INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

u m T Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Artxjr, Ml 48106-1346 USA 800-521-0600

NOTE TO USERS

Page(s) missing in number only; text follows. Microfilmed as received.

110

This reproduction is the best copy available.

UMI

PEPTIDE DEFORMYLASE: CHARACTERIZATION

AND ANTIBACTERIAL DRUG DESIGN

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Bv

Yaoming Wei, M.S.

*****

The Ohio State University^

1999

Dissertation Committee:

Professor Dehua Pei, Adviser Approved by

Professor Ming-Daw Tsai Adviser Professor Russ Hille Ohio State Biochemistry Program UMX Number: 9941457

Copyright 1999 by Wei, Yaoming

All rights reserved.

UMI Microform 9941457 Copyright 1999, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 ABSTRACT

Peptide deformylase catalyzes the specific cleavage of formyl moiety from newly synthesized proteins in bacteria. It is highly selective for N-formyl peptides over N-acetyl peptides. To probe the mechanism of this high selectivety, more fluoroacetyl peptides were synthesized and characterized using a continuous deformylase assay developed previously. Kinetic constants of the substrates suggest that both steric and electronic effects play a role. In addition, factors contributing to the low Km of N-formyl-Met-Leu- p-nitroanilide were investigated.

Deformylase in vivo is capable of deformylating most of the polypeptides in bacterial cells bearing diverse N-terminal sequences. To systematically examine the sequence specificity of peptide deformylase, a peptide library that contained all possible

N-terminally formylated tetrapeptides had been constructed and screened in this lab. In this work, detailed kinetic studies were carried out with representative substrates selected from the library. The results confirmed that, although a broad-specificity enzyme, peptide deformylase cleaves formyl peptides at drastically different rates. It was found that formyl-Met-X-Z-Tyr (X = any amino acid except for aspartate and glumate; Z = lysine, arginine, tyrosine or phenylalanine) and formyl-Met-Phe-Tyr-(Phe/Tyr) peptides are optimal substrates.

ii Deformylation is a conserved and distinctive feature of protein biosynthesis in bacteria and deformylase is essential for bacterial survival. Its absence in eukaryotes makes it an attractive target for antibacterial chemotherapy. Rational design of deformylase inhibitors was undertaken by replacing scissile bond of high-affinity substrate with thiol, carboxylic and aldehyde functionalities. Competitive inhibitors were obtained among which the most potent one had a Ki of 81 nM. To discover inhibitors with different structural features, a combinatorial library of deformylase inhibitors was constructed on solid phase by split-pool synthesis. The structures of most active inhibitors were deduced by deconvolution.

Alternative to inhibitor design, a novel approach to antibacterial drug design was developed, in which a pharmacologically inactive prodrug was selectively processed by the bacterial deformylase to release a cytotoxic drug inside the bacterial pathogen. The prodrugs were a series of 5’-dipeptidyl derivatives of 5-fluoro-2’-deoxyuridine (FdU). In the presence of peptide deformylase, the formyl group at the N-termini of the dipeptides is removed, leading to an intramolecular cyclization and release of FdU. The broad substrate specificity of peptide deformylase allows for the use of unnatural amino acids in the prodrugs, which have improved in vivo stability. These compounds should be biologically inactive because 5'-acylation blocks the 5’-phosphorylation of FdU, a necessary step for FdU to become an active antimetabolite. Because the active drug is produced only inside the bacterial pathogen, these prodrugs should have low cytotoxicity to the host.

Ill Dedicated to my mother

IV ACKNOWLEDGMENTS

I would like to thank Dr. Dehua Pei for his instruction and guidance on my research in his group. I wish to express my appreciation to Drs. Ming-Daw Tsai, Russ

Hille and Smita Patel for serving on my advisory committee and for their help of my career in graduate school.

I am indebted to Ravi Rajagopalan and Ying Zhou for generously providing me with peptide deformylase they purified. I was blessed to have some fantastic guys as my colleagues over the years. 1 am grateful to Dr. Wen-Lian Wu, a talented organic chemist who taught me basic experimental skills in organic synthesis. I wish to thank Kirk Beebe,

Jinu Shin and Dr. Xiao-Chuan Guo, the kind and gentile associates who would never hesitate to extend their helps to others, for making work environment more pleasant and life enjoyable. 1 acknowledge all my lab mates for friendship, assistance and collaboration.

My special gratitude goes to my wife, Rang-Rong Ma for her love and encouragement that helped me endure and survive the hardship of graduate school.

Finally, support from Department of Chemistry, Ohio State Biochemistry Program and the Ohio State University is gratefully acknowledged. This work was supported by the Ohio State University and National Institutes of Health. VITA

November 25, 1968 ...... Bom, Hubei, P. R. China

1986-1992...... School of Medicine, School of Pharmaceutical Sciences, Beijing Medical University

1993-199 4...... University Fellow, The Ohio State University

1994-presen t ...... Graduate Teaching and Research Assistant, The Ohio State University

VI PUBLICATIONS

1. Y. Wei, S.F. Conrad, M.D. Lainnore, and P.T.P. Kaumaya. “De novo design and immunogenicity of a conformation dependent HTLV-I gp46 epitope.” Peptides: Chemistry, Structure and Biology. Mayflower Scientific Ltd. 1996, 565-566.

2. Y. Wei and D. Pei. “Continuous spectrophotometric assay of peptide deformylase.” Analytical Biochemistry 1997, 250, 29-34.

3. Y. Wei, Y. Yan, D. Pei, and B. Gong. “A photoactivated prodrug.” Bioorganic and Medicinal Chemistry Letters 1998, 2419-2422.

4. Y.J. Hu, Y. Wei, Y. Zhou, P.T.R. Rajagopalan and D. Pei. “Determination of substrate specificity for peptide deformylase through the screening of a combinatorial library.” Biochemistry 1999, 38, 643-650.

5. Y. Wei and D. Pei. “Activation of antibacterial prodrugs by peptide deformylase.” Submitted to Bioorganic and Medicinal Chemistry.

FIELDS OF STUDY

Major Field: Biochemistry

Vll TABLE OF CONTENTS

Page Abstract ...... ii

Dedication ...... iv

Acknowledgments ...... v

Vita...... vi

List o f Tables ...... x

List of Figures ...... xi

Chapters:

1 .General Introduction ...... l 1.1 Biological Function of Peptide Deformylase ...... 1 1.2 Research History ...... 3 1.3 Structure and Mechanism ...... 5 1.4 Sequence analysis ...... 9 1.5 Peptide Deformylase as Antibiotic Drug Target ...... 10

2. Substrate Specificity of Peptide Deformylase ...... 11 2.1 Introduction ...... 11 2.2 Experimental Procedures ...... 15 2.2.1 Materials...... 15 2.2.2 Analytical Methods ...... 15 2.2.3 Deformylase Assays ...... 16 2.2.4 Substrate preparation ...... 17 2.3 Results and Discussion ...... 21 2.3.1 Origin of Specificity for N-formyl Peptides ...... 21 2.3.2 Factors Contributing to High Affinity of f-Met-Leu-pNA ...... 25 2.3.3 Sequence Specificity of Peptide Deformylase ...... 26

3. Peptide Deformylase Inhibitors ...... 35 3.1 Introduction ...... 35 3.2 Experimental Procedures ...... 38

viii 3.2.1 Materials...... 38 3.2.2 Synthesis ...... 39 3.2.3 Construction of Inhibitor Library ...... 55 3.2.4 Screening of Inhibitor Library ...... 57 3.2.5 Inhibition Assay ...... 59 3.3 Results and Discussion ...... 59 3.3.1 Initial Design of Deformylase Inhibitors ...... 59 3.3.2 Aldehyde: A Transition State Analog as Deformylase Inhibition ...... 62 3.3.3 Sulfhydryl Deformylase Inhibitors ...... 67 3.3.4 Synthesis of 3-acetyIthio-2-alkylpropionic acids ...... 75 3.3.5 Design and Construction of Inhibitor Library ...... 77 3.3.6 Screening of Inhibitor Library ...... 82

4. Prodrug Activation by Peptide Deformylase ...... 90 4.1 Introduction ...... 90 4.2 Experimental Procedures ...... 92 4.2.1 Material and Methods ...... 92 4.2.2 Synthesis ...... 92 4.2.3 Kinetic and Biological Assays ...... 105 4.3 Results...... 108 4.3.1 Design and Synthesis of Prodrugs ...... 108 4.3.2 Deformylase Assay of Prodrugs ...... 116 4.3.3 Cyclization of Deformylated Prodrugs ...... 118 4.3.4 Chemical and Enzymatic Stability of Prodrugs ...... 122 4.3.5 Activation o f Prodrugs by E. coli Cell Lysate...... 124 4.3.6 Inhibition of Bacterial Cell Growth ...... 127 4.3.7 Studies of 3’-(f-Met-Pro)-FdU ...... 132 4.4 Discussion ...... 135

List of References ...... 138

Appendix: Proton NMR Spectra ...... 145

IX LIST OF TABLES

Table Page

2.1 Kinetic constants of acetyl peptides as deformylase substrates ...... 23

2.2 Kinetic constants of deformylase substrates with different sequences ...... 29

3.1 Constants of carboxyl and sulfhydryl deformylase inhibitors ...... 61

4.1 Summary of kinetic data of prodrug activation ...... 121 LIST OF FIGURES

Figure Page

1.1 Protein biosynthesis in bacteria ...... 2

1.2 Crystal structure of E. coli peptide deformylase ...... 7

1.3 Proposed catalytic mechanism for peptide deformylase ...... 8

2.1 Scheme of aminopeptidase-coupled deformylase assay ...... 12

2.2 Plot of deformylase reaction rate vs. substrate concentration ...... 24

2.3 Scheme of formate dehydrogenase-coupled deformylase assay ...... 27

2.4 Linear correlation between OD 36S and [formate] ...... 28

3.1 Carboxylic and sulfhydryl deformylase inhibitors ...... 60

3.2 Initial synthesis of an aldehydic deformylase inhibitor ...... 63

3.3 Conversion of bromide to hydroxyl functionality ...... 64

3.4 Revised synthesis of aldehydic deformylase inhibitor ...... 6 6

3.5 Synthesis of 3-acetylthio-2-(2-methylthioethyl)-propionic acid ...... 69

3.6 Final steps of sulfhydryl deformylase inhibitors ...... 70

3.7 Plot of deformylase reaction rate vs. substrate concentration in the presence of inhibitor ...... 73

3.8 A. Lineweaver-Burk plot of deformylase inhibition ...... 74 B. Plot of Km app vs. [I] ...... 74

3.9 Preparation of 2-alkyl-3-mercaptopropionic acids ...... 76

3.10 Scheme of construction of inhibitor library ...... 79

xi 3.11 Structures of P 3 ’ building blocks ...... 80

3.12 Deconvolution of P 3 ’ position of inhibitor library ...... 87

3.13 Deconvolution with 2 -aminoanthracene as P 3 ’ moiety ...... 8 8

3.14 Deconvolution with 1 -aminoanthracene as P 3 ’ moiety ...... 89

4.1 Prodrug activation by peptide deformylase ...... 91

4.2 Initial synthesis of f-Met-Pro-dU ...... I l l

4.3 Revised synthesis of f-Met-Pro-dU ...... 112

4.4 Attempted synthesis of tBoc-(a-methyl)Ala-dU ...... 113

4.5 Final synthetic scheme of prodrugs ...... 115

4.6 Plot of deformylation rate vs. prodrug concentration ...... 117

4.7 HPLC analysis of activation of prodrug 56a...... 119

4.8 First order kinetics of cyclization of deformylated prodrugs ...... 120

4.9 HPLC analysis of prodrug degradation ...... 123

4.10 HPLC analysis of activation of prodrug 56a by E. coli lysate...... 126

4.11 Inhibition of E. coli cell growth by prodrugs ...... 128

4.12 Inhibition of staphylococcus epidermidis cell growth by prodrugs ...... 129

4.13 Inhibition of £. coli pET22b-def cell growth by prodrugs ...... 130

4.14 Synthesis of 3’-(f-Met-Pro)-FdU ...... 134

XU CHAPTER 1

GENERAL INTRODUCTION

1.1 Biological Function of Peptide Deformylase

Methionine is the universal start signal for de novo protein synthesis in all organisms. Translation initiation and subsequent processing of N-termini of nascent polypeptides in prokaryotic cells involve a series of enzymatic actions (Figure 1,1): 1.

Methionyl-tRNA synthetase adds methionine to the hydroxyl group of the terminal adenosine in tRNAf^^^ (Heinrikson and Hartley, 1967). 2. Transformylase then transfers a formyl group from N'°-formyl-tetrahydrofolate to the amino group of methionine esterified to tRNAf^^^ (Adams & Capecchi, 1966; Dickerman et al., 1967; Kahn et al.,

1980; Blanquet et al., 1984). 3. Following translation initiation, peptide deformylase cleaves the N-formyl group from nascent peptides (Adams, 1968; Livingston & Leder,

1969; Takeda & Webster, 1968). 4. N-terminal methionine from certain deformylated proteins is then removed by methionine aminopeptidase (Ben-Bassat et al., 1987; Miller et al., 1987).

Translation machinery in prokaryotes, with only three translation initiation factors in the instance of E. coli, is simpler than that in eukaryotes. N-terminal formylation is thought to confer Met-tRNA'^^'^V its initiator identity by favoring its participation in the Methionine (Met) tRNA^et^ Aminoacyl-tRNA synthetase Met-tRNAMBtf Y N^°-formyltetrahydrofo(ate Transformylase N-formyl-Met-tRNA**®*^ I protein synthesis at ribosome

N-formyl-Met-AA^-AA2 AA"-tRNA

formate Peptide deformylase Met-AA^-AA2 AA"-tRNA Methionine Met aminopeptidase AA1-AA2 AA"-tRNA 1 AA^-AA2 AA" (mature protein)

Figure 1.1. Protein biosynthesis in bacteria. initiation process and by preventing its recognition by elongation apparatus (Varshney &

RajBhandary, 1992; Guillon et ai., 1993). Formylation of Met-tRNAf^®‘ was shown to be required for the initiation o f translation process (Bianchetti et al., 1977; Kung et al. 1979).

As a result, all nascent polypeptides synthesized in bacteria start with N- formylmethionene. Nonetheless mature proteins of bacteria were shown to have lost the

N-formyl group (Marcher & Sanger, 1964). As the function to remove N-terminal formyl group from polypeptides following translation initiation, peptide deformylase activity was first discovered in the late sixties. At least one function of deformylation is to free up the

N-terminus of peptides so that methionine aminppeptidase can further process N-terminal methionine, which may be required for proper protein folding and function.

1.2 Research History

Early characterization of peptide deformylase with crude cell lysate established that deformylase cleaves formyl group via a hydrolytic pathway, as both formate and deformylated peptides are the products of the enzymatic reaction (Adams, 1968). It was found that deformylase was quite labile under a variety of purification conditions and was very sensitive to EDTA and thiols such as 2-mercaptoethanoI and dithiothreitol (Adams,

1968; Livingston & Leder, 1969). Despite intense interest in its function during protein synthesis and maturation, peptide deformylase has eluded further purification and studies for decades due to its extraordinary instability.

Last several years has witnessed significant progress on the research of peptide deformylase since the gene was cloned and characterized from Escherichia coli (Meinnel & Blanquet, 1993; Mazel et al., 1994) and Thermus thermophilus (Meinnel & Blanquet,

1994). The cloned deformylase was then inserted to expression vectors for overproduction. In our and other labs, purification procedures have been developed to produce high yields of homogeneous enzyme (Meinnel & Blanquet, 1993; Rajagopalan et al., 1997). The availability of deformylase in large quantities has since accelerated the mechanistic studies of peptide deformylase.

Meinnel group carried out initial characterization of £. coli deformylase over­ expressed and found that the enzyme contained zinc metal and was sensitive to 1 , 1 0 - phenanthroline (Meinnel & Blanquet, 1995). Coupled with the occurrence of an HEXXH motif, where X stands for any amino acids, they classified deformylase as a member of zinc metallopeptidase super family. Their subsequent work with site-directed mutagenesis established that two histidines. His 132 and His 136 from the conserved motif and Cys90 were zinc ligands (Meinnel et al. 1995). The mutagenesis also identified the importance of Glul33, another residue from the conserved HEXXH motif, as the catalytic base.

The dispute over the identity of metal ion at enzyme active site ensued when deformylase purified by our group, which is several hundred fold more active than that reported for Zn^^ enzyme, was shown to contain iron by metal analysis (Rajagopalan et al., 1997). In contrast to the highly stable zinc protein, the purified Fe^^-containing deformylase, in agreement with the record of lability of this enzyme, remains highly unstable. Iron was therefore proposed to be the physiological metal of peptide deformylase. Although not ready to take this claim, Meinnel group later on conceded that

Zn^^ could not possibly be the metal to endow deformylase with an activity high enough to sustain ceil growth (Ragusa et al., 1998). However, they argued that deformylase containing Ni^"^ is also sufficiently active and would fulfill the duty required of this enzyme inside cells as well. Now it has been independently discovered by another group that iron is the native metal in deformylase and that can readily displace Fe^^ ion, rendering the enzyme inactive (Groche et al., 1998). Since zinc binds to the enzyme active site more tightly and could only be removed via protein unfolding, zinc found in deformylase preparations is probably an artifact and should be dismissed as physiologically relevant.

The decades-old mystery surrounding the extreme instability of deformylase was recently unraveled (Rajagopalan & Pei, 1999; Groche et al., 1998). It was demonstrated that enzyme inactivation was attributable to Fe^^ oxidation by omnipresent oxygen in aqueous solution, and that the stability could be much improved when oxygen or active oxygen species was depleted from deformylase reaction by another enzyme or a chemical reagent.

1.3 Structure and Mechanism

Structure of the stable Zn^^ deformylase has been determined by nuclear magnetic resonance spectroscopy and X-ray crystallography (Meinnel et al., 1996; Chan et al.,

1997). The full-length deformylase consists of a single domain of 168 amino acid residues. In an arrangement as shown in Figure 1.2, the protein contains three major a- helixes, three p-sheet regions and a 3-10 helix. The structure also revealed that the metal ion was tetrahedrally coordinated with His 132, His 136 and Cys90 and water molecule and supported the role of Glul33 as the catalytic base.

In addition to Zn^^, Fe^^, and deformylase described above, our group has prepared a Co^^-containing deformylase that is nearly fully active while being stable.

Structures of these deformylase variants have all been determined (Becker et al., 1998 a

& b; Dardel et al., 1998; Hao et al., 1999). Although catalytic efficiency strongly depended on the identity of the bound metal ion, there are no significant differences in three-dimensional structures of the enzymes. It was observed by Becker et al (1998b) that the tetrahedral volume spanned by four metal ligands around the zinc ion, due to its tighter binding, was smaller than in the case of Fe^"^ or They suggested that this subtle difference be held accountable for the inactivity of Zn'^-containing deformylase because of the metal ion’s reluctance to change from 4-coordination to 5-coordination state in the course of catalysis.

Based on the structure of free deformylase and deformylase complexed with inhibitor or reaction product, several overall similar versions of mechanism have been proposed (Hao et al., 1999; Becker et al., 1998b). Glu-133 deprotonates the a water molecule to generate hydroxide ion, a better nucleophile which then attacks the formyl carbon. The proton abstracted by Glu-133 is subsequently used to protonate the amide of the substrate, making it a better leaving group. Hao’s model suggested that the metal ion facilitates the generation of OH* instead of functioning as a Lewis acid to activate formamide bond, and that rather Leu-91 amide NH is the Lewis acid to help activate the formyl oxygen (Figure 1.3). 0 2

0 5

Figure 1.2. Crystal structure of E. coli peptide deformylase (reproduced from Chan et al., Biochemistry, 1997). V GIrvSOüirvî Glu-133 \ t>in-aGln-50 Glu-133 >=o--

Leu-91— ( NH

l-H

H H-N"^'^ly-4S X "''H " H-N^ - 'Gly-^S r OH île-44 -N lle-44 HzO

Gln-50 Glu-133 \ uin-5Gln-50 >=o. >=o-- Glu-133 / V a

' T'-V" ''H H-N ' ^ - ♦ 5

lla-44 lle-44

Figure 1 .3. Proposed catalytic mechanism for deformylase reaction. 1.4 Sequence Analysis

More deformylase genes have now been characterized either through genetic

screening taking advantage of conditional viability of the def mutant of E. coli (Maze! et

al., 1997), or through systematic bacterial genome sequencing (Fleischmann et al., 1995;

Fraser et al., 1995). A total of 33 deformylase sequences can now be retrieved from data

bank. To acquire information about structure and function relationship, alignment of deformylase sequences has been done with the available sequences at different times

(Meinnel et al., 1997; Mazel et al., 1997; Dardel et al., 1998). Enzymes from various bacterial origins display strong homologies to each other. Besides a few scattered conserved residues, there are three consensus stretches of amino acids including HEXXH, the signature motif of zinc metalloprotease, EGCLS where the third metal ligand resides, and GXGXAAXQ. When examined in the context of 3D structure, these conserved motifs build the active site around metal ion of deformylase. Site-directed mutagenesis confirmed the importance of these residues in maintaining the active and stable enzyme times (Meinnel et al., 1997; Dardel et al., 1998). As shown in the co-crystal structure of deformylase and inhibitor, these conserved residues are involved in interaction with the inhibitor (Hao et al., 1999). The fact that deformylase family adopts a constant tertiary structure and active site indicates that a competitive inhibitor of a given deformylase could be effective against others and have wide applicability. 1.5 Peptide Deformylase as Antibiotic Drug Target

Protein synthesis in cytoplasm of eukaryotic cell initiates with methionine as

opposed to N-formyl-methionine in prokaryotes and there is no such a process as

deformylation of polypeptides following translation initiation and prior to aminopeptidase

action (Kozak, 1983). On the other hand, mitochondrial protein synthesis in eukaryotic

cells does begin with formyl methionine (Bianchetti et al., 1971), however, deformylase

activity is absent from the organelle also and virtually all proteins synthesized inside

mitochondria retain the N-formyl methionyl moiety. Thus, deformylation of nascent

proteins is a conserved and distinctive feature of bacterial system.

As an essential activity for cell survival, deformylase activity is indispensable for

cell viability. It has been demonstrated that disruption of deformylase gene alone indeed

carried lethal effect to bacteria (Mazel et al, 1994). This conserved and distinctive

metabolic trait of bacteria can be exploited as target for antibacterial chemotherapy. On

one hand, nontoxic molecules that can be converted to toxic compounds by deformylase

action should specifically kill bacteria. On the other hand, deformylase inhibitors will be able to block protein processing and arrest bacterial growth.

The work described in this dissertation is an extension of the research activities summarized in my thesis for Master’s Degree. The work described in the Master’s thesis is not covered again in this dissertation but should be counted as part of my fulfillment for the Ph.D. Degree at the Ohio State University.

10 CHAPTER 2

SUBSTRATE SPECIFICITY OF PEPTIDE DEFORMYLASE

2.1 Introduction

One of the most intriguing features of peptide deformylase is the high specificity

for N-formyiated substrates. It was originally reported that it has no detectable activity toward N-acetylmethionyl peptides (Adams, 1968; Meinnel & Blanquet, 1995).

Previously, we developed a coupled assay using N-formyl-Met-Leu-p-nitroanilide (f-Met-

Leu-pNA) as the substrate, which, after being deformylated, is further hydrolyzed by aminopeptidase to release the chromophore p-nitroaniline (Wei & Pei, 1997). This sensitive, continuous assay was especially valuable for peptide deformylase, an unstable enzyme that mandates very short reaction time for accurate determination of deformylase reaction rates. The high sensitivity of this assay also allowed quantitative measurement of acylase activity of deformylase (Figure 2.1). Our results unambiguously showed that N- acetylmethionyl peptides were indeed deformylase substrates although the reaction rate was more than 4 orders of magnitude lower than that of the formylated counterpart. Steric or electronic effect has been proposed to be responsible for this exquisite selectivity. To gain insight into the origin of the preference of formyl peptides by deformylase, fluoroacetyl peptides were synthesized and tested as potential substrates (Figure 2.1).

11 o O

/)—NO, H O

peptide deformylase

O

HgN /V-NG, O

aminopeptidase

+ + H.N NO, OH OH

O O (Monitored at 405 nm)

1, R = H; 2, R = CH3 ; 3. R = FCHg; 4, R = FgCH; 5, R = F 3 C

Figure 2.1. Scheme of aminopeptidase-coupled deformylase assay.

1 2 The substrate f-Met-Leu-pNA developed for deformylase assay has been one of the best substrates characterized so far. Compared to all other N-formyl peptides that often do not show saturation, its relatively low Km (20 jrM) is quite unusual. To investigate the factors that contributed to the low Km, analogs of f-Met-Leu-pNA with different sequences were synthesized and kinetic studies carried out.

The other interesting feature of deformylase is its sequence specificity. As the essential activity to hydrolyze formyl group fi'om proteins which bear a wide variety of N- terminal sequences in bacterial cells, peptide deformylase has evolved as an enzyme with broad sequence specificity. However, not all N-formyl groups are removed from polypeptides following translation and there have been observations that some proteins remain formylated in their functional states. Examples include several Escherichia coli ribosomal proteins (Hauschild-Rogat, 1968), several E. coli chemotactic peptides

(Marasco et al., 1984), and Salmonella typhimurium aspartate chemoreceptor (Milligan et al., 1990). There is also another group of proteins which normally undergo complete deformylation but become partially formylated when overproduced in E. coli cells

(Warren et al., 1996). Increased production of peptide deformylase has been employed to achieve complete formylation of these proteins (Warren et al., 1996). The question arises whether the state of formylation is a result of the substrate specificity of peptide deformylase. Previously, kinetic characterization of peptide deformylase has been carried out using N-formylated peptide substrates (Adams, 1968; Meinnel & Blanquet, 1995;

Rajagopalan et al., 1997). It was found that deformylase does not have stringent requirement beyond N-formylmethione in substrate sequences and that most of the N-

13 formylmethionyl peptides tested so far are deformylated by peptide deformylase at comparable rates. While these studies have demonstrated some level of substrate specificity, they are practically limited by the number of peptides that can be synthesized and tested.

To thoroughly study the substrate specificity with substrates of all possible sequences, a deformylase substrate library was synthesized and screened by Dr. Yun-Jin

Hu. A consensus sequence preferred by deformylase was concluded to be f-MX(F/Y)Y

(X = any amino acid except for Asp and Glu) by analysis of the frequency of occurrence of amino acids at each given position. Specifically, at the Pi’ site, methionine

(norleucine) was the most preferred amino acid, followed by glycine, histidine, and phenylalanine or tyrosine. Little selectivity was loimd at the P 2 ’ site. All amino acids, with the exception of glycine and the negatively charged aspartate and glutamate, appeared at similar frequencies (with slightly higher frequency for tyrosine). The P 3 ’ site was highly selective for an aromatic residue. The P 4 ’ position strongly preferred a tyrosine residue (Hu et al., 1999). The sequences selected from the library were grouped based on the identity of their N-terminal residues (Hu et al., 1999). In this chapter, detailed kinetic studies of representative substrates were conducted to confirm the results of library screening and to gain further insight to deformylase-substrate interaction. The knowledge will help shed light on the environment of deformylase active site and aid rational design of deformylase inhibitors.

14 2.2 Experimental Procedures

2.2.1 Materials

Escherichia coli peptidase deformylase was overexpressed in E. Coli and purified

to apparent homogeneity in this laboratory by Rajagopalan and Zhou as described

(Rajagopalan & Pei, 1998). Aeromonas proteolytica aminopeptidase and Pseudomonas

oxalaticas formate dehydrogenase were purchased from Sigma Chemical Company (St.

Louis, MO). All other chemicals were purchased from either Sigma Chemical Company

or Aldrich Chemical Company (Milwaukee, WI).

2.2.2 Analytical Methods

'H NMR spectra of synthetic compounds were recorded on Bruker NMR 200, 250

or 300 MHz spectrometers. The internal standard for *H-NMR spectra determined in

CDCI3 was tetramethylsilane; for those determined in DMSG and D 2O, the solvent was

used as the reference at 5 2.49 and 4.63 respectively. Chemical shifts (5) were reported in

the unit of parts per million (ppm) downfield from tetramethylsilane. Proton NMR

splitting patterns were designated as: s, singlet; d, doublet; t, triplet; q, quartet; m,

multiplet and all coupling constants are given in hertz (Hz). Mass spectrometric analysis was performed at Campus Chemical Instrument Center of the Ohio State University.

15 2.23 Deformylase Assays

Method A, continuous deformylase assay. The deformylase reaction is coupled to an aminopeptidase, which further processes the deacylated peptides to release a chromogenic product (Wei & Pei, 1997). This method was used for all of the N- acetylated peptides. All assays were carried out at room temperature in 1 ml volume in polystyrene cuvettes. The final concentrations of buffer were 50 mM sodium phosphate

(pH 7.0) and 100 pM ethylene glycol bis(J5-aniinoethyl ether)-N,N’-tetraacetic acid

(EGTA). The reaction with various substrate concentrations and 0.8 U of aminopeptidase were initiated by the addition of deformylase and monitored at 405 nm in a Perkin-Elmer

13 UV/VIS spectrophotometer. The initial rates were calculated from the slope of the early part of the reaction progression curves. The absorbance was converted to concentration using extinction coefficient of p-NA (s = 1.06 x 1 O'* M'*cm'').

Method B, which monitors the release of formate by a formate dehydrogenase as previously described (Lazennec & Meinnel, 1997; Rajagopalan et al., 1997), was used for most of the N-formylated substrates. A standard line was generated as follows: 20 to 200 pM of sodium formate was oxidized with 0.25 U of Pseudomonas oxalaticas formate dehydrogenase in the presence of 1 mM NAD\ 50 mM sodium phosphate (pH 7.0) in a total volume of 1 ml. The reactions were incubated at 0 °C for 25 min. The absorbance of each reaction at 365 nm was plotted against the concentration of sodium formate.

Deformylase reactions in a volume of 0.5 ml with 0 — 2 mM formylated peptides and various amount of Fe^^ deformylase was allowed to proceed for 30 sec to achieve a

16 <20% product conversion. After the reaction was quenched by addition of 10 pi of 10

mM of ethanedithiol, 10 pi of 100 mM of NAD^ and 0.25 U of formate dehydrogenase

were added. The total volume was then adjusted to 1 ml and the reaction was incubated

at 0 °C for 20 - 25 min before OD 365 was measured. The amount of formate released by

deformylase reaction was calculated.

Method C was used for peptides that are poor substrates of the deformylase.

Typically, deformylase reaction in a volume of 100 pi in an eppendorf tube was allowed

to proceed to achieve ~10% product conversion. The reaction was terminated by addition

of an equal volume of 100 mM TFA or 10 mM H 2O2, and the reaction mixture was

analyzed by reversed-phase HPLC (monitored at 214 or 315 nm). Integration of the

product peak and the remaining substrate peak gave the percentage conversion, from

which the initial rates were calculated.

2.2.4 Substrate Preparation

t-Boc-Methlonylleucyl-p-nitroanillde 6().

To t-Boc-methionine (0.50 g,

^ 2 . 0 mmole) and leucyl-p-nitroanilide in

H II = H 2 0 ml of dichloromethane was added

DCC (0.42 g, 2.0 mmole). The reaction was kept at room temperature for 1 h and dicyclohexylurea precipitate was filtered out.

17 The filtrate was washed with 5% NaHCOs (20 ml x 3), 2% HCl (20 ml x 3), water (20 ml) and saturated NaCl solution (20 ml). Evaporation of evaporation gave 0.97 g of product (quantitative yield). ‘H NMR (200 MHz, CDCI 3): ô 9.24 (s, IH), 8.15 (d, J = 9.2

Hz, 2H), 7.79 (d, J = 9.1 Hz, 2H), 6.80 (d, J = 7.7 Hz, IH), 5.38 (d, J = 5.2 Hz, IH), 4.47

- 4.68 (m, IH), 4.12-4.35 (m, IH), 2.60 (t, J = 6.9 Hz, 2H), 2.10 (s, 3H), 1.45-2.25 (m,

5H), 1.43 (s, 9H), 0.97 (d, J = 6.4 Hz, 3H), 0.92 (d, 3H, 6.3 Hz, 3H).

N-(FluoroacetyI)-Met-Leu-p-nitroaniiide (3).

g'' t-Boc-Met-Leu-pNA (0.100 g,

9 y 9 0.207 mmole ) was dissolved in acetic FH H H i H acid / anisole / 38% HCl (2 ml / 1 ml / 0.3

ml). After 30 min at room temperature, the reaction was evaporated and triturated with diethyl ether (10 ml x 3) to give white solid

(0.060g, 0.143 mmole) in 69% yield. The HCl salt of Met-Leu-pNA (63 mg, 0.15 mmol) and sodium monofluoroacetate (15 mg, 0.15 mmol) were suspended in 1 0 ml of CH 2 CI2, and then dicyclohexylcarbodiimide (30 mg, 0.15 mmol) was added. The mixture was stirred for 1 2 h at room temperature. Dicyclohexylurea was then removed by filtration, and the filtrate was washed with 5% sodium carbonate (3x10 ml), 5% HCl (3x10 ml), water (10 ml), and saturated NaCl solution (10 ml). Evaporation of the solvent afforded

65 mg of a white solid in quantitative yield. 'H NMR (200 MHz, DMSO): Ô 10.64 (s,

IH), 8.33 (d, J = 7.5 Hz, IH), 8.12-8.29 (m, 3H), 7.85 (d, J = 9.3 Hz, 2H), 4.83 (d, / =

18 47.0 Hz, 2H), 4.35-4.55 (m, 2H), 2.45 (t, 7.9 Hz, 2H), 2.04 (s, 3H), 1.42-2.00 (m,

5H), 0.86-0.94 (m, 6 H).

Propiolyl-Met-Leu-p-nitroanilide (8 ).

To propioiic acid (0.040 g, 0.57

mmole) and Met-Leu-pNA (0.20 g, 0.52

H mmole) in 5 ml of dichloromethane, DCC o ^ I was added. After 2 h at room temperature, white dicyclohexylurea precipitate was removed by filtration. The filtrate was concentrated in vacuo, residue redissolved in 2 0 ml of ethyl acetate, and insoluble filtered out. The ethyl acetate was extracted with 5% NaHCO], (20 ml x 3), 2% HCl (20 ml x 3) and dried over MgS 0 4 . Evaporation of solvent afforded 0.088 g of white solid in 37 % yield. ‘H NMR (200 MHz, DMSO); 5 10.64 (s, IH), 8.94 (d, J = 7.5 Hz, IH), 8.30 ( d, J =

7.3 Hz, IH), 8.22 (d, J = 9.2 Hz, 2H), 7.84 (d, J = 9.3 Hz, 2H), 4.25-4.60 (m, 2H), 4.18 (s,

IH), 2.44 (t, J = 7.8 Hz, 2H), 2.20 (s, 3H), 1.40-2.00 (m, 5H), 0.91 (d, J = 6.7 Hz, 3H0,

0.87 (d, J = 6.9 Hz, 3H).

N-Formyl-Met-Leu-anilide (9).

3 ^ To tBoc-Leucine monohydrate (0.498 g, 2.0

mmole) and aniline (0.279 g, 3.0 mmole) in 10 ml of

dichloromethane was added DCC (0.412 g, 2.0

mmole). The reaction was stirred overnight and

19 dicyclohexylurea removed by filtration. The filtrate was diluted with 20 ml of

dichloromethane and washed twice with 5% HCl, twice with 5% NaHCOs. Evaporation

of solvent gave 0 . 8 6 g of white solid (quantitative yield).

The above product was dissolved in TFA / phenol / H 2 O / ethanedithiol /

thioanisole (16.5 ml/lml/lml/ 0.5 ml / 1 ml). After 1 h at rt, solvent was evaporated

in vacuo and the residue suspended in 30 ml of 5% HCl. After extraction with ethyl

acetate (30 ml x 3), the aqueous phase was adjusted to basic pH with NaiCOs and extracted with 30 ml of ethyl acetate. Evaporation of solvent gave 0.263 g of colorless oil

(59% yield).

To Leu-anilide (0.263 g, 1.17 mmole) and formyl-Met (0.208 g, 1.17 mmole) in 5 ml dichloromethane was added DCC (0.266 g, 1.29 mmole) was added. After 6 h at rt, dicyclohexylurea was removed by filtration. The filtrate was extracted with 5% NaHCOs and 5% HCl. Evaporation of dichloromethane gave 0.15 g of white solid (34 % yield). *H

NMR (300 MHz, DMSO): 5 9.95 (s, IH), 8.15-8.52 (m, 2H), 8.03 (s, IH), 6.59-7.70 (m,

5H), 4.30-4.55 (m, 2H), 2.25-2.55 (m, 2H), 2.02 (s, 3H), 1.40-2.00 (m, 5H), 0.75-0.95 (m,

6 H). Note: extensive racemization of methionine was observed in the last step of coupling.

N-Formyl-Met-Leu-amide (10).

3 ^ N-Formeyl-Met (0.088g, 0.50 mmole) was

dissolved in 6 ml of tetrahydrofuran. Leucinamide

MH 2 hydrochloric acid (0.091 g, 0.55 mmole), triethylamine

20 (0.060 g, 0.60 mmole) and DCC (0.113 g, 0.55 mmole) were added in this order with stirring and the suspension was left overnight. Precipitate was removed by filtration and the filtrate evaporated. Ethyl acetate was added to the residue upon which the product crystallized as a white solid. The product was collected by filtration and washed with a small amount of ethyl acetate. An amount of 40 mg was obtained (28% yield). NMR

(200 MHz, DMSO): 5 8.30 (d, J = 7.7 Hz, IH), 8.00 (s, IH), 7.98 (d, J = 7.8 Hz, IH),

7.28 (s, IH), 6.98 (s, IH), 4.30-4.50 (m, IH), 4.07-4.30 (m, IH), 2.42 (t, J = 8.0 Hz, 2H),

2.02 (s, 3H), 1.40-1.95 (m, 5H), 0.86 (d, J = 6.4 Hz, 3H), 0.82 (d, J = 6.3 Hz, 3H).

2.3 Results and Discussion

2.3.1 Origin of Specificity for N-formyi Peptides

Peptide deformylase is highly specific for formyl group (Adams 1968). Using a sensitive continuous deformylase, we were able to accurately determine the kcai/K^ value for N-acetyl-Met-Leu-p-NA. It turned out to be 41 M"'s"' and was more than 4 orders of magnitude lower than that of formyl substrate (Wei & Pei, 1997). This high selectivity could be due to either steric hindrance of the larger acetyl group or the inherently lower electrophilicity of acetamide compared to formamide. To determine which factor or whether both factors are causing the observed rate difference, substrate analogs with acetyl groups substituted with electron-withdrawing fluorine atoms were synthesized, deacetylation rate of these compounds by E. coli deformylase was studied using this continuous assay. The kcat/KM value for N-trifluoroacetyl-Met-Leu-p-nitroanilide, 136 M'

2 1 ‘s'* is on the same order of magnitude as that of N-acetyl substrate, but 10"* fold lower

than the formyl substrate, respectively. Although the exact Km could not be obtained

because of the limit imposed by the solubility, neither acetyl nor trifluoroacetyl substrates

showed saturation kinetics within the concentration range measured (40 pM). The

increased Km values for both substrates compared to the N-formyl compound suggest that

steric effect plays an important role in the substrate selectivity of deformylase. The acetyl

groups may be too bulky to fit into the enzyme active site and dislodge the carbonyl group

firom a proper position to allow an efficient nucleophilic attack by the enzyme. This may also be the reason why propiolyl-Met-Leu-pNA was not a deformylase substrate. It is

likely that the large size of the propiolyl group forbids enzyme binding even though the alkynyl group attached to the carbonyl carbon is electron withdrawing.

While peptide deformylase had very poor activity toward Ac-Met-Leu-pNA, substitution of fluorine for the -H ’s increased that activity significantly with the kcat/KM values of mono fluoroacetyl and difluoroacetyl peptides are 15, and 180 fold higher respectively (Table 2.1). The fact that substitution of one and two fluorines progressively increased the catalytic rate indicates that the lower rate for acetyl peptides is due, at least in part, to the lower electrophilicity of the acetyl group.

While the acetyl- and trifluoroacteyl- substrates showed no sign of saturation, interestingly, the monofluoroacetyl and difluoroacetyl substrates can saturated the enzyme at relatively low concentrations. The decrease in the Km value upon mono- or difluorination is probably due to the introduction of specific interactions between the fluorine atoms and the protein (e.g., H bonding). However, trifluoromethyl group is

2 2 significantly larger in size than a methyl group (White & Coville, 1994; Datta &

Majumdar, 1991), and the size may become a more predominant factor in the efficiency of catalytic hydrolysis by deformylase. The large size of trifluoroacetyl group may preclude proper enzyme binding required for efficient reaction, and explain the lack of saturation kinetics and decreased kcV Km value for trifluoroacetyl peptide.

Substrate Km (liM) kcat(s') kcat/KM (M''s-‘)

1 20.3 ± 1.3 37.8 ± 1.8 1.87 X 10^

2 ------—— 42

3 25.0 ± 2.4 0.016 ± 0 . 0 0 1 640 4 3.9 ± 0.3 0.3 ±0.01 7.8 X 10"^ 5 ---- —— 136

Table 2.1. Kinetic constants of acetyl peptides as deformylase substrates.

23 If 20“ "oj ■5 15 £

> - FAc-Met-Leu-pNA

20 40 50 [s], mM

0.8

3 ) 0.6

Io E 0.4 c COo X 0.2 > F Ac-Met-Leu-pNA 20

Figure 2.2. Plot of deformylase reaction rate vs. substrate concentration.

24 23.2 Factors Contributing to High Affinity of f-Met-Leu-pNA

It was a surprise when f-Met-Leu-pNA, developed for aminopeptidase-coupled assay was found to have a low Km- Considering the non-saturation kinetics of many natural deformylase substrates, further studies are warranted to dissect the factors that might have contributed to the high binding affinity of this molecule. The information may shed light on the general structural features favored by deformylase active site and help deformylase inhibitor design.

To investigate the role of Leu-pNA moiety in enzyme binding, structural variants in this part of f-Met-Leu-pNA molecule were synthesized and assayed. Removal of the C- terminal pNA group reduces the enzyme’s affinity to f-Met-Leu-NHa by -40 fold, indicating p-nitroanilide group at the C-terminal end is clearly critical for binding (Wei &

Pei, 1997). Next, to find out if the nitro group on the benzene ring is dispensable, f-Met-

Leu-anilide was synthesized. Kinetic assay was carried out by Ying Zhou with formate dehydrogenase assay and Km value was shown to be >400 pM. This indicated that the nitro- group is essential in increasing the binding affinity of the substrate, probably by engaging in hydrogen bonding to enzyme active site resides. In addition, introduction of carboxyl group onto the benzene ring was previously shown to decrease binding affinity by 20 fold (Wei, 1997). All the above evidence proves the importance of p-nitroanilide moiety in favorable enzyme binding, although it had not been meant for this purpose, but rather as a chromophore designed for a coupled deformylase assay in the first place.

An analog of f-Met-Leu-pNA with the leucine left out, N-formyl-Met-p- nitroanilide was also synthesized (Wei, 1997) and characterized with a discontinuous

25 An analog of f-Met-Leu-pNA with the leucine left out, N-formyl-Met-p-

nitroanilide was also synthesized (Wei, 1997) and characterized with a discontinuous

aminopeptidase-coupled assay. It showed non-saturation kinetics as well as drastically reduced kcat/Kvi value. It seems that the presence of the second residue following methionine is also important for favored enzyme recognition, although the specific identity of the residue may not be important since a variety of amino acids can be tolerated at this position.

23.3 Sequence Specificity of Peptide Deformylase

Deformylase reaction rates were determined by coupling deformylase reaction to formate dehydrogensase from Pseudomonas oxalaticus (Figure 2.3). This enzyme is able to oxidize formate to CO 2, and simultaneously reduces one equivalent amount of NAD^ to NADH, which is accompanied by an increase in absorbance at 365 nm. Figure 2.4 shows a standard line generated using OD 365 of FDH reaction solution containing various known concentrations of sodium formate (slope = 2.96 ODs^s/mM formate). The amount of formate produced by FDH-coupled deformylase reaction can then be determined by comparing the absorbance at 365 nm with this standard line.

It should be noted that inconsistency in the quality of the commercial FDH added difficulty to this assay. Various conditions have been tried out and the optimized procedure was as described earlier in this chapter, which may have to be adapted further for other batches of FDH with different qualities. Typically, FDH reaction was performed at 0 °C to avoid an unknown white precipitate that would form at room temperature. A

2 6 reaction time of 20 - 30 minutes is normally sufficient for complete FDH reaction.

Reaction longer than that is not recommended because of the NADH-consuming oxidase

impurities in commercial FDH. This assay was used to characterize most of the

substrates. For N-formyl peptides with extremely low activity, this assay became not

practical and HPLC analysis instead was used.

HgO O NAD+ V^formylase V FDH N—peptide A N HCOOH ------^ CO H A ' NADH (monitored at 365 nm) N—peptide

O

Figure 2.3. Scheme of formate dehydrogenase-coupled deformylase assay.

27 0.6

0.5

0.4

o * 0.3 O 0.2

0 50 100 150 200 [formate], |iM

Figure 2.4. Linear correlation between OD 365 and [formate].

2 8 entry substrate kctCs') Km (M) kcat/KM (xlO^M'' s'")

1 f-MAFYBR 1030 ± 100 950± 150 108

2 f-MAYYBR - - 99

3 f-MNYYBR - - 32 4 f-FYFHBR 204 ± 12 490 ± 60 42 5 f-YPMYBR 1.7 ±0.1 1070± 100 0.16

6 f-HVYYBR 0.26 ± 0 . 0 1 1390± 130 0.019 7 f-GAFYBR 0.51 ± 0.08 3300 ± 650 0.015

8 f-IAYYBR 1 1 ± 1 2420 ± 270 0.44 9 f-KAYYBR 0.014 ±0.001 560± 150 0.0024

1 0 f-RRYABBRM -- ND

1 1 f-MNYLBBRM 136± 18 920 ± 180 15

1 2 f-MDYLBBRM - > 2 0 0 0 2 . 6 13 f-MAKYBR 580 ± 56 320 ± 90 180

14 f-MARYBR - - 150 15 f-MARLBR 360 ± 9 404 ± 26 89

16 f-MAYKBR 680 ± 34 780 ± 70 8 6 17 f-MAYRBR 280 ± 15 310 ± 54 90

18 f-MAYLBBRM - - 16 19 f-MAFLBBRM - - 16

2 0 f-HVYLBBRM 0.058 ± 0.003 470 ± 36 0 . 0 1 2

2 1 f-GAFLBBRM 0.051 ±0.003 1420 ± 150 0.0036

22 f-FYFSBBRM - - 1 0

23 f-MDDDBR - - 0.0090

Table 2.2. Kinetic constants of deformylase substrates with different sequences. M, methionine; B, alanine; -, limited solubility prevented determination of kcm and ATm values; ND, no detectable activity.

29 Kinetic assays with individual sequences representative of the five classes (Table

entries 1-7) was conducted. As expected, class I peptides, which represent the majority of

the selected peptides and have norleucine as the N-terminal residue, are all excellent

substrates. The kcat/Kw values are 1.08 x 9.9 x 10^, and 3.2 x IQ^ M"' s'*,

respectively, for f-MAFY, f-MAYY and f-MNYY. This is consistent with that

methionine is the N-terminal residue of all nascent polypeptides, the physiological

substrates of peptide deformylase. Earlier results based on a small set of peptides that

have also shown that methionine and norleucine are the preferred amino acids at the N-

terminus among the amino acids tested (Adams & Capecchi, 1966; Rajagopalan et al.,

1997). These represent some of the best peptide substrates so far identified, and their

activities are 2-3 orders of magnitude higher than those of some randomly chosen N-

formylmethionyl peptides, which generally have kcat/KM values in the range of 1 0 ^-1 0 '^ NT

’s'* (Rajagopalan et al., 1997). A class II peptide (f-FYFH) with an N-terminal

phenylalanine, which is structurally distinct from methionine, is also an excellent

substrate (kcat/KM = 4.2 x lO^ M’’ s'*). Y-Formyltyrosyl peptides, also selected from the

library, are significantly poorer substrates (kcat/KM = 1.6 x 10^ M'* s'* for f-YPMY). The class III and IV peptides that contain N-terminal glycine or histidine (e.g., f-GAFY and f-

HVYY) are very poor substrates (kcat/KM < 200 M'* s'*). The class V peptides were not individually characterized due to their sequence diversity but are likely to be poor substrates based on the earlier studies by this laboratory and others (Adams, 1968;

Livingston & Leder, 1968; Meinnel & Blanquet, 1995; Rajagopalan et al., 1997). Thus, the results showed that the deformylase has strict structural requirement at the Pf site,

30 with phenylalanine being the only natural amino acid that can replace methionine while

retaining substantial deformylase activity.

Because isoleucine, lysine, and arginine are potentially important for deformylase

recognition but were eliminated from the library, peptides were synthesized and assayed

with these amino acids incorporated into various positions to test their role in substrate

binding/catalysis. Other substitutions were also made to confirm the importance of the

selected amino acids at each position. Replacement of the N-terminal methionine in f-

MAYY by isoleucine resulted in a marginally active substrate, with a kcJKu value of 4.4

X 10^ M'* s'* (Table 2.2, entry 8 ). Substitution of lysine or arginine at this position,

however, rendered the peptides essentially inactive toward the E. coli deformylase (Table

2.2, entries 9 and 10). At the penultimate position, most of the amino acids are tolerated,

with the exception of negatively charged aspartate and glutamate. Indeed, replacement of

the asparagine of f-MNYL by an aspartyl residue resulted in a 6 -fold reduction in activity

(Table 2.2, entries 11 vs. 12). At the P 3 ’ position, substitution of lysine or arginine for

F/Y in the consensus sequence actually increased the activity by 1.5-5.6-fold (Table 2.2,

compare entries 1 and 2 vs. 13 and 14; 15 vs 18). At the P 4 ’ position, replacement of

tyrosine by lysine or arginine slightly decreased the enzymatic activity (Table 2.2,

compare entries 2 vs 16 and 17), whereas substitution of leucine or serine resulted in

significantly poorer substrates (Table 2.2, compare entries 1 vs 19; 2 vs 18; 6 vs 20; 7 vs

21; 4 vs 22). Finally, when all three residues at the P 2 ’-P4 ’ positions were replaced by

negatively charged aspartate, the resulting peptide, f-MDDD (Table 2.2, entry 23), was a very poor substrate, consistent with the absence of aspartate or glutamate among the

31 selected sequences (Hu et al., 1999). Therefore, an optimal substrate should have the

sequence f-MXZY (X = any amino acid except for Asp and Glu; Z = K, R, Y, or F) or f-

FY(F/Y)X.

Peptide deformylase has evolved to act as a broad-specificity enzyme that

deformylates polypeptides bearing diverse N-terminal sequences. However, it is clear

from the kinetic studies that it does so at drastically different rates. The most prominent

feature of the preferred sequences is amino acids with hydrophobic side chains at all four

positions, consistent with the fact that the deformylase active/binding site is comprised of

largely hydrophobic residues (Meinnel & Blanquet, 1996; Chan et al., 1997; Becker et al.,

1998). The observed sequence specificity can be explained by the structural studies of the

deformylase bound with an inhibitor, (S)-2-0-(H-phosphonoxy)caproyI-L-leucyl-/?-

nitroanilide, which mimics the tetrahedral intermediate of the deformylase-catalyzed

reaction (Hao et al., 1998). In this enzyme-inhibitor complex, the side chain of the ?i’

residue (n-butyl) is deeply buried in a hydrophobic pocket formed by both protein side

chains and the aromatic ring of the / 7 -nitroanilide group. The intimate hydrophobic

interactions between the protein and the P f side chain likely make key contributions to the overall affinity and anchor the formamide moiety for nucleophilic attack by a metal- bound water molecule/hydroxide ion. This limits the types of amino acid side chains that can fit into the pocket. Some side chains may be able to fit into the pocket with high affinity, but may place the scissile formamide group in a nonoptimal position or orientation so that the hydrolysis rate (kcat) is slow. Note that several peptides in Table 2.2 have relatively high affinity (as indicated by their low ATm values) to the deformylase but

32 have very low kcat values (e.g., f-GAFL, f-KAYY, and f-HVYL). The side chain of the

P2 ’ residue, L-Ieucine, is engaged to hydrophobic interactions with the solvent-exposed

Leu-91 and the protein backbone on one face, and the other face, is completely exposed

to the solvent. The lack of specific interactions suggests that many amino acids would be

tolerated at this position, and this is indeed the case. Negatively charged residues are

disfavored, presumably because of their unfavorable interactions with the hydrophobic

Leu-91 and/or the protein backbone. At the Pg'/P#' positions, the p-nitroanilide group sits in a shallow hydrophobic pocket formed by hydrophobic side chains of the protein and the side chain of the Pi’ residue of the inhibitor. The aromatic side chains of the P 3 ’ and

P4 ’ residues in the selected peptides likely occupy the same pocket. The shallow pocket permits many other hydrophobic side chains to bind, albeit with less optimal interactions.

This feature is important, for it allows the deformylase to accept a diverse array of sequences as substrates. Indeed, besides phenylalanine and tyrosine, the selected peptides also contain histidine, leucine, norleucine, and glutamine at the P 3 ’ and P 4 ’ positions.

Lysine and arginine appear to be even more favorable than phenylalanine or tyrosine at the P 3 ’ position. This may be explained by the amphipathic nature of these two side chains. They have hydrophobic methylene groups to interact with the hydrophobic patch on the protein surface and have a positively charged terminal group to interact with the hydrophilic solvent molecules. The f-FY(F/Y)X peptides may bind to the deformylase in a different manner.

A-formylglycyl and Y-formylhistidyl peptides have very low activities toward the

E. coli deformylase (kcat/Kw < 200 M’’ s'*) and reasons for their selection from the library

33 was not clear. However, the f-GIy and f-His peptides were shown to have higher affinity to the deformylase relative to many f-methionyl peptides that failed to saturate the deformylase at 2 mM (Rajagopalan et al. 1997). Resin beads that carry high-affinity peptides might recruit more enzyme molecules to their surface so that the total amount of deformylated peptides was high enough to render the beads positive (Vmax = kcat[E]) even though the kcat value determined in solution phase was low.

In summary, detailed kinetic characterization of individual N-formylpeptides confirmed and complemented the results of a previous library work, and expanded our knowledge of interaction between deformylase and substrate. The sequence specificity information obtained in this work will serve investigation of the catalytic mechanism and in vivo function of peptide deformylase as well as designing selective deformylase inhibitors.

34 CHAPTERS

PEPTIDE DEFORMYLASE INHIBITORS

3.1 Introduction

When penicillin was introduced during the Second World War, it was a medical miracle, rapidly vanquishing the biggest wartime killer-infected wounds. Over the last half-century, antibiotics have helped save millions of lives. However, the selective pressure exerted by antimicrobial agents causes some bacteria to develop resistance to protect themselves against future attack (Neu, 1992). Inappropriate use of antibiotics has accelerated the creation of tougher bacteria that no longer respond to many medications

(Davies, 1996). Today, resistance has been developed to virtually all antibiotics in use and patients infected with some multi-drug resistant bacteria can no longer be treated effectively with any antibiotic therapy available today (Neu, 1992). There is a pressing need for new antibacterial agents with novel mechanisms of action to replace the existing ones that have lost their effectiveness. As a promising target for novel broad-spectrum antibiotics, deformylase is being pursued for drug development by a number of academic and industrial research groups. In our lab, we undertook the work of rational design of deformylase inhibitors that include both transition-state analogs and metal-chelators.

35 Inhibitors specific for peptide deformylase may selectively block protein maturation in

bacterial cells and serve as a new generation of broad-spectrum antibiotics.

Metalloproteases are a group of widely distributed and functionally important

enzymes. For instance, angiotension-converting enzyme is an important element in

regulation of blood pressure, carboxypeptidase A participates in digestion of food in

gastrointestinal tract and collagenase is the destructive agent in arthritis. Many

metalloproteases have been extensively studied and inhibitors designed in order to

intervene in the physiological and disease processes involving these enzymes.

Metalloproteases are characterized by the presence of an essential metal ion and

sensitivity to chelating reagents such as small thiols and 1 , 1 0 -phenanthroline.

Substituting metal-chelating functionalities for the scissile peptide bond of substrates

generally produces potent and specific inhibitors. This approach has been remarkably

successful in designing inhibitors for angiotensin converting enzyme (Cushman et al

1977) and carboxypeptidases A and B (Ondetti et al, 1979). Some of the inhibitors have

been used as commercial drugs.

Although there are considerable differences in structure, substrate specificity and mechanism among proteases, proteolytic hydrolysis of peptide bonds entails a common key step: the generation of a tetrahedral intermediate by a nucleophilic attack on the scissile carbonyl carbon. Aldehydes and ketones, which give tetrahedral adducts upon hydration in aqueous milieu, and phosphonates are common mimetics of the intermediate.

Replacing the scissile amide in enzyme substrates with the aldehyde (Westerik &

36 Wolfenden, 1972; Thompson, 1973) and phosphonate (Jacobsen & Bartlett, 1981) has

given rise to inhibitors for various proteases.

Besides the aforementioned tetrahedral and metal chelating functional groups, a

potent inhibitor also requires certain structural determinants that would interact favorably

with other areas of enzyme active site. These structural determinants are critical for high

binding affinity and necessary selectivity of prospective inhibitors. As the enzyme to

deformylate all nascent proteins in bacterial cells which bear a wide variety of sequences

at the N-termini, peptide deformylase has broad substrate specificity. All N-formyl

peptides studied so far, including the optimal substrates selected from the library

described in previous chapter, either show non-saturation kinetics or at best, have Km

value in high micromolar range. Deformylase inhibitors based on these substrates will

probably also have low binding affinity to the enzyme. Previously, an analog of

deformylase substrate f-Met-Ala-Ser was synthesized and tested as a potential inhibitor.

The failure of this compound to inhibit deformylase underscores of the importance of

information concerning enzyme substrate recognition in inhibitor design.

An unnatural deformylase substrate, f-Met-Leu-pNA, with relatively low Km was later discovered coincidentally in the course of assay development (Wei & Pei, 1997). As described in previous chapter, the specific interactions that are responsible for the high affinity of this molecule are not yet fully understood. However, it is now possible to undertake design of deformylase inhibitors that are analogues of this substrate. Indeed, a phosphonate analog has been found to have a Ki of 37 pM (Hu et al., 1998). In this chapter, more rationally designed derivatives of f-Met-Leu-pNA were synthesized and

37 evaluated as deformylase inhibitors. Besides being potential leads for antibiotics that may be eventually marketed commercially, the inhibitors designed in our lab are usehd probes of deformylase mechanism.

Because of the limited number of compounds one can synthesize and test individually, the scope of compounds chosen for investigation is constrained.

Combinatorial library approach has emerged as a powerful tool to drug discovery that drastically expands the number of molecules surveyed in a reasonable time frame. In order to identify more lead compounds with various structural features for further drug development, a deformylase inhibitor library was constructed on solid phase and screened.

3.2 Experimental Procedures

3.2.1 Materials

4-Mercaptobutyric acid was purchased from Karl Industries Inc. (Aurora, OH).

TentaGel S NH 2 resin, l-hydroxybenzotriazole (HOBt), and 2-(l//-benzotriazole-l-yl)-

1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) was purchased from Advanced

ChemTech (Louisville, KY). Tert-butyl esters of amino acids hydrochloride were from

Bachem. (2-Tricarboxylethyl)phosphine (TCEP) was purchased from Boehringer

Mannheim. P 3 ’ building moiety of inhibitor library, which are all amines were obtained from either Aldrich, Fluka or Lancaster. All other reagents were purchased from either

Aldrich or Sigma.

38 3.2.2 Synthesis

Succiny lieucy 1-p-nitroanilide (11).

O O Leu-pNA (0.062 g, 0.25 mmole) and

succinic anhydride (0.030g, 0.30 mmole)

were dissolved in THF. The reaction was

stirred for 2 h at rt and refluxed for 0.5 h. The cooled reaction was then distributed

between 5% hydrochloric acid and ethyl acetate. Evaporation of the organic phase gave

white solid in quantitative yield.

(2-Butyl)malonylleucyl-p-nitroaniiide (13).

Dimethyl (2-butyl)malonate was treated

O with NaOH to generate 2-butylmalonate which can HO. ^ i Ü be extracted with ethyl acetate after acidification. 0 0 = " After desiccation over P 2O5, the acid crystallized

as a white solid, which was then coupled to Leu-pNA with DCC in dichloromethane.

Following filtration to remove DCU precipitate, the reaction was extracted with 5%

NaHCO]. The aqueous layer was shown to be a single spot on TLC and was used directly

for deformylase inhibition assay.

General Procedure for Synthesis of Mercaptoacyi-Leu-pNA. All the coupling

reactions between acids and Leu-pNA were carried out with HBTU. Products were purified by silica gel column and eluted with ethyl acetate/hexanes. 2^'-Dithiobis(-acetylIeucyl-p-aitroanilide) (14).

Q Mercaptoacetic acid was oxidized to

"àf dimer with hydrogen peroxide before the

coupling reaction. 'H NMR (200 MHz, DMSO):

Ô 10.73 (s, IH), 8.44 (d, J = 7.5 Hz, IH), 8.21 (d, J = 9.2 Hz, 2H), 7.84 (d, J = 9.2 Hz,

2H), 4.35 - 4.58 (m, IH), 3.56 (d, J = 2.7 Hz, 2H), 1.40 - 1.80 (m, 3H), 0.80 - 0.99 (m,

6 H). FAB-MS, 649. Calculated mass 648.

3-lVIercaptopropionyHeucyl-p-nitroaniIide (15).

O 'h NMR (250 MHz, CDCI 3): Ô 9.55 (s. H ÏÏ HS IH), 8.08 (d, J = 9.2 Hz, 2H), 7.61 (d, J - 9.2

Hz, 2H), 6.44 (d, J = 7.9 Hz, IH), 4.60- 4.79 (m,

IH), 2.72 - 2.95 (m, 2H), 2.50 - 2.72 (m, 2H), 1.54 - 1.93 (m, 4H), 0.80 - 1.12 (m, 6 H).

FAB-MS, 340. Calculated mass, 339.

4-MercaptobutanoyUeucyi-p-nitroanilide (16).

Q The commercial 4-mercaptobutyric

acid, which contained a large amount of = H O thiolactone and oxidized dimer, was treated with NaOH to open the thiolactone ring, and triphenylphosphine to reduce the disulfide bond before use. 'H NMR (300 MHz, CDCI 3 ): Ô 9.46 (s, IH), 8.00 - 8.19 (m, 2H), 7.53 -

7.77 (m, 2H), 6.15 (d, J = 7.9 Hz, IH), 4.52 - 4.75 (m, IH), 2.25 - 2.72 (m, 4H), 1.98 (t, J

40 = 8.2 Hz, IH), 1.55 - 1.90 (m, 5H), 0.83 - 1.10 (m, 6 H). FAB-MS, 354. Calculated mass,

353.

2-n-Butyl-y-butyrolactone (17).

y-Butyrolactone (0.86 g, 10 mmole) was added dropwise to a

O solution of lithium diisopropylamide ( 8 ml x 1.5 M) in 30 ml of

newly distilled THF at —78 °C. The reaction was kept at this temperature for another 30

minutes before n-iodobutane (2.2 Ig, 12 mmole) in HMPA (2.15 g, 12 mmole) was added.

After 2 h at —78 °C, the reaction was warmed to rt slowly, poured into saturated NH 4 CI

solution (50 ml) and extracted with diethyl ether (50 mi x 2). The organic phase was

concentrated and chromatographed on silica gel column (elution with ethyl acetate /

hexanes, 1:2) to give 0.40 g of colorless oil (28% yield). *H NMR (200 MHz, CDCI 3): Ô

4.00-4.43 (m, 2H), 2.20-2.60 (m, 2H), 1.60-2.05 (m, 2H), 1.10-1.55 (m, 5H), 0.70-1.00

(m, 3H). Calctd mass: 142.0. FABMS (MH^: m/z 143.1.

4-Bromo-2-n-butylbutanoyileueyl-p-nitroanilide (18).

'H NMR (200 MHz, CDCI3): Ô 9.52 (s,

O IH), 7.95-8.15 (m, 2H), 7.56 (d, J = 9.3 FIz,

^ N 2H), 6.18 (d, J = 7.8 Hz, IH), 4.50-4.80 (m,

IH), 3.15-3.60 (m, 2H), 2.30-2.62 (m, IH),

2.08-2.30 (m, IH), 1.85-2.08 (m, IH), 1.58-1.85 (m, 3H), 1.10-1.58 (m, 6 H), 0.65-1.10

(m, 9H).

41 4-HydroxyI-2-n-butylbutanoyi-Leu-pNA (22).

2-butyl-y-butyrolactone (0.142 g, 1.0 NO ^ mmole) was heated in 1.25 ml of 1 M NaOH

and 2.5 ml of dioxane at 60 °C for 1 h. The

reaction was then cooled, acidified and extracted with ethyl acetate. Desiccation over P 2O5 afforded 4-hydroxyi-2-butyIbutanoic acid 23 as white solid in quantitative yield.

The above solid was dissolved in 10 ml of dichloromethane and imidazole (0.204 g, 3.0 mmole) and t-butyl-dimethylsilyl chloride (0.450 g, 3.0 mmole) were added. The reaction mixture was stirried overnight. White precipitate was filtered out and the filtrate concentrated. The residue was resuspended in diethyl ether (20 ml) and washed with water (30 ml). The organic phase was concentrated and redissovled in methanol/THF (10 ml/3 ml). After 0.5 g of K 2CO3 in 3 ml of H 2O was added, the reaction was stirred at rt for 4 h. After evaporation of solvents, the residue was dissolved in brine (10 ml), extracted with etliyl acetate (10 ml x 2). The aqueous was cooled to 0 °C, acidified carefully to pH3 with citric acid and extracted with ethyl acetate (10 ml x 2). The organic phase was dried over MgS 0 4 and evaporated to give crude 4-(t-butyl-dimethylsilyloxy)-2- butylbutanoic acid 24 as colorless oil 0.192 g (70% yield).

The above product (0.192 g, 0.70 mmole) was coupled to Leu-pNA (0.176 g,

0.70 mmole) with DCC (0.159 g, 0.77 mmole) in 5 ml of CH 2CI2 . After filtration to remove dicyclohexylurea precipitate, the reaction solution was concentrated and

42 chromatographed on silica gel column (elution with ethyl acetate / hexanes, 1:4), 0.127 g

of white solid (25) was obtained (36% yield).

To 65 mg of the above product 25 in 5 ml of ethanol was added 4 drops of

concentrated hydrochloric acid. After 30 min at room temperature, the reaction was

concentrated and subjected to silica gel chromatography (elution with ethyl acetate, 1 : 1 ).

The two diasteroisomers can be easily resolved and 23 mg and 26 mg of white solids

respectively were obtained for isomers that came off column first (2 2 a) and second (2 2 b)

(total 97 % yield for two isomers). 22a 'H NMR (200 MHz, CDCI3 ): ô 9.81 (s, IH), 8.04

(d, J = 9.1 Hz, 2H), 7.63 (d, J = 9.2 Hz, 2H), 6.44 (d, J = 7.8 Hz, IH), 4.57-4.80 (m, IH),

3.55-3.88 (m, 2H), 2.30-2.55 (m, IH), 1.55-2.05 (m, 6 H), 1.10-1.55 (m, 6 H), 0.70-1.05

(m, 9H). 22b 'H NMR (300 MHz, CDCI3 ) Ô 9.55 (s, IH), 8.12 (d, J = 9.1 Hz, 2H), 7.64

(d, J = 9.2 Hz, 2H), 6.30 (d, J = 7.8 Hz, IH), 4.50-4.70 (m, IH), 3.55-3.80 (m, 2H), 2.32-

2.52 (m, IH), 1.95-2.20 (broad singlet, IH), 1.55-1.95 (m, 5H), 1.05-1.55 (m, 6 H), 0.65-

1.05 (m, 9H).

2-n-Butyl-4-oxobutanoyl-Leu-pNA (19).

To compound 22a (23 mg) in 0.5 ml of

O l_l 9 ^ dichloromethane was added pyridinium

i H chlorochromate (PCC) (25 mg). After stnrring

at rt for 1 h, the suspension was applied to a silica gel column and eluted with ethyl acetate / hexanes ( 1 : 1 ) to give 1 0 mg of white

solid. ‘H NMR (200 MHz, CDCI3 ): 5 9.80 (s, IH), 9.12 (s, IH), 8.12 (d, J = 9.1 Hz, 2H),

43 7.87 (d, J = 9.1 Hz, 2H), 6.05 (d, J = 7.7 Hz, IH), 4.55-4.75 (m, IH), 2.35-2.85 (m, 3H),

1.15-2.05 (m, 9H), 0.75-1.08 (m, 9H). Note, NMR data of the major product was reported.

2-MethyIthioethyl iodide (26).

Nal (9.00 g, 60 mmole) was dried by heating for 3 h. The Nal and

(methylthio)ethyl chloride (4.24 g, 40 mmole) is added to 40 ml of anhydrous acetone in a flask that had been purged with argon. The reaction was reflux under argon for 3 h in dark and cooled to room temperature. After the precipitate was removed by filtration, the yellow filtrate was concentrated and resuspended in 50 mi of ice-cold water. The suspension was extracted with dichloromethane (25 ml x 3) and the combined organic phase was washed with concentrated NaHSOs (50 ml), HiO (30 ml x 2) and dried over

Na2 S0 4 . Evaporation of solvent gave 5.98 g of oil with very light yellow color (74% yield). 'H NMR (300 MHz, CDCb) 5 3.31 (t, J = 7.1 Hz, 2H), 2.94 (t, J = 7.1 Hz, 2H),

2.15 (s,3H).

t-Butyl 4-methyithiobutanoate (27).

A flask containing LDA (24 ml x 1.5 M) in 80 ml of

THF was cooled to —78 °C under argon. t-Butyl acetate (3.48 g, 30 mmole) was added dropwise and stirred for 30 min at -78 °C. A solution of 2- methylthioethyl iodide (6.23 g, 30 mmole), HMPA (5.91 g, 33 mmole) in 20 ml of THF was added dropwise and continued for 2 h before it was raised to room temperature. The

44 reaction was quenched with 100 ml o f 10% NH4CI and extracted with ether (100 ml x 3).

The combined ether layer is washed with 200 ml of water and 100 ml of saturated NaCI solution. The organic phase was dried with Na 2 S0 4 , concentrated and purified with silica gel chromatography (eluted with a gradient of 0 - 5% ehthyi acetate in hexanes) to give

6.31 g o f colorless oil (45% yield). 'H NMR (250 MHz, CDCI3) Ô 2.54 (t, J = 6.9 Hz,

2H), 2.35 (t, J = 6.9 Hz,2H), 2.11 (s, 3H), 1.80-1.98 (m, 2H), 1.48 (s, 9H).

t-Butyl 2-(2-trimethylsilylethoxy)methyl-4-methylthiobutanoate (28).

g / A flask containing LDA (16 ml x 1.5 M) in 80

ml of THF was cooled to -78 °C and t-butyl 4-

methylthiobutyrate (3.80 g, 20 mmole) in 15 ml of THF

was added dropwise. The reaction was allowed to continue for 1 h before 4 g of 2-trimethyIsilylethoxymethyl chloride (SEM-Cl) (4.00 g, 24 mmole) was added dropwise. After 2 h at —78 °C, the reaction was warmed to room temperature and quenched with 100 ml of 10% NH 4 CI and extracted with 200 ml of chloroform. The organic phase was washed with 50 ml of H 2O and 50 ml of saturated

NaCl solution, dried with Na 2 S0 4 and concentrated. Purification with silica gel chromatography (eluted with a gradient of 0 — 8 % ethyl acetate in hexanes) gave 4.28 g of colorless oil (67% yield). 'H NMR (300 MHz, CDCI3) 5 3.38-3.65 (m, 4H), 2.60-2.75 (m,

IH), 2.45-2.60 (m, 2H), 2.10 (s, 3H), 1.70-1.90 (m, 2H), 1.45 (s, 9H), 0.80-1.00 (m, 2H),

0.02 (s, 9H). Calctd mass: 320.18. EIMS (M^: m/z 320.18.

45 t-Butyl 2-(2-methylthioethyl)acrylate (30).

II To 2.9 g of the compound 28 in 20 ml of anhydrous

THF was added a solution of n-Bi^NF (30 ml x 1 M) in THF

and the reaction was refluxed for 6 h. The reaction was cooled and purified by silica gel

column chromatography (eluted with a gradient of ethyl acetate 0 - 8 % in hexanes) to

give 1.60 g of colorless oil (93% yield). ‘H NMR (250 MHz, CDCI 3) Ô 6.12 (s, IH), 5.54

(s, IH), 2.45-2.70 (m, 4H), 2.13 (s, 3H), 1.51 (s, 9H).

t-Butyl 3-acetylthio-2-(2-methylthioethyi)propionate (31).

The compound 30 (1.50 g, 6.82 mmole) was mixed 5

ml of thioacetic acid without solvent, heated at 65 °C for 3 h

cooled to room temperature. Silica gel chromatography O O ' yielded 1.30 g of colorless oil ( 6 8 % yield). 'H NMR (250

MHz, CDCI3) Ô 2.98-3.18 (m, 2H), 2.57-2.70 (m, IH), 2.52 (t, J = 6.9 Hz, 2H), 2.33 (s,

3H), 2.10 (s, 3H), 1.70-2.05 (m, 2H), 1.45 (s, 9H).

3-AcetyIthio-2-(2-methylthioethyI)propionyl-Leu-pNA (34).

The above t-butyl ester 31 was

' ^ ^ 2 deprotected with TFA/thioanisole/ anisole

(3:2:1) at room temperature for 1.5 h. After

evaporation of solvent, the residue was dissolved in 20 ml of 5% NaHC 0 3 , extracted with chloroform (10 ml x 3). The aqueous

46 phase was acidified and extracted again with chloroform (10 ml x 2). Evaporation of the

combined organic phase gave 3-acetylthio-2-(2-methylthioethyl)propionic acid as a

yellowish oil in quantitative yield. The acid (0.045 g, 0.21 mmole) were coupled to Leu-

pNA (0.050 g, 0.20 mmole) using DCC (0.044 g, 0.21 mmole) in 2 ml of

dichloromethane. Silica gel chromatography yielded 0.042 g of white solid (51 % yield).

3-Mercapto-2-(2-methylthio)ethylpropionyl-Leu-pNA (36).

To the 20 mg of 34 dissolved in 5 ml of

O anhydrous methanol was added 20 mg of (>10

equivalent) sodium borohydride. After stirring at

rt for 2 h, the reaction was complete and it was

quenched with 5% HCI and extracted with ethyl acetate. The organic phase was dried

over MgS 0 4 and evaporated to give product as white solid in quantitative yield.

3-Acetylthio-2-n-butyIpropionyi-Leu-pNA (35).

The compound was prepared from

9 ^ ^ 3-acetythio-2-n-but>'Ipropionic acid and

H Leu-pNA with similar procedure as

described above. NMR (300 MHz,

CDCI3) 5 9.47 & 9.42 (s, IH), 8.02-8.17 (m, 2H), 7.51-7.70 (m, 2H), 6.18 & 6.14 (d, J =

8.0 Hz, IH), 4.52-4.75 (m, IH), 2.90-3.15 (m, 2H), 2.35-2.52 (m, IH), 2.22 & 2.32 (s,

3H), 1.47-1.90 (m, 5H), 1.15-1.45 (m, 4H), 0.70-1.09 (m, 9H).

47 3-Mercapto-2-ii-butylpropionyl-Leu-pNA (37).

The compound was prepared from 3-

acetylthio-2-n-butyIpropionyl-Leu-pNA and

H NaBHt with similar procedure as described

above. ‘H NMR (300 MHz, CDCI 3) 5 9.56 &

9.61 (s, IH), 8.00-8.15 (m, 2H), 7.50-7.70 (m, 2H), 6.31 & 6.25 (d, J = 7.6 Hz, IH), 4.60-

4.80 (m, IH), 2.52-2.90 (m, 2H), 2.28-2.50 (m, IH), 1.43-1.94 (m, 6 H), 1.13-1.43 (m,

4H), 0.65-1.08 (m, 9H). Calctd mass: 395.5. FABMS (MH^: 396.2.

Dimethyl 2-butylmalonate (38).

Dimethyl malonate (30 mmole, 3.96 g) and sodium

methoxide (1.62 g, 30 mmole) in 10 ml of DMF was heated at 55

°C for 30 min before n-butyl iodide (3.68 g, 20 mmole) in 10 ml of

DMF was added. After 3 h at 120 °C, the reaction was cooled to room temperature, diluted with 1 0 0 ml of icy water, and extracted with 1 0 0 ml of ether.

The ether layer was concentrated and purified by silica gel column (eluted with ethyl acetate / hexane, 1:3) to give product as colorless oil 2.32 g (62% yield). *H NMR (200

MHz, CDCI3) Ô 3.71 (s, 6 H), 3.34 (t, J = 7.5 Hz, IH), 1.80-1.97 (m, 2H), 1.15-1.45 (m,

4H), 0.87 (t, J = 6.7 Hz, 3H).

48 Dimethyl 2-isobutylmalonate (39).

Dimethyl malonate (5.81 g, 44 mmole) and sodium

methoxide (4.75 g, 8 8 mmole) in 100 ml of anhydrous methanol

Q Q was refluxed for 20 min and isobutyl iodide in 50 ml methanol was added. After another 20 min of refluxing, the reaction was cooled to room temperature and concentrated. The residue was distributed in saturated NH 4CI solution 50 ml of diethyl ether (50 ml x 2). The ether layer was concentrated and purified with silica gel column (eluted with ethyl acetate / hexane 1:3) to give 0.82 g colorless oil (11% yield).

'H NMR (200 MHz, CDCI3) Ô 3.73 (s, 6 H), 3.45 (t, J = 7.6 Hz, IH), 1.80 (t, J = 7.4 Hz,

2H), 1.40-1.68 (m, IH), 0.91 (d, J = 6.5 Hz, 6 H).

General procedure for synthesis of 2-alkylacrylic acids.

2-Alkylacrylic acids were prepared following procedure adapted from literature

(Duhamel et al., 1996). A solution of 2-alkylmalonate (10 mmole) and NaOH (0.88 g, 22 mmole) in 15 ml H 2 O/I 5 ml MeOH was refluxed for 1 h. After being cooled to room temperature, the reaction solution was diluted with H 2 O (30 ml) and extracted with ethyl acetate (30 ml). The aqueous layer was cooled to 0 °C, acidified to pH = 1 with concentrated hydrochloric acid and extracted with ethyl acetate (20 ml x 3). The combined ethyl acetate layer was dried with MgS 0 4 , filtered and evaporated to give 2 - alkylmalonic acids as white solids.

The acids obtained above was dissolved in 20 ml of ethyl acetate and cooled to 0

°C when dimethylamine (20 mmole) and paraformaldehyde (15 mmole) were added. The

49 reaction was stirred at room temp for 1 0 min, refluxed for 2 h, upon which it became a homogeneous solution from the initial suspension, and cooled to room temperature. After

30 ml of H 2O was added, the reaction was cooled to 0 °C, acidified to pH = 1 with concentrated hydrochloric acid and extracted with ethyl acetate (30 ml x 2). The combined organic phase was dried with MgS 0 4 , filtered and solvent evaporated to give products.

2-n-Butylacrylic acid (43).

The overall yield is 95%. 'H NMR (300 MHz, CDCI 3); 5 6.28 (s,

IH), 5.64 (s, IH), 2.31 (t, J = 7.5 Hz, 2H), 1.22-1.58 (m, 4H), 0.92 (t, J = OH 7.2 Hz, 3H). O

2-Isobutyiacrylic acid (44).

The overall yield is 44%. *H NMR (200 MHz, CDCI 3): 5 6.32 (s,

IH), 5.61 (s, IH), 2.17 (d, J = 7.0 Hz, 2H), 1.70 - 1.97 (m, IH), 0.89 (d, J OH

g = 6 . 6 Hz, 6 H).

2-BenzyIacrylic acid (45).

The overall yield is 45%. 'H NMR (200 MHz, CDCI3): S 7.12 - OH 7.60 (m, 5H), 6.38 (s, IH), 5.58 (s, IH), 3.63 (s, 2H). O

50 2-n-Butyl-2-dimethyiaminomethyl-malonic acid hydrochloride (42).

When water was used as the solvent for the reaction of 2-n-

butylmalonic acid (and other 2 -alkylmalonic acids as well) with OH paraformaldehyde and dimethyl amine, the Manmch base was O O / \ HCI obtained as the major final product. Briefly, after refluxing for 2 h, the reaction solution was cooled and acidified, upon which a large amount of white precipitate formed. The white solid was collected by filtration and washed with cold water to give title compound. ‘H NMR (200 MHz, D 2 O) 6 3.58 (s, 2H), 2.85 (s, 6 H), 1.76

(t, J = 7.7 Hz, 2H), 1.05-1.48 (m, 4H), 0.82 (t, J = 6 . 6 Hz, 3H).

General procedure for synthesis of 3-acetylthio-2-alkylpropionc acids. 2- alkyl-acrylic acids ( 1 0 mmole) were mixed with 15 mmole of thioacetic acid without solvent and refluxed for 2 h and cooled to room temperature. Excess thioacetic acid was evaporated to give the products with quantitative yields. The products contained trace impurities that were from commercial thioacetic acids. The products were of quality sufficient for direct use for further synthesis. However, most (but not all) of the impurities can be removed by the following work-up if needed. The products were dissolved in 50 ml of 5% NaHCO] and extracted with ethyl acetate (50 ml x 2). The aqueous solution was cooled to 0 °C, carefully acidified to pH = 2 with concentrated hydrochloric acid, and extracted with ethyl acetate (50 ml). The ethyl acetate layer was dried with MgS 0 4 , filtered and solvent evaporated to give purer products.

51 3-Acetylthio-2-n-butylpropionic acid (47).

NMR (200 MHz, CDCI3) Ô 2.95-3.20 (m, 2H), 2.52-2.70 OH (m, IH), 2.33 (s, 3H), 1.52-1.80 (m, 2H), 1.20-1.45 (m, 4H), 0.89 V'o (t, J = 6 . 6 Hz, 3H).

3-Acetyithio-2-isobutylpropionic acid (48).

'H NMR (200 MHz, CDCI3) ô 2.87-3.23 (m, 2H), 2.57- OH Q 2.78 (m, IH), 2.34 (s, 3H), 1.52-1.83 (m, 2H), 1.28-1.50 (m, IH),

0.93 (d, J = 6 . 6 Hz, 6 H).

3-Acetylthio-2-benzylpropionic acid (49).

‘H NMR (200 MHz, CDCI3) Ô 7.10-7.40 (m, 5H), 2.80- OH 3.20 (m, 5H), 2.33 (s, 3H). V'o

3-Acetylthio-2-methylpropioiiic acid (50).

^OH ‘H NMR (200 MHz, CDCI3) Ô 2.95-3.20 (m, 2H), 2.60-

2.85 (m, IH), 2.33 (s, 3H), 1.28 (d, J = 7.1 Hz, 3H).

52 3-(2-Pyridyldithio)propionic acid (52).

To a solution of 2, 2’-dithiopyridine (3.96 g, 18

mmole) in glacial acetic acid (0.4 ml) and 10 ml of absolute O ethanol was added dropwise a solution of 3- mercaptopropionic acid in 10 ml ethanol with vigorous stirring. The reaction was left overnight and then concentrated and applied to an alumina gel column equilibrated with dichloromethane / Ethanol (60/40, v/v). The column was eluted with the same solvent under which condition all by products and excess reactant came out. The column was then washed with a mixture of dichloromethane / ethanol / acetic acid (60/40/4) to elute desired product. The product fractionates contained a small amount of column bed material and appeared milky. The factions were collected, concentrated and subjected to a silica gel column (elution with ethyl acetate / hexanes / acetic acid, 50/50/4) to remove

AI2O3 . Solvent was removed by exhaustive evaporation in vacuo to afford 1.35 g of product as white solid (70% yield). 'H NMR (200 MHz, CDCI 3) ô 8.40-8.58 (m, IH),

8.15 (broad singlet, IH), 7.55-7.75 (m, 2H), 7.03-7.22 (m, IH), 3.05 (t, J = 6 . 8 Hz, 2H),

2.79 (t, J = 6 . 8 Hz, 2H).

3-Acetylthio-2-a-butylpropioayl-Leu-2-aathraceaeainide (53).

tBoc-Leucine (0.125 g, 0.50

l_l yr H T T 1 mmols) was coupled to 2 -anthracene

= H (0.097 g, 0.50 mmole) with DCC in 8

53 ml of dichloromethane/DMF (1:1) at room temperature for 1 h. The precipitate was

removed by filtration and filtrate was diluted with ether ( 2 0 ml) and washed with H 2O ( 2 0

ml X 2). The organic phase was concentrated and chromatographed (elution with ethyl

acetate / hexanes, 1:2) on silica gel column to give 0.152 g of off-white solid (75% yield).

The above product (0.106 g, 0.25 mmole) was dissolved in TFA / anisole (1.5 ml /

1.5 ml) and the solution instantly became blue and black color. After 1 h at rt, the reaction

was concentrated in vacuo, and the residue triturated with hexanes with solid formed

immediately. Upon standing overnight, the dark color of the suspension disappeared and an off-white solid formed. Solvent was pipeted out and the solid was washed with hexanes two more times.

The above compound was coupled to compound 3-acetylthio-2-n-butyIpropionic

(0.052 g, 0.25 mmole) with HBTU (0.095 g, 0.025 mmole) and triethylamine (0.075 g,

0.75 mmole) in 3 ml of DMF. After 1 h, the reaction was diluted with 10 ml o f diethyl ether and washed with H 2O (10 ml x 3). The organic phase was concentrated and chromatographed on silica gel column (elution with ethyl acetate / hexanes, 1 :2 ) gave

0.119 g of product (diastereomeric mixture) as off-white solid (93% yield for two steps).

*H NMR (300 MHz, CDCI 3) Ô 9.05-9.25 (m, IH), 7.30-8.43 (m, 9H), 6.40-6.60 (m, IH),

4.74-4.90 (m, IH), 2.95-3.20 (m, 2H), 2.38-2.58 (m, IH), 2.13 & 2.32 (s, 3H), 1.45-2.00

(m, 5H), 1.10-1.45 (m, 4H), 0.65-1.10 (m, 9H).

54 3-Mercapto-2-ii-butylpropionyl-Leu-2-anthraceneainide (54).

Title compound was prepared in

quantitative yield from 3-acetylthio-2-

butyIpropionyl-Leu-2-anthraceneamide by

treatment with NaBH* in similar procedure described above. 'H NMR (300 MHz, CDCI 3 ) 5 8.77-9.00 (m, IH), 7.30-8.40

(m, 9H), 6.15 -6.47 (m, IH), 4.68-4.85 (m, IH), 2.73-2.95 (m, IH), 2.50-2.73 (m, IH),

2.23-2.45 (m, IH), 1.40-2.05 (m, 6 H), 1.12-1.40 (m, 4H), 0.65-1.12 (m, 9H). Calctd mass: 450.6. FABMS (M H^: 451.2.

3.2.3 Construction of Inhibitor Library

Linker attachment. TentaGel S NH 2 resin (80-100 pm, 0.30 mmol/g loading,

2.86 X 10^ beads/g) was used as the solid support for the peptide libraries. Synthesis was carried out on a 2.0-g scale on a homemade peptide synthesis apparatus. To a suspension of the resin in 10 ml of dichloromethane, three equivalents of 3 -(2-pyridyIdithio) propionic acid and diisopropylcarbodiimide were added. The reaction was complete after shaking at room temperature for 1 h as indicated by ninhydrin test. After washing 5 times with DMF, the resin was evenly divided into 5 separate reaction vessels.

First randomized building block. The first randomized building block was introduced by thiol exchange. The 3-acetylthio-2-alkyl-propionic acids (0.6 mmole) were

55 treated with 2 N NHs/MeOH (5 ml) at room temperature for 2 h to cleave the thioester bond. The reactions was evaporated to dryness, redissolved in O.IM / pH 7.0 phosphate buffer / acetonitrile (4.5 ml/0.5 ml) and added to the resin. After 2 h at room temperature, the reactions were drained and concentration of 2 -thiopyridone in the solution determined by measuring OD at 343 nm (Grassetti & Murray, 1967). All the reactions were complete as judged by the amount of released 2 -thiopyridone that could be calculated from its reaction volume.

After washing with acetic acid/dichloromethane/methanol (5:4:1), the beads were incubated in the same mixed solvents with shaking for 1 h to protonate the carboxylate group. The beads were then washed with large amount of dichloromethane, incubated in dichloromethane with shaking, drained and washed again with large amount of dichloromethane to ensure complete removal of residual acetic acid and methanol.

The beads were dried by vacuum suction and 50 mg from each reaction was weighed out and saved for deconvolution purpose later. The rest of the beads from 5 vessels were combined and mixed thoroughly and divided evenly to 5 portions.

Second randomized building block. Five carboxyl-protected (as t-butyl esters) amino acids were coupled to the resin in the five vessels using a 3-fold excess of reagents by standard peptide chemistry with HOBt/HBTU/NMP in DMF for 1 h. Coupling reactions was repeated once to ensure complete reaction. The resin was washed with

DMF and dichloromethane and treated with TFA / anisole (1:1) to deprotect t-butyl ester.

TFA was completely removed by washing with a large amount of methanol, shaking in

56 methanol for 1 h, washing with a large amount of dichloromethane, shaking in

dichloromethane for 1 h and again washed with a large amount of dichloromethane. A

portion of 50 mg of resin from each vessel was saved for later screening and the

remainder from all vessels was combined, thoroughly mixed and redistributed evenly into

30 reaction vessels.

Third randomized building block. All thirty building blocks at this randomized

position are amines. Aromatic amines were coupled to the resin with 3 equivalents of

diisopropylcarbodiimide and 0.3 equivalents of DMAP in dichloromethane and DMF

(1:1) for 1 h. All other amines were coupled with 3 equivalents of HBTU and HOBt in

the presence NMP in DMF. Each coupling reaction was repeated once. The resin was

washed with DMF.

Cleavage of inhibitors from solid phase. The resin was suspended in MeOH /

0.1 M sodium phosphate buffer (pH 7.0) (8:1, v/v) and 3 equivalents of TCEP was added.

After incubation at rt for I h, reaction solutions were drained and used directly for

deformylase inhibition assay. The solutions were stored at -20 °C until use.

3.2.4 Screening of inhibitor library.

Most of molecules of the library also inhibited Aeromonos aminopeptidase,

another metallopeptidase, and a stopped version of the aminopeptidase-coupled deformylase assay was employed to screen the library (Wei & Pei, 1997). Briefly,

57 deformylase reaction of 50 piM of f-Met-Leu-pNA in the presence of a total concentration of 25 jxM inhibitors was allowed to proceed for 30 sec before being quenched with 10 mM of hydrogen peroxide which inactivates both the deformylase and the inhibitors. 100 pM of p-chloromercuribenzoic acid was added and incubated for 5 min to scavenge any trace thiols that may interfere with subsequent aminopeptidase reactions. After 0.4 U aminopeptidase was added and the reaction was incubated for another 5 min, OD 405 was measured. Percentage of activity remaining was calculated by dividing the absorbance of deformylase reaction in the presence of inhibitor by that of controls. Reactions were carried out in triplets.

After each round of screening, the pool that showed the most inhibition was selected for next round of deconvolution. Following the coupling of 30 amines at P3’ position and cleavage, the 30 individual reaction solutions were assayed for deformylase inhibition to identify the most active portion (round 1). The specific amine used for this portion was coupled to the five fractions of resin that had been saved after attachment of

5 amino acids of P 2 ’ position. After cleavage, the five reaction solutions were again tested for activity to find out the optimal residue at P 2 ’ position (round 2). Finally, the five portions of resin that had been saved after attachment of 3-mercaptopropanoic acids were coupled to the specific amino acid as P 2 ’ residue, and the amine as P3’ residue. The five reactions were cleaved and assayed to determine the most active P f structure (round 3).

58 3.2.5 Inhibition assay

For aldehydic and carboxyl types of inhibitors, kinetic studies were carried out with continuous deformylae assay coupled to Aeromonas proteolytica aminopeptidase.

Inhibitors of thiol type were assayed using a stopped version of this assay, which is similar to the procedure described for inhibitor library screening. Deformylase reaction rates of f-Met-Leu-pNA were determined in the presence of various concentrations of inhibitors and the apparent Km values (Km app)were plotted vs. inhibitor concentrations.

From the slope of the plot (Km app/KM = 1 + |T|/K,), inhibitor constant K; values were calculated.

In case where only a limited number of data points were taken, Ki values were calculated from the equation below and an average was reported.

V W = (Km + [s]) / {(1 + [1]/K,)Km + [s]}

Where V’ and V are deformylase reaction rates determined with or without inhibitors, respectively, and [s], [ 1] are the substrate and inhibitor concentrations used, respectively.

Km value of 20 pM was used for f-Met-Leu-pNA for calculation.

3.3 Results and Discussion

3.3.1 Initial Design of Deformylase Inhibitors

Peptide deformylase contains a metal ion at the active site that is essential for its catalytic activity. The metal ion probably coordinates with the carbonyl oxygen of bound substrate and assists amide bond cleavage. It is therefore postulated that deformylase

59 inhibitors may be designed by incorporation of metal chelating functional groups in place of carbonyl group in deformylase substrates. To test the validity of this hypothesis, two types of amino acid derivatives, carboxyalkanoyl and mercaptoaikanoyl peptides, were prepared with chemical reagents readily available in the lab. The former were synthesized by reacting Leu-pNA with corresponding acid anhydrides, or carboxylic acid and DCC; the later by coupling corresponding mercaptoalkanoic acid to Leu-pNA with HBTU.

NO, O

HO O 11

NOg

o o 12

NO,

HO

13

Figure 3.1. Carboxylic and sulfhydryl deformylase inhibitors.

60 Carboxylate inhibitors K,(pM) Thiol inhibitors Ki (pM)

2 -mercaptoacetyl- Succinyl-Leu-pNA 222 ± 38 >190 Leu-pNA

3-mercaptopropionyl Glutaryl-Leu-pNA ND 34.3+2.1 -Leu-pNA

(2-n-butyl)maionyI- 4-mercaptobutanoyl 6 6 ± 3 7.5 ± 0.5 Leu-pNA -Leu-pNA

Table 3.1. Constants of carboxylate and sulfhydryl deformylase inhibitors.

Ali compounds were shown to be inhibitors of deformylase and the inhibition was time independent. Succinyl-Leu-pNA was moderately active with Ki of 222 ±38 jxM. Its analog 13 had improved potency (Ki = 6 6 ± 3 p.M). The increased affinity may be attributed to the fact that the latter has one less methylene between two carbonyl groups and/or that it has a butyl substitution on a-carbon. It would be interesting to test malonyl and glutaryl analogs to find out the optimal chain length of the carboxyalkanoyl group.

Replacement of the carboxyl group by sulfhydryl group leads to increased potency. The K[ values of 3-mercaptopropionyl- (15) and 4-mercaptobutanoyl-inhibitors

(16) are 34 ± 2 pM and 7.5 ± 0.5 pM respectively. The remarkably higher affinity of the mercapto-derivatives may be explained by the fact that a thiolate is a much better ligand for the metal ion than a carboxylate group.

61 4-mercaptobutanoyl-Leu-pNA is >7 times more potent than 3-mercaptopropionyl derivative, whereas Ki of the shorter 2-mercaptoacetyI analog (14) is estimated to be >190 fiM. This suggests that the optimal chain length for mercaptoaikanoyl inhibitors is that of

4-mercaptobutanoyl inhibitor. This length probably allows for the most favorable alignment of thiol with metal ion, and other inhibitor functionalities with their respective sites on deformylase. Potency may be improved by introducing alkyl substitution on a- carbon, which would interact with the Pi’ binding pocket.

3.3.2 Aldehyde: A Transition State Analog as Deformylase Inhibitor

The target molecule 19 in Figure 3.2 is an analog of deformylase substrate, f-Met-

Leu-pNA, with an aldehyde group in the place of formylamide. In aqueous solution, aldehydes exist in equilibrium with hemiacetals, which mimic the sp3 hybridization of the scissile carbon in the transition state of proteolytic reactions. Thus, compound 13 was designed as a potential deformylase inhibitor. Figure 3.2 outlines the initial attempts toward its synthesis. N-Butyl was introduced to the lactone ring through alkylation. The reaction suffered very low yields for unknown reasons. The resulting compound 2-butyl-

4-butyrolactone was treated with boron tribromide to open the lactone ring followed by aminolysis with Leu-pNA to give 2-n-butyl-4-bromobutanoyl—Leu-pNA (Yazawa et al

1974), which was aimed to generate the desired aldehyde by oxidation. There are many examples of this type of reactions in literature (Epstein & Sweat 1967). However, the bromide underwent extensive decomposition when various oxidation conditions were

6 2 applied which generally involved extended heating at high temperature and base treatment.

1. LDA 1. BBr3 2. Leu-pNA

17

NO. NO.,

oxidation N = H O H

18

Figure 3.2. Initial synthesis of an aldehydic deformylase inhibitor.

Attempt was then made to convert the above alkyl bromide functionality to hydroxyl group, which could then be oxidized to aldehyde under much milder conditions.

Bromobutanoylanilide, prepared conveniently from 4-butyrolactone and aniline, produced the corresponding alcohol in high yields after brief heating with sodium carbonate and silver carbonate. Despite this success, 2-n-butyl-4-bromobutanoyl-Leu-pNA again decomposed, giving a mixture of unidentified species.under similar conditions.

63 Na^COg/AggCO^ 70 =C

NO, NO,

HO = H 5 H

18 22

Figure 3.3. Conversion of bromide to hydroxyl functionality.

An alternative route as outlined in Figure 3.4 was tried out. The lactone ring was opened with NaOFI to produce 2-buyl-4-hydroxylbutanoic acid. The product had to be used within several hours as it recyclized to form lactone over time after preparation.

Coupling of this compound to Leu-pNA was first attempted with DCC and HBTU. In both cases, extensive lactonization of the 4-hydroxyIbutanoic acid was observed, and no product isolated. To prevent the cyclization, the hydroxyl group had apparently to be protected. Thus, right after opening of the lactone ring, the product treated with excess t- butyldimethylsilyl chloride to protect the -OH and -carboxyl as ether and ester respectively (Corey & Venkateswarlu, 1972). The ester was selectively deprotected in aqueous potassium carbonate solution to release the carboxyl group (Morton &

Thompson, 1978). The acid was then coupled to Leu-pNA and the ether bond of the

64 product was be easily cleaved with diluted hydrochloric acid solution in ethanol to liberate the hydroxyl group (Wetter & Oertle, 1985; Cunico & Bedell, 1980). Oxidation of hydroxyl group gave the final aldehyde product.

Deformylase inhibition by the compound was studied with continuous aminopeptidase-coupled assay. It exhibits time-independent, competitive inhibition of deformylase with an inhibition constant of 35 ± 6 pM. Since the structure of the aldehye deviates from the substrate f-Met-Leu-pNA by only one isosteric substitution (-CH? vs. —

NH), the two molecules probably bind to the enzyme active site with similar mode.

Therefore, it is not surprising to see that the binding affinity of the two molecules to deformylase is very close (35 pM vs. 20 pM).

65 1. NaOH/H,0/dioxane 1. TBDMS-CI, imidazole O OH J HO 17

NO, S i-O OH -S i-O DCC a H

24 25

Eton

Figure 3.4. Revised synthesis of aldehydic deformylase inhibitor.

66 3.3.3 Sulfhydryl Deformylase Inhibitors

As discussed earlier in this chapter, 3-mercaptopropanoyl-Leu-pNA was a fairly

potent inhibitor of deformylase and the activity is expected to improve when proper alkyl

groups similar to methionine side chain are introduced to a-carbon. Figure 3.5 illustrates

the synthesis toward such a molecule. The 2-methylthioethyl iodide used was prepared

from chloride by heating with sodium iodide in acetone. The iodide was very unstable and as a result, the halide exchange reaction as well as subsequent work-up was done in

dark. The product was used for alkylation immediately after preparation. Another step alkylation introduced 2 -trimethylsilylethoxymethyl substitution that was aimed to produce hydroxymethyl group after deprotection, and eventually the target molecule in several steps. Interestingly, treatment with tetrabutylammonium fluoride produced an a,

P-unsaturated ester. Apparently, an P-elimination had occurred probably because in anhydrous organic solvent fluoride ion is a base strong enough to deprotonate a-carbon.

The unexpected outcome of the above reaction actually turned out to be useful for the synthesis. The a, P-unsaturated ester is a Michael acceptor and heating with thiolacetic acid afforded the thiol ester. After deprotection of t-butyl ester, the acid may be coupled to various peptides to produce potential deformylase inhibitors.

The acid prepared above was coupled to dipeptide Ala-Ser by solid phase synthesis. Following cleavage with TFA, free thiol of the peptide was liberated ammonia hydroxide. Unfortunately, the compound did not show appreciable inhibition of deformylase. Considering non-saturation kinetics for natural deformylase substrates, the low binding affinity of the f-Met-Ala-Ser analogue could be understood. It is therefore

67 critical to incorporate optimal peptide sequences that interact with other binding areas other than the metal ion at the enzyme active site.

Since f-Met-Leu-pNA has the lowest Km among all substrates characterized for far, it was chosen as the parent molecule for further inhibitor design. As a result, the acid

32 was coupled to Leu-pNA. Free thiol group was released by reductive cleavage using sodium borohydride as strong acid or base hydrolysis would damage the leucyl-p-anilide bond.

The resulting molecule was assayed with two independent assays, aminopeptidase- and formate dehydrogenase-coupled assay and was shown to be a potent inhibitor with K[ of 0.19 ± 0.2 nM. Compared to 3-mercaptopropionyl-Leu-pNA (Ki = 34

± 2 pM), incorporation of methionine side chain indeed drastically improved the inhibitor activity.

6 8 1. LDA, -78 °C ° / 1.LDA O ^ 2. 2. SEM-CI

26 27 O

'O O n-Bu^NF O' o

28 HO 29 .Si 1. mesyl chloride 2. thiolacetic acid |n - Bu^NF o

SH 8 O O O O 30 31

TFA/anlsole S OH O O 32

Figure 3.5. Synthesis of 3-acetylthio-2-(2-methyIthioethyI)propionic acid.

69 OH

^OH solid phase synthesis NH3 HS OH E H O H2O 32 33

NO.

S OH DCC Y' E H O O

32. X = S 34. X = s 47. X = CH, 35. X = CH

NO N aBH . HS

36. X = S 37. X = CH,

Figure 3.6. Final steps of synthesis of sulfhydryl deformylase inhibitors.

70 There are several drawbacks with the synthesis of the methionine analogue with masked thiol group, 3-acetylthio-2-(2-methyIthioethyi)propionic acid 32, which is the requisite building block needed for the synthesis of other inhibitors of this type. Some of reagents used were expensive and preparation of 2 -methylthioethyl iodide was tricky given its extreme instability. The two alkylation steps requiring strict anhydrous conditions were rather inconvenient and gave only moderate yields. A superior synthetic route of 2-substituted acrylic acids was adopted involving decarboxylation of Mannich base that could be conveniently prepared from respective malonic acids (Figure 3.9). A norleucine rather than a methionine analog was also chosen in order to expedite the synthesis and to avoid intense odor of small sulfhydryl compounds. Thus, 2-butyl-3- acetylthiomalonic acid was synthesized and used to prepare more deformylase inhibitors, including 2-n-butyl-3-mercaptopropionyl-Leu-pNA (Figure 3.6).

The diastereomeric mixture of this inhibitor was resolved on HPLC and kinetic characterization was conducted in detail. Lineweaver-Burk plot of the kinetic data showed the inhibition to be competitive. Kinetic characterization showed that the more active isomer of 37 had K[ of 81 nM, and the other isomer with a K, of 0.65 ± 0.06 pM was about 8 fold less active. It is likely that the activity of the less active isomer was caused by contamination of a small amount of the more active one. The differential potency of the two isomers of 37 is in good agreement with the fact that formyl-(D)- methionine peptides are not deformylase substrates. Inhibitor 37 was >100 fold more active than inhibitor 15 that lacks the n-butyl side chain at P f position, indicating again the importance of PF side chain in enzyme binding. A Ki = 0.19 ± 0.2 pM was reported

71 earlier in this chapter for a closely related inhibitor (36) that had methionine instead norleucine side chain substitution. This K| of 36 was determined using a diastereomeric mixture. Since only one of the two isomers likely contributes to the most of the inhibition, the true Kj of the more active isomer of 36 should be roughly half of 0.19 pM.

Therefore just as expected, deformylase inhibitors 36 and 37 possess similar potency.

Recently, inhibitors of the same type were reported but the activity (Kj = 2.5 pM) was much lower (Meiimel et al., 1999). The compounds described in this work represents the most potent inhibitor ever made up to the point when they were developed. Studies reported above exemplify the rational approach that has been and will continue to be an integral part of inhibitor development and optimization. When reasonable amount of knowledge has been accumulated about enzyme substrate specificity and mechanism, high affinity molecules with structural configiuation complementary to enzyme substrate binding sites may be designed.

72 1.4 0 nM 1.2

1 131 nM -Î2 0 5 262 n E o E 394 nM 3. 0.6 525 nM 0.4

0.2

0 0 20 40 60 80 100 120 140 160 [s], mM

Figure 3.7. Plot of deformylase reaction rate vs. substrate concentration in the presence of inhibitor.

73 3.5

2.5

0.5 0 20 40 60 80 100 1/[s], mivr^

8

6

4 (0â

2

0 0 100 200 300 500400 [I], nM

Figure 3.8. A. Lineweaver-Burk plot of deformylase inhibition. B. Plot o f K m app/ K m v s [I].

74 33.4 Synthesis of 3-acetyltbio-2-aU^lpropionic acids.

With the highly acidic nature of a-hydrogen in malonic acid ester, the a-carbon

was easily deprotected with sodium hydroxide or sodium methoxide. The resulting

carbanion was refluxed with appropriate alkyl halides to achieve a-alkylation. Following

hydrolysis of esters, malonic acids were reacted with dimethylamine and

paraformaldehyde to give the corresponding Mannich bases, which upon heating

produced 2-substituted acryhc acids (Duhamel et al. 1996). Interestingly, when the

formation of Mannich base and its subsequent decarboxylation was conducted in water,

the yields were very poor even after prolonged time of heating, while the reaction gave

high yields within 3 hours when ethyl acetate was used as solvent instead. This reaction

seems to be more thermodynamically favored in organic solvents than in aqueous solution

due to the more polar nature of reactants than products. The 1, 4-addition of the a, p-

unsaturated acids with thiolacetic acid afforded 3-acetylmercaptopropionic acids.

The above compounds were used to prepare inhibitors in solution phase by coupling to peptides and subsequent cleavage of the thiol ester bond. They were also the starting materials for combinatorial inhibitor synthesis on solid phase. In later case, the compounds were treated with methanolic ammonia to unmask the thiol group, followed by evaporation of the solvent and neutralization with phosphate buffer. The free sulfhydryl group of the resulting 2-alkyl-3-mercaptopropionic acids allows ready attachment to pretreated TentaGel beads through thiol exchange.

75 o o O R11, NaOH or or NaOMe R O ^ O R RO OR R = CH, R1 38. R = methyl, R1 = n-butyl 39. R = methyl, R1 = isobutyl 40. R = ethyl, R1 = benzyl

O 1. (CHzO),, O HN(CH3)2 43. R1 = n-butyl 1. NaOH 4 4 . R1 = isobutyl ► HO OH OH 4 5 = benzyl 2 . H ' 2 . reflux RI R1 46. R1 = methyl 41

O O O CH3 COSH 1. NH3/MeOH OH ► HS ONa 2. pH7.0 buffer RI R1 4 7 . R1 = n-butyl 51a. RI = n-butyl 48. R1 = isobutyl 51b. R1 = isobutyl 49. R1 = benzyl 51c. R1 = benzyl 50. RI = methyl 5id . R1 = methyl

Figure 3.9. Preparation of 2-alkyI-3-mercaptopropionic acids.

76 33.5 Design and Construction of Inhibitor Library

The inhibitor library was generated on TentaGel S resin (Figure 3.10). A linker, 3- mercaptopropionic acid, after activation of its thiol group with 2, 2’-dithiodipyridine, was hooked to the amino group of solid resin beads by amide bond formation. The linker was added to allow convenient attachment of Pi’ moieties through disulfide bond and to add some flexibility to the peptides, making them more accessible to chemical reagents used in subsequent reactions.

The randomized positions were generated using a split-pool synthesis method

(Lam et al., 1991; Furka et al., 1991). This method ensures an equal representation of all building blocks used at each of the randomized positions. In this work, randomized position at P |’ site included five 3-mercaptopropionic acids with a-alkyl substitutions that resemble the side chains of Gly, Ala, Leu, Norleucine and Phe [designated as (HS)Gly,

(HS)Ala, (HS)Leu, (HS)nLeu and (HS)Phe respectively]. They were chosen based on our previous finding that only certain amino acids with hydrophobic side chains may substitute Met in deformylase substrate while retaining good activities. Amino acids at randomized P 2 ’ site consist of (L)-Ala, (D)-Ala, (a-methyl)Ala, Leu and Phe. These amino acids with simple aliphatic side chains were used because of commercial availability, ease of synthesis and because residues Ala, Leu, Phe did appear at the penultimate position in some of good deformylase substrates. Unnatural amino acids (D)-

Ala, (a-methyl)Alanine were used in order to obtain inhibitors with higher in v/vo stability. At P3’ and P4’ position, 30 amines with aromatic rings were used because amino acids with aromatic side chains are highly favored by deformylase (Figure 3.11). The P3’

77 building blocks are designated in bold italic font to distinguish from the bold font generally used for numbering molecules throughout this dissertation.

Therefore, the theoretical diversity of the library is 5 x 5 x 30 or 750 different sequences. The decision to randomize these positions was based on our previous observation that only the four N-terminal four residues are important for deformylase binding and catalysis; extension of the peptide beyond the fourth position does not increase the kinetic constant kcat/Kvi (Rajagopalan, et al., 1997). The attachment of and P 3 ’ building block was all executed with standard peptide chemistry.

78 O DCC, 52 N NH,

R1 ONa O R1 HS OH

O

O H^N O R1 O O H R3 R2 N, O HBTU, HOBt, NMP O R3 R2

O R1 O 1. TFA R4 2. H2 NR4 , DIPC or O R3 R2 HBTU/HOBt/NMP

R1 O TCEP H HS N, R4 O R3 R2

R1 = n-butyl R2 = CH 3 , R3 = H = i-butyl R2 = H, R3 = CH 3

I R2 = CH3 , R3 = CH3 ^methyl r 2 = i-butyl, R3 = H R2 = benzyl, R3 = H

Figure 3.10. Scheme of construction of inhibitor library.

79 NH

NH'2 NH'2 1 3

H .N. vii NH

N'

NO. 2 CF. 6

-ÏJ, H ^N H , N. N'

NO. 11 12 10 NH.

XVI

NH HgN ^ 0

13 14 16 15

Figure 3.11. Structures of P 3 ’ building blocks.

80 (Continued from previous page)

NH. .NH.

— NH. W N 17 18 19

NH NH NH

21 22 NH.

.NH. NHNH HN' HOI HN' HO

N ^ S

NO. NO. OH OH 23 CH, 24 25 26

H .N, .NH. 'OH NH, HN' NH. N HOI

W/O F ^ 27 28 29 F 30 ^^3

81 3.3.6 Screening of Inhibitor Library

An iterative strategy for active sequence identification was employed (Blake &

Litzi-Davis, 1992). Following cleavage from the resin, each of the 30 individual pools has a specific P 3 ’ moiety while Pi’and P 2 ’ are a mixture of 25 (5 x 5) discrete entities. While many of the 30 pools inhibited deformylase to some extent, the order of preference for P3 ’ residues is 14 > 13 > 6, 8 > 15. The order of activity was confirmed by an independent deformylase assay wherein the free amine of f-Met-Ala released by deformylase reacts with 2, 4, 6 -trinitrosulfonic acid to give a product that can be monitored on a spectrophotometer at 420 nm. It is worth pointing out the resemblance of the structural features of the selected amines, especially between 14 and 13, or 6, 8 and 15.

For the next round of screening, 2-Aminoanthracene, the P3 ’ moiety for the most active pool 14 was chosen. It was coupled to the five portions of resin that had been saved after attachment of P 2 ’ amino acids and deprotection of t-butyl group. Cleavage with

TCEP furnished 5 pools of inhibitors, each being a mixture of inhibitors with 5 different mercaptoacyl moieties at PT position. Evaluation indicated that the pool with Leu as the

P2 ’ residue is the most active followed by Phe, while others showed no significant inhibition.

For the third round of screening. Leu and 2-aminoanthracene as defined P 2 ’ and

P3 ’, respectively, were coupled to the 5 portions of resin that had been saved following the attachment of 2-alkyl-3-mercaptopropionic acids. Activity assay of the five individual inhibitors cleaved from resin showed that n-butyl, followed by benzyl, is the optimal side

8 2 chain at the Pi’ position. This is consistent with the results from our substrate specificity studies that norleucine and phenylanine are the preferred terminal residues.

Deconvolution of the library was also carried out with 1 -aminoanthracene, the second most favored structure at P 3 ’ position. The results were very similar to that obtained with 2-aminoanthracene. Leu and Phe at P 2 ’, while norleucine and Phe analogs at Pi’ position, in that order, remain as the matter of choice for active deformylase inhibitors. Thus, the overall trend in the preference for Pi’ and P%’ moiety was reproduced with two different amines as the P 3 ’ block. The striking similarity between results of the two independent sets of library screening indicated that inhibitors with 2 - aminoanthracene and 1 -aminoanthracene bind deformylase in similar modes.

Alternatively, the interactions between inhibitor Pi’, P%’ and P 3 ’ moieties and corresponding enzyme Si’, 8 2 ’ and S3 ’ sites are relatively independent of each other. The similar profile of two deconvolution outcomes is also a strong piece of evidence that the results are not due to error or artifacts, but are genuine and reliable.

The inhibitor identified by library deconvolution 3-mercapto-2-butylpropionyl-

Leucyl-2-antharceneamide, were synthesized from 3-acetylthio-2-butylpropionic acid and

Leu-2-anthraceneamide in similar procedure to that described for other sulfhydryl inhibitors. Due to the poor solubility of 2-aminoanthracene and certain synthetic intermediates, DMF had to be used to in two coupling reactions and it could be removed by ether / water extraction during work-up. The synthesis furnished the product in good yield whose identity was confirmed by NMR characterization. Activity assay was performed in order to confirm the results of library synthesis and screening, and to obtain

83 accurate inhibition constant. Unfortunately, due to the high hydrophobicity, the inhibitor precipitated out at the concentration above 5 pM. The poor solubility prevented detailed characterization and as a result, assay was conducted only with low concentrations of the inhibitor. The compound clearly inhibits deformylase and K, of 5.4 ± 0.3 pM was calculated from multiple data points with 1-5 pM of the inhibitions. The high potency of the selected inhibitor indicated that our design and execution of the library synthesis and screening is valid and sound.

The results produced from the above library work can be rationalized with the information generated by X-ray crystallographic work of deformylase-inhibitor complex

(Hao et al., 1999). The structure reveals that one important factor contributing to the binding affinity of the PCLNA inhibitor is its ability to fill the hydrophobic cavity located near deformylase metal site. The binding results in the reduction of the unfavorable surface energy and the increase of the favorable hydrophobic contacts. The selection of 3- mercapto-2-butylpropanoyl-Leucyl-2-antharceneamide, a highly hydrophobic molecule, may be explained by the highly hydrophobic nature of deformylase active site.

Some important hydrophobic interactions between the protein and inhibitor have been mapped out by crystal structure of the complex (Hao et al., 1999). First, at the Sf site, a hydrophobic side-chain analogous to methionine is important to fill a hydrophobic pocket. This agrees with our finding that inhibitors with norleucine and Phe analogs at this position bind best. Second, A large aromatic group is needed at the S 3' position to help to bury the hydrophobic Sf residue in a hydrophobic cleft of the protein. The best

84 Ps’ moiety identified from library screening, 2- or 1-aminoanthracene are indeed very large aromatic rings.

The crystal structure predicts that a variety of groups could be tolerated at S 2' position and previous substrate library screening showed that deformlyase does not have selectivity at this site. However, deconvolution of inhibitor library indicated that S 2' residue is also important for optimal recognition. Crystal structure shows that Leu, the S 2 ' residue of PCLNA inhibitor, is involved in van der Waals contact with protein. The fact that amino acids with big aliphatic side chains (Leu, Phe) were selected, while those with small ones were not supports the existence of this interaction and suggests its importance for high affinity binding.

It should be noted, however, that even this most potent inhibitor from the library turned out to be at least 10 fold less active than 2-n-butyl-3-mercaptopropionyl-Leu-pNA, while the only difference of the two inhibitors is the P3 ’ aromatic amines. From our previous studies, it is clear that the nitro group is essential for the high affinity of f-Met-

Leu-pNA. In this inhibitor library synthesis, some p-nitroaniline analogs were intentionally included as P 3 ’ biulding blocks. The fact that none of them emerged through activity screening further suggests the importance in high affinity binding of the relative location of, as well as distance between nitro and amine groups on the benzene ring. It is obvious that a large aromatic ring distinguishes the selected inhibitor from other molecules in the library, but it is not sufficient. To obtain inhibitors with activities surpassing that of 2-n-butyl-3-mercaptopropionyl-Leu-pNA, additional binding energy is

85 needed, possibly by incorporation of a small group at a proper location on the aromatic ring capable of binding to forming hydrogen bonds with other active-site residues.

Recently, it was reported that positively charged residues at P 2 ’ site are favored by deformylase (Meinnel et al., 1999). Work from this lab also showed that replacement of

Leu in 2-n-but>4-3-mercaptopropionyl-Leu-pNA with lysine increases inhibition potency by ~5 fold. This further indicated the importance of P 2 ’ residue in deformylase binding and the importance may be rationalized by the amphipathic nature of lysine and arginine side chains whose hydrophobic chain is able to interact with protein surface while the charged terminus with solvent water. In light of the highly hydrophobic nature predicted for deformylase inhibitors, charged residues are highly desired at P 2 ’ site, not only to enhance potency, but to boost solubility.

In summary, a deformylase inhibitor library was successfully synthesized and deconvoluted, leading to the discovery of some potent molecules. From the information generated about the structural features favored by deformylase active site, guidelines are suggested for future design of even more active deformylase inhibitors.

8 6 100 P deconvolution

p 70

2 3 4 5 6 7 8 9 10 11 12 13 14 15

110 P deconvolution 3

y 80

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Figure 3.12. Deconvolution of P 3 ’ position of inhibitor library.

87 110- P ' deconvolution 2

100 -

> ü < c m 2 (U Û.

Aia (D)AIa Leu Phe (a-me)Ala

P ' deconvolution

100 -

> tj < "c g

(HS)Gly (HS)Ala (HS)Leu (HS)nLeu (HS)Phe

Figure 3.13. Deconvolution with 2-aminoanthracene as P 3 ’ moiety.

8 8 110 p deconvolution 2

tj 90-

y 80-

Ala (D)Ala Leu Phe (a-me)Ala

P deconvolution

100-

. f

5 804 c g ÛL

(HS)Gly (HS)Ala (HS)Leu (HS)nLeu (HS)Phe

Figure 3.14. Deconvolution with I-aminoanthracene as P 3 ’ moiety.

89 CHAPTER 4

PRODRUG ACTIVATION BY PEPTIDE DEFORMYLASE

4.1 Introduction

The emergence of bacterial pathogens that are resistant to multiple classes of

existing antibiotics has created an urgent demand for new antibacterial agents with novel mechanisms of action (Neu, 1992; Davis, 1994; Spratt, 1994; Gold & Moellering, 1996).

Parallel to deformylase inhibitor design, an alternative approach to the development of antibacterial drugs was devised in this lab. In this approach, pharmacologically inactive prodrugs are selectively processed by bacterial deformylase to trigger the release of an active drug. Model prodrugs developed in this work involve a clinical antibiotic and anticancer drug 5-fluoro-2’-deoxyuridine, which is masked by acylation at 5'-OH group with N-formylated dipeptidyl units (Figure 4.1). When bacterial cells take up these compounds, their peptide deformylase will remove the N-formyl group, initiating a cyclization reaction (Gisin & Merrifield, 1972; Khosla et al., 1972; Rothe & Mazanek,

1974; Pedroso et al., 1986; Wipf et al., 1991) to release FdU. FdU is phosphorylated by thymidine kinase in vivo to generate 5-fluorodeoxyuridine 5’-phosphate, which in turn acts as an mechanism-based inhibitor of thymidylate synthetase (Santi et al., 1987).

90 O R 3 R 2 5 6 w

peptide deformylase

' Ï u r HgN O R 3 R 2 w OH

spontaneous cyclization O R1 NH 1 R2 .N ^ / A, r" R 3 N

^ 0 H w OH 58 5 9

a: R1-R2 = -CHsCHjCHj-. R3 = H b; R1-R3 = -CH 2CH 2CH 2-, R 2 = H c: R1 = CH 3. R 2 = GH2CH(CH 3)2. R3 = H d: R1 = H, R2 = R3 = CH3

Figure 4.1. Prodrug activation by peptide deformylase.

91 Since 5’-phosphorylation is necessary for FdU’s antimetabolic activity, acylation

of the 5’-OH is expected to prevent 5’-phosphorylation of FdU and render it biologically

inactive. Because of the unique presence of peptide deformylase in bacterial ceils, active drugs should be produced only inside the bacterial cells, thus minimizing the cytotoxicity of these prodrugs to the host cells.

4.2 EXPERIMENTAL PROCEDURES

4.2.1 Material and Methods

E. coli peptide deformylase was purified as described (Rajagopalan et al., 1997).

Staphylococcus epidermidis was kindly provided by Professor Neil Baker at Department of Microbiology, the Ohio State University. Fetal bovine serum which had been dialyzed to remove molecules with molecular weights <1000 was obtained from Sigma Chemical

Company (St. Louis, MO). FdU was from ICN Pharmaceuticals, Inc. (Costa Mesa, CA).

All other chemicals were purchased from Aldrich (Milwaukee, WT), Sigma, or Advanced

ChemTech, Inc. (Louisville, KY).

4.2.2 Synthesis

tBoc-(D)-Proline (60). t-BOC-(D)-Proline was prepared as described (Itoh et al.

1977). To a solution of (D)-proline (2.30 g, 20 mmole) and triethylamine (3.03 g, 30 mmole) in methanol/dioxane/HiO (30 ml / 10 ml / 20 ml) was added t-

92 butoxycarbonyloxyimino-2-phenylacetonitrile (5.42 g, 22 mmole). The reaction became homogeneous in 0.5 h and was continued for a total of 2 h at room temperature. It was then concentrated in vacuo and ICO ml of 5% NaHCO] was added. After extraction with benzene (50 ml x 3), the aqueous phase was acidified with citric acid to pH 2-3 and extracted with ethyl acetate 100 ml. The organic phase was dried over MgS 0 4 and evaporated to give 3.20 g of tBoc-(D)Proline as white solid (74.4% yield).

N-FormyI-(D)-Met (61) and N-formyl-(L)-Ala (62). N-Formyl amino acids were prepared by formylation of amino acids in a mixture of acetic anhydride and formic acid followed by recrystallization from water (Sheehan & Yang, 1958).

5’-(4% 4’-Dimethoxytrityl)-2’-deoxyuridlne (5’-DMT-dU, 64).

O 2'-Deoxyuridine (0.23 g, 1.0 mmole) was dried by co-

f[ evaporation with anhydrous pyridine (5 ml x 3) and then DMT-O-i dissolved in 10 ml of pyridine. To the solution was added 4, 4’-

V"^ dimethoxytrityl chloride (0.41 g, 1.2 mmole) and 4- OH dimethylaminopyridine (0.12 g, 1.0 mmole). The reaction mixture was stirred for 5 h at rt and then concentrated. Purification by silica gel column chromatography (elution with dichloromethane / methanol, 15/1) gave 0.53 g of white solid (quantitative yield).

93 5’-(4,4’-Dimethoxytrityl)-3’-(trimethyIsilylethoxymethyl)-2’-deoxyuridine

(S -DMT-3-SEM dU, 65).

O To a solution of 5’-DMT-dU (0.44 g, 0.83 mmole) in 5

f| ml of dichloromethane was added N, N-diisopropylethylamine DMT-O—V ^ N ' ^ 0 V n | (0.47 g, 2.83 mmole) and 2-(trimethylsilyl)ethoxymethyl

chloride (SEM-Cl, 0.62 g, 4.84 mmole). After 12 h at rt, the O-SEM reaction was concentrated and purified by silica gel column chromatography (elution with

0-3% methanol in dichloromethane) to give crude product.

t-Boc-Pro-dU (71).

O 2’-Deoxyuridine (1.14 g, 5.0 mmole) was

\ ^ 0 NH dehydrated by co-evaporation with pyridine (5 ml

\J X 3). The pretreated 2’-deoxyuridine along with t-

° wI Boc-Pro (1.08 g, 5.0 mmole) was desiccated over OH

P2 O 5 under vacuum for 1 h. To a solution of t-Boc-Proline in 15 ml of pyridine was added

2, 4, 6-triisopropylbenzenesulfonyl chloride (1.51 g, 5.0 mmole) with vigorous stirring.

After 1 h at room temperature, the reaction solution was added dropwise to a solution of

2’-deoxyuridine in 10 ml of pyridine that had been cooled to -10 °C. After stirring at this temperature for 24 h, the reaction was warmed to room temperature and solvent evaporated. Purification by silica gel chromatography (elution with 0-5% linear gradient of methanol inchloroform) gave 0.42 g of white solid (20% yield). *H NMR (200 MHz,

CDCI3 ): Ô 8.45-8.8 (m, IH), 7.44 (d, J = 8.1 Hz, 7.44), 6.23 (d, J = 6.4 Hz, IH), 5.70-5.87

94 (m, IH), 4.00-4.55 (m, 5H), 3.59-3.78 (broad, IH), 3.30-3.59 (m, 2H), 2.35-2.57 (m, IH),

2.10-2.35 (m, 2H), 1.70-2.10 (m, 3 H), 1.33-1.57 (m, 9H). Calctd mass: 425.4. F ASMS

(MH^: m/z 426.2.

5’-(N-Formyl-methionylprolyl)-2’-deoxyuridine (f-Met-Pro-dU, 68).

t-BOC-Pro-dU (0.10 g, 0.24 mmole) O ^ was dissolved in TEA / tMophenol (5 ml / 1

ml). After 0.5 h at room temperature, the

O solvent was evaporated and the residue

OH triturated with diethyl ether to give white solid. To a solution of N-formyl-Met (0.035 g, 0.20 mmole) and N-hydroxylsuccinimide

(0.023 g, 0.20 mmole) in 5 ml of tetrahydrofuran was added dicyclohexylurea (0.045 g,

0.22 mmole). After 1 h at rt, dicyclohexylurea precipitate was removed by filtration and the filtrate concentrated. The residue was dissolved in 5 ml of dichloromethane and transferred to the TEA salt of 5’-Prolyl-dU prepared above. After triethylamine (0.030 g,

0.3 mmole) was added, the reaction mixture was stirred for 2.5 h. Concentration and silica gel column chromatography (elution with chloroform / methanol, 15:1) gave 0.040

g of white solid (41% yield). ‘H NMR (200 MHz, CDCI3 ): 6 9.10 - 9.34 (s, IH), 8.10 -

8.25 (m, IH), 7.40 - 7.58 (m, IH), 7.08 - 7.20 (m, IH), 6.10 - 6.30 (m, IH), 5.78 (t, J =

7.5 Hz, IH), 4.95 - 5.15 (m, IH), 4.50 - 4.68 (m, IH), 4.23 - 4.50 (m, 2H), 4.01 - 4.23

(m, 2H), 3.60 - 4.00 (m, 2H), 2.35 - 2.67 (m, 3H), 1.75 - 2.35 (m, lOH). Calctd mass:

484.5. FABMS (MH"): m/z 485.3.

95 General Procedure for Synthesis of t-BOC-Pro-FdU, t-BOC (D)Pro-FdU, t-

BOC-(N-methyl)Leu-FdU, t-BOC-(a-methyl)AJa-FdU.FdU (0.49 g, 2.0 mmol) and t-

BOC-amino acid (2 mmol) were dried over P 2O5 for 3 h. The predried FdU and

txiphenylphosphine (0.52 g, 2.0 mmol) were dissolved in 25 ml of freshly distilled dioxane. With vigorous stirring, the predried t-Boc-amino acid and diethyl

azidodicarboxylate (0.35 g, 2.0 mmol) dissolved in 5 ml of dioxane were added dropwise over a period of 30 min. After stirring for 2 h at room temperature, the reaction mixture was concentrated in vacuo and subjected to silica gel flash chromatography. The column was eluted with a gradient of 0-5% methanol in chloroform. All products were crystalline white solids. The yields were 76 - 89 %.

t-BOC-Pro-FdU (78a).

O

NH ‘H NMR (200 mHz, CDCI3): 5 9.68 - 10.07

(m, IH), 7.47 - 7.69 (m, IH), 6.24 (t, J = 6.2 Hz,

IH), 4.06 - 4.54 (m, 5H), 3.30 - 3.58 (m, 2H), 2.36 OH

- 2 . 5 9 (m, IH), 1.78 - 2.36 (m, 6H), 1.30 - 1.52 (m, 9H). Calctd mass: 443.4. FABMS

(MH^: m/z 444.2.

96 t-BOC-(D)Pro-FdU (78b).

*H NMR (200 mHz, CDCI 3): 5 9.71 (s,

O IH), 7.52 - 7.86 (m, IH), 6.25 (t, J = 5.8 Hz, IH),

O N _ / IH), 4.40 - 4.60 (m, IH), 4.12 - IT % n N O O \ / O g 4.38 (m, 2H), 3.97 - 4.13 (m, IH), 3.84 (s, IH),

3.29 -3.64 (m, 2H), 2.34 - 2.56 (m, IH), 1.77 - OH

2 . 3 4 (m, 5H), 1.30 - 1.52 (m, 9H). Calctd mass; 443.4. FABMS (MH^: m/z 444.2.

t-BOC-(N-methyl)Ieucine-FdU (78c).

O ‘h NMR (200 mHz, CDCI 3): 5 8.71 -

1 ° 'i T 8.93 (m, IH), 7.51 (d, J= 6.0 Hz, IH), 6.21 (t, .1 = n N O X £ O 5.2 Hz, IH), 4.04 - 4.77 (m. 5H), 3.14 (s, IH),

f r ~ 2 . 7 7 - 2.92 (m, 3H), 2.36 - 2.57 (m, IH), 2.03 - I OH

2 . 2 2 (m, IH), 1.48 - 1.90 (m, 3H), 1.44 (s, 9H), 0.95 (d, J = 5.6 Hz, 3 H), 0.92 (d, J = 5.9

Hz, 3H). Calctd mass: 473.5. FABMS (MH^: m/z 474.3.

t-BOC-(a-methyi)alanine-FdU (78d).

O ‘h NMR (200 mHz, CDCI 3): Ô 9.38 (s,

O ; IH), 7.74 (d, J = 6.2 Hz, IH), 6.15 - 6.29 (m,

N X ° IH), 5.14 (s, IH), 4.52-4.69 (m, IH). 4.34-4.51 O ' W I (m, IH), 4.14 - 4.32 (m, IH), 4.01 - 4.14 (m, IH), OH

97 3.76 (s, IH), 2.37 - 2.56 (m, IH), 2.08 - 2.27 (m, IH), 1.50 (s, 3H), 1.49 (s, 3H), 1.40 (s,

9H). Calctd mass: 431.4. FABMS (MH^: m/z 432.3.

General Procedure for Synthesis of f-Met-Pro-FdU, f-Met-(D)Pro-FdU, f-

Met-(a-methyi)AIa-FdU. Compound 78a, 78b, or 78d (1.0 mmol) was dissolved in 4 ml of TFA and 1.0 ml of anisole and incubated at room temperature for 15 min. TEA was then evaporated in vacuo and the resulting residue was triturated with diethyl ether (4 x

10 ml) to afford white solids.

N-formylmethionine (0.18 g, 1.0 mmol) and N-hydroxysuccinimide (0.12 g, 1.0 mmol) were dissolved in 5 ml of tetrahydrofuran. After addition of dicyclohexyl- carbodiimide (DCC) (0.23 g, 1.1 mmol), the reaction was allowed to proceed for 1 h. The precipitate was removed by filtration and the filtrate was concentrated in vacuo to give N- formyl-methionine N-hydroxysuccinimide ester 72. Ester 72, the TFA salt of the amino acid-FdU conjugates from above, and triethylamine (1.2 mmol) were dissolved in 10 ml of dichloromethane and the mixture was stirred for 1 to 3 h. The reaction products were concentrated in vacuo and purified by silica gel flash chromatography. The column was eluted with methanol/dichloromethane (1:15). The overall yield (from 78) was 96%, 85% and 65%, respectively, for 56a, 56b, and 56d. All products were hygroscopic crystalline white solids.

98 f-Met-(L)-Pro-FdU (56a).

g / ‘h NMR (200 mHz, DMSO): Ô 11.89 O (s, IH), 8.45 (d, J = 8.1 Hz, IH), 7.98 (s, IH), NH 7.90 (d, J = 6.9 Hz, IH), 6.13 (t, J = 6.7 Hz,

IH), 5.47 (s, IH), 4.67 - 4.88 (m, IH), 4.10 -

4.43 (ra, 4H), 3.83 - 3.98 (m, IH), 3.51 - 3.83

(m, 2H), 2.35 - 2.56 (m, 2H), 1.59 - 2.30 (m, IIH). Calctd mass: 502.5. FABMS (MFC): m/z 503.2.

f-Met-(D)-Pro-FdU (56b).

‘H NMR (200 mHz, DMSO): Ô 11.87 O (d, J = 5.0 Hz, IH), 8.38 (d, J = 8.7 Hz, IH), I N '^ O 7.82 - 8.10 (m, 2H), 6.14 (t, J = 6.7 Hz, IH), O ° 5.44 (s, IH), 4.68 - 4.90 (m, IH), 4.07 - 4.48

(m, 4H), 3.80 -4.02 (m, IH), 3.48 - 3.80 (m,

2H), 2.33 - 2.55 (m, 2H), 1.57 - 2.33 (m, 1 IH).

f-Met-(a-methyl)alanine-FdU (56d).

O ‘H NMR (200 mHz, DMSO): Ô 11.86

(d, J = 5.0 Hz, IH), 8.42 IH), 8.21 (d, J =

O ^ O g ° 8.3 Hz, IH), 7.98 (s, IH), 7.77 - 7.93 (m, IH),

^ 6.01 - 6.18 (m, IH), 5.38 (s, IH), 4.32 - 4.53 OH 99 (m, IH), 3.96 -4.32 (m, 3H), 3.79 - 3.95 (m, IH), 2.39 (t, J = 7.9 Hz, 2H), 2.03 - 2.23 (m,

2H), 2.00 (s, 3H), 1.60 - 1.95 (m, 2H), 1.35 (s, 3H), 1.34 (s, 3H). Calctd mass: 490.5.

FABMS (MHT): m/z 491.4.

f-Met-(N-methyi)Ieucine-FdU (56c).

Q t-BOC-(N-methyI)Ieucine-PdU (78c)

. Q was deprotected with TFA / anisole as I ti [1^ ^ described above. The resulting TFA salt of

' (N-methyl)leucine-FdU (0.24 g, 0.50 mmol), OH N-formylmethionine (0.088 g, 0.50 mmol), and N, N-diisopropylethylamine (0.16 g, 1.2 mmol) were dissolved in 5 ml acetonitrile.

After addition of 0-benzotriazol-l-yl-N,N,N\N’-tetramethyluronium hexafluorophosphate (HBTU) (0.19 g, 0.50 mmol), the reaction was stirred for 30 min at room temperature followed by concentration in vacuo. The residue was purified twice by silica gel flash chromatography. The column was eluted with methanol/dichloromethane

(1:15). The product was obtained as a hygroscopic, crystalline white solid at 31% overall yield (from 78c). ‘H NMR (200 mHz, DMSO): Ô 11.87 (d, J = 5.0 Hz, IH), 8.45 (d, J =

8.1, IH), 7.98 (s, IH), 7.88 (d, J = 7.0, IH), 6.12 (t, J = 6.8 Hz, IH), 5.42 (s, IH), 4.84 -

5.21 (m, 2H), 4.06 - 4.37 (m, 3H), 3.77 - 3.98 (m, IH), 2.88 - 3.06 (m, 3H), 2.30 - 2.49

(m, 2H), 1.30 - 2.30 (m,lOH), 0.65 -0.98 (m, 6H). Calctd mass: 532.6. FABMS (MFT): m/z 533.4.

1 0 0 Synthesis of Control Prodrug Analogues. Prodrug analogues used for control tests of bacterial growth inhibiton including formyl-(D)-Met-Pro-FdU, acetyI-Met-(a- methyl)alanine-FdU and formyl-Ala-(a-methyl)Ala-FdU were synthesized as the general procedure described for prodrugs.

Fomiyl-(D)-Met-Pro-FdU (79).

Calctd mass: 502.5. FABMS (MH^ 503.3.

Acetyl-Met-(a-methyl)alanine—FdU (80).

O 'H NMR (200 MHz, DMSO): Ô 11.89

O 7.75-8.12 (m, 2H), 6.11

0 ^ 0 N O (t, J = 6.7 Hz, IH), 5.39 (d, J = 4.3 Hz, IH),

4.00-4.30 (m, 4H), 3.75-4.00 (m, IH), 2.27- OH 2.55 (m, 2H), 1.55-2.27 (m, lOH), 1.35 (s,

3H), 1.33 (s, 3H). Calctd mass: 504.5. FABMS (MH^ 505.3.

F0 rmyI-Ala-(a-methy 1) Ala-FdU (81).

O ‘ h NMR (200 MHz, DMSO): 5 11.8

II I H M H T (broad, IH), 8.36 (s, IH). 8.18 (d, J = 7.9 IH),

H ^ / \ 7.92 (s, IH), 7.86 (d, J = 7.0 Hz, IH), 6.10 (t, J

= 6.1 Hz, IH), 5.38 (d, J = 4.2 Hz, IH), 3.97- OH

1 0 1 4.50 (m, 4H), 3.80 - 3.95 (m, IH), 1.95-2.20 (m, 2H), 1.10-1.50 (m, 9H). )- Calctd mass:

430.4. FABMS (MtT) 431.2.

General Procedures for Oxidation of Prodrugs. Sulfur of the methionine residue in prodrugs was oxidized to sulfone to generate compounds used as controls for tests of bacterial growth inhibition. To a solution of 0.2 mmole prodrug in 5 ml of chloroform was added 0.8 mmole of 3-chloroperoxybenzoic acid (m-CPBA). After stirring at room temperature for Ih, the reaction solution was concentrated and purified by silica gel column chromatography (elution with dichloromethane/methano 1, 10:1). All compounds were isolated as a single spot on TLC plates, characterized by NMR or mass spectrometry.

Oxidized formyl-Met-Pro-FdU (82).

Calctd mass: 534.5. FABMS (MFT^: m/z 535.3.

Oxidized formyl-Met-(D)-Pro-FdU (83).

^ , 'H NMR (200 MHz, DMSO): Ô 11.86 V*o o (d, J = 5.1 Hz, IH), 8.51 (d, J = 8.2 Hz, IH), NH

6.5 Hz, IH), 4.70 - 4.95 (m, IH), 4.10 - 4.45

OH (m, 4H), 3.80 - 4.02 (m, IH), 3.55 - 3.78 (m.

102 2H), 3.00 - 3.25 (m, 4H), 2.96 (s, 3H), 1.75 - 2.85 (m, 6H). Calctd mass: 534.5. FABMS

(MH^: m/z 535.3.

Oxidized formyl-lVIet-(a-methyl)alanine-FdU (84).

Calctd mass: 506.5. FABMS (MFC): m/z 507.2.

5’-(4,4’-Dimethoxytrityl)-5-fluoro-2’-deoxyuridine (5’-DMT-FdU, 85).

O 5-FIuoro-2’-deoxyuridine (0.221 g, 0.90 mmole) was

NH dried by co-evaporation with anhydrous pyridine and desiccation DMT-O over phosphorous pentoxide. The dried FdU and DMT chloride

(0.610 g 1.80 mmole,) in 8 ml of pyridine were incubated for 12 OH h at room temperature. The reaction was concentrated and subjected to silica gel flash column chromatography (0 - 5% MeOFF in CFFiCli). The unreacted FdU was recovered and reused. Crystalline solid (0.260 g) with slightly yellow color was obtained (52% yield). ‘H NMR (200 mHz, CDCI3 ): 5 9.41 (s, IFF), 7.83 (d, J = 6.0 Hz, IH), 6.75 - 7.42

(m, 13 H), 6.30 (t, J = 5.8 Hz, IFF), 4.45 - 4.63 (m, IFF), 4.03 - 4.14 (m, IH), 3.78 (s, 6H),

3.41 (m, 2H), 2.40 - 2.66 (m, 2FF), 2.14 - 2.35 (m, IH).

103 5’-Dimethoxytrityl-3’-(t-BOC-proIyI)-5-fluoro-2’-deoxyuridine (86).

O tBOC-ProIine (0.236 g, 1.10 mmole) and 4-

dimethylaminopyridine (0.268 g, 2.20 mmole) in 4 _ " > o N O U n| ml of CH 2CI2 was cooled to 0 °C and 2, 4, 6 - XT^ ^ _Q triisopropylsulfonyl chloride (0.332 g, 1.10 mmole) O in 2 ml of CH 2CI2 was added dropwise. The reaction

solution was warmed to room temperature over 1 h and 5’-DMT-FdU (0.494 g, 0.90

mmole) in 2 ml of CH 2 CI2 was added. After another 1 h at room temperature, the reaction

mixture was diluted with 20 ml of CH 2CI2 and extracted with 20 ml of 5 % NaHCO]

twice. The organic phase was dried over MgSO^ and concentrated. Silica gel column chromatography (elution with 0 - 5% MeOH in CH 2CI2) gave off-white solid 0.581 g

(71% yield). 'H NMR (200 mHz, CDCI 3): 5 8.85 - 9.12 (m, IH), 7.80 -7.90 (m, IH), 6.75

-7.45 (m, 13H), 6.25 -6.40 (m, IH), 5.40 - 5.65 (m, IH), 4.08 - 4.40 (m, 2H), 3.78 (s, 6 H),

3.25 - 3.68 (m, 4H), 2.10 - 2.70 (m, 3H), 1.77- 2.05 (m, 3H), 1.28 - 1.54 (m, 9H). Calctd

mass: 745.4. FABMS (MH^: m/z 746.1.

3’-(N-formyl-methionylprolyI)-FdU (3’-f-Met-Pro-FdU, 88).

O Compound 8 6 (0.381 g, 0.51 mmole) was

dissolved in TFA / anisole / MeOH (4.5 ml / 1.5

ml / 0.2 ml). After 15 min at room temperature.

\ q ' solvents were evaporated. Trituration of the O residue with diethyl ether gave crystalline white

104 solid. To the product was added 3 ml of acetonitrile and triethylamine (0.060 g, 0.60 mmole). AAer stirring for 5 min, the TFA salt of 3’-prolyl-FdU dissolved. Crude N- hydroxysuccinimide ester of N-formy 1-methionine 72 (0.50 mmole) in acetonitrile was added. After 2.5 h at room temperature, the reaction was concentrated and subjected to silica gel column chromatography (elution with MeOH / CH 2CI2 , 1:15) to give 0.172 g of white solid (69% yield). 'H NMR (200 mHz, DMSO): Ô 11.80 - 12.05 (m, IH), 7.84 -

8.70 (m, 3H), 6.02 - 6.28 (m, IH), 5.10 - 5.50 (m, 2H), 4.65 - 4.88 (m, IH), 3.54 - 4.55

(m, 6H), 1.60 - 2.60 (m, 13H). Calctd mass: 502.5. FABMS (MH^: m/z 503.3.

Oxidized 3’-(formyi-Met-Pro)-FdU (89). The title compound was prepared by oxidation of 88 with m-CPBA as described previously. Calctd mass: 534.5. FABMS

(MH^: m/z 535.3.

4.2.3 Kinetic and Biological Assays

Deformylase Assay.The activity of E. coli deformylase toward prodrugs 56a-d was determined using formate dehydrogenase-coupled assay as described previously.

Peptide deformylase reactions were carried out with 100 pM to 2 mM substrates in a 50 mM sodium phosphate buffer (pH 7.4) with a total reaction volume of 500 pi. The amounts of E. coli peptide deformylase used in each reaction was 1.3, 33.5, 6.7 and 0.67 pg, respectively, for compounds 56a-d. The reactions were allowed to proceed for 1 min before they were quenched by addition of 10 pi of 10 mM ethanedithiol. After 10 pi of

100 mM NAD^ and 0.25 unit of Pseudomonas formate dehydrogenase were added, the

105 reaction volume was adjusted to 1.0 ml by adding distilled H 2O and incubated at 0 °C for

20 min. The absorbance of each reaction at 365 nm was determined and concentration of formate generated by peptide deformylase reaction was calculated using the standard line.

The substrate to product conversion was kept at <20%.

Cyclization Kinetics. Prodrugs 56a-d (1 mM) were completely deformyiated within 1-5 min with 67 pg of peptide deformylase in 20 mM sodium phosphate (pH 7.4) in a total reaction volume of 100 pi. Cyclization of the deformyiated intermediates (57a- d) was allowed to proceed at room temperature for various amounts of time and 10 pi aliquots of the reaction mixture were withdrawn and quenched by the addition of 10 pi of

100 mM TFA. The terminated reaction was injected onto a reversed-phase HPLC instrument equipped with a Cjg column and was eluted with a linear gradient of acetonitrile (0% for 5 min and then 0 - 40% in 20 min) in H 2 O (containing 0.05% TFA).

The absorbance at 260 nm (due to FdU) was monitored. The ratio of uncyclized intermediate (57a-d) to the total amount of the intermediate before cyclization (A/Ao) can be calculated by equation:

A/Ao = Iw-X-FdU / (Iw-X-FdU + Udu) where Ivi-x-Fdu is the integrated area under the peak corresponding to 57a-d and Ipdu is the integrated area under the FdU peak. Plot of ln(A/Ao) vs. time generated a series of straight lines. The half-lives (ti/ 2) of 57a-d were calculated from the slopes of these lines

(k): ti/2 = 0.693/k.

106 Stability Tests of Prodrugs. Prodrugs S6a-d (final 100 fj.M) were dissolved in

100 (j1 of 20 mM sodium phosphate, pH 7.4 or the buffer plus 5% (v/v) freshly thawed

fetal bovine serum. The resulting solution was incubated at 37 °C for 13-18 h and

analyzed by reversed-phase HPLC using the gradient conditions described above. The

percentage of degradation was quantitated by comparing the integrated areas under the

individual peaks (monitored at 214 nm and 260 nm).

In Vitro Activation of Prodrugs by Cell Lysate. An overnight culture of E. coli

BL21(DE3) cells was diluted 100-fold into 500 ml of fresh LB medium and was grown at

37 °C to ODeoo = 0.67. The cells were collected by centrifugation (10 min 5000 rpm),

resuspended in 10 ml of a 50 mM sodium phosphate solution (pH 7.0), and passed through a french pressure cell twice (set at 12,000 psi). The crude lysate was centrifuged at 15,000 rpm for 10 min and the clear supernatant was aliquoted, frozen in liquid nitrogen, and stored at -70 °C until use. For prodrug activation, typically 10 pi of 10 mM prodrug solution was added to 100 pi of the crude cell lysate. The resulting mixture was

incubated at room temperature for 60 min and then analyzed by reversed-phase HPLC as described previously.

Cell Growth Inhibition. An overnight culture of E. coli or Staphylococcus epidermidis cells was diluted into 5 ml of fresh LB medium in 15-ml culture tubes.

Various concentrations of prodrugs and control compounds were added at the time of dilution or at mid-log phase of cell growth. The resulting culture was incubated in a

107 rollerdrum at 37 °C for 4-10 h and the cell density was monitored on a UV-vis

spectrophotometer at 600 nm.

4.3 RESULTS

4.3.1 Design and Synthesis of Prodrugs

5-FIuoro-2’-deoxyuridme (FdU) is an expensive reagent and the natural

nucleoside 2’-deoxyuridine was used in the initial attempt to work out the general

conditions for prodrug synthesis. Proline was chosen as the second residue in the first

prodrug to be made because of the lower cost and better-established chemistry of natural

amino acids. Illustrated in Figure 4.2 was the original synthetic scheme. There were two —

OH groups in dU and it was a concern that the two hydroxyl groups may compete in ester

bond formation with amino acid moiety. In order to block 3’-OH, dU was first treated

with DMT-chloride to protect only 5’-hydroxyl group as an ether. The free 3’-OH was

then blocked as a different type of ether when the resulting compound 64 was treated with

SEM-chloride (Lipshutz & Tegram, 1980). The DMT ether o f 5’-0H was readily

unmasked with trichloroacetic acid. In the subsequent steps that were not carried out, the

5’-OH of the 3’-protected dU would be coupled to tBoc-Pro. After deprotection of proline

amino group and coupling to formyl methionine, the 3’-ester could be cleaved with

tetrabutylammonium fluoride to afifurd the final prodrug. Although not technically

challenging, the reactions were not trivial and in most of the steps column purification was required to separate product from excess or unreacted reagents. Furthermore, yields

108 of some of the steps were not perfect as would be expected for such common reactions.

As a result, the synthetic route was reexamined for modification. In this relatively short scheme, there were a total of four steps of protection and deprotection of 5’- and 3’- hydroxyl groups. They were likely the ones that can be possibly cut out in order to improve the efficiency.

Since numerous prodrug analogs were needed for biological studies, a great amount of time and effort would be saved if the synthesis of each prodrug was to be shortened by a few steps. With 5’- hydroxyl group of dU being a primary alcohol while the 3’- one secondary, it was reasonable to expect that different reactivity of the two groups would offer certain degree of selectivity in the estérification. The above synthetic route was thus abandoned and directe estérification of 5’-OH of dU was attempted.

Coupling of dU and tBoc-Pro with DCC gave a messy reaction and low yield, a result an apparent low selectivity. To achieve specific estérification of 5’-OH with high yield, the coupling reagent was critical. The t-Boc amino acid was then activated with 2,

4, 6-triisopropylbenzenesulfonyl chloride (TIPS-Cl) to form a mixed anhydride wherein the steric hindrance imposed by two isopropyl groups on the benzene rings was expected to lead to a better discrimination of primary and second alcohols. Indeed, desired product

5’-(t-Boc-Pro)-dU was obtained by this method. The final prodrug was produced by amide bond formation of deprotected proline to N-hydroxysuccinimide ester of methionine. This activated ester was used to minimize racemization of methionine during the reaction (Figure 4.3).

109 NOTE TO USERS

Page(s) missing in number only; text follows. Microfilmed as received.

110

This reproduction is the best copy available.

UMI NH NH NH

HO D M T -0 DMT-O DMT-CI, DMAP SEM-CI

pyridine (i-pr)zNEt OH OH O-SEM 6 3 6 4 6 5

NH NH CCUCOOH tB O C -P ro HO

66 O -SE M 6 7 O SE M

NH TFA coupling reagent N H O

68 OH

Figure 4.2. Initial synthesis of f-Met-Pro-dU.

I ll s-ci

Y O OH O 69 70

NH 2'-deoxyuridine

; ^ OH

NH TFA 72 A H N H ELN O

68 OH

Figure 4.3. Revised synthesis of f-Met-Pro-dU.

A problem was encountered when the above methodology developed was applied to synthesize prodrug containing (a-methyl)alanine as the second residue. Estérification with TIPS-Cl as coupling reagent did not produce the desired product, instead, 5’-(2,4, 6- triisopropylbenzenesulfonyl)-FdU was isolated as the major product (Figure 4.4). It was likely that the mixed anhydride did not form at all when tBoc-(a-methyl)Ala was treated

112 with TIPS-Cl due to the steric hindrance of a-carbon of (a-methyl)alanine , and that

TIPS-Cl simply reacted with FdU when it was added later. It was also possible that the anhydride did form, but the steric hindrance of a-carbon of (a-methyl)alanine prevented a nucleophilic attack of carboxyl carbon by 5'-OH of dU, which rather reacted with sulfonyl in the mixed anhydride.

S-CI O H H N N OH O- O pyridine O 73 74

NH NH H N O O

75 OH 76 OH

Figure 4.4. Attempted synthesis of tBoc-(methyl)Ala-dU.

It was later found that Mitsunobu condensation was a method of choice for estérification of our purpose (Mitsunobu, 1981). It not only produced carboxylic esters with high yields and selectivity, but had general utility for various amino acids. The final

113 synthesis of all prodrugs was thus established (Figure 4.5). After the chemistry was confirmed for all four designed prodrugs 56a-d using dU. the reactions were repeated with more costly reagent FdU. Following this more expedient route, prodrugs were synthesized in overall yields of 25-73% from FdU. While this is a facile method to prepare prodrugs, there were two major drawbacks. First, despite the care exercised to suppress racemization of methionine in the last step of coupling, it still occurred for all four prodrug compounds. The extent varied greatly depending on the identity of the second amino acid in the specific prodrug. Second, all the prodrugs were extremely polar molecules and mixed solvents of chloroform or dichloromethane with methanol were used as eluant for column purification. As a result separation of prodrugs from other components (triethylamine and diisopropylethylamine in particular) following last step was troublesome.

In hindsight, it seems that prodrugs could also be made by direct Mitsunobu condensation of FdU with the formyl dipeptides, provided that the peptidyl units can be easily made. From my own experience, f-Met-Pro seems to be more difficult to synthesize than formyl-Met- dipeptides with other residues at the second position.

All compounds synthesized were characterized with NMR spectroscopy. It should be pointed out that some prodrugs might exist in different conformers that are in slow equilibrium. It was noticed that tBoc hydrogen atoms in t-Boc-Pro and t-Boc-(N- methyl)leucine were two distinctive, well-separated peaks in contrast to the one single peak for tBoc in other tBoc-amino acids. The same factor also complicated the NMR

114 spectra of some of final prodrug compounds. In addition, the identity and purity of all compounds was confirmed mass spectrometry and HPLC analysis.

O

R1 O NH OH + HO^ ^ N O > f V " ' DEAD, PPha I O R3 R2 '"VoJ y 77 OH 59

O

R1 O NH A. Et3 N, for a, b, d ® • V " " Or \ DOR3 DOR2 Y N f-Met, HBTU, Et3 N for c r 78 O

O R1 N H O R3 R2 a: R1-R2 = -CHzCHgCHz-, R3 = H y b: R1-R3 = -CH 2 CH2 CH2 -, R2 = H 56 OH c; R1 = CH3 , R2 = CH2 CH(CH3 )2 , R3 = H

d: R1 = H. R2 = R3 = CH3

Figure 4.5. Final synthetic scheme of prodrugs.

115 4.3.2 Deformylase assay of Prodrugs

The assay showed that all prodrugs 56a-d were deformylase substrates with non­ saturation kinetics in the concentrations measured. The values of kcat/Kw range from 750 to 2.9 X 10^ M‘'s‘* (Table 4.1). It is noteworthy that these activities are comparable to those reported for short N-formylmethionyl peptide substrates (Rajagopalan et al., 1977).

Deformylation in vivo is believed to be co-translational and therefore the physiological substrates of peptide deformylase are likely unfolded peptides (Pine, 1969). Our previous work has shown that amino acid residues beyond the fourth position from the N-terminus have little effect on the deformylase reaction rate (Rajagopalan et al., 1977). These results suggest that prodrugs 56a-d should compete favorably with the endogenous substrates in bacterial cells for deformylase and be rapidly deformylated.

116 5 60 4 Cû cû a 40 e 3 2 : 20 î > 1 0 0 0.5 1 1.5 2 0 0.5 1 1.5 2 [f-Met-(L)Pro-FdU], mM [f-Met-(D)Pro-FdU], mM

8 150

= 120 6 Sû =0 4 60

0 0 0.5 1 1.5 2 0 0.5 1 1.5 2 [f-Met-(N-methyl)Leu-FdU], mM [f-Met-(a-methyl)Ala-FdU], mM

Figure 4.6. Plot of deformylation rate vs. prodrug concentration.

117 43.3 Cyclization of Deformylated Prodrugs

Following deformylation, the intermediates 57a-d undergo facile, spontaneous intramolecular aminolysis to release diketopiperazines 58a-d (retention time = 17.0 min for 58a) and FdU (retention time = 7.3 min) (Figure 4.7). The release o f FdU displays first-order kinetics (ti/? = 2-51 min), consistent with an intramolecular reaction (Table

4.1). Diketopiperazine 58a was collected from HPLC runs and its identity confirmed by mass spectrometry.

The incorporation of unnatural amino acids into 56b-d exerts some interesting effects on the kinetics of deformylation and cyclization. The use of a D-proline reduces the deformylation rate by 16-fold (relative to 56a) but increases the cyclization rate by

3.2-fold (ti /2 - 4.7 min) (Gisin & Merrifield, 1972). Similarly, N-methylleucine (56c) slows down the deformylase reaction by 7.3-fold, but increases the cyclization reaction rate by 7.3-fold {l\a = 2.1 min) (Khosla et al., 1972). On the other hand, the use of a- methylalanine (56d) increases the deformylase reaction rate by 2.4-fold, but reduces the cyclization rate by 3.4-fold (ti /2 = 51.1 min). These results suggest that it may be possible to generate prodrugs with desired activation rates by varying the structure of the penultimate amino acid.

118 10 15 20 25 Retention Time (min)

Figure 4.7. HPLC analysis of activation of prodrug 56a. At time 0, 67 pg of deformylase was added to ~1 mM 56a in 100 pL of 20 mM sodium phosphate (pH 7.4). At various times, 10 pL aliquots were withdrawn, quenched by the addition of an equal volume of 100 mM TFA, and loaded onto an HPLC instrument (C|g column; monitored at 214 nm). a, t = 60 min control without deformylase; b, t = 1 min; c, t = 15 min; d, t = 60 min; and e, FdU standard. The retention times of 56a, 57a, 58a, and FdU are 19.7, 18.2, 17.0, and 7.3 min, respectively. The small peak resistant to deformylation (t = 20.1 min) is f- (D)Met-Pro-FdU.

119 -0.5

1 ° c

-2.5

0 20 40 60 80 100 120 140 160 Time (min)

f-Met-(L)Pro-FdU -Ar— f-Met-(D)Pro-FdU -■— f-Met-(N-methyl)Leu-FdU -♦— f-Met-(a-methyl)Ala-FdU

Figure 4.8. First order kinetics of cyclization of deformylated prodrugs. Ao, concentration of 57a-d at time zero (before cyclization); A, concentration of 57a-d at time t.

120 Deformylation kcai/K^ Cyclization t'/, prodrug (X 10^ M'‘ s ') (min)

56a 12.1 ±0.4 15.2

56b 0.75 ± 0.03 4.68

56c 1.65 ±0.06 2.07

56d 28.9 ± 1.4 51.1

Table 4.1. Summary of kinetic data of prodrug activation (at pH 7.4).

121 4.3.4 Chemical and Enzymatic Stability of Prodrugs

Prodrug 56a. which contains an L-proline as the penultimate residue, is relatively stable under neutral aqueous conditions (20 mM sodium phosphate, pH 7.4); incubation for 18 h at 37 °C led to approximately 16% degradation, as judged by the amount of released free FdU (Figure 4.9). This is presumably due to a combination of nonenzymatic hydrolysis of the ester linkage and background deformylation followed by cyclization.

Prodrug 56a was then tested for stability against proteolysis by adding to the phosphate buffer 5% (final) fetal bovine serum. Significant degradation (-55%) into FdU and two other species was observed after 18 h of incubation (Figure 4.9). Judging from the elution pattern of the products, the enzymatic cleavage appears to occur primarily at the ester linkage to produce FdU (t = 6.5 min) and formyl-Met-Pro (t = 13.9 min), although detailed characterization of the products was not conducted.

To improve the stability of prodrugs against potential protease action, unnatural amino acids were employed to construct prodrugs 56b-d (Figure 4.1). An L-methionine was still used as the N-terminal residue, because the peptide deformylase is highly specific for an L-methionine at this position (Adams, 1968; Livingston, 1969;

Rajagopalan et al., 1997). At the penultimate position, unnatural amino acids D-proline

(56b), N-methyl-L-leucine (56c), and a-methylalanine (56d) were substituted for the L- proline. Previous studies have shown that peptide deformylase has broad substrate specificity at the penultimate position and can tolerate D-amino acids (Rajagopalan et al.,

1997).

122 8 12 16 20 Retention time (min)

Figure 4.9. HPLC analysis (monitored at 214 nm) of prodrug degradation (18 h at 37 °C). a, 56a + 20 mM phosphate buffer (pH 7.4); b, 56a + phosphate buffer + 5% fetal bovine serum; c, 56d + phosphate buffer; d, 56d + phosphate buffer + 5% serum. The smaller peak at retention time of 14.7 min (in c and d) is due to racemization of 56d at the methionyl residue during synthesis.

123 As expected, the use of D-proline and N-methyl-L-Ieucine rendered prodrugs 56b and 56c much more stable against protease action; incubation for 13 h at 37 °C in the presence or absence of 5% fetal bovine serum resulted in essentially the same pattern of degradation (data not shown). Unfortunately, for reasons not yet clear, 56b and 56c are more susceptible to nonenzymatic hydrolysis of the formyl group than the parent compound, 56a. Incubation in 20 mM phosphate buffer for 13 h at 37 °C resulted in 64% and 54% degradation for 56b and 56c, respectively. Prodrug 56d, which contains an a- methylalanine at the penultimate position, shows dramatically improved stability against both nonenzymatic and enzymatic hydrolysis (Figure 4.9c and d). After 18 h of incubation at 37 °C in the presence or absence of 5% fetal bovine serum, only a trace amount of FdU was observed. Integration of the HPLC peaks gives an estimated upper limit of ~I% of degradation. In addition, when prodrugs 56b-d were treated with crude E. coli cell lysate, no side reaction other than deformylation and cyclization was observed.

4.3.5 Activation of Prodrugs by E. coli Cell Lysate

Compounds 56a-d were treated with crude bacterial cell lysate prepared from wild-type BL21(DE3) cells that had been grown to the mid-log phase and the cleavage products then analyzed by reversed-phase HPLC. All four compounds were rapidly deformylated by the endogenous deformylase followed by cyclization. Since 56d was most stable against background or proteolytic hydrolysis, it was characterized in more detail. As shown in Figure 4.10, treatment of Id with crude cell lysate for 1 h resulted in complete disappearance of the Id peak (t = 19.0 min). In the mean time, a new peak

124 appeared at t = 15.1 min when monitored at 214 nm, which was not visible when monitored at 260 nm. This species is assigned as diketopiperazine 58d, as it is not derived from the cell lysate, coelutes with 58d derived from treatment of 56d with purified deformylase, and has a similar retention time to 58a, which had been carefully characterized by mass spectrometry. Note that there is no sign of any degradation of 56d by cellular enzymes other than the desired deformylation and cyclization reactions.

Interestingly, a peak corresponding to free FdU (which has a retention time of 7.3 min under the conditions) was not observed (Figure 4.10). Careful examination of the HPLC chromatograms revealed an increase in intensity for a peak at t = 2.8 min (peak 3) as the

56d peak decreased in intensity. Peaks 1-5 were individually collected, treated with alkaline phosphatase, and then reanalyzed on HPLC. While phosphatase treatment of peaks 1, 2, 4, and 5 produced no change of the unknown species, treatment of peak 3 led to a peak at t = 7.3 min, the characteristic retention time of FdU under the conditions (the four natural nucleosides all have different retention times). Thus, as FdU was generated from the cyclization reaction, it was converted by thymidine kinase in the crude lysate into the more hydrophilic FdU 5’-monophosphate, the active antimetabolite.

125 1 2 '

3

" s Ü& 1 1

______«

0 10 15 20 25

Retention Time (min)

Figure 4.10. HPLC analysis of activation of prodrug 56d by E. coli lysate (214 nm). a. 56d only; b. 56d treated with cell lysate for 1 h, followed by 3 h incubation at room temperature (for cyclization to complete); and c. treatment of peak 3 derived from b with alkaline phosphatase. The small peak with retention t = 19.5 min (in a and b) is due to racemization of 56d at the methionyl residue; the resulting isomer is resistant to deformylase reaction.

126 4.3.6 Inhibition of Bacterial Cell Growth.

Activity of compounds 56a-d to inhibit bacterial cell growth was tested. Compounds

56a-d indeed inhibited £. coli cell growth, but with relatively high IC 50 values of 100-

1000 (xM (see Figure 4.11 for an example). Upon detailed examination, it was found that

this value varied depending on conditions for cell growth. For instance, 20 pM of f-Met-

(D)Pro-FdU completely stopped E. coli cell growth in minimum medium, while having

very little effect when the experiment was done with E. coli grown in Lauria broth.

Generally, under conditions where cells are healthy and fast-growing, significantly higher

concentrations of prodrugs were required to show any inhibitory effect. At high prodrug

concentrations and over extended period of time, background hydrolysis of prodrugs and

other unrelated factors might have played a role in the inhibition observed. The concern

was addressed with proper controls using prodrug analogues that were not deformylase

substrates. If the observed inhibition had been due to specific activation by deformylase action as designed, the non-substrate control molecules would remain intact inside

bacterila cells and thus harmless to the cells.

It was known that when the sulfur atom of methionine side chain in deformylase

substrate is oxidized to sulfone, the peptide would no longer be deformylated. Sulfone analogues of prodrugs were furnished with treatment of prodruga with m-CPBA. At high concentrations, these sulfone compounds also showed inhibition of cell growth less than corresponding prodrugs 56a-d. One may argue that the sulfones could be reduced inside cells and the compounds could become again deformylase substrates. To rule out this possibility, f-(D)Met-Pro-FdU was synthesized and tested. With D-methionine as the P f

127 residue, this compound should not be a deformylase substrate, however, it showed similar degree of inhibition as other compounds. Therefore, the observed cell growth inhibition by prodrugs was not due to intended deformylase action, but hydrolytic release of FdU.

1.6 no drug 1.4 20 pM f-Met-(D)Pro-FdU 100 pM f-Met-(D)Pro-FdU 1.2 500 pM f-Met-(D)Pro-FdU

1 § Q 0.8 O 0.6

0.4

0.2

0 0 1 2 3 4 5 6 Time (hour)

Figure 4. II. Inhibition of E. coli cell growth by prodrugs. E. coli (DH5a) cell culture was grown overnight and diluted ICO fold to fresh LB medium containing prodrugs. Cell density was monitored at 600 nm.

128 50 |iM f-met-(a-methyl)Ala-FdU 50 |o,M oxidized f-Met-(a-methyi)Ala-FdU ♦— 50 fiM Ac-Met-(a-methyl)Ala-FdU A— no drug

Q 0.3

3 4 5 Time (hour)

0.9 — 100 pIVI f-Met-(a-methyl)Ala-FdU 0.8 ■— 100 |aM Ac-Met-(a-methyi)Ala-Fdu ^ — 100 [J.M f-Ala-(a-methyl)Ala-FdU 0.7

0.6 §

8 ” 0.4

0.3

0.2

0 1 2 3 4 Time (hour)

Figure 4.12. Inhibition o f staphylococcus epidermidis cell growth by prodrugs. Prodrugs were added at beginning of dilution or midlog phase of cell growth.

129 no IPTG, no prodrug no IPTG, 100|.iM f-Met-(D)Pro-FdU 100 i^M IPTG, no prodrug 1.5 - 100 pIVI IPTG, 100pMf-Met-(D)Pro-FdU

g û ” 1

2 3 Time (hour)

0.7 100 pM IPTG, 100|iM f-Met-Pro-FdU 0.6 100 i^M IPTG, 100|j.M oxidized f-Met-P-FdU, 100 p,M IPTG, no drug 0.5 S Q 0.4 O 0.3

0.2

0.1

Om * * 4 6 8 Time (hour)

Figure 4.13. Inhibition of E. coli pET22b-def cell growth by prodrugs. An overnight culture of E. coli BL21[DE3] carrying plasmid pET22b-def was diluted 1000 fold to fresh LB medium containing prodrugs and cell density was monitored at 600 nm.

130 To minimize nonspecific hydrolysis of prodrug during activity tests, the stable prodrug 56d was chosen for more extensive studies. In addition to sulfone. two more control molecules, acetyl-Met-(a-methyI)alanine-FdU and formyI-Ala-(a-methyI)alanine-

FdU, were synthesized and tested alongside 56d. Gram-positive bacteria staphylococcus epidermidis and Bacillus subtilis were used as the target because Gram-positive cells are generally more susceptible to drug penetration and are therefore more sensitive to antibiotic treatment. As shown in figure 4.12, inhibition did occur at lower prodrug concentration. However, inhibition profile again strongly depended on experimental conditions and there was no appreciable difference between 56d and controls.

One or a combination of two factors could cause the ineffectiveness of growth inhibition. First, the prodrugs may not be efficiently deformylated by the deformylase due to competition with the endogenous substrates. Second, the prodrugs may poorly penetrate the bacterial cells. We therefore tested the prodrugs with BL21(DE3) cells transformed with the overproducing plasmid pET-22b-def (Rajagopalan et al., 1997). In the presence of 100 pM IPTG, these cells produce peptide deformylase as -50% of total cellular proteins. Thus, these cells should be able to completely deformylate any prodrugs that have entered the cell cytoplasm. However, the growth inhibition profile of these overproducing cells was similar to that of BL2l(DE3) cells without plasmid (Figure

4.13).

Taken together, the above results suggest that inefficient deformylation is unlikely the reason for the relatively low potency of 56a-d for cell growth inhibition. Rather, it is

131 the poor transport of prodrugs 56a-d across the bacterial cell wall/membrane that has

limited the intracellular concentration of the prodrugs.

4.3.7 Studies of 3’-(f-Met-Pro)-FdU

While the prodrugs developed above were unable to cross cell wall or membrane,

the active drug FdU is a clinical drug capable of exerting its cytotoxic effect even at very

low concentration. The high potency of FdU indicated that FdU is actively transported

into the cells, possibly by the transporter used by cells to sequester its natural analog dU.

Apparently, modification at 5’-0H of FdU impedes active transport of prodrugs. To develop prodrugs that are able to penetrate into bacterial cells, compounds with the N- formyl-dipeptide moiety attached to the 3’-OH of FdU were designed. It was hoped that these molecules would be recognized by the transporter protein that may interact with structural features at the a-face of 2’-deoxyribose. On the other hand, attachment of a peptide moiety at 3’-hydroxyl may prevent FdU from approaching the active site of its target enzyme and make the molecule not or less toxic.

Synthesis of 3’-(f-Met-Pro)-FdU was carried out as outlined in Figure 4.15.

Following protection of 5’-0H of FdU as DMT-ether, 3’-OH was linked to tBoc-proline that had been activated by TIPS-Cl. Treatment of 86 with TFA deblocked trityl and tBoc protection simultaneously. Amide bond formation between formyl-methionine and 87 afforded final product 89.

Deformylation rate of the compound was determined using formate dehydrogenase assay. Like all other prodrugs, it is a substrate of deformylase. Value of

132 kcat/Kw is 5.07 x I O'* about 4 fold higher than 5’-(f-Met-Leu)-FdU. Once deformylated, the molecule underwent cyclization to release FdU. The cyclization rate was not accurately determined but was expected in the range reported for other prodrugs.

Finally, 3'-(f-Met-Pro)-FdU was tested with bacterial cell culture. Unfortunately, it failed to show potent inhibition of cell growth as compared to controls. The overall profile of inhibition tests was no different form those of other prodrugs. This inefficiency was again probably due to problem in drug transport.

133 NH NH

HO DMT-O pyridine TIPS-Ci, DMAP

OH OH 85 59 O

NH NH

' n A q HO ' n J TFA TFAHN XT O 86 87 O

NH o~ HO O A H N H

88

Figure 4.15. Synthesis of 3’-(f-Met-Pro)-FdU.

134 4.4 DISCUSSION

Because peptide deformyiase is an essential and unique activity to eubacteria, it is being actively pursued as a target for designing novel antibiotics. The approach that has been taken by pharmaceutical groups is to rationally design and/or screen for deformyiase inhibitors, which are expected to selectively block bacterial cell growth. In this work we have conceptually demonstrated a novel approach to antibacterial chemotherapy which takes advantage of the unique bacterial peptide deformyiase to convert a biologically inactive prodrug into an active antibacterial agent. This prodrug approach offers several advantages over the inhibitor approach. First, the prodrug is designed to be biologically inactive and can be converted into an active drug only through deformyiase action.

Therefore, the active drug is generated only inside a bacterial pathogen where active peptide deformyiase is present. Such drugs should have much lower toxicity to an eukaryotic host. Second, this approach is extremely versatile. In principle, any molecule that can block cell growth and that requires a free hydroxyl group (or other nucleophilic groups) for toxicity can be masked by an acyl group to eliminate or reduce its toxicity.

Once taken up by a bacterial pathogen, the acyl group is removed to release the toxic compound that can target any one of the cytoplasmic proteins essential for bacterial survival. Third, prodrugs of this type are more likely to be broad-spectrum antibiotics as compared to deformyiase inhibitors. Molecular cloning and genomic sequencing have so far revealed 34 deformyiase genes; their peptide sequences show a wide range of homologies (20% to 65%) (Mazel et al., 1997, Meinnel et al., 1997). Although the active site structure seems to be well conserved, the sequence diversity suggest that a potent

135 inhibitor for one deformyiase may not be so for its counterpart from a different pathogen.

Our earlier studies have shown that the deformyiase requires interactions with the four N- terminal residues for optimal binding and activity (Rajagopalan et al., 1997). This structural diversity is much less of a concern for the prodrugs. This is because that each peptide deformyiase has evolved to be a broad-specifrcity enzyme so that it can deformylate thousands of bacterial proteins which bear a wide variety of N-terminal sequences. Thus, a prodrug that can be activated by the deformyiase in one organism is likely to do the same in another. Fourth, bacterial cells are less likely to become resistant to the prodrugs than to the inhibitor drugs. The sequence diversity of the deformylases predicts that a mutation in the substrate binding site (not the active site) may be tolerated, although such a mutation might change the relative substrate specificity. Such a mutation would undoubtedly affect or even abolish the inhibitor binding, making the pathogen resistant to the inhibitor. Cells with such a mutation would, however, still be sensitive to the prodrugs, which are substrates of the deformyiase. If a cell encounters a mutation that prevents prodrug activation, the mutant enzyme would have difficulty in deformylating the cell’s own proteins, resulting in a fatal consequence. Finally, our preliminary results show that 1,2-ethanedithiol, a potent inhibitor of peptide deformyiase (Rajagopalan et al.,

1997), arrests E. coli cell growth when added directly to the medium, but upon removal of the inhibitor cells resume rapid growth (unpublished results). This may suggest that a deformyiase inhibitor acts as a bacteriostatic agent. The prodrugs can, however, be made as either bacteriostatic or bactericidal drugs by varying the structure of the active drug.

136 A potential disadvantage of the prodrug approach lies in the difficulty in its

design. First, a prodrug must be a good substrate of the deformyiase, as the physiological

substrates are natural antagonists of the prodrug. Second, after deformyiase action, the cyclization must be sufficiently facile to generate a critical concentration of the active drug. By using dipeptides as the acylating unit, prodrugs 56a-d showed deformylation rates comparable to those of the natural substrates. Indeed, a crude E. coli cell lysate was very effective in deformylating 56a-d. The cyclization step is also rapid, with Ua o f 2-51 min. In addition, by varying the structure of the penultimate residue, we can control the rates of both the enzymatic deformylation and the subsequent cyclization steps. Another potential concern with the use of peptidyl units in drug design is the in vivo stability against protease action. Although 56a, which contained only L-amino acids, was unstable against proteases, prodrugs containing unnatural amino acids (56b-d) are essentially resistant to proteases. It may also be possible to use nonpeptidic acyl groups as the acylating unit.

In summary, our preliminary results conceptually demonstrate that the prodrug activation is a viable approach to antibacterial chemotherapy. Although the prodrugs described here have not been very potent in inhibiting in vivo cell growth due to poor cell permeability, redesign with more hydrophobic drugs or those that can be taken up by active transport systems will likely provide a new class of potent antibiotics.

137 List of References

Adams, J. M. (1968) On the release of the formyl group from nascent protein. J. Mol. Biol. 33, 571-589.

Adams, J.M. & Capecchi, M. (1966) The role of N-formylmethionine-tRNA as the initiator of protein synthesis. Proc. Natl. Acad. Sci. USA 55, 147 - 155.

Anderson, G.W., Zimmerman, J.E. & Callahan, P.M. (1964) The use of esters of N- hydroxysuccinimide in peptide synthesis. J. Am. Chem. Soc. 86, 1839-42.

Ben-Bassat, A., Bauer, K., Chang, S.Y., Myambo, K., Boosman, A. & Chang, S. (1987) Processing of the initiation methionine from proteins: properties of the Escherichia coli methionine aminopeptidase and its gene structure. J. Bacterial. 169, 751 — 757. a. Becker, A., Schlichting, 1., Kabsch, W„ Schultz, S., Wagner, A.F. (1998) Structure of peptide deformyiase and identification of the substrate binding site. J. Biol. Chem. 273, 11413-6. b. Becker, A., Schlichting, L, Kabsch W„ Groche, D., Schultz, S., Wagner, A.F. (1998) Iron center, substrate recognition and mechanism of peptide deformyiase. Nature Structural Biology 5 (12): 1053-8.

Bianchetti, R., Lucchini, G., Crosti, P., Tortora, P. (1977) Dependence of mitochondrial protein synthesis initiation on formylation of the initiator methionyl-tRNAf. J. Biol. Chem. 252, 2519-23.

Bianchetti, R., Lucchini, G. & Sartirana, M.L. (1971) Endogenous synthesis of formyl- methionine peptides in isolated mitochondria and chloroplasts. Biochem. Biophys. Res. Commun. 42, 97-102.

Blake, J & Litzi-Davis, L. (1992) Evaluation of peptide libraries: An interactive strategy to analyze the reactivity of peptide mixtures with antibodies. Bioconjugate Chem. 3, 510- 513.

Bodanszky, M. (1993) Principles o f peptide synthesis, 2"^ ed.. Springer-Verlag, Germany.

138 Campbell, D.A. & Bermak, J.C. (1994) Solid-phase sjTithesis of peptidylphosphonates. J. Am. Chem. Soc. 116, 6039-40.

Chan, M.K., Gong, W., Rajagopalan, P.T., Hao, B., Tsai, C.M. & Pei, D. (1997) Crystal structure of the Escherichia coli peptide deformyiase. Biochemistry 36,13904-9.

Corey, E.J. & Venkateswarlu A. (1972) Protection of hydroxyl groups as tert- butyldimethylsilyl derivatives./. Am. Chem. Soc. 64, 6190-1.

Corey, E.J., Weinshenker, N.M., Schaaf, T.K., Huber. W. (1969) J. Am. Chem. Soc. 91. 5675.

Cunico, R.F. & Bedell, L. (1980)/. Org. Chem. 45, 4797.

Cushman, D.W., Cheung, H.S., Sabo, E.F. & Ondetti, M.A. (1977) Design of potent competitive inhibitors of angiotensin-converting enzyme. Carboxyalkanoyl and mercaptoalkanoyl amino acids. Biochemistry 16, 5484-91.

Dardel F. Ragusa S. Lazennec C. Blanquet S. Meinnel T. (1998) Solution structure of nickel-peptide deformyiase./. Mol. Biol. 280,501-13.

Datta, D. & Majumdar, D. (1991)/. Phys. Org. Chem. 4, 611-617.

Davies, J. (1996) Bacteria on the rampage. Nature 383, 219.

Davis, J. D. (1994) Science 264, 375-382.

Dickerman, H.W., Steers Jr, E., Redfield, B.G. & Weissbach, H. (1967) Methionyl soluble ribonucleic acid transformylase. I. Purification and partial characterization. / Biol. Chem. 242, 1522-1525.

Duhamel, P., Duhamel, L., Danvy, D., Monteil, T., Lecomte, J.-M. & Schwartz, J.- C. (1996) Procédé de synthèse d’acides acryliques alpha-substitues et leur application. European Patent 0 729 936 Al (in French).

Epstein W.W. & Sweat, F.W. (1967) Dimethyl sulfoxide oxidations. Chemical Review 67 (3), 247-260.

Fleishman, R. D., Adams, M. D., White, O., Clayton, R. A., Kirkness, E. F., Kerlavage, A. R., Bult, C. J., Tomb, J.-F., Dougherty, B. A. & Merrick, J. M. (1995) Whole-genome random sequencing of Haemophilus influenzae Rd. Science 269,496-512.

139 Fraser, C. M., Gocayne, J. D., White, O., Adams, M. D., Clayton, M. D., Fleishmann, R. D., Bult, C. J., Kerlavage, A. R., Sutton, G. & Kelley, J. M. (1995) The minimal gene complement of Mycoplasma genitalium. Science 270, 397-403.

Furka, A., Sebestyen, F., Asgedom, M.m & Dibo, G. (1991) Int. J. Peptide Protein Res. 37, 487-493.

Gisin, B. F. & Merrifield, R. B. (1972) J. Am. Chem. Soc. 94, 3100-3106.

Gold, H. S. & Moellering Jr., R. C. (1996) New Eng. J. Med. 335, 1445-1453.

Grassetti, D.R. & Murray Jr., J.F. (1967) Determination of sulfhydryl groups with 2,2’- or 4,4’-dithiodipyridine. Archives o f Biochemistry and Biophysics 119, 41-49.

Groche, D., Becker, A., Schlichting, I., Kabsch, W., Schultz, S., Wagner, A.F. (1998) Isolation and ciystallization of functionally competent Escherichia coli peptide deformyiase forms containing either iron or nickel in the active site. Biochem. Biophys. Res. Comm. 246, 342-6.

Guillon, J.M., Mechulam, Y., Schmitter, J.M., Blanquet, S. & Fayat, G. (1992) Disruption of the gene for Met-tRNA^^V formytransferase severely impairs the growth of Escherichia coli. J. Bacterial. 174, 4294-430.

Hao, B., Gong, W., Rajagopalan, P.T., Zhou, Y., Pei, D. & Chan, M.K (1999) Structural basis for the design of antibiotics targeting peptide deformyiase. Biochemistry 38, 4712-9.

Hauschild-Rogat, P. (1968) N-formylmethionine as a N-terminal group of E. coli ribosomal protein. Mol. Gen. Genet. 102, 95-101.

Heinrikson, R.L. & Hartley, B.S. (1967) Purification and properties of methionyl- transfer-ribonucleic acid synthetase from Escherichia coli. Biochem. J. 105, 17-24.

Hu, Y.-J., Rajagopalan, P.T.R. & Pei, D. (1998) H-phosphonate derivatives as novel peptide deformyiase inhibitors. Bioorg. Med. Chem. Lett. 8, 2479-2482.

Hu, Y.J., Wei, Y., Zhou, Y., Rajagopalan, P.T.R. & Pei, D. (1999) Determination of substrate specificity for peptide deformyiase through the screening of a combinatorial peptide library. Biochemistry 38, 643-50.

Itoh, M., Hagiwara, D., & Kamiya, T. (1977) Bull. Chem. Soc. Japan. 50, 718-721.

Jacobsen, N.E. & Bartlett. P.A. (1981) A phosphonamidate dipeptide analogue as an inhibitor of carboxypeptidase A. J. Am. Chem. Soc. 103, 654-7.

140 Khosla, M. C., Smeby, R., & Bumpus, F. M. (1972) J. Am. Chem. Soc. 94, 4721-4724.

Kozak, M. (1983) Comparison of initiation of protein synthesis in prokaryotes, eukaryotes, and organelles. Microbiol. Rev. 47, 1-45.

Kung, H.F., Eskin, B., Redfield, B. & Weissbach, H. (1979) DNA-directed in vivo synthesis of galactosidase: Requirement for formylation of methionyl-tRNAMetf. Arch. Biochem. Biophys. 195, 396 —400.

Lam, K.S.. Salmon, S.E., Hersh, E. M., Hruby, V.J., Kazmierski, W.M. & Knapp, R.J. (1991) Nature 354, 82-84.

Lazennec, C., Meinnel, T. (1997) Formate dehydrogenase-coupled spectrophotometric assay of peptide deformyiase. Analytical Biochemistry 244, 180-2.

Lipshutz, B. H. & Tegram, J.J. (1980) Tetrahedron Letter 21, 3343.

Livingston, D. M. & Leder, P. (1969) Deformylation and protein synthesis. Biochemistry 8, 435-443.

Marasco, W. A., Phan, S. H., BCrutzsch, H., Showell, H. J., Feltner, D. E., Naim, R., Becker, E. L., and Ward, P. A. (1984) J. Biol. Chem. 259, 5430-5439.

Marcker, K. & Sanger, F. (1964) N-Formyl-methionyl-tRNA. J. Mol. Biol. 8, 835-840.

Mazel, D., Coic, E., Blanchard, S., Saurin, W., & Marliere, P. (1997) A survey of polypeptide deformyiase function throughout the eubacterial lineage. J. Mol. Biol. 266, 939-949.

Mazel, D., Pochet S., & Marliere, P. (1994) Genetic characterization of polypeptide deformyiase, a distinctive enzyme of eubacterial translation. EMBO J. 13, 914-923;

Mazel, D., Coic, E., Blanchard, S., Saurin, W. & Marliere P. (1997) A survey of polypeptide deformyiase function throughout the eubacterial lineage. JOURNAL OF MOLECULAR BIOLOGY266, 939-49.

Meinnel, T. & Blanquet, S. (1993) Evidence that peptide deformyiase and methionyl- tRNA(fMet) formyltransferase are encoded within the same operon in Escherichia coli. Journal o f Bacteriology 175, 7737-40.

Meinnel, T. & Blanquet, S. (1994) Characterization of the Thermus thermophilus locus encoding peptide deformyiase and methionyl-tRNA(fMet) formyltransferase. J. Bacteriol. 176, 7387-7390.

141 Meinnel, T. & Blanquet, S. (1995) Enzymatic properties of Escherichia coli peptide deformyiase. Jbwrwa/ o f Bacteriology 177, 1883-7.

Meinnel, T., Blanquet, S. & Dardel, F. (1996) A new subclass of the zinc metalloproteases superfamily revealed by the solution structure of peptide deformyiase. Journal o f Molecular Biology. 262, 375-86.

Meinnel T. Lazennec C. Blanquet S. (1995) Mapping of the active site zinc ligands of peptide d&ïormy\asQ. Journal o f Molecular Biology. 254, 175-83.

Meinnel T. Lazennec C. Dardel F. Schmitter JM. Blanquet S. (1996) The C-terminal domain of peptide deformyiase is disordered and dispensable for activity. FEBS Letters. 385,91-5.

Meinnel, T., Lazennec, C., Villoing, S., & Blanquet, S. (1997) Structure-function relationships within the peptide deformyiase family. Evidence for a conserved architecture of the active site involving three conserved motifs and a metal ion. J. Mol. Biol. 267, 749-761.

Meinnel, T., Mechulam, Y., & Blanquet, S. (1993) Methionine as translation start signal: a review of the enzymes of the pathway in Escherichia coli. [Review] Biochimie 75, 1061-1075.

Meinnel, T., Patiny, L., Ragusa, S. & Blanquet, S. (1999) Design and synthesis of substrate analogue inhibitors of peptide deformyiase. Biochemistry 38, 4287-95.

Miller, C.G., Strauch, K.L., Kukral, A.M., Miller, J.L. Wingfield, P.T., Mazzei, G. J., Werlen, L.C., Graber, P. & Mowa, N.R. (1987) N-terminal methionine specific peptidase m. Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 84, 2718-2722.

Milligan, D. L., and Koshland, D. E., Jr. (1990) The amino terminus of the aspartate chemoreceptor is formylmethionine. J. Biol. Chem. 265, 4455-4460.

Mitsunobu, O. (19S\) Synthesis, 1-28.

Morton, D.R. & Thompson, J.L. (1978) Total synthesis of 3-Oxa-4,5,6-trinor-3,7-inter-m- phenylene prostaglandins. 2. Conjugate addition approach. J. Org. Chem. 43, 2102-6.

Neu, H. C. (1992) The crisis in antibiotic resistance. Science 257, 1064-1073.

Ondetti, M.A., Condon, M.E., Reid, J., Sabo, E.F., Cheung, H.S. & Cushman, D.W. (1979) Design of potent and specific inhibitors of carboxypeptidases A and B. Biochemistry 18, 1427-30.

142 Pedroso, E., Grandas, A., de las Heras, X., Eritja, R., & Giralt, E. (1986) Deketopiperazine formation in solid phase peptides synthesis using p-alkoxybenzyl ester resins and Fmoc amino acids. Tetrahedron Letter 27, 743-746.

Pine, M. J. (1969) Biochim. Biophys. Acta 174, 359-372.

Ragusa, S., Blanquet, S., Meinnel, T. (1998) Control of peptide deformyiase activity by metal cations. Journal o f Molecular Biology 280, 515-23.

Rajagopalan, P. T. R., Datta, A., & Pei, D. Purification, characterization, and inhibition of peptide deformyiase from Escherichia coli. (1997) Biochemistry 36, 13910-13918.

Rajagopalan, P. T. R., Yu, X. C., & Pei, D. (1997) J! Am. Chem. Soc. 119, 12418-12419.

Rajagopalan, P.T.R., & Pei, D. (1998) Oxygen-mediated inactivation of peptide deformyiase. J. Biol. Chem.273, 22305-10.

Rothe, M. & Mazanek, J. (1974) Liebigs Ann. Chem., 439-459.

Santi, D. V., McHenry, C. S., Raines, R. T., & Ivanetich, K. M. (1987) Kinetics and thermodynamics of the interaction of 5-fiuoro-2'-deoxyuridylate with thymidylate synthetase. Biochemistry 26, 8606-8613.

Schmitt, E., Guillon, J.M., Meinnel, T., Mechulam, Y., Dardel, F. & Blanquet S. (1996) Molecular recognition governing the initiation of translation in Escherichia coli. A review. Biochimie 78, 543-54.

Sheehan, J.C. & Yang, D.-D. H. (1958) The use of N-formylamino acids in peptide synthesis. J. Am. Chem. Soc. 80, 1154-8.

Spratt, B. G. (1994) Science 264, 388-393.

Takeda, M. & Webster, R.E. (1968) Protein chain initiation and deformylation in B. subtilis homogenates. Proc. Natl. Acad. Sci. USA 60, 1487-1494.

Thompson, R.C. (1973) Use of peptide aldehydes to generate transition-state analogs of elastase. Biochemistry. 12, 47-51.

Varshney, U. & RajBhandary, U.L. (1992) Role of methionine and formylation of initiator tRNA in initiation of protein synthesis in Escherichia coli. J. Bacteriol. 174, 7819-7826.

143 Warren, W.C., Bentle, K.A., Schlittler, M.R., Schvvane, A.C., O'Neil, J.P. & Bogosian, G. (1996) Increased production of peptide deformyiase eliminates retention of formylmethionine in bovine somatotropin overproduced in Escherichia coli. Gene 174. 235-8.

Wei, Y. (1997) Master thesis: Development of a continuous deformyiase assay and site- directed mutagenesis of Escherichia, coli deformyiase. The Ohio State University.

Wei, Y. & Pei, D. (1997) Continuous spectrophotometric assay of peptide deformyiase. Analytical Biochemistry 250, 29-34.

Westerik, J.O. & Wolfenden R. (1972) Aldehydes as inhibitors of Papain. J. Biol. Chem. 247,8195-7.

Wetter, H. & Oertle, K. (1985) Tetrahedron Lett. 26, 5515.

White, D. & Coville, N.J. (1994) Adv. Organomet. Chem. 36, 95-158.

Wipf P., Li, W., & Sekhar, V. (1991) Syntheses of chemoreversible prodrugs of ara-C. Bioorg. Med. Chem. Lett. 12, 745-750.

Yazawa, H., Tanaka, K. & Kariyone K. (1974) The reaction of carboxylic esters with boron tribromide, a convenient method for the synthesis of amides and transestérification. Tetrahedron Letters 46. 3995-6.

144 APPENDIX: PROTON NMR SPECTRA

145 (200 MHz, CDCy

4^ o\

(

y i ^ / / // _A_ L . J L rv. A. 0 0 V y

' 1 ' —J r 'i -r *t“*| "1" I I I I 'I 9,5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 NO.

FKC

(200 MHz, DMSO)

wi,w #,*### iwi I**'**

10.0 PPM NO,

(200 MHz. DMSO)

LJ

10.0 PPM NH,

(200 MHz, DMSO)

10.0 PPM (200 MHz, DMSO)

L/\ O

______L

10.0 PPM NO.

(200 MHz, DMSO)

1.0 10.0 9.0 8.0 7 0 6.0 5.0 4,0 3 0 2.0 1 .0 PPM NO. HS

0

(250 MHz, C D C y

U\ w NO.

HS

(300 MHz, C D C y

10.0 PPM ( 2 0 0 M H z.C D C y

PPM NO.

(200 MHz. CDCy

PPM U\ o>

\n /

PPM NO,

HO

tjjj

PPM NO,

H

(200 MHz. CDCI3 )

PPM 9,0 6.9 6.9 6.0 5.0 4.5 4.0 3.9 3.0 PPM 0 .

0 (250 MHz. CDCy

OON

jULUL_u

9 0 B . 0 7 . 0 6 0 5 0 4 0 3 . 0 2 . 0 1 0 PPM (250 MHz, CDCI3 )

7 0 A . 0 3 0 2 0 1 0 0 . 0 PPM (250 MHz. CDCI3 )

S

en in rvj m

PPM a

n - ’ -'- ’’ I ' T - T - r - - r - i - r - T - - T ~ j~ r- | - r T [-'i n I I T" p i ■ I I I I ■' V ' ' ' I ' r 8.5 8.0 7.5 7.0 6.5 6.0 5 0 4.5 4 03.5 3.0 2.5 2 0 1.5 1 0 0 0 PPM NO,

(300 MHz. CDCy

u

PPM NO.

HS

(300 MHz, C D C y

v v

PPM (300 MHz, CDCy

On

PPM 0 ^

(300 MHz. CDCy

O n

KW

4 0 3.5 3 0 PPM OH

0

(200 MHz. CDCI3 )

00

u ;

9,5 9 0 8,5 8,0 7.5 6,5 6,0 5,5 5,0 4,5 4,0 3,5 3.0 2,57,0 2,0 1.51,0 5 PPM 0 0

(200 MHz. CDCI3 )

On NO

/ /

U in 10 03 ic N GO xr (V) ru

g . 0 8 . 0 7,0 6 , 0 5.0 4 .0 3 .0 2 .0 1 . 0 PPM 0 . .0 YY \ 0 0

(200 MHz. CDCI3 )

o

Y J l l J v v V

' t -T -p-r-T -r ->-7-1- I ' I ■ 9.5 9 0 0.5 6.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1 .5 1 0 PPM HO OH

/ ' x HCI (200 MHz, DgO)

PPM OH

(200 MHz, CDCI3 )

to

U V 1

' I ' ' I ' ' I ' T-j-r -T-l-r • T - p - r T - j - r T-T-T -T-pr ’-r -pT 9.5 9.0 8.5 8.0 7.5 7 0 6 5 5 5 5.0 4.5 4 .0 3.5 3 0 ? ,5 2.0 15 ,5 PPM (200 MHz, CDCI3 )

U kn 10 oc in T 1,5 1.0 .5 ^ O H

0 (200 MHz, C D C y

9.5 9 0 8.58.0 7.57.0 8,5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1 .5 JO 5 PPM .OH V' 0 0 (200 MHz, CDCy

y Di Ül - ^ I

cv 10 ’ I ' T 9.5 9.0 8.5 8.0 7 5 7.0 6.5 6.0 5,5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 .5 PPM OH

-j 0\

9.5 9.0 0.5 8.0 7.5 7.0 6.5 6.0 5.04.0 3.5 3 .0 2.5 2.0 I .51.0 .5 PPM s OH

0 0

(200 MHz. CDCI3 )

'-T ' ■ ' 1 ' '"l ' ' 1 ' ' I ' -t-T-r- -"T' 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 1.5 1.0 .5 PPM OH

(200 MHz, C D C y

PPM 0

NH

(200 MHz, CDCy OH

V V

8,5 PPM 0 A NH H

(200 MHz, DMSO) OH

00 o

L kl V V AA

9 .5 9.0 8 5 8 0 r-p--, 7 5 7 0 B 5 5.5 4 5 4 0 2 ,5 2 0 1 ,5 1 0 0

NH

(2 OOMHZ.CDCI3 ) OH

00

10.0 9.0 8.0 7.0GO 5.0 4,0 3,0 2.0 1.0 PPM N " 0

(200 MHz. CD CÜ OH

A > v A/ l X v U

■ I ■ -r-y-r- 10.0 9.0 0.0 7.0 6 .0 5.0 4.0 3.0 2 .0 1.0 PPM 0

OH

00 U)

PPM 0

NH

(200 MHz. CDCI3 ) OH

V u

PPM NH

(200 MHz. DMSO) OH

1/100

12.0 11.0 10.0 PPM Hdd 0 01 o u 021

VO 00

HO (OSIAJQ 'ZHIAI OOZ)

HN NH

OH (200 MHz, DMSO)

00

12.0 11.0 10.0 PPM NH

(200 MHz. DMSO) OH

00 00

r 12.0 11.0 10.0 PPM NH

(200 MHz. DMSO) OH

12.0 11.0 10.0 PPM NH

(200 MHz, DMSO) OH

12.0 n .0 10.0 PPM 0

NH

(200 MHz, DMSO) OH

12.0 11 .0 10.0 PPM 0

NH

DM T-0

OH (200 MHz. CDCy

9.0 B.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2 5 2.0 1 .59.5 \ .0 5 PPM 0

NH DMT-0

(200 MHz, CDCy

-A

9 0 8.5 8 .0 7.57.0 6.5 6.0 5 59.5 5.0 4 5 4.0 3.5 3.0 2.5 2.0 1 5 1.0 5 PPM 0

HO

(200 MHz, DMSO)

12.0 no 10,0 PPM