Abstract of the Dissertation

UNIVERSITY OF CALIFORNIA, SAN DIEGO

Aminoglycoside Derivatives and Mimetics as RNA Binders

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

Chemistry

Fang Zhao

Committee in charge:

Professor Yitzhak Tor, Chair Professor Jeffery D. Esko Professor Thomas Hermhann Professor Joseph M. O’Connor Professor Susan S. Taylor

2007

Copyrights©

Fang Zhao, 2007

The dissertation of Fang Zhao is approved, and it is acceptable in quality and form for publication on microfilm.

Chair

University of California, San Diego

2007

iii

With love dedicated to Wenxiang Zhao and Zhiqin Wang, my parents.

TABLE OF CONTENTS

Signature Page ...... iii

Dedication ...... iv

Table of Contents ...... v

List of Symbols and Abbreviations...... vii

List of Figures ...... ix

List of Tables ...... xiii

Acknowledgement ...... xiv

Curriculum Vitae ...... xv

Abstract of the Dissertation ...... xvi

Chapter 1 ...... 1

1.1 Introduction ...... 2

1.2 RNA as a Drug Target ...... 7

1.3 Aminoglycosides ...... 12

1.4 Design of RNA Binders ...... 22

1.5 Goals ...... 30

References ...... 31

Chapter 2 ...... 37

2.1 Introduction ...... 38

2.2 Design of Conformationally Restricted Aminoglycosides ...... 41

2.3 Synthesis ...... 46

2.4 Spectroscopic Characterizations...... 50

2.5 15N NMR Studies and pKa measurement ...... 54

2.6 Ligand Binding to the A– Site RNA ...... 57

2.7 Crystal Structure ...... 60

2.8 Discussion ...... 66

References ...... 69

Chapter 3 ...... 73

3.1 Introduction ...... 74

3.2 Synthesis of the 4, 5 substituted amino–cyclohexanecarboxylic acids ...... 78

3.3 Conformation Analysis for Cyclohexyl Oxiranes ...... 91

References ...... 98

Chapter 4 ...... 100

4.1 Synthetic Experimental Section ...... 101

4.1.1 Materials ...... 101 4.1.2 NMR Instrumentation ...... 101 4.1.3 Synthesis of conformationally-constrained Neomycin B (2a) ...... 101 4.1.4 Synthesis of conformationally-constrained Neomycin B (2b) ...... 112 4.1.5 Synthesis of 2-aminocyclohexancarboxylic acids...... 118

References ...... 127

LIST OF SYMBOLS AND ABBREVIATIONS

Å angstrom

 chemical shift

 difference o degree atm. atmosphere ave. average

Boc tert-butyloxycarbonyl br broad

Bu butyl

Bz benzyl calcd calculated

DMAP 4-(dimethylamino)pyridine

DMF N,N-dimethylformamide

DMSO dimethyl sulfoxide equiv. equivalent

ESI MS electrospray ionization mass spectrum

Et ethyl

FAB MS fast atom bombardment mass spectrum

HPLC high performance liquid chromatography

HR-MS high resolution mass spectrometry iPr iso-propyl

vii

K kelvin kcal kilo-calorie

L liter

M mol L-1

Me methyl mg mili-gram mL mili-liter mmol mili-mole mol mole

NMR nuclear magnetic resonance

Ph phenyl s second t-Bu tertial-butyl

TEA triethylamine tert tertial

Tf triflate

THF tetrahydrofuran

TLC thin layer chromatography

UV ultraviolet

Vis visible

viii

LIST OF FIGURES

Figure 1.1 The central dogma of biology...... 2

Figure 1.2 Schematic view of protein synthesis...... 3

Figure 1.3 Steps at which RNAs have been found to modulate gene expression...... 4

Figure 1.4 a) RNA secondary structure feature. b) Tertiary structure of tRNA...... 9

Figure 1.5 Representative RNA–binding small molecules...... 11

Figure 1.6 Representative aminoglycoside antibiotics ...... 14

Figure 1.7 The ribosomal 16S RNA sequence and the aminoglycoside binding sites...... 15

Figure 1.8 Structure of natural and synthetic deoxy–tobramycin derivatives...... 18

Figure 1.9 Modeling shows the structurally electrostatic complementarity between hammerhead RNA and neomycin B...... 20

Figure 1.10 Structure of Acridine–Neomycin conjugates...... 23

Figure 1.11 Structure of aminoglycosides and guanidinoglycosides...... 24

Figure 1.12 Structure of Pt–aminoglycoside conjugates...... 25

Figure 1.13 Structure of tobramycin and pyranmycin analogs...... 26

Figure 1.14 Structure of neamine derivative for targeting oncogenic RNA ...... 28

Figure 1.15 a) Representative designed piperidine glycoside ligand. b) Three– dimensional model of the designed piperidine glycoside superimposed on paromamine, showing their conformational similarity. c) Model of the piperidine glycoside 1 docked in the three–dimensional structure of the bacterial decoding– site RNA in complex with paromomycin...... 29

Figure 2.1 a) Chemical structures of the neomycin class aminoglycosides studied. b) The conformation of paromomycin bound to A– site11, neomycin bound to the TAR8, a SELEX– derived aptamer9 and the Tau exon.10 ...... 40

Figure 2.2 Conformationally constrained neomycin (2a) and paromomycin (2b). ... 43

Figure 2.3 Torsional angles of aminoglycosides free in solution or bound to A– site or the TAR RNAs...... 44

Figure 2.4 Synthesis of Conformationally Constrained Analogues 2a...... 47

Figure 2.5 Synthesis of Conformationally Constrained Analogues 2a...... 48

Figure 2.6 1H NMR (top) and 1H– 13C gHSQC (bottom) spectrums of restricted neomycin 2a ...... 52

Figure 2.7 1H– 13C gHMBC spectrum of restricted neomycin B 2a (left) and an expansion (right) highlighting the cross– peak from 5''–Hb to 2'–C...... 53

Figure 2.8 15N NMR spectra of restricted neomycin 2a and neomycin 1a at pH 10.7 ...... 54

Figure 2.9 a) RNA construct A– site 2AP (1492) containing the ribosomal decoding site used in this study. b) The structure of fluorescent nucleoside 2– aminopurine (2AP) that is incorporated at position 1492. c) Sample binding isotherms of paromomycin (1b) and restricted paromomycin (2b) binding to A– site 2AP (1492) at pH 7.5. d) The dependence of dissociation constants on pH...... 58

Figure 2.10 Secondary structure of the decoding– site oligonucleotide used for X– ray crystallography and structure determination...... 61

Figure 2.11 Crystal structures of neomycin 1a (a) and restricted neomycin 2a (b) complexed with decoding– site RNA. c) Stereo view of a superimposition of neomycin (1a, green) and its restricted derivative (2a, yellow) bound to the decoding site. Ligand superimposition of neomycin (1a, green) is shown in (d) with restricted neomycin (2a, yellow) and in (e) with paromomycin (1b, orange)...... 62

Figure 2.12 Overlay of the crystal structures of the unliganded decoding-site RNA and the neomycin complex ...... 63

Figure 2.13 Interactions of rings I, II, and IV in restricted neomycin (2a) with the decoding– site RNA...... 65

Figure 3.1 a) Structure of the cyclohexyl–peptide. b) Proposed 4,5–substituted amino– cyclohexanecarboxylic acid subunit...... 75

Figure 3.2 Schematic view of helical trans–ACHC oligomer. The intramolecular hydrogen–bonds are highlighted in red...... 76

Figure 3.3 a) Stereo image of a preferred solution conformation of a –(1,2) tripeptide module overlaid on the solution structure of paromomycin b) Structure of paromomycin and the modeled tripeptide, roughly in the orientation depicted in the overlay...... 76

Figure 3.4 The general synthetic pathway for the 4,5–substituted–trans–ACHC derivatives...... 78

Figure 3.5 Synthesis of 2R–amino–4R–amino–5R–hydroxyl–cyclohexancarboxylic acid series...... 80

Figure 3.6 Crystal structure of compound 6 (a) and 11 (b). Unrelated Hydrogen atoms have been omitted for clarity...... 82

Figure 3.7 Mechanism of trans diaxial effect for cyclohexyl oxirane opening...... 84

Figure 3.8 Possible conformers of compound 8 and their corresponding ring opening reactions...... 85

Figure 3.9 a) Compound 18 and 23. b) Two possible outcomes of epoxide opening reaction of compound 23 ...... 86

Figure 3.10 Synthesis of compound 18. Reagents and conditions: (a) Boc2O, DMAP; (b) NaOH; (c) (i) ClCO2Et, (ii) NaN3, (iii) , MeOH, p–TsOH; (d) m–CPBA. . 87

Figure 3.11 Synthesis of compound 23. Reagents and conditions: (a) m–CPBA (b) 1 atm. H2 (g), 10% Pd/C, Boc2O, i–Pr2NEt...... 88

Figure 3.12 Comparison of the transition structures for m–CPBA epoxidation of the trans and cis amino–cyclohexencarboxylates...... 88

Figure 3.13 a) Synthesis of tetra–equatorial amino–cyclohexancarboxylic acids from compound 23...... 90

Figure 3.14 Definition of torsion angle Tor1 (left) and Tor2 (right)...... 91

Figure 3.15 Torsion angles of compound 23...... 93

Figure 3.16 Modeling structure of compound 23. The most stable B–type conformation ① (left) and the most stable A–type conformation ⑤ (right)...... 94

Figure 3.17 Compound I, II and III...... 95

Figure 3.18 a) Chelatation effect of Li+ on cyclohexyl oxirane conformation. b) Possible intramolecular hydrogen bonding effect on cyclohexyl oxirane conformation...... 96

Figure 3.19 Steric hindrance of conformation A (left) and B (right)...... 97

1 Figure 4.1 H NMR spectrum of Boc5–constrained neomycin B (6)...... 104

Figure 4.2 gCOSY NMR spectrum of Boc5–constrained neomycin B (6)...... 105

Figure 4.3 gHSQC spectrum of Boc5–constrained neomycin B(6)...... 106

Figure 4.4 gHMBC NMR spectrum of Boc5–constrained neomycin B (6)...... 107

Figure 4.5 1H NMR spectrum of constrained neomycin B (2a)...... 108

Figure 4.6 gCOSY NMR spectrum of constrained neomycin B (2a)...... 109

Figure 4.7 gHSQC NMR spectrum of constrained neomycin B (2a)...... 110

Figure 4.8 gHMBC NMR spectrum of constrained neomycin B (2a) ...... 111

Figure 4.9 1H NMR spectrum of constrained paromomycin (2b) ...... 115

Figure 4.10 gCOSY NMR spectrum of constrained paromomycin (2b) ...... 116

Figure 4.11 gHSQC NMR spectrum of constrained paromomycin (2b) ...... 117

Figure 4.12 gHMBC NMR spectrum of constrained paromomycin (2b) ...... 118

xii

LIST OF TABLES

Table 2.1 Distance between C5'' and N2' for the neomycin/paromomycin when complexed with representative RNAs...... 42

15 Table 2.2 a) N NMR Determination of pKa Values for All Amine Groups. aTolerances indicate the standard error determined from curve fitting. b) Comparison of positive charge number carried by 1a and 2a under different pH conditions. a Tolerances indicate the standard error determined from curve fitting...... 56

Table 2.3 Target Binding Affinities of Aminoglycosides. a Tolerances indicate the standard deviation of at least three independent determinations. bFrom Alper et al..25 ...... 59

Table 3.1 Local minimum conformations of compound 33 from a conformational search (using HyperChem7.5, 300 iterative rounds of conformation search followed by AMBER filed energy minimization). Conformations are listed energetically from low to high. Type A conformers are highlighted in blue; type B in black. Boat/semi boat conformations are in red...... 92

xiii

ACKNOWLEGEMENT

Chapter 2 contains the material as it appears in Molecular Recognition of

RNA by Neomycin and a Restricted Neomycin Derivative. Zhao, F.; Zhao, Q.;

Blount, K.; Han, Q.; Tor. Y.; Hermann,T. Angew Chem, Int. Ed. 2005, 44, 2-6 and

Conformational Constraint as a Means for Understanding RNA-Aminoglycoside

Specificity. Blount, K.; Zhao, F.; Hermann, T.; Tor, Y. J. Am. Chem. Soc. 2005,

127(27), 9818-29. The dissertation author was the primary investigator and author of these papers.

xiv

CURRICULUM VITAE

March 2007 University of California, San Diego Doctor of Philosophy, Chemistry

June 2000 Beijing Normal University Bachelor of Science, Chemistry

Publications

Molecular Recognition of RNA by Neomycin and a Restricted Neomycin Derivative. Zhao, F.; Zhao, Q.; Blount, K.; Han, Q.; Tor. Y.; Hermann,T. Angew Chem, Int. Ed. 2005, 44, 2-6.

Conformational Constraint as a Means for Understanding RNA-Aminoglycoside Specificity. Blount, K.; Zhao, F.; Hermann, T.; Tor, Y. J. Am. Chem. Soc. 2005, 127(27), 9818-29.

ABSTRACT OF THE DISSERTATION

Aminoglycoside Derivatives and Mimetics as RNA Binders

Fang Zhao

Doctor of Philosophy in Chemistry

University of California, San Diego, 2007

Professor Yitzhak Tor, Chair

Ribonucleic acids (RNAs) are functionally sophisticated molecules that play key roles in essential biological processes such as protein biosynthesis, splicing, transcriptional regulation and retroviral replication. Due to the functional diversity of these molecules, a considerable amount of attention has been focused on developing

RNA–targeting therapeutics. Aminoglycoside antibiotics are a a family of structurally related natural products that have been found to bind and modulate the function of a variety of therapeutically significant RNA targets, including 16S ribosomal RNA, tRNA, HIV RRE and Tar RNAs, and group I introns. The target promiscuity of the aminoglycosides is likely due to their electrostatically driven binding mode and their conformational adaptability.

xvi

To study the role of the conformational flexibility of aminoglycosides in RNA binding and to improve the aminoglycosides RNA specificity, a type of novel conformationally constrained aminoglycoside derivatives designed to specifically targeting ribosomal A-site RNA were synthesized. In addition, a series of - cyclohexyl peptides aminoglycoside mimics were designed, which, with improved rigidity and established peptide chemistry, are suitable for screening RNA binders with high RNA specifies. In pursuance of this objective, a variety of amino acid building blocks were synthesized.

xvii

Chapter 1

1 2

1.1 Introduction

The central dogma of molecular biology describes the two–step process of gene expression (transcription and translation) through which genetic information is transferred from DNA to proteins by the mediation of RNA.1, 2

Transcription Translation DNA RNA Protein Reverse Transcription Replication

Figure 1.1 The central dogma of biology.

Despite its pivotal role in the gene expression process, RNA was only viewed as a passive information carrier for many years. The discovery of ribosome has changed this view and facilitated the first major paradigm shift.3-5 For the first time scientists realized that RNAs can possess catalytic properties previously thought to be restricted to protein–based enzymes only, making them active participants in the chemistry of life. For example, RNA molecules play key roles in the process of protein biosynthesis.6 The ribosome, a ribosomal RNA–protein complex, acts as the ―cellular factory‖ catalyzing the coupling of amino acids into with proteins under the direction of messenger RNA (mRNA). Ribosomal RNA (rRNA) accounts for 80% of the total cellular RNA and provides both the structural scaffold and enzymatic function for protein biosynthesis. Meanwhile, transfer RNA (tRNA), another specialized RNA

making up 15% of the total cellular RNA, serves as the raw material carrier responsible for transferring cognate amino acids to the ribosome (Figure 1.2).

Figure 1.2 Schematic view of protein synthesis.

A second paradigm shift of ―RNA knowledge‖ has been emerging more recently.

In the last twenty years, a large number of small noncoding RNAs that act as gene expression regulators has been identified.7 The importance of regulatory RNAs is becoming increasingly apparent. These RNAs impact almost all steps of the gene expression pathway, including transcriptional regulation, RNA transportation, RNA stabilizing, and splicing (Figure 1.3).

Figure 1.3 Steps at which RNAs have been found to modulate gene expression. Arrows and bars represent positive and negative regulations respectively. Abbreviations: small interfering RNA (siRNA); small nucleolar RNA (snoRNA); small nuclear RNA (snRNA); microRNA (miRNA).

Among all the regulatory RNA research, RNA interference (RNAi) has attracted great interest. RNAi is a phenomenon of post–transcriptional gene silencing (PTGS) induced by endogenous production or artificial introduction of double–stranded RNA

(dsRNA) with sequences complementary to the targeted gene into the cell.8-10 RNAi is distinct from other gene silencing phenomena in that silencing can spread from cell to cell and generate heritable phenotypes in first generation progeny when used in

Caenorhabditis elegans. Therefore, using dsRNA to induce RNAi may become a powerful method for the suppression of gene expression. The first RNAi application to reach clinical trials is in the treatment of macular degeneration.11 RNAi has also been shown to be effective in the complete reversal of induced liver failure in mouse models. Another speculative use of dsRNA is in the repression of essential genes in eukaryotic human pathogens or viruses that are dissimilar from any human genes; and this may hold the promises for the development of gene–specific therapeutics in future. 12

Riboswitchs are another type of regulatory RNAs.13 They are structures that form in mRNA which can directly bind a small target molecule. This binding of the target molecule triggers the riboswitch and affects (may activate or repress) the gene’s activity. Thus an mRNA with a riboswitch is directly involved in regulating its own activity, depending on the presence or absence of its target molecule. 14

In addition to its enzymatic and regulatory functions, RNA can also act as genetic material. About 65% of known viral families are defined as RNA viruses

(retroviruses).15 They either use RNA as their genetic material (i.e. HIV, SARS, avian influenza) or their genetic material passes through an RNA intermediate during replication (i.e. Hepatitis B virus). The replication of RNA viruses is an error–prone process due to their lack of DNA polymerases to find and edit out mistakes, which

can result in RNA viruses’ high mutation rates under selective pressure such as antiviral drugs.

In summary, RNA molecules are functionally sophisticated and play key roles in essential biological processes such as protein biosynthesis, splicing, transcriptional regulation and retroviral replication. Due to the functional diversity, a considerable amount of attention has been focused on developing RNA–targeting therapeutics.16-19

1.2 RNA as a Drug Target

Many human diseases are associated with malfunction of RNA–processing.

Fifteen percent of human genetic diseases result from RNA splicing defects.20 In addition, RNA–based viruses have further fueled the interest in the development of potential RNA inhibitors.15 Compared to proteins, which are the conventional drug target, RNA targeting therapy can modulate the disease related pathway at an earlier stage and prevent harmful protein biosynthesis.

Several oligonucleotide–based approaches have been developed to modulate cellular RNA level for biological studies or therapeutic purposes.16 The major methods currently in use are antisense oligonucleotides (ODNs),21-23 bioengineered ribozymes24, 25 and RNA interference (RNAi).8-10, 26 As these methods target specific sequences, the therapeutic agents are usually easy to design. However, the problem still remains is delivery efficiency, oligonucleotide stability, off–target effects and identification of sensitive sites in the target RNAs.16, 27 Recent reviews by Gewirtz and Rossi discussed the current advances in oligonucleotide–based approaches for mRNA knockdown. 16, 27

Compared to the delivery problems suffered by oligonucleotide–based technologies, cell permeable small molecules seem to be attractive RNA targeting agents. RNA shares some structural similarities with both DNA and proteins, making

RNA a unique biomolecule. In contrast to double helical DNA, RNA molecules exist in living cells mostly as single strands with a rich diversity of secondary and tertiary structures. RNAs fold back on themselves to form short double helices through

Watson–Crick base pairings. The short helices are connected by single strand regions such as hairpin loops, bulges, internal loops, junctions and pseudo knots (Figure1.4).

The helices, loops, bulges and junctions are all defined as RNA secondary structures.

The spatial arrangement of various secondary structures elements gives the complex three dimensional structures of RNA. RNA helical regions usually adopt A–form duplex conformations. Compared to B–form DNA, A–form helices contain a narrower and deeper major groove as well as a wider and shallower minor groove.

This difference may serve as one of the recognition elements for RNA/DNA selective ligands. The single stranded regions (loops, bulges and junctions) distort the neighboring helix and enlarge the major/minor groove by mismatched base pairings and unstacked bases, providing potential binding clefts and pockets as recognition sites for small molecules. The property of forming a vast array of complex three– dimensional structures makes RNA–ligand recognition similar to protein inhibition, in which small molecules recognize the shape of RNA instead of primary sequences as in the case of oligonucleotide–based therapies. However, this also makes RNA— ligand design more challenging.

a b

stem hairpin pseudo knot

bulge internal loop branch

Figure 1.4 a) RNA secondary structure feature. b) Tertiary structure of tRNA.

Theraputic examples of small molecules that bind RNA and interrupt important biological function include clinically–used antibiotics such as aminoglycosides28, macrolides29 and oxazolidinoes30 (Figure 1.5, compounds 1–3). These compounds exert their bactericidal activities by targeting bacterial ribosomal RNA and inhibiting protein biosynthesis.

Several natural or synthetic small molecules have been reported to bind RNA and regulate gene expression processes in vivo. In 1998, Green demontrated that small molecules can target a synthetic messenger RNA and modulate gene expression by trigging a mRNA riboswitch.11, 14 Reading from the 5' to 3' end, the mRNA contains an RNA aptamer (the switch), the start condon and a gene sequence. As the

switch is in front of the start codon, it will not be translated in any cases. However, this aptamer can bind to the small molecule and form a rigid complex. The rigid small molecule–RNA complex sits next to the start condon and prevents the ribozyme from scanning the messenger RNA, resulting the surpression of the gene expression.31

Therefore, the gene expression can be regulated depebnding on the present or anscent of the small molecule. Recent studies by Breaker have shown that natural systems use small molecule–RNA binding, which leads to the structural rearangement of the mRNA, to directly modulate mRNA translational efficiencies. 32-34

Natural and synthetic molecules targeting other disease–related RNA have also been discovered and developed (shown in Figure 1.5). All of these above studies have shown that small molecule–RNA interactions can be usefull in regulating biological processes and represent a promising therapeutic approach.

O NH2 O HO HO NH2 HO OH O NH2 OH O NH N HO O 2 N HO H OH O N N O O O O O

NH2 O F O O HO O OMe O OH NH2 OH O 3 HO 2 1

OH N OH H N HN OH OH N NH N 2 N HO N HO OH O OH O O N N N HO

4 5 6

N H2N NH2 R N N N N N HN R NH2 NH2 N N HN H2N N O OMe N N N N NH2 R R N O O2N Cl N

7 8 9 10

Figure 1.5 Representative RNA–binding small molecules: for bacterial rRNA1–5; for thymidylate synthase, 6; for HIV–TAR 7–9; for HIV–RRE10.

Among all the small molecules known to bind RNAs, aminoglycoside antibiotics are one of the most studied RNA binders.

1.3 Aminoglycosides

Aminoglycosides are a family of structurally related broad–spectrum bacterial antibiotics.35 They are natural products produced by gram positive bacteria as a defense mechanism against other bacteria for resource competition. Since the discovery of the first aminoglycosides, streptomycin, in 1944, a series of related antibiotics including kanamycin, neomycin, gentamicin and tobramycin have been successfully introduced to the clinic for the treatment of aerobic gram–negative bacterial infection. Recently, studies have shown that aminoglycosides also have antiviral activities.

The antibiotic and antiviral activities of aminoglycosides arise from their RNA binding ability. They are known to bind and modulate the function of a variety of therapeutically significant RNA targets. To date, aminoglycosides are the best studied RNA ligands with the most structural data available.36-49 RNA– aminoglycoside interactions are viewed as a paradigm that facilitates our understanding of how RNA and small molecules interact. The knowledge gained can be translated into the design and synthesis of ligands aimed at specific RNA targeting.

Aminoglycoside antibiotics are a group of structurally–diverse amino–sugars connected glycosidically with an aminocyclitol. Most of these antibiotics contain the aminocyclitol functionality of 2–deoxystreptamine (2–DOS), with a few exceptions

such as streptidine in streptomycin. Depending on the substituted position of 2–DOS, aminoglycosides fall into two categories, the 4, 5–disubstituted–2–DOS class and the

4, 5–disubstituted–2–DOS class (Figure 1.6).

OH 2 OH OH R R2 O 1 HO R OH Me O 1 3 NH O R R 2 HO NH2 MeHN O R5 HO HO HO O O 4 3 O O R R HN NH2 NH NH2 2

R1 R2 R3 R1 R2 R3 R4 R5 Kanamycin A OH OH H Gentamycin B OH OH OH H NH2 Kanamycin B NH2 OH H G-418 NH2 OH OH CH3 OH Tobramycin NH2 H H Gentamycin C1 NH2 H H CH3 NHCH3 Amikancin OH OH COCH(OH)(CH2)2NH2 Gentamycin C2a NH2 H H CH3 NH2

Gentamycin C2b NH2 H H H NHCH3

Gentamycin C1a NH2 H H H NH2 R1 O Kanamycin HO 2 R NH NH2 2 O HO O NH2 O OH NH NH 2 O 2 O HO HO NH2 HO HO O 3 NH2 HO NH2 OR NH2 NH2 O OH HO HO NH2 OH R1 R2 R3

Paromomycin OH OH H Neamine 2-Deoxystreptamine Neomycin B NH2 OH H

Livodomycin A OH H Man

OH NH O 2 H2N O OH OH HO OH HO HN HO OH HN NH2 O NH2 O O O O HO O NH2 O NHCH3 H2N O OH HO NH 2 O O CH3 NH HO 2 HO HO HO

Ribostamycin Spectinomycin Apramycin

H2N OH NH O NH HO HO OH OH OH HO OH H2N O HN O NH2 H2N O O Me O OHC HO MeHN O O OH HO O NH2 HO NH H N O O 2 2 NH H C NH2 2 3 OH HO O HO HO OH O HO NHMe HO

Sisomycin Streptomycin Hygromycin B

Figure 1.6 Representative aminoglycoside antibiotics

The mechanism of aminoglycosides’ bactericidal activities is not totally understood.50 Studies have shown that the small molecules exert their activities by binding to the ribosomal RNA Footprinting experiments51, 52 have shown that these antibiotics primarily target the prokaryotic 16S ribosomal RNA (Figure 1.7). For the paromomycin and kanamycin families, the binding site is located near the decoding site (A–site). Upon these antibiotics binding, ribosomal RNA undergoes a global conformational change, similar to that induced by the cognate tRNA, therefore leading to miscoding during translation and ultimate bacterial cell death.

Figure 1.7 The ribosomal 16S RNA sequence and the aminoglycoside binding sites.

In addition to the 16S RNA, aminoglycosides were also found to inhibit viral

RNA–protein interactions such as the HIV Rev peptide–RRE RNA interaction53 and the HIV Tat peptide–TAR RNA interaction.54 Since these RNA–protein interactions are essential for viral replication and gene expression, aminoglycosides have become of therapeutic interest for potential antiviral treatment. NMR and footprinting experiments have shown that neomycin inhibits the RRE–Rev interaction via a competitive mechanism. The binding site of Rev and aminoglycosides are both located at the purine–rich internal loop on RRE.55 Neomycin also inhibits TAR–Tat binding, however, in an allosteric fashion.40 The structural transition of TAR induced by neomycin binding is incompatible with the binding of Tat. The binding site of neomycin was identified by ribonuclease protection experiments to be the stem immediately below the three–nucleotide bulge that serves as the primary identity element for Tat.40

A number of other catalytic RNAs have been identified to be inhibited by aminoglycosides through the mechanism of aminoglycosides displacing essential metal ions. Among them are the self–splicing group I introns, hammerhead, human hepatitis delta virus, and hairpin ribozymes, as well as the tRNA processing activity of RNase P RNA.

The RNA affinity of aminoglycosides originates from their combined structural features, namely the polycationic functional groups, the various sugar rings and the connecting glycosidic bonds. The antibiotics interact with RNA mainly through electrostatic interactions between the positively charged ammonium groups on the small molecule and the negatively charged phosphate groups on the RNA. Natural abundance 15N NMR titrations have revealed that the amino groups of free aminoglycosides are predominantly positively charged under physiological conditions

(pH 7.4).56 These ammonium groups are believed to play a crucial role in RNA binding. However, there are a small portion of uncharged amino groups remaining on the free form of aminoglycosides. Due to the limitation of NMR and X–ray crystallography technology, we could not identify the protonation state of these amino groups in the bound form aminoglycoside. Recent thermodynamic experiments have revealed that some of these non–protonated amino groups undergo proton uptake upon binding and participate in the electrostatic interaction as well. This protonation process alone is energetically unfavorable; however it adds more positive charges on the small molecule and maximizes the RNA–ligand electrostatic interaction. Hence the protonation is a spontaneous and energetically favorable process in total.57 Due to the electrostatically driven nature, there is usually a correlation between RNA affinity and the number of amino groups on the ligand. For instance, by converting a hydroxyl group to an amino group, tsynthetic ―amino–aminoglycosides‖ have a higher inhibitory activity for hammerhead self–splicing and the Rev–RRE.58 Similar

trends are observed in naturally occurring kanamycin A vs. kanamycin B, paromomycin vs. neomycin and within gentamycin family.

Hydroxyl groups also have a significant effect on RNA affinity not only through their hydrogen–bonding but also through their impact on electrostatic interaction. A systematic study of deoxygenated tobramycin derivatives has shown that when a hydroxyl group proximal to an amine group is removed, up to twenty fold increase in

RNA inhibiting activities were observed.59 These observations were attributed to the increased basicity of an amino group upon removal of the neighboring hydroxyl group and further supported the critical role of electrostatic interactions in aminoglycoside–RNA binding.

R1 R4 2 R O H N R5 2 O NH2 NH R3 HO 2 O O H2N NH2

R1 R2 R3 R4 R5

Kanamycin B OH OH OH OH OH

Tobramycin OH OH OH H OH

Diberkacin OH OH OH H H

6"-Deoxytobramycin H OH OH H OH

4"-Deoxytobramycin OH H OH H OH

2"-Deoxytobramycin OH OH H H OH

Figure 1.8 Structure of natural and synthetic deoxy–tobramycin derivatives.

However, the RNA–aminoglycosides recognition phenomenon is far more complicated than simple ionic interactions, as simple polyamines (i.e. spermine), some aminoglycosides (i.e. apramycin) and other structurally unrelated antibiotics (i.e. viomycin) that possess comparable number of amines do not have RNA affinities.

NMR and X–ray crystallography provide valuable insight into the interaction mode.36-49 Despite the variability in the aminoglycoside–binding sequence and secondary structure, most of binding sites are located at the major groove of partial helical regions (double helixes with /near bulge, mismatch base pairs or loop residues), with the exception being the TAR–neomycin binding located in the minor groove at the lower duplex stem region of TAR.40 This major groove preference may be attributed to two major factors: (a) Shape complementarity. The glycosidic bonds that link the saccharide building blocks provide limited flexibility and allow the small molecule to adopt certain conformations that are favorable to A–form RNA with a ―loosened‖ major groove. This hypothesis is supported with DOCK computational experiments by identifying aminoglycosides as A–form major groove binders as opposed to the B–form minor groove binders.60 (b) Electrostatic complementarity. The major groove represents a region of deep negative electrostatic potential where maximum electrostatic interactions can form as the binding driving force. Molecular–dynamic simulations have revealed a striking congruence between the position of Mg2+ in the free ribozyme and position of the positive ammonium groups on neomycin. The three–dimensional framework of sugar rings together with

flexible glycosidic bonds provide a proper spatial arrangement of positively charged ammonium groups, which fit well in the negatively charged metal–ion–binding pockets created by the electrostatic field around the RNA (Figure 1.7).61 It is worth noticing that the limited flexibility around the glycosidic bonds is one of the major aminoglycosides structural features, which enables the complementarities of both the shape and electrostatic field between aminoglycosides and a variety of structure– different–RNAs .

a b c

Figure 1.9 Modeling shows the structurally electrostatic complementarity between hammerhead RNA and neomycin B. a) White sphere depict the position of magnesium ions in the crystal structure and green spheres show the ammonium groups in the aminoglycoside; b) Projection of charge densities of hammer head and neomycin (negative in red and positive in red), light blue sphere shows the position of Mg2+; c) charge densities of neomycin B, field lines (yellow) illustrate the gradient of the electrostatic field.

Solvation is also an important part in RNA–small molecule interactions. In solution, RNAs and aminoglycosides are surrounded by highly–order water molecules. These water molecules play a ―cushion‖ like role in mediating interactions around RNA.43-45 By adopting the optimal distance and orientation, the

solvent molecule assure the maximum inter–molecular and intra–molecular hydrogen bonding contacts of RNA. Meanwhile, water molecules can also adjust the electrostatic field by their orientations to a very fine extent and thus provide the optimal fit for ligand–RNA electrostatic interactions. This water mediating effects make RNA more adaptable to different binders and a more promiscuous target for various small molecules.

1.4 Design of RNA Binders

Aminoglycosides show excellent specificity for RNA over DNA, however, they are rather promiscuous RNA binders. They bind to and inhibit the function of a wide range of unrelated RNAs with moderate activities (IC50 range between 0.1–100M).

The multiple therapeutic effects and low–moderate bactericidal potency exhibited by these compounds may originate from their non–selective RNA affinities. Meanwhile, due to their widespread use and misuse, the emergence of bacterial resistance has become another important issue. Large amount of research has been done to improve their affinity, specificity and to overcome their bacterial resistance.

A). In designing RNA ligands, the ―conjugating strategy‖ of tethering known

RNA/DNA/protein binding elements to aminoglycosides is commonly used. The methodology has proved to be effective in improving RNA affinity and selectivity and even may produce RNA ligands with certain new quality such as Pt (II) conjugates shown later in this section.

1. Neomycin–acridine compounds are an example of effectively using the

conjugation approach to design RNA binders (Figure 1.8).62 By combining ionic

and intercalating binding modes, these molecules are designed to target both the

known neomycin binding site and the very close–by (if not overlapped) large

surface potential intercalation site on the RRE construct. Three linkers with

different length were used to seek the optimal fit. All three synthetic conjugates

have similar apparent Ki values around 3nM, which is approximately the same

affinity as that of the Rev peptide and among the strongest competitive inhibitors

of Rev–RRE binding to date. As expected, the linker length does have an effect

on selectivity, with the trend being the shorter the linker the more selective the

molecule is for RRE over DNA and tRNA. This observation can be explained by

the better adaptability of the long linker molecule compared to that of the short

linker one. Therefore, for the best affinity and selectivity, this approach is likely

to best work for short linker conjugates targeting RNA/protein which has the two

type binding elements close to each other as in the case of RRE.

NH2 NH2 NH2 O O O HO HO HO HO HO HO NH2 H2N NH2 H2N NH2 H2N O NH O NH O NH O 2 S O 2 H S O 2 O HO N N O HO N N S O HO N N H H O OH O OH O OH NH2 NH2 NH2 O O O OH OH OH H2N H2N H2N HO HO HO

Neo-N-acridine Neo-S-acridine Neo-C-acridine

Figure 1.10 Structure of Acridine–Neomycin conjugates.

2. Arginine–rich motifs are one of the key elements for RNA–peptide recognition

(i.e. RRE–Rev and Tar–Tat interaction). Inspired by this, the Tor research group

has synthesized ―guanidinoglycosides‖ in which all amino groups of natural

aminoglycosides were converted into guanidino groups (Figure 1.9).63 A

substantial increase in RRE affinity was observed for all the synthetic glycosides.

The selectivity of RNA over DNA is maintained and the RRE specificity was

significantly increased in most cases. The exception being a neomycin derivative,

which has six guanidino groups and may pass the threshold of the positive charge

number required for specific binding.

OH R1 R2 O HO O R1 OH HO 3 O 2 R HN NHR3 R HN NHR2 HO HO R2HN O 2 O 3 O NHR R3HN NHR HO O O OH

NHR2 O HO R1 R2 R3 O 2 R2HN NHR Kanamycin A OH OH H HO

1 2 Kanamycin B NH2 OH H R R

Tobramycin NH2 H H Paromomycin OH NH2

Guanidino-Kanamycin A OH OH NH(C=NH)NH2 Neomycin B NH2 NH2

Guanidino-Kanamycin B NH(C=NH)NH2 OH NH(C=NH)NH2 Guanidino-Paromomycin OH NH(C=NH)NH2

Guanidino-Tobramycin NH(C=NH)NH2 H NH(C=NH)NH2 Guanidino-Neomycin B NH(C=NH)NH2 NH(C=NH)NH2

Figure 1.11 Structure of aminoglycosides and guanidinoglycosides.

3. Cisplatin is a well–known anticancer drug that targets nucleotides, however with

a binding preference to DNA over RNA. In contract to the reversible interaction

of aminoglycosides with RNA, cisplatin reacts with DNA covalently, resulting in

permanent damage. Tethering platinum(II) to neomycin/guanidinoneomycin

renders the Pt (II) conjugates not only irreversible binders but also successfully

maintains the RNA selectivity (Figure 1.10).64

HN NH2 NH NH2 O HN O HO HO HO HO NH2 NH NH HN 2 H2N HN NH O NH NH O NH2 O 2 NH 2 NH NH 2 NH O NH O HO 2 Cl Cl O HO Pt Pt NH O NH 2 O OH NH NH2 O Cl 2 O OH Cl N NH O 2 OH H H N H O 2 OH HO H2N N HO NH

Pt-neomycin conjugate Pt-guanidinylneomycin conjugate

Figure 1.12 Structure of Pt–aminoglycoside conjugates.

B) The ―core derivative approach‖ keeps the aminoglycoside core unit intact while derivatizing different non–sugar functional groups or sugar moieties at various positions is also a widely used strategy for efficiently screening specific RNA binders as well as possibly overcoming existing bacterial resistance.65-75 This approach usually comes with a series of similar compounds with systematic variation of functionality and substituted positions, valuable information for understanding the structural requirements of specific RNA interactions and for directing the design of

RNA ligands.

Neamine is a commonly conserved moiety in many aminoglycosides. It was found to be essential for the bactericidal activities and cell permeability of aminoglycosides.76 A few libraries of neamine derivatives have been designed and synthesized by different research groups. Tobramycin65, 66 and pyranmycin67-69

analogs are neamine derivatives with various sugar moieties added at either the 6 or 5 position of the 2–DOS ring. They were designed for better RNA selectivities and affinities based on the observation that the RNA affinity and specificity of aminoglycosides varies when the composition or the linking positions of the additional sugar moiety change. Knowledge of correlations between amino–sugar assembling and RNA affinity origin is gained and provides valuable information for the future design of RNA binders.

NH2 NH2 O O HO HO HO HO NH2 NH2 NH2 NH2 O NH O NH HO 2 O 2 O OH R R n =1,2 n

Tobramycin Anologs Pyranmycin Anologs

Figure 1.13 Structure of tobramycin and pyranmycin analogs.

Neamine derivatives with various non–sugar functional groups were also studied.70-73 Among them, the Wong group has developed a series of compounds as potential antitumor agents that target oncogenic RNA.70 It is the first time that aminoglycoside derivatives act as high–affinity small–molecule ligands for the oncogenic sequence Bcr–Abl and PAX3–FKHR single stranded mRNA breakpoints.

The experiments have demonstrated the potential of aminoglycosides as gene expression regulators.

Similar to the neamine derivatives, libraries based on aminoglycoside core motifs such as the –hydroxyamine74 and ribostamycin75 have also been designed and explored as better RNA binders.

NH2 NH2 NH2 O O O HO HO HO HO HO HO NH2 H2N NH2 H2N NH2 H2N O O O O NH2 O NH2 NH2 HO HO RO HO

O R R 4-13 14-19 20-26

NH NH NH NH NH NH NH

OH OMe H3C F MeO

4 5 6 7 8 9 10

NH NH NH NH NH OH CH3 N N N N H H 11 12 13 14 (4:4:1)

NH N NH NH NH N NH N NH NH N N H H N

15 16 17 18 19(1:3)

CH3 H H H H H2N N N H N N H N N N 2 2 HO OH OH OH

20 21 22 23

N N H2N N N O H N H2N NH2

24 25 26

Figure 1.14 Structure of neamine derivative for targeting oncogenic RNA

C) ―Aminoglycoside mimics‖ are another method to develop new RNA binders. By keeping some of the aminoglycosides’ structural features, this approach may yield novel compounds with comparable or improved RNA affinities and specificities without being compromised by existing bacterial resistance mechanism.

Piperidine glycosides77 are a type of novel aminoglycoside mimics (Figure 1.13).

The design concept uses 3–aminomethyl–piperidine scaffolds to mimic the unique spatial arrangement of functional groups in 2–DOS which is the conserved moiety in many aminoglycosides. Piperidine mimics are less potent than natural aminoglycosides, but they have higher bacterial translation potency than other mimics with acyclic 2–DOS replacements. Furthermore, unlike the acyclic aminoglycoside mimics, some of which displayed promiscuous inhibition of eukaryotic translation, the active piperidine glycosides exclusively inhibit of the bacterial system. It was suggested that the conformationally restricted flexibility of the 3–aminomethyl– piperidine scaffold might be responsible for both the higher activity and specificity of piperidine glycosides towards the bacterial target.

a b c

1 H N 6 2

3 H2N NH2 5 4 6' O NH2

H2N OH

Figure 1.15 a) Representative designed piperidine glycoside ligand. b) Three–dimensional model of the designed piperidine glycoside (yellow) superimposed on paromamine (blue), showing their conformational similarity. c) Model of the piperidine glycoside 1 (yellow) docked in the three– dimensional structure of the bacterial decoding–site RNA in complex with paromomycin (blue; paromamine core only).

1.5 Goals

This thesis explores aminoglycoside–RNA interactions with a focus on aminoglycosides’ conformational effect on RNA specificity. Questions addressed are

(1) What role does the conformation of aminoglycosides play in RNA binding? (2)

How does one design aminoglycoside derivatives/mimics with improved RNA specificity through impacting its conformation.

To answer these questions we have a) Designed and synthesis conformationally constricted aminoglycosides. Florescent A–site binding affinity test and X–ray crystal gray experiments were done to study the impact of comformational constrain.

bpeptide aminoglycoside mimic were designed. The amino acid building blocks were synthesized and characterized.

References

(1) Crick, F. H. Symp Soc Exp Biol 1958, 12, 138-163.

(2) Crick, F. Nature 1970, 227, 561-&.

(3) Cech, T. R. Science 1987, 236, 1532-1539.

(4) Cech, T. R. Proc Natl Acad Sci U S A 1986, 83, 4360-4363.

(5) Cech, T. R. Science 2000, 289, 878-879.

(6) Green, R.; Noller, H. F. Annu Rev Biochem 1997, 66, 679-716.

(7) Storz, G.; Altuvia, S.; Wassarman, K. M. Annu Rev Biochem 2005, 74, 199- 217.

(8) Cogoni, C.; Macino, G. Curr Opin Genet Dev 2000, 10, 638-643.

(9) Gura, T. Nature 2000, 404, 804-808.

(10) Hammond, S. M.; Caudy, A. A.; Hannon, G. J. Nat Rev Genet 2001, 2, 110- 119.

(11) Cerutti, H.; Casas-Mollano, J. A. Current Genetics 2006, 50, 81-99.

(12) Dzitoyeva, S.; Dimitrijevic, N.; Manev, H. Proceedings of the National Academy of Sciences of the United States of America 2003, 100, 5485-5490.

(13) Vitreschak, A. G.; Rodionov, D. A.; Mironov, A. A.; Gelfand, M. S. Trends in Genetics 2004, 20, 44-50.

(14) Tucker, B. J.; Breaker, R. R. Current Opinion in Structural Biology 2005, 15, 342-348.

(15) Knipe, D. M.; Howley, P. M.; Griffin, D. E. Fundamental virology, 4th ed.; Lippincott Williams & Wilkins: Philadelphia, 2001.

(16) Opalinska, J. B.; Gewirtz, A. M. Nat Rev Drug Discov 2002, 1, 503-514.

(17) Hermann, T. Angew Chem Int Ed Engl 2000, 39, 1890-1904.

(18) Tor, Y. Chembiochem 2003, 4, 998-1007.

(19) Wilson, W. D.; Li, K. Curr Med Chem 2000, 7, 73-98.

(20) Krawczak, M.; Reiss, J.; Cooper, D. N. Hum Genet 1992, 90, 41-54.

(21) Lyer, R. P.; Roland, A.; Zhou, W.; Ghosh, K. Curr Opin Mol Ther 1999, 1, 344-358.

(22) Dias, N.; Stein, C. A. Mol Cancer Ther 2002, 1, 347-355.

(23) Branch, A. D. Hepatology 1996, 24, 1517-1529.

(24) Wilson, D. S.; Szostak, J. W. Annu Rev Biochem 1999, 68, 611-647.

(25) Sullenger, B. A.; Gilboa, E. Nature 2002, 418, 252-258.

(26) Fire, A.; Xu, S.; Montgomery, M. K.; Kostas, S. A.; Driver, S. E.; Mello, C. C. Nature 1998, 391, 806-811.

(27) Scherer, L. J.; Rossi, J. J. Nat Biotechnol 2003, 21, 1457-1465.

(28) Davis, B. D. Microbiol Rev 1987, 51, 341-350.

(29) Hansen, J. L.; Ippolito, J. A.; Ban, N.; Nissen, P.; Moore, P. B.; Steitz, T. A. Mol Cell 2002, 10, 117-128.

(30) Zurenko, G. E.; Gibson, J. K.; Shinabarger, D. L.; Aristoff, P. A.; Ford, C. W.; Tarpley, W. G. Curr Opin Pharmacol 2001, 1, 470-476.

(31) Werstuck, G.; Green, M. R. Science 1998, 282, 296-298.

(32) Winkler, W.; Nahvi, A.; Breaker, R. R. Nature 2002, 419, 952-956.

(33) Winkler, W. C.; Breaker, R. R. Chembiochem 2003, 4, 1024-1032.

(34) Nahvi, A.; Sudarsan, N.; Ebert, M. S.; Zou, X.; Brown, K. L.; Breaker, R. R. Chem Biol 2002, 9, 1043.

(35) Draper, D. E. Annu Rev Biochem 1995, 64, 593-620.

(36) Jiang, L.; Patel, D. J. Nat Struct Biol 1998, 5, 769-774.

(37) Francois, B.; Russell, R. J.; Murray, J. B.; Aboul-ela, F.; Masquida, B.; Vicens, Q.; Westhof, E. Nucleic Acids Res 2005, 33, 5677-5690.

(38) Francois, B.; Szychowski, J.; Adhikari, S. S.; Pachamuthu, K.; Swayze, E. E.; Griffey, R. H.; Migawa, M. T.; Westhof, E.; Hanessian, S. Angew Chem Int Ed Engl 2004, 43, 6735-6738.

(39) Jiang, L.; Suri, A. K.; Fiala, R.; Patel, D. J. Chem Biol 1997, 4, 35-50.

(40) Faber, C.; Sticht, H.; Schweimer, K.; Rosch, P. J Biol Chem 2000, 275, 20660-20666.

(41) Fourmy, D.; Recht, M. I.; Blanchard, S. C.; Puglisi, J. D. Science 1996, 274, 1367-1371.

(42) Jiang, L.; Majumdar, A.; Hu, W.; Jaishree, T. J.; Xu, W.; Patel, D. J. Structure 1999, 7, 817-827.

(43) Vicens, Q.; Westhof, E. J Mol Biol 2003, 326, 1175-1188.

(44) Vicens, Q.; Westhof, E. Chem Biol 2002, 9, 747-755.

(45) Vicens, Q.; Westhof, E. Structure 2001, 9, 647-658.

(46) Mikkelsen, N. E.; Johansson, K.; Virtanen, A.; Kirsebom, L. A. Nat Struct Biol 2001, 8, 510-514.

(47) Lynch, S. R.; Puglisi, J. D. J Mol Biol 2001, 306, 1037-1058.

(48) Varani, L.; Spillantini, M. G.; Goedert, M.; Varani, G. Nucleic Acids Res 2000, 28, 710-719.

(49) Yoshizawa, S.; Fourmy, D.; Puglisi, J. D. Embo J 1998, 17, 6437-6448.

(50) Kotra, L. P.; Haddad, J.; Mobashery, S. Antimicrob Agents Chemother 2000, 44, 3249-3256.

(51) Moazed, D.; Noller, H. F. Nature 1987, 327, 389-394.

(52) Moazed, D.; Stern, S.; Noller, H. F. J Mol Biol 1986, 187, 399-416.

(53) Zapp, M. L.; Stern, S.; Green, M. R. Cell 1993, 74, 969-978.

(54) Mei, H. Y.; Galan, A. A.; Halim, N. S.; Mack, D. P.; Moreland, D. W.; Sanders, K. B.; Truong, H. N.; Czarnik, A. W. Bioorganic & Medicinal Chemistry Letters 1995, 5, 2755-2760.

(55) Lacourciere, K. A.; Stivers, J. T.; Marino, J. P. Biochemistry 2000, 39, 5630- 5641.

(56) Botto, R. E.; Coxon, B. Journal of the American Chemical Society 1983, 105, 1021-1028.

(57) Pilch, D. S.; Kaul, M.; Barbieri, C. M. DNA Binders and Related Subjects 2005, 253, 179-204.

(58) Wang, H.; Tor, Y. Angewandte Chemie-International Edition 1998, 37, 109- 111.

(59) Wang, H.; Tor, Y. Journal of the American Chemical Society 1997, 119, 8734-8735.

(60) Chen, Q.; Shafer, R. H.; Kuntz, I. D. Biochemistry 1997, 36, 11402-11407.

(61) Hermann, T.; Westhof, E. Biopolymers 1998, 48, 155-165.

(62) Luedtke, N. W.; Liu, Q.; Tor, Y. Biochemistry 2003, 42, 11391-11403.

(63) Luedtke, N. W.; Baker, T. J.; Goodman, M.; Tor, Y. Journal of the American Chemical Society 2000, 122, 12035-12036.

(64) Boer, J.; Blount, K. F.; Luedtke, N. W.; Elson-Schwab, L.; Tor, Y. Angewandte Chemie-International Edition 2005, 44, 927-932.

(65) Liang, F. S.; Wang, S. K.; Nakatani, T.; Wong, C. H. Glycobiology 2004, 14, 1195-1196.

(66) Hanessian, S.; Tremblay, M.; Swayze, E. E. Tetrahedron 2003, 59, 983-993.

(67) Elchert, B.; Li, J.; Wang, J. H.; Hui, Y.; Rai, R.; Ptak, R.; Ward, P.; Takemoto, J. Y.; Bensaci, M.; Chang, C. W. T. Journal of Organic Chemistry 2004, 69, 1513-1523.

(68) Chang, C. W. T.; Hui, Y.; Elchert, B.; Wang, J. H.; Li, J.; Rai, R. Organic Letters 2002, 4, 4603-4606.

(69) Wang, J. H.; Li, J.; Tuttle, D.; Takemoto, J. Y.; Chang, C. W. T. Organic Letters 2002, 4, 3997-4000.

(70) Sucheck, S. J.; Greenberg, W. A.; Tolbert, T. J.; Wong, C. H. Angewandte Chemie-International Edition 2000, 39, 1080-+.

(71) Haddad, J.; Kotra, L. P.; Llano-Sotelo, B.; Kim, C.; Azucena, E. F.; Liu, M. Z.; Vakulenko, S. B.; Chow, C. S.; Mobashery, S. Journal of the American Chemical Society 2002, 124, 3229-3237.

(72) Greenberg, W. A.; Priestley, E. S.; Sears, P. S.; Alper, P. B.; Rosenbohm, C.; Hendrix, M.; Hung, S. C.; Wong, C. H. Journal of the American Chemical Society 1999, 121, 6527-6541.

(73) Park, W. K. C.; Auer, M.; Jaksche, H.; Wong, C. H. Journal of the American Chemical Society 1996, 118, 10150-10155.

(74) Wong, C. H.; Hendrix, M.; Manning, D. D.; Rosenbohm, C.; Greenberg, W. A. Journal of the American Chemical Society 1998, 120, 8319-8327.

(75) Alper, P. B.; Hendrix, M.; Sears, P.; Wong, C. H. Journal of the American Chemical Society 1998, 120, 1965-1978.

(76) Eggleston, D. S.; Prescott, C. D.; Pearson, N. D. The many faces of RNA; Academic Press: San Diego, 1998.

(77) Simonsen, K. B.; Ayida, B. K.; Vourloumis, D.; Winters, G. C.; Takahashi, M.; Shandrick, S.; Zhao, Q.; Hermann, T. Chembiochem 2003, 4, 886-890.

Chapter 2

37 38

2.1 Introduction

Aminoglycoside antibiotics such as neomycin (1a) and paromomycin (1b) bind to ribosomal RNA (rRNA) at the decoding site and thereby interfere with the accuracy of protein synthesis, ultimately leading to bacterial cell death (Figure 2.1).

In addition to the decoding site in 16S rRNA, several other RNA motifs form well– defined complexes with individual aminoglycosides, which makes these antibiotics excellent model ligands for the study of RNA recognition.1-6 The target promiscuity of the aminoglycosides can be attributed to two major factors: 1) their highly charged nature, which is responsible for their electrostatically driven RNA– binding mode and 2) their conformational adaptability. It is well established that the majority of oligosaccharides have some degree of flexibility around their glycosidic linkages and as a result can adopt several distinct low energy minima on their conformational surface. This limited flexibility explains some adaptability toward diverse RNA targets.4-6 The restricted conformational flexibility attenuates the contribution of charged interaction between RNA and the aminoglycosides, resulting the formation of well defined drug complexes that are distinct from nonspecific interactions of nucleic acids with flexible polyamines such as spermidine. For example, neomycin B, 1a, or the closely related paromomycin, 1b, (Figure 1a) assumes different conformations, when bound to different RNA targets (Figure 1b).7-

11 Although the relative positions of the individual neomycin rings differ among the complexes, the affinity with which neomycin or paromomycin binds each target is not dramatically different.

To circumvent this conformational flexibility and the resulting RNA target promiscuity, we proposed the ―conformationally constrained aminoglycosides‖.

Covalently "freezing" the aminoglycoside conformation may, under ideal circumstances, yield the following advantageous features: (a) increased affinity to the desired target due to limited entropy losses upon binding and (b) increased selectivity by locking the aminoglycoside skeleton in an unfavorable orientation for binding to competing targets.12, 13

This chapter describes the design, synthesis, crystal structures and thermodynamic studies of two new restricted aminoglycoside analogues, 2a and 2b

(Figure 2.2), which are specifically designed to exhibit enhanced selectivity for the

A– site relative to other RNAs.

OH 6' NH2 4' 5' O O HO II HO HO I/II HO 1' NH NH 2' 2 I/II 2 3' NH2 4 3 2 NH2 O I O NH HO 5'' O NH2 HO O 2 6 1 II/III 5 OH II/III OH O 1'' O 4'' III 3'' 2'' III/IV NH2 NH2 O HO 6''' O HO III/IV 5''' 1''' O O IV OH NH OH NH 2 3''' 2 2''' HO 4''' HO

1a 1b B) C) ) )

Paromomycin/A– Neomycin/Tar site E) D) ) )

Neomycin/ exon A Neomycin/aph(3’)III

– site A– site

Figure 2.1 a) Chemical structures of the neomycin class aminoglycosides studied. Neomycin B (1a) and paromomycin (1b) differ only at the 6' position. Ring numbers are shown in roman numerals. Torsional angles for the glycosidic linkages defined as I/II (O1'– C1'– O4– C4), I/II (C1'– O4– C4– C3), I/III (C5– O5– C1''– O1''), I/III (C4– C5– O5– C1''), III/IV (O1'– C1'– O4– C4) and III/IV (C1'– O4– C4– C3). b) The conformation of paromomycin bound to A– site11, neomycin bound to the TAR8, a SELEX– derived aptamer9 and the Tau exon.10

2.2 Design of Conformationally Restricted Aminoglycosides

In designing the constrained molecules, we asked several questions: Which

RNA is the best target for ―constrained aminoglycosides‖? What aminoglycosides are the most suitable candidates to be constrained? Where and how are we going to constrain the molecules? To answer these questions, we examined all the RNA– aminoglycoside X–ray crystallography and NMR structures available. 7-11 These structural data revealed some interesting information listed below:

A) The relative orientation of rings I and II is very similar in all the RNA complexes,

whereas the conformation around the linkages to rings III and IV is significantly

variable depending on the RNA target.12 The apparent rigidity of the ring I/II

system underlines the importance of this module for RNA recognition, attested

by the fact that the 2– DOS and ring II moieties participate in key interactions

that are responsible for target binding in decoding– site complexes with

aminoglycosides.7, 11, 14 Whereas higher thermal factors of rings III and IV in the

crystal structure of paromomycin might suggest that these sugars generally

contribute less to target– specific interactions.7, 11

B) Paromomycin adopts compact conformation when complex with A– site RNA;

to conformationally constrain itself, paromomycin forms an intramolecular

hydrogen bond. All three published structures of paromomycin bound to the A–

site reveal a similar structure, with a very compact arrangement of the four rings

and rings II and III in spatial proximity.7, 11, 14 To maintained this compact shape,

a naturally occurring intramolecular hydrogen bond between the 2'– amino group

of ring II and the 5''C atom of ring III ( distance at 3.7 Å) is observed in the

2.5Å resolution Westhof’s structure.10, 11

C) In contrast, when bound to other RNA molecules, the antibiotics are more

dynamic and assume a more extended conformation.8-10 For instance, the

published structure of TAR– bound neomycin includes a collection of 17 refined

NMR structures and reveals two primary conformations.8 In both conformations,

rings II and III are more distal to one another, with rings II and III distal to one

another and the distance between 2'–amino group and the 5''–C being 6.9 ± 0.6 Å

(standard deviation). Notably, the torsional distance between C5'' and N2' for the

neomycin/paromomycin complexed with RNAs.angles between rings I and III in

both TAR– bound conformations are very different from the torsional angles in

any of the A– site bound paromomycin structures. Table 2.1 lists the distance

between C5'' and N2' for the neomycin/paromomycin when complexed with

various RNAs.

Table 2.1 Distance between C5'' and N2' for the neomycin/paromomycin when complexed with representative RNAs.

RNA A-site Tar  Exon 10 Sre Aph(3')III O C5''to N2' distance(A) 3.7 6.4 5.5 4.0

Intrigued by these observations, we reasoned that a short covalent link between

2'– amino group and the 5''– C being 6.4Å should constrain the aminoglycoside into a structure resembling the A– site bound form, thereby altering its RNA affinity and target selectivity, identifying compounds 2a and 2b as our target molecules (Figure

2.2). Since the newly formed covalent bond would span a length of approximately

1.5 Å, these restricted aminoglycosides were expected to bind more readily to the A– site (3.7 Å in crystal structure) than to the TAR (averagely 6.9 Å in NMR structure).

NH2 OH O O HO HO HO HO NH2 NH2 NH2 NH2 O O NH O NH2 O 2 OH O OH O

NH2 NH2 O HO O HO

O O OH OH NH2 NH2 HO HO 2a 2b

Figure 2.2 Conformationally constrained neomycin (2a) and paromomycin (2b).

To predict the conformational effect induced by the covalent linker, we performed molecular dynamics simulations of 1a and 2a in water. Using the

AMBER molecular dynamics simulation program, 100 iterative rounds were performed of phases of heating to 5000 K followed by cooling. Following energy minimization, the torsional angles between each of the first three rings of the cooled conformation were recorded, and their distributions were examined (Figure 2.3).

Interestingly, the conformational space available to the restricted 2a is very similar to that available to 1a. The primary difference is that the distribution of different conformations is somewhat more scattered for the nonrestricted 1a, particularly around the glycosidic bond connecting rings I and III (Figure 2.1). This difference is consistent with the expected reduction of conformational flexibility enforced by the covalent linkage. Importantly, the torsional angles in the A– site– bound paromomycin structures (triangles) are very similar to those predicted in this simulation for neomycin and restricted neomycin (open and filled circles, respectively). The slight differences that are observed likely represent minor structural adaptations that occur upon RNA binding.

Figure 2.3 Torsional angles of aminoglycosides free in solution or bound to A– site or the TAR RNAs. A) Torsional angles encountered during molecular dynamics simulations in water are shown for neomycin 1a (open circles) and restricted neomycin 2a (closed circles). Note the tighter distribution of 2a conformations. Torsional angles of TAR– bound neomycin are shown in blue. Torsional angles of paromomycin 1b when bound to the A– site as determined by NMR7 or crystallography11 are shown in green or red, respectively.

It is informative to locate the TAR– bound neomycin conformation within this map. The published structure of TAR– bound neomycin includes a collection of 17 refined models which best fit the observed NMR data. The range of torsional angles between rings I and III within these models (blue, Figure 2.3) represents a conformational family that is distinctive from that encountered in the solvent simulation for either 1a or 2a. Consistent with our design concept, this predicts that it would be energetically disfavored for the restricted 2a to assume a conformation necessary for binding to the TAR. Thus, both the conformations obtained from the molecular dynamics simulations and the 2'N/5''C distance constraint imposed by the linker predict that 2a and 2b should bind less well than the parent 1a and 1b to the

TAR, thereby altering target selectivity.

2.3 Synthesis

For the synthesis of restricted aminoglycoside analogues 2a and 2b we devised an intramolecular and regioselective cyclization strategy of an unprotected 5''– activated neomycin intermediate under high– dilution conditions that exploits the proximity of the 2'–amino and 5''–hydroxymethyl groups (Figure 2.4). A fully Boc– protected neomycin15 was activated at the 5'' position with 2,4,6– triisopropylbenzenesulfonyl chloride to give 4. TFA– mediated cleavage of all Boc groups, followed by dilution and neutralization with Et3N, facilitated an intramolecular cyclization to give the desired 2a. To simplify isolation, this highly polar product was first protected as the penta– Boc derivative 6 and column purified.

Acidic deprotection of all Boc groups followed by reversed– phase purification afforded 2a in 12% overall yield (based on 4).

Conformationally restrained paromomycin derivative 2b was prepared in a fashion similar to 2a (Figure 2.5). To ensure 5''– selective activation of Boc paromomycin 7 that contains two primary alcohols, the 6'– hydroxyl was protected as the isopropylidene ketal. The activation of the 5'' position proceeded smoothly to yield 9. The ketal group was then conveniently removed under the same conditions utilized for Boc deprotection, providing the 5''– activated paromomycin. After neutralization and high dilution cyclization, followed by a similar protection/purification/deprotection scheme, the conformationally restrained paromomycin 2b was obtained in 20% overall yield (based on 9).

NH2 NHBoc O O HO OH HO OH NH2 NHBoc H2N BocHN O O HO O NH2 OH O NHBoc OH OH O a O

NH2 NHBoc O HO O HO

O O OH OH NH2 BocHN HO OH 1 3

+ NH3 O NHBoc HO O HO NH + +H N 3 HO O 3 HO O NH + NHBoc Ar S O O 3 O BocHN OH O NHBoc O O Ar S O O OH c O O + NH3 O HO NHBoc O HO O + OH H3N O HO BocHN OH HO 5 4

NH + 3 NHBoc O HO O HO HO NH + +H N 3 HO NHBoc 2 HN O + O NH3 O O NHBoc O OH O OH e + NH3 O HO NHBoc O HO O f O +H N OH 3 BocHN OH HO HO 2a 6

Figure 2.4 Synthesis of Conformationally Constrained Analogues 2a. Reagents and conditions: a)

Boc2O, NEt3, methanol, 82%; b) 2,4,6– triisopropylbenzene–sulfonyl chloride, pyridine, 70%; c) TFA, CHCl3; d) NEt3, DMF; e) Boc2O, NEt3, methanol, 12% (c –e ); f) TFA, CHCl3, 97%. Ar = 2,4,6– triisopropylbenezene.

OH OH O O HO HO HO HO NH2 NHBoc H2N BocHN O O HO O NH2 HO O NHBoc OH OH O a O

NH2 NHBoc O HO O HO

O O OH OH NH2 BocHN HO HO 1b 7 b

O O O O O O HO HO NHBoc NHBoc O BocHN BocHN O O NHBoc Ar S O O NHBoc HO O OH OH O c

NHBoc NHBoc O HO O HO

O O OH BocHN OH BocHN HO HO 9 8

OH OH OH O O O HO HO HO + + HO HO + NH3 HO + NH3 H2N NHBoc O H3N + HN + O NH3 O NH3 O NHBoc Ar S O O O O OH O OH OH O O e f O

+ NH + NH3 3 O HO NHBoc O HO g O HO O O O + OH + OH H3N OH H3N BocHN HO HO HO 10 2b 11

Figure 2.5 Synthesis of Conformationally Constrained Analogues 2a. Reagents and conditions: a)

Boc2O, Et3N, methanol ; b) 2,2– dimethoxypropane, p– toluene sulfonic acid, 79%; c) 2,4,6– triisopropylbenzene– sulfonyl chloride, pyridine, 71%; d) TFA, CHCl3; e) NEt3, DMF; f) Boc2O,

NEt3, methanol, 20% (d– f); g) TFA, CHCl3, 99%. Ar = 2,4,6– triisopropylbenezene.

It is important to note that protection of both conformationally constrained analogues 2a and 2b has always yielded (n–1) protected derivatives 6 and 11, respectively (where n is the total number of amines), as confirmed by NMR and MS analysis.16 In both cases the least reactive amine is located at the the newly formed bridging secondary amine. This lower nucleophilicity may be a result of steric or electronic effects. These reactivity observations are pertinent to the RNA binding characteristics of the new derivatives as elaborated below.

2.4 Spectroscopic Characterizations

The intramolecular cyclization imposes a significant conformational constraint on the aminoglycosidic skeleton. In terms of composition, however, the cyclized structures 2a and 2b can essentially be viewed as "dehydrated" antibiotics (i.e., the two structures differ only by a molecule of water). To unequivocally prove the formation of the N– C bond between the amine at the 2' position and the 5'' carbon on the D– ribose, a series of NMR experiments has been conducted to assign all resonances and establish the cyclization site.

A qualitative comparison of the 1H NMR spectra of the restricted neomycin derivative 2a and its parent natural product 1a reveals intriguing and informative changes (Figure 2.6 top). The 1H NMR spectrum of the cyclized derivative, under the same pH and counteranion conditions, shows a dramatic upfield shift of the 5'' methylene group, when compared to neomycin's spectrum. This observation is in full agreement with the replacement of an oxygen by a nitrogen functional group replacement. The change in chemical shift pulls the 5''– Hb peak out of a crowded spectral region and significantly facilitates its assignment. Interestingly, the rest of the proton signals of the D– ribose system (ring III) experience a dramatic downfield shift. The anomeric proton (H1'') shifts further downfield to H1', a rare observation for neomycin derivatives. These extraordinary changes in chemical shifts are likely due to a significant conformational distortion of the D– ribose ring enforced by the intramolecular cyclization.

The full assignment of compound 2a (Figure 2.6, assignments of protons and

Carbons are shown in 1H–13C gHSQC spectrum) was accomplished using 1D proton

NMR as well as 2D gradient–assisted correlation spectroscopy (gCOSY), gradient– assisted heteronuclear single quantum coherence (1H–13C gHSQC) and heteronuclear multiple bond correlation (1H–13C gHMBC) spectroscopies. The anomeric protons of rings II, III, IV ( 5.0– 6.0) and the methylene protons of ring I ( = 2.58 and 1.94) provide a convenient starting point, and gCOSY spectra assist in assigning the proton system of each ring (see Supporting Information for all spectra). Phase sensitive gHSQC spectra provide a definitive assignment for all methylene groups at positions

2, 6', 5'' and 6''' (see Figure 2.1 for numbering). Edited gHSQC spectra differentiate the carbons connected to oxygen ( > 55.0) from the carbon centers connected to nitrogen or other carbon atoms ( < 55.0), with one exception: 2'– C at 60.2 ppm

(Figure 2.6).17 This significant downfield shift is consistent with converting the primary amine at this position to a secondary amine.18 Full assignment was not possible until 1H– 13C gHMBC spectroscopy unequivocally secured the assignment of ring II and ring IV (both are 2,6–amino pyranose) by showing correlations across the glycosidic bonds (i.e., correlations were observed from ring II to ring I as 4–H to 1'–C,

4–C to 1'–H, from ring IV to ring III as 3''– H to 1'''–C, 3''–C and 1'''–H).

Figure 2.6 1H NMR (top) and 1H– 13C gHSQC (bottom) spectrums of restricted neomycin 2a. Secondary carbon crosspeaks show up as negative phase in gHSQC, which makes them easy to be identified. The key assignment of 5''H–5''C and 2'H–5''C crosspeaks are highlighted in.

It is important to note that 1H–13C gHMQC is the method of choice to confirm the cyclization site by demonstrating a correlation between the two nuclei across the newly formed secondary amine. Indeed, a cross– peak between 5''–Hb and 2'–C observed for both 2a and 2b unambiguously cements the ring connectivity (Figure

2.7).19

Figure 2.7 1H– 13C gHMBC spectrum of restricted neomycin B 2a (left) and an expansion (right) highlighting the cross– peak from 5''–Hb to 2'–C.

2.5 15N NMR Studies and pKa measurement

To further cement the proposed structure and to evaluate the basicity of the new restricted antibiotics, 15N NMR spectra of the TFA salts of compound 2a and neomycin 1a were recorded. Comparing the spectra at pH > 10.5 shows that while most amines exhibit similar chemical shifts, the 15N signal of the 2' amine in 2a is approximately 5 ppm upfield shifted compared to the corresponding amine in the parent neomycin B 1a (Figure 2.8). This observation further supports the higher substitution of this particular nitrogen and is in agreement with the 13C NMR and the reactivity data discussed above.

Figure 2.8 15N NMR spectra of restricted neomycin 2a and neomycin 1a at pH 10.7 show the distinct chemical shift of the 2'– nitrogen. The chemical shifts of each nitrogen are shown below as a function of pH. Symbols highlight individual nitrogens: 1 (open triangles), 2 (closed triangles), 2' (blue filled squares), 6' (filled circles), 2''' (open squares), and 6''' (open circles).

The pKa values were determined by fitting the data to the equation 2.1 shown below using Kaleidagraph:

pH-pKa  NH2-NH3   NH pH-pK 2 ······················ 2.1  a

Recording the 15N NMR spectra as a function of pH results in well– defined titration curves that can be fitted to yield the pKa value of each amine (Figure 2.8 and Table 2.2).20, 21 The newly formed bridging secondary amine is found to have significantly attenuated basicity. With a pKa value of 6.37, it is almost 2 pKa units below the amine at the same position in neomycin B (pKa 8.14). The lower basicity of this amine may be due to its crowded environment that hinders suitable solvation of the corresponding ammonium ion (and is in agreement with the low nucleophilicity exhibited by this functional group). Interestingly, the lower basicity of this center renders all other amines in 2a more basic when compared to their corresponding amines in neomycin 1a, as would be expected based on intramolecular electrostatic considerations. For example, the pKa of the NH2 at position 1 of 2a is 0.4 units higher than the analogous amine in neomycin 1a. While these absolute pKa values are likely to depend on the counterions, the experimentally determined values clearly reflect the impact of cyclization on the basicity of the 2'–amine. 21

Based on the pKa value obtained, we calculated the total positive charges carried by neomycin 1a and constrained neomycin 2a at three pH conditions (7.5, 6.8, 5.8).

At pH 7.5, the constrained analog has 4.0 positive charges, which is 0.6 less than its

56 parent compound. As pH goes lower, the difference is diminished to 0.1 at pH 5.8.

This result indicates that compared to the unconstrained compound 1a at physiological conditions (pH 7.5), the RNA binding activity of constrained analogue

2a might suffer from charge loss due to the 2'N modification. As the pH goes lower enough, this charge difference and the possible resulting affinity difference could diminished.

15 a Table 2.2 a) N NMR Determination of pKa Values for All Amine Groups. Tolerances indicate the standard error determined from curve fitting. b) Comparison of positive charge number carried by 1a and 2a under different pH conditions. a Tolerances indicate the standard error determined from curve fitting.

a) neomycin (1a) restricted neomycin (2a)

     Amine NH ppm NH ppm  NH ppm  ppm pK 2 3 pKa 2 NH3 a 1 33.5 ± 0.2 40.5 ± 0.2 8.50 ± 0.07 32.9 ± 0.1 40.2 ± 0.1 8.89 ± 0.04 3 35.8 ± 0.1 42.0 ± 0.3 6.59 ± 0.08 35.9 ± 0.1 41.8 ± 0.1 6.68 ± 0.04 2' 25.1 ± 0.1 36.6 ± 0.1 8.14 ± 0.03 29.7 ± 0.1 40.9 ± 0.2 6.37 ± 0.03 6' 17.2 ± 0.4 28.7 ± 0.2 9.36 ± 0.08 16.0 ± 0.1 28.4 ± 0.1 9.45 ± 0.03 2''' 18.7 ± 0.1 32.3 ± 0.1 8.07 ± 0.02 18.9 ± 0.1 32.0 ± 0.1 8.34 ± 0.03 6''' 18.9 ± 0.3 29.0 ± 0.2 9.65 ± 0.07 18.9 ± 0.2 29.3 ± 0.1 9.65 ± 0.04

b) restricted neomycin  charge neomycin (1a) (2a)

pH 7.5 4.0 4.6 - 0.6

pH 6.8 4.7 5.3 - 0.6 pH 5.8 5.7 5.8 - 0.1

2.6 Ligand Binding to the A– Site RNA

Two recent contributions report the incorporation of the fluorescent nucleoside analogue 2– aminopurine (2AP) into the A– site as a means for detecting and quantifying aminoglycoside binding.22, 23 The 2AP substitution at A1492 accurately reports the unstacking of this nucleotide upon binding of the neomycin– class antibiotics. Notably, the magnitude of the 2AP fluorescence increase upon binding is relatively constant among different aminoglycosides and is strongly correlated with the specific structural changes that occur upon binding.22, 23 Capitalizing on these reports, the binding of the aminoglycosides discussed here was measured by titration into a fixed concentration (~200 nM) of A– site 2AP (1492) (Figure 2.9).24 Using the directly measured 1:1 stoichiometry of paromomycin binding to the A– site, the fluorescence increase upon binding of the aminoglycoside was converted to fractional saturation and plotted against the concentration of the unbound aminoglycoside. The binding isotherm fits well to a two– state binding model and yields a reproducible measure of binding affinity. Using this assay, paromomycin binds to the A– site at pH 7.5 with a Kd of 170 nM (Table 2.3), which compares favorably with literature reports.25-27 Much like previous studies, 25-28 under the conditions used to measure paromomycin binding, neomycin binding to the A– site does not adhere to a simple two– state model and does not yield reproducible affinity values from this assay.

Therefore, for the purpose of comparison, the approximate dissociation constant for the first neomycin binding event is compared here to the value of 19 nM at pH 7.5 measured by surface plasmon resonance.25 Notably, restricted– neomycin 2a does

58 show single– state binding, much like paromomycin. This may implicate the 2'– amine as a contributor to neomycin's nonspecific binding.

a b U C U C U G U G C G 1488 C G C G C G N N A U A U C G 1492 C G N N NH O 2 A A A A A A O C G 1494 C G U U U U O OH 14 0 5 G C G C C G C G G C G C 2- aminopurine 1492 140 2 G C 1499 G C 5' 3' 5' 3'

c d

Figure 2.9 a) RNA construct A– site 2AP (1492) containing the ribosomal decoding site used in this study. Model schematicallyillustrates the conformational changes trigged by aminoglycosides (blue ball) binding to the decoding site. The sequence in bold is identical to the bacterial decoding site. Position 1492 is highlighted in red. b) The structure of fluorescent nucleoside 2– aminopurine (2AP) that is incorporated at position 1492. c) Sample binding isotherms of paromomycin (1b) and restricted paromomycin (2b) binding to A– site 2AP (1492) at pH 7.5. d) The dependence of dissociation constants on pH. Error bars indicate plus or minus one standard deviation of three independent measurements.

Table 2.3 Target Binding Affinities of Aminoglycosides. a Tolerances indicate the standard deviation of at least three independent determinations. bFrom Alper et al..25

Kda pH 5.8 (M) Kd pH 6.8 ( M)) Kd pH 7.5 ( M) neomycin B 1a ND ND 0.02b restricted neomycin 2a 0.07 ± 0.02 0.16 ± 0.01 0.42 ± 0.03 paromomycin 1b 0.05 ± 0.02 0.06 ± 0.01 0.17 ± 0.01 restricted paromomycin 2b 0.12 ± 0.03 0.35 ± 0.07 2.4 ± 0.4

At pH 7.5, the restricted– neomycin 2a and restricted– paromomycin 2b bind the

A–site with 22–and 14–fold lower affinity, respectively, compared to neomycin and paromomycin, the parent natural products. This decreased binding affinity is likely to be largely due to the lower overall charge of the cyclized derivatives due to the diminished basicity of the 2'–amine in the restricted molecules. Indeed, upon decreasing the pH to 5.8 to more fully protonate this amine, the affinities of restricted– neomycin and restricted– paromomycin are only 3.8– and 2.3– fold lower than the parent aminoglycosides. Thus, under conditions where the overall protonation states are similar, the conformational restriction of these neomycin– class antibiotics only modestly decreases their binding affinity to the A– site.29

2.7 Crystal Structure

In collaboration with Dr. Thomas Hermann and his coworkers, we studies the

X–ray crystal structures of the conformationally restricted analogue neomycin 2a and its parent compound neomycin 1a to exam weather the cross–linking between 5'' and

2' positions may yield an aminoglycoside that retains the decoding–site–bound conformation of the parent drug.30 Herein, we analyze the structural characteristics of both RNA–small–molecule complexes and discuss the structural basis of aminoglycoside ligand affinity along with general implications on the understanding of RNA recognition.

Both neomycin 1a and restricted neomycin 2a were co–crystallized with an oligonucleotide that contains the decoding–site sequence and flanking bases, which facilitated crystal packing (Figure 2.10). Similar small RNAs have been shown to provide authentic model systems that retain the structural and dynamic characteristics of the ribosomal decoding site.7, 11, 31, 32 Three–dimensional structures of the RNA complexes of 1a and 2a were determined by X–ray diffraction by using anomalous dispersion of the halogen atom in the 5–Br–U1487 residue. For comparison, the structure of the unbound decoding site was solved as well.30

Figure 2.10 Secondary structure of the decoding– site oligonucleotide used for X– ray crystallography and structure determination. The box indicates the region that corresponds to the bacterial RNA sequence. Flanking nucleotides were added to enhance crystal packing. Residues are numbered according to the Escherichia coli 16S rRNA sequence. For multiwavelength anomalous diffraction experiments, U1487 was replaced by 5– Br– U.

The electron density of the decoding– site complexes clearly revealed the identity of the bound aminoglycoside ligands (Figure 2.11 a,b). Specifically, the 5''– hydroxyl– methyl group in neomycin 1a gave rise to a characteristic extension of the electron density envelope which was lacking for the cyclized derivative 2a. Both 1a and 2a bind to the decoding site at the same position as paromomycin 1b.11, 14

Complex formation induces a more compact structure of the RNA around the ligand– binding site as indicated by the base of A1408 and the sugar–phosphate backbone of

A1492 and A1493 pulled closer to the aminoglycoside (Figure 2.12). The hydrogen– bonding contacts between neomycin and the RNA are identical to those of paromomycin, including the interaction of the distinctive 6' substituent (Figure 2.13).

The N1 atom of A1408 in the decoding site forms a hydrogen bond with either the 6'–

62 hydroxyl group of paromomycin or the 6'–amino substituent of neomycin which helps stabilize the base–pair–like interaction between ring II and the A1408 residue. 11

a) 3’ b) 3’ 5’ 5’

I A1493 I A1493 IV IV II II A1492 A1492

Neomycin restricted Neomycin

c) 3’ 3’ 5’ 5’

I A1493 I A1493

II II A1492 A1492 A1408 A1408

d) e)

IV I IV I III III

II II

Figure 2.11 Crystal structures of neomycin 1a (a) and restricted neomycin 2a (b) complexed with decoding– site RNA. Aminoglycoside rings are numbered according to the scheme in Figure 2.1a. Electron density (2Fo–Fc, contoured at 1.0 around the ligands) reveals the identity of the aminoglycosides, specifically the extended 5‖– hydroxymethylene group in neomycin which is absent in the restricted ligand (arrows). c) Stereo view of a superimposition of neomycin (1a, green) and its restricted derivative (2a, yellow) bound to the decoding site (superimposition based on phosphate coordinates; only RNA of the neomycin complex shown). The link between the 2’–and 5‖– positions in 2 is highlighted in ball–and–stick representation. Ligand superimposition of neomycin (1a, green) is shown in (d) with restricted neomycin (2a, yellow) and in (e) with paromomycin (1b, orange).

3’ A1492

A1493 5’

G1494 A1408

Figure 2.12 Overlay of the crystal structures of the unliganded decoding-site RNA (orange) and the neomycin complex (blue) (superimposition based on phosphate coordinates). View along the helix axis with ribbon representation of the RNA backbone and stick model for key bases. Neomycin ligand is in green. In the neomycin complex, the RNA structure is more compact around the ligand binding site, as indicated by the base of A1408 and the sugar phosphate backbone of the unpaired residues A1492 and A1493 pulled towards the aminoglycoside.

The position of rings I and II within restricted neomycin 2a relative to the decoding– site RNA coincides with the binding site of these rings in the neomycin–

RNA complex (Figure 2.11c). Consequently, the hydrogen– bonding network between the RNA and rings I and II is identical for the restricted derivative and neomycin as well as paromomycin (Figure2.13). Distinct patterns are, however, observed for the interactions of rings III and IV. Intramolecular cyclization imparts to 2a with a slightly more compact overall structure than that of the parent neomycin, reflected by differences in the positioning of rings III and IV (Figure 2.11 c,d). Ring

IV of 2a interacts with the phosphate group of G1491, whereas hydrogen bonds are formed to the phosphate group of G1405 in the opposite RNA strand for the complexes formed with neomycin and paromomycin (Figure 2.13).11 The covalent link between rings II and III in 2a pulls ring III away from the bottom of the RNA deep groove which leads to disruption of a hydrogen– bond interaction to C1407 that involves the 2''–hydroxy group; this interaction is present for complexes with neomycin and paromomycin. Another contact of the 5'' hydroxy substituent in the natural aminoglycosides with G1491 is absent for the restricted neomycin.

Interestingly, the contacts to rings III and IV, which were considered to play a secondary role for decoding– site binding, are sacrificed upon binding of 2a to the decoding site. This is not the case for the interactions with rings I and II, the importance of which in RNA target recognition is underscored once again.

G1491

OH NH2 P IV U1406 O O O O N O 3" H N H N OH U1495 O H OH O H A1492 N N 3" 1" P O I H 5" O 5" H O O 1" O III N H 4 O H N O O H 4 O 1' 1' N HO H O H O O II H H H N N H P N H O N N H N A1493 N H N N N N N N O H N A1408 C1407 H G1494

Figure 2.13 Interactions of rings I, II, and IV in restricted neomycin (2a) with the decoding– site RNA. Unlike in the neomycin (1a) complex, the furanose moiety III does not participate in direct contacts with the target. The hydrogen bonding network in the neomycin (1a) complex is identical to that previously observed in decoding– site complexes of paromomycin (1b).7, 11 Whereas there is a considerable range of pKa values for the amino groups in 1a and 2a,12 protonated states dominate at the buffer conditions used (pH6.2) which contribute an important electrostatic component to hydrogen- bonding interactions with the RNA.12 4-6

Comparison of the binding affinities of neomycin 1a with its cyclized derivative

2a for the decoding site revealed a 20–fold lower Kd value for the restricted derivative at pH 7.5,12 in line with a similarly decreased antibacterial potency under physiological conditions.33,34 The overall loss of direct hydrogen bonds between ring

III and the RNA might be another major reason for the lower target affinity of 2a.

2.8 Discussion

Currently, numerous examples exist in the literature of low molecular weight ligands which bind to and alter the function of therapeutically important RNA targets.35-39 Although much work has focused on ways to enhance the affinity of these interactions, comparatively little research has examined the issue of target selectivity.40 Ultimately, for these molecules to be clinically successful, one will need to understand how to modulate their binding to a desired target, to the exclusion of competing targets within a cell. At present, the aminoglycosides are rather promiscuous binders, owing to both their electrostatic binding mode and their ability to conformationally adapt to diverse RNA folds. In an attempt to circumvent the latter, we have synthesized conformationally restricted aminoglycoside analogues designed as specific prokaryotic ribosomal A– site binder.

As elaborated previously, the 2'–nitrogen and 5''–carbon of paromomycin are approximately 3.6 Å apart when bound to the A– site, and a hydrogen bond is likely to exist between the proton of the 2'–nitrogen and the 5'–oxygen.11 Thus, a newly formed 2'–N/5''–C covalent bond should constrain rings II and III in proximity and preorganize paromomycin into a conformation resembling the A– site bound form. In theory, this preorganization could be predicted to enhance the A–site binding affinity, due to a smaller change in the conformational entropy upon binding.41 In practice, however, the decreased conformational entropy can often be compensated by enthalpic penalties of the slight conformational discrepancy between the constrained

67 carbohydrates and the unconstrained active ligand conformation that allows target recognition.42-44

Our X–ray crystal structural studies on RNA decoding– site complexes of neomycin 1a and the restricted derivative 2a demonstrate that the intramolecular 2'–

5''cross– link introduced into the natural product is compatible with target binding.

Comparison of the crystal structures of 1a and 2a complexed to the decoding site reveals the exquisite sensitivity of aminoglycoside target recognition toward even slight modifications of the architecture of the natural product. Whereas key interactions of the aminoglycoside ligand are undisturbed by the modification, minimal differences in the overall conformation of 2a relative to 1a disable hydrogen– bond contacts to the target.

A–site fluorescent binding assay revealed that a 20–fold lower binding affinity for the constrained neomycin 2a compared to its parent compound 1a at pH 7.5. At a lower pH value 5.8 the Kd value of 2a decreased as a result of increasing protonation of the amino groups and thus domination of the ligand–RNA interaction by electrostatic forces. Similar binding activity trend were observed for paromomycin

1b and its constrained analogue 2b. Interestingly, the basicity of the secondary 2'– amino group, which covalently links rings II and III, is the lowest (pKa 6.4 in 2a, versus pKa 8.1 for the corresponding amine in 1a). Under physiological conditions, it is likely a combination of the decreased number of hydrogen–bond interactions between 2a and the decoding site, and diminished electrostatic contribution by the

68 only partially protonated 2'–amino group that might cause the decreased binding affinity and antibacterial potency of the restricted neomycin derivative.

Chapter 2 contains the material as it appears in Molecular Recognition of

RNA by Neomycin and a Restricted Neomycin Derivative. Zhao, F.; Zhao, Q.;

Blount, K.; Han, Q.; Tor. Y.; Hermann,T. Angew Chem, Int. Ed. 2005, 44, 2-6 and

Conformational Constraint as a Means for Understanding RNA-Aminoglycoside

Specificity. Blount, K.; Zhao, F.; Hermann, T.; Tor, Y. J. Am. Chem. Soc. 2005,

127(27), 9818-29. The dissertation author was the primary investigator and author of these papers.

References

(1) Vicens, Q.; Westhof, E. Biopolymers 2003, 70, 42-57.

(2) Tor, Y. Chembiochem 2003, 4, 998-1007.

(3) Vicens, Q.; Westhof, E. Chembiochem 2003, 4, 1018-1023.

(4) Tor, Y.; Hermann, T.; Westhof, E. Chemistry & Biology 1998, 5, R277-R283.

(5) Hermann, T.; Westhof, E. Journal of Molecular Biology 1998, 276, 903-912.

(6) Hermann, T.; Westhof, E. Biopolymers 1998, 48, 155-165.

(7) Fourmy, D.; Recht, M. I.; Blanchard, S. C.; Puglisi, J. D. Science 1996, 274, 1367-1371.

(8) Faber, C.; Sticht, H.; Schweimer, K.; Rosch, P. Journal of Biological Chemistry 2000, 275, 20660-20666.

(9) Jiang, L. C.; Majumdar, A.; Hu, W. D.; Jaishree, T. J.; Xu, W. K.; Patel, D. J. Structure 1999, 7, 817-827.

(10) Varani, L.; Spillantini, M. G.; Goedert, M.; Varani, G. Nucleic Acids Research 2000, 28, 710-719.

(11) Vicens, Q.; Westhof, E. Structure 2001, 9, 647-658.

(12) Blount, K. F.; Zhao, F.; Hermann, T.; Tor, Y. Journal of the American Chemical Society 2005, 127, 9818-9829.

(13) Hermann, T. Angewandte Chemie-International Edition 2000, 39, 1891-1905.

(14) Carter, A. P.; Clemons, W. M.; Brodersen, D. E.; Morgan-Warren, R. J.; Wimberly, B. T.; Ramakrishnan, V. Nature 2000, 407, 340-348.

(15) Michael, K.; Wang, H.; Tor, Y. Bioorganic & Medicinal Chemistry 1999, 7, 1361-1371.

(16) The hydrogens on the all Boc carbamate ( H-N-Boc) are well resolved in aceton/d6-DMSO/d6 (3:1). Full assignment shows that unlike other Boc protected carbamate and the 2'N dose not have the carbamate proton peak.

(17) Similar chemical shift changes have been predicted and observed fothe corresponding carbon in cyclohexylammonium (51.6ppm) vs N-methyl- cyclohexylammonium (59.4ppm).

(18) Sarneski, J. E.; Surprenant, H. L.; Molen, F. K.; Reilley, C. N. Analytical Chemistry 1975, 47, 2116-2124.

(19) The cross-peak between 5''-Ha and 2'-C was not observed. As with H-H coupling, the value of this 3-carbon coupling constant is dependent on the dihedral angle. The absence of cross peaks is not uncommon and can be understood as the distortion of compound 2a and 2b from a common confromation.

(20) Botto, R. E.; Coxon, B. Journal of the American Chemical Society 1983, 105, 1021-1028.

(21) Kaul, M.; Barbieri, C. M.; Kerrigan, J. E.; Pilch, D. S. Journal of Molecular Biology 2003, 326, 1373-1387.

(22) Kaul, M.; Barbieri, C. M.; Pilch, D. S. Journal of the American Chemical Society 2004, 126, 3447-3453.

(23) Shandrick, S.; Zhao, Q.; Han, Q.; Ayida, B. K.; Takahashi, M.; Winters, G. C.; Simonsen, K. B.; Vourloumis, D.; Hermann, T. Angewandte Chemie- International Edition 2004, 43, 3177-3182.

(24) The fluorescent binding affinity measurements was done by Dr. Kenneth F. Blount.

(25) Alper, P. B.; Hendrix, M.; Sears, P.; Wong, C. H. Journal of the American Chemical Society 1998, 120, 1965-1978.

(26) Griffey, R. H.; Hofstadler, S. A.; Sannes-Lowery, K. A.; Ecker, D. J.; Crooke, S. T. Proceedings of the National Academy of Sciences of the United States of America 1999, 96, 10129-10133.

(27) Fourmy, D.; Recht, M. I.; Puglisi, J. D. Journal of Molecular Biology 1998, 277, 347-362.

(28) Kaul, M.; Pilch, D. S. Biochemistry 2002, 41, 7695-7706.

(29) As the pH is decreased, the difference in affinity between neomycin and paromomycin decrease slightly. This suggests the possibility that the cause for convergence of the affinities of the constrained and nonconstrained aminoglycpsides could include smaller, additional term besides the pKa of the 2'-amine.

(30) Zhao, F.; Zhao, Q.; Blount, K. F.; Han, Q.; Tor, Y.; Hermann, T. Angewandte Chemie-International Edition 2005, 44, 5329-5334.

(31) Purohit, P.; Stern, S. Nature 1994, 370, 659-662.

(32) Miyaguchi, H.; Narita, H.; Sakamoto, K.; Yokoyama, S. Nucleic Acids Research 1996, 24, 3700-3706.

(33) Antibacterial potency of 1a and 2a was tested by the serial dilution method described in: S. Barluenga, K. B. Simonsen, E. S. Littlefield, B. K. Ayida, D. Vourloumis, G. C. Winters, M. Takahashi, S. Shandrick, Q. Zhao, Q. Han, T. Hermann, Bioorg. Med. Chem. Lett. 2004, 14, 713-718. The minimum inhibitory concentration (MIC) was determined as the lowest compound concentration that prevented bacterial growth after 18 h of incubation at 37°C. MIC values in g mL-1 for 1a/2a: 1/32 (Escherichia coli ATCC-25922 standard strain), 0.25/8 (Staphylococcus aureus ATCC-25923 standard strain), 0.5/16 (streptomycin-resistant S. aureus BAA-38), 64/>512 (gentamicin-resistant S. aureus BAA-40), 16/>512 (spectinomycin-resistant S. aureus BAA-44)

(34) Caution must be exercised in correlating antibacterial potency with in vitro target binding affinity and mode of interaction. Direct correlation with

molecular parameters is difficult, as the apparent antibacterial potency might be affected by numerous parameters, including compound permeability and off-target effects.

(35) Stage, T. K.; Hertel, K. J.; Uhlenbeck, O. C. Rna-a Publication of the Rna Society 1995, 1, 95-101.

(36) Wang, H.; Tor, Y. Journal of the American Chemical Society 1997, 119, 8734- 8735.

(37) Zapp, M. L.; Stern, S.; Green, M. R. Cell 1993, 74, 969-978.

(38) Mei, H. Y.; Mack, D. P.; Galan, A. A.; Halim, N. S.; Heldsinger, A.; Loo, J. A.; Moreland, D. W.; SannesLowery, K. A.; Sharmeen, L.; Truong, H. N.; Czarnik, A. W. Bioorganic & Medicinal Chemistry 1997, 5, 1173-1184.

(39) Vonahsen, U.; Davies, J.; Schroeder, R. Nature 1991, 353, 368-370.

(40) Luedtke, N. W.; Liu, Q.; Tor, Y. Biochemistry 2003, 42, 11391-11403.

(41) Novotny, J.; Bruccoleri, R. E.; Saul, F. A. Biochemistry 1989, 28, 4735-4749.

(42) Bundle, D. R.; Alibes, R.; Nilar, S.; Otter, A.; Warwas, M.; Zhang, P. Journal of the American Chemical Society 1998, 120, 5317-5318.

(43) Navarre, N.; Amiot, N.; van Oijen, A.; Imberty, A.; Poveda, A.; Jimenez- Barbero, J.; Cooper, A.; Nutley, M. A.; Boons, G. J. Chemistry-a European Journal 1999, 5, 2281-2294.

(44) Wacowich-Sgarbi, S. A.; Bundle, D. R. Journal of Organic Chemistry 1999, 64, 9080-9089.

Chapter 3

73 74

3.1 Introduction

Aminoglycoside antibiotics have been shown to bind a variety of therapeutically important RNAs, such as ribosomal RNA 30S, HIV viral RNA RRE,

TAR etc1-4. These natural antibiotics, however, are rather promiscuous RNA binders as discussed in previous chapters. Antibiotic resistance is a growing problem and contributes to increase the rates of clinical treatment failure and poor prognosis5, 6.

Based on our understanding of how small molecules interact with RNA we designed a novel type of aminoglycoside mimic–the cyclohexyl –peptides (Figure 3.1). The

peptides are composed of 4,5–substituted amino–cyclohexanecarboxylic acid

(trans–ACHC) subunits. We speculate that this cyclohexyl –peptide may still maintain RNA affinity but display higher RNA selectivity, as will be discussed later.

The potential availability of numerous building blocks, and established peptide chemistry will allow us to systematically synthesize, screen the resulting RNA binders and generate structure–activity relationships. It is also worth noting that contemporary literature suggests high stability of  and  peptide against proteases.7

Compared to aminoglycosides, the proposed –peptides possess similar yet improved structural features:

a) b)

R R O R R NH2 R 3 O NH 4 2 R R 2 O 5 1 NH 6 R COOH NH

H2N

R = NH2 or OH R = NH2 or OH

Figure 3.1 a) Structure of the cyclohexyl–peptide. b) Proposed 4,5–substituted amino– cyclohexanecarboxylic acid subunit.

a) Shape similarity: At least 15 residues of –amino acids are required for the

formation of a stable –helix in protic solvent. On the other hand, much shorter

–peptides can adopt ordered conformations due to their more rigidified

backbones. Studies of the synthetic trans–ACHC hexamers have shown that

the oligomers adopt a very stable helix in protic solvent.8-10 This result reflects

the limited degrees of freedom along the C–C bond imposed by the

cyclohexane ring. The helices are defined by fourteen–member-ring hydrogen

bonds that are formed between the backbone carbonyl oxygen and the backbone

NHs, two residues toward the N–terminus (14–helix, Figure 3.2). Our

preliminary molecular modeling suggests that the –(1,2) peptides display

considerable similarity to the neomycin class of aminoglycosides (Figure 3.3).

O N

O N H O H O N

Figure 3.2 Schematic view of helical trans–ACHC oligomer. The intramolecular hydrogen–bonds are highlighted in red.

b) H2N HO

NH2 HO O H2N HO O OH H2N O NH2 H2N NH HN O OH O NH2 O H N O OH 2

O HO HO NH2 O HO NH2 HO OH

Figure 3.3 a) Stereo image of a preferred solution conformation of a –(1,2) tripeptide module (dark balls and sticks) overlaid on the solution structure of paromomycin (colored sticks), showing the shape complementarity between the molecules. b) Structure of paromomycin (left) and the modeled tripeptide, roughly in the orientation depicted in the overlay.

77 b) Positive charges and hydrogen bond donors/acceptors: The proposed building

blocks (trans–ACHC) are substituted with amino and hydroxyl groups at the 4

and 5 positions, which are important for electrostatic interactions and hydrogen

bonding. Further more, one can, in principle, link additional charged residue or

RNA recognition elements to the terminal carboxyl groups. As in the natural

aminoglycosides, the cyclohexane–peptide core provides a scaffold that

projects the positively charged and hydrogen bond donor/accepter groups at the

periphery in spatially well–defined directions and distances. c) Rigidity: The connective amide bonds, compared with glycosidic bonds, are

less flexible. As we discussed in chapter I and II, flexibility is very likely to be

one of the major reasons for the promiscuousity of aminoglycosides. The

replacement of glycosidic bond with amide can hopefully lead to some RNA

binders with more rigidity and specificity.

3.2 Synthesis of the 4, 5 substituted amino–cyclohexanecarboxylic acids

The general synthetic pathway for the 4,5–substituted–trans–ACHC derivatives is shown below (Figure 3. 4). Starting from the commercially available cis–1,2,3,6– tetrahydrophtalic anhydride, subsequent cleavage of 4,5–epoxyl–2–amino– cyclohexanecarboxylic acid with different nucleophiles yields a range of –amino acid building blocks.

Nu HO NHR 3 NHR 3 NHR2 2 2 4 2 4 2 + O 5 1 5 1 OR OR 6 OR1 6 1 1 HO Nu O O O

O NHR2 O OR1 O O

Figure 3.4 The general synthetic pathway for the 4,5–substituted–trans–ACHC derivatives.

2R–amino–4R–amino–5R–hydroxyl–cyclohexancarboxylic acid series

(diaxial–diequatorial series): Cis di–methyl ester 12 was obtained in 96% yield by the hydrolysis of commercially available cis–1,2,3,6–tetrahydropthalic anhydride, followed by treatment with SOCl2 in MeOH (Figure 3.5). The symmetric diester 12 was subjected to stereoselective hydrolysis by porcine liver esterase (PLE) to give the optically pure mono acid 13 in a yield of 98%.11 The protected amino carboxylate 14 was synthesized via a Curtius rearrangement of 3, followed by treatment with benzyl alcohol in the present of triethyl amine.12 The exposure of the cis isomer 14 to

79 sodium methoxide in refluxing methanol led to the more stable trans isomer 15, in a ratio of 4:1 (15 vs. 14). Compound 5 was purified by flash chromatography giving a

50% yield over two steps (from 13 to 15). Treatment of 15 with m–CPBA generated two epoxides 16 and 17 in a ratio of 7:1. The major isomer 18 was carried to the next step, where conversion of the Cbz to Boc was accomplished in a one pot reaction to give compound 18 in 92% yield.13

O O O c NHCbz a OMe b OH O OMe OMe OMe O O O O 14 12 13

e d NHCbz NHCbz NHCbz O + O OMe OMe OMe O O O 15 16 17

N3 NHBoc OMe HO O 19 g h

j CbzHN H2N NHBoc NHBoc f NHBoc i 16 O OMe OMe OMe HO HO k O O O 20 21 18 NC NHBoc OMe HO O 22

H2NH2C NHBoc CbzHNH2C NHBoc m OMe OMe HO HO O O 23 24

Figure 3.5 Synthesis of 2R–amino–4R–amino–5R–hydroxyl–cyclohexancarboxylic acid series. Reagents and conditions: (a) (i) MeOH, (ii) SOCl2, MeOH; (b) porcine liver esterase, pH 8.0 ; (c) (i) ClCO2Et, (ii) NaN3, (iii) Δ, phCH2OH; (d) MeONa, MeOH; (e) m–CPBA; (f) 1 atm. H2 (g), 10% Pd/C, Boc2O, i–Pr2NEt; (g) NaN3/NH4Cl; (h) 1 atm. H2 (g), 10% Pd/C; (i) Sat. NH3/MeOH; (j) CbzCl, Et3N; (k) KCN/NH4Cl; (l) 1 atm. H2 (g), Raney Ni, Sat. NH3/MeOH; (m) CbzCl, Et3N.

An X–ray crystal structure of compound 18 was obtained and revealed that the epoxide is located on the same side as the carbomate (Figure 3.6a). This result was not surprising as it is well known that adjacent hydroxyl or amino groups have a directing effect on epoxidation through hydrogen bonding with m–CBPA.14 As most cyclohexyl oxiranes, compound 18 adopted a half chair conformation. It is worth noting that there is exclusively one conformer, of which both the ester and carbamate groups are in the pseudo–equatorial position; the diaxial conformer was not observed in the crystal structure. Compound 18 was treated with saturated methanolic ammonia to cleave the epoxide ring. The reaction had an excellent regioselectivity

(>19:1) and gave only one single product 20. To further purify and identify the structure of the ring opening product, 20 was protected with a Cbz group, resulting in compound 21. Using 2D–COSY, the newly introduced Cbz carbamate was located at the 4 position on the cyclohexane ring, which also indicated that the epoxide opening favored the 4 over the 5 position. This conclusion is further confirmed by X–ray analysis of compound 21 (Figure 3.6b). Both NaN3 and KCN opened the epoxide 18 in a similar fashion as NH3, with regioselectivy towards the 4 position, leading to compound 19 and 22, respectively.

Compound 19, with the nitrile group, is ready to be used for peptide coupling.

Alternatively, hydrogenation over 10% Pd/C and Cbz protection gave compound 21.

The cyano group of 22 was selectively reduced by H2 (g) using Raney Ni in saturated

NH3/MeOH. The resulting amine 23 was protected with Cbz group yielding compound 24. It is worth noting that all the amino acid building blocks 19, 21 and

24 produced from this synthetic route are diaxial–diequatorial substituted cyclohexane derivatives.

Figure 3.6 Crystal structure of compound 6 (a) and 11 (b). Unrelated Hydrogen atoms have been omitted for clarity.

The mechanisms of epoxide opening have been well studied.15, 16 Depending on the solvents, substrates, nucleophiles, catalysts and temperature, the reaction can proceed through mechanistic pathways via SN1 or SN2; the regiochemical course can be Markovnikov or ant–Markovnikov type of cleavage; the stereochemical outcome

83 may be complete inversion or complete retention of configuration. Since 1980’s,

Sharpless and other chemists have developed varies of catalysts for epoxide opening reactions. These catalysts are mostly metal containing compound, such as Ti(O–i–

17-19 Pr)4, Yb(OTf)3 and LiClO4. They facilitate the reaction and sometimes improve the regioselectivities mostly through chelating effect.

Under the basic conditions that we used, nucleophiles attack at one of the carbon atoms to induce an SN2 substitution, which causes inversion at that center, resulting in α–hydroxyl products. Two products are possible, depending on which carbon atom is attacked by the nucleophile. The transition state of SN2 reaction requires that the incoming nucleophile be in line with the attacked carbon and the oxirane oxygen. In the case of cyclohexyl oxiranes, which usually adopt half chair conformation, this transition state requirement leads to a product with trans–diaxial substitutes in the related conformation (named as Fürst–Plattner rule, also called trans diaxial effect, Figure 3.7).20 Even though the diequatorial product is more stable, the reaction is under kinetic control, not thermodynamic control. It is worth pointing out that the cyclohexyl oxirane may adopt two different half chair conformations and each conformer will give its own diaxial product. In the case of cyclohexyl oxiranes, the two products are enantiomeric; however, when the cyclohexyl oxirane is asymmetrically substituted, as in our case, the two products are different regioisomers.

O O OH

Nu Nu Nu O

O O HO Nu Nu Nu

O O OH Nu Nu Nu O

O O OH

Nu Nu Nu

Figure 3.7 Mechanism of trans diaxial effect for cyclohexyl oxirane opening.

There are two possible outcomes of our reaction (Figure 3.8). Note that one conformer leads to the diaxial–diequatorial products and the other one leads to the tetra–equatorial products. Both products are important and useful as building blocks for –peptide aminoglycoside mimics. The fact that all products 9, 11 and 14 produced from ring opening of 8 are diaxial–diequatorial suggested that there is one dominant conformer in the solution. From Figure 3.8, we can see that this energy favored conformer is the one with both the carbomate and carboxylate at pseudo– equatorial positions (on the left). Not surprisingly, this outcome is in agreement with the X–ray structure of compound 8.

NHBoc O O COOMe

BocHN NH3 NH3

COOMe

BocHN OH BocHN OH COOMe

MeOOC NHCbz NHCbz

NHBoc BocHN

COOMe OH HO

CbzHN COOMe NHCbz

Figure 3.8 Possible conformers of compound 8 and their corresponding ring opening reactions.

2R–amino–4S–hydroxyl–5S–amino–cyclohexancarboxilic acid series (tetra- equatorial series): 2–DOS, a cyclohexane derivative with multiple amino and hydroxyl groups, is a conserved structure core for aminoglycosides. Interestingly, all the substitutions on 2–DOS are positioned equatorial. In fact, six–member–ring structure with per–equatorial functional groups are one of the most common building blocks found in aminoglycosides (e.g. glucosamine in the tobramycin and neomycin family). To pursue the tetra–equatorial building blocks, which are essential for making the –peptide aminoglycoside mimic, we developed an alternative synthetic route. As we discussed, the reason we only got the single product from the trans epoxide 18 is the dominance of one conformer over the other (Figure 3.10). Inspired

86 by this, we decided to examine the ring opening of the cis substituted epoxide, of which the carbamate and the carboxilate substitutes would be at pseudo–equatorial and pseudo–axial position one each. By doing this, we are hoping to shift the equilibrium of the two possible conformations and see the opening products from both sides (Figure 3.9). To examine the effect of the protecting groups, we designed two compounds (28 and 33)21 with protecting groups of different sizes.

a) COOBut NHBoc

O O

NHCOOMe COOMe 28 33

b) COOMe NHBoc O ? O

BocHN NH3 COOMe NH3

COOMe BocHN HO BocHN OH COOMe

NHCbz NHCbz

COOMe BocHN CbzHN COOMe HO BocHN OH NHCbz

BocHN COOMe HO NHCbz

Figure 3.9 a) Compound 18 and 23. b) Two possible outcomes of epoxide opening reaction of compound 23

To synthesize 28, compound 13 was protected as the tert–butyl ester 25 by Boc anhydride in the presence of DMAP (Figure 3.10). Selective hydrolysis of the methyl ester in basic condition afforded compound 26. The Curtius rearrangement of

26 followed by treatment with methanol led to 27. Interestingly the reaction gave a much higher yield under acidic conditions (catalytic amount of p–TsOH) rather than the basic conditions used for compound 13. Epoxidation of 27 using m–CPBA provided 28 with high regioselectivity (>19:1).

O COOBut COOBut a b OH OMe OMe COOH O O 25 26

t t COOBu c COOBu d O

NHCOOMe NHCOOMe 27 28

Figure 3.10 Synthesis of compound 18. Reagents and conditions: (a) Boc2O, DMAP; (b) NaOH; (c) (i) ClCO2Et, (ii) NaN3, (iii) , MeOH, p–TsOH; (d) m–CPBA.

To synthesize 33, the trans–aminohexenylcarboxylate 14 was treated with m–

CPBA to provide compound 31 with excellent regioselectivity (>19:1). Again, the

Cbz protecting group was replaced by Boc to give 33 (Figure 3.11).

NHCbz a NHCbz b NHBoc O O OMe OMe OMe O O O 31 14 33

Figure 3.11 Synthesis of compound 23. Reagents and conditions: (a) m–CPBA (b) 1 atm. H2 (g), 10% Pd/C, Boc2O, i–Pr2NEt.

The epoxidation of both 27 and 31 have better selectivity compared to the trans substrate. This is very likely due to the cis structure, enabling the carbamate to be pseudo–axial and facilitated the hydrogen bonding with m–CPBA from a more convenient position (Figure 3.12).

Ar Ar

O R O O H O H N H O H COOMe O

N R COOMe

Figure 3.12 Comparison of the transition structures for m–CPBA epoxidation of the trans and cis amino–cyclohexencarboxylates.

The epoxide 33 was cleaved using different conditions such as NaN3, KCN and sat. NH3 in MeOH. Unexpectedly, all three reactions produced single products 34, 35 and 38, respectively (Figure 3.13). NMR characterization revealed that all of them were ring opening products from the conformer with pseudo–axial carbamate and pseudo– equatorial ester (Figure 3.10, on the right). The Cbz protection of 35 followed by treatment with sodium methoxide in methanol yields target 37, the tetra–

89 equatorial –amino acid building block. Reduction of 38 followed by the Cbz protection led to compound 40. However, the epimerization of 40 was not successful and only resulted in decomposition.

Compound 28 was cleaved in sat. NH3/MeOH to give only compound 29

(regioselectivity >19:1). It was proven that 29 came from the conformer that also has the carbamate at pseudo–axial position. After protection with Cbz (compound 30), the product was epimerized to the tetra equatorial final product 42.

a) HO NHBoc OMe N3 34 O

a b NHBoc HO HO O HO NHBoc d NHBoc e NHBoc OMe c OMe OMe OMe O H N CbzHN CbzHN 2 O O 33 f O 35 36 37

HO NHBoc g HO NHBoc HO NHBoc OMe h NC OMe OMe O H2NH2C CbzHNH2C 38 O O 39 40

HO NHBoc OMe CbzHNH2C O 41 b)

t t t COOBu a H2N COOBu b CbzHN COOBu O NHCOOMe HO NHCOOMe HO NHCOOMe

28 30 29

CbzHN COOBut

HO NHCOOMe

Figure 3.13 a) Synthesis of tetra–equatorial amino–cyclohexancarboxylic acids from compound 23. Reagents and conditions: (a) NaN3/NH4Cl; (b) 1 atm. H2 (g), 10% Pd/C; (c) Sat. NH3/MeOH; (d) CbzCl, Et3N; (e) NaOMe, MeOH; (f) KCN/NH4Cl; (g) 1 atm. H2 (h), Raney Ni, Sat. NH3/MeOH; (i) CbzCl, Et3N. b) Synthesis of tetra–equatorial amino–cyclohexancarboxylic acids from compound 18. Reagents and conditions: (a) Sat. NH3/MeOH; (b) CbzCl, Et3N; (c) NaOMe, MeOH.

3.3 Conformation Analysis for Cyclohexyl Oxiranes

To understand the regioselectivity observed, we performed conformational studies on epoxides 33 and related compounds. The cyclohexane ring is defined by two dihedral angles Tor1 and Tor2 (Figure 3.14). Using the AMBER conformational analysis program, 300 iterative rounds were performed.22 In each round, random variation of one or both dihedral angels is applied to the molecule, followed by energy minimization. Each local minimum conformer and the corresponding energy were recorded. The 300 rounds yielded totally fifteen minimum conformers (table

3.1).

NHBoc NHBoc 3 3 4 2 4 2 O O 5 1 5 1 6 6 COOMe COOMe

Tor : C - C -C -C 1 6 1 2 3 Tor2: C1-C2-C3-C4

Figure 3.14 Definition of torsion angle Tor1 (left) and Tor2 (right).

Analysis of the fifteen outcome local minima provides the information below:

1. Half chair is more stable than boat / semi boat conformation. All the local

minima obtained are half chair conformations except ○14 and ○15 , which are

represented as boat/semi boat shape. These two are the least stable (highest

energy) conformers among all fifteen conformations and only accounts for

0.06% of population.

Energy Tor1 Tor2 NHBoc COOMe % (Kcal/mol)

1 152.9 -61.4 47.0 a e 49.99 2 153.5 -61.6 47.8 a e 16.93 3 153.7 -61.2 46.9 a e 12.79 4 153.9 -61.6 47.0 a e 8.61 5 154.2 60.6 -53.1 e a 5.48 6 154.7 61.2 -53.6 e a 2.40 7 154.9 58.7 -44.1 e a 1.61 8 154.9 -61.7 47.9 a e 1.61 9 156.2 60.3 -54.4 e a 0.18 10 156.6 57.8 -41.9 e a 0.10 11 156.6 60.6 -54.7 e a 0.10 12 156.7 58.4 -41.9 e a 0.08 13 156.9 -61.2 47.5 a e 0.06 14 157.3 21.7 -63.2 n.a. n.a. 0.03 15 157.4 -15.6 -37.3 n.a. n.a. 0.03

2. Two types of half chair conformations. Analysis of the torsion angles suggests

that the half chair conformers can fall into two major clusters (Figure 3.15).

Conformers within each group have a very similar shape of cyclohexyl oxirane

frame and relatively flexible substitutes. As we predicted, one group of

molecules has pseudo–equatorial carbamate and pseudo–axial ester

(conformation A) and the other group (conformation B) is the opposite.

A-type 60 B-type boat/semi boat

0 Tor2 -90 -60 -30 0 30 60 90

-30

-60

-90 Tor1

Figure 3.15 Torsion angles of compound 23.

3. B type conformers are the dominant conformation. The conformation analysis

revealed that the first four most stable conformers are all B type. The most

stable A type conformer ○5 is 1.3 kcal/mol higher than ○1 , the most stable B

type conformer (Figure 3.16). Based on Van’t Hoff’s equation:

G = - R T ln K

We calculated the equilibrium constant K1/4 as 9.1. This means that the

concentration of conformer ○1 is almost 10 time as that of ○5 . The overall

distributions of each conformer are 9.95% and 89.99% for A and B type,

respectively. From this we can conclude that B is the dominant conformation,

which agrees well with our experimental data.

Figure 3.16 Modeling structure of compound 23. The most stable B–type conformation ① (left) and the most stable A–type conformation ⑤ (right).

4. Steric hindrance is the major cause of the preference for the B conformation. To

understand the higher stability of the B conformation, we did a series of

conformational analysis on related compounds I, II and III (Figure 3.17). Not

surprisingly both mono substituted compounds I and II prefer having the

substitution at a pseudo–equatorial position. This indicated that putting the

substitutes (NHBoc or COOMe) at the pseudo–axial position will cause steric

hindrance which is energetically unfavorable. Since compound 33 is cis di–

substituted, the molecule has to keep one substitution at the axial position.

Experimentally and computationally, compound 33 chose the axial group as the

carbamate. This preference may arise from two reasons: a) Axial NHBoc causes

less steric hindrance compared to axial COOMe. b) There is some extra driving

force such as intramolecular hydrogen bonding that keeps NHBoc at axial

position.

NHBoc O O

O O O O

COOMe COOMe

I II III

Figure 3.17 Compound I, II and III

It has been reported that Li+ affects cyclohexyl oxirane conformation by

chelating the epoxide oxygen atom and substituted ether group (Figure 3.17).19

Since some minimized conformations of 33 suggest the proximity between NH

group and the oxirane oxygen (e.g. distance from N to O is ~ 3.03 Å for ○1 ),

it is possible that there is an intramolecular hydrogen bond holding the

carbomate at the pseudo–axial position. To test this hypothesis, we designed

compound III which is a carbonate version of 33. If the intramolecular

hydrogen bond does exist and plays an important role in maintaining the B

conformation in 33, compound III which dose not have the H atom and is not

able to form the H–bond, would show less or even no preference for type B

conformation. However, the conformational analysis of III is very similar to

that of 33, with even more preference towards B conformation.23 The

energetically difference between most stable B and A conformers are 2.56

kcal/mol (K=73.27). This result clearly excludes the possibility of

intramolecular hydrogen bonding effect.

a) R Li+ b) Boc H O N

O O

COOMe

Figure 3.18 a) Chelatation effect of Li+ on cyclohexyl oxirane conformation. b) Possible intramolecular hydrogen bonding effect on cyclohexyl oxirane conformation.

Converting an equatorial substitute to the axial position causes elevation of

conformational energy of the molecule. This energy disfavor is caused by steric

hindrance, mostly by atoms crowdedly sitting right above the same side of the

six member ring. These atoms repel each other and try to keep as much distance

as possible from each other. They include the oxygen atom of the epoxide, axial

hydrogens on the same side and the axial substituent. For conformation A and B,

these atoms are almost the same except the axial substituents. It is worth noting

that the closer the atoms are to the cyclohexane, the more severe steric clash they

cause. Therefore, the axial substitution atoms that are closely attached to the

cyclohexane are the most important factors to compare (Figure 3.18). For A

type conformation, these atoms are O–C=O and for B type they are H–N–C. We

can see from the structure, the size of O–C=O is overall bigger than H–N–C.

More importantly, H–N–C is more flexible and L shaped which allows the H and

C (attached with the tert–butyl group) substituent to face away from the

cyclohexane as seen in most of the minimized conformations above (Figure

3.16). Conversely, O–C=O is planer and rigid, making it difficult to avoid the

space clash. For instance, in conformation ○5 (Figure 3.16), the most stable A

conformer of 33, the ―O=‖ is forced to lean towards the cyclohexane so the ―O–

CH3‖ can be projected away from the whole molecule.

tBu O Me

C O O O H N C

O H H O H H

COOMe BocHN

Figure 3.19 Steric hindrance of conformation A (left) and B (right).

References

(1) Moazed, D.; Noller, H. F. Nature 1987, 327, 389-394.

(2) Moazed, D.; Stern, S.; Noller, H. F. J Mol Biol 1986, 187, 399-416.

(3) Zapp, M. L.; Stern, S.; Green, M. R. Cell 1993, 74, 969-978.

(4) Mei, H. Y.; Galan, A. A.; Halim, N. S.; Mack, D. P.; Moreland, D. W.; Sanders, K. B.; Truong, H. N.; Czarnik, A. W. Bioorganic & Medicinal Chemistry Letters 1995, 5, 2755-2760.

(5) Davies, J. Nature 1996, 383, 21 9-220.

(6) Levy, S. B. Scientific American 1998, 278, 46-53.

(7) Frackenpohl, J.; Arvidsson, P. I.; Schreiber, J. V.; Seebach, D. Chembiochem 2001, 2, 445-455.

(8) Appella, D. H.; Barchi, J. J.; Durell, S. R.; Gellman, S. H. Journal of the American Chemical Society 1999, 121, 2309-2310.

(9) Appella, D. H.; Christianson, L. A.; Klein, D. A.; Richards, M. R.; Powell, D. R.; Gellman, S. H. Journal of the American Chemical Society 1999, 121, 7574-7581.

(10) Appella, D. H.; Christianson, L. A.; Karle, I. L.; Powell, D. R.; Gellman, S. H. Journal of the American Chemical Society 1999, 121, 6206-6212.

(11) Kobayashi, S.; Kamiyama, K.; Iimori, T.; Ohno, M. Tetrahedron Letters 1984, 25, 2557-2560.

(12) The removal of excess benzyl alcohol from 4 was extremely difficult, and therefore benzyl alcohol was carried to the next step. The yield was calculated over two steps (from compound 3 to 5).

(13) This protecting group change is for the compatibility of the future chemistry. Using tert-butanol after the Curtius rearrangement gave very low yield Boc- protected product.

(14) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chemical Reviews 1993, 93, 1307- 1370.

(15) Parker, R. E.; Isaacs, N. S. Chemical Reviews 1959, 59, 737-799.

(16) Buchanan, J. G.; Sable, H. I. In selective Organic Transformations; Willey- Intescience: New york, 1972.

(17) Caron, M.; Sharpless, K. B. Journal of Organic Chemistry 1985, 50, 1557- 1560.

(18) Chini, M.; Crotti, P.; Favero, L.; Macchia, F.; Pineschi, M. Tetrahedron Letters 1994, 35, 433-436.

(19) Crotti, P.; Favero, L.; Gardelli, C.; Macchia, F.; Pineschi, M. Journal of Organic Chemistry 1995, 60, 2514-2525.

(20) Furst, A.; Plattner, P. A. Helvetica Chimica Acta 1949, 32, 275-275.

(21) Because of the synthesis, compound 8 (without ester and carbamate protecting groups) is the enantiomers of compound 23, which may provides the accessibility of synthesis the group of optically pure building blocks. Please note that this will not affect the regioselectivities of the epoxide opening reactions.

(22) Program Hyperchem 7.5

(23) This result is likely because that cpompound III does not have the carbamate hydrogen as compound 23, which resulte in less steric hindrance.

Chapter 4

Experimental Section

100 101

4.1 Synthetic Experimental Section

4.1.1 Materials

Commercial reagents and solvents were used as received.

4.1.2 NMR Instrumentation

NMR Spectroscopy. 1H, 13C and 2D NMR spectra were recorded on a Varian

Mercury 400 MHz spectrometer. 15N NMR spectra were recorded at 50.7 MHz at

23˚C on a Varian Unity 500 spectrometer using a 5 mm inverse broad band probe

15 with an acquisition time of 1 s and a recycle delay of 0.5 s. NH4Cl (1M, 10% enriched) in 85:15 (v:v) H2O/D2O contained in a coaxial 2–mm inner cell (from New

15 Era) was used as an external reference ( = 24.00). The NH4Cl standard was purchased from Cambridge Isotope Laboratories, Inc. NMR solutions were prepared by dissolving aminoglycosides (TFA salt form, 340 mg each) in 0.5 ml 85:15 (v:v)

H2O/D2O to yield the final concentration of 0.5M–0.6M. The solution pH was adjusted by 31% HCl and 40% KOH in 85:15 (v:v) H2O/D2O. All pH measurements were acquired with a Corning 320 pH meter interfaced with a micro stem glass/calomel combination electrode (from Aldrich).

4.1.3 Synthesis of conformationally-constrained Neomycin B (2a)

Compound 3 (Boc–neomycin B) and Compound 4 (Tips–Boc–neomycin B) were synthesized as published procedures.1

102

Compound 6 (Constrained Boc–neomycin B). Compound 4 (4.0 g, 2.7 mmol) was treated with 50% TFA/CHCl3 (100 ml) for 30 minutes at room temperature. The volatiles were removed under reduced pressure. Water (20 ml) was added and the solution was freeze-dried. The white solid obtained was dissolved in DMF (500 ml) and treated with freshly distilled triethylamine (4.6 ml, 32.4 mmol). The solution was stirred at room temperature for 10 days and then concentrated. The residue was partitioned between water (400 ml) and ethyl acetate (400 ml). The aqueous layer was extracted with ethyl acetate (2 × 200 ml). The aqueous layer was concentrated to yield a white solid.2 The crude reaction mixture was dissolved in methanol (200 ml).

Triethylamine (4.6 ml, 32.4 mmol) and Boc2O (7.67 g, 32.4 mmol) were added and the solution was stirred for 16 hours at room temperature. After concentration, the residue was partitioned between dilute HCl (pH = 3, 500 ml) and ethyl acetate (750 ml). The aqueous layer was extracted with ethyl acetate (2 × 250 ml). The combined organic layer was washed with 1M NaHCO3, brine and dried over MgSO4. The concentrated crude was purified by gravity column chromatography over silica gel

(3% methanol/dichloromethane, with 0.3% concentrated ammonia) to give a white solid (340 mg, 12%). 1H NMR (acetone-d6: DMSO-d6 3:1, 400 MHz): δ 6.61 (t, J =

5.6, HN6’’’), 6.30 (m, HN1), 6.13 (m, HN6’), 6.11 (m, HN3), 6.01 (d, J = 9.6,

HN2’’’), 5.75 (s, H1’’), 5.34 (d, J = 3.2, HO3’’’), 5.24 (d, J = 4.4, HO4’’’), 5.08 (d, J

= 2.4, H1’), 4.88 (s, H1’’’), 4.81 (d, J = 5.2, HO6), 4.68 (s, HO4’), 4.30 (m, H4’’),

4.28 (m, HO2’’), 4.27 (m, H2’’), 4.09 (d, J = 5.2, H3’’), 3.86 (s, 2H, H5’’’ and H3’’’),

3.75 (m, 2H, H2’’’ and H5’), 3.59 (t, J = 9.2, H5), 3.54 (m, H3), 3.48 (m, H4’’’), 3.44

103

(m, 2H, H6’ab), 3.43 (m, H6), 3.38 (m, H4), 3.35 (m, H1), 3.34 (m, H3’), 3.27 (m,

H6’’’), 3.22 (m, H4’), 2.73 (m, H5’’a), 2.55 (m,H5’’b), 2.26 (m, H2’), 2.05 (m, H2e),

1.36 (m, 45H, Boc). 13C NMR (acetone-d6 : DMSO-d6 3:1, 100 MHz, assignment based on gHSQC): δ 110.3 (C1’’), 101.4 (C1’’’), 98.0 (C1’), 83.4 (C3’’), 83.1 (C4’’),

81.4 (C5), 81.1 (C4), 78.5 (C2’’), 76.3 (C6), 72.8 (C5’’’ or C3’’’), 71.6 ( C3’), 71.5

(C4’), 70.7 (C5’), 70.5 (C3’’’or C5’’’), 67.6 (C4’’’), 62.9 (C2’), 52.6 (C2’’’), 52.1

(C1), 50.0 (C3), 49.0 (C5’’), 41.8 (C6’), 40.7 (C6’’’), 35.4 (C2). MS (ESI-positive):

+ m/z = 1097.28 [M+H] , calcd for C48H84N6O22 1096.56.

104

1 Figure 4.1 H NMR spectrum of Boc5–constrained neomycin B (6).

105

Figure 4.2 gCOSY NMR spectrum of Boc5–constrained neomycin B (6).

106

Figure 4.3 gHSQC spectrum of Boc5–constrained neomycin B(6).

107

Figure 4.4 gHMBC NMR spectrum of Boc5–constrained neomycin B (6).

Compound 2a (constrained neomycin B). The Boc protected derivative 6 (340 mg, 0.31 mmol) was treated with 50% TFA/CHCl3 (5 ml) for 30 minutes at room temperature. The volatiles were removed in vacuo. Reverse phase C-18 flash column chromatography (0.1% TFA/water) afforded the desired product 2a as a white

1 solid (389 mg, 97%). H NMR (D2O, 400M Hz): δ 5.77 (d, J = 3.2, H1’’), 5.73 (d, J

= 3.6, H1’), 5.36 (d, J = 1.6, H1’’’), 4.73 (m, H2’’), 4.67 (dd, J1 = 12.8, J2 = 3.2,

H4’’), 4.62 (d, J = 5.6, H3’’), 4.33 (t, J = 4.8, H5’’’), 4.25 (t, J = 3.2, H3’’’), 4.07 (m,

H3’), 4.04 (d, J = 5.6, H4), 4.01 (m, H5’), 3.95 (m, H5), 3.87 (m, H6), 3.84 (m,

H4’’’), 3.65 (m, H5’’a), 3.64 (m, H2’’’), 3.62 (m, H3), 3.53 (m, H4’), 3.47 (m, H6’

108 a), , 3.43 (m, H2’), 3.42 (m, 3H, H6’’’ a, H6’’’b and H1), 3.35 (m, H6’ b), 3.21 (t, J =

13.2, H5’’ b), 2.57 (dt, J1 = 12.4, J2 = 4.4, H2e), 1.93 (dd, J1 = J2 = 12.4, H2a). 13C

NMR (D2O, 100 MHz): δ 163.0 (q, J = 160.0, C1 of TFA ), δ 116.5 (q, J = 1161.6,

C2 of TFA ), δ 110.1 (C1’’), 96.7 (C1’’’), 95.7 (C1’), 80.1 (C4’’), 79.4 (C5), 79.3

(C3’’), 77.7 (C4), 77.1 (C2’’), 73.1 (C6), 70.8 (C4’), 70.3 ( C5’’’), 69.8 (C5’), 67.6

(C3’’’), 67.3 (C4’’’), 66.8 (C3’), 60.2 (C2’), 50.8 (C2’’’), 49.7 (C1), 48.4 (C3), 47.5

(C5’’), 40.4 (C6’’’), 40.0 (C6’), 28.3 (C2). HRMS (ESI-positive): m/z = 597.3097

+ [M+H] , calcd for C23H44N6O12.

Figure 4.5 1H NMR spectrum of constrained neomycin B (2a).

109

Figure 4.6 gCOSY NMR spectrum of constrained neomycin B (2a).

110

Figure 4.7 gHSQC NMR spectrum of constrained neomycin B (2a).

111

Figure 4.8 gHMBC NMR spectrum of constrained neomycin B (2a)

112

4.1.4 Synthesis of conformationally-constrained Neomycin B (2b)

Compound 7 (Boc–paromomycin). Paromomycin sulfate (1.0 g, 1.2 mmol) was dissolved in a mixture of dioxane/water (2:1, 15 ml). Triethylamine (1.7 ml, 12 mmol) was added. The solution was sonicated and Boc2O (2.6 g, 12 mmol) was added. The reaction was stirred for 4 hours, and then concentrated. The residue was partitioned between water (300 ml) and ethyl acetate (400 ml). The aqueous was extracted with ethyl acetate (2 × 150 ml). The combined organic layer was washed with brine and dried over MgSO4. The concentrated crude was purified by flash column chromatography over silica gel (2.5% methanol / dichloromethane) to give 7 as a white solid (1.15 g, 85%). 1H NMR (methanol-d4, 400 MHz): δ 5.32 (s, 1H), 5.16 (s,

1H), 4.92 (s, 1H, masked by HOD peak), 4), 4.18 (s, 2H), 3.96 (s, 1H), 3.92–3.78 (m,

5H), 3.78–3.58 (m, 6H), 3.57–3.47 (m, 4H), 3.46 (s, 1H), 3.42–3.22 (m, 3H), 1.93 (m,

1H), 1.66–1.21 (m, 46H). MS (ESI-positive): m/z = 1138.35 [M+Na]+, calcd for

C48H85N5O24 1115.56.

Compound 8. To a solution of 7 (1.0 g, 0.9 mmol) in DMF (10 ml), p-toluene sulfonic acid (20 mg, 0.12 mmol) was added followed by 2,2-dimethoxypropane (3 ml, 16 mmol). The reaction was stirred for 3 hours, and then neutralized with 1M

NaHCO3. The resulting solution was partitioned between ethyl acetate (400 ml) and water (300 ml). The aqueous layer was extracted with ethyl acetate (2 × 150 ml). The combined organic layer was washed with brine and dried over MgSO4. Flash column chromatography (2% methanol/dichloromethane) afforded the desired product 8 as a

113 white solid (820 mg, 79%). 1H NMR (methanol-d4, 400 MHz): δ 5.46 (s, 1H), 5.12 (s,

1H), 4.88 (s, 1H), 4.20 (m, 1H), 4.15 (m, 1H), 3.96 (m, 1H), 3.92-3.83 (m, 3H), 3.83-

3.57 (m, 7H), 3.57-3.41 (m, 5H), 3.41-3.19 (m, 4H), 1.94 (m, 1H), 1.66-1.21 (m,

+ 52H). MS (ESI-positive): m/z = 1097.28 [M+H] , calcd for C48H84N6O22 1096.56.

Compound 9. 2,4,6-triisopropylbenzenesulfonyl chloride (6.7 g, 22.1 mmol, 32 equiv) was added to the solution of 8 (0.8 g, 0.7 mmol) in pyridine (12 ml). The reaction was stirred for 36 hours and then concentrated. The residue was taken up in ethyl acetate (300 ml), washed with water (4 × 250 ml), 1M NaHCO3 (25 ml) and brine. The organic layer was dried over MgSO4, and concentrated. The residue was purified by flash chromatography (2% Methanol/CH2Cl2) to yield 5 as a white solid

(0.70 g, 71%). 1H NMR (methanol-d4, 400 MHz): δ 7.28 (s, 2H), 5.54 (s, 1H), 5.13

(s, 1H), 4.87 (s, 1H, masked by HOD peak), 4.39–4.01 (m, 7H), 4.01-3.83 (s, 2H),

3.83–3.10 (m, 14H), 2.95 (m, 2H), 1.94 (m, 1H), 1.61–1.18 (m, 70H). MS (ESI-

+ positive): m/z = 1444.39 [M+Na] , calcd for C66H111N5O26S 1421.72.

Compound 11. The fully protected and 5’’-activated derivative 9 (500 mg, 0.351 mmol) was treated with 50% TFA/ CHCl3.for 30 minutes at room temperature. The volatiles were removed in vacuo. Water (5 ml) was added to the desired product and the solution was freeze-dried. The white solid obtained (0.5 g, 0.34 mmol) was dissolved in DMF (300 ml) and was treated with freshly distillated triethylamine (0.5 ml, 3.5 mmol). The reaction was stirred at room temperature for 10 days and then concentrated. The residue was partitioned in water (200 ml) and ethyl acetate (200

114 ml). The aqueous layer was extracted with ethyl acetate (2 × 150 ml). The aqueous layer was concentrated to yield a white solid. To assist purification, the crude compound was Boc protected and separated as described below. The crude reaction mixture was dissolved in methanol (25 ml). Triethylamine (0.5 ml, 3.5 mmol) was added followed by Boc2O (0.77 g, 3.5 mmol). The reaction was stirred for 16 hours, and then concentrated. The residue was partitioned between dilute HCl (pH = 3, 300 ml) and ethyl acetate (400 ml). The aqueous was extracted with ethyl acetate (2 × 150 ml). The combined organic layer was washed with 1M NaHCO3, brine and dried over MgSO4. The concentrated crude was purified by gravity chromatography over silica gel (4.5% methanol/dichloromethane, with 0.45% concentrated ammonia) to give 6 as a white solid (70 mg, 20%). 1H NMR (methanol-d4, 400 MHz): δ 5.76 (s,

1H), 5.15 (d, J = 3.2, 1H), 4.90 (s, 1H), 4.56 (s, 1H), 4.43–4.33 (m, 2H), 4.11 (d, J =

6.4, 1H), 3.92– 3.86 (m, 2H), 3.88 (s, 1H), 3.70 (s, 1H), 3.60 (t, J = 8,1H), 3.57–3.29

(m, 9H), 3.21 (dd, J1 = J2 = 7.2), 2.78 (m, 1H), 2.63 (t, J = 11.6, 1H), 2.44 (m, 1H),

2.18 (m, 1H), 1.58-1.22 (m, 32H). MS (ESI-positive): m/z = 998.25 [M+H]+, calcd for

C43H75N5O21 997.50.

Compound 2b. The Boc protected derivative 11 (30 mg, 0.03 mmol) was treated with 50% TFA/CHCl3.for 30 minutes at room temperature. The volatiles were removed in vacuo. Reverse phase C18 flash column chromatography (0.1%

1 TFA/water) afforded 2b as a white solid (35 mg, 99%). H NMR (D2O, 400 MHz): δ

5.70 (d, J = 2.0, H1’’), 5.54 (d, J = 4.0, H1’), 5.30 (d, J = 1.6, H1’’’), 4.65 (dd, J1 =

5.6, J2 = 2.0, H2’’), 4.60 (m, H3’’), 4.56 (m, H4’’), 4.30 (t, J = 4.8, H5’’’), 4.20 (t, J

115

= 3.2, 3’’’), 3.95 (m, H4), 3.93 (m, H6’a), 3.92 (m, H3’), 3.91 (m, H5), 3.82 (m,

H4’’’), 3.79 (m, H5’), 3.75 (m, H6), 3.70 (m, H6’b), 3.65 (m, H3), 3.59 (m, H2’’’),

3.57 (m, H5’’a), 3.50 (m, H2’), 3.46 (m, H4’), 3.40 (m, 2H, H6’’’a and H6’’’b), 3.38

(m, H1), 3.23 (t, J = 12.4, H5’’b), 2.51 (dt, J1 = 12.8, J2 = 4.4, H2e), 1.84 (dd, J1 =

13 J2 = 12.8, H2a). C NMR (D2O, 100 MHz): δ 163.0 (q, J = 160.0, C1 of TFA ), δ

116.5 (q, J = 1161.6, C2 of TFA ), δ 110.3 (C1’’), 96.6 (C1’’’), 96.3 (C1’), 81.1 (C4),

80.0 (C3’’), 79.9 (C4’’), 79.8 (C5), 76.4 (C2’’), 74.1 (C5’), 73.1 (C6), 70.2 (C5’’’),

69.4 (C4’), 68.3 (C3’), 67.6 (C3’’’), 67.4 (C4’’’), 60.4 (C6’), 59.8 (C2’), 50.8 (C2’’’),

49.2 (C1), 49.0 (C3), 47.7 (C5’’), 40.5 (C6’’’), 28.4 (C2). HRMS (ESI-positive): m/z

+ = 597.3097 [M+H] , calcd for C23H43N5O13 597.29.

Figure 4.9 1H NMR spectrum of constrained paromomycin (2b)

116

Figure 4.10 gCOSY NMR spectrum of constrained paromomycin (2b)

117

Figure 4.11 gHSQC NMR spectrum of constrained paromomycin (2b)

118

Figure 4.12 gHMBC NMR spectrum of constrained paromomycin (2b)

4.1.5 Synthesis of β-aminocyclohexancarboxylic acids.

Compound 11-15 were synthesized as published procedure. 3

Compound 16. Compound 15 (15.000 g, 51.6 mmol) was dissolved in CH2Cl2

(200 ml) and phosphate buffer (0.2M, 200 ml) was added to the solution in a 500ml round bottom flask. m–CPBA (17.4 g, 101.4 mmol, purified) was added. After stirred for 16hour at room temperature, the solution was diluted with dichloromethane (200 ml). The organic phase was separated and washed with saturated NaHCO3 (100 ml

119

x2), brine and dried over MgSO4, filtered and evaporated. The crude material was chromatographed with EtOAc and hexane (1:3) to afford the final compounds as white solid 16 (11.22 g, 36.6 mmol, 72%) and 17 (1.50 g, 4.92 mmol, 10%). 1H

NMR (CDCl3, 400MHz) δ 7.31-7.19 (m, 5H), 5.57 (d, J = 9.2, 1H), 5.00 (s, 2H),

3.99- 3.92 (m, 1H), 3.56 (s, 3H), 3.17-3.14 (m, 1H), 3.08 (app t, 1H), 2.62-2.57 (m,

13 1H), 2.42- 2.11 (m, 3H), 1.86-1.80 (m, 1H); C NMR (CDCl3, 100 MHz) δ 173.47,

155.44, 136.51, 128.42 (2C), 128.01 (3C), 66.50, 51.95, 51.36, 51.20, 46.63, 41.20,

+ 28.96, 24.70. MS (ESI-positive) calcd for C16H19NNaO5 [M+Na] 328.12, found

328.10.

1 Compound 17: H NMR (CDCl3, 400MHz) δ 7.35-7.29 (m, 5H), 5.05 (s, 2H),

4.91 (d, J = 6.8, 1H), 3.90 (J1 = 18, J2 = 8.8, J3 = 4.8, 1H), 3.58 (s, 3H), 3.22-3.19

(m, 1H), 3.13 (t, J = 4.4, 1H), 2.57-2.46 (m, 2H), 2.39-2.33 (m, 1H), 2.20 (ddd, J1 =

13 15.2, J2 = 6.4, J3 = 4.4, 1H), 1.87-1.81 (m,1H); C NMR (CDCl3, 100MHz) δ

173.59, 155.60,136.70, 128.80 (2H), 128.42 (2H), 128.38, 66.948, 52.68, 52.23,

50.24, 40.12, 43.52, 30.76, 20.08. MS (ESI-positive) m/z calcd for C16H19NNaO5

+ [M+Na] 328.12, found 328.17.

Compound 18: To a solution of 16 (9.0 g, 29.4 mmol) in anhydrous DMF (360 ml) were added diisopropylethylamine (16.2 ml, 90.6 mmol), Boc2O (19.2 g, 90.6 mmol) and 10% Pd/C (1.550 mg, 5% mol). The solution was kept under 1atm H2 at room temperature for 4hr, and then evaporated under vacuo. The crude mixture was

120 purified by flash chromatography with EtOAc and hexane (1:3) to afford white solid

1 18 (7.38 g, 27.0 mmol, 92%). H NMR (CDCl3, 400 MHz) δ 5.11 (d, J = 8, 1H),

4.01-3.91 (m, 1H), 3.64 (s, 3H), 3.23 (app t, J = 3.2, 1H), 3.14 (app t, J = 3.2, 1H),

2.62 (app ddd, J1 = 12.4, J2 = 6.0, 1H), 2.33-2.25 (m, 1H), 2.21-2.12 (m, 2H), 1.87

13 (app dd, J1 = 15.2, J2 = 5.5, 1H), 1.370 (s, 9H); C NMR (CDCl3, 100MHz) δ

173.83, 155.14, 79.64, 52.25, 51.84, 51.40, 46.21, 41.46, 28.86, 28.56 (3H) 24.20.

+ MS (ESI-positive) m/z calcd for C13H21NNaO5 [M+Na] 294.13, found 293.99.

Compound 19: To a solution of NH4Cl (385 mg, 7.2 mmol) and NaN3 (1.170 g,

18.0 mmol) in H2O (10 ml) were added a solution of 18 (1.00 g, 3.6 mmol) in MeOH

(40 ml). The solution was kept for 72hrs at room temperature. After neutralized by aqueous HCl (pH = 2) to pH 7, the solvent was removed under vacuo and partitioned between EtOAc (400 ml) and water (400 ml). The aqueous phase was extracted with

EtOAc (250 ml x 2). The combined organic phase was washed with brine and dried over MgSO4, filtered and evaporated. The crude material was purified by flash chromatography with EtOAc and hexane (1: 3) to afford white solid 19 (750 mg, 2.4

1 mmol, 67%). H NMR (CDCl3, 400MHz) δ4.91 (br, 1H), 4.02 (br, 1H), 3.82-3.78

(m, 1H), 3.67 (s, 3H), 2.85-2.79 (m, 2H), 2.17-2.12 (m, 1H), 2.00-1.93 (m, 1H), 1.90-

13 1.77 (m, 2H), 1.41 (s, 9H); C NMR (CDCl3, 100MHz) δ173.78, 155.14, 79.90,

67.64, 61.69, 52.23, 46.71, 43.60, 31.25, 30.18, 28.43 (3C). HRMS (EI) m/z calcd for

C13H22N4O5 314.1585, found 314.1585.

121

Compound 21: NH3 gas was bubbled through a solution of 18 (1.000 g, 3.6 mmol) in MeOH (20 ml) till saturation in a 30 ml pressure tube at 0ºC. The solution was caped and kept at 10ºC for 10 days. The solution was evaporated to dryness carefully under vacuo and diluted with MeOH and water (5:1, 50ml). CbzCl (1.6 ml, 11.3 mmol) and triethylamine (1.6 ml, 11.2 mmol) were added to the solution at 0ºC. The solution was warmed up to room temperature and kept for 16hrs, and then evaporated.

The mixture was partitioned between EtOAc (400 ml) and aqueous HCl (pH = 2, 500 ml). The aqueous phase was extracted with EtOAc (250 ml x 2). The combined organic phase was washed with brine, dried over MgSO4, filtered and evaporated.

The crude product was purified by flash chromatography with EtOAc and hexane (1:

1 3) to yield white solid 21 (1.31 mg, 3.1 mmol, 85% over two steps). H NMR (CDCl3,

400MHz) δ7.37-7.30 (m, 5H), 5.30 (br, 1H), 5.074 (m, 2H), 4.89 (d, J = 5.6), 4.00

(br, 1H), 3.79 (br, 1H), 3.73 (br, 1H), 3.66 (s, 3H), 2.82 (ddd, J1 = J2 = 8, J3 = 4.4,

1H ), 2.2 (br, 1H), 2.16-2.10 (m, 1H), 1.95 (ddd, J1 = 14, J2 = 8, J3 = 4.4, 1H), 1.85,

13 1.73 (m, 2H), 1.42 (s, 9H); C NMR (CDCl3, 100MHz) δ174.07, 156.95, 155.40,

136.48, 128.89 (2H), 128.56 (2H), 128.48, 80.16, 68.59, 67.39, 52.40, 52.35, 47.17,

43.87, 32.44, 30.47, 28.68 (3C). HRMS (EI) m/z calcd for C21H30N2O7 422.2048, found 422.2053.

Compound 22: To a solution of NH4Cl (385 mg, 7.2 mmol) and KCN (1.170 g,

18.0 mmol) in H2O (10 ml) were added a solution of 18 (1.000 g, 3.6 mmol) in

MeOH (40 ml). The solution was kept for 72hrs at room temperature. After

122 neutralized by aqueous HCl (pH = 2) to pH 7, the solvent was removed under vacuo and partitioned between EtOAc (400 ml) and water (400 ml). The aqueous phase was extracted with EtOAc (250 ml x 2). The combined organic phase was washed with brine, dried over MgSO4, filtered and evaporated. The crude material was purified by flash chromatography with EtOAc and hexane (1: 3) to afford white solid 22 (650 mg,

1 2.1 mmol, 58%). H NMR (CDCl3, 400MHz) δ 4.21 (d, 1H, J = 8.1), 4.16-4.11 (m,

2H), 3.96 (br, 1H), 3.66 (s, 3H), 2.98-2.94 (m, 1H), 2.85-2.79 (m, 1H), 2.17-1.90 (m,

13 4H), 1.38 (s, 9H); C NMR (CDCl3, 100 MHz) δ173.77, 155.30, 120.06, 79.99,

65.08, 52.32, 50.54, 47.16, 43.30, 32.58, 31.49, 28.36 (3C). MS (ESI-positive) m/z

+ calcd for C14H22N2NaO5 [M+Na] 321.14, found 321.00

Compound 24: To a solution of saturated NH3 / MeOH (15 ml) was added Raney nickel (50 mg, 20% weight), washed with water and MeOH) and 22 (250 mg, 0.85 mmol). The solution was stirred vigorously at room temperature under 1atm H2 for

16 hrs, filtered and evaporated to dryness. The crude was dissolved in MeOH / water

(1:5, 3 ml), CbzCl (0.20 ml, 1.50 mmol) and triethylamine (0.20 ml, 1.50 mmol) were added at 0ºC. The solution was warmed up to room temperature and kept for 16hrs, and then evaporated. The mixture was partitioned between EtOAc (50 ml) and aqueous HCl (pH = 2, 50 ml). The aqueous phase was extracted with EtOAc (50 ml x

2). The combined organic phase was washed with brine, dried over MgSO4, filtered and evaporated. The crude product was purified by flash chromatography with EtOAc and hexane (1: 3) to yield white solid 24 (213 mg, 0.49 mmol, 98%). 1H NMR

123

(CDCl3, 400MHz) δ 7.36-7.31 (m, 5H), 5.20 (app s, 1H), 5.10 (s, 2H), 4.74 (app s,

1H), 4.19 (app s, 1H), 3.67 (s, 3H), 3.62-3.57 (m, 1H), 3.54-3.49 (m, 1H), 3.03-3.95

13 (m, 1H), 2.91 (app s, 1H), 2.28-2.22 (m, 1H). C NMR (CDCl3, 100MHz) δ173.62,

157.89, 155.26, 136.59, 128.91 (2C), 128.58, 128.44 (2C), 68.05, 67.41, 52.35, 46.53,

44.50, 43.47, 40.86, 30.94, 30.40, 30.04, 28.70 (3C). HRMS (EI) m/z calcd for

C22H32N2O7 436.2210, found 436.2215.

Compound 25-27 were synthesized as published procedure. 3

Compound 28: The compound 28 was obtained by the same procedure described above for compound 16 from 15. The product was obtained in 63% yield as colorless

1 oil. H NMR (CDCl3, 400MHz) δ5.63 (d, J = 10, 1H), 4.07-4.00 (m, 1H), 3.56 (s,

3H), 3.14-3.10 (m, 2H), 2.48 (app dd, J1 = 15.6, J2 = 7.6, 1H), 2.34 (ddd, J1 = J2 =

6.4, J3 = 3.6, 1H), 2.16-2.06 (m, 2H), 2.014 (ddd,, J1 = 15.6, J2 = 6.0, J3 = 3.2, 1H),

13 1.39 (s, 9H). C NMR (CDCl3, 100MHz) δ171.91, 156.46, 81.39, 51.15, 46.12,

+ 41.24, 29.40, 28.17 (3C), 24.88. MS (ESI-positive) m/z calcd for C8H13NO3 [M+H]

171.09, found 171.07.

Compound 30: The compound 30 was obtained by the same procedure described above for compound 21 from 18. The product was obtained in 85% yield as white

1 solid. H NMR (CDCl3, 400MHz) δ 7.46-7.41 (m, 5H), 5.95 (d, J = 9.6, 1H), 4.98

(d, J = 6.4, 1H), 3.94 (app s, 1H), 3.74 (s, 3H), 3.61 (app s, 2H), 3.00 (br, 1H), 2.89 -

124

2.84 (m, 1H), 2.49 (app d, 1H), 2.24- 2.17 ( m, 1H), 1.95 ( app q, 1H), 1.59 (s, 10H);

13 C NMR (CDCl3, 100MHz) δ172.22, 157.21, 156.71, 136.43, 128.91 (2C), 128.61 ,

128.53 (2C), 82.49, 72.59, 67.44, 53.44, 52.47, 48.43, 43.93, 36.76, 31.39, 28.37 (3C).

+ MS (ESI-positive) m/z calcd for C16H22N2O5 [M+H] 322.15, found 321.97.

Compound 31: The compound 31 was obtained by the same procedure described above for compound 16 from 15. The product was obtained in 70% yield as white

1 solid. H NMR (CDCl3, 400MHz) δ7.32-7.25 (m, 5H), 5.77 (d, J = 10, 1H), 5.03 (s,

2H), 4.1-4.03 (m, 1H), 3.8 (s, 3H), 3.16-3.12 (m, 2H), 2.57 (app dd, J1 = 16, J2 = 3.4,

1H), 2.46 (ddd, J1 = J2 = 6.8, J3 = 3.2, 1H), 2.18-2.12 (m, 2H), 2.06 (ddd, J1 = 15.6,

13 J2 = 6.0, J3 = 3.2, 1H), 2.0 (s, 9H). C NMR (CDCl3, 100MHz) δ172.29, 154.77,

135.67, 127.57 (2C), 127.15, 127.11 (2C), 67.71, 51.00, 50.67, 49.99, 45.07, 39.34,

+ 28.02, 23.61. MS (ESI-positive) m/z calcd for C16H20NO5 [M+H] 306.13, found

305.94.

Compound 33: The compound 33 was obtained by the same procedure described above for compound 18 from 16. The product was obtained in 89% yield as white

1 solid. H NMR (CDCl3, 400MHz) δ 5.44 (d, J = 10.4, 1H), 3.99-3.92 (m, 1H), 3.61

(s, 3H), 3.14-3.09 (m, 2H), 2.53 (app dd, J1 = 15.6, J2 = 3.4, 1H), 2.46 (ddd, J1 = J2

= 6.8, J3 = 3.6, 1H), 2.15-2.07 (m, 2H), 2.01 (ddd,, J1 = 15.6, J2 = 6.0, J3 = 3.2, 1H),

13 1.33 (s, 9H); C NMR (CDCl3, 100MHz) δ 173.51, 155.28, 79.38, 52.02, 51.83,

125

51.05, 45.51, 40.54, 29.18, 28.49 (3C), 24.57. MS (ESI-positive) m/z calcd for

+ C13H22NO5 [M+H] 272.1498, found 271.79.

Compound 34: The compound 33 was obtained by the same procedure described above for compound 19 from 18. The product was obtained in 73% yield as white

1 solid H NMR (CDCl3, 400MHz) δ 5.64 (d, J = 6.8), 3.86 (br, 1H), 3.71 (s, 3H),

3.60 (br, 1H), 3.39 (br, 1H), 2.92 (app dd, J1 = 8.8, J2 = 4.4, 1H), 2.76 (br, 1H), 2.36

(ddd, J1= 14.4, J2 = J3= 4.4, 1H), 2.06 (ddd, J1 = 13.2, J2 = J 3 = 4.0, 1H), 1.84-1.76

(m, 1H), 1.58 ( ddd, J1 = 14.4, J2 = 10.8, J3 = 4.8, 1H), 1.4 (s, 9H). 13C NMR

(CDCl3, 100MHz) δ173.50, 155.52, 80.04, 71.64, 62.79, 52.44, 47.63, 42.74, 35.42,

28.63 (3c). HRMS (EI) m/z calcd for C13H22N4O5 314.1585, found 314.1583

Compound 36: The compound 36 was obtained by the same procedure described above for compound 19 from 21. The product was obtained in 78% yield as white

1 solid H NMR (CDCl3, 400MHz) δ5.59 (d, J = 9.2), 3.81 (br, 1H), 3.70 (s, 3H), 3.27

(br, 1H), 2.91 (app s, 1H), 2.51 (app s, 1H), 2.47-2.30 (m, 3H), 2.26 (app d, J = 13.2,

1H), 2.06 ( app d, J = 11.6 1H), 1.78 (ddd, J1 = J2 = J3 = 12, 1H), 1.41 (s, 9H). 13C

NMR (CDCl3, 100MHz) δ174.07, 155.56, 79.82, 73.93, 53.13, 52.24, 48.24, 43.90,

36.53, 34.24, 28.69 (3C). HRMS (EI) m/z calcd for C21H30N2O7 422.2048, found

422.2052

126

Compound 37: The compound 33 was obtained by the same procedure described above for compound 14 from 15. The product was obtained in 71% yield as white

solid HRMS (EI) m/z calcd for C21H30N2O7 422.2048, found 422.2053

Compound 38: The compound 38 was obtained by the same procedure described above for compound 22 from 18. The product was obtained in 62% yield as white

1 solid. H NMR (CDCl3, 400MHz) δ5.70 (d, J = 8, 1H), 4.00-3.88 (m, 2H), 3.68 (s,

3H), 2.91-2.87 (m, 1H), 2.70 (app, 1H), 2.46-2.40 (m, 1H), 2.10-2.04 (m, 1H), 1.83-

13 1.73 (m, 2H), 1.37 (s, 9H). C NMR (CDCl3, 100MHz) δ172.92, 155.45, 120.48,

80.09, 68.53, 52.46, 47.41, 42.37 (2C), 35.91, 33.64, 28.56 (3C). HRMS (EI) m/z calcd for 298.1523, found 298.1529

Compound 40: The compound 40 was obtained by the same procedure described above for compound 22 from 24. The product was obtained in 68% yield as white

1 solid over two steps (from 38 to 40). H NMR (CDCl3, 400MHz) δ7.34-7.25 (m,

5H), 5.71 (br, 1H), 5.60 (d, J = 9.2, 1H), 5.07, 5.02 (AB quartet, Jab = 12, 2H), 3.80-

3.59 (m, 4H), 3.53 (app d, J = 13.2, 1H), 3.34 (br triplet, 1H), 2.97 (app d, J = 13.6,

1H), 2.82 (s, 1H), 2.03-1.95 ( m, 2H), 1.76 (q, J = 12, 1H), 1.38 (s, 10H); 13C NMR

(CDCl3, 100MHz) δ173.98, 157.88, 155.53, 136.55, 128.71 (2C), 128.36, 128.27

(2C), 79.70, 70.16, 67.12, 52.04, 48.35, 43.42, 42.96, 41.44, 37.16, 29.08, 28.56 (3C).

HRMS (EI) m/z calcd for C22H32N2O7 436.2204, found 436.2201

127

References

(1) Michael, K.; Wang, H.; Tor, Y. Bioorganic & Medicinal Chemistry 1999, 7, 1361-1371.

(2) The white solid is the mixture compound 2a and the starting material compound 4 and some other byproducts. To assist purification, the crude product was Boc protected and purified as pure compound 6.

(3) Kobayashi, S.; Kamiyama, K.; Iimori, T.; Ohno, M. Tetrahedron Letters 1984, 25, 2557-2560.