STUDY OF PVCA AND PVCB, TWO ENZYMES

INVOLVED IN MAKING ISONITRILE-CONTAINING NATURAL

PRODUCTS IN BACTERIAL PATHOGENS

______

A Dissertation

presented to

the Faculty of the Graduate School

at the University of Missouri-Columbia

______

In Partial Fulfillment

of the Requirements for the Degree

Ph.D. in Biochemistry

______

By

Jing Zhu

Dr. Peter A. Tipton, Dissertation Advisor

DEC 2016

The undersigned, appointed by the dean of the Graduate School,

have examined the thesis entitled

STUDY OF PVCA AND PVCB, TWO ENZYMES

INVOLVED IN MAKING ISONITRILE-CONTAINING NATURAL

PRODUCTS IN BACTERIAL PATHOGENS

Presented by Jing Zhu

A candidate for the degree of doctor of philosophy in Biochemistry,

And hereby certify that, in their opinion, it is worthy of acceptance.

Dr. Peter A. Tipton

Dr. Gerald Hazelbauer

Dr. Judy D. Wall

Dr. Kent Gates

Dr. Timothy Glass

ACKNOWLEDGEMENTS

The experience living and studying in Columbia, Missouri will become the most precious memory in my whole life and I want to express my gratitude to many people who ever helped me.

I would like to thank the Department of Biochemistry at University of Missouri-

Columbia for admitting me as a graduate student. Otherwise I would not have this chance to experience different cultures and meet a lot of wonderful people here in the past five years.

I would like to extend a special thank you to my thesis advisor Dr. Peter Tipton.

I feel so lucky and so appreciative to have a mentor that is very knowledgeable, kind and patient on everyone and everything. His great passion for science and persistence in doing research impressed and inspired me a lot, especially when I felt lost and depressed in the past. He provided me with millions of brilliant ideas about my project but at the same time, taught me how to be an independent thinker and scientist. I have learned a lot from him, both scientifically and personally. I will be thankful for my time spent in Tipton lab forever.

I was fortune to work with two very nice ladies Krista Arnett and Dr. Emma

Farrell when I joined the lab as a green hand. Krista helped me adapt to all the laboratory techniques and gave me valuable advice on how to use the instruments and how to keep everything organized. Although I only worked with Emma for one year, I was impressed by her kindness, elegance and creativity. I would never

ii forget her returning to the lab after she just got home, to open the door for me a couple of times. Thank both of you for guiding and helping me.

I want to express my gratitude to the undergraduate students who ever worked in the lab for their companionship. Thank Kennady Gee for his hard work on my project and for teaching me how to be a good mentor.

I want to thank my committee members, Dr. Gerald Hazelbauer, Dr. Judy Wall,

Dr. Kent Gates and Dr. Timothy Glass for the suggestions and advice they have provided throughout my graduate study. They challenged me to become a better scientist.

I would like to thank Yufei Li, Qinyi Wang, Ning Bi, Weiwei Wu, Zhenwei Song,

Zhe Li, and Lingyan Jiang for their friendship, support and advice. Without them, life would be boring.

Finally, I would like to express my deepest gratitude to my parents, Lianjin Zhu and Yanwei Tang. Their trust, encouragement, support and constant love make me brave, confident and happy. Thank you for being the most important mentors and friends in my life. And thank you for missing me so much like I miss you when I am

20,000 kilometers away from you.

iii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF FIGURES ...... ix

LIST OF TABLES ...... xii

LIST OF SCHEMES ...... xiii

LIST OF ABBREVIATIONS ...... xiv

ABSTRACT ...... xviii

Chapter 1

General Introduction ...... 1

1.1 Isonitriles in Organic Chemistry ...... 2

1.2 Isonitrile-Containing Natural Products ...... 3

1.3 Isonitrile Biosynthesis ...... 6

1.4 Gene Cluster Related to Isonitrile Biosynthesis ...... 8

1.5 Functions of the pvcA/pvcB Gene Clusters and Their Related Metabolites ...... 15

1.6 Structures of the PvcA and PvcB Proteins from P. aeruginosa ...... 18

1.7 Non-Heme Fe(II) and α-Ketoglutarate Dependent Oxygenases ...... 21

1.8 Predicted Activities of the PvcA and PvcB Proteins...... 23

1.9 Overview of Investigations ...... 24

iv

Chapter 2

Characterization of Three PvcB Homologous Proteins, the Non-Heme Fe2+, α-Ketoglutarate Dependent Oxygenases...... 27

2.1 Published Work ...... 27

2.2 Summary of Published Work ...... 28

2.3 Supplemental Results ...... 30

2.3.1 Heterologous Expression of PaPvcB, XnPvcB and EaPvcB in E. coli 30

2.3.2 Rapid-Mixing Quench Flow Analysis of the PvcB Reactions ...... 32

2.3.3 Stopped-Flow UV-visible Spectroscopic Analysis of the PvcB Reactions ...... 36

2.3.4 Organic Synthesis and 1H-NMR Characterization of 4-Nitro Phenylalanine Isonitrile and 3-Nitro Tyrosine Isonitrile ...... 39

2.3.5 PvcB Activity Assays with 4-Nitro Phenylalanine Isonitrile and 3-Nitro Tyrosine Isonitrile ...... 42

2.4 Conclusion and Discussion ...... 45

2.5 Material and Methods ...... 51

2.5.1 SDS-PAGE Analysis of PvcB Expression in E. coli ...... 51

2.5.2 Rapid Quench Flow Analysis of the PvcB Reactions ...... 51

2.5.3 Stopped-Flow UV-visible Analysis of the PvcB Reactions ...... 52

2.5.4 Synthesis of 4-Nitro Phenylalanine Isonitrile ...... 52

2.5.5 Synthesis of 3-Nitro Tyrosine Isonitrile ...... 54

2.5.6 PvcB Activity Assays with 4-Nitro Phenylalanine Isonitrile ...... 55

2.5.7 PvcB Activity Assays with 3-Nitro Tyrosine Isonitrile ...... 56

2.5.8 HPLC Analyses ...... 56

v

Chapter 3

Characterization of the Reaction Catalyzed by PvcA, an Isonitrile Synthase ...... 57

3.1 Introduction...... 57

3.2 Results ...... 62

3.2.1 Heterologous Expression of PaPvcA and XnPvcA in E. coli ...... 62

3.2.2 Reconstitution of the PvcA Reactions in Vitro ...... 64

3.2.3 Reconstitution of the Isocyanovinylphenol Biosynthetic Pathway in E. coli ...... 65

3.2.4 Identification of the Product of the PvcA Reaction ...... 69

3.2.5 Identification of Tyrosine as One of the Substrates for the PvcA Reaction ...... 73

3.2.6 Seeking the Substrate Providing the Carbon for the Isonitrile Group .. 78

3.2.6.1 Tracking of the Isonitrile Carbon in E. coli Expressing AhPvcA- XnPvcB ...... 78

3.2.6.2 Searching for the Substrate in a Phosphorylated Metabolite Library and the Major Metabolic Pathways ...... 82

3.2.6.3 Looking for the Substrate in the Metabolite Pool of E. coli ...... 83

3.3 Conclusion and Discussion ...... 94

3.4 Material and Methods ...... 108

3.4.1 Cloning of PvcA Genes ...... 108

3.4.2 Heterologous Expression of the PvcA Proteins ...... 109

3.4.3 Purification of the PvcA Proteins ...... 109

3.4.4 PvcA Activity Assays ...... 110

3.4.4.1 Activities of the Purified PvcA Enzymes...... 110

3.4.4.2 In Vitro PvcA-XnPvcB Coupled Reactions ...... 111

vi

3.4.4.3 Cofactor and Co-substrate Requirements...... 111

3.4.4.4 Activities of the Cell Lysates Hosting PvcA Proteins ...... 112

3.4.5 Expression of AhpvcA-XnpvcB/pETDuet-1 in E. coli and Metabolite Detection ...... 113

3.4.6 Reaction Product Identification ...... 113

3.4.7 Substrate Identification...... 115

3.4.7.1 Tracking of Tyrosine in the Biosynthetic Pathway of Isocyanovinylphenol in E. coli ...... 115

3.4.7.2 Tracking of the Isonitrile Carbon in the Biosynthetic Pathway of Isocyanovinylphenol in E. coli ...... 116

3.4.7.3 Looking for the Substrate in a Phosphorylated Metabolite Library ...... 117

3.4.7.4 Searching for the Substrate in the Major Metabolic Pathways .... 120

3.4.7.5 Looking for the Substrate in the Metabolite Pool ...... 123

3.4.7.5.1 Metabolites Extracted from E. coli Cells ...... 123

3.4.7.5.2 Metabolites in the E. coli Cell Lysates ...... 124

3.4.7.5.3 MS Analyses of the E. coli Cell lysate Before and After the PvcA Reaction ...... 125

3.4.7.5.4 Fractionation of the E. coli Cell Lysate Hosting XnPvcB ...... 125

3.4.8 HPLC Analyses ...... 129

3.4.9 MS Analyses ...... 129

vii

Chapter 4

Thesis Summary and Future Directions ...... 131

4.1 Thesis Summary ...... 131

4.1.1 Characterization of Three PvcB Homologous Proteins, the Non-Heme Fe2+, α-Ketoglutarate Dependent Oxygenases...... 131

4.1.2 Characterization of the Reaction Catalyzed by PvcA, an Isonitrile Synthase ...... 133

4.2 Future Directions ...... 135

4.2.1 Mechanistic Studies of the PvcB Reactions ...... 135

4.2.1.1 Determine the Intermediates in the PvcB Reactions ...... 135

4.2.1.2 Investigate Substrate Tolerance for the PvcB Enzymes ...... 136

4.2.1.3 Probe the Reason that Accounts for the Different Activities Exhibited by PvcB Homologs ...... 136

4.2.2 Future Studies of the PvcA Proteins ...... 138

References ...... 141

VITA ...... 158

viii

LIST OF FIGURES

Chapter 1

Figure 1.1 Best structural representation of the isonitrile group ...... 3

Figure 1.2 Examples of Isonitriles from nature ...... 4

Figure 1.3 Isonitriles produced by bacterial pathogens ...... 5

Figure 1.4 Origin of the isonitrile group in different metabolites ...... 7

Figure 1.5 PvcA-containing gene clusters in some common bacterial pathogens ...... 14

Figure 1.6 Structures of the phenoloxidase inhibitors: (A) rhabduscin and (B) byelyankacin ...... 16

Figure 1.7 Crystal structure of (A) PvcA and (B) PvcB from P. aeruginosa ...... 19

Chapter 2

Figure 2.1 SDS-PAGE analysis of the BL21(DE3) cell lysates harboring PaPvcB, EaPvcB, and XnPvcB ...... 31

Figure 2.2 HPLC spectra of the (A) PaPvcB reaction and (B) XnPvcB reaction quenched by HCl at different times ...... 34

Figure 2.3 Intermediate and product formation in (A) PaPvcB and (B) XnPvcB-catalyzed reactions with tyrosine isonitrile as the substrate ...... 35

Figure 2.4 UV-vis spectra of the (A) PaPvcB reaction and (B) XnPvcB reaction monitored at 340 nm (blue dots) and 520 nm (red dots) ...... 38

Figure 2.5 1H-NMR spectrum of (A) 4-nitro phenylalanine isonitrile and (B) 3-nitro tyrosine isonitrile ...... 41

Figure 2.6 Spectroscopic analyses of the PvcB reactions with (A, C) 4-nitro phenylalanine isonitrile and (B, D) tyrosine isonitrile ...... 43

Figure 2.7 Spectroscopic changes during PvcB-catalyzed reactions with (A, C) 3-nitro tyrosine isonitrile and (B, D) 3-nitro tyrosine ...... 44

ix

Chapter 3

Figure 3.1 Coomassie-stained SDS-PAGE analysis of the E. coli BL21(DE3) cell lysates harboring XnPvcA and PaPvcA ...... 63

Figure 3.2 Coomassie-stained SDS-PAGE analysis of the E. coli BL21(DE3) cell lysates harboring AhPvcA and XnPvcB ...... 67

Figure 3.3 HPLC analyses of the metabolites produced from different cultures ...... 68

Figure 3.4 1H-NMR spectrum of (A) Fraction 19 collected from the size-exclusion column, and (B) authentic tyrosine isonitrile ...... 71

Figure 3.5 1H-NMR spectrum of the XnPvcB reaction with (A) fraction 19 and with (B) tyrosine isonitrile ...... 72

Figure 3.6 Incorporation of 3,5-ring-D2 tyrosine into isocyanovinylphenol in the growing culture that expressed AhPvcA and XnPvcB ...... 76

Figure 3.7 HPLC chromatogram of the metabolites isolated from the culture medium before adding m-fluoro-tyrosine (blue); 1 hour after adding m-fluoro-tyrosine (red) and 16 hours after adding m-fluoro-tyrosine (green) ...... 77

Figure 3.8 13C-NMR spectrum of the metabolite isocyanovinylphenol isolated from the culture fed 13 13 13 13 13 13 13 with D-[1- C], D-[2- C], D-[3- C], D-[4- C], D-[5- C], D-[6- C] and D-[UL- C6] glucoses (top to bottom) ...... 80

Figure 3.9 Incorporation of the 13C atom into isocyanovinylphenol isolated from the cultures fed with D-[2-13C] (red dots) and D-[3-13C] (green dots) glucoses, respectively...... 81

Figure 3.10 HPLC chromatogram of the product extracted from the reaction containing only the E. coli cell lysate that hosted XnPvcB (blue trace); the cell lysate + XnPvcA enzyme (red trace); and the cell lysate + authentic tyrosine isonitrile (green trace)...... 85

Figure 3.11 Activities of the lysates from E. coli cells expressing no plasmid (blue), empty pET15b vector (red), XnPvcB (gray), PGDH (yellow), PqsA (pink) and ACP2 (green) in the XnPvcA-XnPvcB coupled reactions ...... 86

Figure 3.12 Fractionation of the E. coli cell lysate by size-exclusion chromatography ...... 91

Figure 3.13 Schematic of partial glycolysis and the pentose phosphate pathway showing fate of (A) glucose C3 and (B) glucose C2 ...... 102

x

Figure 3.14 Proposed function of the unknown macromolecule E and the relationship between the PvcA substrate X and the unknown small compound A ...... 107

xi

LIST OF TABLES

Chapter 3

Table 3.1 Effect of the PvcA reaction on metabolite concentrations in the cell lysate ...... 89

Table 3.2 Cofactors and co-substrates examined in the in vitro PvcA reaction ...... 112

Table 3.3 List of the compounds in the phosphorylated metabolite library ...... 118

Table 3.4 List of the metabolites examined in the in vitro PvcA reaction ...... 122

xii

LIST OF SCHEMES

Chapter 1

Scheme 1.1 Biosynthetic pathway of isonitrile derivative of tryptophan ...... 9

Scheme 1.2 Proposed synthesis of pseudoverdine and pyoverdine chromophore from tyrosine under the function of the pvc operon in P. aeruginosa...... 11

Scheme 1.3 Function of the pvc operon in P. aeruginosa ...... 12

Scheme 1.4 Reactions catalyzed by various Fe2+, α-ketoglutarate dependent dioxygenases, resulting in (A) protein modification, (B) repair of alkylated DNA/RNA, (C) biosynthesis of antibiotics, and (D) biosynthesis of plant metabolites ...... 22

Chapter 2

Scheme 2.1 Organic synthesis of (A) 4-nitro phenylalanine isonitrile and (B) 3-nitro tyrosine isonitrile ...... 40

Scheme 2.2 General mechanism of the reactions catalyzed by the Fe2+, α-ketoglutarate dependent oxygenases ...... 46

Chapter 3

Scheme 3.1 In vitro synthesis of indole isonitrile by (A) purified IsnA/IsnB; (B) purified AmbI1/AmbI3; and (C) E. coli lysates hosting WelI1 and WelI3, respectively ...... 60

Scheme 3.2 Proposed biosynthesis of paerucumarin and rhabduscin in Pseudomonas aeruginosa and Xenorhabdus nematophila ...... 61

Scheme 3.3 Possible mechanisms involved in isonitrile group formation ...... 104

Scheme 3.4 Enzymatic reactions to synthesize metabolites in the Embden-Meyerhof-Parnas, the pentose phosphate pathway, the Entner-Doudoroff pathway and glycerol metabolism ...... 123

xiii

LIST OF ABBREVIATIONS

Amino Acids Abbreviations and One Letter Codes

Arg R Arginine

Asp D Aspartic acid

Glu E Glutamic acid

Gln Q Glutamine

His H Histidine

Lys K Lysine

Ser S Serine

Tyr Y Tyrosine

Others

Ah Aeromonas hydrophila

α-KG α-ketoglutarate

Alk Alkane

Amb Ambiguine

ATP Adenosine Triphosphate

BLAST Basic Local Alignment Search Tool

C Celsius

cm Centimeter

δ Delta, chemical shift (NMR)

d Doublet (NMR)

dd Double Doublets (NMR)

xiv

DMF N,N-Dimethylformamide

DMSO Dimethyl Sulfoxide

Ea or E. amylovora Erwinia amylovora

E. coli Escherichia Coli e.g. Exempli Gratia, “for example”

ESI Electrospray Ionization et al Et alia, “and others”

ExPASy Expert Protein Analysis System

FTMS Fourier Transform Mass Spectrometry g Gram

HEPES 2-[4-(2-HydroxyEthyl)Piperazin-1-yl]EthaneSulfonic acid

HPLC High-performance Liquid Chromatography

IPTG Isopropyl β-D-1-Thiogalactopyranoside

Isn Isonitrile or Isocyanide

K Kilo kDa Kilodalton

KDPG 2-Keto-3-Deoxy-6-Phosphogluconate

L Liter

LB Luria Bertani

LC-MS Liquid Chromatography-Mass Spectrometry

L. pheumophila Legionella pneumophila

LysR Lysine Regulator m Multiplet (NMR)

xv

M Molar

MCS Multiple Cloning Site mg Milligram min Minute mL Milliliter mm Millimeter mM Millimolar mmol Millimole ms Millisecond

MS Mass Spectrometry m/z Mass-to-Charge Ratio

NCBI National Center for Biotechnology Information

μL Microliter

μM Micromolar nm Nanometer nM Nanomolar

NMR Nuclear Magnetic Resonance

ORF Open Reading Frame

Pa or P. aeruginosa Pseudomonas aeruginosa

PLP Pyridoxal 5’-Phosphate

P. luminescens Photorhabdus luminescens ppm Parts per Million (NMR)

PPP Pentose Phosphate Pathway

xvi psi Pounds per Square Inch

Pvc Pyoverdine Chromophore rpm Revolutions per Minute

R.T. Retention Time s Second or Singlet (NMR)

SAM S-Adenosyl-Methionine

S. coelicolor Streptomyces coelicolor

SDS-PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis t Time

TCEP Tris(2-CarboxyEthyl)Phosphine

THF Tetrahydrofolate or Tetrahydrofuran

TLC Thin Layer Chromatography

TMS-CHN2 Trimethylsilyldiazomethane

TPP Thiamine Pyrophosphate

Tris Tris(hydroxymethyl)aminomethane

UV-vis UV-visible v/v Volume/Volume

Wel Welwitindolinone

Xn or X. nematophila Xenorhabdus nematophila

xvii

ABSTRACT

Isonitriles, with an unusual functionality, are very versatile intermediates widely involved in organic synthesis. Interestingly, a variety of isonitriles were encountered in both marine and terrestrial systems as the naturally occurring metabolites. Many of them were isolated as antibiotics; many of them were identified to play important roles in host-pathogen interactions; and a lot more remained uncharacterized. The chemical and functional diversity of these novel compounds may offer important synthetic implications and pharmaceutical potential.

The gene clusters related to the biosynthesis of isonitrile-containing natural products have been identified in many bacteria. They commonly encode two novel proteins, designated PvcA and PvcB. Early functional studies of the proteins identified PvcA as an isonitrile synthase and PvcB as an oxygenase. But the in vitro activities of these proteins remained undetermined until recently.

To demonstrate PvcB as an Fe2+, α-ketoglutarate dependent oxygenase, and as a source of the diversity in isonitrile-containing metabolites, reactions catalyzed by the PvcB homologous proteins from Pseudomonas aeruginosa, Xenorhabdus nematophila and Erwinia amylovora were characterized and compared.

Recombinant PvcB proteins were heterologously expressed and purified to homogeneity. Catalytic activities of three PvcB enzymes were evaluated in the presence of the synthesized substrate tyrosine isonitrile and the co-substrates Fe2+,

xviii

α-ketoglutarate and O2. UV-visible spectra revealed turnovers in all of the three reactions, as well as the difference between the XnPvcB reaction and the other two. Subsequent reaction product determination by 1H-NMR spectroscopy and LC-

MS (liquid chromatography-mass spectrometry) demonstrated that XnPvcB catalyzed an oxidative decarboxylation reaction on tyrosine isonitrile, while the

PaPvcB and EaPvcB reactions retained the carboxyl group. Steady-state and transient-state kinetics were performed to determine the catalytic efficiency and the mechanisms of the PvcB enzymes. And substrate docking and homology modeling were used to compare the active sites of the proteins. Our data suggested the various activities exhibited by the PvcB homologs from different pathways might result from the subtle differences of the proteins in the active sites.

PvcA is a novel isonitrile synthase and is predicted to catalyze the conversion of the amino group into the isonitrile functionality. To determine its in vitro activity and gain a deeper understanding of the mechanism underlying the isonitrile group synthesis, PvcA proteins from Pseudomonas aeruginosa, Xenorhabdus nematophila and Aeromonas hydrophila were expressed and purified. Activity assays were performed with the putative substrates tyrosine and ribulose-5- phoshpate. However, reactions monitored by UV-visible spectroscopy and HPLC

(high-performance liquid chromatography) did not show substrate transformation.

No success was made in the reactions involving crude lysates instead of purified enzymes, either. To determine whether correct substrates were used, in vivo

xix feeding and labeling experiments were performed in the biosynthetic pathway of isocyanovinylphenol to track the product backbone and the isonitrile carbon. Our data demonstrated tyrosine to be one of the substrates for the PvcA enzyme, and suggested the other substrate, which provides the isonitrile carbon might be a metabolite from the pentose phosphate pathway. In the meantime, by overexpressing XnPvcA in E. coli, we detected the PvcA product in both the cell lysate and the culture medium, and identified it as tyrosine isonitrile, further supporting the conclusion that tyrosine is the substrate providing the product backbone. In order to determine the isonitrile carbon precursor, a phosphorylated metabolite library, the central carbon metabolic pathways and the metabolite pool of E. coli cells were examined in search of this substrate, but the substrate was only found in the E. coli cell lysate. Fractionation of the lysate by filtration, dialysis, size-exclusion and ion exchange chromatography identified a macromolecule, like a metabolic enzyme, and a small compound, probably phosphorylated, were both required for PvcA activity.

xx

CHAPTER 1

General Introduction

This chapter introduces the properties of a fascinating functional group in organic chemistry – the isonitrile group (Section 1.1), and the discovery of isonitriles in biological systems (Section 1.2). Investigation of the biosynthesis of the isonitrile group using isotopic labeling studies is discussed (Section 1.3), along with the discovery of the isonitrile biosynthetic gene clusters containing pvcA/B in various organisms (Section 1.4). Functions of the pvcA-containing clusters and the isonitrile metabolites produced by these clusters are introduced

(Section 1.5). Crystal structures of PvcA and PvcB from Pseudomonas aeruginosa are described (Section 1.6), as well as the diverse activities and functions of the non-heme Fe2+，α-ketoglutarate dependent oxygenase family of enzymes to which PvcB belongs (Section 1.7). The predicted catalytic activities of the PvcA and PvcB proteins are discussed (Section 1.8) and a brief overview of the investigations described in Chapter 2 and Chapter 3 is provided (Section

1.9).

1

1.1 Isonitriles in Organic Chemistry

Isonitriles, also called isocyanides or carbylamines, are organic molecules that contain a nitrogen carbon triple bond moiety (Figure 1.1). Isonitriles are unique since they form the only stable entity besides carbon monoxide whose carbon is bonded to a single atom. The isonitrile group can display nucleophilic or electrophilic character, so they have diverse and important roles in organic synthesis[1-3]. For example, isonitriles are widely involved in radical reactions [4,

5], and are active participants in multicomponent reactions, such as the Ugi reaction [6-8]. Nowadays, isonitriles are powerful tools used in the synthesis of heterocyclic compounds and peptides [9, 10]. For instance, radical insertion reactions into isonitriles became an efficient strategy to make the nitrogen heterocycles, recently [11]. In addition, usage of isonitriles to furnish the amide bond has been applied to the synthesis of the important cyclic polypeptide cyclosporine A [9]. Despite the attractions of this group of molecules, some disadvantages exist: isonitriles are relatively unavailable commercially; isonitriles are not simple to synthesize effectively [12]; and isonitriles possess an odor that is described as “horrible” and “extremely distressing”, leaving them unpleasant to handle [13, 14]. These properties have hindered the use of isonitriles and the further study of their chemistry. Thus, researchers are eager to investigate new and effective ways to prepare isonitriles, and also to search for novel isonitriles that are less offensive [15].

2

R N C

Figure 1.1 Best structural representation of the isonitrile group.

1.2 Isonitrile-Containing Natural Products

For a long time, isonitriles were regarded as non-natural compounds. The first reported naturally occurring isonitrile, xanthocillin [16] (Figure 1.2 A), was an antibiotic isolated from Penicillium notatum in 1957. From that day on, more and more isonitriles were encountered in biological systems, including both marine and terrestrial environments. The marine isonitrile metabolites are built on terpene scaffolds (Figure 1.2 B), while most of the terrestrial compounds are derived from amino acids [17-20] (Figure 1.2 A). Many of them are known to have a strong antibiotic, fungicidal, or antimalarial effect [21-23] (Figure 1.2 C).

3

A B CH C 3 OH N H H CH3 N C H HO 2 C N C CH3

Xanthocillin Axisonitrile-1

C H H C C

N N C H

Antibiotic B 371

Figure 1.2 Examples of Isonitriles from nature. (A) Terrestrial isonitrile Xanthocillin. (B) Marine isonitrile Axisonitrile-1. (C) Antibiotic B 371 isolated from Pseudomonas.

Many isonitriles are produced by bacterial pathogens as secondary metabolites [24-26]. For examples, paerucumarin (Figure 1.3 A), an isonitrile functionalized coumarin, was found to be produced by the human pathogen,

Pseudomonas aeruginosa [27]. Rhabduscin (Figure 1.3 B ) is made by the insect pathogen Xenorhabdus nematophila and concentrates at the bacterial cell surface

[24]. Most of these products contain a backbone of tyrosine or tryptophan.

4

Additional modifications make their structures chemically diverse. The existence of the isonitrile-containing products in a wide range of organisms stimulated the interest of researchers to investigate more synthetic possibilities of isonitriles, as well as to explore the bioactivity and the pharmaceutical potential of these natural products [28-30].

A B O

NH OH C O HO N HO HO O HO O O N C

Paerucumarin Rhabduscin

Figure 1.3 Isonitriles produced by bacterial pathogens. (A) Paerucumarin made by Pseudomonas aeruginosa. (B) Rhabduscin produced by Xenorhabdus nematophila.

5

1.3 Isonitrile Biosynthesis

Considering the tedious work required and challenges included to synthesize isonitriles in chemistry, biosynthesis of the isonitriles with various modifications becomes more fascinating. In addition, unraveling the knowledge underlying their biosynthesis may allow for making large quantities of the isonitrile-containing natural products, which would further facilitate their functional studies.

The biogenic origin of the isonitrile group has triggered a lot of interests since the first natural product was isolated and reported [16]. Traditionally, experiments involving feeding cell cultures with isotopically labeled precursors were used to study the isonitrile biosynthesis [31-34]. But consistent data were not obtained with bacteria, fungi, and sponges. Achenbach and Grisebach [31] identified the amino group of L-tyrosine as the source for the isonitrile nitrogen in xanthocillin through the metabolic labeling studies. However, a variety of observations were made regarding the origin of the isonitrile carbon in different natural products.

Bornemann et al. [35] found the intact C2-N group of [2-13C,15N] glycine was incorporated into the isonitrile functionality of the cyanobacterial metabolite hapalindole A. Paur et al. [33] proved L-[methyl-13C] methionine could label the isonitrile carbon of the hezimicins in bacterium Mircomonospora. And Garson [34] demonstrated [14C] cyanide was the source for the isonitrile carbon in di- isocyanoadociane, which is the marine sponge metabolite. But neither methionine

6 nor cyanide was incorporated into the isonitrile group of xanthocillin [31]. Results of these labeling experiments are presented in Figure 1.4.

C C OH N N

C l N HO C N H Xathocillin Hapalindole A

L-tyrosine derived Glycine derived

Not derived from formate and methionine Remains unknown

O C H NH2 N C H H HO HO N H N C C NH H N C 2 C O Di-isocyanoadociane Hazimi cin factor

Methionine derived Cyanide derived

Figure 1.4 Origin of the isonitrile group in different metabolites. Proposed sources of isonitrile N and C are highlighted.

7

Although some information about the origin of the isonitrile group was provided by the early labeling experiments, the low uptake of some precursors into the cells, the scrambling of the label from the precursor due to metabolic pathway crossing, and cell death caused by toxicity have hindered such investigations.

Thus, identifying the genes which are responsible for isonitrile biosynthesis is very crucial.

1.4 Gene Cluster Related to Isonitrile Biosynthesis

In 2005, Brady and Clardy [36] reported the isolation of a novel compound: an isonitrile-containing indole which was detected by screening a soil-derived metagenomic E. coli library for antibiotic activity against Bacillus subtilis. Two genes were identified to be responsible for the synthesis of this compound: isnA and isnB, named after isocyanide or isonitrile. Growth of Escherichia coli cells that independently expressed IsnA or IsnB identified IsnA as the isonitrile synthase that performed the C-N condensation. IsnB operated downstream of IsnA and catalyzed an oxidative decarboxylation reaction.

Feeding studies using IsnA/B expressed in E. coli amino acid autotrophs demonstrated the precursor of this indole as tryptophan [37]. In vivo metabolic labeling studies carried out in E. coli with mutagenesis of genes that code for primary metabolic enzymes identified that the isonitrile carbon was originated from

C2 of the metabolites in the pentose phosphate pathway. Subsequent in vitro

8 reconstitution experiments with purified IsnA, IsnB, tryptophan, α-ketoglutarate and most of the sugars in this pathway (xylulose-5-phosphate was not examined due to unavailability) revealed both ribose-5-phosphate and ribulose-5-phosphate led to the isonitrile production. The tautomerization of ribose-5P and ribulose-5P was used to explain the use of both sugars in the synthesis of the indole isonitrile

(Scheme 1.1). However, only a small quantity of the product was made in the reconstitution system, indicating additional cofactors were needed or better substrates were available.

OH OH H H H N N N - IsnB C OPO3 IsnA O OH O O Ribulose-5P

HO HO N N NH2 O OH C C Tryptophan - CH OPO3 OH OH Ribose-5P

Scheme 1.1 Biosynthetic pathway of isonitrile derivative of tryptophan. IsnA incorporates the C2 carbon of ribulose-5-P into the isonitrile group. IsnB then catalyzes an oxidative decarboxylation reaction.

9

Several years later, Brady described additional biosynthetic clusters containing isnA homologs. They were found in environmental DNA samples and some well-studied organisms [38]. Along with the isnA gene, there was always isnB found. The clusters were expressed heterologously in E. coli or P. aeruginosa, and 9 novel isonitrile metabolites were isolated and characterized. The metabolites were all derived from tryptophan or tyrosine. But they were further modified, as a result of the transformations by different biosynthetic enzymes present in the clusters.

Among the clusters characterized was the P. aeruginosa pvc operon, which was first described in a 1996 publication entitled “Novel pyoverdine biosynthesis gene(s) of Pseudomonas aeruginosa PAO” [39]. This paper linked the operon to the biosynthesis of the pyoverdine chromophore, and the authors named it pvc after pyoverdine chromophore. Another compound called pseudoverdine was found as a by-product in this pathway (Scheme 1.2). Studies in recent years found that pvc deletion strains could still accomplish pyoverdine synthesis under some conditions [27, 40] and the true product of the pvc operon was paerucumarin (2- isocyano-6,7-dihydroxycoumarin), an isonitrile containing compound [27] (Scheme

1.3).

10

HO HO COOH COOH HO NH NH2 2 Tyrosine

HO OH COOH

HO NH2

HOOC HO O O HO N N HO NHCHO HO NH2 Pseudoverdine Chromophore of pyoverdine

Scheme 1.2 Proposed synthesis of pseudoverdine and pyoverdine chromophore from tyrosine under the function of the pvc operon in P. aeruginosa.

11

A

Pseudomonas aeruginosa pvc operon

pvcA pvcB pvcC pvcD

B HO HO HO COOH COOH COOH PvcA NH2 N PvcB N C C Tyrosine PvcC/D

O HO O O COOH

HO N O N C C

Paerucumarin

Scheme 1.3 Function of the pvc operon in P. aeruginosa. (A) pvc operon encodes 4 genes, pvcA, pvcB, pvcC and pvcD. (B) Proposed biosynthesis of paerucumarin by the pvc operon. PvcA converts the amino group of tyrosine into the isonitrile functionality. PvcB catalyzes an oxidative reaction. PvcC/D are responsible for the ring hydroxylation.

12

To clarify the confusion of using both pvc and isn as the names for the isonitrile biosynthetic genes, pvc will be used in this dissertation to designate them.

The pvcA/pvcB homologs are found in a variety of bacterial pathogens, including the plant pathogens Erwinia amylovora and Erwinia carotavora; the insect pathogens Xenorhabdus nematophila and Photorhabdus luminescens; the fish pathogens Vibrio anguillarum, Aeromonas salmonicida and Renibacterium salmoninarum; and the human pathogens Pseudomonas aeruginosa,

Burkholderia mallei and Legionella pneumophila. The synthetic operons ranged from simple systems containing only the pvcA-pvcB pair to more complicated ones which encoded additional biosynthetic enzymes, including monooxygenases, cytochrome C and glycosyltransferases (Figure 1.15). The existence of the gene clusters in a wide range of pathogens indicates that the isonitrile-containing natural products may play an important role in interacting with eukaryotic hosts.

13

PvcA PvcB

Photorhabdus luminescens

Aeromonas salmonicida UGT

Erwinia carotovora UGT

Erwinia amylovora PheA

Xenorhabdus nematophila UGT

Pseudomonas aeruginosa PheA CytC

Burkholderia mallei PheA AsnB QOR CytC

PheA – Monooxygenase UGT – Glycosyltransferases QOR – Quinone oxidoreductase

CytC – Cytochrome C AsnB – Aminotransferases

Figure 1.5 PvcA-containing gene clusters in some common bacterial pathogens. pvcA-pvcB is always found as a pair. Additional biosynthetic enzymes are present in more complex operons.

Genes are annotated with an abbreviation for the gene prediction of an ORF.

14

1.5 Functions of the pvcA/B Clusters and Their Related Metabolites

pvcA and pvcB homologs are widely found in bacteria, as stated above. Even though many of these bacteria are well-studied systems, less is known about the biological functions of the pvcA/B gene clusters and their isonitrile-containing products. In some already characterized examples, the pvcA/B cluster of the human pathogen L. pheumophila is upregulated during biofilm formation [41]. The pvc operon in P. aeruginosa makes paerucumarin, which positively regulates the expression of the cup gene clusters [42], resulting in biofilm development in this bacterium [43, 44]. A biofilm is an assemblage of surface-associated microbial cells that is enclosed in an extracellular polymeric matrix. Biofilm-associated microorganisms have been implicated in over 80% of chronic inflammatory and infectious diseases [45-47]. More recently it has been noted that bacterial biofilms may impair cutaneous wound healing and reduce antibacterial efficiency in healing infected skin wounds [48, 49]. The pvc operon is also adjacent to and regulated by ptxR [50], which is a LysR family transcriptional activator that regulates the production of other virulence factors [51-53]. As such, the pvc gene clusters are tightly related to the pathogenicity of the pathogens.

The two Gram-negative insect pathogens, X. nematophila and P. luminescens, produce the isonitrile rhabduscin (Figure 1.3 B, 1.6) as a strong inhibitor of phenoloxidase [24], which is a copper-dependent metalloenzyme that plays an

15 important role in the invertebrate immunity system [54, 55]. Rhabduscin binds at the periphery of bacterial cells at high concentration. Its location and the potent phenoloxidase-inhibitory activity contribute substantially to virulence. A structurally related molecule, byelyankacin (Figure 1.6), which was first found in Enterobacter sp. B20 [56] and later in metagenomic libraries [38], exhibits strong melanogenesis–inhibitory activity by inhibiting the activity of tyrosinase, which is a phenoloxidase-type enzyme [57]. As such, both rhabduscin and byelyankacin effectively defend against the phenoloxidase activity, a response to the invading pathogens, and thereby support the bacteria’s pathogenic lifestyle.

A B C O N NH OH HO O O

HO O HO O HO OH N C Rhabduscin Byelyankacin

Figure 1.6 Structures of the phenoloxidase inhibitors: (A) rhabduscin and (B) byelyankacin.

16

The mechanisms underlying the bioactivities of isonitrile-containing products were examined in some systems. Cyclohexylisocyanide exhibits strong inhibition of [Fe]-hydrogenase by binding to the iron center of this enzyme with high affinity and thus blocking H2 binding [58]. The marine metabolites di-isocyanoadociane and axisonitrile-3 show potent antimalarial activity by forming a coordination complex with the heme iron, which results in heme detoxification prevention [59].

12-epi-hapalindole E isonitrile is able to inhibit RNA synthesis in bacteria by interacting with the bacterial RNA polymerase and acting as the substrate mimic, which prevents RNA chain elongation [60]. Therefore, the importance of the natural isonitriles may rely on their abilities to bind to the active sites of some key enzymes, to some cofactors or to some important metabolites and thus affect the metabolism in the organism.

The abundance of microbial genomes that contain orthologs for the genes required for the isonitrile synthesis suggests a lot more natural products containing the isonitrile group remain to be recognized, and the functions of many already identified isonitrile metabolites need to be assigned. Characterization of the isonitrile biosynthetic enzymes will facilitate both the identification of the unknown metabolites and the exploration of the biological roles of these compounds.

17

1.6 Structures of the PvcA and PvcB Proteins from P. aeruginosa

Pseudomonas aeruginosa is a common Gram-negative bacterium that can cause disease in plants and animals, including humans. It causes infections in immunocompromised patients and chronic infections in the lungs of cystic fibrosis patients [61, 62], and is associated with a lot of diseases in individuals predisposed to infections as the result of severe burns, wounds, urinary tract or corneal injury

[63-65]. Pseudomonas aeruginosa frequently creates biofilms in the areas that it infects, which make the infection much harder to cure and more antibiotic resistant.

The pvc operon in P. aeruginosa encodes 4 gene products: PvcA, PvcB, PvcC and PvcD (Scheme 1.3). The recombinant overexpression of the pvcABCD operon in P. aeruginosa results in the production of paerucumarin, as described above.

Three-dimensional crystal structures of PvcA and PvcB (Figure 1.7) were determined by Drake and Gulick in 2008 [40], both in the unliganded states.

18

A

Q257 S253 H255 K263

Phosphate R251

B

H259 Succinate H110 Fe2+ D112

Figure 1.7 Crystal structure of (A) PvcA and (B) PvcB from P. aeruginosa. Ribbon diagrams of

PvcA and PvcB are shown with green helices and strands. (A) The PvcA protein is co-crystallized with a phosphate ion. The ion interacts with 5 residues that are highly conserved within the

IsnA/PvcA family of proteins. (B) No ligands are present in the experimental PvcB structure. But

Fe2+ and succinate from the structure of asparagine oxygenase from S. coelicolor, are modeled into the PvcB structure. The side chains in the active sites: His110, Asp112 and His259 interact with iron and are shown in stick representation.

19

The PvcA protein exhibits a Rossmann-like fold in which central 5-stranded parallel β-sheet is surrounded by α-helices on both sides, and a small C-terminal domain contains a second β-sheet with a single helix. The PvcA protein was co- crystallized with a phosphate ion derived from the crystallization cocktail, indicating the enzyme uses a phosphorylated substrate [66, 67]. Also, this phosphate suggests the location of the PvcA active site. It interacts with 5 well-conserved residues: Gln257, Arg251, Ser253, His255 and Lys263, and provides the foundation for the experimental identification and examination of the catalytic mechanism.

Whereas the PvcA protein has no homologs that are characterized structurally,

PvcB belongs to a well-known family of enzymes. The PvcB protein exhibits a jelly roll fold composed of 8 β-strands forming two four-stranded β-sheets, which is typical of the non-heme Fe2+, α-ketoglutarate dependent oxygenase family [68-70].

To identify the active site of the protein, the molecular model of asparagine oxygenase from Streptomyces coelicolor (PDB:2OG7) [71] was superimposed on the structure of PvcB. The side chains in the active sites: His110, Asp112 and

His259 interact with iron and are highly conserved among this family of enzymes.

20

1.7 Non-Heme Fe(II) and α-Ketoglutarate Dependent Oxygenases

The Fe2+, α-ketoglutarate dependent dioxygenases are the largest family of non-heme iron-containing oxygen-activating enzymes [72, 73]. They catalyze an amazing diversity of transformations including hydroxylations, desaturations and oxidative ring closures [74, 75] with important biological implications in both prokaryotic and eukaryotic metabolism. These reactions are widely involved in protein side-chain modifications [76], repair of DNA damaged by methylating reagents [77], oxygen-dependent gene regulation [78], biosynthesis of antibiotics and various plant products [79], and biodegradation of a variety of compounds including herbicides, sulfonates, lipids, and nucleotides [80-83]. Although diverse in their overall chemistry, these enzymes are united by a common mechanism in which dioxygen and α-KG act as the co-substrates and Fe2+ as the cofactor; the oxidation of the primary substrate is coupled to conversion of α-KG into succinate and carbon dioxide [75, 84]. Examples of the reactions catalyzed by the Fe2+, α- ketoglutarate dependent dioxygenases are shown in Scheme 1.4.

21

A O R O R R Prolyl 4-hydroxylase N R N O CO2 2 OH α-KG succinate

NH B NH3 2 CH N N N 3 N

N N N N AlkB HO HO O O H H H H O2 HCHO H H H H O H O H α-KG CO 2 - O P O- succinate O P O O- O- C O O O Clavaminate N Clavaminate N synthase N O COOH synthase O COOH COOH O2 CO2 O CO2 OH 2 α-KG succinate α-KG succinate H O H2N NH 2 NH2 H2O 2

OH D OH HO O Flavone synthase I HO O

O2 CO2 α-KG succinate OH O OH O H2O

Scheme 1.4 Reactions catalyzed by various Fe2+, α-ketoglutarate dependent dioxygenases, resulting in (A) protein modification, (B) repair of alkylated DNA/RNA, (C) biosynthesis of antibiotics, and (D) biosynthesis of plant metabolites.

22

1.8 Predicted Activities of the PvcA and PvcB Proteins

The structure determined for P. aeruginosa PvcB protein is consistent with the early prediction of its function as an oxygenase, but PvcA is a novel protein with only weak homology to the Dit1 protein of Saccharomyces cerevisiae [85]. Dit1 catalyzes the formation of an uncharacterized tyrosine-containing precursor for a spore wall-specific dityrosine-containing macromolecule [86].

In prediction, PvcA is responsible for the conversion of the amino group of tyrosine or tryptophan into the isonitrile group. The isonitrile product is further oxidized by PvcB. Various activities of the PvcB proteins from different pathways, along with the additional functional biosynthetic enzymes in the operons, may result in the organism-specific metabolites.

Although isonitrile-containing metabolites were isolated from the cell cultures expressing the pvcA/B clusters [24, 27, 38], the enzymes were not characterized individually and biochemically in vitro. Little was known about their catalytic activities and chemical mechanisms. Understanding the mechanism of isonitrile biosynthesis may help to foster new developments in isonitrile chemistry and biochemistry. Studying the different reactions catalyzed by PvcB homologs may allow for the exploration of the structural and mechanism relationship of these proteins.

23

1.9 Overview of Investigations

Research presented in this dissertation uses three PvcA homologs from

Pseudomonas aeruginosa, Xenorhabdus nemotophila and Aeromonas hydrophila, and three PvcB homologs from Pseudomonas aeruginosa, Xenorhabdus nemotophila and Erwinia amylovora to study the activities of the PvcA and PvcB proteins and the mechanisms of the reactions they catalyze. These organisms are chosen to provide diversity of enzyme targets within reasonably well-characterized systems. The redundancy of the PvcA proteins will enable experiments to proceed with alternate targets when difficulties arise. The parallel studies with multiple PvcB enzymes will enable the correlation between protein structure and catalytic activity to be determined. For simplicity, these enzymes will be referred to as species name + protein name. For example, PaPvcB is the PvcB enzyme from P. aeruginosa and XnPvcA is the PvcA protein from X. nematophila.

Chapter 2 describes the activity and kinetic studies of three PvcB homologs.

To perform the in vitro experiments, the putative substrate, tyrosine isonitrile was synthesized and examined in the PvcB reactions with the purified enzymes.

Product formation was observed by UV-visible spectroscopy and HPLC. Different products were identified from the PvcB reactions by 1H-NMR spectroscopy and

LC-MS. Steady-state and transient-state kinetic studies were performed to explore the mechanisms behind the different activities of the PvcB homologs. In addition, a couple of substrate analogs were synthesized successfully and tested in the

24

PvcB reactions. One of them will be used in the inhibition study of the PvcB reactions in the future.

Chapter 3 describes the efforts made to characterize the PvcA reaction both in vivo and in vitro. The metabolite tyrosine isonitrile, a relatively unstable compound, was isolated from the culture medium with overexpression of the PvcA protein. Tyrosine isonitrile was demonstrated in Chapter 2 as the substrate for the

PvcB reaction and here as the product for the PvcA reaction. In vitro reconstitution assays failed when multiple PvcA proteins were incubated with the proposed substrates (tyrosine and ribulose-5-phosphate), indicating the substrate(s) was

(were) incorrect. In order to identify the right substrates, a dual expression vector

AhpvcA-XnpvcB/pETDuet-1 was generated for co-expression of PvcA and PvcB in vivo and for constant production of the XnPvcB product, which was stable and well characterized as described in Chapter 2. The creation of the system allowed

13 for in vivo labeling studies using ring-3,5-D2 labeled L-tyrosine and C labeled glucose to track the isonitrile backbone and the origin of the isonitrile carbon. In addition, activities of the purified PvcA enzymes were examined with the metabolites from a phosphorylated metabolite library, the central carbon metabolic pathways, as well as the metabolite pool of E. coli cells. Activities were detected when the PvcA proteins were incubated with the E. coli cell crude lysate but two molecules, a macromolecule and a small compound were both required for these activities. Identities of these two molecules should be further determined in the

25 future.

Together, the presented studies are part of a larger research goal to identify:

1. Mechanism of the isonitrile biosynthetic enzymes in different pathways.

2. Structural basis for the different catalytic activities of PvcB homologs.

3. More physical functions of the pvcA/pvcB clusters and their related metabolites.

26

CHAPTER 2

Characterization of Three PvcB Homologous Proteins, the Non-

Heme Fe2+, α-Ketoglutarate Dependent Oxygenases

2.1 Published Work

Portions of this investigation were reprinted with permission from Jing

Zhu, Geoffrey M. Lippa, Andrew M. Gulick, and Peter A. Tipton. Examining

Reaction Specificity in PvcB, a Source of Diversity in Isonitrile-Containing Natural

Products. Biochemistry, 2015, 54 (16), pp 2659-2669. Copyright 2015 American

Chemical Society.

The Biochemistry article is attached below but is not counted in the number of pages of this thesis.

27

Article

pubs.acs.org/biochemistry

Examining Reaction Specificity in PvcB, a Source of Diversity in Isonitrile-Containing Natural Products † ‡ ‡ § † Jing Zhu, Geoffrey M. Lippa, Andrew M. Gulick, , and Peter A. Tipton*, † Department of Biochemistry, University of Missouri, Columbia, Missouri 65211, United States ‡ Hauptman-Woodward Medical Research Foundation, Buffalo, New York 14203, United States § Department of Structural Biology, University at Buffalo, Buffalo, New York 14203, United States

*S Supporting Information

ABSTRACT: Many bacteria produce isonitrile-containing natu- ral products that are derived from aromatic amino acids. The synthetic clusters that control biosynthesis most commonly encode two enzymes, designated PvcA and PvcB, as well as additional enzymes that direct synthesis of the natural product. The PvcA enzyme installs the isonitrile moiety at the amino group of either tyrosine or tryptophan, as dictated by the particular pathway. The common pathway intermediate produced by PvcA is directed toward different ultimate products by PvcB, a member of the family of Fe2+, α-ketoglutarate-dependent oxygenases. To continue our investigation of the structural and functional properties of the isonitrile biosynthetic pathways, we present here a study of the PvcB homologues from three organisms. Two pathways, derived from Pseudomonas aeruginosa and Xenorhabdus nematophila, produce known products. A third PvcB homologue from Erwinia amylovora is part of an uncharacterized pathway. Our results demonstrate the diversity of reactions catalyzed. Although all PvcB enzymes catalyze the hydroxylation of the tyrosine isonitrile substrate, the elimination of the hydroxyl in Pseudomonas and Erwinia is driven by deprotonation at Cα, resulting in the initial production of an unsaturated tyrosine isonitrile product that then cyclizes to a coumarin derivative. PvcB from Xenorhabdus, in contrast, catalyzes the same oxygenation, but loss of the hydroxyl group is accompanied by decarboxylation of the intermediate. Steady-state kinetic analysis of the three reactions and a docking model for the binding of the tyrosine isonitrile substrate in the PvcB active site highlight subtle differences between the PvcB homologues.

wide variety of isonitrile-containing natural products has aeruginosa showed that pvc mutants are still capable of A been identified from marine and terrestrial microorgan- pyoverdine production and growth under iron limitation,8,9 isms. Alkaloids, terpenes, and amino acids are among the parent casting doubt on the functional assignment of the pvc operon. compounds that are modified to contain isonitriles, and Following the identification of an isonitrile derivative of 1,2 biological activities include antibacterial and cytotoxic effects. tryptophan produced by a homologous gene cluster from an While attention has naturally focused on the origin of the environmental DNA source,4 characterization of the P. isonitrile functionality, the remaining enzymes in the operon 3 aeruginosa pvc operon demonstrated that the true biosynthetic can diverge, resulting in a wide variety of natural products. product is paerucumarin (2), or 2-isocyano-6,7-dihydroxycou- Most commonly, the synthetic operon contains a gene that is 3 2+ α marin, which is derived from tyrosine. Paerucumarin does in annotated as an Fe , -ketoglutarate-dependent oxygenase and fact share some features with the pyoverdine chromophore and is often labeled pvcB. The PvcA enzymes catalyze installation of was detected in its hydrated form in cultures of pvc- the isonitrile moiety into either tyrosine or tryptophan. The overexpressing cells.7 Analysis of homologous proteins has PvcB oxygenases function downstream of the isonitrile 4 demonstrated that PvcA catalyzes the installation of the synthase and may be responsible for the chemical diversity 4 ff isonitrile. Only a few products of pvc biosynthetic operons of the products. Much e ort has been devoted to unraveling the 3 iron-dependent chemistry that leads to substrate hydroxylation have been chemically characterized, including the glycosylated in reactions catalyzed by enzymes within this family.5 derivative rhabduscin (3), that is produced by Xenorhabdus The pvc operon of Pseudomonas aeruginosa, harboring four nematophila (Scheme 1). genes, was originally implicated in the maturation of the chromophore of the siderophore pyoverdine,6,7 resulting in the Received: March 9, 2015 use of pvc for pyoverdine chromophore in the names of these Revised: April 9, 2015 and homologous genes. Subsequent genetic analysis in P.

© XXXX American Chemical Society A DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

Scheme 1 colonies were picked and used to inoculate 10 mL cultures, which grew overnight at 37 °C, and were then used to inoculate flasks containing 1 L of LB medium supplemented with 100 μg/mL carbenicillin. The cultures were grown at 37 °C with rotary shaking (250 rpm) to an optical density of 0.4−0.6 and then placed on ice to lower the temperature to ∼16 °C. IPTG was added to a final concentration of 0.3 mM to induce recombinant protein expression, and the growth was continued at 16 °Cfor16−20 h. The cells were harvested by centrifugation at 3993g for 20 min, yielding 4−5 g of cell paste per liter of medium. The cell paste was stored at −80 °C until it was needed. Frozen cell paste (typically 4−5 g) was suspended in 20 mL of buffer composed of 50 mM Tris (pH 7.5), 250 mM NaCl, 20 mM imidazole, and 0.2 mM TCEP. The suspension was kept on ice and mixed with 1 mg/mL lysozyme for 30 min. The cells were lysed by sonication, and the cell debris was removed by ultacentrifugation at 40000 rpm for 40 min. The clarified supernatant was loaded onto a 5 mL The amino acid sequence and the structure of P. aeruginosa Ni-NTA column (GE HisTrap HP) pre-equilibrated in the 2+ α PvcB suggest that it is an Fe , -ketoglutarate-dependent same buffer that was used for lysing the cells and washed with enzyme.8 Similarly, a PvcB orthologue is found in the X. ff 10 25 mL of bu er at a rate of 2 mL/min. A step gradient nematophila gene cluster responsible for rhabduscin synthesis. consisting of 65 mL of lysis buffer containing 50 mM imidazole, Neither PvcB orthologue has been characterized biochemically, followed by 55 mL of lysis buffer with 300 mM imidazole, was although the ultimate product from the enzymes encoded by used to elute the protein. The purity of the protein was assessed the operon is known. We report here the reactions catalyzed by by SDS−PAGE, and the pooled protein fractions were dialyzed these PvcB enzymes. Additionally, a PvcB orthologue in an against 50 mM Tris (pH 7.5), 250 mM NaCl, and 0.2 mM uncharacterized operon from Erwinia amylovora, a Gram- fi TCEP. TEV protease (1 mg per 50 mg of target) was included negative plant pathogen that causes re blight in certain fruit with the sample during dialysis, to cleave the affinity tag. The trees, was also chosen for study on the basis of the differences protein was concentrated using an Amicon Ultra 10K MWCO in the operon organization, and to build a diverse pool of centrifugation filter and loaded back onto the Ni-NTA column enzyme targets. The identification of the product of this cluster to remove the protease and cleaved tag. The flow-through is being pursued separately. We have previously determined the sample was purified with a Superdex 200 size-exclusion column structures of PvcA and PvcB from P. aeruginosa in the absence ff of ligands.8 Subsequent efforts to grow crystals of PvcB with (GE HiLoad 16/600 Superdex 200 pg) using bu er containing informative ligands were unsuccessful; however, a higher- 40 mM Tris (pH 7.5), 50 mM NaCl, and 0.2 mM TCEP, to resolution structure was obtained that is included herein and remove contaminants. The single peak was concentrated to 5 was used for molecular docking experiments. For the sake of mg/mL for crystallization. Protein concentrations were simplicity, the enzymes from P. aeruginosa, X. nematophila, and determined spectrophotometrically using an extinction coef- ficient of 40000 M−1 cm−1 at 280 nm.8 E. amylovera will be termed PaPvcB, XnPvcB, and EaPvcB, fi respectively. The expression and puri cation of XnPvcB and EaPvcB followed the same procedures that have been described for ■ MATERIALS AND METHODS PaPvcB. The appropriate extinction coefficients were calculated 11 Buffers and common biochemicals were purchased from Fisher from the amino acid sequences using ProtParam. Scientific or Sigma and were used without further purification. Synthesis of Tyrosine Isonitrile, 1. Tyrosine isonitrile was synthesized by N-formylation of tyrosine, followed by Organic solvents were stored over molecular sieves, and 12,13 imidazole used in column chromatography was recrystallized dehydration of the N-formyl group with Burgess reagent. L- Tyrosine (10 mmol, 1.8 g) was dissolved in formic acid (5 mL), from ethyl acetate. 14 Cloning of PvcB Genes. The cloning and construction of and acetic formic anhydride (40 mmol, 3.5 g) was added. The the expression vector for PaPvcB have been described reaction vessel was capped with a drying tube and placed on ice previously.8 The genes for XnPvcB and EaPvcB were for 4 h. The reaction was quenched by addition of 20 mL of synthesized commercially (Genewiz, Inc.); sequences encoding water, and the solution was concentrated to dryness under NdeI and XhoI restriction sites were incorporated at the 5′ and reduced pressure. Ice-cold hydrochloric acid (1 N, 20 mL) was 3 ′ ends, respectively. Sequences of XnPvcB added to the residue to dissolve the remaining tyrosine. The 1 (WP_010845413.1) and EaPvcB (WP_004157574.1) were product was collected by filtration and washed with water: H δ identified through an NCBI BLAST search originating from the NMR (DMSO-d6) 7.96 (s, 1H), 7.07 (d, J = 8.4 Hz, 2H), PaPvcB sequence (AAC21672.1). The synthetic genes were 6.76 (d, J = 8.4 Hz, 2H), 4.43 (m, 1H), 2.94 (dd, J = 13.8, 5.0 subcloned into the pET-15bTEV plasmid. Escherichia coli 10G Hz, 1H), 2.74 (dd, J = 13.8, 9.0 Hz, 1H). cells were transformed with the expression vectors and stored Formyl tyrosine (1 mmol, 0.22 g) was methylated with in 20% glycerol at −80 °C. trimethylsilyldiazomethane (5 mmol of a 2.0 M solution in Expression and Purification of PaPvcB for Crystal- diethyl ether) and MeOH (12 mmol) in dry dichloromethane lization. The PaPvcB expression vector was used to transform (8 mL) at room temperature for 10 min.15 The solvents were 1 δ E. coli BL21(DE3) cells. Transformants were selected by removed by rotary evaporation: H NMR (DMSO-d6) 7.98 growth on LB plates containing carbenicillin. Individual (s, 1H), 6.99 (d, J = 6.6 Hz, 2H), 6.67 (d, J = 6.6 Hz, 2H), 4.51

B DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

(m, 1H), 3.61 (s, 3H), 2.93 (dd, J = 13.8, 5.7 Hz, 2H), 2.78 mechanisms, and eq 3 is for rapid equilibrium ordered (dd, J = 13.8, 9.0 Hz, 2H). mechanisms. ∼ The remaining residue ( 1 mmol of formyl tyrosine methyl VAB ester) was dissolved in dry dichloromethane (8 mL). Burgess v = +++ reagent [(methoxycarbonylsulfamoyl)triethylammonium hy- KKia b KA b KB a AB (1) droxide, inner salt] (1 mmol, 0.24 g) was added and the VAB solution heated to reflux for 2 h.13 The product was purified by v = KA++ KB AB column chromatography using 50 g of silica gel equilibrated in ba (2) CHCl3. After the sample had been loaded, the column was VAB washed with 100 mL of CHCl and then 200 mL of a CHCl / v = 3 3 KK++ KA AB MeOH mixture (98:2). Tyrosine isonitrile methyl ester eluted ia b b (3) when the column was washed with 100 mL of the CHCl3/ Analysis of the product spectra obtained in the PaPvcB 1 δ MeOH mixture (97:3): H NMR (DMSO-d6) 7.03 (d, J = 8.4 reaction utilized singular-value decomposition, as implemented Hz, 2H), 6.71 (d, J = 8.4 Hz, 2H), 5.02 (m, 1H), 3.73 (s, 3H), in KinTek Explorer, version 4.0.16 Details of the analysis are 3.08 (dd, J = 14.1, 5.0 Hz, 1H), 2.97 (dd, J = 14.1, 7.5 Hz, 1H). provided in the Supporting Information. The material from the previous step (∼0.5 mmol) was Reaction Product Identification. Large scale PvcB dissolved in 1.5 mL of THF; 0.5 mL of 2.0 M LiOH was added, reactions were conducted in 2 mL solutions containing 10 and the solution was incubated at room temperature for 16 h to mM tyrosine isonitrile, 10 mM α-ketoglutarate, 5 mM ascorbic ∼ μ hydrolyze the methyl ester. Attempts to desalt tyrosine acid, 0.5 mM (NH4)2Fe(SO4)2, and 50 M PvcB in 50 mM isonitrile by anion exchange and size-exclusion chromatography HEPES (pH 7.8) and 150 mM NaCl. The XnPvcB reaction resulted in hydration of the isonitrile group, so the solution was mixture was incubated at room temperature for 2 h and then lyophilized and the product used without further purification: extracted with two 4 mL aliquots of ethyl acetate. The organic 1H NMR (D O) (Figure S1 of the Supporting Information) δ phases were combined, and the solvent was removed by rotary 2 μ 7.07 (d, J = 7.5 Hz, 2H), 6.98 (d, J = 7.5 Hz, 2H), 4.32 (m, evaporation. The residue was dissolved in 600 L of methanol- − 13 δ d and analyzed by 1H NMR spectroscopy. A sample of the 1H), 3.05 2.86 (m, 2H); C NMR (D2O) 172.8, 155.6 4 (isonitrile C), 153.1, 130.6, 127.3, 115.6, 60.9, 37.6; FT-IR ethyl acetate extract of the XnPvcB reaction mixture was also − − (KBr pellet) 2170 cm 1 (isonitrile group). prepared in methanol and analyzed by LC MS (described in PvcB Activity Assay. Activities of the purified PvcB detail in the Supporting Information). The PaPvcB and EaPvcB proteins were evaluated in the presence of synthetic tyrosine reactions were quenched 5 min after addition of the enzyme by isonitrile and the cosubstrates α-ketoglutarate and Fe2+. All introducing the methylating reagent TMS-CHN2.The reactions were conducted under aerobic conditions at 25 °C. methylation reaction was performed by injecting 0.5 mL of Typical assay mixes contained 200 μM tyrosine isonitrile, 100 the PvcB reaction solution into 2 mL of ethyl acetate μ α μ containing 500 mM methanol and 250 mM TMS-CHN2. M -ketoglutarate, 10 M (NH4)2Fe(SO4)2, 50 mM HEPES (pH 7.8), 150 mM NaCl, and PvcB. The reactions were The solution was vortexed for 10 min, and after the phases had separated, the organic phase was removed and evaporated to initiated by adding PvcB and were monitored spectrophoto- dryness. The residue was dissolved in 0.5 mL of methanol for metrically at 240−500 nm. Scans of the reaction mixtures were LC−MS analysis. To detect the formation of succinate in the acquired every 0.5 min for ∼10 min at room temperature. PvcB reactions, another large scale reaction was conducted, and Control reactions were performed by omitting one of the after incubation at room temperature for 2 h, aliquots of CHCl3 substrates or the enzyme from the reaction mixture. PvcB were added to denature the protein. Following removal of the activities were also tested using formyl tyrosine or tyrosine in ff enzyme by centrifugation, the remaining solution was passed place of tyrosine isonitrile. In addition, the e ects of ascorbic through a short Chelex 100 (Bio-Rad) column to remove the acid, TCEP, and dithiothreitol on the reactions were examined. 3+ fl 1 Fe . The ow-through was dried and dissolved in D2O for H Kinetic Analysis. Steady-state kinetic assays were con- NMR analysis. ducted in solutions containing 50 mM HEPES (pH 7.8), 150 Determination of the Structure of PaPvcB and μ μ α mM NaCl, 10 M (NH4)2Fe(SO4)2,50 M ascorbic acid, - Computational Modeling. − μ − μ Initial crystallization conditions ketoglutarate (5 30 M), and tyrosine isonitrile (5 50 M), were established by Drake and Gulick.8 Using similar ° thermostated at 25 C. Reactions were monitored using a conditions, we grew multiple crystals in the presence of ligands − Carey 50 Bio UV vis spectrophotometer. Product formation or probed the crystals with soaking conditions; no density was detected by the increase in absorbance at 310 nm for the consistent with bound ligands was ever observed. A higher- XnPvcB reaction and at 360 nm for the PaPvcB and EaPvcB resolution model that is used for molecular docking experi- ffi −1 −1 reactions. An extinction coe cient of 4300 M cm was used ments is presented. Crystals around 200 μm × 200 μm were to convert the absorbance to product concentration in the cryoprotected by being soaked for 10−15 min in mother liquor −1 −1 XnPvcB reaction, and a value of 2885 M cm was used in the [0.1 M BTP (pH 7.5), 5% PEG 20000, and 0.1 M sodium PaPvcB and EaPvcB reactions. citrate tribasic] with 34% glycerol. Data collection took place at Data Analysis. Initial velocity kinetic data were fit to the beamline 23ID-B of the Advanced Photon Source (APS) using appropriate functions using GraFit version 5.0 (Erithacus a MAR 300 detector. Data were indexed, integrated, and scaled Software). In the following functions, A and B are the with iMOSFLM and AIMLESS.17 The PaPvcB structure was concentrations of α-ketoglutarate and tyrosine isonitrile, determined by molecular replacement using the protein atoms respectively, and Ka and Kb are their Michaelis constants. Kia of Protein Data Bank (PDB) entry 3EAT. This model was is the dissociation constant for α-ketoglutarate, and V is the subjected to iterative model building18 and refinement with maximal velocity. Equation 1 describes reactions that follow a both REFMAC and PHENIX19 using the native data to 2.05 Å steady-state sequential mechanism; eq 2 is for ping-pong resolution. The final model contains 2207 protein atoms, 150

C DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article waters, and three molecules of glycerol. The diffraction data statistics and refinement statistics are listed in Table 1.17 The final coordinates and structure factors have been deposited in the PDB as entry 4YLM.

Table 1. Diffraction Data Statistics and Refinement Statistics of PaPvcB

PDB entry 4YLM beamline APS 23-ID-B wavelength (Å) 1.0323 space group P6322 unit cell dimensions [a, b, c (Å)] 125.0, 125.0, 107.2 no. of molecules per asymmetric unit 1 resolution range (Å) 40−2.0 (2.07−2.0)a no. of observations 414157 no. of unique reflections 32890 multiplicity 12.6 completeness (%) 100 (100)a ⟨I/σ⟩ 12.6 (4.6)a a Rmerge (%) 12.2 (53.8) a Rpim (%) 5.0 (16.4) CC(1/2) 0.996 (0.925) Refinement a Rfactor (%) 18.73 (21.03) a Rfree (%) 21.88 (22.18) no. of protein molecules/water atoms 2207/150 root-mean-square deviation for bond distances (Å) 0.007 root-mean-square deviation for bond angles (deg) 1.055 Wilson B factor (Å2) 30.2 average B factor (Å2) protein (all, main, side) 34.7, 33.4, 35.4 glycerol 47.8 solvent 37.0 Ramachandran analysis (%) favored 98.1 allowed 1.9 outliers 0 Molprobity ClashScore 4.82 (98th percentile) aValues in parentheses are for the highest-resolution shell (2.07−2.02 Å for diffraction data and 2.11−2.05 Å for refinement). Figure 1. Spectroscopic changes during PvcB-catalyzed reactions. Each reaction mixture contained 200 μM tyrosine isonitrile, 100 μM α- 20 μ Docking was performed with AutoDock Vina and ketoglutarate, 10 M (NH4)2Fe(SO4)2, 150 mM NaCl, and 50 mM DOCK,21 implemented through the Chimera suite of HEPES (pH 7.8). The spectrum acquired at time zero is colored red in programs.22 The PvcB protein was aligned with clavaminate each panel; subsequent spectra shown were acquired at 30 s and 1 min synthase (PDB entry 1DRT) containing Fe2+, α-ketoglutarate, and then at 1 min intervals up to 10 min: (A) 0.46 μM PaPvcB, (B) μ μ and proclavaminic acid.23 The Fe2+ and α-ketoglutarate were 0.26 M XnPvcB, and (C) 0.61 M EaPvcB. extracted from this file and incorporated into the PvcB structure, which served as the receptor for the docking. A failed to show any spectral changes. Some turnover was search space surrounded the active site pocket with dimensions observed in the absence of added Fe2+; the extent of the of 25 Å × 25 Å × 25 Å. The top docking poses were examined reaction was stimulated by addition of ascorbate, suggesting for score and positioning of the Cβ protons toward the iron that the enzyme is isolated with some Fe2+ bound. The PaPvcB atom and comparison to the site of hydroxylation within reaction was explored using formyl-tyrosine in place of tyrosine proclavaminic acid of PDB entry 1DRT. isonitrile, and no reaction was observed. 1H NMR spectro- scopic analysis of competent reaction mixtures revealed that ■ RESULTS succinate was formed from α-ketoglutarate during the reaction. Identification of the Substrates for PvcB Proteins. The Thus, as predicted from the sequence and crystal structure of synthesis of authentic tyrosine isonitrile, 1, allowed us to test PaPvcB, the PvcB proteins are α-ketoglutarate-dependent whether it was the correct substrate for the PvcB proteins. enzymes. Spectroscopic analysis demonstrated dramatic spectral changes Identification of the Products of the PvcB Reactions. when each PvcB enzyme was incubated with tyrosine isonitrile, The spectral changes that accompanied the PaPvcB and Fe2+, and α-ketoglutarate under aerobic conditions (Figure 1). EaPvcB reactions were more complex than those observed in Control reactions lacking tyrosine isonitrile or α-ketoglutarate the XnPvcB reaction, with a peak at 430 nm becoming

D DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article prominent at longer reaction times. When a relatively large the SVD analysis were used to calculate reaction velocities in amount of PaPvcB was utilized in the reaction, all of the the steady-state kinetic studies. The species whose maximal tyrosine isonitrile was turned over within a few seconds, and absorbance is at 280 nm is the true product of the PaPvcB the subsequent spectral changes were monitored. SVD analysis reaction, and it decays nonenzymatically to the species with an revealed that two species were predominant as the solution absorbance peak at 430 nm. aged. Details of the analysis are provided in the Supporting Identification of the products formed in each reaction from Information. The reconstructed spectra of the two species are tyrosine isonitrile was achieved using 1H NMR spectroscopy ffi shown in Figure 2, and the extinction coe cients determined in and LC−mass spectrometry. The XnPvcB product was stable and readily extracted into ethyl acetate, facilitating its identification. The 1H NMR spectrum of the isolated product is shown in Figure 3 and has four aromatic protons that appear as doublets centered at 7.31 and 6.78 ppm (J = 9.0 Hz) and two doublets, each corresponding to one proton, centered at 6.96 and 6.45 ppm (J = 15.0 Hz). This matches the published spectrum of (E)-4-(2-isocyanovinyl)phenol, 5.3 The positive ion mass spectrum showed peaks at m/z 146.18, 164.05, and 178.10 (Figure 3). The low-mass peak is from isocyanovinyl- phenol; the second peak is either a water adduct or the product in which the isonitrile group has become hydrated, and the high-mass peak is presumed to arise from the methanol adduct of isocyanovinylphenol. The instability of the product of the PaPvcB and EaPvcB reactions hindered its identification. Initial attempts to extract it into ethyl acetate failed, as did attempts to accumulate sufficient Figure 2. Component spectra of species in the reaction catalyzed by material for a 1H NMR spectrum. When the reaction was PaPvcB, determined by SVD analysis of time course spectra. The spectrum of the product of the PaPvcB reaction is colored green, and quenched with a methylating reagent, however, the product was the spectrum of the species to which it decays nonenzymatically is stabilized and rendered sufficiently hydrophobic to be extracted colored blue. The spectrum of tyrosine isonitrile (red) is shown for into organic solvent. LC−MS analysis of the PaPvcB and comparison. EaPvcB reactions yielded very similar results (Figure 4). A single species at m/z 204.0 was observed in each instance, which is assigned to 3-(4-hydroxyphenyl)-2-isocyanoacrylate

fi 1 Figure 3. Identi cation of the product of the XnPvcB reaction. (A) For the H NMR spectrum (MeOH-d4 solvent), the reaction was conducted and the product isolated as described in the text. (B) Mass spectrum of the isolated product.

E DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

Figure 4. Mass spectra of products trapped by methylation of the (A) PaPvcB and (B) EaPvcB reactions. methyl ester, indicating that the product of the PaPvcB and High-Resolution Structure of PaPvcB and Molecular EaPvcB reactions is 3-(4-hydroxyphenyl)-2-isocyanoacrylate, 6. Docking. We have previously presented a crystal structure of Steady-State Kinetic Characterization of PvcB Reac- PaPvcB.8 We have continued to optimize crystals, partly in an tions. The steady-state kinetic parameters governing the attempt to identify the structure of the protein bound to reactions catalyzed by the three PvcB proteins characterized substrates or substrate analogues. Despite numerous soaking in this study are listed in Table 2. Complete kinetic and cocrystallization experiments with tyrosine, tyrosine isonitrile, iron, and α-ketoglutarate, none of the data sets that Table 2. Kinetic Parameters of PvcB Reactions were collected and processed gave evidence of ligands in the PvcB active site. Through these efforts, however, a higher- α Km, -ketoglutarate Km, tyrosine isonitrile −1 μ μ resolution data set was collected. This structure is similar to the enzyme kcat (s ) ( M) ( M) a ± ± ± earlier structure, containing disordered loops at positions PaPvcB 2.3 0.2 23 49315 Asp99, Ser169−Tyr173, and Val200−Glu211. The least- XnPvcBb 5.1 ± 0.3 5.7 ± 1.1 13 ± 2 c ± ± ± squares displacement of the two structures, calculated for 270 EaPvcB 0.22 0.01 4.8 0.8 16 2 Cα positions, is 0.2 Å. Although it is globally similar to the aKinetic parameters derived from a fittoeq1.bKinetic parameters fi c fi earlier structure, we used this higher-resolution structure for derived from a ttoeq2. Kinetic parameters derived from a ttoeq modeling and docking with the tyrosine isonitrile substrate. 3. Because we were unable to obtain crystals with protein bound to ligand, we used computational docking to identify a characterization was not conducted, but the concentrations of position for binding the substrate tyrosine isonitrile. We first α-ketoglutarate and tyrosine isonitrile were systematically aligned PaPvcB with homologous structures to identify the varied. The data were visually inspected before being fitto binding position of the iron and α-ketoglutarate. Ligand the appropriate function. Interestingly, the double-reciprocal positions were taken from the structure of clavaminate synthase plots showed systematic differences that are often indicative of (PDB entry 1DRT). With these ligands fixed in the PaPvcB distinct kinetic mechanisms (Figure 5). The XnPvcB reaction structure, we docked a molecule of tyrosine isonitrile. yielded a family of lines intersecting to the left of the origin. Comparison of liganded structures of homologous proteins The lines were parallel in the case of PaPvcB, and in the illustrates variation in the mode of substrate binding. We EaPvcB reaction, the lines intersected on the y-axis. compared the docking poses from multiple runs with the

F DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

aromatic ring of the ligand binds in a hydrophobic pocket formed by Met114, Tyr115, and Leu116, with potential contributions from Trp83, as well. The carboxyl group of the tyrosine isonitrile does not interact directly with any positively charged residues. Arg168 is present nearby at the start of a disordered loop and could interact with the carboxylate. The pro-S hydrogen from the Cβ position is directed toward the iron, potentially identifying the site of hydroxylation in the proposed PvcB mechanism. The distance from Cβ to the docked iron is 3.4 Å. Given the limitations of the accuracy from molecular docking, this distance is similar to 4.4 Å, the value observed for the distance between the iron and the analogous carbon atom in the clavaminate synthase substrate analogue N- acetyl arginine.23 This structure (PDB entry 1DRY) contains a water molecule coordinated to the iron trans to the His259 equivalent residue. The pose of the docked tyrosine isonitrile is oriented such that hydroxylation at the pro-S position, and subsequent anti elimination upon abstraction of the α-proton, would result in the E configuration around the double bond, which is required for cyclization to form the coumarin ring in the PaPvcB-catalyzed reaction. ■ DISCUSSION Paerucumarin, 2, and rhabduscin, 3, are examples of natural products derived from amino acids in which the amino group has been converted to an isonitrile. Paerucumarin appears to act as a signal molecule used by the bacteria to coordinate events in biofilm synthesis.26 Rhabduscin is a potent inhibitor of insect phenoloxidase and plays a role in overcoming the innate immune system of insect hosts that the bacteria encounter.27 In broad terms, both compounds are involved in intercellular communication; the abundance of microbial genomes that contain orthologues for the genes required for isonitrile synthesis suggests that many novel compounds or novel applications of known isonitrile-containing natural products remain to be recognized. Characterization of the biosynthetic enzymes involved will facilitate attempts to gain a deeper understanding of the biological roles of these compounds. Operons that encode enzymes predicted to be involved in isonitrile metabolite synthesis have been identified on the basis of homology to a putative isonitrile synthase, e.g., IsnA or PvcA.4,9 In each case, the operon also includes a gene encoding a putative α-ketoglutarate-dependent enzyme that is believed to act on the product of the isonitrile synthase reaction. In no instance has the activity of one of these enzymes been demonstrated. In this work, we have characterized the catalytic activity of three Fe2+, α-ketoglutarate-dependent enzymes found in operons demonstrated or believed to be responsible for synthesis of isonitrile-containing metabolites in bacteria. The crystal structure of PaPvcB revealed its similarity to known Fe2+, α-ketoglutarate-dependent dioxygenases; however, no ligands were present in the structure that was determined, and the catalytic activity had not been tested.8 Given that Figure 5. Double-reciprocal plots of the reactions catalyzed by PvcB tyrosine isonitrile is the expected product of the PvcA reaction, enzymes. The concentration of tyrosine isonitrile was varied between 5 it seemed reasonable to investigate whether tyrosine isonitrile and 50 μMat(○)5,(●) 10, (□) 20, and (■)50μM α-ketoglutarate. was a substrate for PvcB. Dramatic spectroscopic changes were (A) PaPvcB, fit to eq 2. (B) XnPvcB, fit to eq 1. (C) EaPvcB, fittoeq evident when the enzyme was incubated under aerobic 3. conditions in the presence of tyrosine isonitrile, Fe2+, and α- ketoglutarate. No change was observed in control reactions structures of clavaminate synthase23,24 and carbapenem lacking tyrosine isonitrile or α-ketoglutarate; some activity was synthase.25 From the potential docking poses identified, one evident when Fe2+ was omitted, and the activity was enhanced appeared to be most reasonable in terms of the free energy of by addition of ascorbate (data not shown). The stimulation of binding and chemical plausibility (Figure 6). In this model, the activity by ascorbate has been observed frequently in other

G DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

Figure 6. Stereorepresentation of the model of tyrosine isonitrile docked into the active site of PaPvcB. The Fe and α-ketoglutarate (yellow) from clavaminate synthase are included in the model. Tyrosine isonitrile is colored magenta; the pro-S hydrogen at Cβ is colored cyan. Disordered loops connecting Arg168 to Ser174 and Pro199 to Phe212 are shown with dashed lines.

Fe2+, α-ketoglutarate-dependent dioxygenases.28,29 PaPvcB was The instability of the product of the PaPvcB reaction, and the not active when tyrosine isonitrile was replaced in the reaction fact that it could not be extracted into organic solvent, mixture with tyrosine or formyl-tyrosine. Finally, the assign- suggested that the carboxyl group in tyrosine isonitrile was still ment of PaPvcB as a member of the Fe2+, α-ketoglutarate- present. To generate sufficient material for characterization, a dependent dioxygenase superfamily was confirmed by 1H NMR relatively large amount of PaPvcB was added to the substates, spectroscopic analyses of samples, which clearly showed the and the reaction was quenched within a few minutes by conversion of α-ketoglutarate to succinate. addition of trimethylsilyldiazomethane. Treatment of the The more interesting question was the identity of the reaction solution with the methylating agent both stabilized products formed from tyrosine isonitrile by each of the three the product and rendered it capable of being extracted into enzymes. The XnPvcB reaction was spectroscopically distinct ethyl acetate. We did not isolate enough material for 1H NMR from the PaPvcB and EaPvcB reactions, and monitoring the analysis, but mass spectrometry was consistent with identi- reaction by HPLC and UV−visible spectroscopy indicated that fication of the esterified product as 3-(4-hydroxyphenyl)-2- the product was stable. 1H NMR and mass spectrometric isocyanoacrylate methyl ester. Thus, the product of the analyses were consistent with the assignment of the product as enzymatic reaction is 6, formed by oxidation of tyrosine 5, which arises from oxidative decarboxylation of tyrosine isonitrile. isonitrile. The natural product rhabduscin that is made by X. The UV−visible spectroscopic properties of the EaPvcB nematophila contains the isocyanovinylphenol moiety, so there reaction product were identical to those of the PaPvcB reaction is satisfying congruence between the in vitro XnPvcB reaction product, as was its behavior on HPLC. Mass spectrometry was product and the final product of the metabolic pathway in also consistent with the conclusion that the product (trapped as which it participates. its methyl ester) was 6. EaPvcB was identified by database Paerucumarin results from oxidative cyclization of tyrosine mining, and the product of the pathway in which it operates is isonitrile, and the PaPvcB reaction yielded a species with unknown. The gene is found in an operon along with a pvcA absorbance at long wavelengths, as would be expected for a homologue and an open reading frame that is annotated as a coumarin. However, close examination of reactions conducted flavin-dependent monooxygenase. It seems likely, then, that the with small amounts of enzyme revealed a lag in the appearance product of the pathway will prove to derive from oxidative of the species that absorbed at long wavelengths; conducting cyclization of the PvcB product. Of the PvcB orthologues the reaction with a large amount of enzyme showed that characterized in this work, the two that are present in operons tyrosine isonitrile was rapidly converted to another species that without a glycosyltransferase catalyze decarboxylation of did not absorb at long wavelengths but subsequently reacted to tyrosine isonitrile. XnPvcB catalyzes decarboxylation of tyrosine form the species distinguished by its absorbance peak at 430 isonitrile, and rhabduscin, the product of the pvc operon in X. nm. The time course of this conversion was monitored over the nematophila, is the glycosyl conjugate of the PvcB product. wavelength range from 240 to 400 nm, and singular-value Similarly, byelyankacin is a natural product isolated from decomposition was employed to deconvolute the spectra. In Enterbacter sp. B20; it contains the same isocyanovinylphenol addition to determining the spectra of the components of the moiety as rhabduscin and is also glycosylated. It will be reaction solution, the SVD analysis revealed the kinetic interesting to determine whether the correlation between PvcB relationship between them. The species that absorbed at 430 reaction specificity and glycosylation of the pathway product nm formed in a first-order reaction from another species holds as more pvc operons are characterized. characterized by an absorbance peak at 280 nm. In other words, Fe2+, α-ketoglutarate-dependent dioxygenases typically cata- PaPvcB converted tyrosine isonitrile into a relatively unstable lyze hydroxylation of unactivated alkyl C−H bonds or aromatic species with maximal absorbance at 280 nm, which rings. The mechanisms of these reactions are the subject of spontaneously decomposed into a compound with absorbance several excellent reviews.5,30 However, the products of the PvcB maxima at 360 and 430 nm. reactions characterized here are not hydroxylated. Scheme 2

H DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article

Scheme 2 TrEMBL database shows that this aspartic acid is present in more than half the members of the family. The apparently different steady-state kinetic mechanisms exhibited by the three enzymes were unexpected. The Km values for tyrosine isonitrile vary <10-fold and those for α- ketoglutarate <5-fold. However, the kinetic patterns for PaPvcB, XnPvcB, and EaPvcB were those typically seen for ping-pong, sequential, and rapid equilibrium ordered reactions, respectively. Because it is unlikely that these enzymes operate by fundamentally different mechanisms, it is worth exploring how the observed kinetic patterns could arise. Examination of eqs 1−3 reveals that eq 1 is the most general, in the sense that illustrates how the paradigmatic reaction can be applied to it contains all the terms also found in eq 2 or 3. The rapid explain formation of the observed products. Hydroxylation at equilibrium ordered kinetic pattern described by eq 3 and β C of tyrosine isonitrile generates an intermediate 4 that is exhibited by EaPvcB is observed when the K B term is absent. ’ ff a common to all PvcB s. The di erent products arise from This occurs when the binding and dissociation of α- different ways that the hydroxyl group is expelled. In the α ketoglutarate are rapid compared to the rest of the catalytic XnPvcB reaction, decarboxylation at C drives the hydroxyl cycle; that is, its binding effectively comes to equilibrium. In group to leave. In the PaPvcB and EaPvcB reaction, the terms of the kinetic mechanism shown in Figure 5, the rapid expulsion of the hydroxyl group is caused by abstraction of a equilibrium ordered pattern is observed when k or k is much proton from the α-carbon. Since no atom from molecular 5 7 smaller than the other rate constants. In fact, k for EaPvcB is oxygen is found in the tyrosine isonitrile-derived product of the cat lower than that of the other reactions, although it is not known PvcB reaction, strictly speaking, the enzymes are not dioxygenases, but consideration of the underlying reaction whether chemistry or product release is rate-limiting. The mechanism clearly demonstrates their place in the superfamily. sequential pattern observed in the XnPvcB reaction indicates that all the terms in the denominator of eq 1 contribute The arrangement of tyrosine isonitrile shown in Figure 6 fi emerged from docking studies and satisfies regiochemical and signi cantly to the velocity, and the pattern is a family of lines stereochemical constraints. Cβ is in the proximity of the Fe2+ that intersect to the left of the y-axis. The ping-pong pattern atom, making hydroxylation of that position feasible. Further, exhibited by PaPvcB would seem to indicate a kinetic hydroxylation at the pro-S position would result in an mechanism that is distinct from those of the other two 32 intermediate that yields 6. Formation of the E isomer is enzymes, but as has been mentioned previously, seemingly required, so that the carboxyl group is positioned appropriately parallel patterns in double-reciprocal plots can arise in for cyclization to form the coumarin product. That the pro-S sequential mechanisms when the experimental data lack position is hydroxylated is a testable hypothesis, which we sufficient precision to detect small slope effects, which are propose on the basis of the observation that enzyme-catalyzed contributed by the KiaKb term in the denominator of the rate dehydrations that involve abstraction of a proton α to a equation. In this instance, the slope effects will tend to zero carboxylate typically undergo anti elimination.31 when the dissociation constant for α-ketoglutarate is much The sequences of PaPvcB and EaPvcB are 56% identical; smaller than its Km. In an ordered bi-bi sequential mechanism, fi XnPvcB is 42% identical with PaPvcB and 41% identical with the dissociation constant for the rst substrate is k2/k1 and the EaPvcB. Only PaPvcB has been characterized structurally. We Km is (k5k7)/[k1(k5 + k7)]. Thus, when chemistry or product analyzed the docked model of tyrosine isonitrile in the context release is rapid relative to the preceding steps, the term in the of sequence comparison. The residues of PaPvcB that form the velocity equation that introduces the slope effect can become pocket for the aromatic group in the docked model are very small and the double-reciprocal plots will approximate a conserved with the exception of Leu116, which is replaced with family of parallel lines. These interpretations of the steady-state an Arg in XnPvcB. Arg168 of PaPvcB that is partly disordered kinetics data will be evaluated by transient-state kinetic in the structure is positioned near the carboxylate in the docked investigations to pinpoint the rate-limiting portions of the model. This arginine is conserved in EaPvcB and replaced with reaction cycles in the three enzymes. a threonine in XnPvcB. Although we note the presence of The explosion of sequence information has allowed the arginine and lysine before and after this threonine residue in virtual exploration of the metabolic capabilities of myriad XnPvcB, this difference may be partly responsible for the organisms. However, the difficulties presented by misannota- differences in the PvcB reactions catalyzed by these different enzymes. These changes, both involving differences in tion in databases are widely recognized, and verifying the positively charged residues, may be partly responsible for predictions of protein function that are suggested by synteny differences in the enzyme active site that dictate differences in and sequence remains important. In this case, the PvcB proteins 2+ α activity between PaPvcB and XnPvcB. were correctly assigned as Fe , -ketoglutarate-dependent The mechanisms proposed in Scheme 2 require all three oxygenases, but it remained necessary to identify the products enzymes to have an active site general acid to facilitate of the catalytic reactions in the laboratory. The unique expulsion of the hydroxyl group from Cβ in compound 4.We properties of XnPvcB, which cause it to catalyze decarbox- note that residues Ser169−Tyr173 and Val200−Glu211 are ylation of its substrate, were not evident from sequence analysis disordered in the structure of PaPvcB. Neither loop is especially alone and, in fact, remain to be identified; however, as more well-conserved; however, all three proteins carry an aspartic PvcB proteins are experimentally characterized, the sequence acid residue homologous to Asp203 of PaPvcB that could be signatures that are correlated with particular chemical trans- important. A survey of 582 PvcB homologues through the formations should become evident.

I DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article ■ ASSOCIATED CONTENT Involved in the Synthesis of 2-Isocyano-6,7-dihydroxycoumarin. J. Mol. Biol. 384, 193−205. *S Supporting Information − (9) Clarke-Pearson, M. F., and Brady, S. F. (2008) Paerucumarin, a Additional methods and Figures S1 S4. This material is New Metabolite Produced by the pvc Gene Cluster from Pseudomonas available free of charge via the Internet at http://pubs.acs.org. aeruginosa. J. Bacteriol. 190, 6927−6930. Accession Codes (10) Crawford, J. M., Kontnik, R., and Clardy, J. (2010) Regulating Coordinates have been deposited in the Protein Data Bank as Alternative Lifestyles in Entomopathogenic Bacteria. Curr. Biol. 20, − entry 4YLM. 69 74. (11) Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M. fi ■ AUTHOR INFORMATION R., Appel, R. D., and Bairoch, A. (2005) Protein Identi cation and Analysis Tools on the ExPASy Server. In The Proteomics Handbook Corresponding Author (Walker, J. M., Ed.) Humana Press, Totowa, NJ. *E-mail: [email protected]. Fax: 573-882-5635. Tele- (12) Burgess, E. M., Penton, H. R., and Taylor, E. A. (1973) Thermal phone: 573-882-7968. reactions of alkyl N-carbomethoxysulfamate esters. J. Org. Chem. 38, − Funding 26 31. (13) Hemantha, H. P., and Sureshbabu, V. V. (2010) Isoselenocya- This research was supported by a grant from the National nates derived from amino acid esters: An expedient synthesis and Science Foundation (MCB 1158169 to A.M.G. and P.A.T.). application to the assembly of selenoureidopeptidomimetics, unsym- Notes metrical selenoureas and selenohydantoins. J. Pept. Sci. 16, 644−651. The authors declare no competing financial interest. (14) Muramatsu, I., Murakami, M., Yoneda, T., and Hagitani, A. (1965) The formylation of amino acids with acetic formic anhydride. ■ ACKNOWLEDGMENTS Bull. Chem. Soc. Jpn. 38, 244−246. (15) Leggio, A., Liguori, A., Perri, F., Siciliano, C., and Viscomi, M. C. We thank Professor Timothy Glass (Department of Chemistry, (2009) Methylation of α-amino acids and derivatives using University of Missouri) for helpful discussions. trimethylsilyldiazomethane. Chem. Biol. Drug Des. 73, 287−291. (16) Johnson, K. A., Simpson, Z. B., and Blom, T. (2009) Global ■ ABBREVIATIONS Kinetic Explorer: A new computer program for dynamic simulation − BTP, 1,3-bis[tris(hydroxymethyl)methylamino]propane; and fitting of kinetic data. Anal. Biochem. 387,20 29. ′ (17) Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., HEPES, N-(2-hydroxyethyl)-N -(2-ethane)sulfonic acid; Emsley, P., Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G., HPLC, high-performance liquid chromatography; IPTG, and McCoy, A. (2011) Overview of the CCP4 suite and current isopropyl β-D-1-thiogalactopyranoside; LC−MS, liquid chro- − − developments. Acta Crystallogr. D67, 235 242. matography mass spectrometry; MWCO, molecular weight (18) Emsley, P., and Cowtan, K. (2004) Coot: Model-building tools cutoff; NMR, nuclear magnetic resonance; NTA, nitrilotriacetic for molecular graphics. Acta Crystallogr. D60, 2126−2132. acid; SDS−PAGE, sodium dodecyl sulfate−polyacrylamide gel (19) Adams, P. D., Afonine, P. V., Bunkoczi,́ G., Chen, V. B., Davis, I. electrophoresis; SVD, singular-value decomposition; TCEP, W., Echols, N., Headd, J. J., Hung, L.-W., Kapral, G. J., and Grosse- tris(2-carboxyethyl)phosphine; TMS-CHN2, (trimethylsilyl)- Kunstleve, R. W. (2010) PHENIX: A comprehensive Python-based diazomethane. system for macromolecular structure solution. Acta Crystallogr. D66, 213−221. ■ REFERENCES (20) Trott, O., and Olson, A. J. (2010) AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, (1) Micallef, M. L., Sharma, D., Bunn, B. M., Gerwick, L., efficient optimization, and multithreading. J. Comput. Chem. 31, 455− Viswanathan, R., and Moffitt, M. C. (2014) Comparative analysis of 461. hapalindole, ambiguine and welwitindolinone gene clusters and (21) Lang, P. T., Brozell, S. R., Mukherjee, S., Pettersen, E. F., Meng, reconstitution of indole-isonitrile biosynthesis from cyanobacteria. E. C., Thomas, V., Rizzo, R. C., Case, D. A., James, T. L., and Kuntz, I. BMC Microbiol. 14, 213. D. (2009) DOCK 6: Combining techniques to model RNA−small (2) Scheuer, P. J. (1992) Isocyanides and cyanides as natural molecule complexes. RNA 15, 1219−1230. products. Acc. Chem. Res. 25, 433−439. (22) Yang, Z., Lasker, K., Schneidman-Duhovny, D., Webb, B., (3) Brady, S. F., Bauer, J. D., Clarke-Pearson, M. F., and Daniels, R. Huang, C. C., Pettersen, E. F., Goddard, T. D., Meng, E. C., Sali, A., (2007) Natural Products from isnA-Containing Biosynthetic Gene and Ferrin, T. E. (2012) UCSF Chimera, MODELLER, and IMP: An Clusters Recovered from the Genomes of Cultured and Uncultured − Bacteria. J. Am. Chem. Soc. 129, 12102−12103. integrated modeling system. J. Struct. Biol. 179, 269 278. (4) Brady, S. F., and Clardy, J. (2005) Cloning and Heterologous (23) Zhang, Z., Ren, J., Stammers, D. K., Baldwin, J. E., Harlos, K., Expression of Isocyanide Biosynthetic Genes from Environmental and Schofield, C. J. (2000) Structural origins of the selectivity of the DNA. Angew. Chem. 44, 7063−7065. trifunctional oxygenase clavaminic acid synthase. Nat. Struct. Mol. Biol. − (5) Krebs, C., GalonićFujimori, D., Walsh, C. T., and Bollinger, J. M. 7, 127 133. (2007) Non-Heme Fe(IV)−Oxo Intermediates. Acc. Chem. Res. 40, (24) Zhang, Z., Ren, J.-s., Harlos, K., McKinnon, C. H., Clifton, I. J., 484−492. and Schofield, C. J. (2002) Crystal structure of a clavaminate − − − − (6) Stintzi, A., Cornelis, P., Hohnadel, D., Meyer, J.-M., Dean, C., synthase Fe(II) 2-oxoglutarate substrate NO complex: Evidence Poole, K., Kourambas, S., and Krishnapillai, V. (1996) Novel for metal centered rearrangements. FEBS Lett. 517,7−12. pyoverdine biosynthesis gene(s) of Pseudomonas aeruginosa PAO. (25) Clifton, I. J., Doan, L. X., Sleeman, M. C., Topf, M., Suzuki, H., Microbiology 142, 1181−1190. Wilmouth, R. C., and Schofield, C. J. (2003) Crystal structure of (7) Stintzi, A., Johnson, Z., Stonehouse, M., Ochsner, U., Meyer, J.- carbapenem synthase (CarC). J. Biol. Chem. 278, 20843−20850. M., Vasil, M. L., and Poole, K. (1999) The pvc Gene Cluster of (26) Qaisar, U., Luo, L., Haley, C. L., Brady, S. F., Carty, N. L., Pseudomonas aeruginosa: Role in Synthesis of the Pyoverdine Colmer-Hamood, J. A., and Hamood, A. N. (2013) The pvc Operon Chromophore and Regulation by PtxR and PvdS. J. Bacteriol. 181, Regulates the Expression of the Pseudomonas aeruginosa Fimbrial 4118−4124. Chaperone/Usher Pathway (Cup) Genes. PLoS One 8, e62735. (8) Drake, E. J., and Gulick, A. M. (2008) Three-dimensional (27) Crawford, J. M., Portmann, C., Zhang, X., Roeffaers, M. B. J., Structures of Pseudomonas aeruginosa PvcA and PvcB, Two Proteins and Clardy, J. (2012) Small molecule perimeter defense in

J DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Biochemistry Article entomopathogenic bacteria. Proc. Natl. Acad. Sci. U.S.A. 109, 10821− 10826. (28) Myllyla,̈ R., Majamaa, K., Günzler, V., Hanauske-Abel, H. M., and Kivirikko, K. I. (1984) Ascorbate is consumed stoichiometrically in the uncoupled reactions catalyzed by prolyl 4-hydroxylase and lysyl hydroxylase. J. Biol. Chem. 259, 5403−5405. (29) Saari, R. E., and Hausinger, R. P. (1998) Ascorbic Acid- Dependent Turnover and Reactivation of 2,4-Dichlorophenoxyacetic Acid/α-Ketoglutarate Dioxygenase Using Thiophenoxyacetic Acid. Biochemistry 37, 3035−3042. (30) Simmons, J. M., Müller, T. A., and Hausinger, R. P. (2008) Fe II/α-ketoglutarate hydroxylases involved in nucleobase, nucleoside, nucleotide, and chromatin metabolism. Dalton Trans., 5132−5142. (31) Schwab, J. M., Klassen, J. B., and Habib, A. (1986) Stereochemical course of the hydration reaction catalysed by β- hydroxydecanoylthioester dehydrase. J. Chem. Soc., Chem. Commun., 357−378. (32) Ning, J., Purich, D. L., and Fromm, H. J. (1969) Studies on the Kinetic Mechanism and Allosteric Nature of Bovine Brain Hexokinase. J. Biol. Chem. 244, 3840−3846.

K DOI: 10.1021/acs.biochem.5b00255 Biochemistry XXXX, XXX, XXX−XXX Supplementary Information for:

Examining Reaction Specificity in PvcB, a Source of Diversity in Isonitrile-containing

Natural Products

Jing Zhu, Geoffrey M. Lippa, Andrew M. Gulick, and Peter A. Tipton

1

Synthesis of tyrosine isonitrile The route for synthesis of tyrosine isonitrile is given in Scheme S1. 1H NMR spectrum of the product is shown in Figure S1.

Liquid chromatography-mass spectrometry 10 l of the samples were injected into an integrated Thermo-Finnigan LC system equipped with a 150 x 4.6 mm C18 reverse phase column (Sigma-Aldrich). An LC method with a mobile phase of H 2O supplemented with 0.1% formic acid (Solvent A) and MeOH (Solvent B) was used: t = 0 min, 5%B; t = 5 min, 5%B; t = 30 min, 30%B; t = 35 min, 30%B; t = 45 min, 80%; t = 50 min, 80%B; t = 55 min, 5%B; t = 60 min, 5%B. HPLC separation of different components in the sample was followed by the mass analysis using a Finnigan TSQ7000 triple-quadrupole mass spectrometer equipped with an Atmospheric Pressure Chemical Ionization (APCI) source operating in positive ion mode.

2 Figure S1

c

d

a b

d

c

b

a

3 SVD Analysis of PaPvcB product decomposition

Tyrosine isonitrile (0.2 mM), α-ketoglutarate (2 mM), ascorbate (50 M), and (NH 4)2Fe(SO 4)2 (10 M) were combined in 50 mM HEPES, pH 7.8, 150 mM NaCl in a total volume of 1.0 mL. Reaction was initiated by the addition of 4 M PaPvcB. Spectra were collected every 0.1 min for 30 min. Spectroscopic data were filtered so that only those collected 30 s apart were imported into KinTek Explorer v. 4.0 1, where singular value decomposition was performed. The amplitudes of the first five eigenvectors derived from SVD are plotted in Figure S2. Based on visual inspection of these data, all but the first two eigenvectors were ignored for subsequent data analysis, and it was determined that 2 species were sufficient to account for the spectra. The time evolution of the SVD spectral components and amplitudes were fitted to a model in which one species was converted irreversibly to another in a first order process.

In this model P 1 is the true product of the PaPvcB reaction, and P 2 is the species to which it -1 converts nonenzymatically. The value of k 1 was determined to be 0.12 ± 0.05 min . The time evolution of the first two eigenvectors were fitted to the model (Figure S3). Spectra reconstructed from the SVD spectra, the two eigenvectors, and the calculated value of k 1 are compared with the experimental data in Figure S4.

4

Figure S2: Amplitudes of the first 5 eigenvectors derived from singular value decomposition of spectra obtained during decomposition of the PaPvcB reaction product.

5

Figure S3: Time dependence of the first two eigenvectors derived from SVD analysis. Experimental values shown as circles (
) and fitted curves shown as solid lines.

6

Figure S4: Comparison between selected experimental spectra and reconstructed spectra from SVD analysis of the PaPvcB reaction. Experimental conditions are described in the manuscript; black solid lines are experimental, red dashed lines are reconstructed. Spectra were acquired at the times indicated on the figure.

References

[1] Johnson, K. A., Simpson, Z. B., and Blom, T. (2009) Global Kinetic Explorer: A new computer program for dynamic simulation and fitting of kinetic data, Analytical Biochemistry 387 , 20-29.

7 2.2 Summary of Published Work

In the above attached article, we present a study of the PvcB protein, which is predicted to be an Fe2+, α-ketoglutarate dependent oxygenase and to be responsible for the chemical diversity of the isonitrile metabolites produced in different pathways. In this investigation, the homologous pvcB genes from the genome of the human pathogen Pseudomonas aeruginosa, the insect pathogen

Xenorhabdus nematophila and the plant pathogen Erwinia amylovora were cloned into the appropriate expression vectors by Eric Drake. Recombinant pvcB genes were heterologously expressed in E. coli and the protein products were purified to homogeneity with metal ion affinity chromatography. The putative substrate

(tyrosine isonitrile) of the PvcB enzyme is not commercially available, so it was synthesized. The in vitro activities of three PvcB enzymes were examined in the presence of the primary substrate tyrosine isonitrile, and the co-substrates α- ketoglutarate and Fe2+ under aerobic conditions, and were characterized by UV- visible spectroscopy. UV-vis spectra of the PaPvcB reaction revealed there was more than one product present in this reaction, thus singular value decomposition analysis was performed by Dr. Peter Tipton to determine the components in the reaction solution, the kinetic relationship between them and the extinction coefficient of each component. Product formed in each of the PvcB reactions was determined by HPLC, 1H-NMR spectroscopy and LC-mass spectrometry

(performed by Dr. Leigh Nathan). Steady-state kinetics were performed to

28 determine the parameters Km and kcat of the PvcB reactions. In addition, a higher resolution structure of the PaPvcB protein was obtained and was used for molecular docking by Dr. Geoffrey Lippa.

To conclude, the work dedicated to synthesize the putative substrate tyrosine isonitrile, to determine the in vitro activities of three PvcB homologous proteins and to identify the products of the reactions demonstrated PvcB is an Fe2+, α- ketoglutarate dependent oxygenase; PvcB proteins encoded by different operons carry out pathway-specific oxidations, resulting in the chemical diversity of the isonitrile-containing natural products; and the structural differences between the

PvcB enzymes may partially account for the various catalytic activities observed.

Apart from the studies described above, additional experiments were conducted to investigate the mechanisms underlying the PvcB reactions. Two analogs of tyrosine isonitrile were synthesized and examined in the PvcB activity assays, in order to determine the substrate tolerance of the PvcB catalysts and to search for potential enzyme inhibitors. Transient-state kinetics, including rapid- mixing quench flow and stopped-flow studies were performed to investigate the reactions on a short timescale, so as to detect and identify the intermediate(s) generated in the PvcB reactions.

29

2.3 Supplemental Results

2.3.1 Heterologous Expression of PaPvcB, XnPvcB and EaPvcB in E. coli

Recombinant pvcB genes were expressed in E. coli cells under 18 °C. Higher temperature resulted in the loss of soluble proteins. Robust expression of PaPvcB or XnPvcB was observed in the cell lysate, while EaPvcB was obtained at low quantity (Figure 2.1).

PvcB proteins were co-expressed with a His6-tag, which facilitated their purification. Purification using metal ion affinity chromatography yielded PvcB proteins that were >95% pure, based on Coomassie-stained SDS-PAGE. Samples supplemented with 10% (v/v) glycerol and stored at -80 °C were stable for several months.

The EaPvcB reaction was not further explored in the following sections due to the low-yield problem of this enzyme. But studying the PaPvcB reaction in more detail will help us gain insight into the EaPvcB reaction, based on the similar activities exhibited by these two proteins.

30

1 2 3 4 5 6 7

kDa

55

35 PaPvcB EaPvcB XnPvcB

Figure 2.1 SDS-PAGE analysis of the BL21(DE3) cell lysates harboring PaPvcB, EaPvcB, and

TM XnPvcB. Lane 1 shows the PageRuler Plus Prestained protein ladder. Lane 2, 4 and 6 are the lysates from the non-IPTG induced cultures. Lane 3, 5 and 7 come from the IPTG-induced cultures expressing PaPvcB, EaPvcB and XnPvcB, respectively. The sizes of the proteins PaPvcB: ~33 kDa, EaPvcB: ~33 kDa, XnPvcB: ~35 kDa are estimated using ExPASy’s “Protparam” tool.

31

2.3.2 Rapid-Mixing Quench Flow Analysis of the PvcB Reactions

As stated earlier, the XnPvcB enzyme catalyzes an oxidative decarboxylation reaction on the substrate, while the PaPvcB reaction only introduces a double bond into tyrosine isonitrile and retains the carboxyl group. A reasonable mechanism was proposed to explain the product formation in two reactions, in which a common intermediate, hydroxylated tyrosine isonitrile, was generated first, and then the hydroxyl group was expelled in different ways, resulting in the products observed.

In order to demonstrate this mechanism, we tried to capture the hydroxylated intermediate by rapid-mixing chemical quench flow and to identify it by HPLC and mass spectrometry. Both the XnPvcB and PaPvcB reactions were carried out under single turnover conditions within short periods of time. HPLC analyses of the

PaPvcB reactions quenched with 1 M HCl at different lengths of time (2 ms, 10 ms,

30 ms, 50 ms, 100 ms, 200 ms and 500 ms) revealed a compound, which eluted at 12.5 min, started to show up in the 10 ms reaction, accumulated and then disappeared from the 50 ms reaction (Figure 2.2 A), indicating it might be an intermediate made in the PaPvcB reaction. The peak with a retention time of 15 min was also generated in this reaction, and was preserved in relatively long-time reactions, suggesting it might be the product. The XnPvcB reaction, however, only generated a species that was eluted at 19 min which was assigned to the product.

No other new peaks were observed in HPLC spectra of the XnPvcB reactions quenched at different times (Figure 2.2 B). It should be noted that by introducing

32

HCl into the reactions, the substrate (tyrosine isonitrile) that remained in the reaction and the corresponding intermediate(s) or product(s) accumulated in the solution were converted to their formyl counterparts.

Areas of the two peaks of interest (potential intermediate and product peaks) shown in the HPLC spectra of the PaPvcB reactions represented the relative amounts of two species, and were plotted as a function of the reaction time (Figure

2.3 A). A plot displaying the product formation in the XnPvcB reaction was generated in the same way (Figure 2.3 B). Plots of both reactions showed a decrease in the amounts of the products in the longer-time reactions, suggesting the products were decomposed gradually after they were formed, probably due to the long duration (1 hour) of each analysis by HPLC and the labile environment in water.

33

A

intermediate product

No enzyme

2 ms 10 ms

30 ms 50 ms

100 ms 200 ms

500 ms

B

product No enzyme

2 ms

10 ms

30 ms 50 ms 100 ms

200 ms

Figure 2.2 HPLC spectra of the (A) PaPvcB reaction and (B) XnPvcB reaction quenched by HCl at different times. X-axis shows the retention time and Y-axis shows the relative UV absorbance at

310 nm. Intermediate and product peaks are indicated by arrows.

34

A

B

Figure 2.3 Intermediate and product formation in (A) PaPvcB and (B) XnPvcB-catalyzed reactions with tyrosine isonitrile as the substrate. X-axis shows the reaction time and Y-axis shows the relative amounts of the species in the quenched reactions based on the HPLC analyses. The curves are drawn to aid the eye.

35

2.3.3 Stopped-Flow UV-visible Spectroscopic Analysis of the PvcB Reactions

An Fe(IV)-oxo intermediate which is highly reactive and essential to the oxidation of the substrate was trapped and characterized in the reaction catalyzed by taurine dioxygenase (TauD), the most representative member in the Fe2+, α-KG dependent oxygenase family. This intermediate, which exhibits an absorption maximum at 318 nm in the TauD reaction, is proposed to be involved in the conserved mechanism used by this family of enzymes [87, 88]. Thus, we attempted to determine the presence of this intermediate in the PvcB-catalyzed reactions by using stopped-flow UV-visible spectroscopy. In addition to the Fe(IV)- oxo intermediate, attempts to detect the αKG-Fe(II)-substrate complex (λmax: 520 nm in the TauD reaction [89]) were made. PaPvcB and XnPvcB reactions were performed under single turnover reactions and the time course of the reaction was monitored by absorbance measurements at 520 and 345 nm for detecting the species tyrosine isonitrile-αKG-Fe(II)PvcB and the Fe(IV)-oxo intermediate, respectively.

UV-vis spectra of the PaPvcB and XnPvcB reactions examined within 0.5 s are shown in Figure 2.4. In either case, an increase followed by a decrease of the absorbance at 345 nm was observed in 0.1 s, revealing the formation and decomposition of the Fe(IV)-oxo species in one catalytic cycle. On the contrary, the tyrosine isonitrile-αKG-Fe(II)PvcB complex exhibited a backward pattern, indicated by the decrease of the absorbance at 520 nm first. This result was

36 consistent with the proposed order of the steps taken place during the reaction: addition of O2 into the iron center converts the tyrosine isonitrile-αKG-Fe(II)PvcB species to Fe(III) and Fe(IV)-oxo intermediates. Subsequent substrate hydroxylation and product releasing restore the initial complex (described in

Section 2.4).

37

A B

Figure 2.4 UV-vis spectra of the (A) PaPvcB reaction and (B) XnPvcB reaction monitored at 340 nm (blue dots) and 520 nm (red dots). Spectra were obtained after mixing 100 μM α-ketoglutarate,

200 μM tyrosine isonitrile in 50 mM HEPES, pH 7.8, 150 mM NaCl with an equal volume of buffer

containing 100 μM PvcB, 100 μM (NH4)2Fe(SO4)2 and 100 μM ascorbic acid. Random data points obtained at the beginning might be due to some artificial effects.

38

2.3.4 Organic Synthesis and 1H-NMR Characterization of 4-Nitro

Phenylalanine Isonitrile and 3-Nitro Tyrosine Isonitrile

4-Nitro phenylalanine isonitrile and 3-nitro tyrosine isonitrile, two analogs of tyrosine isonitrile were synthesized following Scheme 2.1. Products synthesized from each step were all characterized and confirmed by TLC, HPLC and 1H-NMR spectroscopy. 1H-NMR spectra of these two compounds are shown in Figure 2.6.

They were tested in the PvcB reactions directly without further purification.

39

A Synthesis of 4-nitro phenylalanine isonitrile

O OCH3 O OCH O OCH O OH 3 3 Imidazole Burgess reagent LiOH H2N OHCHN N N C C DMF CH2Cl2 THF/H2O

NO2 NO2 NO2 NO2

B Synthesis of 3-nitro tyrosine isonitrile

O OH O OH O OCH3 O OCH3 Formic acetic anhydride TMS-diazomethane Burgess reagent H2N OHCHN OHCHN N Formic acid CH2Cl2 CH2Cl2 C

NO2 NO 2 NO2 NO2 OH OH OH OH

LiOH

O OH

N C

NO2 OH

Scheme 2.1 Organic synthesis of (A) 4-nitro phenylalanine isonitrile and (B) 3-nitro tyrosine isonitrile.

40

A O OH c N d C b b a b a a

NO2 d

B O OH d N e C b a a c c NO2 b e OH

Figure 2.5 1H-NMR spectrum of (A) 4-nitro phenylalanine isonitrile and (B) 3-nitro tyrosine isonitrile.

The protons in the structures and the corresponding peaks are indicated as a, b, c, d and e. Proton

peak of c in (A) and of d in (B) were covered by the H2O peak which was at 4.69 ppm.

41

2.3.5 PvcB Activity Assays with 4-Nitro Phenylalanine Isonitrile and 3-Nitro

Tyrosine Isonitrile

The PvcB reactions were performed with alternative substrates in order to determine the substrate tolerance of the enzymes. The assays were carried out with the primary substrate and the co-substrates α-ketoglutarate and Fe2+ under aerobic conditions. PvcB enzymes were added to initiate the reactions and the reaction solutions were monitored spectrophotometrically at 240-500 nm for about

10 minutes. When tyrosine isonitrile was replaced with 4-nitro phenylalanine isonitrile in either PaPvcB or XnPvcB reaction, the reaction failed to exhibit any absorbance change, indicating 4-nitro phenylalanine isonitrile could not serve as the substrate for the PvcB enzymes (Figure 2.6). However, incubating the PvcB protein with another compound 3-nitro tyrosine isonitrile, which retains the para- hydroxy group of tyrosine isonitrile, showed spectroscopic changes in both

PaPvcB and XnPvcB reactions. Similar to the reactions conducted with tyrosine isonitrile, the PaPvcB reaction using 3-nitro tyrosine isonitrile as the substrate was characterized by yielding a species with maximal absorbance at 280 nm first, which was then converted to another species absorbing at longer wavelength (420 nm).

In the XnPvcB reaction with 3-nitro tyrosine isonitrile, a species with maximal absorption at 310 nm was generated very rapidly. Control reactions replacing 3- nitro tyrosine isonitrile with 3-nitro tyrosine could not support the PvcB activity

(Figure 2.7).

42

A BB

C D D

Figure 2.6 Spectroscopic analyses of the PvcB reactions with (A, C) 4-nitro phenylalanine isonitrile and (B, D) tyrosine isonitrile. Each reaction contained 200 μM primary substrate, 100 μM α-

ketoglutarate, 10 μM (NH4)2Fe(SO4)2, 150 mM NaCl and 50 mM HEPES, pH 7.8. The spectrum of the reaction with no enzyme is shown in red, and of the reaction right after enzyme addition is shown in black. Subsequent spectra were acquired every 1 (or 0.5) minute up to 10 (or 5) minutes in the reactions containing (A, B) 0.35 μM PaPvcB and (C, D) 0.21 μM XnPvcB.

43

A B

C D

Figure 2.7 Spectroscopic changes during PvcB-catalyzed reactions with (A, C) 3-nitro tyrosine isonitrile and (B, D) 3-nitro tyrosine. Each reaction contained 200 μM primary substrate, 100 μM α-

ketoglutarate, 10 μM (NH4)2Fe(SO4)2, 150 mM NaCl and 50 mM HEPES, pH 7.8. The spectrum of the reaction with no enzyme is shown in red, and of the reaction right after enzyme addition is shown in black. Subsequent spectra were acquired every 1 (or 0.5) minute up to 10 (or 5) minutes in the reactions containing (A, B) 0.35 μM PaPvcB and (C, D) 0.21 μM XnPvcB.

44

2.4 Conclusion and Discussion

The PvcB protein belongs to the non-heme Fe2+, α-ketoglutarate dependent oxygenase family. Enzymes in this family catalyze a wide range of oxidative

2+ transformations [72, 75, 84], which always involve Fe , α-ketoglutarate and O2.

There has been much research effort directed towards determining the mechanism of these reactions, even though it’s difficult because the active sites of the enzymes do not exhibit intense spectral features [90]. Activity of this family of enzymes always requires Fe2+, which is weakly bound by three amino acid side chains that typically occur in a (His)2-(Asp/Glu) motif and are highly conserved within the family members [91, 92]. α-Ketoglutarate chelates the iron center and further provides stability by interacting with other side chains. Presence of the primary substrate facilitates oxygen addition to the iron center, followed by generation of an Fe(III)- superoxo species and a high-valent Fe(IV)-oxo intermediate. At the meantime, decomposition of α-KG forms CO2 and succinate. Finally, the high-valent Fe(IV)- oxo species abstracts a hydrogen atom from the primary substrate and forms a substrate radical plus an Fe(III)-hydroxide complex. Subsequent recombination results in hydroxylated substrate and restores the Fe(II) form of the enzyme

(Scheme 2.2) [72, 89, 93, 94].

45

Scheme 2.2 General mechanism of the reactions catalyzed by the Fe2+, α-ketoglutarate dependent oxygenases.

One famous member of this family is taurine dioxygenase (TauD) which catalyzes the hydroxylation of taurine and yields sulfite and aminoacetaldehyde.

Its mechanism has been studied in detail [89, 94], becoming a model system for the whole family. Two iron based intermediates were detected in the mechanistic study of the TauD reaction by stopped-flow absorption and freeze-quench

Mössbauer spectroscopy. One of them is a highly reactive Fe(IV)-oxo complex, which has an absorption maximum at 318 nm and exhibits a large substrate deuterium kinetic isotope effect (2H-KIE) on its decay, indicating it is the species that abstracts the H atom from the primary substrate, resulting in substrate

46 hydroxylation in most cases. An intermediate with the identical spectroscopic features was trapped in the reaction catalyzed by a prolyl-4-hydroxylase, which hydroxylates proline residues, suggesting a conserved mechanism utilized by these enzymes for substrate hydroxylation [95].

To study the pre-steady state of PvcB-catalyzed reactions, we tried first to detect this high valent Fe(IV)-oxo intermediate in the PvcB reactions by stopped- flow UV-visible spectroscopy. Reactions were carried out under single turnover conditions for better detection of short-lived intermediates. The Fe(IV)-oxo species was detected at 345 nm instead of 318 nm (maximal absorbance of the intermediate in the TauD reaction) to prevent the interference by product formation, which exhibited increased absorbance at 310 nm in the XnPvcB and PaPvcB reactions. The UV-vis spectra of both reactions monitored for 0.5 s revealed an absorbance increase at 345 nm, representing the formation of one intermediate during the reaction. The decay of this intermediate was observed as the subsequent absorbance decrease at this wavelength. However, the absorbance was not returned to the basal level, probably due to the accumulation of the products in the reactions.

The PvcB reactions were also monitored at 520 nm to detect the decay and formation of the substrate-αKG-Fe(II)enzyme complex. The UV-vis spectrum showed a decrease followed by an increase in absorbance at this wavelength, indicating this complex was converted to other species, and after one catalytic

47 cycle, was formed again. In fact, this transformation was stimulated by addition of

O2 into the active site. But within this investigation, all the reactions were conducted in fresh buffers (O2 rich); concentration of O2 in the reactions was not controlled.

O2 plays an important role in initiating the enzymatic turnovers, and some turnovers of the Fe2+, α-KG dependent enzymes were observed in the absence of the primary substrate, resulting in self-hydroxylation and enzyme inactivation [96, 97].

Regarding this, PvcB reaction conditions could be further optimized by controlling the O2 concentrations, especially in transient-state kinetic studies. This may be fulfilled by creating an anaerobic environment first, and carrying out reactions in buffers saturated with O2. Effects of the O2 concentration on the intermediate and product formation in the PvcB reactions could be also explored in this way.

The XnPvcB and PaPvcB reactions were proposed to generate different products through a common intermediate, hydroxylated tyrosine isonitrile. To demonstrate this mechanism, we attempted to capture and identify the intermediate in PvcB reactions by rapid-mixing chemical quench flow coupled with

HPLC and mass spectrometry. HPLC analyses of the reactions quenched at different times revealed the formation and decay of one species in the PaPvcB reaction. This species had an absorption maximum at around 300 nm (not 280 nm or longer wavelength), and was assigned to the intermediate. The identity of this intermediate should be further confirmed by mass spectrometry. However, the same intermediate was not observed in the XnPvcB reaction. Whether the

48 hydroxylated intermediate is not accumulated in the reaction or the XnPvcB enzyme converts the substrate in a different way needs to be further examined.

Other than tyrosine isonitrile, alternative isonitriles which are analogous to tyrosine isonitrile were synthesized and examined in the PaPvcB and XnPvcB reactions as the substrates. Organic syntheses of m-fluoro tyrosine isonitrile, 3,4- dihydroxyl phenylalanine isonitrile, 4-nitro phenylalanine isonitrile and 3-nitro tyrosine isonitrile were all tried, but only the latter two were synthesized successfully. When the PvcB enzymes were incubated with 4-nitro phenylalanine isonitrile and 3-nitro tyrosine isonitrile separately in the presence of the co-

2+ substrates, Fe and O2, turnovers were only observed in the reaction with the latter compound. And the UV-vis spectra of the reactions with 3-nitro tyrosine were similar to those with tyrosine isonitrile, indicating 3-nitro tyrosine served as a substrate in the PvcB-catalyzed reactions, while 4-nitro phenylalanine could not.

Comparing the structures of the unsuccessful substrate (4-nitro phenylalanine isonitrile) and the successful ones (3-nitro tyrosine isonitrile and tyrosine isonitrile) found the difference in a hydroxyl group on the benzene ring. It seemed the absence of this hydroxyl group in 4-nitro phenylalanine isonitrile resulted in the failure of transformation by PvcB. The importance of the hydroxyl group on the C4 position of the benzene ring should be further determined by testing more analogs that contain hydroxy group at other positions or have no hydroxyl group in the PvcB reactions.

49

Attempt to synthesize the m-fluoro tyrosine isonitrile failed, but the successful conversion of m-fluoro tyrosine into the corresponding XnPvcB product in the

XnPvcA-XnPvcB coupled reactions (described in Chapter 3) suggested this compound as the substrate for the XnPvcA protein and its isonitrile product as the substrate for the XnPvcB enzyme. It is also interesting to examine 3,4-dihydroxyl phenylalanine isonitrile in the PvcB reactions since the isonitrile metabolite paerucumarin (2-isocyano-6,7-dihydroxy coumarin) produced by P. aeruginosa contains this backbone. The PvcC protein encoded by the same operon in P. aeruginosa is proposed to catalyze the hydroxylation of the C3 position of the tyrosine ring [27, 40]. If 3,4-dihydroxyl phenylalanine isonitrile could serve as the substrate in the PaPvcB reaction, ring closure of the corresponding product will result in paerucumarin formation. This will provide an alternative route for the synthesis of paerucumarin.

It is difficult to synthesize tyrosine isonitrile analogs due to the instability of these compounds when having an isonitrile group close to a carboxyl group. More time could be invested in looking for better methods for the synthesis of tyrosine isonitrile analogs. By synthesizing and examining various substrate analogs, we can gain a deeper understanding of the structure-related chemical mechanism of the PvcB reaction. In addition, we have a chance to find potential enzyme inhibitors, which will play important roles to modulate the biosynthesis of the isonitrile- containing natural products in bacteria.

50

2.5 Material and Methods

2.5.1 SDS-PAGE Analysis of PvcB Expression in E. coli

Recombinant pvcB genes were expressed in E. coli as described earlier. To check protein expression in the cells, aliquots were removed from the growing culture (expressing PaPvcB, XnPvcB or EaPvcB) right before adding 1 mM IPTG and 16 hours after IPTG induction. Aliquots were incubated with 2 x SDS-PAGE loading buffer at 70 °C for 30 minutes with vortex for several times. After centrifugation at 21380 x g for 20 minutes, supernatants were loaded onto SDS-

PAGE for analysis.

2.5.2 Rapid Quench Flow Analysis of the PvcB Reactions

Rapid-mixing chemical quench experiments were conducted using a Kintek

RQF-3 instrument under single turnover conditions. The first reactant syringe contained 800 μΜ tyrosine isonitrile and 400 μΜ α-ketoglutarate. The second reactant syringe contained 500 μΜ PaPvcB or XnPvcB, 500 μΜ (NH4)2Fe(SO4)2 and 500 μΜ ascorbate. Solutions in both syringes were buffered with 50 mM

HEPES pH 7.8, 150 mM NaCl. Equal volumes of solutions from two syringes were mixed in the chemical-quench apparatus thermostated at 25 °C, and the reactions were terminated at fixed times ranging from 2 to 500 ms by the addition of 1 M HCl.

The pooled samples were lyophilized, resuspended in 50 μL of H2O and analyzed by HPLC to detect the presence of the intermediate and product. These samples were also prepared for MS analysis in order to identify the intermediate of the

51

PaPvcB and XnPvcB reaction.

2.5.3 Stopped-Flow UV-visible Analysis of the PvcB Reactions

Pre-steady-state kinetic studies of the PaPvcB and XnPvcB reactions were performed using a KinTek SF-2000 spectrophotometer equipped with a 2.6 cm path length observation cell and a photodiode for single wavelength absorbance detection. The instrument dead-time was determined to be 4 ms [98]. For each reaction 500 data points were collected within 500 ms; 6-10 traces were averaged to obtain time courses that were used for fitting. Single-turnover reactions were conducted by mixing 100 μM tyrosine isonitrile and 200 μM α-ketoglutarate in 50 mM HEPES, pH 7.8, 150 mM NaCl with 100 μM XnPvcB, 100 μM (NH4)2Fe(SO4)2 and 100 μM ascorbate in the same buffer. The time course of the reaction was monitored by absorbance measurements at 520 and 345 nm for detecting the species tyrosine isonitrile-αKG-Fe(II)PvcB and an Fe(IV)-oxo intermediate, respectively.

2.5.4 Synthesis of 4-Nitro Phenylalanine Isonitrile

4-Nitro phenylalanine methyl ester hydrochloride (1 mmol, 0.26 g) was combined with imidazole (2 mmol, 0.14 g) in 1 mL dimethylformamide (DMF). The reaction was stirred at 60 °C for 48 hours. DMF was removed from the reaction by partition into 5 mL ethyl acetate/5 mL 20% sodium chloride. The organic phase was washed with 5 mL brine 5 times in order to get rid of DMF, and then was dried

1 by rotary evaporation. Yield: 55%; H-NMR (d6-DMSO) δ, ppm 8.15 (d, J=9.0, 2H),

52

7.97 (s, 1H), 7.50 (d, J=9.0, 2H), 4.66 (m, 1H), 3.64 (s, 3H), 3.21 (dd, J=15.0, 6.0,

1H), 3.05 (dd, J=15.0, 9.0, 1H).

4-Nitro formyl phenylalanine methyl ester (~0.5 mmol, 0.13 g) was dissolved in dry dichloromethane (5 mL). Burgess reagent ((methoxycarbonylsulfamoyl) triethylammonium hydroxide, inner salt) (0.5 mmol, 0.12 g) was added and the solution was stirred at 50 °C for 2 hours. The product was purified by column chromatography using 30 g silica gel equilibrated in CHCl3. After loading the sample, the column was washed with 100 mL CHCl3, then 100 mL CHCl3: MeOH

(98:2). 4-Nitro formyl phenylalanine methyl ester eluted when the column was

1 washed with 100 mL CHCl3: MeOH (97:3). Yield: 62%; H-NMR (CDCl3) δ, ppm

8.24 (d, J=9.0, 2H), 7.47 (d, J=9.0, 2H), 4.55 (m, 1H), 3.84 (s, 3H), 3.40 (dd, J=13.5,

5.0, 1H), 3.30 (dd, J=13.5, 7.5, 1H).

4-Nitro isonitrile phenylalanine methyl ester obtained from the previous step

(~0.3 mmol) was dissolved in tetrahydrofuran (1.5 mL) and to which 0.5 mL of 2.0

M LiOH was added. The solution was stirred at room temperature for 16 hours to hydrolyze the methyl ester. THF was removed from the reaction by rotary evaporation and the remaining solution was lyophilized to dryness. Yield: 95%; 1H-

NMR (D2O) δ, ppm 8.13 (d, J=9.0, 2H), 7.44 (d, J=9.0, 2H), 3.26-3.14 (m, 2H); FT-

IR (KBr pellet): 2160 cm-1 (isonitrile group).

2.5.5 Synthesis of 3-Nitro Tyrosine Isonitrile

Synthesis of 3-nitro tyrosine isonitrile was similar to that of tyrosine isonitrile.

53

3-Nitro-L-tyrosine (1 mmol, 0.23 g) was dissolved in formic acid (5 mL) and acetic formic anhydride (4 mmol, 0.35 g) was added. The reaction vessel was capped with a drying tube and placed on ice for 2 hours and then incubated with stirring at room temperature overnight. 10 mL water was added to quench the reaction and the solution was concentrated to dryness under reduced pressure. The residue was dissolved in water and the product was collected by filtration. Yield: 81%; 1H-

NMR (d6-DMSO) δ, ppm 7.97 (s, 1H), 7.73 (s, 1H), 7.40 (d, J=9.0, 1H), 7.05 (d,

J=9.0, 1H), 4.50 (m, 1H), 3.06 (dd, J=12.0, 6.0, 1H), 2.85 (dd, J=12.0, 9.0, 1H).

3-Nitro formyl tyrosine (0.8 mmol, 0.21 g) was methylated with the reagent trimethylsilyldiazomethane (2 mmol of 2.0 M solution in diethyl ether) in dry dichloromethane (5 mL). MeOH (6 mmol) was added to the reaction and it was stirred at room temperature for 30 minutes. The solvents were removed by rotary evaporation. Yield: 78%.

The residue (~0.5 mmol 3-nitro formyl tyrosine methyl ester) obtained from the previous step was dissolved in dry dichloromethane (5 mL) and Burgess reagent ((methoxycarbonylsulfamoyl) triethylammonium hydroxide, inner salt) (0.5 mmol, 0.12 g) was added. The solution was heated to reflux for 2 hours. The product was purified by column chromatography using 30 g silica gel equilibrated in CHCl3. After loading the sample, the column was washed with 100 mL CHCl3, then 100 mL CHCl3: MeOH (98:2), followed by 100 mL CHCl3: MeOH (97:3), during which the product eluted. Solvents were removed by rotary evaporation. Yield:

54

1 58%; H-NMR (CDCl3) δ, ppm 8.02 (s, 1H), 7.52 (d, J=9.0, 1H), 7.18 (d, J=9.0,

1H), 4.51 (m, 1H), 3.85 (s, 3H), 3.27 (dd, J=13.5, 4.5, 2H), 3.18 (dd, J=13.5, 7.5,

2H).

To hydrolyze the methyl ester, the material from the silica gel column (~0.3 mmol) was dissolved in 1.5 mL THF. 0.5 mL of 2.0 M LiOH was added and the solution was stirred at room temperature for 16 hours. THF was evaporated under vacuum and the remaining solution was lyophilized to dryness. The residue was

1 not further purified. Yield: 93%; H-NMR (D2O) δ, ppm 7.71 (s, 1H), 7.17 (d, J=6.0,

1H), 6.66 (d, J=9.0, 1H), 4.32 (m, 1H), 3.00-2.85 (m, 2H); FT-IR (KBr pellet): 2158 cm-1 (isonitrile group).

2.5.6 PvcB Activity Assays with 4-Nitro Phenylalanine Isonitrile

Synthesized 4-nitro phenylalanine isonitrile was examined in the reactions catalyzed by PaPvcB and XnPvcB. Enzymatic activity was measured spectrophotometrically under aerobic conditions at 25 °C. Assays were performed in 1 mL quartz cuvettes containing 200 µM 4-nitro phenylalanine isonitrile, 100 µM

α-ketoglutarate, 10 µM (NH4)2Fe(SO4)2, 50 mM HEPES (pH 7.8), 150 mM NaCl, and PaPvcB or XnPvcB. The reactions were initiated by addition of enzyme and were monitored at 240-500 nm. Scans of the reaction mixtures were acquired every 0.5 minute for about 10 minutes at room temperature. Positive control reactions were conducted with tyrosine isonitrile, the confirmed substrate of the

PvcB reactions, under the same condition.

55

2.5.7 PvcB Activity Assays with 3-Nitro Tyrosine Isonitrile

PaPvcB and XnPvcB reactions with 3-nitro tyrosine isonitrile as the substrate were conducted in the same way as described in Section 2.3.6. Negative control reactions were performed with 3-nitro tyrosine, which is an analog of 3-nitro tyrosine isonitrile and is the starting material in the reactions to synthesize 3-nitro tyrosine isonitrile.

2.5.8 HPLC Analyses

HPLC was utilized to monitor every step in the synthesis of tyrosine isonitrile derivatives, as well as the intermediate and product formation in the PvcB reactions. For analysis, 10 µL samples were injected into a Shimadzu LC-10 Ai

Series separations module equipped with a Shimadzu SPD-20A UV-vis detector and a 150 x 4.6 mm C18 reverse phase column (Sigma-Aldrich). The overall system was controlled using LC Solutions software. The UV detector was set at

280 or 310 nm. Compounds were separated by using a gradient method in which a mobile phase of H2O supplemented with 0.1% formic acid (Solvent A) and MeOH

(Solvent B) was used: t = 0 min, 5%B; t = 5 min, 5%B; t = 30 min, 30%B; t = 35 min, 30%B; t = 45 min, 80%; t = 50 min, 80%B; t = 55 min, 5%B; t = 60 min, 5%B.

56

CHAPTER 3

Characterization of the Reaction Catalyzed by PvcA, an Isonitrile

Synthase

3.1 Introduction Bacteria are characterized by their ability to produce and secret secondary metabolites [99, 100]. These structurally diverse molecules are often associated with extensive biological and pharmacological activities [101-103]. A group of natural products containing the isonitrile functionality are widely distributed in biological systems and interest in their biosynthesis and functions has grown in the past three decades [20, 38]. The big breakthrough on identifying the isonitrile biosynthetic gene (e.g. isnA or pvcA) in bacteria was achieved by Brady in 2005

[36]. Early functional studies of the PvcA protein and identification of its homologs in a variety of organisms were discussed in Chapter 1. The pvcA containing clusters always encode another protein, designated IsnB or PvcB. The catalytic activities of the PvcB proteins from three bacterial pathogens were examined and presented in Chapter 2.

Compared to the PvcB reactions, the ones catalyzed by the PvcA enzymes appear to be more mysterious and complicated. The crystal structure of the PvcB protein from Pseudomonas aeruginosa provides strong evidence that it belongs to a family of well-studied enzymes [40], while P. aeruginosa PvcA is a novel protein

57 that only exhibits weak homology to an uncharacterized yeast protein [104]. In addition, PvcB homologs catalyze oxidative reactions which are commonly studied in other areas, but PvcA carries out formation of the isonitrile functionality, probably in a single step by using an amino acid and a phosphorylated sugar as the substrates [37]. This kind of reaction has not been investigated before, let alone the mechanism underlying it, making it the most fascinating and interesting part of this study.

Currently, the in vitro activities of IsnA, AmbI1 and WelI1, the isonitrile synthases derived from an environmental DNA sample, the pathway of Fischerella ambigua and the pathway of Westiella intricate respectively, have been reported

[37, 105, 106] (Scheme 3.1). In all cases, tryptophan builds up the skeleton of the indole product and provides the N atom for the isonitrile group. C2 of ribose-5- phosphate or ribulose-5-phosphate appears to be the source of the isonitrile carbon. The isonitrile synthase and oxidase reactions were always coupled in these studies because the instability of the intermediate generated by the synthase hindered its isolation and detection; in other words, the product of the isonitrile synthase reaction was only inferred but not directly demonstrated.

Pathways for isonitrile-containing natural products in Pseudomonas aeruginosa and Xenorhabdus nematophila have also been proposed [24, 27]

(Scheme 3.2). Instead of using tryptophan as the substrate, these PvcA homologs prefer tyrosine. The heterologous production of the final metabolites paerucumarin

58 and rhabduscin were confirmed in genetically transformed E. coli. However, the activity and function of each individual enzyme, as well as the structure of each intermediate in the pathways has not been demonstrated in vitro. In Chapter 2, we already determined the activities of the PvcB enzymes and validated the reactions they catalyze. So in this chapter, we will focus on determining the in vitro activity of the PvcA protein, and if possible, studying the mechanism of this kind of novel reaction.

59

A H H H N OH OH N N IsnB - IsnA C OPO3 O O OH O Ribulose-5P HO HO N NH2 N C C Tryptophan O OH trans indole isonitrile - CH OPO3 OH OH Ribose-5P

B H H H N OH OH N N AmbI3 - AmbI1 C OPO3 O O OH O C N HO HO NH2 Ribulose-5P N cis indole isonitrile C Tryptophan

C H H H H N N O OH N N WelI3 - WelI1 CH OPO3 O OH OH O C N HO HO N NH2 Ribose-5P N cis indole isonitrile C C Tryptophan trans indole isonitrile

Scheme 3.1 In vitro synthesis of indole isonitrile by (A) purified IsnA/IsnB; (B) purified AmbI1/AmbI3; and (C) E. coli lysates hosting WelI1 and WelI3, respectively. Cis or trans Indole isonitrile was derived from tryptophan and ribose-5P or ribulose-5P.

60

P. aeruginosa P. aeruginosa HO O OH HO O O PvcB PvcCD N HO N C C OH OH O HO O OH OH HO O OH paerucumarin P C O O PvcA NH O OH N 2 C tyrosine ribulose-5-phosphate Glycosyltransferase O O HO PvcB X. nematophila HN OH N C X. nematophila N OH C O rhabduscin

Scheme 3.2 Proposed biosynthesis of paerucumarin and rhabduscin in Pseudomonas aeruginosa and Xenorhabdus nematophila. Tyrosine and ribulose-5P are the putative substrates for the PvcA reactions in these pathways.

61

3.2 Results

3.2.1 Heterologous Expression of PaPvcA and XnPvcA in E. coli

Recombinant pvcA genes from P. aeruginosa, X. nematophila and E. amylovora were overexpressed in E. coli cells grown at 18 °C. Higher temperature resulted in the loss of soluble protein. Expression of PaPvcA and XnPvcA was checked by SDS-PAGE and the results are shown in Figure 3.1. The sizes of the proteins were estimated using ExPASy’s “Protparam” tool (PaPvcA: ~37 kDa,

XnPvcA: ~38 kDa). Soluble EaPvcA was not obtained from the cells. pvcA from another species Aeromonas hydrophila was also subcloned into the appropriate vector for overexpression. Redundancy of the protein targets would allow the experiments to proceed when labile enzyme or low yield problem was found.

Purification using metal ion affinity chromatography yielded proteins that were >95% pure, based on Coomassie-stained SDS-PAGE analysis. Samples supplemented with 10% (v/v) glycerol and stored at -80 °C were stable for several months.

62

1 2 3 4 5 kDa

55

35 XnPvcA PaPvcA

Figure 3.1 Coomassie-stained SDS-PAGE analysis of the E. coli BL21(DE3) cell lysates harboring

XnPvcA and PaPvcA. Lane 1 shows the PageRulerTM Plus Prestained protein ladder. Lane 2 and

3 are IPTG uninduced and induced cell lysates hosting XnPvcA, respectively. Lane 4 and 5 are

IPTG uninduced and induced cell lysates hosting PaPvcA, respectively.

63

3.2.2 Reconstitution of the PvcA Reactions in Vitro

Purified PvcA enzymes (XnPvcA, PaPvcA and AhPvcA) were incubated with the putative substrates tyrosine and ribose-5-phosphate/ribulose-5-phosphate.

However, monitoring the reactions by UV-vis spectroscopy did not reveal any absorbance change in a range of 200-500 nm. In addition, substrate consumption or product formation was not observed in HPLC analyses of the reaction mixtures at different times.

Tyrosine isonitrile, the predicted product of the PvcA reaction, is relatively unstable and is not spectroscopically distinct from tyrosine, one of the expected substrates. Therefore, the PvcB reaction was coupled to the PvcA reaction. The

XnPvcB reaction, which uses tyrosine isonitrile as the substrate and generates a stable and well characterized product [107], was used in these studies. However, the PvcB product isocyanovinylphenol was not made in the coupled reactions with purified enzymes, indicating the PvcA reaction failed to generate tyrosine isonitrile.

Next, experiments were carried out to look for possible cofactor and co- substrate requirements. Some common cofactors, including various metal ions

(Mn2+, Mg2+, Ca2+, K+, Zn2+), thiamine pyrophosphate, pyridoxal phosphate, three forms of tetrahydrofolate (N5-formyl-THF, N5, N10-methenyl-THF, N10-formyl-THF) and ATP were examined in the PvcA-XnPvcB coupled reactions with the proposed substrates tyrosine and ribose-5-phosphate/ribulose-5-phosphate. S-adenosyl methionine, the co-substrate involved in methyl group transfers, was also tried.

64

However, no activity was observed in any of the reactions after inclusion of these molecules.

The purified PvcA protein was also replaced in the coupled reactions with the lysate from cells overexpressing the PvcA protein, in case the enzyme lost its activity during purification. But similarly, the reactions conducted with tyrosine and ribose-5-phosphate/ribulose-5-phosphate were not able to make the product.

3.2.3 Reconstitution of the Isocyanovinylphenol Biosynthetic Pathway in E. coli

The initial attempt to detect the activity of the PvcA protein by incubating the enzyme with the putative substrates tyrosine and ribose-5-phosphate/ribulose-5- phosphate in vitro failed, prompting us to try to confirm its activity in vivo. The metabolite generated by the PvcA reaction was not stable, thus the PvcB reaction was used to convert it to a stable compound.

The activities of three PvcB homologs were described in Chapter 2. The

XnPvcB reaction generates the most stable and well characterized product, isocyanovinylphenol. Therefore, study of the PvcA activity in the pathway of synthesizing isocyanovinylphenol is preferred. In order to reconstitute this biosynthetic pathway in vivo, we constructed a system where the PvcA and

XnPvcB proteins could be co-expressed. This system not only allowed us to obtain large quantities of the isonitrile product isocyanovinylphenol, but also enabled us to perform feeding and labeling experiments to study this pathway in more detail.

65

The dual expression vector AhpvcA-XnpvcB/pETDuet-1 we constructed was transformed into E. coli BL21(DE3) cells for expression at 18 °C. Co-expression of

AhPvcA and XnPvcB was observed and the expression levels of the two proteins were comparable (Figure 3.2). Expression of two genes from one duet vector was more reproducible and stable than that of two genes from two different vectors co- transformed into the cells. Consistent with the SDS-PAGE result, HPLC analysis showed that the metabolite isocyanovinylphenol was isolated from the culture medium successfully (Figure 3.3), further demonstrating the co-expression of the two proteins in the cells. By contrast, the culture expressing no protein, or only

AhPvcA, or only XnPvcB was not able to make the product.

66

1 2 3 4 5

kDa

42

AhPvcA 32 XnPvcB

Figure 3.2 Coomassie-stained SDS-PAGE analysis of the E. coli BL21(DE3) cell lysates harboring

AhPvcA and XnPvcB. Lane 1 shows the PINK stain TM protein ladder; Lane 2-5 show the lysates from cells hosting: no plasmid (Lane 2); AhpvcA/pETDuet-1 (Lane 3); XnpvcB/pACYCDuet-1 (Lane

4) and AhpvcA-XnpvcB /pETDuet-1 (Lane 5). The sizes of the AhPvcA and XnPvcB are 37 kDa and 35 kDa, respectively, based on the ExPASy’s “Protparam” estimation.

67

OH

OH

NC

NHCHO AhPvcA (+), XnPvcB (+)

AhPvcA (-), XnPvcB (+)

AhPvcA (+), XnPvcB (-)

AhPvcA (-), XnPvcB (-)

Figure 3.3 HPLC analyses of the metabolites produced from different cultures. The chromatograph with a UV detector showing that the metabolite isocyanovinylphenol was produced in the culture expressing both AhPvcA and XnPvcB proteins (orange), but not made by the cells that expressed no protein (blue) or only AhPvcA (red) or only XnPvcB (green). The isonitrile group is easily hydrated in the acidic condition so the formyl product is also present.

68

3.2.4 Identification of the Product of the PvcA Reaction

Although the in vitro PvcA reaction was not reconstituted successfully, expressing recombinant PvcA protein in E. coli cells allowed us to detect the product in both the cell lysate and the culture medium. The detection was facilitated by the XnPvcB reaction. Incubation of the XnPvcB enzyme with either the cell lysate or the culture medium from the culture overexpressing XnPvcA resulted in isocyanovinylphenol production, indicating the XnPvcA product was located in both parts. Presence of the PvcA product in the culture medium suggested secretion of this metabolite occurred after it was formed inside the cells.

In order to isolate and identify the product, the cell lysate or the culture medium was fractionated. Dialysis of the cell lysate in 6,000-8,000 MWCO tubing (Fisher) found that the product passed through the membrane into the dialysate, indicating the XnPvcA product was a relatively small molecule. Size-exclusion chromatography allowed us to further separate the compounds by their sizes.

Separated molecules were examined by the XnPvcB reaction and the fractions which supported the XnPvcB reaction were dried and analyzed by 1H-NMR spectroscopy. 1H-NMR spectrum of one of such fractions (Fraction 19) is shown in

Figure 3.4 A. It is compared to the spectrum of authentic tyrosine isonitrile (Figure

3.4 B) we synthesized. All the proton peaks shown in the tyrosine isonitrile spectrum are present in the fraction 19 spectrum, strongly suggesting tyrosine isonitrile is one component of fraction 19. Incubation of this NMR sample with the

69

XnPvcB enzyme and the co-substrates did not affect the peaks around 7.2-7.3 ppm and 3.8 ppm, but the peaks centered at 6.8 ppm and 7.1 ppm were converted to the new peaks which corresponded to the XnPvcB product isocyanovinylphenol

(Figure 3.5) [107]. Thus, we concluded that tyrosine isonitrile was the true product of the XnPvcA reaction. Considering the weak stability of tyrosine isonitrile, we did not further purify the fraction.

70

A

B C N c a b d COOH

a b

c d OH tyrosine isonitrile

Figure 3.4 1H-NMR spectrum of (A) Fraction 19 collected from the size-exclusion column, and (B)

authentic tyrosine isonitrile. Samples were prepared in D2O. Protons and the corresponding peaks in tyrosine isonitrile are assigned in (B) and similar peaks in (A) are indicated by arrows.

71

A

B

c a

b d

Figure 3.5 1H-NMR spectrum of the XnPvcB reaction with (A) fraction 19 and with (B) tyrosine isonitrile. Reaction products were extracted by ethyl acetate and the dried extracts were dissolved

1 in 600 μL methanol-d4 for H-NMR analysis. Protons in the product structure and the corresponding peaks are assigned as a, b, c and d in (B), and similar peaks in (A) are indicated by arrows.

72

3.2.5 Identification of Tyrosine as One of the Substrates for the PvcA

Reaction

The failure to reconstitute the PvcA activity in vitro led us to re-examine the substrates used. To determine whether tyrosine was the right substrate, feeding studies were carried out using AhpvcA/XnpvcB, the biosynthetic pathway for isocyanovinylphenol, expressed in E. coli. The cells containing the plasmid

AhpvcA-XnpvcB/pETDuet-1 were grown in LB supplemented with 3,5-ring-D2 tyrosine. Metabolites isolated from the cultures grown for different lengths of time were analyzed by mass spectrometry. The positive ion mass spectrum showed peaks with m/z values of 146.06 and 148.07, corresponding to the normal product and the labeled product, respectively (Figure 3.6 A). The ratio of the peak intensity

148.07/146.06 in each sample was plotted as a function of the feeding time (Figure

3.6 B). The plot suggested 3,5-ring-D2 tyrosine was incorporated into the product over time and the incorporation was very rapid, with no lag observed following addition of labeled tyrosine.

Similar experiments utilizing m-fluoro tyrosine, an analog of tyrosine were conducted. Metabolites were isolated from the culture medium and analyzed by

HPLC. In the HPLC chromatogram, the normal product and the m-fluoro product eluted at 35.5 min and 36.5 min, respectively (Figure 3.7), allowing us to observe the production of the m-fluoro product over time.

The ability of the PvcA/PvcB enzymes to make the product by using

73 isotopically labeled tyrosine and a tyrosine analog help us to confirm that tyrosine is one of the substrates for the PvcA protein.

74

A

10 minutes

20 minutes

1 hour

2 hours

3 hours

5 hours

B

75

Figure 3.6 Incorporation of 3,5-ring-D2 tyrosine into isocyanovinylphenol in the growing culture that expressed AhPvcA and XnPvcB. (A) Mass spectrum of the products isolated from the culture

medium at different times after addition of 3,5-ring-D2 tyrosine. (B) Plot of the peak intensity ratio of 148.07/146.06 obtained from mass spectrum.

76

OH F

OH

NC

NC

16 hours

1 hour

0 hour

Figure 3.7 HPLC chromatogram of the metabolites isolated from the culture medium before adding m-fluoro-tyrosine (blue); 1 hour after adding m-fluoro-tyrosine (red) and 16 hours after adding m- fluoro-tyrosine (green).

77

3.2.6 Seeking the Substrate Providing the Carbon for the Isonitrile Group

We have confirmed tyrosine to be one of the substrates for the PvcA protein in Section 3.2.5. Next, we tried to identify the source of the isonitrile carbon.

3.2.6.1 Tracking of the Isonitrile Carbon in E. coli Expressing AhPvcA-

XnPvcB

Labeling experiments were carried out in E. coli that expressed both AhPvcA and XnPvcB proteins. The cell cultures were grown in LB supplemented with 13C labeled glucose. The metabolite isocyanovinylphenol was isolated from the culture medium and examined by 13C-NMR spectroscopy. Considering the low natural abundance of 13C compared to 12C, 13C-NMR spectroscopy is very sensitive to detect the presence of the 13C atom in a molecule. 13C-NMR spectrum of

13 isocyanovinylphenol obtained from the culture fed with D-[UL- C6] glucose showed that the 13C was incorporated into the isonitrile carbon, which was centered at 163.9 ppm. Based on this result, we next examined glucoses with single-13C label at different positions. 13C-NMR results revealed isocyanovinylphenol isolated from cultures supplemented with D-[2-13C] and D-[3-

13C] glucoses contained the labeled isonitrile group. Other mono-13C glucoses did not label the product within a certain period of time (Figure 3.8).

To determine the time course of the 13C incorporation, the metabolites were extracted from the cultures which were incubated with D-[2-13C] and D-[3-13C] glucoses for 0 minute, 10 minutes, 20 minutes, 40 minutes and 2 hours after

78 glucose addition, and were analyzed by mass spectrometry. Peaks at 146.06 and

147.06, the m/z values for the normal product and mono-13C labeled product, respectively, were present in the positive ion mass spectrum. The ratio of the peak intensity 147.06/146.06 in each sample was plotted as a function of the feeding time (Figure 3.9). The plot showed rapid incorporation of 13C into the product in the culture supplemented with D-[2-13C] glucose. A little slower incorporation was found in the culture incubated with D-[3-13C] glucose.

79

[1-13C]

[2-13C]

[3-13C]

[4-13C]

[5-13C]

[6-13C]

13 [UL- C6]

Figure 3.8 13C-NMR spectrum of the metabolite isocyanovinylphenol isolated from the culture fed

13 13 13 13 13 13 13 with D-[1- C], D-[2- C], D-[3- C], D-[4- C], D-[5- C], D-[6- C] and D-[UL- C6] glucoses (top to bottom). Isonitrile carbon is indicated by arrow.

80

C2

C3

Figure 3.9 Incorporation of the 13C atom into isocyanovinylphenol isolated from the cultures fed with D-[2-13C] (red dots) and D-[3-13C] (green dots) glucoses, respectively. Ratio of the intensity of

147.06 to 146.06 shown in the mass spectrum is plotted against the feeding time.

81

3.2.6.2 Searching for the Substrate in a Phosphorylated Metabolite Library and the Major Metabolic Pathways

Ribose-5-phosphate and ribulose-5-phosphate are the substrates expected to be the precursor of the isonitrile carbon in the PvcA reaction, and thus were examined first. However, neither of these carbohydrates supported the PvcA reaction (Section 3.2.2). Next, we tried to search broadly for the compound that provides the isonitrile carbon.

We rapidly screened a phosphorylated metabolite library (kindly donated by

Dr. Karen Allen) containing 167 phosphorylated compounds either from commercial source or synthesized, by incubating these compounds individually with tyrosine, α-ketoglutarate, Fe2+ and purified PaPvcA and PaPvcB enzymes.

The PaPvcB reaction was characterized by generating a product with maximal UV absorption at 430 nm [107], thus the reaction mixture turned bright yellow in several minutes. However, no color change was observed in the PaPvcA-PaPvcB coupled reactions with the compounds from the library.

In addition, we searched for the substrate in the central carbon metabolic pathways. Most of the metabolites involved in glycolysis, the pentose phosphate pathway and glycerol metabolism were checked in the PvcA and XnPvcB coupled reactions one by one, or more often, prepared as cocktails and checked together.

However, no activity was detected from the reactions when the enzymes were incubated with tyrosine and these metabolites, indicating none of these

82 compounds were reacting with tyrosine in the presence of the PvcA enzyme.

3.2.6.3 Looking for the Substrate in the Metabolite Pool of E. coli

The failure to find the substrate in the phosphate ester library and the central carbon metabolic pathways prompted us to search in a broader background – the metabolite pool of the growing cells. Metabolites were extracted from E. coli cells during the growth phase with organic solvents and were examined in the PvcA-

XnPvcB coupled reactions, but there was no product formation. Interestingly, when the PvcA enzyme along with its confirmed substrate tyrosine was added to the lysate from cells that expressed the XnPvcB protein, isocyanovinylphenol was produced. Control reaction lacking the PvcA enzyme was not able to make such product (Figure 3.10), indicating the PvcA protein acted on the unknown substrate from the lysate and yielded the product tyrosine isonitrile, which was then converted by XnPvcB to isocyanovinylphenol.

This activity was first detected in the E. coli cell lysate hosting XnPvcB. To eliminate the possibility that the unknown substrate was only induced in the cells that expressed the Pvc proteins, normal cell lysate and lysates hosting other recombinant proteins, including 6-phosphogluconate dehydratase (PGDH), anthranilate-coenzyme A ligase (PqsA) and acyl carrier protein 2 (ACP2) were examined. XnPvcA and XnPvcB enzymes were added to different cell lysates and the product was formed in all cases. HPLC analyses of the reaction extracts showed that there was little product made in the normal BL21(DE3) cell lysate

83 while cell lysate containing the plasmid ACP2/pET14b generated the largest amount of the product (Figure 3.11). The presence of the unknown substrate in various cell lysates indicated it might be a common metabolite or intermediate in the E. coli metabolism.

84

Lysate + tyr-NC

Lysate + XnPvcA

Lysate alone

Figure 3.10 HPLC chromatogram of the product extracted from the reaction containing only the E. coli cell lysate that hosted XnPvcB (blue trace); the cell lysate + XnPvcA enzyme (red trace); and the cell lysate + authentic tyrosine isonitrile (green trace).

85

Relative amount of isocyanovinylphenol of amount Relative

E. coli with plasmids

no plasmid pET15b XnPvcB/pET15b PGDH/pET20b pqsA/pET14b ACP2/pET14b

Figure 3.11 Activities of the lysates from E. coli cells expressing no plasmid (blue), empty pET15b vector (red), XnPvcB (gray), PGDH (yellow), PqsA (pink) and ACP2 (green) in the XnPvcA-XnPvcB coupled reactions. Y-axis shows the relative absorption of the product at 310 nm and is used to represent the relative amount of the product.

86

The potential substrate from the cell lysate was used by the PvcA enzyme in the reaction, thus the amount of this metabolite would decrease to some extent after the reaction and this change might be detected by mass spectrometry. To prepare the samples for MS analysis, the lysate from E. coli cells overexpressing

XnPvcB was divided by two. Half of the lysate was incubated with the XnPvcA enzyme for 3 hours at room temperature and the other half remained untreated.

Macromolecules were then roughly removed from both samples and the remaining solutions were dried. Dried samples were sent to the West Coast Metabolomics

Center (University of California, Davis) for analysis.

The mass peak intensities of various metabolites in two samples were compared. Some metabolites showed more than 50% decrease in regarding to their relative amounts after the PvcA treatment and these metabolites are listed in

Table 3.1. Tyrosine is one of such compounds with its amount decreasing by 57.94% in the PvcA reaction, consistent with the fact that it is one of the substrates of the

PvcA enzyme. Other than tyrosine, glucose, xylose, ribose-5-phosphate and ribose all exhibited huge differences in concentrations before and after the PvcA treatment. Xylose and ribose, in particular, were present in large quantities in the untreated lysate sample, so a 70% decrease after the treatment seemed to be significant. Ribose-5-phosphate is the initially proposed substrate; the other three sugars are close to ribose-5-P in the metabolism and could serve as the precursor of it. Based on the MS results, the effects of glucose, xylose, ribose-5-phosphate

87 and ribose on the lysate + XnPvcA reaction were determined by adding them to the reaction. However, product formation in the reaction was not induced by any of these sugars evidently, indicating none of the compounds was the direct substrate for the PvcA reaction. Three other compounds O-succinylhomoserine, 2- deoxytetronic acid and fumaric acid are also in the list. Considering the substrate structure-related-mechanism of the isonitrile group formation (discussed in Section

3.3), effects of these metabolites on the product formation in the XnPvcA reaction were not determined at this time. Many other mass peaks exhibit decreasing patterns in the PvcA reaction but they were not able to be assigned to the verified metabolites.

88

Table 3.1 Effect of the PvcA reaction on metabolite concentrations in the cell lysate

Sample Lysate no XnPvcA Lysate + XnPvcA Decrease after PvcA Peak intensity

Metabolite treatment (%)

Glucose 2864 530 81.49

Xylose 11099 2737 75.34

Ribose-5-phosphate 1569 424 72.98

Ribose 60680 18043 70.27

O-Succinylhomoserine 3490 1289 63.07

2-Deoxytetronic acid 1025 391 61.85

Tyrosine 251007 105583 57.94

Fumaric acid 113483 50771 55.26

89

The lysate from E. coli cells hosting XnPvcB was fractionated in order to isolate and identify the PvcA substrate. After normal centrifugation, the substrate was located in the supernatant (cell free extract) instead of the pellet (cell debris).

The cell free extract was then centrifuged in an Amicon Ultra centrifugal filter (9K

MWCO) and the solution retained above the membrane, not the portion passing through the membrane showed activity in the XnPvcA-XnPvcB coupled reactions.

The active fraction (solution above the membrane) was then dialyzed in 6,000-

8,000 MWCO tubing to further remove the small molecules. After overnight dialysis, the fraction lost most of the activity in the XnPvcA-XnPvcB coupled reactions. In addition, purification of the cell free extract by size-exclusion chromatography revealed that at least two fractions, one eluting with macromolecules and one eluting with small molecules, must be combined to support the PvcA reaction

(Figure 3.12). These results were combined to suggest that a macromolecule and a small compound were both crucial to PvcA activity.

90

A

I

II III I (fraction 10-17) II (fraction 18-24) III (fraction 36-42) IV IV (fraction 48-53) V (fraction 56-60)

V

B

III + XnPvcA + XnPvcB

I + II + V + XnPvcA + XnPvcB

I + II + IV + XnPvcA + XnPvcB

I + II + III + XnPvcA + XnPvcB

I + II + XnPvcA + XnPvcB

Figure 3.12 Fractionation of the E. coli cell lysate by size-exclusion chromatography. (A) A280 of the fractions collected from the size-exclusion column. (B) HPLC analyses of the product formation in the XnPvcA-XnPvcB coupled reactions with different combinations of the fractions.

91

To further characterize the macromolecule, spin concentrators with different molecular weight cut-offs (9K, 20K, 50K and 100K) were used to fractionate the cell free extract. After centrifugation, the top solution (molecules retained above the membrane) and the bottom solution (molecules passing through the membrane) from each column were checked in the XnPvcA-XnPvcB coupled reactions. The results showed that the top solutions from all the spin columns exhibited activities in the coupled reactions, but only the bottom solution from the 100K spin concentrator supported the XnPvcA reaction. Cation and anion exchange chromatography were utilized to further purify the macromolecule. However, the macromolecule seemed to be lost or inactivated during fractionation on the ion exchange columns. 30% and 70% ammonium sulfate precipitations of the proteins in the cell free extract revealed the presence of the active macromolecule in the

70% pellet. Combining these results, we concluded that the macromolecule essential to PvcA activity might be an enzyme present in the E. coli metabolism and it had a size larger than 50 kDa.

To further isolate the small compound, size-exclusion chromatography followed by anion exchange chromatography were performed to fractionate the bottom solution (small molecules passing through the 9K MWCO membrane) from the spin concentrator. Purification of the small molecules, which were isolated from the size-exclusion column, by anion exchange chromatography showed that the target compound was capable of binding to the column, indicating it was negatively

92 charged. 31P-NMR analysis of the active fraction revealed at least one molecule with a phosphate group was present. Further treatment of this active fraction with alkaline phosphatase resulted in loss of its activity in the PvcA reaction, suggesting the important role of a phosphoryl group in the reaction. This result was consistent with the early prediction of the PvcA substrate to be a phosphorylated carbohydrate.

93

3.3 Conclusion and Discussion

Many isonitrile-containing natural products are produced by bacteria as secondary metabolites and most of them are derived from amino acids. A lot of these compounds have been identified to show antibacterial [108, 109] and antimalarial activities [59, 110], determining their promising importance in therapeutic applications. In order to gain a deeper understanding of and have easier access to these bioactive metabolites, it is important to unravel the knowledge underlying their biosynthesis. The gene clusters responsible for the isonitrile biosynthesis always encode an isonitrile synthase (IsnA or PvcA) and an oxygenase (IsnB or PvcB). PvcA converts the amino group into the isonitrile functionality and PvcB carries out a pathway-specific two-electron oxidative reaction on the isonitrile product.

The in vitro activities of the PvcB homologs from Pseudomonas aeruginosa,

Xenorhabdus nematophila, and Erwinia amylovora have been determined and the product formed in each reaction has been identified previously (Chapter 2) [107].

The X. nematophila PvcB enzyme catalyzes an oxidative decarboxylation reaction on tyrosine isonitrile, while the PvcB reactions in P. aeruginosa and E. amylovora retain the carboxyl group, demonstrating the diverse catalytic activities and mechanisms of the PvcB homologs. Although additional enzymes are encoded by the synthetic operons to further modify the isonitrile products to make them mature and bioactive, the PvcB reaction is undoubtedly a source of diversity in isonitrile-

94 containing metabolites.

PvcA performs upstream of the PvcB reaction, and catalyzes the formation of the isonitrile functionality. The PvcA enzymes from the above three pathways have not been characterized in vitro. But studies of the PvcA homologs in other pathways are known. The first identified isonitrile synthase IsnA, was derived from environmental DNA, and was responsible for making an indole isonitrile [36].

Labeling and feeding studies performed in E. coli followed by in vitro reconstitution assays suggested tryptophan and ribulose-5-phosphate as the precursors of the isonitrile product [37]. Recently, the activities of two other isonitrile synthases,

AmbI1 and WelI1, which are found in the cyanobacteria Fischerella ambigua and

Westiella intricate respectively, were demonstrated in vitro [105, 106]. They shared a common mechanism with IsnA by using tryptophan and ribose-5-phosphate or ribulose-5-phosate as the substrates and yielding an isonitrile-containing intermediate. This intermediate was described as unstable in all cases and was converted by the oxygenses (IsnB, AmbI3 and WelI3) to the stable indole isonitrile products. Thus, the product of the isonitrile synthase reaction was only predicted instead of demonstrated. Therefore, identifying the product of the PvcA reaction became one of the objectives of this investigation.

By overexpressing the XnPvcA protein in E. coli, we were able to detect the unstable metabolite in both the cell lysate and the culture medium by the XnPvcB reaction. Fractionation of the culture medium by size-exclusion chromatography

95 allowed us to isolate the product and further characterize it by 1H-NMR spectroscopy. The identical properties of this metabolite and the authentic tyrosine isonitrile exhibited in both 1H-NMR spectra and the XnPvcB reactions demonstrated that the unstable product was tyrosine isonitrile, which was the confirmed substrate in the PvcB reaction.

Next, we tried to reconstitute the in vitro activities of the PvcA enzymes from the biosynthetic pathways of P. aeruginosa and X. nematophila, where paerucumarin and rhabduscin are produced, respectively. Paerucumarin and rhabduscin both contain a tyrosine backbone. The structures of these two metabolites, as well as the studies of the indole isonitrile biosynthesis in the

PvcA/B homologous pathways led to the proposal that tyrosine and ribulose-5-P were the substrates of PaPvcA and XnPvcA. Incubation of these two compounds with the purified PvcA proteins (PvcA from Pseudomonas aeruginosa,

Xenorhabdus nematophila and Aeromonas hydrophila), however, did not support the product formation. The failure of the in vitro reconstitution assays might result from the cofactor or co-substrate missing, the inactive enzymes obtained after purification, or the incorrect substrate(s) used.

To determine whether an important cofactor or co-substrate was required for the PvcA reaction, some cofactors and potential co-substrates were added to the reactions conducted with the putative substrates tyrosine and ribose-5-P/ribulose-

5-P. Metal ions are important modulators of the biological response and are

96 required by various of enzymes for activity in the key cellular processes [111, 112].

Thus, the effects of the most common monovalent ions K+ and Na+, and the divalent metal ions Mg2+, Mn2+, Ca2+, Fe2+ and Zn2+ on the PvcA reaction were explored. However, none of the ions stimulated the activity of the PvcA enzyme.

Since the N and C atoms of the isonitrile group are derived from two different compounds, N-C bond formation and C-C bond cleavage must take place in isonitrile synthesis. Regarding this, tetrahydrofolate, a cofactor widely used in the metabolism of amino acids and nucleic acids by acting as a donor of a group with one carbon atom [113], was included in the PvcA reaction. Thiamine pyrophosphate (TPP) and pyridoxal phosphate (PLP), two effective cofactors involved in C-C-cleavage and synthesis [114, 115], were also tried. However, the product was not made successfully. In addition, the process to make the isonitrile group may involve energy-requiring steps. Thus, adenosine triphosphate (ATP) was added to the PvcA reaction, but also failed to support the activity. Finally, the molecule S-adenosyl methionine (SAM), which is a common co-substrate involved in methyl group transfers, transsulfuration, and aminopropylation [116], was examined in the PvcA reaction but was definitely not the right key. Although effects of a lot more cofactors remained undetermined so conclusions could not be drawn at this time, we went on to examine the other possibilities accounting for the inactive PvcA reaction.

Other than cofactors or coenzymes, the activity of PvcA could be affected by

97 many factors including temperature, buffer system, and natural stability. For instance, Tris buffer caused precipitations of all PvcA enzymes in our hands. The functions of the PvcA proteins might also get lost during purification even in the appropriate buffer. To ensure that active and functional PvcA enzymes were utilized in the reactions, the lysate from E. coli cells overexpressing the PvcA protein was tried as the enzyme source. Any cofactor or co-substrate required for

PvcA activity would also be present in the lysate. However, incubation of the putative substrates with the lysate hosting the PvcA protein failed to generate the product, either.

The failure to reconstitute the PvcA activity by incubating multiple PvcA enzymes, either purified or in the crude lysates, with tyrosine and ribose-5-

P/ribulose-5-P in the presence of various common cofactors led us to conclude that the substrate(s) proposed for the reaction was (were) incorrect. Therefore, we went back to re-examine the conditions in the reactions of IsnA, AmbI1 and WelI1, which used tryptophan and ribulose-5-P as the substrates. In contrast to the robust isonitrile metabolite production by the cells that heterologously expressed

IsnA/IsnB, substrates (tryptophan/ribulose-5-P)-directed synthesis of indole isonitrile by the purified IsnA/IsnB enzymes yielded small quantities of product after the reaction was incubated at 30 °C for 5 hours or longer. The yields of the products in the AmbI1/AmbI3 and WelI1/WelI3 reactions were not specified, but product isolation was also conducted from the long time reactions. We did not know the

98 catalytic efficiency of the oxygenases in those pathways, but the homologous PvcB enzymes we characterized exhibited relatively rapid turnovers, raising a question herein: whether the activities of the isonitrile synthases were very weak or better substrates might exist. Regarding this, we went on to investigate whether the correct substrates were used in our PvcA reactions.

The reconstitution of the isocyanovinylphenol pathway in vivo by co- expressing AhPvcA and XnPvcB in E. coli allowed us to perform feeding and labeling studies to track the origin of the product skeleton, as well as the source of the isonitrile carbon. Growing the cells in LB medium supplemented with 3,5-ring-

D2 tyrosine for different times showed that the product with a labeled backbone was produced very rapidly upon 3,5-ring-D2 tyrosine addition, indicating the added tyrosine was used directly by the AhPvcA enzyme in the cells, instead of converting to another compound first. Similarly, the cultures grown in medium fed with m- fluoro tyrosine generated the product with a m-fluoro benzene ring, suggesting other than tyrosine, some tyrosine analogs could serve as the substrates for the

PvcA/PvcB enzymes. This phenomenon was also observed in the study of the isonitrile biosynthetic enzymes from the hapalindole-type alkaloid pathway [117].

The WelI1/WelI3 proteins had the ability to convert 7 different tryptophan derivatives into the indole isonitrile product. In other words, both the isonitrile synthase and the oxygenase display substrate promiscuity, which will allow biosynthetic and functional studies of more isonitrile metabolites with different

99 modifications in the future. Incorporation of the tyrosine backbone into the isonitrile product indicated tyrosine is one of the substrates of the PvcA enzyme. In order to further demonstrate that the isonitrile N is derived from the amino group of tyrosine,

15N-tyrosine can be utilized in the future studies.

After confirming tyrosine to be one of the substrates, we suspected the other substrate (ribose-5-P/ribulose-5-P) which provides the isonitrile carbon was incorrect. In order to get more information about this substrate, labeling studies were performed to track the isonitrile carbon. E. coli cells co-expressing the

AhPvcA and XnPvcB proteins were grown in LB medium supplemented with glucoses with mono-13C label at different positions (D-[1-13C], D-[2-13C], D-[3-13C],

D-[4-13C], D-[5-13C] and D-[6-13C] glucoses). Isocyanovinylphenol with the 13C isonitrile group was only produced from the cultures incubated with D-[2-13C] and

D-[3-13C] glucoses for 5 hours, based on the 13C-NMR results. MS analyses of the products isolated from the cultures grown for different times revealed the 13C from either D-[2-13C] or D-[3-13C] glucose was incorporated into the product over time, and the incorporation occurred more rapidly in the culture supplemented with [2-

13C] glucose. The ability of both glucoses to label the isonitrile group indicated the source of the isonitrile carbon could be metabolized from both C2 and C3 of glucose; the rapid incorporation suggested the isonitrile carbon precursor was not far away from glucose in the metabolism.

100

Glucose, the important energy used by organisms, is metabolized via different pathways. Upton catabolism to pyruvate through glycolysis, glucose C1 and C6;

C2 and C5; and C3 and C4 are scrambled when 3-carbon metabolites are produced. However, in our labeling studies, the isonitrile products isolated from the cultures supplemented with D-[4-13C] and D-[5-13C] glucoses did not have a labeled isonitrile group, in contrast to those produced in the cultures with D-[3-13C] and D-[2-13C] glucoses, suggesting the isonitrile carbon comes from a metabolite before the glucose is converted to 3-carbon compounds. Glucose is also metabolized through the pentose phosphate pathway (PPP), which is parallel to glycolysis and generates 5-carbon sugars including ribose-5-phosphate, a precursor for the synthesis of nucleotides. The reactions taking place in the oxidative phase and nonoxidative phase of the PPP can be used to explain our results. Glucose C3 moves into ribulose-5-P C2 in the oxidative phase of the PPP, ribulose-5-P is further metabolized to ribose-5-P and xylulose-5-P in the non- oxidative phase (Figure 3.13 A). Thus, C2 of ribuloseo-5-P, ribose-5-P and xylulose-5-P could be labeled by [3-13C] glucose. Interestingly, C2 of ribulose-5-P can also be derived from glucose C2 through the transketolase reaction which converts fructose-6-P and glyceraldehyde-3-P to erythrose-4-P and xylulose-5-P

(Figure 3.13 B). In this regard, C2 of ribulose-5-P, ribose-5-P and xylulose-5-P could be labeled by both [2-13C] and [3-13C] glucoses.

101

A H2C OH C O HO C H H C OH C HO CHO O C O C OH 2- CH2OPO3 H C OH H C OH H C OH H C OH H2C OH HO C H HO C H HO C H O HO C H C O xylulose-5-P H C OH H C OH H C OH H C OH H C OH H C OH H C OH H C H C OH H C OH 2- 2- 2- 2- CHO CH2OH CH2OPO3 CH2OPO3 CH2OPO3 CH2OPO3 H C OH glu cose 6-phosphogluconate ribulose-5-P glucose-6-P 6-phosphogluconolactone H C OH H C OH CH OPO 2- 2 3 ribose-5-P

B H2C OH C O H C OH H C OH fructose-6-P 2- CH2OPO3 H C OH 2 ribulose-5-P C O H2C OH CHO CHO HO C H C O H C OH H C OH CHO H C OH HO C H HO C H HO C H H C OH H C OH H C OH H C OH H C OH H C OH 2- 2- CH OPO CH2OPO3 H C OH H C OH 2 3 H C OH 2- xylulose-5-P CH OPO 2- CH2OH CH2OPO3 + 2 3 ribose-5-P glucose glucose-6-P + H C O H H C OH C O CH OPO 2- 2 3 H C OH H C OH glyceraldehyde-3-P 2- CH2OPO3 erythrose-4-P

Figure 3.13 Schematic of partial glycolysis and the pentose phosphate pathway showing fate of

(A) glucose C3 and (B) glucose C2. C2 labeled ribulose-5-P, ribose-5-P and xylulose-5-P are highlighted.

102

C2, the ketone carbon of ribulose-5-P, is more favorable to be the isonitrile carbon source according to the possible chemical mechanism of the isonitrile group formation. Although little is known about this mechanism, the first step of synthesizing the isonitrile group should be the N-C bond formation. There are several ways of forming a nitrogen-carbon bond. Here, Schiff base formation or imine formation becomes a very suitable method (Scheme 3.3 A), involving tyrosine with its amino group and an aldehyde with a good leaving group. The amine nitrogen acts as a nucleophile and attacks the aldehyde carbon of the second substrate. The intermediate formed is then dehydrated, leaving us with a

C=N double bond. The leaving group is further expelled to make the nitrogen carbon triple bond. Scheme 3.3 B illustrates how a ketone (ribulose-5-P) could be used as the substrate to make the isonitrile group. Another reasonable mechanism is shown in Scheme 3.3 C, which mimics the formation of the isonitrile group in tyrosine isonitrile synthesis. A formyl group is generated first, followed by the dehydration reaction, resulting in the isonitrile group. However, adding formyl tyrosine either into the culture expressing AhPvcA-XnPvcB or into the reaction with purified PvcA-XnPvcB did not generate the isonitrile product, suggesting the PvcA enzyme did not make the isonitrile product through the formyl intermediate.

103

A H A H A OH H LG C O C H C LG N C N LG H Tyr H N Tyr Tyr N Tyr H :B H :B

B

H C OH H2C OH Tyr 2 H2C O H :B Tyr H H-A C N C O N C C N H H C OH H C OH N Tyr H C OH H C OH H2C OH Tyr 2- CH2 CH OPO 2- CH2OPO3 HC OH 2 3 2- O CH2OPO3

C

O O O H O H2O Tyr OH CH H OH + NH 3 C C C 2 H C O Tyr N C H O CH3 N NH N Tyr Tyr Tyr H H A

Scheme 3.3 Possible mechanisms involved in isonitrile group formation. (A) Imine formation followed by isonitrile group formation. (B) Proposed reactions between ribulose-5-P and tyrosine to make tyrosine isonitrile. (C) Synthesis of tyrosine isonitrile from formyl tyrosine.

104

Both ribose-5-P and ribulose-5-P were tried as the substrates in the PvcA reactions, but neither of them supported the reactions, as described earlier.

Xylulose-5-P is not commercially available, so was not examined at this time.

Instead of focusing on a specific compound, efforts were made in searching broadly for the substrate. Firstly, a phosphorylated metabolite library [118] containing 167 various types of phosphorylated compounds, including sugars, nucleotides, amino acids and others, was screened in the PaPvcA-PaPvcB reactions. The PaPvcB reaction solution turns bright yellow quickly in the presence of tyrosine isonitrile, α-KG and Fe2+, but no color change was observed in any of the reactions, suggesting none of the compounds in the library was converted by

PaPvcA (in the presence of tyrosine) to tyrosine isonitrile. Secondly, most of the metabolites from the central carbon metabolic pathways (glycolysis, the PPP, glycerol metabolism) were examined in the PvcA-XnPvcB reactions. However, there was no sign of product formation based on HPLC analyses of the reaction extracts, indicating the right substrate was not present among the compounds we tried so far. Finally, the entire metabolite pool of E. coli cells was examined in the

PvcA reaction in search of the substrate. Metabolites extracted from the E. coli cells failed to support the PvcA activity, but when PvcA enzyme was incubated with the cell lysate, product formation occurred (detected by the XnPvcB reaction). All the three PvcA enzymes (XnPvcA, PaPvcA and AhPvcA) exhibited such activities when added to the cell lysate, demonstrating all of them were functional after

105 purification. Interestingly, there were more turnovers observed in the XnPvcA-

XnPvcB reactions than in the AhPvcA-XnPvcB and PaPvcA-XnPvcB reactions when the same amounts of the enzymes were utilized and the reactions were incubated for the same time. These three PvcA homologs were expected to catalyze the isonitrile synthesis by using the same substrates and yielding the same products. So whether the different catalytic efficiency between them was due to their intrinsic characters (fast or slow enzyme), or because of the special interactions between the PvcA and PvcB proteins, needs to be explored further.

We tried to isolated the unknown substrate from the crude lysate for identification. To our surprise, fractionation of the cell lysate by dialysis, size- exclusion chromatography, ion exchange chromatography and some other methods revealed that a macromolecule, probably an E. coli metabolic enzyme, and a small compound, probably with a negative charge were both required for

PvcA activity. The identities of the macromolecule and the small compound remain unknown, and both of them need to be further purified and characterized in the future.

An assumption was made to explain the results obtained so far. The real substrate for the PvcA reaction is generated by another enzyme (the unknown macromolecule described above) in a reversible reaction, but the equilibrium for this reaction lies far to the left, which produces the unknown small compound

106 described above. When the macromolecule (E) and the small compound (A) are combined in the presence of PvcA, then the PvcA substrate (X) can be drawn off by the PvcA reaction thus is continuously being generated (Figure 3.14). If so, the best way to determine the real substrate of the PvcA enzyme is to identify the small compound A first, and then we may get an idea about what E and X are.

E PvcA A X tyrosine isonitrile tyrosine

Figure 3.14 Proposed function of the unknown macromolecule E and the relationship between the

PvcA substrate X and the unknown small compound A.

107

3.4 Material and Methods

Buffers and common biochemicals were purchased from Fisher Scientific or

Sigma, and were used without further purification. Organic solvents were stored over molecular sieves, and imidazole used in column chromatography was recrystallized from ethyl acetate.

3.4.1 Cloning of pvcA GenesA [40]

The PapvcA gene was amplified from Pseudomonas aeruginosa genomic

DNA using primers to incorporate the restriction sites at 5’ and 3’ ends, and subcloned into a modified pET15bTEV plasmid [40]. Sequences of XnpvcA and

AhpvcA genes were identified through an NCBI BLAST search originating from the

PapvcA sequence. The genes for XnPvcA and AhPvcA were synthesized commercially (Genewiz, Inc.) including added 5’ (NdeI for XnpvcA, XcoI for AhpvcA) and 3’ (BamHI) sequences. The synthetic genes were subcloned into the pET15bTEV plasmid for expression, similarly as PapvcA.

AhpvcA was also cloned into the multiple cloning site 1 (MCS1) of the pETDuet-1 vector with the restriction sites XcoI and BamHI, while XnpvcB

(described in Chapter 2) was inserted into the multiple cloning site 2 (MCS2) of this plasmid using NdeI and XhoI. This dual expression vector was constructed for expression of two proteins at the same time.

______

A Plasmids containing the pvcA genes were constructed by Eric Drake and Geoff Lippa

108

E. coli 10G cells were transformed with these expression vectors and stored in 10% (v/v) glycerol at -80 °C.

3.4.2 Heterologous Expression of the PvcA Proteins

Plasmids containing the pvcA genes were isolated from 10G cells with

Promega Wizard kit and stored at -20 °C. E. coli BL21(DE3) cells were transformed with the isolated plasmid and transformants were selected by growth on LB plates containing carbenicillin. Single colonies were picked and used to inoculate 5 mL

LB, which was incubated at 37 °C with shaking at 250 rpm overnight. Overnight cultures were then added to a flask containing 800 mL LB supplemented with 100

µg/mL carbenicillin. The cultures were grown at 37 °C with rotary shaking until an optical density of ~0.6 at 600 nm was achieved, and then cold shocked on ice to

~18 °C. 1 mM IPTG was added to the cultures in order to induce recombinant protein expression. The cultures were then shaken at 18 °C for 16 hours. After centrifugation at 3993 x g for 20 minutes, 4-5 g cells were harvested from one liter of medium. The cell paste was stored at -80 °C until needed.

3.4.3 Purification of the PvcA Proteins

Purification of the PvcA proteins was accomplished by immobilized-metal affinity chromatography. Frozen cell paste (typically 3-4 g) was suspended in 20 mL of buffer composed of 50 mM Na2HPO4, pH 7.8, 250 mM NaCl, 10 mM imidazole, 0.2 mM TCEP and 10% (v/v) glycerol. The cell suspension was thawed on ice and before lysis 250 units of benzonase (Novagen) and 0.5 mM

109 phenylmethylsulfonyl fluoride were added. The cells were lysed by two passages through a French Press at 10,000 psi. The cell debris was removed by centrifugation at 26892 x g for 35 minutes. The supernatant was applied to a 1 mL

HisPurTM Ni-NTA column (Thermo Scientific) which was pre-equilibrated in the buffer that was used to lyse the cells. After sample loading, the column was washed with 15 mL lysis buffer, and a step gradient consisting of 25 mL of lysis buffer containing 50 mM imidazole, followed by 25 mL of lysis buffer with 300 mM imidazole was used to elute the protein. The purity of the protein was checked by

SDS-PAGE, and the pooled protein fractions were dialyzed against 50 mM HEPES, pH 7.8, 50 mM NaCl and 10% (v/v) glycerol overnight at 4 °C. Proteins were then concentrated using an Amicon Ultra 9K MWCO centrifugation filter. 10% (v/v) glycerol was added afterwards and the concentration was determined spectrophotometrically using an extinction coefficient of 38,000 M-1cm-1 for

PaPvcA, 41,000 M-1cm-1 for XnPvcA and 27,000 M-1cm-1 for AhPvcA at 280 nm.

The protein sizes (PaPvcA ~37 kDa, XnPvcA ~38 kDa, AhPvcA ~37 kDa) and extinction coefficients were estimated using ExPASy’s “Protparam” tool.

Concentrated proteins (~10 mg/mL) were aliquoted and stored at -80 °C.

3.4.4 PvcA Activity Assays

3.4.4.1 Activities of the Purified PvcA Enzymes

Activities of the purified PvcA proteins were evaluated in the presence of the proposed substrates tyrosine and ribose-5-phosphate/ribulose-5-phosphate.

110

Typical reaction mixtures contained 500 μM D, L-tyrosine, 500 μM D-(+)-ribose-5- phosphate, 5 units of phosphoriboisomerase (Sigma) and 300 nM PvcA enzyme

(all three PvcA homologs were examined individually) in 1 mL 50 mM HEPES, pH

7.8, 150 mM NaCl. Ribulose-5-phosphate was generated in situ from ribose-5- phosphate in the reaction catalyzed by phosphoriboisomerase. Short time reactions were monitored spectrophotometrically (200-500 nm) using a Carey 50

Bio UV-vis spectrophotometer. Scans of the reaction mixtures were acquired every

0.5 minute for about 20 minutes at room temperature. Long time reactions were monitored by HPLC at 280 nm to determine whether the substrate tyrosine was consumed and the product was formed over time. Aliquots were removed from the reaction mixture every 1 hour in the first 5 hours and 16 hours after its incubation at room temperature, and were analyzed by HPLC.

3.4.4.2 In Vitro PvcA-XnPvcB Coupled Reactions

The XnPvcB reaction was also coupled to the PvcA reaction, in order to detect the product formation in the PvcA reaction more easily. The purified XnPvcB enzyme (500 nM) with its co-substrates α-ketoglutarate (500 μM) and

(NH4)2Fe(SO4)2 (10 μM) were added to the PvcA reactions described in Section

3.4.4.1 and the reaction mixtures were monitored spectrophotometrically and by

HPLC, as described above.

3.4.4.3 Cofactor and Co-substrate Requirements

To look for any cofactor or co-substrate requirements in the PvcA reaction,

111 some common cofactors and co-substrates listed in Table 3.2 were added into the

PvcA-XnPvcB coupled reactions at appropriate concentrations. Various metal ions

MgCl2(H2O)6, MnCl2(H2O)4, CaCl2(H2O)2, KCl, and ZnCl2 (100 μM) were prepared as cocktails and checked. (NH4)2Fe(SO4)2 and NaCl are required for the PvcB reaction so were always present. TPP (500 μM), PLP (500 μM), three forms of tetrahydrofolate (500 μM), ATP (500 μM) and SAM (500 μM) were examined individually.

Table 3.2 Cofactors and co-substrates examined in the in vitro PvcA reaction

Cofactor Metal ions Mg2+, Fe2+, Mn2+, Ca2+, K+, Na+, Zn2+

Others Thiamine pyrophosphate, Pyridoxal phosphate, N5-Formyl-

THF, N5, N10-Methenyl-THF, N10-Formyl-THF, ATP

Co-substrate S-Adenosyl methionine

3.4.4.4 Activities of the Cell Lysates Hosting PvcA Proteins

Reactions were also conducted with the E. coli cell lysates harboring the PvcA proteins instead of the purified PvcA enzymes. Tyrosine and ribose-5- phosphate/ribulose-5-phosphate were added to the lysate from BL21(DE3) cells expressing PaPvcA, XnPvcA or AhPvcA. The XnPvcB enzyme with its co- substrates Fe2+ and α-ketoglutarate were included to convert the PvcA product, to

112 a readily detectable species. The reactions were incubated at room temperature for 5 hours or overnight and analyzed by HPLC.

3.4.5 Expression of AhpvcA-XnpvcB/pETDuet-1 in E. coli and Metabolite

Detection

The pETDuet-1 vector containing the AhpvcA and XnpvcB genes was expressed in E. coli BL21(DE3) cells as described in Section 3.4.2. As a control, the cells transformed with no plasmid, the pETDuet-1 vector containing only the

AhpvcA gene and the pACYCDuet-1 vector containing the XnpvcB gene were also grown. To check for protein expression in the cells with different plasmids, aliquots were removed from the growing cultures 5 hours after IPTG induction. The aliquots were incubated with 2 x SDS-PAGE gel loading buffer at 70 °C for 30 minutes with vortex for several times. After centrifugation at 21380 x g for 20 minutes, supernatants were loaded onto SDS-PAGE for analysis.

To examine the metabolites produced by the cells harboring no plasmid or different plasmids, the culture medium from each culture was collected after the cells were spun down. 20 mL culture medium was extracted with 2 x 20 mL ethyl acetate. Organic phases were combined and dried by rotary evaporation. The residues were dissolved and diluted in methanol for HPLC analysis.

3.4.6 Reaction Product Identification

In order to identify the PvcA product, the plasmid XnpvcA/pET15b was overexpressed in E. coli BL21(DE3) cells as described in Section 3.4.2. The

113 expression of the XnPvcA protein is required for product formation, but whether the product would stay inside the cells or move out of the cells was unknown. Both the cell paste and culture medium were collected and saved. The cell lysate (0.3 g cells per mL HEPES buffer) or the culture medium (20 mL concentrated to 1 mL) was incubated with the purified XnPvcB enzyme (10 μM) and the additional substrates (500 μM α-KG and 100 μM (NH4)2Fe(SO4)2) in 50 mM HEPES, pH 7.8,

150 mM NaCl with a total volume of 2 mL at room temperature for 2 hours. The reactions were then extracted with the same volume of ethyl acetate twice. The extracts were combined, dried and dissolved in 500 μL methanol for HPLC analysis.

Both the cell lysate and the culture medium were fractionated. Dialysis of the cell lysate was performed to determine whether the product was a small molecule.

The cell lysate was dialyzed in dialysis tubing (6,000-8,000 MWCO) against deionized H2O overnight at room temperature. The dialysate was frozen at -80 °C for 1 hour and lyophilized to dryness afterwards. The concentrated dialysate or the lysate retained inside the dialysis tubing was incubated with the purified XnPvcB enzyme and its co-substrates (α-KG and Fe2+) at room temperature for 2 hours.

The product was extracted by ethyl acetate and detected by HPLC.

Size-exclusion chromatography was performed to isolate the metabolite from the culture medium. 3 mL culture medium (concentrated from 30 mL) was fractionated on a 105 cm x 1 cm column filled with Bio-Gel P2 resin (Bio-Rad) at

0.2 mL/min with deionized H2O as the eluent. ~35 fractions were collected and

114 each fraction contained 5 mL eluent. 500 μL aliquots from each fraction were incubated with 100 μM α-KG, 100 μM ascorbic acid, 2 μM (NH4)2Fe(SO4)2 and 400 nM XnPvcB in 500 μL buffer (50 mM HEPES pH 7.8, 150 mM NaCl) at room temperature and were monitored spectrophotometrically for 5 minutes. The fractions which triggered the absorbance increase at 310 nm were lyophilized and further analyzed by 1H-NMR spectroscopy. For NMR analysis, the dried material was dissolved in 600 μL D2O.

To further examine the components in the fraction, 1 mM α-KG, 1 mM ascorbic acid, 50 μM (NH4)2Fe(SO4)2 and 10 μM XnPvcB were added to the NMR sample described above. The reaction was incubated at room temperature for 1 hour and then quenched by introducing the methylating reagent TMS-CHN2. 1 mL ethyl acetate containing 100 mM methanol and 50 mM TMS-CHN2 was added to the 600

μL reaction solution. The solution was vortexed for 5 minutes. The organic phase was then taken and evaporated to dryness. The residue was dissolved in 600 μL

1 methanol-d4 and analyzed by H-NMR spectroscopy.

3.4.7 Substrate Identification

3.4.7.1 Tracking of Tyrosine in the Biosynthetic Pathway of

Isocyanovinylphenol in E. coli

The vector AhpvcA-XnpvcB/pETDuet-1 was expressed in E. coli as described in section 3.4.5. Two hours after induction by 1 mM IPTG, 3,5-ring-D2 tyrosine was added to a 100 mL culture to a final concentration of 0.5 mM. The culture was

115 shaken at 18 °C and aliquots (5 mL) were removed from the growing culture 10 minutes, 20 minutes, 1 hour, 2 hours, 3 hours and 5 hours after addition of the labeled tyrosine, and were then extracted with 8 mL ethyl acetate. After the phases separated, the organic phases were dried by rotary evaporation and dissolved in

500 μL methanol for MS analysis.

The tyrosine analog, m-fluoro tyrosine was added to the culture that expressed AhPvcA and XnPvcB proteins, similarly. Metabolites were extracted from the culture medium (5 mL) before adding m-fluoro tyrosine, 1 hour, 2 hours,

5 hours, and 16 hours after its addition by ethyl acetate, and analyzed by HPLC.

3.4.7.2 Tracking of the Isonitrile Carbon in the Biosynthetic Pathway of

Isocyanovinylphenol in E. coli

The pETDuet-1 vector containing the AhpvcA and XnpvcB genes was transformed into BL21(DE3) cells for expression as described in section 3.4.5.

After 1 mM IPTG was added to the culture to induce protein expression, the growth of the cells was continued at 18 °C for 2 hours. Then the cultures were supplemented with 0.5% D-[1-13C], D-[2-13C], D-[3-13C], D-[4-13C], D-[5-13C], D-[6-

13 13 13 C] and D-[UL- C6] glucoses separately. To isolate enough material for C-NMR analysis, 5 mL cultures supplemented with various 13C glucoses were grown at

18 °C for another 5 hours, followed by metabolite extraction by 8 mL ethyl acetate.

The organic phases were dried by rotary evaporation and the residues were

13 dissolved in 600 μL methanol-d4 for C-NMR analysis. To determine the time

116 course of the incorporation of 13C into the product, 1 mL aliquots were removed from 5 mL cultures 10 minutes, 20 minutes, 40 minutes and 2 hours after adding

13C labeled glucoses, and were extracted with 1 mL ethyl acetate twice. Organic layers were combined, dried under a stream of N2 gas, dissolved in 300 μL methanol and analyzed by mass spectrometry.

3.4.7.3 Looking for the Substrate in a Phosphorylated Metabolite Library

A library containing 167 phosphorylated compounds was kindly donated by Dr.

Karen Allen [118]. The names of the compounds are listed in Table 3.3. These compounds were spread in two 96 well plates and each well contained 20 μL compound with a concentration of 2 mM. To rapidly screen this library by the PvcA reaction, 5 μL compound from each well was mixed with 95 μL mixture containing

100 μM D, L-tyrosine, 100 μM MgCl2, 100 μM ascorbic acid, 200 μM α- ketoglutarate, 1.5 μM PaPvcA and 4.7 μM PaPvcB. All of the reactions were incubated at room temperature for 5 hours and the color change of the reaction solutions was observed. The experiment was repeated three times.

117

Table 3.3 List of the compounds in the phosphorylated metabolite library

# Name # Name 1 p-Nitrophenylphosphate (pNPP) 85 Allose-6-phosphate 2 Pyridoxal-5'-phosphate 86 Ribitol-1-phosphate 3 Pyrophosphate (Ppi or POP) 87 Ribonate-5-phosphate 4 Glycerol-phosphate (GP) 88 Arabinonate-5-phosphate 5 Coenzyme A (CoA) 89 Sorbitol-1-phosphate 6 O-Phosphorylethanolamine 90 Mannitol-6-phosphate 7 Acetyl-phosphate 91 D-Mannonate-6-phosphate 3-Phosphoglyceric acid (glycerol 3 8 phosphate) 92 D-Allitol-6-phosphate 2-Phosphoglyceric acid (glycerol 2- 9 phosphate) 93 D-Allonate-6-phosphate 10 DL-Glyceraldehyde 3-phosphate 94 D-Xylitol-5-phosphate 11 Carbamoyl phosphate 95 D-2-Deoxy-ribitol-5-phosphate 12 Imidodiphosphate (PNP) 96 D-2-Deoxy-ribonate-5-phosphate 13 Thiamine pyrophosphate 97 D-Xylulose-5-phosphate 14 Phosphonoformic acid (PFA) 98 D-Threonate-4-phosphate 15 N-Phosphonomethyl glycine (PMG) 99 D-Galactitol-6-phosphate DL-2-Amino-3-phosphonopropionic acid 16 (APPA) 100 D-Erythronate-4-phosphate 17 Phospho(enol)pyruvic acid (PEP) 101 Iso-erythritol-4-phosphate 18 Riboflavin-5-phosphate (FMN) 102 D-Ribulose-5-phosphate β-Nicotinamide adenine dinucleotide 19 phosphate (NADP) 103 D-Threitol-4-phosphate 20 Phosphocholine 104 D-Lyxonate-5-phosphate 21 Erythrose 4-phosphate 105 D-Tagatose-6-phosphate 22 Arabinose 5-phosphate 106 L-Ribulose-5-phosphate 23 D-Ribose 5-phosphate 107 L-Erythronate-4-phosphate 24 2-Deoxyribose 5-phosphate 108 D-2-Keto-glucose-6-phosphate 25 5-Phosphoribosyl-1-pyrophosphate 109 D-3-Deoxy-glucose-6-phosphate 26 α-D-Glucose-1-phosphate 110 D-3-Deoxy-gluconate-6-phosphate 27 D-Glucose-6-phosphate 111 D-3-Deoxy-sorbitol-6-phosphate 28 2-Deoxy-D-glucose 6-phosphate 112 D-Altronate-6-phosphate 29 α-D-Glucose 1,6-bisphosphate 113 D-Psicose-6-phosphate 30 α D-Galactose 1-phosphate 114 L-Sorbose-1-phosphate 31 D-Fructose 6-phosphate 115 D-Iditol-6-phosphate 32 D-Fructose 1,6-bisphosphate 116 D-Sedoheptulose-7-phosphate 33 α-Mannose 1-phosphate 117 D-Glucose-3-phosphate

118

34 Mannose 6-phosphate 118 D-Glucitol-3-phosphate 35 Trehalose-6-phosphate 119 D-Gluconate-3-phosphate 36 Sucrose 6-phosphate 120 Adenosine 3',5'-diphosphate 37 O-Phospho-L-serine 121 D-Arabitol-5-phosphate 38 O-Phospho-L-threonine 122 D-Xylose-3-phosphate 39 O-Phospho-L-tyrosine 123 D-Xylitol-3-phosphate 40 6-Phosphogluconic acid 124 D-Xylonate-3-phosphate 41 ATP 125 D-Allose-3-phosphate 42 ADP 126 D-Allitol-3-phosphate 43 5'-AMP 127 D-Allonate-3-phosphate 44 3'-AMP 128 D-Glucose-3,5-diphosphate 45 dAMP 129 D-Glucitol-3,5-diphohsphate 46 UTP 130 Dihydroxyacetone phosphate (DHAP) 47 UDP 131 Glycolic Acid-O-P 48 UMP 132 L-Ribose-5-P 49 dUMP 133 L-Ribitol-5-P 50 TTP 134 L-Ribonate-5-P 51 TDP 135 L-Arabinose-5-P 52 TMP (dTMP) 136 L-Arabitol-5-P 53 CTP 137 L-Arabinonate-5-P 54 CDP 138 L-Arabitol-1-P 55 CMP 139 D-Xylose-5-P 56 dCMP 140 D-Xylonate-5-P 57 GTP 141 L-Xylose-3-P 58 GDP 142 L-Xylonate-3-P 59 GMP 143 L-Xylitol-3-P 60 dGMP 144 L-Xylose-5-P 61 ITP 145 L-Xylonate-5-P 62 IDP 146 L-Xylitol-5-P 63 IMP 147 D-Mannose-2-P 64 2'AMP 148 D-Mannitol-2-P 65 dATP 149 L-Sorbose-4-P 66 dCTP 150 D-Glucuronic acid-5-P 67 dGTP 151 D-Glucuronicitol-5-P 68 dTTP 152 D-Glucurononate-5-P 69 Glucosamine 6 phosphate 153 D-Lyxose-5-P 2-Keto-3-deoxy-6-phosphogluconate 70 (KDPG) 154 L-Lyxose-5-P 71 Galactonate-6-phosphate 155 L-Lyxitol-5-P 72 Beta-Glucose-1,6-bisphosphate 156 D-Arabitol-1-P

119

D-Glycero-beta-D-manno-heptose-1,7- 73 bisphosphate 157 D-Arabitol-4-P 2-Keto-3-deoxy-D-glycero-D- 74 galactonononic acid-9-phosphate 158 D-Glucitol-2-P 75 N-Acetylneuraminic acid-9-phosphate 159 L-Glucose-3-P D-Glycero-alpha-D-manno-heptose-1,7- 76 bisphosphate 160 L-Glucitol-3-P 77 Galactose-6-phosphate 161 L-Gluconate-3-P 3-Deoxy-D-manno-2-octoulosonate-8- 78 phosphate 162 D-Galactitol-1-P a-D-Glucosamine-1-phosphate (not 79 stable) 163 D-Galactitol-5-P 80 Histidinol-phosphate (not stable) 164 D-Mannitol-1-P 81 N-Acetyl-mannosamine-6-phosphate 165 D-Mannitol-4-P 82 Uridine-5'-diphosphoglucuronic acid 166 D-Mannitol-5-P L-Gulonic acid-2-Methylene Hydroxy 83 2-Deoxy-6-phophoglucitol 167 phosphoric acid 84 2-Deoxy-6-phosphogluconate

3.4.7.4 Searching for the Substrate in the Major Metabolic Pathways

The substrate which provides the isonitrile carbon in the PvcA reaction was sought in the major metabolic pathways, including the Embden-Meyerhof-Parnas

(the most common type of glycolysis), the pentose phosphate pathway, and the

Entner-Doudoroff pathway (a variant of glycolysis in some organisms). The metabolites related to glycerol metabolism were also tested. In the PvcA-XnPvcB assays described in Section 3.4.4.2, ribose-5-phosphate/ribulose-5-phosphate was replaced with the compounds listed in Table 3.4 and the other components remained unchanged.

Some of the metabolites in Table 3.4 were purchased from Sigma and were

120 added to the PvcA-XnPvcB coupled reactions to a final concentration of 1 mM.

These compounds included glucose, glucose-6-phosphate, fructose-6-phosphate, fructose-1,6-bisphosphate, 2-phosphoglycerate, phosphoenolpyruvate, 6- phosphogluconate, ribose-5-phosphate, erythrose-4-phosphate and glycerol.

Other compounds were synthesized from the reactions catalyzed by the purchased enzymes (Scheme 3.4). Products formed in these enzymatic reactions were not further purified and the final concentrations of them were not determined. Instead, the crude reaction mixtures were directly added to the PvcA-XnPvcB coupled reactions for examination.

121

Table 3.4 List of the metabolitesB examined in the in vitro PvcA reaction

Embden-Meyerhof- Entner-Doudoroff Pentose phosphate Glycerol

Parnas pathway pathway metabolism

Glucose 6-Phosphogluconolactone Ribulose-5-P Glycerol

Glucose-6-P 6-Phosphogluconate Ribose-5-P Glycerol-3-P

Fructose-6-P KDPG Erythrose-4-P

Fructose-1,6-bisP

Glyceraldehyde-3-P

Dihydroxyacetone-P

1,3-Bisphosphoglycerate

3-Phosphoglycerate

2-Phosphoglycerate

Phosphoenolpyruvate

Pyruvate

B The metabolites involved in two or more pathways are only listed once. For example, 6-P- gluconolactone is a common intermediate involved in the Entner-Doudoroff and the pentose phosphate pathways, but is listed in the former one. glyceraldehyde-3-P is a cross point of the

Embden-Meyerhof-Parnas and glycerol metabolism, and is shown in the former one.

122

Glycerol kinase Glycerol Glycerol-3-phosphate Mg2+ ATP ADP Glycerolphosphate dehydrogenase/ Triosephosphate isomerase Glycerol-3-phosphate Dihydroxyacetone-P + Glyceraldehyde-3-P NAD+ + Pi NADH + H+

Glyceraldehyde-3-P dehydrogenase Glyceraldehyde-3-phosphate 1,3-Bisphosphoglycerate

NAD+ + Pi NADH + H+

Phosphoglycerate kinase 1,3-Bisphosphoglycerate 3-Phosphoglycerate Mg2+ ADP ATP

Pyruvate kinase Phosphoenolpyruvate Pyruvate Mg2+ ADP ATP

Glucose-6P dehydrogenase Glucose-6-phosphate 6-Phosphogluconolactone

NADP+ NADPH

Phosphoriboisomerase Ribose-5-phosphate Ribulose-5-phosphate

Phosphogluconate dehydratase 6-Phosphogluconate 2-Keto-3-deoxy-6-phosphogluconate (KDPG)

Scheme 3.4 Enzymatic reactions to synthesize metabolites in the Embden-Meyerhof-Parnas, the pentose phosphate pathway, the Entner-Doudoroff pathway and glycerol metabolism. Enzymes are colored red.

3.4.7.5 Looking for the Substrate in the Metabolite Pool

3.4.7.5.1 Metabolites Extracted from E. coli Cells

E. coli BL21(DE3) cells were grown in 500 mL LB medium at 37 °C to an optical density (600 nm) of 1.0 and were then centrifuged at 3993 x g for 20 minutes.

The supernatant was discarded and the cell pellet was suspended in 5 mL 50 mM

123

HEPES, pH 7.8, 150 mM NaCl. 20 mL cold (-20 °C ) acetonitrile/methanol/water

(40/40/20) was added to the cell suspension, which was vortexed and incubated at -20 °C for 15 minutes, followed by centrifugation at 26892 x g and 4 °C for 15 minutes. The supernatant was transferred to a new tube and the cell pellet was extracted by the same solvent again at 4 °C for 10 minutes. After 10 min centrifugation at 4 °C, the supernatant was removed and combined with the previous one. Organic solvents were evaporated under vacuum and the remaining solution was frozen at -80 °C for 1 hour and then lyophilized to dryness.

Dried material was dissolved in 1 mL 50 mM HEPES, pH 7.8, 150 mM NaCl and to which 2 mM D, L-tyrosine, 2 mM α-KG, 0.2 mM (NH4)2Fe(SO4)2, 1.7 μM

AhPvcA or XnPvcA, and 1.5 μM XnPvcB were added. The reaction was incubated at room temperature for 1 hour and then extracted with 1 mL ethyl acetate. The extract was dried and dissolved in 100 μL methanol for HPLC analysis.

3.4.7.5.2 Metabolites in the E. coli cell lysates

E. coli BL21(DE3) cells were transformed with different plasmids including pET15b, XnPvcB/pET15b, PGDH/pET20b, PqsA/pET14b and ACP2/pET14b, respectively. Transformants as well as the cells containing no plasmid were grown in 100 mL LB medium at 37 °C to OD600 ~0.6. 1 mM IPTG was added to each culture to induce protein expression and the growth was continued at 18 °C for 5 hours. Cells were harvested by centrifugation at 3993 x g for 20 minutes and stored at -80 °C until needed. The cell pastes were suspended in 50 mM HEPES, pH 7.8,

124

150 mM NaCl (0.25 g cell per mL buffer) and lysed by one passage through a

French Press at 10,000 psi. Cell debris was removed by centrifugation. 800 μL supernatants were combined with 500 μM tyrosine, 500 μM α-KG, 100 μM ascorbic acid and 50 μM (NH4)2Fe(SO4)2. 3.5 μM XnPvcA and 3.7 μM XnPvcB were added to initiate the reactions. The reactions were incubated at room temperature for 5 hours, followed by 1 mL ethyl acetate extractions. The extracts were dried under a stream of N2 gas, dissolved in 100 μL methanol and analyzed by HPLC.

3.4.7.5.3 MS Analyses of the E. coli Cell Lysate Before and After the PvcA

Reaction

10 mL lysate (0.3 g cell per mL lysate) from E. coli cells expressing XnPvcB was prepared as described in Section 3.4.7.4.2. 10 μM XnPvcA was added to 5 mL lysate and the other 5 mL lysate was left untreated. Both samples were incubated at room temperature for 3 hours, followed by centrifugation in an Amicon

Ultra centrifugal filter (9K MWCO). Small molecules passing through the filter membrane from each sample were collected, frozen and lyophilized. Dried samples were sent to the West Coast Metabolomics Center (University of

California, Davis) for MS analysis.

3.4.7.5.4 Fractionation of the E. coli Cell Lysate Hosting XnPvcB

The lysate from E. coli cells expressing XnPvcB was fractionated to isolate the substrate. The cell pastes were suspended in HEPES buffer and were lysed by French Press as described in Section 3.4.7.4.2. The lysate was then centrifuged

125 at 26892 x g for 40 minutes. The cell free extract (supernatant) and the cell debris

(pellet) were saved and examined in the XnPvcA-XnPvcB coupled reactions, separately. The cell free extract was then applied to an Amicon Ultra 9K MWCO centrifugal filter for separation of macromolecules and metabolites. The spin columns were centrifuged at 4 °C and 10760 x g for 45 minutes. The flow through and the solution retained above the filter membrane were collected and checked in the XnPvcA-XnPvcB coupled reactions, separately. Centrifugal filters with different molecular weight cut-offs (3K, 20K, 50K and 100K) were also tried.

The XnPvcA-XnPvcB coupled assays were conducted to check samples from each step of fractionation. Generally, to ~500 μL buffer containing the different parts of fractionation, 500 μM tyrosine, 500 μM α-KG, 100 μM ascorbic acid, 50 μM

(NH4)2Fe(SO4)2, 3.5 μM XnPvcA and 3.7 μM XnPvcB were added. The reactions were incubated at room temperature for 4 hours, and then extracted by 1 mL ethyl acetate. Extracts were dried and dissolved in 100 μL methanol for HPLC analysis.

Preparation of the macromolecules by dialysis

To further remove the small molecules from the solution retained above the membrane of the centrifugal filter described above, the solution (5-10 mL) was put in dialysis tubing (6,000-8,000 MWCO) and dialyzed against 2 x 1 L 10 mM HEPES, pH 7.8, 30 mM NaCl at 4 °C overnight. The dialyzed cell free extract was used as the source of the macromolecules in the XnPvcA reaction.

Fractionation of the proteins by ammonium sulfate precipitation

126

Proteins in the cell lysate were susceptible to degradation by proteases over time so the dialyzed cell free extract could not be kept long. In order for easier access to the macromolecules, the dialyzed cell free extract was further treated by ammonium sulfate. The proteins in the cell free extract were first precipitated by

30% ammonium sulfate. Solid ammonium sulfate was added slowly to the lysate solution to a final concentration of 30% (179 g per liter of solution) and the solution was stirred at room temperature for 15 minutes, followed by centrifugation at 4 °C and 26892 x g for 20 minutes. The supernatant was separated from the pellet and a 70% ammonium sulfate precipitation was performed with the supernatant by slowly adding 273 g of solid ammonium sulfate per liter of supernatant solution.

The solution was stirred and centrifuged as above. The pellets from the 30% and

70% ammonium sulfate precipitations were resuspended in HEPES buffer, dialyzed against the buffer at 4 °C for several hours and then tested in the XnPvcA-

XnPvcB coupled reactions.

Purification of the macromolecules by ion exchange chromatography

Ion exchange chromatography was performed to further purify the macromolecule required for the PvcA reaction. After dialyzing away most of the small molecules, the cell free extract was loaded onto a 1 mL HiTrapTM Q HP anion exchange or a 1 mL HiTrapTM SP HP cation exchange column for purification. The anion exchange column was pre-equilibrated and washed with 20 mL 20 mM Tris, pH 7.8. 100 mL Tris buffer containing a gradient of NaCl (0 mM-1 M) was used to

127 eluent the macromolecules that bind to the column. Similarly, the cation exchange column was tried and 50 mM HEPES, pH 7.8 with NaCl was used as the elution buffer during the purification. 500 μL aliquots from each fraction were combined with small molecules and examined in the XnPvcA-XnPvcB coupled reactions.

Separation of the molecules by size-exclusion chromatography

Size-exclusion chromatography was performed to separate macromolecules and small molecules, and more importantly, to isolate the small compound required for the PvcA reaction. The cell free extract was fractionated on a 105 cm x 1 cm column filled with Bio-Gel P2 resin (Bio-Rad). The column was pre-equilibrated in

10% methanol and run at 0.2 mL/min with the same solvent. 3 mL fractions were collected and 300 μL aliquots from each fraction were checked in the XnPvcA-

XnPvcB reactions. The fractions which supported the PvcA reaction were frozen and lyophilized. Dried material was dissolved in 600 μL D2O for NMR analysis and was further purified by anion exchange chromatography.

Isolation of the small molecules by anion exchange chromatography

The active small molecules isolated from the size-exclusion column described above were further purified by anion exchange chromatography. Dried material

TM was dissolved in 2 mL deionized H2O and fractionated on 1 mL HiTrap Q HP anion exchange column. The column was pre-equilibrated and washed with deionized H2O, and a gradient of NaCl (50 mL) from 0 mM to 1 M was used to elute the molecules that bind to the column. 3 mL fractions were collected and 300

128

μL aliquots from each fraction were checked in the XnPvcA-XnPvcB reactions. The fractions which supported the PvcA reaction were lyophilized to dryness and further analyzed by NMR spectroscopy and mass spectrometry.

3.4.8 HPLC Analyses

HPLC was utilized to detect the product formation in the PvcA reaction and the PvcA-XnPvcB coupled reactions, as well as the metabolite produced by the cells overexpressing AhPvcA and XnPvcB proteins. For analysis, 10 µL samples were injected into a Shimadzu LC-10 Ai Series separations module equipped with a Shimadzu SPD-20A UV-vis detector and a 150 x 4.6 mm C18 reverse phase column (Sigma-Aldrich). The overall system was controlled using LC Solutions software. The UV detector was set at 280 or 310 nm. Compounds were separated by using a gradient method in which a mobile phase of H2O supplemented with

0.1% formic acid (Solvent A) and MeOH (Solvent B) was used: t = 0 min, 5%B; t =

5 min, 5%B; t = 30 min, 30%B; t = 35 min, 30%B; t = 45 min, 80%; t = 50 min,

80%B; t = 55 min, 5%B; t = 60 min, 5%B.

3.4.9 MS AnalysesC

All the samples were analyzed by direct-infusion positive-ion ESI on the LTQ

Orbitrap. Samples were prepared in 80% methanol/1.25% formic acid and were drawn into a 250 μL Hamilton syringe and infused at 3 μL/min. Selected ion

______

C MS analyses were performed by Brian Mooney

129 monitoring FTMS data were collected over 2 minutes of infusion. Data were examined in xcalibur.

130

CHAPTER 4

Thesis Summary and Future Directions

4.1 Thesis Summary

A remarkable variety of isonitrile-containing natural products have been found in both terrestrial and marine environments. Many of them are produced by bacterial pathogens as secondary metabolites. The gene clusters related to isonitrile biosynthesis have been identified in a lot of species, and they commonly encode an isonitrile synthase (IsnA or PvcA) and an Fe2+, α-ketoglutarate dependent oxygenase (IsnB or PvcB). Although functions of the PvcA and PvcB proteins have been assigned, the catalytic and chemical mechanisms of these novel enzymes remained uncharacterized.

4.1.1 Characterization of Three PvcB Homologous Proteins, the Non-Heme

Fe2+, α-Ketoglutarate Dependent Oxygenases

Chapter 2 described an investigation of the reactions catalyzed by three PvcB proteins from the homologous pathways in P. aeruginosa, E. amylovora and X. nematophila. They were hypothesized to carry out oxidative reactions on the tyrosine isonitrile and generate different products. By synthesizing the putative substrate tyrosine isonitrile and heterologously expressing the proteins in E. coli, we were able to reconstitute and characterize the in vitro activities of these three

PvcB proteins by UV-visible spectroscopy and HPLC. Product formed in each

131 reaction was identified by 1H-NMR spectroscopy and LC-mass spectrometry.

Steady-state and transient-state kinetic studies were performed to determine the mechanisms of these reactions. Substrate analogs were synthesized and examined in the reactions to look for potential enzyme inhibitors.

This study provides the in vitro data to support the functional assignment of

PvcB as a non-heme Fe2+, α-ketoglutarate dependent oxygenase to act on the putative substrate tyrosine isonitrile. In contrast to the PaPvcB and EaPvcB reactions which catalyzed desaturation of the substrate, the XnPvcB carries out an oxidative decarboxylation on tyrosine isonitrile. Further examination of three enzymes in the active sites revealed subtle differences between XnPvcB and

PaPvcB or EaPvcB. Although different products were made in the PaPvcB and

XnPvcB reactions, a common intermediate was proposed to be generated by both enzymes. Failure to trap this intermediate revealed its absence in the XnPvcB reaction. The PvcB proteins exhibited tolerance for the substrate analog 3-nitro tyrosine isonitrile, but were not able to convert the other analog 4-nitro phenylalanine isonitrile, which might be used in the future inhibition studies of the

PvcB reactions.

132

4.1.2 Characterization of the Reaction Catalyzed by PvcA, an Isonitrile

Synthase

Chapter 3 described a study of the PvcA protein, a novel isonitrile synthase proposed to convert the amino group of tyrosine or tryptophan into the isonitrile group. The in vitro activity assays were performed with the PvcA proteins which were from the homologous pathways in P. aeruginosa, X. nematophila and A. hydrophila and were heterologously expressed in E. coli. Purified proteins or the crude lysates were used as the enzyme source in the reactions with the putative substrates tyrosine and ribulose-5-phosphate, but no activities were found.

Expression of the XnPvcA protein in E. coli revealed the presence of the product in both the cell lysate and the culture medium. The product was further identified as tyrosine isonitrile, the predicted product of the PvcA reaction. Co-expression of the AhPvcA and XnPvcB proteins in E. coli resulted in isocyanovinylphenol production. In vivo feeding studies using 3,5-ring-D2 tyrosine and m-fluoro tyrosine in this pathway revealed the rapid incorporation of the tyrosine benzene ring into isocyanovinylphenol, demonstrating tyrosine to be one of the substrates in the

PvcA reaction. Labeling studies using 13C labeled glucose in the same pathway revealed the ability of both glucose C2 and C3 to be incorporated into the isonitrile group, implying that the source of the isonitrile carbon came from the pentose phosphate pathway. This unknown substrate, which provides the isonitrile carbon, was sought broadly in a phosphorylated metabolite library, the central metabolic

133 pathways and the metabolite pool of E. coli cells, and it was only found in the E. coli crude lysate. Fractionation of the cell lysate identified two molecules, one is a macromolecule and one is a small compound, that were both required for PvcA activity.

134

4.2 Future Directions

4.2.1 Mechanistic Studies of the PvcB Reactions

The remaining questions in regarding to the PvcB reactions in this investigation include:

A. Identification of the intermediate observed in short-time PaPvcB reactions by

mass spectrometry; determination of the reason for its absence in the XnPvcB

reaction.

B. Preparation of more substrate analogs to determine substrate tolerance for the

PvcB enzymes; identification of the substrate features that enable

transformations in the PvcB reactions.

C. Investigation of the mechanisms underlying different activities of the PaPvcB

and XnPvcB enzymes.

4.2.1.1 Determine the Intermediates in the PvcB Reactions

An intermediate was observed in HPLC spectra of the PaPvcB short-time reactions and needs to be determined by LC-mass spectrometry. In order to generate enough material for identification, several quenched samples will be combined. Time scales of the substrate consumption, the intermediate formation and decay, and the product generation in the PvcB reactions can be further explored by analyzing a series of samples quenched at different times. But in order to have accurate measurements, the species in the reactions should be stabilized.

One way to achieve this goal is to isolate the compounds from the reaction

135 mixtures by ethyl acetate extraction. HPLC methods can also be modified to allow shorter time of each analysis. Since the intermediate was not observed in the

XnPvcB reaction, a standard of this proposed intermediate can be synthesized and tested in both PvcB reactions to determine whether it is a common intermediate used by the PvcB enzymes.

4.2.1.2 Investigate the Substrate Tolerance for the PvcB Enzymes

3-nitro tyrosine isonitrile could be converted by the PvcB enzymes while 4- nitro phenylalanine failed. To better interpret this result, more tyrosine isonitrile analogs, with different substitutions on the benzene ring will be synthesized and tested in the PvcB reactions. The importance of the hydroxyl group on the benzene ring will be determined in this way. In addition, compounds which can serve as the substrates in the PvcB reactions could be tested for turnover efficiency, and compared to each other. This may provide insight into the structure and chemical mechanism of the PvcB enzyme.

4.2.1.3 Probe the Reason that Accounts for the Different Activities Exhibited by PvcB Homologs

The substrate docking and homology models generated to compare the active sites of PaPvcB, EaPvcB and XnPvcB proteins revealed subtle differences.

XnPvcB has an arginine residue (position 108) whereas PaPvcB and EaPvcB have a leucine (position 116). Arg168 of PaPvcB is conserved in EaPvcB but replaced by a Lysine in XnPvcB. The differences in the active sites may be an important

136 determinant of reaction specificity. To check this possibility, mutations on Leu116 or

Arg168 of PaPvcB can be made. The structures of the mutants will be examined to ensure the correct folding of the proteins, and the functions of the mutants will be analyzed biochemically. In addition, mutagenesis of other key residues (highly conserved) in the active sites of the PvcB proteins will allow us to gain a deeper understanding of the chemical mechanisms unitized by this family of enzymes.

Last, but not least, the structures of the PvcB enzymes could be compared to those of the IsnB homologs which use tryptophan isonitrile as the substrates.

137

4.2.2 Future Studies of the PvcA proteins

Although a lot of efforts were made to reconstitute the PvcA reaction in vitro, the results obtained so far indicate the substrate which provides the isonitrile carbon has not been identified successfully. The latest experiments revealed two molecules from the E. coli lysate, a macromolecule, which is probably a common metabolic enzyme, and a small molecule, possibly with a phosphorylated group are both essential to PvcA activity. The possible roles that these two molecules play in the PvcA reaction were proposed in Chapter 3, and need to be demonstrated next. Compared to the macromolecule identification, it’s more feasible to determine the small molecule first. In fact, we have already performed

MS analysis to identify the small molecule a couple of times, but no good results were obtained because of the complexity of the samples, the low concentration of the compound in the samples and the unpredictable behavior of this compound in

MS. Efforts should be made to further purify and enrich the small compound by size-exclusion and ion-exchange chromatography. Different solvents could be utilized to optimize the separation, but stability of the small compound under different conditions should be noted. The detection of the small compound is facilitated by the PvcA reaction, in which a decrease of the amount of this compound could be monitored. But the reaction is only working in the presence of the macromolecule. Therefore, to make the reaction system as simple as possible, further purification of the macromolecule is also required. Fractionation of the

138 macromolecule by ion-exchange chromatography resulted in the loss of the active macromolecule, for some unknown reasons. Thus, size-exclusion chromatography followed by ammonium sulfate precipitation could be used to get purer macromolecules. What’s more, the conditions of the PvcA reactions could be optimized. Effects of some factors including temperature, reaction time, buffer and the concentrations of the added substrates and enzymes on the final product formation should be determined. Overall, identifying the small molecule will provide insight into the real substrate of the PvcA reaction, and thus is the most important mission for now.

After the substrate for the PvcA enzyme is identified, in vitro reconstitution assays could be performed in the pure substrate-directed reactions with different

PvcA enzymes. The activities will be characterized by UV-visible spectroscopy.

Since the product of the PvcA reaction is unstable, the XnPvcB reaction will still be coupled to facilitate product detection. Following activity assays, transient kinetics will be used to determine the catalytic constants Km, kcat and Kcat/Km in different

PvcA reactions and the turnover efficiency of each PvcA enzyme will be compared.

Transient kinetics will be performed to detect and identify the intermediates in the reactions. In addition, mutagenesis studies and crystallography investigations would be performed to unravel the knowledge underlying the mechanism of the isonitrile group formation.

The reactions catalyzed by IsnA, WelI1 and AmbI1 utilize tryptophan and

139 ribulose-5-P as the substrates, while the substrates for our PvcA enzymes are tyrosine and an unknown compound. Thus, it is interesting to investigate the reason for the substrate specificity of these enzymes, probably through the crystallography studies.

140

References

1. Veiderma, M., Chemistry of the isocyanides and their multicomponent

reactions, including their libraries--the initiatives of Ivar Ugi. Estonian

Academy of Sciences: Chemistry, 2007. 56(2): pp. 98-103.

2. Wilson, R.M., et al., A Fascinating Journey into History: Exploration of the

World of Isonitriles En Route to Complex Amides. Angewandte Chemie

(International ed. in English), 2012. 51(12): pp. 2834-2848.

3. Ugi, I., B. Werner, and A. Dömling, The chemistry of isocyanides, their

multicomponent reactions and their libraries. Molecules, 2003. 8(1): pp. 53-

66.

4. Curran, D.P., S.B. Ko, and H. Josien, Cascade Radical Reactions of

Isonitriles: A Second ‐ Generation Synthesis of (20S) ‐ Camptothecin,

Topotecan, Irinotecan, and GI‐147211C. Angewandte Chemie International

Edition in English, 1996. 34(23‐24): pp. 2683-2684.

5. Ryu, I., N. Sonoda, and D.P. Curran, Tandem radical reactions of carbon

monoxide, isonitriles, and other reagent equivalents of the geminal radical

acceptor/radical precursor synthon. Chemical reviews, 1996. 96(1): pp.

177-194.

6. Dömling, A., Recent developments in isocyanide based multicomponent

reactions in applied chemistry. Chemical reviews, 2006. 106(1): pp. 17-89.

7. Dömling, A. and I. Ugi, Multicomponent reactions with isocyanides.

141

Angewandte Chemie International Edition, 2000. 39(18): pp. 3168-3210.

8. Koopmanschap, G., E. Ruijter, and R.V.A. Orru, Isocyanide-based

multicomponent reactions towards cyclic constrained peptidomimetics.

Beilstein Journal of Organic Chemistry, 2014. 10: pp. 544-598.

9. Wu, X., et al., Total synthesis of cyclosporine: access to N-methylated

peptides via isonitrile coupling reactions. Journal of the American Chemical

Society, 2010. 132(12): pp. 4098-4100.

10. Yuan, Y., et al., Preparation and reactions of N-thioformyl peptides from

amino thioacids and isonitriles. Tetrahedron letters, 2009. 50(20): pp. 2329-

2333.

11. Zhang, B. and A. Studer, Recent advances in the synthesis of nitrogen

heterocycles via radical cascade reactions using isonitriles as radical

acceptors. Chemical Society Reviews, 2015. 44(11): pp. 3505-3521.

12. Linderman, R.J., S. Binet, and S.R. Petrich, Enhanced diastereoselectivity

in the asymmetric Ugi reaction using a new “convertible” isonitrile. The

Journal of Organic Chemistry, 1999. 64(2): pp. 336-337.

13. Pirrung, M.C. and S. Ghorai, Versatile, fragrant, convertible isonitriles.

Journal of the American Chemical Society, 2006. 128(36): pp. 11772-11773.

14. Ugi, I., et al., Isonitrile Syntheses. Newer methods of preparative organic

chemistry, 1968. 4: pp. 37.

15. Santra, S., et al., CONVERTIBLE ISOCYANIDES: APPLICATION IN

142

SMALL MOLECULE SYNTHESIS, CARBOHYDRATE SYNTHESIS, AND

DRUG DISCOVERY. Domino and Intramolecular Rearrangement

Reactions as Advanced Synthetic Methods in Glycoscience, 2016: pp. 121.

16. Hagedorn, I. and H. Tonjes, Structural explanation of xanthocillin; a new

antibiotic. Die Pharmazie, 1957. 12(9): pp. 567.

17. Minale, L., R. Riccio, and G. Sodano, Acanthellin-1, an unique isonitrile

sesquiterpene from the sponge Acanthella acuta. Tetrahedron, 1974. 30(11):

pp. 1341-1343.

18. Scheuer, P.J., Isocyanides and cyanides as natural products. Accounts of

chemical research, 1992. 25(10): pp. 433-439.

19. Garson, M.J. and J.S. Simpson, Marine isocyanides and related natural

products–structure, biosynthesis and ecology. Natural product reports,

2004. 21(1): pp. 164-179.

20. Edenborough, M.S. and R.B. Herbert, Naturally occurring isocyanides.

Natural product reports, 1988. 5(3): pp. 229-245.

21. Ghisalberti, E. and K. Sivasithamparam, Antifungal antibiotics produced by

Trichoderma spp. Soil Biology and Biochemistry, 1991. 23(11): pp. 1011-

1020.

22. Gloer, J.B., et al., Structure of brassicicolin A: a novel isocyanide antibiotic

from the phylloplane fungus Alternaria brassicicola. The Journal of Organic

Chemistry, 1988. 53(16): pp. 3758-3761.

143

23. Evans, J., E.J. Napier, and P. Yates, Isolation of a new antibiotic from a

species of Pseudomonas. The Journal of antibiotics, 1976. 29(8): pp. 850-

852.

24. Crawford, J.M., et al., Small molecule perimeter defense in

entomopathogenic bacteria. Proceedings of the National Academy of

Sciences, 2012. 109(27): pp. 10821-10826.

25. Crawford, J.M., R. Kontnik, and J. Clardy, Regulating alternative lifestyles

in entomopathogenic bacteria. Current Biology, 2010. 20(1): pp. 69-74.

26. Fujiwara, A., et al., Fermentation, isolation and characterization of isonitrile

antibiotics. Agricultural and Biological Chemistry, 1982. 46(7): pp. 1803-

1809.

27. Clarke-Pearson, M.F. and S.F. Brady, Paerucumarin, a new metabolite

produced by the pvc gene cluster from Pseudomonas aeruginosa. Journal

of bacteriology, 2008. 190(20): pp. 6927-6930.

28. Pedras, M.S.C. and E.E. Yaya, The first isocyanide of plant origin expands

functional group diversity in cruciferous phytoalexins: synthesis, structure

and bioactivity of isocyalexin A. Organic & biomolecular chemistry, 2012.

10(18): pp. 3613-3616.

29. Avilés, E. and A.D. Rodríguez, Monamphilectine A, a potent antimalarial β-

lactam from marine sponge Hymeniacidon sp: Isolation, structure,

semisynthesis, and Bioactivity. Organic letters, 2010. 12(22): pp. 5290-

144

5293.

30. Akritopoulou-Zanze, I., Isocyanide-based multicomponent reactions in drug

discovery. Current opinion in chemical biology, 2008. 12(3): pp. 324-331.

31. Achenbach, H. and H. Grisebach, Zur biogenese des xanthocillins.

Zeitschrift für Naturforschung B, 1965. 20(2): pp. 137-140.

32. Herbert, R.B. and J. Mann, The incorporation of C 1 units in the biosynthesis

of tuberin and xanthocillin. Journal of the Chemical Society, Chemical

Communications, 1984(22): pp. 1474-1475.

33. Puar, M., et al., The biosynthesis of hazimicins: Possible origin of isonitrile

carbon. The Journal of antibiotics, 1985. 38(4): pp. 530-532.

34. Garson, M.J., Biosynthesis of the novel diterpene isonitrile

diisocyanoadociane by a marine sponge of the Amphimedon genus:

incorporation studies with sodium [14 C] cyanide and sodium [2-14 C]

acetate. Journal of the Chemical Society, Chemical Communications,

1986(1): pp. 35-36.

35. Bornemann, V., G.M. Patterson, and R.E. Moore, Isonitrile biosynthesis in

the cyanophyte Hapalosiphon fontinalis. Journal of the American Chemical

Society, 1988. 110(7): pp. 2339-2340.

36. Brady, S.F. and J. Clardy, Cloning and heterologous expression of

isocyanide biosynthetic genes from environmental DNA. Angewandte

Chemie International Edition, 2005. 44(43): pp. 7063-7065.

145

37. Brady, S.F. and J. Clardy, Systematic investigation of the Escherichia coli

metabolome for the biosynthetic origin of an isocyanide carbon atom.

Angewandte Chemie International Edition, 2005. 44(43): pp. 7045-7048.

38. Brady, S.F., et al., Natural products from isnA-containing biosynthetic gene

clusters recovered from the genomes of cultured and uncultured bacteria.

Journal of the American Chemical Society, 2007. 129(40): pp. 12102-12103.

39. Stintzi, A., et al., Novel pyoverdine biosynthesis gene (s) of Pseudomonas

aeruginosa PAO. Microbiology, 1996. 142(5): pp. 1181-1190.

40. Drake, E.J. and A.M. Gulick, Three-dimensional structures of Pseudomonas

aeruginosa PvcA and PvcB, two proteins involved in the synthesis of 2-

isocyano-6, 7-dihydroxycoumarin. Journal of molecular biology, 2008.

384(1): pp. 193-205.

41. Hindre, T., et al., Transcriptional profiling of Legionella pneumophila biofilm

cells and the influence of iron on biofilm formation. Microbiology, 2008.

154(1): pp. 30-41.

42. Qaisar, U., et al., The pvc operon regulates the expression of the

Pseudomonas aeruginosa fimbrial chaperone/usher pathway (Cup) genes.

PloS one, 2013. 8(4): pp. e62735.

43. Vallet, I., et al., The chaperone/usher pathways of Pseudomonas

aeruginosa: identification of fimbrial gene clusters (cup) and their

involvement in biofilm formation. Proceedings of the National Academy of

146

Sciences, 2001. 98(12): pp. 6911-6916.

44. Vallet, I., et al., Biofilm formation in Pseudomonas aeruginosa: fimbrial cup

gene clusters are controlled by the transcriptional regulator MvaT. Journal

of bacteriology, 2004. 186(9): pp. 2880-2890.

45. Hall-Stoodley, L., J.W. Costerton, and P. Stoodley, Bacterial biofilms: from

the natural environment to infectious diseases. Nature Reviews

Microbiology, 2004. 2(2): pp. 95-108.

46. Parsek, M.R. and P.K. Singh, Bacterial biofilms: an emerging link to disease

pathogenesis. Annual Reviews in Microbiology, 2003. 57(1): pp. 677-701.

47. Burmølle, M., et al., Biofilms in chronic infections–a matter of opportunity–

monospecies biofilms in multispecies infections. FEMS Immunology &

Medical Microbiology, 2010. 59(3): pp. 324-336.

48. Schierle, C.F., et al., Staphylococcal biofilms impair wound healing by

delaying reepithelialization in a murine cutaneous wound model. Wound

Repair and Regeneration, 2009. 17(3): pp. 354-359.

49. Guo, S. and L.A. DiPietro, Factors affecting wound healing. Journal of

dental research, 2010. 89(3): pp. 219-229.

50. Stintzi, A., et al., The pvc Gene Cluster ofPseudomonas aeruginosa: Role

in Synthesis of the Pyoverdine Chromophore and Regulation by PtxR and

PvdS. Journal of bacteriology, 1999. 181(13): pp. 4118-4124.

51. Hamood, A.N., et al., Isolation and characterization of a Pseudomonas

147

aeruginosa gene, ptxR, which positively regulates exotoxin A production.

Molecular microbiology, 1996. 21(1): pp. 97-110.

52. Carty, N., The effect of ptxR on the production of exotoxin A, exoenzyme U,

and the pyoverdine chromophore in the Pseudomonas aeruginosa strain

PA103. 2002.

53. Sperandio, V., C.C. Li, and J.B. Kaper, Quorum-sensing Escherichia coli

regulator A: a regulator of the LysR family involved in the regulation of the

locus of enterocyte effacement pathogenicity island in enterohemorrhagic

E. coli. Infection and immunity, 2002. 70(6): pp. 3085-3093.

54. González‐Santoyo, I. and A. C órdoba‐Aguilar, Phenoloxidase: a key

component of the insect immune system. Entomologia Experimentalis et

Applicata, 2012. 142(1): pp. 1-16.

55. Kanost, M.R., M.J. Gorman, and N. Beckage, Phenoloxidases in insect

immunity. Insect immunology, 2008. 1: pp. 69-96.

56. Takahashi, S., et al., Byelyankacin: A novel melanogenesis inhibitor

produced by Enterobacter sp. B20. Journal of Antibiotics, 2007. 60(11): pp.

717.

57. Chang, T.-S., Natural melanogenesis inhibitors acting through the down-

regulation of tyrosinase activity. Materials, 2012. 5(9): pp. 1661-1685.

58. Shima, S. and K. Ataka, Isocyanides inhibit [Fe]-hydrogenase with very high

affinity. FEBS Letters, 2011. 585(2): pp. 353-356.

148

59. Wright, A.D., et al., Inhibition of heme detoxification processes underlies the

antimalarial activity of terpene isonitrile compounds from marine sponges.

Journal of medicinal chemistry, 2001. 44(6): pp. 873-885.

60. Doan, N.T., P.R. Stewart, and G.D. Smith, Inhibition of bacterial RNA

polymerase by the cyanobacterial metabolites 12-epi-hapalindole E

isonitrile and calothrixin A. FEMS microbiology letters, 2001. 196(2): pp.

135-139.

61. Li, Z., et al., Longitudinal development of mucoid Pseudomonas aeruginosa

infection and lung disease progression in children with cystic fibrosis. Jama,

2005. 293(5): pp. 581-588.

62. Defez, C., et al., Risk factors for multidrug-resistant Pseudomonas

aeruginosa nosocomial infection. Journal of Hospital Infection, 2004. 57(3):

pp. 209-216.

63. Ferroni, A., et al., Outbreak of nosocomial urinary tract infections due to

Pseudomonas aeruginosa in a paediatric surgical unit associated with tap-

water contamination. Journal of Hospital Infection, 1998. 39(4): pp. 301-307.

64. Ramphal, R., M. McNiece, and F. Polack, Adherence of Pseudomonas

aeruginosa to the injured cornea: a step in the pathogenesis of corneal

infections. Annals of ophthalmology, 1981. 13(4): pp. 421-425.

65. Fergie, J.E., et al., Pseudomonas aeruginosa bacteremia in

immunocompromised children: analysis of factors associated with a poor

149

outcome. Clinical infectious diseases, 1994. 18(3): pp. 390-394.

66. Leiros, I., S. McSweeney, and E. Hough, The reaction mechanism of

phospholipase D from Streptomyces sp. strain PMF. Snapshots along the

reaction pathway reveal a pentacoordinate reaction intermediate and an

unexpected final product. Journal of molecular biology, 2004. 339(4): pp.

805-820.

67. Wise, E., et al., Homologous (β/α) 8-barrel enzymes that catalyze unrelated

reactions: orotidine 5'-monophosphate decarboxylase and 3-keto-l-

gulonate 6-phosphate decarboxylase. Biochemistry, 2002. 41(12): pp.

3861-3869.

68. Aik, W., et al., Role of the jelly-roll fold in substrate binding by 2-oxoglutarate

oxygenases. Current opinion in structural biology, 2012. 22(6): pp. 691-700.

69. Ozer, A. and R.K. Bruick, Non-heme dioxygenases: cellular sensors and

regulators jelly rolled into one? Nature chemical biology, 2007. 3(3): pp.

144-153.

70. Ryle, M.J. and R.P. Hausinger, Non-heme iron oxygenases. Current opinion

in chemical biology, 2002. 6(2): pp. 193-201.

71. Strieker, M., et al., Mechanistic and structural basis of stereospecific Cβ-

hydroxylation in calcium-dependent antibiotic, a daptomycin-type

lipopeptide. ACS chemical biology, 2007. 2(3): pp. 187-196.

72. Hausinger, R.P., Fe (II)/α-ketoglutarate-dependent hydroxylases and

150

related enzymes. Critical reviews in biochemistry and molecular biology,

2004. 39(1): pp. 21-68.

73. Wu, L.-F., S. Meng, and G.-L. Tang, Ferrous iron and α-ketoglutarate-

dependent dioxygenases in the biosynthesis of microbial natural products.

Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics, 2016.

1864(5): pp. 453-470.

74. Yin, X. and T.M. Zabriskie, VioC is a Non‐Heme Iron, α‐Ketoglutarate‐

Dependent Oxygenase that Catalyzes the Formation of 3S‐Hydroxy‐L‐

Arginine during Viomycin Biosynthesis. ChemBioChem, 2004. 5(9): pp.

1274-1277.

75. Schofield, C.J. and Z. Zhang, Structural and mechanistic studies on 2-

oxoglutarate-dependent oxygenases and related enzymes. Current opinion

in structural biology, 1999. 9(6): pp. 722-731.

76. Pihlajaniemi, T., R. Myllylä, and K.I. Kivirikko, Prolyl 4-hydroxylase and its

role in collagen synthesis. Journal of hepatology, 1991. 13: pp. S2-S7.

77. Falnes, P.Ø., R.F. Johansen, and E. Seeberg, AlkB-mediated oxidative

demethylation reverses DNA damage in Escherichia coli. Nature, 2002.

419(6903): pp. 178-182.

78. Braliou, G.G., et al., 2-Oxoglutarate-dependent oxygenases control

hepcidin gene expression. Journal of hepatology, 2008. 48(5): pp. 801-810.

79. Sleeman, M.C., et al., Biosynthesis of carbapenem antibiotics: new

151

carbapenam substrates for carbapenem synthase (CarC). ChemBioChem,

2004. 5(6): pp. 879-882.

80. Fukumori, F. and R. Hausinger, Purification and characterization of 2, 4-

dichlorophenoxyacetate/alpha-ketoglutarate dioxygenase. Journal of

Biological Chemistry, 1993. 268(32): pp. 24311-24317.

81. Eichhorn, E., et al., Characterization of α-ketoglutarate-dependent taurine

dioxygenase from Escherichia coli. Journal of Biological Chemistry, 1997.

272(37): pp. 23031-23036.

82. Mukherji, M., et al., The chemical biology of branched-chain lipid

metabolism. Progress in lipid research, 2003. 42(5): pp. 359-376.

83. Smiley, J.A., et al., Genes of the thymidine salvage pathway: thymine-7-

hydroxylase from a Rhodotorula glutinis cDNA library and iso-orotate

decarboxylase from Neurospora crassa. Biochimica et Biophysica Acta

(BBA)-General Subjects, 2005. 1723(1): pp. 256-264.

84. Hewitson, K., et al., Oxidation by 2-oxoglutarate oxygenases: non-haem

iron systems in catalysis and signalling. Philosophical Transactions of the

Royal Society of London A: Mathematical, Physical and Engineering

Sciences, 2005. 363(1829): pp. 807-828.

85. Briza, P., et al., Isolation of two developmentally regulated genes involved

in spore wall maturation in Saccharomyces cerevisiae. Genes &

Development, 1990. 4(10): pp. 1775-1789.

152

86. Lesage, G. and H. Bussey, Cell wall assembly in Saccharomyces cerevisiae.

Microbiology and Molecular Biology Reviews, 2006. 70(2): pp. 317-343.

87. Krebs, C., et al., Non-heme Fe(IV)-oxo intermediates. Accounts of Chemical

Research, 2007. 40(7): pp. 484-492.

88. Martinez, S. and R.P. Hausinger, Catalytic Mechanisms of Fe(II)- and 2-

Oxoglutarate-dependent Oxygenases. J Biol Chem, 2015. 290(34): pp.

20702-11.

89. Grzyska, P.K., et al., Steady-state and transient kinetic analyses of

taurine/alpha-ketoglutarate dioxygenase: Effects of oxygen concentration,

alternative sulfonates, and active-site variants on the fe(IV)-oxo

intermediate. Biochemistry, 2005. 44(10): pp. 3845-3855.

90. Decker, A. and E.I. Solomon, Dioxygen activation by copper, heme and non-

heme iron enzymes: comparison of electronic structures and reactivities.

Current opinion in chemical biology, 2005. 9(2): pp. 152-163.

91. Kovaleva, E.G. and J.D. Lipscomb, Versatility of biological non-heme Fe (II)

centers in oxygen activation reactions. Nature chemical biology, 2008. 4(3):

pp. 186-193.

92. Dong, Y., et al., Crystal Structure Analysis of a Synthetic Non‐Heme Diiron

O2 Adduct: Insight into the Mechanism of Oxygen Activation. Angewandte

Chemie International Edition in English, 1996. 35(6): pp. 618-620.

93. Knauer, S.H., et al., The Fe(II)/alpha-ketoglutarate-dependent taurine

153

dioxygenases from Pseudomonas putida and Escherichia coli are tetramers.

FEBS J, 2012. 279(5): pp. 816-31.

94. Ryle, M.J., R. Padmakumar, and R.P. Hausinger, Stopped-flow kinetic

analysis of Escherichia coli taurine/alpha-ketoglutarate dioxygenase:

Interactions with alpha-ketoglutarate, taurine, and oxygen. Biochemistry,

1999. 38(46): pp. 15278-15286.

95. Hoffart, L.M., et al., Direct spectroscopic detection of a CH-cleaving high-

spin Fe (IV) complex in a prolyl-4-hydroxylase. Proceedings of the National

Academy of Sciences, 2006. 103(40): pp. 14738-14743.

96. Yang, Z., et al., Characterization of LipL as a non-heme, Fe(II)-dependent

alpha-ketoglutarate:UMP dioxygenase that generates uridine-5'-aldehyde

during A-90289 biosynthesis. J Biol Chem, 2011. 286(10): pp. 7885-92.

97. Counts, D.F., G.J. Cardinale, and S. Udenfriend, Prolyl hydroxylase half

reaction: peptidyl prolyl-independent decarboxylation of alpha-ketoglutarate.

Proceedings of the National Academy of Sciences, 1978. 75(5): pp. 2145-

2149.

98. Peterman, B., Measurement of the dead time of a fluorescence stopped-

flow instrument. Analytical biochemistry, 1979. 93: pp. 442-444.

99. Malik, V., Microbial secondary metabolism. Trends in Biochemical Sciences,

1980. 5(3): pp. 68-72.

100. Bode, H.B., Entomopathogenic bacteria as a source of secondary

154

metabolites. Current opinion in chemical biology, 2009. 13(2): pp. 224-230.

101. Demain, A., Pharmaceutically active secondary metabolites of

microorganisms. Applied microbiology and biotechnology, 1999. 52(4): pp.

455-463.

102. Pietra, F., Secondary metabolites from marine microorganisms: bacteria,

protozoa, algae and fungi. Achievements and prospects. Natural product

reports, 1997. 14(5): pp. 453-464.

103. Carmichael, W., Cyanobacteria secondary metabolites—the cyanotoxins.

Journal of applied bacteriology, 1992. 72(6): pp. 445-459.

104. Briza, P., M. Eckerstorfer, and M. Breitenbach, The sporulation-specific

enzymes encoded by the DIT1 and DIT2 genes catalyze a two-step reaction

leading to a soluble LL-dityrosine-containing precursor of the yeast spore

wall. Proceedings of the National Academy of Sciences, 1994. 91(10): pp.

4524-4528.

105. Hillwig, M.L., Q. Zhu, and X. Liu, Biosynthesis of ambiguine indole alkaloids

in cyanobacterium Fischerella ambigua. ACS Chem Biol, 2014. 9(2): pp.

372-7.

106. Micallef, M.L., et al., Comparative analysis of hapalindole, ambiguine and

welwitindolinone gene clusters and reconstitution of indole-isonitrile

biosynthesis from cyanobacteria. Bmc Microbiology, 2014. 14.

107. Zhu, J., et al., Examining Reaction Specificity in PvcB, a Source of Diversity

155

in Isonitrile-Containing Natural Products. Biochemistry, 2015. 54(16): pp.

2659-2669.

108. Young, R.M., Synthetic and bioactivity studies of antiplasmodial and

antibacterial marine natural products. 2011.

109. Mo, S., et al., Antimicrobial ambiguine isonitriles from the cyanobacterium

Fischerella ambigua. Journal of natural products, 2009. 72(5): pp. 894-899.

110. König, G.M., A.D. Wright, and C.K. Angerhofer, Novel potent antimalarial

diterpene isocyanates, isothiocyanates, and isonitriles from the tropical

marine sponge Cymbastela hooperi. The Journal of Organic Chemistry,

1996. 61(10): pp. 3259-3267.

111. Faulhammer, D. and M. Famulok, Characterization and divalent metal-ion

dependence of in vitro selected deoxyribozymes which cleave DNA/RNA

chimeric oligonucleotides. Journal of molecular biology, 1997. 269(2): pp.

188-202.

112. Andreini, C., et al., Metal ions in biological catalysis: from enzyme

databases to general principles. JBIC Journal of Biological Inorganic

Chemistry, 2008. 13(8): pp. 1205-1218.

113. Vorholt, J.A., Cofactor-dependent pathways of formaldehyde oxidation in

methylotrophic bacteria. Archives of microbiology, 2002. 178(4): pp. 239-

249.

114. Bettendorff, L. and P. Wins, Thiamin diphosphate in biological chemistry:

156

new aspects of thiamin metabolism, especially triphosphate derivatives

acting other than as cofactors. FEBS journal, 2009. 276(11): pp. 2917-2925.

115. Eliot, A.C. and J.F. Kirsch, Pyridoxal phosphate enzymes: mechanistic,

structural, and evolutionary considerations. Annual review of biochemistry,

2004. 73(1): pp. 383-415.

116. Struck, A.W., et al., S ‐ Adenosyl ‐ Methionine ‐ Dependent

Methyltransferases: Highly Versatile Enzymes in Biocatalysis, Biosynthesis

and Other Biotechnological Applications. ChemBioChem, 2012. 13(18): pp.

2642-2655.

117. Ittiamornkul, K., et al., Promiscuous indolyl vinyl isonitrile synthases in the

biogenesis and diversification of hapalindole-type alkaloids. Chem. Sci.,

2015. 6(12): pp. 6836-6840.

118. Huang, H., et al., Panoramic view of a superfamily of phosphatases through

substrate profiling. Proceedings of the National Academy of Sciences, 2015.

112(16): pp. E1974-E1983.

157

VITA

Jing Zhu was born January 1st, 1989 in Changzhou, JiangSu, China. After completing high school at Changzhou Senior High School in 2007, Jing attended

East China Normal University in Shanghai, China. Four years later, she received a Bachelor of Arts Degree in the major of Biotechnology, in 2011. Jing came to the

United States and joined the Department of Biochemistry at University of Missouri-

Columbia in the Fall of 2011. Jing joined Dr. Tipton’s lab in March 2012. She received her PhD in Biochemistry in Dec 2016.

158