The Chemistry of Salvia divinorum

Thomas Anthony Munro

Submitted in total fulfilment of the requirements

of the degree of Doctor of Philosophy

April 2006

Department of Chemistry

The University of Melbourne 2 Abstract

Salvia divinorum is a hallucinogenic sage used to treat illness by the Mazatec Indians of Mexico. Salvinorin A (1a), a neoclerodane diterpenoid isolated from the plant, is a potent, selective agonist at the κ opioid receptor (KOR), and is the first non-nitrogenous opioid. The plant is used recreationally as a hallucinogen, but is unpopular due to its dysphoric effects. 1a has been prohibited in Australia under an invalid systematic name.

An early report of psychoactive alkaloids in S. divinorum proved to be irre- producible. Similarly, tests in mice suggesting the presence of psychoactive compounds other than 1a were confounded and therefore unreliable.

O O O

O 1 1 O O OR O R H H H H 2 H H O R 2 2 8 O O R

O O O O O OR3 R1 R2 R1 R2 R3

1a 1d Ac OH 28a OH H H 1e H OAc 28b OH OH Me 1f H H 28c H OAc H

In this work, an improved isolation method for 1a was developed, using fil- tration through activated carbon to decolourise the crude extract. Six new diterpenoids were isolated: salvinorins D–F (1d–1f) and divinatorins A–C (28a–28c). Five known terpenoids not previously reported from this species were also isolated.

3 4

The structure–activity relationships of 1a were evaluated via selective mod- ifications of each functional group. Useful synthetic methods are reviewed, including the first thorough review of furanolactone hydrogenations. Testing of the derivatives at the KOR suggests that the methyl ester and furan ring of 1a are required for activity, but that the lactone and ketone functionali- ties are not. Other compounds from S. divinorum did not bind to the KOR, suggesting that 1a is the plant’s active principle.

The structure of the 8-epimer of 1a, reported previously without supporting evidence, was firmly established. This epimerisation proved to be a general phenomenon among salvinorins and related furanolactones, occurring via eno- lisation of the lactone. The more complex mechanism proposed by Koreeda and co-workers was inconsistent with subsequent data. Under strongly basic conditions, autoxidation of 1a occurred to give the enedione 59 as the major product. A previously proposed structure was shown to be incorrect.

Salvinorins and divinatorins were tested and found to be inactive against in- sects, bacteria, fungi, HIV, tumour cell lines and protein synthesis.

O

OH O H O O

O O 59 Declaration

This is to certify that

1. the thesis comprises only my original work except where indicated in the preface,

2. due acknowledgement has been made in the text to all other material used,

3.thethesisislessthan100,000wordsin length, exclusive of tables, maps, bibliographies and appendices.

Thomas Munro

5 6 Preface

Some of the bioassays described in Chapter 4 were performed by others.

• Opioid receptor binding assays were performed in the laboratories of Bryan Roth, Case Western Reserve University (Cleveland, Ohio), by Glenn Goetchius, Beth Ann Toth, Feng Yan, and Timothy Vortherms.

• Protein synthesis inhibition assays were performed in the laboratories of Jerry Pelletier, McGill University (Montreal, Canada).

• HIV replication assays (NL4.3 and AD8 strains) were performed in the laboratories of Sharon Lewin, Monash University, by Ajantha Solomon.

• Other HIV assays were performed at Southern Research Institute (Fred- erick, Maryland), under the direction of Dr Stephen Turk, U. S. National Institute of Allergy and Infectious Diseases (NIAID).

• Tumour cell growth inhibition assays were performed by the U. S. Na- tional Cancer Institute (NCI).

Other assays were performed collaboratively:

• Antibacterial and antifungal assays were performed in the laboratories of Professor Roy Robins-Browne, University of Melbourne, with Andrea Bigham.

• Insect antifeedant assays were performed using supplies and facilities provided by David Heckel and Charles Robin, University of Melbourne.

7 8

Some photographs were provided by others, as credited.

Parts of this work have been published previously:

• Munro, T. A.; Rizzacasa, M. A. Salvinorins D–F, New Neoclerodane Diterpenoids from Salvia divinorum, andanImprovedMethodforthe Isolation of Salvinorin A. J. Nat. Prod. 2003, 66, 703–705. http://dx. doi.org/10.1021/np0205699

• Bigham, A. K.; Munro, T. A.; Rizzacasa, M. A.; Robins-Browne, R. M. Divinatorins A–C, New Neoclerodane Diterpenoids from the Controlled Sage Salvia divinorum. J. Nat. Prod. 2003, 66, 1242–1244. http://dx. doi.org/10.1021/np030313i

• Munro, T. A.; Rizzacasa, M. A.; Roth, B. L.; Toth, B. A.; Yan, F. Studies toward the Pharmacophore of Salvinorin A, a Potent κ Opioid Receptor Agonist. J. Med. Chem. 2005, 48, 345–348. http://dx.doi. org/10.1021/jm049438q

• Munro,T.A.;Goetchius,G.W.;Roth,B.L.;Vortherms,T.A.;Rizza- casa, M. A. Autoxidation of Salvinorin A under Basic Conditions. J. Org. Chem. 2005, 70, 10,057–10,061. http://dx.doi.org/10.1021/jo051813e Acknowledgments

Thanks to the Commonwealth Government for an Australian Postgraduate Award.

Thanks Mark for betting tight resources on a risky project. I’m glad it paid off. Thanks to the man with the coolest name in chemistry, Leander Jerome Julian Valdés III, for generously sharing ideas and advice. Daniel Siebert for helpful advice on lots of stuff, and for introducing me to Bryan Roth. Torsten for introducing me to Daniel; Carl Turney for introducing me to Torsten (and everything else); and Erik for introducing me to Carl. Six degrees of separation. Mike and Heike, Antoine and Lara for being so generous. Frances for changing my life, and for the world’s coolest lab coat and bestest present. TK for teaching me NMR, from setting the trash hole to running a NOESY. Les Gamel for all that masterful glassblowing. Carl Schiesser for the radical initiator that dare not speak its name. Sammy for letting me use windoze. Danny for letting me use Adobe Creative Suite. Ben for introducing me to Hoye’s NMR papers. The man with the second-coolest name in chemistry, Carlos Rodríguez, for translating Díaz. Vic Iwanov for letting me use the safe, and filling out those annoying manifests. Max Hem for the photographs, glorious as always. A man with another cool name, Slava Olcheski, for the Oaxaca shot. Scott Crawford for getting those stunning shots out of the SEM. Richard Westkaemper for providing the binding model data. Tom for proofreading – I owe you one big fella. Same goes for Caroline. Cheese, Gromit! Thanks Dad for keeping life interesting. Mojave, Ilulisaat, Peshawar, terror australis, silver shark, helibagging. What can I say: my dad’s better than yours. Thanks mum

9 10 for the sacrifices you made and the love and hard slog you put into raising kids and working full time. And for helping with the move so I could keep writing this till five days before leaving for my postdoc! Contents

Abstract 3

Declaration 5

Preface 7

Acknowledgments 9

List of Figures 17

List of Schemes 23

List of Tables 25

Acronyms 27

1 Introduction. 31

1.1Botany...... 31

1.2Ethnopharmacology...... 33

1.3Chemistry...... 36

1.3.1 Terpenoids...... 36

1.3.2 Allegedalkaloids...... 43

1.4Pharmacology...... 47

11 12 CONTENTS

1.4.1 AnimalTesting...... 47

1.4.2 HumanTesting...... 51

1.4.3 In vitro Testing...... 53

1.4.4 Mechanism: κ Opioids...... 54

1.5Toxicology...... 61

1.6Socialimpact...... 62

1.6.1 RecreationalUse...... 62

1.6.2 LegalStatus...... 63

1.7Summary...... 66

2Isolation. 67

2.1IsolationProcedure...... 67

2.1.1 ExtractionConditions...... 67

2.1.2 ProblemsCausedbyPigments...... 69

2.1.3 UseofActivatedCarbon...... 72

2.1.4 SeparationofTerpenoids...... 82

2.2StructureElucidation...... 90

2.2.1 RevisedNMRAssignmentsforSalvinorinA(1a)..... 90

2.2.2 RevisedNMRAssignmentsforSalvinorinC(1c)..... 92

2.2.3 OtherKnownDiterpenoids...... 92

2.2.4 KnownTriterpenoids...... 96

2.2.5 SalvinorinsD-F(1d-1f)...... 98

2.2.6 DivinatorinsA-C(28a-28c)...... 110

2.2.7 Subsequent isolations...... 118 CONTENTS 13

3 Synthesis. 119

3.1Knownderivatives...... 119

3.2 Epimerisation at C-8 under Basic Conditions...... 120

3.2.1 PreviousReports...... 120

3.2.2 8-epi-SalvinorinsAandB(37aand37b)...... 121

3.2.3 Control of Epimerisation and Separation of Epimers. . . 122

3.2.4 8-epi-Salvinorin C (37c) and Related Compounds. ....124

3.2.5 ChromatographicIdentificationofEpimers...... 125

3.2.6 Mechanism...... 126

3.2.7 Attempted Deacetylation under Acidic Conditions. . . . 128

3.3SimpleDerivatives...... 128

3.3.1 Esters(46and47)...... 128

3.3.2 AttemptedBenzylEtherFormation(48)...... 129

3.3.3 17-Deoxy Compounds (49 and 50)...... 130

3.3.4 TetrahydrosalvinorinA(51)...... 131

3.3.5 (+)-Hardwickiic Acid (ent-29a)...... 134

3.4ModificationoftheMethylEster...... 134

3.4.1 RelevantResultsfromPreviousWork...... 134

3.4.2 TreatmentofSalvinorinAwithKOHinMeOH...... 135

3.4.3 O-Demethylsalvinorin A (67a)...... 144

3.4.4 O-Demethyl-18-deoxysalvinorinA(77)...... 150

3.5ModificationoftheKetone...... 153

3.5.1 AttemptedMethylenation...... 153

3.5.2 AttemptedDirectDeoxygenation...... 154

3.5.3 IndirectDeoxygenation...... 155 14 CONTENTS

4 Bioassays. 161

4.1InsectAntifeedantActivity...... 161

4.2EukaryoticProteinSynthesisInhibition...... 163

4.3AntimicrobialActivity...... 164

4.3.1 BacteriaandFungi...... 164

4.3.2 HIV-1...... 165

4.4NCIAnticancerScreen...... 168

4.5 Activity at the κ Opioid Receptor...... 170

4.5.1 OtherSalvinorinsandDivinatorins...... 171

4.5.2 ModificationoftheKetone...... 173

4.5.3 ModificationoftheAcetoxyGroup...... 174

4.5.4 ModificationoftheMethylEster...... 175

4.5.5 ModificationoftheLactone...... 176

4.5.6 ModificationoftheFuranRing...... 176

4.5.7 IncorporationintoaRevisedBindingModel...... 177

4.5.8 Subsequent Results...... 179

5 Experimental. 181

5.1GeneralConditions...... 181

5.1.1 InstrumentsandProcedures...... 181

5.1.2 Reagents...... 183

5.1.3 PlantMaterials...... 183

5.1.4 Assays...... 184

5.2Isolation...... 185

5.2.1 Extraction of Commercial S. divinorum...... 185 CONTENTS 15

5.2.2 Extraction of Australian S. divinorum...... 187

5.2.3 SalvinorinA(1a)...... 188

5.2.4 SalvinorinB(1b)...... 189

5.2.5 SalvinorinC(1c)...... 189

5.2.6 SalvinorinD(1d)...... 190

5.2.7 SalvinorinE(1e)...... 191

5.2.8 SalvinorinF(1f)...... 192

5.2.9 DivinatorinA(28a)...... 194

5.2.10DivinatorinB(28b)...... 195

5.2.11DivinatorinC(28c)...... 196

5.2.12(–)-HardwickiicAcid(29a)andmethylester29b.....197

5.2.13OleanolicAcid(31)...... 198

5.2.14PresqualeneAlcohol(32)...... 198

5.2.15Peplusol(33)...... 199

5.3Synthesis...... 199

5.3.1 Salvinorin C (1c) via acetylation of salvinorin D (1d). . . 199

5.3.2 Salvinorins D (1d) and E (1e) via acetylation of 1h. . . . 200

5.3.3 Salvinorins C (1c) and E (1e) via acetylation of 1h. . . . 200

5.3.4 DideacetylsalvinorinC(1h)from1c...... 201

5.3.5 (+)-Hardwickiic acid (ent-29a)...... 202

5.3.6 SalvinorinAlactol(35)...... 203

5.3.7 (4R)-3,4-DihydrosalvinorinC(36c)...... 205

5.3.8 (4R)-3,4-DihydrosalvinorinE(36e)...... 206

5.3.9 (4R)-Dideacetyl-3,4-dihydrosalvinorinC(36h)...... 207

5.3.10 8-epi-SalvinorinA(37a)...... 208 16 CONTENTS

5.3.11 8-epi-SalvinorinB(37b)...... 210

5.3.12 8-epi-SalvinorinC(37c)...... 211

5.3.13 8-epi-SalvinorinD(37d)...... 212

5.3.14 8-epi-SalvinorinE(37e)...... 214

5.3.15 8-epi-DideacetylsalvinorinC(37h)...... 215

5.3.16SalvinorinBformate(46)...... 216

5.3.17 Dideacetylsalvinorin C 2-O-(4-bromobenzoate) (47). . . . 218

5.3.1817-DeoxysalvinorinA(49)...... 219

5.3.198,17-Didehydro-17-deoxysalvinorinA(50)...... 220

5.3.2013,14,15,16-TetrahydrosalvinorinA(51)...... 222

5.3.21Autoxidationof1ainKOH/MeOH...... 223

5.3.22 NaBH4 reductionof59...... 228

5.3.23 O-Demethyl-18-deoxysalvinorinA(77)...... 229

5.3.241-DeoxysalvinorinA(81a)...... 233

Bibliography 239 List of Figures

1.1 Flowering specimen of S. divinorum.1 ...... 31

1.2 Location of Oaxaca and the Sierra Mazateca within Mexico. . . 32

1.3 Young Salvia divinorum plant, Oaxaca.8 ...... 33

1.4 S. divinorum flower in bud (stereoview).1 ...... 33

1.5 Ortega et al’s X-ray structure of 1a (stereoview)...... 37

1.6 Salvinorins B (1b)andC(1c)...... 38

1.7Knownterpenoids...... 38

1.8 Clerodin (5)andstandardclerodanenumbering...... 39

1.9Biosyntheticprecursorsofditerpenoids...... 40

1.10 Peltate glandular trichome (SEM stereoview).39 ...... 42

1.11 Underside of S. divinorum leaves (SEM stereoview).39 ...... 43

1.12 Activity of pure 1a versus mixtures in mice (open field assay). . 48

1.13 Morphine (11)andderivatives...... 54

1.14 Arylacetamide κ opioids...... 56

1.15 Structurally diverse κ opioids...... 59

1.16RelevantIUPACfused-ringnumberingschemes...... 64

40 2.1 Siebert’s TLC analysis of crude CHCl3 extracts...... 69

2.2Representativemajorplantpigments...... 71

17 18 LIST OF FIGURES

2.3 Ortho- and non-ortho-substitutedPCBs...... 76

2.4ApparatusforFiltrationthroughActivatedCarbon...... 78

2.5Flavonoids...... 80

2.6 SEM image of salvinorin A crystals (blade morphology).39 ... 82

2.7 Other salvinorin A crystal morphologies (stereoview).39 ..... 84

2.8 Terpenoids isolated from S. divinorum...... 85

2.9 TLC data of isolated compounds...... 86

2.10 Isolation of terpenoids from commercial S. divinorum...... 87

2.11 Isolation of terpenoids from Australian S. divinorum...... 88

1 2.12 H NMR spectrum of 1a (800 MHz, CDCl3)...... 91

2.13 HSQC spectrum of 1a (800 MHz, CDCl3)...... 91

2.14 Revised NMR assignments for 1a (stereoview)...... 92

1 2.15 H NMR spectrum of 1c (400 MHz, CDCl3)...... 93

2.16 Revised NMR assignments for 1c (stereoview)...... 93

2.17 HMQC and HMBC spectra of 1c (400 MHz, CDCl3)...... 94

2.18 Single-crystal X-ray structure of 29a (stereoview)...... 95

2.19 (E)-Phytol...... 96

2.20Oleanolicacid...... 96

2.21Presqualenealcoholandpeplusol...... 97

1 2.22 H NMR spectrum of 1d (400 MHz, CDCl3)...... 99

2.23 NMR assignments and key 2D correlations for 1d (stereoview). . 100

2.24 HMQC spectrum of 1d (400 MHz, CDCl3)...... 100

2.25 HMBC spectrum of 1d...... 101

1 2.26 H NMR spectrum of 1e (400 MHz, CDCl3)...... 102

2.27 NMR assignments and key 2D correlations for 1e (stereoview). . 102 LIST OF FIGURES 19

2.28 HMQC spectrum of 1e...... 103

2.29 HMBC spectrum of 1e...... 103

1 2.30 H NMR spectrum of 1h (400 MHz, CDCl3)...... 104

2.31 NMR assignments and key 2D correlations for 1h (stereoview). . 105

2.32 HMQC and HMBC spectra of 1h...... 105

2.33 NOESY spectrum of 1h...... 105

1 2.34 H NMR spectrum of 1f (400 MHz, CDCl3)...... 108

2.35 NMR assignments and key 2D correlations for 1f (stereoview). . 109

2.36 HMQC spectrum of 1f...... 109

2.37 HMBC spectrum of 1f...... 109

2.38 Divinatorins A–C (28a-28c) and hardwickiic acid (29a).....110

1 2.39 H NMR spectrum of 28a (400 MHz, CDCl3)...... 110

2.40 NMR assignments and key 2D correlations for 28a (stereoview). 111

2.41 HMBC spectrum of 28a...... 111

2.42 NOESY spectrum of 28a...... 112

1 2.43 H NMR spectrum of 28b (400 MHz, CDCl3)...... 113

2.44 NMR assignments and key 2D correlations for 28b (stereoview). 113

2.45 HMBC spectrum of 28b...... 114

2.46 NOESY spectrum of 28b...... 114

1 2.47 H NMR spectrum of 28c (400 MHz, CDCl3)...... 115

2.48 NMR assignments and key 2D correlations for 28c (stereoview). 115

2.49 HMBC spectrum of 28c...... 116

2.50 NOESY spectrum of 28c...... 116

2.51 Subsequently isolated compounds...... 118

3.1 Key NOESY correlations for 37a (stereoview)...... 123 20 LIST OF FIGURES

3.2 TLC comparison of epimers using vanillin/H2SO4...... 126

3.3Esterandetherderivatives...... 129

3.4 Furanolactones 53, 54 and 55...... 131

3.5 (+)-Hardwickiic acid (ent-29a)...... 134

3.6 Key HMBC correlations of 59 and 60a...... 137

3.7 UV/Visible spectra of 1a, 1c and 59 inMeCN...... 138

3.8 Diol 36h andproposedautoxidationproduct62...... 139

3.9 O-DemethylsalvinorinsAandB...... 145

3.10Usefulnon-hydrogenbonddonorsolvents...... 145

3.11 Methylenated target compound 79...... 154

3.12 RP-LCMS traces of early fractions versus 81b/82b...... 159

3.13 1-Deoxysalvinorin A (81a)...... 160

4.1 Luciferase assay results for salvinorins and divinatorins (50 µM). 163

4.2(-)-HardwickiicacidanddivinatorinsA-C...... 164

4.3 Disk and microdilution assays for ent-29a and crude extract.428 165

4.4HIV-1replicationassays(NL43andAD8strains)...... 166

4.5HIV-1replicationassays(ROJOisolate)...... 167

4.6NCI60celllineresultsforsalvinorinsanddivinatorins.....169

4.7CNScelllineresultsfordivinatorinBandsalvinorinB...... 170

4.8KORbindingaffinityandpotencyofsalvinorinA...... 171

4.9KORbindingaffinitiesofsalvinorinsanddivinatorins...... 172

4.10KORactivityafterketonemodifications...... 174

4.11KORactivityafteracetoxygroupmodifications...... 174

4.12KORactivityaftermethylestermodifications...... 175

4.13KORactivityafterlactonemodifications...... 176 LIST OF FIGURES 21

4.14KORactivityafterfuranmodifications...... 176

4.15Westkaemper’soriginalbindingmodel...... 177

4.16Westkaemper’srevisedbindingmodel(stereoview)...... 178

4.17 KOR binding affinities and potencies of recent derivatives. . . . 179 22 LIST OF FIGURES List of Schemes

1.1 Cyclization of 7 to the trans-neoclerodane skeleton(stereoview). 41 1.2 Cyclization of 7 to the cis-neoclerodaneskeleton...... 41 2.1 Interconversion of 1c-1h...... 104 3.1Preparationofknownderivatives...... 119 3.2 Formation of 8-epi-salvinorinsAandB...... 121 3.3 Synthesis of 37c-37h...... 125 3.4 Koreeda et al’s proposedmechanismofepimerisation...... 127 3.5Epimerisationofrelatednaturalproductswithbase...... 127 3.6 Brown’s deuteration of 1b...... 128 3.7 Deoxygenation of lactol 35...... 130 3.8 Hydrogenation of 1a andotherfuranolactones...... 131

3.9 LiAlH4 and Li/NH3 reductions of 1a...... 135 3.10 Autoxidation of 1a...... 136 3.11Proposedmechanismoftheautoxidation...... 141 3.12 Unexpected oxidation product 65...... 141 3.13 Attempted reductions of 59...... 143

3.14 BAl2andBAc2estercleavagemechanisms...... 144 3.15 Formation of mixed anhydride 68...... 147

3.16 Some previously reported BH3·THFreductions...... 151 3.17 Borane reduction of 67a...... 152 3.18Ketonedeoxygenationviaatosylhydrazone...... 154 3.19 Formation of cyclic thionocarbonate 80...... 156 3.20 Radical deoxygenation of 80...... 158

23 24 LIST OF SCHEMES List of Tables

1.1 Relative potency of 1a at cloned κ opioid receptors...... 60

1.2 Alternate systematic names for 1a...... 64

2.1 Effects of solvent and temperature on recovery of 1a...... 68

2.2 Data and sources used to identify known compounds...... 94

3.1 Coupling constants (Hz) at H-8 for 1a, 1b and8-epimers.....122

3.2Somepreviouslyreportedfuranolactonehydrogenations.....133

3.3 Summary of results - nucleophilic cleavage of 1a methyl ester. . 148

3.4 Unsuccessful treatment of 1a withexcesstosylhydrazide.....155

4.1Antifeedanttestresults...... 162

4.2KORradioligandandfunctionalassayresults...... 172

5.1 Yields and TLC data (hRf ) of isolated compounds...... 187

25 26 LIST OF TABLES Acronyms

AIBN Azobisisobutyronitrile [2,2’-Azobis(2-methylpropionitrile)]

AZT Azidovudine

Borsm Based on recovered starting material

CNS Central nervous system

2−4 COSY Correlation spectroscopy 2D NMR ( JHH)

DEPT Distortionless enhancement by polarization transfer NMR (13Cmul- tiplicity)

DMAP 4-Dimethylaminopyridine

DMF Dimethylformamide

DMPU 1,3-dimethyltetrahydropyrimidin-2-one

EC50 Concentration causing 50% of maximal efficacy

GC/MS Gas chromatography/mass spectrometry

2−3 HMBC Heteronuclear multiple bond correlation 2D NMR ( JCH )

1 HMQC Heteronuclear multiple quantum coherence 2D NMR ( JCH)

HMPA Hexamethylphosphoric triamide (a.k.a. HMPT)

HPLC High performance liquid chromatography

HRESIMS High resolution electrospray ionisation mass spectrometry

27 28 LIST OF TABLES hRf = Rf × 100

1 HSQC Heteronuclear single quantum coherence 2D NMR ( JCH)

IC50 Concentration causing 50% inhibition

InChI IUPAC International Chemical Identifier

K i Receptor binding affinity constant

KOR κ (kappa) opioid receptor

LC/MS Liquid chromatography/mass spectrometry

LD50 Dose lethal to 50% of test animals

MIC Minimum inhibitory concentration

NDPSC National Drugs and Poisons Schedule Committee

NMR Nuclear magnetic resonance nOe Nuclear Overhauser effect (through–space signal enhancement)

NOESY Nuclear Overhauser enhancement spectroscopy (through–space 2D NMR)

PCB Polychlorinated biphenyl

PCR Polymerase chain reaction

R a f Retardation factor f [distances of analyte (a) and solvent front (f) from origin]

ROESY Rotating frame Overhauser enhancement spectroscopy (through– space 2D NMR)

RP Reverse phase (nonpolar stationary phase, polar mobile phase)

SEM Scanning electron microscopy / standard error of the mean

THF Tetrahydrofuran LIST OF TABLES 29

TLC Thin layer chromatography 30 LIST OF TABLES Chapter 1

Introduction.

Figure 1.1: Flowering specimen of S. divinorum.1

1.1 Botany.

The genus Salvia, containing over 900 species internationally, is one of the largest in the family Lamiaceae (or Labiatae). The largest subgenus is the

31 32 CHAPTER 1. INTRODUCTION.

N

Mexico Sierra Mazateca

Oaxaca

0300km100 200

Figure 1.2: Location of Oaxaca and the Sierra Mazateca within Mexico.

South American Calosphace (or Jungia),2 of which over 300 species are found in Mexico.3 Among these is Salvia divinorum Epling & Játiva.4, 5 Discovered in the highlands of northern Oaxaca (Figure 1.2), the plant is a perennial shrub growing to about 1.5 m, preferring moist shady sites at high elevations.6 While the flowers are distinctive (Figure 1.1), the plant’s appearance is nondescript during vegetative growth (Figure 1.3), apart from the unusual square stem. In the original botanical description, based on a dried specimen and witness reports, the flower (corolla) was erroneously described as being blue.4 This error was incorporated into several colour botanical illustrations.7 The error originated in observations by Robert Gordon Wasson and Albert Hofmann, who obtained the type specimen but were not botanists.6 In fact, the corolla is white, emerging from a violet calyx (Figure 1.4). This error has been corrected in an amended botanical description,6 which also has the advantage of being in English rather than Latin.

Several others had investigated the species and collected specimens,9 but Was- son and Hofmann were the first to obtain a flowering specimen, which is essen- tial for species identification. It should be pointed out that, although Wasson and Hofmann were credited as the collectors of the type specimen,4 it was in fact given to them by Natividad Rosa, a Mazatec healing woman.10, 11 Wasson and Hofmann have also been incorrectly credited with bringing live plants to 1.2. ETHNOPHARMACOLOGY. 33

Figure 1.3: Young Salvia divinorum plant, Oaxaca.8 the U.S.A., which was actually done by botanist Sterling Bunnell.12 Most of the S. divinorum plants in cultivation internationally are clones of this mis- named “Wasson and Hofmann” strain.12

Figure 1.4: S. divinorum flower in bud (stereoview).1

There has been no report of growing the plant from seed: propagation is ex- clusively vegetative.6,13 Moreover, there are no published first-hand reports of wild populations;7 despite thorough searches in the region by several workers, all known stands appear to have been planted.6, 14

1.2 Ethnopharmacology.

The leaves of S. divinorum are used as a traditional medicine by the Mazatec Indians of the Oaxaca region. The disorders treated include gastrointestinal 34 CHAPTER 1. INTRODUCTION. problems, headaches, rheumatism, anaemia and swelling of the stomach.15, 11, 13 An infusion is prepared by crushing or rubbing fresh leaves in water; a frothy infusion is considered a sign of potency. The leaf residue may also be applied to the patient’s forehead as a poultice afterwards.11

In addition, the healers (curanderos) themselves drink the infusion to induce visions.5 They believe these visions allow them to divine the cause of the illness. Hence the name S. divinorum, meaning “sage of the seers.”4 For these effects, larger doses of the infusion are used, or the leaves themselves are slowly chewed and eaten. Eating the leaves is very difficult, due to their sickeningly bitter taste, and often induces vomiting. First-hand accounts of the effects vary from barely perceptible (an overlay of “dancing colours”)5 to powerful experiences, in which awareness of reality is lost and bizarre visions are perceived as real:

I saw a pulsating purplish light that changed to an insect-like shape, perhaps a bee or a moth, and then into a pulsating sea anemone. It expanded into a desert full of prickly pear cacti, and remained so for several minutes. During the first session and throughout the night, my visions had all appeared to be something like a cross between a silent moving picture and a cartoon. I felt myself to be an observer of these mute visions, rather than being an actual part of them. Suddenly, however, I was in a broad meadow with brightly colored flowers. I had just crossed a stream by way of a small wooden bridge. Next to me was something that seemed to be the skeleton of a giant model airplane made of rainbow colored inner tubing. The sky was bright blue and I could see ... woods in the distance. I found myself talking to a man in a shining white robe who was either shaking my hand, or else holding on to it. It was an amazing hallucination, as I truly believed I was in the meadow.14

This is in marked contrast to the pseudohallucinations induced by compounds 1.2. ETHNOPHARMACOLOGY. 35 such as LSD. The visions were accompanied by impaired physical coordination and slurred speech.16 Notwithstanding the intensity of some such experiences, the Mazatecs consider the plant the weakest of their visionary substances.13 Theyrefertoitasška María pastora (the leaves of Mary shepherdess), appar- ently from an obscure Catholic term for the Virgin Mary.17

Ott has argued7 that the lack of an indigenous name indicates that the plant is not indigenous to the region – that, like sheep and Catholicism themselves, it came to the Mazatecs after European settlement. He points out that the introduced European hallucinogenic mushroom Psilocybe cubensis also has no indigenous name, unlike its indigenous equivalents, and is regarded as an in- ferior substitute. S. divinorum is regarded in the same way. Furthermore, the Mazatecs consider the plant as part of the same family as two introduced Coleus species. As further evidence, he points out that the Mazatec belief that drying the leaves destroys their potency has been disproven. Similarly, drink- ing an infusion has been shown to be very inefficient; much stronger effects are produced by retaining the infusion,18 or alternatively a “quid” of chewed leaves,7 in the mouth. Ott argues that each of the above points suggests a lack of cultural tradition, and concludes that the plant was adopted relatively recently from another tribe.

Ott has also endorsed Wasson’s suggestion19 that S. divinorum may represent the divinatory plant pipiltzintzintli cultivated by the Aztecs.7 The Aztecs pre- pared an infusion from the plant, including the leaves, and also applied it as a poultice; no other plant was used in both these ways. The only Mexican plant whose leaves are presently so used is S. divinorum. Accounts of pipiltz- intzintli make no mention of seeds, unlike other Aztec visionary plants; and S. divinorum is effectively seedless, unlike other contemporary Mexican visionary plants. Here, then, we have a plant without an indigenous name, and an in- digenous name referring to an unknown plant. One has no present, the other has no past, and they have several very distinctive characteristics in common.

Others reject this hypothesis. Díaz has pointed out that male and female gen- 36 CHAPTER 1. INTRODUCTION. ders of pipiltzintzintli were reported, that the plant was dried before use, and that the infusion was prepared from the entire plant, none of which is true of S. divinorum.20 Ott counters that the Aztecs may have been using gender metaphorically, as the Mazatecs do today. Drying does not affect the plant’s potency, and the addition of other plant parts, while superfluous, would not affect the infusion. Moreover, Díaz’s preferred candidate, Cannabis sativa,can- not be correct, since it was introduced after European settlement.7 Valdés also rejects Wasson’s hypothesis, preferring Beltrán’s proposal that pipiltzintzintli was a synonym for the morning glory, ololiuhqui.13 Ott has shown, however, that these were explicitly described as different plants, and that S. divinorum is thus

... the only Mexican entheogenic plant which fits the criteria for pipiltzintzintli, and ... remains our best guess for the identity of the lost Aztec entheogen.7

Both parties in this surprisingly bitter dispute concede that the evidence is inconclusive, and both suspect that S. divinorum has been used by tribes other than the Mazatecs. Valdés notes early anecdotal reports of use by the Cuicatec and Otomí tribes, whose lands adjoin the Sierra Mazateca.21, 9

1.3 Chemistry.

1.3.1 Terpenoids.

1.3.1.1 Isolation.

The first compound isolated from S. divinorum was 1a, a clerodane diter- penoid discovered by Ortega et al in 1982 and named salvinorin.22 The ter- minology of these terpenoids will be discussed in the next section. Extraction of the dried leaves in refluxing CHCl3, chromatography on “Tonsil” activated clay and crystallisation from MeOH gave 1a in unstated yield. The structure 1.3. CHEMISTRY. 37

O

O O O H H O O

O O 1a

Figure 1.5: Ortega et al’s X-ray structure of 1a (stereoview). was elucidated spectroscopically, and confirmed by X-ray crystallography.22 Absolute stereochemistry was tentatively assigned by comparing the circular dichroism spectrum with known compounds.

Subsequently, Valdés isolated the same compound. Unaware of Ortega’s work, he named the compound divinorin A in his thesis.15 He also isolated the deacetyl analogue 1b, which he named divinorin B. When the work was pub- lished, this oversight was corrected, and the compounds were named salvi-

23 norins A and B. Valdés’s isolation procedure was more complex. The Et2O extract was dissolved in MeOH and washed with hexanes. Repeated chro- matography on silica gel and repeated recrystallisation from EtOH gave 1a in 1.8 g/kg yield based on dry weight. The yield of 1b was much lower (74 mg/kg). Spectroscopic structure elucidation of 1a was again confirmed by X-ray crystallography. The fit of the model was in excellent agreement with the diffraction data (R-factor = 8.7 %), though not quite as good as Ortega et al’s (5.2 %). 38 CHAPTER 1. INTRODUCTION.

O O

O O O O O O H H H H HO O O O

O O 1b O O 1c

Figure 1.6: Salvinorins B (1b)andC(1c).

Ortega et al’s model also has the advantage of including the positions of hydro- gen atoms. Unfortunately, the absolute stereochemistry of Valdés et al’s model is incorrect, although it was correctly assigned in the paper, again on the basis of circular dichroism. This stereochemistry has since been definitively con- firmed by the use of exciton chirality circular dichroism,24 and more recently X-ray crystallography,25 on suitable derivatives. This absolute stereochemistry is common to all clerodanes isolated from the Lamiaceae.26

Valdés et al subsequently isolated salvinorin C (1c).27 Repeated chromatog- raphy of the Et2O extract followed by HPLC gave 1c in 78 mg/kg yield. Spectroscopic structure elucidation was supported by partial synthesis of ana- logues.

O HO O H 2 H

H H HO 3 4

Figure 1.7: Known terpenoids.

In addition to these new diterpenoids, several known terpenoids were detected. The monoterpenoid loliolide (2) was isolated and fully characterised by Valdés, inthesamemanneras1c, in 4.4 mg/kg yield.28 GC/MS analysis by Giroud 1.3. CHEMISTRY. 39 et al also detected compounds whose MS data were consistent with the nor- triterpenoid stigmasterol (3) and the diterpenoid neophytadiene (4).29

1.3.1.2 Terminology.

15 O 16 H 14 H 13 O 12 11 H H 1 9 17 2 10 8

5 20 3 7 O 4 6 OAc 19 OAc 5 18

Figure 1.8: Clerodin (5) and standard clerodane numbering.

Clerodane diterpenoids30 are named after clerodin (5). Conclusively establish- ing the structure of this compound was a tortuous process, which has been lucidly reviewed by Rodriguez-Hahn et al.26 The absolute stereochemistry was initially proposed as ent-5. Compounds with stereochemistry matching 5 were therefore termed ent-clerodane for many years. Extensive crystallographic and degradation studies ultimately proved the true structure of clerodin to be 5. To avoid ambiguity, the term “neoclerodane” was therefore coined for com- pounds matching 5.31 Compounds previously known as clerodanes would be termed ent-neoclerodane. For further detail, consult Rodriguez-Hahn et al.26

Clerodanes are further subdivided into trans-andcis- varieties, according to the configuration of C-5 relative to C-10 in the standard32 numbering scheme (Figure 1.8). All clerodane diterpenoids isolated from S. divinorum to date are trans-neoclerodane.

1.3.1.3 Biosynthesis.

Terpenoids are synthesised from C5 “isoprene” units, derived from isopen- tenyl diphosphate (6).33 Assembly of four isoprene units gives geranylger- anyl diphosphate (7), the final common biosynthetic intermediate of all diter- 40 CHAPTER 1. INTRODUCTION.

O O P P = OPP O O OH OH OH 6

OPP

7

Figure 1.9: Biosynthetic precursors of diterpenoids.

penoids.34 Formation of clerodanes begins with protonation at C-14 of 7 (Scheme 1.1). This initiates a cationic cascade leading to the labdane in- termediate 8. A sequence of 1,2-hydride and methyl shifts then gives the trans-clerodane skeleton 9.

It is unknown whether these 1,2- shifts are concerted. In the formation of cis- clerodanes, a discrete halimane intermediate (10) is formed (Scheme 1.2).34 These pathways have been substantiated by feeding plants isotopically labelled mevalonic acid, a precursor to 6. Notably, labelling of the terminal carbon of 7, which becomes C-18 in labdane intermediate 8, results in labelling at C-18 in trans-clerodanes,35 but C-19 in cis-clerodanes,36 as expected (Schemes 1.1 and 1.2). Tritium labelling also confirms the hydride shift from H-5 to H-10.36

Mevalonic acid was used in these studies because it was long assumed to be the sole precursor to 6. Recently, revolutionary work has overturned this as- sumption.37, 38 An alternate pathway, via 1-deoxyxylulose 5-phosphate, has been established. Indeed, the mevalonic acid pathway makes a negligible con- tribution to diterpenoid biosynthesis. Nonetheless, as reflected in the above results, the pathways are not mutually exclusive. Some “crosstalk” occurs, which is increased by the feeding of precursors.38 Nonetheless, incorporation of mevalonate into diterpenoids is very low: below 0.01% in some cases.36 1.3. CHEMISTRY. 41

OPP OPP

14 14

15 15 H+ H+ CH OPP CH OPP 2 2 7 7

+ +

8 8 18 19 18 19 CH OPP CH OPP 2 2

+ +

18 18 19 9 19 9

Scheme 1.1: Cyclization of 7 to the trans-neoclerodane skeleton(stereoview).

OPP OPP OPP OPP

H 9 H 8 10 + 10 5 H + + + H 19 H 18 19 18 7 810 Scheme 1.2: Cyclization of 7 to the cis-neoclerodane skeleton. 42 CHAPTER 1. INTRODUCTION.

0 20 µm 0 20 µm

Figure 1.10: Peltate glandular trichome (SEM stereoview).39

1.3.1.4 Distribution.

In a series of simple and elegant experiments, Siebert has demonstrated that the salvinorins are not evenly distributed through the tissues of S. divinorum, but are localised in particular structures: peltate glandular trichomes (Fig- ure 1.10).40 These are found particularly on the undersides of the leaves - densely packed on newly formed leaves, more sparsely distributed on mature ones (Figure 1.11). Other types of trichome, glandular and non-glandular, are also visible. Terpenoid accumulation in glandular trichomes is typical of the Lamiaceae family.40

This finding is significant for future isolation work. Siebert found that dipping fresh leaves in CHCl3 (30 seconds × 3) gave nearly complete recovery of salvi- norins.40 Powdering the leaves, as has been done in all isolation procedures to date, is therefore unnecessary (see Section 2.1.1 on page 67). 1.3. CHEMISTRY. 43

0 200 µm 0 200 µm

0 200 µm 0 200 µm Above: immature leaf (∼1mmwide);below: mature leaf (∼10 cm wide).

Figure 1.11: Underside of S. divinorum leaves (SEM stereoview).39

1.3.2 Alleged alkaloids.

1.3.2.1 Summary of the Original Report.

The first investigation into the chemistry of S. divinorum was reported by José Luis Díaz in 1975.41 The report is in Spanish, but key sections have been translated by Valdés, who also added flowcharts clarifying Díaz’s procedures.42

The report was not peer reviewed, and the workers involved were not named. Valdés, who later visited the lab, reports that the chemistry was performed 44 CHAPTER 1. INTRODUCTION. by “undergraduate biology and botany students with a minimal chemistry background.”43 Many essential details were omitted. For instance, in some cases TLC data were given without the solvent system. Administration of certain extracts reportedly caused abnormal behaviour and posture in cats. These are, however, not evident in the accompanying photographs. Moreover, there were apparently “great variations” and “inconsistency” in the results, which are not specified: “the described behaviour is not always present.” The assay results are reported simply as “active”, “inactive” or “dubiously active” without further detail, and without specifying the number of subjects or trials. No positive or negative controls were used to validate the assay.

Given these serious deficiencies, the procedures and results will not be repro- duced here. The interested reader will find Valdés’s translation helpful.42 One important omission should be noted: the isolation procedure Valdés describes as method 2 was performed twice: on the first occasion, as he notes, one of the fractions was active. On the second occasion, no fraction was active, yet another “inconsistency.”

The results can be briefly summarised as follows: certain fractions of the plant extract, soluble in aqueous acid but insoluble in aqueous base, some- times caused cats to behave abnormally. Other fractions never showed clear activity. The active, acid-soluble fractions contained at least four compounds which gave positive reactions to Dragendorff’s reagent,44 a standard alkaloid visualisation reagent, in both standard and modified (Lüdy-Tenger)45, 46 forms. On this basis, Díaz concluded that “several alkaloids exist in Salvia divinorum, two of them apparently psychoactive.” In 1977 he reported that the structures of the two compounds were under study.20 In 1979, however, he reported that:

It has been particularly difficult to identify the substance(s) respon- sible for these interesting effects. There exists a great variability or instability in the constituents of S. divinorum, which has im- peded the consistent reproduction of the mental or behavioural al- terations, preventing the identification of the active fraction. Some 1.3. CHEMISTRY. 45

initial observations indicated the presence of nitrogenous compounds, possibly amino acids or amines, although they now appear to be of no pharmacological interest.47

These remarks, while vague, are not consistent with the original report. Pre- sumably attempts were made to replicate the original experiments, but the results proved irreproducible.

1.3.2.2 Discussion.

The phytochemistry of the genus Salvia has been thoroughly studied,3 yielding hundreds of terpenoids.2 Yet extensive literature searches48, 49, 50 revealed no report of an alkaloid from an American Salvia species. The saps of several American Salvias have given positive results to Dragendorff’s reagent,51 but far more tested negative. Furthermore, Dragendorff’s gives false positives with numerous non-nitrogenous compounds.44

The American species S. reflexa gave false positives to several alkaloid test

52 + reagents. The compound responsible proved to be choline [Me3NEtOH] . Being a quaternary ammonium compound rather than an alkaloid, choline could not be extracted from basic aqueous solution by chloroform.52 This is also true of non-nitrogenous Dragendorff’s-positive compounds,44 and thus none of these compounds can account for Díaz’s results. Díaz41 cites a sec- ondary source53 reporting that histamine occurs in the genus Salvia.The primary source54 cited there stresses that histamine is a primary metabolite, formed by decarboxylation of the amino acid histidine. It is therefore described as a “biogenic amine” rather than a true alkaloid, which in the strict sense of the word are secondary metabolites.

Recent work on some Mexican Salvia species has found a very close chemotaxo- nomic relationship with Chinese species,2 some of which have yielded alkaloids. In particular, several seco- and nor-abietane diterpenoids previously known only from the Chinese species S. miltiorrhiza have been isolated from Mexican 46 CHAPTER 1. INTRODUCTION.

Salvias.2 S. miltiorrhiza has also yielded alkaloids.55 Thus, the presence of alkaloids in S. divinorum cannot be dismissed out of hand.

Subsequent work has failed to replicate any of Díaz’s findings, however. Valdés found no compound in the crude extract which gave a positive reaction to Dra- gendorff’s or other alkaloid-specific test reagents.56, 23, 57 He has hypothesised that Díaz’s group actually used mislabelled Erlich’s reagent, which reacts with alkaloids but also furanolactones, and thus mistook the salvinorins for alka- loids.57

This hypothesis, however, is inconsistent with the published claims; all defi- nitely active fractions were soluble in aqueous acid, unlike 1a.Thus,iftheac- tive fractions contained the salvinorins then not only the visualisation reagent, but the fractions themselves, must have been misidentified. Moreover, the ac- tivity reported by Díaz was dramatically different from that later observed in cats by Valdés. Díaz reported “intense attention”, “reactions of fear and attack” and “fury”, lasting 10 minutes after intravenous injection.41, 42 The ef- fects observed by Valdés were almost the opposite. Subcutaneous injection of an extract caused erratic eye movements rather than intense attention, and loss of physical coordination (the cat could not walk, much less assume postures of fear and attack). The effects lasted over 24 hours rather than 10 minutes.58

Chemical investigations by several groups have now yielded a total of 20 ter- penoids, none of which is soluble in aqueous acid (see below). Furthermore, GC/MS29 and LC/MS59 analyses of the crude extract gave no indication of the presence of alkaloids. Compounds containing an odd number of nitrogens give characteristic molecular ions and fragments, identifiable by the nitrogen rule.60

In summary, the claim of biologically active alkaloids was implausible and irreproducible, and has been abandoned by its author. The claim was evidently false. 1.4. PHARMACOLOGY. 47 1.4 Pharmacology.

1.4.1 Animal Testing.

1.4.1.1 Cats and Rats.

As discussed above, Díaz’s tests in cats yielded no useful results. The next worker to study the plant’s pharmacology was Valdés. Although he consumed the traditional infusion during his initial work in Oaxaca, and later tested it for activity after freeze-drying, further use of human subjects was “precluded”61 (i.e. forbidden by his supervisor).

For the purposes of bioassay-guided fractionation, Valdés therefore had to develop an animal assay. This presented an enormous challenge. Previous attempts to identify the active principles of hallucinogenic plants using ani- mal assays, even when sustained and well funded, had failed.7 In each case the active principle was later identified, quickly and cheaply, using human tests. Examples include mescaline (from Lophophora williamsii), psilocybin (Psilocybe cubensis) and lysergic acid amides (Ipomoea and Turbina spp.)62 Similarly, initial testing of LSD in mice caused no apparent effect other than “disquiet”.10 To these must of course be added Díaz’s work with S. divinorum.

Evidently, and unsurprisingly, it is practically impossible to tell if an animal is hallucinating. In Valdés’s preliminary trials, standard hallucinogen assays in rats and cats were not sensitive to the effects of the extract.63 However, the cats exhibited impaired motor coordination, which reminded him of the im- paired physical coordination and slurred speech caused by the infusion.16 The duration of action was much longer, however (24 hours). Given the difficulty of detecting psychological states in animals, Valdés decided to choose an assay sensitive to this physical effect. 48 CHAPTER 1. INTRODUCTION.

1.4.1.2 Mice.

A standard assay of impaired locomotor function in mice, the inverted screen test,64 proved sensitive to the crude extract, but insufficiently so. Next, a modified65 version of the open field assay66 was tested. The movements of mice on a printed grid were recorded over 15 minutes. Three measures of activity were recorded: lines crossed, number of rearings, and time spent immobile.

This assay showed unambiguous effects, with a clear dose-response relation- ship. Assay-guided fractionation led to the identification of 1a as the active principle. No activity was seen with 1b, the only other pure compound iso- lated. However, various mixed fractions gave puzzling results.

60 300 lines crossed rearings 16 immobility 14 250 50 12 200 40 10 150 30 8 6 100 rearings 20

lines crossed 4 50 10 2 time immobile (min) 0 0 0 10 20 30 40 50 100 10 20 30 40 50 100 10 20 30 40 50 100 dose (mg/kg) dose (mg/kg) dose (mg/kg) log scale log scale log scale

1a impure 1a (~10% 1c) mixtures

Figure 1.12: Activity of pure 1a versus mixtures in mice (open field assay).

Using the open field assay, impure 1a (later found to contain ∼10% 1c)was found to be “significantly more potent”27 than the pure compound.65 It was therefore long suspected13, 27, 67 that 1c was also psychoactive.

The open field assay results are shown in Figure 1.12. The potency of the fraction containing ∼10% 1c was at least 10× greater than pure 1a,byall three measures (Figure 1.12). Thus, if Valdés’s interpretation68 were correct, 1c would be at least 100× more potent than 1a,makingitamongthemost potent psychoactive compounds ever discovered.

Three other mixed fractions containing 1a also gave higher efficacies than the pure compound, despite containing as little as 10% 1a.65 At 100 mg/kg, the dilute fractions almost completely immobilised the mice, while pure 1a only reduced activity. Despite their higher efficacy at high doses, however, these 1.4. PHARMACOLOGY. 49 fractions were not in fact more potent: the lines of best fit suggest that pure 1a had comparable efficacy at intermediate doses, and higher efficacy at low doses (Figure 1.12). Thus, these results are not consistent with the presence of a more potent compound in the mixed fractions, but appear to be confounded. The anomaly was not specific to the open field assay: the same pattern was evident69 in the inverted screen assay.

The confounding factor was probably differences in absorption, caused by the unusual method of administration. The test fractions were injected in an emulsion of vegetable oil, water and Tween 80 surfactant.70 The use of oil/water/surfactant emulsions as drug vehicles has proven effective for some hydrophobic drugs,71 and 1a is indeed hydrophobic. However, to be effectively delivered in an emulsion, drugs must also be highly lipophilic:

Generally the most difficult drugs are those which have limited solubility in both water and lipids (typically with log P values of approximately 2). It is unlikely that lipid formulation will be of value for such drugs.71

The solubility of 1a is negligible in hexanes, and presumably in oil (predicted log P = 1.8).72 Thus, this vehicle would be expected to give poor absorption; 1a may have been administered not in solution, but as a suspension.

The 1a tested by Valdés was crystalline. Two of the impure fractions were explicitly described as “oily solids.”73 The other two were not described; however, they were mixtures, and were obtained by evaporation from 10%

73 MeOH/CHCl3. Evaporation of 1a from chlorinated solvents gives an amor- phous solid even when pure. Hence, all four of the impure fractions shown in Figure 1.12 were amorphous. This would result in differences in solubility; the amorphous state of a solid typically has 2-10× higher solubility than the crystalline state.74 This results from the energy barrier to dissolution of a crystalline solid: disruption of the ordered crystal lattice requires additional

74, 75 energy (the enthalpy of fusion, ∆Hf ). Also, in a vigorously agitated emul- sion, the amorphous solids could form a dispersion of microscopic particles, 50 CHAPTER 1. INTRODUCTION. with a higher surface area than macroscopic crystals. This in turn would in- crease the rate of dissolution (kinetics of solubility).75 Thus, faster dissolution and greater absorption of the amorphous fractions would be expected.

Other compounds present in the mixtures might also influence absorption. Many instances have been reported of enhanced absorption of active com- pounds from a crude extract relative to the pure compound.76 For instance, some terpenoids are known to act as permeation enhancers,77 increasing trans- dermal absorption of co-administered drugs up to 90×.Thus,inactivecom- pounds in the crude fractions may have increased absorption of 1a.

Recent evidence conclusively establishes that the emulsion was very poorly absorbed. When injected in solution (EtOH/surfactant/H2O), 1 mg/kg of 1a reduced locomotor activity significantly over 30 minutes.78 Reinvestigation by Valdés et al confirmed this, showing that 0.5 mg/kg caused maximal im- pairment in the inverted screen test;79 quadrupling the dose did not increase efficacy. By contrast, in the original tests using the emulsion, maximal im- pairment only occurred above 1500 mg/kg.69 Thus, absorption of 1a from the emulsion was clearly negligible.

The use of an emulsion also seems to prolong the effect of 1a; performance on the inverted screen test remained impaired after 30 minutes.15 By comparison, the effects of the solution peaked within 5 minutes, and were undetectable by 15 minutes.79 Consistent with this, a recent study found a strong analgesic effect of 1 mg/kg injected 1a solution at 10 minutes,80 while a previous study which began testing 20-25 minutes after injection found very little analgesic effect even at 40 mg/kg.81 Another study reported analgesic effects at 0.6 mg/kg, but unfortunately contained no detail on administration or timing.82

In summary, the open field results may have been confounded by differences in bioavailability. The vehicle used gave very poor absorption, and superior absorption of amorphous solids is typical. Thus, the differences in apparent potency between pure 1a and the impure fractions do not provide reliable evidence of the presence of other active compounds. Evidence that 1c and 1.4. PHARMACOLOGY. 51 other compounds in the plant are inactive will be presented in the next section, and in section 4.5.1 on page 171.

1.4.1.3 Rhesus Monkeys (Postscript).

Studies administering 1a to rhesus monkeys have recently been reported.83, 84 Interestingly, even the highest dose tested (32 µg/kg) produced only “slight overt behavioural effects”,83 described as “sedation-like” in both studies, con- firming the difficulty of establishing animal models of psychoactivity.

1.4.2 Human Testing.

As mentioned above, Valdés’s open field assay results indicated that 1a was the active principle of S. divinorum. Further testing with other compounds in the open field indicated that 1a was approximately equal in potency to mescaline.13, 14 However, this conclusion remained tentative. As Valdés had written earlier,

... there is no definite evidence that divinorin A is an hallucinogen ... the results are as yet unclear. And they will probably remain so until the divinorins are tested in human beings.68

The question was finally resolved by Siebert in 1994.18 He reported that 1a was hallucinogenic when vaporised and inhaled. Activity was percepti- ble with doses as low as 200 µg. The effects commenced within seconds, and the strongest effects lasted 5 – 10 minutes. No effects were detectable after 30 minutes, suggesting a half-life of under 10 minutes. Siebert also explored the effect of the route of administration. Swallowing encapsulated 1a in very large doses (10 mg) produced no effect. This also held for the plant: an in- fusion prepared from ten fresh leaves, when swallowed, produced no effect in any subject. The same amount, when held in the mouth for 10 minutes and spat out, produced definite effects in all subjects. Thus, absorption through 52 CHAPTER 1. INTRODUCTION. the oral mucosa is clearly far more efficient than through the gastrointestinal system. Nonetheless, an ethanolic solution of 1a gave inconsistent effects sub- lingually. The onset of effects is slower by the sublingual route, and the effects last longer.

Siebert’s results were soon confirmed by others.7, 85 Ott and Gartz86 also re- ported that sublingual application of 1a was effective in acetone or Me2SO, with potency comparable to inhalation. Thus, far from being equipotent with mescaline as the open field assay indicated, 1a is in fact ∼1000× more po- tent.14, 7 Indeed, it is the most potent naturally-occurring hallucinogen yet isolated.7 The rapid metabolism of 1a has been confirmed in vitro87 and in vivo.88

Threshold doses of 1a produce visions of coloured patterns overlaid on reality, reminiscent of the pseudohallucinations induced by indole and phenethylamine hallucinogens. These patterns are faint, and only perceptible under dark and quiet conditions.14 Higher doses, however, produce intense and unique effects like those described earlier (Section 1.2 on page 33). Awareness that the visions are drug-induced is lost, and true dreamlike hallucinations occur. These expe- riences frequently involve certain themes not found with other hallucinogens. The distinction between self and surroundings is lost; subjects often feel that they are blending into, or have become, inanimate objects.18 Similarly, the distinction between past and present is weakened; subjects will relive events, often from childhood, rather than simply remembering them. Siebert’s initial reports of these unique themes have again been confirmed by others.85, 89

Further research has revealed no evidence of active compounds other than 1a. Siebert confirmed that 1b is inactive.40 He also found that self-administration of 3 mg90 of 1c, sublingually in acetone, had no noticeable effect;40 this is 10× a threshold dose of 1a by that route. Evidence that 1c and other terpenoids in the plant are inactive in vitro will be presented in Section 4.5.1 on page 171.

In conclusion, although definitive proof awaited human testing, Valdés nonethe- less correctly identified the active principle of S. divinorum using animal as- 1.4. PHARMACOLOGY. 53 says. The magnitude of this achievement is not widely appreciated. As noted above, previous attempts to identify plant hallucinogens using animal assays had invariably failed, while human assays had succeeded, quickly and cheaply. Valdés’s supervisor had thus forbidden the only demonstrably effective tech- nique available. Moreover, in this case even human testing had failed. And it had been performed, independently, by arguably the two greatest authorities in the field, Alexander Shulgin91 and Albert Hofmann92, 10 (twice in the latter case).93 Consider also the prevailing consensus when Valdés began work: the active principle, whose effects were barely perceptible, was water-soluble and well-absorbed orally, but unstable and destroyed by drying. This consensus proved false in every detail. The active compound(s) also appeared to be al- kaloid(s), which was also false. Sadly, despite overcoming these myths, Valdés considers the time he spent on animal testing wasted, believing it cost him priority on the discovery of 1a.94

1.4.3 In vitro Testing.

After establishing that 1a was psychoactive, Siebert submitted the compound for in vitro screening against potential molecular targets.18 The NovaScreen assay tested for radioligand binding inhibition at targets including receptors for small-molecule and peptide neurotransmitters, as well as ion channels and enzymes. No binding was detected at 10 µM.

Subsequently, the compound was screened by Roth et al against a much larger battery of targets, the National Institute of Mental Health’s Psychoactive Drug Screening Program. This revealed that 1a bound with high affinity to the κ

67 opioid receptor (K i =4nM). Functional testing showed that it activated the receptor with full efficacy and high potency (EC50 =1nM).Thus,1a is a potent full agonist at the κ opioid receptor. This result has since been replicated by other groups.82, 81, 95 Further confirmation has come from in vivo testing. The effects of 1a in mice are blocked by selective κ opioid antagonists;81, 78, 80 rhesus monkeys trained to discriminate κ opioids recognised 1a as such, and 54 CHAPTER 1. INTRODUCTION. the effects were blocked by a nonselective opioid antagonist.83

No binding was apparent in vitro to any of 48 other CNS targets at 10 µM; 1a is thus extremely selective compared to most other psychoactive compounds.67 However, the strength of this conclusion is contingent on the number of targets tested; many orphan receptors exist for which affinity testing cannot yet be performed.96 However, in vivo confirmation of the selectivity of 1a is available. Siebert found that the effects of 1a were blocked by naloxone, a nonselective opioid antagonist.97 This makes it highly unlikely that a nonopioid mechanism contributes independently to the compound’s effects.

1.4.4 Mechanism: κ Opioids.

1.4.4.1 Discovery.

The strongest painkiller available in the preindustrial era was opium, a milky secretion of the opium poppy Papaver somniferum. The isolation of morphine (11) from opium was reported by Sertürner in 1806; he also showed the com- pound to be a potent analgesic.98

N H

11 HO O H OH

N O N N OH

O HO HO HO O 12 13 14

Figure 1.13: Morphine (11) and derivatives.

Morphine has had an immense impact on science. It was the first drug – the first known pharmacologically active compound.99 It was the first alkaloid; 1.4. PHARMACOLOGY. 55 indeed, the word alkaloid was coined to describe it.98 Its structure elucidation took over a century, which is unsurprising since morphine’s discovery predated the concept of molecular structure by several decades. Indeed it predated the publication of Dalton’s atomic hypothesis, which was regarded with skepticism for decades.100

Morphine’s very strong analgesic effect is accompanied by serious side effects: nausea, respiratory depression and constipation.101 Intense euphoria also oc- curs in some, making the drug highly addictive. The pursuit of an analgesic lacking these side effects gave rise to medicinal chemistry. The first morphine derivatives were synthesised in the 1850s;98 the structure-activity relationships of morphine have since been explored more thoroughly than those of any other compound, with thousands of derivatives synthesised.101 Simplified structures were found to retain activity: morphinans such as butorphanol (12), ben- zomorphans such as ketazocine (13) and remarkably even phenylpiperidines such as pethidine (meperidine, 14).

Some of these compounds (such as 14) closely mimic morphine’s actions. Some, however, proved to be antagonists, which have proven immensely valu- able in reversing opioid overdose. Other derivatives caused quite distinct be- havioural effects, but were not antagonists and did not exhibit cross-tolerance. The study of these differences led to the discovery of opioid receptor subtypes. The µ and κ subtypes were named after the archetypes morphine (11)and ketazocine (13), while δ comes from the vas deferens, in which that subtype was discovered.101, 102 These receptors are commonly abbreviated as MOR, KOR and DOR. Neither of these terminologies is endorsed by the International

102, 78 Union of Pharmacology, but the officially sanctioned names, OP1−3, have not gained wide acceptance. The study of opioid receptors led to the discovery of the endogenous ligands, the endorphins.101, 102 Remarkably, recent work has proven that morphine is endogenous in humans.103

Opioids are compounds which act at opioid receptors, and whose effects are reversed by the antagonist naloxone.101 This terminology lends itself readily to 56 CHAPTER 1. INTRODUCTION. selective compounds; thus, 1a will be referred to below as a “κ opioid.” This is synonymous with the common but unwieldy tautology “κ opioid receptor agonist.”

1.4.4.2 Development.

N N

N 16 R O O N

O Cl Cl Cl Cl 15 O 17 18

Figure 1.14: Arylacetamide κ opioids.

U50,488 (15) was devised as a structurally simplified morphine derivative.104, 105 In animal tests, U50,488 was an effective analgesic, whose effects were reversed by naloxone. Remarkably, however, the compound did not produce physical dependence like morphine, and was not self-administered. Initially referred to as a non-µ opioid, 15 was soon found to be the first selective agonist at the κ opioid receptor.

These remarkable findings inspired extensive research; numerous derivatives were synthesised, some of which proved to be even more potent and selective.105 Examples include U69,593 (16), spiradoline (U62,066, 17) and enadoline (CI- 977, 18). Testing of these compounds confirmed the findings with 15:atlastit appeared that the long-awaited nonaddictive opioid analgesics had been found.

1.4.4.3 CNS Effects.

Addiction can be studied using animal models. The µ opioids are positively reinforcing. This means test animals trained to self-administer them will do 1.4. PHARMACOLOGY. 57 so compulsively, at the expense of social activity, food and sleep. By contrast, κ opioids (including 1a)78 are negatively reinforcing, or aversive. Test animals will not self-administer them; if administered when an animal behaves in a certain way, the animal will avoid that behaviour. In other words, κ opioids act as a punishment, where µ opioids act as a reward. This is a vast field of research in its own right, and a thorough review is available.106 The mechanism of these effects, believed to be mediated by dopamine, has also been thoroughly studied.78, 107

Since the objective had been to eliminate the euphoria associated with opioids, the aversive nature of κ opioids was initially regarded as desirable. Human tests soon revealed a problem: rather than euphoria, κ opioids produced dys- phoria. This lowering of mood was accompanied by “psychotomimetic” effects: confusion, hallucinations and depersonalisation.

Enadoline (18), for instance, caused side effects such as somnolence, hallucina- tions, anxiety, depersonalisation, confusion and abnormal thinking; the sever- ity of these effects led to termination of clinical development.108 Subsequent tests with higher doses reported visual, auditory and tactile hallucinations, along with impaired coordination and recall. One subject felt waves moving through the floor, and felt his body was blending into them109 (cf. 1a).18 Spi- radoline (17) caused altered perceptions, impaired coordination and slurred speech; subjects reported being more irritable, anxious and sad.110 Responses to other κ opioids range from “personality disorders and mild confusion”111 to “depersonalisation”, “dreamlike” experiences and “episodes of unmotivated and uncontrolled laughter”112 (again cf. 1a).18

In one study, heroin users were given various substances and asked “do you like the drug?” and “does the drug have any good effects?” They consistently answered yes after administration of a µ opioid, but no after enadoline (18).109 The opposite was true when asked about “bad effects.” Similar results were reported in tests of nonselective κ opioids such as 13113 and 12.114 Subjects with extensive histories of illicit drug use reported dysphoria, hallucinations, 58 CHAPTER 1. INTRODUCTION. paranoia and confusion after the κ opioids, but euphoria after morphine. Thus, both drug-naïve subjects and experienced illicit drug users overwhelmingly find κ opioids dysphoric and aversive.

1.4.4.4 Therapeutic Potential.

Clinical development of these compounds as analgesics was eventually aban- doned due to these psychological effects.115, 105 Some κ opioids which do not cross the blood-brain barrier, and are therefore not psychoactive, remain in de- velopment for arthritis116 and abdominal pain.117 This is interesting in light of Mazatec use of S. divinorum infusion against abdominal pain and swelling.11, 15 Besides analgesia, many other therapeutic uses have been proposed for κ opi- oids.115, 118, 119, 120 None of these therapies has reached clinical use.

There has been a report of an antidepressant effect of U50,488 (15)inananimal model.121 However, there was at the time no validated protocol for that model (learned helplessness in mice),122 and compounds of known activity were not used as controls. In another model, the forced swim test in rats, one study reported that a κ opioid had no effect.123 However, other studies reported a prodepressant effect,124 one of which used 1a.107 This conclusion, which is consistent with the aversive and dysphoric effects discussed above, is also more consistent with prior results and recent theoretical insights.107, 125 Thus, it seems that κ opioids exacerbate depression in animal models. Surprisingly, however, there have been reports of antidepressant effects from S. divinorum use.126, 127 However, the plant was used only three times weekly. It may be that, just as the acute euphoria caused by µ opioids is followed by prolonged dysphoria, the converse may be the case with κ opioids.

1.4.4.5 Claimed Subtypes.

There have been claims that certain κ opioids lack dysphoric effects. TRK 820 (19, Figure 1.15 on the facing page) allegedly produces “moderate be- havioural/psychological side effects, but not psychotomimetic activity”, but 1.4. PHARMACOLOGY. 59 this is based on “unpublished data”.128 It is also often claimed that 19 is not aversive in animals,128 which is untrue.129 This is reminiscent of earlier claims that enadoline (18) “appears to be devoid of psychotomimetic activity,”130 which was subsequently and spectacularly discredited as mentioned above.

One proposed mechanism for differences between κ opioids is that there are subtypes of κ opioid receptor.131 Differences between κ opioids might reflect different selectivities for these subtypes. The evidence for subtypes is purely pharmacological131 — the gene encoding the receptor has now been cloned from several species, with no subtypes detected. For this reason, and because of the lack of selective ligands for each proposed subtype, the claim has not won general acceptance.132 The apparent subtypes may in fact represent dimer formation between different receptors.131, 132

1.4.4.6 Structure, Potency and Selectivity.

N O N OH O

HO O H N HO 19 O 13

phenylalkylamine moiety

N N N O F HN O O N S

O 18 20

Figure 1.15: Structurally diverse κ opioids.

The major structural classes of κ opioid are peptides such as dynorphin A, mor- phine derivatives such as 19, benzomorphans such as 13 and arylacetamides such as 18 (Figure 1.15). The first major departure from these categories was the benzodiazepine tifluadom (20).105 The discovery of such compounds, and 60 CHAPTER 1. INTRODUCTION. similarly diverse ligands at other subtypes, eventually rendered generalisations about structure-activity relationships impossible.101

Nonetheless, as diverse as these compounds may appear, they are all alka- loids. 1a was the first non-nitrogenous opioid reported. Indeed, it was the first opioid lacking the phenylethylamine moiety or its propyl homologue (Fig- ure 1.15). These moieties are near-ubiquitous in hallucinogens,133 and indeed in psychoactive compounds generally. In a random sample of compounds from the Merck Index, 82% of psychoactive compounds contained a phenylalky- lamine moiety, versus 8% of non-psychoactive compounds.134 Conversely, 58% of phenylalkylamines were psychoactive, versus 3% of other compounds. Be- ing non-nitrogenous and lacking a benzene ring, 1a is thus not merely unique among opioids, but extremely unusual among psychoactive compounds in gen- eral.

EC50 (nM) Ref 1a 15 16 19 11.2 67 72413 135 4.5 4.5 95 45 207 136 3.1 3.9 137 4.6 2.2 0.025 81 1.1 0.0048 138 16 0.15 139 Table 1.1: Relative potency of 1a at cloned κ opioid receptors.

The binding affinity and potency of 1a in vitro are close to those of U50,488 (15) and U69,593 (16). Potencies are shown in Table 1.1. The most potent κ opioid to date appears to be TRK 820 (19). However, few groups have studied this compound, and their binding affinity data are wildly discordant

81 139 (K i =75pM vs 3.5 nM). The compound also has lower µ/κ selectivity than U50,488138 and other arylacetamides,104 which are in turn less selective than 1a.67 Indeed, there has been no report of 1a showing any affinity at the

µ opioid receptor (K i >10µM).

In summary, κ opioids have not lived up to initial expectations as analgesics. 1.5. TOXICOLOGY. 61

Ultimately, the enormous effort to improve upon morphine has failed. Two hundred years after its discovery, morphine remains “the opioid of first choice” for severe pain under World Health Organization guidelines.140 There is no superior analgesic:

Morphine remains the most widely used opioid for the manage- ment of pain and the standard against which other opioids are compared.141

Opioids remain the topic of extensive research, however, for other therapeutic purposes, as mentioned above.

1.5 Toxicology.

There has been little toxicological research on S. divinorum and 1a.Valdés made several incidental observations on the topic. One of his rats died during oral administration of an extract, but this was due to choking on the large volume of liquid rather than toxicity.142

The crude aqueous extract proved highly toxic by injection in cats. Subcu- taneous injection of 0.7 g/kg, approximately equivalent to a human dose of the infusion, caused a sterile abscess at the injection site.58 Ahigherdose(1.3 g/kg) caused kidney failure in two cats, one of which died. The other recovered with a sterile abscess. Based on the known toxicity of polyphenols in crude ex- tracts,143 and veterinary examinations suggesting this was the cause,58 Valdés prepared a polyphenol-free extract by partitioning between CH2Cl2/MeOH and water. The resulting organic fraction was not toxic, even in much higher doses, estimated to be equivalent to 635× the human dose.58

Further toxicity testing was performed in mice. The ether extract of the leaves was dissolved in aq. MeOH and washed with hexanes. The resulting methano- lic fraction, which was presumably free of polyphenols, was nevertheless still

69 toxic enough to kill the mice within a week (LD50 = 340 mg/kg i.p.) This is 62 CHAPTER 1. INTRODUCTION. approximately equal to the toxicity of vitamin B3.144 Pure 1a caused no fatal- ities at 1 g/kg,13 the largest dose which could be administered as an emulsion, and more than 105 × the psychoactive dose in humans.

As discussed above, however, the emulsion was very poorly absorbed. This would reduce the immediate systemic dose of the extracts, giving a mislead- ingly low impression of acute toxicity. On the other hand, the undissolved material at the injection site may have had toxic effects through another mech- anism, such as shock. It would also presumably slowly dissolve, causing chronic rather than acute toxicity, especially to the liver. Surfactants such as Tween 80 can also have adverse effects.145 These data therefore cannot be used to make reliable comparisons with other substances.

Mowry et al later published a study specifically devoted to the toxicology of 1a.146 Injection of 6.4 mg/kg/day for two weeks caused no fatalities; post- mortem examination of organs and tissues showed no apparent changes. Un- fortunately, since the drug was administered intraperitoneally as a “fine sus- pension,” and behavioural effects were not described, the same issue of bioavail- ability arises as with Valdés’s results. It is also regrettable that larger doses were not administered to determine the LD50.

Human toxicological data are not available. There have been no published reports of toxic effects or hospitalisation related to S. divinorum or salvinorin A.147

1.6 Social impact.

1.6.1 Recreational Use.

S. divinorum came to the attention of recreational drug users even before chemical investigations began. Books published in 1973 described its use148 and cultivation,149 listing a mail-order source for live cuttings. By 1975, Ott reports seeing some young Mexicans smoking the dried leaves recreationally.7 1.6. SOCIAL IMPACT. 63

Although he claims Díaz also reported this at the time, the latter was in fact referring to a different plant.41

The plant nonetheless remained extremely obscure until Siebert’s findings were popularised in 1996 by Turner.85 This immediately overturned the plant’s reputation as a weak substitute for illicit drugs:

Salvinorin A is the most potent naturally occurring psychedelic known ... it frequently induces experiences of an intensity [...] be- yond those experienced with any other psychedelic ...85

The plant’s notoriety then began to grow. Online companies began selling live cuttings, dried leaves, extracts and tinctures for recreational use. Interestingly, however, after being widely available for a decade, the drug has not had any noticeable social impact (as for instance LSD and MDMA rapidly did). As Turner noted:

Many who have used salvinorin A find the experience extremely unnerving, frightening and overly intense. Most have no desire to repeat the experience.85

This is consistent with the dysphoric and aversive qualities discussed above.

1.6.2 Legal Status.

1.6.2.1 Australia.

In 2002, Australia became the first country to prohibit S. divinorum and salvi- norin A. The committee responsible, the National Drugs and Poisons Schedule Committee (NDPSC), justified the decision “on the basis of high potential for abuse and risk to public health and safety.”150

Salvinorin A was prohibited under a purportedly systematic name: 64 CHAPTER 1. INTRODUCTION.

8-METHOXYCARBONYL-4A,8A-DIMETHYL-6-ACETOXY- 5-KETO-3,4,4B,7,9,10,10A-SEPTAHYDRO-3-(4-FURANYL)- 2,1-NAPHTHO[4,3-E]PYRONE.150

This name is in fact meaningless. Compare the one used by Chemical Abstracts Service:

2H -Naphtho[2,1-c]pyran-7-carboxylic acid, 9-(acetyloxy)-2-(3-furanyl)dodecahydro-6a,10b-dimethyl-4,10-dioxo-, methyl ester, (2S,4aR,6aR,7R,9S,10aS,10bR)-48

NDPSC CAS - 2S,4aR,6aR,7R,9S,10aS,10bR 8-METHOXYCARBONYL- 7-carboxylic acid, methyl ester 4A,8A-DIMETHYL- 6a,10b-dimethyl- 6-ACETOXY 9-(acetyloxy)- 5-KETO 4,10-dioxo- 3-(4-FURANYL) 2-(3-furanyl) 3,4,4B,7,9,10,10A-SEPTAHYDRO dodecahydro- 2,1-NAPHTHO[4,3-E]PYRONE 2H-Naphtho[2,1-c]pyran Table 1.2: Alternate systematic names for 1a.

Enumerating all the errors in the NDPSC name would be beyond the scope of this work, but a brief selection will be given here. The names are contrasted in order of functional group in Table 1.2. Note especially the nonexistent terms “4-furanyl” and “septahydro.” The latter was presumably intended to mean heptahydro, which is incorrect.

3 2 4 2 1 O 3 5 4a 10 1 10a 4 6 9 10b 4b 10a 4a 10 5 7 8a 8 6a 8 9 7 6 phenanthrene 2H-naphtho[2,1-c]pyran

Figure 1.16: Relevant IUPAC fused-ring numbering schemes.

The name given by the NDPSC originated in a post to the internet newsgroup alt.drugs by William E. White: 1.6. SOCIAL IMPACT. 65

I tried to get an IUPAC name out of this; unfortunately, I gagged when trying to decide whether it was a naphthopyrone, or a modi- fied phenanthrene, or whether it should start from cyclohexane in- stead of aromatic rings, or whatever. The fact that phenanthrene is numbered funny didn’t help. The best I could do is

8-methoxycarbonyl-4a,8a-dimethyl-6-acetoxy-5-keto-3,4,4b,7,9,10, 10a-septahydro-3-(4-furanyl)-2,1-naphtho[4,3-e]pyrone.

(whew!) Sorry, I haven’t done o-chem since the Reagan adminis- tration.151

This explains the incorrect numbering of 1a: White named it as a naph- thopyran derivative, but numbered it as a phenanthrene derivative (see Figure 1.16).152 The resulting name was then copied without acknowledgment by the NDPSC, with the sole change being capitalisation.

After these criticisms153 of their decision attracted publicity,154 the NDPSC revisited the issue in late 2002. Conceding that the name used was incorrect, they decided to adopt the CAS name.155 However, they rearranged the sub- stituents in nonalphabetical order, in conformance with their own unspecified “naming conventions.” The new name is therefore also incorrect.

The NDPSC is Australia’s peak body for the regulation of chemicals, whose chief task is to list controlled chemicals by name. It is surprising and dis- turbing, therefore, that no member of this body understands basic chemical nomenclature. White’s posting also raises another serious issue. The news- group alt.drugs156 is a forum for users, dealers and manufacturers of illicit drugs to share ideas and advice. Publications which “promote, incite or in- struct in matters of crime” cannot be legally imported into Australia, as for instance by downloading from the internet.157 How, then, did the committee legally obtain this material? Committee members repeatedly refused to reveal the source to a federal shadow minister.158, 159 Similarly, documents released under a freedom of information request160 contained nothing on this topic. 66 CHAPTER 1. INTRODUCTION.

A disturbing question thus remains unanswered: how did obvious errors in a prohibited publication find their way into Australian law?

1.6.2.2 Other Countries.

In 2003, Denmark became the second country to prohibit S. divinorum and salvinorin A.161 William White’s erroneous name was used again, with the locants capitalised. This new and peculiar error suggests that the authors copied the Australian version, changing the substituents to lower case, but leaving the locants in capitals. Ironically, the NDPSC had by this time already admitted the name was incorrect and adopted a new one.

S. divinorum and/or salvinorin A have since been prohibited in other countries, including Belgium, Italy, South Korea and Spain.162 There have been similar moves in several US states, but a 2002 bill163 to prohibit the plant nationally met with well-organised opposition,147 and was not enacted.

1.7 Summary.

Previous work had thus established that 1a was a psychoactive κ opioid. How- ever, virtually nothing was known of its structure-activity relationships, and only three other compounds had been isolated from S. divinorum. The suspi- cion that other psychoactive compounds were present persisted. Chapter 2

Isolation.

2.1 Isolation Procedure.

2.1.1 Extraction Conditions.

The isolation methods reviewed above suffer from a number of drawbacks. Both Valdés et al23 and Ortega et al22 employed refluxing solvents for the ini- tial extraction of the dried leaves. Subsequently, however, Gruber164 reported that extraction at room temperature was superior, based on HPLC analysis. He found that steeping for several days at room temperature gave markedly better recoveries of 1a than refluxing for two hours, in either CHCl3 or MeOH (see Table 2.1 on the next page). In optimising the room-temperature ex- traction, he found that highest recoveries of 1a were achieved after less than 30 minutes; levels then declined steadily (by nearly 50% over two days).164 Gruber speculated that this decline might be due to 1a precipitating out of solution as other compounds accumulated. This explanation is implausible since 1a is freely soluble in CHCl3 (evaporation gives an amorphous resin rather than a crystalline precipitate), and the solutions were extremely di- lute. Furthermore, the decline is faster at reflux, when solubility is greater. A more plausible explanation of the decline is decomposition, with the rate proportional to temperature. This conjecture was confirmed in the course of

67 68 CHAPTER 2. ISOLATION. this work. When chromatographically pure 1a was recrystallised from EtOH, and the resulting mother liquor recrystallised in turn, new higher spots be- came apparent by TLC (70% Et2O/petrol). After another recrystallisation, the mother liquor (18% of the original fraction) consisted almost entirely of the higher spots, with only traces of 1a. The decomposition products were not characterised. In contrast, however, Valdés reports that he recovered 1a

165 unchanged after refluxing in MeOH or MeOH/H2O for 2 weeks. It may be that decomposition only occurs in the presence of oxygen.

recovery (mg/g) TtimeCHCl3 MeOH rt 4 d 1.75 1.56 reflux 100 min 1.41 0.44 Table 2.1: Effects of solvent and temperature on recovery of 1a.

After his optimisation experiments, Gruber adopted a standard procedure of steeping the dried, powdered leaves in CHCl3 for 30 minutes. For scale-up, however, this presents the problem of handling and evaporating large volumes of this carcinogenic solvent. Therefore, for this work a less toxic substitute was sought, of lower polarity than MeOH (given the inferior recoveries achieved in that solvent). Acetone was selected as an affordable solvent meeting both of these criteria. Dried, powdered leaves were steeped for 30–60 minutes in room-temperature acetone (× 3). TLC analysis (5% MeOH/CH2Cl2)showed that the second extract was much lower in 1a than the first; the third extract showed no detectable 1a. A further steep in CHCl3 also showed no 1a.This extraction procedure was therefore adopted for all of the work described below.

Further improvements are possible. Powdering the leaves, for instance, appears to be unnecessary. Pseudonymous reports166 on the internet that 1a can be extracted in high yield from intact leaves have been confirmed: Siebert found that dipping fresh leaves in CHCl3 (30 seconds × 3) gave excellent recovery of terpenoids with minimal contamination by pigments.40 Drying and powdering the leaves, followed by a further extraction, gave a pigment-rich extract with negligible terpenoids. Thus powdering the leaves before extraction is unneces- 2.1. ISOLATION PROCEDURE. 69 sary, and indeed counterproductive. While this analytical procedure has not yet been tested on a preparative scale or on dried leaves, these results strongly suggest that extraction procedures based on brief steeping of intact leaves will prove superior to the protocol given above. Claims166 that extraction in chilled solvents gives a cleaner extract also deserve investigation.

2.1.2 Problems Caused by Pigments.

2.1.2.1 Interference with Previous Isolations.

1a 1b

youngest oldest

Developed in 50% EtOAc/petrol, visualised in vanillin/H2SO4.

40 Figure 2.1: Siebert’s TLC analysis of crude CHCl3 extracts.

The extract from the above procedure was a blackish-green tar. The numer- ous pigments present in this extract rendered TLC analysis difficult. Relative to the colourless terpenoids of interest, the pigments were present in larger quantities and stained more vividly, obscuring the terpenoid spots. The prob- lem is illustrated in Figure 2.1. Daniel Siebert analyzed chloroform extracts of leaves of varying age, which had been dried and powdered.40 In all cases, the intensities of pigment spots were equal to or greater than 1a,andmuch greater than 1b. The pigments thus present a major obstacle to the isolation of other terpenoids, which are present in lower levels. For instance, 1c is not even detectable in the figure; in this system it appears immediately above 1a, which is itself nearly obscured in the youngest leaf extract.

Valdés et al had addressed this problem by partitioning the crude extract be- tween aqueous MeOH and hexanes.23 Using his partitioning procedure proved difficult: both phases were black, and could not be distinguished even by close 70 CHAPTER 2. ISOLATION. inspection in direct sunlight. The interface was ultimately located by holding a torch against the separatory funnel and looking directly into the beam, which was dimly visible through the MeOH phase. The hexane phase was totally opaque. Another drawback to the procedure was the tendency of the MeOH phase to spatter during evaporation, but the water content (10%) was far too high to use a drying agent such as MgSO4. After partitioning, while there was some reduction in mass, the extract remained intensely coloured.

Gruber also encountered difficulties with Valdés et al’s procedure, obtaining greenish crystals of 1a even after partitioning, repeated chromatography and recrystallisation.167 HPLC analysis indicated purity of approximately 85% based on UV detection at 208 nm;168 however this cannot be translated to percentage by mass, since the coloured impurities would be expected to have higher molar absorptivities. Gruber found that the trace pigments could be re- moved by rinsing with cold MeOH.167 It was presumably to remove these trace pigments that Valdés recrystallised 1a asecondtime,forwhichnoexplanation was given.23 This repeated recrystallisation is clearly a major drawback to the procedure, adding complexity and reducing recovery.

Others have simplified Valdés et al’s procedure, omitting the chromatography and simply washing the crude extract with less polar solvents to remove the pigments.86, 166 These procedures, while ingeniously simple, have not yet been replicated in the peer-reviewed literature. A recent paper took the opposite approach, omitting partitioning and recrystallisation in favour of chromatog- raphy. Tidgewell et al169 isolated 1a in very high yield as a green powder, after repeated chromatography of the crude extract on silica gel. Despite the colouration, the material exhibited a reasonably sharp melting point range of 3 ◦C; other evidence of purity was not reported.

In summary, the pigments present in the extract caused considerable prob- lems in the isolation of 1a. They therefore posed even greater obstacles to the isolation of the other diterpenoids present in S. divinorum, since these are present in much smaller quantities than 1a. Indeed, as noted at the start 2.1. ISOLATION PROCEDURE. 71 of this section, the pigments made the mere detection of other compounds by TLC difficult. It was therefore critical to this work that the crude ex- tract itself be decolourised, rather than individual compounds after isolation. Valdés et al’s partitioning procedure was unsatisfactory, as was chromatog- raphy on silica gel. Gruber found that reverse-phase chromatography (using a C-18 solid-phase extraction cartridge), while removing some pigmentation, was likewise inadequate.167 Both of these techniques rely upon differences in polarity to achieve separation. Since the pigments exhibited a very wide range of polarities, spreading from baseline to solvent front on TLC (Figure 2.1), another basis for separation was clearly necessary.

2.1.2.2 Chemistry of Plant Pigments.

O O NN O Mg 21 O N N O

OH 22 HO O OH

O OH O OH O 23 OH OH

Figure 2.2: Representative major plant pigments.

The overwhelming majority of plant pigments fall into three categories: chloro- phylls (eg. chlorophyll-a, 21, Figure 2.2), carotenoids (eg. β-carotene, 22)and flavonoids (usually present as glycosides eg. quercitrin, 23).170, 171 In addition 72 CHAPTER 2. ISOLATION. to these ubiquitous classes, there are many minor pigments confined to partic- ular taxa, which have been reviewed elsewhere.170 The common factor among these pigments is an extended π system, often but not always aromatic, which results in intense absorption of visible light. As a result, pigments are gener- ally large molecules. They vary widely in polarity, from extremely hydrophilic (eg. 23) to extremely hydrophobic (eg. 22). The chlorophylls themselves vary: one of the earliest procedures for separating chlorophylls a and b was partitioning between aq. MeOH and petrol.171 Thus, it is unsurprising that Valdés’s partitioning procedure, and polarity-based fractionations in general, are of limited effectiveness in decolourising plant extracts.

2.1.3 Use of Activated Carbon.

The standard method for decolourisation of a solution is adsorption, and the most common adsorbent is activated carbon, often referred to as decolouris- ing carbon.172 Activated carbon has been in widespread use, both scientifi- cally and industrially, for centuries. The characteristics of compounds which affect adsorption are well-established empirically.173 Within a homologous se- ries of compounds, adsorption increases with length. In addition, aromatic compounds are strongly adsorbed, especially polycyclic aromatic compounds. Generally, increased polarity reduces adsorption, but this effect is smaller, and easily overcome by size. Polyphenols, for instance, are strongly adsorbed de- spite being freely soluble in water. Many other variables affect adsorption to lesser extents.173 While these general trends are well-established, the underly- ing mechanisms remain in dispute.

2.1.3.1 Mechanism of Action.

The surface chemistry of different types of activated carbon varies, from highly oxidised forms with numerous polar functional groups distributed across the surface, to graphitised forms with very little functionality.173 The most im- portant common characteristic of all forms is their enormous surface area, 2.1. ISOLATION PROCEDURE. 73 estimated at up to 1500 m2/g in some cases,174 resulting from the material’s extremely porous structure. Recent work suggests that this is the key factor controlling adsorption. In “solvophobic” theories,175, 176 adsorption is treated as being driven by two key factors: the (unfavourable) formation of a sol- vent cavity around the solute, and (favourable) van der Waals attraction be- tween the solute and the carbon surface. As a result, “the theory predicts the adsorbability to depend linearly on the nonpolar surface area of the adsor- bate”.176 This simple model predicts very accurately the adsorption data for a wide range of compounds.175, 176 Another recent study, based on a much larger data set, used a more complex model (the “linear solvation energy relationship” model).177 Interestingly, however, despite being based on different parameters, the study reached very similar conclusions. Planar molecules were found to be more strongly adsorbed than nonplanar molecules. This was rationalised as follows:

Since the dispersion force of attraction is very sensitive to the dis- tance of separation between the surface of the activated carbon and the center of the solute molecule, it is reasonable that a pla- nar molecule would be adsorbed more strongly than an otherwise similar globular molecule.177

That is, consistent with the solvophobic model, surface area available for con- tact is the key factor in determining adsorbability. Interestingly, this effect was found not only for aromatic compounds; planar olefins were just as strongly adsorbed. A similar effect had been observed in earlier work,178 based on an- other model, and used to predict that square planar organometallic complexes would adsorb more strongly than octahedral complexes. This prediction was confirmed experimentally.178 This also helps explain the strong adsorbability of carotenoids such as 22, which although aliphatic can assume planar confor- mations which present a much larger surface area to the carbon surface than less conjugated hydrocarbons. Indeed, in this work non-conjugated terpenoids 74 CHAPTER 2. ISOLATION. comparable in size to 22 were easily desorbed from carbon (Section 2.2.4.2 on page 97), while 22 and other carotenoids were not.

These works establish a strong theoretical and empirical relationship between surface area and adsorbability. Nonetheless, other factors affect adsorption, and the underlying mechanisms remain controversial, even for particular classes of compound.179

2.1.3.2 Use during Recrystallisation.

The most common application of activated carbon by chemists is in remov- ing coloured impurities from crude compounds. Typically, a small amount of powdered activated carbon (1–2% w/w relative to the crude compound) is added during recrystallisation; the decolourised solution is then filtered before cooling.172 In many cases, however, this is not effective. Indeed, during the preliminary work described above, flash column chromatography of the crude extract gave an orange solid which was rich in 1a by 1HNMR.Duringre- crystallisation from MeOH, the bright yellow colour of the solution was not diminished by activated carbon, even at 10% w/w. Upon cooling, the result- ing crystalline 1a remained faintly yellow, and the mother liquor vividly so. While this might be dealt with by using larger amounts of activated carbon, this would also result in increased adsorption of 1a itself, and therefore reduced recovery.

Even if the procedure had been effective, moreover, it would still only be useful for removal of trace pigments from individual compounds. As noted above, it was essential to this work that the crude extract itself be decolourised. This would therefore require the use of a large amount of activated carbon. The compounds of interest would then have to be desorbed, in what would be “effectively a chromatographic procedure.”172 2.1. ISOLATION PROCEDURE. 75

2.1.3.3 Use in Chromatography.

Activated carbon is rarely employed as a stationary phase in chromatography. It is not suited to general use; for most separations, other adsorbents offer superior selectivity and resolution.180 Nonetheless, it has found use in certain niche applications. Perhaps the most common use has been the isolation of antibiotics181 and enzymes182 from microbial fermentation broths. Activated carbon remains a standard part of such procedures, although alternatives have been proposed.183, 184

Another application where activated carbon has proven superior to conven- tional adsorbents is in the separation of fullerenes,185, 186 giving better reso- lution and recovery than alumina, along with higher speed and lower cost.

186 The larger ellipsoidal fullerenes, above C70, are very difficult to desorb. Activated carbon has also proven valuable in separation of polychlorinated biphenyls (PCBs).187, 188, 189 Commercial PCB mixtures are very complex, con- taining hundreds of PCBs differing in the number and position of chlorine sub- stituents.189 Many of these compounds differ very little in polarity and volatil- ity, resulting in incomplete resolution by either HPLC or GC. This problem can be resolved by multi-dimensional GC, but activated carbon permits sepa- ration on a different basis: extent of ortho-substitution.189 Non-o-substituted PCBs (such as 24, Figure 2.3) can assume a coplanar conformation, and are therefore very strongly adsorbed (as expected). Each o-substituent presents a steric barrier to this conformation, and thus reduces adsorption, reaching a minimum with fully o-substituted PCBs (such as 25). PCBs can therefore be eluted from a carbon column in decreasing degree of o-substitution.190, 188 These fractions are greatly simplified relative to the initial mixture, and can be further purified using conventional techniques.187, 188 This method also has the attraction that degree of o-substitution is of great biological significance. In addition to their toxicity, non-o-substituted PCBs such as 24 are extremely teratogenic; this effect is reduced by one o-substituent, and abolished by more than one.187 76 CHAPTER 2. ISOLATION.

Cl Cl Cl Cl Cl Cl Cl Cl Cl 24 Cl 25

Figure 2.3: Ortho- and non-ortho-substituted PCBs.

A surprisingly uncommon use of chromatography on activated carbon is for decolourisation of crude extracts during isolation191, 192 or analysis193 of natural products. In most cases this is used, as here, to permit analysis of the crude mixture rather than purification of individual compounds.

As with any adsorbent, eluent selection is critical to achieving effective sepa- ration. The eluotropic series on activated carbon given by Gordon180 is:

i H2O

In general, then, the larger and less polar the solvent, the greater its elut- ing strength. That benzene is the strongest eluent in the series is unsurpris- ing given the strong adsorption of aromatic compounds. In recent decades, toluene has been used as a less toxic and equally effective substitute for ben- zene.189, 185 Indeed, toluene appears on the balance of the evidence to adsorb more strongly than benzene,177 which is unsurprising given its larger surface area. Consistent with this, recent work has shown that o-dichlorobenzene is a much stronger eluent again.186 Based on their measured adsorption con- stants,177 p-xylene would probably offer comparable eluting strength, with the advantage of a lower boiling point (138 vs 178 ◦C). Another approach to elu- tion of very strongly adsorbed compounds is exhaustive extraction with hot solvent, using Soxhlet apparatus or similar.187

A major obstacle to practical chromatographic use of activated carbon powder is that it packs very tightly; it is almost impermeable to solvent even under pressure.180 One solution to this is to “fluidise” the activated carbon, using a stream of nitrogen to remove the finest particles.187 Similarly, industrial-scale water treatment usually employs granulated activated carbon.174 Using larger 2.1. ISOLATION PROCEDURE. 77 particles in this way, however, has the undesired effect of reducing available surface area. A more common solution is therefore to use mixed adsorbents.180 The most common choice is diatomite filter aid (eg. Celite),187 but other possibilities include cellulose powder193 and silica gel.189, 185

2.1.3.4 Application to S. divinorum Extract.

For this work, the mixed adsorbent approach was taken. Diatomite filter aid was selected due to its low cost, safety and ready availability. The proce- dure was performed as standard flash column chromatography, monitored by TLC.194 Initial testing on a small scale indicated that MeOH did not elute 1a, while acetone did so rapidly. Additionally, 1a is far more soluble in acetone. Some tailing was evident, so a stepwise gradient was employed based on Gor-

180 don’s eluotropic series, through Et2O, EtOAc and EtOAc/petrol. Petrol was not used alone due to the insolubility of 1a. Gradient elution helped to minimise tailing, but did not eliminate it – 1a was detectable from near the solvent front into the EtOAc/hexane fractions.

No pigments were eluted when sufficient carbon was used; a carbon:solute mass ratio of 20:1 was found adequate. At a ratio of 3:1, the adsorbent was overloaded, and orange pigments were eluted even by acetone. After the sub- sequent isolation of pure 1a (detailed in Section 2.1.4 on page 82), the recovery of this compound from carbon was tested. A microscale test column on 15 mg of 1a at 100:1 carbon ratio gave approximately 80% recovery. The loss was more likely due to the small scale than irreversible adsorption, since stripping the column in 50% toluene/petrol gave no additional 1a. Nonetheless, use of minimal carbon is clearly advisable to minimise any losses. Interestingly, the elution with toluene/petrol gave an orange fraction, presumably carotenoids. Chlorophylls, which are more strongly adsorbed than carotenoids,195 were not eluted.

A number of practical points should be noted:

• Activated carbon powder can contain much finer particles than flash 78 CHAPTER 2. ISOLATION.

chromatography grade silica; these can contaminate fractions and block glass frits. A layer of filter aid below the carbon prevents this.

• As noted by others,185 the carbon/filter aid mixture shrinks and cracks when dry, which may contaminate the decolourised fractions with crude extract. To prevent this, it is critical that the solvent level never be allowed to touch the surface of the carbon. Since it is difficult to locate the interface between the very dark crude extract and the black carbon, it is helpful to add a thick layer of sand on top. This layer should never be allowed to run dry.

• In later work196 filtration under vacuum proved more convenient and practical than flash chromatography. This preferred approach is shown in Figure 2.4. Fractions are collected by closing the separatory funnel tap and applying vacuum. When the desired amount has been collected, vacuum is released and the fraction dispensed.

• A gradient from 50% to 20% EtOAc/petrol gave faster elution and less tailing, without eluting pigments. All fractions containing terpenoids by TLC were pooled.

Sand Activated carbon/filter aid Filter aid Glass frit

Vacuum adaptor

Separatory funnel

Figure 2.4: Apparatus for Filtration through Activated Carbon.

While some separation between compounds in the decolourised fractions was 2.1. ISOLATION PROCEDURE. 79 apparent (see Section 2.1.4.2 on page 85), separation of individual compounds proved unworkable due to tailing. Thus, since the pigments were not eluted and the eluted compounds were not separated, in practice the preferred procedure resembled filtration more than chromatography.

2.1.3.5 General Applicability of the Method.

This procedure has some attractive features for use in isolating biologically active compounds from plants. Activated carbon filtration provides a much greater reduction in mass and complexity of the extract than more common partitioning procedures. Thus, as seen with PCBs, analysis is simplified by utilising an extra “dimension” for separation (surface area/planarity), com- pared to the typical reliance on polarity differences. Using the procedure de- scribed here, as discussed below, diterpenoids and triterpenoids were rapidly eluted, while the tetraterpene carotenoids were completely removed.

Moreover, this added dimension (size) is of biological relevance. It is well established pharmacologically that large molecules show reduced absorption and permeation. This is embodied in Lipinski’s famous “rule of 5”,197 which states among other things that compounds with a relative molecular mass over 500 will be less readily absorbed in vivo, and therefore less likely to show pharmacological activity. Thus, compounds eluted first from activated carbon are more likely to be responsible for biological activity.

Nonetheless, as indicated previously, size is not the sole determinant of ad- sorption. Compounds of interest may not be recovered using the procedure described here if they are highly planar, or strongly adsorbed for some other reason. Of particular relevance here are flavonoids, some of which are known to be psychoactive. Apigenin (26a), isolated from S. officinalis, binds weakly

198 to the benzodiazepine site of the GABAA receptor (IC50 =30µM). Other flavones isolated from plants have shown much greater potency at this site;

199 27 binds with affinity comparable to diazepam (K i =6nM). Interestingly, some flavonoids appear to share the anxiolytic properties of benzodiazepines, 80 CHAPTER 2. ISOLATION. without the amnesic200 and sedating201 qualities.

There has been no published report of flavonoids from S. divinorum. However, in unpublished work, Valdés isolated the known apigenin derivative 5-hydroxy- 7,4’-dimethoxyflavone (26b)202 from S. divinorum in 20 mg/kg yield.203 This compound has been isolated from several Salvia species; identity was con- firmed by comparison of 1Hand13CNMR,IR,MS(CI)andmpwithliterature values.202 In other recent unpublished work, Claudio Medana detected a com- pound whose ESIMS data were consistent with quercitrin (23) or a stereoiso- mer in a crude extract of S. divinorum.204 The aglycone of 23, quercetin, activates several CNS receptors.205 Quercetin, but not 23 itself, has been proposed to have antidiarrhoeal effects;206 if so, this might conceivably con- tribute to the known effects of S. divinorum on gastric motility.207 However, quercetin’s gastrointestinal effects remain in dispute.208

OR HO 2' OH O HO O HO O RO O OH

O O R OH O OH O OH OH O O 26a H 26b Me 27 23 OH OH

Figure 2.5: Flavonoids.

Flavonoids are not recovered using the isolation procedure described above. However, they are unlikely to play a detectable role in the plant’s effects. The binding affinity of 26a is more than three orders of magnitude below that of 1a.23 Moreover, Valdés isolated 1a in two orders of magnitude greater yield than 26b.

In other plants, however, where flavonoids or other strongly adsorbed com- pounds are responsible for activity, the above procedure would clearly be in- adequate. Nonetheless, this obstacle is readily overcome. If bioassay-guided fractionation showed loss of activity in the decolourised extract, further elu- tion with stronger eluents as outlined in Section 2.1.3.3 could be used. The 2.1. ISOLATION PROCEDURE. 81 active fraction would then be greatly simplified relative to the initial extract, providing another dimension of separation, as with PCBs.

2.1.3.6 Other Decolourisation Adsorbents.

Although extremely cheap for laboratory use, use of activated carbon can be expensive on an industrial scale. For this reason, cheaper alternative adsor- bents have been investigated. These are generally waste products: agricultural (leaves, straw, husks, pulp) or industrial (ash, slag).174 One adsorbent which has been used extensively is acid-activated clay.174 Activated clays (or “ac- tivated earths”) are widely used, for instance, to decolourise cooking oils.209 Although not as effective by weight as activated carbon,195 the lower cost makes industrial use attractive.

The adsorbent “Tonsil” used by Ortega et al in the original isolation22 of 1a is a type of activated clay used for decolourisation.209, 210 Unfortunately, this was not explained; the material was described (in a footnote) merely as “bentonitic earth” composed of silica (72.5%), alumina and other minerals.22 The body of the paper and the experimental procedure make no mention of pigments or colouration. Thus, readers may assume that Tonsil is simply a silica substitute. It is perhaps for this reason that no subsequent worker has tested this adsorbent. Nonetheless, Dr Ortega has confirmed that Tonsil was selected in order to decolourise the crude extract.211 He reports that the procedure was highly effective: the collected fractions were almost colourless. Tonsil was used for other isolations,212 but proved reactive towards certain functional groups, especially epoxides,213 and therefore unsuited to general use. Indeed, this substance has since found wide synthetic use as an acid catalyst.214 Activated carbon is milder and more generally applicable. 82 CHAPTER 2. ISOLATION.

0 100 µm

Figure 2.6: SEM image of salvinorin A crystals (blade morphology).39

2.1.4 Separation of Terpenoids.

2.1.4.1 Crystallisation of 1a.

TLC indicated that the major component of the decolourised extract (Sec- tion 2.1.3.4 on page 77) was 1a, which was obtained as fine colourless needles by crystallisation from MeOH. 1H NMR showed other trace components  1% each. Another crystallisation of the mother liquor gave additional 1a.There- sulting mother liquor was no longer predominantly 1a by TLC. In other tests, EtOH was also an effective recrystallisation solvent (Lee et al subsequently used acetone).137

Serendipitously, trituration in Et2O also proved effective for isolation of 1a from the decolourised extract. In early tests, a chromatographic fraction rich in

1a was dissolved in Et2O for transfer; an amorphous, flaky precipitate formed instantly. The filtrate proved, on further chromatography, to be devoid of 1a. The precipitate consisted of 1a which, despite its inferior appearance, was 1 highly pure by H NMR. Thus, trituration in cold Et2O represents a promising alternative to recrystallisation from boiling alcohols – easier, safer and (more significantly) much faster. Moreover, the risk of decomposition in heated solu- tions (Section 2.1.1 on page 67) is avoided. This procedure, which was unfortu- 2.1. ISOLATION PROCEDURE. 83 nately given insufficient emphasis in our initial publication,215 clearly deserves further exploration.

The fine needles of 1a were examined by scanning electron microscopy (SEM) (Figure 2.6). Shorter prisms were also observed, as well as plates or flakes (Fig- ure 2.7), and close examination of crystals showed small defective, semicrys- talline particles adhering. 84 CHAPTER 2. ISOLATION.

0 200 0 200 µm µm

0 20 µm 0 20 µm

0 20 0 20 µm µm

Figure 2.7: Other salvinorin A crystal morphologies (stereoview).39 2.1. ISOLATION PROCEDURE. 85

2.1.4.2 Chromatography.

O O O

1 1 O O OR O R H H H H H H RO R2 O O R2

O O O O O OR3 R R1 R2 R1 R2 R3

1a Ac 1c Ac OAc 28a OH H H 1b H 1d Ac OH 28b OH OH Me 1e H OAc 28c H OAc H 1f H H 29a H H H

OH H OH 30 O H HO H 31 H

32 OH

OH 33

Figure 2.8: Terpenoids isolated from S. divinorum.

The mother liquor from crystallisation of 1a was an unexpectedly complex mixture. Although at the time only four compounds had been reported from S. divinorum, TLC showed many more. The spots were poorly resolved, mak- ing a precise count impossible. Valdés et al had reported27 that 1a and 1c were resolved in EtOAc/hexanes, but not in MeOH/CHCl3; this information proved invaluable. 2D TLC using these systems gave much greater resolution than either system alone. Many compounds not resolved by one system were 86 CHAPTER 2. ISOLATION. ) 2

Cl 80 2

33 70 31

(10% acetone/CH 1c f 60 30 hR 1a

50 1e 29a 1f 40 28c 1b 32 28b 30 1d

20 28a

10 10 20 30 40 50 60 70 80

hRf (70% Et2O/petrol)

Figure 2.9: TLC data of isolated compounds. resolved by the other. These differences in selectivity proved to be general for petrol-based versus chlorinated solvent-based systems. After experimen- tation with various petrol-based systems, Et2O gave similar resolution of 1a and 1c to EtOAc, but better separation from other compounds. Similarly, in CH2Cl2-based systems, acetone resolved 1a and 1c from other compounds better than MeOH. Also, these new systems spread the mixture over a wider range, from near the baseline to near the solvent-front. These systems were therefore adopted for general use.

Extensive column chromatography yielded six new diterpenoids in addition to 1a–1c: salvinorins D–F (1d–1f) and divinatorins A–C (28a–28c, Figure 2.8 on the previous page). Five known terpenoids were also isolated: (–)-hardwickiic acid (29a), (E)-phytol (30), oleanolic acid (31), presqualene alcohol (32) and peplusol (33). 2.1. ISOLATION PROCEDURE. 87

Dried leaves (860 g)

Powder, steep in acetone (× 3)

Crude extract (30.5 g)

Column on activated carbon acetone to petrol gradient

Decolourised extract (5.7 g)

Recrystallise (MeOH, EtOH) 1a (2.6 g)

Combine Column on silica

5-50% acetone/CH2Cl2

Series A Series B Series C Series D 656 mg 150 mg 359 mg 77 mg

Column Columns 70-90% Et2O/petrol Column 1b (13 mg) 28a (36 mg) Triturate (Et2O) 50-80% Et2O/petrol

1d (75 mg) 1c (219 mg) 29a (7 mg) Combine Column 1a (2.9 g) 60-100% Et2O/petrol

Fraction C1 Fraction C2 Fraction C3 55 mg 119 mg 57 mg 1d (114 mg)

Column Columns

60-100% Et2O/petrol Columns 28b (41 mg) Combine HPLC 28c (23 mg) 31 (3 mg) 60% EtOAc/petrol

1e (3 mg) 1f (1 mg)

Figure 2.10: Isolation of terpenoids from commercial S. divinorum.

Figure 2.9 on the facing page shows the TLC data for all compounds isolated in this work. The most dramatic differences between the two systems are 88 CHAPTER 2. ISOLATION.

Dried leaves (224 g)

Powder, steep in acetone (× 3)

Crude extract (7 g)

Vacuum filtration on activated carbon 50-20% EtOAc/petrol

Series E Series F Recrystallise (MeOH × 2) 97 mg 279 mg

Columns on silica

33 (1 mg) 32 (23 mg) 30 (12 mg) 1a (126 mg)

Figure 2.11: Isolation of terpenoids from Australian S. divinorum. emphasised by dashed lines. For instance, separation of 1c from 28b or of 1a from 1f is not possible in Et2O/petrol, but trivial in acetone/CH2Cl2.The converse is true for separating 1f from 28c or 1a from 30.

With these solvent systems in place, separation of the components of the mother liquor was straightforward, albeit time-consuming. Repeated column chromatography, alternating between Et2O/petrol and acetone/CH2Cl2,re- solved most compounds. In deciding which fractions to pool, TLC analysis in each of the main solvent systems proved invaluable. For particular separa- tions, other solvent mixtures were sometimes used when TLC indicated better resolution. The overall procedure, as applied to the extraction of commercial dried S. divinorum leaves, is summarised in Figure 2.10 on the previous page. For more detail, see Experimental Section 5.2.1 on page 185. Yields expressed in mg/kg are given in Table 5.1 on page 187.

A similar procedure was used on a sample of Australian-grown S. divinorum leaves (Figure 2.11), resulting in the isolation of several additional compounds. Partial separation of these compounds was achieved in the initial filtration through activated carbon, but the severe tailing exhibited by all compounds 2.1. ISOLATION PROCEDURE. 89 prevented full separation. Since chromatography on silica was ultimately re- quired, attempts to separate compounds on carbon are not recommended.

2.1.4.3 Crystallisation of new compounds.

As with 1a, trituration in Et2O proved effective for purification of 1d.At- tempted dissolution of a fraction containing ≈ 33% 1d in boiling Et2Ogave white crystals of the pure compound. Also, as reported by Tidgewell et al,169 trituration in cold MeOH is effective for purification of 1b.

The ease of crystallising these compounds varies widely. Evaporation of a so- lution of 1b gave crystals irrespective of the solvent. 1a and 1d gave crystals from some solvents, but an amorphous resin from others (especially CH2Cl2

27 and CHCl3). Although Valdés reported the crystallisation of 1c, attempts to replicate this were unsuccessful. Attempts to recrystallise the other new compounds, under a variety of conditions, were also unsuccessful. On one oc- casion, 28b gave crystals from cold CH2Cl2 on addition of petrol, but repeated attempts to replicate this were unsuccessful.

2.1.4.4 Losses during isolation.

The recoveries of these compounds do not accurately reflect their actual pro- portions in the crude extract. The main objective in this work was to obtain the maximum number of pure compounds; yield was a much lower priority except in the case of 1a. Traces of one compound were often discarded during the isolation of another. In the most extreme case, all of the 29a isolated was derived from a small portion (13%) of the crude extract. The actual content of this compound in the crude extract, therefore, was much higher than the reported yield.

There were also losses due to other causes. In one case (separation of 1e and 1f), despite good separation by TLC (Figure 2.9 on page 86), repeated column chromatography gave incomplete resolution. Ultimately, HPLC gave 90 CHAPTER 2. ISOLATION. full resolution, but unfortunately the recovery after this lengthy process was very poor. From a 34 mg fraction containing only 1e and 1f by TLC and 1H NMR, the final combined recovery of these two compounds was only 4 mg.

Similarly, yield of 1b (15 mg/kg) was very low compared to earlier work (75 mg/kg).23 This was probably largely due to losses in chromatography. During subsequent synthetic work, it became apparent that column chromatography of 1b in petrol-based systems gave very poor recoveries. Stripping the column in acetone/CH2Cl2 gave an improvement, but full recovery was only achieved using MeOH/CH2Cl2. Additionally, recrystallisation of 1b-rich fractions from MeOH gave poor recoveries, and TLC showed extensive decomposition. As noted in Section 2.1.1 on page 67, this decomposition also occurred with 1a.

2.2 Structure Elucidation.

2.2.1 Revised NMR Assignments for Salvinorin A (1a).

The published NMR assignments23 of 1a (based on decoupling experiments) were verified on the basis of 800 MHz 1H NMR (Figure 2.12 on the next page), along with HSQC, HMBC, DEPT and nOe experiments. These new data supported the original 1Hand13C assignments in all cases except H-6 and -7. The HSQC spectrum showed cross peaks from C-6 (δ 38.1 ppm) to multiplets at δ 1.57 & 1.78 ppm, and from C-7 (δ 18.1 ppm) to δ 1.63 & 2.15 ppm (Figure 2.13 on the facing page). The 800 MHz 1H NMR spectrum also gave well-resolved first-order multiplets for H-6β,7α and 11β, which overlap to form a single multiplet (δ 1.54 – 1.68 ppm) at 400 MHz. Resolution enhancement revealed all coupling constants in these multiplets, allowing unambiguous assignment of all peaks (Figure 2.14 on page 92). 2.2. STRUCTURE ELUCIDATION. 91 7.402 7.384 6.369 5.521 5.138 3.724 2.745 2.504 2.304 2.178 2.163 2.074 1.792 1.642 1.574 1.447 1.113

O

O O O H H O O

OO 1a 2.35 2.33 2.31 2.29

1.661.64 1.62 1.60 1.58 1.56 ppm 2.762.74 ppm

6.37 6.36

2.18 2.16 2.14 1.801.78 ppm

7.40 7.39 ppm

5.53 5.52 5.51 5.165.15 5.14 5.13 ppm 2.52 2.51 2.50 2.49 2.06

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 ppm 1.63 1.59 0.82 0.85 1.00 3.25 1.02 0.88 2.27 5.30 1.54 1.51 1.28 1.94 2.96 2.70

1 Figure 2.12: H NMR spectrum of 1a (800 MHz, CDCl3).

C-7 20 80

90

100

110 30

120

130

140 C-6 40 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2

50

60

3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 ppm

Figure 2.13: HSQC spectrum of 1a (800 MHz, CDCl3). 92 CHAPTER 2. ISOLATION.

NOESY NOESY

7.38 t 7.38 t (REVISED ASSIGNMENT) (REVISED ASSIGNMENT) δH δH δH δH 143.7 143.7 δC δC 108.3 139.4 108.3 139.4 6.37 dd 7.41 dt 6.37 dd 7.41 dt

2.16 s 125.2 2.16 s 125.2 1.57 ddd 1.57 ddd 72.0 72.0 20.5 2.50 dd 20.5 2.50 dd 43.3 43.3 5.52 ddd 5.52 ddd 169.9 169.9 5.14 ddt 2.18 br s 2.07 dd 5.14 ddt 2.18 br s 2.07 dd 171.1 171.1 35.4 35.4 202.0 202.0 75.0 64.0 51.3 75.0 64.0 51.3 15.2 15.2 2.74 dd 1.45 s 2.74 dd 1.45 s 1.58 td 1.58 td 30.9 42.0 18.1 30.9 42.0 18.1 2.17-2.14 m 2.17-2.14 m 53.5 38.1 53.5 38.1 2.31-2.28 m 1.64 tdd 2.31-2.28 m 1.64 tdd 16.4 16.4 171.5 1.79 dt 171.5 1.79 dt

1.11 s 1.11 s

51.9 51.9

3.72 s 3.72 s

Figure 2.14: Revised NMR assignments for 1a (stereoview).

2.2.2 Revised NMR Assignments for Salvinorin C (1c).

The published structure and NMR assignments27 of 1c were verified on the basis of 1H NMR (Figure 2.15 on the facing page), decoupling, HMQC, HMBC and DEPT experiments. These data were consistent with the published struc- ture, but the assignments of H-19, H-20, C-17 and the acetates were revised as shown in Figure 2.16 on the next page. Contrary to the original assignments, the HMQC spectrum showed strong cross peaks between C-19 (21.8 ppm) and H-19 (1.71 s), as well as C-20 (15.7 ppm) and H-20 (1.21 s) - see Figure 2.17a. This was confirmed by HMBC cross peaks from H-19 to C-4 (142.6 ppm), and H-20 to C-11 (44.2 ppm) - Figure 2.17b.

HMBC data also allowed the assignment of the acetate signals. The C-2- O-acetate carbonyl (169.8 ppm), originally assigned to C-17,27 showed cross peaks to H-2 (5.54 ppm) and 2.04 ppm; its C-1-O- counterpart (170.5 ppm) showed cross peaks to H-1 (5.75 ppm) and 2.12 ppm - Figure 2.17c,d.

2.2.3 Other Known Diterpenoids.

In addition to 1a-1c, five known compounds not previously reported from S. divinorum were isolated. The data and sources used to identify them are given 2.2. STRUCTURE ELUCIDATION. 93 7 6 6 0 2 1 2 3 1 4 2 0 7 7 1 . . . . . 1 2 2 1 3

21. 2 2 . 1.8 1.7 7.4 O 1.3 11.2 1. 0 6 2

. O 7 O O 11 O H H O 17 1 2 O 20

4 19 1c OO

6.5 6.4 5.8 5.7 5.6 5.5 2.6 2.4

7 6 5 4 3 2 ppm

0.67 0.78 1.97 1.03 5.06 1.14 1.17 4.54 0.70 0.88 1.00 3.00 1.00 2.95 4.43

1 Figure 2.15: H NMR spectrum of 1c (400 MHz, CDCl3).

7.42 t 7.42 t HMBC HMBC 143.9 143.9 δ δ δ δ C H C H 108.4 139.4 108.4 139.4 6.41 dd 6.41 dd δ (REVISED ASSIGNMENT) δ (REVISED ASSIGNMENT) C/H C/H 7.45 m 7.45 m 125.4 125.4

2.04 s 1.68 dd 2.04 s 1.68 dd 71.8 71.8 2.48 dd 2.48 dd 5.54 dd 44.2 5.54 dd 44.2 20.7 5.75 br d 5.53 dd 20.7 5.75 br d 5.53 dd 169.8 169.8 1.49 br s 2.13 dd 171.4 1.49 br s 2.13 dd 171.4 64.1 64.1 69.2 37.2 69.2 37.2 52.6 52.6 51.8 51.8 132.4 1.21 s 132.4 1.21 s 6.46 dd 1.23 td15.7 6.46 dd 1.23 td 15.7 18.4 18.4 38.1 38.1 170.5 2.17-2.10 m 170.5 2.17-2.10 m 21.1 142.6 37.0 21.1 142.6 37.0 2.12 s 165.7 1.85-1.75 m 2.12 s 165.7 1.85-1.75 m 2.59 dt 2.59 dt 21.8 21.8 1.71 s 1.71 s 51.8 51.8

3.74 s 3.74 s

Figure 2.16: Revised NMR assignments for 1c (stereoview).

in Table 2.2 on the following page. 94 CHAPTER 2. ISOLATION.

δH (ppm) 1.15 1.2 a) HMQC 1.20 b) HMBC 1.3 1.25 1.4 1.30 1.5 1.35 1.6 1.40 1.7 1.45 1.50 1.8 1.55 1.9 1.60 2.0 1.65 2.1 1.70 2.2 1.75 23 22 21 20 19 18 17 16 15 140 130 120 110 100 90 80 70 60 50 40

d) HMBC 1.98 c) HMBC 5.50 2.00 2.02 5.55 2.04 2.06 5.60 2.08 5.65 2.10 2.12 5.70 2.14 2.16 5.75 2.18

2.20 5.80 171.5 170.5 169.5 168.5 172.5 172.0 171.5 171.0 170.5 170.0 169.5 169.0 168.5 δC(ppm)

Figure 2.17: HMQC and HMBC spectra of 1c (400 MHz, CDCl3).

1 13 Compound HNMR CNMR FTIR [α]D HRESIMS EIMS TLC 29a sa, d216, 217 cs 29b* sa, d218 d218 d218 d218, 217 d217 cs 30 sp,219 d220 d219, 220 d219, 220 31 d221 d221 c 32 d,222 sp223 d223 d222 d223 33 d224 d224 d224 d224

* 29a was methylated with CH2N2 for full characterisation. c consistent with calculated value. cs cospotted with authentic material in petrol- and CH2Cl2-based systems. d consistent with published data. sa superimposable with spectrum of authentic material. sp superimposable with published spectrum.

Table 2.2: Data and sources used to identify known compounds.

2.2.3.1 (–)-Hardwickiic acid (29a).

(–)-Hardwickiic acid (29a) was first isolated from Hardwickia pinnata by Misra et al,225 and the structure (including absolute stereochemistry) elucidated by 2.2. STRUCTURE ELUCIDATION. 95

O

H H

R

29a H O OR 29b Me

Figure 2.18: Single-crystal X-ray structure of 29a (stereoview).

spectroscopic methods and extensive degradation studies.226, 227 Recently, X- ray crystallography has confirmed this structure (Figure 2.18).228 29a has since been isolated from many other plant species, including Salvia regla.229 29a has been reported to exhibit insecticidal230 and antimicrobial217 activity. Interest- ingly, its enantiomer (+)-hardwickiic acid (ent-29a) has been isolated from Copaifera officinalis231 and related species.232 For definitive identification, au- thentic ent-29a was isolated as the methyl ester ent-29b from commercial Copaiba balsam.218 The material from S. divinorum, after methylation to give 29b, cospotted by TLC and gave a superimposable 1H NMR spectrum, but the opposite sign of optical rotation. 96 CHAPTER 2. ISOLATION.

30 OH

Figure 2.19: (E)-Phytol.

2.2.3.2 (E)-Phytol (30).

(E)-Phytol (30) is ubiquitous in photosynthetic organisms, as the sidechain of the chlorophylls (eg. 21 on page 71). 30 displays anticancer activity,233 and may play a role in the health benefits of green-yellow vegetables.234 30 also exhibits antiplasmodial,235 antimycobacterial236, 237 and antiteratogenic238 activities.

2.2.4 Known Triterpenoids.

2.2.4.1 Oleanolic Acid (31).

H OH

O H HO H 31

Figure 2.20: Oleanolic acid.

Oleanolic acid (31) occurs commonly in plants, including many Salvia species.239 It exhibits an extremely wide range of potential therapeutic activities, which have been reviewed elsewhere:240, 241, 239 anti-inflammatory, hepatoprotective, gastroprotective, cardiovascular, hypolipidemic and immunoregulatory effects. Of more significance is anti-tumour activity, which appears to occur via several independent mechanisms (inhibition of tumourigenesis, inhibition of tumour promotion, and induction of tumour cell differentiation).240 Similarly, 31 is active against HIV-1 by two mechanisms: inhibition of HIV protease242 and reverse transcriptase243 enzymes. The compound also exhibits activity against 2.2. STRUCTURE ELUCIDATION. 97 many other pathogens: herpes simplex virus,244 a range of bacteria245, 246 (in- cluding mycobacteria)247 and Leishmania spp,248 as well as being an insect antifeedant.249

2.2.4.2 Presqualene Alcohol (32) and Peplusol (33).

H

32 OH

OH 33

Figure 2.21: Presqualene alcohol and peplusol.

The diphosphate of presqualene alcohol (32) is a precursor in the biosynthesis of sterols (including cholesterol), and has therefore been the subject of intense synthetic223 and biosynthetic250 interest. The isolation of 32 from a plant is unusual.

The closely related compound peplusol (33) was isolated from Euphorbia pe- plus by Giner et al.224 In a paper submitted a year later, Connolly et al reported the same compound from E. lateriflora, naming it anhydrobisfar- nesol.251 The earlier paper by Giner et al was mentioned, but only in a foot- note; priority was claimed252 based on a preliminary communication.253 That communication, however, contained no experimental or characterisation data, and the absolute configuration was not determined even in the full paper. Additionally, the ambiguous semisystematic name anhydrobisfarnesol leaves regio- and stereochemistry unspecified. Thus, Giner et al clearly have priority, and their name is used here. Connolly and co-workers confirmed the proposed structure by synthesis, obtaining the racemic compound in 6% yield over six steps.251 However, they failed to cite a clearly superior synthesis (50% over two steps)254 published fifteen years earlier. Connolly has proposed 33 as a biosynthetic precursor to 32 itself.252 98 CHAPTER 2. ISOLATION.

It is interesting to note that 33 and 32 are approximately equal in size (22- carbon chain) to the carotenoid 22 (p.71, 24 carbons), but were easily eluted from activated carbon in the above chromatography procedure. Since, un- like 22, these compounds are not conjugated, this again confirms the strong influence of planarity on adsorption.

2.2.5 Salvinorins D-F (1d-1f).

2.2.5.1 Salvinorins D and E (1d and 1e).

Inspection of the 1H NMR spectra of 1d-1f suggested that they were deriva- tives of 1c.Comparedto1c,the1H NMR spectra of 1d (Figure 2.22) and 1e (Figure 2.26 on page 102) showed only one acetyl peak, and gained one new peak (δ 2.01 in 1d; δ 1.94 in 1e) which exchanged with D2O. The presence of a hydroxyl group was confirmed by IR spectroscopy (3475 cm−1 in 1d; 3510 cm−1 in 1e). HRESIMS established the molecular formula of both compounds as C23H28O8, consistent with the loss of one acetyl group. These data indicated that the compounds were monoacetates of 1c.

Relative to 1c, the H-2 signal of 1d was shifted upfield from δ 5.55 to δ 4.44, and coupled to the hydroxylic proton (2.01 ppm) as well as to H-1 (5.70 ppm) and H-3 (6.54 ppm). The HMBC spectrum (Figure 2.25 on page 101) showed correlations between H-2 and C-4 (141.2 ppm), as well as H-1 and C-5 (37.6 ppm). These quaternary carbons, not apparent in the HMQC (Figure 2.24 on page 100) and DEPT spectra, were in turn located by HMBC correlations. This established the structure of 1d as 2-deacetylsalvinorin C.

Relative to 1c, the H-1 signal of 1e was shifted from δ 5.76 to δ 4.46. Irra- diation sharpened the H-10 singlet. The HMBC spectrum (Figure 2.29) again showed a correlation between H-2 (5.40 ppm) and C-4 (143.4 ppm), not ap- parent in the HMQC spectrum (Figure 2.28). Thus 1e is 1-deacetylsalvinorin C, as shown. The quaternary carbon at 169.8 ppm was incorrectly assigned to C-17 in our original publication;215 this has been reassigned on the basis of its 2.2. STRUCTURE ELUCIDATION. 99

O

O

O O H H HO 1 2 O 5 3 4

O O 1d 0 5 8 6 2 3 5 8 7 1 6 2 . . . . 1 3 2 1

7.5 7.4 2.2 2.2 2.0 1.8 1.6 0 6 2 . 7 5.7 5.6 5.5

1.31.2 1.1

6.6 6.5 6.4

4.5 4.4 2.6 2.5

7 6 5 4 3 2 ppm

0.83 0.87 0.99 3.00 4.98 1.05 4.43 0.99 0.85 0.92 0.98 1.00 2.00 0.97 4.07

1 Figure 2.22: H NMR spectrum of 1d (400 MHz, CDCl3).

HMBC crosspeak with the acetate protons (see Figure 2.27 on page 102). 100 CHAPTER 2. ISOLATION.

7.42 t 7.42 t

J (Hz) 143.8 J (Hz) 143.8 δ δ δ δ H H H H 108.4 139.4 108.4 139.4 HMBC 6.40 dd HMBC 6.40 dd δ 7.44 s δ 7.44 s C 125.4 C 125.4

1.64 ddd 1.64 ddd 71.6 71.6 2.54 dd 2.54 dd 4.44 ddd 43.9 4.44 ddd 43.9 5.70 dt 5.53 dd 5.70 dt 5.53 dd 2.01 br d 2.13 dd 2.01 br d 2.13 dd 1.42 br s 171.6 1.42 br s 171.6 68.6 37.0 68.6 37.0 2.4 Hz 66.5 2.4 Hz 66.5 51.8 51.8 1.3 Hz 1.3 Hz 52.5 52.5 135.7 1.20 td15.6 1.22 s 135.7 1.20 td 15.6 1.22 s 6.54 dd 6.54 dd 18.3 18.3 37.6 2.17-2.09 m 37.6 2.17-2.09 m 171.6 171.6 141.2 37.0 141.2 37.0 2.15 s 21.2 2.15 s 21.2 1.78 dtd 1.78 dtd 2.56 dt 2.56 dt 166.2 166.2 21.6 21.6

1.69 s 1.69 s 51.7 51.7 3.74 s 3.74 s

Figure 2.23: NMR assignments and key 2D correlations for 1d (stereoview).

δ H δ (ppm) H (ppm)

4.5 1.2 1.4 5.0 1.6 1.8 2.0 5.5 2.2 2.4 6.0 2.6 2.8 6.5 3.0 3.2 7.0 3.4 3.6 7.5 3.8

140 130 120 110 100 90 80 70 50 45 40 35 30 25 20 15

δC(ppm) δC(ppm)

Figure 2.24: HMQC spectrum of 1d (400 MHz, CDCl3). 2.2. STRUCTURE ELUCIDATION. 101

δH (ppm)

1.5 2.0

2.5 4.36 3.0 4.40

4.44 5.45 3.5 5.50 4.48 5.55 4.0 4.52 5.60 5.65 4.5 4.56 5.70 143 142 141 140 139 138 137 136 135 134 5.75 5.0 45 44 43 42 41 40 39 38 37 36 5.5 6.0 6.5 7.0 7.5 160 140 120 100 80 60 40 20

δC(ppm)

1 Doublets resulting from direct coupling ( JCH) have been removed digitally.

Figure 2.25: HMBC spectrum of 1d. 102 CHAPTER 2. ISOLATION.

O

O OH O H H O O

O O 1e 7 3 6 9 0 0 6 2 2 6 7 5 1 4 2 5 7 7 ...... 1 1 3 7 2 1

2.2 2.1

1.9 1.8 1.7 1.6 7.4 1.3 1.2 1.1

5.6 5.5 5.4

6.4 4.5 4.4

2.6 2.5 2.4

7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 ppm

0.84 0.92 1.05 3.02 5.19 0.97 3.16 1.53 0.91 0.93 1.03 1.06 2.08 0.87 1.18 1.20 3.12

1 Figure 2.26: H NMR spectrum of 1e (400 MHz, CDCl3).

7.42 t 7.42 t J (Hz) J (Hz) 143.9 143.9 δ δ δ δ H H H H HMBC HMBC δ (REVISED ASSIGNMENT) 108.4 139.3 δ (REVISED ASSIGNMENT) 108.4 139.3 C C 6.41 dd 6.41 dd δ 7.44 br s δ 7.44 br s C C 125.8 125.8

2.17 s 1.62 dd 2.17 s 1.62 dd 71.7 71.7 2.46 dd 2.46 dd 5.40 dd 5.40 dd 44.4 5.60 ddd 44.4 21.0 4.46 ddd 21.0 4.46 ddd 5.60 ddd 169.8 2.18-2.07 m 169.8 2.18-2.07 m 1.30 br s 171.8 1.30 br s 171.8 72.3 37.5 72.3 37.5 2.4 Hz 64.3 2.4 Hz 64.3 1.6 Hz 51.7 1.6 Hz 51.7 54.0 54.0 131.5 1.47 s 131.5 1.47 s 6.43 dd 1.19 td16.2 6.43 dd 1.19 td 16.2 1.94 dd 18.4 1.94 dd 18.4 37.8 2.18-2.07 m 37.8 2.18-2.07 m 143.4 37.0 143.4 37.0 166.0 1.84 dtd 166.0 1.84 dtd 2.52 ddd 2.52 ddd 21.9 21.9 1.72 s 1.72 s 51.8 51.8

3.73 s 3.73 s

Figure 2.27: NMR assignments and key 2D correlations for 1e (stereoview). 2.2. STRUCTURE ELUCIDATION. 103

δH (ppm) δH (ppm)

4.5 1.4 1.5 1.6 5.0 1.8 2.0 5.5 2.2 2.4 6.0 2.6 2.8 6.5 3.0 3.2 7.0 3.4 3.6 7.5 3.8 140 130 120 110 100 90 80 70 55 50 45 40 35 30 25 20 15 δ (ppm) δC(ppm) C

Figure 2.28: HMQC spectrum of 1e.

δH (ppm)

1.6 5.4 1.3 1.8 5.6 1.4 2.0 1.5 5.8 2.2 1.6 6.0 1.7 2.4 6.2 1.8 2.6 6.4 1.9 2.8 6.6 2.0 3.0 6.8 2.1 3.2 7.0 2.2 3.4 7.2 2.3 2.4 3.6 7.4 2.5 3.8 170 165 160 155 150 145 140 135 130 125 165 160 155 150 145 140 135 130 125 120 115 110 70 65 60 55 50 45 40 35 30 25 20 15

δC(ppm)

1 Doublets resulting from direct coupling ( JCH) have been removed digitally.

Figure 2.29: HMBC spectrum of 1e. 104 CHAPTER 2. ISOLATION.

2.2.5.2 Interconversion of 1c-1e via Diol 1h.

O

1 OR O H H R2O R2 R1 O 1c Ac Ac a 1d HAc a 1e Ac H b O O a 1h HH

a) Ac2O, pyridine, DMAP b) Na2CO3,CH2Cl2/MeOH (1:1), rt. Scheme 2.1: Interconversion of 1c-1h. 1 3 7 . 3 3 6 2 0 7 0 0 6 6 4 2 7 . . . . 1 1 1 7 (+D2O)

2.50 2.45 2.40 2.35 2.30 2.25 2.20 2.15 2.10 2.05 7.5 7.4

(+D2O)

5.7 5.6 1.90 1.85 1.80 1.75 1.70 1.65 1.60 1.55

(+D2O)

1.20 1.15 1.10

6.5 6.4 4.4 4.3 4.2

7 6 5 4 3 2 ppm

0.78 0.82 1.01 3.00 1.95 1.15 3.18 0.99 0.78 0.83 0.95 1.03 2.06 2.14 3.16 1.31

1 Figure 2.30: H NMR spectrum of 1h (400 MHz, CDCl3).

To confirm the structures of 1c-1e,diol1h was prepared by deacetylation of

1c. Tidgewell et al’s standard conditions for deacetylation of 1a (Na2CO3 in minimal MeOH)169 caused extensive epimerisation (see Section 3.2.4). How- ever, epimerisation could be suppressed by addition of CH2Cl2,affording1h 2.2. STRUCTURE ELUCIDATION. 105

7.42 t 7.42 t J (Hz) J (Hz) 143.9 143.9 δ δ δ δ H H H H 108.4 NOESY 139.3 NOESY 108.4 139.3 HMBC 6.40 dd HMBC 6.40 dd δ 7.43 br s δ 7.43 br s C 125.8 C 125.8

1.60 ddd 1.60 ddd 71.8 71.8 2.49 dd 2.49 dd 4.28 dd 44.4 4.28 dd 44.4 4.32 br d 5.60 dd 4.32 br d 5.60 dd 2.40-2.30 m 2.14-2.10 m 2.40-2.30 m 2.14-2.10 m 1.22 d 172.0 1.22 d 172.0 69.6 37.5 69.6 37.5 65.6 65.6 2.4 Hz 51.8 2.4 Hz 51.8 2.40 - 54.1 1.47 s 2.40 - 54.1 1.47 s 135.3 2.30 m 16.3 135.3 2.30 m 16.3 6.48 dd 1.16 tdd 18.4 6.48 dd 1.16 tdd 18.4 37.5 2.09 dq 37.5 2.09 dq 142.2 37.0 142.2 37.0 1.82 dtd 1.82 dtd 166.6 2.50-2.45 m 166.6 2.50-2.45 m

22.1 22.1

51.7 1.70 s 51.7 1.70 s

3.73 s 3.73 s

Figure 2.31: NMR assignments and key 2D correlations for 1h (stereoview).

δH (ppm) 1.5 2.0 1.8 2.5 a) HMQC 2.0 b) HMBC 3.0 2.2 3.5 2.4 4.0 2.6 4.5 2.8 5.0 5.5 3.0 6.0 3.2 6.5 3.4 7.0 3.6 7.5 3.8 130 110 90 70 50 30 10 170 165 160 155 150 145 140 δC(ppm)

Figure 2.32: HMQC and HMBC spectra of 1h.

4.3 2.1 4.4 4.5 4.6 2.2 4.7 4.8 nOe effect 4.9 2.3 direct coupling 5.0 5.1 5.2 2.4 5.3 5.4 2.5 5.5 5.6 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 ppm

Figure 2.33: NOESY spectrum of 1h. from 1c or 1d in up to 86% yield. K2CO3 was equally effective.

Analysis of the NMR data (Figures 2.30 on the facing page and 2.32) was straightforward. Carbons C-17 and -18 were distinguished on the basis of 106 CHAPTER 2. ISOLATION.

HMBC correlations (Figure 2.32), and H-1 and 2 on the basis of coupling constants to H-3 (Figure 2.31 on the preceding page). These couplings were only resolved after D2O exchange. The relative stereochemistry was confirmed using NOESY data (Figure 2.33 on the previous page). H-1 showed cross peaks to H-2, -10 and -11α,whileH-8correlatedtoH-10and-11β,andH-12 correlated to H-20 (Figure 2.31 on the preceding page). This confirms the con- figurations shown for 1h, and therefore of 1c-1e, through the interconversions summarised in Scheme 2.1 on page 104.

The proposed structure of 1d was verified by acetylation, which proceeded smoothly to give 1c, identical in all respects with the isolated material - TLC cospot, 1Hand13C NMR, IR, and optical rotation (Scheme 2.1 on page 104). Insufficient 1e was available for acetylation. In our initial publication,215 we assumed that 1e would be too hindered for direct acetylation under standard conditions. This proved incorrect; the 1-hydroxy group of 1h is readily acety- lated. However, the crude product after three hours consisted of 1d and 1e in approximately 1:2 ratio, with only traces of 1c; thus, the 1-position of 1h is (as expected) less reactive than the 2-position, and each of the monoacetates is less reactive than the diol. At 45 ◦C, the starting material was consumed in approximately 1 hour, giving a mixture of 1c and 1e, confirming that 1e is the less reactive of the monoacetates. It also appears to be less stable; after acetylation under these standard conditions, it was contaminated by an insep- arable impurity, apparently an isomer other than the 8-epimer. This did not occur with 1d. 1e was also much more prone to decomposition in storage than 1d.

During the isolation of 1c, Valdés detected traces of what appeared to be 1d, 1e,and1h in S. divinorum;255 while we subsequently confirmed the presence of 1d and 1e,215 no-one has yet isolated 1h. It may be that the compound occurs in the plant in very low levels; this would be expected if its formation were the rate-limiting step on the path to 1c–1e. Alternatively, current iso- lation procedures may give poor recoveries. Indeed, like 1b, 1h precipitates when loaded on silica gel in most solvent systems. Stripping the column with 2.2. STRUCTURE ELUCIDATION. 107

MeOH/CH2Cl2 was required to achieve satisfactory recovery.

Asampleof1h was supplied to Claudio Medana for LCMS comparison with the crude extract. This proved difficult, since 1b and 1h were not resolved by reverse-phase HPLC, either C-18 functionalised silica or polymeric C-18. The compounds were ultimately resolved on a porous graphitic carbon column, eluting with a MeOH to CH2Cl2 gradient. Under these conditions, a crude

256 CH2Cl2 extract showed no detectable 1h.

2.2.5.3 Salvinorin F (1f).

The molecular formula of compound 1f,C21H26O6, was established by HRES- IMS. In contrast to 1d and 1e, 1H NMR (Figure 2.34 on the next page) showed no acetyl peak, only one oxymethine signal, and a diastereotopic methylene (δ 2.35 and 2.60) coupling to H-1 and H-3 (Figure 2.35 on page 109). This implied the 2-deoxy structure shown. Protons H-2α and H-2β were distin-

257 guished by their coupling constants: molecular modelling predicted J 1,2β

= 5.4 Hz (observed: 5.5 Hz) and J 1,2α = 1.5 Hz (observed: 1.1 Hz). In our initial publication,215 this latter coupling was not observed, but subsequent resolution enhancement allowed direct measurement. H-2β also showed the expected HMBC cross peaks to H-4 and -10 (Figure 2.37 on page 109). 108 CHAPTER 2. ISOLATION.

O

OH O H H 2 1 10 O

3 4

O O 1f 6 6 7 0 0 7 0 6 2 5 4 7 7 5 2 . . . . . 1 1 7 3 1

2.6 2.5 2.4 2.3 7.5 7.4 1.8 1.7 1.6 2.2 2.1

6.7 6.6 6.5 6.4

1.3 1.2 1.1 4.6 4.5 4.4 5.7 5.6 5.5

7 6 5 4 3 2 ppm

0.82 0.90 1.06 2.08 2.16 3.07 1.04 3.19 1.28 0.85 0.97 1.01 3.02 1.01 1.05 1.02 1.11 1.31

1 Figure 2.34: H NMR spectrum of 1f (400 MHz, CDCl3). 2.2. STRUCTURE ELUCIDATION. 109

7.42 t 7.42 t J (Hz) J (Hz) 143.8 143.8 δ δ δ δ H H H H 108.4 108.4 6.41 dd 139.3 6.41 dd 139.3 HMBC HMBC δ 7.44 br s δ 7.44 br s C 125.9 C 125.9

1.62 ddd 1.62 ddd 5.5 Hz 71.7 5.5 Hz 71.7 2.46 dd 2.46 dd 2.60 ddd 44.4 5.60 ddd 2.60 ddd 44.4 5.60 ddd 1.1 Hz 4.51 br dd 1.1 Hz 4.51 br dd 2.17-2.08 m 1.25 br s 172.1 1.25 br s 2.17-2.08 m 172.1 37.7 37.7 38.0 63.9 38.0 2.35 ddt 2.35 ddt 63.9 52.2 52.2 54.8 1.48 s 54.8 1.48 s 133.4 1.18 tdd16.4 133.4 1.18 tdd 16.4 18.6 18.6 6.67 ddd 1.29 dd 2.17-2.08 m 6.67 ddd 1.29 dd 2.17-2.08 m 140.6 36.6 37.3 140.6 36.6 37.3 1.82 dtd 1.82 dtd 166.9 2.53 dt 166.9 2.53 dt

21.6 21.6 1.71 s 1.71 s 51.5 51.5 3.72 s 3.72 s

Figure 2.35: NMR assignments and key 2D correlations for 1f (stereoview).

δH (ppm)

4.5 1.2 1.4 5.0 1.6 1.8 5.5 2.0 2.2 6.0 2.4 2.6 6.5 2.8 3.0 7.0 3.2 3.4 7.5 3.6 3.8

140 130 120 110 100 90 80 70 δC(ppm) 55 50 45 40 35 30 25 20 15

Figure 2.36: HMQC spectrum of 1f.

δH (ppm) 1.5 1.1 2.0 1.2 1.3 2.5 1.4 3.0 1.5 3.5 1.6 4.0 1.7 4.5 1.8 5.0 1.9 2.0 5.5 2.1 6.0 2.2 6.5 2.3 7.0 2.4 7.5 2.5

160 150 140 130 120 110 δC(ppm) 70 65 60 55 50 45 40 35 30 25 20 15

1 Doublets resulting from direct coupling ( JCH) have been removed digitally.

Figure 2.37: HMBC spectrum of 1f. 110 CHAPTER 2. ISOLATION.

2.2.6 Divinatorins A-C (28a-28c).

2.2.6.1 Structure Elucidation.

O

R1 H H R1 R2 R3

R2 28a OH H H 28b OH OH Me 18 28b H OAc H O OR3 29a H H H

Figure 2.38: Divinatorins A–C (28a-28c) and hardwickiic acid (29a).

O

OH H H

O OH 28a

2.1 2.0 1.9 1.8 7.3 7.2 2.6 2.5 2.4 2.3 1.72 1.70 1.68 1.66

6.92 6.90 6.88 1.6 1.5 1.4 1.3 1.2 4.51 4.50 4.49 4.48 4.47

6.26 6.25

0.87 0.92 1 1.1 4.2 2.1 3.2 0.88 0.99 1 3.1 1.1 2.1 1.1 3 7 6 5 4 3 2 1 ppm

1 Figure 2.39: H NMR spectrum of 28a (400 MHz, CDCl3). 2.2. STRUCTURE ELUCIDATION. 111

7.36 t 7.36 t J (Hz) J (Hz) 142.8 142.8 δ δ δ δ H H 138.4 H H 138.4 110.9 7.20 td 110.9 7.20 td NOESY NOESY HMBC HMBC δ 6.25 dt 125.2 δ 6.25 dt 125.2 C C 18.2 18.2 2.05 ddd 2.34 td 2.05 ddd 2.34 td 1.67 ddd 1.67 ddd 1.85 ddd 39.1 1.85 ddd 39.1 4.49 br d 4.49 br d 4.9 Hz 4.9 Hz 1.60-1.54 m 1.60-1.54 m 2.56 ddd 1.45 br s 2.56 ddd 1.45 br s 39.7 15.7 39.7 15.7 64.7 64.7 37.1 0.84 d 37.1 0.84 d 2.43-2.33 m 38.1 49.0 1.15 s 2.43-2.33 m 38.1 49.0 1.15 s 19.8 19.8 2.7 Hz 2.7 Hz 4.8 Hz 1.23-1.17 m 27.4 4.8 Hz 1.23-1.17 m 27.4 136.2 37.4 1.47-1.42 m 136.2 37.4 1.47-1.42 m 6.90 ddd 38.6 6.90 ddd 38.6 140.8 1.60-1.54 m 140.8 1.60-1.54 m 21.4 21.4 2.43-2.33 dt 2.43-2.33 dt 171.8 171.8

1.64 s 1.64 s

Figure 2.40: NMR assignments and key 2D correlations for 28a (stereoview).

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

140 130 120 110 100 90 80 70 60 50 40 30 20 ppm

Figure 2.41: HMBC spectrum of 28a.

The structures of 28a-28c were elucidated mainly by NMR spectroscopy (1H, 13C, DEPT, HMQC, HMBC, COSY, and NOESY). The 1H NMR spectra suggested that they were derivatives of hardwickiic acid (29a). The molecular formula of 28a (C20H28O4 by HRESIMS) differed from that of 29a by one oxygen atom, suggesting the presence of a hydroxy group, which was confirmed 112 CHAPTER 2. ISOLATION.

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

nOe effect 2.1 direct coupling 2.2

2.3

2.4

2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 ppm

4.47

4.48

4.49

4.50

4.51

4.52 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 ppm

Figure 2.42: NOESY spectrum of 28a. by a strong IR absorption at 3392 cm−1. This was located at C-1 based on 1H NMR: the oxymethine at δ 4.49 showed couplings to H-2, -3 & -10 (COSY), and C-3, -9 and -10 (HMBC) - see Figure 2.41 on the preceding page. The 1α configuration was confirmed by an H-1 to H-11 crosspeak in the NOESY spectrum (Figure 2.42). NOESY cross peaks also permitted stereospecific assignment of the rotatable H-11 and -12 methylenes, as shown in Figure 2.40 on the preceding page. In the favoured conformation thus established, C-12 is approximately anti-periplanar to the C-20 methyl group, as found in the crystal structure of 29a228 (Figure 2.18 on page 95). Along with the cross peaks from H-20 to H-17 and -19, these NOESY data confirm the relative configuration of all stereogenic centres. 2.2. STRUCTURE ELUCIDATION. 113

O

OH 7.3 7.2 H H OH 2.1 2.0 1.9 1.8

OO 28b

1.60 1.55

6.68 6.66 6.64

2.5 2.4 2.3 3.4 3.85 6.26 6.24 4.45

0.76 0.81 1 0.93 3 3.1 4.6 4.1 0.77 0.92 0.89 3 0.93 1 2.9 0.92

7 6 5 4 3 2 ppm 1

1 Figure 2.43: H NMR spectrum of 28b (400 MHz, CDCl3).

7.35 t 7.35 t

142.8 142.8 J (Hz) J (Hz) 138.4 138.4 δ δ 110.8 7.20 dd δ δ 110.8 7.20 dd H H 6.25 dd H H 6.25 dd 124.9 124.9 NOESY NOESY HMBC HMBC δ δ C 18.2 2.42 dddd C 18.2 2.42 dddd 2.08 dddd 2.08 dddd 1.77 ddd 1.77 ddd 1.90 ddd 1.90 ddd 4.46 dq 38.8 4.46 dq 38.8 5.1 Hz 3.84 dd 5.1 Hz 3.84 dd 1.64-1.49 m 1.49 br s 1.64-1.49 m 1.49 br s 2.53 ddd 1.45 br s 2.53 ddd 1.45 br s 39.1 39.1 64.3 63.9 64.3 63.9 2.31 ddt 44.8 2.31 ddt 44.8 38.0 48.7 1.18 s 38.0 48.7 1.18 s 20.9 3.38 dd 20.9 3.38 dd 2.8 Hz 1.49 br s 2.8 Hz 1.49 br s 4.9 Hz 4.9 Hz 1.19-1.13 m 21.9 1.19-1.13 m 21.9 133.2 37.1 1.85 dq 133.2 37.1 1.85 dq 6.65 ddd 38.0 6.65 ddd 38.0 141.4 1.64-1.49 m 141.4 1.64-1.49 m 21.4 21.4 2.36 dt 2.36 dt 167.3 167.3

1.66 s 1.66 s 51.3 51.3

3.71 s 3.71 s

Figure 2.44: NMR assignments and key 2D correlations for 28b (stereoview).

The 1H NMR spectrum of 28b suggested a methyl ester (δ 3.71) with two hy- droxy groups (δ 1.49, 2H, D2O-exchangeable). IR showed a strong absorption −1 at 3434 cm . The molecular formula, C21H30O5 (HRESIMS), was consistent with this. One of the hydroxy groups was again located at C-1, showing the same couplings as in 28a. The second was located at C-17, based on the cou- 114 CHAPTER 2. ISOLATION.

3.4

3.5

3.6

3.7

3.8

3.9

4.0

4.1

4.2

4.3

4.4

130 120 110 100 90 80 70 60 50 40 30 20 ppm

Figure 2.45: HMBC spectrum of 28b.

3.4 1.2 3.5 3.6 1.3 3.7 3.8

1.4 3.9 4.0

1.5 4.1 nOe effect 4.2 direct coupling 1.6 4.3 4.4

1.7 4.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 ppm

Figure 2.46: NOESY spectrum of 28b. plings of the oxymethylene signals (δ 3.38 and 3.84) to H-8 (COSY) and to C-7, -8 and -9 (HMBC) — see Figure 2.45. The NOESY spectrum showed H-1 to H-11 and H-17 to H-20 cross peaks, confirming the configuration at these centres (Figure 2.46). Again, NOESY data permitted full assignment of the diastereotopic H-11, -12 and -17 methylene protons.

The 1H NMR spectrum of 28c showed an acetyl methyl signal (δ 2.03). Com- 2.2. STRUCTURE ELUCIDATION. 115

O

O H H O

O OH 28c 3 3 7 0 2 8 . . . 1 0 2

2.5 2.4 2.3 2.2

7.3 7.2

1.8 1.7 1.6 1.5 1.4

1.2 1.1

6.91 6.90 6.89 6.88 6.87

4.25 3.80 3.75

6.29 6.28 0.81 0.87 1 2.1 3 3.3 1.2 0.8 1 1 0.99 2 5.4 3.8 2.9

7 6 5 4 3 2 1 ppm

1 Figure 2.47: H NMR spectrum of 28c (400 MHz, CDCl3).

7.35 t 7.35 t

δ δ 142.8 δ δ 142.8 H H H H 138.5 138.5 NOESY 7.22 dd NOESY 7.22 dd HMBC 110.9 HMBC 110.9

δ 125.2 δ 125.2 C 6.28 dd 2.03 s C 6.28 dd 2.03 s

18.3 21.0 18.3 21.0 2.40 td 2.40 td 1.64 ddd 171.2 1.64 ddd 171.2 38.9 38.9 2.24-2.16 m 1.82-1.67 m 4.26 dd 2.24-2.16 m 1.82-1.67 m 4.26 dd 1.82-1.67 m 1.82-1.67 m 1.42 br d38.4 66.1 1.42 br d38.4 66.1 17.0 17.0 40.9 40.9 46.8 46.8 2.35 dt 27.4 0.83 s 3.79 dd 2.35 dt 27.4 0.83 s 3.79 dd 1.54-1.44 m 19.0 1.54-1.44 m 19.0 1.15 td 22.3 1.15 td 22.3 140.3 37.4 1.82-1.67 m 140.3 37.4 1.82-1.67 m 6.89 dd 35.2 6.89 dd 35.2 141.2 1.54-1.44 m 141.2 1.54-1.44 m 20.5 20.5 2.53 dt 2.53 dt 171.9 171.9

1.27 s 1.27 s

Figure 2.48: NMR assignments and key 2D correlations for 28c (stereoview). 116 CHAPTER 2. ISOLATION.

1.0

1.5

2.0

2.5

3.0

3.5

4.0

70 60 50 40 30 20 ppm

Figure 2.49: HMBC spectrum of 28c.

3.8 0.9 1.0 3.9 1.1 1.2 4.0 nOe effect 1.3 direct coupling 1.4 4.1 1.5 1.6 4.2 1.7

4.3 1.8 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 ppm

Figure 2.50: NOESY spectrum of 28c. pared to 28b, the C-1 oxymethine was absent, while the oxymethylene signals were shifted downfield to δ 3.79 and 4.26, with the same COSY and HMBC couplings (Figure 2.49), establishing the 17-acetoxy structure shown. This was consistent with the molecular formula, C22H30O5 (HRESIMS). As with 28b, NOESY data (Figure 2.50) confirmed the configuration of all stereogenic cen- tres and permitted full assignment of the H-11, -12 and -17 methylene protons.

The neoclerodane absolute stereochemistry shown for the salvinorins and div- inatorins is common to all clerodanes isolated from the Lamiaceae,26 including 2.2. STRUCTURE ELUCIDATION. 117

1a24, 25 and 29a.226

2.2.6.2 Derivation of the Name Divinatorin.

The inspiration for the name “divinatorin” was a comment by Albert Hofmann, who obtained the type specimen of S. divinorum for botanical identification:

It was determined at the Botanical Department at Harvard that it was a new species of Salvia and it got the name Salvia divinorum. It is a wrong name, bad Latin; it should be actually Salvia divinato- rum. They do not know very good Latin, these botanists. I was not very happy with the name because Salvia divinorum means “Salvia of the ghosts” whereas Salvia divinatorum, the correct name, means “Salvia of the priests” ...258

The original name was coined by Carl Epling, a botanist fluent in Latin, and was intended to mean “of the seers.”4 According to classicist Dr Parshia Lee- Stecum, the original name is in fact correct, being taken from the adjective divinus:259

divinus ... can be (and was as early as classical antiquity) used as a substantive, meaning “soothsayer/prophet/seer”. Hence divino- rum: “of the prophets/soothsayers/seers”.

Hofmann’s alternate rendering, derived from the medieval Latin noun “div- inator”, would have had the same meaning in the medieval period. However, this word was not used in antiquity, and “divinatorum” would have been in- terpreted as a form of the verb divino (I foresee):

To someone of ... Virgil’s, Ovid’s or even Tacitus’ time, Salvia div- inatorum would thus mean “salvia of those who have been foretold” ...

Nonetheless, in honour of Hofmann’s key role in the scientific identification of the plant, compounds 28a-28c were named divinatorins. 118 CHAPTER 2. ISOLATION.

2.2.7 Subsequent isolations.

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

4

O O O O R O O R 1g 28d CH2OAc 34a OMe 28e CHO 34b OMe

Figure 2.51: Subsequently isolated compounds.

Five new diterpenoids have been isolated from S. divinorum since the initial publication of this work. Lee et al isolated salvinorin G (1g) and divinatorins DandE(28d & 28e)137 after decolourisation of the crude extract as described above. The structures of the three compounds, elucidated using 2D NMR, were incorrectly drawn with tetrahedral geometry at C-4 in the original publication; this has been corrected in Figure 2.51. Harding et al isolated salvinicins A and B(34a & 34b)25 without decolourisation. The structure of 34a was elucidated using 2D NMR, and definitively confirmed by X-ray crystallography. Chapter 3

Synthesis.

In order to study the structure-activity relationships of salvinorin A (1a)at opioid receptors, various synthetic modifications were made to the compound.

3.1 Known derivatives.

O O O

O O O O O O OR1 O H H H H H H O O R2O OH O O

O O O O O O

R2 R1 ab 35 1a 36h HH c 36e Ac H 36d HAc d c 36c Ac Ac i ◦ Conditions: a) Bu2AlH, THF, –78 C, 25 min, 65% (81% borsm); b) NaBH4,EtOH,40 ◦ ◦ C, 56%; c) Ac2O, pyridine, DMAP, rt; d) i) trimethyl orthoacetate, AcOH, 105 C; ii) AcOH, cat. HCl, THF/H2O, rt, 74% over 2 steps. Scheme 3.1: Preparation of known derivatives.

Several known derivatives of 1a were prepared by published procedures with i minor modifications. Lactol 35 was prepared using Bu2AlH following Brown’s

119 120 CHAPTER 3. SYNTHESIS. procedure,260 albeit in lower yield. Brown reported a yield of 89% after 6 minutes at –78 ◦C. In my hands, the reaction was not complete even after 25 minutes; yield plateaued at 65% (81% based on recovered starting material). Although warming to –40 ◦C forced the reaction to completion, the yield was not improved due to side reactions.

Diol 36h was originally prepared by Valdés from 1a using 0.6 equivalents of

i 23, 15 261 NaBH4 in PrOH; as found by Brown, EtOH proved equally satisfactory; also, the precise stoichiometry of NaBH4 proved unimportant, although a large excess (10 eq) reduced the yield.

Acetylation of 36h under standard conditions gave dihydrosalvinorin E (36e).23 Several trace byproducts could not be completely removed, even by HPLC; this problem was solved by HPLC purification of the starting diol 36e itself.

Using Ac2O/pyridine/DMAP, direct formation of dihydrosalvinorin C (36c) from 36h occurs in negligible yield, even after prolonged reflux.262 Valdés et al’s indirect route via the orthoacetate and 36d27 was therefore used. Satis- factory purification of 36c required HPLC (40% EtOAc/petrol). Harding et al have since reported that 36h can be directly diacetylated using NEt3 instead of pyridine.82

3.2 Epimerisation at C-8 under Basic Condi- tions.

3.2.1 Previous Reports.

In the NaBH4 reduction of 1a described above, Valdés isolated equal amounts of 36h and a byproduct which “appears to be stereoisomeric at C-8 and/or C-9”.23 Acetylation and oxidation of this compound gave “a thus far undeter- mined stereoisomer” of 1a. Subsequently, Brown identified this compound as 8-epi-salvinorin A (37a, Scheme 3.2 on the facing page), and also reported that deacetylation of 1a under basic conditions gave 8-epi-salvinorin B (37b).263 3.2. EPIMERISATION AT C-8 UNDER BASIC CONDITIONS. 121

No characterisation data was reported for any of these compounds. Further references to these epimerisations appeared over the following decade,24, 14 but no data was published until 2001, when Valdés et al characterised diol 38h.27 No basis was given for this proposed structure, however.

O

O O H H HO 8 O O Ac O / pyr O B- 2

ROH O O O O O O O O H H H H O 37b O O 8 O

NaHCO3, DMF O O 1a O O 37a

O O PCC NaBH4

O OH O OH O H H H H HO O 8 O 8 O

Ac O / pyr O O 2 O O 38h 38e Scheme 3.2: Formation of 8-epi-salvinorins A and B.

3.2.2 8-epi-Salvinorins A and B (37a and 37b).

Deacetylation of 1a under mildly basic conditions (saturated NaHCO3 in MeOH)gaveamixtureof1b and another compound. The new compound proved, as Brown had claimed, to be 8-epi-salvinorin B (37b). Acetylation with Ac2O/pyridine gave 8-epi-salvinorin A (37a).

The structures of 37a and 37b were elucidated using 2D NMR (HMQC, 122 CHAPTER 3. SYNTHESIS.

1a 11.6, 3.1 37a 5.0, 2.2 1b 11.7, 3.0 37b 4.8, 2.2

Table 3.1: Coupling constants (Hz) at H-8 for 1a, 1b and 8-epimers.

HMBC, COSY and NOESY). The configuration of H-8 was apparent from the loss of the trans-diaxial coupling constant, establishing an equatorial configu- ration (Table 3.1). Also, irradiation of H-12 gave a strong nOe enhancement of H-8 (Figure 3.1 on the next page). The corresponding experiment on 1a caused an enhancement of H-20 rather than H-8 (Figure 2.14 on page 92). The coupling constants and NOESY correlations shown in Figure 3.1 also establish an approximately trans-diaxial relationship between H-11β and H-12. This suggests the preferred conformation is approximately as shown, with the furan equatorial and the C ring in a pseudoboat conformation. Consistent with this, X-ray crystallography of stereoisomeric furanolactones has shown that in the solid state, the furan is invariably equatorial264, 265, 266 even where the C ring is forced into a pseudoboat conformation.267 NMR analysis, based on coupling constants and NOESY spectra, suggests that this also holds in solution.266 It should be noted that both 37a and 37b lack the distinctive infrared car- bonyl absorption reported for pseudoboat δ–lactones (≈ 1760 cm−1).268 How- ever, those values were obtained in solution, and are subject to solvent effects; recording the spectrum as a mull gives typical carbonyl absorptions ( ≤ 1730 cm−1).269 The spectra of 37a and 37b were recorded as neat films.

Our initial publication270 contained characterisation data for 37a but not for 37b. Harding et al82 and Lee et al271 subsequently published 1HNMRdata for 37b which is inconsistent with our own.

3.2.3 Control of Epimerisation and Separation of Epimers.

Brown reported that treating 1a with KCN in refluxing MeOH/THF gave 1b in quantitative yield,272 but in my hands 37b was the major product, as had

273 beenthecasewithNaHCO3. Valdés noted the same result. Additionally, the epimers were not resolved on silica using Brown’s solvent system (3% 3.2. EPIMERISATION AT C-8 UNDER BASIC CONDITIONS. 123

δH δH δH δH

NOESY NOESY

1.50 dd 2.36 dd (15.0, 2.2 Hz) 1.50 dd 2.36 dd (15.0, 2.2 Hz) (15.0, 12.0 Hz) 5.25 dd (15.0, 12.0 Hz) 5.25 dd (12.0, 2.2 Hz) (12.0, 2.2 Hz)

2.24 br s 2.24 br s 2.45 dd 2.45 dd (5.0, 2.2 Hz) (5.0, 2.2 Hz)

Figure 3.1: Key NOESY correlations for 37a (stereoview).

MeOH/CH2Cl2). Petrol-based systems resolved the epimers, but gave poor recoveries: some precipitation occurred when loading in these systems. Thus, effective separation required elution with petrol-based systems for separation, followed by stripping with MeOH/CH2Cl2 to achieve full recovery of 1b. Yield was generally below 50%, making this an unsatisfactory route to 1b.

Both of these problems were elegantly overcome by Tidgewell et al,169 using a suspension of 1a and Na2CO3 in minimal MeOH at room temperature. In my hands, very little 37b was evident by TLC under these conditions. This was removed by trituration in MeOH, giving 1b in 76% yield without chromatography. The supernatant from trituration was a complex mixture containing very little 37b.

These results were published without discussion, but demand explanation. It is puzzling that Na2CO3 appears to cause less epimerisation than the much weaker base NaHCO3 (the conjugate acid of Na2CO3). When Tidgewell et al’s procedure was repeated in oxygen-free MeOH, the supernatant contained almost pure 37b,whichwasobtainedasanamberresinin23%yieldbyevap- oration. The precipitate consisted again of crystalline 1b (74%). Thus, 1b is nearly insoluble in room-temperature MeOH, while 37b is freely soluble. Additionally, these results confirm that as expected, epimerisation does occur 124 CHAPTER 3. SYNTHESIS. under Tidgewell et al’s conditions, but is usually not apparent due to decom- position in oxygenated solution. Nonetheless, the yield of 1b is not affected, so the reaction can be performed without precautions against oxygen or moisture.

Apparently, then, the improved epimeric ratio results from performing the reaction as a suspension in minimal MeOH. The reaction solution must rapidly become saturated with 1b; as additional 1b is generated, it precipitates, unlike 37b. Thus, although an equilibrium mixture of the epimers is present in solution, this represents only a small proportion of the total reaction mixture. When the deacetylation is performed entirely in solution, for example by the use of cosolvents such as THF or CH2Cl2, the epimeric mixture ultimately comprises the entire reaction mixture. Similarly, as will be discussed below, the procedure causes extensive epimerisation when applied to 1c, where there are no dramatic differences in MeOH solubility among starting material and products.

Substituting NaHCO3 for Na2CO3 under Tidgewell et al’s conditions gave no reaction after one day. Conversely, Na2CO3 in MeOH/CH2Cl2 gave a complex mixture; the distinctive H-12 signals of 8-epimers were apparent by 1HNMR. The above deacetylations proceed via base-catalysed solvolysis; the use of po- lar “aprotic” solvents (strictly, non-hydrogen bond donor solvents) permits epimerisation without deacetylation. Thus, treatment of 1a with NaHCO3 in dry DMF or DMPU at 150 ◦C gives approximately 50% epimerisation to 37a (addition of water caused deacetylation). In commercial DMPU (98% purity), epimerisation occurred in the absence of NaHCO3, presumably due to a ba- sic impurity; heating in distilled DMPU or DMF alone gave no reaction. Of course, this reaction can also be used to regenerate the natural compounds from the 8-epimers.

3.2.4 8-epi-Salvinorin C (37c) and Related Compounds.

As discussed above (Section 2.2.5.2), application of Tidgewell et al’s deacety-

169 lation conditions to 1c (Na2CO3 in minimal MeOH) gave predominantly 3.2. EPIMERISATION AT C-8 UNDER BASIC CONDITIONS. 125 the 8-epi-diol 37h (in a 4:1 ratio with 1h) — see Scheme 3.3. Again, KCN in MeOH gave the same result. Both epimers were too soluble in MeOH to permit trituration. The 8-epimer 37h was freely soluble in Me2SO, but poorly soluble in acetone and chloroform, with a tendency to precipitate during preparation of NMR samples.

O O

O O 1 O O OR O H H H 2 H O R O 1 O O R2 R 37c Ac Ac 37d HAc c b O O O O 37e Ac H b 1c 37h HH a

Conditions: a) Na2CO3,MeOH,rt× 4h. b) Ac2O, pyridine, (± DMAP) c) CDCl3,rt overnight Scheme 3.3: Synthesis of 37c-37h.

Acetylation of 37h under mild conditions gave the 2-acetate 37e in high yield.

Surprisingly, standing 37e in CDCl3 overnight gave 50% conversion to the 1- acetate 37d, presumably due to traces of DCl. Filtration of either compound through a plug of basic Al2O3 also caused this unexpected acetate migration. Alumina-catalyzed acyl migrations have been reported previously274 in 1,2- dihydroxy terpenoids. No such migration was ever observed in CDCl3 solutions of the natural epimers 1d and 1e. Acetylation of 37d or 37e under forcing conditions (with catalytic DMAP at 50 ◦C) gave 37c.

3.2.5 Chromatographic Identification of Epimers.

The epimers gave different colours when visualised with vanillin/H2SO4 in EtOH. After brief heating, the natural H-8β compounds slowly developed a pink/purple colour, while the 8α compounds turned blue. These colours were observed for salvinorins A–E (1a, 1b and 1c–1e)andthediol1h. Addition- ally, in each case, the 8-epimer gave a higher Rf in Et2O or EtOAc/petrol. 126 CHAPTER 3. SYNTHESIS.

1a 37a 1b37b 1h 37h

eluent: Et2OEt2O 50% EtOAc/petrol

Figure 3.2: TLC comparison of epimers using vanillin/H2SO4.

These relationships also held for many derivatives to be discussed below; thus, configuration at C-8 can be confidently inferred from TLC alone. Given the tendency for salvinorins to epimerise at C-8 under basic conditions, this infor- mation should prove useful for future synthetic work.

3.2.6 Mechanism.

Koreeda and co-workers have on several occasions23, 24, 263 proposed a complex mechanism for the epimerisation of 1b, involving cleavage of the C-8/9 bond (see Scheme 3.4 on the next page). No explanation was given for rejecting the obvious mechanism of lactone enolate formation, followed by protonation from the opposite face. Presumably Koreeda and co-workers rejected this simpler mechanism because the α-protons of ketones are more acidic than those of esters, by several orders of magnitude.275 Why, therefore, should epimerisation occur at C-8 rather than C-2 or 10? The more complex mechanism reconciles this apparent contradiction: abstraction of H-10 inverts the configuration of 3.2. EPIMERISATION AT C-8 UNDER BASIC CONDITIONS. 127

H-8 indirectly. This mechanism has been tentatively endorsed by Lee et al.271

O O O B- H+ O O O O O O H H H H RO RO RO 9 - O 10 8 O O

O O O O O O Scheme 3.4: Koreeda et al’s proposed mechanism of epimerisation.

This mechanism is not consistent with the above results. Since epimerisation of 1c occurs under identical conditions, the ketone is not essential to the process. This is also true of many derivatives of 1a lacking the ketone, which will be discussed below. Other furanolactone terpenoids undergo epimerisation at C-8 under basic conditions,268, 276, 277 despite lacking the ketone (Scheme 3.5). Several of the structures have been definitively established by X-ray crystallography: columbin (39),265 isocolumbin (40),267 palmarin (42)266 and jateorin (43).264

O O

R R R' O O R R' H H 41 8 O 8 O O O O a H H R 42 O O OH OH H H H 39 43 a a H H H 40 44 Conditions: a) ROH, OH−, ∆. Scheme 3.5: Epimerisation of related natural products with base.

Conclusive evidence against the mechanism can be found in Brown’s own re- sults, performed under Koreeda’s guidance. Refluxing 1a with KCN in CD3OD gave 45,278 deuterated at C-2, -8 and -10 (Scheme 3.6 on the following page). This establishes that exchange occurs at H-8 in a methanolic solution of CN− 128 CHAPTER 3. SYNTHESIS.

2− 275 (a weaker base than CO3 ). Thus their proposed mechanism, in which this deprotonation does not occur, is incorrect. Both the ketone and lactone are enolised; evidently, however, inversions at C-2 or -10 are thermodynamically unfavourable relative to C-8.

O O

O O O O O H H D D D O HO O 2 10 8 O KCN, CD3OD

O O 1a O O 45 Scheme 3.6: Brown’s deuteration of 1b.

3.2.7 Attempted Deacetylation under Acidic Conditions.

Given the instability of 1a under basic conditions, deacetylation under acidic conditions was investigated. In collaboration with us, Ken G. Holden treated

1a with 5% methanesulfonic acid in CH2Cl2/MeOH at room temperature. Monitoring by TLC showed initial formation of 1b and byproducts; after 2 days, these compounds were consumed, giving two barely resolved spots. NMR analysis showed that each spot consisted of two compounds; deacetyla- tion appeared to have been accompanied by lactone methanolysis (additional methoxy peaks and an upfield shift of H-12). Since this route is clearly inferior, the products were not characterised further.

3.3 Simple Derivatives.

3.3.1 Esters (46 and 47).

279 280 Treatment of 1b with HCO2H/Ac2O in pyridine gave formate 46.Treat-

281 ment in neat HCO2H proved too vigorous, giving inseparable byproducts. NMR assignments of 46 were inferred from the near-identical spectra of 1a. 3.3. SIMPLE DERIVATIVES. 129

O O O

O O O O OH O O O H H H H H H O O 1 O O 2 O O Br

O O 46 O O 47 O O 48

Figure 3.3: Ester and ether derivatives.

In the hope of obtaining crystallographic confirmation of the structure and ab- solute stereochemistry of 1h,thepara-bromobenzoate 47 was prepared using the acyl chloride, DMAP and NEt3. As found in benzoylation of 3,4-dihydro analogues,282, 24 no dibenzoylated compound was detected. Unfortunately, 47 was obtained as a powder. Recrystallisation from MeOH, and slow evaporation of a CH2Cl2/Et2O solution, failed to give crystals suitable for X-ray analysis.

3.3.2 Attempted Benzyl Ether Formation (48).

Inspired by a report that salvinorin B benzoate is a µ opioid agonist,82 for- mation of the benzyl ether 48 was attempted. Benzyl trichloroacetimidate

283, 284 and catalytic Me3SiOTf in CH2Cl2 gave a complex, inseparable mixture which did not appear to contain 48 by 1H NMR. An alternative route, benzyl

285 286 bromide with freshly prepared Ag2OinCH2Cl2, gave negligible reaction

287 after 15 hours, despite both reagents being in excess. Addition of Bu4NI gave completion in 4 hours. Again a complex mixture of benzylated compounds resulted, which appeared to contain traces of 48.

Subsequently, Béguin et al reported successful preparation of 48 using BnBr

95 and Ag2OinMeCN. This is surprising, since in previous work MeCN was the worst solvent tested, giving hydrolysis of BnBr and decreased alkylation rates.286 130 CHAPTER 3. SYNTHESIS.

3.3.3 17-Deoxy Compounds (49 and 50).

17-Deoxysalvinorin A (49) was synthesised by deoxygenation of lactol 35 using

288 Et3SiH and BF3·OEt2. The enol ether (50) was also formed as a byproduct (Scheme 3.7). To improve the yield of 49, other routes289 were explored. Use

290 of Et3SiH with Amberlyst 15 sulfonic acid resin instead gave 50 exclusively in 76% yield. Lactol 35 proved extremely prone to acid-catalysed elimination; ◦ storage overnight in CDCl3 at -20 Cgave52%yieldof50. Such dehydrations typically require much harsher conditions;291, 292 however, low-temperature de-

293 hydration of a hemiacetal with BF3·OEt2 has been reported. In rare cases, lactols may be so prone to dehydration that they cannot be isolated.294 In this case, as with epimerisation at C-8, the elimination may be driven by relief of steric interactions or strain in the natural H-8β configuration. The furan substituent may also stabilise the oxonium intermediate through electron do- nation.

1H 13C O O O HMBC

O O O O O O O O O H H H H H O O 17 O OH 8 7 6 O O O O O O

49 50 Et3SiH, BF3•OEt2, 35 48%+ 23% CH2Cl2, 0 °C

Et3SiH, Amberlyst 15 76% CH2Cl2 Scheme 3.7: Deoxygenation of lactol 35.

By 1H NMR, the H-17 oxymethylene of 49 appeared as a doublet, δ 3.58, coupling to H-8 (COSY crosspeak). As with 1a (Figure 2.14 on page 92), irradiation of H-12 gave a strong nOe enhancement of H-20 rather than H-8, 3.3. SIMPLE DERIVATIVES. 131 confirming the configuration at C-8. The quaternary C-8 peak of 50 showed HMBC correlations to H-6, -7 and -17 (Scheme 3.7). A long range coupling (1.8 Hz) was evident between one of the H-7 protons and H-17.

3.3.4 Tetrahydrosalvinorin A (51).

O O

13 H H H O O O OH O O H H H H O O O O

O O 94% O O 59%

H2, Pd/C H2 (4 atm), Rh/C 52 1a 51 MeOH MeOH/CH2Cl2

O O O O O

H2 + + cat. O O R OH R O R OH

O R' O R' O R' O O ABC D Scheme 3.8: Hydrogenation of 1a and other furanolactones.

O O O H H O O O H O H O O O H O OH O H O O OAc O O OAc O H

Limonin (53) Montanin C (54) Teucrin A (55)

Figure 3.4: Furanolactones 53, 54 and 55.

To explore the role of the furan ring in the effects of 1a, tetrahydrosalvinorin 132 CHAPTER 3. SYNTHESIS.

A(51) was prepared. Valdés had reported15, 23 that catalytic hydrogenation of 1a over palladium on carbon gave hexahydrosalvinorin A (52)innear- quantitative yield after 24 hours (Scheme 3.8). Saturation of the furan ring was accompanied by hydrogenolysis of the lactone. This is typical of the pseu- dobenzylic bond of furanolactones (A): although high yields of tetrahydro compounds C have been reported with Pd/C, hexahydro compounds D gen- erally predominate (see Table 3.2 on the facing page). In some cases, use of

Pd/BaSO4 also gave B. The highest reported yields of C have been achieved with PtO2 in acetic acid; however, results are highly variable. This is true of other catalysts: different groups report very different results for the same sub- strate under similar conditions (compare the divergent results for 53 and 54). This may be due to differences in the catalyst; for instance, some batches of Pd/C are acidic.295 Generally, acidic conditions favour hydrogenolysis, while bases (especially nitrogenous bases)296 suppress it.297 Consistent with this, hydrogenation of 1a at rtp over Pd/C in 0.1% H2SO4/MeOH gave complete conversion to 52 in 10 minutes. The substrate also affects the outcome: differ- ent compounds sometimes give dramatically different results under identical conditions (compare 55 and 54 in Ref. 298).

Rhodium on carbon is reputed172, 319, 297 to cause less hydrogenolysis than pal- ladium catalysts. There have been reports of selective reduction of furanolac- tones using this catalyst, giving C without D, albeit in low yields.317, 318 When applied to 1a, Rh/C gave negligible progress at atmospheric pressure. At 4 atm, the reaction proceeded smoothly to give 51 in unusually high yield (59%). For characterisation, the less polar 13-epimer was separated by repeated HPLC (baseline resolution was not achieved). By 1H NMR, H-12 showed a new cou- pling to H-13, but its coupling constants to H-11 were scarcely affected, sug- gesting little change of conformation in the lactone.

Determination of the configuration at the tetrahydrofuran C-13 position is challenging. This has recently been achieved for the tetrahydro derivative C of limonin (53) by comparison of nOe (ROESY) data for each epimer with predictions from molecular modelling.307 Since only one epimer of 51 was 3.3. SIMPLE DERIVATIVES. 133

Catalyst Solvent Time BC D Ref. %% % 30% Pd/C EtOAc 22 00299 10% Pd/C MeOH 24 h* 65 298(55) 24 h* 53 20 298(54) 14 78 300(54) 8h 10 76 301 35 min 689268 1h 10 84 268 1h 785268 EtOH 30 min 17 56 302 20 80 303 45 min 29 67 304 AcOH 5 h 50 305 24 h 10 59-69 306(53) 12 41 276 844276 CH2Cl2 24 h* 48 307(53) dioxane 459308 EtOAc 36 h 74 309 H2O 30 310 5% Pd/C AcOH 17 h 20 77 311(53) EtOH 6 h 251312(53) 10% Pd/BaSO4 MeOH 22 24 39 313 18 20 35 313 EtOH n.s. n.s. 314(55) AcOH 55 314 PtO2 AcOH 5 h 34 10 315 20 h 42 <55 316 7h 53 305 30 h 30 299 AcOH/dioxane 48 h 88 311(53) 96 h 86 311(53) 5% Rh/C AcOH 10 h 42 317(55) EtOAc 48 h 36 318

*=2Atm. n.s. = not stated Table 3.2: Some previously reported furanolactone hydrogenations.

obtained pure, this was not attempted in this case. Since this work, Béguin et al have synthesised 51 using Pt/C catalyst, albeit in lower yield (36%).125 134 CHAPTER 3. SYNTHESIS.

O

H H

R ent-29a H O OR ent-29b Me

Figure 3.5: (+)-Hardwickiic acid (ent-29a).

3.3.5 (+)-Hardwickiic Acid (ent-29a).

(+)-Hardwickiic acid (ent-29a) occurs in commercially-available copaiba bal- sam. The compound is difficult to separate from other acids in the crude mixture, and was therefore isolated as the methyl ester ent-29b following a published procedure.218 Cleavage of the methyl ester proved challenging. Heating in KOH/MeOH172 under reflux gave only slow decomposition, but

320 microwave irradiation on KF/Al2O3 provided ent-29a in low yield. The reaction did not proceed in the absence of KF.

3.4 Modification of the Methyl Ester.

3.4.1 Relevant Results from Previous Work.

To explore the pharmacophore of 1a, a selective modification of the C-18 methyl ester was desirable. It was apparent from previous work that hydride reduction was unsuitable. As discussed above, Valdés and Brown had achieved selective reductions of the ketone and lactone respectively (Scheme 3.1 on

i 260 page 119). Varying the solvent in the Bu2AlH reduction was of no benefit. ◦ Use of LiAlH4 at –78 C gave the triol 56 (Scheme 3.9). Reduction of the methyl ester was not observed in any of these reactions, demonstrating its very low reactivity. Béguin et al have recently reported full reduction of all

125 carbonyl groups with LiAlH4 at room temperature, giving pentaol 57. 3.4. MODIFICATION OF THE METHYL ESTER. 135

1a LiAlH Li/NH , -100 °C 4 LiAlH4 rt 3 -78 °C O O O

OH O OH OH OH OH H H H H H H HO HO OH OH O

O O 56 OH 57 OH 58

Scheme 3.9: LiAlH4 and Li/NH3 reductions of 1a.

Similarly, Brown reduced the methyl ester using lithium in ammonia,321 but this was accompanied by deoxygenation at C-2, epimerisation at C-1, and cleavage of the lactone, giving 58. Brown gives two conflicting structures for 58; the hybrid structure shown here incorporates those features consistent with the 1H NMR data. The furan peaks downfield of 6 ppm rule out saturation of the furan ring. The C-18 oxymethylene δ and J values are nearly identical to those of 18-hydroxy derivative 77 (see Scheme 3.17 on page 152), ruling out epimerisation at C-4.

Low reactivity at the methyl ester was also evident under other conditions.

As noted earlier, KCN/CD3OD gave deuteration at C-2, -8 and -10, but not C-4, α to the methyl ester (see Scheme 3.6 on page 128). Clearly none of these approaches offered the possibility of a selective reaction at C-18. As an alternative approach, cleavage of the C-18 methyl ester to the acid would permit selective borane reduction to the hydroxyl, as in 57 and 58,inthe presence of other carbonyls.

3.4.2 Treatment of Salvinorin A with KOH in MeOH.

The first route attempted for methyl ester cleavage was basic hydrolysis. Since the methyl ester was unaffected by Na2CO3 in MeOH, 1a was treated with 1M KOH in MeOH . The solution turned a deep orange, and the starting material was rapidly consumed. The major product, found in the neutral fraction, 136 CHAPTER 3. SYNTHESIS. was enedione 59 (Scheme 3.10). The base-soluble fraction was difficult to analyze, smearing on TLC and giving a poorly-resolved 1H NMR spectrum. Surprisingly, at least eight peaks were apparent in the methoxy region. After

322 1 methylation with Me3SiCHN2, TLC showed only a single spot. HNMR analysis, however, revealed three major compounds (60a, 60b and 60c), which were separated with difficulty by HPLC. Baseline resolution was not achieved, necessitating repeated repurification and poor recoveries.

3.4.2.1 Structure Elucidation of the Products.

O O O

O R R' O O OH O + O O H H 1. KOH / MeOH H R R' O O H H O O O 10 8 O 60a

2. Me3SiCHN2 O H H 60b O O O O O 47% O O 53% H H 1a 59 60c Scheme 3.10: Autoxidation of 1a.

The 1H spectrum of 59 showed two new singlets at δ 6.91 and 6.99 ppm. The peak at 6.91 showed no COSY or HMQC crosspeak, and exchanged with D2O. Such strongly deshielded exchangeable peaks are typical of cyclic α-diones, whose enol tautomers are stabilised by internal H-bonds.323 Consistent with this, the compound exhibited strong IR absorptions at 3373 and 1651 cm−1 (OH and enol C=C). The structure was further elucidated by analysis of the HMBC spectrum. The quaternary C-10 peak, located unambiguously by its correlations to the H-19 and 20 methyls, showed a correlation to the enolic proton, placing the enol at C-1 and the ketone at C-2. The vinylic H-3 peak showed the expected correlations to C-1, 4, 5 and 18 (see Figure 3.6). UV absorptions at 215, 249 and 324 nm confirmed an extended π system, as in methyl 4-oxopentenoate (61, Figure 3.7 – 222 and 324 nm).324 Compare the spectra of 1a and 1c. Note that the longest-wavelength peak approaches vis- ible wavelengths. The absorptions of α-diones exhibit a bathochromic shift 3.4. MODIFICATION OF THE METHYL ESTER. 137 towards visible wavelengths in basic solution:325, 326, 327 the orange colour ob- served during formation of 59 in basic MeOH is consistent with this.

O O 1H 13C

H O O O O 1 H H H O 10 2 O O O 20 5 3 O H 4 19 18 O H 60a O O 59 O O

Figure 3.6: Key HMBC correlations of 59 and 60a.

HRESIMS confirmed the molecular formula. The remainder of the structure, unchanged from the salvinorins, was fully elucidated and assigned by NMR experiments (DEPT, COSY, HMQC, HMBC, and nOe). The H-12 coupling constants were closer to those of 37a than of 1a, suggesting that epimerisation at C-8 might have occurred. However, the β configuration of H-8 was evidenced by a trans-diaxial coupling constant (9.7 Hz); in addition, irradiation of H-12 gave an nOe enhancement of H-20. The structure of 59 thus established is remarkably similar to salvinorin G (1g, Figure 2.51 on page 118).137

The major product from the base-soluble fraction was identified as 1,2-secotriester 60a (Figure 3.6) based on extensive NMR experiments. The H-10 singlet showed an HMBC correlation to the new C-1 ester carbonyl (Figure 3.6). H-4 showedcorrelationstoC-3,5and10,andformedanisolatedspinsystemwith the two deshielded H-3 peaks. This confirmed the location of the new methyl esters at C-1 and 2, although the three esters were not sufficiently resolved in the 2D spectra to allow individual assignment. The remaining NMR data was very similar to 1a. The chemical shift and coupling constants of H-12 were near-identical to those of 1a, confirming the configuration at C-8. HRESIMS confirmed the molecular formula. The second major product was identified as the 8-epimer 60b based on the shifts and coupling constants of H-8 and 12 (near-identical to those of 37a).270 Assignment of the remaining data was straightforward. 2D NMR showed the same correlations as 60a;HRESIMS 138 CHAPTER 3. SYNTHESIS.

20000 O O

18000 O O O 16000 H H O O O 59 O O 14000 O 61 O O O 12000 H H O O O O O

ε 10000 1c 1a OH O H 8000 O O O O 6000 1c 59 4000 1a O O 2000

0 200 220 240 260 280 300 320 340 360 380 400 λ (nM)

Figure 3.7: UV/Visible spectra of 1a, 1c and 59 in MeCN. again confirmed the molecular formula. Interestingly, although 60a and 60b cospotted by TLC, they gave the expected colours with vanillin/H2SO4 in EtOH (Figure 3.2 on page 126).

The third (minor) compound decomposed in CDCl3 before characterisation was completed, but was tentatively assigned as 60c (Scheme 3.10 on page 136). HRESIMS established that the compound was also an isomer of 60a,andthe appearance of the same couplings in the COSY spectrum suggested another stereoisomer. The coupling constants of H-8 and 12 established that the C- ring configurations matched those of 60a. Indeed, all of the coupling constants determined were close to those in 60a, whereas many chemical shifts showed large changes. This implied a change in the electronic environment of the coupling protons, without a change in their configuration. The most plausible candidate structure was therefore the 10-epimer 60c, since H-10 is not coupled. Placing a large substituent in an axial configuration would be expected to affect the conformation of both remaining rings, and hence the chemical shifts around those rings. By contrast, inversion at C-4 would not be expected 3.4. MODIFICATION OF THE METHYL ESTER. 139 to dramatically alter the conformations of the rings, but would be expected to alter the coupling constants with the H-3 protons. These couplings were scarcely changed, while the chemical shifts of H-4, 7, 8, 10, 11, and 12 (but not H-3) were dramatically altered. Thus 60c is the more plausible structure. The particularly large change at H-4, shifted downfield by 0.83 ppm, might be due to falling within the deshielding region of the C-1 carbonyl. Also, as mentioned earlier, H-10 is much more readily exchangeable than H-4. Brown reported278 that when 1a was refluxed with KCN in CD3OD, deuterium exchange occurred at H-2, 8 and 10 but not H-4 (Scheme 3.6 on page 128). In the absence of nOe data, however, the proposed structure 60c must remain tentative.

3.4.2.2 Comparison with Previous Reports.

These results conflict with those published previously. Tidgewell et al169 re- ported that heating 1a with NaOH in MeOH caused cleavage of the methyl ester and opening of the lactone, without giving further detail. Since lactone hydrolysis would be reversed upon neutralisation, this presumably refers to methanolysis. Ester cleavage may have been inferred from the formation of a base-soluble fraction, and lactone methanolysis from the methoxy peaks in the 1H NMR spectrum. As shown above, however, the acidic fraction and its methoxy peaks result from cleavage of the α-hydroxy ketone. When the 1,2-diol 36h23 (Figure 3.8) was refluxed in KOH/MeOH for 30 min, an epimeric mix- ture at C-8 was recovered in near-quantitative yield, confirming that neither methyl ester hydrolysis nor lactone methanolysis occur under these conditions.

O O

OH O O H H H HO O O O

O O 36h O O 62

Figure 3.8: Diol 36h andproposedautoxidationproduct62. 140 CHAPTER 3. SYNTHESIS.

More recently, Lee et al treated 1a with Ba(OH)2 in MeOH, reportedly obtain- ing 62 in 75% yield.271 The 1Hand13CNMRdataquotedfor62 are identical to those of 59, apart from the omission of the H-6 multiplet at δ 1.77-1.67 ppm; they are evidently the same compound. Their proposed structure 62 is not consistent with the additional data presented here. Specifically, HRMS es- tablished the molecular formula as C21H22O7 (59)ratherthanC21H22O6 (62). Further, the singlet at 6.91 cannot be attached to C-1, since it exchanges with

D2O, has no HMQC crosspeak, and lacks the expected HMBC correlations to C-3, 5, and 9. A corresponding methine peak is also absent from the DEPT spectra. The red colour of the reaction mixture125 again suggests the distinc- tive red shift of α-diones. The presence of the C-1 ketone was confirmed by reduction (Section 3.4.2.5 below). Finally, deoxygenation of a ketone would not be expected under these conditions.

3.4.2.3 Proposed Mechanism.

Autoxidation of α-hydroxy ketones (“acyloins” or “α-ketols”) to α-diones (“diosphe- nols”) under basic conditions is well-established.325, 328, 329 The reaction con-

329 sumes one equivalent of O2, generating H2O2. While the autoxidation of unsubstituted ketones requires stronger bases such as tBuOK, α-hydroxy ke- tones are more readily enolised, and the reaction proceeds with KOH.328 A proposed mechanism, via saturated dione 63, is shown in Scheme 3.11.

The alternate pathway, to seco-diester 64, also has numerous precedents.328, 330 We based our proposed mechanism on the generally-accepted formation of hy- droperoxide intermediates,328 although this mechanism has been disputed.330 Tautomerisation of the enolate or radical will give the regioisomeric diester, which along with epimerisation at C-8 and 10 explains the numerous methoxy peaks in the 1H NMR spectrum of the crude product.

Dehydrogenation of 63 to form 59 is unusual. While there have been sev- eral reports of dehydrogenation of 1,4-diones in alcoholic KOH,331 literature searches48 revealed no such reaction involving a 4-ketoester. However, α-diones 3.4. MODIFICATION OF THE METHYL ESTER. 141

O O O

O O - O O O O 2 O O H H H H H H O -OH HO HO O O O2 O

O O 1a O O O O

- MeOH MeOH/MeO

O O O

O O O O O O O H H HOO H H H H O HO O O HO O HO O

-OH O O 63 O O O O

- 1) OH 2) O2

O O 59 O

MeOH O O O O O O H H H H H H O O O O O O HO HOO O2 H O O O O O O O 64 -OH Scheme 3.11: Proposed mechanism of the autoxidation.

O O O

O O O O O O H H H H H H HO CrO O 3 O O -H2O O O

pyridine HO

O O 1b O O 65 O O 66 Scheme 3.12: Unexpected oxidation product 65. are much more readily enolised than unsubstituted ketones; in the case of 59, no trace of the 1-keto tautomer was detectable by NMR. It is therefore plausi- ble that 63 should be extremely reactive, and unsurprising that this compound 142 CHAPTER 3. SYNTHESIS. was not isolated. Consistent with this, Brown’s attempts to prepare 63 via PCC oxidation of 1b gave no isolable product.332 More recently, Harding et al oxidised 1b using similar conditions (CrO3 in pyridine), unexpectedly ob- taining the lactone 65136 (Scheme 3.12) with the loss of C-2, rather than the expected 63. No mechanism was proposed. It is interesting to note that lactols analogous to 66 have been reported from autoxidation328, 333 or ozonolysis with oxidative workup334 of α-diones, and that dehydration of 66 would give 65.

However, if 66 were formed in the presence of excess CrO3, oxidation rather than elimination would be expected.

3.4.2.4 Variation of Reaction Conditions.

The yield and selectivity of autoxidations of this type can be subject to strong solvent effects.328 The reaction was therefore repeated in EtOH, iPrOH and tBuOH. The resulting neutral and acidic fractions were noticeably more com- plex.Ineachcase,59 was contaminated by inseparable impurities (presumably including the 8-epimer); the original selection of MeOH thus proved fortuitous. The reaction proceeded with only traces of oxygen, for instance when per-

328, 330 formed under N2 in MeOH pre-saturated with N2. This is again typical. Nonetheless, the reaction was faster and more consistent when the solution was saturated with O2. Anther useful refinement was the use of dilute KOH rather than NaHCO3 to extract the extremely hydrophobic acid diesters during workup.

3.4.2.5 Attempted Reductions of 59.

The unexpected installation of the 3,4-double bond in 59 suggested the com- pound might provide a route to diol 1h. However, reduction with NaBH4 in EtOH/CH2Cl2 was accompanied by conjugate addition and epimerisation, giving 8-epi-diol 38h27 in low yield (Scheme 3.13). Attempted Luche reduc-

335 336 tion with NaBH4–CeCl3 in MeOH (with or without sonication) was also unsuccessful, giving a complex mixture whose unstable major components re- 3.4. MODIFICATION OF THE METHYL ESTER. 143

O O

OH O OH O H H H H HO HO O O

O O O O NaBH4 NaBH4 38h 59 1h CeCl3 X Scheme 3.13: Attempted reductions of 59. tained the characteristic enedione peaks at δ 6.8 and 7.0 ppm. These products cospotted with the starting material in petrol-based systems, making the reac- tion difficult to follow, but were resolved by acetone/CH2Cl2 and gave a darker purple with vanillin. There are precedents for ketones which are smoothly re-

335 duced by NaBH4 alone, but not in the presence of CeCl3. Note that use

337 of “forcing conditions” (reflux for 12 hours) was futile. NaBH4 decomposes to unreactive tetramethoxyborate within 5 minutes under the standard condi-

336 tions (with CeCl3 in MeOH at room temperature). Prolonging or heating the reaction will therefore achieve nothing. Continued reaction would require frequent additions of fresh reagent.

Commonly used drying procedures for CeCl3 can cause partial decomposi-

338 tion. The quality of the CeCl3 used was verified by reduction of 2-cyclohexen- one,336 which gave 2-cyclohexenol rapidly and quantitatively.

The enols of α-diones form complexes with a wide variety of metal salts such

339 as FeCl3; it appears likely that such an enolic complex forms with 59 in preference to the desired ketone–solvent–Ce3+ complex,336 and the reaction therefore does not follow the desired course. The only successful precedent located for this reduction involved non-enolisable enediones.340 It might be possible to prevent this problem by protection of the enol, but an attempt at

Et3Si protection was unsuccessful. 144 CHAPTER 3. SYNTHESIS.

3.4.3 O-Demethylsalvinorin A (67a).

To summarise the above results, previous work showed the C-18 methyl ester to be the least reactive carboxyl of 1a under a variety of conditions. The few reactions vigorous enough to attack this position gave undesired side reac- tions at other positions; no selective transformation had been reported. The preliminary results above were consistent with this.

The selective cleavage of methyl esters has been the topic of extensive research; thorough reviews are available.341, 342 The most effective procedures involve nucleophilic substitution at the alkoxy (or carbinol) carbon, via an SN2mech- anism, with the carboxylate as leaving group. This mechanism (BAl2ester cleavage) contrasts with the BAc2 mechanism typical of basic hydrolysis, in- volving attack at the carboxyl carbon (Scheme 3.14).343 The tendencies of different nucleophiles to act via these mechanisms can be understood in terms of hard/soft acid-base theory.342 Soft nucleophiles such as I− attack the soft electrophilic carbinol via the BAl2 mechanism, while hard nucleophiles such − as F favour attack at the hard electrophilic carboxyl carbon, via the BAc2 mechanism. Thus soft nucleophiles are less affected by hindrance around the carboxyl carbon, but more affected by hindered alkyl chains; hence the se- lectivity for methyl esters. Given the severe hindrance exhibited by the C-18 carboxyl carbon of 1a, the desirability of methyl ester selectivity, and the dis- astrous effects of hydroxide, the use of soft nucleophiles was clearly preferable.

Nu- BAl2 O H O + H - CH3Nu R O H R O

BAc2 O -O O O + CH OH R R R - 3 O OH O -OH

Scheme 3.14: BAl2andBAc2 ester cleavage mechanisms. 3.4. MODIFICATION OF THE METHYL ESTER. 145

3.4.3.1 Cleavage with Iodide and Cyanide Reagents.

Perhaps the most widely used source for a soft nucleophile is LiI. Early re- ports used pyridine and 2,6-lutidine344 as solvent, but subsequent work found more polar non-hydrogen bond donor solvents such as DMF345 and especially

346 HMPA (Figure 3.10) to be superior, as would be expected for an SN2 reac- tion.347 Anhydrous LiI has been reported to give greater selectivity for methyl esters, but lower yield, than the hydrate.348

O O O

O O O O O O O O H H H H H H O O HO O O O

O O 1a O OH 67a O OH 67b

Figure 3.9: O-Demethyl salvinorins A and B.

An initial trial of anhydrous LiI in dry DMF, after refluxing for 24 h, gave a 29% yield of the epimerised acid 67a. The dark brown colour and fishy odour of the reaction mixture suggested decomposition. The neutral fraction was a complex mixture containing little starting material.

Adding sodium acetate has been reported to lower the required temperature and reaction time for this procedure.349 However, in this case the reaction showed no progress after 6 hours at 130 ◦C, and the yield after refluxing overnight was not improved.

O O O N P N NN N N N

2,6-lutidine DMF DMPU HMPA

Figure 3.10: Useful non-hydrogen bond donor solvents.

The use of HMPA was avoided due to its carcinogenicity,347 but work with other nucleophiles has found that the safe substitute DMPU350 (Figure 3.10) 146 CHAPTER 3. SYNTHESIS. gives comparable rate enhancements.351, 352 Substituting this solvent for DMF (150 ◦C for 25 h) gave no improvement in absolute yield (27%), but a much cleaner neutral fraction of epimerised 1a, giving a 79% yield based on recovered starting material. The reaction mixture again showed a dark colour and a fishy smell suggestive of decomposition.

A run at a higher temperature (190 ◦C) gave a different base-soluble product which was not identified. The NMR spectrum showed an epimeric mixture lacking the acetyl peaks of 67a.Itdidnotmatch67b, however, and was not acetylated by Ac2O/DMAP in pyridine.

The use of NaI in DMPU at 150 ◦C for 23 h gave a light-coloured reaction mixture and a greatly improved yield of 67a (73%). This was surprising, since lithium has been found to be superior to sodium as a counterion with iodide345 and other353 nucleophiles. Unfortunately, two careful attempts to replicate the reaction gave the same yield as LiI.

Sodium cyanide has been reported to give superior results to lithium halides, including LiI, especially in HMPA.346, 343 Treatment of 1a with KCN in DMPU at 90 ◦C for 27 h, however, gave a complex mixture with only traces of 67a. With NaCN the reaction proceeded more slowly, but again gave a complex mixture after consumption of starting material (70 h).

3.4.3.2 Cleavage with Thiolates.

Thiolates (RS−) are another group of soft nucleophiles commonly used for methyl ester cleavage. A recent study reported excellent results using arylthiols and catalytic base at 190 ◦C, generating thiolate in situ.351 Attempts to apply this procedure were unsuccessful: treatment of 1a with 4-methylbenzenethiol with K2CO3 in DMPU gave no reaction after 10 minutes. Prolonged reaction (16 h) gave decomposition, with no indication of acid formation.

Aliphatic thiolates have found greater use. Lithium methanethiolate (LiSMe),354 as well as the ethyl355 and propyl353 homologues all cleave methyl esters at room temperature, whereas halide and cyanide reagents are typically used above 100 3.4. MODIFICATION OF THE METHYL ESTER. 147

◦C. Again lithium353 appears to be a superior counterion to sodium,356 and HMPA353 and DMPU352 superior solvents to DMF.

Alkylthiolate solutions are unstable.353, 355 However, LiSMe is easily prepared and air-stable;354 ethanethiolates are now commercially available. For this work, LiSEt was chosen since ethanethiol (bp 35 ◦C) is much less volatile than methanethiol (bp 6 ◦C). The sickening thiol odour is therefore greatly reduced. The salt was prepared from ethanethiol and nBuLi using a simplified version of published procedures for LiSMe354 and LiSPr.357 Again, DMPU was substituted for the carcinogen HMPA.

O O

O O O O O H H H H HO Ac2O, DMAP O O O DMPU

O OH 67b O O 68

O Scheme 3.15: Formation of mixed anhydride 68.

The reaction between 2a and LiSEt showed little progress after 6 hours at room temperature. After heating to 55 ◦C for 23 hours, the reaction went to completion, with little change in colour. This was in marked contrast to the high-temperature procedures. Ester cleavage was accompanied by deacetyla- tion, giving 67b; standard acetylation conditions gave the epimeric acids 67a in good yield (typically 73% over two steps).

On one occasion the acetylation was performed in one pot, after quenching the thiolate with acetic acid (the orange colour faded to faint yellow). Addition of Ac2O/DMAP gave 67a in four hours, unfortunately accompanied by the mixed anhydride 68 (crude 1H NMR of the neutral fraction showed additional characteristic peaks at δ 2.24 and 2.25 ppm; Scheme 3.15). The anhydrides proved remarkably stable. Some starting material remained after reflux in

THF/H2O for 1 hour; addition of NaHCO3 was necessary for complete hydrol- 148 CHAPTER 3. SYNTHESIS. ysis to 67a. This process was more laborious than a separate acetylation step, and was therefore abandoned.

The results of the various cleavage attempts are summarised in Table 3.3. Yield of 67a Nu Base Solvent T time Isolated borsm* ◦Ch%% LiI - DMF 150 24 29 LiI NaOAc 130-150 21 26 LiI - DMPU 150 25 27 79 LiI - 190 24 0 NaI - 150 14 23 NaCN - 90 70 0 KCN - 90 27 0 p-MePhSH cat. K2CO3 200 16 0 LiSEt - 55 23 73 (after acetylation) * = based on recovered starting material Table 3.3: Summary of results - nucleophilic cleavage of 1a methyl ester.

3.4.3.3 Confirmation of Structure of 67a.

The mixed acids 67a streaked on TLC, and reduction by BH3 was also ex- pected to cause epimerisation, so chromatographic separation was not at- tempted. The structure of 67a was definitively confirmed by methylation with 1 CH2N2, giving a mixture of 1a and 37a by H NMR. The products cospotted with authentic material in both petrol– and CH2Cl2–based TLC systems and gave identical colours with vanillin dip.

The 1H NMR spectrum of 67a lacked the methoxy peak, but was otherwise near-identical to a mixture of 1a and 37a. One interesting difference was the first-order H-4 multiplet at 2.80 (dd, J = 5.2, 3.5). The H-4 multiplet of 1a is slightly non-first order due to the almost coinciding H-3 peaks, but at 800 MHz can be approximated as 2.74 (dd, J = 11.3, 5.6). Apparently the formation of a carboxylic acid dimer (as with hardwickiic acid, Figure 2.18) pushes the C-4/18 bond away from the equatorial position, altering the conformation of the A ring. Thus the trans-diaxial coupling constant for H-4 is lowered, and the H-3 multiplets are no longer coincident. 3.4. MODIFICATION OF THE METHYL ESTER. 149

3.4.3.4 Subsequent Developments and Discussion.

Since this work, Me3SnOH has been reported to cleave methyl esters selec- tively under mild, neutral conditions in complex substrates too sensitive for previous methods.358 While the method appears very promising, it has several drawbacks. Firstly, the compound is extremely neurotoxic.359 Secondly, like other oxytin reagents, Me3SnOH is a hard nucleophile, attacking the carboxyl carbon.360 The reactivity of other oxytin reagents is greatly reduced against hindered carboxyls, including terpenoids resembling 1a.360

Also since the publication of the above work,270 Lee et al have reported an improved yield of 1a via the LiI method (56%125 or 72%),361 by refluxing in pyridine for 36 hours. The LiI was presumably a hydrate since no drying is mentioned. If the higher yield proves reproducible, this method offers clear advantages over the thiolate route above: the acetate remains intact, and the reagents are common, stable and odourless. This result is surprising in light of the early results on solvent effects discussed in Section 3.4.3.1 on page 145.

By the same token, the alkylthiolate route has not been optimised. This route offers the inherent advantage of proceeding at or near room temperature.

Given the accumulation of the basic carboxylate during BAl2 ester cleavage, and the base-sensitivity of 1a, this should permit higher maximum yields.

The deacetylation observed with LiSEt is puzzling, since alkylthiolates have been found previously to cause less deacetylation than LiI.353 One potential explanation is contamination, since the thiolate was not purified before use. On several occasions after the cleavage and acetylation, 2-hydroxyethyl ac- etate362 was unexpectedly isolated. If one of the reagents used (presumably the thiol) was contaminated with ethanediol, the resulting thiolate would be contaminated with the dihydroxide. Also, any LiOH present in the nBuLi used (from exposure to moisture) would also have been present in the thiolate. The nBuLi solution used was labelled 2.5 M, but titration363 gave a value of 2.1 M. Finally, as discussed in Section 3.2.3 on page 124, the commercial DMPU used subsequently proved to contain an impurity capable of causing epimerisation. 150 CHAPTER 3. SYNTHESIS.

However, this was not accompanied by deacetylation, even at 150 ◦C. One of these possible contaminants may have been responsible for the deacetylation observed, and may also have lowered the yield. The use of demonstrably pure solvents and reagents may improve the outcome of this reaction.

3.4.4 O-Demethyl-18-deoxysalvinorin A (77).

3.4.4.1 Borane Reduction of 67a.

With the acid 67a in hand, the planned borane reduction could be attempted. ◦ BH3·THF reduces carboxylic acids rapidly at low temperatures (0 Corbelow in many cases), even when hindered (69, Scheme 3.16).364 Esters and lactones are generally reduced at a much lower rate,365 allowing mild, selective reduc- tions of polyfunctional acids such as 70366 and 71.367 However, some acids are much less reactive. For example, many aromatic acids such as 72368 re-

369 quire excess BH3, longer reaction times and higher temperatures. In some

370 extreme cases, refluxing for several hours with excess BH3 is required (73 and 74).371

Conversely, some lactones (75 and 76) are rapidly reduced to lactols under mild conditions, and even to cyclic ethers and acyclic diols with prolonged reaction.372 Thus, the selectivity achieved depends on the substrate and precise conditions. An excellent review is available.365

Slightly different protocols have been used, but generally borane is added drop- wise at low temperature, then the mixture is warmed gradually to the desired reaction temperature.364, 373 During the initial addition, deprotonation of the

374 acid gives intermediate acyloxyboranes (with visible evolution of H2), which are reduced much faster than other functionalities, accounting for the selectiv- ity of the reagent. The low temperature and slow addition of BH3 minimises side reactions until these intermediates are formed.

Unfortunately, attempts to apply the mildest possible protocol to 67a were ◦ unsuccessful. Addition of BH3·THF at -25 C gave no visible evolution of H2. 3.4. MODIFICATION OF THE METHYL ESTER. 151

X OH O X OH O

O O OH Br O O X 69 70 71 1.3 eq, 0°C to rt, 1h, 95% 1 eq, 0 °C, 3h, 86% 1.3 eq, 0°C, 30 min, 66%

O O X O X X O O X OH BH3•THF HO OH H2

HO X 73 72 4 eq, reflux, 4h, 94% 3.5 eq, rt, 48h, 71% X OR OR OH HO OR OR X 4 eq, reflux, 3h, 77% 74 (R= Me/Et)

H OH H H H

H H X O 75 X O 76 H

X O 1 eq, rt, 30 min, 80% HO H2 3 eq, rt, 3 days, 55%

Scheme 3.16: Some previously reported BH3·THF reductions.

Gradual warming to 0 ◦C and addition of excess borane had no effect. Warming to room temperature, then 45 ◦C, over several days gave no reaction. Starting material was recovered almost quantitatively. In later trials, borane was added at room temperature, giving the first visible signs of H2 evolution. The actual 152 CHAPTER 3. SYNTHESIS. reduction, however, required a large excess of borane. Small amounts of the 8-epi-alcohol 78 were detected (Scheme 3.17), but not the desired alcohol 77. Refluxing with excess borane gave 78 accompanied by unidentified byproducts. Therefore a smaller excess was tried at intermediate temperature. The most successful conditions are shown in Scheme 3.17, giving the epimeric products in 48% total yield.

O O 1H 1H COSY O O O O O O H H H R O O O O R BH3•THF (1.3 eq) H O OH dropwise, rt x 5 min, OH 23% 77 then 55 °C x 90 min H 67a 25% 78 Scheme 3.17: Borane reduction of 67a.

These conditions are not optimised. None of the target 77 was isolated until the last few experiments, when it was found to cospot with the starting mate- rial in the TLC system used previously (1% AcOH in 20% acetone/CH2Cl2). Thus, in earlier experiments, clean formation of the desired product may have gone undetected, while decomposition of that product may have been mis- taken for consumption of starting material. An alternate solvent system, 1%

NEt3/EtOAc, resolved all compounds.

Earlier trials may have therefore been more successful than they appeared, and milder conditions warrant reinvestigation. Nonetheless, it is apparent from the lack of hydrogen evolution at 0 ◦Cthat67a is an exceptionally unreactive substrate.374 This is unsurprising given the near-inertness of the C-18 position to all other procedures tried - hydride reagents, deuteration, strong base and soft nucleophiles. On the other hand, given the side reactions apparent at reflux with BH3, 67a is also a sensitive substrate. It is evident that the optimal conditions will be intermediate between the mild, original protocols364, 373 and the harsh conditions required for stubborn, resilient substrates such as 74. 3.5. MODIFICATION OF THE KETONE. 153

3.4.4.2 Structure Elucidation of 77.

Although the H-4 multiplet of 77 (δ 1.89 ppm) was too complex for determi- nation of coupling constants, the COSY spectrum showed cross peaks with the diastereotopic H-18 oxymethylene (δ 3.94 & 3.49 ppm). The chemical shifts and coupling constants for H-8 and H-12 were near-identical to those of 1a, and irradiation of H-12 gave a strong nOe enhancement of H-20 rather than H-8 (compare Figure 2.14 on page 92), confirming the configuration at C-8. A strong infrared absorption at 3468 cm−1 confirmed the presence of a hydroxyl. HRMS confirmed the molecular formula.

3.5 Modification of the Ketone.

Selective reduction of the C-1 ketone in 1a had already been reported (36e).15, 23 Since the 1α-hydroxy of this compound was axial, however, any change in bi- ological activity would be ambiguous; attributable either to the loss of the ketone, or the presence of a new protruding H-bond donor. Other modifica- tions of the ketone were therefore investigated. One obvious alternative was deoxygenation. Methylenation was also attempted, since this would preserve the sp2 hybridisation of C-1 and thus have less effect on the conformation of ring A, albeit with a much larger substituent.

3.5.1 Attempted Methylenation.

Initially, Wittig olefination was attempted. Treatment of 1a with the ylide

375 n ◦ formed from Ph3PMeBr and BuLi (1.6 eq, THF, 35 C, 20 hours) gave only epimerisation and partial deacetylation. No trace of the desired methylenated compound 79 was detected.

The Wittig reaction often fails with hindered substrates, and epimerisation is also common due to the basicity of the ylide.376 The Tebbe reagent and related titanium compounds377 allow the methylenation of some carbonyl compounds 154 CHAPTER 3. SYNTHESIS.

O

O O H H O O

O O 79

Figure 3.11: Methylenated target compound 79.

for which Wittig conditions fail. Treatment of 1a with Zn-CH2Br2-TiCl4

378 reagent (room temperature, CH2Cl2, 1 hour) gave an extremely viscous, black reaction mixture. After standard workup, however, only starting mate- rial was detectable (72% recovery). The reagent was not tested for activity against a known substrate, but was freshly prepared379 and had the reported grey colour and thick consistency. On addition to water, the reagent blackened and effervesced vigorously.

3.5.2 Attempted Direct Deoxygenation.

O S O NH O N TsNHNH2 NaBH4 RR RR RR Scheme 3.18: Ketone deoxygenation via a tosylhydrazone.

The next modification of the ketone to be attempted was deoxygenation. The Clemmensen reduction (Zn/HCl) was not suitable, since α-acetoxy groups are eliminated,380 even under the mildest conditions.381 An alternative mild ap- proach is via a tosylhydrazone, which can be reduced to the hydrocarbon by hydride reagents (Scheme 3.18).382, 383 In attempts to form the tosylhydrazone, 1a was treated with tosylhydrazide384 under a variety of conditions (Table 3.4). Since no reaction occurred in solution (even with sonication), microwave irradi- 3.5. MODIFICATION OF THE KETONE. 155 ation of the neat reagents was attempted. Under normal conditions, this is not effective for ketones.385 However, some ketones react readily when the flask is supported in an alumina bath, generating much higher temperatures.386 Since these conditions proved ineffective, a procedure for oxime formation on silica gel387 was adapted. Basic and acidic alumina were also tried as substitutes. Unreacted starting material was recovered in all cases. It is interesting to note that 1a is stable at room temperature in acetic acid (compare formic acid, Section 3.3.1 on page 128).

Solvent Catalyst Conditions Time Ref. THF - reflux 6 h 388 AcOH - rt 18 h 389 AcOH - ultrasound, rt 2 h 390 EtOH basic Al2O3 reflux 3 h 391 - - microwave, rt to 115 ◦C5min 386 - silica gel microwave, rt to 115 ◦C5min 387 ◦ -basicAl2O3 microwave, rt to 115 C5min 387 ◦ -acidicAl2O3 microwave, rt to 115 C5min 387 Ultrasound: 40 kHz, 50 W transmitted. Microwave: 1400 W (700 W output). Table 3.4: Unsuccessful treatment of 1a with excess tosylhydrazide.

The tosylhydrazide used was prepared by a published procedure,384 omitting the optional recrystallisation, which caused decomposition. Identity and purity (especially the absence of the typical384 impurity ditosylhydrazide)392 were confirmed by TLC and 1HNMR.393

No other promising methods of tosylhydrazone formation were located. El- Sayed’s recent review of sulfonohydrazides,394 although published in 2004, ap- pears to have been written decades earlier. The only post-1971 references cited are the author’s own, and these are cited only in the final sentence.

3.5.3 Indirect Deoxygenation.

3.5.3.1 Formation of Cyclic Thionocarbonate (80).

Given the failure of the direct reductions, deoxygenation of 36h was at- tempted. The usual approaches to hydroxyl deoxygenation,395 involving either 156 CHAPTER 3. SYNTHESIS.

O O S S OH O N N O O H H N N H H HO O O O DMF, 90 °C

O O 36h O O 80

67% from 1a Scheme 3.19: Formation of cyclic thionocarbonate 80.

hydride reduction of sulfonate derivatives or radical reduction of thiocarbonyl derivatives,396 would require derivatisation of the extremely hindered and un- reactive 1α-hydroxyl group. While direct acetylation has been achieved using

82 NEt3, this is ineffective for benzoylation (as seen with the less-hindered 1h, Section 3.3.1 on page 128). 1b has been mesylated at C-2 under mild condi- tions;82 however, the yield was very low (32%), and C-1 is much more hindered.

Thus, direct installation of hindered functionalities at C-1 is likely to be dif- ficult. Given that borohydride reduction of 1a proceeds exclusively from one face, and sluggishly, the reduction step is also likely to be challenging. Of the two approaches, ionic versus radical reduction, radical reductions are less susceptible to hindrance.396 Inspired by the indirect diacetylation of 36h via the 1,2-orthoacetate (Section 3.1),27 radical deoxygenation of a cyclic thiono- carbonate was attempted.

Treatment of 36h with 1,1’-thiocarbonyldiimidazole (Scheme 3.19) in DMF gave 80 in high yield. Since this gave inseparable 8-epimers, the reaction was subsequently performed in two steps from 1a without separation of interme- diate 36h and 38h.By1H NMR, the H-1 and -2 signals of 80 were shifted downfield, and the characteristic 13C peak (191 ppm) of cyclic thionocarbon- ates was present. 3.5. MODIFICATION OF THE KETONE. 157

3.5.3.2 Unsuccessful Radical Deoxygenation Attempts.

Although radical reductions have traditionally been performed with organos- tannanes, these compounds are highly toxic, and purification of the products tends to be difficult.397 This has led to extensive and fruitful research into al- ternatives, such as silanes and phosphites.396 Perhaps the most promising396 of these alternatives is hypophosphorous acid, H3PO2. Typically, this is buffered

398 with NEt3, and AIBN serves as radical chain initiator. However, other initiators and bases have been successfully substituted.399 The standard396 conditions of excess H3PO2 and NEt3 in refluxing 1,4-dioxane were used; how- ever, benzoyl peroxide was substituted for AIBN because the latter was not

396 at hand. (BzO)2 gives superior results to AIBN with alkyl phosphites. De- spite adding seven portions of (BzO)2 over five hours at reflux, most starting material was recovered (70%), along with 15% deprotected diols 36h/38h.No deoxygenated products were detected.

One report used intriguingly simple conditions: magnesium in methanol.400 Although the method had only been proven for cyclic thionocarbonates of 2,3-dihydroxy esters, its extreme simplicity and nontoxicity were attractive. Treatment of 80 with excess Mg turnings in MeOH at reflux for 40 minutes gave no reaction. Addition of activated401 Mg turnings had no effect after a further two hours’ reflux. Starting material was recovered (94%).

n 3.5.3.3 Radical Deoxygenation using Bu3SnH.

n The traditional Bu3SnH/AIBN route was then attempted. This is the best-- established route (especially for cyclic thionocarbonates),402, 403, 404 but often gives side reactions,405, 406 sometimes to the exclusion of the desired deoxy products.407 An indispensible paper by Kanemitsu et al thoroughly describes the theory and practice of the procedure, including control of side reactions.405

n 408 The Bu3SnH used was prepared by a published procedure with reduced reaction time (10 minutes),409 distilled and stored in darkness under argon at 158 CHAPTER 3. SYNTHESIS.

O O O

O O OH O O O H R H H H H HO O O O O n 1) Bu3SnH 80 ++ AIBN toluene O O R O O O O 80 °C, 6 h H 81b 22% 36f detected 83 4% 2) silica gel by LCMS H 82b 25% Scheme 3.20: Radical deoxygenation of 80.

-20 ◦C. A mixture410 of this and AIBN was added, in small portions, to 80 in deoxygenated405 toluene at 80 ◦C over 6 hours. Flash chromatography gave the desired 81b (25%) and its 8-epimer 82b (22% — see Scheme 3.20). Further chromatography gave a small amount of the cyclic carbonate 83, a common405 byproduct of this procedure. The expected 2-deoxy regioisomers 36f were not isolated; however, the less polar fractions eluted first contained a complex mixture of products, heavily contaminated with organotin compounds. Analy- sis of this mixture by reverse-phase liquid chromatography/mass spectrometry (RP-LC/MS) was performed by Claudio Medana at Turin University. Com- parison of the early fractions with 81b and 82b confirmed the presence of two less polar compounds isomeric with 81b (Figure 3.12). Surprisingly, the presumed 36f coeluted with the tin contaminants on C-18 reverse-phase, as on silica gel. Also, total ion count detection showed only a single broad organotin peak. Mass-selective detection, however, gave well-resolved peaks for the four products, illustrating the power of LCMS; none of the peaks were apparent by UV detection. The presence of 81b/82b in the early fractions shows that these were the major products, at least 2:1 relative to 36f.

3.5.3.4 Potential Improvements.

Given the incomplete recovery of 81b/82b, the yield could clearly be in- creased. One simple and effective way of removing organostannane reagents 3.5. MODIFICATION OF THE KETONE. 159

18.60 total ion count 100 50 ≤ ≤ 700 (early fractions) 80 m/z

60

40

20

0 100 19.03 15.95 80 m/z = 377 (early fractions) 19.39 60 15.21

40

20

relative abundance 0

82b 100 15.95 m/z = 377 (pure 81b/82b) 80 81b

60 15.24 40

20

0 0 2 4 6 8 101214161820222426283032343638404244 Time (min)

Figure 3.12: RP-LCMS traces of early fractions versus 81b/82b.

and byproducts is by washing an acetonitrile solution with hexanes.411, 410 However, in this case, the early fractions remained complex after washing, despite a dramatic reduction in organostannanes (mass reduced by a third). An alternative method is C-18 reverse phase chromatography.412 In this case,

C-18 RP-TLC in 80% MeCN/H2O confirmed the effectiveness of the wash: the petrol layer consisted of baseline material which was absent from the MeCN layer. However, the desired products were not resolved by RP-TLC, smear- ing badly. It would not be advisable to apply either of these methods to the crude reaction mixture: the initial reaction products are themselves alkyltin- substituted thiocarbonates, which are cleaved on silica to give the desired alcohols.405 Two more complex methods of organostannane removal, involv-

413 397 ing treatment with NaBH3CN or I2 and KF, are harsher and have the potential for side reactions. 160 CHAPTER 3. SYNTHESIS.

Thus, the organostannane route used above has serious disadvantages: the requirement for freshly prepared reagents, the difficulty of purification, the n low recovery of products and most importantly toxicity. While Bu3SnH is less toxic than other organostannanes,414 it is prepared from the highly

n 415 toxic ( Bu3Sn)2O, and the toxicity of the organostannane byproducts is unknown. This route is therefore not recommended. There are now many alternatives,398, 416, 417 some of which have been successfully applied to cyclic thionocarbonates.418, 419

3.5.3.5 1-Deoxysalvinorin A (81a).

O

O O H H O 1 2 10 O

O O 81a

Figure 3.13: 1-Deoxysalvinorin A (81a).

Acetylation of 81b (Ac2O/pyridine/DMAP, room temperature) gave 81a. Both H-2 (δ 4.74, tt) and H-10 (δ 1.10, dd) showed new trans-diaxial cou- plings to H-1α (δ 1.50, td), and smaller couplings to H-1β (δ 1.95-1.89, m). The HMQC spectrum correlated these peaks with a new 13C peak at 26.7 ppm. The DEPT spectrum confirmed C-1 as a methylene, and HRMS confirmed the molecular formula. The remaining data was extremely similar to 1a. Chapter 4

Bioassays.

4.1 Insect Antifeedant Activity.

A number of neoclerodane diterpenoids, very similar in structure to the salvi- norins, display a broad range of activities against insects.2 Some act as in- sect antifeedants, or appetite suppressants, rather than insecticides. Hundreds of clerodane diterpenoids have been tested; a thorough review is available.420 Salvinorins C-F (1c-1e and 1f) fit an empirical pharmacophore for antifeedant activity against Tenebrio molitor:421 an α,β-unsaturated carbonyl approxi- mately 10 Å from the oxygen of a conformationally constrained furan. By contrast, a saturated C-18 carbonyl, as found in salvinorins A and B (1a and 1b), seems to confer activity against Spodoptera littoralis.422

A selection of these compounds were therefore screened for antifeedant ac- tivity. This work was done in collaboration with Dr Charles Robin and Dr David Heckel in the Department of Genetics at the University of Melbourne. The species selected was Helicoverpa armigera, which has been widely used in antifeedant tests.420 A standard choice assay employing sweetened fibreglass discs422, 423 was used, with one modification. In early tests, the larvae refused to eat dry discs; the discs were therefore moistened, which gave satisfactory results. Larvae were presented with two discs: both were sweetened with su- crose, and one was treated with test compound. Tests were terminated when

161 162 CHAPTER 4. BIOASSAYS. more than half of one disc had been eaten. For full detail, see the Experimental Section.

The change in masses of the control (C) and treated (T) discs allow the cal- culation of the antifeedant index, AI:

100 × (∆C − ∆T ) AI = ∆C +∆T

Antifeedant indices fall between +100 (maximal antifeedant effect) and -100 (maximal phagostimulant effect); a value of zero indicates no effect.

The results are shown in Table 4.1.

Compound AI SEM salvinorin A 1a 030 salvinorin C 1c 10 26 divinatorin A 28a 611 divinatorin B 28b 223 divinatorin C 28c 16 33 Table 4.1: Antifeedant test results.

None of these compounds had a statistically significant effect. While the fact that four of the five compounds had a slightly positive AI suggests a weak (nonsignificant) effect, the standard error of the mean (SEM) is in each case larger than the mean itself. By contrast, potent antifeedants often have AI values of close to 100, with SEM < 5.420 The low means and high variabil- ity observed here suggest that these compounds are not potent antifeedants against this species. Nonetheless, the value of this experiment was limited by the unavailability of a proven antifeedant as a positive control compound. Also, there are significant differences between species; compounds inactive against one species are often active against others.422, 420 The most common larvae observed feeding on S. divinorum in Oaxaca were of the genus Chryocerinae.41 4.2. EUKARYOTIC PROTEIN SYNTHESIS INHIBITION. 163 4.2 Eukaryotic Protein Synthesis Inhibition.

Dr Jerry Pelletier at McGill University (Montreal) tested salvinorins A-D (1a- 1d) and divinatorins A-C (28a-28c) for inhibition of eukaryotic protein syn- thesis. The assay used has been described in detail elsewhere.424 Briefly, lu- ciferase, the protein responsible for chemiluminescence in fireflies, is produced in vitro in a cell culture. Addition of a protein synthesis inhibitor (such as anisomycin) causes a reduction in light output relative to the control. The light output thus provides a simple measure of protein synthesis. The results are shown in Figure 4.1. Luciferase from two species (firefly and Renilla) was used. None of the compounds tested caused significant inhibition at 50 µM (significant inhibition is defined as a relative light output of 30% or less). The positive control, anisomycin, caused complete inhibition at 10 µM.

anisomycin (+ve control) Firefly Renilla 1a

1b

1c

1d

28a

28b

28c

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Relative light output (%)

Figure 4.1: Luciferase assay results for salvinorins and divinatorins (50 µM). 164 CHAPTER 4. BIOASSAYS. 4.3 Antimicrobial Activity.

4.3.1 Bacteria and Fungi.

O

R1 H H R1 R2 R3

R2 28a OH H H 28b OH OH Me 18 28b H OAc H O OR3 29a H H H

Figure 4.2: (-)-Hardwickiic acid and divinatorins A-C.

(-)-Hardwickiic acid (29a) has been reported to display potent, broad-spectrum activity against bacteria and fungi.217 Divinatorins A-C (28a-28c)werethere- fore screened against standard antibiotic susceptible strains of Escherichia coli, Staphylococcus aureus, Bacillus subtilis,andCandida albicans, using broth mi- crodilution425, 426 and disk diffusion427 assays in each case.196 This work was done in collaboration with Professor Roy Robins-Browne and Andrea Bigham, in the Department of Microbiology and Immunology at the University of Mel- bourne.

No activity was apparent against any of the test organisms at 100 µg/mL or 100 µg/disk. These data extend the stringent structure-activity requirements of 29a.217 In particular, a hydrogen bond donor at C-18 seems to be a necessary, but not sufficient, condition of activity.217

To probe this further, we decided to screen (+)-hardwickiic acid (ent-29a). Ent-29a proved active against Staph. aureus (minimum inhibitory concen- tration (MIC) 25 µg/mL) and B. subtilis (MIC 12.5 µg/mL, 10 mm zone of inhibition), but much less potent than its enantiomer (MIC 0.78 µg/mL against B. subtilis).217 The assays are shown in Figure 4.3. A very small zone of inhibition is apparent around ent-29a in the disk diffusion assay.

An undergraduate investigation published without peer review429 reported that 4.3. ANTIMICROBIAL ACTIVITY. 165 streptomycin sulphate acetone (positive control) (negative control) (+)-hardwickiic acid (ent-29a) µg/mL 100 50 E. coli Staph. aureus MIC 25 MIC 12.5 6.25 3.12 1.56 0.78 0.39 C. albicans B. subtilis 0.19

(+)-hardwickiic acid (ent-29a) Crude extract B. subtilisE. coli S. aureus

Figure 4.3: Disk and microdilution assays for ent-29a and crude extract.428 the acetone extract of S. divinorum was active against a wide range of bacteria. We were unable to confirm these results. The acetone extract of the commercial material, as well as 1a, showed no activity at 100 µg/mL or 100 µg/disk. Apparently insufficient 29a and oleanolic acid (31) were present to elicit an effect (31, like 29a, is active against B. subtilis and Staph. aureus).245

4.3.2 HIV-1.

Several κ opioids have been shown to inhibit HIV replication in vitro in several human cell types.430 They appear to act as viral entry inhibitors, by causing downregulation of the CXCR4 co-receptor, used by the virus to attach to the cell.431 HIV entry inhibitors are currently the focus of intensive research.432 The use of synergistic drug cocktails has already produced dramatic progress, and the addition of a further mechanism is expected to offer still greater syn- ergy. KOR antagonists, which do not themselves inhibit viral replication, have been shown to enhance the effects of the standard therapy azidovudine (AZT).433

Beside their effect on viral replication, κ opioids also appear to counteract the 166 CHAPTER 4. BIOASSAYS. neurotoxic effects of HIV in infected cells.434, 435 Finally, since nausea is a com- mon side effect of standard therapies,436 the possible antiemetic activity437, 438 of κ opioids may be advantageous.

4.3.2.1 NL4.3 and AD8 Strains.

107 HIV strain:

106 NL4.3 cells 6 105 AD8

104

103 1a (10-6 M) HIV copies / 10 HIV copies Me2SO only 102 control 101 02468 Days post infection

Figure 4.4: HIV-1 replication assays (NL43 and AD8 strains).

Based on the above rationale, we submitted salvinorin A (1a)fortesting against HIV in vitro. Testing was performed by Dr Sharon Lewin and Ajantha Solomon at Monash University’s medical department (Alfred Hospital). Two strains of HIV-1 were tested: NL4.3 (T cell tropic) and AD8 (macrophage tropic). HIV was quantitated using real-time PCR.439, 440 Incubation of pe- ripheral blood mononuclear cells (PBMCs) with 1 µM salvinorin A had no effect on viral replication after one week (Figure 4.4). There was also no change in expression of CXCR4, CCR5 or CD4 receptors after one day, as determined using FACS staining.440

4.3.2.2 ROJO and TEKI Strains.

Two other HIV-1 strains were tested for the US National Institute of Allergy and Infectious Diseases (NIAID), under the direction of Dr Stephen Turk. 4.3. ANTIMICROBIAL ACTIVITY. 167

The isolates tested were ROJO (syncytium inducing/lymphocyte tropic) and TEKI (non-syncytium inducing/monocyte tropic). Again, salvinorin A (up to 230 µM) had no effect on viral replication in PBMCs after one week of incubation (eg. Figure 4.5).441 The positive control, AZT, strongly inhibited viral replication (IC50 = 13 nM). Tests were performed in triplicate.

160 160 PERCENT OF CELL CONTROL

140 salvinorin A (1a) 140

120 120

100 100

80 80

60 60 % VIRUS CONTROL 40 40 % CELL CONTROL 20 20

0.00 0.02 0.07 0.23 0.73 2.30 7.27 23.0 72.7 230 CONCENTRATION (µM) 140 140 AZT PERCENT OF CELL CONTROL 120 120

100 100

80 80

60 60

40 % VIRUS CONTROL 40 % CELL CONTROL 20 20 PERCENT OF VIRUS CONTROL PERCENT OF VIRUS CONTROL 0 0 0.00 0.10 0.32 1.00 3.16 10.0 31.6 100 316.2 1000 CONCENTRATION (nM)

Figure 4.5: HIV-1 replication assays (ROJO isolate).

4.3.2.3 Discussion.

These results were surprising given the established activity of the κ opioids U50,488 and U69,593, which are of comparable potency to 1a. However, the extremely potent endogenous κ opioid dynorphin A showed no activity;442, 443 this has been speculatively attributed to metabolism by peptidases in the cell culture. Consistent with this, the potent fragment dynorphin A1−13, which has a half life in plasma of less than one minute,444 caused negligible inhibition of HIV-1 replication.442 This suggests that the rapid metabolism87, 88 of 1a may account for its lack of activity.

However, this hypothesis cannot account for other evidence. The intact peptide

444 dynorphin A1−17, which has a half life in plasma of three hours, also caused 168 CHAPTER 4. BIOASSAYS. no inhibition.442 Moreover, only brief exposure to U50,488 (< 30 minutes) is necessary to inhibit HIV-1 replication.431 Thus, there appear to be other factors at work. Wang et al found that 1a causes much less κ opioid receptor internalisation than U50,488, despite their similar potencies.81 Such differences may involve the apparent subtypes445 of κ opioid receptors.

There is an anecdotal report of improved health following S. divinorum use by an AIDS patient.446 While the above data suggest that 1a itself has no activity against HIV, oleanolic acid (31) is known to be active by two complementary mechanisms,243, 242 as discussed in Section 2.2.4.1 on page 96. Given a sufficient dosage, therefore, the crude plant extract would probably exhibit some activity. No relevant data was located regarding other compounds in the plant.

4.4 NCI Anticancer Screen.

Salvinorins B and C (1b and 1c) and divinatorins A-C (28a-28c) were tested by the US National Cancer Institute, in the standard in vitro assay against 60 tumour cell lines.447 The standard measure of tumour cell growth inhibition is the GI50 – the drug concentration at which growth is reduced to 50% of the control value. In most cases, this degree of inhibition was not achieved. Where −5 substantial inhibition occurred, the GI50 was above 10 Minallcases(Figure −6 4.6). In the one apparent exception, 28b exhibited a GI50 of just under 10 M against the SF-539 brain tumour cell line. However, a significant inhibition of growth was already apparent at 10−8 M (59% of control), which scarcely increased up to 10−5 M (45%), a 1000-fold increase in concentration (Figure

4.7). The GI50 value was therefore clearly artefactual. In all other cases, inhibition of growth increased sharply at higher doses: see Figure 4.6. GI50 values this high are strongly predictive of poor performance in subsequent assays,448 and the NCI therefore did not select the compounds for further testing.

Salvinorin B (1b) completely inhibited growth, giving negative growth rates, 4.4. NCI ANTICANCER SCREEN. 169

All Cell Lines 100

50 divinatorin A (28a) 0

(NSC D737768) -50

-100 All Cell Lines -8 -7 -6 -5 -4 100 100

50 50 salvinorin B (1b) divinatorin B (28b) 0 0 % Growth (NSC D737715) (NSC D737769) -50 -50

-100 -100 -8 -7 -6 -5 -4 -8 -7 -6 -5 -4 100 100

50 50 salvinorin C (1c) divinatorin C (28c) 0 0 % Growth (NSC D737716) (NSC D737770) -50 -50

-100 -100 -8 -7 -6 -5 -4 -8 -7 -6 -5 -4

Log10 Concentration (M) Log10 Concentration (M)

Figure 4.6: NCI 60 cell line results for salvinorins and divinatorins in 8 cell lines. The “Total Growth Inhibition” concentration (TGI) was above 10−5 M in each case. No other compound gave this degree of inhibition. Only one cell line gave below -50% growth, the standard threshold for clear cyto- −4 toxicity: the SNB-75 CNS tumour line, with an LC50 slightly below 10 M (Figure 4.7). It is interesting that CNS tumour cells responded more strongly to 1b than the other categories. Nonetheless, the potencies even in these cases are unremarkable.

Salvinorin A (1a) was not accepted even for in vitro testing. Given the ap- 170 CHAPTER 4. BIOASSAYS.

100 100

50 50

divinatorin B (28b) salvinorin B (1b) 0 0 % Growth (NSC D737769) (NSC D737715)

-50 -50

-100 -100 -8 -7 -6 -5 -4 -8 -7 -6 -5 -4

Log10 Concentration (M) Log10 Concentration (M)

SF-268 CNS tumour cell line: SF-295 SF-539 SNB-75 U251 SNB-19

Figure 4.7: CNS cell line results for divinatorin B and salvinorin B. parently rapid deacetylation of 1a in cell culture, however, it seems likely that the results would in any case have been close to those of 1b. Surprisingly, the reason for the rejection of 1a was that it had already been tested in vivo more than 20 years earlier.449 The results have apparently not been published previously. Using a standard protocol,450 mice were implanted with P388 tu- mour cells (lymphocytic leukaemia); test animals were injected with 1a daily for 5 days. Doses of 1a up to 76 mg/kg had no effect, while 152 mg/kg was slightly toxic (survival time 79% of control).451 The value of 0% given on the NCI website449 is a typographical error.451

4.5 Activity at the κ Opioid Receptor.

Previous work found that 1b was inactive at the κ opioid receptor, while re- placement of the acetoxy group in 1a with more hindered esters dramatically reduced binding affinity.135 No information was available about other func- tional groups.

In order to further explore salvinorin A’s structure-activity relationships, iso- lated compounds and derivatives were tested in vitro for binding affinity at cloned opioid receptors using radioligand assays. Those compounds with sub- micromolar affinity were also screened for agonist potency and efficacy using 4.5. ACTIVITY AT THE κ OPIOID RECEPTOR. 171 a functional assay (see Experimental Section for details). The testing was done under the direction of Dr Bryan Roth, under the auspices of the National Institute of Mental Health’s Psychoactive Drug Screening Program.

No compound showed affinity for µ or δ subtypes (K i >1µM), and the following discussion will therefore discuss the κ opioid receptor exclusively. The raw data are shown in Table 4.2; structures can be found in the discussion below. The following discussion will express affinities and potencies relative to 1a:

Ki(x) Krel(x)= Ki(1a) and EC50(x) ECrel(x)= EC50(1a)

A value of 10 thus signifies a potency or affinity 10 × lower than 1a.

O

O O O H H 1a O O Ki 4 ± 1 nM

EC50 46 ± 8 nM

O O

Figure 4.8: KOR binding affinity and potency of salvinorin A.

4.5.1 Other Salvinorins and Divinatorins.

4.5.1.1 Radioligand Binding Results.

There have been conflicting results on salvinorin B (1b). Initial work by the

135 Roth group found it inactive (K i >10µM), but subsequent work by Béguin

137 271 125 et al reported high affinities (K i =66, 111 or 155 nM). Retesting by the Roth group indicated very weak affinity (3.1 µMversus1a control 44 nM).452 172 CHAPTER 4. BIOASSAYS.

K i s.e.m. K rel EC50 s.e.m. ECrel E max s.e.m. nM nM % 1a 41 46 8 100 19 1b 3,153* 71* 1c 1,022 262 255 1d >10,000 >2,500 1e >10,000 >2,500 28a >10,000 >2,500 28b >10,000 >2,500 28c >10,000 >2,500 46 18 2 4.5 315 35 6.8 108 11 37a 163 50 41 244 102 5.3 78 17 36e 1,125 281 36c >10,000 >2,500 51 156 18 39 126 36 2.7 108 14 35 59 11 15 78 21 1.7 107 5 49 611.5223 60 4.8 103 13 50 621.5624 200 14 116 10 77 347 53 87 >10,000 >217 81a 18 2 4.5 141 43 3 122 27 59 >10,000 >2,500 60a 2,900 400 725 * = direct comparison to 1a (44 nM) against [3H]diprenorphine Table 4.2: KOR radioligand and functional assay results.

O O O

1 O O OR O R1 H H H H HO R2O H H O O R2

O O O O O OR3

2 1 1 2 3 Krel R R Krel R R R Krel

1b 71 1c Ac Ac 255 28a OH H H >2,500 1d H Ac >2,500 28b OH OH Me >2,500 1e Ac H >2,500 28c H OAc H >2,500

Figure 4.9: KOR binding affinities of salvinorins and divinatorins.

The discrepancy is not as great as it appears, since the two groups obtained

137 different absolute affinities for 1a.Béguinet al reported values of K rel = 66,

271 125 452 85 or 120. The Roth group’s latest result (K rel = 71) is concordant. 4.5. ACTIVITY AT THE κ OPIOID RECEPTOR. 173

The nonconcordant earlier result (K rel > 500) may have resulted from the use of a different radioligand – [3H]bremazocine rather than [3H]diprenorphine. Moreover, in all cases the binding affinity of 1b was negligible relative to 1a.

Salvinorin C (1c) also showed negligible binding affinity compared to 1a (K rel = 255, Figure 4.9). Salvinorins D (1d)andE(1e) and divinatorins A-C (28a-

28c)showednoaffinity(K rel > 2,500). Based on these results, we tentatively concluded270 that 1a is the sole κ opioid present in the plant. Recently, Lee et al have reported that salvinorins B (1b)andG(1g) and divinatorin D (28d, Figure 2.51 on page 118) bind to the KOR, but with much lower affinities

137 and potencies than 1a ( K rel = 66 – 418). Since the concentrations of these compounds are also orders of magnitude lower, their contribution to the activation of the KOR by S. divinorum is negligible. Nine other isolated compounds were inactive. Thus, these results strongly suggest that 1a is effectively the sole active principle of S. divinorum.

The weak activity of 28d shows that the C ring can be cleaved without total loss of affinity. Apart from this greater conformational freedom, 28d not only lacks the 2-acetoxy function, but also possesses a 1α-hydroxy group, both of which drastically reduce activity in analogues of 1a, as will be discussed be- low. The structure-activity relationships of this compound are thus markedly different from those of 1a. Given that its deacetyl analogues 28b and 28e137 are inactive, perhaps the 17-acetoxy function substitutes for the furan ring in binding. However 28c, which also possesses this function, is inactive. Further exploration of this productive series of compounds is clearly warranted.

4.5.2 Modification of the Ketone.

Reduction of the ketone to an α-hydroxy group (giving 36e) dramatically reduced binding affinity (Figure 4.10). Acetylation (giving 36c) abolished binding entirely. These results were surprising, since molecular modelling had indicated that these modifications would not affect binding.67 Although the very low affinities of 1e, 1c, 36e,and36c might appear to suggest that the 174 CHAPTER 4. BIOASSAYS.

O O O

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

O O 36e O O 36c O O 81a

Krel 281 Krel >2,500 Krel 4.5

ECrel 3

Figure 4.10: KOR activity after ketone modifications. ketone is part of the pharmacophore, the relatively high affinity and potency of 1-deoxy compound 81a show that it is not. This suggests that 1α-hydroxy or acetoxy groups interact unfavourably with the KOR.

Curiously, saturating the 3,4-double bond in 1e (inactive) gives 36e (active), while the opposite is true of 1c (active) and 36c (inactive). However, none of these compounds show appreciable (sub-micromolar) affinity. The effect of this double bond itself therefore remains unclear.

4.5.3 Modification of the Acetoxy Group.

O O O

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

O O O 46 O O 59 O O 60a

Krel 4.5 Krel >2,500 Krel 725

ECrel 6.8

Figure 4.11: KOR activity after acetoxy group modifications.

It had previously been shown that substituting more hindered esters for the 2-acetoxy group reduced binding affinity.135 This suggested that the less- 4.5. ACTIVITY AT THE κ OPIOID RECEPTOR. 175 hindered formate 46 (Figure 4.11 on the preceding page) might prove more potent, but in fact both affinity and potency were reduced. The acetoxy group therefore appears to be the optimal alkyl chain length.

The autoxidation products 59 and 60a were also screened. 59 was inactive, as Lee et al reported for the incorrect structure 62 (Figure 3.8 on page 139).271

Surprisingly, 60a showed weak affinity at the KOR (K rel = 725). This provides further evidence that the 2-acyloxy function in 1a is not essential for binding, but that modifying this function usually reduces affinity dramatically.

4.5.4 Modification of the Methyl Ester.

O O

O O O O O O H H H H O O O O

OH O OH 77 67a

Krel 87 -

ECrel >217

Figure 4.12: KOR activity after methyl ester modifications.

The role of the methyl ester in binding was strongly confirmed by the results for 18-hydroxy derivative 77, which appeared to be an antagonist, since it bound but did not activate the receptor. However, no functional tests of antagonist potency were performed.

The precursor acid 67a was not tested, since the 8-epimers streaked on silica, so separation was not attempted. Lee et al subsequently separated the epimers using repeated chromatography. They reported, surprisingly, that the natural H-8β epimer was inactive at the KOR, while the H-8α epimer had high affinity

361 (K i =48nM). 176 CHAPTER 4. BIOASSAYS.

4.5.5 Modification of the Lactone.

O O O

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

35 49 50 O O O O O O Krel 15 Krel 1.5 Krel 1.5

ECrel 1.7 ECrel 4.8 ECrel 14

Figure 4.13: KOR activity after lactone modifications.

The high affinities of lactol 35,ether49, and enol ether 50 (all full agonists) show that the lactone carbonyl is not essential for binding or activity. How- ever, the latter two compounds, especially 50, showed reduced potency in the functional assay.

4.5.6 Modification of the Furan Ring.

O O

13 H H O O O O O O H H H H O O 8 O O

O O 37a O O 51 Krel 41 Krel 39

ECrel 5.3 ECrel 2.7

Figure 4.14: KOR activity after furan modifications.

The substantially reduced affinity of tetrahydrofuran 51 suggests that the fu- ran ring is involved in binding. As described earlier, one 13-epimer of 51 was isolated for characterisation. The binding affinity did not differ signif- icantly from the 1:1 mixture (123 nM vs 156 nM). The reduced affinity of 4.5. ACTIVITY AT THE κ OPIOID RECEPTOR. 177

8-epi-salvinorin A (37a) also supports the role of the furan; since the lactone carbonyl is evidently not necessary for binding, the orientation of the furan is the salient difference between this compound and 1a. The reduction in potency of these compounds was far smaller than the reduction in affinity, particularly in 51.

4.5.7 Incorporation into a Revised Binding Model.

H N Gln 115 H O O Tyr 313 O H O O O H H O O H O

Tyr 312 O H O O Tyr 139

Figure 4.15: Westkaemper’s original binding model.

The original report of the activity of 1a at the KOR67 included a proposed model of its interactions with the receptor (Figure 4.15). As shown, the furan oxygen and the acetyl, methyl ester and lactone carbonyls act as hydrogen bond acceptors. The model, created by Dr Richard Westkaemper, was derived from a model of U69,593 binding reported previously.

While the proposed roles of the furan, acetoxy group and methyl ester were consistent with our data above, the proposed interaction with the lactone seemed unlikely given the high affinities of 35, 49 and 50 mentioned above.

Westkaemper subsequently proposed a revised model,453 which drew upon the above data on the roles of each functional group in binding, although this was regrettably not acknowledged in the finished paper. Additional data on the role of the 2-acetoxy group was provided by others.453 Complementing these chemical data, site-directed mutagenesis in the Roth group allowed selective 178 CHAPTER 4. BIOASSAYS. modification of the receptor itself, identifying the residues involved in binding to 1a.453 The revised model is shown in Figure 4.16, visualised from coordi- nates kindly provided by Dr Westkaemper. For clarity, hydrogen atoms are not shown.

Tyr 313Tyr 320 Tyr 313 Tyr 320

Tyr 119- - - H-bonds* Tyr 119 - - - H-bonds*

Ile 294 Ile 294 - - - hydrophobic - - - hydrophobic interactions interactions Glu 297 Glu 297

*not present simultaneously *not present simultaneously

Figure 4.16: Westkaemper’s revised binding model (stereoview).

The revised model has only one residue in common with the original: tyrosine 313. The proposed interaction is a hydrophobic one with the acetyl methyl group, however, rather than the original H bond to the carbonyl oxygen. This was suggested by the surprising result that mutation of this tyrosine residue to phenylalanine (lacking the hydroxyl) did not affect binding affinity.453 Three of the five proposed interactions in the new model are hydrophobic. The mutagenesis data suggest that the furan can form an H bond to either of tyrosines 199 or 320.453 While these bonds are both shown in Figure 4.16 for clarity, they cannot be present simultaneously, since only one nonbonded pair is available on the furan oxygen.

One interesting aspect of this model is that three of the five residues involved (isoleucine 294, glutamic acid 297 and tyrosine 313) are unique to the κ sub- type. This provides a plausible explanation for the striking κ selectivity of 4.5. ACTIVITY AT THE κ OPIOID RECEPTOR. 179

1a and analogues; very few derivatives show submicromolar affinity δ or µ subtypes, and none of these are selective.25, 82, 136

4.5.8 Subsequent Results.

O O O O HO O HO H O O O O H H O O O O RO H H H H O O O O O

O O O OH O O 67a (H-8b) 34a O O O Krel >770 ECrel 102

84 85 O 86

Krel 6 0.31 O Ki (m) 12

ECrel 4 0.13 Krel 0.13 (m/k)

Figure 4.17: KOR binding affinities and potencies of recent derivatives.

Since publication of these results, numerous 1a analogues have been tested in vitro by other groups. These results again confirm that replacement of the 2-acetoxy group (with other esters, ethers, carbonates, amides etc) almost in- variably reduces binding affinity.95, 271, 82 There are exceptions, however. The high affinity and potency of ethyl ether 8495 indicate that the acetyl carbonyl is not essential for activity. This result provides strong support for the hy- drophobic interaction at C-2 in Westkaemper’s revised model (Figure 4.17). Another notable derivative is methoxymethyl ether 85,271 to date the only derivative significantly more potent than 1a. Surprisingly, benzoate 8682 and some related derivatives activate the µ opioid receptor, but are not selective. Other results indicate that epimerisation at C-2 does not abolish activity in all derivatives.125

There is less data on modifications at C-18,361, 95 but all esters and amides 180 CHAPTER 4. BIOASSAYS. tested showed large reductions in affinity and potency. The loss of affinity on O-demethylation (giving 67a)361 strongly supports the carbinol hydrophobic interaction in Westkaemper’s model over the more intuitive carbonyl H bond.

Even less data has been reported on furan modifications, but the greatly re- duced potency of salvinicin A (34a)25 confirms that hindrance in this region interferes with binding. Some synthetic modifications of the furan ring have been prepared,136 but biological data have not been reported. Chapter 5

Experimental.

5.1 General Conditions

5.1.1 Instruments and Procedures.

Flash Chromatography: Flash column chromatography was performed ac- cordingtoLeonard’s194 procedure using Scharlau silica gel 60 (particle size 0.04 - 0.06 mm) or Merck silica gel 60. Mass ratios up to 400:1 versus

crude product were used for difficult separations (∆hRf <5).Theac- tivated carbon used was Merck aktivkohle 2183 powder (20:1 mass ratio versus crude product).

HPLC: Spherex 5 µm silica column (250 × 10 mm), flow rate 2 mL min−1 with refractive index detection. Solvent front reached detector at 6.8 min.

HRESIMS: Bruker 4.7T BiOAPEX FTMS.

InChIs: IUPAC International Chemical Identifiers454 were created using winChI version 1.455

IR: Bio-Rad FTS 165 FT-IR (running WIN-IR v 4.14) and Shimadzu FTIR 8400 (running HYPER IR v 1.57), using thin films on NaCl discs.

181 182 CHAPTER 5. EXPERIMENTAL.

LCMS: Phenomenex Luna 3 µm C-18 column (150 × 2 mm). Gradient elu-

tion: 0.05 % HCO2H in 20-100% MeCN/H2O over 40 minutes. ESI in- terface with ion trap detector. Detailed conditions have been published elsewhere.59

NMR: Varian Inova 400, Inova 500 and Unity Plus 400 (running Vnmr 6.1; some processing on VnmrJ 1.1D, some on iNMR). Bruker US2 800 (run- ning TopSpin 1.3). 1H NMR are 400 MHz, and 13C NMR 100 MHz, un- less otherwise stated. 13C multiplicities are based on DEPT experiments. Assignments are given according to the standard32 clerodane numbering scheme below. Where 13C assignments are given, all assignments (1Hand 13C) are based on 2D NMR data (COSY, HMQC, HMBC). In other cases, 1H assignments were made by comparison with related compounds for which such data was available. Stereochemical assignments were based on coupling constants where possible, and NOESY data elsewhere (eg. pro-R vs. pro-S in rotatable methylenes). Peaks whose stereochemistry could not be unambiguously assigned on these bases are listed as “a” and “b”. Complex first-order multiplets were analysed using Hoye’s al- gorithms,456, 457 with resolution enhancement using the line-broadening window function where necessary (not shown in the reproduced spectra). Abbreviations: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, quin = quintet, sext =sextet, sept = septet, b = broad. The δ

values of solvent residual peaks (eg. CHCl3 = 7.26 ppm) and impurities were taken from Gottlieb et al.458

15

16 14 13

11 12

1 9 17 2 10 8

5 20 3 7 4 6 19 18

Polarimetry: JASCO DIP-1000. Concentration c is in g/l00 mL; the units 5.1. GENERAL CONDITIONS 183

of the specific rotation are (◦·mL·g−1·dm−1).

SEM: Samples were sputter-coated with gold. Fresh leaves were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer at rt for 24 h. The samples were rinsed in the buffer for 30 min (× 3)andthendehydratedinincreasing concentrations of aq. EtOH for 30 min each (10, 30, 50, 70, 90, and 100%). After two further rinses in fresh 100% EtOH, the leaf samples were dried in a critical point dryer.

TLC: Merck silica gel 60 F254 plates, visualised with phosphomolybdic acid

in EtOH and heated unless otherwise indicated. “hRf ”=Rf × 100.

UV: Shimadzu UV-2401PC (quartz cell).

5.1.2 Reagents.

DMPU and EtSH (Aldrich, 98%) were stored over 4Å sieves.

“Petrol” refers to the fraction boiling at 40-60 ◦C.

Reactions “under Ar” were performed using freshly distilled solvents unless otherwise indicated.

5.1.3 Plant Materials.

Commercial material Dried S. divinorum leaves, cultivated in Oaxaca Mex- ico, were purchased from Salvia Space Ethnobotanicals (Berkeley, Cal- ifornia). Voucher specimens were deposited at the National Herbarium of Victoria (accession number MEL 2101361) and the University of Mel- bourne Herbarium (MELU s.n.)

Australian material Additional S. divinorum plants were cultivated in Mel- bourne. A voucher specimen of the dried leaves was deposited at the National Herbarium of Victoria (accession number MEL 2145478). 184 CHAPTER 5. EXPERIMENTAL.

Copaiba balsam was donated by Australian Botanical Products (Hallam, Victoria).

5.1.4 Assays.

Radioligand Binding Assays performed as previously detailed459, 67 using cloned receptors stably expressed in HEK 293 cells.

κ: rat KORs with [3H]diprenorphine (50 Ci/mmol, PerkinElmer Inc) or [3H]U69,593 (41.4 Ci/mmol, PerkinElmer Inc) as radioligand.

δ: humanDORswith[3H]DADLE (51.5 Ci/mmol, PerkinElmer Inc) as radi- oligand.

3 µ: humanMORswith[H]diprenorphine as radioligand. K i values were calcu- lated using Prism 4.01 (GraphPad Software, Inc) as the mean ±SEM of quadruplicate (n ≥ 4) determinations. Nonspecific binding was defined using 10 µM naloxone.

Calcium Flux Functional Assay. Performed as previously detailed135, 460 using cloned rat KORs stably expressed in HEK 293 cells, cotransfected 2+ with the universal G protein Gα16.Ca mobilisation was quantified using a 96-well FlexStationII with the calcium flux assay kit (Molecular

Devices Corp, Sunnyvale, CA). EC50 and Emax values were calculated using Prism 4.01 (GraphPad Software, Inc), as the mean ±SEM of qua- druplicate (n ≥ 4) determinations.

Antimicrobial Tests. The extract and compounds were tested against Es- cherichia coli (ATCC 25922), Staphylococcus aureus (ATCC 29213), Bacillus subtilis (ATCC 6633), and Candida albicans (ATCC 90028) us- ing standard broth microdilution426, 425 (100 - 0.19 µg/mL using two-fold serial dilutions) and disc-diffusion427 assays (100 µg/disk). All measure- ments were performed in duplicate. Streptomycin sulfate and ampho- tericin B were used as positive controls. 5.2. ISOLATION 185

Insect Antifeedant Tests. A standard choice assay employing sweetened fi- breglass discs423 was used, except that the discs were moistened. Final instar H. armigera larvae were selected on the basis of head size. Squares of Whatman GF/A filter (3.6 cm2) were treated with 100 µLof0.05M sucrose solution. Test discs were then treated with 10 µgoftestcom- pound in acetone, usually423 expressed as 100 µL of a 100 ppm (100 mg L−1) solution. Control discs were treated with the same volume of ace- tone. Discs were dried, numbered and weighed. Before performing the assay, the discs were moistened with distilled water. After more than half of one disc had been eaten, the discs were dried and reweighed. Tests were performed with 5 replicates per compound.

HIV-1 Replication Assays.440 Alfred Hospital: PBMC were stimulated for 72 hours, and then infected with NL4.3orAD8(30ng/mlvirus)for2 hours. Cells were then washed and seeded in 24 well plates. Salvinorin

A was dissolved in Me2SO before addition. Me2SO at the same concen- tration was also compared to the control. Cells were lysed on day 0, 2, 5 and 7 of infection. HIV was quantitated using real-time PCR.439 NIAID/Southern Research Institute: detailed assay conditions can be found elsewhere.441

5.2 Isolation

5.2.1 Extraction of Commercial S. divinorum.

Dried S. divinorum leaves (860 g) were powdered and steeped for 1 h in acetone (3 × 1 L). Filtration and evaporation under reduced pressure gave a dark green tar (30.5 g). This was purified by flash column chromatography on an equal mixture of activated carbon and diatomite filter aid, eluting with a stepwise gradient from acetone to petrol, to give an amber semicrystalline syrup (5.73 g). Recrystallisations from MeOH and EtOH gave 1a (2.64 g). The mother liquor was purified by flash column chromatography on silica gel 186 CHAPTER 5. EXPERIMENTAL.

(5-50% acetone/CH2Cl2 gradient). This was divided, based on TLC (10% acetone/CH2Cl2), into four series: A (656 mg), B (150 mg), C (359 mg) and D(77mg).

TLC: hRf of series A B C D

10% acetone/CH2Cl2 62 36-53 18 12

Series A: flash column chromatography, eluting with a gradient from 50-80%

Et2O/petrol, gave 1c (total yield 219 mg, 0.25 g/kg) and additional 1a,which was recrystallised from EtOH (total yield 2.9 g, 3.4 g/kg). Further elution gave 29a (7 mg).

Series B: flash column chromatography on silica gel (35 g) in 70-90% Et2O/petrol, and recrystallisation from MeOH, gave 1b (13 mg, 0.015 g/kg).

Series C: Trituration in hot Et2Ogave1d (75 mg). Flash column chromatog- raphy of the mother liquor (60-100% Et2O/petrol) gave four fractions based on TLC (70% Et2O/petrol): C1 (55 mg), C2 (119 mg), C3 (57 mg) and C4 (39 mg).

TLC: hRf of fraction C1 C2 C3 C4

70% Et2O/petrol 40-43 28-21 17 13

Fraction C1: Repeated flash column chromatography (20% acetone/petrol and 40-60% Et2O/petrol) gave 28c (23 mg) and 31 (3 mg).

Fraction C2: Repeated flash column chromatography (25% acetone/petrol and 60-100% Et2O/petrol) gave 28b (41 mg).

Fraction C3: Extensive flash column chromatography (Et2O/petrol, ace- tone/petrol and EtOAc/petrol) gave additional 28b (total yield 41 mg) and a mixture of 1e and 1f. Final purification by HPLC (60% EtOAc/petrol) gave 1e (2.8 mg) and 1f (1.1 mg).

Fraction C4 gave additional 1d (total yield 114 mg, 0.13 g/kg).

Series D: Repeated flash column chromatography (60% Et2O/petrol and 4%

MeOH/CH2Cl2)gave28a (36 mg). 5.2. ISOLATION 187

5.2.2 Extraction of Australian S. divinorum.

Dried, powdered S. divinorum leaves (224 g) were steeped in acetone for 30 min (3 × 250 mL). Filtration and evaporation under reduced pressure gave a dark green tar (7 g). This was purified by vacuum filtration through a 1:1 mixture of activated carbon (75 g) and diatomite filter aid, eluting with a gradient from 50-20% EtOAc/petrol, to give series E (97 mg) and F (279 mg) basedonTLC(70%Et2O/petrol).

TLC: hRf of series E F

70% Et2O/petrol 55-68 16

Series E: Repeated flash column chromatography (1% acetone/CH2Cl2 and

20% Et2O/petrol) gave 33 (1 mg). Further flash column chromatography

(0.75% MeOH/CH2Cl2 and 1% EtOH/CHCl3)gave32 (23 mg) and 30 (12 mg).

Series F: Two recrystallisations from MeOH gave 1a (126 mg).

yield 70% Et2O/ 10% acetone/

(mg/kg) petrol CH2Cl2 Salvinorin A 1a 3,400 24 57 B 1b 15 14 37 C 1c 254 31 60 D 1d 132 18 25 E 1e 32347 F 1f 12440 Divinatorin A 28a 42 37 15 B 28b 48 31 31 C 28c 27 50 39 (–)-Hardwickiic acid 29a 86445 Oleanolic acid 31 36665 Presqualene alcohol 32 34434 Peplusol 33 47573 (E)-Phytol 30 35958

Table 5.1: Yields and TLC data (hRf ) of isolated compounds. 188 CHAPTER 5. EXPERIMENTAL.

5.2.3 Salvinorin A (1a).

InChI=1/C23H28O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8,11,14-17,19H,5,7,9-10H2,1- 4H3/t14-,15-,16-,17-,19-,22-,23-/m0/s1

O

O O O H H O O

O O 1a

colourless crystals, mp (from EtOH) 236-238 ◦C; lit. (from MeOH) 238-240 ◦C;22

TLC: See Table 5.1 on the previous page.

UV (MeCN): λmax (log ε)208(3.76)nm;

23 lit. UV (MeOH): λmax (log ε) 211 (3.72) nm;

1 H NMR (800 MHz, CDCl3): δ 7.41 (1H, dt, J = 1.7, 0.9 Hz, H-16), 7.38 (1H, t, J =1.7Hz,H-15),6.37(1H,dd,J = 1.9, 0.9 Hz, H-14), 5.52 (1H, ddd, J = 11.7, 5.2, 0.8 Hz, H-12), 5.14 (1H, ∼ddt, J ≈ 11.9, 9.0, 0.9 Hz, H-2), 3.72

(3H, s, CO2CH 3), 2.74 (1H, ∼dd, J ≈ 11.3, 5.6 Hz, H-4), 2.50 (1H, dd, J = 13.5, 5.2 Hz, H-11α), 2.31-2.28 (2H, m, H-3), 2.18 (1H, br s, H-10), 2.17-2.14

(1H, m, H-7β), 2.16 (3H, s, OCOCH 3), 2.07 (1H, dd, J = 12.0, 3.1 Hz, H-8), 1.79 (1H, dt, J = 13.4, 3.1 Hz, H-6α), 1.64 (1H, tdd, J = 13.5, 12.1, 3.4 Hz, H-7α), 1.58 (1H, td, J = 13.5, 0.9 Hz, H-6β), 1.57 (1H, ddd, J = 13.5, 11.7, 0.8 Hz, H-11β), 1.45 (3H, s, H-20), 1.11 (3H, s, H-19);

13 22 CNMR(CDCl3) data matched previously reported values. 5.2. ISOLATION 189

5.2.4 Salvinorin B (1b).

InChI=1/C21H26O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7,10,12-15,17,22H,4,6,8-9H2,1-3H3/ t12-,13-,14-,15-,17-,20-,21-/m0/s1

O

O O H H HO O

O O 1b

◦ colourless crystals, mp (from hot MeOH) 239-240 C; (from cold Et2O/petrol) 244-245 ◦C; lit. (from hot MeOH) 213-216 ◦C;23 251-254 ◦C;15, 260 (from cold MeOH) 211- 214◦C;169

1 13 23 Hand CNMR(CDCl3) data matched previously reported values.

5.2.5 Salvinorin C (1c).

InChI=1/C25H30O9/c1-13(26)32-18-10-17(22(28)30-5)24(3)8-6-16-23(29)34-19 (15-7-9-31-12-15)11-25(16,4)21(24)20(18)33-14(2)27/h7,9-10,12,16,18-21H,6,8, 11H2,1-5H3/t16-,18-,19-,20-,21-,24-,25-/m0/s1

O

O O O O H H O O

O O 1c 190 CHAPTER 5. EXPERIMENTAL. clear resin;

TLC: See Table 5.1 on page 187.

16 [α]D +70 (c 0.6, CHCl3);

22 27 [α]D +49 (c 0.6, CHCl3) lit;

UV (MeCN): λmax (log ε)208(4.10)nm;

FTIR (film): ν
max 3145, 2952, 1740, 1643, 1506, 1435, 1372, 1315, 1226, 1174, 1142, 1075, 1041, 961, 950, 910, 875, 789, 754, 695, 667 cm−1;

1 13 27 Hand CNMR(CDCl3) data matched previously reported values.

5.2.6 Salvinorin D (1d).

InChI=1/C23H28O8/c1-12(24)30-18-16(25)9-15(20(26)28-4)22(2)7-5-14-21(27) 31-17(13-6-8-29-11-13)10-23(14,3)19(18)22/h6,8-9,11,14,16-19,25H,5,7,10H2, 1-4H3/t14-,16-,17-,18-,19-,22-,23-/m0/s1

O

O

O O H H HO O

O O 1d

◦ fine colourless crystals, mp (from cold Et2O) 185-187 C;

TLC: See Table 5.1 on page 187.

17 [α]D +67 (c 1.0, CH2Cl2);

FTIR (film): ν
max 3475, 3146, 2952, 2861, 1723, 1505, 1435, 1371, 1315, 1228, 1195, 1174, 1142, 1071, 1056, 1027, 949, 875, 788, 744, 699 cm−1;

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.54 (1H, dd, J = 2.4, 1.3 Hz, H-3), 6.40 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.70 5.2. ISOLATION 191

(1H, dt, J = 5.5, 1.3 Hz, H-1), 5.53 (1H, dd, J = 11.2, 5.8 Hz, H-12), 4.44

(1H, ddd, J = 6.7, 5.5, 2.4 Hz, H-2), 3.74 (3H, s, CO2CH 3), 2.56 (1H, dt, J = 13.3, 3.4 Hz, H-6α), 2.54 (1H, dd, J = 13.1, 5.9 Hz, H-11α), 2.17-2.09 (1H, m, H-7β), 2.15 (3H, s, OCOCH 3), 2.13 (1H, dd, J = 13.5, 3.5 Hz, H-8), 2.01 (1H, br d, J =6.7Hz,OH), 1.78 (1H, dtd, J = 15.1, 13.3, 3.5 Hz, H-7α), 1.69 (3H, s, H-19), 1.64 (1H, ddd J = 13.1, 11.2, 1.0 Hz, H-11β), 1.42 (1H, br s, H-10),1.22(3H,s,H-20),1.20(1H,td,J = 13.5, 3.5 Hz, H-6β);

13 CNMR(CDCl3): δ 171.57 (C, C-17/OCOCH3), 171.55 (C, C-17/OCOCH3), 166.2 (C, C-18), 143.8 (CH, C-15), 141.2 (C, C-4), 139.4 (CH, C-16), 135.7 (CH, C-3), 125.4 (C, C-13), 108.4 (CH, C-14), 71.6 (CH, C-12), 68.6 (CH,

C-2), 66.5 (CH, C-1), 52.5 (CH, C-10), 51.8 (CH, C-8), 51.7 (CH3,CO2CH3),

43.9 (CH2, C-11), 37.6 (C, C-5), 37.03 (C, C-9), 36.97 (CH2, C-6), 21.6 (CH3,

C-19), 21.2 (CH3,OCOCH3), 18.3 (CH2, C-7), 15.6 (CH3, C-20);

+ + HRESIMS: [M + Na] m/z 455.1672 (calcd for C23H28O8Na , 455.1676).

5.2.7 Salvinorin E (1e).

InChI=1/C23H28O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8-9,11,14,16-19,25H,5,7,10H2, 1-4H3/t14-,16-,17-,18-,19-,22-,23-/m0/s1

O

O OH O H H O O

O O 1e clear resin;

TLC: See Table 5.1 on page 187.

HPLC: tR (min) 1e 1f 60% EtOAc/petrol 9.8 10.7 192 CHAPTER 5. EXPERIMENTAL.

17 [α]D +46 (c 0.14, CHCl3);

FTIR (film): ν
max 3510, 3144, 2952, 2858, 1722, 1641, 1505, 1436, 1374, 1316, 1228, 1142, 1070, 1029, 949, 935, 897, 875, 805, 788, 763, 708, 690, 679 cm−1;

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.43 (1H, dd, J = 2.4, 1.6 Hz, H-3), 6.41 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.60 (1H, ddd, J = 11.0, 5.8, 0.9 Hz, H-12), 5.40 (1H, dd, J = 4.9, 2.4 Hz, H-2),

4.46 (1H, ddd, J = 4.7, 1.6, 1.3 Hz, H-1), 3.73 (3H, s, CO2CH 3), 2.52 (1H, ddd, J = 12.5, 3.0, 2.6 Hz, H-6α), 2.46 (1H, dd, J = 13.1, 6.0 Hz, H-11α),

2.17 (3H, s, OCOCH 3), 2.18-2.07 (2H, m, H-7β & 8), 1.94 (1H, dd, J =2.3, 1.5 Hz, OH), 1.84 (1H, dtd, J = 14.2, 12.0, 3.7 Hz, H-7α), 1.72 (3H, s, H-19), 1.62 (1H, dd, J = 13.1, 11.2 Hz, H-11β), 1.47 (3H, s, H-20), 1.30 (1H, br s, H-10), 1.19 (1H, td, J = 13.3, 3.7 Hz, H-6β);

13 CNMR(CDCl3): δ 171.8 (C, OCOCH3), 169.8 (C, C-17), 166.0 (C, C-18), 143.9 (CH, C-15), 143.4 (C, C-4), 139.3 (CH, C-16), 131.5 (CH, C-3), 125.8 (C, C-13), 108.4 (CH, C-14), 72.3 (CH, C-2), 71.7 (CH, C-12), 64.3 (CH, C-

1), 54.0 (CH, C-10), 51.8 (CH3,CO2CH3), 51.7 (CH, C-8), 44.4 (CH2, C-11),

37.8 (C, C-5), 37.5 (C, C-9), 37.0 (CH2, C-6), 21.9 (CH3,C-19),21.0(CH3,

OCOCH3), 18.4 (CH2, C-7), 16.2 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 455.1687 (calcd for C23H28O8Na , 455.1676).

5.2.8 Salvinorin F (1f).

InChI=1/C21H26O6/c1-20-8-6-14-19(24)27-16(12-7-9-26-11-12)10-21(14,2)17 (20)15(22)5-4-13(20)18(23)25-3/h4,7,9,11,14-17,22H,5-6,8,10H2,1-3H3/t14-,15 -,16-,17-,20-,21-/m0/s1 5.2. ISOLATION 193

O

OH O H H O

O O 1f

clear resin;

TLC: See Table 5.1 on page 187.

HPLC: see table on page 191.

16 [α]D −20 (c 0.05, CHCl3);

FTIR (film): ν
max 3514, 3147, 2951, 2857, 1712, 1637, 1505, 1436, 1372, 1318, 1232, 1194, 1144, 1070, 1028, 978, 947, 896, 875, 797, 756, 686 cm−1;

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.67 (1H, ddd, J = 4.7, 3.0, 1.0 Hz, H-3), 6.41 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.60 (1H, ddd, J = 11.2, 5.8, 0.9 Hz, H-12), 4.51 (1H, br dd, J = 5.5, 4.7 Hz,

H-1), 3.72 (3H, s, CO2CH 3), 2.60 (1H, ddd, J = 20.1, 5.5, 3.0 Hz, H-2β), 2.53 (1H, dt, J = 13.2, 3.2 Hz, H-6α), 2.46 (1H, dd, J = 13.2, 5.8 Hz, H-11α), 2.35 (1H, ddt, J = 20.1, 4.6, 1.1 Hz, H-2α), 2.17-2.08 (2H, m, H-7β, 8), 1.82 (1H, dtd, J = 15.0, 13.2, 3.5 Hz, H-7α), 1.71 (3H, s, H-19), 1.62 (1H, ddd, J = 13.2, 11.4, 0.9 Hz, H-11β), 1.48 (3H, s, H-20), 1.29 (1H, dd, J = 4.7, 0.7 Hz,

OH), 1.25 (1H, br s, w 1 = 3.8 Hz, H-10), 1.18 (1H, tdd, J = 13.3, 3.6, 0.9 Hz, 2 H-6β);

13 CNMR(CDCl3): δ 172.1 (C, C-17), 166.9 (C, C-18), 143.8 (CH, C-15), 140.6 (C, C-4), 139.3 (CH, C-16), 133.4 (CH, C-3), 125.9 (C, C-13), 108.4 (CH, C- 14), 71.7 (CH, C-12), 63.9 (CH, C-1), 54.8 (CH, C-10), 52.2 (CH, C-8), 51.5

(CH3,CO2CH3), 44.4 (CH2,C-11),38.0(CH2, C-2), 37.7 (C, C-9), 37.3 (CH2,

C-6), 36.6 (C, C-5), 21.6 (CH3, C-19), 18.6 (CH2, C-7), 16.4 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 397.1610 (calcd for C21H26O6Na , 397.1622). 194 CHAPTER 5. EXPERIMENTAL.

5.2.9 Divinatorin A (28a).

InChI=1/C20H28O4/c1-13-6-9-20(3)15(18(22)23)4-5-16(21)17(20)19(13,2)10- 7-14-8-11-24-12-14/h4,8,11-13,16-17,21H,5-7,9-10H2,1-3H3,(H,22,23)/t13-,16+, 17-,19+,20+/m1/s1/f/h22H

O

OH H H

O OH 28a amber resin;

TLC: See Table 5.1 on page 187.

19 [α]D −53 (c 1.8, CH2Cl2);

FTIR (film): ν
max 3392, 2927, 2874, 2648, 1684, 1634, 1503, 1456, 1411, 1386, 1262, 1245, 1221, 1162, 1102, 1066, 1025, 1003, 966, 924, 894, 874, 782, 760, 703, 670 cm−1;

1 HNMR(CDCl3): δ 7.36 (1H, t, J = 1.6 Hz, H-15), 7.20 (1H, td, J =2.5, 1.0 Hz, H-16), 6.90 (1H, ddd, J = 4.8, 2.7, 1.0 Hz, H-3), 6.25 (1H, dt, J = 1.8, 1.0 Hz, H-14), 4.49 (1H, br d, J = 4.8 Hz, H-1), 2.56 (1H, ddd, J = 20.1, 5.1, 2.8 Hz, H-2β), 2.43-2.33 (2H, m, H-2α,6α), 2.34 (1H, br td, J = 13.5, 4.2 Hz, H-12-pro-R), 2.05 (1H, br ddd, J = 14.3, 12.9, 4.7 Hz, H-12-pro-S), 1.85 (1H, ddd, J = 14.7, 12.8, 4.7 Hz, H-11-pro-S), 1.67 (1H, ddd, J = 14.8, 12.8, 4.4 Hz, H-11-pro-R), 1.64 (3H, s, H-19), 1.60-1.54 (2H, m, H-7α, 8), 1.47-1.42 (1H, m, H-7β), 1.45 (1H, br s, H-10), 1.23-1.17 (1H, m, H-6β), 1.15 (3H, s, H-20), 0.84 (3H, d, J =6.2Hz,H-17);

13 CNMR(CDCl3): δ 171.8 (C, C-18), 142.8 (CH, C-15), 140.8 (C, C-4), 138.4 (CH, C-16), 136.2 (CH, C-3), 125.2 (C, C-13), 110.9 (CH, C-14), 64.7 (CH,

C-1), 49.0 (CH, C-10), 39.7 (C, C-9), 39.1 (CH2, C-11), 38.6 (CH2, C-6), 38.1 5.2. ISOLATION 195

(CH2, C-2), 37.4 (C, C-5), 37.1 (CH, C-8), 27.4 (CH2, C-7), 21.4 (CH3, C-19),

19.8 (CH3, C-20), 18.2 (CH2, C-12), 15.7 (CH3,C-17);

+ + HRESIMS: [M + Na] m/z 355.1864 (calcd for C20H28O4Na , 355.1880).

5.2.10 Divinatorin B (28b).

InChI=1/C21H30O5/c1-20(9-6-14-8-11-26-13-14)15(12-22)7-10-21(2)16(19(2 4)25-3)4-5-17(23)18(20)21/h4,8,11,13,15,17-18,22-23H,5-7,9-10,12H2,1-3H3/ t15-,17-,18+,20-,21-/m0/s1

O

OH H H OH

O O 28b amber resin;

TLC: See Table 5.1 on page 187.

20 [α]D −54 (c 2.1, CHCl3);

FTIR (film): ν
max 3434, 2930, 2881, 1714, 1503, 1461, 1437, 1385, 1356, 1236, 1196, 1162, 1097, 1067, 1025, 979, 943, 921, 874, 781, 759, 734 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.35 (1H, t, J = 1.7 Hz, H-15), 7.20 (1H, dd, J = 1.6, 1.0 Hz, H-16), 6.65 (1H, ddd, J = 4.9, 2.8, 1.0 Hz, H-3), 6.25 (1H, dd, J = 1.8, 0.9 Hz, H-14), 4.46 (1H, dq, J = 5.1, 1.3 Hz, H-1), 3.84 (1H, dd, J =

10.5, 3.5 Hz, H-17-pro-R), 3.71 (3H, s, CO2CH 3), 3.38 (1H, dd, J = 10.5, 8.0 Hz, H-17-pro-S), 2.53 (1H, ddd, J = 19.9, 5.1, 2.8 Hz, H-2β), 2.42 (1H, dddd, J = 14.5, 12.7, 4.7, 1.1 Hz, H-12-pro-R), 2.36 (1H, dt, J = 13.0, 3.4 Hz, H-6α), 2.31 (1H, ddt, J = 19.9, 4.9, 1.4 Hz, H-2α), 2.08 (1H, dddd, J = 14.5, 12.5, 4.8, 1.1 Hz, H-12-pro-S), 1.90 (1H, ddd, J = 15.0, 12.4, 4.9 Hz, H-11-pro-S), 1.85 (1H, dq, J = 13.2, 3.4 Hz, H-7β), 1.77 (1H, ddd, J = 15.0, 12.6, 4.6 Hz, 196 CHAPTER 5. EXPERIMENTAL.

H-11-pro-R), 1.66 (3H, s, H-19), 1.64-1.49 (2H, m, H-8 & 7α), 1.49 (2H, br s, OH), 1.45 (1H, br s, H-10), 1.19-1.13 (1H, m, H-6β), 1.18 (3H, s, H-20);

13 CNMR(CDCl3): δ 167.3 (C, C-18), 142.8 (CH, C-15), 141.4 (C, C-4), 138.4 (CH, C-16), 133.2 (CH, C-3), 124.9 (C, C-13), 110.8 (CH, C-14), 64.3 (CH, C-

1), 63.9 (CH2, C-17), 51.3 (CH3,CO2CH3), 48.7 (CH, C-10), 44.8 (CH, C-8),

39.1 (C, C-9), 38.8 (CH2, C-11), 37.97 (CH2, C-2/6), 37.96 (CH2, C-2/6), 37.1

(C, C-5), 21.9 (CH2, C-7), 21.4 (CH3, C-19), 20.9 (CH3, C-20), 18.2 (CH2, C-12);

+ + HRESIMS: [M + Na] m/z 385.1988 (calcd for C21H30O5Na , 385.1985).

5.2.11 Divinatorin C (28c).

InChI=1/C22H30O5/c1-15(23)27-14-17-8-11-22(3)18(20(24)25)5-4-6-19(22)21 (17,2)10-7-16-9-12-26-13-16/h5,9,12-13,17,19H,4,6-8,10-11,14H2,1-3H3,(H,24, 25)/t17-,19+,21-,22-/m0/s1/f/h24H

O

O H H O

O OH 28c amber resin;

TLC: See Table 5.1 on page 187.

25 [α]D −110 (c 1.1, CHCl3);

FTIR (film): ν
max 2960, 2938, 2873, 2650, 1738, 1681, 1629, 1502, 1459, 1420, 1385, 1367, 1236, 1161, 1064, 1025, 1000, 975, 873, 785, 759, 730 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.35 (1H, t, J = 1.6 Hz, H-15), 7.22 (1H, dd, J =1.6,0.8Hz,H-16),6.89(1H,dd,J = 4.8, 2.9 Hz, H-3), 6.28 (1H, dd, J 5.2. ISOLATION 197

= 1.8, 0.9 Hz, H-14), 4.26 (1H, dd, J = 11.0, 4.1 Hz, H-17-pro-R), 3.79 (1H, dd, J = 11.0, 8.4 Hz, H-17-pro-S), 2.53 (1H, dt, J = 13.2, 3.2 Hz, H-6α), 2.40 (1H, td, J = 13.8, 4.2 Hz, H-12-pro-R), 2.35 (1H, dt, J = 20.2, 5.3 Hz, H-2α),

2.24-2.16 (2H, m, H-2β &12-pro-S), 2.03 (3H, s, OCOCH 3), 1.82-1.67 (4H, m, H-1β,7β,8,11-pro-S), 1.64 (1H, ddd, J = 15.0, 12.6, 4.3 Hz, H-11-pro-R), 1.54-1.44 (2H, m, H-1α &7α), 1.42 (1H, br d, J = 12.1 Hz, H-10), 1.27 (3H, s,H-19),1.15(1H,td,J = 13.2, 3.6 Hz, H-6β), 0.83 (3H, s, H-20);

13 CNMR(CDCl3): δ 171.9 (C, C-18), 171.2 (C, OCOCH3), 142.8 (CH, C-15), 141.2 (C, C-4), 140.3 (CH, C-3), 138.5 (CH, C-16), 125.2 (C, C-13), 110.9 (CH,

C-14), 66.1 (CH2, C-17), 46.8 (CH, C-10), 40.9 (CH, C-8), 38.9 (CH2, C-11),

38.4 (C, C-9), 37.4 (C, C-5), 35.2 (CH2, C-6), 27.4 (CH2, C-2), 22.3 (CH2,

C-7), 21.0 (CH3,OCOCH3), 20.5 (CH3, C-19), 19.0 (CH3,C-20),18.3(CH2,

C-12), 17.0 (CH2,C-1);

+ + HRESIMS: [M + Na] m/z 397.1989 (calcd for C22H30O5Na , 397.1985).

5.2.12 (–)-Hardwickiic Acid (29a) and methyl ester 29b.

InChI=1/C20H28O3/c1-14-7-10-20(3)16(18(21)22)5-4-6-17(20)19(14,2)11-8- 15-9-12-23-13-15/h5,9,12-14,17H,4,6-8,10-11H2,1-3H3,(H,21,22)/t14-,17-,19+, 20+/m1/s1

O

H H

R

29a H O OR 29b Me

1 216, 217 HNMR(CDCl3) matched previously reported values; the spectrum was superimposable with that of authentic (+)-hardwickiic acid (ent-29a), prepared as detailed in Section 5.3.5 on page 202. Methylation with CH2N2 in

Et2O gave the methyl ester 29b; 198 CHAPTER 5. EXPERIMENTAL.

InChI=1/C21H30O3/c1-15-8-11-21(3)17(19(22)23-4)6-5-7-18(21)20(15,2)12- 9-16-10-13-24-14-16/h6,10,13-15,18H,5,7-9,11-12H2,1-4H3/t15-,18-,20+,21+ /m1/s1

18 [α]D −86 (c 0.04, CHCl3);

23 217 [α]D −104 (c 1.1, CHCl3) lit;

1 13 Hand CNMR(CDCl3), FTIR (film) and EIMS (70 eV) matched previ- ously reported values.218, 217 The 1H NMR spectrum was superimposable with that of authentic (+)-methyl hardwickiate (ent-29b), prepared as detailed in Section 5.3.5 on page 202.

5.2.13 Oleanolic Acid (31).

InChI=1/C30H48O3/c1-25(2)14-16-30(24(32)33)17-15-28(6)19(20(30)18-25) 8-9-22-27(5)12-11-23(31)26(3,4)21(27)10-13-29(22,28)7/h8,20-23,31H,9-18H2, 1-7H3,(H,32,33)/t20-,21-,22+,23-,27-,28+,29+,30-/m0/s1

H OH

O H HO H 31

+ + HRESIMS: [M + Na] m/z 479.3516 (calcd for C30H48O3Na , 479.3496);

1 13 221 Hand CNMR(CDCl3) matched previously reported values.

5.2.14 Presqualene Alcohol (32).

InChI=1/C30H50O/c1-23(2)13-9-15-25(5)17-11-18-27(7)21-28-29(22-31)30(28, 8)20-12-19-26(6)16-10-14-24(3)4/h13-14,17,19,21,28-29,31H,9-12,15-16,18,20,22 H2,1-8H3/b25-17+,26-19+,27-21+/t28-,29-,30-/m1/s1 5.3. SYNTHESIS 199

H

OH 32

21 [α]D +45 (c 1.2, CHCl3);

20 223 [α]D +49 (c 4.8, CHCl3) lit;

1 222 13 223 222 HNMR(CDCl3), CNMR(C6D6) and FTIR (film) matched previ- 1 ously reported values. The H NMR spectrum in C6D6 was superimposable with a previously published spectrum.223

5.2.15 Peplusol (33).

InChI=1/C30H50O/c1-24(2)13-9-15-26(5)17-11-18-28(7)21-22-30(23-31)29(8) 20-12-19-27(6)16-10-14-25(3)4/h13-14,17,19,21,30-31H,8-12,15-16,18,20,22-23 H2,1-7H3/b26-17+,27-19+,28-21+/t30-/m0/s1

OH 33

20 [α]D −6(c 0.07, CHCl3);

25 i 224 [α]D −18 (c 0.74, PrOH) lit;

1 13 Hand CNMR(CDCl3) and FTIR (film) matched previously reported val- ues.224

5.3 Synthesis

5.3.1 Salvinorin C (1c) via acetylation of salvinorin D (1d).

Ac2O (250 µL, 2.6 mmol) was added to a solution of 1d (11.0 mg, 25.4 µmol) in dry pyridine (2.5 mL) under Ar. After stirring for 3.5 h, TLC 200 CHAPTER 5. EXPERIMENTAL.

(10% acetone/CH2Cl2) indicated completion. The reaction mixture was di- luted with ice water and extracted with Et2O(× 3). The organic phase was washed with saturated NaHCO3, saturated CuSO4, water and brine. Drying

(MgSO4), evaporation in vacuo and flash column chromatography on silica gel

(55% Et2O/petrol) gave 1c as a clear resin (8.4 mg, 70%);

TLC: See Table 5.1 on page 187. Cospotted with isolated material.

16 [α]D +69 (c 0.4, CHCl3).

Other spectra (FTIR, 1Hand13C NMR) superimposable with those of the isolated material.

5.3.2 Salvinorins D (1d) and E (1e) via acetylation of 1h.

1h (2.8 mg, 7.2 µmol), Ac2O(7µL, 74 µmol) and catalytic DMAP (1mg) were stirred in pyridine (3 mL) for 3 h, when TLC (10% acetone/CH2Cl2, visualised with KMnO4) showed no starting material. Evaporation in vacuo gave a ≈1:2 mixture of 1d and 1e.

TLC: See Table 5.1 on page 187 and table on the facing page. Cospotted with isolated materials.

1H NMR spectrum superimposable with those of the isolated materials.

5.3.3 Salvinorins C (1c) and E (1e) via acetylation of 1h.

1h (4.1 mg, 10.5 µmol), Ac2O(8µL, 85 µmol) and catalytic DMAP (1 mg) were stirred in pyridine (500 µL) at 45 ◦C for 75 min, when TLC (10% acetone/CH2Cl2, visualised with KMnO4/H2SO4 dip) showed no starting ma- terial. Evaporation in vacuo and flash column chromatography (2.5 – 10% acetone/CH2Cl2)gave1c and 1e; the latter was contaminated by an insepa- rable byproduct. 5.3. SYNTHESIS 201

TLC: See Table 5.1 on page 187 and table on this page. Cospotted with isolated materials.

1H NMR spectra superimposable with those of the isolated materials.

5.3.4 Dideacetylsalvinorin C (1h) from 1c.

To a solution of 1c (5.8 mg, 12.2 µmol) and Na2CO3 (5.1 mg, 41.1 µmol) in

CH2Cl2 (1 mL) was added MeOH (1 mL), and the solution stirred at rt for 2 h, when TLC (10% acetone/ CH2Cl2) showed considerable starting material. After heating at 45 ◦C for a further 90 min, TLC indicated completion. The solution was partitioned between brine (acidified with 10% HCl) and CH2Cl2

(× 3). Drying (MgSO4), evaporation in vacuo, and flash column chromatogra- phy (loaded in CH2Cl2, eluted with 33 - 50% EtOAc/petrol, then 25% MeOH/

CH2Cl2)gave1h as a resin (4.1 mg, 86%);

InChI=1/C21H26O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7-8,10,12,14-17,22-23H,4,6,9H2,1-3H3 /t12-,14-,15-,16-,17-,20-,21-/m0/s1

O

OH O H H HO O

O O 1h

TLC: hRf 1c 1h

10% acetone/CH2Cl2 60 18

18 [α]D +27 (c 0.2, CH2Cl2);

FTIR (film): ν
max 3456, 2951, 1714, 1504, 1435, 1379, 1314, 1229, 1177, 1144, 1075, 1049, 1027, 949, 875, 788, 736, 685 cm−1; 202 CHAPTER 5. EXPERIMENTAL.

1 HNMR(CDCl3): δ 7.43 (1H, m, H-16), 7.42 (1H, t, J = 1.8 Hz, H-15), 6.48 (1H, dd, J = 2.4, 1.6 Hz, H-3), 6.40 (1H, dd, J = 1.9, 0.9 Hz, H-14), 5.60 (1H, dd, J = 11.1, 5.9 Hz, H-12), 4.32 (1H, br d,∗ J = 5.1 Hz, H-1), 4.28 (1H, dd,∗

J = 5.1, 2.4 Hz, H-2), 3.73 (3H, s, CO2CH 3), 2.49 (1H, dd, J = 13.2, 6.0 Hz, H-11α), 2.50-2.45 (1H, m, H-6α), 2.40-2.30 (2H, m, OH), 2.14-2.10 (1H, m, H-8), 2.09 (1H, dq, J = 14.6, 3.6 Hz, H-7β), 1.82 (1H, dtd, J = 15.0, 13.2, 3.3 Hz, H-7α), 1.70 (3H, s, H-19), 1.60 (1H, ddd, J = 13.0, 11.1, 0.8 Hz, H-11β), 1.47 (3H, s, H-20), 1.22 (1H, d, J = 1.0 Hz, H-10), 1.16 (1H, tdd, J = 13.3, 3.6, 0.9 Hz, H-6β);

13 CNMR(CDCl3): δ 172.0 (C, C-17), 166.6 (C, C-18), 143.9 (CH, C-15), 142.2 (C, C-4), 139.3 (CH, C-16), 135.3 (CH, C-3), 125.8 (C, C-13), 108.4 (CH, C- 14), 71.8 (CH, C-12), 69.6 (CH, C-2), 65.6 (CH, C-1), 54.1 (CH, C-10), 51.8

(CH, C-8), 51.7 (CH3,CO2CH3), 44.4 (CH2, C-11), 37.49 (C, C-5/9), 37.45

(C, C-5/9), 37.0 (CH2, C-6), 22.1 (CH3, C-19), 18.4 (CH2, C-7), 16.3 (CH3, C-20);

+ + HRESIMS: [M + Na] m/z 413.1588 (calcd for C21H26O7Na , 413.1571).

5.3.5 (+)-Hardwickiic acid (ent-29a).

Following a modified version of Costa et al’s procedure,218 the acid fraction of copaiba balsam was methylated with CH2N2 in Et2O, and the methyl ester (ent-29b) was isolated. This ester (32 mg, 97 µmol) was dissolved in acetone.

KF/Al2O3 (220 mg, 40% w/w KF) was added and the acetone evaporated under reduced pressure. This powder was irradiated in a microwave oven (650 W) at 100% power for 8 minutes, then cooled. Minimal water was added and stirred for 5 minutes, then filtered, and the filter cake was rinsed with water

(× 2). Rinsing of the filter cake with CHCl3 (× 3) gave, after drying (MgSO4) and evaporation, starting material ent-29b (11 mg). The aqueous filtrate was acidified with 10% HCl and extracted with CHCl3 (× 4). The pooled organic extracts were dried (MgSO4) and evaporated to give 14 mg crude product.

∗ After D2O exchange. 5.3. SYNTHESIS 203

Flash column chromatography in 80% Et2O/petrol gave ent-29a (5 mg, 16 µmol) as a semicrystalline film;

InChI=1/C20H28O3/c1-14-7-10-20(3)16(18(21)22)5-4-6-17(20)19(14,2)11-8- 15-9-12-23-13-15/h5,9,12-14,17H,4,6-8,10-11H2,1-3H3,(H,21,22)/t14-,17-,19+, 20+/m0/s1/f/h21H

O

H H

R ent-29a H O OR ent-29b Me

TLC: hRf ent-29a ent-29b

5% acetone/CH2Cl2 31 76 19 [α]D +81 (c 0.2, CHCl3);

23 217 [α]D −85 (CHCl3) lit. value for 29a;

FTIR, 1Hand13CNMRdata217 matched reported values.

5.3.6 Salvinorin A lactol (35).

Following the published procedure,260 1a (15.8 mg, 36.5 µmol) was warmed in ◦ i THF (1 mL) under Ar until fully dissolved, then cooled to -78 C. Bu2AlH (1MinTHF,0.5mL,500µmol) was added dropwise. The solution was stirred for 25 minutes, then quenched (sat. aq. NH4Cl dropwise), evaporated in vacuo until thick, diluted in water and extracted into Et2O(× 3). Washing (brine), drying (MgSO4), evaporation in vacuo, and flash column chromatography (6-

260 10% acetone/CH2Cl2 gradient) gave 35 as a clear resin (10.4 mg, 65% (81% borsm));

InChI=1/C23H30O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8,11,14-17,19,21,27H,5,7,9-10H 2,1-4H3/t14-,15-,16-,17-,19-,21u,22-,23-/m0/s1 204 CHAPTER 5. EXPERIMENTAL.

O

O O O H H O OH

O O 35

TLC: hRf 1a 35

10% acetone/CH2Cl2 57 25

FTIR (film): ν
max 3446, 2953, 2256, 1730, 1503, 1438, 1379, 1272, 1237, 1203, 1161, 1123, 1053, 1010, 978, 914, 875, 785, 732, 647 cm−1;

1 HNMR,major(17β) anomer (CDCl3): δ 7.36 (1H, br s, H-16), 7.34 (1H, t, J = 1.8 Hz, H-15), 6.38 (1H, dd, J = 1.8, 0.9 Hz, H-14), 5.14-5.09 (1H, m, H-2), 4.87 (1H, dd, J = 11.6, 2.4 Hz, H-12), 4.80 (1H, d, J = 8.7 Hz, H-17),

3.70 (3H, s, CO2CH 3), 2.76-2.72 (1H, m, H-4), 2.28-2.23 (2H, m, H-3), 2.14

(3H, s, OCOCH 3), 2.11 (1H, dd, J = 13.2, 2.4 Hz, H-11α), 2.07 (1H, d, J = 0.9Hz,H-10),1.80(1H,dq,J = 13.9, 3.3 Hz, H-7β), 1.70 (1H, dt, J = 13.5, 3.2 Hz, H-6α), 1.64-1.57 (1H, m, H-6β), 1.38 (3H, s, H-20), 1.43-1.32 (1H, m, H-7α), 1.21 (1H, ddd, J = 13.2, 11.6, 0.9 Hz, H-11β), 1.15-1.09 (1H, m, H-8), 1.08 (3H, s, H-19); these data are broadly consistent with incomplete data published previously;260

13 CNMR,major(17β) anomer (CDCl3): δ 202.5 (C, C-1), 171.9 (C, C-18),

169.9 (C, OCOCH3), 143.0 (CH, C-15), 139.1 (CH, C-16), 126.2 (C, C-13), 108.8 (CH, C-14), 94.2 (CH, C-17), 75.0 (CH, C-2), 66.2 (CH, C-12), 65.4

(CH, C-10), 53.6 (CH, C-4), 52.1 (CH, C-8), 51.8 (CH3,CO2CH3), 44.7 (CH2,

C-11), 42.4 (C, C-5), 38.8 (CH2, C-6), 35.6 (C, C-9), 30.8 (CH2, C-3), 20.6

(CH3,OCOCH3), 17.6 (CH2, C-7), 16.7 (CH3,C-19),15.0(CH3,C-20). 5.3. SYNTHESIS 205

5.3.7 (4R)-3,4-Dihydrosalvinorin C (36c).

Following a published procedure,27 formation of the orthoacetate of 36h (which did not go to completion in 10 h), followed by acid-catalyzed hydrolysis, gave 36d which was used in the next reaction without purification;

InChI=1/C23H30O8/c1-12(24)30-18-16(25)9-15(20(26)28-4)22(2)7-5-14-21(27) 31-17(13-6-8-29-11-13)10-23(14,3)19(18)22/h6,8,11,14-19,25H,5,7,9-10H2,1-4H3 /t14-,15-,16-,17-,18-,19-,22-,23-/m0/s1

O

O

O O H H HO O

O O 36d

TLC: hRf 36h Orth.Ac. 36d

Et2O336029

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.41 (1H, t, J =1.7Hz,H-15), 6.41 (1H, br d, J = 2.0 Hz, H-14), 5.61 (1H, br s, H-1), 5.47 (1H, dd, J = 11.5,

5.3 Hz, H-12), 3.76-3.72 (1H, m, H-2), 3.69 (3H, s, CO2CH 3), 2.49 (1H, dd, J

= 13.1, 5.6 Hz, H-11α), 2.22-2.08 (3H, m), 2.16 (3H, s, OCOCH 3), 1.86-1.77 (3H, m), 1.71-1.61 (2H, m), 1.35 (3H, s, H-20), 1.35-1.24 (1H, m), 1.18 (3H, s, H-19),1.13(1H,d,J =1.9Hz,H-10);

1 27 HNMR[(CD3)2CO] matched values previously reported for “CDCl3”.

27 Acetylation of 36d with Ac2O/pyridine and purification by HPLC (40% EtOAc/ petrol) gave 36c;

InChI=1/C25H32O9/c1-13(26)32-18-10-17(22(28)30-5)24(3)8-6-16-23(29)34- 19(15-7-9-31-12-15)11-25(16,4)21(24)20(18)33-14(2)27/h7,9,12,16-21H,6,8,10- 11H2,1-5H3/t16-,17-,18-,19-,20-,21-,24-,25-/m0/s1 206 CHAPTER 5. EXPERIMENTAL.

O

O O O O H H O O

O O 36c

TLC: hRf 36d 36c

10% acetone/CH2Cl2 33 77 (developed in KMnO4).

HPLC: tR = 18.5 min (40% EtOAc/petrol).

1 HNMR(CDCl3): δ 7.45 (1H, br s, H-16), 7.41 (1H, t, J = 1.8 Hz, H-15), 6.41 (1H, dd, J = 2.0, 0.9 Hz, H-14), 5.67 (1H, dt, J = 3.5, 1.5 Hz, H-1), 5.46 (1H, dd, J = 11.6, 5.5 Hz, H-12), 4.79-4.74 (1H, m, H-2), 3.70 (3H, s, CO2CH 3), 2.43 (1H, dd, J = 13.3, 5.4 Hz, H-11α), 2.31-2.24 (2H, m), 2.16-2.09 (2H, m),

2.14 (3H, s, OCOCH 3), 1.98 (3H, s, OCOCH 3), 1.84-1.79 (2H, m), 1.72-1.53

(m, obscured by H2O), 1.69 (1H, ddd, J = 13.4, 11.6, 0.9 Hz, H-11β), 1.40-1.31 (2H,m),1.38(3H,s,H-20),1.20(1H,d,J = 1.9 Hz, H-10), 1.17 (3H, s, H-19); these data are broadly consistent with those published previously82 (assign-

1 23 13 27 ments differ). HNMR[(CD3)2CO] and CNMR(CDCl3) matched pre- viously reported values.

5.3.8 (4R)-3,4-Dihydrosalvinorin E (36e).

23 27 Following the published procedure, acetylation in Ac2O/pyridine of 36h (which had been purified by HPLC in EtOAc) gave 36e;

InChI=1/C23H30O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8,11,14-19,25H,5,7,9-10H2,1-4 H3/t14-,15-,16-,17-,18-,19-,22-,23-/m0/s1 5.3. SYNTHESIS 207

O

O OH O H H O O

O O 36e

TLC: hRf 36h 36e

50% EtOAc/petrol 20 48 (developed in KMnO4).

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.41 (1H, dd, J = 1.8, 0.9 Hz, H-14), 5.55 (1H, dd, J = 11.4, 5.4 Hz, H-12), 4.70 (1H, ddd, J = 11.7, 4.6, 3.2 Hz, H-2), 4.29 (1H, br s, H-1), 3.68 (3H, s,

CO2CH 3), 2.41 (1H, dd, J = 13.2, 5.4 Hz, H-11α), 2.31 (1H, q, J = 12.6 Hz, H-3α), 2.20(1H, dd, J = 13.2, 2.4 Hz, H-4), 2.14-2.08 (1H, m, H-8), 2.09 (3H, s, OCOCH 3), 1.90 (1H, br s, w 1 =11Hz,OH), 1.82 (1H, dddd, J = 12.5, 2 4.9, 2.5, 1.1 Hz, H-3β), 1.76 (1H, dq, J = 13.3, 3.2 Hz, H-7β), 1.74-1.69 (1H, m, H-6a), 1.64 (1H, ddd, J = 13.0, 11.6, 0.9 Hz, H-11β), 1.45 (3H, s, H-20), 1.38 (3H, s, H-19), 1.38-1.26 (2H, m, H-6b,7α), 1.00 (1H, br s, H-10); these data are consistent with incomplete data published previously.23 13C

27 NMR (CDCl3) matched previously reported values.

5.3.9 (4R)-Dideacetyl-3,4-dihydrosalvinorin C (36h).

A slight modification of a published procedure was used:23, 260 to 1a (37.6 mg,

86.9 µmol) and NaBH4 (4.4 mg, 116 µmol) was added CH2Cl2 (200 µL), fol- lowed by EtOH (3 mL), and the cloudy solution stirred under Ar at 40 ◦C.

After 4 h, TLC (Et2O) indicated completion. The solution was cooled to 0 ◦ C, and 0.5% H2SO4/MeOH added dropwise until effervescence ceased. The solution was concentrated to ≈ 500 µL in vacuo, then partitioned between brine (acidified with 10% HCl) and CH2Cl2 (× 3). Drying (MgSO4), evapora- tion in vacuo and flash column chromatography (40-66% EtOAc/petrol) gave 36h15, 23 (19.1 mg, 56%); 208 CHAPTER 5. EXPERIMENTAL.

InChI=1/C21H28O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7,10,12-17,22-23H,4,6,8-9H2,1-3H3/t 12-,13-,14-,15-,16-,17-,20-,21-/m0/s1

O

OH O H H HO O

O O 36h

TLC: hRf 1a 36h 38h

Et2O451425

HPLC: tR =8.7min(EtOAc).

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.42 (1H, t, J =1.7Hz,H-15), 6.41 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.56 (1H, ddd, J = 11.5, 6.1, 0.9 Hz,

H-12), 4.20 (1H, br s, H-1), 3.68 (3H, s, CO2CH 3), 3.58 (1H, ddd, J = 11.6, 5.2, 3.4 Hz, H-2), 2.48 (1H, dd, J = 13.2, 5.6 Hz, H-11α), 2.25-2.07 (4H, m),

1.91 (2H, br s, w 1 =70Hz,OH), 1.77-1.69 (3H, m), 1.62 (1H, ddd, J = 13.3, 2 11.6, 1.0 Hz, H-11β), 1.46 (3H, s, H-20), 1.37 (3H, s, H-19), 1.29 (1H, tdd, J = 13.6, 3.1, 1.0 Hz, H-6α), 0.92 (1H, d, J = 1.9 Hz, H-10);

1 23 HNMR([CD3]2CO) matched previously reported values.

5.3.10 8-epi-Salvinorin A (37a).

Distilled DMPU (60 ◦C / 0.1 mmHg) was added to 1a (21.4 mg, 49.5 µmol) ◦ and NaHCO3 (30.1 mg, 358 µmol), and stirred at 150 C for 2 h. The amber solution was diluted with EtOAc, neutralised dropwise with 10% HCl, and washed (10% HCl × 4, then brine). Drying (MgSO4) and evaporation in vacuo followed by flash column chromatography (30% - 50% EtOAc/petrol gradient) monitored by TLC (Et2O) gave 37a as a clear resin (10.8 mg, 51% (81% borsm)); 5.3. SYNTHESIS 209

InChI=1/C23H28O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8,11,14-17,19H,5,7,9-10H2,1-4 H3/t14-,15+,16+,17+,19+,22+,23+/m1/s1

O

O O O H H O 8 O

O O 37a

TLC: hRf 1a 37a

Et2O3950

vanillin/H2SO4 purple blue

13 [α]D −53 (c 0.6, CHCl3);

FTIR (film): ν
max 3018, 2951, 2883, 1732, 1504, 1450, 1437, 1375, 1323, 1238, 1202, 1161, 1124, 1084, 1047, 1024, 997, 970, 937, 876, 783, 756, 667, 601 cm−1;

1 HNMR(CDCl3): δ 7.43 (1H, br s, H-16), 7.38 (1H, t, J =1.7Hz,H-15), 6.37 (1H, d, J = 1.7 Hz, H-14), 5.25 (1H, dd, J = 12.0, 2.2 Hz, H-12), 5.09

(1H, ∼dd, J ≈ 10.9, 9.3 Hz, H-2), 3.69 (3H, s, CO2CH 3), 2.79-2.72 (1H, m, H-4), 2.45 (1H, dd, J = 5.0, 2.2 Hz, H-8), 2.36 (1H, dd, J = 15.0, 2.2 Hz, H-11α), 2.29-2.22 (2H, m, H-3), 2.24 (1H, br s, H-10), 2.19 (1H, dq, J = 14.3,

3.1 Hz, H-7β), 2.15 (3H, s, OCOCH 3), 2.00 (1H, td, J = 13.7, 3.9 Hz, H-6β), 1.83 (1H, tdd, J = 14.2, 5.0, 3.9 Hz, H-7α), 1.62 (3H, s, H-20), 1.54 (1H, dt, J = 13.7, 3.4 Hz, H-6α), 1.50 (1H, dd, J = 15.0, 12.0 Hz, H-11β), 1.07 (3H, s, H-19);

13 CNMR(CDCl3): δ 202.3 (C, C-1), 173.4 (C, C-17), 171.8 (C, C-18), 169.8

(C, OCOCH3), 143.6 (CH, C-15), 139.7 (CH, C-16), 123.3 (C, C-13), 108.5 (CH, C-14), 75.2 (CH, C-2), 70.1 (CH, C-12), 64.1 (CH, C-10), 52.9 (CH, C-

4), 51.8 (CH3,CO2CH3), 48.0 (CH2, C-11), 45.2 (CH, C-8), 42.2 (C, C-5),

34.7 (C, C-9), 33.9 (CH2, C-6), 30.6 (CH2, C-3), 24.6 (CH3, C-20), 20.5 (CH3,

OCOCH3), 17.6 (CH2, C-7), 15.2 (CH3,C-19); 210 CHAPTER 5. EXPERIMENTAL.

+ + HRESIMS: [M + Na] m/z 455.1683 (calcd for C23H28O8Na , 455.1676).

5.3.11 8-epi-Salvinorin B (37b).

MeOH (2 mL) was refluxed under Ar for 15 minutes, and cooled to rt. 1a (32.3 mg, 74.7 µmol) and Na2CO3 (30.4 mg, 245 µmol) were added, and the resulting suspension stirred under Ar at rt for 4.5 h, when TLC indicated completion. The reaction was quenched with 10% HCl, then partitioned between 10% HCl and CH2Cl2 (× 4). Drying (MgSO4) and evaporation in vacuo gave an off- white powder. Addition of minimal MeOH (≈ 0.3 mL) and centrifugation gave 1b as a white powder (21.5 mg, 74%). The supernatant was evaporated under reduced pressure to give 37b as an amber resin (6.6 mg, 23%);

InChI=1/C21H26O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7,10,12-15,17,22H,4,6,8-9H2,1-3H3/t 12-,13+,14+,15+,17+,20+,21+/m1/s1

O

O O H H HO 8 O

O O 37b

TLC: hRf 1a 1b 37b

Et2O402440 50% EtOAc/petrol 37 23 37

10% acetone/CH2Cl2 69 52 54

vanillin/H2SO4 purple purple blue

13 [α]D −27 (c 0.3, CHCl3);

FTIR (film): ν
max 3475, 2952, 1728, 1504, 1438, 1386, 1274, 1201, 1156, 1120, 1051, 1026, 997, 970, 876, 786, 755 cm−1; 5.3. SYNTHESIS 211

1 HNMR(CDCl3): δ 7.44 (1H, m, H-16), 7.40 (1H, t, J = 1.8 Hz, H-15), 6.37 (1H, dd, J = 1.9, 0.8 Hz, H-14), 5.29 (1H, dd, J = 11.8, 2.2 Hz, H-12), 4.00

(1H, ddd, J = 11.9, 7.6, 1.3 Hz, H-2), 3.69 (3H, s, CO2CH 3), 2.70 (1H, dd, J = 13.5, 3.2 Hz, H-4), 2.47 (1H, dd, J = 4.8, 2.2 Hz, H-8), 2.44 (1H, ddd, J = 13.5, 7.6, 3.2 Hz, H-3β), 2.42 (1H, dd, J = 14.8, 2.2 Hz, H-11α), 2.23 (1H, d, J = 1.3 Hz, H-10), 2.21 (1H, dq, J = 14.4, 3.1 Hz, H-7β), 2.01 (1H, td, J = 13.5, 11.9 Hz, H-3α), 1.99 (1H, td, J = 13.8, 0.8 Hz, H-6β), 1.85 (1H, tdd, J = 14.0, 4.8, 3.7 Hz, H-7α), 1.65 (3H, s, H-20), 1.54 (1H, dt, J = 13.5, 3.5 Hz, H-6α), 1.46 (1H, ddd, J = 14.8, 11.7, 0.7 Hz, H-11β), 1.06 (3H, s, H-19); these data are inconsistent with those of Harding et al82 and Lee et al.271

13 CNMR(CDCl3): δ 209.1 (C, C-1), 173.4 (C, C-17), 172.1 (C, C-18), 143.7 (CH, C-15), 139.6 (CH, C-16), 123.5 (C, C-13), 108.4 (CH, C-14), 74.5 (CH, C-

2), 70.0 (CH, C-12), 63.7 (CH, C-10), 52.4 (CH, C-4), 51.7 (CH3,CO2CH3),

48.3 (CH2, C-11), 45.3 (CH, C-8), 42.7 (C, C-5), 34.6 (C, C-9), 34.3 (CH2,

C-3), 33.9 (CH2, C-6), 24.7 (CH3, C-20), 17.6 (CH2, C-7), 15.3 (CH3, C-19);

+ + HRESIMS: [M + Na] m/z 413.1573 (calcd for C21H26O7Na , 413.1571).

5.3.12 8-epi-Salvinorin C (37c).

37d (1.7 mg, 3.9 µmol), Ac2O (0.25 mL) and catalytic DMAP (1mg) were stirred in pyridine (0.5 mL) at 50 ◦C for 90 min, when TLC (10% acetone/CH2Cl2, visualised with KMnO4/H2SO4 dip) showed no starting ma- terial. In a separate flask, 37e ( 2.8 mg, 6.5 µmol) was subjected to the same conditions for 4 h, when TLC (same conditions) showed no starting material. The crude products, which cospotted by TLC, were pooled and evaporated in vacuo. Flash column chromatography (2-4% acetone/CH2Cl2)gave37c (4.7 mg, 95%) as a colourless resin;

InChI=1/C25H30O9/c1-13(26)32-18-10-17(22(28)30-5)24(3)8-6-16-23(29)34- 19(15-7-9-31-12-15)11-25(16,4)21(24)20(18)33-14(2)27/h7,9-10,12,16,18-21H,6, 8,11H2,1-5H3/t16-,18+,19+,20+,21+,24+,25+/m1/s1 212 CHAPTER 5. EXPERIMENTAL.

O

O O O O H H O 8 O

O O 37c

TLC: hRf 37c 37d 37e

3% acetone/CH2Cl2 35826

vanillin/H2SO4 blue blue blue

23 [α]D +9 (c 0.2, CH2Cl2);

FTIR (film): ν
max 2926, 2854, 1741, 1716, 1645, 1556, 1504, 1435, 1371, 1332, 1229, 1198, 1172, 1159, 1093, 1032, 962, 875, 802 cm−1;

1 HNMR(CDCl3): δ 7.50 (1H, dd, J = 1.6, 0.8 Hz, H-16), 7.43 (1H, t, J = 1.7 Hz, H-15), 6.42 (2H, m, H-3 & 14), 5.59 (1H, ddd, J = 4.7, 1.3, 0.9 Hz, H-1), 5.49 (1H, dd, J = 4.8, 2.3 Hz, H-2), 5.28 (1H, br d, J = 11.0 Hz, H-12),

3.72 (3H, s, CO2CH 3), 2.49 (1H, dd, J = 5.3, 2.3 Hz, H-8), 2.33 (1H, dt, J = 13.4, 3.4 Hz, H-6α), 2.26-2.16 (1H, m, H-7β), 2.14 (1H, dd, J = 14.3, 1.6 Hz,

H-11α), 2.10 (3H, s, OCOCH 3), 2.02 (3H, s, OCOCH 3), 1.97 (1H, tdd, J =

14.8, 5.5, 3.9 Hz, H-7α), 1.70 (3H, s, H-19), 1.61-1.50 (m, H-11β,6β,H2O), 1.35 (3H, s, H-20), 1.25 (1H, br s, H-10);

13 CNMR(CDCl3): δ 173.6, 170.5, 169.7, 165.7, 143.7, 142.6, 139.7, 132.1, 123.4, 108.4, 69.7, 69.6, 65.1, 51.74, 51.72, 48.9, 45.6, 38.2, 36.0, 33.5, 25.1, 21.14, 21.13, 20.6, 18.2;

+ + HRESIMS: [M + Na] m/z 497.1788 (calcd for C25H30O9Na , 497.1782).

5.3.13 8-epi-Salvinorin D (37d).

Standing 37e in CDCl3 at rt overnight gave a 1:1 mixture with 37d,which was isolated by flash column chromatography (5-10% acetone/CH2Cl2) as an amber resin; 5.3. SYNTHESIS 213

InChI=1/C23H28O8/c1-12(24)30-18-16(25)9-15(20(26)28-4)22(2)7-5-14-21(2 7)31-17(13-6-8-29-11-13)10-23(14,3)19(18)22/h6,8-9,11,14,16-19,25H,5,7,10 H2,1-4H3/t14-,16+,17+,18+,19+,22+,23+/m1/s1

O

O

O O H H HO 8 O

O O 37d

TLC: See table on the preceding page.

23 [α]D +15 (c 0.1, CH2Cl2);

FTIR (film): ν
max 3402, 2953, 2926, 1738, 1716, 1503, 1462, 1440, 1382, 1368, 1333, 1302, 1228, 1199, 1173, 1158, 1099, 1055, 1025, 981, 875, 842, 798, 784, 732 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.49 (1H, br s, H-16), 7.43 (1H, t, J =1.8 Hz, H-15), 6.49 (1H, dd, J = 2.3, 1.5 Hz, H-3), 6.41 (1H, dd, J = 2.0, 0.8 Hz, H-14), 5.53 (1H, br d, J = 5.0 Hz, H-1), 5.29 (1H, br d, J = 11.5 Hz, H-12),

4.41 (1H, dd, J = 5.1, 1.4 Hz, H-2), 3.72 (3H, s, CO2CH 3), 2.48 (1H, dd, J = 5.3, 2.3 Hz, H-8), 2.30 (1H, dt, J = 13.5, 3.8 Hz, H-6α), 2.22 (1H, dtd, J = 14.7, 3.5, 2.1 Hz, H-7β), 2.17 (1H, dd, J = 14.1, 1.5 Hz, H-11α), 2.13 (3H, s, OCOCH 3), 1.96 (1H, tdd, J = 14.2, 5.3, 3.9 Hz, H-7α), 1.70-1.64 (1H, m,

H-11β), 1.67 (3H, s, H-19), 1.62-1.52 (m, H-6β,H2O),1.37(3H,s,H-20),1.25 (1H, br s, H-10);

13 CNMR(CDCl3): δ 173.7, 171.8, 166.1, 143.7, 141.4, 139.7, 135.0, 123.6, 108.4, 69.6, 69.3, 67.6, 51.7, 51.6, 49.1, 45.7, 37.8, 36.0, 33.5, 25.2, 21.3, 21.0, 18.1;

+ + HRESIMS: [M + Na] m/z 455.1681 (calcd for C23H28O8Na , 455.1676). 214 CHAPTER 5. EXPERIMENTAL.

5.3.14 8-epi-Salvinorin E (37e).

37h (6.6 mg, 16.9 µmol) was dissolved in pyridine (2 mL) and Ac2O(250

µL), and stirred at rt for 3 h, when TLC (10% acetone/CH2Cl2)showedno starting material. H2O (10 mL) was added, and the solution extracted with

Et2O(× 2). The pooled organic layers were washed (sat. NaHCO3,sat.

CuSO4, and brine). Drying (MgSO4), evaporation in vacuo and flash column chromatography (80% Et2O/petrol) gave 37e (5.3 mg, 72%) as an amber resin;

InChI=1/C23H28O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h6,8-9,11,14,16-19,25H,5,7,10H2,1 -4H3/t14-,16+,17+,18+,19+,22+,23+/m1/s1

O

O OH O H H O 8 O

O O 37e

TLC: See tables on page 212 and on the next page.

22 [α]D +14 (c 0.1, CH2Cl2);

FTIR (film): ν
max 3514, 2951, 1737, 1722, 1503, 1461, 1435, 1373, 1322, 1257, 1227, 1199, 1155, 1126, 1092, 1047, 1031, 931, 900, 837, 843, 806, 780, 735 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.49 (1H, br s, H-16), 7.42 (1H, t, J =1.9Hz, H-15), 6.41 (1H, dd, J = 2.0, 1.0 Hz, H-14), 6.39 (1H, dd, J = 2.4, 1.5 Hz, H-3), 5.35 (1H, dd, J = 4.7, 2.3 Hz, H-2), 5.30 (1H, br d, J = 11.5 Hz, H-12),

4.29 (1H, dd, J = 4.4, 1.7 Hz, H-1), 3.71 (3H, s, CO2CH 3), 2.49 (1H, dd, J = 5.2, 2.1 Hz, H-8), 2.36 (1H, dt, J = 13.3, 3.6 Hz, H-6α), 2.21 (1H, dtd, J =

14.5, 3.7, 2.1 Hz, H-7β), 2.14 (3H, s, OCOCH 3), 2.12 (1H, dd, J = 13.7, 1.5 Hz, H-11α), 1.99 (1H, tt, J = 14.3, 4.4 Hz, H-7α), 1.89 (1H, dd, J = 2.5, 1.3 5.3. SYNTHESIS 215

Hz, OH), 1.70 (3H, s, H-19), 1.64 (1H, dd, J = 13.9, 11.5 Hz, H-11β), 1.62 (3H, s, H-20), 1.54 (1H, td, J = 13.1, 3.4 Hz, H-6β), 1.30 (1H, t, J =0.9Hz, H-10);

13 CNMR(CDCl3): δ 174.0, 169.7, 166.0, 143.7, 143.4, 139.6, 131.2, 123.8, 108.4, 72.5, 69.6, 65.0, 53.1, 51.6, 49.4, 45.6, 37.9, 36.2, 33.4, 25.8, 21.2, 21.0, 18.1;

+ + HRESIMS: [M + Na] m/z 455.1675 (calcd for C23H28O8Na , 455.1676).

5.3.15 8-epi-Dideacetylsalvinorin C (37h).

1c (11.0 mg, 23.2 µmol) and KCN∗ (7.5 mg, 115 µmol) were dissolved in MeOH

(1.5 mL) and stirred at rt for 30 min, when TLC (10% acetone/CH2Cl2)showed no starting material. The solution was evaporated in vacuo, and the residue partitioned between H2OandCH2Cl2 (× 3). Recrystallisation from acetone, repeated on the mother liquor, gave 37h (6.6 mg, 73%) as colourless crystals;

InChI=1/C21H26O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7-8,10,12,14-17,22-23H,4,6,9H2,1-3H3 /t12-,14+,15+,16+,17+,20+,21+/m1/s1

O

OH O H H HO 8 O

O O 37h

TLC: hRf 1c 37h 1h

10% acetone/CH2Cl2 63 23 18 50% EtOAc/petrol - 28 19

vanillin/H2SO4 purple blue purple

∗ Use of this poisonous reagent is not recommended. K2CO3 and other bases are equally effective. 216 CHAPTER 5. EXPERIMENTAL. mp (from acetone) 243-245 ◦C;

22 [α]D +15 (c 0.04, CH2Cl2);

FTIR (film): ν
max 3461, 2952, 2926, 1716, 1507, 1459, 1437, 1378, 1259, 1228, 1200, 1180, 1156, 1095, 1052, 875, 803 cm−1;

1 HNMR(CDCl3): δ 7.48 (1H, m, H-16), 7.42 (1H, t, J = 1.6 Hz, H-15), 6.44 (1H, dd, J = 2.4, 1.6 Hz, H-3), 6.40 (1H, dd, J = 1.9, 0.9 Hz, H-14), 5.30 (1H, br d,J= 11.5 Hz, H-12), 4.25 (1H, dd, J = 4.9, 2.4 Hz, H-2), 4.16 (1H, ddd,

J = 4.9, 1.6, 0.7 Hz, H-1), 3.71 (3H, s, CO2CH 3), 2.48 (1H, dd, J = 5.5, 2.2 Hz, H-8), 2.23 (1H, dt, J = 13.5, 2.2 Hz, H-6α), 2.23-2.17 (1H, m, H-7β), 2.15 (1H, dd, J = 13.9, 1.5 Hz, H-11α), 1.98 (1H, tdd, J = 14.5, 5.3, 3.8 Hz, H-7α),

1.81 (2H, br s, w 1 ≈ 50 Hz, OH), 1.68 (3H, s, H-19), 1.64 (3H, s, H-20), 1.62 2 (1H, dd, J = 13.8, 12.1 Hz, H-11β), 1.51 (1H, td, J = 12.9, 3.8 Hz, H-6β), 1.21 (1H, br s, H-10);

13 CNMR(CDCl3): δ 174.2, 166.4, 143.7, 142.2, 139.6, 134.9, 123.9, 108.4, 69.9, 69.6, 66.5, 53.2, 51.6, 49.4, 45.6, 37.6, 36.2, 33.4, 26.0, 21.3, 18.1;

+ + HRESIMS: [M + Na] m/z 413.1571 (calcd for C21H26O7Na , 413.1571).

5.3.16 Salvinorin B formate (46).

◦ A mixture of Ac2O (0.25 mL) and HCO2H (0.7 mL) was stirred at 45 C for 40 minutes. Meanwhile 1b (18.0 mg, 46.1 µmol) was added to dry pyri- dine (1 mL), warmed to 45 ◦C until fully dissolved, then cooled to 0 ◦C. The cooled anhydride mixture was added dropwise by pipette, with violent bubbling. The solution was warmed to room temperature and stirred for 30 minutes, when TLC (20% acetone/CH2Cl2, visualised in KMnO4) indicated completion. The reaction mixture was cooled to 0 ◦C, diluted dropwise with water, and extracted into EtOAc. The organic layer was washed (1% HCl, water, 5% NaHCO3 and brine). Drying (MgSO4), evaporation in vacuo and

flash column chromatography (2-4% MeOH/CH2Cl2 gradient) gave 46 as a clear resin (13.8 mg, 72%); 5.3. SYNTHESIS 217

InChI=1/C22H26O8/c1-21-6-4-13-20(26)30-16(12-5-7-28-10-12)9-22(13,2)18 (21)17(24)15(29-11-23)8-14(21)19(25)27-3/h5,7,10-11,13-16,18H,4,6,8-9H2,1-3 H3/t13-,14-,15-,16-,18-,21-,22-/m0/s1

O

O O O H H O O

O O 46

TLC: hRf 1b 46

20% acetone/CH2Cl2 65 85

26 [α]D −54 (c 0.6, CH2Cl2);

FTIR (film): ν
max 3147, 2952, 2854, 1726, 1504, 1451, 1439, 1396, 1372, 1278, 1224, 1163, 1108, 1072, 1024, 1008, 941, 899, 875, 786, 735, 701 cm−1;

1 HNMR(CDCl3): δ 8.14 (1H, s, CHO),7.41(1H,brs,H-16),7.40(1H,t,J = 1.8 Hz, H-15), 6.38 (1H, dd, J = 1.8, 0.9 Hz, H-14), 5.54 (1H, dd, J = 11.7, 5.2

Hz, H-12), 5.25 (1H, ddt, J = 11.2, 8.7, 1.1 Hz, H-2), 3.73 (3H, s, CO2CH 3), 2.79-2.74 (1H, m, H-4), 2.51 (1H, dd, J = 13.5, 5.2 Hz, H-11α), 2.37-2.32 (2H, m, H-3), 2.18 (1H, d, J = 0.8 Hz, H-10), 2.19-2.15 (1H, m, H-7α), 2.08 (1H, dd, J = 11.5, 3.0 Hz, H-8), 1.80 (1H, ∼dt, J ≈ 12.9, 2.9 Hz, H-6α), 1.65 (1H, tdd, J = 13.3, 11.8, 3.2 Hz, H-7β), 1.63-1.58 (1H, m, H-6β), 1.60 (1H, ddd, J = 13.5, 11.7, 0.8 Hz, H-11β), 1.46 (3H, s, H-20), 1.13 (3H, s, H-19);

13 CNMR(CDCl3): δ 200.8 (C, C-1), 171.4 (C, C-18), 171.0 (C, C-17), 159.4 (CH, CHO), 143.7 (CH, C-15), 139.4 (CH, C-16), 125.1 (C, C-13), 108.3 (CH, C-14),74.5(CH,C-2),72.0(CH,C-12),64.1(CH,C-10),53.5(CH,C-4),52.0

(CH3,CO2CH3), 51.3 (CH, C-8), 43.4 (CH2, C-11), 42.1 (C, C-5), 38.1 (CH2,

C-6), 35.5 (C, C-9), 30.6 (CH2, C-3), 18.1 (CH2, C-7), 16.4 (CH3, C-19), 15.2

(CH3, C-20);

+ + HRESIMS: [M + Na] m/z 441.1525 (calcd for C22H26O8Na , 441.1520). 218 CHAPTER 5. EXPERIMENTAL.

5.3.17 Dideacetylsalvinorin C 2-O-(4-bromobenzoate) (47).

Toasolutionof1h (3.8 mg, 9.7 µmol), 4-bromobenzoyl chloride (8.7 mg, 39.6

µmol) and DMAP (3.1 mg, 25.4 µmol) in dry CH2Cl2 (1.5 mL) was added

NEt3 (100 µL). The bright yellow solution was stirred under Ar for 90 min, when TLC (Et2O, visualised in KMnO4) indicated completion. The solution was diluted in Et2O, washed with 10% HCl (× 3), sat. NaHCO3 and brine.

Drying (MgSO4) and evaporation in vacuo gave 47 as a white powder (5.0 mg, 90%);

InChI=1/C28H29BrO8/c1-27-10-8-18-26(33)37-21(16-9-11-35-14-16)13-28(18, 2)23(27)22(30)20(12-19(27)25(32)34-3)36-24(31)15-4-6-17(29)7-5-15/h4-7,9,11 -12,14,18,20-23,30H,8,10,13H2,1-3H3/t18-,20-,21-,22-,23-,27-,28-/m0/s1

O

O OH O H H O O Br

O O 47

TLC: hRf 1h 47

10% acetone/CH2Cl2 40 78

◦ mp (from cold Et2O) 161-166 C;

18 [α]D +86 (c 0.4, CH2Cl2);

FTIR (film): ν
max 3509, 2950, 1717, 1590, 1504, 1484, 1434, 1398, 1269, 1233, 1175, 1143, 1104, 1071, 1012, 946, 875, 849, 786, 758, 679 cm−1;

1 HNMR(CDCl3): δ 7.93 (2H, d, J = 8.6 Hz, H-2’), 7.62 (2H, d, J =8.6Hz, H-3’), 7.44 (1H, br s, H-16), 7.42 (1H, t, J = 1.9 Hz, H-15), 6.54 (1H, t, J = 2.0Hz,H-3),6.42(1H,dd,J = 1.9, 0.9 Hz, H-14), 5.63 (1H, dd, J = 4.8, 2.4 Hz, H-2), 5.59 (1H, dd, J = 10.9, 5.8 Hz, H-12), 4.61 (1H, br d, J =4.6Hz,

H-1), 3.74 (3H, s, CO2CH 3), 2.54 (1H, dt, J = 13.4, 3.6 Hz, H-6α), 2.48 (1H, 5.3. SYNTHESIS 219 dd, J = 13.3, 6.1 Hz, H-11α), 2.17 (1H, dd, J = 12.0, 3.5 Hz, H-8), 2.12 (1H, dq, J = 12.0, 3.5 Hz, H-7β), 1.96 (1H, br s, w 1 = 14.6 Hz, OH), 1.92-1.80 2 (1H, m, H-7α), 1.76 (3H, s, H-19), 1.67 (1H, dd, J = 12.9, 11.2 Hz, H-11β), 1.47 (3H, s, H-20), 1.38 (1H, s, H-10), 1.22 (1H, td, J = 13.1, 3.7 Hz, H-6β);

13 CNMR(CDCl3): δ 171.8, 166.0, 164.9, 143.9, 143.6, 139.3, 132.0, 131.4, 131.2, 129.0, 128.0, 125.7, 108.4, 73.2, 71.7, 64.4, 53.9, 51.8, 51.7, 44.4, 37.9, 37.6, 37.0, 22.0, 18.4, 16.2;

+ + HRESIMS: [M + Na] m/z 595.0938, 597.0920 (calcd for C28H29O8BrNa , 595.0938, 597.0917).

5.3.18 17-Deoxysalvinorin A (49).

Et3SiH (20 µL, 125 µmol) was added to a stirred solution of lactol 35 (18.3 mg, 42.1 µmol) in dry CH2Cl2 (1 mL) under Ar. The solution was cooled to ◦ 0 C, and BF3·Et2O(10µL, 79 µmol) was added. The light brown solution ◦ was stirred at 0 C for 2 h, when TLC (10% acetone/CH2Cl2)showedno35.

The reaction was quenched (0.5 mL sat. NaHCO3), and partitioned between

Et2O and brine. Drying (MgSO4), evaporation in vacuo and flash column chromatography (0-4% acetone/CH2Cl2 gradient) gave enol ether 50 (4.1 mg, 23%) along with 49 as a clear resin (8.4 mg, 48%);

InChI=1/C23H30O7/c1-13(24)30-17-9-16(21(26)27-4)22(2)7-5-15-12-29-18(14 -6-8-28-11-14)10-23(15,3)20(22)19(17)25/h6,8,11,15-18,20H,5,7,9-10,12H2,1- 4H3/t15-,16-,17-,18-,20-,22-,23-/m0/s1

O

O O O H H O

O O 49 220 CHAPTER 5. EXPERIMENTAL.

TLC: hRf 35 49 50

10% acetone/CH2Cl2 38 72 72

2% acetone/CH2Cl2 -2442

20 [α]D −81 (c 0.4, CH2Cl2);

FTIR (film): ν
max 2950, 2927, 2857, 1730, 1506, 1236, 1201, 1161, 1107, 1042, 1018, 932, 889, 875, 784, 736 cm−1;

1 HNMR(CDCl3): δ 7.33 (2H, m, H-15 & 16), 6.34 (1H, t, J =1.4Hz,H-14), 5.14 (1H, dd, J = 10.7, 9.4 Hz, H-2), 4.70 (1H, dd, J = 11.6, 2.5 Hz, H-12),

3.70 (3H, s, CO2CH 3), 3.58 (2H, d, J = 7.6 Hz, H-17), 2.79-2.75 (1H, m, H-4),

2.29-2.24 (2H, m, H-3), 2.15 (3H, s, OCOCH 3), 2.12 (1H, dd, J = 13.1, 2.6 Hz, H-11α), 2.09 (1H, d, J = 1.0 Hz, H-10), 1.68-1.62 (2H, m, H-6), 1.54-1.43 (1H, m, H-8), 1.38 (3H, s, H-20), 1.35-1.26 (2H, m, H-7), 1.19 (1H, ddd, J = 13.1, 11.6, 1.0 Hz, H-11β), 1.08 (3H, s, H-19);

13 CNMR(CDCl3): δ 202.5 (C, C-1), 171.9 (C, C-18), 169.9 (C, OCOCH3), 142.9 (CH, C-15), 138.9 (CH, C-16), 127.0 (C, C-13), 108.7 (CH, C-14), 75.0

(CH, C-2), 67.5 (CH, C-12), 67.1 (CH2, C-17), 65.6 (CH, C-10), 53.8 (CH,

C-4), 51.8 (CH3,CO2CH3), 46.4 (CH, C-8), 45.4 (CH2, C-11), 42.7 (C, C-5),

39.0 (CH2, C-6), 34.6 (C, C-9), 30.8 (CH2, C-3), 20.6 (CH3,OCOCH3), 19.6

(CH2, C-7), 16.8 (CH3, C-19), 13.7 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 441.1881 (calcd for C23H30O7Na , 441.1884).

5.3.19 8,17-Didehydro-17-deoxysalvinorin A (50).

Et3SiH (25 µL, 156 µmol) and Amberlyst 15 resin (22 mg) were added to a solution of lactol 35 (15.1 mg, 34.7µmol) in CH2Cl2 (1 mL). The sealed flask was stirred at rt for 24 h. Filtration, evaporation and flash column chromatography (1% acetone/CH2Cl2)gave50 as a clear resin (11.0 mg, 76%);

InChI=1/C23H28O7/c1-13(24)30-17-9-16(21(26)27-4)22(2)7-5-15-12-29-18(14- 5.3. SYNTHESIS 221

6-8-28-11-14)10-23(15,3)20(22)19(17)25/h6,8,11-12,16-18,20H,5,7,9-10H2,1-4H 3/t16-,17-,18-,20-,22-,23-/m0/s1

O

O O O H O

O O 50

TLC: See table on the facing page.

23 [α]D −60 (c 0.5, CH2Cl2);

FTIR (film): ν
max 2927, 2855, 1731, 1664, 1504, 1437, 1380, 1275, 1237, 1203, 1164, 1113, 1098, 1044, 1024, 1000, 962, 944, 875, 786, 737 cm−1;

1 HNMR(CDCl3): δ 7.38 (1H, br s, H-16), 7.36 (1H, t, J =1.9Hz,H-15), 6.36 (1H, dd, J = 1.9, 0.8 Hz, H-14), 6.26 (1H, d, J = 1.8 Hz, H-17), 5.12 (1H, dd, J = 11.0, 9.0 Hz, H-2), 4.78 (1H, dd, J = 11.7, 2.1 Hz, H-12), 3.70 (3H, s, CO2CH 3), 2.74-2.69 (1H, m, H-4), 2.33 (1H, dd, J = 13.6, 2.1 Hz, H-11α),

2.32-2.24 (1H, m, H-7α), 2.29-2.24 (2H, m, H-3), 2.16 (3H, s, OCOCH 3), 2.12 (1H, d, J = 0.8 Hz, H-10), 1.94 (1H, ddd, J = 14.6, 4.6, 2.6 Hz, H-7β), 1.69 (1H, ddd, J = 13.2, 4.4, 2.6 Hz, H-6α), 1.52 (3H, s, H-20), 1.50 (1H, td, J = 13.2, 4.6 Hz, H-6β), 1.40 (1H, ddd, J = 13.6, 11.6, 0.8 Hz, H-11β), 1.15 (3H, s, H-19);

13 CNMR(CDCl3): δ 202.9 (C, C-1), 172.0 (C, C-18), 169.9 (C, OCOCH3), 143.2 (CH, C-15), 139.3 (CH, C-16), 137.1 (CH, C-17), 125.8 (C, C-13), 117.0 (C, C-8), 108.7 (CH, C-14), 75.2 (CH, C-2), 66.6 (CH, C-12), 65.5 (CH, C-10),

53.5(CH,C-4),51.7(CH3,CO2CH3), 44.4 (CH2, C-11), 42.8 (C, C-5), 40.1

(CH2, C-6), 34.1 (C, C-9), 30.6 (CH2, C-3), 22.8 (CH3, C-20), 22.6 (CH2,C-7),

20.6 (CH3,OCOCH3), 15.4 (CH3,C-19);

+ + HRESIMS: [M + Na] m/z 439.1728 (calcd for C23H28O7Na , 439.1727). 222 CHAPTER 5. EXPERIMENTAL.

5.3.20 13,14,15,16-Tetrahydrosalvinorin A (51).

5% Rh/C (25.3 mg) was added to a solution of 1a (20.3 mg, 46.9 µmol) in

50% CH2Cl2/MeOH (6 mL). The suspension was agitated under H2 (4 atm) at room temperature for 90 minutes, when TLC indicated completion (hRf =

46 (51), 74 (1a) in 20% acetone/CH2Cl2). The solution was filtered through diatomite filter aid and evaporated in vacuo. Flash column chromatography

(10% acetone/CHCl3)gave51 (13-epimers, 1:1) as a clear resin (12 mg, 59%). For characterisation, a portion of the less polar epimer was separated by HPLC in EtOAc;

InChI=1/C23H32O8/c1-12(24)30-16-9-15(20(26)28-4)22(2)7-5-14-21(27)31-17 (13-6-8-29-11-13)10-23(14,3)19(22)18(16)25/h13-17,19H,5-11H2,1-4H3/t13u,1 4-,15-,16-,17-,19-,22-,23-/m0/s1

O

13 H H O O O H H O O

O O 51

TLC: hRf 51 1a

20% acetone/CH2Cl2 46 74

HPLC: tR (min) 51 EtOAc 11.4, 11.75

19 [α]D −39 (c 0.4, CHCl3);

FTIR (film): ν
max 2953, 2879, 1730, 1452, 1438, 1379, 1277, 1235, 1203, 1165, 1110, 1078, 1050, 1005, 951, 912, 895, 755 cm−1;

1 HNMR(CDCl3): δ 5.14 (1H, dd, J = 11.4, 8.6 Hz, H-2), 4.44 (1H, ddd, J = 11.7, 6.9, 5.0 Hz, H-12), 3.85 (1H, td, J = 8.5, 5.0 Hz, H-15a), 3.78 (1H, dd, J = 9.0, 7.6 Hz, H-16a), 3.74 (1H, dt, J = 8.5, 7.4 Hz, H-15b), 3.72 (3H, 5.3. SYNTHESIS 223 s, CO2CH 3), 3.54 (1H, dd, J = 9.0, 6.7 Hz, H-16b), 2.73 (1H, ∼dd, J ≈ 11.4, 5.4 Hz, H-4), 2.43 (1H, sext, J = 7.4 Hz, H-13), 2.31-2.26 (2H, m, H-3), 2.19

(1H, dd, J = 13.3, 5.0 Hz, H-11α), 2.17 (3H, s, OCOCH 3), 2.17-2.12 (1H, m, H-7α), 2.12 (1H, br s, H-10), 2.08 (1H, dddd, J = 12.6, 8.7, 7.4, 5.0 Hz, H-14a), 1.95 (1H, dd, J = 11.5, 3.1 Hz, H-8), 1.81 (1H, ddt, J = 12.6, 8.3, 7.4 Hz, H-14b), 1.78-1.74 (1H, m, H-6α), 1.65-1.56 (1H, m, H-7β), 1.60-1.50 (1H, m, H-6β), 1.35 (3H, s, H-20), 1.26 (1H, ddd, J = 13.3, 11.7, 0.8 Hz, H-11β), 1.09 (3H, s, H-19);

13 CNMR(CDCl3): δ 202.0 (C, C-1), 171.5 (C, C-18), 171.3 (C, C-17), 169.9

(C, OCOCH3),78.2(CH,C-12),75.0(CH,C-2),68.8(CH2, C-16), 68.0 (CH2,

C-15), 64.0 (CH, C-10), 53.5 (CH, C-4), 52.0 (CH3,CO2CH3), 51.4 (CH, C-8),

45.1 (CH, C-13), 42.0 (C, C-5), 41.2 (CH2,C-11),38.1(CH2, C-6), 35.1 (C,

C-9), 30.8 (CH2, C-3), 28.2 (CH2, C-14), 20.6 (CH3,OCOCH3), 18.1 (CH2,

C-7), 16.3 (CH3, C-19), 15.1 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 459.1984 (calcd for C23H32O8Na , 459.1989).

5.3.21 Autoxidation of 1a in KOH/MeOH.

Oxygen was bubbled through a solution of KOH in MeOH (1 M, 2 mL) for 5 min. This was then added to a solution of 1a (21.4 mg, 49.5 µmol) in minimal

CH2Cl2 (≈ 250 µL), and oxygen bubbled through the resulting orange solution for 20 minutes, when TLC (Et2O) showed no 1a or 1b. 10% HCl was added dropwise until an opaque white colour persisted. The solution was diluted in

0.05 M NaOH and extracted into CH2Cl2 (× 3). Drying (MgSO4), evapora- tion in vacuo and flash column chromatography (20 – 50% Et2O/petrol) gave enedione 59 as a resin (9.0 mg, 47%);

InChI=1/C21H22O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7-8,10,12,15,23H,4,6,9H2,1-3H3/t12-, 15-,20-,21-/m0/s1 224 CHAPTER 5. EXPERIMENTAL.

O

OH O H O O

O O 59

TLC: hRf 1a 1b 59

Et2O 544269

14 [α]D +58 (c 0.7, CH2Cl2);

UV (CH3CN): λmax (log ε) 215 (4.30), 249 (3.77), 324 (3.57) nm;

FTIR (film): ν
max 3373, 3149, 2955, 1726, 1651, 1601, 1504, 1456, 1435, 1408, 1380, 1331, 1244, 1203, 1162, 1068, 1022, 911, 875, 793, 736, 703 cm−1;

1 HNMR(CDCl3): δ 7.48 (1H, br s, H-16), 7.41 (1H, t, J =1.8Hz,H-15), 6.99 (1H, s, H-3), 6.91 (1H, s, OH), 6.42 (1H, dd, J = 2.0, 0.9 Hz, H-14), 5.44

(1H, dd, J = 12.3, 2.9 Hz, H-12), 3.85 (3H, s, CO2CH 3), 3.11 (1H, ddd, J = 14.8, 2.9, 1.2 Hz, H-11α), 2.98 (1H, ddd, J = 9.7, 5.4, 1.2 Hz, H-8), 2.53 (1H, ddd, J = 14.1, 7.7, 6.3 Hz, H-6a), 2.24 (1H, dtd, J = 14.6, 7.4, 5.3 Hz, H-7a), 2.02 (1H, dd, J = 15.0, 12.2 Hz, H-11β), 1.98 (1H, dddd, J = 14.3, 9.7, 7.7, 6.4 Hz, H-7b), 1.77-1.67 (1H, m, H-6b), 1.72 (3H, s, H-19), 1.67 (3H, s, H-20);

13 CNMR(CDCl3): δ 180.7 (C, C-2), 173.2 (C, C-17), 165.4 (C, C-18), 157.5 (C, C-4), 145.0 (C, C-1), 143.6 (CH, C-15), 140.0 (C, C-10), 139.6 (CH, C-16), 128.2 (CH, C-3), 124.5 (C, C-13), 108.4 (CH, C-14), 70.9 (CH, C-12), 52.6

(CH3,CO2CH3), 44.9 (CH, C-8), 42.3 (C, C-5), 37.7 (C, C-9), 36.8 (CH2,

C-11), 30.3 (CH3,C-19),28.3(CH2, C-6), 24.4 (CH3, C-20), 21.9 (CH2,C-7);

+ + HRESIMS: [M + Na] m/z 409.1265 (calcd for C21H22O7Na , 409.1258).

The aqueous layer was acidified with 10% HCl until opaque white, then ex- tracted into CH2Cl2 (× 3), which was dried (MgSO4) and filtered. MeOH (10 mL) and Me3SiCHN2 in Et2O (2.0 M, 200 µL) were added. The yellow solution was stirred for 30 min, then evaporated under reduced pressure. Flash column 5.3. SYNTHESIS 225 chromatography (50% EtOAc/petrol) gave a mixture of the seco triesters 60a, 60b and 60c (12.0 mg, 53%). Repeated HPLC (36% EtOAc/petrol) gave a compound which decomposed before full characterisation but was tentatively identified, based on HRESIMS and NMR (1H and COSY), as 60c;

InChI=1/C23H30O9/c1-22(15(19(25)29-4)10-17(24)28-3)8-6-14-20(26)32-16 (13-7-9-31-12-13)11-23(14,2)18(22)21(27)30-5/h7,9,12,14-16,18H,6,8,10-11H2, 1-5H3/t14-,15-,16-,18+,22-,23-/m0/s1

O

O O H H

O 10 O O

O 60c O O (tentative)

TLC: hRf 59 60a 60b 60c

80% Et2O/petrol 50 41 41 41

HPLC: tR (min) 60c 60a 60b 36% EtOAc/petrol 17.0 17.4 18.3

1 HNMR(CDCl3): δ 7.46 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.42 (1H, dd, J = 1.9, 0.9 Hz, H-14), 5.19 (1H, dd, J = 11.7, 5.4 Hz, H-12),

3.72 (3H, s, CO2CH3), 3.67 (3H, s, CO2CH3), 3.65 (3H, s, CO2CH3), 3.55 (1H, dd, J = 12.3, 2.9 Hz, H-4), 2.80 (1H, dd, J = 16.6, 12.3 Hz, H-3a), 2.59 (1H, dd, J = 14.9, 5.5 Hz, H-11α), 2.56 (1H, dd, J = 16.6, 2.9 Hz, H-3b), 2.53 (1H, br s, H-10), 2.39 (1H, dd, J = 12.6, 3.7 Hz, H-8), 1.94-1.83 (1H, m, H-7α), 1.86 (1H, dd, J = 14.9, 11.8 Hz, H-11β), 1.78 (1H, dq, J = 14.4, 3.8

Hz, H-7β), 1.60-1.41 (2H, m, H-6 [obscured by H2O]), 1.41 (3H, s, H-19), 1.20 (3H, s, H-20);

+ + HRESIMS: [M + Na] m/z 473.1781 (calcd for C23H30O9Na , 473.1782).

Further elution gave 60a; 226 CHAPTER 5. EXPERIMENTAL.

InChI=1/C23H30O9/c1-22(15(19(25)29-4)10-17(24)28-3)8-6-14-20(26)32-16 (13-7-9-31-12-13)11-23(14,2)18(22)21(27)30-5/h7,9,12,14-16,18H,6,8,10-11H2, 1-5H3/t14-,15-,16-,18-,22-,23-/m0/s1

O

O O H H O O O

O O O 60a

TLC, HPLC: see table on the previous page.

24 [α]D +6 (c 0.1, CH2Cl2);

FTIR (film): ν
max 2953, 1732, 1506, 1436, 1393, 1373, 1261, 1226, 1202, 1163, 1137, 1079, 1025, 875, 792 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.44 (1H, br s, H-16), 7.43 (1H, t, J =1.9Hz, H-15), 6.41 (1H, m, H-14), 5.46 (1H, dd, J = 11.6, 5.4 Hz, H-12), 3.69 (3H, s,

CO2CH3), 3.68 (3H, s, CO2CH3), 3.65 (3H, s, CO2CH3), 2.82 (1H, dd, J = 15.9, 11.8 Hz, H-3a), 2.72 (1H, dd, J = 11.8, 2.0 Hz, H-4), 2.43 (1H, dd, J = 15.9, 2.0 Hz, H-3b), 2.25 (1H, s, H-10), 2.17-2.08 (2H, m, H-7β, 8), 2.01 (1H, dd, J = 13.7, 5.5 Hz, H-11α), 1.84 (1H, ddd, J = 13.7, 12.0, 0.8 Hz, H-11β), 1.74-1.63 (2H, m, H-6a, 7α), 1.52-1.48 (1H, m, H-6b), 1.38 (3H, s, H-19), 1.30 (3H, s, H-20);

13 CNMR(CDCl3): δ 174.0 (C, C-2/18), 172.5 (C, C-2/18), 171.7 (C, C-1), 171.0 (C, C-17), 143.8 (CH, C-15), 139.6 (CH, C-16), 125.1 (C, C-13), 108.5

(CH, C-14), 71.6 (CH, C-12), 58.6 (CH, C-10), 52.0 (CH3,CO2CH3), 51.9

(CH, C-4), 51.7 (CH3,CO2CH3), 51.4 (CH3,CO2CH3), 50.5 (CH, C-8), 44.6

(CH2, C-11), 38.7 (C, C-5), 36.9 (C, C-9), 35.1 (CH2, C-6), 32.4 (CH2,C-3),

18.8 (CH3, C-19), 18.2 (CH2, C-7), 15.6 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 473.1783 (calcd for C23H30O9Na , 473.1782). 5.3. SYNTHESIS 227

Further elution gave 60b;

InChI=1/C23H30O9/c1-22(15(19(25)29-4)10-17(24)28-3)8-6-14-20(26)32-16 (13-7-9-31-12-13)11-23(14,2)18(22)21(27)30-5/h7,9,12,14-16,18H,6,8,10-11H2, 1-5H3/t14-,15+,16+,18+,22+,23+/m1/s1

O

O O H H

O 8 O O

O O O 60b

TLC, HPLC: see table on page 225.

23 [α]D +4 (c 0.4, CH2Cl2);

FTIR (film): ν
max 2953, 1733, 1504, 1436, 1392, 1361, 1247, 1200, 1166, 1091, 1062, 1025, 998, 968, 909, 875, 849, 794, 735 cm−1;

1 H NMR (500 MHz, CDCl3): δ 7.47 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz, H-15), 6.42 (1H, dd, J = 1.8, 0.7 Hz, H-14), 5.26 (1H, dd, J = 12.0, 2.1 Hz,

H-12), 3.67 (3H, s, CO2CH3), 3.66 (3H, s, CO2CH3), 3.64 (3H, s, CO2CH3), 2.84 (1H, dd, J = 15.9, 11.9 Hz, H-3a), 2.76 (1H, dd, J = 11.9, 1.6 Hz, H-4), 2.53 (1H, dd, J = 4.6, 2.8, H-8), 2.39 (1H, dd, J = 16.0, 1.8 Hz, H-3b), 2.28 (1H, br s, H-10), 2.19-2.14 (1H, m, H-7β), 2.02 (1H, dd, J = 14.5, 11.9 Hz, H-11β), 1.94-1.86 (2H, m, H-6β,7α), 1.80 (1H, dd, J = 14.4, 2.1 Hz, H-11α), 1.48 (3H, s, H-20), 1.32-1.27 (1H, m, H-6α), 1.29 (3H, s, H-19);

13 CNMR(CDCl3): δ 173.8 (C, C-18), 173.2 (C, C-17), 172.6 (C, C-2), 172.2 (C, C-1), 143.7 (CH, C-15), 139.7 (CH, C-16), 123.6 (C, C-13), 108.4 (CH,

C-14), 69.8 (CH, C-12), 57.8 (CH, C-10), 51.9 (CH3,CO2CH3), 51.7 (CH3,

CO2CH3), 51.7 (CH, C-4), 51.5 (CH3,CO2CH3), 46.5 (CH2, C-11), 44.4 (CH,

C-8), 38.5 (C, C-5), 35.8 (C, C-9), 32.2 (CH2, C-3), 31.6 (CH2, C-6), 25.1

(CH3, C-20), 18.2 (CH3, C-19), 18.2 (CH2,C-7);

+ + HRESIMS: [M + Na] m/z 473.1780 (calcd for C23H30O9Na , 473.1782). 228 CHAPTER 5. EXPERIMENTAL.

5.3.22 NaBH4 reduction of 59.

Enedione 59 (41.3 mg, 107 µmol) and NaBH4 (10 mg, 264 µmol) were dissolved in CH2Cl2 (500 µL), followed by EtOH (2 mL), and stirred under Ar at 40 ◦C. The initial orange colour faded to faint yellow within 1 h. After 4 h, ◦ TLC (Et2O) indicated completion. The solution was cooled to 0 C, and 0.5%

H2SO4/MeOH added dropwise until effervescence ceased. The solution was concentrated to ≈ 500 µL in vacuo, then partitioned between brine (acidified with 10% HCl) and CH2Cl2 (× 3). Drying (MgSO4), evaporation in vacuo

27 and flash column chromatography (70-100% Et2O/petrol) gave 38h (15.7 mg, 37%);

InChI=1/C21H28O7/c1-20-6-4-12-19(25)28-15(11-5-7-27-10-11)9-21(12,2)17 (20)16(23)14(22)8-13(20)18(24)26-3/h5,7,10,12-17,22-23H,4,6,8-9H2,1-3H3/t 12-,13+,14+,15+,16+,17+,20+,21+/m1/s1

O

OH O H H HO 8 O

O O 38h

TLC: hRf 38h 59

Et2O5369

1 HNMR(CDCl3): δ 7.48 (1H, dt, J = 1.7, 0.9 Hz, H-16), 7.42 (1H, t, J = 1.7Hz,H-15),6.41(1H,dd,J = 1.9, 0.9 Hz, H-14), 5.29 (1H, dd, J = 11.7,

1.5 Hz, H-12), 4.07 (1H, br s, H-1), 3.65 (3H, s, CO2CH 3), 3.54 (1H, ddd, J = 11.1, 4.7, 3.2 Hz, H-2), 2.45 (1H, br d, J = 4.7 Hz, H-8), 2.23-2.10 (4H, m),

1.90 (1H, tdd, J = 14.4, 5.5, 4.0 Hz, H-7α), 1.73-1.53 (m, obscured by H2O& OH), 1.66 (3H, s, H-20), 1.32 (3H, s, H-19), 0.90 (1H, d, J =1.6Hz,H-10);

1 13 27 HNMR([CD3]2CO) and CNMR(CDCl3) matched literature values. 5.3. SYNTHESIS 229

5.3.23 O-Demethyl-18-deoxysalvinorin A (77).

To dry EtSH (1.5 mL, 20.2 mmol) at 0 ◦C, stirred rapidly under a stream of Ar, was added nBuLi in hexanes (2.1 M, 8 mL, 16.9 mmol). Immediate, violent gas evolution was accompanied by the sudden formation of a white solid. Rapid stirring was continued while the remaining nBuLi was added, swirling the flask when necessary to free the stir bar. After warming to room temperature, the solution was evaporated under reduced pressure and dried under high vacuum at 50 ◦C for 30 min, giving LiSEt as a white powder. This was stored at room temperature in a sealed flask under Ar for up to a year without losing activity.

1a (42.2 mg, 97.6 µmol) and LiSEt (13.7 mg, 201 µmol) under Ar were dis- solved in DMPU (1 mL). The yellow solution was stirred at 55 ◦C for 23 h, when TLC (1% H2SO4/10% MeOH/CH2Cl2) indicated consumption of 1a and intermediate 1b. The cooled orange solution was diluted with EtOAc and washed (10% HCl × 3, then water), then extracted into 1% NaHCO3 (× 3). The pooled aqueous fractions were acidified at 0 ◦C with 10% HCl, then extracted into CH2Cl2 (× 3). Drying (MgSO4) and evaporation in vacuo gave an amber resin, which was treated with Ac2O (0.4 mL) in pyridine (1 mL) and catalytic DMAP at room temperature for 17 h. After cooling to 0 ◦Cand quenching (water), the solution was diluted in EtOAc and washed (10% HCl and sat. NH4Cl). Drying (MgSO4) and evaporation in vacuo gave the mixed acids 67a (H-8α : β, ∼1.4:1) as an amber resin (29.7 mg, 73% over two steps);

InChI=1/C22H26O8/c1-11(23)29-15-8-14(19(25)26)21(2)6-4-13-20(27)30-16 (12-5-7-28-10-12)9-22(13,3)18(21)17(15)24/h5,7,10,13-16,18H,4,6,8-9H2,1- 3H3,(H,25,26)/t13u,14-,15-,16-,18-,21-,22-/m0/s1/f/h25H

O

O O O H H O O

O OH 67a 230 CHAPTER 5. EXPERIMENTAL.

TLC: hRf 1a 1b 67a 77 78

1% H2SO4/10% MeOH/CH2Cl2 71 63 0-20

1% NEt3/EtOAc 0-30 56 56

Et2O2030

vanillin/H2SO4 purple purple purple blue

1 ∗ HNMR(CDCl3): δ 7.43 (1H, br s, H-16 ), 7.41 (1H, br s, H-16), 7.39-7.38 (2H, m, H-15, 15∗), 6.37 (2H, m, H-14, 14∗), 5.52 (1H, dd, J = 11.6, 5.2 Hz, H-12), 5.26 (1H, dd, J = 11.9, 1.8 Hz, H-12∗), 5.16 (1H, dd, J = 12.4, 7.6 Hz, H-2), 5.11 (1H, dd, J = 12.4, 7.2 Hz, H-2∗), 2.80 (1H, dd, J = 5.2, 3.5 Hz, H-4), 2.76 (1H, dd, J = 5.2, 3.5 Hz, H-4∗), 2.49 (1H, dd, J = 13.4, 5.2 Hz, H-11α), 2.45 (1H, dd, J = 4.9, 2.0 Hz, H-8∗), 2.37 (1H, dd, J = 14.8, 2.0 Hz, H-11α∗), 2.37-1.94 (m), 2.28 (1H, br s, H-10∗), 2.20 (1H, br s, H-10), 2.17 ∗ (3H, s, OCOCH 3), 2.15 (3H, s, OCOCH 3 ), 1.84 (1H, tt, J = 14.2, 4.2 Hz, H-7α∗), 1.73 (1H, dtd, J = 13.7, 3.5, 0.8 Hz, H-6α), 1.67-1.52 (m), 1.63 (3H, s, H-20∗), 1.50 (1H, dd, J = 15.1, 12.1 Hz, H-11β∗), 1.45 (3H, s, H-20), 1.13 (3H, s, H-19), 1.09 (3H, s, H-19∗).

The mixed acids 67a (35.2 mg, 84.0 µmol) were stirred in dry THF (1 mL) at rt, under Ar, in a flask fitted with a reflux condenser. BH3·THF (1.0 M, 110 µL, 110 µmol) was added dropwise; the solution was heated to 55 ◦C and stirred at this temperature for 90 minutes, when TLC (1% NEt3/EtOAc) indicated completion. The solution was cooled to room temperature, quenched with water (dropwise), and evaporated under reduced pressure to a cloudy paste.

This was diluted with sat. NaHCO3 and extracted into CH2Cl2 (×3). Drying

(MgSO4), evaporation in vacuo and flash column chromatography (10-25% acetone/Et2O gradient) monitored by TLC (Et2O) gave 78 as a clear resin (8.4 mg, 25%);

InChI=1/C22H28O7/c1-12(24)28-16-8-14(10-23)21(2)6-4-15-20(26)29-17(13- 5-7-27-11-13)9-22(15,3)19(21)18(16)25/h5,7,11,14-17,19,23H,4,6,8-10H2,1- 3H3/t14-,15+,16-,17-,19-,21-,22-/m0/s1

∗Unnatural (H-8α)epimer. 5.3. SYNTHESIS 231

O

O O O H H O 8 O

OH 78

TLC: see table on the preceding page.

24 [α]D −31 (c 0.3, CH2Cl2);

FTIR (film): ν
max 3547, 3148, 2942, 2880, 1725, 1504, 1468, 1450, 1440, 1376, 1238, 1202, 1159, 1085, 1050, 1024, 950, 934, 875, 802, 784, 734, 703 cm−1;

1 HNMR(CDCl3): δ 7.44 (1H, br s, H-16), 7.38 (1H, t, J =1.8Hz,H-15), 6.38 (1H, dd, J = 1.9, 0.9 Hz, H-14), 5.25 (1H, dd, J = 12.0, 2.2 Hz, H-12), 5.09 (1H, ddd, J = 12.3, 7.2, 1.0 Hz, H-2), 3.93 (1H, dd, J = 10.5, 3.9 Hz, H-18a), 3.46 (1H, ddd, J = 10.8, 8.0, 0.6 Hz, H-18b), 2.52 (1H, ddd, J = 12.3, 7.0, 2.8 Hz, H-3a), 2.45-2.43 (1H, m, H-8), 2.36 (1H, dd, J = 15.0, 2.2 Hz, H-11α), 2.24 (1H, d, J = 0.9 Hz, H-10), 2.22-2.09 (2H, m, H-6β,7β), 2.15 (3H, s, OCOCH 3), 1.91-1.72 (4H, m, H-3b, 4, 6α,7α), 1.64 (3H, s, H-20), 1.52 (1H, dd, J = 14.8, 12.0 Hz, H-11β), 0.92 (3H, s, H-19);

13 CNMR(CDCl3): δ 204.0, 173.6, 169.9, 143.6, 139.7, 123.4, 108.6, 76.2, 70.1, 64.5, 61.7, 50.3, 48.1, 45.4, 42.1, 34.6, 33.9, 31.6, 24.8, 20.6, 17.7, 15.5;

+ + HRESIMS: [M + Na] m/z 427.1725 (calcd for C22H28O7Na , 427.1727).

Further elution gave 77 as a clear resin (7.7 mg, 23%);

InChI=1/C22H28O7/c1-12(24)28-16-8-14(10-23)21(2)6-4-15-20(26)29-17(13- 5-7-27-11-13)9-22(15,3)19(21)18(16)25/h5,7,11,14-17,19,23H,4,6,8-10H2,1- 3H3/t14-,15-,16-,17-,19-,21-,22-/m0/s1 232 CHAPTER 5. EXPERIMENTAL.

O

O O O H H O O

OH 77

TLC: see table on page 230.

25 [α]D −19 (c 0.3, CH2Cl2);

FTIR (film): ν
max 3468, 2944, 1727, 1504, 1452, 1376, 1237, 1162, 1095, 1053, 1023, 952, 875, 794, 735, 703 cm−1;

1 HNMR(CDCl3): δ 7.41 (1H, br s, H-16), 7.39 (1H, t, J = 1.9 Hz, H-15), 6.38 (1H, dd, J = 1.9, 0.8 Hz, H-14), 5.52 (1H, dd, J = 11.7, 5.3 Hz, H-12), 5.15 (1H, dd, J = 12.3, 7.6 Hz, H-2), 3.94 (1H, dd, J = 10.3, 3.9 Hz, H-18a), 3.49 (1H, dd, J = 10.3, 8.0 Hz, H-18b), 2.54 (1H, ddd, J = 12.3, 7.0, 2.6 Hz, H-3a), 2.49 (1H, dd, J = 13.3, 5.3 Hz, H-11α), 2.17 (1H, d, J = 0.9 Hz, H-10), 2.16

(3H, s, OCOCH 3), 2.20-2.13 (1H, m, H-7β), 2.06 (1H, dd, J = 12.1, 3.0 Hz, H-8), 1.99 (1H, dt, J = 13.3, 3.3 Hz, H-6α), 1.94-1.87 (1H, m, H-4), 1.80 (1H, q, J = 12.3 Hz, H-3b), 1.64 (1H, dddd, J = 14.0, 13.5, 12.1, 3.4 Hz, H-7α), 1.58 (1H, ddd, J = 13.3, 11.7, 0.9 Hz, H-11β), 1.45 (3H, s, H-20), 1.43 (1H, td, J = 13.2, 3.7 Hz, H-6β), 0.96 (3H, s, H-19);

13 CNMR(CDCl3): δ 203.6 (C, C-1), 171.3 (C, C-17), 170.0 (C, OCOCH3), 143.7 (CH, C-15), 139.4 (CH, C-16), 125.2 (C, C-13), 108.4 (CH, C-14), 76.0

(CH, C-2), 72.1 (CH, C-12), 64.5 (CH, C-10), 61.7 (CH2, C-18), 51.5 (CH,

C-8), 50.8 (CH, C-4), 43.4 (CH2, C-11), 41.9 (C, C-5), 38.1 (CH2, C-6), 35.3

(C, C-9), 31.8 (CH2, C-3), 20.6 (CH3,OCOCH3), 18.1 (CH2, C-7), 16.7 (CH3,

C-19), 15.3 (CH3, C-20);

+ + HRESIMS: [M + Na] m/z 427.1729 (calcd for C22H28O7Na , 427.1727). 5.3. SYNTHESIS 233

5.3.24 1-Deoxysalvinorin A (81a).

AsolutionofNaBH4 (11.6 mg, 307 µmol) and 1a (108 mg, 249 µmol) in dry ◦ EtOH (9 mL)/CH2Cl2 (2 mL) was stirred under Ar at 40 C. The cloudy solution gradually cleared. At 4 h, TLC (Et2O) indicated completion (for hRf values, see table on page 208). The solution was evaporated in vacuo,and the residue partitioned between brine and CH2Cl2 (× 3). Drying (MgSO4) and evaporation in vacuo gave an approximately equal mixture of 36h23 and 38h27 (total 103 mg), which was used without purification.

This crude mixture and 1,1’-thiocarbonyldiimidazole (118 mg, 662 µmol) were dissolved in dry DMF and stirred under Ar at 90 ◦Cfor6h,whenTLC

(40% acetone/ CH2Cl2) indicated completion. The cooled solution was diluted in EtOAc/Et2Oandwashed(10%HCl× 3, then brine). Drying (MgSO4) and evaporation in vacuo followed by flash column chromatography (1-5% acetone/CH2Cl2 gradient) gave the cyclic thionocarbonates 80 (H-8α : β, ≈ 1:2), as a clear resin (72 mg, 67% over two steps);

InChI=1/C22H26O7S/c1-21-6-4-12-19(24)27-15(11-5-7-26-10-11)9-22(12,2)17 (21)16-14(28-20(30)29-16)8-13(21)18(23)25-3/h5,7,10,12-17H,4,6,8-9H2,1-3H3 /t12u,13-,14-,15-,16-,17-,21-,22-/m0/s1

O

S O O H H O O

O O 80

TLC: hRf 36h/38h 80 83 81b 82b 81a

Et2O 47404774

10% acetone/CH2Cl2 55 70 20 20

40% acetone/CH2Cl2 62 86

vanillin/H2SO4 purple blue purple 234 CHAPTER 5. EXPERIMENTAL.

1 ∗ HNMR(CDCl3): δ 7.49 (1H, br s, H-16 ), 7.45 (1H, br s, H-16), 7.43-7.42 (2H, m, H-15, 15∗), 6.40 (1H, br s, H-14, 14∗), 5.58 (1H, dd, J = 11.2, 5.2 Hz, H-12), 5.34 (1H, br d, J = 10.8 Hz, H-12∗), 5.03 (1H, dd, J = 6.4, 2.8 ∗ ∗ Hz, H-1), 4.93 - 4.78 (3H, m, H-2, 1 ,2), 3.70 (3H, s, CO2CH 3), 3.67 (3H, ∗ ∗ s, CO2CH 3 ), 2.56 (1H, dd, J = 5.3, 1.9 Hz, H-8 ), 2.45 (1H, dd, J = 12.8, 5.2 Hz, H-11α), 2.27 - 2.10 (m), 1.94 (1H, tdd, J = 14.2, 4.8, 3.9 Hz, H-7α∗), 1.82 - 1.69 (m), 1.71 (1H, dd, J = 13.5, 11.0 Hz, H-11β), 1.64 (3H, s, H-20∗), 1.59-1.49 (m), 1.47 (3H, s, H-20), 1.37 - 1.20 (m), 1.33 (3H, s, H-19), 1.30 (3H, s, H-19∗);

13 CNMR(CDCl3): δ 191.0, 190.9, 173.4, 171.6, 171.4, 170.8, 144.0, 143.8, 139.7, 139.4, 125.1, 123.3, 108.22, 108.19, 79.8, 79.0, 78.9, 78.8, 71.5, 69.6, 53.9, 53.1, 52.1, 52.0, 51.85, 51.79, 51.2, 48.0, 45.9, 43.5, 38.7, 37.5, 36.4, 36.2, 35.9, 34.9, 26.4, 26.3, 26.2, 18.3, 17.6, 16.5, 16.2, 14.8. n 408 Bu3SnH was prepared by a published procedure with reduced reaction time (10 minutes),409 distilled (80 ◦C/0.5 mmHg) and stored in the dark under argon at -20 ◦C (when used, it remained clear and unclouded). The thionocarbonate mixture was dissolved in dry toluene (2 mL). Ar was bubbled through the resulting cloudy solution for 2 min; it was then stirred under Ar at 80 ◦C. A

n 408 solution of Bu3SnH (150 µL, 550 µmol) and AIBN (5.4 mg, 33 µmol) in dry toluene (2 mL, also deoxygenated) was added in small portions over 4 h; the solution gradually cleared. After a further 2 h, TLC (10% acetone/ CH2Cl2) indicated completion. The solution was cooled and loaded directly onto silica gel, rinsing the flask with CH2Cl2 (× 2). Repeated flash column chromatog- raphy (50-70% EtOAc/petrol gradient) monitored by TLC (Et2O) gave two fractions, A and B. Fraction A was subjected to flash column chromatography

(10% acetone/ CH2Cl2)togivecarbonate83 as a clear resin (2.7 mg, 4%);

InChI=1/C22H26O8/c1-21-6-4-12-19(24)28-15(11-5-7-27-10-11)9-22(12,2)17 (21)16-14(29-20(25)30-16)8-13(21)18(23)26-3/h5,7,10,12-17H,4,6,8-9H2,1-3H3 /t12-,13-,14-,15-,16-,17-,21-,22-/m0/s1

∗Unnatural (H-8α)epimer. 5.3. SYNTHESIS 235

O

O O O H H O O

O O 83

TLC: see table on page 233.

22 [α]D +5 (c 0.1, CH2Cl2);

FTIR (film): ν
max 2952, 1797, 1727, 1502, 1452, 1441, 1370, 1276, 1229, 1200, 1163, 1073, 1043, 1023, 956, 875, 790, 736 cm−1;

1 HNMR(CDCl3): δ 7.46 (1H, br s, H-16), 7.43 (1H, t, J = 1.9 Hz, H-15), 6.40 (1H, dd, J = 1.9, 0.9 Hz, H-14), 5.58 (1H, dd, J = 11.4, 5.4, H-12), 4.90 (1H, dd, J = 6.3, 2.9 Hz, H-1), 4.67 (1H, tdd, J = 7.9, 6.2, 2.2 Hz, H-2), 3.70 (3H, s, CO2CH 3), 2.46-2.41 (2H, m), 2.27-2.10 (4H, m), 1.87-1.67 (m, overlapping with H2O), 1.42 (3H, s, H-20), 1.36-1.24 (m), 1.34 (3H, s, H-19), 1.27 (1H, d, J = 2.9 Hz, H-10);

13 CNMR(CDCl3): δ 171.6, 170.8, 154.1, 144.0, 139.4, 125.2, 108.2, 75.0, 74.6, 71.5, 54.5, 52.2, 52.0, 43.7, 38.8, 37.5, 35.9, 26.9, 24.3, 18.3, 16.4, 16.3;

+ + HRESIMS: [M + Na] m/z 441.1515 (calcd for C22H26O8Na , 441.1520).

Further elution gave 82b as a clear resin (15.5 mg, 25%);

InChI=1/C21H28O6/c1-20-6-4-14-19(24)27-16(12-5-7-26-11-12)10-21(14,2)17 (20)9-13(22)8-15(20)18(23)25-3/h5,7,11,13-17,22H,4,6,8-10H2,1-3H3/t13-,14+, 15-,16-,17-,20-,21-/m0/s1

O

O H H HO 8 O

O O 82b 236 CHAPTER 5. EXPERIMENTAL.

TLC: see table on page 233.

21 [α]D −4(c 0.1, CH2Cl2);

FTIR (film): ν
max 3432, 2948, 2871, 1728, 1502, 1450, 1438, 1390, 1363, 1277, 1255, 1199, 1167, 1091, 1054, 1022, 1000, 875, 840, 785, 727 cm−1;

1 HNMR(CDCl3): δ 7.47 (1H, br s, H-16), 7.42 (1H, t, J = 1.7 Hz, H-15), 6.41 (1H, dd, J = 2.0, 0.9 Hz, H-14), 5.24 (1H, dd, J = 11.8, 1.5 Hz, H-12), 3.69 -

3.60 (1H, m, H-2), 3.64 (3H, s, CO2CH 3), 2.48 (1H, dd, J = 5.0, 1.6 Hz, H-8), 2.17 (1H, dtd, J = 14.4, 3.5, 2.1 Hz, H-7β), 2.12 (1H, dd, J = 13.2, 3.6 Hz, H-4), 1.99 (1H, dd, J = 14.0, 1.6 Hz, H-11α), 1.91 (1H, dddd, J = 12.9, 5.0, 3.3, 1.9 Hz, H-1β), 1.87 - 1.58 (m), 1.81 (1H, dd, J = 13.0, 11.0 Hz, H-11β), 1.52 (1H, dt, J = 13.6, 3.2 Hz, H-6α), 1.43 (1H, td, J = 12.6, 10.7 Hz, H-1α), 1.28-1.24 (1H, m, H-6β), 1.24 (3H, s, H-19), 1.08 (3H, s, H-20), 1.00 (1H, dd, J = 12.7, 2.2 Hz, H-10);

13 CNMR(CDCl3): δ 174.1, 173.3, 143.7, 139.7, 123.8, 108.4, 70.1, 69.7, 53.8, 52.4, 51.3, 49.1, 44.8, 36.4, 35.8, 34.1, 34.0, 31.3, 24.1, 17.9, 14.1;

+ + HRESIMS: [M + Na] m/z 399.1771 (calcd for C21H28O6Na , 399.1778).

Fraction B gave 81b as a resin (13.5 mg, 22%);

InChI=1/C21H28O6/c1-20-6-4-14-19(24)27-16(12-5-7-26-11-12)10-21(14,2)17 (20)9-13(22)8-15(20)18(23)25-3/h5,7,11,13-17,22H,4,6,8-10H2,1-3H3/t13-,14-, 15-,16-,17-,20-,21-/m0/s1

O

O H H HO O

O O 81b

TLC: see table on page 233. 5.3. SYNTHESIS 237

1 HNMR(CDCl3): δ 7.43 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.40 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.50 (1H, ddd, J = 11.4, 5.7, 0.8 Hz,

H-12), 3.70-3.63 (1H, m, H-2), 3.67 (3H, s, CO2CH 3), 2.32 (1H, dd, J = 13.4, 5.8 Hz, H-11α), 2.13 (2H, br dd, J = 12.9, 3.2 Hz, H-4, 8), 2.09 (1H, dq, J = 14.1, 3.5 Hz, H-7β), 1.96-1.82 (3H, m, H-1β, 3), 1.74 (1H, dt, J = 13.6, 3.2

Hz, H-6α), 1.65-1.58 (m, overlapping with H2O), 1.44 (1H, td, J = 12.6, 10.9 Hz, H-1α), 1.28 (1H, td, J = 13.6, 4.0 Hz, H-6β), 1.10 (3H, s, H-19), 1.07 (3H, s, H-20), 1.04 (1H, dd, J = 12.8, 2.5 Hz, H-10);

13 CNMR(CDCl3): δ 173.0, 172.0, 143.8, 139.3, 125.7, 108.4, 71.8, 70.0, 54.4, 53.0, 51.5, 51.2, 44.0, 38.0, 36.9, 36.2, 33.7, 30.4, 18.2, 15.0, 14.6.

This was fully characterised as the acetate by dissolving in pyridine (0.3 mL) and Ac2O (0.3 mL) with a crystal of DMAP. The solution was stirred at rt for 2.5 h, when TLC (Et2O, visualised in KMnO4) indicated completion, then evaporated in vacuo. Flash column chromatography (60% Et2O/n-pentane) gave 81a as a clear resin (12.3 mg, 82%);

InChI=1/C23H30O7/c1-13(24)29-15-9-17(20(25)27-4)22(2)7-5-16-21(26)30-18 (14-6-8-28-12-14)11-23(16,3)19(22)10-15/h6,8,12,15-19H,5,7,9-11H2,1-4H3/t15 -,16-,17-,18-,19-,22-,23-/m0/s1

O

O O H H O 1 2 10 O

O O 81a

TLC: see table on page 233.

21 [α]D −8(c 0.1, CH2Cl2);

FTIR (film): ν
max 2952, 1729, 1503, 1450, 1434, 1365, 1319, 1245, 1202, 1153, 1099, 1024, 954, 901, 875, 783, 734 cm−1; 238 CHAPTER 5. EXPERIMENTAL.

1 HNMR(CDCl3): δ 7.43 (1H, br s, H-16), 7.42 (1H, t, J =1.8Hz,H-15), 6.41 (1H, dd, J = 1.9, 1.0 Hz, H-14), 5.47 (1H, ddd, J = 11.2, 5.6, 0.7 Hz,

H-12), 4.74 (1H, tt, J = 11.0, 5.5 Hz, H-2), 3.66 (3H, s, CO2CH 3), 2.28 (1H, dd, J = 13.5, 5.9 Hz, H-11α), 2.18 (1H, dd, J = 10.9, 3.5 Hz, H-4), 2.15 (1H, dd, J = 10.0, 3.5 Hz, H-8), 2.09 (1H, dq, J = 14.2, 3.4 Hz, H-7β), 2.05 (3H, s,

OCOCH 3), 1.95-1.89 (1H, m, H-1β), 1.92-1.83 (2H, m, H-3), 1.74 (1H, dt, J = 13.5, 3.3 Hz, H-6α), 1.68-1.57 (1H, m, H-7α), 1.64 (1H, dd, J = 13.6, 11.5 Hz, H-11β), 1.50 (1H, td, J = 12.8, 11.2 Hz, H-1α), 1.31 (1H, td, J = 13.6, 3.7 Hz, H-6β), 1.11 (3H, s, H-19), 1.10 (1H, dd, J = 12.9, 2.2 Hz, H-10), 1.06 (3H, s, H-20);

13 CNMR(CDCl3): δ 172.7 (C, C-18), 171.9 (C, C-17), 170.5 (C, OCOCH3), 143.8 (CH, C-15), 139.4 (CH, C-16), 125.6 (C, C-13), 108.4 (CH, C-14), 71.82 (CH, C-2/12), 71.77 (CH, C-2/12), 54.2 (CH, C-4), 52.9 (CH, C-10), 51.5

(CH3,CO2CH3), 51.2 (CH, C-8), 43.9 (CH2, C-11), 38.0 (CH2, C-6), 36.9 (C,

C-9), 36.2 (C, C-5), 29.6 (CH2, C-3), 26.7 (CH2, C-1), 21.3 (CH3,OCOCH3),

18.2 (CH2, C-7), 15.0 (CH3, C-19), 14.6 (CH3,C-20);

+ + HRESIMS: [M + Na] m/z 441.1893 (calcd for C23H30O7Na , 441.1884). Bibliography

[1] Photo by Max Hem, Melbourne.

[2] Esquivel, B.; Sánchez, A. A.; Aranda, E.: Natural products of agrochem- ical interest from Mexican Labiatae. In Phytochemicals and Phytophar- maceuticals, F. Shahidi; C.-T. Ho, eds., AOCS Press, Champaign, IL, pp. 371–385 2000.

[3] Rodriguez-Hahn, L.; Esquivel, B.; Cardenas, J.: Neo-clerodane diter- penoids from American Salvia species. Recent Adv. Phytochem. 1995, 29, 311–332.

[4] Epling, C.; Játiva-M., C. D.: A New Species of Salvia from Mexico. Bot. Mus. Leafl. Harvard Univ. 1962, 20, 75–76. Unofficial transcription available online. URL http://www.sagewisdom.org/epling&jativa.html

[5] Wasson, R. G.: A New Mexican Psychotropic Drug from the Mint Fam- ily. Bot. Mus. Leafl. Harvard Univ. 1962, 20, 77–84. Unofficial transcrip- tion available online. URL http://www.sagewisdom.org/wasson1.html

[6] Reisfield, A. S.: The Botany of Salvia divinorum (Labiatae). Sida 1993, 15, 349–366. HTML adaptation available online. URL http://www.sagewisdom.org/reisfield.html

[7] Ott, J.: Ethnopharmacognosy and Human Pharmacology of Salvia di- vinorum and Salvinorin A. Curare 1995, 18, 103–129. See pp. 112 (use by Mexican teenagers), 116 - 118 (pipiltzintzintli), 126 (failure of animal testing). Unofficial transcription available online. URL http://www.sagewisdom.org/ott2.html

[8] Photo by Slava Olcheski, Virginia.

[9] Valdés, III, L. J. J.: The early history of Salvia divinorum. Entheogen Rev. 2001, 10, 73–75,80. URL http://www.sagewisdom.org/earlysdhistory.html

[10] Hofmann, A.: LSD: My Problem Child. McGraw-Hill, New York 1980. Translated by Jonathan Ott. See Chapters 1 and 2 (animal tests of LSD),

239 240 BIBLIOGRAPHY

6(S. divinorum). Unofficial transcription available online. URL http://www.druglibrary.org/schaffer/LSD/child.htm

[11] Valdés, III, L. J. J.; Díaz, J. L.; Paul, A. G.: Ethnopharma- cology of ska Maria Pastora (Salvia divinorum, Epling and Játiva- M.). J. Ethnopharmacol. 1983, 7, 287–312. http://dx.doi.org/10.1016/ 0378-8741(83)90004-1. Unofficial transcription available online. URL http://www.sagewisdom.org/valdes83.html

[12] Siebert, D. J.: The History of the First Salvia divinorum Plants Culti- vated Outside of Mexico. Entheogen Rev. 2003, 12, 117–118. URL http://www.sagewisdom.org/salviahistory.html

[13] Valdés, III, L. J. J.; Hatfield, G. M.; Koreeda, M.; Paul, A. G.: Studies of Salvia divinorum (Lamiaceae), an Hallucinogenic Mint from the Sierra Mazateca in Oaxaca, Central Mexico. Econ. Bot. 1987, 41, 283–291. Unofficial transcription available online. URL http://www.sagewisdom.org/valdes87.html

[14]Valdés,III,L.J.J.: Salvia divinorum and the Unique Diterpene Hal- lucinogen, Salvinorin (Divinorin) A. J. Psychoactive Drugs 1994, 26, 277–283. Unofficial transcription available online. URL http://www.sagewisdom.org/valdes94.html

[15]Valdés,III,L.J.J.: The Pharmacognosy of Salvia divinorum (Epling And Játiva-M): an Investigation of Ska María Pastora.Ph.D.thesis, University Of Michigan, Ann Arbor, MI 1983. ProQuest Publication Number: AAT 8402393 (document ID: 750217621). See p. 49 - 50 (“re- medial use”). URL http://wwwlib.umi.com/dissertations/fullcit/8402393

[16] Ref 15, pp. 80, 129.

[17] Valdés, III, L. J. J.: La Divina Pastora (The Divine Shepherdess), a Background to Studies of Salvia divinorum. Salvia Divinorum Magazine 2003, 1(2), 6–9. URL http://salvia.us/magazine.html

[18] Siebert, D. J.: Salvia divinorum and salvinorin A: New pharmaco- logic findings. J. Ethnopharmacol. 1994, 43, 53–56. http://dx.doi.org/ 10.1016/0378-8741(94)90116-3. Free transcription available. URL http://www.sagewisdom.org/jep.html

[19] Wasson, R. G.: Notes on the Present Status of Ololiuhqui and the Other Hallucinogens of Mexico. Bot. Mus. Leafl. Harvard Univ. 1963, 20, 161– 212. Unofficial transcription available online. URL http://pages.prodigy.com/GBonline/liquix.htm BIBLIOGRAPHY 241

[20] Díaz, J. L.: Ethnopharmacology of Sacred Psychoactive Plants Used by the Indians of Mexico. Annu. Rev. Pharmacol. Toxicol. 1977, 17, 647– 675. See p. 656. URL http://dx.doi.org/10.1146/annurev.pa.17.040177.003243

[21] Ref 15, pp. 4 - 5.

[22] Ortega, A.; Blount, J. F.; Manchand, P. S.: Salvinorin, a New trans- Neoclerodane Diterpene from Salvia divinorum (Labiatae). J. Chem. Soc., Perkin Trans. 1 1982, pp. 2505–2508. URL http://dx.doi.org/10.1039/P19820002505

[23] Valdés, III, L. J. J.; Butler, W. M.; Hatfield, G. M.; Paul, A. G.; Koreeda, M.: Divinorin A, a Psychotropic Terpenoid, and Divinorin B from the Hallucinogenic Mexican Mint Salvia divinorum. J. Org. Chem. 1984, 49, 4716–4720. URL http://dx.doi.org/10.1021/jo00198a026

[24] Koreeda, M.; Brown, L.; Valdés, III, L. J. J.: The absolute stereochem- istry of salvinorins. Chem. Lett. 1990, pp. 2015–2018.

[25] Harding, W. W.; Tidgewell, K.; Schmidt, M.; Shah, K.; Dersch, C. M.; Snyder, J.; Parrish, D.; Deschamps, J. R.; Rothman, R. B.; Prisinzano, T. E.: Salvinicins A and B, New Neoclerodane Diterpenes from Salvia divinorum. Org. Lett. 2005, 7, 3017–3020. URL http://dx.doi.org/10.1021/ol0510522

[26] Rodriguez-Hahn, L.; Esquivel, B.; Cardenas, J.: Clerodane diterpenes in Labiatae. Prog. Chem. Org. Nat. Prod. 1994, 63, 107–196. See p. 110 (absolute stereochemistry).

[27] Valdés, III, L. J. J.; Chang, H. M.; Visger, D. C.; Koreeda, M.: Salvi- norin C, a New Neoclerodane Diterpene from a Bioactive Fraction of the Hallucinogenic Mexican Mint Salvia divinorum. Org. Lett. 2001, 3, 3935–3937. URL http://dx.doi.org/10.1021/ol016820d

[28] Valdés, III, L. J. J.: Loliolide from Salvia divinorum. J. Nat. Prod. 1986, 49, 171. URL http://dx.doi.org/10.1021/np50043a031

[29] Giroud, C.; Felber, F.; Augsburger, M.; Horisberger, B.; Rivier, L.; Man- gin, P.: Salvia divinorum: an hallucinogenic mint which might become a new recreational drug in Switzerland. Forensic Sci. Int. 2000, 112, 143–150. URL http://dx.doi.org/10.1016/S0379-0738(00)00180-8

[30] Merritt, A. T.; Ley, S. V.: Clerodane diterpenoids. Nat. Prod. Rep. 1992, 9, 243–287. URL http://dx.doi.org/10.1039/NP9920900243 242 BIBLIOGRAPHY

[31] Rogers, D.; Unal, G. G.; Williams, D. J.; Ley, S. V.; Sim, G. A.; Joshi, B. S.; Ravindranath, K. R.: The crystal structure of 3-epicaryoptin and the reversal of the currently accepted absolute configuration of clerodin. J. Chem. Soc., Chem. Commun. 1979, pp. 97–99. URL http://dx.doi.org/10.1039/C39790000097 [32] McCrindle, R.; Overton, K. H.: The chemistry of cyclic diterpenoids. In Advances in Organic Chemistry: Methods and Results., R. A. Raphael; E. C. Taylor; H. Wynberg, eds., Wiley Interscience, New York, vol. 5, pp. 47–105 1965. [33] Cane, D. E.: Isoprenoid Biosynthesis: Overview. In Comprehensive Nat- ural Products Chemistry, D. E. Cane; D. H. R. Barton; K. Nakanishi; O. Meth-Cohn, eds., Elsevier, New York, vol. 2, pp. 1–14 1999. URL http://www.amazon.com/gp/reader/0080431542 [34] MacMillan, J.; Beale, M. H.: Diterpene Biosynthesis. In Comprehensive Natural Products Chemistry,D.E.Cane;D.H.R.Barton;K.Nakanishi; O. Meth-Cohn, eds., Elsevier, New York, vol. 2, pp. 221–243 1999. URL http://www.amazon.com/gp/reader/0080431542 [35] Nabeta, K.; Ishikawa, T.; Kawae, T.; Okuyama, H.: Biosynthesis of het- eroscyphic acid A in cell cultures of Heteroscyphus planus: nonequivalent labelling of C-5 units in diterpene biosynthesis. J. Chem. Soc., Chem. Commun. 1995, pp. 681–682. URL http://dx.doi.org/10.1039/C39950000681 [36] Akhila, A.; Rani, K.; Thakur, R. S.: Biosynthesis of the clerodane furano-diterpene lactone skeleton in Tinospora cordifolia. Phytochem- istry 1991, 30, 2573–2576. URL http://dx.doi.org/10.1016/0031-9422(91)85103-7 [37] Rohmer, M.: A Mevalonate-independent Route to Isopentenyl Diphos- phate. In Comprehensive Natural Products Chemistry,D.E.Cane; D. H. R. Barton; K. Nakanishi; O. Meth-Cohn, eds., Elsevier, New York, vol. 2, pp. 45–68 1999. URL http://www.amazon.com/gp/reader/0080431542 [38] Eisenreich, W.; Bacher, A.; Arigoni, D.; Rohdich, F.: Biosynthesis of isoprenoids via the non-mevalonate pathway. Cell. Mol. Life Sci. 2004, 61, 1401–1426. URL http://dx.doi.org/10.1007/s00018-004-3381-z [39] SEM image created in collaboration with Dr. Simon Crawford, University of Melbourne. [40] Siebert, D. J.: Localization of Salvinorin A and Related Compounds in Glandular Trichomes of the Psychoactive Sage, Salvia divinorum. Ann. Bot. 2004, 93, 763–771. URL http://dx.doi.org/10.1093/aob/mch089 BIBLIOGRAPHY 243

[41] Díaz, J. L.: Etnofarmacología de algunos psicotrópicos vegetales de México. In Etnofarmacología de Plantas Alucinógenas Latinoamericanas, J. L. Díaz, ed., Centro Mexicano de Estudios en Farmacodependencia, Mexico, vol. 4 of Cuadernos Científicos CEMEF, pp. 135–201 1975.See pp. 149 - 153 (isolation and testing), pp. 153 - 154 (Leonurus sibiricus as Cannabis substitute), pp. 196 - 197 (English abstract).

[42] Ref 15, pp. 20 - 26. Note: “HAc” is intended to refer to AcOH.

[43] Valdés, III, L. J. J. Email, 18th July 2003.

[44] Habib, A. A.: False-positive alkaloid reactions. J. Pharm. Sci. 1980, 69, 37–43. And references therein.

[45] Lüdy-Tenger, F.: Versuche zur mikrochemischen Identifizierung einiger Pyridinderivate. Pharm. Acta Helv. 1953, 28, 22–26.

[46] McConnell Davis, T. W.; Farmilo, C. G.; Genest, K.: Analysis of an impure heroin seizure. Bull. Narc. 1962, 14, 47–57. URL http://www.unodc.org/unodc/bulletin/bulletin_1962-01-01_3_ page006.html

[47] Díaz, J. L.: Ethnopharmacology and taxonomy of Mexican psy- chodysleptic plants. J. Psychedelic Drugs 1979, 11, 71–101.

[48] Scifinder Scholar (Chemical Abstracts Service). Searched under the exact phrases “salvia” along with “alkaloid” or “nitrogenous” or “nitrogen- containing”. Accessed March 2006. URL http://www.cas.org/SCIFINDER/scicover2.html

[49] Web of Science (Thomson Scientific). Searched under “salvia” along with “alkaloid” or “nitrogenous” or “nitrogen-containing”. Accessed March 2006. URL http://scientific.thomson.com/products/wos/

[50] Dictionary of Natural Products. Chapman and Hall/CRC Press, web version 2004(2) ed. 2004. Searched for “N” in molecular formula field, and “salvia” in “biological source” field. Accessed March 2006. URL http://www.chemnetbase.com/scripts/dnpweb.exe

[51]Raffauf,R.F.:Plant Alkaloids: A Guide to Their Discovery and Distri- bution. Food Products Press, Binghamton, NY 1996. See pp. 111 - 113 (S. apiana, azurea, ballotaeflora and leucophylla).

[52] Williams, C. H.; Hines, H. J. G.: The toxic properties of Salvia reflexa. Aust. Vet. J. 1940, 16, 14–20.

[53] Raffauf, R. F.: Handbook of Alkaloids and Alkaloid-Containing Plants. Wiley Interscience, New York 1970. 244 BIBLIOGRAPHY

[54] Werle, E.; Raub, A.: Über Vorkommen, Bildung Und Abbau Biogener Amine Bei Pflanzen Unter Besonderer Berücksichtigung Des Histamins. Biochem. Z. 1948, 318, 538–553. English abstract available via Scifinder (CAN 43:15521).

[55] Don, M.-J.; Shen, C.-C.; Lin, Y.-L.; Syu, Wan, J.; Ding, Y.-H.; Sun, C.- M.: Nitrogen-Containing Compounds from Salvia miltiorrhiza. J. Nat. Prod. 2005, 68, 1066–1070. URL http://dx.doi.org/10.1021/np0500934

[56] Ref 15, p. 131.

[57] Valdés, III, L. J. J.: Comments on the Report of Alkaloids in Salvia divinorum by J. L. Díaz in 1975. Salvia divinorum Magazine 2003, 1(2), 11. URL http://salvia.us/magazine.html

[58] Ref 15, pp. 104, 127 - 128.

[59] Medana, C.; Massolino, C.; Pazzi, M.; Baiocchi, C.: Determination of salvinorins and divinatorins in Salvia divinorum leaves by liquid chro- matography/multistage mass spectrometry. Rapid Commun. Mass Spec- trom. 2006, 20, 131–136. URL http://dx.doi.org/10.1002/rcm.2288

[60] Silverstein, R. M.; Webster, F. X.; Kiemle, D.: Spectrometric Identifica- tion of Organic Compounds. Wiley, New York, 7th ed. 2005.Seep.14 (nitrogen rule and corollary).

[61] Ref 15, pp. 101, 103.

[62] Ott, J.: Pharmacotheon: Entheogenic drugs, their plant sources and his- tory. Natural Products Co., Kennewick, WA 1993. See esp. p. 153 (fail- ure of animal testing).

[63] Ref 15, pp. 103 - 105.

[64] Coughenour, L. L.; McLean, J. R.; Parker, R. B.: A new device for the rapid measurement of impaired motor function in mice,. Pharmacol. Biochem. Behav. 1977, 6, 351–353. URL http://dx.doi.org/10.1016/0091-3057(77)90036-3

[65] Ref 15, pp. 160 - 167. Data from Table 3, p. 164. Fractions plotted are divinorin A, TLC 2 (impure fraction), TLC 3 (much lower in 1a), “CHCl3/MeOH” and 90% MeOH. The 1a contents of the mixed frac- tions were estimated as follows. 1) “CHCl3/MeOH” fraction (actually CH2Cl2/MeOH): as noted on p. 165, this was the “oily solids” from Fig. 12, p. 122. Preparative TLC of 1.03 g this (Fig. 25, p. 158) gave TLC 2 (66 mg) and 3 (60 mg). The 1a was contained in these fractions; the level was “much higher” in TLC 2 than TLC 3 (p. 159). The CH2Cl2/MeOH BIBLIOGRAPHY 245

fraction (“oily solids”) therefore contained between 60 and 126 mg of 1a (6–12%), probably closer to 6%. 2) 90% MeOH Fraction: as noted on p. 165, see the partitioning on p. 137. Chromatography on 9.36 g of this fraction gave 1.51 g of yellowish crystals, which yielded 892 mg of 1a after recrystallization. The 90% MeOH fraction therefore contained 9.5–16% of 1a.

[66] Brimblecombe, R. W.; Green, A. L.: Effect of Monoamine Oxidase In- hibitors on the Behaviour of Rats in Hall’s Open Field. Nature 1962, 194, 983. URL http://dx.doi.org/10.1038/194983a0

[67] Roth, B. L.; Baner, K.; Westkaemper, R.; Siebert, D. J.; Rice, K. C.; Steinberg, S.; Ernsberger, P.; Rothman, R. B.: Salvinorin A: A potent naturally occurring nonnitrogenous κ-opioid selective agonist. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 11934–11939. URL http://dx.doi.org/10.1073/pnas.182234399

[68] Ref 15, pp. 166 - 167.

[69] Ref 15, pp. 144 - 148.

[70] Ref 15, p. 140.

[71] Pouton, C. W.: Lipid formulations for oral administration of drugs: non-emulsifying, self-emulsifying and ‘self-microemulsifying’ drug deliv- ery systems. Eur. J. Pharm. Sci. 2000, 11, S93–98. URL http://dx.doi.org/10.1016/S0928-0987(00)00167-6

[72] ACD/Labs v8.14 for Solaris. See Scifinder (Ref. 48). URL http://www.acdlabs.com

[73] Ref 15. See the CH2Cl2/MeOH (p. 122) and 90% MeOH fractions (p. 135).

[74] Hancock, B. C.; Parks, M.: What is the True Solubility Advantage for Amorphous Pharmaceuticals? Pharm. Res. 2000, 17, 397–404. See also references therein. URL http://dx.doi.org/10.1023/A:1007516718048

[75] Griesser, U.; Stowell, J. G.: Solid-state analysis and polymorphism. In Pharmaceutical Analysis, D. C. Lee; M. Webb, eds., Blackwell, Oxford, pp. 240–294 2003. URL http://www.amazon.com/gp/reader/0849328144

[76] Eder, M.; Mehnert, W.: Bedeutung Pflanzlicher Begleitstoffe in Extrak- ten. Pharmazie 1998, 53, 285–293. URL http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve& db=PubMed&dopt=Citation&list_uids=9631497 246 BIBLIOGRAPHY

[77] Ghafourian, T.; Zandasrar, P.; Hamishekar, H.; Nokhodchi, A.: The ef- fect of penetration enhancers on drug delivery through skin: a QSAR study. J. Controlled Release 2004, 99, 113–125. See also references therein. URL http://dx.doi.org/10.1016/j.jconrel.2004.06.010

[78] Zhang, Y.; Butelman, E.; Schlussman, S.; Ho, A.; Kreek, M.: Effects of the plant-derived hallucinogen salvinorin A on basal dopamine levels in the caudate putamen and in a conditioned place aversion assay in mice: agonist actions at kappa opioid receptors. Psychopharmacology 2005, 179, 551–558. URL http://dx.doi.org/10.1007/s00213-004-2087-0

[79]Fantegrossi,W.E.;Kugle,K.M.;Valdés,III,L.J.J.;Koreeda,M.; Woods, J. H.: Kappa-opioid receptor-mediated effects of the plant- derived hallucinogen, salvinorin A, on inverted screen performance in the mouse. Behav. Pharmacol. 2005, 16, 627–633. URL http://www.behaviouralpharm.com/pt/re/bpharm/abstract. 00008877-200512000-00005.htm

[80] McCurdy, C. R.; Sufka, K. J.; Smith, G. H.; Warnick, J. E.; Nieto, M. J.: Antinociceptive profile of salvinorin A, a structurally unique kappa opi- oid receptor agonist. Pharmacol. Biochem. Behav. 2006, 83, 109–113. URL http://dx.doi.org/10.1016/j.pbb.2005.12.011

[81] Wang, Y.; Tang, K.; Inan, S.; Siebert, D. J.; Holzgrabe, U.; Lee, D. Y.; Huang, P.; Li, J.-G.; Cowan, A.; Liu-Chen, L.-Y.: Comparison of Phar- macological Activities of Three Distinct κ-ligands (Salvinorin A, TRK- 820 and 3FLB) on κ Opioid Receptors in vitro and their Antipruritic and Antinociceptive Activities in vivo. J. Pharmacol. Exp. Ther. 2005, 312, 220–230. URL http://dx.doi.org/10.1124/jpet.104.073668

[82] Harding, W. W.; Tidgewell, K.; Byrd, N.; Cobb, H.; Dersch, C. M.; Butelman, E. R.; Rothman, R. B.; Prisinzano, T. E.: Neoclerodane Diterpenes as a Novel Scaffold for µ Opioid Receptor Ligands. J. Med. Chem. 2005, 48, 4765–4771. Mesylate = compound 15. URL http://dx.doi.org/10.1021/jm048963m

[83] Butelman, E. R.; Harris, T. J.; Kreek, M. J.: The plant-derived hallu- cinogen, salvinorin A, produces κ-opioid agonist-like discriminative ef- fects in rhesus monkeys. Psychopharmacology 2004, 172, 220–224. URL http://dx.doi.org/10.1007/s00213-003-1638-0

[84] Schmidt, M. D.; Schmidt, M. S.; Butelman, E. R.; Harding, W. W.; Tidgewell, K.; Murry, D. J.; Kreek, M. J.; Prisinzano, T. E.: Phar- macokinetics of the plant-derived κ-opioid hallucinogen salvinorin A in nonhuman primates. Synapse 2005, 58, 208–210. URL http://dx.doi.org/10.1002/syn.20191 BIBLIOGRAPHY 247

[85] Turner, D. M.: Salvinorin - The Psychedelic Essence of Salvia Divino- rum. Panther Press, San Francisco, CA 1996. URL http://www.lavondyss.com/donut/scov.html

[86] Gartz, J.: Salvia divinorum: Die Wahrsagesalbei. Nachtschatten Verlag, Solothurn, Switzerland 2001. See pp. 39 - 39 (extraction procedure), 62 - 63 (sublingual 1a in DMSO).

[87] Schmidt, M. S.; Prisinzano, T. E.; Tidgewell, K.; Harding, W.; Butel- man, E. R.; Kreek, M. J.; Murry, D. J.: Determination of Salvinorin A in body fluids by high performance liquid chromatography-atmospheric pressure chemical ionization. J. Chromatogr. B. 2005, 818, 221–225. URL http://dx.doi.org/10.1016/j.jchromb.2004.12.041

[88] Pichini, S.; Abanades, S.; Farré, M.; Pellegrini, M.; Marchei, E.; Pacifici, R.; de la Torre, R.; Zuccaro, P.: Quantification of the plant-derived hal- lucinogen Salvinorin A in conventional and non-conventional biological fluids by gas chromatography/mass spectrometry after Salvia divinorum smoking. Rapid Commun. Mass Spectrom. 2005, 19, 1649–1656. URL http://dx.doi.org/10.1002/rcm.1970

[89] Aardvark, D., ed.: Salvia Divinorum and Salvinorin A: the Best of the Entheogen Review, 1992-2000. Entheogen Review, Sacramento, 2nd ed. 2001.

[90] Siebert, D. J. Email, 16th March 2004.

[91] Shulgin, A. Email, 13th Dec. 2005.

[92] Hofmann, A.: Mexikanische Zauberdrogen und ihre Wirkstoffe. Planta Med. 1964, 12, 341–352. URL http://www.erowid.org/references/refs_view.php?A=ShowDoc1& ID=2715

[93] Ref 15, p. 19.

[94] Valdés, III, L. J. J. Email, 19th Aug. 2002.

[95] Béguin, C.; Richards, M. R.; Wang, Y.; Chen, Y.; Liu-Chen, L.-Y.; Ma, Z.; Lee, D. Y. W.; Carlezon, W. A.; Cohen, B. M.: Synthesis and in vitro pharmacological evaluation of salvinorin A analogues modified at C(2). Bioorg.Med.Chem.Lett.2005, 15, 2761–2765. URL http://dx.doi.org/10.1016/j.bmcl.2005.03.113

[96] Armbruster, B. N.; Roth, B. L.: Mining the receptorome. J. Biol. Chem. 2005, 280, 5129–5132. URL http://dx.doi.org/10.1074/jbc.R400030200

[97] Sheffler, D. J.; Roth, B. L.: Salvinorin A: the “magic mint” hallucinogen finds a molecular target in the kappa opioid receptor. Trends Pharmacol. 248 BIBLIOGRAPHY

Sci. 2003, 24, 107–109. URL http://dx.doi.org/10.1016/S0165-6147(03)00027-0

[98] Sneader, W.: Drug Discovery: A History. Wiley, New York 2005.See esp. pp. 90, 91, 115. URL http://books.google.com/books?ie=UTF-8& vid=ISBN0471899801

[99] Blakemore, P. R.; White, J. D.: Morphine, the Proteus of organic molecules. Chem. Commun. 2002, pp. 1159–1168. URL http://dx.doi.org/10.1039/b111551k

[100] Buckingham, J.: Chasing the molecule. Sutton, Stroud, UK 2004.

[101] Rees, D. C.; Hunter, J. C.: Opioid receptors. In Comprehensive medicinal chemistry : the rational design, mechanistic study and therapeutic appli- cation of chemical compounds,J.C.Emmett;C.Hansch;P.G.Sammes; J. B. Taylor, eds., Pergamon Press, New York, vol. 3, pp. 805–846 1990.

[102] Dhawan, B.; Cesselin, F.; Raghubir, R.; Reisine, T.; Bradley, P.; Por- toghese, P.; Hamon, M.: International Union of Pharmacology. XII. Clas- sification of opioid receptors. Pharmacol. Rev. 1996, 48, 567–592. URL http://pharmrev.aspetjournals.org/cgi/reprint/48/4/567

[103] Poeaknapo, C.; Schmidt, J.; Brandsch, M.; Drager, B.; Zenk, M. H.: Endogenous formation of morphine in human cells. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 14091–14096. URL http://dx.doi.org/10.1073/pnas.0405430101

[104] Szmuszkovicz, J.: U-50,488 and the κ receptor. A personalized account covering the period 1973 to 1990. Prog. Drug Res. 1999, 52, 167–195. URL http://books.google.com/books?ie=UTF-8& vid=ISBN376435979X

[105] Szmuszkovicz, J.: U-50,488 and the κ receptor. Part 2. 1991-1998. Prog. Drug Res. 1999, 53, 1–51. URL http://books.google.com/books?ie=UTF-8& vid=ISBN3764360283

[106]vanRee,J.M.;Gerrits,M.A.F.M.;Vanderschuren,L.J.M.J.: Opioids, Reward and Addiction: An Encounter of Biology, Psychology, and Medicine. Pharmacol. Rev. 1999, 51, 341–396. URL http://pharmrev.aspetjournals.org/cgi/content/abstract/51/2/ 341

[107] Carlezon Jr., W. A.; Béguin, C.; DiNieri, J. A.; Baumann, M. H.; Richards, M. R.; Todtenkopf, M. S.; Rothman, R. B.; Ma, Z.; Lee, D. Y. W.; Cohen, B. M.: Depressive-like Effects of the κ-Opioid Recep- tor Agonist Salvinorin A on Behavior and Neurochemistry in Rats. J. BIBLIOGRAPHY 249

Pharmacol. Exp. Ther. 2005, 316, 440–447. And references therein. URL http://dx.doi.org/10.1124/jpet.105.092304 [108] Pande, A. C.; Pyke, R. E.; Greiner, M.; Wideman, G. L.; Benjamin, R.; Pierce, M. W.: Analgesic efficacy of enadoline versus placebo or morphine in postsurgical pain. Clin. Neuropharmacol. 1996, 19, 451–456. URL http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve& db=PubMed&dopt=Citation&list_uids=8889289 [109] Walsh, S. L.; Strain, E. C.; Abreu, M. E.; Bigelow, G. E.: Enadoline, a selective kappa opioid agonist: comparison with butorphanol and hy- dromorphone in humans. Psychopharmacology 2001, 157, 151–162. URL http://dx.doi.org/10.1007/s002130100788 [110] Chappell, P. B.; Leckman, J. F.; Scahill, L. D.; Hardin, M. T.; Anderson, G.; Cohen, D. J.: Neuroendocrine and behavioral effects of the selective kappa agonist spiradoline in Tourette’s syndrome: a pilot study. Psychi- atry Res. 1993, 47, 267–280. URL http://dx.doi.org/10.1016/0165-1781(93)90084-T [111] Gadano, A.; Moreau, R.; Pessione, F.; Trombino, C.; Giuily, N.; Sinnas- samy, P.; Valla, D.; Lebrec, D.: Aquaretic effects of niravoline, a κ-opioid agonist, in patients with cirrhosis. J. Hepatol. 2000, 32, 38–42. URL http://dx.doi.org/10.1016/S0168-8278(00)80187-7 [112] Pfeiffer, A.; Brantl, V.; Herz, A.; Emrich, H. M.: Psychotomimesis me- diated by κ opiate receptors. Science 1986, 233, 774–776. URL http://dx.doi.org/10.1126/science.3016896 [113] Kumor, K. M.; Haertzen, C. A.; Johnson, R. E.; Kocher, T.; Jasinski, D.: Human psychopharmacology of ketocyclazocine as compared with cyclazocine, morphine and placebo. J. Pharmacol. Exp. Ther. 1986, 238, 960–968. URL http://jpet.aspetjournals.org/cgi/content/abstract/238/3/960 [114] Greenwald, M. K.; Stitzer, M. L.: Butorphanol agonist effects and acute physical dependence in opioid abusers: comparison with morphine. Drug Alcohol Depend. 1998, 53, 17–30. URL http://dx.doi.org/10.1016/S0376-8716(98)00104-5 [115] Barber, A.; Gottschlich, R.: Novel developments with selective, non- peptidic kappa-opioid receptor agonists. Expert Opin. Invest. Drugs 1997, 6, 1351–1368. URL http://www.ingentaconnect.com/content/apl/eid/1997/00000006/ 00000010/art00004 [116] Bush, K. A.; Kirkham, B. W.; Walker, J. S.: The κ-opioid agonist, asi- madoline, alters cytokine gene expression in adjuvant arthritis. Rheuma- tology (Oxford). 2001, 40, 1013–1021. 250 BIBLIOGRAPHY

URL http://rheumatology.oxfordjournals.org/cgi/content/abstract/40/ 9/1013

[117] Riviere, P. J. M.: Peripheral kappa-opioid agonists for visceral pain. Br. J. Pharmacol. 2004, 141, 1331–1334. URL http://dx.doi.org/10.1038/sj.bjp.0705763

[118] Dortch-Carnes, J.; Potter, D. E.: Bremazocine: a κ-opioid agonist with potent analgesic and other pharmacologic properties. CNS Drug Rev. 2005, 11, 195–212. URL http://www.nevapress.com/cnsdr/contents/1102.html

[119] Sorbera, L. A.; Castaner, J.; Leeson, P. A.: Nalfurafine hydrochloride. Drugs Fut. 2003, 28, 237–242. URL http://dx.doi.org/10.1358/dof.2003.028.03.723584

[120] Wadenberg, M. L.: A review of the properties of spiradoline: a potent and selective κ-opioid receptor agonist. CNS Drug Rev. 2003, 9, 187– 198. URL http://www.nevapress.com/cnsdr/contents/0902.html

[121] Ukai, M.; Suzuki, M.; Mamiya, T.: Effects of U-50,488H, a κ-opioid receptor agonist, on the learned helplessness model of depression in mice. J. Neural Transm. 2002, 109, 1221–1225. URL http://dx.doi.org/10.1007/s00702-002-0764-x

[122] Chourbaji, S.; Zacher, C.; Sanchis-Segura, C.; Dormann, C.; Vollmayr, B.; Gass, P.: Learned helplessness: Validity and reliability of depressive- like states in mice. Brain Res. Protoc. 2005, 16, 70–78. URL http://dx.doi.org/10.1016/j.brainresprot.2005.09.002

[123] Broom, D. C.; Jutkiewicz, E. M.; Folk, J. E.; Traynor, J. R.; Rice, K. C.; Woods, J. H.: Nonpeptidic δ-opioid receptor agonists reduce immobility in the forced swim assay in rats. Neuropsychopharmacology 2002, 26, 744–755. URL http://dx.doi.org/10.1016/S0893-133X(01)00413-4

[124] Mague, S. D.; Pliakas, A. M.; Todtenkopf, M. S.; Tomasiewicz, H. C.; Zhang, Y.; Stevens, W. C., J.; Jones, R. M.; Portoghese, P. S.; Carlezon, W. A., J.: Antidepressant-Like Effects of κ-Opioid Receptor Antagonists in the Forced Swim Test in Rats. J. Pharmacol. Exp. Ther. 2003, 305, 323–330. URL http://dx.doi.org/10.1124/jpet.102.046433

[125] Béguin, C.; Carlezon, W.; Cohen, B. M.; He, M.; Lee, D. Y.-W.; Richards, M. R.: Salvinorin derivatives and uses thereof 2005.Inter- national Patent WO2005089745. See pp. 64 (hydrogenation), 65 (LiAlH4 reduction), 66 [Ba(OH)2 in MeOH]. URL http://www.patentmatic.com/downloadform.php BIBLIOGRAPHY 251

[126] Hanes, K. R.: Antidepressant effects of the herb Salvia divinorum: a case report. J. Clin. Psychopharmacol. 2001, 21, 634–635. URL http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve& db=PubMed&dopt=Citation&list_uids=11763023

[127] Hanes, K. R.: Salvia divinorum: Clinical and Research Potential. MAPS Bulletin 2003, 13, 18–20. URL http://www.maps.org/news-letters/v13n1/13118han.pdf

[128] Endoh, T.; Tajima, A.; Izumimoto, N.; Suzuki, T.; Saitoh, A.; Narita, M.; Kamei, J.; Tseng, L. F.; Mizoguchi, H.; Nagase, H.: TRK-820, a se- lective κ-opioid agonist, produces potent antinociception in cynomolgus monkeys. Jpn. J. Pharmacol. 2001, 85, 282–290. URL http://dx.doi.org/10.1254/jjp.85.282

[129] Mori, T.; Nomura, M.; Nagase, H.; Narita, M.; Suzuki, T.: Effects of a newly synthesized κ-opioid receptor agonist, TRK-820, on the discrimi- native stimulus and rewarding effects of cocaine in rats. Psychopharma- cology 2002, 161, 17–22. URL http://dx.doi.org/10.1007/s00213-002-1028-z

[130] Tortella, F. C.; Rose, J.; Robles, L.; Moreton, J. E.; Hughes, J.; Hunter, J. C.: EEG Spectral Analysis of the Neuroprotective Kappa Opioids Enadoline and PD117302. J. Pharmacol. Exp. Ther. 1997, 282, 286– 293. URL http://jpet.aspetjournals.org/cgi/content/abstract/282/1/286

[131] Butelman, E. R.; Ko, M. C.; Traynor, J. R.; Vivian, J. A.; Kreek, M. J.; Woods, J. H.: GR89,696: a potent κ-opioid agonist with subtype selec- tivity in rhesus monkeys. J. Pharmacol. Exp. Ther. 2001, 298, 1049– 1059. And references therein. URL http://jpet.aspetjournals.org/cgi/content/full/298/3/1049

[132] Connor, M.; Kitchen, I.: Has the sun set on κ3-opioid receptors? Br. J. Pharmacol. 2006, 147, 349–350. URL http://dx.doi.org/10.1038/sj.bjp.0706603

[133] Schultes, R.; Hofmann, A.: The Botany and Chemistry of Hallucinogens. Charles C. Thomas, Springfield, IL, 2nd ed. 1980. See esp. p. 26 re structural types. Note that the phenylalkylamine skeletal types contain far more active compounds than the others. URL http://www.amazon.com/gp/reader/0398038635

[134] Lloyd, E. J.; Andrews, P. R.: A common structural model for central nervous system drugs and their receptors. J. Med. Chem. 1986, 29, 453– 462. Figures of 82% and 58% taken from the appendix. Other figures calculated based on table IV. URL http://dx.doi.org/10.1021/jm00154a005 252 BIBLIOGRAPHY

[135] Chavkin, C.; Sud, S.; Jin, W.; Stewart, J.; Zjawiony, J. K.; Siebert, D. J.; Toth, B. A.; Hufeisen, S. J.; Roth, B. L.: Salvinorin A, an Ac- tive Component of the Hallucinogenic Sage Salvia divinorum Is a Highly Efficacious κ-Opioid Receptor Agonist: Structural and Functional Con- siderations. J. Pharmacol. Exp. Ther. 2004, 308, 1197–1203. URL http://dx.doi.org/10.1124/jpet.103.059394

[136] Harding, W. W.; Schmidt, M.; Tidgewell, K.; Kannan, P.; Holden, K. G.; Gilmour, B.; Navarro, H.; Rothman, R. B.; Prisinzano, T. E.: Synthetic Studies of Neoclerodane Diterpenes from Salvia divinorum:Semisyn- thesis of Salvinicins A and B and Other Chemical Transformations of Salvinorin A. J. Nat. Prod. 2006, 69, 107–112. URL http://dx.doi.org/10.1021/np050398i

[137] Lee, D. Y. W.; Ma, Z.; Liu-Chen, L.-Y.; Wang, Y.; Chen, Y.; Carlezon, W. A.; Cohen, B.: New neoclerodane diterpenoids isolated from the leaves of Salvia divinorum and their binding affinities for human κ-opioid receptors. Bioorg.Med.Chem.2005, 13, 5635–5639. URL http://dx.doi.org/10.1016/j.bmc.2005.05.054

[138] Nagase, H.; Hayakawa, J.; Kawamura, K.; Kawai, K.; Takezawa, Y.; Matsuura, H.; Tajima, C.; Endo, T.: Discovery of a structurally novel opioid κ-agonist derived from 4,5-epoxymorphinan. Chem. Pharm. Bull. 1998, 46, 366–369. URL http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve& db=PubMed&dopt=Citation&list_uids=9501472

[139] Seki, T.; Awamura, S.; Kimura, C.; Ide, S.; Sakano, K.; Minami, M.; Na- gase, H.; Satoh, M.: Pharmacological properties of TRK-820 on cloned µ-, δ-andκ-opioid receptors and nociceptin receptor. Eur. J. Pharma- col. 1999, 376, 159–167. URL http://dx.doi.org/10.1016/S0014-2999(99)00369-6

[140] Zech, D. F.; Grond, S.; Lynch, J.; Hertel, D.; Lehmann, K. A.: Valida- tion of World Health Organization Guidelines for cancer pain relief: a 10-year prospective study. Pain 1995, 63, 65–76. URL http://dx.doi.org/10.1016/0304-3959(95)00017-M

[141] Australian and New Zealand College of Anaesthetists: Acute Pain Man- agement: Scientific Evidence. National Health and Medical Research Council, Canberra, 2nd ed. 2005. URL http://www.nhmrc.gov.au/publications/_files/cp104.pdf

[142] Ref 15, pp. 103, 112.

[143] Ref 15, p. 165.

[144] Lewis, Sr, R. J.: Niacin (NCQ900): record number 16623. In Sax’s Dan- gerous Properties of Industrial Materials, Wiley, New York, 10th ed. BIBLIOGRAPHY 253

2000. URL http://www.knovel.com/knovel2/Toc.jsp?BookID=707

[145] ten Tije, A. J.; Verweij, J.; Loos, W. J.; Sparreboom, A.: Phar- macological effects of formulation vehicles : implications for cancer chemotherapy. Clin. Pharmacokinet. 2003, 42, 665–685. URL http://www.ingentaconnect.com/content/adis/cpk/2003/ 00000042/00000007/art00005

[146] Mowry, M.; Mosher, M.; Briner, W.: Acute physiologic and chronic histologic changes in rats and mice exposed to the unique hallucinogen salvinorin A. J. Psychoactive Drugs 2003, 35, 379 – 382. Manuscript available online. Accessed March 2006. URL http://www.sagewisdom.org/mowryetal.pdf

[147] Boire, R. G.; Russo, E.; Fish, A. R.; Bowman, J.: Salvia divinorum: in- formation concerning the plant and its active principle. Center for Cog- nitive Liberty and Ethics, Davis, CA, 2nd ed. 2002. URL http://www.cognitiveliberty.org/pdf/salvia_dea.pdf

[148] Gottlieb, A.: Legal Highs. High Times/Level Press, San Francisco 1973.

[149] Grubber, H.: Growing the Hallucinogens. High Times/Level Press, San Francisco 1973.

[150] Record of the Reasons, 33rd Meeting, 20-22 November 2001.National Drugs and Poisons Schedule Committee, Therapeutic Goods Adminis- tration, Commonwealth Department of Health and Ageing, Canberra 2001. Accessed March 2006. URL http://tga.health.gov.au/ndpsc/record/rr200111.pdf

[151] White, W. E.: All About Salvia Divinorum. v 1.0 ed. 1995.Postedto alt.drugs, 17th October 1995. Accessed March 2006. URL http://groups.google.com.au/group/alt.drugs/browse_thread/ thread/93775d77997065f9

[152] Moss, G. P.: Nomenclature of fused and bridged fused ring systems. Pure Appl. Chem. 1998, 70, 143–216. Accessed March 2006. URL http://www.chem.qmul.ac.uk/iupac/fusedring/

[153] Munro, T.: The Prohibition of Diviner’s Sage 2002. Accessed March 2006. URL http://www.thomasmunro.com/salvia.htm

[154] Leask, E.: Should Diviner’s Sage be banned? Australian Vital 2002, 1(3), 149–151.

[155] Record of the Reasons, 36th Meeting, 15-17 October 2002.National Drugs and Poisons Schedule Committee, Therapeutic Goods Adminis- tration, Commonwealth Department of Health and Ageing, Canberra 254 BIBLIOGRAPHY

2002. Accessed March 2006. URL http://tga.health.gov.au/ndpsc/record/rr200210.pdf

[156] alt.drugs (usenet group). Accessed March 2006. URL http://groups.google.com/group/alt.drugs?hl=en

[157] Guidelines for the Classification of Publications. Office of Film and Lit- erature Classification, Canberra 2005. See p. 16. Accessed March 2006. URL http://www.oflc.gov.au/resource.html?resource=63&filename=63. pdf

[158] Andrews, K.: Salvinorin A (Question No. 1963), 11th August 2003. In Parliamentary Debates, House of Representatives, Commonwealth of Australia, Canberra, pp. 18142–18143. Accessed March 2006. URL http://www.aph.gov.au/hansard/reps/dailys/dr110803.pdf

[159] McEwen, J.: Letter to Lindsay Tanner MP, 8th December 2003.

[160] Hartley, M.: Freedom of Information Request No: 117/04. Letter to Karl Hanes, 7th January 2005.

[161] BEK nr 714 af 14/08/2003 (Gældende). Retsinformation, Legal Infor- mation Division, Ministry of Justice, Denmark. URL http://www.retsinfo.dk/delfin/html/b2003/0071405.htm

[162] Siebert, D. J.: The legal status of Salvia divinorum 2006. Accessed March 2006. URL http://www.sagewisdom.org/legalstatus.html

[163] United States. Hallucinogen Control Act of 2002, HR 5607, 107th Congress, 2nd session, 10th October 2002. URL http://thomas.loc.gov/cgi-bin/query/z?c107:H.R.5607:

[164] Gruber, J. W.: Quantification of Salvinorin A from Tissues of Salvia divinorum (Epling & Játiva-M.). Master’s thesis, Philadelphia College of Pharmacy and Science, Philadelphia, PA 1997. ProQuest document ID: 740502451. Publication Number: AAT 1385726. See pp. 46 - 50 (effect of temperature on recovery). URL http://wwwlib.umi.com/dissertations/fullcit/1385726

[165] Valdés, III, L. J. J. Email, 22nd July 2005.

[166] ‘Sphere’: Salvia Divinorum Extractions using Chilled Acetone. Accessed March 2006. URL http://photos.imageevent.com/sphere/ salviadivinorumextractiontech/Chilled_acetone_extraction.pdf

[167] Ref 164, pp. 35-37 (isolation procedure).

[168] Ref 164, pp. 52-54 (purity of 1a). BIBLIOGRAPHY 255

[169] Tidgewell, K.; Harding, W. W.; Schmidt, M.; Holden, K. G.; Murry, D. J.; Prisinzano, T. E.: A facile method for the preparation of deuterium 2 labeled salvinorin A: synthesis of [2,2,2- H3]-salvinorin A. Bioorg. Med. Chem. Lett. 2004, 14, 5099–5102. URL http://dx.doi.org/10.1016/j.bmcl.2004.07.081

[170] Davies, K. M., ed.: Plant pigments and their manipulation. Blackwell, Oxford 2004. See esp. pp. 1 - 22 (general introduction).

[171] Goodwin, T. W., ed.: Chemistry and Biochemistry of Plant Pigments. Academic Press, London 1965. See vol.1 (general introduction), vol. 2 p.19 (partitioning chlorophylls).

[172] Vogel, A. I.; Furniss, B. S.; Hannaford, A. J.; Rogers, V.; Smith, P. W. G.; Tatchell, A. R.: Vogel’s Textbook of Practical Organic Chemistry, Including Qualitative Organic Analysis. Longman, London, 4th ed. 1978. See p. 64 (Rh vs Pd catalysts), 109 - 110 (use of decolourising carbon), p. 493 (KOH/MeOH ester solvolysis).

[173] Hassler, J. W.: Purification with Activated Carbon: industrial, commer- cial, environmental. Chemical Pub. Co., New York, 3rd ed. 1974.See esp. pp. 15 ,16 (general factors governing adsorption).

[174] Sanghi, R.; Bhattacharya, B.: Review on decolorisation of aqueous dye solutions by low cost adsorbents. Color. Technol. 2002, 118, 256–269. Accessed March 2006. URL http://www.ingentaconnect.com/content/sdc/ct/2002/00000118/ 00000005/art00009

[175] Belfort, G.; Altshuler, G. L.; Thallam, K. K.; Feerick, C. P.; Woodfield, K. L.: Selective adsorption of organic homologues onto activated carbon from dilute aqueous solutions: Solvophobic interaction approach: Part IV. Effect of simple structural modifications with aliphatics. AIChE J. 1984, 30, 197–207. URL http://dx.doi.org/10.1002/aic.690300205

[176] Vailaya, A.; Horváth, C.: Retention in reversed-phase chromatography: partition or adsorption? J. Chromatogr., A 1998, 829, 1–27. See esp. pp. 9, 11. URL http://dx.doi.org/10.1016/S0021-9673(98)00727-4

[177] Luehrs, D. C.; Hickey, J. P.; Nilsen, P. E.; Godbole, K. A.; Rogers, T. N.: Linear Solvation Energy Relationship of the Limiting Partition Coeffi- cient of Organic Solutes between Water and Activated Carbon. Environ. Sci. Technol. 1996, 30, 143–152. See esp. Table 1, entries 129, 160, 165 and 174; 277 and 289; 348, 358, 361 and 366 (adsorption constants of benzene, toluene, p-xylene and o-dichlorobenzene from 3 studies - their refs. 14,17 and 27). URL http://dx.doi.org/10.1021/es950200o 256 BIBLIOGRAPHY

[178] Chiou, C. C. T.; Manes, M.: Application of the Polanyi Adsorption Potential Theory to Adsorption from Solution on Activated Carbon. IV. Steric Factors, as Illustrated by the Adsorption of Planar and Octahedral Metal Acetylacetonates. J. Phys. Chem. 1973, 77, 809–813. URL http://dx.doi.org/10.1021/j100625a015

[179] Dąbrowski, A.; Podkościelny, P.; Hubicki, Z.; Barczak, M.: Adsorption of phenolic compounds by activated carbon - a critical review. Chemosphere 2005, 58, 1049–1070. URL http://dx.doi.org/10.1016/j.chemosphere.2004.09.067

[180] Gordon, A. J.; Ford, R. A.: The chemist’s companion: a handbook of practical data, techniques, and references. Wiley, New York 1972.See pp. 371 - 376 (chromatography on activated carbon).

[181] Wu, Z.; Xie, L.; Xia, G.; Zhang, J.; Nie, Y.; Hu, J.; Wang, S.; Zhang, R.: A new tetrodotoxin-producing actinomycete, Nocardiopsis dassonvillei, isolated from the ovaries of puffer fish Fugu rubripes. Toxicon 2005, 45, 851–859. URL http://dx.doi.org/10.1016/j.toxicon.2005.02.005

[182] Naundorf, A.; Ajisaka, K.: Purification of α-N-acetyl-galactosaminidase from Aspergillus niger and its use in the synthesis of GalNAc-α-(1→O)- serine. Enzyme Microb. Technol. 1999, 25, 483–488. URL http://dx.doi.org/10.1016/S0141-0229(99)00073-3

[183] Aksu, Z.; Tunç, O.: Application of biosorption for penicillin G removal: comparison with activated carbon. Process Biochem. 2005, 40, 831–847. URL http://dx.doi.org/10.1016/j.procbio.2004.02.014

[184] Li, S. Z.; Li, X. Y.; Cui, Z. F.; Wang, D. Z.: Application of ultrafiltration to improve the extraction of antibiotics. Sep. Purif. Technol. 2004, 34, 115–123. URL http://dx.doi.org/10.1016/S1383-5866(03)00185-0

[185] Scrivens, W. A.; Bedworth, P. V.; Tour, J. M.: Purification of Gram Quantities of C60. A New Inexpensive and Facile Method. J. Am. Chem. Soc. 1992, 114, 7917 – 7919. URL http://dx.doi.org/10.1021/ja00046a051

[186] Scrivens, W. A.; Cassell, A. M.; North, B. L.; Tour, J. M.: Single Column Purification of Gram Quantities of C70. J. Am. Chem. Soc. 1994, 116, 6939–6940. URL http://dx.doi.org/10.1021/ja00094a060

[187] Athanasiadou, M.; Jensen, S.; Klasson Wehler, E.: Preparative Fraction- ation of a Commercial PCB Product. Chemosphere 1991, 23, 957–970. URL http://dx.doi.org/10.1016/0045-6535(91)90123-U BIBLIOGRAPHY 257

[188] Kočan, A.; Petrík, J.; Chovancová, J.; Drobná, B.: Method for the group separation of non-ortho-, mono-ortho- and multi-ortho- substituted polychlorinated biphenyls and polychlorinated dibenzo- p-dioxins/polychlorinated dibenzofurans using activated carbon chro- matography. J. Chromatogr., A 1994, 665, 139–153. URL http://dx.doi.org/10.1016/0021-9673(94)87042-X

[189] Sericano, J. L.; El-Husseini, A. M.; Wade, T. L.: Isolation of planar poly- chlorinated biphenyls by carbon column chromatography. Chemosphere 1991, 23, 915–924. URL http://dx.doi.org/10.1016/0045-6535(91)90096-V

[190] Zupančič-Kralj, L.; Jan, J.; Marsel, J.: Fractionation of chloroorganic compounds on a carbon cartridge. Chemosphere 1991, 23, 841–843. URL http://dx.doi.org/10.1016/0045-6535(91)90089-V

[191] Lombardo, M. E.; Viscelli, T. A.; Mittelman, A.; Hudson, P. B.: The Removal of Non-steroidal Pigments of Urinary Extracts by Adsorption on Charcoal. J. Biol. Chem. 1955, 212, 353–360. URL http://intl.jbc.org/cgi/reprint/212/1/353

[192] Murai, M.; Takenaka, T.; Nishibe, S.: Iridoids from Plantago major. Nat. Med. 1996, 50,306.

[193] MacRae, H. F.; McKinley, W. P.: Chromatographic Identification of Some Organophosphate Insecticides in the Presence of Plant Extracts. J. Agric. Food Chem. 1963, 11, 174–178. URL http://dx.doi.org/10.1021/jf60126a023

[194] Leonard, J.; Lygo, B.; Procter, G.: Advanced Practical Organic Chem- istry. Blackie Academic & Professional, London, 2nd ed. 1995. See pp. 208 - 214.

[195] Ribeiro, M. H. L.; Lourenco, P. A. S.; Monteiro, J. P.; Ferreira-Dias, S.: Kinetics of selective adsorption of impurities from a crude vegetable oil in hexane to activated earths and carbons. Eur. Food Res. Technol. 2001, 213, 132–138. URL http://dx.doi.org/10.1007/s002170100347

[196] Bigham, A. K.; Munro, T. A.; Rizzacasa, M. A.; Robins-Browne, R. M.: Divinatorins A-C, New Neoclerodane Diterpenoids from the Controlled Sage Salvia divinorum. J. Nat. Prod. 2003, 66, 1242–1244. URL http://dx.doi.org/10.1021/np030313i

[197] Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.: Experi- mental and computational approaches to estimate solubility and perme- ability in drug discovery and development settings. Adv. Drug Delivery Rev. 1997, 23, 3–25. URL http://dx.doi.org/10.1016/S0169-409X(96)00423-1 258 BIBLIOGRAPHY

[198] Kavvadias, D.; Monschein, V.; Sand, P.; Riederer, P.; Schreier, P.: Con- stituents of sage (Salvia officinalis) with in vitro affinity to human brain benzodiazepine receptor. Planta Med. 2003, 69, 113–117. URL http://dx.doi.org/10.1055/s-2003-37712

[199] Huen, M. S. Y.; Hui, K.-M.; Leung, J. W. C.; Sigel, E.; Baur, R.; Wong, J. T.-F.; Xue, H.: Naturally occurring 2’-hydroxyl-substituted flavonoids as high-affinity benzodiazepine site ligands. Biochem. Pharmacol. 2003, 66, 2397–2407. URL http://dx.doi.org/10.1016/j.bcp.2003.08.016

[200] Salgueiro, J. B.; Ardenghi, P.; Dias, M.; Ferreira, M. B. C.; Izquierdo, I.; Medina, J. H.: Anxiolytic natural and synthetic flavonoid ligands of the central benzodiazepine receptor have no effect on memory tasks in rats. Pharmacol. Biochem. Behav. 1997, 58, 887–891. URL http://dx.doi.org/10.1016/S0091-3057(97)00054-3

[201] Zanoli, P.; Avallone, R.; Baraldi, M.: Behavioral characterization of the flavonoids apigenin and chrysin. Fitoterapia 2000, 71, S117–S123. URL http://dx.doi.org/10.1016/S0367-326X(00)00186-6

[202] González, A. G.; Aguiar, Z. E.; Luis, J. G.; Ravelo, A. G.; Vázquez, J. T.; Domínguez, X. A.: Flavonoids from Salvia texana. Phytochemistry 1989, 28, 2871–2872. URL http://dx.doi.org/10.1016/S0031-9422(00)98114-7

[203] Valdés, III, L. J. J. Email, 1st Oct. 2002.

[204] Medana, C. Email, 2nd Nov. 2005.

[205] Goutman, J. D.; Waxemberg, M. D.; Donate-Oliver, F.; Pomata, P. E.; Calvo, D. J.: Flavonoid modulation of ionic currents mediated by GABAA and GABAC receptors. Eur. J. Pharmacol. 2003, 461, 79–87. URL http://dx.doi.org/10.1016/S0014-2999(03)01309-8

[206] Lutterodt, G. D.: Inhibition of gastrointestinal release of acetylcholine by quercetin as a possible mode of action of Psidium guajava leaf extracts in the treatment of acute diarrhoeal disease. J. Ethnopharmacol. 1989, 25, 235–247. URL http://dx.doi.org/10.1016/0378-8741(89)90030-5

[207] Capasso, R.; Borrelli, F.; Capasso, F.; Siebert, D. J.; Stewart, D. J.; Zjawiony, J. K.; Izzo, A. A.: The hallucinogenic herb Salvia divinorum and its active ingredient salvinorin A inhibit enteric cholinergic transmis- sion in the guinea-pig ileum. Neurogastroenterol. Motil. 2006, 18, 69–75. URL http://dx.doi.org/10.1111/j.1365-2982.2005.00725.x

[208] Singh, A.; Naidu, P. S.; Kulkarni, S. K.: Quercetin, a bioflavonoid, reverses development of tolerance and dependence to morphine. Drug BIBLIOGRAPHY 259

Dev. Res. 2002, 57, 167–172. URL http://dx.doi.org/10.1002/ddr.10119

[209] Ma, M.-H.; Lin, C.-I.: Adsorption kinetics of β-carotene from soy oil using regenerated clay. Sep. Purif. Technol. 2004, 39, 201–209. URL http://dx.doi.org/10.1016/j.seppur.2003.12.007

[210] Tanin, S.; Gurgey, I.: Bleaching of cottonseed and sunflower oils by active carbons obtained from rice hulls. Chim. Acta Turc. 1988, 16, 209–219.

[211] Ortega, A. Email, 15th Oct. 2005.

[212] Delgado, G.; Romo de Vivar, A.; Ortega, A.; Cárdenas, J.; Schlemper, E. O.: Diterpenoids from Viguiera insignis. Phytochemistry 1983, 22, 1227–1230. URL http://dx.doi.org/10.1016/0031-9422(83)80227-1

[213] Jimenez, M.; Ortega, A.; Navarro, A.; Maldonado, E.; Van Calsteren, M. R.; Jankowski, C. K.: Reaction of the Molluscicide Glaucolide B with Bentonite. J. Nat. Prod. 1995, 58, 424–427. URL http://dx.doi.org/10.1021/np50117a012

[214] Ruvalcaba, R. M.; Arroyo Razo, G. A.; Carrillo, G. P.; Reyes, F. D.; Ortiz, A. C.; Toledano, C. A.; Salazar, M. S.: Preparative heterocyclic chemistry using tonsil, a bentonitic clay: 1981 to 2003. Trends Hetero- cycl. Chem. 2003, 9, 195–235.

[215] Munro, T. A.; Rizzacasa, M. A.: Salvinorins D-F, New Neoclerodane Diterpenoids from Salvia divinorum, and an Improved Method for the Isolation of Salvinorin A. J. Nat. Prod. 2003, 66, 703–705. URL http://dx.doi.org/10.1021/np0205699

[216] Ohsaki, A.; Yan, L. T.; Ito, S.; Edatsugi, H.; Iwata, D.; Komoda, Y.: The isolation and in vivo potent antitumor activity of clerodane diterpenoid from the oleoresin of the Brazilian medicinal plant, Copaifera langsdorfii Desfon. Bioorg. Med. Chem. Lett. 1994, 4, 2889–2892. URL http://dx.doi.org/10.1016/S0960-894X(01)80834-9

[217] McChesney, J. D.; Clark, A. M.; Silveira, E. R.: Antimicrobial diterpenes of Croton sonderianus, 1. Hardwickic and 3,4-secotrachylobanoic acids. J. Nat. Prod. 1991, 54, 1625–1633. URL http://dx.doi.org/10.1021/np50078a021

[218] Costa, M.; Tanaka, C. M. A.; Imamura, P. M.; Marsaioli, A. J.: Iso- lation and synthesis of a new clerodane from Echinodorus grandiflorus. Phytochemistry 1999, 50, 117–122. URL http://dx.doi.org/10.1016/S0031-9422(98)00464-6

[219] Spectral Database for Organic Compounds (SDBS), National Institute of Advanced Industrial Science and Technology (AIST), Japan. (E)-Phytol 260 BIBLIOGRAPHY

= compound 7578 (accessed Oct. 2005). URL http://www.aist.go.jp/RIODB/SDBS/cgi-bin/cre_index.cgi

[220] Boxer, S. G.; Closs, G. L.; Katz, J. J.: Effect of magnesium coordination on the 13Cand15N magnetic resonance spectra of chlorophyll a. The Relative energies of nitrogen nπ* states as deduced from a complete assignment of chemical shifts. J. Am. Chem. Soc. 1974, 96, 7058–7066. URL http://dx.doi.org/10.1021/ja00829a038

[221] Huang, K.-F.; Hsu, C.-J.: Constituents of stem bark of Erythrina ar- borescens. J. Chin. Med. (Taiwan) 2001, 12, 61 – 67. (accessed Oct. 2005). URL http://www.nricm.edu.tw/jcm/012/12-1-06.pdf

[222] Altman, L. J.; Kowerski, R. C.; Laungani, D. R.: Studies in terpene biosynthesis. Synthesis and resolution of presqualene and prephytoene alcohols. J. Am. Chem. Soc. 1978, 100, 6174–6182. URL http://dx.doi.org/10.1021/ja00487a037

[223] Rogers, D. H.; Yi, E. C.; Poulter, C. D.: Enantioselective Synthesis of (+)-Presqualene Diphosphate. J. Org. Chem. 1995, 60, 941–945. URL http://dx.doi.org/10.1021/jo00109a026

[224] Giner, J.-L.; Berkowitz, J. D.; Andersson, T.: Nonpolar Components of the Latex of Euphorbia peplus. J. Nat. Prod. 2000, 63, 267–269. URL http://dx.doi.org/10.1021/np990081g

[225] Misra, R.; Pandey, R. C.; Dev, S.: The chemistry of the oleo resin from hardwickia pinnata: a series of new diterpenoids. Tetrahedron Lett. 1964, 5, 3751–3759. URL http://dx.doi.org/10.1016/S0040-4039(01)89373-4

[226] Misra, R.; Pandey, R. C.; Dev, S.: Absolute stereochemistry of hard- wickiic acid and its congeners. Tetrahedron Lett. 1968, 9, 2681–2684. URL http://dx.doi.org/10.1016/S0040-4039(00)89672-0

[227] Misra, R.; Pandey, R. C.; Dev, S.: Higher isoprenoids–VIII: Diterpenoids from the oleoresin of hardwickia pinnata part 1: hardwickiic acid. Tetra- hedron 1979, 35, 2301–2310. URL http://dx.doi.org/10.1016/0040-4020(79)80125-8

[228] Chaichantipyuth, C.; Muangsin, N.; Chaichit, N.; Roengsumran, S.; Petsom, A.; Watanabe, T.; Ishikawa, T.: Crystal structure of (-)- hardwickiic acid, C19H27OCOOH. Z. Kristallogr. - New Cryst. Struct. 2004, 219, 111–113. (accessed Oct 2005). URL http://www.oldenbourg.de/verlag/zkristallogr/mn-ncsc0402. htm#p111 BIBLIOGRAPHY 261

[229] Ortega, A.; Cárdenas, J.; Gage, D. A.; Maldonado, E.: Abietane and clerodane diterpenes from Salvia regla. Phytochemistry 1995, 39, 931– 933. URL http://dx.doi.org/10.1016/0031-9422(95)00083-J

[230] Bandara, B. M. R.; Wimalasiri, W. R.; Bandara, K. A. N. P.: Isolation and insecticidal activity of (-)-hardwickiic acid from Croton aromaticus. Planta Med. 1987, 53,575.

[231] Cocker, W.; Moore, A. L.; Pratt, A. C.: Dextrorotatory hardwickiic acid, an extractive of Copaifera officinalis. Tetrahedron Lett. 1965, 6, 1983–1985. URL http://dx.doi.org/10.1016/S0040-4039(01)83897-1

[232] Cascon, V.; Gilbert, B.: Characterization of the chemical composition of oleoresins of Copaifera guianensis Desf., Copaifera duckei Dwyer and Copaifera multijuga Hayne. Phytochemistry 2000, 55, 773–778. URL http://dx.doi.org/10.1016/S0031-9422(00)00284-3

[233] Komiya, T.; Kyohkon, M.; Ohwaki, S.; Eto, J.; Katsuzaki, H.; Imai, K.; Kataoka, T.; Yoshioka, K.; Ishii, Y.; Hibasami, H.: Phytol induces programmed cell death in human lymphoid leukemia Molt 4B cells. Int. J. Mol. Med. 1999, 4, 377–380.

[234] Tomita, Y.: Immunological role of vitamin A and its related substances in prevention of cancer. Nutr. Cancer 1983, 5, 187–194.

[235] Köhler, I.; Jenett-Siems, K.; Kraft, C.; Siems, K.; Abbiw, D.; Bien- zle, U.; Eich, E.: Herbal remedies traditionally used against malaria. Part 7. Herbal remedies traditionally used against malaria in Ghana: Bioassay-guided fractionation of Microglossa pyrifolia (Asteraceae). Z. Naturforsch.,C:J.Biosci.2002, 57, 1022–1027. URL http://www.znaturforsch.com/sc/57c/s57c1022.pdf

[236] Rajab, M. S.; Cantrell, C. L.; Franzblau, S. G.; Fisher, N. H.: An- timycobacterial activity of (E)-phytol and derivatives. A preliminary structure-activity study. Planta Med. 1998, 64,2–4.

[237] Saludes, J. P.; Garson, M. J.; Franzblau, S. G.; Aguinaldo, A. M.: An- titubercular constituents from the hexane fraction of Morinda citrifolia Linn. (Rubiaceae). Phytother. Res. 2002, 16, 683–685. URL http://dx.doi.org/10.1002/ptr.1003

[238] Arnhold, T.; Elmazar, M. M. A.; Nau, H.: Prevention of Vitamin A Ter- atogenesis by Phytol or Phytanic Acid Results from Reduced Metabolism of Retinol to the Teratogenic Metabolite, All-trans-retinoic Acid. Toxi- col. Sci. 2002, 66, 274–282. URL http://dx.doi.org/10.1093/toxsci/66.2.274 262 BIBLIOGRAPHY

[239] Topçu, G.: Bioactive Triterpenoids from Salvia Species. J. Nat. Prod. 2006, 69, 482–487. URL http://dx.doi.org/10.1021/np0600402

[240] Ovesná, Z.; Vachálková, A.; Horváthová, K.; Táthová, D.: Pentacyclic triterpenoic acids: new chemoprotective compounds. Neoplasma 2004, 51, 327–333. URL http://www.neoplasma.sk/abstrakt.asp?idn=645

[241] Liu, J.: Pharmacology of oleanolic acid and ursolic acid. J. Ethnophar- macol. 1995, 49, 57–68. URL http://dx.doi.org/10.1016/0378-8741(95)01310-5

[242] Ma, C.-m.; Nakamura, N.; Hattori, M.; Kakuda, H.; Qiao, J.-c.; Yu, H.- l.: Inhibitory effects on HIV-1 protease of constituents from the wood of Xanthoceras sorbifolia. J. Nat. Prod. 2000, 63, 238–242. URL http://dx.doi.org/10.1021/np9902441

[243] Watanabe, M.; Kobayashi, Y.; Ogihara, J.; Kato, J.; Oishi, K.: HIV-1 reverse transcriptase-inhibitory compound in Salvia officinalis. Food Sci. Technol. Res. 2000, 6, 216–220.

[244] Ikeda, T.; Yokomizo, K.; Okawa, M.; Tsuchihashi, R.; Kinjo, J.; Nohara, T.; Uyeda, M.: Anti-herpes virus type 1 activity of oleanane-type triter- penoids. Biol. Pharm. Bull. 2005, 28, 1779–1781. URL http://dx.doi.org/10.1248/bpb.28.1779

[245] Woldemichael, G. M.; Singh, M. P.; Maiese, W. M.; Timmermann, B. N.: Constituents of antibacterial extract of Caesalpinia paraguariensis Burk. Z.Naturforsch.,C:J.Biosci.2003, 58, 70–75. URL http://www.znaturforsch.com/s58c/s58c0070.pdf

[246] Weimann, C.; Goransson, U.; Pongprayoon-Claeson, U.; Claeson, P.; Bohlin, L.; Rimpler, H.; Heinrich, M.: Spasmolytic effects of Baccharis conferta and some of its constituents. J. Pharm. Pharmacol. 2002, 54, 99–104. URL http://www.aapspharmaceutica.com/search/view.asp?ID=5187

[247] Wachter, G. A.; Valcic, S.; Flagg, M. L.; Franzblau, S. G.; Montenegro, G.; Suarez, E.; Timmermann, B. N.: Antitubercular activity of penta- cyclic triterpenoids from plants of Argentina and Chile. Phytomedicine 1999, 6, 341–345.

[248] Tan, N.; Kaloga, M.; Radtke, O. A.; Kiderlen, A. F.; Öksüz, S.; Ulubelen, A.; Kolodziej, H.: Abietane diterpenoids and triterpenoic acids from Salvia cilicica and their antileishmanial activities. Phytochemistry 2002, 61, 881–884. URL http://dx.doi.org/10.1016/S0031-9422(02)00361-8 BIBLIOGRAPHY 263

[249] Mallavadhani, U. V.; Mahapatra, A.; Raja, S. S.; Manjula, C.: An- tifeedant activity of some pentacyclic triterpene acids and their fatty acid ester analogs. J. Agric. Food Chem. 2003, 51, 1952–1955. URL http://dx.doi.org/10.1021/jf020691d [250] Blagg, B. S. J.; Jarstfer, M. B.; Rogers, D. H.; Poulter, C. D.: Re- combinant Squalene Synthase. A Mechanism for the Rearrangement of Presqualene Diphosphate to Squalene. J. Am. Chem. Soc. 2002, 124, 8846–8853. URL http://dx.doi.org/10.1021/ja020411a [251] Faure, S.; Connolly, J. D.; Fakunle, C. O.; Piva, O.: Structure and Syn- thesis of Anhydrobisfarnesol from Euphorbia lateriflora and Asymmetric Synthesis of (R)-Sesquilavandulol. Tetrahedron 2000, 56, 9647–9653. URL http://dx.doi.org/10.1016/S0040-4020(00)00913-3 [252] Connolly, J. D.: Natural products from around the world. Pure Appl. Chem. 2001, 73, 567–571. URL http://www.iupac.org/publications/pac/2001/pdf/7303x0567.pdf [253] Connolly, J. D.: Natural products from around the world. Rev. Lati- noamer. Quím. 1997, 25, 77–85. [254] Van Tamelen, E. E.; Leopold, E. J.: Mechanism of presqualene pyrophosphate-squalene biosynthesis. II. Synthesis of bifarnesol. Tetra- hedron Lett. 1985, 26, 3303–3306. URL http://dx.doi.org/10.1016/S0040-4039(00)98283-2 [255] Valdés, III, L. J. J.: Divinorin C, a new neoclerodane diterpene from a bioactive TLC fraction of Salvia divinorum. Entheogen Rev. 2000, 9, 141. Reprinted in Ref. 89. URL http://www.sagewisdom.org/divinorinc.html [256] Medana, C. Emails, 10th February, 4th and 5th March 2005. [257] PC Model v5.04 for Mac; Serena Software: Bloomington, IN, 1994. URL http://www.serenasoft.com [258] Grof, S.: Stanislav Grof interviews Dr. Albert Hofmann. MAPS Bulletin 2001, 11, 22–35. See p. 29. URL http://www.maps.org/news-letters/v11n2/11222gro.pdf [259] Lee-Stecum, P. Email, 7th May 2003. [260] Brown, L.: The Stereocontrolled Synthesis of Optically Active Vitamin E Side Chains. II. Benzoyl Triflate and its Application in the Determina- tion of the Absolute Configuration of Divinorin A and B, and Terrecyclic acid. Ph.D. thesis, University of Michigan, Ann Arbor, MI 1984.Pro- Quest Publication Number: AAT 8422201 (Document ID: 748992941). See pp. 78 – 80, 196 – 200 (hydride reductions). URL http://wwwlib.umi.com/dissertations/fullcit/8422201 264 BIBLIOGRAPHY

[261] Ref 260, pp. 84, 85 (compound 123).

[262] Ref 260, p. 76 (direct diacetylation).

− [263]Ref260,pp.72–75(BH4 reductions, epimerizations and proposed mech- anism).

[264] Goddard, R.; Akhtar, F.: The structure of jateorin. Acta Crystallogr., Sect.C:Cryst.Struct.Commun.1986, C42, 1217–1220. URL http://dx.doi.org/10.1107/S0108270186092831

[265] Swaminathan, K.; Sinha, U. C.; Ramakumar, S.: Structure of columbin, a diterpenoid furanolactone from Tinospora cordifolia Miers. Acta Crys- tallogr., Sect. C: Cryst. Struct. Commun. 1989, C45, 300–303. URL http://dx.doi.org/10.1107/S0108270188010583

[266] Yonemitsu, M.; Fukuda, N.; Kimura, T.; Komori, T.; Lindner, H. J.; Habermehl, G.: Crystal structure and NMR spectrometric analysis of palmarin. Liebigs Ann. Chem. 1989, pp. 485–487.

[267] Cheung, K. K.; Melville, D.; Overton, K. H.; Robertson, J. M.; Sim, G. A.: The stereochemistry of isocolumbin: x-ray analysis of the 1-p- iodophenyl-3-phenylpyrazoline adduct of isocolumbin. J. Chem. Soc., B 1966, pp. 853–861. URL http://dx.doi.org/10.1039/J29660000853

[268] Overton, K. H.; Weir, N. G.; Wylie, A.: The stereochemistry of the colombo root bitter principles. J. Chem. Soc., C 1966, pp. 1482–1490. See isocolumbin, dihydrocolumbin and dihydroisocolumbin (hydrogena- tions); table 2 (furanolactone infrared absorptions). URL http://dx.doi.org/10.1039/J39660001482

[269] Yonemitsu, M.; Fukuda, N.; Kimura, T.; Komori, T.: Studies on the constituents of Jateorhiza palmata Miers (colombo root). I. Separation and structure of a new furanoid diterpene glucoside (palmatoside A). Liebigs Ann. Chem. 1986, pp. 1327–1333. See compounds 3, 4 (nujol mulls) and 6 (mull vs CHCl3).

[270] Munro, T. A.; Rizzacasa, M. A.; Roth, B. L.; Toth, B. A.; Yan, F.: Studies toward the Pharmacophore of Salvinorin A, a Potent κ-Opioid Receptor Agonist. J. Med. Chem. 2005, 48, 345–348. URL http://dx.doi.org/10.1021/jm049438q

[271] Lee, D. Y. W.; Karnati, V. V. R.; He, M.; Liu-Chen, L.-Y.; Kondareti, L.; Ma, Z.; Wang, Y.; Chen, Y.; Béguin, C.; Carlezon, W. A.; Cohen, B.: Synthesis and in vitro pharmacological studies of new C(2) modified salvinorin A analogues. Bioorg.Med.Chem.Lett.2005, 15, 3744–3747. Epimerization is discussed on p. 3746 c.1. URL http://dx.doi.org/10.1016/j.bmcl.2005.05.048 BIBLIOGRAPHY 265

[272] Ref 260, p. 83 (deacetylations with KCN); cf. p. 202 (yield).

[273] Valdés, III, L. J. J. Email, 28th Apr. 2003.

[274] Khong, P. W.; Lewis, K. G.: Acetyl migration in urs-12-ene-3α,24-diol. Aust. J. Chem. 1975, 28, 201–206. See also references 1 - 6 therein. URL http://dx.doi.org/10.1071/CH9750201

[275] Smith, M. B.; March, J.: March’s advanced organic chemistry: reactions, mechanisms, and structure. Wiley, New York, 5th ed. 2001.SeeTable 8.1, pp. 330 - 331 (pKa of conjugate acids), Section 12-3 pp. 773 - 775 (keto/enol tautomerism).

[276] Barton, D. H. R.; Overton, K. H.; Wylie, A.: Diterpenoid bitter princi- ples. IV. Investigations on the constitution of palmarin. J. Chem. Soc. 1962, pp. 4809–4815. See hydrogenations of palmarin and isojateorin. URL http://dx.doi.org/10.1039/JR9620004809

[277] Balasubramanian, S. K.; Barton, D. H. R.; Jackman, L. M.: Diterpenoid bitter principles. V. Constitution of palmarin and its congeners. J. Chem. Soc. 1962, pp. 4816–4820. URL http://dx.doi.org/10.1039/JR9620004816

[278] Ref 260, pp. 90 – 92 (deuteration). The NMR data given are consistent with this structure. The 13Cpeakatδ 171.6 is incorrectly assigned to the lactone, however, on the basis of an apparent reduction in intensity rel- ative to 1a. The apparent reduction was probably due to an inadequate signal-to-noise ratio; five peaks were not observed. The claimed yield of 89% is implausible given the epimerization we and Valdés observed with KCN.

[279] Strazzolini, P.; Giumanini, A. G.; Cauci, S.: Acetic formic anhydride: a review. Tetrahedron 1990, 46, 1081–1118. URL http://dx.doi.org/10.1016/S0040-4020(01)86676-X

[280] Reber, F.; Lardon, A.; Reichstein, T.: Teilsynthese von 11-Epi- corticosteron. Bestandteile der Nebennierenrinde und verwandte Stoffe, 86. Mitteilung. Helv. Chim. Acta 1954, 37, 45–58. See compounds XII and XVIII. URL http://dx.doi.org/10.1002/hlca.19540370107

[281] Oliveto, E. P.; Gerold, C.; Rausser, R.; Hershberg, E. B.: 11-Oxygenated Steroids. XII. The Preparation of 17α-Hydroxycorticosterone 21-Acetate (Kendall’s Compound F Acetate) via 11β-Formates. J. Am. Chem. Soc. 1955, 77, 3564 – 3567. See compound V. URL http://dx.doi.org/10.1021/ja01618a042

[282] Ref 260, pp. 84 – 87 (benzoylation). 266 BIBLIOGRAPHY

[283] Danklmaier, J.; Hoenig, H.: Regioselektive benzylierungen an Hydroxyalkyl-3-morpholinonen. Liebigs Ann. Chem. 1989, pp. 665–669.

[284] Eckenberg, P.; Groth, U.; Huhn, T.; Richter, N.; Schmeck, C.: A useful application of benzyl trichloroacetimidate for the benzylation of alcohols. Tetrahedron 1993, 49, 1619–1624. URL http://dx.doi.org/10.1016/S0040-4020(01)80349-5

[285] Tanabe, M.; Peters, R. H.: (R,S)-Mevalonolactone-2-13C. Org. Synth. 1990, 60, 92. Also Coll. Vol. 7, p. 386. See note 4. URL http://www.orgsyn.org/orgsyn/pdfs/CV7P0386.pdf

[286] Bouzide, A.; Sauve, G.: Highly selective silver (I) oxide mediated mono- protection of symmetrical diols. Tetrahedron Lett. 1997, 38, 5945–5948. URL http://dx.doi.org/10.1016/S0040-4039(97)01328-2

[287] Charette, A. B.: Tetra-n-butylammonium iodide. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wiley,NewYork 1995. URL http://dx.doi.org/10.1002/047084289X.rt018

[288] Kraus, G. A.; Frazier, K. A.; Roth, B. D.; Taschner, M. J.; Neuen- schwander, K.: Conversion of lactones into ethers. J. Org. Chem. 1981, 46, 2417–2419. URL http://dx.doi.org/10.1021/jo00324a050

[289] Lundt, I.: Oxidation, reduction and deoxygenation. In Glycoscience: chemistry and chemical biology,B.O.Fraser-Reid;K.Tatsuta;J.Thiem, eds., Springer, New York, vol. 1, pp. 501–531 2001.

[290] Hansen, M. C.; Verdaguer, X.; Buchwald, S. L.: Convenient Two-Step Conversion of Lactones into Cyclic Ethers. J. Org. Chem. 1998, 63, 2360–2361. URL http://dx.doi.org/10.1021/jo9716082

[291] Kawamura, K.; Hinou, H.; Matsuo, G.; Nakata, T.: Efficient strategy for convergent synthesis of trans-fused polycyclic ethers based on an in- tramolecular SmI2-promoted cyclization of iodo ester. Tetrahedron Lett. 2003, 44, 5259–5261. See compound 11. URL http://dx.doi.org/10.1016/S0040-4039(03)01277-2

[292] Miki, Y.; Ohta, M.; Hachiken, H.; Takemura, S.: A simple synthesis of naphtho[1,8-bc]pyran. Synthesis 1990,p.312. URL http://dx.doi.org/10.1055/s-1990-26860

[293] Kourafalos, V. N.; Marakos, P.; Pouli, N.; Townsend, L. B.: The Syn- thesis of 4-Deazaformycin A. J. Org. Chem. 2003, 68, 6466–6469. See compound 7. URL http://dx.doi.org/10.1021/jo026715x BIBLIOGRAPHY 267

[294] Barbier, M.; Devys, M.; Parisot, D.: Synthesis of anhydronectri- achrysone, an extensively conjugated γ-pyrone. Synth. Commun. 1993, 23, 1481–1488. [295] Ikawa, T.; Sajiki, H.; Hirota, K.: Unexpected deprotection of silyl and THP ethers induced by serious disparity in the quality of Pd/C catalysts and elucidation of the mechanism. Tetrahedron 2004, 60, 6189–6195. URL http://dx.doi.org/10.1016/j.tet.2004.05.040 [296] Sajiki, H.; Hattori, K.; Hirota, K.: Highly chemoselective hydrogenation with retention of the epoxide function using a heterogeneous Pd/C- ethylenediamine catalyst and THF. Chem. Eur. J. 2000, 6, 2200–2204. See also references therein. URL http://dx.doi.org/10.1002/1521-3765(20000616)6:12<2200:: AID-CHEM2200>3.0.CO;2-3 [297] Rylander, P. N.: Hydrogenation methods. Academic Press, London 1985. See pp. 10 (use of base), 157+ (catalyst effects). [298] Bruno, M.; Rosselli, S.; Pibiri, I.; Piozzi, F.; Simmonds, M.: Hydro- genation derivatives of neo-clerodanes. Heterocycles 2000, 55, 599–612. See hydrogenations of montanin C (4) and teucrin A (6). Site accessed February 2006. URL https://www2.heterocycles.jp/FMPro?-db=journalred.fp5& -format=/w2/charge.html&-lay=data&-recid=35852&-Token. 3=COM-99-8777&-Script=citingAndDuplicate&-Find [299] Housley, J. R.; King, F. E.; King, T. J.; Taylor, P. R.: Chemistry of hardwood extractives. XXXIV. Constituents of Guarea species. J. Chem. Soc. 1962, pp. 5095–5104. See dihydrogedunene hydrogenations. URL http://dx.doi.org/10.1039/JR9620005095 [300] Malakov, P. Y.; Papanov, G. Y.; Mollov, N. M.; Spassov, S. L.: Montanin-C, a New Furanoid Diterpene from Teucrium montanum L. Z. Naturforsch., B: Anorg. Chem., Org. Chem. 1978, 33B, 789–791. [301] Brieskorn, C. H.; Pfeuffer, T.: Labiatenbitterstoffe: Pikropolin und ähn- liche Diterpenoide aus Poleigamander. Chem. Ber. 1967, 100, 1998– 2010. See picropolin acetate. URL http://dx.doi.org/10.1002/cber.19671000629 [302] Adesogan, E. K.: The structure of penduliflaworosin, a new furanoid diterpene from Croton penduliflorus. J. Chem. Soc., Perkin Trans. 1 1981, pp. 1151–1153. URL http://dx.doi.org/10.1039/P19810001151 [303] Bui, A. M.; Parello, J.; Potier, P.; Janot, M. M.: Sur la structure du collybolide, nouvelle substance sesquiterpénique extraite du Colly- bia maculata Alb. et Sch. ex Fries (Basidiomycètes). C. R. Acad. Sci., C. 1970, 270, 1022–1026. See 2 and 2’. 268 BIBLIOGRAPHY

[304] Chan, W. R.; Taylor, D. R.; Willis, C. R.: Terpenoids from the Eu- phorbiaceae. I. The structure of crotonin, a norditerpene from Croton lucidus. J. Chem. Soc., C 1968, pp. 2781–2785. URL http://dx.doi.org/10.1039/J39680002781 [305] Chatterjee, A.; Kundu, A. B.; Chakrabortty, T.; Chandrasekharan, S.: Extractives of Aphanamixis polystachya wall (Parker). Structures and stereochemistry of aphanamixin and aphanamixinin. Tetrahedron 1970, 26, 1859–1867. See hydrogenations of XXIX. URL http://dx.doi.org/10.1016/S0040-4020(01)92762-0 [306] Emerson, O. H.: Bitter principles in citrus. III. Some reactions of limonin. J. Am. Chem. Soc. 1952, 74, 688–693. URL http://dx.doi.org/10.1021/ja01123a031 [307] Ruberto, G.; Renda, A.; Tringali, C.; Napoli, E. M.; Simmonds, M. S. J.: Citrus limonoids and their semisynthetic derivatives as antifeedant agents against Spodoptera frugiperda larvae. A structure-activity rela- tionship study. J. Agric. Food Chem. 2002, 50, 6766–6774. URL http://dx.doi.org/10.1021/jf020607u [308] De Alvarenga, M. A.; Gottlieb, H. E.; Gottlieb, O. R.; Magalhães, M. T.; Da Silva, V. O.: Diasin, a diterpene from Croton diasii. Phytochemistry 1978, 17, 1773–1776. URL http://dx.doi.org/10.1016/S0031-9422(00)88692-6 [309] Rodriguez-Hahn, L.; O’Reilly, R.; Esquivel, B.; Maldonado, E.; Or- tega, A.; Cardenas, J.; Toscano, R. A.; Chan, T.-M.: Tilifodiolide, tetraline-type diterpenoid of clerodanic origin from Salvia tiliaefolia. J. Org. Chem. 1990, 55, 3522–3525. URL http://dx.doi.org/10.1021/jo00298a026 [310] Hori, T.; Kiang, A. K.; Nakanishi, K.; Sasaki, S.; Woods, M. C.: Struc- tures of fibraurin and a minor product from Fibraurea chloroleuca. Tetra- hedron 1967, 23, 2649–2656. URL http://dx.doi.org/10.1016/0040-4020(67)85129-9 [311] Melera, A.; Schaffner, K.; Arigoni, D.; Jeger, O.: Zur Konstitution des Limonins I. Über den Verlauf der alkalischen Hydrolyse von Limonin und Limonol. Helv. Chim. Acta 1957, 40, 1420–1437. See hydrogenations of I and II, pp. 1431 - 1432. URL http://dx.doi.org/10.1002/hlca.19570400529 [312] Rosenfeld, R. S.; Hofmann, K.: The chemistry of limonin. J. Am. Chem. Soc. 1951, 73, 2491–2493. URL http://dx.doi.org/10.1021/ja01150a024 [313] Ida, Y.; Kubo, S.; Fujita, M.; Komori, T.; Kawasaki, T.: Struktur der Diosbulbine-D, -E, -F, -G und H. Justus Liebigs Ann. Chem. 1978, pp. 818–833. See hydrogenation of compounds D and F. BIBLIOGRAPHY 269

[314] Popa, D. P.; Reinbol’d, A. M.: Structure of teucrin A. Chem. Nat. Comp. 1975, 9, 27–30. Translation of Khim. Prir. Soedin., 1973, 9, 31-35. URL http://dx.doi.org/10.1007/BF00580883

[315] Pailer, M.; Schaden, G.; Spiteller, G.; Frenzl, W.: Die Konstitution des Fraxinellons aus Dictamnus albus L. Monatsh. Chem. 1965, 96, 1324– 1346. URL http://dx.doi.org/10.1007/BF00904284

[316] Arndt, R. R.; Baarschers, W. H.: Structure of phragmalin. Meliacin with a norbornane part skeleton. Tetrahedron 1972, 28, 2333–2340. URL http://dx.doi.org/10.1016/S0040-4020(01)93576-8

[317] Kouzi, S. A.; McMurtry, R. J.; Nelson, S. D.: Hepatotoxicity of German- der (Teucrium chamaedrys L.) and One of Its Constituent Neoclerodane Diterpenes Teucrin A in the Mouse. Chem. Res. Toxicol. 1994, 7, 850– 856. URL http://dx.doi.org/10.1021/tx00042a020

[318] MacKinnon, S.; Durst, T.; Arnason, J. T.; Angerhofer, C.; Pezzuto, J.; Sanchez-Vindas, P. E.; Poveda, L. J.; Gbeassor, M.: Antimalarial Activity of Tropical Meliaceae Extracts and Gedunin Derivatives. J. Nat. Prod. 1997, 60, 336–341. See compound 10. URL http://dx.doi.org/10.1021/np9605394

[319] Siegel, S.: Rhodium on Alumina. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wiley,NewYork1995. URL http://dx.doi.org/10.1002/047084289X.rr003

[320] Kabalka, G. W.; Wang, L.; Pagni, R. M.: Potassium fluoride doped alumina: an effective reagent for ester hydrolysis under solvent free con- ditions. Green Chem. 2001, 3, 261–262. URL http://dx.doi.org/10.1039/b106423c

+ [321] Ref 260, p. 81 (Li/NH3 reduction); cf. p. 201 (structure). The [M+H] value given in the mass spectral data (265) is evidently a typographical error, being smaller than several daughter ions. The correct value is 365.

[322] Presser, A.; Hüfner, A.: Trimethylsilyldiazomethane - A Mild and Effi- cient Reagent for the Methylation of Carboxylic Acids and Alcohols in Natural Products. Monatsh. Chem. 2004, 135, 1015–1022. URL http://dx.doi.org/10.1007/s00706-004-0188-4

[323] Ref 60, p. 153 (cyclic α-diones).

[324] Raymond, S.: The Structure of Methyl Acetylacrylate. J. Am. Chem. Soc. 1950, 72, 4304–4306. URL http://dx.doi.org/10.1021/ja01165a537 270 BIBLIOGRAPHY

[325] Caton, M. P. L.; Darnbrough, G.; Parker, T.: Prostaglandins VI: Base- Catalyzed Autoxidation of α-Hydroxycyclopentanones and the Synthesis of 9,10-Diketoprostanoic Acids. Tetrahedron Lett. 1980, 21, 1685–1686. URL http://dx.doi.org/10.1016/S0040-4039(00)77786-0

[326] Sundt, E.; Jeger, O.; Prelog, V.: Conversion of Cevagenin into a Diosphe- nol. Chem. Ind. 1953, pp. 1365–1366.

[327] Lavie, D.; Shlomo, S.: The Constituents of Ecballium elaterium L. II. α-Elaterin. J. Am. Chem. Soc. 1958, 80, 707–710. URL http://dx.doi.org/10.1021/ja01536a046

[328] Lack, R. E.; Ridley, A. B.: Autoxidation of 2α-Hydroxy-5α-Cholestan- 3-One in Methanolic Potassium Hydroxide. J. Chem. Soc., C 1968, pp. 3017–3020. URL http://dx.doi.org/10.1039/J39680003017

[329] Weissberger, A.; Schwarze, W.: Über die Autoxydation des 1-Oxy-α- tetralons. Justus Liebigs Ann. Chem. 1931, 487, 53–61.

[330] Slavikova, B.; Kasal, A.; Budesinsky, M.: Autoxidation vs hydrolysis in 16α-acyloxy steroids. Collect. Czech. Chem. Commun. 1999, 64, 1125– 1134. URL http://dx.doi.org/10.1135/cccc19991125

[331] Dauben, W. G.; Boswell, G. A.; Templeton, W.: Base-catalyzed autox- idation of cyclic 1,4-diketones to enediones. J. Org. Chem. 1960, 25, 1853–1855. URL http://dx.doi.org/10.1021/jo01081a006

[332] Ref 260, p. 82 (PCC oxidation).

[333] González, A. G.; Aguiar, Z. E.; Luis, J. G.; Ravelo, A. G.: New diter- penes from Salvia texana. Chemical and biogenetic aspects. Tetrahedron 1989, 45, 5203–5214. See compounds 13a, b. URL http://dx.doi.org/10.1016/S0040-4020(01)81097-8

[334] Ganguly, A. K.; Govindachari, T. R.; Mohamed, P. A.: Oxidation of ring A in lupeol. Tetrahedron 1966, 22, 3597–3599. URL http://dx.doi.org/10.1016/S0040-4020(01)92548-7

[335] Marchand, A. P.; Reddy, G. M.: Ultrasound-promoted sodium boro- hydride reduction of pentacyclo[5.4.0.02,6.03,10.05,9]undecane-8,11-dione (PCUD-8,11-dione) and of 4,4-dimethoxy-2,3,5,6-tetrachloro-PCUD- 8,11-dione. Org. Prep. Proced. Int. 1990, 22, 528–531. 59 was sonicated at 35 kHz, 320 W output.

[336] Gemal, A. L.; Luche, J. L.: Lanthanoids in organic synthesis. 6. Reduc- tion of α-enones by sodium borohydride in the presence of lanthanoid chlorides: synthetic and mechanistic aspects. J. Am. Chem. Soc. 1981, BIBLIOGRAPHY 271

103, 5454–5459. On decomposition, see p. 5456 c. 2, esp. fig. 3. URL http://dx.doi.org/10.1021/ja00408a029

[337] Marchand, A. P.; LaRoe, W. D.; Sharma, G. V. M.; Suri, S. C.; Reddy, D. S.: Facile stereoselective reductions of enediones and cage diketones using NaBH4-CeCl3. J. Org. Chem. 1986, 51, 1622–1625. See attempted reduction of 12 and 16. URL http://dx.doi.org/10.1021/jo00359a054

[338] Dimitrov, V.; Kostova, K.; Genov, M.: Anhydrous cerium(III) chloride - effect of the drying process on activity and efficiency. Tetrahedron Lett. 1996, 37, 6787–6790. URL http://dx.doi.org/10.1016/S0040-4039(96)01479-7

[339] Zsakó, J.; László, T.: Complex compounds of some diosphenolic type carotenoids. Astacene complexes. Spectrophotometric study of the sys- tem iron(III)-astacene. Rev. Roum. Chim. 1981, 26, 237–243.

[340] Rubin, Y.; Knobler, C. B.; Diederich, F.: Precursors to the Cy- clo[n]carbons: From 3,4-Dialkynyl-3-cyclobutene-1,2-diones and 3,4- Dialkynyl-3-cyclobutene-1,2-diols to Cyclobutenodehydroannulenes and Higher Oxides of Carbon. J. Am. Chem. Soc. 1990, 112, 1607–1617. See Scheme V. URL http://dx.doi.org/10.1021/ja00160a047

[341] McMurry, J. E.: Ester cleavages via SN2-type dealkylation. Org. React. 1976, 24, 187–224. URL http://dx.doi.org/10.1002/0471264180.or024.02

[342] Salomon, C. J.; Mata, E. G.; Mascaretti, O. A.: Recent developments in chemical deprotection of ester functional groups. Tetrahedron 1993, 49, 3691–3734. URL http://dx.doi.org/10.1016/S0040-4020(01)90225-X

[343] Müller, P.; Siegfried, B.: SN2 reactions with carboxylic esters. Selective cleavage of methyl esters. Helv. Chim. Acta 1974, 57, 987–994. URL http://dx.doi.org/10.1002/hlca.19740570404

[344] Elsinger, F.; Schreiber, J.; Eschenmoser, A.: Notiz über die Selektiv- ität der Spaltung von Carbonsäuremethylestern mit Lithiumjodid. Helv. Chim. Acta 1960, 43, 113–118. URL http://dx.doi.org/10.1002/hlca.19600430116

[345] Dean, P. D. G.: Halogenolysis of methyl glycyrrhetate with lithium iodide-dimethylformamide. J. Chem. Soc. 1965, p. 6655. URL http://dx.doi.org/10.1039/JR9650006655

[346] Müller, P.; Siegfried, B.: Decarboxylation of β-keto esters in hexam- ethylphosphoric triamide. Tetrahedron Lett. 1973, pp. 3565–3568. URL http://dx.doi.org/10.1016/S0040-4039(01)86971-9 272 BIBLIOGRAPHY

[347] Dykstra, R. R.: Hexamethylphosphoric Triamide. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wiley,NewYork 1995. URL http://dx.doi.org/10.1002/047084289X.rh020

[348] Elsinger, F.: 2-Benzylcyclopentanone. Org. Synth. 1965, 45,7.Also Coll. Vol. 5, p. 76. URL http://www.orgsyn.org/orgsyn/pdfs/CV5P0076.pdf

[349] McMurry, J. E.; Wong, G. B.: Improved method for the cleavage of methyl esters. Synth. Commun. 1972, 2, 389–394.

[350] Beck, A. K.; Seebach, D.: N,N -Dimethylpropyleneurea. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wiley,NewYork 1995. URL http://dx.doi.org/10.1002/047084289X.rd366

[351] Sharma, L.; Nayak, M. K.; Chakraborti, A. K.: A mild and chemoselec- tive method for ester O-alkyl cleavage using in situ generated potassium thiophenoxide from catalytic quantities of base. Tetrahedron 1999, 55, 9595–9600. URL http://dx.doi.org/10.1016/S0040-4020(99)00505-0

[352] Misner, J. W.; Kennedy, J. H.; Biggs, W. S.: Integration of a highly selective demethylation of a quaternized ergoline into a one-pot synthesis of pergolide. Org. Process Res. Dev. 1997, 1, 77–80. URL http://dx.doi.org/10.1021/op9600015

[353] Bartlett, P. A.; Johnson, W. S.: Improved reagent for the O-alkyl cleav- age of methyl esters by nucleophilic displacement. Tetrahedron Lett. 1970, 11, 4459–4462. URL http://dx.doi.org/10.1016/S0040-4039(01)83950-2

[354] Kelly, T. R.; Dali, H. M.; Tsang, W. G.: Lithium thiomethoxide: a convenient mercaptide reagent. Tetrahedron Lett. 1977, 18, 3859–3860. URL http://dx.doi.org/10.1016/S0040-4039(01)83373-6

[355] Feutrill, G. I.; Mirrington, R. N.: Reactions with thioethoxide ion in dimethylformamide. II. O-Alkyl cleavage of methylene ethers, methyl esters, and aryloxyacetates. Aust. J. Chem. 1972, 25, 1731–1735. URL http://dx.doi.org/10.1071/CH9721731

[356] Vaughan, W. R.; Baumann, J. B.: Reactions of Alkyl Carboxylic Esters with Mercaptides. J. Org. Chem. 1962, 27, 739–744. URL http://dx.doi.org/10.1021/jo01050a010

[357] Ireland, R. E.; Thompson, W. J.: An approach to the total synthesis of chlorothricolide: the synthesis of the top half. J. Org. Chem. 1979, 44, 3041–3052. See p. 3048 compound 13 (preparation of LiSPr). URL http://dx.doi.org/10.1021/jo01331a017 BIBLIOGRAPHY 273

[358] Nicolaou, K. C.; Estrada, A. A.; Zak, M.; Lee, S. H.; Safina, B. S.: A Mild and Selective Method for the Hydrolysis of Esters with Trimethyltin Hydroxide. Angew. Chem. Int. Ed. 2005, 44, 1378–1382. URL http://dx.doi.org/10.1002/anie.200462207

[359] Ishida, N.; Akaike, M.; Tsutsumi, S.; Kanai, H.; Masui, A.; Sadamatsu, M.; Kuroda, Y.; Watanabe, Y.; McEwen, B. S.; Kato, N.: Trimethyltin syndrome as a hippocampal degeneration model: temporal changes and neurochemical features of seizure susceptibility and learning impairment. Neuroscience 1997, 81, 1183–1191. URL http://dx.doi.org/10.1016/S0306-4522(97)00220-0

[360] Mascaretti, O. A.; Furlan, R. L. E.: Esterifications, transesterifica- tions, and deesterifications mediated by organotin oxides, hydroxides, and alkoxides. Aldrichimica Acta 1997, 30, 55–68.

[361] Lee, D. Y.; He, M.; Kondaveti, L.; Liu-Chen, L. Y.; Ma, Z.; Wang, Y.; Chen, Y.; Li, J. G.; Béguin, C.; Carlezon, W. A.; Cohen, B.: Synthesis and in vitro pharmacological studies of C(4) modified salvinorin A ana- logues. Bioorg. Med. Chem. Lett. 2005, 15, 4169–4173. URL http://dx.doi.org/10.1016/j.bmcl.2005.06.092

[362] Oikawa, M.; Wada, A.; Okazaki, F.; Kusumoto, S.: Acidic, Selective Monoacylation of vic-Diols. J. Org. Chem. 1996, 61, 4469–4471. See compound 38. URL http://dx.doi.org/10.1021/jo960081a

[363] Burchat, A. F.; Chong, J. M.; Nielsen, N.: Titration of alkyllithiums with a simple reagent to a blue endpoint. J. Organomet. Chem. 1997, 542, 281–283. URL http://dx.doi.org/10.1016/S0022-328X(97)00143-5

[364] Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy, S.; Stocky, T. P.: Selective reductions. XIX. Rapid reaction of carboxylic acids with borane-tetrahydrofuran. Remarkably convenient procedure for the selec- tive conversion of carboxylic acids to the corresponding alcohols in the presence of other functional groups. J. Org. Chem. 1973, 38, 2786–2792. URL http://dx.doi.org/10.1021/jo00956a011

[365] Lane, C. F.: Reduction of organic compounds with diborane. Chem. Rev. 1976, 76,773–799. URL http://dx.doi.org/10.1021/cr60304a005

[366] Corey, E. J.; Sachdev, H. S.: 2,4-Dinitrobenzenesulfonylhydrazine, a useful reagent for the Eschenmoser α,β cleavage of α,β-epoxy ketones. Conformational control of halolactonization. J. Org. Chem. 1975, 40, 579 – 581. See compound IV. URL http://dx.doi.org/10.1021/jo00893a008 274 BIBLIOGRAPHY

[367] McCorkindale, N. J.; Roy, T. P.; Hutchinson, S. A.: Isolation and syn- thesis of 3-chlorogentisyl alcohol–a metabolite of Penicillium canadense. Tetrahedron 1972, 28, 1107–1111. See compound 5. URL http://dx.doi.org/10.1016/0040-4020(72)80170-4

[368] Hagishita, S.; Kuriyama, K.: Optical activity of C2 symmetrical 9,10- dihydro-9,10-ethanoanthracenes. Tetrahedron 1972, 28, 1435–1467. See compound 5. BH3 was generated in situ as per reference 5 therein. Stoi- chiometry inferred from limiting reagent NaBH4. URL http://dx.doi.org/10.1016/0040-4020(72)88028-1

[369] Brown, H. C.; Heim, P.; Yoon, N. M.: Selective reductions. XV. Re- action of diborane in tetrahydrofuran with selected organic compounds containing representative functional groups. J. Am. Chem. Soc. 1970, 92, 1637 – 1646. URL http://dx.doi.org/10.1021/ja00709a037

[370] Doxsee, K. M.; Feigel, M.; Stewart, K. D.; Canary, J. W.; Knobler, C. B.; Cram, D. J.: Host-guest complexation. 42. Preorganization strongly en- hances the tendancy of hemispherands to form hemispheraplexes. J. Am. Chem. Soc. 1987, 109, 3098 – 3107. See compound 24. URL http://dx.doi.org/10.1021/ja00244a037

[371] Artz, S. P.; Cram, D. J.: Host-guest complexation. 28. Hemispherands with four self-organizing units. J. Am. Chem. Soc. 1984, pp. 2160–2171. See compounds 32 and 37. URL http://dx.doi.org/10.1021/ja00319a042

[372] Dias, J. R.; Pettit, G. R.: Reduction of δ-lactones and hindered esters with diborane. J. Org. Chem. 1971, 36, 3485–3489. See compounds 7 and 9. URL http://dx.doi.org/10.1021/jo00822a003

[373] Kende, A. S.; Fludzinski, P.: Ethyl 4-hydroxycrotonate. Org. Synth. 1986, 64, 104. Also Coll. Vol. 7, p. 221. URL http://www.orgsyn.org/orgsyn/pdfs/CV7P0221.pdf

[374] Brown, H. C.; Stocky, T. P.: Selective reductions. 24. Acyloxyboranes in the controlled reaction of carboxylic acids with borane-tetrahydrofuran. Acyloxyboranes as intermediates in the fast reduction of carboxylic acids by borane-tetrahydrofuran. J. Am. Chem. Soc. 1977, 99, 8218 – 8226. URL http://dx.doi.org/10.1021/ja00467a016

[375] Lee, K. C.: Methyltriphenylphosphonium Bromide. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wiley,NewYork 1995. URL http://dx.doi.org/10.1002/047084289X.rm273 BIBLIOGRAPHY 275

[376] Pine, S. H.; Shen, G. S.; Hoang, H.: Ketone methylenation using the Tebbe and Wittig reagents - a comparison. Synthesis 1991, pp. 165– 167. URL http://dx.doi.org/10.1055/s-1991-26406

[377] Pine, S. H.: Carbonyl methylenation and alkylidenation using titanium- based reagents. Org. React. 1993, 43, 1–91. URL http://dx.doi.org/10.1002/0471264180.or043.01

[378] Lombardo, L.: Methylenation of carbonyl compounds with zinc- dibromomethane-titanium tetrachloride. Applications to gibberellins. Tetrahedron Lett. 1982, 23, 4293–4296. Note: due to the thick con- sistency of this reagent, transfer was very difficult using 0.5 mm I.D. (21 gauge) needles, and a larger gauge would be advisable. URL http://dx.doi.org/10.1016/S0040-4039(00)88728-6

[379] Lombardo, L.: Methylenation of carbonyl compounds: (+)-3-methylene- cis-p-menthane. Org. Synth. 1987, 65, 81. Also Coll. Vol. 8, p. 386. URL http://www.orgsyn.org/orgsyn/pdfs/CV8P0386.pdf

[380] Vedejs, E.: Clemmensen reduction of ketones in anhydrous organic sol- vents. Org. React. 1975, 22, 401–422. See table p. 418 (eliminations). URL http://dx.doi.org/10.1002/0471264180.or022.03

[381] Yamamura, S.; Toda, M.; Hirata, Y.: Modified Clemmensen Reduc- tion: Cholestane. Org. Synth. 1973, 53, 289. See discussion re α- acetoxycholestanone. URL http://www.orgsyn.org/orgsyn/pdfs/CV6P0289.pdf

[382] Chamberlin, A. R.; Sheppeck, II, J. E.: p-Toluenesulfonylhydrazide. In Encyclopedia of reagents for organic synthesis,L.A.Paquette,ed.,Wi- ley, New York 1995. URL http://dx.doi.org/10.1002/047084289X.rt137

[383] Caglioti, L.: Reduction of ketones by use of the tosylhydrazone deriva- tives: androstan-17β-ol. Org. Synth. 1972, 52, 122. URL http://www.orgsyn.org/orgsyn/pdfs/CV6P0062.pdf

[384] Friedman, L.; Litle, R. L.; Reichle, W. R.: p-Toluenesulfonylhydrazide. Org. Synth. 1960, 40, 93. Also Coll. Vol. 5, p. 1055. See note 6 re recrys- tallisation. URL http://www.orgsyn.org/orgsyn/pdfs/CV5P1055.pdf

[385] Bandgar, B. P.; Sadavarte, V. S.; Uppalla, L. S.; Govande, R.: Chemose- lective preparation of oximes, semicarbazones, and tosylhydrazones with- out catalyst and solvent. Monatsh. Chem. 2001, 132, 403–406. URL http://dx.doi.org/10.1007/s007060170126 276 BIBLIOGRAPHY

[386] Ješelnik, M.; Varma, R. S.; Polanc, S.; Kočevar, M.: Solid-state syn- thesis of heterocyclic hydrazones using microwaves under catalyst-free conditions. Green Chem. 2002, 4, 35–38. URL http://dx.doi.org/10.1039/b108029f

[387] Hajipour, A. R.; Mallakpour, S. E.; Imanzadeh, G.: A Rapid and Con- venient Synthesis of Oximes in Dry Media under Microwave Irradiation. J. Chem. Res. 1999, pp. 228–229. URL http://dx.doi.org/10.1039/a806359a

[388] Bertz, S. H.; Dabbag, G.: Improved preparations of some arenesulfonyl- hydrazones. J. Org. Chem. 1983, 48, 116–119. URL http://dx.doi.org/10.1021/jo00149a022

[389] Iida, T.; Tamura, T.; Matsumoto, T.; Chang, F.: Improved Conditions for Preparation and Reductive Cleavage of Steroidal Ketone Tosylhydra- zones. Synthesis 1984, pp. 957–959. URL http://dx.doi.org/10.1055/s-1984-31036

[390] Jarikote, D. V.; Deshmukh, R. R.; Rajagopal, R.; Lahoti, R. J.; Daniel, T.; Srinivasan, K. V.: Ultrasound promoted facile synthesis of arylhy- drazones at ambient conditions. Ultrason. Sonochem. 2003, 10, 45–48. URL http://dx.doi.org/10.1016/S1350-4177(02)00100-1

[391] Vinczer, P.; Novak, L.; Szantay, C.: A mild method for the formation of tosylhydrazone derivatives from keto esters. Synth. Commun. 1984, 14, 281–288.

[392] Hertz, H. S.; Coxon, B.; Siedle, A. R.: Disproportionation and pyrolysis of p-toluenesulfonylhydrazine. J. Org. Chem. 1977, 42, 2508–2509. URL http://dx.doi.org/10.1021/jo00434a036

[393] Pouchert, C. J.; Behnke, J.: The Aldrich library of 13Cand1HFT-NMR spectra, vol. 2. Aldrich Chemical Co., Milwaukee 1993. See p. 1634A.

[394] El-Sayed, R. A.: Review on the chemistry of sulfonohydrazides and sul- fonoazides. Phosphorus, Sulfur Silicon Relat. Elem. 2004, 179, 237–266. Compare references 1-138 (pre-1972) with 139-158 (self citations). URL http://dx.doi.org/10.1080/10426500490274673

[395] Sutherland, A. G.: One or More CH Bond(s) Formed by Substitution: Reduction of C-Halogen and C-Chalcogen Bonds. In Comprehensive Or- ganic Functional Group Transformations, A. R. Katritzky; O. Meth- Cohn; C. W. Rees, eds., Pergamon, Oxford, vol. 1, pp. 1–26, 1st ed. 1995.

[396] Barton, D. H. R.; Ferreira, J. A.; Jaszberenyi, J. C.: Free radical de- oxygenation of thiocarbonyl derivatives of alcohols. In Preparative Car- bohydrate Chemistry, S. Hanessian, ed., Marcel Dekker, New York, pp. 151–172 1997. BIBLIOGRAPHY 277

[397] Salomon, C. J.; Danelon, G. O.; Mascaretti, O. A.: A Practical Method for the Disposal of Organotin Residues from Reaction Mixtures. J. Org. Chem. 2000, 65, 9220–9222. See also references therein. URL http://dx.doi.org/10.1021/jo000525+

[398] Barton, D. H. R.; Jang, D. O.; Jaszberenyi, J. C.: The invention of rad- ical reactions. 32. Radical deoxygenations, dehalogenations, and deami- nations with dialkyl phosphites and hypophosphorous acid as hydrogen sources. J. Org. Chem. 1993, 58, 6838–6842. URL http://dx.doi.org/10.1021/jo00076a054

[399] Takamatsu, S.; Katayama, S.; Hirose, N.; Naito, M.; Izawa, K.: Radi- cal deoxygenation and dehalogenation of nucleoside derivatives with hy- pophosphorous acid and dialkyl phosphites. Tetrahedron Lett. 2001, 42, 7605–7608. URL http://dx.doi.org/10.1016/S0040-4039(01)01617-3

[400] Rho, H.-S.; Ko, B.-S.: Regioselective deoxygenation of the cyclic thiono- carbonates of 2,3-dihydroxy esters with magnesium in methanol. Synth. Commun. 1999, 29, 2875–2880.

[401] Brown, A. C.; Carpino, L. A.: Magnesium in methanol: substitute for sodium amalgam in desulfonylation reactions. J. Org. Chem. 1985, 50, 1749–1750. URL http://dx.doi.org/10.1021/jo00210a035

[402] Fukukawa, K.; Ueda, T.; Hirano, T.: Nucleosides and nucleotides. VL. Facile deoxygenation of neplanocin A and nucleosides by the use of trib- utyltin hydride. Chem. Pharm. Bull. 1983, 31, 1842–1847.

[403] Patroni, J. J.; Stick, R. V.; Engelhardt, L. M.; White, A. H.: The deoxy- genation of some carbohydrate diols via the derived cyclic thiocarbonate. Aust. J. Chem. 1986, 39, 699–711. URL http://dx.doi.org/10.1071/CH9860699

[404] Rho, H.-S.: Tributyltin hydride-induced free radical deoxygenation of the cyclic thionocarbonates of threo-2,3-dihydroxy esters and ketones. Synth. Commun. 1997, 27, 3887–3893.

[405] Kanemitsu, K.; Tsuda, Y.; Haque, M. E.; Tsubono, K.; Kikuchi, T.: Re- action of cyclic thioxocarbonates with tributyltin hydride. Chem. Pharm. Bull. 1987, 35, 3874–3879.

[406] Tsuda, Y.; Kanemitsu, K.; Kakimoto, K.; Kikuchi, T.: Utilization of sugars in organic synthesis. XIX. Trialkyltin radical used to catalyze the O,S-rearrangement of cyclic thionocarbonates. A new entry to thio sugars. Chem. Pharm. Bull. 1987, 35, 2148–2150. 278 BIBLIOGRAPHY

[407] Patroni, J. J.; Stick, R. V.; Matthew, D.; Tilbrook, G.; Skelton, B. W.; White, A. H.: The synthesis and reactivity of cyclic thiocarbonates de- rived from some carbohydrate 1,2-diols. Aust. J. Chem. 1989, 42, 2127– 2141. URL http://dx.doi.org/10.1071/CH9892127

[408] Szammer, J.; Ötvös, L.: A convenient preparation of tri-n-butyltin hy- dride. Chem. Ind. 1988, p. 764.

[409] McAlonan, H.; Stevenson, P. J.: Convenient One-Pot Synthesis of Hexa- n-butylditin from Bis(tri-n-butyltin) Oxide. Organometallics 1995, 14, 4021–4022. URL http://dx.doi.org/10.1021/om00008a058

[410] Giese, B.; Gröninger, K. S.: 1,3,4,6-Tetra-O-acetyl-2-deoxy-α-D- glucopyranose. Org. Synth. 1990, 69, 66–71. Also Coll. Vol. 8, p. 583. URL http://www.orgsyn.org/orgsyn/pdfs/CV8P0583.pdf

[411] Berge, J. M.; Roberts, S. M.: Recommended work-up procedure for reductions employing tri-n-butyltin hydride. Synthesis 1979, pp. 471– 472. URL http://dx.doi.org/10.1055/s-1979-28726

[412] Farina, V.: A simple chromatographic technique for the purification of organic stannanes. J. Org. Chem. 1991, 56, 4985–4987. URL http://dx.doi.org/10.1021/jo00016a037

[413] Crich, D.; Sun, S.: A Practical Method for the Removal of Organotin Residues from Reaction Mixtures. J. Org. Chem. 1996, 61, 7200–7201. URL http://dx.doi.org/10.1021/jo960751c

[414] Rajan Babu, T. B. V.; Bulman Page, P. C.; Buckley, B. R.: Tri-n- butylstannane. In Encyclopedia of reagents for organic synthesis,L.A. Paquette,ed.,Wiley,NewYork1995. URL http://dx.doi.org/10.1002/047084289X.rt181

[415] Sano, H.: Bis(tri-n-butyltin) Oxide. In Encyclopedia of reagents for or- ganic synthesis,L.A.Paquette,ed.,Wiley,NewYork1995. URL http://dx.doi.org/10.1002/047084289X.rb195

[416] Jang, D. O.; Cho, D. H.; Barton, D. H. R.: Radical deoxygenation of alcohols via their S-methyl dithiocarbonate derivatives with di-n- butylphosphine oxide as hydrogen-atom donor. Synlett 1998, pp. 39–40. URL http://dx.doi.org/10.1055/s-1998-1582

[417] Jang, D. O.; Cho, D. H.: Radical deoxygenation of alcohols and vicinal diols with N-ethylpiperidine hypophosphite in water. Tetrahedron Lett. 2002, 43, 5921–5924. URL http://dx.doi.org/10.1016/S0040-4039(02)01247-9 BIBLIOGRAPHY 279

[418] Jang, D. O.; Song, S. H.: Facile synthesis of enantiopure (R)-malates. Tetrahedron Lett. 2000, 41, 247–249. URL http://dx.doi.org/10.1016/S0040-4039(99)02057-2 [419] Cho, D. H.; Song, S. H.; Jang, D. O.: A method for preparation of un- natural (R)-malic acid derivatives with phenylsilanes. Synth. Commun. 2003, 33, 515–519. URL http://dx.doi.org/10.1081/SCC-120015803 [420] Klein Gebbinck, E. A.; Jansen, B. J. M.; de Groot, A.: Insect antifeedant activity of clerodane diterpenes and related model compounds. Phyto- chemistry 2002, 61, 737–770. See table 6 (H. armigera). URL http://dx.doi.org/10.1016/S0031-9422(02)00174-7 [421] Enriz, R. D.; Baldoni, H. A.; Zamora, M. A.; Jáuregui, E. A.; Sosa, M. E.; Tonn, C. E.; Luco, J. M.; Gordaliza, M.: Structure-Antifeedant Activity Relationship of Clerodane Diterpenoids. Comparative Study with Withanolides and Azadirachtin. J. Agric. Food Chem. 2000, 48, 1384–1392. URL http://dx.doi.org/10.1021/jf990006b [422] Simmonds, M. S. J.; Blaney, W. M.; Esquivel, B.; Rodriguez-Hahn, L.: Effect of clerodane-type diterpenoids isolated from Salvia on the feeding behavior of Spodoptera littoralis. Pestic. Sci. 1996, 47, 17–23. URL http://dx.doi.org/10.1002/(SICI)1096-9063(199605)47:1<17:: AID-PS378>3.0.CO;2-I [423] Simmonds, M. S. J.; Blaney, W. M.; Fellows, L. E.: Behavioral and elec- trophysiological study of antifeedant mechanisms associated with poly- hydroxy alkaloids. J. Chem. Ecol. 1990, 16, 3167–3196. See p. 3171 (choice assay). URL http://dx.doi.org/10.1007/BF00979618 [424] Novac, O.; Guenier, A. S.; Pelletier, J.: Inhibitors of protein synthesis identified by a high throughput multiplexed translation screen. Nucleic Acids Res. 2004, 32, 902–915. See p. 904 under “dual-luciferase reporter assay.”. URL http://dx.doi.org/10.1093/nar/gkh235 [425] Reference Method for Broth Dilution Antifungal Susceptibility Testing of Yeasts; Approved Standard M27-A2. National Committee for Clinical Laboratory Standards, Wayne, PA, 2nd ed. 2002. [426] Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically; Approved Standard M7-A6.NationalCommitteefor Clinical Laboratory Standards, Wayne, PA, 6th ed. 2003. [427] Performance Standards for Antimicrobial Disk Susceptibility Tests; Ap- proved Standard M2-A8. National Committee for Clinical Laboratory Standards, Wayne, PA, 8th ed. 2003. 280 BIBLIOGRAPHY

[428] Photos by Andrea Bigham, University of Melbourne.

[429] Rovinsky, S. A.; Cizadlo, G. R.: Salvia divinorum Epling et Játiva-M. (Labiatae): An Ethnopharmacological Investigation. The McNair Schol- arly Review 1998, 3, 142–156. College of St Scholastica (Duluth, MN). Unofficial transcription available online. Accessed March 2006. URL http://www.sagewisdom.org/rovinsky.html

[430] Peterson, P. K.; Gekker, G.; Lokensgard, J. R.; Bidlack, J. M.; Chang, A. C.; Fang, X. G.; Portoghese, P. S.: κ-opioid receptor agonist suppres- sion of HIV-1 expression in CD4+ lymphocytes. Biochem. Pharmacol. 2001, 61, 1145–1151. See also references therein (other cell types). URL http://dx.doi.org/10.1016/S0006-2952(01)00574-3

[431] Lokensgard, J. R.; Gekker, G.; Peterson, P. K.: κ-opioid receptor ago- nist inhibition of HIV-1 envelope glycoprotein-mediated membrane fu- sion and CXCR4 expression on CD4+ lymphocytes. Biochem. Pharma- col. 2002, 63, 1037–1041. URL http://dx.doi.org/10.1016/S0006-2952(02)00875-4

[432] Reeves, J. D.; Piefer, A. J.: Emerging drug targets for antiretroviral therapy. Drugs 2005, 65, 1747–1766. URL http://www.ingentaconnect.com/content/adis/dgs/2005/ 00000065/00000013/art00002

[433] Gekker, G.; Lokensgard, J. R.; Peterson, P. K.: Naltrexone potentiates anti-HIV-1 activity of antiretroviral drugs in CD4+ lymphocyte cultures. Drug Alcohol Depend. 2001, 64, 257–263. URL http://dx.doi.org/10.1016/S0376-8716(01)00140-5

[434] Sheng, W. S.; Hu, S. X.; Lokensgard, J. R.; Peterson, P. K.: U50,488 inhibits HIV-1 Tat-induced monocyte chemoattractant protein-1 (CCL2) production by human astrocytes. Biochem. Pharmacol. 2003, 65, 9–14. URL http://dx.doi.org/10.1016/S0006-2952(02)01480-6

[435] Chao, C. C.; Hu, S.; Gekker, G.; Lokensgard, J. R.; Heyes, M. P.; Pe- terson, P. K.: U50,488 protection against HIV-1-related neurotoxicity: involvement of quinolinic acid suppression. Neuropharmacology 2000, 39, 150–160. URL http://dx.doi.org/10.1016/S0028-3908(99)00063-5

[436] O’Brien, M. E.; Clark, R. A.; Besch, C. L.; Myers, L.; Kissinger, P.: Patterns and correlates of discontinuation of the initial HAART regimen in an urban outpatient cohort. J. Acquir. Immune Defic. Syndr. 2003, 34, 407–414. URL http://www.jaids.org/pt/re/jaids/abstract. 00126334-200312010-00008.htm BIBLIOGRAPHY 281

[437] Blancquaert, J. P.; Lefebvre, R. A.; Willems, J. L.: Emetic and antiemetic effects of opioids in the dog. Eur. J. Pharmacol. 1986, 128, 143–150. URL http://dx.doi.org/10.1016/0014-2999(86)90760-0

[438] Kohl, R. L.; MacDonald, S.: New pharmacologic approaches to the pre- vention of space/motion sickness. J. Clin. Pharmacol. 1991, 31, 934–946. URL http://jcp.sagepub.com/cgi/content/abstract/31/10/934

[439] Lewin, S. R.; Vesanen, M.; Kostrikis, L.; Hurley, A.; Duran, M.; Zhang, L.; Ho, D. D.; Markowitz, M.: Use of real-time PCR and molecular beacons to detect virus replication in human immunodeficiency virus type 1-infected individuals on prolonged effective antiretroviral therapy. J. Virol. 1999, 73, 6099–6103. URL http://jvi.asm.org/cgi/reprint/73/7/6099

[440] Solomon, A.; Lane, N.; Wightman, F.; Gorry, P. R.; Lewin, S. R.: Enhanced replicative capacity and pathogenicity of HIV-1 isolated from individuals infected with drug-resistant virus and declining CD4+ T-cell counts. J. Acquir. Immune Defic. Syndr. 2005, 40, 140–148. URL http://www.jaids.com/pt/re/jaids/abstract. 00126334-200510010-00004.htm

[441] Project Report A694-38: Anti-HIV Activity and Cytotoxicity of One NI- AID Compound (9966) Tested in PBMC Cell-Based Assays.Southern Research Institute, Frederick, Maryland 2004. Available from the au- thor.

[442] Chao, C. C.; Gekker, G.; Hu, S.; Sheng, W. S.; Shark, K. B.; Bu, D. F.; Archer, S.; Bidlack, J. M.; Peterson, P. K.: κ opioid receptors in human microglia downregulate human immunodeficiency virus 1 expression. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 8051–8056. URL http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve& db=PubMed&dopt=Citation&list_uids=8755601

[443] Chao, C. C.; Gekker, G.; Hu, S. X.; Kravitz, F.; Peterson, P. K.: κ-opioid potentiation of tumor necrosis factor-α-induced anti-HIV-1 activity in acutely infected human brain cell cultures. Biochem. Pharmacol. 1998, 56, 397–404. URL http://dx.doi.org/10.1016/S0006-2952(98)00161-0

[444] Chou, J. Z.; Chait, B. T.; Wang, R.; Kreek, M. J.: Differential biotrans- formation of dynorphin A (1-17) and dynorphin A (1-13) peptides in human blood, ex vivo. Peptides 1996, 17, 983–990. URL http://dx.doi.org/10.1016/0196-9781(96)00154-4

[445] Childers, S. R.; Xiao, R.; Vogt, L.; Sim, L. J.: Lack of evidence of κ2- selective activation of G-proteins: κ opioid receptor stimulation of [35S] GTPγS binding in guinea pig brain. Biochem. Pharmacol. 1998, 56, 282 BIBLIOGRAPHY

113–120. URL http://dx.doi.org/10.1016/S0006-2952(98)00123-3

[446]Bosworth,D.T.:Salvia, AIDS, and the healing journey 2005. URL http://www.sagewisdom.org/aidshealing.html

[447] Holbeck, S. L.: Update on NCI in vitro drug screen utilities. Eur. J. Cancer 2004, 40, 785–793. URL http://dx.doi.org/10.1016/j.ejca.2003.11.022

[448] Johnson, J. I.; Decker, S.; Zaharevitz, D.; Rubinstein, L. V.; Venditti, J. M.; Schepartz, S.; Kalyandrug, S.; Christian, M.; Arbuck, S.; Holling- shead, M.; Sausville, E. A.: Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br. J. Cancer 2001, 84, 1424–1431. URL http://dx.doi.org/10.1054/bjoc.2001.1796

[449] US National Cancer Institute, Developmental Therapeutics Program Data Search. Salvinorin A = NSC 604585. URL http://dtp.nci.nih.gov/docs/dtp_search.html

[450] In Vivo Cancer Models 1976-1982. National Institutes of Health, Washington, DC 1984. NIH Publication # 84-2635. See p. 27 (P388 protocol). URL http://dtp.nci.nih.gov/branches/btb/pdf/In_Vivo_Cancer_ Models.pdf

[451] Schultz, R. J. Drug Synthesis and Chemistry Branch, National Cancer Institute. Email, 5th March 2006.

[452] Roth, B. L. R. Email, 23rd August 2005.

[453] Yan, F.; Mosier, P. D.; Westkaemper, R. B.; Stewart, J.; Zjawiony, J. K.; Vortherms, T. A.; Sheffler, D. J.; Roth, B. L.: Identification of the Molecular Mechanisms by Which the Diterpenoid Salvinorin A Binds to κ-Opioid Receptors. Biochemistry 2005, 44, 8643–8651. URL http://dx.doi.org/10.1021/bi050490d

[454] Rovner, S. L.: Chemical ‘Naming’ Method Unveiled. Chem.Eng.News 2005, 83, 39–40. (accessed Jan. 2006). URL http://pubs.acs.org/isubscribe/journals/cen/83/i34/html/ 8334sci1.html

[455] IUPAC International Chemical Identifier (accessed Oct. 2005). URL http://www.iupac.org/inchi

[456] Hoye, T. R.; Hanson, P. R.; Vyvyan, J. R.: A Practical Guide to First- Order Multiplet Analysis in 1H NMR Spectroscopy. J. Org. Chem. 1994, 59, 4096–4103. URL http://dx.doi.org/10.1021/jo00094a018 BIBLIOGRAPHY 283

[457] Hoye, T. R.; Zhao, H.: A Method for Easily Determining Coupling Con- stant Values: An Addendum to “A Practical Guide to First-Order Multi- plet Analysis in 1H NMR Spectroscopy”. J. Org. Chem. 2002, 67, 4014– 4016. URL http://dx.doi.org/10.1021/jo001139v

[458] Gottlieb, H. E.; Kotlyar, V.; Nudelman, A.: NMR Chemical Shifts of Common Laboratory Solvents as Trace Impurities. J. Org. Chem. 1997, 62, 7512–7515. URL http://dx.doi.org/10.1021/jo971176v

[459] Glennon, R. A.; Lee, M.; Rangisetty, J. B.; Dukat, M.; Roth, B. L.; Savage, J. E.; McBride, A.; Rauser, L.; Hufeisen, S.; Lee, D. K.: 2- Substituted tryptamines: agents with selectivity for 5-HT6 serotonin receptors. J. Med. Chem. 2000, 43, 1011–1018. URL http://dx.doi.org/10.1021/jm990550b

[460] Pauwels, P. J.; Colpaert, F. C.: Heterogeneous ligand-mediated Ca++ responses at wt and mutant α2A-adrenoceptors suggest multiple ligand activation binding sites at the α2A-adrenoceptor. Neuropharmacology 2000, 39, 2101–2111. URL http://dx.doi.org/10.1016/S0028-3908(00)00040-X

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: MUNRO, THOMAS ANTHONY

Title: The chemistry of Salvia divinorum

Date: 2006-04

Citation: Munro, T. A. (2006). The chemistry of Salvia divinorum. PhD thesis, Science: School of Chemistry, University of Melbourne.

Publication Status: Unpublished

Persistent Link: http://hdl.handle.net/11343/39135

File Description: The chemistry of Salvia divinorum

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.