Copyright by Charles Evans Melançon, III 2006

The Dissertation Committee for Charles Evans Melançon, III certifies that this is the approved version of the following dissertation:

Investigation and Engineering of Macrolide Antibiotic Sugar Biosynthesis and Glycosylation Pathways of Actinomycetes

Committee:

Hung-wen Liu, Supervisor

Walter Fast

David E. Graham

Stephen F. Martin

Christian P. Whitman Investigation and Engineering of Macrolide Antibiotic Sugar Biosynthesis and Glycosylation Pathways of Actinomycetes

by

Charles Evans Melançon, III, B.S.; B.A.

Dissertation Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

The University of Texas at Austin December 2006

Dedication

This dissertation is dedicated to my loving and supportive wife Dr. Lin Hong, who has been with me throughout this work, as a coworker, a friend, girlfriend, spouse, and as the mother of our daughter. I also dedicate this work to my beautiful daughter, Samantha Mei-wen Melançon, who was conceived, was born, and has lived her first few months while this dissertation was taking shape. Lastly, but certainly not least, this work is dedicated to my parents, Charles E. Melançon Jr. and Patricia S. Knight, who never stopped believing that I could achieve anything I put my mind to.

Acknowledgements

The time leading up to writing this dissertation and even more so, the time spent writing it, has been a time of intense reflection for me. Where have I come from? Where am I now? Where am I going? Who and what have been influential in my life during graduate school? My adolescent and early adult years were a time of rapid change for me, and a trying time for my parents. Looking back, I have my parents to thank primarily for the fact that I made it to graduate school in the first place. They never doubted my capabilities even when I doubted them, and always encouraged me to pursue my dreams. They provided me with a stable environment to grow up in, and encouraged my curiosity from a young age. I am eternally grateful to them for all they have done for me. Since my arrival in Austin in July, 2001, my life has changed immeasurably. Joining the Liu Research Group has been a large contributing factor to those changes. I feel like The Liu Group is to science what the Marines are to the military. The training is rigorous, the expectations are high, and the change is permanent. When I started graduate school, I was very enthusiastic about doing research. Dr. Liu’s dedication to his work has inspired me to maintain that enthusiasm. It was also in the Liu Group where I met my wonderful wife, Lin, who has changed my life for the better in innumerable ways.

In the Liu Group, I have learned a great deal about natural product biosynthesis and genetic engineering and have gained an appreciation for mechanistic enzymology. The degree to which we are immersed in scientific research in the Liu Group, from the intense three-and-a-half hour literature reviews to the strong encouragement we regularly

v receive from Dr. Liu to “namely, work hard, and be serious,” has created an environment conducive to my maximal growth and development as a scientist. Dr. Liu has, from the beginning, shown that he believed in me and has encouraged me along the way. He has also given me the latitude to develop my own ideas and to work independently. This freedom has been instrumental in my development as a scientist, and has also given me a tremendous sense of satisfaction in my work. Some of the most important things I’ve learned have been non-scientific. I have learned the importance of being persistent. I have learned that attention to detail that might be considered unnecessary and even inhibitory in other fields can be a tremendous asset in scientific research. I have identified my strengths and my weaknesses and have learned how to maximize the use of my strengths. I have learned how to be assertive. I have learned how to learn as well as how to teach. I have learned how to work effectively and harmoniously in a group setting. All of these skills will no doubt serve me well for the remainder of my life.

Some of my most influential teachers have been the other students and postdocs in the Liu Group. I have learned something from each person with whom I have interacted in the group. I have Sveta to thank for most of my initial training. During the first two years in the lab, she graciously answered hundreds of questions I had, and was instrumental in helping me to develop my laboratory skills. Haruko also taught me a lot

about natural product characterization, and gave me a crash course in NMR interpretation. Lin, in addition to being a great wife and mother to our child, gave me a lot of helpful instruction on protein purification, TDP-sugar preparation, and enzyme assay. Yasushi (the all-knowledgeable) has also been extremely helpful in teaching me everything from computer skills and bioinformatics to molecular biology and chemistry.

vi But everyone in the group, both past and present, has played a role in helping me to get the most out of my graduate experience. I am forever grateful to everyone in the group. During my time in the Liu Group, many of my coworkers have become my friends. The bond shared among coworkers in the Liu Group is unique. We have spent thousands of hours together, shared meals, stories, embarrassing moments, joys, and disappointments. We have grown up together. We have learned each others’ languages. We have influenced each other in ways we may not even realize. Many of the memories we have shared will remain with me for the rest of my life. I hope both the personal and professional bonds we have forged here will remain after we go our separate ways, and that life will bring us together many more times.

In the broadest sense, I am also grateful to Nature itself for the existence of fascinating biochemical systems worthy of detailed study. Without these beautifully intricate phenomena, I would not have nearly as interesting a job as I have now, and would need to find another passion. I feel privileged to live in this time in history when so many interesting questions at the chemistry/biology interface remain unanswered, yet many of the tools to answer them are readily available. I am also thankful to the many scientists who made the discoveries on which our work today is based. As I look to the future, I take comfort in the fact that there will always be new frontiers to explore, and feel confident that I can contribute in a meaningful way to the

collective efforts to explore them. I wish everyone in the Liu Group much success and happiness in the future.

vii Investigation and Engineering of Macrolide Antibiotic Sugar Biosynthesis and Glycosylation Pathways of Actinomycetes

Publication No.______

Charles Evans Melançon, III, Ph.D. The University of Texas at Austin, 2006

Supervisor: Hung-wen Liu

Natural products are an important source of bioactive lead compounds used in drug development. The diverse sugar moieties found in natural product structures are often critical to their bioactivity. Therefore, advances in our understanding of natural product sugar biosynthesis and glycosyltransfer and in our ability to synthesize natural product derivatives through manipulation of the biosynthetic machinery are important and can impact the treatment of human diseases. The work described in this dissertation focuses on the functional elucidation of enzymes involved in biosynthesis and glycosyltransfer of the deoxysugar D-mycaminose, which is a structural component of the macrolide antibiotic tylosin, and on the use of genes encoding the biosynthesis and attachment of D-mycaminose, D-desosamine, and other deoxysugars for the engineered production of macrolide derivatives with altered sugar structures. First, the functions of TylM3 as the activator protein for the glycosyltransferase TylM2 in tylosin biosynthesis in Streptomyces fradiae, and the function of the homologous protein MydC as the activator protein for the glycosyltransferase MycB in

viii mycinamicin biosynthesis in Micromonospora griseorubida were elucidated by expression of combinations of their encoding genes in engineered Streptomyces venezuelae hosts. These studies also showed that these glycosyltransferases have relaxed substrate specificity. During this work, a failed attempt to reconstitute the mycaminose biosynthetic pathway in an S. venezuelae mutant resulted in the discovery of a novel hexose 3,4-ketoisomerase, Tyl1a, which is involved in formation of TDP-D-mycaminose. Discovery of Tyl1a allowed reconstitution of the mycaminose pathway in S. venezuelae, and demonstration that Tyl1a alone could convert the desosamine pathway to a mycaminose biosynthesizing pathway. This work resulted in synthesis of several glycosylated macrolide derivatives. The enzymatic activity of purified recombinant

Tyl1a was characterized in vitro by 1H NMR product analysis and steady state kinetics, and the substrate specificity of Tyl1a was found to be relaxed. Finally, three S. venezuelae mutants expressing hybrid deoxysugar biosynthetic pathways were constructed, one of which resulted in formation of non-natural deoxysugar-bearing macrolides. This work has provided important functional information on the sugar biosynthesis enzymes and glycosyltransferases studied, and has illustrated the feasibility of constructing complex engineered deoxysugar biosynthesis pathways in the macrolide producer S. venezuelae.

ix Table of Contents

List of Tables xv

List of Figures xvi

List of Abbreviations xxii

Chapter 1. Background and Significance 1 1. Introduction 1 2. Deoxysugar Biosynthesis in Natural Product Pathways 7 3. Natural Product Glycosyltransferases 21 4. Natural Product Bioengineering using Glycosyltransferases and Sugar Biosynthesis Enzymes 32 5. Summary and Thesis Statement 48

Chapter 2. Functional Analysis of tylM2/tylM3 and mycB/mydC Pairs Required for Efficient Glycosyltransfer in Macrolide Antibiotic Biosynthesis 50 1. Introduction 50 2. Experimental Procedures 55 General 55 Plasmids and Vectors 56 Bacterial Strains 56 Instrumentation 56 Preparation of Competent Cells 57 PCR Amplification of DNA 58 Construction of desVII Disruption Plasmid pDesVII-K2 59 Construction of Expression Plasmids pCM1 and pCM1b 60 Construction of pCM17 62 Construction of pCM18 62 Construction of pCM4 63

x Construction of pCM13 63 Construction of pCM21 63 Construction of pCM25 64 Construction of pCM26 64 Construction of pCM27 65 Construction of desVII/pAX617 65 Construction of pCM8 66 Conjugal Transfer of pDesVII-K2 into Wild-Type S. venezuelae 66 Conjugal Transfer of pDesVII-K2 into KdesI-80 S. venezuelae 68 Screening for Double-Crossover Mutants 68 Conjugal Transfer of Expression Plasmids into KdesVII and KdesI/VII S. venezuelae 69 Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains 71 Southern Blot Analysis of KdesVII and KdesI/KdesVII Mutants 71 Small-Scale Isolation and Analysis of Metabolites Produced by Mutants 72 HPLC Analysis of Samples from S. venezuelae Mutants 74 Analysis of Time Course for Metabolite Production by KdesI/VII/pCM25 and KdesI/VII/pCM26 Fed Tylactone 75 Large-Scale Isolation and Analysis of Extracts of KdesVII-2-1 and KdesVII/pCM4 S. venezuelae Mutants 75 Large-Scale Isolation and NMR Analysis of Glycosylated Tylactone Derivatives from S. venezuelae Mutants Fed Tylactone 76 3. Results and Discussion 78 Disruption of desVII in Wild-Type and KdesI-80 S. venezuelae 78 Analysis of Metabolites Produced by KdesVII-2-1 80 Design, Construction, and Metabolite Analysis of KdesVII/desVII/pAX617 and KdesI/VII/pCM8 81 Construction of Expression Plasmid pCM1 81 xi Design, Construction, and Metabolite Analysis of KdesVII/pCM4 and KdesVII/pCM13 82 Design, Construction, and Metabolite Analysis of KdesI/VII/pCM21 and KdesI/VII/pCM27 84 Design, Construction, and Metabolite Analysis of KdesI/VII/pCM25 and KdesI/VII/pCM26 86 Analysis of the Time Course of Metabolite Production in KdesI/VII/pCM25 and KdesI/VII/pCM26 in the Presence of Tylactone 90 Large-Scale Isolation and Structure Determination of 5-O- Mycaminosyl-Tylactone and 2’-Glucosyl-5-O-Mycaminosyl- Tylactone 91 Design, Construction, and Metabolite Analysis of KdesVII/pCM17 and KdesVII/pCM18 92 4. Conclusions 94

Chapter 3. Discovery of the TDP-4-Keto-6-Deoxy-D-Glucose 3,4- Ketoisomerase Tyl1a from Streptomyces fradiae and its Use for In Vivo Reconstitution of the D-Mycaminose Biosynthetic Pathway in Streptomyces venezuelae 97 1. Introduction 97 2. Experimental Procedures 102 General 102 Plasmids and Vectors 102 Bacterial Strains 103 Instrumentation 103 Preparation of Competent Cells 103 PCR Amplification of DNA 103 Construction of desVII Disruption Plasmid pDesVII-K2 103 Construction of Expression Plasmids pCM1, pCM1b, and pCM1d 103 Construction of pCM7b 105 Construction of pCM21 105 Construction of pCM23 105 xii Construction of pCM24 106 Construction of pCM30 106 Construction of KdesI-80 S. venezuelae 107 Conjugal Transfer of pDesVII-K2 into KdesI-80 S. venezuelae and Screening for Double-Crossover Mutants 107 Conjugal Transfer of Expression Plasmids into KdesI and KdesI/VII S. venezuelae 107 Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains 108 Southern Blot Analysis of KdesI/KdesVII Mutant 108 Small-Scale Isolation and Analysis of Metabolites Produced by Mutants 108 Large-Scale Isolation and Analysis of Extracts of KdesI/VII/pCM7b and KdesI/pCM23 S. venezuelae Mutants Grown in Vegetative Media 109 Large-Scale Isolation and Analysis of Extracts of KdesI/VII/pCM7b Mutant Grown in PGM Media 110 3. Results and Discussion 111 Design, Construction, and Analysis of KdesI/VII/pCM7b S. venezuelae 111 Identification of ORF 1a and Reconstitution of the Mycaminose Design, Construction, and Analysis of KdesI/pCM23 to Test DesVII Donor Substrate Flexibility 117 Design, Construction, and Analysis of KdesI/pCM24 and KdesI/pCM30 119 4. Conclusions 123

Chapter 4. In vitro Characterization of Tyl1a: Product Identification, Steady-State Kinetic Analysis, and Substrate Specificity Studies 125 1. Introduction 125 2. Experimental Procedures 130 General 130 xiii Plasmids and Vectors 130 Bacterial Strains 130 Instrumentation 130 SDS-PAGE and Determination of Protein Concentration 131 Gene Amplification and Cloning of tyl1a 132 Growth of E. coli BL21-tyl1a/pET28b(+) Cells 133

Purification of N-His6-Tyl1a Protein 133

Cleavage of N-His6-tag from Tyl1a by Thrombin 134 Molecular Mass Determination 135 Enzymatic Synthesis of TDP-4-Keto-6-Deoxy-D-Glucose 135 Preparation of Enzymes used in Enzymatic Synthesis of TDP-4-Keto-6-Deoxy-D-Glucose 137 HPLC Activity Assay for Tyl1a 139 In situ 1H NMR Assay for Tyl1a and Characterization of Products 139 Determination of Kinetic Parameters for Tyl1a 140 HPLC Assay of Coupled Tyl1a-TylB Reaction 141 Enzymatic Preparation of TDP-4-Keto-2,6-Dideoxy-D-Glucose 142 HPLC Analysis of Incubation Mixture Containing Tyl1a with TDP-4-Keto-2,6-Dideoxy-D-Glucose 142 Coupled Assay of Tyl1a-TylB with TDP-4-Keto-2,6-Dideoxy-D-Glucose 143 HPLC Analysis of Incubation Mixture Containing Tyl1a and CDP-4-Keto-6-Deoxy-D-Glucose 144 Coupled Assay of Tyl1a, TylB with CDP-4-Keto-6-Deoxy-D-Glucose 145 3. Results and Discussion 145 Purification and Characterization of Tyl1a 145 Catalytic Properties of Tyl1a 146 In situ 1H NMR Analysis of Tyl1a Reaction and Identification of Reaction Products 148 Determination of Kinetic Parameters for Tyl1a-Catalyzed Reaction 150 xiv Tyl1a Substrate Specificity 150 TylB Substrate Specificity 154 Discussion of the Genomics of Hexose Ketoisomerases 155 Implications for Tyl1a Mechanism and Structure 156 Biosynthesis of 3-Amino-2,3,6-Trideoxyhexoses 157 Nucleotide Specificity of Sugar Biosynthetic Enzymes 158 4. Conclusions 160

Chapter 5. Construction of Non-Natural Sugar Biosynthesis Pathways in Engineered S. venezuelae Hosts 162 1. Introduction 162 2. Experimental Procedures 170 General 170 Plasmids and Vectors 170 Bacterial Strains 170 Instrumentation 171 Preparation of Competent Cells 171 Growth and Maintenance of M. megalomicea 171 Growth and Maintenance of A. thermoaerophilus 171 Colony PCR Using A. thermoaerophilus 172 PCR Amplification of DNA 172 Construction of Expression Plasmid pCM1d 172 Construction of pCM42 172 Construction of pCM43 173 Construction of pCM45 174 Construction of KdesI-80 S. venezuelae 174 Construction of KdesII S. venezuelae 174 Conjugal Transfer of Expression Plasmids into KdesI and KdesII S. venezuelae 175 Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains 175 Small-scale Isolation and Analysis of Metabolites Produced xv by Mutants 175 Large-Scale Isolation and Analysis of Extracts of KdesII/pCM42, KdesI/pCM43, and KdesI/pCM45 S. venezuelae Mutants Grown in Vegetative Media 176 3. Results and Discussion 178 Design, Construction, and Analysis of KdesII/pCM42 178 Design, Construction, and Analysis of KdesI/pCM43 181 Design, Construction, and Analysis of KdesI/pCM45 187 4. Conclusions 193

Bibliography 194

Vita 212

xvi List of Tables

Table 5-1: Comparison of the Chemical Shifts and Coupling Constants of Sugar Signals of 3-O-Mycaminosyl Methynolide and

3-O-(4-epi-D-Mycaminosyl) Neomethynolide 190

xvii List of Figures

Figure 1-1: Structures of D-Ribose and D-Deoxyribose 2 Figure 1-2: Structures of the Nine Common Eukaryotic Sugars 2 Figure 1-3: Early Steps in Formation of TDP-6-Deoxyhexoses 4 Figure 1-4: Selected UDP-Sugar and CDP-Sugar Biosynthesis Reactions 9

Figure 1-5: GDP-D-Mycosamine Biosynthesis 10

Figure 1-6: TDP-L-Rhamnose Biosynthesis 11 Figure 1-7: Desosamine-Containing Macrolide Antibiotics 12 Figure 1-8: Hypothetical Pathways for Desosamine Biosynthesis 13 Figure 1-9: Novel Macrolide Compounds Produced by KdesI, KdesII, KdesV, and KdesVI S. venezuelae Mutants 14

Figure 1-10: TDP-D-Mycaminose Formation in Tylosin Biosynthesis 15 Figure 1-11: Biochemically Characterized TDP-Deoxysugar Pathways 17 Figure 1-12: TDP-Deoxysugar Structures whose Production is Encoded by Sequenced Gene Clusters 20 Figure 1-13: Phylogenetic Analysis of TDP-Deoxysugar Aminotransferases 21 Figure 1-14: Glycosyltransferase Mechanisms 23 Figure 1-15: Order of Glycosylation Events in Tylosin Biosynthesis 27

Figure 1-16: Functional Elucidation of OleG1 and OleG2 by Heterologous Expression in Saccharopolyspora erythraea 29 Figure 1-17: Structures of Calicheamicin and Vancomycin 31 Figure 1-18: Hybrid Compounds made by Heterologous Expression of Actinorhodin Genes 36 Figure 1-19: Structures of Erythromycin A and 6-Deoxy Erythromycin A 36

xviii Figure 1-20: Engineered Glycosylation of A47934 by GtfE' 37 Figure 1-21: Engineered Biosynthesis of Rhamnosyl Derivatives of Methymycin and Narbomycin in S. venezuelae 39 Figure 1-22: Engineered Biosynthesis of Epirubicin in S. peucetius 40 Figure 1-23: Natural and Engineered Biosynthesis of Urdamycins 41 Figure 1-24: Creation of Elloramycin Analogues by Combinatorial Glycosylation 43 Figure 1-25: In vitro Glycorandomization of the Vancomycin Aglycon 46

Figure 1-26: Chemoenzymatic Synthesis of Vancomycin Analogues 47 Figure 2-1: Biosynthesis of Methymycin, Neomethymycin, and Pikromycin in Wild-Type S. venezuelae 51 Figure 2-2: Proposed Role of DesVIII and Analysis of the KdesVIII Mutant 52

Figure 2-3: TDP-D-Mycaminose Formation in Tylosin Biosynthesis 53 Figure 2-4: Glycosylation by MycB in Mycinamicin Biosynthesis 54 Figure 2-5: Plasmid pDesVII-K2 60 Figure 2-6: Plasmids pCM17 and pCM18 62 Figure 2-7: Plasmids pCM4 and pCM13 63 Figure 2-8: Plasmids pCM21 and pCM25 64 Figure 2-9: Plasmids pCM26 and pCM27 65

Figure 2-10: Plasmids desVII/pAX617 and pCM8 66 Figure 2-11: Engineered Glycosylation in KdesVII and KdesI/VII S. venezuelae 79 Figure 2-12: Southern Blot Analysis of KdesI/VII S. venezuelae 80 Figure 2-13: Macrolides Potentially Formed by KdesVII/pCM4 and KdesVII/pCM13 82 Figure 2-14: TLC Analysis of Extracts of KdesVII/pCM4 and

xix KdesVII/pCM13 Fed Tylactone 83 Figure 2-15: Macrolides Potentially Formed by KdesI/VII/pCM21, KdesI/VII/pCM25, KdesI/VII/pCM26, and KdesI/VII/pCM27 84 Figure 2-16: TLC Analysis of Extracts of KdesI/VII/pCM21 and KdesI/VII/pCM27 in the Presence and Absence of Tylactone 85 Figure 2-17: TLC Analysis of Extracts of KdesI/VII/pCM25 and KdesI/VII/pCM26 in the Presence and Absence of Tylactone 88 Figure 2-18: HPLC Traces of Metabolite Extracts of KdesIVII/pCM25

and KdesI/VII/pCM26 in the Absence of Tylactone 89 Figure 2-19: Time Course for TylM2-Dependent Glycosylation of Tylactone in KdesI/VII/pCM25 and KdesI/VII/pCM26 91 Figure 2-20: HPLC Traces of Metabolite Extracts of KdesI/VII/pCM25 and KdesI/VII/pCM26 in the Presence of Tylactone 91 Figure 2-21: Glycosylation of 10-Deoxymethynolide and Tylactone in KdesVII/pCM17 and KdesVII/pCM18 92 Figure 2-22: TLC Analysis of Extracts of KdesVII/pCM17 and KdesVII/pCM18 in the Presence and Absence of Tylactone 93 Figure 2-23: HPLC Traces of Metabolite Extracts of KdesVII/pCM17 and KdesVII/pCM18 in the Absence of Tylactone 94

Figure 3-1: TDP-D-Mycaminose Formation in Tylosin Biosynthesis 98 Figure 3-2: Biosynthesis of Methymycin, Neomethymycin, and Pikromycin in Wild-Type S. venezuelae 99 Figure 3-3: Engineered Glycosylation of Tylactone in KdesI/VII/pCM7b and KdesI/VII/pCM21 101 Figure 3-4: Plasmids pCM7b and pCM21 106

xx Figure 3-5: Plasmids pCM24 and pCM30 107 Figure 3-6: Macrolactones Produced by KdesI/VII/pCM7b 112 Figure 3-7: TLC Analysis of Metabolites produced by KdesI/VII/pCM7b in the Presence and Absence of Exogenous Tylactone 113 Figure 3-8: Formation of Quinovosylated Macrolides by KdesI/VII/pCM7b Fed Tylactone and by KdesI-80 114 Figure 3-9: The tylIBA Region of the Tylosin Biosynthetic Gene Cluster in Streptomyces fradiae Showing the Location of ORF 1a 115

Figure 3-10: TLC Analysis of the Metabolites Produced by the KdesI/VII/pCM21 Mutant Grown in the Presence of Tylactone 116 Figure 3-11: Formation of Glycosylated Tylactone Derivatives by KdesI/VII/pCM21 117 Figure 3-12: TLC Analysis of Metabolites Produced by KdesI/pCM23 118 Figure 3-13: Formation of Glycosylated Methymycin Derivatives by KdesI/pCM23 119 Figure 3-14: Formation of Glycosylated Methymycin Derivatives by KdesI/pCM24 and KdesI/pCM30 120 Figure 3-15: TLC analysis of metabolites produced by KdesI/pCM24 121 Figure 3-16: TLC analysis of metabolites produced by KdesI/pCM30 122

Figure 4-1: TDP-D-Mycaminose Formation in Tylosin Biosynthesis with FdtA Reaction Inset 126 Figure 4-2: Engineered Glycosylation of Tylactone in KdesI/VII/pCM7b and KdesI/VII/pCM21 128 Figure 4-3: Alternate Substrates Accepted by Tyl1a and TylB 129 Figure 4-4: Plasmid tyl1a/pET28b(+) 132

xxi Figure 4-5: SDS-PAGE Gel of Purified Tyl1a-N-His6 and Thrombin-Cleaved Tyl1a 145 Figure 4-6: HPLC Traces of Product Formation in the Tyl1a and TylB Reactions 146 Figure 4-7: HPLC Traces of Tyl1a Reaction Time Course 147 Figure 4-8: Summary Graph of Tyl1a Reaction Time Course Data 148 Figure 4-9: 1H NMR Stack Plot of Tyl1a-Catalyzed Reaction 149 Figure 4-10: Proposed Mechanism of Tyl1a Product Degradation 149

Figure 4-11: Steady-State Kinetic Data for Tyl1a-Catalyzed Reaction 150 Figure 4-12: Plot of the Integration of Select 1H NMR Signals in Tyl1a- Catalyzed Reaction 151

Figure 4-13: Enzymatic Synthesis of TDP-3-Amino-2,3,6-Trideoxy-D-Glucose 151 Figure 4-14: HPLC Traces of Product Formation in the Tyl1a and TylB Reactions Using Alternate Substrates 152 Figure 4-15: Summary Graph of Tyl1a Reaction Time Course Data Using

TDP-4-Keto-2,6-Dideoxy-D-Glucose as Substrate 153

Figure 4-16: Enzymatic Synthesis of CDP-3-Amino-3,6-Dideoxy-D-Glucose 153 Figure 4-17: Summary Graph of Tyl1a Reaction Time Course Data Using

CDP-4-Keto-6-Deoxy-D-Glucose as Substrate 154 Figure 4-18: Proposed Mechanisms of the Tyl1a- and FdtA-Catalyzed Reactions 157 Figure 4-19: Natural and Engineered Biosynthesis of TDP-3-Amino-2,3,6-Trideoxyhexoses 159 Figure 5-1: Structures of Methymycin, Neomethymycin, Pikromycin,

D-Mycaminose, and D-Quinovose 163 Figure 5-2: Proposed Engineered Biosynthetic Pathways for Non-Natural

xxii Sugars in KdesII/pCM42, KdesI/pCM43, KdesI/pCM45 164 Figure 5-3: Functions of TylM1, FdtA, SpnS, TylX3, and MegDII in their Respective Pathways 165 Figure 5-4: Possible Outcomes of Sugar Biosynthesis and Glycosylation in KdesI/pCM43 167 Figure 5-5: Compounds Detected in KdesII/pCM42 and KdesI/pCM43 168 Figure 5-6: Biosynthesis of Non-Natural Sugar-Containing Macrolides in KdesI/pCM45 169

Figure 5-7: Plasmids pCM42 and pCM43 173 Figure 5-8: Plasmid pCM45 174 Figure 5-9: Glycosylated Macrolide Biosynthesis in KdesII and KdesII/pCM42, and the SpnS-Catalyzed Reaction 179 Figure 5-10: TLC Analysis of Metabolites Produced by KdesII/pCM42 180

Figure 5-11: 4-N-monomethylamino-6-deoxy-D-glucosyl Narbonolide 181

Figure 5-12: TDP-D-Angolosamine Biosynthesis in Medermycin Pathway 182 Figure 5-13: Functions of TylM1, TylX3, and MegDII and Possible Outcomes of Sugar Biosynthesis and Glycosylation in KdesI/pCM43 184 Figure 5-14: TLC Analysis of Metabolites Produced by KdesI/pCM43 185 Figure 5-15: Quinovosylated Macrolides Produced by KdesI/pCM43 186

Figure 5-16: Non-Natural Sugar Biosynthesis in KdesI/pCM45 188 Figure 5-17: TLC Analysis of Metabolites Produced by a Small-Scale Culture of KdesI/pCM45 188

Figure 5-18: 1H NMR Structure Assignment of 4-Epi-D-Mycaminose 189

Figure 5-19: Biosynthesis of 4-Epi-D-Mycaminosyl and

xxiii 3-N-Monomethylamino-6-Deoxy-D-Galactosyl Macrolides in KdesI/pCM45 190

Figure 5-20: ESI-MS-MS Spectra of 4-Epi-D-Mycaminosyl and 3-N-

Monomethylamino-6-Deoxy-D-Galactosyl Macrolides 192

xxiv List of Abbreviations

ADP Adenosine diphosphate Apr Apramycin ATCC American Type Culture Collection BCIP 5-Bromo-4-chloro-3-indolyl-phosphate BSA Bovine serum Albumen CDP Cytosine diphosphate CI Chemical ionization CMP Cytosine monophosphate COSY Correlation spectroscopy DIG Digoxigenin DMSO Dimethylsulfoxide DNA Deoxyribonucleic acid dUTP Deoxyuridine triphosphate EDTA Ethylenediamine tetraacetic acid EI Electron ionization ESI Electrospray ionization FPLC Fast protein liquid chromatography

Gal Galactose GalNAc N-Acetylgalactosamine GDP Guanosine diphosphate GlcA Glucuronic acid GlcNAc N-Acetylglucosamine Glu Glucose

xxv GT Glycosyltransferase GTP Guanosine triphosphate HMBC Heteronuclear multiple bond coherence HPLC High performance liquid chromatography HSQC Heteronuclear single quantum correlation IPTG Isopropyl-β-D-thiogalactoside IVG In vitro glycorandomization Kan Kanamycin

LB Luria-Bertani broth LPS Lipopolysaccharide Man Mannose MOPS Morpholinopropane sulphonic acid MS Mass spectrometry NADPH Nicotine adenine dinucleotide phosphate (reduced) NBT Nitro blue tetrazolium NDP Nucleotide diphosphate NeuAc N-Acetyl neuraminic acid NMP Nucleotide monophosphate NMR Nuclear magnetic resonance

NOESY Nuclear Overhauser effect spectroscopy OD Optical density OMT O-mycaminosyl tylonolide ORF Open reading frame PAGE Polyacrylamide gel electrophoresis PCR Polymerase chain reaction

xxvi PCS Polycloning site PKS Polyketide synthase PLP Pyridoxal 5’-phosphate PMSF Phenylmethylsulfonyl fluoride RBS Ribosome binding site RNA Ribonucleic acid SAM S-Adenosylmethionine SDS Sodium dodecyl sulfate

SN2 Bimolecular nucleophilic substitution SPA Sporulating agar TDP Thymidine diphosphate Thio Thiostrepton TLC Thin layer chromatography TPCK Tosyl phenylalanyl chloromethyl ketone TSB Tryptic soy broth TTP Thymidine triphosphate UDP Uridine triphosphate UV Ultraviolet Xyl Xylose

xxvii Chapter 1. Background and Significance

1. INTRODUCTION

Glycosylated molecules are ubiquitous in biological systems and serve a myriad of functions, including information storage and transfer, energy storage, maintenance of structural integrity, molecular recognition, signaling, virulence, and chemical defense. Several types of human diseases are associated with aberrant protein glycosylation patterns,1, 2 and initiation of certain viral infections involves recognition of specific cell surface protein glycoforms.3 Bacterial virulence is often mediated by cell surface polysaccharides,4 and many bacteria use glycosylated small molecules (antibiotics) as chemical weapons with which to gain a selective advantage.5 The sugar moieties of a variety of biomolecules have been shown to be important and often critical for the activity of the parent compounds through both functional studies of molecules with altered glycosylation patterns and through structural studies of glycosylated compounds bound to their targets. Furthermore, it has been demonstrated that alteration of the structures of the sugar moieties of bioactive glycosylated compounds can have profound effects on their activity, selectivity, and pharmacokinetic properties.6, 7 Many of the glycosylated small molecules that bacteria use for defense are clinically important for the treatment of bacterial and fungal infections, cancer, and other human diseases. Glycosylated biomolecules can conveniently be divided into the following categories: nucleic acids, glycoproteins, glycolipids, polysaccharides, and glycosylated

natural products. Nucleic acids possess D-ribose (1) and 2-deoxy-D-ribose (2) moieties which serve structural roles and are involved in formation and maintenance of higher order structures in DNA and RNA, and are important for catalysis by RNA. Eukaryotic

glycoproteins and glycolipids, which are composed of only nine monomers (D-glucose 1 HO HO (Glu, 3), D-galactose (Gal, 4), 2-N-acetyl-2- O O OH OH deoxy-D-glucose (GlcNAc, 5), 2-N-acetyl-

OH OH OH 2-deoxy-D-galactose (GalNAc, 6), D- 1 2 mannose (Man, 7), L-fucose (Fuc, 8), D- Figure1-1 glucuronic acid (GlcA, 9), D-xylose (Xyl, 10) and N-acetyl-neuraminic acid (also known as sialic acid, NeuAc, 11), (Figure 1-2),8 derive the structural diversity of their surface glycans from a combination of the number, position, and identity of the sugars of which they are composed. Prokaryotic cell surface polysaccharides possess a wider array of modified sugar structures than their eukaryotic counterparts, comprising a number of other highly modified sugars in addition to the nine

OH OH OH OH

HO O O HO O HO HO HO OH OH OH OH AcNH OH 345 OH OH OH OH OH O O HO Me O HO HO OH AcNH OH OH OH OH 6 78 OH OH HO OH O O O OH HO HO HO O HO HO AcHN O OH OH OH OH HO 9 10 11

Figure 1-2 common eukaryotic sugars. Therefore their structural diversity derives more from the identities of the component sugars than do eukaryotic glycoconjugates. Interestingly, glycosylated natural products possess an even more staggering array of sugar structures than prokaryotic polysaccharides, making the structural diversity of this class of

2 compounds even more reliant on the variation of individual sugars than either eukaryotic or prokaryotic polysaccharides. In fact, many natural product sugars bear structural modifications that are unique to this class of biomolecules. Recent research on natural product glycosylation has focused on elucidating the enzymatic processes by which the structures of the diverse sugar components are biosynthesized and transferred to aglycon scaffolds with two primary goals in mind: to understand the often unique chemical mechanisms of enzymes involved in formation of these unusual sugars and to lay the groundwork for developing enzymatic methods to change the glycosylation patterns of bioactive natural products. Because the identities of individual sugars, their positions on aglycon scaffolds, and the number of sugar moieties in a chain are all important determinants of natural product bioactivity, the ability to manipulate these factors has important implications for the development of new and more effective drugs. A necessary prerequisite for manipulation of natural product sugar biosynthetic and glycosylation pathways is an understanding of how these pathways operate. This involves identification of genes encoding enzymes involved in the biosyntheses of diverse sugar structures as well as those involved in attachment of these sugars to aglycon scaffolds.

The advent of modern molecular biological techniques has greatly facilitated investigation of biosynthetic processes leading to the formation of a variety of

glycosylated natural products. Many glycosylated natural products are produced by the gram-positive bacteria of the order Actinomycetales, particularly by organisms of the genus Streptomyces. Although several genome sequences of organisms of order Actinomycetales, commonly referred to as Actinomycetes, have been determined, the majority of sequence information on genes involved in the biosyntheses of glycosylated natural products has been determined not from genome sequencing, but from targeted

3 homology-based identification and sequencing of genes encoding their formation. This process is greatly facilitated by the fact that genes involved in formation of a given natural product are almost always clustered in a single locus in Actinomycetes. The results of bioinformatic, genetic, and biochemical studies on natural product deoxysugar biosynthesis in Actinomycetes, which will be covered in more detail in Section 2 of this chapter, indicate that the majority of natural product sugar biosynthetic pathways are likely to proceed via thymidine 5’-triphosphate (TTP)-dependent activation

of D-glucose-1-phosphate (12) forming TDP-D-glucose (13), which then undergoes

conversion to the 6-deoxysugar TDP-4-keto-6-deoxy-D-glucose (14) by TDP-D-glucose 4,6-dehydratase. Compound 14 is the last intermediate common to all 6-deoxyhexose pathways, and is used as the precursor for a diverse set of TDP-sugar products (Figure 1- 3). Extensive biochemical characterization of the enzymes involved in deoxysugar

OH OH O Me O O O diverse HO HO TDP-sugar HO HO HO products OH = OH OH OPO3 OTDP OTDP 12 13 14 Figure 1-3

biosynthetic pathways has revealed some general trends about the reaction sequences to make these sugars which will be discussed in Section 2 of this chapter, and has also allowed the inference that Nature employs a “combinatorial biosynthetic” strategy for generating sugar structural diversity. Namely, only a limited set of nucleotide-sugar modifying enzymes are required for the biosynthesis of diverse sugars, with each enzyme having evolved the necessary substrate specificity to catalyze the specific step in the formation of a target sugar. After the sequencing of many TDP-sugar biosynthetic gene clusters and biochemical characterization of enzymes involved in deoxysugar biosynthesis pathways, our understanding of the formation of these sugars has advanced 4 such that we can now begin to use comparative genomic approaches to accurately predict the biosynthetic pathways of newly discovered sugars. Glycosyltransferases (GTs) are enzymes which catalyze the coupling of sugars, usually supplied in an NDP-activated form, to a diverse set of acceptor molecules. Because of the importance of sugar moieties for natural product bioactivity, these enzymes play critical roles in the biosynthetic pathways of glycosylated natural products. Disruption of GT-encoding genes results in formation of unglycosylated products which generally have reduced bioactivity. GTs are important targets for efforts to alter the glycosylation patterns of natural products through engineering because the specificities of GTs toward both NDP-sugar donor and aglycon acceptor control the structures of final products that can be formed. Thus, GTs may be considered the “gatekeepers” controlling engineered glycosylation of secondary metabolites. Recent studies have shown that several natural product GTs are substrate-flexible, allowing them to be used both in vivo

and in vitro to catalyze the attachment of alternate sugars to natural product aglycons.9 In fact, the exhibition of some degree of substrate flexibility appears to be common among natural product GTs, suggesting the feasibility of using GTs for the synthesis of a wide array of natural product analogues. Because glycosylation plays a vital role in conferring bioactivity to many natural products, studying the factors controlling GT activity and specificity is an important endeavor which can lead to the effective use of

GTs in the synthesis of new glycoforms of natural products with interesting biological properties. Recent characterization of a number of the enzymes involved in sugar biosynthesis and glycosyltransfer has led to the development of several complementary strategies for changing the sugar moieties of natural products. One example is precursor- directed biosynthesis, which involves feeding an analogue of a natural product precursor

5 to a producing strain. Incorporation of this analogue in place of the natural precursor leads to alteration of natural product structure. Another example is that of in vitro chemoenzymatic synthesis, which involves use of a substrate flexible GT to couple chemically or enzymatically synthesized non-native sugar substrates to acceptor molecules.10 In vitro glycodiversification is a variation of this strategy in which non- natural TDP-sugars are generated chemoenzymatically from synthetic sugar-1- phosphates using a substrate-flexible nucleotidylyltransferase and are subsequently used as substrates in glycosyltransfer reactions.11 Strategies for sugar structure alteration involving in vivo manipulation of endogenous sugar biosynthesis pathways through gene disruption and heterologous expression of foreign sugar biosynthetic genes are commonly

referred to as metabolic pathway engineering or combinatorial biosynthesis.10 All of these strategies have proven effective in generating natural product analogs with altered sugar structures, and each has distinct advantages and disadvantages which will be discussed in Section 4 of this chapter. Knowledge and practical experience in the identification of genes encoding enzymes involved in sugar biosynthesis and glycosyltransfer, elucidation of sugar biosynthetic pathways, and manipulation of these pathways to generate natural products with altered glycosylation patterns have grown tremendously in recent years. These studies have brought to light Nature’s ingenuity in creating diverse natural product

glycoforms through combinatorial biosynthetic. By learning Nature’s strategies for synthesizing and transferring modified sugars to aglycon scaffolds and by demonstrating the feasibility of making structural alterations to the sugar components of natural products, we are taking the first steps toward realization of the full potential of this work for the betterment of mankind. Sections 2 through 4 of this chapter will expand upon the concepts and background information introduced here, covering in more detail the

6 subjects of deoxysugar biosynthesis in natural product pathways, glycosyltransferase activity, and strategies for structural alteration of natural product sugar moieties. This background information is not intended to be an exhaustive survey of what is known about these topics, but rather to summarize the major discoveries in each area, highlight gaps in knowledge that are filled by the work presented in this dissertation, and provide a perspective on the field as a whole. The relevance of this work in the study and exploitation of natural product sugar biosynthesis pathways will also be described.

2. DEOXYSUGAR BIOSYNTHESIS IN NATURAL PRODUCT PATHWAYS

As mentioned previously, sugars are key components of many natural product structures. The majority of sugars incorporated into natural product structures are unusual sugars, having one or more hydroxyl groups replaced with a hydrogen atom or other functional group, such as an amine. Sugar biosyntheses, the processes by which sugar precursors derived from primary metabolism are converted to modified sugar structures, will be reviewed in this section. The common strategy that Nature employs to facilitate the transfer of sugars to the acceptor substrates is to activate the sugar donor by adding a nucleotide 5’- diphosphate (NDP) or occasionally a nucleotide 5’-monophosphate (NMP) group to the anomeric carbon of the sugar. The nucleotide phosphate serves as a good leaving group during the glycosyltransfer reaction. TDP-, ADP-, GDP-, CDP-, CMP-, and UDP- activated sugars are known to occur in biological systems. Most NDP-sugars are derived from the corresponding sugar-1-phosphates with the exception of CMP-activated sugars, which are formed by direct coupling to the anomeric hydroxyl group of the parent sugar. Normally, nucleotide activation occurs early in a sugar biosynthetic pathway, with CMP- activation pathways as exceptions.12-14 Transfer of a nucleotide phosphate to a sugar-1- 7 phosphate is accomplished by the nucleotidylyltransferase enzyme. Nucleotidylyltransferases are members of the same enzyme family and thus share sequence homology to each other. As mentioned in Section 1, the majority of sugar

donors used in glycosylation of natural product aglycon are derived from TDP-D-glucose (13) which is biosynthesized by thymidylylation of glucose-1-phosphate (12) (Figure 1- 3). However, cases are known where UDP-sugars are likely to be the donors in natural product formation.15, 16 Likewise, the intermediacy of a GDP-sugar in natural product biosynthesis has been reported.17 As the majority of sugars which are components of natural products are deoxygenated at C-6, the next step in natural product sugar biosynthesis is almost always the formation of an NDP-4-keto-6-deoxyhexose by the action of an NDP-sugar 4,6- dehydratase. Enzymes of this class which catalyze 4,6-dehydration of TDP-, UDP-,

CDP-, and GDP-sugars are known, and belong to a distinct enzyme family.18-21 Thus far, only TDP-sugar 4,6-dehydratases have been shown to be involved in formation of natural product sugar precursors, although those acting on GDP-sugars are suspected to exist in certain natural product pathways.17 Because enzymes of this family display significant sequence identity to one another, it is not possible to know with certainty the nucleotide specificities of these enzymes from gene sequence alone. However, based on biochemical work done on NDP-sugar processing enzymes to date, it is possible to make educated guesses with regard to the identity of the nucleotide used to activate a sugar in a given natural product pathway. Interestingly, all known examples involving 4,6-dehydration of UDP-sugars

employ UDP-2-N-acetyl-2-deoxy-α-D-glucosamine (UDP-GlcNAc, 15) as the substrate, but C-6 deoxygenated products of UDP-GlcNAc have not yet been observed or proposed to be precursors of glycosylated natural products. These are found only in the cell

8 surface polysaccharides of various bacterial species.13, 22 However, there is at least one example in which a UDP-activated sugar is proposed to be an intermediate in the biosynthesis of a glycosylated natural product. As shown in Figure 1-4, UDP-L-lyxose (16) is believed to be the sugar donor in the biosynthesis of the orthosomycin class

antibiotic avilamycin. Compound (16) has been proposed to derive from UDP-D-glucose

(17) via sequential conversion of 17 to UDP-D-glucuronic acid (18), 18 to UDP-D-xylose

(19), 19 to UDP-D-arabinose (20), and finally, 20 to 16 (Figure 1-4). An enzyme encoded by a gene in the avilamycin gene cluster, AviE2, was shown to be essential for

avilamycin formation, catalyzing conversion of 18 to 19.23 The related compound,

evernimicin, possesses an L-lyxose moiety as well, and therefore the biosynthesis of this compound also likely involves 16 as an intermediate.24 Biosynthesis of 16 does not,

however, involve a 4,6-dehydration reaction. The reaction converting CDP-D-glucose

(21) to CDP-4-keto-6-deoxy-D-glucose (22) catalyzed by CDP-glucose 4,6-dehydratase is known to occur in the biosynthesis of several unusual sugars which are components of bacterial cell surface polysaccharides.25 However, there is no evidence indicating that CDP-activated sugar intermediates are involved in natural product pathways.

OH OH OH O O AviE2 O HO O HO O HO HO HO HO HO HO AcNH OH OH OH OUDP OUDP OUDP OUDP 15 17 18 19

OH HO HO O Me O O HO O O HO HO HO OH OH OH OH OUDP OCDP OCDP OH OUDP 21 22 16 20 Figure 1-4

All known GDP-activated sugars are derived from GDP-D-mannose (23) which is formed by the action of a GTP-specific nucleotidylyltransferase. Compound 23 is

9 converted to GDP-4-keto-6-deoxy-D-mannose (24) by GDP-D-mannose 4,6- dehydratase.26 Although it has not been confirmed experimentally, 24 is proposed to be

an intermediate in the formation of GDP-D-mycosamine (25, Figure 1-5), the proposed sugar donor in the biosynthesis of several closely related polyene macrolide antibiotics. This proposal was based on the presence of an open reading frame (ORF) with high sequence identity to known GDP-D-mannose 4,6-dehydratases in each of the four polyene biosynthetic gene clusters. The lack of an obvious candidate to catalyze the C-2 epimerization reaction, which would be required if glucose rather than mannose were the precursor to mycosamine, also supports the proposal. Compound 24 is believed to be

converted to GDP-3-keto-6-deoxy-D-mannose (26) by an as yet unidentified ketoisomerase, and 26 is proposed to undergo C-3 transamination to give 25.27

OH OH Me OH Me OH Me OH O O O O O HO HO HO HO HO H2N O OGDP OGDP OGDP OGDP 23 24 26 25 Figure 1-5

The first TDP-activated sugar whose biosynthesis was investigated was TDP-β-L-

rhamnose (27). L-Rhamnose is commonly found in bacterial cell surface polysaccharides28 and in the sugar appendages of many natural products.29, 30 Enzymatic activities catalyzing formation of 27 were first observed in cellular extracts of Pseudomonas aeruginosa in 1961.31 Genetic characterization of a gene cluster involved in the biosynthesis of the O-antigen of Salmonella enterica led to the identification of four genes, rmlABCD, that are essential for the formation of 27.32 It was not until 1993 that purification and biochemical characterization of the glucose-1-phosphate thymidylyltransferase RfbA (RmlA) from S. enterica, was reported.33 Subsequent studies demonstrated the activities of each of the four purified enzymes in the pathway,

10 establishing that the route for the biosynthesis of 27 involves formation of 14 from 13 by

RmlB, 3,5-epimerization of 14 by RmlC to give TDP-4-keto-6-deoxy-L-mannose (28), and stereospecific 4-ketoreduction of 28 by RmlD to yield 27 (Figure 1-6).34, 35

OH OH Me O O HO RmlA HO O RmlB O HO HO HO OH = OH OH OPO3 OTDP OTDP 12 13 14

RmlC

Me Me O OTDP RmlD O OTDP HO HO O HO OH OH 27 28 Figure 1-6

Several other TDP-deoxysugar biosynthetic pathways have been studied in detail during the past ten years using both genetic and biochemical techniques, and over thirty more have been proposed based on sequence homologies between genes found in glycosylated natural product gene clusters and their well-characterized homologues. Great advances have also been made in our understanding of the biosyntheses of TDP- deoxysugars found as components of bacterial cell surface polysaccharides such as lipopolysaccharides (LPS) and S-layer polysaccharides. With very few exceptions, one of them being 27, the identities of the TDP-deoxysugars used in natural product biosynthesis and those used in polysaccharide biosynthesis are distinct.

The 3-N,N-dimethylamino-3,4,6-trideoxyhexose D-desosamine (29) is a component of a number of macrolide antibiotics, including methymycin (30), narbomycin (31), pikromycin (32), mycinamicin II (33), erythromycin A (34), megalomicin (35), and oleandomycin (36). The organism Streptomyces venezuelae is a producer of 30 as well as the structurally related compounds neomethymycin (37), 31, and 32, each of which

11 contains desosamine (Figure 1-7). The gene cluster encoding production of these compounds was identified and sequenced. Within the cluster are a conserved set of genes, DesI-VIII, which were predicted to be involved in desosamine biosynthesis and

O O Me Me O NMe2 R1 Me HO Me Me R NMe2 O Me O Me Me O NMe Me HO 2 O R2 O 29 O O Me Me HO O O O Me Me Me Me Me 29 O HO HO O Me 29 O 30 R1 = OH, R2 = H O O OMe O OMe 37 R1 = H, R2 = OH Me Me 31 R = H 33 32 R = OH

O O OH O O Me Me Me Me O Me Me HO OH HO NMe2 Me O Me Me NMe Me OH NMe2 OH 2 Me OH NMe2 Me HO Me HO Me HO Me O O O O Me Me O O O Me Me O O Me 29 Me 29 O O 29 O OMe O O O Me OH Me Me Me Me OMe OH OH O O OH O Me Me Me 34 35 36

Figure 1-7 glycosyltransfer. DesIII and DesIV display significant similarity to previously characterized glucose-1-phosphate thymidylyltransferase RmlA and TDP-D-glucose 4,6- dehydratase RmlB, and were assigned the functions of conversion of 12 to 13, and 13 to 14, respectively. Enzymological and structural studies of DesIV demonstrate that it does indeed catalyze conversion of 13 to 14.36 DesI and DesV show homology to PLP- dependent enzymes found in other sugar biosynthetic pathways. DesII shows homology to putative SAM-radical-dependent enzymes, and DesVI is similar to SAM-dependent methyltransferases. DesVII is homologous to glycosyltransferases (GTs), and DesVIII appears to be distantly related to P-450 enzymes, yet lacks the conserved cysteine residue which coordinates to the heme iron. Initially, it was proposed that DesVIII catalyzes 3,4- ketoisomerization of 14 to form TDP-3-keto-6-deoxy-D-glucose (38). Next, the proposed C-4 dehydrase DesI and reductase DesII together were thought to catalyze C-4 12 deoxygenation of 38 to form TDP-3-keto-4,6-dideoxy-D-glucose (39). DesV would then

act as a C-3 aminotransferase, converting 39 to TDP-3-amino-3,4,6-trideoxy-D-glucose (40). Finally, DesVI was thought to catalyze the final 3-N,N-dimethyltransfer reaction in

the pathway to form TDP-D-desosamine (41) (Figure 1-8, Path A). Gene disruption experiments were performed on several of the proposed desosamine pathway genes in

Me O H2N HO OH DesI OTDP DesII 48

OH PathB Me Me O Me O O O DesV HO O DesIV H N HO HO 2 O OH OH OH OH OTDP OTDP OTDP OTDP 13 14 Path A 39 40

Me DesIII DesVIII O DesI DesVI HO DesII OH O OH OTDP Me O 38 HO HO O HO Me2N OH = OH OPO3 OTDP 12 Figure 1-8 41

order to test their proposed functions. NMR analysis of the macrolides produced by the gene disruption mutants of desI, desII, desV, and desVI showed that each mutant produced a small amount of particular glycosylated methymycin derivatives each with an altered sugar structure. In the case of the DesVI disruption mutant, KdesVI, 3-N- acetylamino-3,4,6-trideoxy-D-glucosyl derivatives of methynolide (42) and neomethynolide (43, Figure 1-9) were produced. Formation of these compounds was consistent with the predicted role of DesVI as the 3-N,N-dimethyltransferase catalyzing

the final step in the biosynthesis of TDP-D-desosamine (41). The acetylation of the 3- amino group of 40 was suggested to be catalyzed by a promiscuous cellular acetyltransferase.37 Subsequent in vitro assay of purified DesVI using chemically synthesized 40 as substrate confirmed its function as the 3-N,N-dimethyltransferase in the pathway.38 Disruption of desV led to the accumulation of 4,6-dideoxy-D-glucosyl

13 O O O O Me OH OH Me Me Me Me Me O DesIII HO HO O Me Me Me Me HO Me O O O Me Me HO HO O O HO HO HO OTDP O O Me O O O O O O OPO = Me Me DesIV 3 Me Me Me Me Me ABC D

O Me Me Me Me Me O O O O O DesI H2N DesII DesV DesVI HO HO H2N Me2N HO HO HO HO HO OTDP OTDP O OTDP OTDP OTDP 14 48 39 40 41

DesVII DesVII DesVII DesVII DesVII

Me Me Me Me Me O O HO O AcHN O O R HO R HO R HO R AcHN Me2N R OH OH OH OH OH

46 R = C 44 R = A 42 R = A 30 R = A 49 R = C 47 R = D 45 R = B 43 R = B 37 R = B KdesI KdesII KdesV KdesVI Wild type

Figure 1-9 derivatives of methynolide (44) and neomethynolide (45), supporting the prediction that C-3 aminotransfer is catalyzed by DesV and that this reaction occurs after C-4 deoxygenation. A stereospecific reduction of the C-3 keto group of 39 by an endogenous reductase was proposed to account for the presence of the hydroxyl group at C-3 of 44 and 45.39 Surprisingly, disruption of desII led to accumulation of 4-N-acetylamino-4,6-

dideoxy-D-glucosyl derivatives of 10-deoxymethynolide (46) and narbonolide (47). This finding led to a reexamination of the proposed functions of DesI and DesII. Subsequent in vitro characterization of DesI showed that it was incapable of acting on the originally proposed substrate 38, but instead catalyzed C-4 aminotransfer of 14 to form TDP-4- amino-4,6-dideoxy-D-glucose (48, Figure 1-8, Path B).40 Furthermore, disruption of desI resulted in formation of quinovosyl 10-deoxymethynolide (49), consistent with the role of DesI as a C-4 aminotransferase.41 From these results, it became clear that DesII likely catalyzes an oxidative deamination reaction converting 48 to the 3-keto sugar 39. Indeed, using purified DesII and DesV as the catalysts, the DesI product 48 was converted to a

14 new compound with high resolution mass spectrum consistent with it being 39.42 Thus,

the biosynthesis of TDP-D-desosamine (41) has now been completely elucidated, and proceeds via Path B shown in Figure 1-8.

The 3-N,N-dimethylamino-3,6-dideoxyhexose D-mycaminose (49) is found as a substituent of a number of 16-membered ring macrolide antibiotics, including leucomycins, carbomycins, maridomycins, platenomycins, midecamycins, and spiramycins. It is also the first sugar attached to tylactone (50), a 16-membered macrolactone, in the formation of the macrolide antibiotic tylosin (51) in Streptomyces fradiae (Figure 1-10). The biosynthetic pathway of TDP-D-mycaminose (52) in S. fradiae has been studied for nearly ten years. Extensive genetic and phenotypic complementation studies revealed the genetic organization of the tylosin (tyl) biosynthetic gene cluster in which the tylG region harbors the polyketide synthase (PKS) genes for making tylactone and the flanking tylLM, tylIBA and tylCK regions contain the

O Me

Me Me

Me Me O O O OH Me Me Me Me CHO O O OH HO Me Me NMe Me NMe2 O 2 HO OH HO O 50 TylM2 OMe OH Me Me O Me OMe Me O Me O O O O 49 Me OH Me O OH O OH O O Me HO Me Me Me2N 51 OH 52 OTDP

TylM1 OH OH Me Me O Me O O TylM3 TylA1 O HO TylB HO O TylA2 HO O HO H2N HO HO HO OH O OH OH OTDP OTDP OH OH = 53 38 14 OTDP 13 OTDP 12 OPO3 Figure 1-10

genes for unusual sugar formation.43 The tylLM, tylIBA and tylCK regions were sequenced in previous studies, and 17 open reading frames (ORFs) were identified within these regions.44 Sequence alignments with other sugar biosynthetic genes, especially those reported by Cundliffe and coworkers who had also sequenced the tylIBA and tylLM 15 regions of the tyl cluster,45 led to the assignment of tylA1, tylA2, tylB, tylM1, tylM2, and tylM3 as genes involved in mycaminose formation and attachment. The tylA1, tylA2, and tylM2 genes all show high sequence identity with their well- characterized counterparts in other sugar biosynthetic pathways, and thus were assigned

the following functions: tylA1 encodes an α-D-glucose-1-phosphate thymidylyltransferase

responsible for conversion of 12 to 13, tylA2 encodes a TDP-D-glucose 4,6-dehydratase converting 13 to 14, and tylM2 encodes a glycosyltransferase responsible for the attachment of 49 to tylactone (50). The tylB and tylM1 genes encode a putative PLP- dependent aminotransferase and an SAM-dependent methyltransferase, respectively. TylB and TylM1 are homologues of DesV and DesVI from the desosamine pathway, respectively, and are predicted to convert TDP-3-keto-6-deoxy-D-glucose (38) to TDP-3-

amino-3,6-dideoxy-D-glucose (53), and 53 to 52, respectively. Subsequently, TylB and TylM1 were purified and assayed in vitro using chemically synthesized 53 as substrate. TylM1 was shown to convert 53 to 52 using SAM, and TylB was shown to catalyze the reverse aminotransferase reaction, converting 53 to 38 using α-ketoglutarate as the

cosubstrate.44, 46 The process by which TDP-4-keto-6-deoxy-D-glucose (14) isomerizes to 38 remains unknown. On the basis of two early reports in which a portion of 14 was transformed to 38 during purification by Dowex-1 ion exchange chromatography, a reversible non-enzymatic ketoisomerization between 14 and 38 was thought to occur.47, 48 However, in the study of TylB, when the reaction was run in reverse using 53 and α- ketoglutarate as substrates, only 38 was produced and no trace of 14 was detected. Also, no product was formed upon incubation of 14 with TylB. These results indicated that, at least under the in vitro conditions used, there was no chemical isomerization between 14 and 38, suggesting that the 3,4-ketoisomerization is more likely enzyme-catalyzed.46 This activity was tentatively assigned to the tylM3 gene product, which displays low

16 sequence similarity to P-450 enzymes but lacks the conserved cysteine residue that coordinates the heme iron. Interestingly, TylM3 is a homologue of DesVIII, which was proposed to catalyze ketoisomerization of 14 to 38 in the desosamine pathway (Figure 8, Path A), but whose role in the pathway was questionable after establishing the functions of DesI and DesII. Thus, the enzyme responsible for the proposed ketoisomerization step in the biosynthesis of 52 is unlikely to be TylM3. The biosyntheses of several other natural product deoxysugars, such as those of

TDP-L-mycarose (54),49-52 TDP-D-forosamine (55),53, 54 TDP-L-eremosamine (4-epi-L-

vancosamine, 56),55 and TDP-L-noviose (57)56 have also recently been studied in vitro, resulting in elucidation of the details of the biosynthetic route to 54-56, and preliminary evidence in support of a biosynthetic pathway for 57 (Figure 1-11). Several important general principles have been learned from these studies. First, in each of these pathways, as well as those of 41 and 52, the sugars are TDP-activated, and C-6 deoxygenation of 13 to form 14 occurs initially after thymidylylation. Second, a variety of enzymatic

Me Me NovW Me NovU NovS Me O OTDP Me O OTDP O OTDP HO O HO HO HO OH O OH OH 57

OH O Me Me Me Me Me O O DesI O O HO H2N DesII DesV O DesVI O HO HO HO H2N Me2N OH OH OH O OH OH OH OTDP OTDP OTDP OTDP OTDP 13 14 48 39 40 OTDP 41

OH Me Me Me O TylM3 HO TylB HO O TylM1 HO O HO O H2N Me2N HO O OH OH OTDP OH OH = OTDP OTDP OPO 38 53 52 12 3

TylX3 Me SpnO O O Me O Me O O O OH OH EvaA TylC1 TylC3 Me TylC2 TylK O OTDP Me OTDP HO O HO O OTDP OH OTDP Me OTDP O Me Me 54

Me Me O O Me Me SpnN O SpnQ O SpnR O SpnS H2N Me N O HO 2 OTDP OTDP OTDP 55 OTDP

Me O Me Me Me EvaB O EvaC O EvaD EvaE O Me O OTDP Me O OTDP Me H2N HO O NH2 NH2 OTDP NH OTDP 2 56 Figure 1-11

17 activities which operate downstream of formation of 14 have been identified, and their substrates confirmed. These include enzymes involved in C-2 deoxygenation (TylX3, SpnO, EvaA), C-3 deoxygenation (SpnQ), C-4 deoxygenation (DesII), C-3 transamination (DesV, TylB, EvaB), C-4 transamination (DesI, SpnR), C-3 ketoreduction (TylC1, SpnN), C-4 ketoreduction (TylC2, EvaE), C-3 methylation (TylC3, EvaC), 3- N,N-dimethylation (DesVI, TylM1), 4-N,N-dimethylation (SpnS), C-5 epimerization (TylK, EvaD), and C-3, C-5 epimerization (RmlC). These enzymatic activities account for most of those expected to exist in unusual sugar biosynthetic pathways based on the sugar structures found in natural products discovered thus far. Types of natural product sugar biosynthetic enzymes predicted to exist but have not yet been functionally verified in vitro include 3-epimerases, 4-N-monomethyltransferases, and 5-C-methyltransferases. Thus, a relatively small set of enzymatic activities seem to be sufficient to generate these diverse structures.

Also, based on the substrate specificities of these characterized enzymes, some general principles about the order of steps in these sugar biosynthesis pathways can be inferred. For instance, in all 2,6-dideoxysugar pathways, C-2 deoxygenation occurs immediately after C-6 deoxygenation. In the case of 2,3,4,6-tetradeoxysugar 55, C-3 deoxygenation occurs after the C-2 deoxygenation/C-3 ketoreduction step. For the 4,6- dideoxysugar 41, C-4 deoxygenation by DesII requires prior 4-aminotransfer by DesI,

and occurs after C-6 deoxygenation. Thus, from the information obtained so far, the order of deoxygenation steps is C-6ÆC-2 for 2,6-deoxysugars, C-6ÆC-2ÆC-3 for 2,3,6- trideoxysugars, and C-6ÆC-4 for 4,6-dideoxysugars. Transamination can occur at C-3 or C-4 using substrates with varying degrees of deoxygenation, and in the cases studies so far gives products carrying an equatorial amine group. Ketoreduction can also occur at C-3 or C-4, and generates either of the two possible stereochemical outcomes,

18 depending on the specificity of the enzyme. C-4 ketoreduction seems to occur as the final step of the pathways characterized thus far, probably due to the necessity of the C-4 keto group for the chemical transformations of many of the enzymes in these pathways. 3-C-methyltransfer in the two cases studied thus far proceeds with inversion of stereochemistry at the methylated center. N,N-dimethyltransfer can occur at either a C-3 or C-4 amine, and usually takes place at a late stage in the pathway. C-5 epimerization also occurs at a late stage of the pathway, typically followed by C-4 ketoreduction. These observations can be used as guidelines for the de novo prediction of natural product sugar biosynthetic pathways.

In addition to the aforementioned sugar biosynthesis pathways that have been characterized in vitro, there are more than 30 other unusual TDP-sugars used in natural product biosynthesis for which gene clusters have been identified and sequenced. This number is currently increasing at a rate of several pathways per year. Figure 1-12 shows a collection of TDP-sugar structures predicted to be the final products of the biosynthetic pathways encoded in sequenced natural product gene clusters. Multiple gene clusters encoding the production of some of these sugars are also known. As a result, it is now possible to perform comparative genomic analysis of sugar biosynthetic genes in these clusters to make predictions about gene functions, substrate specificities, and the order of enzymatic steps in these sugar biosynthetic pathways. An illustration of comparative genomics as a functional prediction tool applied to a class of natural product sugar biosynthetic enzymes is presented here. Within the sugar aminotransferase class, of which DesI, DesV, TylB, EvaB, and SpnR have been characterized biochemically, multiple sequence alignment of all predicted sugar aminotransferases found in natural product biosynthetic gene clusters and subsequent phylogenetic tree construction (Figure 1-13) reveals subclasses within this group which correlate well with predicted substrate

19 Me Me O O HO Me O OTDP Me O OTDP Me O OTDP HO Me O OTDP HO HO HO HO OH HO OH OH OH OTDP OH HO OTDP TDP-6-deoxy-D-allose 27 TDP-2-deoxy-L-fucose TDP-L-olivose TDP-D-olivose TDP-L-amicetose HO HO OH HO Me HO O Me O Me Me O Me OTDP Me O OTDP O O HO HO OTDP HO HO HO HO O OTDP OTDP OTDP OH TDP-D-fucose TDP-D-fucofuranose TDP-L-digitoxose TDP-D-oliose TDP-4-keto-2,6- TDP-L-rhodinose dideoxy-D-glucose Me Me Me OH OH O O Me OTDP O Me OTDP Me O OTDP O HO Me O OTDP O HO HO HO Me MeO Me HO OH OH OH OTDP Me Me OTDP OH TDP-6-deoxy-3-C- TDP-4-keto-2,6-deoxy-3- 54 TDP-D-mycarose TDP-L-axenose TDP-L-oleandrose methyl-L-mannose methyl-D-glucose Me Me Me Me Me Me Me O Me O OTDP Me O OTDP Me O OTDP O OTDP Me O OTDP HO HO HO HO HO OH OH O NH2 NH2 NH OH HO 2 Me OTDP TDP-L-chromose TDP-L-oxo- TDP-L-vancosamine TDP-2-deoxy-D-evalose 57 vancosamine 56 Me Me Me Me Me O O O O Me OTDP HO Me O OTDP MeHN H2N HO O HO Me2N HO Me HO NMe2 NH2 OTDP OTDP HO OTDP NO2 OTDP TDP-4,6-deoxy-D-glucose TDP-3-N,N-dimethyl- TDP-D-angolosamine TDP-L-daunosamine TDP-D-vicenisamine TDP-2,3,6-trideoxy-3-nitro-3-C- L-vancosamine methyl-4-amino-D-glucose Me NMe Me Me 2 Me O Me O OTDP Me OTDP Me OTDP O Me OTDP Me2N HO O O O O Me N 2 Me N HO Me2N OH 2 OH OH NMe2 NMe2 HO OTDP HO OTDP HO OTDP TDP-6-deoxy-5-C-methyl- TDP-2-deoxy- TDP-L-rhodosamine 41 TDP-L-megosamine 55 4-N,N-dimethyl-L-allose 52 L-nogalamine Figure 1-12 specificities. The clade in which DesV falls (Figure 1-13, top) is made up of seven aminotransferase homologues found in gene clusters that encode formation of desosamine-containing natural products, suggesting these seven enzymes catalyze the identical reaction as DesV, 3-aminotransfer to the 4,6-dideoxy substrate 39. Similarly, TylB and the members of its clade are found in gene clusters for natural products containing mycaminose, suggesting these catalyze the same reaction as TylB. In general, members of each subgroup in the phylogenetic tree are predicted to catalyze the same reaction using the same substrate, and their positions within the tree jibe well with the final sugar structure found in each natural product and the coincidence of other biosynthetic genes thought to be involved in formation of that sugar. A similar type of analysis performed on other classes of natural product sugar biosynthetic enzymes can also be helpful for informed de novo prediction of biosynthetic pathways, although it is necessary to point out that this type of analysis is not an adequate substitute for in vitro

20 characterization. Nevertheless, the results generated from this type of analysis need to be considered when predicting natural product sugar biosynthetic pathways.

A. erythreum EryCI (0.1613) EryCI (0.1546) DesV (0.0326) 4,6-dideoxysugar NbmG (0.0386) 3-aminotransferases MydD (0.1537) OleN2 (0.1645) MidC (0.1008) Spi28 (0.0883) 6-deoxysugar TylB (0.1785) 3-aminotransferases BusR (0.0297) SpnR (0.0338) 2,3,6-trideoxysugar Spi43 (0.1940) 4-aminotransferases Kij7 (0.2269) ChmCIV (0.0307) GerB (0.0360) MydE (0.1661) OleN1 (0.1545) 6-deoxysugar EryCIV (0.1525) 4-aminotransferases Lkm38 (0.1457) DesI (0.0364) NbmK (0.0479) SgcA4 (0.1872) 2,6-dideoxysugar VinF (0.1843) 4-aminotransferases MegDII (0.1381) AknZ (0.0944) DvaB (0.0425) EvaB (0.0415) StaI (0.0710) SnogI (0.1018) 2,6-dideoxysugar DnmJ (0.1284) 3-aminotransferases Hed20 (0.0895) Med20 (0.0799) Kij2 (0.0886) RubN4 (0.1259) Figure 1-13

As one can see, great progress has been made in our understanding of how nucleotide-activated sugars used in natural product glycosylation are biosynthesized. It is apparent after reviewing the immense structural diversity of nucleotide-activated sugars from the natural product pathways whose clusters have been sequenced thus far and the comparatively few types of enzymatic activities used to generate them that Nature has evolved to adopt a “combinatorial biosynthetic” approach to generate a wide variety of sugar products which impart diverse biological activities on the parent compounds.

3. NATURAL PRODUCT GLYCOSYLTRANSFERASES

Glycosyltransferases (GTs) are enzymes which catalyze the coupling of activated sugar donors, usually in the form of NDP-activated sugars, to a diverse set of 21 nucleophilic acceptor molecules to form O-, N-, S-, and C-glycosidic linkages. GTs specific for both saccharide and non-saccharide acceptor molecules are well-known. These enzymes are found widely in Nature, and are involved in the formation of peptidoglycan and LPS in bacteria, protein glycosylation and glycoconjugate formation in eukaryotes, cell wall formation in plants, and glycosylation of natural products in bacteria and plants. More than 12,000 glycosyltransferase genes have been identified to date, including more than one hundred in various natural product biosynthetic gene clusters.57 Owing to their widespread involvement in important biological processes in many organisms, there has been much interest in studying glycosyltransferase mechanisms, structures, and substrate specificities. In particular because of interest in altering the glycosylation patterns of natural products, one of the recent focuses in natural product GT research is to identify GTs with relaxed specificity. These GTs would be useful for alteration of the glycosylation patterns of natural product structures. There is also interest in broadening or altering the specificities of natural product GTs using structural information or random mutagenesis.

GTs can be classified into two groups, inverting and retaining GTs, based on the stereochemistry at the anomeric center of the glycosylated product relative to that of the donor sugar.58 Studying the mechanisms by which GTs catalyze glycosyltransfer is important for active site engineering strategies to broaden or alter specificity. Studies on

inverting GTs support an SN2-like mechanism. As shown in Figure 1-14A, an active site base facilitates catalysis by deprotonation of the acceptor nucleophile which then attacks the anomeric carbon, releasing the NDP group. For retaining GTs, a double displacement mechanism involving attack of a nucleophilic active site residue forming a covalent sugar-enzyme intermediate and releasing NDP had been proposed. This is followed by the attack of the acceptor nucleophile from the opposite face of the anomeric carbon to

22 A B BH+ NDP H O OR O HO HO OR

ONDP

B Nuc Nuc NDP O O O HO HO Nuc HO

ONDP OR H OR BH+ B

C NDP O O HO HO

OR NDPO H OR Figure 1-14

release the enzyme nucleophile to give product (Figure 1-14B). This mechanism is analogous to that of the well-studied retaining glycosidases. However, the crystal structural analysis and biochemical and mutagenesis studies of the retaining GT LgtC from Neisseria meningitidis failed to locate the active site nucleophile. Proposals involving use of the donor substrate C-6 hydroxyl, an active site glutamine, or active site

aspartate as the nucelophile have been put forth, as has an SNi mechanism involving direct nucleophilic attack by the acceptor from the same face as the NDP moiety (Figure 1-14C).59, 60 Further experiments are necessary to discern which mechanism is operative. Until recently, structural studies of GTs were scarce. By 1999, only one GT structure had been determined. By the end of 2000, there were five. Currently, the structures of at least 25 different GTs, including three which are involved in bacterial natural product glycosylation, are known.57 The structures of GTs fall into two superfamilies, GT-A and GT-B. Recently, the structure of the sialyltransferase CstII, 23 which displays structural similarity to GT-A superfamily members yet possesses significant differences, was solved, and was classified as the first representative of a new structural superfamily.61 The GT-A superfamily is characterized by a single domain with α/β/α sandwich topology which resembles a Rossman fold. The GT-A superfamily members also possess a conserved DXD motif which bind a catalytically essential divalent metal, usually Mn2+ to coordinate with the diphosphate of the donor substrate. GT-B superfamily members have two Rossman fold-like domains with a cleft between domains where substrates bind. They are metal-ion independent, and lack any universally conserved motifs. Despite extremely low sequence similarity (>10% identity) among GTs within either GT-A or GT-B superfamily, the three dimensional structures of GTs within a superfamily are quite similar. Nearly all bacterial natural product GTs are believed, based on sequence analysis, to be members of the GT-B superfamily. The three natural product GTs whose X-ray crystal structures have been solved are GtfA, GtfB, and GtfD, which catalyze the glycosylation of the glycopeptide antibiotics vancomycin and chloroeremomycin. These antibiotics are effective against Gram-positive bacterial infections, including methicillin-resistant Staphylococcus aureus, and are considered the “antibiotics of last resort” for the treatment of multi-drug resistant bacterial infections. The structure of free GtfB was the first of these to be solved.62 As typical for a GT-B superfamily member, the enzyme has two domains separated by a flexible linker region, forming a deep cleft between the two domains. Molecular modeling of the donor and acceptor substrate binding to GtfB revealed that the N-terminal domain contains the aglycon binding site, and the C-terminal domain contains the sugar binding site. Since these two domains appear to be well-separated, it may be possible to create chimeric GTs with new glycosyltransfer specificities. Structures of GtfA and GtfD, each with acceptor substrate and TDP bound, were later determined.63, 64 It was found that two

24 conformational states are possible for the GtfA enzyme: open and closed. The open state was seen when acceptor substrate was bound, and the closed form was observed when both acceptor and donor substrates were bound, suggesting that TDP-sugar binding triggers the formation of the closed complex. There were no interdomain contacts observed for the GtfA structure. In contrast, the structure of GtfD in the closed conformation revealed several interdomain contacts. These results indicated that creation of chimeric or engineered GT variants may be more complicated than was surmised from information obtained from the GtfB structure alone. To date there is only a single example in which natural product GT engineering was successful in altering substrate specificity. This example will be discussed in Section 4.

The first in vitro demonstration of the activity of a GT that glycosylates a bacterial natural product was the study of the resistance GT OleI from Streptomyces antibioticus.65 This enzyme catalyzes the coupling of D-glucose to the C-2 hydroxyl of the desosamine moiety of the macrolide antibiotic oleandomycin, forming an inactive product as part of a self-protection mechanism. Another resistance GT from the same strain, OleD, was also characterized in vitro and shown to catalyze the same reaction as OleI, but with broader specificity for the acceptor substrate, presumably in order to protect the cell from attack by macrolides produced by other strains.66 In 2001, the in vitro activities of four GTs, GtfB, C, D, and E, which recognize vancomycin and

teichoplanin aglycons, were demonstrated.67 During the course of the research described in this thesis, the in vitro activities of seven other natural product GTs have been demonstrated. These are DesVII, the desosaminyltransferase from the biosynthetic pathway of the macrolide antibiotic methymycin of S. venezuelae, whose genetic inactivation will be described in this thesis,68 the desosaminyltransferase EryCIII from the biosynthetic pathway of the macrolide antibiotic erythromycin of Saccharopolyspora

25 erythraea,69 the 2-deoxy-L-fucosyltransferase AknK from the biosynthetic pathway of the anthracycline antibiotic aclacinomycin,70 the L-daunosaminyltransferase AknS, also from

the aclacinomycin pathway,71 the L-noviosyltransferase NovM from the novobiocin

biosynthetic pathway of Streptomyces sphaeroides,72 and the L-rhamnosyltransferase CalG1 and the aminopentosyltransferase CalG4 from the calicheamycin biosynthetic pathway of Micromonospora echinospora.73 Work in this area was hampered by the difficulty in obtaining the proper NDP-sugar and aglycon substrates, which must be prepared via chemical or enzymatic synthesis or by isolation from the appropriate mutant. This situation was further complicated by difficulties in protein overexpression and purification. It should be mentioned that the activities of a number of plant natural product GTs have been demonstrated in vitro in recent years.74 Gene disruption, complementation, heterologous expression, and bioconversion experiments have been used to confirm the functions of a number of natural product GTs. For example, tylosin (51, Figure 1-15), produced by S fradiae, possesses three sugars in its structure and three GT-encoding genes, tylM2, tylN, and tylCV, exist in the tylosin biosynthetic gene cluster. Disruption of each of these genes followed by product analysis revealed which GT catalyzes the attachment of which sugar, and the preferred order of attachment. Disruption of tylM2 yielded a small amount of tylactone (50), yet the mutant strain was still able to convert exogenously fed O-mycaminosyl tylonolide (OMT, 58), demycinosyl tylonolide (DMT, 59), and desmycosin (60) to 51. These observations strongly suggesting that TylM2 catalyzes the coupling of D-mycaminose (49) to 50 and that 49 is the first sugar to be attached to 50.75 Disruption of tylN led to the accumulation of (59). It was also found that the fed 58 could be converted to 59, and 60 could be

converted to 51, indicating that TylN catalyzes the coupling of the D-deoxyallose (61) moiety which is likely the second sugar to be attached in the preferred biosynthetic

26 O O Me Me CHO Me Me Me TylM2 NMe2 Me Me HO OH O OH Me O HO O O Me O OH 49 O OH Me 50 Me TylCV 58 TylN

O O Me Me Me CHO CHO O HO Me Me NMe HO NMe2 O 2 HO O HO OH Me O OH OMe OMe Me O Me O O Me O O Me 62 O OH O OH O OH Me Me 63 Me 59 60

TylN TylCV O Me Me CHO O HO Me O NMe2 HO O OMe OMe Me O OH O O Me 62 Me O OH O OH Me Me 63 51 Figure 1-15

pathway, although the GT that attaches L-mycarose (63) can act on 58 that accumulates in the tylN knockout strain.76 Subsequent O-methylation at the 2- and 3- positions forms

the D-mycinose (62) moiety of 51. Disruption of tylCV led to accumulation of 60, supporting the role of TylCV as the mycarosyltransferase responsible for the final glycosylation step.77 The combined use of gene disruption and bioconversion of exogenously fed intermediates in the study of tylosin biosynthesis is an elegant demonstration of GT functional analysis in vivo. Similar techniques have also been used successfully for the elucidation of GT function in a number of other natural product systems. However, these methods can sometimes lead to inconclusive results due, in general, to the nature of the in vivo system used. Specifically, gene disruption can have polar effects on downstream genes, complicating the interpretation of results. In

27 addition, certain Actinomycete strains are refractory to genetic manipulation, making gene disruption impossible. Also, as seen in the tylosin pathway, biosynthetic pathways can proceed in a branched rather than a linear fashion, with several intermediate steps proceeding in a random or non-obligatory order. This can also complicate the interpretation of the results. Some of these complications can be circumvented by using heterologous expression or complementation strategies to deduce gene function. One example of the use of interspecies complementation experiments to provide support for GT function is that of the functional studies of the GTs OleG1 and OleG2 from the oleandomycin (36) producer Streptomyces antibioticus. These two GTs were thought to catalyze the

attachment of D-desosamine (29) and L-oleandrose (64) to oleandolide (65), but it was not known which GT catalyzed which step (Figure 1-16A). In order to address this question, complementation studies in Sa. erythraea, the producer of the macrolide antibiotic erythromycin A (34) which is structurally very similar to 36, were carried out. In previous work, disruption mutants of each of the two GTs-encoding genes involved in erythromycin biosynthesis, eryBV and eryCIII, had been made, and had been found to produce erythronolide B (66) and 3-O-mycarosyl erythronolide B (67), respectively. These results confirmed that EryBV is the mycarosyltransferase attaching the first sugar to 66 and EryCIII is the desosaminyltransferase attaching the second sugar to 67.78, 79 Expression constructs for OleG1 and OleG2 were made and each was transferred into the eryBV and eryCIII Sa. erythraea mutants, respectively. It was found that expression of OleG1 complemented the eryCIII mutant, restoring production of 34, and expression of OleG2 complemented the eryBV mutant, leading to formation of a doubly glycosylated erythromycin derivative (68) bearing L-rhamnose (69) rather than L-mycarose (70, Figure 1-16B). These results support the conclusion that OleG1 is the desosaminyltransferase in

28 O O A O O Me Me Me Me Me Me OH NMe2 OH HO Me O Me Me O O Me Me O OH 29 O O O OH Me 64 Me OMe O OH 65 Me 36 OH Me OTDP B O O O O Me Me HO Me Me Me Me Me OH OH 54 OH EryCIII or HO Me Me Me Me OH Me OH OleG1 Me OH NMe2 Me Me Me HO O OH O OH O O O Me Me EryBV Me Me O OH O O O O OH OMe Me Me Me 70 Me Me OH 66 67 O OH 34 O Me Me

O O O Me Me Me Me Me Me EryCIII or Me Me Me Me Me Me OH OleG1 OH NMe OH 2 OleG2 Me Me HO Me O OH O O O O OH Me Me Me Me O O O O OH O Me Me 69 Me OH OH Me O OTDP HO O OH O OH OH Me OH Me HO OH 68 27 C Me O OTDP O O MeO Me Me Me Me HO OH OH Me 71 Me OH Me OH Me Me Me O OH O OH Me OleG2 Me O OH O O Me Me OMe OH 66 72 O Me Figure 1-16 the biosynthesis of 36 that is responsible for attaching the second sugar, and OleG2 is the oleandrosyltransferase catalyzing the coupling of the first sugar to 65. The incorporation of the 69 moiety, rather than 70, into 68 may be ascribed to the incapability of OleG2 to accept TDP-L-mycarose (54). Apparently it prefers endogenous cellular TDP-L-

29 rhamnose (27) due to its similarity to TDP-L-oleandrose (71), which is the natural substrate for OleG2.80 The use of heterologous expression of a GT gene to determine the function of the corresponding GT is exemplified by the study of OleG2 from the biosynthetic pathway of 36. Data from the complementation experiments in Sa. erythraea described above supported the proposed function of this GT as the oleandrosyltransferase catalyzing the first glycosylation step in the biosynthesis of 36. In order to provide more definitive evidence about its function, oleG2 was integrated into the chromosome of a non- antibiotic producing strain Streptomyces albus under the control of a strong constitutive promoter, ermE. The genes thought to be responsible for formation of TDP-L-oleandrose (71) were then cloned into an expression vector and transferred to the S. albus mutant expressing OleG2. The resulting mutant strain was capable of converting fed erythronolide B (66), an aglycon structurally similar to 65 that is an intermediate in the biosynthesis of erythromycin A (34), to 3-O-oleandrosyl erythronolide B (72). These findings provided additional evidence that OleG2 is indeed the L-oleandrosyltransferase in the oleandomycin biosynthetic pathway (Figure 1-16C).81 Such complementation and heterologous expression strategies have also been used to study GT function in a number of other natural product systems in vivo. It is clear when assay of the purified proteins is difficult or impractical, in vivo methods are useful for studying gene function.

An exciting finding in the field of natural product GTs was recently reported.73 In this work, it was serendipitously discovered that natural product GTs can catalyze the reverse reactions under appropriate conditions. Normally, a GT uses an NDP-sugar and an acceptor molecule as substrates and produces glycosylated product and NDP. But in the reverse reaction, NDP and glycosylated natural product can be used as substrates to produce NDP-sugar and aglycon. This strategy was demonstrated in vitro using GTs

30 from the calicheamicin (73) and vancomycin (74) biosynthetic pathways (Figure 1-17). This finding has important implications for future determination of the functions of GTs, for the preparation of NDP-sugar and aglycon substrates from natural products, and for the alteration of natural product glycosylation patterns, which will be discussed in the next section. As far as using this strategy to elucidate GT function, one can envision

HO NH2 O O Me O MeS OH S HO OMe NH Me O OH S O O OH Me O Me Cl I O Me S O O NH O HO HO OH O OMe OH Cl H O OMe O N H O O Me Me N O O N NH HO H Me H N Me N O H NH OMe NH O HO OH MeO O O NH2 73 O OH Me HO OH Me 74 Figure 1-17

isolating a glycosylated natural product containing one or more sugars and cloning, overexpressing, and purifying the GT(s) predicted to be involved in formation of the compound. One could then assay each GT in the presence of the natural product and NDP. Observation of NDP-sugar and aglycon products in a given assay mixture would be evidence that a particular GT is responsible for a given glycosylation step.

In summary, GTs are ubiquitous enzymes involved in a myriad of glycosylation

processes in a variety of organisms. In spite of the wide variety of substrates they employ and the high degree of divergence of their amino acid sequences, GTs fall into two main structural classes, GT-A and GT-B. Natural product GTs are members of the GT-B superfamily. Structural studies of several glycopeptide antibiotic GTs have shed light on the features of substrate binding and the mechanism of catalysis of natural product GTs. These results hold promise for future enzyme engineering work aimed at

31 creating GTs with novel synthetic capabilities. Only recently have the activities of a handful of bacterial natural product GTs been demonstrated in vitro. This work was hampered by difficulties in obtaining the appropriate NDP-sugar and aglycon substrates and purified soluble enzymes. A number of in vivo strategies, such as gene disruption, bioconversion, complementation, and heterologous expression, have been developed for the elucidation of GT function. These methods are particularly useful in situations when obtaining the appropriate substrates and enzymes for in vitro assays is problematic.

4. NATURAL PRODUCT BIOENGINEERING USING GLYCOSYLTRANSFERASES AND SUGAR BIOSYNTHESIS ENZYMES

Natural product bioengineering, which is the use of biosynthetic genes or their encoded enzymes to construct non-natural metabolites in vivo or in vitro, is a new approach for the creation of natural product analogues. Natural products that possess bioactivity are often not suitable for clinical applications due to problems with solubility, bioavailability, toxicity, or ineffectiveness against resistant organisms, making synthesis of their analogues important in clinical drug discovery. Traditionally, natural product analogues are prepared by chemical synthesis or chemical derivatization of isolated natural products (semi-synthesis). But because the structures of many natural products

are quite complex, often having several chiral centers and many reactive functional groups, total synthesis is generally impractical, and even semi-synthesis can be very challenging. The bioengineering approach offers an alternative to chemical synthesis or semi-synthesis with distinct advantages: access to chemical derivative structures not easily obtainable by synthetic means, ease of compound production, reduced production of toxic byproducts, scalability of the process, and high potential for performing combinatorial biosynthesis as provided by the specifics of the biosynthetic route. 32 As mentioned in Section 1, the feasibility of natural product bioengineering requires detailed knowledge about the biosynthetic pathway(s) to be engineered. In the case of glycosylated natural products, a thorough understanding of sugar biosynthesis routes and the functions of GTs involved is essential for bioengineering work. Section 2 outlined what is currently known about deoxysugar biosynthesis in natural product pathways, and Section 3 dealt with the GT reactions and elucidation of functions of GTs using biochemical and genetic techniques. In this section, recent advances in using what is known about deoxysugar biosynthetic enzymes and GTs to perform natural product bioengineering will be discussed. Several of the cases described in Sections 2 and 3 will be revisited, as some of the studies are also examples of metabolic pathway engineering.

Natural product biosynthetic engineering can be conveniently divided into three main categories: precursor-directed biosynthesis, metabolic pathway engineering, and chemoenzymatic/enzymatic synthesis. Precursor directed biosynthesis involves replacement of a building block known to be used in the biosynthesis of a natural product with a structural analogue to effect a change in the structure of the final product. This is usually accomplished by adding the analogue to the growth media. A variation of this, called mutasynthesis, combines precursor directed biosynthesis with gene disruption to aid in non-natural building block incorporation. Metabolic pathway engineering, also referred to as combinatorial biosynthesis, involves the manipulation of the genes responsible for the biosynthesis of a given natural product in order to alter its structure. As was seen in Sections 2 and 3, genes can be inactivated and/or heterologously expressed in order to achieve the desired engineering objective. Chemoenzymatic and enzymatic syntheses involve the use of chemically or enzymatically prepared substrates, or libraries thereof, as substrates for a natural product biosynthetic enzyme. Normally the enzyme chosen is one which catalyzes a coupling reaction in a biosynthetic pathway,

33 such as a GT, because the combinatorial biosynthetic potential of such an enzyme is greater than that of a single substrate enzyme. Variations of this strategy involving enzymes that have been engineered to broaden substrate specificity have been demonstrated. All of these strategies rely on the relaxed substrate specificities of one or more biosynthetic enzymes in the engineered pathway. Each of these strategies has inherent advantages and disadvantages which make it more or less suitable to a specific engineering task. Of the three methods mentioned, precursor-directed biosynthesis is both the best- established and the simplest. It has been used effectively to generate structural variations of a variety of classes of natural products. Several cases have been reported in which precursor-directed biosynthesis was used to make non-natural glycosylated natural products. These experiments involved feeding chemically modified aglycons or aglycons of structurally related compounds to a producing strain to change the glycosylation patterns. Successful examples in aminoglycoside, macrolide, anthracycline, and

glycopeptide producing organisms have been documented.82 The advantage of this method is that no genetic engineering or enzyme purification is needed, and no detailed information about the steps in the biosynthetic pathway is required. However, a suitable substrate must be generated synthetically or semi-synthetically. There are also several disadvantages to this method. First, the compound must be fed to the cell, limiting the

scale of the process to the molar amount of fed precursor. Also, the compound needs to be uptaken by the cell, although cell free preparations have been used. Third, one is limited to using the enzymes present in the wild-type cell. Nevertheless, precursor- directed biosynthesis is a powerful method for making natural product analogues. A variation on precursor-directed biosynthesis, called mutational biosynthesis, or “mutasynthesis,” is also very useful. In this method, a gene involved in a step whose

34 product is being replaced by an analogue is disrupted. The non-natural building block, or mutasynthon, is then fed to the mutant to replace the missing intermediate, resulting in formation of a new analogue. The advantage of this method over precursor-directed biosynthesis is that the mutasynthon, which is often not a very competent substrate, need not outcompete the natural building block. This method has been used effectively to change the structures of several different classes of natural products, including polyketides, non-ribosomal peptides, aminoglycosides, and others.83 However, some of the same drawbacks of precursor-directed biosynthesis, such as scalability and uptake, remain as issues in the use of mutasynthesis.

Metabolic pathway engineering, or the genetic manipulation of a biosynthetic pathway to change the structure of the final product, is a versatile method which has been enjoying rapid development in the natural product bioengineering field in recent years. The first report of the use of this technology to make natural product analogues was in 1985, when subcloned DNA fragments from the actinorhodin (75) gene cluster of Streptomyces coelicolor were expressed in the strains producing the structurally similar compounds medermycin (76) and granaticin (77). These experiments resulted in the production of three novel compounds, mederrhodin A (78), dihydrogranaticin (79), and dihydrogranatirhodin (80, Figure 1-18) which had hybrid structures and were presumably formed by heterologously expressed components of the actinorhodin biosynthetic machinery working with those of medermycin or granaticin pathways.84 Although the functions of the individual genes in these pathways were not known at the time, this pioneering work demonstrated the potential of genetic engineering of biosynthetic pathways to generate novel natural products. Cluster sequencing work and gene functional prediction allowed targeted disruption of genes, as exemplified by the 1991 report in which a P450 gene in the erythromycin A (34) biosynthetic cluster was

35 Me O OH OH O Me Me OH O Me HO O O O O O Me2N O HO OH O OH OH O OH O O 75 O 78 Me OH O Me HO O OH OH O Me Me2N O Me O O OH O O O OH O 76 O 79 OH OH O Me Me OH OH O Me O Me OH O O O OH O O O OH O 77 O 80 Figure 1-18

disrupted, resulting in formation of the novel compound 6-deoxy erythromycin A (81, Figure 1-19), which has enhanced acid stability.85 Currently, more than one hundred natural product gene clusters have been sequenced, and biosynthetic studies of prototypical members of many classes of compounds have been done, making increasingly sophisticated natural product pathway engineering possible.

O O Me Me Me Me HO OH HO Me Me Me OH NMe2 Me OH 6 NMe2 Me HO Me HO O O O Me O O O Me Me Me O O OMe O O OMe Me Me Me Me O OH O OH Me Me 34 81 Figure 1-19

Engineering of the sugar structures and glycosylation patterns of glycosylated natural products began in the mid-1990’s. Simple cases of GT gene disruption leading to

36 erythromycin analogues lacking one or more sugars, which was briefly discussed in Section 3, are some of the earliest examples of altering natural product glycosylation patterns by bioengineering.78, 79 However, these experiments were originally performed to elucidate GT function rather than to generate structurally modified compounds. Another early example is one in which GtfE’, a glucosyltransferase from the vancomycin (74) biosynthetic pathway of the Actinomycete Amycolatopsis orientalis, was expressed in Streptomyces toyocaensis, the producer of a non-glycosylated non-ribosomal peptide antibiotic A47934 (82), which is structurally similar to 74. This resulted in regiospecific glucosylation of A47934, forming the novel compound glucosyl-A47934 (83, Figure 1-

20).86 The success of this work depended upon the relaxed substrate specificity of GtfE’ which allowed it to glycosylate the A47934 aglycon

HO NH2

Me O OH Me O OH GtfE' O O Cl OH

HO OH Cl H O N H O O N Me N NH O H N NH NH H O HO O O NH2 O OH Me HO OH Me

74 OH O OH O OH Cl O Cl OH

HO OH HO OH Cl Cl H O O H O O N H GtfE' N H O N H O N H N N N N H NH H NH NH O O NH O O HO O HO O HO NH2 HO NH2 O Cl O O Cl O OH OH HO OH - HO OH - O3SO 83 O3SO 82 Figure 1-20

37 A more systematic study of GT substrate flexibility leading to natural products with altered sugar structures was carried out using the desosamine (29) biosynthetic pathway involved in methymycin/pikromycin (30/32) formation in S. venezuelae, which was discussed extensively in Section 2 of this chapter. In these studies, each of the four desosamine biosynthesis genes, desI, desII, desV, and desVI, was disrupted. In each mutant, analogues of 30 with altered sugar structure (42-47, 49) were generated (See Section 2, Figure 1-9). Subsequent engineering work in which an expression plasmid containing the genes strM and strL, proposed to catalyze the conversion of 14 to 28 and

28 to a mixture of TDP-L-rhamnose (27) and TDP-L-dihydrostreptose (84) in the streptomycin (85) pathway of Streptomyces griseus, was transferred into the KdesI mutant, resulted in formation of four new methymycin/narbomycin analogues (86-89)

bearing the sugar L-rhamnose (69, Figure 1-21).87 As described previously, these studies, together with in vitro work on desosamine pathway enzymes, were instrumental in elucidating the desosamine biosynthetic pathway. But more importantly, these studies were successful in generating 11 novel glycosylated analogues of 30/32 (42-47, 49, 86- 89), clearly revealing that the GT DesVII is highly substrate flexible, tolerating TDP- sugar substrates 27, 90-93 (Figure 1-21).37, 39-41 The substrate flexibility of DesVII suggests that this enzyme can be used for combinatorial glycosylation both in vivo and in vitro.

Another interesting early example of pathway engineering to produce novel glycosylated natural products is the engineering of the daunorubicin (94) biosynthetic pathway of Streptomyces peucetius. Epirubicin (95), an analogue of 94 in which the

stereochemistry of the C-4 hydroxyl group of the sugar L-daunosamine (96) is reversed, is an important semi-synthetic antitumor drug with improved therapeutic properties compared to 94 (Figure 1-22). Compound 95 is produced by a multi-step chemical route

38 O Me O DesI HO OH OTDP 14 Me StrM HO O HO HO OH O StrL Me O OTDP O 90 OTDP Me Me OH OTDP R1 Me OH OH O HO Me O OH Me OH Me 84 28 O OH AcHN O R 69 2 O O HO StrL OH Me Me 91 OTDP 86 R = H, R = H Me DesVII 1 2 H2N O OTDP 87 R1 = OH, R2 = H HO 88 R = H, R = OH Me OH 1 2 O N H NH HO HN OH O H N HO 27 OH O NH Me OH OHC 2 Me 92 OTDP O OH Me O OH Me OH Me OH OH O Me 69 Me O O Me O Me O O O AcHN HO NHMe OH Me HO 89 93 OTDP 85 Figure 1-21

from 94. Although the biosynthetic route to the supposed activated sugar intermediate,

TDP-L-daunosamine (97), was not known, the gene cluster encoding production of 94 had been isolated, and several genes were shown through gene disruption to be involved in the formation or attachment of 96. One of these genes, dnrV, was proposed based on sequence analysis to encode the sugar 4-ketoreductase in the biosynthesis of 97. This gene was inactivated and replaced with avrB and eryBIV from the avermectin and erythromycin (34) biosynthetic pathways, respectively. Each of these was predicted to encode an enzyme that catalyzes sugar 4-ketoreduction with the opposite stereochemistry as DnrV based on the sugar structures in avermectin and 34. Impressively, both replacement mutants made significant quantities of 95.88 This example is interesting for

several reasons. First, the sugar generated, 4-epi-L-daunosamine (98), is a sugar that does not exist in nature, illustrating the potential of pathway engineering to construct non- natural molecules using the proper combination of natural biosynthetic enzymes. This remains the only report to date of the creation of a glycosylated natural product bearing a 39 O O OH O OH O Me Me OH OH

MeO OH O O MeO O OH O Me O 96 Me O 98 HO NH2 HO NH2 94 95

DnrS DnrS

Me Me O OTDP O OTDP HO NH NH HO 2 2 97 DnrV AvrB or EryBIV wild-type KdnrV S. peucetius S. peucetius

O Me O Me O OTDP HO OH O NH OTDP 2 14 Figure 1-22

sugar that does not exist in nature. Second, it is amazing that in spite of the fact that the pathway for the formation of 97 was not known, the correct open reading frame (ORF) was chosen for disruption. In fact, the biosynthetic pathway proposed for 97 in the literature (not shown) is likely incorrect based on biochemical studies of homologues in the well-studied sugar pathways discussed in Section 2.

During the period of this dissertation work, natural product pathway engineering work has become increasingly sophisticated. Several interesting developments have been reported during this time. First, the first and only example to date of creation of an engineered natural product GT which is able to catalyze formation of a novel compound was reported.89, 90 Two GTs involved in urdamycin A (99) biosynthesis in Streptomyces fradiae Tu2717, UrdGT1b and UrdGT1c, were chosen as the starting points for

40 engineering efforts. These GTs are 91% identical in sequence to each other, yet show different specificities with respect to both the nucleotide sugar and the acceptor substrate.

UrdGT1c attaches L-rhodinose (100) to the 3-position of the D-olivose (101) moiety of aquayamycin (102) to form 12b-derhodinosyl urdamycin G (103), while UrdGT1b

attaches a D-olivose (101) moiety to the 4-position of the L-rhodinosyl moiety of 103 to form derhodinosyl urdamycin A (104, Figure 1-23). The fact that these two enzymes share such a high degree of sequence identity yet catalyze different reactions presents a

Me Me O O O O RO OH RO OH Me Me O UrdGT1c O HO OH HO OH HO O 101 OH O OH O Me O 102 R = H 103 R = H OH 100 engineered UrdGT1b/1c UrdGT1b variants Me O Me O O O RO OH RO OH Me Me Me O O O OH HO O OH HO OH Me HO O O O 101 OH O OH O Me O Me O Me 105 O A OH HO O HO 99 R = A 104 R = H 101

Figure 1-23

unique engineering opportunity. First, a series of chimeric enzymes were made by exchanging portions of UrdGT1b and UrdGT1c sequences, and were transferred to appropriate S. fradiae mutant hosts to identify specificity-conferring regions of the enzymes. These experiments resulted in a high proportion of functional chimeras which displayed specificity of one or the other parental enzyme, and showed that a particular 30 amino acid region of these enzymes was responsible for conferring specificity.90 This 30 amino acid region contains 18 non-identical amino acids. Mutagenesis performed on sets of these 18 amino acids revealed that 8 of these were not important for specificity. A library of 2048 mutant genes with randomized combinations of parental sequences at 41 these 10 positions was created, and partially screened using TLC and HPLC methods. Enzymes having a single parental activity, both parental activities, exclusively novel activity, or combinations of novel and parental activities were observed. The novel

activity observed was D-olivosylation of the C-4 hydroxyl of the D-olivose moiety of 103, forming the novel branched-chain derivative urdamycin P (105). This remarkable demonstration of manipulating GT function to create a novel natural product derivative illustrates the potential to create new GT activity through engineering. However, this example is a special case. The occurrence of two enzymes with 91% sequence identity that catalyze different reactions is rare, and greatly simplifies the engineering problem, making the requisite library size manageable for TLC-based screening methods. Most natural product GTs that act on acceptor substrate from the same structural class but have different specificities are typically about 50% identical in sequence, making thorough library coverage impossible even with a high-throughput GT activity screen. Recent developments in the area of high-throughput screens for GT activity 91, 92 may soon lead to other successful examples of engineered GTs with non-natural activities and specificities like those created in this work. Another example is that of the establishment of the substrate flexible rhamnosyltransferase ElmGT from the elloramycin (106) producer Streptomyces olivaceus and exploration of the breadth of its sugar substrate flexibility using various

combinations of sugar biosynthetic genes. Initial work done by expressing a cosmid, 16F4, containing a 25 kb portion of the biosynthetic gene cluster for 106, which was thought to contain most of the 106 biosynthetic genes, in the urdamycin A (99) producer S. fradiae Tu2717 and the mithramycin producer Streptomyces argillaceus resulted in the

formation of four novel elloramycin derivatives (Figure 1-24) bearing D-olivose (101), L- rhodinose (100), D-mycarose (107), and a 1,3-linked disaccharide of 101.93, 94 Control

42 OMe O Me OH O O Me Me O Me HO O MeO Me O HO O HO O HO HO O OMe 101 OH HO 100 101 101 O OH Me O MeO Me OH MeO O Me Me Me OMe 106 HO O O O Me HO HO HO OH Me OH OH OH R = 107 70 108 OH 69 O Me OH O O OH MeO Me Me HO O Me O O HO Me O HO RO OMe HO HO OH OH HO O OH 109 110 111 112 OH Me O HO 113

Figure 1-24

experiments confirmed that the GT activity was present on 16F4, as were all genes needed to form the elloramycin aglycon. Clearly the endogenous TDP-sugars produced by the expression hosts were used by the elloramycin cluster GT for glycosylation of the elloramycin aglycon. Subsequent work identified the gene encoding the GT, which was named ElmGT.95 After these initial successes, 16F4 was expressed in the non-producing strain Streptomyces albus. A flexible cloning strategy was also employed to create a series of plasmids expressing different combinations of sugar biosynthetic genes, which were co-transferred to S. albus with 16F4. These experiments resulted in the production of previously observed elloramycin derivatives bearing 100 and 101 as well as novel derivatives bearing L-rhamnose (69), L-mycarose (70), L-olivose (108), D-glucose (109),

L-amicetose (110), D-amicetose (111), 4-deacetyl-L-chromose B (112), and L-digitoxose

(113).96-99 Especially interesting is the fact that several of the constructs which generate the TDP-sugar precursors used for glycosylation of these engineered compounds are made using genes from several different sugar biosynthetic pathways. Thus, in addition to making new compounds, these results provide valuable information about the

43 functions and substrate specificities of the sugar biosynthetic enzymes expressed from these constructs, and illustrates the feasibility of constructing hybrid sugar biosynthetic pathways. Since the biosynthetic pathways for 110 and 111 have not yet been identified, some of the enzymes assembled to synthesize these sugars must be substrate flexible in order for the pathway to synthesize the desired sugar product. These experiments illustrate the potential for modular construction of engineered sugar biosynthetic pathways such that genes needed to construct target sugars are systematically assembled in a series of plasmids. This strategy could be employed to construct small libraries of compounds, as was done here.

Modern biosynthetic engineering strategies have been applied to natural product systems other than sugar biosynthesis, such as polyketide biosynthesis, non-ribosomal peptide biosynthesis, indolocarbazole biosynthesis, and others. Additionally, there are other examples of sugar pathway engineering which were not covered in this Section. The cases presented are intended to provide a broad survey of the progress made in the field and the strategies used in pathway engineering. Chemoenzymatic and enzymatic syntheses of natural product analogues have recently become practical due to the availability of a number of biochemically characterized natural product biosynthetic enzymes, including sugar biosynthetic enzymes, GTs, and others, as well as the availability of the appropriate substrates and

substrate analogues. As far as enzymatic synthesis of natural product analogues is concerned, the majority of work has focused on the use of GTs because, as was mentioned earlier, the sugar moieties of natural products are usually critical to bioactivity. But importantly, GTs and other coupling enzymes are excellent targets for in vitro combinatorial syntheses because they use two substrates which can be separately altered, dramatically increasing potential library size and structural diversity. The

44 advantages of in vitro glycosylation over other methods is that if a purified GT and a collection of substrate analogues are available, screening substrates for coupling by a GT is straightforward. This can potentially yield libraries of compounds provided that the GT is reasonably tolerant of substrate analogues. Also, substrate concentrations can be manipulated such that unfavorable reaction processes can be driven to completion. In the assay process, information about specific structural features of substrate analogues can also be obtained, guiding future use of a given GT in synthetic applications. Disadvantages with this method arise mainly from the difficulty in preparing appropriate substrates and enzymes. Therefore, efforts in this field of research have focused on finding creative ways of generating substrate libraries.

As described in Section 3, only a handful of bacterial natural product GTs have been demonstrated active in vitro. The first demonstration of activity, as well as the first demonstration of natural product GT substrate flexibility came in 2001 with studies of the vancomycin GTs GtfB, GtfC, GtfD, and GtfE. The ability of the glucosyltransferase

GtfE to transfer D-glucose (114) to the acceptor substrate analogue, teicoplanin aglycon (115), and that of the epivancosaminyltransferase GtfD to glycosylate the substrate analogue 116 were tested. In both cases, these GTs were able to recognize and process the acceptor analogue (Figure 1-25A).67 Subsequent studies explored the substrate flexibilities of GtfD and GtfE in depth using a technique called in vitro

glycorandomization (IVG) for the preparation of a library of TDP-sugars. In IVG, a small library of α-sugar-1-phosphates are made chemically and thymidylylated using a substrate-flexible nucleotidylyltransferase, resulting in a library of TDP-sugars to be used as GT donor substrates. Since the TDP coupling step is notoriously unreliable, often resulting in low yields, IVG is advantageous for simplifying the preparation of TDP- sugars. A total of 34 natural and unnatural TDP-sugars were tested as substrates for the

45 HO NH2

Me O A 114 OH OH Me HO OH O OH O O OH Cl O Cl OH O Cl OH

HO OH HO OH HO OH Cl Cl Cl H O O H O O H O O N H N H N H H O H O H O N N N N N GtfD N N N NH GtfE N NH NH H H O H NH O O NH O O NH O HO O HO O HO O HO NH2 HO NH2 HO NH2 O O O O O O OH OH OH HO OH HO OH HO OH 115 HO 116 HO HO HO NH2 117 O B Me OH Me O R O R O OH Cl O Cl O Cl OH IVG library HO OH HO OH Cl HO OH Cl Cl H O O H O N H N H O H O O O N Me O N Me N H N N O N Me H N NH NH GtfD N O NH H N NH NH NH H O H O H N NH O NH O NH H HO GtfE HO O O O HO NH O O 2 O NH2 NH O O O 2 OH O Me OH Me OH HO OH Me OH Me HO Me HO OH Me 118 R = H 119 R = NH Figure 1-25 2

glucosyltransferase GtfE, of which 31 showed at least 25% conversion to glycosylated product under the conditions used, indicating that GtfD is highly flexible toward the sugar substrate (Figure 1-25B). Furthermore, GtfE was tested for its ability to catalyze

transfer of L-vancosamine (117) to two of the non-natural products of the GtfE reaction, and was found to accept both, forming vancosaminylated derivatives 118 and 119.100, 101 These experiments not only illustrate the flexibility of these GTs, particularly GtfD, but also demonstrate the potential of IVG to simplify the synthesis of TDP-sugars. Other work focusing on the chemoenzymatic preparation of vancomycin analogues with altered sugar structures has also been performed using the GTs GtfA,

GtfC, and GtfD. GtfA transfers a 4-epi-L-vancosamine (120) moiety onto the vancomycin pseudoaglycon (121) to form chloroorienticin B (122), which is epivancosaminylated at a different site by GtfC to form chloroeremomycin (123). GtfC can also act on 121 directly. GtfD transfers L-vancosamine (117) to 121, forming vancomycin (74, Figure 1-26). A systematic investigation of the abilities of these three

46 OH OH HO OH HO OH O O NH2 O OH O OH Cl HO Cl Me O HO OH 120 Me O OH Cl Cl H O H O N H O N H O O N Me GtfA O N Me N NH N NH O H N NH O H N NH NH H O NH H O HO HO O O O NH2 O NH2 O O OH Me OH Me HO OH 121 Me HO OH 122 Me

GtfD GtfC

HO NH2 NH2 HO O Me O Me OH OH Me 117 Me O OH 120 O OH O O NH2 OH O OH O Cl HO Cl Me O Me HO OH O OH Cl Cl H O H O O N H O N H O N Me O N Me N N NH H N NH H N O NH O H NH NH H O NH O HO HO O O NH O NH2 O 2 O O OH Me OH Me HO OH 74 Me HO OH 123 Me Figure 1-26

GTs to transfer sugars from a ten member chemically synthesized TDP-sugar library onto the various acceptor substrates in vitro was undertaken. These studies revealed that each of the GTs has a unique degree of substrate flexibility, with GtfA being the least flexible and GtfC and GtfD being more flexible.102 Even GtfC and GtfD, however, were unable to accept all ten TDP-sugars. These studies provide a detailed survey of the substrate promiscuity of these three enzymes which will guide future use of these enzymes in synthetic applications. Other in vitro natural product GT substrate flexibility tests have been performed using the GT NovM from the novobiocin pathway,103 EryCIII from the erythromycin pathway,69 and CalG1 from the calicheamicin pathway.73 In summary, precursor directed biosynthesis, metabolic pathway engineering, and chemoenzymatic/enzymatic synthesis are all effective means of altering the sugar structures of bioactive natural products. Each method has advantages and disadvantages, which have been described. The specific properties of the biological system(s) from which components are being used will influence the choice of the best method to 47 accomplish a given engineering goal. All methods rely on substrate promiscuity on the part of pathway enzymes. Generally, precursor-directed biosynthesis required less information about the details of the pathway to be utilized, while metabolic pathway engineering and chemoenzymatic/enzymatic synthesis require detailed information about the roles of specific enzymes in the pathway being engineered. Pathway engineering has the advantages of requiring no enzyme or substrate purification, and of the resulting analogue-producing mutants are a renewable and scalable source of compounds. Chemoenzymatic/enzymatic syntheses are advantageous because interpretation of results is straightforward, and reaction conditions can be manipulated to drive unfavorable reactions to form the product.

5. SUMMARY AND THESIS STATEMENT

Since natural products are the primary source of bioactive lead compounds used in drug development, advances in natural product biosynthesis and derivatization are ultimately important for the treatment of human diseases. The past twenty years have seen dramatic changes in the natural products field. The advent of modern molecular biological techniques has revolutionized the way natural product biosynthesis is studied as well as how natural product derivatization is performed. The identification and sequencing of over one hundred natural product biosynthetic gene clusters and the elucidation of key steps in the biosynthesis of many natural products have made it possible to use the natural product biosynthetic machinery in vivo or in vitro for the creation of natural product analogues. The diverse sugar moieties found in natural product structures are often critical to their bioactivity. Because of this, both the study of sugar biosynthesis and glycosylation and the use of sugar biosynthetic and GT genes and enzymes in natural product 48 derivatization have become important research areas. The work described in this dissertation focuses on the functional elucidation of enzymes involved in biosynthesis

and glycosyltransfer of the deoxysugar D-mycaminose, which is a structural component of the macrolide antibiotic tylosin, and on the use of genes encoding the biosynthesis and attachment of D-mycaminose, D-desosamine, and other deoxysugars for the engineered production of macrolide derivatives with altered sugar structures. All in vivo work was performed in the model Actinomycete host S. venezuelae. In vivo functional elucidation of TylM3 as a GT activator protein for the GT TylM2, and of its homologue MydC as the corresponding activator protein for the GT MycB from the mycinamicin (33) in the host S. venezuelae described in Chapter 2 has led to a better understanding of the activities of macrolide GTs and has facilitated use of these GTs to catalyze formation of several novel glycosylated macrolide derivatives. Discovery of a hexose 3,4-ketoisomerase, Tyl1a, which is involved in formation of TDP-

D-mycaminose (52), allowed successful reconstitution of the 49 biosynthesis and attachment pathway in an engineered S. venezuelae host. In vivo and in vitro characterization of Tyl1a are described in Chapter 3 and 4, respectively. The design and construction of several S. venezuelae mutants expressing engineered deoxysugar biosynthetic pathways, one of which resulted in the formation of non-natural deoxysugar- bearing macrolides, is described in Chapter 5. This work has contributed important

information on the activities and specificities of the sugar biosynthesis and GT enzymes studied, and has illustrated the feasibility of constructing complex engineered deoxysugar biosynthesis pathways in the macrolide producer S. venezuelae.

49 Chapter 2. Functional Analysis of tylM2/tylM3 and mycB/mydC Pairs Required for Efficient Glycosyltransfer in Macrolide Antibiotic Biosynthesis

1. INTRODUCTION

As described in Chapter 1, the presence of deoxy- and aminosugars in the structures of many secondary metabolites, including many antibiotics, is often essential for their biological activities.5-7 It is well known that alteration of the sugar appendage(s) of these compounds can have a profound impact on the spectrum and/or potency of their biological activities. Since the sugar components are attached to the parent aglycons by the action of glycosyltransferases (GTs), understanding the factors controlling the activity and specificity of these glycosyltransferases is critical for the generation of new glycosylated natural products having improved or novel biological properties with potential clinical applications. Recently, studies of GTs involved in the biosynthesis of secondary metabolites have attracted much attention. In particular, in vivo gene disruption and heterologous gene expression have been successfully used to probe glycosyltransferase specificity in bacterial strains which produce glycosylated antibiotics. As a result, a number of "hybrid" glycoconjugates have been generated by genetic engineering of the sugar

biosynthetic machinery in the producing strains.104, 105 Progress has also been made in exploiting the relatively relaxed specificity of these glycosyltransferases toward sugar substrates in vitro to derivatize specific aglycons with diverse glycoforms.10 In the biosynthetic pathway for the macrolide antibiotics methymycin (30), neomethymycin (37), and pikromycin (32) in the Actinomycete Streptomyces venezuelae, the desVII gene was proposed based on sequence homology to encode a GT which

50 catalyzes the transfer of the D-desosamine (29) moiety from TDP-D-desosamine (41) to either the 12-membered ring macrolactone 10-deoxymethynolide (124) or the 14- membered macrolactone ring narbonolide (125), producing YC-17 (126) and narbomycin (31) respectively. 126 and 31 are then hydroxylated by the P-450 hydroxylase PikC,106 resulting in formation of the final antibiotic structures methymycin (30), neomethymycin

OH OH Me O Me O O O DesIII O H2N DesI DesIVHO HO HO HO HO HO HO HO HO HO OTDP OTDP OTDP OPO = 48 14 13 12 3

DesII O O O Me Me Me Me R1 O DesVII Me Me Me Me NMe Me 2 Me NMe2 O DesVIII O O HO PikC HO HO O R2 O OTDP O OH O O Me O O O Me 39 Me Me Me Me Me Me 124 126 30 R1 = OH, R2 = H DesV 37 R1 = H, R2 = OH Me Me O O DesVI H2N Me2N HO HO OTDP OTDP 40 41 O O O Me Me Me HO Me Me Me Me NMe Me Me 2 PikC NMe2 Me Me HO HO O Me O OH DesVII O O Me O O O Me Me Me Me O O DesVIII O O O O Me Me Me 125 31 32

Figure 2-1

(37) and pikromycin (32) (Figure 2-1). Repeated attempts to reconstitute DesVII activity in vitro using purified enzyme and substrates 124 and 41 failed. The biosynthetic gene cluster for 30/37/32 was examined in order to identify other genes which might be involved in glycosylation in the pathway. One gene, desVIII, stood out as a potential candidate for this role. DesVIII and its homologues, which are conserved in a number of glycosylated antibiotic gene clusters, were originally proposed to be hexose ketoisomerases that convert TDP-4-keto-6-deoxy-

D-glucose (14) to TDP-3-keto-6-deoxy-D-glucose (38, Figure 2-2). However, studies of

D-desosamine biosynthesis indicate that conversion of 14 to 38 is not required in the 51 pathway, and enzymes catalyzing each step O Me Me O DesVIII ? HO O in the pathway have been identified and HO OH O OH OTDP OTDP functionally characterized.37-42 Disruption of 14 38 O desVIII in S. venezuelae led to accumulation Me R1 Me Me of non-glycosylated macrolactones 124, KdesVIII O mutant R2 methynolide (127), and neomethynolide O OH Me Me

(128), indicating that DesVIII is important 124 R1 = H, R2 = H 127 R1 = OH, R2 = H D for either formation or attachment of - 128 R1 = H, R2 = OH 107 desosamine (29). However, it is unclear Figure 2-2 what the exact role of DesVIII and its homologues is in glycosyltransfer. In order to test whether DesVIII is involved in macrolide glycosylation, desVIII was cloned, and the DesVIII protein was overexpressed and purified. Inclusion of DesVIII in DesVII assay mixture with 124 and 41 remarkably led to efficient formation of a new product which was shown to be 126, indicating that DesVIII was important for the glycosyltransfer reaction.107 There is another example of a GT, the eukaryotic GT β- 1,4-galactosyltransferase, that requires a regulatory protein for full activity.108 However, DesVIII does not show homology to this regulatory protein, suggesting that it acts by a different mechanism. At the time of this report, this was the only example of a natural product GT which requires an auxiliary protein for activity. Concurrent with the in vitro studies of DesVII, heterologous expression of the biosynthetic pathway for TDP-D-mycaminose (52) from the tylosin (51) gene cluster of Streptomyces fradiae in S. venezuelae was also carried out to probe the substrate specificity of DesVII and other macrolide GTs in vivo. As described in Chapter 1, the biosynthesis of 53 and its use in glycosylation of tylactone (50) is proposed to occur as

52 O Me

Me Me

Me Me O O O OH Me Me Me Me CHO O O OH HO Me NMe Me NMe Me 2 O 2 HO OH HO O 50 TylM2 OMe OH Me Me O Me OMe Me O Me O O O O 49 Me OH Me O OH O OH O O Me HO Me Me Me2N 58 51 OH 52 OTDP

TylM1 OH OH Me Me O Me O O TylM3 TylA1 O HO TylB HO O TylA2 HO O HO H2N HO HO HO OH O OH OH OTDP OTDP OH OH = 53 38 14 OTDP 13 OTDP 12 OPO3

Figure 2-3

shown in Figure 2-3. The functions of the aminotransferase TylB and the methyltransferase TylM1 have been verified in vitro using chemically synthesized substrates.44, 46 The functions of TylA1, TylA2, and TylM2 as glucose-1-phosphate thymidylyltransferase, TDP-glucose 4,6-dehydratase, and glycosyltransferase were presumed based on high sequence identity to characterized homologues in other pathways. Disruption of tylM2 resulted in accumulation of a small amount of 50, yet the tylM2 disruption mutant was capable of converting O-mycaminosyl tylactone (58) to 51, indicating that TylM2 is likely the mycaminosyltransferase catalyzing the attachment of 52 to 50 to form 58. Originally, the interconversion of 14 and 38 was thought to be spontaneous, but studies of the reverse TylB reaction failed to yield any 14, suggesting that this reaction was enzyme-catalyzed. This enzymatic step was tentatively assigned to TylM3 based on gene disruption studies which implicated TylM3 in the formation or attachment of D-mycaminose (49)109 and on the homology of TylM3 to DesVIII, which was originally assigned a ketoisomerase function as well. However, after the observation that the activity of DesVII was dependent on the presence of DesVIII, the proposed function of TylM3 was reexamined. TylM3 and DesVIII share 33% sequence identity and 45% sequence similarity, suggesting that 53 TylM3 may serve an analogous function in the tylosin pathway as DesVIII does in the methymycin pathway. Furthermore, a BLAST search of DesVIII and TylM3 revealed the presence of a number of homologous protein sequences encoded in various other macrolide and anthracycline antibiotic gene clusters. The genes encoding two of these homologues, eryCII and dnrQ, were shown by gene disruption to be critical to formation

or attachment of deoxysugars in their respective pathways.78, 110 Another homologue of DesVIII and TylM3 is MydC from the mycinamicin (33) pathway of the Actinomycete Micromonospora griseorubida. The gene cluster encoding production of 33 had recently

been sequenced,111 yet no experiments on the function of MydC had yet been performed. MydC is 45% identical, 56% similar to DesVIII. Interestingly, desVIII, tylM3, and mydC, as well as most of their homologues, lie immediately upstream of GT-encoding genes, suggesting a functional link between them and GTs. The desVIII gene is upstream of desVII, tylM3 lies upstream of tylM2, and mydC lies upstream of mycB, which was

predicted based on sequence homology to encode the GT catalyzing the attachment of D- desosamine (29) to protomycinolide IV (129) in the biosynthesis of 33 (Figure 2-4). Thus, it would be interesting to test whether TylM3 and MydC serve analogous roles to that of DesVIII for their respective GTs, TylM2 and MycB. Information obtained from these studies would help to establish the general function of proteins in this class as well as to determine the factors which are important for the activities of macrolide GTs. Me Me Me

O O NMe2 O NMe2 HO HO OH O O Me O O Me Me Me Me MycB 29 O Me Me Me Me O O O HO HO O Me O O O OMe O OMe Me Me Me 129 33 Figure 2-4

We sought to test whether TylM3 and MydC are important for the activity of their proposed partner GTs. This would be accomplished by heterologous expression of each 54 of the GTs TylM2 and MycB, as well as the enzymes needed to form the appropriate sugar substrate for each GT, with or without the proposed auxiliary protein, in an engineered S. venezuelae strain lacking the endogenous glycosyltransferase DesVII. Pairs of mutant strains differing only in the presence or absence of the auxiliary gene can then be grown in parallel, and their metabolites extracted and subjected to analysis by HPLC. Under such conditions, the effect of the auxiliary protein on the activity of the corresponding GT can be determined. The results of this analysis indicate that the activities of the GTs TylM2 and

MycB are significantly enhanced by coexpression of the appropriate DesVIII homologues, TylM3 and MydC respectively, from each pathway. It was also found that TylM2 and MycB show relaxed substrate specificity, as their substitution for DesVII in various S. venezuelae mutants led to the isolation of several new glycosylated macrolides. These results have significant implications for future strategies of glycodiversification through a combinatorial biosynthetic approach.

2. EXPERIMENTAL PROCEDURES

General. Enzymes and molecular weight standard used for molecular cloning experiments were products of Invitrogen (Carlsbad, CA) or New England Biolabs (Ipswich, MA), except for pfu DNA polymerase, which was purchased from Stratagene (La Jolla, CA). The digoxigenin (DIG)-labeling and probing kit for the Southern Blot assays was purchased from Boehringer Mannheim (Mannheim, Germany). Antibiotics and chemicals were purchased from Sigma-Aldrich Chemicals, Co. (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA). Oligonucleotide primers for cloning of Tyl1a were prepared by either Integrated DNA Technologies (Coralville, IA) or Invitrogen (Carlsbad, CA). Growth media components were obtained from Becton Dickinson (Sparks, MD). 55 Silica gel used in chromatography was obtained from Sorbent Technologies (Atlanta, GA). Analytical thin-layer chromatography (TLC) was carried out using 0.25 mm

Polygram Sil G/UV254 plates obtained from Macherey-Nagel (Easton, PA). Plasmids and Vectors. Cosmid pLZ4,112 containing the desosamine biosynthetic cluster, was used as the template for amplification by the polymerase chain reaction (PCR) of all genes and gene fragments from S. venezuelae. Cosmid pHJL309 was used as the template for PCR amplification of tylM3, tylM2, and tylM1, while cosmid pSET552 was used as the template for PCR amplification of tyl1a and tylB. Both of these cosmids were generously provided by Dr. Eugene Seno of Eli Lilly Research Laboratories. Cosmid pMG160, kindly provided by Dr. Fumio Kato at Toho University, Japan, was used as the template for PCR amplification of mydC and mycB. The Streptomyces expression plasmid pAX617, derived from pOJ446, as well as pIJ702, the template used for PCR amplification of the thiostrepton resistance gene, were acquired from Dr. David Sherman at the University of Minnesota (now at the University of Michigan). The gene disruption plasmid pKC1139 was a gift from Dr. Leonard Katz of the Abbott Laboratories (now at KOSAN Biosciences, Hayward, CA). Methods and protocols for recombinant DNA manipulation were obtained from Sambrook et al.,113 and those involving the use of Streptomyces strains were obtained from Kieser et. al.114 Bacterial Strains. E. coli DH5α from Invitrogen (Carlsbad, CA) was used as the host for routine cloning. E. coli S17-1 was the donor strain used in S. venezuelae

conjugal transfer experiments.115 S. venezuelae ATCC 15439 was obtained from the American Type Culture Collection (Rockville, MD) as a freeze-dried sample, which was reconstituted according to the instructions from ATCC. Instrumentation. pH values were determined using a Corning pH meter 240 from Fisher Scientific. Agarose gel electrophoresis was performed using a mini-sub cell GT

56 apparatus from Biorad (Richmond, CA) which was powered by either a FB600 or a FB300 power supply from Fisher Scientific. Centrifugation procedures were performed using either an Avanti J-25 or Avanti J-E unit from Beckman-Coulter (Arlington Heights, IL) for large volumes, or using an Eppendorf 5415C microcentrifuge from Brinkmann Instruments, Inc (Westbury, NY) for small volumes. Photography of agarose gels was done using a Kodak EDAS 290 apparatus connected to a PC running Kodak 1D 3.5 software using a FBTIV-88 transilluminator from Fisher Scientific. PCR was performed using an Eppendorf Mastercycler Gradient from Brinkman Instruments. HPLC separations were performed using a Beckman 366 instrument (Beckman Instruments,

Fullerton, CA). Analytical C18 HPLC column was purchased from Varian (Palo Alto,

CA), while the preparative C18 HPLC column was obtained from Alltech (Deerfield, IL). NMR spectra were acquired on either a Varian Unity 300 or 500 MHz spectrometer, and

1 13 chemical shifts (δ in ppm) are given relative to that of Me4Si (for H and C). Coupling constants are reported in hertz (Hz). DNA sequencing was performed by the Core Facilities of the Institute of Cellular and Molecular Biology at the University of Texas at Austin. Mass spectra were obtained by the Mass Spectrometry Core Facility in the Department of Chemistry and Biochemistry at the University of Texas at Austin. Preparation of Competent Cells. Competent cells were made using the RbCl method.113 A single fresh colony of the appropriate E. coli strain was used to inoculate 2 mL of the Luria-Bertani (LB) liquid medium and the resulting culture was grown overnight at 37 °C with shaking at 250 rpm. A 500 µL aliquot of an overnight culture was used to inoculate 50 mL of the LB medium in an Erlenmeyer flask. When the cell growth

at 37 °C reached an OD600 of approximately 0.4, the culture was transferred into a pre- chilled polypropylene tube and incubated on ice for 30 min. After centrifugation at 3,000 g for 5 min, the supernatant was discarded, and the cell pellet was gently resuspended in

57 one-third the original culture volume of ice-cold RF1 solution (100 mM RbCl, 15% glycerol, 50 mM MnCl2, 30 mM potassium acetate, 10 mM CaCl2, pH 5.8, filter sterilized by passage through a 0.22 µm membrane). After incubation on ice for 15 min, the cell suspension was centrifuged at 3,000 g, 4 °C for 5 min and the resulting cell pellet was resuspended in two twenty-fifths of the original culture volume of ice-cold RF2 solution

(10 mM RbCl, 10 mM MOPS, 75 mM CaCl2, 15% glycerol, pH 6.8). The cells were then aliquoted into 100 µL portions in pre-chilled microcentrifuge tubes and frozen at –80 °C. PCR Amplification of DNA. Although the procedure used for successful PCR amplification of each DNA fragment varied slightly, a general procedure for primer design and PCR amplification is given here. Two oligonucleotide primers complementary to the sequences at each end of the DNA fragment to be amplified were designed and the appropriate restriction enzyme sites were incorporated when appropriate. Several primer design rules were obeyed. First, the melting temperatures

(Tm) of the portion of the primer pair complementary to the target sequence and of the entire primer sequence of the primer pair were matched as closely as possible to ensure efficient annealing throughout the reaction. If restriction enzyme recognition sites were engineered into the primers, a four base pair variable sequence was added to the 5’ end of the primer. For engineered NdeI sites, a seven base pair variable sequence was added to the 5’ end. Each primer was terminated with one or two G or C bases. Additionally, primer pairs were checked using software available on the IDT company website to ensure that stable secondary structures, self-dimers and hetero-dimers did not occur in the chosen sequences. A polymerase-mediated amplification was routinely carried out in a 0.5 mL thin-walled microcentrifuge tube. A typical 100 µL reaction mixture consisted of 62 µL of deionized water, 10 µL of 10X pfu polymerase buffer, 10 µL of deoxyribonucleotide triphosphate (dNTP) mix (2.5 mM each dNTP), 10.0 µL of DMSO,

58 2.5 µL of each of the primers (20 µM stock concentration), 1.0 µL of the template (approximately 0.1 µg), and 2.0 µL of the cloned pfu polymerase (2.5 units, Invitrogen). The reaction mixture was mixed thoroughly by pipette and subjected to the following thermal cycling conditions: (1) 1 cycle of incubation at 95 °C for 5 min; (2) 3 cycles of incubation at 95 °C for 30 s, “X” °C for 30 s, and 72 °C for slightly more than one minute

per kb of target DNA length, where “X” is the Tm of the complementary region of the

primer with the lowest Tm; (3) 27 cycles of incubation at 95 °C for 30 s, “Y” °C for 30 s,

and 72 °C for slightly more than one minute per kb of target DNA, where “Y” is the Tm of the entire primer with the lowest Tm, (4) 72 °C for 10 min. The tubes were held at 4 °C prior to being removed from the thermal cycler. Construction of desVII Disruption Plasmid pDesVII-K2. Plasmid pDesVII-K2 was constructed as follows. A 1025 bp region of S. venezuelae DNA upstream of desVII, referred to as desVII-1, and a 1022 bp region downstream of desVII (desVII-2) were amplified with EcoRI/XbaI and PstI/HindIII restriction sites introduced at the 5'- and the 3'-end via the primer sequences, respectively. These PCR products served as the regions to flank the thiostrepton resistance marker which allow for double-crossover recombination with the desVII locus in the S. venezuelae chromosomal DNA. The primers used for amplification of the desVII-1 fragment were desVII-1-up, 5'- GGCCGAATTCCTGGAGCGCGAGCAGGTG-3' and desVII-1-down, 5'-GCGCTCT-

AGAGAACGAGGTCAGCAGGAC-3'. The primers used for amplification of the desVII-2 fragment were desVII-2-up, 5'-GGCCCTGCAGGATCGTCCCCGAGCTGGA- G-3' and desVII-2-down, 5'-GCGCAAGCTTCAGCAGGGTGCCGACGAC-3'. The introduced restriction sites are shown in bold. The 5'-end of desVII-1 corresponds to a position 21 nucleotides downstream from the desVII start codon, while the 3'-end of desVII-2 corresponds to a position 64 nucleotides upstream from the desVII stop codon.

59 Thus, small portions of the upstream and downstream ends of desVII remain after recombination. The thiostrepton resistance marker and its promoter region, used to replace desVII, was amplified with introduced XbaI and PstI restriction sites using the primers thio2-X-up, 5'-

GGCCTCTAGACGA- ATACTTCATATGC-3' and thio-P- down, 5'-GCGCCTGCAGCTTATC- GGTTGGCCGC-3'. These three fragments were then digested with the appropriate restriction enzymes and sequentially subcloned into pUC119 digested with the appropriate restriction enzymes. After all three fragments had been cloned into pUC119,

the EcoRI-HindIII fragment containing desVII-1, thior, and desVII-2, in that order, was excised from pUC119 and cloned into pKC1139 digested with EcoRI and HindIII to give pDesVII-K2 (Figure 2-5).

Construction of Expression Plasmids pCM1 and pCM1b. S. venezuelae expression plasmid pAX617 has only EcoRI and XbaI sites available for cloning. In order to facilitate cloning of multiple genes using this vector, pAX617 was modified, resulting in pCM1. A 161 bp fragment of pUC119 containing the polycloning site (PCS), with an introduced downstream SpeI restriction site (shown in bold in the primer), was amplified using start primer 119PCS-2-up, 5'-GGACGGCCAGTGAATTCGAGCTCG-3', and halt primer 119PCS2-S-down, 5'-GGCCACTAGTTTATGCTTCCGGCTCGTATG- 3'. The EcoRI site, found at the upstream end of the pUC119 PCS, was retained in the

60 upstream primer, and is shown in the primer in italics. The digested EcoRI-SpeI fragment was ligated into pAX617 digested with EcoRI and XbaI to give pCM1. Since XbaI and SpeI leave complementary digested ends, but are not isoschizomers, ligation gave a nonfunctional site at their junction. Because many of the restriction sites present in the pUC119 polycloning site are also present in pAX617, pCM1 gains only one additional useful restriction site, a HindIII site, which lies downstream of the XbaI site in the pUC119 PCS. Thus, the order of useful sites in the pCM1 PCS is 5’-EcoRI-XbaI- HindIII-3’.

pAX617 was further modified in order to increase its general utility to give pCM1b, which contains four unique restriction sites in the PCS which can be used in cloning. While construction of the plasmids described in this chapter did not require the use of the BamHI site made useable in pCM1b, this vector was used for the construction of pCM27, and therefore its construction is described. pCM1b was created primarily to facilitate the construction of other plasmids which were being made concurrently with those in this chapter, but are not described here. pAX617 contains a single BamHI site outside the PCS. If this site were removed from the vector and re-inserted into the PCS, it could be used for cloning. To do this, desVII/pAX617, which was constructed by Dr. Svetlana Borisova of this group,116 was modified by digestion with BamHI, treatment with Klenow fragment, and blunt end ligation, resulting in deletion of the BamHI site. The product of this step was designated pre-pCM1b. Next, a 161 bp fragment containing the pUC119 PCS, which has a BamHI site between the EcoRI and XbaI sites, was amplified with upstream EcoRI and downstream SpeI sites using the 119PCS-2-up and 119PCS2-S-down, the same primers used for the construction of pCM1. The digested EcoRI-SpeI pUC119 fragment was ligated into pre-pCM1b digested with EcoRI and XbaI to give pCM1b. As in construction of pCM1, ligation of XbaI and SpeI ends gives a non-

61 functional site. pCM1b has a PCS with four useable sites in the order 5’-EcoRI-BamHI- XbaI-HindIII-3’. pCM17. All fragments used for construction of expression plasmids pCM4, pCM8, pCM13, pCM17, pCM18, pCM21, pCM25, pCM26, and pCM27 were amplified containing their native Shine-Dalgarno sequences. Plasmid pCM17 was constructed by amplifying a 1354 bp fragment containing mycB with upstream EcoRI and downstream XbaI restriction sites introduced using the start primer mycB-E-up, 5'-GGCCGAATTC- ATCAGCGCCTGACCA-3', and the halt primer mycB-X-down, 5'-GCGCTCTAGAGT-

CGATCAAGGTCAGCGC-3' (where the restriction sites are shown in bold). The EcoRI, XbaI digested mycB-containing fragment was ligated into EcoRI, XbaI digested pCM1 to give pCM17 (Figure 2-6). pCM18. pCM18 was constructed by amplifying a 2629 bp fragment containing both mydC and mycB with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer mydC-E-up, 5'-GGCCGAATTCGGGTGACCGACAGATT- CA-3' (where the EcoRI restriction site is shown in bold), and the halt primer mycB-X- down, the same primer used for amplification of the mycB-containing gene fragment of pCM17. The EcoRI, XbaI digested mydC, mycB-containing fragment was ligated into

62 EcoRI, XbaI digested pCM1 to give pCM18 (Figure 2-6). pCM4. pCM4 was constructed by amplifying a 1346 bp fragment containing tylM2 with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer tylM2-E-up, 5’-GGCCGAATTCGACCGGAAGGGAAAC-3’ and the halt primer tylM2-X-down, 5’-GCGCTCTAGAGTGGGGTCTACCTTTC-3’ (where restriction sites are shown in bold). The EcoRI, XbaI digested tylM2-containing fragment was ligated into EcoRI, XbaI digested pCM1 to give pCM4 (Figure 2-7). pCM13. pCM13 was constructed by amplifying a 2773 bp fragment containing tylM3 and tylM2 with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer tylM3-2-E-up, 5’-GGCCGAATTCACGATGATGACCGAAC-3’ and the halt primer tylM2-X-down, which was also used for amplification of tylM2 in pCM4. The EcoRI, XbaI digested tylM3, tylM2-containing fragment was ligated into EcoRI, XbaI digested pCM1 to give pCM13 (Figure 2-7).

pCM21. pCM21 was constructed by amplifying a 3359 bp fragment containing tylM3, tylM2, and tylM1 with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer tylM3-2-E-up, which was also used in the amplification of the insert of pCM13, and the halt primer tylM1-X-down 5’-GCGCTCT-

63 AGAGATCTCGGGGTCCCACTTC-3’ (where XbaI restriction site is shown in bold). Also amplified was a 1653 bp fragment containing tyl1a and tylB with introduced upstream XbaI and downstream HindIII restriction sites using the start primer tyl1a-X-up, 5'-GGCCTCTAGAACGAAGGAGCGAGGAGC-3' and the halt primer tylB-H-down, 5'-GCGCAAGCTTTCACGGGCCTTCCT-3' (where the XbaI and HindIII restriction sites are shown in bold). These two fragments were digested with the appropriate restriction enzymes and sequentially cloned into pCM1 digested with the same restriction enzymes to give pCM21 (Figure 2-8).

pCM25. pCM25 was constructed using the tylM3, tylM2-containing fragment from pCM13 and the tyl1a, tylB-containing fragment from pCM21. These were digested with the appropriate restriction enzymes and sequentially cloned into pCM1 digested with the same restriction enzymes to give pCM25 (Figure 2-8).

pCM26. pCM26 was constructed using the tylM2-containing fragment from pCM4 and the tyl1a, tylB-containing fragment from pCM21. These were digested with the appropriate restriction enzymes and sequentially cloned into pCM1 digested with the same restriction enzymes to give pCM26 (Figure 2-9).

64 pCM27. pCM27 was constructed by first amplifying a 2132 bp fragment containing tylM2 and tylM1 with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer tylM2-E-up, which was also used in the amplification of the insert of pCM4, and the halt primer tylM1-X-down, which was also used in the amplification of the tylM3, tylM2, tylM1-containing insert of pCM21. This fragment and the tyl1a, tylB-containing fragment from pCM21 were digested with the appropriate restriction enzymes and sequentially cloned into pCM1b digested with the same restriction enzymes to give pCM27 (Figure 2-9). All constructed plasmids were sequenced using the same primers that were employed to amplify the inserts contained in each plasmid as well as an upstream sequencing primer which binds in the pikromycin promoter region of pCM1, pCM1-seq-up and downstream sequencing primer M13rev which binds in the cloned region of pUC119 downstream of the PCS. Errors in the originally published sequences of tylB, mydC, and mycB were uncovered during the sequencing process.

desVII/pAX617. desVII/pAX617 was constructed by Dr. Svetlana Borisova of this laboratory. It’s construction is described in Dr. Borisova’s dissertation (Figure 2- 10).116 65 pCM8. pCM8 was constructed by first excising an EcoRI/XbaI fragment containing desI and desII from desI,II/pAX617, which was obtained from Dr. Svetlana Borisova, and cloning it into pCM1 digested with the same enzymes. Next, a 1356 bp fragment containing desVII with introduced upstream XbaI and downstream HindIII restriction sites was amplified using the start primer desVII-X-up, 5’-GGCCTCTAGAC- AAGGAAGGACACGACG-3’ and the halt primer desVII-H-down, 5’-GCGCAAGCT- TAGATACAGGGGTGAGGC-3’ (where restriction sites are shown in bold). This fragment was digested with XbaI and HindIII and cloned into the pCM1-derived intermediate containing desI and desII digested with XbaI and HindIII to give pCM8 (Figure 2-10).

Conjugal Transfer of pDesVII-K2 into Wild-Type S. venezuelae. These experiments were performed using the methods published by Bierman et al.117 with minor modifications, which are described in detail by Zhao et al.118 The disruption plasmid pDesVII-K2 was isolated from E. coli DH5α for use in the conjugal transfer experiment. All S. venezuelae cultures were grown in glass tubes containing glass beads. Two 1 mL aliquots of previously prepared mycelia of wild-type S. venezuelae (recipient) was inoculated into 9 mL of tryptic soy broth (TSB) and grown at 29 °C for 18 h. In parallel, E. coli S17-1 competent cells (donor) were transformed with pDesVII-K2, and

66 transformants were selected by overnight growth on Luria-Bertani (LB) agar plates containing apramycin (50 µg/mL) and streptomycin (10 µg/mL) at 37 °C. pDesVII-K2 contains an apramycin resistance gene and E. coli S17-1 itself is streptomycin resistant. Subsequently, 2 mL aliquots of each of the two recipient cultures were re-inoculated into 18 mL of TSB and allowed to grow for another 18 h at 29 °C. Before re-inoculation, the culture was briefly sonicated to avoid clumping. In parallel, 2 single colonies of donor cells were inoculated into 2 mL of TSB containing apramycin (50 µg/mL) and streptomycin (10 µg/mL), and allowed to grow at 37 °C overnight in a rotary shaker at

250 rpm. Next, 1 mL of each recipient culture was transferred into 9 mL of TSB and grown for 3 h at 29 °C. In parallel, 20 µL of each overnight donor culture was transferred to 2 mL of TSB containing apramycin (50 µg/mL) and streptomycin (10 µg/mL), and were allowed to grow at 37 °C for 3 h. The more turbid of the two donor and recipient cultures were chosen for further use. Both donor and recipient cells were recovered by centrifugation and washed twice with TSB to remove antibiotics. A 2 mL aliquot of TSB was then used to resuspend the recipient and donor cells separately. Three mixtures of the donor and recipient cells were prepared in 9:1, 1:1, and 1:10 ratios, respectively. Aliquots of 100 µL of the mixed cultures were spread on freshly prepared AS1 agar plates, which was prepared as follows. A mixture of 1 g of yeast extract, 0.2 g of L- alanine, 0.5 g of L-arginine, 5 g of soluble starch, 2.5 g of NaCl, 10 g of Na2SO4, and 20 g of agar were dissolved in 1 L of deionized water, pH was adjusted to 7.5, and the mixture was autoclaved. After autoclaving, MgCl2 was added to a final concentration of 10 mM.119 Three to five replicate plates were normally made for each donor/recipient mixture. In addition, two plates were inoculated with recipient cells and one plate was inoculated with donor cells as controls. The plates were incubated at 29 °C overnight prior to overlaying with the appropriate antibiotics. For conjugation plates, 1-2 mL of

67 aqueous solution of nalidixic acid (500 µg/mL), apramycin (500 µg/mL) and thiostrepton (500 µg/mL) was applied to the plate surface. For negative controls, one recipient plate was overlaid with all three antibiotics (500 µg/mL each), and one donor plate was overlaid with nalidixic acid (500 µg/mL). For a positive control, one recipient plate was overlaid with nalidixic acid (500 µg/mL). The plates were incubated at 29 °C for 7 to 10 days, by which time small recombinant S. venezuelae colonies appeared. Conjugal Transfer of pDesVII-K2 into KdesI-80 S. venezuelae. This experiment was performed similarly to conjugal transfer of pDesVII-K2 into wild-type S. venezuelae, which is described in detail above, with the following changes. KdesI-80 S. venezuelae is kanamycin resistant, and thus all recipient cultures were grown in the presence of kanamycin (50 µg/mL). Also, the antibiotic solution used in overlay of conjugation plates and recipient negative control plate contained kanamycin (500 µg/mL) in addition to nalidixic acid (500 µg/mL), apramycin (500 µg/mL), and thiostrepton (500 µg/mL), and recipient positive control plate was overlayed with kanamycin (500 µg/mL) in addition to nalidixic acid (500 µg/mL). Screening for Double-Crossover Mutants. Colonies from conjugal transfer of pDesVII-K2 to wild-type S. venezuelae that grew on AS1 plates overlayed with apramycin, thiostrepton, and nalidixic acid were picked and streaked onto SPA plates containing apramycin (50 µg/mL) and thiostrepton (50 µg/mL). SPA agar was prepared as follows. A mixture of 1 g of yeast extract, 1 g of beef extract, 2 g of tryptone, 10 g of glucose, several crystals of FeSO4, and 15 g of agar were dissolved in 1 L of deionized water and autoclaved. Three spore-to-spore passages were carried out on SPA plates without antibiotics to facilitate the homologous recombination between the disruption plasmid and the S. venezuelae chromosome as well as the natural loss of the plasmid by S. venezuelae. The resulting spores were inoculated into 5 mL of SGGP medium without

68 antibiotics and grown at 29 °C overnight. SGGP media is made as follows. A mixture of 4 g of peptone, 4 g of yeast extract, 4 g of casamino acids, 2 g of glycine, 0.25 g of

MgSO4•7 H2O were dissolved in 1 L of deionized water, and autoclaved. After

autoclaving, glucose and KH2PO4 to final concentrations of 1% w/v and 10 mM, respectively, were added to the media. Overnight cultures were diluted 105- to 107-fold with SGGP and plated on SPA supplemented with thiostrepton (50 µg/mL). Single colonies were then replicated on SPA plates containing apramycin (50 µg/mL). Colonies which showed thiostrepton resistant (ThioR) and apramycin sensitive (AprS) characteristics were identified as double-crossover mutants with genotype KdesVII. The protocol for screening double-crossover mutants with KdesI/KdesVII genotype from the conjugal transfer of pDesVII-K2 into KdesI-80 S. venezuelae was similar to that used for KdesVII, except for the following changes. SPA plates used for initial transfer from AS1 plates contained kanamycin (50 µg/mL) in addition to apramycin (50 µg/mL) and thiostrepton (50 µg/mL), diluted SGGP cultures were plated on SPA plates supplemented with kanamycin (50 µg/mL) and thiostrepton (50 µg/mL) and replicated on SPA plates containing apramycin (50 µg/mL), and double-crossover mutants with KdesI/VII genotype were identified on the basis of their thiostrepton and kanamycin resistant (ThioR, KanR) and apramycin sensitive (AprS) characteristics. Individuals with the correct genotypes were grown on SPA plates containing the appropriate antibiotics, and spores from these plates were used directly for the preparation of spore suspensions.

Conjugal Transfer of Expression Plasmids into KdesVII and KdesI/VII S. venezuelae. pCM4, pCM13, pCM17, pCM18, and desVII/pAX617 were transferred to KdesVII S. venezuelae, and pCM8, pCM21, pCM25, pCM26, and pCM27 were transferred to KdesI/VII S. venezuelae. These experiments were performed similarly to

69 conjugal transfer of pDesVII-K2 into wild-type S. venezuelae, which is described in detail above, with the following changes. KdesVII S. venezuelae is thiostrepton resistant, and thus all KdesVII recipient cultures were grown in the presence of thiostrepton (50 µg/mL). Also, the antibiotic solution used in overlay of KdesVII conjugation plates and recipient negative control plate contained thiostrepton (500 µg/mL) in addition to nalidixic acid (500 µg/mL) and apramycin (500 µg/mL), and KdesVII recipient positive control plate was overlayed with thiostrepton (500 µg/mL) in addition to nalidixic acid (500 µg/mL).

KdesI/VII S. venezuelae is kanamycin and thiostrepton resistant, and thus all KdesI/VII recipient cultures were grown in the presence of kanamycin (50 µg/mL) and thiostrepton (50 µg/mL). Also, the antibiotic solution used in overlay of KdesI/VII conjugation plates and recipient negative control plate contained kanamycin (500 µg/mL) and thiostrepton (500 µg/mL) in addition to nalidixic acid (500 µg/mL) and apramycin (500 µg/mL), and KdesI/VII recipient positive control plate was overlayed with kanamycin (500 µg/mL) and thiostrepton (500 µg/mL) in addition to nalidixic acid (500 µg/mL). After conjugal transfer of expression plasmids into the appropriate S. venezuelae hosts, colonies were picked from AS1 plates and streaked on SPA plates containing the appropriate antibiotics. For KdesVII/pCM4, KdesVII/pCM13, KdesVII/pCM17, and KdesVII/pCM18, SPA plates containing thiostrepton (500 µg/mL) and apramycin (500 µg/mL) was used, whereas for KdesI/VII/pCM21, KdesI/VII/pCM25, KdesI/VII/pCM26, and KdesI/VII/pCM27, SPA plates containing kanamycin (500 µg/mL), apramycin (500 µg/mL), and thiostrepton (500 µg/mL) were used. Spores from these plates were used directly for the preparation of spore suspensions.

70 Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains. The wild type S. venezuelae and its mutant strains were stored as spore suspensions in 20% glycerol at –80 °C. Colonies of S. venezuelae strains were first streaked out on SPA plates containing the appropriate antibiotics and allowed to grow for up to one week at 29 °C. To prepare spore suspensions, 2 mL of a sterile solution of 20% glycerol in deionized water was added to each plate, and the spores were gently scraped off the plate surface with sterile Q-tips. The resulting suspensions were passed through a pipette tip loosely packed with cotton wool into sterile Eppendorf tubes and stored at –80 °C.

To prepare frozen mycelia of S. venezuelae strains, a 10-25 µL aliquot of spore suspension was inoculated into 25 mL of SGGP medium containing the appropriate antibiotics in a 250 mL baffled flask and allowed to grow for 28 h at 29 °C in a rotary shaker. Cell cultures were transferred to 50 mL conical tubes and centrifuged at room temperature for 10 min at 1,700 g. The supernatant was discarded and cell pellet was washed twice with 15 mL of sterile 10.3% sucrose. The washed mycelial pellet was resuspended in 10 mL of sterile 10.3% sucrose and stored in 2 mL aliquots at –20 °C. Southern Blot Analysis of KdesVII and KdesI/KdesVII Mutants. To verify that the desVII gene had been replaced by the thiostrepton resistance gene in the KdesVII and KdesI/KdesVII mutants, the genomic DNA from four individuals of the KdesVII genotype and five individuals of the KdesI/VII genotype were isolated using an established procedure.120 Wild-type S. venezuelae genomic DNA was used as a negative control for analysis of both KdesVII and KdesI/VII samples, and KdesI-80 S. venezuelae genomic DNA was used as a negative control for analysis of KdesI/VII samples. A portion of each genomic isolate was digested with PstI, and another portion digested with BamHI. Digested DNA samples were separated by agarose gel electrophoresis and transferred to a nylon membrane using established procedures.113 Two probes, desVI and

71 thior, were made by PCR amplification using digoxigenin-labeled deoxyuridine triphosphate (DIG-dUTP) in place of dUTP. The desVI probe was prepared using primers desVII-2-up, 5'-GGCCCTGCAGGATCGTCCCCGAGCTGGA-G-3' and desVII- 2-down, 5'-GCGCAAGCTTCAGCAGGGTGCCGACGAC-3', which were also used in the construction of the downstream flanking fragment of pDesVII-K2. The desVI probe, when allowed to hybridize with PstI-digested genomic DNA, would give a signal at 12.7 kb for sequences lacking the thiostrepton resistance gene (thior), and a 3.5 kb signal for those having the gene incorporated in place of desVII due to a PstI site at the 3’ end of

the thior gene. The thior probe was prepared using the same primers used in the amplification of the thior gene during construction of pDesVII-K2. The thior probe, when allowed to hybridize with BamHI-digested genomic DNA, would give no signal for sequences lacking the resistance gene, and a signal at 5.9 kb for those having the gene incorporated in place of desVII. The desVI probe was allowed to hybridize to membranes with PstI-digested KdesVII and KdesI/VII DNA, and the thior probe was allowed to hybridize to membranes with BamHI-digested KdesVII and KdesI/VII DNA. DIG-labeled probes were detected by Western Blot using anti-DIG antibody conjugated to alkaline phosphatase, followed by colorimetric detection with nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP). Small-scale Isolation and Analysis of Metabolites Produced by Mutants. Analysis

was carried out by growing the mutants under the conditions favoring the formation of 12-membered ring macrolides.121 Small-scale growth of mutants involved inoculation of a 5 mL culture of the seed medium with 10-25 µL of spore suspension of the mutant to be analyzed, and growth at 29 °C for 48 h at 250 rpm in a rotary shaker. Seed medium was prepared by the addition of 20 g of glucose, 15 g of soybean flour, 5 g of CaCO3, 1 g of

NaCl, and 0.002 g of CoCl2•6H2O to 1L of deionized water, adjustment of the pH to 7.2,

72 and autoclaving. The culture was vortexed briefly to prevent clumping, and 500 µL of the seed culture was transferred to 25 mL of vegetative medium and allowed to grow for another 48-60 h at 29 °C, and shaken at 250 rpm in a rotary shaker. If tylactone (50) was fed, 500 µg (5 µL of 100 mg/mL solution in ethanol) was added to the vegetative culture when it was started. Vegetative media was prepared by the addition of 20 g of glucose,

30 g of soybean flour, 2.5 g of CaCO3, 1 g of NaCl, and 0.002 g of CoCl2•6H2O to 1L of deionized water, adjustment of the pH to 7.2, and autoclaving. The mycelia and insoluble media components were removed by centrifugation at 8,000 rpm for 15 min. The

supernatant was then collected, adjusted to pH 9.5 with 10 M KOH, and extracted with an equal volume of chloroform. After evaporation of solvent, a crude yellow oil was obtained. A small amount of this was analyzed by TLC using a 90:9.9:0.1

chloroform/methanol/25% NH4OH solvent system. The compounds were visualized by vanillin staining followed by heating with a heat gun or under short wavelength UV light. Vanillin stain solution was prepared by dissolving 0.75% w/v vanillin, in methanol with

1.5% v/v concentrated H2SO4. The aglycon 10-deoxymethynolide (124) appeared as a dark blue color, methymycin (30) and other 12-membered ring macrolides hydroxylated at C-10 as an olive green color, and neomethymycin (37) and other 12-membered ring macrolides hydroxylated at C-12 as an orange color. Small-scale preparative TLC of metabolite produced by S. venezuelae mutants for mass spectrometric analysis was performed by the separation of the entire crude extracts of a 25 mL culture by TLC, identification of the spot of interest by vanillin staining of the left and right edges of the TLC, and scraping the silica containing the spot of interest from the central portion of the TLC plate. This silica gel was then crushed and extracted three times with 100 µL of methanol, and the solution was concentrated to 50 µL in a glass mass spectrometry sample tube.

73 Mass spectrometric data for 5-O-desosaminyl-tylactone (130) isolated from

+ KdesVII/pCM13: High resolution CI-MS: C31H54NO7 (M + H) calculated 552.3900, found 552.3881. Mass spectrometric data for 5-O-desosaminyl-tylactone (130) and 2’-glucosyl-5- O-desosaminyl-tylactone (131) isolated from KdesVII/pCM18: High resolution CI-MS

+ (130): C31H54NO7 (M + H) calculated 552.3900, found 552.3925, High resolution CI-MS

+ (131): C37H64NO12 (M + H) calculated 714.4429, found 714.4430. Mass spectrometric data for 5-O-mycaminosyl-tylactone (58) and 2’-glucosyl-5-

O-mycaminosyl-tylactone (132) isolated from KdesI/VII/pCM21: High resolution CI-

+ MS (58): C31H54NO8 (M + H) calculated 568.3849, found 568.3859, High resolution CI-

+ MS (132): C37H64NO13 (M + H) calculated 730.4378, found 730.4378. Mass spectrometric data for 2’-glucosyl-5-O-mycaminosyl-tylactone (132)

+ isolated from KdesI/VII/pCM26: High resolution CI-MS: C37H64NO13 (M + H) calculated 730.4378, found 730.4377. HPLC Analysis of Samples from S. venezuelae Mutants. Samples used in HPLC analysis were prepared by withdrawing a 25 mL sample of the growing culture, subjecting the sample to centrifugation to remove cells and insoluble media components,

adding 20 µL of a 10 mg/mL tylosin solution (as an internal standard) to the supernatant, adjusting the pH of the supernatant to 9.5 with concentrated KOH, and extracting the metabolites in the supernatant with an equal volume of chloroform. After evaporation, the compounds were dissolved in 1 mL of 50% acetonitrile, 50% aqueous 14 mM

triethylamine, pH 2.5. A 15 µL aliquot of the solution was injected into the analytical C18 column. The HPLC solvent system and time program used were identical to those reported by Butler et al.122 Buffer A was 14 mM aqueous triethylamine adjusted to pH 2.5 with trifluoroacetic acid, Buffer B was acetonitrile. Time program consisted of

74 isocratic elution using 26% Buffer B until 14 min, then increase in %B from 26% to 60% over 11 min, followed by increase in %B from 60% to 100% over 10 min. The detector was set at 282 nm for detection of tylactone (50) and its glycosylated derivatives, and at 235 nm for detection of methynolide (127) and neomethynolide (128) and their glycosylated derivatives.

Analysis of Time Course for Metabolite Production by KdesI/VII/pCM25 and KdesI/VII/pCM26 Fed Tylactone. A 40 mL culture of KdesI/VII/pCM25 and a 40 mL culture of KdesI/VII/pCM26 were inoculated in seed media containing the appropriate antibiotics using 25 µL of spore suspension, and grown in parallel for 48 h at 29 °C, 250 rpm. A 20 mL aliquot of each culture was used to inoculate 1 L of the vegetative media containing the appropriate antibiotics and 10 mg of tylactone. These cultures were grown in parallel for 24 h, during which time 25 mL aliquots of culture were removed every 2 h, and the metabolites were extracted and prepared for HPLC analysis as described above. The sum of the integration of 58 and 132 peaks were plotted versus time in Figure 2-19.

Large-Scale Isolation and Analysis of Extracts of KdesVII-2-1 and KdesVII/pCM4 S. venezuelae Mutants. A 50 mL culture of KdesVII-2-1 was grown in seed media containing the appropriate antibiotics at 29 °C, 48 h, and was shaken at 250 rpm in a rotary shaker. A 20 mL aliquot of seed culture was used to inoculate each of two 1 L flasks of vegetative media, which were grown at 29 °C, 250 rpm for 48 h. Culture broth was obtained by removal of cells and insoluble components as described previously for small-scale isolation. Compounds were extracted from the aqueous layer three times using an equal volume of chloroform each time. After evaporation of solvent, compounds were separated by silica gel flash chromatography using a gradient of 0-20% methanol in chloroform and analyzed by TLC and 1H NMR. An identical procedure was used for the analysis of extracts of KdesVII/pCM4, except that a 150 mL seed culture

75 was used, and 20 mL of seed culture was used to inoculate each of six 1 L flasks of vegetative media.

Large-scale Isolation and NMR Analysis of Glycosylated Tylactone Derivatives from S. venezuelae Mutants Fed Tylactone. For isolation and NMR characterization of 5- O-mycaminosyl-tylactone (58) and 2'-glucosyl-5-O-mycaminosyl-tylactone (132), a 100 mL culture of KdesI/VII/pCM25 was grown in seed media containing the appropriate antibiotics at 29 °C, 48 h, and was shaken at 250 rpm in a rotary shaker. A 20 mL aliquot of seed culture was used to inoculate each of four 1 L flasks of vegetative media which

contained 10 mg/L of tylactone (50). It was previously observed that after 24 h of growth under these conditions, 58 was the predominant product, whereas after 48 h of growth, 132 predominated. Therefore, 2 L of the fermentation broth were harvested after 24 h, and the other 2 L were harvested after 48 h. Culture broth was obtained by removal of cells and insoluble components as described previously for small-scale isolation. Compounds were extracted from the aqueous layer three times using an equal volume of chloroform each time. After evaporation of solvent, compounds of interest were partially purified by silica gel flash chromatography using a gradient of 0-28% methanol in

chloroform. Compound 58 was further purified by preparative scale HPLC on a C18 column by isocratic elution using 60% acetonitrile in 57 mM aqueous ammonium acetate buffer. Compound 132 was purified in the same manner, but using isocratic elution with

80% acetonitrile in 57 mM aqueous ammonium acetate buffer. Yield of purified 58 was 3.0 mg, and that of 132 was 6.8 mg. Analysis of each compound by 1H, 13C, HSQC, HMBC, COSY, and NOESY spectroscopies allowed conclusive assignment of the structure of each compound.

1 Spectral data for 5-O-mycaminosyl-tylactone (58): H NMR (500 MHz, CD3OD)

δ 7.23 (1H, d, J = 15.3, 11-H), 6.43 (1H, d, J = 15.5, 10-H), 5.66 (1H, d, J = 10.4, 13-H),

76 4.68 (1H, br dt, J = 9.6, 2.4, 15-H), 4.29 (1H, d, J = 7.4, 1'-H), 3.75 (1H, d, J = 9.6, 5-H), 3.71 (1H, d, J = 9.8, 3-H), 3.41 (1H, dd, J = 10.5, 7.5, 2'-H), 3.27 (1H, d, J = 6.0, 5'-H), 3.23 (1H, d, J = 9.6, 4'-H), 2.78 (1H, tq, J = 10.2, 6.6, 14-H), 2.70 (1H, m, 3'-H), 2.69

(6H, s, N-CH3), 2.63 (1H, m, 8-H), 2.49 (1H, dd, J = 17.6, 9.7, 2-H), 2.03 (1H, d, J = 17.5, 2-H), 1.88 (1H, m, 16-H), 1.84 (3H, s, 12-Me), 1.72 (1H, m, 7-H), 1.68 (1H, m, 4- H), 1.57 (1H, m, 6-H), 1.57 (1H, m, 16-H), 1.57 (2H, m, 19-H), 1.44 (1H, m, 7-H), 1.25 (3H, d, J = 5.8, 5'-Me), 1.19 (3H, d, J = 7.0, 8-Me), 1.07 (3H, d, J = 6.6, 14-Me), 1.03 (3H, d, J = 6.8, 4-Me), 0.93 (3H, t, J = 7.3, 16-Me), 0.85 (3H, t, J = 7.3, 19-Me). 13C

NMR (125 MHz, CD3OD) δ 206.9 (C-9), 175.0 (C-1), 149.7 (C-11), 147.8 (C-13), 134.9 (C-12), 119.9 (C-10), 104.9 (C-1'), 80.0 (C-15), 79.5 (C-5), 74.1 (C-5'), 72.0 (C-3’), 72.0

(C-4'), 71.2 (C-2'), 68.0 (C-3), 46.5 (C-8), 42.2 (C-4), 42.2 × 2 (N-Me), 41.2 (C-2), 40.0 (C-14), 34.9 (C-7), 28.8 (C-6), 25.7 (C-16), 22.2 (C-19), 17.9 (C-6'), 17.9 (C-21), 16.3 (C-23), 13.1 (C-22), 12.5 (C-20), 10.0 (C-17), 9.6 (C-18). High resolution CI-MS:

+ C31H53NO8 (M + H) calculated 568.3849, found 568.3841. Spectral data for 2'-glucosyl-5-O-mycaminosyl-tylactone (132): 1H NMR (500

MHz, CD3OD) δ 7.23 (1H, d, J = 15.3, 11-H), 6.44 (1H, d, J = 15.5, 10-H), 5.66 (1H, d, J = 10.2, 13-H), 4.67 (1H, dt, J = 9.5, 2.3, 15-H), 4.44 (1H, d, J = 7.4, 1'-H), 4.39 (1H, d, J = 7.8, 1"-H), 3.92 (1H, dd, J = 11.7, 2.1, 6"-H), 3.79 (1H, d, J = 9.8, 5-H), 3.71 (1H, d, J = 11.4, 3-H), 3.70 (1H, d, J = 11.6, 6"-H), 3.39 (1H, m, 2'-H), 3.38 (1H, m, 4'-H), 3.36

(1H, m, 4"-H), 3.29 (1H, m, 5"-H), 3.27 (1H, m, 3"-H), 3.26 (1H, m, 5'-H), 3.25 (1H, m, 2"-H), 2.86 (1H, t, J = 10.2, 3'-H), 2.78 (1H, tq, 10.2, 6.5, 14-H), 2.68 (6H, s, N-Me), 2.61 (1H, m, 8-H), 2.48 (1H, dd, J = 17.6, 9.9, 2-H), 2.03 (1H, d, J = 17.5, 2-H), 1.84 (1H, m, 16-H), 1.84 (3H, s, 12-Me), 1.79 (1H, m, 4-H), 1.76 (1H, m, 7-H), 1.57 (1H, m, 16-H), 1.57 (1H, m, 6-H), 1.50 (2H, m, 19-H), 1.36 (1H, m, 7-H), 1.24 (3H, d, J = 6.0, 5'- Me), 1.19 (3H, d, J = 7.5, 8-Me), 1.08 (3H, d, J = 7.0, 4-Me), 1.07 (3H, d, J = 6.6, 14-

77 Me), 0.93 (3H, t, J = 7.3, 16-Me), 0.84 (3H, t, J = 7.2, 19-Me). 13C NMR (125 MHz,

CD3OD) δ 207.0 (C-9), 175.2 (C-1), 149.8 (C-11), 147.8 (C-13), 135.0 (C-12), 119.9 (C- 10), 107.6 (C-1"), 103.0 (C-1'), 81.8 (C-2'), 80.0 (C-15), 79.5 (C-5), 79.1 (C-5"), 77.7 (C- 4"), 76.2 (C-2"), 73.9 (C-5'), 72.6 (C-4'), 71.7 (C-3"), 71.6 (C-3'), 68.3 (C-3), 63.2 (C-6"),

46.6 (C-8), 42.2 (C-4), 42.2 × 2 (N-Me), 41.1 (C-2), 40.1 (C-14), 34.8 (C-7), 30.8 (C-6), 25.7 (C-16), 22.3 (C-19), 18.0 (C-21), 17.6 (C-6'), 16.3 (C-23), 13.1 (C-22), 12.5 (C-20),

+ 10.0 (C-17), 10.0 (C-18). High resolution CI-MS: C37H63NO13 (M + H) calculated 730.4378, found 730.4369.

3. RESULTS AND DISCUSSION

Disruption of desVII in Wild-Type and KdesI-80 S. venezuelae. In order both to obtain evidence on the functions of the putative GT activator proteins TylM3 and MydC and to probe the donor and acceptor specificities of the corresponding partner GTs TylM2 and MycB, an in vivo approach was used. This involved creation of two desVII gene disruption mutants in S. venezuelae. The desVII gene is presumed to encode the GT responsible for the attachment of desosamine (29) to 10-deoxymethynolide (124) during the biosynthesis of methymycin/neomethymycin (30/37) in S. venezuelae (Figure 2-1). Disruption of desVII in wild-type S. venezuelae would create an in vivo system suitable for testing the activities and specificities of exogenous GTs and their proposed activator

proteins. However, in this system, one is limited to the use of TDP-D-desosamine (41) that is expected to accumulate in the mutant, as a GT donor substrate. Since the native

donor substrate of TylM2 is TDP-D-mycaminose (52), another disruption mutant that would allow heterologous expression of genes responsible for the biosynthesis of 52 along with a GT and its proposed activator protein, was designed. The KdesI-80 S.

78 venezuelae mutant, which was previously constructed by Dr. Svetlana Borisova of this group,41 was used as the starting point for disruption of desVII, to construct a

KdesI/desVII double mutant which would accumulate TDP-4-keto-6-deoxy-D-glucose (14) and 124 (Figure 2-11). Sugar biosynthesis enzymes from the mycaminose pathway or other pathways could then be expressed in this double mutant strain, to produce the target TDP-sugar which could serve as a substrate for the co-expressed GT and proposed activator protein.

KdesVII S. venezuelae KdesI/VII S. venezuelae O O Me Me

Me Me Me Me Me O Me O O O O O OH Me2N O OH HO HO HO Me OTDP Me Me OTDP Me 41 124 14 124

mycaminose biosynthetic GT, activator genes protein Me HO O Me2N glycosylated macrolides HO OTDP 52

GT, activator protein

glycosylated macrolides

Figure 2-11

To this end, the desVII disruption plasmid pDesVII-K2 (Figure 2-5) containing a thiostrepton resistance gene flanked by the two regions flanking the desVII genomic locus was constructed using the E. coli/S. venezuelae shuttle vector pKC1139. pDesVII-K2 is designed to be introduced into S. venezuelae by conjugation from donor strain E. coli S17-1 and to undergo double crossover recombination with the desVII locus, replacing all but the extreme 5’- and 3’-ends of the desVII gene with the thiostrepton resistance marker, including promoter. A number of S. venezuelae colonies possessing the correct

79 antibiotic resistance phenotypes were obtained from both the KdesVII and KdesI/VII construction experiments. Southern blot analysis of four colonies of the KdesI/VII genotype (Figure 2-12) and five colonies of the KdesVII genotype (data not shown) in comparison with wild-type and KdesI-80 clearly showed that each individual colony underwent double crossover recombination resulting in disruption of desVII. One individual of each phenotype, designated KdesVII-2-1 and KdesI/VII #1, were chosen for use in all subsequent experiments.

Figure 2-12. Southern blot analysis of KdesI/VII S. venezuelae. M = marker lane, 1 = wild-type, 2 = KdesI-80, 3-6 = KdesI/VII #1-4, respectively. Results of Southern blot analysis of KdesVII were identical to those shown here.

Analysis of Metabolites Produced by KdesVII-2-1. As a confirmation of DesVII function and to verify that disruption of desVII abolished all GT activity in KdesVII-2-1, extracts from a 2 L culture of KdesVII-2-1 were subjected to separation by silica gel chromatography and fractions were analyzed by TLC and 1H NMR. As expected, the major product 124 (130 mg), along with traces of 127 and 128, were detected. However, no glycosylated macrolides were found, confirming the function of DesVII as the desosaminyltransferase in the pathway for biosynthesis of 30, 32, and 37.

80 Design, Construction, and Metabolite Analysis of KdesVII/desVII/pAX617 and KdesI/VII/pCM8. In order test whether disruption of desVII in the KdesVII-2-1 mutant or the KdesI/VII #1 mutant caused a polar effect on the expression of other genes in the cluster, two control experiments were performed. In the first experiment, the plasmid desVII/pAX617, which contains desVII, was transferred to KdesVII-2-1 and the metabolite profile of resulting mutant, KdesVII/desVII/pAX617, was analyzed by TLC. Results clearly showed substantial production of wild-type metabolites 30 and 32, confirming that disruption of desVII had not significantly impaired expression of other genes in the cluster. In the second experiment, the plasmid pCM8, which contains desI, desII, and desVII, was introduced into KdesI/VII #1, giving mutant KdesI/VII/pCM8. Previous work in our group has established that disruption of desI in KdesI-80 has a polar effect on expression of desII. Thus, expression of desI, desII, and desVII from pCM8 in KdesI/VII #1 should restore wild-type metabolite production. Indeed, TLC analysis of small-scale extracts of KdesI/VII/pCM8 clearly showed that presence of wild-type metabolites 30 and 32, confirming that disruption of desVII had also not significantly impaired expression of other genes in the KdesI/VII mutant. Construction of Expression Plasmid pCM1. pCM1 is a modified version of the plasmid pAX617 (pDHS617)123 which is an E. coli/Streptomyces shuttle vector with a strong promoter (Ppik) from the pikromycin cluster of S. venezuelae upstream of the cloning site capable of high-level expression of cloned genes in S. venezuelae. pCM1 was constructed from pAX617 in order to facilitate cloning of several genes. pAX617 contained EcoRI and XbaI cloning sites downstream of Ppik. pCM1 has an additional HindIII site downstream of the XbaI site, allowing direct cloning of two fragments into the vector without the need for subcloning. pCM1 was used as the starting point for

81 construction of all expression plasmids described in this chapter except for desVII/pAX617.

Design, Construction, and Metabolite Analysis of KdesVII/pCM4 and KdesVII/pCM13. pCM4, which contains tylM2, and pCM13, which contains tylM3 and tylM2, were designed for expression in KdesVII S. venezuelae to test the importance of the proposed GT activator protein TylM3 in TylM2-dependent glycosylation (Figure 2- 7). A previous report showed that TylM2 was capable of attaching desosamine (29) to tylactone (50) in an engineered mutant of Saccharopolyspora erythraea.124 Therefore,

exogenous 50 could be fed to the culture, and TylM2 activity could be detected by the formation of desosaminyl tylactone (130). Comparison of the amount of 130 formed by KdesVII/pCM4 and KdesVII/pCM13 mutants could provide evidence for the proposed function of TylM3. Also, if 50 were not fed to these mutants, the ability of TylM2 to transfer 29 to 124 i.e. to complement DesVII, could also be tested (Figure 2-13).

O Me

Me Me

Me Me O O OH Me Me O OH TylM2 Me Me NMe2 50 TylM3 HO Me Me O Me O O 29 Me O OH O Me Me2N OH 130 41 OTDP

O Me

Me Me O O O Me Me O OH R1 R1 Me Me Me TylM2 Me NMe2 Me Me O 124 TylM3 HO O R2 O R O O Me 2 O OH Me 29 Me Me Me Me O 30 R1 = OH, R2 = H 37 R = H, R = OH 127 R1 = OH, R2 = H Me2N 1 2 HO 128 R1 = H, R2 = OH OTDP 41 Figure 2-13

82

Figure 2-14. TLC analysis of extracts of KdesVII/pCM4 and KdesVII/pCM13 fed 50. 1 = KdesVII/pCM4 0-4% MeOH elution, 2 = KdesVII/pCM4 4-15% MeOH elution, 3 = KdesVII/pCM13 0-4% MeOH elution, 4 = KdesVII/pCM13 4-15% MeOH elution. Initial TLC analysis of metabolites isolated from small-scale cultures of KdesVII/pCM4 and KdesVII/pCM13 mutants grown in the absence of 50 failed to reveal any obvious new spots on TLC. Subsequent large-scale analysis of extracts from KdesVII/pCM4, which was performed by Dr. Haruko Takahashi of this group, detected only macrolide derivatives 124, 127, and 128, confirming that this mutant was incapable of forming any glycosylated compounds. However, TLC analysis of metabolites produced by a small-scale culture of KdesVII/pCM13 in the presence of 50 revealed formation of a small amount of a new product (Figure 2-14). Preparative TLC isolation of the new spot from a 25 mL culture fed 2.5 mg of 50 and subsequent high resolution mass spectrometric analysis identified this compound as desosaminyl tylactone (130). Production of the compound was not very efficient (>5% conversion) as estimated by the ratio of remaining 50 to the new product 130. Interestingly, TLC analysis of extracts of KdesVII/pCM4 fed 50, which was performed in parallel with KdesVII/pCM13, failed to reveal any new spot corresponding to 130, suggesting that TylM2 was more active in the 83 presence of TylM3. This supports the proposed function of TylM3 as an activator protein for TylM2. Detection of 130 further confirms that TylM2 is capable of accepting 41 as a substrate, although the low yield suggests that 41 is not as competent a substrate of TylM2 as its natural substrate 52. Failure to detect any glycosylation of 124 by either mutant in the absence of 50 suggests that TylM2 is not capable of recognizing the alternative macrolactone substrate 124.

Design, Construction, and Metabolite Analysis of KdesI/VII/pCM21 and KdesI/VII/pCM27. pCM21, which contains tylM3, tylM2, tyl1a, tylB, and tylM1, encodes all enzymes necessary for conversion of TDP-4-keto-6-deoxy-D-glucose (14) to TDP-D- mycaminose (52) (Tyl1a, TylB, and TylM1) as well as the GT TylM2 and the proposed activator protein, TylM3 (Figure 2-8, Figure 2-9). Tyl1a is a hexose ketoisomerase responsible for conversion of 14 to 38 whose identification and functional analysis will be discussed in Chapters 3 and 4. pCM27, which contains tylM2, tyl1a, tylB, and tylM1, is identical to pCM21 except that it lacks tylM3. These expression plasmids are designed for introduction into KdesI/VII in order to test the function of TylM3 in the TylM2-

O Me NH2 O HO OH O Me HO R O OH OTDP 133 14 NHAc O HO OH Me R O Me Tyl1a O 134 Me Me O O Me Me O Me Me HO Me Me Me O OH TylM2 Me Me O OH NMe2 Me OTDP O OH (TylM3) HO OH 38 OH DesG Me Me O Me Me O Me Me O O O O 50 49 O NMe2 TylB O OH TylM1 O OH HO O Me or Me Me Me O HO O DesVI HO 58 OH Me N H2N 2 HO HO OH OH O OTDP 52 OTDP 53 TylM2 Me 132 (TylM3) R1 Me Me NMe O O 2 HO OH Me R Me 2 O O O Me O R1 Me Me Me Me 49 H2N O OH OTDP R 135 R1 = OH, R2 = H 40 2 O OH 136 R1 = H, R2 = OH Me Me 124 Figure 2-15 84 catalyzed glycosyltransfer reaction using the natural substrates of TylM2, 52 biosynthesized by the mycaminose pathway enzymes, and exogenous 50. Also, if 50 is not added to the media, the ability of TylM2 to use 10-deoxymethynolide (124) as an acceptor substrate would be tested using the natural TylM2 substrate, 52 (Figure 2-15).

Figure 2-16. TLC analysis of extracts of KdesI/VII/pCM21 and KdesI/VII/pCM27 in the presence and absence of 50. 1 = KdesI/VII/pCM21, 2 = KdesI/VII/pCM21 + 50, 3 = KdesI/VII/pCM27, 4 = KdesI/VII/pCM27 + 50.

Initial TLC analysis of metabolites isolated from KdesI/VII/pCM21 and KdesI/VII/pCM27 mutants in the absence of exogenous 50 failed to reveal any obvious new spots on TLC. However, TLC analysis of metabolites isolated from a 25 mL culture of KdesI/VII/pCM21 grown in the presence of 50 revealed two new polar spots with the characteristic maroon color of tylactone derivatives, whereas TLC analysis of metabolites from a culture of KdesI/VII/pCM27 also grown in the presence of 50 showed no corresponding new spots (Figure 2-16). Preparative TLC of each of these compounds revealed that the less polar of the two is mycaminosyl tylactone (58) and the more polar

85 has mass consistent with it being a glucosyl derivative of mycaminosyl tylactone (132). There is precedent for glucosylation of macrolides in S. venezuelae. A GT, DesG, which is capable of glucosylation of endogenous 12-membered ring macrolides 30/32 at the C-2 position of the desosamine moiety has been identified and partially characterized.116 This modification is thought to be a self-protection mechanism for the host against the effects of the antibiotic, as glucosylated 30/32 lack antibiotic activity. A secreted glycosidase DesR that removes the glucose residue, has also been characterized from S. venezuelae.118 These results have several implications. First, they suggest that the mycaminose biosynthetic pathway is indeed functional in KdesI/VII/pCM21, and that efficient glycosylation of 50 is occurring in this mutant. Second, the fact that no glycosylated derivatives of 50 were produced by KdesI/VII/pCM27, which lacks TylM3, suggests that TylM3 is critical for the activity of TylM2 in this mutant. This suggestion, however, turns out to be incorrect as shown by experiments which will be discussed in the next section of this chapter. Third, the fact that no glycosylated compounds were found in either mutant in the absence of 50 suggests that TylM2 is not capable of transferring mycaminose to 124. This observation also turns out to be incorrect in light of experiments discussed in the next section of this chapter. Finally, the production of 132 by KdesI/VII/pCM21 shows that DesG can glucosylate 58.

Design, Construction, and Metabolite Analysis of KdesI/VII/pCM25 and KdesI/VII/pCM26. pCM25, which contains tylM3, tylM2, tyl1a, and tylB, and pCM26, which contains tylM2, tyl1a, and tylB, are identical to pCM21 and pCM27, respectively, except that both lack tylM1, which encodes the N,N-dimethyltransferase catalyzing the final step in formation of 52 (Figure 2-8, Figure 2-9). These expression plasmids, when introduced into KdesI/VII, were expected to function similarly to pCM21 and pCM27, respectively, allowing assessment of the importance of TylM3 in the TylM2-catalyzed

86 glycosylation reaction. However, because tylM1 is not included in these constructs, it is possible that the N,N-dimethylation step in 52 biosynthesis would not occur and

macrolide derivatives bearing the 3-amino-6-deoxy-D-glucose (133) or 3-N-acetyl-6-

deoxy-D-glucose (134) might be obtained (Figure 2-15). Recall that macrolides bearing a 3-N-acetylamino sugar were observed upon disruption of desVI in S. venezuelae, indicating that a promiscuous sugar N-acetyltransferase must be present in the strain. It is also possible that the endogenous N,N-dimethyltransferase DesVI would be capable of substituting for TylM1 in the mycaminose pathway, leading to the formation of mycaminose-bearing tylactone derivatives. Previous in vivo work demonstrated that tylM1 complements a desVI disruption mutant, although no experiments designed to test the converse scenario (complementation of tylM1 by desVI) have yet been conducted. Also, in vitro substrate specificity studies on TylM1 and DesVI demonstrated that both are capable of turnover of the TylM1 substrate TDP-3-amino-3,6-dideoxy-D-glucose (53)

and the DesVI substrate TDP-3-amino-3,4,6-trideoxy-D-glucose (40), although turnover of 53 by DesVI was 22-fold slower than with its natural substrate 40 (Figure 2-15).125 TLC analysis of small-scale culture extracts of KdesI/VII/pCM25 grown in the presence of 50 revealed, as expected, the presence of two glycosylated derivatives of 50 (Figure 2- 17), which were identified by preparative TLC and high resolution mass spectrometry as 58 and 132. This confirms that DesVI is capable of substituting for TylM1 in the

biosynthetic pathway of 52. However, analysis of small-scale extracts of KdesI/VII/pCM25 grown in the absence of 50 unexpectedly revealed the presence of small quantities of two new polar spots with the characteristic olive green and orange colors of methynolide (127) and neomethynolide (128) derivatives (Figure 2-17). These two compounds were identified as 3-O-mycaminosyl methynolide (135) and 3-O- mycaminosyl neomethynolide (136) by TLC and HPLC analyses in which these

87 Figure 2-17. TLC analysis of extracts of KdesI/VII/pCM25 and KdesI/VII/pCM26 in the presence and absence of 50. 1 = KdesI/VII/pCM25, 2 = KdesI/VII/pCM26, 3 = KdesI/VII/pCM25 + 50, 4 = KdesI/VII/pCM26 + 50.

compounds displayed identical Rf values and retention times as authentic standards of 135 and 136 obtained from other experiments that will be discussed in Chapter 3 (Figure 2-18). Similar TLC (Figure 2-17) and HPLC (Figure 2-18) analyses of small-scale extracts of KdesI/VII/pCM26 grown in the absence of 50 revealed no detectable 135 and 136, whereas analysis of small extracts of this mutant grown in the presence of 50

surprisingly revealed two glycosylated derivatives of 50 with Rf values identical to those of 58 and 132 (Figure 2-17). The more polar of these was identified by preparative TLC and high resolution mass spectrometry as 132. At first glance, results obtained from analysis of KdesI/VII/pCM25 in the absence of 50 and KdesI/VII/pCM26 in the presence of 50 seem to be at odds with results obtained from previously described experiments on KdesI/VII/pCM21 and KdesI/VII/pCM27 which suggested that TylM3 was absolutely required for TylM2- dependent glycosylation and that TylM2 was incapable of glycosylation of 124. However, in light of results derived from work with KdesI/VII/pCM25 and

88 0.06 0.06

0.05 0.05

0.04 0.04 135 0.03 136 0.03 AU A U 0.02 0.02 A 0.01 0.01 B 0.00 0.00

7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 11.50 Minutes

Figure 2-18. HPLC traces of metabolite extracts of (A) KdesI/KdesVII/pCM25 and (B) KdesI/KdesVII/pCM26 showing formation of 3-O-mycaminosyl methynolide (135) and 3-O-mycaminosyl neomethynolide (136) after growth for 60 h.

KdesI/VII/pCM26, the following interpretations seem to be the most likely. First, after identification of 135 and 136, it is clear that TylM2 is indeed capable of transferring 49 to 124, although not very efficiently, as only ~4% of 124 was converted to 135 and 136 by TylM2 as estimated by HPLC peak integration. This would imply that 135 and 136 are also produced by KdesI/VII/pCM21, although they were not detectable by TLC analysis. Second, the presence of 58 and 132 in extracts from KdesI/VII/pCM26 fed 50 demonstrates that TylM3 is not absolutely required for TylM2-dependent glycosylation of 50. However, the amount of 58 and 132 produced by KdesI/VII/pCM26 was much less when compared to that produced by KdesI/VII/pCM25 under identical conditions by TLC analysis, suggesting that the presence of TylM3 does enhance glycosylation by TylM2. The fact that 135 and 136 were produced by KdesI/VII/pCM25 but were not detected by TLC or HPLC analysis of extracts of KdesI/VII/pCM26 supports this assertion. However, the observation that KdesI/VII/pCM27 and KdesI/VII/pCM26 display different metabolite profiles in the presence of exogenous 50 is perplexing. The

89 most likely explanation for this is that KdesI/VII/pCM27 possesses a defect in sugar biosynthesis or macrolide glycosylation acquired either during construction of pCM27 or during genetic manipulations following conjugal transfer of pCM27 to KdesI/VII. Resolution of this issue was not explored further. Subsequent characterization of the effect of TylM3 on TylM2 activity focused on the KdesI/VII/pCM25 and KdesI/VII/pCM26 mutant pair, which seemed to display behavior that was more easily interpretable.

Analysis of the Time Course of Metabolite Production in KdesI/VII/pCM25 and KdesI/VII/pCM26 in the Presence of Tylactone. In order to obtain quantitative data of the effect of TylM3 on the activity of TylM2 in vivo, monitoring the production of 58 and 132 by the KdesI/VII/pCM25 and KdesI/VII/pCM26 mutants grown in parallel in the presence of 50 was carried out. A 1 L vegetative culture of each mutant was grown in parallel containing 10 mg of 50, and samples of each culture were taken at 2 h time intervals for 24 h and processed to obtain crude metabolites. These were subjected to immediate HPLC analysis in order to quantify formation of glycosylated products 58 and 132. Data normalized to a tylosin (51) internal standard added to the culture broth before extraction (Figure 2-19) clearly shows that the KdesI/VII/pCM25 mutant produces more glycosylated compounds at a faster rate than the KdesI/VII/pCM26 mutant. At the 14 h time point, KdesI/VII/pCM25 has produced 34-fold more glycosylated compounds than KdesI/VII/pCM26 (Figure 2-19, Figure 2-20). This establishes that TylM3 does have a stimulatory effect on TylM2-dependent glycosylation. However, KdesI/VII/pCM26 is still capable of producing significant amounts of 58 and 132, demonstrating that under the conditions of the assay, TylM3 is not absolutely required for TylM2 to be active. This is consistent with earlier TLC comparison between KdesI/VII/pCM25 and

90 KdesI/VII/pCM26 (Figure 2-17) which showed formation of 58 and 132 by both mutants, but more production by KdesI/VII/pCM25 than by KdesI/VII/pCM26.

Tylactone Glycosylation by KdesI/VII/pCM25 and KdesI/VII/pCM26 3.50E+06

t KdesI/VII/pCM25

a 3.00E+06 n KdesI/VII/pCM26 o

i 2.50E+06 t a m 2.00E+06 gr 2 n e

t 1.50E+06 n 28 i

k 1.00E+06 a 5.00E+05 pe 0.00E+00 0 5 10 15 20 25 30 time (h)

Figure 2-19. Time course for TylM2-dependent glycosylation of 50 in KdesI/VII/ pCM25 and KdesI/VII/pCM26.

Large-Scale Isolation and Structure Determination of 58 and 132. In order to more conclusively identify the two novel derivatives of 50 produced by KdesI/VII/pCM25, large-scale growth of this mutant in the presence of 50 was

50 0.30 0.3

58 0.2

0.20 0.2

A 0.1 U

0.10 132 0.1 A 0.0 B 0.00 0.0

18 20 22 24 26 28 30 32 34 time (min)

Figure 2-20. HPLC traces of metabolite extracts of (A) KdesI/KdesVII/pCM25 and (B) KdesI/KdesVII/pCM26 showing formation of 5-O-mycaminosyl tylactone (58) and 2’-glucosyl-5-O-mycaminosyl tylactone (132) after growth for 14 h in the presence of 50.

91 performed. Purification of each compound and NMR analysis allowed conclusive assignment of the structures of 58 and 132.

Design, Construction, and Metabolite Analysis of KdesVII/pCM17 and KdesVII/pCM18. pCM17, which contains mycB, and pCM18, which contains mydC and mycB (Figure 2-6), were designed for expression in KdesVII S. venezuelae to test the effect of the proposed activator protein, MydC, on the activity of the GT MycB, and to test the ability of MycB to accept non-natural aglycon substrates 124 and 50 (Figure 2- 21). In fact, addressing the former issue requires MycB to accept at least one of the non-

natural aglycon substrates, as the natural MycB aglycon, 129, was not available. TLC (Figure 2-22) and HPLC (Figure 2-23) analyses of metabolites isolated from a small- scale extracts of KdesVII/pCM18 showed near quantitative (99%) conversion of 124 to a mixture of 30, 37, and another less polar compound which was not identified, but is likely to be either 126 or 31. But analysis of KdesVII/pCM17 showed no obvious glycosylated spots on TLC (Figure 2-22), and HPLC analysis (Figure 2-23) showed ~5% conversion of 124 to 30 and 37 as estimated by HPLC peak integration. Compounds 30 and 37 were O O Me Me MycB R1 Me Me Me Me NMe2 O MydC O HO R2 O OH O O O Me Me Me Me Me

124 30 R1 = OH, R2 = H Me 37 R1 = H, R2 = OH O

Me2N HO OTDP O 41 O O Me Me Me Me Me Me Me Me Me NMe2 HO Me Me Me DesG Me Me O Me Me Me O O OH O O O O 29 O NMe2 O OH MycB O OH O OH Me MydC HO O 50 Me Me 130 OH

HOHO Figure 2-21 131

92 Figure 2-22. TLC analysis of extracts of KdesVII/pCM17 and KdesVII/pCM18 in the presence and absence of 50. 1 = KdesVII/pCM17, 2 = KdesVII/pCM18, 3 = KdesVII/pCM17 + 50, 4 = KdesVII/pCM18 + 50.

identified in HPLC by comparison to authentic standards of these compounds. TLC analysis of small- scale extracts of KdesVII/pCM17 and KdesVII/pCM18 grown in the presence of 50 also showed that two polar derivatives of 50 were produced by KdesVII/pCM18, whereas no derivatives of 50 were detected in metabolite produced by KdesVII/pCM17 (Figure 2-22). Preparative TLC isolation of these compounds and analysis by high resolution mass spectrometry established, as expected, that the less polar of the new compounds was 130 and the more polar was a glucosylated derivative of 130 which is likely to be 2’-glucosyl-5-O-desosaminyl tylactone (131). These results clearly show that MycB is capable of transferring 29 to aglycons 124 and 50, and that MydC has a significant activating effect on MycB. They also demonstrate that DesG is capable of glucosylating 130. Interestingly, MycB more efficiently accepts the 12-membered macrolactone 124 than the 16-membered macrolactone 50, which is structurally more similar to its natural substrate 129.

93 0.35

0.30 0.30 37 0.25 0.20 0.20

30 A U AU 0.15

0.10 0.10 A 0.05 B 0.00 0.00

9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 time (min) Figure 2-23. HPLC traces of metabolite extracts of (A) KdesVII/pCM18 and (B) KdesVII/pCM17 showing formation of methymycin (30) and neomethymycin (37) after growth for 60 h.

4. CONCLUSIONS

The results obtained in these studies provide compelling support for the functional assignments of TylM3 and MydC as activator proteins for their respective GTs, TylM2 and MycB, and provides further support for the notion that, in general, DesVIII and its homologues act as GT activator proteins in natural product pathways. A recent search of the Protein Databank using the DesVIII sequence as the query identified a total of 20 DesVIII homologues in the databank, all of which occur in either macrolide or anthracycline antibiotic biosynthesis pathways. The knowledge that these activator proteins are important for reconstituting the activity of their partner GTs is critical to the use of these GTs in combinatorial biosynthetic applications. This work also explores the substrate specificity of TylM2 and MycB. TylM2 was shown to be capable of employing the alternate donor substrate 41 and alternate acceptor substrate 124, and MycB was shown to accept alternate acceptors 50 and 124. This flexibility resulted in the production of the engineered macrolide compounds 58, 130, 131, 132, 135, and 136. Interestingly, in studies of both the TylM2/TylM3 and MycB/MydC pairs, it was 94 observed that the activator protein has a stimulatory effect on GT activity, but that it is not absolutely required. There are two plausible explanations for this phenomenon, one or both of which could be correct. First, it is possible simply that TylM3 and MydC are not absolutely required for the activity of their respective partner GTs. Although in vivo studies suggest that DesVIII is absolutely required for DesVII activity, as no glycosylated

compounds are observed in an S. venezuelae mutant in which desVIII was disrupted,107 a low level of DesVII activity could be detected in vitro in the absence of DesVIII, suggesting that the requirement is not absolute.116 Similar to the in vivo results of

DesVIII, disruption of tylM3 in S. fradiae resulted in no detectable production of 51.109 Recent studies of the natural product GT AknS, which requires the DesVIII homologue AknT for full activity in vitro, show that AknS possesses low level activity even in the absence of AknT.71 The activity of the macrolide GT EryCIII, which is also enhanced by the presence of DesVIII homologue EryCII in vitro, displays even less of a dependence on the presence of its activator protein.126 Therefore it is plausible that although these GTs have evolved to depend on their respective activator proteins for full activity, each GT may display a different level of tolerance for the absence of its activator protein. A second possible explanation for this phenomenon is that DesVIII, which is present in the engineered S. venezuelae strains used in these experiments, could partially substitute for TylM3 or MydC. Recent work exploring the ability of several DesVIII homologues to

substitute for DesVIII in vivo revealed that one of these proteins, DnrQ, is capable of restoring partial DesVII activity.116 Also, AknT was shown in vitro to enhance EryCIII activity, although not as well as EryCII.126 These results show that it is possible for an activator protein from one biosynthetic pathway to partially substitute for the natural activator protein in another pathway, and imply that DesVIII could be acting in this manner in mutants lacking TylM3 and MydC. Further experiments are required to better

95 address this issue. The exact functions of DesVIII and its homologues are not yet known, but several proposals have been put forth. The first suggestion was that these proteins function as chaperones which ensure proper folding of their partner GTs.107 Another idea was that they act as regulatory subunits which bind to and reversibly activate their partner

GTs.71 Recently it was demonstrated that EryCIII can be activated by either co- expression or co-incubation with EryCII, and retains activity even after EryCII has been almost completely removed. This suggests that activator proteins enhance GT activity by inducing a conformational change of the GT from an inactive to an active

conformation.126 This is certainly plausible, but further studies are needed to clarify the functions of these proteins.

96 Chapter 3. Discovery of the TDP-4-Keto-6-Deoxy-D-Glucose 3,4- Ketoisomerase Tyl1a from Streptomyces fradiae and its Use for In Vivo Reconstitution of the D-Mycaminose Biosynthetic Pathway in Streptomyces venezuelae

1. INTRODUCTION

As described in Chapters 1 and 2, deoxyaminosugars, such as 3-N,N- dimethylamino-3,6-dideoxy-D-glucose (D-mycaminose, 49), are present in many biologically active secondary metabolite structures, and are frequently critical for the bioactivity of the parent compounds.5, 7, 127 Because of the importance of deoxysugars in natural product bioactivity, much effort has been given to the study of their biosynthesis. This work is done in part in the hope that the information gained can be used to engineer these biosynthetic pathways to produce specific sugar structures which can be coupled to natural product scaffolds by GTs.

D-Mycaminose (49) is a component of several structurally related macrolide antibiotics, one of which is tylosin (51) produced by S. fradiae. In this organism, 49 is known to be produced in a TDP-activated form, TDP-D-mycaminose (52), and is coupled to the 16-membered aglycon tylactone (50) by the GT TylM2. The entire tylosin biosynthetic gene cluster has previously been identified and sequenced.43, 128 Analysis of this gene cluster allowed initial speculation that 52 was biosynthesized from glucose-1- phosphate (12) in five steps (Figure 3-1).44, 129 The first two steps, thymidylylation of 12 by TylA1 to form TDP-D-glucose (13), and 4,6-dehydration of 13 by TylA2 to form

TDP-4-keto-6-deoxy-D-glucose (14) were predicted based on high sequence identity between these enzymes and the well-characterized homologues from other pathways. The final two steps of the pathway were proposed to be catalyzed by the aminotransferase TylB and N,N-dimethyltransferase TylM1, respectively. In vitro studies of TylB and 97 O Me

Me Me

Me Me O O O OH Me Me Me Me CHO O O OH HO Me Me NMe Me NMe2 O 2 HO OH HO O 50 TylM2 OMe OH Me Me O Me OMe Me O Me O O O O 49 Me OH Me O OH O OH O O Me HO Me Me Me2N 58 51 OH 52 OTDP

TylM1 OH OH Me Me O Me O O TylM3 TylA1 O HO TylB HO O TylA2 HO O HO H2N HO HO HO OH O OH OH OTDP OTDP OH OH = 53 38 14 OTDP 13 OTDP 12 OPO3

Figure 3-1

TylM1 demonstrated that they indeed catalyze the conversion of TDP-3-keto-6-deoxy-D-

glucose (38) to TDP-3-amino-3,6-dideoxy-D-glucose (53) and the conversion of 53 to 52, respectively.44, 46 However, the process by which 14 is isomerized to 38 was not understood. An observation that 14 spontaneously isomerizes to 38 by passage over Dowex-1 anion exchange resin led to the suggestion that this process occurred non- enzymatically in vivo.47, 48 However, in vitro studies of TylB in which the enzyme was assayed in the reverse direction in the presence of α-ketoglutarate failed to produce any 14 from 38, suggesting that interconversion of these two compounds must be enzyme- catalyzed.46 This function was tentatively assigned to TylM3, a protein of unknown function with sequence homology to P-450 enzymes, but lacks the conserved cysteine

residue which coordinates the heme iron. Based on the available information on the biosynthesis and attachment of 49, genetic manipulation aimed at the engineered biosynthesis of 52 in Streptomyces venezuelae was initiated in order to test the abilities of TylM2 and the macrolide GT DesVII from S. venezuelae to catalyze formation of hybrid compounds using substrates from both pathways. As was discussed previously, S. venezuelae produces macrolide compounds methymycin (30), neomethymycin (37), and pikromycin (32) bearing the 98 deoxysugar D-desosamine (29). The enzymes responsible for the biosynthesis of these compounds, in particular those involved in formation of the sugar moiety, have been

well-characterized. TDP-D-desosamine (41), the activated precursor for the glycosyltransfer reaction, is biosynthesized from glucose-1-phosphate (12) in six enzymatic steps as shown in Figure 3-2. Compound 12 is thymidylylated by DesIII to

form TDP-D-glucose (13) which is then converted to TDP-4-keto-6-deoxy-D-glucose (14) by the 4,6-dehydratase DesIV. Compound 14, which is the last intermediate common to both desosamine and mycaminose pathways, is then converted to TDP-4-

amino-4,6-dideoxy-D-glucose (48) by DesI. Subsequent oxidative deamination by DesII, C-3 aminotransfer by DesV, and 3-N,N-diemthyltransfer by DesVI complete the formation of 41. The GT DesVII then transfers the desosamine moiety from 41 to the 12- membered macrolactone 10-deoxymethynolide (124) and the 14-membered macrolactone narbonolide (125) to form YC-17 (126) and narbomycin (31), which are hydroxylated by the P450 enzyme PikC to give the final compounds 30, 37, and 32. Methods for genetic OH OH Me O Me O O O DesIII O H2N DesI DesIVHO HO HO HO HO HO HO HO HO HO OTDP OTDP OTDP OPO = 48 14 13 12 3

DesII O O O Me Me Me Me R1 O DesVII Me Me Me Me NMe Me 2 Me NMe2 O DesVIII O O HO PikC HO HO O R2 O OTDP O OH O O Me O O O Me 39 Me Me Me Me Me Me 124 126 30 R1 = OH, R2 = H DesV 37 R1 = H, R2 = OH Me Me O O DesVI H2N Me2N HO HO OTDP OTDP 40 41 O O O Me Me Me HO Me Me Me Me NMe Me Me 2 PikC NMe2 Me Me HO HO O Me O OH DesVII O O Me O O O Me Me Me Me O O DesVIII O O O O Me Me Me 125 31 32 Figure 3-2 99 manipulation of S. venezuelae are well-established, yields of macrolide compounds are high, and compound production time is short compared to other Actinomycetes, making S. venezuelae a good choice for engineering work. Engineered S. venezuelae mutants which could test the abilities of TylM2 to glycosylate endogenous aglycon 124 and the ability of DesVII to use 52 as a donor were designed. The first of these, which tests the ability of TylM2 to transfer 49 from its native substrate 52 to the 12-membered macrolide 124, was constructed using the previously made KdesI-80 mutant41 as the starting point. First, the desVII gene was

disrupted in this mutant, eliminating the endogenous macrolide GT activity. This mutant accumulates sugar intermediate 14, which would be converted to 52 by heterologously expressed mycaminose biosynthetic enzymes TylM3, TylB, and TylM1. Heterologously expressed TylM2 could then be tested for the ability to accept the alternate acceptor substrate 124. Construction of the KdesI/VII double mutant followed by expression of plasmid pCM7b containing genes tylM3, tylM2, tylM1, tylB in this mutant failed to produce any glycosylated compounds, suggesting that TylM2 was incapable of accepting 124 as a substrate. As a control to check for gene expression, the natural aglycon substrate of TylM2, tylactone (50), was fed to this mutant, resulting in formation of a glycosylated tylactone derivative presumed to be 5-O-mycaminosyl tylactone (58). However, mass spectrometric analysis of this compound revealed that it was not 58, but rather was the novel compound quinovosyl tylactone (137) (Figure 3-3, left). In light of this result, which suggested that the isomerization of 14 to 38 might not be occurring in this mutant, as well as concurrent in vitro work on the TylM3 homologue DesVIII showed that it was an activator protein for the GT DesVII, the mycaminose biosynthetic pathway was reconsidered. Careful analysis of the tylosin gene cluster revealed a

100 previously overlooked orphan ORF, 1a, with homology to a recently characterized hexose 3,4-isomerase.130 Subsequent expression of a new plasmid pCM21, containing 1a in addition to tylM3, tylM2, tylM1, and tylB, in KdesI/VII resulted in efficient conversion of 50 to 58, supporting the proposed role of the enzyme encoded by 1a (hereafter referred to as tyl1a) as the hexose 3,4-isomerase operative in the mycaminose biosynthetic pathway (Figure 3- 3, right). With the requisite genes for reconstitution of the mycaminose pathway identified, a second set of experiments designed to test the ability of DesVII to transfer 49 from 52 to its native aglycon substrate 124 were carried out. This work showed that DesVII is indeed capable of efficient transfer of 49 to 124, generating the novel

OH OH O Me O DesIII O HO HO O DesIV HO HO HO OH = OH OH 12 OPO3 13 OTDP 14 OTDP

KdesI/VII S. venezuelae KdesI/VII S. venezuelae pCM7b (tylM3, tylM2, tylM1, tylB) pCM21 (tylM3, tylM2, tylM1, tyl1a, tylB)

Tyl1a

O Me Me O HO O HO TylB OH O O OH OTDP OTDP Me 14 Me Me 38 HO O Me H2N OH OTDP Me Me 53 Me O OH Me HO O HO O O OH TylM1 HO Me2N OH Me 50 OH OTDP 52 OTDP

TylM2 TylM2 TylM3 TylM3 O O Me Me Me Me Me OH Me HO OH NMe2 Me Me O Me HO OH O O Me Me O Me O O O OH O OH Me 137 Me 58 Figure 3-3 101 macrolide derivatives bearing 49. Other constructs were designed to test whether endogenous S. venezuelae homologues of the aminotransferase TylB and N,N- dimethyltransferase TylM1, DesV and DesVI, respectively, were capable of functioning in the mycaminose pathway. Expression of these constructs in the appropriate S. venezuelae mutant revealed that DesV can replace TylB and that DesVI can replace

TylM1, demonstrating in the process that expression of Tyl1a alone can convert the D-

desosamine pathway to a D-mycaminose pathway and thus providing strong evidence for the function of Tyl1a as the 3,4-ketoisomerase converting 14 to 38. These studies have important implications for the construction of natural products bearing 49 as well as for the use of GTs DesVII and TylM2 in combinatorial applications. Tyl1a is the first example to be identified of a hexose ketoisomerase participating in natural product biosynthesis, and is only the second characterized member of this new class of enzymes.

2. EXPERIMENTAL PROCEDURES

General. Most materials used for work described in this chapter have already been mentioned in the Experimental Procedures section of Chapter 2. Klenow (-) Exo enzyme was purchased from USB (Cleveland, OH) and used in “fill-in” reaction according to the manufacturer’s protocol. QuikChange Multi-Site Directed Mutagenesis Kit was purchased from Stratagene (La Jolla, CA). Plasmids and Vectors. All plasmids and vectors used for work described in this chapter have been mentioned in the Experimental Procedures section of Chapter 2. pLZ4 was used as the template for PCR amplification of all fragments of derived from S. venezuelae, pHJL309 was used as the template for amplification of tyl1a and tylB, and pSET552 was used as the template for amplification of tylM3, tylM2, and tylM1.

102 Bacterial Strains. All bacterial strains used for work described in this chapter have been mentioned in the Experimental Procedures section of Chapter 2. Instrumentation. All instrumentation used for work described in this chapter has been mentioned in the Experimental Procedures section of Chapter 2. Preparation of Competent Cells. The procedure used to prepare E. coli competent cells was described in the Experimental Procedures section of Chapter 2. PCR Amplification of DNA. The design of oligonucleotide primers and the general procedure for PCR amplification of DNA fragments was described in the

Experimental Procedures section of Chapter 2. Construction of desVII Disruption Plasmid pDesVII-K2. The construction of plasmid pDesVII-K2, which was used in the disruption of desVII in KdesI-80 to give the KdesI/VII double mutant was described in the Experimental Procedures section of Chapter 2. Construction of Expression Plasmids - pCM1, pCM1b, and pCM1d. The construction of expression plasmids pCM1 and pCM1b, which are modified versions of pAX617 designed to provide additional restriction sites in the polycloning site (PCS) to facilitate direct cloning of two or more inserts, were described in Chapter 2. pAX617 was further modified in order to increase its general utility to give pCM1d, which contains six unique restriction sites in the PCS for cloning. While construction of most of the plasmids described in this chapter did not require the use of restriction sites made useable in pCM1b and pCM1d, these two vectors were created primarily to facilitate the construction of other plasmids which were made concurrently with those in this chapter, but are not described here. pAX617 contains a single BglII site and a single NdeI site outside the PCS. These sites occur rarely in Streptomyces DNA, and thus would be useful in cloning. Previously

103 in the construction of pCM1b, the BamHI site of desVII/pAX617 was removed, resulting in a plasmid designated desVII/pAX617(-)BamHI. This plasmid was used as the starting point for further modification eventually resulting in creation of pCM1d, which lacks BglII and NdeI sites outside the PCS. First, desVII/pAX617(-)BamHI was modified by digestion with BglII, treatment with Klenow fragment, and blunt end ligation, resulting in deletion of the BglII site. The product of this step was designated pre-pCM1c. This process was repeated with the NdeI site of pre-pCM1c to give a product lacking both BglII and NdeI sites, designated pre-pCM1d. Next, the PCS of pUC119 was mutagenized

in order to insert BglII and NdeI sites into it. This was accomplished using two-site simultaneous site-directed mutagenesis as described in the Stratagene QuikChange Multi- Site Directed Mutagenesis Kit. Primers pUC119PCS-SacI-to-NdeI-2, 5’-CGGCCAGTG- AATTCCATATGGGTACCCGGGGATG-3’ and pUC119PCS-SalI-to-BglII-2 5’-GGA- TCTAGAAGATCTCTGCAGGCATGC-3’, which bind to the same strand of pUC119, were used in the mutagenic PCR reaction. Introduced BglII and NdeI sites are shown in bold in the primers. Screening transformants by individual digestion with BglII and NdeI

revealed a single clone with both sites present. This clone, designated pUC119+, was used as the template for amplification of a 158 bp fragment containing the mutagenized PCS with EcoRI and SpeI ends using start primer pUC119PCS-SacI-to-NdeI-2, which was used in the mutagenesis reaction described above, and halt primer 119PCS2-S-down,

5'-GGCCACTAGTTTATGCTTCCGGCTCGTATG-3'. The EcoRI site in pUC119PCS- SacI-to-NdeI-2 is shown in italics, and the engineered SpeI site in 119PCS2-S-down is shown in bold. The digested EcoRI-SpeI fragment was ligated into pre-pCM1d digested with EcoRI and XbaI to give pCM1d. As in construction of pCM1 and pCM1b, ligation of XbaI and SpeI ends gives a non-functional site. pCM1d has a PCS with six useable sites in the order 5’-EcoRI-NdeI-BamHI-XbaI-BglII-HindIII-3’.

104 pCM7b. All fragments used for construction of expression plasmids pCM7b, pCM21, pCM23, pCM24, and pCM30 were amplified containing their native Shine- Dalgarno sequences. pCM7b was constructed by amplifying 3359 bp fragment containing tylM3, tylM2, and tylM1 with introduced upstream EcoRI and downstream XbaI restriction sites using the start primer tylM3-2-E-up and halt primer tylM1-X-down, which were also used in amplification of the tylM3-tylM2-tylM1-containing insert of pCM21 in Chapter 2. Also amplified was a 1226 bp fragment containing tylB with introduced upstream XbaI and downstream HindIII restriction sites using the start primer tylB-X-up, 5'-GGCCTCTAGAGGACTACGACGAGTTCCTGA-3' and halt primer tylB-H-down, which was used in amplification of the tyl1a-tylB-containing fragment of pCM21. The XbaI restriction site in tylB-X-up is shown in bold. These two fragments were digested with the appropriate restriction enzymes and sequentially cloned into pCM1 digested with the same restriction enzymes to give pCM7b (Figure 3-4). pCM21. The construction of pCM21 was described in the Experimental Procedures section of Chapter 2. pCM23. pCM23 was constructed by amplifying an 861 bp fragment containing tylM1 with introduced upstream BamHI and downstream XbaI restriction sites using the start primer tylM1-B-up, 5’-GGCCGGATCCACCGCACTCCACAGGAAC-3’ and halt primer tylM1-X-down, which was also used in amplification of the tylM3-tylM2-tylM1- containing insert of pCM21 in Chapter 2 (where the BamHI restriction site in tylM1-B-up is shown in bold). Also amplified was a 1653 bp fragment containing tyl1a and tylB with introduced upstream XbaI and downstream HindIII restriction sites using the start primer tyl1a-X-up and halt primer tylB-H-down, which were also used in the amplification of the same tyl1a-tylB-containing insert of pCM21. These two fragments were digested

105 with the appropriate restriction enzymes and sequentially cloned into pCM1b digested with the same restriction enzymes to give pCM23 (Figure 3-4).

pCM24. pCM24 was constructed by amplifying a 1653 bp fragment containing tyl1a and tylB with introduced upstream XbaI and downstream HindIII restriction sites using the start primer tyl1a-X-up and halt primer tylB-H-down, which were also used in the amplification of the same tyl1a-tylB-containing insert of pCM21. This fragment was digested with XbaI and HindIII and ligated into pCM1b digested with the same restriction enzymes to give pCM24 (Figure 3-5). pCM30. pCM30 was constructed by amplifying a 533 bp fragment containing tyl1a with introduced upstream XbaI and downstream HindIII restriction sites using the start primer tyl1a-X-up, which was also used in amplification of the tyl1a-tylB-containing insert of pCM21, and tyl1a-H-down, 5’-GCGCAAGCTTAGATCGTGGAACGGCAC- 3’ (where the HindIII restriction site in tyl1a-H-down is shown in bold). This fragment was digested with XbaI and HindIII and ligated into pCM1d digested with the same restriction enzymes to give pCM30 (Figure 3-5).

106 Construction of KdesI-80 S. venezuelae. KdesI-80 S. venezuelae mutant was constructed by Dr. Svetlana Borisova of this group, and its construction has been

reported.41

Conjugal Transfer of pDesVII-K2 into KdesI-80 S. venezuelae and Screening for Double-Crossover Mutants. These experiments are described in the Experimental Procedures section of Chapter 2.

Conjugal Transfer of Expression Plasmids into KdesI and KdesI/VII S. venezuelae. pCM23, pCM24, and pCM30 were transferred to KdesI-80 S. venezuelae, and pCM7b was transferred to KdesI/VII S. venezuelae. The transfer of pCM7b into

KdesI/VII was performed identically to other conjugal transfer experiments involving transfer of pCM1 derivatives into KdesI/VII which are described in the Experimental Procedures section of Chapter 2. The transfer of pCM23, pCM24, and pCM30 into KdesI-80 were performed similarly to other conjugal transfer experiments described in the Experimental Procedures section of Chapter 2, with the following changes. KdesI-80 S. venezuelae is kanamycin resistant, and thus all KdesI recipient cultures were grown in the presence of kanamycin (50 µg/mL). Also, the antibiotic solution used in overlay of 107 KdesI-80 conjugation plates and recipient negative control plate contained kanamycin (500 µg/mL) in addition to nalidixic acid (500 µg/mL) and apramycin (500 µg/mL), and KdesI-80 recipient positive control plate was overlayed with kanamycin (500 µg/mL) in addition to nalidixic acid (500 µg/mL).

Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains. These experiments were performed as described in the Experimental Procedures section of Chapter 2. Southern Blot Analysis of KdesI/KdesVII Mutant. This experiment was performed

as described in the Experimental Procedures section of Chapter 2. Small-scale Isolation and Analysis of Metabolites Produced by Mutants. These experiments were performed as described in the Experimental Procedures section of Chapter 2. Mass spectrometric data for 5-O-quinovosyl tylactone (137) isolated from

+ KdesI/VII/pCM7b: High resolution CI-MS: C29H49O9 (M + H) calculated 541.3377, found 541.3379. Mass spectrometric data for 5-O-mycaminosyl tylactone (58) and 2’-glucosyl-5- O-mycaminosyl tylactone (132) isolated from KdesI/VII/pCM21 were reported in the Experimental Procedures section of Chapter 2. Mass spectrometric data for 3-O-mycaminosyl 10-deoxymethynolide (138), 3-O-

mycaminosyl neomethynolide (136), and 3-O-mycaminosyl novamethynolide (139)

+ isolated from KdesI/pCM23: High resolution CI-MS (138): C25H44NO7 (M + H)

+ calculated 470.3118, found 470.3109; high resolution CI-MS (136): C25H44NO8 (M + H)

+ calculated 486.3067, found 486.3079; high resolution CI-MS (139): C25H44NO9 (M + H) calculated 502.3016, found 502.2999.

108 Mass spectrometric data for 3-O-mycaminosyl 10-deoxymethynolide (138), 3-O- mycaminosyl methynolide/neomethynolide (135/136), 3-O-mycaminosyl narbonolide (140), and 3-O-mycaminosyl pikronolide (141) isolated from KdesI/pCM30: High

+ resolution CI-MS (138): C25H44NO7 (M + H) calculated 470.3118, found 470.3118, high

+ resolution CI-MS (135/136): C25H44NO8 (M + H) calculated 486.3067, found 486.3088,

+ high resolution CI-MS (140): C28H48NO8 (M + H) calculated 526.3380, found 526.3367,

+ high resolution CI-MS (141): C28H48NO9 (M + H) calculated 542.3329, found 542.3311. Large-Scale Isolation and Analysis of Extracts of KdesI/VII/pCM7b and KdesI/pCM23 S. venezuelae Mutants Grown in Vegetative Media. These experiments were performed nearly identically to that of the large-scale isolation of compounds from the KdesVII/pCM4 mutant described in Chapter 2. A 150 mL culture of KdesI/VII/pCM7b and KdesI/pCM23 were grown in seed media containing the appropriate antibiotics at 29 °C for 48 h with shaking at 250 rpm in a rotary shaker. For each mutant, 20 mL of seed culture was used to inoculate each of six 1 L flasks of vegetative media, which were grown at 29 °C with shaking at 250 rpm for 48 h. Culture broth was obtained by removal of cells and insoluble components as described previously for small-scale isolation. Compounds were extracted from the aqueous layer three times using an equal volume of chloroform each time. After evaporation of solvent, compounds from KdesI/VII/pCM7b were separated by silica gel flash chromatography

using a gradient of 0-20% methanol in chloroform and analyzed by TLC and 1H NMR. Separation of compounds isolated from KdesI/pCM23 was performed by silica gel flash chromatography using a gradient of 0-40% methanol in chloroform, and fractions were analyzed by TLC and 1H NMR. Fractions from silica gel column of KdesI/pCM23 were further separated by preparative scale HPLC on a C18 column by isocratic elution using 30% or 40% acetonitrile in 57 mM aqueous ammonium acetate buffer. Compounds 136,

109 138, and 139 were partially purified and analyzed by 1H NMR. Compound 135 was purified and its structure determined by 1H, 13C, HSQC, HMBC, COSY, and NOESY spectroscopies.

Spectral data for 3-O-mycaminosyl methynolide (135): 1H NMR (500 MHz,

CDCl3) δ 6.59 (1H, d, J = 16.0, 9-H), 6.34 (1H, d, J = 16.0, 8-H), 4.76 (1H, dd, J = 10.9, 2.3, 11-H), 4.28 (1H, d, J = 7.3, 1'-H), 3.65 (1H, dd, J = 10.4, 1.0, 3-H), 3.51 (1H, dd, J = 10.4, 7.3, 2'-H), 3.29 (1H, dq, J = 8.8, 6.1, 5'-H), 3.09 (1H, brt, J = 9.4, 4'-H), 2.81 (1H, dq, J = 10.4, 6.9, 2-H), 2.57 (1H, ddq, J = 12.6, 7.1, 4.1, 6-H), 2.52 (6H, s, N-Me), 2.40

(1H, brt, J = 9.4, 3'-H), 1.94 (1H, ddq, J = 14.1, 7.4, 2.3, 12-H), 1.63 (1H, brt, J = 13.1, 5- H), 1.51 (1H, ddq, J = 14.1, 10.9, 7.4, 12-H), 1.38 (1H, m, 5-H), 1.38 (3H, d, J = 6.9, 2- Me), 1.37 (3H, s, 10-Me), 1.30 (3H, d, J = 6.1, 5'-Me), 1.25 (1H, m, 4-H), 1.18 (3H, d, J = 7.1, 6-Me), 1.00 (3H, d, J = 6.6, 4-Me), 0.91 (3H, t, J = 7.4, 12-Me). 13C NMR (125

MHz, CDCl3) δ 204.0 (C-7), 174.8 (C-1), 148.6 (C-9), 125.8 (C-8), 104.5 (C-1'), 84.9 (C- 3), 76.6 (C-11), 74.5 (C-10), 73.3 (C-5'), 71.1 (C-2'), 70.9 (C-4'), 70.3 (C-3'), 45.0 (C-6),

44.1 (C-2), 41.7 × 2 (N-Me), 34.0 (C-5), 33.7 (C-4), 21.3 (C-12), 19.6 (C-10-Me), 17.9 (C-5'-Me), 17.7 (C-6-Me), 17.4 (C-4-Me), 16.4 (C-2-Me), 10.7 (C-12-Me). High

+ resolution CI-MS: C25H43NO8 (M + H) calculated 486.3067, found 486.3070. Large-Scale Isolation and Analysis of Extracts of KdesI/VII/pCM7b Mutant Grown in PGM media. The major macrolide products of S. venezuelae grown in PGM

media are 14-membered macrolides. PGM media is prepared by combining 10 g of

glucose, 10 g of glycerol, 10 g of peptone, 5 g of beef extracts, 5 g of NaCl, 2 g of CaCO3 in 1 L of tap water, adjusting the pH to 7.3 using NaOH, and autoclaving. A 150 mL culture of KdesI/VII/pCM7b was grown in seed media containing the appropriate antibiotics at 29 °C for 48 h with shaking at 250 rpm in a rotary shaker. A 20 mL aliquot of the seed culture was used to inoculate each of six 1 L flasks of PGM media, which

110 were grown at 29 °C with shaking at 250 rpm for 4 days. Procedures for obtaining culture broth, extraction, separation, and analysis of compounds are identical to those used for large-scale isolation and analysis of compounds obtained from S. venezuelae grown in vegetative media.

3. RESULTS AND DISCUSSION

Design, Construction, and Analysis of KdesI/VII/pCM7b S. venezuelae. In order to construct an S. venezuelae mutant suitable for heterologous expression of the enzymes involved in D-mycaminose (49) biosynthesis and attachment, the previously constructed KdesI mutant41 was used as the starting point for creation of a KdesI/KdesVII double mutant. With both desI and desVII genes disrupted in the mutant, the desosamine (29) biosynthetic pathway would stop at TDP-4-keto-6-deoxy-D-glucose (14), and no glycosylated product is expected due to the absence of DesVII GT. An expression plasmid, pCM7b (Figure 3-4), carrying all the proposed genes involved in 49 biosynthesis and attachment, tylM3, tylM2, tylM1 and tylB, was constructed and subsequently introduced into the KdesI/VII mutant. It was expected that this mutant would produce significant amounts of TDP-D-mycaminose (52), the natural donor substrate of GT TylM2. Thus, the ability of TylM2 to recognize the endogenous macrolactone 10-deoxymethynolide (124) would be tested. If TylM2 were capable of using 124 as an acceptor substrate, methymycin/neomethymycin (30/37) derivatives bearing D-mycaminose (49) in place of D-desosamine (29) would be generated. TLC analysis of small-scale extracts from five colonies with the KdesI/VII/pCM7b genotype failed to show any obvious glycosylated macrolide compounds. Therefore, analysis of the compounds made by growth of a large-scale (6 L) culture of one KdesI/VII/pCM7b

111 colony was performed in order to identify any minor glycosylated compounds that might be present. However, this analysis revealed only non-glycosylated macrolides, primarily 10-deoxymethynolide (124), with small amounts of methynolide (127), and neomethynolide (128) also present (Figure 3-6). This result suggested that TylM2 was incapable of transfer of 49 to 124. Reasoning that the 12-membered macrolactone 124 may be structurally too different from the 16-membered ring macrolactone tylactone (50) to be accepted by TylM2, the extracts from a large-scale (6 L) culture of the KdesI/VII/pCM7b mutant

grown in PGM media, which favors production of the 14-membered ring macrolactone narbonolide (125) were isolated and analyzed. However, this analysis again revealed production of only non-glycosylated macrolides, primarily 125, with small amounts of pikronolide (142) (Figure 3-6). This suggests that TylM2 is also not capable of glycosylating 125. An alternative explanation for the results of large-scale analysis of the KdesI/VII/pCM7b mutant might be that one or more of the genes on pCM7b was not expressed. To test this possibility, a control experiment in which tylactone (50) was fed KdesI/VII/pCM7b to a small culture of O O Me Me KdesI/VII/pCM7b was performed. R1 R Me Me Me Me If all genes for biosynthesis and O Me R2 attachment of 49 were expressed O OH O OH Me Me Me O O from pCM7b, one would expect 124 R = H, R = H Me 1 2 efficient conversion of 50 to 5-O- 127 R1 = OH, R2 = H 125 R = H, 128 R1 = H, R2 = OH 142 R = OH mycaminosyl tylactone (58). TLC vegitative media PGM media analysis of extracts from the Figure 3-6 KdesI/VII/pCM7b culture fed 50

112 revealed formation of a new polar spot with the characteristic maroon color of a derivative of 50 that was absent in extracts of KdesI/VII/pCM7b grown in the absence of 50 (Figure 3-7). However, high-resolution mass spectrometric analysis of this compound isolated by preparative TLC revealed that it was not 58, but rather was quinovosyl tylactone (137) (Figure 3-8). Another quinovosylated compound, quinovosyl 10- deoxymethynolide (143), was previously isolated from the KdesI-80 S. venezuelae mutant.41 Compound 143 was thought to be formed by stereospecific C-4 ketoreduction

of TDP-4-keto-6-deoxy-D-glucose (14), which accumulates in the KdesI-80 mutant, by

an unidentified endogenous ketoreductase, to form TDP-D-quinovose (144). This

compound was then used as a donor for the transfer of D-quinovose to 124 by DesVII, forming 143 (Figure 3-8). The result obtained from the tylactone feeding experiment supports the idea that TylM2 is functional in this mutant and that it is capable of

Figure 3-7. Metabolites produced by KdesI/VII/pCM7b in the presence and absence of exogenous 50. 1 = KdesI/VII/pCM7b grown in vegetative media, 2 = KdesI/VII/pCM7b grown in vegetative media in the presence of 50. 113 OH OH O DesIII O HO HO HO HO HO HO OPO = OTDP 12 3 13 DesIV

O Me O O KdesI/VII/pCM7b HO O KdesI-80 Me HO Me OTDP Me Me 14 Me Me reductase O Me Me O OH Me O OH O OH HO O Me Me Me HO 124 50 HO O OTDP Me 144 O Me TylM2 DesVII Me Me OH TylM3 DesVIII HO OH Me Me OH Me Me O Me O O O HO OH O O O Me O OH Me Me 145 Me 137 143 Figure 3-8 accepting the alternate substrate 144. By analogy to the formation of 143 by KdesI-80, the conversion of 50 to 137 byKdesI/VII/pCM7b suggested that the mycaminose pathway might be halted at an early step, such as ketoisomerization, causing accumulation of 14 which was reduced to 144 and used by TylM2 to glycosylate 50. Importantly, the activity of the proposed ketoisomerase TylM3 had not yet been directly tested, making its functional assignment questionable. Interestingly, work done by Dr. Svetlana Borisova of our group that was concurrent to studies of KdesI/VII/pCM7b established that DesVIII, which is a homologue of TylM3 (33% sequence identity), acts as an activator protein for the GT DesVII.107 Taken together, analysis of KdesI/VII/pCM7b and Dr. Borisova’s results on the function of DesVIII forced reconsideration of which enzyme might be catalyzing ketoisomerization in the mycaminose pathway.

114 Identification of ORF 1a and Reconstitution of the Mycaminose Pathway. The tylosin biosynthetic gene cluster, which has been fully sequenced,128 contains several ORFs with unassigned function. In order to explore the possibility that one of these genes might encode a sugar ketoisomerase, each ORF in the gene cluster whose function had not been assigned was used as a query in a BLAST search of the Protein Databank. Incredibly, a small 440 bp ORF, 1a, (Figure 3-9) which had not been assigned as an ORF in the original gene sequencing work, encoded a putative protein which showed homology to a recently characterized enzyme, FdtA that was shown to catalyze the

conversion of 14 to TDP-3-keto-6-deoxy-D-galactose, the C-4 epimer of the TylB substrate 38.130 FdtA participates in the formation of TDP-3-acetamido-6-deoxy-D- galactose as part of S-layer polysaccharide biosynthesis in the thermophilic Bacillus Aneurinibacillus thermoaerophilus. Publication of this report, which was the first to identify an enzyme having hexose 3,4-ketoisomerase activity, occurred in June of 2003, six months prior to the observation that KdesI/VII/pCM7b can convert 50 to 143. ORF 1a was found directly upstream of and translationally coupled to the aminotransferase tylB in the tylosin gene cluster, suggesting a functional link between the two genes. This exciting discovery of ORF 1a in the tylosin gene cluster required testing of a new hypothesis that ORF 1a encodes a ketoisomerase responsible for conversion of 14 to 38 in the mycaminose biosynthetic pathway.

Figure 3-9. The tylIBA region of the tylosin biosynthetic gene cluster in Streptomyces fradiae showing the location of ORF 1a.

115 In order to test the proposed function of ORF 1a as the missing ketoisomerase in the mycaminose pathway, expression plasmid pCM21 (Figure 2-8), containing ORF 1a in addition to tylM3, tylM2, tylM1, and tylB, was constructed and expressed in the KdesI/VII S. venezuelae mutant. When a small-scale culture of KdesI/VII/pCM21 was grown in the presence of 50, two new polar spots were detected (Figure 3-10) which were isolated by preparative TLC and determined by high-resolution mass spectrometry to have elemental compositions consistent with 5-O-mycaminosyl tylactone (58) and 2’-glucosyl-5-O- mycaminosyl tylactone (130). Compound 130 is believed to derive from 58 by the

endogenous resistance GT DesG as part of an antibiotic self-protection mechanism

(Figure 3-11).118 This result strongly suggests that ORF 1a, hereafter referred to as tyl1a, encodes a sugar ketoisomerase which converts 14 to 38 in the mycaminose biosynthetic pathway, and shows that the mycaminose pathway can be reconstituted in KdesI/VII by the introduction of pCM21.

Figure 3-10. TLC analysis of the metabolites produced by the KdesI/VII/pCM21 mutant grown in the presence of 50.

116 OH OH O Me Me O DesIII O O HO HO O DesIV Tyl1a HO HO HO HO OH = OH O OH OH OTDP 12 OPO3 13 OTDP 14 OTDP 38

O TylB Me Me TylM2 Me Me Me O O NMe2 TylM3 HO TylM1 HO HO OH Me N Me Me O 2 H2N Me OH OH O O OTDP 52 53 OTDP O OH O Me Me 58 Me Me DesG Me Me O OH O O OH Me Me Me Me Me OH Me Me O O O O NMe2 O OH HO O Me OH

HOHO

132 Figure 3-11

Design, Construction, and Analysis of KdesI/pCM23 to test DesVII Donor Substrate Flexibility. After identification of tyl1a and reconstitution of the mycaminose biosynthetic pathway, it was then possible to create a construct to test the ability of

DesVII to transfer D-mycaminose (49) from 52 to 124 in vivo. pCM23 (Figure 3-4), which contains mycaminose biosynthetic genes tyl1a, tylB, and tylM1, was constructed for this purpose. Transfer of pCM23 to the KdesI S. venezuelae mutant should result in formation of 52 which can potentially be used by DesVII. Construction of KdesI/pCM23 and subsequent TLC analysis of metabolites isolated from a small-scale culture of this mutant clearly revealed the presence of three prominent new polar compounds with the characteristic colors of derivatives of 10-deoxymethynolide (124), methynolide (127), and neomethynolide (128) (Figure 3-12).

117 Figure 3-12. TLC analysis of metabolites produced by KdesI/pCM23.

In order to characterize these compounds in more detail, a large-scale (6 L) culture of KdesI/pCM23 was grown in vegetative media and the products were extracted from the culture broth, separated by silica gel chromatography and HPLC, and analyzed by TLC, mass spectrometry, and NMR. Seven distinct glycosylated macrolides were observed. The predominant compound (280 mg) was purified and shown by detailed structural characterization using 1H, 13C, and various 2D NMR techniques to be 3-O- mycaminosyl methynolide (135). Three other compounds were identified by partial 1H NMR characterization and high resolution mass spectrometry as 3-O-mycaminosyl 10- deoxymethynolide (138, 20 mg), 3-O-mycaminosyl neomethynolide (136, 180 mg), and 3-O-mycaminosyl novamethynolide (139, 35 mg) (Figure 3-13). Three other glycosylated macrolide compounds present in minute quantities were too little to be identified. Formation of 12-membered macrolides which are hydroxylated at both C-10

and C-12, as is 139, by wild-type S. venezuelae, was only recently reported.131 Glycosylated macrolide compounds accounted for about 70% of total macrolides

118 produced in KdesI/pCM23, indicating that DesVII efficiently catalyzes attachment of 49 to 124.

OH OH O Me Me O DesIII O O HO HO O DesIV Tyl1a HO HO HO HO OH = OH O OH OH OTDP 12 OPO3 13 OTDP 14 OTDP 38

O TylB Me R 1 DesVII Me Me Me Me NMe DesVIII HO O TylM1 HO O O 2 OH Me N R HO 2 H2N 2 O OH OH O O Me OTDP 52 53 OTDP Me Me 49 O Me 138 R1 = H, R2 = H 135 R1 = OH, R2 = H Me Me O 136 R1 = H, R2 = OH 139 R1 = OH, R2 = OH O OH Me Me 124

Figure 3-13

Design, Construction, and Analysis of KdesI/pCM24 and KdesI/pCM30. Having shown that DesVII is capable of efficient transfer of 49 to endogenous macrolactone 124, it was now possible to test whether desosamine biosynthetic enzymes DesV and DesVI could substitute for mycaminose biosynthetic enzymes TylB and TylM1, respectively. The substrates of TylB and TylM1 differ from those of DesV and DesVI by the addition of the C-4 hydroxyl group. A previous experiment in which a desV disruption mutant of S. venezuelae was transformed with a plasmid containing tylB showed that tylB could restore efficient production of wild-type macrolides 30 and 37, demonstrating that TylB is capable of accepting a 4-deoxysugar analogue of its natural substrate.132 A similar complementation experiment in which tylM1 was introduced into a desVI disruption mutant also showed that tylM1 can efficiently accept a 4-deoxy substrate analogue.132 In vitro experiments with TylM1 mirrored in vivo studies of TylM1, showing that the

TylM1 was capable of turnover of the DesVI substrate with a comparable kcat value as its 119 natural substrate (7.2 min-1 for DesVI substrate compared to 9.9 min-1 for natural substrate). However, studies of the ability of DesVI to accept the TylM1 substrate

revealed a 22-fold reduction in kcat for the 4-hydroxy substrate analogue compared to the natural substrate (4.2 min-1 compared to 92 min-1).125 Therefore it was not clear whether DesV or DesVI would be competent in a mycaminose biosynthetic pathway in vivo. To test the abilities of DesV and DesVI to participate in the engineered mycaminose pathways in which they are used to replace TylB and TylM1, respectively, two constructs were made. The first, pCM24, contained tyl1a and tylB which, when OH OH O Me O DesIII O HO HO O DesIV HO HO HO OH = OH OH 12 OPO3 13 OTDP 14 OTDP

KdesI/pCM24 KdesI/pCM30 tyl1a, tylB tyl1a Tyl1a Tyl1a

Me O Me O HO Me HO O

O OH Me OTDP Me O OH 38 O 38 OTDP O OH TylB Me Me DesV 124 Me Me HO O O O HO H N 2 H2N OH Me OH 53 OTDP OTDP Me Me 53

DesVI Me DesVI O OH Me Me Me HO O O O HO O Me Me2N Me2N OH OH 52 OTDP 125 52 OTDP

DesVII DesVII DesVIII DesVIII O O O Me Me Me R1 R R1 Me Me Me Me NMe NMe2 Me Me NMe 2 2 O HO OH O HO OH Me HO OH R2 O O R2 O O Me O O Me O O O Me Me Me 49 Me 49 Me Me 49 O O 138 R1 = H, R2 = H Me 135 R1 = OH, R2 = H 135 R = OH, R = H 1 2 140 R = H 136 R1 = H, R2 = OH 136 R = H, R = OH 1 2 141 R = OH Figure 3-14

120 introduced into KdesI S. venezuelae, would test the ability of DesVI to catalyze the final step of the mycaminose pathway, namely conversion of 53 to 52 (Figure 3-14). After introduction of pCM24 into KdesI-80 by conjugal transfer, TLC analysis of the metabolites produced by a small scale culture of this mutant showed that it produces comparable quantities of 135 and 136 to the KdesI/pCM23 mutant (Figure 3-15), demonstrating that DesVI can function efficiently in place of TylM1 in the mycaminose pathway. Interestingly, the metabolite profiles of KdesI/pCM23 and KdesI/pCM24 are noticeably different. While KdesI/pCM23 produces significant quantities of 138,

KdesI/pCM24 produces little or no 138 as judged by TLC. However, the reason for this difference is not clear.

Figure 3-15. TLC analysis of metabolites produced by KdesI/pCM24.

The second construct, pCM30 was designed to simultaneously test the abilities of DesV and DesVI to catalyze the final two steps in mycaminose biosynthesis. pCM30 contains only tyl1a, which converts 14 accumulating in the KdesI mutant to 38. If DesV is capable of converting 38 to 53, DesVI would be expected to convert 53 to 52, as it did in the KdesI/pCM24 mutant, resulting in formation of mycaminosylated macrolides

121 (Figure 3-14). Introduction of pCM30 into KdesI-80 by conjugal transfer followed by TLC analysis of the metabolites produced by the resulting mutant, KdesI/pCM30 (Figure 3-16) showed that this mutant does indeed produce significant quantities of glycosylated compounds of comparable polarity to those observed in extracts of KdesI/pCM23 and KdesI/pCM24. High resolution mass spectrometric analysis of compounds in the polar fraction of crude extracts from KdesI/pCM30 revealed the presence of compounds with elemental composition consistent with 135/136 and 138 as well as mycaminosylated 14- membered ring macrolides mycaminosyl narbonolide (140) and mycaminosyl pikronolide

(141). As judged by TLC, yields of glycosylated compounds in the KdesI/pCM30

Figure 3-16. TLC analysis of metabolites produced by KdesI/pCM30. mutant were comparable to those observed in KdesI/pCM23 and KdesI/pCM24. This result conclusively shows that both DesV and DesVI are capable of functioning efficiently as the surrogate aminotransferase and N,N-dimethyltransferase in the mycaminose biosynthetic pathway, respectively. This demonstration shows that in spite of the significantly reduced in vitro kinetics of DesVI turnover of 53, DesVI in vivo activity is sufficient to lead to efficient formation of mycaminosylated macrolides. Also, 122 results obtained from the KdesI/pCM30 mutant are the first demonstration that DesV is capable of turnover of 38. While information on the kinetics of this process awaits further investigation, this result has implications for future pathway engineering work using aminotransferases DesV and TylB. Finally, results of the analysis of KdesI/pCM30 provide further evidence of the key role of Tyl1a as the ketoisomerase converting 14 to 38 in the mycaminose pathway, demonstrating that expression of Tyl1a in KdesI-80 efficiently converts a non-functional desosamine biosynthetic pathway to a mycaminose pathway.

4. CONCLUSIONS

In summary, the results reported provide strong support for the catalytic role of

Tyl1a as TDP-4-keto-6-deoxy-D-glucose 3,4-isomerase. This finding is significant because it assigns a function to an orphan ORF whose encoded protein catalyzes an important conversion in unusual sugar biosynthesis. Because its homologues likely catalyze similar isomerization reactions in sugar formation, their existence in a gene cluster will shed light on the encoded biosynthetic pathway. With this gene identified, the mycaminose biosynthetic pathway has now been fully elucidated, and a valuable addition to our collection of sugar biosynthetic enzymes which can be used for combinatorial applications has been found. In the process of this work, several new compounds 137-141 were generated, demonstrating the applicability of Tyl1a to the synthesis of new macrolide derivatives. The tolerance of DesV and DesVI for 4-hydroxy substrate analogues were also shown in vivo. Finally, the substitution of Tyl1a for DesI and DesII was shown to be able to convert the desosamine pathway to a mycaminose biosynthesizing pathway. The relative simplicity of this transformation is encouraging

123 for efforts to construct diverse nucleotide sugars for glycodiversification of secondary metabolites.

124 Chapter 4. In vitro Characterization of Tyl1a: Product Identification, Steady-State Kinetic Analysis, and Substrate Specificity Studies

1. INTRODUCTION

As mentioned in previous chapters, deoxysugars, such as D-mycaminose (49), are present in the structures of many secondary metabolites possessing antigenic, antibiotic, and chemotherapeutic properties.133 These unusual sugars have been shown to be important for the biological activities of the parent compounds by functional studies of natural products with altered glycosylation patterns,6, 7 and by structural studies of glycosylated natural products complexed with their targets.134 Due to the direct role of deoxysugars in conferring natural product bioactivity, there has been much interest in elucidating deoxysugar biosynthetic pathways.42, 53, 107, 135-137 Recently, the feasibility of manipulating the sugar biosynthetic machinery to generate new glycosylated natural products has been demonstrated.10, 107, 138, 139 However, to further exploit this strategy, enzymes involved in the formation of a diverse set of deoxysugars must be identified, their activities demonstrated, preferred substrates identified, and tolerance for other substrates assessed. A detailed understanding of the biochemical properties of these enzymes is important because any rational attempt to effectively utilize a specific enzyme

in biosynthetic applications requires an understanding of the details of its catalytic process.

The biosynthesis of D-mycaminose (49) has been studied for more than ten years. This 3-N,N-dimethylamino-3,6-dideoxyhexose is found as a substituent on a number of 16-membered ring macrolide antibiotics.140 It is also the first sugar attached to tylactone (50), a 16-membered macrolactone, in the formation of tylosin (51) in Streptomyces fradiae (Figure 4-1). Extensive genetic and phenotypic complementation studies

125 O Me

Me Me

Me Me O O O OH Me Me Me (TylM2) Me CHO O O OH HO TylM3 Me Me NMe Me NMe2 O 2 TylM2 HO OH HO O 50 OMe OH Me Me O Me OMe Me O Me O O O O 49 Me OH Me O OH O OH O O Me HO Me Me Me2N 58 51 OH 52 OTDP

TylM1 (TylM3) OH OH Me Me O Me O O Tyl1a TylA2 TylA1 O HO TylB HO O HO O HO H2N HO HO HO OH O OH OH OTDP OTDP OH OH = 53 38 14 OTDP 13 OTDP 12 OPO3

HO O Me Me O FdtA O HO OH O OH OTDP OTDP 14 153

Figure 4-1

Figure 4-1. The TDP-D-mycaminose pathway in the biosynthesis of tylosin in Streptomyces fradiae. The original assignments of enzymes catalyzing each step are shown in parentheses, and the revised assignments are shown in boldface. The reaction catalyzed by FdtA from Aneurinibacillus thermoaerophilus is shown in the inset.

revealed the genetic organization of the tylosin (tyl) biosynthetic gene cluster in which the tylG region harbors the polyketide synthase (PKS) genes for making tylactone and the flanking tylLM, tylIBA and tylCK regions contain the genes for unusual sugar formation.43 The tylLM, tylIBA and tylCK regions were sequenced in previous studies, and 17 open reading frames (ORFs) were identified within these regions.44 Sequence similarities with other sugar biosynthetic genes, especially those reported by Cundliffe and coworkers who

had also sequenced the tylIBA and tylLM regions of the tyl cluster,45 led to the assignment of tylA1, tylA2, tylB, tylM1, tylM2, and tylM3 as genes involved in mycaminose formation and attachment. The tylA1, tylA2, and tylM2 genes all show high sequence identity with their well- characterized counterparts in other sugar biosynthetic pathways, and thus were assigned

126 the following functions: tylA1 encodes α-D-glucose-1-phosphate thymidylyltransferase

responsible for the conversion of 12 to 13, tylA2 encodes a TDP-D-glucose 4,6- dehydratase converting 13 to 14, and tylM2 encodes a glycosyltransferase responsible for the attachment of 49 to tylactone (50). The tylB and tylM1 genes encode a pyridoxal 5'- phosphate (PLP)-dependent aminotransferase and an S-adenosylmethionine (SAM)- dependent methyltransferase, respectively. As depicted in Figure 4-1, TylB catalyzes the

C-3 transamination of TDP-3-keto-6-deoxy-D-glucose (38) to form TDP-3-amino-3,6-

dideoxy-D-glucose (53), and TylM1 catalyzes the N,N-dimethylation of 53 to give TDP-

D-mycaminose (52).44, 46 While most of the steps in the proposed mycaminose biosynthetic pathway are supported by sequence alignment data or biochemical evidence, the process by which

TDP-4-keto-6-deoxy-D-glucose (14) isomerizes to 38 remained unknown until recently. On the basis of two early reports in which a portion of 14 was transformed to 38 during purification by Dowex-1 ion exchange chromatography, a reversible non-enzymatic ketoisomerization between 14 and 38 was thought to occur.47, 48 However, in the study of TylB, when the reaction was run in reverse using 53 and α-ketoglutarate as substrates, only 38 was produced and no trace of 14 was detected. Also, no product was formed upon incubation of 14 with TylB. These results indicated that, at least under the in vitro conditions used, there was no chemical isomerization between 14 and 38, suggesting that

the 3,4-ketoisomerization is more likely enzyme-catalyzed.46 This activity was tentatively assigned to the tylM3 gene product, which displays low sequence similarity to P-450 enzymes but lacks the conserved cysteine residue that coordinates the heme iron. An attempt to reconstitute the mycaminose biosynthetic pathway by expression of plasmid pCM7b in the KdesI/VII S. venezuelae mutant, which was described in Chapter 3, showed that TylB, TylM1, TylM2, and TylM3 are not capable of converting 14 to

127 TDP-D-mycaminose (52) and coupling 52 to tylactone (50). Feeding exogenous 50 to KdesI/VII/pCM7b led to formation of quinovosyl tylactone (137) rather than the anticipated mycaminosyl tylactone (58) (Figure 4-2), suggesting that conversion of 14 to 38 did not occur in the recombinant strain. This result led to the identification of an ORF, tyl1a, in the tylosin gene cluster, which exhibits modest sequence homology (34% identity, 52% similarity) to the recently reported TDP-4-keto-6-deoxy-D-glucose-3,4- ketoisomerase FdtA from A. thermoaerophilus.130 Subsequent heterologous expression of tyl1a together with tylB, tylM1, tylM2, and tylM3 in the KdesI/VII S. venezuelae mutant resulted in the quantitative conversion of 50 to 5-O-mycaminosyl-tylactone 58

OH OH O Me O DesIII O HO HO O DesIV HO HO HO OH = OH OH 12 OPO3 13 OTDP 14 OTDP

KdesI/VII S. venezuelae KdesI/VII S. venezuelae pCM7b (tylM3, tylM2, tylM1, tylB) pCM21 (tylM3, tylM2, tylM1, tyl1a, tylB)

Tyl1a

O Me Me O HO O HO TylB OH O O OH OTDP OTDP Me 14 Me Me 38 HO O Me H2N OH OTDP Me Me 53 Me O OH Me HO O HO O O OH TylM1 HO Me2N OH Me 50 OH OTDP 52 OTDP

TylM2 TylM2 TylM3 TylM3 O O Me Me Me Me Me OH Me HO OH NMe2 Me Me O Me HO OH O O Me Me O Me O O O OH O OH Me 137 Me 58 Figure 4-2

128 (Figure 4-2). These findings identified Tyl1a as the TDP-4-keto-6-deoxy-D-glucose-3,4- ketoisomerase in the mycaminose pathway. In this chapter the overexpression, purification, and biochemical characterization of Tyl1a is described. In situ 1H NMR spectroscopic analysis was used to show that Tyl1a converts 14 to 38, which could then be converted to 53 by incubation with the next enzyme in the mycaminose pathway, TylB. These results firmly establish Tyl1a as the 3,4-ketoisomerase in the mycaminose pathway. The steady-state kinetic parameters of the Tyl1a-catalyzed reaction were determined using a discontinuous HPLC-based assay. The substrate specificity of Tyl1a was also tested and it was found that Tyl1a can process

the alternate substrate TDP-4-keto-2,6-dideoxy-D-glucose (145), and can also act on

CDP-4-keto-6-deoxy-D-glucose (146), albeit at a much reduced rate (Figure 4-3). Additionally, we demonstrated that TylB is able to convert the products generated by

Tyl1a using 145 and 146 as substrates to TDP-3-amino-2,3,6-trideoxy-D-glucose (147) and CDP-3-amino-3,6-dideoxy-D-glucose (148), respectively (Figure 4-3). These findings have important implications for deoxysugar pathway engineering efforts and for the functional elucidation and characterization of other Tyl1a and FdtA homologues. O Me Me Me O Tyl1a HO O TylB HO O HO H2N O OH OH OH OTDP OTDP 14 OTDP 38 53

O Me Me Me O Tyl1a HO O TylB HO O HO H2N O OTDP OTDP 145 OTDP 147

Me O Me Me O Tyl1a HO O TylB HO O HO H2N O OH OH OH OCDP OCDP 146 OCDP 148

Figure 4-3 129 2. EXPERIMENTAL PROCEDURES

General. Materials used for molecular cloning experiments were identical to those described in Chapter 2. Antibiotics and chemicals were products of Sigma-Aldrich Chemical Co. (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA). Oligonucleotide primers for cloning of Tyl1a were prepared by Integrated DNA Technologies (Coralville, IA). Ni-NTA agarose was obtained from Qiagen (Valencia, CA). Growth media components were acquired from Becton Dickinson (Sparks, MD). Bio-gel P2 resin and all reagents for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) were purchased from Bio-Rad (Hercules, CA), with the exception of the prestained protein molecular weight marker, which was obtained from New England Biolabs. Amicon YM-10 filtration products were purchased from Millipore (Billerica, MA). Sephadex G-10 resin was acquired from Amersham (Piscataway, NJ). Kinetic data were analyzed by non-linear fit using Grafit5 (Erithacus Software Ltd., UK).

Plasmids and Vectors. The tyl1a and tylB genes were amplified from the cosmid pSET552, generously provided by Dr. Eugene Seno of Eli Lilly Research Laboratories. Vector pET28b(+) and the overexpression hosts E. coli BL21 and BL21(DE3) were purchased from Novagen (Madison, WI). The general methods and protocols for recombinant DNA manipulations followed those described by Sambrook et al.113 Bacterial Strains. Escherichia coli strain DH5α and Salmonella enterica serovar Typhimurium LT2 (ATCC 15277) were purchased from Bethesda Research Laboratories (Gaithersburg, MD) and the American Type Culture Collection (Manassas, VA), respectively. Overexpression strains E. coli BL21 and BL21(DE3) were purchased from Novagen. Instrumentation. Determination of pH values, agarose gel electrophoresis, centrifugation, photography of agarose gels, PCR, and NMR spectroscopy were 130 preformed using identical equipment to that described in Chapter 2. Chemical shifts (δ in ppm) for NMR spectroscopy are reported relative to that of dimethylsulfoxide (DMSO, δ 2.54 for 1H NMR). HPLC separations were performed using a Beckman 366 instrument (Beckman Instruments, Fullerton, CA). Analytical and semi-preparative CarboPac PA1 HPLC columns were obtained from Dionex (Sunnyvale, CA). FPLC was obtained from Amersham Biosciences and Mono-Q H/R 16/10 and Superdex 200 HR 10/30 FPLC columns were obtained from Pharmacia (Uppsala, Sweden). Mini-PROTEAN II vertical system used for SDS-PAGE, GelAir gel drying system, and the accessories and reagents used for SDS-PAGE were products of BioRad. Ultraviolet–visible spectra were obtained using a Beckman DU-650 spectrophotometer. Cell disruption was performed using a Fisher 550 Sonic Dismembrator. DNA sequencing was performed by the Core Facilities of the Institute of Cellular and Molecular Biology at the University of Texas at Austin. Mass spectra were obtained by the Mass Spectrometry Core Facility in the Department of Chemistry and Biochemistry at the University of Texas at Austin.

SDS-PAGE and Determination of Protein Concentration. The relative molecular mass and purity of enzyme samples were determined using SDS-PAGE as described by

Laemmli.141 The separating gel and stacking gel were 12% and 4% polyacrylamide, respectively. Prior to electrophoresis, protein samples were mixed with an equal volume of loading buffer and heated for 5 min at 100 °C. Loading buffer consisted of 62.5 mM

Tris-HCl buffer (pH 6.8), 10% glycerol, 2% SDS, 5% β-mercaptoethanol, and 0.0025% bromophenol blue. Electrophoresis was conducted in 25 mM Tris-HCl, 192 mM glycine, and 0.1% SDS (pH 8.3). Gels were stained with Coomassie blue (2.5 g/L of Coomassie Brilliant Blue G-250 in acetic acid : water : methanol, 1:4:5 by volume) and de-stained in ethanol : acetic acid : water (4:5:41 by volume).142 Protein concentrations were

131 determined according to Bradford143 using bovine serum albumin (BSA) as the standard and dye reagent from BioRad (Hercules, CA). Gene Amplification and Cloning of tyl1a. Two oligonucleotide primers complementary to the sequence at each end of tyl1a were prepared to amplify the tyl1a gene from the cosmid pSET552. These two primers, Tylorf1a-24/28-N-up (5'- GGAATTCCATATGGCGGCGAGCACTACGACGGAGGG-3') and Tylorf1a-28-H- down (5'-GCGCAAGCTTTCACGGGTGGCTCCTGCC-3'), were designed to amplify tyl1a with engineered 5' NdeI and 3' HindIII restriction sites (shown in bold) to be cloned

into pET28b(+), and thereby encode expression of Tyl1a with an N-terminal His6-tag. The PCR-amplified tyl1a gene was purified, digested with NdeI and HindIII restriction enzymes, and ligated into the NdeI/HindIII-digested vector, pET28b(+), to give the recombinant plasmid tyl1a/pET28b(+) (Figure 4-4). This plasmid was used to transform E. coli BL21 for protein overexpression.

132 Growth of E. coli BL21-tyl1a/pET28b(+) Cells. An overnight culture of E. coli BL21-tyl1a/pET28b(+) grown in Luria-Bertani (LB) media containing 50 μg/mL

kanamycin at 37 °C, was used (2 mL each) to inoculate 6 L (in 6 × 1 L aliquots) of LB culture containing 35 μg/mL kanamycin. These cultures were incubated at 37 oC until the

OD600 reached 0.6. Protein expression was then induced with addition of isopropyl β-D- thiogalactoside (IPTG) to a final concentration of 0.2 mM, and the cells were allowed to grow at 37 °C for an additional 5 h. After this time, OD600 had reached 1.5. The cells were harvested by centrifugation at 6000 g for 10 min and stored at -80 °C until lysis. A total of 18 g (wet weight) of cells was obtained.

Purification of N-His6-Tyl1a Protein. All purification steps were carried out at 4 °C, except the FPLC step, which was carried out at 25 °C. Thawed cells were suspended

in lysis buffer containing 15% glycerol (37 mL). Two 1000× protease inhibitor cocktails were used in the preparation: one containing 16 mg/mL each of phenylmethylsulfonylfluoride (PMSF) and N-p-tosyl-L-phenylalanine (TPCK) in 1:1 DMSO/i-PrOH, and the other containing 2.4 mg/mL each of leupeptin and lima bean

trypsin inhibitor in water. These two solutions were added to the cell suspension at 1× final concentration every 30 min prior to Ni-NTA chromatography. To this cell suspension were also added EDTA and lysozyme to 0.3 mM and 1 mg/mL final concentrations, respectively. The mixture was stirred gently for 30 min on ice. Next,

360 μg (360 μL of 1 mg/mL) of each DNase and RNase (20 μg/g of cells) were added to the mixture, and the resulting solution was incubated for an additional 15 min on ice with

gentle stirring. Subsequent sonication was performed using 12 × 20 s pulses with 30 s pauses between each pulse. The lysate was centrifuged at 15,000 g for 30 min and the supernatant was subjected to Ni-NTA chromatography.

133 The general purification procedure is a modified version of that described in the QIAexpressionist handbook provided by Qiagen for use with Ni-NTA agarose resin. Specifically, Ni-NTA slurry (10 mL) was washed twice with lysis buffer before use. The soluble protein fraction was incubated with 10 mL of Ni-NTA beads, which had been washed with lysis buffer, on a rotator at 4 °C for 1 h. Lysate and beads were then loaded onto a column, which was allowed to drain, and then washed with wash buffer containing 15% glycerol (75 mL). Bound protein was eluted using elution buffer containing 15%

glycerol (4 mL) and collected in 4 × 1 mL portions, which were pooled and dialyzed

against 3 × 1 L of 50 mM NaH2PO4 buffer, 300 mM NaCl, 15% glycerol, pH 8.0. Tyl1a was greater than 90% pure after Ni-NTA chromatography, and the yield was 45 mg per liter of cell culture. Further purification of Tyl1a was performed on an FPLC Mono-Q 16/10 column using a linear gradient of 0 to 50% buffer B in buffer A as eluant. Buffer A was 20 mM Tris•HCl, pH 7.5, and buffer B was 20 mM Tris•HCl, 250 mM NaCl, pH 7.5. The detector was set at 280 nm and the flow rate was 3 mL/min. Collected fractions were analyzed by SDS-PAGE, the desired fractions pooled, mixed with an equal volume of 20 mM Tris•HCl buffer, 250 mM NaCl, 30% glycerol, pH 7.5, and concentrated to 1-2 mg/mL using an Amicon YM-10 filter. On the basis of SDS-PAGE analysis, the isolated Tyl1a was estimated to be greater than 95% pure. The purified Tyl1a was flash frozen and stored at -80 °C.

Cleavage of N-His6-tag from Tyl1a by Thrombin. Tyl1a expressed from pET28b(+) contains a thrombin cleavage site between the His6-tag and the first amino acid of Tyl1a. The Novagen Thrombin Cleavage Capture Kit containing biotinylated thrombin and streptavidin agarose for removal of thrombin was used for the preparation

of N-His6-cleaved Tyl1a. Small-scale trials showed that Tyl1a-N-His6 could be efficiently cleaved after a 2 h incubation with 0.5 units of thrombin per mg Tyl1a.

134 Accordingly, Tyl1a-N-His6 (5 mg) was incubated with 2.5 units of thrombin at 25 °C overnight. Thrombin was then removed using streptavidin agarose according to the manufacturer's protocol. Cleaved Tyl1a with remaining non-native N-terminal sequence (GSH), was further purified by FPLC using a Mono Q column as described above. The desired fractions (>95% pure) were pooled, mixed with an equal volume of 20 mM Tris•HCl buffer containing 250 mM NaCl, 30% glycerol, pH 7.5, and concentrated to 1-2 mg/mL using an Amicon YM-10 filter. The purified Tyl1a was flash frozen and stored at -80 °C.

Molecular Mass Determination. The native molecular mass of Tyl1a was determined by gel filtration performed on an FPLC equipped with a Superdex 200 HR

10/30 column. The proteins were eluted using 50 mM NaH2PO4 buffer, 150 mM NaCl, pH 7.0, at a flow rate of 0.5 mL/min. The system was calibrated with protein standards

(Sigma), and the void volume (V0) was measured using blue dextran. The data were analyzed by the method of Andrews.144 The following protein standards (Sigma MW- GF-200) were injected into FPLC (250 μL each): cytochrome c (12.4 kDa, 4 mg/mL), carbonic anhydrase (29 kDa, 6 mg/mL), bovine serum albumin (BSA, 66 kDa, 10 mg/mL), alcohol dehydrogenase (150 kDa, 10 mg/mL), and β-amylase (200 kDa, 8 mg/mL). Blue dextran (2000 kDa, 250 μL of 2 mg/mL) was used to determine the void volume of the column. Thrombin-cleaved Tyl1a (500 μL) used in the analysis was of a 2

mg/mL concentration.

Enzymatic Synthesis of TDP-4-keto-6-deoxy-D-glucose (14). The large-scale enzymatic preparation of Tyl1a substrate was initiated by coupling thymidine with glucose-1-phosphate (12) to make TDP-D-glucose (13), which was then converted to

TDP-4-keto-6-deoxy-D-glucose (14). Preparation of TDP-D-glucose from thymidine and glucose-1-phosphate (12) was conducted in a two-stage, "one-pot" reaction.137 In the first

135 stage, a mixture containing 76.2 mM phosphoenolpyruvate (PEP), 24 mM thymidine, 1.6

mM ATP, 27 mM MgCl2, 25 μM thymidine kinase (TK), 25 μM thymidylate kinase (TMK), 25 μM nucleoside diphosphate kinase (NDK), and 1000 units of rabbit muscle pyruvate kinase (PK) in 17 mL of 45 mM Tris•HCl buffer, pH 7.5, was incubated at 37 °C for 4 h, generating thymidine triphosphate (TTP). The enzymes were removed by

filtration through an Amicon YM-10 membrane, and glucose-1-phospate (12), MgCl2, and α-D-glucose-1-phosphate thymidylyltransferase (RfbA) from S. enterica LT2 were added to the filtrate to give final concentrations of 28 mM, 50 mM, and 36 μM, respectively. The mixture was incubated for 16 h at 37 °C, centrifuged at 5,000 g for 10 min to remove precipitate, and filtered through an Amicon YM-10 membrane to remove enzymes. The crude product (13), with a theoretical yield of 228 mg, was stored at 4 °C.

The enzyme-free filtrate from the previous step was loaded onto a Bio-gel P2

column (25 mm × 100 cm) pre-washed with water, and run at a flow rate of 12 mL/h with water as the eluant, with 8 mL fractions collected. Fractions showing UV absorption at 267 nm were lyophilized, and the identity and purity of the compounds in each fraction

assessed by 1H and 31P NMR spectroscopy. TDP-D-glucose (13) containing fractions (total weight 170 mg), which varied in purity from 25-70%, were pooled according to

their purities. Further purification of TDP-D-glucose was performed by FPLC equipped with a Mono Q 16/10 column. A linear gradient of 0 to 40% of a solution of 400 mM

NH4HCO3 in water was used as the eluant. The detector was set at 280 nm, and the flow rate was 5 mL/min. Fractions containing the major peak were lyophilized individually,

redissolved in water, and lyophilized again to remove NH4HCO3. The purities of these fractions (total weight 123 mg), which ranged from 50-90%, were assessed by 1H and 31P

NMR spectroscopy. From these fractions, 23 mg of 90% pure TDP-D-glucose (13) was

1 obtained. H NMR (300 MHz, D2O) δ 1.78 (3H, s, 5"-Me), 2.22 (2H, m, 2'-H), 3.28-3.40

136 (2H, m, 2-H, 3-H), 3.59-3.78 (2H, m, 4-H, 5-H), 3.95-4.06 (3H, m, 4’-H, 5'-H), 4.45 (1H, m, 3'-H), 5.44 (1H, dd, J = 6.9, 3.3, 1-H), 6.20 (1H, t, J = 6.9, 1'-H), 7.59 (1H, s, 6"-H).

TDP-D-glucose (13, 23 mg) obtained from the previous step was dissolved in 47

mM KH2PO4 buffer, pH 7.5, to give a final concentration of 29 μM. This solution was

incubated with TDP-D-glucose 4,6-dehydratase (RfbB) from S. enterica LT2 (18 μM) at 37 °C for 2 h, after which RfbB was removed by filtration through an Amicon YM-10 membrane. The filtrate was loaded onto a Sephadex G-10 column pre-washed with water and run at a flow rate of 1 mL/min using water as the eluant, with 10 mL fractions collected. Those fractions exhibiting absorption at 267 nm were lyophilized and their

purities assessed by 1H and 31P NMR spectroscopy. Those fractions containing pure

TDP-4-keto-6-deoxy-D-glucose (14, 12.3 mg, >90% purity) were combined and the concentration of 14 in the solution was determined spectrophotometrically at 267 nm, (ε

-1 -1 1 = 9600 M cm ). H NMR (300 MHz D2O) of 14 (a mixture of hydrate and keto forms): δ 1.08 (3H, d, J = 6.5, 5-Me of hydrate form), 1.12 (3H, d, J = 6.5, 5-Me of keto form), 1.79 (3H, s, 5"-Me), 2.09-2.26 (2H, m, 2'-H), 3.48 (1H, m, 2-H of hydrate form), 3.64 (1H, d, J = 10.0, 3-H of hydrate form), 3.68 (1H, m, 2-H of keto form), 3.96 (1H, q, J = 6.5, 5-H of hydrate form), 4.01-4.07 (3H, m, 4'-H, 5'-H), 4.48 (1H, m, 3'-H), 5.41 (1H, dd, J = 7.3, 3.8, 1-H of hydrate form), 5.59 (1H, dd, J = 7.0, 3.0, 1-H of keto form), 6.20 (1H, t, J = 6.9, 1'-H), 7.60 (1H, s, 6"-H).

Preparation of Enzymes used in Enzymatic Synthesis of 14. The TK, TMK, and NDK used in this synthesis were prepared as described by Takahashi et al.137 Rabbit muscle pyruvate kinase was purchased from Sigma as a 400-800 units/mg ammonium sulfate precipitate. This ammonium sulfate precipitate was dissolved in water to a

concentration of 2500 units/mL, dialyzed against 50 mM NaH2PO4 buffer, 300 mM NaCl, pH 8.0, to remove ammonium sulfate, and stored at -80 °C.

137 RfbA was prepared by Dr. Yung-nan Liu of this laboratory as follows. The rfbA gene was amplified from Salmonella enterica serovar Typhimurium LT2 genomic DNA using PCR. The start primer contained an engineered BamHI restriction site (in italics), ribosomal binding sequence (underlined), and AT rich region upstream of the native start codon (shown in bold), with sequence 5'-CGGGATCCGAAGGAGATATATAATGAA- AACGCGTAAGGGC-3'; and the halt primer contained an engineered PstI restriction site (in italics) and a C-terminal His5-tag (underlined) immediately downstream of the stop codon (in bold), with sequence 5'-CTTGCATGCCTGCAGTTAATGATGATGAT-

GATGTAAACCTTTCACCATC-3'. The PCR-amplified gene was purified, digested with BamHI and PstI, and ligated into BamHI/PstI digested pUC18. The resulting construct was used to transform E. coli BL21(DE3). Overexpression was achieved by growth of the transformed host in LB media in the presence of 100 μg/mL ampicillin at 37 oC overnight. Protein was purified from the harvested cells by Ni-NTA affinity chromatography in an identical manner to that used for the purification of Tyl1a. RfbB used in the synthesis of 14 was prepared by Dr. Yung-nan Liu in the following manner. The rfbB gene was amplified from Salmonella enterica serovar Typhimurium LT2 genomic DNA by PCR. The start primer contained an engineered EcoRI restriction site (in italics), ribosomal binding sequence (underlined), and AT rich region upstream of the native start codon (shown in bold), with sequence 5'-

GGAATTCGAAGGAGATATATAATGGTGAAGATACTTATTACTGG-3'. The halt

primer contained an engineered BamHI restriction site (in italics) and a His5-tag sequence (underlined) immediately downstream of the stop codon (in bold), with sequence 5'- CGGGATCCTTAATGATGATGATGATGCTGGCGTCCTTCATAGTTC-3'. The PCR-amplified gene was purified, digested with EcoRI and BamHI, and ligated into EcoRI/BamHI digested pUC18. The resulting construct was used to transform E. coli

138 BL21(DE3). Overexpression was achieved by growth of the transformed host in LB medium supplemented with 100 μg/mL ampicillin at 37 oC overnight. Protein was purified from the harvested cells by Ni-NTA affinity chromatography in an identical manner to that used for the purification of Tyl1a. HPLC Activity Assay for Tyl1a. A reaction mixture (35 μL) containing 2.85 μM

Tyl1a (with or without N-His6-tag) and 1 mM TDP-4-keto-6-deoxy-D-glucose (14) in 50

mM KH2PO4 buffer, pH 7.5, was incubated at 25 °C. Aliquots of 5 μL were removed at various time points, quenched by flash freezing in liquid nitrogen, thawed at 4 °C, diluted

by addition of 20 μL of 50 mM KH2PO4 buffer, pH 7.5, filtered through a Microcon YM- 10 membrane to remove enzyme, and the filtrate flash frozen until HPLC analysis. HPLC analysis was performed using a Dionex Carbopac PA1 column with 9.5 μL of sample for each injection. The sample was eluted by a gradient of water as solvent A and

500 mM NH4OCOCH3 (adjusted to pH 7.0 with aqueous NH3) as solvent B where the gradient ran from 5 to 20% B over 15 min, then 20 to 60% B over 20 min, 60 to 100% B over 2 min, 3 min wash at 100% B, 100-5% B over 5 min, and re-equilibration at 5% B for 15 min. The flow rate was 1 mL/min, and the detector was set at 267 nm. The

retention times were 35.3 min for TDP-4-keto-6-deoxy-D-glucose (14), 39.0 min for

TDP-3-keto-6-deoxy-D-glucose (38), 41.9 min for TDP, and 1.8 min for the degradation product, (2R,3R)-2-methyl-3,5-dihydroxy-4-keto-2,3-dihydropyran (149). The substrate

and product ratios were calculated from the integration of the corresponding peaks on the HPLC chromatogram.

In situ 1H NMR Assay for Tyl1a and Characterization of Products. A reaction

mixture (600 μL) containing 10 mM TDP-4-keto-6-deoxy-D-glucose (14), 50 mM

KH2PO4 buffer, pH 7.5, 5% (v/v) DMSO-d6 (as a reference) was prepared in an NMR tube. After shimming and initial peak integration in a 500 MHz NMR spectrometer, the

139 sample was removed from the tube, mixed thoroughly with glycerol-free His6-tagged Tyl1a (6 or 10 μM final concentration), and returned to the NMR tube. Data were

acquired at 5-min intervals for 150 min. Spectral data for TDP-3-keto-6-deoxy-D- glucose (38) and the degradation product (2R,3R)-2-methyl-3,5-dihydroxy-4-keto-2,3- dihydropyran (149) were assigned from the spectra generated during the in situ assay. 1H

NMR (DMSO-d6) of 38: δ 1.23 (3H, d, J =5.5, 5-Me), 1.73 (3-H, s, 5"-Me), 2.15-2.20 (2H, m, 2'-H), 3.92-4.01 (5H, m, 4-H, 5-H, 4'-H, 5'-H), 4.00-4.05 (1H, m, 2-H), 4.45-4.50 (1H, m, 3'-H), 5.68 (1H, dd, J =7.0, 4.5, 1-H), 6.15 (1H, t, J = 7.0, 1'-H), 7.55 (1H, s, 6"-

1 H). H NMR (DMSO-d6) of 149: δ 1.28 (3H, d, J = 6.0, 5-Me), 4.03 (1H, d, J = 12.3, 4- H), 4.10 (1H, dq, J = 12.3, 6.0, 5-H), 7.32 (1H, s, 1-H). Determination of Kinetic Parameters for Tyl1a. The steady state kinetic parameters of the Tyl1a-catalyzed reaction were determined by the HPLC activity assay

as described above. Reactions containing 100 nM N-His6-tagged Tyl1a and varied

amounts of TDP-4-keto-6-deoxy-D-glucose (14) (3 μM to 1 mM) in 50 mM KH2PO4 buffer, pH 7.5, were incubated at 25 °C. Larger reaction volumes were used for the lower substrate concentrations to facilitate HPLC analysis. Aliquots were taken at four different time points for each of 16 substrate concentrations. Incubation mixtures with lower substrate concentrations were monitored for shorter periods of time (2 to 3 min), and those with higher substrate concentrations were monitored for periods as long as 7 min. Since the concentration of substrate in each sample was known, the percent conversion determined by HPLC could be used to calculate μmoles of product formed for each time point. The amount of product formed at the four time points for a given substrate concentration were plotted against time, and the slope of each line was plotted versus substrate concentration. The resulting data were fit to the Michaelis–Menten equation by non-linear regression using Grafit 5 to determine the kcat and Km values.

140 As a comparison, data obtained from the HPLC time course study and those obtained from the in situ NMR assay by following changes in integration of individual proton signals over time (for example, the 5-methyl signal of 14, 38, and the 2-methyl

signal of 149), were also used to calculate the apparent kcat. In each case, peak integrations were normalized to initial substrate concentration and fit to exponential equations by non-linear regression analysis. The rate constant for the disappearance of

substrate from these data was used to calculate the apparent kcat in each experiment. HPLC Assay of Coupled Tyl1a-TylB Reaction. The C-3 aminotransferase TylB used in this coupled assay was prepared by Dr. Hiroshi Yamase of this group according

to published procedures.46 The chemically synthesized standard of the TylB product,

TDP-3-amino-3,6-dideoxy-D-glucose (53), was prepared by Dr. Siu-man Yeung of this group according to published procedures.44 The reaction mixture (100 μL) contained

28.5 μM Tyl1a, 10 μM TylB, 1 mM TDP-4-keto-6-deoxy-D-glucose (14), 10 mM L- glutamate, 50 μM pyridoxal-5'-phosphate (PLP), in 50 mM KH2PO4 buffer, pH 7.5. Aliquots were removed at various time points and analyzed by HPLC in the same manner as used for Tyl1a activity assays. The HPLC retention time for TDP-3-amino-3,6-

dideoxy-D-glucose (53) was 13.7 min. For isolation and MS characterization of the TylB product (53) and Tyl1a degradation product (149), 300 μg of TDP-4-keto-6-deoxy-D- glucose (14) at a concentration of 1 mM was incubated with 3 μM Tyl1a, 10 μM TylB,

50 μM PLP, and 10 mM L-glutamate in 50 mM KH2PO4 buffer, pH 7.5. The total reaction volume was 550 μL. The reaction was incubated at 25 °C for 70 min, quenched by flash freezing in liquid nitrogen, thawed on ice, and enzymes removed by filtration through Microcon YM-10 at 4 °C. The sample was then separated using a semi- preparative Dionex CarboPac PA-1 column and a gradient elution program identical to that used for the analytical Dionex CarboPac PA-1 column, with a flow rate of 5 mL/min.

141 Compounds 53 and 149, which eluted at 12.5 and 1.8 min, respectively, were collected manually and lyophilized to dryness. HPLC analysis showed 35% conversion of 14 to 53 and 65% conversion of 14 to 149 (the degradation product) and TDP. Compound 53 was resuspended in water to a concentration of 0.1 mg/mL and 149 was resuspended in methanol to a concentration of 1 mg/mL. High resolution ESI-MS of 53: calculated for

C16H27N3O14P2 (M - H) 546.0883, found 546.0885; 149: calculated for C6H4O4 (M + H) 145.0501, found 145.0500.

Enzymatic Preparation of TDP-4-keto-2,6-dideoxy-D-glucose (145). This

compound was prepared enzymatically by Dr. Lin Hong of this group from TDP-D- glucose (13) using purified enzymes RfbB from S. enterica LT2, TylX3 from

Streptomyces fradiae,145 and SpnN, the TDP-3,4-diketo-2,6-dideoxy-D-glucose 3- ketoreductase from the spinosyn biosynthetic pathway of Saccharopolyspora spinosa.54

A typical reaction mixture containing 28 mM TDP-D-glucose (13), 34 μM RfbB, 17 mM NADPH, 10 μM TylX3, 17 μM SpnN, in 50 mM Tris•HCl buffer, pH 7.5, 10% glycerol, was incubated at 25 °C for 4 h. Enzymes were removed using a Centricon YM-10 microconcentrator, and the filtrate was separated on a Bio-gel P2 gel filtration column

(25 mm × 100 cm) pre-washed with 25 mM NH4HCO3, and run at a flow rate of 12 mL/h

with 25 mM NH4HCO3 as the eluant. Fractions (8 mL each) showing UV absorbance at 267 nm were lyophilized, and the identity and purity of the compounds in each fraction assessed by 1H and 31P NMR spectroscopy. The 1H-NMR spectrum of the purified TDP-

4-keto-2,6-dideoxy-D-glucose (145) was identical to that previously reported.146

HPLC Analysis of Incubation Mixture Containing Tyl1a with TDP-4-keto-2,6- dideoxy-D-glucose (145). The reaction mixture (70 μL) contained 2.85 μM Tyl1a and 1

mM 145 in 50 mM KH2PO4 buffer, pH 7.5, and was incubated at 25 °C. Aliquots were removed at various time points, and analyzed by HPLC in the same manner as used for

142 other Tyl1a activity assays. The retention time of the substrate 145 was 33.4 min, that of

the product, TDP-3-keto-2,6-dideoxy-D-glucose (150), was 37.0 min, and that of the degradation product, (2R,3R)-2-methyl-3-hydroxy-4-keto-2,3-dihydropyran (151), was 1.7 min. Coupled Assay of Tyl1a-TylB with TDP-4-keto-2,6-dideoxy-D-glucose (145). The reaction mixture (50 μL), which was incubated at 25 °C, contained 2.85 μM Tyl1a, 28.5

μM TylB, 1 mM 145, 28.5 mM L-glutamate, and 142.5 μM PLP in 50 mM KH2PO4 buffer, pH 7.5. Aliquots were removed at various time points and subjected to HPLC analysis as described for the Tyl1a activity assays. The retention time of the product,

TDP-3-amino-2,3,6-trideoxy-D-glucose (147), was 9.5 min. For isolation and MS characterization of the TylB product (147) and Tyl1a degradation product (151), 1.0 mg

of the SpnN product TDP-4-keto-2,6-dideoxy-D-glucose (145) at a concentration of 2

mM was incubated with 1 μM Tyl1a, 30 μM TylB, 250 μM PLP, and 50 mM L-glutamate in 50 mM KH2PO4 buffer, pH 7.5. The total reaction volume was 940 μL. The reaction was incubated at 25 °C for 12 h and enzymes removed by filtration through Microcon YM-10 at 4 °C. The sample was then separated using a semi-preparative Dionex CarboPac PA-1 column and a gradient elution program identical to that used for the analytical Dionex CarboPac PA-1 column, with a flow rate of 5 mL/min. Compounds 147 and 151, which eluted at 8.9 and 1.8 min, respectively, were collected. HPLC

analysis showed 15% conversion of 145 to 147 and 85% conversion of 145 to 151 (the degradation product) and TDP. Compound 151 was lyophilized to dryness and resuspended in methanol to a concentration of 1 mg/mL. Compound 147, which is unstable when lyophilized to dryness, or when stored at 25 °C, was lyophilized to reduce its volume tenfold, and stored at -80 °C. High resolution EI-MS of 151: calculated for

143 + C6H4O3 (M ) 128.0473, found 128.0486. High resolution ESI-MS 147: calculated for

C16H27N3O13P2 (M - H) 530.0941, found 530.0941. HPLC Analysis of Incubation Mixture Containing Tyl1a and CDP-4-keto-6-

deoxy-D-glucose (146). Preparation and quantitation of CDP-4-keto-6-deoxy-D-glucose (146) was performed by Dr. Yung-nan Liu of our group as described by Chen et al.147 Preparation of enzymes used to make 146 was performed by Dr. Yung-nan Liu as described by Thorson et al.148, 149 The reaction mixture (100 μL) contained 100 μM

Tyl1a and 1 mM 146 in 50 mM KH2PO4 buffer, pH 7.5, and the incubation was carried out at 25 °C. Aliquots were removed at various time points and analyzed by HPLC in the same manner as used for other Tyl1a activity assays. The retention times of 146, the product CDP-3-keto-6-deoxy-D-glucose (152), and CDP were 33.1, 36.6, and 42.3 min, respectively.

Coupled Assay of Tyl1a, TylB with CDP-4-keto-6-deoxy-D-glucose (146). The reaction mixture (100 μL), which was incubated at 25 °C, contained 35 μM Tyl1a, 35 μM

TylB, 1 mM 146, 35 mM L-glutamate, and 175 μM PLP in 50 mM KH2PO4 buffer, pH 7.5. Aliquots were removed at various time points and subjected to HPLC analysis as described for the Tyl1a activity assays. The retention time of the product, CDP-3-amino-

3,6-dideoxy-D-glucose (148), was 11.4 min. For isolation and MS characterization of the

TylB product (148), 1.8 mM CDP-4-keto-6-deoxy-D-glucose (146) was incubated with

35 μM Tyl1a, 35 μM TylB, 175 μM PLP, and 35 mM L-glutamate in 100 μL of 50 mM

KH2PO4 buffer, pH 7.5. The reaction mixture was incubated at 25 °C for 48 h and enzymes were removed by filtration using a Microcon YM-10 at 4 °C. The sample was then further purified using a semi-preparative Dionex CarboPac PA-1 column and a gradient elution program identical to that used for the analytical Dionex CarboPac PA-1 column, with a flow rate of 5 mL/min. Compound 148, which was eluted at 11.1 min,

144 was collected manually, lyophilized to dryness and resuspended in water to a concentration of 0.1 mg/mL. High resolution ESI-MS of 148: calculated for

C15H25N4O14P2 (M - H) 547.0843, found 547.0841. HPLC analysis showed 46% conversion of 146 to 148, with the remaining 54% consisting of 146 and 152.

3. RESULTS AND DISCUSSION

Purification and Characterization of Tyl1a. The tyl1a gene128 was amplified, cloned, and overexpressed to give N-terminal His6-tagged Tyl1a. Production of soluble protein was quite efficient, with 275 mg of Tyl1a obtained from a 6 L culture after the Ni-NTA chromatographic step. Further purification by FPLC using a Mono Q column improved the purity to >95% as assessed by SDS-PAGE. Subunit molecular mass was estimated to be 19 kDa based on SDS-PAGE analysis (Figure 4-5), which agrees well

with the calculated mass of 18,806 Da for the N-terminal methionine-cleaved His6-tagged Tyl1a. Thrombin cleavage was also carried out to generate Tyl1a without the N-terminal

His6-tag for comparative kinetic studies. Gel filtration analysis revealed a native molecular mass of 31.0 kDa for Tyl1a, suggesting a homodimeric structure in solution.

Figure 4-5. SDS-PAGE gel (18%) of (A) purified N- His6-tagged Tyl1a, and (B) purified Tyl1a with His6- tag removed by thrombin cleavage.

145 Catalytic Properties of Tyl1a. Because Tyl1a exhibits modest sequence identity (34% identity, 52% similarity) with FdtA, which catalyzes the conversion of TDP-4-keto-

6-deoxy-D-glucose (14) to TDP-3-keto-6-deoxy-D-galactose (153, Figure 4-1 inset),130 and was implicated by the in vivo results described in Chapter 3 as a ketoisomerase in the mycaminose pathway, we thought it likely that Tyl1a catalyzes conversion of 14 to TDP-

3-keto-6-deoxy-D-glucose (38). The predicted substrate of Tyl1a, TDP-4-keto-6-deoxy-

D-glucose (14), was therefore prepared enzymatically from thymidine and glucose-1- phosphate (12) using six enzymes in a two-stage, one-pot reaction.137 Subsequent HPLC

analysis of an incubation mixture containing 14 and N-His6-tagged Tyl1a showed time- dependent depletion of substrate and the appearance of three new products (Figure 4-6, traces a, b). Identical results were also observed using thrombin-cleaved Tyl1a under the

0.16 14 TMP 0.12 a 38 TDP 0.08 149 b AU 53 0.04 c

d 0.00 0 5 10 15 20 25 30 35 40 45 retention time (min) Figure 4-6. HPLC traces showing product formation in the Tyl1a and TylB reactions: (a) incubation mixture containing TDP-4-keto-6-deoxy-D-glucose (14, 1 mM) in 50 mM KH2PO4 buffer (pH 7.5) without Tyl1a; (b) incubation mixture containing 14 (1 mM) and Tyl1a (2.85 μM) in the same phosphate buffer as in (a); (c) incubation mixture containing 14 (1 mM) and Tyl1a (28.5 μM) in the presence of TylB (10 μM), PLP (50 μM), and L-glutamate (10 mM), in the same phosphate buffer as in (a). Note that TMP and TDP peaks are not visible due to the adjustment of the scaling to keep the strong peaks of 53 and 149 in scale; (d) chemically synthesized TDP-3-amino-3,6- dideoxy-D-glucose (53).

146 same reaction conditions, indicating that the N-terminal His6-tag has no effect on Tyl1a activity. Thus, the N-His6-tagged Tyl1a was used for all subsequent work. Analysis of the full reaction time course revealed that 14 (retention time = 35.3 min) was first converted to an intermediate (retention time = 39.0 min), which was swiftly depleted with concomitant formation of TDP (retention time = 41.9 min) and a new product (retention time =1.8 min) (Figure 4-7). Integration of the peak corresponding to the substrate and each of the three new peaks over time gave the traces shown in Figure 4-8. Evidently, the immediate product having a retention time of 39.0 min is unstable, degrading to TDP

149 TDP 0.12 TMP 38 14 120’ 0.08 90’ AU 60’ 40’ 20’ 0.04 10’ 5’ 3’ 2’ 0.00 0 5 10 15 20 25 30 35 40 45 50 retention time (min)

Figure 4-7. HPLC traces of Tyl1a reaction time course study (2 – 120 min) using 1 mM 14 and 2.85 μM Tyl1a in 50 mM KH PO buffer (pH 7.5). 2 4 and a new species with a retention time of 1.8 min. In an attempt to divert the transient intermediate to a more stable product, TylB, which catalyzes the subsequent step in the mycaminose pathway, was added to the incubation mixture. As expected, the addition of

TylB along with PLP and L-glutamate to the reaction mixture led to a new product (retention time = 13.7 min), which co-eluted with the chemically synthesized TDP-3- amino-3,6-dideoxy-D-glucose (53) (Figure 4-6, traces c, d). The identity of this new product was further confirmed to be 53 by high resolution ESI mass spectrometry.

147 Clearly, the unstable intermediate generated in the Tyl1a reaction is TDP-3-keto-6-

deoxy-D-glucose (38), which is converted to 53 in the presence of TylB.

1

0.75

0.5 normalized 0.25

peak integration at nm 267 0 050100150 time (min) Figure 4-8. Time course of Tyl1a (2.85 μM)-catalyzed reaction using 14 (1 mM) as substrate. Integrations of the HPLC peaks of substrate, product, and degradation products were plotted versus time. (○) represents 14, (●) 387, (■) 149, and (□) TDP.

In situ 1H NMR Analysis of Tyl1a Reaction and Identification of Reaction Products. In order to directly characterize the unstable Tyl1a reaction product, in situ 1H

NMR analysis was performed. In this experiment, TDP-4-keto-6-deoxy-D-glucose (14) was incubated with glycerol-free Tyl1a in an NMR tube, and the reaction was monitored at 5-min intervals for 150 min. A stack plot of the resulting spectra in which 6 μM Tyl1a was used is shown in Figure 4-9. Disappearance of the proton signals for 14 is seen along with the appearance of a new set of signals for the Tyl1a product. These signals reached a maximum intensity at about 50 min, and then diminished. A third set of signals

corresponding to the degradation product also appeared after a short lag. Comparison of these spectra to those of 14 and the chemoenzymatically synthesized 3846 enabled assignment of the 1H NMR signals for 38. Thus, the Tyl1a product is indeed TDP-3-

keto-6-deoxy-D-glucose (38) which confirms that Tyl1a is a TDP-4-keto-6-deoxy-D- glucose-3,4-ketoisomerase. Signals corresponding to the degradation product were also identified. The chemical shifts and coupling constants of these signals are consistent with

148 150

100

time (min) 50

5

7 6 5 4 3 2 1

chemical shift (ppm)

Figure 4-9. 1H NMR stack plot of Tyl1a reaction (10 mM 14, 6 μM Tyl1a in 50 mM KH2PO4 buffer, pH 7.5) monitored over 150 min. The signals correspond to each compound are labeled as (○) from 14, (●) from 38, and (□) from 149.

the structure of (2R,3R)-2-methyl-3,5-dihydroxy-4-keto-2,3-dihydropyran (149), which is likely formed by C-2 deprotonation of 38 followed by elimination of TDP to give the 1,2- unsaturated ketone (Figure 4-10). High resolution CI-MS analysis of the 1.8 min peak collected from the HPLC assay confirms the assigned structure.

Me O Me Me H O Tyl1a HO O O HO HO OH O OH OTDP OTDP 14 38 TDP O OH 149 Figure 4-10

149 Determination of Kinetic Parameters for Tyl1a-Catalyzed Reaction. To determine the steady state kinetic parameters for the Tyl1a-catalyzed conversion of 14 to

38, a discontinuous HPLC assay was developed and performed. The plot of v0 (initial velocity) versus [S] (3 μM-1 mM) (Figure 4-11) was fitted to the Michaelis-Menten

-1 equation by non-linear regression to yield the values of kcat of 6.1 min and Km of 27 μM for 14. Data obtained from the HPLC time course study (Figure 4-8) and those obtained

from the in situ NMR assay (Figure 4-12) were also used to calculate apparent kcat values.

-1 A kcat of ~ 7.0 min was obtained from the HPLC data, and an apparent kcat of 2.4 ± 0.6 min-1 was obtained from the in situ 1H NMR assays. These values are in good agreement

with the kcat value determined by the HPLC initial velocity assays.

0.6 )

-1 0.4 min

-1 M μ 0.2 Vo ( Vo

0 0 100 200 300 400 500 [S] (μM)

Figure 4-11. Plot of v0 versus [S] determined via HPLC assay from which the steady state kinetic constants for the Tyl1a reaction were determined.

Tyl1a Substrate Specificity. To determine the substrate specificity of Tyl1a, the 2-

deoxy analogue of 14, TDP-4-keto-2,6-dideoxy-D-glucose (145), and the CDP version of

14, CDP-4-keto-6-deoxy-D-glucose (146), were tested as possible substrates. Compound

145 was prepared enzymatically from TDP-4-keto-6-deoxy-D-glucose (14) using TylX3145 from the mycarose biosynthetic pathway of S. fradiae, and SpnN54

150 10

7.5

5

2.5 concentration (mM)

0 050100150 time (min)

Figure 4-12. Plot of the integration of the 1H NMR signals (see Figure 3) of the 5- methyl group of 14 (○) and 38 (●), and that of 2-methyl group of 149 (□) during the in 1 situ H NMR assay (10 mM 14 and 10 μM Tyl1a in 50 mM KH2PO4 buffer, pH 7.5).

from the forosamine biosynthetic pathway of S. spinosa, as the catalysts (Figure 4-13). HPLC analysis of an incubation mixture containing 145 (1 mM) and Tyl1a (2.85 μM) showed time-dependent disappearance of the substrate peak at 33.4 min, accumulation of a small amount of a new peak at 37.0 min, and the tandem formation of TDP and another new peak at 1.7 min (Figure 4-14, trace a, b). Based on the retention times and confirmed identities of products from the assay of Tyl1a with 14, the peaks at 37.0 and

1.7 min likely correspond to TDP-3-keto-2,6-dideoxy-D-glucose (150) and its degradation product (2R,3R)-2-methyl-3-hydroxy-4-keto-2,3-dihydropyran (151),

Me O Me O Me O TylX3 O O SpnN O HO HO OH O OTDP 14 OTDP 155 OTDP 145

Tyl1a Me Me Me TDP O HO O TylB HO O HO H2N O O OTDP OTDP 147 150 151 Figure 4-13 151 145 a 0.40 150 151 b 0.3 c 147 AU 0.20 146 d 149 152 0.10 e

148 f 0.0 0 5 10 15 20 25 30 35 40 45

Figure 4-14. HPLC traces showing conversion of alternate substrates 145 and 146 by Tyl1a and TylB: (a) 1 mM 145 without Tyl1a, (b) 1 mM 145 + 2.85 μM Tyl1a, (c) 1 mM 145 + 2.85 μM Tyl1a, 28.5 μM TylB, 142.5 μM PLP, and 28.5 mM L-glutamate, (d) 1 mM 146 without Tyl1a, (e) 1 mM 146 + 100 μM Tyl1a, (f) 1 mM 146 + 35 μM Tyl1a, 35 μM TylB, 175 μM PLP, and 35 mM L-glutamate.

respectively (Figure 4-13). While no spectral data was collected for 150 due to its low yield and instability, high resolution EI-MS data of the 1.7 min peak is consistent with the assigned structure of 151. By plotting the peak integration of substrate, product, and degradation products of this reaction versus time (Figure 4-15), it was found that 150 is more prone to degradation than 38, as less of it accumulates during the course of the reaction, and TDP and 151 are formed significantly more rapidly. By fitting the substrate depletion data from this time course to a single exponential equation, the apparent kcat for the conversion of 145 to 150 by Tyl1a was estimated to be 7.9 min-1 (Figure 4-15). These analyses indicate that 145 and 14 are comparable substrates for Tyl1a.

We next examined whether Tyl1a can accept CDP-4-keto-6-deoxy-D-glucose (146) as a substrate. This compound was generated enzymatically from CTP and

glucose-1-phosphate (12) by the action of α-D-glucose cytidylyltransferase (Ep) and

CDP-D-glucose 4,6-dehydratase (Eod) from Yersinia pseudotuberculosis as described 152

1

0.75

0.5

at nm 267 0.25

normalized peakintegration 0 050100150 time (min)

Figure 4-15. Time course of Tyl1a (2.85 μM)-catalyzed reaction using 145 (1 mM) as substrate. Integrations of the HPLC peaks of substrate, product, and degradation products were plotted versus time. (○) represents 145, (●) 150, (■) 151, and (□) TDP. previously (Figure 4-16).147 The assay was initially performed using 1 mM 146 and 1 μM Tyl1a, and monitored by HPLC as before. Although formation of a new peak at 36.6 min was discernible, its formation was very slow, with only 2% conversion observed after a 24 h incubation. The assay was repeated with an increased amount of Tyl1a (100 μM), and the time-dependent disappearance of substrate at 33.1 min, the formation of product at 36.6 min, and the degradation of the product to TDP and a new product at 1.8 min were clearly noted, with product levels reaching a maximum of 30% conversion after

OH OH O O O Me O Ep O Eod O HO HO HO OH OH OH = = OCDP 12 OPO3 OCDP 146

Tyl1a Me Me Me CDP O HO O TylB HO O HO H2N OH O OH O OH OCDP OCDP 148 152 149 Figure 4-16

153 8 h (Figure 4-14, traces d, e). The pattern of product formation and degradation is similar to that observed for the Tyl1a reaction with its natural substrate, suggesting that the

product in this case is likely CDP-3-keto-6-deoxy-D-glucose (152) and the degradation product seen at 1.8 min is the same pyran (149) formed in the reaction of Tyl1a with 14 (Figure 4-16). By fitting the substrate depletion data from this time course to a single exponential, the apparent kobs value for the conversion of 146 to 152 by Tyl1a under these conditions was calculated to be 0.015 min-1 (Figure 4-17). Thus, Tyl1a appears to also be able to catalyze a 3,4-ketoisomerization of 146, although the efficiency of conversion is low, requiring a high concentration of enzyme and a long incubation time to achieve significant turnover.

1

0.8

0.6

0.4 normalized

integration at 267 nm 0.2

0 0 200 400 600 time (min)

Figure 4-17. Time course of Tyl1a (100 μM)-catalyzed reaction using 146 (1 mM) as substrate. Integrations of the HPLC peaks of substrate, product, and degradation products were plotted versus time. (○) represents 146, (●) 152, and (□) 149.

TylB Substrate Specificity. Incubation of 145 with Tyl1a (2.85 μM) and TylB (28.5 μM) was also carried out to test whether TylB could accept 150 as a substrate. As expected, HPLC analysis of the incubation mixture showed substrate depletion as well as formation of the degradation products, TDP and 151. Interestingly, the time-dependent formation of a new peak at 9.5 min was also observed. The conversion of 145 to this new

154 species was less than 2% as estimated from peak integration. Inclusion of less Tyl1a (1 μM) and more TylB (50 μM) along with a longer incubation time (12 h) resulted in a 15% overall conversion (Figure 4-14, trace c). In view of the effect of the increased TylB

concentration, this new product is likely TDP-3-amino-2,3,6-trideoxy-D-glucose (147, Figure 4-13). High resolution ESI-MS data of this product purified by HPLC is consistent with the assigned structure (147). These results strongly suggest that TylB is capable of converting 150 to 147 in spite of the inherent instability of 150. Encouraged by the observation that Tyl1a was able to convert 150 to 147, we

subsequently incubated CDP-4-keto-6-deoxy-D-glucose (146) (1 mM) with Tyl1a (35 μM) and TylB (35 μM) to test the ability of TylB to accept 152 as a substrate. As expected, we observed slow substrate depletion and concomitant formation of Tyl1a product 152. However, we also observed formation of a new peak at 11.4 min. Incubation of the reaction mixture for 32 h led to 34% conversion of the starting material to this new product (Figure 4-14, trace f). Based on the dependence of product formation on the presence of TylB, this compound is likely to be CDP-3-amino-3,6-dideoxy-D- glucose (148, Figure 4-16). High resolution ESI-MS data of this product purified by HPLC is consistent with the assigned structure (148). These results strongly suggest that TylB is capable of converting 152 to 148. Discussion of the Genomics of Hexose Ketoisomerases. At the time when the

function of FdtA from A. thermoaerophilus was first verified, fewer than 10 ORFs encoding homologous proteins had been identified. There are currently at least 65 ORFs exhibiting homology to fdtA in the NCBI database, the vast majority of which were uncovered as part of whole genome sequencing projects. These ORFs exist exclusively in bacteria, and are often clustered with genes proposed to be involved in outer membrane polysaccharide biosynthesis. A significant portion (18%) of these

155 homologues are found to encode the C-terminal domains of putative bifunctional sugar ketoisomerase/N-acetyltransferases. One example of this type of bifunctional enzyme, WxcM from Xanthomonas campestris pv. campestris, has been shown through genetic studies to be involved in lipopolysaccharide (LPS) biosynthesis.150 Interestingly, tyl1a and the Spi seq 25 gene from the spiramycin biosynthetic cluster of Streptomyces ambofaciens are the only two homologues found in natural product biosynthetic gene clusters. The structure of spiramycin and the organization of its biosynthetic cluster bear significant similarity to those of tylosin. Since a mycaminose moiety is also present in the structure of spiramycin, the protein encoded by Spi seq 25 likely serves an analogous role to Tyl1a in mycaminose formation in S. ambofaciens. It is worth noting that most fdtA/tyl1a homologous are found in close proximity to genes encoding sugar aminotransferases, with the two often being co-transcribed. In view of the instability of the Tyl1a product, this close linkage between 3,4-ketoisomerase and aminotransferase may be advantageous, allowing coordinate expression of these two genes in order to minimize degradation of the unstable 3-keto sugar product.

Implications for Tyl1a Mechanism and Structure. The mechanism for the 3,4- ketoisomerization catalyzed by both Tyl1a and FdtA could proceed with deprotonation at C-3 of 14 to form an enediol (or enediolate) intermediate (i.e., 154 in Figure 4-18) followed by reprotonation at C-4 to give the 3-keto product (38 or 153). Protein fold and structure analysis of Tyl1a and FdtA using the Phyre program predicts that these ketoisomerases belong to the RmlC-like cupin superfamily.151 Interestingly, RmlC, the

TDP-4-keto-6-deoxy-D-glucose-3,5-epimerase involved in L-rhamnose biosynthesis, is a dimeric protein which also processes 14. In fact, many members in this superfamily are NDP-4-ketohexose epimerases responsible for inversion of the C-3 and/or C-5 centers of their substrates. The epimerization catalyzed by these enzymes is thought to involve two

156 sequential cycles of abstraction of the proton α to the keto group followed by reprotonation at the opposite face of the enolate intermediate.152 Thus, there are apparent parallels between the mechanism of RmlC and that proposed for Tyl1a. It is also

interesting to note that Tyl1a and FdtA use the same substrate, TDP-4-keto-6-deoxy-D- glucose (14), yet form products 38 and 153 (Figure 4-18), respectively, which are C-4 epimers of each other. This observation suggests that deoxysugar products with opposite C-4 stereochemistry can be produced by replacing Tyl1a with FdtA (or vice versa). Mechanistically, the stereospecificity of Tyl1a- and FdtA-catalyzed reactions may be determined by the position of the proton donating residue in the active site relative to C-4 of the sugar substrate. Further studies are required to examine the predicted mechanistic and structural similarities between Tyl1a and RmlC-like epimerases, and to decipher the molecular basis for the difference in stereospecificity of Tyl1a- and FdtA-catalyzed reactions.

Me H+ HO Me O Tyl1a HO O H+ OH O OH O Me H O OTDP OTDP O 154 38 HO OH Me H OTDP HO HO Me 14 O O H+ FdtA OH O OH H O OTDP OTDP 154 153

Figure 4-18

Biosynthesis of 3-amino-2,3,6-trideoxyhexoses. The presence of 3-amino-2,3,6- trideoxyhexoses in a number of bioactive natural products prompted us to investigate the ability of Tyl1a and TylB to process 2-deoxysugar substrates. The only example of an NDP-3-amino-2,3,6-trideoxyhexose whose biosynthesis has been fully characterized is 157 TDP-L-eremosamine (56) from the chloroeremomycin pathway of Amycolatopsis orientalis.55 This pathway contains a specific aminotransferase, EvaB, capable of acting

on C-3 of the highly unstable intermediate TDP-3,4-diketo-2,6-dideoxy-D-glucose (155)

to give TDP-3-amino-4-keto-2,3,6-trideoxy-D-glucose (156), which may serve as a general precursor for 3-amino-2,3,6-trdeoxyhexoses (Figure 4-19, route A). In the case of TDP-L-eremosamine (56), compound 156 is further modified by a methyltransferase and an epimerase before reduction of the C-4 keto group. Having verified the function of Tyl1a, we envisioned that a pathway including a 3,4-ketoisomerization step may be an alternative biosynthetic route to 3-amino-2,3,6- trideoxyhexoses. This pathway, starting from 14, involves C-2 deoxygenation and subsequent C-3 ketoreduction to give 145, followed by 3,4-ketoisomerization and C-3 transamination catalyzed by homologues of Tyl1a and TylB, respectively, to give 147 (Figure 4-19, route B). The feasibility of such a pathway could be determined by examining the ability of Tyl1a and TylB to act on 2-deoxysugar substrates. The fact that

we were able to demonstrate the formation of 147 enzymatically from 14 via a 14 → 155 → 145 → 150 → 147 route in vitro suggests that this pathway is plausible. However, due to the instability of the 3-keto-2,6-dideoxy sugar 150, an excess of TylB relative to Tyl1a is required to yield appreciable amounts of 147. Since each gene cluster encoding production of 3-amino-2,3,6-trideoxyhexoses discovered thus far possesses an aminotransferase gene closely related to that encoding EvaB, it seems that nature has adopted a more efficient route to make 3-amino-2,3,6-trideoxyhexoses by evolving an aminotransferase that captures the unstable intermediate 155 (route A) rather than one that captures 150 (route B).

Nucleotide Specificity of Sugar Biosynthetic Enzymes. While most deoxysugars used in the biosynthesis of secondary metabolites are TDP-derivatives, deoxysugar

158 O Me Me O Me O OTDP H2N HO NH2 3-aminotransfer OTDP 156 56 Me O Me Route A O 2,3-dehydration O O HO OH 14 OTDP O OTDP 155 Route B O Me Me Me O 3-ketoreduction O 3,4-isomerization HO 3-aminotransfer HO O HO H2N O OTDP OTDP 145 OTDP 150 147 Figure 4-19

structures bearing CDP-, GDP-, and UDP- groups are not uncommon. Studies

demonstrating the ability of the sugar thymidylyltransferase RmlA (Ep) and mannose-1- phosphate guanylyltransferase ManC to use NTPs other than their natural substrates,153,

154 as well as studies demonstrating the ability of GTs OleD, FucT-III, and LgtA to use various NDP-activated sugars as donor substrates,116, 155 have been performed. However, few studies of this type have been conducted, and none involving deoxysugar biosynthetic enzymes other than nucleotidylyltransferases and GTs have been performed. Identification of deoxysugar biosynthetic enzymes having relaxed specificity with respect to the nucleoside portion of their NDP-sugar substrates would be useful for preparing NDP-activated sugars. Here, we demonstrated that Tyl1a is capable of turnover of 146, the CDP version of its natural substrate, but at a significantly reduced rate (about 400- fold slower than 14), requiring high concentrations of enzyme for significant product formation. However, large amounts of Tyl1a are readily obtainable so that in vitro synthesis of CDP-sugars involving a 3,4-ketoisomerization reaction in their biosynthesis may be feasible using Tyl1a. TylB is also capable of turnover of a CDP-activated substrate. More extensive testing of deoxysugar biosynthetic enzymes for NDP promiscuity will be necessary to assess the synthetic feasibility of this approach. As the number of X-ray crystal structures of deoxysugar biosynthetic enzymes increases, more sophisticated approaches involving mutagenesis may be employed to generate NDP-

159 promiscuous enzymes for use in the construction of a variety of NDP-activated deoxysugars. Another interesting observation about Tyl1a-catalyzed turnover of the CDP- activated substrate 146 was that the rates of its isomerization to 152 and the decomposition of 152 to 149 were slower than, but roughly proportional to the rates of the isomerization of 14 to 38 and the decomposition of 38 to 149. This suggests either that 152 is somehow more stable then its TDP-activated counterpart 38, or that the decomposition of these 3-keto intermediates is actually catalyzed by Tyl1a, with the rate of decomposition being related to the ability of the 3-ketosugar product (152 or 38) to re- bind in the Tyl1a active site. While this issue was not investigated further, it is more likely that Tyl1a does indeed catalyze decomposition of the 3-ketosugar product.

However, earlier observation of the slow degradation of chemically synthesized 3846 suggests that degradation by both non-enzymatic and enzymatic processes does occur, with Tyl1a enhancing the rate of degradation.

4. CONCLUSION

The work described herein has established the function of Tyl1a as the TDP-4-

keto-6-deoxy-D-glucose-3,4-ketoisomerase in the mycaminose biosynthetic pathway in S. fradiae. Tyl1a is only the second example of an enzyme from this newly discovered class to be characterized in vitro. The biochemical characterization of Tyl1a has significant implications for the correct functional assignment of its many uncharacterized homologues, and for future investigation of the structure and mechanism of this group of enzymes. The identification and in vitro characterization of Tyl1a and the successful reconstitution of the reactions by Tyl1a and TylB to convert 14 to 53 make possible the

enzymatic preparation of TDP-3-amino-3,6-dideoxyhexoses, such as 53 and TDP-D- 160 mycaminose (52). The availability of Tyl1a and its demonstrated substrate flexibility are important for the in vitro and in vivo preparation of a variety of NDP-deoxysugars and thus, the glycodiversification of natural products.

161 Chapter 5. Construction of Non-Natural Sugar Biosynthesis Pathways in Engineered S. venezuelae Hosts

1. INTRODUCTION

The realization that deoxysugars are essential for the activities of many important biomolecules such as natural products5-7, 127 and cell surface polysaccharides,4, 28 has led to the recent surge of investigation of the biosynthesis of deoxysugars at the molecular level.9, 135 These studies have in turn facilitated rational manipulation of deoxysugar biosynthetic pathways to generate new antibiotic structures with potential clinical applications. The success of these engineering strategies hinges on the identification and exploitation of the substrate flexibilities of pathway enzymes. The substrate flexibilities of several natural product glycosyltransferases (GTs) have recently been demonstrated and subsequently exploited for the synthesis of novel drug analogues.10, 156, 157 The in vitro investigations of unusual sugar biosynthesis and work on the construction of hybrid deoxysugar biosynthetic pathways in engineered hosts have also brought to light the fact that substrate flexibility appears to be a general trait for many sugar biosynthetic enzymes.10, 51, 53, 157 One can envision using the inherent substrate flexibility of particular sugar biosynthesis enzymes for the construction of engineered pathways capable of

biosynthesizing sugar structures that have not yet been observed in nature, thus increasing the diversity of sugar structures available for natural product glycodiversification.

Specifically, the goal of work described in this chapter is to construct engineered biosynthetic routes which to make non-natural sugar-bearing macrolide derivatives in

vivo. The D-desosamine (29)-bearing macrolide antibiotics methymycin (30), neomethymycin (37), and pikromycin (32) produced by S. venezuelae (Figure 5-1),

162 O O Me Me R1 HO Me Me Me Me NMe2 NMe O 2 HO HO R Me 2 O O O Me O O O Me Me Me 29 Me 29 O O 30 R1 = OH, R2 = H Me 37 R1 = H, R2 = OH 32

NMe2 OH HO OH HO OH O O O Me O Me 49 157 Figure 5-1

whose biosynthesis has been described in detail in previous chapters, comprise an important class of compounds which are effective antibacterial agents. As described in Chapters 2 and 3, combinations of genes from the mycaminose biosynthetic pathway have been expressed in several mutants of S. venezuelae, resulting in the formation of new macrolide derivatives bearing D-quinovose (157) and D-mycaminose (49) moieties.

These results demonstrate the feasibility of re-routing the D-desosamine (29) pathway to make other sugars which can be attached to the endogenous macrolactones by the promiscuous GT DesVII in S. venezuelae. In this chapter, three biosynthetic routes designed to synthesize new macrolides each bearing an unusual sugar (158-160, Figure 5-2) were constructed in mutants of S. venezuelae. These routes were assembled using enzymes imported from several sugar

pathways, including the TDP-D-forosamine (55) pathway of Saccharopolyspora spinosa,

29 the TDP-L-mycarose (54)52 and TDP-D-mycaminose (52) pathways of S. fradiae, the

TDP-L-megosamine (161) pathway of Micromonospora megalomicea,158 and the TDP-3-

N-acetylamino-6-deoxy-D-galactose (162) pathway of A. thermoaerophilus130 (Figure 5-

163 3). In each case, one or more enzymes in the engineered pathway are required to display substrate promiscuity in order for the engineered pathway to biosynthesize its target.

O Me O HO OH KdesII/pCM42 14 OTDP KdesI/pCM45

FdtA DesI KdesI/pCM43 HO Me TylX3 Me O O H2N HO Me OH O OH O OTDP 48 OTDP O 173

O OTDP DesV SpnS 165

HO Me Me MegDII O Me N O 2 H2N HO OH OH O Me OTDP 163 OTDP O H N 2 DesVI DesVII 166 OTDP DesVIII reductase HO Me Me O O Me2N O Me Me2N HO OH OH HO O OTDP 158 H2N OTDP DesVII DesVIII TylM1 HO Me Me O O HO O Me2N OH Me2N 160 167 OTDP DesVII DesVIII

Me HO O O Me2N 159

Figure 5-2

Specifically, the first example, the S. venezuelae mutant KdesII/pCM42, was designed to synthesize macrolides bearing the non-natural sugar 4-N,N-dimethylamino-

4,6-dideoxy-D-glucose (158) by halting the desosamine pathway at TDP-4-amino-4,6- dideoxy-D-glucose (48) and expressing the 4-N,N-dimethyltransferase SpnS from the 164 forosamine pathway, which must convert 48 to TDP-4-N,N-dimethylamino-4,6-dideoxy-

D-glucose (163). SpnS normally catalyzes 4-N,N-dimethyltransfer to TDP-4-amino-

2,3,4,6-tetradeoxy-D-glucose (164), which differs from 48 by the lack of hydroxyl groups at C-2 and C-3 (Figure 5-3). Therefore, the ability of SpnS to accept a 2,3-hydroxy analogue of its natural substrate, as well as the ability of the GT DesVII to recognize 163, are being tested in KdesII/pCM42.

Me Me HO O TylM1 HO O H2N Me2N HO HO OTDP OTDP 53 52

HO HO Me Me FdtA O O AcHN OH HO O OTDP OTDP 173 162

O Me Me SpnS Me O O O H2N Me2N HO OH 14 OTDP 164 OTDP 55 OTDP

Me OH TylX3 O O Me O OTDP HO O OTDP Me 165 54 MegDII NMe2 O Me Me OTDP O O

H2N HO 166 OTDP 161

Figure 5-3

The second example, KdesI/pCM43, was designed to synthesize macrolides

bearing D-angolosamine (159). While this sugar does occur naturally as a constituent of the natural products medermycin,159 hedmaycin,160 and others, the engineered biosynthesis of 159 in KdesI/pCM43 was to be performed without using genes from the known angolosamine pathways. The plasmid pCM43 contains the genes tylX3, which 165 encodes a 2,3-dehydratase converting 14 to TDP-6-deoxy-3,4-diketo-D-glucose (165); megDII, which is predicted to encode a 3-aminotransferase converting 165 to TDP-4-

keto-3-amino-2,3,6-trideoxy-D-glucose (166); and tylM1, which encodes a 3-N,N- dimethyltransferase converting 53 to 52. In order for this pathway to synthesize macrolides bearing 159, 14 that accumulates in the KdesI mutant must be converted to

166 by TylX3 and MegDII, and 166 must be converted to TDP-D-angolosamine (167) by tandem action of TylM1 and the unidentified endogenous 4-ketoreductase whose activity was observed in the KdesI S. venezuelae mutant. Conversion of 166 to 167 can be envisioned to occur with either the reductase or TylM1 acting first (Figure 5-4), yet both scenarios would require substrate promiscuity of the enzymes involved. Finally, the synthesis of macrolides bearing 159 would require the ability of DesVII to use 167 as a substrate in the glycosyltransfer reaction. Note that even if the unidentified 4- ketoreductase and/or TylM1 are incapable of turnover of non-natural substrates, DesVII may still be able to catalyze glycosyltransfer using accumulating sugar intermediates (166, 168, or 169 are all possible as substrates). Thus, novel glycosylated macrolides bearing sugars 170, 171, or 172, or even 3-N-acetylated derivatives of 170 or 172 similar to those observed in KdesII and KdesVI mutants, might be obtained (Figure 5-4).

The third example, KdesI/pCM45, was designed to synthesize macrolides bearing

the non-natural sugar 3-N,N-dimethylamino-3,6-dideoxy-D-galactose (4-epi-D- mycaminose, 160). As described in Chapter 3, the KdesI/pCM30 mutant efficiently produces macrolides bearing D-mycaminose (49), demonstrating that substitution of DesI

with the 3,4-ketoisomerase Tyl1a converts a D-desosamine pathway to a D-mycaminose pathway. The gene encoding Tyl1a was identified through its homology to FdtA, a 3,4-

ketoisomerase involved in the biosynthesis of TDP-3-N-acetylamino-3,6-dideoxy-D- galactose (162), which is part of S-layer polysaccharide formation in A.

166 DesVII O Me DesVIII O O H2N O Me 170 O HO DesVII OH O Me O Me 14 OTDP O DesVIII O O Me2N Me2N TylX3 168 OTDP 171

Me TylM1 reductase Me O Me Me DesVII O O MegDII O HO O DesVIII HO O Me N O H2N Me2N 2 O OTDP OTDP 167 OTDP 159 165 166 reductase TylM1

Me Me O HO HO O O H2N H N DesVII 2 169 OTDP DesVIII 172

Figure 5-4

thermoaerophilus.130 As mentioned in Chapter 4, FdtA uses the same substrate as Tyl1a,

14, yet converts it to the C-4 epimer of the Tyl1a product, TDP-3-keto-6-deoxy-D- galactose (173) (Figure 5-3). Therefore introduction of pCM45, which contains fdtA, into KdesI, is designed to divert 14 to 173. If DesV, DesVI, and DesVII are all capable of tolerating substrates with the opposite stereochemistry at C-4, the strain would produce macrolides bearing 160 (Figure 5-2). The relaxed substrate specificity of pathway enzymes is not the only prerequisite for the success of this engineered pathway. A potential obstacle is the differences in codon usage between S. venezuelae and A. thermoaerophilus. Actinomycetes, such as S. venezuelae, have GC-rich DNA (about 70% GC), whereas A. thermoaerophilus DNA is AT-rich (about 30% GC). This difference in codon usage leads to a high frequency of rare codons when A. thermoaerophilus DNA is expressed in S. venezuelae which could prevent expression. After creation of these engineered S. venezuelae mutants, TLC analysis of small- scale cultures of each revealed no obvious glycosylated macrolides in extracts from KdesII/pCM42 and KdesI/pCM43, but showed several prominent polar spots in extracts

167 from KdesI/pCM45. Large-scale cultures of KdesII/pCM42 and KdesI/pCM43 were then carried out and extracts from these cultures were subjected to separation and detailed analysis by TLC, 1H NMR, and mass spectrometry to identify any glycosylated compounds which might be present in minute quantities. This analysis revealed the

presence of 4-N-acetylamino-4,6-dideoxy-D-glucosylated derivatives of 10- deoxymethynolide (174) and narbonolide (175) and a minor compound 4-N-

monomethylamino-4,6-dideoxy-D-glucosyl narbonolide (176) in extracts of KdesII/pCM42, and quinovosyl 10-deoxymethynolide (143) and minor amounts of other quinovosylated compounds (177-180) in extracts of KdesI/pCM43 (Figure 5-5).

O O Me Me R1 Me Me Me OH Me OH O O HO NHAc HO OH R2 O O O O Me O O Me Me Me Me Me

174 143 R1 = H, R2 = H 177 R = OH, R = H O 1 2 178 R1 = H, R2 = OH Me O Me Me OH Me NHR Me HO R O O Me Me O Me OH Me HO OH O O Me O O O Me Me Me 175 R = Ac O O 176 R = Me Me 179 R = H KdesII/pCM42 180 R = OH KdesI/pCM43

Figure 5-5

Compounds bearing all of these sugars except 4-N-monomethylamino-4,6-

dideoxy-D-glucose had been found previously in KdesII and KdesI mutants, respectively, indicating that the engineered pathways were non-functional in these mutants. The presence of 176 was deemed inconclusive as will be discussed in the Results and Discussion section of the chapter. However, analysis of extracts from a large-scale 168 culture of KdesI/pCM45 revealed three major and at least seven minor glycosylated compounds. Separation and characterization of these compounds by mass spectrometry and NMR spectroscopy unambiguously demonstrated that the major compounds were 4- epi-D-mycaminose (160)-bearing methynolide, neomethynolide, and pikronolide derivatives (181, 182, and 183, respectively) (Figure 5-6). Interestingly, evidence obtained based on mass spectrometric fragmentation patterns of the minor compounds revealed that at least three of them bear the novel sugar 3-N-monomethylamino-6-deoxy-

D-galactose (184), indicating that alteration of C-4 stereochemistry perturbed the in vivo kinetics of one or more enzymes involved in this engineered pathway such that a small amount of TDP-3-N-monomethylamino-6-deoxy-D-galactose (185) is used by DesVII before being converted to TDP-4-epi-D-mycaminose (186) by DesVI (Figure 5-6).

O OH Me MeHN R1 OH HO Me Me Me2N O O Me O HO R 184 2 O O O Me Me Me 160 DesVII DesVII DesVIII 181 R1 = OH, R2 = H DesVIII 182 R1 = H, R2 = OH HO Me HO Me O O MeHN Me N OH 2 OTDP OH 185 186 OTDP DesVII O DesVIII Me HO Me Me OH Me2N Me HO O O O Me Me O O 160 Me 183 Figure 5-6

169 These above experiments were designed to test the practical limits of deoxysugar biosynthetic engineering, with each experiment simultaneously testing the promiscuity of several sugar pathway enzymes. The fact that only one of these three experiments was successful illustrates both the complex nature of these biosynthetic engineering endeavors and the value in stepwise establishment of the desired functions of enzymes in an engineered pathway. Nevertheless, the positive results obtained with the KdesI/pCM45 mutant showed that even engineering work whose feasibility depends on a stringent set of criteria can sometimes work, and provided useful information about the promiscuities of the enzymes used in this pathway. Both the failures and successes of this work are valuable for future sugar biosynthetic engineering work using the S. venezuelae system.

2. EXPERIMENTAL PROCEDURES

General. All materials used for work described in this chapter have already been mentioned in the Experimental Procedures section of Chapter 2 with the exception of Saccharopolyspora spinosa genomic DNA for PCR amplification of spnS, which was obtained from Dr. Hyung-jin Kwon of this laboratory. Plasmids and Vectors. All plasmids and vectors used for work described in this chapter have been mentioned in the Experimental Procedures section of Chapter 2. pHJL309 was used as the template for PCR amplification of tylM1 and pSET552 was used as the template for amplification of tylX3. Bacterial Strains. Most bacterial strains used for work described in this chapter have been mentioned in the Experimental Procedures section of Chapter 2. Micromonospora megalomicea subsp. nigra (ATCC 27598) was obtained as a freeze- dried sample from the American Type Culture Collection (Manassas, VA).

170 Aneurinibacillus thermoaerophilus (ATCC 700303) was obtained as a freeze-dried sample from Dr. Paul Messner of the University of Natural Resources and Applied Life Sciences, Vienna, Austria. Instrumentation. All instrumentation used for work described in this chapter has been mentioned in the Experimental Procedures section of Chapter 2. Preparation of Competent Cells. The procedure used to prepare E. coli competent cells was described in the Experimental Procedures section of Chapter 2. Growth and Maintenance of M. megalomicea. M. megalomicea was grown at 29 °C and maintained using ATCC medium #172 (either as a liquid or solid), which were prepared as follows: A mixture of 10 g of glucose, 20 g of soluble starch, 5 g of yeast

extract, 5 g of N-Z amine, and 1 g of CaCO3 were dissolved in 1 L of deionized water and the mixture was autoclaved to make liquid media. Solid media was prepared similarly, but with the addition of 15 g agar. Genomic DNA was isolated from a culture of M. megalomicea grown in media #172 at 29 °C with shaking at 250 rpm for 7 days according to the procedure given in Kieser et al.114 A spore suspension of M. megalomicea was prepared for long-term storage by growth on #172 agar for 1 week at 29 °C. During growth, cells changed color from orange to black, and were almost exclusively black after one week. The surface of the plate was overlayed with a sterile 20% glycerol solution and spores were gently suspended in the solution by rubbing the

surface of the plate with a sterile cotton swab. Spore suspension was stored at -80 °C. Growth and Maintenance of A. thermoaerophilus. A. thermoaerophilus was grown at 55 °C and maintained in Nutrient Broth or Nutrient Agar, which were prepared as follows: 3 g of beef extract and 5 g of peptone were dissolved in 1 L of deionized water and autoclaved. Solid media was prepared similarly, but with the addition of 15 g agar. A cell stock for long-term storage was prepared in the following manner. Cells

171 were grown on Nutrient Agar overnight and a single colony was used to inoculate a 50 mL Nutrient Broth culture in a 125 mL Erlenmeyer flask, which was grown at 55 °C with shaking at 125 rpm overnight. A 1 mL aliquot of the overnight culture was used to inoculate a prewarmed 50 mL Nutrient Broth culture in a 125 mL Erlenmeyer flask. This culture was grown at 55 °C with shaking at 125 rpm until the OD600 reached 0.6-1.0. Cells were collected by centrifugation, resuspended in 1 mL of sterile 80% glycerol water, and stored at -80 °C. Colony PCR using A. thermoaerophilus. Several colonies picked from a Nutrient Agar plate were suspended in 20 mM NaOH and the mixture was incubated at 37 °C for 1 h. The mixture was then centrifuged, the supernatant discarded, and the precipitated cell debris containing genomic DNA kept. The PCR reaction mixture was added directly to the tube containing the precipitate and mixed well before the PCR reaction.

PCR Amplification of DNA. The design of oligonucleotide primers and the general procedure for PCR amplification of DNA fragments was described in the Experimental Procedures section of Chapter 2. Construction of Expression Plasmid pCM1d. The construction of pCM1d, which was used as the precursor for pCM42, pCM43, and pCM45 had been described in the Experimental Procedures section of Chapter 3. pCM42. The selected genes used for the construction of the expression plasmids

pCM42 and pCM43 were amplified containing their native Shine-Dalgarno sequences. pCM42 was constructed by amplifying a 794 bp fragment containing spnS with introduced upstream EcoRI and downstream NdeI restriction sites using the start primer spnS-E-up, 5’-GCCGAATTCACCACTCCGAAAGTG-3’ and halt primer spnS-N- down, 5’-GAATTCATATGGGGCTCAGTTGCGGATTC-3’ (where restriction sites are

172 shown in bold). This fragment was digested with EcoRI and NdeI and ligated into pCM1d digested with the same restriction enzymes to give pCM42 (Figure 5-7). pCM43. pCM43 was constructed first by amplifying three DNA fragments: a 1539 bp fragment containing tylX3 with introduced upstream EcoRI and downstream NdeI restriction sites using the start primer tylX3-E-up, 5’-GGCCGAATTCGTGCTCA- CTGACATC-3’ and halt primer tylX3-N-down, 5’- GGAATTCATATGGGCAGGAC- CGGATCAGTAGAG-3’; an 1196 bp fragment containing megDII with introduced upstream NdeI and downstream BamHI restriction sites using the start primer megDII-N- up, 5’-GGAATTCATATGGCCGGTAAGAAGGACGGTG-3’ and halt primer megDII- B-down, 5’-GCGCGGATCCTCTTCGCTGAC-3’; and an 861 bp fragment containing tylM1 with introduced upstream BamHI and downstream XbaI sites using the start primer tylM1-B-up and halt primer tylM1-X-down, which were identical to those used in amplification of the same tylM1-containing fragment for the construction of pCM23. All introduced restriction sites in the primers are shown in bold. Next, each of these fragments was digested with the appropriate restriction enzymes and sequentially cloned into pCM1d digested with the same restriction enzymes. Insertion of all three fragments resulted in the assembly of pCM43 (Figure 5-7).

173 pCM45. pCM45 was constructed by amplifying a 465 bp fragment containing fdtA with introduced upstream EcoRI and ribosome binding site (RBS) and downstream HindIII sites using the start primer fdtA-E-RBS-up, 5’- ATATGAATTCGAGGAGCAA- AACTATGGAAAATAAAGTTATTAACTTC-3’ and halt primer fdtA-H-down, 5’-CG- GCAAGCTTCGAAATTACAGTTTATCCTTC-3’. Restriction sites are shown in bold and the engineered RBS is shown in italics. The published sequence containing fdtA does not include the sequence upstream of the gene where the RBS would be, so an RBS identical to that present upstream of the native tyl1a sequence was engineered into fdtA- E-RBS-up seven nucleotides upstream of the start codon (shown underlined in the primer sequence). The sequence between the RBS and start codon was chosen based on its inability to cause stable secondary structures, homodimers, or heterdimers in fdtA-E-RBS-up and between the primer pair. This fragment was digested with EcoRI and HindIII and ligated into pCM1d digested with the same restriction enzymes to give pCM45 (Figure 5-8).

Construction of KdesI-80 S. venezuelae. KdesI-80 S. venezuelae mutant was constructed by Dr. Svetlana Borisova of this group, and its construction has been reported.41 Construction of KdesII S. venezuelae. KdesII S. venezuelae mutant was constructed by Dr. Svetlana Borisova of this group, and its construction has been reported.40

174 Conjugal Transfer of Expression Plasmids into KdesI and KdesII S. venezuelae. pCM42 was transferred to KdesII S. venezuelae, and pCM43 and pCM45 were transferred to KdesI S. venezuelae. The KdesII mutant, like KdesI-80, is kanamycin resistant. The transfer of each of these plasmids into the appropriate S. venezuelae host was performed based on an identical procedure used in the transfer of pCM1 derivatives into KdesI which are described in the Experimental Procedures section of Chapters 2 and 3.

Preparation of Spore Suspensions and Frozen Mycelia for S. venezuelae Strains. These experiments were performed as described in the Experimental Procedures section of Chapter 2. Small-scale Isolation and Analysis of Metabolites Produced by Mutants. These experiments were performed as described in the Experimental Procedures section of Chapter 2.

Mass spectrometric data for 4-N-acetyl-6-deoxy-D-glucosyl 10-deoxymethynolide

(174), 4-N-acetyl-6-deoxy-D-glucosyl narbonolide (175), 4-N-monomethylamino-6-

deoxy-D-glucosyl narbonolide (176) partially purified from extracts of KdesI/pCM42:

+ High resolution CI-MS of 174: C25H42NO8 (M + H) calculated 484.2910, found

+ 484.2905; high resolution CI-MS of 175: C28H46NO9 (M + H) calculated 540.3173,

+ found 540.3158; high resolution CI-MS of 176: C27H46NO8 (M + H) calculated 512.3223, found 512.3219. Mass spectrometric data for quinovosyl 10-deoxymethynolide (143), quinovosyl narbonolide (187), quinovosyl methynolide/neomethynolide (188/189), and quinovosyl

pikronolide (190) isolated from KdesI/pCM43: High resolution CI-MS of 143: C23H39O8

+ (M + H) calculated 443.2645, found 443.2640; high resolution CI-MS of 187: C26H42O9 (M + H)+ calculated 498.2829, found 498.2831; high resolution CI-MS of 188/189:

175 + C23H38O9 (M + H) calculated 458.2516, found 458.2515; high resolution CI-MS of 190:

+ C26H42O10 (M + H) calculated 514.2778, found 514.2758.

Mass spectrometric data for 3-O-(4-epi-D-mycaminosyl) methynolide (181), 3-O-

(4-epi-D-mycaminosyl) pikronolide (183), 3-O-(4-epi-D-mycaminosyl) novamethynolide

(191), 3-O-(4-epi-D-mycaminosyl) narbonolide (192), 3-O-(4-epi-D-mycaminosyl) 10-

deoxymethynolide (193), 3-O-(3-N-monomethylamino-6-deoxy-D-galactosyl)

novamethynolide (194), 3-O-(3-N-monomethylamino-6-deoxy-D-galactosyl) methynolide

(195), and 3-O-(3-N-monomethylamino-6-deoxy-D-galactosyl) neomethynolide (196)

+ isolated from KdesI/pCM45: High resolution CI-MS of 181: C25H44NO8 (M + H)

calculated 486.3067, found 486.3065; high resolution CI-MS of 183: C28H44NO9 (M +

+ H) calculated 542.3329, found 542.3323; high resolution CI-MS of 191: C25H44NO9 (M

+ + H) calculated 502.3016, found 502.3018; high resolution CI-MS of 192: C28H48NO8 (M + H)+ calculated 526.3380, found 526.3375; high resolution CI-MS of 193:

+ C25H44NO7 (M + H) calculated 470.3118, found 470.3115; high resolution CI-MS of

+ 194: C24H41NO9 (M + H) calculated 488.2860, found 488.2867; high resolution CI-MS

+ of 195: C24H42NO8 (M + H) calculated 472.2910, found 472.2913; high resolution CI-

+ MS of 196: C24H42NO8 (M + H) calculated 472.2910, found 472.2914. Large-Scale Isolation and Analysis of Extracts of KdesII/pCM42, KdesI/pCM43, and KdesI/pCM45 S. venezuelae Mutants Grown in Vegetative Media. These experiments were performed nearly identically to those of other large-scale isolation experiments described in Chapters 2 and 3. Cultures of each mutant were grown in 150 mL of seed media containing the appropriate antibiotics at 29 °C, 48 h with shaking at 250 rpm in a rotary shaker. For each mutant, 20 mL of seed culture was used to inoculate each of six 1 L flasks of vegetative media, which were grown at 29 °C with shaking at 250 rpm for 48 h. Culture broth was obtained by removal of cells and insoluble

176 components as described previously for small-scale isolation. Compounds were extracted from the aqueous layer three times using an equal volume of chloroform each time. After evaporation of solvent, compounds from KdesII/pCM42 and KdesI/pCM43 were separated by silica gel flash chromatography using a gradient of 0-20% methanol in chloroform and analyzed by TLC and 1H NMR. Separation of compounds isolated from KdesI/pCM45 was performed by silica gel flash chromatography using a gradient of 0- 40% methanol in chloroform, and fractions were analyzed by TLC and 1H NMR. Fractions obtained from silica gel chromatography of KdesI/pCM43 were further purified by preparative scale HPLC on a C18 column by isocratic elution using 30% or 40% acetonitrile in 57 mM aqueous ammonium acetate buffer. The more polar fractions of interest obtained from silica gel chromatography of KdesI/pCM45 were purified by

preparative scale HPLC on a C18 column by isocratic elution using 28% acetonitrile in 57 mM aqueous ammonium acetate buffer. The less polar fractions of interest obtained from silica gel chromatography of KdesI/pCM45 were purified by preparative scale HPLC on a C18 column by the following program (A = 57 mM aqueous ammonium acetate buffer, B = acetonitrile): isocratic elution using 42% B from 0-20 min, then linear gradient elution using 42-60% B from 20-35 min, then isocratic elution using 60% B from 35-40 min, followed by 100% B wash and re-equilibration. Compounds 143, 174, 175, 181, and 183 were partially purified and analyzed by 1H NMR. Compound 182 was purified

and its structure determined by 1H, 13C, HSQC, HMBC, COSY, and NOESY spectroscopies.

Spectral data for 3-O-(4-epi-D-mycaminosyl) neomethynolide (182): 1H NMR

(500 MHz, CDCl3) δ 6.72 (1H, dd, J = 15.9, 5.3, 9-H), 6.40 (1H, dd, J = 15.9, 1.0, 8-H), 4.77 (1H, dd, J = 8.9, 2.0, 11-H), 4.27 (1H, d, J = 7.5, 1'-H), 4.14 (1H, br d, J = 2.5, 4’- H), 3.96 (1H, dd, 11.0, 7.5, 2’-H), 3.85 (1H, dq, J = 8.9, 6.3, 12-H), 3.59 (1H, br d, J =

177 10.2, 3-H), 3.52 (1H, br q, J = 6.5, 5'-H), 3.14 (1H, dd, J = 11.0, 2.5, 3’-H), 3.03 (1H, m, 10-H), 2.97 (6H, s, N-Me), 2.79 (1H, dq, J = 10.2, 7.1, 2-H), 2.49 (1H, ddq, J = 12.7, 7.1, 4.1, 6-H), 1.57 (1H, br t, J = 12.7, 5a-H), 1.37 (1H, br dt, J = 12.7, 4.1, 5b-H), 1.30 (3H, d, J = 7.1, 2-Me), 1.21 (1H, m, 4-H), 1.21 (3H, d, J = 6.5, 5’-Me), 1.15 (3H, d, J = 7.1, 6- Me), 1.15 (3H, d, J = 6.3, 12-Me), 1.12 (3H, d, J = 7.0, 10-Me), 0.96 (3H, d, J = 7.0, 4-

13 Me). C NMR (125 MHz, CDCl3) δ 205.0 (C-7), 174.4 (C-1), 147.4 (C-9), 126.0 (C-8), 103.9 (C-1'), 85.6 (C-3), 75.7 (C-11), 71.8 (C-5'), 68.3 (C-4'), 68.0 (C-3'), 67.4 (C-2),

66.0 (C-12), 45.0 (C-6), 43.5 (C-2), 42.5 × 2 (N-Me), 35.4 (C-10), 34.1 (C-5), 33.5 (C-4), 20.8 (C-12-Me), 17.6 (C-6-Me), 17.3 (C-4-Me), 16.0 (C-2-Me), 15.8 (C-5’-Me), 9.7 (C-

+ 10-Me). High resolution CI-MS: C25H44NO8 (M + H) calculated 486.3067, found 486.3065.

3. RESULTS AND DISCUSSION

Design, Construction, and Analysis of KdesII/pCM42. SpnS is the 4-N,N- dimethyltransferase which catalyzes the final step in the biosynthesis of the highly

deoxygenated sugar TDP-D-forosamine (55) as part of the formation of the spinosyn family of natural products in S. spinosa. The entire pathway for the biosynthesis of 55 has recently been characterized in vitro54 and the function of SpnS has been confirmed as part of this work. SpnS is the only sugar 4-N,N-dimethyltransferase to have been identified, making it a potentially useful tool in biosynthetic engineering work. Previous work on the D-desosamine biosynthetic pathway established that DesI catalyzes 4- aminotransfer using 14 as a substrate, forming 48. The KdesII mutant of S. venezuelae is thought to accumulate the DesI product 48 which is acetylated at the C-4 amine by an unknown endogenous acetyltransferase, resulting in the formation of 197. This compound can be used as a donor substrate by the GT DesVII, resulting in the formation 178 of 174 and 175 by the KdesII mutant. Therefore, the KdesII mutant presents the opportunity for testing the substrate flexibility of SpnS through heterologous expression of spnS. If SpnS is capable of 4-N,N-dimethyltransfer using 48 as a substrate, and if DesVII is capable of using the resulting non-natural TDP-sugar, TDP-4-N,N-

dimethylamino-4,6-dideoxy-D-glucose (163) as a substrate, novel macrolides bearing 4-

N,N-dimethylamino-4,6-dideoxy-D-glucose (158) would be produced (Figure 5-9). Compound 48 differs from the natural SpnS substrate 164 by the presence of C-2 and C-3 hydroxyl groups, but the stereochemistry of the C-4 amine functional group that is methylated in the SpnS-catalyzed reaction is preserved in 48, suggesting that it is possible for SpnS to use 48 as a substrate. O Me O HO HO OTDP 14 KdesII DesI KdesII/pCM42

Me Me Me acetyltransferase O SpnS O AcHN O H2N Me2N HO HO HO HO OH HO OTDP OTDP 197 OTDP 48 163 DesVII DesVII DesVIII DesVIII O Me Me O Me2N Me Me OH HO O O OH HO NHAc O O O Me 158 Me Me 174 O Me SpnS Me Me O O H2N Me2N Me Me OH NHAc OTDP OTDP Me HO 152 55 O O O Me Me O O Me 175 Figure 5-9

179 The expression plasmid pCM42 (Figure 5-7), which contains spnS, was constructed and conjugally transferred into the KdesII mutant, resulting in KdesII/pCM42. The crude metabolites of several individuals of this genotype were first examined by TLC, but no obvious new metabolites were produced by these individuals (Figure 5-10). In order to detect minor metabolites, 6 L of one individual of KdesII/pCM42 was grown in vegetative media, and the crude metabolites produced by this mutant were prepared. After separation by silica gel chromatography, a polar spot with an appropriate Rf value on TLC for a glycosylated macrolide was observed (Figure 5-10). 1H NMR analysis showed that fractions containing this spot actually contained 2 to 3 glycosylated macrolides. However, mass spectrometric analysis of this fraction

Figure 5-10. TLC analysis of metabolites produced by KdesII/pCM42. 1 = KdesII/pCM42 extracts from small-scale culture, 2 = KdesII/pCM42 large-scale crude extracts, 3 = KdesII/pCM42 silica gel chromatographic fractions containing glycosylated macrolides

showed that the major compounds present in this sample were the same 4-N-acetylamino-

6-deoxy-D-glucosyl derivatives of 10-deoxymethynolide (174) and narbonolide (175) that had been identified previously in the KdesII mutant. Interestingly, a minor component of

180 this fraction had elemental composition consistent with 4-N-monomethylamino-6-deoxy-

D-glucosyl narbonolide (176, Figure 5-11). Unfortunately, attempts to isolate this compound by reverse-phase HPLC failed. Thus, this result is inconclusive and requires O further experiments for verification. The Me presence of 174 and 175 suggests that if SpnS Me Me OH NHMe is capable of turnover of 48, this activity is Me HO O O O Me not very efficient. However, it is also Me O O Me possible that SpnS is capable of turnover of 176 48, but that DesVII is either inefficient at or Figure 5-11 incapable of turnover of 163.

Design, Construction, and Analysis of KdesI/pCM43. The sugar D-angolosamine (159), which is present in the structures of several natural products, is the 2-deoxy

analogue of D-mycaminose (49). Results described in Chapters 2 and 3 established that DesVII is quite efficient at transferring of 49 to its natural acceptor substrates, and is thus tolerant of structural modifications at C-4 of the donor substrate. Previous results with other S. venezuelae mutants demonstrated that DesVII is capable of accepting other non- natural donor sugars with alterations at C-3 and C-4 as substrates. However, sugars with changes at the C-2 position had not yet been tested or observed. Because 159 is identical to 49 except for the lack of the C-2 hydroxyl group, construction of a pathway to

biosynthesize 159 presented a unique opportunity to directly test the effect of removal of the C-2 hydroxyl group of the donor substrate on substrate recognition by DesVII. At the time of the design of this experiment, only one gene cluster containing genes for the

biosynthesis of TDP-D-angolosamine (167) had been identified. It is the medermycin cluster of Streptomyces sp. AM-7161.159 Based on genes present in the medermycin cluster, 167 was predicted to be biosynthesized in four steps from 14 (Figure 5-12). First

181 NMe2 NMe2 HO OH OH O O O Me O Me 49 159

Me O Me O Me O Med16 O O Med20 O

HO H2N OH O OTDP 14 OTDP 165 166 OTDP

Med14

Me Me HO O Med15 HO O Me2N H2N 167 OTDP 169 OTDP Figure 5-12

the 2,3-dehydratase Med16 and C-3 aminotransferase Med20 are expected to convert 14 to 166 via the unstable 3,4-diketo intermediate 165 in a manner identical to that observed

in TDP-L-vancosamine (56) biosynthesis.55 Next, the 4-ketoreductase Med14 is predicted to stereospecifically reduce the C-4 ketone to a hydroxyl group, resulting in formation of 169, which then undergoes 3-N,N-dimethyltransfer catalyzed by Med15 to form 167. Unfortunately, the medermycin-producing strain is not commercially available, precluding heterologous expression of the angolosamine biosynthetic enzymes. However, homologues of Med16 and Med20 which catalyze identical reactions have been characterized, and homologues of Med14 and Med15 which catalyze similar reactions are also known. Therefore, it might be possible to construct an artificial angolosamine biosynthetic pathway using genes from related pathways. If successful, this experiment would be a powerful demonstration of the use of pathway engineering to construct a sugar structure of interest. The genes tylX3, megDII, and tylM1 were chosen for the construction of the expression plasmid pCM43 (Figure 5-7), which when expressed in the KdesI S.

182 venezuelae mutant, is expected to produce 167. TylX3 is a homologue of Med16 from

the TDP-L-mycarose (54) biosynthetic pathway in the tylosin gene cluster of S. fradiae, and has been shown to catalyze the conversion of 14 to 165.49 MegDII is a homologue of

the proposed aminotransferase Med20 from the TDP-L-megosamine (161) biosynthetic pathway in the megalomicin (35) gene cluster of Micromonospora megalomicea.158 Although the function of MegDII has not been established experimentally, its homology to EvaB from the vancosamine biosynthetic pathway, which has been shown to catalyze conversion of 165 to 166, strongly suggests that MegDII catalyzes the identical reaction. Expression of these two genes alone in KdesI S. venezuelae may result in conversion of accumulated 14 to 166, which could potentially be a substrate for DesVII (Figure 5-13). However, once 166 is formed, it is likely that the endogenous reductase which stereospecifically reduces the C-4 ketone of 14 in the KdesI mutant to form TDP-D- quinovose (156), will reduce the corresponding ketone of 166, forming 169. The

substrate flexible 3-N,N-dimethyltransferase TylM1 from the D-mycaminose (49) pathway was chosen to catalyze the transformation of 169 to 167, the final step in the engineered angolosamine pathway. Although the two step conversion of 166 to 167 is predicted to occur with ketoreduction preceding methyltransfer, the latter two steps could also occur in the reverse order. The ability of this pathway to synthesize macrolides bearing 159 depends on the flexibilities of the endogenous 4-ketoreductase, TylM1, and

DesVII. However, if either the ketoreductase or TylM1 or both enzymes were unable to catalyze the requisite steps in the pathway, various other TDP-sugars, such as 166, 167, or 169, or N-acetylated versions of 166 or 169, might accumulate in KdesI/pCM43 (Figure 5-13). If DesVII were capable of utilizing one of these sugars, novel macrolides could be formed. Thus, there are a number of scenarios involving expression of the genes on pCM43 which might lead to a variety of macrolides bearing novel sugar

183 Me Me HO O TylM1 HO O H2N Me2N HO HO OTDP OTDP 53 52

Me Me O TylX3 OH O O O Me O OTDP HO OH HO 14 OTDP O OTDP Me 165 54 MegDII NMe2 O Me Me OTDP O O

H2N HO 166 OTDP 161

DesVII O Me DesVIII O O H2N O Me 170 O HO OH DesVII O Me O Me 14 OTDP O DesVIII O O Me2N Me2N TylX3 168 OTDP 171

Me TylM1 reductase Me O Me Me DesVII O O MegDII O HO O DesVIII HO O Me N O H2N Me2N 2 O OTDP 166 OTDP 167 OTDP 159 165 reductase TylM1

Me Me O HO HO O O H2N H N DesVII 2 169 OTDP DesVIII 172

Figure 5-13 structures. Any positive results from this experiment (i.e. formation of a new glycosylated macrolide) could be used to guide future work to construct additional engineered pathways using components from pCM43. After construction of pCM43 and its subsequent conjugal transfer into KdesI S. venezuelae, analysis of small-scale culture extracts from several individuals of this genotype failed to detect any obvious new spots on TLC which might correspond to new glycosylated macrolides (Figure 5-14). In order to look for minor compounds produced

184 by this mutant, a 6 L culture of one KdesI/pCM43 individual was grown in vegetative media and its crude metabolites extracted. Separation of the crude extracts by silica gel chromatography and analysis of fractions by 1H NMR showed the presence of at least three minor compounds which appeared to be glycosylated macrolides (Figure 5-14). Further purification of these compounds by preparative HPLC and analysis of each

compound by high resolution mass spectrometry and 1H NMR revealed the presence of at least four glycosylated macrolide derivatives. However, these compounds had elemental compositions consistent with quinovosyl derivatives of 10-deoxymethynolide (143), narbonolide (187), methynolide/neomethynolide (188/189), and pikronolide (190) (Figure 5-15). Compound 143 was previously identified in the extracts of the KdesI-80 mutant, but 187-190 are new compounds not previously observed. The presence of quinovosyl macrolides in the KdesI/pCM43 mutant implies that significant amounts of 14 were still accumulating in this mutant and suggests that an early step of the engineered pathway, most likely 2,3-dehydration by TylX3, was not operative.

Figure 5-14. TLC analysis of metabolites produced by KdesI/pCM43. 1 = KdesI-80, 2 = KdesI/pCM43 #1-5 small-scale extracts, 3 = KdesI/pCM43 large-scale silica gel chromatographic fractions containing 143, 4 = KdesI/pCM43 large-scale silica gel chromatographic fractions containing 188, 189.

185 O O Me Me R1 R Me Me Me Me OH OH O HO OH HO OH R2 O Me O O Me O O O Me Me Me Me O O 143 R1 = H, R2 = H Me 188 R1 = OH, R2 = H 187 R = H 189 R1 = H, R2 = OH 190 R = OH

Figure 5-15

The most obvious explanation for this result is that a mutation in the coding sequence of pCM43 may have been acquired during cloning which prevented expression of one of the genes. However, sequence analysis of pCM43 failed to show any deleterious mutations in the coding region of the construct, although the entire insert was not sequenced. Another possible explanation for this result comes from an observation about the sequence of the region upstream of tylX3. In Streptomyces gene clusters, the coding regions of genes which are co-transcribed are commonly found with little or no intervening sequence between the stop codon of the upstream gene and the start codon of the downstream gene. In the tylosin gene cluster, tylX3 and the gene directly upstream of it are separated by a 126 bp non-coding sequence. This sequence was amplified along with tylX3 during cloning in order to include any important regulatory elements in this region that might be important for expression of tylX3. However, after the results of the KdesI/pCM43 experiment were obtained, the possibility that this non-coding region may actually prevent expression of tylX3 was considered. Analysis of the sequence of this non-coding region using an RNA secondary structure prediction algorithm revealed the presence of strong secondary structure in this region. Although it is unclear what the function of this non-coding region might be, it is quite possible that, at least in the context

186 of heterologous expression in S. venezuelae, this region may have undesired effects on downstream gene expression. However, further experiments to resolve this issue were not performed. Interestingly, a recent report demonstrated the engineered biosynthesis of

novel macrolides bearing the 2-deoxysugar L-olivose (108) in S. venezuelae,161 demonstrating that DesVII is capable of accepting sugars lacking the C-2 hydroxyl group. It is therefore likely that DesVII would tolerate 167 as a substrate. Design, Construction, and Analysis of KdesI/pCM45. FdtA is a hexose 3,4- ketoisomerase involved in formation of TDP-3-N-acetylamino-3,6-dideoxy-D-galactose (162) as part of S-layer polysaccharide biosynthesis in A. thermoaerophilus. Interestingly, this enzyme uses 14, the same substrate as Tyl1a, but forms 173, which is the C-4 epimer of the Tyl1a product. Experiments with the KdesI/pCM30 mutant, which were described in Chapter 3, showed that substitution of DesI with Tyl1a converted a desosamine pathway to a mycaminose pathway. The success of KdesI/pCM30 provided a unique opportunity to test the feasibility of constructing a non-natural sugar biosynthetic pathway using FdtA. Expression of FdtA in KdesI S. venezuelae should lead to the formation of 173, which might then be acted upon by endogenous aminotransferase DesV and 3-N,N-dimethyltransferase DesVI. If both of these enzymes and the GT DesVII were capable of accepting substrates with an axial C-4 stereochemistry, macrolides bearing the non-natural sugar 4-epi-D-mycaminose (160) would be produced

(Figure 5-16). The expression plasmid pCM45 containing fdtA was constructed. The published sequence of fdtA did not include the native ribosome binding site, so one identical in sequence to that found upstream of tyl1a was engineered into the upstream primer. The overall GC content and codon usage of A. thermoaerophilus and S. venezuelae genes are markedly different, which might result in problems with expression of FdtA in S.

187 HO Me HO Me O Me O FdtA O DesV O HO H2N OH O OH OH 14 OTDP 173 OTDP OTDP DesVI

HO DesVII Me HO O DesVIII Me O O Me2N OH Me2N OH 160 OTDP

Figure 5-16 venezuelae. However, no codon optimization of fdtA was performed. After construction of pCM45 and its introduction into the KdesI mutant by conjugal transfer, analysis of small-scale extracts of several individuals revealed the presence of prominent highly polar compounds (Figure 5-17).

Figure 5-17. TLC analysis of metabolites produced by a small-scale culture of KdesI/pCM45.

188 In order to identify these compounds, a 3 L culture of KdesI/pCM45 was grown in vegetative media and the crude metabolites were extracted from the broth and separated by silica gel chromatography. TLC and 1H NMR analysis of the column fractions revealed the presence of three major and at least seven minor glycosylated macrolides accounting for about 60% of the total weight of macrolide compounds found in the extracts. High resolution mass spectrometry of partially purified compounds showed that these compounds have the same elemental compositions as the desired 4-epi-

D-mycaminosyl derivatives of the endogenous 12- and 14-membered ring macrolides. Further purification of these fractions by preparative HPLC afforded one of the major

compounds in pure form. This compound was structurally characterized using 1H-, 13C-, and several two dimensional NMR techniques as well as high resolution mass spectrometry and shown to be 3-O-(4-epi-D-mycaminosyl) neomethynolide (182). The

H-4’ signal of 182 was a broad doublet with J3,4 = 2.5 Hz and J4,5 ≈ 1 Hz, consistent with a stereo-relationships between the axial H-3’ and H-5’ protons and the equatorial H-4’ proton (Figure 5-18). Interestingly, comparison of the spectral data for the sugar moiety of 182 with those of the sugar moiety of 3-O-mycaminosyl methynolide (135) previously isolated from KdesI/pCM23 (Table 5-1) showed significant downfield shifts of the H-2’,

H-3’, H-4’, and H-5’ signals of the 4-epi-D-mycaminose moiety relative to the

Me OH H Me looking down looking down 11.0 Hz N 4' C4'-C3' bond C5'-C4' bond 7.5 Hz 2' H 3' Me HO OH OH O Me O O C2' N Me C5'-Me 5' 1' H H ' 6.5 Hz H4' 4 H H ~1 Hz C5' C3' 0 H3' ~70o H5' near 90 2.5 Hz

Figure 5-18 189 D-mycaminosyl methynolide 4-epi-D-mycaminosyl (135) neomethynolide (182) proton ppm shift splitting J (Hz) ppm shift splitting J (Hz) 1’-H 4.28 d 7.3 4.27 d 7.5 2’-H 3.51 dd 10.4, 7.3 3.96 dd 11.0, 7.5 3’-H 2.40 brt 9.4 3.14 dd 11.0, 2.5 3’ N-Me 2.52 s n/a 2.97 s n/a 4’-H 3.09 brt 9.4 4.14 brd 2.5 5’-H 3.29 dq 8.8, 6.1 3.52 brq 6.5 5’-Me 1.30 d 6.1 1.21 d 6.5

Table 5-1. Comparison of chemical shifts and coupling constants of sugar signals of 135 and 182.

corresponding signals of D-mycaminose. The other two major compounds were obtained

in 90% and 75% purity, respectively, and were identified as 3-O-(4-epi-D-mycaminosyl) methynolide (181) and 3-O-(4-epi-D-mycaminosyl) pikronolide (183) by high resolution mass spectrometry and comparison of their 1H NMR spectra to that of 182 (Figure 5-19).

O Me R1 OH Me Me MeHN O HO O O O Me O Me Me 184 Me R1 OH Me 195 R1 = OH, R2 = H Me Me2N O 194 R1 = H, R2 = OH HO R2 O 196 R1 = OH, R2 = OH O O Me Me Me 160 DesVII DesVII 193 R = H, R = H DesVIII DesVIII 1 2 181 R1 = OH, R2 = H HO 182 R = H, R = OH Me HO Me 1 2 O O 191 R1 = OH, R2 = OH MeHN Me N OH 2 OTDP OH 185 186 OTDP DesVII O DesVIII Me HO Me Me OH Me2N Me HO O O O Me Me O O 160 Me 192 R = H Figure 5-19 183 R = OH 190 HPLC separation of the glycosylated macrolide fractions of KdesI/pCM45 extracts afforded six of the minor compounds in pure or nearly pure form. While the yields of these compounds were not sufficient for NMR characterization, high resolution mass spectrometry and HPLC analysis showed that three of the compounds had elemental

composition and polarity consistent with 3-O-(4-epi-D-mycaminosyl) novamethynolide

(191), 3-O-(4-epi-D-mycaminosyl) narbonolide (192), and 3-O-(4-epi-D-mycaminosyl) 10-deoxymethynolide (193). However, the other three minor compounds displayed elemental compositions and polarities consistent with desmethyl analogues of 181/182 and 191. Reasoning that these compounds could bear 3-N-monomethylated sugars, ESI- MS-MS fragmentation analysis was carried out on the purified desmethyl analogues of 182 and 191 in order to determine whether the aglycon or the sugar lacked a methyl group. Comparison of the ESI-MS-MS fragmentation patterns of 182 and 191 (Figure 5- 20A, C, respectively) and their corresponding desmethyl analogues (Figure 5-20B, D, respectively) clearly revealed that the sugar moiety of each of these analogues lacks a methyl group. These results strongly suggest that these analogues are novel 3-N-

monomethylamino-6-deoxy-D-galactosyl derivatives of neomethynolide (196) and novamethynolide (194, Figure 5-20). The presence of these compounds was unexpected, as no macrolides bearing 3-N-monomethylated derivatives of desosamine or mycaminose have ever been detected in wild-type or engineered S. venezuelae strains. The production

of these compounds by the KdesI/pCM45 mutant is likely the result of interception of a

portion of TDP-3-N-monomethyl-6-deoxy-D-galactose (185), the product of the first DesVI-catalyzed methyltransfer reaction, by the glycosyltransferase DesVII before it can

be converted to TDP-4-epi-D-mycaminose (186) by DesVI. Apparently, the ability of DesVI to act on 185 is impaired due to the change in C-4 stereochemistry, allowing

191 sufficient accumulation of 185 for DesVII to couple it to 10-deoxymethynolide (124), resulting in the formation of 194-196.

O OH Me 174 Me2N A 100 HO O Me Me Me + O 486 75 Exact Mass: 174.113 HO O O O Me 50 HO Me Me Me N 2 OH 25 170 Exact Mass: 485.2989 0 200 300 400 500 m/z O 160 MeHNOH Me B 100 HO O Me Me Me 75 + O Exact Mass: 160.0974 HO O O O Me 50 HO Me Me MeHN OH 472 25 184 Exact Mass: 471.2832 0 200 300 400 500 O OH m/z 174 Me2N Me C 100 HO HO O Me Me + Me O 75 Exact Mass: 174.113 HO O O O Me relative abundance 50 190 Me HO Me Me N 2 OH 502 25 179 Exact Mass: 501.2938 0 200 300 O 400 500 m/z Me MeHNOH 160 HO D 100 HO Me Me O Me O + HO 488 75 Exact Mass: 160.0974 O O O Me Me HO 50 Me MeHN OH 487 25 182 Exact Mass: 487.2781 470 0 200 300 400 500 m/z

Figure 5-20. ESI-MS-MS spectra of (A) 182, (B) desmethyl analogue of 182, (C) 191, and (D) desmethyl analogue of 191. Signals with m/z = 174 correspond to ions of 4- epi-D-mycaminose (160), whereas those with m/z = 160 correspond to ions of 4-N- monomethylamino-6-deoxy-D-galactose (184).

The results obtained with the KdesI/pCM45 mutant are intriguing for several reasons. First, they demonstrate the design and successful assembly of a pathway for the biosynthesis and attachment of a non-natural deoxysugar, TDP-4-epi-D-mycaminose (186), generating several new macrolides 181-183, 191-196. The success of this strategy relied on the tolerance of four desosamine pathway enzymes, DesV, DesVI, DesVII, and DesVIII, for substrates with altered C-4 stereochemistry. Second, this engineering work 192 serendipitously led to the creation of at least three additional novel macrolide derivatives (194-196) bearing the non-natural deoxysugar 3-N-monomethyl-6-deoxy-D-galactose (184). Formation of 194-196 relied on the subtle differences in the proficiencies of two desosamine pathway enzymes, DesVI and DesVII, for turnover of non-natural substrates. These differences were only brought to light after interrogation of these enzymes with non-natural substrates generated by pathway engineering, illustrating the influence of subtle enzymological effects on the outcome of engineered pathways.

4. CONCLUSIONS

With the increasingly rapid discovery of unusual sugar pathway-encoding genes in both natural product and polysaccharide biosynthetic gene clusters, new components for pathway construction are continually becoming available to the biosynthetic engineer. These new “glycosyl tools” expand the number of feasibly constructed sugar structures, making it possible to assemble sugar biosynthetic pathways using enzymes from several pathways and to create custom-designed pathways to make sugars that do not exist in nature, such as 158 and 160. While the inconclusive and negative results obtained with the KdesII/pCM42 and KdesI/pCM43 mutants illustrate some of the challenges faced in biosynthetic engineering work, the nine non-natural sugar-bearing compounds (181-183, 191-196) generated by a selected single gene substitution in the KdesI/pCM45 mutant demonstrates the power of this approach for synthesizing non-natural deoxysugar-bearing antibiotics.

193 References

1. Durand, G.; Seta, N., Protein glycosylation and diseases: blood and urinary oligosaccharides as markers for diagnosis and therapeutic monitoring. Clin. Chem. 2000, 46, (6 Pt 1), 795-805. 2. Martin-Rendon, E.; Blake, D. J., Protein glycosylation in disease: new insights into the congenital muscular dystrophies. Trends Pharmacol. Sci. 2003, 24, (4), 178-83. 3. Este, J. A., Virus entry as a target for anti-HIV intervention. Curr. Med. Chem. 2003, 10, (17), 1617-32. 4. Schnaitman, C. A.; Klena, J. D., Genetics of lipopolysaccharide biosynthesis in enteric bacteria. Microbiol Rev 1993, 57, (3), 655-82. 5. Thorson, J. S., Hosted, T. J., Jr., Jiang, J., Biggins, J. B., Ahlert, J., Nature's carbohydrate chemists: the enzymatic glycosylation of bioactive bacterial metabolites. Curr. Org. Chem. 2001, 5, 139-167. 6. Kren, V., Martinkova, L., Glycosides in medicine: "the role of glycosidic residue in biological activity". Curr. Med. Chem. 2001, 8, 1303-1328. 7. Weymouth-Wilson, A. C., The role of carbohydrates in biologically active natural products. Nat. Prod. Rep. 1997, 14, 99-110. 8. Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., Marth, J., Essentials of Glycobiology. Cold Spring Harbor Laboratory Press Cold Spring Harbor, New York, 1999. 9. Hallis, T. M., Liu, H.-w., Learning nature's strategies for making deoxy sugars: pathways, mechanisms, and combinatorial applications. Acc. Chem. Res. 1999, 32, 579-588. 10. Blanchard, S.; Thorson, J. S., Enzymatic tools for engineering natural product glycosylation. Curr. Opin. Chem. Biol. 2006, 10, (3), 263-71. 194 11. Yang, J.; Hoffmeister, D.; Liu, L.; Fu, X.; Thorson, J. S., Natural product glycorandomization. Bioorg. Med. Chem. 2004, 12, (7), 1577-84. 12. Jelakovic, S., Schulz, G. E., The Structure of CMP:2-keto-3-deoxy-manno- octonic acid synthetase and of its complexes with substrates and substrate analogues. J. Mol. Biol. 2001, 312, 143-155. 13. Schoenhofen, I. C., McNally, D. J., Vinogradov, E., Whitfield, D., Young, N. M., Dick, S., Wakarchuk, W. W., Brisson, J.-R., Logan, S. M. , Functional characterization of dehydratase/aminotransferase pairs from Helicobacter and Campylobacter. J. Biol. Chem. 2006, 281, (2), 723-732. 14. Tanner, M. E., The enzymes of sialic acid biosynthesis. Bioorg. Chem. 2002, 33, 216-228. 15. Hofmann, C.; Boll, R.; Heitmann, B.; Hauser, G.; Durr, C.; Frerich, A.; Weitnauer, G.; Glaser, S. J.; Bechthold, A., Genes encoding enzymes responsible for biosynthesis of L-lyxose and attachment of eurekanate during avilamycin biosynthesis. Chem. Biol. 2005, 12, (10), 1137-43. 16. Sosio, M.; Stinchi, S.; Beltrametti, F.; Lazzarini, A.; Donadio, S., The gene cluster for the biosynthesis of the glycopeptide antibiotic A40926 by Nonomuraea species. Chem. Biol. 2003, 10, (6), 541-9. 17. Nedal, A.; Zotchev, S. B., Biosynthesis of deoxyaminosugars in antibiotic- producing bacteria. Appl. Microbiol. Biotechnol. 2004, 64, (1), 7-15. 18. Vogan, E. M.; Bellamacina, C.; He, X.; Liu, H. W.; Ringe, D.; Petsko, G. A., Crystal structure at 1.8 A resolution of CDP-D-glucose 4,6-dehydratase from Yersinia pseudotuberculosis. Biochemistry 2004, 43, (11), 3057-67. 19. Mulichak, A. M.; Bonin, C. P.; Reiter, W. D.; Garavito, R. M., Structure of the MUR1 GDP-mannose 4,6-dehydratase from Arabidopsis thaliana: implications for ligand binding and specificity. Biochemistry 2002, 41, (52), 15578-89. 20. Schoenhofen, I. C.; McNally, D. J.; Vinogradov, E.; Whitfield, D.; Young, N. M.; Dick, S.; Wakarchuk, W. W.; Brisson, J. R.; Logan, S. M., Functional characterization of dehydratase/aminotransferase pairs from Helicobacter and

195 Campylobacter: enzymes distinguishing the pseudaminic acid and bacillosamine biosynthetic pathways. J. Biol. Chem. 2006, 281, (2), 723-32. 21. Hegeman, A. D.; Gross, J. W.; Frey, P. A., Concerted and stepwise dehydration mechanisms observed in wild-type and mutated Escherichia coli dTDP-glucose 4,6-dehydratase. Biochemistry 2002, 41, (8), 2797-804. 22. Mulrooney, E. F., Poon, K. K. H., McNally, D. J., Brisson, J.-R., Lam, J. S., Biosynthesis of UDP-N-acetyl-L-fucosamine, a precursor to the biosynthesis of lipopolysaccharide in Pseudomonas aeruginosa serotype O11. J. Biol. Chem. 2005, 280, (20), 19535-19542. 23. Hofmann, C., Boll, R., Heitmann, B., Hauser, G., Durr, C., Frerich, A., Weitnauer, G., Glaser, S. J., Bechthold, A., Genes encoding enzymes responsible for biosynthesis of L-lyxose and attachment of eurekanate during avilamycin biosynthesis. Chem. Biol. 2005, 12, (10), 1137-1143. 24. Hosted, T. J., Wang, T. X., Alexander, D. C., Horan, A. C., Characterization of the biosynthetic gene cluster for the oligosaccharide antibiotic, evernimicin, in Micromonospora carbonacea var. africana ATCC 39149. J. Ind. Micro. Biotech. 2001, 27, 386-392. 25. Russell, R., Liu, H.-w., Stereochemical and mechanistic studies of CDP-D- glucose oxidoreductase isolated from Yersinia pseudotuberculosis. J. Am. Chem. Soc. 1991, 113, 7777-7778. 26. Mulichak, A. M., Bonin, C. P., Reiter, W.-D., Garavito, R. M., Structure of the MUR1 GDP-mannose 4,6-dehydratase from Arabidopsis thaliana: implications for ligand binding and specificity. Biochemistry 2002, 41, 15578-15589. 27. Aparicio, J. F., Caffrey, P., Gil, J. A., Zotchev, S. B., Polyene antibiotic biosynthesis gene clusters. Appl. Microbiol. Biotech. 2003, 61, 179-188. 28. Samuel, G.; Reeves, P., Biosynthesis of O-antigens: genes and pathways involved in nucleotide sugar precursor synthesis and O-antigen assembly. Carbohydr. Res. 2003, 338, (23), 2503-19. 29. Waldron, C.; Matsushima, P.; Rosteck, P. R., Jr.; Broughton, M. C.; Turner, J.; Madduri, K.; Crawford, K. P.; Merlo, D. J.; Baltz, R. H., Cloning and analysis of

196 the spinosad biosynthetic gene cluster of Saccharopolyspora spinosa. Chem. Biol. 2001, 8, (5), 487-99. 30. Ahlert, J.; Shepard, E.; Lomovskaya, N.; Zazopoulos, E.; Staffa, A.; Bachmann, B. O.; Huang, K.; Fonstein, L.; Czisny, A.; Whitwam, R. E.; Farnet, C. M.; Thorson, J. S., The calicheamicin gene cluster and its iterative type I enediyne PKS. Science 2002, 297, (5584), 1173-6. 31. Glaser, L., Kornfield, S., The enzymatic synthesis of thymidine-linked sugars. J. Biol. Chem. 1961, 236, (6), 1795-1799. 32. Jiang, X. M., Neal, B., Santiago, F., Lee, S. J., Romana, L. K., Reeves, P. R. , Structure and sequence of the rfb (O antigen) gene cluster of Salmonella serovar typhimurium (strain LT2). Mol. Microbiol. 1991, 5, (3), 695-713. 33. Lindquist, L., Kaiser, R., Reeves, P. R., Lindberg, A. A., Purification, characterization and HPLC assay of Salmonella glucose-1-phosphate thymidylyltransferase from the cloned rfbA gene. Eur. J. Biochem. 1993, 211, 763-770. 34. Stein, A., Kula, M.-R., Elling, L., Combined preparative enzymatic synthesis of dTDP-6-deoxy-4-keto-D-glucose from dTDP and sucrose. Glycoconj. J. 1998, 15, 139-145. 35. Graninger, M., Nidetzky, B., Heinrichs, D. E., Whitfield, C., Messner, P., Characterization of dTDP-4-dehydrorhamnose 3,5-epimerase and dTDP-4- dehydrorhamnose reductase, required for dTDP-L-rhamnose biosynthesis in Salmonella enterica serovar typhimurium LT2 J. Biol. Chem. 1999, 274, (35), 25069-25077. 36. Allard, S. T.; Cleland, W. W.; Holden, H. M., High resolution X-ray structure of dTDP-glucose 4,6-dehydratase from Streptomyces venezuelae. J. Biol. Chem. 2004, 279, (3), 2211-20. 37. Zhao, L., Sherman, D. H., Liu, H.-w., Biosynthesis of desosamine: construction of a new methymycin/neomethymycin analogue by deletion of a desoamine biosynthetic gene J. Am. Chem. Soc. 1998, 120, 10256-10257.

197 38. Chang, C.; Zhao, L.; Yamase, H.; Liu, H., DesVI: A new member of the sugar N,N-dimethyltransferase family involved in the biosynthesis of desosamine. Angew. Chem. Int. Ed. 2000, 39, (12), 2160-2163. 39. Zhao, L., Que, N. L. S., Xue, Y., Sherman, D. H., Liu, H.-w., Mechansitic studies of desosamine biosynthesis: C-4 deoxygenation precedes C-3 transamination. J. Am. Chem. Soc. 1998, 120, 12159-12160. 40. Zhao, L.; Borisova, S.; Yeung, S. M.; Liu, H., Study of C-4 deoxygenation in the biosynthesis of desosamine: evidence implicating a novel mechanism. J. Am. Chem. Soc. 2001, 123, (32), 7909-10. 41. Borisova, S. A.; Zhao, L.; Sherman, D. H.; Liu, H. W., Biosynthesis of desosamine: construction of a new macrolide carrying a genetically designed sugar moiety. Org. Lett. 1999, 1, (1), 133-6. 42. Szu, P. H.; He, X.; Zhao, L.; Liu, H. W., Biosynthesis of TDP-D-desosamine: identification of a strategy for C4 deoxygenation. Angew. Chem. Int. Ed. 2005, 44, (41), 6742-6. 43. Baltz, R. H., Seno, E. T., Genetics of Streptomyces fradiae and tylosin biosynthesis. Annual Rev. Microbiol. 1988, 42, 547-574. 44. Chen, H., Guo, Z., Liu, H.-w., Expression, purification, and characterization of TylM1, an N,N-dimethyltransferase involved in the biosynthesis of mycaminose. J. Am. Chem. Soc. 1998, 120, 9951-9952. 45. Cundliffe, E., Bate, N., Butler, A., Fish, S., Gandecha, A., Merson-Davies, L., The tylosin biosynthetic genes of Streptomyces fradiae. Antonie van Leeuwenhoek 2001, 79, 229-234. 46. Chen, H., Yeung, S.-M., Que, N. L. S., Muller, T., Schmidt, R. R., Liu, H.-w., Expression, purification, and characterization of TylB, an aminotransferase involved in the biosynthesis of mycaminose. J. Am. Chem. Soc. 1999, 121, 7166- 7167. 47. Naudorf, A., Klaffke, W., Substrate specificity of native dTDP-D-glucose-4,6- dehydratase: chemo-enzymic syntheses of artificial and naturally occurring deoxy sugars. Carbohydr. Res. 1996, 285, 141-150.

198 48. Stein, A., Kula, M.-R., Elling, L., Versack, S., Klaffke, W., Synthesis of dTDP-6- deoxy-4-ketoglucose and analogs with native and recombinant dTDP-glucose-4,6- dehydratase. Angew. Chem. Int. Ed. 1995, 34, 1748-749. 49. Chen, H., Agnihotri, G., Guo, Z., Que, N. L. S., Chen, X., Liu, H.-w., Biosynthesis of mycarose: isolation and characterization of enzymes involved in the C-2 deoxygenation. J. Am. Chem. Soc. 1999, 121, 8124-8125. 50. Chen, H., Zhao, Z., Hallis, T. M., Guo, Z., Liu, H.-w., Insights into the branched- chain formation of mycarose: methylation catalyzed by an (S)- adenosylmethionine-dependent methyltransferase. Angew. Chem. Int. Ed. 2001, 40, (3), 607-610. 51. Takahashi, H., Liu, Y.-n., Chen, H., Liu, H.-w., Biosynthesis of TDP-L-mycarose: the specificity of a single enzyme governs the outcome of the pathway. J. Am. Chem. Soc. 2005, 127, 9340-9341. 52. Takahashi, H., Liu, Y.-n., Liu, H.-w., A two-stage one-pot enzymatic synthesis of TDP-L-mycarose from thymidine and glucose-1-phosphate. J. Am. Chem. Soc. 2006, 128, 1432-1433. 53. Zhao, Z., Hong, L., Liu, H.-w., Characterization of the protein encoded by spnR from the spinosyn gene cluster of Saccharopolyspora spinosa: mechanistic implications for forosamine biosynthesis. J. Am. Chem. Soc. 2005, 127, (21), 7692-7693. 54. Hong, L. Biochemical studies of the enzymes involved in deoxysugar D- forosamine biosynthesis. The University of Texas at Austin, Austin, TX, 2004. 55. Chen, H., Thomas, M. G., Hubbard, B. K., Losey, H. C., Walsh, C. T., Burkart, M. G., Deoxysugars in glycopeptide antibiotics: enzymatic synthesis of TDP-L- epivancosmaine in chloroeremomycin biosynthesis. Proc. Natl. Acad. Sci. 2000, 97, (22), 11942-11947. 56. Thuy, T. T. T., Lee, H. C., Kim, C.-G., Heide, L., Sohng, J. K., Functional characterizations of novWUS involved in novobiocin biosynthesis from Streptomyces sphaeroides. Arch. Biochem. Biophys. 2005, 436, 161-167.

199 57. Breton, C.; Snajdrova, L.; Jeanneau, C.; Koca, J.; Imberty, A., Structures and mechanisms of glycosyltransferases. Glycobiology 2006, 16, (2), 29R-37R. 58. Lairson, L. L.; Withers, S. G., Mechanistic analogies amongst carbohydrate modifying enzymes. Chem. Commun. (Camb.) 2004, (20), 2243-8. 59. Persson, K.; Ly, H. D.; Dieckelmann, M.; Wakarchuk, W. W.; Withers, S. G.; Strynadka, N. C., Crystal structure of the retaining galactosyltransferase LgtC from Neisseria meningitidis in complex with donor and acceptor sugar analogs. Nat. Struct. Biol. 2001, 8, (2), 166-75. 60. Lairson, L. L.; Chiu, C. P.; Ly, H. D.; He, S.; Wakarchuk, W. W.; Strynadka, N. C.; Withers, S. G., Intermediate trapping on a mutant retaining alpha- galactosyltransferase identifies an unexpected aspartate residue. J. Biol. Chem. 2004, 279, (27), 28339-44. 61. Chiu, C. P.; Watts, A. G.; Lairson, L. L.; Gilbert, M.; Lim, D.; Wakarchuk, W. W.; Withers, S. G.; Strynadka, N. C., Structural analysis of the sialyltransferase CstII from Campylobacter jejuni in complex with a substrate analog. Nat. Struct. Mol. Biol. 2004, 11, (2), 163-70. 62. Mulichak, A. M.; Losey, H. C.; Walsh, C. T.; Garavito, R. M., Structure of the UDP-glucosyltransferase GtfB that modifies the heptapeptide aglycone in the biosynthesis of vancomycin group antibiotics. Structure 2001, 9, (7), 547-57. 63. Mulichak, A. M.; Losey, H. C.; Lu, W.; Wawrzak, Z.; Walsh, C. T.; Garavito, R. M., Structure of the TDP-epi-vancosaminyltransferase GtfA from the chloroeremomycin biosynthetic pathway. Proc. Natl. Acad. Sci. U S A 2003, 100, (16), 9238-43. 64. Mulichak, A. M.; Lu, W.; Losey, H. C.; Walsh, C. T.; Garavito, R. M., Crystal structure of vancosaminyltransferase GtfD from the vancomycin biosynthetic pathway: interactions with acceptor and nucleotide ligands. Biochemistry 2004, 43, (18), 5170-80. 65. Quiros, L. M.; Salas, J. A., Biosynthesis of the macrolide oleandomycin by Streptomyces antibioticus. Purification and kinetic characterization of an oleandomycin glucosyltransferase. J. Biol. Chem. 1995, 270, (31), 18234-9.

200 66. Quiros, L. M., Carbajo, R. J., Brana, A. F., Salas, J. A., Glycosylation of macrolide antibiotics: purification and kinetic studies of a macrolide glycosyltransferase from Streptomyces antibiotics. J. Biol. Chem. 2000, 275, (16), 11713-11720. 67. Losey, H. C.; Peczuh, M. W.; Chen, Z.; Eggert, U. S.; Dong, S. D.; Pelczer, I.; Kahne, D.; Walsh, C. T., Tandem action of glycosyltransferases in the maturation of vancomycin and teicoplanin aglycones: novel glycopeptides. Biochemistry 2001, 40, (15), 4745-55. 68. Borisova, S. A.; Zhao, L.; Melancon, I. C.; Kao, C. L.; Liu, H. W., Characterization of the glycosyltransferase activity of desVII: analysis of and implications for the biosynthesis of macrolide antibiotics. J. Am. Chem. Soc. 2004, 126, (21), 6534-5. 69. Lee, H. Y.; Chung, H. S.; Hang, C.; Khosla, C.; Walsh, C. T.; Kahne, D.; Walker, S., Reconstitution and characterization of a new desosaminyl transferase, EryCIII, from the erythromycin biosynthetic pathway. J. Am. Chem. Soc. 2004, 126, (32), 9924-5. 70. Lu, W.; Leimkuhler, C.; Oberthur, M.; Kahne, D.; Walsh, C. T., AknK is an L-2- deoxyfucosyltransferase in the biosynthesis of the anthracycline aclacinomycin A. Biochemistry 2004, 43, (15), 4548-58. 71. Lu, W.; Leimkuhler, C.; Gatto, G. J., Jr.; Kruger, R. G.; Oberthur, M.; Kahne, D.; Walsh, C. T., AknT is an activating protein for the glycosyltransferase AknS in L- aminodeoxysugar transfer to the aglycone of aclacinomycin A. Chem. Biol. 2005, 12, (5), 527-34. 72. Freel Meyers, C. L.; Oberthur, M.; Anderson, J. W.; Kahne, D.; Walsh, C. T., Initial characterization of novobiocic acid noviosyl transferase activity of NovM in biosynthesis of the antibiotic novobiocin. Biochemistry 2003, 42, (14), 4179- 89. 73. Zhang, C.; Griffith, B. R.; Fu, Q.; Albermann, C.; Fu, X.; Lee, I. K.; Li, L.; Thorson, J. S., Exploiting the reversibility of natural product glycosyltransferase- catalyzed reactions. Science 2006, 313, (5791), 1291-4.

201 74. Lim, E.-K., Plant glycosyltransferases: their potential an novel biocatalysts. Chem. Eur. J. 2005, 11, 5486-5494. 75. Fish, S. A.; Cundliffe, E., Stimulation of polyketide metabolism in Streptomyces fradiae by tylosin and its glycosylated precursors. Microbiol. 1997, 143 ( Pt 12), 3871-6. 76. Wilson, V. T.; Cundliffe, E., Characterization and targeted disruption of a glycosyltransferase gene in the tylosin producer, Streptomyces fradiae. Gene 1998, 214, (1-2), 95-100. 77. Bate, N.; Butler, A. R.; Smith, I. P.; Cundliffe, E., The mycarose-biosynthetic genes of Streptomyces fradiae, producer of tylosin. Microbiol. 2000, 146 ( Pt 1), 139-46. 78. Summers, R. G.; Donadio, S.; Staver, M. J.; Wendt-Pienkowski, E.; Hutchinson, C. R.; Katz, L., Sequencing and mutagenesis of genes from the erythromycin biosynthetic gene cluster of Saccharopolyspora erythraea that are involved in L- mycarose and D-desosamine production. Microbiol. 1997, 143 ( Pt 10), 3251-62. 79. Salah-Bey, K.; Doumith, M.; Michel, J. M.; Haydock, S.; Cortes, J.; Leadlay, P. F.; Raynal, M. C., Targeted gene inactivation for the elucidation of deoxysugar biosynthesis in the erythromycin producer Saccharopolyspora erythraea. Mol. Gen. Genet. 1998, 257, (5), 542-53. 80. Doumith, M.; Legrand, R.; Lang, C.; Salas, J. A.; Raynal, M. C., Interspecies complementation in Saccharopolyspora erythraea : elucidation of the function of oleP1, oleG1 and oleG2 from the oleandomycin biosynthetic gene cluster of Streptomyces antibioticus and generation of new erythromycin derivatives. Mol. Microbiol. 1999, 34, (5), 1039-48. 81. Aguirrezabalaga, I.; Olano, C.; Allende, N.; Rodriguez, L.; Brana, A. F.; Mendez, C.; Salas, J. A., Identification and expression of genes involved in biosynthesis of L-oleandrose and its intermediate L-olivose in the oleandomycin producer Streptomyces antibioticus. Antimicrob. Agents. Chemother. 2000, 44, (5), 1266- 75.

202 82. Thiericke, R., Rohr, J., Biological variation of microbial metabolites by precursor- directed biosynthesis. Nat. Prod. Rep. 1993, 10, (3), 265-289. 83. Weist, S., Sussmuth, R. D., Mutational biosynthesis - a tool for the generation of structural diversity in the biosynthesis of antibiotics. Appl. Microbiol. Biotech. 2005, 68, 141-150. 84. Hopwood, D. A.; Malpartida, F.; Kieser, H. M.; Ikeda, H.; Duncan, J.; Fujii, I.; Rudd, B. A.; Floss, H. G.; Omura, S., Production of 'hybrid' antibiotics by genetic engineering. Nature 1985, 314, (6012), 642-4. 85. Weber, J. M.; Leung, J. O.; Swanson, S. J.; Idler, K. B.; McAlpine, J. B., An erythromycin derivative produced by targeted gene disruption in Saccharopolyspora erythraea. Science 1991, 252, (5002), 114-7. 86. Solenberg, P. J.; Matsushima, P.; Stack, D. R.; Wilkie, S. C.; Thompson, R. C.; Baltz, R. H., Production of hybrid glycopeptide antibiotics in vitro and in Streptomyces toyocaensis. Chem. Biol. 1997, 4, (3), 195-202. 87. Yamase, H., Zhao, L., Liu, H.-w., Engineering a hybrid sugar biosynthetic pathway: production of L-rhamnose and its implication on dihydrostreptose biosynthesis. J. Am. Chem. Soc. 2000, 122, 12397-12398. 88. Madduri, K.; Kennedy, J.; Rivola, G.; Inventi-Solari, A.; Filippini, S.; Zanuso, G.; Colombo, A. L.; Gewain, K. M.; Occi, J. L.; MacNeil, D. J.; Hutchinson, C. R., Production of the antitumor drug epirubicin (4'-epidoxorubicin) and its precursor by a genetically engineered strain of Streptomyces peucetius. Nat. Biotechnol. 1998, 16, (1), 69-74. 89. Hoffmeister, D.; Wilkinson, B.; Foster, G.; Sidebottom, P. J.; Ichinose, K.; Bechthold, A., Engineered urdamycin glycosyltransferases are broadened and altered in substrate specificity. Chem. Biol. 2002, 9, (3), 287-95. 90. Hoffmeister, D.; Ichinose, K.; Bechthold, A., Two sequence elements of glycosyltransferases involved in urdamycin biosynthesis are responsible for substrate specificity and enzymatic activity. Chem. Biol. 2001, 8, (6), 557-67.

203 91. Love, K. R.; Swoboda, J. G.; Noren, C. J.; Walker, S., Enabling glycosyltransferase evolution: a facile substrate-attachment strategy for phage- display enzyme evolution. Chembiochem 2006, 7, (5), 753-6. 92. Aharoni, A.; Thieme, K.; Chiu, C. P.; Buchini, S.; Lairson, L. L.; Chen, H.; Strynadka, N. C.; Wakarchuk, W. W.; Withers, S. G., High-throughput screening methodology for the directed evolution of glycosyltransferases. Nat. Methods. 2006, 3, (8), 609-14. 93. Decker, H., Haag, S., Udvarnoki, G., Rohr, J., Novel genetically engineered tetracenomycins. Ang. Chem. Int. Ed. 1995, 34, 1107-1110. 94. Wohlert, S.-E., Blanco, G., Lombo, F., Fernandez, E., Brana, A. F., Reich, S., Udvarnoki, G., Mendez, C., Decker, H., Frevert, J., Salas, J. A., Rohr, J., Novel hybrid tetracenomycins through combinatorial biosynthesis using a glycosyltransferase encoded by the elm genes in cosmid 16F4 and which shows broad sugar substrate specificity. J. Am. Chem. Soc. 1998, 120, 10596-10601. 95. Blanco, G.; Patallo, E. P.; Brana, A. F.; Trefzer, A.; Bechthold, A.; Rohr, J.; Mendez, C.; Salas, J. A., Identification of a sugar flexible glycosyltransferase from Streptomyces olivaceus, the producer of the antitumor polyketide elloramycin. Chem. Biol. 2001, 8, (3), 253-63. 96. Rodriguez, L., Aguirrezabalaga, I., Allende, N., Brana, A. F., Mendez, C., Salas, J. A., Engineering deoxysugar biosynthetic pathways from antibiotic-producing microorganisms: a tool to produce novel glycosylated bioactive compounds. Chem. Biol. 2002, 9, 721-729. 97. Fischer, C., Rodriguez, L., Patallo, E. P., Lipata, F., Brana, A. F., Mendez, C., Salas, J. A., Rohr, J., Digitoxosyltetracenomycin C and glucosyltetracenomycin C, two novel elloramycin analogues obtained by expolring the sugar donor substrate flexibility of glycosyltransferass ElmGT. J. Nat. Prod. 2002, 65, 1685- 1689. 98. Lombo, F.; Gibson, M.; Greenwell, L.; Brana, A. F.; Rohr, J.; Salas, J. A.; Mendez, C., Engineering biosynthetic pathways for deoxysugars: branched-chain

204 sugar pathways and derivatives from the antitumor tetracenomycin. Chem. Biol. 2004, 11, (12), 1709-18. 99. Perez, M.; Lombo, F.; Zhu, L.; Gibson, M.; Brana, A. F.; Rohr, J.; Salas, J. A.; Mendez, C., Combining sugar biosynthesis genes for the generation of L- and D- amicetose and formation of two novel antitumor tetracenomycins. Chem. Commun. (Camb.) 2005, (12), 1604-6. 100. Losey, H. C.; Jiang, J.; Biggins, J. B.; Oberthur, M.; Ye, X. Y.; Dong, S. D.; Kahne, D.; Thorson, J. S.; Walsh, C. T., Incorporation of glucose analogs by GtfE and GtfD from the vancomycin biosynthetic pathway to generate variant glycopeptides. Chem. Biol. 2002, 9, (12), 1305-14. 101. Fu, X.; Albermann, C.; Jiang, J.; Liao, J.; Zhang, C.; Thorson, J. S., Antibiotic optimization via in vitro glycorandomization. Nat. Biotechnol. 2003, 21, (12), 1467-9. 102. Oberthur, M.; Leimkuhler, C.; Kruger, R. G.; Lu, W.; Walsh, C. T.; Kahne, D., A systematic investigation of the synthetic utility of glycopeptide glycosyltransferases. J. Am. Chem. Soc. 2005, 127, (30), 10747-52. 103. Albermann, C.; Soriano, A.; Jiang, J.; Vollmer, H.; Biggins, J. B.; Barton, W. A.; Lesniak, J.; Nikolov, D. B.; Thorson, J. S., Substrate specificity of NovM: implications for novobiocin biosynthesis and glycorandomization. Org. Lett. 2003, 5, (6), 933-6. 104. Trefzer, A., Blanco, G., Remsing, L., Kunzel, E., Rix, U., Lipata, F., Brana, A. F., Mendez, C., Rohr, J., Bechthold, A., Salas, J. A., Rationally designed glyfcosylated premithramycins: hybrid aromatic polyketides using genes from three different biosynthetic pathways. J. Am. Chem. Soc. 2002, 124, 6056-6062. 105. Gaisser, S.; Martin, C. J.; Wilkinson, B.; Sheridan, R. M.; Lill, R. E.; Weston, A. J.; Ready, S. J.; Waldron, C.; Crouse, G. D.; Leadlay, P. F.; Staunton, J., Engineered biosynthesis of novel spinosyns bearing altered deoxyhexose substituents. Chem. Commun. (Camb.) 2002, (6), 618-9.

205 106. Xue, Y.; Wilson, D.; Zhao, L.; Liu, H.; Sherman, D. H., Hydroxylation of macrolactones YC-17 and narbomycin is mediated by the pikC-encoded cytochrome P450 in Streptomyces venezuelae. Chem. Biol. 1998, 5, (11), 661-7. 107. Borisova, S. A., Zhao, L., Melançon, C. E. III, Kao, C-L., Liu, H.-w., Characterization of the glycosyltransferase activity of DesVII: analysis of and implication for the biosynthesis of macrolide antibiotics. J. Am. Chem. Soc. 2004, 126, 6534-6535. 108. Ramakrishnan, B.; Boeggeman, E.; Ramasamy, V.; Qasba, P. K., Structure and catalytic cycle of beta-1,4-galactosyltransferase. Curr. Opin. Struct. Biol. 2004, 14, (5), 593-600. 109. Butler, A. R.; Flint, S. A.; Cundliffe, E., Feedback control of polyketide metabolism during tylosin production. Microbiol. 2001, 147, (Pt 4), 795-801. 110. Otten, S. L.; Liu, X.; Ferguson, J.; Hutchinson, C. R., Cloning and characterization of the Streptomyces peucetius dnrQS genes encoding a daunosamine biosynthesis enzyme and a glycosyl transferase involved in daunorubicin biosynthesis. J. Bacteriol. 1995, 177, (22), 6688-92. 111. Anzai, Y.; Saito, N.; Tanaka, M.; Kinoshita, K.; Koyama, Y.; Kato, F., Organization of the biosynthetic gene cluster for the polyketide macrolide mycinamicin in Micromonospora griseorubida. FEMS Microbiol. Lett. 2003, 218, (1), 135-41. 112. Xue, Y.; Zhao, L.; Liu, H. W.; Sherman, D. H., A gene cluster for macrolide antibiotic biosynthesis in Streptomyces venezuelae: architecture of metabolic diversity. Proc. Natl. Acad. Sci. U S A 1998, 95, (21), 12111-6. 113. Sambrook, J., Fritsch, E. F., Maniatis, T., Molecular cloning: a laboratory manual. 1989. 114. Kieser, T., Bibb, M. J., Buttner, M. J., Chater, K. F., Hopwood, D. A., Practical Streptomyces genetics. The John Innes Foundation: Norwich, UK, 2000. 115. Simon, R., Priefer, U., Puhler, A., A broad host range mobilization system for in vivo genetic engineering: transportation mutagenesis in gram negative bacteria. Biotech. 1983, 1, (9), 784-791.

206 116. Borisova, S. A. Genetic and biochemical studies of the biosynthesis and attachment of D-desosamine, the deoxy sugar component of macrolide antibiotics produced by Streptomyces venezuelae. The University of Texas at Austin, Austin, TX, 2004. 117. Bierman, M., Logan, R., O'Brien, K., Seno, E. T., Rao, R. N., Schoner, B. E., Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene 1992, 116, 43-49. 118. Zhao, L.; Beyer, N. J.; Borisova, S. A.; Liu, H. W., Beta-glucosylation as a part of self-resistance mechanism in methymycin/pikromycin producing strain Streptomyces venezuelae. Biochemistry 2003, 42, (50), 14794-804. 119. Baltz, R. H., Genetic recombination by protoplast fusion in Streptomyces. Dev. Ind. Microbiol. 1980, 21, 43-54. 120. Hopwood, D. A., Bibb, M. J., Kieser, T., Bruton, C. J., Kieser, H. M., Lydiate, D. J., Smith, C. P., Ward, J. M., Schrempf, H., Genetic manipulation of Streptomyces: a laboratory manual. The John Innes Foundation: Norwich, UK, 1985. 121. Lambalot, R. H., Cane, D. E., Isolation and characterization of 10- deoxymethynolide produced by Streptomyces venezuelae. J. Antibiot. 1992, 45, 1981-1982. 122. Butler, A. R.; Bate, N.; Kiehl, D. E.; Kirst, H. A.; Cundliffe, E., Genetic engineering of aminodeoxyhexose biosynthesis in Streptomyces fradiae. Nat. Biotechnol. 2002, 20, (7), 713-6. 123. Xue, Y.; Sherman, D. H., Biosynthesis and combinatorial biosynthesis of pikromycin-related macrolides in Streptomyces venezuelae. Metab. Eng. 2001, 3, (1), 15-26. 124. Gaisser, S.; Reather, J.; Wirtz, G.; Kellenberger, L.; Staunton, J.; Leadlay, P. F., A defined system for hybrid macrolide biosynthesis in Saccharopolyspora erythraea. Mol. Microbiol. 2000, 36, (2), 391-401. 125. Chen, H.; Yamase, H.; Murakami, K.; Chang, C. W.; Zhao, L.; Zhao, Z.; Liu, H. W., Expression, purification, and characterization of two N,N-

207 dimethyltransferases, TylM1 and DesVI, involved in the biosynthesis of mycaminose and desosamine. Biochemistry 2002, 41, (29), 9165-83. 126. Yuan, Y.; Chung, H. S.; Leimkuhler, C.; Walsh, C. T.; Kahne, D.; Walker, S., In vitro reconstitution of EryCIII activity for the preparation of unnatural macrolides. J. .Am. Chem. Soc. 2005, 127, (41), 14128-9. 127. He, X. M.; Liu, H. W., Formation of unusual sugars: mechanistic studies and biosynthetic applications. Annu. Rev. Biochem. 2002, 71, 701-54. 128. Cundliffe, E.; Bate, N.; Butler, A.; Fish, S.; Gandecha, A.; Merson-Davies, L., The tylosin-biosynthetic genes of Streptomyces fradiae. Antonie Van Leeuwenhoek 2001, 79, (3-4), 229-34. 129. Gandecha, A. R.; Large, S. L.; Cundliffe, E., Analysis of four tylosin biosynthetic genes from the tylLM region of the Streptomyces fradiae genome. Gene 1997, 184, (2), 197-203. 130. Pfoestl, A., Hofinger, A., Kosma, P., Messner, P., Biosynthesis of dTDP-3- acetamido-3,6-dideoxy-α-D-galactose in Aneurinibacillus thermoaerophilus L420-91T. J. Biol. Chem. 2003, 278, (29), 26410-26417. 131. Zhang, Q.; Sherman, D. H., Isolation and structure determination of novamethymycin, a new bioactive metabolite of the methymycin biosynthetic pathway in Streptomyces venezuelae. J. Nat. Prod. 2001, 64, (11), 1447-50. 132. Zhao, L. Biosynthetic studies of D-desosamine and engineered biosynthesis of methymycin/pikromycin analogs carrying modified deoxysugars. University of Minnesota, Minneapolis, MN, 2000. 133. Johnson, D. A., Liu, H.-w., Deoxysugars: occurrence, genetics, and mechanisms of biosynthesis Pergamon: 1999; Vol. 3, p 311-365. 134. Hermann, T., Drugs targeting the ribosome. Curr. Opin. Struct. Biol. 2005, 15, 355-366. 135. Trefzer, A.; Salas, J. A.; Bechthold, A., Genes and enzymes involved in deoxysugar biosynthesis in bacteria. Nat. Prod. Rep. 1999, 16, (3), 283-99. 136. He, X., Liu, H.-w., Formation of unusual sugars: mechanistic and biosynthetic applications. Annual Rev. Biochem. 2002, 71, 701-754. 208 137. Takahashi, H., Liu, Y-n., Liu, H-w., A two-stage one-pot enzymatic synthesis of TDP-L-mycarose from thymidine and glucose-1-phosphate. J. Am. Chem. Soc. 2006, 128, 1432-1433. 138. Salas, J. A., Mendez, C., Biosynthesis pathways for deoxysugars in antibiotic- producing Actinomycetes: isolation, characterization and generation of novel glycosylated derivatives. J. Mol. Microbiol. Biotech. 2005, 9, 77-85. 139. Hong, J. S. J., Park, S. H., Choi, C. Y., Sohng, J. K., Yoon, Y. J., New olivosyl derivatives of methymycin/pikromycin from an engineered strain of Streptomyces venezuelae. FEMS Microbiol. Lett. 2004, 238, 391-399. 140. Bryskier, A. J., Butzler, J.-P., Neu, H. C., Tulkens, P. M., Macrolides: chemistry, pharmacology, and clinical uses. Arnette Blackwell: Paris, 1993; p 698. 141. Laemmli, U. K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, 227, (5259), 680-685. 142. Vesterberg, O., Staining of protein zones after isoelectric focusing in polyacrylamide gels. Biochim. Biophys. Acta 1971, 243, (2), 345-8. 143. Bradford, M. M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248-254. 144. Andrews, P., Estimation of the molecular weights of proteins by Sephadex gel- filtration. Biochemical J. 1964, 91, (2), 222-233. 145. Chen, H., Agnihotri, G., Guo, Z., Que, N. L. S., Chen, X. H., Liu, H-w., Biosynthesis of mycarose: isolation and characterization of enzymes involved in the C-2 deoxygenation. J. Am. Chem. Soc. 1999, 121, 8124-8125. 146. Draeger, G., Park, S.-H., Floss, H. G., Mechanism of the C-2 deoxygenation step in the biosynthesis of the deoxyhexose moieties of the antibiotics granaticin and oleandomycin. J. Am. Chem. Soc. 1999, 121, 2611-2612. 147. Chen, X. M. H., Ploux, O., Liu, H.-w., Biosynthesis of 3,6-dideoxyhexoses: in vivo and in vitro evidence for protein-protein interaction between CDP-6-deoxy-

L-threo-D-glycero-4-hexulose-3-dehydrase (E1) and its reductase (E3). Biochemistry 1996, 35, 16412-16420.

209 148. Thorson, J. S., Lo, S. F., Ploux, O., He, X., Liu, H.-w., Studies of the biosynthesis of 3,6-dideoxyhexoses: molecular cloning and characterization of the asc (ascarylose) region from Yersinia pseudotuberculosis serogroup VA. J. Bact. 1994, 176, (17), 5483-5493. 149. Thorson, J. S., Kelly, T. M., Liu, H.-w., Cloning, sequencing, and overexpression in Escherichia coli of the α−D-glucose-1-phosphate cytidylyltransferase gene isolated from Yersinia pseudotuberculosis. J. Bact. 1994, 176, (7), 1840-1849. 150. Vorholter, F.-J., Niehaus, K., Puhler, A., Lipopolysaccharide biosynthesis in Xanthomonas campestris pv. campestris: a cluster of 15 genes is involved in the biosynthesis of the LPS O-antigen and the LPS core. Mol. Genet. Genom. 2000, 266, 79-95. 151. Dunwell, J. M., Culham, A., Carter, C. E., Sosa-Aguirre, C. R., Goodenough, P. W., Evolution of functional diversity in the cupin superfamily. Trends Biochem. Sci. 2001, 26, (12), 740-746. 152. Jakimowicz, P., Tello, M., Freel Meyers, C. L., Walsh, C. T., Buttner, M. J., Field, R. A., Lawson, D. M., The 1.6-A resolution crystal structure of NovW: a 4-keto-6-deoxy sugar epimerase from the novobiocin biosynthetic gene cluster of Streptomyces spheroides. Proteins: Structure, Function, and Bioinformatics 2006, 63, 261-265. 153. Jiang, J., Biggins, J. B., Thorson, J. S., A general enzymatic method for the synthesis of natural and "unnatural" UDP- and TDP-nucleotide sugars. J. Am. Chem. Soc. 2000, 122, 6803-6804. 154. Yang, Y.-H., Kang, Y.-B., Lee, K.-W., Lee, T.-H., Park, S.-S., Hwang, B.-Y., Kim, B.-G., Characterization of GDP-mannose pyrophosphorylase from Escherichia coli O157:H7 EDL933 and its broad substrate specificity. J. Mol. Catal. B: Enz. 2005, 37, 1-8. 155. Khaled, A.; Ivannikova, T.; Auge, C., Synthesis of unnatural sugar nucleotides and their evaluation as donor substrates in glycosyltransferase-catalyzed reactions. Carbohydr. Res. 2004, 339, (16), 2641-9.

210 156. Borisova, S. A.; Zhang, C.; Takahashi, H.; Zhang, H.; Wong, A. W.; Thorson, J. S.; Liu, H. W., Substrate specificity of the macrolide-glycosylating enzyme pair DesVII/DesVIII: opportunities, limitations, and mechanistic hypotheses. Angew. Chem. Int. Ed. 2006, 45, (17), 2748-53. 157. Salas, J. A.; Mendez, C., Biosynthesis pathways for deoxysugars in antibiotic- producing Actinomycetes: isolation, characterization and generation of novel glycosylated derivatives. J. Mol. Microbiol. Biotech. 2005, 9, (2), 77-85. 158. Volchegursky, Y.; Hu, Z.; Katz, L.; McDaniel, R., Biosynthesis of the anti- parasitic agent megalomicin: transformation of erythromycin to megalomicin in Saccharopolyspora erythraea. Mol. Microbiol. 2000, 37, (4), 752-62. 159. Ichinose, K.; Ozawa, M.; Itou, K.; Kunieda, K.; Ebizuka, Y., Cloning, sequencing and heterologous expression of the medermycin biosynthetic gene cluster of Streptomyces sp. AM-7161: towards comparative analysis of the benzoisochromanequinone gene clusters. Microbiol. 2003, 149, (Pt 7), 1633-45. 160. Bililign, T.; Hyun, C. G.; Williams, J. S.; Czisny, A. M.; Thorson, J. S., The hedamycin locus implicates a novel aromatic PKS priming mechanism. Chem. Biol. 2004, 11, (7), 959-69. 161. Hong, J. S. J., Park, S. H., Choi, C. Y., Sohng, J. K., Yoon, Y. J., New olivosyl derivatives of methymycin/pikromycin from an engineered strain of Streptomyces venezuelae. FEMS Microbiol. Lett. 2004, 238, 391-399.

211

Vita

Charles E. (Chad) Melançon III was born in Tulsa, Oklahoma on August 18th, 1975 to Charles E. Melançon Jr. and Patricia S. Briggs Melançon. The Melançon family moved to the suburbs of New Orleans, Louisiana when Chad was one year old. He attended Westbank Montessori School from age two through four, Country Day Academy from kindergarten through third grade, and St. Andrew the Apostle School from fourth through seventh grade. Chad attended Jesuit High School from eighth through twelfth grade, graduated Summa Cum Laude in 1993, and was a National Merit Semi-finalist and Finalist. He received the Dean’s Honor Scholarship to Louisiana State University, and attended LSU from 1993 until 1998, during which time he accumulated more than 150 credit hours under several declared and undeclared majors, including History, Psychology, French, Poultry Science, Biochemistry, and Geography. In 1999, Chad transferred to The University of New Orleans, where he majored in Biology. During his time at UNO, he conducted undergraduate research in the laboratories of Dr. Wendy Schluchter in Biology and Dr. Mark Trudell in Chemistry, and was the president of Beta Beta Beta Biological Honors Society for the 2000-2001 school year. Chad completed Bachelor of Science and Bachelor of Arts degrees at UNO in Biology and Chemistry, respectively, in May of 2001. He was admitted to the Biochemistry graduate program in the Department of Chemistry and Biochemistry at The University of Texas at Austin in 2001, and received the College of Natural Sciences Dean’s Excellence Award. He joined the Ben Liu Research Group in October of 2001 and was a Teaching Assistant

212 for Biochemistry Laboratory (CH369L) for five semesters and for Principles of Biochemistry (CH369) for one semester. During his graduate career, Chad was the recipient of the Lewis Award, Ravel Fellowship, and the UT Continuing Fellowship. Chad was a contributing author on four research articles and one review article by the time of submission of this dissertation, with several other scholarly articles for which he will be a contributing author currently in preparation.

Permanent address: 11282 Taylor Draper Ln., Apt. 626 Austin, TX 78759 This dissertation was typed by Charles E Melançon III.

213