The Pennsylvania State University

The Graduate School

Eberly College of Science

PART I. TOTAL SYNTHESIS OF CELOGENTIN C VIA A STEREOSELECTIVE C-H

FUNCTIONALIZATION

PART II. DIVERGENT TOTAL SYNTHESES OF THE UNUSUAL MONOTERPENOID

INDOLE ALKALOIDS ALSTILOBANINE A, E AND ANGUSTILODINE

A Dissertation in

Chemistry

by

Yiqing Feng

© 2014 Yiqing Feng

Submitted in Partial Fulfillment of the Requirements for the degree of

Doctor of Philosophy

May 2014

ii

The dissertation of Yiqing Feng was reviewed and approved* by the following:

Steven M. Weinreb Russell and Mildred Marker Professor of Natural Products Chemistry Dissertation Co-Advisor Co-Chair of Committee

Gong Chen Assistant Professor of Chemistry Dissertation Co-Advisor Co-Chair of Committee

Raymond L. Funk Professor of Chemistry

Alexander Radosivech Assistant Professor of Chemistry

Yan-Ming Wang Associate Professor of Biology and Molecular Biology

Barbara J. Garrison Shapiro Professor of Chemistry Head of the Department of Chemistry

*Signatures are on file in the Graduate School iii

Abstract

Part I

A total synthesis of the antimitotic bicyclic peptide celogentin C (4) has been completed.

The unique Trp C6-Leu Cβ single bond linkage of the target molecule was constructed via a highly regio- and stereoselective palladium-catalyzed aminoquinoline carboxamide-directed C-H functionalization using an 8-aminoquinoline α-phthaloylleucinamide 86 and a fully functionalized 6-iodotryptophan derivative 122 as coupling partners. The α-amino group in the coupling product 125 was converted to an azido group to reduce steric hindrance around the leucine carbonyl in azide 126. The 8-aminoquinoline auxiliary was subsequently removed via a

N-Boc activation/LiOH-cleavage sequence to afford the key azido acid 128, which was further elaborated to the left-hand ring peptide acid 135 of celogentin C (4). Compound 135 was further converted to 137 by peptide coupling with a proline benzyl ester. The key His-N/Trp C2 linkage embedded in the right-hand ring of the target molecule was furnished by adopting Castle’s strategy which involved an NCS-mediated oxidative C-N σ bond formation of the Trp-C2 of intermediate 137 with the His-N1 of an Arg/His dipeptide. The right-hand ring of the natural product celogentin C (4) was eventually formed via a macrolactamization to afford celogentin C

(4) in 23 steps from simple α-amino acid building blocks.

iv

Part II

Divergent total syntheses of the unusual monoterpenoid indole alkaloids alstilobanine A

(15), E (14) and angustilodine (13) in racemic form have been completed. The synthesis started with a highly efficient Michael addition of a dianion of indole diester 44 to α-chloro-3- piperidone-derived nitrosoalkene 25, which forged the C15-C16 σ bond, producing a pair of diastereomers 47 and 82. Subsequent functional group manipulation including a C15/C16 epimerization afforded the key C15/C16 anti-ketoacid 57. A Romo nucleophile-promoted aldol- lactonization (NPAL) of 57 furnished the advanced pentacyclic intermediate 72 with the requisite C15, C19 and C20 stereochemical configuration for the target alkaloids. The C16 hydroxymethyl group was then introduced stereoselectively via alkylation of the enolate of NH indole β-lactone ester 89a with monomeric formaldehyde to afford C16 hydroxymethylated β- lactone 99, which was conveniently converted to the corresponding C17 OTBS diol 117. A two- step Appel iodination/palladium-catalyzed reductive deiodination sequence converted intermediate 117 to the C18 methyl compound 108 and utimately provided racemic alstilobanine

A (15) after tosyl and TBS protecting group removal. After numerous unproductive attempts to access angustilodine and alstilobanine E, a silver(I)-promoted cyclic hemithioketal formation/palladium-catalyzed silane-mediated desulfurization sequence was established using

C16 hydroxymethyl dithioketal 165, which provided an efficient access to the key oxepane 136.

Syntheses of the target alkaloids were completed using oxepane 136, which afforded racemic alstilobanine E (14) in 21 steps in 16% overall yield, and racemic angustilodine (13) in 22 steps in 13% total yield.

v

Table of Contents

LIST OF FIGURES...... ………………………...... xii

LIST OF TABLES...... xiii

ACKNOWLEGEMENTS...... xiv

Part I. Total Synthesis of Celogentin C via Stereoselective C-H Functionalization………..1

Chapter 1. Introduction and Background……………………………………………………...... 1

1.1. Isolation and Structure Elucidation of the Celogentin and Moroidin Peptides...... 1

1.2. Tubulin Polymerization Inhibitory Activities of Moroidin and Celogentin Compounds...... 3

1.3. Previous Synthetic Studies on Moroidin/Celogentin Compounds...... 4

1.3.1. Previous Synthetic Studies on His N1-Trp C2 Linkage in the Right-Hand Ring of

Moroidin and Celogentin C...... 5

1.3.1.1. Synthetic Studies toward the Right-Hand Ring of Moroidin...... 5

1.3.1.2. Synthesis of Celogentin C Right-Hand Ring via a His-Trp Oxidative

Coupling...... 7

1.3.2. Previous Synthetic Studies toward the Central Tryptophan Moiety of Celogentin C and

Synthesis of Stephanotic Acid Methyl Ester...... 8

1.3.2.1. Catalytic Asymmetric Synthesis of the Central Tryptophan Residue of

Celogentin C...... 8

1.3.2.2. Asymmetric Synthesis of the Central Tryptophan Residue of Stephanotic vi

Acid...... 9

1.3.2.3. Synthesis of Stephanotic Acid Methyl Ester...... 10

1.3.3. First Total Syntheis of Celogentin C...... 13

Chapter 2. Results and Discussion...... 17

2.1. Synthetic Plan...... 18

2.2. Background on Auxiliary-Directed Palladium-Catalyzed sp3 C-H Arylation...... 20

2.2.1. Seminal Work on Regioselective sp3 C-H Arylation...... 20

2.2.2. Palladium-Catalyzed Regio- and Stereoselective sp3 C-H Functionalization of Amino

Acid Derivatives...... 23

2.3 Application of a Palladium-Catalyzed Regio- and Stereoselective sp3 C-H Arylation in Total

Synthesis of Celogentin C: Model Studies...... 27

2.3.1 Model Studies on β C-H Functionalization of Leucine Derivatives...... 27

2.3.2 Model Studies on Removal of the Aminoquinoline Auxiliary...... 28

2.4. Total Synthesis of Celogentin C...... 33

2.4.1. Preparation of the C-H Functionalization Precursors...... 33

2.4.2. Construction of the Key Leu-Trp C-C Linkage via Regio- and Stereoselective C-H

Functionalization...... 35

2.4.3. Removal of the Aminoquinoline Auxilliary...... 36 vii

2.4.4. Synthesis of the Left-hand Ring of Celogentin C...... 37

2.4.5. Completion of the Total Synthesis of Celogentin C...... 38

2.5. Concluding Remarks...... 40

Chatper 3. Experimental Section...... 41

References...... 65

Part II. Divergent Total Syntheses of the Unusual Monoterpenoid Indole Alkaloids

Alstilobanines A, E and Angustilodine...... 68

Chapter 1. Introduction and Backgrounds...... 68

1.1. Monoterpenoid Indole Alkaloids: Background...... 68

1.2. The Alstilobanine-Angustilodine Family: Unusual Monoterpenoid Indole Alkaloids.

Discovery, Structure Elucidation and Preliminary Bioactivity Studies...... 73

1.3. Synthetic Studies on the Angustilodine-Type Alkaloids...... 75

1.3.1. Original Synthesis Plan Toward the Angustilodine-Type Alkaloids...... 76

1.3.2. Previous Synthetic Work Toward the Angustilodine-Type Alkaloids in the Weinreb

Group...... 77

1.3.2.1. Background on Nitrosoalkene Conjugate Additions...... 77

1.3.2.2. Studies on Nitrosoalkene Conjugate Additions with Indole-2-acetate

Enolates...... 80

1.3.2.3. Synthesis of the Key Romo NPAL Cyclization Precursor...... 84

1.3.2.4. The Romo Nucleophile-Promoted Aldol-Lactonization (NPAL)...... 86 viii

1.3.2.5. Construction of the Cis-Fused Azadecalin Moiety of the Angustilodine-Type

Alkaloids...... 89

1.3.2.6. Studies on Introduction of the C16 Hydroxymethyl Group...... 90

Chapter 2. Results and Discussion...... 94

2.1. Problems Associated with the First Generation Majireck Approach and Some

Solutions...... 94

2.1.1. Deoximation of Intermediate 54...... 94

2.1.2. Stereochemistry of the Nitrosoalkene Conjugate Addition Step...... 95

2.1.2.1. Discovery of the C15/C16 Diastereomer Problem...... 95

2.1.2.2. Diastereoselectivity of the Nitrosoalkene Conjugate Addition...... 98

2.1.3. Reexamination of the Majireck Synthesis and Observations of Problems with the

C15/C16 Syn-Diastereomeric Series...... 101

2.1.3.1. Reexamination of the Majireck Approach Using the Pure C15/C16 anti-

Diastereomeric Series...... 102

2.1.3.1.1. Preparation of C15/C16 anti-Ketoacid 57...... 102

2.1.3.1.2. Romo NPAL of C15/C16 anti-Ketoacid 57...... 103

2.1.3.2. Reexamination of the Majireck Approach Using the C15/C16 syn-Diastereomeric

Series...... 104

2.1.3.2.1. Preparation of C15/C16 syn-Ketoacid 83...... 104 ix

2.1.3.2.2. Romo NPAL of the C15/C16 syn-ketoacid 83...... 105

2.1.3.2.3. Solution to the C15/C16 syn-diastereomer Problem...... 107

2.2. Modified Strategies for the Divergent Synthesis of the Angustilodine-Type

Alkaloids...... 108

2.2.1. C16 Hydroxymethylation...... 108

2.2.1.1. A Modified C16 Hydroxymethylation Strategy...... 108

2.2.1.2. C16 Hydroxymethylation Studies...... 110

2.2.2. C18 Deoxygenation and Total Synthesis of Alstilobanine A...... 117

2.2.2.1. The First Generation C18 Deoxygenation Approach...... 117

2.2.2.2. A C18 Deoxygenation /C16 Hydroxymethylation Approach...... 120

2.2.2.3. The Optimal End-Game Route and Total Synthesis of Racemic

Alstilobanine A...... 123

2.2.3. Total Syntheses of Alstilobanine E and Angustilodine...... 126

2.2.3.1. Attempted Oxepane Formation from ɛ-Lactone 102...... 126

2.2.3.1.1. Attempted Lewis Acid-Catalyzed ɛ-Lactone Reduction...... 127

2.2.3.1.2. Attempted Silane Radical-Mediated ɛ-Lactone Reduction...... 128

2.2.3.1.3. Thiolactonization/Reduction Approach...... 128 x

2.2.3.2. Oxepane Formation via Intramolecular Etherification...... 130

2.2.3.2.1. Attempted Intramolecular Etherification via Mitsunobu Reaction of

Triol 103...... 130

2.2.3.2.2. Attempted Intramolecular Etherifications via Mitsunobu Reactions Using

Selectively Protected Triols...... 131

2.2.3.2.3. Attempted Oxepane Formation via Intramolecular Etherification of

Halo-Alcohols ...... 133

2.2.3.3. Attempted Oxepane Formation via Intramolecular C16 Alkylation...... 136

2.2.3.4. Silver(I)-Promoted Cyclic Hemithioketal Formation/Reduction

Approach...... 138

2.2.3.4.1. Background and a Revised End-Game Synthetic Plan...... 138

2.2.3.4.2. Optimized Preparation of Hydroxy Aldehyde 104...... 140

2.2.3.4.3. Preparation of the Ag(I)-Promoted Cyclization Precursor

Dithioacetal 173...... 143

2.2.3.4.4. Ag(I)-Promoted Cyclization of C16 Hydroxymethyl Dithioacetal 173 and

Formation of the Key Cyclic Hemithioacetal...... 144

2.2.3.4.5. Reduction of the Thiohemiketals to the Oxepane...... 146

2.2.3.5. Completion of the Total Syntheses of Racemic Alstilobanine E xi and Angustilodine...... 151

2.3. Concluding Remarks...... 153

Chapter 3. Experimental Section...... 157

References...... 205

xii

List of Figures

Part I.

Figure 1. Structure of Moroidin (1)...... 1

Figure 2. Structures of the Celogentin C Natural Products...... 2

Figure 3. Moroidin Derivatives Used in the Antimitotic Activity Studies...... 3

Figure 4. Stephanotic Acid (10)...... 4

Part II.

Figure 1. The Corynanthe, Iboga and Aspidosperma Backbones in Monoterpenoid Indole

Alkaloids...... 68

Figure 2. Representative Monoterpenoid Indole Alkaloids Used As Therapeutic Agents...... 70

Figure 3. Some Representative MIAs Having Potent Biological Activity...... 71

Figure 4. Angustilodine Alkaloids: Structures and Conformations...... 73

Figure 5. Key HMBC and NOE Correlations of Hydroxymethyl β-Lactone 99...... 115

xiii

List of Tables

Part I.

Table 1. Model Studies of Removal of the Aminoquinoline Auxiliary ...... 30

xiv

Acknowledgements

I am very grateful to my advisors, Professors Steven M. Weinreb and Gong Chen, who have given me meticulous guidance and inspiration throughout my years in the Penn State

Chemistry Department. I cannot overstate my appreciation for the freedom of thought and encouragement I was given in both groups. I would also like to thank all the great individuals I worked with at Penn State who provided numerous insightful discussions and help in both chemistry and life.

I am blessed to pursue my interest in chemistry with the unconditional support from my parents. Without their support, I would not have been able to survive all the difficulties that come in research and life. This dissertation is dedicated to my beloved family.

Part I. Total Synthesis of Celogentin C via a Stereoselective C-H

Functionalization

Chapter 1. Introduction and Background

1.1. Isolation and Structure Elucidation of the Celogentin/Moroidin Peptides

Moroidin (1) (Figure 1) is a bicyclic peptide originally isolated from the stinging hairs of the eastern Australian bushy plant Laportea moroides by Williams, et al., in 1986,1 and was found to be partially responsible for the intense pain, local sweating, piloerection and arteriolar dilation experienced upon contact with the host plant.

Figure 1. Structure of moroidin (1)

More recently, Kobayashi and coworkers reisolated moroidin (1) from the seeds of

Celosia argentea, together with three previously unreported bicyclic peptides termed as celogentin A (2), B (3), and C (4).2 The same group later isolated additional members of the celogentin family such as celogentins D-H (5-7),3 from the same extract of seeds of C. Argentea

(Figure 2).

1

The structures of moroidin (1) and the celogentin compounds (2-7) were elucidated through amino acid residue sequencing, extensive NMR and MS studies,2-4 as well as computer assisted structural analysis,5 and were found to consist entirely of L-amino acids. The moroidin and celogentin compounds all share a common 17-membered left-hand ring as a common structural element. Celogentin A (2) and B (3) lack the glycine residue found in moroidin’s right- hand ring, while celogentin B (3) contains an extra histidine residue affixed to the original histidine residue on its right-hand ring. Celogentin C (4) contains an extra proline residue embedded in its right-hand ring that is not found in celogentins A and B.

Figure 2. Structures of the celogentin natural products

The moroidin and celogentin compounds both contain two unusual non-peptide linkages that are rarely observed in natural cyclic peptides: on the left-hand ring, the tryptophan C6 is

2 linked with a leucine β-C through a C-C σ bond; on the right-hand ring, the histidine N1 is directly linked to the tryptophan C2 through a C-N single bond.

1.2. Tubulin Polymerization Inhibitory Activities of the Moriodin and Celogentin

Compounds

In their study of the effect of moroidin against tubulin polymerization,6 Morita and coworkers found the metabolite’s inhibitory activity (IC50, 3.0 μM) to be more potent than that of the classical antimitotic agent colchicine (IC50, 10 μM). The bis-macrocyclic structure of moroidin was shown to be central to this activity via control experiments using a hydrolysate of moroidin (cleaved between the Arg and Gly residues using α-chymotrypsin) which showed only modest inhibition of tubulin polymerization. The pyrimidine derivative 8, prepared via treatment with 2,4-pentanedione, showed no significant decrease in tubulin polymerization inhibitory activity (IC50, 6.0 μM). Likewise, the methyl ester derivative of moroidin (9) showed similar levels of activity against tubulin polymerization (IC50, 7.0 μM). These results suggest that neither the charged free guanidine moiety in the Arg residue nor the free carboxylic acid functionality in the histidine play any major role in the antimitotic activity of 1.2

Figure 3. Moroidin derivatives used in the antimitotic activity studies

3

Kobayashi and coworkers have studied the antimitotic properties of the celogentin compounds and found that celogentins A and B display less potent tubulin polymerization inhibitory activity (IC50, 20 μM and 30 μM, respectively) than moroidin. On the other hand, celogentin C showed the most potent antimitotic activity of any of these bis-macrocyclic peptides (IC50, 0.8 μM). Kobayashi and coworkers also evaluated stephanotic acid (10), a cyclic peptide with a left-hand ring that resembles that of both the moroidin and celogentin compounds, but found no appreciable tubulin polymerization inhibition.

Figure 4. Stephanotic acid

Although the cellular binding site and specific mechanism for their antimitotic activity remain unclear, celogentins A, B, and C nevertheless represent a new class of bicyclic natural peptides that show great promise as potential chemotherapeutic agents. Further investigation of the interactions between the celogentin/moroidin compounds and the mechanisms underlying tubulin polymerization in vivo could eventually lead to the discovery of an entirely new type of cancer treatment.

1.3. Previous Synthetic Studies on the Moroidin/Celogentin Compounds

The unconventional non-peptide σ-bond connections between Leu-Trp and His-Trp, common among the moroidin and celogentin compounds, are not common structural features of

4 natural peptides. However, N-linked histidine residues are known to exist in other macrocyclic peptides. On the other hand, the Leu-Trp C-C linkage is extremely rare, and presents a significant synthetic challenge. The unusual architecture of the moroidin/celogentin compounds, along with their significant biological activities have prompted a number of synthetic studies which have particularly focused on the construction of the key Leu-Trp C-C σ-bond.

1.3.1. Previous Synthetic Studies on His N1-Trp C2 Linkage in the Right-Hand Ring of

Moroidin and Celogentin C

1.3.1.1. Synthetic Studies Toward the Right-Hand Ring of Moroidin7

Moody and co-workers have synthesized the fully protected right-hand ring of moroidin using a Horner-Wadsworth-Emmons (HWE) olefination, followed by an asymmetric hydrogenation and macrocyclization. Their synthesis began with the displacement of the C2 chlorine from aldehyde 12 by the imidazole N1 of histidinol derivative 11 in the presence of the strong base NaHMDS, forging the key His N1-Trp C2 linkage in good yield (Scheme 1).

Subsequent HWE olefination of the resulting aldehyde 13 with the Cbz-protected aminophosphonate in the presence of DBU afforded the desired dehydrotryptophan 15 in excellent yield with exclusive Z-selectivity. The alkene moiety of indole 15 was asymmetrically hydrogenated using a catalytic amount of (R,R)-DIPAMP-Rh, which gave the desired enantiomer

16 in good yield. The Boc group on the Trp-His fragment of 16 was cleaved to afford free amine

17, which was subjected to an HOAt/EDCI-mediated peptide coupling with the protected Arg-

Gly dipeptide 18 to afford cyclization precursor 19 in 61% yield. The terminal carboxyl and amino groups of this “tetrapeptide” were deprotected, and the resulting amino acid was subjected

5 to a diphenylphosphoryl azide (DPPA)-mediated macrolactamization to afford macrocycle 20, which represents the fully-protected right-hand ring of moroidin.

Scheme 1.

6

1.3.1.2. Synthesis of the Celogentin C Right-Hand Ring via a His-Trp Oxidative Coupling

Castle and co-workers likewise prepared the right-hand macrocyclic moiety of celogentin

C (4) using a pivotal intermolecular oxidative coupling between Trp-Pro dipeptide 21 and protected His-Arg dipeptide 22 (Scheme 2).8

Scheme 2.

Thus, treatment of 21 and 22 with NCS resulted in linear tetrapeptide 23, which contains the His imidazole and Trp indole N1-C2 σ bond, in moderate yield. It is believed that upon reaction with NCS, the Trp indole moiety of 21 was converted to a β-chloroiminium ion such as

21a, which was trapped by the imidazole N1 of the histidine residue, followed by rearomatization-driven elimination of HCl to generate the coupled product 23. The synthesis continued with a one-pot removal of both the benzyl ester and N-Cbz protecting groups using a palladium-catalyzed hydrogenolysis, followed by an HBTU-mediated macrocyclization to

7 generate macrocyclic tetrapeptide 24, which comprises the right-hand ring of celogentin C, in excellent yield.

1.3.2. Previous Synthetic Studies Toward the Central Tryptophan Moiety of Celogentin C and Synthesis of Stephanotic Acid Methyl Ester

1.3.2.1. Catalytic Asymmetric Synthesis of the Central Tryptophan Residue of Celogentin

C

Castle and coworkers9 have developed catalytic methodology to prepare the central tryptophan residue of celogentin C (4) that employed an electron deficient cinchonidine-derived catalyst 27 to induce chirality in the alkylation of glycinate Schiff base 25 with TES-propargyl bromide (26) (Scheme 3).

Scheme 3.

8

The reaction was found to provide optimal yield and enantioselectivity when performed at -20 oC, affording the alkylation product 28 in 79% yield and 95% ee. The Schiff base 28 was successively hydrolyzed with aqueous HCl, followed by N-Cbz protection to generate TES- propargyl derivative 29 in good yield. A palladium-catalyzed Larock indole synthesis between the propagyl derivative 29 and aromatic substituted iodoaniline 30 afforded the central tryptophan residue 31 in moderate yield and with good retention of the absolute stereochemistry of the tryptophan moiety.

1.3.2.2. Asymmetric Synthesis of the Central Tryptophan Residue of Stephanotic Acid

The Moody group has synthesized the central tryptophan-valine-isoleucine tripeptide of stephanotic acid (10), which occupies the same position as the isoleucine-leucine fragment in the left-hand ring of the moroidin peptides, using a Horner-Wadsworth-Emmons olefination/asymmetric hydrogenation strategy similar to their route to the tryptophan residue in the right-hand ring of moroidin10 (cf. Scheme 1). Thus, isoleucine-valinamide 32 was elaborated to the phosphonoglycine 34 necessary for the HWE olefination using a chemoselective rhodium- catalyzed N-H insertion of the metallocarbene derived from trimethyl diazophosphonoacetate

(Scheme 4). The olefination of 6-bromoindole-3-aldehyde (35) was conducted with phosphonoglycine 34 in the presence of DBU to afford (Z)-dehydrotryptophan peptide 36 in good yield. Dehydrotryptophan 36 was asymmetrically hydrogenated using a catalytic amount of

[(COD)Rh((S,S)-Et-DuPHOS)]OTf to afford the Trp-Val-Ile tripeptide 37 in excellent yield but with only mediocre diastereoselectivity. Further attempts to elaborate the Trp-Leu crosslink using a 6-bromoindole motif were unsuccessful.

9

Scheme 4.

1.3.2.3. Synthesis of Stephanotic Acid Methyl Ester

In 2006, Moody and co-workers reported the first total synthesis of stephanotic acid methyl ester (50).11 Due to the unsatisfactory results for elaborating the Trp C6-Leu Cβ crosslinkage via typical palladium-catalyzed coupling reactions in their previous studies,10

Moody and co-workers opted instead to employ a less direct route to install the hindered dehydroamino acid using a method developed by Hoppe12 (Scheme 5). The synthesis started with the readily accessible Boc-protected indole ketone 38, which was first treated with isothiocyanatoacetate in the presence of NaH to give the thioxo-oxazolidine 39. The free NH in

39 was then protected as the Cbz derivative and the resulting carbamate 40 underwent fragmentation via loss of COS upon treatment with potassium tert-butoxide to afford the dehydroamino acid 41 as a separable E/Z mixture. Since the asymmetric reduction of dehydroamino acid 41 was unsuccessful, an achiral reduction sequence in which the E/Z mixture of 41 was treated with Mg/MeOH to give a racemic diastereomeric mixture 42 (d.r. of desired to undesired = 2:1) was instead utilized. The major diastereomer was isolated and subjected to

10 formylation with concomitant N-Boc cleavage using dichloromethyl methyl ether and titanium(IV) chloride to afford the leu-indole aldehyde 43.

Scheme 5.

To continue with the synthesis (Scheme 6), Moody, et al. next subjected indole aldehyde ethyl ester 44 to a basic hydrolysis followed by PyBOP-mediated peptide coupling with isoleucine tert-butyl ester to afford tripepetide precursor 45 in good yield. Following a strategy similar to that employed in their earlier approach to the central tryptophan residue of stephanotic acid,10 indole aldehyde 45 was N-Boc protected in order to activate the aldehyde moiety, and then subjected to a Horner-Wadsworth-Emmons olefination with phosphonoglycine 46 to afford a separable 1:1 mixture of Z-dehydroamino acid 47 diastereomers. Both diastereomers of 47 were subjected to Rh-catalyzed asymmetric reduction, since the compound possessing the desired configuration for elaboration to the natural product could not be identified at this stage.

This reaction proceeded in nearly quantitative yield affording both isomers of the saturated

11 tetrapeptide 48 in excellent de. Following N-Boc removal of 48, a HATU-mediated macrolactamization was carried on to afford the cyclic peptide 49 in moderate yield from both diastereomers. Hydrogenolytic N-Cbz cleavage over Pearlman’s catalyst followed by direct coupling with pyroglutamic acid then provided the stephanotic acid methyl ester 50 in 14% overall yield from 49 after extensive HPLC purification to remove the undesired epimer after the macrolactamization step.

Scheme 6.

12

1.3.3. First Total Syntheis of Celogentin C

In 2009, the Castle group reported the first total synthesis of celogentin C (4).13 Their approach, which was based upon many previous synthetic studies, tackles the construction of the

Leu-Trp C-C cross-link at an early stage using a pivotal Knoevenagel condensation of nitroacetamide 52 with the indole C6 aldehyde 51 using titanium(IV) chloride and N- methylmorpholine to deliver α,β-unsaturated α-nitro amide 53 (Scheme 7). After successful installation of the C-C cross link, the original route forward involved using a chiral Lewis acid- promoted radical conjugate addition to introduce the isopropyl motif in 54 stereoselectively.

However, disappointingly it was discovered that conducting the conjugate addition in the presence of Mg-DBFOX (DBFOX = 4,6-dibenzofurandily-2,2’-bisoxazoline) or a second generation DBFOX catalyst14 did not provide any diastereoselectivity. An exploration of alternative conditions eventually led to a substrate-controlled reaction with the achiral Lewis acid,

Zn(OTf)2, which afforded a mixture of four diastereomers (54a-d) in a 1:2.9:2.0:1.2 ratio following reduction of the nitro group. This mixture, although not totally separable, did favor the desired diastereomer 54b. Thus, the two major isomers 54b, c were carried on to following peptide coupling with pyroglutamic acid to afford pentapeptide 55 in excellent yield. After removal of both N-Cbz and O-benzyl protecting groups using palladium catalyzed hydrogenolysis, the diastereomeric mixture of the free amino acids 56b, 56c was finally separated to afford the desired isomer 56b in moderate yield.

13

Scheme 7. Castle’s first total synthesis of celogentin C (4)

The requisite free amino acid 56 was next subjected to macrolactamization using a

HOBt/HBTU-mediated process to afford cyclic peptide 57 as a single diastereomer in good yield

(Scheme 8). Simultaneous tert-butyl and TES removal in the presence of B-bromocatecholborane delivered the free acid form of the left-hand ring 58 of celogentin C in good yield. The NMR

14 spectra of the cyclic acid methyl ester 59 were virtually identical to that of stephanotic acid methyl ester (50) reported earlier by the Moody group.11

Scheme 8.

To complete the total synthesis of celogentin C, free acid 58 was coupled with proline benzyl ester in the presence of EDCI and HOBt to afford the hexapeptide precursor 60 (Scheme

9). A NCS mediated intermolecular oxidative coupling with His-Arg dipeptide 61 was then performed in the presence of excess proline benzyl ester (to suppress extensive chlorination on the proline residue of the substrate), which successfully forged the key His N1-Trp-C2 cross- linkage. Subsequent N-Cbz and O-Bn deprotection afforded free amino acid octapeptide 62. The

15 closure of the right-hand ring of celogentin C was then achieved through an HOBt/HBTU- mediated macrolactamization. Scission of both N-Pbf and O-tBu protecting groups was achieved via treatment with TFA to afford celogentin C (4) in good yield.

Scheme 9.

16

Chapter 2. Results and Discussion

Previous synthetic attempts directed toward the celogentin and moroidin families were largely inefficient, requiring lengthy routes with low stereo-/diastereoselectivities for some key transformations, particularly the construction of the unique leucine Cβ-tryptophan C6 cross linkage (vide supra). In Moody’s total synthesis of stephanotic acid (cf. Scheme 6), for example, the pivotal Leu-Trp C-C linkage required four linear steps for its installation and gave a 39% yield with only a 25% de, requiring extensive chromatographic separation to obtain the desired diastereomer. Similarily, the Castle total synthesis of celogentin C (cf. Scheme 7) relied upon a low-yielding four-step linear Knoevenagel condensation/conjugate addition approach to the Cβ-

C6 linkage that produced a mixture of four diastereomers, with the desired isomer only slightly favored.

Given the challenge that the unusual Leu-Trp C-C linkage has posed to synthetic chemists, our goal was to develop a more efficient total synthesis of celogentin C with a particular focus on streamlining the installation of this pivotal structural motif, which resulted in the total synthesis of celogentin C in 2009 (vide infra).15

17

2.1. Synthetic Plan

Our retrosynthesis of celogentin C involved a macrolactamization of the Pro/Arg of intermediate 63 similar to that utilized by Castle, et al.13 in order to close the right-hand ring of the target molecule(Scheme 10). The unusual Trp C2-His N1 cross linkage would be constructed using Castle’s NCS-mediated intermolecular oxidative coupling of the left-hand cyclic peptide

65 with His-Arg dipeptide 64. Construction of the left-hand macrocycle 66 would be achieved via macrolactamization at the Val-Trp residues following installation of the pyroglutamic acid and Val-Leu dipeptide residues 68 using classic peptide coupling methods. Our planned synthesis would tackle the challenging Leu-Cβ/Trp-C6 cross linkage at an early stage using a palladium-catalyzed aminoquinoline-directed regio- and stereoselective leucine β C-H functionalization strategy to join an aminoquinoline leucinamide 70 and an 6-iodotryptophan derivative such as 71.

18

Scheme 10: Retrosynthetic analysis of celogentin C (4)

19

2.2. Background on Auxiliary-Directed Palladium-Catalyzed Sp3 C-H Arylation

2.2.1. Seminal Work on Regioselective Sp3 C-H Arylation

In 2005, Daugulis and co-workers reported the first successful C-C σ-bond formation using a palladium-catalyzed regioselective sp3 C-H arylation.16, 17,18 This seminal discovery was a milestone for C-H functionalization, since until that time all reported sp3 C-H functionalizations were limited to those forming a heteroatom-carbon single bond.19

The Daugulis group used 8-aminoquinoline as a directing auxiliary, which was embedded in a series of alkyl carboxamides 72 (Scheme 11). Upon heating these amides 72 with excess aryl iodide 73 in the presence of a catalytic amount of palladium acetate and a stochiometric amount of silver acetate under solvent-free conditions, the arylation products 74a- d resulting from β-C-H activation were isolated in good yields and excellent regio-selectivity

(entries 1-4). Daugulis, et al. found that secondary C-H bonds reacted faster than primary, resulting in tetraarylation product 74d (entry 4). Thus, the Daugulis methodology serves as a synthetic surrogate for the 1,4-conjugate addition of an aryl organometallic species to an α,β- unsaturated acid.

20

Scheme 11.

The Daugulis group also reported a similarly unprecedented regioselective γ-arylation of amines (Scheme 12), in which the 2-picolinamide embedded in 74 was used as the directing auxiliary. These γ-arylations proceeded more slowly relative to the aminoquinoline-directed β-C-

21

H activation, but gave better selectivity for monoarylations, with primary C-H bonds being more reactive than secondary C-H bonds (entries 1-2). Aryl bromides were much less reactive.

Scheme 12.

In studies to elucidate the mechanism of these transformations, Daugulis and coworkers isolated the key palladium amide intermediate 76, which was characterized by X-ray crystallography. They hence postulated that these highly regioselective C-H functionaliztions commenced with formation of a palladium amide species 78 from substrates 77, which subsequently underwent γ-C-H bond cleavage to preferentially form a five-membered

22 palladacycle such as 79. The aryl iodide then undergoes oxidative addition to the palladium center resulting in a Pd (IV) species and then ultimately leads to the arylated product upon reductive elemination. The Pd(OAc)2 catalyst is subsequently regenerated from the unreactive iodide bonded Pd(II) species.

Scheme 13. Proposed mechanistic pathway of the C-H arylation

2.2.2. Palladium-Catalyzed Regio- and Stereoselective Sp3 C-H Functionalization of Amino

Acid Derivatives

Based on the aforementioned discovery made by Daugulis, et al., Corey and co-workers developed a highly regio- and stereoselective β-C-H functionalization of α-amino acid derivatives (Scheme 14).20 In this report, a variety of α-N-phthaloylamino acid 8-aminoquinoline amides 81 were subjected to oxidative acetoxylation with Mn(OAc)2, Oxone and acetic anhydride in nitromethane in the presence of palladium acetate as catalyst. After heating the reaction mixture at 80 oC for 22 h, the corresponding β-acetoxy compounds 82 were isolated in moderate yields with good to excellent diastereoselectivities favoring the erythro products.

23

Scheme 14. Palladium-catalyzed stereoselective β-acetoxylation of amino acid derivatives

A mechanism based on concerted palladation-deprotonation pathway17g is likely operative here (Scheme 15) leading to the formation of the five-membered palladacycle intermediate 84. Corey, et al. attributed the observed erythro-selectivity to the energetic favorability of the putative trans-palladacycle intermediate 84 over the corresponding cis-isomer

(vide infra). Oxidation of the resulting palladacycle intermediate to a new higher valent Pd- complex and the subsequent reductive elimination should lead to the acetylated product. Similar oxidative addition/reductive elimination with aryl iodide is likely responsible for the Pd- catalyzed C-H arylation reactions.

Scheme 15. Mechanistic pathway for Pd(II)-catalyzed β-acetoxylation proposed by Corey

24

This same series of α-N-phthaloylamino acid 8-aminoquinoline amides were also shown to undergo smooth β-arylation with aryl iodides using similar conditions to those employed by

Daugulis, et al., affording the β-arylated compounds 87 in good yields (Scheme 16) under solvent-less conditions. Secondary β C-H bonds were found to be more reactive than primary β

C-H bonds, leading to the diarylation product 89 (Scheme 16 (b)). In the case of tertiary β C-H bonds (Scheme 16 (c)), preferrential γ-arylation was observed, presumably due to a significant steric effect that renders the β-CH inaccessible to the palladium amide species (cf. Scheme 13).

The observed diastereoselectivity for the γ-arylation products 91 and 93 can again be rationalized using a palladacycle intermediate similar to 84, in which the trans-isomer is energetically favored. The β-arylation products of the leucine aminoquinoline series were of particualr interest to us, since they possess same relative stereochemistry as the Leu-Trp motif embedded in celogentin C.

Scheme 16 (a).

Scheme 16 (b)

25

Scheme 16 (c)

2.3 Application of a Palladium-Catalyzed Regio- and Stereoselective Sp3 C-H Arylation in

Total Synthesis of Celogentin C: Model Studies

2.3.1 Model Studies on β C-H Functionalization of Leucine Derivatives

We began our preliminary investigation of β C-H functionalization as a method to install the Leu/Trp Cβ-C6 σ bond in celogentin C (4) by first conducting a model study with 8- aminoquinoline leucinamide 86 and N-protected 6-iodoindole 94 (Scheme 17). Corey and coworkers had conducted their Pd-catalyed β-arylations of amino acid derivatives under solvent- free conditions to afford the arylated products excellent yield (cf. Scheme 16 (a)). However, the high melting point of the iodoaryl coupling partner in our model system dictated the use of a solvent. After extensive screening, a minimal amount of t-BuOH was found to be the optimal solvent choice since it did not interfere with the C-H funtionalization. Thus, a mixture of leucine derivative 86 and iodoindole 94 in a 2:1 ratio were heated at 110 oC in a sealed tube, together

26 with a catalytic amount of palladium acetate (20 mol%) and the iodide scavenger silver acetate

(1.5 eq) in a minimal volume of t-BuOH solvent for 16 h. Gratifyingly, the desired β-indolynated leucine product 95 was isolated in 80% yield as a single diastereomer with the desired relative configuration.

Scheme 17. Model Pd(II)-catalyzed β-indolynation of 8-aminoquinoline leucinamide 86

2.3.2 Model Studies on Removal of the Aminoquinoline Auxiliary

In order for our approach to be feasible, a facile cleavage of the aminoquinoline directing auxiliary needed to be performed under mild conditions following the arylation step. One such mild procedure for the cleavage of an amide 96 that attracted our attention was the carbamate activation method developed by Grieco, et al. (Scheme 18). 21 In this protocol, a secondary amide NH is activated with a sterically demanding Boc group to form imide 97 which directs an external nucleophile to attack the amide carbonyl, resulting in expulstion of the incipient tert- butyl carbamate 99 and the nucleophile-derived carbonyl product 98.

Scheme 18. Grieco’s amide cleavage protocol

27

Thus, arylated leucine 95 was treated with Boc anhydride along with a catalytic amount of DMAP to furnish the N-carbonylated substrate 100 in high yield (Scheme 19). N-Boc compound 100 was then treated with the nucleophile LiOH in THF/H2O in an attempt to deliver the free acid 101 necessary for the subsequent peptide coupling step. This reaction, however, resulted in a complex mixture of polar products.

Scheme 19.

We then screened a variety of other common nucleophiles in this reaction, none of which provided any detectable quantity of the desired cleavage products (Table 1). We reasoned that the electrophillic nature of phthalimide-protecting group was complicating the reaction. In fact, the reaction of 100 with excess ethanethiol and LiHMDS yielded phthalimide ring opening product 102 with concomitant loss of the Boc group (Entry 1, Table 1). As a control experiment, the aminoquinoline leucinamide 86 was converted to the N-Boc derivative 103 (cf. Scheme 19).

The Greico acyl cleavage of 103 afforded the desired thioester 104 instead of Boc cleavage, with concomitant phthalimide ring opening (entry 2, Table 1). In an attempt to remove the complications caused by the phthalimide protecting group, the indolynated compound 100 was treated with ethylenediamine in n-BuOH (entry 3, Table 1) to afford free amine 105 resulting from complete cleavage of the phthalimide group. Interestingly, simultaneous N-Boc cleavage was again observed. We next tested the Grieco protocol on 2,5-dimethylpyrrole-protected

28 compound 106, which we prepared from 105 (entry 4, Table 1). Although phthalimide opening had successfully been suppressed by this maneuver, the desired Boc-aminoquinoline cleavage did not occur, instead resulting in clean N-Boc removal once again. Thus, it appeared to us that the nucleophile was selectively attacking the Boc carbonyl group rather than that of the desired leucinamide carbonyl motif of indolynated compounds 100 and 106 in a manner contrary to what was generally observed in the Greico amide cleavage (cf. Scheme 18).

Table 1. Model studies of removal of the aminoquinoline auxiliary

29

In the usual Grieco secondary amide cleavage protocol, the Boc carbamate carbonyl is usually less reactive toward nucleophillic attack than the amide carbonyl due to the resonance electron donating effect of the adjacent t-butoxy functional group. In addition, the bulky tert- butoxy moiety usually provides sufficient steric bias to inhibit nucleophilic attack on the carbamate motif, thus further favoring cleavage of the amide over the carbamate. However, in the case of Leu-Trp substrate 100, the indole moiety installed via C-H functionalization reverses the selectivity of the Grieco amide cleavage. We hypothesize that the observed reactivity is due to the indole, where steric hindrance around the leucinamide carbonyl has become greater than that near the Boc group.

To solve this problem, as well as to avoid the complications caused by the phthalimide group, we decided to decrease the steric bulk of the leucine α-amino protecting group. However, common protecting groups such as Cbz were ruled out since the C-H functionalization reaction was observed to be totally suppressed if the α-amino group of the leucinamide substrate was not bis-protected, such as the problems in using substrate 108 (Scheme 20). Further complicating matters, we reasoned that during aminoquinoline amide Boc activation, a second Boc group would be installed on the leucine α-carbamate NH, thus making the steric hindrance around the leucinamide moiety even worse. Thus, it appeared most practical to perform a protecting group manipulation after the C-H functionalization step.

30

Scheme 20.

The azido group is a commonly used surrogate for a free amine functionality in peptide and carbohydrate synthesis. Wong and co-workers have developed convenient diazo-transfer methodology22 in which free amines are readily converted to azides via treatment with triflic azide in the presence of triethylamine (Scheme 21). We thought that the small size of the azido functionality might solve the steric problems we encountered during the course of the aminoquinoline auxiliary removal.

Scheme 21. Diazo-transfer sequence

Thus, the phthalimide aminoquinoline compound 95 was first treated with ethylene diamine in butanol at 90 oC to smoothly cleave the phthalimide group (Scheme 22). The subsequent diazo-transfer reaction of the free amine 112 with triflic azide and triethylamine afforded the azido compound 113 in good yield. Activation of the aminoquinoline amide NH was performed using Boc anhydride in the presence of DMAP, which afforded the N-Boc derivative 114 in quantitative yield. We then tested the feasibility of aminoquinoline auxiliary

31 removal on azido substrate 114 using LiOH/H2O2. Gratifyingly, desired α-azido acid 115 was isolated along with the N-Boc aminoquinoline (116) in excellent yield.

Scheme 22.

2.4. Total Synthesis of Celogentin C

2.4.1. Preparation of the C-H Functionalization Precursors

With the directing auxiliary removal problem solved, we embarked on the total synthesis of celogentin C. Our synthesis commenced with preparation of the 6-iodoindole counterpart for the pivotal C-H activation step (Scheme 23). To this end, commercially available α-N-Boc- tryptophan (117) was first converted to the t-butyl ester 118 with t-butyl bromide and the phase transfer catalyst triethylbenzylammonium chloride in good yield. The indole moiety of t-butyl ester 118 was then nitrated using nitric acid in CH2Cl2. This reaction produced an inseparable mixture of nitro regioisomers, as well as products of Boc and/or t-butyl ester cleavage, the product mixture was treated with sodium hydride and tosyl chloride in DMF to furnish the desired 6-nitro-N-tosyl indole ester 120 in modest yield, which was conveniently separated from other regioisomers and products by silica gel column chromatography.

32

Scheme 23.

To introduce the 6-iodine substituent, we planned to effect a reduction/Sandmeyer sequence on nitro compound 120 (Scheme 24). Thus, the nitro indole 120 was readily converted to amino indole ester 121 using a palladium-catalyzed hydrogenation. Aminoindole 121 was then subjected to a standard Sandmeyer iodination using sodium nitrite and sulfuric acid. However, this Sandemeyer reaction resulted in the isolation of a significant amount of reduction product

123, which was unfortunately inseparable from the desired 6-iodoindole ester 122. We reasoned that this undesired reduction product arose from rapid decomposition of the unstable diazonium intermediate 124 via a radical pathway.

After extensive screening and optimization, we eventually solved this problem by adopting the modified Sandmeyer method developed by Campagne and co-workers.23 To this end, aminoindole ester 121 was first treated with nitrosonium tetrafluoroborate in acetonitrile, followed by the addition of a large excess of a solid mixture of iodine/KI. The desired 6- iodoindole ester 122 was isolated in moderate yield without contamination by the reduced

33 product 123. We believe that the tetrafluoroborate anion helps to stabilize the diazonium intermediate, which is then converted to the iodide 122 via a radical pathway.

Scheme 24.

2.4.2. Construction of the Key Leu-Trp C-C Linkage via Regio- and Stereoselective C-H

Functionalization

With 6-iodoindole ester 122 in hand, we were ready to test the pivotal C-H functionalization on this fully functionalized synthetic intermediate. Thus, iodoindole ester 122 and leucine aminoquinoline compound 86 were admixed with Pd(OAc)2 and AgOAc in a minimal amount of t-BuOH and the mixture was heated in a sealed vessel at 110 oC for 36 hours

(Scheme 25). To our delight, the desired coupling product 125 was isolated as a single diastereomer in 85% yield. This C-H functionalization process was reproducible on a relatively large scale, with a four-gram reaction proceeding smoothly to give product 125 without any decrease in yield or stereoselectivity.

34

Scheme 25. C-H Functionalization of substrate 86 with iodoindole 122

2.4.3. Removal of the Aminoquinoline Auxiliary

Although we had successfully removed the aminoquinoline directing auxiliary in a model system (cf. Scheme 22), we needed to determine whether that strategy would work on the real system. We were primarily concerned that incorporation of the fully protected tryptophan side chain in 125 might change the steric environment around the leucinamide carbonyl. Thus, the N- phthaloyl group of 125 was first removed using ethylene diamine in n-butanol followed by reaction with triflic anhydride to afford the azido compound 126 in good yield (Scheme 26).

Further N-protection of 126 was carried out using Boc anhydride and DMAP at 70 oC in acetonitrile, which resulted in the desired aminoquinoline amide NH carbomoylation and bis-Boc protection of the tryptophan α-amino group to afford globally protected substrate 127 in satisfactory yield. This Boc activated compound was then subjected to the LiOH/H2O2 amide cleavage conditions which, to our delight, cleaved the N-Boc aminoquinoline moiety to provide the desired azido acid 128 in quantitative yield.

35

Scheme 26.

2.4.4. Synthesis of the Left-Hand Ring of Celogentin C

Our synthesis of celogentin C (4) continued by converting azido acid 128 to an activated

N-hydroxysuccinimide ester, which underwent direct peptide coupling with NH2Leu-Val-OH to furnish acid 130 (Scheme 27). Removal of both Boc groups from the Trp motif was performed using hydrogen chloride in dioxane at room temperature which provided the macrolactamization precursor 131 in good yield. Amino acid 131 was then treated with EDCI/HOOBt, which furnished the desired cyclic peptide 132 in high yield.

36

Scheme 27.

At this stage, azido compound 132 was subjected to a palladium-catalyzed hydrogenation to furnish the free amine 133. Interestingly, when this reaction was performed in methanol, only the N,N-dimethylated product was isolated. The cause of this side reaction is unknown. However, switching the solvent to ethyl acetate cleanly afforded the desired free amino compound 133 in good yield. To complete the left-hand ring of celogentin C, cyclic amino peptide 133 was combined with pyroglutamic acid in an EDCI-mediated peptide coupling to provide pentapeptide

134 in excellent yield. The indole N-tosyl group of 134 was subsequently cleaved using magnesium turnings in anhydrous methanol with sonication at room temperature. Subsequent t-

Bu ester cleavage with TFA provided macrocyclic acid 135. Carboxylic acid 135 was then converted to the corresponding methyl ester 136 using thionyl chloride and methanol, which was identical in proton and carbon-13 NMR to the advanced celogentin C intermediate reported by the Castle group. 13

37

Scheme 28.

2.4.5. Completion of the Total Synthesis of Celogentin C

To finish the synthesis, macrocyclic acid 135 was coupled with benzyl prolinate to furnish benzyl ester 137 (Scheme 29). Using an approach developed by the Castle group, benzyl ester 137 was then subjected to an NCS-mediated oxidative coupling with the dipeptide

CbzNHArg(Pbf)-HisOtBu, which installed the pivotal His-N1-Trp-C2 linkage in compound 138.

After hydrogenolytic removal of the Cbz and benzyl groups, the right-hand ring of celogentin C was formed via an HBTU/HOBt-mediated peptide coupling. Removal of the Pbf group on the guanidine moiety using aqueous TFA then furnished celogentin C (4), which was identical in every respect to natural celogentin C and to synthetic samples provided by Castle.24

38

Scheme 29.

2.5. Concluding Remarks

In conclusion, we have completed a stereoselective total synthesis of the antimitotic macrocyclic peptide celogentin C (4) relying on a highly efficient palladium catalyzed sp3 β C-H functionalization of an aminoquinoline leucinamide and a fully functionalized tryptophan to stereoselectively forge the unique Trp C6-Leu Cβ σ bond.

Our synthesis demonstrates the efficiency of C-H functionalization as a powerful method in complex natural product synthesis. Although C-H functionalization has been the recent focus of numerous methodological studies, the application of such a directed unactivated sp3 C-H indonylation in complex natural product total synthesis has hitherto not been reported.

39

We expect that future applications of this useful methodology in complex target molecule synthesis will emerge. However, we are also aware that the requirement for directing-auxiliaries nessesitates both pre-functionalization and removal strategies that might cause unforeseen problems in synthetic work (cf. Scheme 19, Table 1). Thus the development of new C-H functionalization methodologies in which an easily cleavable intermediate bearing directing groups such as imine or enamine would no doubt add to the efficiency of such strategies.

40

Chapter 3. Experimental Section

General Methods. All non-aqueous reactions were carried out in oven- or flame-dried glassware under an argon atmosphere. All chemicals were purchased from commercial vendors and used as recieved, unless otherwise specified. Tetrahydrofuran (THF) and dichloromethane

(CH2Cl2) were obtained from a solvent purification system (Glass Contour). Reactions were magnetically stirred and monitored by thin layer chromatography (TLC) with 250 μm EMD 60

F254 precoated silica gel plates. Flash column chromatography was performed using Silicycle silica gel P60 (230-400 mesh). Sonication reactions were performed in a Branson 1510 sonicator.

1H and 13C NMR spectral data were recorded on Bruker DPX-300, AMX-360 or DRX 400 MHz spectrometers. Chemical shifts are reported relative to chloroform (δ 7.26), methanol (δ 3.31), or

DMSO (δ 2.50) for 1H NMR and chloroform (δ 77.2), methanol (δ 49.0), or DMSO (δ 39.5) for

13C NMR.

41

Synthesis of Arylated Compound 95. To a 4 mL glass vial with a PTFE-lined cap was added compound 2 (63 mg, 0.16 mmol, 2 eq.), 3 (32 mg, 0.08 mmol, 1 eq), Pd(OAc)2 (3.7 mg,

0.016 mmol, 20 mol%), AgOAc (21 mg, 0.12 mmol, 1.5 eq), and 0.5 mL of t-BuOH. The vial was capped with or without argon flushing, covered with aluminum foil, and heated at 110 °C for 16 h. The reaction mixture was then cooled to rt and diluted with dichloromethane (10 mL).

The mixture was filtered through a pad of Celite and the filtrate was concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 20-50%

23 EtOAc/hexanes) to furnish compound 95 as a white solid (43 mg, 81%). [α] D -49° (c 1.4,

1 CHCl3); H NMR (300 MHz, CDCl3) δ 9.72 (s, 1H), 8.54 (dd, 1H, J = 3.1, 5.6 Hz), 8.16 (s, 1H),

7.96 (m, 2H), 7.92 (d, 1H, J = 8.9 Hz), 7.85 (d, 1H, J = 2.6 Hz), 7.78 (m, 2H), 7.63 (d, 2H, J =

8.6 Hz), 7.58 (d, 2H, J = 8.0 Hz), 7.40-7.30 (m, 3H), 7.12 (m, 1H), 6.72 (d, 1H, J = 3.4 Hz), 6.65

(d, 2H, J = 7.9 Hz), 5.76 (d, 1H, J = 12.3 Hz), 4.34 (dd, 1H, J = 2.8, 12.1 Hz), 2.08 (m, 1H), 1.89

13 (s, 3H), 0.84 (d, 3H, J = 6.7 Hz), 0.74 (d, 3H, J = 6.7 Hz); C NMR (75 MHz, CDCl3) δ 168.7,

166.2, 147.7, 144.6, 138.1, 135.3, 134.7, 134.3, 133.1, 132.0, 130.3, 129.3, 127.4, 126.9, 126.7,

126.6, 123.8, 121.6, 121.5, 121.2, 116.8, 108.8, 57.8, 48.9, 29.6, 21.6, 21.1, 16.1; IR (film) 3281,

-1 + + 2953, 1709, 1527, 1381, 1172 cm ; HRMS (m/z): [M+H ] calcd for C38H33N4O5S , 657.2166; found 657.2171.

42

Synthesis of Amine 112. Compound 95 (82 mg, 0.125 mmol) was dissolved in 5 mL of n-BuOH at rt. Ethylenediamine (85 μL, 1.27 mmol) was then added and the resulting bright

o yellow solution was heated at 90 C and stirred under N2 for 1 h. n-BuOH was evaporated in vacuo to give a residue which was purified by flash chromatography on silica gel (50%

23 1 EtOAc/hexanes) to afford 112 as a white foam (56 mg, 85%). [α] D-15°; H NMR (300 MHz,

CDCl3) δ 8.84 (dd, J = 7.2, 1.5 Hz, 1H), 8.72 (dd, J = 4.1, 1.4 Hz, 1H), 8.12 (dd, J = 8.3, 1.3 Hz,

1H), 7.96 (s, 1H), 7.70 (d, J = 8.3 Hz, 2H), 7.50 (m, 2H), 7.46 (d, J = 3.6 Hz, 1H), 7.40 (m, 1H),

7.36 (d, J = 8.2 Hz, 1H), 7.16 (d, J = 8.1 Hz, 1H), 7.06 (d, J = 8.1 Hz, 2H), 6.54 (d, J = 3.6 Hz,

1H), 3.95 (d, J = 4.1 Hz, 1H), 3.14 (dd, J = 10.0, 4.0 Hz, 1H), 2.60 (m, 1H), 2.25 (s, 3H), 1.90

13 (br s, 2H, NH), 1.16 (d, J = 6.5 Hz, 3H), 0.74 (d, J = 6.5 Hz, 3H); C NMR (75 MHz, CDCl3) δ

172.6, 148.4, 144.8, 139.0, 138.6, 136.1, 135.2, 135.1, 134.4, 129.7, 129.6, 129.0, 128.2, 128.0,

127.3, 126.7, 126.2, 124.6, 121.6, 121.4, 121.3, 116.5, 113.8, 109.0, 58.7, 57.8, 29.0, 21.6, 21.5,

21.4; IR (film) 3277, 2949, 1665, 1523, 1365, 1164 cm-1; HRMS (m/z): [M+H+] calcd for

+ C30H31N4O3S , 527.2111; found 527.2107.

43

Synthesis of Azide 113. To a solution of triflic azide (1.06 mmol) in dichloromethane

(0.5 mL) was added amine 112 (56 mg, 0.11 mmol) and triethylamine (20 µL, 0.16 mmol). The resulting solution was stirred at rt for 2 h before the solvent was removed in vacuo. Flash chromatography of the residue on silica gel (gradient 20-30% EtoAc/hexanes) afforded azide 113

23 1 as a white solid (55 mg, 95%). [α] D+7.5°; H NMR (300 MHz, CDCl3) δ 10.25 (s, 1H, NH),

8.64 (dd, J = 5.3, 3.6 Hz, 1H), 8.60 (dd, J = 4.1, 1.4 Hz, 1H), 8.11 (dd, J = 8.3, 1.4 Hz, 1H), 7.97

(s, 1H), 7.71 (d, J = 8.3 Hz, 2H), 7.51 (s, 1H), 7.49 (d, J = 1.8 Hz, 1H), 7.44 (d, J = 3.6 Hz, 1H),

7.39 (m, 1H), 7.32 (d, J = 7.3 Hz, 1H), 7.12 (d, J = 8.2 Hz, 1H), 7.06 (d, J = 8.1 Hz, 2H), 6.50 (d,

J = 3.5 Hz, 1H), 4.57 (d, J = 6.0 Hz, 1H), 3.25 (dd, J = 8.0, 6.3 Hz, 1H), 2.58 (m, 1H), 2.25 (s,

13 3H), 1.08 (d, J = 6.6 Hz, 3H), 0.80 (d, J = 6.7 Hz, 3H); C NMR (75 MHz, CDCl3) δ 167.2,

148.4, 144.8, 138.5, 136.1, 136.0, 135.1, 135.0, 133.5, 129.9, 129.7, 127.8, 127.1, 126.7, 126.4,

124.4, 122.2, 121.6, 121.2, 116.8, 114.2, 109.0, 67.9, 55.2, 29.0, 21.5, 21.4, 20.3; IR (film) 3320,

-1 + + 2105, 1674, 1526, 1369, 1167 cm ; HRMS (m/z): [M+H ] calcd. for C30H29N6O3S , 553.2016; found 553.2022.

44

Synthesis of Carbamate 114. Azide 113 (41 mg, 0.074 mmol) and Boc2O (81 mg, 0.37 mmol) were dissolved in 0.5 mL of MeCN at rt with stirring. DMAP (20 mg, 0.16 mmol) was added in one portion, and stirring was continued for another 45 min at rt. The brown-red reaction mixture was evaporated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) to afford carbamate 114 as a light yellow foam

23 1 (46 mg, 96%). [α] D +64.9 °; H NMR (300 MHz, CDCl3) δ 8.71 (s, 1H, NH), 8.05 (d, J = 8.2 Hz,

1H), 7.95 (s, 1H), 7.76 (d, J = 8.2 Hz, 2H), 7.60 (d, J = 8.5 Hz, 1H), 7.56 (d, J = 3.6 Hz, 1H),

7.47 (d, J = 8.1 Hz, 1H), 7.31 (m, 1H), 7.16 (d, J = 8.1 Hz, 1H), 7.06 (d, J = 8.1 Hz, 2H), 6.94 (s,

1H), 6.66 (d, J = 3.5 Hz, 1H), 5.80 (d, J = 10.5 Hz, 1H, Leu-α), 3.38 (dd, J = 10.4, 4.6 Hz, 1H,

Leu-β), 2.47 (m, 1H), 1.21 (s, 3H), 0.93 (d, J = 1.9 Hz, 3H), 0.91 (d, J = 1.9 Hz, 3H); 13C NMR

(75 MHz, CDCl3) δ 173.2, 152.7, 150.3, 144.8, 143.9, 136.0, 135.5, 135.1, 134.9, 134.7, 129.8,

129.7, 128.4, 127.8, 127.6, 126.9, 125.8, 125.0, 121.4, 120.7, 115.3, 108.8, 83.4, 62.8, 53.7, 29.0,

27.5, 21.5, 18.0; IR (film) 2965, 2100, 1725, 1374, 1168, 1125 cm-1; HRMS (m/z): [M+H+] cald.

+ for C35H37N6O5S , 653.2541, found 653.2544.

45

Synthesis of Azido Acid 115. To a solution of carbamate 114 (36 mg, 0.055 mmol) in

THF/H2O (4:1, 1 mL) at 0 °C was added H2O2 (30%, 30 µL, 0.28 mmol) and LiOH (4.6 mg,

0.11 mmol). The mixture was then allowed to warm to rt and stirred vigorously for 2 h. Na2SO3

(150 mg in 3 mL of H2O) was added and the mixture was stirred for 5 min before acidification to pH 2 with 0.5 M HCl. The reaction mixture was then extracted with dichloromethane. The organic layers were dried over Na2SO4 and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes) to afford azido

23 1 acid 115 as a white solid (22 mg, 94%). [α] D -31°; H NMR (300 MHz, CDCl3) δ 7.87 (s, 1H),

7.75 (d, J = 8.1 Hz, 2H), 7.52 (d, J = 3.6 Hz, 1H), 7.40 (d, J = 8.0 Hz, 1H), 7.26 (s, 1H), 7.19 (d,

J = 8.0 Hz, 2H), 7.04 (d, J = 10.8 Hz, 1H), 6.60 (d, J = 3.5 Hz, 1H), 4.28 (d, J = 6.8 Hz, 1H, Leu-

α), 3.05 (t, J = 7.1 Hz, 1H, Leu-β), 2.37 (m, 1H), 2.32 (s, 3H), 0.98 (d, J = 6.6 Hz, 3H), 0.75 (d, J

13 = 6.7 Hz, 3H); C NMR (75 MHz, CDCl3) δ 145.0, 135.4, 135.0, 134.9, 130.0, 129.8, 126.8,

126.6, 124.7, 121.1, 113.7, 108.9, 54.0, 29.1, 21.5, 21.1, 19.7; IR (film) 2961, 2104, 1713, 1370,

-1 + + 1168, 1113 cm ; HRMS (m/z): [M+NH4 ] calcd. for C21H26N5O4S , 444.1700; found 444.1723.

46

Synthesis of Nitro Tryptophan 120. To a solution of (S)-tert-butyl 2-(tert- butoxycarbonylamino)-3-(1H-indol-3-yl)propanoate (118, 62.6 g, 0.167 mol) and glacial acetic acid (50 mL, 0.83 mol) in CH2Cl2 (1000 mL) at 0 °C was added 70% nitric acid (16.5 mL, 0.29 mol) dropwise over a period of 10 min. The resulting solution was stirred at 0 °C for another 20 min. The ice bath was then removed, the light yellow solution was stirred at rt for another 2 h, and then washed with 200 mL of saturated NaHCO3 solution. The organic layer was separated, dried over Na2SO4 and evaporated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) to give an inseparable mixture of nitro regioisomers (17.1 g).

Without further purification, the crude mixture was dissolved in DMF (340 mL) and cooled to 0 °C, NaH (60% dispersion in mineral oil, 2.55 g, 63.8 mmol) was added and the mixture was stirred at 0 °C for 20 min, followed by addition of tosyl chloride (15.1 g, 79.4 mmol). The ice bath was removed and stirring was continued at rt. After 20 min, DMF was evaporated in vacuo to give a slurry, which was triturated with EtOAc (600 mL), washed with water, dried, and concentrated in vacuo. Purification of the residue by flash chromatography on silica gel (gradient 20-30% EtOAc/hexanes) furnished compound 120 as a yellow solid (25.2 g,

23 1 24% over two steps). [α] D -7.9 ⁰; H NMR (400 MHz, CDCl3) δ 8.85 (s, 1H), 8.13 (d, J = 8.7

Hz, 1H), 7.80 (d, J = 8.1 Hz, 2H), 7.66 (s, 1H), 7.64 (d, J = 9.3 Hz, 1H), 7.28 (d, J = 8.0Hz, 2H),

5.13 (d, J = 6.5 Hz, 1H), 4.50 (s, 1H, NH), 3.26 (d, J =14.4 Hz, 1H), 3.16 (d, J = 14.4Hz, 1H),

47

13 2.37 (s, 3H), 1.48 (s, 9H), 1.36 (s, 9H); C NMR (100 MHz, CDCl3) δ 170.4, 155.2, 146.0,

145.4, 136.0, 134.8, 133.8, 130.6, 127.3, 127.2, 127.1, 127.0, 118.5, 117.7, 83.0, 80.3, 28.4, 28.1,

-1 + 21.8; IR (film) 3352, 2977, 2373, 1694, 1322, 1145 cm ; HRMS (m/z): [M+NH4 ] calcd. for

+ C27H37N4O8S , 577.2327, found 577.2333.

Synthesis of Amino Tryptophan 121. Pd/C (10%, 1.0 g) was suspended in a solution of nitro compound 120 (8.80 g, 15.8 mmol) in methanol (650 mL) and the mixture was stirred at rt under H2 (1 atm). After 2.5 h, the reaction mixture was filtered through a pad of Celite, and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel

23 1 (gradient 30-50% EtOAc/hexanes) to afford 121 as a pale solid (7.75 g, 78%). [α] D +8.9 ⁰; H

NMR (400 MHz, CDCl3) δ7.72 (d, J = 8.1 Hz, 2H), 7.29-7.22 (m, 4H), 7.18 (s, 1H), 6.64 (d, J =

8.3 Hz, 1H), 5.08 (d, J = 6.9 Hz, 1H, Trp-α), 4.50 (d, J = 4.4 Hz, 1H), 3.80 (br s, 2H, NH), 3.10

13 (app. s, 2H, Trp-β), 2.36 (s, 3H), 1.46 (s, 9H), 1.36 (s, 9H); C NMR (100 MHz, CDCl3) δ 170.7,

155.1, 144.6, 144.4, 136.5, 135.4, 129.8, 126.7, 123.7, 121.9, 120.4, 117.8, 112.6, 103.3, 99.4,

82.3, 79.8, 54.0, 28.3, 27.9, 21.6; IR (film) 3372, 2981, 2357, 1701, 1370, 1156 cm-1; HRMS

+ + (ESI+) calcd. for C27H36N3O6S [M+H ] 530.2319, found 530.2333.

48

Synthesis of Iodo Tryptophan 122. Over a period of 10 min, a solution of amino compound 121 (4.80 g, 9.1 mmol) in MeCN (10 mL) was added dropwise to a solution of nitrosonium tetrafluoroborate(NOBF4) (1.17 g, 10 mmol) in MeCN (40 mL) at -40 °C. The resulting dark red solution was stirred at -40 °C for 30 min, then warmed to 0 °C and maintained there for 10 min. The mixture was then cooled to -40 ⁰C and a mixture of powdered potassium iodide (3.0 g, 18 mmol) and iodine (2.3 g, 9.0 mmol) was added portionwise over a period of 45 min. After the addition was complete, the reaction mixture was warmed to rt and stirred for another 15 min. The reaction mixture was then diluted with CH2Cl2 (400 mL) and washed with

NaHCO3(aq) and Na2S2O3(aq), dried over Na2SO4 and concentrated in vacuo. The residue was subjected to flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) to furnish

23 1 iodo compound 122 as a pale yellow solid (2.47 g, 42%). [α] D +11 °; H NMR (400 MHz,

CDCl3) δ 8.32 (s, 1H), 7.73 (d, J = 8.0 Hz, 2H), 7.52 (d, J = 8.0 Hz, 1H), 7.31 (s, 1H), 7.26 (m,

3H), 5.08 (d, J = 6.0 Hz), 4.48 (s, 1H), 3.15-3.10 (m, 2H), 2.36 (s, 3H), 1.43 (s, 9H), 1.35 (s, 1H);

13 C NMR (100 MHz, CDCl3) δ 170.4, 155.0, 145.2, 135.8, 135.0, 132.0, 130.5, 130.0, 129.9,

127.6, 126.8, 124.5, 122.4, 121.4, 117.5, 89.0, 83.0, 82.5, 79.9, 54.0, 28.3, 28.1, 27.9, 27.7, 21.6;

-1 + + IR (film) 3376, 2961, 1694, 1362, 1160 cm ; HRMS (ESI+) calcd. for C27H37IN3O6S [M+NH4 ]

658.1442, found 658.1447.

49

Synthesis of Arylated Compound 125. In a sealable 75 mL pressure vessel, iodo compound 122 (4.00 g, 6.25 mmol), leucine derivative 86 (4.84 g, 12.5 mmol, 2 eq), silver acetate (1.56 g, 9.4 mmol, 1.5 eq) and palladium acetate (280 mg, 1.25 mmol, 20 mol%) were suspended in 12 mL of t-BuOH under argon. The reaction vessel was covered with aluminum foil and heated at 110 ⁰C for 36 h. The reaction mixture was then cooled to rt, diluted with dichloromethane, sonicated and filtered through Celite. The filtrate was concentrated and the residue was purified by flash chromatrography on silica gel (gradient 20-40% EtOAc/hexanes) to

23 furnish arylated compound 125 as an off white solid (4.78 g, 85%). [α] D -22.1 ° (c 2.9, CHCl3);

1 H NMR (400 MHz, CDCl3) δ 9.87 (s, 1H, NH), 8.52 (d, 1H, J = 6.9 Hz), 8.13 (s, 1H), 7.97-7.93

(m, 4H), 7.79-7.77 (m, 2H), 7.63 (d, J = 8.0 Hz), 7.54 (d, J = 7.5 Hz, 1H), 7.42 (s, 1H), 7.37-7.19

(m, 3H), 7.19 (m, 1H), 6.71 (d, J = 8.1 Hz), 5.75 (d, J = 12.0 Hz, 1H), 5.10 (d, J = 7.3 Hz, 1H),

4.51 (d, J = 6.4 Hz, 1H), 4.38 (d, J = 12.2 Hz, 1H), 3.19 (s, 2H), 2.06 (br s, 1H), 1.96 (s, 3H),

1.40 (s, 9H), 1.25 (s, 9H), 0.83 (d, J = 5.5 Hz, 3H), 0.74 (d, J = 5.5 Hz, 3H); 13C NMR (100 MHz,

CDCl3) δ 170.6, 168.6, 166.1, 147.8, 144.5, 138.2, 135.4, 135.3, 134.7, 134.3, 131.9, 129.4,

127.5, 126.9, 126.8, 124.4, 123.8, 121.6, 121.4, 119.9, 117.5, 116.8, 110.7, 109.6, 105.8, 82.4,

79.9, 57.8, 54.0, 48.7, 29.5, 28.3, 27.8, 27.6, 21.6, 21.2, 16.1; IR (film) 2961, 2369, 1709, 1366,

-1 + + 1160 cm ; HRMS (ESI+) calcd. for C50H54N5O9S [M+H ] 900.3637, found 900.3635.

50

Synthesis of Azido Compound 126. Compound 125 (3.72 g, 4.14 mmol) was dissolved in 100 mL of n-butanol at rt. Ethylenediamine (2.76 mL, 41 mmol) was added and the resulting solution was stirred at rt for 10 h. The volatiles were removed in vacuo to furnish crude amino compound, which was carried on without further purification for the diazo transfer reaction using the same procedure as for compound 113. The product from the diazo transfer reaction was purified by flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) afforded azido

23 compund 126 as a light yellow foam (2.70 g, 82% over two steps). [α] D +10.4 ° (c 2.22, CHCl3);

1 H NMR (400 MHz, CDCl3) δ 10.31 (s, 1H, NH), 8.67 (s, 1H), 8.65 (s, 1H), 7.96 (s, 1H), 7.70 (d,

J = 7.6 Hz, 2H), 7.51 (s, 1H), 7.50 (d, J = 1.9 Hz, 1H), 7.41 (m, 1H), 7.32 (d, J = 8.2 Hz, 1H),

7.28 (d, J = 9.4 Hz, 2H), 7.14 (d, J = 8.1 Hz, 1H), 7.07 (d, J = 8.1 Hz, 2H), 5.05 (d, J = 6.6 Hz,

1H, Trp-α), 4.55 (d, J = 5.0 Hz, 1H, Leu-α), 4.44 (br s, 1H, NH), 3.28 (app. s, 1H), 3.03 (app. s,

2H), 2.57 (m, 1H), 2.25 (s, 3H), 1.40 (s, 9H), 1.20 (s, 9H), 1.07 (d, J = 5.8 Hz, 3H), 0.78 (d, J =

13 5.8 Hz, 3H); C NMR (100 MHz, CDCl3) δ 170.6, 167.1, 155.1, 148.4, 144.8, 138.6, 136.4,

136.1, 135.2, 135.1, 133.6, 129.7, 127.9, 127.1, 126.8, 124.3, 124.2, 122.2, 121.7, 119.7, 117.6,

116.8, 114.3, 110.7, 108.0, 82.4, 80.0, 68.0, 55.1, 53.9, 29.0, 28.3, 27.7, 21.5, 21.4, 20.2; IR

-1 + (film) 3321, 2969, 2108, 1694, 1520, 1362, 1156 cm ; HRMS (ESI+) cald. for C42H52N5O7S

[M+H+] 796.3488, found 796.3480.

51

Synthesis of bis-Boc Compound 127. Azido compound 126 (520 mg, 0.65 mmol) and

Boc anhydride (2.85 g, 13.1 mmol) were dissolved in 0.3 mL of MeCN. DMAP (240 mg, 1.9 mmol) was added and the resulting solution was stirred vigorously and heated at 70 °C for 15 min. The volatile organics were removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) to furnish the bis-Boc

23 1 compound 127 as a yellow foam (582 mg, 89%). [α] D +27.8 ° (c 5.08, CHCl3); H NMR (400

MHz, CDCl3) δ 8.06 (d, J = 8.0 Hz, 1H), 7.89 (s, 1H), 7.73 (d, J = 7.9 Hz, 2H), 7.63 (d, J = 7.8

Hz, 1H), 7.45 (s, 1H), 7.41 (s, 1H), 7.26 (m, 1H), 7.12 (d, J = 8.2 Hz, 1H), 7.08 (d, J = 7.8 Hz,

2H), 5.78 (d, J = 10.1 Hz, 1H, Trp-α), 5.11 (m, 1H), 3.45 (app. d, J = 14.8 Hz, 1H), 3.33 (m, 2H),

2.45 (m, 1H), 2.24 (s, 3H), 1.48 (s, 9H), 1.36 (s, 18H), 1.24 (s, 9H), 0.88 (app. s, 6H); 13C NMR

(100 MHz, CDCl3) δ 169.3, 152.7, 152.1, 150.3, 144.5, 143.9, 136.0, 135.6, 135.1, 134.9, 130.0,

129.9, 128.5, 128.0, 127.7, 126.9, 125.9, 124.6, 121.4, 118.9, 118.8, 114.9, 109.6, 83.4, 82.9,

81.7, 63.0, 58.5, 53.4, 28.9, 27.9, 27.8, 27.5, 25.0, 21.5, 18.0; IR (film) 2969, 2361, 2100, 1729,

-1 + + 1362, 1145 cm ; HRMS (ESI+) calcd. for C52H66N7O11S [M+H ] 996.4536, found 996.4531.

52

Synthesis of Azido Acid 128. To a solution of bis-Boc compound 127 (2.29 g, 2.3 mmol) in THF/H2O (4:1, 24 mL) at 0 °C was added hydrogen peroxide (30%, 1.26 mL, 11.5 mmol) and lithium hydroxide (96 mg, 2.3 mmol). The resultant emulsion was warmed to rt and stirred for 3 h, diluted with water (150 mL), treated with Na2SO3, acidified to pH 2 with 0.6 M HCl, and extracted three times with EtOAc. The combined organic layers were washed with brine, dried over Na2SO4 and concentrated in vacuo. The residue was purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes, then 5% MeOH/ CH2Cl2) to yield azido acid 18 as a

23 1 white solid (1.72 g, 97%). [α] D -29.3 ° (c 2.22, CHCl3); H NMR (400 MHz, CDCl3) δ 7.82 (s,

1H), 7.72 (d, J = 10.9 Hz, 2H), 7.40 (d, J = 10.8 Hz, 1H), 7.19 (d, J = 10.8 Hz, 2H), 7.06 (d, J =

10.8 Hz, 1H), 5.06 (dd, J = 13.2, 6.8 Hz, 1H, Trp-α), 4.24 (d, J = 7.9 Hz, 1H, Leu-α), 3.34 (m,

2H, Trp-β), 3.06 (t, J = 8.8 Hz, 1H, Leu-β), 2.38 (m, 1H), 2.31 (s, 3H), 1.54 (s, 9H), 1.33 (s,

13 18H), 0.94 (d, J = 8.6 Hz, 3H), 0.70 (d, J = 8.8 Hz, 3H); C NMR (100 MHz, CDCl3)δ 174.0,

169.2, 152.0, 144.8, 135.8, 135.2, 135.0, 130.2, 129.8, 129.7, 126.9, 126.4, 124.9, 124.6, 119.3,

118.8, 113.5, 82.9, 81.8, 64.9, 58.3, 53.7, 29.1, 27.9, 27.8, 25.0, 21.5, 21.0, 19.7; IR (film) 2973,

-1 + + 2108, 1733, 1367, 1133 cm ; HRMS (ESI+) calcd. for C38H55N6O10S [M+NH4 ] 787.3695, found 787.3716.

53

Synthesis of Ester 129. A solution of azido acid 128 (1.68 g, 2.18 mmol) in dichloromethane (30 mL) was treated with N-hydroxysuccinimide (0.33 g, 2.87 mmol) and N,N’- dicyclohexylcarbodiimide (DCC, 0.50 g, 2.43 mmol). The mixture was stirred at rt for 1 h. The dichloromethane was evaporated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes) to furnish ester 129 as a pale

23 1 solid (1.75 g, 92%). [α] D -81⁰(c 2.10, CHCl3); H NMR (400 MHz, CDCl3) δ 7.84 (s, 1H), 7.74

(d, J = 8.2 Hz, 2H), 7.44 (d, J = 8.1 Hz, 1H), 7.37 (s, 1H), 7.19 (d, J = 8.1 Hz, 2H), 7.15 (d, J =

8.4 Hz, 1H), 5.08 (dd, J = 8.8, 5.5 Hz, 1H, Trp-α), 4.57 (d, J = 5.2 Hz, 1H, Leu-α), 3.35 (m, 2H,

Trp-β), 3.12 (dd, J = 8.8, 5.5Hz, 1H, Leu-β), 2.83 (s, 4H), 2.45 (m, 1H), 2.31 (s, 3H), 1.47 (s,

9H), 1.35 (s, 18H), 1.07 (d, J = 6.5 Hz, 3H), 0.72 (d, J = 6.6 Hz, 3H); 13C NMR (100 MHz,

CDCl3)δ 169.2, 168.2, 152.0, 144.6, 135.3, 135.2, 135.1, 130.5, 129.8, 126.9, 125.0, 124.1, 119.6,

118.8, 114.1, 82.8, 81.7, 63.4, 58.3, 54.9, 34.0, 29.0, 28.0, 27.8, 25.6, 25.0, 24.9, 21.5, 21.1, 20.4;

-1 + IR (film) 2973, 2108, 1733, 1366, 1137 cm ; HRMS (ESI+) calcd. for C42H58N7O12S

[M+NH4+] 884.3859, found 884.3863.

54

Synthesis of Acid 130. A solution of ester 129 (1.69 g, 1.95 mmol) and NH2ValLeuOH

(0.95 g, 4.1 mmol) in 100 mL of DMF was treated with aqueous NaHCO3 (0.2 M, 20 mL) and the mixture was stirred at rt for 16 h. The solution was then acidified to pH 3 with 0.5 M HCl, diluted with water, and extracted with EtOAc-ether (1:1, 3 X 150 mL). The organic layers were combined, washed with water (2 X 100 mL), dried over Na2SO4 and evaporated to dryness.

Purification of the residue by flash chromatography on silica gel (gradient 20-50%

23 EtOAc/hexanes, then EtOAc) afforded acid 130 as an off white solid (1.61 g, 84%). [α] D -20 °

1 (c 0.2, MeOH); H NMR (300 MHz, CDCl3) δ 7.77 (s, 1H), 7.73 (d, J = 8.2 Hz, 2H), 7.41 (s, 1H),

7.38 (s, 1H), 7.20 (d, J = 8.1 Hz, 2H), 7.04 (d, J = 8.0 Hz, 1H), 6.88 (d, J = 8.4 Hz, 1H), 6.59 (d,

J = 8.0 Hz, 1H), 5.06 (dd, J = 9.7, 5.5 Hz, 1H, Trp-α ), 4.47 (dd, J = 8.4, 4.7 Hz, 1H), 4.39 (d, J

= 5.1 Hz, 1H, N3-α), 4.33 ( m, 1H), 3.40 (dd, J = 15.1, 4.8 Hz, 1H), 3.26 (m, 1H), 2.98 (m, 1H),

2.49 (m, 1H), 2.31 (s, 3H), 2.19 (m, 1H), 1.47 (s, 9H), 1.35 (s, 18H), 1.15 (m, 3H), 1.14 (d, J =

13 6.1 Hz, 3H), 1.04 (d, J = 6.4 Hz, 6H), 0.90 (d, J = 6.6 Hz, 9H); C NMR (75 MHz, CDCl3) δ

174.3, 171.6, 169.2, 168.6, 152.0, 114.8, 135.6, 135.1, 135.0, 130.1, 129.8, 126.9, 124.7, 124.2,

119.3, 118.9, 114.0, 82.9, 81.8, 66.6, 58.4, 57.1, 55.1, 51.8, 40.3, 31.0, 28.8, 27.9, 27.8, 24.9,

24.3, 22.6, 21.9, 21.5, 21.4, 20.5, 18.9, 17.6; IR (film) 3328, 3265, 2977, 2100, 1745, 1374, 1160

-1 + + cm ; HRMS (ESI+) calcd. for C49H75N8O12S [M+NH4 ] 999.5220, found 999.5217.

55

Synthesis of Amino Acid 131. Bis-Boc compound 130 (1.51 g, 1.54 mmol) was treated with HCl in dioxane (4 M, 30 mL) at 0 °C, and the mixture was stirred for 3 h. Dioxane was evaporated in vacuo to give a residue which was purified by flash chromatography on silica gel

(gradient 2-15% MeOH/ CH2Cl2) to afford amino acid 131 as its hydrochloride salt (1.02 g,

23 1 85%). [α] D -29.4 ° (c 0.17, 1:1 MeOH/CHCl3); H NMR (300 MHz, CD3OD) δ 7.84 (d, J = 8.3

Hz, 2H), 7.79 (s, 1H), 7.60 (s, 1H), 7.52 (d, 8.2 Hz, 1H), 7.31 (d, J = 8.1 Hz, 2H), 7.13 (d, J = 8.2

Hz, 1H), 4.35 (d, J = 9.6 Hz, 1H), 4.28 (m, 2H), 3.92 (d, J = 5.7 Hz, 1H, N3-α), 3.30 (m, 3H),

3.27 (m, 2H), 3.20 (dd, J = 9.9, 5.0 Hz, 1H), 2.38 (m, 1H), 2.34 (s, 3H), 1.92 (m, 1H), 1.42 (m,

13 3H), 1.29 (s, 9H), 0.87-0.69 (m, 18H); C NMR (100 MHz, CD3OD) δ 175.3, 173.5, 170.7,

169.2, 147.0, 136.4, 136.3, 136.0, 131.2, 130.7, 128.3, 126.6, 125.5, 120.2, 116.7, 116.5, 85.3,

66.3, 59.6, 54.1, 53.8, 53.0, 41.8, 31.7, 29.7, 28.0, 27.3, 25.6, 23.4, 22.0, 21.5, 19.4, 18.4, 18.2;

-1 IR (film) 3328, 3285, 2953, 2353, 2088, 1737, 1666, 1504, 1366, 1160 cm ; HRMS (ESI+) calcd.

+ + for C39H56N7O8S [M+H ] 782.3906, found 782.3922.

56

Synthesis of Macrocyclic Peptide 132. A solution of amino acid 131 (300 mg, 0.36 mmol) in DMF (180 mL) was treated with HOOBt (90 mg, 0.55 mmol) and EDC (132 µL, 0.74 mmol), and the resulting yellow solution was stirred at rt for 12 h. Water (350 mL) was added and the mixture was cooled to 0 °C. The white precipitate which formed was collected, washed with ice cold water and dried. This solid was further purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes) to afford cyclic peptide 132 as a white solid (225 mg,

23 1 82%). [α] D -23 ° (c 0.26, MeOH); H NMR (400 MHz, CD3OD) δ 7.68 (d, J = 7.5 Hz, 2H), 7.67

(s, 1H), 7.49 (d, J = 7.9 Hz, 1H), 7.36 (s, 1H), 7.27 (d, J = 7.6 Hz, 2H), 7.04 (d, J = 7.9 Hz, 1H),

5.33 (app. s, 1H), 4.21 (t, J = 6.9 Hz, 1H), 3.90 (d, J = 11.9 Hz, 1H), 3.55 (d, J = 6.28 Hz, 1H),

3.39 (m, 1H), 3.16 (m, 2H), 2.35 (s, 3H), 2.34 (m, 1H), 2.01 (m, 1H), 1.59 (m, 1H), 1.52 (s, 9H),

13 1.43 (m, 2H), 0.97-0.85 (m, 18H); C NMR (75 MHz, CD3OD) δ 173.5, 172.2, 171.8, 170.6,

147.0, 137.3, 136.6, 135.4, 131.4, 130.9, 128.2, 127.2, 123.4, 122.2, 119.5, 118.7, 83.8, 66.6,

59.9, 54.3, 53.9, 51.6, 44.1, 32.4, 31.2, 30.3, 28.7, 28.5, 26.4, 23.8, 22.4, 22.1, 22.0, 19.5, 19.0,

18.7; IR (film) 3325, 2961, 2404, 2282, 2096, 1654, 1516, 1362, 1160 cm-1; HRMS (ESI+)

+ + C39H54N7O7S [M+H ] 764.3800, found 764.3804.

57

Synthesis of Amine 133. To a solution of azide 132 (200 mg, 0.26 mmol) and glacial

HOAc (63 µL, 1.05 mmol) in EtOAc (40 mL) was added Pd/C (10%, 100 mg) at rt. The reaction vessel was evacuated and back filled with hydrogen gas from a balloon (3 times) and the mixture was stirred under H2 at rt for 6 h. The reaction mixture was then filtered through Celite. The filtrate was concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (30% EtOAc/hexanes, then gradient 2-5% MeOH/ CH2Cl2) to furnish amine 133 as

23 1 an off white solid (163 mg, 85%). [α] D -47.5 ° (c 1.01, MeOH); H NMR (300 MHz, CD3OD) δ

7.66 (d, J = 7.6 Hz, 2H), 7.61 (s, 1H), 7.44 (d, J = 8.4 Hz, 1H), 7.33 (s, 1H), 7.26 (d, J = 7.7 Hz,

2H), 7.05 (J = 8.2 Hz, 1H), 5.28 (app. t, J = 8.4, 7.8 Hz, 1H), 4.23 (t, J = 7.7, 7.4 Hz, 1H), 3.62

(m, 2H), 3.41 (dd, J = 15.0, 6.0 Hz, 1H), 3.14 (dd, J = 15.3, 10.7 Hz, 1H), 2.90 (d, J = 10.9 Hz,

1H), 2.46 (m, 1H), 2.34 (s,3H), 2.01 (m, 1H), 1.59 (m, 1H), 1.52 (s, 9H), 1.42 (t, J = 7.4 Hz, 2H),

13 0.95-0.85 (m, 15H), 0.75 (d, J = 6.5 Hz, 3H); C NMR (75 MHz, CD3OD) δ 175.8, 173.6, 172.2,

172.1, 147.0, 137.1, 136.6, 131.4, 130.8, 128.2, 126.9, 124.1, 121.5, 119.5, 118.5, 83.8, 59.9,

59.7, 57.1, 54.2, 51.9, 44.8, 32.6, 28.7, 28.4, 26.1, 23.5, 22.8, 21.9, 19.5, 19.3, 17.5; IR (film)

-1 + + 2973, 2941, 2361, 1721, 1366, 1216, 1014 cm ; HRMS (ESI+) calcd. for C39H56N5O7S [M+H ]

738.3895, found 738.3911.

58

Synthesis of Pentapeptide 134. Amine 133 (150 mg, 0.20 mmol), and pyroglutamic acid

(39 mg, 0.30 mmol) were dissolved in 15 mL of DMF, followed by treatment with HOOBt (54 mg, 0.30 mmol) and EDC (57 µL, 0.32 mmol). The reaction mixture was stirred at rt for 12 h, then partitioned between EtOAc/ether (1:1, 250 mL) and water (100 mL). The aqueous layer was extracted with EtOAc/ether (1:1, 100 mL). The organic layers were combined, dried over

Na2SO4 and concentrated in vacuo to give a residue which was purified by flash chromatography

23 1 to afford compound 134 as an off white solid (171 mg, 99%). [α] D -139 ⁰(c 0.67, MeOH); H

NMR (300 MHz, CD3OD) δ 7.68 (d, J = 3.3 Hz, 2H), 7.64 (s, 1H), 7.50 (d, J = 8.1 Hz, 1H), 7.34

(s, 1H), 7.25 (d, J = 8.0 Hz, 2H), 7.16 (d, J = 8.2 Hz, 1H), 5.30 (dd, J = 9.6, 5.5 Hz, 1H), 4.28

(dd, J = 8.5, 3.6 Hz, 1H), 4.20 (dd, J = 10.6, 3.9 Hz, 1H), 3.64 (d, J = 7.5 Hz, 1H), 3.41 (dd, J =

15.8, 6.4 Hz, 1H), 3.22 (dd, J = 12.5, 4.0 Hz, 1H), 2.44 (m, 1H), 2.33 (s, 3H), 2.23 (m, 1H), 1.98

13 (m, 2H), 1.52 (s, 9H), 1.46-1.28 (m, 3H), 0.92-0.77 (m, 18H); C NMR (75 MHz, CD3OD) δ

175.0, 174.0, 172.5, 172.2, 172.1, 147.0, 137.2, 136.6, 135.7, 131.4, 128.2, 127.0, 124.2, 121.7,

119.7, 118.9, 83.9, 58.1, 57.7, 54.3, 52.0, 44.3, 32.7, 30.8, 29.3, 27.4, 26.0, 24.0, 22.6, 22.0, 19.5,

19.3, 18.1; IR (film) 3285, 2965, 1642, 1516, 1374, 1160 cm-1; HRMS (ESI+) calcd. for

+ + C44H61N6O9S [M+H ] 849.4215, found 849.4213.

59

Synthesis of Free NH Indole 134a. To a solution of compound 134 (171 mg, 0.20 mmol) in methanol (30 mL) was added magnesium turnings (1.35 g, 56 mmol) and the mixture was sonicated at rt for 2 h. The reaction was then quenched by addition of saturated aqueous NH4Cl

(60 mL) and water (30 mL), and extracted with EtOAc (4X50 mL). The organic layers were dried over Na2SO4, and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (5% MeOH/ CH2Cl2 CH2Cl2) to afford the indole 134a as a white

23 1 solid (121 mg, 87%). [α] D -65⁰(c 0.17, MeOH); H NMR (300 MHz, CD3OD) δ 7.48 (d, J = 8.4

Hz, 1H), 6.98 (m, 3H), 5.28 (dd, J = 9.6, 5.5 Hz, 1H), 4.26 (m, 2H), 3.84 (d, J = 7.1, 1H), 3.45

(dd, J = 14.8, 5.7 Hz, 1H), 3.17 (dd, J = 14.4, 9.0 Hz, 1H), 3.10 (dd, J = 11.7, 3.8 Hz, 1H), 2.49-

2.20 (m, 4H), 2.07 (m, 1H), 1.96 (m, 1H), 1.50 (s, 9H), 1.45-1.28 (m, 3H), 0.98 (d, J = 6.8 Hz,

3H), 0.91 ( d, J = 5.1 Hz, 6H), 0.86 (d, J =6.1 Hz, 3H), 0.84 (d, J = 6.4 Hz, 3H), 0.80 (d, J = 6.4

13 Hz, 3H); C NMR (75 MHz, CD3OD) δ 181.9, 174.9, 174.4, 173.1, 172.7, 172.2, 138.7, 131.7,

127.6, 125.7, 120.2, 119.7, 116.4, 110.5, 83.5, 60.1, 58.1, 58.0, 54.2, 54.0, 53.4, 44.6, 32.8, 30.8,

29.5, 29.2, 28.7, 27.4, 26.0, 24.0, 22.6, 22.0, 19.5, 19.2, 18.5; IR (film) 3561, 2961, 1686, 1504,

-1 + + 1160 cm ; HRMS (ESI+) calcd. for C37H55N6O7 [M+H ] 695.4127, found 695.4104.

60

Synthesis of Acid 135. A solution of indole 134a (120 mg, 0.172 mmol), trifluoroacetic acid (5 mL), triisopropylsilane (TIPS, 150 µL) and water (100 µL) were stirred at rt for 1.5 h.

The volatiles were removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 5-20% MeOH/ CH2Cl2) to afford acid 135 as its TFA salt

23 1 (white solid, 143 mg, 96%). [α] D -78⁰ (c 0.09, MeOH); H NMR (300 MHz, CD3OD ) δ 8.53 (d,

J = 8.7 Hz, 1H), 8.07 (d, J = 9.7 Hz, 1H), 7.87 (d, J = 8.9 Hz, 1H), 7.52 (d, J = 8.3 Hz, 1H), 6.98

(m, 3H), 6.73 (d, J = 8.0 Hz, 1H), 5.38 (m, 1H), 4.30-4.26 (m, 2H), 3.88 (t, J = 7.6 Hz, 1H),

3.55 (dd, J = 15.0. 5.5 Hz, 1H), 3.21 (dd, J = 14.9, 9.4 Hz, 1H), 3.10 (dd, J = 11.8, 4.0 Hz, 1H),

2.49-2.20 (m, 4H), 2.12-1.93 (m, 2H), 1.48-1.36 (m, 3H), 0.97 (d, J = 6.8 Hz, 3H), 0.92-0.83 (m,

13 12H), 0.80 (d, J = 6.3 Hz, 3H); C NMR (75 MHz, CD3OD) δ 182.0, 175.3, 175.0, 174.4, 173.2,

173.1, 172.3, 138.6, 131.6, 127.8, 125.6, 120.2, 119.7, 116.4, 110.6, 60.0, 58.1, 54.2, 54.0, 52.8,

44.6, 32.8, 30.8, 29.5, 29.0, 27.4, 26.0, 24.0, 22.6, 22.0, 19.5, 19.1, 18.5; IR (film) 3317, 2945,

-1 + + 2357, 1630, 1516, 1354, 1216 cm ; HRMS (ESI+) calcd. for C33H47N6O7 [M+H ] 639.3501, found 639.3515.

61

Synthesis of Methyl Ester 136. A solution of acid 135 (5.0 mg, 0.0058 mmol) in methanol (0.5 mL) was treated with SOCl2 (15 µL) at 0 °C for 2 h. The reaction was then quenched by addition of aqueous NaHCO3 and extracted with CHCl3 (3X3 mL). Organic layers were combined, dried and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 2-5% MeOH/CH2Cl2) to afford methyl ester 136 as a white solid (3.1 mg, 82%). 1H NMR (400 MHz, DMSO-d6) δ 10.73 (s, 1H), 8.58 (d, J = 8.8 Hz,

1H), 8.42 (d, J = 9.3 Hz, 1H), 7.88 (s, 1H), 7.38 (d, J = 7.3 Hz, 2H), 7.01 (s, 1H), 6.98 (d, J = 8.0

Hz, 2H), 6.91 (s, 1H), 6.84 (d, J = 7.7 Hz, 1H), 5.25 (m, 1H), 4.87 (t, J = 10.5 Hz, 1H), 4.12 (d, J

= 8.5 Hz, 1H), 4.05 (m, 1H), 3.89 (dd, J = 7.1, 6.6Hz, 1H), 3.67 (s, 3H), 3.28 (d, J = 14.5 Hz,

1H), 3.16 (dd, J = 14.4, 8.0 Hz, 1H), 2.99 (d, J = 11.2 Hz, 1H), 2.29-2.07 (m, 4H), 1.69 (m, 1H),

1.38 (m, 1H), 1.25-1.18 (m, 2H), 0.93-0.86 (m, 9H), 0.80-0.77 (app t, J = 6.2 Hz, 6H), 0.71 (d, J

= 5.8 Hz, 3H); 13C NMR (100 MHz, DMSO-d6) δ178.0, 172.8, 172.6, 172.2, 170.4, 170.3, 136.5,

130.2, 126.0, 124.0, 119.5, 118.2, 114.9, 108.8, 79.6 (CF3), 58.6, 55.5, 52.5, 52.3, 52.1, 51.7,

43.3, 31.1, 29.4, 27.5, 27.2, 26.0, 24.2, 23.7, 22.3, 21.3, 19.1, 18.1, 17.7; HRMS (ESI+) calcd.

+ + for C34H49N6O7 [M+H ] 653.3657, found 653.3671.

62

Synthesis of Compound 137. At 0 °C, a solution of acid 137 (63 mg, 0.073 mmol) and

L-proline benzyl ester hydrochloride (18 mg, 0.073 mmol) in THF (7 mL) was treated with

DIPEA (38 µL, 0.218 mmol), HOBt (15 mg, 0.11 mmol) and EDCI (21 mg, 0.11 mmol). After stirring at rt for 12 h, the reaction mixture was washed with NaHCO3 (sat. aq., 18 mL) and extracted with dichloromethane (3 X 18 mL). The organic layers were combined, dried over

Na2SO4, and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 2-10% MeOH/ CH2Cl2) to afford compound 137 as a

23 1 beige solid (45 mg, 75%). [α] D -53 ° (c 0.34, CH2Cl2); H NMR (400 MHz, CD3OD) δ 7.51 (d,

J = 8.3 Hz, 1H), 7.41-7.33 (m, 5H), 6.99 (m, 3H), 5.58 (m, 1H), 5.23 (d, J = 12.2 Hz, 1H), 5.16

(d, J = 12.2 Hz, 1H), 4.85 (m, 1H), 4.62 (m, 1H), 4.29-4.26 (m, 2H), 4.03 (d, J = 6.4 Hz, 1H),

3.85 (m, 2H), 3.38-3.42 (m, 1H), 3.08 (app d, J = 11.9 Hz, 2H), 2.58-2.14 (m, 5H), 2.12-1.88 (m,

4H), 1.57-1.21 (m, 3H), 0.98 (d, J = 6.8 Hz, 3H), 0.92-0.79 (m, 15H); IR (film) 3285, 3068, 2914,

-1 + + 1769, 1694, 1528, 1374, 1164, 1089 cm ; HRMS (ESI+) calcd. for C45H60N7O8 [M+H ]

826.4498, found 826.4495.

63

Synthesis of Celogentin C (4). From compound 137, celogentin C (4) was prepared according to the reported procedure.13 The final product was purified by semipreparative HPLC

(Agilent 1100, C18 60x250 Zorbax column, 15-50% CH3CN/H2O with 0.1% TFA, over 30 min, 4 mL/min). Celogentin C (4) trifluoroacetic acid salt: white solid, 1H NMR (600 MHz, DMSO-d6)

δ 11.56 (s, 1H), 8.77(d, J = 9.6 Hz, 1H), 8.52 (d, J = 9.6 Hz, 1H), 8.33 (d, J = 10.2 Hz, 1H), 8.29

( d, J = 8.4 Hz, 1H), 7.86 (s, 1H), 7.54 (d, J = 8.4, 2H), 7.50 (br s, 1H), 7.40 (t, J = 5.4 Hz, 1H),

6.98 (d, J = 8.4 Hz, 2H), 6.89 (s, 1H), 6.72 (d, J= 6.0 Hz, 1H), 5.70-5.65 (m, 1H), 4.91 (td, J =

11.4, 1.8 Hz, 1H), 4.83 (t, J = 10.5 Hz, 1H), 4.20-4.14 (m, 2H), 4.10 (dd, J = 8.7, 3.6 Hz, 1H),

4.04-3.95 (m, 2H), 3.88-3.77 (m, 1H), 3.61 (t, J = 7.8 Hz, 1H), 3.29(d, J = 16.2 Hz, 1H), 3.24 (dd,

J = 15.0, 4.8 Hz, 1H), 3.10-3.05 (m, 3H), 2.85 (dd, J = 16.2, 12.0 Hz, 1H), 2.61 (t, J = 13.8, 1.8

Hz, 1H), 2.28-2.21 (m, 2H), 2.16-2.13 (m, 1H), 2.13-2.07 (m, 2H), 2.06-2.00 (m, 2H), 1.87-1.83

(m, 1H), 1.80-1.75 (m, 2H), 1.73-1.62 (m, 2H), 1.50-1.39 (m, 4H), 1.17 (td, J = 11.4, 3.6 Hz,

1H), 0.87 (d, J = 6.6 Hz, 3H), 0.82 (d, J = 6.6 Hz, 3H), 0.77 (d, J = 6.6 Hz, 3H), 0.75-0.71 (m,

9H); 13C NMR (150 MHz, DMSO-d6) δ 177.5, 172.2, 171.5, 171.4, 171.3, 171.0, 169.4, 169.2,

158.0, 157.8, 156.6, 137.2, 132.8, 127.3, 125.0, 119.6, 114.1, 102.9, 61.6, 57.3, 55.1, 54.7, 52.2,

52.1, 51.3, 49.9, 47.2, 46.9, 41.5, 40.6, 31.2, 29.8, 29.0, 26.6, 25.6, 25.1, 24.0, 23.9, 23.1, 21.9,

+ + 20.9, 18.7, 18.3, 17.0; HRMS (ESI+) calcd. for C50H71N14O10 [M+H ] 1027.5472, found

1027.5474.

64

References

(1) Leung, T. W. C.; Williams, D. H.; Barna, J. C. J.; Foti, S.; Oelrichs, P. B.

Tetrahedron 1986, 42, 3333.

(2) Kobayashi, J.; Suzuki, H.; Shimbo, K.; Takeya, K.; Morita, H. J. Org. Chem.

2001, 66, 6626.

(3) Suzuki, H.; Morita, H.; Iwasaki, S.; Kobayashi, J. Tetrahedron 2003, 59, 5307.

(4) Suzuki, H.; Morita, H.; Shiro, M.; Kobayashi, J. Tetrahedron 2004, 60, 2489.

(5) Kahn, S. D.; Booth, P. M.; Waltho, J. P.; Williams, D. H. J. Org. Chem. 2000, 65,

8406.

(6) Morita, H.; Shimbo, K.; Shigemori, H.; Kobayashi, J. Bioorg. Med. Chem. Lett.

2000, 10, 469.

(7) Harrison, J. R.; Moody, C. J. Tetrahedron Lett. 2003, 44, 5189.

(8) He, L. W.; Yang, L. P.; Castle, S. L. Org. Lett. 2006, 8, 1165.

(9) Castle, S. L.; Srikanth, G. S. C. Org. Lett. 2003, 5, 3611.

(10) Bentley, D. J.; Moody, C. J. Org. Biomol. Chem. 2004, 2, 3545.

(11) Bentley, D. J.; Slawin, A. M. Z.; Moody, C. J. Org. Lett. 2006, 8, 1975.

(12) Hoppe, D.; Follmann, R. Chem. Ber. 1976, 109, 3062.

(13) Ma, B.; Litvinov, D. N.; He, L. W.; Banerjee, B.; Castle, S. L. Angew. Chem., Int.

Ed. 2009, 48, 6104.

(14) Banerjee, B.; Capps, S. G.; Kang, J.; Robinson, J. W.; Castle, S. L. J. Org. Chem.

2008, 73, 8973.

(15) Feng, Y. Q.; Chen, G. Angew. Chem., Int. Ed. 2010, 49, 958.

(16) Zaitsev, V. G.; Shabashov, D.; Daugulis, O. J. Am. Chem. Soc. 2005, 127, 13154.

65

(17) For reviews on C-H functionalizations, see: (a) Crabtree, R. H. Chem. Rev. 2010,

110, 575. (b) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147. (c) Wencel-

Delord, J.; Droge, T.; Liu, F.; Glorius, F. Chem. Soc. Rev. 2011, 40, 4740. (d) Arockiam, P. B.;

Bruneau, C.; Dixneuf, P. H. Chem. Rev. 2012, 112, 5879. (e) Chen, D. Y. K.; Youn, S. W.

Chem. Eur. J. 2012, 18, 9452. (f) Yamaguchi, J.; Yamaguchi, A. D.; Itami, K. Angew.

Chem., Int. Ed. 2012, 51, 8960. (g) Lapointe, D.; Fagnou, K. Chem. Lett. 2010, 39, 1119.

(18) For reviews on the application of C-H functionalization in total synthesis, see: (a)

Gutekunst, W. R.; Baran, P. S. Chem. Soc. Rev. 2011, 40, 1976. (b) McMurray, L.; O'Hara, F.;

Gaunt, M. J. Chem. Soc. Rev. 2011, 40, 1885.

(19) For selected examples of recent development of C-H functionalization methodologies, see: (a) Feng, Y. Q.; Wang, Y. J.; Landgraf, B.; Liu, S.; Chen, G. Org. Lett. 2010,

12, 3414. (b) Shabashov, D.; Daugulis, O. J. Am. Chem. Soc. 2010, 132, 3965. (c) Dai,

H. X.; Stepan, A. F.; Plummer, M. S.; Zhang, Y. H.; Yu, J. Q. J. Am. Chem. Soc. 2011, 133,

7222. (d) He, G.; Chen, G. Angew. Chem., Int. Ed. 2011, 50, 5192. (e) He, G.; Zhao, Y.

S.; Zhang, S. Y.; Lu, C. X.; Chen, G. J. Am. Chem. Soc. 2012, 134, 3. (f) Zhang, S. Y.; He,

G.; Zhao, Y. S.; Wright, K.; Nack, W. A.; Chen, G. J. Am. Chem. Soc. 2012, 134, 7313. (g)

Zhang, S. Y.; He, G.; Nack, W. A.; Zhao, Y. S.; Li, Q.; Chen, G. J. Am. Chem. Soc. 2013, 135,

2124. (h) Zhang, S. Y.; Li, Q.; He, G.; Nack, W. A.; Chen, G. J. Am. Chem. Soc. 2013, 135,

12135.

(20) Reddy, B. V. S.; Reddy, L. R.; Corey, E. J. Org. Lett. 2006, 8, 3391.

(21) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. J. Org. Chem. 1983, 48, 2424.

(22) Nyffeler, P. T.; Liang, C. H.; Koeller, K. M.; Wong, C. H. J. Am. Chem. Soc.

2002, 124, 10773.

66

(23) Michaux, J.; Retailleau, P.; Campagne, J. M. Synlett 2008, 1532.

(24) We thank Professor Castle for providing a sample of synthetic celogentin C for comparison.

67

Part II. Divergent Total Syntheses of the Unusual Monoterpenoid Indole

Alkaloids Alstilobanines A, E and Angustilodine

Chapter 1. Introduction and Background

1.1. Monoterpenoid Indole Alkaloids: Background

Monoterpenoid indole alkaloids (MIAs) comprise a class of secondary metabolites derived from higher plants that are particularly abundant in members of the Apocynaceae,

Loganiaceae, and Rubiaceae families. To date, over 2,000 MIAs have been isolated, making this group the largest known family of nitrogen-containing secondary metabolites1.

Monoterpenoid indole alkaloids feature basic skeletons in which a C9 or C10 terpene fragment is affixed to a tryptamine nucleus. Based on the connectivities of their terpene segments, MIAs are generally classified into three different categories: Corynanthe, Iboga and

Aspidosperma, which derive their names from their respective host-plant genera (Figure 1).

Figure 1. The Corynanthe, Iboga and Aspidosperma backbones in monoterpenoid indole alkaloids

Despite the structural variability among the three MIA subclasses, these molecules all likely originate from the same central biosynthetic intermediate, strictosidine (4). Higher plants biosynthesize strictosidine (4) from the condensation of tryptamine (1) and the monoterpene precursor secologanin (2) via an enzymatic Pictet-Spengler-type electrophilic aromatic

68 substitution catalyzed by strictosidine synthase (STR) (Scheme 1). The glucose moiety in 4 is subsequently cleaved by β-glucosidase to afford intermediate 5, which then undergoes a series of structural modifications via reduction, oxidation and/or rearrangement to produce the three main subclasses of MIAs. The terpenoid moiety in the Corynanthe scafold is derived from its biogenetic precursor, secologanin (2). The Iboga and Aspidosperma frameworks could then result from rearrangement of the Corynanthe scafold, likely traversing the common biosynthetic intermediate stemmadenine (6).2 The tryptamine aminoethyl side chain is rarely modified in the biosynthesis of MIAs. Thus, the different skeletal features of each subclass mostly originate from structural modifications of the terpenoid moiety.

Scheme 1. Biogenetic pathway for in vivo synthesis of monoterpenoid indole alkaloids

69

The great structural diversity of the MIAs is complemented by a wide spectrum of substantial biological activity including antihypertensive, antiinflammatory, and antitumor properties. Many members of the MIA family have been used in the medical realm as important therapeutics to treat a variety of human diseases.

Figure 2. Representative monoterpenoid indole alkaloids used as therapeutic agents

3 Yohimbine (7), a pre- and post-synaptic α2-receptor inhibitor, has enjoyed commercial success as a prescription drug for the treatment of male impotence. Reserpine (8) is clinically used to relieve hypotensive symptoms.4 More recently, vinblastine (9) and other bis- monoterpenoid indole alkaloids isolated from Madagasgar periwinkle were found to possess high inhibitory activity against tumor cells.5 Bis-monoterpene indole alkaloid 9 is now widely used as a chemotherapeutic agent against non-small-cell lung cancer, breast cancer, Hodgkin’s lymphoma and other types of white blood cell cancers, functioning via an antimicrotubule mechanism.

70

Research on MIAs can trace its origins back almost 200 years to the milestone isolation of strychnine (10) by Pelletier and Caventou.6 Despite the apparent maturity of the field, the search for novel MIAs is still ongoing, bolstered by the growing demand for novel nature- derived medical treatments. With the aid of modern analytical techniques, many new MIAs are being discovered at an unprecedented pace every year. Some of these novel MIAs have shown promise as potential therapeutic agents. For instance, calophyline A (11), isolated in 2012, 7 is a

Corynanthe-type alkaloid featuring an unprecedented rearranged monoterpenoid segment (Fig.

3). In addition to its intriguing structure, preliminary studies have shown that 11 exhibits aggressive selective cytotoxicity toward the human prostate cancer cell line PC-3 (IC50 37µM) while remaining benign to peripheral blood mononuclear cells (PBMC) at 50 µM. Another novel

MIA, turpiniside (12), isolated in 2010,8 contains a pre-glucose degradation secologanin appendage on the oxidatively-cleaved tryptamine indole motif, and exhibits a relaxative effect on super coiled pBR322 plasmid DNA (DNA scission) in the presence of Cu (II) cation.

Figure 3. Some representative MIAs having potent biological activity

The aforementioned structural diversity and therapeutic promise of MIAs has prompted the interest of synthetic organic chemists. Since these natural products are usually produced in minimal quantities by the host plants, the scarcity of these compounds at times renders detailed biological studies extremely difficult. Thus, total synthesis from readily available starting materials serves as an attractive alternative to laborious isolations from natural sources to access

71 these complex MIAs. These total syntheses can also provide insight into the structural subtleties of the natural products. In fact, synthetic studies have often led to the revision of erroneously proposed structures, indicating the shortcomings of even the most sophisticated modern instrumentation methods.9 Moreover, total synthesis allows for structural modification/derivatization of the natural products, potentially providing leads in the development of novel drugs as well as providing information regarding the mechanisms of biological activity of MIAs.

To the synthetic chemists’ advantage, the pursuit of complex natural products, including monoterpenoid indole alkaloids, has been a major impetus toward the discovery of novel chemical transformations and methodologies. Certain MIAs have become classical showcases for synthetic organic chemists to illustrate the utility of novel methods. For example, the

Corynanthe-type MIA strychnine (10) which was first synthesized by Woodward in 1954,10 ushered in the modern era in the history of organic synthesis. After Woodward’s 26-step synthesis, twelve other total syntheses,11 each premised upon distinct key steps, have emerged in an evolving trend toward an ultimate level of atom-economy and conciseness, thanks to the development of new chemical transformations. Indeed, the most recent route, which was completed in 2011 by Vanderwal, et al. requires only six linear steps to afford the target molecule.11c

72

1.2. The Alstilobanine-Angustilodine Family: Unusual Monoterpenoid Indole

Alkaloids. Discovery, Structure Elucidation and Preliminary Bioactivity

Studies

In 2004, Kam and Choo12 isolated the novel monoterpenoid indole alkaloid angustilodine

(13) from the leaf extract of the Malayan plant Alstonia angustiloba. Angustilodine is a pentacyclic Corynanthe-type MIA. Unlike other Corynanthe alkaloids, 13 features a unique rearranged skeleton, containing a bridged seven-membered cyclic ether, in which the tryptamine aminoethyl moiety has been highly modified.

Figure 4. Angustilodine alkaloids: structures and conformations

In 2008, Morita and coworkers13 isolated several novel MIAs from the leaf extracts of A. angustiloba including alstilobanines A and E. Alstilobanine E (14) is the de-methyl congener of angustilodine. Alstilobanine A (15), on the other hand, contains an indole nucleus appended to a cis-azadecalin framework which lacks the seven-membered ether bridge. In this work, the term

“angustilodine-type alkaloids” will be utilized in reference to alkaloids 13, 14 and 15.

73

Morita, et al. fully characterized alstilobanines A and E through 2-D NMR experiments

(Figure 4). A strong NOE correlation of H14-H21 in alstilobanine E (13) was observed. Such a

1,4-flagpole interaction is a good indication that the piperidine moiety of alstilobanine E occupies a boat conformation. The piperidine moiety of alstilobanine A (15) was assigned as being in a chair conformation based on normal 1,3-diaxial H-H NOE correlations, as well the lack of any observed 1,4-flagpole hydrogen NOE interactions.

On the other hand, in their original isolation paper, Kam and Choo depicted the N-methyl congener of 14, angustilodine (13), as having the piperidine ring disposed in a chair conformation.12 However, no NOE data were reported to support this assertion. It appeared to us, however, that the N-methyl group of angustilodine should not cause change in the piperidine conformation relative to alstilobanine E (14). In addition, the reported hydrogen coupling constants of the piperidine moiety of angustilodine are in good agreement with those of alstilobanine E as reported by Morita, et al., suggesting that the piperidine moiety of angustilodine is in fact in a boat conformation. Indeed Professor Kam has recently informed us that the chair conformation shown for the angustilodine piperidine moiety in his paper was simply assumed.14

Preliminary biological activity studies have shown that alstilobanine A (15) has moderate relaxant activity against phenylephrine-induced thoracic rat aortic ring contraction with endothelium.13 However, no additional biological studies have been published to date.

Morita and coworkers have proposed a plausible biogenetic pathway for the formation of alstilobanines A and E from the common precursor stemmadenine (6) (Scheme 2). Thus, oxidation of the tertiary piperidine nitrogen in 6 followed by a Grob-type fragmentation of the

74 tryptamine side chain would afford key iminium ion intermediate 17. The chemistry of compound 17 then bifurcates in vivo towards either the oxepine 18 or epoxide 20. The epoxide

20 can be opened at the less substituted position via intramolecular attack of the indole C3 to provide alstilobanine E (14). In a similar fashion, intramolecular indole C3 attack on epoxide 20 results in the formation of alstilobanine A (15).

Scheme 2. Plausible biosynthetic pathway for the angustilodine-type alkaloids

1.3. Synthetic Studies on the Angustilodine-Type Alkaloids

The novel angustilodine-type alkaloids are densely functionalized molecules which are challenging synthetic targets. Prior to our published total synthesis of racemic alstilobanine A

(vida infra), no syntheses or synthetic approaches to any of these alkaloids had been reported.

75

Perhaps owing to the scarcity of material from the isolation, biological activity studies on these alkaloids have also been very limited. The challenging structures, together with the potential medicinal value of these alkaloids, prompted us to design a synthetic route to prepare these targets.

1.3.1. Original Synthesis Plan Toward the Angustilodine-Type Alkaloids

The original retrosynthetic analysis towards a stereoselective synthesis of the angustilodine- type alkaloids relied on a divergent endgame strategy leading to the three congeners, angustilodine (13), alstilobanine A (15) and alstilobanine E (14), from the common indole β- lactone 21, which contains the cis-fused azadecalin motif found in the target molecules (Scheme

3). To generate the C19-C20 single bond in 21, as well as the three contiguous stereogenic centers at C15-C19-C20 of the cis-fused azadecalin moiety, the plan was to apply an intramolecular nucleophile-promoted aldol lactonization (NPAL) developed by Romo,15 using keto-acid 21. Recognizing the 1,4-dicarbonyl motif embedded in keto-acid 21, we planned to unmask the ketone carbonyl in oxime ester intermediate 23. Finally, we intended to implement a pivotal C15-C16 bond formation via an intermolecular conjugate addition of the indole-2-acetate enolate 24 to the 3-piperidone-derived nitrosoalkene compound 25 using methodology recently developed by the Weinreb group (vide infra).

76

Scheme 3. Original retrosynthetic analysis of angustilodine-type alkaloids

1.3.2. Previous Synthetic Work Toward the Angustilodine-Type Alkaloids in the Weinreb

Group

Former Weinreb group member Max M. Majireck carried out the initial efforts towards a total synthesis of the angustilodine-type alkaloids. As planned in our retrosynthetic analysis (vide supra), the first pivotal step to forge the C15-C16 single bond would hinge on a conjugate addition of an indole ester enolate to the transient nitrosoalkene species 25 generated in situ.

Thus, the synthesis of the angustilodine-type alkaloids commenced with a search for appropriate counterparts for the proposed nitrosoalkene conjugate addition.

1.3.2.1. Background on Nitrosoalkene Conjugate Addition

Nitrosoalkenes are transient and highly reactive intermediates with lifetimes on the order of seconds. These reactive species are involved in an array of synthetically useful reactions. For instance, pericyclic reactions, including inverse electron demand [4+2]-cycloadditions have been observed for these intermediates.16 Since the electron-deficient nitrosoalkenes are electronically

77 similar to α,β-unsaturated carbonyl compounds, conjugate additions occur when species such as

27 are exposed to appropriate nucleophiles, resulting in an overall transformation involving α- substitution of α-halooxime derivative 26 to afford 28 (Scheme 4). Conjugate additions of nucleophiles to nitrosoalkenes, therefore, represent a useful method for the preparation of α- functionalized carbonyl compounds, in which the nitrosoalkene acts as an enolonium ion surrogate for a ketone or aldehyde.17 Suprisingly, this methodology has been only sporadically utilized. However, in recent years, the Weinreb group has undertaken a systematic investigation geared toward the further development and applications of nitrosoalkene conjugate additions, especially those using enolates as Michael donors.

Scheme 4.

Two common procedures have been used to generate nitrosoalkenes such as 27 in situ for use in conjugate additions. Base-promoted 1,4-hydrogen chloride elimination of an α- chlorooxime 26 (R = H) is the most commonly reported method.16a This procedure involves treatment of 26 with two equivalents of a nucleophile, with the first acting as a base for the initial elimination step. Despite its widespread usage, this method is inefficient when valuable nucleophiles are used. The less commonly employed, yet synthetically attractive Denmark protocol enables nitrosoalkene conjugate addition using only one equivalent of nucleophile.18 In this method, the 1,4-elimination of an α-chloro-O-silyloxime 26 (R = SiR3) to the corresponding nitrosoalkene 27 is implemented via the addition of a fluoride source. Thus, the Denmark procedure is potentially more practical in instances where precious nucleophiles are utilized. In

78 recent years, the Weinreb group has developed both intra- and intermolecular variants of methodology involving enolate conjugate additions to nitrosoalkenes generated via the Denmark method (Schemes 5, 6). 19 For example, deprotonation of a series of cyclic α-chloro-O- silyloximes 29 with an amide base, usually KHMDS, generated the stabilized carbanions 30.

Subsequent treatment of 30 with NBu4F (TBAF) following the Denmark protocol generated the nitrosoalkene intermediates 31 which readily underwent intramolecular Michael addition with the carbanions to afford a series of bridged bicyclic carbocycles 32 in good yields (Scheme 5).19a

Scheme 5. Intramolecular nitrosoalkene Michael additions19a

79

The Weinreb group has also reported a number of intermolecular nitrosoalkene Michael additions using an experimental procedure similar to the intramolecular variant to generate the corresponding adducts in good yields (Scheme 6).19c Entries (g) and (l) are of particular relevance to the total synthesis of the angustilodine-type alkaloids, since the phenylacetate enolate employed represents a Michael donor electronically related to the indole-2-acetate enolate 24 we proposed to employ in our synthetic plan for the natural products (cf. Scheme 3).

Scheme 6. Intermolecular nitrosoalkene Michael additions19c

1.3.2.2. Studies on Nitrosoalkene Conjugate Additions with Indole-2-acetate Enolates

Majireck’s exploratory studies toward a total synthesis of the angustilodine-type alkaloids commenced with investigation on the feasibility of the conjugate addition of an indole-2-acetate

80 enolate such as 36 to the piperidone-derived nitrosoalkene 2520 (Scheme 7). The Denmark protocol was initially selected for nitrosoalkene generation due to its potential efficiency relative to the 1,4-dehydrohalogenation of a free α-chlorooxime. Thus, one equivalent of N-Boc indole acetate 36 was treated with LiHMDS to afford the corresponding enolate 37. One equivalent of

OTBS chlorooxime 38 was then added, followed by subsequent slow addition of the fluoride source, TBAF. Unfortunately, this reaction failed to provide any of the desired Michael adduct

40. As an alternative, two equivalents of the indole acetate enolate 36 were reacted with one equivalent of the α-chlorooxime 39. While this operation did result in the formation of the desired Michael adduct 40, the yield was very low.

Scheme 7.

It was later discovered by Majireck that reaction of two equivalents of the monoanion of unprotected indole acetate 4121 with one equivalent of α-chlorooxime 39 resulted in the formation of the desired Michael adduct 43 in 59% yield (Scheme 8), along with some recovered starting ester 41 (66%). It was proposed that deprotonation of 41 likely generated the C-centered

81 monoanion 42b which exists in equilibrium with the N-monoanion 42a. This proposed facile equilibrium between the two monoanions 42a and 42b can be rationalized based on analysis of some model pKa data. Although the exact pKa values for the indole NH and ester α-H in compound 41 are not available, comparison of the pKa value for the NH of indole itself (ca. 21 in DMSO22) with that of the α-H of methyl phenylacetate (ca. 23 in DMSO22) provided an estimated Keq that supports the possibility of such an equilibrium. The observed Michael adduct

43 results from reaction of the C-centered monoanion 42b with the nitrosoalkene 25 derived from 39, with no adducts being detetected resulting from N-alkylation (which in fact might be reversible) or from indole C3-alkylation.

Scheme 8.

Since all attempts to further elaborate 43 via introduction of an acetic acid unit onto the indole C3-position of the oxime-protected adduct were unproductive, it was decided to explore an analogous nitrosoalkene Michael addition with indole diester 44, which already contains a

C3-substituent.

82

Scheme 9.

Thus, indole methyl ester 41 was treated with oxalyl chloride, followed by esterification with 2-trimethylsilylethanol to afford indole diester 44 in good yield. Indole diester 44 (Scheme

9) was then treated with two equivalents of LiHMDS to generate dianion 45, which was reacted with one equivalent of α-chlorooxime 39 to provide the desired Michael adduct 47 in nearly quantitative yield. It was proposed that this transformation begins with dehydrochlorination of 39 by dianion 45, generating the transient nitrosoalkene intermediate 25 along with the indole ester monoanions 46a and 46b, which likely exist as an equilibrium mixture as observed for anion

42a/b (cf. Scheme 8). The conjugate addition once again took place exclusively via the carbon- centered anion 46b. C15-C16 anti-Adduct 47 was initially believed to be formed via the nitrosoalkene Michael addition as single diastereomer at C15/C16 but as a mixture of E-oxime

83

47a and Z-oxime 47b which was not separated. However, it was later found that the stereochemical assignment of this conjugate addition was incorrect (vide infra).

1.3.2.3. Synthesis of the Key Romo NPAL Cyclization Precursor

With the nitrosoalkene Michael adduct 47a/b in hand, the synthesis continued toward the preparation of the key ketoacid precursor needed to explore the Romo NPAL cyclization (cf.

Scheme 3). It was necessary to first elaborate the indole C3-oxalyl moiety to the requisite acetate

(Scheme 10). Thus, the free oximes 47a/b (unseparated mixture) was treated with TBS chloride and imidazole to afford a mixture of the O-TBS-oximes 48, which was then subjected to a stepwise reduction/acetylation/hydrogenation sequence similar to that reported by Hlasta, et al.23 through acetate intermediates 49 to afford the deoxygenated compounds 50a/b (thought to be an

E/Z mixture of O-silyloximes) in high total yield. Slow evaporation of a solution of this mixture of NH indoles 50a/b in EtOAc provided crystals which were analyzed by X-ray crystallography.

The product 50a which was partially crystallized from the mixture was found to have the relative stereochemistry at C15-C16 as shown, as well as the E-geometry of the oxime motif (vide infra).

The material which did not crystallize could be separated for characterization from 50a on a silica gel column, 17 and was erroneously thought to be C15/C16 anti-Z-oxime 50b based on

Majireck’s original assumptions about the stereochemical outcome of the NA-Michael addition

(cf. Scheme 9). However, the unseparated mixture of what was believed to be 50a and 50b was used in subsequent steps by Majireck.

84

Scheme 10.

In an earlier model study by Majireck, an attempt had been made to deoximate oxime 43 to unmask the ketone moiety, but the incipient indole-NH ketone underwent intramolecular cyclization to afford the corresponding tetracyclic indolopyrrole 52 as the major product

(Scheme 11).

Scheme 11. Model deoximation of 43 and the formation of indolopyrrole 52

Thus, the indole NH of the mixture of 50a/b was first masked with a Cbz group to afford protected indole 53 (Scheme 12). The O-TBS group in 53 was then removed using TBAF and the resulting free oxime(s) 54 was converted to the corresponding oxime(s) pivalate 55.24

85

Scheme 12.

This intermediate 55 was then subjected to an Fe0-mediated reductive deoximation protocol developed in our laboratory25 to afford ketodiester 56 in variable yield (vide infra). The

TMS-ethyl ester moiety of 56 was next smoothly removed using TFA to afford the desired ketoacid 57 in high yield which at this time was still erroneously believed by Majireck to be a single C15/C16 diastereomer.

1.3.2.4. The Romo Nucleophile-Promoted Aldol-Lactonization (NPAL)

With ketoacid 57 now available, the feasibility of installing the cis-azadecalin motif for the angustilodine-type alkaloids using a pivotal nucleophile-promoted aldol-lactonization (NPAL) step developed by Romo and coworkers was next explored (cf. Scheme 3). 15a,15c,d In 2001,15a

Romo and coworkers reported that treatment of an aldehyde acid such as 58 with Mukaiyama’s

86 reagent (59) in the presence of a tertiary amine such as triethylamine led to clean production of the bicyclic β-lactones 60 in generally acceptable yields (Scheme 13).

Scheme 13. Romo NPAL using aldehyde acids for β-lactone preparation

In the mechanism proposed for this transformation, activation of the carboxylic acid with

Mukaiyama’s reagent 59 provides acyl species 61, which undergoes elimination to furnish a ketene intermediate 62 (Scheme 14). Upon attack by an nucleophilic promoter such as the triethylamine, followed by subsequent α-proton abstraction by the amine base, ketene 62 is converted to the key ammonium enolate aldehyde intermediate 63. Next, an intramolecular aldol reaction occurs that could, in principle, reversibly give either the anti-aldolate 65 or the syn- aldolate 66. While the anti-aldolate simply equilibrates back to enolate aldehyde 64 via a retro-

87 aldol pathway, the alkoxy and carbonyl moieties of the syn-aldolate 66 are properly aligned for further cyclization to occur, which ultimately leads to the cis-fused bicyclic β-lactone 68 and regeneration of the nucleophilic promoter.

Scheme 14. Mechanistic pathway of the Romo nucleophile catalyzed aldol lactonization (NCAL)

The original Romo NPAL methodology15a was limited in scope to aldehyde-acid substrates. However, further experimentation15c,d revealed that using a stoichiometric amount of

4-pyrrolidinopyridine (PPY) as the nucleophilic promoter, along with modified Mukaiyama’s reagent 70 with triflate as the counterion rather than the more nucleophilic iodide, expanded the scope of the methodology to include keto-acids such as 69 (Scheme 15). Romo, et al. were able to demonstrate the conversion of a variety of keto-acids to the corresponding bicyclic and tricyclic β-lactones all with excellent levels of diastereoselectivity. Of particular significance to our planned synthesis of the angustilodine-type alkaloids was the example constructing the cis-

88 fused decalin framework in the tricyclic β-lactone product 71 in entry d, since the angustilodine- type alkaloids also contain a similar cis-fused 2-azadecalin moiety which could, in principle, be installed via C19-C20 σ-bond formation using this methodology (cf. Scheme 3).

Scheme 15. Romo NPAL using keto acids as substrates

1.3.2.5. Construction of the Cis-Fused Azadecalin Moiety of the Angustilodine-Type

Alkaloids

The indole ketoacid 57 in the Majireck synthesis was subjected to the key aldol- lactonization sequence using the protocol developed by Romo, et al. To this end, ketoacid 57

89 was treated with 2-bromo-N-propylpyridinium triflate (70), PPY and DIPEA at room temperature to afford the pentacyclic β-lactones 72 and 73 in excellent yield, with the desired cis-azadecalin system 72 preferentially formed in about a 7:1 diastereomer ratio relative to the trans-azadecalin product 73 (Scheme 16). The products 72 and 73 were inseparable at this stage, and the mixture was carried on to the next step. However, the level of diastereoselectivity of this transformation was irreproducible, and the stereochemical course of this reaction proved to be more complex than originally thought (vide infra).

Scheme 16.

1.3.2.6. Studies on Introduction of the C16 Hydroxymethyl Group

With β-lactone intermediate 72 in hand, introduction of the C16 hydroxymethyl moiety was explored by Majireck. The first approach hinged on an intermolecular C16 hydroxymethylation using an advanced substrate derived from β-lactone 72. Thus, the β-lactone ring of the mixture of 72 and 73 was reductively cleaved with DIBAL-H to afford diol 74 in high yield, which was chromatographically separated from the corresponding diol product produced from the trans-2-azadecalin 73. Diol 74 was then treated with acetaldehyde dimethyl acetal under acidic conditions to provide acetal 75 as a single diastereomer. Acetal ester 75 was then subjected to a variety of alkylation conditions using a number of amide bases as deprotonation reagents, followed by alkylation with electrophiles such as monomeric formaldehyde, BOM-Cl

90 and MOM-Cl. Unfortunately all these attempts only resulted in recovery of the starting acetal ester 75. It was reasoned that a steric interaction of the N-Cbz group with the methyl ester moiety was preventing proper allignment of the C16-H with the ester carbonyl π* bond, which ultimately caused the failures in the attempted deprotonations to form an enolate.

Scheme 17.

However, treatment of acetal ester 75 with excess NaH and paraformaldehyde in DMF resulted in C16 hydroxymethylation with concomitant loss of the N-Cbz group, affording the indole hydroxymethyl product 76 in high yield. Unfortunately, the stereochemistry of the C16 hydroxymethyl group of 76 was found to be incorrect by extensive 2-D NMR studies. This transformation is believed to occur by initial cleavage of the Cbz group by adventitious NaOH

91 contained in the NaH, followed by ester enolization to afford dianion 77a/b. This dianion is probably in the more stable conformation 77b rather than 77a, which is highly destablized by

A1,3 strain.17 The subsequent alkylation with formaldehyde appears to occur via this conformation 77b from the face opposite to the bulky cyclic acetal moiety.

In view of these disappointing results, an intramolecular alkylation strategy was explored using diol 74 in an attempt to install the C16 hydroxymethyl group with the requisite stereochemistry for the target alkaloids (Scheme 18). Thus, the primary hydroxyl group in 74 was selectively protected as a silyl ether using tert-butyldimethylsilyl triflate/2,6-lutidine, and the N-Cbz group was cleaved via a palladium-catalyzed hydrogenolysis to afford the free indole

78. A chloromethylsilyl moiety was then installed onto the C-20 tertiary hydroxyl group of 78 and the resulting chloromethylsilane was further elaborated to the iodomethylsilane 79 via a

Finklestein reaction. Subsequent deprotonation of indole 79 with two equivalents of KHMDS generated the corresponding dianion which cleanly underwent intramolecular cyclization upon warming to room temperature to furnish the pentacyclic silyl ether 80.

Scheme 18.

92

With bridged silyl ether 80 in hand, numerous attempts17 were made to effect a Fleming-

Tamao oxidation to provide diol intermediate 81. Unfortunately, all Tamao-Fleming protocols attempted resulted in significant decomposition of the starting material, mostly through a retro- aldol loss of the hydroxymethyl group under the basic reaction conditions. The Woerpel modification26 of the Tamao-Fleming oxidation, which employs a mildly basic oxidant, was similarly unsuccessful in providing any of the desired diol 81. In addition to the retro-aldol problems, the electron rich free indole moiety was also likely incompatible with the oxidative

Tamao-Fleming and Woerpel conditions.27 Thus, in light of these failures, it became necessary to modify the C16-hydroxymethylation strategy and abandon the Tamao-Fleming approach (vide infra).

93

Chapter 2. Results and Discussion

2.1. Problems Associated with the First Generation Majireck Approach and

Some Solutions

In addition to the C16 hydroxymethylation problems encountered in the synthetic studies of the angustilodine-type alkaloids by Majireck (vide supra), several other problematic steps arose after the author began work on this project, all of which significantly handicapped progress.

These problematic steps included: (1) cleavage of intermediate oxime 54, which was irreproducible and sensitive to scaleup, (2) the nitrosoalkene conjugate addition which was initially incorrectly proposed to be exclusively C15/C16-diastereoselective, and (3) the Romo

NPAL which provided irreproducible and unreliable diastereoselectivity. The author’s initial work on this project was focused on solving these problems.

2.1.1. Deoximation of Intermediate 54

The pivalation-reduction deoximation approach employed in the Majireck route was rather inefficient, generally providing the 3-piperidone product 56 in only 50% overall yield from oxime 54 (cf. Scheme 12). This deoximation method was also difficult to reproduce on a large scale, affording ketoester 56 in variable but generally poor yields, thereby preventing efficient accumulation of the more advanced intermediates necessary for subsequent exploratory experiments towards the targets. Thus, the development of a more reliable deoximation protocol to unmask the C20 carbonyl in oxime 54 was necessary.

In 1959, Ponder and co-workers reported a mildly acidic method for oxime hydrolysis using levulinic acid and dilute HCl at or near ambient temperature that was effective on a multi-

94 gram scale.28 The reaction is presumably facilitated by intramolecular acid-promoted oxime cleavage. Thus, E-oxime isomer 54a was prepared from the crystalline indole NH silyloxime 50a, whose structure had previously been determined by X-ray analysis, via the N-Cbz protection/O- desilylation sequence used in the Majireck work (cf. Scheme 12). Gratifyingly, subjection of this

E-oxime isomer 54a to 9:1 (v/v) levulinic acid/1 M HCl at 30 oC cleanly afforded the desired ketone 56 in excellent yield (Scheme 19), and without any detectable erosion of the C15-C16 stereochemistry (vide infra). The reaction was run on multiple occasions on a seven-gram scale with no decrease in yield.

Scheme 19. Deoximation of E-oxime 54a using levulinic acid/HCl

2.1.2. Stereochemistry of the Nitrosoalkene Conjugate Addition Step

2.1.2.1. Discovery of the C15/C16 Diastereomer Problem

Originally it was believed that the nitrosoalkene conjugate addition of the enolate of indole-2-acetate 44 with piperidone-derived nitrosoalkene 25 exclusively formed a single C15-

C16 diastereomer but as a pair of E/Z oxime geometric isomers 47a and 47b (cf. Scheme 9). This proposal appeared suspect to the author, since a preliminary conformational analysis suggests that the preferred axial attack of the essentially planar indole acetate enolate 44 onto the equally stable half-chair conformers 25a and 25b of the 3-piperidone-derived nitrosoalkene19d is likely

95 from either conformer (Scheme 20). Thus it appeared a priori that this conjugate addition should lead to a mixture of C15/C16 anti- and syn-diastereomers 47 and 82, respectively, rather than exclusively giving the former diastereomer.

Scheme 20. Conformational analysis of the nitrosoalkene Michael addition with enolate 45

Indeed, Majireck’s assumption about stereochemistry of the nitrosoalkene Michael addition (NA-Michael) was found to be incorrect. Evidence for this erroneous conclusion came from detailed reactivity and product distribution analysis in the course of post-NA-Michael transformations using pure putative “Z”-oxime 54b-derived compounds. To prepare the pure

“Z”-oxime 54b, the supposed mixture of “E/Z”-oxime geometric isomers of the NA-Michael adducts 47a/b was converted to the corresponding acetates 49a and 49b via similar conditions used in Majireck’s approach (Scheme 21). The “Z”-acetates 49b were conveniently separated from the E-acetates 49a on a silica gel column, and were further converted to the free “Z”-oxime

54b through a reduction/oxime silylation/N-Cbz protection sequence used in the Majireck work.

96

Scheme 21. Preparation of pure “Z”-oxime isomer 54b

Thus, putative “Z”-oxime 54b was subjected to the same levulinic acid deoximation procedure used for the E-isomer (Scheme 22). Interestingly, this transformation took three days at 60 oC to afford the ketone product 56*, but in only 65% yield, contrary to the smooth and high yielding deoximation of the E-isomer 54a (cf. Scheme 19). Although this supposed “Z”-oxime- derived ketone product 56* has identical thin layer polarity as the authentic E-oxime-derived ketone 56, the 1H NMR spectra of these ketone products had some differences. However, although we suspected that this ketone ester was different from the one derived from oxime 54a, the NMR data were difficult to analyze since the spectra were complicated by the presence of

Cbz rotomers.

97

Scheme 22. Levulinic acid-mediated deoximation of “Z”-oxime 54b

Moreover, when the pure putative “Z”-oxime-derived ketoester 56* was subjected to the

TFA-mediated TMS-ethyl ester cleavage, a partial carbonyl epimerization appeared to occurr to afford an inseparable mixture of the anti- and syn-ketoacids 57 and 83 in a ratio which could not be determined by NMR. This epimerization was not observed for the authentic pure E-oxime derived ketoester 56 (cf. Scheme 26, vide infra), thus strongly suggesting that the ketoester 56* derived from the putative “Z” oxime series in fact had a different C15/C16 configuration relative to the E-oxime-derived compound 56.

Scheme 23. TMS-ethyl removal of putative “Z”-oxime derived ketoester 56*

2.1.2.2. Diastereoselectivity of the Nitrosoalkene Conjugate Addition

Since the source of the C15/C16 diastereomerism was not clear at this point, our focus next shifted to tracing the step which gave rise to the C15/C16 syn-diastereomer mixture. We were aware that the KHMDS-mediated N-Cbz installation step (cf. Scheme 12) might have

98 caused a C15/C16 epimerization due to the strong basicity of the conditions. In order to test if this step was the source of the C15/C16 diastereomer mixture, we decided to test a mild N-Cbz installation procedure which does not involve strong base (Scheme 24). Thus, both the E- and supposed “Z”- isomers of free indole 50a and 50b were N-Cbz protected under nearly neutral conditions using Cbz2O in the presense of a catalytic amount of DMAP to afford the corresponding N-Cbz derivatives 53a and 53b. Both of the resulting N-Cbz products 53a and

53b were spectroscopically identical to the corresponding compounds prepared using the original

KHMDS/CbzCl method. Therefore, it seemed likely that no C15/C16 epimerization had occurred during the KHMDS-mediated Cbz installation step (cf. Scheme 12).

Scheme 24. N-Cbz protection under mild conditions

Since it appeared that other post-NA-Michael transformations in the Majireck route (cf.

Schemes 10 and 12) were unlikely to cause epimerization, we thus believed that the C15/C16 syn-diastereomer had resulted from the key nitrosoalkene Michael addition. To confirm this postulate, the NA-Michael step was carefully reexamined (Scheme 25).

99

Thus, the NA-Michael addition was performed under identical conditions as those employed by Majireck, which provided a mixture of two thin-layer-separable products in a 1.2:1 ratio based on 1H NMR. For the purposes of characterization, these two products were separated on a silica gel column and each was individually elaborated to the NH silyloxime compounds

50a and 84. It was found that the NH silyloxime derived from the less-polar NA-Michael adduct had 1H and 13C NMR spectra identical to those of Majireck’s sample of 50a, which had been isolated by partial crystallization and unambiguously identified by X-ray crystallography as the

C15/C16-anti-E-silyloxime (cf. Scheme 10).

The NH silyloxime 84 derived from the more-polar NA-Michael adduct had spectroscopic characteristics very different from 50a that could not easily be attributed to oxime geometric isomerism (i.e. the chemical shifts of C15-H and C16-H of 84 were different by >0.3 ppm from those of 50a), and a later Romo NPAL reaction using the ketoacid derived from this compound 84 further supported the case for C15/C16 diastereoisomerism between 50a and 84

(vide infra).

Thus, this evidence supported the supposition that, rather than forming a single C15/C16 anti-diastereomer in the form of a pair of E/Z-oxime geometric isomers as proposed by Majireck, the nitrosoalkene Michael addition in fact provided a mixture of C15/C16 anti- and syn- diastereomers 47 and 82, with the anti-diastereomer 47 preferentially formed in a slight excess

(Scheme 25). Since the C15/C16 anti-NA-Michael adduct 47 has the E-oxime geometry (as shown by X-ray), we presume the oxime moiety of the C15/C16 syn-adduct 82 also has the E- geometry, although this point is irrelevant to the synthesis.

100

Scheme 25. Diastereoselectivity of the nitrosoalkene conjugate addition

2.1.3. Reexamination of the Majireck Synthesis and Observations of Problems with the

C15/C16 Syn-Diastereomeric Series

With the C15/C16 diastereomerism issue in mind, the Majireck synthesis was reexamined on a large scale. For practical convenience of large scale preparations, the diastereomeric NA

Michael products 47a and 82 were not separated. Instead, following a similar procedure used previously (cf. Scheme 21), oximes 47a and 82 from the nitrosoalkene Michael reaction were carried through as a mixture to the corresponding acetate precursor stage for the Hlasta reduction.

The two C15/C16 diastereomers 49a and 49b were then separated conveniently at this stage by silica gel column chromatography (cf. Scheme 21).

101

2.1.3.1. Reexamination of the Majireck Approach Using the Pure C15/C16 Anti-

Diastereomeric Series

2.1.3.1.1. Preparation of C15/C16 Anti-Ketoacid 57

The purified C15/C16 anti-acetate 49a was then converted to the corresponding ketoacid

57. In the first step, C15/C16 anti-isomer 49a was subjected to a large scale palladium-catalyzed reduction to 50a as employed in the Majireck synthesis (Scheme 26). Initially, this step could only be performed in high yield on a less than 200 mg scale at room temperature. However, by warming the reaction mixture at 30 oC for four days, we were able to perform this reduction on a

12-gram scale to form product 50a in excellent yield. The indole NH of anti-50a was then Cbz- protected using the Cbz2O/DMAP method (cf. Scheme 24), and the TBS group was removed with acetic acid-buffered NBu4F (TBAF) (cf. Scheme 12), followed by the levulinic acid- mediated deoximation (cf. Scheme 19), all of which proceeded smoothly on large scales to afford ketoester 56 in excellent yield (Scheme 26). TMS-ethyl ester of ketoester 56 was then cleaved using TFA to provide C15/C16 anti-ketoacid 57 as a pure single diastereomer.

Scheme 26.

102

2.1.3.1.2. Romo NPAL of C15/C16 Anti-Ketoacid 57

Reexamination of the Romo NPAL of anti-ketoacid 57 was then conducted. In several initial trials, the Romo NPAL of diastereomerically pure 57 proceeded with rather poor diastereoselectivity, with the cis-/trans-azadecalin product ratio of some runs being as low as 1:1.

We suspected that due to the prolonged reaction time, the diisopropylethylamine and 4- pyrolidinopyridine might be effecting a carbonyl epimerization, which in turn might be causing the poor diastereoselectivity of the cyclization. By accident, we noticed on one occasion that a sample of ketoacid 57 (purified using 1% acetic acid in ethyl acetate/hexanes on silica gel) that contained residual acetic acid, when subjected to the Romo cyclization, showed very high diastereoselectivity in formation of the β-lactone. We thus performed several small scale Romo

NPAL reactions using ketoacid 57 that had been stringently purified to exclude any residual acetic acid, with or without acetic acid as an additive. While the trials without the acetic acid additive afforded the β-lactone products again in a ~1:1 cis/trans diastereomer ratio, a reaction run with the addition of 1.2 equivalents of acetic acid furnished the desired cis-diastereomer 72 in a very favorable 13:1 dr, even after 16 h at ambient temperature. Furthermore, we later discovered that the NPAL reaction only took 1 h to go to completion and, in fact, shortening the reaction time further improved the dr to 97:3 without any decrease in yield (Scheme 27).

Scheme 27. Optimized Romo NPAL using C15/C16 anti-ketoacid 57

103

2.1.3.2. Reexamination of the Majireck Approach Using the C15/C16 Syn-Diastereomeric

Series

After reexamination and optimization of the synthetic sequence using the pure C15/C16 anti-diastereomeric series through the Romo NPAL step (vide supra), we next examined the same reactions using the pure C15/C16 syn-acetate 85 we had obtained earlier as shown in

Scheme 21 ( previously numbered as 49b).

2.1.3.2.1. Preparation of C15/C16 Syn-Ketoacid 83

C15/C16 syn-Acetate 85 was subjected to a synthetic sequence similar to that used for the anti-isomer (Scheme 28). Thus, transformation of acetate 85 to N-Cbz free oxime 86 (previously

54b) went in good yields similar to those using the C15/C16 anti-diastereomeric series (cf.

Schemes 26). As previously noted, the deoximation of 86 was sluggish (cf. Scheme 22), but provided the pure C15/C16 syn-ketoester 87 in an acceptable yield.

Scheme 28. Preparation of pure C15/C16 syn-ketoester 87

104

The TFA-mediated TMS ethyl ester cleavage using pure syn-ketoester 87 was then re- examined (Scheme 29). As noted earlier, the syn-ketoester had been found to undergo carbonyl epimerization upon treatment with TFA (cf. Scheme 23). In order to obtain a pure sample of

C15/C16 syn-ketoacid 83, we attempted to optimize the deesterification reaction and subsequent purification conditions. Although this ester cleavage proceeded in reliably good overall yields, under no conditions were we able to avoid the carbonyl epimerization, nor were we able to separate the syn-ketoacid 83 from the anti-isomer 57 from the ester cleavage product mixture. In addition, the ratio of these two diastereomers could not be determined by NMR in any of the mixtures.

Scheme 29. TFA-mediated TMS-ethyl ester cleavage of C15/C16 syn-ketoester 87

2.1.3.2.2. Romo NPAL of the C15/C16 syn-Ketoacid 83

Although we could not obtain a pure sample of C15/C16 syn-ketoacid 83, we decided to explore the Romo NPAL with the mixture of C15/C16 diastereomers. Thus, the TFA-ester- cleavage product mixture 57/83 was subjected to the Romo NPAL reaction using the optimal conditions found for pure 57 with HOAc as an additive (cf. Scheme 27) (Scheme 30). However, contrary to the excellent cis-azadecalin selectivity achieved using pure C15/C16 anti-isomer 57

(cf. Scheme 27), this reaction furnished a complex, inseparable mixture of what appeared to be cis-fused β-lactones 72 and the trans-fused β-lactones 73 (Scheme 30).

105

Scheme 30. Detailed product distribution analysis of the Romo NPAL using inseparable C15/C16 diastereomeric mixture of ketoacids 83/57

To assess the diastereoselectivity of this transformation, the inseparable mixture of N-Cbz

β-lactones 72/73 was subjected to a palladium catalyzed hydrogenolysis to afford a separable mixture of NH indole β-lactones 88 and 89 each as a pair of C16 ester mixture as determined by extensive 2-D NMR experiments (Scheme 30). The configuration of each isomer was determined primarily by using NOE correlations between H15, H16 and H19. The ratio of α/β ester epimers 88a/88b in the trans-β-lactones was 2.5:1, while that in the cis-β-lactones was 1:26.

Interestingly, the undesired trans-fused β-lactones 88a/b were the major products of this cyclization. Thus, the configuration at C15/C16 clearly plays an important role in directing the

106 azadecalin-stereochemistry in the pentacyclic β-lactone products, but at present time it is difficult to assess the reason for this observation.

2.1.3.2.3. Solution to the C15/C16 Syn-diastereomer Problem

Since the trans-decalin major product 73 in the Romo NPAL of C15/C16 syn-ketoacid 83 has the incorrect azadecalin configuration for the angustilodine-type alkaloids, it became important to find a way to use the C15/C16 syn-diastereomeric series which leads to a complex mixture of β-lactones (vide supra).

Fortunately, the syn-epimer of free indole TBS oxime 84 was found to undergo a smooth

C16 epimerization upon treatment with KHMDS at rt without significant substrate decomposition to afford the desired anti-diastereomer 50a in good yield (Scheme 31). Therefore, the C15/C16 syn-diastereomeric series could also be used in the synthesis. Interestingly, this epimerization did not occur when the mixture of 84 and KHMDS was kept and quenched at -78 oC. Thus, it appeared to us that the anti-50a is the thermodynamically favored diastereomer, but the reason for this is unclear at the present time.

Scheme 31.

107

2.2. Modified Strategies for the Divergent Synthesis of the Angustilodine-Type

Alkaloids

2.2.1. C16 Hydroxymethylation

2.2.1.1. A Modified C16 Hydroxymethylation Strategy

Although we had initially envisaged that a Fleming-Tamao-fashioned oxidation of the bridged pentacyclic silylether 80 would deliver the diol 81, which could be further elaborated to either the oxepane moiety or to the C19 methyl in the angustilodine type alkaloids, we were forced to abandon this approach due to numerous problems encountered in this transformation

(cf. Scheme 18). Thus, we turned to effecting a C16 hydroxymethylation strategy involving a stereoselective α-alkylation of the NH β-lactone ester 89a, prepared previously by Cbz removal of the major Romo NPAL product 72 (cf. Scheme 30), with an appropriate electrophile to introduce the C16 hydroxymethyl moiety in 91 (Scheme 32). Molecular models suggested that the approach of an electrophile should occur from the less hindered face of the pentacyclic enolate dianion 92, leading to the desired C16-stereochemistry in 91. The resulting lactone 91 would in turn be elaborated to the desired diol 90 via a reductive cleavage of the β-lactone moiety. Subsequent intramolecular cyclization of diol 90 would ultimately lead to the oxepane- containing angustilodine-type alkaloids 13 and 14, and reduction of the C18 hydroxymethyl group would afford alstilobanine A (15).

108

Scheme 32.

Prior to our work, Overman and co-workers had reported an α-alkylation of a similar indole ester enolate with monomeric formaldehyde in their total synthesis of actinophyllic acid

(94) (Scheme 33).29 In their approach, the NH indole ester 93 was enolized by treatment with

LDA, followed by addition of monomeric formaldehyde to afford the α-hydroxymethylated product 93a, which upon acid treatment cyclized to the hemiacetal with concomitant methyl ester hydrolysis to afford the alkaloid 94.

109

Scheme 33.

2.2.1.2. C16 Hydroxymethylation Studies

Before we commenced testing the feasiblity of the key alkylation reaction of the requisite substrate 89a, a preliminary screening of potentially competent electrophiles was carried out using indole ketodiester 44 as a model substrate. Thus, treatment of indole diester 44 with two equivalents of LiHMDS at -78 oC to form the dianion, followed by dropwise addition of monomeric formaldehyde prepared following the procedure developed by Schlosser, et al.,30 resulted in the desired C-hydroxymethylated product 95 in moderate yield (Scheme 34). A low temperature quench of the resulting reaction mixture with acetic acid was found to be crucial to the success of these reactions, since using a standard aqueous ammonium chloride work-up led only to decomposition products. Electrophiles other than monomeric formaldehyde, such as

BOMCl and MOMCl, were also found to react as desired under these conditions.

110

Scheme 34.

To test the alkylation on the actual substrate 89a, the N-Cbz group of the inseparable 97:3 mixture of intermediates 72/73 (prepared as described in section 2.1.3.1.2. from C15/C16 anti- ketoacid 57 using a Romo NPAL) was first removed via palladium-catalyzed hydrogenolysis to afford the free indole β-lactone 89a in excellent yield (Scheme 35). This compound was easily separated by silica gel column chromatography from a small amount of the trans-azadecalin product 88b that arose from the minor Romo NPAL diastereomer 73. The reaction was carried out in non-nucleophilic ethyl acetate as solvent containing a catalytic amount of glacial acetic acid (as proton source necessary for facilitating the hydrogenolysis) in order to avoid β-lactone

111 ring opening by the nucleophilic protic solvents such as methanol that are typically used for such a hydrogenolysis.

Scheme 35.

The alkylation of 89a was examined first under the same conditions used in the model studies (cf. Scheme 34). Thus, free indole β-lactone 89a was treated with two equivalents of

LiHMDS to generate the dianion, followed by addition of the same three electrophiles used previously (Scheme 36). Surprisingly, only the starting ester 89a was recovered in all cases. We suspected that LiHMDS might not be a strong enough base to fully deprotonate 89a, hence the recovery of the starting ester. Thus, we next screened a variety of strong bases including

KHMDS, NaH and KH (Scheme 36, entries 4-8) in combination with different electrophiles, most of which unfortunately resulted only in decomposition of the ester substrate. Treatment of

89a with KHMDS, followed by paraformaldehyde or monomeric formaldehyde30 and subsequent warming of the reaction mixture to room temperature led primarily to the undesired N- hydroxymethylated product 98 (entry 4).

112

Scheme 36.

Since our trials with strong sodium or potassium bases were unproductive, and generally led to decomposition products, based on the work of Overman (cf. Scheme 33), we decided to reinvestigate the use of a lithium base. We believed that perhaps the use of a lithium base might be crucial to this alkylation due to either the greater stability of a lithium enolate compared to potassium and sodium, and/or the lithium counterion which chelates with the formaldehyde better than the potassium and sodium in the corresponding Zimmerman-Traxler transition state of the aldol step that might ultimately facilitate the reaction. Thus, we decided to revisit the reaction of β-lactone ester 89a with LiHMDS and monomeric formaldehyde, and thoroughly

113 investigate the feasibility of effecting such an alkylation by fine-tuning reaction parameters such as temperature and work-up conditions.

Our failed LiHMDS-mediated alkylation attempts of ester 89a (cf. Scheme 36, entries 1-

3), which resulted in complete recovery of starting ester 89a, were all carried out at -78 oC. We hoped that a slightly higher temperature might trigger the aldol reaction. Indeed, after optimizing the reaction temperature as well as the work-up procedure, we were able to effect the desired

C16 hydroxymethylation to provide hydroxy lactone 99 as a single diastereomer in acceptable yield (Scheme 37). Although the initial indole ester dianion formation had to be performed at -78 oC to avoid decomposition, we discovered that the reaction of the dianion of 89a with monomeric formaldehyde required a slight temperature elevation to -40 oC. At temperatures higher than -40 oC, significant N-hydroxymethylation of both starting substrate 89a and product

99, as well as decomposition of the starting material, were observed. However, the N-alkylation problem could be suppressed by quenching the reaction with glacial acetic acid at -40 oC within

8 min of warming to that temperature from -78 °C.

Scheme 37.

Extensive 2-D NMR experiments confirmed that the hydroxymethyl moiety is connected to C16 (confirmed by key HMBC signals: 3J C2-H17, 3J C17-H15 and 3J H17-methyl ester carbonyl) and is on the same side of the pentacycle as the β-lactone moiety (key nOe correlation includes H17-H15), thus providing the requisite stereochemical arrangement for elaboration to

114 the angustilodine-type natural products (Fig. 5). The strong nOe crosspeak caused by a 1,3- diaxial interaction between H15 and H21a also suggested a chair conformation of the piperidine ring of 99.

Figure 5. Key HMBC and nOe correlations of hydroxymethyl β-lactone 99

An interesting lactone rearrangement of C16 hydroxymethyl β-lactone 99 was found during some preliminary attempts to access the oxepane moiety contained in angustilodine (13) and alstilobanine E (14). We had initially planned to prepare the oxepane ring via a cyclization of a diol intermediate such as 90 (cf. Scheme 32). To prepare an intermediate similar to 90, we first attempted to form a C17 tosylate from hydroxymethyl indole β-lactone 99 (Scheme 38).

Suprisingly, the expected tosylate 100 was not produced under a variety of conditions, but rather

ɛ-lactone product 102 formed in the presence of various bases or Lewis acids such as Yb(OTf)3.

We also attempted to install a TBS group on the primary alcohol of 99 using TBSCl and imidazole, but instead, the same rearranged ɛ-lactone 102 was observed. Extensive 2-D NMR and mass spectrometric analyses established unambiguously that 102 is indeed the ɛ-lactone product resulting from translactonization of the β-lactone 99. This result further confirmed that the alkylation of the enolate 92 with monomeric formaldehyde had occured stereoselectively from the requisite face (cf. Scheme 32), furnishing the hydroxymethylated β-lactone 99 with the appropriate stereochemistry for the angustilodine natural products.

115

However, despite screening of a number of different bases, attempts to effect this translactonization cleanly was unsuccessful, with a typical 35% yield of 102 being the optimal result using triethylamine at room temperature along with recovery of the starting β-lactone.

Recycling of unreacted starting β-lactone 99 was also only partly successful, presumably due to losses from ester hydrolysis and hydroxymethyl cleavage via a retro-aldol reaction. Nevertheless, by recycling 99 once, we were able to generate the ɛ-lactone in an acceptable 58% yield.

Scheme 38.

Another noteworthy finding from the 2D NMR experiments on the ɛ-lactone 102 was a strong nOe correlation between H14 and H21 caused by a flagpole-like interaction (Scheme 39), which suggested that the piperidine ring in 102 is in a boat conformation similar to that reported for alstilobanine E (14) (cf. Figure 4),13 which posesses essentially the same skeletal connectivity as pentacycle 102. On the other hand, the piperidine ring of β-lactone 99 is in a chair conformation (vide supra, cf. Figure 5). Thus, translactonization of β-lactone 99 to ɛ-lactone 102 would require a conformational inversion of the piperidine ring in the transition state. In addition, although the ring strain of the β-lactone moiety is partially released in the ɛ-lactone product, the energetically penalizing boat-like piperidine ring of ɛ-lactone 102 counterbalances the favorable

β-lactone ring-strain release. Due to these factors, β-lactone 99 and ɛ-lactone 102 are likely in equilibrium at room temperature in the presence of base, which in turn prevents the translactonization to go to completion.

116

Scheme 39.

2.2.2. C18 Deoxygenation and Total Synthesis of Alstilobanine A

2.2.2.1. The First Generation C18 Deoxygenation Approach

During some preliminary studies on the feasibility of forming the oxepane moiety of angustilodine and alstilobanine E by by a lactone-to-cyclic-ether reduction of the ɛ-lactone moiety of 102 (vide infra), we discovered that treatment of lactone 102 with DIBAL-H and a

o catalytic amount of boron trifluoride etherate in CH2Cl2 at -78 C afforded triol 103 along with some aldehyde diol product 104 (Scheme 40). Although we were not able to improve the yield of this partial reduction to afford aldehyde 104 cleanly, we decided to investigate synthesis of alstilobanine A (15) using this intermediate since it appeared that the aldehyde functionality would allow access to the C18 methyl group found in this alkaloid.

Scheme 40.

117

Thus, the aldehyde moiety in 104 was converted to dithiane 105 with 1,2-ethanedithiol in the presence of boron trifluoride etherate, which was then subjected to desulfurization with freshly prepared W-2 Raney nickel31 in ethanol at 60 oC (Scheme 41). To our delight, the desired

C18 methyl compound 106 was formed in acceptable overall yield. It should be noted that, due to over-reduction and purification problems involved in the preparation of aldehyde 104 from ɛ- lactone 102, dithiane 105 was difficult to fully purify.

Scheme 41.

Intermediate 106 was then treated with sodium naphthalenide in THF at -78 oC in an effort to remove the N-tosyl group to furnish alstilobanine A. Although Majireck had successfully applied these detosylation conditions to a simpler model substrate,17 compound 106 decomposed, yielding a complex mixture of highly polar products. We also attempted the detosylation of 106 using some milder methods including sonication with maganesium turnings in anhydrous methanol,32 as well as treatment with hydrogen fluoride/pyridine complex.33

However, the C16 hydroxymethyl group proved labile toward these conditions, resulting in loss of formaldehyde through a retro-aldol pathway to give methyl indole 107 as the sole isolable product.

118

Scheme 42.

From these results, protection of the C17 primary alcohol appeared to be necessary to avoid retro-aldol decomposition of 106 (Scheme 43). Thus, upon treatment with tert- butyldimethylsilyl triflate and 2,6-lutidine at rt, the primary alcohol moiety of the indole diol 106 was selectively protected to afford silyl ether 108. Sonication of 108 with magnesium turnings in anhydrous methanol at room temperature smoothly cleaved the N-tosyl group without causing any side reactions to afford NH piperidine indole 109 in moderate yield.

Scheme 43.

Preliminary attempts at TBS removal from 109, however, did not give any alstilobanine

A (15). However, due to a lack of material, we were only able to test a limited number of O-TBS cleavage conditions, including TBAF/HOAc, all of which failed. The scarcity of the material was caused by inefficient transformations in our first generation route. In fact, the best yield of 108 was only 8% over 5 steps from β-lactone 99. Specifically, the translactonization of 99 to 102 and the subsequent reduction to aldehyde 104 are particularly low yielding (cf. Schemes 40-41).

Difficulties in reproducing the formation of dithiane 105 from aldehyde 104, as well as problems

119 with the W-2 Raney nickel-mediated desulfurization of 105 provided additional impetus for us to redesign the endgame route toward alstilobanine A.

2.2.2.2. A C18 Deoxygenation /C16 Hydroxymethylation Approach

Due to these problems, we thus sought to develop a route which avoids the use of aldehyde 104. To this end, a C18 deoxygenation of diol 74 was explored followed by installation of the C16 hydroxymethyl moeity (Scheme 45). We hoped that the stereochemical outcome of the C16 hydroxymethylation of 110 would take a similar couse as that of β-lactone 99.

Scheme 45.

Since some previous studies by Majireck17 had indicated that the classic Barton-

McCombie deoxygenation conditions were not compatible with intermediate 74, we instead decided to investigate a deoxygenation of the corresponding alkyl iodide derived from 74 employing a palladium-catalyzed hydrogenation which would simultaneously remove the N-Cbz, enabling subsequent ester enolization. Thus, diol 74 was first converted to mesylate 111 which was then subjected to a Finkelstein reaction in order to prepare the C18 alkyl iodide 112 (Scheme

46). However, mesylate 111 was resistant to iodination, providing only a small amount of iodide

112 even after prolonged reaction times.

120

Scheme 46.

We next attempted to prepare the C18-iodide via an Appel reaction34 of diol 74 (Scheme

47). However, we were unable to isolate any of the desired iodide 112 upon treatment of diol 74 with PPh3, imidazole and iodine in acetonitrile. Instead the major product was oxazoline 113 which arises via participation of the acetonitrile solvent in the reaction.

Scheme 47.

However, by changing the solvent to dichloromethane, we managed to suppress the oxazoline formation and isolate the desired C18 iodide 112 in good yield (Scheme 48). We next subjected iodide 112 to a palladium-catalyzed hydrogenolysis using hydrogen gas.23,35 A 1:1 t-

BuOH/EtOAc solvent system containing a catalytic amount of acetic acid was employed here in order to suppress the possible the formation of undesired C18 ether side products that could result from the use of the more standard nucleophilic hydrogenation solvents such as methanol or ethanol. After 16 h at room temperature, deiodination was achieved with simultaneous N-Cbz cleavage, affording the free NH indole deoxygenated compound 110 in excellent yield.

121

Scheme 48.

With the C18 methyl compound 110 in hand, our plan was to introduce the C16 hydroxymethyl moiety via alkylation with LiHMDS and monomeric formaldehyde in a manner similar to that used for β-lactone ester 89a (cf. Scheme 37). Thus, ester 110 was treated with excess LiHMDS at -78 oC, followed by freshly prepared monomeric formaldehyde30 in THF at -

40 oC (Scheme 49). This reaction, however, produced exclusively N-hydroxymethyl derivative

114 in high yield. The desired transformation would probably require the formation of a trianion

110b rather than dianion 110a, which could be problematic. Thus, we reasoned that protection of the tertiary alcohol would solve this problem.

Scheme 49. Attempted C16 hydroxymethylation

122

Thus, C19 methyl indole 110 was refluxed with trimethylsilyl cyanide and imidazole in dichloromethane in a sealed tube to afford the silyl ether product 115, which was then subjected to the alkylation with monomeric formaldehyde (Scheme 50). Unfortunately, the product of alkylation was again exclusively the N-hydroxymethylindole 116 in moderate yield, regardless of whether LiHMDS or LDA was employed as base. We reasoned that the steric effects of the C19 methyl and OTMS groups were preventing the C16 hydroxymethylation reaction from proceeding as desired.

Scheme 50.

2.2.2.3. The Optimal End-Game Route and Total Synthesis of Racemic Alstilobanine A (15)

With the problems hitherto encountered in the route toward alstilobanine A in mind, we decided to explore a different strategy toward the alkaloid, which would utilize the mild and efficient Pd-catalyzed deiodination step previously tested to form a methyl group (cf. Scheme

48), as well as integrate a late stage TBS protection of the C16 hydroxymethyl motif (cf. Scheme

43) necessary for the sulfonamide removal (Scheme 51). Thus, the palladium-catalyzed hydrogenolysis of C18 iodide 118 would lead to the deoxygenated C18 methyl derivative 108 in a manner similar to our previous study (cf. Scheme 48). The alkyl iodide moiety in 118 would be introduced by an Appel reaction of diol 117, which would in turn be prepared via a reduction of the C17-O-TBS β-lactone 101.

123

Scheme 51.

Our investigation thus commenced with preparation of the C17-O-TBS β-lactone 101.

Previous efforts to protect the C16 hydroxymethyl moiety in lactone 99 as the O-TBS ether failed due to the rapid competing translactonization in the presence of a variety of bases (cf.

Scheme 38). On the other hand, hydroxymethyl β-lactone 99 is stable at -78 oC in the presence of excess LiHMDS (cf. Scheme 37), indicating that the translactonization requires a higher reaction temperature. Thus, we hoped that silyation of the free C17 hydroxyl group of 99 would occur at a low temperature in the presence of a weak base that would serve as an acid scavenger rather than generating the C17 alkoxide.

The TBSOTf/lutidine-mediated conditions we had employed earlier (cf. Scheme 43) met most of our criteria: TBSOTf is a very reactive silylation reagent, and 2,6-lutidine is a weaker and more hindered base than triethylamine, imidazole or potassium carbonate, all of which caused translactonization. Thus, the silylation of C16 hydroxymethyl β-lactone 99 was performed at both room temperature and at 0 oC (Scheme 52). Indeed, the reaction performed at

0 oC gave predominantly the desired C17 OTBS β-lactone 101 in high yield, while the same

124 reaction performed at room temperature gave a mixture of C17 OTBS β-lactone 101 and C20

OTBS ε-lactone 119 in about a 1:1 ratio.

Scheme 52.

With the C17 primary alcohol now protected as the TBS ether, we needed a mild reducing agent that would convert the β-lactone motif to diol 117 without concomitant OTBS cleavage (cf. Scheme 59, vide infra). We were pleased to discover that LiBH4 efficiently and chemoselectively reduced the β-lactone ester to diol ester 117 in good yield (Scheme 53). The subsequent Appel reaction of diol 117 furnished iodide 118 in excellent yield. Iodide 118 was then smoothly dehalogenated via a palladium-catalyzed hydrogenation to the corresponding C19 methyl compound 108, which was identical to the intermediate from our previous but inefficient route (cf. Scheme 43). This new streamlined route afforded 108 reproducibly in 58% yield over four steps from β-lactone 99, compared to 8% yield over five steps from β-lactone 99 in our previous synthesis.

After clean N-tosyl removal of compound 108 using magnesium turnings with sonication at room temperature, we tested a series of mild conditions to cleave the C17 TBS group from

OTBS methyl piperidine 109. It was found that dry hydrogen chloride in methanol/chloroform was the most effective reagent. Following a convenient work-up by simply evaporating the

125 reaction mixture under high vacuum, the hydrochloride salt of racemic alstilobanine A (15) was isolated in quantitative yield. The synthetic sample was found to have proton and carbon NMR spectra identical to those reported for the natural product.13,36,37

Scheme 53.

126

2.2.3. Total Syntheses of Alstilobanine E (14) and Angustilodine (13)

2.2.3.1. Attempted Oxepane Formation from ɛ-Lactone 102

Since ε-lactone 102 already contains the desired seven-membered ring structurally related to the oxepane moiety found in angustilodine and alstilobanine E, we next investigated whether direct chemoselective deoxygenation of the lactone to the 7-membered ring cyclic ether could be effected without any concomitant ring opening.

2.2.3.1.1. Attempted Lewis Acid-Catalyzed ɛ-Lactone Reduction

We first tested a series of Lewis acid-catalyzed lactone reduction conditions on ɛ-lactone

102 (Scheme 54). Unfortunately, all of these trials resulted in conversion to the over-reduced triol 103. The success of the chemoselective lactone deoxygenation relies upon the stability of the lactol intermediate such as 120 resulting from the initial complexation/reduction step. Thus, the observed over reduction to the triol 103 was not suprising, since NMR analysis (cf. Scheme

40) revealed that the hydroxymethyl aldehyde 104 was strongly favored over the ɛ-lactol 120, presumably due to the high energy double boat conformation of the latter species (cf. Scheme

39).

127

Scheme 54.

2.2.3.1.2. Attempted Silane Radical-Mediated ɛ-Lactone Reduction

Given our failure to effect the desired Lewis acid-catalyzed reduction of ɛ-lactone 102, we next attempted to apply the chemoselective lactone-to-cyclic-ether reduction method reported by Nakao, et al. which employs trichlorosilyl radical generated via irradiation of trichlorosilane in the presence of benzoyl peroxide with a flood lamp.38 To avoid potential undesired side reactions, we first protected the C20 tertiary hydroxyl of ɛ-lactone 102 as an OTBS ether 119. ε-

Lactone 119 was then converted to the N-Boc derivative 121 using Boc anhydride and DMAP

(Scheme 55). A mixture of ɛ-lactone 121, trichlorosilane and a catalytic amount of benzoyl peroxide in THF was then irradiated with a flood lamp in a sealed tube for 1 h. However, the

128 only product detected was the free tertiary alcohol 122 which arose via desilylation of 121 by

HCl generated in situ. Longer reaction times also did not provide any of the desired ether product.

Scheme 55.

2.2.3.1.3. Thiolactonization/Reduction Approach

During the course of the synthetic studies toward the brevetoxin natural products,

Nicolaou and co-workers developed an elegant two-step method for the preparation of medium sized cyclic ethers from the corresponding lactones (Scheme 56).39 In this process, the lactone

123 was first converted to the thiolactone 124 using Lawesson’s reagent. Thiolactone 124 was then partially reduced to the cyclic methylthioacetal 125 via treatment with a reducing agent such as LiBHEt3 or a Grignard reagent, followed by methyl iodide. Thioacetal 125 was then either desulfurized using a tin hydride radical procedure to afford the cyclic ether 126.

Scheme 56.

Although we realized that applying the Nicolaou methodology to our route toward angustilodine and alstilobanine E would add a few more steps than desirable, its promise as a

129 good pathway for the formation of medium sized cyclic ethers could not be ignored.

Unfortunately, treatment of the fully protected ε-lactone 121 with Lawesson’s reagent at elevated temperatures resulted in substantial decomposition (Scheme 57). A milder thiolactonization procedure using the acid scavenger N,N,N’,N’-tetramethylthiourea along with Belleau’s reagent was similarly unsuccessful, resulting only in O-TBS cleavage and decomposition.

Scheme 57. Attempted thiolactone formation from ɛ-lactone 121

2.2.3.2. Oxepane Formation via Intramolecular Etherification

2.2.3.2.1. Attempted Intramolecular Etherification via Mitsunobu Reactions of Triol 103

Since our attempts to form the oxepane moiety using ɛ-lactone 102 were unproductive, we instead opted to use triol intermediate 103 which we hoped could be selectively cyclized via etherification of the C17-C18 primary alcohols. Thus, hydroxymethyl β-lactone 99 was treated with DIBAL-H at room temperature to afford the triol 103 in moderate yield (Scheme 58). Triol

103 was then subjected to Mitsunobu reaction conditions in an attempt to furnish the bridging oxepane of the angustilodine-type alkaloids. However, this strategy was unsuccessful. The major product of treatment of triol 103 with one equivalent of DIAD/PPh3 at room temperature was the five-membered cyclic ether derivative 128, which contains a free hydroxyl group at C18.40 In

130 addition, a small amount of oxetane 12741 was isolated as a minor side product, which is the result of intramolecular etherification of C20 hydroxyl with the C18 primary alcohol motif.

In a different trial, when triol 103 was treated with an excess of DIAD/PPh3 at room temperature, the result was formation of cyclopropane 129.42 We also tried to perform the cyclization at a higher temperature in an attempt to force the oxepane formation. Thus, triol 103 was refluxed in benzene with one equivalent of DEAD/PPh3. However, this experiment led only to a product selectivity in favor of the oxetane product 127 along with minimal amount of both

128 and 129. The desired oxepane product was never observed under any of these Mitsunobu conditions.

Scheme 58. Attempted oxepane formation from triol 103 through Mitsunobu reactions

131

2.2.3.2.2. Attempted Intramolecular Etherifications via Mitsunobu Reactions Using

Selectively Protected Triols

Given the failures we encountered using triol 103, we therefore elected to selectively protect the tertiary C20 hydroxyl group as the t-butyldimethylsilyl ether in order to try to alleviate some of these problems (Scheme 59). Thus, ɛ-lactone 102 was treated with excess tert- butyldimethylsilyl triflate and 2,6-lutidine in dichloromethane at room temperature to afford the silyl ether derivative 119. However, the ɛ-lactone moiety was quite resistant toward reduction, and required the use of excess DIBAL-H at room temperature, which resulted in loss of the C20

TBS protecting group to give triol 103.43

Scheme 59.

Since the reduction of ɛ-lactone 119 was problematic, we next attempted protection of triol 103 using trimethyl orthobenzoate and a catalytic amount of p-toluenesulfonic acid in refluxing toluene to selectively mask the C20-C18 diol as the benzylidene acetal 130 (Scheme

60). We had hoped to partially reduce the benzylidene ring of 130 to unmask the C18 primary hydroxyl group while converting the C20 tertiary alcohol to the benzyl ether 131. However, benzyl acetal 130 proved to be unreactive in all attempts at partial cleavage under any of the typical Lewis acid-catalyzed conditions,44 perhaps due to steric encumbrance of the benzylidene acetal moiety.

132

Scheme 60.

As an alternative, triol 103 was treated with neat benzyl bromide in the presence of silver oxide, which selectively protected both primary alcohol motifs to afford dibenzyl ether 132

(Scheme 61). Since TBSOTf/lutidine was not effective in protecting the C20 tertiary hydroxyl of dibenzyl ether 132, the trimethylsilyl ether 133 was instead prepared by heating 132 with trimethylsilyl cyanide and imidazole in dichloromethane in a sealed tube for 3 days. The benzyl ether moieties in the fully protected dibenzyl TMS ether 133 were then cleaved using hydrogenolysis in the presence of 10% palladium on charcoal at 1 atm to afford diol 134. Diol

134 was then treated with DEAD/PPh3 at temperatures ranging from room temperature to reflux in benzene in attempts to cyclize to the oxepane. However, disappointingly, full recovery of the starting material occurred regardless of the conditions used.

133

Scheme 61.

2.2.3.2.3. Attempted Oxepane Formation via Intramolecular Etherification of Halo-

Alcohols

In our route to alstilobanine A (15), we had prepared C18 iodo OTBS ether 118 (cf.

Scheme 53). We believed it might be possible to effect OTBS cleavage of 118 in the presence of a metal salt promoter, such as Ag(I), to generate a C-17 silver alkoxide which might react with the iodomethyl moiety to deliver the cyclic ether 136 (Scheme 62).

Scheme 62.

To this end, OTBS iodide 118 was treated with silver triflate and with silver tetrafluoroborate in dichloromethane at room temperature (Scheme 63). We had hoped that a trace of triflic acid or hydrogen tetrafluoroborate contained in the silver salts would cleave the

134

C17 O-TBS group, while the silver(I) metal cation would activate the C18 alkyl iodide to form the C-O σ bond. However, the desired O-TBS cleavage was not observed, but only led to cyclopropylimine 137 in excellent yield.

Scheme 63.

As an alternative to the above reaction, O-TBS iodide 118 was treated with TBAF in dichloromethane in an effort to generate the C17 alkoxide. However, a retro-aldol reaction occurred following silyl ether cleavage, resulting in the formation of indoline cyclopropane derivative 138, whose structure was established by extensive 2-D NMR experiments, in nearly quantitative yield (Scheme 64).

Scheme 64.

At this point, it appeared to us that the desired seven-membered cyclic ether formation using halo alcohols such as 118 would not be viable unless we could suppress intramolecular cyclopropanation and retro-aldol side reactions. We postulated that protection of the indole nitrogen with an electron-withdrawing group might avoid both pathways by rendering the indole

135 nitrogen less capable of assisting in C3 nucleophilic attack as well as suppressing the formation of an ester enolate as was the case with ester 75 (cf. Scheme 17).

Thus, O-TBS β-lactone 101 was treated with CbzCl and KHMDS to afford the N-Cbz product 139 in high yield (Scheme 65). N-Protected β-lactone 139 was then reduced to diol 140 with lithium borohydride in acceptable yield. An Appel bromination reaction was carried out using diol 139 and carbon tetrabromide after the corresponding iodination failed to produce any of the desired iodide. This transformation, although low yielding, installed the C18 bromine and serendipitously also cleaved the TBS ether, affording the bromomethyl diol 141 in one operation.

We then attempted an intramolecular cyclization of bromo alcohol 141 with silica gel in nitromethane,45 but isolated a single product which was identified as the undesired oxetane 142 after N-Cbz removal to give previously prepared compound 127. Therefore, while we were able to suppress the retro-aldol and cyclopropanation pathways by protecting the indole nitrogen, the substrate was still reluctant to cyclize to the desired oxepane.

Scheme 65.

136

Thus, given our failure to isolate even a trace amount of a seven-membered ether product during the above studies, we were forced to abandon this approach to the oxepane ring of alstilobanine E and angustilodine.

2.2.3.3. Attempted Oxepane Formation via Intramolecular C16 Alkylation

In view of the aforementioned failures, we devised an alternative ring closure strategy in which we envisioned effecting an intramolecular alkylation of the C16 ester enolate with a pendant chloromethyl ether motif as in 144 to furnish the bridging oxepane ring 143. The chloromethylether cyclization precursor 144 would be prepared from the corresponding methylthiomethyl ether (MTM) compound 145, which would, in turn, be conveniently prepared from diol 146 (Scheme 66).

Scheme 66.

The revised route commenced with the removal of the Cbz group of diol 74 using a Pd- catalyzed hydrogenolysis, which afforded the free indole diol 146 quantitatively (Scheme 67).

Diol 146 was transformed to MTM ether 147 using acetic anhydride in anhydrous DMSO46 but only in modest yield. The free tertiary hydroxyl group of MTM ether 147 was protected as the O-

TBS ether 148 in good yield. MTM/silyl ether 148 was treated with fresh sulfuryl chloride47 and immediately with LiHMDS in an effort to induce ring closure of the presumably labile17 intermediate chloromethyl ether 149. Although a variety of conditions were screened, neither the

137 chloromethyl ether nor the oxepane product was observed, but rather a complex mixture was produced.

Scheme 67.

2.2.3.4. Silver(I)-Promoted Cyclic Hemithioacetal Formation/Reduction Approach

2.2.3.4.1. Background and a Revised End-Game Synthetic Plan

During investigative studies toward the brevetoxin natural products, Nicolaou and co- workers45 reported a protocol in which a hydroxyl thioacetal such as 150 had been converted to the medium sized cyclic hemithioacetal 152 upon treatment with a silver(I) salt and sodium bicarbonate in nitromethane in the presence of 4 Å molecular sieves and silica gel (Scheme 68).

This reaction presumably proceeded through the sulfonium intermediate 151. Hemithioacetal

152 was subsequently converted to the corresponding cyclic ether 153 by a radical-mediated tin hydride reduction. Compound 152 could be further elaborated to cyclic ether derivatives such as

155 via a two-step oxidation/substitution sequence involving oxidation to sulfone 154 and subsequent displacement of the sulfone with various nucleophiles in the presence of a Lewis acid.

It was discovered that the counteranion of the silver(I) promoter, as well as the solvent employed,

138 affected the reaction outcome significantly. Silver perchlorate was found to give the optimal yields and solvents with high dielectric constants such as nitromethane facilitated the cyclization, while other common less polor aprotic solvents like dichloromethane inhibited the reaction almost completely. Although the formation of a seven-membered cyclic ether product was not investigated, Nicolaou and co-workers were able to prepare a series of eight and nine membered cyclic hemithioacetals smoothly using this methodology. Despite our earlier failed attempts to form the bridging seven membered ether moiety of the angustilodine natural products using a

Ag(I)-promoted etherification (cf. Scheme 65), Nicolaou’s methodology appeared worth attempting.

Scheme 68.

Our revised retrosynthetic plan would involve a tin hydride radical-mediated reduction of hemithioacetal 157 to afford the oxepane moiety in 156 (Scheme 69). The key silver(I) promoted cyclic hemithioacetal formation would be effected from C16 hydroxymethyl dithioacetal 159 through a nucleophillic attack of the C17 hydroxyl to the sp2-hybridized electrophilic C18 of the

139 sulfonium intermediate 158. In fact, an electronically similar cyclization was previously exemplified by the translactonization of β-lactone 99 to ɛ-lactone 102 (cf. Scheme 38-39). The thioacetal precursor 159 would be prepared via reaction of aldehyde 160 with a thiol in the presence of a Lewis acid. Aldehyde 168 would be readily accessible by functional group modification of the hydroxymethyl aldehyde 104 we had used for earlier synthetic studies (cf.

Scheme 41).

Scheme 69.

2.2.3.4.2. Optimized Preparation of Hydroxy Aldehyde 104

Before embarking on an investigation of this revised end-game strategy, we needed to optimize the preparation of hydroxymethyl aldehyde 104, which had been previously prepared via partial reduction of ɛ-lactone 102 with DIBAL-H in the presence of boron trifluoride etherate at low temperature (cf. Scheme 40). As discussed previously, this “first generation” approach suffered from a lack of reproducibility due to overreduction to the triol 103, with the optimal yields of aldehyde 104 in the range of 15-35%. Furthermore, the translactonization of β-lactone

140

99 to ɛ-lactone 102 was also rather low yielding, which further handicapped accumulation of material for subsequent exploratory studies.

In order to combat the overreduction issue, we first attempted to prepare a C18 N-methyl-

N-methoxyamide derivative such as 162, which could then be chemoselectively reduced to the desired aldehyde (Scheme 70).48 Thus, both the indole nitrogen and primary hydroxyl moiety of intermediate 99 were Boc protected to avoid interference with the organoaluminum reagent employed in the amide formation. The fully protected β-lactone 169 was treated with N,O- dimethylhydroxylamine hydrochloride and dimethylaluminum chloride,49 which unfortunately led to complete decomposition of the starting materials.

Scheme 70.

As an alternative to the failed N-methyl-N-methoxyamide formation, we next investigated the feasibility of converting a β-lactone species to the corresponding thioester, followed by a chemoselective Fukuyama reduction, in which the thioester would be reduced to the aldehyde using a silane in the presence of a palladium catalyst (Scheme 71).50 To avoid a potential translactonization side reaction (cf. Scheme 38) that might arise under the thioesterification conditions, we chose OTBS β-lactone 101 (cf. Scheme 52) as the starting substrate. Attempted thioester formation on the O-TBS β-lactone 101 using tert-butylthiol and ethanethiol with trimethylaluminum,51 however, resulted in complex product mixtures. After a

141 screening of other thioesterification conditions, a serviceable route to 163 was discovered by reacting O-TBS β-lactone 101 with ethanethiol and triethylamine to furnish the corresponding thioester 163 in acceptable yield. Addition of a small amount of acetic acid, which presumably facilitates proton transfer significantly increased the reaction rate. The resultant thioester 163 was treated with triethylsilane and 10% Pd/C at room temperature for three days, during which time additional palladium catalyst and silane were added periodically to give the desired OTBS aldehyde 164 in 80% yield on a 13 mg scale. However, this reduction was not amenable to scaleup. In one case on a 50 mg scale, for instance, despite running the reaction for four days, the aldehyde product was isolated in an unsatisfactory 42% yield along with recovered starting material 163.

Scheme 71.

This reproducibility problem was eventually solved by adopting β-lactone reduction/selective oxidation strategy. We had initially avoided any oxidations of a free NH indole substrate such as 117 due to the potential for competing indole oxidation. However, recently Zhu and co-workers reported the successful oxidation of an indole alcohol to the corresponding aldehyde using IBX in DMSO in a total synthesis of alstoscholarine.52 This report inspired us to investigate the feasibility of using IBX as the oxidation agent on our indole alcohol substrates. Gratifyingly, after 12 h at room temperature, O-TBS diol 117 was cleanly converted to aldehyde 164 via treatment with IBX53 in anhydrous DMSO (Scheme 72). The TBS group was

142 cleaved from aldehyde 164 by stirring with 1 M HCl in ethyl acetate to afford hydroxymethyl aldehyde 104 in 94% yield over two steps. Compared to our previous translactonization/partial reduction approach, which provided aldehyde 104 in 11-26% yield over two steps from β-lactone

99, this second generation approach furnished aldehyde 104 from β-lactone 99 in a reproducible

61% yield over four steps. Furthermore, the transformations in the new route are amenable to scaleup.

Scheme 72.

2.2.3.4.3. Preparation of the Ag(I)-Promoted Cyclization Precursor Dithioacetal 173

With a reliable route to prepare sufficient quantities of aldehyde 104 in hand, we could continue our studies toward exploiting the Nicolaou methodology to access the seven-membered bridging ether in angustilodine and alstilobanine E. Thus, aldehyde 104 was first treated with ethanethiol in the presence of boron trifluoride etherate (Scheme 73), conditions we had employed successfully for the preparation of cyclic dithiane 105 from 104 using ethanedithiol (cf.

Scheme 41). Surprisingly, we were unable to isolate the desired dithioacetal 165, with the reaction providing a mixture of at least five different products. Attempted isolation of these products via column chromatography on silica gel led to decomposition of some of the compounds and thus we were unable to ascertain whether any useful transformation(s) had taken place.

143

We postulated that the Lewis acid, boron trifluoride etherate, might be too reactive for an acyclic dithioacetal formation from a sensitive substrate such as 104. In fact, Nicolaou and co- workers had previously reported similar difficulties in generating dithioacetals from precursors which contained hydroxyl groups and aldehyde functionality six or seven atoms removed from one another.45 However, it was found that titanium tetrachloride was the only effective Lewis acid in these cases, with all other reagents screened, including boron trifluoride etherate, failing to give any of the desired products.

Indeed, replacement of boron trifluoride with titanium tetrachloride,54 led to the formation of hydroxymethyl dithioacetal 165 from aldehyde 104 in excellent yield. Dithioacetal

165 was also found to be stable on silica gel chromatography.

Scheme 73.

2.2.3.4.4. Ag(I)-Promoted Cyclization of C16 Hydroxymethyl Dithioacetal 173 and

Formation of the Key Cyclic Hemithioacetals

With compound 165 in hand, we decided to first test the Nicolaou cyclization on the unprotected indole diol substrate. Thus, the free NH indole hydroxymethyl dithioacetal 165 was

144 stirred with activated 4Å molecular sieves, dry silica gel and sodium bicarbonate at room temperature in nitromethane, then treated with a solution of silver perchlorate in nitromethane

(Scheme 74). The slightly yellow reaction mixture turned dark red shortly after the addition of the silver perchlorate, followed by a complete consumption of the starting material as monitored by TLC within an hour. Gratifyingly, the desired seven-membered cyclic hemithioacetal was isolated as a chromatographically separable 3:1 mixture of C18 diastereomers 167a and 167b in

86% total yield. The structures of 167a and 167b were established unambiguously through extensive 2-D NMR experiments. This transformation cleanly afforded the desired seven- membered ring products, without any detectable amount of four-membered or three-membered ring products resulting from attack of the C20 hydroxyl group or C3 of the indole on sulfonium ion intermediate 166.

Scheme 74.

It is noteworthy that the work-up procudure was found to be crucial to the success of this cyclization. In a typical operation, once consumption of starting dithioacetal 165 was indicated

145 by TLC, the reaction mixture was diluted with dichloromethane and ethyl acetate. This dilution caused precipitation of excess silver perchlorate from the reaction mixture, which could subsequently be removed together with molecular sieves and sodium bicarbonate via filtration of the mixture through a pad of Celite. Failure to perform this dilution resulted in removal of only the molecular sieves and sodium bicarbonate through filtration while leaving the silver perchlorate in the filtrate with the cyclized products 167a/b. A rapid ring opening reaction of

167a/b was observed under these conditions, which was probably promoted by silver perchlorate and trace amounts of water, affording hydroxymethyl aldehyde 104 (Scheme 75).

Scheme 75.

2.2.3.4.5. Reduction of the Thiohemiacetals to the Oxepane

We planned to prepare the seven-membered bridging ether from cyclic hemithioacetals

167a/b through a radical mediated reduction similar to that which was employed by Nicolaou.45

Since the C18 diastereoisomerism of 167a/b was inconsequential in the synthesis, these diastereomers were used either as a mixture or as a single diastereomer in subsequent transformations, all of which gave essentially identical results. Thus, 167 was heated at 120 °C in toluene with triphenyltin hydride and AIBN in a sealed tube (Scheme 76). However, the starting thiohemiacetals were unreactive under these conditions, and none of the desired reduction product 136 was observed.

146

Scheme 76.

We then turned our attention to another method Nicolaou and co-workers had used to desulfurize hemithioacetal systems via an oxidation/reduction method, in which the thioethyl group was first converted to a sulfone derivative and then subsequently reduced under Lewis acid catalysis.45 To this end, hemithioacetal 167 was first oxidized with excess mCPBA at 0 oC to afford sulfone 168 in acceptable yield (Scheme 77). The free indole moiety did not interfere with the desired oxidation. However, we were disappointed to find that treatment of 168 with DIBAL-

H and a catalytic amount of boron trifluoride etherate, identical to the conditions employed by

Nicolaou, led to reduction of the methyl ester to the corresponding primary alcohol while the sulfone moiety remained untouched, thus affording hydroxymethyl sulfone 169 as the only isolable product. Replacement of DIBAL-H with triethylsilane led to a complex mixture of products. Reduction with lithium naphthalenide in THF was similarly unsuccessful, leading to decomposition of the starting material. Likewise, the mild detosylation conditions involving

Mg0/MeOH that we had used successfully in our total synthesis of alstilobanine A (15) (cf.

Scheme 52) was applied to substrate 168 in an attempt to cleave both the ethylsulfonyl and tosyl moieties in one operation. However, sonication of 168 with excess magnesium turnings in anhydrous methanol afforded only the detosylation product 170 which contained the ethylsulfonyl group untouched.

147

Scheme 77.

We next attempted a hemithioacetal desulfurization method using Raney nickel55 under conditions that had been successfully utilized for the desulfurization of cyclic dithiane 105 in our earlier synthetic studies toward alstilobanine A (15) (cf. Scheme 41). To this end, hemithioacetal

167 was stirred with freshly prepared W-2 Raney nickel31 in ethanol at room temperature, which resulted only in recovery of most of the starting material. Raising the reaction temperature to 60 oC led to a product that was identified by NMR and mass spectrometry as the unusual hexacyclic acetal 171 (Scheme 78). Acetal 171 undoubtedly arises via the oxonium intermediate 172 which is generated from the initial ethylthio cleavage occurred on the surface of the Ni catalyst.

Scheme 78.

148

We had hoped that the oxonium intermediate 172 which might be generated from 167 using Lewis acid-catalysis could be trapped and converted to the desired oxepane by reducing agents. Thus, hemithioacetal 167 was treated with a variety of Lewis acid and reductant combinations. However, none of these reactions gave any of the desired product 136, resulting in most cases only in recovery of the starting material (Scheme 79). Attempted reduction of cyclic acetal 171 via zinc boronhydride in ethanol was similarly unsuccessful.

Scheme 79.

In the hope of reducing the intermediate oxonium ion 172, the hemithioacetal mixture

167 was treated with W-2 Raney nickel in a hydrogen atmosphere, but at 500 psi for 16 h until all the starting material had been consumed. Indeed, from this reaction we were able to isolate the desired oxepane product 136 in moderate yield, along with some of the acetal side product

171. However, further efforts to optimize this desulfurization reaction were unsuccessful. In one run, we could isolate the desired oxepane compound 136 in 72% yield along with a minimal amount of acetal side product (Scheme 80). However, using a different batch of Raney nickel prepared in apparently the same fashion afforded predominantly the acetal product 171. The nature of the Raney nickel appeared to be critical to the success of the desulfurization, but was found to be very difficult to control. Various forms of Raney nickel catalysts including W-1, W-

149

2, W-4, W-5 and triethylsilane-poisoned W-2 and W-5 were also prepared and tested. However, in no case were we able to reproduce the 72% yield of 136 but rather most runs afforded acetal product 171 or led to decomposition of the starting material.

Scheme 80.

In our earlier attempts to optimize the preparation of aldehyde 104, we had successfully reduced the C18 thioester 163 to aldehyde 164 using the Fukuyama reduction protocol using triethylsilane in the presence of 10% palladium on charcoal50 (cf. Scheme 71). We thus sought to apply this procedure to hemithioacetal 167 (Scheme 81). Since the Fukuyama reduction procedure generally utilizes triethylsilane and 10% palladium on charcoal in THF, we subjected hemithioacetal 167 to these conditions, again resulting in starting material recovery. Since it is known that the Fukuyama reduction is sometimes solvent dependant,50 we next carried out the reduction of 167 in the more polar solvent acetone at room temperature. Gratifyingly, oxepane

136 was isolated in 30% yield along with recovery of 50% of the starting material 167. Most importantly, the acetal side product 171 was not observed in this reaction. Optimal conditions were eventually found involving heating the reaction mixture at reflux along with portionwise addition of 10% palladium on charcoal and triethylsilane over 6-8 hours, affording oxepane 136 in high 94% yield.

150

Scheme 81.

2.2.3.5. Completion of the Total Syntheses of Racemic Alstilobanine E and Angustilodine

With oxepane 136 in hand, we were now able to complete the total syntheses of alstilobanine A and angustilodine (Scheme 82). Removal of the tosyl group from the piperidine nitrogen was accomplished in good yield by sonicating oxepane 136 with magnesium turnings in anhydrous methanol at room temperature. Treatment of the free piperidine 14 with TFA produced a TFA salt which was identical in NMR spectra to the natural alstilobanine E (14) isolated by Morita.13

Reductive amination of alstilobanine E (14) using paraformaldehyde in ethanol in the presence of sodium borohydride and a catalytic amount of acetic acid56 afforded the N- methylated derivative, angustilodine (13), in 81% yield. Synthetic angustilodine (13) had identical 1H and 13C NMR spectra to natural material isolated by Kam and co-workers as determined by a direct comparison with authentic spectra provided by Professor Kam.57 A

NOESY experiment using synthetic 13 revealed a strong flagpole interaction between H14 and

H21 on the piperidine moiety, which confirmed its disposition in a boat conformation in accordance with Morita’s report on alstilobanine E (14).13

151

Scheme 82. Completion of the total syntheses of alstilobanine E (14) and angustilodine (13)

152

2.3. Concluding Remarks

In summary, we have completed the first total syntheses of the unusual monoterpenoid indole alkaloids angustilodine (13), alstilobanine E (14) and alstilobanine A (15) in racemic form.

The overall route is summarized in Scheme 83.

Our divergent route toward the angustilodine-type alkaloids hinged on an unprecedented intermolecular conjugate addition of the enolate of a functionalized indole ester 44 to the transient nitrosoalkene intermediate 25 derived from a α-chloropiperidone oxime 39 to construct the C15/C16 σ bond, affording a diastereomeric mixture of adducts 47 and 82. These Michael adducts were elaborated to the C15/C16 anti-ketoacid 57 through a series of highly efficient transformations including a reduction of the indole-3-oxalyl to a C3 acetate moiety, an epimerization of the C15/C16 syn-diastereomer to the corresponding anti-isomer, a C20 deoximation, and a TFA-mediated TMSE-ester cleavage. The cis-fused azadecalin moiety with the requisite C15, C19, C20 configuration found in the target alkaloids was formed from ketoacid 57 via a key Romo NPAL reaction, which afforded β-lactone 72 in excellent yield and with high cis-azadecalin diastereoselectivity. The C16 hydroxymethyl group was installed with the requisite configuration using the ester enolate of indole NH β-lactone 89a and monomeric formaldehyde to afford C16 hydroxymethyl β-lactone 99. β-Lactone 99 was subsequently tranformed into the C17 OTBS diol 117 which serves as a key common intermediate towards these alkaloids.

153

Scheme 83. Summary of total syntheses of racemic angustilodine, alstilobanine A and E

154

In the synthesis of racemic alstilobanine A (15), diol 117 was C18 deoxygenated using an

Appel iodination/Pd-catalyzed reduction sequence to afford the C19 methyl compound 108.

Smooth tosyl and TBS removal afforded alstilobanine A (15) in 19 steps overall from indole 44 in 21% total yield.

After numerous unproductive attempts to form the oxepane moiety found in angustilodine and alstilobanine E, a viable and efficient route to this substructure was eventually established. Thus, diol 117 was first converted to the C16 hydroxymethyl dithioacetal 165 using an IBX-oxidation/OTBS removal/dithioacetal formation sequence. Dithioacetal 165 was subjected to a silver(I) promoted cyclic hemithioacetal formation procedure similar to that developed by Nicolaou, et al., which afforded the seven-membered hemithioacetals 167 in excellent yield. Following a Pd-catalyzed-silane-mediated reductive thioethyl cleavage, the oxepane moiety found in 13 and 14 was sucessfully installed, affording the key oxepane 136 in excellent yield. Tosyl removal of 136 resulted in racemic alstilobanine E (14) in 21 steps from indole 44 in 16% total yield. N-Methylation of 14 via reductive amination afforded racemic angustilodine (13) in 22 steps from 44 in 13% total yield.

In summary, we have shown that nitrosoalkene conjugate additions can be a powerful

C-C bond forming strategy. The hitherto underdeveloped nitrosoalkene Michael addition of enolates constitutes a powerful umpolung methodology which we expect will find increasing use in synthesis of complex natural products, especially those bearing embedded 1,4-dicarbonyl moieties.

155

Scheme 83. (continued)

156

Chapter 3. Experimental Section

General Methods. All non-aqueous reactions were carried out in oven- or flame-dried glassware under an argon atmosphere. All chemicals were purchased from commercial vendors and used as is, unless otherwise specified. Tetrahydrofuran (THF) and dichloromethane (CH2Cl2) were obtained from a solvent purification system (Glass Contour). Reactions were magnetically stirred and monitored by thin layer chromatography (TLC) with 250 μm EMD 60 F254 precoated silica gel plates. Flash column chromatography was performed using Silicycle silica gel P60 (230-400 mesh). Sonication reactions were performed in a Branson 1510 sonicator. 1H and 13C NMR spectral data were recorded on Bruker DPX-300, AMX-360 or DRX 400 MHz spectrometers. Chemical shifts are reported relative to chloroform (δ 7.26), methanol (δ 3.31), or

DMSO (δ 2.50) for 1H NMR and chloroform (δ 77.2), methanol (δ 49.0), or DMSO (δ 39.5) for

13C NMR.

Synthesis of Nitrosoalkene Michael Adducts anti-42 and syn-87. To a -78 °C solution of indole 5 (8.18 g, 22.6 mmol) was added LiHMDS (47.5 mL, 47.5 mmol, 1.0 M in THF) and the resulting solution was stirred for 30 min. A solution of α-chlorooxime 7 (9.6 g, 31.7 mmol) in

THF (69 mL) was added via cannula over 10 min. The reaction mixture was stirred at -78 °C for

2 h and then diluted with aqueous NH4Cl. The organic layer was separated and the aqueous layer

157 was extracted with EtOAc. The combined organic layers were dried over MgSO4 and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel

(gradient 20-50% EtOAc/hexanes) to yield diastereomeric Michael adducts 42 and 87 (14.01 g,

99%, ~1.2 : 1 by 1H NMR) which were carried on to the next step without separation. 1H NMR

(300 MHz, CDCl3) δ 10.31 (s, 0.5H), 10.23 (s, 0.5H), 8.90 (s, 0.5H), 8.66 (s, 0.5H), 7.79-7.58 (m,

3H), 7.39-7.17 (m, 5H), 5.59 (d, J = 5.9 Hz, 0.5H), 5.13 (d, J = 8.6 Hz, 0.5H), 4.93 (d, J = 14.5

Hz, 0.5H), 4.60-4.50 (m, 2H), 3.69-4.47 (m, 4H) 3.37 (p, J = 5.7 Hz, 0.5H), 3.20 (d, J = 14.6 Hz,

0.5 H); 3.10 (q, J = 7.3 Hz, 0.5H), 2.97-2.85 (m, 1H), 2.44-2.28 (m, 3H), 2.08-2.03 (m, 0.5H),

13 1.78-1.76 (m, 0.5H), 1.41-1.31 (m, 0.5H), 0.11-0.07 (m, 9H); C NMR (75 MHz, CDCl3) δ

183.6, 183.0, 172.6, 171.9, 166.5, 166.5, 152.8, 152.7, 144.61, 144.57, 143.5, 143.4, 135.83,

135.80, 130.4, 130.3, 128.06, 127.96, 125.8, 125.6, 124.4, 124.2, 123.4, 123.2, 120.0, 119.6,

112.9, 112.7, 111.2, 111.1, 65.33, 65.29, 53.3, 53.2, 45.4, 44.9, 44.2, 43.4, 42.9, 42.7, 42.3, 41.9,

27.9, 27.8, 22.0, 21.9, 17.8, 14.6, 14.2, -1.1; LRMS-ES+ m/z (relative intensity) 666 (M+K+,

+ 100); HRMS-ES+ (C30H41N4O8SSi) calcd 645.2414 (M+NH4 ), found 645.2445.

For characterization of the isomers, a small amount of the mixture was separated by flash chromatography on silica gel (gradient 2-10% Et2O/ CH2Cl2) to afford 42 (less polar) as an

1 orange foamy solid. H NMR (400 MHz, CDCl3) δ 9.96 (s, 1H, NH), 7.91 (s, 1H), 7.62(d, J =

8.0 Hz, 3H), 7.38 (m, 1H), 7.26-7.19 (m, 4H), 5.52 (d, J = 6.1.1 Hz, 1H), 4.87 (d, J = 14.6 Hz,

1H), 4.50 (t, J = 8.2 Hz, 2H), 3.64 (s, 3H), 3.30 (m, 1H), 3.18 (d, J = 14.7 Hz, 1H), 2.82 (m, 1H),

2.43 (m, 1H), 2.34 (s, 3H), 2.02 (dd, J = 3.4, 12.8 Hz, 1H), 1.38 (m, 1H), 1.15 (t, J = 8.7 Hz, 2H),

13 0.07 (s, 9H); C NMR (75 MHz, CDCl3) δ 183.2, 171.5, 166.0, 153.1, 144.2, 135.3, 133.0,

130.0, 127.7, 125.3, 123.9, 122.9, 119.6, 112.4, 111.1, 64.9, 52.9, 45.0, 43.0, 42.6, 42.0, 27.5,

1 21.6, 17.5, -1.4. 87 (more polar, brown foamy solid), H NMR (300 MHz, CDCl3) δ 9.84 (s, 1H),

158

7.97 (s, 1H), 7.61-7.57 (m, 3H), 7.29-7.23 (m, 2H), 1.17-7.14 (m, 3H), 5.00 (d, J = 8.4 Hz, 1H),

4.50-4.39 (m, 3H), 3.67 (s, 3H), 3.45-3.38 (m, 3H), 2.94 (m, 1H), 2.82 (m, 1H), 2.34 (s, 3H),

13 1.69 (m, 2H), 1.09-1.05 (m, 2H), 0.00 (s, 9H); C NMR (75 MHz, CDCl3) δ 182.9, 172.8, 166.2,

152.9, 144.5, 143.5, 135.7, 133.3, 130.3, 128.1, 125.8, 124.4, 123.4, 120.2, 122.6, 111.2, 65.3,

53.3, 45.0, 44.1, 43.3, 42.0, 28.0, 22.0, 17.8, -1.1.

Synthesis of α-Ketoester 48. To a solution of ester oxime mixture 42/87 (5.66 g, 9.01 mmol) and imidazole (2.47 g, 36.3 mmol) in CH2Cl2 (240 mL) was added TBSCl (4.10 g, 27.2 mmol). The resulting suspension was stirred for 12 h at rt and then diluted with 1 M HCl. The organic layer was separated and the aqueous layer was extracted with CH2Cl2. The combined organic layers were dried over MgSO4 and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (30% EtOAc/hexanes) yielding O-TBS-oximes 48

1 (5.98 g, 89%) as an orange solid. H NMR (300 MHz, CDCl3) δ 9.91 (s, 0.5H), 9.79 (s, 0.5H),

7.77-7.61 (m, 3H), 7.41-7.21 (m, 5H) 5.56 (d, J = 5.7, 0.5H), 5.06-4.97 (m, 1H), 4.84 (d, J = 15.0,

0.5H), 4.54-4.45 (m, 2H), 3.69-3.60 (m, 4H), 3.41 (dt, J = 7.1, 12.6 Hz, 0.5H), 3.30 (d, J = 15.0,

0.5H), 3.18 (d, J = 14.9, 0.5H), 3.00-2.92 (m, 1H), 2.85-2.70 (m, 0.5H), 2.42 (s, 1.5H) 2.32 (s,

1.5H), 2.06-1.98 (m, 0.5H), 1.85-1.70 (m, 0.5H), 1.65-1.55 (m, 0.5H), 1.40-1.35 (m, 1H), 1.20-

13 1.13 (m, 2H), 0.99 (m, 9H), 0.31-0.20 (m, 6H), 0.10-0.07 (m, 9H); C NMR (75 MHz, CDCl3) δ

183.4, 182.8, 172.6, 171.0, 166.3, 166.1, 157.9, 156.8, 144.40, 144.37, 143.6, 143.4, 135.5, 135.4,

159

133.9, 133.6, 130.20, 130.15, 128.0, 127.9, 126.0, 125.7, 124.4, 124.2, 123.5, 123.2, 120.6, 120.0,

112.4, 112.2, 111.3, 65.1, 53.2, 53.0, 45.2, 44.1, 43.5, 43.3, 42.9, 28.1, 27.7, 26.34, 26.28, 21.93,

21.85, 18.3, 18.2, 17.79, 17.76, -1.1, -4.2, -4.6, -4.8; LRMS-ES+ m/z (relative intensity) 780

+ + (M+K , 100); HRMS-ES+ (C36H52N3O8SSi2) calcd 742.3014 (MH ), found 742.3005.

Synthesis of Acetates 49a and 85. To a stirred solution of α-ketoesters 48 (21.5 g, 29 mmol) in MeOH (55 mL) and THF (425 mL) cooled to 0 °C was added NaBH4 (1.32 g, 34.7 mmol). The resulting solution was stirred for 1 h at 0 °C, and then diluted with aqueous NH4Cl.

The organic layer was separated and the aqueous layer was extracted with CH2Cl2. The combined organic layers were washed with water and brine and then dried over MgSO4.

Removal of the solvent in vacuo provided the indole-3-hydroxyl acetate products as a white powder.

Without further purification, this mixture was dissolved in a 1:1 (v/v) mixture of

Ac2O:pyridine (292 mL) and the solution was stirred at rt for 12 h. After removing the volatiles in vacuo, 49a and 85 were separated by flash chromatography on silica gel (gradient 15-40%

EtOAc/hexanes). 85 (more polar isomer, orange foamy solid, 12.60 g, 55%, ~3:1 mixture of

1 acetoxy diastereomers): H NMR (400 MHz, CDCl3) δ 9.07 (s, 0.75H), 9.03 (s, 0.25H), 7.75-

7.71 (m, 1H), 7.64 (d, J = 8.2 Hz, 2H), 7.32-7.13 (m, 5H), 6.21 (s, 0.25H), 6.18 (s, 0.75H), 4.97-

160

4.88 (m, 1H), 4.52 (d, J = 5.9 Hz, 0.25H), 4.48 (d, J = 5.5 Hz, 0.75H), 4.28-4.20 (m, 1H), 4.09-

4.02 (m, 1H), 3.67-3.55 (m, 4H), 3.41-3.31 (m, 2H), 3.00-2.93 (m, 1H), 2.39-2.36 (m, 3H), 2.14

(s, 3H), 1.95-1.70 (m, 3H), 1.60-1.40 (m, 2H), 0.98 (s, 9H), 0.29-0.18 (m, 6H), 0.00--0.02 (m,

13 9H); C NMR (90 MHz, CDCl3) δ 171.0, 170.8, 170.6, 170.5, 169.1, 157.3, 144.0, 135.3, 135.2,

133.7, 133.5, 132.2, 131.7, 129.8, 127.6, 126.2, 126.1, 122.7, 120.3, 119.3, 111.2, 108.0, 107.8,

67.7, 64.2, 53.4, 52.5, 44.8, 43.2, 42.9, 42.7, 42.4, 42.1, 28.0, 26.0, 25.9, 21.5, 20.9, 20.7, 17.9,

17.3, -1.56, -1.59, -4.8, -5.1; 49a (less polar isomer, orange foamy solid, 8.29 g, 36%, ~2:1

1 mixture of acetoxy diastereomers): H NMR (300 MHz, CDCl3) δ 8.73-8.71 (m, 1H), 7.76 (d, J

= 7.6 Hz, 1H), 7.66 (d, J = 8.2 Hz, 1H), 7.35-7.14 (m, 5H), 6.24-6.22 (m, 1H), 5.02 (t, J = 7.1 Hz,

1H), 4.34-4.07 (m, 3H), 3.66-3.53 (m, 4H), 3.15-3.07 (m, 2H), 2.72 (d, J = 11.3, 3.2 Hz, 1H),

2.45 (s, 3H), 2.20-2.18 (m, 3H), 1.58-1.39 (m, 2H), 1.00-0.89 (m, 12H), 0.26 (s, 3H), 0.20 (s,

13 3H), 0.02--0.04 (m, 9H); C NMR (75 MHz, CDCl3) δ 173.1, 173.0, 170.9, 170.8, 169.3, 169.2,

156.7, 144.4, 136.1, 136.0, 133.8, 133.6, 132.4, 132.2, 130.2, 128.1, 126.7, 123.5, 123.4, 121.1,

120.2, 111.5, 108.54, 108.50, 68.4, 68.0, 64.7, 64.5, 53.9, 53.0, 52.9, 45.4, 44.4, 44.1, 42.8, 42.6,

42.5, 28.5, 28.0, 26.4, 22.0, 21.2, 21.1, 18.4, 17.7, 17.6, -1.2, -4.75, -4.89, -4.92; LRMS-ES+

(mixture of diastereomers 49a/85) m/z (relative intensity) 824 (M+K+, 25); HRMS-ES+

+ (C38H56N3O9SSi2) calcd 786.3276 (MH ), found 786.3286.

161

Synthesis of Indole Diester 84. To a solution of acetate 85 (2.94 g, 3.75 mmol) in t-

BuOH (80 mL) was added 10% Pd/C (1.20 g). The resulting mixture was evacuated and backfilled with H2 from a balloon and TEA (10.0 mL) was then added. The resulting mixture was warmed to 30 °C and stirred for 4 days until all the starting material was consumed as judged by TLC. The reaction mixture was then diluted with EtOAc, filtered through a pad of

Celite and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 10-25% EtOAc/hexanes) to afford indole 84 (2.47 g, 91%) as an off-white

1 foam. H NMR (360 MHz, CDCl3) δ 8.74 (br s, 1H), 7.62 (d, J = 7.2 Hz, 2H), 7.57 (d, J = 7.8 Hz,

1H), 7.28-7.24 (m, 3H), 7.17 (dd, J = 7.1, 7.5 Hz, 1H), 7.09 (dd, J = 7.3, 7.4 Hz, 1H), 4.82 (d, J

= 15.1 Hz, 1H), 4.33 (d, J = 6 Hz, 1H), 4.11-4.07 (m, 2H), 3.64 (s, 3H), 3.60 (d, J = 4.4 Hz, 2H),

3.58-3.55 (m, 1H), 3.36 (d, J = 15.2 Hz, 1H), 3.30-3.25 (m, 1H), 2.90 (td, J = 4.0, 11.9 Hz, 1H),

2.36 (s, 3H), 1.89-1.84 (m, 1H), 1.59-1.54 (m, 1H), 0.94(s, 9H), 0.21(s, 3H), 0.15(s, 3H); 13C

NMR (75 MHz, CDCl3) δ 171.7, 171.6, 157.2, 144.0, 135.4, 133.7, 130.0, 129.9, 129.5, 128.9,

127.7 (2C), 122.3, 119.7, 118.9, 111.0, 107.6, 63.2, 52.5, 44.7, 43.0, 42.7, 42.3, 30.8, 27.6, 26.0,

21.6, 18.0, 17.4, -1.4, -4.8, -5.0; LRMS-ES+ m/z (relative intensity) 766 (M+K+, 75); HRMS-

+ ES+ (C36H57N4O7SSi2) calcd. for 745.3487 (M+NH4 ), found 745.3478.

162

Synthesis of Indole Diester 50a. Following a similar procedure described for preparation of 84, acetate 49a (3.95 g, 5.03 mmol) was reduced by catalytic hydrogenation to afford indole

1 diester 50a as an off white foamy solid (3.16 g, 86%): H NMR (400 MHz, CDCl3) δ 8.19 (s,

1H), 7.65 (d, J = 8.5 Hz, 2H), 7.57 (d, J = 7.7 Hz, 1H), 7.32 (d, J = 8.0 Hz, 1H), 7.26 (d, J = 7.6

Hz, 1H), 7.17 (dd, J = 7.0, 7.1 Hz, 1H), 7.11 (dd, J = 7.1, 7.2 Hz, 1H), 5.03 (d, J = 14.4 Hz, 1H),

4.17-4.13 (m, 2H), 4.06 (d, J = 10.7 Hz, 1H), 3.64 (app. s, 2H), 3.62 (s, 3H), 3.57 (m, 1H), 3.06-

3.00 (m, 2H), 2.66 (td, J = 3.8, 11.5 Hz, 1H), 2.44 (s, 3H), 1.46 (m, 1H), 1.39 (m, 1H), 0.97 (s,

13 9H), 0.23 (s, 3H), 0.16 (s, 3H); C NMR (75 MHz, CDCl3) δ 173.0, 171.4, 156.5, 144.0, 135.7,

133.4, 129.8, 128.0 (2C), 127.8, 122.7, 120.1, 119.2, 111.0, 108.3, 63.3, 52.5, 45.2, 43.8, 42.5,

42.0, 30.6, 27.8, 26.1, 25.9, 21.7, 18.1, 17.5, -1.4, -5.1, -5.2; HRMS-ES+ (C36H54N3O7SSi2) calcd 728.3221 (MH+), found 728.3224.

Epimerization of Ester 84 to Ester 50a. To a solution of diester 84 (4.45 g, 6.12 mmol) in THF (81 mL) cooled to -78 oC was added dropwise KHMDS solution (0.5 M in toluene, 13.5 mL, 6.73 mmol, 1.1 equiv.). After the addition was complete, the cooling bath was removed and the reaction mixture was stirred at rt for 30 min. To the resultant olive green solution was added glacial acetic acid (0.40 mL) followed by saturated aqueous NH4Cl. The mixture was extracted with EtOAc. The organic phase washed with brine, dried over Na2SO4 and concentrated. The residue was purified by flash chromatography on silica gel (30 % EtOAc/hexanes) to afford

163 diester 50a as a slightly yellow foam (3.23 g, 73%), which was identical to ester 50a prepared from hydrogenation of 49a.

Synthesis of N-Cbz Derivative 53a. A stirred solution of indole 50a (1.05 g, 1.45 mmol) in acetonitrile (40 mL) was heated to 90 oC. Dibenzyl dicarbonate (2.07 g, 7.24 mmmol) was added, followed immediately by DMAP (0.53 g, 4.33 mmol). After gas evolvution stopped (~1 min), the reaction mixture was heated at 90 oC for another 2 min, and cooled to rt. The solvent was removed in vacuo to give a pale brown residue, which was purified by flash chromatography on silica gel (5 to 20 % EtOAc/hexanes) to afford N-protected indole 53a as a white foam (1.23 g,

1 99%, ~2:1 Cbz rotamers). H NMR (400 MHz, CDCl3) δ 8.11 (d, J = 7.6 Hz, 0.7H), 8.00 (m,

0.3H), 7.64 (d, J = 6.9 Hz, 2H), 7.57 (d, J = 6.4 Hz, 0.7H), 7.46-7.26 (m, 9H), 5.68 (d, J = 10.1

Hz, 0.3H), 5.17 (d, J = 11.7 Hz, 0.7H), 5.52-5.46 (m, 0.7H), 5.29 (d, J = 13.1 Hz, 0.7H), 5.16 (d,

J = 9.3 Hz, 1.7H),4.14-4.10 (m, 2H), 3.90 (d, J = 15.9 Hz, 0.3H), 3.65-3.53 (m, 2.7H), 3.48 (s,

3H), 3.11 (m, 1H), 2.90 (d, J = 13.2 Hz, 0.3H), 2.68 (d, J = 13.2 Hz, 0.7H), 2.60 (m, 0.3H), 2.49-

2.44 (m, 3H), 2.17 (m, 0.7H), 1.40 (m, 0.3H), 1.12 (m, 1.7H), 0.98 (m, 9H), 0.22 (s, 2H), 0.19-

13 0.17 (m, 4H),0.02 (s, 9H); C NMR (75 MHz, CDCl3) δ 171.7, 171.1, 170.8, 170.1, 155.8, 155.6,

151.8, 151.3, 143.8, 135.9, 135.6, 134.8, 133.4, 132.5, 132.2, 129.8, 129.4, 129.2, 129.0, 128.8,

127.9, 125.3, 123.5, 119.3, 118.9, 118.6, 117.2, 115.8, 69.3, 68.6, 63.6, 63.4, 52.1, 45.7, 42.8,

42.4, 42.0, 41.3, 40.7, 30.9, 30.2, 28.6, 27.5, 26.2, 25.6, 21.7, 18.3, 17.5, -1.5, -5.1; LRMS-ES+

164

+ + + m/z (relative intensity) 862 (MH , 90); HRMS-ES (C44H60N3O9SSi2) calcd 862.3589 (MH ), found 862.3585.

Synthesis of Oxime 54a. To a stirred solution of O-TBS-oxime 53a (12.9 g, 14.9 mmol) in THF (600 mL) was added AcOH (6.6 mL) followed by TBAF (22.1 mL, 22.1 mmol, 1.0 M in

THF). The resulting solution was stirred at rt overnight and then diluted with aqueous NH4Cl.

The organic layer was separated and the aqueous layer was extracted with EtOAc. The combined organic layers were dried over MgSO4 and concentrated in vacuo to give a residue which was purified by flash column chromatography on silica gel (30% EtOAc/hexanes) to yield free oxime

1 54a as a pale foam (11.1 g, 100%, ~2:1 Cbz rotamers). H NMR (400 MHz, CDCl3) δ 8.10 (m,

0.7H), 8.0-7.8 (m, 1.3H), 7.65 (d, J = 7.1 Hz, 2H), 7.54 (m, 0.7H), 7.45 (m, 3.5H), 7.35-7.26 (m,

6H), 5.68 (m, 0.3H), 5.57 (d, J = 11.5 Hz, 0.7H), 5.47 (s, 0.7H), 5.18 (m, 1.4H), 5.07 (m, 1H),

4.13 (m, 2H), 3.87 (m, 0.3H), 3.65-3.59 (m, 3H), 3.48 (s, 3H), 3.09 (m, 1H), 2.91 (0.3H), 2.68 (d,

J = 13.1 Hz, 0.7H), 2.60 (m, 0.3H), 2.48-2.44 (m, 3H), 2.19 (m, 0.7H),, 1.64 (m, 0.4H), 1.40 (m,

13 0.3H), 1.17 (1.60H), 0.94 (t, J = 8.8 Hz, 2H), 0.01 (s, 9H); C NMR (75 MHz, CDCl3) δ 172.0,

171.5, 170.8, 170.5, 152.9, 151.8, 151.4, 143.9, 135.8, 135.5, 134.7, 133.0, 132.3, 132.0, 129.8,

129.4, 129.2, 129.1, 129.0, 128.8, 128.6, 127.9, 127.8, 127.1, 125.3, 124.9, 123.5, 123.2, 119.3,

118.6, 117.2, 116.1, 115.9, 69.5, 68.7, 65.4, 64.5, 63.8, 63.4, 52.6, 52.3, 45.7, 42.4, 41.8, 41.0,

40.7, 40.5, 31.0, 30.7, 30.3, 28.5, 27.6, 21.6, 21.1, 19.2, 17.3, 13.8, -1.5; LRMS-ES+ m/z

165

+ + (relative intensity) 748 (MH , 75); HRMS-ES+ (C38H46N3O8SSi2) calcd 748.2724 (MH ), found

748.2690.

Synthesis of Ketone 56. Oxime 54a (11.14 g, 14.9 mmol) was added to a mixture of levulinic acid and 1 M HCl (334 g, 9:1 v/v) and the mixture was stirred at 30 oC for 4.5 h. The reaction mixture was diluted with water (1 L), and was extracted with dichloromethane. The organic layer was washed with sat. NaHCO3, water and brine, dried over Na2SO4, and concentrated. The residue was purified by flash column chromatography on silica gel (gradient

20-30% EtOAc/hexanes) to afford ketone 56 as a slightly pink foam (10.04 g, 92%). 1H NMR

(400 MHz, CDCl3) δ 8.06 (d, J = 7.6 Hz, 1H), 7.62 (d, J = 8.1 Hz, 2H), 7.55-7.26 (m, 10H), 5.54

(d, J = 11.8 Hz, 1H), 5.20 (d, J = 11.6 Hz, 1H), 4.91 (d, J = 6.6 Hz, 1H), 4.18-4.07 (m, 2H), 4.21

(d, J = 13.6 Hz, 1H), 3.62 (s, 2H), 3.48 (s, 3H), 3.32 (q, J = 9.3 Hz, 1H), 3.20 (d, J = 13.6 Hz,,

1H), 2.48 (s, 3H), 2.46 (m, 1H), 1.46-1.42 (m, 2H), 0.95 (t, J = 8.6 Hz, 2H), 0.02 (s, 9H); 13C

NMR (75 MHz, CDCl3) δ 201.6, 171.2, 170.2, 151.5, 144.3, 135.4, 134.7, 132.2, 131.7, 130.0,

129.3, 129.1, 128.9, 128.0, 125.4, 123.7, 119.3, 117.4, 115.9, 68.9, 63.8, 56.1, 52.4, 48.8, 45.3,

+ 40.0, 30.8, 27.5, 21.7, 17.4, -1.5; LRMS-ES+ m/z (relative intensity) 750 (M+NH4 , 75);

+ HRMS-ES+ (C38H48N3O9SSi) calcd 750.2881 (M+NH4 ), found 750.2878.

166

Synthesis of Keto Acid 57. To a stirred solution of TMSE-ester 56 (3.95 g, 5.38 mmol) in CH2Cl2 (138 mL) was added TFA (34 mL). The resulting solution was stirred at rt for 9 h and the solvent was removed in vacuo to give a residue which was purified by flash column chromatography on silica gel (40% EtOAc/hexanes + 1% AcOH) to give keto acid 57 (3.39 g,

1 100%). H NMR (300 MHz, CDCl3) δ 8.07 (d, J = 6.9 Hz, 1H), 7.62 (d, J = 8.0 Hz, 2H), 7.50-

7.16 (m, 10H), 5.54 (d, J = 11.8 Hz, 1H), 5.21 (d, J = 11.1 Hz, 1H), 4.87 (d, J = 6.2 Hz, 1H),

4.00 (13.4 Hz, 1H), 3.68 (s, 2H), 3.46 (s, 3H), 3.36-3.29 (m, 1H), 3.18 (d, J = 13.9 Hz, 1H), 2.48

13 (s, 3H), 1.41 (m, 2H); C NMR (75 MHz, CDCl3) δ 201.6, 175.3, 171.1, 151.4, 144.3, 137.9,

135.4, 134.5, 132.2, 131.9, 129.9, 129.3, 129.1, 128.8, 128.2, 127.9, 125.5, 125.3, 123.7, 119.1,

116.6, 115.9, 68.9, 56.0, 52.4, 48.7, 45.2, 40.0, 30.0, 27.5, 21.6; LRMS-ES+ m/z (relative

+ + intensity) 633 (MH , 100); HRMS-ES+ (C33H36N3O9S) calcd 650.2172 (M+NH4 ), found

650.2142.

Synthesis of β-Lactones 72 and 73. To a stirred suspension of 4-PPY (1.09 g, 7.35 mmol), 2-bromo-N-propylpyridinium triflate (2.57 g, 7.35 mmol) and DIPEA (1.7 mL, 9.8 mmol) in CH2Cl2 (79 mL) at rt was added a solution of keto acid 57 (3.00 g, 4.90 mmol) and glacial acetic acid (0.35 mL, 6.1 mmol) in CH2Cl2 (18 mL) over 1 h via a syringe pump. The resultant

167 orange solution was stirred for another 3 h at rt. Solvent was then removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 30-40%

EtOAc/hexanes) to give an inseparable mixture of β-lactones 72 and 73 as an off white solid

-1 1 (2.71 g, 94%). FTIR (film) 1835 cm ; H NMR (400 MHz, CDCl3, ~97:3 mixture of diastereomers determined by 1H NMR; only the major diastereomer peaks are reported) δ 8.02 (d,

J = 7.2 Hz, 1H), 7.66 (d, J = 7.8 Hz, 2H), 7.60 (d, J = 7.4 Hz, 1H), 7.45 (m, 2H), 7.42-7.31 (m,

7H), 5.46 (d, J = 10.8 Hz, 1H), 5.32 (d, J = 12.0 Hz, 1H), 4.76 (s, 1H), 4.41 (d, J = 5.7 Hz, 1H),

3.88 (d, J = 12.0 Hz, 1H), 3.50 (m, 1H), 3.50 (s, 3H), 2.95 (d, J = 11.8 Hz, 1H), 2.55 (dd, J = 7.3,

13 14.5 Hz, 1H), 1.60 (m, 2H); C NMR (75 MHz, CDCl3) δ 170.7, 165.4, 151.8, 144.5, 136.6,

134.5, 132.3, 130.1, 129.9, 129.0, 128.9, 128.8, 127.8, 126.9, 125.8, 123.8, 118.8, 115.6, 110.6,

75.6, 69.4, 53.0, 52.2, 50.9, 44.8, 43.3, 38.5, 25.1, 21.7; LRMS-ES+ m/z (relative intensity) 615

+ (MH+, 100); HRMS-ES+ (C33H34N3O8S) calcd 632.2067 (M+NH4 ), found 632.2053.

Synthesis of NH-Indole Esters 89a and 88b. 10% Pd/C (0.54 g) was suspended in a solution of the mixture of N-Cbz β-lactones 72 and 73 (2.71 g, 4.41 mmol) in EtOAc (280 mL).

One drop of glacial acetic acid was added to the mixture, followed by three evacuation-backfill cycles with hydrogen gas from a balloon. The reaction mixture was stirred under a balloon of H2 for 1.5 h at rt. The mixture was filtered through a pad of Celite, concentrated and the residue was purified by flash chromatography on silica gel (gradient 2-5% Et2O/ CH2Cl2) to afford indole

168

1 89a (1.99 g, 94%) as a white solid. H NMR (400 MHz, CDCl3) δ 9.50 (br s, 1H), 7.66 (d, J =

8.0 Hz, 2H), 7.62 (d, J = 7.6 Hz, 1H), 7.42 (d, J = 7.8 Hz, 1H), 7.36 (d, J = 8.0 Hz, 2H), 7.25-

7.18 (m, 2H), 4.86 (s, 1H), 4.21 (d, J = 4.5 Hz, 1H), 4.09 (d, J = 11.6 Hz, 1H), 3.86 (s,3H), 3.72

(dd, J = 2.0, 9.6 Hz, 1H), 2.78 (d, J = 11.6 Hz, 1H), 2.71 (dt, J = 4.5, 12.6 Hz, 1H), 2.46 (s, 3H),

2.30 (td, J = 2.5, 12.0 Hz, 1H), 1.58 (m, 1H), 1.40 (qd, J = 4.5, 13.0 Hz, 1H); 13C NMR (75 MHz,

CDCl3) δ 171.5, 166.1, 144.5, 136.1, 132.6, 130.1, 129.1, 127.9, 125.7, 122.8, 120.7, 118.3,

+ 111.7, 101.4, 53.6, 53.0, 51.7, 45.4, 39.5, 39.1, 24.5, 21.7; HRMS (m/z): [M + NH4] calcd for

C25H28N3O6S, 498.1693; found 498.1663.

Indole 88b (65 mg, 3 %) was obtained as a white solid. 1H NMR (400 MHz, DMSO-d6) δ

7.69 (d, J = 8.1 Hz, 2H), 7.61 (d, J = 7.8 Hz, 1H), 7.45 (d, J = 8.1 Hz, 2H), 7.40 (d, J = 8.0 Hz,

1H), 7.15 (dd, J = 7.1, 7.4 Hz, 1H), 7.08 (dd, J = 7.4, 7.5 Hz, 1H), 4.84 (s, 1H), 4.03-4.00 (m,

2H), 3.51 (br s, 4H), 2.94 (dd, J = 3.6, 12.6 Hz, 1H), 2.88 (d, J = 11.6 Hz, 1H), 2.46-2.43 (m,

1H), 2.41 (s, 3H), 1.95 (br d, J = 10.1 Hz, 1H), 1.08 (qd, J = 4.3, 12.9 Hz, 1H); 13C NMR (75

MHz, DMSO-d6) δ 170.9, 166.9, 143.9, 136.2, 132.6, 130.0, 127.5, 125.4, 121.6, 119.3, 117.9,

111.5, 100.2, 76.8, 52.4, 52.1, 50.7, 44.8, 42.8, 38.8, 30.7, 28.7, 21.0; HRMS (m/z): [M + H]+ calcd for C25H25N2O6S, 481.1428; found 481.1395.

169

Synthesis of N-Hydroxymethyl Indole β-Lactone 98. To a solution of NH indole lactone 89a (25.0 mg, 0.052 mmol) in THF (5 mL) at -78 oC was added dropwise a solution of

KHMDS in toluene (0.5 M, 0.26 mL, 0.13 mmol). The resulting red-orange solution was stirred at -78 oC for 15 min, and then treated with a suspension of paraformaldehyde (5.0 mg) in THF

(0.5 mL). The reaction mixture was stirred for another 12 h and allowed to warm to rt. NH4Cl(aq.) was then added, and the mixture was extracted with EtOAc. Organic layers were combined, washed with brine and dried over Na2SO4. The solvent was removed in vacuo to provide a residue which was purified by flash chromatography on silica gel (30% EtOAc/hexanes) to

1 afford the title compound 98 as a colorless gum (7.0 mg, 26%). H NMR (400 MHz, CDCl3) δ

7.67 (d, J = 7.0 Hz, 2H), 7.61 (d, J = 7.7 Hz, 1H), 7.54 (d, J = 8.0 Hz, 1H), 7.36 (d, J = 7.4 Hz,

2H), 7.31 (d, J = 7.6 Hz, 1H), 7.23 (m, 1H), 5.49 (t, J = 12.1 Hz, 1H), 5.14 (d, J = 12.0 Hz, 1H),

4.84 (s, 1H), 4.26 (d, J = 4.6 Hz, 1H), 4.06 (d, J = 11.7 Hz, 1H), 3.89 (m, 1H), 3.85 (s, 3H), 3.74

(m, 1H), 2.81 (d, J = 11.7 Hz, 1H), 2.56 (m, 1H), 2.46 (s, 3H), 2.31 (app t, J = 11.2 Hz, 1H), 1.55

13 (m, 2H); C NMR (100 MHz, CDCl3) δ 173.4, 166.1, 144.8, 138.4, 132.8, 130.4, 128.1, 127.7,

125.7, 124.0, 121.7, 118.9, 112.2, 110.5, 109.6, 103.4, 75.9, 68.2, 53.7, 51.6, 45.4, 40.8, 39.9,

25.0, 21.9; LRMS-ES+ m/z (relative intensity) 511 (MH+, 100).

170

Synthesis of α-Hydroxymethyl Ester 99. To a solution of indole β-lactone 89a (255 mg,

0.531 mmol) in THF (25 mL) cooled to -78 oC was added a solution of LiHMDS (1.0 M in THF,

1.60 mL, 1.60 mmol) dropwise with stirring. The resulting orange red solution was stirred at -78 oC for another 30 min. A solution of freshly distilled monomeric formaldehyde30 in THF (~0.5 M,

10.6 mL, 5.3 mmol) was added dropwise. The resulting brownish red solution was stirred at -78 oC for 5 min, then warmed to -40 oC and stirred for another 15 min. The reaction mixture was then quenched at -40 oC by addition of glacial acetic acid (0.10 mL, 1.6 mmol). The bright yellow solution was diluted with dichloromethane, quickly washed with ice cold water and dried over Na2SO4. After concentration of the solution, the residue was purified by column chromatography on silica gel (5% Et2O/ CH2Cl2, then 50% EtOAc/hexanes) to afford α- hydroxymethyl ester 99 as an off-white solid (160 mg, 59%). FT-IR (ATR) 3403, 1824 cm-1; 1H

NMR (400 MHz, CDCl3) δ 9.43 (s, 1H, NH), 7.62-7.66 (m, 4H), 7.34-7.42 (m, 4H), 7.16-7.25

(m, 2H), 4.86 (s, 1H), 3.97-4.02 (m, 3H), 3.83 (s, 3H), 3.68 (app. d, J = 15.6Hz , 1H), 2.74 (d, J

= 15.7 Hz, 1H), 2.57 (dd, J = 6.4, 15.9 Hz, 1H), 2.46 (s, 3H), 2.26 (td, J = 4.2, 15.7 Hz, 1H), 2.13

13 (t, J = 8.2 Hz, 1H), 1.43-1.58 (m, 2H); C NMR (100 MHz, CDCl3) δ 172.4, 167.1, 144.7, 136.6,

132.9, 131.6, 130.2, 128.0, 125.7, 123.2, 120.8, 118.6, 111.8, 101.0, 69.5, 55.4, 54.1, 53.1, 51.7,

45.3, 40.7, 26.8, 21.8. ESI MS (m/z): [M + H]+ 511.3; HRMS-ES (m/z): [M + H]+ calcd for

C26H27N2O7S, 511.1539; found, 511.1538.

171

Synthesis of ε-Lactone 102. To a solution of β-lactone 99 (200 mg, 0.39 mmol) in dichloromethane (2.6 mL) at rt was added triethylamine (10.6 mL, excess). The resulting slightly yellow solution was stirred at rt for 2 h, evaporated to dryness and the residue was purified by flash column chromatography on silica gel (5% Et2O/ CH2Cl2, then 50% EtOAc/hexanes) to afford ε-lactone 102 as an off white solid (65 mg, 33%). The recovered β-lactone 99 was subjected to another translactonization under the same conditions. After 2 runs, 40 mg of β- lactone 99 was recovered and ε-lactone 102 was isolated as an off white foam (92 mg, BRSM =

-1 1 58%). FT-IR (ATR) 3407, 1725 cm ; H NMR (400 MHz, CDCl3) δ 9.61 (s, 1H, NH), 7.61 (d, J

= 8.2 Hz, 2H), 7.53 (d, J = 7.5 Hz, 1H), 7.41 (d, J = 8.0 Hz, 1H), 7.25 – 7.27 (m, 4H), 7.21 (t, J =

7.4 Hz, 1H), 7.15 (t, J = 7.5 Hz), 4.90 (d, J = 11.7 Hz, 1H), 4.37 (d, J = 11.7 Hz, 1H), 4.16 (s,

1H), 3.93 (s, 3H), 3.65 (br s, 1H), 3.58 (d, J = 13.5 Hz, 1H), 3.52-3.58 (m, 1H), 3.01 (m, 1H),

2.81 (dd, J = 3.8, 13.6 Hz, 1H), 2.59 (d, J = 13.6 Hz, 1H), 2.37 (s, 3H),1.40 (m, 1H), 0.76 (m,

13 1H); C NMR (100 MHz, CDCl3) δ 171.0, 167.8, 144.2, 135.9, 134.4, 132.0, 130.2, 127.4,

124.9, 122.8, 120.9, 118.0, 113.7, 112.1, 107.0, 76.0, 73.6, 53.7, 53.0, 52.8, 52.2, 50.6, 42.4, 23.8,

+ + + 21.7; ESI MS (m/z): [M + H] 511.3; HRMS (m/z): [M + NH4] calcd for C26H30N3O7S ,

528.1799; found, 528.1797.

172

Synthesis of Hydroxy Aldehyde 104. Method A. To a solution of ɛ-lactone 102 (5.5 mg,

o 0.011 mmol) in CH2Cl2 (0.2 mL) at -78 C was added BF3·OEt2 (2.7 µL, 0.022 mmol), followed by dropwise addition of DIBAL-H solution in toluene (1.5 M, 14 µL, 0.022 mmol). The reaction mixture was stirred at -78 oC for 1 h, and then quenched by addition of 10% Rochelle’s salt and

Na2HPO4/KH2PO4 buffer (pH 7). The mixture was extracted with CH2Cl2. The organic layer was washed with brine and dried over Na2SO4. The solvent was removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 30-50% EtOAc/hexanes) to

1 afford aldehyde 104 as an off-white foam (1.2 mg, 22%). H NMR (400 MHz, CDCl3) δ 10.3 (s,

1H), 8.94 (s, 1H), 7.67 (m, 3H), 7.41 (d, J = 8.2 Hz, 1H), 7.35 (d, J = 8.0 Hz, 2H), 7.24 (m, 1H),

7.14 (dd, J = 7.4, 7.5 Hz, 1H), 4.51 (s, 1H), 4.19 (d, J = 11.1 Hz, 1H), 4.15-4.00 (m, 3H), 3.78 (s,

3H), 3.70 (m, 2H), 2.46 (s, 3H), 2.26 (m, 2H), 2.11 (td, J = 4.2, 11.4 Hz, 1H), 1.00-0.86 (m, 2H);

13 C NMR δ (75 MHz, CDCl3) δ 203.1, 173.4, 144.5, 137.4, 133.2, 130.8, 130.3, 128.2, 126.4,

123.1, 120.5, 120.0, 112.1, 105.2, 72.5, 69.4, 55.5, 55.2, 53.2, 49.6, 48.7, 45.9, 26.8, 22.0;

+ HRMS (m/z): [M + H] calcd for C26H29N2O7S, 513.1695; found 513.1696.

Method B. OTBS aldehyde 164 (233 mg, 0.36 mmol) was dissolved in a solution of HCl in EtOAc (1.0 M, 5 mL) and the mixture was stirred at rt for 1 h. The volatiles were removed in vacuo to afford hydroxy aldehyde 104 as a red solid (183 mg, 100%), which was used in the next step without further purification.

173

Synthesis of Dithiane 105. To a stirred solution of aldehyde 104 (10 mg, 0.020 mmol) in CH2Cl2 (1.5 mL) at rt was added BF3·OEt2 (2.5 µL, 0.020 mmol) and 1,2-ethanedithiol (3.5

µL, 0.042 mmol). The reaction mixture was stirred at rt for 12 h, then diluted with CH2Cl2 and phosphate buffer (pH 7). The organic layer was separated, and aqueous layer was extracted with

CH2Cl2. The combined organic layers were washed with brine, dried over Na2SO4 and concentrated in vacuo to give a residue, which was purified by column chromatography on silica gel (gradient 30-50% EtOAc/hexanes) to afford dithiane 105 as a white foam (10 mg, 87%). 1H

NMR (400 MHz, CDCl3) δ 9.05 (br s, 1H), 7.68-7.60 (m, 3H), 7.44 (d, J = 7.8 Hz, 1H), 7.38 (d,

J = 8.9 Hz, 2H), 7.23-7.12 (m, 2H), 5.87 (br s, 1H), 4.88 (s, 1H), 4.28 (d, J = 11.2 Hz, 1H), 4.17-

4.02 (m, 3H), 3.87 (s, 3H), 3.70 (m, 1H), 3.05 (m, 1H), 2.78 (d, J = 11.7 Hz, 1H), 2.45 (s, 3H),

2.30 (m, 3H); LRMS-ES+ m/z (relative intensity) 589 (MH+).

Synthesis of Methyl Diol 106. To a suspension of freshly prepared W-2 Raney nickel in ethanol (ca 2 mL, excess) was added a solution of dithiane 105 (12.0 mg, 0.020 mmol) in ethanol (5 mL). The reaction mixture was heated at 60 oC untill TLC indicated consumption of

174 starting dithiane (ca 45 min). Raney nickel was filtered off through a pad of Celite, and the filtrate was evaporated to dryness in vacuo to give the methyl compound 106 which is pure based

1 1 on H NMR, as a white powder (7.5 mg, 74%). H NMR (400 MHz, CDCl3) δ 8.64 (br s, 1H),

7.72 (d, J = 7.8 Hz, 1H), 7.67 (d, J = 8.1 Hz, 2H), 7.35 (m, 3H), 7.20 (dd, J = 7.2, 7.2 Hz, 1H),

7.12 (dd, J = 7.2, 7.2 Hz, 1H), 4.11 (d, J = 11.2 Hz, 1H), 4.01 (d, J = 11.2 Hz, 1H), 3.93 (d, J =

10.6 Hz, 1H), 3.75 (s, 3H), 3.67 (q, J = 7.1 Hz, 1H), 2.45 (s, 3H), 2.27 (m, 1H), 2.13 (dd, J = 3.7,

12.5 Hz, 1H), 2.08 (d, J = 11.6 Hz, 1H), 1.56 (d, J = 7.0 Hz, 3H), 0.83 (m, 2H); 13C NMR (75

MHz, CDCl3) δ 174.1, 137.9, 133.3, 130.2, 128.1, 122.5, 120.9, 120.0, 111.9, 78.8, 72.7, 53.0,

47.8, 46.1, 31.0, 30.1, 26.8, 22.0, 13.3; LRMS-ES+ m/z (relative intensity) 499 (MH+).

Synthesis of Iodide 112. To a solution of PPh3 (59 mg, 0.225 mmol) and I2 (57 mg,

0.225 mmol) in CH2Cl2 (2 mL) was added solid imidazole (26 mg, 0.38 mmol). The resulting dark brown suspension was stirred at rt for 15 min, and then a solution of indole diol 74 (31 mg,

0.050 mmol) in CH2Cl2 (1.5 mL) was added dropwise. The resulting bright yellow suspension was stirred at rt for 16 h, then diluted with CH2Cl2. The solution was washed with 5% Na2S2O3, brine, and dried over Na2SO4. After concentration of the solution, the residue was purified by flash chromatography on silica gel (gradient 10-30% EtOAc/hexanes) to afford iodide 112 as a

1 slightly yellow foam (26 mg, 72%). H NMR (300 MHz, CDCl3) δ 8.10 (m, 1H), 7.66 (d, J = 8.0

Hz, 2H), 7.61 (m, 1H), 7.48-7.29 (m, 10H), 5.47 (d, J = 12.0 Hz, 1H), 5.30 (d, J = 12.0 Hz, 1H),

4.55 (m, 1H), 4.17-3.86 (m, 4H), 3.68 (m, 1H), 3.36 (s, 3H), 2.65 (m, 1H), 2.45 (s, 3H), 2.26 (m,

3H), 0.99-0.85 (m, 2H); LRMS-ES+ m/z (relative intensity) 728 (MH+, 100).

175

Synthesis of Methyl Compound 110. 10% Pd/C (100 mg) was suspended in a solution of N-Cbz iodide 112 (26 mg, 0.036 mmol) in a mixture of EtOAc/t-BuOH (1:1 v/v, 25 mL), followed by addition of glacial acetic acid (20 µL, 0.33 mmol). The reaction mixture was evacuated and backfilled with hydrogen gas from a balloon (3 times), and was stirred under hydrogen for 24 h. The reaction mixture was then filtered through a pad of Celite. The filtrate was concentrated, and the residue was purified by flash chromatography on silica gel (30%

EtOAc/hexanes) to afford the methyl compound 110 as a white solid (16 mg, 95%). 1H NMR

(400 MHz, CDCl3) δ 8.97 (br s, 1H), 8.10 (m, 1H), 7.66 (d, J = 8.0 Hz, 2H), 7.61 (m, 1H), 7.48-

7.29 (m, 10H), 5.47 (d, J = 12.0 Hz, 1H), 5.30 (d, J = 12.0 Hz, 1H), 4.55 (m, 1H), 4.17-3.86 (m,

4H), 3.68 (m, 1H), 3.36 (s, 3H), 2.65 (m, 1H), 2.45 (s, 3H), 2.26 (m, 3H), 1.48 (d, J = 7.0 Hz,

3H), 0.99-0.85 (m, 2H).

Synthesis of TMS Ether 115. A sealed tube was charged with tertiary alcohol 110 (10 mg, 0.021 mmol), TMSCN (0.45 mL, excessive) and CH2Cl2 (1 mL). The tube was then sealed and heated at 50 oC for 48 h. The volatiles were removed in vacuo to give a dark red residue which was purified by column chromatography on silica gel (20% EtOAc/hexanes) to afford the

176

1 O-TMS compound 115 as a brown gum (10 mg, 87%). H NMR (400 MHz, CDCl3) δ 8.90 (s,

1H), 7.68 (d, J = 8.0 Hz, 2H), 7.64 (d, J = 7.8 Hz, 1H), 7.38-7.34 (m, 3H), 7.15 (dd, J = 7.3, 7.5

Hz, 1H), 7.08 (dd, J = 7.6, 7.4 Hz, 1H), 4.35 (d, J = 2.0 Hz, 1H), 4.08 (d, J = 10.7 Hz, 1H), 3.80

(s, 3H), 3.75 (m, 1H), 3.54 (q, J = 6.7 Hz, 1H), 2.47 (s, 3H), 2.29 (t, J = 11.7 Hz, 1H), 2.21 (d, J

= 10.8 Hz, 1H), 1.49-1.65 (m, 3H), 1.42 (d, J = 6.8 Hz, 3H), 0.98-0.84 (m, 2H), 0.11 (s, 9H);

LRMS-ES+ m/z (relative intensity) 541 (MH+, 100).

Synthesis of N-Hydroxymethyl Compound 116. To a solution of indole O-TMS ether

115 (3.5 mg, 0.0065 mmol) in THF (0.5 mL) at -78 oC was added dropwise a solution of

LiHMDS (1.0 M, 20 μL, 0.020 mmol). The resulting orange red solution was stirred at -78 oC for

30 min, then treated with a solution of freshly distilled monomeric formaldehyde in THF (~0.5

M, 0.13 mL, 0.065 mmol). The reaction mixture was stirred at -78 oC for 5 min, then warmed to

-40 oC and stirred for another 10 min. The reaction was quenched by addition of glacial HOAc (2

μL) at -40 oC. The volatiles were removed in vacuo to afford a residue which was purified by flash chromatography on silica gel (gradient 20-30% EtOAc/hexanes) to afford compound 116 as

1 a slightly brown gum (1.5 mg, 41%). H NMR (400 MHz, CDCl3) δ 7.67 (m, 3H), 7.53 (d, J =

10.9 Hz, 1H), 7.37 (d, J = 10.9 Hz, 2H), 7.24 (m, 1H), 7.13 (dd, J = 10.6, 10.1 Hz, 1H), 5.45 (t, J

= 16.4 Hz, 1H), 5.10 (dd, J = 3.1, 16.1 Hz, 1H), 4.34 (dd, J = 3.0, 7.3 Hz, 1H), 4.15-4.01 (m, 1H),

3.83 (m, 1H), 3.80 (s, 3H), 3.70 (dd, J = 3.6, 16.6 Hz, 1H), 3.57 (dd, J = 3.0, 9.1 Hz, 1H), 2.47 (s,

3H), 2.25 (m, 1H), 2.20 (d, J = 14.7 Hz, 1H), 2.15 (m, 1H), 1.41 (d, J = 11.4 Hz, 3H), 0.09 (s,

9H); LRMS-ES+ m/z (relative intensity) 571 (MH+, 100).

177

Synthesis of O-TBS β-Lactone 101. To a solution of hydroxymethyl compound 99 (100

o mg, 0.196 mmol) in CH2Cl2 (15 mL) at 0 C was added 2,6-lutidine (226 μL, 1.95 mmol) and freshly distilled TBSOTf (225 μL, 0.98 mmol) with stirring. The colorless reaction mixture was stirred at 0 oC and monitored by TLC. Once the reaction was complete (~40 min), the solvent was removed in vacuo. The residue was purified by flash chromatography on silica gel (gradient

5-30% EtOAc/hexanes) to afford O-TBS ether 101 as a white foam (105 mg, 86%). 1H NMR

(400 MHz, CDCl3) δ 9.32 (br s, 1H, NH), 7.66-7.63 (m, 3H), 7.40 (d, J = 8.0 Hz, 1H), 7.36 (d, J

= 7.7 Hz, 2H), 7.23 (m, 1H), 7.19 (dd, J = 7.5, 7.6 Hz, 1H), 4.84 (s, 1H), 4.03 (d, J = 10.1 Hz,

1H), 4.00 (d, J = 12.4 Hz, 1H), 3.88 (d, J = 9.2 Hz, 1H), 3.81 (s, 3H), 3.70 (m, 1H), 2.74 (d, J =

11.6 Hz, 1H), 2.55 (dd, J = 4.7, 12.1 Hz, 1H), 2.46 (s, 3H), 2.27 (td, J = 3.0, 11.8 Hz,, 1H), 1.55-

13 1.47 (m, 2H), 0.80 (s, 9H), -0.17 (s, 3H),, -0.33 (s, 3H); C NMR (75 MHz, CDCl3) δ 172.5,

167.4, 144.8, 136.4, 132.7, 132.4, 130.4, 128.2, 125.7, 123.1, 120.7, 118.7, 111.7, 100.4, 70.3,

55.8, 54.4, 53.0, 51.9, 45.5, 41.3, 27.1, 26.1, 25.9, 22.0, 18.4, -5.6; HRMS (m/z): [M + H]+ calcd for C32H41N2O7SSi, 625.2404; found 625.2390.

178

Synthesis of Indole Diol Ester 117. To a solution of O-TBS β-lactone 101 (120 mg,

0.193 mmol) in THF at rt was added LiBH4 (22 mg, 1.0 mmol). The reaction mixture was stirred at rt for 12 h, diluted with aqueous NH4Cl and extracted with CH2Cl2. The extract was dried over

Na2SO4 and the solvent was removed in vacuo to give a residue which was purified by flash chromatography on silica gel (2% to 5% Et2O/ CH2Cl2) affording diol 117 as a white solid (90

1 mg, 75%). H NMR (400 MHz, CDCl3) δ 8.76 (br s, 1H), 7.69-7.66 (m, 3H), 7.39 (d, J = 7.9 Hz,

1H), 7.34 (d, J = 7.7 Hz, 2H), 7.22 (dd, J = 7.6, 7.8 Hz, 1H), 7.14 (dd, J = 7.1, 7.4 Hz, 1H), 4.87

(d, J = 12.2 Hz, 1H), 4.44 (d, J = 12.9 Hz, 1H), 4.23 (d, J = 8.7 Hz, 1H), 4.08 (d, J = 9.5 Hz, 2H),

3.71 (s, 3H), 3.67 (m, 1H), 2.45 (s, 3H), 2.25 (dd, J = 10.5, 10.8 Hz, 1H), 2.17 (d, J = 11.8 Hz,

1H), 1.92 (d, J = 12.2 Hz, 1H), 1.59 (m, 1H), 1.40 (m, 1H), 0.87 (s, 9H), -0.05 (s, 3H), -0.11 (s,

13 3H); C NMR (75 MHz, CDCl3) δ 173.4, 144.3, 137.0, 135.0, 133.2, 130.2, 128.1, 125.8, 122.3,

120.2, 119.2, 111.9, 103.8, 74.5, 71.8, 61.0, 55.8, 54.9, 52.5, 47.6, 46.1, 37.4, 26.4, 26.2, 22.0,

+ 18.5, -5.2, -5.5; HRMS (m/z): [M + H] calcd for C32H45N2O7SSi, 629.2717; found 629.2717.

179

Synthesis of Iodo Alcohol 118. To a solution of PPh3 (229 mg, 0.87 mmol) and I2 (222 mg, 0.87 mmol) in CH2Cl2 (9 mL) at rt was added imidazole (99 mg, 1.48 mmol). The resulting yellow suspension was stirred at rt for 10 min. A solution of diol 117 (122 mg, 0.195 mmol) in

CH2Cl2 (15 mL) was added dropwise, and the bright yellow suspension was stirred at rt for 5 h.

The mixture was diluted with CH2Cl2 and washed with 5% Na2S2O3. The aqueous layer was extracted with CH2Cl2. The combined organic layers were washed with brine, dried over Na2SO4, and the solvent was removed in vacuo. The residue was purified by flash column chromatography on silica gel (2% Et2O/ CH2Cl2) to afford iodo alcohol 118 as a white foamy

1 solid (133 mg, 93%). H NMR (400 MHz, CDCl3) δ 8.80 (br s, 1H), 7.67 (d, J = 8.1 Hz, 2H),

7.63 (d, J = 7.7 Hz, 1H), 7.37-7.33 (m, 3H), 7.20 (dd, J = 7.1, 7.5 Hz, 1H), 7.14 (dd, J = 7.4, 7.4

Hz, 1H), 4.17-4.09 (m, 2H), 4.00-3.96 (m, 1H), 3.90-3.88 (m, 2H), 3.68 (s, 3H), 3.30 (s, 1H),

2.44 (s, 3H), 2.30 (m, 1H), 2.18 (m, 1H), 2.08 (dd, J = 3.1, 11.5 Hz, 1H), 1.63 (m, 1H), 1.43 (m,

13 1H), 0.84 (s, 9H), -0.08 (s, 3H), -0.10 (s, 3H); C NMR (75 MHz, CDCl3) δ 172.9, 144.0, 136.5,

132.7, 132.3, 132.1, 129.9, 128.7, 127.9, 126.0, 122.0, 119.6, 111.4, 108.9, 72.9, 71.0, 54.8, 54.3,

52.3, 45.6, 36.4, 25.8, 21.6, 18.2, 4.8, 1.1, -5.6, -5.7; HRMS (m/z): [M + H]+ calcd for

C32H44IN2O6SSi, 739.1734; found, 739.1764.

180

Synthesis of Methyl Compound 108. 10% Pd/C (158 mg) was suspended in a stirred solution of iodide 118 (29 mg, 0.039 mmol) in t-BuOH/EtOAc (1:1 v/v, 30 mL). The reaction mixture was evacuated and backfilled with H2 three times from a balloon and stirred under a H2 atmosphere at rt. After 3.5 h, another portion of 10% Pd/C (32 mg) was added and the reaction mixture was stirred under H2 at rt for another 12 h. The Pd/C was filtered off through a pad of

Celite, and the solvent was removed in vacuo to afford a yellow foamy solid. This material was purified by flash column chromatography on silica gel (CH2Cl2, then 2% Et2O/ CH2Cl2) to afford

1 methyl compound 108 as a white solid (23 mg, 96%). H NMR (400 MHz, CDCl3) δ 8.51 (br s,

1H), 7.72 (d, J = 7.9 Hz, 1H), 7.67 (d, J= 8.1 Hz, 2H), 7.34-7.32 (m, 3H), 7.17 (dd, J = 7.1, 7.6

Hz, 1H), 7.08 (dd, J= 7.2, 7.4 Hz, 1H), 4.13 (d, J = 9.4 Hz, 1H), 4.00 (d, J = 9.4 Hz, 1H), 3.96 (d,

J = 11.8 Hz, 1H), 3.72 (s, 3H), 3.66 (q, J = 7.0 Hz, 1H), 3.09 (s, 1H), 2.45 (s, 3H), 2.26 (t, J =

10.4 Hz, 1H), 2.04 (d, J = 13.0 Hz, 1H), 2.00 (dd, J= 3.7, 12.6 Hz, ,1H), 1.63 (qd, J = 4.4, 12.9

Hz, 1H), 1.57 (d, J = 6.9 Hz, 3H), 0.80 (s, 3H), -0.07 (s, 3H), -0.12 (s, 3H); 13C NMR (75 MHz,

CDCl3) δ 173.3, 143.9, 137.1, 133.0, 130.0, 129.9, 129.8, 127.9, 126.7, 121.7, 120.5, 119.1,

111.4, 111.2, 72.5, 70.7, 55.1, 53.4, 52.3, 48.2, 46.0, 30.5, 26.5, 25.8, 25.7, 21.7, 18.1, 12.8, -5.7,

+ -5.8; HRMS (m/z): [M + H] calcd for C32H45N2O6SSi, 613.2768; found, 613.2772.

181

Synthesis of Piperidine 109. Magnesium turnings (399 mg, excess) were added to a solution of compound 108 (21.6 mg, 0.035 mmol) in methanol, and the mixture was sonicated at rt until all the magnesium turnings dissolved (~45 min). The reaction mixture was then poured

o into aqueous NH4Cl at 0 C, and extracted with CHCl3. The organic extract was dried over

Na2SO4, and evaporated to dryness in vacuo to give a white foamy solid. This material was purified by flash chromatography on silica gel (2% to 5% MeOH/CHCl3, then 1% NEt3 in 5%

1 MeOH/CHCl3) to afford piperidine 109 as a white solid (15.6 mg, 96%). H NMR (400 MHz,

CD3OD) δ 7.90 (br s, 1H), 7.66 (d, J = 7.8 Hz, 1H), 7.33 (d, J = 8.0 Hz, 1H), 7.08 (dd, J = 7.2,

7.4 Hz, 1H), 6.98 (dd, J = 7.2, 7.3 Hz, 1H), 4.34 (d, J = 9.3 Hz, 1H), 4.20 (d, J = 9.4 Hz, 1H),

3.77 (s, 3H), 3.42 (q, J = 6.7 Hz, 1H), 3.28 (m, 1H), 3.03 (q, J = 7.2 Hz, 1H), 2.94 (br d, J = 12.6

Hz, 1H), 2.69 (m, 1H), 2.50 (d, J = 12.7 Hz, 1H), 2.29 (m, 1H), 1.49 (d, J = 6.7 Hz, 3H), 0.83 (s,

13 9H), -0.02 (s, 3H), -0.10 (s, 3H); C NMR (75 MHz, CD3OD) δ 175.1, 138.6, 131.7, 127.9,

122.2, 120.8, 119.4, 112.7, 112.0, 79.6, 72.4, 71.6, 56.9, 54.0, 52.7, 47.8, 45.4, 32.1, 26.4, 19.2,

+ 14.5, -5.4, -5.5; HRMS (m/z): [M + H] calcd for C25H39N2O4Si, 459.2674; found, .

182

Synthesis of (±)-Alstilobanine A (15). A solution of O-TBS indole 109 (16.5 mg, 0.036

o mmol) in CHCl3 (3.3 mL) at 0 C was treated with a solution of hydrogen chloride in MeOH

(1.25 M, 3.3 mL). The resultant colorless solution was stirred at rt for 2 h, and the volatiles were removed under high vacuum to afford (±)-alstilobanine A (15) hydrochloride salt as a white solid

1 (12.2 mg, 100%). H NMR (400 MHz, CD3OD) δ 7.55 (d, J = 7.9 Hz, 1H), 7.34 (d, J = 8.0 Hz,

1H), 7.08 (dd, J = 7.4, 7.5 Hz, 1H), 6.98 (dd, J = 7.4, 7.5 Hz, 1H), 4.14 (d, J = 10.9 Hz, 1H), 4.04

(d, J = 10.8 H, 1H), 3.78 (s, 3H), 3.54 (d, J = 12.5 Hz, 1H), 3.26 (q, J = 7.1 Hz, 1H), 3.09 (m,

1H), 2.93 (m, 1H), 2.87 (d, J = 12.6 Hz, 1H), 2.52 (m, 1H), 2.06 (m, 1H), 1.87 (m, 1H), 1.43 (d,

13 J = 6.7 Hz, 3H); C NMR (75 MHz, CD3OD) δ 174.9, 138.9, 130.9, 127.2, 122.5, 119.6, 119.5,

113.3, 112.2, 71.0, 68.7, 55.2, 53.0, 50.5, 42.4 (2C), 34.4, 22.6, 15.7; HRMS (m/z): [M + H]+ calcd for C19H25N2O4, 345.1809; found, 345.1809.

183

Synthesis of Triol 103. To a solution of β-lactone 99 (10.0 mg, 0.020 mmol) in THF

(1.7 mL) at rt was added dropwise a solution of DIBAL-H in toluene (1.5 M, 0.16 mL, 0.24 mmol). The reaction mixture was stirred at rt for 1 h, then quenched with aqueous NH4Cl and 10%

Rochelle’s salt, and extracted with EtOAc. The organic layer was washed with brine, dried over

Na2SO4 and concentrated in vacuo to give a residue which was purified by column chromatography on silica gel (gradient 30-50% EtOAc/hexanes) to afford triol 103 as a white

1 solid (5.5 mg, 54%). H NMR (400 MHz, CDCl3) δ 8.94 (br s, 1H), 7.66 (d, J = 8.1 Hz, 3H),

7.41 (d, J = 8.1 Hz, 1H), 7.35 (d, J = 8.0 Hz, 2H), 7.22 (dd, J = 7.4, 7.4 Hz, 1H), 7.14 (dd, J =

7.4, 7.4 Hz, 1H), 5.11 (br s, 1H), 4.78 (d, J = 12.9 Hz, 1H), 4.47 (t, J = 9.7 Hz, 1H), 4.19 (dd, J =

4.2, 11.1 Hz, 1H), 4.13-4.06 (m, 2H), 3.76 (s, 3H), 3.69 (m, 2H), 2.71 (m, 1H), 2.46 (s, 3H), 2.27

(m, 1H), 2.21 (d, J = 12.0 Hz, 1H), 2.10 (dd, J = 3.7, 12.2 Hz, 1H), 1.44 (m, 1H), 0.86 (m, 1H);

13 C NMR (100 MHz, CDCl3) δ 174.1, 144.3, 137.3, 133.2, 130.0, 128.0, 122.6, 120.5, 119.3,

117.6, 112.1, 111.2, 74.4, 71.3, 61.2, 55.0, 52.9, 26.5, 22.0.

Synthesis of ɛ-Lactone TBS Ether 119. To a solution of ɛ-lactone 102 (24 mg, 0.047 mmol) in CH2Cl2 (4.6 mL) was added TBSOTf (0.22 mL, 0.96 mmol) and 2, 6-lutidine (0.22 mL,

1.9 mmol). The reaction mixture was stirred at rt for 1 h. The volatiles were removed in vacuo to

184 give a residue which was purified by column chromatography on silica gel to afford the OTBS ɛ-

-1 1 lactone 119 as a white foam (15 mg, 51%). FTIR (film) 1730 cm ; H NMR (400 MHz, CDCl3)

δ 9.48 (br s, 1H), 7.58 (d, J = 7.8 Hz, 2H), 7.43 (d, J = 7.2 Hz, 1H), 7.35 (m, 1H), 7.23 (m, 3H),

7.19 (dd, J = 6.1, 6.1 Hz, 1H), 4.42 (d, J = 10.8 Hz, 1H), 4.28 (s, 1H), 4.14 (d, J = 11.7 Hz, 1H),

3.85 (s, 3H), 3.82 (d, J = 10.9 Hz, 1H), 3.45 (d, J = 10.6 Hz, 1H), 2.64 (d, J = 11.9 Hz, 1H), 2.43

(m, 4H), 2.15 (t, J = 11.7 Hz, 1H), 1.35 (d, J = 13.1 Hz, 1H), 0.88 (s, 9H), 0.80-0.76 (m, 2H),

0.07 (s, 3H), -0.01 (s, 3H); LRMS-ES+ m/z (relative intensity) 625 (MH+, 100).

Synthesis of N-Boc O-TBS ε-Lactone 121. To a solution of O-TBS NH indole lactone

119 (9.0 mg, 0.0144 mmol) in MeCN (0.5 mL) was added Boc2O (56 mg, 3.1 mmol) and DMAP

(11 mg, 0.091 mmol). The resulting orange solution was stirred at rt for 16 h. The volatiles were then removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 10-20% EtOAc/hexanes) to furnish compound 121 as a white solid (7.0 mg, 67%).

1 H NMR (400 MHz, CDCl3) δ 7.92 (m,, 1H), 7.60 (d, J = 10.9 Hz, 2H), 7.54 (m, 1H), 7.32-7.29

(m, 4H), 4.44 (d, J = 14.9 Hz, 1H), 4.34 (d, J = 14.9 Hz, 1H), 4.23 (s, 1H), 4.10 (d, J = 17.6 Hz,

1H), 3.72 (s, 3H), 3.48 (m, 1H), 2.59 (d, J = 15.7 Hz, 1H), 2.42 (s, 3H), 2.09-2.00 (m, 2H), 1.66

(s, 9H), 0.88 (s, 9H), 0.11 (s, 3H), 0.00 (s, 3H); LRMS-ES+ m/z (relative intensity) 725 (MH+,

100).

185

Synthesis of Oxetane 127 and Cyclic Ether 128. To a solution of PPh3 (8.5 mg,

0.032 mmol) and triol 103 (12.0 mg, 0.024 mmol) in THF (5 mL) at rt was added DIAD (5 µL,

0.016 mmol) dropwise. The resulting orange solution was stirred at rt for 1 h. The volatiles were then removed in vacuo to give an orange residue which was purified by flash chromatography on silica gel (gradient 20-30% EtOAc/hexanes) to afford bridged cyclic ether 128 as the major

1 product (white foam, 6.1 mg, 51%). H NMR (400 MHz, CDCl3) δ 9.04 (br s, 1H), 7.69 (d, J =

7.7 Hz, 2H), 7.60 (d, J = 7.6 Hz, 1H), 7.36 (m, 1H), 7.18 (dd, J = 7.1, 7.8 Hz, 1H), 7.13 (dd, J =

7.1, 7.8 Hz, 1H), 4.97 (d, J = 4.0 Hz, 1H), 4.73 (d, J = 10.8 Hz, 1H), 4.38 (d, J = 11.7 Hz, 1H),

4.21-4.18 (m, 3H), 3.97 (d, J = 9.9 Hz, 1H), 3.86 (s, 3H), 3.72 (d, J = 5.1 Hz, 1H), 2.45 (s, 3H),

2.41 (m, 1H), 2.29 (d, J = 12.0 Hz, 1H), 1.69-1.54 (m, 1H), 1.49 (dd, J = 4.3, 12.8 Hz, 1H); 13C

NMR (100 MHz, CDCl3) δ 171.4, 144.0, 136.1, 134.3, 132.8, 130.0, 127.5, 126.6, 122.1, 120.2,

119.6, 113.9, 111.7, 106.9, 106.1, 83.1, 81.3, 62.1, 54.7, 54.0, 53.2, 51.8, 45.7, 39.4, 29.9, 22.8,

21.8; LRMS-ES+ m/z (relative intensity) 497 (MH+, 100); HRMS (m/z): [M + H]+ calcd for

C26H29N2O6S, 497.1746; found 497.1729.

Oxetane 127 was isolated as a side product (white solid, 2.8 mg, 23%). 1H NMR (400

MHz, CDCl3) δ 9.24 (br s, 1H), 7.69-7.23 (m, 8H), 5.13 (t, J = 6.1 Hz, 1H), 4.42 (m, 1H), 4.31

(m, 1H), 4.12 (m, 1H), 4.04 (dd, J = 3.5, 6.4 Hz, 1H), 3.96 (dd, J = 3.3, 5.5 Hz, 1H), 3.83 (s, 3H),

3.60 (m, 1H), 2.45 (s, 3H), 2.39 (d, J = 11.4 Hz, 1H), 2.23-2.00 (m, 2H), 1.39 (m, 1H); 13C NMR

(100 MHz, CDCl3) δ 173.7, 144.2, 136.8, 133.2, 132.7, 132.4, 131.7, 130.3, 128.7, 128.0, 126.1,

186

122.7, 120.0, 116.1, 110.4, 85.1, 75.3, 71.7, 55.7, 54.2, 52.9, 45.6, 43.0, 33.8, 26.0, 21.8; HRMS

+ (m/z): [M + H] calcd for C26H29N2O6S, 497.1746; found 497.1752.

Oxetane 127 could also be prepared as a major product using the following procedure: to

o a solution of PPh3 (4.3 mg, 0.016 mmol) and triol 103 (6.5 mg, 0.013 mmol) in benzene at 0 C was added dropwise a solution of DEAD in toluene (40 wt%, 6.5 µL, 0.014 mmol). The reaction mixture was stirred at 0 oC for 5 min, and then heated at reflux for 1 h. The mixture was cooled to rt and concentrated in vacuo to give a residue which was purified by preparative TLC (30%

EtOAc/hexanes, plate developed 4 times) to afford oxetane 127 as a white solid (4.4 mg, 70%), together with cyclopropane 129 (<0.5 mg, ~8%) and bridged cyclic ether 128 (<0.5 mg, ~8%).

Synthesis of Cyclopropane 129. To a solution of PPh3 (21 mg, 0.080 mmol) and triol

103 (8.4 mg, 0.016 mmol) in THF (1.5 mL) at 0 oC was added DIAD (16 µL, 0.050 mmol) dropwise. The resulting orange solution was stirred at 0 oC for 10 min, then allowed to warm to rt and stirred for 5 h. The volatiles were removed in vacuo to give an orange residue which was purified by flash chromatography on silica gel (gradient 20-30% EtOAc/hexanes) to afford

1 cyclopropane 129 as a white foam (6.6 mg, 86%). H NMR (400 MHz, CDCl3) δ 7.73 (d, J = 7.8

Hz, 1H), 7.68 (m, 2H), 7.63 (d, J = 8.4 Hz, 2H), 7.47 (m, 3H), 4.25 (d, J = 9.2 Hz, 1H), 4.21 (d, J

= 10.2 Hz, 1H), 4.08 (d, J = 9.5 Hz, 1H), 3.86 (s, 3H), 3.81 (m, 1H), 3.17 (dd, J = 5.6, 8.5 Hz,

1H), 2.64 (t, J = 5.4 Hz, 1H), 2.47 (d, J = 10.5 Hz, 1H), 2.44 (s, 3H), 2.21 (m, 1H), 2.15 (dd, J =

3.2, 12.7 Hz, 1H), 2.06 (m, 1H), 1.93 (m, 1H), 0.95 (m, 1H), 0.77 (m, 1H); 13C NMR (75 MHz,

187

CDCl3) δ 170.2, 132.4, 132.3, 132.2, 130.1, 128.8, 128.7, 127.7, 127.2, 125.2, 121.8, 119.0,

116.6, 112.8, 107.9, 106.3, 78.9, 72.8, 60.5, 58.6, 53.5, 53.0, 46.0, 42.2, 24.3, 22.0, 21.8; LRMS-

+ + ES+ m/z (relative intensity) 479 (MH , 100); HRMS (m/z): [M + H] calcd for C26H27N2O5S,

479.1641; found 479.1631.

Synthesis of Benzylidene Compound 130. A solution of indole triol 103 (14.0 mg,

0.027 mmol), TsOH (1 mg, cat.) and benzaldehyde dimethyl acetal (12 µL, 0.081 mmol) in toluene (15 mL) was heated at reflux with a Dean-Stark apparatus for 1 h. The reaction mixture was then treated with aqueous NaHCO3, extracted with EtOAc. The organic layers were dried over Na2SO4. The volatiles were removed in vacuo to give a residue which was purified by flash chromatography on silica gel (40% EtOAc/hexanes) to afford benzylidene compound 130 as a

1 yellow solid (7.5 mg, 46%). H NMR (300 MHz, CDCl3) δ 8.88 (br s, 1H), 7.70 (d, J = 8.2 Hz,

2H), 7.62 (d, J = 7.8 Hz, 1H), 7.41-7.29 (m, 8H), 7.23-7.12 (m, 2H), 5.13 (d, J = 12.2 Hz, 1H),

4.81 (d, J = 11.4 Hz, 1H), 4.59 (dd, J = 3.3, 12.0 Hz, 1H), 4.22 (m, 1H), 4.04 (m, 1H), 3.77 (s,

3H), 3.57 (d, J = 2.9 Hz, 1H), 2.47 (s, 3H), 2.37 (m, 1H), 2.28 (m, 2H) 2.09 (m, 1H), 1.41 (m,

1H), 0.93 (m, 1H).

188

Synthesis of Fully Protected Triol 133. A mixture of triol 103 (78 mg, 0.15 mmol) and tetrabutylammonium iodide (14 mg, 0.038 mmol) was stirred in neat benzyl bromide (0.60 mL) at rt for 10 min. Solid silver oxide (153 mg, 0.66 mmol) was then added. The reaction mixture was stirred at rt for 4 h, then diluted with CH2Cl2 and filtered through a pad of Celite. The filtrate was evaporated in vacuo to provide the crude bisbenzyl ether 132 as a brown viscous oil, which was used without further purification.

The crude bisbenzyl ether product 132, trimethysilyl cyanide (TMSCN, 1.9 mL, excess) and imidazole (8 mg, 0.12 mmol) were dissolved in CH2Cl2 (3 mL). The mixture was heated at reflux in a sealed tube for 72 h. The volatiles were removed in vacuo to provide a residue which was purified by flash chromatography on silica gel (gradient 5-30% EtOAc/hexanes) to afford the fully protected bisbenzyl TMS ether 133 as a colorless gum (28 mg, 24% over 2 steps). 1H

NMR (400 MHz, CDCl3) δ 8.80 (br s, 1H), 7.65 (d, J = 7.4 Hz, 2H), 7.62-7.50 (m, 3H), 7.44-

7.48 (m, 2H), 7.34-7.30 (m, 6H), 7.14 (app d, J = 7.8 Hz, 4H), 7.02 (dd, J = 7.6, 7.7 Hz, 1H),

4.76 (d, J = 12.3 Hz, 1H), 4.67 (d, J = 12.2 Hz, 1H), 4.54 (d, J = 10.1 Hz, 1H), 4.44 (d, J = 12.0,

1H), 4.35 (d, J = 12.1 Hz, 1H), 4.19 (d, J = 8.9 Hz, 1H), 4.03 (d, J = 7.7 Hz, 1H), 3.92 (d, J = 7.8

Hz, 1H), 3.86 (s, 1H), 3.80 (d, J = 8.9 Hz, 1H), 3.72 (s, 3H), 3.72-3.62 (m, 1H), 2.45 (m, 1H),

2.43 (s, 3H), 2.20 (m, 1H), 2.11 (d, J = 10.0 Hz, 1H), 1.96 (d, J = 12.4 Hz, 1H), 1.50 (m, 1H),

13 0.00 (s, 9H); C NMR (100 MHz, CDCl3) δ 173.4, 144.1, 139.2, 138.2, 136.8, 133.4, 132.6,

130.0, 128.7 (2C), 128.5, 128.3, 128.1, 128.0, 127.9, 127.7, 126.4, 121.8, 120.5, 119.2, 111.5,

189

107.2, 79.0, 76.5, 73.8, 73.3, 70.9, 56.4, 54.2, 52.6, 46.2, 37.8, 26.7, 22.0, 3.1; LRMS-ES+ m/z

(relative intensity) 767 (MH+).

Synthesis of Diol Silylether 134. To a solution of bisbenzyl ether 133 (5.3 mg, 0.0069 mmol) in t-BuOH/EtOAc (1:3 v/v, 3 mL) was added 10% Pd/C (6 mg) and glacial acetic acid

(0.4 µL, 0.0069 mmol). The reaction vessel was evacuated, backfilled with hydrogen gas from a balloon three times, and stirred under hydrogen gas from a balloon at rt for 12 h. The reaction mixture was then filtered through a pad of Celite and the filtrate was concentrated in vacuo. The residue was purified by flash chromatography on silica gel (gradient 30-40% EtOAc/hexanes) to

1 afford diol 134 as a white foamy solid (2.1 mg, 52%). H NMR (400 MHz, CDCl3) δ 8.79 (br s,

1H), 7.73 (d, J = 8.0 Hz, 1H), 7.68-7.64 (m, 3H), 7.37 (d, J = 8.0 Hz, 2H), 7.20 (d, J = 7.4 Hz,

1H), 7.13 (d, J = 7.9 Hz, 1H), 4.38-4.17 (m, 4H), 4.05 (m, 1H), 3.86 (m, 1H), 3.78 (s, 3H), 3.76

(m, 1H), 3.69 (m, 1H), 2.47 (s, 3H), 2.25 (m, 2H), 1.61 (m, 1H), 0.97 (m, 1H), 0.20 (s, 9H);

LRMS-ES+ m/z (relative intensity) 587 (MH+).

Synthesis of Cyclopropane TBS Ether 137. To a stirred solution of AgOTf (2.6 mg,

0.010 mmol) in CH2Cl2 (0.8 mL) at rt was added dropwise a solution of iodide 118 (4.0 mg,

190

0.0054 mmol) in CH2Cl2 (0.2 mL). After 15 min, the colorless solution became a yellow suspension. The reaction mixture was then filtered through a pad of Celite. The filtrate was concentrated in vacuo to give cyclopropane 137 as a white solid (3.0 mg, 91%). 1H NMR (400

MHz, CDCl3) δ 7.65 (d, J = 7.7 Hz, 1H), 7.57 (d, J = 8.1 Hz, 2H), 7.36 (dd, J = 7.5, 7.6 Hz, 1H),

7.30 (d, J = 8.1 Hz, 2H), 7.23 (m, 1H), 7.06 (d, J = 7.4 Hz, 1H), 4.16 (d, J = 9.7 Hz, 1H), 3.99 (d,

J = 9.9 Hz, 1H), 3.52 (m, 2H), 3.48 (br s, 1H), 2.83 (s, 3H), 2.72-2.70 (m, 2H), 2.62 (t, J = 7.9

Hz, 1H), 2.37 (s, 3H), 2.30 (m, 1H), 1.77 (dd, J = 4.8, 8.0 Hz, 1H), 0.82 (m, 1H), 0.75 (s, 9H), -

0.05 (s, 3H), -0.15 (s, 3H); LRMS-ES+ m/z (relative intensity) 611 (MH+, 100); HRMS (m/z):

+ [M + H] calcd for C32H42N2O6SSi, 611.2611; found 611.2602.

Synthesis of Cyclopropane Indoline 138. To a stirred solution of OTBS iodomethyl indole 118 (28 mg, 0.038 mmol) in CH2Cl2 (5.6 mL) at rt was added a solution of TBAF in THF

(1.0 M, 71 µL, 0.071 mmol). The reaction mixture was stirred at rt for 1.5 h, diluted with aqueous NH4Cl and extracted with CH2Cl2. The organic layer was washed with brine, dried over

Na2SO4, and concentrated in vacuo. The residue was purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes) to afford cyclopropane indoline 138 as a white solid (17

1 mg, 97%). H NMR (400 MHz, CDCl3) δ 9.57 (br s, 1H), 7.56 (d, J = 8.0 Hz, 2H), 7.27 (m, 2H),

7.17 (dd, J = 7.5, 7.6 Hz, 1H), 6.93 (dd, J = 7.5, 7.6 Hz, 1H), 6.85 (d, J = 7.7 Hz, 2H), 3.70 (s,

3H), 3.59 (d, J = 11.4 Hz, 1H), 3.47 (d, J = 11.9 Hz, 1H), 3.18 (s, 1H), 2.98 (d, J = 14.5 Hz, 1H),

2.56 (br s, 1H), 2.40 (s, 3H), 2.37-2.32 (m, 2H), 2.23-2.13 (m, 2H), 2.06 (m, 1H)1.40 (m, 1H);

13 C NMR (100 MHz, CDCl3) δ 169.2, 162.5, 144.2, 143.9, 133.6, 131.0, 130.1, 127.7, 127.6,

191

121.2, 118.7, 116.7, 114.0, 111.5, 109.4, 85.9, 69.4, 51.7, 50.8, 42.7, 34.4, 32.6, 32.5, 29.8, 23.4,

21.7, 20.6; LRMS-ES+ m/z (relative intensity) 467 (MH+, 100); HRMS (m/z): [M + H]+ calcd for

C25H27N2O5S, 467.1641; found 467.1671.

Synthesis of Fully Protected β-Lactone 139. To a solution of NH indole β-lactone

101 (28 mg, 0.045 mmol) in THF (2.5 mL) cooled to -78 oC was added a solution of KHMDS in toluene (0.5 M, 0.14 mL, 0.070 mmol). After the resulting yellow solution was stirred at -78 oC for 30 min, CbzCl (neat, 51 µL, 0.36 mmol) was added dropwise. The reaction mixture was allowed to warm to rt, stirred for another 5 h, then diluted with CH2Cl2. The organic phase was washed with aqueous NH4Cl, dried over Na2SO4, and concentrated in vacuo to give a residue which was purified by column chromatography on silica gel (CH2Cl2) to afford the N-Cbz β-

1 lactone 139 as a colorless gum (31 mg, 91%). H NMR (400 MHz, CDCl3) δ 7.98-7.95 (m, 1H),

7.66 (m, 3H), 7.48 (m, 2H), 7.42-7.29 (m, 6H), 5.46 (d, J = 12.0 Hz, 1H), 5.35 (d, J = 12.0 Hz,

1H), 4.71 (m, 2H), 4.41 (d, J = 10.1 Hz, 1H), 3.97 (d, J = 10.1 Hz, 1H), 3.82 (m, 1H), 3.47 (m,

1H), 3.39 (s, 3H), 2.95 (m, 2H), 2.46 (s, 3H), 2.43 (m, 1H), 1.64-1.59 (m, 1H), 0.73 (s, 9H), -

0.04 (s, 3H), -0.11 (s, 3H).

192

Synthesis of Diol 140. To a solution of N-Cbz-β-lactone 139 (10.0 mg, 0.013 mmol) in

THF (2.0 mL) at rt was added solid LiBH4 (2.0 mg, 0.091 mmol). The reaction mixture was stirred at rt for 3 h, diluted with aqueous NH4Cl, and extracted with CH2Cl2. The organic layer was dried over Na2SO4, and concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 2-5% Et2O/CH2Cl2) to afford diol 140 as a white solid

1 (6.4 mg, 64%). H NMR (400 MHz, CDCl3) δ 8.02 (br m, 1H), 7.87 (m, 1H), 7.65 (m, 2H), 7.46

(m, 2H), 7.40-7.30 (m, 7H), 5.46 (d, J = 13.2 Hz, 1H), 5.37 (d, J = 12.0 Hz, 1H), 4.71 (s, 1H),

4.43 (m, 1H), 4.35 (m, 1H) 4.22 (m, 2H), 3.78 (m, 1H), 3.56 (s, 1H), 3.36 (s, 3H), 2.44 (s, 3H),

2.12 (m, 2H), 0.65 (s, 9H), -0.03 (s, 3H), -0.20 (s, 3H); LRMS-ES+ m/z (relative intensity) 728

+ + (MH , 100); HRMS (m/z): [M + H] calcd for C40H51N2O9SSi, 763.3085; found 763.3121.

Synthesis of Bromo Alcohol 141. To a solution of PPh3 (39 mg, 0.15 mmol) and CBr4

(45 mg, 0.14 mmol) in CH2Cl2 at rt was added dropwise a solution of N-Cbz diol 140 (41 mg,

0.054 mmol) in CH2Cl2 (8.2 mL). The resulting cloudy reaction mixture was stirred at rt for 20 h.

The mixture was evaporated in vacuo to give a residue which was purified by preparative TLC

193 on silica gel (5% Et2O/CH2Cl2) to afford the bromo alcohol 141 as a white solid (10 mg, 26%).

1 H NMR (400 MHz, CDCl3) δ 8.05 (m, 1H), 7.67 (m, 3H), 7.48 (d, J = 7.6 Hz, 2H), 7.43-7.30

(m, 7H), 5.48 (d, J = 12.0 Hz, 1H), 5.35 (d, J = 12.0 Hz, 1H), 4.41 (br s, 2H), 4.22-4.11 (m, 2H),

4.03 (d, J = 6.6 Hz, 1H), 3.97 (d, J = 11.2 Hz, 1H), 3.66 (m, 1H), 3.33 (s, 3H), 2.45 (s, 3H), 2.28-

13 2.14 (m, 2H), 1.49 (m, 1H), 1.00-0.86 (m, 2H); C NMR (100 MHz, CDCl3) δ 175.0, 144.4,

135.0, 133.0, 130.2, 129.4, 129.3, 129.2, 128.4, 128.2, 125.7, 123.8, 120.2, 116.4, 112.3, 110.0,

+ 70.9, 69.5, 54.9, 52.0, 31.2, 27.3, 22.0; HRMS (m/z): [M + H] calcd for C34H36N2O8SBr,

711.1376; found 711.1417.

Synthesis of Oxetane Alcohol 142. To a solution of bromide 141 (1.6 mg, 0.0023 mmol) in CH2Cl2 (0.8 mL) was added activated 4Å molecular sieves (5.2 mg), silica gel (5.2 mg), solid

NaHCO3 (3.0 mg, 0.036 mmol) and solid AgOTf (12.2 mg, 0.047 mmol). The reaction mixture was stirred at rt for 2 h, then diluted with CH2Cl2 and filtered through a pad of Celite. The filtrate was concentrated in vacuo to afford N-Cbz oxetane 141 as a white solid (1.2 mg, 85%). 1H

NMR (400 MHz, CDCl3) δ 8.05 (d, J = 10.8 Hz, 1H), 7.67 (d, J = 7.2 Hz, 2H), 7.55-7.15 (m,

10H), 5.51 (d, J = 14.4 Hz, 1H), 5.37 (d, J = 14.4 Hz, 1H), 5.13 (m, 1H), 4.52 (d, J = 11.0 Hz,

1H), 4.43 (d, J = 11.0 Hz, 1H), 4.31 (t, J = 7.2 Hz, 1H), 3.98 (m, 1H), 3.86 (m, 2H), 3.66 (m,

1H), 3.37 (s, 3H), 2.46 (s, 3H), 2.35 (t, J = 7.2 Hz, 1H), 2.17-2.07 (m, 2H).

For the purpose of characterization, N-Cbz oxetane 141 (1.2 mg) and glacial HOAc (1.0

μL) were dissolved in EtOAc (1.0 mL) and hydrogenated with 10% Pd/C (5 mg) under hydrogen

194 gas from a balloon (1 atm) at rt for 1.5 h. The reaction mixture was then filtered through a pad of

Celite and the filtrate was evaporated to dryness in vacuo to give a white solid (0.8 mg, 85%), which was identical to oxetane 127.

Synthesis of Methylthiomethyl (MTM) Ether 147. To a solution of indole diol 74 (33.0 mg, 0.068 mmol) in anhydrous DMSO (0.21 mL) was added acetic anhydride (0.15 mL, excess) and glacial acetic acid (27 µL). The reaction mixture was stirred at rt for 22 h, quenched with

NaHCO3(aq.), and extracted with EtOAc. The organic layer was washed with water, brine, and dried over Na2SO4. The solvent was removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 30-50% EtOAc/hexanes) to afford the MTM ether

1 147 as an off-white foamy solid (12.5 mg, 34%). H NMR (400 MHz, CDCl3) δ 9.10 (br s, 1H),

7.68 (m, 3H), 7.37 (m, 3H), 7.17 (dd, J = 9.3, 10.7 Hz, 1H), 7.10 (dd, J = 9.2, 9.7 Hz, 1H), 4.75-

4.65 (m, 1H), 4.61 (d, J = 2.8 Hz, 1H), 4.60 (m, 1H), 4.56 (m, 1H), 4.28 (d, J = 11.7 Hz,, 1H),

4.18 (d, J = 15.3 Hz, 1H), 3.79 (s, 3H), 3.72 (m, 1H), 2.45 (s, 3H), 2.20 (app s, 1H), 1.85 (s, 3H),

0.96-0.85 (m, 2H); LRMS-ES+ m/z (relative intensity) 545 (MH+).

195

Synthesis of TBS Ether 148. MTM ether 147 (12.5 mg, 0.023 mmol), TBSOTf (0.11 mL, 0.46 mmol) and 2,6-lutidine (0.10 mL, 0.91 mmol) were dissolved in CH2Cl2 (2 mL). The reaction mixture was heated at reflux in a sealed tube for 48 h. The mixture was evaporated to dryness in vacuo to provide a residue which was purified by flash chromatography on silica gel

(gradient 2-5% Et2O/CH2Cl2, then 30% EtOAc/hexanes) to afford TBS ether 147 as a colorless

1 gum (12.0 mg, 81%). H NMR (300 MHz, CDCl3) δ 9.04 (br s, 1H), 7.82 (d, J = 7.6 Hz, 1H),

7.68 (d, J = 8.0 Hz, 2H), 7.36 (app d, J = 7.8 Hz, 3H), 7.19-7.09 (m, 2H), 4.87 (d, J = 11.4 Hz,

1H), 4.76 (d, J = 11.4 Hz, 1H), 4.55 (d, J = 10.5 Hz, 1H), 4.40 (m, 1H), 4.33 (dd, J = 4.3, 10.1

Hz, 1H), 3.93 (app t, J = 8.0 Hz, 1H), 3.79 (s, 3H), 3.68 (m, 1H), 2.46 (s, 3H), 2.31 (s, 3H), 0.81

+ (s, 9H), 0.18 (s, 3H), 0.12 (s, 3H); LRMS-ES+ m/z (relative intensity) 676 (M+NH4 , 100).

Synthesis of Thioester 163. To a solution of OTBS β-lactone 101 (55 mg, 0.088 mmol) in CH2Cl2 (10 mL) was added NEt3 (0.385 mL, 2.82 mmol) and ethanethiol (0.41 mL, 5.69 mmol), along with glacial HOAc (11μL, 0.19 mmol). The solution was stirred at rt for 36 h until

TLC indicated the consumption of starting material. The volatiles were removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 30-50%

196

EtOAc/hexanes) to afford thioester 163 (42 mg, 70%) as a white foam. 1H NMR (400 MHz,

CDCl3) δ 8.62 (br s, 1H), 7.67 (d, J = 7.9 Hz, 2H), 7.63 (d, J = 7.9 Hz, 1H), 7.33 (m, 3H), 7.17

(dd, J = 7.3, 7.5 Hz, 1H), 7.06 (dd, J = 7.3, 7.5 Hz, 1H), 4.82 (s, 1H), 4.45 (s, 1H), 4.28 (d, J =

8.8 Hz, 1H), 4.05 (d, J = 9.0 Hz, 1H), 3.91 (d, J = 11.2 Hz, 1H), 3.70 (s, 3H), 3.68 (m, 1H), 3.21-

3.11 (m, 2H), 2.44 (s, 3H), 2.147 (m, 2H), 1.88 (m, 2H), 1.64-1.55 (m, 2H), 1.47 (t, J = 7.2 Hz,

13 3H), 0.88 (s, 9H), -0.03 (s, 3H), -0.07 (s, 3H); C NMR (75 MHz, CDCl3) δ 205.1, 173.1, 144.4,

136.6, 133.4, 132.2, 130.1, 128.5, 125.8, 122.4, 119.9, 119.8, 111.6, 103.6, 72.7, 71.6, 55.6, 55.3,

52.5, 50.4, 47.1, 46.5, 31.1, 26.2, 26.0, 25.9, 25.1, 22.0, 19.6, 18.5, 18.4, 14.6, 14.2, -5.2, -5.5.

Synthesis of Aldehyde 164. Method A. To a solution of thioester 163 (24 mg, 0.035 mmol) in acetone (3 mL) was added 10% Pd/C (225 mg) and Et3SiH (1.1 mL, 6.9 mmol) in three equivalent portions over 6 h. The reaction mixture was then filtered through a pad of Celite. The filtrate was concentrated in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 5-40% EtOAc/hexanes) to afford OTBS aldehyde 164 as a white foam (18 mg, 80%).

Method B. Freshly prepared IBX (46 mg, 0.16 mmol) was added to a solution of diol

117 (97 mg, 0.16 mmol) in dry DMSO (5.3 mL) and the mixture was stirred at rt for 20 h. The reaction mixture was diluted with water and extracted with ether. The organic phase was washed with aqueous NaHCO3, brine, and dried over MgSO4. The solvent was removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 20-40%

197

EtOAc/hexanes) to afford aldehyde 164 as a white solid (91 mg, 94%). 1H NMR (300 MHz,

CDCl3) δ 10.26 (s, 1H), 8.74 (br s, 1H, NH), 7.68 (d, J = 8.0 Hz, 2H), 7.42-7.32 (m, 4H), 7.21

(dd, J = 7.5, 7.5 Hz, 1H), 7.11 (dd, J = 7.5, 7.5 Hz, 1H), 4.70 (br s, 1H), 4.42 (s, 1H), 4.20 (d, J =

9.5 Hz, 1H), 4.13-4.08 (m, 1H), 4.02 (d, J = 9.5 Hz, 1H), 3.74 (s, 3H), 3.70 (m, 1H), 2.44 (s, 3H),

2.29 (m, 1H), 2.23 (m, 1H), 2.00 (dd, J = 3.8, 11.5 Hz, 1H), 1.54 (dd, J = 4.2, 12.8 Hz, 1H), 1.43

13 (m, 1H), 0.79 (s, 9H), -0.09 (s, 3H), -0.14 (s, 3H); C NMR (75 MHz, CDCl3) δ 203.6, 172.8,

144.0, 136.9, 133.0, 131.4, 129.9, 127.9, 125.8, 122.4, 120.0, 119.4, 111.5, 103.7, 72.4, 70.5,

55.2, 55.1, 52.6, 48.9, 48.5, 45.7, 26.4, 25.7, 21.7, 18.1, -5.7, -5.8; HRMS (m/z): [M + H]+ calcd for C32H43N2O7SSi, 627.2560; found 627.2542.

Synthesis of Dithioacetal 165. To a solution of hydroxymethyl aldehyde 104 (16 mg,

o 0.032 mmol) and ethanethiol (69 µL, 0.96 mmol) in CH2Cl2 (2.5 mL) at -50 C was added a solution of TiCl4 in CH2Cl2 (1.0 M, 0.10 mL, 0.10 mmol). The resulting dark orange-brown mixture was stirred while allowed to warm slowly to rt over 15 h. The reaction mixture was then diluted with CH2Cl2, treated with aqueous NaHCO3, and extracted with CH2Cl2. The organic layers were washed with aqueous NH4Cl and brine, and dried over Na2SO4. The solvent was removed in vacuo and the residue was purified by flash chromatography on silica gel (gradient

10-50% EtOAc/hexanes) to afford dithioacetal 165 as a pink solid (18 mg, 94%). 1H NMR (300

MHz, CDCl3) δ 9.09 (br s, 1H), 8.15 (d, J = 6.3 Hz, 1H), 7.64 (d, J = 8.0 Hz, 2H), 7.36-7.33 (m,

3H), 7.22-7.12 (m, 2H), 4.70 (s, 1H), 4.19 (s, 3H), 4.13 (m, 1H), 3.63 (br m, 3H), 2.91 (m, 1H),

198

2.73 (q, J = 7.2 Hz, 4H), 2.44 (s, 3H), 2.23-2.12 (m, 3H), 1.31 (t, J = 7.2 Hz, 3H), 1.20 (t, J = 7.2

13 Hz, 3H); C NMR (75 MHz, CDCl3) δ 174.2, 144.5, 137.1, 132.7, 132.2, 130.2, 128.5, 127.0,

122.6, 122.0, 119.9, 111.8, 109.2, 77.7, 70.1, 56.1, 54.5, 52.9, 52.5, 28.7, 27.4, 22.0, 14.7 (2C);

+ HRMS (m/z): [M + Na] calcd for C30H38N2O6S3Na, 641.1790; found 641.1792.

Synthesis of Cyclic Hemithioacetal 167a/b. To a solution of dithioacetal 165 (130 mg,

0.21 mmol) in nitromethane (10 mL) at rt was added a mixture of 4Å MS and flame dried silica gel (1:1 wt. mixture, 390 mg), followed by solid NaHCO3 (79 mg, 0.94 mmol). The mixture was stirred at rt for 40 min and then was treated dropwise with a solution of anhydrous AgClO4 (131 mg, 0.63 mmol) in nitromethane (1.7 mL). The resulting brown suspension was stirred at rt until

TLC indicated completion of the reaction (~30 min). The mixture was diluted with CH2Cl2 and

EtOAc, filtered through a pad of Celite, and the filtrate was evaporated to dryness to give a yellow powder. This powder was stirred with 10 mL of CH2Cl2 at rt for 10 min and filtered to give the major C18 diastereomer 167a as a white solid (77 mg, 66%). 1H NMR (400 MHz,

DMSO-d6) δ 10.84 (br s, 1H), 7.44 (d, J = 7.5 Hz, 3H), 7.37 (d, J = 7.6 Hz, 1H), 7.17 (d, J = 7.9

Hz, 2H), 7.04 (app t, J = 7.0 Hz, 1H), 6.98 (app t, J = 7.4 Hz, 1H), 5.49 (d, J = 3.1 Hz, 1H), 5.26

(br s, 1H), 4.07 (d, J = 11.4 Hz, 1H), 3.89 (d, J = 11.4 Hz, 1H), 3.78 (s, 3H), 3.30 (d, J = 12.2 Hz,

2H), 3.22 (d, J = 3.1 Hz, 1H), 3.07 (t, J = 12.7 Hz, 1H), 2.46-2.36 (m, 3H), 2.30 (s, 3H), 2.37 (d,

J = 12.5 Hz, 1H), 1.51 (d, J = 11.0 Hz, 1H), 1.10 (t, J = 7.4 Hz, 3H), 0.57 (m, 1H); 13C NMR

199

(100 MHz, CDCl3) δ 172.2, 143.6, 136.6, 136.3, 135.8, 130.2, 127.6, 126.9, 121.2, 119.9, 118.1,

113.1, 111.7, 110.2, 82.4, 77.5, 68.3, 54.9, 53.5, 51.9, 48.3, 47.5, 42.2, 25.7, 25.4, 21.8, 16.0.

The filtrate was concentrated to give a residue which was purified by flash chromatography on silica gel (gradient 20-50% EtOAc/hexanes) to afford the minor C18

1 diastereomer 167b as an off-white solid (23 mg, 20%). H NMR (400 MHz, CDCl3) δ 9.48 (s,

1H), 7.63 (d, J = 8.1 Hz, 2H), 7.43 (d, J = 8.0 Hz, 1H), 7.36 (d, J = 7.2 Hz, 1H), 7.20 (d, J = 8.0

Hz, 2H), 7.11-7.16 (m, 2H), 5.58 (d, J = 4.4 Hz, 1H), 4.44 (d, J = 11.3 Hz, 1H), 4.02 (d, J = 11.3

Hz, 1H), 3.85 (s, 3H), 3.59 (d, J = 13.5 Hz, 1H), 3.30-3.36 (m, 2H), 3.28 (d, J = 4.5 Hz, 1H),

2.65 (d, J = 13.5 Hz, 1H), 2.62 (dd, J = 4.6, 13.0 Hz, 1H), 2.41-2.48 (m, 2H), 2.37 (s, 3H), 1.52

13 (m, 1H), 1.16 (t, J = 7.4 Hz, 3H), 0.97 (dd, J = 5.7, 13.5 Hz, 1H); C NMR (100 MHz, CDCl3) δ

172.5, 143.3, 136.0, 135.5, 135.2, 129.6, 127.5, 126.6, 121.9, 120.4, 117.9, 115.6, 112.8, 111.8,

109.9, 83.5, 77.9, 66.9, 54.6, 53.0, 50.5, 49.8, 46.8, 41.9, 25.4, 25.1, 21.6, 15,3; HRMS (m/z): [M

+ + H] calcd for C28H33N2O6S2, 557.1775; found 557.1780.

Hemithioacetals 167a and 167b exhibited identical reactivities in subsequent transformations. Whether using as a mixture or as a single diastereomer did not affect the chemical outcome of the subsequent transformations.

Synthesis of Indole Sulfone 168. To a suspension of hemithioacetal 167 (20 mg, 0.036

o mmol) in CH2Cl2 (5.0 mL) at 0 C was added a solution of m-CPBA (62 mg, 0.036 mmol) in

o CH2Cl2 (1.0 mL). The reaction mixture was stirred at 0 C for 1.5 h, then was diluted with

200

CH2Cl2 and treated with aqueous NaHCO3. The organic layer was separated and the aqueous layer extracted with CH2Cl2. The combined organic layer was washed with brine, dried over

Na2SO4, and concentrated in vacuo. The residue was purified by flash chromatography on silica gel (gradient 30-50% EtOAc/hexanes, then EtOAc) to give sulfone 168 as a white solid (13 mg,

1 62%). H NMR (400 MHz, CDCl3) δ 9.54 (s, 1H), 7.60 (app d, J = 8.0 Hz, 3H), 7.37 (d, J = 7.0

Hz, 1H), 7.23 (d, J = 8.0 Hz, 2H), 7.14 (m, 2H), 5.43 (br s, 1H), 4.09 (app br s, 2H), 3.96 (s, 1H),

3.87 (s, 3H), 3.53 (d, J = 13.5 Hz, 1H), 3.47 (m, 1H), 3.11 (dd, J = 5.9, 10.2 Hz, 1H), 2.88 (m,

1H), 2.75 (m, 1H), 2.58 (d, J = 13.5 Hz, 1H), 2.52 (dd, J = 4.3, 13.2 Hz, 1H), 2.36 (s, 3H), 1.41

13 (m, 1H), 1.22 (t, J = 7.5 Hz, 3H), 0.78 (m, 1H); C NMR (100 MHz, CD3OD) δ 172.0, 143.9,

135.6, 134.7, 132.3, 130.0, 127.2, 125.7, 122.0, 120.5, 118.6, 111.8, 106.7, 85.7, 76.1, 71.3, 53.9,

+ 53,4, 53.1, 52.0, 44.1, 41.8, 41.1, 25.0, 21.6, 5.4; HRMS (m/z): [M + NH4] calcd for

C28H36N3O8S2, 606.1944; found 606.1939.

Synthesis of Oxetane Acetal 171. Hemithioacetal 167 (1.8 mg, 0.0032 mmol) was heated with a slurry of freshly prepared W-2 Raney nickel in ethanol (~0.5 mL, excess) at 60 oC for 1 h. The reaction mixture was diluted with CH2Cl2 and filtered through a pad of Celite. The filtrate was concentrated to give oxetane acetal 171 as a colorless gum (1.0 mg, 63%). 1H NMR

(360 MHz, CDCl3) δ 9.49 (br s, 1H), 7.63 (d, J = 7.2 Hz, 2H), 7.44 (d, J = 7.0 Hz, 1H), 7.39 (d, J

= 7.2 Hz, 1H), 7.17 (m, 3H), 6.80 (m, 1H), 4.36 (m, 1H), 4.17 (d, J = 11.0 Hz, 1H), 4.03 (m, 2H),

3.85 (s, 3H), 3.58 (m, 2H), 3.32 (m, 2H), 3.06 (app s, 1H), 2.68 (d, J = 12.0 Hz, 1H), 2.60 (m,

201

+ 1H), 2.38 (s, 3H), 1.44 (m, 2H); HRMS (m/z): [M + H] calcd for C26H27N2O6S, 495.1590, found 495.1572.

Synthesis of Oxepane 136. To a solution of hemithioacetals 167 (13.6 mg, 0.0245 mmol) in acetone (5.0 mL) was added 10% Pd/C (75 mg) and triethylsilane (0.7 mL, excess). The reaction mixture was heated at reflux. Additional Pd/C and triethylsilane were added every 2 h until TLC indicated completion of the reaction (6 additions of 10% Pd/C and Et3SiH were required in total). The reaction mixture was then filtered through Celite and concentrated in vacuo to give a residue which was purified by flash column chromatography on silica gel

(gradient 5-10% EtOAc/hexanes, then 30% EtOAc/hexanes, then 5% Et2O/CH2Cl2, then 50%

EtOAc/hexanes) to afford oxepane 136 as a white foamy solid (11.4 mg, 94%). 1H NMR (400

MHz, CDCl3) δ 9.49 (s, 1H), 7.62 (d, J = 8.1 Hz, 2H), 7.44 (d, J = 7.6 Hz, 1H), 7.39 (d, J = 7.9

Hz, 1H), 7.21 (d, J = 8.1 Hz, 2H), 7.11-7.17 (m, 2H), 4.38 (dd, J = 3.7, 11.4 Hz, 1H), 4.17 (d, J =

11.4 Hz, 1H), 3.84 (s, 3H), 3.83 (d, J = 11.4 Hz, 1H), 3.58 (d, J = 13.4 Hz, 1H), 3.54 (dd, J = 1.8,

11.3 Hz, 1H), 3.25-3.37 (m, 2H), 3.06 (t, J = 1.8 Hz, 1H), 2.67 (d, J = 13.3 Hz, 1H), 2.60 (dd, J =

13 4.5, 13.1 Hz, 1H), 2.36 (s, 3H). 1.48 (m, 1H), 0.95 (m, 1H); C NMR (100 MHz, CDCl3) δ

173.0, 143.7, 135.8, 135.6, 135.4, 130.0, 127.8, 125.5, 122.0, 120.3, 117.9, 114.4, 112.5, 112.0,

74.2, 68.0, 55.4, 53.2, 52.7, 51.6, 42.4, 42.3, 25.5, 21.9, 18.4; HRMS (m/z): [M + H]+ calcd for

C26H29N2O6S, 497.1746; found 497.1753.

202

Synthesis of (±)-Alstilobanine E (14). Magnesium turnings (451 mg, excess) were added to a solution of oxepane 136 (20 mg, 0.040 mmol) in anhydrous methanol and the mixture was sonicated at rt until all the magnesium had dissolved (~45 min). The reaction mixture was

o then poured into aqueous NH4Cl at 0 C and extracted with CHCl3. The organic extract was dried over Na2SO4, and evaporated to dryness in vacuo. The residue was purified by flash chromatography on silica gel (gradient 1-5% MeOH/CH2Cl2, then 1% NEt3 in 5%

MeOH/CH2Cl2, then 1% NEt3 in 10% MeOH/ CH2Cl2) to afford racemic alstilobanine E (14) as

1 an off-white solid (13 mg, 95%). H NMR (400 MHz, CD3OD, 1 µL of TFA added) δ 7.90 (s,

1H), 7.47 (d, J = 8.6 Hz, 1H), 7.43 (d, J = 9.0 Hz, 1H), 7.10 (dd, J = 7.0, 7.4 Hz, 1H), 7.03 (dd, J

= 7.0, 7.3 Hz, 1H), 4.31 (dd, J = 3.0, 11.4 Hz, 1H), 4.14 (d, J = 11.4 Hz, 1H), 3.90 (s, 3H), 3.84

(d, J = 11.4 Hz, 1H), 3.57 (dd, J = 3.0, 11.6 Hz, 1H), 3.24 (m, 1H), 3.20 (m, 1H), 2.96 (d, J =

13.0 Hz, 1H), 2.74 (dd, J = 5.2, 13.0 Hz, 1H), 2.61 (d, J = 13.1 Hz, 1H), 1.69-1.74 (m, 1H), 1.19

13 (dddd, J = 5.9, 13.1, 13.3, 13.9 Hz, 1H); C NMR (100 MHz, CD3OD) δ 172.9, 137.8, 135.3,

126.4, 122.4, 120.6, 118.0, 113.0, 112.5, 75.6, 75.4, 67.2, 56.1, 53.2, 51.5, 49.0, 43.9, 39.8, 23.3;

+ HRMS (m/z): [M + H] calcd for C19H23N2O4, 343.1652; found 343.1658.

203

Synthesis of (±)-Angustilodine (13). Alstilobanine E (13) (10.4 mg, 0.0304 mmol) and paraformaldehyde (6.4 mg, ~7 equiv.) were stirred in methanol (3.0 mL) and treated with glacial acetic acid (7.5 µL) at rt. The resulting white suspension was stirred at rt for another 30 min, and solid NaBH4 was added every 60 min (4.0 mg, 0.105 mmol in each portion, 3 h total) until TLC

(1% NEt3 in 5% MeOH/CH2Cl2) indicated completion of the reaction. The reaction mixture was treated with aqueous NH4Cl, and was extracted with CHCl3. The combined organic layers were washed with aqueous NaHCO3, brine, and dried over Na2SO4. The solvent was removed in vacuo to give a residue which was purified by flash chromatography on silica gel (gradient 1-5%

MeOH/CH2Cl2, then 1% NEt3 in 5% MeOH/CH2Cl2) to afford angustilodine (13) as an off-white

1 gum (8.8 mg, 81%). H NMR (400 MHz, CDCl3) δ 9.55 (s, 1H), 7.47 (d, J = 7.6 Hz, 1H), 7.42 (d,

J = 7.8 Hz, 1H), 7.17 (app t, J = 8.0 Hz, 1H), 7.12 (app t, J = 7.6 Hz, 1H), 4.39 (dd, J = 2.1, 11.3

Hz, 1H), 4.12 (d, J = 11.4 Hz, 1H), 3.86 (d, J = 11.4 Hz, 1H), 3.84 (s, 3H), 3.55 (dd, J = 3.3, 11.3

Hz, 1H), 3.06 (app t, J = 3.0 Hz, 1H), 3.03 (m, 1H), 2.66 (dd, J = 4.4, 13.1 Hz, 1H), 2.36 (d, J =

11.9 Hz, 1H), 2.31 (d, J = 11.3 Hz, 1H), 2.28 (m, 1H), 2.26 (s, 3H), 1.32 (m, 1H), 0.98 (m, 1H);

13 C NMR (212.5 MHz, CDCl3) δ 173.3, 135.5, 135.0, 125.6, 121.2, 119.7, 117.4, 111.7, 111.3,

76.5, 72.9, 66.2, 61.0, 55.2, 52.6, 52.3, 49.9, 45.4, 41.9, 24.8; HRMS (m/z): [M + H]+ calcd for

C20H25N2O4, 357.1809, found 357.1806.

204

References

(1) (a) Ishikura, M.; Abe, T.; Choshi, T.; Hibino, S. Nat. Prod. Rep. 2013, 30, 694.

(b) Leonard, J. Nat. Prod. Rep. 1999, 16, 319. (c) Cordell, G. A. Introduction to

Alkaloids: A Biogenetic Approach; Wiley-Interscience: New York, 1981. (d) Herbert, R. B. In

The Chemistry of Heterocyclic Compounds; Saxton, J. E., Ed.; Wiley: New York, 1983.

(2) Qureshi, A. A.; Scott, A. I. J. Chem. Soc. Chem. Comm. 1968, 945.

(3) (a) Zheng, H. Y.; Rinaman, L. Euro. J. Neurosci. 2013, 37, 1340. (b) Yokota, S.;

Hikasa, Y.; Shimura, I.; Kusunose, S. Am. J. Vet. Res. 2013, 74, 395. (c) Sandberg, M.;

Pettersson, U.; Henriksnas, J.; Jansson, L. Horm. Metab. Res. 2013, 45, 252. (d) Porst, H.;

Burnett, A.; Brock, G.; Ghanem, H.; Giuliano, F.; Glina, S.; Hellstrom, W.; Martin-Morales, A.;

Salonia, A.; Sharlip, I.; Med, I. S. C. S. J. Sex. Med. 2013, 10, 130.

(4) (a) Aboulenein, H. Y.; Muhtadi, F. J. Drugs Exp. Clin. Res. 1982, 8, 97. (b)

Kallioma.Jl; Kasanen, A. Curr. Ther. Res. Clin. E. 1969, 11, 344. (c) Mcqueen, E. G.; Doyle, A.

E.; Smirk, F. H. Nature 1954, 174, 1015.

(5) (a) Silvestri, R. J. Med. Chem. 2013, 56, 625. (b) Kavoi, B. M.; Makanya,

A. N.; Kiama, S. G. Anat. Histol. Embryol. 2012, 41, 374. (c) Tanaka, H.; Matsushima, H.;

Mizumoto, N.; Takashima, A. J. Invest. Dermatol. 2006, 126, 38. (d) Greig, N. H.; Soncrant, T.

T.; Shetty, H. U.; Momma, S.; Smith, Q. R.; Rapoport, S. I. Cancer Chemoth. Pharm. 1990, 26,

263. (e) Magnus, P.; Ladlow, M.; Elliott, J. J. Am. Chem. Soc. 1987, 109, 7929.(f) Gupta, R. S.

Cancer Treat. Rep. 1985, 69, 515. (g) Fujita, J.; Tokuda, H.; Sugawara, K.; Yoshida, O.

Toxicol. Lett. 1984, 22, 261.

(6) Caventou, J. B.; Pelletier, P.-J. Ann. Chim. Phys. 1818, 8, 323.

205

(7) Li, L. M.; Yang, T.; Liu, Y.; Liu, J.; Li, M. H.; Wang, Y. T.; Yang, S. X.; Zou, Q.;

Li, G. Y. Org. Lett. 2012, 14, 3450.

(8) Wu, M.; Wu, P.; Xie, H. H.; Wu, G. J.; Wei, X. Y. Planta Med. 2011, 77, 284.

(9) (a) Weinreb, S. M. Acc. Chem. Res. 2003, 36, 59. (b) Weinreb, S. M. Chem.

Rev. 2006, 106, 2531.

(10) (a) Woodward, R. B.; Cava, M. P.; Hunger, A.; Ollis, W. D.; Daeniker, H. U.;

Schenker, K. Tetrahedron 1963, 19, 247. (b) Woodward, R. B.; Cava, M. P.; Ollis, W. D.;

Hunger, A.; Daeniker, H. U.; Schenker, K. J. Am. Chem. Soc. 1954, 76, 4749.

(11) (a) Cannon, J. S.; Overman, L. E. Angew. Chem., Int. Ed. 2012, 51, 4288. (b)

Roth, K. Chem. Unserer Zeit 2011, 45, 202. (c) Martin, D. B. C.; Vanderwal, C. D. Chem. Sci.

2011, 2, 649. (d) Mori, M. Heterocycles 2010, 81, 259. (e) Beemelmanns, C.; Reissig, H. U.

Angew. Chem., Int. Ed. 2010, 49, 8021. (f) Zhang, H. J.; Boonsombat, J.; Padwa, A. Org.

Lett. 2007, 9, 279. (g) Ohshima, T.; Xu, Y. J.; Takita, R.; Shibasaki, M. Tetrahedron 2004,

60, 9569. (h) Kaburagi, Y.; Tokuyama, H.; Fukuyama, T. J. Am. Chem. Soc. 2004, 126,

10246. (i) Ohshima, T.; Xu, Y. J.; Takita, R.; Shimizu, S.; Zhong, D. F.; Shibasaki, M. J. Am.

Chem. Soc. 2003, 125, 2014. (j) Ohshima, T.; Xu, J. Y.; Takita, R.; Shimizu, S.; Zhong, D. F.;

Shibasaki, M. J. Am. Chem. Soc. 2002, 124, 14546. (k) Nakanishi, M.; Mori, M. Angew. Chem.,

Int. Ed. 2002, 41, 1934. (l) Bodwell, G. J.; Li, J. Angew. Chem., Int. Ed. 2002, 41, 3261.

(m) Sole, D.; Bonjoch, J.; Garcia-Rubio, S.; Peidro, E.; Bosch, J. Chem. Eur. J. 2000, 6,

655. (n) Eichberg, M. J.; Dorta, R. L.; Lamottke, K.; Vollhardt, K. P. C. Org. Lett. 2000, 2,

2479. (o) Vollhardt, K. P. C.; Eichberg, M. J.; Dorta, R. L.; Lamottke, K. Abstr. Pap. Am. Chem.

Soc. 1999, 218, U36. (p) Sole, D.; Bonjoch, J.; Garcia-Rubio, S.; Peidro, E.; Bosch, J. Angew.

Chem., Int. Ed. 1999, 38, 395. (q) Knight, S. D.; Overman, L. E.; Pairaudeau, G. J. Am.

206

Chem. Soc. 1995, 117, 5776. (r) Kuehne, M. E.; Xu, F. J. Org. Chem. 1993, 58, 7490. (s)

Knight, S. D.; Overman, L. E.; Pairaudeau, G. J. Am. Chem. Soc. 1993, 115, 9293.

(12) Kam, T. S.; Choo, Y. M. Helv. Chim. Acta 2004, 87, 366.

(13) Koyama, K.; Hirasawa, Y.; Zaima, K.; Hoe, T. C.; Chan, K. L.; Morita, H. Bioorg.

Med. Chem. 2008, 16, 6483.

(14) Kam, T. S., Personal Communication.

(15) (a) Cortez, G. S.; Tennyson, R. L.; Romo, D. J. Am. Chem. Soc. 2001, 123, 7945.

(b) Cortez, G. S.; Oh, S. H.; Romo, D. Synthesis 2001, 1731. (c) Henry-Riyad, H.;

Lee, C.; Purohit, V. C.; Romo, D. Org. Lett. 2006, 8, 4363. (d) Oh, S. H.; Cortez, G. S.; Romo,

D. J. Org. Chem. 2005, 70, 2835.

(16) (a) Gilchrist, T. L. Chem. Soc. Rev. 1983, 12, 53. (b) Adam, W.; Krebs, O.

Chem. Rev. 2003, 103, 4131. (c) Gowenlock, B. G.; Richter-Addo, G. B. Chem. Rev. 2004, 104,

3315. (d) de los Santos, J. M.; Ignacio, R.; Rubiales, G.; Aparicio, D.; Palacios, F. J. Org. Chem.

2011, 76, 6715.

(17) Majireck, M. M. Ph.D Dissertation, The Pennsylvania State University, 2011.

(18) (a) Denmark, S. E.; Dappen, M. S. J. Org. Chem. 1984, 49, 798. (b) Denmark,

S. E.; Dappen, M. S.; Sternberg, J. A. J. Org. Chem. 1984, 49, 4741.

(19) (a) Korboukh, I.; Kumar, P.; Weinreb, S. M. J. Am. Chem. Soc. 2007, 129, 10342.

(b) Li, P. H.; Majireck, M. M.; Witek, J. A.; Weinreb, S. M. Tetrahedron Lett. 2010, 51,

2032. (c) Kumar, P.; Li, P. H.; Korboukh, I.; Wang, T. L.; Yennawar, H.; Weinreb, S. M. J.

Org. Chem. 2011, 76, 2094. (d) Sengupta, R.; Witek, J. A.; Weinreb, S. M. Tetrahedron 2011,

67, 8229. (e) Witek, J. A.; Weinreb, S. M. Org. Lett. 2011, 13, 1258.

(20) Chauhan, P. S.; Majireck, M. M.; Weinreb, S. M. Heterocycles 2012, 84, 577.

207

(21) Mahboobi, S.; Bernauer, K. Helv. Chim. Acta 1988, 71, 2034.

(22) Evans, D. A. evans.harvard.edu/pdf/evans_pka_table.pdf; Vol. 2013.

(23) Hlasta, D. J.; Luttinger, D.; Perrone, M. H.; Silbernagel, M. J.; Ward, S. J.;

Haubrich, D. R. J. Med. Chem. 1987, 30, 1555.

(24) Davidson, J. P.; Corey, E. J. J. Am. Chem. Soc. 2003, 125, 13486.

(25) Majireck, M. M.; Witek, J. A.; Weinreb, S. M. Tetrahedron Lett. 2010, 51, 3555.

(26) Smitrovich, J. H.; Woerpel, K. A. J. Org. Chem. 1996, 61, 6044.

(27) Nakagawa, M.; Kato, S.; Kataoka, S.; Hino, T. J. Am. Chem. Soc. 1979, 101, 3136.

(28) Depuy, C. H.; Ponder, B. W. J. Am. Chem. Soc. 1959, 81, 4629.

(29) (a) Martin, C. L.; Overman, L. E.; Rohde, J. M. J. Am. Chem. Soc. 2010, 132,

4894. (b) Martin, C. L.; Overman, L. E.; Rohde, J. M. J. Am. Chem. Soc. 2008, 130, 7568.

(30) Schlosser, M.; Jenny, T.; Guggisberg, Y. Synlett 1990, 704.

(31) Mozingo, R. Org. Synth. 1941, 21, 15.

(32) Alonso, D. A.; Andersson, P. G. J. Org. Chem. 1998, 63, 9455.

(33) Oppolzer, W.; Bienayme, H.; Genevoisborella, A. J. Am. Chem. Soc. 1991, 113,

9660.

(34) Huscroft, I. T.; Carlson, E. J.; Chicchi, G. G.; Kurtz, M. M.; London, C.; Raubo,

P.; Wheeldon, A.; Kulagowski, J. J. Bioorg. Med. Chem. Lett. 2006, 16, 2008.

(35) (a) Zunk, M.; Kiefel, M. J. Tetrahedron Lett. 2011, 52, 1296. (b) Kato, H.;

Ohmori, K.; Suzuki, K. Tetrahedron Lett. 2000, 41, 6827.

(36) (a) Feng, Y. Q.; Majireck, M. M.; Weinreb, S. M. Angew. Chem., Int. Ed. 2012,

51, 12846. (b) Feng, Y. Q.; Majireck, M. M.; Weinreb, S. M. J. Org. Chem. 2014, 79, 7.

208

(37) We thank Prof. Morita for providing proton and carbon spectra of natural alstilobanine A.

(38) Nakao, R.; Tsurugi, J.; Fukumoto, T. J. Org. Chem. 1972, 37, 76.

(39) Nicolaou, K. C.; Mcgarry, D. G.; Somers, P. K.; Kim, B. H.; Ogilvie, W. W.;

Yiannikouros, G.; Prasad, C. V. C.; Veale, C. A.; Hark, R. R. J. Am. Chem. Soc. 1990, 112, 6263.

(40) Prange, T.; Rodriguez, M. S.; Suarez, E. J. Org. Chem. 2003, 68, 4422.

(41) Popsavin, V.; Benedekovic, G.; Sreco, B.; Francuz, J.; Popsavin, M.; Kojic, V.;

Bogdanovic, G.; Divjakovic, V. Tetrahedron 2009, 65, 10596.

(42) Johansen, J. E.; Christie, B. D.; Rapoport, H. J. Org. Chem. 1981, 46, 4914.

(43) Corey, E. J.; Jones, G. B. J. Org. Chem. 1992, 57, 1028.

(44) (a) Schreiber, S. L.; Wang, Z. Y.; Schulte, G. Tetrahedron Lett. 1988, 29, 4085.

(b) Hatakeyama, S.; Sakurai, K.; Saijo, K.; Takano, S. Tetrahedron Lett. 1985, 26, 1333.

(c) Jiang, L.; Chan, T. H. Tetrahedron Lett. 1998, 39, 355. (d) Hernandez-Torres, J. M.;

Achkar, J.; Wei, A. J. Org. Chem. 2004, 69, 7206.

(45) Nicolaou, K. C.; Prasad, C. V. C.; Hwang, C. K.; Duggan, M. E.; Veale, C. A. J.

Am. Chem. Soc. 1989, 111, 5321.

(46) Pojer, P. M.; Angyal, S. J. Aust. J. Chem. 1978, 31, 1031.

(47) Fleming, F. F.; Zhang, Z. Y.; Wang, Q. Z.; Steward, O. W. J. Org. Chem. 2003,

68, 7646.

(48) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.

(49) Levin, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun. 1982, 12, 989.

(50) (a) Fukuyama, T.; Lin, S. C.; Li, L. P. J. Am. Chem. Soc. 1990, 112, 7050. (b)

Kimura, M.; Seki, M. Tetrahedron Lett. 2004, 45, 3219.

209

(51) Hatch, R. P.; Weinreb, S. M. J. Org. Chem. 1977, 42, 3960.

(52) Gerfaud, T.; Xie, C. S.; Neuville, L.; Zhu, J. P. Angew. Chem., Int. Ed. 2011, 50,

3954.

(53) Boeckman, R. K., Jr.; Shao, P.; Mullins, J. J. Org. Synth. 2000, 77, 141.

(54) (a) Bulman-page, P. C.; Roberts, R. A.; Paquette, L. A. Tetrahedron Lett. 1983,

24, 3555. (b) Kumar, V.; Dev, S. Tetrahedron Lett. 1983, 24, 1289.

(55) Blom, P.; Ruttens, B.; Van Hoof, S.; Hubrecht, I.; Van der Eycken, J.; Sas, B.;

Van hemel, J.; Vandenkerckhove, J. J. Org. Chem. 2005, 70, 10109.

(56) Schonafinger, K.; Yasenchak, C. M.; Vollman, A.; Ong, H. H. J. Heterocycl.

Chem. 1988, 25, 535.

(57) We thank Professor Kam for providing authentic proton and carbon NMR spectra of natural angustilodine (13).

210

Vita

Yiqing Feng

Yiqing “Jerry” Feng was born and raised in Shanghai, China. He attended Fudan

University and obtained his bachelor’s degree in chemistry in 2005. After a brief research associate appointment at the School of Pharmacy, Jiaotong University, he became a graduate student in Department of Chemistry at the Pennsylvania State University in 2007. At Penn State, he joined Prof. Gong Chen’s group, where he completed a total synthesis of celogentin C, developed C-H activation methodology for benzannulation and was involved in a total synthesis of mannopeptimycin. In 2011, he continued his graduate research with Prof. Steven M. Weinreb, and completed total syntheses of alstilobanines A, E and angustilodine. After graduation, he will begin a postdoctoral position with Prof. Dale L. Boger at the Scripps Research Institute in La

Jolla, California.