K. pneumoniae qad-encoded quinate dehydrogenase avoided the common difficulty associated with heterologous expression. This meant that promoter compatibility and codon usage were completely avoided as factors requiring consideration in route to achieving adequate overexpression of the enzyme that reduced 3-dehydroquinic acid to quinic acid. The construction of plasmid pKD12.138 began with PCR amplification of a 1.2-kb Pumb—aroE fragment from pIA321.27 The aroE with its own promoter PM; was under the transcriptional control of a tac promoter in pIA321. Cloning the PumEaroE into pKL4.ZOBZ4“‘ which already bears an aroFFER locus on a pSU18 vector afforded the 4.8-kb plasmid pKD12.036A. The orientation of the PumEaroE locus is in the opposite direction as that of aroFFBR. The 1.9-kb serA locus was excised from pD262528 and subsequently cloned into pKD12.036A to afford pKD12.047A. The orienation of the serA locus is in the same direction as that of PumEaroE. The ,B-Iac gene was PCR amplified from pSU18 and inserted into pKD12.047A resulting in the 7.7-kb plasmid pKD12.112.7 A 2.2-kb tktA fragment excised from pMF51A29 was ligated into pKD12.112 using T4 ligase afforded the 9.9-kb plasmid pKD12.138. The tktA gene is transcribed in the same direction as that of serA gene?"5 Overexpression of tktA-encoded transketolase is hypothesized to increase the availability of D-erythrose 4-phosphate.30 54 pMF51A BamHl digest Ncol BamHI 2.2-kb BamHI tktA Hindlll (Smal) i) Hindlll digest ii) Klenow treatment Klenow treatment iii) CIAP treatment Fast-LinkTM DNA ligase (Hind|l|)(Smal) Figure 23. Construction of plasmid pKD12.138. 55 Biosynthesis of quinic acid by glucose-limited fed-batch fermentation. The quinate-synthesizing E. coli QPl.1/pKD12.l38 were cultured under fermentor-controlled conditions at 33 °C, pH 7.0, with dissolved oxygen maintained at a set point of 10% air saturation. Plasmid maintenance relied on nutritional pressure as opposed to resistance to antibiotics. Glucose addition was controlled by dissolved 02 concentration with the rate of glucose addition dictated by a proportional-integral-derivative (PID) control loop. When dissolved oxygen levels exceeded the set point value indicating decreased microbial metabolism, the rate of glucose addition was increased and conversely the rate of glucose addition was decreased when dissolved oxygen levels declined below the set point value indicating increased microbial metabolism. A proportional gain (Kc) on the glucose PID control loop of 0.1 was used for culturing E. coli QPl.1/pKDl2.138. These conditions maintained a steady-state concentration of glucose of approximately 0.2 mM.26 Fed-batch fermentations of E. coli QPl.1/pKD12.138 synthesized 49 g/L of quinic acid in 20% (mol/mol) yield in 48 h (Figure 24). The titer is more than a 10-fold increase relative to the account of quinic acid production by E. coli AB2848/pKD136/pTW8090A, and required neither heterologous gene expression nor a multiple plasmid expression system. In addition to quinic acid, 3-dehydroquinic acid was accumulated at low levels in fermentation supernatant. Concentrations of 3- dehydroquinic acid steadily increased in the culture supernatant of E. coli QPl.1/pKD12.138 reaching a maximum concentration of 11.2 g/L at 24 h followed by a steadily decline to 5.5 g/L at 48 h. 56 A 01 O O O O O.) O 20 a -L O I r QA, DHQ, dry cell mass (g/L) O O 12 18 24 30 36 42 48 time (h) Figure 24. E. coli QPl.1/pKD12.138 cultured under glucose-limited, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. The yield (entry 1, Table 2) of quinic acid synthesized by E. coli QPl.1/pKDl2.138 reflects both the amount of glucose consumed to form biomass as well as the amount of glucose consumed to synthesize quinic acid. The theoretical maximum yield for synthesis of quinic acid from glucose with E. coli using phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) for glucose transport is 43% (mol/mol).248 The detailed analysis of theoretical maximum yield is deferred to Chapter 3. Cultures of E. coli QPl.1/pKD12.138 needed to be supplemented with aromatic amino acids and aromatic vitamins due to the aroD mutation rendering 3-dehydroquinate dehydratase catalytically inactive. E. coli QPl . 1/pKD12. 138 supplemented with partially purified shikimic acid produced 45 g/L of quinic acid in 23% molar yield.31 57 The accumulation of substantial concentrations of 3-dehydroquinic acid during the fermentation raised a question whether the shikimate dehydrogenase activity was sufficient in E. coli QPl.1/pKD12.138 to channel all the 3-dehydroquinic acid forward to quinic acid. Shikimate dehydrogenase specific activities were measured in the reverse direction at 25 °C with the oxidation of shikimic acid to 3-dehydroshikimic acid by monitoring the reduction of NADP+ at 340 nm (a: 6.18 x 10'3 M‘1 cm").32 Table 2 shows the shikimate dehydrogenase activities increased linearly from 9.2 units/mg to 19 units/mg during the fermentation runs (entry 1, Table 2). The steady decrease of 3- dehydroquinic acid concentration from 24 h to 48 h also raised an intriguing possibility that 3-dehydroquinic acid previously synthesized and exported into the culture medium could be transported back into the cytoplasm and subsequently be reduced to quinic acid. Table 1. Concentrations and yields of quinic acid and 3-dehydroquinic acid synthesized by various E. coli strains under different culture conditions. [QAJC QA [DHQ]" DHQ Total entry construct j/L yield‘ g/L yieldf yield‘ 1 QPl . l/pKD12. 138“ 49 20% 5.5 2.2% 22% 2 QPl.1/pNR4.230“ 46 20% 4.8 2.0% 22% 3 QPl.1/pKDlS.O7la 49 18% 11 4.1% 22% 4 QPl.1/pNR4.272° 56 20% 7.7 2.8% 23% 5 QPl.1/pKDl2.l38b 25 8.2% 46 16% 24% 6 QPl.1/pNR4.230b 39 12% 31 9.3% 21% 7 QPI . lptsG/pKD12J38b 26 9.8% 36 14% 24% “glucose-limited fermentor-controlled conditions. ”glucose-rich fermentor-controlled conditions. CQA: quinic acid. “DHQ: 3-dehydroquinic acid. ‘(mol QA)/(mol glucose consumed) f(mol DHQ)/(mol glucose consumed) “(mol QA + mol DHQ)/(mol glucose consumed) 58 Table 2. Shikimate dehydrogenase specific activities for various strains cultured under different conditions. shikimate delydrogenase (U/mg) entry commas 12 h 24 h 36 h 48 h 1 QPl.1/pKD12.138“ 9.2 10 15 19 2 QPl.1/pNR4.230“ 22 27 32 46 3 QPl.1/pKD15.071“ 9.5 12 14 17 4 QPl.1/pNR4.272“ 23 32 45 51 5 QPl.1/pKD12.138” 8.5 11 10 11 6 QP] . 14119124230” 19 22 22 23 “glucose-limited fermentor-controlled conditions. ”glucose-rich fermentor-controlled conditions. Table 3. DAHP synthase specific activities for various strains cultured under different conditions. DAHP synthase (U/mg) entry commas 12 h 24 h 36 h 48 h 1 QPl.1/pKD12.l38“ 0.40 0.33 0.22 0.16 2 QPl.1/pNR4.230“ 0.71 0.29 0.20 0.10 3 QPl.1/pKD15.07l" 0.65 0.73 0.43 0.30 4 QPl.1/pNR4.272“ 0.94 1.29 0.80 0.44 5 QPl.1/pKD12.l38b 0.46 1.05 0.90 0.74 6 QPl.1/pNR4.230b 0.44 1.73 1.57 0.73 “glucose-limited fermentor-controlled conditions. ”glucose-rich fermentor—controlled conditions. Increasing aroE-encoded shikimate dehydrogenase activity seems to be essential to increase quinic acid production. In plasmid pKD12.138, the PumEaroE gene was derived from plasmid pIA321,27 where the transcription of aroE is under the control of Pm),- and Pm tandom promoters. Plasmid pLZl.l6933 contained the same set of gene as that of pKD12.112 except pLZl.l69 expressed the aroE open reading frame directly under the control of a Pm promoter. Plasmid pNR4.230 was then constructed by cloning a 2.2-kb tktA gene excised from plasmid pSK4.203 into pLZl . 169 (Figure 25). 59 pSK4.203 l BamHl digest Ncol BamHl 2.2-kb BamHI tktA i) BamHl digest ii) CIAP treatment Fast-LinkTM DNA liagse BamHl (Smal) Figure 25. Construction of plasmid pNR4.230. 6O Culturing E. coli QPl.1/pNR4.230 under same conditions synthesized 46 g/L of quinic acid and 4.8 g/L of 3-dehydroquinic acid at 48 h (entry 2, Table 1). The maximum concentration of 3-dehydroquinic acid was 7 g/L at 24 h (Figure 26). Comparison of the specific activities obtained for QPl.1/pKD12.138 and QPl.1/pNR4.230 revealed that changing from PumharoE to PmaroE resulted in more than a two-fold increase in shikimate dehydrogenase activity under the same fermentation conditions (entres 1 and 2, Table 2). DAHP synthase specific activities were also about the same in these two constructs (entries 1 and 2, Table 3). The slight decrease in quinic acid concentration synthesized by QPl.1/pNR4.230 compared to QPl.1/pKD12.138 may reflect the metabolic burden caused by the higher expression level of shikimate dchydogenase. 60 350.. . m _ (I) g 40 «L g — _ 8 30. 8 d‘ 20 ~- ' I D <£10» 0 C 01 iii 0 12 18 24 30 36 42 48 time (h) Figure 26. E. coli QPl.1/pNR4.230 cultured under glucose-limited, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. 61 Circumvent phosphoenolpyruvate consumption by PTS-mediated glucose transport. Microbes such as E. coli employ the phosphoenolpyruvate:carbohydrate phosphotransferase (PT 8) system for glucose transport with the expenditure of one molecule of phosphoenolpyruvate for one molecule of glucose transported into the cytoplasm.34 The PI“ S system is composed of general PTS proteins and carbohydrate- specific proteins. The general PTS proteins include soluble cytoplastic enzyme I (EI) encoded by ptsI gene and histidine protein (HPr) encoded by ptsH gene that participate in transferring a phospho group from phosphoenolpyruvate to carbohydrates-specific PTS proteins. The glucose specific PTS proteins are composed of soluble err-encoded HAG“: and ptsG-encoded membrane-bound permease IICBGlc (Figure 27). glucose periplasm c o lasm yt p glucose 6-phosphate 1, p-® IIAG'c PEP P® P® pyruvateX @ X ® G) Figure 27. Schematic of glucose transport and phosphorylation using PTS system. Abbreviations: PEP, phosphoenolpyeuvate; El, PTS enzyme 1; HPr, PTS HPr protein; HAG“, glucose-specific PTS soluble component; IICBG'°, glucose-specific PTS membrane component. As shown in Figure 27, the phosphoryl group is transferred from phosphoenolpyruvate to El, then to HPr, and finally to IIAG'°. Phosphorylated IIAGlc 62 phosphorylates the membrane-bound IICBGlc that subsequently catalyzes the transport and phosphorylation of glucose. Utilization of phosphoenolpyruvate by PT S for glucose transport has been known to limit the yields of biosynthesis of shikimate pathway metabolites.” The PTS-generated pyruvate is further metabolized and not recycled to phosphoenolpyruvate under normal aerobic culture conditions. Phosphoenolpyruvate synthase-mediated pyruvate recycling to alleviate phosphoenolpyruvate limitation, first demonstrated by Patnaik et a1,36 has been successively exploited to improve the titers and yields of shikimate pathway products such as DAHP,36 3-dehydroshikimic acid,37 and shikimic acid.38 E. coli ppsA-encoded phosphoenolpyruvate synthase catalyzes the conversion of pyruvate with adenosine triphosphate (ATP) to form phosphoenolpyruvate, adenosine monophosphate (AMP) and inorganic phosphate.36 When phosphoenolpyruvate synthase fully recycles PTS- generated pyruvate to phosphoenolpyruvate, the theoretical maximum yield for the synthesis of 3-dehydroquinic acid from glucose will be doubled to 86% (mol/mol). The detailed analysis of this yield enhancement is deferred to Chapter 4. In an effort to increase the intercellular availability of phosphoenolpyruvate, plasmid pKD15.071 was constructed by plasmid localization of the ppsA gene into pKD12.138 (Figure 30).31 The transcription of ppsA was under the control of its own promoter. E. coli QP1.1 transformed with pKD15.071 was cultivated under fed-batch fermentation conditions for 48 h (Figure 28). The resulting 49 g/L quinic acid that was synthesized was identical to that observed in E. coli QPl.1/pKD12.138 (entry 3 vs. entry 1, Table 1). The final 3-dehydroquinic acid concentration increased to 11 g/L, resulting in a final QAzDHQ molar ratio of 4.5. Although the total production of quinic acid and 63 3-dehydroquinic acid was increased, the improvement was not reflected in the concentration of quinic acid synthesized, but in the accumulation of more 3- dehydroquinic acid in culture supernatant. The total yield of quinic acid and 3- dehydroquinic acid remained unchanged relative to that achieved with E. coli QPl.1/pKD12.138 (entry 3 vs. entry 1, Table 1). 60 3501 o ' 1 a”: 9 '— 3404 __ = C 8 304 a 'U 0' 20 4 ' I D 6 10+ i1 0 O o. 0 1 2 1 8 24 30 36 42 48 time (11) Figure 28. E. coli QPl.1/pKD15.07l cultured under glucose-limited, fermentor- controlled conditions. Quinic acid ([3), DHQ (I), dry cell mass (0) in g/L. The increased 3-dehydroquinic acid accumulation observed in fermentations of E. coli QPl.1/pKD15.071 prompted examination of the impact of phosphoenolpyruvate synthase overexpression in E. coli construct QPl.1/pNR4.230 which showed a 2-fold higher shikimate dehydrogenase activity (entry 2 vs. entry 1, Table 2). The 3.0-kb ppsA gene was cloned into pNR4.230 in the same fashion as in pKD15.071 (Figure 31). Culturing QPl.1/pNR4.272 under the same conditions resulted in 56 g/L of quinic acid in 20% (mol/mol) yield and 7.7 g/L of 3-dehydroquinic acid in 2.8% (mol/mol) yield (Figure 29 and entry 4, Table 1). .1 ,_J 0) b O O O O N O QA, DHQ, dry cell mass (g/L) 3 o 1 - 1 0 12 18 24 30 36 42 48 time (h) 0 Figure 29. E. coli QPl.1/pNR4.272 cultured under glucose-limited, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. The increase in the concentration of quinic acid and the increase in yield achieved by E. coli QPl.1/pNR4.272 relative to E. coli QPl.1/pKD12.l38 represent a modest improvement attendant with doubling the shikimate dehydrogenase activity and overexpression of phosphoenolpyruvate synthase. Increased shikimate dehydrohenase expression level also has a negative impact on phosphoenolpyruvate synthase activity. The measured phosphoenolpyruvate synthase activity is lower in construct E. coli QPl.1/pNR4.272 than that in E. coli QPl.1/pKD15.07l (entry 1 vs. 2, Table 4). Table 4. Phosphoenolpyruvate synthase specific activity during fermentation runs. a PEP synthase (U/mg) entry commas 12 h 24 h 36 h 48 h 1 QPl.1/pKD15.071 0.05 0.05 0.04 0.06 2 QPl.1/pNR4.272 0.03 0.04 0.02 0.01 “glucose-limited fermentor-controlled conditions. 65 pKL1 .878 Hindlll and EcoRI digest EcoRl Hindlll L 3.0-kb A l . y PpsA (Hindlll)(Smal) i) Ncol digest ii) Klenow treatment Klenow treatment iii) CIAP treatment Fast-LinkTM DNA liagse (Hindlll)(Smal) Figure 30. Construction of plasmid pKD15.071. 66 pKL1.87B Hindlll and EcoRI digest EcoRl Hindlll I 3.0-kb i l PPSA BamHl (Smal) i) Ncol digest ii) Klenow treatment Klenow treatment iii) CIAP treatment Fast-LinkTM DNA liagse pNR4.272 11 .5-kb BamHl (Smal) Figure 31. Construction of plasmid pNR4.272. 67 Another reported strategy to avoid expenditure of phosphoenolpyruvate by PTS is utilization of a non-PTS system for glucose transport. Non-PTS glucose transport includes Zymomonas mobilis glf—encoded glucose facilitator” and E. coli galP—encoded galactose-proton symport.40 After entry, glucose is phosphorylated by Z. mobilis or E. coli glk-encoded glucokinase to produce glucose 6-phosphate (Figure 32). It has been demonstrated that the Z. mobilis glucose facilitator and glucokinase can complement glucose transport and phosphorylation in E. coli strain lacking a functional native PTS system.39 glucose glucose + H+ periplasm GaIP cytoplasm glucose glucose + H+ ATP le ADP glucose 6-phosphate Figure 32. Schematic of non-PTS glucose transport and phosphorylation. The effect of expression of Z. mobilis glf and glk genes in a non-PTS E. coli host for the production of quinic acid was evaluated. E. coli QPl.1pts was constructed by disruption of the ptsH, ptsI and crr genes involved in the PTS-mediated glucose transport system by P1 phage transduction of E. coli QP1.1 using E. coli TP2811"l (A(ptsH, ptsI, 68 crr)::KanR) as the donor strain. Plasmid pSC6.09038 that was prepared by inserting a [CC glfglk cassette under the transcriptional control of a P promoter into plasmid pKD12.138 was then transformed into E. coli QPl.1pts. However, E. coli QPl.1pts/pSC6.090 cultivated under glucose-rich fermentor-controlled conditions grew slowly and failed to reach the stationary phase of growth. This is in contrast with a previously examined shikimate-synthesizing E. coli SP1.lpts/pSC6.090 which synthesized 71 g/L of shikimic acid in 27% (mol/mol) yield from glucose.38 3-Dehydroquinic acid recapture under glucose-limited conditions. The profile of 3-dehydroquinic acid synthesized under glucose-limited, fermentor-controlled conditions suggested an intriguing possibility that E. coli QP1.1 could transport 3- dehydroquinic acid previously synthesized and exported into the culture medium back into the cytoplasm and subsequently reduce it to quinic acid. To verify this hypothesis, E. coli QPl.1/pNR4.276, a construct incapable of de novo synthesizing quinic acid from glucose, was constructed. The aroFFER gene in the plasmid pNR4.230 was inactivated by a four-nucleotide frame shift to create the plasmid pNR4.276 (Figure 33). As a result, E. coli QPl .1/pNR4.276 expressed DAHP synthase only from its chromosomal aroF, aroG, and aroH loci. All of the DAHP synthase activity expressed in E. coli QPl.1/pNR4.276 was then sensitive to feedback inhibition by L-tyrosine, L-phenylalanine, and L- tryptophan added into the fermentation medium. E. coli QPl.1/pNR4.276 was cultured under glucose-limited conditions with the addition of L-tyrosine (0.35 g), L-phenylalanine (0.35 g), and L-tryptophan (0.175 g) every 6 h staring from 12 h and continuing until 36 h to inhibit native DAHP synthase 69 21C activities. The absence of quinic acid and 3-dehydroquinic acid in a control experiment indicated an effective inhibition of DAHP synthase under this condition. BamHl(Smal) i) 391" digest ii) Klenow treatment iii) Fast-LinkTM DNA liagse pNR4.276 9.5-kb BamH|(Smal) Figure 33. Construction of plasmid pNR4.276. 70 3-Dehydroquinic acid was synthesized from quinic acid in 5 steps according to a method described previously (Figure 34).42 The trans-diol of methyl quinate was selectively protected with 2,3-butane bisacetal43 followed by transesterification to afford the benzyl ester of protected quinate 7. Oxidation of compound 7 with KIO3 in the presence of a catalytic amount of RuCl3 gave fully protected dehydroquinate 8. Acid hydrolysis of the 2,3-butane bisacetal protecting group followed by catalytic hydrogenation of the benzyl ester 9 afforded 3-dehydroquinic acid in 36% total yield from quinic acid. A recently reported synthesis of 3-dehydroquinic acid utilized the similar reaction sequence for conversion of quinic acid to 3-dehydroquinic acid.44 HO,. COZH HO,‘ COZCH3 HO,“ COan 3 0 ——»“ ——> HO“. i OH Ho“ ,_ o OCH Ho“ i o OCH \ \ OH O \ CH3 3 O:&CH3 3 H30 OCH3 HaC OCH3 quinic acid 6 7 HO. COan HO,. COan HO,. COZH —> ——> 1 ——-> o i o OCH 0 i OH 0 i OH \ HaC OCH3 3-dehydroquinic 8 9 acid Figure 34. Synthesis of 3-dehydroquinic acid from quinic acid. Keys: (a) i) Dowex- 50 (H*), MeOH, reflux; ii) trimethylorthoformate, 2,3-butanedione, (:)-10- camphosulfomic acid, 66%; (b) (BuZSnCl)O(Bu2SnOH) (0.05 equiv.), BnOH (2 equiv.), toluene, reflux; (c) RuCl3 (0.03 equiv.), KIO3 (3 equiv.), benzyltrimethyl ammonium chloride (0.02 equiv.), room temperature, 67%; (d) CF3C02H/H20 (20:1, v/v), CHZClz, 0 °C, 2 h, 82%; (e) Pd/C (0.05 equiv.), H2, 50 psi, THF/H20 (1:1, v/v), 97%. 71 The synthesized 3-dehydroquinic acid (5.0 g) was added into E. coli QPl.1/pNR4.276 culture medium at 18 h again with the same aromatic amino acid supplements as in the control experiment to inhibit E. coli native DAHP synthase activities. Formation of quinic acid was observed 6 h after the addition of 3- dehydroquinic acid (Figure 35). At 48 h, 2.5 g/L of quinic acid and 2.1 g/L of 3- dehydroquinic acid were present in the culture supernatant. This result indicates that recapture of 3-dehydroquinic acid from culture medium into E. coli cytoplasm and subsequent reduction by shikimate dehydrogenase operate in this construct under glucose-limited fermentor-controlled condition. 7 . 6O 6 O O , 450 -. 3 1:5 we 93 __ to c4 308 I E 0.3+ =5 < o 02.- .205 1- . 71° 04—. . i0 0 12 48 time (h) Figure 35. Biosynthesis of quinic acid from the added 3-dehydroquinic acid by E. coli QPl.1/pNR4.276 under glucose-limited, fermentor-controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. 72 Biosynthesis of quinic acid under glucose-rich, fermentor-controlled conditions. Glucose-rich fermentation where glucose is constantly added to the fermentation vessel at a rate sufficient to maintain a glucose concentration between 50 mM and 160 mM is an effective method for increasing the flow of carbon directed into the shikimate pathway.” However, culturing E. coli QPl.1/pKDlZ.l38 at 33°C, pH 7.0, DC. 10% under the glucose-rich conditions, only led to the formation of 25 g/L of quinic acid and 46 g/L of 3-dehydroquinic acid by 48 h (Figure 36). Examination of the DAHP synthase specific activities under this condition revealed higher activities compared to that of the same construct cultivated under glucose-limited condition (entry 5 vs. entry 1, Table 3). However, the shikimate dehydrogenase activities declined significantly. It remained at a stable level of approximately 10 units/mg throughout the fermentation run under glucose-rich condition, as compared to a gradual increase from 9 to 19 units/mg under glucose-limited condition (entries 5 and 1, Table 2). O) O 01 O .h C N 0 QA, DHQ, dry cell mass (g/L) 8 8 O Figure 36. E. coli QPl.1/pKD12.138 cultured under glucose-rich, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. 73 Cultivation of E. coli QPl.1/pNR4.230 (Figure 37) under glucose-rich, fed-batch conditions synthesized 39 g/L of quinic acid in 12% (mol/mol) yield and 31 g/L of 3- dehydroquinic acid in 9.3% (mol/mol) yield (entry 6, Table 2) by 48 h. The increase of 14 g/L of quinic acid relative to QPl.1/pKD12.l38 corresponded to a decline of 15 g/L of 3-dehydroquinic acid. Comparison of the enzyme activities revealed a two-fold increase in shikimate dehydrogenase activities (entry 6 vs. entry 5, Table 2) and similar DAHP synthase activities (entry 6 vs. entry 5, Table 3). The two-fold increased expression of shikimate dehydrogenase increased the concentration of synthesized quinic acid significantly under glucose-rich conditions. 0) O N (D h 01 O O O O 0 QA, DHQ, dry cell mass (g/L) 3 O 30 36 42 48 time (h) Figure 37 . E. coli QPl.1/pNR4.230 cultured under glucose-rich, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. Biosynthesis of quinic acid in a E. coli QPl.1 mutant devoid of catabolite repression by glucose. The accumulation of substantial concentrations of 3- dehydroquinic acid under glucose-rich fermentation conditions can not be solely 74 attributed to the lower expression of shikimate dehydrogenase since the shikimate dehydrogenase activities of QPl.1/pNR4.230 under glucose-rich fermentation conditions were even higher than those of OP1.1/pKD12.138 under glucose-limited conditions (entry 6 vs. entry 1, Table 2). One possibility is that E. coli can not transport back and reduce 3-dehydroquinic acid previously synthesized and exported into the fermentation medium due to catabolite repression by glucose that is present in high concentrations under glucose-rich condition. In addition to its role in carbohydrate transport, the PTS system is also known to be involved in the regulation of catabolite repression.34 In an attempt to disrupt catabolite repression in E. coli QP1.1 strain by glucose, a mutation in E. coli ptsG locus was introduced into E. coli QP1.1 by Pl-phage mediated transduction of ptsG::Tn5 from E. coli IT1168.46 As described previously, ptsG encodes a transmembrane IICBGLC protein in the E. coli phosphoenolpyruvate-dependent glucose phosphotransferase (PTS) system for transport and phosphorylation of glucose to glucose 6-phosphate (Figure 26).47 The AGLC IICBGLC protein is phosphorylated by phosphorylated II protein. In the presence of high glucose concentration, the IIAGLC protein is subsequently dephosphorylated. The dephosphorylated IIAGLC protein binds to and inactivates a number of non-PTS carbohydrates permeases.34 Disruption of the ptsG- gene has been shown to result in a loss of catabolite repression by glucose.48 E. coli QPl.1ptsG formed blue colonies on X- gal indicator plate containing glucose and lactose. The formation of blue colonies indicated expression of ,B—galactosidase in the presence of glucose due to the inactivation of catabolite repression. The effect of introducing the ptsG mutation on quinic acid production was tested on E. coli QPl.1ptsG/pKD12.138, which was cultivated under 75 glucose-rich fed-batch fermentation conditions. Over a period of 48 h, E. coli QPl.1ptsG/pKD12.138 synthesized 36 g/L of 3-dehydroquinic acid and 26 g/L of quinic acid (Figure 38). The concentration and yield of quinic acid synthesized by E. coli QPl.1/pKDlZ.l38 and E. coli QPl.1ptsG/pKD12.l38 were essentially the same (entry 5 vs. entry 7, Table l). The decline in total production relative to E. coli QPl.1/pKD12.138 may reflect a slower rate of glucose transport by inactivating the major glucose transport protein. Glucose transport in E. coli QP1.1 ptsG could occur less efficiently via one or more of the receptors belonging to the glucose, mannitol or mannose family of proteins that are not specific for glucose transport.34 0) O N (10 h 01 O O O 0 QA, DHQ, dry cell mass (g/L) 8 O 0 12 18 24 30 36 42 48 time (h) Figure 38. E. coli QPl.1ptsG/pKD12.138 cultured under glucose-rich, fermentor- controlled conditions. Quinic acid (El), DHQ (I), dry cell mass (0) in g/L. In summary, a microbial route to quinic acid by E. coli QPl.1/pNR4.272 has achieved a titer of 56 g/L and 20% (mol/mol) yield from glucose under glucose-limited, fermentor-controlled condition. 76 {‘0 01 ill Hydroquinone Toxicity. The toxicity of hydroquinone toward ethanologenic E. coli cultured on xylose under fermentative conditions has been analyzed from the perspective of hydroquinone’s inhibition of sugar catabolism and damage to the plasma membrane.49 To gauge the toxicity of hydroquinone toward E. coli cultured aerobically 248 on glucose, a 3-dehydroquinic acid-synthesizing E. coli QPl.1/pKL4.33 was used. E. coli QPl.1/pKL4.33 differs from E. coli QPl.1/pKDl2.138 only in the absence of plasmid-localized aroE encoding shikimate dehydrogenase and tktA encoding transketolase. Utilization of E. coli QPl.1/pKL4.33 that only synthesizes 3- dehydroquinic acid for analysis of hydroquinone’s toxicity avoided the complication with E. coli QPl.1/pKD12.l38 that synthesized a mixture of quinic acid and 3-dehydroquinic acid. The fermentor conditions used to culture 3-dehydroquinate-synthesizing E. coli QPl.1/pKL4.33 were based on the parameters used to culture quinate-synthesizing E. coli QPl.1/pKDl2.138. A sterile, aqueous solution of hydroquinone was added to the fermentor run to a final concentration of 2 g/L at 12 h. E. coli QPl.1/pKL4.33 was able to grow and synthesize 3-dehydroquinic acid in the presence of added hydroquinone. However, 3-dehydroquinic acid synthesis dropped by approximately 50% upon addition of hydroquinone (Figure 398). Less cell mass was also formed (Figure 39a) and increased amounts of acetic acid were produced (Figure 39b) in the presence of added hydroquinone. The specific activity of DAHP synthase did not significantly change over the course of the fermentor run when hydroquinone was added. As the first enzyme in the common pathway, the specific activity of DAHP synthase significantly determines carbon flow directed into synthesis of 3-dehydroquinic acid.50 The reduced concentration of synthesized 3-dehydroquinic acid, reduction in cell 77 mass, and increased production of acetic acid observed for E. coli QPl.1/pKL4.33 indicate that hydroquinone concentrations as low as 2 g/L are toxic to E. coli growth and metabolism. 50 40 -~ 35 ~- 30 ~- 25 4- 20 4 15 4» O 10 4 DHQ, dry cell mass (g/L) I) o o 12 18 24 30 36 42 48 time (h) 2.5 1.5 -» ACOH (SJ/L) 0.5 .. , ‘ 0 . ‘ I l I 12 18 24 30 36 42 48 time (h) Figure 39. 3-Dehydroquinate-synthesizing E. coli QPl.1/pKL4.33. (a) 3- Dehydroquinic acid (DHQ) synthesis and E. coli cell mass in the absence (D) vs. presence (I) of added hydroquinone; (b) Acetic acid formation in the absence (D) vs. presence (I) of added hydroquinone. 78 Chemical synthesis of hydroquinone from quinic acid. The literature method that used magnesium (IV) dioxide for oxidation of quinic acid to hydroquinone6'5' resulted in a low yield and formation of salts as byproducts. New chemical methodology was needed to significantly increase the yield and eliminate the byproducts. Reagents were needed that were sufficiently robust for use in fermentation broths while being mild enough to avoid overoxidation of hydroquinone to benzoquinone. The conversion of quinic acid to hydroquinone can be viewed to proceed via an oxidative decarboxylation followed by dehydration and subsequent aromatization to hydroquinone. Chemical methods for oxidative decarboxylation of a-hydroxycarboxylic acids involves the use of a variety of oxidants including periodate,52 permanganate,53 chromic acid,54 lead acetate"’5 and hypochlorite.56 Among these, hypochlorite is easily available, cheap and an environmentally acceptable chemical (bleaching agent). Hypochlorite has been well established as a versatile oxidant in organic chemistry. Diols, diones and a—hydroxyketones are easily cleaved oxidatively with hypochlorites.57 Hypochlorite-induced oxidative decarboxylation of a—hydroxycarboxylic acids and a- keto acids have been reported in the literature. Nwaukwa et al.58 reported conversions of a—hydroxy carboxylic acids into lower carboxylic acid homologs with calcium hypochlorite in aqueous acetonitrile-acetic acid solutions. Carlsen56 reported oxidation of a-hydroxyacids exclusively to the corresponding aldehydes with sodium hypochlorite in a two-phase ether-water solvent system without the addition of acetic acid. Elmore et al.59 extended the hypochlorite-induced oxidative decarboxylation to trisubstituted acetic acid compounds. 79 Hypochlorite oxidations of quinic acid. A typical fermentation broth of E. coli QPl.1/pNR4.272 contains 290 mM ammonium quinate, 44 mM ammonium 3- dehydroquinate, approximately 1 M ammonium salts, 43 mM potassium phosphate and around 50 g/L of cell mass. Ammonium hydroxide along with 1 M HZSO4 were added to the culture medium to maintain the pH at 7 during fermentation runs. The quinic acid fermentation broth was partially purified to remove E. coli cells, proteins and ammonium salts. E. coli cells were removed by centrifugation. Heating the culture supernatant to reflux followed by acidification resulted in precipitation of proteins, which were removed by centrifugation. The clarified fermentation broth was then decolorized with activated charcoal. 3-Dehydroquinic acid could be removed from quinic acid broth by converting it into protocatechuic acid upon reflux with acid and was subsequently removed during the charcoal decolorization step. Initial attempts at oxidative decarboxylation of quinic acid with sodium hypochlorite (commercial bleach) were carried out in the decolorized fermentation broth. However, hypochlorite oxidation of quinic acid was not observed in lieu of ammonium ion removal. Hence, clarified, decolorized fermentation broth was passed through a strong cation-exchange resin (Dowex 50 H") to remove ammonium ion prior to hypochlorite oxidation. Sodium hypochlorite (3 equivalents) was added to clarified, decolorized, ammonium ion-free fermentation broth. The reaction was then acidified to pH 1.5 - 2 with simultaneous addition of H2S04 (2 M, 1.25 equivalents) and reacted at room temperature to give 3(R),5(R)-trihydroxycyclohexanone (10, Figure 40) based on ‘H NMR analysis of the crude reaction solution. Excess hypochlorite was then quenched with 2-propanol. Without purification, the resulting solution was refluxed under an inert 80 atmosphere for 10 h. The concentration of cyclohexanone 10 decreased rapidly, the concentration of hydroquinone increased, and a,/3-unsaturated enones 11 and 12 accumulated as intermediates (Figure 41). HO,‘ 002H OH HO“ :5 ‘OH OH OH quinic acid hydroquinone H20 002% a \1 b o — 0 o - b 6 6. —__’ + H0“ 2 OH Ho“ 2 i OH OH OH OH 10 _ enone 11 enone 12 _ Figure 40. Conversion of quinic acid to hydroquinone. Keys: (a) i) NaOCl, room temperature, ii) 2-propanol, room temperature; (b) reflux. Identification and quantification of enone intermediates 1 1 and 12 was accomplished by 1H NMR analysis of aliquots withdrawn from the reaction solution and comparison with the 1H NMR spectra of enones 11 and 12 independently synthesized from 3(R),5(R)—trihydroxycyclohexanone (10) and butane 2,3-bisacetal-protected methyl quinate, respectively (Figure 42). Extraction of the dehydration/aromatization reaction solution with tert—butylmethyl ether followed by sublimation afforded hydroquinone in 87% overall isolated yield from quinic acid. 81 70 604F7 A ‘ A A A A5 -_ 2 0 ‘ é c40-~ .9 A §30 E " o “8 820‘“ o ' . o 10“" [TI 0 o 0 OJ- 1+H4fil1lfltlfia'gaato— O 1 2 3 4 5 6 8 10 time(h) Figure 41. Dehydration and aromatization of 3(R),5(R)-trihydroxycyclohexanone (10) to hydroquinone. 3(R),5(R)-trihydroxycyclohexanone (10) (E1), 4(S),5(R)- - dihydroxy-Z-cyclohexen-1-one (l l) (O), 4(R),5(R)-dihydroxy-2-cyclohex-l-one ( 12) (O), hydroquinone (A). The enone 11 was obtained in a three-step synthesis from quinic acid. The intermediate 13 has been used as a chiral building block in a number of natural product syntheses.”“""60 Synthesis of compound 13 via the 3(R),5(R)-trihydroxycyclohexanone 10 provided a shorter route compared to the reported synthesis from quinic acid.“ Removal of the acetonide protecting group of 13 gave cis-diol 11 in 23% total yield from quinic acid. Synthesis of enone intermediate 12 utilized selective protection of the trans- diol of the methyl quinate with the butane 2,3-bisacetal protecting group.43 Reduction of the methyl ester with lithium aluminum hydride followed by cleavage of the resulting vicinal diol by NaIO4 afforded the ,B-hydroxyketone 16. The hydroxy group was eliminated by acetylation followed by treatment with diisopropylethylamine. Removal of 82 the butane 2,3-bisacetal protecting group with trifluoroacetic acid afforded trans-diol 12 in 40% overall yield from quinic acid. HO, COZH O -——» 6. —+ 6 ——»" 6 H0“. H0“ OH 0“ Ho“' 3 OH quinic OHacid 1O 13 11 1. HO co3CH3e ; 0 HO\\. H09. §—> .0. 9‘!OH OH 0.10CH3 O.\OCH3 O.\OCH3 CH H CH H300 CCH3 3 H300 EC}? 3 HaOC CCH3 3 14 15 16 12 Figure 42. Syntheses of enone intermediates 11 and 12 from quinic acid. Keys: (a) NaOCl, H2803, rt., 76%; (b) i) acetone, TsOH, rt., 79%; ii) AczO, (i-Pr)2NEt, DMAP, CH2C13, 0 °C, 91%; (c) CF3C03H/I-IZO (2:1, v/v), 0 °C, 43%; (d) i) CH3OH, Dowex 50(H+), reflux; ii) 2,3-butanedione, CH(OCH3)3, CH3OH, CSA, reflux, 79% (e) LiAlH3, THF, 0 °C, rt., 93% (f) NaIO3, phosphate buffer (pH 7), 0 °C, room temp., 72%; (g) i) ACZO, (i-Pr)3NEt, DMAP, CHzClz, 0 °C, 100%; ii) CF3C03H/CH2C13/HZO (9:1:1, v/v/v), 0 °C, 75%. Electrochemical oxidation. Quinic acid was also converted into 3(R),5(R)~ trihydroxycyclohexanone (10, Figure 40) by electrochemical oxidation.‘52 The electrolysis was performed at room temperature in a 50 cm3 electrolysis cell (Figure 43) fitted with a pair of Pt electrodes (2 x 1.35 cmz). Clarified, decolorized, ammonium ion- free quinic acid fermentation broth was adjusted to pH 10 by addition of l N aqueous 83 NaOH prior to electrolysis. Electrolysis at a current density of 400 mA/cm2 for 4 h afforded 3(R),5(R)-ttihydroxycyclohexanone 10 in 24% yield along with an 8% yield of formic acid. A total of 25% of unreacted quinic acid remained in the solution. The rest of the quinic acid was likely to be oxidized to CO2 during the electrolysis. Electrolysis at a higher current density (600 mA/cmz) or electrolysis for a longer period of time resulted in increased formation of formic acid without increased yields of ketone 10. Protek DC power supply ’3 6 ' ' : Pt L - 3......3 D Pt (‘Ub uuuuu > Figure 43. Diagram of electrochemical oxidation apparatus. Transition metal-mediated oxidation. High-yielding conversion of quinic acid to hydroquinone by transition metal-mediated oxidative decarboxylation was achieved using stoichiometric amounts of Ce4+ and VS“ salts (Table 1). Oxidation of quinic acid in clarified, decolorized, ammonium ion-free fermentation broth at room temperature with 84 ceric ammonium sulfate ((NH3)3Ce(SO4)3) followed by refluxing the reaction solution for 10 h afforded hydroquinone in 91% isolated yield (entry 1, Table 5). Addition of vanadium pentoxide (V203) and H380, to quinate-containing fermentation broth, which had been clarified, decolorized, and rendered ammonium ion-free led to an 85% isolated yield of hydroquinone (entry 2, Table 5) after heating the reaction solution to 50 °C and final reaction at reflux for 8 h. The oxidations and associated reaction conditions summarized in Table 5 are noteworthy in the absence of benzoquinone formation resulting from oxidation of the initially formed hydroquinone. Catalytic oxidative decarboxylation. O2 is an environmental-friendly oxidant in organic chemistry. Catalytic aerobic oxidative decarboxylation of a-hydroxyacids by a Co(III) ortho—phenylene—bis-(N ’-methyloxamidate) complex in the presence of pivalaldehyde has been reported.63 However, this reaction was performed in organic solvent and required use of a stoichiometric amount of pivalaldehyde. In route to identification of metals that could be used in catalytic amounts, CuCl with 02 as a cooxidant and FeSO,3 or RuCl3 with 11202 as the cooxidant were examined.64 No oxidative decarboxylation of quinic acid in clarified, decolorized, ammonium ion-free fermentation broth was observed. However, use of substoichiometric, catalytic amounts of Ag3PO4 along with potassium persulfate (K38303) as the cooxidant did lead to oxidative decarboxylation. For these Ag3PO4-catalyzed oxidations, removal of ammonium ions was critical. The presence of ammonium ions led to formation of a silver mirror and an absence of hydroquinone formation. Oxidation of quinic acid in partially purified culture supernatant with a catalytic amount of Ag3PO4 (10 mol %) and cooxidant K2S203 heated initially at 50 °C followed by heating of the reaction solution at 85 reflux afforded an 85% yield of hydroquinone (entry 3, Table 5). Quantities of Ag3PO3 catalyst as low as 2 mol % and 1 mol % relative to quinic acid could be used (entries 4 and 5, Table 5) although significant decreases in the yield of hydroquinone obtained were observed. Table 5. Reaction conditions and yields for chlorine-free oxidation of quinic acid. entry oxidant/catalyst (equiv.)" conditions Hygsggugne l (NH3)3Ce(SO,3)3 (2.4) rt., 30 min; reflux, 10 h 91 2 V305 (1.1) 50 °C, 4 h; reflux, 8 h 85 3 K3S303 (1.2)/Ag3PO3 (0.10) 50 °C, 4 h; reflux, 8 h 85 4 K28303 (1.2)/Ag3PO4 (0.02) 50 °C, 4 h; reflux, 8 h 74 5 K2S203 (l.2)/Ag3PO4 (0.01) 50 °C, 4 h; reflux, 8 h 51 “ Units: mol oxidant/mol quinic acid. b reported yields are based on isolated, purified hydroquinone. Discussion Microbial synthesis of quinic acid from glucose. The first reported microbial synthesis of quinic acid relied on heterogeneous expression of the Klebsiella pneumoniae qad gene encoding NAD+-requiring quinate dehydrogenasef Synthesis of quinic acid utilizing the E. coli aroE-encoding NADP-dependent shikimate dehydrogenase is noteworthy given that 3-dehydroquinic acid is not its native substrate. Overexpression of the E. coli aroE gene instead of the K. pneumoniae qad gene in E. coli constructs avoids the difference in codon usage and protein folding between these two organisms. The shikimate dehydrogenase activity in this study was measured using oxidation of shikimic acid as the substrate.65 The Michaelis constant (Km = 1.2 mM) for the unnatural substrate 3-dehydroquinic acid by shikimate dehydrogenase is roughly ten times higher than for 3- 86 dehydroshikimic acid (Km = 0.11 mM), and the vmax is slightly lower for shikimate dehydrogenase-catalyzed reduction of 3-dehydroquinic acid than 3-dehydroshikimic acid (0.096 mmol/L/min vs. 0.11 mmol/L/min).7 Therefore, the specific activities determined during fermentation runs do not represent the in vivo activities with 3-dehydroquinic acid as substrate. A two-fold increase in shikimate dehydrogenase activity resulted from replacing the P, 3,033 tandom promoter in E. coli QPl.1/pKD12.138 with a single Pm promoter for aroE in E. coli QPl.1/pNR4.230 but did not lead to an increase in either the concentration or yield of biosynthesized quinic acid. The approximately 5 g/L of 3- dehydroquinic acid accumulated in fermentation broth of E. coli QPl.1/pKD12.138 at 48 h under the glucose-limited conditions could be attributed to feedback inhibition of shikimate dehydrogenase by quinic acid in light of the precedent that shikimate dehydrogenase is feedback-inhibited by shikimic acid.28 When ppsA-encoding phosphoenolpyruvate synthase was co-expressed with tktA—encoding transketolase in E. coli QPl.1/pKD15.071 and E. coli QPl.1/pNR4.272, the flow of carbon into the common pathway was elevated. E. coli QPl.1/pKD15.071 synthesized the same amount of quinic acid (49 g/L) as E. coli QPl.1/pKD12.138, which possibly reflects a limitation imposed by inadequate shikimate dehydrogenase activity. Accordingly, increasing shikimate dehydrogenase specific activity by two-fold in E. coli QPl.1/pNR4.272 led to an increase in synthesized quinic acid. The steady decrease from 24~48 h of 3-dehydroquinic acid, which initially increased in concentration over the first 24 h during fermentor-controlled cultivation, raised an intriguing possibility that E. coli was capable of recapturing 3-dehydroquinic 87 acid in the culture medium and subsequently reducing it to quinic acid under glucose- limited conditions. A putative quinate transport protein has been found in Aspergillus nidulans.66 However, a quinate transport protein has not been identified in E. coli. To establish the recapture of initially exported 3—dehydroshikimic acid, chemically synthesized 3-dehydroquinic acid was added into fermentation medium of E. coli QPl.1/pNR4.276, which lacks an active feedback-insensitive DAHP synthase. The three native DAHP synthase isozymes encoded by aroF, aroG and aroH in the E. coli QP1.1 genome are subjected to the feedback inhibition by three aromatic amino acids, L- tyrosine, L-phenylalanine and L-tryptophan. Addition of three aromatic amino acids into the fermentation medium effectively blocked the biosynthesis of shikimate pathway metabolites. E. coli QPl.1/pNR4.276 catalyzed the conversion of 3-dehydroquinic acid added to its culture medium into a 1.2:] molar ratio of quinic acidz3-dehydroquinic acid within 36 h. Formation of quinic acid supports the hypothesis that E. coli is capable of recapturing 3-dehydroquinic acid in the culture medium and reducing it to quinic acid. The importance of recapturing 3-dehydroquinic acid was demonstrated when E. coli biocatalysts were cultured under fermentor-controlled, glucose-rich conditions. Culturing E. coli QPl.1/pNR4.230 under glucose-rich condition resulted in the synthesis of 39 g/L of quinic acid and 31 g/L of 3-dehydroquinic acid. By contrast, E. coli QPl.1/pKD12.l38 under glucose-limited condition synthesized 49 g/L of quinic acid and 5.5 g/L of 3-dehydroquinic acid, although the latter construct possessed lower shikimate dehydrogenase specific activities (entry 6 vs. entry 1, Table 2). Catabolite repression by glucose likely contributed to the dramatic shift in product profile under glucose-rich culture conditions. Excess glucose present in the fermentation 88 medium may inhibit the transport of 3-dehydroquinic acid synthesized and exported into the culture medium back into the E. coli cytoplasm. For E. coli QPl.1/pKD12. 138 and E. coli QPl.1/pNR4.230, shikimate dehydrogenase specific activities were significant lower under glucose-rich conditions relative to the activities observed under glucose-limited conditions. Meanwhile, DAHP synthase activity is generally higher under glucose-rich condition than that under glucose-limited condition. The combination of these effects resulted in the accumulation of significant concentrations of 3-dehydroquinic acid byproduct in the culture medium. Inactivating catabolite repression by glucose could enable E. coli to transport 3- dehydroquinate into the cytoplasm in the presence of glucose. One such catabolite repression-insensitive mutant, E. coli QPl.1ptsG/pKD12.138, was constructed and examined under glucose-rich culture conditions but did not result in an increase in quinic acid synthesis. One straightforward way to increase quinic acid synthesis without reestablishing the 3-dehydroquinic acid recapture mechanism under glucose-rich culture conditions is to further increase the shikimate dehydrogenase activity to compete with 3- dehydroquinate export. E. coli QPl.1/pNR4.230 synthesized a higher concentration of quinic acid relative to E. coli QPl.1/pKD12.138, which possessed lower shikimate dehydrogenase specific activities under glucose-rich, fermentor-controlled conditions (entry 5 vs. 6, Table 1). However, further increasing shikimate dehydrogenase activities by expressing aroE from a stronger promoter such as T7 promoter might not be beneficial to quinic acid production due to the metabolic burden associated with high levels of protein expression. 89 One possible solution to this problem is to increase the catalytic efficiency of the aroE-encoding shikimate dehydrogenase toward the unnatural substrate 3-dehydroquinic acid by directed evolution or site-directed mutagenesis. After completion of this work, a ydiB-encoded shikimate dehydrogenase isozyme in E. coli was reported and characterized.“68 The YdiB is a dual specificity quinate/shikimate dehydrogenase using either NAD+ or NADP” as a cofactor. The K, for quinate with NAD+ as cofactor is 41 11M which is 10 times lower than with NADP” as cofactor (K, = 555 M). The Km value for shikimate is 20 11M with NAD“ and 120 11M with NADP+ as cofactor. By contrast, aroE-encoded shikimate dehydrogenase isozyme uses NADP” exclusively with a Km value of 65 ,uM for shikimate.67 However, the catalytic efficiency of aroE-encoded shikimate dehydrogenase (k = 14,200 min“) is 2000-4000 fold higher than the ydiB- cat encoded dual-substrate dehydrogenase (k = 3-7 min"). YdiB shares a 50% similarity with the AroE counterpart. Both enzymes display a similar architecture with two 01/13 domains separated by a wide cleft. Replacing aroE-encoded shikimate dehydrogenase with yd iB-encoded quinate/shikimate dehydrogenase in the quinate-synthesizing E. coli constructs may not be able to achieve high-titer synthesis of quinic acid due to its low km, value. However, site-specific mutagenesis or directed evolution of the aroE-encoded shikimate dehydrogenase isozyme could be utilized to create a mutant dehydrogenase that exhibits both high catalytic efficiency and high specificity toward 3-dehydroquinic acid. The availability of X-ray crystal structures for both AroE and YdiB potentially could simplify and accelerate the mutagenesis and evolution process. Since the structures of AroE and YdiB are similar, comparison of their substrate-binding motif may provide valuable 90 information to guide the site-directed mutagenesis or saturation mutagenesis in aroE- encoded shikimate dehydrogenase. Screening methods based on measuring dehydrogenase activity toward reduction of 3-dehydroquinic acid can be exploited to identify the mutants with improved activity. Alternatively, a selection strategy can be envisioned to take advantage of the reversibility of these dehydrogenase-catalyzed conversions by selecting mutants capable of using quinic acid as a sole aromatic amino acids supplement in an E. coli strain lacking DAHP synthase. An improved quinate dehydrogenase is expected to eliminate the 3-dehydroquinate accumulation mentioned above and channel all of the carbon directed into the shikimate pathway into quinic acid. Chemical synthesis of hydroquinone from quinic acid. With the improvements achieved in the biocatalytic methodology for conversion of glucose into quinic acid, improving the companion chemical methodology for conversion of quinic acid into hydroquinone becomes critical to make this route practical. Reaction methodology previously employed for the chemical conversion of quinic acid to hydroquinone (Figure 18) by refluxing purified quinic acid in an aqueous solution with MnO2 synthesized hydroquinone in 10% yield. 6 Utilization of MnO2 has the same problem associated with the commercial route to hydroquinone where stoichiometric amounts of MnO2 are used to oxidize aniline (Figure 15).1 Oxidative decarboxylation of quinic acid by reactions with hypochlorite afforded 3(R),5(R)-trihydroxycyclohexanone (10, Figure 40) rapidly and quantitatively at room temperature. The mild reaction conditions and the stability of 3(R),5(R)- trihydroxycyclohexanone under these conditions are critical for a high-yielding conversion of quinic acid into hydroquinone. Dehydration and aromatization of ketone 91 10 to hydroquinone during oxidative decarboxylation of quinic acid would likely result in overoxidation of hydroquinone to benzoquinone or chlorination of hydroquinone. Of all the methods developed for chlorine-free oxidative decarboxylation of quinic acid, electrochemical oxidation of quinic acid was the simplest and most environmental-friendly method. However, this method suffered from overoxidation of either quinic acid or its decarboxylation product 10 as suggested by the formation of formic acid. Only approximately 50% of the mass balance could be accounted for after partial electrochemical oxidation of quinic acid. The lack of improved 3(R),5(R)- trihydroxycyclohexanone (10) yields realized with use of electrochemical oxidation led to the examination of oxidative decarboxylation of quinic acid with Ce‘“ and VS“. Oxidative decarboxylation of a-hydroxycarboxylic acids with Ce4+ and V+5 has been reported.‘59 The metal-mediated oxidation of the a-hydroxycarboxylate quinic acid afforded favorable yields (Table 5) of hydroquinone formed with (NH4)2Ce(SO3)3 and V305 as oxidants. Moving from use of stoichiometric to catalytic amounts of metal for oxidation of quinic acid took advantage of the reported ability of Ag2+ formed from peroxydisulfate 7° Use of insoluble oxidation of Ag1+ to accelerate the oxidation of carboxylic acids. Ag3PO4 as catalyst followed from inorganic phosphate being the dominant oxyanion present in fermentation broth. Overoxidation of hydroquinone to benzoquinone was not observed when only a small molar excess of K28203 relative to quinic acid was used. Benzene-free synthesis of hydroquinone. Exposure to benzene, which has been linked to both acute myeloid leukemia and non-Hodgkin's lymphoma,7| continues to create challenges to the chemical industry. Besides the fact that benzene is derived from 92 nonrenewable resources, high costs have been cited by the US. chemical industry to be a major impediment to reducing exposure limits for benzene.72 Ultimately, the most effective way of addressing benzene’s human health risk may be the development of fundamentally new syntheses for aromatic chemicals and products from non-toxic, non- volatile and renewable starting materials to replace the existing synthetic routes from benzene and its derivatives. The synthesis of hydroquinone via chemical oxidation of microbe-synthesized quinic acid can be viewed as being part of this process. E. coli QPl.1/pNR4.272 synthesizes 56 g/L of quinic acid in 20% yield from glucose. Fermentor—controlled conditions for this conversion are amenable to scale-up for larger volume cultivation. Nonetheless, significant improvement in the yield and concentration of quinic acid microbially synthesized from glucose is needed. Oxidation of quinic acid with NaOCl or Ag3PO4/KZSZO3 allows quinic acid to be converted into hydroquinone in high yield and without overoxidation to benzoquinone. Synthesis of hydroquinone from glucose clearly has aspects that need to be further improved before this route can be used on an industn'al scale. However, the microbial synthesis of quinic acid and chemical oxidation of glucose-derived quinic acid described in this chapter moves the synthesis of hydroquinone from glucose from a proof-of-concept conversion to a route that may ultimately supplant currently employed syntheses of hydroquinone where benzene serves as the starting material. There is also an economic opportunity associated with producing chemicals from glucose and other carbohydrates. As declining reserves and increasing consumption lead to continuing increases in the price of petroleum,73 the costs of using petroleum-derived benzene will likewise increase. By contrast, the increasing availability of glucose derived 93 from corn starch and the development of corn fiber and lignocellulose as carbohydrate feedstocks suggest that glucose and other carbohydrates will become an increasingly valuable source of carbon for synthesis of chemicals. As the world moves to national CO2 budgets set by international treaty, hydroquinone synthesized from carbohydrates such as glucose may also be important CO2 credits, given the consideration that the glucose used as a starting material is essentially an immobilized form of C03. 94 Reference ' Krumenacker, L.; Costantini, M.; Pontal, P.; Sentenac, J. In Kirk-OIhmer Encyclopedia of Chemical Technology, 4’” ed.; Wiley: New York, 1995, Vol. 13, p. 996-1014. 2 Brand, T. Hydroquinone set to expand as new uses push demand. Chemical Market Reporter 1997 , 251, 5. 3 (a) Yoshikawa, A.; Yoshida, S.; Terao, I. Microbial preparation of hydroquinone. Bioprocess Technol. 1993, 16, 149-161. (b) Hall, M. C. Microbiological preparation of hydoquinone. European Patent 0 073 134 A2, 1982. (c) Higgins, I. J .; Hammond, R. C.; Sariaslani, F. 8.; Best, D.; Davies, M. M.; Tryhom, S. E.; Taylor, F. Biotransformation of hydrocarbons and related compounds by whole organism suspensions of methane—grown Methylosinus trichosporium OB 3b. Biochem. Biophys. Res. Commun. 1979, 89, 671-677. 4 (a) Barker, J. L.; Frost, J. W. Microbial synthesis of p-hydroxybenzoic acid from glucose. Biotechnol. Bioeng. 2001, 76, 376-390. (b) Amaratunga, M.; Lobos, J. H.; Johnson, B. F.; Williams, D. US. Patent 6030819, 2000. 5 (a) van Berkel, W. J. H.; Eppink, M. H. M.; Middelhoven, W. J .; Vervoort, J.; Rietjens, I. M. Catabolism of 4-hydroxybenzoate in Candida parasilosis proceeds through initial oxidative decarboxylation by a FAD-dependent 4-hydroxybenzoate 1-hydroxylase. FEMS Microbiol. Lett. 1994, 121, 207-215. (b) Eppink, M. H. M.; Boeren, S. A.; Vervoort, J .; van Berkel, W. J. H. Purification and properties of 4-hydroxybenzoate 1- hydroxylase (decarboxylating), a novel flavin adenine dinucleotide-dependent monooxygenase from Candida parapsilosis CBS604. J. Bacteriol. 1997, 179, 6680-6687. (c) Eppink, M. H.; Cammaart, E.; Van, Wassenaar. D.; Middelhoven, W. J .; van Berkel, W. J. Purification and properties of hydroquinone hydroxylase, a FAD-dependent monooxygenase involved in the catabolism of 4-hydroxybenzoate in Candida parapsilosis CBS604. Eur. J. Biochem. 2000, 267, 6832-6840. 6 Draths, K. M.; Ward, T. L.; Frost, J. W. Biocatalysis and nineteenth century organic chemistry: conversion of D-glucose into quinoid organics. J. Am. Chem. Soc. 1992, 114, 9725-9726. 7 Draths, K. M.; Knop, D. R.; Frost, J. W. Shikimic acid and quinic acid: replacing isolation from plant sources with recombinant biocatalysis. J. Am. Chem. Soc. 1999, 121 , 1603-1604. 8 Eliel, E. L.; Ramirez, M. B. (-)-Quinic acid: configurational (stereochemical) descriptors. Tetrahedron: Asymmetr. 1997, 8, 3551-3554. 9 Hofmann, F. C. Crell ’s Chemische Annalen, 1790, 2 314. 95 ‘0 Fisher, H. 0., L.; Dangschat, G. Ber. Dtsch. Chem. Gas. 1932, 65, 1009. ” Phoon, C. W.; Abell, C. Use of quinic acid as template in solid-phase combinatorial synthesis. J. Comb. Chem. 1999, I, 485-492. ‘2 Federspiel, M.; Fischer, R.; Hennig, M.; Mair, H.-J.; Oberhauser, T.; Rimmler, G.; Albiez, T.; Bruhin, J .; Estermann, H.; Gandert, C.; Goeckel, V.; Goetzoe, S.; Hoffmann, U.; Huber, G.; Janatsch, G.; Lauper, S.; Roeckel-Staebler, O.; Trussardi, R.; Zwahlen, A. G. Industrial synthesis of the key precursor in the synthesis of the anti—influenza drug Oseltamivir phosphate (Ro 64-0796/002, GS-4104-02): Ethyl (3R,4S,SS)-4,5-epoxy-3-(1- ethyl-propoxy)-l-cyclohexene-l-carboxylate. Org. Proc. Res. Develop. 1999, 3, 266-274. ‘3 Matsuo, K.; Sugimura, W.; Shimizu, Y.; Nishiwaki, K.; Kuwajima, H. Synthesis of (-)- sugiresinol dimethyl ether utilizing (-)-quinic acid. Heterocycles 2000, 53, 1505-1513. ‘4 Murray, L. M.; O’Brien, P.; Taylor, R. J. K. Stereoselective reactions of a (-)-quinic acid-derived enone: application to the synthesis of the core of scyphostatin. Org. Lett. 2003, 5, 1943-1946. ‘5 Barros, M. T .; Maycock, C. D.; Ventura, M. R. Enantioselective total synthesis of (+)- Eutypoxide B. J. Org. Chem. 1997, 62, 3984-3988. ‘6 Ono, K.; Yoshida, A.; Saito, N.; Fujishima, T.; Honzawa, S.; Suhara, Y.; Kishimoto, S.; Sugiura, T.; Waku, K.; Takayama, H.; Kittaka, A. Efficient synthesis of 2-modified 101,25-dihydroxy-l9-norvitamin D3 with Julia olefination: high potency in induction of differentiation on HL-60 cells. J. Org. Chem. 2003, 68, 7407-7415. '7 Su, Z.; Paquette, L. A. Conversion of D-(-)-quinic acid into an enantiopure C-4 functionalized 2-iodocyclohexenone acetal. J. Org. Chem. 1995, 60, 764-766. ‘8 Ulibarri, G.; Nadler, W.; Skrydstrup, T.; Audrain, H.; Chiaroni, A.; Riche, C.; Grierison, D. S. Construction of the bicyclic core structure of the enediyne antibiotic esperamicin-Al in either enantiomeric form from (-)-quinic acid. J. Org. Chem. 1995, 60, 2753-2761. ‘9 Wang, Z.-X.; Shi. Y. A new type of ketone catalyst for asymmetric epoxidation. J. Org. Chem. 1997, 62, 8622-8623. 2° Haslam, E. In Shikimic Acid: Metabolism and Metabolites; Wiley & Sons: New York, 1993,p.56. 2‘ Bestmann, H. J .; Heid, H. A. Stereospecific synthesis of optically pure quinic acid and shikimic acid from D-arabinose. Angew. Chem, Int. Ed. Engl,. 1971, 10, 336-337. 96 22 Hiroya, K.; Ogasawara, K. A concise enantio- and diastereo-controlled synthesis of (-)- quinic acid and (-)-shikimic acid. Chem. Commun. 1998, 2033-2034. 23 Pittard, J .; Wallace, B. J. Distribution and function of genes concerned with aromatic biosynthesis in Escherichia coli. J. Bacteriol. 1966, 91, 1494-1508. 24 (a) Li, K.; Mikola, M. R.; Draths, K. M.; Worden, R. M.; Frost, J. W. Fed-batch fermentor synthesis of 3-dehydroshikimic acid using recombinant Escherichia coli. Biotechnol. Bioeng. 1999, 64, 61-73. (b) Weaver, L. M.; Hermann, K. M. Cloning of an aroF allele encoding a tyrosine-insensitive 3-deoxy-D-arabino-heptulosonate 7- phosphate synthase. J. Bacteriol. 1990, I 72, 6581-6584. 25 Snell, K. D.; Draths, K. M.; Frost, J. W. Synthetic modification of the Escherichia coli chromosome: enhancing the biocatalytic conversion of glucose into aromatic chemicals. J. Am. Chem. Soc. 1996, 118, 5605-5614. 26 Ran, N.; Knop, D. R.; Draths, K. M.; Frost, J. W. Benzene-free synthesis of hydroquinone. J. Am. Chem. Soc. 2001, 123, 10927-10934. 27 Anton, I. A.; Coggins. J. R. Sequencing and overexpression of the Escherichia coli aroE gene encoding shikimate dehydrogenase. Biochem. J. 1988, 249, 319-326. 28 Dell, K. A.; Frost, J. W. Identification and removal of impediments to biocatalytic synthesis of aromatics from D-glucose: rate-limiting enzymes in the common pathway of aromatic amino acid biosynthesis. J. Am. Chem. Soc. 1993, 115, 11581-11589. 29 Farabaugh, M. M. S. Thesis, Michigan State University, 1996. 3° (a) Draths, K. M.; Pompliano, D. L.; Frost, J. W.; Berry, A.; Disbrow, G. L.; Staversky, R. J .; Lievense, J. C. Biocatalytic synthesis of aromatics from D-glucose: the role of transketolase. J. Am. Chem. Soc. 1992, 114, 3956-3962. (b) Patnaik, R.; Liao, J. C. Engineering of Escherichia coli central metabolism for aromatic metabolite production with near theoretical yield. Appl. Environ. Microbiol. 1994, 60, 3903-3908. 3‘ Knop, D. R. Doctorate Dissertation, Michigan State University, 2002. 32 Chaudhuri, S.; Anton, I. A.; Coggins, J. R. Shikimate dehydrogenase from Escherichia coli. Methods Enzymol. 1987, 142, 315-320. 33 Zhu, L. Y.; Frost, J. W. Unpublished result. 3" Postma, P. W.; Lengeler, J. W.; Jacobson, G. R. Phosphoenolpyruvate:carbohydrate phosphotransferase systems. In Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd ed.; Neidhardt, F. C., Ed.; ASM press: Washinfton, DC, 1996: p. 1149-1174. 97 35 (a) Forberg, C.; Eliaeson, T.; Haggstrom, L. Correlation of theoretical and experimental yields of phenylalanine from non-growing cells of a rec Escherichia coli strain. J. Biotech. 1988, 7, 319-322. (b) Frost, J. W.; Draths, K. M. Biocatalytic syntheses of aromatics from D-glucose: renewable microbial sources of aromatic compounds. Ann. Rev. Microbial. 1995, 49, 557-579. 36 Patnaik, R.; Spitzer, R. G.; Liao, J. C. Pathway engineering for production of aromatics in Escherichia coli: confirmation of stoichiometric analysis by independent modulation of AroG, TktA, and Pps activities. Biotechnol. Bioeng. 1995, 46, 361—370. 37 Yi, J.; Li, K.; Draths, K. M.; Frost, J. W. Modulation of phosphoenolpyruvate synthase expression increases shikimate pathway product yields in E. coli. Biotechnol. Prog. 2002, 18,1141-1148. 38 Chandran, S. C.; Yi, J.; Draths, K. M.; von Daeniken, R.; Weber, W.; Frost, J. W. Phosphoenolpyruvate availability and the biosynthesis of shikimic acid. Biotechnol. Prog. 2003, 19, 808-814. 39 Parker, C.; Barnell, W. 0.; Snoep, J. L.; Ingram, L. O.; Conway, T. Characterization of the Zymomonas mobilis glucose facilitator gene product (glf) in recombinant Escherichia coli: examination of transport mechanism, kinetics and the role of glucokinase in glucose transport. Mol. Microbiol. 1995, 15, 795-802. 4° (a) Saier, M. H. Jr.; Bromberg, F. G.; Roseman, S. Characterization of constitutive galactose permease mutants in Salmonella typhimurium. J. Bacteriol. 1973, 113, 512- 514. (b) Flores, N .; Xiao, J.; Berry, A.; Bolivar, F.; Valle, F. Pathway engineering for the production of aromatic compounds in Escherichia coli. Nat. Biotechnol. 1996, 14, 620- 623. ‘" Levy, 8.; Zeng, G.; Danchin, A. Cyclic AMP synthesis in Escherichia coli strains bearing known deletions in the pts phosphotransferase operon. Gene 1990, 86, 27-33. 42 Tian, F. Doctorate Dissertation, Michigan State University, 1998. 43 Montchamp, J .-L.; Tian, F.; Hart, M. E.; Frost, J. W. 2,3-Butane bisacetal protection of vicinal diequatorial diols. J. Org. Chem. 1996, 61, 3897-3899. 44 Sann, C. L.; Abell, C.; Abell, A. D. A conventient method for the synthesis of dehydroquinic acid. Synth. Commun. 2003, 33, 527-533. 45 Knop, D. R.; Draths, K. M.; Chandran, S. S.; Barker, J. L.; von Daeniken, R.; Weber, W.; Frost, J. W. Hydroaromatic equilibration during biosynthesis of shikimic acid. J. Am. Chem. Soc. 2001, 123, 10173-10182. 98 46 Kimata, K.; Takahashi, H.; Inada, T.; Postma, P.; Aiba, H. CAMP receptor protein- cAMP plays a crucial role in glucose-lactose diauxie by activating the major glucose transporter gene in E. coli. Proc. Natl. Acad. Sci. USA 1997, 94, 12914-12919. 47 (a) Erni, B.; Zanolari, B. Glucose-permease of the bacterial phosphotransferase system. Gene cloning, overproduction, and amino acid sequence of enzyme 116'“. J. Biol. Chem. 1986, 261 , 16398-16403. (b) Bouma, C. L.; Meadow, N. D.; Stover, E. W.; Roseman, S. II-BG'°, a glucose receptor of the bacterial phosphotransferase system: Molecular cloning of ptsG and purification of the receptor from an overproducing strain of Escherichia coli. Proc. Natl. Acad. Sci. USA 1987, 84, 930-934. 48 (a) Donnelly, M. I.; Sanville-Millard, C.; Chatterjee, R. US Patent, 6,159,738, 2000. (b) Nichols, N. N .; Dien, B. S.; Bothast, R. J. Use of catabolite repression mutants for fermentation of sugar mixtures to ethanol. Appl. Microbiol. Biotechnol. 2001, 56, 120- 125. 49 Zaldivar, J.; Martinez, A.; Ingram, L. 0. Effect of alcohol compounds found in hemicellulose hydrolysate on the growth and fermentation of ethanologenic Escherichia coli. Biotechnol. Bioeng. 2000, 68, 524-530. 5° Ogino, T.; Garner, C.; Markley, J. L.; Herrmann, K. M. Biosynthesis of aromatic compounds: 13C NMR spectroscopy of whole Escherichia coli cells. Proc. Natl. Acad. Sci. USA. 1982, 79, 5828-5832. 5‘ Woskresensky, A. Justus Liebiegs Ann. Chem. 1838, 27, 257. 52 Yahuka, Y.; Katz, R.; Sarel, S. Bile acid chemistry. III. Stepwise side-chain shortening by way of sodium per-iodate oxidation of a-hydroxy bile acids into corresponding aldehydes. Tetrahedron Lett. 1968, 1725-1728. 53 Bhale, V. M.; Sant, P. G.; Bafna, S. L. Acid permanganate oxidation of tartaric acid. J. Sci. Ind. Res. 1956, 153, 45-46. 5‘ Bakore, G. V.; Narian, S. Kinetic investigation of the oxidation of some a-hydroxy- carboxylic acids by chromic acid. J. Chem. Soc. 1963, 3419-3424. 55 Sheldon, R. A.; Kochi, J. K. Oxidative decarboxylation of acids by lead tetraacetate. Org. React. 1972, 19, 279-421. 56 Carlsen, P. H. J. Oxidation of a-hydroxy carboxylic acid with sodium hypochlorite. Acta Chem. Scand. B. 1984, 38, 343-344. 57 (a) Carlsen, P. H. J.; Misund, K.; Roe, J. Synthesis of optical active glyceric acid derivatives from ascorbic acids. Acta Chem. Scand. 1995, 49, 297-300. (b) Stevens, R. 99 V.; Chapman, K. T.; Stubbs, C. A.; Tam, W. W.; Albizati, K. F. Further studies on the utility of sodium hypochlorite in organic synthesis. Selective oxidation of diols and directed conversion of aldehydes to esters. Tetrahedron Lett. 1982, 23, 4647-4650. 58 Nwaukwa, S. O.; Keehn, P. M. Oxidative cleavage of a-diols, a-diones, a- hydroxyketones and a-hydroxy- and a-keto acids with calcium hypochlorite [Ca(OCl)3]. Tetrahedron Lett. 1982, 23 , 3135-3138. 59 Elmore, P. R.; Reed, R. T.; Terkle-Huslig, T.; Welch, J. S.; Young, S. M.; Landolt, R. G. Hypochlorite-induced oxidative decarboxylation of trisustituted acetic acids. J. Org. Chem. 1989, 54, 970-972. 60 (a) Barros, M. T .; Maycock, C. D.; Ventura, M. R. The effect of DMSO on the borohydride reduction of a cyclohexanone: a formal enantioselective synthesis of (+)- epibatidine. Tetrahedron Lett. 1999, 40, 557660 (b) Kamenecka, T. M.; Overman, L. E. An enantioselective approach to the synthesis of manzamine A. Tetrahedron Lett. 1994, 35, 4279-4282. 6' (a) Audia, J. E.; Boisvert, L.; Patten, A. D.; Villalobos, A.; Danishefsky, S. J. Synthesis of two useful, enantiomerically pure derivatives of (S)-4-hydroxy-2-cyclohexenone. J. Org. Chem. 1989, 54, 3738-3340. (b) Trost, B. M.; Romero, A. G. Synthesis of optically active isoquinuclidines utilizing a diastereoselectivity control element. J. Org. Chem. 1986, 51, 2332-2342. 62 (a) Renaud, P.; Hurzeler, M.; Seebach, D. Electrochemical oxidation of (S)-malic-acid- derivatives: a route to enantiomerically pure alkylmalonaldehydic esters. Helv. Chim. Acta 1987, 70, 292-298. (b) Baizer, M. M. Recent developments in organic synthesis by electrolysis. Tetrahedron 1984, 40, 935-969. 63 Blay, G.; Fernandez, 1.; Formentin, P.; Pedro, J. R.; Rosellé, A. L.; Ruiz, R.; Journaux, Y. Catalytic aerobic oxidative decarboxylation of a—hydroxy-acids. Methyl mandelate as a benzoyl anion equivalent. Tetrahedron Lett. 1998, 39, 3327-3330. 64 (a) Toussaint, 0.; Capdevielle, P.; Maumy, M. Oxidative decarboxylation of aliphatic carboxylic acids by the copper(I)/oxygen system. Tetrahedron Lett. 1984, 25, 3819-3822. (b) Barak, G.; Dakka, J .; Sasson, Y. Selective oxidation of alcohols by a HzOz-RuCl3 under phase-transfer conditions. J. Org. Chem. 1988, 53, 3553-3555. (c) Goebel, W. F. Derivatives of citraconic acid: I. the synthesis of methyltartaic acid and the decomposition of the dihydroxymaleic acid. J. Am. Chem. Soc. 1925, 47, 1990-1998. 65 Chaudhuri, S.; Coggins, J. R. The purification of shikimate dehydrogenase from Escherichia coli. Biochem. J. 1985, 226, 217-223. 66 Whittington, H. A.; Grant, S.; Roberts, C. F.; Lamb, H.; Hawkins, A. R. Identification 100 and isolation of a putative permease gene in the quinic acid utilization (QUT) gene cluster of Aspergillus nidulans. Curr. Genet. 1987, 12, 135-139. 67 Michel, G.; Roszak, A. W.; Sauve, V.; Maclean, J.; Matte, A.; Coggins, J. R.; Cygler, M.; Lapthom, A. J. Structure of shikimate dehydrogenase AroE and its paralog YdiB. J. Biol. Chem. 2003, 278, 19463-19472. 68 Benach, J .; Lee, 1.; Edstrom, W.; Kuzin, A. P.; Chiang, Y.; Acton, T. B.; Montelione, G. T.; Hunt, J. F. The 2.3-A crystal structure of the shikimate 5-dehydrogenase orthologue YdiB from Escherichia coli suggests a novel catalytic environment for a NAD-dependent dehydrogenase. J. Biol. Chem. 2003, 278, 19176-19182. 69 (a) Ho, T.-L. Ceric ion oxidation in organic chemistry. Synthesis 1973, 347-354. (b) Amjad, Z.; McAuley, A.; Gomwalk, U. D. Metal-ion oxidations in solution. Part 16. The oxidation of a-hydroxycarboxylic acids by cerium(IV) in perchloric acid media. Chem. Soc. Dalton Trans. 1977, 82-88. (c) Jones, J. R.; Waters, W. A.; Littler, J. S. Oxidation of organic compounds with quinquevalent vanadium. VII. The kinetic resemblance between oxidations of some a-hydroxy acids and that of pinacol and of its monomethyl ether. J. Chem. Soc. 1961, 630-633. (d) Kalidoss, P.; Srinivasan, V. S. Estimated distribution between one- and two-equivalent reaction paths in vanadium(V)-induced electron transfer. Chem. Soc. Dalton Trans. 1984, 2631-2635. 7° (a) Anderson, J. M.; Kochi, J. K. Silver(I)-catalyzed oxidative decarboxylation of acids by peroxydisulfate. Role of Silver(II). J. Am. Chem. Soc. 1970, 92, 1651-1659. (b) Walling, C.; Camaioni, D. M. Role of Silver(II) in silver-catalyzed oxidations by peroxydisulfate. J. Org. Chem. 1978, 43, 3266-3271. (c) Kumar, A.; Neta, P. Complexation and oxidation of glycine and related compounds by Ag (11). J. Am. Chem. Soc. 1980, 102, 7284. (d) Minisci, F.; Citterio, A.; Giordano, C. Electron-transfer processes: peroxydisulfate, a useful and versatile reagent in organic chemistry. Acc. Chem. Res. 1983, 16, 27-32. 7‘ (a) O’Connor, S. R.; Farmer, P. B.; Lauder, I. Benzene and non-Hodgkin's lymphoma. J. Pathol. 1999, 189, 448-453. (b) Farris, G. M.; Everitt, J. I.; Irons, R.; Popp, J. A. Carcinogenicity of inhaled benzene in CBA mice. F undam. Appl. Toxicol. 1993, 20, 503- 507. (c) Huff, J. E.; Haseman, J. K.; DeMarini, D. M. Multiple-site carcinogenecity of benzene in Fischer 344 rats and B6C3F1 mice. Environ. Health Perspect. 1989, 82, 125- 163. 72 Trade Secrets, A Moyers Report. http://www.pbs.org/tradesecrets/transcripthtml. 73 httpz/lwww.eia.doe.gov/emeu/steo/pub/contents.html Short-term energy outlook- February 2004. 101 CHAPTER THREE Analysis of Gallic Acid Biosynthetic Pathway in Escherichia coli with [5-‘80]-3-Dehydroshikimic Acid Introduction Gallic acid (3,4,5-trihydroxybenzoic acid) is a central precursor for hydrolysable tannins, the abundant secondary metabolites in higher plants. Commercial availability of gallic acid relies on its isolation and hydrolysis of tannins from insect carapices (gall nuts) harvested in China, and from the seed pods of Coulteria tinctoria tree found in Peru.1 Gallic acid and its esters have found a diverse range of industrial uses, such as antioxidants in food and cosmetics, and used as a material for inks and paints.1 In the pharmaceutical industry, gallic acid is used for the synthesis of antibiotic trimethoprim (Figure 44).1 Thermal decarboxylation of gallic acid in copper autoclaves afforded pyrogallol (1,2,3-trihydroxybenzene, Figure 44), which is used in the production of dyes and photographic developers and also in laboratories for absorbing oxygen.l O O\/\ OH (10” HO OH OH OCH3 OH trimethoprim propyl gallate pyrogallol Figure 44. Industrial chemicals derived from gallic acid. 102 Tannins, together with lignins, are the most widespread and abundant polyphenols in plants.2 The involvement of tannins in the protection of plants against insert and mammalian herbivores is well established.3 Tannins can form insoluble complexes with proteins that subsequently reduce the feeding value of the plant. Similarly, commercial uses of tannins in the leather industry are also based on their binding with proteins.4 Plant tannins are classified as condensed tannins and hydrolysable tannins. The latter is characterized by a central fi-D—glucose moiety whose hydroxy] groups are esterified with one to five gallic acid molecules. The fully galloylated glucose derivative, 1,2,3,4,6-O- pentagalloylglucose, is regarded as an immediate precursor to gallotannins and ellagitannins, two subclasses of hydrolysable tannins.2b The initial steps in enzymatic synthesis of a wide range of complex gallotannins and ellagitannins are comprised of series of reactions from gallic acid and UDP-glucose to 1,2,3,4,6-O-pentagalloyglucose via l-O-galloylglucose 4 (Figure 45).2b’5 l,2,3,4,6-O- Pentagalloylglucose can be converted to gallotannins by the addition of as many as 10-12 additional galloyl residues to the 1,2,3,4,6-O-pentagalloylglucose core by using l-O- galloylglucose 4 as galloyl donor. Ellagitannins are formed by oxidative processes that yield C-C linkages between adjacent galloyl groups of the 1,2,3,4,6-O- pentagalloylglucose, followed by subsequent formation of dimeric and oligomeric derivatives that are connected via C-C or C-O-C bonds between the galloyl residues?!"5 103 OH O stew- COZH UDP- -glucose -§%O > H HO OH 90 l OH UDP glucose gallic acid 1-0—galloylglucose (4) HO OH HO OH OH OH HO 00 0 OH OH OH O 0 OH HO 0 0 o O OH inflow “4* 0 0 OH OH O HO OH 1 ,6-O-digalloylglucose 1 ,2,3,4,6-O-pentagalloylglucose HO HO OH OH Figure 45. Biosynthesis of l,2,3,4,6-O-pentagalloylglucose. Biosynthesis of gallic acid, the principle phenolic unit of hydrolysable tannins, is not yet fully clarified. In addition to higher plants, gallic acid is also observed in cultures 6 7 of Phycomyces blakesleeanus, Pseudomonas fluorescens,7 Entorobacter cloacae, Aspergillus terreus8 and recombined E. coli.9 Three alternative pathways leading to gallic acid have been proposed as illustrated in Figure 46.10 104 H203PO\/'\/"\ H aroF aroG H04 COZH E4P 6” aroH aroB Q —> —> C 02H H203PO OH 3d h (:H . . PEP DAHP " e geirdOQUII’IIC 00le /COZH COZH OH 3-dehydrgshikimic L-phenylalanine caffeicH acid 30 . aroZ/ A ‘ /CO:H COZH COzH protocaotechuic gallic acid 3,4,5-trihydroxy acud cmnamlc aCId Figure 46. Three proposed pathways for biosynthesis of gallic acid. Pathway A, via 3- dehydroshikimic acid; Pathway B, via protocatechuic acid; Pathway C, via caffeic acid derived from L-phenylalanine. In the proposed pathways A and B, gallic acid is biosynthesized from 3- dehydroshikimic acid. 3-Dehydroshikimic acid could be enzymatically converted into gallic acid by oxidoreductase-catalyzed dehydrogenation followed by subsequent aromatization (Pathway A) or dehydratase-catalyzed conversion to protocatechuic acid (PCA) followed by hydroxylation of the protocatchuic acid to form gallic acid (Pathway 105 B). On the other hand, gallic acid could be derived from L-phenylalanine via caffeic acid intermediacy involving side-chain degradation of 3,4,5-trihydroxycinnamic acid (Pathway C), since it is generally accepted that benzoic acids are produced in higher plants by side-chain degradation of cinnamic acids.” Ambiguous results regarding these alternative pathways were obtained after feeding carboxyl-MC-labeled shikimic acid to Rhus and Acer leaves.10c Experimental evidence that favored the directed biosynthesis of gallic acid from 3-dehydroshikimic acid in higher plants came from induction experiments with the herbicide glyphosate12 (N-(phosphonomethyl)glycine). Glyphosate inhibits a shikimate pathway enzyme, 5- enolpyruvylshikimate-3-phosphate synthase, and causes a reduction in the synthesis of aromatic amino acids and phenylpropanoids. However, the level of gallic acid was higher in higher plants treated with glyphosate, which indicated that gallic acid was derived from a shikimate pathway intermediate before shikimate 3-phosphate, the substrate for 5-enolpyruvylshikimate-3-phosphate synthase. Strong evidence supporting this View has been obtained by feeding l3C-glucose to cultures of fungus Phycomyces blakesleeanus and to young leaves of dicotyloneous tree Rhus typhina, followed by NMR spectroscopic analyses of isotope distributions of the isolated gallic acid and aromatic amino acids.13 Cultures of P. blakesleeanus were grown in medium containing [1-‘3C]-glucose (99.5% l3C enrichment) as carbon source. Gallic acid and aromatic amino acids L- phenylalanine, L-tyrosine and L-tryptophan were isolated from culture media and cell mass. Figure 47 showed the abbreviated conversions of glucose to L-phenylalanine, L- tyrosine and L-tryptophan and their respective l3C-enrichment on each carbon atom. Due 106 to the symmetry of the aromatic rings of gallic acid, L-phenylalanine and L-tyrosine, the ring carbon atom at 2/6 and 3/5 yielded average l3C abundances, although they had different biosynthetic origins. The 13'C-labeling pattern of the aliphatic side-chain that was derived from phosphoenolpyruvate was used to reconstruct the labeling pattern of phosphoenolpyruvate. On the other hand, L-tryptophan reflected the 13C distributions of the original shikimate ring system, and the labeling pattern of D-erythrose 4-phosphate can be reconstructed based on the L-tryptophan biosynthetic pathway. The labeling patterns of phosphoenolpyruvate, D-erythrose 4-phosphate and L- phenylalanine were used to predict the labeling patterns of gallic acids biosynthesized via intermediacy of cafferic acid, 3-dehydroshikimic acid, or protoctechuic acid. As shown in Figure 48, the observed gallic acid labeling pattern of gallic acid closely agreed with the predictedlabeling patterns for biosynthesis involving intermediacy of 3- dehydroshikimic acid or protocatechuic acid, but not with the predicted labeling patterns for biosynthesis proceeding via the phenylpropanoid pathway.14 Labeling experiments with a mixture of [U-‘3C]-glucose and unlabeled glucose on young leaves of R. typhina also agreed with the hypothesis that gallic acid was derived from 3-dehydroshikimic acid or protocatchuic acid.13 This ruled out L-phenylanine as major precursor for gallic acid biosynthesis. However, the intermediacy of protocatechuic acid still remained an open question due to the fact that protocatechuic acid had the same carbon skeleton as 3- dehydroshikimic acid. Earlier work with cell-free systems from mung bean seedlings15 and leaves from Pelagonium inquinan16 suggested the involvement of the protocatechuic acid in biosynthesis of gallic acid, but these preliminary results were never corroborated by detailed investigations. 107 OH COZH a —*—-> JL HO OH OH [1 -‘3C]-glucose chorismic acid V \° COZH H020, COZH ' NH2 0 OH prephenic acid anthranilate 1 .6 COQH 002H 1 .9 NH2 23.4 NH2 0" 1 .1 C-6 C-2 20.2 + 05 C-3 2.4 c-4 5.4 L-phenylalanine 5 a 18.5' - PEP 54.: OH Figure 47. 13C-Abundance (in percentage) of L-phenylalanine, L-tyrosine and L-tryptophan from P. blakesleeanus cultured with [1-‘3C1-glucose. Keys: (a) see Figure 7, Chapter 2; (b) chorismate mutase (pheA); (c) anthranilate synthase (trpD, trpE); (d) L-phenylalanine and L-tyrosine biosynthesis enzymes;l7 (e) L-tryptophan biosynthesis enzymes.17 The labeling pattern of phosphoenolpyruvate (PEP) and D-erythrose 4- phosphate (E4P) were reconstructed from aromatic amino acids based on established mechanisms of the shikimate pathway. 108 COZH 2 0 1 5 2. 6 23 4 /\OPO3H2 H203P?8\5}\/U\H PEP E4P 0“ ll C02H 1.9 23 4 NH2 COQH COZH 1 1 20 2 2 4 O OH HO 5'4 OH OH L-phenylalanine DHS PCA 23.4 002H 0902H 20902H 1.1 2. 0 20.2 HOEEng mfi1>22232 2.4 H0 5.4 OH 1 5 0H 0.2 H09 OH predicted via L-phenylalanine predicted via DHS 0r PCA observed Figure 48. Predicted and observed labeling patterns of gallic acid synthesized by P. blakesleeanus cultured with [1-13C]-glucose. Distinguishing the two pathways can be achieved using [5-‘80]-3- dehydroshikimic acid (Figure 49). If gallic acid was biosynthesized via the dehydrogenation of 3-dehydroshikimic acid, the resulting gallic acid would contain an 18O-labeled oxygen atom, which can be determined by mass spectrometry. If gallic acid was formed via dehydration to protocatechuic acid followed by hydroxylation of protocatechuic acid with 02, the product gallic acid would not contain 18O-labeled hydroxyl group (Figure 49). 109 COZH HO 18OH OH A [3-‘801-gallic acid COZH 9:“m WW 0“ Y COZH COZH [5-‘801-0Hs B .——.> HO HO OH OH OH protocatechuic gallic acid acid Figure 49. Strategy for distinguishing proposed gallic acid biosynthetic pathways with [5-1801-3-dehydroshikimic acid. A. Oxidoreductase-catalyzed oxidation of 3- dehydroshikimic acid; B. Hydroxylase-catalyzed hydroxylation of protocatechuic acid. Synthesis of [5-1801-3-dehydroshikimic acid. The [5-‘8O]-3-dehydroshikimic acid was synthesized from shikimic acid via a chemoenzymatic route (Figure 50). Berchtold’s procedure was adapted for the synthesis of the hydroxyl epoxide 21.l8 Acid catalyzed epoxide opening19 with 2 equivalents HZ'SO provided the [5-‘8O]-shikimate methyl ester 22, which was purified by reverse phase HPLC method in 37% yield. Base-catalyzed hydrolysis converted 22 to [5-‘8O]-shikimic acid. Conversion of [5-‘80]-shikimic acid to [5-‘8O]-3-dehydroshikimic acid was carried out by an enzymatic oxidation using shikimate dehydrogenase and N ADP”. Shikimate dehydrogenase was purified from E. coli strain ABZ834/pIA321 that contained a plasmid-localized aroE encoding shikimate dehydrogenase as previously described.20 To reduce the cost of the cofactor, the NADP+ required for oxidation of 110 [5-‘8O]-shikimic acid was regenerated by reductive amination of a-ketoglutarate with NADPH and ammonia catalyzed by L-glutamate dehydrogenase.21 Excess amounts of a- ketoglutarate and L-glutamate dehydrogenase were used to drive the reaction to completion. Subsequent purification of the reaction mixture using an anion exchange column (Bio-Rad AG-l-X8, acetic form) gave the [5-'BO]-3-dehydroshikimic acid in 84% yield (Figure 50). COZH COZCHa COZCH3 C. __,. O -->“ fl Ho“' _,_‘ OH H0“. 75 HO“. é 18OH OH OH shikimic acid 21 [5-‘801-22 COQH COZH ——>° O. " £1 How i 180H 0 _5_ 180H OH OH [5-‘8O]-shikimic acid [5-‘BO]-DHS Figure 50. Synthesis of [5-‘8O]-3-dehydroshikimic acid from shikimic acid. Keys: (a) i) CH3OH, Amberlite 1R120 (H*), reflux, 24 h, 100%; ii) PPh3, DEAD, THF, 0 °C—rt., 1.5 h, 80%; (b) H2'80 (2 equiv.), CF3SO3H (0.1 equiv.), CH3CN, 0 °C—rt. 48 h, 37%; (c) NaOH, THF/H20 (1:1, v/v), rt., 5 h, 100%; (d) shikimate dehydogenase, NADP*, a- ketoglutaric acid, NH4C1, L-glutamate dehydrogenase, potassium phosphate buffer (pH 8.0), rt., 5 h, 84%. Biosynthesis of gallic acid in E. coli. In addition to higher plants and fungi, recombinant E. coli constructs are capable of biosynthesis of gallic acid under fermentor-controlled conditions.9 Gallate- synthesizing E. coli KL3 was constructed from E. coli AB2834aroE by homologous 111 recombination of an aroB gene into the chromosomal serA locus.921 The absence of aroE- encoded shikimic dehydrogenase activity that catalyzes the conversion of 3- dehydroshikimate to shikimate resulted in the accumulation of 3-dehydroshikimate in culture supernatant. E. coli KL3 expressing plasmid-localized aroFFBR, which encodes a feedback-insensitive isozyme of DAHP synthase, biosynthesized gallic acid along with 3- dehydroshikimic acid during fermentation runs in a ratio of around 1:6 - 1:11 (mol gallic acid/mol 3-dehydroshikimic acid). However, no enzyme activity that catalyzed the conversion of 3-dehydroshikimic acid to gallic acid was detected in E. coli crude cell extracts.22 The lack of assayable activity might suggest that the responsible enzyme was labile or 3-dehydroshikimic acid was not the immediate precursor. [5-180]-3- Dehydroshikimic acid was then used to determine unequivocally whether the gallic acid biosynthetic pathway proceeded directly from 3-dehydroshikimic acid or via protocatechuic acid intermediacy. [5-180]-3-Dehydroshikimic acid (0.75 g) was mixed with unlabeled 3- dehydroshikimic acid (10.25 g) to prepare [5-180]-enriched 3-dehydroshikimic acid. The levels of 18o-em-icluiient in the [5-180]-3-dehydroshikimic acid and the isolated product gallic acid were determined by fast atom bombardment (FAB) mass spectrometry. Although 13C NMR analysis of 18O incorporation is common in natural product biosynthesis studies,23 the relatively low 18O-enrichment levels used in this study and difficulties encountered on quantitative interpretation prevented its utility in this study. Mass spectrometry permited direct observation and quantitative determination of the 18O- labeling in gallic acid. However, direct measurement of 18O-enrichment level in 3- dehydroshikimic acid using F AB(-) mass spectrometry was unsuccessful. Converting 3- 112 dehydroshikimic acid to shikimic acid alleviated the interference but still gave an unsatisfactory mass spectrum due to a weak molecular ion signal. To circumvent this problem, 3-dehydroshikimic acid was first chemically oxidized to gallic acid by Cu(OAc)2 using a previously established method,24 and then its 18O-enrichment was measured by FAB(-) mass spectrometry to represent the 18O-enrichment level in 3- dehydroshikimic acid. Entry 2 in Table 6 shows ion abundances at m/z 169, m/z 170, and m/z 171 from the FAB(-) mass spectrometry data (100:12.00:8.65). The ion signal at m/z 169 was the molecular ion peak (C7H505, M-H*), which was set to 100. The ion signal at m/z 170, (M-H++1), corresponds to ions containing one 13C atom (C6‘3CHSOS). The ion peak at m/z 171, (M-H*+2), is contributed by ions containing either one 180 atom or ions bearing two 13C atoms (C7HSO4‘SO and C5‘3C2H505). For unlabeled gallic acid, its ratios of the abundance of the molecular ion (M-H+) at m/z 169 to (M-H++1) peak at m/z 170 to (M- H++2) signal at m/z 171 were 100:11.74:2.21 (entry 1, Table 6) and was used as background. To calculate 18O-enrichment in gallic acid, the background contributions due to naturally occurring 180 atoms and ions bearing two l3C atoms were subtracted from isotope peaks at m/z 171. The excess 18O-derived portion at m/z 171 relative to the total (sum of ion intensities at m/z 169, 170 and 171) was calculated as the ‘80- enrichment. Therefore, the 18O-enrichment in the gallic acid corresponding to the prepared [5-‘8O]-enriched 3-dehydroshikimic acid was 5.34% ((8.65- 2.21)/(100+12.9+8.65)). 113 Table 6. 18O-Enrichment determined by FAB(-) mass spectrometry. Ion distribution (average of three) 180 18O entry compound m/z m/z m/z enrichment incorporation 169 170 171 1 GA 100 1 1.74 2.21 - _ (unlabeled) 2 DHS added 100 12.00 8.65 5.34% 100% (”O-enriched) 3 GA 100 12.92 8.96 5.54% 104% (E. coli KL3/pRC1.55B) 4 GA 100 11.76 2.32 0.10% 1.9% (E. coli KL7/pSK6.76) 5 GA 100 12.02 7.47 4.40% 82.4% (air oxidation) 6 DHS recovered 100 l 1.71 8.46 5.20% 97.4% (E. coli KL3/pRC1.55B) Abbreviations: 3-dehydroshikimic acid (DHS); galic acid (GA). The [5-lSO]-enriched 3-dehydroshikimic acid was then added to a culture medium of E. coli KL3/pRC1.55B with glucose addition under fermentor-controlled conditions. Plasmid pRCl.SSB contained a serA gene encoding 3-phosphoglycerate dehydrogenase, a critical enzyme for biosynthesis of L-serine. Incorporation of a serA gene complemented the chromosomal serA inactivation and thus enabled E. coli KL3/pRC1.55B to grow in minimal salts medium without L-serine supplementation. Disruption of the aroE gene also interrupted the biosyntheses of aromatic amino acids and aromatic vitamins. Therefore, growth of E. coli KL3/pRC1.55B required supplementation with L-tyrosine, L-tryptophan, L-phenylalanine, p-hydroxybenzoic acid, p-aminobenzoic acid and 2,3-dihydroxybenzoic acid in culture medium. Another purpose of adding aromatic amino acids into the culture medium was to inhibit native DAHP synthases encoded by aroF, aroG and aroH loci in the E. coli KL3/pRC1.55B 114 genome. The effective inhibition of DAHP synthases was critical in this study due to interference from newly synthesized unlabeled 3-dehydroshikimic acid from glucose. A shiA-encoded transport protein responsible for transporting 3-dehydroshikimate and shikimate in E. coli has been identified.25 Expression of the shiA gene is not regulated by TyrR repressor protein. Therefore, the [5-180]-enriched 3-dehydroshikimate could be transported into the cytoplasm in the presence of aromatic amino acids in the culture medium. E. coli KL3/pRC1.55B was cultivated in 100 mL of M9 medium supplemented with aromatic amino acids and aromatic vitamins for 10-12 h at 37°C with agitation at 250 rpm and then transferred to a fermentor which contained 700 mL of culture medium when its OD600 reached 2-3. Cultivation under fermentor-controlled conditions was performed in a 2.0 L capacity Biostat MD B-Braun fermentor. The temperature was maintained at 36°C, and the pH at 7.0 by addition of concentrated NH4OH or 2 N H2804. Dissolved oxygen was maintained at 20% air saturation throughout the fermentation process. [5-'BO]-Enriched 3-dehydroshikimic acid (5.0 g) was added to the culture at 20 h when the fermentation culture entered the stationary growth phase of its growth. Aromatic amino acids L-tyrosine (0.7 g), L-phenylalanine (0.7 g) and L-tryptophan (0.35 g) were added at the beginning of the fermentation as required for growth and at 18 h and 30 h to inhibit the native DAHP synthase activity. The effective inhibition of native DAHP synthase by supplementing aromatic amino acids was confirmed by the absence of shikimate pathway metabolites in E. coli KL3/pRC1.55B culture medium in a separate experiment without addition of 3-dehydroshikimic acid under fermentor-controlled conditions. 115 During the fermentation, the concentration of 3-dehydroshikimic acid slowly decreased, gallic acid (0.39 g/L) and 3-dehydroquinic acid (DHQ, 0.89 g/L) were produced at 42 h (Figure 51). After removal of cells and proteins, the supernatant (450 mL) was extraction with ethyl acetate, the organic phase was dried and gallic acid (0.0462 g) and a small amount of protocatechuic acid (0.0082 g) were isolated by a separation employing reverse phase HPLC. Compared with the 18O-enrichment (5.34%) in the added 3-dehydroshikimic acid (entry 2, Table 6), the level of 18O-enrichment (5.54%) in the isolated gallic acid (entry 3, Table 6) clearly indicates that the resulting 3- dehydroshikimic acid was converted into gallic acid with retention of 18O-labeled oxygen. 7 45 6+ — $40 0 A 35 d 5JL 1. _. F— A 22 .- 30 E. z? 4.. a .25 5’3 O . g <' . -20 _ (9 3i 3’ as .. 15 I 2-~ . E D 10 1‘ i1'5 04—. r I J1— -0 0 12 18 24 30 36 42 time (h) Figure 51. Gallic acid and 3-dehydroquinic acid production upon addition of [5- 18O]-enriched 3-dehydroshikimic acid by E. coli KL3/pRC1.55B. 3-Dehydroshikimic acid (DHS, open bar), gallic acid (GA, black bar), 3-dehydroquinic acid (DHQ, gray bar), dry cell mass (0). 116 The remaining 3-dehydroshikimic acid in fermentation broth was also extracted with ethyl acetate from fermentation supernatant using a continuous liquid-liquid extraction apparatus. The 18O-enrichment of 5.20% (entry 6, Table 6) indicated that the l8O-labeling in 3-dehydroshikimic acid was stable during the fermentation. An alternative biocatalytic route for converting glucose to gallic acid has been reported that utilized the K. pneumoniae aroZ-encoded DHS dehydratase26 to dehydrate 3-dehydroshikimic acid to protocatechuic acid followed by hydroxylation of protocatechuic acid by the P. aeruginosa pobA*-encoded p-hydroxybenzoate hydroxylase (Figure 52).27 E. coli KL7/pSK6.161 synthesized 20 g/L of gallic acid in 12% yield from glucose under fermentor-controlled conditions. E. coli KL7 was constructed by homologous recombination of an aroBaroZ insert into chromosomal serA locus in the aroE strain AB3248. Plasmid pSK6.161 contained p0bA*, aroFFBR and serA genes required for biosynthesis of gallic acid.24 DHS dehydratase is present in Neurospora crassa28 and other microorganisms as well.29 Isotopic labeling studies showed that DHS dehydratase from N. crassa catalyzed the syn elimination of water from C-5 hydroxyl group of 3-dehydroshikimic acid in which the hydroxyl oxygen atom at C-5 was lost.” Mutagenesis of the pobA gene, which encodes p-hydroxybenzoate hydroxylase in Pseudomonas aeruginosa, produced a mutant p-hydroxybenzoate hydroxylase encoded by pobA *, which was capable of hydroxylating protocatechuic acid to form gallic acid.31 117 0P03H2 PESOZH H3PO4 HO) 002“ H3PO4 HO,“ COZH OH 0 aroF aroB m aroG OH O 5. OH H arOH H203PO OH OH H203PO OH DAHP 3-dehydroquinic acid E4P COZH COZH COZH aroD "fi’ob‘fi aroZ pobA * OH? H HO OH H200 H20 0 HNADPH, NADP”, OH 02 H20 3-dehydroshikimic protocatechuic gallic acid acid acid Figure 52. Microbial synthesis of gallic acid by exploiting pobA* activity. Intermediates (abbreviations): phosphoenolpyruvate (PEP), D-erythrose 4-phosphate (E4P), 3—deoxy-D-arabino-heptulosonic acid 7-phosphate (DAHP). Genes (enzymes): aroF, aroG, aroH, DAHP synthase; aroB, 3-dehydroquinate synthase; aroD, 3- dehydroquinate dehydratase; aroZ, 3-dehydroshikimate dehydratase; p0bA*, p- hydroxybenzoate hydroxylase. Plasmid pSK6.76 contained aroZ, serA and pobA* genes. E. coli KL7/pSK6.76 was capable of converting 3-dehydroshikimic acid into gallic acid by a mechanism that represented the proposed gallic acid biosynthetic pathway via protocatechuic acid. Therefore, E. coli KL7/pSK6.76 was used as model for the alternate pathway to gallic acid that proceeds via 3-dehydroshikimic acid intermediacy. Due to the absence of a feedback-insensitive DAHP synthase, E. coli KL7/pSK6.76 was incapable of de novo biosynthesis of shikimate pathway metabolites when aromatic amino acids were added into the fermentation medium to inhibit the native, genome-encoded DAHP synthases. 118 \1 4:. 01 6 fl _ o -_ 40 ,2 0 ~_ 35 <1 5 4 Q g . '1" 30 8 < -o-0 (D 4 1r 4- 25 '5: <' ' “3’ 8 3 “i , a a ‘s o “ O 5 U 1 l5 0 O 4 g 0 12 18 24 30 36 time (h) Figure 53. Gallic acid and 3-dehydroquinic acid production upon addition of [5- 18O]-enriched 3-dehydroshikimic acid by E. coli KL7/pSK6.76. 3-Dehydroshikimic acid (DHS, C1), protocatechuic acid (PCA, stripped bar), gallic acid (GA, I), dry cell mass (0). KL7/pSK6.76 was cultured under fermentor-controlled conditions where the temperature was maintained at 36 °C, pH was maintained at 7 and dissolved oxygen was controlled at 20% of air saturation. [5-‘SO]-Enriched 3-dehydroshikimic acid (5.0 g) was added at 12 h, aromatic amino acids, L-tyrosine (0.7 g), L-phenylalanine (0.7 g) and L- tryptophan (0.35 g) was added at the beginning of fermentation and again at 18 b. Although the lac! repressor gene was not present in the plasmid, isopropyl fi-D- thiogalactopyranoside (IPTG, 15 mg) was added every 6 h after 12 h. Consumption of 3- dehydroshikimic acid and accumulation of protocatechuic acid were observed simultaneously during the first 12 h after addition of 3-dehydroshikimic acid. Complete consumption of 3-dehydroshikimic acid resulted in synthesis of 2.3 g/L of protocatechuic acid and 1.2 g/L of gallic acid at 12 h after addition (Figure 53). Protocatechuic acid 119 concentration was then decreased from 2.3 g/L at 24 h to a final concentration at 0.75 g/L at 36 h, while the concentration of gallic acid concentration increased from 1.24 g/L at 24 h to a final concentration of 3.1 g/L at 36 h. This profile was consistent with the biocatalytic route to gallic acid via protocatechuic acid as an intermediate. The gallic acid produced by E. coli KL7/pSK6.76 was isolated from culture supernatant and analyzed using FAB(-) mass spectrometry. The ratio of the 18O-labeled gallic acid signal (m/z 171) over its unlabeled counterpart (m/z 169, 170) was 2.32 to 111.76 (100+11.76), which meant the 18O-enrichment in the gallic acid was 0.10% (entry 4, Table 6). The very low incorporation (1.9%) of '80 atom in the final product gallic acid indicates the 18O-labeled hydroxyl group of 3-dehydroshikimic acid was lost and the newly added hydroxyl group was from 1602 as predicted by the reaction mechanism. For comparison, gallic acid synthesized by E. coli KL3/pRC1.55B fully retained the '80 labeling (entry 3, Table 6) and clearly was not biosynthesized via protocatechuic acid intermediacy. Kambourakis et al.24 reported a Cu2+/Zn2+ mediated air oxidation of 3- dehydroshikimic acid in acetic acid solution to afford gallic acid in 70% yield. Air- oxidation of 3-dehydroshikimic acid in phosphate-buffered water (1.0 M, pH 6.5) led to a mixture of gallic acid (14%), protocatechuic acid (12%), tricarballylic acid (14%) and pyrogallol (3%) at 40 °C after 50 h. A plausible mechanism32 for air oxidation of 3- dehydroshikimic acid was also proposed that involved initial tautomerization to an enediol form 23. Enediol 23 could either form protocatechuic acid by elimination of a water molecule or further react with O2 to yield gallic acid, tricarballytic acid and pyrogallol (Figure 54). Evidence for the presence of enediol 23 intermediacy was 120 obtained after isolation of dihydrogallic acid from an anaerobic KZHPO4 (1 M) solution of 3-dehydroshikimic acid. Inorganic phosphate could function as a general base to catalyze the initial tautomerization of 3-dehydroshikimic acid in phosphate-buffered water. COZH H20 COZH COaH )7 HO OH ‘_ protocatechuic O s OH HO OH acid OH OH 3-dehydroshikimic enediol 23 COQH acid COZH / HO OH OH gallic acid 0 OH OH 02 COZH enediol 24 \ // HOZC COZH tricarballylic COZH acid HO OH OH HO’E JOH dihydrogallic OH acid pyrogallol Figure 54. Mechanism for base-catalyzed air oxidation of 3-dehydroshikimic acid. 121 COZH COZH H2180 COZH OJ: ‘180H HO 18OH H5 OH OH _ OH _ [5-‘80]-DHS [5-‘30]-enediol 25 COZH COzH —-‘—- —- (1 HO OH 0 i OH OH OH enediol23 DHS Figure 55. Proposed mechanism for 18O-exchange of 3-dehydroshikimic acid in inorganic phosphate buffered water. Although dilute inorganic phosphate buffer (43 mM) was used in the fermentation medium, the possibility that gallic acid formed in the E. coli fermentation was the result of air oxidation of biosynthesized 3-dehydroshikimate could not be ruled out. It was not clear whether the hydroxyl oxygen atom at C-5 was exchangable during tautomerization process. The air oxidation of [5-‘80]-enriched 3-dehydroshikimic acid was carried out in potassium phosphate buffer (1 M, pH 6.5) at room temperature. After 48 h, gallic acid (19.7%), PCA (22.7%) and tricarballylic acid (25.0%) were produced as reported previously. Gallic acid produced by air oxidation was purified and analyzed by mass spectrometry as mentioned before. Mass spectra data showed an l8O-enrichment of 4.40% (entry 5, Table 6), which indicated only 82.4% of the 18O-labeling in 3- dehydroshikimic acid was incorporated into gallic acid as a result of air oxidation catalyzed by inorganic phosphate. The loss of 17.6% 180 atom during air oxidation might result from the [5-‘8O]-hydroxyl exchange with H2160 in inorganic phosphate buffer. A 122 A ‘-—- A »«._._-‘.'a-¢.-._..~-o_—L—~.—__._.. :—._... .-_.-a—,,, ._ . possible mechanism for this exchange is proposed in Figure 55. The first step involves tautomerization of 3-dehydroshikimic acid to its enediol form, same as in the air oxidation process that leads to a mixture of gallic acid, protocatechuic acid, tricarballylic acid and pyrogallol. Elimination of the 5-hydroxy group leads to a a,fi-unsaturated intermediate 25. H2160 could add to intermediate 25 through 1,4-addition reaction. Discussion Conversion of 3-dehydroshikimic acid into gallic acid has been narrowed to two pathways. 3-Dehydroshikimic acid could be directly oxidized to gallic acid which would be indicated by retention of C-5 hydroxyl group in 3-dehydroshikimic acid in the C-3 hydroxyl of gallic acid. Alternatively, 3-dehydroshikimic acid could undergo initial dehydration to protocatchuic acid involving elimination of the 5-hydroxy group, followed by hydroxylation to afford gallic acid. [5-‘8O]-3-Dehydroshikimic acid was chemically synthesized and [this labeled 3-dehydroshikimic acid was added into a culture of E. coli KL3/pRC1.55B in which in vivo biosynthesis of shikimate pathway metabolites was inhibited. Gallic acid was produced along with 3-dehydroquinic acid. Since E. coli KL3/pRC1.SSB is incapable of producing shikimate pathway metabolites under the fermentation conditions, the production of 3-dehydroquinic acid indicates the added 3- dehydroshikimic acid was transported into the E. coli cytoplasm presumably by a shiA- encoded shikimate transport protein and was converted to 3-dehydroquinic acid by the reversible, aroD-encoded 3-dehydroquinate dehydratase. The gallic acid produced by E. coli KL3/pRC1.55B and the remaining 3- dehydroshikimic acid in the culture medium were isolated and analyzed by mass 123 spectrometry. The full 18O-labeling retention in the gallic acid synthesized indicates that the 5-hydroxyl group in 3-dehydroshikimic acid was completely incorporated into gallic acid. By contrast, the 18O-labeled S-hydroxyl group in 3-dehydroshikimic acid was almost completely lost when protocatechuic acid is the intermediate in gallic acid synthesis by KL7/pSK6.76. These results clearly indicate indicated that 3- dehydroshikimic acid was directly oxidized to gallic acid without protocatechuic acid intermediacy in recombinant E. coli KL3. In addition to E. coli, this methodology can also be applied to probe the gallic acid biosynthesis in higher plant and fungi, which might lead to the discovery of a 3- dehydroshikimate dehydrogenase. However, the reported inorganic phosphate-catalyzed aerobic oxidation of 3-dehydroshikimic acid needed to be examined carefully. Abiotic aerobic oxidation could also result in retention of "‘0 atom in product gallic acid. The lack of assayable enzyme activity for conversion of 3-dehydroshikimic acid to gallic acid in cell-free extract adds more ambiguity to the existence of an oxidoreductase for 3- dehydroshikimate oxidation in nature. Air oxidation of 3-dehydroshikimic acid is always accompanied by the formation of nearly equal quantities of protocatechuic acid and tricaballylic acid, which is not consistent with the much higher ratio of gallic acid to protocatechuic acid observed in the [5380]-3-dehydroshikimate feeding experiment (6.2:1 by E. coli KL3/pRC55). Additional evidence that disfavors the abiotic aerobic oxidation of 3-dehydroshikimic acid comes from the results that the 18O-labeling was partially washed out in gallic acid produced by aerobic oxidation of 3-dehydroshikimic acid in inorganic phosphate- buffered water. Therefore, examination of the extent of label washout in 18O-enriched 3- 124 dehydroshikimic acid after fermentation could provide additional insights into the role of aerobic oxidation in the conversion of 3-dehydroshikimic acid to gallic acid. The remaining 3-dehdroshikimic acid was isolated from the E. coli KL3/pRCl .55B fermentation supernatant and analyzed by mass spectrometry. The data reveals a slightly decline in 18O-incorporation in 3-dehydoshikimic acid after fermentation (97.4%) compared with the added 3-dehydroshikimic acid (100%). This result indicates that the exchange reaction occurred at a much slower rate in the fermentation medium possibly due to low concentration (43 mM) of inorganic phosphate and an oxygen level of 20% air saturation used in the fermentation conditions. These results suggest that abiotic, aerobic oxidation is unlikely to be the primary route for gallic acid formation although a small portion of gallic acid is possibly formed by this mechanism. After the work was completed, Ossipov et a1.33 isolated the first 3- dehydroshikimate dehydrogenase that catalyzes the direct oxidation of 3-dehydroshikimic acid to gallic acid from leaves of mountain birch (Betula pubescens ssp. Czerepanovii). The 3-dehydroshikimate dehydrogenase was found to be a NADP-dependent enzyme with a Km value of 0.008 mM for NADP” and a K m value of 0.49 mM for 3- dehydroshikimate. The discovery of 3-dehydroshikimate dehydrogenase also raises an interesting question about the biosynthesis of its substrate, 3-dehydroshikimic acid in higher plants, since it is generally accepted that the third and fourth steps of the shikimate pathway are catalyzed by a bifunctional enzyme 3-dehydroquinate dehydratase-shikimate dehydrogenase in higher plants plastids.34 The metabolite channeling mechanism prevents the accumulation of 3-dehydroshikimate, the immediate precursor to gallic acid. 125 Therefore, the biosynthesis of 3-dehydroshikimate is proposed to proceed by monofunctional enzymes.35 The protein sequence of the 3-dehydroshikimate dehydrogenase from birch was not obtained.36 In theory, heterologous expression of the 3-dehydroshikimate dehydrogenase in a 3-dehydroshikimate producing E. coli strain could constitute a better biosynthesis of gallic acid than current synthesis by E. coli KL7/pSK6.16127 via protocatechuic acid intermediacy. In E. coli KL7/pSK6.161, the conversion of protocatechuic acid to gallic acid catalyzed by the pobA*-encoded p-hydroxybenzoate hydroxylase generates H202 as a byproduct. This H202 is capable of damaging E. coli cells and therefore reducing the yield and concentration of gallic acid synthesized by E. coli. 126 Reference ' Leston, G. In Kirk-Othmer Encyclopedia of Chemical Technology; Kroschwitz, J. I., Howe-Grant, M., Eds.; Wiley: New York, 1996; Vol. 19, p 778. 2 (a) Scalbert, A.; Mila, 1.; Expert, D.; Marmolle, F.; Albrecht, A.-M.; Hurrell, R.; Huneau, J.-F.; Tome, D. Polyphenols, metal ion complexation and biological consequences. Basic Life Sci. 1999, 66, 545-554. (b) Gross, G. G. Biosynthesis, biodegradation and cellular localization of hydrolysable tannins. Recent Advances Phytochem. 1998, 33, 185-213. 3 Harbome, J. B.; Grayer, R. J .; Flavonoids and inserts. In The F lavonoids-Advances in Research Since 1986. Harborne, J. B. Ed., Chapman and Hall, London, 1994, p. 589. 4 Haslam, E. In Practical Polyphenolics: From Structure to Molecular Recognition and Physiology Action. Cambridge University Press, Cambridge, 1998. 5 Grundhofer, P.; Niemetz, R.; Schilling, G.; Gross, G. G. Biosynthesis and subcellular distribution of hydrolysable tannins. Phytochem. 2001, 5 7, 915-927. 6 Haslam, E.; Haworth, R. D.; Knowles, P. F. Gallotannins. IV. The biosynthesis of gallic acid. J. Chem. Soc. 1961, 1854-1859. 7 Korth, H. Formation of gallic acid from quinic acid by Enterobacter cloacae and Pseudomonasfluorescens. Arch. Microbiol. 1973, 89, 67-72. 8 Kawakubo, J.; Nishira, H.; Aoki, K.; Shinke, R. Studies on phenolic-compounds production by molds. 2. Isolation of a gallic acid-producing microorganism with sake cake medium and production of gallic acid. Biosci. Biotech. Biochem. 1993, 5 7, 1360- 1361. 9 (a) Li, K.; Mikola, M. R.; Draths, K. M.; Worden, R. M.; Frost, J. W. Fed-batch fermentor synthesis of 3-dehydroshikimic acid using recombinant Escherichia coli. Biotechnol. Bioeng. 1999, 64, 61-73. (b) Yi, J.; Li, K.; Draths, K. M.; Frost, J. W. Modulation of phosphoenolpyruvate synthase expression increases shikimate pathway product yields in E. coli. Biotechnol. Prog. 2002, 18, 1141-1148. (c) Yi, J.; Draths, K. M.; Li, K.; Frost, J. W. Altered glucose transport and shikimate pathway product yields in E. coli. Biotechnol. Prog. 2003, 19, 1450-1459. 1° (a) Comthwaite, D.; Haslam, E. Gallotannins. IX. Biosynthesis of gallic acid in Rhus typhina. J. Chem. Soc. 1965, 3008-3011. (b) Saijo, R. Pathway of gallic acid biosynthesis and its esterification with catechins in young tea shoots. Agric. Biol. Chem. 1983, 47, 455-460. (c) Ishikura, N .; Hayashida, S.; Tazaki, K. Biosynthesis of gallic and ellagic acids with l4C-labeled compounds in Acer and Rhus leaves. Bot. Mag. Tokyo 1984, 97, 355-367. 127 ” Loscher, R.; Heide, L. Biosynthesis of p-hydroxybenzoate from p-coumarate and p- coumaroyl-coenzyme A in cell-free extracts of Lithospermum erithrorhizon cell cultures. Plant Physiol. 1994, 106, 271-279. ‘2 (a) Lydon, J .; Duck, S. O. Glyphosate induction of elevated levels of hydroxybenzoic acids in higher plants. Agric. Food Chem. 1988, 36, 813-818. (b) Becerril, J.; Duke, S.; Lydon, J. Glyphosate effects on shikimate pathway products in leaves and flowers of velvetleaf. Phytochem. 1989, 28, 695-700. ‘3 Werner, 1.; Bacher, A.; Eisenreich W. Retrobiosynthetic NMR studies with l3C-labeled glucose: formation of gallic acid in plants and fungi. J. Biol. Chem. 1997, 272, 25474- 25482. ” Dixon, R. A.; Achnine, L.; Kota, P.; Liu, C.-J.; Reddy, M. S. S.; Wang, L. The phenylpropanoid pathway and plant defense: a genomics perspective. Mol. Plant Pathol. 2002, 3, 371-390. ‘5 T ateoka, T. N. Formation of protocatechuic acid from 5-dehydroshikimic acid in the extract of mung bean seedlings. Bot. Mag. Tokyo 1968, 81, 103-104. '6 Kato, N.; Shiroya, M.; Yoshida, S.; Hasegawa, M. Biosynthesis of gallic acid by a homogenate of the leaves of Pelargonium inquinans. Bot. Mag. Tokyo 1968, 81 , 506-507. '7 Pittard, A. J. Biosynthsis of aromatic amino acids. In Escherichia coli and Salmonella: Cellular and Molecular Biology, 2“d ed.; Neidhardt, F. C., Ed.; ASM press: Washinfton, DC, 1996: p. 458-484.. '8 (a) McGowan, D. A.; Berchtold, G. A. (—)-Methyl cis-3-hydroxy-4,5-oxycyclohex-1- enecarboxylate: stereospecific formation from and conversion to (-)-methyl shikimate; Complex formation with bis(carbomethoxy)hydrazine. J. Org. Chem. 1981, 46, 2381. (b) Tan, D. S.; Foley, M. A.; Stockwell, B. R.; Shair, M. D.; Schreiber, S. L. Synthesis and preliminary evaluation of a library of polycyclic small molecules for use in chemical genetic assays. J. Am. Chem. Soc. 1999, 121, 9073-9087. ‘9 Gustin, D. J.; Hilvert, D. J. Chemoenzymatic synthesis of isotopically labeled chorismic acids. J. Org. Chem. 1999, 64, 4935-4938. 2° Chaudhuri, S.; Anton, I. A.; Coggins, J. R. Shikimate dehydrogenase from Escherichia coli. Methods Enzymol. 1987, 142, 315-320. 2' (a) Tolbert, T. J .; Williamson, J. R. Preparation of specifically deuterated and 13C- labeled RNA for NMR studies using enzymatic synthesis. J. Am. Chem. Soc. 1997, 119, 12100-12106. (b) Carrea, G.; Bovara, R.; Cremonesi, P.; Lodi, R. Enzymatic preparation of 12-ketochenodeoxycholic acid with N ADP regeneration. Biotechnol. Bioeng. 1984, 26, 128 560-563. (c) Wong, C.-H.; McCurry, S. D.; Whitesides, G. M. Practical enzymatic syntheses of ribulose 1,5-bisphosphate and ribose 5-phosphate. J. Am. Chem. Soc. 1980, 102, 7938-7939. ((1) Chenault, H. K.; Whitesides, G. M. Regeneration of nicotinamide cofactors for use in organic synthesis. Applied Biochem. Biotechnol. 1987, 14, 147-197. 22 Kambourakis, S. Doctorate Dissertation, Michigan State University, 2000. 23 Vederas, J. C. Biosynthetic studies using oxygen-18 isotope shifts in carbon-13 nuclear magnetic resonance. Can. J. Chem. 1982, 60, 1637-1642. 24 Kambourakis, S.; Frost, J. W. Synthesis of gallic acid: Cu2+ mediated oxidation of 3- dehydroshikimc acid. J. Org. Chem. 2000, 65 , 6904-6909. 25 Whipp, M. J.; Camakaris, H.; Pittard, A. J. Cloning and analysis of the shiA gene, which encodes the shikimate transport system of Escherichia coli K-12. Gene 1998, 209, 185-192. 26 Draths, K. M.; Frost, J. W. Environmentally compatible synthesis of catechol from D- glucose. J. Am. Chem. Soc. 1995, 117, 2395-2400. 27 Kambourakis, S.; Draths, K. M.; Frost, J. W. Synthesis of gallic acid and pyrogallol from glucose: replacing natural product isolation with microbial catalysis. J. Am. Chem. Soc. 2000, 122, 9042-9043. 28 Schweizer, M.; Case, M. B.; Dykstra, C. C.; Giles, N. H.; Kushner, S. R. Cloning and quinic acid (qa) gene cluster from Neurospora crassa: Identification of recombinant plasmids containing both qa-2+ and qa-3+. Gene 1981, 14, 23-32. 29 Hawkins, A. R.; Francisco daSilva, A. J .; Roberts, C. F. Cloning and characterization of the three enzyme structural genes QU TB, QU TC and QU TE from the quinic acid utilization gene cluster in Aspergillus nidulans. Curr. Genetics 1985, 9, 305-311. 3° (a) Gross, S. R. Enzymic conversion of 5-dehydroshikimic acid to protocatechuic acid. J. Biol. Chem. 1958, 233, 1146-1151. (b) Floss, H. G.; Scharf, K. H.; Zenk, M. H.; Onderka, D. F.; Carroll, M. Stereochemistry of the enzymic and nonenzymic conversion of 3-dehydroshikimate into protocatechuate. J. Chem. Soc., Chem. Commun. 1971, 14, 765-766. 3‘ (a) Entsch, B.; Palfey, B. A.; Ballou, D. P.; Massey, V. Catalytic function of tyrosine residues in para-hydroxybenzoate hydroxylase as determined by the study of site- directed mutants. J Biol. Chem. 1991, 266, 17341-17349. (b) Eschrich, K.; van der Bolt, F. J. T.; de Kok, A.; van Berkel, W. J. H. Role of Tyr201 and Tyr385 in substrate activation by p-hydroxybenzoate hydroxylase from Pseudomonas fluorescens. Eur. J. Biochem. 1993, 216, 137-146. 129 32 Richman, J. E.; Chang, Y.-C.; Kambourakis, S.; Draths, K. M.; Almy, E.; Snell, K. D.; Strasburg, G. M.; Frost, J. W. Reaction of 3-dehydroshikimic acid with molecular oxygen and hydrogen peroxide: products, mechanism, and associated antioxidant activity. J. Am. Chem. Soc. 1996, 118, 11587-11591. 33 Ossipov, V.; Salminen, J .-P.; Ossipov, S.; Haukiojia, E.; Pihlaja, K. Gallic acid and hydrolysable tannins are formed in birch leaves from an intermediate compound of the shikimate pathway. Biochem. System. Ecol. 2003, 31, 3-16. 34 Hermann, K. M.; Weaver, L. M. The shikimate pathway. Ann. Rev. Plant. Physiol. Plant. Mol. Biol. 1999, 50, 473-503. 35 Mousdale, D.; Coggins, J. Subcellular localization of the common shikimate pathway enzymes in Pisum sativum L. Planta 1985, 163, 241-249. 36 Ossipov, V. personal communication. 130 CHAPTER FOUR Creation of a Pyruvate-Based Shikimate Pathway in Escherichia coli Introduction Improving the synthesis of hydroquinone from glucose discussed in Chapter 2 is part of a larger effort to improve microbial syntheses of shikimate pathway products. A variety of strategies have been employed for improving the yields of shikimate pathway metabolites in E. coli."2 These strategies are primarily focused on expression of mutant isozymes of DAHP synthase that are insensitive to feedback inhibition by aromatic amino acids and increasing the availability of D-erythrose 4-phosphate (E4P) and phosphoenolpyruvate (PEP). E4P and PEP are the substrates of DAHP synthase, which is the first enzyme in the shikimate pathway. In vivo activity of the first enzyme plays an essential role in directing carbon flow into the shikimate pathway.3 Feedback inhibition of the DAHP synthase isozymes in vivo by aromatic amino acids constitutes the major regulatory control of aromatic amino acids biosynthesis.3a Several alleles that encode feedback-insensitive DAHP synthase have been obtained by mutation of aroF,3"'” aroG4 and aroH.5 The use of DAHP synthase isozymes that are insensitive to feedback inhibition by aromatic amino acids has been a centerpiece of virtually all efforts to synthesize shikimate pathway products in high yield. D-Erythrose 4-phosphate is derived from the non-oxidative pentose phosphate pathway. Draths et al.6 demonstrated in 1990 that amplified expression of transketolase resulted in a two-fold increase in concentrations of biosynthesized shikimate pathway metabolites presumably by increasing the availability of D-erythrose 4-phosphate. 131 Transketolase catalyzes the conversion of D-fructose 6-phosphate into D-erythrose 4- phosphate. Transketolase also serves to generate the substrate D-sedoheptulose-7- phosphate for conversion to D-erythrose 4-phosphate catalyzed by tal-encoded transaldolase. Overexpression of transaldolase also relieves D-erythrose 4-phosphate limitation in the presence of overexpressed phosphoenolpyruvate synthase. No further improvements in the biosynthesis of shikimate pathway metabolites has been observed when both transketolase and transaldolase are overexpressed.7 Phosphoenolpyruvate is derived from the Embden-Meyerhof-Parnas glycolytic pathway. Besides being a substrate for the first enzyme in the shikimate pathway, phosphoenolpyruvate is also used as a phosphate donor in the phosphoenolpyruvate:carbohydrate phosphotransferase (PT S) system for microbial transport and phosphorylation of glucose, as a substrate in 3-deoxy-D-manno- octulosonate biosynthesis and peptidoglycan biosynthesis, and participates in ATP- generating reactions catalyzed by pyruvate kinase.8 The resulting competition between the shikimate pathway and other enzymes for the cellular supplies of phosphoenolpyruvate limits the concentrations and yields of natural products synthesized by way of the shikimate pathway. In wild-type E. coli, the major consumer of phosphoenolpyruvate is the glucose phosphotransferase system (PTS) for transport and phosphorylation of glucose.9 PEP- mediated phosphoryl group transfer both energetically drives glucose transport and generates the glucose 6-phosphate. One molecule of phosphoenolpyruvate is converted into pyruvate for every molecule of glucose transported into the cytoplasm by the PTS system. PTS-generated pyruvate is oxidized through the tricarboxylic acid cycle (TCA 132 cycle) to carbon dioxide and is apparently not recycled to phosphoenolpyruvate under normal aerobic culture conditions. The theoretical maximum total yield for shikimate pathway products synthesized from glucose is 43% (mol/mol) based on theoretical flux analysis1d (Figure 56). The yield reflects the fundamental limitation imposed on microbial synthesis by the PTS—mediated glucose transport. Glucose PYR<7 =l= G6P 1 Ribulose-SP W 71 , l/ \1 F6P f. st R5P 61 2 1 Tkt GAP S7P 1,6-FDP m 9< 2 fi/ \2 2 1 Tal /1 6 L>- F6P DHAP ——>GAP 10 l 7 N <— \J DAHP synthase DAHP Figure 56. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains reliant on PTS for glucose transport. The numbers are the relative fluxes needed to convert 7 mol of glucose into DAHP. Enzymes: Pps, PEP synthase; Tkt, transketolase; Tal, transaldolase. Metabolites: G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; 1,6-FDP, 1,6-fructose diphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde 3-phosphate; RSP: ribose 5-phosphate; X5P, xylulose S-phosphate; S7P, sedoheptulose 7-phosphate; PYR, pyruvate. The theoretical maximum yield is calculated by assuming that the branching pathways are blocked and the carbon flow is directed by the most efficient pathways with minimal loss to carbon dioxide, other metabolites and biomass. Under these conditions, 133 the relative flux through each step at the steady state can be calculated by balancing the input (production) and output (consumption) fluxes from each metabolite pool. If E. coli transports glucose by PTS without recycling pyruvate to phosphoenolpyruvate, the input of 7 mol of glucose can produce 3 mol of DAHP (43% molar yield) and 7 mol of pyruvate (Figure 56), which is further metabolized to CO2 and acetyl-CoA for the TCA cycle. Different strategies have been described for increasing the cytoplasmic concentration of phosphoenolpyruvate by way of circumventing the expenditure of phosphoenolpyruvate by the PTS-mediated glucose transport employed by E. coli and many other microbes.‘0 Those strategies comprised of recycling of the PTS-generated pyruvate back to phosphoenolpyruvate using amplified expression of ppsA-encoded phosphoenolpyruvate synthase1d and replacing PTS-driven glucose transport by facilitated diffusion mediated by the Zymomonas mobilis glf-encoded glucose facilitator‘1 and galactose-proton symport mediated by E. coli galP-encoded galactose permease (Figure 57).23 In addition, employing a different source of carbon such as D-xylose and L-arabinose for microbial growth and shikimate pathway products production would circumvent the PT S-mediated glucose transport since the transport of these pentoses in E. coli are driven by a high affinity permease driven by the conversion of ATP to ADP.""12 Unfortunately, commercial D-xylose and L-arabinose streams that are pure, abundant, and. inexpensive are not currently available. An alternative strategy employs glucose adjuncts that can be readily converted to phosphoenolpyruvate. For example, equivmolar amounts of succinic acid can be added to glucose-limited cultures of E. coli KL3/pKL6.218A. This construct converted succinic acid by way of the TCA cycle into oxaloacetate. 134 Overexpressed phosphoenolpyruvate carboxylase then converted the oxaloacetate into PEP. Glucose-limited cultures of E. coli construct KL3/pKL6.218A overexpressing phosphoenolpyruvate carboxykinase in the presence of DAHP synthase and transketolase led to synthesis of 30% higher concentrations of 3-dehydroshikimic acid.” The yield (29%, mol/mol), however, was almost the same as using glucose as sole carbon source (28%, mol/mol). ATP AMP + P, Pps OPO3H2 O COZH ACOzH PEP pyruvate OH OH .,OH “TS/L ,.OH OH Glt or Gall: le HO OH ATP ADP H203PO OH glucose glucose 6-phosphate Figure 57. Previously examined strategies to increase phosphoenolpyruvate availability. Metabolite: PEP, phosphoenolpyruvate. Enzymes: Pps, phosphoenolpyruvate synthase; Glf, glucose facilitator; GalP, galactose permease; le, glucose kinase. This work explores whether pyruvate can replace phosphoenolpyruvate in an enzyme-catalyzed condensation with D-erythrose 4-phosphate to form 3-deoxy-D- arabino-heptulosonate 7-phosphate (DAHP, Figure 58). The centerpiece of this created shikimate pathway variant is the directed evolution of 2-keto-3-deoxy-6- phosphogalactonate (KDPGal, Figure 58) aldolase.l4 This constitutes a fundamental departure from all previous strategies employed to increase phosphoenolpyruvate 135 availability in E. coli. The pyruvate-based shikimate pathway utilizes the pyruvate byproduct of PT S-mediated glucose transport instead of competing with PT S-mediated glucose transport for cellular supplies of phosphoenolpyruvate. pyruvate phosphoenolpyruvate , OH H203PO OH DAHP \ aromatic COZH H04. 002” amino acids m d Q ‘—— aromatic O ; OH O ; OH VltamlnS OH OH 3-dehydroshikimate 3-dehydroquinate Figure 58. Proposed pyruvate-based shikimate pathway and the native phosphoenolpyruvate-based shikimate pathway. “ Metabolites: E4P, D-erythrose 4- phosphate; DAHP, 3-deoxy-D-arabino-heptulosonic acid 7-phosphate. b Enzymes (genes): (a) KDPGal aldolase (dgoA); (b) DAHP synthase (aroF, aroG, aroH); (c) 3- dehydroquinate synthase (aroB); (d) 3-dehydroquinate dehydratase (aroD). An input of 7 mol of glucose can thus produce 6 mol of DAHP under the condition where 1 mol of pyruvate generated by the PTS system is converted to phosphoenolpyruvate (Figure 59). This in theory can be accomplished by converting 136 pyruvate into oxalacetate by way of the TCA cycle followed by PEP carboxykinase- catalyzed conversion to phosphoenolpyruvate. The theoretical maximum yield for biosynthesis of a shikimate pathway metabolite such as 3-dehydroshikimic acid via a pyruvate-based shikimate pathway is 86% (mol/mol) as compared with 43% (mol/mol) using the phosphoenolpyruvate-based shikimate pathway with PTS-mediated glucose transport (Figure S9). Glucose / 71 3 66" 2 Ribulose-SP \ 7 2 2 l 2 / \ F6P K” X5P R5P 51 4 2 T1“ 1,6-FDP Tkt 9< GAP S7P 5 4 / \i 4 2 Ta! )2 DHAP -—5—>GAP "’_ FGP 6 l 7 PEP 6 1 LLfi-a DAHP Figure 59. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains employing pyruvate-based shikimate pathway. The numbers are the relative fluxes needed to convert 7 mol of glucose into DAHP. Enzymes: Pps, PEP synthase; Tkt, transketolase; Tal, transaldolase. Metabolites: G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; 1,6-FDP, 1,6-fructose diphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde 3-phosphate; R5P: ribose S-phosphate; X5P, xylulose 5- phosphate; S7P, sedoheptulose 7-phosphate; PYR, pyruvate. 137 According to Lowry et al.,15 the concentration of pyruvate in E. coli cells grown on glucose is about 0.9 umol/g (dry weight), while the phosphoenolpyruvate concentration is about 0.21 umol/g (dry weight). Utilization of the more abundant pyruvate as the substrate for the first enzyme of the shikimate pathway instead of phosphoenolpyruvate might therefore lead to the synthesis of higher concentrations of shikimate pathway products. A pyruvate-based shikimate pathway is also potentially important form the standpoint of controlling fed-batch cultures. When glucose is added too rapidly to E. coli during fed-batch cultures under fermentor-controlled conditions, large amounts of pyruvate are generated by the microbe resulting in the export and accumulation of acetic acid in the culture supernatant.” To avoid the accumulation of acetic acid to toxic levels, the rate of glucose addition needs to be closely matched to the rate of glucose consumed by the microbe.” In a pyruvate-based shikimate pathway, the glucose addition rate may not need to be tightly controlled since the pyruvate generated during PTS-mediated glucose transport is being consumed to make shikimate pathway products. Although relieved from feedback inhibition by aromatic amino acids, the AroFFBR and AroGFBR feedback-insensitive isozymes of DAHP synthase are prone to a decline in specific activity over the course of an E. coli culture. Tribe and Pittard reported a 7 h'1 half-life for the specific activity of AroG and a 5.5 h'1 half-life for the specific activity of AroF.l7 Unlike the native DAHP synthase isozymes which are feedback-inhibited by aromatic amino acids, KDPGal aldolase is unlikely to be subject to feedback inhibition by aromatic amino acids. Replacing the proteolysis-labile DAHP synthase with KDPGal 138 aldolase would be beneficial to the biosynthesis of shikimate pathway products by maintaining stable enzyme activity. Beyond the attendant biocatalytic implications, a shikimate pathway variant based on condensation of pyruvate with D-erythrose 4-phosphate may be important as a theoretical construct. Minimizing expenditure of the high-energy phosphoenolpyruvate by the shikimate pathway might be a metabolic advantage under certain growth conditions. The shikimate pathway variant outlined in this chapter may thus serve as a model of a naturally occurring aromatic biosynthetic pathway that remains to be discovered. Previously examined syntheses of shikimate pathway products by E. coli constructs with phosphoenolpyruvate synthase overexpression. Patnaik et a1ld were the first to demonstrate the positive effect on shikimate pathway product yields attendant with recycling of pyruvate back to phosphoenolpyruvate by amplified expression of p p sA-encoded PEP synthase. PEP synthase catalyzes the generation of phosphoenolpyruvate from pyruvate with the expenditure of high-energy ATP to form AMP.l8 If one molecule of pyruvate generated from phosphoenolpyruvate for glucose transport and formation of glucose 6-phosphate is recycled back to phosphoenolpyruvate instead of being oxidized to C02, the maximum theoretical yield for shikimate pathway products synthesized from glucose would be 86% (mol/mol). Figure 60 shows that if pyruvate is fully recycled to phosphoenolpyruvate by ppsA-encoded PEP synthase, a total of 6 mol of DAHP can be produced from 7 mol of glucose (86% molar yield). It is also noteworthy from the flux distribution in Figure 60 that 4 mol out of 6 mol input E4P is derived from the transketolase-catalyzed reaction by 139 coupling Embden-Meyerhof pathway intermediate fructose-6-phosphate and glyceraldehyde 3-phosphate. In reality, this is not likely to happen without amplified expression of transketolase. Glucose PYR = 7 =l= 1 7* J \ GGP 2 Ribulose—SP 71 2/ \2 2 F6P r» X5P RSP 51 4 2 Tkt 1.6-FDP T” X GAP 4 i/ \i 4 2 Tal j DHAP —5—>GAP *- FBP 101 7 6 \ 7 6 DAHP synthase DAHP Figure 60. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains with fully recycling pyruvate to phosphoenolpyruvate. The numbers are the relative fluxes needed to convert 7 mol of glucose into DAHP. Enzymes: Pps, PEP synthase; Tkt, transketolase; Tal, transaldolase. Metabolites: G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; 1,6-FDP, 1,6-fructose diphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde 3-phosphate; R5P: ribose 5-phosphate; X5P, xylulose 5- phosphate; S7P, sedoheptulose 7-phosphate; PYR, pyruvate. Using an E. coli aroB construct with overexpression of a feedback-insensitive isozyme of DAHP synthase (AroGFBR), PEP synthase and transketolase, 3-deoxy-D- arabino-heptulosonic acid (DAH) was synthesized in 90% (mol/mol) yield from glucose as reported by Patnaik et al.”"‘ However, the yield likely overestimated the actual 140 “ H..-—a .- —.-*~ .- . o conversion of glucose into DAH since the aroB construct was initially grown in rich medium followed by resuspension in minimal salts medium containing glucose. Previously examined shikimate pathway synthesis by E. coli constructs employing non-PTS glucose transport. In another approach, expenditure of PEP during glucose transport was completely avoided by a non-PTS glucose transport mechanism. Non-PT S glucose transport includes facilitated diffusion mediated by the Zymomonas mobilis glf-encoded glucose facilitator” and galactose-proton symport mediated by the E. coli galP-encoded galactose permease.20 Glf is a low-affinity glucose facilitator with an apparent Km for glucose of approximately 1.1-2.9 mM.19 GalP is a member of the major facilitator family (MFS).21 E. coli constructs were assembled where PTS-mediated glucose transport was replaced with heterologous expression of the Z. mobilis glf-encoded glucose facilitator or the E. coli galP-encoded galactose permease. Genomic E. coli glk-encoded glucokinase alone or in combination with plasmid-localized Z. mobilis glk-encoded glucokinase phosphorylates glucose to glucose-6-phosphate using ATP as the phosphoryl group donor. On the basis of the flux analysis shown in Figure 61, the maximum theoretical yield for shikimate pathway products synthesized from glucose would also be 86% (mol/mol). Ingram and coworkers” were the first to demonstrate that E. coli mutants lacking PTS-mediated glucose transport and phosphorylation can grow on glucose with heterologous expression of the Z. mobilis glf-encoded glucose facilitator protein and Z. mobilis glk-encoded glucokinase. Heterologous expression of the Z. mobilis glf-encoded glucose facilitator22 increased the concentrations of L-phenylalanine synthesized by various E. coli constructs. The impact of expression of Z. mobilis glf and Z. mobilis glk 141 on synthesis of shikimic acid in E. coli SP1.1pts/pSC6.090B, which is deficient in PT S- mediated transport and phosphorylation of glucose, has been examined.23 SP1.1pts/pSC6.090B synthesized 71 g/L of shikimic acid in 27% (mol/mol) yield which represented the highest yield achieved for shikimic acid synthesis with no supplement of yeast extract. Glucose vi 66" 2 Ribulose-SP \ 7 2 2 l 2 / \ F6P r» X5P R5P 1,6—FDP Tkt —< GAP S7P 5 4 4.x 4 Ex 5 DHAP —>GAP - FGP synth 6 l 6 6 DAHP ase DAHP Figure 61. Reaction pathways for maximal conversion of glucose to DAHP for E. coli strains employing non-PTS glucose transport. The numbers are the relative fluxes needed to convert 7 mol of glucose into DAHP. Enzymes: Pps, PEP synthase; Tkt, transketolase; Tal, transaldolase. Metabolites: G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; 1,6-FDP, 1,6-fructose diphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde 3-phosphate; R5P: ribose 5-phosphate; X5P, xylulose 5-phosphate; S7P, sedoheptulose 7-phosphate; PYR, pyruvate. Several research groups have studied the impact of shikimate pathway product yields in E. coli attendant with replacement of PTS-mediated glucose transport with glucose transport mediated with by the GalP galactose-proton symport system.2 Valle 142 and coworkers examined DAHP production in E. coli NF9 which was devoid of PTS- mediated glucose transport and relied on GalP-mediated glucose transport.211 E. coli NF9/pRW5tkt synthesized 2.4-fold more DAHP under shake-flask culture conditions relative to E. coli PBlO3/pRW5tkt, which retained PTS-mediated glucose transport.28 Plasmid pRWStkt contained tktA encoding transketolase and aroGFBR encoding feedback- insensitive DAHP synthase. Synthesis of DAHP in 71% (mol/mol) yield from glucose by E. coli NF9aroB harboring feedback-insensitive aroGFBR and tktA genes, and a detailed analysis of carbon metabolism have recently been reportedm‘ for E. coli constructs where GalP was recruited for glucose transport. In a contradictory report, Bailey and coworkers2c examined L-phenylalanine synthesis in E. coli PPA316/pSY130-14, which also relied on GalP-mediated glucose transport with inactivation of PTS-mediated glucose transport. Plasmid pSY130-l4 carried aroFFBR encoding feedback-insensitive DAHP synthase and pheAFBR encoding chorismate mutase-prephenate dehydratase, which are the critical regulated steps in L-phenylalanine biosynthesis. In contrast to Valle’s result, E. coli PPA316/pSY130-14 under fermentor-controlled culture conditions produced less L-phenylalanine in the range of O - 67% of that produced by the PTS strain PPA305/pSY130-14.2° The divergent impact of GalP-mediated glucose transport on the synthesis of L-phenylalanine relative to synthesis of DAHP may be the result of different E. coli strains and different cultivation conditions. The Bailey study also didn’t amplify expression of either transketolase or transaldolase. Increased availability of phosphoenolpyruvate in E. coli is not reflected in increased shikimate pathway product yield until the availability of D-erythrose 4-phosphate is increased with amplified expression of transketolase.lc 143 DAHP synthase (3-deoxy-D-arabino-heptulosonate-7-phosphate synthase). DAHP synthase catalyzes the first committed step in the shikimate pathway. The two substrates, phosphoenolpyruvate and D-erythrose 4-phosphate, condense in a stereospecific aldol-like reaction to form a molecule of 3-deoxy-D-arabin0-heptulosonate 7—phosphate (DAHP). Addition occurs to the re-face of the carbonyl group of D- erythrose 4-phosphate and to the si-face of the double bond in the phosphoenolpyruvate. In E. coli, there are three isozymes of DAHP synthase. Each isozyme is specifically feedback—regulated by one of the three aromatic amino acids end products. The crystal structure of the phenylalanine-regulated DAHP synthase from E. coli has been described and represents the first reported protein structure of this family.24 The data show a tetrameric enzyme complex to be made of a dimer of two tight dimmers with the monomers each possessing a single (,B/a)8-barrel domain. The active site of the enzyme is situated at the carboxy terminal end of the (Weds-barrel and possesses an overall net positive potential which complements the negatively charges substrates, phosphoenolpyruvate and D-erythrose 4-phosphate. The crystal structure of the phenylalanine-regulated DAHP synthase complexed with its inhibitor L-phenylalanine, phosphoenolpyruvate and its metal cofactor Mn2+ has been determined recently to elucidate the allosteric inhibition mechanism by L-phenylalanine.25 L-Phenylalanine is bound in a cavity located on the outer side of the DAHP synthase (flak-barrel near its N- amino terminus in each of the four monomers and about 20 A from the closet active site. Upon binding with L-phenylalanine. the enzyme loses the ability to bind D-erythrose 4- phosphate and binds phosphoenolpyruvate in a flipped orientation. 144 KDPGal aldolase (2-keto-3-deoxy-6-phosphogalactonate aldolase). The key enzyme in the proposed pyruvate-based shikimate pathway is dgoA-encoded KDPGal aldolase. By catalyzing the reversible cleavage of KDPGal to pyruvate and D- glyceraldehyde 3-phosphate (GAP, Figure 62), KDPGal aldolase enables E. coli to use D- galactonate as a sole carbon and energy source. D-Galactonate is transported into the E. coli cytoplasm by the dgoT-encoded transport protein. D-Galactonate catabolism in E. col 1' cytoplasm involves the conversion of D-galactonate into pyruvate and glyceraldehydes 3-phosphate by the sequential actions of dgoD-encoded galactonate dehydratase, dgoK-encoded 2-keto-3-deoxygalactonate kinase and dgoA-encoded 2-keto- 3-deoxy-6-phosphogalactonate aldolase (KDPGal aldolase, Figure 62).”5 The D- galactonate catabolic enzymes are tightly regulated and only expressed in the presence of D-galactonate. The absence of these enzyme activities in E. coli grown on galactose suggests that D-galactonate catabolic enzymes do not participate in galactose catabolism in E. coli. Galactose is metabolized in enteric bacteria including E. coli by the conversion of galactose to glucose 6-phosphate. The catabolism is initiated by phosphorylation of galactose to galactose-l-phosphate catalyzed by galK-encoded galactokinase, followed by conversion to UDP-galactose catalyzed by galT—encoded galactose-l-phosphate uridylyltransferase, then to UDP-glucose catalyzed by galE-encoded UDP-galactose-4— epimerase. UDP-glucose is converted into glucose-6-phosphate by galU-encoded glucose-l-phosphate uridylytransferase and pgm-encoded phosphoglucomutase. On the other hand, the catabolism of galactose in Pseudomonas is via an Entner-Doudoroff pathway27 in which galactose is oxidized to D-galactonate by a galactose dehydrogenase. 145 D-Glactonate is then converted to pyruvate and glyceraldehydes 3-phosphate following the same catalytic route as in the E. coli galactonate catabolic pathway (Figure 62). This galactose catabolic pathway is also found in Caulobacter cresem‘us,28 Azotobacter vinelana'ii,29 Rhizobium meliloti,30 Gluconobacter liquefaciens31 and nonpathogenic Mycobacteria.32 Consequently, KDPGal aldolase activity has been detected in all the microbes mentioned above. In all instances, KDPGal aldolase is an inducible enzyme, expressed only when cells are grown on galactose or galactonate as the sole carbon source. QH OH 9H 0 ' d 00 - d 0K HOWCOZH ‘1.— HOWCOZH _q_> OH OH OH D-galactonate D-2-keto-3-deoxygalactonate 9H 0 d 0A 0 O - 9 H203POWCOZH H203PO’\.:/ILH + )LCOZH OH OH D-2-keto-3-deoxy-6- D-glyceraldehyde phosphogalactonate 3-phosphate pyruvate Figure 62. Pathway for D-galactonate catabolism in E. coli. Proteins (genes): galatonate permease (dgoT); galactonate dehydratase (dgoD); 2-keto-3-deoxygalactonate kinase (dgoK); 2-keto-3-deoxy-6-phosphoga1actonate aldolase (dgoA). KDPG aldolase33 (2-keto-3-deoxy-6-phosphogluconate aldolase), catalyzes a reaction similar to that catalyzed by KDPGal aldolase (Figure 63a). KDPG aldolase is part of the Entner-Doudoroff pathway, which is an alternative to the Embden-Myerhof- Pamas pathway for glucose catabolism and enables E. coli to use gluconic acid as a sole carbon and energy source. KDPG aldolase catalyzes the reversible cleaveage of KDPG 146 to pyruvate and glyceraldehyde 3-phosphate. Both KDPG aldolase and KDPGal aldolase accept a broad range of short chain unnatural electrophilic aldehydes.34'35 KDPG aldolase catalyzes the si-face addition of pyruvate to an eletrophilic aldehyde to obtain 4-hydroxy- 2-ketobutyrates with the S-configuration at the newly formed C-4 stereogenic center (Figure 63a,b). By contrast, KDPGal aldolase generates the aldol adducts with the R- configuration at the new stereogenic center (Figure 63c). O O OH O KDPG aldolase a. H203PO/\ZLH + )LCOZH = H203POWCOZH OH OH D-glyceraldehyde pyruvate 2-keto-3-deoxy-6- 3-phosphate phosphogluconate 0H 0 0 KDPG aldolase 0“ 0H 0 H 0 P0 = = H 0 P0 2 3 MH + ACOZH 2 3 i COZH OH OH D-erythrose pyruvate 3-deoxy-D-ribo-heptulosonate 4-phosphate 7-phosphate (DRHP) OH O O OH OH O KDPGaI aldolase -_- Ho\/'\i/ILH + )LCOZH —_ HO i COZH OH OH D-erythrose pyruvate 3-deoxy-D-arabino-heptulosonic acid (DAH) Figure 63. KDPGal aldolase (dgoA) and KDPG aldolase (eda) catalyzed reactions. Systematic studies from two research groups have found that KDPG aldolase from Pseudomonas putida, E. coli and Zymomonas mobilis could catalyze the condensation of pyruvate with D-erythrose 4-phosphate.36 However, the product of this condensation is 3-deoxy-D-ribo-heptulosonic acid 7-phosphate (DRHP). DRHP is the 147 diastereomer of 3-deoxy-D-arabino-heptulosonic acid 7-phosphate (DAHP), which is needed in the shikimate pathway (Figure 63). DAHP would be formed in a KDPGal aldolase-catalyzed condensation of pyruvate with D-erythrose 4-phosphate (Figure 63c). D-Erythrose has been previously observed to be a poor substrate for KDPGal aldolase.348 Aldose phosphates were known to lead to KDPGal aldolase activities lOO-fold higher than with the corresponding aldoses,34 although there was no direct precedent for KDPGal aldolase-catalyzed condensation of pyruvate with D-erythrose 4-phosphate. Both KDPG aldolase and KDPGal aldolase belong to the type I aldolase family whose reaction mechanism involves formation of a Schiff base intermediate between a lysine residue and a substrate carbonyl in the active site. The X-ray crystal structure of KDPG aldolase has been solved at 1.95 A resolution,37 while the crystal structure of KDPGal aldolase is not currently available. The E. coli KDPG aldolase structure indicates that pyruvate underwent a nucleophilic attack with Lys-l33, forming a protonated carbinolamine intermediate, a Schiff base precursor, which was stabilized by hydrogen bonding with active site residues Glu-45, Thr-73, Arg-49 and a water molecule. The pyruvate C-3 methyl group was stabilized by Phe-135 through hydrophobic interactions. The Lys-126 residue was also proposed to be the active site of E. coli KDPGal aldolase.33 Directed evolution of KDPGal aldolase Expression and purification of E. coli KDPGaI aldolase. Meloche and O’Connell39 reported an isolation of KDPGal aldolase from P. saccarophila that involved five chromatographic steps with an overall purification yield of 14%. The final KDPGal 148 aldolase had a specific activity of 130 U/mg. The laborious separation was due to difficulty to achieve complete separation of KDPGal aldolase and KDPG aldolase. Toone and coworkers34 purified KDPGal aldolase from an eda‘ mutant of P. cepacia that was devoid of KDPG aldolase. KDPGal aldolase was purified to a specific activity of 4.2 U/mg with an overall yield of 86% in a single chromatographic step on Sepharose CL-4B derivatized with a Procion Navy H-ER triazine dye. Their attempt to purify E. coli KDPGal aldolase from an eda‘ mutant strain DF214 harboring a plasmid pTC19O containing eda gene using the same chromatographic column, however, was unsuccessful. In Toone’s experiment, KDPGal aldolase was expressed from the chromosomal ngA gene. In this work, E. coli KDPGal aldolase was expressed from a plasmid- localized dgoA gene. E. coli KDPGal aldolase-encoding dgoA gene is localized in the dgo cluster at min 83.4 in the E. coli chromosome. All the genes involved in D- galactonate catabolism including a regulator protein dgoR, 2-keto-3-deoxy-galatonate kinase-encoding dgoK, 2-keto-3-deoxy-6-phosphogalatonate aldolase-encoding dgoA, D- galatonate dehydratase-encoding ngD and D-galatonate permease-encoding dgoT are localized in this cluster sequentially under the control of a single promoter. The E. coli dgoA gene sequence was obtained from the National Center for Biotechnology Information (NCBI) and sequence errors were corrected according to Babbitt et al.33 The published dgoA gene from NCBI contains a fragment of 1763-nt DNA, which comprises the open reading frames of the 618-nt dgoA and the 1149-nt dgoD gene. The three nucleotides sequence errors cause a frame shift in the published amino acid sequences. 149 To explore the catalytic activity of KDPGal aldolase towards phosphorylated D- erythrose (E4P, Figure 58), the 618-nt E. coli ngA was cloned into a cloning vector pCR2.1-TOPO with transcription under the control of a lac promoter to prepare plasmid pNR5.223 (Figure 66). E. coli AB3248,40 in which all three DAHP synthase isoenzymes encoded by aroF, aroG and aroH have been inactivated by three successive rounds of random chemical mutagenesis, was transformed with plasmid pNR5.223. Expression of KDPGal aldolase was induced by addition of 0.2 mM isopropyl fl—D-thioglucopyranoside (IPTG) in Luria-Bertani (LB) medium containing ampicillin. The specific activity of KDPGal aldolase in crude cell-free extract was 30 U/mg. The cell-free extract, after ammonium sulfate treatment, was applied to a DEAE-cellulose column to afford the partially purified KDPGal aldolase in an overall 4% yield and a final specific activity of 87 U/mg. Table 7. Purification of E. coli KDPGal aldolase. total specific entry steps protein activity 5:2: puringtlon yield (mg) w/mg) 1 crude extract 1274 30 38661 1.0 100% 2 (NH4)2SO4 25 % _ 65 % 630 30 18684 1.0 29% 3 DEAE-cellulose 18 87 1600 2.9 4. 1 % Synthesis of 2-keto-3-deoxy-6-phosphogalactonate (KDPGal). KDPGal was chemically synthesized following the synthesis reported by Toone and coworkers (Figure 64).34b The synthesis proceeded from commercially available D-galactonate-l,4-lactone. The lactone was treated with a catalytic amount of sulfuric acid in acetone to afford the 5,6-0-isopropylidine 16 as the sole product in 90% yield. Compound 16 underwent 150 fi—elimination to yield 17 in 34% yield after treated with 10 equivalents of pyridine and 5 equivalents of acetic anhydride at 50 °C for 48 h. The major byproduct of this reaction was the gluco-epimer 18 (20% yield) presumably arising from the corresponding furan (Figure 65). Selective phosphorylation of the primary alcohol following removal of the isopropylidene was carried out with 1.3 equivalents of dibenzyl-N,N- diethylphosphoramidite and 1 equivalent of lH-tetrazole in THF at 0 °C. Without purification, the reaction was cooled to -60 °C, and the intermediate was oxidation with 2 equivalents of 85% m-CPBA in methylene chloride to afford the phosphate 20 in 37% yield. The reported procedure used two equivalents of lH-tetrazole, however, following the procedure resulted in phosphorylation of both the primary and the secondary alcohols in our hand. Reducing the amount of IH-tetrazole to one equivalent led to the formation of the monophosphorylated product. Phosphorylation of 19 was also attempted with tetrabenzylpyrophosphate (TBPP) in presence of sodium hydride in THF at 0 °C. No phosphorylation was observed and the starting material was recovered. Catalytic hydrogenation removed the benzyl ether phosphate protecting groups and yielded the C-2 acetate-protected KDPGal 6—lactone. After removal of THF, the (Ll-lactone was dissolved in water and the pH was adjusted to 7.5 with addition of 50 mM aqueous LiOH solution. Simultaneous hydrolysis of the C-2 acetate and 6—lactone were completed after 48 h to afford the KDPGal lithium salt in 70% yield. 151 O .. O 03¢?" M £2” . “a" ———-> 03‘ —> — + 30H CH OH6 OH OAc D- -galactonate-1, 4-lactone o 0 o )(of..<=26:c HOG—7 (£22: 0 Ho]..(_ Z‘ 0 OAc Ac (BnO)2P- 0 OAc 1a 20 L0 ('15) 0 OH 9H 0 I ' ' . —’e (m 0 . -——“ LiHospo/WLou COle OH HO KDPGaI Figure 64. Synthesis of 2-keto-3-deoxy-6-phosphogalactonate. Keys: (a) acetone, H2804, rt, 4 h, 90%; (b) acetic anhydride, pyridine, CHzClz, 50 °C, 48 h, 34%; (c) 80% glacial acetic acid, rt, 36 h, 86%; (d) i) Dibenzyl-N,N—diethylphosphoramidite (DDP), lH-tetrazole, THF; ii) m-CPBA, 37%; (e) i) H2, 10% Pt/C, THF; ii) aqueous 50 mM LiOH, 70%. 8'] - o 1 0 may): x. {6/9' =7