Synthetic Studies of the Yunnaneic Acids

Daniel R. Griffith

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2013

© 2013 Daniel R. Griffith All rights reserved ABSTRACT

Synthetic Studies of the Yunnaneic Acids

Daniel R. Griffith

Chapter 1. Introduction.

Caffeic acid and its metabolites are widely distributed throughout the plant kingdom.

Nature exploits the reactivity of the catechol and unsaturated acid functionalities of multiple caffeic acid units to create a diverse library of molecules, many of which are biologically active.

Like other plant-based polyphenols, this biological activity often manifests as anti-oxidant activity. The reason for such activity usually lies in the catechol group common to all of these natural products. Biosynthetic pathways to these compounds from caffeic acid, which itself arises via the shikimic acid pathway, are discussed. The proposed biosynthesis of the yunnaneic acids is also discussed, and a possible alternative is put forward, informed by the recent isolation of rufescenolide. Selected total syntheses of some of these caffeic acid metabolites are described, providing a backdrop for a discussion of the unique synthetic challenges presented by the yunnaneic acids.

Chapter 2. Synthetic Studies of Yunnaneic Acids C and D

The total syntheses of yunnaneic acids C and D were completed using a lead-mediated oxidative dearomatization/Diels–Alder reaction cascade to forge the bicyclo[2.2.2]octene core of the molecules in one step from simple precursors. The route featuring this key step was developed after a first-generation route using hypervalent iodine to effect dearomatization had been explored and was found to be fraught with inefficiencies. The tricyclic key intermediate resulting from the lead-mediated cascade sequence was also exploited to complete the total synthesis of rufescenolide. Chapter 3. Synthetic Studies of Yunnaneic Acids A and B

Model studies towards the dimeric yunnaneic acids A and B were undertaken, which resulted in the characterization of several novel dimeric compounds whose connectivity and/or stereochemistry did not correspond to that found in the natural products. Although the total synthesis of yunnaneic acids A and B could not be realized, a selective pseudodimerization between model versions of yunnaneic acids C and D was accomplished, with no competing homodimerization, by exposing the two dimerization partners to hexafluoroisopropanol, a protic, non-nucleophilic solvent. TABLE OF CONTENTS

Chapter 1. Introduction ...... 1

1.1 Caffeic Acid Metabolites ...... 2

1.2 Biological Significance of Plant-Based Polyphenols and General Mechanism of

Action ...... 5

1.3 Biosynthesis of Caffeic Acid and Some of its Derivatives ...... 8

1.3.1 Caffeic Acid ...... 8

1.3.2 Rosmarinic Acid ...... 9

1.3.3 Podophyllotoxin ...... 11

1.3.4 Lithospermic Acid ...... 11

1.3.5 Helicterin Natural Products ...... 12

1.3.6 Yunnaneic Acids ...... 16

1.4 Synthetic Studies of Caffeic Acid Metabolites–Selected Examples ...... 22

1.4.1 Podophyllotoxin ...... 22

1.4.2 Lithospermic Acid ...... 24

1.4.3 Helicterin Natural Products ...... 26

1.5 Yunnaneic Acids: A Unique Set of Synthetic Challenges ...... 28

1.6 References ...... 30

Chapter 2. Synthetic Stuides of Yunnaneic Acids C and D ...... 34

2.1 Early Synthetic Explorations ...... 35

2.2 Second Generation Synthesis of Model Yunnaneic Acid C and Successful Approach

to Model Yunnaneic Acid D ...... 45

i 2.3 Total Synthesis of Rufescenolide ...... 53

2.4 Total Synthesis of Yunnaneic Acids C and D ...... 55

2.5 Conclusion ...... 62

2.6 References ...... 65

2.7 Experimental Section ...... 68

Chapter 3. Synthetic Stuides of Yunnaneic Acids A and B ...... 186

3.1 Dimerization Strategy and Literature Precedent ...... 187

3.2 Dimerization Studies Towards Yunnaneic Acid A ...... 194

3.3 Dimerization Studies Towards Yunnaneic Acid B ...... 199

3.4 Conclusion ...... 200

3.5 References ...... 202

3.6 Experimental Section ...... 204

ii ACKNOWLEDGMENTS

First and foremost, I would like to acknowledge Prof. Scott Snyder for being such a supportive mentor. Total synthesis can be a real battle, especially when working on a project alone. Whenever I thought that I was not equal to the task, Scott always responded with encouragement and with total belief in my abilities as a chemist. Scott also provided me with total freedom to pursue the ideas that excited me the most, which helped immeasurably in my growth as a scientist. After five years as a member of the Snyder lab, I can confidently say that pursuing a Ph.D. in chemistry was the best decision that I ever made, and that could not have been the case without a great mentor to guide those five years.

I would like to thank my family–Mom, Dad, and Allison–for their love and encouragement. I could not have made it to this point without their support. My parents’ continual assertion that “everything will work out” became my mantra throughout graduate school. I am also grateful to Allison for deciding to come to Columbia for her Ph.D. studies.

Having a supportive family member so nearby the past few years has enhanced my graduate school experience tremendously. I have also had the privilege of seeing her scientific development up close, which has made me one proud older brother.

I would like to thank Dr. Lorenzo Botta for his help in making the synthesis of yunnaneic acid C a reality and his diligent efforts in the dimerization studies towards yunnaneic acid B.

After three years of struggling on my own, it was a tremendous relief to have his help on this project. I could not have had a nicer guy to work with, and he made the difficulties associated with this project much more bearable. I would also like to thank Tyler St. Denis for being such an enthusiastic undergraduate and for his work on rufescenolide. Tyler was always eager to listen to my half-baked ideas and then do the necessary literature searching to make them work in the

iii lab. I would also like to acknowledge Dr. Ferenc Kontes for his help in the early stages of the project and, most importantly, for training me when I started in the group and serving as an invaluable source of advice throughout graduate school.

I would also like to acknowledge the members of the Snyder group for being such an awesome group of co-workers. I could not have put in the commitment to research that I did without such a great environment in which to do chemistry. In particular, I would like to acknowledge the following individuals: Myles Smith (for being a great hood-mate and sharing his enthusiasm for chemistry along with his encyclopedic knowledge of the literature); Dr. Adel

ElSohly (for being essentially a second advisor; I look forward to reuniting in Berkeley); Trevor

Sherwood (a true brother-in-arms; BMS is a lucky company); Dr. Daniel Treitler and Nathan

Wright (for being great company during the early mornings in lab and for being great friends).

I would like to thank the members of the Parkin group for their help with X-ray diffraction. I would truly be lost without them. In particular, I would like to thank Ahmed al-

Harbi, Yi Rong, Ava Kreider-Mueller, and, last but not least, Drs. Aaron and Wesley Sattler.

Aaron and Wes are both exceptionally passionate about chemistry, which definitely rubbed off on me. We are all lucky that there are two of them.

I would like to thank Prof. James Leighton and Prof. Dalibor Sames for serving on my committee and their feedback over the past five years. I would also like to thank Prof. Tristan

Lambert and Prof. Jack Norton for helpful discussions about chemistry and my future career.

I would also like to thank Dr. John Decatur, Dr. Yasahiro Itagaki, and Luis Avila for their help with NMR, mass spectrometric, and infrared analysis, respectively.

Finally, I thank the Department of Defense for a fellowship.

iv

Dedicated to my family

v

“Is not the true purpose of education to help you find out, so that as you grow up you can give

your whole mind, heart, and body to what you really love to do?”

–J. Krishnamurti

“Sometimes you climb out of bed in the morning and think, I’m not going to make it, but you

laugh inside–remembering all the times you’ve felt that way.”

–Charles Bukowski

vi LIST OF ABBREVIATIONS

Ac acetyl

BDE bond dissociation energy

Bn benzyl brsm based on recovered starting material calcd calculated

Cy cyclohexyl

DCC N,N’-dicyclohexylcarbodiimide

DFT density functional theory

DIAD diisopropyl azodicarboxylate

DIBAL-H diisobutylaluminum hydride

4-DMAP 4-dimethylaminopyridine

DMF N,N-dimethylformamide

DNA deoxyribonucleic acid

DOSP (N-dodecylbenzenesulfonyl)prolinate

EDC N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride equiv equivalents eV electron volts

EWG electron withdrawing group

FAB fast atom bombardment

Fc ferrocenyl

FMO frontier molecular orbital h hours

vii HAT hydrogen atom transfer

HFIP 1,1,1,3,3,3-hexafluoro-2-propanol

HIV human immunodeficiency virus

HOMO highest occupied molecular orbital

HPLC high pressure liquid chromatography

HPPR hydroxyphenylpyruvate reductase

HRMS high resolution mass spectrometry

Hz hertz

Ile isoleucine

IR infrared

LRMS low resolution mass spectrometry

LUMO lowest unoccupied molecular orbital

MALDI matrix assisted laser desorption ionization min minutes

MOB masked ortho-benzoquinone

NBS N-bromosuccinimide

NIH National Institutes of Health

NMO N-methylmorpholine-N-oxide

NMR nuclear magnetic resonance

PAL L-phenylalanine ammonia lyase

PIFA phenyliodine bis(trifluoroacetate) py pyridine

ROS reactive oxygen species

viii SET single electron transfer

TAT tyrosine aminotransferase

TBAF tetra-n-butylammonium fluoride

TBS tert-butyldimethylsilyl

TEMPO (2,2,6,6-tetramethylpiperidin-1-yl)oxy radical

Tf trifluoromethanesulfonyl

TFA trifluoroacetic acid

TfBn p-(trifluoromethyl)benzyl

THF tetrahydrofuran

TLC thin layer chromatography

TMS trimethylsilyl

TPAP tetra-n-propylammonium perruthenate

ix 1

CHAPTER 1

Introduction

2

1.1 Caffeic Acid-based Metabolites

Caffeic acid (1) is a ubiquitous catechol found throughout the plant kingdom, most likely reflecting its role in the biosynthesis of lignin, an integral component of plant cell walls.1

Because of its wide distribution in plants, it is unsurprising that human consumption of this compound is common. In particular, coffee beans are the most prevalent source of caffeic acid in the human diet (the blood plasma levels of caffeic acid in regular coffee drinkers has been measured to be as high as 505 ng/L).2 In nature, caffeic acid is combined in myriad ways by plants to generate a diverse array of molecular structures, some notable examples of which are shown in Figure 1.

Many of these compounds are biosynthesized by plants that have been used in traditional medicine.3 Most of these compounds have anti-oxidant activity, as is true for many plant-derived polyphenols, but many of them also shown promising biological activity in other ways (see

Section 1.2 for more on the biological activity of these compounds). One of the most widely distributed of these caffeic acid derivatives is rosmarinic acid (2), which is formally a caffeic acid dimer.4 It is found in numerous different plant families, but it derives its name from

Rosmarinus officinalis, the plant from which it was first isolated in pure form.5 There have been numerous biological activities described for this compound, including anti-inflammatory and antiviral activity.4 3

OH OH O CO H CO2H 2 OH HO O O O OH HO O OH OH HO OH CO2H OH O CO2H

CO2H OH HO 1: caffeic acid O OH OH OH OH OH O O 3: lithospermic acid O 4: rabdosiin O CO2H O O OH OH OH OH OH OH HO MeO OMe OH OMe 2: rosmarinic acid O O O O 5: podophyllotoxin O O MeO C MeO C 2 CO Me 2 OH 2 MeO2C O OH O O CO2Me MeO C H O 2 O O OH OMe O OR O O HO HO MeO2C OH OH OH OH OH OH OH MeO2C O OH 9: helisterculin A O O OH 7: R = Me, helicterin A

8: R = H, helicterin B HO OH HO OH

6: helisorin HO OH HO OH OH OH HO OH

CO2H CO2H CO2H

CO2H O O O O O HO O OH O O HO HO OH HO O HO HO O H HO O O R O CO2H OH CO2H CO2H

CO H O 11: R = H, yunnaneic acid A 2 13: yunnaneic acid D 10: yunnaneic acid C 12: R = OH, yunnaneic acid B

Figure 1. Selected examples of caffeic acid-derived natural products

4

One of the most notable caffeic acid metabolites is lithospermic acid (3), which has also

been isolated from multiple plant sources.6,7 One is the dried roots of Salvia miltiorrhizae, known

in the world of traditional Chinese medicine as “Dan shen” and has been used widely to treat

cardiovascular disease.8-10 In addition, Lithospermum ruderale, a plant that also contains

lithospermic acid, has been used by certain Native American tribes to make a contraceptive tea.6

Indeed, extracts from this and other plants containing lithospermic acid have been shown to have

antihormonal activities. Perhaps most intriguingly, this molecule has also been shown to inhibit

HIV-1 integrase,11 an enzyme used by the virus to integrate its DNA into the host genome, which

is also viewed as a valuable therapeutic target.

In contrast to lithospermic acid, which is a neolignan, rosmarinic acid can also dimerize

to give lignan-type structures such as that of rabdosiin (4).12,13 Much like its neolignan cousin,

rabdosiin has shown broad biological activity. It is a constituent of numerous traditional

medicines used in Japan and China for the treatment of gastrointestinal disorders.12,14 In a more modern setting, rabdosiin has been shown to possess anti-HIV activity14 and to inhibit DNA

topoisomerases I and II.15

Podophyllotoxin (5) is a tumor-inhibiting lignan whose structure, derived from two

molecules of ferulic acid (the 3-methylated variant of caffeic acid), features a trans-lactone and

four contiguous stereocenters.16 This lignin has been shown to inhibit microtubule assembly and has served as a model for multiple clinical anticancer agents.17

The members of the helicterin family (6–9) represent an even greater degree of molecular complexity in the broader family of caffeic acid derivatives. These molecules, isolated from the fruits of the shrub Helicteres isora,18,19 can be envisioned as Diels–Alder adducts of two

molecules of rosmarinic acid. These Diels–Alder adducts can then dimerize to form helicterins A 5

(7) or B (8), which are formally octamers of caffeic acid. The fruits of the shrubs from which

these natural products were isolated have been used as anti-convulsants, and the natural products

themselves possess weak to mild inhibitory activity against avian myeloblastosis virus reverse

transcriptase.19 This antiviral activity is notable, given that extracts from the fruits of this shrub have been reported to have anti-HIV activity.20

Finally, along with the helicterins, the yunnaneic acids (10–13), isolated in 1996 from

Salvia yunanensis in southern China by Tanaka and Kouno, et al.,21 represent some of the most structurally ornate caffeic acid metabolites isolated to date (for an extended discussion of yunnaneic acid structure, see Section 1.5). Although Salvia yunnanensis has been used as an alternative to Salvia miltorrhizae in traditional Chinese medicine,22 there is scant information on

the biological activity of the yunnaneic acids in the literature.23 In addition, unlike the molecules

described above, no total synthesis of any of the yunnaneic acids has been reported. Thus, the

structural complexity and the lack of a previous synthesis served as two of the reasons for our

interest in these molecules from a synthetic standpoint. More specifically, we realized that the

subtle structural differences between the yunnaneic acids and the helicterins would require the

development of novel synthetic strategies quite different than those deployed in the total

syntheses of the helicterins (see Section 1.4.3; see Section 1.5 for a detailed discussion of these

differences). The results of our synthetic studies of these molecules are described in Chapters 2

and 3 of this dissertation.

1.2 Biological Significance of Plant-Based Polyphenols and General Mechanism of Action

Plants often produce polyphenols to provide resistance against microbial pathogens and

animal herbivores, such as insects, as well as to protect from damage from solar radiation.3 As 6

alluded to in Section 1.1, caffeic acid-derived polyphenols possess a variety of biological

activities. The main source of such activity is the ability of polyphenols to act as antioxidants.

Indeed, the phenol functional group makes these compounds excellent at scavenging free

radicals and reactive oxygen species (ROS) that are formed under conditions of oxidative stress.

Phenols generally react with ROS either by hydrogen atom transfer (HAT) to generate a

phenoxy radical or (15) via a single electron transfer (SET) mechanism to generate a stabilized phenol radical cation (17, Scheme 1).24 The operative pathway is largely determined by the bond

dissociation energy (BDE) of the ArO–H bond compared to the ionization potential of the

phenol. In the case of caffeic acid metabolites, the HAT mechanism is most common because the

presence of an ortho-hydroxy group on the phenolic ring lowers the BDE of the O–H bond by

stabilizing the phenoxy radical through hydrogen bonding. Polyphenols that lack such an

arrangement of hydroxy groups, such as the trihydroxystilbene resveratrol (18), are believed to

follow the SET pathway.25

OH ROS O HO OH

HAT H R OH R O 14 15

OH ROS OH SET R R OH 16 17 18: resveratrol

Scheme 1. Two mechanisms for ROS scavenging by polyphenols

Perhaps paradoxically, the same catechol functional group that imbues these metabolites

with antioxidant activity can also result in pro-oxidative activity under certain conditions. This

can occur when the catechol forms a chelate with iron(III) or copper(II) ions. The initial 7

oxidation of the catechol results in reduction of the metal ion (e.g. CuIIàCuI) and formation of a

26 –• semiquinone radical (21, Scheme 2). This species can reduce O2 to the ROS O2 , which, in

I I turn, is further reduced by the newly generated Cu to H2O2. Further action by Cu can convert

• H2O2 into the damaging HO radical via a Fenton-type reaction. In addition, the o-quinone (22a) generated from the O2 reduction can cause further damage via covalent modification of DNA.

intramolecular OH CuII O SET O -H+ O CuII CuI + R OH -H R O R O R O H H 14 19 20 21

O2 CuII CuI CuII CuI

- • HO + HO H2O2 O2 2 H+

O O

R O R O 22a 22b

Scheme 2. Proposed mechanism of pro-oxidative action of catechols

Although this pro-oxidant activity would seem to preclude the use of catechol-based molecules in medicine, it is generally hoped that careful tuning of the structure could furnish a drug that, present in low concentration in normal cells, would have a protective effect. On the other hand, in cancer cells, where conditions of oxidative stress are more prevalent, a higher concentration of polyphenol would work in tandem with metal ions present to generate more

ROS, resulting in cell death.27

In spite of notable biological activities, polyphenols have seen little use as drug

molecules, and with reason. A common critique of polyphenols is that they often engage in non-

specific interactions with proteins.3 In addition, the phenolic OH groups can readily be 8

glucoronidated in the liver,28 leading to poor pharmacokinetics. Nevertheless, plant-based

polyphenols have served as useful probes in the study of important biological pathways. For

example, resveratrol and other flavonoids have been utilized to better understand the mechanism

by which the sirtuin family of enzymes might potentially influence longevity.29,30

1.3 Biosynthesis of Caffeic Acid and Some of Its Derivatives

The same phenolic function that is the source of much of the biological activity of caffeic

acid and its metabolites is also exploited by plants as a biosynthetic handle to generate diverse

structures through a variety of pathways. Indeed, phenols can be easily oxidized to form a

reactive phenoxy radical, which can combine with other phenoxy radicals to forge new carbon-

carbon or carbon-oxygen bonds. In this section, the biosyntheses of several caffeic acid

metabolites are described, and the implications of these pathways for the biosynthesis of the

yunnaneic acids will be discussed.

1.3.1 Caffeic acid (1)

Caffeic acid is biosynthesized from phenylalanine (24), which, like other aromatic acids, arises

via the shikimic acid pathway.16 The first step from phenylalanine involves the elimination of

ammonia, mediated by the enzyme L-phenylalanine ammonia lyase (PAL), to give cinnamic acid

(25, Scheme 3). The para-hydroxy group is next installed via the formation of an arene oxide

(26) by a cytochrome P450 enzyme. This intermediate then undergoes a 1,2-hydride shift with

epoxide opening (called an NIH shift31) to give phenol 27 following tautomerization. This phenol, named p-coumaric acid, can undergo a second hydroxylation to give caffeic acid. This 9 second oxidation does not involve an NIH shift, but the intermediacy of an arene oxide has been suggested.

CO H CO H CO2H 2 2 PAL NH2

HO OH OH

23: shikimic acid 24: phenylalanine 25: cinnamic acid

[O]

CO2H CO2H CO2H

[O] NIH shift

HO O OH OH H 1: caffeic acid 27: p-coumaric acid 26

Scheme 3. Biosynthesis of caffeic acid

1.3.2 Rosmarinic acid (2)

The biosynthetic pathway leading to rosmarinic acid has been well studied.4 These studies have shown that the two halves of rosmarinic acid arise from different amino acids

(Scheme 4). Namely, the caffeic acid portion comes from phenylalanine and the dihydroxyphenyllactic acid portion comes from tyrosine (28). The phenylalanine-derived portion arises by the mechanism described in Section 1.3.1: formation of cinnamic acid followed by hydroxylation. The formation of the central ester linkage has been shown to occur before the second oxidation to form the catechol domains. Prior to this ester formation, p-coumaric acid must be activated as a coenzyme A thioester (30). The other half of the molecule arises in the 10 following manner: (1) tyrosine is converted into 4-hydroxyphenylpyruvic acid (29) by the enzyme tyrosine aminotransferase (TAT) working in concert with pyridoxal phosphate; (2) reduction of the ketone in 29 by hydroxyphenylpyruvate reductase (HPPR). The resulting hydroxyphenyllactic acid (31) is then coupled with the activated hydroxycinnamic acid molecule to give ester 32, which is then hydroxylated twice to give rosmarinic acid.

CO2H CO2H NH NH 2 HO 2 24: phenylalanine 28: tyrosine

TAT

CO2H CO2H O HO HO

27: p-coumaric acid 29: 4-hydroxyphenylpyruvic acid

HPPR O

CO2H SCoA OH HO HO 30: p-coumaroyl-CoA 31: 4-hydroxyphenyllactic acid

OH O CO2H O

HO 31

[O] (x 2)

OH O CO2H HO O OH

HO 2: rosmarinic acid

Scheme 4. Biosynthesis of rosmarinic acid 11

1.3.3 Podophyllotoxin (5)

Podophyllotoxin arises from the coupling of two molecules of ferulic acid (33), which plants obtain by methylating the 3-hydroxy group of caffeic acid. Initially, the coupling produces trans-lactone 34, and further oxidations furnish yatein (35, Scheme 5). Yatein can then be oxidized to form a p-quinone methide-like intermediate (36), which is capable of undergoing cyclization. Finally, benzylic oxidation of 37 gives podophyllotoxin.16

OH MeO O OMe O O HO O O O

MeO OMe CO2H MeO OH OMe 33: ferulic acid 34 35: yatein

OH H H H O O O O O O O O O H O H O H O

MeO OMe MeO OMe MeO OMe OMe OMe OMe 5: podophyllotoxin 37 36

Scheme 5. Podophyllotoxin biosynthesis

1.3.4 Lithospermic acid (3)

Lithospermic acid is formally a trimer of caffeic acid and might appear to form from rosmarinic acid reacting with caffeic acid. However, its formation may well involve the dimerization of two molecules of rosmarinic acid followed by cleavage of one of the dihydroxyphenyllactic acid side chains since such a dimer has been isolated (lithospermic acid 12

B).32 However, the mechanism by which such a dimerization occurs has not been elucidated in

detail. Interestingly, reaction of methyl isoferulate (38) or methyl ferulate (44) with a variety of

oxidants gave a mixture of oxidative coupling products (39–43; 45), none of which resemble the connectivity found in lithospermic acid (Scheme 6).33,34 Thus, these results may argue for enzymatic participation in the conversion of rosmarinic acid into lithospermic acid.

Cotelle and Vezin (2003):

CO2Me CO2Me CO2Me OMe OMe HO HO Ag2O + CO2Me O O O OMe OMe OMe

39 CO2Me 40 CO2Me 3% yield 3% yield OH OMe 38: methyl FeCl3 CO2Me isoferulate CO2Me CO2Me CO2Me CO2Me OMe HO HO + + MeO MeO O OH OH OH OMe OMe OMe 41 42 43 CO2Me 6% yield 2% yield 3% yield

Snyder and Kontes (2009):

CO2Me MeO2C CO2Me AgOAc OMe

O OMe OH OH OMe 45 44: methyl ferulate 20% yield only characterizable product

Scheme 6. Attempted biomimetic oxidations

1.3.5 Helicterin natural products (6–9)

The helicterin natural products contain a bicyclo[2.2.2]octene core, and can be envisioned to arise biosynthetically from two molecules of rosmarinic acid in the case of the “monomeric” 13

helisterculin A and helisorin. There are multiple ways to explain the formation of the bicycle,

although direct evidence for one pathway being operative does not yet exist.

One mechanistic possibility (Scheme 7) involves a Diels-Alder reaction between an o-

quinone formed via oxidation of rosmarinic acid (diene) and the unsaturated ester of another

rosmarinic acid molecule (dienophile).19 Such a pathway would explain the endo orientation of the ester group on the bicycle. However, given the lack of control provided by remote stereocenters in Diels-Alder reactions of this type,34,35 it would seem that an enzyme may be

necessary to control the stereochemistry relative to the chiral center on the side chain. However,

there exists only scattered evidence for so-called “Diels-Alderase” enzymes in the literature

despite decades of searching. Indeed, given the close resemblance between most Diels-Alder

adducts and the transition states leading up to them, the prospect of product inhibition makes

catalysis of true, concerted Diels-Alder reactions a difficult feat for an enzyme.36

An alternative pathway that has been proposed involves coupling of two rosmarinic acid-

derived radicals (48a and 48b) followed by an aldol-type condensation between the resulting

enol and p-quinone methide (49, Scheme 7).19 This pathway does not require invocation of a

Diels-Alderase, but it would seem that rapid aromatization of the initial radical coupling product

would preclude the second bond formation unless it were very rapid. Such reservations have

been borne out in laboratory experiments with methyl ferulate and various oxidizing agents (see

Section 1.3.4). 14

CO R CO2R 2 CO2R Diels- [O] Alder RO C RO2C O 2 O

OH O O OH

46 OH OH OH OH 47 CO2Me

R = OH CO2R

CO2R CO2R OH

radical coupling RO2C O

OH O OH O OH 48a 48b OH O 49

Scheme 7. Alternate biosynthetic pathways for helicterin-type natural products proposed by Tezuka et al.

An alternative proposal was recently put forward37 involving a diene containing a cis diol

(50), which could be envisioned to arise from a shikimic acid derivative (Scheme 8). A Diels-

Alder reaction with such a diene could then achieve facial control without the need for an enzyme. Intriguingly, the product of such a Diels-Alder reaction would match the structure for helisterculin B (51).19 However, in laboratory studies with model compounds, such a Diels-Alder

reaction proved impossible to achieve without manipulating the electronics of the diene, and only

then in modest yield under harsh reaction conditions.37 Thus, much uncertainty remains

regarding the biogenetic origin of the bicyclic ring system of these natural products. 15

CO2R

CO2R CO2R CO2H ? RO C OH Diels-Alder 2 + OH HO OH OH OH OH OH OH 23: shikimic acid 50: diene 46: dienophile OH OH 51: helisterculin B

Scheme 8. Alternative Diels-Alder hypothesis proposed by Snyder and Kontes

Helicterins A and B have been proposed to arise from dimerization of an appropriate

hydroxyketone (52, Scheme 9).18 Although certainly plausible, a model version of 52 was not found to dimerize34 in such a way under a wide variety of reaction conditions (see section 1.5;

for more on synthetic studies of this family of compounds, see Section 1.4.3).

CO2R CO2R CO2R

RO2C O RO2C O dimerization O CO2R OH OMe OR HO OH OH OH OH OH 52 7: R = Me helicterin A

8: R = H helicterin B

Scheme 9. Proposed biosynthesis of helicterins A and B

16

1.3.6 Yunnaneic Acids (10–13)

The biosynthesis for yunnaneic acids C and D proposed by Tanaka and Kouno, et al. is

similar to the Diels-Alder pathway advanced for the helicterins as described in Section 1.3.5.21

Thus, yunnaneic acid C would arise from a Diels-Alder cycloaddition between a rosmarinic acid-

derived o-quinone (53) and caffeic acid (Scheme 10). As with the helicterins, it would seem that enzymatic assistance would be needed to control the stereochemistry of the Diels-Alder reaction relative to the distal chiral center on the phenyllactic acid side chain of 53. Furthermore, the orientation of the carboxyl group of the natural product (10) would necessitate an exo selective

Diels-Alder reaction, which has not, to the best of our knowledge, ever been observed among the numerous examples of intermolecular Diels-Alder reactions involving similar dearomatized starting materials.38 In short, such a straightforward cycloaddition almost certainly could not occur spontaneously and would require an enzyme. For yunnaneic acid D, the same isolation team proposed a similar Diels-Alder reaction between a dearomatized tautomer of rosmarinic acid (54) and caffeic acid. In addition to raising the same concerns for the yunnaneic acid C biosynthesis, such a dearomatized intermediate would seem to prefer rearomatization to undergoing a Diels-Alder reaction. Once again, to the best of our knowledge, there is no evidence of such dearomatizations occurring through simple tautomerization of a phenol in biosynthetic pathways. Furthermore, accessing such an intermediate via reduction of the corresponding o-quinone would need to be followed by an exceedingly rapid Diels-Alder reaction that outpaces aromatization to the catechol, which also seems unlikely. 17

a) Diels-Alder hypothesis

OH O RO2C RO2C OH O HO HO

O HO HO O

CO H O O CO2R CO2R 2 CO2H 2 53 10: yunnaneic acid C

OH RO2C RO2C O HO HO OH OH HO HO

CO H O O 2 CO2H CO2R 54 13: yunnaneic acid D b) Phenoxy radical hypothesis

OH OH OH O RO2C OH OH O OH O –2H•

HO OH

CO H O CO R CO H CO R 2 2 2 2 CO2H 56 2 1 48b 55 C-C bond rearomatization formation

RO C CO R 2 2 CO2R OH HO O CO2H OH CO2H HO O OH O O OH CO2H OH OH 57 10: yunnaneic acid C 3: lithospermic acid

Scheme 10. Biosynthetic hypotheses for yunnaneic acids involving both Diels–Alder and radical chemistry

18

A radical pathway similar to that proposed by Tezuka et al. for the biosynthesis of the

helicterin natural products19 could conceivably be applied to the yunnaneic acids (Scheme 10).

As in the case outlined in Section 1.3.5.2, the final bond formation would need to compete with

the rearomatization of the initial radical adduct (56). If the rearomatization occurred first, the

resulting phenolic OH group would likely attack the p-quinone methide within 57 to give a

dihydrobenzofuran. Interestingly, such a reaction would result in lithospermic acid, whose

closely related cousin (lithospermic acid B) was isolated alongside the yunnaneic acids.21 While

it may be tempting to postulate that lithospermic acid and the yunnaneic acids arise from the

same reactive intermediate (56) in nature, it would seem to follow that one or more yunnaneic

acids might be found in any other plant that produces lithospermic acid. Given the wide range of

plant species from which lithospermic acid has been isolated, it seems unlikely that yunnaneic

acids would go unnoticed in so many plant extracts.

Recently, David et al. reported the isolation of rufescenolide (58)39 from the Brazilian shrub Cordia rufescens, which has been used in traditional medicine as an abortifacient and anti- inflammatory agent. This compound features the same bicyclo[2.2.2]octene framework as the yunnaneic acids with one key difference–the presence of a lactone ring. This lactone could arise biosynthetically from a reduced yunnaneic acid derivative via lactonization of the free acid. An alternative possibility could involve a tethering of the dienophile prior to cycloaddition. This tethering might occur via some kind of oxidative dearomatization (Scheme 11). Indeed, such

dearomatized intermediates are known to be involved in arene hydroxylations (see Section

1.3.1).16 If arene oxide 60 could be attacked by the carboxyl group of caffeic acid at the appropriate position, the resulting intermediate (59) would be primed to undergo a spontaneous

Diels-Alder reaction to give the core structure of rufescenolide. In contrast to the biosynthetic 19 hypotheses for the helicterins centered on Diels-Alder chemistry, the laboratory analog of this reaction (i.e. an intramolecular Diels-Alder reaction following tethering via a dearomatization event) proceeds almost instantaneously at room temperature (see Chapter 2).

MeO2C OH HO ? O OH HO H MeO2C O Diels-Alder HO O O HO 58: rufescenolide 59

[O] O

MeO C MeO C 2 61 2 60 HO

CO2H HO 1

Scheme 11. Biosynthetic hypothesis for rufescenolide

If such a hypothesis were true for rufescenolide, it may also be operative in the biogenesis of the yunnaneic acids, in which case hydrolysis of the lactone followed by oxidation state adjustment would give yunnaneic acid C or D. The advantages of this hypothesis are that it explains the stereochemistry observed in the natural products without invoking an elusive “Diels-

Alderase” and that, as described in Chapter 2, analogous, mild laboratory conditions have been demonstrated.

Moreover, one might be able to test this hypothesis through feeding experiments. For instance, Cordia rufescens could be fed a labeled precursor (such as caffeic acid or 20 phenylalanine) in which the carboxylic acid contains two 18O isotopes. If rufescenolide were to arise through an intramolecular Diels-Alder reaction, some of the natural product isolated from the plant fed with the labeled compounds should contain both of the heavy oxygen atoms (58a,

Scheme 12). The molecule arising from an intermolecular cycloaddition (or a corresponding series of addition reactions) followed by lactonization would only contain one labeled oxygen

(58b). Such an experiment with yunnaneic acid biosynthesis might also shed light on the biosynthesis of those molecules, although the results could be harder to interpret due to complicating factors. For instance, the diketone of yunnaneic acid C readily forms a hydrate.

Thus, the diketone form is in equilibrium with the hydrated form such that any 18O incorporated into the ketone could be easily exchanged with unlabeled water.

MeO2C HO

OH HO H IMDA pathway 18O OH 18O OH 58a

and/or NH2

18O 18 18 18 OH O OH MeO2C 1: caffeic acid 24: phenylalanine HO

intermolecular OH HO pathway H

O 18O 58b

Scheme 12. Proposed feeding experiment to elucidate rufescenolide biogenesis

Yunnaneic acid A arises from the formation of a spiroketal linkage between yunnaneic acids C and D. As mentioned in Section 1.3.6.4, 1,2-diketones such as the one found in 21

yunnaneic acid C readily undergo addition of water or alcohols.40 Thus, it would appear that the alcohol of yunnaneic acid D could add into such an electrophilic diketone, which would be followed by spontaneous attack of the oxygen of the resulting hemiketal onto the ketone of the yunnaneic acid D fragment (Scheme 13). The factors that would control the regio- and

stereochemistry of such a ketalization are unclear.

Yunnaneic acid B comes from two yunnaneic acid C molecules, and would appear to

come from a similar ketalization reaction. In this case, the diketone would be attacked by a

hydrated diketone. In both cases, it seems reasonable that these dimerizations could occur

spontaneously without the mediation of enzymes, provided that the structure of the natural

products could guide the regio- and stereochemistry of these processes.

RO C RO C HO 2 2 OH HO OH RO C RO C HO 2 2 HO O HO HO O HO O O R OH CO2H CO2H R O O CO2H CO2H 10 10a: R = OH 62 10b: R = H

RO C RO C HO 2 2 OH HO O HO O O R CO2H OH CO2H

11: R = H, yunnaneic acid A 12: R = OH, yunnaneic acid B

Scheme 13. Dimerization accounting for biogenesis of yunnaneic acids A and B

22

1.4 Synthetic Studies of Caffeic Acid Metabolites–Selected Examples

Many of the natural products described in Section 1.1 have attracted the interest of

synthetic organic chemists because of their biological activities and unique structural features.

Furthermore, as illustrated in Scheme 6, biomimetic experiments in which nature’s starting

material is treated with a simple oxidant often result in complex mixtures of products that may

not even contain the desired natural product. Thus, more sophisticated synthetic approaches are

required. In certain cases, these natural products have served as proving grounds for novel

synthetic methodologies, demonstrating their usefulness in constructing complex targets. Several

syntheses of these natural products are discussed briefly in this Section.

1.4.1 Podophyllotoxin (5)

The development of the clinical antitumor agents etoposide and teniposide,17 both based on

podophyllotoxin, spurred a great deal of synthetic interest in the parent natural product in hopes

of developing other, more active, analogs that might not be so easily accessible semi-

synthetically from the isolated natural product itself.41 Two of the more concise syntheses of this natural product are summarized in Scheme 14. Bush and Jones42 completed an asymmetric synthesis of (–)-podophyllotoxin in eight steps and 15% overall yield. Their synthesis featured a diastereoselective Diels–Alder reaction between pyrone 63 and dienophile 64, in which the menthyl group was used as a chiral auxiliary. In addition to controlling the stereochemistry of the

Diels–Alder reaction, the steric bulk of the menthyl group was exploited to control the facial selectivity of a later hydrogenation to secure the trans-fused lactone. 23

Bush and Jones (1993):

O iPr HO C O 2 H OMenthyl Me O O 1) MeCN, 50 °C O O O + O O 2) AcOH, 50 °C O 65 Ar MeO OMe O 64 OMe H2, Pd/C, 63 HO EtOAc H O HO C O 2 H OMenthyl O O H O O O H Ar O MeO OMe 66 OMe 5: podophyllotoxin Bhat et al. (1996): Ph O S n-BuLi, THF, O -78 °C, O O Ph O O O 68 HO 1) TFA S H O 5 O then ArCHO O 2) HgO, 67 BF3•OEt2 MeO OMe 69 OMe

Scheme 14. Selected total syntheses of podophyllotoxin

A few years later, Bhat et al. reported a seven-step asymmetric total synthesis.43 These

workers used a chiral sulfoxide (67, installed using Sharpless epoxidation-like conditions

developed by Kagan et al.44) to control the stereochemistry of a one-pot conjugate addition/aldol reaction. Podophyllotoxin was obtained after acid-mediated ring closure and a Pummerer rearrangement.

24

1.4.2 Lithospermic Acid (3)

In a similar vein, the anti-HIV activity of lithospermic acid has prompted much synthetic

interest.45-47 Among the flurry of synthetic work that has been reported in the past decade, the asymmetric syntheses reported by Ellman, Bergman, et al. and Yu and Wang stand out for their efficiency and the application of different powerful C–H functionalization methods developed by the respective groups.48,49

The Ellman/Bergman synthesis of (+)-lithospermic acid, reported in 2005, was achieved

in ten steps and 6% overall yield. This work featured a diastereoselective Rh-catalyzed C-H

alkylation reaction developed by the Ellman group50 to form the dihydrobenzofuran core, in which the stereochemistry was controlled by a chiral imine auxiliary (Scheme 15). Following cleavage of the auxiliary, dihydrobenzofuran 71 was obtained in 88% yield and 73% ee (56% yield, 99% ee after recrystallization).

Ellman, Bergman, et al. (2005):

CO2H CHO CO2Me O O 1) [RhCl(coe)2]2 (10 mol %) HO N FcPCy (30 mol %) 2 OMe OH MeO2C PhMe, 75 °C O CO2H OH 2) HCl, H2O OMe OMe OMe OH O 71 88%, 73% ee O OMe OMe (56%, 99% ee 70 after OH recrystallization) 3: lithospermic acid

Scheme 15. Total synthesis of lithospermic acid by Ellman, Bergman, et al.

Six years later, Yu and Wang completed a 12 step synthesis of (+)-lithospermic acid in 11

% overall yield. The synthesis included two steps that relied on C–H functionalization chemistry, nicely demonstrating the synthetic utility of these modern methods. As shown in Scheme 16, the dihydrobenzofuran was constructed in a diastereoselective fashion via a chiral auxiliary-

51 controlled insertion of a Rh carbenoid (generated using Davies’s Rh2(S-DOSP)4 catalyst). The 25 chiral side chain was then installed using the group’s Pd-catalyzed C–H olefination chemistry,52 in which the functionalization is accelerated by an Ac-Ile-OH ligand.

Yu and Wang (2011): O C–H insertion O O CO H N 2 1) Rh2(S-DOSP)4 (0.5 mol %) OMe N2 O OMe 2) Ba(OH)2 O OMe OMe OMe 73 OMe 8.5 : 1 dr, 73 % 72

CO2Me Pd(OAc)2, KHCO3, O O O2, Ac-Ile-OH, MeO tAmOH, 85 °C OMe C–H olefination 74

CO Me CO2H 2 O O O O HO MeO OH OMe 1) Me SnOH CO H OMe CO2H OH 3 2

OH 2) TMSI-quinoline OMe O O OH OMe 75 93 % 3: lithospermic acid

Scheme 16. Total synthesis of lithospermic acid by Yu and Wang

26

1.4.3 Helicterin Natural Products

In 2009, Snyder and Kontes reported the total syntheses of helicterin B (8), helisorin (6),

and helisterculin A (9).34 The bicyclic core was assembled via a retro-Diels-Alder/Diels-Alder

cascade in which heating of compound 76 initiated a retro-Diels-Alder reaction to give masked

o-benzoquinone (MOB) 77, which then underwent a Diels-Alder reaction to give the desired

bicycle (Scheme 17). This process is believed to occur under thermodynamic control. The

synthesis also featured a skillful manipulation of hydroxyketone stereochemistry to obtain

hydroxyketal 81, which then underwent dimerization upon exposure to BF3•OEt2 to give a C2- symmetric dimer. Synthetic helicterin B was obtained following removal of the p- trifluoromethylbenzyl ethers (a newly-introduced protecting group that resists cleavage by

BF3•OEt2 but is cleaved by slightly stronger Lewis acids). The synthesis of helisorin featured an intramolecular Friedel–Crafts-type reaction of 79 to give the unique tricyclic structure of that

natural product. 27

CO2R CO R CO2R 2 CO2R

MeO OMe 1) NaBH OMe Diels–Alder 4 H O 220 °C OMe 2) HCl RO C RO2C 2 OMe retro CO2R OH O OMe OMe Diels– O O Alder OMe 78 O CO2R OTfBn OTfBn OTfBn 76 77 OTfBn OTfBn (6.7 equiv) OTfBn 79 80

1) Me4NBH(OAc)3 2) TBSOTf, Et3N 3) Dess–Martin OH 4) HCl, MeOH, OH HC(OMe) OH 3 HO

CO2R O O TfBnO OTfBn O O 1) BF •OEt MeO2C 3 2 OMe CO2Me 2) BBr RO2C 3 OH R = MeO2C O O O CO2Me OMe O O O MeO2C OMe OH OTfBn HO HO OTfBn 81 OH OH OH OH OH OH 8: helicterin B OH OH

CO2R O O

OMe MeO C RO2C 1) BF3•OEt2 2 2) BBr OMe 3 MeO2C O O O O OH OTfBn OTfBn HO 79 OH HO OH 6: helisorin

Scheme 17. Syntheses of helicterin B and helisorin by Snyder and Kontes (TfBn = p-trifluoro- methylbenzyl)

28

CO2Me CO2Me CO2Me

MeO C O 2 NaH MeO C O CO Me OH 2 2 O OH OH H OMe OMe MeO OMe OMe OMe 82 83

Scheme 18. Unusual dimerization observed during studies towards helicterins A and B

1.5 Yunnaneic Acids: A Unique Set of Synthetic Challenges

The yunnaneic acids first caught our attention when, in the pursuit of the total synthesis of helicterin B, my former colleague, Dr. Ferenc Kontes, observed that, upon exposure to various acids and bases, model hydroxyketone 82 dimerized to form an unsymmetrical dimer (83) whose spiroketal moiety linking the two monomers resembles those of yunnaneic acids A and B

(Scheme 18).34 Despite the superficial resemblance to this dimer and the helicterin family of

natural products more broadly, the structures of the yunnaneic acids present significant synthetic

challenges that would require the deployment of methods quite different from those employed in

the helicterin synthesis. First, whereas dimer 83 is the result of homodimerization of a single hydroxyketone, yunnaneic acid A requires the selective union of two distinct monomeric fragments (i.e. yunnaneic acids C and D; Scheme 19A). Second, yunnaneic acid B presents a similar challenge in that its synthesis would likely require the nucleophilic attack of a hydrated diketone (85b) on a non-hydrated diketone (i.e. yunnaneic acid C). Thus, the union of two subtly different monomers would once again be required. The third key difference between the yunnaneic acids and the helicterin family of natural products lies in the bicyclo[2.2.2]octene 29 core. Although the structures of both families appear to result from the Diels-Alder reaction between two derivatives of caffeic acid, the substituents decorating the bicycle differ in both regio- and stereochemistry. Whereas the bicycles typical of the helicterins are formally endo

Diels–Alder products, the yunnaneic acids are formally exo Diels–Alder adducts. Thus, a successful synthesis of these molecules would require circumvention of the inherent reactivity of an appropriate Diels–Alder precursor (Scheme 19B).

CO2R A) RO2C RO2C RO2C

Ar ? O O Ar O Ar HO Ar O O CO2H CO2H R O R CO2H OH CO2H 84 85a: R = H 86a: R = H 85b: R = OH 86b: R = OH Is control achievable? B) typical reactivity:

OH EWG OR 1 1 EWG2 EWG + [O] 2 O OMe MeOH R2 R OR EWG1 2 1 87 88 89 desired reactivity: endo OH EWG1 OR1 EWG 2 [O] R + 2 O OMe MeOH R2 EWG OR EWG1 2 1 90 87 88 exo

Scheme 19. A) Challenge presented by anticipated dimerization reaction; B) challenge presented by yunnaneic acid bicyclic architecture

The synthetic strategies developed towards meeting these challenges will be described in the remainder of this dissertation. These investigations ultimately resulted in the development of 30 a cascade reaction enabling a rapid construction of the yunnaneic acids’ bicyclic core, paving the way towards the first total syntheses of yunnaneic acids C and D, the subject of Chapter 2. The dimerization studies described in Chapter 3 shed much light on the reactivity of the monomeric yunnaneic acids. These studies culminated in the discovery of reaction conditions that facilitated a selective heterodimerization with no competing homodimerization.

1.6 References

(1) Boerjan, W.; Ralph, J.; Baucher, M. Ann. Rev. Plant Biol. 2003, 54, 519.

(2) Manach, C.; Williamson, G.; Morand, C.; Scalbert, A.; Remesy, C. Am. J. Clin. Nutr. 2005, 81, 2305.

(3) Quideau, S.; Deffieux, D.; Douat-Casassus, C.; Pouysegu, L. Angew. Chem. Int. Ed. 2011, 50, 586.

(4) Petersen, M.; Simmonds, M. S. J. Phytochemistry 2003, 62, 121.

(5) Scarpati, M. L.; Oriente, G. Ric. Sci. 1958, 28, 2329.

(6) Kelley, C. J.; Mahajan, J. R.; Brooks, L. C.; Neubert, L. A.; Breneman, W. R.; Carmack, M. J. Org. Chem. 1975, 40, 1804.

(7) Kelley, C. J.; Harruff, R. C.; Carmack, M. J. Org. Chem. 1976, 41, 449.

(8) Lu, Y. R.; Foo, L. Y. Phytochemistry 2002, 59, 117.

(9) Jiang, R. W.; Lau, K. M.; Hon, P. M.; Mak, T. C. W.; Woo, K. S.; Fung, K. P. Curr. Med. Chem. 2005, 12, 237.

(10) Zhou, L. M.; Zuo, Z.; Chow, M. S. S. J. Clin. Pharmacol. 2005, 45, 1345.

(11) Abd-Elazem, I. S.; Chen, H. S.; Bates, R. B.; Huang, R. C. C. Antivir. Res. 2002, 55, 91. 31

(12) Agata, I.; Hatano, T.; Nishibe, S.; Okuda, T. Chem. Pharm. Bull. 1988, 36, 3223.

(13) Agata, I.; Hatano, T.; Nishibe, S.; Okuda, T. Phytochemistry 1989, 28, 2447.

(14) Kashiwada, Y.; Nishizawa, M.; Yamagishi, T.; Tanaka, T.; Nonaka, G.; Cosentino, L. M.; Snider, J. V.; Lee, K. H. J. Nat. Prod. 1995, 58, 392.

(15) Kashiwada, Y.; Bastow, K. F.; Lee, K. H. Bioorg. Med. Chem. Lett. 1995, 5, 905.

(16) Herbert, R. B. The Biosynthesis of Secondary Metabolites; 2nd ed.; Chapman and Hall: London; New York, 1989. p. 96.

(17) Keller-Juslen, C.; Kuhn, M.; Wartburg, A. V.; Stahelin, H. J. Med. Chem. 1971, 14, 936.

(18) Tezuka, Y.; Terazono, M.; Kusumoto, T. I.; Hatanaka, Y.; Kadota, S.; Hattori, M.; Namba, T.; Kikuchi, T.; Tanaka, K.; Supriyatna, S. Helv. Chim. Acta 2000, 83, 2908.

(19) Tezuka, Y.; Terazono, M.; Kusumoto, T. I.; Kawashima, Y.; Hatanaka, Y.; Kadota, S.; Hattori, M.; Namba, T.; Kikuchi, T.; Tanaka, K.; Supriyatna, S. Helv. Chim. Acta 1999, 82, 408.

(20) Otake, T.; Mori, H.; Morimoto, M.; Ueba, N.; Sutardjo, S.; Kusumoto, I. T.; Hattori, M.; Namba, T. Phytother. Res. 1995, 9, 6.

(21) Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Young, T. J. J. Nat. Prod. 1996, 59, 843.

(22) Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Yang, C. R. Chem. Pharm. Bull. 1997, 45, 1596.

(23) Yokozawa, T.; Chen, C. P.; Dong, E.; Tanaka, T.; Nonaka, G. I.; Nishioka, I. Biochem. Pharmacol. 1998, 56, 213.

(24) Wright, J. S.; Johnson, E. R.; DiLabio, G. A. J. Am. Chem. Soc. 2001, 123, 1173.

(25) Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. J. Phys. Chem. A 2004, 108, 4916. 32

(26) Mira, L.; Fernandez, M. T.; Santos, M.; Rocha, R.; Florencio, M. H.; Jennings, K. R. Free Radical Res. 2002, 36, 1199.

(27) Fan, G. J.; Jin, X. L.; Qian, Y. P.; Wang, Q.; Yang, R. T.; Dai, F.; Tang, J. J.; Shang, Y. J.; Cheng, L. X.; Yang, J.; Zhou, B. Chem. Eur. J. 2009, 15, 12889.

(28) Shangari, N.; Chan, T. S.; O'Brien, P. J. Methods in Enzymology. Phase II: Conjugation Enzymes and Transport Systems 2005, 400, 342.

(29) Howitz, K. T.; Bitterman, K. J.; Cohen, H. Y.; Lamming, D. W.; Lavu, S.; Wood, J. G.; Zipkin, R. E.; Chung, P.; Kisielewski, A.; Zhang, L. L.; Scherer, B.; Sinclair, D. A. Nature 2003, 425, 191.

(30) Quideau, S. Chem. Bio. Chem. 2004, 5, 427.

(31) Guroff, G.; Daly, J. W.; Jerina, D. M.; Renson, J.; Witkop, B.; Udenfriend, S. Science 1967, 157, 1524.

(32) Tanaka, T.; Morimoto, S.; Nonaka, G. I.; Nishioka, I.; Yokozawa, T.; Chung, H. Y.; Oura, H. Chem. Pharm. Bull. 1989, 37, 340.

(33) Cotelle, P.; Vezin, H. Tetrahedron Lett. 2003, 44, 3289.

(34) Snyder, S. A.; Kontes, F. J. Am. Chem. Soc. 2009, 131, 1745.

(35) Gagnepain, J.; Castet, F.; Quideau, S. Angew. Chem. Int. Ed. 2007, 46, 1533.

(36) Stocking, E. M.; Williams, R. M. Angew. Chem. Int. Ed. 2003, 42, 3078.

(37) Snyder, S. A.; Kontes, F. Isr. J. Chem. 2011, 51, 378.

(38) Liao, C. C.; Peddinti, R. K. Acc. Chem. Res. 2002, 35, 856.

(39) do Vale, A. E.; David, J. M.; dos Santos, E. O.; David, J. P.; de Silva, L. C. R. C.; Bahia, M. V.; Brandao, H. N. Phytochemistry 2012, 76, 158. 33

(40) Scharf, H.-D.; Kuesters, W. Chem. Ber. 1972, 105, 564.

(41) Review: Ward, R. S. Synthesis 1992, 719.

(42) Bush, E. J.; Jones, D. W. J. Chem. Soc. Chem. Commun. 1993, 1200.

(43) Hadimani, S. B.; Tanpure, R. P.; Bhat, S. V. Tetrahedron Lett. 1996, 37, 4791.

(44) Pitchen, P.; Dunach, E.; Deshmukh, M. N.; Kagan, H. B. J. Am. Chem. Soc. 1984, 106, 8188.

(45) Fischer, J.; Savage, G. P.; Coster, M. J. Org. Lett. 2011, 13, 3376.

(46) Ghosh, A. K.; Cheng, X.; Zhou, B. Org. Lett. 2012, 14, 5046.

(47) Varadaraju, T. G.; Hwu, J. R. Org. Biomol. Chem. 2012, 10, 5456.

(48) O'Malley, S. J.; Tan, K. L.; Watzke, A.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc. 2005, 127, 13496.

(49) Wang, D. H.; Yu, J.-Q. J. Am. Chem. Soc. 2011, 133, 5767.

(50) Colby, D. A.; Bergman, R. G.; Ellman, J. A. Chem. Rev. 2010, 110, 624.

(51) Davies, H. M. L.; Hansen, T. J. Am. Chem. Soc. 1997, 119, 9075.

(52) Wang, D. H.; Engle, K. M.; Shi, B. F.; Yu, J.-Q. Science 2010, 327, 315, and references cited therein.

34

CHAPTER 2

Synthetic Studies of Yunnaneic Acids C and D

35

2.1 Early Synthetic Explorations

HO HO OH OH

CO2H CO2H

O O O O

HO HO OH HO O HO H O O CO2H CO2H

1: yunnaneic acid C 2: yunnaneic acid D

Figure 1. Structures of yunnaneic acids C and D

The first problem we hoped to tackle in the synthesis of yunnaneic acids C and D1 (1, 2, Figure

1) was the construction of the bicyclic core with the proper regio- and stereochemistry. Given the success of the retro-Diels–Alder/Diels–Alder strategy in the synthesis of the helicterins2 (Scheme

1A), we wondered if a modified version of that strategy might provide access to the yunnaneic acid architecture. In particular, we were curious to see if changing the regiochemistry of the starting material could influence the regioselectivity of a subsequent Diels–Alder reaction.

Specifically, we hoped that the MOB ketal derived from 6 would fashion the requisite bicycle from a Diels–Alder reaction with the desired stereo- and regiochemistry (Scheme 1B). 36

A) Snyder and Kontes (2009): CO2Me

OH OMe OMe MeO C 1) PhI(OAc)2, MeOH 2

2) 220 °C OMe O CO2Me CO2Me 3: methyl MeO 4 OMe ferulate OMe OMe 5 B) Similar strategy for the yunnaneic acids: OMe MeO2C OH MeO 1) PhI(OAc) , MeOH 2 OMe 2) 220 °C MeO OMe CO2Me CO Me O CO2Me 2 MeO 4 7 6: methyl OMe isoferulate ?

Scheme 1. Initial retro-Diels–Alder/Diels–Alder strategy inspired by the total synthesis of the helicterins

To probe this question experimentally, methyl isoferulate (6) was first reacted with

PhI(OAc)2 in MeOH to give the expected [4+2] dimerization product 8 (Scheme 2).

Unfortunately, heating of this dimer in the presence of excess dienophile 4 did not provide any

characterizable products. Apparently, altering the substitution pattern on the diene of such a

Diels–Alder reaction rendered it unreactive towards the desired dienophile. Altering the

electronics of the dienophile by reducing the ester of 4 to the alcohol, thus raising the energy of

the highest occupied molecular orbital (HOMO), also failed to result in any Diels–Alder

reaction.*

* For further discussion of this reaction in the context of frontier molecular orbital (FMO) theory, see Section 2.2 37

CO2Me

OMe MeO2C OH MeO OMe mesitylene MeO a) PhI(OAc)2 O O OMe 220 °C OMe MeOH MeO OMe OMe OMe MeO C 8 MeO O CO2Me 2 CO2Me (regiochemistry unknown) 6 4 7

CO2Me

Reagents and conditions: a) PhI(OAc)2 (1.5 equiv), MeOH/CH2Cl2 (5:1), 0!23 °C, 15 min, 97%

Scheme 2. Attempted retro-Diels–Alder/Diels–Alder approach to yunnaneic acid core

In order to work around the apparent unreactive nature of the MOB ketal derived from 6,

we thought that perhaps the proper regio- and stereochemistry of the bicyclo[2.2.2]octene would

be better enforced via an intramolecular Diels–Alder reaction. Thinking along these lines

retrosynthetically (Scheme 3), the monomeric yunnaneic acids could arise from a common

intermediate 9. This common intermediate could come from an intramolecular Diels–Alder

reaction that would follow an oxidative dearomatization of an appropriate phenol 12 in which,

rather than a methoxy group, a dienophile-bearing carboxylic acid (11) would be attached. The

diene and dienophile would then be poised to undergo the desired Diels–Alder reaction (10) in

such a way that enforces the proper orientation of the cycloaddition partners. Similar reactions

with hypervalent iodine as the oxidant have appeared in the literature, albeit with simple acids

such as acrylic acid that can be used as a co-solvent.3 However, such dearomatizations with

cinnamic acid derivatives have not been reported. Unfortunately, attempts to adapt such

conditions to our system failed, with intractable mixtures of products being the only result, even

when a large excess (up to 20 equivalents) of the dienophilic carboxylic acid and/or more

reactive hypervalent iodine sources such as PIFA [PhI(OCOCF3)2] were used. A possible reason 38

for these failures was the insolubility of the carboxylic acids in solvents commonly used for such

transformations.

RO2C RO2C R'O R'O

O R'O R'O O OMe O CO2H O O 1: yunnaneic acid C 9

Diels-Alder

OR' OH OR' OMe RO2C dearomat- ization R'O

R'O O OMe CO H CO R 2 2 O O 11 12 10

Scheme 3. Retrosynthetic analysis of yunnaneic acid C

To work around the apparent inability of the phenol (the latent diene) and the dienophile

to couple under dearomatization conditions, we pursued a strategy of tethering the dienophile to the phenol prior to the dearomatization event. Thus, rather than requiring a poorly soluble carboxylic acid to attack, the nucleophile would be the solvent itself (i.e. MeOH), as in the case of the [4+2] homodimerization (Scheme 2). To test this strategy, the p-hydroxy group of methyl caffeate (13) was selectively blocked by reacting it with AllylBr and NaH (Scheme 4). Under carefully optimized conditions (two equivalents of AllylBr added to a solution of methyl caffeate and NaH at –50 °C followed by gradual warming to room temperature), the desired allyl ether could be obtained in 57% yield along with some diallylated material (10-20%). This selectively 39

protected material was then coupled with carboxylic acid 14 using DCC4 followed by Pd-

mediated removal of the allyl protecting group5 to give phenol 15.

With the desired phenol in hand, we then turned to the task of oxidative dearomatization,

which would hopefully be followed by a spontaneous Diels–Alder reaction. Unfortunately, the

only product that could be recovered from the reaction of 15 with PhI(OAc)2 in MeOH was the

cleaved dienophile (4) resulting from methanolysis of the ester linkage. Presumably, oxidation of

the phenol resulted in significant activation of the ester linkage (potentially via the intermediacy

of 17). The carbonyl group, having enhanced electrophilicity, was then attacked by MeOH in

preference to the dearomatized ring. Such an attack would result in the methyl ester observed and

an o-quinone (18), which could then decompose.

OH OMe OH MeO2C OH O PhI(OAc) MeO a) NaH, AllylBr OMe 2 O b) DCC, 14 MeOH MeO O c) Pd(PPh3)4, 15 OMe AcOH O CO2Me CO2Me O 13 MeO CO2H 16 MeOH " OMe " OMe MeO O O 14 OMe O O OMe O 17

CO2Me CO2Me MeO O 18 4 Reagents and conditions: a) NaH (1.05 equiv), KI (0.1 equiv), AllylBr (2.0 equiv), DMF, -50!25 °C, 16 h, 57 %; b) DCC (1.5 equiv), 4-DMAP (0.1 equiv), 14 (1.5 equiv), CH2Cl2, 24 h; c) Pd(PPh3)4 (0.5 equiv), AcOH, 23 °C, 36 h, 67 % (2 steps)

Scheme 4. Attempted intramolecular Diels-Alder with ester-linked dienophile

In light of this failure, we sought a substrate with a more robust tether that would not

prove so fragile upon oxidation of the phenol. Thus, we turned our attention to the synthesis of a substrate in which the phenol and dienophile were linked by an ether. The desired ether (20, 40

Scheme 5) was constructed via a Mitsunobu reaction of phenol 19 with allylic alcohol 23. In addition to the desired ether, several side products also resulted, which made purification difficult. Chief among these is a compound that has been tentatively assigned as the SN2’ addition product, though rigorous characterization was not possible due to difficulties with purification. In any event, the allyl ether could then be deprotected via palladium-catalyzed

6 deallylation using Et2NH as an allyl scavenger to give the desired Diels–Alder precursor 20 in

26% yield over two steps. Although the yield for these two steps was disappointing, extensive

screening of reaction conditions for both steps failed to result in any improvement. Indeed, it is

perhaps unsurprising that selective deallylation of a substrate containing two allylic ethers

proved troublesome. However, no other protecting group investigated could be both selectively

installed at the 4-position of the catechol and removed under sufficiently mild conditions.

Fortunately, despite these issues, ample amounts of 20 (i.e. 1-3 grams) could be obtained from

this sequence to investigate the anticipated dearomatization/Diels–Alder cascade.

OAllyl OMe OH MeO2C OH a) DIAD, 21 c) PhI(OAc)2, O ROH MeO PPh3 OMe

b) Pd(OAc)2, Ph3P, MeO O Et2NH OR O CO2Me CO2Me 20 19 21: R = H MeO CO Me d) DIBAL-H MeO 22: R = Me 2 OH

MeO MeO 4 23

Reagents and conditions: a) DIAD (1.5 equiv), Ph3P (1.5 equiv), 23 (1.2 equiv), THF, 0!25 °C, 24 h; b) Pd(OAc)2 (0.2 equiv), Ph3P (0.4 equiv), Et2NH/THF/H2O (3:6:1), 25 °C, 1 h, 26% (2 steps); c) PhI(OAc)2 (1.1 equiv), ROH, 1,4-dioxane, 75 °C, 5 min, 67% (for R = H), 74% (for R = Me); d) DIBAL-H (3.2 equiv), THF, -78!25 °C, 18 h, 99%

Scheme 5. Dearomatization/Diels–Alder with ether-linked dienophile

41

To our delight, when subjected to PhI(OAc)2 in the presence of MeOH or H2O at 75 °C,

20 underwent a smooth dearomatization and Diels–Alder reaction in the same pot, furnishing

tricycle 21 or 22 in reasonable yield (65-75%), especially considering that the reaction forms

three new bonds. The above synthetic sequence could provide sufficient amounts (i.e. hundreds of milligrams) of the Diels–Alder product to pursue its elaboration into model versions of yunnaneic acids C and D.

Next, to open the hemiketal or ketal of 21 or 22, we hoped to establish an equilibrium between the open and closed forms of the ketal (21 and 24), followed by interception of the free alcohol with an oxidizing agent (Scheme 6A). To our dismay, the ketal or hemiketal of both 21 and 22 proved very unreactive. Our attempts at rupturing this functional group with acids and/or oxidants were met with failure (Scheme 6B), with either recovered starting material or decomposition being the most common results. Even methods that had provided success in cases of similarly “locked” hemiketals for others (such as excess Dess–Martin periodinane in refluxing benzene7) failed in our hands. After this string of failures, we concluded that too high a kinetic barrier existed to establish an equilibrium between the two ring-chain tautomers. The reason for such a high barrier may have been the instability of the 1,2-diketone present in the ring-opened species. 42

A) MeO2C MeO2C MeO2C MeO MeO MeO [O] O O MeO O MeO MeO OH O O O HO O 21 24 25

H2O

MeO2C MeO2C MeO2C MeO MeO MeO [O] or O O MeO O MeO MeO OH OH O O CO2H O 27 O 28 HO 26

B) MeO2C MeO2C MeO2C MeO a) Me NBH(OAc) 4 3 Ar Ar MeO O OH OH OH OH O O O O B(OAc)2 21 29 30 H numerous methods NaBH4 MeO2C MeO2C MeO C 2 MeO MeO Ar MeO OH MeO O OH O B(OAc) O 2 CO H O HO HO H 2 27 32 31

C) MeO2C MeO2C MeO2C MeO MeO MeO

b) NaBH4 MeO O MeO OAc MeO OAc OMe OMe c) Ac2O, Et3N O O O CO2H 22 33 34

Reagents and conditions: a) Me4NBH(OAc)3 (5.0 equiv), MeCN/AcOH (2:1), 0!25 °C, 2 h, 77%; b) NaBH4 (2.0 equiv), CH2Cl2/MeOH (1:1), 0 °C, 45 min; c) Ac2O (1.1 equiv), Et3N (1.5 equiv), 4-DMAP (cat.), CH2Cl2, 25 °C, 2 h, 81% (2 steps)

Scheme 6. A) Anticipated ring-chain tautomerism of Diels–Alder product 21; B) Successful opening of hemiketal of 21; C) Attempted opening of partially reduced intermediate 33 43

We thought that we might be able to work around this problem if the neighboring ketone

were reduced, which would make opening of the ketal more facile since the resulting

hydroxyketone intermediate might not present the same issues of instability as the corresponding

diketone. Thus, 21 was reduced with excess Me4NBH(OAc)3 in AcOH/MeCN, which resulted in triol 32 in 77% yield (Scheme 6B).8 This product presumably arose from initial reduction of the

ketone from the exo face followed by rupture of the hemiketal of 29 and a second reduction

event, directed by the endo-disposed alcohol. The intramolecular nature of the second reduction

appears to be crucial to trapping the open form of the hemiketal since reaction with NaBH4 only reduced the ketone of 20, leaving the hemiketal intact. Thus, the open form of the hemiketal might be so short-lived that hydride delivery must occur intramolecularly. It is also worth noting that the presence of a neighboring ketone does not appear to be the chief reason for the recalcitrance of the hemiketal since the NaBH4 reduction product (22) proved similarly resistant to opening (Scheme 6C).

With the hemiketal having been successfully pried open, we next turned our attention to adjusting the oxidation state of 32. This task required a carefully orchestrated set of oxidation steps in order to prevent the alcohol from re-closing (Scheme 7). Such a sequence involved: 1) selective oxidation of the primary alcohol under the catalytic action of TEMPO with PhI(OAc)2

as a co-oxidant;9 2) Pinnick oxidation10 of the resulting aldehyde; 3) esterification of the newly

formed carboxylic acid with TMSCHN2; and 4) oxidation of trans diol 35 with Dess–Martin periodinane.11-13 This sequence resulted in a model version of the diketone of yunnaneic acid C

(36). 44

MeO2C MeO2C MeO2C MeO MeO MeO a) TEMPO, PhI(OAc) 2 d) Dess–Martin O MeO OH MeO OH MeO b) Pinnick [O] [O] OH OH c) TMSCHN2 O CO2Me CO2Me HO 32 35 36

Reagents and conditions: a) TEMPO (0.1 equiv), PhI(OAc)2 (1.1 equiv), CH2Cl2, 25 °C, 3 h, 60%; b) NaClO2 (5.0 equiv), NaH2PO4•2H2O (10.0 equiv), 2-methyl-2-butene (10.0 equiv), t-BuOH, H2O, 25 °C, 2 h; c) TMSCHN2 (2.0 equiv), THF/MeOH (10:1), 0 °C, 30 min, 62% (2 steps); d) Dess-Martin periodinane (4.0 equiv), NaHCO3 (7.5 equiv), CH2Cl2, 0!25 °C, 2 h, 35%

Scheme 7. Synthesis of model yunnaneic acid C

Our attention next turned to converting triol 32 into a hydroxyketone corresponding to the framework of yunnaneic acid D. At the time, we were unsure of the stereochemistry of the trans diol. Specifically, we were uncertain whether the exo alcohol was on the same carbon atom as the one in yunnaneic acid D, with the stereochemical assignment up until this point chiefly based on mechanistic reasoning. In order to access a compound with the proper oxidation state, the relatively unhindered exo-disposed hydroxyl group of 35 was selectively protected as a TBS ether and the free endo alcohol was then oxidized using Dess–Martin periodinane (Scheme 8).

Removal of the TBS group of 37 gave a hydroxyketone, a molecule whose regiochemistry was determined by synthesis of the crystalline p-bromobenzoate derivative 38. X-ray crystallographic analysis of 38 showed that, as expected, the regiochemistry of the hydroxyketone did not match that of yunnaneic acid D. 45

Scheme 8. Synthesis of hydroxyketone with undesired regiochemistry

In addition to the lack of obvious inroads to the desired hydroxyketone, the route

described above suffered from several other drawbacks. First, the Mitsunobu reaction and

deallylation sequence was low yielding and the crude Mitsunobu product was difficult to purify

because of the mixture of products formed (in addition to the usual purification difficulties often

associated with this reaction). Second, the diol oxidation to give the diketone was also low-

yielding and was somewhat unreliable. Third, the route required excessive and non-strategic redox manipulations (oxidative dearomatization followed by reduction followed by three oxidations). These drawbacks combined to seriously compromise the efficiency of the synthesis and, from a more practical standpoint, made it difficult to obtain even modest quantities of model compounds.

2.2 Second-generation Synthesis of the Yunnaneic Acid C Model and Successful Approach to the Yunnaneic Acid D Model

In light of these deficiencies, we reconsidered our initial goal of obtaining the desired

Diels–Alder adduct directly from an appropriate carboxylic acid (dienophile) and phenol (latent 46

diene) and began searching for an alternative oxidizing agent that could effect such a transformation. A promising candidate was Pb(OAc)4, which is known to dearomatize

appropriately substituted phenols to MOB ketals14 in a reaction known as the Wessely oxidation.15 Furthermore, it had been shown by multiple investigators that replacement of the

acetate ligands on the PbIV metal center with an α,β−unsaturated carboxylate could result in a

dearomatization event followed by a Diels–Alder reaction to generate a tricyclic product of the

type sought.16,17 To our delight, this protocol could be adapted to the synthesis of the yunnaneic acid bicycle (Scheme 9). In the event, the desired PbIV reagent could be generated by vigorously

stirring a suspension of carboxylic acid 14 with Pb(OAc)4 in CH2Cl2 followed by concentration

and co-evaporation with toluene to remove the AcOH evolved during the ligand exchange.

Subsequent addition of a solution of methyl ferulate (3) to a solution of the Pb(IV) reagent in

1,4-dioxane at 25 °C resulted in the nearly instantaneous formation of the desired product 16

along with a small amount of the α-acetoxylated product 41; the structure of 16 was confirmed

by X-ray crystallography. Interestingly, previous reports of these reactions involved performing

the initial oxidation, solvent removal, dissolution in a higher boiling solvent, and prolonged

heating to effect the cycloaddition, with few exceptions.17,18 In contrast, this reaction occurred spontaneously and rapidly at ambient temperature–neither the dearomatized intermediate 40 nor starting material could be detected by TLC analysis immediately after phenol addition was complete. 47

Scheme 9. Modified Wessely oxidation/Diels–Alder cascade

A plausible reaction mechanism is shown in Scheme 9. Wessely oxidations are known to selectively deliver the carboxylate ligand to the more electron rich ortho position of the phenol, implicating an intramolecular delivery mechanism after the phenol coordinates to the PbIV (39).19

The dearomatized intermediate 40 is then poised to undergo an intramolecular [4+2] reaction.

The notable speed of this reaction, especially in comparison to some similar reactions previously reported, deserves further comment. Dory, Deslongchamps et al. have performed calculations20 showing that the lowest unoccupied molecular orbitals (LUMOs) of dienones such as 43, which are believed to be the FMOs that participate in the cycloadditions, are quite delocalized thanks to participation by the ketal carbon (as represented by the canonical resonance form of 43). These calculations showed that the LUMO of 43 qualitatively resembles the LUMO 48

of o-quinone itself (44), which has a low energy LUMO and is highly reactive (Scheme 10A). †

This information may explain the contrast in reactivity of the dienone intermediate 40 from the

transformation of 3 to 16 with that from the transformation reported by Yates and Auksi, in which a tertiary acyloxy group is present instead of a ketal (46, Scheme 10B).17 Thus, the LUMO should be less delocalized, higher in energy, and, therefore, less reactive. The Njardarson group reported a similar dearomatization/Diels–Alder sequence in which the Diels–Alder step required

heating.18 In this case, the ketal carbon of 49 should be able to participate in the delocalization of

the LUMO, which would seem to contradict the argument that such delocalization should

enhance reactivity. Perhaps the attenuated reactivity could be explained by a deactivating effect

of the tosylate substituent of 49, which was found to stabilize the intermediate dienone. In addition, this example involves a Diels–Alder cycloaddition between a trisubstituted dienophile

(tiglic acid) and trisubstituted diene, which could be more sterically demanding than disubstitution on both partners, as in our case.

† Interestingly, Diels–Alder reactions of these electron-poor dienones do not appear to behave like straightforward inverse electron-demand Diels–Alder reactions. For instance, dienes related to 43 react faster with methyl vinyl ketone than with ethyl vinyl ether.20 49

O A) O O O O O O O O O O

42 43 43' 44 –1.93 eV –2.94 eV –3.59 eV B)Yates and Auksi (1979):

OH O O Me acrylic acid, Me O Me Me Me Pb(OAc) C H Me 4 O 6 6 reflux O O 45 46 47 Njardarson et al. (2009):

OTs OH tiglic acid, O Me O OMe TsO OMe Pb(OAc)4, TsO PhMe Me OMe O 80 °C O CH2Cl2, 48 0 25 °C O Me O 50 ! 49 Me

Scheme 10. A) Calculated LUMO energies for various dienones using DFT (B3LYP/6-31Gd) by Dory, Deslongchamps, et al. (see ref. 20); B) Examples of Wessely oxidation/Diels–Alder cascades requiring heating to favor Diels–Alder reaction With the successful development of these reaction conditions, we hoped that the resultant tricycle would prove more amenable to the manipulations necessary to form the yunnaneic acid

C and D cores. Pleasingly, the ketal-lactone of 16 proved more labile than the hemiketal of 21,

with the desired hydrolysis occurring in good yield upon exposure of 16 to aqueous TFA at 25

°C for 14 h to afford a model version of yunnaneic acid C as a bright yellow foam (27).‡

‡ Diketone 27 appears to slowly undergo cyclization, in which the carboxyl group attacks the nearby ketone to form a hemiketal (essentially 16 without a methyl group), as indicated by resonances at 197.7 ppm (ketone) and 97.6 ppm (hemiketal) in the 13C NMR of a months-old sample of the material. The NMR of freshly prepared 27 did not have such a hemiketal peak. Strangely, the carbonyls of the diketone also did not appear (carbonyls of similar bicyclic diketones often appear at 200–210 ppm) in the spectrum of freshly prepared material, either. It is unclear why these signals are so weak. 50

Conveniently, the ketal-lactone of 16 serves as a perfect internal protecting group for that

ketone, allowing the other to be chemoselectively reduced to give the proper oxidation pattern of

yunnaneic acid D. However, the desired alcohol stereochemistry would necessitate attack of a

hydride from the concave face of the polycycle. The outcome of this reduction proved highly

dependent on the reducing agent used. For instance, NaBH4 gave a mixture of highly polar products that could not be separated. These products probably resulted from ketal-lactone ring opening and multiple unselective reductions. By contrast, use of a less active reducing agent such as NaBH(OAc)3 cleanly afforded a mixture of exo and endo alcohols 51 and 52 without disturbing the ketal-lactone (the alcohol stereochemistry was determined by NOESY experiments). Interestingly, a slight preference (ca. 1.3:1) for the desired exo alcohol (51) was observed. This result is in contrast with that observed in the synthesis of the helicterin natural products by my former colleague, Dr. Ferenc Kontes, in which the reduction proceeded exclusively from the less hindered face of a similar ketone.2

MeO2C MeO2C MeO2C MeO MeO MeO a) NaBH(OAc) O 3 OH + MeO MeO MeO OMe OMe OMe H H OH O O O O 16 O 51 NOE O 52 1.3 : 1

b) TFA, H2O c) TFA, H2O

MeO2C MeO2C MeO MeO

OH MeO O MeO

O O CO2H CO2H 27 53

Reagents and conditions: a) NaBH(OAc)3 (5.0 equiv), THF/AcOH (1:1), 25 °C, 3 h, 71%; b) TFA, H2O, CH2Cl2, 25 °C, 16 h, 73%; c) TFA, H2O, CH2Cl2, 25 °C, 3 h, 95%

Scheme 11. Elaboration of Diels–Alder product 16 to model yunnaneic acids C and D 51

In hopes of enhancing the selectivity for the exo alcohol, we next investigated Luche-type

conditions known to effect reduction from the more hindered face of bicyclic ketones.21 Results of these experiments are summarized in Table 1. Although higher selectivities could often be observed, the NaBH(OAc)3-mediated reduction proved more reliable in terms of yield and reproducibility. In any case, following careful chromatographic separation of the two alcohols, the endo isomer (52) could be recycled to the ketone via a Ley–Griffith oxidation22 (the Swern oxidation was also successful, but gave inconsistent results with the real system, vide infra).

Table 1. Attempted diastereoselective reductions under Luche-type conditions MeO2C MeO2C MeO2C MeO MeO MeO

O NaBH4 (1.3 equiv) (54) OH MeO MeO + MeO OMe OMe OMe Ln salt (1.3 equiv) (55) H H OH O CH2Cl2, MeOH, -78 °C O O O 16 O 51 O 52

Entry Order Yield (%) 51:52 Ln salt of addn.

1 CeCl3 16, 55, 54 27 (64 brsm) 4:1

2 CeCl3 54, 55, 16 52 1.6:1

3 CeCl3 55, 54, 16 80 1.8:1

4 CeCl3•7H2O 16, 55, 54 40 (84 brsm) 2.8:1

5 CeCl3•7H2O 55, 54, 16 72 2.2:1

6 SmCl3•6H2O 55, 54, 16 48 2.8:1

7 CeCl3 (2.6 equiv) 55, 54, 16 35 2.2:1

Having obtained the desired alcohol, the ketal-lactone could be opened with aqueous

TFA to give the desired hydroxyketone 53 (Scheme 11). The stereo- and regiochemistry of the hydroxyketone could be determined by X-ray crystallography, although only an initial solution 52

could be obtained due to the presence of disordered solvent molecules in the unit cell

(subsequent attempts at re-growing X-ray quality crystals were unsuccessful). Significantly, the ratio of TFA and H2O in the reaction proved to be quite important. With smaller amounts of

water present, opening of the ketal-lactone was sluggish and a small amount of the dimeric side

product 56 (as determined by X-ray crystallography) was isolated in addition to the

hydroxyketone (Figure 2). In addition, the reaction required careful monitoring, as extended

reaction times resulted in the isomerization of the hydroxyketone to a mixture of regio- and

stereoisomers.

Figure 2. Dimeric side product observed following ketal-lactone hydrolysis

In summary, the above sequence allowed access to significantly larger amounts

(hundreds of milligrams in a single run) of model versions of the monomeric natural products

than was previously possible. Importantly, the model dearomatization/Diels–Alder reaction

could proceed from commercially available materials (instead of material arising from a series of

low-yielding steps) and provided an intermediate that was much easier to synthetically

manipulate in the desired fashion.

53

2.3 Total Synthesis of Rufescenolide

As mentioned in Section 1.3.6, we were intrigued by the recent isolation of

rufescenolide23 (58) because of its potential implications for the biosynthesis of the yunnaneic acids. It further piqued our interest because of its striking resemblance to Diels-Alder product 16.

Indeed, the only operations that would be required from 16 would be reduction of the ketone, reductive removal of the methoxy group on the ketal-lactone, and removal of the methyl ether protecting groups on the catechol.

CO2Me MeO C MeO C 2 2 MeO CO2Me MeO MeO OMe a) TMSOTf, MeO OMe OH Et3SiH OH MeO MeO + O OMe H O O O O O O O O 51 57 O 56

b) BBr3

MeO2C HO

OH HO H

O O 58: rufescenolide

Reagents and conditions: a) TMSOTf (3.0 equiv), Et3SiH (20 equiv), CH2Cl2, 25 °C, 2.5 h, 54% (57), 25% (56); b) BBr3 (6.0 equiv), CH2Cl2, 0 °C, 10 min, 55%

Scheme 12. Total synthesis of rufescenolide

Alkoxy groups such as the one found in 51 could normally be removed by activation with a Lewis acid to generate an intermediate oxacarbenium ion, which can be reduced by a hydride donor. However, the oxacarbenium ion generated in this case (59, Scheme 13A) would likely be unstable because of poor overlap between a lone pair of the lactone oxygen and the orbital on the neighboring carbon bearing the positive charge (i.e. an “anti-Bredt” oxacarbenium ion). 54

Despite this reservation, 51 reacted with TMSOTf in the presence of Et3SiH to give the

24 desired product (57, Scheme 12). A major side product was the C2-symmetric dimer 56 that

had been previously characterized, presumably resulting from attack on the cationic intermediate

by a second molecule of 51 instead of the hydride. This side reaction could be minimized

(although not completely suppressed) by using a large excess of Et3SiH and performing the reaction under more dilute conditions (0.03 M). The hypothesized instability of oxacarbenium 59 led us to consider an alternative mechanism (Scheme 13), such as one in which oxacarbenium 60 would be formed instead, followed by reduction by Et3SiH to methyl ether 61. The methoxy group of this ether could then be ionized after complexation with another equivalent of the Lewis acid to give secondary carbocation 62; this intermediate could then be intercepted by the carboxylate to ultimately deliver 57. Interestingly, a trace product tentatively assigned as methyl ether 63 was also isolated from the reaction. However, resubjection of 63 to TMSOTf did not result in formation of lactone 57, suggesting that the methyl ether was simply a side product resulting from competitive ring opening and is not an intermediate on the desired reaction pathway. Thus, we favor the seemingly more straightforward mechanistic route via

oxacarbenium 59. From lactone 57, rufescenolide could be obtained following deprotection of

the methyl ethers with BBr3, which proceeded uneventfully. Overall, the synthesis of rufescenolide required just four steps from methyl ferulate (3).

55

A) MeO2C MeO2C MeO2C MeO MeO MeO

OH TMSOTf OH Et SiH OH MeO MeO 3 MeO OMe -MeOTMS O O O O 51 O 59 O 57

TMSOTf

MeO2C MeO2C MeO2C MeO MeO MeO

Et3SiH OH OH TMSOTf OH MeO MeO MeO -MeOTMS OMe OMe TMSO O TMSO O TMSO O 60 61 62

B) MeO2C MeO2C MeO MeO OH TMSOTf MeO OH MeO OMe HO O O O 63 57

Scheme 13. A) Proposed mechanisms for transformation of 46 to 52. B) Attempted conversion of putative intermediate to 57

2.4 Total Synthesis of Yunnaneic Acids C and D

With a reliable route to model versions of yunnaneic acid C and D established, our attention turned to adapting that strategy to the natural products themselves. The first task was synthesizing fully functionalized Diels–Alder precursors with an appropriate set of protecting groups on the catechols and the carboxyl group of the phenyllactic acid side-chain. We ultimately determined that the protecting groups on the catechols would need to be orthogonal to that placed on the carboxyl group (vide infra) and would need to be readily removable without cleaving the side chain of the natural products. After exploring several protecting group 56

ensembles,§ it was found that the combination of benzyl ethers and an allyl ester was ideal.

Indeed, this combination had been used in a prior synthesis of rosmarinic acid (64).25

Synthesis of phenolic Diels–Alder precursor 72 commenced with the esterification of commercially available, enantiopure rosmarinic acid (64) with TMSCHN2 followed by benzylation of the four phenols of the crude methyl ester (Scheme 14). The central ester linkage of 65 was then cleaved by NaOMe to give two fragments, α, β-unsaturated ester 66 and chiral alcohol 67. The methyl ester of 66 could be hydrolyzed by LiOH to give 69, the dienophile of the planned Diels–Alder reaction. The methyl ester of 67 was exchanged for an allyl ester by simple hydrolysis (which occurred without any observed racemization, as determined by observation of a single diastereomer upon coupling of the alcohol with chiral, non-racemic O-acetyl mandelic acid) and alkylation of the acid. The resulting chiral allyl ester (70) was then coupled with carboxylic acid 71 (synthesized in two steps from ferulic acid2) with EDC and 4-DMAP.

Desilylation of the crude coupling product then gave the desired phenol 72. This straightforward sequence (adapted from previous syntheses of caffeic acid metabolites2,26) could reliably deliver

multi-gram quantities of 72. Of note, because an excess of the dienophile would be needed, our

supply of carboxylic acid 69 derived from rosmarinic acid could be supplemented by

perbenzylating caffeic acid and then hydrolyzing the benzyl ester (68).

§ Several catechol protecting groups (e.g. MOM ether) and carboxylic acid protecting groups (e.g. trimethylsilylethyl) failed to undergo the dearomatization/Diels–Alder cascade while some other protecting groups (e.g. allyl ethers, methyl ester) underwent the key cascade reaction but presented difficulties in their removal. 57

OH OBn OH OBn

BnO CO2Me

BnO 66 a) TMSCHN2 c) NaOMe O O O O + b) K2CO3, BnBr CO H 2 CO2Me BnO OH

CO Me BnO 2 67 OH OBn OH OBn d) LiOH 64: rosmarinic 65 e) K2CO3, AllylBr acid

BnO CO2R i) LiOH BnO CO2H BnO OH

CO Allyl BnO BnO BnO 2 66 R = Me 69 70 68 R = Bn f) EDC, 4-DMAP MeO CO2H 71 g) TBAF, AcOH TBSO 71 OH OMe AllylO2C O BnO O

OBn h) Pb(OAc)4 BnO BnO 69 O O O OMe CO2Allyl O O 73a/73b OBn OBn 72

Reagents and conditions: a) TMSCHN2 (0.95 equiv), THF/MeOH (10:1), -78!25 °C; b) BnBr (6.0 equiv), K2CO3 (6.0 equiv), DMF, 55 °C, 16 h; c) NaOMe (1.0 equiv), CH2Cl2/MeOH (1:1), 25 °C, 95% (for 66 over 3 steps), 81% (for 67 over 3 steps); d) LiOH (1.4 equiv), THF/MeOH/H2O (3:1:1), 45 °C, 6 h; e) K2CO3 (1.5 equiv), AllylBr (1.5 equiv), DMF, 25 °C, 2 h; f) EDC (2.0 equiv), DMAP (1.6 equiv), 71 (3.0 equiv), CH2Cl2, 25 °C, 3h; g) TBAF (3.0 equiv), AcOH, THF, 0 °C, 1 h, 82% (4 steps); h) 70 (7.0 equiv), Pb(OAc)4 (1.1 equiv), CH2Cl2, 25 °C, 1 h; then 1,4-dioxane, 25 °C, 30 min; then 72 (1.0 equiv), 25 °C, 5 min, 50%; i) LiOH (1.4 equiv), THF/MeOH/H2O (3:1:1), 45 °C, 6 h, 76% (two steps from caffeic acid), 65% (from 66)

Scheme 14. Synthesis of cascade precursor and successful modified Wessely oxidation/Diels–Alder cascade 58

Pleasingly, upon subjection of this fully functionalized material to the

dearomatization/Diels–Alder protocol, the desired adduct was obtained in a reasonable 50%

yield, albeit as a 1:1 mixture of diastereomers 73a and 73b. This lack of stereoselectivity is

consistent with previously reported dearomatizations of arenes bearing only a distant chiral center.2,27 Nevertheless, this result constitutes, to the best of our knowledge, the most complex

example of a non-dimeric adduct produced by an oxidative dearomatization/Diels–Alder cascade in which the diene and dienophile were not tethered prior to the reaction.28

In order to complete the syntheses of yunnaneic acid C and D, we needed to identify a

method to separate the mixture of two diastereomers afforded by the Diels–Alder reaction. To our dismay, no solvent system proved capable of doing so chromatographically, even if preparative TLC or semi-preparative HPLC was used. Fortunately, treatment of the mixture of

73a and 73b with NaBH(OAc)3, as in the model system, effected a clean reduction of the ketone, resulting in four diastereomers (i.e. two exo alcohols, 74a/b, and two endo alcohols, 75a/b), which could be separated via careful preparative TLC. The two undesired endo alcohols (75a and 75b) could be subjected to a Ley–Griffith oxidation to give the Diels-Alder products 73a and

73b as two separate diastereomers. 73b could be carried forward to yunnaneic acid C** (1) (and

73a to the epimer of yunnaneic acid C’s enantiomer) by a three-step sequence. First, the four

benzyl ethers of 73b were cleaved with BCl3 in CH2Cl2 at -78 °C in 79% yield. The allyl ester

was then removed under palladium catalysis29 (Meldrum’s acid proved to be the optimal allyl

scavenger). Finally, the ketal-lactone could be opened with aqueous TFA to afford yunnaneic

acid C (1).

** Both diastereomers were carried through the described synthetic sequence, and it was found that the 1H NMR spectra of the diketone (1) and quinoxaline (77) derived from the more polar endo alcohol (75b) more closely matched those reported by Tanaka et al.1 (See Experimental Section) 59

AllylO2C OBn AllylO2C OBn AllylO2C OBn O O O O O O BnO OBn BnO OBn BnO OBn a) NaBH(OAc)3 BnO BnO BnO + OH O OMe OMe OMe OH O O O O O O 75b 73a/73b 74b e) TPAP, NMO b) BCl 3 f) BCl3 c) Pd(PPh ) , Meldrum's acid 3 4 g) Pd(PPh3)4, d) HCl, H2O Meldrum's acid h) TFA, H2O HO2C OH HO2C OH O O HO OH O O HO OH HO HO OH O CO H O 2 O CO2H 2: yunnaneic acid D 1: yunnaneic acid C

Reagents and conditions: a) NaBH(OAc)3 (5.0 equiv), THF/AcOH (1:1), 25 °C, 82% (1.5:1 74b:75b); b) BCl3 (8.0 equiv), CH2Cl2, -78 °C, 5 min; c) Pd(PPh3)4 (5 mol %), Meldrum's acid (1.5 equiv), THF, 25 °C, 15 min, 55% (2 steps); d) HCl, H2O, CH2Cl2, 25 °C, 2 h, 95%; e) TPAP (5 mol %), NMO (2.0 equiv), 4 Å mol. sieves, CH2Cl2, 0 °C, 1 h, 67%; f) BCl3 (8.0 equiv), CH2Cl2,-78 °C, 79%; g) Pd(PPh3)4 (10 mol %), Meldrum's acid (1.5 equiv), THF, 25 °C, 15 min; h) TFA, H2O, CH2Cl2, 25 °C, 16 h, 50% (2 steps)

Scheme 15. Completion of total syntheses of yunnaneic acids C and D

Tanaka and Kouno et al. characterized yunnaneic acid C (1) as its quinoxaline derivative

(77) because the diketone was readily hydrated, leading to complicated NMR spectra and a broad peak on HPLC analysis.1 This derivative was prepared by these workers via condensation of the diketone with o-phenylenediamine (76) under acidic conditions. Subjection of our synthetic yunnaneic acid C to the same conditions (although a longer reaction time than that reported was necessary in our case) yielded a quinoxaline derivative (77) whose spectral data matched that of the naturally-derived compound (Scheme 16). 60

HO2C OH HO2C OH NH2 O O O O HO OH HO OH

NH2 HO 76 HO

O EtOH, AcOH, N 60 °C, 19 h O N CO2H 19 % CO2H 1: yunnaneic acid C 77

Scheme 16. Derivatization of yunnaneic acid C

The exo alcohol 74b obtained from the reduction could be subjected to a similar sequence to afford yunnaneic acid D (2).†† The only difference in this case was that synthetic yunnaneic acid D was found to readily isomerize, presumably forming a different hydroxyketone stereo- or regioisomer upon standing in the presence of residual TFA that lingered following the final ketal-lactone cleavage step. This isomerization occurred even during the course of attempted removal of trace TFA via azeotroping with toluene. Thus, the selection of an appropriate acid for ketal-lactone cleavage was crucial. HCl proved to be well suited to this transformation in that it was sufficiently strong to carry out the reaction in a reasonably short time and also easy to remove in a non-basic aqueous work-up. Spectral data for synthetic yunnaneic acid D (2) agreed

with that reported by the isolation chemists.1

The presence of an orthogonal protecting group on the side-chain carboxylate proved to be critical, since it was found that premature unmasking of this group had deleterious effects. For example, when the acid was also protected as a benzyl ester (as in 83, Scheme 17), significant side-chain cleavage was observed upon exposure to Lewis acids. A methyl ester would have been preferable to the allyl ester that was ultimately used since it would shorten the synthesis by

†† Both 74a and 74b were carried forward to yunnaneic acid D and the epimer of its enantiomer, and comparison of 13C NMR data of the two diastereomers with that reported by Tanaka et al.1 showed that the material derived from 74b corresponded to yunnaneic acid D. 61

two steps. Although debenzylation of the methyl ester derivative proceeded smoothly,

30 demethylation under mild conditions such as Me3SnOH was unsuccessful.

A slightly different route to yunnaneic acids C and D involving a different order of events

was developed in parallel to that described above. The purpose of this route was to provide a way

to obtain compound 83 as a single diastereomer since we had been experiencing difficulties with

separating the diastereomers arising from the Diels–Alder reaction using traditional

chromatographic techniques. Thus, we targeted carboxylic acid 81 with the hopes that

diastereomeric salts could be formed with a chiral amine and separated to provide 81 as a single

enantiomer to couple with the enantiopure chiral alcohol 82 (Scheme 17). The carboxylic acid

would need to be protected prior to the dearomatization event, and an allyl ester was deemed

suitable since it could easily be removed under mild conditions (the ketal-lactone was found to

be unstable to nucleophilic bases).

The desired Diels-Alder precursor 79 was synthesized by subjecting ferulic acid (78) to

Na2CO3 and allyl bromide. The sodium counterion was critical, as the stronger K2CO3 resulted in competitive alkylation of the phenol. The dearomatization/Diels–Alder cascade proceeded without incident, and the allyl ester of 80 could be deprotected under the action of palladium catalysis. The resulting racemic carboxylic acid (81) was subjected to a number of different chiral amine bases, but, unfortunately, no enantiomerically enriched material was obtained from these mixtures of diastereomeric salts. Nevertheless, racemic 81 could be coupled with chiral alcohol 82 using EDC and 4-DMAP to give 83a/b (as a mixture of diastereomers), thus providing an alternative route to yunnaneic acids C and D. Ultimately, this route was discarded in favor of the one described previously because the EDC coupling reaction required two 62 equivalents of the carboxylic acid. Although the excess acid could be recovered, it made the route less efficient in terms of overall material throughput.

OH OH AllylO2C OMe OMe BnO

a) Na2CO3, b) Pb(OAc)4 BnO O AllylBr OBn OMe OBn O CO2H CO2Allyl O 69 80 78: ferulic acid 79

CO2H c) Pd(PPh3)4 Meldrum's acid

BnO2C O HO C BnO O 2 BnO OBn d) EDC, DMAP BnO BnO BnO O O BnO OH OMe OMe CO Bn BnO 2 O O 82 O 81 O 83a/83b

Reagents and conditions: a) Na2CO3 (1.2 equiv), AllylBr (1.1 equiv), DMF, 25 °C, 9 h, 88%; b) 69 (7.0 equiv), Pb(OAc)4 (1.1 equiv), CH2Cl2, 25 °C, 1 h; then 1,4-dioxane, 25 ° C, 30 min; then 79 (1.0 equiv), 25 °C,10 min, 58%; c) Pd(PPh3)4 (5 mol %), Meldrum's acid (1.5 equiv), THF, 25 °C, 15 min, 95%; d) EDC (1.0 equiv), DMAP (0.8 equiv), 82 (0.5 equiv), CH2Cl2, 25 °C, 16 h, 84% (based on 82)

Scheme 17. Alternate route to yunnaneic acids involving later-stage side chain coupling

2.5 Conclusion

In conclusion, the first total syntheses of yunnaneic acids C and D were achieved in 13 and 12 steps, respectively, with the key step being an oxidative dearomatization/Diels–Alder cascade on a complex substrate. This cascade completely controlled the relative stereo- and regiochemistry of the carbon substituents on the bicyclic core. Furthermore, the product of this cascade sequence (84) could be leveraged to access other caffeic acid metabolites. In addition to 63

rufescenolide (58), yunnaneic acid F (85) would appear readily accessible via formal addition of acetic acid to the diketone of yunnaneic acid C, and yunnaneic acid E (86) could conceivably

arise from an oxidative cleavage of the same diketone.31

HO2C HO2C O O HO O HO O

OH OH HO HO HO HO OH O

O O CO2H CO2H 1: yunnaneic acid C 2: yunnaneic acid D R2O CO2R1

R2O O OMe [O] O HO2C O 84 HO CO H O 2 +AcOH HO O HO HO OH OH HO HO2C H O CO2H HO O CO H O 2 O CO2H HO OH 58: rufescenolide HO 86: yunnaneic acid E

O O OH CO2HOH

85: yunnaneic acid F

Scheme 18. Potential access to multiple yunnaneic acids and related natural products from Diels- Alder product 84

In addition to further demonstrating the synthetic utility of MOB ketals as reactive

intermediates, this work also underscores the need for stereoselective dearomatization

reactions.32 Current methods using chiral iodides in catalytic amounts along with a

stoichiometric co-oxidant hold great promise.33 However, these methods are limited in scope and often require that the nucleophile be tethered to the arene. Given that Pb(IV) can coordinate four 64

ligands and only two sites are needed for the reacting species, a chiral bidentate ligand34,35 could conceivably influence the stereochemistry of the dearomatization, especially if an intramolecular

carboxylate transfer step is operative (see Section 2.2). Preliminary experimentation with

camphoric acid as a chiral bidentate dicarboxylate ligand did not show any stereoselectivity, but

the dearomatization/Diels–Alder reaction did, nevertheless, proceed in decent yield (50%,

unoptimized). Thus, perhaps further screening of chiral, non-racemic bidentate ligands could

result in useful levels of stereoselectivity, although the prospect of a catalytic process seems

poor.

In addition, as alluded to in Chapter 1, the spontaneous nature of the Diels–Alder reaction

to form the bicyclic core of the yunnaneic acids would seem to support an argument that the

biosynthesis of these molecules proceeds through a similar intramolecular Diels–Alder reaction.

Such a biosynthetic Diels–Alder reaction would explain the observed stereochemistry and would

also not require the invocation of an elusive “Diels–Alderase.”

Finally, this efficient route to the monomeric yunnaneic acids also provided access to

ample stockpiles of model versions of yunnaneic acids C and D and derivatives thereof to facilitate our synthetic efforts towards yunnaneic acids A and B, which are the subject of the next chapter.

Postdoctoral fellow Dr. Lorenzo Botta assisted me in the final stages of the work

presented in this chapter. In particular, he assisted in determining the proper protecting group

ensemble and helped determine the proper conditions for their removal. Tyler St. Denis, an

undergraduate working under my supervision, performed the total synthesis of rufescenolide and

I optimized the reductive demethoxylation reaction. Dr. Ferenc Kontes, a fellow graduate

student, developed the initial, first generation route to model diketone 36, and he proved an 65 invaluable source of advice throughout the project because of his experience with the helicterins.

In addition, I wish to thank several members of the Parkin group, who performed the X-ray analyses: Ava Kreider-Mueller (16), Dr. Aaron Sattler (38), Dr. Wesley Sattler (53), and Ahmed al-Harbi (56).

2.6 References

(1) Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Young, T. J. J. Nat. Prod. 1996, 59, 843.

(2) Snyder, S. A.; Kontes, F. J. Am. Chem. Soc. 2009, 131, 1745.

(3) Chu, C. S.; Lee, T. H.; Rao, P. D.; Song, L. D.; Liao, C. C. J. Org. Chem. 1999, 64, 4111.

(4) Perez, A.; Gimeno, N.; Vera, F.; Ros, M. B.; Serrano, J. L.; De la Fuente, M. R. Eur. J. Org. Chem. 2008, 826.

(5) Luo, S. Y.; Jang, Y. J.; Liu, J. Y.; Chu, C. S.; Liao, C. C.; Hung, S. C. Angew. Chem. Int Ed. 2008, 47, 8082.

(6) Brotin, T.; Devic, T.; Lesage, A.; Emsley, L.; Collet, A. Chem. Eur. J. 2001, 7, 1561.

(7) Nicolaou, K. C.; Baran, P. S.; Zhong, Y. L.; Fong, K. C.; Choi, K. S. J. Am. Chem. Soc. 2002, 124, 2190.

(8) Evans, D. A.; Chapman, K. T. Tetrahedron Lett. 1986, 27, 5939.

(9) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem. 1987, 52, 2559.

(10) Overman, L. E.; Paone, D. V. J. Am. Chem. Soc. 2001, 123, 9465.

(11) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155. 66

(12) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277.

(13) Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549.

(14) Andersson, G. Acta. Chem. Scand. B 1976, 30, 403.

(15) Wessely, F.; Lauterbachkeil, G.; Sinwel, F. Monatsh. Chem. 1950, 81, 811.

(16) Bichan, D. J.; Yates, P. Can. J. Chem. 1975, 53, 2054.

(17) Yates, P.; Auksi, H. Can. J. Chem. 1979, 57, 2853.

(18) Morton, J. G. M.; Draghici, C.; Kwon, L. D.; Njardarson, J. T. Org. Lett. 2009, 11, 4492.

(19) Harrison, M. J. J. Chem. Soc. C 1970, 728.

(20) Dory, Y. L.; Roy, A. L.; Soucy, P.; Deslongchamps, P. Org. Lett. 2009, 11, 1197.

(21) Krief, A.; Surleraux, D. Synlett 1991, 273.

(22) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J. Chem. Soc. Chem. Comm. 1987, 1625.

(23) do Vale, A. E.; David, J. M.; dos Santos, E. O.; David, J. P.; de Silva, L. C. R. C.; Bahia, M. V.; Brandao, H. N. Phytochemistry 2012, 76, 158.

(24) Alonso, D.; Perez, M.; Gomez, G.; Covelo, B.; Fall, Y. Tetrahedron 2005, 61, 2021.

(25) Eicher, T.; Ott, M.; Speicher, A. Synthesis 1996, 755.

(26) O'Malley, S. J.; Tan, K. L.; Watzke, A.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc. 2005, 127, 13496.

(27) Gagnepain, J.; Castet, F.; Quideau, S. Angew. Chem. Int. Ed. 2007, 46, 1533. 67

(28) Roche, S. P.; Porco, J. A. Angew. Chem. Int. Ed. 2011, 50, 4068.

(29) Roush, W. R.; Sciotti, R. J. J. Am. Chem. Soc. 1998, 120, 7411.

(30) Nicolaou, K. C.; Estrada, A. A.; Zak, M.; Lee, S. H.; Safina, B. S. Angew. Chem. Int. Ed. 2005, 44, 1378.

(31) Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Yang, C. R. Chem. Pharm. Bull. 1997, 45, 1596.

(32) Magdziak, D.; Meek, S. J.; Pettus, T. R. R. Chem. Rev. 2004, 104, 1383.

(33) Dohi, T.; Takenaga, N.; Nakae, T.; Toyoda, Y.; Yamasaki, M.; Shiro, M.; Fujioka, H.; Maruyama, A.; Kita, Y. J. Am. Chem. Soc. 2013, 135, 4558.

(34) Moloney, M. G.; Paul, D. R.; Prottey, S. C.; Thompson, R. M.; Wright, E. J. Organomet. Chem. 1997, 534, 195.

(35) Moloney, M. G.; Paul, D. R.; Thompson, R. M.; Wright, E. J. Organomet. Chem. 1998, 560, 77.

68

2.7 Experimental Section

General Procedures. All reactions were carried out under an argon atmosphere with dry solvents under anhydrous conditions, unless otherwise stated. Dry methylene chloride (CH2Cl2), diethyl ether (Et2O), and tetrahydrofuran (THF) were obtained by passing commercially

available pre-dried, oxygen-free formulations through activated alumina columns; triethylamine

(Et3N) was distilled from KOH; acetone, methanol (MeOH), and dimethylformamide (DMF)

were purchased in anhydrous form from Sigma-Aldrich and used as received. Yields refer to

chromatographically and spectroscopically (1H and 13C NMR) homogeneous materials, unless

otherwise stated. Reagents were purchased at the highest commercial quality and used without

further purification, unless otherwise stated. Reactions were magnetically stirred and monitored

by thin-layer chromatography (TLC) carried out on 0.25 mm E. Merck silica gel plates (60F-

254) using UV light and an aqueous solution of cerium ammonium sulfate and ammonium

molybdate and heat as visualizing agents. Preparative TLC was carried out on 0.50 mm E. Merck

silica gel plates (60F-254). SiliCycle silica gel (60 Å, academic grade, particle size 40-63 µm)

was used for flash column chromatography. NMR spectra were recorded on Bruker DRX-300,

DRX-400, DRX-500, and 500 ASCEND instruments and calibrated using residual undeuterated

solvent as an internal reference. The following abbreviations are used to explain multiplicities: s

= singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad, app = apparent. IR

spectra were recorded on a Perkin-Elmer Spectrum Two FT-IR spectrometer. High resolution

mass spectra (HRMS) were recorded in the Columbia University Mass Spectral Core facility on

a JOEL HX110 mass spectrometer using FAB (Fast Atom Bombardment). Optical rotations were 69

recorded on a Jasco DIP-1000 digital polarimeter. Characterization data are not provided for

known compounds.

Diels–Alder Homodimer (8). To a solution of 6 (2.00 g, 9.60 mmol, 1.0 equiv) in

MeOH/CH2Cl2 (5:1, 90 mL) was added PhI(OAc)2 (4.57 g, 14.4 mmol, 1.5 equiv) at 0 °C. The

resulting deep yellow solution was stirred for 20 min and then concentrated directly. The residue

was purified by flash column chromatography (silica gel, 3:2 hexanes:EtOAc) to give the

product as a yellow solid, (2.07 g, 97 %, 5:1 mixture of regioisomers). 8: 1H NMR (400 MHz,

CDCl3) δ 7.24 (d, J = 16.0 Hz, 1 H), 7.15 (d, J = 16.0 Hz, 1 H), 6.44 (d, J = 7.0 Hz, 1 H), 6.32

(d, J = 16.2, 1 H), 6.13 (s, 1 H), 5.61 (d, J = 15.9 Hz, 1 H), 3.84 (s, 3 H), 3.72 (s, 3 H), 3.71–3.67

(m, 1 H), 3.55 (t, J = 2.3 Hz, 1 H), 3.48 (s, 3 H), 3.42 (s, 3 H), 3.35 (dd, J = 8.5, 1.7 Hz, 1 H),

3.27 (dd, J = 7.0, 1.8 Hz, 1 H), 3.24 (s, 3 H), 3.03 (s, 3 H).

Allyl-protected Phenol (19). Methyl caffeate (13) (0.890 g, 4.59 mmol, 1.0 equiv) was dissolved in DMF (24 mL) and solid KI (0.083 g, 0.50 mmol, 0.1 equiv) was added at 25 °C.

The resultant solution was then cooled to –50 °C and NaH (60 % dispersion in mineral oil, 0.193 g, 5.20 mmol, 1.05 equiv) was added. The resulting yellow solution was stirred for 5 min at –50

°C, allyl bromide (0.75 mL, 9.9 mmol, 2.0 equiv) was added, and the resulting mixture was allowed to gradually warm to 25 °C and stirred for an additional 16 h. Upon completion, the reaction mixture was poured into 1 M HCl (50 mL) and extracted with Et2O (3 × 50 mL). The

combined organic layers were then washed with water (3 × 50 mL), dried (MgSO4), filtered, and concentrated. The resultant crude product was purified by flash column chromatography (silica gel, hexanes:EtOAc, 9:1à1:1) to give the desired phenol (0.605 g, 57% yield, 4.5:1 mixture of 70

regioisomers) as a white solid alongside the diallylated compound (0.228 g, 18% yield) and

recovered starting material (0.120 g, 13 %). 19: Rf = 0.31 (hexanes:EtOAc, 4:1); IR (film) νmax

3387, 2960, 2865, 1694, 1629, 1609, 1580, 1504, 1426, 1316, 1267, 1163, 1123, 1023, 979, 934,

–1 1 860, 800 cm ; H NMR (500 MHz, CDCl3) δ 7.62 (d, J = 16.0 Hz, 1 H), 7.17 (d, J = 2.0 Hz, 1

H), 7.03 (dd, J = 8.5, 2.0 Hz, 1 H), 6.87 (d, J = 8.5 Hz, 1 H), 6.32 (d, J = 16.0 Hz, 1 H), 6.12–

6.04 (m, 1 H), 5.69 (s, 1 H), 5.44 (dd, J = 17.5, 1.5 Hz, 1 H), 5.37 (dd, J = 10.5, 1.0 Hz, 1 H),

13 4.67 (app dt, J = 4.0, 1.5 Hz, 1 H), 3.82 (s, 3 H); C NMR (125 MHz, CDCl3) 167.7, 147.5,

146.1, 144.7, 132.3, 128.2, 121.6, 118.7, 115.9, 113.3, 111.9, 69.8, 51.6; HRMS (FAB) calcd for

+ + C13H14O4 [M] 234.0892, found 234.0904.

OMe OAllyl O OMe O

87 CO2Me

DCC Coupling Product (87). Adapted from A. Perez; F. Vera; M. B. Ros; J. L. Serrano; M. R.

De la Fuente, Eur. J. Org. Chem. 2008, 826-833. To a solution of 19 (0.480 g, 2.10 mmol, 1.0

equiv) and 14 (0.667 g, 3.20 mmol, 1.5 equiv) in CH2Cl2 (42 mL) was added sequentially DCC

(0.660 g, 3.20 mmol, 1.5 equiv) and 4-DMAP (0.026 g, 0.200 mmol, 0.1 equiv) at room temperature. The resulting solution was stirred for 24 h. The mixture was then filtered and the filtrate was concentrated. The white solid was purified by flash column chromatography (silica gel, 2:1 hexanes:EtOAc) to give the ester as a white amorphous solid (1.05 g, contained some

1 dicyclohexylurea impurity). 87: H NMR (300 MHz, CDCl3) δ 7.83 (d, J = 15.9 Hz, 1 H), 7.62

(d, J = 15.9 Hz, 1 H), 7.40–7.28 (m, 2 H), 7.22–7.07 (m, 2 H), 6.97 (d, J = 8.5 Hz, 1 H), 6.90 (d,

J = 8.3 Hz, 1 H), 6.54 (d J = 15.9 Hz, 1 H), 6.30 (d, J = 15.9 Hz, 1 H), 5.97 (ddt, J = 17.2, 10.5, 71

4.9 Hz, 1 H), 5.38 (dd, J = 17.3, 1.6 Hz, 1 H), 5.24 (dd, J = 10.6, 1.5 Hz, 1 H), 4.61 (dd, J = 5.0,

1.7 Hz, 2 H), 3.93 (s, 6 H), 3.79 (s, 3 H).

Phenol (15). Adapted from S.-L. Luo; Y.-J. Jiang; J.-Y. Liu; C.-S. Chu; C.-C. Liao; S.-C. Hung,

Angew. Chem. Int. Ed. 2008, 47, 8082-8085. To a solution of allyl ether 87 (0.020 g, 0.048

mmol, 1.0 equiv) in AcOH (1.0 mL) was added Pd(PPh3)4 (0.006 g, 0.005 mmol, 0.1 equiv). The solution was heated to 80 °C and stirred for 40 min. The mixture was filtered through Celite and concentrated. The crude solid was purified by flash column chromatography to give the product

1 as a white solid (0.009 g, 67 % over two steps). 15: H NMR (300 MHz, CDCl3) δ 7.85 (d, J =

15.9 Hz, 1 H), 7.60 (d, J = 16.0 Hz, 1 H), 7.46–7.30 (m, 2 H), 7.17 (d, J = 8.3 Hz, 1 H), 7.10 (dd,

J = 3.4, 1.9 Hz, 1 H), 7.03 (d, J = 8.3 Hz, 1 H), 6.89 (d, J = 8.3 Hz, 1 H), 6.50 (d, J = 15.9 Hz, 1

H), 6.30 (d, J = 15.9 Hz, 1 H), 3.93 (s, 6 H), 3.78 (s, 3 H).

Allylic Alcohol (23).1 Adapted from S. B. Wan; Q. P. Dou; T. H. Chan, Tetrahedron, 2006, 62,

5897-5904. To a solution of 4 (1.00 g, 4.50 mmol, 1.0 equiv) in THF (38 mL) was added

DIBAL-H (1.0 M in toluene, 7.20 mL, 7.20 mmol, 3.2 equiv) at -78 °C. The solution was slowly allowed to warm to room temperature and stirred for 24 h. Upon completion, the solution was cooled to 0 °C and then poured into a stirring mixture of hexanes (65 mL) and saturated aqueous

MgSO4 (2.5 mL). The mixture was stirred for 2 h, and the resulting cloudy suspension was filtered and the solid was washed with EtOAc (65 mL). The combined filtrates were then concentrated to give a cloudy colorless oil that solidified upon standing (0.900 g, 99%). This crude material could be carried forward without further purification.

1 Block, E.; Stevenson, R. J. Org. Chem. 1971, 36, 3453. 72

Phenol 20. To a solution of allyl ether 19 (0.380 g, 1.60 mmol, 1.0 equiv), allylic alcohol 23

(0.373 g, 1.90 mmol, 1.2 equiv), and Ph3P (0.628 g, 2.40 mmol, 1.5 equiv) in THF (16 mL) at 0

°C was added DIAD (0.47 mL, 2.40 mmol, 1.5 equiv). The resultant orange solution was then allowed to warm slowly to 25 °C and stirred for 24 h. Upon completion, the reaction mixture was concentrated directly and the resulting oil was purified by flash column chromatography

(silica gel, hexanes:EtOAc, 7:3→1:1) to give the etherification product as a clear, colorless oil.

Pressing forward, this intermediate was dissolved in THF (7.5 mL) and Ph3P (135 mg, 0.51

mmol, 0.4 equiv) and H2O (1.4 mL) were added sequentially at 25 °C. Et2NH (3.75 mL) was then added, and the resultant solution became yellow-green in color. Next, Pd(OAc)2 (57 mg,

0.26 mmol, 0.2 equiv) was added at 25 °C, and during the next few minutes of stirring, the

solution became very dark green in color. The reaction mixture was then stirred for an additional

1 h at 25 °C. Upon completion, the reaction contents were concentrated directly, and the

resultant residue was then diluted with EtOAc (40 mL) and washed with H2O (50 mL) and brine

(50 mL). The organic layer was then dried (MgSO4), filtered, and concentrated to give a dark, green oil. This crude product was then purified by flash column chromatography (silica gel, hexanes:Et2O, 1:1→2:3) to give the desired phenol (0.150 g, 26% yield over two steps) as an off- white solid. 20: Rf = 0.32 (hexanes:EtOAc, 3:2); IR (film) νmax 3359, 2951, 2932, 1698, 1601,

–1 1 1511, 1438, 1262, 1158, 1108, 1025 cm ; H NMR (400 MHz, CDCl3) δ 7.62 (d, J = 15.6 Hz, 1

H), 7.10–7.07 (m, 2 H), 6.98–6.93 (m, 3 H), 6.84 (d, J = 8.8 Hz, 1 H), 6.68 (d, J = 15.6 Hz, 1 H),

6.32–6.24 (m, 2 H), 6.02 (br s, 1 H), 4.77 (d, J = 6.0 Hz, 2 H), 3.92 (s, 3 H), 3.89 (s, 3 H), 3.80

13 (s, 3 H); C NMR (125 MHz, CDCl3) δ 168.6, 150.4, 150.1, 149.2, 146.8, 145.8, 135.3, 129.9,

127.9, 124.0, 122.0, 121.0, 116.1, 115.9, 112.1, 112.0, 109.9, 70.9, 56.9, 56.8, 52.5. 73

Diels–Alder Product (21). To a suspension of phenol 20 (0.054 g, 0.15 mmol, 1.0 equiv) in

1,4-dioxane/H2O (2:1, 1.5 mL) at 75 °C was added a solution of PhI(OAc)2 (0.052 g, 0.16 mmol,

1.1 equiv) in CH2Cl2 (0.5 mL) dropwise under an ambient atmosphere. The resulting dark red

solution was stirred at 75 °C for 5 min and then was cooled to 25 °C. The reaction mixture was then diluted with EtOAc (10 mL), washed with saturated aqueous NaHCO3 (10 mL), and the

aqueous layer was then extracted with EtOAc (2 × 5 mL). The combined organic layers were

then dried (MgSO4), filtered, and concentrated to give a crude brown oil. Purification of this residue by flash column chromatography (silica gel, hexanes:EtOAc, 1:1) to give the desired

Diels–Alder product (0.039 g, 67% yield) as a pale yellow oil. 21: Rf = 0.20 (hexanes:EtOAc,

–1 1 1:1); IR (film) νmax 3421, 2953, 1741, 1631, 1516, 1438, 1314, 1256, 1178, 1144, 1026 cm ; H

NMR (400 MHz, CDCl3) δ 7.43 (d, J = 15.6 Hz, 1 H), 6.76 (d, J = 8.4 Hz, 1 H), 6.54 (d, J = 2.4

Hz, 1 H), 6.47 (dd, J = 8.4, 2.0 Hz, 1 H), 6.29–6.21 (m, 2 H), 4.38 (dd, J = 8.0, 3.2 Hz, 1 H),

4.11 (s, 1 H), 3.93 (d, J = 8.0 Hz, 1 H), 3.85 (s, 3 H), 3.82 (s, 3 H), 3.80 (s, 3 H), 3.73 (dd, J =

4.0, 2.0 Hz, 1 H), 3.51 (dd, J = 7.2, 2.4 Hz, 1 H), 3.43 (br s, 1 H), 2.82 (br s, 1 H); 13C NMR

(100 MHz, CDCl3) δ 200.6, 167.0, 148.9, 148.3, 141.2, 139.1, 133.6, 131.9, 119.7, 118.9, 111.4,

+ 111.2, 96.4, 73.8, 55.9 (2 C), 53.8, 51.9, 47.2, 45.9, 44.8; HRMS (FAB) calcd for C21H22O7

[M]+ 386.1366, found 386.1370.

Diels-Alder Product (22). To a solution of 20 (0.330 g, 0.890 mmol, 1.0 equiv) in MeOH (10 mL) was added a solution of PhI(OAc)2 (0.316 g, 0.980 mmol, 1.1 equiv) in CH2Cl2 (1.0 mL)

dropwise at 60 °C. The resulting solution was stirred at 60 °C for 5 min and then cooled to 25

°C. The mixture was poured into saturated aqueous NaHCO3 (20 mL) and extracted with EtOAc 74

(3 x 20 mL). The combined organic layers were washed with brine (20 mL) and then dried

(MgSO4), filtered, and concentrated to give a crude brown oil. Purification of the residue by flash column chromatography (silica gel, hexanes:EtOAc, 8:2à7:3) to give the desired Diels–

Alder product (0.218 g, 74% yield) as a white amorphous solid. 22: Rf = 0.24 (silica gel,

1 hexanes:EtOAc, 3:2); H NMR (400 MHz, CDCl3) δ 7.38 (d, J = 15.9 Hz, 1 H), 6.70 (d, J = 8.3

Hz, 1 H), 6.48 (d, J = 2.2 Hz, 1 H), 6.40 (dd, J = 8.3, 2.1 Hz, 1 H), 6.26 (d, J = 8.1 Hz, 1 H), 6.14

(d, J = 15.9 Hz, 1 H), 4.26 (dd, J = 8.3, 3.3 Hz, 1 H), 3.94 (d, J = 8.3 Hz, 1 H), 3.84–3.80 (m, 1

H), 3.78 (s, 3 H), 3.76 (s, 3 H), 3.74 (s, 3 H), 3.71 (dd, J = 4.4, 2.2 Hz, 1 H), 3.50 (s, 3 H), 3.41–

3.28 (m, 2 H), 2.77 (br s, 1 H).

Triol (32). To a solution of 21 (0.350 g, 0.91 mmol, 1.0 equiv) in MeCN/AcOH (2:1, 13.5 mL)

was added Me4NBH(OAc)3 (1.19 g, 4.50 mmol, 5.0 equiv) at 0 °C. The resulting solution was

stirred for 2.5 h. Upon completion, 0.25 M aqueous Rochelle’s salt (15 mL) was added and the

mixture was stirred at 25 °C for 45 min. This mixture was then diluted with EtOAc (60 mL). The

layers were separated and the organic layer was washed with saturated aqueous NaHCO3 (60

mL). The organic layer was then dried (MgSO4), filtered, and concentrated to give a cloudy oil.

The crude product was purified by flash column chromatography (silica gel, EtOAc:MeOH,

1 95:5) to give triol 32 (0.273 g, 77 %) as a white solid. H NMR (300 MHz, acetone-d6) δ 7.45 (d,

J = 15.7 Hz, 1 H), 6.92–6.56 (m, 4 H), 6.06 (d, J = 15.7 Hz, 1 H), 4.03 (m, 1 H), 3.88–3.68 (m,

11 H), 3.65 (d, J = 3.7 Hz, 1 H), 3.50 (t, J = 2.9 Hz, 1 H), 3.34 (d, J = 2.3 Hz, 1 H), 2.98 (d, J =

8.1 Hz, 1 H), 2.72 (dd, J = 6.9, 3.2 Hz, 1 H).

75

Acetates (33). To a solution of 22 (0.200 g, 0.50 mmol, 1.0 equiv) in CH2Cl2/MeOH (1:1, 5.0 mL) at 0 °C was added NaBH4 (0.038 g, 1.00 mmol, 2.0 equiv) in a single portion. The resulting solution was stirred for 45 mL. The reaction mixture was warmed to 25 °C and diluted with

EtOAc (15 mL) and poured into saturated aqueous NH4Cl (15 mL). The layers were separated,

and the aqueous layer was re-extracted with EtOAc (3 x 15 mL). The combined organic layers

were washed with brine (50 mL), dried (MgSO4), filtered, and concentrated to give the crude alcohol as a mixture of diastereomers. The crude material was then dissolved in CH2Cl2 (1.0 mL) and Et3N (0.10 mL, 0.75 mmol, 1.5 equiv), Ac2O (0.05 mL, 0.55 mmol, 1.1 equiv), and DMAP

(1 flake, catalytic) were added sequentially at 25 °C. The resulting solution was stirred for 3 h and then MeOH (0.2 mL) was added. The resulting mixture was stirred for 15 min and then diluted with EtOAc (5.0 mL). The mixture was washed with 0.1 N HCl (5.0 mL) followed by saturated aqueous NaHCO3 (5.0 mL). The combined aqueous layers were re-extracted with

EtOAc (2 x 5.0 mL). The combined organic layers were dried (MgSO4), filtered, and

concentrated to give a clear, colorless residue that was purified by flash column chromatography

(silica gel, hexanes:EtOAc, 4:1) to give two diastereomeric alcohols as clear colorless oils

1 (combined yield: 180 mg, 81 % over two steps, 1.33:1 d.r.). 33a: H NMR (300 MHz, CDCl3)

δ 7.52 (d, J = 15.9 Hz, 1 H), 6.73 (d, J = 8.3 Hz, 1 H), 6.53 (d, J = 2.1 Hz, 1 H), 6.48–6.28 (m, 2

H), 6.12 (d, J = 15.8 Hz, 1 H), 5.07 (d, J = 4.0 Hz, 1 H), 3.85 (s, 3 H), 3.83 (s, 3 H), 3.80 (s, 3

H), 3.48 (d, J = 3.9 Hz, 1 H), 3.40 (s, 3 H), 3.15 (ddd, J = 6.5, 4.2, 2.7 Hz, 1 H), 3.09–2.92 (m, 1

H), 2.58–2.42 (m, 1 H), 2.02 (s, 3 H).

1 33b: H NMR (300 MHz, CDCl3) δ 7.42 (d, J = 15.9 Hz, 1 H), 6.73 (d, J = 7.8 Hz, 1 H), 6.47

(m, 3 H), 6.08 (d, J = 15.9 Hz, 1 H), 4.61 (d, J = 2.4 Hz, 1 H), 4.23 (dd, J = 7.9, 3.8 Hz, 1 H), 76

3.95–3.75 (m, 11 H), 3.55–3.43 (m, 1 H), 3.36 (s, 3 H), 3.27–3.16 (m, 1 H), 2.96–2.83 (m, 1 H),

2.61 (d, J = 2.4 Hz, 1 H), 2.21 (s, 3 H).

MeO2C MeO

MeO OH OH O H 88

Aldehyde (88). To a solution of 32 (0.185 g, 0.47 mmol, 1.0 equiv) in CH2Cl2 (5.0 mL) was

added 0.5 M pH 8.6 buffer followed by a catalytic amount of n-Bu4NCl (weighing this reagent was impossible because of its hygroscopic nature). To this biphasic mixture was added TEMPO

(0.004 g, 0.024 mmol, 0.05 equiv) and PhI(OAc)2 (0.231 g, 0.72 mmol, 1.5 equiv) at 0 °C under

ambient atmosphere. The resulting suspension was vigorously stirred at 0 °C for 3 h. Upon

completion, the solution was diluted with EtOAc (30 mL) and washed with saturated aqueous

Na2S2O3 (30 mL). The aqueous layer was further extracted with EtOAc (3 x 30 mL), and the

combined organic layers were then dried (MgSO4), filtered, and concentrated to give a pale red oil. The crude material was purified by flash column chromatography (100 % EtOAc) to give the

1 aldehyde 88 (0.137 g, 75 %) as a yellow oil. 88: H NMR (300 MHz, CDCl3) δ 9.91 (s, 1 H),

7.54 (d, J = 15.7 Hz, 1 H), 6.78 (d, J = 8.2 Hz, 1 H), 6.75–6.63 (m, 3 H), 6.16 (d, J = 15.7 Hz, 1

H), 3.87 (m, 1 H), 3.86 (s, 6 H), 3.82 (m, 1 H), 3.81 (s, 3 H), 3.61 (br s, 1 H), 3.54 (br s, 1 H),

3.00–2.93 (m, 1 H), 2.64 (dd, J = 7.6, 2.4 Hz, 1 H).

Dihydroxy Ester (35). Aldehyde 88 (0.300 g, 0.77 mmol, 1.0 equiv) was dissolved in t-BuOH

(8.0 mL) and the flask was immersed in a cold water bath. To this solution was added 2-methyl- 77

2-butene (0.82 mL, 7.70 mmol, 10.0 equiv). An aqueous solution (10 mL) of NaH2PO4•2H2O

(1.20 g, 7.70 mmol, 10.0 equiv) and NaClO2 (0.351 g, 3.89 mmol, 5.0 equiv) was then added and the cooled solution was stirred for 1 h. Upon completion, the reaction mixture was poured into saturated aqueous NH4Cl (35 mL) and extracted with EtOAc (3 x 35 mL). The combined organic

layers were dried (MgSO4), filtered, and concentrated to give the crude acid as a light yellow oil,

which was carried forward to the next step without additional purification. This oil was dissolved

in THF/MeOH (4:1, 7.5 mL) and TMSCHN2 (2.0 M in Et2O, 0.74 mL, 1.48 mmol, 2.0 equiv) was added dropwise at 25 °C. The resulting yellow solution was stirred at 25 °C for 1 h. Upon completion, AcOH was added dropwise until the yellow color dissipated and gas evolution ceased, indicating complete consumption of unreacted TMSCHN2. The mixture was then diluted with EtOAc (30 mL) and washed with H2O (30 mL). The aqueous layer was further extracted with EtOAc (2 x 30 mL). The combined organic layers were dried (MgSO4), filtered, and concentrated to give the crude diol as a yellow oil. The crude product was purified by flash column chromatography (silica gel, hexanes:EtOAc, 1:4) to give 35 (0.200 g, 62 % over two

1 steps) as a yellow oil. 35: H NMR (300 MHz, CDCl3) δ 7.51 (d, J = 15.8 Hz, 1 H), 6.77 (d, J =

8.3 Hz, 1 H), 6.73–6.59 (m, 3 H), 6.16 (d, J = 15.8 Hz, 1 H), 3.98 (br s, 1 H), 3.90 (br s, 1 H),

3.86 (s, 6 H), 3.81 (s, 3 H), 3.77 (s, 3 H), 3.65–3.53 (m, 1 H), 3.52–3.39 (m, 1 H), 2.99–2.87 (m,

1 H), 2.70 (dd, J = 6.9, 2.7 Hz, 1 H).

Model Diketone (36). To a solution of 35 (0.066 g, 0.16 mmol, 1.0 equiv) in moist CH2Cl2 (2.0 mL) was added solid NaHCO3 (0.100 g, 1.23 mmol, 7.7 equiv). The suspension was cooled to 0

°C and Dess–Martin periodinane (0.290 g, 0.68 mmol, 4.3 equiv) was then added. The resulting

yellow suspension was then stirred for 2 h under ambient atmosphere. Upon completion, 78

saturated aqueous Na2SO3 (2.0 mL) was added and the mixture was stirred for 15 min. The layers were then separated and the aqueous layer was extracted with EtOAc (3 x 5 mL). The combined organic layers were then washed successively with saturated aqueous NaHCO3 (5 mL)

and brine (5 mL). The organic layers were then dried (MgSO4), filtered, and concentrated to give the crude diketone as a yellow oil. The crude material was purified by flash column chromatography (silica gel, hexanes:EtOAc, 1:1) to give diketone 36 (0.023 g, 35 %) as a yellow

1 foam. 36: H NMR (400 MHz, CDCl3) δ 7.47 (d, J = 15.8 Hz, 1 H), 6.83 (d, J = 8.3 Hz, 1 H),

6.77–6.61 (m, 3 H), 6.26 (d, J = 15.9 Hz, 1 H), 4.13 (br s, 1 H), 3.96–3.74 (m, 14 H), 3.18 (dd, J

= 6.7, 2.7 Hz, 1 H).

MeO2C MeO

MeO OTBS OH CO2Me 89

TBS-Protected Alcohol (89). To a solution of diol 35 (0.280 g, 0.67 mmol, 1.0 equiv) in CH2Cl2

(5.0 mL) at -78 °C were added sequentially 2,6-lutidine (0.39 mL, 3.35 mmol, 5.0 equiv) and

TBSOTf (0.20 mL, 0.87 mmol, 1.3 equiv). The solution was stirred for 2.5 h during which it was

allowed to gradually warm to 25 °C. Upon completion, the reaction mixture was diluted with

EtOAc (15 mL) and poured into H2O (15 mL). The layers were separated and the aqueous layer was re-extracted with EtOAc (2 x 15 mL). The combined organic layers were then washed with saturated aqueous NaHCO3 and then dried (MgSO4), filtered, and concentrated to give the crude product as a clear colorless oil. The crude material was then purified by flash column chromatography (Et3N-deactivated silica gel, hexanes:EtOAc, 1:1) to give the desired silyl ether

1 (89) (0.256 g, 72 %) as a clear colorless oil. 89: H NMR (300 MHz, CDCl3) δ 7.48 (d, J = 15.8 79

Hz, 1 H), 6.85–6.52 (m, 4 H), 6.07 (d, J = 15.8 Hz, 1 H), 3.87 (br s, 1 H), 3.85 (s, 6 H), 3.80 (s, 3

H), 3.76 (s, 3 H), 3.55 (br s, 1 H), 3.30 (d, J = 2.4 Hz, 1 H), 2.85 (ddd, J = 6.8, 3.3, 1.9 Hz, 1 H),

2.64 (d, J = 6.9, 2.6 Hz, 1 H), 2.12 (d, J = 5.6 Hz, 1 H), 0.86 (s, 9 H), 0.11 (s, 6 H).

Siloxyketone (37). To a solution of alcohol 89 (0.256 g, 0.48 mmol, 1.0 equiv) in moist CH2Cl2

(5.0 mL) were added sequentially solid NaHCO3 (0.200 g, 2.40 mmol, 5.0 equiv) and Dess–

Martin periodinane (0.398 g, 0.94 mmol, 1.96 equiv). The resulting yellow suspension was

stirred at 25 °C under ambient atmosphere for 1.5 h. Upon completion, saturated aqueous

Na2SO3 (4.0 mL) was added and the mixture was stirred for 15 min. The mixture was then

diluted with CH2Cl2 (15 mL) and washed successively with saturated aqueous NaHCO3 (15 mL) and brine (15 mL). The organic layer was dried (MgSO4), filtered, and concentrated to give the

desired product (crude yield: 250 mg, 98 %) as a clear colorless oil, which was sufficiently pure

1 to carry forward without further purification. 37: H NMR (300 MHz, CDCl3) δ 7.50 (d, J = 15.9

Hz, 1 H), 6.99 (d, J = 7.7 Hz, 1 H), 6.78 (d, J = 8.2 Hz, 1 H), 6.67 (m, 2 H), 6.53 (d, J = 7.1 Hz,

1 H), 6.18 (d, J = 15.8 Hz, 1 H), 4.06 (d, J = 2.4 Hz, 1 H), 3.86 (s, 6 H), 3.82 (s, 3 H), 3.77 (s, 3

H), 3.73 (br s, 1 H), 3.59 (d, J = 2.4 Hz, 1 H), 3.41 (dd, J = 6.7, 1.5 Hz, 1 H), 2.81 (dd, J = 7.6,

2.5 Hz, 1 H), 0.88 (s, 9 H), 0.14 (s, 3 H), 0.12 (s, 3 H).

MeO2C MeO

MeO O OH

CO2Me 90

Hydroxyketone (90). In a plastic vial, 37 (0.040 g, 0.075 mmol, 1.0 equiv) was dissolved in

MeCN (1.0 mL) and cooled to 0 °C. HF•py (0.22 mL, 2.40 mmol, 32 equiv) was added 80

dropwise, and the solution was stirred for 3 h. Upon completion, saturated aqueous NaHCO3 (5.0 mL) was carefully added to the mixture, which was then extracted with EtOAc (3 x 5 mL). The combined organic layers were then dried (MgSO4), filtered, and concentrated to give the crude product as a clear colorless oil. The crude material was purified by flash column chromatography

(silica gel, hexanes:EtOAc, 100:0à40:60) to give hydroxyketone 90 (0.009 g, 29 %) as a clear

1 colorless oil. 90: H NMR (400 MHz, CDCl3) δ 7.49 (d, J = 15.8 Hz, 1 H), 6.80 (d, J = 8.1 Hz, 1

H), 6.68 (m, 2 H), 6.55 (d, J = 6.3 Hz, 1 H), 6.28 (d, J = 15.8 Hz, 1 H), 4.14 (d, J = 2.4 Hz, 1 H),

3.98–3.66 (m, 14 H), 3.51 (dd, J = 6.8, 1.5 Hz, 1 H), 2.85 (dd, J = 7.8, 2.4 Hz, 1 H), 2.66 (br s, 1

H).

p-Bromobenzoate (38). To a solution of hydroxyketone 90 (0.008 g, 0.019 mmol, 1.0 equiv) in

CH2Cl2/pyridine (3:2, 0.5 mL) was added p-bromobenzoyl chloride (0.021 g, 0.096 mmol, 5.0

equiv) at 0 °C. The resulting mixture was stirred for 2 h. The reaction mixture was diluted with

CH2Cl2 (5 mL) and washed with saturated aqueous NH4Cl. The aqueous layer was re-extracted

with CH2Cl2 (2 x 5 mL). The combined organic layers washed with brine and then dried

(MgSO4), filtered, and concentrated to give a white solid. This crude material was purified by preparative TLC (silica gel, hexanes:EtOAc, 3:2) to give the desired p-bromobenzoate derivative

38 (0.007 g, 60 %) contaminated with p-bromobenzoic acid as a white solid. 38 was crystallized

1 from n-hexane:CH2Cl2 (~2:1) via liquid diffusion. 38: H NMR (400 MHz, CDCl3) δ 7.77 (d, J =

8.6 Hz, 2 H), 7.55 (d, J = 8.6 Hz, 2 H), 7.51 (d, J = 15.9 Hz, 1 H), 6.78 (d, J = 8.2 Hz, 1 H),

6.68–6.64 (m, 3 H), 6.11 (d, J = 15.9 Hz, 1 H), 5.54 (br s, 1 H), 3.98–3.73 (m, 14 H), 3.58 (dd, J

= 6.7, 1.5 Hz, 1 H), 2.90 (dd, J = 7.6, 2.5 Hz, 1 H).

81

Model Diels–Alder Product (16). Solid Pb(OAc)4 (95 %, 1.20 g, 2.61 mmol, 1.1 equiv) was suspended in toluene (5 mL) and concentrated directly to remove residual AcOH. The resulting light brown solid was then dissolved in CH2Cl2 (33 mL) and 3,4-dimethoxycinnamic acid (14)

(3.45 g, 16.6 mmol, 7.0 equiv) was added at 25 °C. The resulting orange suspension was stirred vigorously at 25 °C for 30 min, and then the suspension was concentrated directly. Upon complete removal of the solvent, the flask was backfilled with argon by attaching a balloon to the air inlet of the rotary evaporator. The resultant orange solid was then suspended in toluene

(15 mL) and concentrated directly to remove AcOH resulting from ligand exchange. This azeotroping procedure was then repeated. The subsequent orange solid was then re-suspended in

CH2Cl2 (33 mL) and the above procedure (i.e. stirring for 30 min, concentration, and azeotroping) was repeated. After the final azeotrope, the orange solid was dissolved in anhydrous 1,4-dioxane (33 mL) and the resulting deep red solution was stirred for 30 min at 25

°C. A solution of methyl ferulate (3) (0.493 g, 2.37 mmol, 1.0 equiv) in 1,4-dioxane (3 mL) was then added via syringe, resulting in rapid disappearance of the red solution color. After stirring

for 2 min at 25 °C, ethylene glycol (0.160 mL) was added, resulting in the formation of a thick

slurry which was stirred vigorously for 20 min at 25 °C. Upon completion, the reaction contents

were concentrated directly, and the resulting off-white solid was suspended in EtOAc and

filtered. The solid was then rinsed with EtOAc (3 × 15 mL) and filtered. The combined filtrates

were then washed with saturated aqueous NaHCO3 (3 × 50 mL) and brine (50 mL), dried

(MgSO4), filtered, and concentrated to give a yellow oil. This crude residue was purified by flash column chromatography (silica gel, hexanes:EtOAc, 7:3→2:3) to give the desired Diels–

Alder adduct 16 (0.675 g, 69% yield) as a pale yellow foam along with the undersired Wessely oxidation product 41 (0.074 g, 12% yield) as a yellow oil. [Note: the excess carboxylic acid 82

could be recovered from the aqueous layer by acidifying to pH 2 and extracting twice with

EtOAc]. 16: Rf = 0.44 (hexanes:EtOAc, 1:1); IR (film) νmax 2951, 2839, 1792, 1747, 1710,

–1 1 1633, 1515, 1314, 1193, 1024, 916 cm ; H NMR (500 MHz, CDCl3) δ 7.40 (d, 1 H, J =15.5

Hz), 6.72 (d, J = 8.5 Hz, 1 H), 6.49 (s, 1 H), 6.37 (m, 2 H), 6.23 (d, J = 16 Hz, 1 H), 4.13 (d, J =

5 Hz, 1 H), 3.82–3.73 (m, 10 H), 3.71–3.61 (m, 4 H), 3.27 (d, J = 5 Hz, 1 H); 13C NMR (125

MHz, CDCl3) δ 195.3, 172.9, 166.6, 149.0, 148.7, 140.4, 137.1, 133.5, 131.0, 119.6, 119.3,

111.2 (2 C), 99.5, 56.3, 55.9 (2 C), 53.9, 52.0, 47.4, 44.9, 43.9; HRMS (FAB) calcd for

+ + C22H22O8 [M] 414.1315, found 414.1302.

41: Rf = 0.65 (hexanes:EtOAc, 1:1) ; IR (film) νmax 2953, 1707, 1691, 1436, 1248, 1168, 1104,

–1 1 1087, 1014, 977, 822 cm ; H NMR (500 MHz, CDCl3) δ 7.36 (d, J = 16.5 Hz, 1 H), 7.17 (d, J

= 10 Hz, 1 H), 6.41 (s, 1 H), 6.28 (d, J = 10 Hz, 1 H), 6.24 (d, J = 16 Hz, 1 H), 3.81 (s, 3 H),

13 3.50 (s, 3 H), 2.14 (s, 3 H); C NMR (125 MHz, CDCl3) δ 190.6, 169.6, 166.4, 141.3, 137.1,

136.2, 132.7, 126.9, 120.6, 92.7, 51.9, 51.6, 20.4; HRMS could not be obtained (product not sufficiently stable). A [M + H]+ peak of 267.11 was observed via LRMS (FAB).

Alcohols 51 and 52. Diels–Alder adduct 16 (0.630 g, 1.52 mmol, 1.0 equiv) was dissolved in

THF/AcOH (1:1, 14 mL), NaBH(OAc)3 (1.61 g, 7.61 mmol, 5.0 equiv) was added at 25 °C, and the resulting clear yellow solution was stirred for 3 h at 25 °C. Upon completion, an aqueous solution of Rochelle’s salt (0.5 M; 38 mL) was then added and the mixture was stirred for an

additional 20 min at 25 °C, after which time it was carefully poured into saturated aqueous

NaHCO3 (50 mL; CAUTION: vigorous bubbling was observed) and extracted with EtOAc (50

mL). The organic layer was then washed with saturated aqueous NaHCO3 (50 mL) and the combined aqueous layers were re-extracted with EtOAc (2 × 50 mL). The combined organic 83

layers were then dried (MgSO4), filtered, and concentrated to give a clear, colorless oil. This crude residue was purified by flash column chromatography (silica gel, CH2Cl2:Et2O, 19:1→9:1) to give exo alcohol 51 (0.228 g, 36% yield) as a white solid and endo alcohol 52 (0.221 g, 35% yield) as a cloudy, colorless oil. 51: Rf = 0.34 (hexanes:EtOAc, 3:7); IR (film) νmax 3505, 2949,

–1 1 1777, 1713, 1634, 1518, 1351, 1256, 1197, 1025, 929 cm ; H NMR (500 MHz, CDCl3) δ 7.44

(d, J = 16.0 Hz, 1 H), 6.70 (d, J = 8.0 Hz, 1 H), 6.54 (d, J = 6.5 Hz, 1 H), 6.47 (d, J = 2.5 Hz, 1

H), 6.38 (dd, J = 8.5 Hz, 2.0 Hz, 1 H), 6.08 (d, J = 16 Hz, 1 H), 4.11 (app t, J = 3.5 Hz, 1 H),

3.89 (d, J = 4.5 Hz, 1 H), 3.81 (s, 3 H), 3.79 (s, 3 H), 3.77 (s, 3 H), 3.58 (s, 3 H), 3.45 (app dt,

7.0 Hz, 3.0 Hz, 1 H), 3.36 (br s, 1 H), 2.90 (dd, J = 4.5 Hz, 2.0 Hz, 1 H), 2.51 (d, J = 4.0 Hz, 1

13 H); C NMR (125 MHz, CDCl3) δ 174.0, 167.0, 148.8, 148.2, 141.5, 139.8, 133.3, 132.9, 119.5,

117.0, 111.2, 111.1, 109.8, 74.2, 55.9, 52.9, 51.9, 47.8, 47.3, 44.2, 40.4; HRMS (FAB) calcd for

+ + C22H24O8 [M] 416.1471, found 416.146.

52: Rf = 0.34 (hexanes:EtOAc, 3:7); IR (film) νmax 3477, 3059, 2993, 2951, 2844, 1780, 1716,

–1 1 1633, 1605, 1518, 1255, 1203, 1175, 1026, 948, 735 cm ; H NMR (500 MHz, CDCl3) δ 7.38

(d, J = 16.0 Hz, 1 H), 6.69 (d, J = 10.0 Hz, 1 H), 6.51–6.49 (m, 2 H), 6.36 (dd, J = 8.0 Hz, 2.0

Hz, 1 H), 6.09 (d, J = 16 Hz, 1 H), 3.86 (dd, J = 5.0 Hz, 2.0 Hz, 1 H), 3.80 (s, 3 H), 3.78 (s, 3

H), 3.77 (s, 3 H), 3.68 (app t, J = 2.5 Hz, 1 H), 3.62 (d, J = 2.0 Hz, 1 H), 3.47 (s, 3 H), 3.15 (app dt, J = 7.0 Hz, 3.0 Hz, 1 H), 2.98 (dd, J = 5.0 Hz, 2.0 Hz, 1 H), 2.92 (d, J = 5.5 Hz, 1 H); 13C

NMR (125 MHz, CDCl3) δ 173.9, 167.1, 148.8, 148.1, 141.4, 139.2, 134.3, 133.6, 119.6, 117.8,

111.5, 111.0, 107.7, 72.3, 55.9, 52.1, 51.9, 48.1, 47.5, 42.7, 39.8; HRMS (FAB) calcd for

+ + C22H24O8 [M] 416.1471, found 416.1454.

84

Model Diketone 27. Diels–Alder adduct 16 (0.240 g, 0.58 mmol, 1.0 equiv) was dissolved in

CH2Cl2 (2.5 mL), H2O (1.25 mL) and TFA (1.25 mL) were added sequentially at 25 °C, and the resulting mixture was stirred vigorously at 25 °C for 16 h. Upon completion, the reaction mixture was concentrated directly to give a yellow oil, from which residual TFA was removed by further co-evaporations with toluene (2 × 5 mL). The resultant crude yellow oil was purified by flash column chromatography (silica gel, CH2Cl2:MeOH, 97:3→19:1) to give the desired diketone 27 (0.170 g, 73% yield) as a yellow foam. 27: Rf = 0.61 (CH2Cl2:MeOH, 9:1); IR

(film) νmax 3385, 3056, 3002, 2954, 2831, 1793, 1740, 1717, 1634, 1518, 1258, 1197, 1145, 1026

–1 1 cm ; H NMR (500 MHz, CDCl3) δ 7.32 (d, J = 16.0 Hz, 1 H), 6.66 (d, J = 8.5 Hz, 1 H), 6.44

(br s, 1 H), 6.35–6.31 (m, 2 H), 6.17 (d, J = 15.5 Hz, 1 H), 4.01 (dd, J = 4.0 Hz, 1.5 Hz, 1 H),

3.74 (s, 3 H), 3.72 (s, 3 H), 3.71 (s, 3 H), 3.74–3.68 (m, 2 H), 3.19 (br s, 1 H); 13C NMR (125

MHz, CDCl3) δ 166.8, 149.1, 148.8, 139.8, 138.4, 133.4, 120.1, 119.5, 111.3, 56.0, 55.4, 52.1,

+ + 48.6, 44.7, 40.1; HRMS (FAB) calcd for C21H20O8 [M] 400.1158, found 400.1190.

Luche-Type Reduction (Representative Procedure). A solution of anhydrous CeCl3 (0.061 g,

0.19 mmol, 1.3 equiv) in MeOH (1.0 mL) was cooled to -78 °C and NaBH4 (0.007g, 0.19 mmol,

1.3 equiv) was added. The resulting solution was stirred at -78 °C for 30 min and a solution of 16

(0.061 g, 0.25 mmol, 1.0 equiv) in CH2Cl2 (0.5 mL) was added rapidly. A few drops of AcOH were added to then added to the solution without delay and the flask was immersed in a warm water bath to melt the just-added AcOH. Upon warming to 25 °C, the mixture was poured into brine (5 mL) and extracted with EtOAc (3 x 5 mL). The combined organic layers were then dried

(MgSO4), filtered, and concentrated. The residue was purified by flash column chromatography 85

(silica gel, CH2Cl2:Et2O, 9:1) to give alcohols 51 and 52 (47 mg, 77 % combined yield, 51:52 =

1.8:1).

Model Hydroxyketone (53). Reduction product 51 (0.200 g, 0.49 mmol, 1.0 equiv) was

suspended in CH2Cl2 (2.5 mL). H2O (1.25 mL) and TFA (1.25 mL) were then added sequentially

at 25 °C, and the resulting colorless, cloudy mixture was stirred vigorously at 25 °C for 3 h.

Upon completion, the mixture was concentrated directly, from which residual TFA was removed by further co-evaporations with toluene (2 × 5 mL). The resultant crude white solid was carried forward without any further purification. 53: Rf = 0.19 (CH2Cl2:MeOH, 9:1); IR (film) νmax

3395, 2948, 2913, 2841, 1713, 1630, 1590, 1517, 1314, 1255, 1144, 1025 cm–1; 1H NMR (500

MHz, CD3OD), 7.44 (d, J = 16.0 Hz, 1 H), 6.89–6.79 (m, 3 H), 6.74 (dd, J = 8.5 Hz, 2.0 Hz, 1

H), 6.22 (d, J = 15.5 Hz, 1 H), 4.09 (d, J = 2.5 Hz, 1 H), 3.83–3.78 (m, 10 H), 3.67 (d, J = 4.0

Hz, 1 H), 3.26 (app dt, J = 4.0 Hz, 2.0 Hz, 1 H), 2.82 (dd, J = 5.5 Hz, 2.5 Hz, 1 H); 13C NMR

(125 MHz, CD3OD) δ 177.3, 171.6, 167.5, 149.0, 148.0, 141.1, 140.3, 136.8, 136.2, 119.3,

117.1, 111.7 (2 C), 71.4, 55.1, 55.0, 51.1, 50.9, 50.8, 46.5, 31.3; HRMS (FAB) calcd for

+ + C21H22O8Na [M + Na] 425.1212, found 425.1213.

Protected rufescenolide (57). Hydroxyketal 51 (0.075 g, 0.18 mmol, 1.0 equiv) was dissolved in CH2Cl2 (9.0 mL) (the mixture had to be stirred for 20 min for the starting material to completely dissolve). To this mixture was added Et3SiH (0.57 mL, 3.60 mmol, 20.0 equiv) and

the clear solution was stirred for 5 min at 25 °C. TMSOTf (0.10 mL, 0.54 mmol, 3.0 equiv) was

then added and the resulting yellow mixture was stirred for 2.5 h at 25 °C. Three drops of

pyridine were then added and the color disappeared and gas evolution was observed. The 86

mixture was poured into 1.0 M HCl (35 mL) and extracted with EtOAc (2 × 35 mL). The

combined organic layers were washed with brine (35 mL) and then dried (MgSO4), filtered, and

concentrated to give a waxy, white solid. The crude product was purified by flash column

chromatography (silica gel, CH2Cl2:Et2O, 2:3) to give the desired product 57 (0.037 g, 54 % yield) alongside C2-symmetric dimer 56 (0.017 g, 25 % yield). 57: Rf = 0.28 (hexanes:EtOAc,

- 3:7); IR (film) νmax 3474, 2952, 2914, 1774, 1717, 1625, 1515, 1430, 1252, 1160, 1027, 983 cm

1 1 ; H NMR (400 MHz, CDCl3) δ 7.47 (d, J = 16.0 Hz, 1 H), 6.75 (d, J = 8.4 Hz, 1 H), 6.54 (d, J

= 6.4 Hz, 1 H), 6.51 (d, J = 2.0 Hz, 1 H), 6.43 (dd, J = 8.4, 2.0 Hz, 1 H), 6.18 (d, J = 16.0 Hz, 1

H), 4.34 (d, J = 4.8 Hz, 1 H), 4.15 (d, J = 3.6 Hz, 1 H), 3.97 (app t, J = 4.8 Hz, 1 H), 3.86 (s, 3

H), 3.84 (s, 3 H), 3.82 (s, 3 H), 3.44–3.40 (m, 1 H), 3.29 (br s, 1 H), 2.75 (d, J = 4.0 Hz, 1 H);

13 C NMR (125 MHz, CDCl3) δ 177.7, 167.1, 148.8, 148.2, 141.3, 138.5, 134.7, 133.3, 119.6,

117.5, 111.2, 111.0, 82.2, 73.7, 55.9 (2 C), 51.9, 46.7, 44.4, 44.2, 39.3; HRMS (FAB) calcd for

+ + C21H22O7 [M] 386.1366, found 386.1370.

56: Rf = 0.44 (hexanes:EtOAc, 3:7); IR (film) νmax 2956, 2838, 1784, 1715, 1635, 1516, 1437,

–1 1 1314, 1254, 1188, 1026, 912, 734 cm ; H NMR (400 MHz, CDCl3) δ 7.29 (d, J = 16.0 Hz, 1

H), 6.74 (d, J = 8.4 Hz, 1 H), 6.49 (d, J = 2.0 Hz, 1 H), 6.41 (dd, J = 8.4, 2.0 Hz, 1 H), 6.37 (d, J

= 6.4 Hz, 1 H), 6.01 (d, J = 15.6 Hz, 1 H), 4.29 (d, J = 3.6 Hz, 1 H), 3.86 (s, 3 H), 3.83 (s, 6 H),

3.62 (d, J = 4.4 Hz, 1 H), 3.49 (app quintet, J = 3.2 Hz, 1 H), 3.39 (br s, 1 H), 2.91 (dd, J = 4.4,

13 2.0 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 173.1, 166.8, 148.9, 148.4, 140.8, 136.4, 133.1,

132.5, 119.7, 117.7, 111.2, 111.0, 106.6, 74.2, 55.9 (3 C), 51.9, 47.1, 44.3, 43.6; HRMS (FAB)

+ + calcd for C42H40O14 [M] 768.2418, found 768.2448.

87

Rufescenolide (58). To a solution of 57 (0.012 g, 0.031 mmol, 1.0 equiv) in CH2Cl2 (1.0 mL) was added BBr3 (1.0 M in CH2Cl2; 0.19 mL, 0.19 mmol, 6.0 equiv) at 0 °C. The dark yellow

solution was stirred at this temperature for 10 min. Upon completion, saturated aqueous NaHCO3

(1.0 mL) was added, and the mixture was poured into water and then extracted with EtOAc (2 ×

5 mL). The combined organic layers were washed with brine (5 mL) and dried (MgSO4), filtered, and concentrated to give a clear colorless oil. The crude product was purified by preparative thin layer chromatography (hexanes:EtOAc, 3:7) to give rufescenolide (0.006 g, 55 % yield) as an

amorphous white solid. 58: Rf = 0.56 (EtOAc); IR (film) νmax 3354, 1766, 1696, 1630, 1518,

–1 1 1438, 1271, 1172, 1116, 1074, 986 cm ; H NMR (400 MHz, CD3OD) δ 7.45 (d, J = 16.0 Hz, 1

H), 6.64 (d, J = 8.0 Hz, 1 H), 6.52 (d, J = 6.5 Hz, 1 H), 6.43 (d, J = 2.0 Hz, 1 H), 6.34 (dd, J =

8.4, 2.4 Hz, 1 H), 6.27 (d, J = 16.0 Hz, 1 H), 4.24 (d, J = 5.2 Hz, 1 H), 4.11 (app t, J = 4.5 Hz, 1

H), 4.01 (d, J = 3.6 Hz, 1 H), 3.77 (s, 3 H), 3.35 (m, 1 H), 3.15 (br s), 2.66 (d, J = 4.5 Hz, 1

13 H); C NMR (125 MHz, CD3OD) δ 179.4, 167.9, 144.7, 143.9, 142.1, 140.0, 133.9, 133.2,

118.8, 116.1, 114.7, 114.6, 82.7, 73.5, 50.8, 46.3, 44.2 (2 C), 38.8; HRMS (FAB) calcd for

+ + C19H19O7 [M + H] 359.1131, found 359.1125. All spectroscopic data matched that reported by

David et al.2

Chiral Alcohol (67). Rosmarinic acid (5.00 g, 13.9 mmol, 1.0 equiv) was dissolved in a mixture of THF and MeOH (10:1, 110 mL) at 25 °C and the resultant solution was cooled to –78 °C. To this solution was added TMSCHN2 (2.0 M in Et2O, 6.6 mL total, 0.95 equiv) in 0.55 mL portions

every 5 min until the addition was complete. After the final addition had been achieved, the low-

temperature bath was removed and the reaction mixture was allowed to warm to 25 °C and the

2 do Vale, A. E.; David, J. M.; dos Santos, E. O.; David, J. P.; de Silva, L. C. R. C.; Bahia, M. V.; Brandao, H. N. Phytochemistry 2012, 76, 158. 88

brown solution was stirred at 25 °C for an additional 1 h. Upon completion, the reaction mixture

was quenched by the addition of glacial AcOH (~3 drops; no bubbling was observed, indicating

complete consumption of the TMSCHN2) and concentrated directly. The resulting brown oil was filtered through a pad of silica gel, eluting with CH2Cl2:MeOH (9:1), and the filtrate was

concentrated directly. The resulting brown foam was then dissolved in DMF (70 mL) and solid

K2CO3 (11.5 g, 83.4 mmol, 6.0 equiv) was added followed by benzyl bromide (9.9 mL, 83.4 mmol, 6.0 equiv) at 25 °C. The resultant brown suspension was then heated to 55 °C and stirred vigorously at that temperature for 16 h. Upon completion, the reaction contents were cooled to

25 °C, poured into 1 M HCl (140 mL), and extracted with Et2O (3 × 140 mL). The combined

organic layers were then washed with 1 M HCl (3 × 140 mL) and brine (140 mL), dried

(MgSO4), filtered, and concentrated to give a yellow oil which was carried forward without further purification. Next, this newly obtained oil was dissolved in a mixture of CH2Cl2 and

MeOH (160 mL, 1:1) and solid NaOMe (750 mg, 13.9 mmol, 1.0 equiv) was added at 25 °C.

The resulting dark yellow solution was stirred at 25 °C for 4.5 h. Upon completion, the reaction mixture was poured into saturated aqueous NH4Cl (600 mL) and extracted with EtOAc (3 × 600 mL). The combined organic layers were then washed with brine (600 mL), dried (MgSO4), filtered, and concentrated. The resultant crude white solid was purified by flash column chromatography (silica gel, hexanes:EtOAc, 7:3→3:2) to give α,β-unsaturated ester 66 (4.93 g,

95% yield over three steps) as a white solid along with the desired chiral alcohol 67 (4.41 g, 81%

22 yield over three steps) as a white crystalline solid. 67: Rf = 0.60 (hexanes:EtOAc, 1:1); [α] D =

+7.8° (c = 0.5, CHCl3); IR (film) νmax 3480, 3066, 3032, 2926, 1733, 1606, 1589, 1512, 1459,

–1 1 1427, 1265, 1218, 1137, 1091, 1020, 736, 696 cm ; H NMR (400 MHz, CDCl3) δ 7.46–7.31

(m, 10 H), 6.89 (d, J = 8.0 Hz, 1 H), 6.75 (d, J = 2.0 H, 1 H), 6.74 (dd, J = 8.4, 2.0 Hz, 1 H), 89

5.17 (s, 2H), 5.16 (s, 2 H), 4.42 (ddd, J = 6.4, 4.4, 2.0 Hz, 1 H), 3.74 (s, 3H), 3.05 (dd, J = 14.0,

4.4 Hz, 1 H), 2.89 (dd, J = 14, 6.4 Hz, 1 H), 2.63 (d, J = 6.4 Hz, 1H); 13C NMR (100 MHz,

CDCl3) δ 174.5, 148.8, 148.1, 137.5, 137.4, 129.7, 128.5, 127.8 (2 C), 127.4 (2 C), 122.5, 116.7,

+ + 115.1, 71.3, 52.4, 40.1; HRMS (FAB) calcd for C24H24O5 [M] 392.1624, found 392.1624.

Chiral Allyl Ester (70). Chiral alcohol 67 (3.99 g, 10.2 mmol, 1.0 equiv) was dissolved in

THF/MeOH/H2O (3:1:1, 50 mL) at 25 °C, and the resultant mixture was cooled to 0 °C and solid

LiOH (0.600 g, 14.3 mmol, 1.4 equiv) was added. The ice bath was then removed and the mixture was heated to 45 °C and stirred under an ambient atmosphere for 6 h. Upon completion, the reaction mixture was cooled to 25 °C, poured into 1 M HCl (160 mL), and extracted with

EtOAc (3 x 150 mL). The combined organic layers were then dried (MgSO4), filtered, and

concentrated. The resultant crude white solid was dissolved in DMF (40 mL), and K2CO3 (2.20 g, 15.9 mmol, 1.5 equiv) and allyl bromide (1.40 mL, 15.87 mmol, 1.5 equiv) were added sequentially at 25 °C. The reaction mixture was then stirred at 25 °C for 2 h before being poured into 1 M HCl (160 mL) and extracted with Et2O (3 × 80 mL). The combined organic layers were then washed with H2O (5 × 80 mL), dried (MgSO4), filtered, and concentrated to give the desired allyl ester 70 (4.08 g) as a white solid of sufficient purity to press forward without further

22 purification. 70: Rf = 0.44 (hexanes:EtOAc, 1:1); [α] D = +16.9° (c = 0.5, CHCl3); IR (film)

νmax 3421, 3034, 2945, 1726, 1592, 1519, 1454, 1429, 1386, 1270, 1251, 1241, 1229, 1168,

–1 1 1140, 1024 cm ; H NMR (400 MHz, CDCl3) δ 7.40–7.32 (m, 10 H), 6.90 (d, J = 8.0 Hz, 1 H),

6.87 (d, J = 2.0 Hz, 1 H), 6.76 (dd, J = 8.0, 1.6 Hz, 1 H), 5.96–5.86 (m, 1 H), 5.35 (dd, J = 17.2,

1.2 Hz, 1 H), 5.17 (s, 2 H), 5.16 (s, 2 H), 4.68–4.59 (m, 2 H), 4.44 (dd, J = 10.4, 5.6 Hz, 1 H),

3.07 (dd, J = 14.0, 4.4 Hz, 1 H), 2.95 (dd, J = 14.0, 6.8 Hz, 1 H), 2.69 (d, J = 6.0 Hz, 1 H); 13C 90

(100 MHz, CDCl3) δ 173.8, 148.8, 148.1, 137.5, 137.4, 131.5, 129.6, 128.5 (2 C), 127.8 (2 C),

127.4, 127.3, 122.6, 119.2, 118.7, 115.2, 71.4, 71.3, 66.2, 40.0; HRMS (FAB) calcd for

+ + C26H26O5 [M] 418.1780, found 418.1793.

Benzyl-Protected Dienophile (69). Esters 66 and 68 were subjected to the procedure used for the hydrolysis of 67 to give the dienophile 69 in 65–80 % yield.

Phenol 72. Carboxylic acid 71 was prepared according to the literature (S.A. Snyder; F. Kontes,

J. Am. Chem. Soc. 2009, 131, 1745). To a solution of the crude alcohol 70 (4.08 g, 9.76 mmol,

1.0 equiv) and carboxylic acid 71 (9.00 g, 29.3 mmol, 3.0 equiv) in CH2Cl2 (55 mL) at 25 °C were sequentially added 4-DMAP (1.91 g, 15.6 mmol, 1.6 equiv) and EDCI (3.74 g, 19.5 mmol,

2.0 equiv). The resulting yellow solution was then stirred at 25 °C for 3 h. Upon completion, the reaction mixture was diluted with EtOAc (165 mL) and washed with 1 M HCl (165 mL) and brine (165 mL). The organic layer was then dried (MgSO4), filtered, and concentrated to give the

desired intermediate as a viscous yellow oil. Pressing forward without any additional

purification, this crude yellow oil was dissolved in THF (60 mL), cooled to 0 °C, and AcOH (0.6

mL) and TBAF (1.0 M in THF, 29.3 mL, 29.3 mmol, 3.0 equiv) were added sequentially. The resulting solution was then stirred at 0 °C for 1 h. Upon completion, the reaction mixture was poured into 1 M HCl (180 mL) and extracted with EtOAc (180 mL). The organic layer was then

washed with brine (180 mL), dried (MgSO4), filtered, and concentrated. The resultant crude

yellow solid was purified by flash column chromatography (silica gel, hexanes:EtOAc, 3:2) to

give phenol 72 (4.97 g, 82% yield over four steps) a clear, pale yellow oil. 72: Rf = 0.48

22 (hexanes:EtOAc, 1:1); [α] D = +17.4° (c = 0.5, CHCl3); IR (film) νmax 3419, 3034, 1750, 1713, 91

1632, 1603, 1591, 1455, 1429, 1380, 1314, 1268, 1154, 1029, 984, 738, 697 cm–1; 1H NMR (400

MHz, CDCl3) δ 7.67 (d, J = 15.6 Hz, 1 H), 7.47–7.30 (m, 10 H), 7.09 (dd, J = 8.4, 2.0 Hz, 1 H),

7.03 (d, J = 2.0 Hz, 1 H), 6.95–6.90 (m, 3 H), 6.83 (dd, J = 8.0, 2.0 Hz, 1 H), 6.33 (d, J = 15.6

Hz, 1 H), 5.93 (s, 1 H), 5.92–5.85 (m, 1 H), 5.39–5.27 (m, 2 H), 5.26 (dd, J = 9.2, 1.2 Hz, 1 H),

13 5.16 (s, 2 H), 5.15 (s, 2 H), 4.65–4.63 (m, 2 H), 3.91 (s, 3 H), 3.19–3.12 (m, 2 H); C (100 MHz,

CDCl3) δ 169.6, 166.4, 149.0, 148.3, 148.2, 146.8, 146.2, 137.4, 137.2, 131.5, 129.3, 128.5,

127.8, 127.8, 127.4, 127.3, 126.8, 123.4, 122.5, 118.8, 116.6, 115.2, 114.8, 114.4, 109.4, 72.9,

+ + 71.5, 71.4, 65.9, 56.0, 37.1; HRMS (FAB) calcd for C36H34O8 [M] 594.2254, found 594.2250.

Diels–Alder Adduct 73. Pb(OAc)4 (3.77 g, 8.10 mmol, 1.1 equiv) was suspended in toluene (5 mL) and concentrated directly to remove any trace/adventitious AcOH that might be present.

The resulting light brown solid was then dissolved in CH2Cl2 (100 mL) and dienophile 69 (18.5 g, 51.4 mmol, 7.0 equiv) was added at 25 °C. The resulting orange suspension was stirred vigorously at 25 °C for 30 min. Upon completion, the suspension was concentrated directly.

The flask was then backfilled with argon by attaching a balloon to the air inlet of the rotary evaporator. The orange solid was then suspended in toluene (50 mL) and concentrated to remove any residual AcOH resulting from ligand exchange. This azeotroping procedure was then repeated. The orange solid was then re-suspended in CH2Cl2 and the above procedure (i.e.

stirring for 30 min, concentration, azeotroping) was repeated. Next, the orange solid was

dissolved in anhydrous 1,4-dioxane (100 mL) and the resulting orange suspension was stirred for

30 min at 25 °C. A solution of phenol 72 (4.37 g, 7.30 mmol, 1.0 equiv) in 1,4-dioxane (10 mL)

was then added via syringe, resulting in the disappearance of the orange color of the original suspension and a significant thickening of the mixture. The reaction flask was swirled manually 92

to ensure efficient mixing, and ethylene glycol (0.5 mL, 8.87 mmol, 1.2 equiv) was then added.

The slurry was stirred vigorously at 25 °C for 20 min, after which time the reaction contents

were concentrated directly. The resulting white solid was then suspended in EtOAc (50 mL) and

filtered. The solid was then thoroughly rinsed with EtOAc (3 × 15 mL) and filtered. The

combined filtrates were then washed with brine (500 mL). The organic layer was then dried

(MgSO4), filtered, and concentrated. The resultant crude yellow solid was then suspended in

Et2O (50 mL) and filtered, with the yellow filtrate concentrated. The resultant crude yellow oil

was purified by flash column chromatography (silica gel, hexanes:EtOAc, 4:1→3:2) to give the

desired Diels–Alder product 73 (3.45 g, 50% yield, ~1:1 mixture of inseparable diastereomers)

3 22 as a clear, colorless oil. 73a : Rf = 0.75 (hexanes:EtOAc, 1:1); [α] D = + 24.7° (c = 0.5, CHCl3);

IR (film) νmax 3486, 3066, 3031, 2926, 2149, 1796, 1750, 1721, 1632, 1592, 1513, 1464, 1429,

–1 1 1380, 1262,1163, 1016, 920, 853, 737 cm ; H NMR (400 MHz, CDCl3) δ 7.44–7.40 (m, 6 H),

7.40–7.30 (m, 15 H), 6.89–6.86 (m, 2 H), 6.83–6.77 (m, 2 H), 6.45 (d, J = 2.4 Hz, 1 H), 6.25 (br d, J = 6.0 Hz, 1 H), 6.19 (d, J = 16.0 Hz, 1 H), 5.85 (ddt, J = 17.2, 10.4, 5.6 Hz, 1 H), 5.34–5.22

(m, 3 H), 5.17–5.06 (m, 8 H), 4.65–4.54 (m, 2 H), 3.93 (dd, J = 5.2, 2.0 Hz, 1 H), 3.68 (s, 3 H),

13 3.68 (m, 1 H), 3.61 (dd, J = 6.8, 2.8 Hz, 1 H), 3.19–3.08 (m, 3 H); C NMR (125 MHz, CDCl3)

δ 195.3, 172.6, 171.1, 169.2, 165.5, 148.9, 148.7, 148.6, 148.3, 141.5, 137.2, 137.1, 136.9 (2 C),

136.8, 134.1, 131.4 (2 C), 128.8, 128.6 (2 C), 128.5, 127.9 (2 C), 127.8, 127.4, 127.3, 127.2,

122.5, 120.6, 118.9, 118.5, 116.5, 115.0, 114.6, 99.1, 73.5, 71.6, 71.3 (2 C), 71.1, 66.0, 60.4,

+ + 56.2, 54.4, 47.4, 44.9, 44.3, 36.9; HRMS (FAB) calcd for C59H52O12 [M] 952.3459, found

952.3472.

3 Diastereomers were separated after reducing the ketone to the alcohol. The separate diastereomers were then oxidized to regenerate the ketone (vide infra). 93

22 73b: Rf = 0.75 (hexanes:EtOAc, 1:1); [α] D = –4.0° (c = 1.0, CHCl3); IR (film) νmax 3486, 3066,

3031, 2926, 2149, 1796, 1750, 1721, 1632, 1592, 1513, 1464, 1429, 1380, 1262,1163, 1016, 920,

–1 22 1 853, 737 cm [α] D = –4.0° (c = 1.0, CHCl3); H NMR (400 MHz, CDCl3) 7.44–7.42 (m, 6 H),

7.39–7.28 (m, 15 H), 6.906.87 (m, 2H), 6.84–6.77 (m, 2H), 6.45 (d, J = 1.6 Hz), 6.38 (dd, J =

8.0, 1.6 Hz, 1 H), 6.26–6.21 (m, 1 H), 6.16 (d, J = 16.0 Hz, 1 H), 5.85 (ddt, J =17.2, 10.4, 5.6

Hz), 5.32–5.23 (m, 3 H), 5.15–5.09 (m, 8 H), 4.67–4.57 (m, 2 H), 3.94 (dd, J = 5.2, 2.4 Hz, 1 H),

3.74 (s, 3 H), 3.71 (m, 1 H), 3.61 (dd, J = 6.8, 2.8 Hz, 1 H), 3.21–3.07 (m, 3 H); 13C NMR (125

MHz, CDCl3) 195.3, 172.6, 169.2, 165.4, 148.9, 148.7, 148.6, 148.3, 141.5, 141.3, 137.2, 137.1,

136.9, 136.8, 134.1, 131.4 (2 C), 128.9, 128.6, 128.5, 128.0, 127.9, 127.8, 127.4, 127.3, 127.2,

122.4, 120.6, 118.8, 118.5, 118.4, 116.5, 115.0 (2 C), 114.6, 99.2, 73.6, 73.5, 71.6, 71.5, 71.3,

71.2, 71.1, 66.0 (2 C), 60.4, 56.2, 56.1, 54.3, 47.4, 44.9, 44.1, 36.9; HRMS (FAB) calcd for

+ + C59H52O12 [M] 952.3459, found 952.3472.

Alcohols (74/75). A mixture of Diels–Alder product diastereomers 73 (2.1 g, 2.20 mmol) was

dissolved in THF/AcOH (1:1, 20 mL), NaBH(OAc)3 (2.3 g, 11.0 mmol, 5.0 equiv) was added at

25 °C, and the resulting pale yellow solution was stirred at 25 °C for 4 h. An aqueous solution of

Rochelle’s salt (0.5 M; 50 mL) was then added and the resultant mixture was stirred for 20 min,

after which time it was carefully poured into saturated aqueous NaHCO3 (100 mL; CAUTION: vigorous bubbling was observed) and extracted with EtOAc (100 mL). The organic layer was washed with saturated aqueous NaHCO3 (100 mL) and the combined aqueous layers were re-

extracted with EtOAc (2 × 100 mL). The combined organic layers were then dried (MgSO4), filtered, and concentrated. The resultant clear, colorless residue was purified by flash column chromatography (silica gel, CH2Cl2:Et2O, 19:1→9:1) to give two pairs of diastereomers (i.e. two 94

exo alcohols and two endo alcohols), each as clear colorless oils (1.72 g total, 82% yield, ~1.5:1

74:75). Each pair of diastereomers was then separated by preparative TLC in 30 mg batches

(CH2Cl2:Et2O, 97:3, 4–5 developments per plate). Exo alcohol 74a: Rf = 0.52 (CH2Cl2:Et2O,

22 9:1); [α] D = +94.3° (c = 0.95, CHCl3); IR (film) νmax 3503, 3063, 3032, 2946, 2863, 1779,

1752, 1717, 1633, 1604, 1513, 1454, 1382, 1264, 1206, 1162, 1023, 931, 737, 697 cm–1; 1H

NMR (500 MHz, CDCl3) δ 7.43–7.28 (m, 21 H), 6.88–6.85 (m, 2 H), 6.80–6.76 (m, 2 H), 6.45

(d, J = 2.5 Hz, 1 H), 6.42–6.39 (m, 2 H), 6.06 (d, J = 16.0 Hz, 1 H), 5.83 (ddt, J = 17.5, 10.5, 5.5

Hz, 1 H), 5.32–5.21 (m, 3 H), 5.14–5.06 (m, 8 H), 4.61 (ddt, J = 12.0, 5.5, 1.0 Hz, 1 H), 4.56

(ddt, J = 13.0, 5.5, 1.5 Hz, 1 H), 4.07 (app t, J = 3.5 Hz, 1 H), 3.78 (d, J = 4.5 Hz, 1 H), 3.52 (s,

3 H), 3.38 (app quintet, J = 3.0 Hz, 1 H), 3.29 (br s, 1 H), 3.15 (dd, J = 14.5, 5.0 Hz, 1 H), 3.10

(dd, J = 14.5, 3.0 Hz, 1 H), 2.74 (dd, J = 4.5, 1.5 Hz, 1 H), 2.34 (d, J = 4.5 Hz, 1 H); 13C NMR

(125 MHz, CDCl3) δ 173.8, 169.3, 165.7, 149.0, 148.5, 148.4, 148.3, 142.7, 140.4, 137.3, 137.2,

137.1, 137.0, 133.8, 132.8, 131.4, 128.9, 128.6, 128.5 (3 C), 127.9 (3 C), 127.8, 127.4, 127.3 (2

C), 127.2, 122.5, 120.6, 118.9, 116.7, 116.2, 115.2 (2 C), 114.8, 109.7, 74.1, 73.3, 71.6, 71.4 (2

+ + C), 71.3, 66.0, 52.9, 47.8, 47.1, 44.2, 40.4, 37.0; HRMS (FAB) calcd for C59H54O12 [M]

954.3615, found 954.3619.

22 Exo alcohol 74b: Rf = 0.52 (CH2Cl2:Et2O, 9:1); [α] D = –92.0° (c = 0.75, CHCl3); IR (film)

νmax 3523, 3032, 2949, 1779, 1717, 1632, 1512, 1261, 1205, 1161, 1140, 1080, 1020, 930, 738,

–1 1 698 cm ; H NMR (500 MHz, CDCl3) δ 7.46–7.30 (m, 21 H), 6.92–6.90 (m, 2 H), 6.84–6.80

(m, 2 H), 6.48 (d, J = 2.5 Hz, 1 H), 6.45–6.41 (m, 2 H), 6.07 (d, J = 16.0 Hz, 1 H), 5.87 (ddt, J =

17.5, 10.5, 6.0 Hz, 1 H), 5.36–5.24 (m, 3 H), 5.17–5.09 (m, 8 H), 4.65–4.59 (m, 2 H), 4.10 (app t, J = 3.5 Hz, 1 H), 3.82 (d, J = 4.5 Hz, 1 H), 3.59 (s, 3 H), 3.41 (app quintet, J = 3.0 Hz, 1 H),

3.32 (br s, 1 H), 3.19 (dd, J = 14.5, 4.5 Hz, 1 H), 3.12 (dd, J = 14.0, 8.5 Hz, 1 H), 2.76 (d, J = 95

13 4.5, 1.5 Hz, 1 H), 2.40 (d, J = 4.5 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 173.9, 169.3, 165.7,

149.0, 148.5, 148.3 (2 C), 142.6, 140.5, 137.3, 137.2, 137.1, 137.0, 133.8, 132.8, 131.4, 129.0,

128.5 (5 C), 127.9, 127.8, 127.4, 127.3 (2 C), 127.2 (2 C), 122.5, 120.6, 118.9, 116.6, 116.2,

115.2 (2 C), 114.8, 109.7, 74.1, 73.4, 71.6, 71.4 (2 C), 71.3, 66.0, 53.0, 47.9, 47.0, 44.2, 40.3,

37.0.

22 Endo alcohol 75a: Rf = 0.41 (CH2Cl2:Et2O, 9:1); [α] D = +79.6° (c = 0.75, CHCl3); IR (film)

νmax 3474, 3062, 3024, 2936, 1780, 1754, 1718, 1633, 1513, 1454, 1262, 1204, 1163, 1139,

–1 1 1021, 736, 697 cm ; H NMR (500 MHz, CDCl3) δ 7.46–7.30 (m, 21 H), 6.91–6.88 (m, 2 H),

6.82–6.79 (m, 2 H), 6.50 (d, J = 2.5 Hz, 1 H), 6.44–6.40 (m, 2 H), 6.09 (d, J = 15.5 Hz, 1 H),

5.86 (ddt, J = 17.5, 10.5, 6.0 Hz, 1 H), 5.35–5.23 (m, 3 H), 5.14–5.07 (m, 8 H), 4.63 (dd, J =

13.0, 5.5 Hz, 1 H), 4.59 (dd, J = 13.0, 5.5 Hz, 1 H), 3.78 (dd, J = 5.0, 1.5 Hz, 1 H), 3.65 (dd, J =

5.0, 2.5 Hz, 1 H), 3.55 (br s, 1 H), 3.45 (s, 3 H), 3.17–3.09 (m, 3 H), 2.85 (dd, J = 4.5, 1.5 Hz, 1

13 H), 2.48 (d, J = 5.5 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 173.5, 169.3, 165.8, 149.0, 148.5,

148.3, 148.2, 142.5, 139.8, 137.3, 137.2, 137.1, 137.0, 134.1, 131.4, 129.0, 128.5 (3 C), 127.9 (2

C), 127.8, 127.4 (2 C), 127.3 (2 C), 127.2, 122.5, 120.8, 118.9, 117.0, 116.6, 115.3, 115.2, 114.8,

107.5, 73.3, 72.1, 71.6, 71.4, 71.3 (2 C), 66.0, 53.4, 52.2, 48.1, 47.3, 42.8, 39.9, 37.0; HRMS

+ + (FAB) calcd for C59H54O12 [M] 954.3615, found 954.3622.

22 Endo alcohol 75b: Rf = 0.41 (CH2Cl2:Et2O, 9:1); [α] D = –64.7° (c = 0.80, CHCl3); IR (film)

νmax 3474, 3062, 3024, 2936, 1780, 1754, 1718, 1633, 1513, 1454, 1262, 1204, 1163, 1139,

–1 1 1021, 736, 697 cm ; H NMR (500 MHz, CDCl3) δ 7.46–7.29 (m, 21 H), 6.92–6.89 (m, 2 H),

6.83–6.79 (m, 2 H), 6.50 (d, J = 2.0 Hz, 1 H), 6.44–6.38 (m, 2 H), 6.07 (d, J = 16.0 Hz, 1 H),

5.88 (ddt, J = 12.5, 8.0, 5.5 Hz, 1 H), 5.34–5.24 (m, 3 H), 5.14–5.10 (m, 8 H), 4.67–4.59 (m, 2

H), 3.78 (dd, J = 4.5, 1.5 Hz, 1 H), 3.63 (br s, 1 H), 3.56 (br s, 1 H), 3.48 (s, 3 H), 3.21–3.07 (m, 96

13 3 H), 2.84 (dd, J = 4.5, 1.5 Hz, 1 H), 2.51 (d, J = 5.0 Hz, 1 H); C NMR (125 MHz, CDCl3)

δ 173.5, 169.3, 165.7, 149.0, 148.5, 148.3, 148.1, 142.4, 139.8, 137.3, 137.1 137.0, 134.1, 131.4,

129.0, 128.5 (5 C), 127.9 (3 C), 127.8, 127.4, 127.3 (3 C), 127.2, 122.5, 120.7, 118.9, 117.1,

116.6, 115.4, 115.2, 114.8, 107.5, 73.4, 72.1, 71.6, 71.4, 71.3, 66.0, 52.2, 48.1, 47.2, 42.7, 39.8,

37.0.

HO2C OH

O O HO OH

HO OH OMe O O 91

Carboxylic Acid 91. Exo alcohol 74b (0.107 g, 0.112 mmol, 1.0 equiv) was dissolved in benzene (1.0 mL), concentrated to remove any trace/adventitious H2O, and then dissolved in

CH2Cl2 (2.0 mL) at 25 °C. The resultant solution was cooled to –78 °C and BCl3 (0.9 mL, 1.0 M

in CH2Cl2, 8.0 equiv) was added in a single portion. The resulting yellow solution was then stirred at –78 °C for 5 min, after which time pH 7 phosphate buffer (0.18 M, 1.0 mL) was added and the flask was immediately immersed in a warm water bath to melt the ice that formed upon addition of the buffer solution. This mixture was then diluted with EtOAc (10 mL) and washed with brine (10 mL). The aqueous layer was then re-extracted with EtOAc (2 × 10 mL), and the combined organic layers were dried (MgSO4), filtered, and concentrated. The resultant crude

pale brown oil was then dissolved in THF (2.0 mL) and the mixture was degassed by bubbling

argon through the solution for 10 min. To this mixture was sequentially added Meldrum’s acid

(24 mg, 0.168 mmol, 1.5 equiv) and Pd(PPh3)4 (6 mg, 0.006 mmol, 0.05 equiv) at 25 °C. The resulting yellow solution was then stirred at 25 °C for 15 min, and, upon completion, was 97

concentrated directly. The resulting yellow foam was purified by preparative TLC

(CH2Cl2:MeOH doped with 1% AcOH, 4:1) to give the desired deprotected compound (62 mg,

55% yield over two steps) as a white solid. 91: Rf = 0.07 (CH2Cl2:MeOH doped with 1% AcOH,

22 4:1); [α] D = –40.9° (c = 1.0, MeOH); IR (film) νmax 3233, 2962, 2925, 2361, 1773, 1702, 1528,

–1 1 1599, 1401, 1268, 1206, 1024, 930, 668 cm ; H NMR (400 MHz, DMSO-d6) δ 7.25 (d, J =

15.6 Hz, 1 H), 6.64 (d, J = 1.6 Hz, 1 H), 6.61–6.57 (m, 2 H), 6.53 (d, J = 6.4 Hz), 6.47 (dd, J =

7.6, 1.6 Hz, 1 H), 6.33 (d, J = 2.0 Hz, 1 H), 6.23 (dd, J = 8.0, 2.0 Hz, 1 H), 6.18 (d, J = 16.0 Hz,

1 H), 4.85 (dd, J = 10.4, 2.4 Hz, 1 H), 4.18 (d, J = 4.0 Hz, 1 H), 3.96 (d, J = 2.8 Hz, 1 H), 3.46

(s, 3 H), 3.22 (br s, 1 H), 3.15 (dd, J = 6.0, 3.2 Hz, 1 H), 3.03 (d, J = 12.4 Hz, 1 H), 2.72 (dd, J =

13 14.4, 10.4 Hz, 1 H), 2.65 (d, J = 4.0 Hz, 1 H); C NMR (125 MHz, DMSO-d6) δ 175.4, 173.4,

166.6, 145.7, 145.5, 144.8, 144.1, 141.8, 140.4, 133.4, 133.2, 130.5, 129.4, 128.7, 120.0, 118.8,

118.0, 117.0, 116.2, 116.1, 115.9, 112.1, 77.0, 73.8, 53.3, 48.8, 47.7, 43.8; HRMS (FAB) calcd

+ + for C28H27O12 [M + H] 555.1503, found 555.1512.

Yunnaneic Acid D (2). Carboxylic acid 91 (0.035 g, 0.063 mmol, 1.0 equiv) was suspended in

CH2Cl2/H2O (2:1, 1.5 mL) and HCl (4.0 M in dioxane, 0.5 mL) was added at 25 °C The mixture was stirred at 25 °C for 2 h and then poured into brine (5 mL) and extracted with EtOAc (3 × 5 mL). The combined organic layers were then dried (MgSO4), filtered, and concentrated.

Residual dioxane was removed by co-evaporation with toluene (2 × 2.0 mL), affording

yunnaneic acid D (2, 0.030 g, 88 % yield) as a white powder. [Note: This compound appears to

isomerize readily in the presence of trace acid. Thus, removal of trace EtOAc from the crude

product by prolonged drying could not be performed without compromising the purity of the

sample]. 2: Rf = N/A; IR (film) νmax 3328, 2529, 2919, 2849, 1710, 1626, 1521, 1446, 1255, 98

–1 1 1175, 1115, 1076, 978, 812 cm ; H NMR (500 MHz, acetone-d6) δ 7.39 (d, J = 16.0 Hz, 1 H),

6.90–6.84 (m, 2 H), 6.76–6.72 (m, 3 H), 6.64 (dd, J = 8.0, 2.0 Hz, 1 H), 6.57–6.53 (m, 1 H), 6.22

(d, J = 16.0 Hz, 1 H), 5.21 (dd, J = 8.5, 3.5 Hz, 1 H), 4.01 (d, J = 1.5 Hz, 1 H), 3.87 (br s, 1 H),

3.55 (br s, 1 H), 3.28 (d, J = 6.0 Hz, 1 H), 3.12 (dd, J = 14.5, 3.5 Hz, 1 H), 2.99 (dd, J = 14.5,

13 8.5 Hz, 1 H), 2.88 (d, J = 2.0 Hz, 1 H); C NMR (125 MHz, acetone-d6) δ 207.8, 175.5, 171.0,

166.5, 145.8, 145.6, 143.3, 141.9, 136.8, 136.3, 129.1, 121.7, 119.9, 117.9, 117.3, 116.1 (2 C),

+ + 115.5, 74.1, 72.4, 51.1, 50.1, 48.8, 47.0, 37.4; HRMS (FAB) calcd for C27H24O12Na [M + Na]

563.1166; HRMS data could not be obtained due to facile side-chain cleavage. A molecular ion

peak of 563.11 was observed via LRMS (FAB). All spectroscopic data matched that reported by

Tanaka et al.4

4 Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Young, T. J. J. Nat. Prod. 1996, 59, 843. 99

Table 2. Comparison of 13C NMR chemical shifts of deprotected hydroxyketone diastereomers derived from exo alcohols 74a (non-natural stereochemistry) and 74b (natural stereochemistry) with those of natural yunnaneic acid D

Natural 74a-derived Absolute Difference 74b-derived Absolute Difference 37.4 36.5 0.9 37.4 0.0 47.0 40.4 6.6 47.0 0.0 48.9 49.4 8.5 48.8 0.1 50.1 44.9 5.2 50.1 0.0 51.0 49.7 1.3 51.1 0.1 72.4 71.4 1.0 72.4 0.0 74.0 73.2 0.8 74.0 0.0 115.5 114.6 0.9 115.5 0.0 116.0 115.1 0.9 116.1 0.1 116.0 115.2 0.8 116.1 0.1 117.3 115.5 1.8 117.3 0.0 117.8 116.5 1.3 117.9 0.1 119.9 119.0 0.9 119.9 0.0 121.7 120.8 0.9 121.7 0.0 129.0 128.1 0.9 129.1 0.1 136.3 135.8 0.5 136.3 0.0 136.7 136.6 0.1 136.7 0.0 142.0 141.1 0.9 141.9 0.1 143.4 143.6 0.2 143.3 0.1 144.9 144.0 0.9 144.9 0.0 145.7 144.7 1.0 145.6 0.1 145.8 144.9 0.9 145.8 0.0 166.6 165.7 0.9 166.6 0.0 171.0 170.2 0.8 171.0 0.0 175.6 174.8 0.8 175.5 0.1 207.9 207.1 0.8 207.8 0.1 Average 1.62 0.04

Diels–Alder Product 73b (as single diastereomer). Procedure adapted from Nielsen, T. E.;

Tanner, D. J. Org. Chem. 2002, 67, 6366. To a stirred suspension of pulverized activated 4 Å molecular sieves (0.035 g, 0.5 g/mmol of alcohol) and NMO (0.017 g, 0.15 mmol, 2.0 equiv) in

CH2Cl2 (1.0 mL) at 25 °C was added endo alcohol 70d (0.070 g, 0.073 mmol, 1.0 equiv) as a

solution in CH2Cl2 (0.5 mL). The resulting slurry was stirred at 25 °C for 10 min, cooled to 0 °C, 100

and TPAP (1 mg, 0.003 mmol, 0.05 equiv) was added in a single portion. The resulting black

slurry was stirred at 0 °C for 1 h. Upon completion, the reaction mixture was filtered through a

plug of silica gel (eluted with hexanes:EtOAc, 7:3) to give 73b (0.045 g, 67% yield) as a pale

yellow oil. See above for characterization data.

AllylO2C OH

O O HO OH

HO O OMe O O 92

Tetraphenol 92. The debenzylation procedure described above for the synthesis of 91 was applied to ketone 73b (0.045 g, 0.031 mmol, 1.0 equiv) to give a crude brown oil following concentration of the organic layers. This crude product was purified by flash column chromatography (silica gel, hexanes:EtOAc, 2:3→3:7) to give tetraphenol 92 (0.022 g, 79%

22 yield) as a yellow oil. 92: Rf = 0.59 (CH2Cl2:MeOH, 9:1); [α] D = –44.5° (c = 0.55, MeOH); IR

–1 (film) νmax 3382, 2945, 2916, 1853, 1739, 1711, 1625, 1599, 1523, 1444, 1365, 1282, 1191 cm ;

1 H NMR (400 MHz, acetone-d6) δ 7.48 (d, J = 16.0 Hz, 1 H), 6.80 (d, J = 2.0 Hz, 1 H), 6.77–

6.72 (m, 2H), 6.69–6.62 (m, 3 H), 6.50–6.42 (m, 2 H), 5.91 (ddt, J = 17.2, 10.4, 5.2 Hz, 1 H),

5.36–5.19 (m, 3 H), 4.66 (dd, J = 5.2, 2.0 Hz, 1 H), 4.65 (dd, J =5.2, 2.0 Hz, 1 H), 4.62 (dd, J =

5.2, 0.8 Hz, 1 H), 3.82 (br s, 1 H), 3.73 (dd, J = 6.8, 2.8 Hz, 1 H), 3.64 (s, 3 H), 3.33 (dd, J = 5.2,

0.8 Hz, 1 H), 3.12 (dd, J = 14.4, 4.8 Hz, 1 H), 3.03 (dd, J = 14.4, 8.4 Hz, 1 H); 13C NMR (125

MHz, acetone-d6) δ 195.4, 173.0, 169.0, 165.5, 145.0, 144.6, 144.1, 141.8, 137.5, 135.1, 132.1,

131.0, 128.3, 127.6, 120.7, 119.2, 118.4, 117.5, 116.4, 115.2, 115.1, 115.0, 99.7, 73.6, 65.2, 56.5, 101

+ + 52.7, 47.4, 44.7, 42.9, 36.5; HRMS (FAB) calcd for C31H29O12 [M+1] 593.1659, found

593.1686.

Yunnaneic Acid C (1). To a degassed solution of tetraphenol 92 (22 mg, 0.037 mmol, 1.0 equiv)

in THF (1.0 mL) at 25 °C was sequentially added Meldrum’s acid (8 mg, 0.056 mmol, 1.5 equiv)

and Pd(PPh3)4 (4 mg, 0.004 mmol, 0.1 equiv). The resultant yellow solution was stirred at 25 °C

for 15 min and then was concentrated directly. The so-obtained residue was then purified by

preparative TLC (CH2Cl2:MeOH doped with 1% AcOH, 4:1) to give a clear, colorless wax. This

newly synthesized material was suspended in CH2Cl2/H2O (2:1, 0.75 mL) and TFA (0.5 mL) was added at 25 °C. Over time, the initially colorless mixture became increasingly yellow, with the mixture being stirred at 25 °C for 16 h. Upon completion, the reaction contents were poured into brine (5 mL) and extracted with EtOAc (3 × 5 mL). The combined organic layers were then dried

(MgSO4), filtered, and concentrated to give yunnaneic acid C (1, 10 mg, 50% over two steps) as a yellow solid. 1: Rf = N/A; IR (film) νmax 3282, 2919, 1705, 1687, 1633, 1521, 1442, 1260,

–1 1 1191, 1148 cm ; H NMR (500 MHz, acetone-d6) δ 7.47 (d, J = 16.0 Hz, 1 H), 6.84–6.83 (m, 2

H), 6.76–6.60 (m, 5 H), 6.43 (d, J = 15.5 Hz, 1 H), 5.23 (dd, J = 9.0, 4.0 Hz, 1 H), 4.34 (br s, 1

H), 3.79 (app t, J = 3.0 Hz, 1 H), 3.74 (dd, J = 6.5, 2.5 Hz, 1 H), 3.32–3.25 (m, 1 H), 3.14 (dd, J

13 = 14.5, 4.0 Hz, 1 H), 3.01 (dd, J = 14.5, 9.0 Hz, 1 H); C NMR (125 MHz, acetone-d6) δ 174.6,

171.3, 166.4, 146.0, 145.8, 145.4, 144.9, 139.4, 136.2, 129.0, 121.5, 120.1, 120.0, 117.4, 116.2,

116.1, 115.9, 74.3, 66.1, 56.9, 54.9, 45.7, 37.4. Spectroscopic data matched that reported by

Tanaka et al.3

102

Table 3. Comparison of 1H NMR chemical shifts of deprotected diketone diastereomers derived from endo alcohols 75a (non-natural stereochemistry) and 75b (natural stereochemistry) with those of natural yunnaneic acid C

Natural 75a-derived Absolute Difference 75b-derived Absolute Difference 7.47 7.47 0.00 7.47 0.00 6.84 6.85 0.01 6.84 0.00 6.82 6.77 0.05 6.83 0.01 6.75 6.75 0.00 6.76 0.01 6.69 6.74 0.05 6.74 0.05 6.66 6.65 0.01 6.66 0.00 6.52 6.63 0.11 6.65 0.13 6.42 6.42 0.00 6.43 0.01 5.26 5.23 0.03 5.23 0.03 4.32 4.35 0.03 4.34 0.02 3.79 3.78 0.01 3.79 0.00 3.74 3.73 0.01 3.74 0.00 3.31 3.31 0.00 3.30 0.01 3.14 3.13 0.01 3.14 0.00 3.03 3.02 0.01 3.01 0.02 Average 0.024 0.019

Quinoxaline Derivative (77). Adapted from T. Tanaka; A. Nishimura; I. Kouno, G.-I. Nonaka;

T.-J. Young, J. Nat. Prod. 1996, 59, 843-849. The following was performed on a diastereomeric mixture. Yunnaneic acid C (1) and its diastereomer (~1:1, 0.027 g total, 0.049 mmol, 1.0 equiv) were dissolved in EtOH/AcOH (4:1, 1.6 mL). The light brown solution was degassed via the freeze-pump-thaw method (liquid N2; repeated twice). 1,2-phenylenediamine (71) (0.011 g, 0.10

mmol, 2.0 equiv) was then added. The resulting brown solution was stirred at 60 °C for 19 h.

Upon completion, the mixture was cooled to 25 °C and concentrated directly. The resulting red-

orange oil was taken up in EtOAc/1 M HCl (1:1, 10 mL). The layers were separated, and the

organic layer was washed with 1 M HCl (2 x 5 mL). The organic layer was then dried (MgSO4), filtered, and concentrated to give the crude quinoxaline as a brown solid (1:1 d.r.). This material was separated by reversed-phase semi-preparative HPLC (Shimadzu Epic C18 5µ 250 x 9.6 mm, 103

H2O/MeCN, 95 % à 65 % for 40 min, 65% for 20 min, UV detector at 280 nm, tR, undesired 44.9

22 min, tR, desired = 46.8 min) to afford pure quinoxaline 77 (0.0025 g, 19 % yield). 77: [α] D =

+103.4° (c = 0.05, MeOH); IR (film) νmax 2922, 2851, 1707, 1625, 1521, 1445, 1260, 1020, 799

-1 1 cm ; H NMR (400 MHz, acetone-d6) δ 8.04–7.99 (m, 2 H), 7.79–7.75 (m, 2 H), 7.53 (d, J =

16.0 Hz, 1 H), 7.23 (d, J = 5.2 H, 1 H), 6.88 (d, J = 2.0 Hz, 1 H), 6.83–6.75 (m, 3 H), 6.70–6.65

(m, 2 H), 6.52 (d, J = 16.0 Hz, 1 H), 5.24 (dd, J = 8.4, 4.0 Hz, 1 H), 4.93 (app t, J = 2.0 Hz, 1 H),

4.35 (dd, J = 6.4, 2.0 Hz, 1 H), 3.63 (dd, J = 6.0, 1.2 Hz, 1 H), 3.19 (dd, J = 6.0, 2.4 Hz, 1 H),

3.16 (dd, J = 16.0, 4.0 Hz, 1 H), 3.05 (dd, J = 14.4, 8.4 Hz, 1 H); 13C NMR (125 MHz, acetone-

d6) δ 173.5, 170.8, 166.4, 157.8, 155.2, 145.9, 145.6, 145.1, 144.7 (2 C), 141.7 (2 C), 141.6,

141.3, 134.8, 130.0, 129.7 (3 C), 128.9, 121.6, 120.0, 118.2, 117.3, 116.2, 115.9, 115.4, 74.0,

+ 52.0, 51.5, 48.4, 46.7, 37.4; HRMS (FAB) calcd for C33H27N2O10 [M+H] 611.1666, found

611.1648. All spectroscopic data for this compound matched that reported by Tanaka et al.3

104

Table 4. Comparison of 1H NMR chemical shifts of quinoxaline diastereomers derived from endo alcohols 75a (non-natural stereochemistry) and 75b (natural stereochemistry) with those of the naturally-derived quinoxaline (77)

Naturally-derived 75a-derived Absolute Difference 75b-derived Absolute Difference 8.04 8.05 0.01 8.02 0.02 7.79 7.80 0.01 7.77 0.02 7.53 7.52 0.01 7.53 0.00 7.22 7.23 0.01 7.23 0.01 6.84 6.92 0.08 6.88 0.04 6.81 6.83 0.02 6.82 0.01 6.81 6.81 0.00 6.81 0.00 6.75 6.79 0.04 6.75 0.00 6.67 6.68 0.01 6.68 0.01 6.64 6.67 0.03 6.66 0.02 6.56 6.57 0.01 6.52 0.04 5.21 5.23 0.02 5.24 0.03 4.95 5.01 0.06 4.93 0.02 4.39 4.40 0.01 4.35 0.04 3.61 3.63 0.02 3.63 0.02 3.17 3.23 0.06 3.19 0.02 3.13 3.15 0.02 3.16 0.03 3.02 3.04 0.02 3.05 0.03 0.024 0.020

Allyl Ester (79). To a solution of ferulic acid (78) (2.00 g, 10.3 mmol, 1.0 equiv) in DMF (30 mL) was added Na2CO3 (1.30 g, 12.4 mmol, 1.2 equiv). The resulting solution was stirred at 25

°C for 15 min. Allyl bromide (0.98 mL, 11.3 mmol, 1.1 equiv) was then added and the resulting

solution was stirred for 10.5 h. Upon completion, the reaction mixture was poured into 1 M HCl

(60 mL) and extracted with EtOAc (60 mL). The organic layer was then washed successively

with saturated aqueous NaHCO3 (60 mL) and brine (60 mL). The organic layer was then dried

(MgSO4), filtered, and concentrated to give a pale brown oil. The crude material was purified by flash column chromatography (silica gel, hexanes:EtOAc, 8:2à7:3) to give the desired allyl ester 79 (2.13 g, 88 %) as a yellow oil alongside the doubly allylated derivative (0.354 g, 12 %).

1 79: H NMR (300 MHz, CDCl3) δ 7.64 (d, J = 15.9 Hz, 1 H), 7.08 (dd, J = 8.2, 1.9 Hz, 1 H),

7.03 (d, J = 1.8 Hz, 1 H), 6.92 (d, J = 8.1 Hz, 1 H), 6.32 (d, J = 15.9 Hz, 1 H), 6.11–5.90 (m, 1 105

H), 5.86 (s, 1 H), 5.37 (dq, J = 17.2, 1.6 Hz, 1 H), 5.32–5.21 (m, 1 H), 4.71 (dt, J = 5.7, 1.4 Hz, 2

H), 3.93 (s, 3 H).

Allylated Diels-Alder Product (80). Allyl ester 79 (0.900 g, 4.27 mmol) and dienophile 69

(10.8 g, 29.9 mmol) were subjected to the Wessely oxidation/Diels–Alder protocol as described for the synthesis of compound 73. The crude product was purified by flash column chromatography (silica gel, hexanes:EtOAc, 4:1à3:2) to give 75 (1.36 g, 77 %) as a yellow foam alongside the Wessely acetoxylation product (analogous to 41) (0.269 g, 24 %). 80: 1H

NMR (400 MHz, CDCl3) δ 7.47–7.34 (m, 11 H), 6.84 (d, J = 8.4 Hz, 1 H), 6.48 (d, J = 2.4 Hz, 1

H), 6.42 (dd, J = 8.4, 2.0 Hz, 1 H), 6.29 (dd, J = 6.8, 2.4 Hz, 1 H), 6.22 (d, J = 16.0 Hz, 1 H),

6.01 (ddt, J = 17.2, 10.8, 5.6 Hz, 1 H), 5.41 (dd, J = 17.2, 1.6 Hz, 1 H), 5.32 (dd, J = 10.4, 1.2

Hz, 1 H), 5.17 (s, 2 H), 5.13 (d, J = 5.2, 2 H), 4.77–4.73 (m, 2 H), 3.99 (dd, J = 5.2, 2.4 Hz, 1 H),

3.74 (s, 3 H), 3.72 (br s, 1 H), 3.65 (dd, J = 6.8, 2.8 Hz, 1 H), 3.15 (dd, J = 5.2, 0.8 Hz, 1 H).

Carboxylic Acid (81). To a solution of allyl ester 80 (0.917 g, 1.55 mmol, 1.0 equiv) in THF (15

mL) was added Meldrum’s acid (0.335 g, 2.32 mmol, 1.5 equiv) and the solution was degassed

by bubbling through argon for 10 min. Pd(PPh3)4 (0.090 g, 0.078 mmol, 0.05 equiv) was then

added, and the resulting yellow solution was stirred for 15 min at 25 °C. Upon completion, the

reaction mixture was concentrated directly and the orange residue was purified by flash column

chromatography (silica gel, CH2Cl2:MeOH, 98:2à95:5) to give the acid 81 (0.814 g, 95%) as a

1 white amorphous solid. 81: H NMR (300 MHz, CDCl3) δ 7.50–7.26 (m, 11 H), 6.83 (d, J = 8.4

Hz, 1 H), 6.49 (s, 1 H), 6.43 (d, J = 8.4 Hz, 1 H), 6.29 (d, J = 6.8 Hz, 1 H), 6.22 (d, J = 15.8 Hz, 106

1 H), 5.12 (s, 2 H), 5.11 (s, 2 H), 4.06 (dd, J = 5.3, 2.2 Hz, 1 H), 3.79–3.59 (m, 5 H), 3.17 (dd, J

= 5.5, 1.5 Hz, 1 H).

Coupling Product (83a/b). To a solution of chiral alcohol 82 (prepared in analogy to chiral

alcohol 70) (0.187 g, 0.40 mmol, 1.0 equiv) and carboxylic acid 81 (0.440 g, 0.80 mmol, 2.0

equiv) in CH2Cl2 (4.0 mL) was added 4-DMAP (0.078 g, 0.64 mmol, 1.6 equiv) and EDC (0.154 g, 0.80 mmol, 2.0 equiv) at 25 °C. The yellow solution was stirred at 25 °C for 16 h. Upon completion, the reaction mixture was diluted with EtOAc (16 mL) and washed successively with

1 M HCl (20 mL; some brine was added to improve separation of the layers) and brine (20 mL).

The organic layer was then dried (MgSO4), filtered, and concentrated. The crude product was

purified by flash column chromatography (silica gel, hexanes:EtOAc, 7:3) to give 83 as a pale

1 yellow oil (0.337 g, 84 %, 1:1 d.r.). 83a/b: H NMR (300 MHz, CDCl3—NOTE: because of overlapping peaks of the two diastereomers, multiplet analysis could not be performed) δ 7.50–

7.20 (26 H), 6.93–6.80 (3 H), 6.78–6.64 (1 H), 6.53–6.41 (1 H), 6.41–6.32 (1 H), 6.30–6.11 (2

H), 5.43–5.28 (1 H), 5.19–5.02 (10 H), 4.00–3.88 (1 H), 3.75–3.63 (3 H), 3.63–3.53 (1 H), 3.24–

3.03 (3 H).

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186

CHAPTER 3

Synthetic Studies of Yunnaneic Acids A and B

187

3.1 Dimerization Strategy and Literature Precedent

In general, late-stage dimerization reactions represent an attractive method1 for the construction of natural products with symmetrical (or nearly symmetrical) structures because of the rapid increase in molecular complexity afforded by such a process (essentially doubling the molecular weight). Furthermore, such a strategy also allows the final product to arise from two subunits with either identical or similar structure. Altogether, these advantages inherent to dimerization strategies can result in remarkably efficient syntheses of highly complex molecules.2-7 However, such dimerizations can often prove extraordinarily challenging to reduce

to practice. In many cases, one must find appropriate conditions not only to form a reactive

monomeric intermediate, but also to induce that intermediate to react with a second monomer

through the desired dimerization pathway.

The dimeric yunnaneic acids A (1) and B (2)8 constitute a unique synthetic challenge,

even after accounting for the difficulties typically associated with dimerization reactions

(Scheme 1). As briefly discussed in Chapter 1 (Section 1.5), yunnaneic acid A requires the

selective “pseudo-dimerization” of two similar, but distinct, monomeric fragments–the

hydroxyketone of yunnaneic acid D (4) and the diketone of yunnaneic acid C (3). Yunnaneic

acid B, on the other hand, requires a ketalization event between a diketone (3) and its hydrated

form (5, Scheme 1). 188

HO HO OH OH

CO2H CO2H

O O O O HO HO OH HO OH + OH OH HO O HO H CO H O O 2 CO2H CO2H CO2H O 3 yunnaneic acid C 4 yunnaneic acid D O O O HO OH HO HO HO OH OH O HO O CO H CO H O R 2 2 CO2H OH CO2H O O O O 1 R = H yunnaneic acid A + 2 R = OH yunnaneic acid B HO HO OH HO O HO OH O O CO2H CO2H 3 yunnaneic acid C 5

Scheme 1. Required dimerizations for the synthesis of yunnaneic acids A and B

Although one could readily rationalize mechanisms to account for these dimerizations

(Scheme 2), it was not clear from the outset of our studies why we could expect these

dimerizations to proceed with the desired stereo-, regio-, and chemoselectivity. Nevertheless, we

had reasons to be optimistic that such dimerizations would be favorable processes. First, it

seemed reasonable to surmise that such dimerizations might proceed spontaneously in nature,

rather than requiring enzymes to forge the spiroketal linkages, especially because of the known

electrophilicity of cyclic 1,2-diketones.9,10 Thus, these dimerizations would seem likely to be thermodynamically downhill processes. Second, in an experiment conducted by the isolation chemists, yunnaneic acid A was found to only partially hydrolyze to yunnaneic acids C and D

8 when heated to 90 °C in dilute aqueous H2SO4 for 1.5 h. Thus, it would appear that, at least in 189 the case of yunnaneic acid A, the dimeric linkage should be robust and its formation largely irreversible.

OH HO OH OH

CO2H CO2H R1O2C R1O2C O O O O Ar O Ar HO OH OH OH R O O O HO O CO2H CO2H HO CO H O CO2H 6 2 R

3 4 R = H 5 R = OH

HO OH OH OH

CO2H CO2H O 1: R = H, yunnaneic acid A O O O HO OH 2: R = OH, yunnaneic acid B HO O HO O R CO H O 2 OH CO2H

Scheme 2. Proposed mechanism for dimerizations

In addition, we thought that the presence of the carboxylic acid on the endo face of the bicycle might discourage attack on the undesired carbonyl either through steric hindrance (Figure

1A) or perhaps by attacking the carbonyl, thus providing a form of internal protection (Figure

1B). We also noticed that the relative orientation of the two bicycles (labeled as unit A and unit

B in Figure 1C) in yunnaneic acids A (1) and B (2) and the unsymmetrical dimer (7) unintentionally synthesized previously in our laboratory (vide infra) was the same, with the free 190 hydroxyl group of the hemiketal present in unit B pointing in the general direction of the ketone

(or alcohol) of unit A, rather than away from it (as in 8).

Figure 1. A) Potential steric repulsion resulting from addition; B) Potential internal protection by carboxylic acid (bicycle substituents omitted for clarity); C) Relative orientations of bicycles of yunnaneic acids and unsymmetrical dimer

Although we were reassured by these observations, we were also aware that there were only sparse examples in the literature of similar dimerizations. The principal example of such a dimerization that served as the cornerstone of our planned dimerization strategy came from our own laboratory11 (cf. Section 1.5). In this case, my former colleague, Dr. Ferenc Kontes, showed that hydroxyketone 9 (Scheme 3), upon exposure to acidic or basic conditions (NaH proved to be 191

the optimal reagent), the alcohol (or alkoxide) of the hydroxyketone would engage the ketone of

a second equivalent of the hydroxyketone as shown to give, rather than a C2-symmetric dimer

with the desired dioxane linkage (10), the unsymmetrical dimer 7 (structure determined by X-ray

crystallography, see Figure 1 C), containing a spiroketal resembling those found in the yunnaneic

acids.*

CO2Me CO2Me

MeO2C O O CO2Me CO2Me OH OH MeO OMe OMe MeO2C O OMe 10 OH

CO2Me CO2Me NaH OMe

OMe MeO2C O CO2Me 9 O OH OH H

OMe MeO OMe OMe 7

Scheme 3. Unexpected dimerization hinting at a possible route towards yunnaneic acids

Another relevant example of a similar dimerization, this one involving self-condensation of a hydroxyketone, comes from the total synthesis of idesolide12 (12, Scheme 4), a natural product that has attracted much synthetic interest in recent years. Idesolide was successfully

* Although the dimerization was performed with a racemic mixture of starting material, only one diastereomer of product was isolated in near quantitative yield. Thus, a single pair of monomer enantiomers must react preferentially. 192

synthesized independently by Iwabuchi et al.13 and Kuwahara et al.,14 both in 2010, by dimerizing salicortin (11), the monomeric hydroxyketone, under basic conditions. In addition to informing our dimerization studies from an experimental standpoint, idesolide also served as a cautionary tale, as Snider et al. described their failure to observe the formation of idesolide from salicortin under a variety of conditions.15 Thus, although the dimerization may appear

straightforward, it proved difficult to reduce to practice. This difficulty may have arisen because

of a high kinetic barrier for the dimerization, especially compared to an aromatization pathway

that results in the formation of salicylic acid derivatives (13). In addition, these studies show the

importance of concentration to these dimerization processes, as both groups’ successful dimerization reactions proceeded under solvent-free conditions.†

Iwabuchi et al./Kuwahara et al. (2010)

MeO C CO2Me CO2Me 2 HO base O OH neat O O OH 11: salicortin 12: idesolide

Snider et al. (2007) MeO2C HO CO2Me O

O OH CO2Me 12 OH various conditions O 11 CO2Me

-H2O OH 13

Scheme 4. Synthetic studies of idesolide

† In contrast to the diastereoselective formation of 7, multiple diastereomers of 12 were isolated if racemic 11 was subjected to dimerization. 193

Diketone dimerizations of the type needed to form yunnaneic acid B have been observed

(Scheme 5). Scharf and Kuesters16 observed that, upon acidic hydrolysis of carbonate 14, dimer

16 was observed in 10–20% yield, with diketone 15 being the major product (60%). Klinotova et al. observed a similar dimerization event upon the attempted chromatographic purification of

diketone 17 on silica gel.17

Scharf and Kuesters (1972): OH + O O H3O O O O + Cl O O O OH Cl 60 % 10 – 20% 14 15 16

Klinotova et al. (1993): Me Me O Me O O Me HO O Me Me Me silica gel O O O O O O Me acetone/ Me OHO AcO pet. ether Me Me Me Me 17 18

Scheme 5. Two previously reported diketone dimerizations

Finally, an example of a yunnaneic acid A-type dimerization in which a hydroxyketone

reacted with a diketone was reported by Yates and Langford (Scheme 6).18 They showed that reaction of diketone 19 with 0.5 equiv. of MeLi resulted in formation of dimer 21, in which the anion of the hydroxyketone (20) formed from a single MeLi addition to the diketone attacked an unreacted diketone. Along similar lines, we hoped that deprotonating the alcohol of yunnaneic acid D in the presence of the diketone of yunnaneic acid C could result in formation of yunnaneic acid A. 194

O OLi O MeLi 19 O Me O (0.5 equiv) Me O O OH 19 20 21

Scheme 6. Dimerization observed by Yates and Langford

3.2 Dimerization Studies Towards Yunnaneic Acid A

As discussed above, we hoped to exploit the electrophilic nature of the 1,2-diketone to form the spiroketal of yunnaneic acid A (1). However, this task proved extremely difficult to reduce to practice. Conditions used for all of the examples of similar dimerizations in the literature were explored in an effort to synthesize a model version of yunnaneic acid A. When

model compounds 23 and 25 (methyl ester derivatives were used in initial model studies for ease of handling) were subjected to strong bases, the diketone appears to decompose (Scheme 7). The

hydroxyketone then dimerizes unproductively to give a compound tentatively assigned as 26.

Indeed, the same product is obtained when hydroxyketone 25 alone is treated with NaH

(although we could not confirm the exact structure, HRMS shows a molecular weight twice that

of the hydroxyketone). Under more weakly basic conditions, no reaction occurs. Under acidic

conditions, intractable mixtures of products often resulted. One exception was BF3•OEt2, which

also gave the hydroxyketone homodimer 26. 195

MeO2C MeO2C MeO a) HCl, MeO

MeO O MeOH MeO O OMe O O CO2Me O 22 23

MeO2C MeO2C MeO MeO b) TMSCHN2 OH OH MeO MeO H H O O CO2H CO2Me 24 25

MeO2C NaH MeO2C MeO 23 + 25 MeO OMe OH MeO O c) NaH 25 O CO2Me OH CO2Me 26 (tentative structure)

Reagents and conditions: a) HCl (0.5 M), MeOH, 25 °C, 14 h, 40%; b) TMSCHN2 (1.2 equiv), THF/MeOH (9:1), -78 °C, 15 min, 28%; c) NaH (10 equiv), THF, 0!25 °C, 1 h, 15% (52% brsm)

Scheme 7. Synthesis of model dimerization precursors and attempted dimerization under basic conditions

These failures led us to consider replacing the diketone with a more reactive

oxacarbenium ion generated from a dimethyl ketal. Thus, dimethyl ketal 27 was synthesized and

reacted, along with hydroxyketone 25, with BF3•OEt2 (Scheme 8). To our dismay, the unreacted dimethyl ketal was recovered. This suggests that the barrier to formation of such an unstable oxacarbenium ion is too great even for strong Lewis acids. To circumvent this barrier, we synthesized a reduced form of the dimethyl ketal. As a test case, hydroxyketal 29 (synthesized in 196

one step from siloxyketone 28) was reacted with BF3•OEt2, resulting in the C2-symmetric dimer

30 (structure confirmed by X-ray crystallography), precisely the same dimerization behavior that

was observed for a similar hydroxyketal in the helicterin studies.11 Although this dimer did not correspond to a natural product structure, it did suggest that an oxacarbenium ion, once generated, could react with an alcohol functionality to provide dimeric products. To prevent the unwanted homodimerization, the alcohol was protected as an acetate (31). Unfortunately,

subsequent reaction with hydroxyketone 24 in the presence of BF3•OEt2 resulted only in the ketal hydrolysis product.

Scheme 8. Experiments using dimethyl ketals as dimerization precursors

Through analysis of these experimental results, we noted several common elements.

Specifically, reactions were performed with a base or acid in a polar aprotic solvent. Non-polar 197

solvents could not dissolve the hydroxyketone, and most polar protic solvents were sufficiently nucleophilic to add into the diketone, preventing attack by the hydroxyketone. Eventually, it was found that 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP), a protic, non-nucleophilic solvent, in combination with molecular sieves to exclude water, was unique in providing the proper conditions to effect pseudodimerization of 24 and 32 (Scheme 9), affording a rather unstable dimer alongside a second dimer formed in smaller amounts that likewise proved unstable, hampering our ability to characterize these dimers. Fortunately, esterification of the crude dimerization products with TMSCHN2 allowed for easier isolation of the dimers of interest,

which were stable in their methyl ester forms. These compounds proved amenable to

bromination of their aryl rings upon treatment with NBS, which afforded crystalline compounds

33 and 34, which could be characterized by X-ray crystallography. Unfortunately, neither

structure’s connectivity matched that reported for yunnaneic acid A. Although we had indeed

achieved the selective union of the two distinct monomers with no apparent homodimerization, it

was found that, in the case of the major product 33, the wrong carbonyl of the diketone had been

engaged by the hydroxyketone. In the case of the minor product 34, the desired carbonyl had

undergone attack, but the wrong pair of enantiomers had reacted, resulting in a compound that

was a diastereomer of the structure corresponding to the natural product. 198

Scheme 9. Synthesis of heterodimeric analogs of yunnaneic acid A architecture

Interestingly, the carboxylic acid of 32 appears to be critical for facilitating this dimerization. The corresponding methyl ester (23) failed to react under the same conditions

(Scheme 10). On the other hand, both the free acid (24) and the methyl ester (25) of the hydroxyketone proved to be competent dimerization partners.

MeO CO2Me CO2Me MeO CO Me MeO CO Me 2 2 MeO OMe MeO MeO O HFIP OMe OH O O CO2Me O O O OH CO2Me CO2Me CO2Me H 25 23 35 + MeO MeO CO2Me CO2Me CO2Me MeO CO2Me

MeO MeO MeO OMe OH 1) HFIP O OMe O 2) TMSCHN 2 O OH CO Me O O O CO2Me 2 CO2H MeO2C H 32 25 36

Scheme 10. Experiments demonstrating the importance of the carboxylic acid within 32 to dimer formation 199

3.3 Dimerization Studies Towards Yunnaneic Acid B

Experiments performed concurrently with those described in Section 3.2, aimed at

forging the yunnaneic acid B-type linkage, first focused on adapting the conditions of Scharf and

Kuesters and Klinotova et al. (vide supra). Simply allowing the diketone to stand for extended

periods in the presence of varying amounts of water at various temperatures resulted in no

reaction. Likewise, exposure of the diketone to silica gel in the presence of various solvents and

additives resulted only in recovery of starting material. Ultimately, application of similar

conditions optimized for the pseudodimerization of the hydroxyketone and the diketone (HFIP)

to 32 alone also affords a mixture of dimers (37 and 38), albeit in lower yield (Scheme 11). In

this case, molecular sieves were omitted since 0.5 equivalents of water would be necessary. As

was the case with the heterodimers, these compounds proved rather unstable. Furthermore, these

compounds could not easily be converted to their methyl esters; reaction with TMSCHN2

resulted in a complex mixture of products. The desired methyl esters could be isolated by

reacting the crude dimerization mixture with Cs2CO3 and Me2SO4, but the reaction was low

yielding and X-ray quality crystals could not be obtained through this route.

MeO CO2Me

MeO Two dimers (37 + 38) O HFIP, 50 °C, 12 % and 8 % yield, respectively 16 h O Unknown structures CO2H 32

Scheme 11. Synthesis of two diketone homodimers

Fortunately, one of these dimers was characterized by X-ray crystallography following subjection of a brominated diketone (40) to the dimerization conditions (Scheme 12). Once 200

again, we had obtained a dimer whose structure (41) did not match that of the natural product. In

this case, rather than attack of a water molecule on the diketone as required, the pendant carboxyl

group had attacked the ketone, followed by attack of the resulting hemiketal oxygen on a second

diketone molecule. Furthermore, ketalization had occurred on the undesired carbonyl of the

diketone (syn to the carboxyl group). Unfortunately, X-ray quality crystals of the other isolated dimer proved elusive. As was the case for the pseudodimerization described in Section 3.2, the carboxylic acid of the diketone was necessary for the dimerization to occur since the acid acts as a nucleophile in the formation of dimer 41.

Scheme 12. Synthesis of a crystalline derivative of one of the diketone homodimers

3.4 Conclusion

The experiments described in this chapter seem to suggest that formation of the spiroketal

linkages found in yunnaneic acids A and B is not a spontaneous process. Instead, only mixtures 201

of dimers containing the improper connectivity and/or stereochemistry could be isolated. Thus,

the structures formed in nature do not seem commensurate with the innate reactivity of the

monomers. This may suggest some kind of enzymatic assistance for nature’s dimerization

process. Furthermore, the results of these studies seem consistent with those carried out in our

laboratory in pursuit of the helicterins, in which it was also found that the “obvious” monomer

(i.e. hydroxyketone 9) failed to dimerize spontaneously in the desired manner (see Scheme 3).

Nevertheless, we were successful in developing conditions that resulted in

pseudodimerization of the diketone and hydroxyketone without any apparent competition from

homodimerization. It is not fully clear why HFIP proved to be the ideal solvent for this reaction.

One possibility is its ability to act as a hydrogen bond donor without nucleophilically adding to

the diketone. It is also possible that it is a weak enough acid that it does not lead to uncontrolled

19 reaction pathways and complex product mixtures, but strong enough (pKa = 9.3) to activate the

diketone for attack.

Finally, the work described in this chapter exemplifies the promise of dimerization

strategies to construct highly complex molecules in a concise fashion. At the same time, the

molecules constructed did not match the targeted structures. Thus, this work also underscores the

high degree of difficulty associated with these dimerization strategies. It also highlights the need

for more work in the area of complex, oligomeric natural products to make dimerization

strategies more practicable for future synthetic efforts. Success in this area would allow chemists

to access compounds of far greater complexity from identical or similar precursors.20

Postdoctoral fellow Lorenzo Botta adapted the HFIP-mediated dimerization conditions developed for the synthesis of compounds 33 and 34 to the attempted synthesis of yunnaneic acid B (as described in Section 3.3) and obtained X-ray quality crystals of compound 41. His 202

contribution to this work is gratefully acknowledged. I also wish to thank several members of the

Parkin group, who performed the X-ray analyses: Dr. Wesley Sattler (33), Dr. Aaron Sattler (34),

and Yi Rong (41).

3.5 References

(1) Nicolaou, K. C.; Snyder, S. A. Classics in Total Synthesis II: More Targets, Strategies, Methods; Wiley-VCH: Weinheim, 2003. pp. 351-355 and references cited therein.

(2) Kim, J.; Ashenhurst, J. A.; Movassaghi, M. Science 2009, 324, 238.

(3) Ondrus, A. E.; Movassaghi, M. Chem. Commun. 2009, 4151.

(4) Shoji, M.; Hayashi, Y. Eur. J. Org. Chem. 2007, 3783.

(5) Snyder, S. A.; ElSohly, A. M.; Kontes, F. Angew. Chem. Int. Edit. 2010, 49, 9693.

(6) For a review on oligomeric natural products and their synthesis, see: Snyder, S. A.; ElSohly, A. M.; Kontes, F. Nat. Prod. Rep. 2011, 28, 897.

(7) For a review on synthetic strategies towards C2-symmetric natural products, see: Vrettou, M.; Gray, A. A.; Brewer, A. R. E.; Barrett, A. G. M. Tetrahedron 2007, 63, 1487.

(8) Tanaka, T.; Nishimura, A.; Kouno, I.; Nonaka, G.; Young, T. J. J. Nat. Prod. 1996, 59, 843.

(9) Rubin, M. B. Fortschr. Chem. Forsch. 1969, 13, 251.

(10) Sandris, C.; Ourisson, G. Bull. Soc. Chim. Fr. 1958, 350.

(11) Snyder, S. A.; Kontes, F. J. Am. Chem. Soc. 2009, 131, 1745.

(12) Kim, S. H.; Sung, S. H.; Choi, S. Y.; Chung, Y. K.; Kim, J.; Kim, Y. C. Org. Lett. 2005, 7, 3275. 203

(13) Yamakoshi, H.; Shibuya, M.; Tomizawa, M.; Osada, Y.; Kanoh, N.; Iwabuchi, Y. Org. Lett. 2010, 12, 980.

(14) Nagasawa, T.; Shimada, N.; Torihata, M.; Kuwahara, S. Tetrahedron 2010, 66, 4965.

(15) Richardson, A. M.; Chen, C. H.; Snider, B. B. J. Org. Chem. 2007, 72, 8099.

(16) Scharf, H.-D.; Kuesters, W. Chem. Ber. 1972, 105, 564.

(17) Klinotova, E.; Klinot, J.; Krecek, V.; Hilgard, S.; Budesinsky, M.; Malat, J. Collect. Czech. Chem. Commun. 1993, 58, 1675.

(18) Yates, P.; Langford, G. E. Can. J. Chem. 1981, 59, 344.

(19) Shuklov, I. A.; Dubrovina, N. V.; Boerner, A. Synthesis 2007, 2925.

(20) Snyder, S.A. Nature 2010, 465, 560. 204

3.6 Experimental Section

General Procedures. All reactions were carried out under an argon atmosphere with dry solvents under anhydrous conditions, unless otherwise stated. Dry methylene chloride (CH2Cl2), diethyl ether (Et2O), and tetrahydrofuran (THF) were obtained by passing commercially available pre-dried, oxygen-free formulations through activated alumina columns; triethylamine

(Et3N) was distilled from KOH; methanol (MeOH) was purchased in anhydrous form from

Sigma-Aldrich and used as received. Yields refer to chromatographically and spectroscopically

(1H and 13C NMR) homogeneous materials, unless otherwise stated. Reagents were purchased at the highest commercial quality and used without further purification, unless otherwise stated.

Reactions were magnetically stirred and monitored by thin-layer chromatography (TLC) carried out on 0.25 mm E. Merck silica gel plates (60F-254) using UV light and an aqueous solution of cerium ammonium sulfate and ammonium molybdate and heat as visualizing agents. Preparative

TLC was carried out on 0.50 mm E. Merck silica gel plates (60F-254). SiliCycle silica gel (60 Å, academic grade, particle size 40-63 µm) was used for flash column chromatography. NMR

spectra were recorded on Bruker DRX-300, DRX-400, DRX-500, and 500 ASCEND instruments

and calibrated using residual undeuterated solvent as an internal reference. The following

abbreviations are used to explain multiplicities: s = singlet, d = doublet, t = triplet, q = quartet, m

= multiplet, br = broad, app = apparent. IR spectra were recorded on a Perkin-Elmer Spectrum

Two FT-IR spectrometer. High resolution mass spectra (HRMS) were recorded in the Columbia

University Mass Spectral Core facility on a JOEL HX110 mass spectrometer using FAB (Fast

Atom Bombardment). Optical rotations were recorded on a Jasco DIP-1000 digital polarimeter.

205

Diketone Methyl Ester (23) and Dimethyl Ketal (27). To a solution of ketone 22 (0.100 g, 0.24

mmol, 1.0 equiv) in MeOH (1.0 mL) was added methanolic HCl (1.0 M solution, 1.0 mL) at 25

°C, and the resulting solution was stirred at this temperature for 16 h. Upon completion, the

reaction mixture was concentrated directly. The resulting yellow oil was purified by flash

column chromatography (silica gel, hexanes:EtOAc, 3:2à1:1) to give dimethyl ketal 27 (0.040 g, 36 % yield) as a pale yellow oil and diketone 23 (0.040 g, 40 %) as a yellow foam. 23: 1H

NMR (400 MHz, CDCl3) δ 7.47 (d, J = 15.8 Hz, 1 H), 6.83 (d, J = 8.3 Hz, 1 H), 6.77–6.61 (m, 3

H), 6.26 (d, J = 15.9 Hz, 1 H), 4.13 (br s, 1 H), 3.96–3.74 (m, 14 H), 3.18 (dd, J = 6.7, 2.7 Hz, 1

H).

1 27: H NMR (400 MHz, CDCl3) δ 7.42 (d, J = 15.8 Hz, 1 H), 6.79 (d, J = 8.3 Hz, 1 H), 6.75–

6.64 (m, 3 H), 6.12 (d, J = 15.9 Hz, 1 H), 3.96 (d, J = 6.1 Hz, 1 H), 3.86 (s, 6 H), 3.82 (br s, 1 H),

3.79 (s, 3 H), 3.74 (s, 3 H), 3.44 (s, 3 H), 3.36 (d, J = 6.8 Hz, 1 H), 3.29 (s, 3 H), 2.92 (dd, J =

6.5, 2.6 Hz, 1 H).

Hydroxyketone Methyl Ester (25). To a solution of carboxylic acid 24 (0.116 g, 0.29 mmol,

1.0 equiv) in THF/MeOH (9:1, 3.0 mL) at –78 °C was added TMSCHN2 (2.0 M in Et2O, 0.18 mL, 0.36 mmol, 1.2 equiv). The resultant yellow solution was then stirred at –78 °C for 15 min before AcOH was added dropwise until all bubbling ceased. The resulting pale yellow solution was then diluted with EtOAc (15 mL) and washed with saturated aqueous NaHCO3 (15 mL).

The aqueous layer was then extracted with EtOAc (2 × 15 mL), and the combined organic layers

were dried (MgSO4), filtered, and concentrated. The resultant crude white foam product was purified by flash column chromatography (silica gel, hexanes:EtOAc, 2:3) to give the ester 25

(0.033 g, 28% yield) as a clear colorless oil. 25: Rf = 0.39 (hexanes:EtOAc, 3:7); IR (film) νmax 206

3458, 3069, 2999, 2954, 2841, 1740, 1631, 1518, 1464, 1437, 1314, 1256, 1196, 1174, 1087,

–1 1 1026, 735 cm ; H NMR (500 MHz, CDCl3) δ 7.37 (d, J = 16.0 Hz, 1 H), 6.77 (d, J = 8.5 Hz, 1

H), 6.70 (d, J = 6.0 Hz, 1 H), 6.68 (d, J = 2.0 Hz, 1 H), 6.61 (dd, J = 8.5, 2.0 Hz, 1 H), 6.11 (d, J

= 15.5 Hz, 1 H), 4.09 (d, J = 2.5 Hz, 1 H), 3.85 (s, 3 H), 3.84 (s, 3 H), 3.82 (d, J = 1.5 Hz, 1 H),

3.77 (s, 3 H), 3.73 (s, 3 H), 3.58 (dd, J = 5.5, 2.0 Hz, 1 H), 3.37 (app dt, J = 6.0, 2.0 Hz, 1 H),

13 2.94 (dd, J = 5.5, 2.5 Hz, 1 H), 2.91 (br s, 1 H); C NMR (125 MHz, CDCl3) δ 207.8, 174.2,

166.9, 149.0, 148.3, 140.7, 139.5, 135.6, 135.0, 119.1, 118.4, 111.2, 111.1, 71.5, 55.9 (2 C),

+ + 52.9, 51.9, 49.8, 48.9, 46.4, 46.1; HRMS (FAB) calcd for C22H24O8 [M] 416.1471, found

416.1484.

Hydroxyketone Homodimer (26). To a solution of hydroxyketone 25 (0.041 g, 0.099 mmol, 1.0

equiv) in THF (1.0 mL) was added NaH (60% dispersion in mineral oil, 0.039 g, 0.99 mmol,

10.0 equiv) at 0 °C. The resulting yellow slurry was warmed to 25 °C and stirred for 1 h. Upon

completion, saturated aqueous NH4Cl (1.0 mL) was then carefully added and the mixture was poured into water (5 mL) and extracted with EtOAc (3 × 5 mL). The combined organic layers were then washed with brine (5 mL), dried (MgSO4), filtered, and concentrated. The resultant crude clear, colorless oil was purified by flash column chromatography (silica gel, hexanes:EtOAc, 3:7) to give dimeric product 26 (6 mg, 15% yield) along with recovered starting material (15 mg, 37% yield). 26: Rf = 0.20 (hexanes:EtOAc, 3:7); IR (film) νmax 3420, 2990,

2951, 2914, 2841, 1720, 1631, 1517, 1435, 1313, 1257, 1197, 1027 cm–1; 1H NMR (500 MHz,

CDCl3) δ 7.58 (d, J = 15.0 Hz, 1 H), 7.36 (d, J = 15.5 Hz, 1 H), 6.77 (d, J = 8.5 Hz, 1 H), 6.73–

6.69 (m, 3 H), 6.66 (dd, J = 8.5, 2.0 Hz, 1 H), 6.61 (dd, J = 8.5, 2.0 Hz, 1 H), 6.57 (d, J = 6.0

Hz, 1 H), 6.41 (d, J = 6.0 Hz, 1 H), 6.19 (d, J = 16.0 Hz, 1 H), 5.93 (d, J = 16.0 Hz, 1 H), 4.74 207

(br s, 1 H), 4.14 (d, J = 4.0 Hz, 1 H), 3.85 (s, 9 H), 3.81 (s, 6 H), 3.76 (s, 3 H), 3.67 (s, 3 H), 3.65

(s, 3 H), 3.58 (d, J = 6.5 Hz, 1 H), 3.41 (d, J = 7.5 Hz, 1 H), 3.27 (br s, 1 H), 3.11–3.08 (m, 2 H),

2.56–2.53 (m, 2 H), 2.32 (dd, J = 7.5, 2.0 Hz, 1 H); 13C NMR 174.0, 172.5, 167.2, 148.9, 148.8,

148.1, 147.8, 140.9, 140.6 (2 C), 140.1, 137.5, 136.5, 136.2, 135.9, 118.9, 118.1, 117.8, 111.5,

111.3, 111.2 (2 C), 111.1, 106.2, 85.4, 55.9 (3 C), 55.8, 52.5, 52.0, 51.8, 51.7, 49.5, 49.1, 46.4,

+ + 44.6, 43.3, 43.0, 41.7, 41.0; HRMS (FAB) calcd for C44H48O16 [M] 832.2942, found 832.2960.

Hydroxyketal (29). To a solution of siloxyketone 28 (0.088 g, 0.17 mmol, 1.0 equiv) in

MeOH/HC(OMe)3 (4:1, 2.0 mL) was added HCl (4.0 M solution in dioxane, 0.25 mL) at 25 °C, and the resulting solution was stirred for 16 h. Upon completion, the reaction mixture was concentrated directly, and the resulting crude material was purified by flash column chromatography (silica gel, hexanes:EtOAc, 3:2) to give hydroxyketal 29 (0.075 g, 95 %) as a

1 pale yellow oil. 29: H NMR (300 MHz, CDCl3) δ 7.50 (d, J = 15.8 Hz, 1 H), 6.75 (d, J = 8.3

Hz, 1 H), 6.68 (d, J = 2.0 Hz, 1 H), 6.63 (dd, J = 8.3, 2.1 Hz, 1 H), 6.54 (d, J = 6.5 Hz, 1 H), 6.20

(d, J = 15.7 Hz, 1 H), 4.02 (d, J = 2.0 Hz, 1 H), 3.90 (d, J = 6.5 Hz, 1 H), 3.85 (s, 3 H), 3.84 (s, 3

H), 3.32 (s, 3 H), 3.24 (s, 3 H), 3.06 (dd, J = 6.6, 1.5 Hz, 1 H), 2.56 (dd, J = 7.4, 2.4 Hz, 1 H).

C2-symmetric dimer 30. To a solution of dimethyl ketal 29 (0.075 g, 0.16 mmol, 1.0 equiv) in

CH2Cl2 (1.5 mL) at 0 °C was added BF3•OEt2 (0.09 mL, 0.65 mmol, 4.0 equiv) in a single

portion. The resulting pale orange solution was stirred at 0 °C for 30 min. Upon completion,

saturated aqueous NaHCO3 (0.5 mL) was added and the mixture was warmed to 25 °C. The mixture was then poured into H2O (10 mL) and extracted with EtOAc (3 × 10 mL). The combined organic layers were washed with brine (10 mL), dried (MgSO4), filtered, and 208

concentrated. The resultant crude pale brown oil was purified by flash column chromatography

(silica gel, hexanes:EtOAc, 1:1→3:7) to give 30 as a clear waxy solid (0.025 g, 36% yield)

which crystallized from CH2Cl2/n-hexane (colorless prisms). 30: Rf = 0.47 (hexanes:EtOAc,

–1 1 3:7); IR (film) νmax 2961, 1720, 1631, 1516, 1436, 1265, 1236, 1169, 1132, 1028, 841 cm ; H

NMR (400 MHz, CDCl3) δ 7.37 (d, J = 16.0 Hz, 2 H), 6.75 (d, J = 8.8 Hz, 2 H), 6.67–6.64 (m, 4

H), 6.26 (d, J = 6.4 Hz, 2 H), 6.03 (d, J = 15.6 Hz, 2 H), 3.85 (s, 6 H), 3.84 (s, 6 H), 3.83 (s, 6

H), 3.74 (s, 6 H), 3.69 (d, J = 3.2 Hz, 2 H), 3.62 (br d, J = 6.0 Hz, 2 H), 3.43 (app q, J = 4.4, 2.0

Hz, 2 H), 3.37 (s, 6 H), 3.05 (dd, J = 6.8, 1.2 Hz, 2 H), 2.58 (dd, J = 6.8, 2.8 Hz, 2 H); 13C NMR

(100 MHz, CDCl3) 173.0, 167.7, 148.8, 147.9, 142.2, 140.0, 136.7, 136.3, 119.3, 116.7, 111.1,

111.0, 100.2, 70.2, 55.9 (2 C), 52.3, 51.7, 49.3, 48.9, 44.7, 39.9, 39.2; HRMS (FAB) calcd for

+ + C46H52O16 [Μ] 860.3255, found 860.3265.

Acetate (31). To a solution of hydroxyketal 29 (0.027 g, 0.058 mmol, 1.0 equiv) in CH2Cl2 (0.8

mL) was added Ac2O (0.1 mL, excess), Et3N (0.1 mL, excess), and DMAP (catalytic) at 0 °C.

The resulting yellow solution was slowly allowed to warm to 25 °C and stirred for 14 h. Upon completion, MeOH (0.05 mL) was added, and the solution was diluted with EtOAc (5 mL) and washed successively with 0.1 M HCl (5 mL) and saturated aqueous NaHCO3 (5 mL). The

combined aqueous layers were then extracted with EtOAc (2 x 5 mL). The combined organic

layers were dried (MgSO4), filtered, and concentrated to give a pale yellow oil. The crude material was purified by flash column chromatography to give acetate 29 (0.024 g, 83 %) as a

1 pale yellow oil. 29: H NMR (400 MHz, CDCl3) δ 7.40 (d, J = 15.8 Hz, 1 H), 6.79–6.71 (m, 2

H), 6.69 (dd, J = 8.4, 2.1 Hz, 1 H), 6.62 (d, J = 6.4 Hz, 1 H), 6.13 (d, J = 15.8 Hz, 1 H), 4.72 (d, 209

J = 3.9 Hz, 1 H), 3.91–3.65 (m, 18 H), 3.29 (s, 3 H), 3.14 (s, 3 H), 3.10 (d, J = 5.9 Hz, 1 H), 2.42

(d, J = 7.9 Hz, 1 H), 2.05 (s, 3 H).

Heterodimer Methyl Esters (35 and 36). Diketone 32 (0.081 g, 0.20 mmol, 1.0 equiv) and

hydroxyketone 24 (0.081 mg, 0.20 mmol, 1.0 equiv) were dissolved in CH2Cl2/HFIP (1:1, 0.4

mL) and pulverized 3 Å molecular sieves were added (80 mg). The resulting yellow mixture was

then stirred at 25 °C for 14 h. Upon completion, the sieves were removed by filtration and the

filtrate was concentrated. The resultant yellow foam was dissolved in Et2O/MeOH (10:1, 2.2

mL), cooled to –78 °C, and TMSCHN2 (2.0 M in Et2O, 0.4 mL, 0.8 mmol, 4.0 equiv) was added.

After stirring the resultant yellow mixture at –78 °C for 30 min, AcOH was added dropwise until bubbling ceased. The mixture was then warmed to 25 °C, diluted with EtOAc (10 mL), and washed with water (10 mL). The organic layer was then dried (MgSO4), filtered, and

concentrated. The resultant crude yellow oil was purified by column chromatography (silica gel,

hexanes:EtOAc, 1:1→2:3) to give a mixture of two dimers as a clear colorless oil (0.042 g, 26%

combined yield over two steps, 4:1 mixture of 35:36). These two dimeric products were then

separated by preparative TLC (CH2Cl2:Et2O, 9:1). 35: Rf = 0.35 (hexanes:EtOAc, 3:7); IR (film)

–1 1 νmax 3473, 2952, 2928, 2844, 1722, 1632, 1592, 1517, 1436, 1256, 1027, 734 cm ; H NMR

(500 MHz, CDCl3) δ 7.41 (d, J = 16.0 Hz, 1 H), 7.33 (d, J = 16.0 Hz, 1 H), 6.78–6.73 (m, 3 H),

6.69–6.61 (m, 4 H), 6.46 (d, J = 6.0 Hz, 1 H), 6.16 (d, J = 15.5 Hz, 1 H), 6.07 (d, J = 16.0 Hz, 1

H), 4.43 (d, J = 3.5 Hz, 1 H), 3.87–3.82 (m, 15 H), 3.78 (s, 3 H), 3.78 (s, 3 H), 3.74 (s, 3 H), 3.69

(s, 3 H), 3.61 (d, J = 7.0 Hz, 1 H), 3.28 (app t, J = 5.0 Hz, 1 H), 2.90 (dd, J = 6.5, 2.5 Hz, 1 H),

13 2.74 (dd, J = 6.0, 1.5 Hz, 1 H), 2.45 (dd, J = 7.5, 2.0 Hz, 1 H); C NMR (125 MHz, CDCl3)

δ 206.1, 172.6 (2 C), 167.6, 166.6, 149.1, 148.9, 148.5, 148.0, 141.1, 140.8, 139.2, 138.8, 136.2 210

(2 C), 134.9, 119.4, 119.2, 119.0, 117.5, 111.4, 111.3 (4 C), 107.7, 101.7, 86.3, 56.0 (2 C), 55.9

(2 C), 52.7, 52.1, 52.0, 51.8, 50.6, 50.0, 49.4, 48.1, 44.0, 43.3, 42.2, 40.5; HRMS (FAB) calcd

+ + for C44H47O16 [M + H] 831.2864, found 831.2842.

36: Rf = 0.35 (hexanes:EtOAc, 3:7); IR (film) νmax 3461, 3062, 3009, 2951, 2834, 1724, 1632,

–1 1 1518, 1436, 1313, 1258, 1198, 1028 cm ; H NMR (500 MHz, CDCl3) δ 7.61 (d, J = 15.5 Hz, 1

H), 7.37 (d, J = 15.5 Hz, 1 H), 6.80–6.75 (m, 4 H), 6.71–6.68 (m, 2 H), 6.46 (d, J = 6.5 Hz, 1 H),

6.41 (d, J = 6.0 Hz, 1 H), 6.25 (d, J = 16.0 Hz, 1 H), 6.03 (d, J = 16.0 Hz, 1 H), 4.36 (d, J = 3.0

Hz, 1 H), 4.00 (d, J = 8.0 Hz, 1 H), 3.91 (br s, 1 H), 3.88–3.85 (m, 15 H), 3.80 (s, 3 H), 3.72 (s, 6

H), 3.63 (d, J = 7.0 Hz, 1 H), 3.49–3.47 (m, 2 H), 3.13 (app t, J = 5.0 Hz, 1 H), 2.65 (dd, J = 7.5,

13 2.0 Hz, 1 H), 2.54 (dd, J = 7.0, 2.0 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 203.7, 172.5, 171.4,

167.2, 166.7, 149.1, 149.0, 148.5, 148.1, 143.2, 141.1, 140.8, 139.4, 136.1, 135.7, 134.6, 130.8,

120.0, 119.0, 118.9, 118.1, 111.5, 111.4 (2 C), 111.1, 108.0, 101.1, 86.8, 55.9 (4 C), 54.6, 52.2,

+ + 52.1, 51.8, 51.7, 49.1, 49.0, 44.5, 43.3 (2 C), 40.8, 40.4; HRMS calcd for C44H47O16 [M + H]

831.2864, found 831.2861.

Brominated Heterodimers (33 and 34). To a solution of a mixture of 35 and 36 (~1.7:1 35:36,

0.030 g total, 0.036 mmol, 1.0 equiv) in MeCN (1.0 mL) was added NBS (0.026 g, 0.14 mmol,

4.0 equiv) at 25 °C. The clear colorless solution was stirred at 25 °C for 14 h. Upon completion,

saturated aqueous Na2SO3 (0.5 mL) was added, and the resulting mixture was stirred vigorously

for 15 min. Upon completion, the mixture was diluted with EtOAc (5 mL) and washed with

saturated aqueous NaHCO3 (3 x 5 mL). The organic layer was dried (MgSO4), filtered, and concentrated to give a white solid. The crude material was purified by preparative TLC (silica gel, hexanes:EtOAc, 2:3) to afford 33 (0.012 g) and 34 (0.007 g, 53 % overall yield) as white solids. Both products were crystallized from CHCl3/i-PrOH. 33: Rf = 0.35 (hexanes:EtOAc, 2:3); 211

IR (film) νmax 3463, 3002, 2951, 2844, 1720, 1632, 1506, 1506, 1486, 1376, 1312, 1257, 1195,

–1 1 1164, 1063, 1029, 913, 842, 730 cm ; H NMR (400 MHz, CDCl3) δ 7.63 (d, J = 15.6 Hz, 1 H),

7.37 (d, J = 16.0 Hz, 1 H), 7.07 (d, J = 4.0 Hz, 2 H), 6.40 (d, J = 6.0 Hz, 2 H), 6.34–6.28 (m, 3

H), 6.06 (d, J = 16.0 Hz, 1 H), 4.87 (s, 1 H), 4.64 (d, J = 7.6 Hz, 1 H), 4.48 (d, J = 3.2 Hz, 1 H),

4.27 (d, J = 6.8 Hz, 1 H), 3.99 (br s, 1 H), 3.89 (s, 3 H), 3.87 (s, 6 H), 3.81 (s, 3 H), 3.78 (s, 3 H),

3.76 (s, 3 H), 3.72 (s, 3 H), 3.69 (s, 3 H), 3.53 (br s, 1 H), 3.37 (dd, J = 6.4, 1.2 Hz, 1 H), 3.05–

3.01 (m, 1 H), 2.77 (dd, J = 7.6, 2.0 Hz, 1 H), 2.66 (dd, J = 5.2, 2.4 Hz, 1 H); 13C NMR (100

MHz, CDCl3) δ 204.1, 173.0, 172.0, 168.2, 167.7, 150.1, 149.7 (2 C), 149.5, 144.4, 142.4, 140.1,

136.5, 134.5, 132.9, 131.6, 121.4, 119.5, 117.2, 117.0, 116.4, 116.2, 111.7, 111.4, 108.9, 101.9,

87.6, 57.3, 57.2 (4 C), 55.0, 53.5, 53.3, 52.9, 52.8, 48.7, 48.5, 45.0, 44.4, 40.6, 40.0; HRMS

+ + calcd for C44H44Br2O16Na [M + Na] 1010.0972; HRMS could not be obtained. LRMS

(MALDI), showed peaks at 1010.90 (M + Na) and 1026.87 (M + K).

1 34: Rf = 0.35 (hexanes:EtOAc, 2:3); H NMR (400 MHz, CDCl3) δ 7.47 (d, J = 16.0 Hz, 1 H),

7.36 (d, J = 16.0 Hz, 1 H), 7.06 (d, J = 7.2 Hz, 2 H), 6.54 (d, J = 6.5=4 Hz, 1 H), 6.41–6.38 (m, 3

H), 6.24 (d, J = 15.6 Hz, 1 H), 6.15 (d, J = 15.6 Hz, 1 H), 4.91 (s, 1 H), 4.54 (d, J = 3.6 Hz, 1 H),

4.49 (dd, J = 6.8, 2.0 Hz, 1 H), 4.28 (d, J = 7.2 Hz, 1 H), 3.95 (br s, 1 H), 3.89 (s, 3 H), 3.87 (s, 3

H), 3.81 (s, 6 H), 3.77 (br s, 1 H), 3.77 (s, 9 H), 3.72 (s, 3 H), 3.22–3.19 (m, 1 H), 3.00–2.98 (m,

1 H), 2.67 (dd, J = 6.8, 2.0 Hz, 1 H), 2.63 (dd, J = 6.8, 2.0 Hz, 1 H).

Diketone homodimers (37 and 38). Diketone 32 (0.113 g, 0.28 mmol, 1.0 equiv) was dissolved in HFIP (0.09 mL) and stirred at 50 °C for 16 h. Upon completion, the yellow colored solution was cooled to 25 °C, concentrated, and purified by preparative TLC (CH2Cl2:MeOH, 93:7) to

give 37 (0.014 mg, 12% yield) and 38 (0.009 g, 8% yield) as pale yellow foams. 37: Rf = 0.18 212

(hexanes:EtOAc, 3:7); IR (film) νmax 3385, 3059, 2999, 2953, 2933, 2844, 1804, 1718, 1635,

-1 1 1518, 1464, 1438, 1313, 1257, 1242, 1145, 1027 cm ; H NMR (400 MHz, CDCl3) δ 7.49 (d, J

= 16.0 Hz, 1 H), 7.42 (d, J = 16.0 Hz, 1 H), 6.78–6.75 (m, 2 H), 6.73 (d, J = 8.4 Hz, 1 H), 6.68

(dd, J = 8.0, 1.6 Hz, 1 H), 6.53 (d, J = 6.4 Hz, 1 H), 6.49 (d, J = 2.0 Hz, 1 H), 6.44 (d, J = 6.4

Hz, 1 H), 6.35 (dd, J = 8.4, 2.0 Hz, 1 H), 6.17 (d, J = 16.0 Hz, 1 H), 6.09 (d, J = 16.0 Hz, 1 H),

5.66 (s, 1 H), 4.06 (d, J = 8.0 Hz, 1 H), 3.88–3.79 (m, 19 H), 3.76 (br s, 1 H), 3.60 (dd, J = 6.4,

1.2 Hz, 1 H), 3.52 (br s, 1 H), 3.47 (dd, J = 6.8, 3.2 Hz, 1 H), 3.13 (br d, J = 4.0 Hz, 1 H), 2.75

13 (dd, J = 8.0, 2.0 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 201.2, 172.0, 167.4, 166.7, 149.1,

148.8, 148.6, 148.1, 142.3, 140.9, 139.4, 137.6, 133.9, 133.4, 132.1, 130.9, 120.1, 119.6, 118.8,

118.5, 111.5, 111.3, 111.1, 111.0, 109.9, 105.4, 102.3, 56.0, 55.9 (2 C), 54.3, 53.4, 52.1, 52.0,

+ + 48.8, 48.6, 47.0, 46.6, 44.0, 42.7, 40.0, 29.7; HRMS (FAB) calcd for C42H40O16 [M] 800.2316,

found 800.2295.

1 38: H NMR (400 MHz, CDCl3) δ 7.59 (d, J = 16.0 Hz, 1 H), 7.44 (d, J = 15.6 Hz, 1 H), 6.77 (d,

J = 4.8 Hz, 1 H), 6.75 (d, J = 4.8 Hz, 1 H), 6.65 (d, J = 2.0 Hz, 1 H), 6.58–6.49 (m, 4 H), 6.40

(dd, J = 8.4, 2.0 Hz, 1 H), 6.23 (d, J = 16.0 Hz, 1 H), 5.99 (d, J = 15.6 Hz, 1 H), 3.93 (dd, J =

4.8, 1.6 Hz, 1 H), 3.89 (s, 3 H), 3.863 (s, 3 H), 3.857 (s, 3 H), 3.85 (s, 3 H), 3.83 (s, 3 H), 3.81 (s,

3 H), 3.78–3.75 (m, 2 H), 3.54–3.50 (m, 3 H), 3.14 (dd, J = 5.2, 1.2 Hz, 1 H), 2.77 (dd, J = 7.6,

13 2.0 Hz, 1 H); C NMR (100 MHz, CDCl3) δ 201.9, 166.8, 166.6, 149.2, 148.8, 148.7, 148.2,

141.7, 141.0, 139.4, 138.8, 133.6, 132.8, 132.4, 131.0, 119.9, 119.8, 119.0, 118.6, 111.7, 111.5,

111.1, 110.9, 109.4, 106.5, 101.6, 56.0, 55.9 (2 C), 54.4, 52.0, 51.9, 48.7, 48.0, 47.4, 46.0, 43.8,

42.5, 41.4, 29.7.

213

Brominated Diels–Alder product (39). Following the bromination procedure above, Diels–

Alder product 22 (0.140 g, 0.34 mmol, 1.0 equiv) was reacted with NBS (0.073 g, 0.41 mmol,

1.2 equiv) to give the brominated product as a pale yellow solid (0.080 g, 48%). 39: Rf = 0.39

(hexanes:EtOAc, 1:1); IR (film) νmax 3486, 3065, 2952, 2843, 1794, 1749, 1716, 1633, 1593,

-1 1 1518, 1453, 1316, 1252, 1196, 1088, 1026, 920, 881, 735 cm ; H NMR (400 MHz, CDCl3)

δ 7.46 (d, J = 15.6 Hz, 1 H), 6.78 (d, J = 8.4 Hz, 1 H), 6.51 (d, J = 2.0 Hz, 1 H), 6.41 (dt, J = 6.8,

2.4 Hz, 1 H), 6.25 (d, J = 15.6 Hz, 1 H), 4.06 (dd, J = 5.2, 2.4 Hz, 1 H), 3.89 (s, 3 H), 3.85 (s, 6

H), 3.82 (s, 1 H), 3.77 (s, 3 H), 3.76–3.73 (m, 1 H), 3.30 (d, J = 5.2 Hz, 1 H); 13C NMR (100

MHz, CDCl3) δ 195.4, 172.8, 166.6, 149.1, 148.8, 140.3, 137.1, 133.4, 131.0, 119.9, 119.6,

111.2, 99.4, 56.4, 55.9, 54.2, 52.1, 47.5, 45.1, 44.3.

Brominated diketone (40). To a solution of brominated Diels–Alder product 39 (0.071 g, 0.14

mmol) in CH2Cl2 (0.75 mL) was added water (0.15 mL) and TFA (0.60 mL). The resulting yellow mixture was vigorously stirred at 25 °C under ambient atmosphere for 16 h. Upon completion, the mixture was concentrated directly and the yellow residue was purified by flash column chromatography (silica gel, CH2Cl2:MeOH, 9:1) to give the brominated diketone 40

(0.060 g, 90%) as a yellow solid. 40: Rf = 0.18 (CH2Cl2:MeOH, 9:1); IR (film) νmax 3409, 2955,

2843, 1790, 1743, 1711, 1633, 1517, 1440, 1316, 1249, 1197, 1145, 1025, 981, 929 cm-1; 1H

NMR (400 MHz, CDCl3) δ 7.44 (d, J = 15.6 Hz, 1 H), 6.79 (d, J = 8.0 Hz, 1 H), 6.59 (s, 1 H),

6.46–6.41 (m, 2 H), 6.29 (d, J = 15.6 Hz, 1 H), 4.11 (br d, J = 2.0 Hz, 1 H), 3.88 (s, 3 H), 3.85 (s,

13 3 H), 3.84 (br s, 4 H), 3.32 (br d, J = 4.8 Hz, 1 H); C NMR (125 MHz, CDCl3) δ 173.4, 166.8,

149.1, 148.8, 139.9, 138.2, 133.1, 131.3, 120.2, 119.6, 111.3 (2 C), 56.0 (2 C), 55.0, 52.1, 48.3,

44.7. 214

Brominated diketone homodimer (41). Diketone 40 (0.105 g, 0.22 mmol, 1.0 equiv) was

dissolved in HFIP (0.055 mL). The resulting yellow mixture was stirred at 50 °C under ambient

atmosphere for 16 h. Upon completion, the mixture was concentrated directly, and the yellow

residue was purified by preparative TLC (CH2Cl2:MeOH, 92:8) to give the brominated dimer 41

(0.005 g, 5%) as a white solid. The product was crystallized from CHCl3/i-PrOH (colorless plates). 41: Rf = 0.47 (CH2Cl2:MeOH, 9:1); IR (film) νmax 3407, 2934, 1801, 1717, 1634, 1508,

-1 1 1490, 1330, 1249, 1200, 1166, 1030, 847 cm ; H NMR (500 MHz, CDCl3) δ 7.50 (d, J = 16.0,

1 H), 7.42 (d, J = 16.0 Hz, 1 H), 7.07 (s, 1 H), 7.02 (s, 1 H), 6.43 (s, 1 H), 6.38 (br s, 2 H), 6.21

(d, J = 16.0 Hz, 1 H), 6.12 (d, J = 16.0 Hz, 1 H), 4.62 (d, J = 8.0 Hz, 1 H), 4.18 (s, 1 H), 3.98 (d,

J = 4.5 Hz, 1 H), 3.87 (s, 6 H), 3.83 (s, 6 H), 3.76 (s, 3 H), 3.60 (s, 3 H), 3.68–3.66 (m, 1 H),

3.49 (d, J = 6.5 Hz, 1 H), 3.12 (d, J = 4.5 Hz, 1 H), 2.96 (d, J = 8.5 Hz, 1 H); 13C NMR (125

MHz, acetone-d6) δ 202.5, 171.7, 171.0, 166.6, 166.3 149.6, 149.4, 149.2 (2 C), 148.8, 148.5,

143.3, 140.6, 140.0, 139.8, 136.6, 135.0, 133.4, 132.0, 130.7 (2 C), 120.0, 119.5, 119.2, 116.2,

115.9, 115.8, 114.4, 113.6, 112.9, 111.5, 110.3, 105.0, 102.3, 55.6 (2 C), 55.5, 55.3, 55.2, 54.0,

53.7, 51.1, 50.9, 47.1, 46.7, 46.5, 45.8, 44.4, 44.1, 41.4, 40.0.

215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240