Efforts towards steroid natural products using a sequential Diels-Alder strategy

by

Jason Blair Crawford B. Sc., University o f Victoria, 1991

A Dissertation Submitted in Partial Fulfillment o f the Requirements for the Degree o f

DOCTOR OF PHILOSOPHY in the Department o f Chemistry

We accept this dissertation as conforming to the required standard

Dr. Claude Sniffe Supervisor (Departement de Chimie, Universite de Sherbrooke)

Dr. G erk$ A . Poulton, Departmental Member (Department o f Chemistry)

Dr. Peter C. Wan, Departmental Member (Department o f Chemistry)

Dr. Paul Rom^jniuk, Outside Member (Department o f Biochemistry/Microbiology)

Dr. Edward Piers, External Examiner (Department o f Chemistry, University o f British Columbia)

© Jason Blair Crawford, 1996 University o f Victoria

All rights reserved. This dissertation may not be reproduced in whole or in part, by photocopying or by other means, without the permission o f the author.

ABSTRACT

A novel strategy has been developed for the generation o f the perhydrophenanthrene skeleton through the use o f sequential Diels-AJder reactions on a 1,3,7,9-tetraene. This strategy ailows for the generation o f the equivalent o f the steroidal A/B/C ring-system, in an efficient and stereoselective manner. A similar strategy, also involving sequential Diels-Alder cycloaddition reactions, was employed in the attempted synthesis o f a steroid natural product.

Examiners:

Dr. Claude Spjkfo. Siio^rvisor (Departement de Chimie, Universite de Sherbrooke)

Dr. GeraldA. Poulton. Denartmental Member (Department o f Chemistry)

Dr, Peter C. Wan, Departmental Member (Department o f Chemistry)

Dr, Paul Roi^aniuk, Outside Member (Department o f Biochemistry/Microbiology)

Dr. Edward Piers, External Examiner (Department o f Chemistry, University o f British Columbia)

P age#

Title Page i

Abstract ii

Table o f Contents iii

List o f Tables vi

List o f Figures vii

List o f Schemes ix

List o f Spectra xi

Acknowledgements xiii

List o f Abbreviations xv

Chapter One: Introduction: Page #

1.1: Steroids: General Features, Functions and Historical Perspective 1

1.1.1; Cholesterol 1

1.1.2: Bile Salts 2

1.1.3: Cardiac Aglycones and Sapogenins 3

1.1.4: Sex Hormones and Corticosteroids 4

1.2: Biogenesis o f Steroids 6

1.3: Biological Activity and Clinical Uses o f Androgenic Steroids 11

1.3.1: Biological Activity o f Androgens 11

1.3.2: Clinical Uses o f Androgens 13

1.4: Synthetic Strategies Towards Androgens and other Steroids 16 1.4,1 • Biomimetic Carbocation-Polyolefin Cyclization Approach 16 1.4.2: Aldol or Robinson Annulation Approach 20

Chapter Two: Results and Discussion: Page #

2.1: Retrosynthetic Analysi s 39

2.2: Model Diene Synthesis 41

2.3; Synthesis and Evaluation o f Dienophiles 44

2.3.1: Carbomethoxybutadiene Studies 44

2.3.2: Acrolein Studies 48

2,3.3: Enyne Studies 54

2.4: Research Towards New Acyclic Bis-Dienes 68

2.5: Synthesis o f Bis-Dienes Incorporating the D-Ring 74 2.5.1: 2-Methyl-cyclopentane-1,3 -dione Studies 75 2.5.2: Studies Involving Cycloisomerization Reactions 78 2.5.2.1: Generation o f Cycloisomerization Precursor 78

2.5.22: Cycloisomerization Reaction 81

2.5.2.3: Attempts at the Generation o f the Second 90 Diene Fragment

2.6: Alternate Sequential Diels-Alder Strategy Using Cyclo- 99 Isomerization Product as ‘First’ Diene

2.6.1: Diels-Alder Reaction Between 118 and MVK 103 2.6.2: Subsequent Modification o f the Bicyclic Cycloadduct 106 2.6.3: Attempted IMDAC Using the Newly Generated Diene 119

Chapter Four: Experimental 129

References 182

Appendix 1: 195

A 1.1: 2-Carbomethoxybutadiene as a Diene 195

A 1.1.1: Determination o f Enophilicity o f 45 195 A 1.1.2: Reactivity o f Structural Analogs o f 45 201

Chapter Two: Paee # Table 2-1: Preliminary Attempts at the Generation o f 118 from 111 83

via a Palladium Catalyzed Cycloisomerization

Table 2-2: Attempts at the Conversion o f 111 to 118 using 119 and 86 HOAc as a Catalytic System

Table 2-3: Optimization o f Cycloisomerization o f 111 Using 88 Pd(OAc)z/BBEDA as a Catalytic System

Table 2-4: Predicted Energetic and Dihedral Angle Values for 108 Cycloadducts 143 and 144

Table 2-5: Attempted Conditions for IMDAC o f 157 121

Appendix One: Paee #

Table A -l: Tlsermal Reaction o f 45 With Various Dienes 197 Table A-2: RT Reaction o f 45 With Various Dienes 198 Table A-3: Reaction o f 45 and 165 with Maleic Anhydride 199 Table A-4: Reactivity o f Amide Analogs o f 45 Diels-Alder 203

Reaction with 41

Appendix Two: Paee #

Chapter One: Page # Figure 1-1: Cholesterol, Showing Conventional Carbon Numbering 1

and Ring Nomenclature

Figure 1-2: Digitonin: a steroidal glycoside 3

Figure 1-3: Prediction o f FMO Coefficients for Dienes and 27 Dienophiies

Figure 1-4: Possible Transition States for IMDAC o f 25 35 Figure 1-5: Potential Steroidal A-Rings From a Common Synthetic 36

Intermediate

Figure 1-6: 5a-Dihydrotestosterone, a Potential Synthetic Target 38

Chapter Two: Paae #

Figure 2-1: Potential Dienophile Candidate Molecules 49 Figure 2-2: Transition States for IMDAC Reactions 53 Figure 2-3: Potential Steric Congestion Between Angular Methyl 54

Group and Pendant Diene in IMDAC exo Transition State

Figure 2-4: Structural and Functional Group Requirements o f Enyne 55 Bis-Dienophile

Figure 2-5: ORTEP Diagram o f Major IMDAC Product 68 59 Figure 2-6: 'H NMR Spectrum (expansion) and 'H COSY Spectrum 65

o f M ajor IMDAC Product 69

Figure 2-7: N M R Spectral Evidence for Proposed Structure o f 69a 65 Figure 2-8: Alternate Catalytic System 119 and Ligand 120 for 84

Cycloisomerization Reaction

Figure 2-9: Possible Facial Orientations for DAC o f 118 with MVK 101 Figure 2-10: Portions o f the 'H -i3C Correlated Spectrum and 'H COSY 104

Spectrum o f Cycloadduct 143

Modelling Calculations ,

-Figure 2-12: Expected Transition State Geometry for IMDAC 121

Appendix One:

Chapter One:

Scheme # Paee # Scheme # Pace

ti-Scheme 1-1 3 Scheme 1-13 23 f Scheme 1-2 5 Scheme 1-14 24 Scheme 1-3 6 Scheme 1-15 25 Scheme 1-4 7 Scheme 1-16 28 Scheme lr5 8 Scheme 1-17 29 Scheme 1-6 10 Scheme 1-18 30 Scheme 1-7 17 Scheme 1-19 31 Scheme 1-8 18 Scheme 1-20 32 Scheme 1-9 19 Scheme 1-21 32 Scheme 1-10 20 Scheme 1-22 34 Scheme 1-11 21 Scheme 1-23 37 Scheme 1-12 22 Chapter Two:

Scheme # Paee # Scheme # Paee H

Scheme 2-1 39 Scheme 2-11 57 Scheme 2-2 41 Scheme 2-12 58 Scheme 2-3 42 Scheme 2-13 60 Scheme 2-4 45 Scheme 2-14 62 Scheme 2-5 46 Scheme 2 -15 63 Scheme 2-6 47 Scheme 2-16 63 Scheme 2-7 50 Scheme 2-17 68 Scheme 2-8 51 Scheme 2-18 69 Scheme 2-9 52 Scheme 2-19 69 Scheme 2-10 56 Scheme 2-20 70

Scheme # Paee # Scheme 2-21 71 Scheme 2-41 96 Scheme 2-22 71 Scheme 2-42 97 Scheme 2«23 72 Scheme 2-43 97 Scheme 2-24 73 Scheme 2-44 100 Scheme 2-25 74 Scheme 2-45 102 Scheme 2-26 75 Scheme 2-46 103 Scheme 2-27 76 Scheme 2-47 106 Scheme 2-28 77 Scheme 2-48 109 Scheme 2-29 78 Scheme 2-49 110 Scheme 2-30 77 Scheme 2-50 111 Scheme 2-31 79 Scheme 2-51 112 Scheme 2-32 80 Scheme 2-52 113 Scheme 2-33 82 Scheme 2-53 114 Scheme 2-34 87 Scheme 2-54 115 Scheme 2-35 90 Scheme 2-55 117 Scheme 2-36 91 Scheme 2-56 118 Scheme 2-37 92 Scheme 2-57 119 Scheme 2-38 93 Scheme 2-58 120 Scheme 2-39 94 Scheme 2-59 123 Scheme 2-40 95 Scheme 2-60 126 Appendix One:

Scheme # P a e e # Scheme # Paee

Scheme A -1 195 Scheme A-4 200

Scheme A-2 197 Scheme A-5 202

S n ect'i' ... . P aso #

‘H NMR (250 MHz) and F:.,TR spectra o f 37 206

*H NMR (360 MHz) and 13C l4MR o f 38 207

lH NMR (360 MHz) and 13C NMR o f 40 208

'H NMR (360 MHz) and IR spectra o f 41 209

13C NMR and DEPT spectra o f 41 210

'H NMR (90 MHz) and IR spectra o f 44 211 ’H NMR (360 MHz) and 13C NMR spectra o f 46 212 lH NMR (250 MHz) spectrum o f 47 213 lH NMR (360 MHz) and IR spectra of 55 214 ‘H NMR (3C0 MHz) and IR spectra o f 56 215 'H NMR (250 MHz) spectrum o f 57 216 'H NMR (360 MHz) and IR spectra o f 58 217 'H NMR (360 MHz) and IR spectra o f 59 218

l3C NMR and DEPT spectra o f 59 219

‘H NMR (360 MHz) and IR spectra of 60 220 ‘H NMR (360 MHz) and IR spectra o f 61 221 ‘H NMR (360 MHz) and IR spectra o f 63 222 *H NMR (360 MHz) and IR spectra of 65 223 ‘H NMR (360 MHz) and IR spectra o f 66 224 'H NMR (360 MHz) and IR spectra o f 67 225 ‘H NMR (360 MHz) and IR spectra of 68 226

,3C NMR and DEPT spectra o f 68 227

]H NMR (360 MHz) and IR spectra o f 69 2.78

13C NMR and DEPT spectra o f 69 229

NOESY (top) and COSY spectra o f 69 230

'H NMR (360 MHz) and IR spectra o f 107 23)

'H NMR (360 MHz) and IR spectra o f 109 233

’H NMR (360 MHz) and IR spectra o f 110 234

‘H NMR (360 MHz) and IR spectra o f 111 235

'H NMR (250 MHz) and 13C NMR spectra o f 112 236

’H NMR (300 MHz) and IR spectra o f 118 237

l3C NMR and DEPT spectra o f 118 238

’H NMR (300 MHz) and IR spectra o f 138 239

'H NMR (300 MHz) and IR spectra o f 143 240

l3C NMR and DEPT spectra o f 143 241

‘H /i3C Correlated (top) and COSY spectra o f 143 242

'H NMR (300 MHz) and IR spectra o f 144 243

'H NMR (300 MHz) and IR spectra o f 145 244

'H NMR (300 MHz) and IR spectra o f 146 245

1?C NMR and DEPT spectra o f 146 246

'H NMR (300 MHz) and IR spectra o f 152 247

During the course o f my degree, I feel very fortunate to have had the unwavering support o f my friends and family. I am confident that, without this support, the project would have been much less productive and much less enjoyable. As such, I owe a great deal o f thanks to my parents, my brothers and sisters, and their families. Also, I wish to thank my buddy Geoff, for making sure I didn’t take myself too seriously, and, o f course, a particularly heartfelt thanks to Kathy, whose constant encouragement and caring support was greatly appreciated.

At the University, I was also very fortunate to have had the opportunity, during the course o f my research, to consult with many co-workers, professors, and other graduate students, whose advice, in many cases was extremely helpful. I wish to thank the following for their help and advice: Noah Tu, Gang Liu, Eric Fillion, Brian Eastman, Rob Gossage, Rich Hooper, Dr. ^hom as Fyles, and Dr. Peter Wan.

O f course, the degree required that I would have to obtain a wide variety o f spectral data which, when I couldn’t obtain them myself, were provided by the following, to whom I owe a great deal o f thanks: Mrs. Christine Greenwood (NMR, UVic), Dr, Dave MacGillivray (MS, UVic), Mr. Lea Sohallig (MS, UVic), Dr. Norman Pothier (NMR, Sherbrooke), Mr. Gaston Boulet (MS, Sherbrooke) and Mr. Marc Drouhin (X- Ray, Sherbrooke).

I also owe a great deal o f thanks to Dr. Claude Spino. I feel very lucky to have been associated with a supervisor whose enthusiasm for the project, and for organic chemistry in general provided an atmosphere which allowed for a great deal o f enjoyable learning to take place. I will forever be indebted to Claude for his unselfish sharing o f knowledge and time, which allowed me to develop a much greater understanding and appreciation o f synthetic organic chemistry than !’d anticipated when L started the project.

providing me with funding for my degree. I feel very lucky and appreciative to have had the benefit o f tw o particularly generous scholarships from the University. Also, I wish to thank NSERC for funding Claude’s research, and thus largely allowing for such research to take place at Canadian Universities.

AcCl: Acetyl Chloride

AIBN: Azo-rso-Bis-Butyronitrile

BBEDA: Bis-(benzylidine)ethylenediamine Bn: Benzyl

Bu: Butyl

CDAC: Cross Diels-Alder Cycloaddition CoA: Coenzyme A

COSY: Homonuclear Shift Correlated Spectroscopy DAC: Diels-Alder Cycloaddition

DHEA. Dehydroepiandrosterone

DEPT: Distortionless Enhancement by Polarization Transfer ( ‘H/I3C in this case) An,n: Unsaturation Between Carbon # ’s n and n’

E: Ester

ee: Enantiomeric Excess

ERG: Electron Releasing Group Et: Ethyl

EWG: Electron Withdrawing Group FMO: Frontier Molecular Orbital hGH: Human Growth Hormone HMDS: Hexamethyldisilazane

HOMO: Highest Occupied Molecular Orbital IMDAC: Intramolecular Diels-Alder Cycloaddition IR: Infra Red

LAH: Lithium Aluminum Hydride LDA: Lithium Diisopropylamide

LUMO: Lowest Unoccupied Molecular Orbital Me; Methyl

MVK: Methyl Vinyl Ketone nBuLi: n-Butyllithium NBS: N-Brornosuccinimide

NMR: Nuclear Magnetic Resonance nOe: Nuclear Overhauser Enhancement

NOESY: Homonuclear Nuclear Overhauser Enhancement Correlated Spectroscopy (2D) Ph: Phenyl Pr. Propyl RT: Room Temperature TBDMS: ter/-Butyldimethylsilyl Tf; Trifluoromethanesulfonyl THP: Tetrahydropyran

INTRODUCTION

1.1: General Features, Functions, and Historical Perspective:

Sterols are a diverse family o f modified triterpenoids which are present in most eukaryotic cells. Originally, they were named as the collective group o f solid alcohols obtained from nor. saponifiable portions o f lipid extracts o f tissues': the name itself is based on the Greek word steros, which means 'solid'. Structurally, a tetracyclic ring skeleton (perhydrocyclopentenophenanthrene) is common to all steroids (see Figure 1-1), and the wide ranging biological activities o f these molecules is a result o f the differing functionalities o f the substituents on the rings and the varying degree o f unsaturation o f the ring skeleton.

22 24 _ 23 s H

HO

Figure 1-1: Cholesterol, Showing Conventional Carbon Numbering and Ring Nomenclature

1.1.1: Cholesterol:

Cholesterol, the most abundant steroid in mammalian systems (approximately 2-3g per kg body weight in humans), was first isolated in the early 1800's: the discovery generally being accredited to Michel Eugene Chevreul in 1812.' Although, in North

American society, cholesterol is perhaps thought o f mainly as being a major cause o f heart disease through the formation o f atheroscopic plaques (likely a result o f a high-fat diet), cholesterol is in fact an important and essential compound for survival. Through integration into the phospholipid bilayer o f cell membranes (via hydrophobic association o f the non-polar part o f the cholesterol molecule with the lipid chains and hydrogen bonding o f the cholesterol hydroxyl group to the fatty acid derived ester carbonyl group) cholesterol plays an important role in the mediation o f membrane fluidity. By preventing close, ordered association o f the fatty acid acyl chains, cholesterol serves to prevent crystallization o f the membrane. And, through hydrogen bonding to the acyl chains, cholesterol also hinders rapid movement o f the individual phospholipid groups. Thus,

2

cholesterol prevents the membrane from becoming too fluid or too solid in nature.

1.1.2: Bile Salts:

Shortly after the discovery o f cholesterol, lithocholic acid, a steroidal bile acid, was isolated from ox bile (in 1828 by Leopold Gmelin'). Lithocholic acid (Scheme 1-1), and some twenty other structurally related steroid-based acids, are generally found in the body as amide-acids that are formed through the condensation o f the C24 acid functionality o f the parent steroid with the amino functionality o f an a-am ino acid (often glycine or taurine: see Scheme 1-1): o f course, in the intestine, they exist as the sodium salts o f the respective acids. Through their amphipathic, detergent-like nature, they are capable o f emulsifying fats in the digestive tract into micelles, which enables transport through the

2 small intestinal wall.

HO'

HO'

HOoC

1.1.3: Cardiac Aglycones and Sapogenins:

The next major family o f steroids to be discovered were the cardiotonic glycosides. Perhaps the most well known o f this group is digitonin, which is isolated from the purple foxglove. It exists in nature as a steroidal glycoside: a condensation o f saccharide units with the steroidal skeleton. In the case o f digitonin, a pentasaccharide chain is attached to the oxygen on C3 (Figure 1-2). Although the structure o f the aglycone (without the pentasaccharide) o f digitonin was not elucidated until 1935* the cardiotonic nature o f digitonin was well known in the 19th century. In very small doses (0.1 mg per day),

I 2

digitonin is an effective treatment o f congestive heart failure. ’ A related class o f compounds, the sapogenins (so named because o f the soapy nature o f their aqueous solutions) were also discovered at approximately the same time.'

H O

-(xylose)(galactose)5(glucos9)20 '

^

1.1.4: Sex Hormones and Corticosteroids:

In the 1930's, a new class o f steroids were isolated: the se.: hormones, As was the case with the cardiotonic glycosides, the knowledge o f biologically active compounds extracted from sexual organs (usually dog or bull testicles) preceded the isolation and structural elucidation o f the compounds themselves by several decades. Perhaps the first well documented case o f use o f steroid-based hormones was by Charles Edouard Brown- Sequard, whom in 1889 made the bold claim that by injecting himself with liquid extract o f dog and guinea pig testicles he had reversed his own aging process: an increase in physical strength and intellectual energy were two o f the claimed benefits.1 In 1911, A. Pezard found that the comb o f a male capon grew in direct proportion to the injected dose o f animal testicular extract: perhaps the first documented case with verifiable scientific results. Two decades later, in 1931, A dolf Butenandt isolated 15 mg o f androsterone from 15,000 litres o f male urine. Soon thereafter, in May, 1935, testosterone was also isolated from urine by K.G. David, E. Laqueur and their colleagues. Perhaps more interesting to the synthetic chemist were tw o subsequent syntheses, completed later in the same year, o f testosterone from cholesterol by Butenandt (and co-workers) and by Leopold Ruzicka and A. Wettstein: an achievement for which Butenandt and Ruzicka received the Nobel Prize for Chemistry in 1939. At approximately the same time, estrone and estradiol were isolated from ovaries o f cattle, and some o f the corticosteroids were isolated from the adrenal cortex. Much later, the metabolic origin o f these steroid

2 3

pregnenolone

HO

OH

te sto ste ro n e _progesterone

OH

V

, OH

HO

1.2: Biogenesis of Steroids:

Due to intense scientific interest in steroidal metabolism and biogenesis (particularly cholesterol), a great deal is known about the biological origin o f steroids. In fact, the genesis o f cholesterol and other steroids can be traced all the way back to the individual acetate units ^ As shown in Scheme 1-3, the first step is the generation o f the individual isopentenyl pyrophosphate units, which will later malm up the steroidal skeleton. These units are biosynthesized from mevalonic acid, which, in turn, has

acetyl-. acetyl-. 2-4

coenzyme A as its origin.

A

SCoA

OH v OH 9

OH

H3O6P2Q

H3O6P2O

As shown in Scheme 1-4, the next stage in the biogenesis o f steroids is the condensation o f the five-carbon isoprenoid units into squalene. The newly formed squalene molecule is referred to as a triterpene, as it is created from three ten-carbon terpene units (which each consist o f tw o isoprene units), At this point the carbon skeleton is large enough to create the steroidal skeleton, and in most animal systems, the triterpenoids are the largest natural terpene-based molecules, As a side note, some

forty-carotenoid family o f compounds.

H'

2

6

3

I

2

6

3

^ \ ^ O P 2 0 6H3

Following the generation o f squalene, an enzyme, squalene monooxygenase, serves to epoxidize the C2-C3 bond o f the squalene. The molecule will then undergo a stereospecific cyclization, undoubtedly utilizing another enzyme, which will provide a mode o f acid catalysis and a molecular geometry restriction, to give an intermediate

carbocation.2-* N ote that, as shown in Scheme 1-5, the three six membered rings (A, B and C respectively) are in a chair-boat-chair type array. The intermediate carbocation will undergo a stereospecific rearrangement (made possible by the axial nature of the migrating substituents) to give, as a product in animal systems, lanosterol. A similar cyclization

2,3

occurs in plants and fungi to give stigmasterol and ergosterol respectively. Note that the stereochemistry o f the cyclization and subsequent rearrangement is the cause o f the stereochemical arrangement o f all the ring junctions and substituents. Following the

2 -4

generation o f lanosterol, several metabolic modifications take place to give cholesterol.

S q u alen e

sq u alene-2,3-epoxide

H* / E nzym atic Control

lanosterol

As shown previously in Scheme 1-2, cholesterol (from lanosterol) is the biogenic source (in animals) for the six major classes o f steroids: sterols, sapogenins, cardiac aglycones, bile acids, adrenal steroids and sex hormones. This report will focus on the sex hormones, and more specifically the androgenic male sex hormones. The conversion o f cholesterol to this class o f compounds occurs mainly in the testis, but also occurs in lesser amounts in the adrenal cortex and the ovaries. Cholesterol is first converted to

M

pregnenolone, which is then converted to progesterone. From here, a variety o f androgens can be generated: the term androgen encompasses a group o f male sex hormones which are responsible (at least in part) for the development of secondary sexual characteristics.

-C H 3 C 0 2

H

H

androsterone _{an d ro stan ed io n e} testosterone

androstane-3a, 17|Vdiol androstane-17p-ol-3-one (dihydrotestosterone) Scheme 1-6

As shown in Scheme 1-6, progesterone is hydroxylated at the C l 7 position, which can yield, after oxidation, androstenedione. From this point, testosterone, the most abundant o f the human androgens, can be generated, Several other androstane-based steroids can also be generated through subsequent metabolic modification o f testosterone (see Scheme 1-6).1,2,4

1.3: Biological Activity and Clinical Uses of Androgenic Steroids:

The androgenic steroids are responsible for two majOi types o f biological activities: the androgenic effect (masculinization) and the anabolic effect.' As is well known to the public, the use o f anabolic steroids (synthetic derivatives o f testosterone, usually having an acyl chain attached to the oxygen on C l 7) as performance enhancing drugs provides athletes, both male and female, with a rapid degree o f muscular development that may be accompanied by a variety o f side effects. Recent estimates suggest that the number o f people abusing anabolic steroids in the United States of America may be close to one million.'

1.3.1: Biological Activity o f Androgens:

The body, both male and female, naturally produces testosterone and other androgens throughout life. These compounds are very effective, sometimes exerting their

-9 5,6

desired effect at concentrations as low as 10 moles/litre. The natural level o f androgen excretion is governed largely by peptide-based hormones which are excreted from the

2

pituitary and hypothalamus glands (including human growth hormone and gonadotropin ). When a certain level o f these hormones is detected by the extracellular receptors o f the target organ, which, in this case would be the testis, ovaries or the adrenal cortex, a cascade o f intracellular events occurs which provides for the synthesis o f greater levels o f the androgens. In the case o f testosterone, the major site o f biosynthesis appears to be the Leydig cells in the testis.5

The androgens themselves are then secreted into the bloodstream. Solubility o f these lipid-based molecules in the aqueous extracellular environment is very low; as a result, steroids are sometimes secreted to the bloodstream as sulfates o r glucuronides.1

Another biologically operative mode o f solubilizing the androgens (and other steroids) in

2 5 7 8

the bloodstream is through association with water-soluble plasma proteins.

Once the androgens reach the target organ, they must cross the cell membrane in order 10 exert their desired effect. This can happen in three ways: attachment o f the 'free'

steroid to an extracellular receptor and subsequent incorporation into the cytoplasm, diffusion o f the free steroid across the cell membrane, or 'ingestion' o f a lipoprotein

5,8

complex which contains the steroid. Through either mode, the steroid is able to gain access to the cytoplasm, which will contain a specific protein-based receptor for the androgen. The next step after formation o f the steroid-receptor complex is migration o f

5 8

the complex to, and eventually through the nuclear membrane. ’ Once inside the nucleus, the steroid-receptor complex is capable o f binding specific segments o f chromatin (DNA

5 8

strands) which enhances the rate o f transcription o f certain genes. ’ Following the transcription, the newly formed RNA, through a process o f translation, will generate new proteins and enzymes which will exert their biological effects within the c e l l/8 Eventually, the intranuclear steroid-receptor complex presumably dissociates or is biologically degraded to terminate the increased rate o f transcription.

On a macroscopic level, the androgenic effects o f the steroids are those which result in a development o f the reproductive tract and growth o f facial and body hair. The anabolic effects o f these androgens are those which don't take place in the somatic and reproductive tract tissue. Such effects are an acceleration in growth, with a concomitant decrease in the level o f body fat, and also an enlargement o f the larynx and thickening o f the vocal cords. Perhaps the most well known anabolic effect is the increase in muscle bulk and strength.

Historically, the biological efficacy o f various androgens was measured as a function o f capon comb development as a function o f injected dose. Another similar experiment measured the anabolic efficacy o f androgens by evaluating the increase in nitrogen retention (detected through measurement o f concentrations in urine; increased nitrogen retention being related to increased protein synthesis) as a function o f dose . Through these tests, it was determined that testosterone was one o f the most effective androgenic steroids.' At a later date, the blood plasma levels o f the various androgens could be measured, and the following levels were found in the average adult male5 (expressed as pg per 100 mL): testosterone, 0.7; dehydroepiandrosterone, 0.5; androstenedione, 0.1; 5a-dihydrotestosterone, 0.05. Through the use o f radiolabelled steroids in tissue binding studies, and subsequent recovery o f the intracellular steroids, it was found that the major intracellular metabolite (and strongest binding steroid to the

S 8

intracellular receptor) was 5a-dihydrotestosterone. ’

1.3.2: Clinical Uses o f Androgens:

The clinical uses o f androgens (and anabolics) are quite varied, Since the androgens are responsible for the development o f secondary sexual characteristics, they have often been used as a puberty and growth stimulating factor in boys who are experiencing a significant developmental delay.' Androgens are also used, in conjunction with the proteinacious human growth hormone (hGH), to initiate growth in children who

1.2,5

are hGH deficient. Another use o f the androgens has been for recuperative treatment of chronic debilitating conditions such as those experienced by those who have recently been

9

burned, had surgery, radiation therapy or chemotherapy.

As a result o f the pote. ’t effect o f the androgens on the reproductive tract, they have been heavily studied as a mode o f contraception in both men and women. In fact,

most female oral contraceptives are based on varying ratios o f progesterone and testosterone derivatives (with 17a-et.hynyl-19-nor-testosterone (norethindrone) being one

. 1 .4

o f the most widely used androgen derivatives). In males, testosterone has also been 9

examined as a potential contraceptive. The hypothalamus gland reacts to high levels o f plasma testosterone by reducing the release o f leutinizing hormone-releasing hormone,

9

which, in turn, lowers the production o f sperm. Although such a contraceptive has not yet become widely available, testing has been conducted on humans through the World

9

Health Organization, the results o f which have shown that the treatment is effective.

Perhaps one o f the most interesting clinical potentials o f testosterone and other androgens are as 'anti-aging' compounds. From the time that Brown-Sequard injected himself with the testicular extracts and claimed their various benefits, people have been enamoured with the idea o f the androgens' potential ability to retard (or, even more

9

ambitiously, reverse) the effects o f aging. More recent experiments have used testosterone derivatives, in some cases in the conjunction with hGH, in men over the age o f 54 who had low to normal levels o f natural testosterone: the positive results included an

9

increase in lean body mass and strength and a better spatial perception and word memory. Another, more recent test carried out with eight men and eight women over the age o f 50 at the University o f California, San Diego, used dehydroepiandrosterone (DHEA), The goal o f the experiment was to examine the effects o f restoring DHEA levels to peak levels (usually experienced between the ages o f 25 and 30) on the patients.10 The reported mode o f action o f tiiis steroid is to increase the amount o f insulin-like growth factor 1, which is involved with the regulation o f cellular metabolism and the immune system. In males, the steroid was also found to activate natural killer cells, which are involved in the immune response. Although a well-controlled clinical study involving a large population has not

yet been conducted to verify the results, the sixteen patients in the preliminary study reported an increase in physical and psychological well being.'0

Perhaps, most notoriously, androgenic and anabolic steroids are well known for their anabolic effects. Athletes, and others, both male and female, have used anabolic steroids to increase muscle mass and strength for enhanced appearance or performance since the late 1940's.'° Unfortunately, the purchase and administration o f these steroids, which are sometimes intended for livestock, are often conducted through an international black market which exists largely due to the huge demand (roughly $1-billion US per year) for such drugs. Unfortunately, these drugs are not without side effects. In women, the excess testosterone (and metabolites) can lead to a degree o f masculinization (increased amounts o f facial and body hair, deepening o f voice and increased clitoral size), In men, the increased androgenic levels can lead to decreased sperm p ro d u c tio n .A n o th e r side effect for men, which is well known amongst bodybuilders, is an enlargement o f the nipples and an increase in amounts fatty tissue around the nipple (essentially the early development o f a female-like breast): this is likely a result o f the effects the female sex hormones estrogen and estradiol, which are two o f the metabolites o f testosterone, Another side effect o f excess testosterone, for both men and women, is the development o f pattern baldness on the scalp, which again is a result o f one o f the metabolites o f

5 8

1.4: Synthetic Strategies Towards Androgens and Other Steroids:

With such a wide and varied degree o f biological activity (including those o f clinical, and therefore economical importance), and with such a relatively complex structure, the steroid family has long been a desired synthetic target for chemists. Early experiments using exhaustive oxidative degradation were followed by selenium dehydrogenation in the 1930's, which allowed for the structural elucidation o f many steroids*. By the late 1940's, the stereochemistry o f the steroids had been largely determined, which, along with the determination o f cyclohexane conformation (and axial and equatorial substituents) in 1950,'* provided the necessary structural information to allow for steroidal synthesis.

Many o f the first syntheses o f androgens were based on modifications o f existing steroids (often cholesterol). These partial syntheses were mainly involved with functional group manipulations, and didn't have to deal with the significant problem o f stereoselective genera tion o f the carbocyclic ring skeleton.

Attempted total syntheses o f steroids soon followed, and could largely be broken up into three basic classes (with various sub-classes) based on the strategy chosen for the generation o f the carbocyclic skeleton: (1) those using a biomimetic, carbocation-based 'cascade'; (2) those using condensation reactions, such as the Aldol or the Robinson Annulation; (3) tho ve using pericyclic reactions, such as the Diels-Alder reaction.

1.4.1: Biomimetic Carbocathn-Polyolefin Cyclization Approach:

Chemists have sought to mimic the biological cyclization o f squalene-2,3-oxide (to

12-14

tetracyclic skeleton is generated 'in one step' in a stereoselective, and in biological systems enantioselective manner provides a very attractive target to the synthetic organic chemist, Unfortunately, there is one fundamental problem with this method if squalene epoxide is used as a starting material. As shown in Scheme 1-7, in which the cyclization is shown in a stepwise manner, a relatively unstable secondary carbocation is generated as an intermediate. Without the stabilization which the biological enzyme presumably provides, the cyclization cascade will not lead to lanosterol.

ctS^

<rv

In fact, the first attempted cyclizations o f squalene-2,3-oxide yielded a tricyclic product which resulted from the formation o f a tertiary carbocation at C14 rather than a

12

secondary carbocation at C13. Thus, the 'C'-like ring will be five-membered, and the fourth ring (the D-ring) won't form. This pathway, which gives two different tricyclic alkenes (based on the elimination o f two different protons), is outlined in Scheme 1-8. Unfortunately, experimentation with different Lewis acids, solvents, temperatures and other variables was unsuccessful in altering the regiochemistry o f the cyclization in

non-enzymatic laboratory systems. As a side note, such a pathway has been successfully utilized to generate natural products such as malabaricanediol.'3

Lewis Acid _{L A O}

0

In order to generate a steroid-like structure utilizing the carbocation-polyolefin cascade type pathway, the difficulties associated with the secondary carbocation must somehow be circumvented. The most obvious and common way that this is achieved is through the use o f a starting material in which the five-membered ring (the D-ring) and sometimes also the attached six membered ring (the C-ring) are already present. Two examples o f such a strategy are provided in Scheme 1-9. The first uses a molecule in which the five membered ring is present in the starting material (to give the isoeuphenol type system)15, and the second uses a starting material in which the D-ring is also present (to give the lanosterol-type system).'6 N ote the variance in the B/C ring junction stereochemistiy in the two examples resulting from the differing geometries o f cyclization.

R D Lewis Acid HO isoeuphenol system Lewis Acid

HO'

Although these early biomimetic investigations did not lead immediately (as hoped)

14

to lanosterol, a later study did in fact produce lanosterol from a biomimetic pathway in which the C- and D-rings were present in the starting material (Scheme 1-10). It is important to note, however that the cyclization did not yield lanosterol directly. Following cyclization, the alkene was present between carbon atoms 7 and 8; therefore, a short series o f alkene migrations were required to incorporate the alkene at the correct location (at the B/C ring junction, carbons 8 and 9). Although the olefm-carbocation cyclization route, owing to its potential simplicity, still remains a very attractive route to the steroid skeleton, to date, synthetic chemists have unfortunately been unable to

14

duplicate the economical effectiveness o f the single, natural 83-kDa enzyme by achieving the conversion from squalene-2,3-oxide to lanosterol in a single step. Although research in this area is still being conducted, many alternative routes have been developed which allow for better, more predictable control o f the synthesis o f the steroidal skeleton.

Lewis Acid

AcO’

R

HO1

Scheme 1-10

1.4.2: Aldol o r R obinson A nnulation A pproach:

The first total synthesis o f a steroid skeleton, conducted in 1951 by R.B. W oodw ard,'7 did not use a biomimetic carbocation-olefin cyclization route. In fact, the synthesis can be thought o f as a 'hybrid' route in that it uses both a Diels-Alder reaction and two Robinson Annulation reactions to generate the tetracyclic molecule. However, since the generation o f the A/B and the B/C ring junctions are a result o f the Robinson Annulation reactions, the synthesis should provide an effective example o f the condensation-type synthetic strategy towards the steroids.

As shown in Scheme 1-11, the synthesis commences with a Diels-Alder cyclization (which will be discussed in more detail in section 1.4.3) between butadiene and methoxyltolylquinone to give the bicyclic species 1. Subsequent epimerization o f the ring junction stereochemistry with acid and lithium aluminum hydride reduction o f the dione functionality yields the bicyclic ketone 2. Following removal o f the hydroxyl functionality

18

(with zinc and acid), and condensation (via a Robinson Annulation (see Scheme 1-12)), the tricyclic ketone 3 is obtained. These three rings, from left to right, will constitute, after suitable modification, the B,C, and D-rings o f the steroid.

h eat

2 . LAH O

HO

2. ethyl vinyl Ketone h3c o

0

Scheme 1-11

18

The Robinson Annulation using ethyl vinyl ketone (or similar compounds) has proven to be an efficient mode o f incorporating a six membered ring with a ketone functionality (other methods will be discussed later). As shown in Scheme 1-12, the reaction proceeds in two basic steps, the first being a 1,4-type addition (to generate 2a) and the second being an Aldol condensation (to give 2b), Following dehydration (which is often spontaneous), the a,p-unsaturated ketone 3 is generated, In this synthesis, the newly-formed cyclohexenone ring will become the steroid B-ring. At this stage, two major steps are required to generate the steroidal skeleton: incorporation o f the A-ring, and modification o f the D-ring to give a five, rather than a six-membered ring, The former is achieved through the application o f another Robinson Annulation, As shown in Scheme 1-13, after modification o f 3 (oxidation o f the alkene on the D-ring and subsequent

protection o f the diol then hydrogenation o f the alkene on the C-ring) to give 4, the enolate o f the ketone is reacted with an aniline derivative (in a three step sequence) to give S, which serves the iiinction o f preventing enolization (and subsequent reaction) at that site during the following Robinson Annulation. Thus, when 5 is treated with a base and acrylonitrile, the reaction proceeds regiospecifically. Hydrolysis o f the nitrile functionality (and the protected site a -to the B-ring ketone), and subsequent reaction o f the ethyl ester (formed from the hydrolysis and subsequent esterification o f the nitrile) with methylmagnesium bromide, allows for the Aldol reaction to occur and affords the tetracyclic compound 6. b a s e 1,4-addition

“O'

Note that the stereochemistry o f the newly formed ring junction is shown as a single isomer: in the actual synthesis, the Robinson Annulation reactions yield both isomers. The undesired isomers o f 6 and 3 were discarded, and only the compounds with

the correct stereochemistry were carried through to the next steps. This property o f the Robinson Annulation is unfortunate, as it does not allow for a stereoselective formation o f the desired compounds. For a stereoselective synthesis, unless the 'wrong' stereoisomer can be converted to the 'right' stereoisomer, the overall synthetic efficiency o f the pathway will suffer since a significant fraction o f the material must be discarded at each Annulation step.

NAr

///

N

The last major sequence in the synthesis is involved with the conversion o f the six- membered ring at the upper right side to a five-membered steroidal D-ring, This was achieved reasonably simply through the employment of periodic acid, which will deprotect and oxidatively cleave the diol to give the dialdehyde 7 (thus opening the 'D' ring). As seen in Scheme 1-14, the dialdehyde can then undergo an intramolecular Aldol reaction (followed by a dehydration) to give

8

As a side note, this compound was also used as a starting material for a later synthesis o f cholesterol'9 and also lanosterol.20

b a se 0 CHO 1. Oxidation 0 Scheme 1-14 1.4.3: Diels-Alder Approach:

The Diels-Alder reaction is a thermally induced [4+2]-cycloaddition between a 47t- system (referred to as the diene) and a 27t-system (referred to as the dienophile).21 The cycloaddition occurs in a concerted manner through a boat-like transition state to give a cyclohexene ring as a product. Since the bond formation and bond breakage occur simultaneously, the stereochemistry o f the substituents on the diene or dienophile are reflected in the cycloadduct. Thus, if the geometry o f approach o f the dienophile (with respect to the diene) can be predicted or controlled, then the stereochemistry o f the cycloadduct can thus be predicted or controlled: as shown in Scheme 1-15, up to four contiguous stereocenters may be generated. Fortunately, for the modem synthetic organic chemist, a great deal o f research has been conducted with the Diels-Alder reaction, and as a result it is, in fact, possible to predict (and, thus, with appropriate starting materials,

22

control) the regiochemistry and stereochemistry o f the cycloaddition. In the case o f steroids, or, for that matter, any product in which there are one or more cyclohexene (0 1

cyclohexane) rings present, the Diels-Alder reaction presents a very attractive route to efficiently and predictably generate the carbocyclic skeleton.

d ien e dienophile cycloadduct Scheme 1-15

The three main factors that must be considered when a Diels-Alder reaction is to be employed in a synthesis are the relative reactivity of the diene/dienophile system, the

22

regiochemistry o f the addition, and the stereochemistiy o f the addition. The first factor is

23

perhaps best understood through the use o f Frontier Molecular Orbital Theory. ‘ Since the ‘normal’ Diels-Alder reaction occurs as a result o f the interaction o f the HOMO o f the diene with the LUMO o f the dienophile, it stands to reason that substituents which minimize the energy gap between these two orbitals will cause the cycloaddition to become more facile. Since the LUMO o f the dienophile will be higher in energy than the HOMO o f the diene, the desired substituents should lower the energy o f the dienophile LUMO while raising the energy o f the diene HOMO. Hence, electron-withdrawing substituents (such as carbonyl groups or nitrites) 'activate' dienophiles, whereas electron releasing substituents (alkoxy groups) 'activate' dienes.

The regiochemistry o f the cycloaddition can be predicted based on the location and

22

the nature o f the substituents on both the diene and the dienophile. As discussed in the previous paragraph, these substituents wiil affect the relative energies o f the molecular

23

orbitals, but they will also affect the magnitude o f the frontier orbital coefficients. Since the strongest interaction will occur between orbitals with the largest coefficients (and, since these interactions involve the frontier orbitals), the relative location and electronic effects o f the substituents will also control the regiochemistry o f the reaction. Although the magnitude o f all the FMO coefficients in a molecule may be difficult to predict, one can use a reasonably simple and reliable 'tool' to predict the regiochemistry o f the

24

addition. The 'end' o f the diene (either carbon one or four o f the 1,3 diene) which has the greatest orbital coefficient can be predicted, by simple resonance-like 'arrow-pushing' from the most electron-rich substituent. As shown in Figure 1-3, the electron-releasing group at carbon one will give the largest orbital coefficient at carbon four, and an ERG at carbon tw o will give the largest FMO coefficient at carbon one. Similar arguments with an electron-withdrawing group on the dienophile will show that carbon two is most capable of'accepting' electrons, and will thus have the largest FMO coefficient.

ERG

ERG

4

C -4 h a s largest FMO coefficient

ERG.

“

C-1 h a s largest FMO coefficient

EWG

_EWG

2 | j | - ' C -2 h a s largest FMO coefficient

EW 3=electron-w ithdraw ing group (carbonyl, nitrile etc.) E R G = electron-releasing group (alkoxy, siloxy etc.)

Figure 1-3: Prediction o f FMO Coefficients for Dienes and Dienophiles

The stereochemistry o f the cycloaddition will be a result o f the orientation o f the diene with respect to the dienophile. As shown in Scheme 1-16, the addition can proceed through either an endo or an exo mode. Fortunately, for the organic chemist, the two modes o f cycloaddition are often energetically quite different, thus one product is often formed selectively or exclusively. In the case when the dienophile bears an unsaturated substituent (such as a carbonyl), the substituent can undergo a stabilizing interaction with

22

the Tt-system o f the diene. O f course, this so called secondary orbital interaction would only occur when the substituent on the dienophile is oriented over the diene; thus, the endo transition state would be expected to be energetically favoured. Other rationalizations for the preference o f the endo-type addition have also been made using

25

dipolar interactions and van der Waals interactions as arguments O f course, the degree o f preference for the endo transition state will vary with different systems, but in many cases the preference can be exploited in a synthetic strategy. As well, Lewis acid catalysts

can be employed to enhance the reactivity and the selectivity o f a given diene/dienophile

Since the steroidal skeleton contains three six-membered rings, a synthetic strategy to generate one, tw o, or even all three o f the rings via Diels-Alder cycloaddition reactions could potentially be developed. As shown in Scheme 1-17, an example o f three variations on a basic intramolecular Diels-Alder strategy are shown, which can be used to generate the A (to give 10b), B ( l i b ) , and C (12b) rings respectively. N ote that, in each case, a second ring, adjacent to the one formed via the Diels-Alder reaction, is also formed.

system.

cycloadduct dienophile

cycloadduct Scheme 1-16

-

g

$?

*06?

c i ? —

Scheme 1-17

In most Diels-Alder based strategies towards steroids, it is either the A- or B-ring that is formed via the cycloaddition reaction. Perhaps the simplest example o f such a strategy is employed in the synthesis o f estrogen-based steroids. In such a system, the A- ring is aromatic, which offers tw o advantages to the synthetic chemist: access to the potential formation o f an orf&oquinone dimethide intermediate, an extremely reactive diene, and secondly a relative degree o f simplicity since the A/B ring junction does not contain any stereocenters. An example o f such a strategy is shown in the synthesis of

27

estra-l,3,5(10)-trien-17-one (15). A thermally induced cheletropic elimination of sulfur dioxide from the starting material 13 will generate the o-quinone dimethide diene. Note that the intermediate db ne 14 contains only two stereocenters (which, in the synthesis are racemic, but bear the indicated relative stereochemical relationship to each other). These stereocenters control the approach o f the dienophile with respect to the diene, such that the dienophile must be 'over* the diene (as drawn) during the cycloaddition. Thus, the generation o f the newly formed stereocenters (B/C ring junction) is 'controlled' by the

relative stereochemistry o f the C/D ring junction: this process is generally known as relative asymmetric induction. This aspect o f the Diels-Alder reaction is potentially very powerful as it could allow for the generation o f a number o f stereocenters (which, if the starting material were to be chiral, would also be chiral) from only a few 'directing' centers.

O

Scheme 1-18

Such strategies to generate the B-ring o f a steroidal skeleton are not limited to the estrogen/estrone type steroids. In fact, there are examples o f stereoselective transannular Alder reactions (in macrocyclic systems such as 16) in which a transannular Diels-Alder reaction is used to generate the B-ring, while, at the same time, also generating the

28

A- and C-rings. As shown in Scheme 1-19, the stereochemistry at the ring junctions is controlled by the approach o f the dienophile with respect to the diene and also by the geometry (cis vs. tram ) o f the double bonds. In this case, the A/B/C junction stereochemistry o f the product 17 is cis-anti-trans. Clearly, through changing the stereochemistry o f the double bonds o f the diene and/or the dienophile, the

stereochemistry o f the junctions can be also changed. In fact, th t Deslongchamps group has applied such a strategy to generate a variety o f polycyclic systems with good control

29

over the ring junction stereochemistry.

18CPC

0 16

Scheme 1-19

Another example o f a stereocontrolled intramolecular Diels-Alder reaction, which, in this case, was used to generate the A- (and B-) ring of a steroidal skeleton,30 is shown in Scheme 1-20. This particular synthesis was conducted in an enantioselective manner; thus the tw o newly formed stereocenters in 19 are chiral in nature. The Diels-Alder reaction formed two isomeric products in a 4:1 ratio under the indicated conditions: the major product having the stereochemistry shown in the scheme, and the minor product having a cis A/B ring junction (with a P-hydrogen at carbon 5). Since both the diene and the dienophile are not electronically activated (via appropriate substituents) in this system, the cycloaddition requires a relatively high temperature and a long duration to occur (220°C for 100 hours). In this case, the intermediate 19 was used to generate testosterone and androsterone in an enantioselective manner.30

100 hours

Scheme 1-20

One final example3' o f the utility o f the Diels-Alder reactions in the generation o f steroidal-type carbocyclic skeletons also illustrates a relatively new concept in synthetic organic chemistry: the use o f tandem reactions. In this case, a radical cyclization reaction is used on 20 to generate a five membered ring which contains an exocyclic diene (Scheme 1-21). This intermediate, 21, will, under the same reaction conditions, undergo an intramolecular Diels-Alder reaction with the pendant dienophile, to give the tricyclic species 22 as the product (thus generating the equivalent o f the steroidal B- and C-rings).

BuaSnH / AIBN

(E = C 0 2 C H 3)

Although the product 22 does not contain the entire tetracyclic steroidal skeleton, it could easily be employed as an intermediate in the synthesis o f various steroids or other natural products. Effectively, in this reaction, the B, C, and D-rings are generated in one step. Perhaps if such a strategy were to employ an intact A-ring with a defined junction stereochemistry (where the B-ring will form), it would be possible to use this type o f strategy in a stereo- or enantioselective synthesis o f a steroidal skeleton.

1,5 Project Outline:

The goal o f this particular project is to develop a novel, efficient, versatile and stereoselective method to generate the steroidal carbocyclic skeleton, and then employ the method in a total synthesis o f a steroidal natural product. The basic strategy is to construct the three six membered rings o f the steroidal skeleton in a stereoselective manner - giving a perhydrophenanthrene with the A/B/C stereochemistry being trans-anti­ tram - using a tandem or a sequential Diels-Alder approach.

The basic outline o f the strategy is shown in Scheme 1-22. Ideally, both Diels- Alder reactions could be conducted at the same time, in a tandem fashion. Such a strategy provides an example o f an aspect o f the Diels-Alder reaction that is seldom employed in

32

synthesis: the cross-Diels-Alder cycloaddiiion (CDAC), In such a reaction, tw o (or more) dienes are reacted, with one acting as a dienophile and another as a diene. Clearly, the substituents on the diene and dienophile must be chosen carefully to ensure that the reactions occur in a chemoselective manner; in the first DAC in Scheme 1-22 (between 23 and 24), there are actually 18 possible cycloadducts.

H

H

E=electron-wfthdrawing group, Z=electron-releasing group

One could predict, however, that the first DAC should occur between the most electronically activated diene and dienophile, which, in Scheme 1-22, would be between 23 and 24 as indicated. In order for the A/B ring-junction stereochemistry to be tram in nature, the first intermolecular DAC must be endo-selective. Following the first DAC, the second intramolecular cycloaddition between the unactivated pendant diene and dienophile should proceed in a selective manner to give 26 (see Figure 1-4). As shown in the figure, the B-ring should, by energetic considerations, be in the most stable chair-like conformation in the transition state. The pendant diene can then adopt two possible conformations to enable the necessary formation o f the boat-like DAC geometry in the 'C - ring. O f the two possible conformations, 25a should be more stable, and therefore favoured, because it lacks the steric interaction between the diene and the axial 'E' group in transition state 25b. Thus, one would predict, following the IMDAC, the stereochemistry shown in 26.

25a

Figure 1-4: Possible Transition States for the IMDAC o f 25

One rather attractive aspect o f this strategy, is that, with the appropriate 'Z'- substituent on the diene (most likely trimethylsiloxy), it would be possible to generate, from the cycloadduct, a number o f different A-rings present in steroid natural products (see Figure 1-5). For example, hydrolysis o f the silyl enol ether would lead to the

33

androstane-type A-ring directly. Oxidative desilylation would lead to the ot,p- unsaturated ketone, which is present in the testosterone and progesterone-based systems.

And, finally, oxidative methods could be employed to the unsaturated ketone to yield the aromatic A-ring present in estrogen and estrone-type steroids. This versatility o f the A- ring intermediate should allow for the synthesis o f a wide variety o f steroid natural products from a common intermediate.

A-ring interm ediate _{A ndrostane-type} A-ring

T e sto stero n e an d E strogen/estrone progestin-type type A-ring

A-ring

Figure 1-5: Potential Steroidal A-Rings From a Common Synthetic , Intermediate.

As well, careful choice o f other substituents on the diene and dienophile could lead, with synthetic control, to a number o f structural analogs (for example, 18, and 19- nor steroids) o f the natural products which may have significant biological activity. In fact, many commercial steroid based drugs are analogs o f natural products, so a potential route to the synthesis o f these analogs could be o f significant pharmaceutical importance.

Following the synthesis o f the appropriate 'bis-diene' and 'bis-dienophile1, and stereoselective generation o f the perhydrophenanthrene skeleton, a strategy to incorporate

the five-membered D-ring o f the stf .oidal skeleton would have to be developed. There are two possible modes by which this could be achieved (as shown in Scheme 1-23); incorporation o f the ring into the bis-diene (to give a structure such as 27 following the Diels-Alder reactions), or addition o f the ring (to a structure such as 26) following the two Diels-Alder reactions. Whichever mode is chosen, two major points must be addressed; the stereochemistry o f the C/D ring junction, and the choice o f functionality at C l 7. The C/D junction must be tra m in nature, and in almost all steroids, there exists an angular methyl group attached to C l 3 (see structure 28 in Scheme 1-23). As well, several steroids contain a defined stereocenter at C l 7 (hydroxyl, or alkyl), so care must be taken in the planning stage o f the total synthesis to ensure that the stereochemistry at C l 7 will be in accordance with that o f the rest o f the molecule.

Tricyclic Interm ediate Incorporated D-ring D efined Stereochem istry at C/D junction and Cl 7 Scheme 1-23

Once the five membered ring is incorporated, and the two cycloadditions are done, a short number o f functional group manipulations would have to be accomplished to complete the synthesis o f a steroidal natural product. Most likely, a product which contains a saturated A-ring would be chosen as a synthetic target because o f its relatively simple access via the above strategy. Thus, an androstane-type steroid - such as 5a- dihydrotestosterone, shown in Figure 1-6 - would likely be the first natural product as a synthetic target o f this strategy. Once such a total synthesis is achieved, the next step in

the project would be to attempt the total synthesis on an enantioselective level. At the same time, it would also be possible to test the versatility o f the method through attempting the synthesis o f a variety o f different Jero id natural products. If adaptable to enantioselective techniques, and versatile in its application, the strategy described herein could be o f significant phaimaceutical (and chemical) interest, as it would allow access to a wide variety o f structural analogs o f natural products for biological testing, and even potential clinical use.

OH

0

CHAPTER TWO:

RESULTS AND DISCUSSION

2.1: Retrosvnthetic Analysis;

From the target molecule, dihydrotestosterone (29), the retrosynthetic analysis o f two synthetic strategies are shown in Scheme 2-1. The steroid should be available from the products o f the two Diels-Alder cycloadditions (31 or 30). As shown in the diagram, the stereochemistry o f the A/B/C ring junctions in 30 and 31 should be trans-anti-trans.

RO

RO

OR

DAC - between 32 or 33 and the 'bis-dienophile' - must be endo-selective in nature. Two possible bis-dienes are shown in Scheme 2-1: one which incorporates an intact five- membered ring (32), and another, 33, which is acyclic in nature, but which will allow for the generation o f the D-ring following the DAC’s. The synthesis o f both bis-dienes would be attempted. Synthesis o f 32 may be possible via an acyclic precursor 34 via application o f recently developed palladium-catalyzed cycloisomerization techniques.34

From an experimental standpoint, we had to establish that the intermolecular DAC occurs in an endo-selective (and regioselective) manner. Secondly, determination that the two cycloadditions can be accomplished in a tandem or sequential manner to generate the three six-membered rings must be made. Keeping this in mind, the syntheses were approached with the following chronological goals: (1) Testing o f the reactivity and selectivity o f various dienophiles using a simple model diene. (2) Generation o f a model bis-diene and reaction with the previously evaluated dienophiles to generate a perhydrophenanthrene (three fused six-membered rings) skeleton with the correct trans- anti-tram stereochemistry o f the ring junctions. (3) Synthesis o f one or both bis-dienes (32 and/or 33) (4) Stereoselective development o f the C/D ring junction and generation o f a steroidal natural product such as dihydrotestosterone (29).

In order to evaluate the endo-selectivity o f various dienophiles, a model diene was required which, following a DAC, would generate a cycloadduct that would resemble the A-ring o f a steroidal nucleus. Accordingly, 2-trimethylsiloxy-l,3-pentadiene 35 was synthesized from 3-penten-2-one by treatment with LDA and quenching with TMSCl

35

(Scheme 2-2). Two other model dienes were also made via trapping o f the enolate with different electrophiles: TBDMSC1 and diethyl chlorophosphonate. However, the trimethylsilyl enol ether was chosen as the preferred model due to its relatively easy isolation (via distillation) and ease o f subsequent transformation.

1. LDA/THF

0 2. TMSCl J M S O

35

Scheme 2-2

The next model compound that was required was one that would allow evaluation o f both DAC reactions. Thus, a bis-diene that could react with a bis-dienophile to give the perhydrophenanthrene skeleton was needed. The synthesis o f this bis-diene is outlined in

3 7

Scheme 2-3. From the commercial vinylmagnesium bromide and acrolein, 1,4- pentadien-3-ol (36) was generated in 87% yield. Following isolation o f the product by distillation, the alcohol was reacted with triethyl orthoacetate in refluxing toluene,

3 8

undergoing an orthoester Claisen rearrangement, to yield the ester 37 in a 73% yield with the tra/w-stereochemistry at the internal alkene.

Note: 36 is also available commercially, but due to its high cost, the synthesis was conducted from the less expensive starting materials acrolein and vinylmagnesium bromide,

PhCH3, A

Sw em

TMSCl

TMSO

Scheme 2-3

Generation o f the second diene unit was accomplished via reduction o f the ester 37

3 9

to an aldehyde followed by a Homer-Wadsworth-Emmons reaction. Although the most direct way to accomplish this was to reduce the ester with DIBAL-H, we found it more efficient to instead reduce the ester 37 to the stable alcohol 38 (83 % yield), which could be purified by distillation or chromatography, followed by oxidation to the aldehyde 39

4 0

using the Swem reaction conditions. The aldehyde was used directly, without purification after work-up, in the subsequent Homer-Wadsworth-Emmons reaction to give the ketone 40 in a 79% yield (from the alcohol 39). Treatment o f the ketone with LDA, followed by trapping o f the kinetic enolate with TMSCl resulted in the formation o f the second diene unit, and gave the bis-diene 41 in a 81% yield following distillation.

quantities o f 41) was reasonably easy, since only one chromatographic separation (to purify 40) was necessary in the entire scheme: all other materials could be isolated in pure form via distillation techniques.